Handbook of Electroporation 9783319267791

Citation preview

Treatment Planning for Electrochemotherapy and Irreversible Electroporation of Deep-Seated Tumors Bor Kos

Abstract

Electrochemotherapy and irreversible electroporation can be used to treat deepseated tumors. Key to treatment success is ensuring that the entirety of the target tumor is covered with electric fields of sufficient strength during the treatment. Treatment planning using numerical methods has long been established in radiotherapy, and this chapter presents the necessary tools to realize a similar treatment planning framework also for electrochemotherapy and irreversible electroporation. Treatment planning consists of identifying the target tumor and surrounding tissues on tomographic medical images. This reconstruction can be used to build a numerical model of the region of interest with by assigning correct conductivities to each tissue in the treatment zone. Using the finite element method, electric fields for a given electrode configuration can be determined. By coupling the numerical model of electroporation with appropriate optimization algorithms, the voltage to be delivered to each electrode pair can be determined and the positions of electrodes can be adjusted to ensure successful coverage of the target volume. Possible approaches to execute the treatment according to the prepared treatment plan are also discussed at the end of the chapter. Keywords

Irreversible electroporation • Electrochemotherapy • Deep-seated tumors • Treatment planning • Radiotherapy • Optimization

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Electric Field in Tissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Numerical Methods for Computing Electric Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

B. Kos (*) Faculty of Electrical Engineering, University of Ljubljana, Ljubljana, Slovenia e-mail: [email protected] # Springer International Publishing AG 2017 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_2-1

1

2

B. Kos

Predicting Response to Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Predicting Response Based on Electric Field Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modeling Probability of Cell Survival . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Optimization and Treatment Plan Execution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8 9 11 13 15 16 16

Introduction Electroporation is a physical mechanism in which externally applied electric field pulses cause changes in the cellular membrane, which increase the membrane’s permeability. Depending on the strength, duration, and number of the applied electric field pulses, these changes in the cellular membrane can either be reversible or irreversible, with irreversible electroporation requiring a higher electric field strength, a greater number of pulses, or both. Both principal types of effect can be used to treat deep-seated tumors in various organs with good success rates, more so, when we consider that many of these studies treated patients with no further therapeutic options: liver (Thomson et al. 2011; Edhemovic et al. 2014), kidney (Thomson et al. 2011; Narayanan and Doshi 2016), pancreas (Martin 2013), prostate (Onik et al. 2007), and metastases in bones (Gasbarrini et al. 2015; Bianchi et al. 2016). Reversible electroporation is used in conjunction with certain chemotherapeutic drugs (bleomycin and cisplatin), which greatly potentiates their antitumor effectiveness and has a number of other beneficial effects; this kind of treatment is termed electrochemotherapy (ECT) (Mir et al. 1991; Miklavčič et al. 2014). Irreversible electroporation (IRE), on the other hand, can be used to directly ablate tissue, and directly causes either apoptotic or necrotic cell death; irreversible electroporation and nonthermal irreversible electroporation ablation are terms used in the literature (Davalos et al. 2005; Jiang et al. 2015). For a treatment of a deep-seated tumor to be successful, it is necessary to ensure that the whole tumor is covered with a sufficiently strong electric field, exceeding the reversible or irreversible electroporation threshold, depending on the type of treatment that is being performed, i.e., electrochemotherapy or tissue ablation by irreversible electroporation, respectively. For deep-seated tumors, the electric field pulses are delivered to the target area by needle electrodes, which can be either in a fixed array or as individual electrodes. These electrodes need to be introduced in the immediate vicinity of the tumor or inside the tumor, because the strongest intensity of electric fields is found at the surface of the electrodes. Further away from the electrode surface, the electric field decreases rapidly, but the speed of this decrease is dependent on many factors, such as the shape of the electrode, diameter of the electrode needle, and on the proximity of the counter electrode. If we consider an imaginary point and move it from the surface of one electrode towards the surface of the second electrode, the electric field will first decrease to a certain point and then start increasing again. The values of electric field strength along this path are

Treatment Planning for Electrochemotherapy and Irreversible Electroporation. . .

3

dependent on the path itself, the distance between electrodes, the shape of the electrodes, the applied voltage, and the electrical properties of the tissue(s) in which the electrodes are inserted. Additionally, given the limitations of available electroporation pulse generators in terms of maximum voltage and maximum current also typically requires the use of more than one electrode pair, which further complicates predicting the response of the treatment as a whole. As can be inferred from the number of variables affecting the electric field distribution, the distribution of electric fields in tissue is not necessarily intuitive and easy to determine, therefore it is advisable to use some kind of treatment planning to determine the best way to do a procedure of this kind. Treatment planning is a term that originated in radiotherapy. With the spread of 3D imaging modalities such as ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI) in 1980s, these imaging modalities became ever more commonly used to plan therapeutic radiation therapy (Jaffray et al. 2007). In this approach, medical imaging is first performed on a patient to determine the anatomy of the patient, the location of the tumor to be treated, and its surrounding regions. This allows the physician in charge of the treatment to delineate the clinical target volume (CTV), that is, the gross tumor volume (GTV) surrounded by a clinically relevant safety margin. This safety margin is determined based on the histology and aggressiveness of the tumor. Surrounding critical structures are also defined and then used in the treatment planning step to design the irradiation pattern in such a way that a sufficient dose is delivered to the CTV while limiting the irradiation of the surrounding healthy tissue and especially to any critical tissues present. The treatment is then delivered in multiple fraction over several weeks, while controlling the patient positioning and in image-guided radiation therapy also following the positioning of the CTV itself due to breathing, tissue deformation, etc. (Jaffray et al. 2007; Bujold et al. 2012). While dedicated off-the-shelf treatment planning software for radiotherapy planning is available on the market with appropriate certification as medical devices, this is unfortunately not yet the case for electrochemotherapy and irreversible electroporation. A proof-of-concept tool Visifield has been developed at the University of Ljubljana (www.visifield.com, University of Ljubljana, Slovenia) (Marčan et al. 2015), but this cannot yet be used in the clinical setting. The aim of this chapter is therefore to present the methods and procedures that can be used to make individualized treatment plans for ECT or IRE.

Electric Field in Tissue In this subsection, the effects that different electrode geometry has on electric field distribution in the tissue will be illustrated. Treatment planning in electroporationbased treatments can follow a similar paradigm as for radiation therapy, with some steps modified to take into account the different physics involved in electroporation (Pavliha et al. 2012). One of the essential in-between steps between imaging and computation of electric fields in tissue is image segmentation into different tissues. This allows us to take into account the markedly different electrical properties that

4

B. Kos

different tissues have, and at the same time also serves as the delineation of the GTV for the treatment planning process. Plate electrodes were the most used electrode types in the initial studies of electrochemotherapy. They have a simple geometry, where the tissue to be treated is compressed between two parallel plates. This makes the setup very repeatable, and the electric field delivered to the target tissue is roughly approximated as the ratio V/d, where V is delivered voltage and d is the distance between the internal faces of the electrode. However, this simplification is only true in homogeneous tissues, when the distance between the electrodes is small in comparison to their width and length and if the conductivity of tissue remains constant irrespective of the applied electric fields. The last issue with plate electrodes is that they can only be used on relatively small target areas and cannot be inserted into tissue. Further increasing the complexity of planning ECT and IRE treatments of deepseated tumors is the fact that different tissues in the human body have markedly different electrical properties. Examples of tissues with the lowest conductivity include the outer layer of the skin (stratum corneum), bone, bone marrow, and fat tissues, while tissues with the highest conductivity include various glands, small intestines, and muscle (Gabriel et al. 1996). Tumor tissues generally have a higher conductivity than surrounding tissues as well (O’Rourke et al. 2007; Haemmerich et al. 2009; Peyman et al. 2015). Additionally, tissue conductivity also increases as a function of electric fields, because electroporation causes additional conductive pathways to form through otherwise insulating cell membranes. This is shown in Fig. 1, where the conductivity of liver and tumor tissue is presented as a function of electric field (Corovic et al. 2013). For a situation with plate electrodes and a nonhomogeneous tissue, the electric field strength is dependent on the thickness and the conductivity of each different tissue, but generally the tissue with higher conductivity “sees” a lower electric field than the tissue with a lower conductivity. The exact electric field in these nonhomogeneous tissues quickly gets impossible to calculate analytically, so computational methods have to be used. Figure 2 shows an illustration of the influence of tissue conductivity and electrode shape on the electric field in tissue. For better clarity, only electric fields above 400 V/cm are shown. An illustration of electric fields between plate and needle electrodes is shown. The electrodes are 1 cm apart, and 1000 V is applied between them. The upper row (panels A and B) shows the electric field between plate electrodes. For example with homogeneous tissue, the electric field is the exact value of voltage-to-distance ratio only in the point exactly between the electrodes (coordinates 0,0 on panel A). This is clearly not the case in panel B, where the tissue to the left of the slanted line has lower conductivity by a factor of 3 than the tissue to the right of the dividing line. This conductivity ratio is similar to the ratio between, e.g., liver and muscle (Gabriel et al. 2009); however the difference between tumors and surrounding tissue can be even higher than 5 (Haemmerich et al. 2009), which amplifies this effect even further. The panels in the bottom row (C and D) show the electric field between needle electrodes. The difference between needle electrodes and plate electrodes is that needle electrodes have an even more inhomogeneous

Treatment Planning for Electrochemotherapy and Irreversible Electroporation. . .

5

Fig. 1 The effect of electroporation on tissue conductivity. The initial conductivity of liver and liver tumors is based on the measurements of ex-vivo human liver and liver tumors (Haemmerich et al. 2009). At sufficiently strong electric fields, the conductivity starts to increase and the final factor of conductivity increase in this graph is 3.5 for liver and 3 for tumor (Corovic et al. 2013)

electric field distribution, and the electric fields exactly between electrodes is actually lower than the voltage-to-distance ratio. This effect is even more pronounced in the case with nonhomogeneous tissue (panel D). In tissues, this kind of conductivity change is expected on any interface between two different tissues, such as liver tumor, kidney tumor, or muscle tumor. While the conductivity transition is immediate in Fig. 2, in actual tissue, it would be more gradual across a millimeter or two, therefore the electric field would have a more gradual transition, but it would still be present. Tissue inhomogeneity thus renders electric field in tissue very difficult to predict without computer simulations.

Numerical Methods for Computing Electric Fields Previously the reasons for using numerical methods when planning any kind of electroporation treatment were demonstrated. One of the numerical methods, which has gained the most traction in electroporation research, is finite element method (Sel et al. 2007; Županič and Miklavčič 2010; Zupanic et al. 2012). The finite element method is based on discretizing the model geometry into a mesh of finite elements, on which the unknown quantity (in the case of treatment planning for

Fig. 2 The influence of electrode shape and tissue inhomogeneity on electric field distribution. The figure shows: (a) plate electrodes inside homogeneous tissue; (b) plate electrodes inside a nonhomogeneous tissue; (c) needle electrodes in homogeneous tissue; and (d) needle electrodes in nonhomogeneous tissue. In examples (b) and (d), the following relation holds for conductivities: σ2=3
σ1

6 B. Kos

Treatment Planning for Electrochemotherapy and Irreversible Electroporation. . .

7

electroporation this is the electric potential) is presumed to change according to some simple function (usually a polynomial). The solution of the whole problem is then translated into solving a set of algebraic equations for steady state problems. The computation of the electric field distribution during an electroporation pulse is an example problem, which can be simplified by assuming that it is a steady state problem. This simplification of electric field pulses to a static electrical conductive media problem can be justified as follows. Most electroporation pulses for electrochemotherapy and irreversible electroporation use 100 μs long pulses. These pulses are long compared to the cell membrane charging constant, which means that the induced transmembrane voltage will reach its maximum value long before the end of the pulse (Kotnik et al. 2012). Therefore, it is possible to assume that all the transient phenomena of electroporation occur in the beginning of the pulse and have settled out before the end of the pulse. With this simplification, it is possible to greatly shorten simulation times for electroporation models (Ivorra 2010). The equation to solve then becomes the Laplace equation for electric potential: ∇
ðσ∇V Þ ¼ 0,

(1)

where σ is tissue electric conductivity, ∇ is the gradient operator, and V is the electric potential. In order to take into account the increase in conductivity due to electroporation, the conductivity σ should also be a function of the electric field, which is defined as: E ¼ ∇V:

(2)

The following text will assume that Comsol Multiphysics (COMSOL AB, Stockholm, Sweden) is used for the finite element modeling since it has a relatively accessible user interface, and academic discounts are available for the software, but the same equations and equivalent solvers can also be implemented for example in free open-source finite element method packages FEniCS (https://fenicsproject.org/) (Alnæs et al. 2015) and FreeFem++ (http://www.freefem.org/) (Hecht 2012). The Comsol Multiphysics AC-DC model provides a physics interface for modeling the electric fields and currents in tissue, when pulses are applied through electrodes. The most appropriate mode for modeling electroporation is called electric currents. The first important choice that needs to be made is whether to model in three dimensions or to use a two-dimensional approximation. The former is much more computationally intensive and requires more computer resources, but gives a more representative result in the whole computational domain. On the other hand, a two-dimensional approximation is much faster to compute and less likely to run into issues with computational stability and therefore also allows the user to easily make more experiments. The initial step in numerical treatment planning is the creation of a representative patient geometry. The easiest approach is to use a geometrical representation using a sphere or spheroid to represent the target tumor. This can be done directly in the

8

B. Kos

graphical user interface of the Comsol Multiphysics software. The next step is the inclusion of electrodes – for deep-seated tumors these are typically needles of some kind, which can be adequately modeled by a cylinder. The tissue surrounding electrodes can be simply defined as a box or cylinder with sufficiently large dimensions. This kind of schematic geometry can serve as an initial step or a learning tool for treatment planning, but is still capable of giving valuable results (Garcia et al. 2011, 2014; Denzi et al. 2015). Once the geometry is defined, voltage can be applied to the electrode parts of the geometry via boundary conditions. The most basic way to define that is the Dirichlet boundary condition, which is essentially a fixed potential of the boundary of the electrode. In Comsol Multiphysics AC/DC module, this can be defined as an electric potential boundary condition, which allows the user to specify the fixed voltage at the electrode boundary. An even better option is using the Terminal boundary condition, which also provides a user-friendly interface for the computation of the total current delivered during the pulse. The interface for including electric field-dependent conductivity can be found by defining a smoothed step function (smoothed Heaviside function) in the material properties of the tissue (Garcia et al. 2014). In order to realize a more detailed patient-specific geometry some more steps need to be performed. One approach entails using the segmented medical images to extract the surface of different tissues and combining them with a detailed model of the electrodes. This approach, however, can quickly run into difficulties with generating the mesh for finite element method when attempting to insert the electrodes into the detailed geometry model with complicated geometry. Specialized software tools exist, which can help with that, but the process still involves a considerable amount of manual work. Another approach represents using Comsol Multiphysics in conjunction with Matlab (Mathworks, Natick, MA). This allows the generation of the mesh by taking into account as a solid geometry only the electrodes and a suitably large bounding box. The conductivities of different tissues can then be assigned directly to the free mesh elements by using an interpolation function for conductivity in Comsol Multiphysics (Aström et al. 2009). This method was used to generate the example figures in the continuation of the chapter.

Predicting Response to Treatment The computation of electric field is the first prerequisite of treatment planning which then needs to be evaluated in terms of expected biological outcome. In the first studies of ECT, electric field thresholds have been determined for trains of eight 100 μs pulses, delivered at 1 Hz pulse repetition frequency (Mir et al. 1991). This kind of pulse protocol has also been the most widely used electroporation protocol in clinical practice, but different number of pulses can also achieve the same effect, albeit at different electric field thresholds (Pucihar et al. 2011). For IRE, a higher number of pulses is typically used (Jiang et al. 2015), which has the benefit of

Treatment Planning for Electrochemotherapy and Irreversible Electroporation. . .

9

reducing the electric field strength needed to achieve ablation, but also increases the thermal load of the treatment and increases the treatment duration.

Predicting Response Based on Electric Field Thresholds The total volume of tissue that can be covered with electric field above the electroporation threshold with a single pair of electrodes is limited by the maximum voltage and current that a pulse generator can deliver. Therefore, it is necessary to develop a method of accounting for successivle coverage from more pairs of electrodes. The number of possible electrode pairs is given by Eq. 3, where n is the number of electrodes: Possible pairs ¼

nðn  1Þ : 2

(3)

Currently, the two most widely available pulse generators for ECT and IRE have connections for six individual electrodes, which yield a total of 15 possible pairs. Depending on the positioning of the electrodes, some of the possible pairs might not make sense, such as when given two electrodes are too far apart to produce therapeutically effective electric fields between them. The most conservative method for evaluating cumulative coverage of multiple electrode pairs is considering each electrode pair and the pulses delivered to it as a separate instance, by defining the volume of tissue above the electroporation threshold appropriate for the application as completely treated. The coverage of each successive electrode pair can then be superimposed and the total coverage determined from this superposition. This approach is illustrated on Fig. 3. Another reason that this approach is conservative is that it slightly underestimates the volume of treated tissue, since the conductivity of tissue is increased due to delivery of pulses to one electrode pair and the increase in conductivity persists for a short duration after the pulse delivery switches to the next electrode pair (Ivorra et al. 2009). More insight into the coverage of the target tissue with electric field can be determined by an approach analogous to dose-volume histograms, which are typically used in radiotherapy treatment planning (Zupanic et al. 2012). This kind of graph is shown in Fig. 4 and gives information what volume fraction of the GTV is covered by electric fields of at least the strength indicated on the x-axis. Here the treatment equivalent electric field after the application of pulses to n-th electrode pair (from the total p pairs) is determined as follows:  EEQn ¼


max EEQ n1 , En ; n > 1 , En ; n¼1

(4)

where EEQ n1 is the treatment equivalent field from the electrode pairs 1 through n1. This approach is conservative in the sense that the whole CTV is being

10

B. Kos

Fig. 3 Illustration of the successive coverage of a larger volume by utilizing more electrode pairs. By delivering pulses between more than one electrode pairs, a larger CTV can be covered with sufficiently strong electric fields. Homogeneous tissue is presented in this figure for clarity, and applied voltage is 1000 V between each electrode pair. Area covered by each electrode pair is indicated by different colors; each treatment field represents electric field strength above 400 V/cm

subjected by more than just one train of pulses. And to ensure the complete coverage of the CTV, some overlap between volumes treated by each electrode pair is inevitable, as can be seen in Fig. 3. The consequence of this is that some parts of the CTV are in fact experiencing two, three, or even more pulse trains of comparable electric field strength. The effective electroporation thresholds would therefore be actually even lower, since the electroporation thresholds are reducing with increasing number of pulses (Pucihar et al. 2011). The overlap of coverage can also be seen in Fig. 4, where, e.g., the volume fraction of the tumor after the third electrode pair (between electrodes 2 and 5) is the same as after the second electrode pair at 400 V/cm. At higher electric field strengths, it can be seen that the third pulse sequence increases the volume fraction by about 0.1. Due to the positioning of the electrodes and pair selection, a larger jump in coverage can be seen again for pair number 4 (between electrodes 2 and 6). However, looking only at the integrative coverage of the CTV provides only a descriptive statistic of the total coverage, without indicating where the coverage is lacking. Another approach is to present the results of the model directly on the original patient medical images, which have been used to generate the patientspecific model. This kind of visualization is shown in Fig. 5, and is based on the same case as Fig. 4.

Treatment Planning for Electrochemotherapy and Irreversible Electroporation. . .

11

Fig. 4 Electric field histogram. The cumulative volume fraction which is covered by electric fields of at least the strength indicated on the x-axis is shown after each electrode pair application. The data of electric field coverage was taken from the treatment plan of a patient with colorectal carcinoma metastases in the liver treated with ECT (Edhemovic et al. 2014)

Modeling Probability of Cell Survival The approach based on electric field thresholds is a good approach based on its simplicity and ease of understanding; however it has drawbacks for the task of treatment planning. The first is that the electric field threshold presents a very sharp delineation between “treated” and “untreated” tissue – namely if 400 V/cm is considered as a threshold for reversible electroporation in tumor tissue, and parts of the tumor, where the electric field strength determined by treatment planning would be considered fully treated, while any parts that fall even a little below the threshold would be considered completely untreated. This kind of result is overly simplistic in the fact that actual cells which are exposed to either 399 V/cm or 401 V/cm have a very similar probability of survival. The possible solution to this drawback of the electric field threshold for predicting the tissue response to ECT or IRE is therefore statistical modeling of the cell survival. The cell survival model called Peleg-Fermi model based on the original author of the model used for predicting microbial inactivation after exposure to

12

B. Kos

Fig. 5 Electric fields in tissue after all electroporation pulses have been delivered. The maximum value of electric field is shown in all tissues. Since the electrodes are located in the same axial plane as the MRI scan, the electrode track can be seen in the image. The electrode used in the treatment plan has a diameter of 1.2 mm and a length of the active part of 4 cm

pulsed electric field has already been applied to IRE treatment planning (Garcia et al. 2014) in the following way (Fig. 6). A survival function is defined as follows: S¼

1   E  Ec ð nÞ 1 þ exp Að n Þ

(5)

where S is the probability of cell survival (ranging from 0 to 1), E is the local electric field determined by numerical modeling, Ec(n) is the critical electric field at which 50% of the cells are expected to die, and A(n) is the shape factor, which determines the width of transition zone between fully treated and untreated regions. A larger value of A results in a wider transition area and a lower value of A results in a narrower transition area. The benefit of using this approach is that both Ec and A are a function of the number of pulses delivered to the treated area, and are defined as follows: Ec ðnÞ ¼ Ec0
ek1 n ,

(6)

AðnÞ ¼ A0
ek2 n ,

(7)

Treatment Planning for Electrochemotherapy and Irreversible Electroporation. . .

13

Fig. 6 The landscape of the probability of survival according to the Peleg-Fermi model depending on the number of pulses and electric field strength (for values of the parameters: Ec0 = 750 V/cm, A0 = 15 V/cm, k1 = 0.015, k2 = 0.003). The figure shows that the transition region between fully electroporated and nonelectroporated tissue spans about 200 V/cm of electric field strength at 10 pulses; however at 90 pulses this transition region becomes narrower at 100 pulses

Ec0 and A0 are initial values for the parameter, which are theoretical values at 0 pulses, while the k parameters define the rate at which the values decrease with increasing number of pulses. At this point it is important to stress that the values of the four parameters need to be experimentally determined and validated for different tissues and that they have a limited interval in which they can be used. By increasing the number of pulses above a reasonable number, this equation has the possibility to return clearly erroneous results since the value of S when n approaches infinity is 0. With this limitation in mind however it is possible to apply this equation to the treatment planning problem successfully.

Optimization and Treatment Plan Execution When performing treatment planning, the goal is to generate technically feasible electrode positioning, which will ensure the complete coverage of the CTV, at the same time limit damage to potential critical structures in the vicinity of the CTV, all while maintaining compliance with the hardware limitations of the available pulse

14

B. Kos

generators. Due to these considerations, it is not always the best approach to increase the voltage to electrode pairs, since that can run the risk of exceeding the maximum current limitation of the pulse generator. Additionally, if the electrodes are too far apart, or if the contrast between conductivities of the GTV and the surrounding healthy tissue is too large, it can be impossible to achieve sufficiently strong electric fields to successfully treat the target. The use of optimization algorithms can simplify the search for appropriate electrode positions and voltages. The considerations above need to be formalized into a fitness function, which allows the optimization algorithm to compare candidate solutions. The fitness function has to be designed differently for ECT and IRE, since the two different approaches require different degrees of coverage. ECT requires the whole CTV to be covered with at least with electric fields above the reversible electroporation threshold. Fields above the irreversible electroporation threshold in the CTV are not required, but are not detrimental to the success of the treatment. For IRE however, the whole CTV needs to be covered with fields above the irreversible electroporation threshold. Additionally, for IRE, temperature rise is an important consideration, especially when electrodes are located near sensitive critical structures (Kos et al. 2015). Example fitness functions would be based on the approach described in (Zupanic et al. 2012): sur FECT ¼ 100 V CTV rev  10 V irr ,

(8)

sur FIRE ¼ 100 V CTV irr  10 V irr ,

(9)

In Eqs. 8 and 9, the values of VCTV indicate the volume fraction of the CTV, which is covered by fields above the reversible or irreversible electroporation threshold indicated by the subscripts rev and ire, respectively. The volume of surrounding tissue above the irreversible electroporation threshold is intended to minimize damage to surrounding critical structures. For treatment planning optimization, two general situations can be anticipated when performing optimization. In the first situation, the electrode positions are fixed by anatomical, surgical, or some other constraints and only the voltages can be varied. In the second situation, the optimization algorithm can have the full freedom of electrode positioning and voltages between each electrode pair, therefore the position and direction of each electrode can be varied, as well as the voltage between each electrode pair. The first situation is less complex, since the smaller number of parameters limits the possible number of solutions to a more manageable number. In this case, it is feasible to compute the solution of electric field between each electrode pair for most feasible voltages and then use the optimization algorithm to select the best combination of voltages, which achieve sufficient CTV coverage with electric field and stay within the constraints of the pulse generator. A gradient optimization algorithm can be used in this case. This requires the computation of the partial derivatives of the fitness function with respect to all parameters. The algorithm then changes the parameter, which results in the largest increase of the fitness function. In the next

Treatment Planning for Electrochemotherapy and Irreversible Electroporation. . .

15

iteration, the gradient needs to be determined again, so that the appropriate parameter can be adjusted, until the solution converges to a maximum. With a gradient algorithm it is not possible to guarantee that this maximum is not a small local maximum instead of a true global maximum. When the electrode positions also have to be determined by the optimization algorithm, the best approach is to use a genetic algorithm, which attempts to mimic natural selection (Zupanic et al. 2012). For optimization with a genetic algorithm, all parameters of the treatment need to be encoded in a “chromosome,” which is a numerical variable in the optimization algorithm, which contains the position and direction of all electrodes and the voltage delivered between each electrode pair. Initially a population of candidate solutions is generated at random and the fitness function is evaluated for each member of the population of solutions. A number of best performing solutions are selected to “breed” in order to generate the next generation. The chromosomes of the next population are generated by crossing over the chromosomes of random pairs from the breeding individuals. Finally, mutation is introduced, which is a random change of a single parameter, can occur after the cross-over phase and generate new values for parameters, which have not previously been present in the population. This process is repeated, until a solution which satisfies the criteria is found. This approach is computationally much more intensive, because with changing electrode positions, the whole treatment (all electrode pairs) needs to be computed for every new candidate solution. But the benefit of using the genetic algorithm is also that it is much less likely to get stuck in a local maximum in comparison to a gradient-based algorithm. Because the genetic algorithm changes the parameters randomly (but based on previous successful solutions), it can also find a successful solution faster than a gradient-based algorithm, which has to make small improvements on existing solutions.

Conclusion Treatment of deep-seated tumors with electrochemotherapy (ECT) or irreversible electroporation (IRE) is a challenging procedure, which requires the careful design of the procedure and good control of positioning of the electrodes and careful selection of voltage for each electrode pair. Analogously to radiotherapy, where a physical mechanism (ionizing radiation) is used to elicit a biological response, in electroporation a different physical mechanism (nonionizing pulsed electric fields) is used to elicit a biological response (uptake of cytotoxic drugs in electrochemotherapy, or apoptotic cell death in IRE). The physical laws which govern the distribution of electric fields in tissue are well understood and lend themselves to efficient numerical solutions. The numerical methods can be used in conjunction with tissue segmentation to predict the response to the treatment and to tailor the treatment parameters in a way that will maximize the probability of a successful treatment outcome. Optimization algorithms provide the means to automatically adjust voltages and electrode positions until a treatment plan with sufficient quality and robustness is generated. This can then be translated into practice by following

16

B. Kos

instructions in the treatment plan or by using some kind of navigational assistance device, which guides the performing physician in positioning the electrodes according to the treatment plan.

Cross-References ▶ Cliniporator: Medical Electroporation of Tumors ▶ Electric Field Distribution and Electroporation Threshold ▶ Electrochemotherapy and Its Clinical Applications ▶ Mathematical Models Describing Cell Death due to Electroporation ▶ Modeling Microbial Inactivation by Pulsed Electric Field ▶ Tissue Ablation by Irreversible Electroporation

References Alnæs M, Blechta J, Hake J, et al (2015) The FEniCS project version 1.5. doi: 10.11588/ ans.2015.100.20553 Aström M, Zrinzo LU, Tisch S et al (2009) Method for patient-specific finite element modeling and simulation of deep brain stimulation. Med Biol Eng Comput 47:21–28. doi:10.1007/s11517008-0411-2 Bianchi G, Campanacci L, Ronchetti M, Donati D (2016) Electrochemotherapy in the treatment of bone metastases: a phase II trial. World J Surg 40:3088–3094. doi:10.1007/s00268-016-3627-6 Bujold A, Craig T, Jaffray D, Dawson LA (2012) Image-guided radiotherapy: has it influenced patient outcomes? Semin Radiat Oncol 22:50–61. doi:10.1016/j.semradonc.2011.09.001 Corovic S, Lackovic I, Sustaric P et al (2013) Modeling of electric field distribution in tissues during electroporation. Biomed Eng Online 12:16. doi:10.1186/1475-925X-12-16 Davalos R, Mir L, Rubinsky B (2005) Tissue ablation with irreversible electroporation. Ann Biomed Eng 33:223–231. doi:10.1007/s10439-005-8981-8 Denzi A, Strigari L, Di Filippo F et al (2015) Modeling the positioning of single needle electrodes for the treatment of breast cancer in a clinical case. Biomed Eng Online 14(Suppl 3):S1. doi:10.1186/1475-925X-14-S3-S1 Edhemovic I, Brecelj E, Gasljevic G et al (2014) Intraoperative electrochemotherapy of colorectal liver metastases. J Surg Oncol. doi:10.1002/jso.23625 Gabriel C, Peyman A, Grant E (2009) Electrical conductivity of tissue at frequencies below 1 MHz. Phys Med Biol 54:4863–4878. doi:10.1088/0031-9155/54/16/002 Gabriel S, Lau R, Gabriel C (1996) The dielectric properties of biological tissues: II. measurements in the frequency range 10 Hz to 20 GHz. Phys Med Biol 41:2251–2269 Garcia PA, Davalos RV, Miklavcic D (2014) A numerical investigation of the electric and thermal cell kill distributions in electroporation-based therapies in tissue. PLoS One 9:e103083. doi:10.1371/journal.pone.0103083 Garcia PA, Rossmeisl JH Jr, Neal RE 2nd et al (2011) A parametric study delineating irreversible electroporation from thermal damage based on a minimally invasive intracranial procedure. Biomed Eng Online 10:34. doi:10.1186/1475-925X-10-34 Gasbarrini A, Campos WK, Campanacci L, Boriani S (2015) Electrochemotherapy to metastatic spinal melanoma: a novel treatment of spinal metastasis? Spine 40:E1340–E1346. doi:10.1097/ BRS.0000000000001125

Treatment Planning for Electrochemotherapy and Irreversible Electroporation. . .

17

Haemmerich D, Schutt D, Wright A et al (2009) Electrical conductivity measurement of excised human metastatic liver tumours before and after thermal ablation. Physiol Meas 30:459–466. doi:10.1088/0967-3334/30/5/003 Hecht F (2012) New development in freefem++. J Numer Math 20(3–4):251–256. doi:10.1515/ jnum-2012-0013 Ivorra A (2010) Tissue electroporation as a bioelectric phenomenon: basic concepts. In: Rubinsky B (ed) Irreversible Electroporation. Springer, Berlin/Heidelberg, pp 23–61 Ivorra A, Al-Sakere B, Rubinsky B, Mir L (2009) In vivo electrical conductivity measurements during and after tumor electroporation: conductivity changes reflect the treatment outcome. Phys Med Biol 54:5949–5963. doi:10.1088/0031-9155/54/19/019 Jaffray D, Kupelian P, Djemil T, Macklis RM (2007) Review of image-guided radiation therapy. Expert Rev Anticancer Ther 7:89–103. doi:10.1586/14737140.7.1.89 Jiang C, Davalos R, Bischof J (2015) A review of basic to clinical studies of irreversible electroporation therapy. IEEE Trans Biomed Eng 62:4–20. doi:10.1109/TBME.2014.2367543 Kos B, Voigt P, Miklavcic D, Moche M (2015) Careful treatment planning enables safe ablation of liver tumors adjacent to major blood vessels by percutaneous irreversible electroporation (IRE). Radiol Oncol 49:234–241. doi:10.1515/raon-2015-0031 Kotnik T, Kramar P, Pucihar G et al (2012) Cell membrane electroporation – part 1: the phenomenon. IEEE Electr Insul Mag 28:14–23 Marčan M, Pavliha D, Kos B et al (2015) Web-based tool for visualization of electric field distribution in deep-seated body structures and planning of electroporation-based treatments. Biomed Eng Online 14(Suppl 3):S4. doi:10.1186/1475-925X-14-S3-S4 Martin RCG (2013) Irreversible electroporation of locally advanced pancreatic head adenocarcinoma. J Gastrointest Surg 17:1850–1856. doi:10.1007/s11605-013-2309-z Miklavčič D, Mali B, Kos B et al (2014) Electrochemotherapy: from the drawing board into medical practice. Biomed Eng Online 13:29. doi:10.1186/1475-925X-13-29 Mir LM, Orlowski S, Belehradek J, Paoletti C (1991) Electrochemotherapy potentiation of antitumor effect of bleomycin by local electric pulses. Eur J Cancer 27:68–72 Narayanan G, Doshi MH (2016) Irreversible electroporation (IRE) in renal tumors. Curr Urol Rep 17:15. doi:10.1007/s11934-015-0571-1 Onik G, Mikus P, Rubinsky B (2007) Irreversible electroporation: implications for prostate ablation. Technol Cancer Res Treat 6:295–300 O’Rourke AP, Lazebnik M, Bertram JM et al (2007) Dielectric properties of human normal, malignant and cirrhotic liver tissue: in vivo and ex vivo measurements from 0.5 to 20 GHz using a precision open-ended coaxial probe. Phys Med Biol 52:4707–4719. doi:10.1088/00319155/52/15/022 Pavliha D, Kos B, Županič A et al (2012) Patient-specific treatment planning of electrochemotherapy: procedure design and possible pitfalls. Bioelectrochemistry 87:265–273. doi:10.1016/j.bioelechem.2012.01.007 Peyman A, Kos B, Djokić M et al (2015) Variation in dielectric properties due to pathological changes in human liver. Bioelectromagnetics 36:603–612. doi:10.1002/bem.21939 Pucihar G, Krmelj J, Reberšek M et al (2011) Equivalent pulse parameters for electroporation. IEEE Trans Biomed Eng 58:3279–3288. doi:10.1109/TBME.2011.2167232 Sel D, Lebar A, Miklavcic D (2007) Feasibility of employing model-based optimization of pulse amplitude and electrode distance for effective tumor electropermeabilization. IEEE Trans Biomed Eng 54:773–781. doi:10.1109/TBME.2006.889196 Thomson KR, Cheung W, Ellis SJ et al (2011) Investigation of the safety of irreversible electroporation in humans. J Vasc Interv Radiol 22:611–621. doi:10.1016/j.jvir.2010.12.014 Zupanic A, Kos B, Miklavcic D (2012) Treatment planning of electroporation-based medical interventions: electrochemotherapy, gene electrotransfer and irreversible electroporation. Phys Med Biol 57:5425–5440. doi:10.1088/0031-9155/57/17/5425 Županič A, Miklavčič D (2010) Optimization and numerical modeling in irreversible electroporation treatment planning. In: Rubinsky B (ed) Irreversible Electroporation. Springer, Berlin/ Heidelberg, pp 203–222

Different Approaches Used in Modeling of Cell Membrane Electroporation Clair Poignard, Aude Silve, and Lars Wegner

Abstract

Cell electroporation is a complex phenomenon, which consists in the emergence of defects in cell membranes subjected to electric pulses. Since the end of the 1990s, biophysical models have been developed to explain and predict the conditions for cell electroporation. However the recent biological data, in particular those dealing with the influence of the repetition rate of the pulses, challenge these biophysical models. In this chapter, different approaches to model electropore formation are presented. The simplest equivalent circuit model is first presented. Biophysical approaches, which are extensively presented in chapter “▶ Electropore Energy and Thermodynamics,” are rapidly overviewed. For each approach, advantages and disadvantages are also discussed, in terms of physical meaning and validation with the experimental data. Then phenomenological approach is introduced. Such approaches consist in designing the model on an empirical basis thanks to the experience. Even though the physical bases of such models are still lacking, they provide new interesting views on the electroporation processes, as described by the experiments. The aim of the chapter is to introduce the reader to different ways of modeling cell membrane electroporation and to provide some possible directions to obtain a more reliable theory of electroporation in accordance with the experiments and with a justified theoretical basis.

C. Poignard (*) Team MONC, INRIA Bordeaux-Sud-Ouest, Institut de Mathématiques de Bordeaux, CNRS UMR 5251 & Université de Bordeaux, Bordeaux, France e-mail: [email protected] A. Silve • L. Wegner Karlsruhe Institute of Technology, Institute for Pulse Power and Microwave Technology, Eggenstein-Leopoldshafen, Germany e-mail: [email protected]; [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_3-1

1

2

C. Poignard et al.

Keywords

Cell electroporation modeling • Biophysical and mathematical models for pore formation

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cell Electrical Modeling Before Electroporation: The Linear Regime . . . . . . . . . . . . . . . . . . . . . . . . . . Equivalent Circuit for Cell: The 0D Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3D Maxwell Equations and Their Simplication to the 3D Model . . . . . . . . . . . . . . . . . . . . . . . . . . . Spherical Cell in a Unidirectional Constant Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Basic Concepts in Biophysics for Pore Formation in Liposomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Single-Pore Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . From One Pore to the Total Pore Density Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pore Radii Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The KN Model for Electroporation Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Back to the Electric Equation and Drawbacks of the KN Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . Phenomenological Models for Membrane Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biological Evidences of Electroporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Two-Step Process of Membrane Electroporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Calibration of the Model with Patch-Clamp Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biophysical Basis of the Phenomenological and Link with Pore Radius Evolution . . . . . . . . The Role of the Surface Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Concluding Remarks and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 4 5 7 9 9 12 12 13 14 16 16 17 18 20 20 22 23 23

Introduction Eukaryotic cell is a complex biological entity, which is the main constituent of any biological tissues: it is somehow the base unit of any living organism. These cells are generically composed of cytoplasm, which includes the nucleus, mitochondria, and other organelles that are necessary for life. This cytoplasm is protected from the extracellular stress by the plasma membrane, which is a phospholipid bilayer. This barrier plays a double role of protecting the cell and controlling exchanges between the cytoplasm and the extracellular medium. In the 1970s, it has been observed that electric shock may change transiently the membrane, allowing the entrance of usually non-permeant molecules into the cytoplasm. This phenomenon, called electroporation or electropermeabilization, has then been studied for cancer treatments, by treating cells with cytototoxic drugs – such as bleomycin or cisplatin – while exposing them to pulsed electric fields. This treatment called electrochemotherapy (see the chapters by L.M Mir, Gehl and Serša, and also Cadossi et al.) is now used standardly in more than 40 cancer institutes in Europe for cutaneous tumors, and several clinical studies are driven for deep located tumors. Even though the bases of cell electroporation are well known, several experimental observations are still unexplained, and the modeling of the phenomenon suffers from a lack of accuracy. The aim of this chapter consists in presenting different

Different Approaches Used in Modeling of Cell Membrane Electroporation

σm Eext

Electrical description

Vm

Permeabilisation

εm

3

Ci Transport

ρm

Cout

Fig. 1 Schematic process of electroporation

approaches to model cell membrane electroporation and in discussing the pros and cons of each approach. Generally speaking, the process of electroporation can be schematized by Fig. 1: Exposure to applied external electrical field leads to an increase of the transmembrane voltage Vm, which induces an increase of membrane conductivity σ m, permittivity em, and permeability pm. Increased membrane conductivity and permittivity, in turn, affect membrane voltage til the end of the pulse. Membrane permeability is not intrinsic: it depends on the considered molecule and is detected thanks to transport of molecules (PI, DNA plasmids, etc.). The aim of the chapter is to focus on the electrical description of the cell membrane, without accounting for the transport process, which is still unclear. The chapter is organized into five sections. First, the linear electric model of Schwan et al. (Foster and Schwan 1989) is presented. The cell is composed of a conducting cytoplasm surrounded by a resistive thin layer. Equivalent circuit model is described and complexified by the generic model in which the thin membrane is accounted for by imposing equivalent transmission conditions across the interface between the cytoplasm and the outer medium. Electroporation phenomenon modeling is then addressed. In section “Basic Concepts in Biophysics for Pore Formation in Liposomes,” the biophysical models for electropore formation are briefly introduced, and the model of Krassowska and Neu referred to as the KN model, considered as the most advanced description of electroporation, is described. In section “Phenomenological Models for Membrane Conductivity,” phenomenological approaches which consist in designing a model thanks to the experimental observations are proposed. Phenomenological models avoid the drawbacks of the other models, in particular when considering the effect of pulse repetition. In conclusion, current challenges in cell electroporation modeling are given, in order to obtain models reliable and in complete accordance with the experiments.

Cell Electrical Modeling Before Electroporation: The Linear Regime In the Schwan model (Foster and Schwan 1989), the cell is composed of a homogeneous conducting cytoplasm, which is roughly tens of micrometers in diameter, surrounded by a few nanometers thick, insulating membrane (see Fig. 2).

4

C. Poignard et al.

a

b

Fig. 2 Electrical model of biological cell by Schwan et al. (Foster and Schwan 1989) and its equivalent electric circuit. The cytoplasm Oc is protected thanks to the thin membrane Om, whose thickness h is about a few nanometers. The cell is embedded in an extracellular medium denoted by Oe. The entire domain is denoted by Ω. a Schematic cell and dielectric parameters. b Equivalent electric cell model

Equivalent Circuit for Cell: The 0D Model The simplest way to model the cell is to derive an electric circuit model in which the cell cytoplasm is described by a resistivity Rc, the cell membrane is identified by a capacitor whose capacitance equals Cm, and the ambient medium is described by a resistivity Re as given by Fig. 2b. Kirchhoff’s circuit law writes then

Different Approaches Used in Modeling of Cell Membrane Electroporation

V cell ¼ V m þ Rc Cm

dVm : dt

5

(1)

If a static electric field of magnitude E is applied to the cell of radius rc, one infers 2r c E ¼ V m þ Rc Cm

dVm : dt

(2)

Assuming that Vm is zero at the initial time and that the electric field E is constant, then Vm is given by
 V m ðtÞ ¼ 2r c E 1  et=τ , with τ ¼ Rc Cm :

(3)

Interestingly, this very simple model exhibits a linear dependency of the transmembrane voltage on the radius. Such a linear dependency is intrinsic, and it will be recovered in the 3D model, as explained below. However, this equivalent model is too rough, and in particular it cannot describe the influence of the cell shape or the effect of the direction of the electric field on the membrane voltage. Therefore, it is necessary to use partial differential equations (PDEs), which describe the electric field in the whole cell.

3D Maxwell Equations and Their Simplication to the 3D Model Maxwell equations and the standard constitutive laws for a dielectric material with permittivity e and conductivity σ read curlH ¼ e@t E þ σE, curlE ¼ curlH:

(4)

This 3D vector system is quite complex to solve, especially for high contrast material as biological cells. However, neglecting the time variation of the magnetic field leads to a curl-free electric field E, which then implies that E derives from a scalar potential E ¼ ∇V. Then, taking the divergence of the left equation leads to the following PDE on V called the electroquasistatic formulation: ∇  ðe∇@t V Þ þ ∇  ðσ∇V Þ ¼ 0;

(5)

to which initial and boundary conditions are imposed, for instance, V j@Ω ¼ g, V jt¼0 ¼ V0:

(6)

For the cell cytoplasm and the extracellular medium, the ratio e/σ is about 109s (see the parameters of Fig. 2a), meaning that up to several megahertz, the displacement currents can be neglected. However, due to the high resistivity of the membrane, these currents have to be accounted for in the thin layer (Poignard 2009); hence V is the continuous solution to

6

C. Poignard et al.

  σ@n V Γ þ , Γh

∇  ðσ∇V Þ ¼ 0, in Ohc [ Ohm [ Oe ;     ¼ em @t @n V Γ  , Γþh þ σ m @n V Γ  , Γþh , on Γ and Γ h respectively; V j@Ω ¼ g, V jt¼0 ¼ V 0 ;

(7) (8) (9)

where Γ h and Γ are the respective outer and inner boundaries of the cell membrane, and the normal vectors are taken from the inner to the outer part of the cell. Even though this equation is a rough simplification of the Maxwell vector equations, ∇V describes quite precisely the electric field at low frequency, and thus it is widely used in the electrical bioengineering community. However, due to the high resistivity and the small thickness of the membrane, it is still complex to solve accurately the above equation on V. To perform computations on realistic cell shapes without meshing the cell membrane, Pucihar et al. (2006) propose to replace the membrane by an equivalent condition on the boundary of the cytoplasm (see Fig. 3). Denoting by S0 the surface conductance and by Cm the capacitance of the membrane defined as Cm ¼ em =h, S0 ¼ σ m =h;

(10)

and denoting by Oc the whole cell Oc ¼ Ohc [ Om ;

(11)

the electric potential V is approached by U, the solution to ΔU ¼ 0, in Oc [ Oe ;

(12a)

σ e @n U jΓ þ ¼ σ c @n U jΓ  ;

(12b)

Cm @t V m þ S0 V m ¼ σ c @n U jΓ  , where V m ¼ U jΓþ  U jΓ ;

(12c)

U j@Ω ¼ g, Ujt¼0 ¼ V 0 :

(12d)

It is worth noting that unlike V, the approximate potential U is discontinuous across the interface: this is the effect of the high resistivity and the small thickness of the membrane. Equation (12c) corresponds to a contact resistance model, rigorously justified by Poignard and Perrussel (Perrussel and Poignard 2013). It is a generalization of (2), accounting the cell geometry and the electric field orientation thanks to the Neumann derivative.

Different Approaches Used in Modeling of Cell Membrane Electroporation

7

Fig. 3 Cell model with a zero-thickness membrane. The influence of the membrane is described through its capacitance Cm and its surface conductance S0

Spherical Cell in a Unidirectional Constant Field For simple shapes and simple electric field, it is possible to compute analytically the transmembrane voltage. For instance, let assume that the cell is a sphere of radius rc embedded in the whole space ℝ3 and submitted to the unidirectional electric field E ¼ EðtÞez . Using polar coordinates x ¼ r cos θ sin φ, y ¼ r sin θ sin φ, z ¼ r cos φ – the Kirchhoff law (12c) reads  Cm @t V m þ S0 þ

 2σ e σ c 3σ e σ c EðtÞ cos φ: Vm ¼ r c ð2σ e þ σ c Þ 2σ e þ σ c

(13)

Therefore, for a time-constant electric field, one obtains the explicit expression of Vm: 3 V m ðt, φÞ ¼ r c E 2


 1 1  et=τm cos φ, σ c þ 2σ e r c S0 1þ 2σ c σ e

with

τm ¼

rc Cm : 2σ c σ e þ r c S0 σ c þ 2σ e

(14) In the linear regime, the membrane conductance is of order 1, and since the cell radius is of the order tens of microns, one has 1 r c Cm ðσ c þ 2σ e Þ  1, and τm  : σ c þ 2σ e 2σ c σ e 1þ r c S0 2σ c σ e

(15)

Thus one recovers the linear dependency of the membrane voltage with respect to the cell radius but with a different constant compared with the equivalent circuit model.

8

TMV

C. Poignard et al.

Depolarization

Hyperpolarization

Fig. 4 Left: Schematic angular variation of the transmembrane voltage according to formula (14). Right: Fluorescent measurements and relative fluorescent with respect to the angle (From Poignard and Silve 2014)

One interesting feature of formula (14) lies in the fact that the membrane conductance has to increase dramatically to impact the membrane voltage. Roughly, for values of S0 increasing from 1 to 104 S/m2, the changes on the potential is only about 10%. Therefore, direct measurements of the membrane potential cannot detect any increase of membrane conductance below 1,000 S/m2. Another interesting feature is that the constant time τm is bounded by below by r c Cm ðσ c þ 2σ e Þ=ð2σ c σ e Þ , which is about several tenth of microseconds for cells, and thus the membrane voltage stabilizes within a few microseconds. Formula (14) predicts the sinusoidal behavior of the transmembrane voltage in the linear regime as observed by the experimental data (Fig. 4). Remark 1 (On the Schwann equation and short pulses) As mention above, for square pulses longer than the intrinsic membrane charging time τm, equality (14) leads to the so-called Schwann equation, which gives the transmembrane voltage value at the end of the pulse: V m  t  τm ð3=2Þr c E cos ϕ,   since 1  et=τm rm  1. It is worth noting that such formula does not hold for short pulses, which lasts a few nanoseconds. For such pulses, the membrane voltage is approximated by 3 t 3σ e σ c t MðEÞ cos φ V m ðt, φÞ  t  τm r c MðEÞ cos φ  2 τm σ c þ 2σ e Cm

(16)

where M(E) is the mean value of E along the pulse: MðEÞ ¼

1 t

Z

t 0

EðsÞds:

(17)

Different Approaches Used in Modeling of Cell Membrane Electroporation

9

The important fact is that the transmembrane voltage for short pulses does not depend on the cell radius but on the time of pulse application, and the mean value of the field amplitude is involved.

Basic Concepts in Biophysics for Pore Formation in Liposomes The creation and growth of single pores in vesicles have been studied for many years (Kroeger et al. 2009; Ryham et al. 2011; Sandre et al. 1999; Weaver and Chimazdzhev 1996). At the end of the 1990s, Sandre et al. (1999) have studied the pore creation in vesicles embedded in a highly viscous fluid. The high viscosity of the ambient medium made possible to visualize pore formation in stretched vesicles in real time and thus the comparison of the theory with the experiments. Unfortunately, such experiments have not been performed in electroporation research. Note that changes in the medium viscosity would imply dramatic changes in the medium conductivity and in the lipid properties, which thus would complexify the modeling. However since the end of the 1990s, biophysicists proposed different model for pore formation in membranes subjected to high transmembrane voltage. An extensive and physical presentation of such approaches is given in the first section, chapters “▶ Electropore Energy and Thermodynamics” and “▶ Pore Lifetime and Permeabilization Lifetime in Models” by J. Weaver.

Single-Pore Models Pore radius models are based on the description of the free energy of membranes, thanks to which Langevin-type equation provides the pore radius evolution. More precisely, given the membrane energy Em as a function of the pore radius r, the time evolution of the single-pore radius r behaves as follows accordingly to Weaver and Chizmadzhev (1996): dr D ¼ @ r Em ; dt kB T

(18)

where D, kB, and T hold, respectively, for the diffusion coefficient, the Boltzmann constant, and the temperature. The two next subsections deal with the two main models that describe the single-pore radius that referred to, respectively, as BGS-CW model and DAV model.

The Classical BGS-CW Model Brochart–Wyart, de Gennes, and Sandre (Sandre et al. 1999) on one side and Chizmadzhev and Weaver (Weaver and Chimazdzhev 1996) on the other side proposed the membrane free energy Em given by

10

C. Poignard et al.

  Cs Em ðr Þ ¼ Ef ðr Þ þ Ep ðr Þ ¼ 2πγr  πr 2 σ 0 þ ap V 2m þ 4 : r

(19)

The theoretical basis of this energy is presented in J.C. Weaver’s chapter “▶ Electropore Energy and Thermodynamics.” In the above equation, γ is the line tension, which tends to shrink the pore, while σ 0 is the surface tension of the stretched vesicles. The term Cs holds for the steric repulsion of the lipids. It ensures that a small space r0 between the phospholipids remains at rest and has a very low influence for pore radii above r0. The term ap is the membrane capacitance and Vm is the transmembrane voltage. The pore radius evolution, which is called here BGS-CW model, is then derived thanks to (18) and Stokes–Einstein – that is, kB T=D ¼ 2πηm δ where δ is the membrane thickness – such that     Cs dr 1 ¼ γ  r σ 0 þ ap V 2m  5 : dt ηm δ r

(20)

One can point out a few drawbacks of the model. The first one lies in the fact that pores are assumed cylindrical all along their lives. Such assumption seems justified during the formation, according to the very recent study of Sengel and Wallace (2016). However there is no physical justification of circular shrinkage, in particular the lateral diffusion of membrane lipids is important (a few nm2 per ns) and should change the shape of the defects. In addition pore expansion is exponentially fast above the critical radius rc given by rc 

γ ; σ 0 þ ap V 2m

(21)

which is hardly defensible since the pore radius should be bounded at least by half of the membrane circumference. On the other hand, if the membrane voltage is stopped, the pore shrinks, but the shrinkage is linear as soon as the radius is smaller than γ/σ0, whereas Sandre et al. have reported an acceleration of the shrinkage for small radii (Sandre et al. 1999). More recently, Kroeger et al. (2009 and Ryham et al. (2011) have pointed out that experiments are at odds with the linear closure, and they derived in two different ways a curvature-driven pore model.

Curvature-Driven Pore Closure: The Role of Aqueous Viscosity In Ryham et al. (2011), Ryham et al. suggested that the aqueous viscosity of the ambient medium impacts the pore dynamics. They derived the dominant aqueous viscosity (DAV) model by adding a force Fs, which accounts for the lateral stresses generated on the bilayer: Fs ¼ Cηs r

dr ; dr

(22)

where C is a non-dimension constant, whose value is around C  8 according to (Ryham et al. 2011) and ηs is the viscosity of the solution. Summing up all the forces,

Different Approaches Used in Modeling of Cell Membrane Electroporation

11

the following equation on the pore radius evolution in membrane submitted to a transmembrane voltage holds Cηs r

  Cs dr dr þ ηm δ ¼ γ þ r σ 0 þ ap V 2m þ 5 : dt dt r

(23)

Note that the steric repulsion Cs/r5 is not given in the DAV model as written in Ryham et al. (2011); however it is necessary to prevent negative radii. The main insight of this model, which fits very well the experiments of Portet and Dimova (2010), as shown by Fig. 5 of Ryham et al. (2011), is the predominance of the aqueous viscosity. Actually even if ηs is much smaller than the lipid viscosity, for pore radii bigger than 1 μm, the term Cηsr plays a crucial role in the pore closure since the membrane thickness h is very small of order 10 nm. The term ηeff ðr Þ ¼ ηm h þ Crηs ;

(24)

can be seen as the effective membrane viscosity, which increases linearly with respect to the pore radius. Ryham et al. pointed out that due to the membrane thinness, the aqueous viscosity cannot be neglected, explaining the curvature-driven pore closure. Kroeger et al. used another reasoning based on the electrochemical potential to obtain a similar equation accounting for membrane viscosity (see Eq. 7 of Kroeger et al. 2009), keeping the hypothesis of cylindrical pores.

From One Pore to the Total Pore Density Equation From these models, Smoluchowski equation, which is a drift–diffusion equation on the distribution function of pores of radius r, gives the pore distribution in the space of radii (see chapter ▶ Electropore Energy and Thermodynamics). The very important insight of Krassowska and Neu in the late 1990s was to approach Smoluchowski equation by an ordinary differential equation on the total pore density Nep, thanks to a subtle asymptotic analysis (Neu and Krassowska 1999). Interestingly Kroeger et al. (2009) pointed out that both BGS-CW model and DAV model of pore radius lead to the same ordinary differential equation for the pore density   dNep N ep qV2m =V 2ep V 2m =V 2ep ¼ αe 1 e ; dt N0

(25)

where Vep is the threshold membrane voltage above which electroporation occurs and No is the pore density at rest, when Vm equals 0 and α and q > 1 are ad hoc parameters.

12

C. Poignard et al.

Fig. 5 Numerical simulations of KN model from Fig. 3 of (Smith et al. 2014) (Courtesy of J.C. Weaver). “Electrical and poration response to a 1 kV/cm, 100 μs trapezoidal pulse with 1 μs rise and fall times. a Transmembrane voltage spatially averaged over the polar quadrants, during the initial 5.5 μs and at the end of the pulse from 97 to 103 μs. Red “+” indicates the hyperpolarizing anodic side, and black “” side and black “” indicate the depolarizing cathodic side. b Angular profile of transmembrane voltage at 99 μs. Background-shaded regions indicate the angular extent of EP around the anodic and cathodic sides of the membrane. Red and black dots mark the location of the anodic and cathodic poles on the membrane, respectively. c Equipotentials (black) and pore density, n(θ) (white), at 99 μs”

Pore Radii Evolution In the pore density model, all the pores are created with the same radius, rm. This is due to the averaging and the asymptotic analysis performed to pass from the Smoluchowski equation to the simple differential equation on Nep. In Krassowska

Different Approaches Used in Modeling of Cell Membrane Electroporation

13

and Filev (2007 and Smith et al. (2004), Krassowska et al. add another equation for the pore radius evolution rj, for j = 1, . . ., K, which has the same basis of Eq. 20: dr j D ¼ kB T dt

!   2   Cs ap r j V m  2πγ þ πr j σ eff r j þ 5 ; rj

(26)

where ap and σ eff are defined as   ap r j ¼

  Fmax   , σ eff r j ¼ 2σ 0  1 þ r h = r j þ rT

2σ 0  σ 0 1  14

XK j¼1

r 2j

!2 :

(27)

R2cell

In Krassowska and Filev (2007), the link between K and Nep is not precisely stated, but it seems that the following equality holds: K ðtÞ ¼ 4πR2cell N ep ðtÞ:

(28)

The KN Model for Electroporation Current From the total pore density Nep and from the description of the pore radii evolution, Krassowska and Neu et al. (Krassowska and Filev 2007; Neu and Krassowska 1999; Smith et al. 2004) propose the following nonlinear electroporation current Iep: I ep ðt, V m Þ ¼

K ðtÞ X

  iep r j , V m ;

(29)

j¼1

where iep(rj, Vm) is the current through a single pore of size rj. In Krassowska and Filev’s paper (Krassowska and Filev 2007), the current is given by   iep r j , V m ¼

2σr j   Vm; 1 þ δ= 2πr j

(30)

which is used in recent works by Weaver et al. (Son et al. 2016) and Li and Lin (2011). Note that the way K(t) is computed is still unclear. For numerical purpose, Filev and Krassowska proposed to consider only two kinds of pores: Ksmall for small pores and Klarge large pores. At the beginning, all the pores are created at the same radius, with a density Nep. Since each node is associated to a surface area δA, the number of small pores Ksmall is given by K small ¼ δANep on the nodes of the mesh. Then at any point of the mesh on the membrane, the pore radii evolve, and if the radius increases locally, then the Ksmall is decreased and Klarge increases. One of the drawbacks lies in

14

C. Poignard et al.

the fact that the number of pores depends on the mesh, which means that the numerical method is not mathematically consistent: changing the mesh changes the number of pores passing from Ksmall to Klarge. However since the transmembrane voltage value is quite robust with respect to the membrane conductance, as shown in subsection “Spherical Cell in a Unidirectional Constant Field” of section “Cell Electrical Modeling Before Electroporation: The Linear Regime,” such a variability should not be troublesome, even though it raises modeling issues.

Back to the Electric Equation and Drawbacks of the KN Model In order to link the biophysical model of electroporation to the electric cell model, the Kirchhoff Eq. 12c is changed into Cm @t V m þ S0 V m þ I ep ðt, V m Þ ¼ σ c @n U jΓ :

(31)

KN model provides a biophysical explanation of pore formation. The grounds of the model consist of cylindrical pore formed by transmembrane voltage. Such hypothesis has been recently confirmed, at least during the formation process, by optical signal-channel recording in Sengel and Wallace’s paper (2016), even though more complex phenomena are pointed out such as electropore diffusion and the resealing has not been addressed in this study. Interestingly, KN model provides a link between pores of the membrane and global electric potential in the whole cell. From the numerical point of view, the model is not easy to solve because it is nonlinear and stiff (the dynamic of pore creation is very rapid); therefore, accurate numerical methods combined with very fine mesh is needed. The computational cost is quite high, especially for finite element methods; therefore most of the numerical studies focus on bidimensional simulations. The typical behavior of transmembrane voltage Vm is illustrated by Fig. 5, which comes from Smith et al. (2014). During the first microsecond, the membrane is being charged as a capacitor, but once the transmembrane voltage reaches a threshold value, pores are created, leading to a decrease and then a stabilization of the transmembrane voltage within 2 μs. At the equator, the electric field is tangent to the membrane, and thus the transmembrane voltage does not increase: no electroporation occurs. At the pole, the electric field is normal, and the transmembrane voltage is the highest before electroporation; therefore, this is the privileged location for pore formation, as shown in ▶ Fig. C. 5. Such privileged locations near the poles facing the electrodes have been corroborated qualitatively by the experiments in many studies, in particular from Teissié and Rols’ group (see, for instance, Escoffre et al. 2009; Rols et al. 1998 and reference therein), Miklavčič’s group (Kotnik et al. 2010, 2012), and also Vernier’s team (2003).

Different Approaches Used in Modeling of Cell Membrane Electroporation

a

15

b 1 ms

10 4

5 nA

1 nA

3

1 ms 2 nA 5 nA

16 15

10 4

1

1

5 ms

2.5 2.0 1.5 1.0 0.5 0.0 2

4

6

8 10 12 14 16 18 20 Pulse number

Specific membrane conductance (nS/pF)

Specific membrane conductance (nS/pF)

10 ms 3.0 2.5 2.0 1.5 1.0 0.5 0.0 2

4

6

8 10 12 14 16 18 20 Pulse number

Fig. 6 Experimental patch-clamp measurements on DC3F (Wegner et al. 2015). “Two examples of the current response to repetitive application of double-pulse protocol, consisting of a depolarization to 320 mV command voltage for 10 ms and successive clamp at 80 mV for 25 or 100 ms (a and b, upper; note that part of the trace is omitted in b). Subsequently, the membrane was stepped back to 0 mV for 5 s before the next pulse of the same type was applied. Individual repetitions are superimposed as indicated by numbers. The membrane conductance at the physiological voltage range was calculated from the difference in steady-state current level induced by a voltage step from 80 to 0 mV (redrawn at an enlarged scale in a, b, as indicated by the arrows). Bar graphs (a, b, lower) show the conductance levels for the applied sequence of pulses. The conductance increased stepwise (dotted lines), either in several steps (a, pulses 3, 4, and 10) or in one step (b, pulse 15). These two examples were selected to reflect the variability among individual cells”

KN Model vs Experiments One of the main drawbacks of the KN model lies in the fact that it barely corroborates quantitatively the experiments. One can point out here a few experimental observations non-accounted for by the KN model. • According to Son et al. (2016), the maximum electroporation should be achieved after the first pulse. Patch-clamp experiments on Chinese hamster ovary DC3F performed by Wegner et al. exhibit a cumulative effect of the pulses (Wegner et al. 2015) leading to a persistent permeabilization, in addition to the transient permeabilization accounted for by the models (see Fig. 6). As mentioned by Wegner et al., such a persistent effect could result from a change in lipid property (such as oxidation), which is not accounted for in the KN model. Interestingly, patch-clamp measurements of membrane intensity do no exhibit systematically a

16

C. Poignard et al.

stabilization during the pulse unlike KN model. According to Wegner et al. (2015), for some patch-clamped cells, the current response does not stabilize during the pulse, and moreover the membrane conductance increases with the number of pulses. • Such cumulative effect is also reported in experiments performed by Pakhomova et al. (2011) and Dermol et al. (2016). Whether the pulse repetition rate induces a sensitization or a desensitization is still controversial. Recently Silve et al. reported that high repetition rate of similar pulses is less efficient than the same pulses repeated at a lower rate (Silve et al. 2014). These two studies have different explanations of the same phenomenon: the pulse repetition rate has an impact on the electroporation. • From the modeling point of view, since membrane is a dynamic soft matter (lipids are constantly in motion), it is hard to imagine that no diffusion or motion of pores is involved in the description of pore density. Another important drawback is that saturation of the membrane conductance and pore formation is not accounted for as described, e.g., in Wegner et al. (2015). The KN model has been extensively studied in silico during the past decade, providing interesting explanations of electroporation. However the modeling has several drawbacks making the model hardly predictive. The main drawbacks of such modeling lie in the number of non-measurable parameters and the sensitivity of the results to the parameters which makes the model calibration hardly obtainable. Even though biophysics gives a range for their values, the model is too sensitive to slight modification of many parameters, making hard the parameter calibration with biological data.

Phenomenological Models for Membrane Conductivity In order to compare theoretical models and experimental measurements, an alternative to biophysical models is to develop phenomenological models with the fewest parameters to describe the main features of the experiments. The fact that a few parameters are involved facilitates the parameter identification to obtain results that are quantitatively in accordance with the experiments, the drawbacks being the physical justification of the modeling.

Biological Evidences of Electroporation The phenomenological model of membrane permeabilization proposed by Leguébe et al. in (Leguébe et al. 2014) is based on the following assumptions, which come from experimental observations: • Permeabilization results from a long-term effect of defects in the membrane related to an alteration of phospholipids in the membrane. Such alterations may

Different Approaches Used in Modeling of Cell Membrane Electroporation

17

be due to the combination of high electric field and water molecules inside the membrane. Actually it has been reported by Harakawa et al. (2005) that electric field changes the phospholipid composition, by altering the lipid property. • The dynamics of alteration and reconstruction of the membrane are much longer as observed by the experiments (Rols et al. 1998). The alteration of the lipids is a physical phenomenon, which occurs as long as pores are present on the membrane and which is persistent after the end of the field pulse. On the contrary, the membrane recovery is a biological phenomenon, which takes time: it happens for minutes after the electric shock. • Lipids diffuse along the membrane at a speed dL around 1 μm2/s (Fahey and Webb 1978), which is non-negligible compared to the lapse of time between two pulses (usually of the order of 1 s), and therefore this surface diffusion has to be accounted for.

The Two-Step Process of Membrane Electroporation Leguébe et al. proposed in 2014 a two-step model for membrane electroporation, describing the membrane surface conductance Sm as Sm ðt,Þ ¼ S0 þ S1 X1 ðt,Þ þ S2 X2 ðt,Þ;

(32)

where S0 is the membrane conductance at rest, S1 is the surface conductance of the fully porated membrane, and S2 is membrane conductance due to the long-term effect of electroporation. The non-dimension variables X1 and X2 refer to as the degree of poration during the pulse and the degree of long-term changes in the membrane, respectively, and they satisfy the following equations: @X1 βðV m ðt,ÞÞ  X1 ðt, V m ðt,ÞÞ ¼ , t > 0; τ1 @t h iþ @X2  dΓ ΔΓ X2 ¼ ðX1  X2 Þ=τ2 , t > 0; @t X1 jt¼0 ¼ X01 , X2 jt¼0 ¼ X02

(33a) (33b) (33c)

where []+ denotes the positive part, while τ1 and τ2 are the characteristic times of “pore” creation and changes in the membrane. The function β is a sigmoidal function, for instance, βðλÞ ¼ e

ðV ep =ðλþυ0 ÞÞ

2

   1 þ tan kep jλ þ υ0 j=V ep  1 , or βðλÞ ¼ 2

(34)

The Kirchhoff law is then given by Cm @t V m þ Sm ðt,ÞV m ¼ σ c @n UjΓ :

(35)

18

C. Poignard et al.

Such model involves six parameters: Vep, the two characteristic times τ1, τ2, the diffusion coefficient dΓ, and the two conductances S1 and S2.

Calibration of the Model with Patch-Clamp Experiments Patch-clamp experiments provide interesting direct measurements of the current that flows across the membrane, for different pulses. Considering patch-clamp experiments on DC3F of (Wegner et al. 2015), it is possible to fit the model parameters with the data. The idea is to use the angular invariance of the command voltage g. Assuming that the source potential is applied at the center of the cell – assumed spherical – on a ball of radius Rs, the flux σ c @r Ujr¼Rc reads     Rc Rc Rc   σ c @r U r¼Rc ¼ ðV m þ gÞ= þ 1 ; σ e σ c Rs

(36)

and thus the membrane potential Vm satisfies I ðtÞ :¼ Cm @t V m þ Sm ðt, V m Þ ðV m þ υ0 Þ    Rc Rc Rc ¼ ðV m þ gÞ= þ 1 : σ e σ c Rs

(37)

The feature of the above equation is twofold: on one hand, the patch-clamp technique measures directly the current I, and thus one can compare measured and computed intensities. On the other hand the right-hand side of the formula makes it possible to obtain the transmembrane voltage from the measured intensity. The value of the parameters to obtain the currents of Fig. 8 is given in Table 1 of Leguébe et al. (2014). Phenomenological models being designed to be calibrated with the experiments (see Figure 7), it is natural to try to identify parameters that match the data, the key point being to find a set of parameters such that numerical results fit with as many experiments as possible. Indeed, if different set of parameters are needed for each experiment, it means that the model is not reliable. Interestingly, Fig. 8 shows that it is possible to find a set of parameters so that the numerical simulations fit quite well with the data. The two functions X1 and X2 seem necessary to describe the data: the first X1 increases very fastly during the pulse and vanishes at the end of the pulse, while X2 increases more slowly but remains after the pulse. Fitting the KN model with these patch-clamp data is proved to be difficult. One reason is the huge number of parameters, which prevents a systematic sensitivity analysis of the model. The two-step process exhibited by the data is hard to obtain with the variable Nep.

Different Approaches Used in Modeling of Cell Membrane Electroporation

19

Fig. 7 Comparison between simulations and patch-clamp data. The applied voltage is a rectangular function (as described in Wegner et al. 2015) with a voltage amplitude of 820 mV during 10 ms. Thanks to the model, one can compare the currents but also the transmembrane voltage (TMV). The membrane conductance is increasing during the pulse delivery

Biophysical Basis of the Phenomenological and Link with Pore Radius Evolution Equation (33) of the phenomenological model of Leguébe et al. is very different from the approach of Krassowska et al. More precisely, Eq. 33a can be linked to the pore density Eq. 25, but the reaction–diffusion (33b) involves surface diffusion of lipids which is not accounted for in Krassowska et al. approach. Even though the physical bases of such reaction–diffusion model is still unclear, it can be related to the phase-ordering kinetics approach, as described in the seminal book of Bray (2002). The concept being to look at the membrane as a two-phase medium described thanks to the order-parameter field ϕ ¼ 0: the lipid phase, corresponds to the state ϕ and the water phase to ϕ ¼ 1. The Landau free energy of the membrane subjected to transmembrane voltage Vm is given by the functional H(ϕ):  Z  1 1 2 2 j∇ϕj þ V ðϕÞ þ Cm ðϕÞV m ds; H ð ϕÞ ¼ 2 Γ 2

(38)

20

C. Poignard et al.

Fig. 8 Comparison of the computed and measured currents for different applied voltages to DC3F for a well-chosen set of parameters. Experimental setup is described in (Wegner et al. 2015). Interestingly, the measured current exhibits a two-step increase, as in the model. Such results cannot be obtained easily with the KN model. a Applied voltage amplitude: b applied voltage amplitude: c applied voltage amplitude: 860 mV, 460 mV, and 140 mV

where V is a double-well potential with two minima 0 and 1 and Cm(ϕ) is the membrane capacitance depending on ϕ. For instance, Weaver and Chizmadzev proposed Cm ¼ CW ϕ þ CL ð1  ϕÞ in Weaver and Chimazdzhev (1996). From this Landau free energy, the so-called model A (Bray 2002) describes the evolution of the order parameter: @t ϕ ¼ d

@H ðϕÞ; @ϕ

(39)

leading to the following reaction–diffusion on ϕ: 1 @t ϕ ¼ dΔΓ ϕ  V 0 ðϕÞ þ ðCW  CL ÞV 2m : 2

(40)

The right-hand side of phenomenological model can be seen as a specific model of the term V 0 ðϕÞ  12 ðCW  CL ÞV 2m in the above equation. The main insight of such approach is to avoid the simplistic assumption of cylindrical pores. It is more general, but the equation on single pore can be derived from it. More precisely,

Different Approaches Used in Modeling of Cell Membrane Electroporation

21

making the assumption that only one cylindrical pore of radius r is present in the membrane supposed to be flat and infinite, then the last equation leads to an equation similar to Eq. 20, as described in Bray’s book (2002).

The Role of the Surface Diffusion The main novelty of Leguébe et al. model lies in the surface diffusion, which is numerically difficult to tackle. It is thus important to determine whether or not such surface diffusion plays an important role. Theoretically, the diffusion induces a delay in the spatial response of the membrane; therefore, diffusion should appear when several identical pulses at different frequency are applied. In model without surface diffusion, the maximal electroporation occurs after the first pulse, and then the pulse efficiency decreases as reported by Weaver et al. in Son et al. (2016). In particular the influence of pulse frequency is not accounted for. Surface diffusion induces a complex response, which is a balance between the characteristic time of lipid diffusion and pulse frequency. In order to investigate the role of the surface diffusion, 3D simulations of a spherical cell submitted to ten permeabilizing micropulses (10 μs, 40 kV/m), with various repetition rates from 1 to 1,000 Hz, are presented. The lateral diffusion of the lipids on the membrane is set to dL ¼ 1012 m2 s1 ; which is in the range of the measured lateral diffusion of the lipids in cell membranes (Fahey and Webb 1978). The average permeabilization X2 of the membrane is computed at any simulation time. Figure 9 shows the distribution of X2 on the surface of the cell at different instants of the 1 Hz and 1,000 Hz simulations. For the case of fast repetition rate, the altered lipids do not have time to be evenly spread on the membrane. Since the next pulse alters the same region as the previous one, therefore the total quantity of altered lipids is lower than for the 1 Hz case. Figure 10 presents the average of X2 after each pulse. As expected, the permeabilization is more efficient if enough time is left between pulses to let the lipids diffuse. These simulations corroborate qualitatively the results of high voltage/low voltage experiments (Pucihar et al. 2002; Šatkauskas et al. 2005) that, within the first seconds after the pulses, show a better permeabilization when the lapse of time between two consecutive pulses is longer.

Concluding Remarks and Perspectives In this chapter, biophysical and phenomenological ways to model cell electropermeabilization have been presented, and pros and cons for each approach are discussed. To summarize, biophysical models are more reliable due to their

22

C. Poignard et al.

1 0.8

After pulse 1

0.6

Before pulse 2

After pulse 2

After pulse 5

2 s after pulse 10

0.2 0 X2 (A.U.)

Fig. 9 Numerical illustration of the lateral diffusion effect (Leguébe et al. 2014). Influence of the pulse frequency on the membrane permeabilization X2. The magnitude of each pulse is 40 kV/m during 10 μs. Ten pulses are applied on both cells, but the time between pulses is different: 1 s for the top line, 1 ms for the bottom line. After ten pulses, the average of X2 is around 8  108 for the 1 Hz case and half for the 1,000 Hz case 1

X2 (A.U.)

0.8 1 Hz 10 Hz 100 Hz 1000 Hz

0.6

0.4

0.2

0

1

2

3

4

5 6 7 Pulse number

8

9

10

Fig. 10 Numerical illustration of the increase of the variable X2 (Leguébe et al. 2014). Average permeabilization X2 on the cell after each of the ten pulses of Fig. 9 for different pulse repetition rates

theoretical grounds, but they generally involve too many parameters that prevent quantitative comparisons with the biological data. On the other hand, phenomenological models are designed to account for the biological observations, but they suffer from a lack of mechanistic basis. An interesting challenge in cell electroporation modeling would be to enrich the phenomenological modeling with theoretical grounds, and the phase-ordering kinetics theory could be the appropriate way to tackle this challenges.

Different Approaches Used in Modeling of Cell Membrane Electroporation

23

The long-term goal would consist in providing a model that can predict quantitatively the degree of cell permeabilization and the amount of molecules taken up by electroporation. The fitting of the parameters is thus a crucial point, which is common to these perspectives. Acknowledgments This study has been carried out with financial support from the French State, managed by the French National Research Agency (ANR) in the frame of the “Investments for the future,” Programme IdEx Bordeaux, CPU (ANR-10-IDEX-03-02). Numerical simulations presented in this paper were carried out using the PLAFRIM experimental testbed, being developed under the Inria PlaFRIM development action with support from LABRI and IMB and other entities: Conseil Régional d’Aquitaine, FeDER, Université de Bordeaux, and CNRS (see https://plafrim.bordeaux.inria.fr/). C.P. is partly granted by the Plan Cancer project DYNAMO (9749 Inserm), and the Plan Cancer project NUMEP (11099 Inserm). C.P. is also member of the European lab of CNRS, LEA EBAM, on electroporation.

Cross-References ▶ Electroporation and Electropermeabilization ▶ Electropore Energy and Thermodynamics ▶ Pore Lifetime and Permeabilization Lifetime in Models

References Bray A (2002) Theory of phase-ordering kinetics. Adv Phys 51:481–587 Dermol J, Pakhomova O, Pakhomov A, Miklavčič D (2016) Cell electrosensitization exists only in certain electroporation buffers. PLoS ONE 11:e0159434 Escoffre J, Portet T, Wasungu L, Teissié J, Dean D, Rols M (2009) What is (still not) known of the mechanism by which electroporation mediates gene transfer and expression in cells and tissues. Mol Biotechnol 41:286–295 Fahey PF, Webb WW (1978) Lateral diffusion in phospholipid bilayer membranes and multilamellar liquid crystals. Biochemistry 17(15):3046–3053 Foster K, Schwan H (1989) Dielectric properties of tissues and biological materials: a critical review. CRC Biomed Eng 17(1):25–104 Harakawa S, Inoue N, Hori T, Tochio K, Kariya T, Takahashi K, Doge F, Suzuki H, Nagasawa H (2005) Effects of a 50 Hz electric field on plasma lipid peroxide level and antioxidant activity in rats. Bioelectromagnetics 26(7):589–594 Kotnik T, Pucihar G, Miklavčič D (2010) Induced transmembrane voltage and its correlation with electroporation-mediated molecular transport. J Membr Biol 236:3–13 Kotnik T, Kramar P, Pucihar G, Miklavčič D, Tarek M (2012) Cell membrane electroporation ? Part 1: the phenomenon. IEEE Electr Insul M 28:14–23 Krassowska W, Filev PD (2007) Modelling electroporation in a single cell. Biophys J 92 (2):404–4017 Kroeger JH, Vernon D, Grant M (2009) Curvature-driven pore growth in charged membranes during charge-pulse and voltage-clamp experiments. Biophys J 96(3):907–916 Leguébe M, Silve A, Mir L, Poignard C (2014) Conducting and permeable states of cell membrane submitted to high voltage pulses. mathematical and numerical studies validated by the experiments. J Theor Biol 360:83–94

24

C. Poignard et al.

Li J, Lin H (2011) Numerical simulation of molecular uptake via electroporation. Bioelectrochemistry 82(1):10–21 Neu J, Krassowska W (1999) Asymptotic model of electroporation. Phys Rev E 53(3):3471–3482 Pakhomova O, Gregory B, Khorokhorina V, Bowman A, Xiao S, Pakhomov A (2011) Electroporation-induced electrosen-sitization. Plos One 6:e17100 Perrussel R, Poignard C (2013) Asymptotic expansion of steady-state potential in a high contrast medium with a thin resistive layer. Appl Math Comput 221:48–65 Poignard C (2009) About the transmembrane voltage potential of a biological cell in time-harmonic regime. ESAIM Proc 26:162–179 Poignard C, Silve A (2014) Différence de potentiel induite par un champ electrique sur la membrane d’une cellule biologique. La Rev 3EI 75:11–20 Portet T, Dimova R (2010) A new method for measuring edge tensions and stability of lipid bilayers: effect of membrane composition. Biophys J 84:3263–3273 Pucihar G, Mir L, Miklavčič D (2002) The effect of pulse repetition frequency on the uptake into electropermeabilized cells in vitro with possible applications in electrochemotherapy. Bioelectrochemistry 57:167–172 Pucihar G, Kotnik T, Valič B, Miklavčič D (2006) Numerical determination of transmembrane voltage induced on irregularly shaped cells. Ann Biomed Eng 34(4):642–652 Rols M, Delteil C, Golzio M, Teissié J (1998) Control by ATP and ADP of voltage-induced mammalian-cell-membrane permeabilization, gene transfer and resulting expression. Eur J Biochem 254:382–388 Ryham R, Berezovik I, Cohen F (2011) Aqueous viscosity is the primary source of friction in lipidic pore dynamics. Biophys J 101:2929–2938 Sandre O, Moreaux L, Brochard-Wyart F (1999) Dynamics of transient pores in stretched vesicles. Proc Natl Acad Sci 96:10591–10596 Šatkauskas S, André F, Bureau M, Scherman D, Miklavčič D, Mir L (2005) Electrophoretic component of electric pulses determines the efficacy of in vivo DNA electrotransfer. Hum Gene Ther 16(10):1194–1201 Sengel J, Wallace MI (2016) Imaging the dynamics of individual pores. PNAS 113:5281–5286 Silve A, Giumerá Brunet A, Ivorra A, Mir L (2014) Comparison of the effects of the repetition rate between microsecond and nanosecond pulses: electropermeabilisation-induced electro-desensitization? Biochim Biophys Acta Gen Subj 1840:2139–2151 Smith K, Neu J, Krassowska W (2004) Model of creation and evolution of stable electropores for DNA delivery. Biophys J 86(5):2813–2826 Smith K, Son R, Gowrishankar T, Weaver J (2014) Emergence of a large pore subpopulation during electroporating pulses. Bioelectrochemistry 100:3–10 Son R, Gowrishankar T, Weaver J (2016) Modeling a conventional electroporation pulse train: decreased pore number, cumulative calcium transport and an example of electrosensitization. IEEE Trans Biomed Eng 63:571–580 Vernier T, Sun Y, Marcu L, Salemi S, Craft C, Gundersen M (2003) Calcium bursts induced by nanosecond electric pulses. Biochem Biophys Res Commun 310(2):286–295 Weaver J, Chimazdzhev Y (1996) Theory of electroporation: a review. Bioelectrochem Bioenerg 41:135–160 Wegner L, Frey W, Silve A (2015) Electroporation of dc–3f cells is a dual process. Biophys J 108:1660–1671

Electric Field Distribution and Electroporation Threshold Matej Kranjc and Damijan Miklavčič

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroporation Threshold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Values of Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electric Field Distribution in Biological Tissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Numerical Modeling of Electric Field Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monitoring of Electric Field Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Magnetic Resonance Electrical Impedance Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In Vivo Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plant Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 2 5 6 7 8 9 9 10 11 15 15 16

Abstract

This chapter is dedicated to fundamentals of interaction between cells and tissues exposed to the externally applied electric field. Through experimental work and mathematical modeling, it has been shown that an accurate coverage of tissue with sufficiently large electric field presents one of the most important conditions for successful outcome of electroporation-based applications. The electroporation process as well as cell viability is also governed by other parameters of applied electric pulses and characteristics of targeted tissue; thus, different electroporation threshold values of the electric field for reversible and irreversible electroporation are being reported. Electric pulses and tissue structure also define established electric field distribution which is difficult to predict. Still, numerical modeling M. Kranjc (*) • D. Miklavčič (*) Faculty of Electrical Engineering, University of Ljubljana, Ljubljana, Slovenia e-mail: [email protected]; [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_4-1

1

2

M. Kranjc and D. Miklavčič

has proven to be very efficient in providing simulated maps of electric field distributions for various electrode geometries and tissue structures, especially if they incorporate nonlinear behavior of tissue electrical properties. Electric field distribution in tissues can also be indirectly monitored using magnetic resonance techniques that enable determination of electric field distribution in situ while taking into account nonlinear changes that occur in the tissue due to electroporation. Brief introduction to the monitoring method together with maps of electric field distributions in animal and vegetable tissues obtained by means of magnetic resonance techniques is presented in the last part of the chapter. Keywords

Reversible electroporation • Irreversible electroporation • Electrical conductivity • Numerical modeling • Magnetic resonance electrical impedance tomography

Introduction Most theoretical studies predict that the application of electric pulses to the cell results in structural changes which occur in the membrane in the form of hydrophilic pores. This phenomenon, termed electroporation (sometimes also electropermeabilization), allows various normally nonpermeant molecules to enter the cell by crossing the membrane. When electric parameters (number, shape, duration, and repetition frequency of electric pulses, direction of electric field), electrode geometry, and electrode positions are appropriately chosen and consequently the cell is exposed to adequate electric field, transient structural changes can be attained. After a certain period of time, the membrane reseals and the cell survives. This is termed reversible electroporation as cell preserves its viability. On the contrary, irreversible electroporation leads to cell death as the applied electric field is too strong and the cell does not regain its homeostasis after electroporation. Through experimental work and mathematical modeling, it has been shown that an accurate coverage of tissue with sufficiently large electric field presents one of the most important conditions for successful outcome of electroporation applications such as electrochemotherapy, irreversible electroporation tissue ablation, and pulsed electric field processing, to name a few. In the first part of chapter, relation between transmembrane voltage and electric field is briefly explained, followed by an overview of electroporation outcomes in relation to applied electric field strength. The second part of the chapter is dedicated to electric field distribution, how it is established, and how it depends on tissue properties. In addition, monitoring of electric field distribution is presented together with results in biological tissues.

Electroporation Threshold Cell electroporation occurs when the externally applied electric field is above the electroporation threshold value. The cell can be exposed to the electric field by an application of electric pulses which establish an induced potential difference across

Electric Field Distribution and Electroporation Threshold

3

the cell membrane, i.e., transmembrane voltage (see “▶ Transmembrane Voltage Induced by Applied Electric Fields”), which is added to the resting membrane potential. The cell membrane becomes permeable when transmembrane voltage, a sum of induced potential difference and resting membrane potential, exceeds certain critical value. The standard model of electroporation or model of pore formation describes electroporation phenomenon as a formation of aqueous pores due to applied transmembrane voltage. Aqueous pores are established after transition from the hydrophobic to the hydrophilic state which are then stable due to a local minimum of free energy. This state is reversible until applied voltage remains under critical value, whereas when this value is exceeded, an irreversible breakdown of the membrane occurs. Relation between applied electric field E and transmembrane potential difference ΔV in a spherical cell is described in Schwan’s equation 3 ΔV ¼ Er cos ðφÞ 2

(1)

where r is the radius of the cell and φ is the polar angle measured from the center of the cell with respect to the direction of the field (Grosse and Schwan 1992). Relation between these variables in Eq. 1 shows that induced transmembrane voltage is proportional to the electric field and the cell radius. For irregularly shaped cells, as well as for cells close to each other, the induced transmembrane voltage cannot be derived analytically; thus, numerical approach has to be used. In dense cell suspension and in tissues, cells are positioned much closer together, and therefore each of the cells is not exposed to the same electric field as it is in the case of the low density cell suspension (Fig. 1). Structure of tissues is usually not homogeneous but rather heterogeneous as tissues consist of cells of different types, sizes, shapes, and orientations. In addition, cells in tissues are distributed in different densities and are connected to each other through gap junctions and extracellular matrix that can additionally influence the process of electroporation (Fear and Stuchly 1998). Still, for all types of cells, it is in common that induced transmembrane voltage and electroporation process are strongly correlated with the applied electric field. Exposure of cells to external electric field Eext can lead to four different outcomes of electroporation process defined by three different thresholds (see Fig. 2): • Eext < Erev: No electroporation process occurs since the electric field below reversible electroporation threshold Erev. • Erev < Eext < Eirrev: Electric field strength exceeds Erev and temporary membrane permeabilization is established. Electric field is still below irreversible electroporation threshold Eirrev, and cells remain viable after the end of electric field exposure. • Eirrev < Eext < Etherm: Permanent membrane permeabilization leads to extensive leakage of intracellular content and cell death. Still, no thermal damage is present since electric field is below critical value of Etherm. Here, it is important to note

4

M. Kranjc and D. Miklavčič

Fig. 1 Fraction of electroporated cell, as a function of electric field strength cell densities of 20 vol.%, 28 vol.%, and 33 vol.%. The results are the averages of five random distributions of ternary core–shell tissue models for a given cell density. The curves are the fits of the hyperbolic tangent law to the calculated data. Duration of the electric field was 100 μs (The figure is adapted from Essone Mezeme et al. 2012)

Fig. 2 Reversible electroporation, irreversible electroporation, and thermal damage as functions of electric field strength and pulse duration. Two points with different combinations of pulse parameters but with the same outcome, i.e., reversible electroporation, are depicted

that combining the cell permeabilization effect of electroporation with the effects of electrolysis (see “▶ Combining Electrolysis and Electroporation for Tissue Ablation”) can produce effective ablation even at electric fields below Eirrev (Stehling et al. 2016). • Eext > Etherm: Electric field establishes high electric currents causing temperature increase and thermal damage to the cell.

Electric Field Distribution and Electroporation Threshold

5

The outcomes of electroporation process are not determined solely by electric field strength but also by the duration of exposure to the field (texp): texp ¼ tp  N p ;

(2)

where tp is a duration of a single electric pulse and Np is the number of applied pulses. Each specific application of electroporation requires to some extent different settings of pulse parameters, among which the pulse amplitude, tp, and Np have the largest impact on the outcome of the experiment. In addition, these parameters need to be adjusted for the targeted cell type, density of cells, and electrode geometry. Still, the same outcome of experiment can be obtained by using equivalent pulse parameters which is a compensation of the change in the value of specific pulse parameter by carefully selecting value of the other parameter, e.g., instead of using a number of short, high-voltage pulses, one can use longer pulses with lower voltage. Relation between pulse parameters can be described by simple mathematical expressions, such as power, logarithmic, or exponential functions (Pucihar et al. 2011).

Values of Thresholds The optimal values of electric pulse parameters depend on the sensitivity of the cells or tissue to the electric pulses; thus, different threshold values of the electric field for reversible and irreversible electroporation are being reported in literature. In vitro electroporation thresholds are different in regard to cell lines. For example, reversible threshold was found to be between 400 and 600 V/cm for different cell lines exposed to electric pulses with a duration of exposure to the electric field (texp) of 8  100 μs (Čemažar et al. 1998), whereas the majority of electric field strength used in irreversible electroporation studies fall between 1000 and 2000 V/cm depending on texp (Jiang et al. 2015). Still, electroporation thresholds cannot be regarded as discrete values where all cells are either permeabilized or not affected. In reality, these transitions from non-electroporated to electroporated state (see Fig. 1) and from reversibly to irreversibly electroporated cells are continuous. These so-called permeabilization and survival curves (Canatella et al. 2001) can be described using mathematical models (see “▶ Mathematical Models Describing Cell Death Due to Electroporation In Vitro”) which allow prediction of electroporation outcomes not just for electric parameters that were used for determination of the curve but also for other parameters using interpolation. Some of the best models of cell survival as a function of treatment time were the adapted Gompertz and the Geeraerd models and, as a function of the electric field, the logistic, adapted Gompertz and Peleg–Fermi models. Permeabilization and survival in vivo curves are similarly shaped as in vitro, but they are scaled to the thresholds according to the type of treated tissue. Reversible and irreversible thresholds were mainly determined by overlapping electric field distribution obtained by mathematical models with experimental observations such as histological analysis. There are not many reports on in vivo reversible threshold

6

M. Kranjc and D. Miklavčič

for different tissues. One of the well-known estimation of reversible threshold (460 V/cm) was done on a rabbit liver exposed to electric pulses with (texp) of 8  100 μs (Miklavcic et al. 2000; Sel et al. 2005). The estimation was obtained by overlapping histological results with the electric field distribution obtained by the numerical model of tissue that incorporated electric field-dependent electrical conductivity. There are more reported observations available for irreversible thresholds with threshold ranging from 500 to 1300 V/cm (Jiang et al. 2015; Qin et al. 2013). Similarly as in in vitro studies, it is important to acknowledge that these values depend on texp, i.e., on the number and duration of applied pulses. Pulsed electric field (PEF) also determines the outcome and the efficiency of food processing applications such as improving mass transfer processes. The phenomenon of mass transfer occurs in many operations of the food industry in which the aim is extracting a given substance of interest (fruit juices, sugar, colorants, antioxidants, etc.), removing water from foods (drying), or introducing a given substance into the food matrix (osmotic dehydration, salting). The electric field required for improving mass transfer can be achieved at field strength in the range of 50–500 V/cm and 1–10 kV/cm for texp in the range of 100–10,000 μs and texp < 100 μs, respectively. One of the well-known applications in food industry that employs irreversible electroporation is microbial inactivation. Since the size of microorganism is around 10–100 times smaller than the size of eukaryotic cells, higher electric field strength (10–35 kV/cm) needs to be applied for irreversible disruption of cell membrane integrity of microorganisms (see Eq. 1).

Electric Field Distribution in Biological Tissue Electric field distribution is established in the biological tissue when electric current passes through the tissue. In conventional electroporation, electric current is established by applying electric pulses via electrodes that are in contact with the tissue. This can be performed either via noninvasive plate electrodes, which embrace the tissue, or with invasive needle electrodes, which are inserted into the tissue. In both cases, when electric current passes through the tissue, it is distributed through the tissue by taking the shortest and easiest path between points of highest potential differences. In general, distribution of electric current is represented by the amount of electric current per unit area, also known as current density. In terms of tissue electrical properties, areas with higher current densities are also areas with highest electrical conductivity. Highly perfused tissues, such as muscles, have higher conductivity, whereas tissues such as adipose tissues or bone have lower conductivity. In addition to current density distribution, electrical conductivity also defines electric field distribution but in opposite manner; electric field will be higher in areas with lower conductivity and vice versa (see Fig. 3c, d). Another feature that also defines electric field distribution is the inlet and the outlet of electric current, i.e., electrode geometry as demonstrated in Fig. 3a, b. Electric field strength established between plate electrodes is often estimated using U/d ratio, i.e., voltage applied to the electrodes (U ) divided by the distance

Electric Field Distribution and Electroporation Threshold

7

Fig. 3 Four examples of electric field distribution established in homogeneous and heterogeneous tissue in a section plane perpendicular to plate and needle electrodes. Potential difference between pair of electrodes (colored in white) in all of four examples is 1000 V. Width of plate electrodes (a, c) and diameter of needle electrodes (b, d) are 0.5 mm, while the distance between both types of electrodes is 10 mm. Homogeneous tissue (a, b) consists of single medium with homogeneous electrical conductivity, whereas heterogeneous tissue (c, d) consists of additional circular-shaped structure with 10 times higher electrical conductivity. All four examples of electric field distribution were obtained by numerical modeling

between the electrodes (d ). This estimation is valid only approximately in the area between and away from the edges of the electrodes. Discrepancy between estimation and real value of electric field becomes more evident with larger d and smaller electrode surfaces or if embraced tissue is not homogeneous (see Fig. 3c). Therefore, for valid prediction of electric field strength and distribution, it is necessary to introduce numerical modeling, especially in the case of any other electrode geometry than plate electrodes or if treatment is performed on tissue with heterogeneous electrical conductivity.

Mathematical Model By definition, mathematical model is a description of a system composed of variables and equations that describe relationships between the variables. Mathematical models are frequently used in studies of the effect of the electromagnetic field

8

M. Kranjc and D. Miklavčič

accompanied by coupled phenomena on biological systems such as cells, tissues, and organs. These systems are usually based on rather complex geometries with inhomogeneous tissue properties and advanced boundary conditions which prevent solving mathematical models analytically. Therefore, instead of analytical approach, approximate solution of equations describing complex mathematical models is obtained by numerical modeling. A good mathematical model that is validated with experimental results is a strong tool that offers significant insight into the understanding of observed processes. This is especially true for electroporation where mathematical modeling is crucial for determining electric field distribution established in tissues which electrical conductivity increases due to application of electric pulses.

Numerical Modeling of Electric Field Distribution Mathematical model describing electric field distribution in biological systems can be modeled as steady state by neglecting transient responses that occur during electric pulse rise time since duration of electric pulses is considerably longer than cell membrane charging time. This simplifies set of Maxwell equations to Laplace steady-state equation: ∇ðσ∇V Þ ¼ 0;

(3)

where σ is an electrical conductivity of a biological system and V is established electric potential. The electric field distribution E is then a gradient of electric potential: E ¼ ∇V:

(4)

The numerical method that has proven to be very efficient in solving mathematical models of electromagnetic studies is finite element method, a numerical technique for finding approximate solutions of mathematical model by dividing the geometry of a model into small finite elements. The solution in these elements can be considered to be a simple piecewise polynomial function which can then be solved using standard mathematical techniques. Numerical models have already demonstrated the importance of contact between tissue and plate electrodes and the importance of insertion depth of needle electrodes. It has also been shown that for a certain number and duration of applied electric pulses, the electric field has to be higher than a threshold value for electroporation process to occur. Therefore, modeling of electric field distribution can be used for prediction of the electroporation outcome. Electric field distribution in the tissue can be determined by the position of electrodes and applied pulse parameters, both obtained by so-called treatment planning method (see “▶ Treatment Planning for Electrochemotherapy and Irreversible Electroporation of Deep-Seated Tumors”), which has already proved to have a great potential in clinical use of electrochemotherapy and tissue ablation with irreversible electroporation of deep-seated

Electric Field Distribution and Electroporation Threshold

9

tumors. The outcome of this method is a treatment plan which consists of instructions for physicians on where to position the electrodes and which voltages have to be applied between electrode pairs in order to ensure a successful treatment. The treatment plan is based on a numerical model prepared using CT- or MRI-based tumor images, which are used for segmentation of the tumor and important healthy structures in its surrounding. However, deviations in implementation of the treatment plan can occur during the treatment in the clinic, as it is relatively difficult to insert electrodes precisely according to the anatomy-based numerical treatment plan. With respect to the inherent nonlinear behavior of biological tissues, it is important to include those nonlinearities in the numerical model which contribute to the outcome of calculations. In the case of electroporation, the most important nonlinearity to include is the electrical conductivity of the tissue (see “▶ Electric Field Distribution Modeling in Tissue Considering Tissue Conductivity Increase Due to Electroporation”). Namely, the local electric field in tissues is affected by applied electroporation pulses, which depend on local electrical conductivity, and vice versa, electroporation increases the conductivity and consequently alters the electric field distribution. Also, the applied electric field has to reach reversible electroporation threshold Erev in order to cause conductivity increase. For the duration of the pulse, this increase is permanent, and electrical conductivity cannot decrease to its lower value even in the case in which electric field drops below Erev. This makes it difficult to properly characterize the numerical model of the treated tissue, and predicting the coverage of the treated tissue with accurate electric field is a daunting task, since it relies mostly on the accuracy of the electrical conductivity of the treated tissue used in the numerical model. As there is a lack of tissue-specific experimental data on tissue properties for numerical models that would be able to provide accurate and relevant electric field distribution, electroporation monitoring method that would allow direct measurement of the electric field strength within the sample would be of great importance for electroporation applications.

Monitoring of Electric Field Distribution A method capable of monitoring electric field distribution during electric pulse delivery would enable the detection of insufficient electric field coverage before the end of either reversible or irreversible electroporation treatment. Since electroporation depends on local electric field, monitoring of electric field distribution would increase and assure treatment effectiveness. As there are no available approaches of measuring electric field distribution in situ, an indirect approach using magnetic resonance techniques was suggested.

Magnetic Resonance Electrical Impedance Tomography Magnetic resonance electrical impedance tomography (MREIT) (see “▶ Principles and Use of Magnetic Resonance Electrical Impedance Tomography in Tissue

10

M. Kranjc and D. Miklavčič

Electroporation”) enables reconstruction of electric field distribution by measurement of electric current density distribution and electrical conductivity of the treated subject during application of electric pulses using MRI and numerical algorithms. This enables determination of electric field distribution in situ while taking into account changes that occur in the tissue due to electroporation. MREIT is a relatively new medical imaging modality based on numerical reconstruction of electrical conductivity inside a tissue by means of current density obtained by current density imaging (CDI) (see “▶ Current Density Imaging as Means to Follow Tissue Electroporation”) (Seo and Woo 2011; Joy et al. 1999). The MREIT algorithm applied for reconstruction of electrical conductivity of the tissue is based on solving Laplace’s equation (Eq. 3) through iterative calculation. Electrical conductivity is updated after each iteration (k + 1): σ kþ1 ¼

jJCDI j : j∇uk j

(5)

where JCDI is current density obtained by CDI and ∇uk is a solution of Eq. 3. When the difference between two successive conductivities falls below certain value, electric field distribution can be calculated using Ohm’s law: E¼

JCDI σ

(6)

Still, the main difficulty of using MREIT to determine electric field distribution during electroporation is associated with the limited capability of MRI scanners for their use during application of electroporation such as electrochemotherapy and irreversible electroporation tissue ablation. Hopefully, this could be surpassed in the near future using open MRI scanner (Wang et al. 2010). Another limitation is also that MREIT requires an application of at least two electric pulses with a delay of approximately a second between them in order to deliver complete current density information. This currently puts a frequency limitation on MREIT, although hopefully future improvements of MRI scanners will enable to reduce the required delay between applied pulses. Conductivity changes that occur during the pulse are at the moment also too demanding to asses with MREIT as a function of time. Even though, it is important to be aware that the cumulative effect of electric current on the MRI signal phase is measured. Therefore, this technique yields an electric field distribution, which is a time average of its altering time course so that all the consequences of conductivity alteration, which affect electric current, are not neglected within obtained electric field distribution.

In Vivo Tissues Investigation of the feasibility of MREIT for in situ monitoring of the electric field distribution during in vivo reversible electroporation was performed on mouse tumors (Kranjc et al. 2015). Electroporation was performed by applying two

Electric Field Distribution and Electroporation Threshold

11

sequences of four high-voltage electric pulses with an amplitude of 500 V, duration of 100 μs, i.e., texp = 800 μs, at a pulse repetition rate of 5 kHz via two needle platinum–iridium electrodes inserted into the tumor. Electric field distribution in the tumor during pulse delivery was obtained by means of MREIT. For posttreatment assessment of reversibly electroporated areas in the tumor, the contrast agent gadolinium-tetraazacyclododecanetetraacetic acid (Gd-DOTA) was injected before the application of pulses. After 24 h, T1-weighted images were acquired for the observation of tumor areas where Gd-DOTA molecules were entrapped inside the reversibly electroporated tumor cells (Leroy-Willig et al. 2005). An example of electric field distribution obtained by means of MREIT is shown in Fig. 4a where it is overlaid to T1-weighted image acquired just before the application of electric pulses. As expected, the electric field was the highest around the electrodes where it exceeded irreversible electroporation threshold (900 V/cm), while it remained under the reversible electroporation threshold (400 V/cm) in the areas away from electrodes. The average coverage of tumors with electric field leading to reversible electroporation of tumor cells CMREIT was calculated by dividing the predicted surface area of the reversibly electroporated tumor cells with the surface area of the entire tumor, whereas tumor fraction FGdDOTA was calculated by dividing the surface area of reversibly electroporated tumor cells as obtained by entrapped MR contrast agent with the surface area of the entire tumor. CMREIT  SD and FGdDOTA  SD were 38  9 % and 41  13 %, respectively. Apparent differences can be observed when comparing tumor coverage with electric field leading to reversible electroporation of tumor cells (CMREIT) in Fig. 4b. These differences can be attributed to a varying distance between the electrodes which lead to different electric field distributions in different tumors. Another source of varying electric field distribution is believed to be the heterogeneous nature of tumor electrical conductivity (Muftuler et al. 2006). However, the advantage of MREIT is exactly its ability to compensate the effects of different electrode placements and tumor conductivity heterogeneities which occur in treated tissues, thus yielding accurate electric field distribution.

Plant Tissues Electric field distribution during electroporation was successfully reconstructed in animal tissues and in vegetable tissues. Considering practical value, monitoring of electric field distribution in vegetable tissues seems interesting since it can potentially enable monitoring of pulsed electric field (PEF) applications in which applied electric field mostly determines the outcome and the efficiency of PEF applications. The study was performed on potato tubers (Kranjc et al. 2016) since PEF treatments are already well established in potato industry for reducing cutting forces oil uptake and browning during frying. In addition, electroporated areas in potato tubers become distinctively darker few hours after the treatment and are therefore convenient for observing electroporation effects.

12

M. Kranjc and D. Miklavčič

Fig. 4 (a) The electric field distribution in the tumor obtained by MREIT superimposed to the T-weighted image acquired before the application of electric pulses. (b) Scatterplot of the coverage of five tumors with the electric field of reversible electroporation CMREIT and Gd-DOTA cell entrapment FGdDOTA

Monitoring was performed on yellow-fleshed potato tubers’ (Solanum tuberosum) cultivar “Agata” that were sliced to cylindrical-shaped samples (Fig. 5). Two needle electrodes with a diameter of 1 mm were inserted in the potato sample and placed in a MR microscopy probe. Potato samples exposed to electric pulses were treated inside the MRI scanner equipped with micro-imaging system for MR microscopy. Electroporation treatment of potatoes was performed by applying two sequences of four high-voltage electric pulses with a duration of 100 μs, i.e., texp = 800 μs, a pulse repetition frequency of 5 kHz, and an amplitude of 500 V, 750 V, and 1000 V. Electric field mapping was enabled by two-shot rapid acquisition with relaxation enhancement (RARE) CDI sequence which enabled imaging of current density distribution inside conductive samples. Electric field in potato tubers during application of electric pulses was reconstructed from CDI data by a mathematical algorithm based on solving Laplace’s equation. Potato tubers were also dynamically monitored by multiparametric MRI protocol. The protocol consisted of diffusion-weighted imaging (DWI) based on a pulsed gradient spin-echo (PGSE) sequence for the apparent diffusion coefficient (ADC) mapping and a multi-spinecho (MSE) imaging sequence based on the Carr–Purcell–Meiboom–Gill multiecho train for the T2 mapping. DWI and MSE images were taken every 45 min until 12 h after the PEF treatment. A comparison of digital photographs of treated potatoes taken 18 h after application of electric pulses and the corresponding measured and simulated electric field maps is shown in Fig. 6. The darkened region in the treated potatoes is a result of oxidation that began immediately after the treatment. The extent of regions with high electric field in the measured electric field maps corresponds to the results of the simulations, while the electric field distribution deviates from the simulated. Interestingly, electric field was not distributed evenly on the left and right side of the treated potatoes. The origin of the effect is most likely associated with a heterogeneous potato structure as well as its heterogeneous electrical conductivity, which

Electric Field Distribution and Electroporation Threshold

13

Fig. 5 Illustration of potato tuber sample sliced to cylindrical-shaped samples. Two needle electrodes with a diameter of 1 mm were inserted in the potato sample and placed in a MR microscopy probe

resulted in an uneven distribution of the electric field. According to previous studies related to PEF treatment of potatoes, electrical conductivity of untreated potato tubers is considered homogeneous, and, with the application of electric pulses, conductivity starts to increase following the sigmoid function (Ivorra et al. 2009). However, even with an electric field-dependent conductivity, the electric field should be distributed symmetrically, similar to electric field distributions obtained by simulations in Fig. 6. In results obtained by MREIT, electric field distributions in potatoes were not symmetrical, suggesting that electrical conductivity of potatoes was heterogeneous even before the application of electric pulses (Faridnia et al. 2015). In addition, these differences between the simulated electric field distribution and the electric field distribution obtained by MREIT are becoming more distinct with a higher amplitude of applied electric pulses, presumably due to more pronounced effect of nonlinear relation between electric field and electrical conductivity of the potato tuber. Furthermore, multiparametric MRI allowed additional analysis of induced tissue changes due to application of electric pulses. These changes are associated with an increase of extracellular water content on the account of a decrease of intracellular water content. Differences between extracellular and intracellular water are reflected in T2 relaxation time of water compartments as shown in Fig. 7. Intracellular water is more bound and has therefore shorter T2 than the extracellular water. Gradual decrease of T2 values with time after the application of electric pulses is associated with a decrease of extracellular water content. As shown in Fig. 7, T2 values in potatoes are scattered in three groups. First, lowest values of T2 were measured in untreated potato tubers (E = 0 V/cm) and were consistent with results from other studies (Nott et al. 2003). Second, scatter of T2 was obtained in areas of potatoes exposed to an electric field ranging from 200 V/cm to 400 V/cm. Values of T2 were significantly different compared to untreated potatoes, suggesting that electric field resulted in permeabilization of cells’ membrane and in the release of water content from potato cells to extracellular space. Values of T2 and consequently the amount of extracellular water, however, have not changed significantly within 12 h. Third, scatter of T2 values was measured in areas of potato exposed to an electric field higher than 400 V/cm. Here, however, T2 values were linearly increasing with the electric field; hence the amount of

14

M. Kranjc and D. Miklavčič

Fig. 6 Digital photographs of three potato tubers from group 1 and one potato from the control group taken 18 h after the PEF treatment ( first row), measured electric field distributions during the PEF treatment (second row), and simulations of the electric field distributions obtained by the finite element method (third row). Potato tubers were subjected to electric pulses of amplitudes 0, 500, 750, and 1000 V Fig. 7 Scatter of T2 values (200 V/cm < E < 400 V/cm) of potatoes exposed to 750 V was significantly different compared to untreated potatoes, suggesting that electric field resulted in permeabilization of cells’ membrane and in the release of water content from potato cells. Scatter of T2 values (E > 400 V/cm) was linearly increasing with the electric field; hence the amount of released water increased with the intensity of electric field

released water increased with the intensity of electric field. In contrast to results below 400 V/cm, the amount of extracellular water decreased with time. Since the electric field value of 400 V/cm distinguishes two trends of T2 values, one can speculate that two electroporation processes were induced by an electric field of

Electric Field Distribution and Electroporation Threshold

15

lower and higher value of 400 V/cm, i.e., reversible below 400 V/cm and irreversible electroporation above 400 V/cm. Obtained range of field strength for reversible electroporation (between 200 and 400 V/cm) is in accordance with the study by Galindo et al. in which reversible electroporation was demonstrated by propidium iodide staining of cell nucleus using comparable electric field exposure time of texp = 1 ms (Galindo et al. 2009). Cell viability study also showed that field strength lower than 400 V/cm mostly does not influence the tuber cells, whereas increased cell death was observed when higher field strengths were applied. Also, field strengths higher than 400 V/cm are reported to have a considerable impact on potato tuber microstructure, leakage of ion, and breakdown of the membrane. It seems that reported significant changes also correspond to changes of T2 values.

Conclusion Cell electroporation occurs when the threshold transmembrane potential is reached, that is, when the externally applied electric field is above the electroporation threshold value. This state is reversible until applied voltage remains under critical value, whereas when this value is exceeded, an irreversible breakdown of the membrane occurs. The electroporation process as well as cell viability is governed by the parameters of applied electric pulses and characteristics of targeted tissue. It is important to acknowledge that each specific application of electroporation, e.g., electrochemotherapy, gene transfer, and irreversible electroporation tissue ablation, requires to some extent different settings of pulse parameters in order to obtain desired electroporation process outcome. Electric field distribution, which is established in biological tissue when electric current passes through the tissue, is difficult to predict, especially when targeted tissue has heterogeneous structure. Therefore, for valid prediction of electric field strength and distribution, it is necessary to introduce numerical modeling. Still, as there is a lack of tissue-specific experimental data on tissue properties for numerical models that would be able to provide accurate and relevant electric field distribution, indirect approach of monitoring electric field distribution during electroporation can be accomplished using magnetic resonance techniques.

Cross-References ▶ Combining Electrolysis and Electroporation for Tissue Ablation ▶ Current Density Imaging as Means to Follow Tissue Electroporation ▶ Electric Field Distribution Modeling in Tissue Considering Tissue Conductivity Increase Due to Electroporation ▶ Mathematical Models Describing Cell Death Due to Electroporation In Vitro ▶ Principles and Use of Magnetic Resonance Electrical Impedance Tomography in Tissue Electroporation ▶ Transmembrane Voltage Induced by Applied Electric Fields

16

M. Kranjc and D. Miklavčič

▶ Treatment Planning for Electrochemotherapy and Irreversible Electroporation of Deep-Seated Tumors

References Canatella PJ, Karr JF, Petros JA, Prausnitz MR (2001) Quantitative study of electroporationmediated molecular uptake and cell viability. Biophys J 80(2):755–764 Čemažar M, Jarm T, Miklavčič D, Lebar AM, Ihan A, Kopitar NA, Serša G (1998) Effect of electric-field intensity on electropermeabilization and electrosensitivity of various tumor-cell lines in vitro. Electro- Magnetobiol 17(2):263–272 Essone Mezeme M, Pucihar G, Pavlin M, Brosseau C, Miklavcic D (2012) A numerical analysis of multicellular environment for modeling tissue electroporation. Appl Phys Lett 100(14):143,701 Faridnia F, Burritt DJ, Bremer PJ, Oey I (2015) Innovative approach to determine the effect of pulsed electric fields on the microstructure of whole potato tubers: use of cell viability, microscopic images and ionic leakage measurements. Food Res Int 77:556–564 Fear EC, Stuchly MA (1998) Modeling assemblies of biological cells exposed to electric fields. IEEE Trans Bio-Med Eng 45(10):1259–1271 Galindo FG, Dejmek P, Lundgren K, Rasmusson AG, Vicente A, Moritz T (2009) Metabolomic evaluation of pulsed electric field-induced stress on potato tissue. Planta 230(3):469–479 Grosse C, Schwan HP (1992) Cellular membrane potentials induced by alternating fields. Biophys J 63(6):1632–1642 Ivorra A, Mir LM, Rubinsky B (2009) Electric field redistribution due to conductivity changes during tissue electroporation: experiments with a simple vegetal model. IFMBE Proc 25:59–62 Jiang C, Davalos RV, Bischof JC (2015) A review of basic to clinical studies of irreversible electroporation therapy. IEEE Trans Bio-Med Eng 62(1):4–20 Joy ML, Lebedev VP, Gati JS (1999) Imaging of current density and current pathways in rabbit brain during transcranial electrostimulation. IEEE Trans Bio-Med Eng 46(9):1139–1149 Kranjc M, Markelc B, Bajd F, Čemažar M, Serša I, Blagus T, Miklavčič D (2015) In situ monitoring of electric field distribution in mouse tumor during electroporation. Radiology 274(1):115–123 Kranjc M, Bajd F, Serša I, de Boevere M, Miklavčič D (2016) Electric field distribution in relation to cell membrane electroporation in potato tuber tissue studied by magnetic resonance techniques. Innovative Food Sci Emerg Technol (in press). Doi: 10.1016/j.ifset.2016.03.002 Leroy-Willig A, Bureau MF, Scherman D, Carlier PG (2005) In vivo NMR imaging evaluation of efficiency and toxicity of gene electrotransfer in rat muscle. Gene Ther 12:1434–1443 Miklavcic D, Semrov D, Mekid H, Mir LM (2000) A validated model of in vivo electric field distribution in tissues for electrochemotherapy and for DNA electrotransfer for gene therapy. Biochim Biophys Acta 1523(1):73–83 Muftuler LT, Hamamura MJ, Birgul O, Nalcioglu O (2006) In vivo MRI electrical impedance tomography (MREIT) of tumors. Technol Cancer Res Treat 5(4):381–387 Nott KP, Shaarani SM, Hall LD (2003) The effect of microwave heating on potato texture studied with magnetic resonance imaging. In: Magnetic resonance in food science. Royal Society of Chemistry, Cambridge, pp 38–45 Pucihar G, Krmelj J, Rebersek M, Napotnik TB, Miklavcic D (2011) Equivalent pulse parameters for electroporation. IEEE Trans Bio-Med Eng 58(11):3279–3288 Qin Z, Jiang J, Long G, Lindgren B, Bischof JC (2013) Irreversible electroporation: an in vivo study with dorsal skin fold chamber. Ann Biomed Eng 41(3):619–629 Sel D, Cukjati D, Batiuskaite D, Slivnik T, Mir LM, Miklavcic D (2005) Sequential finite element model of tissue electropermeabilization. IEEE Trans Bio-Med Eng 52(5):816–827 Seo JK, Woo EJ (2011) Magnetic resonance electrical impedance tomography (MREIT). SIAM Rev 53(1):40–68

Electric Field Distribution and Electroporation Threshold

17

Stehling MK, Guenther E, Mikus P, Klein N, Rubinsky L, Rubinsky B (2016) Synergistic combination of electrolysis and electroporation for tissue ablation. PLoS One 11(2), e0148,317 Wang H, Wang Y, Yang W, Wang Z, Hu L (2010) Conductivity image reconstruction of oblique slice with C-shaped open permanent magnet MRI systems. IEEE Trans Appl Supercond 20(3):814–817

Modeling of Electrochemical Reactions During Pulsed Electric Field Treatment Gianpiero Pataro, Giorgio Donsì, and Giovanna Ferrari

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principles of Electrochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrochemical Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrode Potentials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrochemical Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrochemical Behavior of a PEF Treatment Chamber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ionic Double Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Equivalent Electrical Circuit of an Electrochemical Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Metal Release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Metal Release During PEF Treatment of Food Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modeling of Metal Release Phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Case Study: Metal Release in a PEF Treatment Chamber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methodological Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 10 11 12 14 14 15 17 18 18 19 19 20 27 29

Abstract

In pulsed electric field (PEF) treatment, a biological material (cell suspension, plant, animal, or human tissue) is placed between two electrodes of a treatment chamber and here exposed to short (from several nanoseconds to several milliseconds) high intense electric field (0.5–80 kV/cm) pulses, with the final aim to induce the physical electropermeabilization of the cell membranes. However, G. Pataro (*) Department of Industrial Engineering, University of Salerno, Fisciano, SA, Italy e-mail: [email protected] G. Donsì • G. Ferrari Department of Industrial Engineering, University of Salerno, Fisciano, SA, Italy ProdAl Scarl – University of Salerno, Fisciano, SA, Italy e-mail: [email protected]; [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_5-1

1

2

G. Pataro et al.

when typical conditions for PEF processing are applied, undesired electrochemical reactions, especially those involving metal release from the electrodes, unavoidably occur at the electrode–electrolyte interface of a PEF treatment chamber. The occurrence of these electrode reactions is a very complex phenomenon, which is affected by several factors, such as PEF chamber design and electrode material, PEF electrical parameters, as well as composition and chemical–physical properties of the treated biological material. In this chapter, the basic concepts of electrochemistry are reported, and a detailed description of the electrochemical phenomena occurring at the electrode–electrolyte interface of a PEF chamber is given with a special emphasis to the case of PEF application for food pasteurization. Finally, particular attention is devoted to metal release from electrodes in a PEF treatment chamber, and a possible methodological approach to develop a mathematical model describing the phenomena of the migration of the electrode material into the treated biomaterial is also presented. Keywords

Pulsed electric field (PEF) • Electrochemical reactions • Ionic double layer • Metal release • Modeling

Introduction Pulsed electric field (PEF) is a well-established nonthermal technology utilized in several treatments in the area of medicine, biology, biotechnology, and food processing, which are based on the reversible or irreversible permeabilization of the membranes of biological cells (microbial, algae, plant, animal, or human) (Miklavčič et al. 2014; Golberg et al. 2016; Pakhomov et al. 2010). During a PEF treatment, the biological material (cell suspension, plant, animal, or human tissue) is placed in direct contact with the metal electrodes of a treatment chamber, operated either in batchwise or in continuous, and exposed to repetitive (Hz–kHz) very short (from several nanoseconds to several milliseconds) and intense (0.5–80 kV/cm) electric field (E) pulses provided by a pulse generator (Fig. 1a) (Mahnič-Kalamiza et al. 2014). The pulses commonly used in PEF treatments are unipolar or bipolar, with either exponential or square-wave shape (Figs. 1b). The application of electric pulses causes a drastic increase in the permeability of the cell membranes, referred to as electroporation or electropermeabilization, leading to a loss of their semipermeability, either temporarily (reversible electroporation) or permanently (irreversible electroporation) (Kotnik et al. 2012) (Transmembrane Voltage Induced by Applied Electric Fields). The extent of cell membrane permeabilization depends, besides on electric field strength, also on other electrical parameters such as total specific energy input (WT), or number of pulses (n), or treatment time (tPEF), pulse width (or pulse duration) (τp), pulse repetition frequency ( f ), and pulse shape. Treatment time refers to the number

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

a

3

Pulse generator

b V V I STATE

Pulse shapes

A

Unipolar square wave 0.00 0.00 0.00 0.00

kV A kW Hz

STOP

U (V)

Unipolar Exponential decay wave

UNIPOLAR BIPOLAR

τp OFF-ON TRIC

STOP RUN

Bipolar square wave

E

U (V)

U/e

τp

Bipolar Exponential decay wave

0

Time (ms or ms)

Cell subjected to electric field pulses

Fig. 1 (a) Simplified scheme of a PEF system; (b) typical pulse shapes used in PEF treatments

of pulses applied multiplied by the pulse width, which depends on the pulse shape (Fig. 1b) (Raso et al. 2016). In general, at high electric fields (>15 kV/cm), PEF treatments represent a suitable alternative to traditional thermal processing for the inactivation of alterative and pathogenic microorganisms and quality-related enzymes in different liquid food products, which retain, with only slight modification, their organoleptical, nutritional, and health-promoting attributes (Impact of Pulsed Electric Fields on Enzymes; Effect of Pulsed Electric Fields on Food Constituents). The application of low or moderate (0.1–10 kV/cm) electric fields has been also advantageously proposed to facilitate the mass transfer phenomena of target molecules, namely, drugs, solutes, water, juice, and bioactive compounds, through the cell membranes of different biomaterials, such as cell suspension, plant, animal, and human tissues. Thus, new methods for cancer treatments and gene therapy, as well as highly efficient food processing operations, such as extraction, juice expression, drying, and freezing, assisted by PEF technology, have been set up, and some of them are already in use in specific applications (Electrochemotherapy of Cutaneous Metastases; Gene Delivery by Electroporation In Vitro: Mechanisms; Pulsed Electric Fields as Pretreatment for Subsequent Food Process Operations) (Barba et al. 2015; Donsi et al. 2010; Golberg et al. 2016; Miklavčič et al. 2014; Pakhomov et al. 2010).

4

G. Pataro et al.

However, typical PEF treatment operating conditions cause some unavoidable electrochemical reactions due to the flow of high current through the electrode–solution interface of the treatment chamber. The extent of these reactions mostly depends on the PEF treatment intensity, which, in turn, is of key importance in applications such as the pasteurization of liquid foods. Indeed, these undesired electrode reactions should be minimized to guarantee safety and quality of food products and avoid electrode fouling and corrosion. For example, as far as food safety and quality aspects are concerned, the electrode reactions can determine changes of chemical structure and physical properties (e.g., pH) of the liquid in the vicinity of electrode surfaces; cause the production of toxic chemicals, namely, H2O2, HCl, and HClO; and induce the detachment of small particles of electrode material in the product, with a significant loss in flavor acceptance (Evrendilek et al. 2004; Meneses et al. 2011; Morren et al. 2003; Pataro et al. 2014). In addition, the reaction products formed are likely to react with other compounds in the liquid food, leading to the formation of toxic compounds even after the pulse treatment has been completed. Due to the occurrence of electrochemical reactions, a film of food particles at the electrode surfaces may form. This phenomenon is referred as electrode fouling. The deposit of material on the electrode surface during extended processing period is also responsible for several problems, such as local electric field distortion, arcing, contamination of the system, and, in some case, interruption of the flow of the product (Bushnell et al. 1995; Morren et al. 2003). Finally, due to corrosion, the electrodes can be severely damaged, and their surface roughness can increase due to metal release. This, in turn, can again cause local electric field distortion and arcing, drastically shortening the lifetime of the electrodes, thus reducing the technical feasibility of PEF technology (Gad and Jayaram 2012a; Pataro et al. 2014; Roodenburg et al. 2005a, b). Nevertheless, although electrochemical reactions and corrosion are well known in other fields, their importance in PEF treatment systems was studied in detail only recently (Table 1). In 1995, Bushnell et al. (1995), based only on a theoretical analysis, suggested a method to avoid or limit electrochemical reactions and fouling of electrodes in a PEF treatment system by removing all the residual charge at the electrode interface during the time elapsing between two consecutive pulses (zero net charge concept). Few years later, Morren et al. (2003) derived an equivalent circuit to account for electrochemical reactions in a PEF chamber and presented the first experimental results on the release of electrode material into liquid product undergoing a PEF treatment. Since then, several researchers have addressed the problem of electrochemical reactions in PEF treatment systems highlighting that electrode reactions are a very complex phenomena, whose extent depends on many factors, namely, processing parameters, design parameters, and treatment medium characteristics (Table 2). An adequate knowledge of the role played by these factors on the occurrence and extent of the electrochemical reactions at the electrode–solution interface of a PEF chamber is necessary to identify the proper treatment conditions allowing minimizing the undesired side effects of these reactions. Additionally, another important

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

5

Table 1 Summary of the published data on electrochemical phenomena in PEF treatment systems Type of study Theoretical

Treatment chamber –

Experimental

Batch parallel plate configuration with either aluminum or stainless steel electrodes

Treatment medium: Spinner minimum essential medium Treatment conditions: Unipolar/bipolar square pulses, E = 0–2 kV/cm, 1–8 pulses, τp = 100 μs to 1 ms

Theoretical/ experimental

Tubular configuration with stainless steel electrodes

Treatment medium: Aqueous NaCl solution (σ = 4 mS/cm) Treatment conditions: Sinusoidal waveform, 0.15 kV/ cm, 480 mA, f = 1 MHz to 0.001 Hz, 1.5 mL/s

Experimental

Colinear configuration with stainless steel (AISI 316) electrodes

Theoretical/ experimental

Colinear configuration with stainless

Treatment medium: Beer Treatment conditions: Bipolar square pulses, E = 41–44 kV/cm, f = 600 Hz, τp = 4 μs, tPEF = 175 s, flow rate = 1 mL/s Treatment medium: Aqueous NaCl solution (σ = 2.2

Treatment conditions –

Main results A method to reduce electrochemical reactions and the fouling of electrodes by removing the residual charge from one electrode during the discharge period was suggested Bipolar pulses reduced electrolytic contamination (Al3+ and Fe2+/Fe3+) with respect to unipolar pulses Concentrations of Fe2+/Fe3+ above 1.5 mM caused significant loss of viability of DC-3F cells; no effect on cell survival for Al3+ (0–2.5 mM) was detected Development of an equivalent circuit to model the behavior of the double-layer capacitor Fe, Cr, Ni, and Mn were dissolved in the treated liquid below a certain critical frequency (and above a certain pulse width) Fe, Cr, Zn, and Mn were released in the PEF-treated beer samples with significant loss in flavor acceptance and mouth feeling of the products

References Bushnell et al. (1995)

The amount of metal release was related to the transferred

Roodenburg et al. (2005a)

Kotnik et al. (2001)

Morren et al. (2003)

Evrendilek et al. (2004)

(continued)

6

G. Pataro et al.

Table 1 (continued) Type of study

Treatment chamber steel (AISI 316) electrodes

Colinear configuration with stainless steel (AISI 316) electrodes

Experimental

Treatment conditions

Main results

mS/cm) Treatment conditions: Unipolar and nonsymmetrical bipolar pulses, 30 kV/cm, 6.7 pulses, τp =1–5 μs, 100 mL/s Treatment medium: Orange juice (pH = 3.8; σ = 3 mS/cm) Treatment conditions: Nonsymmetrical bipolar pulse, 30 kV/ cm, 6.7 pulses, τp =1 μs, 100 mL/s

charge during the pulse treatment Lower transferred charge and the use of nonsymmetrical pulse shape reduced the release of Fe, Cr, Ni, and Mn

Batch parallel plate configuration with polished stainless steel electrodes

Treatment medium: 154 mM NaCl solution Treatment conditions: Unipolar exponential decay pulses, E = 4 kV/cm, 60–120 pulses, τp =20 μs

Continuous flow chamber made with two stainless steel (type 316) electrodes having opposing convex surface (biconcave processing zone)

Treatment medium: Aqueous NaCl/HCl solution (pH = 3.5–7.0, σ = 2 mS/cm) Treatment conditions: Unipolar exponential decay pulses, 10–30 kV/cm, WT = 0–1000 J/mL, f = 10 Hz, τp = 1.1–1.6 μs,

The metal concentrations in the PEF-treated juice that were found do not exceed the legislation values for fruit juices and the EU Drinking Water Directive Because of the metal release, the lifetime of the electrodes was estimated to be approximately 760 h The roughness of the anode surface increased in a manner depending to the total amount of electric charge passed through the unit area of the electrode No changes of the cathode surface were detected The concentration of dissolved Fe and Ni increased with the number of pulses and peak voltage applied Low-pH solutions resulted in higher concentration of metals

References

Roodenburg et al. (2005b)

Saulis et al. (2007)

Gad and Jayaram (2011)

(continued)

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

7

Table 1 (continued) Type of study

Treatment chamber

Theoretical/ experimental

Batch parallel electrode configuration with cylindrical stainless steel electrodes

Experimental

Continuous flow chamber made with two stainless steel (type 316) electrodes having opposing convex surface (biconcave processing zone) Continuous flow chamber made with two stainless steel (type 316) electrodes having opposing convex surface (biconcave processing zone)

Experimental

Treatment conditions flow rate = 1500 mL/min Treatment medium: Aqueous NaCl solution (pH = 7.0, σ = 1 mS/cm) PBS buffer (pH = 7.1, σ = 1.0 mS/cm) Treatment conditions: Unipolar exponential decay pulses, E = 10–25 kV/cm, WT = 3–208 kJ/kg, τp = 3.4 μs, f = 1–30 Hz Treatment medium: Skimmed milk (pH = 7.0, σ = 5.23 mS/cm) Treatment conditions: Exponential decay pulses, 30–40 kV/ cm, WT = 0–1000 J/mL, f = 21 Hz, τp = 1.1–1.6 μs

Treatment medium: Apple cider (pH = 3.8,σ = 2.78 mS/cm), beer (pH = 4.3, σ = 0.994 mS/cm), milk (pH = 6.9, σ = 5.52 mS/cm), orange juice (pH = 4.1, σ = 4.54 mS/cm) Treatment conditions: Exponential decay pulses, 40 kV/cm, WT = 0–1000 J/mL, f = 21 Hz

Main results

References

Electrode reactions induced pH shifts of up to 4.04 units close the electrodes The local change of pH at the anode (pH = 3.3) resulted in considerable local PPO inactivation. For the cathode (pH = 10.9) and the center (pH = 7) of the treatment chamber, no inactivation of the PPO was detected The amount of Fe and Ni released in the liquid medium increased with increasing the energy density, the field strength, and the pulse width No appreciable differences were detected for chromium

Meneses et al. (2011)

Highly conductive foods, such as milk and orange juice, experienced a higher rate of electrode material migration, which reduced the lifetime of the electrodes The effect of pH appeared to be insignificant

Gad and Jayaram 2012b

Gad and Jayaram (2012a)

(continued)

8

G. Pataro et al.

Table 1 (continued) Type of study Experimental

Treatment chamber Continuous flow parallel plate configuration with stainless steel (AISI 316 L) electrodes

Theoretical/ experimental

Continuous flow parallel plate configuration with stainless steel (AISI 316L) electrodes

Treatment conditions Treatment medium: Citrate–phosphate (McIlvaine) and Trizma-HCl buffer (pH 7, σ = 2 mS/cm) Treatment conditions: Unipolar exponential decay pulses, E = 12–31 kV/cm, WT = 20–100 J/mL, τp = 3.1 μs, flow rates = 2–4 L/h

Treatment medium: Citrate–phosphate (McIlvaine) and Trizma-HCl buffer (pH 7, σ = 2 mS/cm) Treatment conditions: Monopolar exponential decay pulses, E = 12–31 kV/cm, WT = 20–100 J/mL, τp = 3.1 μs, flow rates = 2–4 L/h

Main results The amount of dissolved Fe, Cr, and Ni increased upon increasing WT and decreasing E, as due to the corresponding change in the value of pulse repetition frequency The presence of halides in the treatment of medium (TrizmaHCl) accelerated the corrosion process at the electrode–solution interface A mathematical model describing the phenomenon of the metal release was developed and validated The spatial distribution of the metal concentration in the cross section of the treatment zone was nonhomogeneous due to the combination of hydrodynamic and kinetic effects

References Pataro et al. (2014)

Pataro et al. (2015a)

(continued)

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

9

Table 1 (continued) Type of study Theoretical/ experimental

Treatment chamber Continuous flow parallel plate configuration with stainless steel (316 L) electrodes

Treatment conditions Treatment medium: Citrate–phosphate (McIlvaine) and Trizma-HCl buffer (pH 7, σ = 2 mS/cm) Treatment conditions: Unipolar exponential decay pulses, E = 12–31 kV/cm, WT = 20–100 J/mL, τp = 3.1 μs, flow rates = 2–4 L/h

Main results Pulse frequency, rather than the applied voltage, is the main parameter affecting the occurrence of the metal release The presence of halides in the composition of the treatment medium (Trizma-HCl) markedly increased the faradaic current density, leading to an intense localized metal release from the electrode’s surface (corrosion)

References Pataro et al. (2015b)

Table 2 Main parameters affecting electrochemical reactions during PEF treatments Process parameters Electric field strength Total specific energy Pulsed width Polarity Frequency Pulse shape Flow rate Temperature

Design parameters Chamber configuration Electrode material Electrode area Electrode rugosity ... ... ... ...

Treatment medium properties Composition Electrical conductivity pH Halides ...

factor to be taken into account is that in PEF systems, the current densities occurring at the electrodes are of orders of magnitude higher than in traditional electrochemical systems. This raises the question whether the models commonly used to describe the electrochemical phenomena at the electrodes are applicable to PEF systems. To this purpose, numerical simulations specifically set up for PEF systems might significantly help to improve the understanding of fundamentals of the electrochemical phenomena occurring at the electrode–liquid interface of a PEF chamber, clarifying the effects of the main electrical parameters, chamber design

10

G. Pataro et al.

characteristics, and treatment medium composition on electrode corrosion or release of electrode’s materials (Pataro et al. 2015a, b).

Principles of Electrochemistry Basic principles of electrochemistry and related equations can be traced back before to the discovery of the electron by J. J. Thompson in 1893. In 1799, Alessandro Volta invented the first electrical battery, known as voltaic pile, built by alternating stacks of copper and zinc disks separated by paper soaked in acidic solutions (Ciobanu et al. 2007). Volta’s work stimulated other scientists to further investigations in this field. It is worth mentioning that, in 1800, William Nicholson and Anthony Carlisle used the current generated with a battery to decompose water into hydrogen and oxygen (electrolysis). In 1835, Michael Faraday coined the basic terminology of electrochemistry, namely, anode, cathode, electrode, electrolyte, and ions. More importantly, Faraday discovered the two laws of electrochemistry. The first law states that the amount of chemical changes or decomposition is proportional exactly to the quantity of electric charge passing through the electrolyte solution. The second law affirms that the amount of different substances deposited or dissolved by the same amount of electric charge is proportional to their chemical equivalent weights. Specifically, Faraday’s laws can be summarized by mi ¼

Qi Mi F ni

(1)

where Qi is the charge in Coulombs [C] involved in the deposition or dissolution of the chemical species i, mi is the total mass of the chemical species i deposited or dissolved at an electrode [g], F = 96485 C mol1 is the Faraday constant, Mi is the molar mass of the substance [g mol1], and ni is the valency number of ions of the substance (electrons transferred per ion). The ratio Mi/ni is the equivalent weight of the substance altered. These findings led ultimately to the development of the area of science nowadays known as electrochemistry, which is defined as the study of chemical reactions occurring at the interface between an electrode, usually a solid metal or a semiconductor, and an ionic conductor or electrolyte. These reactions, generally referred to as electrode processes or electrode reactions or electrochemical reactions, involve the transport of ions within the electrolyte, accounting for the electrical conductivity, or ions or electrons across the electrode–electrolyte interface resulting in the interfacial potential differences, which may trigger electrotransfer reactions at the electrodes. Chemical reactions where electrons are transferred directly between chemical species (e.g., molecules, atoms, or ions) changing their oxidation state (i.e., the fictitious charge that an atom would have if all bonds to atoms of different elements were 100 % ionic) are called oxidation–reduction or (redox) reactions.

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

11

A redox reaction can be seen as the simultaneous unwinding of two separate “half-reactions”: • Reduction half-reactions, which correspond to the gain of electrons from a chemical species resulting in a decrease of its oxidation state. The chemical species that gains electrons is called oxidizing agent or oxidant. • Oxidation half-reactions, which correspond to the loss of electrons from a chemical species resulting in an increase of its oxidation state. The chemical species that loses electrons is called reducing agent or reductant. Given two chemical species, namely, 1 and 2, the reduction and oxidation halfreactions can be expressed as follows: Reduction half-reaction Ox 1 þ ne ! Red 1

(2)

Red 2 ! Ox 2 þ ne

(3)

Oxidation half-reaction

where Ox1 and Red1 are, respectively, the oxidized and reduced form of the species 1; Ox2 and Red2 are, respectively, the oxidized and reduced form of the species 2; and n is the number of electrons (e) exchanged during the reaction (the same in both half-reactions). The overall redox reaction is then represented as Ox 1 þ Red 2 ! Red 1 þ Ox 2

(4)

Electrochemical Cells An electrochemical cell is a device capable of either generating electrical energy from chemical reactions or facilitating chemical reactions introducing electrical energy. It generally consists of two half-cells, each containing an electrode in contact with an electrolyte. The two half-cells may use the same electrolyte, or they may use different electrolytes, which are typically placed in ionic contact by a salt bridge. Electrodes can be made from any sufficiently conductive material, such as metals, semiconductors, graphite, and even conductive polymers. Current flows through the electrodes, which are connected by a conductor, via the movement of electrons, while through the electrolyte is carried by ions. In an electrochemical cell, species from one half-cell lose electrons (oxidation) to their electrode, while species from the other half-cell gain electrons (reduction) from

12

G. Pataro et al. Current flow

a

b

Power Supply

V Voltmeter

e-

e-

e- flow

e-

e-

Salt bridge

Anode

X

Current flow

X+

Cathode

Cathode

Anode

Y-(aq)

X

Y X-(aq)

e- flow

Ma-(aq)

Y

Aa+(aq)

Y

Oxidation Reduction X Xa+(aq) + ne- Ya+(aq) + neY

Oxidation A+ neAa-(aq)

Reduction M Ma+(aq) + ne-

Fig. 2 Electrochemical cells. (a) Galvanic cell; (b) electrolytic cell

their electrode. Michael Faraday defined the anode of a cell as the electrode where oxidation occurs and the cathode as the electrode where the reduction takes place. Electrochemical cells are usually classified as either galvanic or electrolytic (Fig. 2). In galvanic cells, reactions occur spontaneously at the electrode–electrolyte interfaces. These cells convert chemical energy in electric energy and are the components of batteries. In electrolytic cells, reactions are forced to occur at the electrode–electrolyte interfaces utilizing an external electrical power source connected to both electrodes. Electric energy is converted in chemical energy of the products of the electrode reactions. Electrolytic cells are used in most cases to decompose chemical compounds, and the process is referred to as electrolysis (e.g., electrolysis of water). Finally, a third type of electrochemical cells, such as rechargeable battery, is also available which acts as galvanic cells when discharged (converting chemical energy into electrical energy) and as electrolytic cells when charged (converting electrical energy into chemical energy).

Electrode Potentials Current and potential are the two electrical variables of major interest in electrochemical cells. Current i), expressed in amperes [A], is related to the electrode reaction rate and is equal to the variation of the charge with time: i¼

dQ dt

(5)

Reduction half-reaction C12(g) + 2eO2(g) + 4H+(aq)+2eFe3+(aq) + eO2(g) + 2H+(aq)+2e2H+(aq) + 2eFe3+(aq) + 3eNi2+(aq) + 2eCo2+(aq) + 2eFe2+(aq) + 2eCr3+(aq) + 3eZn2+(aq) + 2e2H2O(1)+ 2eCr2+(aq) + 2eA13+(aq)+ 3eMn2+(aq)+ 2eNa+(aq)+ e-

2C1-(aq) 2H2O(1) Fe2+(aq) H2O2(aq) H2(g) Fe(s) Ni(s) Co(s) Fe(s) Cr(s) Zn(s) H2(g) + 2OH-(aq) Cr(s) A1(s) Mn(s) Na(s)

E∞red(V) 1.36 1.23 0.77 0.70 0 -0.04 -0.26 -0.28 -0.45 -0.74 -0.76 -0.83 -0.91

13

Oxidizing Power

Reducing Power

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

-1.66 -2.37 -2.71

Fig. 3 Reduction half-reactions with their standard potentials with respect to SHE (Data from Arning and Minteer 2007)

where t, expressed in seconds [s], is the time and Q, expressed in Coulombs [C], is the total charge passed through the cell. Cell potential (Ecell), expressed in volts [V], is a measure of the energy of the cell reaction [1 V = 1 J C1]. The cell potential is the difference in potential of the two half-cells. Each half-cell has a characteristic potential, which depends on the couple electrode–electrolyte, as well as on temperature, pressure, pH, and concentration of the electrolyte. The standard potential (E ) is the maximum potential or equilibrium potential (Eeq), which applies to a half-cell when all the reactants are present at unit activity, that is, when the concentration of the species in the solution is almost 1 M, at a temperature of 298.15 K and at a partial pressure of 100 kPa. The standard potential data are usually listed as reduction potentials (E red) with respect to the standard hydrogen electrode (SHE), to which is arbitrarily assigned a half-cell potential equal to zero (Fig. 3). The reduction potential, E red, is a measure of the tendency of a chemical species to gain electrons and, thereby, to be reduced. On the contrary, the oxidation potential, E ox, which is just the negative of the reduction potential, is a measure of the tendency of a chemical species to lose electrons and, thereby, to be oxidized. However, the reactions are reversible, and the role of a certain electrode in a cell depends on the relative oxidation–reduction potential of both electrodes. Generally, the reaction with the more negative half-cell reduction potential occurs at the anode as oxidation. The reaction with the more positive half-cell reduction potential occurs at the cathode as reduction. The cell potential is then calculated as the difference of

14

G. Pataro et al.

the reduction potentials of cathode and anode or, alternatively, as the sum of the reduction potential for the cathode and the oxidation potential for the anode. Eocell ¼ Eored ðcathodeÞ  Eored ðanodeÞ ¼ Eored ðcathodeÞ þ Eoox ðanodeÞ

(6)

For nonstandard conditions (e.g., difference in temperature, pH, concentration, and pressure), the cell potential will deviate from the standard potential. The actual halfcell reduction potential Ered is a function of the solution concentrations and is related to these latter and to the standard potential E red by the Nernst equation: Ered ¼ Eored þ

RT ½Ox  ln nF ½Red 

(7)

where R is the universal gas constant (8.3145 J mol1 K1); T is the absolute temperature, expressed in Kelvin; and the terms [Ox] and [Red] represent, respectively, the concentrations of the oxidized and reduced species for the generalized reduction half-reaction: Ox þ ne Ð Red

(8)

Electrochemical Reactions Electrochemical Behavior of a PEF Treatment Chamber A PEF chamber, which consists of two metal electrodes placed in direct contact with an electrolytic solution and electrically connected to a pulse generator (Fig. 1a), acts as an electrochemical cell. Specifically, during its operation, a PEF chamber initially made with equal electrodes in contact with an electrolyte with homogeneous concentration can behave in three different ways. Before starting the pulse treatment (i.e., without externally applied voltage), no chemical reactions take place. During pulse delivery, the PEF chamber acts as an electrolytic cell where an external source of electrical energy is responsible for driving electrical current through the cell. This electrical current has an electronic nature (electrons) in the conductive solid electrodes and an ionic nature (ions) in the conductive liquid electrolyte. As a result, electrochemical reactions take place at the electrode–electrolyte interfaces to satisfy the condition of electric current continuity. After pulsing, due to concentration gradients around the electrodes or to surface material changes, the behavior of cell may shift from electrolytic to galvanic. Thus, if the treatment chamber is not dried up, chemical reactions are likely to spontaneously proceed even with no external voltage applied.

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

15

Ionic Double Layer As in any electrochemical cell, at each electrode–solution interface of a PEF treatment chamber, an ionic double layer immediately develops in order to compensate any charge imbalance on the electrode surface, which can occur even if no external voltage is applied (electroneutrality principle). This layer, which consists of charged particles and/or of orientated dipoles, behaves as an electrical capacitor and is often called double-layer capacitor (Fig. 4). It is typically described as a form of different sub-layers. The first layer comprises solvent molecules and (non-hydrated) ions adsorbed directly onto the electrode surface due to several chemical interactions. The plane formed by the electrical centers of this layer is termed inner Helmholtz plane (IHP), which lies χ1 than the electrode surface. The second layer consists of hydrated ions, which are typically located at a distance from the electrode surface no closer than the radius of the inner solvent coordination sphere (χ2). The plane formed by these electrical charges is termed outer Helmholtz plane (OHP). The ionic species in this plane are less compact than the IHP and are distributed in a region called the diffuse layer, which is extended up to the bulk electrolyte region where there is no separation of charge (Ciobanu et al. 2007). The formation of the double layer causes a potential drop (Δϕdl) between the surface and any point of the mass of the electrolyte solution. This potential drop is of the order of magnitude of mV and depends on surface charge and on thickness of the double layer. As shown in Fig. 4, the voltage drop between the electrode and the OHP is linear and tends asymptotically to zero at the imaginary boundary of the double layer. To simplify the analysis of electrochemical kinetics, the contribution of the diffuse layer can be neglected, and it can be assumed that the ionic double layer extends up to the OHP. Hence, the potential drop at the electrode–electrolyte interface at equilibrium (Δϕdle), when there is no net current flow, can be assumed to be equal to the reduction potential of the half-cell reaction (Eored). When a potential difference between the two electrodes is applied, the charge buildup across the double layer takes place (Bockris et al. 2002; Gileadi 1993). In this case, the thickness of the ionic layers at the interface increases leading to an increase of the potential drop Δϕdl. As far as the voltage across the double layer remains below the typical threshold voltage (Uth) of the reaction potential of electrode material (~1–2 V), no electrochemical reactions occur, except some low-level reactions due to the exchange current, as specified later in this chapter. If a high enough voltage is applied for a long enough time to let the voltage across the double-layer capacitor exceed the threshold value, in order to preserve the charge conservation principle, two independent electrochemical (faradaic) half-reactions will occur at the two electrodes: oxidation reactions will take place at the electrode surface behaving as anode (high voltage electrode) and reduction reactions at the electrode surface behaving as cathode (grounded electrode) (Morren et al. 2003). If during PEF treatment unipolar pulses are applied, the electrode connected to high voltage will behave as anode, while the grounded electrode will behave as cathode. If instead bipolar pulses are applied, the cathode and anode interchange places according to the pulse repetition frequency. Therefore, in this latter case, the

Diffuse layer

Bulk electrolyte

OHP

f

Electrode

Fig. 4 Schematic diagram of an ionic double layer at the electrode–electrolyte interface. IHP inner Helmholtz plane, OHP outer Helmholtz plane. χ 1 is the distance between the plane formed by the electrical centers of IHP and the electrode surface; χ 2 is the minimum distance between the solvated ions of the second layer and the electrode surface. Continuous blue line represents the potential profile (ϕ) across the ionic double layer

G. Pataro et al.

IHP

16

Potential (φ)

x1 x2

Solvated Adsorbed (+) ion (+) ion

Distance

Water dipole (arrow points to + charge on dipole)

Non-solvated (-) ion

electrochemical reactions of oxidation and reduction occur alternatively at the same electrode site. The current resulting from this change in oxidation state is referred to as faradaic current (if) because it obeys Faraday’s law. This faradaic current is a direct measure for the rate of the redox reaction. The current necessary to charge the double layer up to the threshold voltage, instead, is called charging current (ic), being a non-faradaic current since it does not obey Faraday’s law, due to the absence of electrochemical reactions associated (Morren et al. 2003). Quite complicated electrode reactions occur consisting of a chain of chemical reactions. They involve mass transport of electro-active species to the electrode, electron transfer across the electrode interface, and the transport of the reaction products back to the solution. Mass transport and charge transfer are two consecutive processes whose rate depends, besides the applied potential, on electrode material, electrolyte, pulse repetition frequency, pulse polarity, temperature, and pH, among others. The slower of the two processes is, therefore, the limiting step and determines the overall rate. For example, the rate of charge transfer is strongly related to the applied potential. When the applied potential is low, charge transfer is low and it is therefore the limiting mechanism. On the contrary, when the potential applied is increased, charge transfer is faster and mass transport is then becoming the limiting mechanism. The mass transfer mechanisms occurring are by diffusion (movements of species due to concentration

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

17

gradients), migration (movement of charged particles due to an electric field), and convection (movement of species induced by stirring, flow, or density gradient) (Wang 2000).

Equivalent Electrical Circuit of an Electrochemical Cell To obtain a relationship between an external electrical perturbation at the terminals of a PEF chamber and the response of the double-layer interface, it is necessary to set up an equivalent electrical circuit (Computational Approach for Electrical Analysis of Biological Tissue Using the Equivalent Circuit Model). In Fig. 5, equivalent electrical circuits for two electrode–electrolyte interfaces are shown for different conditions. In these circuits Rs is the bulk resistance of the solution and Cdl is the double-layer capacitance, while the series of Rct and Zw is the faradaic impedance used to model the faradaic processes occurring at the electrode–solution interface, with Rct the charge transfer resistance and Zw the Warburg impedance used to model the mass transfer limitations (Bockris et al. 2002; Gileadi 1993). These models do not take into account some peculiar factors such as the heterogeneity of electrode surface and the eventual formation of reaction intermediates absorbed at the electrode surface. Moreover, the circuit is nonlinear being Cdl, Rct, and Zw depended on potential (Gileadi 1993). When an external voltage is applied, and the threshold voltage (Uth) at which the reactions typically start is not yet reached, the model is that of Fig. 5a. In this case, the interface is said to be ideally polarizable and only a capacitive (non-faradaic) current is flowing in Cdl. When the threshold voltage (Uth) is reached, a faradaic current starts passing through Rct in the branch of the circuit parallel to Cdl, and the circuits appearing in Fig. 5b or c should be considered, according to the controlling limiting step, namely, charge transfer and mass transfer. The time highlighting for how fast the double layer becomes charged is called threshold time tth and can be evaluated according to Eq. 9: tth ¼

C0dl  U0th i

(9)

where C0dl = (1/2)∙Cdl is the value of the equivalent capacitance and U0th = 2∙Uth the voltage at the electrode–liquid interface (Morren et al. 2003). From Eq. 9, it can be seen that, for a given electrode reaction, the threshold time depends mainly on the current through the chamber and the value of the double-layer capacitance. Thus, PEF treatment chamber configurations featuring high intrinsic electrical resistance (e.g., colinear configuration), which, therefore, operates at a relatively low current, may limit the occurrence of electrochemical reactions. Similarly, the use of high capacitance electrode materials, such as titanium (Cdl = 50 μF/ cm2), glassy carbon (Cdl = 260 μF/cm2), or dimensionally stable anode (DSA)

18

G. Pataro et al.

Fig. 5 Equivalent circuit for the treatment chamber with interface. The two ends of the circuit represent, respectively, the high voltage (left) and the ground (right) electrode. (a) No faradaic reactions; (b) faradaic reactions and charge transfer limitation; (c) faradaic reactions and mass transfer limitation (Adapted from Pataro et al. 2015a, with permission)

E

Anodic interface

Cd1 HV o

Cathodic interface Electrolyte

ic

Cd1

i

ic (a)

Rs Cd1 HV o

ic if

Rct Cd1 HV o

ic

Cd1

i Rs

ic if

Rct Cd1

i

ic

if Rct

Zw

Rs

(b)

if

(c)

Rct Zw

(Cdl = 2000 μF/cm2), instead of the most commonly used stainless steel electrodes (Cdl = 35 μF/cm2), may limit the occurrence of electrode reactions by slowing down the charging process of the double-layer capacitance (Amatore et al. 1998; GóngoraNieto et al. 2002).

Metal Release Metal Release During PEF Treatment of Food Products The migration or release of electrode material into the liquid food under processing is one of the main consequences of the electrochemical reactions occurring at the electrode–solution interface of a PEF treatment chamber. The dissolution of the

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

19

Fig. 6 Corroded electrodes of a PEF treatment chamber after 3 h of operation (21 kV/cm, WT = 100 J/mL) with Trizma-HCl buffer (pH = 7, σ = 2 mS/cm) (Adapted from Pataro et al. 2014, with permission)

anode material due to oxidation of the metal of the electrode is responsible of this phenomenon. The extent of metal release during pulse treatments must be minimized due to the undesired effects occurring, which might limit the exploitation of PEF technology in industrial scale, especially as far as applications for food preservation are concerned. First of all, type and quantity of materials, which are eventually released from electrodes and can be detected in PEF-processed foods, must comply with the health safety rules approved by the food regulatory authorities before being able to introduce the PEF-treated products in the market. Moreover, the organoleptic quality of PEF-processed foods, especially the flavor of the product, should satisfy the requirements of consumers and compete favorably with that of products processed with conventional preservation methods. The presence of any metallic mouth feeling may play against consumers’ acceptance of PEF technology (Evrendilek et al. 2004; Gad and Jayaram 2012a; Roodenburg et al. 2005b; Pataro et al. 2014). Finally, metal release may cause serious damages to the electrode surface (corrosion), drastically lowering the performance of PEF treatment systems and reducing the electrode lifetime (Fig. 6) (Pataro et al. 2014; Roodenburg et al. 2005a, b; Saulis et al. 2007).

Modeling of Metal Release Phenomena Case Study: Metal Release in a PEF Treatment Chamber The setup of a mathematical model describing the electrochemical phenomena occurring at the electrode–solution interface of a PEF treatment chamber is very complex, due to the numerous factors playing an important role in the occurrence of the electrode reactions during pulse treatment (Table 2). Hereafter, a general simplified approach, which can be proposed to develop a mathematical model describing the phenomenon of metal release from the electrodes of a continuous flow PEF treatment chamber, will be illustrated, according to the methodology described by Pataro et al. (2015a, b) and schematically illustrated in Fig. 7.

20

G. Pataro et al.

Fig. 7 Schematic of the methodological approach to set up a model describing the phenomenon of metal release from the electrodes of a PEF chamber

Methodological Approach Figure 8 shows the 3D view of two typical continuous flow PEF treatment chamber configurations: parallel plate (Fig. 8a) and colinear configuration (Fig. 8b). In both model chambers, the high voltage (HV) and grounded (GR) metal electrode are separated by means of a spacer insulator, which bounds a parallelepiped treatment zone (in the X–Y–Z space) in the parallel plate geometry and a cylindrical treatment zone (in the R–Z space) in colinear configuration, through which the treatment medium flows along the Z axis. Regardless of the chamber geometry, the phenomenon of metal release from the electrodes of a PEF treatment chamber into the liquid medium can be regarded as a mass transport process in an electrochemical system, described with Eq. 9: @Ci þ ∇  Ni ¼ r i @t

(10)

Modeling of Electrochemical Reactions During Pulsed Electric Field. . . Fig. 8 3D view of (a) continuous flow parallel plate treatment chamber (Adapted from Pataro et al. 2015a with permission from Elsevier) and (b) colinear treatment chamber

a

21

HV

Anode

Cathode tion

Y

GR

irec

wd

Flo Z

Insulator

X

b

GR

HV r

Flow direction Z

Anode

Insulator

Cathode

where Ci is the concentration of the ionic species i [mol m3] in the treatment medium, ri is the reaction rate [mo m3s1] of the species i, and Ni is the molar flux of the charged species i [mol m2s1], which, in turn, is the sum of three contributions, namely, convection, migration, and diffusion, and is given by the Nernst–Planck equation: Ni ¼ Di ∇Ci  zi um, i FCi ∇φl þ Ci v

(11)

being Di the diffusion coefficient of the species i [m2 s1], zi the charge number of the ionic species i [unitless], φl the electric potential through the electrolyte [V], v the velocity vector [m s1], and um,i the ionic mobility of the charged species i [mols kg1]. To solve Eq. 10, the following initial (I.C.) and boundary (B.C.) conditions are set: I.C.: the initial concentration of each metal species i (e.g., i = Fe, Cr, Ni) in the solution within the PEF treatment zone is zero (or equals to any other initial value Co) Ci ðt ¼ 0Þ ¼ 0

(12)

B.C.1: At the electrode–solution interface, the molar flux Ni of the generic charged species, i, is caused by the faradaic current n  Ni ¼ Ri ¼

vi ðn  ji Þ ne F

(13)

22

G. Pataro et al.

where ji is the faradaic current density associated with the ionic species i and n is the direction normal to the boundary. B.C.2: No flux of charged species at the electrode–insulator and solution–insulator interface occurs: n  Ni ¼ 0

(14)

B.C.3: At the inlet section of the PEF chamber, the concentration of metal ions in the solution is assumed equal to zero (or equals to any other inlet value Cin). Moreover, other information and conditions to be included in order to define completely the model are as follows: • • • •

The electroneutrality of the electrolyte solution The continuity equation of the electric current in the electrolyte solution The continuity equation of the electric current passing through the electrodes The electrical insulation condition at both the electrolyte–insulator and electrode–insulator • Physical and chemical–physical properties of all materials (insulator, electrodes, treatment medium, or electrolyte solution) • Characteristics of the electric pulses generated by the pulse generator (Fig. 1b) As shown in Fig. 7, to quantify the metal release from the electrodes, Eq. 10 should be solved by coupling with the equations of conservation of mass and momentum, primary current, and secondary current distribution.

Fluid Dynamics The distribution of the velocity field v of the liquid product inside the treatment chamber can affect the mass transfer of the metals released from the anode surface in the treatment medium. It can be evaluated by solving the continuity equation (Eq. 15), which accounts for the conservation of mass, and the Navier–Stokes equation (Eq. 16), which accounts for the conservation of momentum: @ρ þ ∇  ðρvÞ ¼ 0 @t  h  i 2 ¼ Dv ¼ ∇P þ ∇  μ ð∇vÞ þ ð∇vÞT þ μð∇  vÞ δ þ ρg ρ Dt 3

(15) (16)

where ρ is the density of the fluid, P is the pressure, μ is the dynamic viscosity, and g is the gravitational acceleration vector. A no-slip condition (v = 0) can be assumed at the fluid–solid boundary interfaces.

Primary Current Distribution The primary current defines the electric potential distribution, φl, in the treatment solution regulating the migration of charged species. Assuming that the distribution

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

23

of the primary current is completely controlled by the resistivity of the electrolytic solution in contact with the electrodes, the equivalent electric circuit representing the behavior of the electrochemical cell is that sketched in Fig. 5a, where the capacitors are assumed to be shortcut circuits. Based on the charge conservation principle, the governing equation for the electric potential U applied at the terminals of the PEF chamber can be written as follows: ∇  J ¼ ∇  ½σ ðT Þ∇U  ¼ 0

(17)

being J the primary current density vector. The relation between the electric field E and the electric potential is as follows: E ¼ ∇U

(18)

The time dependence of the potential U(t) of the high voltage electrode is that imposed by the pulse generator (Fig. 1b), while the other electrode is grounded, i.e., the potential at the boundary surfaces is zero (U = 0). Finally, as initial condition, a zero potential throughout the treatment chamber is considered.

Secondary Current Distribution The secondary current distribution determines the faradaic current density associated with each metallic species ji at the electrode–solution interface, which is a direct measure of the rate of the electrode reactions (ri). It is controlled by both the resistivity of the treatment medium and the faradaic impedance (Fig. 4b or c). The overall current density j flowing across the electrode–solution interface is the sum of the anodic (ja) and the cathodic (jc) current densities: j ¼ ja þ jc

(19)

When no external voltage is applied across the electrodes of the PEF chamber, an equilibrium potential Eeq is reached, which represents the reduction potential Eored of the half-cell reaction, Ox + ne$Red. No net current flow is occurring because the system contains a cathodic current density, given from the half-cell reaction Ox + ne ! Red, balanced by an equal and opposite anodic current density, given from the half-cell reaction Red ! Ox + ne:      jo j¼jja Eeq j¼jjc Eeq 

(20)

being jo the exchange current density, which represents the kinetics of a half-cell reaction at the equilibrium potential (Eeq  Δϕdle), when the so-called overpotential, η (V), is zero. The overpotential is defined by Eq. 21: η ¼ Δ∅dl  Δ∅dle

(21)

24

G. Pataro et al.

Fig. 9 Tafel plot: qualitative behavior for anodic and cathodic process. j is the electrode current density (A/m2); jo is the exchange current density (A/m2); and η is the overpotential (V)

Thus, η indicates the deviation of the potential drop at the electrode–solution interface, Δϕdl, from the equilibrium potential Eeq and represents the driving force for a non-spontaneous half-cell reaction to occur. The sign of η is positive for anodic reactions and negative for cathodic reactions. For each (anodic or cathodic) electrode reaction, the value jo can be derived experimentally from Tafel plots (Fig. 9), which describe for an electrolytic process the dependence of current density from overpotential. This plot shows an anodic branch for positive overpotentials, where η becomes more positive with increasing the current density at the anode, and a cathodic branch for negative overpotentials, where η becomes more negative with increasing the current density at the cathode (Ciobanu et al. 2007). When an external voltage is applied to such an extent such that an overpotential is generated, a net faradaic current density develops because of the occurrence of non-spontaneous half-cell reactions. In that case, the net faradaic current density associated with each metallic element i composing the electrode material ( ji,net) can be calculated by the current density–overpotential equation, known as generalized Butler–Volmer equation: ji,net ¼ jo

   

COx,s αnF CRed,s ð1  αÞnF η  η exp  exp RT RT COx,b CRed,b

(22)

where A [cm2] is the area of the electrode in contact with the electrolyte solution; COx and CRed are the concentrations of the oxidized and reduced species, respectively, at either the surface (subscript s) or at the bulk (subscript b); and α is the charge transfer coefficient, which is a dimensionless parameter with values between 0 and 1 (Ciobanu et al. 2007). In Eq. 22, the first and second terms on the right-hand side represent the rates of the cathodic and anodic parts of the half-cell reaction Ox + ne $ Red. In a well-stirred cell, it can be assumed that the concentrations of the bulk electrolyte and the electrolyte at the electrode surface are equal (i.e., COx,s  COx,b and CRed,s  CRed,b), and the current–overpotential equation (21) reduces to the more familiar Butler–Volmer equation:

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

   

αnF ð1  αÞnF η  exp η j ¼ jo  exp  RT RT

25



(23)

This equation describes the relationship between current density and overpotential when mass transfer limitation is disregarded and the rate of the electrode reaction is entirely controlled by charge transfer (Fig. 5b). It is worth noting that even though the Butler–Volmer equation was derived for one half-cell reaction, it is applicable to all reactions taking place at the anode and the cathode. To calculate the values of η to be introduced in Eqs. 22 and 23, the evaluation of the actual potential drop Δϕdl across the electronic double layer is required, which can be estimated with Eq. 24 obtained by solving the equivalent circuit of Fig. 5b for ideally polarizable electrode–solution interface (Fig. 5a). ΔV Cdl ðtÞ ¼

1 qðtÞ  2 Ceq

(24)

being Ceq = Cdl/2 the equivalent capacitance of the two double-layer capacitors (Stern 1924) and q(t) the charge accumulated on Ceq at the time t = tk after the application of the k-th electric pulse, which can be estimated through an iterative process (Pataro et al. 2015b). The transient of the double-layer potential drop (Eq. 24) induces a transient in the net faradaic current density for each metallic element (Eqs. 22 and 23). From these relationships it is possible to evaluate the average net faradaic current density for k each metallic species i ( ji,av ) during the k-th pulse as ð tk jki,av

¼

tk1

jki dt

ðtk  tk1 Þ

(25)

being the time interval (tktk1) the time occurring between two consecutive pulses applied and represents the reciprocal of the pulse frequency f.

Model Results The set of partial differential equations introduced in the previous paragraphs, namely, the coupled equations of mass and momentum conservation, primary current, secondary current, and mass transfer, with their relative initial and boundary conditions, can be solved using a finite element method (FEM)-based commercial software. From the simulation, it is possible to achieve, for different processing conditions, the dynamics of the spatial distribution of metal concentration in the treatment medium in each cross section of the treatment chamber. As an example, Fig. 10 shows a typical predicted spatial distribution of iron concentration in TrizmaHCl buffer solution (pH = 7, σ = 2 mS/cm) in the cross section at the outlet of a parallel plate PEF chamber (Fig. 8a) at different processing times (U = 5.6 kV, E = 21 kV/cm, WT = 60 kJ/kg, τp = 3.1 μs, f = 21 Hz, Reynolds number Re = 245). The dynamic of the metal release from the anodic surface can be

26

G. Pataro et al. 3.2381x104 4

(ppb) x 10

3

HV

HV

HV

2

4

6

2.5 2 1.5 1 0.5

0

Time (s)

0 0.8367

Fig. 10 Predicted distribution of the iron concentration in the outlet cross section of a parallel plate PEF treatment chamber as a function of the processing time. HV – high voltage (Adapted from Pataro et al. 2015a, with permission from Elsevier)

observed, as well as the characteristic nonhomogeneous spatial distribution of metal concentration in the treatment zone that, for the specific conditions of simulation adopted, is likely ascribable to the hydrodynamic effect (laminar flow) combined to the kinetic effect of metal release at the anodic interface (Pataro et al. 2015a). Moreover, by integrating the spatial distributions of the concentration of metallic elements i) released from the electrodes (Fig. 10), it is possible to evaluate, for each processing time, the bulk concentrations of metals Ci,bulk, at each cross section of the treatment chamber: ðð Ci v  dS Ci ,bulk ¼ ðð v  dS

(26)

where dS is the differential surface vector normal to the direction of the flow. For example, in the most common case of electrodes made in stainless steel AISI316L, the main metallic elements, which are expected to be released from the electrode into the electrolyte solution, are iron, chromium, and nickel. Figure 11 shows a typical release kinetics of these metallic elements in Trizma-HCl buffer solution (pH = 7, σ = 2 mS/cm) evaluated at the outlet cross section of a parallel plate treatment chamber. From this graph it can be seen that, regardless of the metal species, the amount of metal released increases with time and a stationary value is reached after almost 3 s. However, although a similar behavior can be observed for all the metallic

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

27

Fig. 11 Predicted Ci,bulk profiles of Fe, Cr, and Ni in Trizma-HCl buffer solution at the outlet cross section of the parallel plate treatment chamber as a function of the processing time. Simulation conditions: U = 5.6 kV; E = 21 kV/cm, WT = 60 kJ/kg, τp = 3.1 μs, f = 21 Hz, Re = 245 (Adapted from Pataro et al. 2015a, with permission from Elsevier)

elements considered, the final concentrations reached differ, according to the composition of the stainless steel electrodes, as well as the specific release kinetic of each metallic element. Finally, numerical simulations can be also used to get insights on the effect of the different processing parameters determining the extent of the electrode reactions. As an example, Fig. 12 shows a typical release kinetics of iron in Trizma-HCl buffer solution (pH = 7, σ = 2 mS/cm) as a function of either the applied voltage (electric field strength) applied when the energy input is kept constant (Fig. 12a) or the total specific energy input at a fixed applied voltage (electric field strength) (Fig. 12b).

Conclusions Electrochemical reactions unavoidably occur during PEF treatment especially for applications (e.g., microbial inactivation) involving the exchange of large amount of charge at the interface between electrodes and biological material (e.g., cell suspensions, liquid foods). These reactions are undesired and must be minimized, due to serious problems that they may cause, such as the release of metals from the electrodes of the PEF chamber. As a consequence of the occurrence of this phenomenon, loss of cell viability, contamination of treated products, formation of toxic compounds, and fouling and corrosion of the electrode surfaces reducing their lifetime are likely to take place. Knowledge of basic theory of electrochemistry, including terminology, fundamental principles, and equations, as well as of electrochemical cells, are essential for

28

G. Pataro et al.

Fig. 12 Predicted Ci,bulk profiles of Fe in Trizma-HCl buffer solution at the outlet cross section of a parallel plate treatment chamber as a function of the processing time. Simulation conditions: (a) WT = 60 kJ/kg, U = 3.0, 5.6, and 7.6 kV, τp = 3.1 μs, f = 59, 21, and 11 Hz, Re = 245; (b) U = 5.6 kV, E = 21 kV/cm, τp = 3.1 μs, f = 7, 21, and 35 Hz, Re = 245 (Adapted from Pataro et al. 2015a, with permission from Elsevier)

the comprehension of the complex phenomena occurring at the electrode–electrolyte interface of an electrochemical cell, such as a PEF treatment chamber. Moreover, the understanding of the role played by different processing factors, namely, applied voltage, current magnitude or current density, pulse repetition frequency, pulse duration, pulse shape, and fluid composition in affecting the value of the doublelayer capacitor and the rate of its charging process, as well as the value of the overpotential necessary to start the electrochemical reactions, is of utmost importance. Indeed, on the basis of the fundamental theory, it has been demonstrated that

Modeling of Electrochemical Reactions During Pulsed Electric Field. . .

29

the cumulative buildup of charge on the double-layer capacitor must be avoided. This might be done, for example, by selecting electrode material with high capacitance, removing the charge after each pulse, applying bipolar pulses, or applying monopolar pulses at relatively low pulse frequency. However, due to the large number of parameters affecting the extent of electrode reactions, the definition of optimal treatment conditions, as well as the selection of electrode geometry and electrode material able to minimize the occurrence of undesired electrochemical reactions, requires the development of a mathematical model specifically set up for PEF treatment systems. To this purpose, a possible methodological approach to model the phenomena of metal release from the electrodes of a PEF chamber has been described in the last part of this chapter. This approach could be used not only to deeply understand the fundamentals of the electrochemical phenomena taking place at the electrode–solution interface of a PEF chamber but also to clarify the role played by the main relevant design and processing parameters, in order to be able to select optimal PEF treatments, which will allow to prevent or, at least, minimize the occurrence of the undesired electrode reactions.

References Amatore C, Berthou M, Hebert S (1998) Fundamental principles of electrochemical ohmic heating of solutions. J Electroanal Chem 457:191–203 Arning MD, Minteer SD (2007) Electrode potentials. In: Zoski CG (ed) Handbook of electrochemistry. Elsevier, Amsterdam, pp 813–827 Barba FJ, Parniakov O, Pereira SA, Wiktor A, Grimi N, Boussetta N, Saraiva JA, Raso J, MartinBelloso O, Witrowa-Rajchert D, Lebovka N, Vorobiev E (2015) Current applications and new opportunities for the use of pulsed electric fields in food science and industry. Food Res Int 77:773–798 Bockris JO’M, Reddy AKN, Gamboa-Aldeco M (2002) The electrified interface. In: Bockris JO’M, Reddy AKN, Gamboa-Aldeco M (eds) Modern electrochemistry 2A. Foudamentals of electrodics. Kluwer, New York, pp 771–1015 Bushnell AH, Clark RW, Dunn JE, Lloyd SW (1995) Prevention of electrochemical and electrophoretic effects in high strength-electric-field pumpable-food-product treatment systems. US Patent, 5(447), 733. Ciobanu M, Wilburn JP, Krim ML, Cliffel DE (2007) Fundamentals. In: Zoski CG (ed) Handbook of electrochemistry. Elsevier, Amsterdam, pp 3–29 Donsì F, Ferrari G, Pataro G (2010) Applications of Pulsed Electric Field Treatments for the Enhancement of Mass Transfer from Vegetable Tissue. Food Eng Rev, doi 10.1007/s12393010-9015-3. Evrendilek GA, Dantzer S, Li WR, Zhang QH (2004) Pulsed electric field processing of beer: microbial, sensory, and quality analyses. J Food Sci 69:228–232 Gad A, Jayaram SH (2011) Electrode material migration during pulsed electric field (PEF) treatment. In: Conference ESA annual meeting on electrostatics 2011, Cleveland. pp 1–9. http://www.electrostatics.org/esa2011proceedings.html Gad A, Jayaram SH (2012a) Effect of electric pulse parameters on releasing metallic particles from stainless steel electrodes during pulsed electric field processing of milk. In: Conference electrostatics joint conference 2012. Cambridge, ON, pp 1–7. http://www.electrostatics.org/ esa2012proceedings.html

30

G. Pataro et al.

Gad A, Jayaram SH (2012b) Effect of food composition and pH on electrode material migration during PEF application. In: Conference BFE 2012, international conference bio & food electrotechnologies. Salerno, pp 49–52. Gileadi E (1993) Electrode kinetics for chemists, chemical engineers, and material scientists. VCH, New York Golberg A, Sack M, Teissie J, Pataro G, Pliquett U, Saulis G, Stefan T, Miklavcic D, Vorobiev E, Frey W (2016) Energy-efficient biomass processing with pulsed electric fields for bioeconomy and sustainable development. Biotechnol Biofuels, 9:94. doi 10.1186/s13068-016-0508-z. Góngora-Nieto MM, Sepúlveda DR, Pedrow P, Barbosa-Cánovas GV, Swanson BG (2002) Food processing by pulsed electric fields: treatment delivery, inactivation level, and regulatory aspects. Lebensm Wiss Technol 35:375–388 Kotnik T, Miklavčič D, Mir LM (2001) Cell membrane Electropermeabilization by symmetrical bipolar rectangular pulses. Part II. Reduced electrolytic contamination. Bioelectrochem 54:91–95 Kotnik T, Kramar P, Pucihar G, Miklavčič D, Tarek M (2012) Cell membrane electroporation – Part 1: the phenomenon. IEEE Electr Insul Mag 28(5):14–23 Mahnič-Kalamiza S, Vorobiev E, Miklavčic D (2014) Electroporation in food processing and biorefinery. J Membr Biol 247:1279–1304. doi:10.1007/s00232-014-9737-x Meneses N, Jaeger H, Knorr D (2011) pH-changes during pulsed electric field treatments – Numerical simulation and in situ impact on polyphenoloxidase inactivation. Innov Food Sci Emerg Technol 12: 499–504. Miklavčič D, Mali B, Kos B, Heller R, Serša G (2014) Electrochemotherapy: from the drawing board into medical practice. Biomed Eng Online 13:29 Morren J, Roodenburg B, de Haan SWH (2003) Electrochemical reactions and electrode corrosion in pulsed electric field (PEF) treatment chambers. Innovative Food Sci Emerg Technol 4:285–295 Pakhomov AG, Miklavcic D, Markov MS (2010) Advanced electroporation techniques in biology and medicine. CRC Press, Boca Raton Pataro G, Falcone M, Donsì G, Ferrari G (2014) Metal release from stainless steel electrodes of a PEF treatment chamber: effects of electrical parameters and food composition. Innovative Food Sci Emerg Technol 21:58–65 Pataro G, Barca GMJ, Donsì G, Ferrari G (2015a) On the modeling of electrochemical phenomena at the electrode solution interface in a PEF treatment chamber: methodological approach to describe the phenomenon of metal release. J Food Eng 165:34–44 Pataro G, Barca GMJ, Donsì G, Ferrari G (2015b) On the modelling of the electrochemical phenomena at the electrode solution interface of a PEF treatment chamber: effect of electrical parameters and chemical composition of model liquid food. J Food Eng 165:34–44 Raso J, Frey W, Ferrari G, Pataro G, Knorr D, Teissie J, Miklavčič D (2016) Recommendations guidelines on the key information to be reported in studies of application of PEF technology in food and biotechnological processes. Innovative Food Sci Emerg Technol. doi:10.1016/j. ifset.2016.08.003 Roodenburg B, Morren J, Berg HE, de Haan SWH (2005a) Metal release in a stainless steel Pulsed Electric Field (PEF) system. Part I. Effect of different pulse shapes; theory and experimental method. Innovative Food Sci Emerg Technol 6:327–336 Roodenburg B, Morren J, Berg HE, de Haan SWH (2005b) Metal release in a stainless steel pulsed electric field (PEF) system. Part II. The treatment of orange juice; related to legislation and treatment chamber lifetime. Innovative Food Sci Emerg Technol 6:337–345 Saulis G, Rodaitė-Riševičienė R, Snitka V (2007) Increase of the roughness of the stainless-steel anode surface due to the exposure to high-voltage electric pulses as revealed by atomic force microscopy. Bioelectrochem 70:519–523 Stern O (1924) Zur theorie der elektrolytischen doppelschicht. Z Elektrochem 30:508 Wang J (2000) Analytical electrochemistry. Wiley-VCH, New York

Modeling Transport Across the Electroporated Membrane Miao Yu and Hao Lin

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Model Framework: A Continuum Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Electrical Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Permeabilization Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Coupling of the Electrical and the Permeabilization Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . Species Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Numerical Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Exemplary Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 3 8 10 12 15 17 19 21 21

Abstract

Molecular transport is central to many applications of electroporation. With few exceptions, delivering target molecules into the cytoplasm and/or the nucleus defines the very application itself. The pertinent transport phenomena are complex and involve different physical mechanisms which are strongly coupled. Moreover, there is large variability in molecular properties of the target and the protocols used in electroporation practice. Such complexity calls for development of reliable theoretical and modeling tools to assist in understanding, design, and optimization. In this chapter, a theoretical framework based on a continuum approach is presented. The model is built by combining and coupling three key elements: electrodynamics, membrane permeabilization, and species transport. A detailed description of the governing physics and equations, together with aspects M. Yu (*) • H. Lin (*) Department of Mechanical and Aerospace Engineering, Rutgers, The State University of New Jersey, New Brunswick, NJ, USA e-mail: [email protected]; [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_6-1

1

2

M. Yu and H. Lin

of numerical implementation methods, is provided. Exemplary results are given and discussed. The current approach is capable of predicting the permeabilization status of the membrane and the contributions from various transport mechanisms with both spatial and temporal resolutions, thus corroborating experimental results in details. On the other hand, it is limited to simulate the transport of small- to moderate-sized molecules; further development is required to tackle that of macromolecules such as DNA. Keywords

Transport • Electrophoresis • Diffusion • Nernst-Planck equations • Electroporation • Numerical simulation

Introduction Ionic or molecular transport is a key element in both irreversible and reversible electroporation. In the former, excessive exchange of the cytosol with the buffer may contribute significantly to cell death. In the latter, transport plays a central role, as uptake of target molecules is the very goal of its many applications including cell transfection, gene therapy, drug delivery, among others. The targets to be delivered into the tissue or cells range from biologically active agents to fluorescent markers and are very diverse in molecular properties. Meanwhile, the cell or tissue type and the electrotransfer configuration also demonstrate large variability. Extensive experimental efforts have been committed to developing optimized delivery protocols. On the other hand, theoretical and modeling tools are in high demand to offer guidance and assistance in the understanding and optimization. Over the past three decades, significant progresses have been made to develop phenomenological models for membrane permeabilization upon application of a strong electric field (Weaver and Chizmadzhev 1996; Neu and Krassowska 1999). These models are able to predict detailed dynamics of each individual pore, including location on the membrane, creation, destruction, and size evolution. They serve as the foundation for quantitative prediction of cross-membrane transport. Here the term “cross-membrane transport” refers to ions or molecules translocating across the electroporated membrane through the aqueous pores – such is the basic hypothesis throughout this chapter. Evidently, the past research verified that different models are necessary to account for the very different physics involved in the transport of smallto moderate-sized molecules as opposed to large macromolecules. For the former, it is commonly assumed that they can directly traverse electropores; examples include calcium ions, fluorescent markers such as propidium iodide, and most anticancer drugs. The latter are typically long polymer chains that usually have a coil-like effective radius much greater than the typical pore size; examples are deoxyribonucleic acid (DNA), ribonucleic acid (RNA), and other macromolecules. The current state of the art in electroporation research has a relatively strong consensus in the mechanisms transporting smaller molecules: they are diffusion, electrophoresis, or for most cases a combination of the two across the permeabilized

Modeling Transport Across the Electroporated Membrane

3

membrane. On the other hand, the mechanism for DNA electrotransfer remains elusive (Rosazza et al. 2016); possible candidates involve electrophoretic translocation (Yu et al. 2012) and endocytosis. Endocytosis has been strongly indicated by recent experimental work (Wu and Yuan 2011), yet direct translocation via nanochannel-mediated electrophoresis has also been demonstrated in an attempt to rapidly transport DNA and to bypass the slower endocytosis (Boukany et al. 2011). In this chapter, the focus is set to the modeling of the transport of small- to moderate-sized molecules, where both the understanding and model development are relatively mature. A continuum approach shall be adopted, which treats variation (in electric potential, species concentration, pore density, etc.) as continuous and gradual quantitative transitions in space and time. Discontinuities are only allowed across special boundaries (e.g., the cell membrane). Non-continuum approaches, most notably molecular dynamic models, are presented elsewhere (“▶ Molecular Models of Lipid Bilayers and Electropore Formation”) in this handbook. In what follows, a comprehensive description of the theoretical framework is given. This framework is comprised of three different physical models tracking electrodynamics, membrane permeabilization, and molecular transport, respectively. For electrodynamics, the Ohmic equation is solved for conservation of the electric current (section “The Electrical Problem”). For membrane permeabilization, the model is developed based on the Smoluchowski equation in stochastic theory (section “The Permeabilization Model”). For molecular transport, the Nernst-Planck equations are solved to account for electrophoresis, diffusion, and chemical reaction involved in ionic transport (section “Species Transport”). The coupling of these three models, proper boundary conditions (section “The Coupling of the Electrical and the Permeabilization Model”), and a brief introduction of numerical implementation (section “Numerical Implementation”) are also introduced. Exemplary simulation results are then presented (section “Exemplary Results”). Prior to conclusions (section “Conclusions”), other alternative transport models are also briefly introduced (section “Other Models”).

Model Framework: A Continuum Approach The Electrical Problem To begin with, a simple, one-dimensional model is introduced to familiarize the readers with the basic physics. A schematic is shown in Fig. 1. A planar lipid bilayer membrane divides a space filled with an aqueous electrolytic solution of conductivity λ into two compartments. Planar electrodes parallel to the membrane are placed in each compartment. When a voltage difference V0 is applied between the electrodes, an electric field is created in the electrolytic solution. Obviously, the problem can be approximated to be one dimensional mathematically. Define Φ(x, t) as the electric potential field (to be solved), and the electric field strength is E =
@ Φ/@ x, where the minus sign results from the fact that the electric

4

M. Yu and H. Lin

Fig. 1 A schematic of a planar membrane problem. A constant voltage difference V0 is applied on the electrodes. The reservoir is filled with an electrolytic solution of conductivity λ, creating a one-dimensional electric potential field Φ(x, t). The membrane has a capacitance of Cm, as well as a conductance of gm upon electroporation

field directs opposite of the potential gradient. The conservation of Ohmic current λ E (also called the Ohmic equation) in the electrolytes is the governing equation:
 @ @Φ
λ ¼ 0: @x @x

(1)

In most situations, λ can be treated as a constant, so that Eq. 1 is simplified into @2Φ ¼0 @x2

(2)

This equation holds for both regions left and right to the membrane (namely, the bulk). The current balance condition also needs to hold across the membrane. For an intact lipid membrane, the membrane is purely capacitive, and the cross-membrane current balance condition is
λ

@Φ dV m @Φ ¼
λ jx¼0þ ¼
Cm jx¼0
; @x @x dt

(3)

Modeling Transport Across the Electroporated Membrane

5

where Cm is the membrane capacitance, V m ðtÞ  Φðx ¼ 0þ , tÞ
Φðx ¼ 0
, tÞ is the transmembrane potential (TMP), CmdVm/dt is the capacitive current across the membrane, and x = 0+ and 0
refers to the right and the left boundary of the membrane, respectively. The thickness of the membrane ( nm) is ignored. Solving Eqs. 2 and 3 along with an initial condition for Vm(t = 0) provides a transient solution of Φ(x,t) and Vm. Equations 2 and 3 are probably the simplest model for practical electroporation apparatus. For Vm = 0 at t = 0, the solution is a simple exponential decay,   V m ðtÞ ¼ V 0 1
e
t=τcharge ;

(4)

where τcharge ¼ 2Cm L=λ is the electric charging time of the membrane. It is clear Vm will eventually approach V0 given sufficient time. An analytical solution is also available if an alternating current (AC) field of frequency ω and peak voltage V0 instead of the direct current (DC) field is applied across the electrodes: V ðtÞ ¼ V 0 cos ðωtÞ. The solution is V m ðt Þ ¼

V0

 1 þ ωτcharge

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  2  


t=τcharge 1 þ ωτchage cos ωt
arctan ωτcharge
e : 2 (5)

It is straightforward to verify that the solution (Eq. 5) converges to (Eq. 4) in the limit of ω = 0. More conveniently, the solution can be written in a complex form:
iωt  e
e
t=τcharge V m ðtÞ ¼ V 0  Re ; 1 þ iωτcharge

(6)

pffiffiffiffiffiffiffi where Re denotes the real part of a complex function, and i ¼
1 is the imaginary unit. When ωτcharge  1 or ωτcharge  1, the solution (Eq. 5) for Vm sees an evident drop in magnitude as well as a notable phase shift from the original AC field. For both DC and AC electroporation, significant change in behavior of the system is expected when the electric charging time τcharge becomes comparable to the characteristic timescale of the field, namely, the pulse length tp for both DC and AC pulses, and in addition 1/ω for AC pulses. This scaling analysis is useful for problems in all geometries, with τ  CmL/λ where L is the characteristic size of the problem. Electroporation occurs when Vm is greater than a threshold value, typically a fraction of a volt. This produces a significant increase in membrane conductance, gm, which is simply ignored for an intact membrane in Eq. 3. The updated crossmembrane condition becomes
λ

@Φ dV m @Φ
gm V m ¼
λ j þ ¼
Cm jx¼0
: @x x¼0 @x dt

(7)

6

M. Yu and H. Lin

Fig. 2 A schematic of a spherical-shell model with axisymmetry. (r, θ) denotes the spherical coordinate system. x is the axis of rotation and is aligned with the direction of field application. The field strength is denoted by E0. The intra- and extracellular conductivities are denoted by λi and λe, respectively

This equation means that upon the occurrence of electropermeabilization, the crossmembrane electric current becomes a sum of the capacitive current and the conductive current. Noting that gm(t) is a transient value representing the degree of permeabilization, it must be derived from the permeabilization model which is introduced in section “The Permeabilization Model.” A more frequently encountered geometry is the electroporation of a suspending cell of approximately spherical shape. This corresponds to a spherical-shell model as illustrated in Fig. 2. The presumably isolated cell is modeled as a spherical space of radius R, filled with an electrolyte of conductivity λi, encapsulated by an infinitesimally thin membrane of capacitance Cm, and surrounded by an extracellular electrolyte of conductivity λe. A pulse with the constant strength of E0 is applied, and axisymmetry is assumed to be about the direction of the electric field. In principle, this problem has the same physical governing equations as the planar membrane, but it needs to be written under the spherical coordinate system, (r, θ). The Ohmic equation turns out to be   1 @ 1 1 @ @Φ 2 @Φ r sin θ þ ¼ 0: r 2 @r @r r 2 sin θ @θ @θ

(8)

Again the intra- and extracellular conductivities are taken to be constant. Similar to Eq. 3, the cross-membrane current balance (in the direction normal to the membrane) is
λi

@Φ dV @Φ ¼
λe jr¼R
¼
Cm j þ: @r dt @r r¼R

(9)

At r = 1, the field becomes uniform (ambient), so that Φðr ¼ 1Þ ¼
E0 r cos θ. Note that this is usually a good approximation when the size of cell is much smaller than the spacing between the electrodes, but may not be valid in, e.g., patch-clamp experiments. The solution to Eqs. 8 and 9 can easily be obtained as

Modeling Transport Across the Electroporated Membrane

  V m ðθ, tÞ ¼
1:5E0 R cos θ 1
e
t=τcharge ;

7

(10)

where the TMP is defined as V m  Φe ðr ¼ Rþ Þ
Φi ðr ¼ R
Þ, and τcharge ¼ Cm R   1 1 þ 2λe λi . Equation 10 is the most widely used formula to estimate the applied voltage across the cell membrane in electroporation. Typically for Cm  0.01 F/m2, R  10 μm, and λi,e  0.1 S/m, it can be estimated that τcharge  1 μs. This explains why in nanosecond electroporation, a much stronger electric field has to be applied. In such cases, the pulse is too short for the membrane charging process to complete. In order for the TMP to reach the threshold for electroporation, a much higher E0 is   necessary to compensate for the small factor of 1
e
t=τcharge . For conventional electroporation with tens to hundreds of microsecond or longer, the membrane charging often completes and the TMP reaches a steady state:   V m θ, t  τcharge ¼
1:5E0 R cos θ:

(11)

At θ = 0, π, namely, the electrode-facing points, the absolute value of TMP reaches its maximum 1.5E0R where 1.5 is simply a geometric factor. Generally speaking, |Vm|  E0R serves as a good estimate of the TMP for most conventional electroporation applications, noting that here E0 is a characteristic field strength, and R is the characteristic size. If an AC field E(t) = E0 cos(ωt) is applied, the analytical solution for Vm simply becomes


 eiωt
e
t=τcharge V m ðθ, tÞ ¼
1:5E0 Rcosθ  Re : 1 þ iωτcharge

(12)

It is evident that the solution to the spherical geometry is analogous to the planar membrane one. Moreover, the occurrence of electroporation also leads to an increase in membrane conductance, gm, and Eq. 9 needs to be subsequently replaced with
λi

@Φ dV @Φ
gm V m ¼
λe jr¼R
¼
Cm j þ: @r dt @r r¼R

(13)

Again a permeabilization model is necessary to complete the problem. In principle, a continuum model can always be applied to an arbitrary geometry. The general equations are (in three dimensional)   ∇ 
λi,e ∇Φi,e ¼ 0;
λi n  ∇Φi ¼
Cm

dV m
gm V m ¼
λe n  ∇Φe at the membrane: dt

(14) (15)

Here ∇ denotes the gradient of a scalar field, ∇ denotes the divergence of a vector field, and n denotes the unit vector normal to the membrane. With appropriate

8

M. Yu and H. Lin

boundary conditions and numerical approach, the framework is suitable to provide meaningful predictions for cases with much more complex geometries and/or experimental conditions.

The Permeabilization Model The most widely accepted current theory for membrane permeabilization forms from the perspective of minimizing membrane energy. When the transmembrane potential exceeds a certain value, hydrophilic membrane pores greater than a particular size become favorable, because they provide lower energy states than the intact lipid bilayer which holds otherwise more capacitive energy. The permeabilization model is subsequently established based on this theoretical understanding, using the Smoluchowski equation to describe the creation, destruction, and evolution of pores (Barnett and Weaver 1991):
 @nðr, tÞ @ nðr, tÞ @U ðr, V m , tÞ @nðr, tÞ þD

¼ Sðr, V m Þ: @t @r kB T @r @r

(16)

Equation 16 is a partial differential equation (PDE) governing the distribution of pore density n(r, t) as a function of pore radius, r, and time, t. Here D is the diffusion constant for size evolution, U(r, Vm, t) is the pore energy as a function of pore radius and the TMP, kB is the Boltzmann constant, T is temperature, and S(r, Vm) is the source term giving the rate of creation and destruction of pores which also depends on the TMP. Equation 16 is generally cumbersome to deal with. An asymptotic version (the ASE model) of this equation is frequency used, which is given by Neu and Krassowska (1999). Mathematically, the resulting equations are ordinary differential equations (ODE), hence much easier to analyze and solve:
 2 dN ðtÞ N ðtÞ V m =V ep Þ ð ¼ αe 1
: dt N eq ðV m Þ

(17)

Here N(t) is the pore number density per area (m
2), and Neq(Vm) is the equilibrium pore density for a given Vm. It is straightforward to see that pores are created when N < Neq and destroyed when N > Neq. α and Vep are model constants, representing pore creation rate coefficient and characteristic voltage of electroporation, respec2 tively. N ¼ N eqðV m =V ep Þ , where N and q are again model constants. The physical eq

0

0

meaning of these constants and their typical values are listed in Table 1. In deriving Eq. 17 from Eq. 16, several key features of the dynamic system are revealed. First, all hydrophilic pores are created at a specific radius r. Second, the vast majority of pores reach an equilibrium radius rm in typically 5 μs after being created. In other words, for microsecond electroporation where the pulse length is much longer than the characteristic timescale of pore evolution, one can simply

Modeling Transport Across the Electroporated Membrane

9

Table 1 Typical values of parameters used in the ASE model (Krassowska and Filev 2007) Symbol α Vep q N0 r rm D Fmax rh rt β γ σ0 σ0

Value 1  109 m
2 s
1 0.258 V  (rm/r)2 1.5  109 m
2 0.51  10
9 m 0.8  10
9 m 5  10
14 m2 s
1 0.70  10
9 N V
2 0.97  10
9 m 0.31  10
9 m 1.4  10
19 J 1.8  10
11 J m
1 1  10
6 J m
2 2  10
2 J m
2

Definition Creation rate coefficient Characteristic voltage of electroporation Constant for pore creation rate Equilibrium pore density at Vm = 0 Minimum radius of hydrophilic pores Minimum energy radius at Vm = 0 Diffusion coefficient for pore radius Maximum electric force for Vm = 1 V Constant for advection velocity Constant for advection velocity Steric repulsion energy Edge energy Tension of the bilayer without pores Tension of hydrocarbon-water interface

assume that all pores have reached their equilibrium size at the end of the pulse. The value of rm can be obtained by finding the local minima of the pore energy U(r, Vm). If one wishes to capture the pore dynamics in a more accurate manner, another ODE, also derived in an asymptotic manner from the original Smoluchowski equation, is needed (Krassowska and Filev 2007):   dr j ¼ U r r j , V m , σ eff dt

(18)

Here rj is the radius of the j th pore and Ur = @U /@r is the driving force of pore expansion or shrinkage to minimize the pore energy U. Ur is given by 

ur r, V m , σ eff




 r 4 1 D V m Fmax 
2πγ þ 2πσ eff r ; ¼ þ 4β kB T 1 þ r h =ðr þ r t Þ r r

(19)

where D, Fmax, rh, rt, β, r, and γ are model constants. Equation 19 includes four terms, representing the effect of the electric force, the steric repulsion of lipid heads, the line tension on the pore perimeter, and the surface tension of the membrane, respectively. In the last term, σ eff is an effective membrane tension, given by σ eff ¼ 2σ 0


2σ 0
σ 0 ð 1
ρÞ 2

:

(20)

σ 0 and σ 0 are constants, and ρ is the pore area density, namely, the fraction of the local membrane area occupied by aqueous pores. To solve rm, one may simply solve Ur (rm,Vm, σ eff) = 0 for given values of Vm and σ eff. Clearly, the equilibrium pore size rm is a function of the TMP and the membrane tension.

10

M. Yu and H. Lin

Equations 17, 18, 19, and 20 form a complete permeabilization model providing description of pore dynamics (radius, number density, location) given Vm. Briefly, hydrophilic pores will emerge with an initial radius of r, at the number density of N (t,θ) via Eq. 17, and then each begins to evolve in size per Eq. 18. A numerical approach is often taken to obtain the specific solution. A number of model constants are involved, the values of which are either estimated theoretically or measured by experiments. It is worthwhile to mention that this permeabilization model in practice usually predicts a resealing timescale much shorter (within microseconds) (Krassowska and Filev 2007) than what is experimentally observed (tens of milliseconds to minutes and even hours) (Sukharev et al. 1992; Khine et al. 2007). Therefore one should not rely on the model to offer a faithful prediction for the resealing process upon field cessation.

The Coupling of the Electrical and the Permeabilization Model The electrical and the permeabilization models above are derived from electrodynamics and membrane biophysics, respectively. In electroporation, they are coupled together at the membrane via interactions between the transmembrane potential, Vm, and the cross-membrane current density, jm. This is straightforward to understand: the value of Vm commands creation and evolution of aqueous pores on the membrane; the creation of such conductive pores in turn increases jm and “shunts” Vm. Given sufficient time, a new equilibrium between electrodynamics and pore dynamics can be established. Such equilibrium is usually global since the Laplace Equation 8 is mathematically an elliptic PDE. This is to say, a disturbance in the electric current at any membrane locality would influence the entire domain. In the previous subsection, details have been given regarding how to update instantaneous pore information from a local value of Vm(θ, t). The goal of this subsection is to subsequently show how the transient pore dynamics will in turn render quantitative changes in Vm. To fulfill this purpose, one needs to solve the Laplace Equation (8) at every instant with an updated boundary condition at the membrane (13). In Eq. 13, the only variable that is presumably correlated with pore dynamics is the membrane conductance gm (in Sm
2). Upon electrical breakdown of the membrane, the value of gm can increase by several orders of magnitude (Neumann et al. 1989). As a result, the conductive current gmVm in Eq. 13 will correspondingly surge and reshape the electric field distribution. Therefore a formula is needed to calculate gm(θ) from the total number of pores, denoted K(θ), and their radii, denoted rj=1,2,. . .,K (θ), at a membrane locality, θ. Intuitively it can be imagined that the electropores are in parallel with each other, so that gm is the sum of each individual pore conductance Gp divided by the area of a local membrane element ΔA (see more details about ΔA in section “Numerical Implementation”):

Modeling Transport Across the Electroporated Membrane

11

Fig. 3 Schematic of pore resistance Rp and input resistance Ri. The pore is assumed to be cylindrical. The solid lines in the extracellular space are equipotential lines, and the dashed are field lines. Rp is the resistance of the cylindrical pore, and Ri is the resistance of the spreading electrolytic area outside the pore (on both sides). Rp and Ri are in series to give the total resistance of an electropore

gm ¼

K 1 X Gp : ΔΑ j¼1

(21)

Several models exist to calculate Gp from the pore radius and conductivity. A simple and widely accepted approximation is (Powell et al. 1986) Gp ¼

1 : Rp þ Ri

(22)

Here Rp ¼ h=πλr j 2 is the resistance of a cylindrical aqueous pore with thickness h, radius rj, and conductivity λ, and Ri = 1/2λrj is the input resistance (also termed the spreading resistance) resulting from the resistance of the electrolytic solution in regions next to the pore (Fig. 3). The two resistances add to each other as they are in series. In deriving Eq. 22, it is assumed that the conductivities on both sides of a pore are equal. Such constraint can be removed via an improved formula (Li and Lin 2010): Gp ¼

λeff : þ 1=2r j

h=πr 2j

(23)

Here λeff is an effective pore conductivity given by λeff ¼ ðλe
λi Þ=lnðλe =λi Þ. When λe = λi, Eq. 23 converges to Eq. 22. Equations 22 and 23 do not take into account the interaction of ions with the pore walls or the steric hindrance for ions of comparable size to the pore. These effects can be introduced into Eqs. 22 and 23 as prefactors in the formula of Rp: Rp ¼

1 h    : H r s =r j J r j , V m πλr 2j 

(24)

12

M. Yu and H. Lin

In Eq. 24, H is Renkin’s steric hindrance factor, which is a function of the ratio of ion radius rs (for dominant ions contributing to electrical conductivity) to pore radius rj. J is a partitioning coefficient accounting for the electrostatic energy barrier for an ion to enter a pore, which depends on the pore radius and the transmembrane potential. Both factors converge to unit for large pores, and in most cases they only give rise to significant modification of Rp for pores smaller than 1 nm. Interested readers are referred to (Vasilkoski et al. 2006) for details. More complicated pore conductance models have recently been developed using molecular dynamics (Casciola et al. 2016) or a continuum Poisson-Nernst-Planck theory considering the toroidal shape of the pore, pore selectivity for different ionic species, the electric double layer at the lipid-water interface, and the electroostomic flow (Rems et al. 2016). These models are expected to provide improved results at the cost of higher model complexity. In a brief summary, Eqs. 21, 22, 23, and 24 set up approaches for quantification of gm(θ). This closes the dynamic system (Eqs. 8, 13, 14, 15, 16, 17, 18, 19, and 20), and the coupled equations can be solved to obtain temporal and spatial information of the electric field and pore distribution. In principle, the solution can only be obtained with a numerical approach, as details will be given in section “Numerical Implementation.” Throughout this chapter, the effect of the resting potential, Vrest, is neglected. Some authors (e.g., (Debruin and Krassowska 1999)) suggest that a constant Vrest should be added to Vm in Eq. 13 at any moment or place to account for its effects. On the other hand, there still lacks widely accepted evidence to illustrate how the occurrence of electroporation modifies Vrest, which is likely to be nontrivial considering its origin. Readers can add or drop this term – the difference has not been shown to be substantial.

Species Transport In this subsection, the governing equations of electroporation-mediated species transport are presented. Here it is assumed that the transport equation is one-way coupled from the equation system introduced in sections “The Electrical Problem,” “The Permeabilization Model,” and “The Coupling of the Electrical and the Permeabilization Model.” That is, species transport of ions, polymers, and other molecules does depend on the electric field and the permeabilization state of the membrane but does not feed back to interfere with the electrodynamic or permeabilization process. This assumption seems to contradict with the fact that the electrical conductivity derives directly from that of the ionic species. However, in the current model, only one or multiple species (usually the molecules targeted for delivery) whose contributions to conductivity are negligible are considered. Taking, for example, the delivery of propidium iodide, if the extracellular concentration is 100 μM, its contribution to conductivity is 10
4–10
3 S m
1. This is typically much lower than the buffer or cytosol conductivity, e.g., 0.1–1 S m
1. The

Modeling Transport Across the Electroporated Membrane

13

transient dynamics of such species is henceforth unlikely to change the electrical conductivity, based on which the electrical problem is computed. The evolution of the major ionic species composing the buffer and the cytoplasm, on the other hand, does tend to change the bulk electrical conductivities – this is a very subtle issue. Scaling analysis (Lin et al. 2004) demonstrated that the change of the bulk electrical conductivity is primarily mediated by a diffusive process, despite the presence of a driving electric field. The electrophoretic effects between the cations and anions are largely canceled via electroneutrality which is a strong and universal constraint. Since the diffusive timescale is typically much longer than electroporation times, the bulk electrical conductivities being constant are accepted as a valid approximation in this chapter. In the bulk of intra- and extracellular medium, the Nernst-Planck equations are used to describe molecular transport owing to electrophoretic drifting, diffusion, and chemical reaction: @ck ¼ ∇  ðμk Fzk ck ∇ΦÞ þ ∇  ðDk ∇ck Þ þ Sk : @t

(25)

Here ck is the molar concentration of the k th species in mol m
3; μk, the mechanical mobility in mol s kg
1; zk, the valence number of the species; and Dk, the diffusion constant of the species in m2 s
1. According to Einstein’s relation, Dk = μkRT, where R is the gas constant, and T is temperature. F is the Faraday constant and Φ is the electric potential. In Eq. 25, the left-hand side stands for the rate of change of local concentration; the first and the second term on the right-hand side stand for the net inward flux into a locally infinitesimal volume owing to eletrophoretic drifting and molecular diffusion, respectively. Sk is a source term representing local production or consumption of the species from any kinetics. Eq. 25 is again written for the generalized, three-dimensional situation. For a two-dimensional axisymmetric geometry in a spherical coordinate system, the Nernst-Planck equations become
   @ck 1 @ 1 1 @ @Φ 2 @Φ r ck sin θck ¼ μk F zk 2 þ 2 @r r sin θ @θ @t
r @r  @θ  1 @ 1 1 @ @ck 2 @ck r sin θ þDk 2 þ 2 þ Sk : r @r r sin θ @θ @r @θ

(26)

In Eq. 26, Φ(r, θ, t) is available from the models in sections “The Electrical Problem,” “The Permeabilization Model,” and “The Coupling of the Electrical and the Permeabilization Model.” Therefore, the transient species concentration at any position, ck(r, θ, t), can be obtained by integrating Eq. 26, given appropriate boundary conditions at the membrane. This equation can also be used to solve the transport of electrically neutral or immobile molecules, by simply setting to zero pertinent terms. Another benefit of the simplification made above is that now the species transport is decoupled from each other, unless via any chemical kinetics there may exist. The effect of the latter is represented by the source term Sk in each equation. For example,

14

M. Yu and H. Lin

upon entering the cytoplasm, propidium iodide (PI2+) binds to sites on DNA/RNA (denoted simply by a single species B) to emit fluorescence. The binding kinetics is described by kþ

PI2þ þ B Ð PIB k


(27)

where k+ and k
are the association and dissociation rate constants, respectively. This kinetics leads to the following expressions of Sk, which enters Eq. 25 or Eq. 26: 8 < SPI ¼
kþ cPI cB þ k
cPIB S ¼
kþ cPI cB þ k
cPIB : : B SPIB ¼ kþ cPI cB
k
cPIB

(28)

In Eq. 28, the right-hand sides are the corresponding production or depletion rates for each species resulting from the chemical reaction (27). The above equations, similar to the electrical problem, are solved in both the intra- and extracellular spaces. These equations are coupled at the permeabilized membrane to account for cross-membrane transport. An electroporated membrane only provides limited access to ions and molecules. This access, as hypothesized in this chapter, depends exclusively on the pore distribution governed by Eqs. (17, 18, 19, and 20) in section “The Permeabilization Model.” In other words, foreign ions or molecules can only translate across the membrane through electropores. Other transport mechanisms such as ion channels and endocytosis are not accounted for by the current model. Another main hypothesis is that ions or molecules can only go through pores greater than their own sizes, e.g., the hydrated diameter of an ion. A cross-membrane transport model is derived for fkm, the membrane flux density (per area) for the k th species, by solving a one-dimensional Nernst-Planck equation within the pore (Li and Lin 2011):


fm k

Dk ðPek
ln λr Þ ¼ ρk h



 λr
1 cek
cik expðPek Þ : ln λr λr
expðPek Þ

(29)

Here Pek ¼
V m =ðRT=Fzk Þ is an effective Péclet number for each species, λr = λe/λi is the extra- to intracellular conductivity ratio, h is the membrane thickness, and cke, cki are X extra- and intracellular species molar concentration, respectively. πr 2j =ΔΑ is the pore area density summing over pores with radii rj greater ρk ¼ rj >rk

than the effective ionic or molecular radii rk. This term can be conveniently modified to accommodate other models for ion-selective-passage criteria other than simply by size comparison. The flux density is positive when pointing outward and negative when pointing inward. It has to match the local flux density on both sides of the membrane provided by the Nernst-Planck equations:

Modeling Transport Across the Electroporated Membrane

15

e, i fm (30) k ¼ fk   where f ek,i ¼
n  μk Fzk cek, i ∇Φ þ Dk ∇cek, i in a general coordinate system, or f ek, i   ¼
μk Fzk cek, i ∇Φ=@r þ Dk @cek, i =@r in an axisymmetric spherical coordinate

system. In many cases, Eq. 29 can be greatly simplified. If we ignore the conductivity difference across the membrane and set λr = 1, Eq. 29 is reduced to fm k ¼ ρk

Pek Dk cek
cik expðPek Þ : 1
expðPek Þ h

(31)

Considering the situation when Pek = 0 (no electric field or for electrically neutral molecules), Eq. 31 further reduces to a pure diffusive form:  i  e fm k ¼ ρk Dk ck
ck =h:

(32)

On the other hand, when the electric field is on and |Pek|  1 (this is generally true because the transmembrane potential required to induce electroporation is much greater than RT/F 0.025 V), Eq. 31 becomes an upwind scheme as electrophoresis dominates diffusion: 8 μ Fz V > <
ρk k k m cek h fm k ¼ > :
ρk μk Fzk V m cik h

for V m zk > 0 for V m zk < 0

(33)

noting that μkFzkVm is the cross-membrane eletrophoretic velocity of the k th species. Equations 32 and 33 can be used as approximations to Eq. 29 in the specific simplifying limits.

Numerical Implementation Equations 14, 15, 25, 29, 30 for a Cartesian coordinate system, or Eqs. 8, 13, 26, 29 and 30 for an axisymmetric spherical geometry, along with the permeabilization model (17, 18, 19, 20, 21, and 22), provide a complete model description of electroporation-mediated species transport. This offers a powerful tool that is capable of capturing the key dynamic behavior of general interest in electroporation research, such as that of the transmembrane potential, membrane permeability, and molecular transport. Owing to the continuum approach, details on the variables of interest have both spatial and temporal resolution, in contrast to the compartment models which only compute integrated quantities. In general, the equation system can only be solved numerically. The numerical implementation of the Ohmic equation and the Nernst-Planck equations follows

16

M. Yu and H. Lin

standard approaches, typically finite-difference, finite-volume, or finite-element methods. A transport lattice method is also available to tackle the complex cell geometry and field interactions in biologically relevant transport problems (Donald et al. 2005). Noting that both equations to solve are in the form of conservation of flux, flux-conserving algorithms are preferred. The same rule holds across the membrane – the flux has to be conserved on both sides of any membrane element and be equal to the local membrane flux commanded by Eqs. 15, 21 and 22 for the electric current or Eq. 29 for species transport. The far-field boundary condition can be approximated by applying it at a distance far enough from the cell, e.g., 10R from its center. In choosing the time step size, Δt, one should make it small enough to guarantee stability and convergence, but not unnecessarily small to reduce computational cost. The numerical implementation of the permeabilization model needs to be carefully and properly executed. After mesh generation, the membrane is discretized into many membrane elements with an area ΔA (not necessarily uniform). Each area element is associated with a local pore number density, N(θ, t) (in m
2), as governed by Eq. 17. The value of N(θ,t) is updated by the local TMP, Vm(θ). The value of N  Δ A represents the time-dependent number of pores at this membrane element. When this value increases to be greater than one or greater integers, creation of new pores occurs. For example, N  ΔA increasing from 5.5 to 8.3 indicates the creation of three new pores on top of the five existing pores. The radii of newly created pores are initially at r = 0.51 nm and are subsequently tracked, respectively, as rj (the j th pore) which is dynamically governed by Eq. 18. Similarly, if the value of N(t)  Δ A decreases and rounds to a lower integer, destruction of existing pores takes place. For example, when N  ΔA drops from 8.3 to 5.5, three pores disappear locally on ΔA. In practice, one may find three local pores with the smallest size and one may set rj = 0 permanently for these pores, indicating that they have been destroyed. This practical approach brings forth an issue for the selection of ΔA. Clearly, each individual pore is bound within a local membrane element. The value of ΔA should be moderately large so as to contain electropores. For example, a spherical cell of R = 20 μm has a total surface area of around 5,000 μm2. The largest electropore reported in modeling can reach 500 nm in radius or around 1 μm2 in area. Therefore in practice ΔA should be no less than 1 μm2, and a value around 10 μm2 is recommended. In other words, the mesh size cannot be too small in the simulation of membrane electropermeabilization. This subtlety is due to the nature of the ASE model: in fact it combines both continuum and discrete features. Special treatment is also required if the membrane element is not uniform in the area upon mesh generation, which is frequently the case for a spherical cell or more complicated geometries. Spatial variation in ΔA leads to artificial variations in pore creation. For elements with smaller ΔA, pores will be created earlier than those with greater ΔA, as the threshold of N to reach N  ΔA 1 is lower. Conversely, for elements with smaller area, pore creation will be delayed. This result is unphysical, as some part of the membrane is more easily permeabilized than other, with otherwise no physical reason to indicate so. To eliminate this artificial effect, one

Modeling Transport Across the Electroporated Membrane

17

should always prefer to use a uniform ΔA across all membrane elements. However, when nonuniform elements are unavoidable, such as around certain regions where high numerical resolution is demanded for accuracy and stability, work-around approach can also be designed. In this case, one can designate a standard element area ΔA0, and elements smaller than ΔA0 are designated to be its exact fractions. For example, if an element has an area of ΔA0/2, it should be combined with a neighboring element of the same size to evaluate pore dynamics, giving rise to the same effective area ΔA0. The actual element ΔA0/2 can be used in other parts of the model system, such as for the electrical problem and species transport. This approach requires additional care in designing the mesh grid, but is nonetheless straightforward.

Fig. 4 Comparison between simulated and experimental results for electroporation-mediated delivery of calcium ions. The time stamps are the same in all. The top three rows show the simulated evolution of species concentration at the cell center plane. The fourth row shows a convolved concentration of the fluorescent compound CaFluo, in an attempt to take into account a finite optical depth in the measurement system for the observed results. The last row shows experimental images. In the simulated results, red denotes high, and blue denotes low values. The single electric pulse is 6 ms in length and 1 kV/cm in strength (Adapted from (Li and Lin 2011))

18

M. Yu and H. Lin

Exemplary Results Figure 4 provides an example for a whole-cell-level, continuum-based simulation (Li and Lin 2011). The simulation follows the setup of the experimental work (Gabriel and Teissié 1999) which investigates the electrotransfer of calcium ions into a China hamster ovary cell. The simulation uses the same model as introduced in sections “The Electrical Problem,” “The Permeabilization Model,” “The Coupling of the Electrical and the Permeabilization Model,” “Species Transport,” and “Numerical Implementation.” A list of model parameters is given in Table 2. The association and dissociation rate constants are associated with the reaction: kþ

Ca2þ þ Fluo Ð CaFluo; k


(34)

which describes the binding of calcium ions to Fluo-3 to produce the fluorescent compound, CaFluo. The simulation is compared with the fluorescence measurements in which a single electric pulse of 6 ms in length and 1 kV/cm in strength was used. Cells were preloaded with Fluo-3, so that entry of calcium ions into the cytoplasm led to fluorescence emission owing to the binding reaction (34). The evolution of the spatial fluorescence distribution inside the cell was recorded during and after the pulse, as displayed in the bottom row of Fig. 4. The simulation compares a convolved CaFluo concentration profile with the measurements, and a reasonable agreement is found. In Fig. 5, the concentration profiles are plotted along the cell centerline and at the same moments as in Fig. 4. The cell symmetrically spans from x =
8 to 8 μm. Table 2 List of model parameters for the simulation in Fig. 4

Definition Cell radius Membrane thickness Intracellular conductivity Extracellular conductivity Association rate constant Dissociation rate constant Intracellular diffusion coefficient of Ca2+ Extracellular diffusion coefficient of Ca2+ Intracellular diffusion coefficient of Fluo-3 Extracellular diffusion coefficient of Fluo-3 Intracellular diffusion coefficient of CaFluo Extracellular diffusion coefficient of CaFluo Valence number of Ca2+ Valence number of Fluo-3 Valence number of CaFluo Initial intracellular Fluo-3 concentration Initial extracellular Ca2+ concentration

Value 8 μm 5 nm 0.5 S m
1 0.15 S m
1 80 (μMs)
1 90 s
1 250 μm2 s
1 790 μm2 s
1 20 μm2 s
1 90 μm2 s
1 20 μm2 s
1 90 μm2 s
1 +2
5
3 2.2 μM 1 mM

Modeling Transport Across the Electroporated Membrane

19

Figures 4 and 5 display the intricate features of the transport process. Prominently, the intracellular concentration of Ca2+ is much higher than that in the extracellular space. This local concentration elevation results from the fact that the extracellular conductivity in this case is lower than the intracellular conductivity, λe < λi. Noting that the current balance condition across the membrane (13) commands λe Ee jr¼Rþ ¼ λi Ei jr¼R
, this indicates that the electric field strength is higher on the lower-conductivity side; hence, Ee jr¼Rþ > Ei jr¼R
. Subsequently, according to the mass transport balance across the membrane (30), Ee ce jr¼Rþ Ei ci jr¼R
, which holds during the pulse when the electrophoretic flux μkFzkckE dominates over the diffusive flux. This further leads to ce jr¼Rþ < ci jr¼R
, explaining the observed high Ca2+ concentration at the anode-facing (entrance) side of the cell. This effect is due to a well-known electrokinetic mechanism termed field-amplified sample stacking (Chien and Burgi 1991). The analysis suggests that using a lower extracellular conductivity gives higher delivery of small ions or charged molecules in electroporation, which has been confirmed by controlled experimental observation (Sadik et al. 2013). Similar to the “accumulation zone” of Ca2+ on the anode-facing side, a “depletion tail” at the cathode-facing side outside the cell is also seen during pulsation. In addition, calcium ions bind to locally available Fluo-3 almost instantaneously, causing a receding edge in the Fluo-3 profile (Fig. 5b), accompanied by an advancing front in the CaFluo profile (Fig. 5c), and during the pulse. After the pulse ceases, diffusion begins to smear all profiles within the intracellular space. This exemplary case demonstrates the complexity in species transport. The dynamic effects of electric field redistribution, membrane permeabilization, electrophoresis, diffusion, and chemical reaction are shown to work in cohort to give rise to the final profiles observed in measurements. A comprehensive and reliable model is thus highly valuable in connecting fundamental understanding and experimental observations.

Other Models The model framework presented in sections “The Electrical Problem,” “The Permeabilization Model,” “The Coupling of the Electrical and the Permeabilization Model,” “Species Transport,” and “Numerical Implementation” has the strength of resolving detailed spatial evolution. On the other hand, compartment models, although lacking such details, have the appeal of ease in implementation. Before concluding the chapter, they are briefly introduced. A cell-averaged model is proposed by Neumann et al. (1996) to predict the electrodiffusive transport of DNA across the electroporated membrane:


 

zeff je0 dnin D0 nin nout c c ¼
1

Δϕm : kB T Sdt h Vc v

(35)

20

M. Yu and H. Lin

Fig. 5 Species concentrations along the cell centerline at various times in the simulation. The cell symmetrically spans from x =
8 to 8 μm. (a) Calcium ion; (b) Fluo-3; (c) CaFluo; (d) convolved CaFluo. The single electric pulse is 6 ms in length and 1 kV/cm in strength (Adapted from (Li and Lin 2011))

Here ncin is the amount of delivered DNA into the cell, S is the electroporated surface area, h is the membrane thickness, D0 is the apparent cross-membrane diffusion coefficient of DNA, Vc is the cell volume, v is the volume of the bulk solution, nout is the total amount of DNA in the bulk solution, zeff is the effective charge number of the DNA-phosphate group, e0 is the elementary charge, kB is the Boltzmann constant, T is the absolute temperature, and Δϕ m is the mean transmembrane potential across the membrane. Equation 34 is an integrated form of the original NernstPlanck equations (25), assuming that the intra- and extracellular DNA concentration, membrane permeability, and the transmembrane potential are all uniformly distributed. One may further simplify the model by formulae approximating S and Δϕ m instead of obtaining them from the solution. Evidently, Eq. 35 can be used for smallto moderate-sized molecules as well. Diffusion- and endocytosis-driven transport models, such as a pharmacokinetic one (Puc et al. 2003) and a phase-transition one (Miklavčič and Towhidi 2010), were also developed using a volume-averaged manner. These models focus on diffusive transport and tackle the resealing process with various permeabilization theories. The readers are referred to further details in the literature.

Modeling Transport Across the Electroporated Membrane

21

Conclusions Transport phenomena within the context of electroporation generally involve multiple physical, chemical, and biological processes. The intrinsic complexity endows modeling efforts both necessity and difficulty. Over the past decade, significant progresses have been made in this area, which lead to whole-field simulations that are able to directly corroborate experimental measurements with both spatial and temporal resolution. The framework based on a continuum approach is outlined in this chapter. The model composes of three main elements, namely, electrodynamics, membrane permeabilization, and molecular transport. Strong coupling between the three results in complex patterns in all of field distribution, permeabilization status, and species concentration evolution. Numerical simulation of the model can capture much of these features and predict both qualitative and quantitative aspects of the phenomenon. In particular, one important contribution of such modeling efforts is to tease out the main transport mechanisms responsible for molecular delivery. In a broader perspective, the current approach is yet limited to the delivery of small- to moderate-sized molecules (less than a few tens of thousands of Dalton in molecular weight). For macromolecules such as DNA, a different approach has to be taken and is important for the general field of electroporation-mediated molecular transport.

References Barnett A, Weaver JC (1991) Electroporation: a unified, quantitative theory of reversible electrical breakdown and mechanical rupture in artificial planar bilayer membranes. Bioelectrochem Bioenerg 25:163–182 Boukany PE, Morss A, Liao W, Henslee B, Jung H, Zhang X, Yu B, Wang X, Wu Y, Li L, Gao K, Hu X, Zhao X, Hemminger O, Lu W, Lafyatis GP, Lee LJ (2011) Nanochannel electroporation delivers precise amounts of biomolecules into living cells. Nat Nanotechnol 6:747–754 Casciola M, Kasimova MA, Rems L, Zullino S, Apollonio F, Tarek M (2016) Properties of lipid electropores i: molecular dynamics simulations of stabilized pores by constant charge imbalance. Bioelectrochemistry 109:108–116 Chien RL, Burgi DS (1991) Field amplified sample injection in high-performance capillary electrophoresis. J Chromatogr 559:141–152 Debruin KA, Krassowska W (1999) Modeling electroporation in a single cell. i. effects of field strength and rest potential. Biophys J 77(3):1213–1224 Gabriel B, Teissié J (1999) Time courses of mammalian cell electropermeabilization observed by millisecond imaging of membrane property changes during the pulse. Biophys J 76:2158–2165 Khine M, Ionescu-Zanetti C, Blatz A, Wang LP, Lee LP (2007) Single-cell electroporation arrays with real-time monitoring and feedback control. Lab Chip 7:457–462 Krassowska W, Filev PD (2007) Modeling electroporation in a single cell. Biophys J 92(2):404–417 Li J, Lin H (2010) The current-voltage relation for electropores with conductivity gradients. Biomicrofluidics 4:013206 Li J, Lin H (2011) Numerical simulation of molecular uptake via electroporation. Bioelectrochemistry 82:10–21 Lin H, Storey BD, Oddy MH, Chen CH, Santiago JG (2004) Instability of electrokinetic microchannel flows with conductivity gradients. Phys Fluids 16:1922 Miklavčič D, Towhidi L (2010) Numerical study of the electroporation pulse shape effect on molecular uptake of biological cells. Radiol Oncol 44(1):34–41

22

M. Yu and H. Lin

Neu JC, Krassowska W (1999) Asymptotic model of electroporation. Phys Rev E 59(3):3471–3482 Neumann E, Sowers AE, Jordan CA (eds) (1989) Electroporation and electrofusion in cell biology. Plenum Press, New York Neumann E, Kakorin S, Tsoneva I, Nikolova B, Tomov T (1996) Calcium-mediated DNA adsorption to yeast cells and kinetics of cell transformation by electroporation. Biophys J 71:868–877 Powell KT, Derrick ED, Weaver JC (1986) A quantitative theory of reversible electrical breakdown in bilayer membranes. Bioeletrochem Bioenerg 15:243–255 Puc M, Kotnik T, Mir LM, Miklavčič D (2003) Quantitative model of small molecules uptake after in vitro cell electropermeabilization. Bioelectrochemistry 60:1–10 Rems L, Tarek M, Casciola M, Miklavčič D (2016) Properties of lipid electropores ii: comparison of continuum-level modeling of pore conductance to molecular dynamics simulations. Bioelectrochemistry in press. doi:10.1016/j.bioelechem.2016.03.005 Rosazza C, Haberl Meglic S, Zumbusch A, Rols MP, Miklavčič D (2016) Gene electrotransfer: a mechanistic perspective. Curr Gene Ther 16:98–129 Sadik MM, Li J, Shan JW, Shreiber DI, Lin H (2013) Quantification of propidium iodide delivery using millisecond electric pulses: experiments. BBA Biomembr 1828:1322–1328 Stewart DA, Gowrishankar TR, Smith KC, Weaver JC (2005) Cylindrical cell membranes in uniform applied electric fields: validation of a transport lattice method. IEEE Trans Biomed Eng 52(10):1643–1653 Sukharev SI, Klenchin VA, Serov SM, Chernomordik LV, Chizmadzhev YA (1992) Electroporation and electrophoretic DNA transfer into cells: the effect of DNA interaction with electropores. Biophys J 63:1320–1327 Vasilkoski Z, Esser AT, Gowrishankar TR, Weaver JC (2006) Membrane electroporation: the absolute rate equation and nanosecond time scale pore creation. Phys Rev E 74:021,904 Weaver JC, Chizmadzhev YA (1996) Theory of electroporation: a review. Bioelectrochem Bioenerg 41:135–160 Wu M, Yuan F (2011) Membrane binding of plasmid DNA and endocytic pathways are involved in electrotransfection of mammalian cells. PLoS One 6(6):e20,923 Yu M, Tan W, Lin H (2012) A stochastic model for DNA translocation through an electropore. BBA Biomembr 1818(11):2494–2501

Transmembrane Voltage Induced by Applied Electric Fields Tadej Kotnik

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analytical Derivation and Numerical Computation of ITV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spherical Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nonspherical Geometrically Regular Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Irregularly Shaped Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cells in Dense Suspensions and Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . High Field Frequencies and Very Short Pulses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental Determination of ITV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Potentiometric Dyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Image Acquisition and Data Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 4 4 5 6 7 9 10 11 13 14 16 16

Abstract

An exposure of a biological cell to an electric field results in the induced transmembrane voltage (ITV) proportional to the strength of the electric field and superimposed onto the resting transmembrane voltage for the duration of the exposure. The ITV can have a range of effects from modification of the activity of voltage-gated channels to membrane electroporation, and accurate knowledge of spatial distribution and time course of the ITV is important both for the studies of these phenomena and for effectiveness of their applications. Unlike the resting component of the transmembrane voltage, the induced component varies with position on the membrane, it depends on the shape of the cell and its orientation with respect to the electric field, and in dense cell

T. Kotnik (*) Faculty of Electrical Engineering, University of Ljubljana, Ljubljana, Slovenia e-mail: [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_8-1

1

2

T. Kotnik

suspensions and tissues also on the volume fraction occupied by the cells. Inducement of the ITV is a process characterized by a time constant, which amounts to tenths of a microsecond under physiological conditions. As a consequence, the time course of the ITV lags the time course of the electric field that induces it, and for exposures to alternating fields with frequencies above 1 MHz or to pulses with durations below 1 μs, the amplitude of the ITV induced by the field of a given amplitude starts to decrease with further increase of the field frequency or with further decrease of the pulse duration. With field frequencies approaching the GHz range or with pulse durations in the ns range, this attenuation of the ITV comes to a halt, and the voltages induced on the organelle membranes inside the cell can reach the same order of magnitude as the voltage induced by the same field on the plasma membrane, and under certain conditions even exceed it. After the description of methods for analytical derivation and numerical computation of the ITV, the main techniques for experimental determination of ITV are also outlined. Keywords

Induced transmembrane voltage • Electric pulses • Plasma membrane • Organelle membranes • Schwan’s equation • Finite-elements method • Potentiometric dyes • Electroporation

Introduction From the electrical point of view, the cell can roughly be described as an electrolyte (the cytoplasm) surrounded by an electrically insulating shell (the plasma membrane), and the natural surroundings of a cell typically also resemble an electrolyte quite closely. Under such conditions, when a cell is exposed to an electric field, the field locally concentrates within the membrane, which thus shields the cytoplasm from the exposure. For this reason, the internal structure of the cell is unimportant, except with rapidly time-varying electric fields, where, as discussed in section “High Field Frequencies and Very Short Pulses,” the transient exposure of the intracellular structures to the field must also be taken into account for proper analysis, and both the membrane and its surroundings have to be treated as materials with a both nonzero electric conductivity and a nonzero dielectric permittivity. The concentration of the electric field inside the membrane results in a considerable electric potential difference across the membrane, termed the induced transmembrane voltage (ITV), which superimposes onto the resting transmembrane voltage (RTV) resulting from the imbalances between intracellular and extracellular ion concentrations and amounting under physiological conditions to tens of millivolts. After the electric field ceases, so does the ITV, and only the RTV remains on the membrane. Modification of transmembrane voltage through ITV can affect the functioning of voltage-gated membrane channels, thus initiating action potentials and stimulating excitable cells (Bedlack et al. 1994; Cheng et al. 1999; Burnett

Transmembrane Voltage Induced by Applied Electric Fields

3

et al. 2003), and when sufficiently large it can also lead to either reversible or irreversible electroporation of the cells’ plasma membrane (Neumann et al. 1999), with the porated membrane regions closely correlated with the regions of the highest absolute value of transmembrane voltage (Kotnik et al. 2010). It should be noted that not only does the ITV affect electroporation, but once the membrane is electroporated, its significantly increased electric conductivity also affects the ITV, which decreases from the value induced immediately before the onset of electroporation, and generally the more intense the electroporation, the larger is this decrease. Using Ψ to denote electric potential, transmembrane voltage is the difference between the potentials at the inner and the outer membrane surface, Ψint and Ψext, and in physiology RTV is always defined and measured as ΔΨm ¼ Ψint  Ψext ;

(1)

so that a negative value reflects a lower electric potential of the cytosol with respect to the cell exterior. For superposition of the ITV onto the RTV to be applicable also formally, one should adhere to Eq. 1 also when characterizing the ITV, and this chapter does so; thus, the RTV can be superimposed to the ITV by simple mathematical addition to yield the total transmembrane voltage, and this chapter henceforth only deals with the induced transmembrane voltage, using the acronym ITV in the text and the notation ΔΨm in the formulae. From the geometrical point of view, the cell can be characterized as a geometric body (the cytoplasm) surrounded by a shell of uniform thickness (the membrane). For suspended cells, the simplest model of the cell is a homogeneous sphere surrounded by a spherical shell. For augmented generality, the sphere can be replaced by a cylinder (e.g., as a model of the axon of a neuron), a spheroid (e.g., an oblate spheroid as a model of an erythrocyte or a prolate spheroid as a model of a bacillus), or an ellipsoid, but with spheroids and ellipsoids the realistic requirement of uniform membrane thickness complicates the geometrical description of the shell: if its outer surface is characterized as a spheroid or an ellipsoid, its inner surface lacks a simple geometrical characterization and vice versa. The steady-state ITV on the plasma membrane of such cells can, however, still be determined analytically. Spheres, spheroids, and ellipsoids are reasonable models for most suspended cells but not for those in tissues. No simple geometrical body can model a typical cell in a tissue, where furthermore each cell generally differs in its shape from the rest. With cells of irregular shape and/or close to each other, the ITV cannot be derived analytically and cannot be formulated as an elementary mathematical function. This deprives us of some of the insight available from explicit expressions, but with spatial and temporal discretization on modern computers, the ITV induced on an irregular cell of interest can still be estimated rather accurately, provided that the three-dimensional shape of this cell is determined with sufficient accuracy. Numerical computation is generally also required to assess the ITV after the cell is electroporated, as this increases the electric conductivity of the membrane significantly (invalidating its treatment as an insulating shell) and moreover nonuniformly.

4

T. Kotnik

Section “Analytical Derivation and Numerical Computation of ITV” provides a more detailed description of methods for analytical derivation and numerical computation of ITV, while section “Experimental Determination of ITV” outlines the main techniques for experimental determination of ITV.

Analytical Derivation and Numerical Computation of ITV Spherical Cells For an exposure to a homogeneous electric field, the ITV is determined by solving Laplace’s equation. Although biological cells are not perfect spheres, in the simplest theoretical treatment they are considered as such, and for an approximation, the plasma membrane can be treated as nonconductive. Under these assumptions, the process of solving Laplace’s equation with appropriate boundary conditions (Kotnik and Pucihar 2010) yields a description of the ITV often referred to as the steady-state Schwan’s equation (Pauly and Schwan 1959), 3 ΔΨm ¼ E R cos θ ; 2

(2)

where E is the strength of the electric field in the region where the cell is situated (i.e., the local electric field), R is the cell radius, and θ is the angle measured from the center of the cell with respect to the direction of the field. The ITV on the membrane of a spherical cell is thus proportional to the applied electric field and to the cell radius and has extremal values at the points where the field is perpendicular to the membrane, i.e., at θ = 0 and θ = 180 (the “poles” of the cell), while in-between these poles it varies proportionally to the cosine of θ (see Fig. 1, dashed). The value of the ITV given by Eq. 2 is typically established several μs after the onset of the electric field. With exposures to a DC field lasting hundreds of microseconds or more, this formula can safely be applied to yield the maximal, steady-state value of the ITV. To describe its transient behavior during the initial microseconds, one uses the first-order Schwan’s equation (Pauly and Schwan 1959), 3 ΔΨm ¼ E R cos θ ð1  expðt=τm ÞÞ ; 2

(3)

where τm is the time constant of membrane charging, τm ¼

R em σi σe 2d þ Rσ m σ i þ 2σ e

(4)

with σi, σm, and σe the conductivities of the cytoplasm, cell membrane, and extracellular medium, respectively, εm the dielectric permittivity of the membrane, d the membrane thickness, and R again the cell radius.

Transmembrane Voltage Induced by Applied Electric Fields

5

Fig. 1 Normalized steady-state ITV as a function of the polar angle θ for spheroidal cells with the axis of rotational symmetry aligned with the direction of the field. Solid: a prolate spheroidal cell with R2 = 0.2  R1. Dashed: a spherical cell, R2 = R1 = R. Dotted: an oblate spheroidal cell with R2 = 5  R1

In certain experiments in vitro, where artificial extracellular media with conductivities substantially lower than physiological are used, the factor 3/2 in Eqs. 2 and 3 decreases in value, as described in detail in (Kotnik et al. 1997). But generally, Eqs. 3 and 4 are applicable to exposures to sinusoidal (AC) electric fields with frequencies below 1 MHz and to electric pulses longer than 1 μs. To determine the ITV induced by even higher field frequencies or even shorter pulses, the dielectric permittivities of the electrolytes on both sides of the membrane also have to be accounted for. This leads to a further generalization of Eqs. 3 and 4 to the second-order Schwan equation (Grosse and Schwan 1992; Kotnik et al. 1998; Kotnik and Miklavčič 2000a), and the results it yields will be outlined in subsection “High Field Frequencies and Very Short Pulses.”

Nonspherical Geometrically Regular Cells Another direction of generalization is to assume a cell shape more general than that of a sphere. The most straightforward generalization is to a spheroid (a geometrical body obtained by rotating an ellipse around one of its radii, so that one of its orthogonal projections is a sphere and the other two are the same ellipse) and further

6

T. Kotnik

to an ellipsoid (a geometrical body in which each of its three orthogonal projections is a different ellipse). To obtain the analogues of Schwan’s equation for such cells, one solves Laplace’s equation in spheroidal and ellipsoidal coordinates, respectively (Kotnik and Miklavčič 2000b; Gimsa and Wachner 2001; Valič et al. 2003). Besides the fact that this solution is by itself somewhat more intricate than the one in spherical coordinates, the generalization of the shape invokes two additional complications outlined in the next two paragraphs. A description of a cell is geometrically realistic if the thickness of its membrane is uniform. This is the case if the membrane represents the space between two concentric spheres but not with two confocal spheroids or ellipsoids. As a result, the thickness of the membrane modeled in spheroidal or ellipsoidal coordinates is necessarily nonuniform. By solving Laplace’s equation in these coordinates, one thus obtains the spatial distribution of the electric potential in a nonrealistic setting. However, for cells surrounded by a physiological medium and with nonporated membranes, the electric conductivity of the membrane can be neglected (i.e., the membrane is treated as an insulator), and the steady-state ITV obtained in this manner is still realistic, as the electric potential everywhere inside the cytoplasm is constant, and the geometry of the inner surface of the membrane does not affect the potential distribution outside the cell, which is the same as if the cell would be a homogeneous nonconductive body of the same shape; a rigorous discussion of the validity of this approach is given in (Kotnik and Miklavčič 2000b). Figure 1 shows the ITV for a spherical cell, as described by the steady-state Schwan’s equation, and for two spheroidal cells with their axis of rotational symmetry aligned with the direction of the field, as described by the expressions analogous to the Schwan’s equation that are valid for such cell shapes (Kotnik and Pucihar 2010). For nonspherical cells, it is generally more revealing to express the ITV as a function of the arc length along the membrane rather than as a function of the angle θ on the membrane (for a sphere, the two quantities are directly proportional). For uniformity, the normalized version of the arc length is used, denoted by p and increasing from 0 to 1 equidistantly along the arc of the membrane. This is depicted in Fig. 2 for the three cells for which the ITV as a function of θ is shown in Fig. 1, and in this chapter all subsequent plots of the ITV for nonspherical cells will be presented in this manner. The second complication of generalizing the cell shape from a sphere to a spheroid or an ellipsoid is that the ITV now also depends on the orientation of the cell with respect to the electric field. To deal with this, one decomposes the field vector into the components parallel to the axes of the spheroid or the ellipsoid and writes the ITVas a corresponding linear combination of the ITV for each of the three coaxial orientations (Gimsa and Wachner 2001; Valič et al. 2003). Figures 3 and 4 show the effect of rotation of two different spheroids with respect to the direction of the field.

Irregularly Shaped Cells For an irregularly shaped cell, the ITV cannot be expressed as an elementary mathematical function, since for such a geometry Laplace’s equation is not solvable

Transmembrane Voltage Induced by Applied Electric Fields

7

Fig. 2 Normalized steady-state ITV as a function of the normalized arc length p for spheroidal cells with the axis of rotational symmetry aligned with the direction of the field. Solid: a prolate spheroidal cell with R2 = 0.2  R1. Dashed: a spherical cell, R2 = R1 = R. Dotted: an oblate spheroidal cell with R2 = 5  R1

analytically, but using modern computers and the finite-elements method implemented in software packages such as COMSOL Multiphysics, the ITV on a given irregular cell can be determined numerically (Pucihar et al. 2006; Pucihar et al. 2009a). With a sufficiently accurately determined three-dimensional shape of the cell and with sufficiently fine spatial and temporal resolution, the ITVobtained in this manner is quite accurate, but it should be kept in mind that the set of ITV values so obtained over the membrane of the cell under consideration is only applicable to this cell, i.e., for the cell of the particular shape and size for which the threedimensional shape has been determined and on which the ITV has been computed. Figure 5 shows examples of two cells growing in a Petri dish and the ITV computed on their membranes.

Cells in Dense Suspensions and Tissues In natural situations the cells are rarely isolated, and when they are sufficiently close to each other, the mutual distortion of the field caused by their proximity becomes nonnegligible. Often, the cells are also in direct contact, forming two-dimensional (monolayers attached to the bottom of a dish) or three-dimensional (tissues) structures, and they can even be electrically interconnected.

8

T. Kotnik

Fig. 3 Normalized steady-state ITV as a function of p for a prolate spheroidal cell with R2 = 0.2  R1. Solid: axis of rotational symmetry (ARS) aligned with the field. Dashed: ARS at 45 with respect to the field. Dotted: ARS perpendicular to the field

In dilute cell suspensions, the distance between the cells is much larger than the cells themselves, and the local field outside each cell is almost unaffected by the presence of other cells. Thus, for cells representing less than 1 % of the suspension volume (for spherical cells with radii of 10 μm, this corresponds to up to 2 million cells/ml), the deviation of the actual ITV from the one predicted by Schwan’s equation is negligible. However, as the volume fraction occupied by the cells gets larger, the distortion of the local field around each cell by the presence of other cells in the vicinity becomes more pronounced, and the ITV starts to differ noticeably from the prediction yielded by Schwan’s equation (Fig. 6), and an accurate estimation of the ITV must be assessed either numerically of by analytical approximations (Susil et al. 1998; Pavlin et al. 2002). Regardless of the volume fraction the cells occupy, suspended cells float rather freely, so their arrangement is on average rather uniform, resembling a face-centered cubic lattice, which is thus the most appropriate choice for models of dense cell suspensions (Pavlin et al. 2002; Pucihar et al. 2007). For even larger volume fractions of the cells, the electrical properties of the suspension start to approach that of a tissue but only to a certain extent; the arrangement of cells in tissues does not necessarily resemble a face-centered lattice, since cells differ from each other, form specific structures (e.g., layers), and can also be directly electrically coupled (e.g., through gap junctions).

Transmembrane Voltage Induced by Applied Electric Fields

9

Fig. 4 Normalized steady-state ITV as a function of p for an oblate spheroidal cell with R2 = 5  R1. Solid: axis of rotational symmetry (ARS) aligned with the field. Dashed: ARS at 45 with respect to the field. Dotted: ARS perpendicular to the field

High Field Frequencies and Very Short Pulses The time constant of membrane charging (τm) given by Eq. 4 implies that there is a lag between the time courses of the external field and the ITV induced by this field. As mentioned above, τm (and thus the lag) is in tenths of a microsecond under physiological conditions but can be longer when cells are suspended in a medium with electric conductivity substantially below physiological levels. For alternating (AC) fields with the oscillation period much longer than τm, as well as for rectangular pulses much longer than τm, the effect of this lag on the amplitude of the ITV is negligible. However, for AC fields with the period shorter than τm, as well as for pulses shorter than τm, the amplitude of the ITV gets attenuated by the lag, and with further increase in field frequency or decrease in pulse duration, the amplitude of the ITV induced by the field of a given amplitude starts to decrease. Figure 7 shows this attenuation for a spherical cell; the low-frequencies plateau of the amplitude and its intermediate-frequencies decrease are both described by the first-order Schwan’s equation, but the high-frequencies plateau is only described by the second-order version of this equation, for derivation of which all electric conductivities and dielectric permittivities are taken into account (Kotnik et al. 1998; Kotnik and Miklavčič 2000a).

10

T. Kotnik

Fig. 5 Normalized steady-state ITV as a function of p for two irregularly shaped cells growing on the flat surface of a Petri dish, with the inset showing the top view of the cells

With field frequencies approaching the GHz range or with pulse durations in the ns range, this attenuation of the ITV comes to a halt, and the voltages induced on the organelle membranes inside the cell are of the same order of magnitude as the voltage induced by the same field on the plasma membrane. In certain circumstances, particularly if the organelle interior is electrically more conductive than the cytosol or if the organelle membrane has a lower dielectric permittivity than the cell membrane, the ITV on the membrane of such an organelle can temporarily even exceed the ITV on the plasma membrane (Kotnik and Miklavčič 2006). In principle, this could provide a theoretical explanation for reports that very short and intense electric pulses (tens of ns, millions of V/m) can also induce electroporation of organelle membranes (Schoenbach et al. 2001; Beebe et al. 2003; Tekle et al. 2005; Batista Napotnik et al. 2016).

Experimental Determination of ITV An alternative to the analytical and numerical methods for determining the ITV are the experimental techniques – measurements of the ITV with microelectrodes and with potentiometric fluorescent dyes. Microelectrodes were used in pioneering measurements of the action potential propagation, first conventionally (Ling and Gerard 1949) and later as patch-clamp electrodes (Neher and Sakmann 1976),

Transmembrane Voltage Induced by Applied Electric Fields

11

Fig. 6 Normalized steady-state ITV as a function of p for spherical cells in suspensions of various densities (intercellular distances). Solid: The analytical result for a single cell as given by Eq. 2. Dashed: numerical results for cells arranged in a face-centered cubic lattice and occupying (with decreasing dash size) 10 %, 30 %, and 50 % of the total suspension volume

preferred for their simple use and high temporal resolution. However, the invasive nature of measurements and low spatial resolution are considerable shortcomings of this approach. Furthermore, the physical presence of the conductive electrodes during the measurement affects the distribution of the electric field around them and thus also the ITV. In contrast, measurements by means of potentiometric dyes are noninvasive, offer much higher spatial resolution, and do not distort the electric field and consequently the ITV. In addition, a potentiometric measurement can be performed simultaneously on a number of cells. For these reasons, the potentiometric dyes have during the last decades become the predominant method in experimental studies and measurements of the ITV, and the remainder of this chapter focuses exclusively on this approach.

Potentiometric Dyes Based on their response mechanism, potentiometric dyes are divided into two classes: (i) slow potentiometric dyes that are translocated across the membrane by an electrophoretic mechanism, which is accompanied by a fluorescence change and (ii) fast potentiometric dyes that incorporate into the membrane, with energy levels

12

T. Kotnik

Fig. 7 The amplitude of normalized steady-state ITV as a function of the frequency of the AC field. The dashed curve shows the first-order and the solid curve the second-order Schwan’s equation. Note that both axes are logarithmic

of their electrons and consequently their fluorescence properties dependent on transmembrane voltage. Electric pulses used in electrophysiological and electroporation-based applications usually have durations in the range of microseconds to milliseconds. To measure the ITV induced by such pulses, fast potentiometric dyes have to be used. These respond to changes in the ITV within microseconds or less, which makes them suitable even for measurements of the transient effects. Slow dyes, on the other hand, need several seconds to respond to a change in the transmembrane voltage, and while some of them, particularly rhodamine and several carbocyanines, are efficient in the measurements of the RTV, their tardiness makes them largely useless for analysis of the ITV. One of the most widely used fast potentiometric dyes is di-8-ANEPPS (di-8butylamino-naphthyl-ethylene-pyridinium-propyl-sulfonate), developed by Leslie Loew and colleagues at the University of Connecticut (Gross et al. 1986; Loew 1992). This dye is nonfluorescent in water but becomes strongly fluorescent when incorporated into the lipid bilayer of the cell membrane, thereby making the membrane highly visible. This enables the construction of spatial numerical models of cells from a set of cross-section fluorescence images (Pucihar et al. 2006), and thus provides a possibility to compute the ITV on the same cells on which an experiment is carried out.

Transmembrane Voltage Induced by Applied Electric Fields

13

The fluorescence intensity of di-8-ANEPPS varies proportionally to the change of the ITV; the response is linear for voltages ranging from 280 to +250 mV (Lojewska et al. 1989; Cheng et al. 1999). Relatively small changes in fluorescence of the dye, uneven membrane staining, and dye internalization make di-8-ANEPPS less suitable for measurements of the resting membrane voltage, although such efforts were also reported (Zhang et al. 1998). It is, however, well suited for measuring larger changes in membrane voltage, such as the onset of the ITV in nonexcitable cells exposed to external electric fields (Gross et al. 1986; Montana et al. 1989) or action potentials in excitable cells (Bedlack et al. 1994; Cheng et al. 1999). In addition, di-8-ANEPPS allows for determination of the ITV by ratiometric measurements of fluorescence excitation (Montana et al. 1989) or emission (Knisley et al. 2000), which increases the sensitivity of the response.

Image Acquisition and Data Processing As the sensitivity of fast potentiometric dyes to the changes of the ITV is low (typically, a change by 100 mV results in the change of fluorescence intensity by 2–12 %), the fluorescence changes are hardly discernible by the naked eye but become apparent with appropriate image processing and analysis. This procedure is performed in several steps (Pucihar et al. 2009b). First, a pair of images is acquired: the control image (immediately before the exposure to the electric field) and the exposure image (during the delivery of the electric field pulse); to get a more reliable measurement, a sequence of pulses can be applied, with both the control and the exposure image acquired for each pulse. Second, the background fluorescence is subtracted from both images. Third, the region of interest corresponding to the membrane is determined for the cell under investigation, and the fluorescence intensities along this region in the control and pulse image are measured. Fourth, for each pulse, the control data are subtracted from the pulse data, and the result divided by the control data to obtain the relative fluorescence changes; if a sequence of pulses is applied, the values of relative fluorescence changes determined for each pulse can be averaged. Finally, the relative fluorescence changes are translated into ITV using the di-8-ANEPPS calibration curve; a rough estimation of this curve can be obtained from the literature, but for higher accuracy, it must be measured for each particular setup, as shown in Fig. 8. Calibration is performed with either (i) valinomycin, a potassium ionophore, and a set of different potassium concentrations in external medium (Montana et al. 1989; Pucihar et al. 2006, or (ii) patch clamp in voltage-clamp mode (Zhang et al. 1998). For a graphical description of the results, the ITV can be plotted as a function of the relative arc length. A brief summary of the described process in given in Fig. 9, showing a cell stained with di-8-ANEPPS, the processed image reflecting the ITV, and the plot of the ITV along the cell membrane. To remove some of the noise inherent to potentiometric measurements, the ITV curve can be smoothed using a suitable filter (e.g., a moving average).

14

T. Kotnik

Fig. 8 An example of the experimental calibration curve for measurements of the ITV with di-8ANEPPS

Conclusions Knowledge and understanding of transmembrane voltage and the process of its inducement in cells exposed to electric fields are important in electrophysiology, and perhaps even more so in research of electroporation and development of electroporation-based technologies and treatments. As outlined in this chapter, inducement of the ITV is a process, and although for alternating electric fields with frequencies up to the kHz range and for electric pulses lasting tens of μs or longer, it can be approximated as instantaneous, for higher frequencies and shorter pulses the inducement’s time course must also be taken into account for an accurate description of the ITV and its effects. In the context of electroporation, the key aspect is the close correlation between the ITV and the electroporation-mediated transport across the membrane: the membrane regions with the highest absolute value of the ITV are commonly also the ones where the transport proceeds the fastest. While the ITV induced by a given electric field can generally be predicted quite accurately by means of analytical or numerical calculation, a valuable experimental complement in determination of the ITV is found in fast potentiometric dyes that allow, with adequate sample preparation, dye fluorescence calibration, image acquisition, and data processing to determine the actual ITV and compare it to the calculated predictions.

Transmembrane Voltage Induced by Applied Electric Fields

15

Fig. 9 Determination of the steady-state ITV with di-8-ANEPPS. Top left: Raw fluorescence image of a cell stained with di-8-ANEPPS. Top right: Processed image reflecting the ITV. Bottom: steadystate ITV as a function of p determined from the processed image using the calibration curve shown in Fig. 8

Acknowledgment This work was supported by the Slovenian Research Agency (Grant P2-0249) and conducted in the scope of the European Laboratory of Pulsed Electric Fields Applications (LEA EBAM) and within networking efforts of the COST Action TD1104 – European Network for Development of Electroporation-Based Technologies and Treatments (EP4Bio2Med).

16

T. Kotnik

Cross-References ▶ Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Experiments ▶ Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Molecular Models ▶ Electropore Energy and Thermodynamics ▶ Lipid Pores: Molecular and Continuum Models

References Batista Napotnik T, Reberšek M, Vernier PT, Mali B, Miklavčič D (2016) Effects of high voltage nanosecond electric pulses on eukaryotic cells (in vitro): a systematic review. Bioelectrochemistry 110:1–12. doi:10.1016/j.bioelechem.2016.02.011 Bedlack RS, Wei M, Fox SH, Gross E, Loew LM (1994) Distinct electric potentials in soma and neurite membranes. Neuron 13:1187–1193. doi:10.1016/0896-6273(94)90056-6 Beebe SJ, Fox PM, Rec LJ, Willis EL, Schoenbach KH (2003) Nanosecond, high-intensity pulsed electric fields induce apoptosis in human cells. FASEB J 17:1493–1495. doi:10.1096/ fj.02.0859fje Burnett P, Robertson JK, Palmer JM, Ryan RR, Dubin AE, Zivin RA (2003) Fluorescence imaging of electrically stimulated cells. J Biomol Screen 8:660–667. doi:10.1177/1087057103258546 Cheng DK, Tung L, Sobie EA (1999) Nonuniform responses of transmembrane potential during electric field stimulation of single cardiac cells. Am J Physiol 277:H351–H362 Gimsa J, Wachner D (2001) Analytical description of the transmembrane voltage induced on arbitrarily oriented ellipsoidal and cylindrical cells. Biophys J 81:1888–1896. doi:10.1016/ S0006-3495(01)75840-7 Gross D, Loew LM, Webb W (1986) Optical imaging of cell membrane potential changes induced by applied electric fields. Biophys J 50:339–348. doi:10.1016/S0006-3495(86)83467-1 Grosse C, Schwan HP (1992) Cellular membrane potentials induced by alternating fields. Biophys J 63:1632–1642. doi:10.1016/S0006-3495(92)81740-X Knisley SB, Justice RK, Kong W, Johnson PL (2000) Ratiometry of transmembrane voltagesensitive fluorescent dye emission in hearts. Am J Physiol Heart Circ Physiol 279: H1421–H1433 Kotnik T, Miklavčič D (2000a) Second-order model of membrane electric field induced by alternating external electric fields. IEEE Trans Biomed Eng 47:1074–1081. doi:10.1109/ 10.855935 Kotnik T, Miklavčič D (2000b) Analytical description of transmembrane voltage induced by electric fields on spheroidal cells. Biophys J 79:670–679. doi:10.1016/S0006-3495(00)76325-9 Kotnik T, Miklavčič D (2006) Theoretical evaluation of voltage inducement on internal membranes of biological cells exposed to electric fields. Biophys J 90:480–491. doi:10.1529/ biophysj.105.070771 Kotnik T, Pucihar G (2010) Induced transmembrane voltage – theory, modeling, and experiments. In: Pakhomov AG, Miklavčič D, Markov MS (eds) Advanced electroporation techniques in biology and medicine. CRC Press, Boca Raton, pp 51–70 Kotnik T, Bobanović F, Miklavčič D (1997) Sensitivity of transmembrane voltage induced by applied electric fields – a theoretical analysis. Bioelectrochem Bioenerg 43:285–291. doi:10.1016/S0302-4598(97)00023-8 Kotnik T, Miklavčič D, Slivnik T (1998) Time course of transmembrane voltage induced by timevarying electric fields – a method for theoretical analysis and its application. Bioelectrochem Bioenerg 45:3–16. doi:10.1016/S0302-4598(97)00093-7

Transmembrane Voltage Induced by Applied Electric Fields

17

Kotnik T, Pucihar G, Miklavčič D (2010) Induced transmembrane voltage and its correlation with electroporation-mediated molecular transport. J Memb Biol 236:3–13. doi:10.1007/s00232010-9279-9 Ling G, Gerard RW (1949) The normal membrane potential of frog sartorius fibers. J Cell Comp Physiol 34:383–396 Loew LM (1992) Voltage sensitive dyes: measurement of membrane potentials induced by DC and AC electric fields. Bioelectromagn Suppl 1:179–189. doi:10.1002/bem.2250130717 Lojewska Z, Farkas DL, Ehrenberg B, Loew LM (1989) Analysis of the effect of medium and membrane conductance on the amplitude and kinetics of membrane potentials induced by externally applied electric fields. Biophys J 56:121–128. doi:10.1016/S0006-3495(89)82657-8 Montana V, Farkas DL, Loew LM (1989) Dual-wavelength ratiometric fluorescence measurements of membrane-potential. Biochemistry 28:4536–4539. doi:10.1021/bi00437a003 Neher E, Sakmann B (1976) Single-channel currents recorded from membrane of denervated frog muscle fibres. Nature 260:779–802 Neumann E, Kakorin S, Toensing K (1999) Fundamentals of electroporative delivery of drugs and genes. Bioelectrochem Bioenerg 48:3–16. doi:10.1016/S0302-4598(99)00008-2 Pauly H, Schwan HP (1959) Über die Impedanz einer Suspension von kugelförmigen Teilchen mit einer Schale. Z Naturforsch 14B:125–131 Pavlin M, Pavšelj N, Miklavčič D (2002) Dependence of induced transmembrane potential on cell density, arrangement, and cell position inside a cell system. IEEE Trans Biomed Eng 49:605–612. doi:10.1109/TBME.2002.1001975 Pucihar G, Kotnik T, Valič B, Miklavčič D (2006) Numerical determination of the transmembrane voltage induced on irregularly shaped cells. Ann Biomed Eng 34:642–652. doi:10.1007/ s10439-005-9076-2 Pucihar G, Kotnik T, Teissié J, Miklavčič D (2007) Electroporation of dense cell suspensions. Eur Biophys J 36:173–185. doi:10.1007/s00249-006-0115-1 Pucihar G, Miklavčič D, Kotnik T (2009a) A time-dependent numerical model of transmembrane voltage inducement and electroporation of irregularly shaped cells. IEEE T Biomed Eng 56:1491–1501. doi:10.1109/TBME.2009.2014244 Pucihar G, Kotnik T, Miklavčič D (2009b) Measuring the induced membrane voltage with di-8ANEPPS. J Vis Exp 33:1659. doi:10.3791/1659 Schoenbach KH, Beebe SJ, Buescher ES (2001) Intracellular effect of ultrashort electrical pulses. Bioelectromagnetics 22:440–448. doi:10.1002/bem.71 Susil R, Šemrov D, Miklavčič D (1998) Electric field induced transmembrane potential depends on cell density and organization. Electro Magnetobiol 17:391–399 Tekle E, Oubrahim H, Dzekunov SM, Kolb JF, Schoenbach KH, Chock PB (2005) Selective field effects on intracellular vacuoles and vesicle membranes with nanosecond electric pulses. Biophys J 89:274–284. doi:10.1529/biophysj.104.054494 Valič B, Golzio M, Pavlin M, Schatz A, Faurie C, Gabriel B, Teissié J, Rols MP, Miklavčič D (2003) Effect of electric field induced transmembrane potential on spheroidal cells: theory and experiment. Eur Biophys J 32:519–528. doi:10.1007/s00249-003-0296-9 Zhang J, Davidson RM, Wei MD, Loew LM (1998) Membrane electric properties by combined patch clamp and fluorescence ratio imaging in single neurons. Biophys J 74:48–53. doi:10.1016/ S0006-3495(98)77765-3

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and Simulation of Related Phenomena Nikolai Lebovka and Eugene Vorobiev

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroporation of Individual Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroporation of Individual Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pulsed Electric Field Treatment of Suspension of Cells and Plant Tissues . . . . . . . . . . . . . . . . . . . . . Corrections of Schwan’s Equation for Concentrated Cell Ensembles . . . . . . . . . . . . . . . . . . . . . . . Kinetic Models of Inactivation in Suspension of Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plant Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 3 6 7 7 9 10 11 13 16 23 23 24

N. Lebovka (*) Department of Physical Chemistry of Dirperse Minerals, Institute of Biocolloidal Chemistry named after F. D. Ovcharenko, NAS of Ukraine, Kyiv, Ukraine Laboratoire de Transformations Intégrées de la Matière Renouvelable, EA 4297, Centre de Recherches de Royallieu, Sorbonne Universités, Université de Technologie de Compiègne, Compiègne, France e-mail: [email protected] E. Vorobiev (*) Laboratoire de Transformations Intégrées de la Matière Renouvelable, EA 4297, Centre de Recherches de Royallieu, Sorbonne Universités, Université de Technologie de Compiègne, Compiègne, France e-mail: [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_9-1

1

2

N. Lebovka and E. Vorobiev

Abstract

Pulsed electric fields (PEF) are very promising for the treatment of plant (vegetable, fruits) and animal tissues. PEF-assisted processing which gives new perspectives for selective extraction and expression, dehydration, drying, freezing, and osmotic treatment can be marked out. However, the nature of electroporation in food tissues is very complicated and can include different scales starting from single membranes and cells to concentrated suspensions and real plant tissues with complex intrinsic structure and much heterogeneity in different scales. Effects of electroporation depend on the size of cells, their orientation and spatial distribution in space, passive and active electrophysical parameters of cells, pH of media, and the presence in tissue osmotic agents. Optimization of PEF treatment and PEF-assisted processing of plant tissues requires precise adaptation of PEF protocols, selection of optimal electric field strength, pulse duration, repetition time, and temperature of treatment. Mathematical modeling can provide detailed and valuable information for optimization of PEF protocols, power consumption, and preservation of the food quality. Moreover, numerical simulation can be useful for description of PEF effects in the presence of complex transport processes and changes in temperature, electrical conductivity, electric field strength, and spatial distribution of electrolyte inside the tissue during PEF treatment. In recent decades various models and simulations were applied for optimizing PEF treatment of liquid foods, geometrical optimization of the treatment chambers, and prediction of the electric field distribution, flow velocity, and temperature inside the treatment chamber. This chapter discusses various models of PEF treatment and computer simulation of related phenomena, which accompanies PEF-assisted processing of plant tissues. Keywords

Pulsed electric fields • Plant tissues • Mathematical models • Computer simulation

Introduction Pulsed electric fields (PEF) have been demonstrated in many useful capacities for the treatment of plant (vegetable, fruits) tissues. Effects of PEF can be explained by electroporation of cellular membranes and removing the membrane’s barrier functions (Weaver and Chizmadzhev 1996). PEF-assisted processing has nonthermal nature and can be achieved by application of moderate electric fields of 500–1000 V/ cm and short treatment times of 10
4–10
2 s (Barba et al. 2015). Among recent examples of PEF applications, selective extraction of valuable components, retention of health-related compounds, enhancement of juices expression, plant dehydration, drying, freezing, and osmotic treatment can be marked out. Moreover, PEF treatment of liquid foods leads to inactivation of microorganisms and quality-

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . .

3

degrading enzymes and gives extension the shelf life of the products (Barba et al. 2015). However, the nature of electroporation is very complex and can include different scales starting from single membranes and cells to concentrated suspensions and real plant tissues with complex intrinsic structure and much heterogeneity in different scales. The efficiency of electroporation itself as well as different processes assisted by PEF depends critically on many parameters of PEF protocol, variation of these parameters during PEF treatment, temperature and such properties of the treated products as the size of cells, their orientation and distribution in space, passive and active electro-physical parameters of cells, pH of media, the presence in tissue osmotic agents, and other factors. Optimization of PEF treatment and PEF-assisted processing of plant tissues requires for accounting of many parameters, which cannot be provided without mathematical models and numerical simulation. Mathematical models and numerical simulation can provide detailed and valuable information for optimization of treatment time, power consumption, and preservation of the food quality. Moreover, numerical simulation can be useful for description of PEF effects in the presence of complex transport processes and changes in temperature, electrical conductivity, electric field strength, and spatial distribution of electrolyte inside the tissue during PEF treatment. Various models and simulations were already successfully applied for optimizing PEF treatment of liquid foods, geometrical optimization of the treatment chambers, and prediction of the electric field distribution, flow velocity, and temperature inside the treatment chamber (Huang et al. 2012; Gerlach et al. 2008). The current chapter aims to provide a discussion on various models of PEF treatment and computer simulation of phenomena induced by PEF in plant tissues. The different scales of electroporation phenomena are also briefly discussed.

Electroporation of Individual Membranes Models The electroporation occurs when the potential difference across the membrane, or the transmembrane potential, um, exceeds a certain threshold, 0.2–1 V (Weaver and Chizmadzhev 1996). This potential difference corresponds to the electrical field strength across the membrane of the order of 4.105–2.106 for a typical membrane thickness of 5 nm. Electroporation reflects the rearrangements of the membrane lipids leading to increased membrane permeability. The time of membrane pore creation tc decreases with increasing of the transmembrane voltage um (Fig. 1). In general, the time evolution of electroporation includes: – Charging and polarization of the membrane (charging time τc > 1 μs) – Rearrangement of the structure of the membrane and creation of pores (Fig. 1)

4

N. Lebovka and E. Vorobiev

Fig. 1 Mean time of membrane pore creation. Filled squares are the experimental data (Melikov et al. 2001), and the solid line is obtained by simulations based on electroporation theory (Compiled from the data presented in (Vasilkoski et al. 2006))

– Aggregation of different pores and noticeable increase (by several orders of magnitude) in ionic, molecular, and nanoparticle transport through the membrane – Membrane recovery or resealing after the removal of the electric field Different non-pore models and pore models were proposed for explanation of the dielectric breakdown of membranes. The non-pore models account for the electrically induced mechanical, hydrodynamic, wave, osmotic, and viscoelastic instabilities of membranes. For example, the following relations for critical voltages ucm were obtained, respectively, for mechanical compression, hydrodynamic, and wave instability models: ucm ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 0:368Y m h2m =eo em

(1a)

ucm ¼

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 0:5γhm =eo em

(1b)

 1=4 ucm ¼ 24Y m γh3m

(1c)

Here Ym is a constant of elasticity, hm is the membrane thickness, em is a relative dielectric constant, and γ is the surface tension of the membrane. However, these non-pore models fail to describe the stochastic nature of membrane rupture and predict the membrane lifetime dependence from the transmembrane voltage τm(um). Better description of the stochastic nature of membrane rupture can be obtained using the pore model. This model suggests that the observed increase in membrane permeability is due to the formation of pores in the lipid bilayer, and it is based on the

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . . 1 Probability of membrane electoporation, Z

Fig. 2 Probability of membrane damage, Z, (see. Eq. 1) versus the duration of electric field application, t, at different values of um. The membrane lifetime τm was evaluated using Eq. 2 with parameters presented in (Lebedeva 1987)

5

um =1 V 0.8 0.8 V 0.6

0.6 V

0.4

0.5 V

0.2

0

10-6 10-5 10-4 10-3 10-2 10-1 Duration of electric field application, t, s

analysis of the energy balance in the membrane. This model is supported by recent molecular dynamics simulations. For example, the transient aqueous pore model gives the following relation for the lifetime of a plane membrane τm (Weaver and Chizmadzhev 1996): h i 2 τm ¼ τ1 m exp ðW m =kT Þ=1 þ ðum =uo Þ ;

(2)

where W m =kT ¼ πω2 =kTγ is the relative activation energy of electroporation, ω is pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi the line tension of a membrane, uo ¼ 2γ=ðCm ðew =em
1ÞÞ is the voltage parameter, Cm is the specific capacitance of a membrane, ew  80 is the a relative dielectric constant of water, and τ1 m is the limiting value of τm (at um!1 or T!1). For the typical values of parameters (for general lipid membranes at T = 298 K) γ  2.10
3 N/m, ω  1.69.10
11 N, em  2, and Cm  3.5.10
3 F/m2, we have uo  0.17 V and Wm/kT  109. Experimental estimation for plane membrane gave also τ1  3.7.10
7 s (Lebedeva 1987). The probability of membrane damage, Z, during the electric field application with duration of t can be estimated from the first-order kinetic equation: Z ¼ 1
expð
t=tm Þ:

(3)

Figure 2 presents Z(t) dependences at different values of um. Electroporation can be depending on pulsed protocol (mainly on the electric field strength and exposure time). In irreversible electroporation, the pores in the membrane do not reseal. Such type of electroporation is commonly used for killing of microorganisms and for plant cell damage in food and biotechnological applications. From the other hand, in reversible electroporation, the pores can reseal when electric field is switched off.

6

N. Lebovka and E. Vorobiev

Such type of electroporation is used in electrochemotherapy, for transdermal drug delivery and cell fusion (Miklavčič et al. 2014).

Simulation It is believed that the nature of electroporation is rather similar for the planar lipid bilayers and biological cell membranes. Up to date the extensive computer simulations of electroporation of lipid bilayers have been performed. The molecular dynamics models are based on simulation of a trajectory of molecules in time and allow handling of tens of thousands of particles. The different atomistic and coarsegrained approaches allowed studying the processes of the electropore formation, formation of transient water pores induced by ion charge imbalance, influence of cholesterol content in the bilayer on electroporation time, formation of water channels, etc. (Rems and Miklavčič 2016). Simulations evidenced that pore formation in a high intense electric field is driven by local electric field gradients at the water/lipid interface. The first stage of pore formation includes formation of a water bridge across the lipid bilayers. Pores up to 10 nm in diameter can be created depending on the electric field strength and duration of the applied field.

Electroporation of Individual Cells Models In the individual biological cell, the thin membrane with low conductivity σ m separates two high conductive interior and exterior electrolyte media with conductivities σ d and σ, respectively (Fig. 3). The values of transmembrane potential um have been analytically evaluated for individual cells of different geometries. For example, for the spherical cell the Schwan’s equation gives the following expression for static value of um (Weaver and Chizmadzhev 1996): um ¼ 1:5f e ER cos φ

(4)

where R is the radius of the cell, and φ is the angle between direction of E and position of the observation point on the surface of a membrane. The electroporation factor fe(1) depends on the electrical conductivities of all components σ m, σ, and σ d, as well as on the ratio hm/R, where hm (5 nm) is a membrane thickness (Kotnik et al. 1998): f e ¼ ð3hm =RÞσ d σ=½ðσ m þ 2σ Þðσ m þ 0:5σ d Þ
ð1
3hm =RÞðσ
σ m Þðσ d
σ m Þ (5) For the spherical cell, the transmembrane potential and lifetime of the membrane depend upon the local position on the surface of the cell, i.e., on the φ value. The

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . .

7

Fig. 3 Induction of transmembrane potential for spherical cell in the external electric field E External electric field, E

j

R sd s

sm

hm

um(j)

Fig. 4 Relative lifetime, τ* = τ/τ1, versus relative electric field strength, E*, for spherical cell (E* = 1.5ER/uo) and plane membrane (E* = um/uo). See the text for the detail description (Compiled from the data presented in (Lebovka et al. 2002))

average lifetime of a spherical cell, τc, can be estimated accounting for the distribution of um(φ) over the surface of a cell: τ
! c

ðπ

¼ τ
1 m ðφÞd cos φ

(6)

0

Figure 4 presents relative lifetime of the cell evaluated by numerical integration, τ* = τ/τ1, versus relative electric field strength, E* = 1.5ER/uo (Lebovka

8

N. Lebovka and E. Vorobiev

et al. 2002). Here τ1 is the limiting value of τ (Eq. 1). For comparison the τm =τ1 m versus E* = um/u0 dependence for plane membrane is also presented. For the given E*, the value of τc for the spherical cell exceeds the value of τm for the plane membrane. The τ(E) dependence for the spherical cell can be fitted by the formula that is similar to Eq. 1 for the plane membrane: h  i d τc ¼ τ1 c exp ðW c =kT Þ= 1 þ ðE=Eo Þ

(7)

1 with τ1 c  2:01τm , d  1.33, and Wc  0.54 Wm (see Eq. 1) (Lebovka et al. 2002). For the cell with anisotropic shape, the transmembrane potential, um, depends also on the orientation of the cell in the external fields and shape of the cell (Fig. 5). For the elongated cell, the value of um has its maximum or minimum when the longest axis of the cell is parallel or perpendicular to the electric field, respectively. The more discussion on the problem can be found elsewhere (Kotnik and Miklavčič 2000).

Simulation Schwan’s theory predicts the inversely proportional relationship between the cell radius, R, and the threshold magnitude of the applied electric field, E, i.e., E / 1/R. In

Fig. 5 Induced transmembrane voltage um in units of ER1 as a function of the polar angle φ for three spheroidal cells: R2 = 1/5 R1 (solid line), R2 = R1 (spherical cell, dashed line), and R2 = 5R1 (dotted line). Inset: the three cells and the field orientation (From (Kotnik and Miklavčič 2000) with permission)

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . .

a

9

dM e1+m+e2

b Mele,e1 Rele,e1

Cele,e1

Re1/2

c Ich(Um)

Mele,e2 Rele,e2

Rlip

Cele,e2 Re2/2

2Ce1 SW Rm[ep]

Cm Mm

2Ce2

Fig. 6 The two-dimensional model of the individual spherical cell with two electrolytes separated by a membrane (dark curve) (a). Here the charge transport models of exterior electrolyte Me1 (b) and interior electrolyte Me2 (c) and membrane model, Me1 + m + Me2 (d) are presented (From (Gowrishankar and Weaver 2003) with permission)

general, this prediction is consistent with existent experimental data. For example, the experimentally estimated threshold values of electric field strength are of the order of 100–1000 V/cm for large cells in food tissues (R  30–60 μm) and >10 kV/ cm for small microbial cells (R  1–10 μm) (Barba et al. 2015). Simulation with a finite element method gave excellent correspondence with the Schwan’s theory (Zudans et al. 2007). However, the inverse dependence E / 1/R is not always observed or does not exactly fit Schwan’s equation. To explain this inconsistency, the membrane permeabilization and conductivity changes during the electroporation pulse were accounted (Mercadal et al. 2016). The author numerically modeled electroporation of a single cell using the finite element method (FEM) software platform COMSOL Multiphysics 4.4. By modeling the interplay between the cell membrane conductivity, permeability, and transmembrane voltage, the departure from the Schwan’s equation model was observed. It was particularly significant for high level of cell membrane permeabilization during the pulse (for the average relative pore area over the whole membrane higher than 10
4). Computer simulation is useful for evaluation of cell electroporation in the presence of local heterogeneities of membranes related to their passive electrical properties (local resistance and capacitance) and both passive and active interaction mechanisms (ion pumps, channel gating, electroconformational coupling, electroporated membrane regions) (Gowrishankar and Weaver 2003). Figure 6 presents the transport two-dimensional lattice model of the individual spherical cell with two electrolytes separated by a membrane. The model accounted for the nonlinear dependence of the membrane dynamic conductance versus the transmembrane voltage. Transport lattices were solved by Kirchhoff’s laws, using Berkeley SPICE circuit simulator program for nonlinear DC, nonlinear transient, and linear AC analyses. The proposed transport model allowed simulation of cells with irregular geometry and the presence of local regions with different properties.

10

N. Lebovka and E. Vorobiev

Pulsed Electric Field Treatment of Suspension of Cells and Plant Tissues Evaluation of um value for the ensembles of cells (suspension of cells and plant tissues) is more complex and usually requires application of complex mathematical models and numerical calculations. In general case the calculations should account for inhomogeneous distribution in the local electric fields, osmotic flows, resealing of cells, distribution of cell sizes and orientations, their states and local solute concentrations, local electrical conductivity, etc. (Pucihar et al. 2007).

Corrections of Schwan’s Equation for Concentrated Cell Ensembles For concentrated cell ensembles (dense suspensions, plant tissues), the value of um may be altered owing to the deformation of the applied electric field by the adjacent cells. For lattice arrays of closely packed cells, the Schwan’s equation was generalized as (Pavlin et al. 2002; Qin et al. 2005; Ramos et al. 2006; Susil et al. 1998) um ¼ 1:5gðϕÞER cos θ;

(8)

where g(ϕ) is a correction factor, which is a function of volume fraction of particles, ϕ. The finite element method was used to calculate the value of um for different arrangements of spherical cells including a simple cubic lattice (sc), two-dimensional planes, and one-dimensional chains of cells both parallel and perpendicular directions to the applied field (Susil et al. 1998). Numerical simulations using the equivalent circuit method (ECM) were also applied to calculate the value of um for body-centered cubic (bcc) lattice arrangement of spherical cells (Ramos et al. 2006). Figure 7 presents a correction factor g versus normalized volume fraction of particles ϕ/ϕm for spherical cells in sc and bcc packings; here ϕm = π/6 and ϕm = √3 π/8 are the maximum volume fractions for sc and bcc packings, respectively. The value of g decreases practically linear with increasing of ϕm up to g  0.67–0.75 at ϕ = ϕm. The similar calculations were also extended for cell arrangements in bodycentered cubic (bcc) and face-centered cubic (fcc) lattices (Pavlin et al. 2002). Later on the computational results for fcc lattice model were used for explanation of the experimental measurements on the fraction of electroporated cells in dense suspensions (1 %, 18 %, and 36 % by volume) of CHO cells (Pucihar et al. 2007). The following expression for g(ϕ) was analytically evaluated using the meanfield approach (Qin et al. 2005): gð ϕÞ ¼ 

2 þ ð1
ϕÞ

3=2

3 

1þ



1=3 3 ϕ= 4Nπ



(9)

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . .

11

1 Electric filed, E

Correction factor, g

0.9 bcc

sc 0.8

0.7

0

0.2

0.4

0.6

0.8

1

Normalized volume fraction, f/f m Fig. 7 Correction factor g versus normalized volume fraction of particles ϕ/ϕm for spherical cells in sc packing. Here ϕm = π/6 and ϕm = √3 π/8 are the maximum volume fractions for sc and bcc packings, respectively (Compiled from the data presented in (Ramos et al. 2006; Susil et al. 1998))

Here, ϕ is the volume fraction of cells, and n is the number of particles per unit cell. For simple cubic lattice (ϕ = π/6 and n = 1) the above equation gives g = 0.859. Physical modeling of electroporation in close cell-to-cell proximity environments may be useful for estimation of deviations from Schwan’s equation (Gaynor and Bodger 2006). The electrically equivalent model, based on aqueous solution filled thin latex rubber membrane spheroids suspended in a tank filled with water (the so-called “balloon” model), was used to investigate membrane permeabilization. The various two-cell, “couplet” arrangements and a 3  3  3 simple cubic lattice of “cells” were studied. For the case when the external media conductivity is lower than the cell interior conductivity, the electroporation is initiated at substantially decreased value of E in regions of close cell-to-cell proximity as compared to cells in isolation. The experimental study of electroporation in close cell-to-cell proximity environments was performed using optical tweezers for accurate positioning of two cells in fluidic electroporation device (Henslee et al. 2014). The finite element simulations of the electrostatic potential distributions were also performed. The situations of two cells oriented parallel and perpendicular to the electric field were compared. Figure 8 presents the calculated correction factor g versus normalized center-to-center distance d/2R for two cells aligned parallel and perpendicular to the applied field. For

12

N. Lebovka and E. Vorobiev

Correction factor, g

1.04

Electric filed, E

1.02

1

2R

Electric filed, E 0.98

d 0.96

1

1.5

2

2.5

3

3.5

4

Normalized center to center distance, d/2R Fig. 8 Calculated correction factor g versus normalized center-to-center distance d/2R for two cells aligned parallel and perpendicular to applied field (Compiled from the data presented in (Henslee et al. 2014))

parallel geometry, the second cell decreases the field on the first where as the opposite situation is observed for perpendicular geometry. Good theory-experiment agreement was found for the perpendicular geometry. For the parallel geometry, experimentally the measured value g was about a factor of two larger than were obtained in electrostatic calculations. In general, the correction factor is changed ~5–10 % by a nearby, nearly touching second cell.

Kinetic Models of Inactivation in Suspension of Cells The simplest first-order equation for surviving kinetics was proposed by Bigelow (Huang et al. 2012): S ¼ expð
t=τc Þ

(10)

where S is a fraction of non-electroporated cells and τc is a lifetime of cells. Here, the lifetime of cells τc is the empirical parameter that can be obtained by fitting of experimental data on microbial inactivation using Eq. 10. However, in many cases the Bigelow’s equation fails to describe the microbial inactivation by PEF treatment. Different more complicated empirical models were

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . .

13

proposed to simulate the survival curves in microbial inactivation, e.g., Fermi, log-log, log-logistic, and other models. The very popular is the Hulsheger’s model: S ¼ ðt=tc Þ
ðE
Ec Þ=k

(11)

where tc and Ec are the threshold values of treatment time and electric field strength, and k is the parameter. However, this model is purely empirical and there is no theoretical justification for the origin of such behavior. Another model is described by extended exponential (Weibull’s) equation: S ¼ exp
½ðt=τÞn 

(12)

where τ and n are also empirical parameters. The shape parameter n accounts for a concavity of a survival curve. The deviations from first-order kinetics may be explained by different factors such as the distribution of lifetimes and of cell diameters and/or orientation of cells and many other factors. For the model of the distribution of lifetimes, the value of n characterizes the broadness of the distribution. The case n = 1 corresponds to the single lifetime, and the larger the deviation of n from 1 is, the broader the distribution of the lifetimes. The mean lifetime can be estimated from the following relation:   ¼ τΓ n
1 =n;

(13)

where Γ is the Euler gamma function. The Monte Carlo model was developed for estimation of the kinetics of microbial inactivation of cells by PEF (Lebovka and Vorobiev 2004). Initial Gaussian distribution of cell radii r was assumed:   pffiffiffiffiffiffiffiffi f ðRÞ ¼ exp
ðR
< R >Þ2 =2δ2 = 2πδ;

(14)

where and Δ are the mean radius and the standard deviation, respectively. The surviving kinetics was approximated by Eq. 12. The lifetime of cells was evaluated using Eq. 2 accounting for the cell radius dependence of transmembrane potential um(R), Eq. 4. Figure 9 presents some examples of the calculated survivor curves S(t) for a suspension of cells with a normal law distribution of their radiuses. The model predicts the decrease of shape parameter with increase of the standard deviation δ and gives upward concavity, i.e., n < 1. The similar Monte Carlo model was also developed for simulation of inactivation of spheroidal cells by PEF (Lebovka and Vorobiev 2007). The cases of partially and randomly oriented spheroids were analyzed; it was demonstrated non-exponential kinetics can reflect the orientational disorder of cells with anisotropic shape. Figure 10 presents examples of survivor curves S(t) for oblate (a < 1) and prolate (a > 1) spheroids distributed in the space with random orientations for relative

14

N. Lebovka and E. Vorobiev

Fig. 9 Survivor curves S(t) for different values of the relative width of cell radius distribution, Δ/. Here, is the mean radius of cells, τ1 m is the limiting value of τ (Eq. 1), and the calculations were made for relative electric field strength E* = 1.5ER/uo = 10 (Lebovka et al. 2002) (Compiled from the data (Lebovka and Vorobiev 2004))

 1=3 electric field strength E* = E/Eo = 10. Here, Eo ¼ uo =1:5R, R ¼ Rz R2 x is an equivolume radius of a spheroid and a = Rz/Rx is an aspect ratio (major/minor axis). The model predicts upward concavity, n < 1, for prolate cells, and near-exponential kinetics, for oblate spheroids. The most pronounced deviations from the exponential kinetics were observed for disordered suspensions of prolate spheroids at small electric field strength E and large aspect ratio a. For partially oriented suspensions, efficiency of inactivation enhances with increasing of order parameter and field strength.

Plant Tissues The effects of PEF on electroporation of the plant tissue were studied using the Monte Carlo computer simulation model (Lebovka et al. 2001). The tissue was represented by a two-dimensional array of biological cells (Fig. 11). The electrical potential u was applied between the upper and bottom electrodes, and the potential distribution of all cells in the lattice uxy was evaluated numerically by solving the discrete version of Laplace’s equation. For this purpose the successive relaxation scheme was used. The electroporation probability of all cells during the pulse

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . . 100

15

Prolate spheroids

E*=10

Surviving fraction, S

10−1 a=10 10−2

10−3

5

Oblate spheroids 0

10−4

0.2 0.1

10−5 0

50

100

150

Relative time of PEF treatment, t / t m•

200

Fig. 10 Survivor curves S(t) for oblate (a < 1) and prolate (a > 1) spheroids distributed in the space with random orientations. The dashed line corresponds to the simple exponential kinetics for the equivolume with spheroid spherical cell with radius R = (RzR2x)1/3. Here a = Rz/Rx is an aspect ratio (major/minor axis), τ1 m is the limiting value of τ (Eq. 1), and the calculations were made for E* = E/Eo = 10, Eo = uo/1.5R (Compiled from the data (Lebovka and Vorobiev 2007))

application was estimated using the first-order kinetic equation (Eq. 2). The electroporation of cells accompanies with replacing the electrical conductivities σ i ! σ d and updating of the potential uxy in all cells. The effective electrical conductivity of the tissue was calculated using a highly efficient Frank and Lobb algorithm. The degree of disintegration Z was defined as the ratio of electroporated cells and total number of the cell. The application of an external field in the form of an idealized square pulse sequence with a pulse duration ti and a pulse repetition time Δt was assumed. The simulation model accounted also for the possibility of resealing (reversible electroporation) with time constant of τr and moisture transfer processes (diffusion, osmotic flow, and redistribution of moisture inside the sample) with time constant of τd. Figure 12 shows the example of the simulation of PEF-induced electroporation tissue in the presence of resealing and moisture transfer processes (Lebovka et al. 2001). The obtained numerical data were in qualitative correspondences with experimental data for PEF-treated apples. It was demonstrated that PEF-induced damage kinetics is very sensitive to the repetition time Δt. The aggregated groups of the electroporated cell (correlated patterns) were observed for the case when Δt exceeds the time of moisture transfer processes τd. On the contrary, when Δt  τd, the spatial distribution of the

16

N. Lebovka and E. Vorobiev Electrode

Intact cell

si dm PEF

Electropotated cell R

sd

Electrode

Fig. 11 The two-dimensional model of the tissue. Each cell is represented by a node with four conducting bonds. Here, σ d and σ i are the conductivities of intact and electroporated cells, respectively, R is the size of the cell (“radius”), and dm is the thickness of the membrane (Based on the model described in (Lebovka et al. 2001))

electroporated cell was observed. In PEF experiments with long repetition time, Δt = 60 s, the large macroscopic domains of electroporated cells were observed. Considerable differences between the damage kinetics for long and short interpulse distances were observed both in experiments and simulations. The experimental data evidences that PEF protocol with Δt = 60 s was more efficient than PEF protocol with Δt = 0.01 s. Electroporation theory predicts higher electroporation for the tissue with larger radius cells (Eqs. 4 and 8). However, in experiments with selected fruit and vegetable plant tissues (apple, potato, carrot, courgette, orange, and banana), the direct correlation between PEF-induced electroporation and size of the cell was not always observed (Ben Ammar et al. 2011). To explain this anomaly, a Monte Carlo model for simulation of PEF-induced changes in the tissue was developed. Initial Gaussian distribution of cell radii r was assumed (Eq. 14). Before PEF treatment all cells were intact. During the PEF treatment (tPEF > 0), some of the cells remained intact, while others got damaged (electroporated). So, electroporation influenced the distributions of intact and electroporated cells and they were time dependent. The degree of plant tissue damage, Z, was defined as the ratio of volume of the damaged cells and the total volume of cells. In simulation the intact cell was selected randomly. The damage time evolution under the PEF treatment was approximated by the firstorder equation with accounting for the lifetime dependence upon electric field strength E, current damage degree Z, and cell radius r (Eqs. 4, 5, and 7): h  i d τc ¼ τ1 c exp ðW c =kT Þ= 1 þ ð1:5f e ðZ ÞRE=uo Þ

(15)

Relative conductivity,s/si

Disintegration index, Z

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . . 1 0.8

17

tr =10−2 s

0.6 0.4

td (s)

Pulse sequence

1

Dt = 5ti

0.2

10−2

ti =1ms

3 Electroporation

1

Resealing

4

2 Pulse number, n

Fig. 12 The example of the simulation of PEF-induced electroporation tissue in the presence of resealing with time constant of τr and moisture transfer processes with time constant of τd. Degree of disintegration, Z, and relative conductivity, σ/σ i, versus pulse number, n (Lebovka et al. 2001) 1 with τ1 c  2:01τm , d  1.33, and Wc  0.54 Wm (see Eq. 1) (Lebovka et al. 2002). Note that the electroporation factor fe(Z) in Eq. 15 accounts for the changes of electrical conductivity of the extracellular medium σ(Z ) during PEF treatment (see Eq. 5). For estimation purposes, the value of σ d was identified with electrical conductivity of the totally electroporated tissue with highest level of Z (Z  1). Such tissue can be obtained using the PEF treatment in the limit of long treatment time (tPEF  0.1–1 s) and high electric field strength (E > 1000 V/cm). The value of σ increases with increase of Z from σ (Z = 0) = σ i to σ (Z = 1)  σ d in the limit of high tissue damage. The value of σ m was estimated from the serial one-dimensional model of cell packing as σ m  σ i hm/R, where hm (5 nm) is the thickness of the membrane. Figure 13 presents examples of calculated dependencies of electroporation factor fe versus σ/σ d at different values of electrical conductivity contrast, i.e., the ratio between final σ d and initial σ i conductivities, k = σ d/σ i (>1). The value of fe is an increasing function of σ and it attains maximum fmax for the totally electroporated tissue (at σ/σ d = 1 and Z = 1). fmax is an increasing function of k and fmax  1 at k  1. However, for tissues with small values of electrical conductivity contrast the transmembrane potential um, the efficiency of electroporation may be noticeably reduced. Finally, for the current damage degree f (Z ), dependence can be approximated using the general effective medium equation for electrical conductivity:

18

N. Lebovka and E. Vorobiev

Fig. 13 Electroporation factor, fe, versus the electrical conductivity ratio, σ/σ d, at different values of electrical conductivity contrast, k = σ d/σ i (Compiled from the data (Ben Ammar et al. 2011))

        1=s 1=s 1=t 1=t ð1
Z Þ σ i
σ 1=s = σ i þ Aσ 1=s þ Z σ d
σ 1=t = σ d þ Aσ 1=t ¼ 0;

(16)

where s and t are the exponents of the percolation theory, s = 0.73, t = 2.0 for threedimensional random materials, and A = (1
Zc)/Zc, Zc is the percolation threshold corresponding to the transition from low (σ = σ i) to high (σ = σ d) conductivity limits. In experimental works to characterize the extent of electroporation, the conductivity disintegration index is commonly used (Barba et al. 2015): Zc ¼ ðσ
σ i Þ=ðσ d
σ i Þ;

(17)

where the subscripts i and d refer to the conductivities of the untreated (intact) and completely damaged (electroporated) tissue, respectively. Figure 14 shows simulated specific power consumptions W (Zc = 0.8) versus electric field strength E at two values of electrical conductivity contrast, k = σ d/σ i (=10 and = 2). The parameters of simulations are specified in Ben Ammar et al. (2011). The increasing of E accelerates tissue damage in full correspondence with reported experimental data. Moreover, the model allowed explanation of experimentally observed correlation between electroporation efficiency and value

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . .

19

Fig. 14 Specific power consumptions, W, required for relatively high disintegration (Z = 0.8) versus electric field strength, E, at two different values of conductivity ratio σ d/σ i (Compiled from the data (Ben Ammar et al. 2011))

of electrical conductivity contrast. The model predicts also the sensitivity of the evolution of distribution function of intact cells to the values of E and σ d/σ i. The PEF treatment at relatively small fields allowed of preferential damage of larger cells and it can be useful for selective extraction of the intercellular medium from larger cells. The value of W passed through the minimum with increasing of E, and the position of this minimum at E = Eo was dependent upon the value of k = σ d/σ i. The simulated behavior is in good correspondence with existing experimental data (Lebovka et al. 2002). For example, it was estimated Eo  360 V.cm
1 (σ d/σ i  14.3) for potato and Eo  976 V.cm
1 (σ d/σ i  5.6) for banana, i.e., the value of Eo increases with decreasing of σ d/σ i (Ben Ammar et al. 2011). The plant tissue with low σ d/σ i ratio (σ d/σ i  1) requires application of a rather strong electric field (e.g., numerical estimation gives Eo  3000 V/cm for orange, k = σ d/σ i  1.25). The Monte Carlo model for simulation of electroporation efficiency induced by PEF treatment of plant tissues was developed (Lebovka et al. 2014). The PEF treatments with time-variable protocol are applied in combined PEF pressing (hydraulic, roll or belt), when the electric field strength E and electrical conductivity of the sample can change significantly during the operations. The potato was selected as a model tissue. The simulations were done using the constant value of electric field strength E = Ei (model M!) and time-variable protocols with different strategies of E changes:

20

N. Lebovka and E. Vorobiev

exponential increase (model M") or decrease (model M#) of E, persistent (model M""), and anti-persistent (model M"#) evolution of E. In persistent and antipersistent strategies, the direction of the E changes was dependent upon efficiency of tissue electroporation that was characterized by the value of dZ/dW and the sign of the second derivative d2Z/dW2. Note that application of persistent (M"") or antipersistent M"# strategies does not require assignment of the final value of Ef that makes these strategies more attractive for autoregulation of electric field strength E during the PEF treatment. Figure 15 gives examples of simulated specific power consumption, W, (at disintegration index Z = 0.9) versus initial electric field strength, Ei, for different exponential evolution of E. For model M" the electric field strength was increased between Ei and Ef = 2Ei according to E ¼ Ei ð2
expð
t=tPEF ÞÞ

(18a)

For model M# the electric field strength was decreased between Ei and Ef = 0.5Ei according to E ¼ 0:5Ei ð1 þ expð
t=tPEF ÞÞ

(18b)

The data for the constant E model (M!, E = Ei = const) are also presented in Fig. 16 for comparison (Lebovka et al. 2014). The computer simulation predicted that application of protocols with exponential time-variable electric fields would allow optimization of PEF treatment with initial electric field Ei deviating from the optimal value Eo. In general, the W(Ei) dependencies at nonstationary exponential strategies M" and M# were qualitatively similar to that of PEF treatment under stationary conditions, E = Ei = const. The value of Eo was dependent on parameters of the models (i.e., Ef and τE). For PEF treatment at the constant electric field E (model M!), the optimal value was Eo  335 V/cm and the value of Eo increased as τE increased for the model M"(Ef = 2Ei), and the opposite behavior was observed for the model M# (Ef = Ei/2). So, application of nonstationary protocols with exponential increase or decrease of electric field strength (models M" and M#) is more attractive in the cases when Ei < Eo or Ei > Eo, respectively. Simulations for the persistent M"" and anti-persistent M"# strategies revealed significant differences in their electroporation efficiency (Lebovka et al. 2014). For easier adaptation of PEF treatment time-variable protocol, the persistent strategy M"" was more useful than anti-persistent strategy M"#. At low initial value Ei that was smaller than the optimal Eo  335 V/cm, the persistent strategy M"" predicted noticeable benefits in power consumption. For example, the value of W (Z = 0.9) at Ei = 200 V/cm was 3.2 kJ/kg for M"" model and 5.0 kJ/kg for M! model at τE = 3–10τ1. From the other side, the persistent strategy M"" resulted in an increase of power consumption compared to that obtained for M! model at the initial value of Ei, larger than the optimal Eo value. The obtained data have also

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . .

21

Specific power consumption, W, KJ/Kg

5 M↑: E = (Ef -E i )(1-exp(-tPEF /tE)+Ei M↓: E = (Ei -E f)(exp(-tPEF /tE )+Ef

4

3

M↑ Ef = 2Ei

M↓ Ef = Ei /2

Eo

M→ Ei = const 2 100

200

300

400

Z=0.9 500

600

700

100

Fig. 15 Simulated specific power consumption, W, (at disintegration index Z = 0.9) versus initial electric field strength, Ei, for exponential models of E evolution (increase of E, M", and Ef = 2Ei and decrease of E, M#, and Ef = Ei/2), and τE ¼ 3τ1 c , where τ E is the constant characterizing the time dependence of E, and τ1 c is the limiting lifetime of a cell. The data for the constant E model (M!, E = Ei = const) are also presented for comparison (Compiled from the data (Lebovka et al. 2014))

shown that application of anti-persistent M"# strategies did not result in improvement of energy consumption at small values of Ei ( Eo. Finally, the simulation results were in reasonable correspondence with experimental data (Lebovka et al. 2014). The PEF experiments with potato disks at constant values of E gave W(E) dependence that was in good correspondence with theory predictions for the model M!, Ei = const. The minimum power consumption was observed at intermediate electric field strength E = Eo  400 V/cm. PEF experiments with exponential increase (M") and decrease (M#) of E values revealed that M" strategy can be useful at small initial electric field strengths, E < Eo, as far as it allowed significant improvement of PEF treatment efficiency. From the other side, the exponential decrease strategy M" with large initial electric field strength Ei (>Eo) was less efficient than the strategy with constant E at the optimal value of E = Eo  400 V/cm. The computer simulation and mathematical modeling were used for prediction of PEF effects on osmotic dehydration apple slices (Yildiz et al. 2016). The data on the osmotic dehydration times and the final water and solid distributions in apple slices were in good agreement with experimental data.

22

N. Lebovka and E. Vorobiev

It should be mentioned that many attempts were devoted to the studies of electroporation mechanisms in biological tissues in relation with biomedical applications of PEF treatment. In several works the potato tissue was used as a model material (Castellví et al. 2016; Kranjc et al. 2016). Mathematical simulation was used to predict shape and extension of irreversible electroporation (IRE) (Castellví et al. 2016). The sliced potato tubers were PEF treated and a dye solution was used to enhance the IRE area. From digitized images the electroporated area and the resulting treated volume were automatically identified. A three-dimensional reconstruction of the electroporated tissue was generated. The model for reconstruction the electric field distribution during PEF treatment of potato tubers using magnetic resonance electrical impedance tomography (MREIT) was also developed (Kranjc et al. 2016). Data shows that MREIT could be used as an efficient tool for improving the effectiveness of PEF treatment. Simulations have been performed in order to determine the distribution of the electric field and compute the spatial and temporal evolution of electroporation in a biological tissue during the application of pulse (Langus et al. 2016). An electric circuit model for the tissue and a phenomenological model for the electroporation process were used. Data of simulations were in reasonable correspondence with in vivo experiments made on rabbit livers. Numerical simulations based on a finite element model were also applied to study the effects of tissue properties, electrode placement, and electric pulse delivery scenarios in the process of electroporation. The developed model was able to predict the time evolution of an electric pulse current for different PEF protocols.

Conclusions Applications of PEF to processing of plant tissues have many important and promising perspectives in the food industry. The models of PEF treatment and simulation of related phenomena can contribute significantly to the development of efficient PEF protocols and equipments. Nowadays the various approaches on different spatial and temporal scales have been proposed for such aims. However, the models and simulations are still rarely explored in studies of the PEF treatment combined with different food processing techniques such as solid-liquid and pressing extraction, drying, freezing, and osmotic dehydration. These studies are required in the future for facilitation and optimization of the PEF treatment protocol, finding possible synergy of different mode of treatments, improvement of the power consumptions and preservation of food quality. Acknowledgments The authors appreciate the support from the COST Action TD1104 (EP4Bio2Med – European network for development of electroporation-based technologies and treatments; www.electroporation.net).

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . .

23

Cross-References ▶ 3D Tissue Models to Bridge the Gap Between Cell Culture and Tissue in Assessing Electroporation ▶ Application of Numerical Simulation Techniques for Modeling Pulsed Electric Field Processing ▶ Biophysics and Metrology of Electroporation in Tissues ▶ Cell/Tissues Electroporation Modeled using Voronoi Network ▶ Computational Approach for Electrical Analysis of Biological Tissue Using the Equivalent Circuit Model ▶ Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Molecular Models ▶ Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue Modified by Electroporation ▶ Electric Field Distribution Modelling in Tissue Considering Tissue Conductivity Increase Due to Electroporation ▶ Electropore Energy and Thermodynamics ▶ Electroporation and Electropermeabilization ▶ Lattice Model of Electroporation ▶ Lipid Electropore Lifetime in Molecular Models ▶ Lipid Pores: Molecular and Continuum Models ▶ Mathematical Models Describing Cell Death Due to Electroporation in Vitro ▶ Membrane Permeabilization Lifetime in Experiments ▶ Molecular Models of Lipid Bilayers and Electropore Formation ▶ Modeling Microbial Inactivation by Pulsed Electric Field ▶ Pore Lifetime and Permeabilization Lifetime in Models ▶ Pulsed Electric Fields Treatment of Biological Suspensions ▶ Transmembrane Voltage Induced by Applied Electric Fields

References Barba FJ, Parniakov O, Pereira SA, Wiktor A, Grimi N, Boussetta N, Saraiva J, Raso J, MartinBelloso O, Witrowa-Rajchert D, Bals O, Vorobiev E, Lebovka N (2015) Current applications and new opportunities for the use of pulsed electric fields in food science and industry. Food Eng Rev 77:773–798 Ben Ammar J, Lanoisellé J-L, Lebovka NI, Van Hecke E, Vorobiev E (2011) Impact of a pulsed electric field on damage of plant tissues: effects of cell size and tissue electrical conductivity. J Food Sci 76:E90–E97 Castellví Q, Banús J, Ivorra A (2016) 3d assessment of irreversible electroporation treatments in vegetal models. In: 1st World Congress on electroporation and pulsed electric fields in biology, medicine and food & environmental technologies. pp. 294–297 Gaynor PT, Bodger PS (2006) Physical modelling of electroporation in close cell-to-cell proximity environments. Phys Med Biol 51:3175 Gerlach D, Alleborn N, Baars A, Delgado A, Moritz J, Knorr D (2008) Numerical simulations of pulsed electric fields for food preservation: a review. Innov Food Sci Emerg Technol 9:408–417

24

N. Lebovka and E. Vorobiev

Gowrishankar TR, Weaver JC (2003) An approach to electrical modeling of single and multiple cells. Proc Natl Acad Sci 100:3203–3208 Henslee BE, Morss A, Hu X, Lafyatis GP, Lee LJ (2014) Cell-cell proximity effects in multi-cell electroporation. Biomicrofluidics 8:52002 Huang K, Tian H, Gai L, Wang J (2012) A review of kinetic models for inactivating microorganisms and enzymes by pulsed electric field processing. J Food Eng 111:191–207 Kotnik T, Miklavčič D (2000) Analytical description of transmembrane voltage induced by electric fields on spheroidal cells. Biophys J 79:670–679 Kotnik T, Miklavčič D, Slivnik T (1998) Time course of transmembrane voltage induced by timevarying electric fields – a method for theoretical analysis and its application. Bioelectrochem Bioenerg 45:3–16 Kranjc M, Bajd F, Serša I, De Boevere M, Miklavčič D (2016) Electric field distribution in relation to cell membrane electroporation in potato tuber tissue studied by magnetic resonance techniques. Innov Food Sci Emerg Technol doi: http://dx.doi.org/10.1016/j.ifset.2016.03.002 Langus J, Kranjc M, Kos B, Šuštar T, Miklavčič D (2016) Dynamic finite-element model for efficient modelling of electric currents in electroporated tissue. Sci Rep 6:26409 Lebedeva NE (1987) Electric breakdown of bilayer lipid membranes at short times of voltage action. Biol Membr (Biochem Mosc A: Membr Cell Biol Russ) 4:994–998 Lebovka NI, Vorobiev EI (2004) On the origin of the deviation from the first-order kinetics in inactivation of microbial cells by pulsed electric fields. Int J Food Microbiol 91:83–89 Lebovka N, Vorobiev E (2007) The kinetics of inactivation of spheroidal microbial cells by pulsed electric fields http://arxiv.org/abs/0704.2750 Lebovka NI, Bazhal MI, Vorobiev E (2001) Pulsed electric field breakage of cellular tissues: visualisation of percolative properties. Innov Food Sci Emerg Technol 2:113–125 Lebovka NI, Bazhal MI, Vorobiev E (2002) Estimation of characteristic damage time of food materials in pulsed-electric fields. J Food Eng 54:337–346 Lebovka NI, Mhemdi H, Grimi N, Bals O, Vorobiev E (2014) Treatment of potato tissue by pulsed electric fields with time-variable strength: theoretical and experimental analysis. J Food Eng 137:23–31 Melikov KC, Frolov VA, Shcherbakov A, Samsonov AV, Chizmadzhev YA, Chernomordik LV (2001) Voltage-induced nonconductive pre-pores and metastable single pores in unmodified planar lipid bilayer. Biophys J 80:1829–1836 Mercadal B, Vernier PT, Ivorra A (2016) Dependence of electroporation detection threshold on cell radius: an explanation to observations non compatible with Schwan’s equation model. J Membr Biol 1–14, doi: 10.1007/s00232-016-9907-0 Miklavčič D, Mali B, Kos B, Heller R, Serša G (2014) Electrochemotherapy: from the drawing board into medical practice. Biomed Eng Online 13(1) Pavlin M, Pavšelj N, Miklavčič D (2002) Dependence of induced transmembrane potential on cell density, arrangement, and cell position inside a cell system. Biomed Eng IEEE Trans 49:605–612 Pucihar G, Kotnik T, Teissie J, Miklavcic D (2007) Electropermeabilization of dense cell suspensions. Eur Biophys J 36:173–185 Qin Y, Lai S, Jiang Y, Yang T, Wang J (2005) Transmembrane voltage induced on a cell membrane in suspensions exposed to an alternating field: a theoretical analysis. Bioelectrochemistry 67:57–65 Ramos A, Suzuki DOH, Marques JLB (2006) Numerical study of the electrical conductivity and polarization in a suspension of spherical cells. Bioelectrochemistry 68:213–217 Rems L, Miklavčič D (2016) Tutorial: electroporation of cells in complex materials and tissue. J Appl Phys 119:201101 Susil R, Šemrov D, Miklavčič D (1998) Electric field-induced transmembrane potential depends on cell density and organization. Electro- Magnetobiol 17:391–399 Vasilkoski Z, Esser AT, Gowrishankar TR, Weaver JC (2006) Membrane electroporation: the absolute rate equation and nanosecond time scale pore creation. Phys Rev E 74:21904

Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and. . .

25

Weaver JC, Chizmadzhev YA (1996) Theory of electroporation: a review. Bioelectrochem Bioenerg 41:135–160 Yildiz H, Icier F, Eroglu S, Dagci G (2016) Effects of electrical pretreatment conditions on osmotic dehydration of apple slices: experimental investigation and simulation. Innov Food Sci Emerg Technol 35:149–159 Zudans I, Agarwal A, Orwar O, Weber SG (2007) Numerical calculations of single-cell electroporation with an electrolyte-filled capillary. Biophys J 92:3696–3705

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue Modified by Electroporation Samo Mahnič-Kalamiza

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Dual-Porosity Modeling Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Dual-Porosity Approach in Theory: Mathematical Problem Formulation . . . . . . . . . . . . . . The Hydraulic Permeability of the Electroporated Biological Membrane and Estimation of Model Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Analytical Solution of the Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simulated Liquid Pressure Distributions and Consolidation Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . Model Verification and Application by Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 4 5 13 17 18 21 22 24 24

Abstract

One of the basic challenges in applying electroporation to tissues is the understanding of resulting structural property changes and mass transport properties in the treated material. In applying electroporation to plant tissue, the objective is in either achieving improved extraction of intracellular compounds or water out of cells, or facilitating otherwise impractical, impossible, or severely hindered introduction of molecules into cells. Understanding of and the ability to model solute and liquid transport phenomena in treated tissues is therefore of great importance, both in scientific terms as well as for practical purposes that are of interest to the food processing industry or biorefinery. Electrical and process parameters that are characteristic of a particular treatment protocol used to achieve the required mass transport enhancement in tissue should be connected, S. Mahnič-Kalamiza (*) Laboratory of Biocybernetics, Faculty of Electrical Engineering, University of Ljubljana, Ljubljana, Slovenia e-mail: [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_11-1

1

2

S. Mahnič-Kalamiza

theoretically, with the resulting improvement in mass transport and structural changes in tissue. This theoretical link should be established in order to allow for optimization of the treatment protocols, thus saving energy, materials (e.g., extraction reagents or solvents), reducing time for post-treatment processing (e.g., diffusion or maceration stage), and improving the quality and safety of the final product. This chapter presents a modeling approach to understanding mass transport processes in electroporated tissues, whereby the mass transport process is mathematically linked to the structural property changes in tissue resulting from electroporation. The approach has been termed the dual-porosity model and should be considered as an illustration of the fundamental basis to a more comprehensive model connecting effects of electroporation on the biological membrane with its observable effects on the macroscopic level of tissue. To allow the reader immediate application and full comprehension of the model, the complete mathematical derivation and an analytical solution of the fundamental model equations are presented, supplemented by an extensive commentary on the theoretical basis and derivation of individual parameters of the model. Keywords

Mathematical modeling • Tissue electroporation • Plant tissues • Pressure gradients • Liquid flow • Consolidation behavior • Bulk properties

Introduction Models of phenomena in electroporated tissues are scarce in literature. Most concern the electric field or conductivity distribution and changes in tissue following electroporation (Sel et al. 2005; Granot and Rubinsky 2008; Grimi et al. 2010; Čorović et al. 2012; Dymek et al. 2015; Langus et al. 2016) and do not extend the modeled properties to consequent mass transport but are mainly concerned with the objective of ensuring the minimum required coverage of tissue with an electric field of sufficient strength and the (in)homogeneity of the electric treatment applied. Existing mass transport models relating electroporation and tissue are almost exclusively phenomenological in nature, meaning they are developed by fitting a suitably chosen mathematical model according to a best-fit criterion to experimental data, obtained generally by means of a multiparameter experimental study (Lebovka et al. 2000; Fincan and Dejmek 2003; Grimi et al. 2010). Such models may be of great practical aid; however, their construction necessitates a multitude of experiments be performed in numerous repetitions for every varied parameter value from the chosen parameter range. The downside, apart from the relative tediousness and time-consuming effort invested in the endeavor, is the limited applicability of such models in case treatment parameters (electroporation protocol) and/or target tissue properties (e.g., humidity, age, structural and textural properties, etc.) change for a

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

3

variety of reasons. Since tissue properties can depend strongly on the origin of tissue and growth conditions, on what kind of tissue (e.g., roots, leafs, fruit, etc.) is being treated, and can vary considerably between species of plant or indeed between cultivars of the same crop species, this can present a considerable challenge to the usefulness and applicability of such purely phenomenological models. An alternative approach is to build a model from the ground up; a model can be constructed by observing the basic physical laws of the processes involved, while also introducing a minimal sufficient set of parameters that reflect material properties and model variables of the treatment that is applied, a process known as mechanistic modeling (Zhu and Melrose 2003). When modeling effects of electric fields on plant tissues and the resulting mass transport, this rule translates to incorporating, at a minimum, a mass transport and conservation law or set of laws, a parameter reflecting the structural changes induced by the treatment in the target tissue, and a parametrization of the electrical treatment where the link between each of the treatment parameters and the parameter reflecting structural changes must be established either theoretically or empirically. The ultimate goal in trying to understand the effects of treatment on tissue is the establishment of a theoretical relationship between any given treatment protocol parameter and the resulting changes in tissue, a process which remains to date an ongoing effort in the scientific community concerned with basic principles of electroporation. The modeling paradigm presented in this chapter is an attempt at formulating such a relationship by suggesting a theoretical link between one of the most critically important electroporation parameters – the local electric field strength – on the one hand, and the hydraulic resistance of cell membranes to liquid flow in tissue, on the other. This formulation is based on previous works in the biomedical field of electroporation applications and studies done on cell suspensions of animal cells (Smith et al. 2004; Pucihar et al. 2007; Pavlin and Miklavcic 2008; Kotnik et al. 2010; Li and Lin 2011), and the mass transport (liquid flow) equations are grounded in the theory of water relations in soils and fractured rocks, i.e., mechanics of porous materials (Barenblatt et al. 1990). A similar approach has already been developed and proposed for the case of mechanical pressing of oleaginous seeds (Lanoiselle et al. 1996) and beds formed of tissue particles (Petryk and Vorobiev 2013). A completely analogous formulation, from the mathematical point of view, of a diffusion problem can be found in Mahnič-Kalamiza et al. (2014a). The dual-porosity model of liquid expression from tissue that has been electroporated as presented in this chapter can thus be thought of as a demonstration of an approach in building a model based on fundamental principles of electroporation and its known effects on cells in tissue and mass transport laws in porous materials, such as biological (plant) tissues. To keep the account instructive and comprehensible, the model retains a number of parameters that are fundamentally phenomenological in nature; however, their initial approximate values can easily be determined from simple experiments.

4

S. Mahnič-Kalamiza

The Dual-Porosity Modeling Approach In order to study mass transport at the tissue level with relation to electroporation or indeed any treatment, the structural components of tissue that are hindering mass transport and are affected by the treatment must be identified and introduced into a model. This can be done by formulating, as a mathematical relation, the dependence of a property or properties (e.g., permeability, conductivity, etc.) of the identified structure on the relevant (important) treatment parameters. As an example, consider the case of the semipermeable biological membrane. Under normal physiological conditions, the membrane is selectively permeable, meaning that it is not indiscriminately admitting to transmembrane mass transport. With electroporation, the selective permeability of the plasma membrane is either temporarily (case of reversible electroporation) or permanently (in case of irreversible electroporation) disrupted. Any model attempting to relate electroporation effects with the resulting transmembrane, mass transport must therefore introduce the variable permeability of the membrane into the equation(s). Since tissue comprises a multitude of cells of different shapes and sizes arranged in a particular structure, where the spaces between the cells are far from empty, one must account for the specifics of the extracellular matrix as well. This extracellular network also presents a hindrance of its own to mass transport in tissue, and the permeability to solute diffusion or liquid flow in this space will be – if not directly, at least indirectly – a function of the treatment intensity. It is evident that in order to study mass transport phenomena in tissue, one must, at a minimum, consider the permeability of the plasma membrane as well as the permeability of the extracellular space. The dual-porosity approach as described in this chapter follows this logic closely, with the noted exception of neglecting the variable permeability of the extracellular matrix that can be ascribed to electroporation. The reduction of extracellular permeability due to tissue compaction, and therefore its temporal variability, is also not accounted for. The modeled system is schematically illustrated by Fig. 1, giving a highly simplified representation of tissue with the two compartments or spaces – the intracellular and the extracellular. According to the dual-porosity approach, the intracellular and extracellular spaces are each attributed an intrinsic porosity, i.e., a void-to-total (liquid-to-total) volume ratio. The void volume is considered to be occupied (filled) with liquid that can be expressed out of the cells and out of tissue under the driving force of an externally applied pressure, and the remaining volume is considered to comprise biological solids. The model does not concern extracellular air, which is supposed to be expelled out of tissue at the initial stages of the modeled experiment, resulting in either a slight compaction of the tissue or replacement of air pockets with intracellular liquid at the onset of electroporation. The transmembrane flow of liquid is regulated by the membrane, of which hydraulic permeability is a function of electroporation, and the liquid expressed from cells vacates the tissue via the extracellular space. The cell membrane and the extracellular space have different hydrodynamic properties. The assumption that liquid does not reenter cells and does not flow through bulk tissue via the

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

5

Fig. 1 A simplified schematic representation of tissue, introducing the minimal structure necessary for demonstrating the dual-porosity approach. The subfigures correspond to the case of (a) before electroporation or (b) after electroporation. Coloring is for illustrative purposes only and depicts no factual property of the biological material

intracellular pathway is based on the assumption that the membrane is the predominant factor in hindering liquid expression from tissue. If so, the intracellular pressure will always be higher than the extracellular, meaning transmembrane flow is unidirectional. What remains in order to arrive at a model formulation is to write mass conservation laws for the two respective spaces.

The Dual-Porosity Approach in Theory: Mathematical Problem Formulation Constitutive Equations In order to introduce the dual-porosity model in the form of its mathematical formulation, one can consider a particular example of a pressing experiment, set up thus that the mathematical formulation be kept as simple and as instructive as possible. For this purpose, consider the model system schematically depicted in Fig. 2. A sample of plant tissue is supported by a porous support allowing the passage of liquid only, not that of solids, and pressure is applied to the tissue at the top surface in the negative z-axis direction according to this particular choice of a coordinate system. In this model system, one is interested in determining, theoretically, the pressure relations for all times t during pressure application and throughout the tissue sample, i.e., for all z. The problem has been limited to one spatial coordinate along the axis of pressure application based on the assumption that due to sufficiently large surface area perpendicular to the z axis of the tissue sample, the effects of liquid displacement in the x and y directions can be neglected. In general, however, and in cases of strong anisotropy of the material (tissue), it should be noted and considered that the problem is three-dimensional.

6

S. Mahnič-Kalamiza

Fig. 2 A simplified schematic representation of plant (sugar beet tap root) tissue exposed to pressure applied at its top surface while resting on a porous support at its bottom surface; a model system representing liquid expression from tissue and consequent tissue consolidation. The magnifying glass is in aid of illustrating the microscopic structure of the material under pressure

Fig. 3 A simplified schematic representation of tissue highlighting the extracellular space in tissue following electroporation. The dashed line used to represent the membrane of the cell in the middle is in aid of representing the permeabilized state of this cell membrane after electroporation

Based on the established filtration-consolidation theory of bio-solids (Lanoiselle et al. 1996; Petryk and Vorobiev 2007, 2013), the first constitutive equation of the model can be written by observing the law of mass conservation. In the extracellular space (depicted in blue in Fig. 3), the rate of change in porosity (liquid-to-totalvolume ratio) is proportional to the rate at which the liquid is vacating the extracellular space (the extracellular liquid flow velocity) under pressure (outflow), as well as to the rate at which intracellular liquid is flowing through the porous cell membrane into the extracellular space (inflow). The relation just narrated by the

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

7

Fig. 4 A simplified schematic representation of tissue highlighting the intracellular space in tissue following electroporation. As in Fig. 3, the dashed line used to represent the membrane of the cell in the middle is in aid of representing the permeabilized state of this cell membrane after electroporation

preceding sentence can be written more succinctly in the form of a mathematical equation, yielding @ ðρee Þ @ ðρqe Þ ¼ þ ρυie @t @z

(1)

In Eq. 1, which is a first-order partial differential equation, ρ is the liquid medium density; ee the extracellular space porosity; qe is the extracellular liquid flow velocity; and υi-e is the source term that represents the flow of liquid through the cell membrane from the intracellular into the extracellular space. Considering pressure relations in the intracellular space (depicted in red by Fig. 4), one arrives to the second constitutive equation of the dual-porosity model. Again, by observing the mass conservation law, this yields @ ðρei Þ @ ðρqi Þ ¼  ρυie @t @z

(2)

In the above Eq. 2, ρ is, as before, the liquid medium density; ei the intracellular porosity; qi the liquid flow velocity in the intracellular space; and υi-e is the same source term as in Eq. 1 that represents the flow of liquid through the cell membrane from the intracellular into the extracellular space. Note, however, the change of sign from positive to negative in the second term on the right-hand side of Eq. 2 as compared to Eq. 1. The transmembrane flow is depleting the intracellular space of liquid, therefore its negative sign (it is an outflow from the intracellular space perspective).

8

S. Mahnič-Kalamiza

According to estimates that can be found in literature, the permeability of extracellular space is several orders of magnitude greater than that of intact cellular membrane (Tomos 1988; Buttersack and Basler 1991). This should still be a valid assumption under treatment conditions resulting in tissue damage (by electroporation) below the threshold that results in, predominantly, irreversible permeabilization of the cell membranes. If treatment conditions do favor the importance (in the sense of a limiting factor) of the membrane permeability over that of the extracellular space, then the bulk liquid pathway will be primarily from within the cells into the extracellular space and by filtration through the extracellular space out of the tissue block. Assuming constant juice density ρ, and the supposition that finite rate of flow inside the cells can be neglected, Eqs. 1 and 2 simplify to @ee ðz, tÞ @q ðz, tÞ ¼ e þ υie ðz, tÞ @t @z

(3)

@ei ðz, tÞ ¼ υie ðz, tÞ @t

(4)

The above set of equations needs to be reformulated to express the quantities in terms of variables common to both equations with known initial and boundary conditions. These two unknowns are the liquid pressures pi and pe, where pressure pi is the intracellular liquid pressure and pe the extracellular liquid pressure.

On Porosity, Void Ratio, and the Compressibility Modulus Porosities ee and ei featured in Eqs. 3 and 4 are related to solid pressures pe,S = PE – pe and pi,S = PE – pi via the compressibility moduli Ge and Gi and the void ratios ee and ei, where ee = ee/(1 + ee) and ei = ei/(1 + ei). The relationships between void ratios and solid pressures given by the following equations are taken from literature where the multiple-porosity modeling principle has been applied to the problem of pressing of oleaginous seeds (Lanoiselle et al. 1996). These relationships are given as @ee @ee @pe, S @ee @pe 1 @pe ¼ ¼ ¼   Ge @t @t @pe, S @t @pe, S @t

(5)

@ei @ei @pi, S @ei @pi 1 @pi ¼ ¼ ¼   Gi @t @t @pi, S @t @pi, S @t

(6)

In Eqs. 5 and 6, pe,S and pi,S are the pressures of total insoluble solids in the extracellular and intracellular space, respectively. In case of constant-pressure expression, i.e., if @PE/@t = 0, the solid pressures increase in time proportionally to the decrease in respective liquid pressures, i.e., @pe,S/@t = @pe/@t and @pi,S/ @t = @pi/@t. From Eqs. 5 and 6, it is obvious that Ge and Gi, if they can be assumed time- and space-invariant, can be estimated by observing that Ge = @pe,S/@ee and Gi = @pi,S/@ei. This means that compressibility moduli estimates can be based on the slope of the linear function relating the decrease in void ratio with an increase in

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

9

solid pressure in each respective space. In experiments, such as illustrated in Fig. 7, one thus measures the changes in deformation of the tissue sample due to loss of liquid, rather than measuring the liquid pressure itself. Consequently, the fundamental relationship between tissue deformation and loss of liquid pressure should be established by rewriting Eqs. 5 and 6 not for the liquid-to-solid void ratio, but rather for porosities ee and ei, thus @ee @pe, S @ee @p @ee 1 @pe ¼  ¼ e ¼ @t @t @pe, S @t @pe, S Ge, e @t

(7)

@ei @pi, S @ei @p @ei 1 @pi ¼  ¼ i ¼ @t @t @pi, S @t @pi, S Ge, i @t

(8)

The compressibility moduli Ge and Gi as defined in cited literature and Gε,e and Gε,i as defined by Eqs. 7 and 8 are related, as follows from the following relations
 @ee @pe, S @ ee @p @ee 1 1 1 @pe ¼  ¼ ¼ e @t @t @pe, S 1 þ ee @t @pe, S ð1 þ ee Þ2 Ge ð1 þ ee Þ2 @t (9) 1 @pe ¼ Ge, e @t
 @ei @pi, S @ ei @p @ei 1 1 1 @pi ¼  ¼ ¼ i 2 2 Gi ð1 þ ei Þ @t @t @t @pi, S 1 þ ei @t @pi, S ð1 þ ei Þ (10) 1 @pi ¼ Ge, i @t According to expressions given by Eqs. 9 and 10, Gε,e = Ge ∙ (1 + ee) 2 and Gε,i = Gi ∙ (1 + ei) 2. Note, however, that relations ee = ee/(1 + ee) and ei = ei/ (1 + ei) require these quantities to be intrinsic (properties of each respective space) as opposed to bulk properties that are featured in the dual-porosity model equations, Eqs. 3 and 4. Further explanation of this issue can be found in subsection “Cell Volume Fraction and Implications of Its Omission.” By introducing compressibility moduli defined through porosity, a simplification has been made that comes at a cost, a trade-off opted for in order to keep the model simple and comprehensive. Consequently, the model results are valid for small piston displacements in cases where tissue is not severely damaged. The moduli Gε,e and Gε,i that were introduced should be understood as space- and time-averaged values. The approximate values of compressibility moduli can be directly estimated from experiments (see subsection “Estimating Compressibility Moduli Gε,e and Gε,i”), and average values are then determined based on fitting model results to experimental data. Another simplification limiting model applicability to small piston displacements is the assumption of linear elastic deformation of tissue. Compressibility moduli Gε,e and Gε,i are, more strictly following definitions of thermodynamics and stress mechanics, in fact non-normalized bulk elastic moduli. A more in-depth and

10

S. Mahnič-Kalamiza

rigorous treatment in filtration-consolidation theory would also require introduction of material coordinates (see e.g., Petryk and Vorobiev 2013) to account for time- and space-variable porosity, i.e., dzm ¼ ð1  eÞ  dz

(11)

Cell Volume Fraction and Implications of its Omission There is also another important simplification in the model constitutive equations that requires some attention, and that is the omission of the cell volume fraction. Plant tissues normally comprise more than 50 % cells, sometimes – as for instance in the case of fresh sugar beet taproots with high turgidity – much more than 90 % (close to 97–98 %). Constitutive equations 3 and 4 still hold without accounting for this large cell volume fraction, but only if all quantities are understood as bulk tissue properties, i.e., normalized to the entire tissue volume, rather than intrinsic properties, which would be normalized to the volume fraction of each respective space (intra- or extracellular). Introduction of the phenomenologically derived compressibility moduli Gε,e and Gε,i has the advantage of avoiding the necessity to explicitly account for the cell volume fraction. This is advantageous due to the experimental estimation of parameters Gε,e and Gε,i (refer to subsection “Estimating Compressibility Moduli Gε,e and Gε,i”), which are easier to arrive at, as compared to estimates on effective cell volume fraction (note that due to irreversible electroporation, cell volume fraction is a function of treatment intensity). A noteworthy downside of this omission in the approach using bulk tissue properties instead of intrinsic quantities is that the literature-based estimate of the hydraulic permeability coefficient determining extracellular liquid flow rate ke is likely to require an augmentation in order to arrive at its realistic value. In other words, if only a few percent of the total tissue volume is available for extracellular liquid flow while model equations feature bulk-volume normalized parameters, one can expect the estimated intrinsic value of ke will need to be scaled accordingly. Another consequence of working with bulk tissue properties is theoretical; the relations ee = ee/(1 + ee) and ei = ei/(1 + ei) are written for intrinsic properties of tissue, and by consequence, relations in Eqs. 5, 6, 7, 8, 9, 10, 11, and 12 must also be understood as written using intrinsic properties. If one would wish to work with material compressibility moduli Ge and Gi and void ratios ee and ei, one would also have to account for the cell volume fraction. This point is a current challenge in the dual-porosity model development that will be addressed by a future, expanded, and improved formulation of the model. Constitutive Equations Revisited – Darcy Law and the Final Model Formulation Turning the attention to the right-hand side of Eq. 3, the liquid flow velocity in extracellular space qe is given by Darcy law as

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

qe ¼ 

ke @pe μ @z

11

(12)

if assuming unidirectional flow in the direction of the principal axis of applied pressure (z). The source term υi-e(z,t) representing liquid flow through the porous membrane can be written in terms of local pressure difference between the intracellular and extracellular pressure, resulting in α υie ¼ ðpi  pe Þ μ

(13)

Wherein, the proportionality coefficient α is first introduced, the estimation of which is discussed in the following subsection. Combining Eqs. 3, 4, 5, 6, 7, 8, 9, 10, 12, and 13 while dropping the notation of spatial-temporal dependency of both liquid pressures gives the following final form of the dual-porosity model equations
 1 @pe @ ke @pe α   ð pi  pe Þ ¼ 0 Ge, e @t @z μ @z μ

(14)

1 @pi α þ ð p  pe Þ ¼ 0 Ge, i @t μ i

(15)

Note the absence of a spatial derivative in Eq. 15. The above system of equations is no longer a system of two coupled partial differential equations as was the system Eqs. 3 and 4, but rather a coupled integrodifferential system with constant coefficients. Such systems admit readily obtainable analytical solutions. What remain missing in order to obtain a particular analytical solution are the appropriate initial conditions and the boundary conditions for Eq. 14. The initial and boundary conditions for the system of Eqs. 13 and 14 follow from the particular system setup and choice of coordinate system (refer to Fig. 2) and are pe0 ¼ pe0 ðz, 0Þ ¼ PE

(16)

pi0 ¼ pi0 ðz, 0Þ ¼ PE  @pe  ¼0 @z z¼h

(17)

pe jz¼0 ¼ 0

(19)

(18)

The boundary conditions in Eqs. 18 and 19 are known as mixed or NeumannDirichlet boundary conditions. They require some attention and care when

12

S. Mahnič-Kalamiza

assembling the finite difference matrices in the formulation of the numerical solution, which this chapter, however, does not concern. A note on initial condition, Eqs. 16 and 17. Externally applied pressure can be assumed as equally distributed throughout the tissue sample on both the intra- and extracellular liquid phase in case the sample thickness is relatively small (in relation to the number of cell layers and piston-tissue contact surface). If the sample thickness impact to pressure distribution cannot be neglected due to sample dimensions, a more suitable approximation for pressure distribution might be linear, for details, see, e.g., Lanoiselle et al. (1996). In experiments illustrated by Fig. 7, the conditions (use of thin samples) justify the applicability of this initial condition. One can also suppose that external pressure PE redistributes itself equally onto the liquid phase of both the extracellular ( pe0) as well as the intracellular space ( pi0) at the beginning of a pressing experiment, after the extracellular air is eliminated and replaced by liquid at the beginning of the consolidation stage.

From Liquid Pressure to Observable Parameters – The Liquid Pressure Loss Integration and Tissue Deformation/Consolidation In pressing experiments, as set up as shown schematically in Fig. 7, the observed quantity is the deformation of a sample block of tissue. Since the dual-porosity model is describing the dynamics of liquid pressure loss in the extracellular and intracellular space, an expression giving deformation as a function of the cumulative change of pressure throughout the sample must be found in order to enable comparison of model results with those of experiments. The relationship between loss of liquid pressure and deformation is already given in Eqs. 5 and 6. Since the total deformation is the sum of deformations of the extracellular and of the intracellular space, the combined total absolute deformation equals SðtÞ ¼ Se ðtÞ þ Si ðtÞ;

(20)

Total deformation can be expressed as a spatial integral of local infinitesimal differences in void ratio e, therefore Sð t Þ ¼

ð h ð ee ðz, 0Þ 0 ee ðz, tÞ

dee dzþ

ð h ð ei ðz, 0Þ 0

1 dei dz ¼ G e ei ðz, tÞ

ð h ð pe ðz, 0Þ 0

1 dpe dzþ G i pe ðz, tÞ

ð h ð pi ðz, 0Þ 0 pi ðz, tÞ

dpi  dz: (21)

However, since compressibility moduli were defined through porosity e instead of void ratio e, and additionally for reasons of convenience, the relative deformation sε can be introduced and is defined as se ðtÞ ¼

Se ðtÞ 1 ¼ h Ge, e

ð 1 ð pe ðz, 0Þ 0

pe ðz, tÞ

dpe  dz þ

1 Ge, i

ð 1 ð pi ðz, 0Þ 0

pi ðz, tÞ

dpi  dz

(22)

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

13

where h is the initial tissue sample height. Equation 22 gives relative deformation as a function of loss of liquid pressure within the tissue. It can be used to obtain model results from simulated pressure loss profiles, enabling the comparison of the resulting kinetics with experimental data.

The Hydraulic Permeability of the Electroporated Biological Membrane and Estimation of Model Parameters Estimating a – The Hydraulic Permeability of an Electroporated Biological Membrane The proportionality coefficient α is a dimensionless parameter relating the intracellular and extracellular deformation due to transmembrane flux with the pressure drop across the plasma membrane. In accordance with model design and assumptions, it is a function of membrane permeability ki, multiplied by a corrective geometrical factor ξ with unit of m2. The corrective factor ξ accounts for the geometrical configuration of the cell and its porous membrane by relating intracellular space porosity with volume-averaged (bulk) transmembrane flux (further explanation can be found in the Appendix of Mahnič-Kalamiza and Vorobiev 2014). For negligible membrane thickness as compared to the size of the cell, ξ equals the square of specific surface (surface-to-volume ratio), i.e., ξ = (A/V ) 2. On the level of a biological cell, where transmembrane fluid transport occurs, the surface A and volume V are those of a single cell. For an idealized, average, spherical cell of a given radius R, the factor ξ can be calculated as ξ¼

9 R2

(23)

Turning now to the effect of electroporation treatment on the hydraulic permeability coefficient ki, it should be noted that according to the theory of electroporation, electric field of sufficient strength increases the permeability of the plasma membrane. The effect has been attributed to the creation of aqueous pathways or pores in the membrane. These pores nucleate at some initial radius and can expand in both number and size during the application of the electric field. The effect has a transient as well as a long-lasting component, i.e., transient and long-lasting pores are created in the membrane (Pavlin and Miklavcic 2008). It has been demonstrated by experiment, see, e.g., Saulis and Saule (2012), that long-lasting pores permeable to molecules of considerable molecular weight (tens of kDa) can exist in an electroporated membrane for minutes after the application of electric pulses, though they are subject to resealing under favorable physiological conditions. One can, assuming an average stable pore diameter and pore fraction ratio (i.e., the surface fraction of all pores per one cell), theoretically estimate how electroporation changes the hydraulic permeability of the cell membrane.

14

S. Mahnič-Kalamiza

The derivation of the sought relation starts by first relating membrane permeability ki with permeability of a single aqueous pore. The absolute value of membrane flux is, according to Darcy law, jQ m j ¼

ki Am Δp μ l

(24)

However, since the total membrane flux is also the sum of all single-pore fluxes, of which there are as many as there are pores (i.e., Np), it is possible to write   N p kp Ap Δp jQm j ¼ N p Qp  ¼ l μ

(25)

From equating transmembrane flux in Eqs. 24 and 25 follows the relation ki ¼

N p k p Ap ¼ f p kp Am

(26)

where fp = NpAp/Am is the pore surface fraction. Some estimates on the surface fraction of long-lasting pores can be encountered in literature (Pavlin and Miklavcic 2008); however, these estimates are based on experiments on animal cells in vitro and use of a particular treatment protocol commonly used in biomedical applications of electroporation. As extrapolation of these results to plant tissues is difficult, one can, as a general guideline, assume that due to larger plant cells as compared to animal cells and more intense (in the sense of number, amplitude of pulses, delivered energy) treatment protocols used in plant tissue electroporation, the corresponding pore surface fraction will be higher than in experiments reported on in the literature. Since according to the electroporation theory, pore surface fraction is strongly dependent on induced transmembrane voltage, more work is needed to help determine the most important parameters that describe membrane long-term permeability with respect to the treatment protocol. The only remaining uncharacterized parameter from Eq. 26 that requires an estimation is the single pore intrinsic hydraulic permeability kp. To that end, the Hagen-Poiseuille equation for cylindrical pores of length l and radius rp is employed in combination with Darcy law. This yields Δp ¼

8μlQp 8μlQp μlQp ¼ 2 ¼ πr p 4 r p Ap k p Ap

(27)

Based on Eq. 27, kp can be expressed as rp2/8. The estimate on the radius of an average stable pore rp can be obtained from literature or some form of special experiment. Moreover, it allows for coupling of advanced theoretical models of electroporation (pore formation and evolution – see, e.g., Smith et al. 2004;

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

15

Krassowska and Filev 2007) that give estimates on temporal and spatial distribution of a population of pores with a statistical distribution of radii, to the dual-porosity model, which is an interesting prospect for future model development.

Estimating Compressibility Moduli G«,e and G«,i To estimate the compressibility moduli, one can conceptualize two experimental scenarios. In the first experiment, freshly cut intact tissue is subjected to pressing under pressure of insufficient strength to cause cell rupture by itself. Under these conditions, an eventual equilibration between the pressure applied via the piston and the intracellular liquid pressure can be observed. When this equilibrium is reached, tissue sample stops deforming, and the sample sustains a certain measurable, however small, deformation. If the pressure applied is sufficiently large to completely express extracellular fluid while simultaneously not compromising the integrity of the cell plasma membrane, this measured deformation is only due to the compression of extracellular space. The mathematical formulation of this reads s e, 1

1 ¼ se ðt ! 1Þ ¼ Ge,e

ð 1 ð pe ðz, 0Þ 0

1 dpe  dz ¼ G e,e pe ðz, t!1Þ

ð 1 ð PE 0

0

dpe  dz ¼

PE (28) Ge,e

Equation 28 provides means of estimating Gε,e directly from pressing experiments performed using intact, untreated (non-electroporated) tissue. With known deformation at “infinite” time (complete equilibrium) and known applied pressure PE, it is estimated using Ge,e ¼

PE se, 1

(29)

Since Gε,e is a function of the void ratio, which is not constant in time (and is only approximately constant throughout the tissue block along z, provided the sample is thin), the value obtained by Eq. 29 is a rough initial estimate and a good approximation for untreated tissue. It is expected to decrease with increasing treatment intensity, since it is not a true material property but depends on void ratio and the effective cell volume fraction (as established in subsections “On Porosity, Void Ratio, and the Compressibility Modulus” and “Cell Volume Fraction and Implications of Its Omission”). Its average value must be determined by optimization against experimental kinetics. The second conceptual experiment can be proposed for estimating Gε,i. If the cell membranes are permeabilized (by e.g., electroporation), liquid will flow from the intracellular to the extracellular space due to applied pressure and through the extracellular space out of the tissue block. At complete equilibrium (i.e., after “infinite” time), all of the liquid will be expressed from the sample, and the externally applied pressure will be balanced by the sum of solid pressures of the intracellular and the extracellular space. This is written

16

S. Mahnič-Kalamiza

ð ð ð ð 1 1 pe ðz, 0Þ 1 1 pi ðz, 0Þ dpe dzþ dp dz ¼ ... Ge,e 0 pe ðz, t!1Þ Ge,i 0 pi ðz, t!1Þ i   ð ð ð ð 1 1 PE 1 1 PE PE PE PE Ge,i þ Ge,e ¼ dpe dzþ dpi dz ¼ þ ¼ Ge,e 0 0 Ge,i 0 0 Ge,e Ge,i Ge,e Ge,i (30)

s1 ¼ se ðt ! 1Þþsi ðt ! 1Þ ¼

If Gε,e is known, e.g., determined previously according to Eq. 29, and deformation in an experiment is measured when using strongly permeabilized tissue, by expressing Gε,i from Eq. 30, one obtains Ge, i ¼

PE Ge, e s1 Ge, e  PE

(31)

which is a function of either previously known (estimated) or measurable parameters. This estimate gives an approximate value for Gε,i in case of damaged tissue. The average value for untreated or only moderately electropermeabilized tissue is expected to be much higher, as higher means less compressible. Similarly to Gε,e, the value corresponding to the particular degree of electropermeabilization can be determined by obtaining an agreement between model results and experimental data.

Estimating the Intrinsic Hydraulic Permeability Coefficients ke and ki Hydraulic permeability of tissue is almost always measured rather than calculated (Buttersack and Basler 1991), due to high complexity of water pathways within tissue that makes theoretical estimates hard to obtain and the biological diversity, which renders these estimates unreliable across different plant species and even across samples of a single species. Measurements on a number of plant tissues and yeast cells have shown a wide range of permeability, spanning several orders of magnitude, for both tissue as well as plasma membrane of individual cells (Tomos 1988; Buttersack and Basler 1991). It is nevertheless possible to find several accounts of measurement of hydraulic conductivity (Lp) of industrially important plant tissues in literature. The following derivations demonstrate how it is possible to recalculate these measurements in order to estimate the intrinsic hydraulic permeability coefficients required by the dual-porosity model. The hydraulic conductivity Lp found in literature is normally calculated based on an experiment where a tissue sample is subjected to a pressure difference and liquid volume flux is measured, an experiment which, conceptually, does not significantly differ from the one represented in Fig. 2. With known flux and pressure, the hydraulic conductivity is Lp ¼

q Δp

(32)

On the other hand, the Darcy law relates the pressure drop across a conduit of length l with the rate of liquid flow through it, q, as

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

q¼

k Δp μ l

17

(33)

Inserting q from Eq. 32 into Eq. 33 and expressing k gives k ¼ Lp  μ  l

(34)

Equation 34 can be used to calculate the intrinsic hydraulic permeability of tissue from measurements obtained via experiments as described above. Assuming negligible symplastic flow (i.e., only apoplastic), this permeability is an estimate of the hydraulic permeability of extracellular space, ke. The intracellular hydraulic permeability ki is in fact the hydraulic permeability of the plasma membrane of thickness l. This coefficient is expected to change when treatment, be it electrical, mechanical, thermal, chemical, enzymatic or other, is applied to the tissue. Its initial value (i.e., for untreated tissue) can be estimated from pressureprobe experiments. Tables of cell membrane hydraulic permeability can be found in literature on plant physiology for many plants and some yeast species as well.

The Analytical Solution of the Model The particular solution of the integrodifferential system of Eqs. 14 and 15 for initial and boundary conditions Eqs. 16, 17, 18, and 19 can easily be obtained, for instance, by employing the method of separation of variables. A detailed derivation is given for a mathematically completely analogous case of the dual-porosity model applied to the problem of solute diffusion in electroporated tissue (found in Mahnič-Kalamiza et al. 2014a). The final model equations for the pressing case are almost identical to the ones obtained for the corresponding solution of the diffusion problem, after appropriate replacements of parameters have been made. The few additional exceptions to this analogy are: (i) sine replacing the cosine in the Fourier series and (ii) a required update in the eigenvalues due to different boundary conditions. For extracellular and intracellular liquid pressure, the final equations are, respectively, pe ðz,tÞ ¼


 1     4pi0 X 1  ð2n þ 1Þπ γ n, 1 τ þ 1 C1 eγn, 1 t þ γ n, 2 τ þ 1 C2 eγn, 2 t sin z 2h π n¼0 2n þ 1 (35)


 1  4pi0 X 1  ð2n þ 1Þπ γ n, 1 t γ n, 2 t τ1 t pi ðz, tÞ ¼ C1 e z þ C2 e e sin 2h π n¼0 2n þ 1 þ pi0 eτ where

1

t

(36)

18

S. Mahnič-Kalamiza


 pe0  1 τ1  γ n, 2 pi0 C1 ¼ γ n, 1  γ n, 2
 pe0 1 1 τ þ γ n, 1 pi0 C2 ¼ γ n, 1  γ n, 2

(37)

(38)

and

γ n1, 2 ¼

ffi   qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  2  τ1 δ þ λn 2 ν  τ1 δ þ λn 2 ν  4λn 2 ντ1 2

(39)

For the sake of algebra throughout Eqs. 35, 36, 37, 38, and 39, the following replacements were made: ν¼

ke Ge, e μ

(40)

αGe, i μ

(41)

τ1 ¼


Ge, e Ge, i



δ¼

1þ

λn ¼

ð2n þ 1Þπ 2h

(42)

and the eigenvalues λn equal (43)

Simulated Liquid Pressure Distributions and Consolidation Kinetics Employing Eqs. 35, 36, 37, 38, 39, 40, 41, 42, and 43 as given in the preceding subsection yields pressure distribution profiles throughout the tissue sample for all times of the simulated experiment. An example of these profiles along the axis of pressure application is shown in Fig. 5 for six different times taken from the interval between 0 and 45 min of the simulated experiment. There is a marked difference in pressure profiles when comparing the calculated pressure in the extra- with that in the intracellular space. In order to produce these pressure profiles, the value of the parameter reflecting the degree of permeabilization

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

19

Fig. 5 Extracellular (a) and intracellular (b) liquid pressure distributions along the principal axis of pressure application for different times of the simulated experiment

of the membrane, i.e., the surface fraction of pores, has been chosen from the range yielding the membrane hydraulic permeability that is not on a comparable order of magnitude with that of the extracellular space. As a result, there is a noticeable hindrance effect of the plasma membrane to the liquid flow, and it is evident from the comparison of pressure profiles that liquid is being retained within the cells for a longer period while the extracellular space is vacated at a greater rate. The intracellular liquid nevertheless continues replenishing the extracellular space with liquid, and the resulting final extracellular liquid amount at the end of the simulated experiment (at t = 2700 s in this case) is still greater as compared to that which would have been remaining in tissue were it not for the outflow of intracellular juice from the intra- into the extracellular space. The pressure profiles shown in Fig. 5 are then integrated according to Eq. 30 to produce final relative consolidation (i.e., sample deformation) kinetics, an example of which is shown in the right-hand side of Fig. 6. The pressure profiles in the intra- and extracellular space that are integrated, scaled by the compressibility moduli, and then summed to yield the cumulative tissue deformation therefore determine the dynamics of tissue compaction under external pressure. The membrane permeability will determine the particular dynamics of this deformation (the shape of the consolidation kinetics), while the compressibility moduli, still partly phenomenological in nature in the presented version of the dual-porosity model, will determine both the extent of the initial, rapid deformation (see kinetics in Fig. 6 in range of 0–500 s) as well as of the final or total attainable extent of compression of the tissue. In this respect, the compressibility moduli are a reflection of the dewaterability of the tissue, and since dewaterability is dependent on the degree of permeabilization of the membranes in bulk tissue, these parameters will also depend on the parameters of the specific electroporation protocol applied.

Fig. 6 Annotated simulated consolidation kinetics, obtained from liquid pressure distributions as shown in Fig. 5 and according to the integration Eq. 30

20 S. Mahnič-Kalamiza

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

21

Model Verification and Application by Example To demonstrate how the model can be applied, an experiment has been proposed as schematically shown in Fig. 7, whereby a tissue sample is first exposed to treatment by pulsed electric fields, then placed into a specially fabricated treatment chamber where pressure is applied at the top surface and the bottom support is porous, creating a system as depicted by Fig. 2. This system can be described and analyzed with the dual-porosity model, and measured kinetics of tissue consolidation (and thus liquid expression) can be compared against simulated kinetics obtained from the dual-porosity model. Initial model parameter values are based on literature or experimental estimates as described in subsection “The Hydraulic Permeability of the Electroporated Biological Membrane and Estimation of Model Parameters”; however, in order to obtain a good agreement between experimental and modeled kinetics, these parameters were allowed to vary within an acceptable range and their exact values were determined in the loop of an optimization algorithm. The optimization that ensures the model kinetics fit the experimental consolidation curves is also necessary due to a lack of reliable theoretical means of estimating such parameters as pore surface fraction, hydraulic permeability of the plasma membrane (for treated and intact tissue), and extracellular hydraulic permeability (note the issues with estimating these parameters noted in subsections “Cell Volume Fraction and Implications of Its Omission” and “The Hydraulic Permeability of the Electroporated Biological Membrane and Estimation of Model Parameters”). It could be argued that due to the high number of parameters that are allowed to vary and are determined empirically, the model thus applied detracts from its scientific rigor. However, although the optimal model parameter values are sought

Fig. 7 A schematic diagram of the experiment designed for purposes of verifying the model

22

S. Mahnič-Kalamiza

for in a multiparametrical sweep with a high number of degrees of freedom, the parameter values that are determined in this way are then carefully examined and analyzed in light of the electroporation theory and the assumptions stemming from the process of model construction. The necessary changes to parameter values should behave according to the type of tissue that is experimented with, and thus reflect the characteristics of a particular treatment protocol used. As an example, during model construction, an implicit supposition was made that compressibility moduli defined through bulk porosity of each respective space will depend on treatment intensity, due to the (variable) fraction of cells that are reversibly and irreversibly electroporated. Moreover, the effective average pore surface fraction ( fp) should increase non-linearly and experience a sharp sigmoid dependence on treatment intensity (e.g., on applied voltage). These relationships between material and treatment properties and the parameter values that are obtained from the model after fitting to experimental data, if well established, are what verifies the dual-porosity model theory and invites further refinement and improvement of the model with the prospect of eventually arriving at a truly mechanistic model. Figure 8 illustrates the results of a model study designed for the purposes of dualporosity model verification. Experiments were performed according to the procedure shown in Fig. 7 using two types of tissue, apple fruit and sugar beet tap root tissue. In experiments on each type of tissue, only one parameter – the applied voltage to the electrodes – was varied, and consolidation kinetics was recorded. The second part of the study involved running dual-porosity model simulations using estimated values of parameters that were allowed to vary within a predefined range that was deemed acceptable. Figure 8 demonstrates the good agreement that can be obtained between model (solid or dashed lines) and experimental (depicted by symbols) consolidation kinetics. What is more important is the dynamics of parameter value changes that were required to obtain this agreement if comparing the two different materials and the six treatment intensities. A detailed commentary on these results and particular parameter values is outside the scope of this theoretical introductory work and is given in full in Mahnič-Kalamiza et al. (2014b). In summary, the model study provides a means of identifying important parameters in the dual-porosity model that most strongly depend on material properties and those that are functions of PEF treatment intensity. These findings can help with directing future research and model development, and the dual-porosity model of mass transport in tissue can be considered a promising platform for future study of electroporation effects on both tissue structure and mass transport properties.

Conclusion This chapter gives a detailed presentation of the dual-porosity model, which has been developed with the intent of modeling and studying filtration-consolidation behavior of electroporated plant tissues. The paradigm it introduces is that of regarding biological tissue as essentially a two-compartment, two-space, or two-component material of distinct porosity and hydraulic permeability. The objective is to introduce

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

23

Fig. 8 Simulated (lines) and experimental (markers) consolidation kinetics for two model tissues; apple fruit (a) and sugar beet tap root (b) tissue

the concept of a membrane with variable hydraulic properties into a model of mass transport in tissue, thus enabling the coupling of membrane-level electroporation

24

S. Mahnič-Kalamiza

effects to the resulting bulk tissue mass transport processes. More work remains necessary to help determine the most important parameters that describe membrane long-term permeability with respect to the treatment protocol. Nevertheless, the model can at present already be used for evaluation of treatment parameters’ impact to liquid flow in electroporated plant tissues, but perhaps more importantly, it invites future refinement and development for purposes of further elucidating basic principles of pulsed electric field-induced phenomena in biological tissues.

Cross-References ▶ Determination of Pulsed Electric Fields Effects on the Structure of Potato Tubers ▶ Electric Field Distribution and Electroporation ▶ Electric Pulse Parameters Affecting Electroporation Treatment Outcome ▶ Electroporation-Based Applications in Biotechnology ▶ Food Structure Changes Affected by Pulsed Electric Field Treatment ▶ Lipid Pores: Molecular and Continuum Models ▶ Mathematical Models of Pulsed Electric Field Treatment of Plant Tissues and Simulation of Related Phenomena ▶ Modeling Transport Across the Electroporated Membrane ▶ Pulsed Electric Field Treatment for Fruit and Vegetable Processing ▶ Responses of Plant Cells and Tissues to Pulsed Electric Field Treatments ▶ Selective Extraction of Molecules by Pulsed Electric Field Treatment ▶ Techniques to Detect Electroporation in Food Tissues ▶ Water Defects in Phospholipid Bilayers

References Barenblatt GI, Entov VM, Ryzhik VM (1990) Theory of fluid flows through natural rocks. Kluwer, Dordrecht/Boston Buttersack C, Basler W (1991) Hydraulic conductivity of cell-walls in sugar-beet tissue. Plant Sci 76:229–237. doi:10.1016/0168-9452(91)90145-X Čorović S, Mir LM, Miklavčič D (2012) In vivo muscle electroporation threshold determination: realistic numerical models and in vivo experiments. J Memb Biol 245:509–520. doi:10.1007/ s00232-012-9432-8 Dymek K, Rems L, Zorec B et al (2015) Modeling electroporation of the non-treated and vacuum impregnated heterogeneous tissue of spinach leaves. Innovative Food Sci Emerg Technol 29:55–64. doi:10.1016/j.ifset.2014.08.006 Fincan M, Dejmek P (2003) Effect of osmotic pretreatment and pulsed electric field on the viscoelastic properties of potato tissue. J Food Eng 59:169–175. doi:10.1016/S0260-8774(02) 00454-5 Granot Y, Rubinsky B (2008) Mass transfer model for drug delivery in tissue cells with reversible electroporation. Int J Heat Mass Transf 51:5610–5616. doi:10.1016/j. ijheatmasstransfer.2008.04.041 Grimi N, Vorobiev E, Lebovka N, Vaxelaire J (2010) Solid–liquid expression from denaturated plant tissue: filtration-consolidation behaviour. J Food Eng 96:29–36. doi:10.1016/j. jfoodeng.2009.06.039

Dual-Porosity Model of Liquid Extraction by Pressing from Plant Tissue. . .

25

Kotnik T, Pucihar G, Miklavčič D (2010) Induced transmembrane voltage and its correlation with electroporation-mediated molecular transport. J Memb Biol 236:3–13. doi:10.1007/s00232010-9279-9 Krassowska W, Filev PD (2007) Modeling electroporation in a single cell. Biophys J 92:404–417. doi:10.1529/biophysj.106.094235 Langus J, Kranjc M, Kos B et al (2016) Dynamic finite-element model for efficient modelling of electric currents in electroporated tissue. Sci Rep 6:26409. doi:10.1038/srep26409 Lanoiselle JL, Vorobyov EI, Bouvier JM, Piar G (1996) Modeling of solid/liquid expression for cellular materials. AICHE J 42:2057–2068. doi:10.1002/aic.690420726 Lebovka NI, Bazhal MI, Vorobiev E (2000) Simulation and experimental investigation of food material breakage using pulsed electric field treatment. J Food Eng 44:213–223. doi:10.1016/ S0260-8774(00)00029-7 Li J, Lin H (2011) Numerical simulation of molecular uptake via electroporation. Bioelectrochemistry 82:10–21. doi:10.1016/j.bioelechem.2011.04.006 Mahnič-Kalamiza S, Vorobiev E (2014) Dual-porosity model of liquid extraction by pressing from biological tissue modified by electroporation. J Food Eng 137:76–87. doi:10.1016/j. jfoodeng.2014.03.035 Mahnič-Kalamiza S, Miklavčič D, Vorobiev E (2014a) Dual-porosity model of solute diffusion in biological tissue modified by electroporation. Biochim Biophys Acta (BBA) - Biomemb 1838:1950–1966. doi:10.1016/j.bbamem.2014.03.004 Mahnič-Kalamiza S, Miklavčič D, Vorobiev E (2014b) Dual-porosity model of mass transport in electroporated biological tissue: simulations and experimental work for model validation. Innovative Food Sci Emerg Technol. doi:10.1016/j.ifset.2014.09.011 Pavlin M, Miklavcic D (2008) Theoretical and experimental analysis of conductivity, ion diffusion and molecular transport during cell electroporation – relation between short-lived and long-lived pores. Bioelectrochemistry 74:38–46. doi:10.1016/j.bioelechem.2008.04.016 Petryk M, Vorobiev E (2007) Liquid flowing from porous particles during the pressing of biological materials. Comput Chem Eng 31:1336–1345. doi:10.1016/j.compchemeng.2006.12.011 Petryk M, Vorobiev E (2013) Numerical and analytical modeling of solid–liquid expression from soft plant materials. AIChE J 59:4762–4771. doi:10.1002/aic.14213 Pucihar G, Kotnik T, Teissie J, Miklavcic D (2007) Electropermeabilization of dense cell suspensions. Eur Biophys J Biophys Lett 36:173–185. doi:10.1007/s00249-006-0115-1 Saulis G, Saule R (2012) Size of the pores created by an electric pulse: microsecond vs millisecond pulses. Biochim Biophys Acta-Biomemb 1818:3032–3039. doi:10.1016/j. bbamem.2012.06.018 Sel D, Cukjati D, Batiuskaite D et al (2005) Sequential finite element model of tissue electropermeabilization. IEEE Trans Biomed Eng 52:816–827. doi:10.1109/TBME.2005.845212 Smith KC, Neu JC, Krassowska W (2004) Model of creation and evolution of stable electropores for DNA delivery. Biophys J 86:2813–2826 Tomos AD (1988) Cellular water relations of plants. In: Franks F (ed) Water science reviews 3, 1st edn. Cambridge University Press, Cambridge Zhu HX, Melrose JR (2003) A mechanics model for the compression of plant and vegetative tissues. J Theor Biol 221:89–101. doi:10.1006/jtbi.2003.3173

Computational Approach for Electrical Analysis of Biological Tissue Using the Equivalent Circuit Model Airton Ramos and Daniela O. H. Suzuki

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Equivalent Circuit for a Nondispersive Medium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Equivalent Circuit for First-Order Dispersion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dispersion Mechanism and Equivalent Circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Numerical Analysis of the Equivalent Circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modeling Anisotropic, Inhomogeneous, and Nonlinear Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simulation of the Stimulation Effects with Metal Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross‐References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 4 6 10 12 14 19 20 20

Abstract

The equivalent circuit method is a computational approach for electrical analysis in which the properties of the medium are modeled using circuit elements such as conductances and capacitances. Its main application is in the electric field calculation in biological tissues stimulated by electrical potentials applied with metal electrodes in contact with the material. This method allows modeling easily inhomogeneous materials that have dielectric dispersion, anisotropy, and nonlinear electrical behavior. These characteristics are typical of biological tissues. This chapter presents the mathematical foundations of the equivalent circuit

A. Ramos (*) State University of Santa Catarina (UDESC), DEE/CCT, Joinville, Santa Catarina, Brazil e-mail: [email protected]; [email protected] D.O.H. Suzuki (*) Institute of Biomedical Engineering, Federal University of Santa Catarina (UFSC), IEB/EEL/CTC/ UFSC, Florianopolis, Santa Catarina, Brazil e-mail: [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_12-1

1

2

A. Ramos and D.O.H. Suzuki

method and illustrates its main features with a typical example of biological stimulation for the purpose of cell membrane electroporation. Keywords

Equivalent circuit method • Field calculation • Dielectric dispersion • Biological tissues • Electroporation

Introduction The use of computational approaches to solve electromagnetic problems with complex geometries and properties of the materials involved has become common practice in recent decades due to the great development of equipment and software. In the case of biological materials, electromagnetic analysis presents several difficulties that make the computational analysis practically indispensable. In addition to the typical inhomogeneity of biological materials, the following important complex characteristics may be mentioned: biological tissues have strong dielectric dispersion in a wide range of frequencies extending from zero to hundreds of megahertz; biological tissues generally exhibit anisotropy in conductivity and permittivity; biological tissues can present nonlinear electrical behavior as a result of heating or cell membrane electroporation. Among the various numerical methods for electromagnetic analysis, the most commonly used for biological materials are finite difference method (FDM) and finite element method (FEM). The first one has been widely used in the calculation of the absorbed energy (specific absorption rate) for biological materials especially in the range of radio frequencies (Chen and Gandhi 1992; Zhao et al. 2013). In FDM approach, Maxwell’s equations are solved to obtain the electric and magnetic fields distributions in the space discretized with a regular grid of cubic elements. The finite element method is the most used in the field calculation in planning therapies based on biological electroporation: Electrochemotherapy (Corovic et al. 2013; Kos et al. 2010; Miklavcic et al. 2010), irreversible electroporation (Neal et al. 2012; Daniels and Rubinsky 2009), and gene transfer (Lacković et al. 2009; Županič et al. 2010). In the finite elements approach, the geometry under the study is discretized by triangular and tetrahedral elements. This provides greater power to model complex geometries than FDM. Methods based on the modeling by electric circuit elements have also been proposed and used in the analysis of biological materials in recent decades. In the impedance method (IM), the medium is represented by a three-dimensional matrix of impedances modeling the electrical properties of the medium. The analysis is performed in the frequency domain using the phasor method (Nadeem et al. 2003; Ramos 2010). A similar approach is the equivalent circuit method (ECM) that is based on modeling the dielectric relaxation through a network of conductances and capacitances associated with a three-dimensional mesh of orthogonal parallelepipeds

Computational Approach for Electrical Analysis of Biological Tissue Using. . .

3

Fig. 1 Geometric model to obtain the equivalent circuit: Vn, Vm  electric potentials in the center of the volume elements; inm  electric current; A, L  area and distance between the volume elements, respectively

(Ramos et al. 2003). In the ECM, field calculation is converted in the analysis of an electric circuit in the time domain, so that through the Kirchhoff’s first law and continuity equation, the electric potential and accumulated electric charge in each of the mesh elements are calculated in small time steps. The ECM is particularly useful in the analysis of electric process in biological materials, since generally they have high permittivity, low electrical conductivity, and intense dielectric relaxation from extremely low frequencies (ELF) to high frequencies (HF).

Equivalent Circuit for a Nondispersive Medium The modeling using equivalent circuit is based on describing the characteristics of a dielectric material through an electric circuit consisting of conductances and capacitances, which depict the properties associated with the electrical conduction and polarization, respectively, within the material. Consider a small portion of a material divided in two regular halves as shown in Fig. 1. If the volumes are small, an approximate calculation of the conduction and displacement currents can be done by considering uniform electric field in the two volumes and uniform current density at the contact face between the elements. If the portion of material involved in this model is homogeneous and linear in the electric properties, the electric current can be calculated by the simple manner indicated in Eq. 1: A A d inm ¼ σs ðV n  V m Þ þ ε1 εo ðV n  V m Þ L L dt

(1)

Where inm is the electric current, Vm and Vn are the electric potentials at the center of elements m and n, respectively, A is the contact area, L is the distance between the volume elements, σs is the static conductivity, ε1 is the high frequency dielectric constant and εo is the vacuum permittivity. The first term is the conduction current

4

A. Ramos and D.O.H. Suzuki

Fig. 2 Equivalent circuit for a linear, homogeneous, and nondispersive material: Vn, Vm  node potentials; inm  electric current; A, L  area and distance between the volume elements; Gs  static conductance; σs  static conductivity; C1  high-frequency capacitance; ε1  highfrequency dielectric constant; εo  vacuum permittivity

due to the movement of ions. The coefficient that multiplies the electric potential difference is the conductance, whose value depends on the spatial dimensions of the contact between the elements and the conductivity of the material. The second term is the displacement current due to the rate of change of the electric flux inside the elements. Therefore, this term is proportional to the derivative of the electric potential difference in time. The proportionality factor is the capacitance which depends on the physical dimensions and the dielectric constant of the material. The symbols “s” and “1” in the conductivity and dielectric constant indicate that the values are out of the frequency range in which the dielectric dispersion occurs. The first indicates the static conductivity and the second indicates the high-frequency dielectric constant. For this simple system, the equivalent circuit can be constructed with a conductance and a capacitance connected in parallel, as shown in Fig. 2.

Equivalent Circuit for First-Order Dispersion Biological tissues generally have strong dielectric dispersion in the frequency range extending from zero to tens of megahertz mainly related to the space charge polarization due to the accumulation and diffusion of ions on the surface of the biological membranes, but also as a result of the dipolar relaxation of the water and polar macromolecules in the tissue fluids. If the medium has dielectric dispersion, the transient response of the electric current depends on the intensity of the dispersion and the relaxation times of the related processes. The electric polarization of a medium results in a reaction electric field that opposes the applied electric field. Then, taking into account the effect of polarization, the change in current density can be described by the following equation: jd ¼ σd ðE  ER Þ

(2)

Computational Approach for Electrical Analysis of Biological Tissue Using. . .

5

Fig. 3 Equivalent circuit for a first-order dispersion: Vn, Vm  node potentials; idnm  dispersive current; Gd, Cd  dispersion conductance and capacitance, respectively; σd, εd  maximum increases in conductivity and dielectric constant caused by dispersion, respectively

Where jd and σd are increases in the current density and conductivity in relation to the values considered in the Eq. 1. E is the applied electric field, and ER is the reaction field. σd describes how the conductivity of the medium is affected by the electrical polarization. At high frequencies, where the slow polarization processes no longer occur, the reaction field becomes zero, and the conductivity σd adds to the static conductivity of the medium. In a medium whose polarization is proportional to the applied field, the reaction field should be proportional to the accumulated electrical charge in the sample volume. Since Eq. 2 refers to volumetric averaged values, the electric fields E and ER may be easily associated to the potential difference in sample volume. The dispersive current may then be described by the following equation: A id ¼ σd ðΔV  ΔV R Þ L

(3)

Due to the linearity assumption in the polarization process, the reaction potential can be written as being proportional to the accumulated electric charge which, in turn, varies with time at a rate equal to the intensity of the dispersive current: 0

1 ðt 1 id dt0 A id ¼ Gd @ΔV  Cd

(4)

0

where Gd and Cd are the conductance and capacitance related to the dispersive process. This mathematical model is represented by the equivalent circuit shown in Fig. 3. In order to obtain the dispersive current in the time domain to an arbitrary voltage waveform, it can be performed an iterative calculation with small time steps “Δt” compared to the relaxation time of the circuit τ = Cd/Gd. The potential difference across the capacitor can be calculated from the following equation: Gd ðV n  V m  V c Þ ¼ Cd

V c  V 0c Δt

(5)

6

A. Ramos and D.O.H. Suzuki

where Vc0 is the previous value of the potential in the capacitor, i.e., the potential at time t0 = t  Δt. Vc can be obtained by solving Eq. 5: Vc ¼

ðΔt=τÞðV n  V m Þ þ V c 0 ð1 þ Δt=τÞ

(6)

The dispersive current is then obtained in the following manner: 

idnm

ðV n  V m Þ  V c 0 ¼ G d ðV n  V m  V c Þ ¼ G d ð1 þ Δt=τÞ

 (7)

Dispersion Mechanism and Equivalent Circuit In the conductivity and permittivity spectrum of biological tissues, it is generally possible to distinguish two bands, named alpha and beta, in the frequency range that extends from ELF to about 1 MHz. The physical mechanisms involved in the dispersion processes are attributed to diffusion and accumulation of hydrated ions on the surface of plasma membranes. The diffusion of counter ions in the double layer around the cells, according to the model proposed by Schwarz (Schwarz 1962), produces strong dielectric dispersion whose intensity and relaxation time depend on the radius of the cell (R), ion mobility (μ), temperature (T), and concentration of counter ions (no), according to the Eqs. 8 and 9 below (Schwarz 1962): e2 Rno εo K B T

(8)

e R2 2μKB T

(9)

Δεα  τα 

where “e” is the proton electric charge and “KB” is Boltzmann’s constant. With these formulas it can be estimated that the alpha dispersion in cell suspensions can produce variations in the dielectric constant of the order of 105 with relaxation times of the order of milliseconds. Similarly, the accumulation of ions on the inner and outer surfaces of the plasma membrane produces intense dielectric dispersion, which depends on the cell radius, membrane capacitance (Cm), and internal (σi) and external (σo) conductivity according to Eqs. 10 and 11 valid for spherical cells in suspension (Foster and Schwan 1995): Δεβ 

9RCm 4εo

(10)

Computational Approach for Electrical Analysis of Biological Tissue Using. . .

7

Fig. 4 Equivalent circuit of a biological material including first-order dispersions: Vn, Vm  node potentials; inm  total current; is  current in the static conductance Gs; C1  highfrequency capacitance; iα, iβ, iγ  dispersive currents; Gα, Gβ, Gγ  dispersion conductances; Cα, Cβ, Cγ  dispersion capacitances

 τβ  RCm

1 1 þ σi 2σo

 (11)

With these formulas it can be estimated that the beta dispersion produces variations in the dielectric constant of the order of 104 with relaxation times less than a microsecond. The formulas (Eqs. 8, 9, 10, and 11) are strictly valid for dilute cells in suspension but were taken as approximations for tissues when considering the volume fraction very close to unity. In addition to the alpha and beta relaxations, a third dispersive process with relaxation time of tens to hundreds of nanoseconds typically shows significant effects from 1 MHz to higher frequencies in biological tissues. Some different mechanisms may contribute to this gamma dispersion. However, it is estimated that the dipolar relaxation of protein or water molecules bound to proteins can be the most important processes (Foster and Schwan 1995). The circuit in Fig. 4 incorporates all processes described above. The C1 capacitance describes the polarization process with much lower response times than τα, τβ, and τγ. Conductance Gs describes the conduction at very low frequencies. Dilute biological cells suspensions often present relaxation process that can be well modeled by the Debye equation (Foster and Schwan 1995):

8

A. Ramos and D.O.H. Suzuki _

ε ð ωÞ ¼ ε 1 þ

Δε 1 þ jωτd

(12)

where ω is the angular frequency and j = (1)1/2. Dense biological tissues, however, have different behavior. The irregular shape of cells and interstitial spaces and the high density of interfaces result in dispersions that may depart significantly from the Debye formula. In such cases the use of empirical models involving continuous distributions of relaxation times, although with more complex formulation, becomes an alternative. For example, one of the most used models in the interpretation of spectral data in biological materials is the Cole-Cole function (Cole and Cole 1941): Δε

_

ε ðωÞ ¼ ε1 þ

1 þ ðjωτo Þ1α

(13)

where α is a parameter whose value should be adjusted empirically (α < 1), and τo is the central relaxation time of the distribution. A dispersion function based on a continuous distribution of relaxation times can be described as a linear combination of first order dispersions in which the coefficients are defined by a distribution function g(lnτ) as follows (Bottcher and Bordewijk 1978): 1 ð

_

ε ðωÞ  ε1 ¼ Δεr 0

gðlnτÞ dðlnτÞ 1 þ jωτ

(14)

where g(lnτ) meets the normalization condition: 1 ð

gðlnτÞ dðlnτÞ ¼ 1

(15)

0

In the case of Cole-Cole dispersion, g(lnτ) is given by Eq. 16 (Bottcher and Bordewijk 1978): gðlnτÞ ¼

1 senðαπ Þ 2π cosh½ð1  αÞlnðτo =τÞ  cos ðαπ Þ

(16)

Figure 5 shows biological tissue dispersion spectra reproduced from parameters given by Gabriel et al. (Gabriel et al. 1996). The authors performed the adjustment of a model with four dispersion band described with Cole-Cole function to the experimental values obtained in the frequency range from 10 Hz to 100 GHz for many human biological tissues. ε1, σs and Δε, τo and α for each band were determined through a least squares method.

Computational Approach for Electrical Analysis of Biological Tissue Using. . .

9

Fig. 5 Dispersion spectra in human biological tissues according to data published by Gabriel et al. (Gabriel et al. 1996): (a) and (b) skin; (c) and (d) skeletal muscle (not specified orientation); (e) and (f) liver. — Model with three bands of Cole-Cole according to Eq. 17;   model with three Debye bands for each relaxation according to the Eqs. 18 and 19

To obtain the spectra of Fig. 5, it was discarded the band with very fast relaxation of the order of 1011s which not significantly influence the results within the frequency range of interest. The mathematical expression of the dielectric constant of the materials listed in this figure can then be described by the following equation: _

ε ð ωÞ ¼ ε 1 þ

Δεα 1 þ ðjωτα Þ

1α

þ

Δεβ Δεγ  1γ  1β þ 1 þ jωτγ 1 þ jωτβ

(17)

10

A. Ramos and D.O.H. Suzuki

On the other hand, the dashed curves in this figure refers to an approach in which each band is represented by the sum of three Debye dispersions with time constants τo, 30τo e τo/30, and amplitudes determined by the distribution function. For example, in the alpha band, the dielectric constant is calculated as follows: _ ε α ð ωÞ

  Δεα

ahigh ao alow þ þ 1 þ jωτα 1 þ jω30τα 1 þ jωτα =30

gðlnτα Þ gðlnτα Þ þ gðlnð30τα ÞÞ þ gðlnðτα =30ÞÞ gðlnð30τα ÞÞ alow ¼ gðlnτα Þ þ gðlnð30τα ÞÞ þ gðlnðτα =30ÞÞ gðlnðτα =30ÞÞ ahigh ¼ gðlnτα Þ þ gðlnð30τα ÞÞ þ gðlnðτα =30ÞÞ

 (18)

ao ¼

(19)

The choices of the relaxation times for this model with three Debye bands were based on minimizing the mean square error calculated in relation to the equation of the Cole-Cole model. In an equivalent circuit model containing N Debye bands, the dielectric constant and conductivity as functions of frequency are calculated by the following equations: εr ð ωÞ ¼ ε 1 þ

N X i¼1

σðωÞ ¼ σs þ

Δεi 1 þ ω2 τ2i

N X ω2 τ2i Δσi 1 þ ω2 τ2i i¼1

(20)

(21)

where Δεi, Δσi, and τi in each band are related by Eq. 22 (Bottcher and Bordewijk 1978). Δσi ¼

εo Δεi τi

(22)

The graphs in Fig. 5 show good agreement between models up to 106 Hz. Naturally, other ways of adjusting the relaxation times and amplitudes can be implemented to improve the accuracy of the first order model. Anyway, it can be concluded that it is possible to represent the dielectric dispersion by relaxation of a biological tissue through a set of Debye bands. This means that the equivalent circuit of the Fig. 4 can be used in the electrical analysis in a biological tissue if the number of dispersive branches and the values of Gd and Cd parameters are properly chosen for correctly representing the dielectric dispersion in the frequency range of interest.

Computational Approach for Electrical Analysis of Biological Tissue Using. . .

11

Fig. 6 Discretization scheme using in the modeling with the equivalent circuit

Numerical Analysis of the Equivalent Circuit Computational analysis of the electric field distribution and electrical current in biological tissues stimulated by electrical potential applied from electrodes can be performed using the equivalent circuit of the materials applied to a three-dimensional array of volume elements that discretize the space under analysis. Figure 6 shows the discretization scheme with cubic blocks. In the center of each block it is defined a node of the equivalent circuit. There are six neighboring blocks to each block in the mesh. On the boundaries of the domain some connections are replaced according to the boundary conditions. Numerical values of the electric potential in the equivalent circuit are obtained as the solution of the node equations resulting from applying Kirchhoff’s first law on all nodes of the mesh. The node equation to the central node in Fig. 6 can be written as follows:  6  X   d C1nm ðV n  V m Þ þ is þ iα þ iβ þ iγ nm ¼ 0 dt m¼1

(23)

According to the continuity equation, the current summation in Eq. 23 is equal to the negative of the time derivative of the electric charge in the central node: 6  X m¼1

is þ iα þ iβ þ iγ

 nm

¼

dQn dt

(24)

12

A. Ramos and D.O.H. Suzuki

Substituting Eq. 24 into Eq. 23 and integrating both, the basic equations for solution of the equivalent circuit are obtained: " Vn ¼

1 6 X

Qn þ

C1nm

6 X

# C1nm V m

(25)

m¼1

m¼1

Qn ¼ Q0n  Δt

6  X

is þ iα þ iβ þ iγ

 nm

(26)

m¼1

where Q0 n is the value of the charge at time t0 = t  Δt. The current at the static conductance (is) is calculated by Ohm’s law, while the other currents in Eq. 26 are calculated according to Eqs. 6 and 7. In order to obtain the correct convergence in the numerical integration, the time step should be limited to a value lower than the relaxation times of the circuit. The limit value is defined by gamma relaxation time or the time constant τs = ε1/σs of the C1||Gs coupling. The boundary conditions are of two types: electric potential defined on the surface of metal electrodes and null electric field perpendicular on the other domain boundaries. In the internal interfaces between two media with different conductivities and permittivities, the requirement of continuity of potential and current density are met naturally.

Modeling Anisotropic, Inhomogeneous, and Nonlinear Media Anisotropy and inhomogeneity are easily modeled with the equivalent circuit method. In each volume element of the discretization mesh, the material properties (σs, ε1, α, Δεα, τα, β, Δεβ, τβ, γ, Δεγ, τγ) should be specified in each direction of the coordinate system. Thus, the properties can easily vary from one region to another as well they can be modified according to the direction which they refer. Biological tissues usually have anisotropy in electrical properties. Skeletal muscle, for example, has higher conductivity and lower dielectric dispersion in the direction parallel to the muscle fibers than in the perpendicular direction (Foster and Schwan 1995). The most important nonlinear effects in biological tissues are caused by electroporation and heating. The problem of electrical and thermal coupled calculation including the heat flow in the medium is beyond the scope of this chapter. But a simple approach can be used when the electrical stimulation is short. As a result of heating, electrical conductivity in a biological tissue is increased due to the reducing viscosity of the fluid in which the ions move both inside and outside the cells. A simple model is described by Eq. 27 (Neal et al. 2012): σ ð T Þ ¼ σ o ½ 1 þ χ ð T  To Þ 

(27)

Computational Approach for Electrical Analysis of Biological Tissue Using. . .

13

where T is the temperature in each volume element σo is the conductivity at room temperature To at which the material is considered to be in thermal equilibrium prior to application of the electric field and χ is the temperature coefficient of conductivity for the considered medium. Assuming the heat generation by other processes than Joule effect and the heat flow are negligible during the field application, the temperature increases with time according to the Eq. 28: ΔT 

Pdiss Δt cρ

(28)

where Pdiss is the power dissipated in each volume element, c is the specific heat capacity, and ρ is the density of the material. In the equivalent circuit, the elements that are most affected by heating are those who depend directly on fluid conductivities. Such is the case of static conductance Gs, which depends on the external conductivity, and Gβ dispersion conductance, which depends on both the external and internal conductivity. Thus, Eq. 27 can be used to correct the values of these conductances: Gs ¼ Gso þ Gso χ ðT  To Þ

(29)

Gβ ¼ Gβo þ Gβo χ ðT  To Þ

(30)

where Gso and Gβo are the values of the conductances at temperature To. The use of this approach requires that the dissipated power is calculated in every time step Δt and Eqs. 28, 29, and 30 are used to update the values of the conductances in each volume element in the mesh. According to Eqs. 9 and 11, the relaxation times also depend on the temperature, and in a more detailed approach this can be included in the equivalent circuit where they influence the calculation of the dispersive currents according to Eqs. 6 and 7. The other nonlinear effect of most interest is electroporation, which is responsible for increasing the conductivity of the biological tissue as a result of increasing the nonselective plasma membrane permeability. The electroporation evaluated in several biological tissues occurs when the applied electric field is greater than a threshold value between 8 and 40 kV/m and becomes irreversible when the electric field exceeds another threshold between 80 and 120 kV/m (Pavselj et al. 2005; Corovic et al. 2013). When the electroporation occurs, in certain regions of the plasmatic membrane of the cells the membrane conductance increases rapidly and passes to vary in time according to the intensity of the transmembrane potential. Part of the electric current cross the plasmatic membrane through hydrophilic pores created in the electroporation process. Because of this the static conductivity of the biological tissue becomes a function of time and intensity of the applied electric field. Despite intense experimentation around this phenomenon in recent decades, a general mathematical form that correctly expresses the time dependence of tissue conductivity during electroporation has not yet been obtained. Nevertheless, one way in which a dynamic model

14

A. Ramos and D.O.H. Suzuki

of electroporation may be included in the equivalent circuit is expressed by a first order equation in time as follows: Gp ðtÞ ¼ Gp ðt  ΔtÞ þ

A gðE, tÞΔt L

(31)

where Gp is the increase in static conductance of the tissue due to the electroporation process, A and L are explained in Fig. 1, E is the electric field, and the function g(E,t) describes the dynamics of electroporation from the time rate of change of static conductivity: gðE, tÞ ¼

dσs dt

(32)

Starting with a null value, the electroporation conductance Gp is incremented or decremented at each time step according to the intensity of the electric field and its value is added to the static conductance described by Eq. 29: Gs ¼ Gso þ Gso χ ðT  To Þ þ Gp ðtÞ

(33)

Simulation of the Stimulation Effects with Metal Electrodes One of the most common and important analysis refers to biological tissue stimulation with electrodes in the form of needles to produce electroporation of tumors for treatment purposes. Figure 7 shows the geometry used, and Tables 1 and 2 show numerical values used in this example (The data in Table 1 were obtained from the article by Gabriel et al. 1996). The tissue in this case is modeled as a linear and homogeneous medium. Electrodes in the form of metal needles arranged in two rows each containing three needles are used for applying an electric potential difference with pulsed waveform. Three human tissues are used as substrate in these simulations: skin, skeletal muscle, and liver. Its spectra are shown in Fig. 5. A program in C language was built to perform calculations relating to this simulation. The equations for the electric potential were solved with the GaussSeidel method. The simulations were running on a computer with Intel i3TM processor, 4 GB of RAM, and Windows 7TM operational system. Due to the large amount of volume elements in the discretization mesh (880,000) and the high number of time steps (50,000), each simulation took about 4 h. In each simulation, the applied voltage between the two rows of needles has the form of bipolar and symmetrical trapezoidal wave with amplitude suitable to achieve the electric field required for electroporation in a large region in the space between the needles. The rise and fall times of the pulse wave were adjusted aiming to emphasize the effect of dispersion in the transient behavior of the current through the tissue. Figure 8a, b, and c show the waveforms of the total electric current through the tissue in the first period of the applied voltage. The main purpose of this example is

Computational Approach for Electrical Analysis of Biological Tissue Using. . .

15

Fig. 7 Geometry for simulating the electrical stimulation of biological tissue using the equivalent circuit method

Table 1 Tissue properties (Gabriel et al. 1996)

Parameter ε1 σs (S/m) σα (S/m) τα (s) α σβ (S/m) τβ (s) β σγ (S/m) τγ (s) γ

Skin 43 4  104 1.67  104 1.59  103 0.2 0.167 1.59  106 0.16 0.0311 7.96  108 0.001

Muscle 54 0.2 9.73  102 2.274  103 0.001 3.34  102 3.18  104 0.1 0.175 3.54  107 0.1

Liver 43 2  102 1.67  102 1.59  102 0.05 1.946  102 2.274  105 0.2 0.1 5.305e-7 0.2

to show the importance of correct modeling of the dispersive properties of the materials involved in the field calculation in biological tissues. To this end, simulations were performed with different types of approaches: three Cole bands (αβγ Cole), three Debye bands (αβγ Debye), only alpha and beta Debye bands (αβ Debye), only beta Debye band (β Debye), and tissue without dispersion using either Gs or Gs + Gα + Gβ as the conductance. It makes sense to discard a dispersion band if the waveform of the applied voltage contains very different frequencies from the cutoff frequency of that band. Considering the pulsed waveforms generally used in electroporation experiments, the gamma band which has relaxation times of tens to hundreds of nanoseconds may possibly be ignored, and its dispersion capacitance must be added to the capacitance C1 of the model. Likewise, if the alpha band, with relaxation time of tens of milliseconds, is ignored, their dispersion conductance Gα should be added to the static conductance of the model. The waveforms in Fig. 8 show different aspects that should be highlighted. In the case of skin (Fig. 8a) which has very small static conductivity, there is not an adequate conductance value in a model without dispersion which can acceptably

16

A. Ramos and D.O.H. Suzuki

Table 2 Simulation parameters

Parameter Lx (m) Ly (m) Lz (m) Δx (m) Δy (m) Δz (m) d (m) h (m) w(m) ϕ (m) Δt (s) Ns VF (V) tr (s) tf (s) thigh (s) tlow (s) a

Description Domain thickness Domain width Domain length Mesh parameter Mesh parameter Mesh parameter Distance between needle rowsa Needle length Distance between needlesa Needle diameter Time step Number of steps Pulse amplitude Pulse rise time Pulse fall time Pulse duration in high level Pulse duration in low level

Value 8  103 10  103 11  103 1  104 1  104 1  104 5  103 5  103 2  103 4  104 2  109 5  104 500 20  106 20  106 80  106 80  106

Distance center-to-center between needles

represent the dynamic behavior of the current in the tissue. For liver (Fig. 8c), with static conductivity fifty times higher, the curves for models with and without dispersion are comparable but still have significant differences. Only in the case of skeletal muscle (Fig. 8b), which presents very high static conductivity, is obtained a suitable modeling in the simulation conditions with a model without dispersion in which conductance is the sum Gs + Gα + Gβ. Once the spectrum of the applied voltage concentrates 99 % of the intensity in the frequency range from 5 kHz to 100 kHz while the cutoff frequencies of alpha and gamma bands in all three analyzed tissues is below 1 kHz and above 1 MHz, respectively, these bands do not influence substantially the dynamic response of the current. Consequently, in all three cases using only beta dispersion yields a good modeling. Figure 9 refers to another simulation involving the same previous stimulation scheme but now including a small spherical volume between the needles. This volume aims to model a skin tumor with 3 mm diameter located at one millimeter from the surface. The images 9a and 9c refer to the skin model with alpha and beta dispersion while the images 9b and 9d relate to the model without dispersion and using the conductance Gs. It is known that tumors generally have higher conductivity and permittivity than normal tissues. In this case, it was considered that the spherical volume presents five times more conductivity and permittivity than normal skin. The graphs show the distributions of the total electric field in the x and y planes which pass through the center of the tumor at time t = 105 s which corresponds to the moment when the voltage pulse reaches its maximum value of 500 V. It is clear that the field distributions are significantly different between the two models.

Computational Approach for Electrical Analysis of Biological Tissue Using. . .

17

Fig. 8 Total current waveforms obtained in the first cycle of the applied voltage. (a) skin; (b) skeletal muscle (not specified orientation); (c) liver. Each chart contains six curves

18

A. Ramos and D.O.H. Suzuki

Especially inside the tumor where the model with dispersion results in electric field higher than 35 kV/m (minimum value around 37 kV/m) and the model without dispersion results in a large region with lower field (minimum value around 32 kV/m). It was observed that in the end of the positive pulse (t = 9  105 s), the difference between the results obtained with the two models become almost unnoticeable, which shows that the dielectric dispersion affects only the transient electric field distribution. The images show that a relatively large volume of tissue surrounding the needles reaches very intense electric field values (E > 95 kV/m) that can result in irreversible electroporation. It is observed that the tumor is subject to much lower electric field values, but based on the results obtained from the model with dispersion, there should be reversible electroporation inside. Moreover, the results obtained with no dispersion model suggest that would be required to increase the amplitude of the voltage pulses applied between the needles in order to achieve electroporation across

Fig. 9 Electric field distribution in the skin with tumor for the same stimulation scheme of Fig. 7 in the plane x (a and b) and y (c and d) passing through the center of the tumor: (a) and (c) model with alpha and beta dispersion; (b) and (d) model without dispersion using the conductance Gs

ä Fig. 8 (continued) corresponding to the six models used in the simulations. In the graph (a) the curve for Gs + Gα + Gβ is divided by eight in order to be visually comparable to the other curves

Computational Approach for Electrical Analysis of Biological Tissue Using. . .

19

the tumor volume. Thus, it is emphasized that the use of computer simulation with suitable electrical models can predict the effectiveness of the electrical stimulation scheme of a biological tissue for electroporation purposes. These conclusions are corroborated by a similar but more detailed study by Miklavcic and colleagues (Miklavcic et al. 2006). However, one aspect of the problem that was not considered in this analysis due to the lack of sufficient information is the change in conductivity of the tissue due to the electroporation. Once the applied electric field exceeds the threshold for reversible electroporation, the conductivity of the tissue increases intensively and rapidly in a microsecond time scale. Thus, it is expected that the dielectric dispersion has great influence on the intensity and extent of tissue electroporation and the computational models used in the electrical analysis in this case should take this into account. In addition, some studies have shown that due to electroporation, current and electric field distributions in biological tissues are significantly different from those obtained in simulations in which the materials are considered linear regardless of the applied electric field strength (Ramos 2005; Corovic et al. 2013). This makes the use of numerical methods that describe as realistically as possible the dielectric dispersion and the dynamics of electroporation in biological tissues, a basic need for reliable planning therapies based on electroporation.

Conclusion The computational approach for electrical analysis using the equivalent circuit model is a relatively simple technique to implement and allows calculating the spatial distribution of the electric field and electric current in the time domain in materials in which the dielectric dispersion can be adequately described by one or more dispersion Debye bands. When applied to the biological tissues, equivalent circuit method has significant advantages over other numerical techniques, due to the simplicity in construction of realistic models from dispersion spectra of the materials involved and applying the boundary conditions. The example presented in this chapter shows that the dielectric dispersion significantly influences both the temporal response of the electric current as the electric field distribution in the tissue. The modeling based on the equivalent electric circuit can be used to study the electroporation in biological tissues, since the dynamic conductivity changes resulting from the electroporation process is correctly modeled and included as variations of the conductance Gs of the equivalent circuit.

Cross‐References ▶ Application of Numerical Simulation Techniques for Modeling Pulsed Electric Field Processing ▶ Electric Field Distribution and Electroporation Threshold

20

A. Ramos and D.O.H. Suzuki

▶ Electric Field Distribution Modeling in Tissue Considering Tissue Conductivity Increase Due to Electroporation ▶ Electric Pulse Parameters Affecting Electroporation Treatment Outcome ▶ Electroporation and Electropermeabilization ▶ Impedance Measurement as Electroporation Measure ▶ Temperature Increase and Thermal Effects Due to Electroporation in Tissues ▶ Tissue Ablation by Irreversible Electroporation ▶ Transmembrane Voltage Induced by Applied Electric Fields ▶ Treatment Planning for Electrochemotherapy and Irreversible Electroporation of Deep-Seated Tumors

References Bottcher CJF, Bordewijk P (1978) Theory of electric polarization, 2nd edn. Elsevier, Amsterdam Chen JY, Gandhi OP (1992) Numerical simulation of annular-phased arrays of dipoles for hyperthermia of deep-seated tumors. IEEE Trans Biomed Eng 39:209–216 Cole KS, Cole RH (1941) Dispersion and absorption in dielectrics. I. Alternating current characteristics. J Chem Phys 9:341–352 Corovic S, Lackovic I, Sustaric P, Sustar T, Rodic T, Miklavcic D (2013) Modeling of electric field distribution in tissues during electroporation. Biomed Eng Online 12:16 Daniels C, Rubinsky B (2009) Electrical field and temperature model of nonthermal irreversible electroporation in heterogeneous tissues. J Biomech Eng 131:071006–1–12 Foster KR, Schwan HP (1995) Dielectric properties of tissues. In: Polk C, Postow E (eds) Handbook of biological effects of electromagnetic fields, 2nd edn. CRC Press, Boca Raton, pp 27–102 Gabriel S, Lau RW, Gabriel C (1996) The dielectric properties of biological tissues: III. Parametric models for the dielectric spectrum of tissues. Phys Med Biol 41:2271–2293 Kos B, Zupanic A, Kotnik T, Snoj M, Sersa G, Miklavcic D (2010) Robustness of treatment planning for electrochemotherapy of deep-seated tumors. J Membr Biol 236:147–153 Lacković I, Magjarević I, Miklavćić D (2009) Three-dimensional finite-element analysis of joule heating in electrochemotherapy and in vivo gene electrotransfer. IEEE Trans Dielectr Electr Insul 16:1338–1347 Miklavcic D, Corovic S, Pucihar G, Pavselj N (2006) Importance of tumour coverage by sufficiently high local electric field for effective electrochemotherapy. EJC Suppl 4:45–51 Miklavcic D, Snoj M, Zupanic A, Kos B, Cemazar M, Kropivnik M, Bracko M, Pecnik T, Gadzijev E, Sersa G (2010) Towards treatment planning and treatment of deep-seated solid tumors by electrochemotherapy. Biomed Eng Online 9:10 Nadeem M, Thorlin T, Gandhi OP, Person M (2003) Computation of electric and magnetic stimulation in human head using the 3-D impedance method. IEEE Trans Biomed Eng 50:900–907 Neal RE, Garcia PA, Robertson JL, Davalos RV (2012) Experimental characterization and numerical modeling of tissue electrical conductivity during pulsed electric fields for irreversible electroporation treatment planning. IEEE Trans Biomed Eng 59:1076–1085 Pavselj N, Bregar Z, Cukjati D, Batiuskaite D, Mir LM, Miklavcic D (2005) The course of tissue permeabilization studied on a mathematical model of a subcutaneous tumor in small animals. IEEE Trans Biomed Eng 52:1373–1381

Computational Approach for Electrical Analysis of Biological Tissue Using. . .

21

Ramos A (2005) Effect of the electroporation in the field calculation in biological tissues. Artif Organs 29(6):510–513 Ramos A (2010) A new approach to the solution of electromagnetic problems with the impedance method. Math Comput Simul 81:860–874 Ramos A, Raizer A, Marques LB (2003) A new computational approach for electrical analysis of biological tissues. Bioelectrochemistry 59:73–84 Schwarz G (1962) A theory of the low frequency dielectric dispersion of colloidal particles in electrolyte solution. J Phys Chem 66:2636–2642 Zhao Y, Tang L, Rennaker R, Hutchens C, Ibrahim TS (2013) Studies in RF power communication, SAR, and temperature elevation in wireless implantable neural interfaces. PLoS One 8, e77759 Županič A, Čorović S, Miklavčič D, Pavlin M (2010) Numerical optimization of gene electrotransfer into muscle tissue. Biomed Eng Online 9:66

Mathematical Models Describing Cell Death Due to Electroporation Janja Dermol and Damijan Miklavčič

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microbial Inactivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroporation-Based Medical Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Models Used to Describe Cell Death in Suspension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The First-Order Kinetics Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Hülsheger Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Peleg-Fermi Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Weibull Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Logistic Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Adapted/Modified Gompertz Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Geeraerd Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Quadratic Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Peleg-Penchina Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Models Used to Describe Cell Death In Vitro on Attached Cells and Cells in 3D Tissue-Like Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Models of Cell Death Used in the Treatment Chamber and In Vivo . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 3 6 7 8 9 11 11 12 12 12 13 15 15 17 18 19

Abstract

Various models have been developed to describe microbial inactivation by pulsed electric field treatment, and they have just recently been used for describing eukaryotic cell death due to irreversible electroporation. In microbial inactivation, the mathematical models of cell death enable the adaptation of the pulse parameters to achieve sufficient microbial reduction at the lowest energy input while

J. Dermol (*) • D. Miklavčič Faculty of Electrical Engineering, University of Ljubljana, Ljubljana, Slovenia e-mail: [email protected]; [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_13-1

1

2

J. Dermol and D. Miklavčič

preserving flavor and sensitive compounds in the food. For precise prediction, the geometry of the treatment chamber, the fluid flow, the temperature, and the electric field distribution should also be taken into account. In electroporationbased medical treatments, currently, a deterministic critical value of electric field is used to delineate between the destroyed and the unaffected tissue. Consequently, tumor cells which have higher electroporation threshold than the experimentally determined may remain viable and cause incomplete tumor elimination. On the contrary, the more sensitive surrounding tissue could be damaged. Mathematical models of cell death help to achieve sufficient cell death while minimizing the damage to the surrounding vital structures. In this chapter, different models are described which were already used for describing microbial inactivation in liquid foods or eukaryotic cell death (the first order, the Hülsheger, the Peleg-Fermi, the Weibull, the logistic, the Adapted/Modified Gompertz, the Geeraerd, the quadratic, the Peleg-Penchina model). The cell death models have already been used for predicting survival in realistic setups like the treatment chamber in microbial inactivation and different electrode geometries and tissues in irreversible electroporation treatments. In conclusion, cell death models are useful in predicting the treatment outcome. Unfortunately, since the mechanisms of cell death due to electroporation are not yet fully elucidated, the models are empiric. There is no direct connection between the parameters of the models and the biological/electrical parameters. Thus, it is unclear which model is the most appropriate to use. The models have to be optimized for each specific cell type and electric pulses separately. The transferability from the in vitro to the in vivo level is questionable. Keywords

Microbial inactivation • Predictive models • Numerical modeling • Treatment planning • Food pasteurization

Introduction Mathematical modeling is becoming indispensable in the field of life sciences. It enables description and prediction of a response of a biological system knowing the excitation and other parameters, and gives insights into the mechanisms of phenomena. The number of biological experiments can be decreased which decreases the time and costs needed to obtain results. In the field of electroporation, models exist on different levels – molecules, lipid bilayers, cells, and tissues. Cell death or inactivation models are used for the description of microbial inactivation and recently also eukaryotic cell death in irreversible electroporation as nonthermal soft tissue ablation. In this chapter, the statistical models of cell death are described on the level of cells and tissues. Currently, there is no complete explanation of cell death due to electroporation (“▶ Cell Death Due to Electroporation”), neither for prokaryotic nor eukaryotic

Mathematical Models Describing Cell Death Due to Electroporation

3

cells. The cell death is believed to be caused by the following mechanisms. First, high electric pulses can cause membrane destruction and necrotic cell death. Second, there is the apoptotic cell death, reasons for which are not yet clear. Third, in vivo, last of the clonogenic cells are destroyed by the immune system (“▶ Immunological Response During Electroporation”). Thus, the currently used cell death models are empirical and were developed on the basis of the best fit to the experimental results.

Microbial Inactivation In liquid food pasteurization by pulsed electric fields (PEF), short high-voltage electric pulses are applied to decrease the number of microbes and quality-degrading enzymes, and extending the shelf-life while retaining flavor and nutrients (Lelieveld et al. 2007; Sun 2014). In the literature on cell death models, the expressions bacterial and microbial cell death are both used. The microbe is a short expression for microorganisms not visible with a naked eye which includes besides bacteria also protozoa, fungi, and viruses. Thus, bacteria are a subgroup of microorganisms. Since cell death models have been used for describing cell death of various microorganisms (i.e., bacteria, yeast), expression microbial cell death models is used through this chapter. Mathematical modeling of microbial cell death (“▶ Modeling Microbial Inactivation by Pulsed Electric Field”) is necessary to adapt the pulse parameters and achieve sufficient decrease of microorganisms at a minimum energy input while preserving the sensitive compounds in the food (Huang et al. 2012). In the food industry, at least five log reductions of bacteria are required for food pasteurization. Applying too gentle pulses may cause food spoilage and health-related problems while applying too severe pulses may cause Joule heating and decrease the quality of food. The cell death models have been used on a variety of liquid food (e.g., fruit juices, milk, water, liquid egg yolks) or models of such liquids. When predicting the microbial inactivation by the PEF treatment, the geometry of the treatment chamber is modeled, and the electric field, the temperature distribution, and fluid flow are (numerically) calculated (Gerlach et al. 2008). Microbial cell death prediction can be included using the cell death models as a function of the electric field, treatment time, the number of the applied pulses, and/or the pulse repetition frequency (Huang et al. 2013). The inactivation of the quality-degrading enzymes and the degradation of the health-related compounds can also be modeled with similar models as the inactivation of microorganisms due to their similar kinetics.

Electroporation-Based Medical Treatments When treating eukaryotic cells with electroporation, currently, the three main applications are the electrochemotherapy, the irreversible electroporation (“▶ Tissue Ablation by Irreversible Electroporation”), and the gene electrotransfer for gene therapy and DNA vaccination (Yarmush et al. 2014). When treating tumors with electrochemotherapy, pulses of standard parameters and fixed electrode

4

J. Dermol and D. Miklavčič

configurations can be used. However, if the tumor is larger than 2 cm and/or has an irregular shape, a variable electrode configuration should be used. For the irreversible electroporation treatments (Jiang et al. 2015), protocols are already set up in the commercially available electroporators, but the ablated volumes do not always correlate to the predicted ones (Bhutiani et al. 2016). In the electrochemotherapy, irreversible electroporation, and also in gene electrotransfer treatments, the need for patient-specific treatment planning (“▶ Treatment Planning for Electrochemotherapy and Irreversible Electroporation of Deep-Seated Tumors”) is on the rise. The number, the geometry of the electrodes, and the parameters of the electric pulses must be optimized considering the specifics of the cause and the treatment (Županič et al. 2012). In electrochemotherapy, all tumor cells must be reversibly permeabilized, while the surrounding tissue, especially critical structures, should be undamaged. In irreversible electroporation, the tumor must be irreversibly electroporated without significant thermal damage, while the surrounding tissue, especially critical structures like nerves and vessels, should remain largely undamaged. In gene electrotransfer, cells must be reversibly permeabilized, and their viability must remain high. Currently, an experimentally determined critical electric field is used to predict and delineate between the alive and the irreversibly electroporated tissue – the response is modeled by a step function (Fig. 1). The assumption of a step function is, however, too simplistic even for cell suspensions of the same cell line (Canatella et al. 2001; Dermol and Miklavčič 2015) since cells among other differ in size and position in a cell cycle. In tissues, especially in tumors, this inhomogeneity is amplified since cells additionally differ in shape and tissues are comprised of different cell types. In a tumor, in addition to the tumor cells, stromal cells are present in the microenvironment. Therefore, in reality, the transition from alive to irreversibly electroporated state is continuous and distributed over a range of electric field values. The width of this range, as well as the threshold voltage, also depend on the number and length of the applied pulses (Canatella et al. 2001; Garcia et al. 2014). When predicting the survival with the fixed threshold, the cells which have higher critical threshold than the experimentally determined can survive and cause incomplete tumor elimination or its later recurrence. On the other hand, applying more severe electric pulses than necessary can cause tissue necrosis, excessive Joule heating, and thermal damage. By using cell death models to predict the treated volume, the efficiency of the electroporation-based medical treatments and therapies can be increased. Cell death models allow interpolation and predict survival at parameters which were not experimentally determined. The last few of the clonogenic cells are eradicated by the immune system (“▶ Immune Response After Electroporation and Electrochemotherapy”) (Yarmush et al. 2014) which should be considered when determining the tolerated percentage of the survived cells in electroporation-based treatments. A complete eradication of all tumor cells may not be needed. In electrochemotherapy and gene electrotransfer, using cell death models is not enough to correctly predict the treatment outcome. In the electrochemotherapy treatments, the application of electric pulses increases the permeability of the cell

Mathematical Models Describing Cell Death Due to Electroporation

5

Fig. 1 The comparison of the currently used and the proposed way of modeling cell death. The red line shows the currently used step response where below a critical electric field all cells are regarded alive and above all dead. The gray and dark lines show the suggested cell death models which predict a gradual decrease in cell survival. Black circles are the experimentally determined values. It is clear that the experimental values are better described with different cell death models than with a step response

membrane to the chemotherapeutics. Two commonly used chemotherapeutics are cisplatin and bleomycin. Cell death is mostly caused by the cytotoxic effects of chemotherapeutics, although, in the vicinity of the electrodes, cells can be irreversibly electroporated. To correctly predict the affected regions, a model of permeability to cytotoxic drugs and their transport across the membrane as a function of electric pulses should be included in the treatment planning of electrochemotherapy (Dermol and Miklavčič 2014). Namely, when the number of the molecules of chemotherapeutic entering each cell can be calculated, it can be predicted whether the cell will die or not. In the gene electrotherapy, electric pulses increase the permeability of the cell membrane to the DNA. The exact mechanisms of gene electrotherapy are complex and also depend on genes. The presumed steps in the process are electropermeabilization of the cell membrane, electrophoretic migration of the DNA towards membrane, DNA/membrane interaction, DNA translocation across the membrane, intracellular migration of DNA through the cytoplasm, DNA passage through the nuclear envelope, and gene expression (Rosazza et al. 2016). To correctly predict where the transfection will occur, each step should be modeled.

6

J. Dermol and D. Miklavčič

Models Used to Describe Cell Death in Suspension Ideally, cell death model would include all the parameters, important for cell death. A problem arises since the exact mechanisms of cell death due to electroporation are not completely known. Cell death depends on many parameters, among others on the electric field; the duration and the number of the applied pulses; the pulse repetition frequency; the properties of the cells; their size, shape, orientation in the electric field; their concentration in the treated medium; the properties of the treated medium or food; pH; temperature; configuration of the electrodes or the treatment chamber; and the concentration of the chemotherapeutics (in electrochemotherapy). The most important parameters for cell death should be identified and included in the model. It should also be determined which parameters can be excluded to simplify the models and the fitting. In summary, the cell death model due to electroporation should describe the experimental results well, include all the important electrical and biological parameters, have high predictive power, and a minimum number of parameters to be optimized. Various models have been developed and used to describe microbial cell death due to pulsed electric field (PEF) treatment in liquid food pasteurization (Álvarez et al. 2003; Peleg 2006; San Martín et al. 2007; Huang et al. 2012). Only recently, interest to develop and use such models has been expressed in tissue ablation due to irreversible electroporation, i.e., IRE treatment. Most of the models used, however, are empirical, and their parameters have no physical meaning and even have no parameters that could be explicitly linked to treatment parameters such as electric field, pulse duration, the number of pulses, and pulse shape. There is also no solid link demonstrated between cell membrane permeabilization and cell death. In microbial cell death models, the independent variable is usually the treatment time which is easily determined in the case of thermal, irradiation, high pressure, ultrasound, or other continuous treatment. The treatment time of the PEF treatment is more difficult to determine since multiple short pulses are applied at different pulse repetition frequencies. Nevertheless, the treatment time (t) is usually calculated by multiplying the number of pulses with the duration of one pulse (t ¼ NT ), where N denotes the number of the pulses and T the duration of one pulse. This way of calculating the treatment time assumes that only the product of N and T affects the survival but not the duration or the number of the pulses by itself. The effect of the pulse repetition frequency is also neglected. In microbial cell death models, the electric field, the number of pulses, or their repetition frequency were used as independent variables. With the exception of the quadratic model, however, no model so far included more than two independent variables. The electric field is the most important parameter in electroporation-based treatments (Miklavčič et al. 2006). The survival curves as a function of the electric field are usually composed of three parts: a shoulder (upper asymptote) which is then followed by a steep decline in cell number and eventually, a tail (lower asymptote) is reached. The tail can be a caused by either a resistant subpopulation of cells or the reached limit of the survival assay. Several different curves can describe the survival

Mathematical Models Describing Cell Death Due to Electroporation

7

data. When presenting the survival curves and the cell death models, they should be plotted on a semilogarithmic scale. There, deviations and small percentages of survival are more easily noticed. The quality of the fit should be evaluated using various statistical measures: the goodness-of-fit (R2), the root-mean-square error (RMSE), and/or accuracy parameter (Af). R2 is a statistical measure of how close the regression line is to the experimental data; RMSE is a measure of the average deviation between the observed and predicted data; and Af measures the accuracy of the estimates obtained by the models. Other criteria for evaluating the suitability of a model are the number of the models’ parameters (which should be low), the inclusion of more independent variables, and a high predictive power. In continuation, models which were already used for describing microbial or eukaryotic cell death are described.

The First-Order Kinetics Model The first-order kinetics model was developed almost a century ago (Huang et al. 2012). It derives from the assumption that all cells in a population are equally sensitive to the treatment. When used in PEF treatments, it describes cell death as a function of the treatment time: SðtÞ ¼ expðktÞ;

(1)

where S denotes the survival, t the treatment time, and k is a first-order parameter, i.e., the speed of the decrease. The model predicts that the viability decreases immediately – there is no shoulder at short treatment time. The first-order kinetic model did not provide a good fit to electroporation treatment of eukaryotic cells (Dermol and Miklavčič 2015). It has also been observed that it does not describe all microbial inactivation data well. It was stated that the model describes the data sufficiently well only when the data is sampled too scarcely (Peleg 2006). The deviations from the first-order kinetics model could among others be explained by the statistical distribution of cell radii (Lebovka and Vorobiev 2004). The first-order parameter k was found to be temperature dependent which could be modeled in thermal treatments as well as in PEF treatments. In PEF treatments, the temperature of the sample increases due to the Joule heating and consequently the electrical conductivity of the sample is also increased. The parameter k could be modeled by Arrhenius equation as:   EA k ¼ kt exp  RT

(2)

Where kt is the rate constant at the reference temperature, EA the activation energy, R is the gas constant, and T is the sample temperature in Kelvins. Traditionally, the sensitivity of microorganisms to static treatments is described using the decimal reduction time (Dt). It denotes the treatment time needed to obtain

8

J. Dermol and D. Miklavčič

one decimal reduction of the cell population, i.e., 10 % of the initial population survives. It is defined as Dt ¼ 2:303 k since the first-order kinetics model is based on natural and the decimal reduction time on the decimal logarithm. If the treated cell population consists of two populations of which each has its first-order dynamics, the biphasic model can be used: SðtÞ ¼ f ek1 t þ ð1  f Þek2 t ;

(3)

where f denotes the proportion of the first subpopulation in the whole population, and k1 and k2 are the first order parameters of the first and the second subpopulation, respectively. In case there are more subpopulations, the decrease of each subpopulation can be modeled with its own exponential factor which is then added to the (Eq. 3).

The Hülsheger Model Hülsheger studied the effect of electric field on the inactivation of E. coli (Huang et al. 2012). He derived an empirical formula which has two independent variables – the treatment time (t) and the electric field (E). As a function of the electric field, the survival showed a linear decline when the electric field exceeded a certain critical electric field (Ec) which was modeled as: SðEÞ ¼ expðbE ðE  Ec ÞÞ;

(4)

where S is the survival, bE is the regression coefficient, and Ec is the critical value of electric field below which there will be no inactivation (i.e., the lowest electric field that causes inactivation). When E < Ec, the model predicts the survival above one which is incorrect and the survival must be fixed at one. As a function of the treatment time, the results were modeled as:    t SðtÞ ¼ exp bt ln ; tc

(5)

where S is the survival, bt is the regression coefficient, and tc is the extrapolated critical value of treatment time below which there will be no inactivation (i.e., the shortest treatment time which causes inactivation). For treatment times shorter than the critical treatment time, the model predicts the survival to be more than 1. Thus, also here, for t < tc the survival should be fixed at one. Both models Eqs. 4 and 5 were joined in:  EEk c t Sðt, EÞ ¼ tc

(6)

Mathematical Models Describing Cell Death Due to Electroporation

9

where k is the independent inactivation constant, and other parameters have the same meaning as in Eqs. 4 and 5. The parameters tc, Ec, and k were proposed to be microorganism dependent within a certain range of experimental parameters. Although the Hülsheger model includes two independent variables, it cannot universally describe the experimentally determined cell death. With the Hülsheger model, some microbial cell death results were possible to model, and some were not (Huang et al. 2012), while only one study unsuccessfully fit it to the eukaryotic cell death data (Dermol and Miklavčič 2015).

The Peleg-Fermi Model The Peleg-Fermi model has been successfully fit to various microbial inactivation data (Huang et al. 2012) and was also the first to be used for describing eukaryotic cell death due to electroporation (Golberg and Rubinsky 2010). It derives from the Fermi’s equation which is used to describe the behavior of materials at their glass transition temperature. The Peleg-Fermi model is one of the most promising since it describes the data well and includes two independent variables – electric field (E) and the number of the applied pulses (N ): SðE, N Þ ¼

1  E  Ec ð N Þ 1 þ exp k ðN Þ 

(7)

Ec ðN Þ ¼ Ec0 expðk1 N Þ

(8)

kðN Þ ¼ k0 expðk2 N Þ

(9)

where Ec(N ) is the critical electric field where the survival drops to 50 %, k(N ) is the kinetic constant describing the slope of the curve, Ec0 is the intersection of Ec(N ) with the y-axis, k0 (in the same units as the electric field), k1 and k2 are constants which change depending on the parameters of the pulses and the properties of the cells. The critical electric field depends on the number and length of the applied pulses (Pucihar et al. 2011). For values below the critical electric field, the PelegFermi model predicts survival to be 100 %. The Peleg-Fermi model described various experimental data well, although the k(N ) and Ec(N ) did not always change exponentially as a function of the pulse number (Dermol and Miklavčič 2015; Sharabi et al. 2016). Unfortunately, the pulse length and the pulse repetition frequency are not included in the model, and the dependency on them is included indirectly via the parameters of the model. An example of fitting the Peleg-Fermi model to the eukaryotic cell death due to irreversible electroporation in vitro is shown in Fig. 2.

10

J. Dermol and D. Miklavčič

Fig. 2 The Peleg-Fermi model, fit to eukaryotic cell death data in a wide range of pulse number and electric field values. (a) symbols show the experimental values for 100 μs, 1 Hz repetition frequency at different pulse numbers (8, 30, 50, 70, and 90), lines show the fitted Peleg-Fermi model (Eq. 7). (b) symbols are the optimized values of Ec and k for different pulse numbers, lines show the fitted models for Ec(N ) and k(N ) (Eqs. 8 and 9) (Reprinted from Journal of Membrane Biology, Vol 248/Issue 5, Dermol J, Miklavčič D, Mathematical Models Describing Chinese Hamster Ovary Cell Death Due to Electroporation In Vitro, Pages 865–881, Copyright 2015 Springer Science + Business Media New York)

Mathematical Models Describing Cell Death Due to Electroporation

11

The Weibull Model The Weibull model was originally used to describe the time to failure of electronic devices after stress was imposed on them. A parallel can be made when stress, for example, thermal, high pressure, or PEF treatment, is applied to biological cells. Time to cell death after the stress can be described using the Weibull model (San Martín et al. 2007). The Weibull model was successfully used to describe microbial (Huang et al. 2012) and eukaryotic cell death (Dermol and Miklavčič 2015) using various pulse parameters. Although the Weibull model is very adaptable, no clear connection between the electrical/biological parameters and the optimized parameters of the model was made. In some cases, the scale parameter b was exponentially dependent on the electric field. The model is:   x n  SðxÞ ¼ exp  ; b

(10)

where x denotes the treatment time or the electric field, b is the scale (in the same units as x), and n is the shape parameter. The scale parameter determines the characteristic time or the characteristic electric field at which 63 % of the cells die. With different values of the shape parameter, the shape of the survival curve varies between the convex (n < 1), linear (n = 1), and concave curve (n > 1) in semilogarithmic scale.

The Logistic Model The logistic model can be used for describing distributions with a sharp peak and long tails (Cole et al. 1993). When taking into account that survival of cells before the treatment is 100 %, the logistic model can be written as either Eq. 11 or 12, depending on what is chosen as the independent variable. The equations are:   ω SðEÞ ¼ 10^ 1 þ expð4σ ðτ  EÞω1 Þ   ω SðtÞ ¼ 10^ 1 þ expð4σ ðτ  log10 tÞω1 Þ

(11) (12)

where E denotes the electric field, t the treatment time, ω the common logarithm of the lower asymptote, σ the maximum slope, and τ the position of the maximum slope. Survival curves show the cumulative cell death as a function of the independent variable, i.e., applied electric field or treatment time. Thus, each data point includes cells which died due to the corresponding or less severe parameter. The cell death distribution is obtained as a derivative of the cumulative distribution. Only cell death due to the corresponding parameter is plotted. In electroporation treatments, the distribution of cell death with a sharp peak and long tails is obtained when the

12

J. Dermol and D. Miklavčič

independent variable is either the electric field or the common logarithm of the treatment time. The logistic model was successfully fit to inactivation of microorganisms by PEF treatment (Huang et al. 2012) and to eukaryotic cell death (Dermol and Miklavčič 2015).

The Adapted/Modified Gompertz Model The Gompertz model is usually used for describing the growth of a tumor, but in modified form, it can also describe cell death. Originally, it was used for describing bacterial (Linton et al. 1995) cell death due to thermal treatment, but later it was extended to describing bacterial cell death due to pulsed electric field treatment and eukaryotic (Dermol and Miklavčič 2015) cell death due to electroporation. The model is written as:   
 SðxÞ ¼ exp A exp eðB0 þB1 xÞ  A exp eB0

(13)

where x denotes either the treatment time or the electric field, A is the natural logarithm of the survival in the stationary phase (the natural logarithm of the tail), B0 is the length of the shoulder, and B1 the speed of the increase (when it is positive) or decrease (when it is negative) in cell number.

The Geeraerd Model The Geeraerd model (Geeraerd et al. 2000) describes the exponential decrease of cells including a tail which models the resistant cell subpopulation. Assuming the initial survival is 100 %, the Geeraerd model can be written as: SðtÞ ¼ ð1  N res ÞexpðktÞ þ N res

(14)

Where Nres represents the tail and k is the inactivation rate. Since the model does not also include a shoulder, it cannot be used universally for all cell death results. The Geeraerd model was so far successfully fit to experimental data of microbial inactivation after mild heat treatment and eukaryotic cell death due to electroporation.

The Quadratic Model The quadratic model is currently the only model which can model dependency on two or more independent variables. As the independent variable x, various parameters have been used – the electric field, the treatment time, the pulse repetition

Mathematical Models Describing Cell Death Due to Electroporation

13

frequency, the pH, and the concentration of some compound. The model is written as: Sð x Þ ¼ k 1 þ k 2 x þ k 3 x 2

(15)

where the parameter k1 is the central point of the system, k2 is the coefficient of the linear effect, and k3 of the quadratic effect of the independent variable. It is also possible to fit more parameters at once by joining several quadratic models into one: Sðx1 , x2 Þ ¼ k1 þ k2 x1 þ k3 x1 2 þ k4 x2 þ k5 x2 2 þ k6 x1 x2

(16)

where x1 and x2 denote two independent variables which were marked as the x in the (Eq. 15), k1 is the central point of the system, k2 and k4 represent the linear, k3 and k5 the quadratic, and k6 the interactive effects of the independent variables. If some of the terms are nonsignificant under certain experimental conditions, they can be omitted from the model. By quadratic model fitting, the interaction between different experimental parameters can be determined. The quadratic model was successfully fit to microbial inactivation results as well as to the inactivation of the enzymes and health-related compounds like vitamins (Huang et al. 2012).

The Peleg-Penchina Model The Peleg-Penchina model is also an empirical model but can describe only the convex curves (Peleg and Penchina 2000; Álvarez et al. 2003) in the semilogarithmic scale. It is written as: SðtÞ ¼ 10 ^ ðm lnð1 þ ktÞÞ

(17)

where t is the treatment time of the PEF treatment; m and k are the parameters of the model, which have to be optimized; and ln denotes the natural logarithm. Together with the Eq. 17, the authors introduced a way of describing bacterial survival when the intensity of the lethal agent (e.g., temperature, chemical agent, PEF treatment) varies either between treatments or during one treatment. As long as the mechanisms of cell death due to electroporation (PEF treatment) are not known, it is difficult to decide which model is superior to others, and they can all be regarded as equivalent. An example of fitting several cell death models to eukaryotic and bacterial cell death results is shown in Fig. 3.

14

J. Dermol and D. Miklavčič

Fig. 3 (a) Different cell death models as a function of the electric field (the Peleg-Fermi, the Weibull, the Logistic, the Adapted Gompertz model) were fitted to the experimental data of eukaryotic cell death when 90, 100 μs pulses were applied to the cell suspension. Symbols show the experimental values and lines the models. It can be seen that several models describe the experimental data well Observed values are marked by circles, model 1 by a bold line, model 2 by a dotted line, model 3 by a thin line, and model 4 by a dashed line. (Reprinted from Journal of Membrane Biology, Vol 248/Issue 5, Dermol J, Miklavčič D, Mathematical Models Describing Chinese Hamster Ovary Cell Death Due to Electroporation In Vitro, Pages 865–881, Copyright 2015 Springer Science + Business Media New York 2015). (b) Different cell death models as a function of the treatment time were fitted to the survival of E. coli after pulsed electric field treatment. Model 1 is the biphasic model, model 2 is the sigmoidal equation, model 3 is the Weibull model, and model 4 is the Peleg-Penchina model. A good fit was obtained with all tested models (Reprinted from Innovative Food Science & Emerging Technologies, Vol 4/edition number 4, Alvarez I, Virto R, Raso J, Condon S, Comparing predicting models for the Escherichia coli inactivation by pulsed electric fields, Pages No 195–202, Copyright (2003), with permission from Elsevier)

Mathematical Models Describing Cell Death Due to Electroporation

15

Models Used to Describe Cell Death In Vitro on Attached Cells and Cells in 3D Tissue-Like Structure In liquid food pasteurization, the microorganisms are suspended in the food. However, they can also grow in layers, for example, on surgical meshes. Surgical mesh is thin mesh which supports or reinforces damaged tissue. The cell death models were applied to PEF treatments of medical mesh implants, on which 2D layers of bacteria can form as a biofilm (“▶ Electroporation of Biofilms”) (Khan et al. 2016). The authors successfully described the data using the Weibull model. The cell death models have not yet been applied to the attached or 3D eukaryotic cell models. The optimization of the models in vitro would need to be done anew since the critical electric fields are different for the attached than for the suspended cells due to a different shape, density, and connections between cells (Pucihar et al. 2006; Towhidi et al. 2008).

Models of Cell Death Used in the Treatment Chamber and In Vivo Microorganisms grow in suspensions and 2D layers. Thus, there are no reports on using the mathematical models to describe PEF inactivation of microorganisms in vivo. There is, however, an example of modeling the treatment chamber (“▶ Optimization of Pulsed Electric Field Treatment Chamber”) with included suspension of bacteria. Huang et al. predicted the inactivation of bacteria in watermelon juice by first building a 2D numerical model of the treatment chamber and using the Hülsheger model to describe the inactivation (Huang et al. 2013). Authors obtained very good predictions. Cells respond to the electric field to which they are exposed. In treatment chambers (as well as in tissue when performing the IRE treatment), the electric field can be highly inhomogeneous. In future, to obtain the optimal conditions for microbial inactivation, first optimization of the treatment chambers’ geometry and the parameters of electric pulses should be done. There are just a few reports of using cell death models to describe and predict the extent of eukaryotic cell death due to electroporation in vivo. Golberg and Rubinsky were first to suggest using statistical cell death models in vivo (Golberg and Rubinsky 2010). They fit the Peleg-Fermi model to the results of an in vitro study of electroporation of prostate cell line. Then, the fitted Peleg-Fermi model was theoretically applied to the 2D case of tissue electroporation using needle electrodes. The spatial distribution of probability of cell death the authors obtained is shown in Fig. 4. It was shown that there exists an area where the cell death probability ranges from 0 % to 100 % and the deterministic critical electric field is not an optimal choice for predicting the electroporation-based medical treatment outcome. A theoretical study of commercially available bipolar electrodes used in irreversible electroporation treatments was done by Garcia et al. (2014). The authors determined the extent of the cell death caused by the heating and by the electroporation when standard IRE pulses were applied. The thermal damage was evaluated using the Arrhenius integral and the electrical damage using the Peleg-Fermi model.

16

J. Dermol and D. Miklavčič

Fig. 4 The Peleg-Fermi model was fit to the 2D model of prostate tissue treated by irreversible electroporation using needle electrodes. The legend on the right shows the probability of cell death. Around the electrodes exists an area where the probability of cell death is between 0 % and 100 %, thus the deterministic critical electric field is not sufficient for predicting cell death in vivo. (a–c) show the predicted cell death when a different number of 100 μs pulses of the same voltage (1.5 arbitrary unit) are applied: (a) 10 pulses, (b) 50 pulses, (c) 100 pulses (Reprinted from BioMedical Engineering OnLine, Vol 9, Golberg A, Rubinsky B, A statistical model for multidimensional irreversible electroporation cell death in tissue, Page No 13, Copyright 2010 Golberg and Rubinsky; licensee BioMed Central Ltd.)

The electrical damage due to electroporation around the bipolar electrodes is shown in Fig. 5. Different combinations of pulse number and electric field lead to different probabilities of cell death as shown in Fig. 6. A study in vivo was performed by Sharabi et al., who analyzed the brain electroporation using the Peleg-Fermi model to describe cell death as well as the disruption of the blood–brain barrier (Sharabi et al. 2016). The authors concluded that the Peleg-Fermi model could be successfully used to describe the electroporation of rat brain, although they found that the Ec(N ) was better described using power than an exponential function, especially in the range of a high number of pulses (more than 90). As shown in (Eq. 8), Ec(N ) in the Peleg-Fermi model can be expressed as Ec ðN Þ ¼ Ec0 expðk1 N Þ . Sharabi et al. obtained better fit using the equation: Ec ðN Þ ¼ Ec0 N k1 ;

(18)

where Ec0 denotes the critical electric field, N the number of the applied pulses, and k1 is a constant. In electroporation-based medical treatments, the desired use of the models is in tissues in vivo. It is questionable to what extent the in vitro optimized models can be used in vivo. The thresholds for cell death due to electroporation in vivo (Jiang et al. 2015) seem to be different than in vitro (Dermol and Miklavčič 2015). In vivo, the

Mathematical Models Describing Cell Death Due to Electroporation

17

Fig. 5 The probability of cell death due to electroporation around the commercially available bipolar electrodes at different pulse numbers (a – 30, b – 50, c – 70, and d – 90) when applying 100 μs pulses of 3000 V at 1 Hz. The transition zone where the cell death decreases from 100 % to 0 % becomes sharper with increasing pulse number (Reprinted from PLoS ONE, Vol 9/Issue 8, Garcia PA, Davalos RV, Miklavčič D, A Numerical Investigation of the Electric and Thermal Cell Kill Distributions in Electroporation-Based Therapies in Tissue, Pages No e103083, Copyright 2014 Garcia et al.)

cells are connected, they are of different size and shape, their density varies, there are different types of cells in each tissue, and the immune system is present. It seems that the cell death models optimized in vitro may not be directly transferable in vivo but would need to be optimized separately.

18

J. Dermol and D. Miklavčič

Fig. 6 The probability of cell death due to electroporation as a function of pulse number and electric field. The cell death is predicted by the Peleg-Fermi model optimized for the prostate cancer cell death. The contours show the predicted cell death (denoted with numbers in bold) using various combinations of electric field and pulse number. To achieve 99.9 % cell death a certain minimal electric field and a number of pulses should be applied (Reprinted from PLoS ONE, Vol 9/Issue 8, Garcia PA, Davalos RV, Miklavčič D, A Numerical Investigation of the Electric and Thermal Cell Kill Distributions in Electroporation-Based Therapies in Tissue, Pages No e103083, Copyright 2014 Garcia et al.)

Conclusions Using cell death models in treatment planning could increase the efficiency of PEF microbial inactivation and electroporation-based medical treatments. However, currently, it is unclear which model is the most appropriate since they are all empirical, and mostly, there is no clear relation between the electrical/biological parameters and parameters of the cell death models. Although the cell death models describe data well, they must be optimized for each cell type/tissue and different electric pulses separately. A direct translation of in vitro optimized models to an in vivo environment is questionable. In future, cell death models should be based on mechanisms of cell death due to electroporation, include all and only the relevant treatment parameters, and describe and predict the cell death accurately. Acknowledgments This study was supported by the Slovenian Research Agency (ARRS) and conducted within the scope of the Electroporation in Biology and Medicine European Associated Laboratory (LEA-EBAM).

Mathematical Models Describing Cell Death Due to Electroporation

19

Cross-References ▶ Cell Death Due to Electroporation ▶ Electroporation of Biofilms ▶ Immune Response After Electroporation and Electrochemotherapy ▶ Immunological Response During Electroporation ▶ Modeling Microbial Inactivation by Pulsed Electric Field ▶ Optimization of Pulsed Electric Field Treatment Chamber ▶ Tissue Ablation by Irreversible Electroporation ▶ Treatment Planning for Electrochemotherapy and Irreversible Electroporation of Deep-Seated Tumors

References Álvarez I, Virto R, Raso J, Condón S (2003) Comparing predicting models for the Escherichia coli inactivation by pulsed electric fields. Innovat Food Sci Emerg Technol 4:195–202. doi:10.1016/ S1466-8564(03)00004-3 Bhutiani N, Doughtie CA, Martin RCG (2016) Ultrasound validation of mathematically modeled irreversible electroporation ablation areas. Surgery 159:1032–1040. doi:10.1016/j. surg.2015.10.030 Canatella PJ, Karr JF, Petros JA, Prausnitz MR (2001) Quantitative study of electroporationmediated molecular uptake and cell viability. Biophys J 80:755–764. doi:10.1016/S0006-3495 (01)76055-9 Cole MB, Davies KW, Munro G, Holyoak CD, Kilsby DC (1993) A vitalistic model to describe the thermal inactivation of Listeria monocytogenes. J Ind Microbiol 12:232–239. doi:10.1007/ BF01584195 Dermol J, Miklavčič D (2014) Predicting electroporation of cells in an inhomogeneous electric field based on mathematical modeling and experimental CHO-cell permeabilization to propidium iodide determination. Bioelectrochemistry 100:52–61. doi:10.1016/j.bioelechem.2014.03.011 Dermol J, Miklavčič D (2015) Mathematical models describing Chinese hamster ovary cell death due to electroporation in vitro. J Membr Biol 248:865–881. doi:10.1007/s00232-015-9825-6 Garcia PA, Davalos RV, Miklavčič D (2014) A numerical investigation of the electric and thermal cell kill distributions in electroporation-based therapies in tissue. PLoS One 9, e103083. doi:10.1371/journal.pone.0103083 Geeraerd AH, Herremans CH, Van Impe JF (2000) Structural model requirements to describe microbial inactivation during a mild heat treatment. Int J Food Microbiol 59:185–209. doi:10.1016/S0168-1605(00)00362-7 Gerlach D, Alleborn N, Baars A, Delgado A, Moritz J, Knorr D (2008) Numerical simulations of pulsed electric fields for food preservation: a review. Innovat Food Sci Emerg Technol 9:408–417. doi:10.1016/j.ifset.2008.02.001 Golberg A, Rubinsky B (2010) A statistical model for multidimensional irreversible electroporation cell death in tissue. Biomed Eng OnLine 9:13. doi:10.1186/1475-925X-9-13 Huang K, Tian H, Gai L, Wang J (2012) A review of kinetic models for inactivating microorganisms and enzymes by pulsed electric field processing. J Food Eng 111:191–207. doi:10.1016/j. jfoodeng.2012.02.007 Huang K, Yu L, Ling G, Wang J (2013) Coupled simulations in colinear and coaxial continuous pulsed electric field treatment chambers. Trans ASABE 56:1473–1484. doi:10.13031/ trans.56.9167

20

J. Dermol and D. Miklavčič

Jiang C, Davalos RV, Bischof JC (2015) A review of basic to clinical studies of irreversible electroporation therapy. IEEE Trans Biomed Eng 62:4–20. doi:10.1109/TBME.2014.2367543 Khan SI, Blumrosen G, Vecchio D, Golberg A, McCormack MC, Yarmush ML, Hamblin MR, Austen WG (2016) Eradication of multidrug-resistant pseudomonas biofilm with pulsed electric fields: eradication of multidrug-resistant pseudomonas. Biotechnol Bioeng 113:643–650. doi:10.1002/bit.25818 Lebovka NI, Vorobiev E (2004) On the origin of the deviation from the first-order kinetics in inactivation of microbial cells by pulsed electric fields. Int J Food Microbiol 91:83–89. doi:10.1016/S0168-1605(03)00321-0 Lelieveld H, Notermans S, Haan SWHD (2007) Food preservation by pulsed electric fields: from research to application, 1st edn. Woodhead Publishing, Cambridge Linton RH, Carter WH, Pierson MD, Hackney CR (1995) Use of a modified Gompertz equation to model nonlinear survival curves for Listeria monocytogenes Scott A. J Food Prot 58:946–954 Miklavčič D, Čorović S, Pucihar G, Pavšelj N (2006) Importance of tumour coverage by sufficiently high local electric field for effective electrochemotherapy. Eur J Cancer Suppl 4:45–51. doi:10.1016/j.ejcsup.2006.08.006 Peleg M (2006) Advanced quantitative microbiology for foods and biosystems: models for predicting growth and inactivation. Taylor & Francis, Boca Raton Peleg M, Penchina CM (2000) Modeling microbial survival during exposure to a lethal agent with varying intensity. Crit Rev Food Sci Nutr 40:159–172. doi:10.1080/10408690091189301 Pucihar G, Kotnik T, Valič B, Miklavčič D (2006) Numerical determination of transmembrane voltage induced on irregularly shaped cells. Ann Biomed Eng 34:642–652. doi:10.1007/ s10439-005-9076-2 Pucihar G, Krmelj J, Reberšek M, Napotnik T, Miklavčič D (2011) Equivalent pulse parameters for electroporation. IEEE Trans Biomed Eng 58:3279–3288. doi:10.1109/TBME.2011.2167232 Rosazza C, Haberl Meglič S, Zumbusch A, Rols M-P, Miklavčič D (2016) Gene electrotransfer: a mechanistic perspective. Curr Gene Ther 16:98–129. doi:10.2174/ 1566523216666160331130040 San Martín MF, Sepúlveda DR, Altunakar B, Góngora-Nieto MM, Swanson BG, Barbosa-Cánovas GV (2007) Evaluation of selected mathematical models to predict the inactivation of Listeria innocua by pulsed electric fields. LWT- Food Sci Technol 40:1271–1279 Sharabi S, Kos B, Last D, Guez D, Daniels D, Harnof S, Mardor Y, Miklavčič D (2016) A statistical model describing combined irreversible electroporation and electroporation-induced blood–brain barrier disruption. Radiol Oncol 50:28–38 Sun D-W (ed) (2014) Emerging technologies for food processing, 2nd edn. Academic/Elsevier, Amsterdam Towhidi L, Kotnik T, Pucihar G, Firoozabadi SMP, Mozdarani H, Miklavčič D (2008) Variability of the minimal transmembrane voltage resulting in detectable membrane electroporation. Electromagn Biol Med 27:372–385. doi:10.1080/15368370802394644 Yarmush ML, Golberg A, Serša G, Kotnik T, Miklavčič D (2014) Electroporation-based technologies for medicine: principles, applications, and challenges. Annu Rev Biomed Eng 16:295–320. doi:10.1146/annurev-bioeng-071813-104622 Županič A, Kos B, Miklavčič D (2012) Treatment planning of electroporation-based medical interventions: electrochemotherapy, gene electrotransfer and irreversible electroporation. Phys Med Biol 57:5425–5440. doi:10.1088/0031-9155/57/17/5425

Modeling Transdermal Delivery by Electroporation: The Thermodynamic Approach Sid Becker

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Stratum Corneum and the Lipid Barrier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SC Lipid Thermal Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conceptual Description of LTR Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mathematical Description of LTR Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modeling Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Skin as a Composite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Stratum Corneum Defect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermophysical Parameter Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Computational Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 4 7 9 11 12 13 13 14 15 15

Abstract

Low voltage long pulse electroporation of the skin results in localized regions within the stratum corneum of high permeability to mass transfer and electrical conduction. It is believed that these permeability increases result from resistive Joule heating that causes lipid phase changes. This chapter considers the physiology of the stratum corneum and the thermodynamic behavior of the lipid microstructure. A conceptual model of the evolution of the local transport region is presented in order to clearly illustrate the physics underlying this dynamic process. The mathematical equations which govern the conservation of charge, the conservation of thermal energy, and the conservation of drug mass are directly linked to the thermodynamics state of the stratum corneum lipids. The chapter concludes with general modeling considerations to capture increased mass transport associated with electroporation of the skin. S. Becker (*) Mechanical Engineering Department, University of Canterbury, Christchurch, New Zealand e-mail: [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_14-1

1

2

S. Becker

Keywords

Lipid, phase transition • Thermodynamic • Transdermal

Introduction Skin electroporation has been proposed to increase the permeability of the skin to mass transfer (Denet et al. 2004). During skin electroporation the skin is exposed to a series of intense electric pulses. When the drop in electric potential across the skin exceeds a critical value, the skin’s permeability can increase by orders of magnitude. This increase in permeability is attributed to the creation of microscopic aqueous pores in the skin’s outer microstructure. The nature of the changes to the skin’s structure depends on electroporation pulse characteristics such as pulse shape, pulse intensity, and pulse duration (Pavselj et al. 2015). The most common types of pulse shapes used in skin electroporation are the square wave pulse and the exponentially decaying pulse. The square wave pulse has a constant voltage during the entire pulse application. The exponentially decaying pulse amplitude is initially at its maximum value V0 and then decreases exponentially in time as: V(t) = V0∙exp(
t ∕ τ) where the time constant τ regulates the rate of decay. Skin electroporation pulses have been classified into two primary regimes: highvoltage short-duration pulses (HV) and low-voltage long-duration pulses (LV). The two pulsing regimes result in different physics. While for square wave pulses the distinction between HV and LV is relatively clear, for exponentially decaying pulses, this distinction can be less obvious when the tail of the pulse is very long (Pavselj et al. 2015). The short HV electroporation pulse durations can reach 1 ms but are typically 100 μs or less. When the drop in electric potential across the skin reaches some critical value (30–100 V) (Pliquett et al. 1995), skin resistance (both ionic and molecular) drops several orders of magnitude (Denet et al. 2004). In the short pulse regime, it is believed that nm-sized pores develop within the barrier of the skin. The nm-sized pores associated with the short HV electroporation perforate the skin in a more or less homogenous distribution. This result is very helpful when the drug is considered to be small (low molecular weight). This chapter considers the influence of the long duration LV pulses (lasting up to several hundred ms). Such LV pulses make possible two important secondary effects: electrophoresis and Joule heating. The longer duration pulse can result in significant electrophoretic effects that can drive large charged molecules into the skin. The second important effect of long duration skin electroporation pulses is Joule heating which results in localized temperature rises that contribute to increased permeability of the SC by lipid chain melting. Early researchers of skin electroporation used the phrase local transport region (LTR) to describe a phenomenon in which the applied pulses result in “large” μmsized regions of increased permeability (Pliquett et al. 1998). The results of in vitro

Modeling Transdermal Delivery by Electroporation: The Thermodynamic. . .

3

studies show that within the LTR the electrical and mass permeabilities may be many orders of magnitude greater than outside the LTR and that development of LTRs is always associated with thermal effects (Denet et al. 2004). This is believed to be a consequence of the response of the SC lipids to the elevated temperatures that can occur as a result of resistive heating (also called Joule heating) during the applied LV pulse. This chapter describes the current understanding of the physics underlying the relationships between the Joule heating associated with an LV pulse and the permeability increases that are indicative of the growth of the LTR.

The Stratum Corneum and the Lipid Barrier The outer part of the skin is called the epidermis. Its thickness varies depending on the location of the body and varies from person to person. The epidermis is thickest at the palms and soles of the feet reaching about 1.5 mm, while at the eyelids the epidermis is the thinnest at about 0.05 mm. The epidermis is without vasculature and acts as a protective barrier preventing infection and molecular transport. The epidermis can be considered to be an assembly line of the transformation of cells: the epidermal base cells are viable and living, while at the outer surface the cells have transitioned into their dead and flat form. Below the epidermis is the highly vascular dermis (0.3–3 mm thick). Directly below the dermis, at most sites on the body, lies a fatty layer that acts to conserve body heat. At most sites on the body, the skin is perforated by appendages in the form of sweat glands and hair follicles. The thin outermost layer of the epidermis is the Stratum Corneum (SC). The SC is composed of 15–20 layers of corneocytes (flat dead cell shells) which are interconnected by a lipid lamellar bilayer structure in a crystalline-gel phase. The corneocyte-lipid matrix is depicted in Fig. 1. While the thickness of the SC also varies depending on body location, consider that at the forearm it is about 20 μm thick and at the shoulder it is about 10 μm thick. The SC provides the greatest resistance to transport: the permeability to molecular transport can be several orders of magnitude lower in the SC than in the adjacent layers of the skin. Thus the SC receives the most attention in developing transdermal transport protocols which is the subject of this chapter. In order to give a conceptual understanding of the architecture and structure of the SC, a brick and mortar model is sometimes employed. A depiction of a brick and mortar model is presented in Fig. 1. Here the corneocytes are represented by the bricks and the lipid sheets are represented by the mortar filled spaces between the bricks. Note that the lipid structure is interconnected so that, in principle, there is a continuous lipid-filled pathway through the SC. The pathway that weaves around the corneocytes through this lipid structure is called the intercellular route. Because the intercellular route is very indirect and tortuous, the length of the path that a molecule takes through this route can be 10–50 times longer than the thickness of the SC (Lane 2013). By contrast, the much shorter transcellular route passes straight through the corneocytes and the lipids; this path is possible but not common. A final potential

4

S. Becker

Fig. 1 General representation of the stratum corneum, the brick and mortar representation, and the three primary transport routes

path through the SC is called the transappedageal route; as its name indicates, this route follows a pathway along skin appendages such as sweat glands and hair follicles. However, since the skin appendages occupy only about 0.1 % of the total skin area, this is a relatively restricted route (Lane 2013). In practice it is the tortuous intercellular route through the lipid-filled spaces of the SC that most transport follows (Lane 2013). In order to use this intercellular, i.e., lipid route, researchers in the field of transdermal drug delivery have gone to great efforts to understand how to manipulate the lipid microstructure. This chapter considers the manner in which the thermodynamic behavior of the lipid structure can be related to LV electroporation. In order to better understand this relationship, consider next the response of the lipid structure to thermal state.

SC Lipid Thermal Behavior Because LV pulses have been shown to result in localized intense temperature rises at sites on the SC (Denet et al. 2004), this section concerns the behavior of the SC lipid structure at elevated temperatures. The composition of the SC lipid structure is responsible for the skin barrier function. In its undisturbed state, the SC lipid structure is highly organized in a lamellar phase that is composed of sheets of lipid bilayers (see insert Fig. 1). This

Modeling Transdermal Delivery by Electroporation: The Thermodynamic. . .

5

lipid structure is known to become destabilized at elevated temperatures and the structural rearrangement at elevated temperatures has been shown to positively affect the SC’s permeability. Different experimental techniques show this restructuring of the SC lipids at elevated temperatures. For example, x-ray diffraction microscopy shows that the organized lamellar structure of the SC lipids begins to disappear at temperatures between 60  C and 70  C (Silva et al. 2006). When SC lipids are studies using polarized light thermal microscopy, the lipid structures are shown to dissolve at temperatures around 60  C (Silva et al. 2006). The SC has also been shown to be much more permeable to water flux at temperatures around 70  C (Golden et al. 1987). Perhaps the most useful data that describes the thermal behavior of the SC lipids is gained from differential scanning calorimetry (DSC). The motivation behind DCS testing is to identify and define the thermal transitions of the SC. In DSC, a sample of SC is heated so that its temperature increases at a constant rate ( C/min). The principle is that during this controlled temperature rise, the rearrangement of the lipid structures undergoing phase transitions requires additional heating. By carefully monitoring the magnitude of the heating, the phase transition temperatures can be identified. It is believed that within the temperatures in the range 40–130  C, the SC experiences multiple endothermic transitions. Here the high-temperature transitions of the SC are considered by adopting the convention: phase changes E, F, and G. Figure 2a depicts data representative of a DSC study for which the three lipid transition stages are reflected by the steep portions of the heat versus temperature curve. Phase change E which occurs from about 65–72  C is associated with the disordering of the lamellar lipid phase (Tanojo et al. 1999). Phase change E is believed to be responsible for the noticeable increases in the permeability of transport through the intercellular route. Therefore understanding phase change E is critically important when attempting to model the permeability increases associated with thermal effects of skin electroporation. Secondary lipid melting has been associated with phase change F at temperatures around 80  C, but this is believed to be related only to the lipids that are bonded to the corneocytes (Tanojo et al. 1999). Phase transition G is less critical to the barrier function; it is related to irreversible protein denaturation at temperatures above 90  C (Van Duzee 1975). In order to predict the permeability increases, it is necessary to approximate the degree to which lipid phase transition E has taken place. Consider that DSC data can provide the total additional amount of heat required for lipid phase transition E. This is the latent heat ΔHE that is absorbed by the SC as it is heated from TE1 (at which phase transition E begins) to TE2 (the temperature at which phase transition E concludes). One established method that is used to predict the state of the SC represents the degree of SC lipid alteration as the ratio of heat added to the lipids to the heat required for lipid phase transition. The function referred to as the lipid melt fraction is used to describe the degree of lipid disorder. It is mathematically defined as:

6

S. Becker

Fig. 2 (a) Representative differential scanning calorimetry data indicating the three phase changes. (b) Rectangular specific heat versus temperature representation

φ¼

ðH
cSC ðT
T E1 ÞÞ ΔH E

(1)

where cSC is the SC specific heat capacity and ΔHE is the latent heat associated with phase transition E. The total enthalpy, H, is defined as: ðT H¼

cSC, APP dτ

½T E1 < T  T E2 

(2)

T E1

where TE1 and TE2 are the temperatures over which transition E takes place, and cSC, APP is the apparent specific heat of the SC. Equations 1 and 2 can provide an estimate of the lipid state when the temperature dependence of the specific heat is approximated directly from the DSC data. As an alternative to the direct representation of the heat versus temperature curves of Fig. 2a, it is often more convenient to approximate the DSC data using a simple rectangular specific heat versus temperature curve as depicted in Fig. 2b. In this way, the apparent SC specific heat, cSC,APP, can be represented as the sum of the specific heat capacity and the latent specific heat as:

Modeling Transdermal Delivery by Electroporation: The Thermodynamic. . .

cSC, APP ¼ cSC þ cSC, L

7

(3)

where cSC,L is the latent specific heat: cSC, L ¼

ΔH E T E2
T E1

(4)

When the specific heats are treated as constant over the phase transition temperatures (as in Fig. 2b), Eqs. 3 and 4 can be substituted into Eq. 2 to arrive at:  H ¼ cSC þ

 ΔH E ðT
T E1 Þ T E2
T E1

(5)

When Eq. 5 is substituted into the melt fraction of Eq. 1 and simplifies to a ratio of temperature differences, the result is: φ¼

T
T E1 T E2
T E1

(6)

By approximating the DSC data of Fig. 2a as the rectangular shaped specific heat curve of Fig. 2b, it is possible to predict the state of the lipids by the simple ratio of temperature differences of Eq. 6. The lipid melt fraction is a thermodynamically based function that can be used to represent the degree of lipid disorder in the range 0  φ  1. When φ = 0, the lipids are in their unaltered state. In this case, the intercellular route is composed of lipid in their orderly lamellar structures which is very resistive to transport. Additionally this organized structure is associated with a SC that is electrically resistive. When the SC is heated above lipid transition temperatures, the melt fraction increases so that φ > 0, and the SC experiences an increase in electrical conductivity and permeability to mass transport. The limit of these increases occurs at temperatures above TE2 for which φ = 1. At a melt fraction of unity, the lipid architecture is in its completely unstructured state so that the SC is very electrically conductive and is much more permeable to mass transport. A melt fraction of φ > 0 is associated with the permeability increases that are indicative of the LTR; in this way, within the LTR φ > 0, while outside of the LTR φ = 0.

Conceptual Description of LTR Evolution In order to better understand the model of the LTR development during LV electroporation that results as a consequence of thermally induced lipid phase transition, it is helpful to begin with a simple conceptual representation of the process. This process is depicted in Fig. 3. Consider the unaltered SC as in Fig. 3a. Here the electrically resistive SC surrounds a small region that is representative of an appendageal pore or some defect in the SC microstructure. This defect region exists

8

S. Becker

Fig. 3 Conceptual representation of local transport region evolution by electroporation

prior to the application of the pulse and it has a higher electrical conductivity than the surrounding SC. Recall that in the case of skin appendageal macropores, they only occupy a small fraction of the total skin area, so that the existence alone of a defect does not preclude the need for increased permeability. When the LV electric field is applied, the current will prefer the high conductivity pathway through the preexisting pore to the surrounding electrically resistive SC (Fig. 3b). This high current density will result in resistive heating (also called Joule heating). The LV pulse duration is long enough to sustain the temperature rises required for the lipids in the surrounding SC to experience phase transition. As the lipid structures are altered, the SC electrical conductivity within this region increases allowing the electric current to penetrate this part of the SC (Fig. 3c). The current density is proportional to the area affected by lipid phase transitions; thus as the affected area grows, the current density within this area decreases. The area affected by lipid phase

Modeling Transdermal Delivery by Electroporation: The Thermodynamic. . .

9

changes (the LTR) continues to grow either until the end of the electric pulse or until the current density within the LTR is so low that further lipid melting cannot be sustained. This process is representative of the growth of the LTR; within the regions of the SC that have experienced lipid thermal phase transitions, the permeability is much higher than in the surrounding unaffected regions of the SC (Fig. 3d).

Mathematical Description of LTR Evolution As the electric field is applied to the skin, the current attempts to cross the electrically resistive SC. Any SC location with a lower electrical resistance (as would be associated with a sweat gland or hair follicle) will result in a much higher current density. This high current density results in the Joule heating that provides the necessary power for the sudden intense local temperature rises. As the lipids experience phase transition, the electrical conductivity of the SC in these regions increases. This section presents the governing equations that capture this coupled relationship between the electrical field and the thermal field. The temperature distribution is governed by the conservation of thermal energy: ρSC cSC, APP

@T ¼ ∇  ðkSC ∇T Þ þ QJ @t

(7)

where ρ is the density, k is the thermal conductivity, T is the temperature, and t is the time. To fully capture the effect of phase change, the apparent specific heat, cSC,APP, is represented by Eq. 3. The latent specific heat of Eq. 3 depends on the lipid temperatures (it is phase dependent). Recalling the discussion of lipid thermodynamic behavior of section “SC Lipid Thermal Behavior” and using a rectangular specific heat versus temperature approach (Fig. 2b), the latent specific heat may be represented as: 8 0 > > > T
T E1 > > ΔH E > > > T E2
T E1 > < T
T E2 cSC, L ðT Þ ¼ ΔH F > T F2
T E2 > > > T
T F2 > > ΔH G > > T G 2
T F21 > : 0

for T < T E1 for T E1  T  T E2 for T E2  T  T F2

(8)

for T F2  T  T G2 for T F2 < T

Here the subscripts refer to the phase transition E, F, or G. Note that DSC data can provide the heat of transition ΔH values for each phase change as well as the associated temperature ranges which are provided in Table 1. The term, QJ, of Eq. 7 is the Joule heat generated from the electric field. It is the Joule heating that results in the intense temperature rises that are required for thermal phase transitions. The magnitude of the Joule Heating may be represented as:

10

S. Becker

Table 1 Differential scanning calorimetry data of stratum corneum lipid phase change (Adapted from Cornwell et al. (1996) and Silva et al. (2006)) [T1,T2]  C [65,75] [75,90] [90,110]

Phase change E F G

QJ ¼ σ SC j∇ϕj2

ΔH (J/kg) 5500 4000 4700

(9)

Here σ SC is the SC electrical conductivity and ϕ is the electric potential of the applied electric field. Consider that once the local temperature is determined, the extent of lipid melting can be expressed through the melt fraction of Eq. 6 which is rewritten here for emphasis:

φð T Þ ¼

8 > < 0T
T

for T < T E1 E1

> : T E2
T E1 1

for T E1  T  T E2 for T E2 < T

The electric field’s transient non-Ohmic behavior occurs at very short time scales (sub ms) (Chizmadzhev et al. 1998) relative to those of thermal (ms to s) and mass transport (min to hour) and can be neglected for time steps around 1 ms (Chizmadzhev et al. 1998). To model the macroscale electrical behavior of the SC during LV electroporation is often represented by the Laplace equation: ∇  ðσ SC ∇ϕÞ ¼ 0

(10)

The electrical conductivity of the SC has been shown to experience increases during the disruption of the SC architecture. Experimental observations on the thermal dependence of SC electrical resistance have shown that the magnitude of the electrical resistance can drop by two orders of magnitude at lipid thermal phase change temperatures (Craanevanhinsberg et al. 1995). Large changes in the SC electrical conductivity will strongly influence the electric potential of Eq. 10 as well as the Joule heating in Eq. 9 (Becker et al. 2012, 2014; Becker 2011, 2012; Becker and Kuznetsov 2008). To capture this effect, the studies have related the local SC effective electrical conductivity increase directly to lipid melt fraction by a linear relation:   σ SC ¼ σ SC, U þ φ σ SC, M
σ SC, U

(11)

where σ SC,U is the SC electrical conductivity associated with the unaltered lipid structure (φ = 0), and σ SC,M is the electrical conductivity associated with the SC after phase change E is complete (φ = 1).

Modeling Transdermal Delivery by Electroporation: The Thermodynamic. . .

11

Electrically driven transdermal delivery is negotiated by the three modes of transport: electrophoresis, electroosmosis, and diffusion (Denet et al. 2004). Studies focusing on electrically driven transport of large charged molecules show that electroosmotic effects are negligible compared to electrophoresis forces (especially for the relatively short electroporation pulse). Homogenous tissue in vivo electroporation studies show that electrophoretic forces dominate in the transport of large molecules (Pavselj and Preat 2005). The transport of solute through the SC and into the underlying domain is formulated from the Nernst–Planck equation: @C ¼ ∇  ðmSC C∇ϕÞ þ ∇  ðDSC ∇CÞ @t

(12)

Here the term C is the concentration of the drug which is expressed in units of mass per unit volume. The transport coefficient mSC is the effective electrophoretic mobility of the solute (drug) in the SC, and DSC is the effective diffusion coefficient in the SC. The magnitude of the transport coefficients depend on the drug mass, the drug shape, and the drug charge as well as the microstructure characteristics of the medium (here of the SC). Their values can be approximated experimentally or theoretically. In order to show the link between the increase in the SC lipid transport coefficients and the lipid state, consider once again the melt fraction. The experimentally observed increases in permeability to mass transport can be satisfied by relating the transport coefficients directly to the lipid melt fraction of Eq. (6): mL ¼ f ðφÞ and DL ¼ gðφÞ. A linear dependence of transport coefficients on the lipid melt fraction has previously been posed, so that   mSC ¼ mSC, U þ φ mSC, M
mSC, U

(13)

  DSC ¼ DSC, U þ φ DSC, M
DSC, U

(14)

and

The transport coefficients mSC,U and DSC,U correspond to the low magnitudes of mobility and diffusion that are associated with the drug in the SC whose lipids have not experienced any thermal transition φ ¼ 0. The transport coefficients mSC,M and DSC,M represent the highest values of transport coefficient after full lipid melting φ ¼ 1.

Modeling Considerations This chapter has so far focused only on the SC during skin electroporation. In order to model LTR evolution and the subsequent transport of drug, it is necessary to consider the skin as a whole, the associated parameter values, and the computational and numerical methods used to model the conservation equations.

12

S. Becker

Fig. 4 Composite representation of the skin

The Skin as a Composite In practice, electroporation is conducted on the entire skin (not just the SC). Consider the composite representation of the skin presented in Fig. 4 that includes the different layers of the skin as well as an applicator gel at the surface of the SC. In such a case the conservation of charge, the conservation of thermal energy, and the conservation of drug mass must be considered in each of the different composite layers. In the gel layer @T ¼ ∇  ðkG ∇T Þ þ σ G j∇ϕj 2 |fflfflffl{zfflfflffl}QJ @t ∇  ðσ G ∇ϕÞ ¼ 0 @C ¼ ∇  ðmG C∇ϕÞ þ ∇  ðDG ∇CÞ @t ρG c G

(15)

Note that in the conservation of thermal energy, the symbol representing Joule heating, QJ, has been replaced by its definition. In the epidermis @T ¼ ∇  ðkED ∇T Þ þ σ ED j∇ϕj2 @t ∇  ðσ ED ∇ϕÞ ¼ 0 @C ¼ ∇  ðmED C∇ϕÞ þ ∇  ðDED ∇CÞ @t ρED cED

(16)

In the dermis @T ¼ ∇  ðkD ∇T Þ þ σ D j∇ϕj2
ωcb ðT
T b Þ þ QM @t ∇  ðσ D ∇ϕÞ ¼ 0 @C ¼ ∇  ðmD C∇ϕÞ þ ∇  ðDD ∇CÞ @t ρD c D

(17)

Modeling Transdermal Delivery by Electroporation: The Thermodynamic. . .

13

The conservation of thermal energy in the dermis has been structured to include the effects of blood perfusion and of volumetric energy generation. The perfusion rate ω describes the mass flow rate of blood per volume of dermis. While dermis perfusion rates can fluctuate, a good approximation is about ω = 2 kg·m
3s
1. The term Ta is the arterial blood temperature (assigned a value of 37  C), and the parameter cb is the specific heat of blood: cb = 3800 kJ/kgK. In the living tissue layers, the relative contribution of the volumetric metabolic energy generation term, QM, is small compared to the heat supplied by Joule heating so that it is often neglected. Experimentally it has been shown that, depending on pulse intensity, when intense electric pulses are applied to living tissue, the perfusion of blood may be locally temporarily inhibited (80 % reduction) for time scales ranging from minutes to hours (Jarm et al. 2010). Even if perfusion is uninhibited it is likely that during the short time interval of the application of the pulse, the presence of perfusion makes a negligible contribution to the thermal profile.

The Stratum Corneum Defect Recall the conservation of electric charge of Eq. 10 that states ∇  ðσ SC ∇ϕÞ ¼ 0 Note that in the conceptual description of Fig. 3, the electrical conductivity of the SC is higher within the SC defect than in the surrounding SC. This is required in order to achieve the concentrated current density that is needed for the resistive Joule heating of Eq. 9. With this in mind, it is very important to note that this thermodynamic approach requires that the electrical conductivity of the SC prior to the application of the pulse is not uniform throughout the SC and that there is at least one region in the pre-electroporated SC whose electrical conductivity is higher than in the surrounding SC. To account for this, one can simply assign a higher value of electrical conductivity in a small region of the SC. An upper limit of conductivity value in the defect is the conductivity of the applicator gel. In this way: Inside the defect: σ SC ¼ σ G

(18)

In previous studies, the defect region has been modeled as a small cylindrical volume of SC with a diameter of 5–20 μm.

Thermophysical Parameter Values The thermodynamic properties of the skin layers such as density, heat capacity, and thermal conductivity are well established. Electrical properties can be a bit more challenging to find reliable estimates for. This is because the electrical properties of biological tissue are often sensitive to the condition of the tissue (hydration level and

14

S. Becker

Table 2 Representative values of the thermophysical parameters of the skin (Adapted from Zorec et al. (2013)) Thickness, L (μm) Thermal conductivity, k (W/mK) Density, ρ (kg/m3) Heat capacity, c (J/kgK) Electrical conductivity, σ (S/m) Electrophoretic mobility, m (m2/Vs) Diffusion coefficient, D (m2/s)

Gel – 0.6

SC 20 0.2

Epidermis 50 0.21

Dermis 300 0.29

1000 4180 1

1500 3600 σSC-U = 1  10
4 σ SC-M = 5  10
2 mSC-U =1  10
15 mSC-M = 1  10
10 mSC-U =1  10
17 mSC-M = 1  10
12

1110 3600 0.1

1115 3800 0.15

1  10
9

1  10
9

3  10
11

3  10
11

1.5  10
8 1.5  10
10

how it is stored). Generally order of magnitude estimates have been used to represent the conductivity values. The transport coefficients, electrophoretic mobility and the diffusion coefficient, can be strongly influenced by tissue state. These are also very dependent on the drug molecular weight and size. Some representative thermophysical property values are provided in Table 2. The electrical conductivities listed are order of magnitude estimates. The SC conductivity values corresponding to the unaltered lipid state, σ SC,U, and the electrical conductivity associated with total alteration of the SC lipids, σ SC,M, have been chosen to represent the two order of magnitude increase in electrical conductivity with lipid restructuring as suggested by the results of reference (Craanevanhinsberg et al. 1995). The transport coefficient values are also order of magnitude representations and are representative of calcein. The properties of the gel are representative of a phosphate buffer solution (pH 6.5, 100 mM).

Computational Considerations The solution of Eqs. 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, and 17 must be evaluated numerically. In the past, these equations have been recast in discrete finite volume forms. The computational domain representing the space occupied by the skin is divided into individual control volumes, and the electrical charge, the thermal energy, and the mass are conserved in each of these control volume. When the space is discretized, it is reasonable to refine the mesh in regions that are anticipated to experience large variations in parameters values. In the case of LV skin electroporation, refine the mesh within the SC around the defect region. This will be particularly helpful when attempting to resolve the lipid melt fraction and its influence on the transport coefficients of Eqs. 11, 13, and 14 (Becker 2012; Becker and Kunetsov 2007, 2008).

Modeling Transdermal Delivery by Electroporation: The Thermodynamic. . .

15

The conservation of thermal energy is coupled directly to the conservation of charge through the Joule heating term. Within the SC, the transport coefficients (conductivity, mobility, and diffusion coefficient) are temperature dependent through the melt fraction. Because the equations are so strongly coupled, they must be solved iteratively. These coupled equations have previously been solved using a combined explicit-implicit approach. The implicit formulation requires an iterative approach between time steps. To speed convergence it has been previously proposed that all properties within the SC of the current time step are evaluated using the melt fraction of the previous time step (Becker 2012; Becker and Kunetsov 2007, 2008). Note that the Laplace equation that is used to represent the conservation of charge of Eqs. 10, 15, 16, and 17 is not explicitly transient (it has no time derivative). This is generally an accepted approximation because the electric field develops on a time scale that is so much shorter than the thermal field. Numerically what can be done is that the steady Laplace equation is reevaluated at each time step (Sel et al. 2005).

Conclusion Permeability increases resulting from the thermal effects associated with skin electroporation-related Joule heating can be explained in terms of the lipid chain melting. In this way, the transport coefficients can be related directly to the lipid state. By considering differential scanning calorimetry data of stratum corneum thermal behavior, the phase transition E has been proposed to represent the thermal alteration of the lipid microstructure. When the stratum corneum thermal data is recast in a rectangular specific heat versus temperature representation, the degree of lipid melting can be approximated by a ratio of temperature differences. This principle allows the transport coefficients and the electrical conductivity of the stratum corneum to be predicted from the local temperature distribution.

Cross-References ▶ Drug and Gene Delivery Through the Skin ▶ Mass Transfer of Electrolytic Species During Electric-Field-Based Tumor Treatments ▶ Mechanistic Description of Membrane Electropermeabilization ▶ Modeling Transport Across the Electroporated Membrane ▶ Temperature Increase and Thermal Effects Due to Electroporation in Tissues

16

S. Becker

References Becker SM (2011) Skin electroporation with passive transdermal transport theory: a review and a suggestion for future numerical model development. J Heat Transf-Trans ASME 133 (1):011011. doi:10.1115/1.4002362 Becker SM (2012) Transport modeling of skin electroporation and the thermal behavior of the stratum corneum. In: Vafai K (ed) 4th international conference of porous media and its applications in science, engineering, and industry, Potsdam. AIP Conf Proc, pp 365–370 Becker SM, Kunetsov AV (2007) Local temperature rises influence in vivo electroporation pore development: a numerical stratum corneum lipid phase transition model. J Biomech Eng-Trans ASME 129(5):712–721. doi:10.1115/1.2768380 Becker SM, Kuznetsov AV (2008) Thermal in vivo skin electroporation pore development and charged macromolecule transdermal delivery: a numerical study of the influence of chemically enhanced lower lipid phase transition temperatures. Int J Heat Mass Transf 51(7–8):2060–2074. doi:10.1016/j.ijheatmasstransfer.2007.06.010 Becker SM, Pavselj N, Zorec B, Miklavcic D (2012) Transport modeling of skin electroporation and the thermal behavior of the stratum corneum. Paper presented at the 23rd international symposium on transport phenomena, Auckland Becker S, Zorec B, Miklavcic D, Pavselj N (2014) Transdermal transport pathway creation: electroporation pulse order. Math Biosci 257:60–68. doi:10.1016/j.mbs.2014.07.001 Chizmadzhev YA, Indenbom AV, Kuzmin PI, Galichenko SV, Weaver JC, Potts RO (1998) Electrical properties of skin at moderate voltages. Biophys J 74(2):843–856. doi:10.1016/ s0006-3495(98)74008-1 Cornwell PA, Barry BW, Bouwstra JA, Gooris GS (1996) Modes of action of terpene penetration enhancers in human skin differential scanning calorimetry, small-angle X-ray diffraction and enhancer uptake studies. Int J Pharm 127(1):9–26. doi:10.1016/0378-5173(95)04108-7 Craanevanhinsberg WHM, Verhoef JC, Junginger HE, Bodde HE (1995) Thermoelectrical analysis of the human skin barrier. Thermochim Acta 248:303–318. doi:10.1016/0040-6031 (94)01887-m Denet AR, Vanbever R, Preat V (2004) Skin electroporation for transdermal and topical delivery. Adv Drug Deliv Rev 56(5):659–674. doi:10.1016/j.addr.2003.10.027 Golden GM, Guzek DB, Kennedy AH, McKie JE, Potts RO (1987) Stratum-corneum lipid phasetransitions and water barrier properties. Biochemistry 26(8):2382–2388. doi:10.1021/ bi00382a045 Jarm T, Cemazar M, Miklavcic D, Sersa G (2010) Antivascular effects of electrochemotherapy: implications in treatment of bleeding metastases. Expert Rev Anticancer Ther 10(5):729–746. doi:10.1586/era.10.43 Lane ME (2013) Skin penetration enhancers. Int J Pharm 447(1–2):12–21. doi:10.1016/j. ijpharm.2013.02.040 Pavselj N, Preat V (2005) DNA electrotransfer into the skin using a combination of one high- and one low-voltage pulse. J Control Release 106(3):407–415. doi:10.1016/j.jcornel.2005.05.003 Pavselj N, Zorec B, Miklavcic D, Becker S (2015) Experimental factors to be considered in electroporation-mediated transdermal diffusion experiments. J Biomech Eng-Trans ASME 137(12):124501. doi:10.1115/1.4031767 Pliquett U, Langer R, Weaver JC (1995) Changes in the passive electrical-properties of human stratum-corneum due to electroporation. Biochim Et Biophys Acta-Biomemb 1239(2):111–121. doi:10.1016/0005-2736(95)00139-t Pliquett UF, Vanbever R, Preat V, Weaver JC (1998) Local transport regions (LTRs) in human stratum corneum due to long and short ‘high voltage’ pulses. Bioelectrochem Bioenerg 47 (1):151–161. doi:10.1016/s0302-4598(98)00180-9 Sel D, Cukjati D, Batiuskaite D, Slivnik T, Mir LM, Miklavcic D (2005) Sequential finite element model of tissue electropermeabilization. IEEE Trans Biomed Eng 52(5):816–827. doi:10.1109/ tbme.2005.845212

Modeling Transdermal Delivery by Electroporation: The Thermodynamic. . .

17

Silva CL, Nunes SCC, Eusebio MES, Sousa JJS, Pais A (2006) Study of human stratum corneum and extracted lipids by thermomicroscopy and DSC. Chem Phys Lipids 140(1–2):36–47. doi:10.1016/j.chemphyslip.2006.01.004 Tanojo H, Bouwstra JA, Junginger HE, Bodde HE (1999) Thermal analysis studies on human skin and skin barrier modulation by fatty acids and propylene glycol. J Therm Anal Calorim 57 (1):313–322. doi:10.1023/a:1010137807610 Van Duzee BF (1975) Thermal analysis of human stratum corneum. J Invest Dermatol 65 (4):404–408. doi:10.1111/1523-1747.ep12607656 Zorec B, Becker S, Rebersek M, Miklavcic D, Pavselj N (2013) Skin electroporation for transdermal drug delivery: the influence of the order of different square wave electric pulses. Int J Pharm 457(1):214–223. doi:10.1016/j.ijpharm.2013.09.020

Single Cell Electrical Characterization Techniques Olivier Français and Bruno Le Pioufle

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Impedance Measurement of a Single Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrical Characterization Using Electrorotation Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microelectroporation of a Single Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 5 9 11 15 17 17

Abstract

In this chapter the possibility to electrically characterize the electroporation of single cells is discussed. During the electroporation process, the application of pulsed electric field (▶ Transmembrane Voltage Induced by Applied Electric Fields) leads to the partial permeabilization of the cell membrane, opening access to the cytoplasmic compartment for drug delivery, for instance. The electric impedance of the cell is modified by such a treatment. Impedance measurement can thus be employed to characterize the permeabilization effect, on a given angular frequencies range, which leads to an impedance spectrum. The impedance spectrum measurement can be achieved even at the level of single cells. In that case, the distance between the detection microelectrodes and the cell must be considered in order to achieve a satisfactory sensitivity. In addition, when

O. Français ENS de Cachan, CNRS SATIE, Cachan, France e-mail: [email protected] B. Le Pioufle (*) ENS de Cachan, CNRS SATIE, Cachan, France Ecole Normale Supérieure de Cachan, Cachan, France e-mail: bruno.lepioufl[email protected]; bruno.le-pioufl[email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_15-1

1

2

O. Français and B. Le Pioufle

considering the impedance measurement of cells flowing through a microfluidic channel, fast detection is also required, as the cell is present between the measurement electrodes only for a short while. A specific strategy must then be employed to achieve such goal. In this case, instead of screening the whole frequencies spectrum, a few set of frequencies should be selected and mixed to be simultaneously injected prior to synchronous detection. Another way to assess the single cell impedance is the use of indirect methods based on the mechanical behavior of cells that polarize under the application of a stationary or propagative electrical field. In particular, electrorotation experiment provides the electrical characterization of the cell, which is dependent of its level of permeabilization, after the application of electrical field pulses. When such approach is used, the vicinity of electrodes is less critical as the main requirement lays on the homogeneity of the rotating electrical field. Finally, the possibility to focus the electroporation at the subcellular level is discussed in the chapter. Indeed, the use of micro- or nanotechnology permits to localize spatially the electroporation on a tiny portion of the membrane of single cells. Adherent cells can be locally electroporated, using nanosized electrodes, subsequently used to monitor the intracellular potential. Keywords

Bioimpedance • Single Microelectroporation

cell

•

Microfluidics

•

Electrorotation

•

Introduction The impedance measurement is a common way to investigate the level of electroporation of cells (Pavlin et al. 2005). Such measurement can be achieved even at the level of single cells. In that case, the vicinity between the detection microelectrodes and the cell must be considered in order to achieve a satisfactory sensitivity. Indeed the measurement electrodes probe a volume composed of the cell and the medium, the volume ratio between those being thus of prime importance. The impedance spectrum is defined as the ratio between the voltage applied to a pair of electrodes and the current collected by these electrodes, in function of the applied angular frequency. In the case where only medium is contained between the electrodes (distance between the electrodes D, surface of electrode S, see Fig. 1), the impedance expresses: Zm ¼

1 D 1 ¼ σ mS em S S iωem þ iω D D

(1)

where σDmS represents the conductance of the medium between the two electrodes, and em S D its capacitance. In this equation, em represents the permittivity of the medium, σ m its conductivity, ω stands for the angular frequency of the voltage applied to

Single Cell Electrical Characterization Techniques Fig. 1 Single cell immersed within the medium between two electrodes (gap D, surface S). The dielectric properties of the different compartments of the cell (e*memb, e*in) are homogenized to e*cell thanks to the single shell model

3

electrode 1

S (surface)

external medium

e

mbrane me R

D

cytoplasm

electrode 2 electrodes, and i is the complex number (i2 = 1). This impedance can also be expressed in function of the complex permittivity of the medium, defined as em ¼ em i σωm , as seen in Eq. 1. Such spectrum is modified by the presence of a cell between the electrodes, as represented in Fig. 1. The impedance of an elementary volume of medium containing a single cell disposed between those two measurement electrodes (distance between the electrodes D, surface of electrode S, see Fig. 1) can be approximated using the Maxwell mixture theory (Maxwell 1881): Z mix ¼

D 1 D 1 ¼ S σ mix þ iωemix S iωemix

(2)

where e*mix is the complex permittivity of the mixture, composed of the volume of medium containing the single cell, that can be expressed in function of the complex permittivity of the medium itself e*m, the volume fraction Φ of the cell within the medium, and the Claussius-Mossotti factor fCM that characterizes the contrast between the cell and the medium in term of complex permittivities (Jones 2005): 1 þ 2Φ f CM 1  2Φ f CM

(3)

ecell  em ecell þ 2em

(4)

emix ¼ em with f CM ¼

4

O. Français and B. Le Pioufle

and Φ the fraction volume between the cell and the volume between electrodes: 4 πR3 Φ¼3 3 D

(5)

the complex permittivity e* being defined as a function of the permittivity e and conductivity σ of the corresponding material (inner compartment of the cell, cell membrane, medium) as: e ¼ e  i

σ ω

(6)

σ mix and emix relate respectively to the conductivity and permittivity of the equivalent mixture (medium containing the cell). The complex permittivity of the cell e*cell (that appears in the expression of the Claussius-Mossotti factor fCM and thus affects the impedance of the mixture, Eq. 2) depends on many parameters as it relates to the geometry, permittivity, and conductivity of the cell components (membrane, cytoplasm, intracellular components). Anyway the complex permittivity of a spherical-shaped cell can be approximated, using the single shell model (Huang et al. 1992), when considering the cell as an averaged inner compartment having its own dielectrical properties σ in and ein, surrounded by an outer cytoplasmic membrane (σ memb and ememb, see Fig. 1):

ecell

  3 2
 ein  ememb R 3 þ2  6 Re ein þ 2ememb 7 7 ¼ ememb 6 4
 3  e  e 5 in memb R   Re  ein þ 2ememb

(7)

where R stands as the cell radius and e as the thickness of the cell membrane, as depicted in Fig. 1. This chapter will describe firstly the influence of the electroporation on the impedance spectrum characteristic of single cells within miniaturized microfluidic devices (section 1, Impedance Measurement of a single cell). The dependance of the measurement resolution with the volume fraction Φ between the cell and the surrounding medium will be assessed. Methods for the impedance measurement of flowing cells, that can be achieved by choosing a set of excitation frequencies simultaneous injected prior to synchronous detection, will be described and illustrated by some important achievement found in the literature. Section 2 (Electrical Characterization Using Electrorotation Experiment) will describe an alternative method to monitor the impedance spectrum of cells during electroporation that is less dependent on the high volume fraction Φ requirement. This method – the electrorotation spectrum – is based on the analysis of the mechanical angular velocity of cells exposed to a rotationary electric field. Finally the section 3 (Microelectroporation of a Single Cells) will show that the recent improvements of the technology permit the focusing of electrical field in such

Single Cell Electrical Characterization Techniques

a way that the electroporation (microelectroporation).

is

5

localized

on

the

cell

membrane

Impedance Measurement of a Single Cell The electrical characterization of a single cell is defined by the assessment of its dielectrical properties, the conductivity σ and permittivity e of its components (membrane and intracellular components). The single shell model of the cell provides a fair description of its dielectric behavior, through its complex permitivity spectrum, as shown on Fig. 2. Indeed, as introduced in the previous section, the complex permittivity of the cell ecell ¼ ecell  i σωcell involves a real part that relates to the dielectric polarization of the cell (permittivity ecell), while the imaginary part relates to the losses (conductivity σ cell). At the lowest frequencies, the cell membrane is preponderant ( ecell ! Re ememb when ω ! 0, from Eq. 7), while at high frequencies, the interfacial relaxation resonance does appear (see Fig. 2). The latter corresponds to the vibration of ions confined by the cell membrane within the inner part of the cell, which is associated to the loss peak visible in Fig. 2. The impedance measurement should sense both low and high frequencies characteristics (the membrane dielectric characteristic Re ememb at the lowest frequencies Complex permitivity of the cell

4

4

x 10

real part Imaginary part

3.5 membrane like behavior

Complex permitivity

3 2.5 2

Interfacial relaxation

1.5 1 0.5 0 0 10

2

10

4

10

Angular frequency

6

10

8

10

10

10

(Rd/s)

Fig. 2 Real and imaginary parts of the complex permittivity spectrum of a cell. The comparison between the real and imaginary parts of the complex permittivity highlights the fact that the cell membrane is predominant in its dielectric behavior at low frequency and the relaxation resonance induced in the inner compartment confined by the membrane that occurs at the higher frequencies. R = 5 μm, e = 5 nm, σ memb = 108 Sm1, σ in = 1 Sm1, e0 = 8.85 1012, ememb = 11e0, ein = 50e0

6

O. Français and B. Le Pioufle

and the relaxation phenomena induced by the resonance due to the ions confined by the membrane in the inner compartment at higher frequencies) and should thus provide the dielectric characteristics of both the inner compartment and the membrane. Nevertheless, the impedance measurement is affected by the dielectric characteristics of the surrounding medium included between sensing electrodes (see Fig. 1). Using the Maxwell’s mixture model, the impedance spectrum of the composite material including the cell and the surrounding medium, i.e., the dependance of Zmix with the electrical frequency ω, can be evaluated (Sun and Morgan 2010). The impedance of the mixture depends on the dielectric properties of the cell e*cell, on the dielectric properties of the medium e*m, and on the volume ratio between those. As shown in Fig. 3, where the complex resistivity spectrum is represented, the volume ratio Φ determines the sensitivity of the cell impedance measurement (the complex resistivity spectrum ρ*mix(ω) is defined in Eq. 8 as the adimensional equivalent of the mixture impedance Zmix). Indeed the miniaturization of the measurement system (gap and surface of electrodes) advantageously increases the sensitivity of the spectrum to the presence of a single cell between electrodes (10.6 % variation of the impedance, for low frequencies (ω ’ 105 Rd/s), for a characteristic dimension D = 4 Rcell).

Complex resistivity, mixture model 1 D=15 μm no cell, Cdl D=15, Cdl D=20 μm D=25 μm D=30 μm no cell

0.95

0.85 0.8

M in iatu riza tio n

0.9

0.75 0.7 5 10

6

10

Angular frequency

7

10

8

10

ω (Rd/s)

Fig. 3 Impedance spectrum of a single cell within a cubic volume of medium (electrode surface D2, electrode distance D), using the Maxwell’s mixture theory. Plain curves: the cell is present between electrodes that are separated with a distance D varying from 15 μm to 30 μm. Bold red curve: reference impedance spectrum (no cell). Dashed curves: the influence of the double-layer capacitance is shown, with or without cell. R = 5 μm, e = 5 nm, σ memb = 108 Sm1, σ in = 1 Sm1, σ m = 1.4 Sm1, e0 = 8.851012, ememb = 11e0, ein = 50e0, em = 80e0

Single Cell Electrical Characterization Techniques

ρmix ¼

7

1 1 S ¼ ¼ Zmix σ mix þ iωemix iωemix D

(8)

The impedance of the double-layer Zdl ¼ iωC1 D2 , due to the contact between the s σm electrode and the electrolyte, should not be neglected as the cut frequency DC s associated to its effect is also linked to the degree of miniaturization. The doublelayer capacitance is taken into account in Fig. 3 (dot lines, for a double-layer capacitance Cs = 0.5 Fm1, corresponding to black Platinum electrodes (Martinsen and Grimnes 2011)). Anyway this effect becomes negligible, if high surface capacitance materials are chosen for the electrodes, as soon as high electrical frequencies are considered. If this is not the case, differential measurement (as in Holmes et al. (2009), see Fig. 5) where the impedances with and without cells are subtracted will be implemented. Such differential measurement is broadly used as it also permits to get rid of the temperature effect, which occurs in the two measurement zones, and can thus be eliminated by the differentiation. Using a measurement configuration with convenient dimensions (high miniaturization D = 20 μm, for instance, with Rcell = 5 μm), the permeabilization of the membrane (▶ Different Approaches Used in Modeling of Cell Membrane Electroporation) can be monitored with impedance measurement. As shown in Influence of the permeabiliation on the complex resistivity, D=20 μm 0.79

0.77 0.76

−1

10−8 Sm −5 10 Sm −1 10−4 Sm −1 10−3 Sm −1 no cell 10−3, 45°

n

perm meembra abi ne liza tio

0.78

0.75 0.74 0.73 0.72 0.71 5 10

6

10

7

10

8

10

Angular frequency ω (Rd/s)

Fig. 4 Impedance spectrum of a cell, influence of the membrane permeabilization. The distance between electrodes is D = 20 μm. The membrane conductivity σ membrane is the variable parameter. Plain curves: the cell is present between electrodes, and has varying levels of σ memb, from 108 Sm1 to 103 Sm1. Bold pink curve: partial permeabilization of the cell membrane, σ memb = 103 Sm1 on the poles of the cell, with a 45 solid angle, while the rest of the membrane keeps the 108 Sm1 conductivity (this curve is obtained by a 3D finite element calculation). Bold red curve: reference curve, no cell between the electrodes

8

O. Français and B. Le Pioufle

Fig. 4, the permeabilization of the membrane leads to a significant modification of the measured impedance (11 % variation of the impedance for low frequencies (ω ’ 105 Rd/s), for a characteristic dimension D = 4 Rcell, and for an increase of the membrane conductivity σ memb from 108 Sm1 up to 103 Sm1). On this figure is represented the evolution of the mixture complex resistivity, when the whole membrane is permeabilized (pink curve), or when part of it is permeabilized (permeabilization of the poles of the cell facing the electrodes (Fig. 1), on a 45 solid angle section, bold pink curve. This curve is achieved thanks to finite element calculation). The impedance measurement of single cells flowing within a small channel equipped with electrodes was firstly performed by Hoffman and Britt (1979). The velocity of the cell is too fast for a whole impedance spectrum recording, since all frequencies should have to be screened while the cell is passing above the electrodes. For that reason a set of excitation frequencies has to be chosen that might provide the relevant information about the cell status. To do so Hoffman and Britt (1979) injected simultaneously this set of excitation frequencies, the measured spectrum being then demodulated using “synchronous detection” techniques. Gawad et al. (2001) performed the same type of experiment within a micromachined fluidic device. The influence of the medium variations, due to the temperature, for instance, as well as the effect of the double-layer capacitance is compensated thanks to a differential measurement (Morgan et al. 2006). Such a strategy was used successfully to differentiate white blood cells (neutrophils, monocytes, and lymphocytes) from the whole blood sample, after lysis of the erythrocytes (Holmes et al. 2009). To perform those assay, two excitation frequencies where injected for the impedance analysis, and the results were correlated to an optical detection within the same flowing microchannel (see Fig. 5) This approach has also been used to characterize on the same device the impedance variation of single cells flowing forward and backward within a microfluidic channel (Bürgel et al. 2015). Such a device, which structure is shown in Fig. 6, permitted a simultaneous electrical monitoring of the permeabilization thanks to impedance measurement (eight frequencies were simultaneously injected) as well as a microscopy observation of the intracellular content release and cell swelling, in real time and at the single cell level. Interesting possibilities to assess the cytoplasm content are investigated by extending this impedance spectroscopy towards the high frequencies. An example of these possibilities is provided in Haandbæk et al. (2014), where the highfrequency range (up to 500 MHz in this study) allowed to characterize and differentiate wild or mutant yeasts through their intracellular content. The frequency range is still limited by the technology, but more precise characterization of the dielectric properties of cells thanks to impedance spectrometry is expected in the near future.

Single Cell Electrical Characterization Techniques

9

Fig. 5 Microfluidic device for the electrical differentiation of white blood cells thanks to impedance measurement. Two electrical frequencies (503 kHz and 1.7 MHz) were simultaneously applied to contiguous pairs of electrodes for a differential measurement. One pair of electrodes probes the cell, while the other pair of electrodes probes a volume of the same medium without cell as a reference. The differential current measurement is then demodulated using synchronous detection techniques. Fluorescent labeling of the respective population to detect was achieved within the same channel, cell-by-cell in the detection region of the microchannel, using dual laser excitation (Adapted from Holmes et al. (2009), permission from Royal Society of Chemistry Publishing group, Ltd: Lab on Chip, copyright (2009))

Electrical Characterization Using Electrorotation Experiment As recalled previously and shown in Fig. 3 and Eq. 3, the volume fraction of the cell Φ within the measurement zone, i.e., the level of confinement, is crucial in order to achieve an acceptable sensitivity when performing impedance spectroscopy at the

10

Pt electrode Glass

a Flow provided by syringe pump

Fig. 6 On-chip electroporation and impedance spectroscopy. (a) The cell is injected within the microchannel and the impedance is measured before and after the electroporation (8 V amplitude AC voltage applied to platinum planar electrodes spaced 18 μm with a 50 kHz frequency). The cell is then flowed forward and backward between the measurement electrodes (to which superposed 1 V amplitude AC signals, in the range 20 kHz–20 MHz, are imposed). (b) View of the microdevice, and electrical and fluidic connections (From Bürgel et al. (2015), permission from Elsevier Publishing group, Ltd: Sensors and Actuators B, copyright (2015))

O. Français and B. Le Pioufle

Cell

Cell

EIS analysis

Electroporation

V(t) Cell

Cell

t

b Holder PCB

Electrode

Fused silica capillaries

Measurement region

Microfluidic chip

single cell level. An alternative to acquire single cell impedance spectrum, which requires a lower level in term of confinement, is to perform electrorotation experiment (Trainito et al. 2015). In that case the cell is submitted to a homogeneous propagative electric field, which drags the cell due to its polarization (Pohl 1958; Čemažar et al. 2013). In the case of the electrorotation, the propagative electrical field is rotational. The cell experiences a torque, which directly relates to its dielectric contrast with the medium (the contrast being characterized by the ! Claussius-Mossotti factor (Eq. 4)). Indeed this electrorotation torque Γdep depends on the imaginary part of fCM, on the squared modulus of the electric field, and on the volume of the cell (Morgan and Green 2003): ! ! Γdep ¼ 4πR3 em ℑ ðf CM ÞE2 z

(9)

Single Cell Electrical Characterization Techniques

11

!

where z is the unitary vector, normal to the plan of electrodes. Varying the electrical field angular frequency ω, a spectrum of the rotational velocity of the cell can be measured: Ωcell ¼

e 0 e m E2 ℑ ðf CM Þ 2η

(10)

where η stands for the dynamic viscosity of the medium. As seen from Eq. 10, the rotational velocity spectrum is proportional to the imaginary part of the Claussius-Mossotti factor ℑ( fCM), which is affected by the level of permeabilization (Eq. 4). The Fig. 7a points out the dependence of this spectrum on the cell membrane conductivity σ memb. This dependency is particularly obvious for the first peak (the positive one) of ℑ( fCM) function, i.e., for ω < 30 106 Rd/s. Indeed for these low frequencies, the complex permittivity of the mixture is mainly dependent on the membrane properties (ecell  Re ememb , as mentioned in the previous section). The second peak of ℑ( fCM) (the negative one) is sensitive to the dielectric properties of the intracellular compartment (σ in and ein), as seen in Fig. 7b. To estimate the dielectric parameters (ememb, σ memb, ein, σ in), the fitting between the measured experimental electrorotation spectrum and the theoretical spectrum that depends on these dielectric parameters can be achieved using least-square algorithms. In the case of a single cell electrorotation experiment, a strategy consists in (i) trapping the cell thanks to conventional dielectrophoresis forces using a set of electrodes inducing a stationary electrical field and (ii) to superpose to this stationary field a rotational field creating an electrorotation torque (Huang et al. 1992). From the velocity spectrum, the dielectric characteristics of the cell e*cell and finally the dielectric characteristics of the cell components (e*memb for the membrane and e*in for the averaged cytoplasm) can be estimated. This strategy was applied to monitor the permeabilization level of a single cell, trapped by a four-electrodes set, and submitted to such electrorotation torque (Trainito et al. 2015), see Fig. 8.

Microelectroporation of a Single Cell Thanks to the degree of miniaturization achievable with microfluidic technologies, the permeabilization can be localized precisely on the membrane of a single cell (concept of microelectroporation). This approach is a way to focus precisely the access to the cytoplasm and possible drug delivery, and should improve the cell viability, compared to the conventional methods for membrane permeabilization, as the permeabilized zone remains small. In this case, the objective is not the cell bioimpedance measurement but creating an access to the cytoplasm for DNA injection, drug delivery, or intracellular potential recording. The first example showing such an approach is described in Huang and Rubinsky (1999). In Huang

12

O. Français and B. Le Pioufle

a

Imaginary part of Claussius-Mossotti factor 0.6

ne i b ra at memabili z Perme

10−8 Sm −1 −5 10 Sm −1 10−4 Sm −1 −3 10 Sm −1

0.4

0.2

on

0

−0.2

−0.4 5 10

6

10

7

10

Angular frequency

b

8

10

9

10

10

10

ω (Rd/s)

Imaginary part of Claussius-Mossotti factor 0.5 initial low permittivity low conductivity

0.4 0.3

(fCM )

0.2 0.1 0 −0.1 −0.2 −0.3 −0.4 5 10

6

10

7

10

Angular frequency

8

10

9

10

10

10

ω (Rd/s)

Fig. 7 Dependance of the single cell electrorotation spectrum with the dielectric parameters of the cell. (a) The membrane permeabilization mainly affects the first peak of the spectrum, corresponding to the lowest frequencies. The imaginary part of Claussius-Mossotti factor is represented here, with a parametric variation of the membrane conductance (σ memb increasing from 108 Sm1 up to 103 Sm1). (b) The second peak of the characteristic is mostly sensitive to the dielectric properties of the intracytoplasmic compartment σ in and ein. The initial spectrum is represented (σ in = 1 Sm1, ein = 50e0, bold blue curve) as well as the low permittivity for the inner compartment (ein = 20e0, red dots curve) and the low conductivity for the inner compartment (σ in = 0.5 Sm1, light blue curve). For both curves (a) and (b), σ m = 0.3 Sm1 (low conductivity medium)

Single Cell Electrical Characterization Techniques

13

a

b Electrorotation spectrum - B16F10 cell 0.5

Rotational speed [rps]

0 -0.5 -1 -1.5 -2 -2.5 Fitted data BP Experimental data BP Fitted data AP Experimental data AP

-3 -3.5

104

105

106

107

108

Frequency [Hz]

Fig. 8 Influence of the electroporation on the electrorotation spectrum of a single cell. (a) Experimental protocol: the cell captured in the central area using conventional dielectrophoresis is submitted to the electrorotation torque. The electrorotation spectrum is measured before and after electroporation. (b) Experimental electrorotation spectra obtained before and after permeabilization of the single cell (lymphocyte). DEP dielectrophoresis, EROT electrorotation, PEF pulsed electric field, BP before pulses, AP after pulses (Adapted from Trainito et al. (2015), permission from John Wiley and Sons Publishing group, Ltd: Electrophoresis, copyright (2015))

and Rubinsky (2001) the cell is trapped by a microaperture etched within a thin silicon nitride membrane that also focuses the electrical field. The current through the microaperture (and thus through the patched membrane) is measured during the electroporation, which provides a way to monitor electrically the process of cell membrane electroporation. Another version achieved in PDMS (polydimethylsiloxane) was proposed in Khine et al. (2005), where the cell to be electroporated was trapped at the inlet of a tiny microchannel (section = few μm2, see Fig. 9). Real-time monitoring of the

14

O. Français and B. Le Pioufle

a

V/Vo 1

b

0.9 0.8 0.7 0.6 0.5 0.4 0.3

5 µm

0.2 0.1 0 1 0.9 0.8 0.7 V/Vo

0.6 0.5 0.4 0.3 0.2 0.1 0 10

20

30

40

Distance x/µm

Fig. 9 Concept of the microelectroporation. (a) The cell is trapped and immobilized in contact of a microaperture which enhances the transmembrane voltage in the patched zone of the membrane. V corresponds to the potential, while Vo refers to the potential difference applied to electrodes. (b) View of the cell trapped thanks to the horizontal channel micromolded in PDMS technology (Adapted from Khine et al. (2005), permission from Royal Society of Chemistry Publishing group, Ltd: Lab on Chip, copyright (2004))

electroporation, by means of the current measurement, was also achieved in a parallelized version of the device (Khine et al. 2007). Such concept is very similar to the patch-clamp technique (▶ Patch-clamp in Use of Electroporation Mechanisms Studies), firstly proposed by Neher et al. (1978) where the intracellular potential of neurons is measured by the use of a very thin glass pipette patching the cell membrane. The microelectroporation approach proposed in Khine et al. (2007) brings the benefit of possible parallelization which is a key progress for future drug screening applications. Anyway, the patch-clamp technique was later developed in its microfluidic version, known as the planar patch-clamp (Fertig et al. 2003), which is also compatible with the parallelization (Brüggemann et al. 2006). Such similarity in the structure of both types of devices (devices for microelectroporation and devices for the electrophysiological recording) led (Xie et al. 2012) to combine both functions with the same measurement electrodes. Indeed this device integrates nanopillars at the tip of recording electrodes, which are used to permeabilize the membrane very locally, and lets the contact to establish with the intracellular compartment for transmembrane potential recording (Fig. 10).

Single Cell Electrical Characterization Techniques

15

Fig. 10 Concept of the nanoporation: nanopillars are designed on the tip of recording microelectrodes (1.5 μm tall and 150 nm diameter). (c) Once the cell is placed above the electrodes, electric field pulses application leads to nanopore opening on the membrane, which opens transiently the access between the recording electrode and the intracellular medium (recording in (b) versus recording in (a)) (Adapted from Xie et al. (2012), permission from Nature Publishing group, Ltd: Nature Nanotechnology, copyright (2012))

Conclusion This chapter shows that the impedance measurement provides an interesting approach to characterize single cells. In particular the cell permeabilization can be monitored, even at the level of single cells, using therefore devoted microfluidic devices. The degree of miniaturization of such devices is a crucial parameter, as the volume fraction Φ directly influences the sensitivity of the measurement. The electrodes must thus be at the vicinity of the single cell to be characterized (typically the electrode gap should not exceed two times the cell diameter). Under adequate conditions, where the electrodes are in a strong vicinity with the cell, a precise monitoring of the evolution of cell electrical parameters can be achieved during the electroporation. To do so, one should use high-specific surface

16

O. Français and B. Le Pioufle

materials for the electrodes, like electroplated black platinum, as the influence of the double layer due to the interface between the medium and the electrodes might become critical when downscaling the device. Anyway, the influence of this doublelayer capacitance can be minimized using a differential measurement setup, probing a volume of medium without cells as a reference. Moreover, such a differential setup has to be considered in order to get rid of the temperature influence on the impedance measurement. High-throughput electrical characterization of cells can be envisioned in microfluidic devices, cells flowing one by one though the sensing area within a microchannel. In such a configuration, a specific electronics is required for fast and real-time analysis. Indeed the whole frequency spectrum cannot be acquired during the short time while the cell is flowing at the vicinity of electrodes, and a set of discrete excitation frequencies should thus be chosen to be injected simultaneously, prior to synchronous detection. Spreading the characterization spectrum towards high angular frequency is an important technological challenge, as it may permit to assess the dielectric properties of the intracellular content, providing increasing accuracy on the cell status determination. Promising works are ongoing with the high-frequency impedance spectrometry of single cells. Such progress will allow to establish an accurate dielectric signature of the cell as well as a marker of the cell content. In particular, it might be a promising and complementary way to characterize the effect of ultrashort electric field pulses [Measurement and Characterization of Exposure System for High-Frequency/Ultrashort Pulses] (like nanopulses), that permeabilize the intracellular lipid substructures (Schoenbach et al. 2001; Dalmay et al. 2012), as information might be provided on the dielectric properties modifications induced by such electrical treatment. Single cell electrical characterization can also be achieved thanks to electrorotation experiment, exploiting therefore the polarizability of cells immersed within a homogeneous rotary electrical field. Indeed the polarizability of the cell, in a given medium, and thus its mechanical response to a rotating electrical field, is related to the Claussius-Mossotti factor ( fCM, see Eq. 4), that characterizes the dielectric contrast between the cell and the surrounding medium, and depends on the dielectric properties of the cell components (ememb and σ memb for the membrane, ein and σ in for the intracellular compartment, see Eq. 7). Such an approach to estimate the dielectric properties of the cell components does not require high volume fraction Φ, and the required level of confinement is thus less critical. With this alternative way to characterize single cells, it might be also advantageous to increase the characterization spectrum high-frequency limit, as it might permit to assess the cell interior. Finally, the new possibilities in terms of miniaturization achievable with microand nanotechnology open very promising perspectives, like the possibility to target spatially the cell electroporation on a small portion of its membrane, and give electrical access to the cell cytoplasm. Such electrical treatment is less invasive, improving thus the cell viability after treatment. It shows to be a convenient way to transiently get an access through the membrane for action potential recording. The delivery of extracellular compounds through the membrane locally electroporated

Single Cell Electrical Characterization Techniques

17

thanks to this technique might be an interesting approach to be investigated in the near future. Acknowledgments This work has benefited from the financial support of the LabeX LaSIPS (ANR-10-LABX-0040-LaSIPS) managed by the French National Research Agency under the “Investissements d’avenir” program (nANR-11-IDEX-0003-02) and COST TD1104, www.electro poration.net. The authors acknowledge project supports from ENS Cachan-Université Paris Saclay, Institut d’Alembert, and CNRS.

Cross-References ▶ Different Approaches Used in Modeling of Cell Membrane Electroporation ▶ Measurement and Characterization of Exposure System for High-Frequency/ Ultrashort Pulses ▶ Patch-clamp in Use of Electroporation Mechanisms Studies ▶ Transmembrane Voltage Induced by Applied Electric Fields

References Brüggemann A, Stoelzle S, George M, Behrends JC, Fertig N (2006) Microchip technology for automated and parallel patch-clamp recording. Small 2(7):840–846 Bürgel SC, Escobedo C, Haandbæk N, Hierlemann A (2015) On-chip electroporation and impedance spectroscopy of single-cells. Sens Actuators B 210:82–90 Čemažar J, Miklavčič D, Kotnik T (2013) Microfluidic devices for manipulation, modification and characterization of biological cells in electric fields–a review. J Microelectron Electron Compon Mater 43(3):143–161 Dalmay C, De Menorval M, Francais O, Mir L, Le Pioufle B (2012) A microfluidic device with removable packaging for the real time visualisation of intracellular effects of nanosecond electrical pulses on adherent cells. Lab Chip 12(22):4709–4715 Fertig N, George M, Klau M, Meyer C, Tilke A, Sobotta C, Blick RH, Behrends JC (2003) Microstructured apertures in planar glass substrates for ion channel research. Receptors Channels 9(1):29–40 Gawad S, Schild L, Renaud P (2001) Micromachined impedance spectroscopy flow cytometer for cell analysis and particle sizing. Lab Chip 1(1):76–82 Haandbæk N, Bürgel SC, Heer F, Hierlemann A (2014) Characterization of subcellular morphology of single yeast cells using high frequency microfluidic impedance cytometer. Lab Chip 14 (2):369–377 Hoffman R, Britt W (1979) Flow-system measurement of cell impedance properties. J Histochem Cytochem 27(1):234–240 Holmes D, Pettigrew D, Reccius CH, Gwyer JD, van Berkel C, Holloway J, Davies DE, Morgan H (2009) Leukocyte analysis and differentiation using high speed microfluidic single cell impedance cytometry. Lab Chip 9(20):2881–2889 Huang Y, Rubinsky B (1999) Micro-electroporation: improving the efficiency and understanding of electrical permeabilization of cells. Biomed Microdevices 2(2):145–150 Huang Y, Rubinsky B (2001) Microfabricated electroporation chip for single cell membrane permeabilization. Sens Actuators A: Phys 89(3):242–249

18

O. Français and B. Le Pioufle

Huang Y, Holzel R, Pethig R, Wang XB (1992) Differences in the ac electrodynamics of viable and non-viable yeast cells determined through combined dielectrophoresis and electrorotation studies. Phys Med Biol 37(7):1499 Jones TB (2005) Electromechanics of particles. Cambridge University Press, Cambridge Khine M, Lau A, Ionescu-Zanetti C, Seo J, Lee LP (2005) A single cell electroporation chip. Lab Chip 5(1):38–43 Khine M, Ionescu-Zanetti C, Blatz A, Wang LP, Lee LP (2007) Single-cell electroporation arrays with real-time monitoring and feedback control. Lab Chip 7(4):457–462 Martinsen OG, Grimnes S (2011) Bioimpedance and bioelectricity basics. Academic Press, London Maxwell JC (1881) A treatise on electricity and magnetism, vol 1. Clarendon, Oxford Morgan H, Green N (2003) AC electrokinetics: colloids and nanoparticles. Research Studies Press, Philadelphia Morgan H, Sun T, Holmes D, Gawad S, Green NG (2006) Single cell dielectric spectroscopy. J Phys D Appl Phys 40(1):61 Neher E, Sakmann B, Steinbach JH (1978) The extracellular patch clamp: a method for resolving currents through individual open channels in biological membranes. Pflugers Arch 375 (2):219–228 Pavlin M, Kandušer M, Reberšek M, Pucihar G, Hart FX, Magjarevićcacute R, Miklavčič D (2005) Effect of cell electroporation on the conductivity of a cell suspension. Biophys J 88 (6):4378–4390 Pohl HA (1958) Some effects of nonuniform fields on dielectrics. J Appl Phys 29(8):1182–1188 Schoenbach KH, Beebe SJ, Buescher ES (2001) Intracellular effect of ultrashort electrical pulses. Bioelectromagnetics 22(6):440–448 Sun T, Morgan H (2010) Single-cell microfluidic impedance cytometry: a review. Microfluid Nanofluid 8(4):423–443 Trainito CI, Français O, Le Pioufle B (2015) Monitoring the permeabilization of a single cell in a microfluidic device, through the estimation of its dielectric properties based on combined dielectrophoresis and electrorotation in situ experiments. Electrophoresis 36(9–10):1115–1122 Xie C, Lin Z, Hanson L, Cui Y, Cui B (2012) Intracellular recording of action potentials by nanopillar electroporation. Nat Nanotechnol 7(3):185–190

3D Tissue Models to Bridge the Gap Between Cell Culture and Tissue in Assessing Electroporation Laure Gibot

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroporation Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2D Versus 3D Cell Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Human Versus Animal Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tissue Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spheroids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tissue Engineering Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 5 5 6 7 8 12 13 14 14

Abstract

The vast majority of studies on drug delivery by electroporation have been performed in vitro on cell cultures in 2D (cell monolayers or cells in suspension) and in vivo in animal models. The results obtained are often inconsistent. Threedimensional tissue models of increasing complexity are starting to be used to study mechanisms of electrotransfer of cytotoxic molecules and nucleic acids in a tissue context in vitro. These 3D models include collagen gel, normal and/or tumor spheroids, as well as human tissue substitutes produced by tissue engineering. Three-dimensional culture systems allow researchers to address the concept of cell organization in 3D and to take into account dynamic interactions between distinct cell types and surrounding extracellular matrix. Thus, 3D tissue models mimic biological tissue better than classical cell culture. Besides, this L. Gibot (*) Institut de Pharmacologie et de Biologie Structurale, Université de Toulouse, CNRS, UPS, Toulouse, France e-mail: [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_16-1

1

2

L. Gibot

approach is an efficient and ethical alternative to the use of laboratory animals. Moreover, cells can be isolated directly from a patient. This offers the possibility to adapt treatment planning to that person’s biological characteristics, leading to a potentially more effective and better tolerated therapeutic approach. Understanding and mastering the mechanisms of drug delivery by electrotransfer at the tissue scale will enable the development of new strategies to increase transfer efficiency while avoiding damage to surrounding tissue. Finally, this knowledge of mechanisms at the tissue scale will open the door of clinical applications to gene therapy and DNA vaccination and will promote the development of innovative, effective, and personalized therapies. Keywords

3D • Cell sheet • Gel • In vitro • Spheroid • Tissue engineering

Introduction Exposure of cells to high-voltage electric pulses causes membrane electroporation, or electropermeabilization, i.e., transiently permeabilizes the plasma membrane of cells. This phenomenon allows the entry of therapeutic molecules into cells and tissues. The technique is easy to implement, safe, and inexpensive and facilitates tissue targeting since only cells located between the electrodes are affected by the electric field. Multiple developments are made of the methodology in human clinics. Electroporation is a safe technique for transferring nucleic acids into different target organs and tissues, and gene electrotransfer (GET) is a promising method in gene therapy either for systemic secretion of proteins, for vaccination, or in relation with cancer therapy. Electroporation also ensures targeted and massive influx of cytotoxic molecules into electropermeabilized tumor cells. This clinical application, called electrochemotherapy (ECT), is used in many European countries for the treatment of skin cancer and subcutaneous tumors. By potentiating the cytotoxicity of drugs, ECT reduces the doses injected into patients and thus limits their side effects. Irreversible electroporation (IRE) is a novel nonthermal ablation mode based on electro-induced irreversible effect defects on cell plasma membrane. This method offers minimally invasive treatment for selected tumors. Optimizing therapeutic drug or nucleic acid delivery by electroporation is a challenging hurdle in preclinical studies. This is mainly due to differences between animal tissue characteristics and those of human cells as well as the lack of representativeness of conventional 2D cell culture. Since cell physiology is quite different in 2D versus 3D environments, the development of 3D cell culture systems is a real opportunity in biology to study cell behavior after application of stress. Cells classically cultivated in 2D do not provide an accurate representation of in vivo morphology, and cell responses obtained from these cultures are misrepresented in the absence of cell-cell and

3D Tissue Models to Bridge the Gap Between Cell Culture and Tissue in. . .

3

Table 1 3D in vitro model characteristics in the perspective of electroporation studies. ECM, extracellular matrix

Gel

Intercellular junctions No

Spheroid

Yes

Tissue engineering

Yes

Decellularized human tissue

Yes

Type of ECM Coculture cells No Yes Human/ animal No Yes Human/ animal Yes Yes Human/ animal Yes Yes Human

Ease of handling +

Ease of analyses +

Cost ++

+

+

+

++

++

++

++

++

++

cell-matrix interactions. Conventional 2D in vitro models are often considered as highly reductionist, jeopardizing the relevance of the preclinical efficacy data. The mechanisms involved in gene and drug electrotransfer in tissue remain largely unknown, particularly because of the complexity of the 3D architecture. For example, one of the main hurdles for in vivo gene electrotransfer remains the impaired diffusion of large molecules through tissue because of dense extracellular matrix and cell-cell junctions. The gold standard in term of 3D in vitro tissue model has to meet several criteria. Cells have to be organized in the three dimensions within a native extracellular matrix rich. This extracellular matrix should be a representative of in vivo situation, meaning composed of various class of molecules (fibers, glycosaminoglycans, glycoproteins) and not only one such collagen. Coculture of relevant cell types found in the tissue of interest has to be present. Cells should be able to develop intercellular junctions with neighboring cells. Human cell types are preferred. Its cost and time to produce should be reasonable. Its manipulation should be userfriendly. It should be analyzed easily both on fresh and fixed tissue, especially by microscopy techniques in order to visualize processes at cell scale. Nowadays, all the developed 3D models only meet partially these criteria (Table 1) meaning that more work has to be undertaken to propose more physiological 3D models. Compared to classical in vitro studies on monolayers or cells in suspension or in vivo on animal models, only few articles have been published on studies using in vitro 3D models. This chapter discusses 3D studies and underlines the great potential of 3D in vitro tissue models to reach a better understanding and control of drug and gene electrotransfer at tissue scale.

Electroporation Principle When the cell is exposed to high external electric field for sufficient time for plasma membrane charging, the cell membrane becomes permeable for ions, drugs, and molecules as large as plasmid DNA, which are otherwise impermeable. This physical phenomenon of cell permeabilization by external pulsed electric field application is named “electroporation.” If membrane permeability is transient, the cell

4

L. Gibot

survives and the membrane recovers its selective permeability, the electroporation is reversible; if the cell dies, the electroporation is irreversible. The mechanisms of molecule electrotransfer were mainly deciphered thanks to single-cell studies. Electroporation can be described as a three-step process. First, before electric field application, the plasma membrane acts as a physical barrier with selective permeability that prevents the free exchange of hydrophilic molecules between the cytoplasm and the extracellular compartment. Secondly, during electric field application, the transmembrane potential increases, leading to the formation of membrane defects, called pores, even if no experimental data support this hypothesis. These membrane defects allow the exchange of molecules both from the outside to inside of the cell and the opposite. Finally, after electric field application, membrane resealing occurs within about 20 min depending on the cell type. Small molecules such as propidium iodide or antitumor drugs have free access to the cytoplasm of electropermeabilized cells. Their transfer to the cytoplasm occurs both during the external application of the electric pulses and after, as long as the plasma membrane is not resealed. On the other hand, delivery of larger and charged molecules such as plasmid DNA requires distinct consecutive steps: migration of DNA toward the membrane, DNA insertion into the membrane, translocation across the membrane via endocytosis and cytoskeleton implication, DNA migration to the nucleus, and finally transfer of DNA across the nuclear envelope for effective transcription. DNA migration toward the membrane is dependent on DNA electrophoresis, thus trains of long pulses lasting several milliseconds are generally applied for nucleic acid delivery. Due to its selectivity both in terms of time and space, electroporation enabled a development of multiple applications into medical and industrial processes: cancer treatment, gene therapy, wound healing, bacterial disinfection, and electroextraction in agro-industry. Despite their wide range of practical applications in medicine, biotechnology, and process industry, the fundamental mechanisms of cell membrane electroporation are not well understood at tissue scale. Since several decades, a large number of parameters such as electric field intensity, pulse number and duration, frequency of application, media conductivity, temperature, pH, and surface charge have been experimentally investigated, particularly in single-cell context, and theoretically modeled by different research groups. These theoretical developed models are capable of simulating the electroporation phenomena for the simple cases of unique cells in suspension of in monolayer. However, they often fail to predict the cell electroporation phenomena in complex environment such as tissues, where cells don’t display a spherical shape and are embedded in the complex extracellular matrix with nonhomogeneous electrical properties. Furthermore, tissue properties such as perfusion, cell density, and orientation can affect the electric conductivity of the tissue. While theoretical models of cell electroporation in tissues exist, they incorporate only the information about the cells and completely ignore the impact of other tissue components such as extracellular matrix. At tissue scale, the challenge is therefore to decipher the impact of physicochemical properties of extracellular matrix and cell organization in 3D on electric field distribution inside the tissue and cell membrane electroporation

3D Tissue Models to Bridge the Gap Between Cell Culture and Tissue in. . .

5

process. This knowledge gap results in empirical approaches for protocol development with unpredictable electroporation success rates in basic experiments and clinics.

2D Versus 3D Cell Environment Although conventional cell culture monolayers or cells in suspension have provided important scientific and conceptual advances, it is obvious that 2D cell systems lose the architecture of the tissue from which they originate. They also lack mechanical and biophysical signals and intercellular communication (Pampaloni et al. 2007). The 3D environment plays a key role in cell and tissue homeostasis. The cells communicate dynamically and plastically with each other and with the surrounding stroma. It is this permanent and reciprocal exchange between the cells and their extracellular environment which maintains the architecture and the functional organization of the tissues (Fig. 1). Thus, several major studies have shown that the dynamic and reciprocal interactions between the extracellular matrix and the nucleus, mediated by the cytoskeleton, coordinate cellular functions, architecture, and environmental tissue homeostasis (Nelson and Bissell 2006; Zaman 2013). Thus, the concept of “dynamic reciprocity” proposed by Dr Mina Bissell lays on the reciprocal link between the extracellular matrix and the nucleus of a cell where the extracellular compartment provides biological, chemical, and physical signaling cues to maintain tissue integrity. Indeed, complex environmental interactions and forces contribute profoundly to the behavior, phenotype, and evolution of both normal and tumor cells. Since the extracellular compartment dictates cell behavior via a myriad of controlled signals, researchers have to develop ways to mimic these functions in vitro in order to recapitulate tissue functions in both 3D space and over time. In order to study delivery of drugs by electroporation in a human tissue context, it is of utmost importance to produce in vitro biologically relevant 3D models of tissue. These powerful tools have to reproduce biological, biochemical, and biophysical constraints in vitro related to the 3D architecture of human tissue and thus will be much closer to reality than cell culture monolayers or animal models.

Human Versus Animal Model The skin is the preferred tissue for studying the consequences of therapeutic drug electrotransfer, not only because of its accessibility and knowledge gained by clinicians with electrochemotherapy but also because it is rich in presenting cells and thus has immunological properties essential for the development of DNA vaccination (Gothelf and Gehl 2010). The choice of a study model is based on many factors such as availability, cost, ease of operation, and especially on its anatomic/functional similarity to human tissue. Even though porcine skin has the closest anatomy to that of humans, the cost

6

L. Gibot

Fig. 1 Dynamic interactions between cells and extracellular environment in a tissue context: notion of mechanobiology. In a tissue context, there is a physical link between the extracellular matrix and the nucleus via membrane receptors such as integrins and the cytoskeleton. Communication is dynamic and reciprocal since the extracellular environment regulates cell function and gene expression which modulates the architecture of the environment (Reproduced with permission from Macmillan Publishers Ltd: Nature Reviews Molecular Cell Biology, (Wang et al. 2009), Copyright (2009))

per animal and for housing facilities explains why rodents are conventionally used for studies on gene and drug electrotransfer in the skin. However, murine skin differs from human skin in its architecture and its cellular composition. In addition, a transcriptomic study shows that only 30 % of genes associated with the skin are identical between human and mouse (Gerber et al. 2014). The development of tissue engineering for the purposes of regenerative medicine has led to different models of relevant human tissues.

Tissue Engineering Tissue engineering can replace, restore, maintain, or improve the function of human tissues through the laboratory production of biological substitutes for transplantation. In addition to these clinical applications, tissue engineering is also an invaluable tool in basic research. The reconstructed substitutes must provide a biological, biochemical, and biophysical model in which cells survive, differentiate, and organize themselves and are able to perform their biological functions (Badylak et al. 2012). A number of approaches are based on the use of materials (scaffolds), organic or not, to provide a 3D structure yielding specific biological and mechanical properties (Fig. 2).

3D Tissue Models to Bridge the Gap Between Cell Culture and Tissue in. . .

7

Fig. 2 Schema of the three fundamental components of 3D tissue models. Cell culture methods, characteristics of the scaffold, and relevance of biochemical and mechanical signals are primordial to produce a 3D bioactive construct

Gels The extracellular matrix (ECM) is the major component of the 3D cellular environment providing mechanical and biochemical support for cells in a tissue. In order to mimic tissue architecture more closely in vitro, some authors embedded cells in a 3D gel made with extracellular matrix components. Indeed, cells cultured in 3D matrices meet a more physiological environment relative to classical 2D monolayers, especially concerning biomechanical stimulation. Thus, 3D matrices are thought to provide a tissue-like environment. Collagen type I is the most abundant protein in mammals and fulfills a variety of mechanical functions depending on the tissue in the body. Interestingly, its mechanical properties are generally adapted by a modification of the hierarchical structure rather than by a different chemical composition. Collagen I, mostly of bovine and rat tail origin, is the most widely used extracellular matrix protein for 3D cell culture in gel. By providing in vivo-like extracellular matrix structures, it facilitates cell attachment, growth, differentiation, migration, and tissue morphogenesis. Besides its biological relevance, collagen I presents the advantage of being commercially available, even if quite expansive. Collagen is very viscous and a considerable amount gets lost during pipetting for gel preparation. All the protocols last approximately 3 h to produce a usable gel. Generally, collagen comes in an acidic solution and needs to be neutralized with NaOH and made isotonic with 10X PBS before mixing with cells. All the solutions used have to be kept on ice until the cells have been added in order to prevent collagen from gelling. Cells in suspension are then added to the collagen mix. Once placed at 37  C, the polymerization can take place.

8

L. Gibot

A fast polymerization allows optimal cell distribution in the gel. The major drawback of the gel approach is that when seeded with cells, especially fibroblasts, they will contract because of cell remodeling. This means that gels can be analyzed only for days after treatment. Haberl et al. embedded CHO cells in a type I collagen gel and studied the electroporation efficiency with propidium iodide uptake as well as the effects of plasmid concentration and pulse duration on gene electrotransfer (GET) efficiency (Haberl and Pavlin 2010). They confirmed that gene electrotransfer efficiency is comparable to that observed in in vivo studies and depends on plasmid-DNA concentration and pulse duration. Cells grown in agarose gel were recently used to underline the electrosensitization role in cell ablation by a nanosecond pulsed electric field (Muratori et al. 2016). Electrosensitization corresponds to a delayed increase of sensitivity to electroporation for cells that have already been subjected to electroporation. Muratori et al. showed that local ablation was obtained when trains of pulses were split into two identical fractions. They demonstrated that 3D gels containing cells enabled cell electrosensitization to be visualized and were a useful tool to study electro-assisted cell ablation. Thus, electrosensitization may help normal or tumor tissue ablation by reducing the voltage or number of pulses required or by increasing inter-electrode distance without losing ablation efficiency. Tumor cells grown in 3D collagen gels were shown to be a relevant model for studies on irreversible electroporation (IRE)-based cancer therapies and help to improve treatment planning without the use of animal models (Arena et al. 2012). Irreversible electroporation is a nonthermal, local ablation technique that is highly promising in cancer therapy. In order to induce cell death, series of short-duration, high-intensity electric pulses are applied to the targeted tissue. Collagen I hydrogelbased bioengineered tumors were used to generate data for refinement of the algorithms necessary to elucidate electric field distribution in tissue, which forms the basis of current treatment planning algorithms for in vivo applications. Indeed, electrical and thermal fluctuations were monitored during treatment (Fig. 3). Thus, 3D cell culture models can be used for more accurate modeling of IRE-based cancer therapies in an in vitro experimental setting, which will help to spread the clinical use of this method.

Spheroids While the gel strategy uses a biomaterial as a scaffold to provide a 3D structure to the model, the spheroid approach is devoid of exogenous material. In the 1980s, Sutherland et al. proposed a multicellular 3D model named the spheroid (Sutherland 1988). This 3D model is based on cell aggregation properties and exhibits numerous intercellular junctions. Several techniques can be applied to promote spheroid formation (Hirschhaeuser et al. 2010), but two are predominantly used: the hanging drop technique or the seeding of cells in nonadherent round-bottom 96-well plates. In the hanging drop

3D Tissue Models to Bridge the Gap Between Cell Culture and Tissue in. . .

9

Fig. 3 IRE pulse delivery in collagen I gel seeded with tumors cells. (a) Picture of the experimental setup. (b) Geometry and mesh used in the finite element model for simulating the electrical and thermal response of the in vitro tumors to IRE (Reprinted with permission from (Arena et al. 2012), Copyright (2012), with permission from Elsevier)

technique, wells are coated with agarose in order to prevent cell attachment. Then, a small drop (usually 20 μl) of cell culture medium containing 500 cells is deposited on the interior face of the plate lid. When the lid is placed onto the plate, the drop hangs over the agarose-coated well. After 3 days, thanks to gravity, the cells started to attach together and form a small spheroid that has to be transferred into the agarose-coated well in order to grow for 2 more days in large volume of cell culture medium. This technique presents some drawbacks: Some spheroids are lost during the transfer phase; spheroid diameters are not homogenous between all the wells; it is time consuming. The other classical used technique is the growth of cells in nonadherent environment. For this purpose, some plates are conveniently commercially available. Homemade nonadherent plates can be produced using a coating with poly(2-hydroxyethyl methacrylate) (poly-HEMA). The main advantage of the nonadherent approach is that spheroids can be produced in 96-well plates, and they are highly homogenous in size. Usually, spheroids are used around 5 days after seeding and can be followed and analyzed for weeks after treatments. Interestingly, both normal and tumor cells were found to form spheroids. In large proliferating tumor spheroids, gradients of oxygen, nutrients, catabolites, and proliferation are observed, mimicking those established in vivo in avascular regions of solid tumors (Fig. 4). Several publications have demonstrated that a 3D spheroid model is a relevant tool to study and optimize DNA electrotransfer (for review see (Gibot and Rols 2013)). Canatella et al. paved the way for studies on electrotransfer of small molecules in 3D spheroid structures (Canatella et al. 2004). Using calcein, they showed that cellular uptake within a spheroid is spatially heterogeneous after electroporation, with a decreasing gradient from the surface to the core. Indeed, an increased diffusion lag time is required for small molecules to homogeneously diffuse within the spheroid. Besides, dense cell populations exhibit heterogeneous electric properties, which modulate the cell response to electric field application. Finally, depending on their location within the spheroid, cells display a variation in size. Thus, cell

10

L. Gibot Spheroid Characteristics

ate n ctr atio ) La mul l/g cu mo µ ac (

> 10.5 ... ... ... ... ... 1.5 - 3.0 0.0 - 1.5

nt

die

en

a gr

yg

Ox

g)

(

H mm

101 - 120 ... ... ... ... 1 - 20 0

e os n uc utio l G rib t dis

ks ea ) br osis d r an ec str is/n A os lei DN opt uc EL n p N (a e = TU blu n = e gre

g)

l/ mo

(µ

on

uti

P AT

d

rib ist

)

l/g

mo (µ

> 2.1 ... ... ... ... ... 0.3 - 0.6 0.0 - 1.3

> 21 ... ... ... ... ... 3-6 0-3

ry da on osis c se ecr n s y/ l el og c l o e as rph tion ph mo fera So t li s Hi Pro

Fig. 4 Schematic representation of gradients established in a growing tumor spheroid. Dotted line delimits the necrotic core (Reprinted with permission from (Hirschhaeuser et al. 2010), Copyright (2010), with permission from Elsevier)

radius in the core was found to be 19 % smaller than at the surface. Spheroids seem to be a suitable model to reproduce, in vitro, the cell heterogeneity found in tissue. Comparing cell suspensions and spheroids, Chopinet et al. demonstrated that, for the same level of electropermeabilization, GET efficiency was 25-fold higher in cells in suspension than in cells organized in 3D (Chopinet et al. 2012). Furthermore, efficiently transfected cells were solely located at the spheroid surface, especially on the side facing the negative electrode (Fig. 5; Mellor et al. 2006; Wasungu et al. 2009), stressing that large molecule diffusion is a hurdle in GET efficiency, as it is in an in vivo context. Indeed, cell density and contacts between cells (intercellular junctions) seem to limit/prevent uniform DNA distribution. Melanoma spheroids produced in a microgravity environment have proven to be a relevant and useful tool to study gene electrotransfer following intra-spheroid plasmid DNA injection (Marrero and Heller 2012). Large-scale spheroids of

3D Tissue Models to Bridge the Gap Between Cell Culture and Tissue in. . .

11

Fig. 5 GET in spheroid. Cells efficiently electrotransfected with GFP-encoding DNA plasmid (green) are only located at the spheroid surface. (a) Transmitted light images of the spheroid. (b, c) Fluorescent images obtained by confocal microscopy (Reprinted with permission from (Wasungu et al. 2009), Copyright (2009), with permission from Elsevier)

human keratinocytes were injected with melanoma cells before in vitro determination of optimal electric parameters that allow GFP-reporter gene as well as IL-15 gene of interest efficient electrotransfer. Interestingly, these mixed spheroids exhibited distinct sensitivity to transfection between normal cutaneous cells and melanoma tumor cells. Finally, this model demonstrates the predictability of cell response to GET and may facilitate future developments of clinical trials for cancer therapies in vivo. Tumor spheroids in vitro were shown to effectively reproduce processes observed in vivo with electrochemotherapy (ECT) (Gibot et al. 2013), even if some limitations are still present, such as lack of blood circulation or immune system response. Indeed, the concomitant application of an electric field with anticancer drugs conventionally used in ECT, namely bleomycin and cisplatin, results in the total destruction of the spheroid by cell death within 5 days after treatment. This faithfully reproduces results obtained in vivo on small animals. Furthermore, the tumor spheroid was demonstrated to be an effective tool for in vitro screening of new candidate drugs for ECT. Finally, this cell model in 3D will permit adaptation of therapeutic planning to the biological profile of each patient and therefore develop personalized medicine, which is more effective and better tolerated. Indeed, most medical treatments are designed for the average population of patients. Even if it is successful for some patients, it is not for others. Personalized medicine is a promising way to disease prevention and treatment that takes into account differences in patient’s genes, environments, and lifestyles. The final aim is to select treatments that improve chances of survival while reducing adverse effects. Interestingly, mixed spheroids composed of both normal and tumor human cells were shown to be an efficient tool to reproduce in vitro phenomena observed in vivo by clinicians after electrochemotherapy (Gibot et al. 2016). Indeed, after ECT, clinicians observed good tissue healing and excellent esthetic and functional

12

L. Gibot

recovery at the tumor site, meaning that normal tissue located between the electrodes was not affected by the treatment. Experimentally, normal cells within mixed spheroids are spared when treated with ECT. Recently, mixed spheroids also indicated differential effects in normal and malignant cell lines with calcium electroporation (Frandsen et al. 2015). In conclusion, coculture of different cell types within mixed spheroids will better mimic what happens in vivo in tissue submitted to drug electrotransfer.

Tissue Engineering Approaches A number of approaches such as 3D gels are based on the use of exogenous materials as scaffolds to provide a 3D structure and specific mechanical properties to their models. Spheroid models are based on the development of intercellular junctions, but they lack endogenous extracellular matrix. Innovative tissue-engineered approaches focus on the use of living cells secreting their own extracellular matrix. This scaffoldless technique is named the self-assembly method (Athanasiou et al. 2013). Thus, cells stimulated with ascorbic acid, also named vitamin C, can synthesize their own endogenous extracellular matrix rich in native components such as collagens and form a handleable cell sheet. Indeed, ascorbic acid is a cofactor of prolyl hydroxylase and lysyl hydroxylase. These enzymes catalyze the hydroxylation of proline and lysine in hydroxyproline and hydroxylysine, respectively. These hydroxylases, localized in the endoplasmic reticulum of cells, are essential to the intracellular assembly of the collagen triple helix. Indeed, hydroxyproline residues in the procollagen sequence ensure the tertiary structure in triple helix of mature collagen. Over 35 % of procollagen proline residues have to be hydroxylated in order to maintain the collagen triple-helix conformation at the physiological temperature of 37  C. Furthermore, ascorbic acid was shown to act at the level of gene expression by increasing the transcription of collagen I gene and the stabilization of its mRNA. It also increases the excretion rate of extracellular collagen. During production, self-assembled tissue models display biological processes similar to those that occur in vivo. The self-assembly approach leads to reconstructed tissue with a morphological structure and organization close to those of in vivo tissues. Human dermal substitutes have been produced by the self-assembly approach and used to understand the role of each parameter of the experimental setup on the distinct cell types present within the skin during drug electrotransfer. These models were efficiently electropermeabilized with application of millisecond electric pulses (Madi et al. 2015). They are valuable tools to study GET mechanisms in human tissue since they correctly reproduce limiting steps found in vivo such as plasmid DNA mobility within a tissue dense in extracellular matrix (Madi et al. 2016). Thus, 3D tissue models produced by the self-assembly approach appear to be promising to determine optimal electrical parameters for distinct cell types in a tissue. In order to develop 3D tissue models ever closer to native human tissue, a promising and innovative strategy is to use decellularized human tissues. Once

3D Tissue Models to Bridge the Gap Between Cell Culture and Tissue in. . .

13

Fig. 6 Intradermal injection within a recellularized human skin graft. Grafts are easily manipulated. Injection of green food coloring was performed to visualize the location of DNA injection within the dermis (Reproduced from (Bulysheva et al. 2016))

recellularized with human cells, they provide a valuable tool to gain ex vivo assess to drug electrotransfer mechanisms. For example, human skin grafts colonized with human cutaneous cells (keratinocytes and fibroblasts) were efficient ex vivo human models for optimizing gene expression which is a challenging hurdle in preclinical studies (Bulysheva et al. 2016). Interestingly, mechanical integrity and thickness of the tissue allow for its manipulation with forceps and for intradermal injections of plasmid DNA or therapeutic drugs (Fig. 6). In conclusion, this recellularized skin graft structurally and functionally resembled native human skin in tissue histological organization, proving an effective 3D human skin model for preclinical drugs and gene delivery studies.

Conclusion Most studies on therapeutic molecule electrotransfer, either as part of electrochemotherapy or of plasmid DNA vectorization, are carried out on 2D monolayer cultures of cells, or in animal models. These approaches do not allow the putative roles of tissue architecture (extracellular matrix, cell-cell junctions) to be deciphered in drug electrotransfer at the tissue scale. 3D models may be carefully built up in complexity in terms of spatial positioning of different cell types, controlling matrix density, and composition. Such improved biomimetic models are the link between the cells cultivated in 2D and the whole animal body. However, it has to be kept in mind that even the more sophisticated 3D in vitro model will present some limits compared to the human body. Until today, no fluid perfusion was achieved in in vitro produced tissue substitute. Thus, 3D models lack blood and lymphatic circulation. Besides, even if some immune cell types can be included in studies in 3D in vitro, no functional immune system is today available. Finally, despite the development of

14

L. Gibot

“organ-on-chip” technologies, it remains difficult to in vitro reproduce dialogues and communication existing between different tissues and organs. In 1959, Russell and Burch published “The Principles of Humane Experimental Techniques”, and the basic tenet of their report is that “the humanest possible treatment of experimental animals, far from being an obstacle, is actually a prerequisite for a successful animal experiment.” Thus, they proposed the 3Rs’ ethical rule which consist in replacement, reduction, and refinement of animal experiments. Recently, a fourth R was added to this chart for responsibility of the experimenter. Consequently, as a part of the implementation of the 4Rs’ ethical rule, 3D in vitro models are a reasonable alternative to the use of laboratory animals. 3D models can potentially be reconstructed with cells from a patient, which would allow the therapeutic protocol to be adapted to the biological characteristics of the patient. Thus, this approach will permit the development of efficient and personalized medicine. Finally, the exploration of nonviral vectorization of therapeutic drugs in a human tissue context will not only bring many relevant answers on its fundamental mechanisms but also promote crucial advances for medical applications. Thus, a better understanding and control of the electroporation mechanisms at the tissue scale will improve the targeting and effectiveness of drug transfer, whether it is anticancer molecules in electrochemotherapy, genetic information for DNA vaccination, or gene therapy.

Cross-References ▶ 3D Cryogenic Electrospun Scaffolds to Assess Tissue Responses to Electroporation ▶ Electrochemotherapy and Its Clinical Applications ▶ Electroporation and electropermeabilization ▶ Human Spheroids to Assess Electroporation ▶ Irreversible Electroporation and Its Clinical Applications ▶ Principles of Electroporation for Gene Therapy ▶ Tissue Ablation by Irreversible Electroporation ▶ Tissue Engineering with Electroporation

References Arena CB et al (2012) A three-dimensional in vitro tumor platform for modeling therapeutic irreversible electroporation. Biophys J 103(9):2033–2042 Athanasiou KA et al (2013) Self-organization and the self-assembling process in tissue engineering. Annu Rev Biomed Eng 15:115–136 Badylak SF et al (2012) Engineered whole organs and complex tissues. Lancet 379(9819):943–952 Bulysheva AA et al (2016) Recellularized human dermis for testing gene electrotransfer ex vivo. Biomed Mater 11(3):035002

3D Tissue Models to Bridge the Gap Between Cell Culture and Tissue in. . .

15

Canatella PJ et al (2004) Tissue electroporation: quantification and analysis of heterogeneous transport in multicellular environments. Biophys J 86(5):3260–3268 Chopinet L, Wasungu L, Rols MP (2012) First explanations for differences in electrotransfection efficiency in vitro and in vivo using spheroid model. Int J Pharm 423(1):7–15 Frandsen SK et al (2015) Calcium electroporation: evidence for differential effects in normal and malignant cell lines, evaluated in a 3D spheroid model. PLoS One 10(12):e0144028 Gerber PA et al (2014) The top skin-associated genes: a comparative analysis of human and mouse skin transcriptomes. Biol Chem 395(6):577–591 Gibot L, Rols MP (2013) Progress and prospects: the use of 3D spheroid model as a relevant way to study and optimize DNA electrotransfer. Curr Gene Ther 13(3):175–181 Gibot L et al (2013) Antitumor drug delivery in multicellular spheroids by electropermeabilization. J Control Release 167(2):138–147 Gibot L et al (2016) Mixed spheroids as a relevant 3D biological tool to understand therapeutic window of electrochemotherapy. In: Jarm T, Kramar P (eds) 1st world congress on electroporation and pulsed electric fields in biology, medicine and food & environmental technologies. Springer, Singapore, pp 200–203 Gothelf A, Gehl J (2010) Gene electrotransfer to skin; review of existing literature and clinical perspectives. Curr Gene Ther 10(4):287–299 Haberl S, Pavlin M (2010) Use of collagen gel as a three-dimensional in vitro model to study electropermeabilization and gene electrotransfer. J Membr Biol 236(1):87–95 Hirschhaeuser F et al (2010) Multicellular tumor spheroids: an underestimated tool is catching up again. J Biotechnol 148(1):3–15 Madi M, Rols MP, Gibot L (2015) Efficient in vitro electropermeabilization of reconstructed human dermal tissue. J Membr Biol 248(5):903–908 Madi M, Rols MP, Gibot L (2016) Gene electrotransfer in 3D reconstructed human dermal tissue. Curr Gene Ther 16(2):75–82 Marrero B, Heller R (2012) The use of an in vitro 3D melanoma model to predict in vivo plasmid transfection using electroporation. Biomaterials 33(10):3036–3046 Mellor HR et al (2006) Optimising non-viral gene delivery in a tumour spheroid model. J Gene Med 8(9):1160–1170 Muratori C et al (2016) Electrosensitization assists cell ablation by nanosecond pulsed electric field in 3D cultures. Sci Rep 6:23225 Nelson CM, Bissell MJ (2006) Of extracellular matrix, scaffolds, and signaling: tissue architecture regulates development, homeostasis, and cancer. Annu Rev Cell Dev Biol 22:287–309 Pampaloni F, Reynaud EG, Stelzer EH (2007) The third dimension bridges the gap between cell culture and live tissue. Nat Rev Mol Cell Biol 8(10):839–845 Sutherland RM (1988) Cell and environment interactions in tumor microregions: the multicell spheroid model. Science 240(4849):177–184 Wang N, Tytell JD, Ingber DE (2009) Mechanotransduction at a distance: mechanically coupling the extracellular matrix with the nucleus. Nat Rev Mol Cell Biol 10(1):75–82 Wasungu L et al (2009) A 3D in vitro spheroid model as a way to study the mechanisms of electroporation. Int J Pharm 379(2):278–284 Zaman MH (2013) The role of engineering approaches in analysing cancer invasion and metastasis. Nat Rev Cancer 13(8):596–603

Cell Stress Responses to Pulsed Electric Fields Ken-ichi Yano and Keiko Morotomi-Yano

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Overview of Stress Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General Principles of Stress Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Integrated Stress Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alternative Mechanism for Stress-Induced Translational Suppression . . . . . . . . . . . . . . . . . . . . . . Stress Responses Induced by Nanosecond Pulsed Electric Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nanosecond Pulsed Electric Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Activation of eIF2α Phosphorylation by nsPEFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Activation of Stress-Responsive Protein Kinases by nsPEFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effects of nsPEFs on 4E-BP1 and AMPK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Suppression of Protein Synthesis by nsPEFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nucleofection-Induced Stress Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Possible Effects of Electroporation on Cell Physiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 3 4 6 7 7 8 11 12 12 13 14 15 15 16

Abstract

Cells are often exposed to unfavorable conditions and stimuli, which are collectively referred to as stress. To adapt to stress, cells induce a set of reactions, called stress responses, that lead to suppression of translation initiation. Protein synthesis requires considerable amounts of energy and amino acids, and hence stressinduced translational suppression preserves cellular resources and serves as a mechanism for survival. This chapter is aimed at providing an overview of the stress responses induced by pulsed electric fields (PEFs). PEFs are utilized in a broad range of the life sciences, owing to their action on the cell membrane. K.-i. Yano (*) • K. Morotomi-Yano Institute of Pulsed Power Science, Kumamoto University, Kumamoto, Japan e-mail: [email protected]; [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_17-1

1

2

K.-i. Yano and K. Morotomi-Yano

Nanosecond PEFs (nsPEFs) generate small membrane pores that permeate small molecules, such as ions, and longer PEFs yield larger membrane pores suited for DNA transfection and tumor chemotherapy. Irrespective of the pulse width, PEFs are essentially deleterious to cellular homeostasis, because they compromise the membrane integrity and perturb the balance between intra- and extracellular molecules across the cell membrane. Recently, nsPEFs have been shown to induce stress responses in human and mouse cells. nsPEFs elicit stress-responsive signal transduction that comprises two protein kinases, namely, PERK and GCN2, and their substrate eIF2α, leading to translational suppression by eIF2α phosphorylation. nsPEFs affect another signal transduction that controls translation initiation by mTORC1-mediated 4E-BP1 phosphorylation. Collectively, accumulating evidence supports the concept of PEF-induced stress responses that appear to have profound effects on cellular functions such as gene expression and cell death induction.

Keywords

Stress response • Signal transduction • Protein kinase • Protein synthesis • Translational suppression • Nanosecond pulsed electric fields • Nucleofection • Electroporation

Introduction Living cells often encounter physical and chemical stimuli from their external environment and respond to them by evoking intracellular reactions that ultimately lead to changes in cellular physiology and behavior. Cellular responses to external stimuli generally involve the induction of intracellular signaling, which is mediated by sequential protein phosphorylation and is transduced to downstream effector proteins. Consequently, the external stimuli cause alterations in various cellular activities, including gene expression, metabolism, proliferation, differentiation, and cell death induction (Alberts et al. 2014). Among the diverse external stimuli, those with adverse impacts on the cell are referred to as cellular stress. Human cells have an ability to adapt to cellular stress by inducing a distinct set of intracellular reactions that are collectively called stress responses. The stress responses play critical roles in the maintenance of homeostasis under adverse circumstances. Dysfunctions of the stress responses are known to be associated with many diseases, such as inflammation and neurodegenerative diseases, thereby implicating the physiological significance of stress responses (Leprivier et al. 2015). Pulsed electric fields (PEFs) are widely used in the life sciences, because different effects on the cell membrane can be achieved in a pulse width-dependent manner. PEFs on the order of ms to μs are suited for macromolecule transfer, and these PEFs are widely used for DNA transfection (▶ Nucleic Acid Electrotransfer in Mammalian Cells: Mechanistic Description) and electrochemotherapy (▶ Overview and

Cell Stress Responses to Pulsed Electric Fields

3

History of Electrochemotherapy) (Rems and Miklavcic 2016). Nanosecond PEFs (nsPEFs) produce small membrane pores that allow permeation of small molecules, such as ions (Vernier et al. 2006; Pakhomov et al. 2009). PEF-induced membrane permeabilization allows movement of extracellular as well as intracellular molecules across the cell membrane and consequently perturbs cellular homeostasis. Accordingly, PEFs are essentially deleterious for cellular physiology and could be considered as cellular stress. This chapter is intended to provide an overview of the stress responses induced by PEFs. Owing to very limited information currently available on the stress responses evoked during electroporation used for macromolecule transfer, particular emphasis is placed on the cellular responses to nsPEFs and nucleofection. Exposure of human cells to nsPEFs is known to elicit various intracellular reactions, and molecular details of the nsPEF-induced stress responses have been previously demonstrated (Morotomi-Yano et al. 2012b). Nucleofection is a sort of electroporation that is conducted under undisclosed electrical conditions optimized for individual cell types (Gresch et al. 2004). A previous study has shown the induction of stress-responsive signaling by nucleofection (Anderson et al. 2013). In this chapter, first, the general mechanisms of cellular stress responses in human cells are outlined. Second, the nsPEF-induced stress responses are described in detail. An outline of nucleofectioninduced reactions is also given. Finally, possible physiological implications of the PEF-induced stress responses are discussed. Although this chapter focuses on cellular responses in human and mammalian cells, other chapters (▶ Responses of Plant Cells and Tissues to Pulsed Electric Field Treatments; ▶ Stress Response of Plants, Metabolite Production to Pulsed Electric Fields) provide the outline of stress responses in plant cells.

Overview of Stress Responses General Principles of Stress Responses Cells can sense environmental changes and external stimuli and transmit them through intracellular signal transduction that is mainly mediated by a cascade of protein phosphorylation. Protein phosphorylation, a covalent modification with a phosphate group at specific sites, can alter protein conformation, causing either an increase or a decrease in protein activity. Activation of signal transduction results in the phosphorylation of downstream effector proteins, leading to alterations in various cellular functions and physiology. Figure 1 shows a simplified model for stress responses. When cells encounter stressed conditions, phosphorylation-mediated signal transduction is rapidly induced, leading to attenuation of translation initiation and consequent suppression of protein synthesis. Because protein synthesis consumes large amounts of energy and materials, its suppression saves biological resources and thereby serves as a survival mechanism. In addition to translational suppression, stress-induced signal transduction frequently causes altered expression of a specific set of genes, which

4

K.-i. Yano and K. Morotomi-Yano

Fig. 1 Conceptual model for stress response. When the cell encounters adverse environment and stimuli, a stress-sensing protein triggers signal transduction that is mediated by a cascade of protein phosphorylation. Stress-induced signal is transduced to downstream effector proteins and causes attenuated translation initiation and consequent suppression of protein synthesis. Stress-induced signal transduction often leads to alterations of gene expression

contributes to adaptation of cells to stressed conditions. Although Fig. 1 shows a simplified conceptual model for stress responses, human and mammalian cells have multiple mechanisms, in which many cellular proteins are involved as described below.

Integrated Stress Response Outline of Integrated Stress Response Cells have evolved elaborate mechanisms to sense changes in the extracellular environment. When encountering stress, cells induce distinctive signal transduction that attenuates an overall rate of protein synthesis by inhibiting the activity of the eukaryotic initiation factor 2 (eIF2). The stress-induced signal transduction and consequent translational suppression involving the eIF2 inhibition are collectively called the integrated stress response, because cellular responses to diverse forms of stress converge to these reactions (Fig. 2) (Leprivier et al. 2015). The integrated stress response is evolutionarily conserved in a wide range of eukaryotes from yeasts to humans, and translational suppression is a central event in this process. Under normal physiological conditions, cells continuously undergo protein synthesis, which requires substantial amounts of energy and materials. Translational suppression is considered to be beneficial for stressed cells, because it saves cellular resources (Leprivier et al. 2015). Although the integrated stress response generally serves as a prosurvival mechanism, its prolonged activation exerts facilitative effects

Cell Stress Responses to Pulsed Electric Fields

5

Fig. 2 Integrated stress response. Various forms of stress cause the phosphorylation of eIF2α and consequent suppression of translation initiation, which are collectively referred to as the integrated stress response. Human and mammalian cells have four stress-responsive protein kinases, namely, PERK, GCN2, HRI, and PKR, each of which is activated by different external cues. The activated stress-responsive kinases phosphorylate eIF2α, which serves as a convergence point for signal transduction induced by various forms of stress. Phosphorylated eIF2α interferes with the assembly of a functional translational machinery on the 50 cap structure of mRNA. Most mRNA species are translated in a cap-dependent manner; therefore, eIF2α phosphorylation leads to the suppression of general protein synthesis. A complex of GADD34 with PP1 dephosphorylates eIF2α and serves as a negative feedback mechanism for the integrated stress response

on cell death induction (Woehlbier and Hetz 2011; Koromilas 2015). Thus, the attenuation of the stress response in a timely manner is also critical for cell survival. Translational Suppression Mediated by eIF2α Phosphorylation The process of protein synthesis consists of multiple steps from initiation to termination of translation, and it is primarily controlled at the step of translation initiation. Translation initiation of most messenger RNA (mRNA) species relies on the cap structure located at the 50 end of the mRNA. In the cap-dependent translation initiation, initiator methionyl-tRNA, eIF2, and the 40S ribosomal subunit associate to form the 43S preinitiation complex that is recruited to the cap structure bound to the eukaryotic initiation factor 4 (eIF4) and scans for the initiation codon in the mRNA (Baird and Wek 2012). Phosphorylation of the α subunit of eIF2 (eIF2α) inhibits the 43S preinitiation complex formation and consequently suppresses translation initiation (Baird and Wek 2012; Koromilas 2015). In human cells, serine 51 of eIF2α is rapidly phosphorylated under stress conditions (Fig. 2) (Koromilas 2015).

6

K.-i. Yano and K. Morotomi-Yano

The phosphorylation of eIF2α is conducted by stress-responsive protein kinases. Human and mammalian cells have four stress-responsive kinases for eIF2α phosphorylation, each of which is differentially activated by distinct forms of stress (Fig. 2) (Koromilas 2015). Protein kinase RNA-like endoplasmic reticulum kinase (PERK) is activated by endoplasmic reticulum (ER) stress, which is caused by unfolded proteins in the ER. General control nonderepressible 2 (GCN2) is mainly activated by amino acid deprivation. PKR (double stranded RNA-dependent protein kinase) plays a critical role in antiviral responses, and it is activated by double-stranded RNA, which emerges during viral infection. Heme-regulated inhibitor (HRI) has physiological roles particularly in erythroid tissues and is activated by heme deprivation. The phosphorylation of eIF2α at serine 51 is a common consequence of the activation of these stress-responsive kinases and is thus considered a hallmark of the induction of the integrated stress response (Baird and Wek 2012; Koromilas 2015). Recovery from Translational Suppression Caused by eIF2α Phosphorylation When the stress conditions pass over, cells need to restore protein synthesis through dephosphorylation of eIF2α, and growth arrest and DNA damage-inducible protein 34 (GADD34) is involved in this process. GADD34 associates with protein phosphatase 1 (PP1) to form a catalytically active phosphatase that specifically dephosphorylates eIF2α and relieves the translational suppression (Novoa et al. 2001). The activity of GADD34 in the relief of the integrated stress response is regulated at both transcriptional and translational levels. GADD34 gene is expressed at low levels under normal physiological conditions and is transcriptionally activated by a wide variety of stress. Most mRNA species in a stressed cell are poorly translated owing to eIF2α phosphorylation, but GADD34 mRNA can circumvent the eIF2α-mediated translational suppression. Upstream open reading frames (uORFs) in the 50 untranslated region of GADD34 mRNA are critical elements for translation initiation under stress conditions, although its precise mechanism is still elusive (Lee et al. 2009). In addition to GADD34 mRNA, approximately 2.5 % of the total mRNA was estimated to be preferentially translated under stress conditions through uORF-mediated and other related mechanisms (Dang Do et al. 2009; Baird and Wek 2012), which allow the synthesis of a subset of proteins that play critical roles in the control of the integrated stress response. Based on the transcriptional and translational regulation, GADD34 constitutes a negative feedback mechanism for the eIF2α phosphorylation-mediated translational suppression.

Alternative Mechanism for Stress-Induced Translational Suppression Human and mammalian cells have an additional mechanism for stress-induced translational suppression (Fig. 3), which is independent of eIF2α phosphorylation. As described above, the cap structure is important for translation initiation for most mRNA species. eIF4E recognizes the cap structure and recruits eIF4G and other translation initiation factors to form an active translation initiation complex. The inhibitory 4E-binding protein 1 (4E-BP1) competes with eIF4G for binding to eIF4E and thereby inhibits translation initiation (Ma and Blenis 2009). In normal

Cell Stress Responses to Pulsed Electric Fields

7

Fig. 3 mTORC1-mediated translational regulation. mTORC1 is involved in stress-induced translational regulation that is independent of the integrated stress response. On the 50 cap structure of mRNA, eIF4E and other proteins form a complex for translation initiation. 4E-BP1 has the ability to interfere with this protein assembly by binding to eIF4E. (a) Under normal physiological conditions, hyperphosphorylation by mTORC1 sequesters 4E-BP1 from eIF4E, which ensures the cap-dependent translation initiation. (b) Stress conditions cause a reduced mTORC1 activity that in turn results in reduced 4E-BP1 phosphorylation, leading to the suppression of the cap-dependent translation initiation by binding of 4E-BP1 to eIF4E

physiological state, mammalian target of rapamycin complex 1 (mTORC1), a member of the phosphatidylinositol 3-kinase-related kinase family, phosphorylates 4E-BP1 at multiple sites to sequester 4E-BP1 from eIF4E (Fig. 3). Energy deprivation and other stress conditions cause a marked reduction in the catalytic activity of mTORC1, resulting in the decreased phosphorylation and increased eIF4E-binding of 4E-BP1 (Fig. 3) (Ma and Blenis 2009). The catalytic activity of mTORC1 is controlled by several regulatory proteins, one of which is AMP-activated protein kinase (AMPK). AMPK can sense cellular energy status and negatively regulates the mTORC1 activity (Hardie 2011). Because protein synthesis is one of the major energy-consuming processes in the cell, its energy-dependent control involving mTORC1, AMPK, and 4E-BP1 serves as an alternative mechanism for adaptation to the stress conditions. As described above, many proteins function in the integrated stress response and the alternative mechanism for stress-induced translational suppression. Table 1 provides a list of proteins involved in these cellular reactions.

Stress Responses Induced by Nanosecond Pulsed Electric Fields Nanosecond Pulsed Electric Fields PEFs on the order of ms to μs are commonly used for electrotransfer of macromolecules into living cells, because these PEFs efficiently increase membrane permeability, which can be detected by incorporation of membrane-impermeant

8

K.-i. Yano and K. Morotomi-Yano

Table 1 List of proteins involved in stress responses Name eIF2α eIF4E eIF4G 4E-BP1 PERK GCN2 PKR HRI AMPK mTORC1 GADD34 PP1 IRE1 ATF6 CHOP XBP1 GADD45

Function Translation initiation factor Translation initiation factor Translation initiation factor Translation regulator Kinase Kinase Kinase Kinase Kinase Kinase Cofactor of phosphatase Phosphatase Kinase, nuclease Transcription factor Transcription factor Transcription factor Cofactor of kinases

Involvement in stress responses Translational suppression by its phosphorylation Binding to 50 cap of mRNA for translation initiation Assembly of translation initiation complex Inhibition of eIF4E binding to 50 cap of mRNA Activated by ER stress and PEFs Activated by UV, nutrition deprivation, and PEFs Activated by virus infection Activated by heme deprivation Activated by energy deprivation and nsPEFs Suppressed by energy deprivation Dephosphorylate eIF2α Dephosphorylate eIF2α Activated by ER stress Activated by ER stress Downstream factor of ER stress pathway Downstream factor of ER stress pathway Downstream factor of UV-induced stress pathway

fluorescent dyes, such as propidium iodide, into living cells. Compared to these PEFs, nsPEFs have less incidence of membrane permeabilization, as judged by the fluorescent dye incorporation. Instead, nsPEFs elicit calcium influx and cell blebbing, presumably caused by ion imbalance, indicating the generation of small membrane pores that allow permeation of ions and water (▶ Water Defects in Phospholipid Bilayer) (White et al. 2004; Vernier et al. 2006; Pakhomov et al. 2009). Recent studies have documented that nsPEFs induce various intracellular reactions that include stress responses (Morotomi-Yano et al. 2012b).

Activation of eIF2a Phosphorylation by nsPEFs eIF2α plays a critical role in cap-dependent translation initiation, and its activity is negatively controlled by phosphorylation under various stress conditions (Koromilas 2015). The occurrence of eIF2α phosphorylation is regarded as a hallmark of the integrated stress response and can be examined by Western blotting using an antibody specific to phosphorylated eIF2α (Fig. 4). When nsPEFs are applied to cultured cells, eIF2α phosphorylation is easily detected by Western blotting, indicating the induction of the integrated stress response by nsPEFs (Morotomi-Yano et al. 2012b). Phosphorylation of eIF2α can be detected in various cell lines, including HeLa S3, HCT116, Jurkat, and mouse embryonic fibroblasts (MEFs). In the case of HeLa S3, eIF2α phosphorylation occurs within 1 min after nsPEF exposure, persists at high levels for 30 min, and gradually decreases thereafter. Downregulation of eIF2α phosphorylation is known to involve the activation of

Cell Stress Responses to Pulsed Electric Fields

9

Fig. 4 Examples of Western blot analysis of protein phosphorylation. Phosphorylation status of a protein of interest can be examined by Western blotting. When available, a specific antibody against a phosphorylated protein is used for the analysis of protein phosphorylation. If a protein of interest is heavily phosphorylated at multiple sites, such hyperphosphorylation causes a slow migration of the protein in gel electrophoresis and yields shift band(s) in Western blotting. In this case, hyperphosphorylated and unphosphorylated species are distinguishable in a single Western blot. (a) Example of use of a phospho-specific antibody. Western blot analysis of eIF2α phosphorylation was performed using a specific antibody against phosphorylated eIF2α (upper panel) and an antibody reacting with eIF2α irrespective of its phosphorylation status (lower panel). eIF2α exists in both nsPEF-exposed (shown by P) and untreated (shown by -) cells (lower panel), but phosphorylated eIF2α is detected only in the nsPEF-exposed cells (upper panel). (b) Example of a bandshift due to hyperphosphorylation. Western blot analysis of PERK was performed using an anti-PERK antibody. PERK protein is detected as a single band in untreated cells (-). nsPEFs induce hyperphosphorylation of PERK that causes slow electrophoretic migration of PERK protein and yields shifted bands in the nsPEF-exposed cells (P)

GADD34 that constitutes an active eIF2α phosphatase with the catalytic subunit of PP1. Quantitative analysis of reverse transcription-polymerase chain reactions (RT-PCR) demonstrated that GADD34 gene expression is markedly increased at 1 h after nsPEF exposure (Morotomi-Yano et al. 2012b), supporting the idea that GADD34 is involved in the recovery of nsPEF-induced eIF2α phosphorylation (Fig. 5). Recent studies have demonstrated that nsPEFs induce different intracellular reactions in a manner dependent on nsPEF intensity (Fig. 6). Relatively weak nsPEFs, which do not cause obvious growth retardation or cell death, trigger the activation of several signal transduction pathways, including mitogen-activated protein kinase pathways (Morotomi-Yano et al. 2011b; Morotomi-Yano et al. 2011a). Western blot analysis of eIF2α phosphorylation has shown that such weak nsPEFs are insufficient for the induction of the integrated stress response. To achieve eIF2α phosphorylation, modest intensities of nsPEFs are required, which cause reduced proliferation but not gross cell death (Morotomi-Yano et al. 2012b). Intense nsPEFs induce not only eIF2α phosphorylation but also cause cell death (▶ Cell Death Due to Electroporation; ▶ Apoptotic Indicators and Cell Death Following Nanosecond Electroporation) (Morotomi-Yano et al. 2013, 2014). However, it remains unclear whether the nsPEF-induced stress responses serve only as a prosurvival mechanism or have facilitative effects on cell death induction, when overactivated. Further investigation is needed to clarify this point.

10

K.-i. Yano and K. Morotomi-Yano

Fig. 5 nsPEF-induced stress responses. Previous research has demonstrated the effects of nsPEFs on multiple stress-responsive proteins that constitute two independent pathways leading to translational suppression. nsPEFs activate PERK (a) and GCN2 (b) that phosphorylate eIF2α. PERK and GCN2 appear to act on the nsPEF-induced eIF2α phosphorylation in a mutually compensatory manner. (c) nsPEFs cause reduced 4E-BP1 phosphorylation, suggesting an inhibitory effect of nsPEFs on mTORC1. It remains unclear whether nsPEFs affects mTORC1 directly or indirectly. (d) nsPEFs cause reduction in intracellular energy levels and AMPK activation, both of which are known to contribute to negative regulation of the mTORC1 activity and thus may account for the reduced 4E-BP1 phosphorylation. Increased eIF2α phosphorylation and decreased 4E-BP1 phosphorylation serve as two independent mechanisms for translational suppression

Fig. 6 Intensity-dependency of cellular responses to nsPEFs. nsPEFs elicit distinct cellular responses, depending on their intensities. Relatively weak nsPEFs exhibit little effects on cell growth and cell death but can activate several signal pathways, including various MAPK pathways, like JNK, ERK, and p38 pathways, and AMPK pathway. nsPEFs at modest intensities induce the stress responses that include eIF2α phosphorylation, 4E-BP1 dephosphorylation, and translational suppression. Strong nsPEFs cause cell death, and either apoptosis or necrosis is induced in a cell type-dependent manner

Cell Stress Responses to Pulsed Electric Fields

11

Activation of Stress-Responsive Protein Kinases by nsPEFs Human and mammalian cells have four stress-responsive protein kinases for eIF2α phosphorylation, each of which is activated by a distinct form of stress (Koromilas 2015). Activation of these kinases is manifested by their autophosphorylation, which can be detected by Western blotting (Fig. 4). Analysis of phosphorylation status of the stress-responsive kinases by Western blotting demonstrated that PERK and GCN2 are definitely activated by nsPEFs (Fig. 5) (Morotomi-Yano et al. 2012b). Intriguingly, PERK and GCN2 act on the PEF-induced eIF2α phosphorylation in a mutually compensatory manner. Mouse embryonic fibroblast cells lacking either PERK or GCN2 gene do not exhibit any detectable reduction in the nsPEF-induced eIF2α phosphorylation, suggesting the presence of a compensatory mechanism for the lack of either PERK or GCN2 (Morotomi-Yano et al. 2012b). Double knockout cells for both PERK and GCN2 genes exhibit a significant decrease in the nsPEF-induced eIF2α phosphorylation, supporting the idea of the functional compensation between PERK and GCN2 (Morotomi-Yano et al. 2012b), although the molecular details of this phenomenon need to be further investigated. PERK and GCN2 play central roles in the stress response pathways for ER stress and UV irradiation, respectively, and the activation of these kinases by nsPEFs raised a possibility that nsPEFs act in ways similar to ER stress or UV irradiation. It is well established that stress responses caused by ER stress and UV irradiation result in expression of distinct subsets of stress-inducible genes. For example, ER stress activates three upstream signaling proteins, namely, PERK, activating transcription factor 6 (ATF6), and inositol-requiring protein 1 (IRE1), which in turn lead to the induction of CCAAT-enhancer-binding protein homologous protein (CHOP) gene transcription and X-box binding protein 1 (XBP1) mRNA splicing (Szegezdi et al. 2006). UV irradiation elicits GCN2 activation that causes altered expression of growth arrest and DNA damage-inducible 45 (GADD45) and other genes (Takekawa and Saito 1998). In the case of HeLa S3 cells, these alterations in gene expression are detectable within 1 h after treatment with ER stress-inducing agents or UV irradiation. To clarify whether nsPEFs exert their effects as ER stress or not, quantitative RT-PCR analysis of CHOP and XBP1 mRNAs was performed using HeLa S3 cells. However, there were no substantial changes in these mRNAs in nsPEF-exposed cells, indicating that nsPEFs do not act as ER stress despite the obvious PERK activation (Morotomi-Yano et al. 2012b). In addition, RT-PCR analysis performed on the UV-inducible genes indicated no significant changes in these genes in the nsPEF-exposed cells (Morotomi-Yano et al. 2012b). Taken together, these observations suggest that nsPEFs act as a new form of cellular stress distinct from ER stress and UV irradiation. It is noteworthy that PERK and GCN2 are located most upstream in their signaling pathways, suggesting direct effects of nsPEFs on these kinases, although precise mechanisms for their activation by nsPEFs need further investigation.

12

K.-i. Yano and K. Morotomi-Yano

Effects of nsPEFs on 4E-BP1 and AMPK Human cells have an additional mechanism for the stress-induced translational suppression, and 4E-BP1 is involved in this process (Ma and Blenis 2009). Most mRNA species are translated in a cap-dependent manner, and 4E-BP1 can interfere with the assembly of translation factors on the cap structure (Fig. 3). In a normal physiological state, 4E-BP1 is heavily phosphorylated by mTORC1, and its inhibitory activity on the cap-dependent translation initiation is suppressed by hyperphosphorylation. Under stress conditions, particularly during energy deprivation, the catalytic activity of mTORC1 is decreased, leading to decreased 4E-BP1 phosphorylation and consequent translational suppression (Fig. 3) (Ma and Blenis 2009). When nsPEFs are applied to cultured cells, such as HeLa S3 and MEFs, dephosphorylation of 4E-BP1 is induced within 5 min and persists for more than 1 h, suggesting the occurrence of 4E-BP1-mediated translational suppression (Fig. 5) (Morotomi-Yano et al. 2012b). Although the effects of nsPEFs on the catalytic activity of mTORC1 have not been studied yet, the decreased 4E-BP1 phosphorylation appears to reflect the decreased catalytic activity of mTORC1. Recently, electroporation using 100 μs PEFs has been reported to induce energy deprivation in a calcium-dependent manner, which ultimately causes necrotic cell death (Frandsen et al. 2012). Similarly, nsPEFs also induce calcium-dependent necrotic cell death in several cell lines, including HeLa S3, K562, and HEK293 (MorotomiYano et al. 2014). The calcium-dependent energy deprivation in PEF-exposed cells appears to account for the decrease in 4E-BP1 phosphorylation. In addition, nsPEFs have been reported to induce AMPK activation by protein phosphorylation (Morotomi-Yano et al. 2012a). AMPK phosphorylation is rapidly induced within 1 min, persists for 15 min, and then decreases, when nsPEFs are applied to HeLa S3 or Jurkat cells. Contribution of activated AMPK to the negative regulation of mTORC1 has been well documented (Ma and Blenis 2009). Collectively, these observations imply the translational suppression mediated by decreased 4E-BP1 phosphorylation in nsPEF-exposed cells (Fig. 5), although the precise relationship among 4E-BP1, mTORC1, and AMPK in nsPEF-exposed cells remains to be confirmed at molecular levels.

Suppression of Protein Synthesis by nsPEFs The phosphorylation-mediated control of eIF2α and 4E-BP1 activities strongly suggests that nsPEFs activate two independent mechanisms for translational suppression (Fig. 5), which prompted further analysis of protein synthesis rates in nsPEF-exposed cells. Because living cells contain massive amounts of proteins, discrimination of newly synthesized proteins from preexisting ones requires a special experimental approach, namely, metabolic labeling with radioactive amino acids (Fig. 7). In this method, cells are incubated in culture medium containing 35S-labeled methionine and cysteine. During incubation, cells incorporate the radioactive amino acids and use them for protein synthesis, resulting in the

Cell Stress Responses to Pulsed Electric Fields

13

Fig. 7 Measurement of protein synthesis rates by using radioactive amino acids. Radioactive amino acids are used for metabolic labeling of newly synthesized proteins to measure overall rates of protein synthesis. After nsPEF exposure, 35S-labeled amino acids, namely, methionine and cysteine, are added to culture medium. Cells uptake these radioactive amino acids and use them for protein synthesis, yielding radioactive proteins. Following cell extract preparation, proteins are precipitated in the presence of trichloroacetic acid, and radioactivity in a protein fraction is measured by liquid scintillation counting

production of radioactive proteins. Following cell extract preparation, a protein fraction is separated from that of free amino acids, and the radioactivity in the protein fraction is quantified. By applying this method, the suppression of general protein synthesis in nsPEFexposed cells was demonstrated using HeLa S3 cells as well as MEFs (MorotomiYano et al. 2012b). After nsPEF exposure, protein synthesis was quickly attenuated, and the maximum suppression was observed at 30 min. UV irradiation is commonly used as a positive control for stress response induction and elicits persistent translational suppression for several hours. Acute suppression and rapid recovery of protein synthesis are characteristics of the nsPEF-induced stress response, when compared to UV irradiation. Collectively, nsPEFs induce acute suppression of general protein synthesis that is presumably mediated by the phosphorylation of eIF2α and 4E-BP1.

Nucleofection-Induced Stress Responses Nucleofection is a sort of electroporation that is conducted using a specialized device, in which undisclosed electric parameters optimized for individual cell types are preset (Gresch et al. 2004). Nucleofection can achieve efficient transfection and is particularly useful for primary and nondividing cells. In nucleofection-treated cells, phosphorylation status of the signaling molecules in the integrated stress response was analyzed by Western blotting (Anderson et al. 2013). First, eIF2α phosphorylation was detected in MEFs as well as primary human dendritic cells after nucleofection. Increased eIF2α phosphorylation was rapidly induced and detectable up to 2 h. The presence of exogenous DNA was not required for nucleofectioninduced eIF2α phosphorylation, suggesting the causal effects of electric pulses in nucleofection. Similar to nsPEFs, PERK and GCN2 were activated by nucleofection, and GCN2 appears to play a major role in the nucleofection-induced eIF2α phosphorylation. Luciferase activities derived from transfected expression plasmids were decreased after nucleofection, indicative of translational suppression. Collectively,

14

K.-i. Yano and K. Morotomi-Yano

these observations suggest the occurrence of the integrated stress responses by nucleofection and provide insightful information about cellular effects of PEFs used for electroporation (Anderson et al. 2013).

Possible Effects of Electroporation on Cell Physiology In contrast to particular emphasis given to nsPEFs and nucleofection in the above sections, the last part of this chapter is dedicated to PEFs used for electroporation. Because it has become evident that both nsPEF exposure and nucleofection elicit stress responses, it seems reasonable to assume that regular electroporation also induces stress responses. One of the most common applications of electroporation is transfection of exogenous DNA, and PEFs used for DNA transfection may exert their negative effects on transgene expression by stress-induced translational suppression. Accordingly, nucleofection has been shown to cause a reduction in luciferase activity expressed from transfected plasmid DNA (Anderson et al. 2013). In case of nsPEFs, translational suppression is relatively acute and recovers within a few hours (Morotomi-Yano et al. 2012b). These observations suggest the importance of PEF conditions in transfection, and careful optimization of PEF conditions is required for the maximum expression of transgene. Another possible adverse effect of electroporation is the occurrence of cell death, and in fact, intense PEFs can induce massive cell death caused by irreparable damages, which is called irreversible electroporation and is used for cancer treatment. Currently, it remains unclear whether the PEF-induced stress responses are beneficial for cell survival or have a facilitatory influence on cell death. When properly controlled, stress responses collectively serve as a potent mechanism for cell survival, whereas persistent activation of stress responses often promotes cell death, presumably because elimination of overstressed cells is beneficial for the body (Woehlbier and Hetz 2011; Koromilas 2015). Thus, the PEF-induced stress responses may have dual effects depending on the extent and duration of their activation, and future research is required to evaluate the contribution of the stress responses to survival and death in electroporation. In addition to DNA transfection, electroporation is utilized for the introduction of antitumor drugs for cancer chemotherapy, in which the PEF-induced stress responses could have therapeutic advantages. First, tumor cells have mechanisms to counteract the antitumor drugs by degrading or excluding them, and translational suppression could attenuate these cellular activities and augment therapeutic efficacy of the introduced drugs. Second, stress-induced signaling generally leads to the activation of downstream factors, many of which negatively regulate cell growth (Hamanaka et al. 2005; Koromilas 2015). Thus, PEFs used for electroporation may induce stress-induced antiproliferative intracellular reactions. Finally, translational suppression itself is reasonably expected to have inhibitory effects, because cell growth requires synthesis of various proteins. In summary, PEFs used in cancer chemotherapy should have several antiproliferative effects that can act in synergy with antitumor drugs.

Cell Stress Responses to Pulsed Electric Fields

15

Conclusion Despite their usefulness in various life science fields, PEF are essentially detrimental for cellular homeostasis because of their ability to permeabilize the cell membrane. Therefore, PEFs can be regarded as a sort of stress for a living cell, and, indeed, it has become increasingly evident that PEFs elicit stress responses. Exposure of human and mammalian cells to nsPEFs induces two distinct stress response pathways, both of which alter phosphorylation status of translation initiation factors, leading to the suppression of protein synthesis. In addition to nsPEFs, stress responses are also activated by nucleofection, which is a sort of electroporation conducted by a specialized device with pre-optimized electric parameters. Collectively, these observations clearly indicate that PEFs act as stress to induce cascades of intracellular reactions leading to the suppression of protein synthesis. Identification of key intracellular events of PEF-induced stress responses raises next questions to be elucidated. To date, common cell lines, such as HeLa S3 and MEFs, under standardized culture conditions have been utilized as an experimental model for investigating the molecular details of PEF-induced stress responses. Such experimental setups are useful for analysis of intracellular mechanisms but insufficient to assess biological significance of PEF-induced stress responses in the in vivo situations. Further studies employing physiologically relevant conditions are required to advance our understanding of possible contributing factors to PEF-induced stress in vivo, such as an influence of microenvironment and dependency of tissue and cellular contexts. Furthermore, limited information is currently available on how and to what extent the PEF-induced stress responses affect the outcomes of electroporation-based applications. Although PEF-induced stress responses can be speculated to facilitate cancer chemotherapy and to exert negative effects on transgene expression in DNA transfection, experimental validation is required. These future efforts will provide a broader and more integrative view of PEF-induced stress responses, which pave the way for more effective applications of PEFs in the life sciences. Acknowledgment This work was supported by JSPS KAKENHI Grant Number 26350540.

Cross-References ▶ Apoptotic Indicators and Cell Death Following Nanosecond Electroporation ▶ Cell Death Due to Electroporation ▶ Nucleic Acid Electrotransfer in Mammalian Cells: Mechanistic Description ▶ Overview and History of Electrochemotherapy ▶ Responses of Plant Cells and Tissues to Pulsed Electric Field Treatments ▶ Stress Response of Plants, Metabolite Production to Pulsed Electric Fields ▶ Water Defects in Phospholipid Bilayer

16

K.-i. Yano and K. Morotomi-Yano

References Alberts B, Johnson A, Lewis J, Morgan D, Raff M, Roberts K, Walter P (2014) Molecular biology of the cell, 6th edn. Garland Publishing, New York Anderson BR, Kariko K, Weissman D (2013) Nucleofection induces transient eIF2alpha phosphorylation by GCN2 and PERK. Gene Ther 20(2):136–142. doi:10.1038/gt.2012.5 Baird TD, Wek RC (2012) Eukaryotic initiation factor 2 phosphorylation and translational control in metabolism. Adv Nutr 3(3):307–321. doi:10.3945/an.112.002113 Dang Do AN, Kimball SR, Cavener DR, Jefferson LS (2009) eIF2alpha kinases GCN2 and PERK modulate transcription and translation of distinct sets of mRNAs in mouse liver. Physiol Genomics 38(3):328–341. doi:10.1152/physiolgenomics.90396.2008 Frandsen SK, Gissel H, Hojman P, Tramm T, Eriksen J, Gehl J (2012) Direct therapeutic applications of calcium electroporation to effectively induce tumor necrosis. Cancer Res 72 (6):1336–1341. doi:10.1158/0008-5472.CAN-11-3782 Gresch O, Engel FB, Nesic D, Tran TT, England HM, Hickman ES, Korner I, Gan L, Chen S, Castro-Obregon S, Hammermann R, Wolf J, Muller-Hartmann H, Nix M, Siebenkotten G, Kraus G, Lun K (2004) New non-viral method for gene transfer into primary cells. Methods 33 (2):151–163. doi:10.1016/j.ymeth.2003.11.009 Hamanaka RB, Bennett BS, Cullinan SB, Diehl JA (2005) PERK and GCN2 contribute to eIF2alpha phosphorylation and cell cycle arrest after activation of the unfolded protein response pathway. Mol Biol Cell 16(12):5493–5501. doi:10.1091/mbc.E05-03-0268 Hardie DG (2011) AMP-activated protein kinase: an energy sensor that regulates all aspects of cell function. Genes Dev 25(18):1895–1908. doi:10.1101/gad.17420111 Koromilas AE (2015) Roles of the translation initiation factor eIF2alpha serine 51 phosphorylation in cancer formation and treatment. Biochim Biophys Acta 1849(7):871–880. doi:10.1016/j. bbagrm.2014.12.007 Lee YY, Cevallos RC, Jan E (2009) An upstream open reading frame regulates translation of GADD34 during cellular stresses that induce eIF2alpha phosphorylation. J Biol Chem 284 (11):6661–6673. doi:10.1074/jbc.M806735200 Leprivier G, Rotblat B, Khan D, Jan E, Sorensen PH (2015) Stress-mediated translational control in cancer cells. Biochim Biophys Acta 1849(7):845–860. doi:10.1016/j.bbagrm.2014.11.002 Ma XM, Blenis J (2009) Molecular mechanisms of mTOR-mediated translational control. Nat Rev Mol Cell Biol 10(5):307–318. doi:10.1038/nrm2672 Morotomi-Yano K, Akiyama H, Yano K (2011a) Nanosecond pulsed electric fields activate MAPK pathways in human cells. Arch Biochem Biophys 515(1–2):99–106. doi:10.1016/j. abb.2011.09.002 [pii] S0003-9861(11)00312-2 Morotomi-Yano K, Uemura Y, Katsuki S, Akiyama H, Yano K (2011b) Activation of the JNK pathway by nanosecond pulsed electric fields. Biochem Biophys Res Commun 408(3):471–476. doi:10.1016/j.bbrc.2011.04.056 [pii] S0006-291X(11)00641-3 Morotomi-Yano K, Akiyama H, Yano K (2012a) Nanosecond pulsed electric fields activate AMP-activated protein kinase: implications for calcium-mediated activation of cellular signaling. Biochem Biophys Res Commun 428(3):371–375. doi:10.1016/j.bbrc.2012.10.061 Morotomi-Yano K, Oyadomari S, Akiyama H, Yano K (2012b) Nanosecond pulsed electric fields act as a novel cellular stress that induces translational suppression accompanied by eIF2alpha phosphorylation and 4E-BP1 dephosphorylation. Exp Cell Res 318(14):1733–1744. doi:10.1016/j.yexcr.2012.04.016 Morotomi-Yano K, Akiyama H, Yano K (2013) Nanosecond pulsed electric fields induce poly (ADP-ribose) formation and non-apoptotic cell death in HeLa S3 cells. Biochem Biophys Res Commun 438(3):557–562. doi:10.1016/j.bbrc.2013.07.083 Morotomi-Yano K, Akiyama H, Yano K (2014) Different involvement of extracellular calcium in two modes of cell death induced by nanosecond pulsed electric fields. Arch Biochem Biophys 555–556:47–54. doi:10.1016/j.abb.2014.05.020

Cell Stress Responses to Pulsed Electric Fields

17

Novoa I, Zeng H, Harding HP, Ron D (2001) Feedback inhibition of the unfolded protein response by GADD34-mediated dephosphorylation of eIF2alpha. J Cell Biol 153(5):1011–1022 Pakhomov AG, Bowman AM, Ibey BL, Andre FM, Pakhomova ON, Schoenbach KH (2009) Lipid nanopores can form a stable, ion channel-like conduction pathway in cell membrane. Biochem Biophys Res Commun 385(2):181–186. doi:10.1016/j.bbrc.2009.05.035 [pii] S0006-291X(09) 00956-5 Rems L, Miklavcic D (2016) Tutorial: electroporation of cells in complex materials and tissue. J Appl Phys 119(20). Art. no. 201101. doi: 10.1063/1.4949264 Szegezdi E, Logue SE, Gorman AM, Samali A (2006) Mediators of endoplasmic reticulum stressinduced apoptosis. EMBO Rep 7(9):880–885. doi:10.1038/sj.embor.7400779 Takekawa M, Saito H (1998) A family of stress-inducible GADD45-like proteins mediate activation of the stress-responsive MTK1/MEKK4 MAPKKK. Cell 95(4):521–530 Vernier PT, Sun Y, Gundersen MA (2006) Nanoelectropulse-driven membrane perturbation and small molecule permeabilization. BMC Cell Biol 7:37. doi:10.1186/1471-2121-7-37 [pii] 14712121-7-37 White JA, Blackmore PF, Schoenbach KH, Beebe SJ (2004) Stimulation of capacitative calcium entry in HL-60 cells by nanosecond pulsed electric fields. J Biol Chem 279(22):22964–22972. doi:10.1074/jbc.M311135200M311135200 Woehlbier U, Hetz C (2011) Modulating stress responses by the UPRosome: a matter of life and death. Trends Biochem Sci 36(6):329–337. doi:10.1016/j.tibs.2011.03.001

Effects of Electroporation of Mammalian Cells on Cytoskeleton and Intercellular Connections Maja Cemazar

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cytoskeleton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Intercellular Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Role of Cytoskeleton in Cell Electrofusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Effects of Electroporation on Cytoskeleton and Intercellular Connections . . . . . . . . . . . . . . . . The Effects of Electrochemotherapy and Gene Electrotransfer on Cytoskeleton and Intercellular Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 5 6 6 11 13 13 14

Abstract

Electroporation of mammalian cells affects several cellular components, including cytoskeleton and intercellular connections. This chapter describes the effects of electroporation alone or in combination with cytotoxic drugs or plasmid DNA on the components of cytoskeleton. Responses of cells to electroporation depend on the type of cells. In Chinese hamster ovary cells, mainly microtubule filaments were depolymerized following electroporation with microsecond pulses, with a recovery during 60 min, while recently the effect on the actin filaments in the cell cortex was also demonstrated. Electroporation of endothelial cells caused, besides depolymerization of microtubules, also depolymerization of actin filaments, while vimentin intermediate filaments were not affected. In addition, the loss of contractility and vascular endothelial cadherin (VE-cadherin) from the adherens cell junction was also observed. All these changes lead to increased permeability M. Cemazar (*) Department of Experimental Oncology, Institute of Oncology Ljubljana, Ljubljana, Slovenia Faculty of Health Sciences, University of Primorska, Izola, Slovenia e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_18-1

1

2

M. Cemazar

of endothelial cell monolayer. In tumor cells, electroporation does not cause depolymerization of actin filaments, but their rearrangement into stress fibers, lamellipodia, and filopodia. Electroporation with nanosecond pulses also affects actin filaments. Their depolymerization and disassembly of gap junctions were observed following exposure of cells to nanosecond pulses, while tight junctions were not significantly affected. The role of cytoskeletal filaments was also studied following electrochemotherapy and plasmid DNA electrotransfer into the cells demonstrating that cytoskeletal proteins are quicker and more severely affected following electrochemotherapy compared to electroporation alone and that the actin filaments play an active role in gene electrotransfer. Collectively, electroporation greatly affects cytoskeleton and intercellular connections in different types of cells and plays an important role in the transport of plasmid DNA into the cells and in the vascular changes observed after treatment with electrochemotherapy. However, studies are needed to fully elucidate the involvement of cytoskeleton and intercellular connection in different types of electroporation application, which will possibly lead to improved or new treatment options based on electroporation technology. Keywords

Cytoskeleton • Intercellular connections • Electroporation • Nanosecond pulsed electric fields

Introduction Electroporation of different types of cells in vitro is a powerful tool used in many biological applications for introduction of various molecules into the cells or into the membranes of cells, fusion of cells, and extraction of molecules from cells. In electroporation protocols for mammalian cells, most commonly used electric pulses are square wave-shaped electric pulses. The duration of pulses ranges from nanoseconds to milliseconds with electric field densities up to several kV/cm for nanosecond pulses to several tens to hundreds V/cm for micro- and millisecond pulses, depending on the desired biological application. Besides the influence of electrical parameters on the desired outcome of the application, a significant role plays also other physicochemical properties, such as the composition of the electroporation medium (buffer) and its temperature, as well as biological properties of the cells. Among the biological properties of the cells, the size and shape of the cells, whether they are in suspension or adherent, the density of cell suspension, and the type of the cells (fibroblasts, erythrocytes, endothelial cells, different types of tumor cells) are the most studied properties in the electroporation studies. Since early studies on cell electrofusion and cell electropermeabilization for introduction of small molecules into the cells, it is known that mammalian cells behave differently compared to lipid vesicles. Therefore, already in these early studies, the involvement of cytoskeleton in the permeabilization process was suggested (Blangero et al. 1989). Later on

Effects of Electroporation of Mammalian Cells on Cytoskeleton and. . .

3

influence of cell-to-cell junctions on the behavior of cells exposed to electric pulses was also studied (Kanthou et al. 2006). Basic experimental mechanistic studies together with realistic finite element numerical modeling provided further insights into the process of electroporation in cells and tissues and are important for biomedical progress of electroporation into clinics, where it is already used in electrochemotherapy (introduction of cytotoxic drugs into the cells of tumors for tumor destruction), plasmid DNA vaccination protocols (introduction of plasmid DNA encoding antigen either in the skin or muscle for development of immunity), and gene therapy protocols (introduction of plasmid DNA encoding therapeutic protein in different types of tissue (skin, muscle, tumor) depending on the target disease). In this chapter, firstly the basic facts about cytoskeleton and intercellular connections in mammalian cells are described, and the role of cytoskeleton and intercellular connections in electroporation-based methods and treatments, such as cell fusion, electrochemotherapy, and gene electrotransfer, is explained.

Cytoskeleton Cytoskeleton is present in all types of cells: archaea, bacteria, and eukaryotes. Initially, cytoskeleton was thought to be unique to eukaryotic cells; however, in the past 20 years, the presence of cytoskeletal proteins with mainly cytomotive function was also demonstrated in archaeal and bacterial cells (Wickstead and Gull 2011; Backert et al. 2013). In eukaryotes the cytoskeleton contains three major types of protein filaments: microtubules, intermediate filaments, and actin filaments, also called microfilaments. The overall function of cytoskeleton is to support and maintain the shape of the cell, by giving it the mechanical resistance to deformation. In tissues, cytoskeletal proteins are associated with extracellular matrix and also with neighboring cells, with the help of accessory proteins, thus supporting the entire tissue or organ. The cytoskeletal proteins are also involved in many other functions of cells, such as migration, intracellular transport, signaling, endocytosis, segregation of chromosomes during nuclear division, and cytokinesis, the final division of the cell’s cytoplasm. The cytoskeletal proteins are thus very dynamic structures that are capable of rapid growth and disassembly, depending on the type of the cell and the requirements of the cell. In addition, not all types of the cytoskeletal protein are present in all cell types, for example, the certain types of intermediate filaments (Fig. 1) (Wickstead and Gull 2011; Backert et al. 2013). The microtubules are hollow cylinders of outer diameter of 23 nm that are a polymer composed of a dimer of α- and β-tubulin. They originate from the nuclear organization center called centrosome that is located near the cell nucleus and is composed of γ-tubulin. From this centrosome, the microtubules grow outward, and the growing end of the microtubules is the plus end and is oriented toward the site of polarity, for example, to the leading edge of migratory endothelial cell. The energy for the growth of microtubules comes for guanosine triphosphate (GTP), and due to the quick hydrolization of GTP, when bound to β-tubulin, the microtubules can

4

M. Cemazar

Fig. 1 Cytoskeleton and intercellular connections in cells. Blue – microtubules extending from the centrosome toward the cell membrane. Green – intermediate filaments are involved in the desmosomes (green) connecting two neighboring cells and hemidesmosomes (light blue) that are connecting cells to the extracellular matrix. Red – actin filaments are a component of adherens junction and also form a cell cortex, a filamentous network under the cell membrane. Yellow – gap junctions enable communication between cells. Black dots in the apical site of the cells – tight junctions, which are preventing the movement of the proteins embedded into the cell membrane from apical to basal site of the cells. The schematic representation was prepared by Dr. Urska Kamensek

rapidly switch between growing and shrinking, the process known as dynamic instability, which is important for the migration of cells. Another constituent of cytoskeletal proteins are intermediate filaments, which are a class of different types of proteins with average diameter of 10 nm. Intermediate filaments are composed of different types of proteins depending on the type of the cell. Lamins are the only intermediate filaments that are located in the nucleus of all nucleated cells and form the nuclear lamina, which supports the nuclear envelope and helps in disassembly and assembly of the nuclear envelope during the mitosis. In epithelial cells, the intermediate filaments, keratins or cytokeratins, are present and are extended from the nuclear envelope toward the plasma membrane, where they are part of desmosomes and hemidesmosomes. Vimentin is an intermediate filament present in cells of mesenchymal origin, such as fibroblasts, endothelial cells, and leukocytes. Its main role is to maintain cell integrity. Other types of intermediate filaments are neurofilaments, found abundantly in axons, and desmins that are present in sarcomeres of the muscle cells. Actin filaments or microfilaments are the smallest among cytoskeletal proteins, with the diameter of 7 nm. They are composed of G-actin, a globular actin, which is a

Effects of Electroporation of Mammalian Cells on Cytoskeleton and. . .

5

monomer that forms F-actin, a filamentous actin. Actin filaments are very flexible, and their role in the cell is dependent on the accessory proteins that are bound to actin filaments. They are involved in muscle contraction, together with myosin in muscle cells. In non-muscle cells, they are the main constituents of cell cortex, they are involved in signal transduction, they form stress fibers, and they are part of adherent cells in cell-to-cell junction and focal adhesion in cell-to-extracellular matrix adhesion (Fig. 1) (Wickstead and Gull 2011; Backert et al. 2013).

Intercellular Connections Cell-to-cell junctions, also called intercellular bridge, are protein structures that exist between the cells in all multicellular organisms and provide contact between the cells. They are especially abundant in epithelial tissues, where they provide epithelial cell polarization and control paracellular transport. Cell-to-cell junction provides anchoring, communication, and also barrier between the neighboring cells. Besides cell-to-cell junctions, in multicellular organism, also cell-to-extracellular matrix adhesions and non-junctional adhesion mechanisms exist (Fig. 1). Anchoring cell-to-cell junctions are desmosomes and adherens junctions, which are the protein structures between the cells, and hemidesmosomes and focal adhesions that connect cells to extracellular matrix (i.e., basal lamina in the epithelium). The intermediate cytoskeletal proteins, cytokeratin or desmin, are part of desmosomes, where cadherins are the proteins that form the anchor. In hemidesmosomes that connect polarized epithelial cells to extracellular matrix, the cytokeratins are anchored to the extracellular matrix by the involvement of transmembrane integrins. Adherens junctions are composed of bundles of actin filaments. They can appear as bands encircling the epithelial and endothelial cell. Actin bundles are connected across the plasma membrane with the cadherins, like in desmosomes. In focal adhesions, which are adhesion junctions of cell to extracellular matrix in polarized and non-polarized cells, the actin filaments are anchored to the substrate by integrins, like in hemidesmosomes. Cell-to-cell communication is enabled by gap or communication junctions, which enables cell in the epithelium to be in direct contact, where the diffusion of molecules can occur directly, without passage through extracellular matrix. The cytoskeletal proteins are not involved in the formation of the junction. The main proteins that form the gap junction are connexins, which form a pore in the cell membrane called connexon. These junctions are especially important in the heart muscle cell, where they enable heart beating through action potential propagation. Tight junctions are present only in epithelial cells. Their role is dual; they prevent the apical membrane proteins from moving to basolateral surface, thus enabling polarization of the epithelium; and they seal the spaces between epithelial cells to prevent the movement of molecules between the cells. Actin filaments are involved in the formation of the tight junction, together with transmembrane proteins claudins, occludins, and junction adhesion molecules (JAM) (Wickstead and Gull 2011; Backert et al. 2013).

6

M. Cemazar

The Role of Cytoskeleton in Cell Electrofusion Cell electrofusion was a topic on numerous studies, due to the important application of cell fusion in biotechnology and medicine, mainly for production of hybridomas and preparation of cell vaccines. Most of the studies were focusing on the evaluation of different electrical parameters, as well as on other parameters that could influence the cell fusion, such as the composition of the fusion medium, osmolarity of the fusion medium, temperature, and post-pulse incubation of cells. Namely, in the cell electrofusion experiments, the biggest obstacle is the difference in size of the cells to be fused. In fact, in hybridoma technology, small lymphocytes need to be fused with large myeloma cells in order to get the production of antibodies (Jaroszeski et al. 1994; Pucihar et al. 2011; Rems et al. 2013). The study that addressed the cytoskeleton changes after electrofusion was performed by Blangero et al. (Blangero et al. 1989). Chinese hamster ovary (CHO) cells, a clone WTT, have been selected for the experiments. This clone is less anchorage dependent than the parental cell line and can grow also in suspension. However, the plated cells were used for electrofusion experiments, and the cells were immunohistochemically stained for tubulin and actin filaments at different times posttreatment. The protocol for electric pulses used in the study for electrofusion was as follows: five square wave electric pulses of 1200 V/cm and 100 μs duration were applied to the attached cells via two parallel stainless steel electrodes with 5 or 18 mm distance in between. No particular change in the organization of actin filaments was observed during the observation time of 60 min at 37  C. On the other hand, the organization of microtubules was significantly altered already 10 min after the treatment. The fluorescently labeled microtubules were present only around the nucleus, which were clustered in the middle of the fused cell. The observed effect was reversible, and after 60 min the microtubule network was morphologically similar to the network in control untreated cells. In conclusion, the experiments on the involvement of cytoskeleton in cell electrofusion showed that microtubules were transiently depolymerized during cell fusion, while actin filaments were not affected.

The Effects of Electroporation on Cytoskeleton and Intercellular Connections Due to the important role of electroporation in the biomedical research and especially in the medical treatments (electrochemotherapy, gene therapy, DNA vaccination), numerous of preclinical studies are devoted to the elucidation of the molecular changes that electroporation alone or in combination with cytotoxic drugs and nucleic acids (DNA, RNA) is triggering in the cells. These preclinical studies are needed for better planning of electroporation-based treatments. Almost 25 years ago, the study by Rols and Teissie provided experimental evidence that predominantly microtubules are altered by electroporation and that they play a role in the stabilization process of permeabilized cell membranes, presumably due to their interconnections with plasma membrane. These experiments were done on attached CHO

Effects of Electroporation of Mammalian Cells on Cytoskeleton and. . .

7

WTT cells and on suspension of erythrocytes, using ten square wave electric pulses of 100 μs duration and different electric field intensities ranging from 500 to 2400 V/cm (Rols and Teissie 1992; Teissie et al. 1994). Shortly after that, the phenomenon of microtubule depolymerization was demonstrated in chick embryo corneal fibroblast following one exponentially decaying electric pulse of varying voltage-to-distance ratio 500–1000 V/cm and capacitance 125–960 μF. Besides depolymerization of microtubules, perinuclear collapse of intermediate filament vimentin was observed, and, similarly to CHO WTT cells, actin filaments were not affected (Harkin and Hay 1996). The first study on endothelial cell was performed by Kanthou et al. (Kanthou 2006) evaluating the role of cytoskeleton of endothelial cells in the observed vascular effects of electroporation in vivo. In vivo vascular effects of electroporation and electrochemotherapy were demonstrated in tumors and also in muscles after electroporation alone. In vivo studies showed that application of electric pulses causes profound reduction in blood flow, which was reversible. The return of blood flow to pretreatment values was quicker for the normal tissue (muscle) than for tumors (Jarm et al. 2010). Primary human umbilical vein endothelial cells were used in the study. Besides cytoskeletal proteins, microtubules, actin filaments, and vimentin, endothelial monolayer permeability was determined and, also for the first time, the intercellular connections adherens junctions, which constitute a major type of endothelial junction responsible for the control of vascular permeability. Endothelial cells were grown in Transwell inserts, with a pore size of 0.4 μm till they reached confluence, when the experiments were performed. In situ electroporation was performed by a cuvette-type electrode that was designed to encompass the cells growing in the inserts. Due to this setup, a lower electric field strength was needed to effectively permeabilize cells; three square wave electric pulses of 10–80 V with a duration of 100 μm were applied to the cells via the electrodes with 4 mm distance. Immunofluorescence microscopy was used to determine the morphological changes of cytoskeletal filaments and Western blot analysis to determine the phosphorylation status of myosin light chain indicating on the contractility of endothelial cells and passage of fluorescently labeled dextran through the endothelial monolayer to measure the permeability of endothelium. The results of the study showed that electroporation of adherent endothelial cells causes immediate but transient disruption of microtubules and actin filaments. In addition, the loss of contractility and vascular endothelial cadherin (VE-cadherin) from the adherens cell junction was also observed. All these changes lead to the increase in endothelial monolayer permeability. In contrast to previous studies on CHO WTT cells and corneal fibroblast, in the study on endothelial cell, actin filaments were altered. This can be explained by the use of different pulse conditions in the study on fibroblasts and lack of anchorage dependence in CHO WTT cells, while in endothelial cells, actin filaments play a crucial role in maintenance and control of vascular permeability. This is done through the cell contractility that is regulated by the actin-myosin interaction requiring actin polymerization and phosphorylation of myosin light chain and through the integrity of cell-to-cell adherens junction (Kanthou et al. 2006).

8

M. Cemazar

The impact of experimental setup, i.e., adherent, clustered, or cells in suspension and density of the cells and their shape, on the effectiveness of permeabilization and intercellular connections was a topic of several theoretical and experimental studies (Pucihar et al. 2006, 2007a, b, 2008a, b). CHO cells, either parental strain or WTT clone, were used in these studies. A numerical computation of induced transmembrane voltage needed to successful permeabilization of the irregularly shaped cells was performed and the results verified experimentally on CHO cells exposed to five electric pulses of 150 ms duration at 100 V/cm. The finite element numerical model proved to be a valuable tool, especially for computing the induced transmembrane voltage needed for successful permeabilization of irregularly shaped adherent cells or cells in tissue (Pucihar et al. 2006). The importance of cell density on the effectiveness of permeabilization was demonstrated in a study by Pucihar et al. (Pucihar et al. 2007a). CHO WTT cells in suspension were used in these experiments, the density of the cells varied from 1  106 to 400  106 cells/ml, and the authors used different parameters of electric pulses; they applied eight square wave electric pulses of 100 μs duration with electric field strength of 400 or 700 V/cm. Flow cytometric measurement of number of propidium iodide-positive cells, as well as mean fluorescence intensity of these cells, was used to measure the fraction of permeabilized cells and amount of loaded propidium iodide (electroloading). The increased cell density resulted in ~50 % lower fraction of permeabilized cells. Furthermore, the decrease of electroloading was even larger with increased cell density than expected from the difference in the permeabilized cell fraction. This was attributed to the cell swelling that occurs after permeabilization, also due to the rearrangement of cytoskeletal filaments, and reduced availability of propidium iodide to the permeabilized membrane. The same phenomenon affected the resealing of the membrane of cells that was slower in dense cell suspension compared to dilute one (Pucihar et al. 2007a). Furthermore, transport of propidium iodide was studied in adherent cells compared to cells in suspension. Again, CHO WTT cells were used in experiments. The experiments were performed at the same cell densities in both cases (1  105 cells/ ml). Cells were either grown in suspension and just before the experiments spinned down and plated in a chamber slide or they were plated in the chamber slide one day before the experiment, so that they formed attached confluent monolayer. Fluorescence of propidium iodide was used to measure the kinetics of transport into the cells and fluorescence microscope equipped with sensitive photomultiplier (PM) tube. Fluorescence signal entering PM tube was then transformed to a voltage signal, amplified and filtered with a custom-made differential amplifier, and stored in a transient recorded, which was afterward digitalized and processed on a computer using suitable software program. The cells in chamber slides were exposed to single square wave pulse of various amplitudes (350, 500, 650, 800 V) and/or durations (0.1, 0.5, 1.0, 3.0 ms) that were applied through two parallel wire electrodes (diameter 0.5 mm, length 10 mm, interelectrode distance 5 mm). The results of experimental part of the study demonstrated that at the same pulse parameters, transport into the cells growing as confluent monolayer was considerably lower than transport into the cells in suspension (Pucihar et al. 2008a). This result was in

Effects of Electroporation of Mammalian Cells on Cytoskeleton and. . .

9

agreement with previously published numerical and experimental results on dense cell suspension, further supporting the important role of the induced transmembrane voltage in different types of cell culture conditions, where cells have different arrangements of cytoskeleton filaments as well as intercellular connections. The first study that addressed the role of intercellular connections with regard to the induced transmembrane voltage and consequently the effectiveness of electroporation was theoretical and experimental study on adherent CHO cells grown in clusters (Pucihar et al. 2007b, 2008b). Two types of modeling were performed. In the first one, the cells in the cluster were modeled as electrically connected, so the electric current between the cytoplasm of the cells in the contact was not impeded by the plasma membranes separating them. In the second model, the cells in the cluster were considered as electrically insulated, so that the cells would behave as individual cells. Calculated regions of increased membrane conductivity corresponded to the experimentally observed electroporated regions but only if the cells were modeled as electrically insulated. The cells were exposed to one pulse of 400 V and 200 μs duration. On the contrary, when the cells were exposed to low, nonporating electric field, they behaved as electrically connected, implicating on the role of communication channels between the cells, i.e., gap junctions, which were demonstrated that exist in CHO cells (El-Fouly et al. 1987). Further studies are needed to undoubtedly demonstrate the role of gap junctions in the electroporation process. Recently, a destabilization of plasma membranes was analyzed by atomic force microscopy, which enables a direct observation of plasma membranes of fixed and live cells. CHO cells were grown on coverslip for fixed cell experiments and in Petri dishes for living cells observation. Cells were exposed to eight square wave electric pulses of 5 ms duration at 400 V/cm at 1 Hz repetition frequency through stainless steel electrodes that were applied directly onto the glass coverslip or plastic Petri dish. The results of this study for the first time demonstrated that atomic force microscopy can be used for electroporation process investigations. Transient rippling of the membrane surface was visualized, and a decrease in membrane elasticity by 40 % was measured (Chopinet et al. 2013). Further atomic force microscopy studies on live CHO cells growing in Petri dish, exposed to the same electric pulses condition as in previous study, demonstrated that electroporation affects actin filaments also in CHO cells, which was in contrast to previously published results in the same cell line (Chopinet et al. 2014; Rols and Teissie 1992). The results of the study showed that only one form of actin filament is affected, namely, the actin filament that forms the cell cortex, while other cytoplasmatic actin filaments are not destabilized by electroporation. Other types of cytoskeletal filaments were not investigated in this study (Chopinet et al. 2014). As already mentioned, different cell types, as well as different growing conditions, can greatly influence the electroporation-induced response of cells with respect to the involvement of cytoskeletal filament and intercellular connections. Further insights on the specific response of different cell types provided two studies that were performed on different tumor cells and fibroblasts (Xiao et al. 2011; Pehlivanova et al. 2012). In the study of Xiao et al, hepatocellular carcinoma cells HepG2 were selected. This cell line is commonly used in genotoxicological studies

10

M. Cemazar

as it is one of few hepatic cell lines that have retained the activity of metabolic enzymes. The attached cells grown in a glass microscopic chamber were exposed to 30 electric pulses of 450 ns duration, frequency 1 Hz at the field intensity of 8000 V/ cm. It should be noted that in this study, the pulsing condition was different, compared to the studies evaluating effects of electroporation mentioned above. Nevertheless, in this first study on tumor cells evaluating the role of cytoskeleton, only actin cytoskeletal filaments were studied, in a way that they were disrupted before electroporation by treatment of cells with cytochalasin B, which causes actin filaments depolymerization. The actin filaments were not directly observed, but rather apoptosis, necrosis, propidium iodide uptake, and mitochondrial transmembrane potential were measured after exposure of cells to electric pulses. The results of this study demonstrated that depolymerization of actin filaments protects the cells from electroporation-induced apoptosis and necrosis, presumably due to the lack of signal transduction into the cell to mitochondria. On the other hand, depolymerization of actin filaments had no effect on the electroloading of propidium iodide, i.e., permeabilization of the membrane (Xiao et al. 2011). Further support for different responses of different types of cells to electroporation provided a study by Pechlivanova et al., who used in the experiments two different types of human breast cancer cells, MDA-MB-231, which has loosely adhesive properties, and MCF-7, as well as 3T3 fibroblast cell line. Cells were grown either on glass coverslips or directly on the bottom of the 24-well plates. Cells were exposed to eight biphasic electric pulses of 50 μs duration, with 20 μs interval between both phases and 880 μs pause between the pulses. The amplitudes of applied electric pulses were between 200 and 1000 V, and the electric pulses were delivered through two parallel stainless steel electrodes with a 10 mm distance between them. Cell adhesion and survival of cells after electroporation were measured by crystal violet and 3-(4,5-dimethylthiazol-2-yl)-5(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assay, while cytoskeletal rearrangements and cellular interconnections were visualized by immunofluorescence. The main conclusion of this study, based on the results obtained, is that cell adhesion and the type of adhesions (cell to cell or cell to matrix) is cell-type specific, namely, cell adhesion and survival of 3T3 fibroblast and MCF-7 breast carcinoma cells, were not affected significantly by the applied electroporation. On the other hand, the electroporation of invasive MDA-MB-23 cells results in increased adhesion at lower electroporation voltages and decreased cell adhesion at 1000 V/cm. Actin filaments were transiently depolymerized in fibroblasts, while in both types of cancer cells, actin filaments were not depolymerized, but were rearranged in stress fibers, lamellipodia, and filopodia. Furthermore, based on the electroporation-induced adhesion changes, in MDA-MB-231 cells, cell-to-matrix adhesion dominates, while in MCF-7 cells, cell-to-cell adhesion. However, the types of intercellular connections were not studied (Pehlivanova et al. 2012). Only in a few studies, investigators have exposed cells to nanosecond electric pulses electroporation (nanoporation, nanosecond pulsed electric fields) and studied their influence on cytoskeleton, intercellular gap and tight junctions, nuclear membrane, and telomeres (Stacey et al. 2011; Pakhomov et al. 2014; Steuer et al. 2016).

Effects of Electroporation of Mammalian Cells on Cytoskeleton and. . .

11

Three different types of cells were used in the study by Stacey et al., Jurkat leukemia cells, which grow as suspension, and two adherent cell lines, HeLa human epithelioid cervical carcinoma and SV40 fibroblasts. The cells were exposed to one electric pulse of 60 ns duration at the field intensity of 60,000 V/cm. Jurkat cells were more severely damaged by application of electric pulses, demonstrated by induction of breakdown of actin filaments, localization of telomeres out of the nucleus, and shredded nuclear membrane, compared to adherent cells. The authors suggested that the minimal protection of the cytoskeleton offered to cell in suspension allows the effects of the nanosecond pulsed electric fields to be more readily felt at the nucleus, whereas in adherent cell, much of the energy may be absorbed by the extensive cytoskeleton, with no apparent effect on the nucleus (Stacey et al. 2011). In a second study, Pakhomov et al. showed that actin filaments are disrupted by exposure of CHO cells to four 600 ns long pulses of 19,200 V/cm (Pakhomov et al. 2014). In a third study by Steuer et al., rat WB-F344 liver epithelial cells were used (Steuer et al. 2016). Cells grown in monolayers were exposed to 20 nanosecond pulses of 100 ns duration with varying amplitudes between 5,000 and 35,000 V/cm, which are sublethal conditions for nanosecond pulses. Cell viability, actin filament morphology, and gap and tight junctions were studied. The results of the study showed that exposure to nanosecond pulsed electric fields temporarily changes the ability of cells to communicate through gap junctions in a time and field strengthdependent manner. The recovery was observed within 24 h. The immunohistochemical staining of Cx43 protein, a component of gap junctions, demonstrated disassembly of these junctions, while staining of ZO-1 protein in tight junctions demonstrated that tight junctions are not significantly affected. Actin cytoskeleton was reorganized and not depolymerized as demonstrated in a study by Stacey et al. (Steuer et al. 2016; Stacey et al. 2011)

The Effects of Electrochemotherapy and Gene Electrotransfer on Cytoskeleton and Intercellular Connections Electrochemotherapy is a cancer therapy, where electroporation is used to enhance the delivery of cytostatic drugs into the cells. It is already widely used in European cancer centers for treatment of mainly cutaneous cancers. Several mechanisms are involved in the antitumor effectiveness of electrochemotherapy, among others, also vascular disrupting action. To further elucidate the observed vascular effects in vivo on preclinical tumor models and in treatment of patients with electrochemotherapy, a study was performed comparing the effect of electroporation and electrochemotherapy with bleomycin in human endothelial cells. In that study, human microvascular endothelial cells HMEC-1 were selected, which are a good model for human tumor endothelium. The cells were grown in eight chamber slides for determination of morphological changes and in Transwell inserts with pore size of 3 μm for determination of endothelial monolayer permeability. The electrodes used in this study had a gap of 7.3 mm to fit into the chamber of the slide, and the following parameters of electric pulses were used: 50–500 V/7.3 mm and eight

12

M. Cemazar

Fig. 2 The effect of electrochemotherapy with 300 nM bleomycin at different times posttreatment on actin cytoskeleton of attached HMEC-1 cells exposed to eight 100 μs long pulses at various amplitudes. The electric pulses were applied via two stainless steel electrodes 7.3 mm apart. Note that the actin filaments are not affected at 100 V; some cell can recover after electrochemotherapy applied at 300 V, while at 500 V all the cells die within 24 h. The experiment and acquisition of the images were done by Dr. Cecil Meulenberg

pulses of 100 μs duration with repetition frequency of 1 Hz. In the electrochemotherapy experiments, 3–1000 nM bleomycin was used. The results of this study confirmed the results on electroporation-induced changes in cytoskeleton of previous study performed on human umbilical vein endothelial cells. In addition, for the first time, electrochemotherapy-induced changes in cytoskeleton of endothelial cells and permeability of endothelial monolayer were demonstrated. Electrochemotherapy induced quicker onset and more pronounced increase in endothelium monolayer permeability compared to electroporation alone. In addition, also the changes in cytoskeletal filament morphology had quicker onset and were unrecoverable (Fig. 2), while when using electroporation alone, the changes of cytoskeletal filaments were reversible. This study provided further insight into the cellular changes of the vascular disrupting action of electrochemotherapy (Meulenberg et al. 2012). Gene electrotransfer is a method, where electroporation is used to enable the entry of nucleic acids, mainly plasmid DNA, to enter the cells in vitro and also in vivo. This method is used for gene therapy and DNA vaccination. The therapies are not in routine clinical practice as electrochemotherapy, but the progress in this field is very quick. The role of actin filaments during gene electrotransfer was studied in CHO cells growing either in suspension or attached. The cells in suspension were exposed to ten electric pulses of 5 ms duration and field strength of 700 V/cm, while for the attached cells, lower field strength was used, 400 V/cm. The results of the study

Effects of Electroporation of Mammalian Cells on Cytoskeleton and. . .

13

demonstrated that actin patches colocalized with the DNA at the plasma membrane and thus provide, for the first time, evidence that actin filaments are involved in DNA electrotransfer through the plasma membrane (Rosazza et al. 2011). Further studies from the same group, using the same cell line CHO and the same electrical parameters, demonstrated that for the intracellular trafficking of DNA, microtubule filaments with motor proteins kinesins and dyneins are responsible for fast (200–1600 nm/s) and long-range transport of DNA toward the nucleus. Actin filaments together with the myosin might be responsible to transfer the DNA, either in endocytotic vesicle or free plasmid DNA toward the microtubule network (Rosazza et al. 2013).

Conclusion In conclusion, electroporation alone or in combination with cytostatic drug or plasmid DNA affects the cytoskeleton and intercellular connections. These effects differ dependent on the cell type. Using reversible electroporation, with electric pulses of up to 100 μs duration with voltages up to 2000 V/cm, the effects are reversible and, for example, in endothelial cells, cause increase of endothelial monolayer permeability which contributes to the vascular effect of electroporation in vivo. In combination with cytostatic drug bleomycin, these effects are irreversible and resulted in vascular disrupting action of electrochemotherapy. Nanosecond pulsed electric fields have more pronounced effects on cytoskeleton leading to cell death. Sublethal nanosecond pulsed electric fields cause disassembly of gap junction intercellular communication. The same junctions are involved in the observed behavior of clustered cells, where they behave as one giant cell, if exposed to nonporating electric pulses. Furthermore, cytoskeletal filament microtubules and actin filament are also involved in the gene electrotransfer. Altogether, studies on molecular mechanisms that are involved in the process of electroporation are very important; thus, further studies are needed to fully elucidate the involvement of cytoskeleton and intercellular connection in different types of electroporation application, which will possibly lead to further development and improvement of treatment options based on electroporation. Acknowledgements This work was conducted within the scope of LEA EBAM (FrenchSlovenian European Associated Laboratory: Pulsed Electric Fields Applications in Biology and Medicine) and COST Action.

Cross-References ▶ Cancelation Effect of Nanosecond Pulse Electric Fields on Cells in Vitro ▶ Cell Death Due to Electroporation ▶ Cell Stress Responses to Pulsed Electric Fields

14

M. Cemazar

▶ Membrane Permeabilization Lifetime in Experiments ▶ Nucleic Acid Electrotransfer in Mammalian Cells: Mechanistic Description

References Backert S, Boehm M, Wessler S, Tegtmeyer N (2013) Transmigration route of Campylobacter jejuni across polarized intestinal epithelial cells: paracellular, transcellular or both? Cell Commun Signal 11:72. doi:10.1186/1478-811X-11-72 Blangero C, Rols MP, Teissie J (1989) Cytoskeletal reorganization during electric-field-induced fusion of Chinese hamster ovary cells grown in monolayers. Biochim Biophys Acta 981 (2):295–302 Chopinet L, Roduit C, Rols MP, Dague E (2013) Destabilization induced by electropermeabilization analyzed by atomic force microscopy. Biochim Biophys Acta 1828(9):2223–2229. doi:10.1016/j.bbamem.2013.05.035 Chopinet L, Dauge E, Rols MP (2014) AFM sensing cortical actin cytoskeleton destabilization during plasma membrane electropermeabilization. Cytoskeleton 71(10):587–594. doi:10.1002/ cm.21194 El-Fouly M, Trosko JE, Chang CC (1987) Scrape-loading and dye transfer. A rapid and simple technique to study gap junctional intercellular communication. Exp Cell Res 168:422–430 Harkin DG, Hay ED (1996) Effects of electroporation on the tubulin cytoskeleton and directed migration of corneal fibroblasts cultured within collagen matrices. Cell Motil Cytoskeleton 35 (4):345–357 Jarm T, Cemazar M, Miklavcic D, Sersa G (2010) Blood flow-modifying and vascular-disrupting effects of electroporation and electrochemotherapy – implications in treatment of melanoma metastases. Expert Rev Anticancer Ther 10(5):729–746 Jaroszeski MJ, Gilbert R, Fallon PG, Heller R (1994) Mechanically facilitated cell-cell electrofusion. Biophys J 67(4):1574–1581 Kanthou C, Kranjc S, Sersa G, Tozer GM, Zupanic A, Cemazar M (2006) The endothelial cytoskeleton as a target of electroporation based therapies. Mol Cancer Ther 5 (12):3145–3152 Meulenberg CJ, Todorovic V, Cemazar M (2012) Differential cellular effects of electroporation and electrochemotherapy in monolayers of human microvascular endothelial cells. PLoS One 7(12), e52713 Pakhomov AG, Xiao S, Pakhomova ON, Semenov I, Kuipers MA, Ibey BL (2014) Disassembly of actin structures by nanosecond pulsed electric field is a downstream effect of cell swelling. Bioelectrochemistry 100:88–95. doi:10.1016/j.bioelechem.2014.01.004 Pehlivanova VN, Tsoneva IH, Tzoneva RD (2012) Multiple effects of electroporation on the adhesive behaviour of breast cancer cells and fibroblasts. Cancer Cell Int 12:9 Pucihar G, Miklavcic D (2011) A numerical approach to investigate electrofusion of cells of different sizes. In: 5th European conference of the international federation for medical and biological engineering, vol 37 of the series IFMBE Proceedings. pp 1326–1329 10.1007/978-3642-23508-5_342 Pucihar G, Kotnik T, Valic B, Miklavcic D (2006) Numerical determination of transmembrane voltage induced on irregularly shaped cells. Ann Biomed Eng 34(4):642–652 Pucihar G, Kotnik T, Teissie J, Miklavcic D (2007a) Electropermeabilization of dense cell suspensions. Eur Biophys J 36(3):173–185 Pucihar G, Kotnik T, Miklavcic D (2007b) An experimental and numerical study of the induced transmembrane voltage and electroporation on clusters of irregularly shaped cells. IFMBE Proc 16:639–642 Pucihar G, Kotnik T, Miklavcic D, Teissie J (2008a) Kinetics of transmembrane transport of small molecules into electropermeabilized cells. Biophys J 95(6):2837–2848

Effects of Electroporation of Mammalian Cells on Cytoskeleton and. . .

15

Pucihar G, Kotnik T, Miklavcic D (2008b) Time-dependent model of induced transmembrane voltage and electroporation on clusters of cells. IFMBE Proc 20:623–627 Rems L, Usaj M, Kanduser M, Rebersek M, Miklavcic D, Pucihar G (2013) Cell electrofusion using nanosecond electric pulses. Sci Rep 3:3382. doi:10.1038/srep03382 Rols MP, Teissie J (1992) Experimental evidence for the involvement of the cytoskeleton in mammalian cell electropermeabilization. Biochim Biophys Acta 1111(1):45–50 Rosazza C, Escoffre JM, Zumbusch A, Rols MP (2011) The actin cytoskeleton has an active role in the electrotransfer of plasmid DNA in mammalian cells. Mol Ther 19(5):913–921. doi:10.1038/ mt.2010.303 Rosazza C, Buntz A, Rieß T, Wöll D, Zumbusch A, Rols MP (2013) Intracellular tracking of singleplasmid DNA particles after delivery by electroporation. Mol Ther 21(12):2217–2226. doi:10.1038/mt.2013.182 Stacey M, Fox P, Buescher S, Kolb J (2011) Nanosecond pulsed electric field induced cytoskeleton, nuclear membrane and telomere damage adversely impact cell survival. Bioelectrochemistry 82 (2):131–134. doi:10.1016/j.bioelechem.2011.06.002 Steuer A, Schmidt A, Labohá P, Babica P, Kolb JF (2016) Transient suppression of gap junctional intercellular communication after exposure to 100-nanosecond pulsed electric fields. Bioelectrochemistry 112:33–46. doi:10.1016/j.bioelechem.2016.07.003 Teissie J, Rols MP (1994) Manipulation of cell cytoskeleton affects the lifetime of cell-membrane electropermeabilization. Annals of the New York Academy of Sciences 720:98–110, Electrical Injury: A Multidisciplinary Approach to Therapy, Prevention, and Rehabilitation Wickstead B, Gull K (2011) The evolution of the cytoskeleton. J Cell Biol 194(4):513–525. doi:10.1083/jcb.201102065. PMCID: PMC3160578 Xiao D, Tang L, Zeng C, Wang J, Luo X, Yao C, Sun C (2011) Effect of actin cytoskeleton disruption on electric pulse-induced apoptosis and electroporation in tumour cells. Cell Biol Int 35(2):99–104. doi:10.1042/CBI20100464

Nucleic Acid Electrotransfer in Mammalian Cells: Mechanistic Description Muriel Golzio and Marie-Pierre Rols

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Why Choosing Electroporation for Nucleic Acid Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electropermeabilization of the Cell Membrane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plasmid DNA Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electric Field Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DNA/Membrane Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Internalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Intracellular Trafficking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RNAi-Based Small Oligonucleotide Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . siRNA Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modified Oligonucleotides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Toward In Vivo Delivery of Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 4 5 6 7 7 8 9 9 10 11 12 12 13 13

Abstract

The development of new drugs depends greatly on their successful, efficient, low cost, and safe delivery into target cells or tissues. In the case of highly charged macromolecules such as nucleic acids, the therapeutic effectiveness is mainly limited by their bio-distribution within the tissue and the poor permeability of the plasma membrane of cells. For this purpose, electroporation appears as a promising method for nucleic acid delivery. Electroporation is a physical method of

M. Golzio (*) • M.-P. Rols (*) Institut de Pharmacologie et de Biologie Structurale, Université de Toulouse, CNRS, UPS, Toulouse, France e-mail: [email protected]; [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_21-1

1

2

M. Golzio and M.-P. Rols

vectorization that consists of application of electric pulses on cells or tissues. Optimization of the pulses’ parameters leads to the transient permeabilization of the plasma membrane for molecules which otherwise cannot enter the cell. Therefore, the understanding of different principles of drug and gene delivery is necessary and needs to be taken into account according to the specificity of their delivery to tumors and/or normal tissues. This approach has been routinely used in cell biology for more than 30 years for cell transfection and in medicine in a number of clinics and hospitals through Europe to treat cutaneous cancers by increasing the toxicity of anticancer drugs (electrochemotherapy); it is also now under clinical trials for nucleic acid delivery (electrogenotherapy, electrovaccination). The present chapter focuses on electrotransfer of nucleic acids, the nature of nucleic acids (plasmid DNA, mRNA, siRNA, LNA, etc.) which can be electrotransferred, and the mechanism of their electrotransfer.

Keywords

Delivery • Electroporation • Nucleic acid • Plasmid DNA • siRNA

Introduction The development of new drugs depends greatly on their successful, efficient, and safe delivery into target cells or tissues. Physiological barriers and toxicity of carrier systems create multiple challenges to bring molecules to the clinic. An ideal method efficiently delivers therapeutic molecules while being biocompatible, safe, and targeted. In this context, electroporation (EP) is a promising non-viral biophysical method for in vitro and in vivo delivery of various molecules such as drugs (Mir et al. 1998) and nucleic acids (Rols 2010). EP was introduced in 1970s and consists of the direct application of external electric field pulses on target cells or tissues that transiently destabilize the plasma membrane and cause its permeabilization (Neumann et al. 1982). The efficiency and convenience (i.e., easiness, low cost, and speed) of this technique lead to an increasing number of in vivo applications on a large number of both internal and surface organs and tissues (Golzio et al. 2010). Moreover, very few side effects have been reported (mostly superficial burn) emphasizing the innocuousness of this method for clinical use. In addition, no change in the expression profile of major tumor suppressor genes or oncogenes, of genes involved in the stability of DNA, and no promotion of tumor genesis were detected. The expression of metastasis-promoting genes was not increased after electrochemotherapy. To date, several preclinical and clinical studies using EP for cancer treatment have demonstrated encouraging results showing antitumor effectiveness (Sersa et al. 2008; Bodles-Brakhop et al. 2009). In this chapter, the mechanism of nucleic acid delivery is described according to the nature of nucleic acids which can be electrotransferred and the associated clinical applications.

Nucleic Acid Electrotransfer in Mammalian Cells: Mechanistic Description

3

Why Choosing Electroporation for Nucleic Acid Delivery Several options are available for nucleic acid delivery. Broadly, these can be categorized into local or systemic deliveries. Systemic delivery should be avoided as it is associated with the following problems: intravascular degradation (serum endonuclease cleavage, phagocytosis by scavenger macrophages), activation of complement, nonspecific targeting (distribution takes place to all tissues, except for the brain), fast elimination through renal filtration, and potential toxicity (Aigner 2007). In addition, cancer tissue is difficult to reach intravenously as, compared to normal capillaries, tumor blood vessels are immature and display functional defects that lead to restricted intra-tumor blood flow (Jain 2005). Finally, intravenously injected nucleic acids must extravasate and move up to the tissue core. In contrast, using EP method, nucleic acids can be delivered locally and can thus bypass all the mentioned limitations. Moreover, it has been shown that the application of electric pulses on the skin in vivo resulted in a rapid increase in vascular permeability that gradually recovered to basal levels at different times posttreatment, depending on dextran size (Bellard et al. 2012). Simultaneously, the immediate constriction of the blood vessels occurred that was more pronounced for arterioles compared to venules. This vasoconstriction of arterioles results in a transient “vascular lock” (Jarm et al. 2010). The increased permeability of small vessel walls whatever the dextran size associated with delayed perfusion explains the improved delivery of the intravenous injected molecules (i.e., drugs, gene delivery) into the tissues induced by electropermeabilization in vivo (Fig. 1). Systemically injected nucleic acids must avoid intravascular degradation (I), renal clearance (II), and nonspecific targeting (III); access the accessible tissues (IV); extravagate (V); diffuse through the complex extracellular matrix (VI); be up taken by the cell (VII); and escape the endosome (VIII). On the contrary,

Fig. 1 EP overcomes physiological barriers to delivery (Adapted from Chabot et al. (2015))

4

M. Golzio and M.-P. Rols

electrotransferred nucleic acids are injected locally (1), migrate rapidly through the extracellular matrix (2) thanks to the electric field (E) and can also exit from the blood vessel (20 ), enter the cell through the permeabilized membrane (3), and have access to the cytoplasm (4). One other EP advantage is that the delay between the nucleic acid injection and the targeted delivery is very short (Rols 2010). In fact, cells are embedded into a more or less complex extracellular matrix that can prevent nucleic acid diffusion. Therefore, they cannot reach and enter cells easily as they are retained in the stroma where they are more exposed to degradation. By dragging nucleic acids through the extracellular matrix, EP first increases access to target cells and then protects nucleic acids from degradation by the nucleases present, namely, in tumor fluids (Golzio et al. 2010). Nucleic acids must enter into the cells, but the hydrophobic cell membrane creates a barrier for intracellular delivery of highly negatively charged molecules such as nucleic acids. To overcome these delivery hurdles, non-viral and viral strategies have been developed (Zuhorn et al. 2007; Aagaard and Rossi 2007). EP utilizes naked material, which makes it more cost-effective and improves its safety profile compared to carrier systems and virus production. Moreover, the size of the nucleic acid is less limited using EP method than in the case of the virus approach. Although many nucleic acid carriers have been successfully used for in vitro applications, these delivery systems are usually inappropriate for in vivo use due to their high intra- and extracellular toxicity or unsatisfactory tissue delivery. Most of the carrier systems enter cells through endocytosis (Bumcrot et al. 2006). Thus, molecules must undergo endosomal escape to reach its target before acidification in lysosomes. Additionally, retention in endosomes must be avoided as nucleic acids could activate immune response through Toll-like receptors (TLRs) which are present in these sites. Electric pulse application results in the formation of transient permeable structures (Rols and Teissie 1990) also described as pores at the cell plasma membrane level causing transient membrane permeabilization which may avoid the endocytotic pathway at least for RNAi-based small oligonucleotide delivery (as shown in Fig. 1). For larger nucleic acids such as plasmid DNA, a more complex multistep mechanism is involved in the electrotransfer process, as described in detail below.

Electropermeabilization of the Cell Membrane The exposure of living cells to short and intense electric pulses induces positiondependent changes of the transmembrane potential difference (Neumann et al. 1982; Hibino et al. 1991). Cells initially have a resting transmembrane potential difference which is uniform all along their plasma membrane. The application of an electric field superimposes an electro-induced transmembrane potential, Δψi, to the resting

Nucleic Acid Electrotransfer in Mammalian Cells: Mechanistic Description

5

transmembrane potential. The value of Δψi is given by the equation of Schwan considering the cell membrane as a thin spherical dielectric shell: Δψi ¼ f  gðλÞ  r  E  cos ðθÞ; where f is a factor depending on the cell shape; g(λ) a parameter depending on the conductivities λ of the membrane, of the extracellular medium, and of the cytoplasm; r the radius of the cell; E the electric field strength; and θ the angle between the electric field direction and the normal to the membrane at the considered point of the cell surface. Being dependent on the angle θ, the electric field effect is not uniform along the cell membrane, and its maximum effects are present at the poles of the cells facing the electrodes (Rols and Teissie 1990). When the resulting transmembrane potential reaches a threshold value (close to 200–400 mV), transient permeant structures (TPS) are generated at the cell membrane level. The lifetime of these permeant structures also called “electropores” is in the order of minutes (Rols and Teissie 1990). The surface area of the cell membrane and the density of these TPS are controlled by the electric field parameters (Rols and Teissie 1990). The molecular structure of TPS in terms of lipids and/or proteins is not yet elucidated. The basic model suggests the formation of pores which would give the external molecules a direct access to the cytoplasm (Neumann et al. 1982). However, if this model is in agreement with the observed post-pulse-free diffusion of molecules smaller than 4 kDa into the cytoplasm, the delivery of larger molecules such as plasmid DNA cannot be simply described by “electropores” (Rols 2010).

Plasmid DNA Delivery Plasmid DNA electrotransfer was first published in the early 1980s (Neumann et al. 1982) and is, up to now, the most studied and used nucleic acid in electroporation application. It basically consists in mixing naked DNA in an appropriate buffer together with cells and exposing them to a series of well-controlled electric pulses. Plasmid DNA-encoded protein is expressed within a few hours after the electroporation. Whereas initial procedures could cause considerable cell damage, technical developments have led to improvements of the equipment (pulse generators, electrodes) and optimization of protocols. The use of square wave pulse generators allows for the control of the electrical parameters (intensity, duration, frequency, number) independently. The user can therefore adapt them according to the molecule, the cell, and the tissue to load. The possibility to optimize the parameters of the pulses makes the approach very versatile. The number of publications showing the efficiency of the technique has grown exponentially (Rols 2010). Despite the fact that the gene expression level does not reach the one of the viral methods, EP is one of the most efficient non-viral approaches.

6

M. Golzio and M.-P. Rols

Fig. 2 Schematic presentation for the electrotransfer of small molecules. Schematic cartoon showing the processes occurring during application of electric pulses for the delivery of molecules of small molecules ( > Ep, viability loss; (3) best values)

The use of video microscopy allows visualization of the permeabilization phenomenon at the single-cell level. ▶ Fluorescent indicators of membrane permeabilization, such as propidium iodide (PI), are very convenient to detect the electrotransfer of molecules into the cytoplasm. They can simply be added to the cells before application of the electric pulses. In aqueous solution, their fluorescence is very low. When present into the cells, the fluorescence quantum yield of PI increases in such a way that it is not necessary to remove the solution to detect the molecule inside the cell. Once cells are placed on the stage of the microscope and a region of interest is selected (where the electric field will be applied, thanks to the electrodes connected to the pulse generator), the experiments can start. The uptake of the fluorescent dye into the cells will be the signature of membrane electropermeabilization. As explained in more details in other chapters of this handbook and whatever the value of the pulses duration T, permeabilization only appears above a threshold value of pulse intensity E, called Ep. Therefore, the first experiment to perform consists of submitting the cells to increasing values of E and determining the permeabilization efficiency (i.e., the percentage of cells that have been electropermeabilized, cells which nuclei become fluorescent). For E < Ep, which in the example of Fig. 1 is equal to 0.4 kV/cm, no permeabilization occurs. Above E, increasing E leads to the progressive permeabilization of the whole cell population that is obtained at 0.8 kV/cm. Then, the next step is the determination of the cell viability. Cell viability is indeed affected when membranes are electropermeabilized due to different reasons including the release of intracellular compounds such as ATP. For field values higher than 0.9 kV/cm, viability is affected. Once obtained, such kind of results easily allows to define the best conditions for membrane permeabilization and also for gene electrotransfer. In the example shown in Fig. 1, the electric field values that can be used range from 0.6 to 1.0 kV/cm.

Gene Delivery by Electroporation In Vitro: Mechanisms

5

Fig. 2 The different steps involved in membrane electropermeabilization and gene electrotransfer. (1) As directly visualized under a microscope a few seconds following pulses delivery; (2) as drawn from 1. Left: Before pulse application, the membrane is a barrier that prevents the passage of plasmid DNA (blue spots) toward the cell. Middle: During electric pulse application, the plasma membrane is electropermeabilized facing the two electrodes and DNA aggregates are formed. This interaction takes place only on the membrane facing the cathode. Right: 2 h after electric pulse application, DNA molecules have begun to reach the nucleus and expression can be detected

Kinetics of Membrane Permeabilization Electropermeabilization of cells is a fast process that can be detected immediately after the application of electric pulses. Usually, transport across the membrane is not homogeneous on the whole cell membrane. It occurs at the sides of the cells facing the electrodes in an asymmetrical way where it is more pronounced at the anodefacing side of the cells than at the cathode (Fig. 2), i.e., in the hyperpolarized area than in the depolarized area, which is in agreement with both theoretical and experimental considerations as explained in other chapters of this Handbook of Electroporation (▶ Transmembrane Voltage Induced by Applied Electric Fields; ▶ Pore Lifetime and Permeabilization Lifetime in Models; ▶ Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Experiments). Electropermeabilization can be described as a three-step process in respect to electric field: (i) before electropulsation, the plasma membrane acts as a physical barrier that prevents the free exchange of hydrophilic molecules between the cell cytoplasm and external medium; (ii) during electropulsation, when pulses parameters have been correctly defined, E > Ep, the formation of transient permeable structures facing the electrodes allows the exchange of molecules; propidium iodide is observed to rapidly access the cell interior in the region of the cells facing the electrodes, mainly at the anode-facing site; and (iii) after electropulsation, the membrane can stay permeable before resealing occurs (Golzio et al. 2002). Lifetime of permeabilization can be assayed by adding the fluorescent dyes at various times following the pulses. If the cell membrane is still permeable, then the cell will be fluorescent. Resealing varies from a few seconds (when cells are put at 37  C just after pulsation) to several hours (when cells are maintained on ice) according to the experimental conditions

6

M.-P. Rols

(temperature and pulse parameters). However, one has to take into account that viability can be affected since ATP release will occur. It is therefore better to avoid to maintain the cells at low temperature after pulse delivery. Whatever the molecules used to detect permeabilization (if they are small enough and charged), a direct transfer into the cell cytoplasm is observed. When added after electropulsation, molecules can still penetrate into the cells but less efficiently because electric field acts on both the permeabilization of the membrane and on the electrophoretic drag of the charged molecules from the bulk into the cytoplasm. The electrotransfer mechanism involved is indeed specific for the physicochemical properties of the molecule (Paganin et al. 2011). Progress in the knowledge of the involved mechanisms, in particular in the elucidation of membrane structures that are responsible for molecules transfer, is still a biophysical challenge. Hydrophilic pores have been proposed to be created and their formation confirmed by molecular dynamics modeling. But their existence in permeabilized cells has still to be proven. Phospholipid scrambling and changes on lateral mobility of proteins have been observed suggesting that part of the membrane surface is occupied by defects or pores and that these structures propagate rapidly over the cell surface (Escoffre et al. 2014). The fact that the entire cell surface is affected was not so obvious since permeabilization is only induced in specific regions of the cells. So, even if the entire mechanisms of membrane electropermeabilization/electroporation are not fully understood, and the existence of the exact structures responsible for molecules uptake is still a debate, this physical method of vectorization has become one of the most efficient for gene delivery.

Electrotransfer of DNA Molecules into Cells What Is Known About the Process The first electroporation-mediated gene transfer experiment was published more than 30 years ago (Neumann et al. 1982). Translation to the clinic benefited from increased knowledge of the mechanisms involved in the electrotransfer of nucleic acids during the last three decades. As for electropermeabilization, single-cell studies aided in describing the process of DNA electrotransfer. In addition to membrane permeabilization, DNA electrotransfer is dependent on DNA electrophoresis. The oligonucleotide must indeed be present during the pulse to be later on transferred in the cytoplasm. The electrophoretic mobility of pDNA is not dependent on its number of base pairs. Short pulses with high field strength can be used but are less effective than long pulses with lower field strength. Therefore, pulse parameters have to be determined to lead the membrane to be permeable (E > Ep) while preserving as much as possible cell viability (above 30–50 %). Reporter genes (as the green fluorescent protein or β-galactosidase) are useful to optimize the protocol. As for electropermeabilization, single-cell microscopy and fluorescent plasmids can be used to visualize and determine the different steps of electrotransfection. Plasmids can be labeled with fluorescent dyes (as toto, yoyo,

Gene Delivery by Electroporation In Vitro: Mechanisms

7

yopro) to allow visualization of its electrotransfer. DNA molecules, which are negatively charged, migrate electrophoretically when submitted to the electric field. Under electric fields which are too small to permeabilize the membrane (E < Ep), the DNA simply flows around the membrane in the direction of the anode. Beyond the critical field value, above which cell permeabilization occurs (E > Ep), the DNA interacts with the plasma membrane.

DNA/Membrane Interaction Interaction only occurs at the pole of the cell opposite the cathode, and this demonstrates the importance of electrophoretic forces in the initial phase of the DNA/membrane interaction. When the DNA/membrane interaction occurs, the formation of “microdomains” whose dimensions lie between 0.1 and 0.5 μm is observed (Fig. 2). Also seen are clusters or aggregates of DNA which grow during the application of the field. However, once the field is cut, the growth of these clusters stops. DNA electrotransfer can be described as a multistep process: the negatively charged DNA migrates electrophoretically toward the plasma membrane on the cathode side where it accumulates. This interaction, which is observed for several minutes, lasts much longer than the duration of the electric field pulse. Translocation of the plasmid from the plasma membrane to the cytoplasm and its subsequent passage toward the nuclear envelope takes place with a kinetics ranging from minutes to hours.

Dynamic of the Process DNA/membrane interaction and as a direct consequence gene expression depend on electric pulse polarity, repetition frequency, and duration. Both are affected by reversing the polarity and by increasing the repetition frequency or the duration of pulses. These observations revealed the existence of two classes of DNA/membrane interaction: (i) a metastable DNA/membrane complex from which DNA can leave and return to external medium and (ii) a stable DNA/membrane complex, where DNA cannot be removed, even by applying electric pulses of reversed polarity. Only DNA belonging to the second class leads to effective gene expression (Faurie et al. 2010). Dynamics of membrane/complexes formation has been poorly understood because direct observations have been limited to time scales that exceed several seconds. However, experimental measurement of the transport of plasmid DNA and propidium iodide with a temporal resolution of 2 ms has been performed, thanks to the high speed and sensitive camera, and allowed the visualization of the DNA/membrane interaction process during pulse application (Escoffre et al. 2011). Plasmid complexes, or aggregates, start to form at distinct sites on the cell membrane during the first pulse. Increasing the number of pulses does not lead to the creation of new sites, but to the increase in the amount of DNA. The formation of plasmid complexes at fixed sites suggested that membrane domains may be responsible for

8

M.-P. Rols

Table 1 Kinetics of the different steps involved in gene delivery Time scale μs ms

s

Steps involved in DNA electromediated delivery Plasma membrane facing the electrodes is permeabilized Electrophoretic migration of DNA toward the membrane DNA/membrane complex formation

min

Conversion of the metastable form of the DNA/membrane complex to a stable one DNA translocation/diffusion across the membrane

hour

DNA migration toward the nucleus along the cytoskeleton

day

Gene expression

References Golzio et al. (2002) Escoffre et al. (2011) Golzio et al. (2002) Faurie et al. (2010) Escoffre et al. (2011) Rosazza et al. (2011; 2013) Rosazza et al. (2016)

DNA uptake, and their lack of mobility (as directly observed under the microscope or quantify by fluorescence return after photobleaching (FRAP) measurements) could be due to their interaction with the actin cytoskeleton. As will be described later in this chapter, several publications reported evidences for the involvement of cytoskeleton (Rosazza et al. 2011; 2013). The dynamics of the entire process is reported in Table 1. If pulse delivery occurs in a relative short time scale (μs to ms), the subsequent traffic of plasmid DNA occurs during the minutes and hours following pulse delivery.

DNA Transfer Through the Cytoplasm The process of plasmid transfer through the cellular cytoplasm to the nuclear envelope is a complex process (Lechardeur and Lucaks 2006). In principle micro-sized aggregates of DNA or vesicles filled with DNA could be too large to pass through the pores formed by electroporation. However, individual DNA molecules, while they can pass through electropores, have a limited mobility within the cell and may well be totally degraded before reaching the nucleus. It is possible and worth investigating the possibility that the actin cytoskeleton reacts to the presence of DNA aggregates and plays an important role in the subsequent intracellular transport. It seems reasonable that only aggregates beyond a certain size (a few hundred nanometers) can induce a biological cellular response and can be transported by the cell. In addition, the fact that the DNA is in aggregate form means that the DNA in the center of the aggregate is relatively protected from degradation. Therefore, for gene therapy purposes, it is optimal for DNA to enter the cell as single molecules, but the subsequent transport toward the nucleus is, for biological (possibly by inducing a response of the actin cytoskeleton) and physical (diminishing enzymatic degradation) reasons, optimized if the DNA is in a microsized aggregate form.

Gene Delivery by Electroporation In Vitro: Mechanisms

9

Even if the first stage of gene electrotransfection, i.e., migration of plasmid DNA toward the electropermeabilized plasma membrane and its interaction with it, becomes understood, guidelines to improve gene electrotransfer can not only result from the way pulse parameters have been selected. Expression of the pDNA is controlled by the viability of the pulsed population, and successful expression of the plasmid depends on its subsequent migration into the cell. Therefore, the intracellular diffusional properties of plasmid DNA, as well as its metabolic instability and nuclear translocation, represent cell-limiting factors that must be taken into account. The cytoplasm is composed of a network of microfilament and microtubule systems, along with a variety of subcellular organelles present in the cytosol. The mesh-like structure of the cytoskeleton, the presence of organelles, and the high protein concentration mean that there is a substantial molecular crowding in the cytoplasm which hinders the diffusion of plasmid DNA. These apparently contradictory results might be reconciled by the possibility of a disassembly of the cytoskeleton network that may occur during electropermeabilization and are compatible with the idea that the cytoplasm constitutes an important diffusional barrier to gene transfer. In the conditions induced during electropermeabilization, the time a plasmid DNA takes to reach the nuclei is significantly longer than the time needed for a small molecule (hours compared to minutes). Therefore, plasmid DNA present in the cytosol after being electrotransferred can be lost before reaching the nucleus, for example, because of cell division. Finally, after the cytoskeleton, the nuclear envelope will represent the last, but by no means the least important, obstacle for the expression of the plasmid DNA.

Passage Through the Nuclear Envelope and Gene Expression A high transport does not always result in a high level in expression. The relatively large size of plasmid DNA (2–10 MDa) makes it unlikely that the nuclear entry occurs by passive diffusion. Single-particle tracking (SPT) experiments of individual DNA aggregates in living cells showed how electrotransferred DNA is transported in the cytoplasm toward the nucleus. The modes of DNA aggregate motion in CHO cells have been analyzed. Fast active transport of the DNA aggregates occurs over long distances. Tracking experiments in cells treated with different drugs affecting both the actin and the tubulin networks clearly demonstrate that this transport is related to the cellular microtubule network (Fig. 3, Rosazza et al. 2016).

Active Transport of DNA Aggregates Several studies point toward the contribution of endocytosis in the electrotransfer of DNA, but more investigations have to be performed in order to understand what type (s) of endocytosis would be involved. It is necessary to understand as well how electric fields could stimulate such processes. Also notably, any endocytosis model would only explain the internalization of large molecules as it does not support the

10

M.-P. Rols

– Plasmid DNA E +

Clathrin

Caveolin

Adapter protein

Myosin VI

Dynein

Trajectories

Myosin V

Kinesin

Expressed protein

Macropinocytosis

2

Clathrin-ME

3

1

4 Caveolin/Raft-ME

Actin filaments

Rab11 Rab5

5

Early endosome

Recycling endosome Rab9 Lamp1

Late endosome

6

Microtubules

Lysosome

7 8 Nucleus

Fig. 3 Schematic representation of the mechanism of DNA electrotransfer. During the electric pulses, (1) the plasma membrane is permeabilized, (2) DNA is electrophoretically pushed onto the cell membrane, which results in (3) DNA/membrane interactions. After resealing of the membrane, (4) DNA is internalized by endocytosis and other means where actin may take shape of bursts of polymerization. (5) While being actively transported in the cytoplasm by actin and tubulin networks, DNA aggregates pass through the endosomal compartments. Free DNA interacts with adapter protein in order to be transported by motor proteins. For gene expression to occur, (6) DNA has to escape from endosomal compartments. Once in the perinuclear region, (7) DNA crosses the nuclear envelope to be expressed and (8) yield proteins released. Reprinted with permission from Rosazza et al., Current Gene Therapy 2016

Gene Delivery by Electroporation In Vitro: Mechanisms

11

free membrane crossing of small molecules. It has therefore to be considered to occur in parallel to another model valid for small molecule transmembrane exchange. One model that could reconcile all the DNA internalization models would be that DNA accumulates where pores are formed and that its electrophoretically driven insertion in the membrane pulls the pore and the plasma membrane around. This would generate membrane curvature that could be recognized as an emerging endocytic vesicle and induce a similar response from the cell as for an endocytic process, with the recruitment of actin, clathrin, caveolin, dynamin, and other endocytic regulators (Rosazza et al. 2016; Rems and Miklavcic 2016). Electrotransferred DNA trajectories possess portions of active transport interrupted by phases of nearly immobility (Rosazza et al. 2013). During the phases of active transport, DNA aggregates featured a motion on average having a velocity of 250 nm/s, persisting for 6 s and leading to a displacement of 1.3 μm. However, the distributions were rather broad with velocities from 50 to 3400 nm/s, displacements from 0.1 to 12 μm, and active transport durations from 2 to 30 s. These ranges are in agreement with other types of intracellular particle dynamics as observed for viruses, polyplexes, lipoplexes, receptors, endosomes, and mitochondria. Lower velocities were shown to correspond to actin-associated transport. Indeed, after disruption of the microtubules using the nocodazole drug, active transport of the DNA still occurred, and the measured velocities were in the range expected for myosin motors operating on actin – between 50 and 300 nm/s for myosin VI and between 250 and 500 nm/s for myosin V. In addition to motor-driven transport, actin-related movement could be also due to bursts of actin polymerization which were reported to drive viruses, bacteria, or endosomes from the plasma membrane to the cytosol with mean velocities ranging from 50 to 600 nm/s.

New Challenges to Increase Gene Expression As mentioned above, the dense latticework of the cytoskeleton impedes free diffusion of DNA in the intracellular medium. Electrotransferred plasmid DNA, containing specific sequences, could then use the microtubule network and its associated motor proteins to move through the cytoplasm to the nucleus (Vaughan and Dean 2006). Clear limits of efficient gene expression using electric pulses are therefore due to, in addition to the passage of DNA molecules through the plasma membrane, the cytoplasmic crowding and transfer through the nuclear envelope. One of the key challenges for electromediated gene therapy is to pinpoint the ratelimiting steps in this complex process and to find strategies to overcome these obstacles. One of the possible strategies to enhance DNA uptake into cells is to use short (10–300 ns) but high pulse (up to 300 kV/cm) induce effects that primarily affect intracellular structures and functions. As the pulse duration is decreased, below the plasma membrane charging time constant, plasma membrane effects decrease and intracellular effects predominate. An idea, to improve transfection success, is thus to perform classical membrane permeabilization allowing plasmid DNA electrotransfer to the cell cytoplasm and then after, when DNA has reached the

12

M.-P. Rols

nuclear envelope, to specifically permeabilize the nuclei using these short strong nanopulses. Thus, when used in conjunction with classical electropermeabilization, nanopulses gave hope to increase gene expression (Beebe et al. 2003). However this work was not yet replicated. Another idea is to combine electric pulses and ultrasound assisted with gas microbubbles. Although electroporation induced the formation of DNA aggregates into the cell membrane, sonoporation induced its direct propulsion into the cytoplasm. Sonoporation can therefore improve the transfer of electro-induced DNA aggregates by allowing its free and rapid entrance into the cells (Delalande et al. 2013).

Lipid Vesicles and Spheroids as Other Models to Study Gene Electrotransfer Coming back to a mechanistic point of view and due to the complexity of the composition of the plasma membrane, other experimental tools can be useful to characterize the membrane domains observed during gene electrotransfer. For that purpose, giant unilamellar vesicles (GUV) represent a convenient way to study membrane properties such as lipid bilayer composition and membrane tension (Riske and Dimova 2005). They offer the possibility to study and visualize membrane processes due to their cell-like size in the absence of any constraint due to cell cytoskeleton. They can be obtained by simple methods such as electroformation, and their composition can be very simple (one type of phospholipids) or more complex (several lipids including cholesterol). Experiments showed a decrease in vesicle radius which was observed as being due to lipid loss during the permeabilization process. Three mechanisms responsible for lipid loss were directly observed: pore formation, vesicle formation, and tubule formation, which may be involved in molecule uptake. However, no interaction between plasmid DNA and the GUV membrane could be observed; a direct transfer of DNA into the GUVs took place during application of the electric pulses (Portet et al. 2011). That gives clear evidence that “lipid bubble” is not always relevant as a cell and a tissue is not a simple assembly of single cells. Therefore, it is necessary to develop and use different models, from simple lipid vesicles to tumor multicellular tumor spheroids more closed to the in vivo situation, for the understanding of the membrane permeabilization and DNA electrotransfer process in tissues. Each of these models has advantages and limits. Together combined they can help in the study of the full processes (Table 2). Even if the high majority of studies underlying molecule transfer by electric fields have been performed on 2D cell culture in Petri dish or in cells cultured in suspension, 3D multicellular spheroids represent a nice, relevant, cheap, easy-tohandle in vitro model. Upon growth, spheroids display a gradient of proliferating cells. These proliferating cells are located in the outer cell layers and the quiescent cells are located more centrally. This cell heterogeneity is similar to that found in

Gene Delivery by Electroporation In Vitro: Mechanisms

13

Table 2 What models can address about electropermeabilization and gene delivery processes Model GUV

2D Cell culture 3D Cell culture

Membrane permeabilization Direct visualization of membrane permeabilization and its consequences (deformation, lipid loss) Kinetics of permeabilization and its consequences (lateral and transverse mobility of lipids and proteins) Molecule diffusion and transfer that mimic in vivo complex situation (contacts between cells, junctions, and extracellular matrix)

DNA electrotransfer Failed to address DNA/membrane interaction (DNA is directly transferred inside the vesicle) Visualization of DNA/membrane complex formation and access to DNA traffic into the cells Allow to address DNA delivery in 3D and mimic what happens in vivo (decrease in gene expression from the periphery to the core)

avascular micro-regions of tumors (Sutherland 1988). Confocal microscopy allowed to visualize the repartition of permeabilized cells in spheroids submitted to electric pulses. Results revealed that cells were efficiently permeabilized, whatever their localization in the spheroid, even those in the core, mimicking previously observed in vivo situations. Propidium iodide uptake was observed to be present but spatially heterogeneous within the 3D multicellular spheroid after electroporation, with a progressive decrease from peripheral to interior cells. In the case of large molecules as plasmid DNA, spheroids allowed showing that electrophoresis, and not tissue deformation or electroosmosis, is the driving force for interstitial transport. In addition, and at the opposite of cells in 2D cultures, only cells on one side of the outer leaflet expressed the reporter gene (Gibot and Rols 2013). This low expression is in fair agreement with in vivo experiments on tumors. Close contacts between cells and extracellular matrix (ECM) may act as physical barrier that limit/prevent (uniform) DNA distribution and explain the absence of gene expression in the inner region of spheroid. The limited access of plasmid DNA to the central region of spheroid remains a significant barrier to efficient gene delivery in tissues. Taken together, these results, in agreement with the ones obtained by the group of R. Heller (Marrero and Heller 2012), indicate that the spheroid model is more relevant to an in vivo situation than cells cultured as monolayers and therefore can be useful to address the mechanisms of DNA electrotransfer. In order to assess the effects of the extracellular matrix composition and organization, as well as intercellular junctions and communication, other 3D reconstructed human connective tissue model can be used. Cell sheets, reconstructed in vitro by a tissue engineering approach, presents multiple layers of primary dermal fibroblasts embedded in a native, collagen-rich extracellular matrix (ECM) and can be a useful tool to study skin DNA electrotransfer mechanisms. Cells within this standardized 3D tissue can be efficiently electropermeabilized by millisecond electric pulses. A better comprehension of gene electrotransfer in such a model tissue would help improve electrogene therapy approaches such as the systemic delivery of therapeutic proteins and DNA vaccination.

14

M.-P. Rols

Conclusions Classical theories of electropermeabilization present some limits to give a full description of the transport of molecules through membranes. Certain effects of the electric field parameters on membrane permeabilization, and the associated transport of molecules, are well established, but a great deal of what happens at the molecular level remains speculative. ▶ Molecular Models of Lipid Bilayers and Electropore Formation are giving interesting new insight into the process. Electroinduced destabilization of the membrane includes both lateral and transverse redistribution of lipids and proteins, leading to mechanical and electrical modifications which are not yet fully understood. One may suggest that such modifications, which may vary according to the microenvironment, can be involved in the subsequent transport of molecules interacting with them such as the DNA molecules. Experimental verification of the basic mechanisms leading to the electropermeabilization and other changes in the membrane, cells, and tissues remains a priority given the importance of these phenomena for processes in cell biology and in medical applications. In vivo gene electrotransfer (▶ Principles of Electroporation for Gene Therapy) will face other challenges such as the necessity to control electric field distribution (▶ Electric Field Distribution Modelling in Tissue Considering Tissue Conductivity Increase Due to Electroporation) and gene expression both in space (targeted DNA delivery to the cells) and in time. Guidelines for successful DNA delivery are still required but we can be optimistic that further working to improve gene electrotransfer mechanisms will yield effective treatments. Acknowledgment This research was performed in the scope of the EBAM European Associated Laboratory (LEA) and is a result of networking efforts within COST TD1104. It was supported by the Centre National de la Recherche Scientifique (CNRS), the Agence Nationale de la Recherche (ANR), Projet PIERGEN ANR-12-ASTR-0039, and the Direction Générale de l’Armement (DGA).

Cross-References ▶ Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Experiments by Justin Teissié ▶ Electric Field Distribution Modelling in Tissue Considering Tissue Conductivity Increase Due to Electroporation by Damijan Miklavcic ▶ Electroporation and Electropermeabilization by Lluis Mir ▶ Fluorescent Indicators of Membrane Permeabilization by Tina Batiska ▶ Molecular Models of Lipid Bilayers and Electropore Formation by Tarek Mounir ▶ Nucleic Acid Electrotransfer in Mammalian Cells: Mechanistic Description by Muriel Golzio ▶ Pore Lifetime and Permeabilization Lifetime in Models by James Weaver ▶ Principles of Electroporation for Gene Therapy by Loree Heller ▶ Transmembrane Voltage Induced by Applied Electric Fields by Tadej Kotnik

Gene Delivery by Electroporation In Vitro: Mechanisms

15

References Beebe SJ, White J, Blackmore PF, Deng Y, Somers K, Schoenbach KH (2003) Diverse effects of nanosecond pulsed electric fields on cells and tissues. DNA Cell Biol 22:785–796 Daud AI, DeConti RC, Andrews S, Urbas P, Riker AI, Sondak VK, Munster PN, Sullivan DM, Ugen KE, Messina JL, Heller R (2008) Phase I trial of interleukin-12 plasmid electroporation in patients with metastatic melanoma. J Clin Oncol 26:5896–5903 Delalande A, Kotopoulis S, Postema M, Midoux P, Pichon C (2013) Sonoporation: mechanistic insights and ongoing challenges for gene transfer. Gene 525(2):191–199 Escoffre JM, Portet T, Favard C, Teissie J, Dean DS, Rols MP (2011) Electromediated formation of DNA complexes with cell membranes and its consequences for gene delivery. Biochim Biophys Acta 1808:1538–1543 Escoffre JM, Bellard E, Faurie C, Sebai SC, Golzio M, Teissie J, Rols MP (2014) Membrane disorder and phospholipid scrambling in electropermeabilized and viable cells. Biochim Biophys Acta 1838:1701–1709 Faurie C, Rebersek M, Golzio M, Kanduser M, Escoffre JM, Pavlin M, Teissie J, Miklavcic D, Rols MP (2010) Electro-mediated gene transfer and expression are controlled by the life-time of DNA/membrane complex formation. J Gene Med 12:117–125 Gibot L, Rols MP (2013) Progress and prospects: The use of 3D spheroid model as a relevant way to study and optimize DNA electrotransfer. Curr Gene Ther 13:175–181 Golzio M, Teissie J, Rols MP (2002) Direct visualization at the single-cell level of electrically mediated gene delivery. Proc Natl Acad Sci U S A 99:1292–1297 Lambricht L, Lopes A, Kos S, Sersa G, Préat V, Vandermeulen G (2016) Clinical potential of electroporation for gene therapy and DNA vaccine delivery. Expert Opin Drug Deliv 13:295–310 Lechardeur D, Lukacs GL (2006) Nucleocytoplasmic transport of plasmid DNA: a perilous journey from the cytoplasm to the nucleus. Hum Gene Ther 17:882–889 Marrero B, Heller R (2012) The use of an in vitro 3D melanoma model to predict in vivo plasmid transfection using electroporation. Biomaterials 33:3036–3046 Neumann E, Schaefer-Ridder M, Wang Y, Hofschneider PH (1982) Gene transfer into mouse lyoma cells by electroporation in high electric fields. Embo J 1:841–845 Paganin-Gioanni A, Bellard E, Escoffre JM, Rols MP, Teissie J, Golzio M (2011) Direct visualization at the single-cell level of siRNA electrotransfer into cancer cells. Proc Natl Acad Sci U S A 108:10443–10447 Portet T, Favard C, Teissié J, Dean DS, Rols M-P (2011) Insights into the mechanisms of electromediated gene delivery and application to the loading of giant vesicles with negatively charged macromolecules. Soft Matter 7:3872–3881 Rems L, Miklavčič D (2016) Tutorial: electroporation of cells in complex materials and tissue. J Appl Phys 119:201101 Riske KA, Dimova R (2005) Electro-deformation and poration of giant vesicles viewed with high temporal resolution. Biophys J 88(2):1143–1155 Rosazza C, Escoffre JM, Zumbusch A, Rols MP (2011) The actin cytoskeleton has an active role in the electrotransfer of plasmid DNA in mammalian cells. Mol Ther 19:913–921 Rosazza C, Buntz,A, Riess,T, Woll D, Zumbusch,A, Rols MP (2013) Intracellular tracking of single plasmid DNA-particles after delivery by electroporation. Mol Ther 21:2217–2226 Rosazza C, Meglic SH, Zumbusch A, Rols MP, Miklavcic D (2016) Gene electrotransfer: a mechanistic perspective. Curr Gene Ther 16(2):98–129 Serša G, Teissié J, Čemažar M, Signori E, Kamenšek U, Marshall G, Miklavčič D (2015) Electrochemotherapy of tumors as in situ vaccination boosted by immunogene electrotransfer. Cancer Immunol Immunother 64:1315–1327 Sutherland RM (1988) Cell and environment interactions in tumor microregions: the multicell spheroid model. Science 240:177–184

16

M.-P. Rols

Vaughan EE, Dean DA (2006) Intracellular trafficking of plasmids during transfection is mediated by microtubules. Mol Ther 13:422–428 Verma IM, Weitzman MD (2005) Gene therapy: twenty-first century medicine. Annu Rev Biochem 74:711–738 Yarmush ML, Golberg A, Sersa G, Kotnik T, Miklavcic D (2014) Electroporation-based technologies for medicine: principles, applications, and challenges. Annu Rev Biomed Eng 16:295–320

Electroporation of Biofilms Flavien Pillet

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Biofilms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Biofilm Life Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Membrane Permeabilization of Planktonic Bacteria by Pulsed Electric Fields . . . . . . . . . . . . . . . . . Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PEF Disinfection of Biofilms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications In Vitro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications In Vivo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bioelectric Effect on Biofilms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 2 2 3 5 5 5 6 7 7 8 8 9 9 10 12 12 13

Abstract

The extraordinary resistance of bacterial biofilms against the antimicrobials and human immune system raises a grave issue in the medical field. This resistance is due to the confinement of bacteria into a highly dense extracellular matrix. This matrix serves as a filter and increases the concentration needed to kill bacteria from 100 to 1000 times, in comparison with free-floated bacteria (planktonic). F. Pillet (*) CNRS; IPBS (Institut de Pharmacologie et de Biologie Structurale), Toulouse, France Université de Toulouse; UPS; IPBS, Toulouse, France e-mail: fl[email protected]; [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_27-1

1

2

F. Pillet

Within context, the development of new therapeutic strategies is required. To achieve this aim, the pulsed electric fields (PEFs) are a promising technology. The purpose of this chapter is to introduce the complexity of the biofilms and the phenomenon involved in their formation, maturation, and proliferation. The mechanisms of adhesion and the roles of extracellular matrix are specifically discussed. Then, the chapter considers two different strategies which use the PEF technology to inactivate the biofilms. The first approach is direct disinfection of biofilms by PEF. In this case, strong electric fields are applied during a short time. The aim is to induce a bacterial inactivation by irreversible damages of the cell envelope by the production of reactive oxygen species. This method is efficient to prevent in vivo, the biofilm formation on burned wound. The second strategy is to expose biofilm with low electric fields during a long period and in combination with antimicrobials. This technique is called the bioelectric effect and allows removing of the antibiotic resistance of biofilms. The mechanism is not fully understood, but the observed effect is probably due to mass transfer electrophoresis and disruption of the extracellular matrix. Keywords

Bacterial biofilms • Extracellular matrix • PEF disinfection • Bioelectric effects

Introduction Pulsed electric fields (PEFs) are now currently used in cancer therapy to induce a reversible permeabilization of the plasma membrane in cancer cells and induced an entrance of cytotoxic molecules. This treatment is named the electrochemotherapy, and until now, thousands of patients were treated in Europe and the USA. In the food industry, PEFs are employed for the sterilization of pathogenic bacteria, obtained by the irreversible permeabilization of the plasma membrane and the cell-wall disruption. Interestingly, PEFs are not still exploited in medicine to inactivate the bacterial biofilm. Through a few significant examples, the aim of this chapter is to investigate the feasibility of PEF to treat the biofilms.

The Biofilms Background Biofilms are one of the most primitive living form and existing for more than 3.2 billion years. In the classical definition, a biofilm is a community of microorganisms densely self-organized in an extracellular matrix and adhered to an inert or biological support. Most of microorganisms (bacteria, yeasts, and algae) are able to form biofilms and can be found in all environments, even the most hostile, like in extreme conditions of temperatures, pH, or in nutrient-poor medium.

Electroporation of Biofilms

3

This high resistance raises serious problems in the industrial sector, the food industry, and the medical field. In the industrial field, biofilms can be involved in biofouling and biodegradation of metals on pipelines and ships (Schultz et al. 2011). In some instances, the biofilms may be highly pathogenic for human. For example, the Legionella pneumophila colonization in water systems (domestic water supply, cooling towers) causes legionellosis, a fatal respiratory disease. In the food industry, the biofilms promote survival and growth of pathogenic bacteria on food contact surfaces, such as Listeria monocytogenes, the causative agent of listeriosis, a serious infection lethal in 20–30% of cases. In the medical field, biofilms are responsible of 80% of hospital-acquired infections associated with the implantation of medical device (e.g., catheter, stents, orthopedic instrumentation, prosthetic heart valves, pacemakers). Staphylococcus epidermidis and S. aureus are the most frequent origins of these nosocomial infections, and the human and economic cost is considerable. For instance, S. epidermidis causes each year in the USA 100,000 deaths and a cost of $45 billion (Harris et al. 2016). In addition, biofilms may also be involved in chronic lung diseases. Biofilms of Pseudomonas aeruginosa can be found in patients infected by the cystic fibrosis, and the presence of Mycobacterium tuberculosis biofilms is suspected in the case of tuberculosis. These few examples are illustrative of the scale of problem of biofilms in various sectors. The eradication of the biofilms is therefore a critical challenge. In the area of medicine, the highly antibiotic resistance of biofilms is the main obstacle to their eradication. This resistance is due to the secretion of extracellular matrix which acts as a filter. Therefore, the bacteria within biofilms are 100–1,000 times more resistant to antimicrobial agents than bacteria in suspension (Ceri et al. 1999). Consequently, the concentration of antibiotics required is far too high to be used correctly without risk to the patients. An alternative is to remove the infected material, but the surgical operation may be dangerous for the patient, and the exposure to new infection is high. For chronic diseases, like the cystic fibrosis, the ablation of infected tissues is clearly not an option thus no treatment is available.

The Biofilm Life Cycle Biofilm formation starts with the attachment of free bacteria in suspension (planktonic) to a biotic or abiotic surface (Fig. 1). According to the adhesion forces between the cells and the surface, the attachment can be reversible or irreversible. In this latter case, bacteria forms microcolonies and the production of the extracellular matrix is initiated. During the maturation step, the division of bacteria takes place within the biofilm community protected against environment by the matrix. The proliferation step is characterized by the detachment of planktonic bacteria and biofilm fragments, leading to colonization of new surfaces and/or a systemic infection. The adhesion of planktonic bacteria is a complex process that may vary greatly, depending to bacterial species and surface. Bacterial adhesion is mediated by a large diversity of molecular interactions, according to the support. In the initial step

4

F. Pillet

Fig. 1 Steps in biofilm formation

(from seconds to minutes), the attachment of bacteria is facilitated by nonspecific interactions (van der Waals forces, electrostatic forces, and hydrophobicity). For example, several proteins contribute to the hydrophobic properties of bacteria, like AtlE and Bap in S. epidermidis. However, these nonspecific interactions are not the primary driving force toward adhesion, and specific interactions are required to initiate the formation of biofilms. The second step (from minutes to hours) of adhesion is due to a multitude of molecular recognitions between receptors and ligands. To name a few, the family of cell-anchored proteins in S. epidermidis are able to recognize specific molecules on the surface targeted, like the protein Sbp which binds the abiotic surfaces and epithelial cells or the protein SdrF which interacts specifically with the human collagen of type I. In addition, the bacterial appendages (pili, fimbriae) are also involved in specific interactions with the surface to be colonized. During the maturation (few hours to few days), the extracellular matrix is secreted (Flemming and Wingender 2010). Also known as extracellular polymeric substances (EPS), the matrix is a three-dimensional heterogeneous network mostly composed of proteins, carbohydrates, lipids, and extracellular DNA (eDNA). In an infectious context, the main role of the matrix is to induce a high resistance to antimicrobials and immune system. However, the matrix is involved in other phenomena required for the growth and the development of biofilms. The extracellular matrix is crucial to maintain the cell adhesion, the cohesion, the microbial attachment on the surface and provide the mechanical stability of the biofilm. Furthermore, the matrix allows the retention and the accumulation of nutrients and prevents the desiccation. In few cases the biofilms’ matrix operates as an external digestive system by enzymatic activity. Last but not least, the matrix is primordial for bacterial communication inside the biofilm. The most remarkable phenomenon is the quorum sensing; this process is defined as the range of mechanisms of stimuli and the response to control and coordinate the gene expression between bacteria, according to the cell density. An outstanding example is the bacterial programmed cell death observed in biofilms of Escherichia coli. This phenotypic manifestation of self-destruction of bacteria allows the production of extracellular matrix and nutrients to sibling’s bacteria.

Electroporation of Biofilms

5

Another particularity of matrix is also related to improve the horizontal gene transfer between bacteria, leading to acquisition of new antibiotic resistances. The proliferation is the last step required to the biofilm life cycle and the systemic infection. This process is the result of a balance between the biofilm growth and the release of free bacteria, under the regulation of gene expression. However, these mechanisms are not well understood yet, and only few examples are described in the literature. In Staphylococcus family, the upregulation of Agr gene increases the persistence and the virulence of the biofilm. By contrast, the downregulation of this gene induced the dissemination of planktonic bacteria. As a conclusion, the microbial biofilms are communities of bacteria highly organized and dynamic, whose growth and survival depend on a complex machinery, involving the bacterial adhesion, the biofilm maturation including the secretion of extracellular matrix, and eventually the bacterial proliferation. For decades, the research has been focused on the comprehension of these mechanisms, but no effective treatments have been yet found, and the eradication of biofilm infections in human is currently a huge challenge. Within context, new therapeutic approaches must be found.

Membrane Permeabilization of Planktonic Bacteria by Pulsed Electric Fields Background The first example of bacterial inactivation by electric field was described by Fuller in 1898 for the purification of river water. Then, the first industrial application was reported in 1919, with the Electro-Pure Process device developed by Anderson and Finkelstein for pasteurization of milk by ohmic heating. In 1967, Sale and Hamilton observed a nonthermal inactivation by pulsed electric field (PEF) on bacteria by application of micropulses at 25 kV/cm. This surprising observation led the authors to formulate the hypothesis of irreversible PEF permeabilization of bacteria, as a consequence of the alteration of the membrane by electrical breakdown. This hypothesis was improved by Neumann and Rosenheck in 1972 with the demonstration of ATP leakage from vesicles permeabilized by PEF (Neumann and Rosenheck 1972). Today, this versatile method is used in many domains as diverse as cancer treatment (Rols et al. 1998; Marty et al. 2006) (chapter “▶ Biomedical Applications of Electroporation”) or food industry (Saldaña et al. 2014) (chapter “▶ Pulsed Electric Field Treatment for Fruit and Vegetable Processing”).

Principle Membrane electropermeabilization, also called electroporation, is a nonthermal physical process based on the application of pulsed electric fields to induce a

6

F. Pillet

controlled permeabilization of the plasma membrane of eukaryote cells and prokaryotes. Membrane permeabilization is directly correlated with transmembrane voltage (Escoffre et al. 2009) (chapter “▶ Transmembrane Voltage Induced by Applied Electric Fields”). In the absence of electric field, the resting transmembrane potential (ΔΨ0) is homogenous on the cell. In the presence of electric field, the induced transmembrane potential (ΔΨE) depends on the cell shape (f), the medium and cell conductivity (g), the cell radius (r), the intensity of electric field (E), the orientation of the cosθ angle, the duration of exposure (t), and the charging time of the plasma membrane (τ). When the sum in absolute value of ΔΨ0 + ΔΨE is above 200 mV, the plasma membrane is permeabilized. In this case, there is no permeabilization on the membrane exposed perpendicularly to the electric field. By contrast, there is a maximum permeabilization on the membrane exposed in parallel to the electric field. Furthermore, according to the electrical parameters applied (number of pulses, electric field intensity, frequency, etc.), the permeabilization is reversible or irreversible. However, the cell membrane permeabilization is not only the only phenomenon observed during PEF exposure. In a recent study, a cell-wall disorganization was mentioned after PEF in Bacillus pumilus (Pillet et al. 2016). Morphological and mechanical damages as well as a loss of hydrophobicity were observed at the nanoscale.

Applications The reversible permeabilization is not lethal for bacteria and allows the transfer of biomolecules, like the nucleic acids. This technique, also called electrotransformation, is classically used in molecular biology to transfer exogenous DNA into bacteria. The electric field-mediated transfer is a two-step process, with a fast anchored of DNA during PEF exposure, pursued by a slow crossing of the plasma membrane. Furthermore, it has been proved in E. coli that this mechanism is accompanied by a bacterial orientation parallel to the field lines and a considerable cytoplasm ion leakage (Eynard et al. 1998). The irreversible permeabilization leads to the death of bacteria exposed. There is a wide range of applications. First, the PEFs are currently used in the food industry to inactivate pathogenic bacteria. There are many advantages compared to thermic sterilization: the energetic cost is lower and the preservation of nutrients is better. Indeed, the absence of thermal increase prevents the Maillard reactions which are responsible for the carbohydrate and protein degradation during cooking. For example, orange juices sterilized by PEF are commercialized in many countries, like in the Netherlands (McHugh and Toepfl 2016). Second, the irreversible permeabilization of the cell membrane improves the entrance of antimicrobials molecules. In Bacillus cereus, the PEF exposure facilitates the entry of nisin and increases by 60 times the inactivation rate. Third, PEF may be used for the electroextraction of molecules present in the cytoplasm, like the proteins.

Electroporation of Biofilms

7

Fig. 2 Electropermeabilization principle

To conclude, the electropermeabilization of planktonic bacteria by PEF has developed substantially over the past decades, particularly in molecular biology and food industry. By contrast, the inactivation of bacteria by PEF is still not used in a medical context. This is the consequence of two major reasons. The first is that the cell radius of eukaryotic cells is larger than the bacteria, resulting in a higher sensitivity to PEF (Cf. the equation in Fig. 2). In this context, the PEF treatment to kill bacteria in vivo may be dangerous for the patient. The second reason is that there is still a lack of studies to understand the effects of PEF exposure on bacteria in a three-dimensional conformation, like biofilms. The aim of this chapter is to give an overview of the promising studies which described the PEF to prevent and kill bacterial biofilms. This chapter will be focused on two strategies: (i) direct disinfection of biofilms by irreversible permeabilization of the bacterial cell envelope and (ii) the bioelectric effect to enhance antimicrobial penetration within the biofilms.

PEF Disinfection of Biofilms Principle PEF disinfection of biofilms is based on the bacterial inactivation induced by the exposure to moderate or strong electric fields (between 1.5 and 5 kV/cm), with a wide range of pulse duration and frequency. The mechanism of inactivation is not fully understood, but probably due to irreversible permeabilization of the bacterial cell envelope and/or production of reactive oxygen species (ROS). This part focuses on the few studies on biofilm eradication in two contexts, in vitro and in vivo.

8

F. Pillet

Applications In Vitro One of the most common problems of biofilms is their ability to induce the accumulation of microorganisms and biological material on wetted surface. This phenomenon is called the biofouling and concerns various domains in industry and medicine. A previously work described the effects of pulsed electric fields on biofouling, particularly on the development and inhibition of Pseudomonas aeruginosa biofilms (Perez-Roa et al. 2006). The authors used interdigitated electrodes with 29 μm spacing between 22 μm wide electrodes and a voltage between 0.5 and 5 V. These are approximately electric fields from 200 V/cm to 2 kV/cm. Furthermore, alternating current, square-wave pulses were applied. The effects of applied voltage, the frequency, and the pulsing duty ratio (percentage of pulsing time during PEF exposure) were evaluated on the inhibition of biofilm formation. The best result was obtained with a combination of short pulses and high voltages. A 50% reduction of the electrode area covered by the biofilm was measured at 2 kV/cm, 200 Hz, and 1% of duty ratio (i.e., pulses of 50 μs). However, the mechanisms involved are still unknown, and several reasons may be considered: the disruption of the cellular adhesion process, the electrolyte formation, and biocidal effects of the electric fields. This promising study demonstrates the efficiency of PEF to prevent the colonization of bacteria, leading to biofilm formation and biofouling. On the other hand, the eradication of formed biofilm is also a huge challenge. Indeed, in the human body, the majority of bacterial infections are caused by biofilms. Unfortunately, the inactivation of biofilm is further complicated by their resistance to antimicrobials. A recurrent issue is the mesh infection by biofilm. This is a classical pathology which may occur after a ventral hernia repairs and represents a morbidity in 4–16% of patients. In a recent study, the PEF effect on eradication of biofilm-infected mesh was investigated (Khan et al. 2016). For this, the viability of P. aeruginosa on mesh was evaluated after PEF by bioluminescent imaging. The biofilm disruption on mesh was visualized by scanning electron microscopy (SEM). A concentric electrode system was used to apply various PEFs at different electric field strengths and numbers of pulses. The frequency was fixed at 2 Hz and the pulse duration at 50 ms. In the area treated, the conditions required to inactivate 100–80% of bacteria were 300 pulses at 1.2 kV/cm or 150 pulses at 2.3 kV/cm. Furthermore, the inactivation area between the electrodes was proportional to the number of pulses delivered. This work is encouraging to develop a tailored device to treat infected mesh in human.

Applications In Vivo The PEF technology is already used in medicine, as a nonthermal process to destroy solid tumors by irreversible permeabilization (Jiang et al. 2015) (chapters “▶ Irreversible Electroporation and Its Clinical Applications” and “▶ Effects of Reversible, and Irreversible Electroporation on Endothelial Cells and Tissue Blood Flow”).

Electroporation of Biofilms

9

Fig. 3 PEF disinfection in vivo of a burned wound. A third-degree burn area was inflicted on the mouse skin. Immediately after injury, the infection was initiated with multidrug-resistant bacteria. Next, the PEF was applied on the wound infected, using two-plate electrodes (Golberg et al. 2014, 2015)

Recently few studies applied the PEF to eradicate, in mice model, bacteria deposited on burned wound (Golberg et al. 2014, 2015). The principle is summarized in Fig. 3. The authors describe the inactivation of Acinetobacter baumannii, a bacteria involved in hospital-acquired infections on the lung, wound, and burn. The infected areas were treated 30 min after the infection, with pulses of 70 μs at 5 kV/cm and a frequency of 1 Hz. The viability of bacteria on the skin treated was evaluated by bioluminescence. The inactivation rate is correlated with the number of pulses. After 40 pulses, 2.0 log10 of inactivation was measured, whereas 80 pulses led to 5.5 log10 of inactivation. For 80 pulses, the inactivation rate was almost identical after 3 h, with 4.9 log10. It shows the long-term efficacy of PEF to inactivate bacteria on burned wound. These results are interesting and suggested a bright alternative to prevent the biofilm formation in burned wound. Further experiments should be done to evaluate the efficiency of this method to kill already well-established biofilm in vivo.

Bioelectric Effect on Biofilms Principle The main restriction for the biofilm inactivation in vivo is their high resistance to antibiotics and immune system. This resistance is correlated with the presence of extracellular matrix which prevents the entrance of antibiotics and biocides into the biofilm. Within context, an interesting method, named the bioelectric effect, is discussed in this part. It consists to exposition during long periods (few hours to several days) of low electric fields (less than 10 V/cm), in combination with antimicrobials. The treatment induced phenomenon of electrophoresis results in a disruption of the extracellular matrix and increases the antimicrobial concentration

10

F. Pillet

Fig. 4 Principle of bioelectric effects. The drug delivered into the biofilm is facilitated by the disruption of the extracellular matrix due to PEF electrophoresis. The black sticks show the bacteria and the red dots indicate the antimicrobials

delivered into the biofilms. Thus, this method improves the antibiotic efficiency against bacteria packed in biofilm. The principle is described in Fig. 4.

Mechanism The electrical enhancement of efficacy of antimicrobials to eradicate biofilm bacteria was described for the first time in 1992 by the Costerton team (Khoury et al. 1992). The aim was to enhance the antimicrobial penetration into the biofilm, in order to reduce the antibiotic concentration required to kill bacteria. For this, the authors produced biofilms of S. epidermidis and P. aeruginosa with a homemade device composed by a central flow channel, a peristaltic pump, and stainless steel studs considered as electrodes. The biofilms were exposed until 12 h to an electric field of 1.5 V/cm with a current density of 15 μA/cm2. To limit the reactive oxygen species production, the polarity was inverted each 64 s. The bacterial inactivation rate was calculated by colony counting at different exposure times. The combination of low PEF with an antibiotic (tobramycin) was measured for each bacterial strain tested. For S. epidermidis, the exposure to 2.5 mg/ml of tobramycin during 12 h does not lead to significant bacterial inactivation. By contrast, when the tobramycin was combined with the PEF exposure, all bacteria were inactivated after 8 h. Similarly, with P. aeruginosa, the presence of 8 mg/ml induced a loss of viability around 1 log10 after 12 h, whereas this viability decreased by almost 7 log10 with antibiotic within electric field. These results shown that the low voltage fields can be employed to decrease the necessary inhibitory concentration of antibiotics in biofilms. However, the effects of PEF alone versus the bioelectric effect should be evaluated. To answer, the authors made a complementary study on P. aeruginosa (Costerton et al. 1994). They compared during 48 h the viability of biofilm bacteria exposed to electric fields, in the absence or presence of tobramycin. After 48 h, bacteria are not inactivated by the PEF alone. In comparison, the combination of PEF and low concentration of tobramycin (5 mg/ml) led to a bacterial eradication of more than 6 log10. These observations suggested a synergic effect of antibiotics with the pulsed

Electroporation of Biofilms

11

electric fields. To explain this phenomenon, several hypotheses could be formulated. The first evidence is that electric field can facilitate the displacement of charged molecules across the extracellular matrix by electrophoresis, thus inducing a mass transfer (Jass et al. 1995). Another explanation is that the extracellular matrix becomes permeable after PEF exposure. However, the bioelectric effects are not necessarily generalizable for all biofilm organisms and antimicrobial agents. To illustrate this point, a previous study investigated the bioelectric effect of 11 antimicrobial agents from different categories to improve the inactivation of P. aeruginosa, methicillin-resistant Staphylococcus aureus, and S. epidermidis biofilms (del Pozo et al. 2009). The biofilms were cultivated in a continuous flow biofilm growth reactor during 36 h at 37  C. The PEF exposure was applied 24 h at room temperature in the absence or presence of antimicrobials. The inactivation rate was quantified by colony counting. When the antimicrobial agents were used in the absence of electric fields, there was no significant biofilm inactivation, except with the trimethoprim-sulfamethoxazole for S. epidermidis. In the presence of electric fields and antimicrobials, the bioelectric effect was very variable. For example, a bioelectric effect was observed with the daptomycin (an anionic antibiotic) against S. epidermidis but not with methicillinresistant Staphylococcus aureus. On the other hand, the efficacy of vancomycin (an uncharged antibiotic) was enhanced by electrical fields with methicillin-resistant Staphylococcus aureus and S. epidermidis. The overall conclusion is that the bioelectric effect is not generalizable for all bacterial biofilms and antimicrobials. This is a complex process which included not only the mass transfer by electrophoresis because the bioelectric effect seems independent to the charge of antibiotics. Another critical phenomenon during the electric field exposure is the presence of oxygen intermediates or oxygen gas and the influence of pH. In an earlier study, the impact of these components was evaluated on the bioelectric effect with tobramycin against P. aeruginosa biofilms (Stewart et al. 1999). First of all, the bioelectric effect was measured. In the absence of electric fields, a 5 μg/ml of tobramycin led to 2.9 log10 of bacterial inactivation. The inactivation rate increased until 5.6 log10 when an electric field was applied in combination with tobramycin. Then, the influence of pH was tested. Indeed, the pH decreased during PEF, from 7.2 to 4.5 in this study. No increase of antibiotic efficiency was measured when the buffer composition was changed to obtain a pH similar to PEF exposure. Therefore, the low pH during PEF exposure was not responsible for the bioelectric effect. Next, the sodium thiosulfate was used to neutralize oxygen intermediates during PEF exposure. The inactivation rate has not been reduced in the absence of reactive oxygen species. On the other hand, when gaseous oxygen was added in the absence of electric current, the tobramycin efficacy to eradicate biofilms was increased by almost 2 log10. This study demonstrates that the oxygen gas may improve the bioelectric effect. More recently, a very interesting study focused on the gene expression in S. aureus biofilms exposed to gentamicin within electric fields (Zhang et al. 2014). As previously described in the literature, the antibiotic used alone has limited effects in bacterial viability within biofilms. Furthermore, the biofilm inactivation cannot be done by increasing the antibiotic concentration. After 3 days of exposition, only an

12

F. Pillet

inactivation of 2 log10 was measured with 256 mg/L of gentamicin. By contrast, in the presence of low electric fields (5.6 V/cm, 4 h per day), a concentration of 2 mg/ml of gentamicin was sufficient to obtain 6 log10 of inactivation after 3 days of treatment. To get a better understanding of this bioelectric effect, gene expression was evaluated. The authors have shown that the electric fields (alone or with gentamicin) caused an upregulation of positive autolytic regulator genes and downregulation of negative autolytic regulator genes. But, no effect was observed on cell-cell interaction genes. These results suggested that the electric fields are responsible for bacterial autolysis and then increase the sensitivity to gentamicin. The S. aureus autolysis by electric field exposure was confirmed by scanning electron microscopy.

Applications Bioelectric effect applications are still in its infancy and examples are limited. A recent study described the eradication in vitro by electric effect of a dual biofilm which can be involved in periodontitis (Lasserre et al. 2015). This disease may cause the damaging of the tooth-supporting tissue. The infection is provoked by a dual biofilm of Streptococcus gordonii and Porphyromonas gingivalis. In this work, the viability of P. gingivalis was investigated with dual-species biofilms of 7 days old. The inactivation rate was 79.1% with 0.2% of chlorhexidine and 98.9% when the chlorhexidine was supplemented with low electric current (10 mA). This study is promising to develop a new treatment against oral biofilm bacteria. Finally, a very exciting study concerns the inactivation in vivo by bioelectric effects of Pseudomonas aeruginosa in lung infection (Giladi et al. 2010). Mice were infected in the lung by intranasal administration of P. aeruginosa. The authors demonstrate the efficiency to inactivate bacterial growth in vivo of electric fields (12 V/cm) as monotherapy or in combination with ceftazidime. This proof of concepts is promising to treat the pulmonary infections by bioelectric effects.

Conclusion The use of pulsed electric fields is a promising method to inactivate bacterial biofilm in vitro and in vivo. In order to achieve this aim, two different strategies may be employed. The PEF disinfection is the same principle to that used in the food industry and consists to induce an irreversible permeabilization of the bacterial cell envelope. This strategy is efficient to inactivate bacteria in burned wound but may be more complicated to treat in internal tissue, due to the higher sensitivity of the eukaryotic cells. For the bioelectric effect, the main target is the extracellular matrix which protects the bacteria against the antimicrobials. This strategy could be more secure because the electric field strength applied is very low. However, the mechanisms involved in the combination of electric fields and antimicrobials against biofilms are not well known. Further studies should be done to develop the

Electroporation of Biofilms

13

bioelectric effect and inactivate biofilms in the medicine field. In addition, a major part of biofilms are composed of different bacteria which cohabit in symbiosis, thereby increasing the difficulty to develop effective treatments.

References Ceri H, Olson ME, Stremick C, Read RR, Morck D, Buret A (1999) The calgary biofilm device: new technology for rapid determination of antibiotic susceptibilities of bacterial biofilms. J Clin Microbiol 37:1771–1776 Costerton JW, Ellis B, Lam K, Johnson F, Khoury AE (1994) Mechanism of electrical enhancement of efficacy of antibiotics in killing biofilm bacteria. Antimicrob Agents Chemother 38:2803–2809 del Pozo JL, Rouse MS, Mandrekar JN, Sampedro MF, Steckelberg JM, Patel R (2009) Effect of electrical current on the activities of antimicrobial agents against Pseudomonas aeruginosa, Staphylococcus aureus, and Staphylococcus epidermidis biofilms. Antimicrob Agents Chemother 53:35–40. doi:10.1128/AAC.00237-08 Escoffre J-M, Portet T, Wasungu L, Teissié J, Dean D, Rols M-P (2009) What is (still not) known of the mechanism by which electroporation mediates gene transfer and expression in cells and tissues. Mol Biotechnol 41:286–295. doi:10.1007/s12033-008-9121-0 Eynard N, Rodriguez F, Trotard J, Teissié J (1998) Electrooptics studies of Escherichia coli electropulsation: orientation, permeabilization, and gene transfer. Biophys J 75:2587–2596. doi:10.1016/S0006-3495(98)77704-5 Flemming H-C, Wingender J (2010) The biofilm matrix. Nat Rev Microbiol 8:623–633. doi:10.1038/nrmicro2415 Giladi M, Porat Y, Blatt A, Shmueli E, Wasserman Y, Kirson ED, Palti Y (2010) Microbial growth inhibition by alternating electric fields in mice with pseudomonas aeruginosa Lung infection. Antimicrob Agents Chemother 54:3212–3218. doi:10.1128/AAC.01841-09 Golberg A, Broelsch GF, Vecchio D, Khan S, Hamblin MR, Austen WG, Sheridan RL, Yarmush ML (2014) Eradication of multidrug-resistant A. baumannii in burn wounds by antiseptic pulsed electric field. Technology 2:153–160 Golberg A, Broelsch GF, Vecchio D, Khan S, Hamblin MR, Austen WG, Sheridan RL, Yarmush ML (2015) Pulsed electric fields for burn wound disinfection in a murine model. J Burn Care Res 36:7–13. doi:10.1097/BCR.0000000000000157 Harris LG, Murray S, Pascoe B, Bray J, Meric G, Magerios L, Wilkinson TS, Jeeves R, Rohde H, Schwarz S, de Lencastre H, Miragaia M, Rolo J, Bowden R, Jolley KA, Maiden MCJ, Mack D, Sheppard SK (2016) Biofilm morphotypes and population structure among staphylococcus epidermidis from commensal and clinical samples. PLoS One 11:e0151240. doi:10.1371/ journal.pone.0151240 Jass J, Costerton JW, Lappin-Scott HM (1995) The effect of electrical currents and tobramycin on Pseudomonas aeruginosa biofilms. J Ind Microbiol 15:234–242 Jiang C, Davalos RV, Bischof JC (2015) A review of basic to clinical studies of irreversible electroporation therapy. IEEE Trans Biomed Eng 62:4–20. doi:10.1109/TBME.2014.2367543 Khan SI, Blumrosen G, Vecchio D, Golberg A, McCormack MC, Yarmush ML, Hamblin MR, Austen WG (2016) Eradication of multidrug-resistant pseudomonas biofilm with pulsed electric fields. Biotechnol Bioeng 113:643–650. doi:10.1002/bit.25818 Khoury AE, Lam K, Ellis B, Costerton JW (1992) Prevention and control of bacterial infections associated with medical devices. ASAIO J 38:M174–M178 Lasserre JF, Leprince JG, Toma S, Brecx MC (2015) Electrical enhancement of chlorhexidine efficacy against the periodontal pathogen Porphyromonas gingivalis within a biofilm. New Microbiol 38:511–519

14

F. Pillet

Marty M, Sersa G, Garbay JR, Gehl J, Collins CG, Snoj M, Billard V, Geertsen PF, Larkin JO, Miklavcic D, Pavlovic I, Paulin-Kosir SM, Cemazar M, Morsli N, Soden DM, Rudolf Z, Robert C, O’Sullivan GC, Mir LM (2006) Electrochemotherapy – an easy, highly effective and safe treatment of cutaneous and subcutaneous metastases: results of ESOPE (European Standard Operating Procedures of Electrochemotherapy) study. Eur J Cancer Suppl 4:3–13. doi:10.1016/j.ejcsup.2006.08.002 McHugh T, Toepfl S (2016) Pulsed electric field processing for fruits and vegetables. Food Technol 70:73–75 Neumann E, Rosenheck K (1972) Permeability changes induced by electric impulses in vesicular membranes. J Membr Biol 10:279–290. doi:10.1007/BF01867861 Perez-Roa RE, Tompkins DT, Paulose M, Grimes CA, Anderson MA, Noguera DR (2006) Effects of localised, low-voltage pulsed electric fields on the development and inhibition of Pseudomonas aeruginosa biofilms. Biofouling 22:383–390. doi:10.1080/08927010601053541 Pillet F, Formosa-Dague C, Baaziz H, Dague E, Rols M-P (2016) Cell wall as a target for bacteria inactivation by pulsed electric fields. Sci Rep 6:19778. doi:10.1038/srep19778 Rols MP, Delteil C, Golzio M, Dumond P, Cros S, Teissie J (1998) In vivo electrically mediated protein and gene transfer in murine melanoma. Nat Biotechnol 16:168–171. doi:10.1038/ nbt0298-168 Saldaña G, Álvarez I, Condón S, Raso J (2014) Microbiological aspects related to the feasibility of PEF technology for food pasteurization. Crit Rev Food Sci Nutr 54:1415–1426. doi:10.1080/ 10408398.2011.638995 Schultz MP, Bendick JA, Holm ER, Hertel WM (2011) Economic impact of biofouling on a naval surface ship. Biofouling 27:87–98. doi:10.1080/08927014.2010.542809 Stewart PS, Wattanakaroon W, Goodrum L, Fortun SM, McLeod BR (1999) Electrolytic generation of oxygen partially explains electrical enhancement of tobramycin efficacy against Pseudomonas aeruginosa biofilm. Antimicrob Agents Chemother 43:292–296 Zhang J, Neoh KG, Hu X, Kang E-T (2014) Mechanistic insights into response of Staphylococcus aureus to bioelectric effect on polypyrrole/chitosan film. Biomaterials 35:7690–7698. doi:10.1016/j.biomaterials.2014.05.069

Bacterial Cell Envelopes: Composition, Architecture, and Origin Didier Zerbib

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Diderm-Lipopolysaccharide Bacteria Cell Envelope (Glycobacteria) . . . . . . . . . . . . . . . . . . . . . . . . . . The Outer Membrane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Diderm Cell Wall: The Peptidoglycan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Periplasm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Inner Membrane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monoderm Bacteria Cell Envelope (Firmicutes) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monoderm Peptidoglycan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Surface Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Teichoic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Cytoplasmic Membrane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Corynebacterineae Envelope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . “A Phylum Level Perspective on Bacterial Cell Envelope Architecture” . . . . . . . . . . . . . . . . . . . . . . . Dominance of the Diderm-LPS Cell Envelope Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Phyla with Typical Monoderm Cell Envelopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Diderm Exceptions in the Monoderm Phyla . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Diderm Alternative to Diderm LPS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Exceptions in Phyla Spirochaetes, Proteobacteria, and Actininobacteria . . . . . . . . . . . . . . . . . . . . The Phylum Deinococcus-Thermus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Phylum Thermotogae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bacterial S-Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Archaeal Cell Envelope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Origin of Diderm Bacteria; Selective Pressure or Endosymbiosis? . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 4 5 7 8 8 8 10 10 11 11 11 12 12 13 13 13 13 14 14 14 15 15 18 18

D. Zerbib (*) LISBP, Université de Toulouse, CNRS, INRA, INSA, Toulouse, France IPBS, Université de Toulouse, CNRS, UPS, Toulouse, France e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_28-1

1

2

D. Zerbib

Abstract

The bacterial envelope is a complex layered structure, whose primary function is to protect the cell from the environment. The Gram-staining procedure has been a fundamental method to classify the bacteria for more than 100 years. It was based on the effect of the structure and composition of the bacterial envelope on the staining procedure. It has made it possible to classify the bacteria in two main categories: the Gram-positive, which bind the Gram stain, and Gram-negative bacteria, which do not. Currently, the prokaryotes are classified in 30 phyla of Bacteria and in 5 phyla of Archaea. This classification does not consider the Gram-staining properties which was an excessive simplification. The growing amount of data concerning the genomic sequences of bacteria has significantly amended the view of the bacterial phylogeny. In addition, the access to the genetics of the biogenesis of the cell envelope has also allowed envisioning the bacterial Tree of Life in a different way. The bacterial envelopes are now defined with the number of membranes they contain: the cells are either diderms or monoderms. In addition, the presence or the absence of lipopolysaccharides (LPS) in the diderm envelopes is also a fundamental criterion. This chapter is an overview of current knowledge about the composition and architecture of bacterial envelopes in light of recent data showing that the diversity of structures allows reaching the same main objective, the survival and the protection of the bacterium from its environment. Keywords

Bacterial envelopes • Diderm • Monoderm • Cell wall

Introduction The main reasons for the application of electric fields to bacteria are either to cause their complete destruction after cell lysis or to induce a transient and selective permeability of the envelope to allow introduction or extraction of molecules or macromolecules. To achieve these objectives, even when the electroporation of the membrane is effective, it is necessary to take account of the architecture of the bacterial envelope developed to protect the bacteria from their hostile environment. From a “bacterial point of view,” life is associated to the inside and the death to the outside: the envelope must thus be a very effective protective barrier. The bacterial envelope, whose primary function is probably to protect the cell from the environment (Silhavy et al. 2010; Dufresne and Paradis-Bleau 2015), is an active component; it is essential for survival, division, adaptation, morphogenesis, and pathogenesis. In addition to osmotic pressure protection and bacterial shape control, it enables the entry of nutrients, vitamins, and cofactors and the efflux of toxins and unwanted metabolites (Silhavy et al. 2010). The energy production is controlled by the envelope. Electron transport chains induce the electrochemical proton gradient needed to yield the proton motive force in the bacterial membrane

Bacterial Cell Envelopes: Composition, Architecture, and Origin

3

(Taylor 1983) together with cytoplasmic ATP production by oxidative phosphorylation. The cell envelope is able to sense and transduce signals allowing the bacterium to resist and adapt to environmental stresses and changes (Jordan et al. 2008). The bacterial envelope represents the interface between bacteria and hosts; it plays roles in motility, adherence, and secretion (Silhavy et al. 2010). In addition to its multiple active roles, the envelope represents a physical barrier that protects the bacteria from the environment (Holtje 1998; Silhavy et al. 2010). The fence, or the grid, is mainly constituted by the cell wall, giving the bacterial shape, the bacterial resistance to osmotic pressure, and preventing the envelope from rupture. The fence is “porous” and these holes are chemically filled with many constituents like proteins or polysaccharides organized in supramolecular networks. The fence is coated and reinforced by one or two hydrophobic selective barriers: the bacterial membranes. Very schematically, like mentioned above, electroporation generated by electric fields have been used to perform two types of actions on this bacterial barrier: either their destruction to kill bacteria (see, for example, chapters “▶ Irreversible Electroporation and Its Clinical Applications” and “▶ Different Cell Sensitivity to Pulsed Electric Field” or “▶ Cell Death Due to Electroporation”) or their transient and localized disruption (chapter “▶ Electroporation and Electropermeabilization”) to allow either the input (of DNA, for example, “▶ Gene Delivery by Electroporation In Vitro: Mechanisms”) or output (“▶ Impact of Pulsed Electric Field Treatment on Must and Wine Quality”) of interesting compounds. It is clear that the structure, the chemical composition, the organization, and the biophysical and biochemical properties of the fence will greatly influence the effectiveness of the electric field attack. Even if the main wanted effect when applying electric field is to render the membrane permeable, the entire envelope organization has to be taken into account when the global effect of electric field has to be evaluated in terms of effectiveness (Pillet et al. 2016). The electric field is a sort of unique entity facing manifold envelope types. Fortunately, electric field parameters can be adjusted at several levels (intensity, voltage, number of pulses, pulses frequency, pulses sources, pulses shapes) in order to find specific electroporation or lysis protocols (see the chapter “▶ Electric Pulse Parameters Affecting Electroporation Treatment Outcome”). If the bacterial envelope could be summed up simply by one or two lipid membranes that pulsed electric fields (PEF) should destroy or simply make permeable, only one or two universally efficient electroporation protocols would exist. This is very far from being the case. To simplify bacterial cell envelope descriptions, it is important to give clear definitions of their different types in light of microbiological data but also of the bacterium phyla properties and of the evolution of our knowledge of bacterial phylogeny (Sutcliffe 2010). This vision was greatly enlarged when the amount of genome sequencing data increased in databases. The staining procedure developed by Christian Gram (1884), in addition to being a fundamental microbiological diagnostic method for over a century, allowed very early and very conveniently to classify bacteria into two large groups: the Gram-positive and the Gram-negative bacteria. During years, textbooks associated each group to the actual Phyla of Firmicutes (Gram-positive) or Proteobacteria (Gram-negative), each represented

4

D. Zerbib

by the models Bacillus subtilis (or Staphylococcus aureus), with a single membrane and a cell wall, and Escherichia coli (or Salmonella typhymurium), with two membranes, respectively. The principle of Gram staining lies in the effect of the structural differences of the cell envelope of these two main types of bacteria. Firmicutes are Gram positive because they possess a thick layer of peptidoglycan (PG) that retains the primary dye. Proteobacteria are Gram negative because the thin layer of PG, which is protected by an outer membrane (OM), binds only very few dye molecules that are easily removed when the decolorant used in the Gramstaining procedure disrupted the OM (Beveridge 2001). It is now largely admitted that bacterial envelope architecture has to be defined by referring to the number of cellular membrane and not by Gram-staining properties: diderm-bacteria have two membranes, while monoderms have only one (Gupta 1998). Even if the vast majority of diderms are Gram negative (Proteobacteria) and monoderms are often Gram positive (Firmicutes), there are many examples of monoderm Firmicutes displaying aberrant staining (negative or variable) and of some diderm Firmicutes exceptions (Corynebacterineae) that stain as Gram positive. Bacteria are called diderm bacteria when they have two cellular membranes, regardless of lipid composition (Gupta 1998). Among diderm bacteria, diderm-LPS bacteria contain LPS in their OM and have been called Glycobacteria (CavalierSmith 2006, 2010). Cavalier-Smith (2006) also defined Eobacteria as being the counterpart of Glycobacteria, i.e., diderm bacteria without LPS. However, because diderms like Thermotogae or Fibrobacteres were not present in Eobacteria (Cavalier-Smith 2006) while monoderms like Chloroflexi (see below) were indeed included (Cavalier-Smith 2010), the denomination “Eobacteria” will not be used here for non-LPS diderms. In this chapter a brief summary of the overall composition of the envelopes of typical monoderm and diderm bacteria together with selected examples of particular and atypical species with a phylum perspective is presented. It is more focused on the cell wall composition and organization than on the lipid membranes description and dynamics. Electric field effects on membrane are also beyond the scope of this review and will be discussed in detail in other chapters. This chapter provides a simplified scheme of the different bacterial envelope structures as they have to be taken into account when biotechnologists and biophysicists need to evaluate or improve the effects of electric field on bacterial envelopes of various species.

Diderm-Lipopolysaccharide Bacteria Cell Envelope (Glycobacteria) Typical diderm-LPS bacteria, largely represented by Gram-negative members of Proteobacteria phylum, have a cell envelope composed of three layers (Fig. 1). The OM and the inner membrane (IM), also called cytoplasmic membrane (CM), both delineate the hydrophilic periplasm that contains essentially a thin layer of PG constituting the cell wall (CW). The envelopes of typical diderm have a thickness of 35 nm in average, the membrane thickness varies from 5 to 8 nm, and the PG

Bacterial Cell Envelopes: Composition, Architecture, and Origin

5

Fig. 1 Typical diderm-LPS bacteria envelope. The three layers of an envelope of typical didermLPS (Proteobacteria) are shown schematically here. Phospholipids and Lipid A moiety of the lipopolysaccharides are represented in blue. The transmembrane porins and the lipoproteins of outer membrane are in green. The periplasmic proteins and proteins of the inner membrane are in orange. The carbohydrate units of peptidoglycan are either in blue (MurNAc) or in red (GlyNAc). LPS core and O-antigen are both represented in gray. Peri periplasm, OM outer membrane, IM inner membrane, PG peptidoglycan, LPS lipopolysaccharide

between 6 nm (in E. coli) and 2.4 nm in Pseudomonas aeruginosa (Vollmer and Seligman 2010). Both membranes are hydrophobic and semipermeable. In addition to lipids and carbohydrates, the three layers contain proteins.

The Outer Membrane The OM of Proteobacteria is an asymmetric lipid bilayer, with phospholipids only found in the inner leaflet. In E. coli, phospholipids of the OM consist of 90 % of the zwitterionic phospholipid phosphatidylethanolamine (PE), 6 % of the anionic phospholipid phosphatidylglycerol (PG), and 4 % of the anionic phospholipid cardiolipin

6

D. Zerbib

(CL) (or diphosphatidylglycerol) (Lugtenberg and Peters 1976). In comparison to the IM the inner leaflet of the OM contains more saturated fatty acids, is more rigid, and has a smaller net negative charge. In Glycobacteria, i.e., diderm-LPS bacteria (Cavalier-Smith 2006, 2010), the outer leaflet of the OM is mainly composed of the Lipid A, the lipid moiety of anionic glycolipids: LPS (Raetz and Whitfield 2002). LPS is mainly responsible for the human immune system recognition of bacterial infections. Lipid A is also called the endotoxin, and LPS induces endotoxic shock associated with Gramnegative bacteria septicemia. The LPS itself is a glucosamine disaccharide bearing six or seven acyl chains (lipid A), a polysaccharide core, and an extended polysaccharide chain called the O-antigen (Raetz and Whitfield 2002). LPS plays a critical role as an antigen through the O-antigen. Pathogenic E. coli are classified through their O-antigen properties together with their flagellin content, the major protein constituent of flagella, that is referred as H giving, for example, the E. coli pathogenic strain O157:H7. LPS is also involved in the surface net charge of the envelope and in the overall hydrophobicity of diderm-LPS bacteria. The acyl chains of LPS are saturated and thus favor LPS packing, which is reinforced by the neutralization of LPS negative charges with divalent cations like Mg++ and Ca+ + (Nikaido 2003). Because of its properties, the LPS leaflet is a very effective barrier for hydrophobic molecules, a property that is increased by the hydrophilic selectivity of embedded porins for molecules larger than 700 Daltons (Nikaido 2003). Only two main types of proteins are present in the OM: the lipoproteins and the β-barrel proteins, (Silhavy et al. 2010). There are more than 100 different OM-lipoproteins in E. coli (Nakayama et al. 2012) comprising the very abundant tri-acylated Braun’s lipoprotein (Lpp). Lipoproteins have an N-terminal cysteine that is modified by thioether-linked diacylglycerol and amino-linked acyl chains that allow the embedding of the lipid moieties of lipoproteins in the inner leaflet of the OM. β-Barrel proteins are integral transmembrane proteins called outer membrane proteins (OMPs). OMPs are composed of β-sheets that form cylinders. Some OMPs function as porin allowing the passive diffusion of small molecules and defining part of the OM selectivity (Klebba 2005). Porin like OmpF and OmpC are trimers of 16 transmembrane β-strands monomers. They are not specific and allow the diffusion of mono- and disaccharides but also of amino acids. PhoE, having the same topology, is specific for phosphate. Another specific porin is LamB, which is a trimer of 18 transmembrane β-strands that allow the specific diffusion of maltose and maltodextrins across the OM. An interesting example is the one of OmpA, a monomeric conditional porin. OmpA can adopt two different conformations and be either a porin or participate to the structure of the envelope. Finally, some larger OMPs with 20–24 transmembrane β-strands allow the transport of very large compounds like some vitamins and iron chelates (Nikaido 2003). In comparison to OM enzymes, the protease OmpT (Vandeputte-Rutten et al. 2001) or OMPs like the porins OmpF and OmpC are very abundant (around 250,000 copies/

Bacterial Cell Envelopes: Composition, Architecture, and Origin

7

cell) but are not essential. However, the OM is essential by itself and its essentiality is believed to rely on its protective function.

Diderm Cell Wall: The Peptidoglycan The essential peptidoglycan (murein) is the rigid, bag-shaped (sacculus) exoskeleton of bacteria (Weidel and Pelzer 1964) that determines the cell shape (Turner et al. 2014). It is made of glycan strands cross-linked by short peptides (Vollmer and Seligman 2010). PG layers are covering the CM of nearly all bacteria. However, PG is absent in Mycoplasma from Tenericutes, in some Rhickettsia from Proteobacteria (Tamura et al. 1995), and Planctomyces from Planctomycetes, and even if biosynthetic genes are present, it has never been detected in members of the phylum Chlamydiae (Chopra et al. 1998). The integrity of PG prevents bacteria from cell lysis in low osmolarity media. Enzymatic digestion of PG by lysozyme induces the loss of the cell shape and provokes the formation of spheroplast sensitive to cell lysis in low osmolarity media or in water. Any inhibition of PG biosynthesis by mutation or antibiotic action during bacterial growth will provoke cell lysis. The linear glycan strands of PG are made up of alternating N-acetyl glucosamine (GlcNAc) and N-acetyl muramic acid (MurNAc) linked in β-1 ! 4. MurNAc is substituted by unusual pentapeptide containing rare D-amino acids. In nascent PG, the pentapeptide is often L-Ala—D-Glu—mA2pm (meso-diaminopimelic acid)—DAla—D-Ala. The last D-Ala is generally absent from mature PG. Cross-linking of the linear glycan strands occurs between those peptides, generally between the penultimate D-Ala and the mA2pm at position 3, and form a peptide bridge of seven amino acids. This organization, the length of the glycan strands and the degree and nature of cross-linkage are the most common. However, variations exist, from species to species, and depending on strains and growth conditions. Those variations have been used, since more than 40 years, as a taxonomic criterion. PG is attached to the OM through a small lipoprotein (58 residues) anchored in the inner leaflet of the OM: the Braun’s lipoprotein (Lpp) or “murein lipoprotein” (Braun 1975). The Nterminus of Lpp is linked to the lipids embedded in the OM whereas its C-terminal end is covalently linked to the diaminopimelic acid of the PG pentapeptide. Lpp is the most abundant protein in E. coli with more than 500,000 molecules per cell. Lpp mutants produce extracellular OM vesicles indicating that Lpp is the real anchor of the OM on PG. OmpA also provides an additional link between PG and OM. The chemical composition of PG of numerous species is now very well known. The challenge is to observe the real 3D organization of PG, for which several models have been proposed (Vollmer and Seligman 2010). The classical “horizontal layers model” propose an architecture in which the glycan strands are parallel to the CM (Pink et al. 2000). Conversely, in the “vertical scaffold model,” the peptides are parallel to the membrane whereas the glycan strands run perpendicular to the long axis of rod-shaped bacterium (Meroueh et al. 2006). Even if the layer model is

8

D. Zerbib

favored, the actual 3D architecture of PG in different bacteria is still under investigations (Turner et al. 2014).

The Periplasm The periplasm is defined as the aqueous compartment located between the OM and the CM. It consists in a hydrophilic and oxidizing space between the two membranes, containing the cell wall (Mullineaux et al. 2006). The periplasm is rich in proteins such as degradation enzymes (RNase or alkaline phosphatase); periplasmic binding proteins, involved in sugar and amino acid transport and in chemotaxis; and “chaperones” involved in envelope biogenesis. An interesting view of the periplasm has recently been presented as being an OMP quality control space (Lyu and Zhao 2015). It is indeed a challenge for OMP proteins that are synthesized in the cytoplasm to cross the hydrophilic periplasm without any energy producing system: a role of E. coli periplamic proteins like Skp, SurA, and DegP that help OMPs in such a process has been proposed (Mogensen and Otzen 2005).

The Inner Membrane The IM or cytoplasmic membrane (CM) of diderm is a symmetric phospholipid bilayer. In E. coli, the CM is composed of 75 % of the zwitterionic phospholipid phosphoethanolamine (PE), 20 % of the anionic phospholipid phosphatidylglycerol (PG), and 5 % of the anionic phospholipid cardiolipin (CL) and small amounts of phosphatidylserine (PS). The CM of the Diderm E. coli has a smaller negative charge than the one of the monoderm B. subtilis that is rich in PG (van der Does et al. 2000). In Proteobacteria and in monoderm, other minor lipids are present like the polyisoprenoid lipid carrier undecaprenyl-P that transports activated sugar for envelope biogenesis. Moreover, the diderm CM contains integral α-helical proteins together with lipoproteins. In contrary to IM of monoderm, the diderm CM does not contain any teichoic acid (Silhavy et al. 2010). All the membrane associated functions performed by eukaryotic organelles have to be achieved by the bacterial CM: that is, energy production, lipid biosynthesis, protein transport and secretion; all the proteins involved in those functions are found in the CM.

Monoderm Bacteria Cell Envelope (Firmicutes) The main difference between monoderm and diderm bacteria cell envelopes is the absence of OM in the former (Fig. 2). This type of envelope was previously defined as specific of Gram-positive bacteria. This simple view has been modified and the recent discovery of a real OM in the Gram-positive Firmicutes Mycobacteria is a proof of concept attesting that Gram positive and even Firmicutes are not all monoderm bacteria. This section is a description of the main features of the structure

Bacterial Cell Envelopes: Composition, Architecture, and Origin

9

Fig. 2 Typical monoderm bacteria envelope. The two layers of an envelope of typical monoderm (Firmicutes) are shown schematically here. Phospholipids are depicted in blue. Inner membrane proteins are in orange. Carbohydrate basic units from the peptigoglycan are in blue (MurNAc) and red (GlyNAc). Wall teichoic acids (WTA) are in pink with the decorations by carbohydrates in pale blue and by alanine in black. Lipoteichoic acids are in green and their modifications are symbolized as for WTA. Covalently attached or non-covalently attached proteins are depicted in green. PG peptidoglycan, CM cytoplasmic membrane, CW cell wall

of the model Gram-positive Firmicutes envelope represented by B. subtilis or S. aureus and designed as “monoderm bacteria” (Sutcliffe 2010). In Gram-negative diderm-LPS bacteria, the presence of the OM barrier stabilizes the CM, and the PG is relatively thin. In monoderm bacteria, the OM is absent and to protect the cell from the turgor pressure, the layers of PG are several times thicker. The thickness of monoderm envelopes varies from one species to another. For example, it is 40 nm in Staphylococcus aureus and 60 nm in Bacillus subtilis (Vollmer and Seligman 2010). The thickness of the membrane is, as in diderm, relatively constant and close to 6 nm. It is especially the thickness of the PG that is variable: 20 nm in S. aureus to 34 nm in B. subtillis, for example (Vollmer and Seligman 2010 9). In Firmicutes, the cell wall is reinforced by the presence of long anionic polymers, the teichoic acids (TA) (Swoboda et al. 2010). TA are formed by a disaccharide linkage unit and an anionic chain of polyglycerolphosphate or polyribitolphosphate repeats decorated with saccharides and positively charged D-alanyl esters. Wall teichoic acids (WTA) are covalently attached to PG (Brown et al. 2013), while lipoteichoic acids (LTA) are linked to the head groups of CM lipids. TA represent about 60 % of the cell envelope mass and are mainly responsible, together with PG, for the envelope structure and function. Like LPS in Proteobacteria, WTA in Firmicutes are critical determinants of the envelope surface net charge and hydrophobicity.

10

D. Zerbib

The extracellular proteins of monoderm bacteria, because of the absence of OM, are either covalently attached to the PG, partially inserted in the CM, linked to CM lipids anchors, or attached to teichoic acids (Dramsi et al. 2008). Several extracellular proteins of monoderm are orthologues of diderm periplasmic proteins.

Monoderm Peptidoglycan The PG of monoderm Firmicutes is made of glycan strands composed of disaccharide repeats linked together with pentapeptide units. The classical monoderm PG chemical composition is identical to the one of diderm bacteria. The linear glycan strands of PG are made up of alternating N-acetyl glucosamine (GlcNAc) and Nacetyl muramic acid (MurNAc) linked in β-1 ! 4 and are cross-linked together with short specific peptides. The major difference with diderm PG is in the PG thickness; the Gram-positive PG can be 30–100 nm thick whereas the Gram negative is only a few nm thick. Among the Gram-positive bacteria, the differences between PG structures are mainly based on the type of cross-linking of the glycan strands by peptides (Vollmer 2008). In contrary to B. subtilis, S. aureus and many other Grampositive Staphylococcaceae contain branched stem peptides in their PG that are involved in β-lactam resistance (Chambers 2003). In S. aureus, a pentaglycine branch is present on the third amino acid of the stem peptide and involved in glycan strands cross-link. The presence of this specific type of branched peptide is responsible for the S. aureus resistance to β-lactam antibiotics because the target protein, the transpeptidase that performs this specific peptide cross-link (PBP2A) is structurally different in these bacteria and therefore insensitive to inhibition by b-lactams. Branched peptides are also attachment site of specific PG covalently associated proteins.

Surface Proteins The adherence of pathogenic bacteria like S. aureus is an important pathogenicity factor. TA are involved in bacterial adherence together with adhesins, proteins that are able to bind components of the extra cellular matrix like fibronectin, fibrinogen, and elastin. Some adhesins are only interacting with PG or TA, but most of them are covalently attached to the stem peptide of PG (Dramsi et al. 2008). Those proteins are secreted by sortases through a pentapeptide motif present in their C-terminal end. Sortases catalyze the modification of PG precursors by surface proteins prior to their incorporation in mature PG (Dramsi et al. 2008). In S. aureus, sortase A (StrA) possesses more than 20 protein substrates involved in adhesion, which escape from the immune system, phage binding and internalization. Sortase B (StrB) allows the surface display of an iron acquisition protein. The sortases represent the architects of the bacterial surface. Other surface proteins are anchored in the CM through helices inserted in the lipid leaflet. Finally, other surface proteins involved in cell wall modeling are non-covalently linked to either PG or TA.

Bacterial Cell Envelopes: Composition, Architecture, and Origin

11

Teichoic Acids These long anionic polymers are found in many Gram-positive bacteria including the models S. aureus and B. subtilis. They are responsible for the negative surface net charge of those bacteria. The WTA are coupled to PG, and the LTA are anchored in the CM (Fig. 2). WTAs are linked on the C6 of the MurNAc residues of PG glycan strands. The most common WTA are formed by disaccharides modified by more than 60 repeat units of either polyribitolphosphate (polyRboP) or polyglycerolphosphate (polyGroP). WTAs are found “perpendicular” to the PG layers. Depending on species or strains, the RboP or GroP units can be modified with saccharides and positively charged D-alanyl esters. Those variations affect the functions of the WTA. LTA are similar to WTA, they are formed by polyGroP modified by either a D-alanine or a sugar moiety. They are anchored in the CM through embedded glycolipids. Since they contain fewer repeat units than WTA, they do not span beyond the PG. The negative surface net charge of the Grampositive envelope is essential, and when TA are absent from a given species, they are replaced by other polyanionic polymers containing either carboxylate or sulfate groups to provide the negative charges. LTA or WTA are not essential separately, but it is not possible to delete both pathways simultaneously because their primary function is to be anionic and to play a role in cation homeostasis. Cations binding on TA induce the formation of a cationic network that increases the rigidity of the envelope. Moreover, the density of TA negative charges influences the sensitivity of bacteria to cationic antibiotics and to host lytic enzymes.

The Cytoplasmic Membrane The monoderm CM is a symmetrical phospholipid bilayer. The proportions of zwitterionic phospholipid phosphatidylethanolamine (30 % PE) and anionic phospholipid phosphatidylglycerol (70 %, PG) are different than in diderm giving the monoderm CM a higher negative surface charge (van der Does et al. 2000). This bilayer contains undecaprenyl-phosphate necessary for the envelope biogenesis together with α-helical anchored lipoproteins and the saccharide moiety of LTA.

The Corynebacterineae Envelope The suborder of Corynebacterineae, from the phylum Actinobacteria, includes the family of mycobacteriaceae (Mycobacterium), corynebacteriaceae (Corynebacterium), and nocardiaceae (Nocardia). These bacteria were defined as monoderm high G + C Gram-positive bacteria, but they display properties of both Gram-positive and Gram-negative bacteria (Fu and Fu-Liu 2002). The cell envelope of M. tuberculosis has indeed characteristics of both Gram-positive and Gram-negative bacteria (Jankute et al. 2015).

12

D. Zerbib

A common CM is located under a typical thick monoderm PG, but the PG is covalently attached to arabinogalactan (AG) itself linked to very long chain specific fatty acids: the mycolic acids (MA, (Marrakchi et al. 2014)). This giant macromolecule called the PG-AG-MA complex fills the very large periplasm of Mycobacteria. The mycolic acids are β-hydroxylated, α-branched fatty acids and can contain up to 90 carbons. They are responsible for the waxy appearance of mycobacteria; they are involved in the impermeability of the envelope and play an important role in the pathogenicity of Mycobacteria. MA biosynthesis relies on multiple protein complexes (Veyron-Churlet et al. 2004; Carel et al. 2014) that represent targets for efficient antibiotics (Nataraj et al. 2015). Finally, the mycobacterial envelop contains a variety of non-covalently bound cell envelope lipids and glycoconjugates as recently reviewed (Daffé et al. 2014). It has been demonstrated by cryo-electron microscopy of vitreous sections (CEMOVIS) that Corynebacterineae have a OM. The asymmetric OM is supposed to be formed by an inner leaflet of mycolic acids and an outer leaflet of MA comprising inserted extractible lipids (Daffé et al. 2014). The precise structure of the OM remains unclear (Zuber et al. 2008). Like Gram-negative bacteria, mycobacteria have porins in the OM, but those are structurally different from the one of E. coli. Corynebacterineae can be designed as Gram-positive diderm-non-LPS bacteria.

“A Phylum Level Perspective on Bacterial Cell Envelope Architecture” The present understanding of phylogenetic tree allowed classifying prokaryotes in 30 phyla in the domain Bacteria and 5 phyla in the domain Archaea (http://www. bacterio.net/-classifphyla.html) (Sutcliffe 2010). The increasing amount of data concerning the genome sequences together with the precise knowledge of the genetics of the cell envelope biogenesis allowed seeing the bacteria in the Tree of Life in a different manner. It is now clear that the Gram-based classification was an oversimplification and that bacterial envelopes are now to be defined with the number of membranes that they contain (Sutcliffe 2010): bacteria are divided into either monoderm or diderm, and the presence or absence of LPS in diderm is also a fundamental criterion. Among Bacteria, there is a clear dominance of diderm LPS (Glycobacteria) that are not restricted to Proteobacteria but represented in 17 phyla. Typical monoderms exist in Firmicutes and Actinobacteria phyla but also in Tenericutes and Choroflexi. Some other exceptions exist and are briefly listed below.

Dominance of the Diderm-LPS Cell Envelope Architecture Based on 16S rRNA gene sequencing, it has been shown that in addition to the three most known phyla (Firmicutes, Proteobacteria, Actinobacteria), 21 additional phyla could be defined (Cavalier-Smith 2006, 2010; Pace 2009). The enormous amount of data concerning Proteobacteria explains why Gram-negative envelope was

Bacterial Cell Envelopes: Composition, Architecture, and Origin

13

synonymous of diderm-LPS envelope. It is true that many phyla have been shown to contain LPS, as shown either by biochemical characterization or by identifying LPS biosynthesis genes in the genome. This architecture is typical of 17 out of 24 phyla and is thus not restricted to Proteobacteria. Some LPS-containing bacteria do not possess any known component of the LPS biosynthesis pathway (Chlamydiae, Bacteroidetes, Chlorobi, Cyanobacteria) (Sutcliffe 2010).

Phyla with Typical Monoderm Cell Envelopes Only Firmicutes and Actinobacteria are mainly composed of monoderm bacteria with some very important exceptions in the suborder of Corynebacterineae. In addition to the thick cell wall, as mentioned above, the presence of polyanionic polymers is typical of monoderm. The distribution of teichoic acids or teichuronic acids in those two phyla suggests that they are determinants of monoderm polyanionic envelopes. The CM of representative of these phyla also contains either lipoteichoic acids or lipoglycans. The Tree of Life has been modified and other monoderm phyla have been defined. For example, in the 2nd edition of Bergey’s Manual of Systematic Bacteriology, the class Mollicutes (Gram-negative bacteria devoid of cell wall) has been excluded from the phylum Firmicutes and classified in the phylum Tenericutes. In addition, monoderm representatives of a previously proposed phylum (Ktedonobacteria) are now a class of the phylum Chloroflexi that might be a typical monoderm phylum (Sutcliffe 2010).

Diderm Exceptions in the Monoderm Phyla Within the Firmicutes phylum, there are several diderm-LPS bacteria, from the class of Clostridia, and some of them even possess a genuine LPS biosynthesis pathway. As mentioned above, within Actinobacteria, the suborder of Corynebacterineae has the same PG-AG-MA complex even if the mycolic acids are very different from one species to another. Thus, contrary to textbook definitions, Actinobacteria are not all monoderms.

Diderm Alternative to Diderm LPS Exceptions in Phyla Spirochaetes, Proteobacteria, and Actininobacteria Several genera of Spirochaetes phylum have no LPS but have a diderm cell envelope with a specific lipid replacing lipid A in the outer leaflet of the OM. In the Proteobacteria phylum, example exists (Shingomonas) with no LPS and an OM made of glycosphingolipids. In the Actinobacteria phylum, the Corynebacterineae

14

D. Zerbib

possess an asymmetric OM like the diderm-LPS Gram-negative bacteria and can be considered as diderm-MA bacteria (Sutcliffe 2010).

The Phylum Deinococcus-Thermus Deinococci have a complex envelope with a lipid rich layer without LPS and covered by a protein S-layer. LPS seems to be replaced by an unusual lipoglycan (Rothfuss et al. 2006). The S-Layer is formed by a lipoprotein (Hpi) anchored directly in the outer lipid layer. Layered envelope with apparent OM and an S-Layer also exists in Thermus species. Proteins typical of OM like β-barrel proteins and Bam family proteins (β-barrel assembly machinery) have been identified in those bacteria (Brosig et al. 2009).

The Phylum Thermotogae Bacteria from the phylum Thermotogae have an outer envelope (toga) that is essentially composed of proteins rather than lipids. Thermotogae members are Gram negative, contain porins, and do not have any LPS biosynthetic genes (Sutcliffe 2010). The take home message is that there are only few typically monoderm phyla and that the very large majority of known diderm bacteria are diderm-LPS bacteria with some exceptions when LPSs are replaced by other lipid molecules in the OM. Actinobacteria and Firmicutes, which represent 40 % of the recognized bacterial taxa, are not pure monoderms. It seems that only Chloroflexi might be a typical monoderm phylum.

Bacterial S-Layer S-layers (Surface layer) are self-assembled paracrystalline protein lattices found in many bacteria and almost all archaea. As an important component of the bacterial cell envelope, S-layers can fulfill various biological functions and are usually the most abundantly expressed proteins in a cell. In Archae, S-Layer is the only cell-wall component. S-Layers are found in both Gram-positive and Gram-negative bacteria. In Gram positive, the S-Layer is non-covalently attached to PG or other cell-wall components. In Gram negative, the S-Layer is commonly attached to the LPS of the OM. S-layers are mostly composed of one protein or glycoprotein species. In glycoproteins, the degree of glycosylation and glycan composition greatly varies. Single strain can express different S-layers depending on culture conditions. The S-layer proteins are poorly similar; they are usually acidic and hydrophobic. The S-Layer proteins form regular crystalline lattices, which may have an oblique (p1, p2), square (p4), or hexagonal (p3, p6) symmetry. The S-layer proteins possess a heavier domain

Bacterial Cell Envelopes: Composition, Architecture, and Origin

15

that forms the core of the morphological unit cell and a lighter domain that provides connectivity between the units. The central core-forming region is usually oriented toward the cell envelope, giving rise to an overall corrugated inner surface. By contrast, the outer surface appears smooth despite highly variable and speciesspecific ultrastructures. Between 30 % and 70 % of the unit cell volume is occupied by the protein, which leads to the formation of identical and well-defined pores with a diameter of 2–8 nm (Pavkov-Keller et al. 2011).

The Archaeal Cell Envelope The composition of the archaeal cell envelope is very different from the bacterial cell envelope (Albers and Meyer 2011). With the only exception of Ignicoccus, which exhibits an OM and a very thick periplasm, Archaea possess only a single membrane. The CM is enclosed by an S-layer. Archaeal lipids consist of repeating isoprenyl groups linked to a glycerol backbone through an ether linkage. These lipids form diether bilayer membranes. Hyperthermo-acidophiles contain tetraether lipids forming a monolayer membrane. These membranes are very impermeable to protons and enable these Archaea to survive in highly acidic conditions. Another specificity of Archaea is that the extracellular proteins are glycosylated with N- and O-glycosylation. Finally, Archaea do not produce any murein, with the exception of some methanogenic species that produce pseudo-murein.

The Origin of Diderm Bacteria; Selective Pressure or Endosymbiosis? As written in an article’s summary of Norman Pace (Pace et al. 2012), “. . .In 1977, Carl Woese and George Fox published a brief paper in the Proceedings of National Academy of Sciences (USA) that established, for the first time, that the overall phylogenetic structure of the living world is tripartite. . .”. The Tree of Life has been presented as having three main branches: eubacteria, Archaea, and Eucarya (Woese et al. 1990). Since this fundamental discovery based on the analysis of the rDNA, many scientists have collected and interpreted phylogenetic data such as the comparison of the sequences of ribosomal proteins, bacterial physiology, bacterial morphology, conservation of key proteins, and presence of specific indels. in order to grow more and more the universal Tree of Life. While retaining the same backbone than that of the original tree (Woese et al. 1990), a consensual Tree of Life (Fig. 3a) has recently been presented by P. Forterre (2015). The most studied phyla are represented on this tree with an attempt to consensual rooting and grouping in super-phyla (Fig. 3b). The phyla comprising members with LPS-containing envelopes are distributed in all the different super-phyla. On the other hand, they are less represented in the super-phylum Cyanobacteria. However, it is striking to see that the phyla containing the pure monoderms and the atypical diderms are more grouped in the super-phylum Cyanobacteria. In the quest of the

16

D. Zerbib

Fig. 3 Schematic and simplified Universal Tree of Life redrawn with simplifications and modifications from P. Forterre (2015). (a) The backbone of the Universal Tree of Life. The three main clades organized in super-phyla are indicated: Bacteria are in green, Archaea in pale blue, and Eucarya in pale red. Potential rooting on common ancestors (Forterre 2015) are indicated: LUCA Last Unknown Common Ancestor, LBCA -Bacterial-, LACA -Archaeal-, LARCA -Arkaryal-, LECA -Eukaryotic-. (b) Schematic simplified bacterial branch of the Universal Tree of Life from P. Forterre (2015). The names of the most studied phyla are indicated together with the names of super-phyla (in bold). The names of phyla containing a majority of typical diderms are in black, those of phyla containing a majority of typical monoderms are in red. Phyla containing LPS-bacteria are underlined. Phyla in green corresponds to atypical alternative to diderm-LPS bacteria (see text)

last common ancestor (LCA) in bacteria, the division of the Bacteria empire into diderm or monoderm should probably be seen more as a phylogenetics criterion. The origin of diderm bacteria is still debated. Two main theories are opposed in the literature: either their endosymbiotic origin or their selection under antibiotics pressure. The first is supported by James A. Lake (2009) and the second by Radhey

Bacterial Cell Envelopes: Composition, Architecture, and Origin

17

Fig. 4 A cartoon of the development of outer bacterial envelope redrawn with simplifications and modifications from R.S. Gupta (2011). This cartoon of the development of the outer envelope of different bacterial phyla was based on a study of the distribution of inserts in the Hsp60 protein (occurring “between” the simple and typical diderms) and inserts in the Hsp70 protein (occurring “between” typical diderm and Chloroflexi monoderm) (Gupta 2011). IM Inner membrane in red, LPS lipopolysaccharide in either green or grey, PG peptidoglycan in pale green, OM outer membrane in green, MM mycomembrane containing mycolic acids (MA) in blue, T outer toga in grey dashed line. The noncharacterized outer layer of Chloroflexi is in grey

S. Gupta (1998, 2011). James Lake has proposed that diderm came from a former endosymbiosis between an Actinobacteria and a Clostridia (Lake 2009). Radhey Gupta has presented very convincing evidence, based on the study of insertions and deletions in the genes of Hsp60 and Hsp70, showing that probably, the endosymbiosis does not explain the origin of the diderms. Diderms might have evolved from an ancestral monoderm under the selection pressure of antibiotics. It is very interesting because this theory allows understanding how the bacterial envelopes could have evolved (Fig. 4). The typical monoderms may have acquired an outer layer (still not characterized in Chloroflexi) and then an outer cell envelope devoid of LPS (Deinococcus-Thermus). Finally, the presence of LPS in the OM would have given a large part of the typical phyla of diderm-LPS bacteria (Fig. 4). The same ancestral typical monoderm could have acquired either a toga (Thermotogae), either an OM containing mycolic acids (Corynebacterineae) or an atypical external envelope bearing LPS (Fusobacteria, Negativicutes, Synergistetes, Elusimicrobia) (Fig. 4). Trying to develop and analyze the Gupta’s theories in deep phylogeny is far from the purpose of this review. However, it might be interesting, in the aim of understanding more clearly the complexity and variability of bacterial envelopes, to try to

18

D. Zerbib

read the Universal Tree of life as it is interpreted today by Patrick Forterre (2015) (Fig. 3b), in light of the theory on the origin of the diderms of R. Gupta (1998, 2011) (Fig. 4), after an in-depth analysis of the data which have helped to build this chapter and with the knowledge and the vision of the bacterial phylogeny like presented by Iain Sutcliffe (Sutcliffe 2010).

Conclusions The bacterial cell envelopes are complex, dynamic, and play fundamental roles in the adaptation and the protection of bacteria. The peptidoglycan is the most preserved component of bacterial envelopes. It is essential in the maintenance of the rigidity and the shape of bacteria. It is also essential to protect the cell from the very strong osmotic pressure of the cytoplasm. In addition to the PG, the asymmetry of the OM of diderm bacteria, the teichoic acids or their equivalents anionic in monoderm bacteria, and the complex PG-AG-MA of Corynebacterineae ensure a perfect integrity of the cell envelope. The architecture of the cell envelopes of many bacteria is well known, and the biosynthesis of their components is also perfectly decrypted. The challenge is probably to assess the impact of the electrical fields at the molecular level, on each type of envelope, and taking into account all of the biochemical, structural, and architectural characteristics in each bacterium which must be pulsed for a given application on specific species of bacteria.

References Albers SV, Meyer BH (2011) The archaeal cell envelope. Nat Rev Microbiol 9:414–426 Beveridge T (2001) Use of the gram stain in microbiology. Biotech Histochem 76:111–118 Braun V (1975) Covalent lipoprotein from the outer membrane of Escherichia coli. Biochim Biophys Acta 415:335–377 Brosig A, Nesper J, Boos W, Welte W, Diederichs K (2009) Crystal structure of a major outer membrane protein from Thermus thermophilus HB27. J Mol Biol 385:1445–1455 Brown S, Santa Maria JP Jr, Walker S (2013) Wall teichoic acids of gram-positive bacteria. Annu Rev Microbiol 67:313–336 Carel C, Nukdee K, Cantaloube S, Bonne M, Diagne CT, Laval F, Daffe M, Zerbib D (2014) Mycobacterium tuberculosis proteins involved in mycolic acid synthesis and transport localize dynamically to the old growing pole and septum. PLoS One 9, e97148 Cavalier-Smith T (2006) Rooting the tree of life by transition analyses. Biol Direct 1:19 Cavalier-Smith T (2010) Deep phylogeny, ancestral groups and the four ages of life. Philos Trans R Soc Lond B Biol Sci 365:111–132 Chambers HF (2003) Solving staphylococcal resistance to beta-lactams. Trends Microbiol 11:145–148 Chopra I, Storey C, Falla TJ, Pearce JH (1998) Antibiotics, peptidoglycan synthesis and genomics: the chlamydial anomaly revisited. Microbiology 144(Pt 10):2673–2678 Daffé M, Crick DC, Jackson M (2014) Genetics of capsular polysaccharides and cell envelope (glyco)lipids. Microbiol Spectr 2 Dramsi S, Magnet S, Davison S, Arthur M (2008) Covalent attachment of proteins to peptidoglycan. FEMS Microbiol Rev 32:307–320

Bacterial Cell Envelopes: Composition, Architecture, and Origin

19

Dufresne K, Paradis-Bleau C (2015) Biology and assembly of the bacterial envelope. Adv Exp Med Biol 883:41–76 Forterre P (2015) The universal tree of life: an update. Front Microbiol 6:717 Fu LM, Fu-Liu CS (2002) Is Mycobacterium tuberculosis a closer relative to Gram-positive or Gram–negative bacterial pathogens? Tuberculosis 82:85–90 Gupta RS (1998) What are archaebacteria: life’s third domain or monoderm prokaryotes related to Gram-positive bacteria? A new proposal for the classification of prokaryotic organisms. Mol Microbiol 29:695–707 Gupta RS (2011) Origin of diderm (Gram-negative) bacteria: antibiotic selection pressure rather than endosymbiosis likely led to the evolution of bacterial cells with two membranes. Antonie Van Leeuwenhoek 100:171–182 Holtje JV (1998) Growth of the stress-bearing and shape-maintaining murein sacculus of Escherichia coli. Microbiol Mol Biol Rev 62:181–203 Jankute M, Cox JA, Harrison J, Besra GS (2015) Assembly of the mycobacterial cell wall. Annu Rev Microbiol 69:405–423 Jordan S, Hutchings MI, Mascher T (2008) Cell envelope stress response in Gram-positive bacteria. FEMS Microbiol Rev 32:107–146 Klebba PE (2005) The porinologist. J Bacteriol 187:8232–8236 Lake JA (2009) Evidence for an early prokaryotic endosymbiosis. Nature 460:967–971 Lugtenberg EJ, Peters R (1976) Distribution of lipids in cytoplasmic and outer membranes of Escherichia coli K12. Biochim Biophys Acta 441:38–47 Lyu ZX, Zhao XS (2015) Periplasmic quality control in biogenesis of outer membrane proteins. Biochem Soc Trans 43:133–138 Marrakchi H, Laneelle MA, Daffe M (2014) Mycolic acids: structures, biosynthesis, and beyond. Chem Biol 21:67–85 Meroueh SO, Bencze KZ, Hesek D, Lee M, Fisher JF, Stemmler TL, Mobashery S (2006) Threedimensional structure of the bacterial cell wall peptidoglycan. Proc Natl Acad Sci U S A 103:4404–4409 Mogensen JE, Otzen DE (2005) Interactions between folding factors and bacterial outer membrane proteins. Mol Microbiol 57:326–346 Mullineaux CW, Nenninger A, Ray N, Robinson C (2006) Diffusion of green fluorescent protein in three cell environments in Escherichia coli. J Bacteriol 188:3442–3448 Nakayama H, Kurokawa K, Lee BL (2012) Lipoproteins in bacteria: structures and biosynthetic pathways. FEBS J 279:4247–4268 Nataraj V, Varela C, Javid A, Singh A, Besra GS, Bhatt A (2015) Mycolic acids: deciphering and targeting the Achilles’ heel of the tubercle bacillus. Mol Microbiol 98:7–16 Nikaido H (2003) Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev 67:593–656 Pace NR (2009) Mapping the tree of life: progress and prospects. Microbiol Mol Biol Rev 73:565–576 Pace NR, Sapp J, Goldenfeld N (2012) Phylogeny and beyond: scientific, historical, and conceptual significance of the first tree of life. Proc Natl Acad Sci U S A 109:1011–1018 Pavkov-Keller T, Howorka S, Keller W (2011) The structure of bacterial S-layer proteins. Prog Mol Biol Transl Sci 103:73–130 Pillet F, Formosa-Dague C, Baaziz H, Dague E, Rols MP (2016) Cell wall as a target for bacteria inactivation by pulsed electric fields. Sci Rep 6:19778 Pink D, Moeller J, Quinn B, Jericho M, Beveridge T (2000) On the architecture of the gramnegative bacterial murein sacculus. J Bacteriol 182:5925–5930 Raetz CR, Whitfield C (2002) Lipopolysaccharide endotoxins. Annu Rev Biochem 71:635–700 Rothfuss H, Lara JC, Schmid AK, Lidstrom ME (2006) Involvement of the S-layer proteins Hpi and SlpA in the maintenance of cell envelope integrity in Deinococcus radiodurans R1. Microbiology 152:2779–2787

20

D. Zerbib

Silhavy TJ, Kahne D, Walker S (2010) The bacterial cell envelope. Cold Spring Harb Perspect Biol 2:a000414 Sutcliffe IC (2010) A phylum level perspective on bacterial cell envelope architecture. Trends Microbiol 18:464–470 Swoboda JG, Campbell J, Meredith TC, Walker S (2010) Wall teichoic acid function, biosynthesis, and inhibition. Chembiochem 11:35–45 Tamura A, Ohashi N, Urakami H, Miyamura S (1995) Classification of Rickettsia tsutsugamushi in a new genus, Orientia gen. nov., as Orientia tsutsugamushi comb. nov. Int J Syst Bacteriol 45:589–591 Taylor BL (1983) Role of proton motive force in sensory transduction in bacteria. Annu Rev Microbiol 37:551–573 Turner RD, Vollmer W, Foster SJ (2014) Different walls for rods and balls: the diversity of peptidoglycan. Mol Microbiol 91:862–874 van der Does C, Swaving J, van Klompenburg W, Driessen AJ (2000) Non-bilayer lipids stimulate the activity of the reconstituted bacterial protein translocase. J Biol Chem 275:2472–2478 Vandeputte-Rutten L, Kramer RA, Kroon J, Dekker N, Egmond MR, Gros P (2001) Crystal structure of the outer membrane protease OmpT from Escherichia coli suggests a novel catalytic site. EMBO J 20:5033–5039 Veyron-Churlet R, Guerrini O, Mourey L, Daffe M, Zerbib D (2004) Protein-protein interactions within the Fatty Acid Synthase-II system of Mycobacterium tuberculosis are essential for mycobacterial viability. Mol Microbiol 54:1161–1172 Vollmer W (2008) Structural variation in the glycan strands of bacterial peptidoglycan. FEMS Microbiol Rev 32:287–306 Vollmer W, Seligman SJ (2010) Architecture of peptidoglycan: more data and more models. Trends Microbiol 18:59–66 Weidel W, Pelzer H (1964) Bagshaped macromolecules – a new outlook on bacterial cell walls. Adv Enzymol Relat Areas Mol Biol 26:193–232 Woese CR, Kandler O, Wheelis ML (1990) Towards a natural system of organisms: proposal for the domains Archaea, Bacteria, and Eucarya. Proc Natl Acad Sci U S A 87:4576–4579 Zuber B, Chami M, Houssin C, Dubochet J, Griffiths G, Daffe M (2008) Direct visualization of the outer membrane of mycobacteria and corynebacteria in their native state. J Bacteriol 190:5672–5680

3D Culture Models to Assess Tissue Responses to Electroporation Anna A. Bulysheva and Richard Heller

Abstract

Cell and tissue responses to external stimuli are difficult to study in vivo. Traditional monolayer culture conditions allow for observation of cellular response to stimuli in vitro with a great degree of control and manipulation of experimental conditions; however, many studies have shown that cells exhibit different gene expression patterns, drug resistance, and mechanical stress responses in two-dimensional environments, than when cultured in three-dimensional (3D) culture environments or in vivo. Cell-cell and cell-matrix interactions determine many aspects of cellular behavior, including proliferation, metabolism, differentiation potential, and viability. Therefore, many biomimetic strategies exist for 3D cell culture for various applications. This chapter describes 3D culture methods for assessing tissue response to exogenous stimuli, specifically electroporation. These methods include spheroid culture, cell culture on electrospun scaffolds, and cell culture on decellularized human dermal matrices. Spheroid culture is generally recognized as a model system for tumor development and has been used extensively to study electroporation effects. Other 3D culture techniques include using electrospun scaffolds for various tissues such as oral mucosa and head and heck squamous carcinoma, and can be readily adapted to studying electroporation effects. Decellularized human dermis has been recently demonstrated as an excellent substrate for recapitulating human skin and used for electroporation applications. A.A. Bulysheva (*) Frank Reidy Research Center for Bioelectrics, Old Dominion University, Norfolk, VA, USA e-mail: [email protected] R. Heller Frank Reidy Research Center for Bioelectrics, Old Dominion University, Norfolk, VA, USA School of Medical Diagnostics and Translational Sciences, Old Dominion University, Norfolk, VA, USA e-mail: [email protected] # Springer International Publishing AG 2017 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_29-1

1

2

A.A. Bulysheva and R. Heller

Keywords

Electroporation • Electrotransfer • Gene therapy • 3D cell culture • Biomimicry • Spheroid • Extracellular matrix • Electrospinning • Cryogenic electrospinning

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Spheroids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Spinner Flask Suspension Prostate Cancer Spheroid Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Microgravity Melanoma Spheroid Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Hanging Drop Spheroid 3D Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Electrospun Scaffolds for 3D Culture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Electrospinning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Cryogenic Electrospinning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Cryogenic Electrospun Scaffolds for Oral Mucosa and Skin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Recellularized Human Dermis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Introduction Traditional in vitro tissue and cell culture has a long history starting with Harrison’s report of tissue culture of frog nerves in 1907 with the hanging drop method for maintaining neurons for up to 4 weeks (Freshney 2005). It was not until introduction of the proteolytic enzyme trypsin in 1916 that cell culture free of the extracellular matrix was possible resulting in culture and subculture of adherent cells (Freshney 2005). The first human cell line (HeLa) was established in 1952 from a cervical carcinoma (Freshney 2005). Adherent mammalian cell culture has provided a large body of knowledge about cell structure and function. It is a convenient experimental platform for investigating cellular response to their environment and to various perturbations. This two-dimensional (2D) environment, however, does differ from the native environment that various cell types experience in the body. Extracellular cues from the extracellular matrix, cell-cell interaction, and mechanical forces in three dimensions are absent in traditional cell culture of adherent cells to tissue culture plastic or other coated two-dimensional surfaces (Griffith and Swartz 2006). Many studies have demonstrated variability in gene expression profiles of the same cells grown in 2D vs. 3D cultures, including differentiation, proliferation, drug response, and survival. It is hypothesized that tissue specific in vitro 3D culture systems match cell response to perturbations more closely to in vivo responses than can be observed in 2D culture (Griffith and Swartz 2006). Therefore, many efforts have been dedicated to creating 3D culture environments for assessment of cells responding to perturbations such as electroporation. Various 3D culture systems have been developed largely for tissue engineering application to mimic skin, mucosa, blood vessels, liver, bone, and other tissues. The guiding principles remain similar for constructing a suitable model system. A scaffold is created for mimicking the extracellular matrix, and the appropriate cell

3D Culture Models to Assess Tissue Responses to Electroporation

3

type is selected and then seeded on the scaffold. The cells are allowed to replicate their microcellular environment for a set amount of days and the system is tested for specific markers of the desired tissue and cell viability. Once the model is validated, it can then be used for studying tissue responses. This chapter will discuss 3D culture systems and techniques available for studying tissue responses to electroporation, including spheroid culture for tumor modeling, electrospinning techniques for scaffold fabrication, and 3D culture systems of skin and oral mucosa and their potential utility.

Spheroids Spheroid culture has been ubiquitous throughout many disciplines. The hanging drop method, which remains the main method for creating spheroids, has been established initially for developmental biology and embryology studies (Freshney 2005). It was later adapted by cancer biology, stem cell biology, and tissue engineering disciplines (Freshney 2005). Multiple groups have used spheroid culture methods to evaluate effects of electroporation on cells particularly tumor cellular microenvironment, uptake of molecules of various sizes (including dyes, drugs, and DNA), effects on membrane integrity, and cellular viability (Canatella et al. 2004; Mellor et al. 2006; Wasungu et al. 2009; Marrero and Heller 2012; Gibot et al. 2013; Gibot and Rols 2013; Chopinet et al. 2012; Frandsen et al. 2015). Several distinct spheroid models have been developed for studying effects of applied electric fields on tumors in vitro.

Spinner Flask Suspension Prostate Cancer Spheroid Model Canatella et al. (2004) reported using spheroid culture of prostate cancer cells to study the effects of electroporation on molecular uptake in a multicellular environment (Canatella et al. 2004). DU145 cells were suspended in siliconized spinner culture flasks and allowed to form spheroids in suspension culture over a period of multiple days. Spheroids were filtered via nylon meshes and centrifuged to ensure uniform spheroid sizes were used for further experiments. All electroporation experiments on uniformly sized spheroids were performed in 4-mm gap cuvettes. The spheroids were dissociated into single cell suspensions for analysis of molecular uptake after application of electric pulses. Flow cytometry was used for quantitative analysis of fluorescence intensity. Propidium iodine (PI) was used to label and identify nonviable cells. Cells close to the periphery of the spheroids were identified by higher uptake of Hoechst 33342 or calcein blue-AM. Green fluorescence intensity was used to quantify calcein molecular uptake occurring due to electroporation. Molecular uptake by cells in spheroids was reduced compared to molecular uptake by individual cells in suspension, in a manner proportional to spheroid size, with larger spheroids taking up fewer molecules of calcein. Cell uptake of calcein was inversely proportional to the distance from the outer surface of the spheroid. Larger

4

A.A. Bulysheva and R. Heller

field strengths, increased pulse length, and pulse width could enhance molecular uptake by cells residing in the spheroid (Canatella et al. 2004). Modeling of the electric field and predicted molecular uptake profiles were also correlated to experimental results. This study concluded that cells densely packed in three dimensions respond heterogeneously to pulsed electric fields. Cell location, cell size, and extracellular space solute concentration all contribute to this heterogeneity (Canatella et al. 2004).

Microgravity Melanoma Spheroid Model Another example of spheroid culture was developed under simulated microgravity conditions for modeling melanoma tumors (Marrero et al. 2009). The culture system consists of a coculture of keratinocytes and B16F10 melanoma cells or SKMEL-5 melanoma cells that form a large viable spheroid of approximately 1 cm in diameter (Marrero and Heller 2012; Marrero et al. 2009). HaCaT cells were trypsonized and resuspended at a high cell concentration in a rotary microgravity chamber. The chamber was maintained under constant rotation, simulating microgravity conditions. The rate of rotation was adjusted to account for the size of the forming spheroids. After 2 days of culture, the cells adhere to each other forming large spheroids. These ~1 cm in diameter spheroids were then injected with melanoma cells to form a coculture model. This spheroid model was tested for cell viability with a TdT-FragEL DNA fragmentation detection assay. Presence of basement membrane extracellular matrix components was evaluated with immunofluorescence staining for fibronectin. Cell proliferation was assessed with staining for the proliferation marker Ki67. The spheroids injected with melanoma cells remain stable for up to 15 days. Cell viability was determined to be 80%, while cell proliferation was confirmed with positively stained nuclei for Ki67. Keratinocytes were able to produce their own extracellular matrix indicated by positive immunoreactivity with fibronectin antibody (Marrero et al. 2009). This in vitro model of melanoma is particularly useful for evaluating gene electrotransfer parameters for melanoma tumors in vitro due to its macroscopic size enabling DNA injection to the center of the spheroid similar to in vivo protocols (Marrero and Heller 2012). Other important features include presence of keratinocytes, which provide cell-cell interactions and extracellular matrix, which account for cell-matrix interactions, together providing a similar microenvironment normally experienced by the melanoma cells in vivo. It was demonstrated that electrotransfer electrodes that are typically used for in vivo electrotransfer can be easily tested using this in vitro model. Six plate, four plate, 6-needle, and 4-needle electrodes were used to deliver green fluorescent protein (GFP), or interleukin 15 (IL-15) encoding plasmid DNA. GFP and IL-15 expression enhancement was confirmed in electrotransfer groups compared to no pulse controls. Therefore, this spheroid in vitro model of melanoma growth and development can be used to establish treatment regimens for translational therapies of melanoma in vivo (Marrero and Heller 2012; Marrero et al. 2009).

3D Culture Models to Assess Tissue Responses to Electroporation

5

Hanging Drop Spheroid 3D Models The classic hanging drop method for spheroid formation has been explored with human carcinoma and mouse sarcoma cell lines due to their ability to grow in monolayer cultures and form spheroids (Wasungu et al. 2009). Hanging drops of 20 μL containing 500 cells were allowed to aggregate in a handing drop for 72 h. Then the cell aggregates were transferred to wells coated with agar and filled with culture media. The spheroids were allowed to grow for up to 10 days when they reached ~500 μm in diameter. Electric pulse induced permeabilization and molecule delivery efficiency experiments were then conducted with delivery of PI and plasmid DNA encoding the reporter gene GFP. Spheroids were resuspended in pulsing buffer. Both PI and plasmid DNA were dissolved in the pulsing buffer for their respective experiments and were introduced to the spheroids from the outside of the spheroid, in contrast to Marrero et al. (2009), where the spheroids received plasmid DNA injections into the center of the spheroid. Parallel plate electrodes were used to deliver electric pulses establishing a relatively uniform electric field. This study observed that while cell permeability appeared homogenous, gene expression was not homogenously distributed throughout the spheroid. This observation is consistent with observations made by Canatella et al. (2004), with cells located on the outer edges of the spheroids expressing GPF, while inner cells remaining largely unaffected. After a certain threshold, application of higher electric field strengths resulted in lower gene expression, attributed to increased cell death. Cell-cell interactions were attributed to the differences between transfecting single cells compared to cells aggregated in spheroid culture. While the importance of the extracellular matrix role was suggested, no indication of its presence in the spheroid was experimentally presented (Wasungu et al. 2009). This spheroid model offers a well-controlled environment for studying mechanisms of electrotransfer mediated molecule delivery (Wasungu et al. 2009). The hanging drop method for spheroid formation was subsequently used to evaluate gene transfer under common electrotransfer conditions (Chopinet et al. 2012). Human colorectal carcinoma cells were used to form spheroids in the same manner as the prior study (Wasungu et al. 2009). Cells were suspended in a drop for 72 h and transferred to agar-coated wells with media, to allow maturation and growth of large spheroids. The effects of electric pulses on permeabilization and gene transfer were compared between single cells in suspension and cell aggregated in 3-day old spheroids. Spheroids were exposed to PI or plasmid DNA encoding GFP dissolved in pulsing buffer on the outside of the spheroids. Different pulsing conditions were empirically tested to delineate pulse effects on cell permeability, viability, and gene expression. It was shown that detectable DNA interaction with cells occurred only on the external surface of the spheroids. Less than 1% of the cells were found to express the reporter gene indicating low transfection efficiency possibly due to high cell mortality. It was also noted that only the external surface cells interacted with plasmid DNA and therefore internally located cells within the spheroid were shielded from efficient DNA uptake. In conclusion, it was reported

6

A.A. Bulysheva and R. Heller

that this spheroid model can be adapted to various tumor cell lines in order to study potential electrotransfer treatments without the use of animals (Chopinet et al. 2012). The hanging drop in vitro tumor spheroid model was further used to evaluate the relationship between tumor size and sensitivity to applied pulsed electric fields (Gibot and Rols 2013). Two distinct spheroid generation methods were tested, the hanging drop method as described above, and plating the same cell type (human colorectal carcinoma cells) in a nonadherent culture well. Cells were plated in ultralow attachment wells of a 96-well plate then centrifuged at 300 g for 5 min. After centrifugation induced cell aggregation, the cells were cultured for 5 days forming spheroids. Different starting cell numbers yielded different sized spheroids using both methods. Electric pulses were applied with parallel plate electrodes. Spheroid growth curves were meticulously established as a measure of cell viability post electropermeability treatments compared to untreated controls. Spheroid age and size as variables for sensitivity to electric pulses were controlled by varying the spheroid size with different spheroid formation methods, while maintaining the same age. It was determined that spheroid size was the main variable in spheroid sensitivity to electric pulses. Smaller spheroids were found to be more sensitive to equivalent electric pulses than larger spheroids. A higher degree of growth delay was demonstrated by normalized growth curves in smaller spheroids independent of spheroid age (Gibot and Rols 2013). This study serves as another example of how a spheroid model can be used to study effects of electric pulses on cells in a threedimensional environment. The nonadherent spheroid formation technique was used to study tumor sensitivity to calcium electroporation (Frandsen et al. 2015). This technique was expanded to other tumor cell lines, including a human bladder transitional cell carcinoma and a human breast adenocarcinoma. The nonadherent method for spheroid formation was also applied to fabricating human dermal fibroblast spheroids for modeling normal tissue as a control for tumor spheroids modeling tumors (Frandsen et al. 2015). Spheroids were subjected to calcium electroporation or bleomycin electroporation conditions normally used in vivo for tumor treatments. Spheroid size was measured daily after treatment and was normalized to the untreated controls. It was determined that spheroids derived from tumor cell lines were more sensitive to calcium and were more sensitive to bleomycin electroporation treatment than spheroids derived from normal fibroblasts in terms of molecular uptake and viability (Frandsen et al. 2015). It is unclear what tissue a human dermal fibroblast spheroid can represent. Dermal fibroblasts seldom form closely packed clusters in vivo especially in the dermis, where fibroblasts are separated by large collagen bundles with minimal cell-cell contact. However, this study does demonstrate another example of using tumor spheroids in vitro for modeling tumor tissue in vivo and evaluating electroporation effects on tumor cells. Spheroid models are excellent choices for studying electroporation effects on solid tumors of various types due to structural and functional similarities to solid tumors in vivo. Cell-cell interactions can closely resemble those present in tumors in vivo, it is also known that extracellular matrix components are produced in the microcellular environment within the spheroids. However, depending on the tissue

3D Culture Models to Assess Tissue Responses to Electroporation

7

to be studied, the spheroid model may or may not be the appropriate in vitro model. For studies interested in structures such as skin, blood vessels, skeletal muscle, heart muscle, or other sites, it may be beneficial to explore other in vitro models that more closely resemble the structure and function of the tissue of interest.

Electrospun Scaffolds for 3D Culture Electrospinning Electrospinning is a tissue engineering technique frequently used for fabricating scaffolds that are biomimetic, extracellular matrix analogues. These ECM mimicking scaffolds are often used for 3D cell culture in vitro and for developing implantable medical devices for regenerating damaged or dysfunctional tissues and organs. The process of electrospinning depends on electrostatic field application to a charged polymer solution. Figure 1 shows a diagram of the basic process of electrospinning. Typically, a syringe filled with a viscose polymer solution of choice is placed into a static electric field. A static electric field is formed by applying voltage across the tip of the metal needle and a metal collecting target. A typical applied voltage is on the order of 25 kV. Factors such as the distance between the polymer solution and collecting target, the shape of the collecting target, and the amplitude of the applied voltage determine the electrostatic force that acts on the polymer solution. Under optimized conditions, the electrostatic force overcomes the surface tension forces at the tip of the needle and a polymer jet extrudes from the needle towards the collecting target. Dry polymer fibers collect on the collecting target if the solvent is able to evaporate during the travel, therefore the distance to the collecting target has to also be long enough to allow enough time for solvent evaporation. For this reason, solvents that readily evaporate at ambient temperatures are preferred. Fibrous scaffolds that form may consist of nonwoven fibers with diameters in the range between 50 nm and 10 μm. Factors that contribute to scaffold fiber diameters are polymer molecular composition, solvent type, polymer concentration, electrospinning solution viscosity, air gap distance, magnitude of applied voltage, relative humidity during electrospinning, and temperature during electrospinning (Barnes et al. 2007). Natural polymers such as type I collagen, type II collagen, type IV collagen, laminins, elastin, fibrinogen, gelatin, chitosan, and silk fibroin have been successfully electrospun in addition to popular synthetic polymers such as polycaprolactone, polydioxanone, polylactic acid, and polyglycolic acids (Barnes et al. 2007). Silk fibroin, a natural fiber with structural similarities to collagen and excellent biocompatibility and mechanical properties, is frequently used for tissue engineering applications (Bulysheva et al. 2012). Silk fibroin readily electrospins at various concentrations and is soluble in many commonly used solvents. Silk fibroin serves as a favorable substrate for cell adhesion, particularly for keratinocytefibroblast cocultures (Bulysheva et al. 2012, 2013). Electrospun scaffolds can be optimized to have specific mechanical properties with fibers mimicking extracellular matrix fibers in fiber size and cell attachment

8

A.A. Bulysheva and R. Heller

Fig. 1 Traditional electrospinning process

sites. Fiber size is dependent on the concentration of the polymer, higher concentration resulting in larger fiber diameters. Cell attachment sites are provided by the polymer molecular composition. Natural polymers normally found in the mammalian extracellular matrix have abundant mammalian cell binding sites. Synthetic polymers generally lack those binding sites but can be synthetized to mimic cell adhesion sites. Alternatively, synthetic scaffolds may be coated with desired natural polymers to improve cell adhesions. However, traditional electrospinning often results in highly compacted mats of nanofibers with limited void space between fibers. Highly compacted microstructure inhibits cell infiltration into electrospun scaffolds, limiting in vitro and in vivo applications where cell infiltration and vascular ingrowth are desirable, such as 3D cell culture or skin grafts (Brauker et al. 1995; Leong et al. 2009, 2010). Increasing cell infiltration directly without increasing spacing between fibers can be accomplished during the electrospinning process by electrostatically spraying cells into nascent scaffolds (Stankus et al. 2007). This method requires complex protocols with sterile conditions during electrospinning and electrospraying and further considerations for electric field generated force on the fibers and cells. Cell viability is also a major concern. While more cells get integrated throughout the scaffolds, the void space between fibers remains limited. Cell migration and vascular ingrowth remain poor, limited by the densely compacted fibers surrounding the cells (Brauker et al. 1995). Increasing porosity can be accomplished in multiple ways. Inclusion of sacrificial fibers has been proposed as a way to increase porosity. The sacrificial polymer is electrospun in parallel with the desired polymer. Subsequently, sacrificial fibers are selectively dissolved with a solvent that does not dissolve the main fiber material (Jin et al. 2004). This approach does significantly increase porosity but not sufficiently for adequate cell infiltration (Jin et al. 2004) perhaps due to the relatively small size

3D Culture Models to Assess Tissue Responses to Electroporation

9

of the fibers that are washed away, thus leaving relatively small gaps between remaining fibers. A more effective way to increase porosity and cell infiltration has been the cotton-ball electrospinning method. It has been shown that this method leads to scaffolds with high cell infiltration after 1 week of culture (Blakeney et al. 2011). Another effective method for increasing void space between fibers, cell infiltration, and blood vessel ingrowth into electrospun scaffolds is low-temperature electrospinning, which is also known as cryogenic electrospinning and cold-plate electrospinning (Leong et al. 2009, 2010; Simonet et al. 2007).

Cryogenic Electrospinning Low-temperature or cryogenic electrospinning was first described by Simonet et al. (2007) as a simple process aimed at increasing void space between electrospun fibers. The electrospinning process which is shown in Fig. 2 remains largely the same with a small modification to the collection target. The target is maintained at a very cold temperature usually with dry ice, while the rest of the process is maintained at ambient temperature. Relative humidity is maintained at 30% or higher. High humidity (>30% relative humidity) and low temperatures of the collecting target causes quick condensation and ice crystal formation on the collecting target. Ice accumulation and electrospinning are deliberately synchronized to occur simultaneously, thereby, ice crystals prevent tight packing of electrospun fibers into dense scaffolds. Once the scaffold has formed and the electrospinning process is complete, ice crystals can be removed either by freeze drying, allowing them to melt, or washing them away with a warm solvent (Marrero and Heller 2012; Wasungu et al. 2009). Care must be taken to ensure that only water insoluble polymers are raised to melting temperature. Water-soluble polymers must be freeze-dried prior to cross-linking. Scaffolds must remain frozen until sublimation can occur, and freeze thaw-cycles must be avoided. Any melting of the ice crystals prior to cross-linking will result in full or partial destruction of the scaffold, such as electrospun silk fibroin. Once the ice crystals are completely removed, only the polymer fibers remain in the scaffolds, with large void spaces between the fibers. Figure 3 demonstrates a typical electrospinner system.

Cryogenic Electrospun Scaffolds for Oral Mucosa and Skin Cryogenic electrospinning of silk fibroin was first described by Bulysheva et al. (2012) with applications to 3D in vitro culture of epithelial lined tissues such as oral mucosa and head and neck carcinomas (Bulysheva et al. 2012, 2013). Silk fibroin is a natural material widely used clinically as a biocompatible material for implantable devices such as nonresorbable surgical suture. It is also widely used in tissue engineering due to its similarity in structure to collagen I and therefore biomimetic properties for cell attachment, mechanical properties, and nonimmunogenicity (Bulysheva et al. 2012, 2013). However, conventional electrospun silk supports

10

A.A. Bulysheva and R. Heller

Fig. 2 Cryogenic electrospinning process

low cell infiltration due to compacted fiber structure and therefore is difficult to use for true three-dimensional cell culture. Cryogenic electrospun silk is considerably more porous with significantly wider spaces between fibers. This feature was shown to significantly enhance cell infiltration and therefore expanded the applications for electrospun silk (Bulysheva et al. 2012). Conventional and cryogenic electrospun silk scaffolds were fabricated with the same electrospinning parameters, including silk fibroin concentration, applied electric field, distance between the collecting target and the polymer source, as well as translation and rotation speeds of the target. Cryogenic electrospun scaffolds were then freeze-dried to remove any residual water. Both scaffold types were cross-linked and disinfected in gradient ethanol submersion and then washed with PBS and media prior to cell culture. Immortalized dermal fibroblasts were seeded onto conventional and cryogenic electrospun scaffolds and cultured for up to 28 days. It was shown that cryogenic electrospun scaffolds facilitate dermal fibroblast infiltration throughout the entire depth of the scaffold, compared to conventional silk scaffolds, which maintained fibroblasts on the seeding surface only. Cell viability was measured with a Live/Dead stain, quantifying the percentage of viable and dead cells contained within each scaffold. Both electrospun scaffolds maintained high cell viability for the duration of the study up to 4 weeks. Mechanical properties such as strain at break, elasticity modulus, and peak stress were significantly reduced in cryo-spun scaffolds compared to controls, indicating limitations to applications where mechanical strength is not a virtue. Cryogenic electrospun scaffolds were seeded with oral keratinocytes and dermal fibroblasts in a coculture system modeling oral mucosa. In this case, fibroblast were seeded first and allowed to infiltrate throughout the scaffolds for several days prior to keratinocyte cell seeding. Cryogenic electrospun scaffolds supported high cell viability in the coculture system for up to 4 weeks. Proliferation was evaluated

3D Culture Models to Assess Tissue Responses to Electroporation

11

Fig. 3 Electrospinning process. Photograph of a standard electrospinner with a collector, and polymer syringe controlled by a syringe pump inside an insulated cabinet, and a power supply outside of the cabinet (a). Standard electrospun silk (b) compared to cryogenic electrospun silk (c)

with Ki67 staining, while keratinocyte differentiation was evaluated with staining for involucrin. It was shown that keratinocytes derived from normal oral mucosa were able to proliferate and differentiate to terminal differentiation when cultured on cryogenic electrospun scaffolds with dermal fibroblasts at air-liquid interface (Bulysheva et al. 2012). It was also shown, that oral carcinoma cells behaved differently on cryogenic electrospun silk scaffolds from normal oral mucosal keratinocytes, with high proliferation rates, high infiltration into the scaffolds, and low differentiation potential, as normally seen in vivo. Therefore, it was concluded that cryogenic electrospun silk scaffolds provided a suitable 3D microenvironment for oral mucosa studies (Bulysheva et al. 2012). This model was modified to mimic head and neck carcinoma in 3D culture. Studying the morphology of the tumor extracellular matrix under scanning electron microscopy allowed for mimicking its morphology with electrospinning parameters and assembling a silk electrospun scaffold that duplicated the native tumor extracellular matrix in terms of fiber size, orientation, and spacing (Bulysheva

12

A.A. Bulysheva and R. Heller

et al. 2013). Subsequent scaffold seeding with human head and neck carcinoma cells was performed and cell infiltration, viability, proliferation, and differentiation marker expression analysis were performed on these cells grown in 3D culture, in classic 2D culture, and in vivo in a mouse host. Using the same cell line in all three models allowed for a direct comparison between in vitro culture models and in vivo situation. The 3D culture model mimicked in vivo tumorigenesis more closely than the traditional 2D culture, in terms of gene expression profile as well as cell proliferation. This observation was further supported by chemoresistance studies. Three-dimensionally cultured carcinoma cells were largely resistant to up to 25-fold increase in Taxol concentration that would completely inactivate the same cells cultured in traditional 2D culture (Bulysheva et al. 2013). Cryogenic electrospun silk scaffolds were subsequently used for epidermal keratinocyte coculture with dermal fibroblasts (Sheikh et al. 2015) for modeling skin similar to previously established mucosal keratinocyte coculture with dermal fibroblasts (Bulysheva et al. 2012). This skin model may behave similar to the oral mucosal model in terms of fibroblast infiltration and keratinocyte proliferation and differentiation; however, no data was provided to prove presence of keratinocytes after seeding (Sheikh et al. 2015). Further validation would be needed to determine whether different epidermal strata were formed and were distinct from the dermal strata, in order to indicate whether this 3D culture model is representative of normal human skin. These scaffolds can easily be manufactured to a particular thickness (e.g., 200 μm) and cut to convenient pieces for well-plate culture (e.g., 12 mm in diameter for a 24-well plate). The mechanical properties of these scaffolds are such that each disk with cultured cells can be transferred from well to well or from well to electroporation cuvette. It would also be relatively easy to apply electroporation electrodes such as parallel-plate, pin, or needle electrodes typically used for in vivo applications to this in vitro electrospun silk model. While this is a potentially useful 3D culture in vitro model, it has not yet been used for studying electroporation effects.

Recellularized Human Dermis A more complete 3D culture of human skin was recently developed for evaluating tissue response to electroporation and gene delivery (Bulysheva et al. 2016). Donor decellularized human dermis was recellularized with human keratinocytes (HaCat cells) and primary human dermal fibroblasts. Dermal fibroblasts were seeded at a higher concentration on the reticular surface of the dermis and were allowed to infiltrate the dermis for 1 week after seeding. The grafts were flipped over and keratinocytes were then seeded on the papillary surface. There was evidence of an intact basement membrane between the epidermis and the dermis demonstrated by collagen IV staining and supported by distinctly separate epidermal and dermal layers. The epidermis was further shown to contain all distinct layers characteristic of normal human thin skin. This model was used for luciferase encoding plasmid DNA delivery with electroporation to demonstrate its utility in studying electroporation of human skin ex vivo (Bulysheva et al. 2016).

3D Culture Models to Assess Tissue Responses to Electroporation

13

Recellularized dermis or cryogenically electrospun constructs seeded with cells can be readily, manually manipulated to be moved with sterile forceps from well to well, or into an electroporation cuvette or between electrodes for electroporation, without damage to the constructs. The size of these constructs is typically chosen to fit into the wells of a 48 or 24-well plate out of convenience considerations. Thus the constructs are often 10–12 mm in diameter and ~0.2–0.3 mm in thickness. Recellularized dermis constructs can also be injected with ~50 μL of DNA similar to intradermal injections in vivo for studying gene delivery to the skin. These considerations are important for evaluating the practicality of using these 3D in vitro models for studying tissue response to electroporation. It would be difficult to use the standard organotypic skin model for the same experiments with the culture system lacking mechanical integrity.

Conclusion Three-dimensional cell culture is becoming more popular for studying various tissues of the human body. In particular, many 3D culture models have been utilized for studying effects of pulsed electric fields. Primarily, the spheroid models have contributed to improved understanding of electroporation effects on tumors. While spheroid models are wonderful tools for studying solid tumor responses to electroporation, other normal tissues like skin are not well represented by spheroid cell culture methods. Cryogenic electrospun silk scaffolds have been used for 3D culture of oral mucosa, head, and neck squamous carcinoma and skin. These 3D culture systems can be easily used for evaluating tissue response to electroporation. The constructs are representative of their respective tissues, mimicking cell-cell and cell matrix interactions while recapitulating the necessary cellular microenvironment. Another 3D skin model was developed specifically for studying gene transfer parameters to the skin. The recellularized human dermis contains all appropriate epidermal and dermal layers of the skin, as well as dermal extracellular matrix. This 3D skin culture system and many other tissue appropriate models are currently available for studying specific tissue responses to electroporation or gene electrotransfer.

References Barnes CP, Sell SA, Boland ED, Simpson DG, Bowlin GL (2007) Nanofiber technology: designing the next generation of tissue engineering scaffolds. Adv Drug Deliv Rev 59(14):1413 Blakeney B, Tambralli A, Anderson J et al (2011) Cell infiltration and growth in a low density, uncompressed three-dimensional electrospun nanofibrous scaffold. Biomaterials 32(6):1583 Brauker JH, Carr-Brendel VE, Martinson LA, Crudele J, Johnston WD, Johnson RC (1995) Neovascularization of synthetic membranes directed by membrane microarchitecture. J Biomed Mater Res 29(12):1517–1524 Bulysheva A, Bowlin G, Klingelhutz A, Yeudall WA (2012) Low-temperature electrospun silk scaffold for in vitro mucosal modeling. J Biomed Mater Res A 100(3):757–767

14

A.A. Bulysheva and R. Heller

Bulysheva AA, Bowlin GL, Petrova SP, Yeudall WA (2013) Enhanced chemoresistance of squamous carcinoma cells grown in 3D cryogenic electrospun scaffolds. Biomed Mater 8(5):055009. doi:10.1088/1748-6041/8/5/055009 Bulysheva AA, Burcus N, Lundberg C, Edelblute CM, Francis MP, Heller R (2016) Recellularized human dermis for testing gene electrotransfer ex vivo. Biomed Mater 11(3):035002. doi:10.1088/1748-6041/11/3/035002 Canatella PJ, Black MM, Bonnichsen DM, McKenna C, Prausnitz MR (2004) Tissue electroporation: quantification and analysis of heterogeneous transport in multicellular environments. Biophys J 86(5):3260–3268. doi:10.1016/S0006-3495(04)74374-X Chopinet L, Wasungu L, Rols MP (2012) First explanations for differences in electrotransfection efficiency in vitro and in vivo using spheroid model. Int J Pharm 423(1):7–15. doi:10.1016/j. ijpharm.2011.04.054 Frandsen SK, Gibot L, Madi M, Gehl J, Rols MP (2015) Calcium electroporation: Evidence for differential effects in normal and malignant cell lines, evaluated in a 3D spheroid model. PLoS One 10(12):e0144028. doi:10.1371/journal.pone.0144028 Freshney RI 2005 Culture of animal cells: a manual of basic technique, 5th edn. Wiley-Liss, Hoboken, p 642. http://www.loc.gov/catdir/enhancements/fy0626/2005281585-d.html; http:// www.loc.gov/catdir/enhancements/fy0626/2005281585-b.html Gibot L, Rols MP (2013) 3D spheroids’ sensitivity to electric field pulses depends on their size. J Membr Biol 246(10):745–750. doi:10.1007/s00232-013-9535-x Gibot L, Wasungu L, Teissie J, Rols MP (2013) Antitumor drug delivery in multicellular spheroids by electropermeabilization. J Control Release 167(2):138–147. doi:10.1016/j. jconrel.2013.01.021 Griffith L, Swartz M (2006) Capturing complex 3D tissue physiology in vitro. Nat Rev Mol Cell Biol 7(3):211 Jin H, Chen J, Karageorgiou V, Altman G, Kaplan D (2004) Human bone marrow stromal cell responses on electrospun silk fibroin mats. Biomaterials 25(6):1039 Leong M, Rasheed M, Lim T, Chian K (2009) In vitro cell infiltration and in vivo cell infiltration and vascularization in a fibrous, highly porous poly(D,L-lactide) scaffold fabricated by cryogenic electrospinning technique. J Biomed Mater Res A 91(1):231–240 Leong M, Chan W, Chian K, Rasheed M, Anderson J (2010) Fabrication and in vitro and in vivo cell infiltration study of a bilayered cryogenic electrospun poly(D,L-lactide) scaffold. J Biomed Mater Res A 94(4):1141–1149 Marrero B, Heller R (2012) The use of an in vitro 3D melanoma model to predict in vivo plasmid transfection using electroporation. Biomaterials 33(10):3036–3046. doi:10.1016/j. biomaterials.2011.12.049 Marrero B, Messina JL, Heller R (2009) Generation of a tumor spheroid in a microgravity environment as a 3D model of melanoma. In Vitro Cell Dev Biol Anim 45(9):523–534. doi:10.1007/s11626-009-9217-2 Mellor HR, Davies LA, Caspar H et al (2006) Optimising non-viral gene delivery in a tumour spheroid model. J Gene Med 8(9):1160–1170. doi:10.1002/jgm.947 Sheikh FA, Ju HW, Lee JM et al (2015) 3D electrospun silk fibroin nanofibers for fabrication of artificial skin. Nanomedicine 11(3):681–691. doi:10.1016/j.nano.2014.11.007 Simonet M, Schneider OD, Neuenschwander P, Stark WJ (2007) Ultraporous 3D polymer meshes by low-temperature electrospinning: use of ice crystals as a removable void template. Polym Eng Sci 47(12):2020–2026 Stankus JJ, Soletti L, Fujimoto K, Hong Y, Vorp DA, Wagner WR (2007) Fabrication of cell microintegrated blood vessel constructs through electrohydrodynamic atomization. Biomaterials 28(17):2738 Wasungu L, Escoffre JM, Valette A, Teissie J, Rols MP (2009) A 3D in vitro spheroid model as a way to study the mechanisms of electroporation. Int J Pharm 379(2):278–284. doi:10.1016/j. ijpharm.2009.03.035

Effect of Pulsed Electric Fields on Food Constituents Eldevan S. Silva, Shahin Roohinejad, Mohamed Koubaa, Francisco J. Barba, Anet Režek Jambrak, Tomislava Vukušić, Mauro D. Santos, Rui P. Queirós, and Jorge A. Saraiva

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Proteins and Free Amino Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Lipids and Fatty Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Vitamins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Polyphenols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

E.S. Silva (*) Departamento de Química, Universidade Federal de São Carlos, São Carlos, Brazil e-mail: [email protected] S. Roohinejad Department of Food Science, University of Otago,, Dunedin, New Zealand e-mail: [email protected] M. Koubaa Laboratoire Transformations Intégrées de la Matière Renouvelable (UTC/ESCOM, EA 4297 TIMR), Centre de Recherche de Royallieu, Sorbonne Universités, Université de Technologie de Compiègne,, Compiègne, France e-mail: [email protected]; [email protected] F.J. Barba (*) Faculty of Science, Department of Food Science, University of Copenhagen, Frederiksberg C, Denmark Faculty of Pharmacy, Preventive Medicine and Public Health, Food Sciencs, Toxicology and Forensic Medicine Department, Nutrition and Food Science Area, Universitat de València Health Department, València, Spain e-mail: [email protected]; [email protected] A.R. Jambrak Faculty of Food Technology and Biotechnology, Department of Process Engineering, University of Zagreb, Zagreb, Croatia e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_31-2

1

2

E.S. Silva et al.

Abstract

The use of pulsed electric fields (PEF) for food preservation has been increasingly studied in the recent decades by many research groups. This technology has been proven as a good alternative to conventional heat treatments as it better preserves the nutritional and organoleptic characteristics of the products. For instance, several studies have assessed the impact of PEF processing on different nutritional compounds and the quality of the final product. The main molecules that have retained the attention of food scientists and industries are proteins and free amino acids, lipids and free fatty acids, vitamins, and polyphenols. PEF processing of food products demonstrated changes in the structural conformation of proteins and increased amounts of free amino acids. Some research works have reported minor changes in the total fat and either increased or decreased amounts of free fatty acids. Vitamins showed also some modifications and conversions of carotenoid molecules into others. Regarding total phenolic compounds, it was demonstrated that they are resistant to the effects of PEF. The use of PEF processing may thus involve, in addition to the preservation of food products, some changes in the nutritional attributes, which represent an important challenge to replace heat processing in food industries. Some examples that demonstrate the impact of electrical pulses on the different food constituents (macro- and micronutrients as well as bioactive compounds) are shown in this chapter. Keywords

Pulsed electric fields • Proteins and free amino acids • Fat and fatty acids • Vitamins, bioactive compounds • Electroporation

Introduction The use of nonthermal food preservation technologies has increased exponentially in the recent years in response to the consumer’s demands for microbiologically safe food products with higher quality, from both nutritional and physicochemical point of views. The use of high pressure processing (HPP) for nonthermal pasteurization of foods has found its niche into the market, with a growing number of companies using this technology for different food products (e.g., fruit juices, ready-to-eat meals, meat products). In addition to HPP, pulsed electric fields (PEF) technology T. Vukušić Faculty of Food Technology and Biotechnology, Department of Process Engineering, University of Zagreb, Zagreb, Croatia e-mail: [email protected] M.D. Santos • R.P. Queirós • J.A. Saraiva QOPNA, Departamento de Química, Universidade de Aveiro, Campus Universitário de Santiago, Aveiro, Portugal e-mail: [email protected]; [email protected]; [email protected]

Effect of Pulsed Electric Fields on Food Constituents

3

had a major impact on food processing during the 1990s and in the early years of the twenty-first century. Over that time period, several PEF studies carried out were mainly focused on the microbial inactivation of liquid food products (e.g., milk, juices). However, several factors such as the global economic crisis and the resistance to inactivation of some microorganisms by this technology limited its use in food preservation for commercial purposes. In the last decade, the emergence of better designed and more affordable equipment has led numerous food industries to reconsider this technology and to implement it in their existing installations. The main application of this technology relates to food preservation by submitting food, liquid, or semi-solid products, to very short (from nanoseconds to milliseconds) electrical discharge of high-voltage pulses (up to 80 kV/cm) at low/moderate temperatures. PEF treatment potentially inactivates spoilage and pathogenic microorganisms and enzymes with a small impact on nutritional/sensorial properties, unlike thermal treatments that usually lead to higher changes in food properties. To that purpose, food products are placed in a PEF chamber between two electrodes, and the inactivation of microorganisms occurs mainly due to the phenomenon of electroporation in the microorganisms’ membranes since the presence of several ions on foods give to each product a degree of electrical conductivity that allows the electrical current to flow. After PEF treatment, food products should be aseptically packaged and stored under refrigeration, since PEF is not able to inactivate spores. Furthermore, for food preservation purposes, the possibility of combining this technology with heat treatment during short periods in order to improve its efficiency and effectiveness in microbial inactivation has also been indicated in the literature. PEF treatment could be used for several other food applications apart from food preservation, including the improvement of drying and freezing processes, as well as improving juice production, extraction of intracellular valuable compounds, reduction of food contaminants, etc. (Koubaa et al. 2015; Barba et al. 2015). Several studies have demonstrated the promising potential of PEF technology to retain valuable components in food products, better preserving the original color, flavor, texture, and nutritional value. However, this type of treatment may have some effects on food constituents. Overall, as demonstrated by a large number of studies, small molecules from plants are not significantly affected by PEF. Moreover, after PEF treatment, the concentration of many metabolites (e.g., isoflavonoids or polyphenolics) seems to increase in many cases, most likely due to their enhanced release and extractability, because of the permeabilization of the plant cells. In what regards proteins, globally, there are some effects in their structure (particularly on the secondary and tertiary structures) leading to denaturation and subsequent changes in their techno-functional properties, such as solubility, hydrophobicity, and gelation or aggregation. Particularly on enzymes, PEF impact is extremely dependent on the type of enzyme and treatment conditions, some being difficult to inactivate leading sometimes to contradictory results. Taking lipids into account, some electrochemical reactions promoted by PEF can occur, possibly resulting in changes in the lipids structure, leading to their oxidation.

4

E.S. Silva et al.

Although the first studies regarding PEF technology were mainly focused on the inactivation of microorganisms (pathogenic and nonpathogenic), the latter research performed allowed to better understand the real impact of this technology on other molecules and foods characteristics. Thus, the effects of PEF on proteins, lipids, vitamins, and polyphenols will be further discussed, and some of the most relevant results are reviewed in more detail (Table 1).

Table 1 The effect of PEF treatment on food constituents Food constituent Proteins and free amino acids

PEF parameters E: 35 kV/cm

E: 20–40 kV/cm, 200 μs Bipolar electric field pulses of 4 μs width, E: 35 kV/ cm. The pulse repetition rate was 1000 Hz and the total PEF treatment time was 1 ms

Bipolar squarewave pulses of 4 μs, with a frequency of 100 Hz, and 35 kV/ cm field strength during 1500 μs and 1700 μs, respectively

Lipids and fatty acids

E: 35 and 40 kV/ cm, six different treatment times from 40 to 180 μs E: 35 kV/cm, 800 and 1400 μs

PEF treatment main effect PEF treatment induced changes in the bond’s vibration within the side chains of amino acids, antiparallel β-sheets, β-turns, and β-sheet structure A significant increase in the total FAA especially in theanine was observed PEF treatment did not modify the total content of fatty acids and FAA in Parellada grape juice The level of lauric acid significantly decreased after PEF processing The concentration of some amino acids (e.g., histidine, tryptophan, asparagine, phenylalanine, and ornithine) increased after PEF treatment A significant decrease in FAA of tomato juices was observed PEF treatments retained higher amounts (2.5 %) of FAA than thermally treated samples The concentrations of phenylalanine, glutamic acid, valine, serine and alanine increased by 27 %, 6.8 %, 6.3 %, 5.5 %, and 4.8 %, respectively During storage, a higher amino acid content was observed in PEF-treated juices Non-significant changes in the contents of saturated fatty acids, monounsaturated fatty acids, or polyunsaturated fatty acids were observed A significant decrease in elaidic and linoleic acids after PEF treatment was observed PEF-treated samples had higher level of fatty acids during storage compared to the thermally treated samples

References Liu et al. (2011)

Zhao et al. 2009 Garde-Cerdán et al. (2007)

OdriozolaSerrano et al. 2013

Zulueta et al. (2007)

Morales-de la Pena et al. (2011a)

(continued)

Effect of Pulsed Electric Fields on Food Constituents

5

Table 1 (continued) Food constituent Vitamins

PEF parameters E: 28 kV/cm, 50 pulses, treatment temperature less than 34  C E: 40 kV/cm, 57 μs

E: 20–35 kV/cm, 2000 μs E: 30–35 kV/cm, 50–250 Hz, 50–2050 μs E: 0–35 kV/cm, 0.2–2 μs

E: 15–35 kV/cm, 100–1000 μs E: 0.5, 12.5, 25 and 35 kV/cm, frequency: 1 kHz, pulse width: 40 μs, unipolar squarewave E: 15–40 kV/cm, 40–700 μs

E: 22.6 kV/cm, 400 μs E: 35 kV/cm, 750 μs

E: 25, 30, 35, and 40 kV/cm, 30–340 μs

PEF treatment main effect A loss in the concentration of vitamin C (approximately 8 %) was observed

References Cserhalmi et al. (2006)

The concentration of vitamin C in tomato juice reduced as the storage time increased PEF treatment (storage at 4  C for 42 days) significantly retained a higher amount of vitamin C Kinetic models predicted vitamin C retention of PEF-treated juice sample

Min et al. (2003)

PEF treatment reduced vitamin C content of watermelon juice The content of vitamin C in apple juice decreased significantly during PEF treatment The largest vitamin C loss (36.6 %) was observed at 30 kV/cm and 20 μs High losses of vitamin C (2.9–15.7 %) in orange juice and “gazpacho” were observed PEF treatment promoted the conversion of vitamin C from enolform to keto-form

The concentration of water-soluble vitamins remaining in the PEF-treated juice was similar to that found in the heat-treated juice at 84  C No significant changes in riboflavin and thiamine content of milk were observed Neither significant change in the individual carotenoids (β-cryptoxanthin, zeaxanthin, lutein, β-carotene, α-carotene) nor in the content of vitamin A was found PEF processing caused a significant increase in the concentrations of 9-cisviolaxanthin + neoxanthin mixture, antheraxanthin, cis-β-cryptoxanthin and 9-cis-α-carotene

OdriozolaSerrano et al. (2008a) Oms-Oliu et al. (2009) Bi et al. (2013)

Elez-Martínez and MartínBelloso (2007) Zhang et al. 2015

Rivas et al. (2007)

Bendicho et al. (2002) SánchezMoreno et al. (2005)

Torregrosa et al. (2005)

(continued)

6

E.S. Silva et al.

Table 1 (continued) Food constituent Vitamins

Polyphenols

PEF parameters (E: 35 kV/cm, 1500 μs), bipolar pulses of 4 μs at 100 Hz

Moderate-intensity PEF: E: 1 kV/cm, 16 monopolar pulses of 4 μs, frequency of 0.1 Hz High-intensity PEF: E: 35 kV/cm, 1500 μs, bipolar squared-wave pulses of 4 μs, frequency of 100 Hz E: 15, 25, 35, and 40 kV/cm, Treatments times from 40 to 700 μs E: 35 kV/cm, 800 μs and 1400 μs

E: 35 kV/cm, 750 μs (E: 35 kV/cm, 1500 μs), bipolar pulses of 4 μs at 100 Hz

PEF treatment main effect The content of some carotenoids (lycopene, β-carotene and phytofluene) in tomato juice was increased A higher content of carotenoids (lycopene, neurosporene and γ-carotene) and quercetin was observed after PEF treatment during storage compared to thermal treatment samples Moderate PEF processing increased the content of carotenoids in tomato juices, except for the 5-cis-lycopene isomer, which remained unchanged Increase of 63–65 % in 15-cislycopene was reported after moderate PEF treatment High-intensity PEF treated juices had higher carotenoids content (10–20 %) than thermally treated samples The combination of moderate and high intensity PEF treatments led to a higher quality of tomato juice The conversion of zeaxanthin into neoxanthin and violaxanthin, and the conversion of α-carotene to zeaxanthin were observed Biochemical reactions during PEF processing, resulted in the formation of new phenolic compounds Significant effects on the cell membranes or in phenolic complexes with other compounds were observed Some free phenolic compounds were released after PEF processing Possible inactivation of polyphenol oxidase after PEF treatment Further loss of phenolic compounds was prevented No significant changes in the total flavanones, aglycones, hesperidin and naringenin contents were observed A higher retention of total anthocyanins in strawberry juice, after PEF treatment using bipolar mode was reported

References OdriozolaSerrano et al. (2009)

VallverduQueralt et al. (2013)

Zulueta et al. (2010)

Morales-de La Pena et al. (2011b)

SánchezMoreno et al. (2005) OdriozolaSerrano et al. (2009) (continued)

Effect of Pulsed Electric Fields on Food Constituents

7

Table 1 (continued) Food constituent Polyphenols

PEF parameters E: 35 kV/cm, Bipolar pulses of 4 μs, frequency of 1200 pulses E: 35 kV/cm for 1500 μs in bipolar 4 μs pulses at 100 Hz, with an energy density of 8269 kJ/L E = 800 V/cm

Moderate-intensity PEF: E: 1 kV/cm, 16 monopolar pulses of 4 μs, frequency of 0.1 Hz High-intensity PEF: E: 35 kV/cm, 1500 μs, bipolar squared-wave pulses of 4 μs, frequency of 100 Hz E: 1.2–3.0 kV/cm

PEF treatment main effect Application of PEF resulted in the lower loss of total phenolic compounds (32.2 vs. 14.49 %)

References Aguilar-Rosas et al. (2007)

(No significant changes in phenolic compounds were observed: 26.2 mg/ 100 g in PEF-treated sample vs. 26.5 mg/100 g in high pasteurization)

OdriozolaSerrano et al. (2008b)

At fixed value of extraction yield (e.g. Y = 0.6), an increase in the total phenolic compounds (by  1.16) was reported The highest concentrations of total phenolic compounds (CTPC  1310.4 mg GAE/L) were obtained with juice yields of  18 % Tomatoes treated under moderateintensity PEF provided juices with higher content of polyphenol compounds Over storage time, a slight reduction in the polyphenols content was observed in thermal- and PEF-treated juices, except for caffeic acid PEF-treated samples had higher content of ferulic acid, chlorogenic acid, p-coumaric acid, caffeic-Oglucoside acid, rutin, and naringenin following the treatment and during the storage than thermal treated juices Chalcone was formed after the treatment, which was considered as the first step for degradation of anthocyanins together with opening of the pyrilium ring

CarbonellCapella et al. (2016)

VallverduQueralt et al. (2012)

Zhang et al. (2008)

Notes. FAA free amino acids, CTPC concentrations of total phenolic compounds

Proteins and Free Amino Acids Proteins and amino acids are one of the most important nutrients for human health, and it is essential to them from denaturation during food processing. The negative effect of thermal processing on denaturation of proteins has been widely reported in the literature. However, only few studies evaluated the changes of secondary or

8

E.S. Silva et al.

tertiary structure and denaturation of proteins using emerging technologies such as PEF treatment. For instance, Liu et al. (2011) analyzed the changes in protein profile and structure of a soybean solution after applying PEF treatment, and using Fourier transform infrared spectroscopy (FTIR) and differential scanning calorimetry (DSC). For PEF, the pulse intensity applied was ranging from 0 to 50 kV/cm, and the pulse width, frequency, and flow speed were 40 μs, 1.0 kHz, and 10 mL/min, respectively, with a total treatment time of 4.8 ms. PEF treatment induced changes in the bond’s vibration within the side chains of amino acids, antiparallel β-sheets, β-turns, and β-sheet structure, which suggested the unfolding of proteins. In a similar study, PEF treatment (20–40 kV/cm, 200 μs) of green tea infusions caused a significant increase in the total free amino acids (FAA) especially in theanine (Zhao et al. 2009). These results suggested that PEF treatment may be valuable to increase the quality of ready-to-drink liquid foods. Garde-Cerdán et al. (2007) studied the impact of thermal and PEF treatments processing on fatty acids and FAA contents of grape juice. These compounds are important in winemaking as nutritive compounds for the growth of yeast. Neither thermal nor PEF treatments modified the total content of fatty acids and FAA in Parellada grape juice, except for lauric acid, which decreased considerably after PEF processing, and the concentration of some amino acids (histidine, tryptophan, asparagine, phenylalanine, and ornithine), which increased after PEF treatment. Recently, the impact of high-intensity PEF on the amino acids profiles of tomato and strawberry juices was evaluated and compared with thermally treated (90  C for 60 s) juices (Odriozola-Serrano et al. 2013). A continuous flow bench scale system was used for PEF treatment. These authors found a significant decrease in FAA of tomato juices in both PEF- and thermally treated samples compared to untreated samples, although PEF treatments retained higher amounts (2.5 %) of FAA than thermally treated samples. This study also reported that PEF treatment increased the concentrations of phenylalanine, glutamic acid, valine, serine, and alanine, by 27 %, 6.8 %, 6.3 %, 5.5 %, and 4.8 %, respectively. Thermal treatment, on the other hand, reduced the content of most amino acids, comparatively to the untreated fresh juice. It was demonstrated that in both treated tomato juice, total FAA increased significantly during the storage period; however, the concentrations of some minor amino acids (e.g., proline, leucine, valine, isoleucine, arginine, lysine, phenylalanine, and methionine) decreased over the same period. In other words, during storage, a higher amino acid content was observed in PEF-treated juices when compared to the thermally treated samples. The results showed that the concentrations of individual (except for leucine, which decreased by 23 %) and total amino acids have increased during storage. The promising potential of PEF as preservation and processing method to stabilize juices and to increase their nutritional value, by increasing the concentration of amino acids, was demonstrated in this study.

Lipids and Fatty Acids Unsaturated fatty acids are a nutritionally interesting topic because of their positive health benefit effects. Nowadays, the consumption of these compounds in food

Effect of Pulsed Electric Fields on Food Constituents

9

products has attracted both public and industrial attentions. Thus, it is important to preserve these compounds during food processing. The effect of heat treatments on the stability of fatty acids has been studied most and has not been reported to produce a decrease in the content of fatty acids. However, only few studies evaluated the effect of emerging technologies such as PEF treatment on the stability of lipids and fatty acids. For instance, Zulueta et al. (2007) evaluated the effect of PEF treatment on various physicochemical properties and on the fatty acid profile changes of an orange juice-milk beverage. After PEF treatment, nonsignificant changes in the contents of saturated fatty acids, monounsaturated fatty acids, or polyunsaturated fatty acids were observed (only a small reduction in fat content was found). In another study, the effects of PEF (35 kV/cm, 800 and 1400 μs) on fatty acid composition of a fruit juice-soymilk beverage and a subsequent storage during 56 days at 4  C were evaluated and compared with thermal treatments (90  C for 60 s) at equivalent storage time (Morales-de la Pena et al. 2011a). The authors only observed a significant decrease in elaidic and linoleic acids after PEF or thermal treatments. Moreover, PEF-treated samples showed higher level of fatty acids during storage compared to the thermally treated samples.

Vitamins Vitamin C is a very well-known water-soluble vitamin with several health benefits for human health, not only for its antiscorbutic activity, but also due to its several physiological activities, most of them related to its ability to act as a potent free radicals scavenger. Vitamin retention studies to evaluate the impact of food processing on the nutritive value of foods are of great importance as they can be excellent predictors of the subsequent product consumer’s acceptance. For instance, vitamin C has been traditionally used as an indicator of the loss of other nutritionally valuable compounds by applying thermal processing. It must be taken into account that its concentration may decrease during processing, depending on the treatment conditions, such as temperature, oxygen content, and light. Several studies have been published regarding the effects of PEF processing on vitamin C content, and the overall conclusion was that PEF treatment does not significantly decrease the vitamin C content in liquid foods. For instance, the effect of PEF processing applied in a continuous system on physical and chemical properties of freshly squeezed citrus juices (grapefruit, lemon, orange, tangerine) was studied by Cserhalmi et al. (2006). The juices were treated by 50 pulses at 28 kV/cm and the treatment temperature was less than 34  C. They showed a loss in the concentration of vitamin C as a result of PEF treatment of approximately 8 %, while a higher loss (17.2 %) was observed for the heat-pasteurized samples. In another study, the effect of commercial scale PEF processing (40 kV/cm for 57 μs) on the concentration of vitamin C of tomato juices was studied and compared with heat-treated (hot break at 88  C for 2 min or by cold break at 68  C for 2 min) samples (Min et al. 2003). The concentration of vitamin C in tomato juice reduced as the storage time increased regardless of processing methods. However, tomato juice

10

E.S. Silva et al.

treated by PEF processing (storage at 4  C for 42 days) significantly retained a higher amount of vitamin C compared to conventional thermal treatment. Similar results were also obtained by Odriozola-Serrano et al. (2008a) in tomato juice after PEF application (20–35 kV/cm, 2000 μs) using bipolar squared-wave pulses at 250 Hz. Kinetic models predicted vitamin C retention of PEF-treated juice sample. In contrast of these studies, Oms-Oliu et al. (2009) reported that severe PEF treatments (30–35 kV/cm, 50–250 Hz, 50–2050 μs) were reduced vitamin C content of watermelon juice. In another study, the content of vitamin C from apple juice decreased significantly during PEF treatment (0–35 kV/cm, 0.2–2 μs) (Bi et al. 2013) and the largest loss (36.6 %) was observed at 30 kV/cm and 20 μs. Elez-Martínez and Martín-Belloso (2007) also obtained higher losses of vitamin C (2.9–15.7 %) in orange juice and “gazpacho” when high electric field strengths, frequencies, pulse widths, and treatment times were applied (15–35 kV/cm, 100–1000 μs) using bipolar mode. In a recent study, the impact of PEF on vitamin C structure and its antioxidant properties was evaluated (Zhang et al. 2015). In this study, it was concluded that the formation of free hydroxyl radicals (˙OH) after PEF treatment could have a beneficial effect on the antioxidant properties of vitamin C, and this might be attributed to the possible vitamin C structure modifications. The authors observed that PEF treatment promoted the conversion of vitamin C from enol-form to keto-form by determining the fluorescence intensity of vitamin C before and after the treatments, thus indicating the conversion of vitamin C induced by hydroxyl radicals. It has been previously demonstrated that ascorbate and dehydroascorbate molecules are rapidly produced by pairs of ascorbate radicals (Retsky et al. 1993). Nevertheless, no antioxidant capacity is exhibited by dehydroascorbate; thus, it is converted back into ascorbate by adding two electrons (Fig. 1). In this reaction, an electron is donated by an oxidoreductase to the lipid radical in order to terminate the lipid peroxidation chain reaction, thus resulting in an antioxidant effect. Rivas et al. (2007) compared the water-soluble vitamin content (biotin, folic acid, riboflavin, and pantothenic acid) of a PEF treated blended orange and carrot juice (15–40 kV/cm, 40–700 μs) with heat-treated juice (84  C and 95  C, 45 s). The concentration of water-soluble vitamins remaining in the PEF-treated juice was similar to that found in the heat-treated juice at 84  C, but the heat-treated juice at 95  C showed higher losses. Bendicho et al. (2002) studied the modification in vitamin content of milk after PEF processing (22.6 kV/cm, 400 μs) and found no significant changes in riboflavin and thiamine content. With regard to liposoluble vitamins, few studies evaluated the effect of PEF on this kind of compounds. Carotenoids are naturally lipid-soluble pigments in numerous fruits and vegetables. All photosynthetic organisms, as well as a large variety of nonphotosynthetic bacteria and fungi, are able to synthesize carotenoids. These molecules constitute a class of liposoluble tetraterpenes, which originate from the condensation of isoprenyl units, constituting a series of conjugated double bond molecules. According to Eonseon et al. (2003), naturally occurring carotenoids could be classified in two main classes: (1) carotenes (e.g., β-carotene and α-carotene), which are cyclized lycopene, and (2) xanthophylls, which are oxygenated derivatives of carotenes.

Effect of Pulsed Electric Fields on Food Constituents

11

Fig. 1 Proposed mechanism of the radical scavenging activity of vitamin C

Up to now, several works were basically focused on the effect of PEF on the extractability of carotenoids, in which some of them showed provitamin A activity (Roohinejad et al. 2014). The retention of provitamin A is favored by processing at low temperatures, protected from light, exclusion of oxygen, and the presence of antioxidants. Sánchez-Moreno et al. (2005) observed significant change neither in the individual carotenoids (β-cryptoxanthin, zeaxanthin, lutein, β-carotene, α-carotene) nor in the content of vitamin A, when applying PEF (35 kV/cm, 750 μs) in comparison with heat-treated juice. This result differed from that obtained by Torregrosa et al. (2005), who evaluated the effects of PEF processing, with various electric field intensities (25, 30, 35, and 40 kV/cm) and different treatment times (30–340 μs) of carotenoids on an orange-carrot juice mixture (80:20, v/v). In parallel, a conventional heat treatment (98  C, 21 s) was used to the orange-carrot juice, and the results were compared. They concluded that PEF processing generally caused a significant increase in the concentrations of the different identified carotenoids (9-cis-violaxanthin + neoxanthin mixture, antheraxanthin, cis-β-cryptoxanthin and 9-cis-α-carotene), which was proportional to the treatment time. However, when conventional pasteurization was used to process the juice, the concentrations of most of the carotenoids decreased or showed a nonsignificant increase. With PEF treatment at 25 and 30 kV/cm, it was possible to obtain a higher concentration of vitamin A than that found in the thermal pasteurized juice. A possible explanation for this fact is that the formation of radicals during PEF could induce the conversion of carotenoids. A diagrammatic representation of the potential changes is established in Fig. 2. Moreover, Min et al. (2003) found that lycopene concentration was not modified when tomato juice was treated by PEF. Odriozola-Serrano et al. (2008b) and Oms-Oliu et al. (2009) also obtained an increase of lycopene concentration in tomato and watermelon juices after PEF treatments, respectively. In another study, the effect of PEF processing (electric field intensities of 15, 25, 35, and 40 kV/cm, and by treatment times from 40 to 700 μs) on the carotenoid and vitamin A profile of an orange juice-milk beverage was investigated and compared with a pasteurization

Fig. 2 Proposed biosynthetic pathways of carotenes and xanthophylls

12 E.S. Silva et al.

Effect of Pulsed Electric Fields on Food Constituents

13

treatment (90  C, 20 s) (Zulueta et al. 2010). It was demonstrated that the treatment of orange juice-milk by PEF led to the conversion of zeaxanthin into neoxanthin and violaxanthin and the conversion of α-carotene to zeaxanthin (Fig. 2). In addition, PEF-treatment resulted in degradation of some other carotenoids in the sample. Odriozola-Serrano et al. (2009) studied the impact of PEF processing (35 kV/cm, 1500 μs of overall treatment time, bipolar pulses of 4 μs at 100 Hz) and heat pasteurization (90  C for 30 s or 60 s) on the contents of carotenoids and phenolic compounds of tomato juice. The authors investigated the changes in some quality attributes (pH, soluble solids, and color parameters) and compared the obtained results with the untreated juice as control. Results showed that PEF treatment enhanced the content of some carotenoids such as lycopene, β-carotene, and phytofluene, as well as the red color of the juices. Nonetheless, differences in the polyphenols content, pH, and soluble solids were not significant between the treated and the untreated samples. In addition, some health-related compounds slightly increased over time, except for some carotenoids (β-carotene and phytoene) and caffeic acid. A higher content of carotenoids (lycopene, neurosporene and γ-carotene) and quercetin was observed after PEF treatment during storage compared to thermal treatment samples. The authors concluded that PEF processing of tomato juice could increase its nutritional quality and safety. In another recent study conducted by Vallverdu-Queralt et al. (2013), the effect of PEF processing on carotenoids content of tomato juice was investigated. The authors applied first a moderate PEF treatment (Electric field strength of 1 kV/cm using 16 monopolar pulses of 4 μs at a frequency of 0.1 Hz) to raw tomato, which was immediately refrigerated at 4  C during 24 h. Samples were then grinded to obtain the juice, which was treated by high-intensity PEF (electric field strength of 35 kV/cm for 1500 μs using bipolar squared-wave pulses of 4 μs and a frequency of 100 Hz) and then stored under chilled conditions during 56 days. The results showed that moderate PEF processing increased the content of carotenoids in tomato juices, except for the 5-cis-lycopene isomer, which remained unchanged after PEF treatment. The authors reported an increase by 63–65 % in 15-cis-lycopene after moderate PEF treatment. Moreover, over the time, a slight increase was observed in cis-lycopene isomers comparatively to the other carotenoids, which decreased slightly. By comparing high-intensity PEF treatment to thermal processing, it has been reported that the PEF-treated tomato juices maintained higher contents in carotenoids (10–20 %). The authors concluded that the combination of moderate and high-intensity PEF treatments led to a higher quality of tomato juice, not only by increasing the total carotenoids but also by maintaining a higher content of these compounds during the storage.

Polyphenols Over the last decades, polyphenols have attracted the interest of both the scientific community and the food industry due to their health-related benefits and their potential use as food additives. Although several research papers have been

14

E.S. Silva et al.

Fig. 3 Proposed mechanism of radical scavenging activity of gallic acid

published about the effects of PEF on the total phenolic compounds (FolinCiocalteau method), the literature evaluating the PEF effects (electric field strength and treatment time) on individual polyphenols is scarce. Regarding total phenolic compounds, it seems that they are apparently resistant to the effects of PEF, although it is necessary to take into account the interferences associated with the FolinCiocalteau method, which difficult the interpretation of the results. Gallic acid (3,4,5-trihydroxybenzoic acid) is a naturally occurring polyphenol, acting as antioxidant, and it is among the most used molecules as reference to quantify polyphenols using Folin-Ciocalteu method. The antioxidant capacity of gallic acid could be associated to its structural properties. In fact, the presence of a phenolic nucleus and unsaturated side chain leads to forming a resonance stabilized phenoxy radical (Fig. 3), which permits the possible presence of one unpaired electron not only on the oxygen but also throughout the molecule. Cyanidin is another abundant polyphenol in biological systems having the function of preventing lipid peroxidation. In addition, it is considered as one of the largest and most important group of water-soluble pigments naturally occurring. The potential of cyanidin as an electron donor (that stabilizes and inactivates a free radical) from OH groups, attached to the phenolic rings, has been previously reported by Van Acker et al. (1996). During this process, the reducing agent (polyphenol) is changed to an aroxyl radical, a more stable molecule (Fig. 4). In a study conducted by Morales-de La Pena et al. (2011b), the impact of PEF processing (35 kV/cm for 800 μs and 1400 μs) on the total phenolic content of fruit juice-soymilk beverages was investigated. The authors reported some reasons in order to explain the changes observed on the phenolic content after PEF treatment of liquid foods which were: biochemical reactions during the PEF processing, leading to the formation of new phenolic compounds; significant effects on the cell membranes or in phenolic complexes with other compounds, releasing of some free phenolic compounds after PEF processing, and a possible inactivation of polyphenol oxidase after PEF treatment, preventing further loss of phenolic compounds.

Effect of Pulsed Electric Fields on Food Constituents

15

Fig. 4 Proposed mechanism of radical scavenging activity of cyanidin

In another study, Sánchez-Moreno et al. (2005) studied the effect of PEF treatment (35 kV/cm, 750 μs) on flavanones content of orange juice. The authors observed no significant changes in the total flavanones content, aglycones, hesperidin, and naringenin after PEF treatment. In another study, Odriozola-Serrano et al. (2009) indicated a higher retention of total anthocyanins in strawberry juice, after PEF treatment using bipolar mode, rather than monopolar mode. In any case, retention in PEF treated juice was higher than in the thermally treated sample. Aguilar-Rosas et al. (2007) evaluated the effect of thermal (conventional high temperature-short time) and PEF (35 kV/cm) pasteurization on the physicochemical properties and flavor compounds of apple juice samples using bipolar pulses of 4 μs and a frequency of 1200 Hz. Compared to the thermal treatment, application of PEF resulted in the lower loss of total phenolic compounds (32.2 vs. 14.49 %). Similarly,

16

E.S. Silva et al.

Odriozola-Serrano et al. (2008b), who studied the effect of high-intensity PEF treatment (35 kV/cm for 1500 μs in bipolar 4 μs pulses at 100 Hz, with an energy density of 8269 kJ/L) on tomato juice, did not observe significant changes in phenolic compounds (26.2 mg/100 g in PEF-treated sample vs. 26.5 mg/100 g in high pasteurization). Recently, Carbonell-Capella et al. (2016) evaluated the effect of apple pre-treatment by PEF processing (E = 800 V/cm) on juice extraction using freezing-assisted pressing. At fixed value of extraction yield (e.g. Y = 0.6), an increase in the total phenolic compounds (by  1.16) was reported after PEF pre-treatment. The highest concentrations of total phenolic compounds (CTPC 1310.4 mg GAE/L) obtained with juice yields of 18 %. In another study, the impact of PEF processing on the polyphenol profile of tomato juices was evaluated (Vallverdu-Queralt et al. 2012). First, tomatoes were PEF-treated at the moderate intensity and refrigerated at 4  C for 24 h. Then, treated and untreated samples were subjected to PEF treatment with high-intensity or thermal treatment (90  C for 60 s). Compared to the control, tomatoes subjected to PEF-treatment at the moderate intensity provided juices with higher content of polyphenol compounds. Over storage time, a slight decrease in the polyphenols content was observed in thermaland PEF-treated juices, except for caffeic acid. However, compared to the thermal treated juices, PEF-processed tomato samples had higher content of polyphenol compounds (ferulic acid, chlorogenic acid, p-coumaric acid, caffeic-O-glucoside acid, rutin, and naringenin) following the treatment and during the storage. Some studies reported the potential of PEF to induce the polarization of water molecules, thus facilitating its dissociation into ions at high field strength with the subsequent potential formation of free radicals. Although the potential of PEF to form radical species is a fact, there is a lack of studies in the available literature evaluating this phenomenon. For instance, several researchers have evaluated the formation of free radicals after PEF using electron spin resonance (Zhang et al. 2011; Zhao et al. 2011). The generation of reactive species (radicals) after applying PEF can be seen in a positive or negative way depending on the application intended to give to this technology. For instance, apart from cell electroporation phenomena, the inactivation of microorganisms in food systems is attributed to the formation of highly reactive free radicals from chemical species when PEF is applied (Sitzmann 1995). Moreover, PEF can improve the antioxidant properties of some antioxidant compounds due to radical formation. However, PEF radicals could induce the degradation of some polyphenols (e.g., anthocyanin). Zhang et al. (2008) evaluated the effects of PEF processing on purified cyanindin-3-glucoside from red raspberry. Equipment parameters included exponentially decaying wave, 300 μs pulse duration (pulse width), 1 Hz pulse frequency, 0.5 μF capacitor, 1.2, 2.2 and 3.0 kV/cm field intensities, and 300 pulses. The authors demonstrated the formation of chalcone after the treatment, which was considered as the first step for degradation of anthocyanins together with opening of the pyrilium ring. A possible explanation for this fact could be an increase in the concentration of H2O2 or hydroxyl radical, which was shown to enhance the anthocyanin degradation of litchi fruit (Ruenroengklin et al. 2009).

Effect of Pulsed Electric Fields on Food Constituents

17

Conclusion Several studies clearly showed the application of pulsed electric fields (PEF) processing to improve the extractability of valuable bioactive compounds (e.g., amino acids, vitamins, fatty acids, and phenolic compounds) from different food matrices or even induce the generation of certain compounds when treating metabolically active tissues. On the contrary, in some studies, significant effects of PEF treatment on reducing the fatty acid contents, modification of free amino acids (FAA), formation of new phenolic compounds, and reduction of vitamins have been observed. It should be noted that the efficiency of this technology is dependent not only on PEF processing parameters, but also on the treatments before and after PEF and on the product characteristics. Thus, further research works have to be conducted to study the kinetics of generation, retention, modification, or degradation of valuable compounds influenced by PEF treatment as well as to clarify the mechanistic insight of the induced changes.

References Aguilar-Rosas SF, Ballinas-Casarrubias ML, Nevarez-Moorillon GV, Martin-Belloso O, OrtegaRivas E (2007) Thermal and pulsed electric fields pasteurization of apple juice: effects on physicochemical properties and flavour compounds. J Food Eng 83:41–46 Barba FJ, Parniakov O, Pereira SA, Wiktor A, Grimi N, Boussetta N, Saraiva JA, Raso J, MartinBelloso O, Witrowa-Rajchert D, Lebovka N, Vorobiev E (2015) Current applications and new opportunities for the use of pulsed electric fields in food science and industry. Food Res Int 77 (4):773–798 Bendicho S, Espachs A, Arántegui J, Martín O (2002) Effect of high intensity pulsed electric fields and heat treatments on vitamins of milk. J Dairy Res 69:113–123 Bi X, Liu F, Rao L, Li J, Liu B, Liao X, Wu J (2013) Effects of electric field strength and pulse rise time on physicochemical and sensory properties of apple juice by pulsed electric field. Innovative Food Sci Emerg Technol 17:85–92 Carbonell-Capella JM, Parniakov O, Barba FJ, Grimi N, Bals O, Lebovka N, Vorobiev E (2016) “Ice” juice from apples obtained by pressing at subzero temperatures of apples pretreated by pulsed electric fields. Innovative Food Sci Emerg Technol 33:187–194 Cserhalmi Z, Sass-Kiss A, Toth-Markus M, Lechner N (2006) Study of pulsed electric field treated citrus juices. Innovative Food Sci Emerg Technol 7:49–54 Elez-Martínez P, Martín-Belloso O (2007) Effects of high intensity pulsed electric field processing conditions on vitamin C and antioxidant capacity of orange juice and gazpacho, a cold vegetable soup. Food Chem 102:201–209 Eonseon J, Polle JW, Lee HK, Hyun SM, Chang MJ (2003) Xanthophylls in microalgae: from biosynthesis to biotechnological mass production and application. Microb Biotechnol 13:165–174 Garde-Cerdán T, Arias-Gil M, Marsellés-Fontanet AR, Ancín-Azpilicueta C, Martín-Belloso O (2007) Effects of thermal and non-thermal processing treatments on fatty acids and free amino acids of grape juice. Food Control 18:473–479 Koubaa M, Roselló-Soto E, Šic Žlabur J, Režek Jambrak A, Brnčić M, Grimi N, Boussetta N, Barba FJ (2015) Current and new insights in the sustainable and green recovery of nutritionally valuable compounds from Stevia rebaudiana Bertoni. J Agric Food Chem 63:6835–6846 Liu YY, Zeng XA, Deng Z, Yu SJ, Yamasaki S (2011) Effect of pulsed electric field on the secondary structure and thermal properties of soy protein isolate. Eur Food Res Technol 233:841–850

18

E.S. Silva et al.

Min S, Jin ZT, Zhang QH (2003) Commercial scale pulsed electric field processing of tomato juice. J Agric Food Chem 51:3338–3344 Morales-de la Pena M, Salvia-Trujillo L, Rojas-Grau MA, Martin-Belloso O (2011a) Impact of high intensity pulsed electric fields or heat treatments on the fatty acid and mineral profiles of a fruit juice-soymilk beverage during storage. Food Control 22:1975–1983 Morales-de La Pena M, Salvia-Trujillo L, Rojas-Grau MA, Martin-Belloso O (2011b) Changes on phenolic and carotenoid composition of high intensity pulsed electric field and thermally treated fruit juice-soymilk beverages during refrigerated storage. Food Chem 129:982–990 Odriozola-Serrano I, Soliva-Fortuny R, Gimeno-Añó V, Martín-Belloso O (2008a) Modeling changes in health-related compounds of tomato juice treated by high-intensity pulsed electric fields. J Food Eng 89:210–216 Odriozola-Serrano I, Soliva-Fortuny R, Martín-Belloso O (2008b) Changes of health-related compounds throughout cold storage of tomato juice stabilized by thermal or high intensity pulsed electric field treatments. Innovative Food Sci Emerg Technol 9:272–279 Odriozola-Serrano I, Soliva-Fortuny R, Hernandez-Jover T, Martin-Belloso O (2009) Carotenoid and phenolic profile of tomato juices processed by high intensity pulsed electric fields compared with conventional thermal treatments. Food Chem 112:258–266 Odriozola-Serrano I, Garde-Cerdan T, Soliva-Fortuny R, Martin-Belloso O (2013) Differences in free amino acid profile of non-thermally treated tomato and strawberry juices. J Food Compos Anal 32:51–58 Oms-Oliu G, Odriozola-Serrano I, Soliva-Fortuny R, Martín-Belloso O (2009) Effects of highintensity pulsed electric field processing conditions on lycopene, vitamin C and antioxidant capacity of watermelon juice. Food Chem 115:1312–1319 Retsky KL, Freeman MW, Frei B (1993) Ascorbic acid oxidation product(s) protect human low density lipoprotein against atherogenic modification. Anti- rather than prooxidant activity of vitamin C in the presence of transition metal ions. J Biol Chem 268:1304–1309 Rivas A, Rodrigo D, Company B, Sampedro F, Rodrigo M (2007) Effects of pulsed electric fields on water-soluble vitamins and ACE inhibitory peptides added to a mixed orange juice and milk beverage. Food Chem 104:1550–1559 Roohinejad S, Everett DW, Oey I (2014) Effect of pulsed electric field processing on carotenoid extractability of carrot purée. Int J Food Sci Technol 49:2120–2127 Ruenroengklin N, Yang B, Lin H, Chen F, Jiang Y (2009) Degradation of anthocyanin from litchi fruit pericarp by H2O2 and hydroxyl radical. Food Chem 116:995–998 Sánchez-Moreno C, Plaza L, Elez-Martínez P, De Ancos B, Martín-Belloso O, Cano MP (2005) Impact of high pressure and pulsed electric fields on bioactive compounds and antioxidant activity of orange juice in comparison with traditional thermal processing. J Agric Food Chem 53:4403–4409 Sitzmann V (1995) High voltage pulse techniques for food preservation. GW. Gould. New methods for food preservation. London, UK. Blackie Academic and Professional 236–252 Torregrosa F, Cortes C, Esteve MJ, Frigola A (2005) Effect of high-intensity pulsed electric fields processing and conventional heat treatment on orange-carrot juice carotenoids. J Agric Food Chem 53:9519–9525 Vallverdu-Queralt A, Odriozola-Serrano I, Oms-Oliu G, Lamuela-Raventos RM, Elez-Martinez P, Martin-Belloso O (2012) Changes in the polyphenol profile of tomato juices processed by pulsed electric fields. J Agric Food Chem 60:9667–9672 Vallverdu-Queralt A, Odriozola-Serrano I, Oms-Oliu G, Lamuela-Raventos RM, Elez-Martinez P, Martin-Belloso O (2013) Impact of high-intensity pulsed electric fields on carotenoids profile of tomato juice made of moderate-intensity pulsed electric field-treated tomatoes. Food Chem 141:3131–3138 Van Acker SABE, Van Den Berg D-J, Tromp MNJL, Griffioen DH, Van Bennekom WP, Van Der Vijgh WJF, Bast A (1996) Structural aspects of antioxidant activity of flavonoids. Free Radic Biol Med 20:331–342

Effect of Pulsed Electric Fields on Food Constituents

19

Zhang Y, Hu XS, Chen F, Wu JH, Liao XJ, Wang ZF (2008) Stability and colour characteristics of PEF-treated cyanidin-3-glucoside during storage. Food Chem 106:669–676 Zhang S, Yang R, Zhao W, Liang Q, Zhang Z (2011) The first ESR observation of radical species generated under pulsed electric fields processing. LWT- Food Sci Technol 44:1233–1235 Zhang Z-H, Zeng X-A, Brennan CS, Brennan M, Han Z, Xiong X-Y (2015) Effects of pulsed electric fields (PEF) on vitamin C and its antioxidant properties. Int J Mol Sci 16:24159 Zhao W, Yang R, Wang M, Lu R (2009) Effects of pulsed electric fields on bioactive components, colour and flavour of green tea infusions. Int J Food Sci Technol 44:312–321 Zhao W, Yang R, Shi X, Pan K, Zhang S, Zhang W, Hua X (2011) Oxidation of oleic acid under pulsed electric field processing. Food Res Int 44:1463–1467 Zulueta A, Esteve MJ, Frasquet I, Frígola A (2007) Fatty acid profile changes during orange juicemilk beverage processing by high-pulsed electric field. Eur J Lipid Sci Technol 109:25–31 Zulueta A, Barba FJ, Esteve MJ, Frígola A (2010) Effects on the carotenoid pattern and vitamin A of a pulsed electric field-treated orange juice-milk beverage and behavior during storage. Eur Food Res Technol 231:525–534

Pulsed Electric Fields in Combination with Vacuum Impregnation for Improving Freezing Tolerance of Vegetables Federico Gómez Galindo and Katarzyna Dymek

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroporation and Cell Survival . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Influence of Electroporation Protocols on Reversible Electroporation and Survival . . . . . . . . Electrical Properties of Leaf Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Factors Affecting the Electroporation of the Heterogeneous Structures of Leaves . . . . . . . . . Vacuum Impregnation as Strategy to Replace the Air with a Solution of Known Electrical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrical Properties of Vacuum Impregnated Leaves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Metabolic Consequences of Combining Vacuum Impregnation and Pulsed Electric Field . Freezing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Influence of VI and PEF on the Freezing Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 4 6 6 7 8 8 10 12 14 16 16

Abstract

Freezing is a widely used method of preserving food products. Efforts are currently being directed towards improving the quality of sensitive tissues of plant foods such as leaves, after freezing and thawing. One of the methods under investigation is the combination of vacuum impregnation (VI) with cryoprotectants and the application of a pulsed electric field (PEF) to the plant tissue prior to freezing. In this chapter were identify mechanisms for the efficient introduction of a cryoprotectant molecule into the heterogeneous structure of leaf tissue and improve our understanding of the consequences of the introduction of this foreign F.G. Galindo (*) Food Technology, Engineering and Nutrition, Lund University, Lund, Sweden e-mail: [email protected] K. Dymek (*) Optifreeze AB, Lund, Sweden e-mail: [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_32-1

1

2

F.G. Galindo and K. Dymek

molecule into the tissue regarding cell metabolism, freezing point, and ice propagation rate. To obtain precise information on the electroporation of internally located cells, a three-dimensional numerical model of the cross section of a leaf was developed. Validation of the models showed the importance of the wax layer and stomata for the successful electroporation of all cells in the tissue. VI, and the subsequent application of PEF, increased the metabolic activity of the tissue. The increase in metabolic activity after VI was accompanied by the accumulation of trehalose-6-phosphate in the cells. Leaves impregnated with trehalose, sucrose, glucose, and mannitol exhibited significantly lower ice propagation rates and higher freezing temperatures than untreated controls. Leaves subjected to PEF also showed higher freezing temperatures than untreated leaves; however, the ice propagation rate was not influenced by PEF. Keywords

Reversible permeabilization • Cryoprotectants • Leaves • Heterogeneous structure • PEF-induced metabolic responses • Freezing point • Ice propagation rate

Introduction The shelf life of many food products can be extended by freezing as this inhibits microbial growth and slows down other degradation processes. The main disadvantage of freezing is that it can lead to tissue damage due to the formation of ice crystals; the cell membranes are ruptured, and the product becomes soggy after thawing. The freezing and thawing techniques used today are successful for products with low moisture content and products with relatively strong cell walls, for example, green peas, carrots, maize, and peppers. Products with weak cell walls, e.g., leafy vegetables such as spinach or soft fruits such as strawberries, become soggy upon freezing and thawing, they exude liquid, and lose much of their consumer appeal. Efforts are being made to improve methods of freezing in order to widen the application to products that presently cannot be frozen and where material losses in the food chain are high. An innovative method for improving the freezing tolerance of plant tissues has been developed by Dejmek et al. (2013). The method consists of two kinds of pretreatment applied to the plant tissue prior to freezing: vacuum impregnation (VI) and a pulsed electric field (PEF). VI allows the introduction of a cryoprotectant molecule (trehalose) into the extracellular space of the tissue, while the subsequent application of PEF facilitates the distribution of the cryoprotectant in the intracellular space (Phoon et al. 2008). Remarkably, with the combination of these unit operations, cell viability is preserved after a freezing-thawing cycle allowing cells to keep their turgor and, in consequence, avoid tissue collapse after thawing (Fig. 1). The work described in this chapter was carried out on rucola and spinach leaves, which lose their structure upon freezing and thawing due to loss of turgor pressure, plasma membrane damage, and cytorrhysis, i.e., the collapse of the cell wall around

Pulsed Electric Fields in Combination with Vacuum Impregnation for Improving. . .

3

Fig. 1 Spinach leaves treated with a freeze-thaw cycle. (a) The leaf cells were infused with the cryoprotectant trehalose before the freezing-thawing cycle. PEF treatment was used for the delivery of the cryoprotectant into the cytoplasm of the cells. (b) Control sample that did not undergo PEF-assisted infusion of the cryoprotectant before the freezing-thawing cycle

the shrinking protoplasm. Market studies have indicated that the texture of frozen fruits and vegetables is one of the most important parameters determining the quality of frozen/thawed plant food products (Cano 1996). Therefore, improving the texture of sensitive, frozen and thawed plant tissues is of considerable commercial importance. The impregnation of leaves with a cryoprotectant through VI and PEF is associated with several technological challenges due not only to the heterogeneity and complexity of the tissue structure but also to the expected changes in the properties of the tissue when introducing a foreign molecule. Furthermore, evidence has recently been presented that common cryoprotective substances such as sugars affect the metabolic activity of the cells in the period shortly after impregnation (Gómez Galindo and Yusof 2014). The effects of the cryoprotectant on ice formation and propagation, and thus cell damage, upon freezing are largely unknown. Recent advances in the quest for the knowledge required to address these challenges are described in the following sections. In this chapter, VI and PEF are studied separately and in combination for the efficient introduction of a cryoprotectant molecule into the heterogeneous structure of leaf tissue, improving our understanding of the consequences of the introduction of this foreign molecule into the tissue regarding cell metabolism, freezing point, and ice propagation rate.

Electroporation and Cell Survival The first challenges for the application of PEF for improving freezing tolerance of vegetables are to (i) achieve the permeabilization of all kind of cells present in the tissue, which would differ in size and shape, and (ii) achieve cell survival after the application of PEF.

4

F.G. Galindo and K. Dymek

Depending on the parameters, electroporation may be reversible, allowing pore resealing and membrane recovery, or irreversible leading to the death of the cell (Kandušer and Miklavčič 2008). Cells may lose their viability after electroporation due to damage to the plasma membrane or biochemical imbalance resulting from pores in the cell membrane, from which the cell is not able to recover. Electroporation is controlled by a number of parameters such as the electric field strength, pulse duration, the number of pulses, the interval between pulses, the shape of the pulse, and its polarity. For reversible electroporation, the electroporation parameters must be carefully selected to preserve the viability of the cells. Reversible electroporation occurs when pores are created temporarily in the cell membrane, meaning that they reseal and the membrane recovers, and consequently the viability of the cell is preserved. Pore resealing is a relatively long process (often > 1 s) compared to the process of pore creation (milliseconds or microseconds, depending on the duration of the pulse) (Kotnik et al. 2012).

Influence of Electroporation Protocols on Reversible Electroporation and Survival Using electrical resistance measurements, it has been demonstrated that at a certain applied voltage, the electrical resistance of the tissue rapidly decreases with increased number of applied pulses (Dymek et al. 2014). As illustrated in Fig. 2, the decrease of survival of the leaves was mostly associated with the portion of the curve where the electrical resistance does not change further (higher number of pulses). Two separate mechanisms may be responsible for the decrease in survival level: (i) permanent damage to cell membrane (no resealing) which would prevent the buildup of cell potential difference and thereby life or (ii) metabolic consequences provoked by electric pulses, probably occurring after the resealing of pores. Complex metabolic responses of cells exposed to external electric field are not well understood (Gómez Galindo et al. 2008). After resealing, the recovery of cells depends on their ability to achieve biochemical balance after PEF-induced intense influx and efflux of ions and molecules have taken place. If that balance is not attained, cell death may occur (Weaver 2000). Other factors that may contribute to a reduced survival rate are the production of reactive oxygen species and/or the release of intracellular calcium from mitochondria (Buescher and Schoenbach 2003). The survival of the samples may be influenced by the properties of the treated leaf even if leaves were grown under the same conditions in the field. Leaves within the same plant may be at different development stages and therefore have variations in thickness and fine structure of certain tissues, i.e., the cuticle (Riederer and Schönherr 1988). Consequently, they may respond differently in terms of electroporation and survival since sensitivity to electroporation depends inter alia on the physiological state of the cells (Hapala 1997). The air spaces expand during the development of the leaf, and since air is nonconductive, it may influence the general conductivity, the response of the leaf upon application of PEF, and the capacity of the

Pulsed Electric Fields in Combination with Vacuum Impregnation for Improving. . .

5

Fig. 2 Influence of pulse polarity, pulse length, and the interval between pulses on the electrical resistance and cell survival in rucola leaves. The left hand column (a–d) shows the results for 250 μs monopolar (MP) pulses and bipolar (BP) pulses of 125 and 250 μs, with 1 ms intervals between the pulses. The right-hand column (e–h) shows the results when varying the interval between 125 μs BP pulses. Tissue survival was tested with fluorescent microscopy, using fluorescein diacetate (FDA; Sigma Aldrich, USA, λex = 492 nm, λem = 517 nm), as described in Phoon et al. (2008) (Adapted from Figs. 4 and 5 of Dymek et al. 2014, with kind permission from Springer)

leaf to recover after stress. However, it was observed that samples either fully survived the treatment or were fully killed by the same applied parameters. It is premature to speculate about pathways that might be involved in the death or survival of cells in the tissue at certain levels of PEF-induced damage. However, the “all or none” nature of the phenomenon, together with the fact that survival did

6

F.G. Galindo and K. Dymek

not correlate with resistance, suggests a more subtle mechanism than cell resealing, and this is an attractive subject for further research in the field to improve understanding of survival mechanisms of plant tissue upon PEF treatment. The recovery of the cell membrane after electroporation is an important factor when investigating the influence of intervals between pulses on electroporation and survival (Chalermchat 2005). It was observed that for longer intervals between pulses, the survival rate is higher (Fig. 2), suggesting that the longer time between pulses (2 ms) had a beneficial influence on the recovery of cells after electroporation. The distance between pulses may determine the way cells would be affected by a series of pulses (Asavasanti et al. 2011). The pulses could be perceived as separate events or as a long pulse. The dynamics of pores opening and resealing, both during the pulse and the interval, may influence the general recovery and survival after electroporation.

Electrical Properties of Leaf Tissues Biological tissue shows a combination of conducting and insulating properties. When an external electric field is applied to a conductor, electric charges move freely, while in an insulator there is no movement of charges and no electric current is conducted. The electrical properties of biological tissue can thus be described by the conductivity (σ), which is defined as the ability to transport charges, and the permittivity (ε), which is the capacity to trap charges (Miklavčič et al. 2006). When an external electric field is applied to a plant tissue, its response depends on the frequency of the applied field. At low frequencies, the plasma membrane of the cell acts as an insulator with a capacitance and a high resistance and, as the current cannot be conducted through the interior of the cell, it is forced to flow around the cell. Since the entire cell surrounded by the plasma membrane exhibits insulating properties, the overall conductivity of the tissue is decreased. At higher frequencies, the lower capacitance of the plasma membrane allows the current to enter the cell. With increasing frequency, the plasma membrane becomes more “transparent” to the electric field, which passes through the conductive cell interior and, in contrast to low frequencies, the overall conductivity of the tissue increases. The permittivity of the cell membrane shows the opposite behaviour; it decreases with increasing frequency. At low frequencies, the cell is polarized, meaning that the charges are separated at two poles in the cell interior, and thus the cell has a high permittivity. At higher frequencies, the ion separation is less distinguishable, leading to lower permittivity. At low frequencies or under direct current (DC) fields, the current flows almost entirely through the continuous system of the apoplast (Zimmermann 1982).

Factors Affecting the Electroporation of the Heterogeneous Structures of Leaves The internal structure of the leaf lamina is differentiated into easily distinguishable layers, as shown in Fig. 3.

Pulsed Electric Fields in Combination with Vacuum Impregnation for Improving. . .

7

Fig. 3 Microscope image of a cross section of a spinach leaf. UE upper epidermis, PM palisade mesophyll, SM spongy mesophyll, AS air spaces, LE lower epidermis. The image was obtained by staining the leaf for 30 min in 12 μM fluorescein diacetate solution and imaging in a Nikon upright microscope

Using mathematical modeling, Dymek et al. 2015a separated the specific elements of this tissue structure such as cell size, cell arrangement, cuticular wax layer, and stomata and the effects of the connections between cells and the air fraction in the tissue on the creation of pores when PEF was applied under specific conditions. The model was first analyzed in the frequency domain, where alternating voltage and current signal at frequencies from 20 Hz to 1 MHz were used to measure conductivity in the tissue and validate the model. Using this methodology, it was possible to establish important facts regarding the effect of PEF on the leaf. The general conductivity of the leaf is significantly determined by the low conductive wax layer. Since stomata are “holes” in the wax layer, they play a crucial role in the current flow throughout the leaf cross section. Another important factor to consider is the air in the tissue. The average amount of air in the tissue is known, but the location and the size of the air spaces are not uniform in this highly irregular structure. A strategy to avoid the influence of air in the structure on permeabilization as well as have more controlled electrical properties of the extracellular space is replacing the air with a solution of known properties. This can be achieved by vacuum impregnation.

Vacuum Impregnation as Strategy to Replace the Air with a Solution of Known Electrical Properties Vacuum impregnation (VI) is a technique with which external liquids can be introduced into porous structures such as plant tissues. The porous material is immersed in the liquid and subjected to a two-phase pressure change. In the first phase the pressure is decreased, leading to expansion of the gas in the pores of the material and partial flow of the gas out of the pores until mechanical equilibrium is

8

F.G. Galindo and K. Dymek

established. In the second phase, atmospheric pressure is restored causing the compression of the residual gas and the inflow of the liquid into the pores of the extracellular space (Tylewicz et al. 2012). The liquid phase infused into the material may have different properties, and VI can be thus used to change the composition of plant tissues. VI has been shown to improve the quality of frozen-thawed fruits and vegetables by incorporating stabilizers such as pectin or zinc and calcium salts into the structure of the product prior to freezing (Radziejewska-Kubzdela et al. 2014). Impregnation using cryoprotectants, such as antifreeze proteins and sugars such trehalose (Velickova et al. 2013), has been reported to have a positive effect on the quality of the food product after freezing and thawing. During VI, the air inside the tissue, which has high electrical resistance, is replace by a liquid, which has lower electrical resistance, suggesting that the electrical properties of the entire tissue may be changed by VI.

Electrical Properties of Vacuum Impregnated Leaves As mentioned above, VI changes the composition and thus the electrical properties of plant tissue, which may also influence the electrical field distribution within the tissue during electroporation. The electroporation of vacuum impregnated spinach tissue was investigated theoretically, by developing a numerical model of the cross section of the leaf, and experimentally, by measuring the electric current during the pulse at applied voltages from 50 to 500 V (Dymek et al. 2015a). It was concluded that the impregnating solution considerably contributes to the conductive pathways through the leaf cross section. After VI, the current will be markedly affected if stomata are filled with air/gas, cell sap, or impregnated solution. In the impregnated leaves, pores start to be created in the membranes of epidermal cells at low voltages (50 V) and with higher voltages the pore density becomes uniform in the entire leaf cross section (Fig. 4) (Dymek et al. 2015a)

Metabolic Consequences of Combining Vacuum Impregnation and Pulsed Electric Field Substances commonly used for impregnation, such as sugars, affect the metabolic activity of the cells in the tissue on short time scales after impregnation (Panarese et al. 2014). Trehalose, a nonreducing disaccharide, and its metabolic precursor, trehalose-6-phosphate (T6P) are accumulated in some plants as a result of abiotic stress. Moreover, exogenous trehalose has also been shown to act as an elicitor of genes involved in abiotic stress responses. It has been reported that in plants T6P acts as an inhibitor of SnRK1 (sucrose nonfermenting-related kinase 1), a central metabolic regulator of the expression of genes related to energy and carbon supply (Schluepmann et al. 2012).

Pulsed Electric Fields in Combination with Vacuum Impregnation for Improving. . .

9

Fig. 4 Distribution of pore density in the electroporated cells investigated with the model of vacuum impregnated spinach leaf after applying electric pulses with different amplitudes. Models in which the stomata were assumed to be filled with either air or cell sap are compared. The color scale is logarithmic; dark blue representing low pore density, and deep red representing high pore density. Pore density is expressed as pores/m2 (Corresponding to Fig. 8 in Dymek et al. 2015a, with kind permission from Elsevier)

10

F.G. Galindo and K. Dymek

Fig. 5 Concentration of trehalose-6-phosphate determined in untreated spinach leaves, leaves subjected to VI only, and a combination of VI and PEF. The error bars represent the standard deviation of four replicates (Corresponding to Fig. 5 in Dymek et al. 2016, with kind permission from Elsevier)

Figure 5 shows that VI with trehalose causes significant accumulation of T6P in the leaf’s cells. The trehalose may enter the cells through nonspecific mannose transporters (Schluepmann et al. 2012) leading to the measured accumulation of T6P. The application of PEF did not further increase the concentration of T6P. These results indicate that either trehalose is not further accumulated in the cytosol as a result of PEF or that any further increase in trehalose does not lead to further accumulation of T6P after that induced by vacuum impregnation. The metabolic activity of spinach leaves was increased by VI with trehalose and PEF (Fig. 6). The accumulation of T6P and increased metabolic activity resulting from VI might be linked, as T6P inhibits the SnRK1 activity, increasing carbon utilization and thus metabolic activity (Schluepmann et al. 2012). PEF treatment further increased the metabolic activity of the impregnated leaf samples (Fig. 6) but was not accompanied by an increase in T6P concentration. Cell membrane electroporation and resealing involve much more than structural changes in the lipid matrix; this is a complex metabolic process that may involve several stages such as energy release from the movement of ionic species, the hydrolysis of adenosine triphosphate (ATP) to rebuild charge gradients across cell membranes, and/or other physiological events taking place after electroporation and long after resealing (Gómez Galindo et al. 2008).

Freezing Freezing is a well established and widely used method in the food industry for the preservation of fruits and vegetables. Freezing is considered one of the most important methods of maintaining relatively good quality of the food product during

Pulsed Electric Fields in Combination with Vacuum Impregnation for Improving. . .

11

Fig. 6 Typical calorimetric results for the thermal power generated by untreated control spinach samples, spinach samples subjected to VI only, and to VI and PEF. The error bars represent the standard deviation of three replicates at 2000 s, 8900 s, and 15,000 s, which are the first measurement points after treatment. Letters above the error bars indicate statistical differences ( p < 0.05) (Corresponding to Fig. 4 in Dymek et al. 2016, with kind permission from Elsevier)

long-term storage. However, the final quality and appearance of the frozen-thawed product are affected. Natural enzymes in the plant tissues cause changes in the color and flavor. Enzymatic browning occurs during frozen storage in products such as potatoes and mushrooms (Cano 1996). Texture is an important attribute determining the quality of the frozen-thawed product. The texture of plant tissues depends on the arrangement of the cell walls, the intracellular spaces, and the integrity of plasma membranes. Freezing damages the cells irreversibly, leading to rupture of the plasma membranes and extensive leakage of the cell contents, especially in sensitive products such as leaves (Gómez Galindo and Sjöholm 2004). Freezing starts with the nucleation process, which is the merging of water molecules to form a stable ice nucleus. This process may be homogenous, as occurs in pure systems without any nucleating agents, where the molecules spontaneously form a stable interface, or heterogeneous, in which case it is triggered by a foreign molecule, where the water aggregates assemble on a nucleating agent. Nucleation is followed by ice crystal propagation. During slow freezing, ice crystals are first formed in the extracellular spaces in the plant tissue. As the temperature is further reduced, more water is frozen in the extracellular spaces, and the concentration of the solutes consequently increases. Water frozen externally usually results in large ice crystals. An osmotic gradient is created between the cell interior and the extracellular spaces. This gradient causes the movement of water molecules towards the higher solute concentration, i.e., from the interior of the cell through the plasma membrane to the cell exterior. If the tissue continues to be exposed to low temperatures, the cell will dehydrate, the cell walls may tear, and the plasma membrane may rupture. During fast freezing, heat is removed rapidly, and there is no migration of water molecules, hence small ice crystals are formed, uniformly distributed in the intra- and extracellular spaces. The movement of water molecules is controlled by the permeability of the membrane; if

12

F.G. Galindo and K. Dymek

the permeability is high the cell may dehydrate regardless of the freezing rate (Cano 1996). The plasma membrane plays a key role in cell death due to freezing. Therefore, protection of the plasma membrane against dehydration caused by freezing is an important factor in improving the freezing tolerance of plants. This may be achieved by the accumulation of solutes inside the cell stabilizing the membrane either by direct interaction with the membrane or by interactions with surrounding water molecules. Sugars, proline, and beatine have been found to be such membraneprotective solutes (Pearce 2001).

Influence of VI and PEF on the Freezing Process Sugars can act as cryoprotectants in plant tissues through several mechanisms. They can protect cells from freezing damage by supporting the plasma membrane stability by replacing water molecules at the phospholipid interface during dehydration, by acting as cryoprotectants for specific enzymes, or by functioning as osmolytes preventing freezing-induced cell dehydration (Steponkus 1984). In nature, plants accumulate sugars in their cells in response to cold (Pearce 2001). Sugars introduced into the extracellular spaces of the plant tissue artificially using VI may be present in the intracellular space after the application of PEF. High-resolution infrared thermography has been used to visualize the freezing process in order to study the ice propagation rate and the freezing temperature of leaves. The results of using VI with water, sugars, or mannitol in spinach leaves are shown in Fig. 7a, where it can be seen that VI raised the freezing temperature of the leaves, i.e., they froze at a higher temperature than untreated leaves. Trehalose had the greatest effect on the freezing temperature of all the compounds tested. The results obtained when applying PEF to the leaves are shown in Fig. 7b. The application of PEF to the vacuum impregnated tissues did not change the freezing temperatures of leaves impregnated with sugars (Fig. 7a compared to Fig. 7b). However, the leaves subjected to PEF only and to the combination of VI and PEF prior to exposure to freezing froze at higher temperatures than the untreated leaves. Although the impact of sugars is clearly demonstrated, the results are contrary to the expected effect of solutes on freezing point depression and the promotion of supercooling in plant tissues. These results clearly indicate that the presence of sugars in large amounts of water surrounding the cells and filling the intercellular spaces has a different effect than when sugars and water are distributed unevenly and in small amounts in untreated tissues. Another parameter that affects the speed of the freezing process is the ice propagation rate. Low ice propagation rates have been suggested to be beneficial for survival of the tissue upon freezing. Figure 8 shows typical infrared photographs of spinach leaves, showing considerable differences in the ice propagation rate between an untreated leaf and a trehalose impregnated leaf. The ice propagation rate was also compared in leaves subjected to VI with different sugars and leaves treated with the combination of VI and PEF. The results are shown in Fig. 9.

Pulsed Electric Fields in Combination with Vacuum Impregnation for Improving. . .

13

Fig. 7 Freezing temperature of spinach leaves. (a) Leaves were either untreated or vacuum impregnated (VI) with water or isotonic solutions of trehalose, mannitol, sucrose, or glucose. The bars show the means  1 SD (n = 6). Different letters below the error bars indicate significant differences between the means ( p < 0.05). (b) Leaves were either untreated, treated with a pulsed electric field (PEF) only, or vacuum impregnated (VI) with water or isotonic solutions of trehalose, mannitol, sucrose, or glucose and PEF. The bars show the means  1 SD (n = 6). Different letters below the error bars indicate a significant difference between the means ( p < 0.05) (Adapted from Figs. 1 and 2 in Dymek et al. 2015b, with kind permission from Elsevier)

Impregnation with sugars drastically reduced the rate of ice propagation compared to untreated leaves and leaves impregnated with water (insert in Fig. 9). It can be seen from Fig. 9 that PEF only affects the ice propagation rate in leaves previously

14

F.G. Galindo and K. Dymek

Fig. 8 Typical infrared photographs of the freezing of spinach leaves. (a) Untreated leaf. (b) The leaf was vacuum impregnated with isotonic trehalose solution prior to freezing. (Corresponding to Fig. 3 in Dymek et al. 2015b, with kind permission from Elsevier)

impregnated with water and trehalose. Leaves subjected to the combination of VI with sucrose, glucose, and mannitol and PEF exhibited the same ice propagation rates as leaves subjected to VI. PEF causes the leakage of solutes, including sugars, hence the decrease in the ice propagation rate after PEF for leaves impregnated with water is explainable. The increase in ice propagation rate following PEF in the leaves impregnated with trehalose is difficult to explain. Specific sugars may cause different metabolic responses or interact to a different extent with the plasma membrane.

Conclusions The results presented in this chapter reveal the influence of the highly complex and heterogeneous structure of the tissue network of the leaves on the efficiency of VI and PEF treatment. It was also demonstrated that these processes not only modify the composition of the treated tissue but also influence the metabolism of the cells; both of which may influence tissue properties during further processing such as freezing. The most important conclusions are presented below. • Reversible electroporation of rucola leaves is influenced by parameters such as the pulse polarity, number of pulses, and the interval between pulses. Lower survival was observed with monopolar pulses when more than 25 pulses were applied. Longer intervals between bipolar pulses resulted in higher cell survival.

Pulsed Electric Fields in Combination with Vacuum Impregnation for Improving. . .

15

Fig. 9 Influence of PEF on the ice propagation rate of vacuum impregnated spinach leaves. Comparison of ice propagation rate in leaves vacuum impregnated with different solutions and leaves vacuum impregnated and treated with PEF. The bars show the mean  1 SD (n = 6). Different letters above the error bars indicate a significant difference between the means ( p < 0.05) (Corresponding to Fig. 6 in Dymek et al. 2015b, with kind permission from Elsevier)

• The electrical conductivity of the leaf in the frequency range 10 Hz to 1 MHz is not influenced by different cell sizes and cells arrangements but is predominantly determined by the low conductivity of the cuticular wax layer. • Electroporation of vacuum impregnated plant tissue is governed by the stomata. Since the stomata are “holes” in the wax layer, they play a crucial role in the flow of current through the leaf. The current is, therefore, markedly affected depending on whether the stomata are filled with air/gas or cell sap. • Under the studied conditions, pores start to be created in the membranes of epidermal cells at low voltages (50 V), and at higher voltages the pore density becomes uniform in the entire cross section of the leaf. • The metabolic activity of spinach leaves is increased by VI and the combination of VI and PEF. The accumulation of T6P in the cells might lead to an increase in carbon utilization and thus metabolic activity. Electroporating the impregnated leaves further increases their metabolic activity. This was expected due to the cells’ recovery mechanisms, although T6P might not be involved. • Vacuum impregnation and the application of a pulsed electric field influence the freezing temperature and ice propagation rate of spinach leaves. Vacuum impregnation with water and sugars raises the freezing temperature of spinach leaves compared with untreated leaves. Vacuum impregnation with sugars significantly

16

F.G. Galindo and K. Dymek

lowers the ice propagation rate compared with untreated leaves and leaves impregnated with water. PEF treatment induces a significant leakage of solutes and other cytoplasmic constituents into the extracellular spaces in the tissue, causing a significant decrease in the ice propagation rate. The impregnation of trehalose or other cryoprotectants by the combined actions of vacuum impregnation and pulsed electric field is a technology with strong potential for improving the freezing tolerance of leaves and, therefore, keep their fresh-like characteristics after freezing and thawing.

Cross-References ▶ Pulsed Electric Field Treatment for Vegetable Processing ▶ Responses of Plant Cells and Tissues to Pulsed Electric Field Treatments ▶ Stress Response of Plants, metabolite Production to Pulsed Electric Fields

References Asavasanti S, Ristenpart W, Stroeve P, Barrett DM (2011) Permeabilization of plant tissues by monopolar pulsed electric fields: effect of frequency. J Food Sci 76:E98–E111 Buescher ES, Schoenbach KH (2003) Effects of submicrosecond, high intensity pulsed electric fields on living cells – intracellular electromanipulation. IEEE Trans Dielectr Electr Insul 10:788–794 Cano P (1996) Vegetables. In: Jeremiah LE (ed) Freezing effects on food quality. New York: Marcel Dekker, p 247 Chalermchat Y (2005) Effects of pulsed electric fields on plant tissue. PhD thesis, Department of Food Technology, Engineering and Nutrition, Lund University Dejmek P, Gómez Galindo F, Phoon PY, Sjöholm I (2013) Method for the conservation of a plant material. Patent WO 2009/045144 Dymek K, Dejmek P, Gómez Galindo F (2014) Influence of pulsed electric field protocols on the reversible permeabilization of rucola leaves. Food Bioprocess Technol 7:761–773 Dymek K, Retelj L, Zorec B, Pavšelj N, Dejmek P, Gómez Galindo F, Miklavčič D (2015a) Modeling electroporation of the non-treated and vacuum impregnated heterogeneous tissue of spinach leaves. Innov Food Sci Emerg Technol 29:55–64 Dymek K, Dejmek P, Gómez Galindo F, Wisniewski M (2015b) Influence of vacuum impregnation and pulsed electric field on the freezing temperature and ice propagation rates of spinach leaves. LWT–Food Sci Technol 64:497–502 Dymek K, Panarese V, Herremans E, Cantre D, Schoo R, Sastre Toraño J, Schluepmann H, Wadsö L, Verboven P, Nicolai BM, Dejmek P, Gómez Galindo F (2016) Investigation of the metabolic consequences of impregnating spinach leaves with trehalose and applying a pulsed electric field. Bioelectrochem (in press) Gómez Galindo F, Sjöholm I (2004) Applying biochemical and physiological principles in the industrial freezing of vegetables: a case study on carrots. Trends Food Sci Technol 15:39–43 Gómez Galindo F, Yusof NL (2014) New insights into the dynamics of vacuum impregnation of plant tissues and its metabolic consequences. J Sci Food Agric 95:1127–1130 Gómez Galindo F, Wadsö L, Vicente A, Dejmek P (2008) Exploring metabolic responses of potato tissue induced by electric pulses. Food Biophys 3:352–360

Pulsed Electric Fields in Combination with Vacuum Impregnation for Improving. . .

17

Hapala I (1997) Breaking the barrier: methods for reversible permeabilization of cellular membranes. Crit Rev Biotechnol 17:105–122 Kotnik T, Kramar P, Pucihar G, Miklavcic D, Tarek M (2012) Cell membrane electroporation – Part 1: the phenomenon. IEEE Electr Insul Mag 28:14–23 Kandušer M, Miklavčič (2008). Electroporation of biological cell and tissue: An overview. In: Vorobiev E, Lebovka N (eds) Electrotechnologies for extraction from food plants and biomaterials. Springer Science+Business Media, pp 1–28 Miklavčič D, Pavšelj N, Hart FX (2006) Electric properties of tissues. In: Akay M (ed) Wiley encyclopedia of biomedical engineering. Wiley, New York, pp 3578–3589 Panarese V, Rocculi P, Baldi E, Wadsö L, Rasmusson AG, Gómez Galindo F (2014) Vacuum impregnation modulates the metabolic activity of spinach leaves. Innov Food Sci Emerg Technol 26:286–293 Pearce RS (2001) Plant freezing and damage. Ann Bot 87:417–424 Phoon PY, Gómez Galindo F, Vicente A, Dejmek P (2008) Pulsed electric field in combination with vacuum impregnation with trehalose improves the freezing tolerance of spinach leaves. J Food Eng 88:144–148 Radziejewska-Kubzdela E, Biegańska-Marecik R, Kidoń M (2014) Applicability of vacuum impregnation to modify physico-chemical, sensory and nutritive characteristics of plant origin products–a review. Int J Mol Sci 15:16577–16610 Riederer M, Schönherr J (1988) Development of plant cuticles: fine structure and cutin composition of Clivia miniata Reg. leaves. Planta 174:127–138 Schluepmann H, Berke L, Sanchez-Perez GF (2012) Metabolism control over growth: a case for trehalose-6-phosphate in plants. J Exp Bot 63:3379–3390 Steponkus PL (1984) Role of the plasma membrane in freezing injury and cold acclimation. Annu Rev Plant Physiol 35:543–584 Tylewicz U, Lundin P, Cocola L, Dymek K, Rocculi P, Svanberg S, Dejmek P, Gómez Galindo F (2012) Gas in scattering media absorption spectroscopy (GASMAS) detected persistent vacuum in apple tissue after vacuum impregnation. Food Biophys 7:28–34 Velickova E, Tylewicz U, Dalla Rosa M, Winkelhausen E, Kuzmanova S, Gómez Galindo F (2013) Effect of vacuum infused cryoprotectants on the freezing tolerance of strawberry tissues. LWT–Food Sci Technol 52:146–150 Weaver JC (2000) Electroporation of cells and tissues. IEEE Trans Plasma Sci 28: 24–33 Zimmermann U (1982) Electric field-mediated fusion and related electrical phenomena. Biochim Biophys Acta 694:227–277

Electroporation-Based Applications in Biotechnology Saša Haberl Meglič and Tadej Kotnik

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Genetic Transformation of Microorganisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications of Electrotransformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Parameters Affecting Electrotransformation Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inactivation of Microorganisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wastewater Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nonthermal Food and Beverage Pasteurization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Extraction of Biomolecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Unicellular Organisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multicellular Organisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biomass Drying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hardware Considerations and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 5 6 6 8 8 9 10 10 12 14 15 15 16

Abstract

Electroporation is already an established technique in several areas of medicine, but many of its biotechnological applications have only started to emerge; this chapter reviews some of the most promising ones. The introductory section provides an overview, and subsequent sections explore four types of such applications in more detail. The first application described is the most established one – the use of reversible electroporation for heritable genetic modification of microorganisms (electrotransformation); it is described how electrotransformation is used for production of biomolecules, adaptation of microorganisms to diverse conditions, and for basic research, followed by an overview of the parameters affecting the efficiency of electrotransformation. Then, the chapter reviews three S. Haberl Meglič (*) • T. Kotnik (*) Faculty of Electrical Engineering, University of Ljubljana, Ljubljana, Slovenia e-mail: [email protected]; [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_33-2

1

2

S. Haberl Meglič and T. Kotnik

classes of applications that generally aim to upscale to the industrial and/or clinical level, which have only recently started to advance to this stage. Electroporation-based inactivation of microorganisms is first described for wastewater treatment and then for nonthermal pasteurization of foods and beverages. Extraction of biomolecules by means of electroporation (electroextraction) is efficient both in unicellular and multicellular organisms, with the latter class of applications illustrated on the examples of grapes and sugar beet. Electroporation used for fast biomass drying is an emerging technology with several distinctive advantages over the standard techniques, including a much higher energy efficiency. The chapter concludes with a discussion of the main challenges, also from the hardware perspective, and of the future perspectives. Keywords

Pulsed electric field • High-efficiency transformation • Hospital wastewater • Escherichia coli • Plasmid DNA • Industrial scale • Alternative methods • Lipid extraction

Introduction While electroporation is already an established technique in several areas of medicine, many of its biotechnological applications have only started to emerge. There are four general types of such applications, exploiting electroporation either to allow exogenous molecules to enter the cells or for endogenous molecules to be extracted from within the cells (Fig. 1): genetic transformation, inactivation of microorganisms, extraction of biomolecules, and accelerated drying of biomass (Kotnik et al. 2015). In transformation based on electroporation (electrotransformation), exogenous DNA is introduced by means of reversible electroporation, the foreign genes are expressed in their new host cells, and they are inherited upon cell division; this can turn the host microorganisms into “factories” of biomolecules, adapt them to a new environment, or serve to study the role of individual genes (Kotnik et al. 2015). In electroporation-based inactivation, microorganisms are exposed to electric field pulses strong and long enough to inhibit their activity, including their division, growth, and synthesis of toxic substances. This method avoids contamination and is particularly promising in food preservation where radiation or chemicals must be avoided. Today the most widely used method for food preservation is still heating, by which both nutrients and taste are usually affected, decreasing the value of food (Haberl et al. 2013b). Since in electroporation-based inactivation only mild heating occurs, taste and/or nutrients value are not affected. In electroextraction, either microorganisms or multicellular tissues are electroporated to the extent required to release the biomolecules of interest. In some cases, it is also achievable with reversible electroporation, and, in most

Electroporation-Based Applications in Biotechnology

3

Fig. 1 Applications of electroporation in biotechnology. Reprinted from (Kotnik et al. 2015) with permission

cases, it is important to limit electroporation to levels that avoid rapid decomposition of the exposed cells and thus the formation of debris contaminating the extract. Namely, in most frequent method used for extracting substances from microorganism–mechanical disintegration–microorganism membrane is severely damaged (Fig. 2, right panel) and membrane contaminants (such as endotoxins) are released into final sample, thus purification is required. In industry, when producing large volumes, this additional purification step represents up to 80 % of the production costs. However, electroextraction, if pulse treatment conditions are adjusted, does not cause severe damage to microorganism (Fig. 2, middle panel), therefore in large scales could represent more economical method of choice. Finally, when used for facilitating water release from tissues, electroporation is useful in electroporative biomass drying, accelerating the drying process, allowing

4

S. Haberl Meglič and T. Kotnik

Fig. 2 Treatment of bacteria. Scanning electron images of Escherichia coli bacteria: (upper panel) not treated; (middle panel) subjected to electric pulses – a train of 32 pulses with pulse width of 100 μs, electric field strength of 20 kV/cm, and repetition frequency of 1 Hz was delivered; and (bottom panel) mechanically disintegrated. Scale bars represent 200 nm. Image by the author, not published previously

heating to be reduced or avoided, and often also reducing the energy requirements (Kotnik et al. 2015). Sections “Genetic Transformation of Microorganisms,” “Inactivation of Microorganisms,” “Extraction of Biomolecules,” and “Biomass Drying” focus on each of these four types of applications in more detail.

Electroporation-Based Applications in Biotechnology

5

Genetic Transformation of Microorganisms Although some microorganisms can spontaneously transform – take up foreign (heterologous) genes, express and replicate them, and pass them on upon division – the efficiency is often low, and there is ample motivation for controlled artificial transformation. Natural transformation of bacteria was first observed in 1928 in an attempt to develop a pneumonia vaccine. Since then for the purpose of genetic bacteria manipulation artificial gene transfer systems were developed (e.g., transformation using chemicals, sonoporation, freezing, and thawing), from which many could not transform all bacteria species or transformation efficiency was too low. Therefore in order to partially circumvent transformation obstacles in the mid-1980s, electrotransformation has prevailed due to its higher efficiency and applicability to the broadest range of microorganisms. Namely, besides in many very diverse bacteria, electrotransformation is also effective in Archaea, unicellular algae (microalgae), and unicellular fungi (yeasts), although perhaps with some more limitations than in bacteria. Extremophilic features of Archaea make them interesting for biotechnology applications and therefore genetic manipulation. Electroporation as a versatile technique was used for transformation of methanogenic Archaea, where electrotransformation was approximately 50- to 80-fold more efficient than natural transformation, and in hyperthermophilic Archaea. Nevertheless, electroporation is not universally applicable in Archaea. For instance, halophilic Archaea cannot tolerate NaCl concentrations below 1 M, but in suspensions of such salinity the electric field required for electroporation generates organism-damaging heating, and as a consequence, electrotransformation may not be feasible. In some other Archaea, despite attempts at optimization, researchers were unable to detect transformants. Whether these unsuccessful trials were due to biological properties of archaeal species or due to not optimal hardware and experimental setup are difficult to say. Recent studies imply that archaeal lipids with their special moieties, i.e., methyl groups in tails, ether linkages instead of ester linkages, and carbohydrates in the head groups, have much higher stability than other simple lipids studied so far. This could be the reason for unsuccessful electroporation of certain Archaea. In microalgae and yeasts, transformation efficiencies are generally lower than in bacteria and most Archaea, but still sufficient for some applications. Since applications of microalgae cover a broad spectrum, including the food and feed industries, bioenergy, cosmetics, healthcare, and environmental restoration or protection, transformation of these species is a first step in their use for biotechnological applications involving foreign protein expression or molecular modification of metabolic pathways. Numerous microalgae species were successfully electrotransformed, and transformation efficiency via the microfluidic electroporation was up to three orders of magnitude higher than bulk phase electroporation under identical conditions. Typical electric field intensity used to obtain optimal numbers of transformants of most microalgae ranges from 1 to 1.8 kV/cm for pulse durations of 2–26 ms. Highest transformation efficacy of Nannochloropsis sp., fast-growing, microalga capable of

6

S. Haberl Meglič and T. Kotnik

accumulating large amounts of oil was achieved using very high electric field strength of  11 kV/cm field strength (Kilian et al. 2011). To date, successful electrotransformation has been reported for bacteria from at least 13 of the 29 currently recognized taxonomic phyla, for Archaea from at least two of their five phyla, for microalgae from at least three of their six phyla, and for yeasts from both their phyla (Kotnik et al. 2015).

Applications of Electrotransformation Production of Biomolecules Most frequently, electrotransformation is used for synthesis of foreign substances, including antigens, cytokines, enzymes, hormones, and toxins, in host organisms ranging from bacteria to microalgae and yeasts. New transgenic “factories” of biomolecules are emerging at a formidable rate; already in 1995, 4 years after first electrotransformation-mediated transgenic protein production in yeast, a review listed 22 such proteins produced in a single yeast species, and 15 in another (Faber et al. 1995). Adaptation to Diverse Conditions Some microorganisms perform useful functions, for example, probiotics in the intestine, or bioremediators of environmental pollutants, and transformation can adapt them to new antagonists or environments. Thus, the bacterium Deinococcus radiodurans is an efficient bioremediator of uranium, converting it from soluble hexavalent into the insoluble tetravalent state. By that bacteria is preventing leakage from nuclear waste storages, but it cannot survive above 40  C, which is often exceeded in high-density storages. Electrotransfer of its genes encoding uranium chemistry into its thermophilic relative Deinococcus geothermalis produced a strain converting uranium at temperatures up to 55  C (Brim et al. 2003). Basic Research Transfer of a gene into a simpler organism can facilitate analysis of the properties and functions of the encoded protein, which may be obscured in the more complex organism from which the gene originates. If the transgenic protein interacts with host genes or their expression, electrotransformation can also be used to study the genes and proteins of the host organism. Thus, flagella of the bacterium Brachyspira hyodysenteriae consist of two types of proteins encoded by two genes, and transformation that led to inhibition of one of these two gene resulted in thinner flagella (~20 instead of ~26 nm diameter) with reduced motility (Li et al. 2000).

Parameters Affecting Electrotransformation Efficiency Efficiency of electrotransformation is strongly correlated with the extent of electroporation of the bacterial membrane and electrophoresis of DNA through the

Electroporation-Based Applications in Biotechnology

7

electroporated membrane. Thus, the parameters (field strength, duration, and number of pulses delivered) should be set at values rendering membrane permeabilization with minimal viability loss. Field strengths for optimal transformation efficiency range from 1 to 20 kV/cm, and pulse durations from 1 to 30 ms. The optimal values of these two parameters must generally be determined empirically for each particular organism and the DNA molecule to be used for transformation. In general these values are those resulting in the highest extent of electroporation without substantial loss of viability. Other factors influencing electrotransformation efficiency are discussed in subsequent paragraphs.

Organism Envelopes and Growth Phase Transformation efficiency decreases with increasing thickness and number of layers enveloping the DNA of the recipient. Thus, achievable efficiencies are highest for Gram-negative bacteria (107–1010 transformants per μg DNA), lower for Grampositive bacteria and Archaea owing to their thicker cell wall (105–107). The lowest transformation efficiency can be observed for microalgae and yeasts that have a nuclear membrane (104–107) or for organisms that have an outer polysaccharide or slime capsule layer, such as some bacteria and Archaea during particular growth phases (104). The optimum competence of bacteria is genus, strain, and species depended. Meaning that harvest time need to be optimized for each bacteria strain separately in order to obtain maximum electrotransformation efficiency. Most commonly bacteria are harvested at early or late exponential phase, in which the capsular synthesis rate decreases. For some species other effects may also contribute to better electrotransformation also in a late stationary phase, such as autolysis, which could promote the DNA uptake (Kotnik et al. 2015). Organism Size Because the transmembrane voltage induced by exposure to a given field strength is proportional to the size of the organism, the field required for electroporation and thus transformation is larger for smaller organisms. Thus, optimal field strengths are generally higher for bacteria and Archaea (5–20 kV/cm) than for microalgae and yeasts (1–12 kV/cm) (Kotnik 2016). DNA Properties DNA size strongly affects the electrotransformation efficiency, smaller DNA can more easily enter through the membrane, although also as large as 120 kb plasmid DNA can be successfully electrotransformed into bacteria. Transformation efficiency is the highest for supercoiled circular double-stranded (ds) DNA, and increasingly lower for relaxed circular dsDNA, circular single-stranded (ss) DNA, linear dsDNA with homologous ends, and linear dsDNA with nonhomologous ends. Namely, linear form of DNA is most unstable one, easily to be degraded by intracellular enzymes. For DNA concentrations from pg/ml up to μg/ml, transformation efficiency is roughly constant, implying that within this range, and under fixed experimental conditions, the transformation probability for each organism is proportional to the surrounding DNA concentration (Kotnik et al. 2015).

8

S. Haberl Meglič and T. Kotnik

Medium Properties Divalent cations (Ca2+, Mg2+) interact with DNA, and therefore they should generally be avoided. Hyperosmolarity increases the flow from the medium into the organism, and thus generally improves transformation efficiency, but to a limited extent, because substantial influx of salts has a detrimental effect on most organisms. Addition of an osmoprotectant into the medium can reduce this effect, thus further improving the transformation efficiency (Meddeb-Mouelhi et al. 2012). In yeast, electrotransformation efficiency can be improved considerably by chemical pretreatment, for example, with lithium acetate and dithiothreitol, or with thiol compounds (Suga and Hatakeyama 2001).

Inactivation of Microorganisms Microorganisms and their by-products (enzymes and toxins) in food or water can represent serious threats to human health. There is thus ample motivation for reduction of the microorganisms’ pathogenicity, and inactivation of microorganisms by means of irreversible electroporation – in wastewater treatment, as well as in food and beverage industry – is a promising approach to this aim (Haberl et al. 2013b).

Wastewater Treatment Pathogenic microbes in water have long been a concern to the public. Microbial deactivation in water can be achieved by various methods, the most common being chlorine, ozone, and ultraviolet treatments. Inactivation of microorganisms by electroporation has already been demonstrated in the 1960s and proved to be efficient for increasing the shelf life of liquid food. The use of electroporation for microbial inactivation is often termed pulsed electric field (PEF) treatment. Irreversible electroporation is suited for bacterial decontamination of hospital wastewater (Fig. 3), where it also eradicates antibiotic-resistant strains, thus limiting the spread of such bacteria into the environment, which is of general concern nowadays. Bacterial inactivation at an energy input of ~150 kJ/l can reduce the bacterial population by four orders of magnitude with wastewater temperature remaining below 70  C. At those temperature the activity of nucleases are preserved and therefore retain their ability to degrade DNA upon its release from electroporated microorganisms. By that horizontal gene transfer is prevented. Moreover, a combination of mild preheating to 60  C and subsequent electroporation has proved synergistic, leading to the reduction of the required treatment energy for efficient disinfection to ~40 kJ/l (Gusbeth et al. 2009a). This combination was also found to be effective for the inactivation of Gram-positive strains that are harder to inactivate by electroporation alone. Unlike disinfection with ultraviolet light, to which bacteria readily develop tolerance, it was shown that disinfection with electroporation does not lead to bacteria developing tolerance or resistance to the

Electroporation-Based Applications in Biotechnology

9

Fig. 3 Pilot facility for hospital wastewater disinfection at a massflow of 400 l/h. Rectangular pulses of 100 kV/cm amplitude, 1 μs duration, and 10 Hz repetition frequency are provided by a two-stage pulse-forming network generator in the left section of the facility. The power supply is situated in the middle section, while the measurement and control auxiliaries are installed in the closed right section. Photo by Martin Sack, Karlsruhe Institute of Technology, Germany, with permission

treatment for at least 30 generations (Gusbeth et al. 2009b). Upscaling to pilot-scale flow of 400 l/h demonstrated that electroporation as a disinfection technology is also efficient under high mass flow conditions. Still some problems remain to be solved, e.g., deactivating more resistant microbial species, reducing the initial investment cost, and standardizing treatment procedures.

Nonthermal Food and Beverage Pasteurization Following heat deactivation, the food is safe for consumption for a longer period of time, but it often suffers loss of flavor, color, and texture, and/or change of chemical composition. Preservation of food by electroporation maintains color and flavor, and the antioxidant levels are unaffected. The procedure was introduced more than 40 years ago, and is now a promising nonthermal food processing method, competing with ultrasound and high pressure methods. It does not generate by-products, and only mild heating occurs. Success of electroporation as a mechanism of microbial inactivation in foods strongly depends on several factors: electric field strength and duration, energy delivered, electric properties of the treated food, and microbial characteristics, including shape, size, cell wall structure and composition, and growth conditions (Mahnic-Kalamiza et al. 2014). While yeast and bacterial cells are susceptible to

10

S. Haberl Meglič and T. Kotnik

electroporation treatment, bacterial spores are much more resistant to electrical treatment. Therefore, applications of electroporation for microbial inactivation largely aim at food pasteurization rather than sterilization. During the past two decades, electroporation as a method of food preservation has found many applications, and a variety of microorganisms have effectively been inactivated in various beverages and liquid food products, such as fruit and vegetable juices, cider, beer, milk, and soups, as well as in semisolid and solid food products (Evrendilek 2016). In addition, synergistic effects between electroporation and other treatments, for example, nisin, acid, mild heating, low temperature, or high pressure, have been demonstrated (Saulis 2010). The approach of combining high pressure, ultraviolet light, and electric pulses also appears to hold promise in the inactivation of bacterial spores for which each of these mechanisms separately often fails to achieve inactivation.

Extraction of Biomolecules Unicellular Organisms Microorganisms are being recognized as a potential source of diverse biomolecules for industry, pharmacy, or medicine. Established processes to extract these biomolecules include mechanical forces or chemicals, which can be detrimental to the structure and/or integrity of extracted biomolecules. Main drawbacks of chemical methods used for obtaining intracellular compounds from bacteria are: (i) use of expensive chemicals, which are often also toxic and with pharmaceutical-scale production restricted by regulatory bodies; (ii) varying sensitivity of bacteria to these chemicals; (iii) high cost; and/or (iv) relatively long exposure needed for their effectiveness, making the process time consuming. Physical methods are effective for a broad range of microorganisms, but also they have several shortcomings: (i) extensive fragmentation of microorganisms, requiring purification of targeted biomolecules from cellular debris; (ii) nonselective extraction of biomolecules from the microorganisms; (iii) pronounced heating, which can also lead to denaturation of the extracted proteins; and/or (iv) difficulties in handling large volumes. Extraction by electroporation (also termed as electroextraction) offers a promising alternative: it is fast, chemical-free, energy-saving, easily upscalable, and with much less debris due to the limited and delayed disintegration of microorganisms (Fig. 2).

Bacteria Electroextraction of plasmid DNA (pDNA) was long assumed to be inferior in efficiency to the standard extraction method of alkaline lysis and was thus studied primarily for applications where a small yield suffices. Single-pulse electroporation was thus demonstrated as a feasible, though suboptimal method for direct pDNA

Electroporation-Based Applications in Biotechnology

11

Fig. 4 Selective release of bacteria proteins in Gram-negative bacteria. Using electroporation with appropriate electric pulse parameters, only the outer membrane of bacteria with two membranes is permeabilized. As a consequence, proteins located in the space between the two membranes are released into the media surrounding the bacteria cells, and can be collected (Ohshima et al. 2000; Ohshima and Sato 2004). Reprinted from (Haberl et al. 2013b) with permission

transfer from donor bacteria, from which DNA was electroextracted, into recipient bacteria, which were electrotransformed by it (Kotnik et al. 2015). Nevertheless, it was recently shown that, with sufficient optimization, yields with pDNA electroextraction can be comparable or even superior to alkaline lysis (Haberl et al. 2013a). Electroextraction can also be used to obtain bacterial proteins and lipids. For each type of biomolecule, selective size-specific extraction of molecules is generally achievable by optimizing the temperature and the parameters of the electric pulses (Fig. 4) and by performing the procedure in the logarithmic growth phase. The amount of extracted proteins is increasing with electric field strength and pulse duration. For gram negative bacteria electroextraction can also be selective, if pulses are adjusted. Namely, it can easily disrupt outer membrane, but cannot also simultaneously disrupt inner membrane. Thus mainly periplasmic proteins could be extracted (Ohshima et al. 2000). Although the maximum amount of extracted proteins is typically only a third of that optimally obtainable in homogenization with glass beads, electroporation has the advantages of allowing much faster protein extraction. Furthermore, there is no need for beads addition and/or removal, and it was already successfully tested in a preindustrial pilot flow-through system.

Microalgae Microalgae are currently the most productive biomass feedstock, providing ample motivation for developing techniques of extraction of molecules from microalgae, both in batch and flow systems (Flisar et al. 2014). The fractionating characteristics of electroextraction and its high efficiency can overcome current processing hurdles, in particular when targeting biofuel applications with a low added value (Fig. 5), and several companies now utilize electroextraction from microalgae (de Boer et al. 2012). Electroextraction has been also applied to obtain microalgal RNA, proteins, and pigments.

12

S. Haberl Meglič and T. Kotnik

Fig. 5 Pulsed electric field treatment of microalgae biomass exhibits fractionating properties. After treatment, the aqueous fraction is released into extracellular medium, whereas lipid droplets cannot pass the cell boundary owing to their size. This allows new processing route combinations for complete microalgae biomass valorization. Proteins and minerals can therefore be recovered before processing of the lipid-rich biomass. Reprinted from (Kotnik et al. 2015) with permission

Yeasts Electroextraction from yeasts was first used for transfer of their DNA into recipient bacteria (Kotnik et al. 2015). Soon afterwards, electroextraction of proteins in a batch system from different yeast species was described – Saccharomyces cerevisiae, Kluyveromyces lactis, and Schizosaccharomyces pombe. The yield of electroextracted proteins was shown to depend on electric parameters and/or electroporation medium, and different sizes of functional proteins could be electroextracted. It was found that electroextraction technique provokes reversible membrane permeabilization and increases cell wall porosity while preserving yeast viability (Ganeva et al. 2013).

Multicellular Organisms Grape Electroporation of crushed grapes allows fast processing without an adverse influence on taste, as the temperature of the mash increases by at most several  C. The crushed grapes are first pumped through the electroporation chamber and then stored for several hours for extraction to proceed (Turk 2016). Combining electroporation with subsequent fermentation on grape skins gives a more intense color, while combining electroporation with subsequent maceration yields an increased content of polyphenolic compounds in the wine. For white wines, a lighter character of the wine is usually desirable, but for some white grape varieties, electric pulses can be applied advantageously to achieve a more complete extraction and a more intense taste. Although electroporation also releases more tanning substances, must and

Electroporation-Based Applications in Biotechnology

13

Fig. 6 Simplified processing steps for electroporation of sugar beets and alkaline extraction. Washed sugar beets are sliced. The cossettes are immersed in water and treated with a pulsed electric field (PEF) in an electrode system inside the PEF treatment chamber. Liming is carried out inside a cossette mixer. Then the cossettes are transported through an extraction tower. Water is added and sugar is extracted from the cossettes using a countercurrent extraction method. Sugar is refined from the thin juice by the conventional processing steps of juice evaporation, crystallization, and centrifugation. The extracted cossettes are pressed to remove additional juice and fed into a dryer. The dried cossettes are used for the production of animal feed. Reprinted from (Kotnik et al. 2015) with permission

wine have less acidity as a result of chemical buffering, resulting in a slightly smoother taste. More nitrogen available to the yeast in the must also helps to prevent the untypical aging note of the wine (Mahnic-Kalamiza et al. 2014).

Sugar Beet Electroporation of sugar beets provides for considerable energy saving compared to the standard extraction techniques (Vorobiev and Lebovka 2016). Conventionally, sugar beet tissue is disintegrated thermally, typically at a temperature of approximately 72  C, to prepare the cossettes for subsequent countercurrent extraction of the sugar. Treatment by PEF replaces thermal disintegration by electroporation, with a required energy of ~1–1.5 kWh per ton of sugar beet tissue (Mahnic-Kalamiza et al. 2014; Fig. 6). Although such treatment can be performed at ambient temperature, in an industrial process the cossettes need to be kept after the exposure at a temperature of at least ~60  C to prevent mesophilic bacteria from growing. An inverse temperature profile can be applied advantageously during countercurrent extraction by increasing the temperature from ~60  C to ~80  C, the temperature level for the evaporation stages after extraction. Electroporation-assisted extraction results in a purer juice

14

S. Haberl Meglič and T. Kotnik

Fig. 7 Sugar extraction from sugar beets. Left panel: more efficient extraction of juice from sugar beet using electroporation, rather than conventional pressing. Right panel: demonstration plant with mass flow 10 t/h and two Marx generators (1.2 kJ/pulses, 20 Hz). Up to 30 % less energy is required when electroporation extraction is used. Reprinted from (Haberl et al. 2013b) with permission

(because less water is required for extraction) and lower energy consumption during the evaporation stages (Fig. 7).

Biomass Drying Efficient drying contributes significantly to energy saving in electroporation-assisted sugar beet processing. After extraction, cossettes are pressed for additional juice removal and dried for use as animal feed. Combining exposure to high-voltage electric pulses with alkaline extraction results in increased dry matter content of the cossettes – from 35 % to 40 % after pressing (Mahnic-Kalamiza et al. 2014). Adding lime milk to the cossettes for alkaline extraction immediately after electroporation strengthens the cell walls and thus fosters extraction of juice during pressing. As a consequence, less evaporation energy is required in subsequent high-temperature drying. In the conventional process of sugar production, lime milk is only used for purging the juice. Moreover, the alkaline environment associated with lime milk reduces corrosion of steel tubes and thus increases their lifetime. Dry biomass is also required in fuel production from energy crops. However, energy-efficient drying also allows fresh green biomass to be used. As with sugar beet, drying of green biomass might be carried out by combining electroporation, pressing, and drying in an oven. Electroporation-based treatment of green biomass can be performed in an electrified press; mechanical force is applied before and during pulse application, establishing electric contact to the electrodes through

Electroporation-Based Applications in Biotechnology

15

extracted juice without the need to add water (Sack et al. 2009). When drying the biomass in an oven, electroporated material dries two to threefold faster than nonporated material, not only because of decreased water content after pressing but also because of enhanced diffusion of the vapor as a result of cell disintegration (Sack et al. 2008).

Hardware Considerations and Challenges Devices for large-scale electroporation consist of one or more pulse generators connected to an electrode system for continuous pulse delivery to a mass flow. Both must be carefully designed to achieve desired results. Pulse generators equipped with semiconductor switches in series configuration or in Marx configuration, low-scatter spark gap switches in self-breakdown mode, and spark gap switches triggered by a semiconductor-based trigger generator are in use and under ongoing development. For parallel configuration of Marx generators, over-voltage triggering enables long-term operation of spark gap switches without additional wear. Pulse circuits for rectangular, a periodically damped, or strongly damped oscillating pulse shapes are applied. The electrode configuration of the treatment chamber needs to be selected with respect to mechanical and electrical properties of the processed mass, and to the pulse circuit grounding scheme. For energy-efficient operation by automatic adjustment of applied energy, measurements of the processed mass impedance can be used to assess the degree of changes caused by electroporation, but this approach is of limited resolution in materials with high electrical conductivity such as cell suspensions. Other monitoring methods allowing real-time adjustments of exposure conditions are therefore being explored. Integration of electroporation into an existing production line or process must be carefully planned because the cost of required changes may differ considerably for different designs and may even prove to be unacceptable. When rapid processing is required, introduction of a new step involving electroporation may require adjusting other steps of the process to achieve optimal results (de Boer et al. 2012).

Conclusions Each new experimental technique takes time to develop, further time to understand the mechanisms underlying the data, and even more time to have it adopted in routine laboratory or industrial practice. Due to lack of interest, high initial investment costs (i.e., in industry), or concern over personnel safety, many proposed new techniques are abandoned. However, that has not been the case with electroporation. In this chapter, the most widely established and promising applications of electroporation in biotechnology are presented: genetic transformation, microorganism inactivation, extraction of intracellular compounds from microorganisms and tissues, and biomass drying. At the end hardware considerations and challenges are described, since integration of

16

S. Haberl Meglič and T. Kotnik

devices into industry needs to be carefully adjusted in order to avoid high costs or safety issues. All the considerations outlined above point to a need for better knowledge and deeper understanding of the electroporation phenomenon, as well as of its effects on the permeability of the cell membrane and cell wall, not only in cellular aggregates but also in integral tissues. Extraction, in particular for large molecules, is inherently limited by the presence of a cell wall, and the same is true for fluid filtration in tissues. Thus, the ongoing research efforts largely focus on the influence of the cell wall and tissue structure and on new processing combinations for improving mass transport. Acknowledgment This work was supported by the Slovenian Research Agency (Grant P2-0249) and conducted in the scope of the European Laboratory of Pulsed Electric Fields Applications (LEA EBAM) and within networking efforts of the COST Action TD1104 – European Network for Development of Electroporation-Based Technologies and Treatments (EP4Bio2Med).

References Brim H, Venkateswaran A, Kostandarithes HM et al (2003) Engineering Deinococcus geothermalis for bioremediation of high-temperature radioactive waste environments. Appl Environ Microbiol 69:4575–4582. doi:10.1128/AEM.69.8.4575-4582.2003 de Boer K, Moheimani NR, Borowitzka MA, Bahri PA (2012) Extraction and conversion pathways for microalgae to biodiesel: a review focused on energy consumption. J Appl Phycol 24:1681–1698. doi:10.1007/s10811-012-9835-z Evrendilek G (2016) Pulsed electric field treatment for beverage production and preservation. In: Handbook on electroporation. Springer International Publishing Switzerland Faber K, Harder W, Ab G, Veenhuis M (1995) Review – methylotrophic yeasts as factories for the production of foreign proteins. Yeast 11:1331–1344. doi:10.1002/yea.320111402 Flisar K, Haberl Meglic S, Morelj J et al (2014) Testing a prototype pulse generator for a continuous flow system and its use for E. coli inactivation and microalgae lipid extraction. Bioelectrochemistry 100:44–51 Ganeva V, Galutzov B, Teissie J (2013) Evidence that pulsed electric field treatment enhances the cell wall porosity of yeast cells. Appl Biochem Biotechnol 172:1540–1552. doi:10.1007/ s12010-013-0628-x Gusbeth C, Frey W, Schwartz T, Rieder A (2009a) Critical comparison between the pulsed electric field and thermal decontamination methods of hospital wastewater. Acta Phys Pol A 115:1092–1094 Gusbeth C, Frey W, Volkmann H et al (2009b) Pulsed electric field treatment for bacteria reduction and its impact on hospital wastewater. Chemosphere 75:228–233. doi:10.1016/j. chemosphere.2008.11.066 Haberl S, Jarc M, Strancar A et al (2013a) Comparison of alkaline lysis with electroextraction and optimization of electric pulses to extract plasmid DNA from Escherichia coli. J Membr Biol 246:861–867. doi:10.1007/s00232-013-9580-5 Haberl S, Miklavcic D, Sersa G et al (2013b) Cell membrane electroporation-Part 2: the applications. IEEE Electr Insul Mag 29:29–37 Kilian O, Benemann CSE, Niyogi KK, Vick B (2011) High-efficiency homologous recombination in the oil-producing alga Nannochloropsis sp. Proc Natl Acad Sci U S A 108:21265–21269. doi:10.1073/pnas.1105861108

Electroporation-Based Applications in Biotechnology

17

Kotnik T (2016) Transmembrane voltage induced by applied electric fields. In: Handbook on electroporation. Springer International Publishing Switzerland Kotnik T, Frey W, Sack M et al (2015) Electroporation-based applications in biotechnology. Trends Biotechnol 33:480–488. doi:10.1016/j.tibtech.2015.06.002 Li CH, Corum L, Morgan D et al (2000) The spirochete FlaA periplasmic flagellar sheath protein impacts flagellar helicity. J Bacteriol 182:6698–6706. doi:10.1128/JB.182.23.6698-6706.2000 Mahnic-Kalamiza S, Vorobiev E, Miklavcic D (2014) Electroporation in food processing and biorefinery. J Membr Biol 247:1279–1304. doi:10.1007/s00232-014-9737-x Meddeb-Mouelhi F, Dulcey C, Beauregard M (2012) High transformation efficiency of Bacillus subtilis with integrative DNA using glycine betaine as osmoprotectant. Anal Biochem 424:127–129. doi:10.1016/j.ab.2012.01.032 Ohshima T, Sato M (2004) Bacterial sterilization and intracellular protein release by a pulsed electric field. Recent Prog Biochem Biomed Eng Jpn I 760–760 Ohshima T, Hama Y, Sato M (2000) Releasing profiles of gene products from recombinant Escherichia coli in a high-voltage pulsed electric field. Biochem Eng J 5:149–155. doi:10.1016/S1369-703X(00)00055-3 Sack M, Eing C, Berghoefer T et al (2008) Electroporation-assisted dewatering as an alternative method for drying plants. IEEE Trans Plasma Sci 36:2577–2585. doi:10.1109/ TPS.2008.2002440 Sack M, Attmann F, Staengle R et al (2009) Upgrade of the electroporation device KEA-MOBIL. Acta Phys Pol A 115:1081–1083 Saulis G (2010) Electroporation of cell membranes: the fundamental effects of pulsed electric fields in food processing. Food Eng Rev 2:52–73. doi:10.1007/s12393-010-9023-3 Suga M, Hatakeyama T (2001) High efficiency transformation of Schizosaccharomyces pombe pretreated with thiol compounds by electroporation. Yeast 18:1015–1021. doi:10.1002/yea.753. abs Turk M (2016) Pulsed electric fields assisted extraction of valuable compounds from grape pomace. In: Handbook on electroporation. Springer International Publishing Switzerland Vorobiev E, Lebovka N (2016) Pulsed electric fields processing for sugarbeet and whole crops biorefinery. In: Handbook on electroporation. Springer International Publishing Switzerland

Potential Application of Pulsed Electric Fields for Improving Extraction of Plant Pigments Mustafa Fincan

Abstract

Regarding the potential of pulsed electric field (PEF) to extract pigments from plant tissue, research interest has over the last decade steadily grown with novel findings. PEF has repeatedly been referred to as a novel nonthermal tissue disintegration method that can offer less detrimental changes to tissue and to the extraction of heat-labile pigments. Moreover, it was believed that PEFs can improve the pigment extraction processes and color attributes of some plantbased products. Despite the early evidence, the research on PEF-assisted extraction of plant pigments is still at the development stage and requires better understanding in particular of the effects of different pigments, changes at the cellular level, and favorable extraction strategies. So far, the extractability of the red beetroot betanin through PEF has been most widely studied using aqueous solid-liquid extraction systems. Secondarily, the extraction of anthocyanins from various tissues was proved to be enhanced by the PEF treatment, while less research was carried out on the extraction of oil-soluble pigments, chlorophyll, and carotenoids. This work was conducted mainly to give readers an introductory and fundamental knowledge on plant pigment extraction assisted by PEF in light of recent developments. The first part of the work focuses on the role of PEF treatment in line with general extraction principles. In the final part, specific examples of extracting major plant pigments, namely chlorophylls, carotenoids, betalains, and anthocyanins, from selected research works were introduced; thereby, the approaches, methodologies, and major findings were summarized.

M. Fincan (*) Department of Food Engineering, Erciyes University, Melikgazi, Kayseri, Turkey e-mail: mfi[email protected] # Springer International Publishing AG 2017 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_34-2

1

2

M. Fincan

Keywords

Pulsed electric field • Plant pigments • Extraction • Chlorophyll • Carotenoid • Betalain and anthocyanins

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Consumer Interest in Plant Pigments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 PEF-Induced Extraction of Plant Pigments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Main Characteristics of the Different Pigments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Tissue Disintegration Action in the Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Traditional Tissue Disintegration Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 PEF-Induced Tissue Disintegration Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Methodologies and Major Findings on Extracting Plant Pigments with PEF . . . . . . . . . . . . . . . . . . . 7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Introduction In the extraction of intracellular substances from plant tissues, structural properties play a key role. There are numerous hindrances inside plant tissues that negatively affect mass transfer from inside the tissues. As a result, extraction procedures from plant tissues usually comprise multiple processing stages. Each processing stage aims to reduce or eliminate one or more specific resistances, and further processing is added to the extraction system until all resistances are significantly reduced. Among the processing stages, one or more are devoted specifically to the elimination of cell membrane resistance, which acts as the main barrier against the exit of substances from the cell. Various types of conventional cell permeabilization methods, such as heating, freezing-thawing, and solvents, can be used to achieve this goal. However, each of these methods has different drawbacks, such as high energy expenditure, degradation of heat-labile extract, and undesirable solvent residue in the extraction process. Extraction studies in recent years have widely recognized pulse electric fields (PEF) as a nonthermal cell-permeabilization method with a lower energy expenditure (Knorr and Angersbach 1998; Barba et al. 2015; Fincan 2015). In many past examples, PEF-induced permeabilization was demonstrated to facilitate extraction of various intracellular substances, such as sugar, oil, and juice, without significantly affecting the features of the tissue and extract (Soliva-Fortuny et al. 2009; Donsì et al. 2010). Although the PEF-treatment trials have demonstrated beneficial results in the extraction of a wide range of intracellular substances, its effect on pigment extraction has only recently received considerable research attention. A limited

Potential Application of Pulsed Electric Fields for Improving Extraction of. . .

3

number of studies have been conducted so far in this area and from the early results, PEF has been deemed to have significant potential for the improvement and intensification of pigments in various extraction processes. However, more research is needed to better understand the effects of different pigments with further extraction strategies. The aim of the present work is to present introductory knowledge on plant pigment extraction assisted by PEF. The first part of this chapter provides an overview to the readers about the roles of PEF in extraction. It also treated and explained the subject as a new tissue disintegration method compared to conventional ones. In the final part, specific examples of PEF-assisted extraction of plant pigments from selected research were introduced, and the approaches, methodologies, and major findings were summarized.

Consumer Interest in Plant Pigments The current practice of the food, cosmetics, and pharmaceutical industries involves using a wide range of synthetically produced colorants for attaining a product color with desirable characteristics. In contrast, consumers overwhelmingly demand naturally colored products or products colored with naturally extracted pigments. This trend is not only provoked by the putative adverse effect of synthetic colorants but also the versatile health promoting effects of natural pigments. When the biological efficacy of plant pigments was studied using various test methods, including in vitro bioassays, in vivo animal trails, population surveys, and less frequently, clinical trials, many of the findings pointed out the supportive role of plant pigments in disease prevention, therapy, and health prompting. Anthocyanins have been claimed to be associated with the improvement of vision disorders, neuroprotection, cardiovascular disease, and cancers. Carotenoids are associated with the improvement of photooxidative damage, the immune system, cardiovascular disease, and cancers, while betalains and chlorophylls are associated with the improvement of cancers (Andersen et al. 2004). Moreover, considering that there are numerous varieties of pigments within the same pigment category, each variety can also bear a specific bioactive potential. For example, red cabbage contains around 20 types of anthocyanins among more than 250 different types, and its pigments are widely used in the production of beverages, yogurt mixes, confectioneries, and ice creams (Chigurupati et al. 2002). Hagiwara et al. (2002) demonstrated that dietary red cabbage anthocyanins suppressed lesions and reduced colorectal carcinogenesis in rats. In another study, red cabbage anthocyanins were found to have a strong antioxidant activity, and it was suggested that it reduced oxidative stress and exhibited anti-inflammatory effects (Lin et al. 2008; Kolodziejczyk et al. 2011).

4

M. Fincan

PEF-Induced Extraction of Plant Pigments Main Characteristics of the Different Pigments Although literature on the effects of PEF on the plasmalemma of plant cells is readily available, a lot less is known on the inner membranes enclosing the pigments. Because a specific type of pigment can be found in different organelles inside the cell, one can at the first approximation consider electroporation of multiple membranes, the organelle’s membrane, and the plasmalemma to facilitate extraction. Permeabilization of the plasmalemma can lead to destabilization of inner membranes because of osmotic disequilibrium (Luengo et al. 2014b). In particular, the vacuole where betalains and anthocyanins are found is known to be very sensitive to osmosis and can thus be destabilized upon permeabilization of the outer membrane. Chlorophyll and carotenoids, surrounded by outer envelope/membranes of the chloroplasts and chromoplasts, are lipophilic in nature, meaning they are extracted by using organic solvents (Andersen et al. 2004). In designing a PEF-induced extraction experiment, it is also crucial to consider the stability aspects of different pigments and take necessary precautions. Experimental conditions other than PEF can easily affect stability, which in turn leads to erroneous result in the interpretation of the outcome. It is well-known that most pigments are highly susceptible to degradation or lose their color attributes by the intrinsic and extrinsic players. When green vegetables are exposed to mild heat treatment and light, chlorophyll can easily be degraded into many of its derivatives. An acidic environment favors the degradation of green pigment components, chlorophyll a and chlorophyll b, into pheophytin a and pheophytin b, both dark yellowish pigments, and the reaction is further catalyzed with the help of heat. Moreover, this reaction can also be caused by the indirect enzymatic activity of chlorophyllase, that is, chlorophyll a and b are first converted into chlorophyllide by chlorophyllase and then further degraded into pheophytins with the assistance of acid and heat. On the other hand, it is possible that pheophytins, being unstable pigments, can further be decomposed into many other compounds in different colors by the enzymatic and nonenzymatic reactions. Anthocyanins, a group of flavonoids within the polyphenols, are one of the least stable pigments. The color of anthocyanins characteristically varies with the pH of the medium, depending on their chemical structure and the presence of a copigment in the medium. Below pH 2, redness is predominant while at pH 4–5, colorless forms are dominant; however, this is usually not visible in fresh plant tissues because of the presence of copigments that have a stabilizing effect on anthocyanins. Although stability of anthocyanins can vary from one type to another, many anthocyanins are susceptible to degradation by the action of heat, ascorbic acid, peroxides, sugars and their degradation products, enzymes, oxygen, and light. It is worth mentioning that heat treatment among these factors plays the most predominant role and is largely responsible for the loss of color in many processed fruits and vegetables. Besides, ß- glucosidase and α-arabinoside degrade the anthocyanins into unstable anthocyanidines and sugars and convert the anthocyanidines into colorless compounds. It has been suggested that anthocyanins

Potential Application of Pulsed Electric Fields for Improving Extraction of. . .

5

are weak substrates for polyphenol oxidases, but the enzyme can indirectly favor the degradation of the anthocyanins, that is, o-quinone, a product from the activity of polyphenol oxidases on phenolics, can react with the anthocyanins to cause color loss. Stability of carotenoids is most affected by the presence of oxygen, or they can readily be oxidized because of the double bond in their chemical structure. The rate of the reaction increases with an increase in heat and light exposure. The next predominant factor after oxidation is usually the high temperature treatment. Moreover, the light itself can also provoke degradation of many carotenoids. Concerning the degradation by the enzymatic activity, their decomposition is especially caused by lipoxygenases. Color of betalains does not vary with the pH of the medium, but pH has an effect on their stability. For example, betanin, one of the most common members of betalains, is degraded into betalamic acid and cyclodopa-D-glucoside. The reaction can also occur at the heat treatment of the acidic medium. In addition, betalains are susceptible to oxidation reactions, as related to the pH of the medium and exposure of light. The oxidation reaction takes place minimally at a pH range of 4–5. Enzymes are also responsible for the color loss of betalains. Peroxidase, found widely in red beetroot, has a bleaching effect on betacyanins and betaxantines depending on the concentration of hydrogen peroxide. The bleaching induced by peroxidase increases at an acidic medium and reaches a considerably high rate at pH 3–4 (Cemeroğlu et al. 2004).

Tissue Disintegration Action in the Extraction The general extraction procedure for plant pigments shares many common points with those for other types of intracellular substances. After permeabilization of the membranes achieved through any disintegration method, the various types of changes are triggered in the tissue. In some cases, such changes are beneficial from the extraction point of view while undesirable in other cases. This depends highly on the type and intensity of the treatment, as well as the sensitivity of the extract and tissue to the given treatment. Moreover, with the loss of membrane integrity, various substances retained in separate sections within the cell come into contact with each other – enzymes, external solution, and oxygen – and the likelihood for degradation reactions greatly increases.

Traditional Tissue Disintegration Methods Conventional cell disintegration methods, such as heating, freezing-thawing, grinding, and solvent treatment, are widely used in many extraction processes. The changes in the extraction brought about by those methods vary from one to the other. In addition, magnitudes of those changes are influenced significantly by the level of temperature treatment. Thus, the methods are often categorized as thermal or nonthermal, and their effects in the extraction can presumably be foreseen and interpreted as related to the level of temperature treatment. Exemplarily, heat

6

M. Fincan

treatment at about 100  C results in the melting of cellular membranes, denaturation of proteins, and inactivation of enzymes by favoring extractability; on the other hand, the treatment may lead to the destruction of heat-labile substances and the formation of unwanted by-products. Also, this scenario changes when the temperature of the heat treatment is lowered to, for example, 50  C. Freeze-thaw treatment represents an example of a nonthermal tissue disintegration method. Ice growth followed by osmotic drying during slow freezing damages the cellular membranes, and in the post-thawing period, the substance of extraction interest becomes highly susceptible to degradation reactions. Also, many extraction procedures utilize the mixture of various types of solvent. The polarity of the extract and solvent is an important parameter as similar polarities will yield higher solubility, which in turn results in higher diffusion through the tissue. Organic solvents such as ethanol and hexane work in several ways, for example, impair cellular membranes and solubilize nonpolar extract, thus contributing to the mass transfer of the extract into solution; however, they are costly and can leave unwanted residues in the product.

PEF-Induced Tissue Disintegration Method The PEF-assisted pigment extraction can be designed by using different experimental approaches. A size reduction process before PEF treatment is seen to be very common. The tissue is typically processed into a form of small pieces, mash, puree, or pomace. Moreover, a manual PEF application to the sample can also differ. In one approach, only the tissue is submitted to PEF treatment, while in the other approach, the mixture of tissue and solution is treated. The magnitude of pulse variables, namely pulse length, pulse number, interpulse duration, and treatment time, affects the pore formation process, hence the extractability. Characteristics of the pores such as pore size and distribution in the cellular membranes are critically important from the extraction point of view. The pores are deemed to be aqueous channels, which means the polar pigments can easily pass through while nonpolar ones have much less affinity to the passage (Weaver and Chizmadzhev 1996). To maximize extractability, pulse variables should be adjusted to produce irreversible pores, the largest possible in size and most frequent on the membranes. A field strength range of 0.5–10 kV/cm can be typically employed to achieve the irreversible permeabilization of plant tissues. There is a tendency that the longer the pulse width is used, the larger the pores develop. Pulse widths in the range of millisecond result in larger pores as compared to microsecond-scale pulses (Saulis 2010; Saulis and Saule 2012). Similarly, an increase in the number of pulses and interpulse duration can lead to the formation of higher membrane damage or larger pores (Pakhomov et al. 2010; Silve et al. 2014). Furthermore, medium-related parameters, such as temperature, cell/tissue size, and electrical conductivity, can have important effects on the mode of PEF action. The membranes in the lipid nature start to liquefy as temperature increases from 55  C to 65  C (Eshtiaghi and Knorr 2002). PEF application to a highly conductive medium results in a larger current passage across the electrodes, thereby leading to a

Potential Application of Pulsed Electric Fields for Improving Extraction of. . .

7

more joule heating effect as compared to low-conductive medium. As a result, the thermal effect itself can contribute to permeabilization. In particular, this effect is apparent in size reduction processes. The size reduction process can inevitably damage the cells on the cut surface, leaving their conductive intracellular liquid on the surface. In contrast, the effects resulting from PEF-induced permeabilization are more separable if the conductivity around the tissue is lowered through, for example, washing away the conductive debris and/or suspending the sample in a low-conductive medium. Besides, extraction from the PEF-treated samples is often enabled using various solid-liquid extraction processes. Variables of the extraction, such as temperature level, type of solvent, and tonicity of the extracting medium can also affect the extraction process and thus can easily interfere with the effect of PEF. Also, the degree of tissue disintegration achieved by PEF treatment needs to be determined before proceeding to a subsequent extraction stage. This is necessary since there is a direct relationship between the degree of disintegration and the amount of extract, that is, the more cells are permeabilized, the higher extraction is expected. One of the common approaches for the estimation of the ratio of the permeabilized cell is based on the change in the electrical conductivity of the tissue before and after PEF treatment. By dividing the conductivity of the treated tissue with that of the fully permeabilized tissue, various tissue disintegration indices have been developed and used to predict the extent of the disintegration between Z = 0 and 1 (Lebovka et al. 2002; Knorr and Angersbach 1998). Apart from this approach, amount of extract from fully disintegrated tissues can alternatively be measured as a maximum reference and is then associated with the extract amount from the PEF treated tissue.

Methodologies and Major Findings on Extracting Plant Pigments with PEF Although the release of plant pigments after PEF treatment was previously reported (Dörnenburg and Knorr 1993), detailed investigations were undertaken much later. In a prominent study, Fincan et al. (2004) showed that about 90% of the red colorants from the red beetroot were extractable using a PEF-treated tissue in a solid-liquid extraction (SLE) system. Later, more detailed studies were conducted by different researchers, and the role of the variables in the extraction was more widely explored using different approaches, plant tissues, and plant pigments. Thus far, much of the attention has been toward extractability of betalains and anthocyanins in an aqueous SLE system whereas the effects on chlorophyll and carotenoids remained much less explored. Aqueous solvents, especially isotonic ones, do not significantly interact with intact cell membranes; thus, it becomes easier to characterize the PEF-induced effects on the extraction of water-soluble pigments as compared to oil-soluble ones. By contrast, a nonpolar extracting medium, such as ethanol and hexane, can severely damage intact cell membranes during SLE, making evaluation of the PEF effect more difficult. To maximize the PEF-assisted pigment extraction, manipulations of pulse and extraction variables have been a major focus. Table 1 shows a summary of

8

M. Fincan

Table 1 A summary of PEF-assisted plant pigment extraction from different studies PEF-treated plant material Red beetroot pieces

PEF-treatment conditions Rectangular mono pulses, 1 kV/cm, 27–270 pulses, 10 μs of pulse width, 20 ms of interpulse duration

Betanin

Red beetroot pieces

Exponential decay pulses, 1–9 kV/cm, 5 pulses, 2 μs of pulse width, 20 ms of interpulse duration

Betanin

Red beetroot pieces

Betanin

Red beetroot pieces

Near-rectangular, 1–9 kV/cm, 100 μs of pulse width, 100 ms of interpulse duration Rectangular pulses, 0.2–6 kV/cm, 5–100 pulses, 3 μs–60 ms of pulse width, 1 Hz

Anthocyanin

Red cabbage mash

Exponential decay pulses, 2.5 kV/cm, 50 pulses, 15 μs of pulse width, 1 Hz

Anthocyanin

Grape pomace

Rectangular mono pulses, 0–3 kV/cm, 200–2000 pulses, 100 μs of pulse width, 100 ms of interpulse duration

Pigment Betanin

Extraction method and main findings Solid-liquid extraction (SLE) in isotonic water solution: About 90% of the total red pigments extracted at 270 pulses. Solution conductivity correlated with the pigment concentration SLE in McIlvaine buffer solution: Extraction of red pigments up to 90% and higher extraction yield in the extracting medium of pH 3.5 and pH 5 + 1% ascorbic acid at 30–50  C SLE in distilled water: Nearcomplete level of extraction reached at 30  C and 50 min SLE in McIlvaine buffer solution of pH 3.5 at 25  C: Pulses in microsecond range found to be more efficient than in millisecond range SLE in deionized water at 25  C for 4 h: Improvement in the extraction yield with higher efficiencies for nonacylated anthocyanins SLE in a mixture of ethanol and water at 20–50  C for 420 min: Densification of the sample to 1.0 g/cm3 followed by the

References Fincan et al. (2004)

López et al. (2009)

Loginova et al. (2011)

Luengo et al. (2016)

Gachovska et al. (2010)

Brianceau et al. (2015)

(continued)

Potential Application of Pulsed Electric Fields for Improving Extraction of. . .

9

Table 1 (continued) Pigment

PEF-treated plant material

PEF-treatment conditions

Anthocyanin

Grape byproduct

Exponential decay pulses, 3 kV/cm, 30 pulses, 2 Hz

Carotenoid

Tomato peel and pulp

Rectangular pulses, 3–7 kV/cm, 5–100 pulses, 3 μs of pulse width, 1 Hz

Carotenoid

Carrot puree

Bipolar rectangular pulses, 0.1–1 kV/cm, 5–100 pulses, 20 μs of pulse width, 10–75 Hz

Carotenoids and chlorophylls

Microalgae, (Chlorella vulgaris) in McIlvaine buffer solution at pH 7

Rectangular pulses, 10–25 kV/cm, up to 50 pulses, 3 μs of pulse width, 0.5 Hz

Extraction method and main findings PEF-treatment at 1.2 kV/cm increased the extractability most specifically at 20  C and 35  C SLE in a mixture of ethanol and water at 70  C for 1 h: Significant increases in individual and total anthocyanins SLE in mixtures of hexane, acetone, and ethanol at 25  C: PEF treatment at 5 kV/cm increased carotenoid extraction from tomato peel by 39% compared to the control. Higher extraction ratio at acetone (100%), hexane:ethanol (50:50%), and hexane:acetone: ethanol (50:25:25%) than at ethanol (100%) and hexane: acetone (50:50%) SLE in hexane: acetone:ethanol (50:25:25%): PEF treatment of the pomace at 0.1 kV/cm improved the extraction of α- and β-carotene by 20–35% as compared to the untreated sample SLE in 96% ethanol: PEF treatment (50  75 μs at 20 kV/ cm) increased the extraction of carotenoids and chlorophyll a and b. The effect increased with

References

Corrales et al. (2008)

Luengo et al. (2014b)

Roohinejad et al. (2014)

Luengo et al. (2014b)

(continued)

10

M. Fincan

Table 1 (continued) Pigment

Carotenoid

PEF-treated plant material

Red paprika mash

PEF-treatment conditions

Exponential decay pulses, 1.7 kV/cm, 30 pulses, 300 μs of pulse width, 1 Hz

Extraction method and main findings incubation in a McIlvaine buffer solution at pH 7 Mechanical extraction (ME) at 10 MPa for 4 min to expel the juice: About 17% increase in β-carotene content of the juice compared to the untreated

References

AdeOmowaye et al. (2001)

PEF-assisted extraction of plant pigments with the experimental conditions and main findings from different studies. In the experimental approach used by Fincan et al. (2004), the cut red beetroot tissue (1 mm in thickness and 4 mm in diameter) was PEF-treated. Before the treatment, the sample was washed with distilled water in order to lower the conductivity around the tissue during PEF treatment. This action was also deemed necessary as, in a preliminary experiment, it was observed that onion epidermal cells were more easily electropermeabilized in a low-conductive medium with a few pulses. In contrast, the same number of pulses in a high-conductive medium (saline) did not yield permeabilization and further induced electrolysis around the electrodes. For the SLE experiment, an aqueous isotonic level (0.25 M Mannitol) was chosen to minimize the adverse osmotic effects during the SLE period. As shown in Fig. 1a, c, the extractability of pigment and ionic species increased incrementally with the intensity of PEF treatment. The higher degree of variation in the extractability at intermediate PEF intensities is likely the result of the nonhomogenous permeabilization due to different cell sizes. After 270 rectangular pulses of 10 μs at 1 kV/cm (specific energy of 7 kJ/kg), the extraction level was found to be closely comparable to that of the mechanical pressing and freezethawing method (Fig. 2). It was then concluded that this intensity of PEF caused a near-complete level of tissue permeabilization and accordingly resulted in the extraction of about 90% of total red pigments and ionic content. Supportively, when the cross section of the PEF-treated sample after extraction is examined under microscope, most cells appeared colorless or with an insignificant trace of red color (Fig. 3). Worthy of note is the progress of pigment extraction correlated linearly with that of the solution of electric conductivity. The implication of this relationship is that the solution of electric conductivity could be used to predict the progress of pigment extraction unless pigments degrade during the SLE period. As for the relationship between conductivity-based tissue damage index and extraction level of the

Potential Application of Pulsed Electric Fields for Improving Extraction of. . .

11

Fig. 1 Kinetics of color and ionic species extraction after PEF treatments with the indicated number of pulses (a, c) and freezing/thawing (b, d), respectively. Replicate experiments shown (Fincan et al. (2004) with the permission from Elsevier)

Fig. 2 The absorbance and electric conductivity after 1 h of aqueous extraction following PEF treatment (270 pulses), PEF treatment followed by freezing/thawing and mechanical pressing (The absorbance measured at 530 nm, the mannitol blank subtracted). The length of the boxes for each boxplot represents the interquartile range (IQR) divided at median. Vertical whiskers represent largest and smallest values within 1.5 (IQR) of top and bottom of the boxes (Fincan et al. (2004) with the permission from Elsevier)

12

M. Fincan

Fig. 3 Cross section of (a) PEF-treated (270 pulses) and (b) untreated red beet tissue after extraction. Cell walls were stained with tryptophan blue (Fincan et al. (2004) with the permission from Elsevier)

Fig. 4 (a) Absorbance and (b) electric conductivity of the solution at the end of 1 h of aqueous extraction versus relative change in tissue electric conductivity following PEF treatment with the indicated number of pulses (The absorbance measured at 530 nm) (Fincan et al. (2004) with the permission from Elsevier)

pigments or ionic species, the index at its maximal coincided with the highest degree of pigment or ionic extraction. However, the progress of the extraction was correlated linearly only up to the index level of 20–30% (Fig. 4). The ongoing conductivity increases in the tissue during pulsation could be responsible for the nonlinear behavior observed (Knorr and Angersbach 1998), that is, considerable increase in tissue conductivity which was caused by the initial pulses results in a decrease of the pulse efficiency of later pulses. In the extension of this work, the rate of mass transport during the extraction was studied by the Fickian diffusion model (Chalermchat et al. 2004). It has been proposed that the extraction at high degrees of permeabilization achieved by PEF or freezing/thawing could be described by only a (single) fast coefficient, whereas at a low degree of PEF permeabilization, a bimodal diffusion model with a fast and a slow diffusion coefficient accounted for the mass transport.

Potential Application of Pulsed Electric Fields for Improving Extraction of. . .

13

Fig. 5 Influence of the pH of the extracting medium on betanine extraction by PEF in McIlvaine buffer. Treatment conditions: 5 pulses, 5 kV/cm (1.2 kJ/kg) (López et al. (2009) with the permission from Elsevier)

Later, after the work of Fincan et al. (2004), the SLE of betanin from PEF-treated red beetroots was more extensively studied by different researchers. Although the experimental design used in those studies varied to some degree, the major focus was to acquire a better understanding of the role of variables to increase the extractability of red pigments. López et al. (2009) investigated the effects of PEF treatment intensity and the role of some operating conditions of SLE, namely mechanical pressing of the samples, pH, temperature, and ascorbic acid level. In the experiment, the tissue sizes, 5 mm in thickness and 25 mm in diameter, were PEF-treated (Table 1) and then submitted to SLE in a stagnant McIlvaine buffer solution varying in pH from 3 to 6.5 and temperature from 10  C to 60  C. The study reported that it was also possible to extract 90% of the pigments with less specific energy input (2.5 kJ/kg, 5 pulses of 7 kV/cm) than the level reported by Fincan et al. (2004). Consistent with earlier data, the result obtained can be attributed to fact that the ratio of permeabilized cells per specific energy input increases with an increment in field strength. However, the entire extraction takes longer, 300 min, because of a larger sample size and stagnant extracting medium. The study further claimed that the acidity and temperature of the SLE medium had a significant impact on the extraction yield of betanin. Correspondingly, the extraction yield was reported to be considerably higher at pH 3.5 or pH 5 + 1% ascorbic acid, and temperature ranges from 30  C to 50  C (Figs. 5 and 6). It is likely that the stability of betanin is better maintained at those ranges, thus leading to the higher yield, whereas, a temperature level of 60  C appears to negatively impact its stability. Loginova et al. (2011) measured the effect of heat treatment on degradation of red beetroot colorants by using freshly squeezed red beetroot juice. It has been demonstrated that 5 h of thermal treatment at 30  C yielded a degradation index of D  0.2

14

M. Fincan

Fig. 6 Influence of the temperature of the extracting medium on betanine extraction by PEF after 400 min in McIlvaine buffer of pH 3.5. Treatment conditions: 5 pulses, 5 kV/cm (1.2 kJ/kg) (López et al. (2009) with the permission from Elsevier)

whereas 1 h of thermal treatment at 80  C led to a complete degradation (D  1.0) (Fig. 7). The result implies that if the tissue is nonthermally disintegrated, the colorants are still susceptible to degradation even at 30  C by extrinsic and intrinsic factors, such as the active enzymes. Moreover, Loginova et al. (2011) investigated the temperature effects on the SLE of red beetroot pigments by using a different experimental approach. Discular tissue (10 mm in thickness and 41 mm in diameter) whose surface was covered with the beet juice was PEF-treated. During the treatment, an individual pulse train consisting of 20 pulses at 1 kV/cm was repeated to obtain a total treatment time of 0.1 s, and a pause of 25 s between the trains was allowed to avoid heating. This intensity of the PEF treatment was reported to yield a disintegration index of Z = 0.9. Furthermore, the SLE experiment was performed in agitated distilled water varying in temperature range from 30  C to 80  C. And extraction yields of the colorants at Z = 0.9 were determined by measuring the electrical conductivity of the extracting medium and were expressed as the extraction index B of the colorants in Fig. 8b. However, it should be noted here that, although solution conductivity is correlated with the progress of colorant extraction (Fincan et al. 2004), it probably does not supply direct information on the degradation rate of the colorants. Thus, in Fig. 8b, heatinduced colorant degradation appears to be largely invisible. On the other hand, Fig. 8a provides valuable information on the effects of the extracting medium’s tonicity and temperature on the permeabilization behavior of untreated tissue. The increase in extraction index B versus time strongly implies the phenomenon of plasmolysis as affected by the extracting medium’s tonicity and temperature. The

Potential Application of Pulsed Electric Fields for Improving Extraction of. . .

15

Fig. 7 Colorant degradation index of the red beet juice D versus thermal treatment time tT at different temperatures T (Loginova et al. (2011) with the permission from Elsevier)

effect appears to be considerably small after an extraction period of 50 min at 30  C; however, it becomes significantly high at 40  C. If noted to the methodological approaches of the above studies, it can be generalized that the extractability was investigated under the following conditions: (a) the plant tissue to be treated is nearly intact and (b) the medium of the SLE is aqueous. Consequently, the extraction is expressed as influenced by the degree of PEF-induced tissue disintegration. Also, the use of aqueous extracting mediums is necessary both for solubilizing polar pigments and maintaining the integrity of cellular membranes. On the other hand, characterizing the effects of PEF on the extraction of oil soluble pigments is more complex. In past studies done in this area, plant material before PEF treatment is often seen to be preprocessed to mash, pomace, or puree, losing the important percentage of intact cells. In addition, organic solvents used in the SLE of oil soluble pigments are also severely destructive for cellular membranes. In spite of these facts, PEF pretreatment was conformingly reported to have an effect to enhance the extractability of oil-soluble pigment in various studies. Although PEF-induced permeabilization of intact cells remaining in processed plant material can be partly accounted for the increase in oil-soluble pigment, the exact mechanism of the PEF action remains largely unknown. For example, when a completely permeabilized tissue (frozen-thawed grape by-products) was PEF-treated and extraction of anthocyanins was studied using a hydroalcoholic extracting medium at a temperature of 70  C, about a onefold

16

M. Fincan

Fig. 8 Extraction index B of the red beetroot colorants versus extraction time tE at different temperatures T for untreated (a) and PEF pretreated (b) samples of red beet tissue (Loginova et al. (2011) with the permission from Elsevier)

increase in total anthocyanin content was reported (Corrales et al. 2008). In the following section, methodologies of such applications were introduced with the summary of major findings. Luengo et al. (2014a) studied the effect of PEF treatment on the carotenoid extraction from tomato peel and pulp (tomato waste) using different solvent mixtures with a stirring action. Prior to SLE, the degree of tissue disintegration induced by PEF treatments was quantified. It has been reported that the permeabilization of the peel required higher PEF treatment intensity as compared to pulp. The different structural properties of the peel, such as the presence of a hydrophobic waxy layer from the pulp, could possibly be responsible for the higher intensity found. No significant effect on the carotenoid extraction from the pulp was reported. However, PEF treatment of the peel at 5 kV/cm was found to maximally increase the extraction level by 39% as compared to the control. This is likely because of the better infiltration of the solvents through electroporation. Moreover, the PEF-assisted extraction was

Potential Application of Pulsed Electric Fields for Improving Extraction of. . .

17

Fig. 9 Influence of the solvent extraction mixture on the carotenoid extraction yield (CEY) from control (white bars) and PEF-treated (5 kV/cm–90 ms) (black bars) tomato peel after 300 min of extraction. The error bars represent SEM (Luengo et al. (2014a) with the permission from Frontiers)

shown to be significantly affected by the type of solvent and solvent mixtures (Fig. 9). The study concluded that with the PEF treatment, it is possible to reduce the amount of hexan from 45% to 30% in the hexane-ethanol mixture without affecting the carotenoid extraction yield. In the work done by Roohinejad et al. (2014), carrot puree was treated with PEF, and the carotenoid extraction was studied using a solvent mixture, 50% hexane, 25% acetone, and 25% ethanol with stirring action. In the procedure, the PEF-treated samples (Table 1) were centrifuged to separate the pomace (solid phase) and the juice (the liquid phase) and then separately submitted to SLE. The study reported that the PEF treatment of the pomace at 0.1 kV/cm improved the extraction of α- and β-carotene by 20–35% as compared to the untreated sample, and further field strength increment did not yield a considerable increase in the extraction level. In contrast, the PEF treatment at 0.1–1 kV/cm was found to reduce the α- and β-carotene content of the juice. Gachovska et al. (2010) applied PEF treatment to the mashed red cabbage and studied its effects on anthocyanin extraction by using deionized water as extracting medium (Table 1). For the SLE, the PEF-treated mash was rotated end-over-end a

18

M. Fincan

few times in a flask and then placed in a dark box for 4 h at 22  C. The study reported that PEF treatment enhanced the total anthocyanin extraction by 2.12 times, and the degree of enhancement varied with the type of anthocyanin from 1.34 to 3.77 with the higher efficiencies for nonacylated anthocyanins. Moreover, thermal and light stabilities of the PEF-treated samples and control samples were found to be similar, indicating that the PEF treatment did not produce any chemical effects to accelerate light and thermal degradation of anthocyanins. In a study done by Brianceau et al. (2015), fermented grape pomace was treated with PEF in combination with densification treatment, and extractability of anthocyanin was investigated by using a hydroalcoholic SLE (mixture of ethanol and water (50/50, v/v)) system at different temperatures. In the method, the sample was first compressed by applying pressure from 0.8 to 10 bar (yielding a densification from 0.8 to 1.3 g/cm3) and then exposed to different PEF treatments at every compression level. As measured by the percent recovery of polyphenols in the extraction, the highest synergism was obtained at a pressure level of 2 bar (ρ = 1 g/cm3) and a PEF treatment intensity corresponding to E = 1.2 kV/cm and W = 18 kJ/kg (Z = 0.36). The recovery at this level of combined treatment was found higher than the 0.8 and 10 bar. It has been hypothesized that the use of 10 bar caused more mechanical rupture of the cells, thus decreasing the ratio of intact cells to electropermeabilized cells. Correspondingly, when the extraction of anthocyanins at the given combined treatment levels was studied, it has been found that PEF treatment increased anthocyanin extractability regardless of temperature increase (Table 2). Conclusively, the study suggested the PEF can be utilized for the treatments of relatively dry pomace to improve the extraction with lower specific energy consumption. In a recent work, Luengo et al. (2014b) investigated the extraction of carotenoids and chlorophylls from PEF-treated microalgae, Chlorella vulgaris. In the methodology, the microalgal suspension in a citrate phosphate McIlvaine buffer (1 mS/cm; pH 7) was exposed to PEF (Table 1), and the samples, either immediately or after incubation in the buffer, were added to 96% ethanol and vortexed for the pigment extraction. It has been reported that the highest extractability was obtained at 50  75 μs at 20 kV/cm. Interestingly, this study revealed that the incubating of the samples in the buffer post–PEF treatment had a significant impact in the extraction process. Accordingly, PEF treatment with 1 h incubation increased the extraction of carotenoids, chlorophyll a, and chlorophyll b, 1.2, 1.6, and 2.1 times higher than the control, respectively. It has been speculated that the effect of the incubation period could be related to plasmolysis of chloroplast membranes, likely due to osmolytic disequilibrium in the periplasmic space after permeabilization of the cytoplasmic membrane. Ade-Omowaye et al. (2001) investigated the effect of PEF on mechanical extraction of juice from red paprika. In the experiment, paprika mash was either treated with pectolytic enzymes having macerating activity or PEF until a tissue disintegration index value (Z) 0f 0.80–0.90 was obtained. The treated samples were then pressed in a hydraulic press at 10 MPa for 4 min to expel the juice. When the juice was analyzed in terms of color attributes (L, a and b values, and ß-carotene content),

20  C 0 kJ k 0.003a 0.035a 0.113a 0.309a 0.460a

1

18 kJ kg 0.004b 0.045bc 0.135b 0.363b 0.547b

1

35  C 0 kJ kg 1 0.004ab 0.045abc 0.138bc 0.383b 0.570bc 18 kJ kg 0.005b 0.053c 0.158c 0.431c 0.648c

1

50  C 0 kJ kg 0.004ab 0.041ab 0.128ab 0.362b 0.535b 1

18 kJ kg 0.004b 0.045b 0.135b 0.383bc 0.566b

1

SD 0.001 0.005 0.011 0.029 0.046

a–c

For each line, means that are followed by the same letter are not significantly different (P > 0.05). SD: numbers in italics give standard deviation for both control and treated samples for the three studied temperatures. Data in bold are the sum of individual anthocyanins (Deph-3-Glu, Pet-3-Glu, Peo-3-Glu, and Mal-3-Glu). RM raw material

Temperature PEF treatment Deph-3-Glu (g/100 g RM) Pet-3-Glu (g/100 g RM) Peo-3-Glu (g/100 g RM) Mal-3-Glu (g/100 g RM) Total (g/100 g RM)

Table 2 Anthocyanins extracted (expressed in g/100 g RM) after 420 min of extraction at 20  C, 35  C, and 50  C for nontreated and PEF-treated samples (Brianceau et al. (2015) with the permission from Elsevier)

Potential Application of Pulsed Electric Fields for Improving Extraction of. . . 19

20

M. Fincan

the value for coefficient a (signifying redness) from the PEF-treated mash was shown to be significantly higher than that of the enzymatically treated juice. The study also demonstrated that the PEF treatment enriched the ß-carotene content of the juice about 17% higher than the untreated or the enzymatically treated mash. Apart from the SLE-based studies, the study represents another way of PEF treatment utilization in pigment extraction, that is, color or pigment enhancement option for plant-based products through the effects of PEF.

Conclusion As has been evidently demonstrated in many of the studies conducted so far, PEF has potential to improve the extraction and intensification of pigments in a number of extraction processes. The nonthermal way of tissue disintegration action and low-energy consumption provided by PEF fit well with the extraction requirements of plant pigments, most of which are heat-labile. As to the state of the research, PEF-assisted extraction of plant pigment is currently at the beginning level; more studies are needed for PEF to be widely utilized in the extraction processes. Conclusively, water-soluble pigments appear to be more selectively extractable by using aqueous SLE systems following the PEF treatment. Moreover, the effect of PEF is more separable in such extraction systems as membrane integrity is better preserved during the SLE period. On the other hand, necessity of use of organic solvents for the SLE of oil-soluble pigments makes the assessment of PEF’s role in the extraction more difficult as the solvents will also impair the cellular membranes. The mechanism leading to the increase in the extraction of oil-soluble pigments remains largely unknown. Moreover, further research is needed to elucidate the effects of PEF on the stability of pigments in pure solutions, cytoplasmic medium and inner membranes surrounding the pigments.

Cross-References ▶ Electric Pulse Parameters Affecting Electroporation Treatment Outcome ▶ Extraction of Valuable Compounds from Microalgae Using Pulsed Electric Fields ▶ Pulsed Electric Field-Assisted Extraction of Pigments from Chlorella vulgaris ▶ Pulsed Electric Fields Assisted Extraction of Polyphenols from Grape Pomace ▶ Pulsed Electric Fields for Extraction of Secondary Metabolites from Plants ▶ Selective Extraction of Molecules from Biomaterials by Pulsed Electric Field Treatment

Potential Application of Pulsed Electric Fields for Improving Extraction of. . .

21

References Ade-Omowaye BIO, Angersbach A, Taiwo KA, Knorr D (2001) The use of pulsed electric fields in producing juice from paprika (Capsicum annuum L.) J Food Process Preserv 25:353–365. doi:10.1111/j.1745-4549.2001.tb00466.x Andersen OM, Christinet L, Cuttriss AJ et al (2004) Plant pigments and their manipulation. Annu Plant Rev 14:369. doi:10.1093/aob/mci287 Barba FJ, Parniakov O, Pereira SA et al (2015) Current applications and new opportunities for the use of pulsed electric fields in food science and industry. Food Res Int 77:773–798. doi:10.1016/ j.foodres.2015.09.015 Brianceau S, Turk M, Vitrac X, Vorobiev E (2015) Combined densification and pulsed electric field treatment for selective polyphenols recovery from fermented grape pomace. Innovative Food Sci Emerg Technol 29:2–8. doi:10.1016/j.ifset.2014.07.010 Cemeroğlu B, Yemenicioğlu A, Özkan M (2004) Meyve ve Sebze İşleme Teknolojisi. In: Cemeroğlu B (ed) Meyve ve Sebzelerin Bileşimi, 2nd edn. Kültür ve Turizm Bakanlığı, Ankara, pp 88–116 Chalermchat Y, Fincan M, Dejmek P (2004) Pulsed electric field treatment for solid–liquid extraction of red beetroot pigment: mathematical modelling of mass transfer. J Food Eng 64:229–236. doi:10.1016/j.jfoodeng.2003.10.002 Chigurupati N, Saiki L, Gayser C, Dash AK (2002) Evaluation of red cabbage dye as a potential natural color for pharmaceutical use. Int J Pharm 241:293–299. doi:10.1016/S0378-5173(02) 00246-6 Corrales M, Toepfl S, Butz P et al (2008) Extraction of anthocyanins from grape by-products assisted by ultrasonics, high hydrostatic pressure or pulsed electric fields: a comparison. Innovative Food Sci Emerg Technol 9:85–91. doi:10.1016/j.ifset.2007.06.002 Donsì F, Ferrari G, Pataro G (2010) Applications of pulsed electric field treatments for the enhancement of mass transfer from vegetable tissue. Food Eng Rev 2:109–130. doi:10.1007/ s12393-010-9015-3 Dörnenburg H, Knorr D (1993) Cellular permeabilization of cultured plant tissues by high electric field pulses or ultra high pressure for the recovery of secondary metabolites. Food Biotechnol 7:35–48. doi:10.1080/08905439309549844 Eshtiaghi MN, Knorr D (2002) High electric field pulse pretreatment: potential for sugar beet processing. J Food Eng 52:265–272. doi:10.1016/S0260-8774(01)00114-5 Fincan M (2015) Extractability of phenolics from spearmint treated with pulsed electric field. J Food Eng 162:31–37. doi:10.1016/j.jfoodeng.2015.04.004 Fincan M, DeVito F, Dejmek P (2004) Pulsed electric field treatment for solid-liquid extraction of red beetroot pigment. J Food Eng 64:381–388. doi:10.1016/j.jfoodeng.2003.11.006 Gachovska T, Cassada D, Subbiah J et al (2010) Enhanced anthocyanin extraction from red cabbage using pulsed electric field processing. J Food Sci. doi:10.1111/j.1750-3841.2010.01699.x Hagiwara A, Yoshino H, Ichihara T et al (2002) Prevention by natural food anthocyanins, purple sweet potato color and red cabbage color, of 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP)-associated colorectal carcinogenesis in rats initiated with 1,2-dimethylhydrazine. J Toxicol Sci 27:57–68. doi:10.2131/jts.27.57 Knorr D, Angersbach A (1998) Impact of high-intensity electric field pulses on plant membrane permeabilization. Trends Food Sci Technol 9:185–191 Kolodziejczyk J, Saluk-Juszczak J, Posmyk MM et al (2011) Red cabbage anthocyanins may protect blood plasma proteins and lipids. Cent Eur J Biol 6:565–574. doi:10.2478/s11535011-0037-5 Lebovka NI, Bazhal MI, Vorobiev E (2002) Estimation of characteristic damage time of food materials in pulsed-electric fields. J Food Eng 54:337–346. doi:10.1016/S0260-8774(01)00220-5

22

M. Fincan

Lin JY, Li CY, Hwang IF (2008) Characterisation of the pigment components in red cabbage (Brassica oleracea L. var.) juice and their anti-inflammatory effects on LPS-stimulated murine splenocytes. Food Chem 109:771–781. doi:10.1016/j.foodchem.2008.01.039 Loginova KV, Lebovka NI, Vorobiev E (2011) Pulsed electric field assisted aqueous extraction of colorants from red beet. J Food Eng 106:127–133. doi:10.1016/j.jfoodeng.2011.04.019 López N, Puértolas E, Condón S et al (2009) Enhancement of the extraction of betanine from red beetroot by pulsed electric fields. J Food Eng 90:60–66. doi:10.1016/j.jfoodeng.2008.06.002 Luengo E, Álvarez I, Raso J (2014a) Improving carotenoid extraction from tomato waste by pulsed electric fields. Front Nutr 1:12. doi:10.3389/fnut.2014.00012 Luengo E, Condón-Abanto S, Álvarez I, Raso J (2014b) Effect of pulsed electric field treatments on permeabilization and extraction of pigments from Chlorella vulgaris. J Membr Biol 247:1269–1277. doi:10.1007/s00232-014-9688-2 Luengo E, Martínez JM, Álvarez I, Raso J (2016) Effects of millisecond and microsecond pulsed electric fields on red beet cell disintegration and extraction of betanines. Ind Crop Prod 84:28–33. doi:10.1016/j.indcrop.2016.01.016 Pakhomov AG, Miklavcic D, Markov MS (eds) (2010) Advanced electroporation techniques in biology and medicine. CRC Press, Boca Raton Roohinejad S, Everett DW, Oey I (2014) Effect of pulsed electric field processing on carotenoid extractability of carrot puree. Int J Food Sci Technol 49:2120–2127. doi:10.1111/ijfs.12510 Saulis G (2010) Electroporation of cell membranes: the fundamental effects of pulsed electric fields in food processing. Food Eng Rev 2:52–73. doi:10.1007/s12393-010-9023-3 Saulis G, Saule R (2012) Size of the pores created by an electric pulse: microsecond vs millisecond pulses. Biochim Biophys Acta Biomembr 1818:3032–3039 Silve A, Guimerà Brunet A, Al-Sakere B et al (2014) Comparison of the effects of the repetition rate between microsecond and nanosecond pulses: electropermeabilization-induced electro-desensitization? Biochim Biophys Acta Gen Subj 1840:2139–2151. doi:10.1016/j.bbagen.2014.02.011 Soliva-Fortuny R, Balasa A, Knorr D, Martin-Belloso O (2009) Effects of pulsed electric fields on bioactive compounds in foods: a review. Trends Food Sci Technol 20:544–556. doi:10.1016/j. tifs.2009.07.003 Weaver JC, Chizmadzhev YA (1996) Theory of electroporation: a review. Bioelectrochem Bioenerg 41:135–160

Pulsed Electric Fields Bioproduction of Secondary Metabolites in Plant Systems Pedro Elez-Martínez, Robert Soliva-Fortuny, and Olga MartínBelloso

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Plant Secondary Metabolism as Affected by Stress Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Phenolic Secondary Metabolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Terpenoid Secondary Metabolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Sulfur-Containing Secondary Metabolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 PEF Bioproduction of Secondary Metabolites in Plant Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 PEF Bioproduction of Secondary Metabolites in Plant Cell Cultures . . . . . . . . . . . . . . . . . . . . . . . 7 PEF Bioproduction of Secondary Metabolites in Plant Foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Abstract

Plant cell culture has received attention as an effective technology for the production of valuable secondary metabolites to be used as food, nutraceutical, pharmaceutical, and cosmetic ingredients and additives. Moreover, secondary metabolites of fruits and vegetables play a key role in reducing chronic disease risk. Pulsed electric fields (PEF) treatments are currently under study to prospect their potential to induce stress reactions in plant systems, so that bioproduction of certain compounds can be enhanced or stimulated. It has been demonstrated that PEF treatments can be applied as an abiotic stressor to elicit an increase of secondary metabolites in plant systems. Therefore, PEF treatments could be a feasible strategy for increasing the bioproduction of secondary metabolites in plant cell cultures in order to produce natural food additives and nutraceuticals,

P. Elez-Martínez (*) • R. Soliva-Fortuny • O. Martín-Belloso Department of Food Technology, Agrotecnio Center, University of Lleida, Lleida, Spain e-mail: [email protected]; [email protected]; [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_35-1

1

2

P. Elez-Martínez et al.

pharmaceuticals, and cosmetic ingredients and for promoting the antioxidant potential in raw fruits and vegetables prior to processing to the aim of developing healthier plant-based products. This chapter is focused in describing the use of PEF as an abiotic stressor for increasing the bioproduction of secondary metabolites in plant cell cultures and for obtaining fruit and vegetable products with high antioxidant potential.

Keywords

Pulsed electric fields • Secondary metabolites • Bioproduction • Abiotic stress • Plant systems • Fruits and vegetables

Introduction Plant secondary metabolites are of keen interest, because they have shown potential as food additives, nutraceuticals, pharmaceuticals, and cosmetic ingredients (Cai et al. 2012). Plant cell culture has received a lot of attention as an effective technology for the production of valuable secondary metabolites. Therefore, there is a need of exploring methods to increase secondary metabolites accumulation in plant cell culture medium. On the other hand, consumers are more and more concerned about the nutritional and health-related characteristics of fruits and vegetables. Secondary metabolites, such as phenolic compounds and carotenoids, play a key role in reducing chronic disease risk (Boyer and Liu 2004). The antioxidant compounds in some fruits and vegetables can be lost during handling after harvest, even during minimal processing and storage. In this sense, postharvest treatments are needed to preserve the quality and antioxidant potential of fresh produce. The application of postharvest abiotic stresses (i.e., wounding, UV-light radiation, modified atmospheres, exogenous phytohormones) has been proposed in recent years as an effective strategy to activate the secondary metabolism of fruits and vegetables leading to the accumulation of antioxidant compounds with health-promoting benefits (Becerra-Moreno et al. 2015). Pulsed electric fields (PEF) have been extensively studied for designing pasteurization applications for foods, so that they can constitute an alternative to traditional thermal processing to inactivate spoiling and pathogenic microorganisms and quality-related enzymes, with the advantage of retaining or minimally modifying sensorial, nutritional, and health-promoting attributes of liquid food products. Furthermore, PEF can also provide a potential to be used as a pretreatment to improve food processes such as extraction by pressing or solvent diffusion, osmotic dehydration, drying, and freezing. Finally, PEF treatments are currently under study to prospect their potential to induce stress reactions in plant systems or cell cultures, so that bioproduction of certain compounds can be enhanced or stimulated.

Pulsed Electric Fields Bioproduction of Secondary Metabolites in Plant. . .

3

Plant Secondary Metabolism as Affected by Stress Factors Primary metabolism in plant tissues comprises all metabolic pathways underpinning the generation of compounds that are essential to the plant’s survival. In parallel, secondary metabolism is responsible for the production of metabolites that are accumulated within plant tissues because of its biological function but are not essential for the plant survival. Plant secondary metabolites are useful in the long term, in most cases exhibiting biological activity related to the plant defense system. They are very often determining some important characteristics of the plant tissues, including color and aromatic profile. Secondary metabolites may also act as plant hormones, thus intervening in the signaling and regulation of the primary metabolism (Crawley and Harborne 2005). Plants have metabolic pathways involved in the synthesis of tens of thousands of secondary products. These are capable of responding to environmental situations regardless they have a biotic or an abiotic origin. It can be summarized that secondary metabolites help the plant maintain a balance with the surrounding environment, often contributing to the maintenance of tissue homeostasis and adapting to match the environmental needs. Although secondary metabolites are considered to be very specific when eliciting a biological effect, they may exhibit very different functions by acting at different levels. For instance, carotenoids are accumulated in some fruits and exhibit multiple functions; some compounds are related to photoprotection and others to the response against different sources of oxidative stress. Environmental stresses of biotic and abiotic nature have been often regarded as a threat to the production and quality of most agricultural crops. Biotic stress factors include microbial pathogenic agents (virus, bacteria, and fungi), animals, insects, and competition and interaction with other species. Regarding abiotic stress sources, in-field studies have shown that factors such as drought, salinity, light, extreme temperatures, heavy metals, mechanical wounding, and nutrients shortage may elicit different stress responses, thus causing a global reduction of crops and leading to worldwide economic costs (Nakabayashi and Saito 2015). Defense reactions against stress in plant tissue produce reactive oxygen species (ROS) such as singlet oxygen, superoxide anion radicals, hydrogen peroxide, and hydroxyl radicals in several cellular compartments by different mechanisms. Stress situations are counteracted by an increase in radical scavenging processes, which make the plant tissue not only capable to deal with oxygen stress but also to use oxygen activation as a defense system, which results in a respiratory burst (Grassmann et al. 2002). Plants have evolved two different biological processes to cope with ROS: (i) prevention or avoidance of ROS formation and (ii) scavenging of ROS by low molecular weight antioxidants through enzymatic and nonenzymatic processes (Nakabayashi and Saito 2015). Some antioxidant compounds are produced by primary biosynthetic pathways, while some others are generated by secondary metabolic pathways. Hence, very different molecules such as ascorbic acid, glutathione, tocopherols, amino acids, and sugars are presumed to have an antioxidant function in plant tissues, although the reasons that explain such variety of

4

P. Elez-Martínez et al.

compounds have not been yet fully elucidated. These antioxidant compounds are accumulated in the tissues as a consequence of the regular metabolism of plants. However, abiotic stresses also lead to the production of different specialized secondary metabolites. The concentrations of these compounds in different parts of the plant may significantly differ, as they are usually stored in those parts of the tissues that will require protection against oxidative stress. A system for classifying secondary metabolites has not been established yet. However, based on their biosynthetic origin, plant secondary metabolites can be classified into three different groups: (i) flavonoids and allied phenolic and polyphenolic compounds, (ii) terpenoids, and (iii) nitrogen-containing alkaloids and sulfur-containing compounds. Regarding cell structure, secondary metabolites are mainly found in vacuoles and in the inner part of other cytoplasmic organelles, such as chloroplasts, peroxisomes, cytosol, and mitochondria. Regulation of the production of these metabolites has been an important target in the bioengineering research fields. Despite the complexity of secondary metabolism, important advances have been achieved, leading to success stories in engineering the bioproduction of molecules of interest.

Phenolic Secondary Metabolites Phenolic compounds are a large group of secondary metabolites that appear to be mainly associated with plant defense strategies against different sources of stress, either biotic or abiotic. Phenolics are synthetized through three different pathways: (i) the shikimate pathway, which produces phenylpropanoid derivatives; (ii) the acetate pathway, which produces side-chain-elongated phenylpropanoids, including flavonoids and some quinones; and (iii) the acetate/mevalonate pathway, which produces aromatic terpenoids by dehydrogenation reactions (Knaggs 2001). Among phenolic metabolites, phenylpropanoids constitute the most important group of compounds. They are frequently water soluble compounds, as they tend to be conjugated with sugars. All phenylpropanoids are derived from cinnamic acid, which is synthetized from phenylalanine by the action of phenylalanine ammonialyase (PAL), the branch point enzyme between primary (shikimate pathway) and secondary (phenylpropanoid) metabolism. A few cytochrome P450 proteins have been identified to mediate in the biosynthesis of phenolic compounds. This biosynthesis is a good example of metabolic network, where many of the compounds may be obtained from different intermediate products, and a single intermediate product may be transformed into different products depending on the active enzymatic process (Quattrocchio et al. 2006). The phenylpropanoids family includes different groups of compounds, namely, hydroxycinnamic acids, cinnamic aldehydes and monolignols, coumarins, flavonoids, and stilbenoids. Flavonoids are by far the largest class of polyphenols, comprising over 8,000 metabolites that can be further subdivided into six major subclasses, namely, flavones, flavonols, flavanones, flavanols, anthocyanidins, and isoflavones (Harborne and Williams 2000). In plant tissues, the levels of

Pulsed Electric Fields Bioproduction of Secondary Metabolites in Plant. . .

5

phenylpropanoids have been shown to greatly fluctuate in response to changes in numerous environmental factors influenced by light, genetics, germination, ripening, and postharvest characteristics (Pollastri and Tattini 2011). The upregulation of the phenylpropanoid metabolism in response to tissue wounding is one of the hallmarks of biotic stress, being widely observed in plants. Nevertheless, upregulation of the phenylpropanoid metabolism in response to abiotic stress factors has also been reported. The photoprotective role of flavonoids is probably one of the most discussed, but their function against other abiotic stress factors such as nitrogen deficiency and high carbon growth conditions has been also observed. Beyond their function in plant tissues, phenylpropanoids exhibit several biological functions in the human body because of their neuroprotective, anti-inflammatory, analgesic, bactericidal, fungicidal, and spasmolic properties, which have been extensively reviewed in in vitro studies, and their well-documented preventive role against chronic diseases (Harborne and Williams 2000; Quattrocchio et al. 2006). The use of abiotic elicitors, such as chemicals or physical stimuli, has been proposed to stimulate the accumulation of valuable phenolic secondary metabolites in plant tissues. The ability of chemical compounds, namely, metal ions and inorganic compounds, to elicit the biosynthesis of phenolics in plants is well known. However, because elicitors must be added to the tissue or cell culture producing the desired secondary metabolite and needs to be separated thereafter, the use of physical stress factors is being studied for the clean synthesis of these metabolites.

Terpenoid Secondary Metabolites Terpenoids, also called isoprenoid compounds, encompass a group of compounds that greatly differ in both structure and function, thus acting, either internally or externally, as allelopathic agents, repellents, or attractants in plant-plant or plantpathogen/herbivore interactions (Grassmann et al. 2002). Other functions in defense and wound sealing and protection against extreme temperatures and light have been reported. These are highly dependent on the structure or the compound. For instance, carotenoids have a tetraterpenoid structure and are involved in light capture and protection against photooxidative damage, whereas sterols are generally considered to regulate membrane fluidity and permeability. It is not clear whether terpenoids possess an antioxidative function in plants. Some terpenoids or their precursors have though demonstrated radical scavenging capacity against ROS in model systems and particularly against lipid peroxidation as a result of their high lipophilicity. This is consistent with the fact that terpenoids are the basis for many herbal drugs used in the treatments of several degenerative and chronic diseases in which ROS may be involved. Plant terpenoids are majorly biosynthesized through the mevalonic acid pathway, which begins with the conversion of acetyl-coenzyme A to the isoprene precursor, isopentenyl pyrophosphate, from which higher order terpenoid building blocks geranyl pyrophosphate, farnesyl pyrophosphate, and geranylgeranyl pyrophosphate. These in turn may self-condense or undergo internal addition to create the parent

6

P. Elez-Martínez et al.

skeletons of the various terpenoid families through subsequent oxidation, reduction, isomerization, conjugation, or other secondary transformations (McGarvey and Croteau 1995). However, while metabolic pathways are increasingly better described, the effects of environmental factors remain insufficiently understood and quantified. Available studies suggest that terpenoid concentrations in plant tissues may be promoted by appropriately tuning environmental constraints. In this regard, De Pascale et al. (2003) suggested that moderate salinity stress conditions may positively affect the accumulation of lycopene in tomato fruits. Furthermore, in-field experiments have shown that exposure to photooxidative stress of adjacent leaves strongly influences the biosynthesis of carotenoids in orange fruits (Poiroux-Gonord et al. 2013). Regarding postharvest treatments, the available information is even more limited, although preliminary studies seem to point out that abiotic stress factors, namely, mild heat conditions, may be used to successfully increase the concentration of isoprenoid compounds in agricultural crops after harvest (Brecht et al. 2010).

Sulfur-Containing Secondary Metabolites Sulfur-containing compounds are crucial metabolites for the defense of plants against biotic and abiotic stresses. These compounds include elemental sulfur, H2S, glutathione, phytochelatins, sulfur-rich proteins, and various secondary metabolites. Their stress-induced formation is intimately dependent on the regulation of sulfate uptake and assimilation by plants. The main secondary metabolites with interest regarding their health-related biological functions are glucosinolates and S-alkyl cysteine sulfoxides. Although these metabolites are rather unusual in plant tissues, they play a major defensive role in Brassicaceae and Alliaceae plants. For instance, although some insect species may be attracted and stimulated by glucosinolates, their negative biological effects have been reported for a broad range of potential enemies, including insects, bacteria, and fungi. This biological activity has been largely associated to isothiocyanates, the hydrolysis products formed by the action of the enzyme myrosinase, rather than the glucosinolates themselves. These same compounds account for a significant contribution to health-promoting properties in humans. The amount of glucosinolates has been shown to be determined by several environmental factors, including fertirrigation, salinity, light cycling, or extreme temperatures, as well as the occurrence of abiotic or biotic stress sources, such as pathogens, insect herbivores, salicylic acid, and jasmonates (Mithen 2007; Martínez-Ballesta et al. 2013).

PEF Bioproduction of Secondary Metabolites in Plant Systems PEF may cause lethal damage to cells or induce sublethal stress by permeabilizing tissue structures (Soliva-Fortuny et al. 2009). Metabolic responses of plant cells (Cai et al. 2011) and tissues (Galindo et al. 2008, 2009) upon the application of PEF have

Pulsed Electric Fields Bioproduction of Secondary Metabolites in Plant. . .

7

been studied. PEF-induced stresses could include a burst of ROS, which are endogenous signal components required for synthesis of secondary metabolites, which are known to be part of the defense response of plants to stress. Therefore, PEF has been proposed as a promising novel abiotic stressor for stimulating the secondary metabolites biosynthesis and accumulation in plant cells and tissues.

PEF Bioproduction of Secondary Metabolites in Plant Cell Cultures The effects of PEF on secondary metabolite production by Taxus chinensis cell culture were investigated by Ye et al. (2004). Cultured cells in different growth phases were exposed to a PEF (50 Hz, 10 V/m) for various periods of time. A significant increase in intracellular and extracellular accumulation of taxuyunnanine C (Tc), a bioactive secondary metabolite, was observed by exposing the cells in the early exponential growth phase to a 30-min PEF. The Tc content (i.e., the specific production based on dry cell weight) was increased by 30 % after exposure to PEF, without loss of biomass, compared with the control. Production levels of reactive oxygen species, extracellular Tc, and phenolics were all increased, whereas cell capacitance was decreased with PEF treatment. Therefore, these results showed that PEF induced a defense response of plant cells and may have altered the cell/ membrane’s dielectric properties. Gueven and Knorr (2011) investigated the production of aglycone-type isoflavonoids, genistein and daidzein, and β-glycosidic-type isoflavonoids, genistin and daidzin, by the soy plant callus suspension culture in a batch system. Isoflavonoid production in the 7-day-old culture increased with PEF application at 1.6 kV, and aglycone forms were influenced to a greater extent. Further PEF applications were conducted between 1.2 and 2.4 kV so as to find out the influence of PEF on isoflavonoid biosynthesis. PEF application between 1.6 and 2.0 kV increased isoflavonoid concentrations. Further increase in voltage resulted in a decrease in isoflavonoid concentration. These authors attributed the increase of isoflavonoid concentrations in the soy plant callus suspension culture to the influence of PEF on membrane integrity besides protein and enzyme structures and activities. Protein channels located within the cell membrane may have been affected by application of 1.6–2.0 kV resulting in stress conditions and also allowed excretion of the secondary metabolites produced. The effects of PEF (1.6 kV/cm, ten pulses) on secondary metabolism, particularly anthocyanins and phenolic acids synthesis, were investigated by using suspension culture of Vitis vinifera L. cv. Gamay Fréaux (Cai et al. 2011). The treatment of PEF enhanced the total anthocyanins production, especially on days 9 and 14 after the treatment. The stimulating activity was 30 % and 71 %, respectively. The stimulating activity was defined as the ratio of average accumulation titer under treatment to that of control. However, the content of anthocyanins decreased after day 4. Therefore, the biomass was taken into consideration when they tried to get the final production of anthocyanins. For both the control and PEF-treated samples, the content of total extracellular phenolics reached the highest point on day 9 after the treatment and

8

P. Elez-Martínez et al.

declined quickly thereafter. After PEF treatment, extracellular phenolics content was found to accumulate rapidly after day 7 and reached 11 % higher than that of control on the ninth day after PEF treatment. The results indicate that PEF caused a release of secondary metabolites into the medium, which may be related to the changes of the cell membrane dielectric properties. 3-O-glucosyl-resveratrol accumulated to the highest level on day 4 after the treatments. The 3-O-glucosyl-resveratrol content of PEF treatment was 3.6-fold of the control on day 4 after the treatments. However, the content decreased after 4 days. Saw et al. (2012) also studied the production of anthocyanins in cell suspension culture of V. vinifera after treatment with PEF (ten pulses, 1.6 kV/cm, 1 Hz, 0.32 J/kg total specific energy). The maximum anthocyanin synthesis in the PEF-treated samples occurred on day 4 (2.3 mg/g DW), and it decreased after day 4 in both the PEF-treated and untreated samples. On average, the anthocyanin content in the PEF-treated samples was 1.25-fold larger than that in the untreated samples. By day 14, the anthocyanin concentration had decreased by approximately 50 % in both the treated and untreated samples. These results are in concordance with those reported by Cai et al. (2011). Gürsul et al. (2016) observed that PEF can be used to improve phenylalanine ammonia lyase (PAL) activity and thus to increase secondary metabolite biosynthesis by tomato cell cultures. PEF application did not influence PAL activity in secondary subcultures, while its effect on the primary subcultures was significant. In general, PAL activity increased with the increase in the strength and extent of PEF application. Maximum PAL activity was observed for the PEF application of nine pulses at 1.2 kV/cm and 4 h after treatment of the primary subculture. On the other hand, PEF application had no significant effect on biosynthesis of phenolic compounds by the primary subculture in terms of the total phenolic content. Conversely, PEF application had a significant effect on biosynthesis of phenolic compounds by the secondary subcultures. PEF application had no significant effect on the stress reaction time for all subcultures. Maximum amount of total phenolics was obtained for the secondary subcultures at 1.2 kV/cm with one and nine pulses at 4 h after treatment. Total phenolic content of secondary subcultures was higher than the primary subcultures. Thus, subculturing enhanced the cell structure and led to stress resistance against PEF. Consequently, the combination of PEF and subculturing provides a potential to enhance the biosynthesis of phenolic compounds by tomato cell cultures. The effects of combining PEF, as physical stressor and chemical stressors, were also studied for enhancing the secondary metabolite production of plant cell cultures. The addition of 20 g/L sucrose to Taxus chinensis cell cultures immediately after cells were exposed to PEF increased the maximum taxuyunnanine C production up to 40 % in respect to that of untreated (Ye et al. 2004). Therefore, the combination of PEF treatment with sucrose feeding was shown to be effective in improving secondary metabolite production. Cai et al. (2011) and Saw et al. (2012) investigated the effects of PEF and ethephon on the secondary metabolites production in cell cultures of Vitis vinifera L. cv. Gamay Fréaux. Treatment of cell suspension with PEF (1.6 kV/cm, ten pulses) at day 14 of culture resulted in 1.7-fold increase in anthocyanin content when compared to control cells, while treatment with ethephon

Pulsed Electric Fields Bioproduction of Secondary Metabolites in Plant. . .

9

(28 mg/L) resulted in 2.3-fold increase in anthocyanin content. When cells were treated with both ethephon and PEF, 2.5-fold increase in anthocyanin content was observed. The 3-O-glucosyl-resveratrol content of PEF, ethephon, combination of PEF, and ethephon treatments was 3.6-, 14.3-, and 20.1-fold of the control on day 4 after the treatments, respectively. The combination of PEF treatment and ethephon improved secondary metabolites formation in V. vinifera cell cultures.

PEF Bioproduction of Secondary Metabolites in Plant Foods Guderjan et al. (2005) studied the additional production of secondary metabolites in soybeans. The maximum amount of daidzein, reached with 50 pulses of 280 μs at an electric field of 1.3 kV/cm and an energy impact of 1.857 kJ/kg after 24 h of treatment, means an increase of 20 % in comparison to the reference. The maximum amount of genistein could be reached with 20 pulses of 280 μs, 0.743 kJ/kg, and represents a rise of 21 % in comparison to the untreated. Therefore, an increase yield of isoflavonoids in soybeans (20–21 %) was observed after 24 h of application of PEF treatments with intensities in a range of reversible cell permeabilization. The effects of PEF treatments (0.4–2.0 kV/cm, 5–30 pulses, 4 μs pulse width) on the bioactive compounds (total polyphenol, lycopene, and vitamin C content) as well as on the antioxidant capacity of tomato fruit after 24 h of PEF treatment at 4  C were investigated by Vallverdú-Queralt et al. (2012a). A concentration of bioactive compounds higher than that of untreated tomatoes was obtained in PEF-treated tomatoes. A 44 % increase in total polyphenol content was achieved under 30 pulses at 1.2 kV/cm, whereas the maximum increase in lycopene content (32 %) was observed at 1.2 kV/cm and five pulses. The hydrophilic antioxidant capacity was also enhanced by 44 % applying 18 pulses at 1.2 kV/cm, and the lipophilic antioxidant capacity was increased by 37 % under five pulses at 1.2 kV/cm. The maximum overall level of bioactive compounds and antioxidant capacity in the treated tomatoes was obtained under 16 pulses at 1 kV/cm. In the same line, Vallverdú-Queralt et al. (2013a) used a metabolite profiling approach to study the effect of PEF treatments (0.4–2.0 kV/cm, 5–30 pulses, 4 μs pulse width) on the individual polyphenol and carotenoid contents of tomato fruit after refrigeration at 4  C for 24 h. Twenty-four hours after PEF treatments, an increase was observed in hydroxycinnamic acids and flavanones, whereas flavonols, coumaric, and ferulic acid-O glucoside were not affected. Major changes were also observed for carotenoids, except for the 5-cis-lycopene isomer, which remain unchanged after 24 h of PEF treatments. PEF treatments, conducted at 1.2 kV/cm and 30 pulses, led to the greatest increases in chlorogenic (152 %), caffeic acid-O-glucoside (170 %), and caffeic (140 %) acids. On the other hand, treatments at 1.2 kV/cm and five pulses led to maximum increases of α-carotene, 9-cis-lycopene, and 13-cis-lycopene, which increased by 93 %, 94 %, and 140 %, respectively. Vallverdú-Queralt et al. (2012a, 2013a) attributed their results to a PEF-induced stress response but also to an increased permeability of the cellular membrane due to PEF processing, which could potentially made the extraction of bioactive compounds more efficient. PEF

10

P. Elez-Martínez et al.

treatments may induce stress reactions in tomato fruits after 24 h of refrigeration by stimulating metabolic activity and accumulating secondary metabolites. Hence, PEF treatments could contribute to the production of tomatoes with high health-related compounds content. Leong et al. (2016) investigated the influence of PEF (0.5–2 kV/cm, 100 pulses of 20 μs) treatment of wheatgrass (Triticum aestivum L.) seeds, with different water contents, on antioxidant metabolism in the resultant seedlings. The total vitamin C and phenolic contents of seedlings were not significantly influenced by any PEF treatment. Moreover, no significant differences in the total antioxidant capacity were found when extracts from the PEF-enhanced shoots were compared to extracts from control shoots, despite the significant changes in shoot antioxidant metabolism observed. These authors stated that the most likely reason for this is the presence of relatively high levels of chlorophyll, vitamin C, phenolics, and flavonoids in the extracts that would mask any increase in the size of the glutathione pool. The potential of PEF treatments as a feasible strategy for promoting the antioxidant potential in raw materials prior to processing with the aim of developing healthier plant-based products has been studied. Guderjan et al. (2005) reported an increase in the phytosterol content of maize germ oil (up to 32.4 %) when maize germ was further processed after 24 h of PEF treatment (0.6 kV/cm, 120 pulses, 0.62 kJ/kg). Phenolics and carotenoids profile contents of tomato juices made from PEF-treated (1 kV/cm, 16 pulses) tomatoes after 24 h at 4  C were studied during refrigerated storage by Vallverdú-Queralt et al. (2012b, 2013b). In comparison to those untreated, tomatoes subjected to PEF treatments led to juices with a higher content of polyphenol compounds. A slight decrease in polyphenol compounds was observed over time in thermal- and PEF-treated juices made from tomatoes treated by PEF, with the exception of caffeic acid. However, PEF-processed tomato juices had a higher content of polyphenol compounds (ferulic acid, caffeic-O-glucoside acid, p-coumaric acid, chlorogenic acid, rutin, and naringenin) just after processing and through storage than those thermally treated (Vallverdú-Queralt et al. 2012b). PEF treatment of tomatoes increased the content of carotenoid compounds in tomato juices. An enhancement of 63–65 % in 15-cis-lycopene was observed in juices prepared with PEF-treated tomatoes. The content of individual carotenoids in fresh, thermal, and PEF-treated tomato juices underwent a substantial loss during storage (10–40 %), with the exception of cis-lycopene isomers, which were increased. Moreover, PEF-processed tomato juices better maintained the translycopene, lutein, and α- and β-carotene just after processing and during the storage period than heat-treated and fresh juices (Vallverdú-Queralt et al. 2013b).

Conclusions PEF treatments can be applied as an abiotic stressor to elicit an increase of secondary metabolites in plant systems. The stimulation of the secondary metabolism of plant cells and tissues can be optimized under selected PEF conditions in order to enhance the accumulation of secondary metabolites in plant systems. A proper combination

Pulsed Electric Fields Bioproduction of Secondary Metabolites in Plant. . .

11

of PEF treatment intensity and time and temperature after treatment is crucial to achieve positive effects. The extent of the changes depended on the nature of the secondary metabolites at stake, thus revealing the specificity of PEF when applied as an abiotic source of stress. PEF treatments could be a feasible strategy for (i) increasing the bioproduction of secondary metabolites in plant cell cultures in order to produce natural food additives, nutraceuticals, pharmaceuticals, and cosmetic ingredients and (ii) promoting the antioxidant potential in raw fruits and vegetables prior to processing to the aim of developing healthier plant-based products. However, more research and development activities are required to understand, optimize, and apply this complex process to its full potential. Studies about the effects of PEF treatment critical parameters on the secondary metabolite content of plant systems need to be extensively carried out. In-depth research is required in order to study the kinetics of bioproduction, accumulation, or degradation of plant systems secondary metabolites as affected by PEF treatment conditions, as well as to elucidate the mechanistic insight of the induced changes.

References Becerra-Moreno A, Redondo-Gil M, Benavides J, Nair V, Cisneros-Zevallos L, Jacobo-Velázquez DA (2015) Combined effect of water loss and wounding stress on gene activation of metabolic pathways associated with phenolic biosynthesis in carrot. Front Plant Sci 6:837. doi:10.3389/ fpls.2015.00837 Boyer J, Liu RH (2004) Apple phytochemicals and their health benefits. Nutr J 3:1–15 Brecht JK, Cherukuri K, Simonne AH, Tulio AZ, Nunes MCN (2010) Cultivar variability in the response of tomato fruit lycopene synthesis and antioxidant potential to a brief postharvest heat stress. Acta Hortic 11:1205–1214 Cai Z, Riedel H, Saw NMMT, Kütük O, Mewis I, Jäger H, Knorr D, Smetanska I (2011) Effects of pulsed electric field on secondary metabolism of Vitis vinifera L. cv. Gamay Fréaux suspension culture and exudates. Appl Biochem Biotechnol 164:443–453 Cai Z, Kastell A, Knorr D, Smetanska I (2012) Exudation: an expanding technique for continuous production and release of secondary metabolites from plant cell suspension and hairy root cultures. Plant Cell Rep 31:461–477 Crawley MJ, Harborne JB (2005) Plant secondary metabolism. In: Crawley M (ed) Plant ecology. Blackwell, New York, pp 132–155 De Pascale S, Angelino G, Graziani G, Maggio A (2003) Effect of salt stress on water relations and antioxidant activity in tomato. Acta Hortic 614:39–46 Galindo FG, Wadsö L, Vicente A, Dejmek P (2008) Exploring metabolic responses of potato tissue induced by electric pulses. Food Biophys 3:352–360 Galindo FG, Dejmek P, Lundgren K, Rasmusson AG, Vicente A, Moritz T (2009) Metabolomic evaluation of pulsed electric field-induced stress on potato tissue. Planta 230:469–479 Grassmann J, Hippeli S, Elstner EF (2002) Plant’s defence and its benefits for animals and medicine: role of phenolics and terpenoids in avoiding oxygen stress. Plant Physiol Biochem 40:471–478 Guderjan M, Töpfl S, Angersbach A, Knorr (2005) Impact of pulsed electric field treatment on the recovery and quality of plant oils. J Food Eng 67:281–287 Gueven A, Knorr D (2011) Isoflavonoid production by soy plant callus suspension culture. J Food Eng 103:237–243 Gürsul I, Gueven A, Grohmann A, Knorr D (2016) Pulsed electric fields on phenylalanine ammonia lyase activity of tomato cell culture. J Food Eng 188:66–76

12

P. Elez-Martínez et al.

Harborne JB, Williams CA (2000) Advances in flavonoid research since 1992. Phytochemistry 55:481–504 Knaggs AR (2001) The biosynthesis of shikimate metabolites. Nat Prod Rep 18:334–355 Leong SY, Burritt DJ, Oey I (2016) Electropriming of wheatgrass seeds using pulsed electric fields enhances antioxidant metabolism and the bioprotective capacity of wheatgrass shoots. Sci Rep 6:25306 Martínez-Ballesta MC, Moreno DA, Carvajal M (2013) The physiological importance of glucosinolates on plant response to abiotic stress in Brassica. Int J Mol Sci 14:11607–11625 McGarvey DJ, Croteau R (1995) Terpenoid metabolism. Plant Cell 7:1015–1026 Mithen R (2007) Sulfhur-containing compounds. In: Crozier A, Clifford MN, Ashihara H (eds) Plant secondary metabolites: occurrence, structure and role in the human diet. Blackwell, New York, pp 25–46 Nakabayashi R, Saito K (2015) Integrated metabolomics for abiotic stress responses in plants. Curr Opin Plant Biol 24:10–16 Poiroux-Gonord F, Santini J, Fanciullino AL, Lopez-Lauri F, Giannettini J, Sallanon H, Berti L, Urban L (2013) Metabolism in orange fruits is driven by photooxidative stress in the leaves. Physiol Plant 149:175–187 Pollastri S, Tattini M (2011) Flavonols: old compounds for old roles. Ann Bot-Lond 108:1225–1233 Quattrocchio F, Baudry A, Lepiniec L, Grotewold E (2006) The regulation of flavonoid biosynthesis. The science of flavonoids. Springer, New York, pp 97–122 Saw NMMT, Riedel H, Cai Z, Kütük O, Smetanska I (2012) Stimulation of anthocyanin synthesis in grape (Vitis vinifera) cell cultures by pulsed electric fields and ethephon. Plant Cell Tissue Organs Cult 108:47–54 Soliva-Fortuny R, Balasa A, Knorr D, Martín-Belloso O (2009) Effects of pulsed electric fields on bioactive compounds in foods: a review. Trends Food Sci Technol 20:544–556 Vallverdú-Queralt A, Oms-Oliu G, Odriozola-Serrano I, Lamuela-Raventos RM, Martín-Belloso O, Elez-Martínez P (2012a) Effects of pulsed electric fields on the bioactive compound content and antioxidant capacity of tomato fruit. J Agric Food Chem 60:3126–3134 Vallverdú-Queralt A, Oms-Oliu G, Odriozola-Serrano I, Lamuela-Raventos RM, Elez-Martínez P, Martín-Belloso O (2012b) Changes in the polyphenol profile of tomato juices processed by pulsed electric fields. J Agric Food Chem 60:9667–9672 Vallverdú-Queralt A, Oms-Oliu G, Odriozola-Serrano I, Lamuela-Raventos RM, Martín-Belloso O, Elez-Martínez P (2013a) Metabolite profiling of phenolic and carotenoid contents in tomatoes after moderate-intensity pulsed electric field treatments. Food Chem 136:199–205 Vallverdú-Queralt A, Oms-Oliu G, Odriozola-Serrano I, Lamuela-Raventos RM, Elez-Martínez P, Martín-Belloso O (2013b) Impact of high-intensity pulsed electric fields on carotenoids profile of tomato juice made of moderate-intensity pulsed electric field treated tomatoes. Food Chem 141:3131–3138 Ye H, Huang LL, Chen SD, Zhong JJ (2004) Pulsed electric field stimulates plant secondary metabolism in suspension cultures of Taxus chinensis. Biotechnol Bioeng 88:788–795

Microalgal Biorefinery for Bulk and High-Value Products: Product Extraction Within Cell Disintegration P. R. Postma, G. P. ‘t Lam, M. J. Barbosa, R. H. Wijffels, M. H. M. Eppink, and G. Olivieri

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microalgal Biorefinery: Market Opportunities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Process Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Overview of Algal Biorefinery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bead Milling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . High Pressure Homogenization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ultrasound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Supersonic Flow Fluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pulsed Electrical Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microwave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 2 5 5 7 9 10 11 12 14 14 15

P.R. Postma • G.P. ‘t Lam • M.J. Barbosa • M.H.M. Eppink Bioprocess Engineering, AlgaePARC, Wageningen University & Research, Wageningen, The Netherlands e-mail: [email protected]; [email protected]; [email protected]; [email protected] R.H. Wijffels Bioprocess Engineering, AlgaePARC, Wageningen University & Research, Wageningen, The Netherlands Faculty of Biosciences and Aquaculture, Nord University, Bodø, Norway e-mail: [email protected] G. Olivieri (*) Bioprocess Engineering, AlgaePARC, Wageningen University & Research, Wageningen, The Netherlands Dipartimento di Ingegneria Chimica, dei Materiali e della Produzione Industriale, Università degli Studi di Napoli Federico II, Napoli, Italy e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_38-1

1

2

P.R. Postma et al.

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Abstract

Microalgae are a promising source for proteins, lipids, and carbohydrates for the cosmetic, nutraceutical, chemical, food/feed, and biofuel industry. In comparison with soy and palm oil, microalgae can be produced in a more sustainable way. To make microalgae production economically feasible, all biomass ingredients need to be efficiently utilized, similar to petroleum refineries in which oil is fractionated in fuels and a variety of products with higher value. However severe conditions can affect the properties of some components in the biomass. To overcome this, focus needs to be put on biorefinery techniques which are mild and effective. Microalgal biorefinery is a linear process consisting of harvesting, cell disintegration, sequential extraction, and further fractionation. Among these steps, the cell disintegration often represents a bottleneck for the extraction of hydrophilic or hydrophobic components, due to the presence of a tough cell wall in many strains. State of the art knowledge on both novel and classical techniques for product extraction within cell disintegration is presented. Comparison is made on the basis of two main criteria: yield of disintegration and energy consumption. The current work gives also a comprehensive outlook on business cases for microalgae biorefinery. Keywords

Microalgal biorefinery • Cell disintegration • PEF • Protein extraction • Energy consumption

Introduction Microalgal Biorefinery: Market Opportunities Microalgae are very attractive as a feedstock for biobased products due to an aerial productivity superior to traditional agricultural crops: realistic estimates for areal productivity are in the order of magnitude of 40–80 t of dry matter per hectare per year depending on the technology used and location of production (Tredici 2010). In addition algae can have a low water footprint, and production does not compete with agriculture land (Tredici 2010). Microalgae have been of major interest for producing biofuels in the last decade (Chisti 2007). However, at this moment, microalgae production for biofuel production appears to be still too costly with current process (Wijffels and Barbosa 2010). Depending on the species and cultivation conditions, microalgae can accumulate high amounts of lipids, proteins, and carbohydrates, which can be used for different markets such as bulk and high added value products (Fig. 1) (Vigani et al. 2015).

Microalgal Biorefinery for Bulk and High-Value Products: Product Extraction. . .

3

Pharma TAG Glyco-lipids Phospho-lipids

Lipids CO2

Waxes Sterols

Water Proteins Microalgae

Un-soluble proteins

N, P, S, Minerals Energy

Energy

Mono- Oligosaccharides

Carbohydrates

Starch Cellulose

Labour Labour

Pigments

Chlorophyll Carotenoid Phycobilins

Ash/minerals

Food/Feed Paint, Coating

Selling price

Land/ Sunlight

Food additives

Biolubricant Surfactants

Market demand

Soluble proteins

Cosmetic Health care

Biopolymer Bulk chemicals Biodiesel Biokerosene Bioethanol Biobutanol Biomethane

Fig. 1 Overall spectrum of microalgal component and their possible application Table 1 Lipids and proteins supply, import, price, and production data in EU in 2013

Lipids Proteinsa

Total supply 106 MT 27 56

Production

Import

15 18

12 38

Price €/MT 980 310

Cultivated area 106 ha 17 12

Productivity MT/ha 0.9 1.5

Sources: US Dept Agriculture and FAOSTAT Data are referred to whole meal containing from 30 % (sunflower) up to 60 % (rapeseed) of pure proteins

a

Lipids and proteins are the most interesting fractions of the microalgae and their concentration is strongly dependent on the operation conditions during cultivation. Globally the need for lipids and proteins as food, feed, and fuel is especially rising in Europe, where currently 44 % of the lipid and 68 % of the protein requirement is imported (Table 1). In addition, Europe is the only continent where the amount of arable land is strongly decreasing along the years (source: FAOSTAT), indicating microalgae production as a more efficient source of lipids and proteins. However, microalgae are nowadays only produced and commercialized for niche markets, either as whole biomass (food additives and feed for aquaculture) or as extracted valuable components (astaxanthin, beta-carotene, omega-3 fatty acids, and phycobiliproteins), with a very low market volume (10,000 MT/y) (Vigani et al. 2015). When exploiting the whole potential of microalgae components in an overall biorefinery strategy, many different products have to be extracted and purified within

TAG GlycoPhospho-

Cosmetic Health care

Food additives

Food/Feed

Paint, Coating

Boilubricant

Boipolymer

Bulk chemicals

Bioethanol

Biokerosene

1.5 2-3.5

0.37

0.56

0.71

MonounstaredFA

0.95

5-10

Satured FA PolyunsaturedFA Waxes

2.5-15

20-30

23-70

20-40

3-5

4

Sterols

1

1.5

1.5

Chlorophyll Carotenoids

2.5-4.6

2002,000

0.21

Starch Cellulose

0.17

0.22

Mono- Oligosaccharides

1.4

0.75

1

1.5-3.7

Soluble proteins Non-soluble prot.

1.1

2.3

2002,000

€/kg

Biodiesel

P.R. Postma et al.

Fertilizers

4

Phycobilins Ash/minerals

0.025 0.15-0.45 €/kg

1.5-2.5 €/kg

3-10 €/kg

8-10 €/kg

Fig. 2 Selling prices of microalgal components in different market scenarios and derived overall biomass revenue

the process in order to turn the potential selling price of the microalgal biomass higher than the production and extraction costs (Fig. 2) (Wijffels et al. 2010). Downstream processing costs are an important part of the total production costs (Coons et al. 2014). Some species have a tough cell wall, which makes them robust for outdoor cultivation, but also requires hard conditions to extract the intracellular components. Literature addressing the market potential of microalgal components is for a large part focused on production of one specific product from the biomass (e.g., lipids). Therefore, ad-hoc extraction methods were only developed for one specific product and the other available and valuable components in the microalgae were not valorized. To be able to exploit the complete microalgae biomass, it is necessary to use mild cell disintegration techniques. Conventional disintegration, e.g., bead milling, homogenizers, high pressure, heating, osmotic shock, and chemicals, are not considered as mild. In addition, they are mainly used to obtain one final product, while damaging the other fractions. As an example, organic solvents, commonly used in lipid extraction techniques, would harm and denature proteins. In this way large part of soluble proteins should become insoluble and lose their techno-functional properties.

Microalgal Biorefinery for Bulk and High-Value Products: Product Extraction. . .

5

Biorefinery is a procedure that integrates biomass conversion and separation, in which the objective is to obtain several fractions/products by using mild separation from one single source. To do that, the biorefinery techniques appropriate for mild extraction are relatively new and should therefore be studied thoroughly before commercial use is possible (Wijffels et al. 2010).

Process Design Overview of Algal Biorefinery There are algal-production scenarios in which the biomass is considered as an end-product, and there are scenarios in which biorefinery of the biomass into specific components is used. When one or multiple specific components are desired as end product, a further downstream process is required. Such a downstream process typically consists out of a harvesting step, cell disintegration, extraction, and possibly a further fractionation (in particular when an integrated biorefinery is foreseen). In general, in downstream processing both harvesting and cell disintegration are recognized as expensive and therefore the cost limiting process steps. The high harvesting expenditures are often attributed to the low biomass concentrations during cultivation, that commonly range between 0.3 g L 1 and 4 g L 1. In addition, microalgal cells have a small cell size (up to 20 μm) and commonly grow as single cells. Resultantly, a large amount of water needs to be separated from the biomass. Cell disintegration aims to permeabilize or completely break the cell wall and membrane to allow a simple extraction or release of intracellular components. As microalgae commonly grow as single cells, or in small colonies, they typically have a well-developed cell wall and membrane that serves as a protecting boundary. It is for this reason that cell disintegration is next to harvesting an operational (“energy”) intensive process. For both harvesting and cell disintegration, mechanical processes (centrifugation and filtration for harvesting and bead milling and homogenization for cell disintegration) are considered well-established technologies that are able to combine a high throughput, with a constant efficiency. Next to those established technologies, alternative technologies have been proposed such as flocculation for harvesting and pulsed electric field as cell disintegration technology (Toepfl et al. 2006; ‘t Lam et al. 2015). After harvesting and disintegrating the biomass, depending on the foreseen end-product(s), a product separation using extraction and a possible further fractionation are applied. For extracting lipids, several organic solvent are suitable. Generally, methanol/ chloroform shows the best yield due to the best polarity index of the mixture for extracting both the lipid classes (Halim et al. 2012). On industrial scale, hexane is frequently preferred for oil extraction from oilseeds.

6

P.R. Postma et al.

The use of organic solvents has two main drawbacks: (i) it denatures proteins losing their functionality; (ii) large quantity of solvent strongly affects the economy of the process due to the high energy demand for recovering it by distillation. In addition, the extraction has lower performance in case of wet biomass. So, additional cost of pre-drying the biomass should be included in the overall economic balance. On the other hand, pre-drying might cause side effects like decreased solubility of proteins. Besides the already available solvents such as Hexane and “Bligh and Dyer” for hydrophobic components like lipids and pigments, also novel solvents like ionic liquids, supercritical fluids and switchable solvents have been proposed (Herrero and Ibáñez 2015; Yen et al. 2014; Desai et al. 2016). In the study of Desai et al. (2016), successful extraction of the hydrophobic antioxidant astaxanthin from the microalgae Haematococcus pluvialis was reported. The same study also showed the ability to reuse these ionic liquids in subsequent extractions, making them as sustainable solvent. Supercritical CO2 is nowadays the most feasible alternative to organic solvents. About 30 companies are using this technique for lipid extraction from oilseeds. It works at temperatures that are between 40  C and 60  C and pressures ranging from 10 to 70 MPa (Yen et al. 2014; Herrero et al. 2015). The recovery of the solvent is considered to be easy by flashing the mixture at the end and recycling the gas phase. The main advantages are: low temperature, high purity, colorless, odorless, tasteless, low vaporization enthalpy, and less energy consumption. For reducing the extractant cost of the recovery, switchable solvents were developed about a decade ago (Jessop et al. 2012). Switchable solvents are particular liquids that can easily and strongly change their polarity by changing the pH and/or sparging CO2. When they are polar, they can mix with water. When they are nonpolar, they are immiscible with water, but they act as good extractant for lipids. Figure 3 shows the steps of a typical extraction operation carried out with switchable solvents. Next to extracting hydrophobic components, there has been an increased attention for hydrophilic components such as proteins and carbohydrates. As these components are generally fragile components and easily tend to denature (proteins), or degrade at severe conditions such as high temperature, organic solvent usage, or a non-neutral pH, an aqueous extraction of those components is foreseen. As this is an emerging field, currently various technologies like ionic liquids for both hydrophobic and hydrophilic product extraction, and also surfactants, are proposed for separation of those aqueous components. To summarize the efforts reported in literature, two main areas for a breakthrough can be identified in the microalgal biorefinery: (1) to develop an efficient and low energy consuming cell disintegration technique and (2) to improve the yield of extraction of both lipids and proteins. The following section will focus on the current advances in microalgal cell disintegration by providing an overview of current disintegration technologies. In the overview, two types of cell disintegration technologies will be addressed based on their working principle (Fig. 4): physical and nonphysical (Günerken et al. 2015).

Microalgal Biorefinery for Bulk and High-Value Products: Product Extraction. . .

Add water & CO2 product & hydrophobic solvent

7

product hydrophilic solvent & water

decant product

recycle water

decant hydrophobic solvent

hydrophobic solvent water

− CO2 hydrophilic solvent & water

Fig. 3 The process by which a switchable solvent can be separated from a product and recovered without a distillation step (Reproduced from Jessop et al. (2012) with permission of The Royal Society of Chemistry, license number 3904040176494)

The mechanical cell disintegration technologies generally apply a physical force (shear, charge, cavitation, etc.) on the cells leading to an opening of those cells. With chemical cell disintegration, instead of applying physical force, a chemical stress is applied. Due to chemically induced interactions at the cell wall, a weakening of this cell can be induced resulting in an opening of the microalgal cells. After providing an introduction into the various cell disintegration technologies, discussion and future outlook are presented in which various recommendations for the further industrialization of microalgal biorefinery from a cell disintegration perspective will be provided with a focus on pulsed electric field as potential disintegration technology.

Bead Milling Bead mills are homogenizers originally designed for the size reduction of paint and lacquer particles but can also be used to disintegrate biomass suspensions. The basic principle of bead mill is the rapid stirring of small beads in the presence of microorganism suspension. Due to differences in the speed of the beads, high shear forces are created; besides, direct impact of the beads with a microorganism

8

P.R. Postma et al.

Fig. 4 Classification of different cell disintegration technologies (Reprinted from Günerken et al. (2015), Copyright (2015), with permission from Elsevier, license number 3875351355219)

also causes cell disintegration. Bead mill consists of a horizontal or vertical jacketed grinding chamber with a shaft through the center. On this shaft disks, rotors or agitators of different designs can be mounted and will impact the kinetic energy towards the beads. The suspension flows through the grinding chamber while the beads are retained in the chamber by a sieve or axial slot (Fig. 5). Bead mill can be operated both in batch recirculation mode or in continuous mode; in general an external pump is used to create the suspension flow. For batch mode, the suspension flow rate is of minor importance (i.e., only required to assure sufficient recirculation), while the residence time distribution of the suspension inside the grinding chamber is the only important parameter directly influencing the disintegration kinetics. For continuous processing, the flow rate dictates the residence time and therefore the process kinetics. Many other parameters influence the efficiency of bead milling among which are grinding chamber and agitator design, bead size, bead density, bead filling ratio, agitator speed, and biomass concentration (Doucha and Lívanský 2008). Doucha

Microalgal Biorefinery for Bulk and High-Value Products: Product Extraction. . .

9

Fig. 5 Picture of DYNO-Mill Research lab as used by Postma et al. (2015) (a), schematic overview of suspension flow through milling chamber (b) (Pictures kindly provided by Willy A. Bachofen AG Maschinenfabrik, Switzerland)

and Lívanský (2008) found that high bead filling ratios and high bead density are optimal for the disintegration of Chlorella sp. Disintegration levels up to between 58 % and 91 % could be achieved using 2.8–10 kWh kgDW 1. Postma et al. (2015) found that low agitator speeds (6 m s 1) and high biomass concentrations (87.5–145 gDW kg 1) provide the lowest specific energy consumption of 0.8–1.7 kWh kgDW 1 for disintegration of C. vulgaris. Furthermore, water soluble protein yields of 30–35 % were obtained under these conditions using fresh algae.

High Pressure Homogenization Among the mechanical disintegration techniques, high pressure homogenization (HPH) is regarded as a very effective technique and capable of continuous processing. HPH was originally designed to homogenize liquid food products like milk. But it was redesigned for cell disintegration of several microorganisms like, bacteria, yeast, and microalgae (Middelberg 1995; Safi et al. 2014; Ursu et al. 2014). The HPH consists of two main parts, which are the positive displacement pump and a homogenizer valve. The liquid flow direction is changed twice 90 , by first flowing through the center of the valve seat towards the high pressure valve. After which it is forced through the valve opening and will strike the impact ring. By adjusting the valve, the pressure can be set to a desired target (Fig. 6a). Figure 5b shows a production scale HPH with its high pressure pump and valve (Fig. 6c). High liquid shear, turbulence, and cavitation forces are acting on the cells in a short time frame when passing through the homogenizer. It should be considered that these extreme conditions might negatively influence the functional properties of

10

P.R. Postma et al.

Fig. 6 (a) Schematic overview of high pressure homogenizer valve (Authors own drawing), (b) production scale high pressure homogeniser, and (c) high pressure pump and valve detail (Picture (b) and (c) kindly provided GEA Niro Soavi, The Netherlands)

vulnerable products like proteins. Besides, the processing temperature is likely to increase sharply at very high pressures. Successful applications of HPH for the release of enzymes from yeast have been reported. Besides, bacteria like Escherichia coli or Bacillus species have also been successfully disintegrated (Middelberg 1995). Successful application of HPH has been also reported for the release of protein from microalgae. Safi et al. (2014) found that between 41 % and 90 % of the protein content of 5 microalgae species (20 gDW L 1) could be solubilized by two passes of HPH at 270 MPa (7.5 kWh kgDW 1). In addition, Ursu et al. (2014) reported 98 % protein release under alkaline conditions (pH 12) at 2 270 MPa (11.5 kWh kgDW 1) from C. vulgaris (13 gDW L 1). Though it should be considered that in both cases the biomass was frozen before treatment which might have caused cell damage before the application of HPH. Furthermore, when proteins are extracted at a high pH (12), the soluble proteins will precipitate and become no longer soluble at the pH range of 5.5–6 at which functional food proteins are soluble.

Ultrasound During an ultrasonic treatment, the energy of high frequency acoustic waves initiates a cavitation process, and a propagating shock wave forms jet streams in the surrounding medium causing cell disintegration by high shear forces (Mendes-Pinto et al. 2001).

Microalgal Biorefinery for Bulk and High-Value Products: Product Extraction. . .

11

Condensation shock-wave Steam flow

1000m/s

500m/s

5 m/s 5 m/s

Feed

Low pressure supersonic vapour phase

Fig. 7 Schematic representation of Super Sonic Fluid Feed apparatus developed by PDX

The specific energy consumption ranges from 0.06 kWh/kg to 100 kWh kgDW 1 (Günerken et al. 2015). The major drawback of ultrasonication of microalgae biomass is the relatively low cell disintegration efficiency for some microalgae species together with the local and overall heat production. Bubble implosion in acoustic cavitation produce microregions of extreme conditions with estimated temperatures as high as 5000  C and pressures up to 100 MPa. During treatment, the sample temperature can increase significantly from 50 to 90  C (Günerken et al. 2015) and destroy proteins and other intracellular metabolites. Temperature control during treatment can improve product quality; however, the effectiveness of cell disintegration decreases significantly (Sheng et al. 2012). Besides, the energy requirement would increase significantly taking into account the costs for cooling.

Supersonic Flow Fluid Supersonic Flow Fluid Processing (SSFF) can be used for both quick and mild cooking of food and cell disintegration (Fenton et al. 2014). As reported in Fig. 7, steam is introduced into a special annular chamber that is wrapped around the core of the unit and injected through nanopore channels. With increased steam flow, the steam exit velocity becomes supersonic and starts to form a controllable shock wave. This shockwave continues to grow and forms a low density, low temperature, low pressure supersonic velocity zone across the bore diameter, increasing energy transfer and cell disintegration. Although steam is used, the temperature does not exceed 35  C. However, due to the sudden steam condensation during the shockwave, local sharp and fast increase of temperature has to be taken into account. Therefore, this technique does not cause protein or other valuable component to be denatured.

12

P.R. Postma et al.

Pulsed Electrical Field Pulsed Electric Field (PEF) emerged the past decade in the field of cell disintegration with the claim to be promising to permeabilize the microalgal cell membrane and to enhance the release of components (“▶ Extraction of Valuable Compounds from Microalgae Using Pulsed Electric Fields” by Frey, W.) (Toepfl et al. 2006). Up to 27 %, 80 % and 53 % more protein, chlorophyll, and carotenoids, respectively, were released from C. vulgaris. Because the cells are relatively small (3–20 μm), the required field strength is relatively high compared to animal or plant cells. To achieve those electric fields different treatment chambers have been proposed by several research groups who are working on the application of PEF on microalgae (Fig. 8). The most common applied geometries include cross-field, co-field, and co-linear treatment chambers for continuous flow processing. In addition, electroporation cuvettes or similar batch systems have been used as well. A few years before Toepfl et al. (2006), Ganeva et al. (2003) already showed that PEF can be successfully applied to permeabilize yeast and to release relative large protein molecules. The released proteins were up to 250 kDa large with an overall yield up to 50 % and also 70–90 % of the enzyme activity was maintained. More recently, the release of multiple components from microalgae (Auxenochlorella protothecoides) was investigated by Goettel et al. (2013). A strong increase in the conductivity was observed suggesting that the cells were effectively permeabilized. Furthermore, they found that about 8 % of the biomass dry weight was released as carbohydrates whose monomers are small molecules compared to proteins. Instead, the protein release was below 1 % for a specific energy consumption of 0.4 kWh kgDW 1. A major part of the research on the application of PEF on microalgae was conducted with the focus on a single product. Coustets et al. (2013) focused on the release of water soluble protein from C. vulgaris and N. salina. Nevertheless, no protein yield was provided, though the obtained protein concentrations were low (5 h) at elevated temperatures (>35  C) will result in a high energy consumption for heating. When lower valuable chemicals such as lipids for biofuels are aimed to be extracted, there is still a necessity for cost reduction (Gerken et al. 2013). These kind of cost reductions may be achievable with, e.g., immobilizing enzymes. Enzymes have been proposed as well to act as an assisted technology in combination with other cell disintegration technologies (Wang et al. 2015). Next to the existing cost limitations, another limitation of using enzymatic incubation as cell wall disintegration technology is the specificity of these enzymes. Among the various microalgal species, there are large differences in cell wall composition and structure. As a result, for every species, appropriate enzymes should be selected. As up to now, no extensive knowledge on microalgae cell walls is present, applying a rational selection of enzymes is still challenging.

Chemical Cell disintegration can be chemically caused by applying chemicals such as chelating agents, chaotropes, detergents, solvents, hypochlorites, acids, and alkalies depending mainly on the cell wall composition of the microorganism (Middelberg 1995). There are several studies on cell disintegration of microalgae with these agents (Günerken et al. 2015), but the most common are solvent-induced, acid, and alkali. The use of solvents in literature on the microalgae biorefinery is mainly focused on the extraction of specific biochemicals, e.g., astaxanthin and c-phycocyanin. The main issue is related to the organic nature of the solvent that on one hand enhance the extractability of lipids while on the other hand can cause protein denaturation. Some research combines extraction with disintegration in two phase systems (Kleinegris et al. 2011). Harsh acid treatment has been applied to various microalgae biomasses at high temperature (160  C) and generally leads to a higher degree of cell disintegration than the same treatments at lower temperatures (120  C) (Halim et al. 2011). Alkali treatment also requires high temperatures (120  C) causing protein denaturation making this technique less favorable for mild microalgae biorefinery, even if it is still considered the benchmark for total protein (soluble and insoluble) extraction (Halim et al. 2011).

16

P.R. Postma et al.

Conclusions The overview of cell disintegration technologies presented in subsection “Process Design,” illustrated that there are currently various technologies under development or even already applicable at industrial scale. The main aspects of these technologies are summarized in Table 2. Although those technologies have very different mechanisms, they serve the same purpose of enhancing the extractability or release of intracellular components. When cell disintegration in general is developed as a technology that is applicable for various algal downstream process scenarios, it should meet the following two requirements: Costs: The overall costs (“▶ Energy and Cost Analyses of Pulsed Electric Field Applications” by Toepfl, S.) should be minimized. Since there are various scenario’s (energy, fuel, bulk commodities) proposed in which the microalgal biomass represents a fairly low value compared to current cultivation costs, decreasing the energy input during cell disintegration could contribute to the feasibility of those scenarios. Coons et al. (2014) stated, for example, that only 10 % of the total combustion energy of microalgae should be spent to cell disintegration when a biofuel scenario is foreseen. According to their study, this equals about 0.68 kWh/kg biomass. Mildness: A high degree of mildness is a prerequisite. Commonly, during cell disintegration, harsh conditions such as high pressure, temperature, and shear or other extreme environmental conditions like a non-neutral pH are applied. Although those conditions do result in highly efficient cell disintegration with potentially low operational costs, it is likely that next to the cell wall and membrane, intracellular valuables such as proteins also become damaged (denatured). In various applications

Table 2 Summary of cell disintegration technologies for main aspects (Reprinted from Günerken et al. (2015), Copyright (2015), with permission from Elsevier)

Mildness Yes/no

Selective product recovery No

Yes

Yes/no

Optimum biomass concentration Diluted/ concentrated Concentrated

Pulsed electric field Enzymes Supersonic flow fluid Ultrasound

Yes/no

Yes/no

Diluted

Yes Yes/no

Yes N/a

Diluted Diluted

Yes/no

No

Diluted

Microwave Chemical

Yes/no Yes/no

No Yes/no

Diluted Diluted/ concentrated

Disintegration method High pressure homogenization Bead milling

Energy consumption High

Practical scalability Yes

Low/ medium Medium/ high Low Medium/ high Medium/ high High Low/ medium

Yes Yes Yes Yes/no Yes/no Yes/no Yes

Microalgal Biorefinery for Bulk and High-Value Products: Product Extraction. . .

17

Fig. 9 Illustrative overview of protein yields as a function of the specific energy input using PEF and mechanical cell disintegration, bead milling “BM,” and high pressure homogenization “HPH.” When an absolute yield was not provided, estimations were made

such as food and feed applications, but also with an integrated biorefinery, extracting components like proteins in their native form is desired. Therefore, a cell disintegration technology should combine low expenditures with mild conditions to maintain the integrity of all intracellular components. The results by several groups who worked on PEF as microalgal cell disintegration technique have been summarized in Fig. 9. Moreover, two mechanical disintegration methods (bead milling and high pressure homogenization) are provided as a positive control. In this comparison, the protein yield was taken as an illustrative parameter, as the proteins commonly have the highest value when an integrated bulk-commodity biorefinery is foreseen (Wijffels et al. 2010). In addition, proteins are large molecules, so when proteins are released, other (smaller) components such as carbohydrates will be as well. According to the comparison in Fig. 9, applying PEF on different microalgae did not result in yields similar to positive control at an equal or higher energy input than the benchmark technologies. These results show that PEF as single cell disintegration technology is not competitive yet with other technologies for the release of hydrophilic components. However, at the same time, it has been reported that after PEF treatment, the solvent extraction of lipids using the green solvent ethyl-acetate in combination with PEF-treatment resulted in 90 % extraction efficiencies, making

18

P.R. Postma et al.

it competitive to traditional chloroform-methanol extraction (Zbinden et al. 2013). This increased extractability of lipids in combination with the high ion release (>80 %) can suggest that only small pores were made in the cell wall. Azencott et al. (2007) showed that only the cell membrane was harmed during PEF treatment, where the outer cell wall remained unaffected. The results of Zbinden illustrate that PEF is already an interesting technology when only hydrophobic components are foreseen to be extracted using green solvents. However, as a general cell disintegration technology that also enables the release of hydrophilic components, still further research is required. Mahnič-Kalamiza et al. (2014) therefore already envisioned that PEF could be better applied in a multi-stage biorefinery approach. Where PEF would be applied as a first disintegration step followed by aqueous extraction and mild solvent extraction for hydrophobic components. A similar processing route was proposed by Kotnik et al. (2015) to valorise the complete microalgae biomass. Examples of further steps are, for example, optimizing treatment conditions such as the temperature. Postma et al. (2016) reported a 1.7-fold increase in carbohydrate yield when the treatment temperature was increased from 25  C up to 55  C. At a temperature of 55  C, the carbohydrate yield after PEF treatment (EM = 0.55 kWh/ kgdry weight) was 39 %. This yield is in the same order of magnitude as the yield obtained after bead milling (ranging between 48% and 58 %). In addition, Postma et al. (2016) also showed with this study that the treatment temperature during PEF is very well controllable. As high treatment temperature is one of the causes for a non-mild process, this feature illustrates the potentially high degree of mildness of PEF as cell disintegration technology. Despite the increased carbohydrate yields, Postma et al. (2016) did not report a similar increase in protein yield. Nonetheless, this work shows that smart application of operating conditions can result in a better process making PEF a very promising future cell disintegration technology for a microalgae biorefinery approach. Acknowledgements Part of this work is performed within the TKI AlgaePARC Biorefinery program with financial support from the Netherlands’ Ministry of Economic Affairs in the framework of the TKI BioBased Economy under contract nr. TKIBE01009. Part of this work is performed within the IPOP Biorefinery of Wageningen University and Research. Part of this work was conducted in the framework of a COST TD1104 action (Reference code TD1104041262 and TD1104-16869) (www.electroporation.net).

Cross-References ▶ Advantages of Pulsed Electric Field Use for Treatment of Algae ▶ Energy and Cost Analyses of Pulsed Electric Field Applications ▶ Extraction of Valuable Compounds from Microalgae Using Pulsed Electric Fields ▶ Pulsed Electric Fields Assisted Lipid Extraction from Microalgae

Microalgal Biorefinery for Bulk and High-Value Products: Product Extraction. . .

19

References Azencott HR, Peter GF, Prausnitz MR (2007) Influence of the cell wall on intracellular delivery to algal cells by electroporation and sonication. Ultrasound Med Biol 33:1805–1817. doi:10.1016/ j.ultrasmedbio.2007.05.008 Chisti Y (2007) Biodiesel from microalgae. Biotechnol Adv 25:294–306. doi:10.1016/j. biotechadv.2007.02.001 Coons JE, Kalb DM, Dale T, Marrone BL (2014) Getting to low-cost algal biofuels: a monograph on conventional and cutting-edge harvesting and extraction technologies. Algal Res 6, Part B:250–270. doi:10.1016/j.algal.2014.08.005 Coustets M, Al-Karablieh N, Thomsen C, Teissié J (2013) Flow process for electroextraction of total proteins from microalgae. J Membr Biol 246:751–760. doi:10.1007/s00232-013-9542-y Desai RK, Streefland M, Wijffels RH, Eppink MHM (2016) Novel astaxanthin extraction from Haematococcus pluvialis using cell permeabilising ionic liquids. Green Chem 18:1261–1267. doi:10.1039/C5GC01301A Doucha J, Lívanský K (2008) Influence of processing parameters on disintegration of Chlorella cells in various types of homogenizers. Appl Microbiol Biotechnol 81:431–440. doi:10.1007/ s00253-008-1660-6 Fenton MBM, Koroleva O, Gothard MGE, Drake C (2014) An apparatus and method of processing microorganisms. EU Patent 2723849A1, US Patent 20130337523A1 Flisar K, Meglic SH, Morelj J et al (2014) Testing a prototype pulse generator for a continuous flow system and its use for E. coli inactivation and microalgae lipid extraction. Bioelectrochemistry 100:44–51. doi:10.1016/j.bioelechem.2014.03.008 Ganeva V, Galutzov B, Teissié J (2003) High yield electroextraction of proteins from yeast by a flow process. Anal Biochem 315:77–84. doi:10.1016/S0003-2697(02)00699-1 Gerken HG, Donohoe B, Knoshaug EP (2013) Enzymatic cell wall degradation of Chlorella vulgaris and other microalgae for biofuels production. Planta 237:239–253. doi:10.1007/ s00425-012-1765-0 Goettel M, Eing C, Gusbeth C et al (2013) Pulsed electric field assisted extraction of intracellular valuables from microalgae. Algal Res 2:401–408. doi:10.1016/j.algal.2013.07.004 Gonçalves AL, Alvim-Ferraz MCM, Martins FG et al (2016) Integration of microalgae-based bioenergy production into a petrochemical complex: techno-economic assessment. Energies 9:224. doi:10.3390/en9040224 Günerken E, D’Hondt E, Eppink MHM et al (2015) Cell disruption for microalgae biorefineries. Biotechnol Adv 33:243–260. doi:10.1016/j.biotechadv.2015.01.008 Halim R, Gladman B, Danquah MK, Webley PA (2011) Oil extraction from microalgae for biodiesel production. Bioresour Technol 102:178–185. doi:10.1016/j.biortech.2010.06.136 Halim R, Harun R, Danquah MK, Webley PA (2012) Microalgal cell disruption for biofuel development. Appl Energy 91:116–121. doi:10.1016/j.apenergy.2011.08.048 Herrero M, Ibáñez E (2015) Green processes and sustainability: an overview on the extraction of high added-value products from seaweeds and microalgae. J Supercrit Fluids 96:211–216. doi:10.1016/j.supflu.2014.09.006 Jessop PG, Mercer SM, Heldebrant DJ (2012) CO2-triggered switchable solvents, surfactants, and other materials. Energy Environ Sci 5:7240–7253. doi:10.1039/C2EE02912J Kleinegris DMM, Janssen M, Brandenburg WA, Wijffels RH (2011) Two-phase systems: potential for in situ extraction of microalgal products. Biotechnol Adv 29:502–507. doi:10.1016/j. biotechadv.2011.05.018 Kotnik T, Frey W, Sack M et al (2015) Electroporation-based applications in biotechnology. Trends Biotechnol 33:480–488. doi:10.1016/j.tibtech.2015.06.002 Luengo E, Martínez JM, Bordetas A et al (2015a) Influence of the treatment medium temperature on lutein extraction assisted by pulsed electric fields from Chlorella vulgaris. Innov Food Sci Emerg Technol 29:15–22. doi:10.1016/j.ifset.2015.02.012

20

P.R. Postma et al.

Luengo E, Martínez JM, Coustets M et al (2015b) A comparative study on the effects of millisecondand microsecond-pulsed electric field treatments on the permeabilization and extraction of pigments from Chlorella vulgaris. J Membr Biol 248:883–891. doi:10.1007/s00232-015-9796-7 Mahnič-Kalamiza S, Vorobiev E, Miklavčič D (2014) Electroporation in food processing and biorefinery. J Membr Biol 247:1279–1304. doi:10.1007/s00232-014-9737-x Mendes-Pinto MM, Raposo MFJ, Bowen J et al (2001) Evaluation of different cell disruption processes on encysted cells of Haematococcus pluvialis: effects on astaxanthin recovery and implications for bio-availability. J Appl Phycol 13:19–24. doi:10.1023/A:1008183429747 Middelberg AP (1995) Process-scale disruption of microorganisms. Biotechnol Adv 13:491–551 Parniakov O, Barba FJ, Grimi N et al (2015) Pulsed electric field and pH assisted selective extraction of intracellular components from microalgae Nannochloropsis. Algal Res 8:128–134. doi:10.1016/j.algal.2015.01.014 Postma PR, Miron TL, Olivieri G et al (2015) Mild disintegration of the green microalgae Chlorella vulgaris using bead milling. Bioresour Technol 184:297–304. doi:10.1016/j. biortech.2014.09.033 Postma PR, Pataro G, Capitoli M et al (2016) Selective extraction of intracellular components from the microalga Chlorella vulgaris by combined pulsed electric field–temperature treatment. Bioresour Technol 203:80–88. doi:10.1016/j.biortech.2015.12.012 Safi C, Ursu AV, Laroche C et al (2014) Aqueous extraction of proteins from microalgae: effect of different cell disruption methods. Algal Res 3:61–65. doi:10.1016/j.algal.2013.12.004 Sheng J, Vannela R, Rittmann BE (2012) Disruption of Synechocystis PCC 6803 for lipid extraction. Water Sci Technol 65:567. doi:10.2166/wst.2012.879 ‘t Lam GP, Zegeye EK, Vermuë MH et al (2015) Dosage effect of cationic polymers on the flocculation efficiency of the marine microalga Neochloris oleoabundans. Bioresour Technol 198:797–802. doi:10.1016/j.biortech.2015.09.097 Toepfl S, Heinz V, Knorr D (2006) Applications of pulsed electric fields technology for the food industry. In: Raso J, Heinz V (eds) Pulsed electric fields technology for the food industry. Springer US, New York, pp 197–221 Tredici MR (2010) Photobiology of microalgae mass cultures: understanding the tools for the next green revolution. Biofuels 1:143–162. doi:10.4155/bfs.09.10 Ursu A-V, Marcati A, Sayd T et al (2014) Extraction, fractionation and functional properties of proteins from the microalgae Chlorella vulgaris. Bioresour Technol 157:134–139. doi:10.1016/ j.biortech.2014.01.071 Vigani M, Parisi C, Rodríguez-Cerezo E et al (2015) Food and feed products from micro-algae: market opportunities and challenges for the EU. Trends Food Sci Technol 42:81–92. doi:10.1016/j.tifs.2014.12.004 Wang D, Li Y, Hu X et al (2015) Combined enzymatic and mechanical cell disruption and lipid extraction of green alga Neochloris oleoabundans. Int J Mol Sci 16:7707–7722. doi:10.3390/ ijms16047707 Wijffels RH, Barbosa MJ (2010) An outlook on microalgal biofuels. Science 329:796–799. doi:10.1126/science.1189003 Wijffels RH, Barbosa MJ, Eppink MHM (2010) Microalgae for the production of bulk chemicals and biofuels. Biofuels Bioprod Biorefining 4:287–295. doi:10.1002/bbb.215 Yen H-W, Yang S-C, Chen C-H et al (2014) Supercritical fluid extraction of valuable compounds from microalgal biomass. Bioresour Technol 184:291–296. doi:10.1016/j.biortech.2014.10.030 Zbinden MDA, Sturm BSM, Nord RD et al (2013) Pulsed electric field (PEF) as an intensification pretreatment for greener solvent lipid extraction from microalgae. Biotechnol Bioeng 110:1605–1615. doi:10.1002/bit.24829

Mechanistic Description of Membrane Electropermeabilization Justin Teissie

Abstract

Classical cell membrane electropermeabilization of cell membrane is the result of the delivery of electric field pulses on cells. The electric field pulse lasts from submicro- to several milliseconds. The electric field intensity is large enough to induce a dramatic structural local alteration of the cell membrane organization. This results in an enhanced permeabilization of the target cell membrane for molecules otherwise poorly transportable. This structural alteration is indeed a complex process, and its molecular characterization remains an intense field of investigations. The new transient organization of the cell membrane supports a massive transport due to electrophoretic forces and diffusion-driven gradients. This chapter describes the fast events inducing electropermeabilization or the immediate consequences of the field-induced alteration of the membrane and cellular organization. The methods suited to monitor these fast events are critically described as they are key factors in the accuracy of the informations. Three steps are present in cell membrane electropermeabilization: trigger, expansion, and stabilization. The experimental results are discussed in terms of structural information on the new transient membrane organization. Most informations are related to the massive enhanced molecular transport across the membrane and its modulation by the electric field pulse delivery. Keywords

Membrane • Transmembrane Voltage • Transport • Membrane Structure

J. Teissie (*) Institut de Pharmacologie et de Biologie Structurale, CNRS, UPS, Université de Toulouse, Toulouse, France e-mail: [email protected]; [email protected] # Springer International Publishing AG 2017 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_39-1

1

2

J. Teissie

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Electric Field Induces a Transmembrane Voltage (TMV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Resulting TMV Induces Membrane Permeabilization (Electropermeabilization) . . . . . . . . . . Methods Used in Detecting and Follow-up of Membrane Electropermeabilization . . . . . . . . . . . . Conductance Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transport Assayed by Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Permeabilization during the Pulse Delivery Depends on the External Field Strength and Pulse Duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conductance Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Induced Transport across the Membrane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Permeabilization as Described by its Post-Pulse Consequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Control of Cell Membrane Electropermeabilization by the Electrical Parameters . . . . . . . . . . Biochemical Definition of Permeabilized Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lipids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Structural Arrangements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biphotonic Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nuclear Magnetic Resonance (NMR) Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electron Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fluorescence Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lateral Mobility of Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interfacial Modifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modes of Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modulation of Electropermeabilization by Physicochemical Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Buffer Osmotic Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Buffer Ionic Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nucleotides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cytoskeleton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Membrane Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 4 5 5 6 7 7 8 10 10 13 13 13 14 14 14 14 15 15 15 16 16 17 17 18 18 18 19 19 20 21 21

Introduction This chapter describes the events occurring along classical electropermeabilization (electropermeabilization). This is what is observed with a field range of 0 < E < 10 kV/cm for mammalian cells (diameters between 10 and 30 μm) and pulse durations larger than submicros (therefore classical electropermeabilization included some experiments called nsPEF (nanosecond pulsed electric field) when the duration is larger than 100 ns) (Deng et al. 2003; Son et al. 2014). This is obtained with pulse generators with square wave profile rising time larger than 0.1 μs. Electric pulses when delivered to a cell suspension or a tissue result in a membrane permeabilization (electropermeabilization). Classical electropermeabilization results in an enhanced transport of poorly permeable ions and molecules

Mechanistic Description of Membrane Electropermeabilization

3

across the membrane and in a change in its electrical conductivity. Electropermeabilization can be reversible if the pulsing parameters are selected in a proper way. This reversibility is called resealing (resealing). This means that an enhanced membrane permeabilization is transiently induced while preserving the cell viability. More drastic pulsing conditions can make this permeability irreversible (cell lysis would result) (Rems and Miklavcic 2016). This chapter is an attempt to make a state of the art in the field by pointing out the most recent observations. More classical information can be read in the works of the leading groups in the field during the last 40 years (see “Cross-References” at the end of the chapter).

The Electric Field Induces a Transmembrane Voltage (TMV) When a cell is subjected to an external electric field, the cell (where the plasma membrane can be considered as a dielectric shell) can be considered as a spherical capacitor, where the redistribution of electrophoretically driven charged ions in the electrolytes surrounding the membrane, i.e., electric current, leads to an induced transmembrane voltage (TMV) (transmembrane voltage). During the electric field application (electropulsation) (electropulsation) on the cell, an induced transmembrane voltage (ΔVi) is created which is locally associated with the dielectric properties of the plasma membrane. Using a physical model based on a thin, weakly conductive shell (the membrane, conductivity λm), full of an internal conductive medium (the cytoplasm, conductivity λi), and bathed in an external conductive medium (conductivity λe), solution of Laplace’s differential equation gives ΔVi as:
ΔViðM, Ε, tÞ ¼ f gðλÞrE cosθðMÞ 1  expðt=τmÞ

(Eq:1)

where M is the point on the cell that is considered, t is the time after electropulsation is turned on, f is a factor depending on the cell geometry (for a sphere, f = 1.5), r is the radius of the pulsed cell, E is the electric field strength, and θ(M) is the angle between the direction of the field and the normal of the cell surface in M. g(λ) is related to the different conductivities as (Zimmermann et al. 1947) g ð λÞ ¼

h i
2λe 2λm þ λi þ ðλm  λiÞðr  d=rÞ3  3λmðr  d=rÞ h i = ð2λe þ λmÞð2λm þ λiÞ þ 2ðr  d=rÞ3 ðλi  λmÞðλm  λeÞ

(Eq:2)

where d is the thickness of the membrane (a few nm). τm is the characteristic time constant of the membrane charging and can be written as (Kinosita and Tsong 1977): τm ¼ rCmð2λe þ λiÞ=ð2λeλiÞ

(Eq:3)

where Cm (0.5–1.0 μF/cm2) is the specific membrane capacitance. τm is calculated to be in the submicrosecond time range for mammalian cells. It is strongly dependent on the buffer composition as the internal composition is fixed by the cell metabolism.

4

J. Teissie

Fig. 1 Modulation of the TMVof a spherical cell by an applied external field. The TMV point from negative to positive and their lengths are proportional to the voltage. The large arrow on the top represents the electric field lines. On the left, a cell is pictured with its homogeneous resting voltage; the middle is indicative of the field-induced TMV that is highly position dependent along the cell surface. On the right, the resulting modulation of the TMV when the field is present, a highly complex position-dependent TMV, is observed

If the plasma membrane is considered to be a pure spherical dielectric (λm = 0), we obtain g(λ) = 1. In these conditions and when the steady state is reached (t >> τm) as for millisecond pulses, Eq. 1 simplifies to: ΔViðM, EÞ ¼ 1:5rEcosθðMÞ

(Eq:4)

When living cells are electropulsed, ΔVi adds to the resting one ΔV0. This gives an additive effect on one side of the cell and a subtractive one on the opposite. The electric field effect is strongly position dependent (Fig. 1).

The Resulting TMV Induces Membrane Permeabilization (Electropermeabilization) When the new membrane electric potential difference ΔVm (ΔV0 + ΔVi) locally reaches a critical value (ΔVc) (nowadays estimated between 0.25 and 0.6 V for living cells) (Rems and Miklavcic 2016), a local alteration of the membrane structure leads to membrane permeabilization. Molecular processes supporting classical electropermeabilization (electropermeabilization) remain poorly understood (Teissie et al. 2005) in spite of more than 40 years of effort (Neumann and Rosenheck 1972). The membrane is called “porous.” There is a general agreement that it cannot be described by the occurrence of reversible hydrophilic holes, so-called pores (“electroporation”) (electroporation), in the lipid bilayer. A membrane of a cell is a more complex system. It is an out of equilibrium complex from a thermodynamic point of view. It forms a complex with the extracellular matrix and the cytoskeleton. Its transmembrane voltage is dependent on the cellular energy reserves.

Mechanistic Description of Membrane Electropermeabilization

5

If at any point ΔVm gets high enough to induce sufficient number of conductive defects, the defects allow enhanced ionic transport through the membrane and increase the membrane conductivity by several orders of magnitude. This in turn partially discharges the membrane and thereby reduces ΔVm during the pulse. In such case, the membrane conductivity has to be considered as a function of ΔVm, which will be in more details described in other chapters (see “Cross-References” at the end of the chapter). Electropermeabilization is described as a local event in the cell surface. Membrane permeabilization is associated by an enhanced transport of watersoluble small molecules (that are of a low molecular weight, say less than 2 kD) and ions. It is a time-dependent process, and its kinetics can be observed by monitoring the electrical conductivity change in the membrane or the change in the transport parameters. It can be assayed by the inflow of dyes (trypan blue, propidium iodide, and similar) by observation under the microscope or with a flow cytometer. Permeabilization can then be quantified by two parameters: the number of cells in the population where the uptake can be detected and the amount of uploaded dyes (i.e., the distribution of fluorescence intensity in each single cell in the population). To follow the onset of electropermeabilization which occurs during the pulse delivery, the observation is on the submillisecond time range. But in most experiments, it results in an enhanced transport (transport) during the post-pulse step that is observed. But this is affected by the resealing (see “Cross-References” at the end of the chapter). Different biophysical approaches monitor the transport as mentioned above. They are shortly described in a first part of this chapter. They are used to analyze the processes and their control by biophysical and physiological parameters. A lot of studies have been developed on pure lipid assemblies (planar lipid membrane, liposomes). Processes on such physical models are relevant of soft matter physics and would not be considered in the present chapter. Only events occurring in living cell membranes will be approached. Computer-assisted simulations of transvoltage effect on lipid bilayers are not considered relevant of classical electropermeabilization (different pulse duration, much higher equivalent field strength).

Methods Used in Detecting and Follow-up of Membrane Electropermeabilization Conductance Changes Direct Assays by Electrical Measurements By measuring the voltage and the current flowing between the electrodes, one gets access to the conductance of the suspension (conductance). Under steady-state condition, the conductance of a diluted intact cell suspension can be considered controlled by the conductance of the external solution. An ohmic behavior is observed when a voltage pulse is delivered on the cell suspension (as long as the Joule heating is small).

6

J. Teissie

When electropermeabilization appears, two events are present: (i) The dielectric shell has conductive defects that increase during the pulse (ii) There is leak of the internal (cytoplasmic) solution Both events result in an increase in the cell suspension conductivity. This is a very fast process that can be detected on the submicrosecond time scale after correcting from the interfacial electrochemical reactions on the electrodes (Kinosita and Tsong 1979).

Indirect Approaches by Fluorescence Membrane-embedded fluorescent probes have their emission controlled by the transmembrane voltage by a direct effect of the transmembrane field (Hibino et al. 1991). The probe emission change linked to the alteration of the membrane conductance (that affects the transmembrane field) is very fast and closely follows the change in conductance. A very fast detection is needed where short (ns to μs) laser flashes triggered at delayed times after the onset of the field pulses are used to stimulate the emission (Frey et al. 2006; Hibino et al. 1991).

Transport Assayed by Fluorescence Membrane permeabilization is associated to an enhanced diffusion-driven transport of water-soluble molecules and ions following the physical laws of diffusion (Fick’s equation) (diffusion). This is a time-dependent process, and its kinetics can be observed by monitoring the inflow of a fluorescent molecules. In most cases, molecules such as propidium iodide are used as their fluorescence emission is strongly enhanced in the cytoplasm due to their binding to nucleic acids. Transport is assayed online with time resolution down to the us (Pucihar et al. 2008) or imaging on the ms range as shown by video imaging in the late 1990s (Table 1). In most cases, it is the post-pulse transport that is assayed as it is a result of the structural alterations present during the pulse delivery (see “Cross-References”). Table 1 Synthetic view of the different techniques which are used to investigate membrane electropermeabilization (strength and limits, main results) Methods Conductance Patch camp

Temporal resolution Fast (microsecond) Slow (second)

Spatial resolution Low (on population) Low (on single cell)

Fluorescence NMR SEM AFM

Fast (microsecond) very slow (hours) very slow very slow

1 μm on single cell Average information High (but fixed sample) High

Molecular information average size of defects molecular definition of defects transport tilt of lipid headgroups ultrastructural defects ultrastructural defects

Mechanistic Description of Membrane Electropermeabilization

7

Permeabilization during the Pulse Delivery Depends on the External Field Strength and Pulse Duration Conductance Changes Electrical Measurements Conductance changes during the pulse can be observed on cell suspension only when the field is larger than a threshold (critical value Ec) (Kinosita and Tsong 1979) (conductance). The increase is time dependent during the delivery of the pulse. Its magnitude is dependent on the electric field intensity. The relative conductivity changes during the application of a rectangular field pulse between 0.75 and 1.2 kV/cm, respectively, with a duration of 1 ms to a mammalian cell dense suspension (pellet) in a low conductivity isosmotic buffer sucrose are also a very fast increase. The conductance increase can be detected within the first μs of the pulse delivery after correcting for the electrodes interfacial contribution to the electrical signal. A plateau in the conductance change is observed for long pulses in conductive buffers. Imaging Observations The use of a very fast fluorescence imaging microscopy connected to a pulse generator gives access to the conductance changes (Hibino et al. 1991) (imaging). A voltage-sensitive fluorescent dye was used an indicator of the time-dependent transmembrane change induced by the external electric field on sea urchin eggs. But when the field was over a critical value, the transmembrane voltage kept locally a constant value on a cap facing the electrodes. A high membrane conductance of the order of 1 S/cm2 is therefore present within 2 μs after the onset of the external field. The size of the conductive cap is determined by the strength of the external field. This high local membrane conductance affects the electric field-induced transmembrane voltage as predicted by Eq. 1. A steady value is observed in the cap where the transmembrane conductance is obtained (Hibino et al. 1993). Only 0.5 microseconds after the onset of the rectangular electric pulse, the two sides of the cell facing the electrodes are affected, and a high membrane conductance in the order of 1 S/cm2 is measured; an asymmetric process is observed; a higher conductance is detected on the positive side. Along the pulse delivery (the external field remaining constant), a further increase in the conductance is detected, reaching the order of 10 mS/cm by 1 ms. The kinetics of the conductance increase is different on the two opposite sides of the cell, being faster on the negative-electrode side; as a result at the end of a 1 ms pulse, the conductance on the negative side is more than twice that on the positive side. Different processes clearly support electropermeabilization on the opposite caps of a pulsed cell.

8

J. Teissie

Induced Transport across the Membrane The transport of a water-soluble dye (propidium iodide, PI) induced by electropermeabilization was monitored under an ultrafast fluorescence microscope (bandwidth larger than 1 MHz) during the electric pulse on a cell population of mammalian cells as well as at the single cell level (Pucihar et al. 2008) (propidium iodide). The transport became detectable as early as 60 μs after the start of the pulse (transport). Transport is delayed as compared to the enhanced membrane conductance. After the delay of 60 μs, the time course of fluorescence during the pulse was approximately linear, suggesting mainly electrophoretic transport of molecules during the pulse delivery lasting up to 1 ms. Transport through these pathways can occur by different mechanisms during and after a pulse. To determine the time scale of transport and the mechanism(s) by which it occurs, efflux of a fluorescent molecule, calcein, across erythrocyte ghost membranes was measured with a fluorescence microscope photometer with a fast (millisecond) time resolution during and after electric pulses that lasted several milliseconds (Prausnitz et al. 1995). Only a reduced leak was observed during the pulse, and most of the efflux from the ghost was after the pulse. Transport caused by electropermeabilization occurred predominantly by electrophoresis and/or electroosmosis during a pulse, although transport occurred in part or almost completely by diffusion within milliseconds to seconds after the pulse (electrophoresis, electroosmosis). On a shorter time scale (microsecond), the course of electropermeabilization was analyzed during the electric field application using a rapid single cell fluorescent imaging system. Interaction and penetration of propidium iodide, with and across cell membrane, was asymmetrical during electropermeabilization. Localized enhancement of the dye fluorescence was observed during the pulsation on the cell surface and in the cytoplasm. But this was observed only when the amplitude of the external field was higher than a critical value. Specific staining of a limited anodefacing cell volume was observed during the first milliseconds when the pulse was applied. The membrane fluorescence level increased linearly during the pulse delivery, whereas the geometry of the staining was unchanged. The fluorescence in the cap increased linearly during the pulse delivery (lasting up to 20 ms), as expected from an electrophoretic drift. The fraction of the membrane on which structural alterations occurred, Ap, was defined by the electric field strength E. It was a linear function of the reciprocal of the electric field strength. Ap ¼ A0ð1  Ep=EÞ

(Eq:5)

where Ep is the critical field strength needed to trigger membrane electropermeabilization. The size of the cap where permeabilization is present increased with an increase of the field strength as soon as the field was larger than the critical field Ep (Fig. 2). The density of defects supporting the transport across the permeabilized membrane cap was governed by the pulse duration (Fig. 3).

Mechanistic Description of Membrane Electropermeabilization

9

Fig. 2 Effect of the field strength on the size of the electropermeabilized cap. Cells are pictured as a sphere. Field pulses are delivered at a constant pulse duration. No permeabilization is present before the pulse delivery (left side of the cartoon). Low field has no effect on the membrane organization. As soon as the field induces a permeabilizing TMV, i.e., the field strength is larger than Ep, a cap is brought to the “permeabilized state” (a gray cap on the cartoon). The size of this permeabilized cap increases with a further increase of the field strength

Fig. 3 Effect of the pulse duration on membrane electropermeabilization. Cells are pictured as a sphere. No permeabilization is present before the pulse delivery (left side of the cartoon). As soon as the field induces a permeabilizing TMV, a cap is brought to the “permeabilized state” (the gray cap on the cartoon) (trigger step). As long as the field is present with a constant strength, an increase in the defects supporting the permeabilization occurs shown by darker levels in the gray levels of the cap. This is called the expansion step

Electropermeabilization is a localized but asymmetrical process (electropermeabilization). Transport was not the same on the two cell sides. The membrane defects (defects) are created unequally on the two opposing sides of the cell during the pulse, implying a vectorial effect of the electric field on the membrane. Exchanges of calcium ions through electropermeabilized membrane of Chinese hamster ovary cells were found to be asymmetrical. Entry of calcium ions during a millisecond pulse occurred on the anode-facing cell hemisphere. The exchanges during the pulse were mostly by an electrophoretic drift. The membrane regions stained by propidium iodide uptake were the same as those where calcium exchanges occurred. Positively charged species are electrophoretically pushed in the

10

J. Teissie

cell cytoplasm on the anode-facing cell hemisphere. Electropermeabilization is facilitating the electrophoretic movement of any kind of small species across a membrane cap controlled by the field strength during the pulse delivery. Water transport was detected resulting in huge osmotic swelling of the cell when using ms pulses (a high level of permeabilization) (swelling). This is easily prevented by adding large water-soluble additives in the pulsing buffer (such as sucrose or polyethylene glycol, PEG).

Permeabilization as Described by its Post-Pulse Consequences Most studies are not on the events during the pulse but on the observations of cells that had just been pulsed. The consequence is that it is not the direct effect of the onset of electropermeabilization that is observed but its synergy with the resealing process that follows. What is then observed is the percentage that have been permeabilized and the level of the inflow of the reporter molecules (Canatella et al. 2001; Pucihar et al. 2008). The last parameter is indeed strongly controlled by the resealing. A more relevant approach of the onset of permeabilization during the pulse is to observe the subpopulation of permeabilized cells in the total population of pulsed cells. One limit is of course the sensibility of the assay to detect permeabilized cells. Results were observed to be in many cases different by using different assays. The inflow is under the control of the permeability coefficient of the reporter molecule (roughly dependent on the size). The number of molecules that can be detected inside the cytoplasm is controlled by the method of detection. For example, detection is obtained under less stringent conditions when using PI penetration (fluorescence detection) rather than trypan blue labeling (colorimetric method) (detection).

Control of Cell Membrane Electropermeabilization by the Electrical Parameters Field Strength The electric field strength of the pulses delivered on the cell suspension was observed to be a critical parameter to trigger cell membrane permeabilization (detected in most cases by the uptake of a fluorescent dye). Transport is indeed controlled by the field strength E. Fluorescent-positive cells (i.e., permeabilized) induced by the pulse are detected only when the intensity of the pulse is larger than a critical threshold (Ep) (note: in a cell population, a percentage of cells has a leaky membrane where the dye uptake spontaneously takes place. They are observed in the control cells. This must be corrected from the observed permeabilization). For a given cell (size r) and a given pulse duration (T), Ep induces the permeabilization on a cap with a critical size, across which transport can be detected. An accurate definition should be Ep(r,T).

Mechanistic Description of Membrane Electropermeabilization

11

A fairly linear relationship is observed between permeabilization and the reciprocal of E (Eq. 5). This supports the online observation that a cap on the cell surface is brought to the permeabilized state by the pulse with a size controlled by the field. This cap is where the critical permeabilizing TMV was reached along the pulse. At high field, all cells are permeabilized, as Ep(r) is obtained for all values of r in the population.

Pulse Duration When the field strength of the delivered pulses is larger than this critical strength Ep, it is observed that the percentage of cells detected as permeabilized is increased with an increase in the pulse duration. Transport is controlled by the pulse duration T (duration). This increase is detected to level off at a plateau value close to 100% (all cells are detected to be permeabilized when long pulses are delivered). Interestingly, the critical parameter Ep is observed to decrease with an increase in the pulse duration. A linear relationship is observed between Ep(T) and the reciprocal of the pulse duration (T). This gives access by extrapolation to a limiting value of Ep (for an infinite duration). This may result from several factors such as: – The density of defects increases with the pulse duration in the cap where the critical TMV is triggered (defects). The inflow increases as a consequence and detection is obtained with a very limited permeabilized cap. – Due to the electrophoretic drift of membrane proteins, a microdomain is created that is more sensitive to the field-induced TMP. – Due to the electromechanical stress on the cell, membrane stretching occurs, and the cap facing the electrode has a forced curvature making it more fragile. – The transport by the electrophoretic forces during the pulse is high enough to give a detectable accumulation of the reporter dye.

Repetition of Pulses In most studies, not only a single pulse is delivered to the cells, but a train of pulses is applied where two complementary parameters are to take into account: the number of successive pulses and delay between each single pulse (frequency) (delay). Conductance measurements give a direct access of the effect of the repetitive pulses. It is observed that the membrane conductance is further increased by the repetitive pulses but that a partial resealing occurs during the delay (Kinosita and Tsong 1979). More conducting defects are induced along a repetition of pulses. The local defects are considered part of the structural membrane reorganization induced by the external field. They remain present for a few seconds after each single pulse. Their density is dependent on the number of pulses. Cumulative effects are shown by an enhanced transport and are observed when repeated pulses are applied. Transport therefore increases with the number of pulses. More cells are then detected as permeabilized. Again a plateau level is detected when a large number of pulses are delivered. Indeed, this brings a result where all cells are permeabilized.

12

J. Teissie

Delay between Pulses The delay plays a role in the final state of pulsed cells. This is a complex process that is not clearly explained at the present state of our knowledge (so-called electrosensitization) (electrosensitization). At a constant number of pulses/pulse duration product, transfer of molecules is strongly affected by the time between pulses. Permeabilization is observed to be higher when a low number of long pulses are delivered with a long delay rather than a repetition of short pulses. Orientation of Pulsing The electric field being a vector has a given orientation (perpendicular to the electrodes when they are plane parallel). Cell electropermeabilization is triggered by an electric field; its characteristics should depend on its vectorial properties. The direction of the field is easily changed by moving the position of the parallel electrodes relative to fixed cells (either plated or anchored to a dish). Four different pulse sequences were compared: normal (pulses all in one given direction), crossed (pulses delivered in two perpendicular directions), inverted (pulses with opposite polarities), and crossed-inverted (pulses delivered in two perpendicular directions with opposite polarities). Under non-saturating conditions, the normal condition was observed to be the less effective conditions, while the highest effect was obtained under the crossed-inverted one. Permeabilized caps are formed only on one side of the cell under the less effective conditions (field strength close to Ep), while they appear on four different places on the cell surface under the most effective setting. Multiple studies have shown that delayed bipolar (BP) electric pulses in the microsecond range are more effective at permeabilizing cells as compared to monopolar (MP) pulse equivalents. To study permeabilization effectiveness, MP or BP (with no delay in the inversion) pulses were delivered to single Chinese hamster ovary (CHO) cells, and the response of propidium iodide (PI) or calcium uptake was measured by confocal microscopy. Results showed that BP pulses were less effective at increasing intracellular calcium concentration or PI uptake than MP pulses. Flow cytometry analysis of CHO cells after exposure (at 15 min) revealed that to achieve positive PI transport, ten times more BP pulses were required than MP pulses. Overall, unlike longer pulse exposures, BP nsPEF exposures proved far less effective at both membrane permeabilization than MP nsPEF (Ibey et al. 2014). The uptake of calcium ions was decreased by more than an order of magnitude by the delivery of a second one of opposite polarity immediately after a first pulse. This effect reflects the vectorial character of the electrophoretic transport of ions through the electropermeabilized caps in the membrane during the two parts of the bipolar pulse. This observation supports the conclusion that electrophoresis is the dominant transport mechanism during the pulse delivery (where permeabilization is created), rather than diffusion that is driven by the concentration difference. By delivering the second pulse with a delay after the first one, the effect of reverse electrophoresis disappears (electrophoresis). This delay is very short (in the hundred microseconds range) and illustrative of a very fast recovery of the membrane perturbation supporting the electrophoretic transport.

Mechanistic Description of Membrane Electropermeabilization

13

Biochemical Definition of Permeabilized Structures Lipids Lipid unilamellar vesicles (LUV) (vesicles), physical models of cells but composed of pure lipids, are quite suitable systems as their size is close to the one of cells. Their molecular packings are similar to what one expects to be present in the plasma membrane. When submitted to short (where stretching is avoided) high field pulses, LUVare transiently permeabilized. Leakage of trapped sucrose can be detected when the field strength is larger than a critical value (Ep). No global and permanent damages to the vesicle bilayer are detected. Electropermeabilization is therefore obtained in a lipid assembly. Lipid domains when present in cell membrane are putative target of electropermeabilization.

Proteins Electropermeabilization of red blood cells was supported by the associated cell osmotic swelling. It was observed that in a low ionic medium, at least 35% of the conductive defects were related to the opening of Na+/K+ ATPase channels. This conclusion was supported by the partial blockage of the membrane conductance generated by the externally applied electric field pulse by a specific inhibitor, ouabain. A similar inhibition of the conductance was obtained when using a specific cross-linking reagent, Cu++  phenanthroline, of the ATPase. Interestingly this effect was not observed in a high ionic medium as expected for the effect of the cross-linker. A large fraction of the voltage-induced defects that occurred are still unidentified. The effects of large magnitude transmembrane voltage pulses on voltage-gated Na and K channel behavior were studied in frog skeletal muscle membrane of isolated fibers (Chen and Lee 1994). A 4 ms transmembrane potential pulse of 600 mV reduces both Na and K channel conductivities and the ionic selectivity of the K channels against Na + ions. This reflects an electroconformational damage to ionic channels in parallel to the induction of defects in the lipid bilayer. Classical effectors such as tetrodotoxin (TTX) and tetraethylammonium (TEA) binding effects were unaltered. By the whole-cell patchclamp method investigation on neuronal cells, currents of voltage-gated (VG) Ca(2+) and Na(+) channels (I(Ca) and I(Na)) in cultured GH3 and NG108 cells were observed to be inhibited when cells were pulsed with submicrosecond fields at or above 1.5–2 kV/cm. The field pulses caused prolonged inhibition of I(Ca) and I(Na). A “leak” current (I(leak)) was present indicative that other field targets were present in the plasma membrane.

14

J. Teissie

Structural Arrangements Biphotonic Microscopy Molecular center asymmetry is required for the creation of second harmonic generation (SHG) (second harmonic generation) signals. It makes this spectroscopy a powerful technique for visualizing changes in interfacial layers in the plasma membrane of biological cells. Lipophilic SHG probes, such as Di-4ANEPPDHQ (Di-4)(pyridinium, 4-(2-(6-(dibutylamino)-2-naphthalenyl)ethenyl)1-(3-sulfopropyl)-, hydroxide), can detect instantaneous/immediate perturbations in the plasma membrane. Rapid changes in membrane symmetry could be detected using SHG (Moen et al. 2014). Following pulsed electric fields exposure, an instantaneous drop of ~50% in SHG signal was detected only from the anodic pole of the cell. This dramatic loss in SHG signal reflects a local dramatic perturbation in the interfacial nature of the membrane.

Nuclear Magnetic Resonance (NMR) Spectroscopy 31P NMR spectroscopy gives structural informations on the lipid polar head region (nuclear magnetic resonance). Chinese hamster ovary (CHO) cells were reversibly permeabilized by submitting them to short, high-intensity, square wave pulses (1.8 kV/cm, 0.1 ms). Due to these pulsing conditions (high field, short duration), a large cap was electropermeabilized covering the majority of the cell surface, but the population viability was preserved. It was taken advantage from the observation that cells remained in a permeable state without loss of viability for several hours at 4  C. A differential method in the NMR pulse sequence was used in 1987 to analyze the phospholipid head groups and to get rid of the signals coming from other phosphorus groups. Control cells displayed the classical bilayer signature. A new anisotropic peak with respect to control cells was observed on 31P NMR spectroscopic analysis of the phospholipid components. This peak is only present when the cells are permeable, and normal anisotropy is recovered after resealing. This field-induced anisotropic peak was located downfield from the main peak associated to the phospholipids when organized in bilayers, but its localization peak is very different from the one of a hexagonal phase. This new transient signature associated to the electropermeabilized state can be explained by a new erected orientation of the lipid polar heads or by a random distribution of their orientation. This reorganization of the polar head group region results in a weakening of the hydration layer.

Electron Microscopy Ultrastructural changes of the plasma membrane were detected just after the pulse delivery by scanning electron microscopy (scanning electron microscopy). Cells were chemically fixed a few seconds after their exposure to electric pulses in order to

Mechanistic Description of Membrane Electropermeabilization

15

investigate the ultrastructural changes associated with the reversible permeabilization and observed by electron microscopy. By scanning electron microscopy, numerous microvilli and blebs were observed immediately after application of the field. Blebbing was facilitated under hypoosmolar conditions suggesting that they were assisted by the osmotic swelling induced by the membrane permeabilization (Gass and Chernomordik 1990). The appearance of osmotic pressure-dependent “blebs” was indicative of local weakening of the plasma membrane.

Fluorescence Microscopy Studies using fluorescence microscopy with 1-palmitoyl-2-{6-[(7-nitro-2-1,3benzoxadiazol-4-yl)amino]hexanoyl}-sn-glycero-3-phosphocholine (C6-NBD-PC), a phospholipid analog, monitored phospholipid scrambling of mammalian cells (fluorescence microscopy). Millisecond permeabilizing pulses induced membrane disorganization by increasing the translocation of phosphatidylcholines according to an ATP-independent process. These pulses resulted in a rapid phospholipid flip/flop within less than 1 s and were exclusively restricted to the regions of the permeabilized membrane. The electrically mediated local membrane disorganization was not associated with a loss of cell viability. The occurrence of such a flip-flop supports the existence of direct interactions between the movement of membrane zwitterionic phospholipids and the electric field.

Lateral Mobility of Membranes Lateral movement in each layer of the plasma membrane is present in all cell membranes. They are measured by using fluorescence recovery after photodegradation (FRAP). Electropermeabilization by ms long pulses affects the fast lateral mobility of a glycosylphosphatidylinositol (GPI)-anchored protein (Rae-1). The mobile fraction is indicative that 10–20% of the membrane surface is affected by defects under these conditions (where only a moderate field intensity was used to preserve the cell viability). These structures, which support the membrane permeability, propagate rapidly ( tc or E > Ec, including, the magnitude of S(t) or S(E) would be indistinguishable from zero. [This continuous logistic model was used by Fermi to describe the very sharp density distribution in an atom. It is very high and constant within the nucleus radius (rc) and practically zero beyond it.]

Equivalency of a PEF to Thermal Inactivation The widely used concept of F0 in thermal processing notwithstanding the correct criterion of microbial inactivation efficacy of a preservation process is the accomplished number of decades reduction in the targeted microbial population’s size. This

6

M. Peleg

Fig. 2 Construction of an equivalency curve between a PEF treatment and thermal inactivation at a constant reference temperature (Adapted from Corradini and Peleg (2010))

reduction, mathematically expressed as –Log10[Sfinal], is an observed quantity and more importantly model independent. When the survival ratio, or its base ten logarithm, is used as a process’s lethality measure, one can compare the efficacy of treatments of different kinds, thermal or nonthermal. To do the comparison all one needs is to construct an equivalency curve (Corradini and Peleg 2010). A schematic example of which is given in Fig. 2. The figure shows the translation of survival data recorded during a PEF treatment into corresponding times at a reference lethal temperature. Thus if it is known from experience that a certain isothermal (or nonisothermal) heat treatment is safe, one can design a PEF process that will match or exceed the thermal safety level with respect to the particular organism in the particular food or medium. Obviously, this can also be achieved by simply setting the targeted value of –Log10[Sfinal] and check whether it has been surpassed. The advantage of the equivalency plot is that it helps visualize what happens during the examined PEF process. It also provides a lethality measure expressed in term of equivalent time at the chosen temperature, which is familiar and intuitively meaningful to practitioners, i.e., food microbiologists and technologists. This “equivalent time” might be especially useful to highlight the benefits of the PEF technology, which is aimed at reducing the adverse effects of conventional heat treatments on the nutritional value and other quality attributes of thermally processed food.

Continuous Models of Microbial Injury PEF “survival curves” of the kind shown in Fig. 1 depict the three ways relationship between the number of recovered viable cells (colony-forming units or CFU’s), expressed as a survival ratio, the electric field’s intensity, and number of pulses. Recovery of the survivors is done in a favorable growth medium and under optimal growth conditions. Consequently, the count might include injured cells, which in the actual food under realistic storage conditions might not be able to repair the damage

Modeling Microbial Inactivation by Pulsed Electric Field

7

Fig. 3 Screen display of the Wolfram Demonstration where the “subtraction method” is used to determine the fraction of injured cells after an antimicrobial treatment

and would eventually perish. If so, an experimental survival curve so determined represents a worst-case scenario, and its use to establish the process’s safety provides an extra margin of safety. Nevertheless, it is not inconceivable that since the formed colonies are counted after 24–48 h of incubation, injured cells that would have been able to repair the damage if given longer time are not represented in the count. At least theoretically, therefore, such uncounted cells might still pose a health or spoilage hazard and it would be prudent to take this possibility into account. Apart from direct microscopic observation, the number of injured cells after a lethal treatment has been traditionally determined by counting the cells recovered under stressful conditions, which presumably only intact cells can withstand. This number is subtracted from that of the cells recovered under optimal conditions and the result is sought number of the injured cells. This method is based on the assumption that the injured cells are unable to form colonies under less than optimal conditions. The principle of the method is demonstrated in Fig. 3. The figure shows

8

M. Peleg

the screen display of an interactive Wolfram Demonstration posted on the Internet where the user can generate the two Weibullian survival curves: “total” – under optimal conditions, and “intact” – under stressful conditions, and the resulting (dashed) “injury curve.” To access the demonstration open: http://demonstrations. wolfram.com/InjuryInMicrobialInactivation/. [The program CDF Player which runs the demonstration (and over 10,850 other demonstrations to date) is freely downloadable following instructions on the screen.] The posted demonstration is based on the assumption that both the inclusive survival curve, i.e., that of the injured as well as intact cells and that of the intact cells only can be described by the Weibullian (“stretched exponential”) model. The validity of this assumption in thermal inactivation has been demonstrated by the analysis of published microbial mortality and injury data (Corradini and Peleg 2007). The analysis also indicated that most of the survivors were injured to at least some extent. Notice that the “subtraction method” on which this demonstration is based lumps together all cell injuries regardless of the damage severity. Theoretically at least, this weakness of the method can be remedied by recovering the surviving cells under conditions that create different stress levels. If and when done, the results would enable the researcher to quantify the numbers or fraction of surviving cells that sustain different degrees of injury. An example of a program that simulates cell injuries of various levels is given in Fig. 4. It shows the screen display of a different Wolfram Demonstration, especially written for the purpose and can be found at: http://demonstrations.wolfram.com/DegreesOfMicrobialInjuryAndSurvival/. Here too the assumption is that the inclusive survival curves, i.e., those of the injured (at the different levels) as well as the intact cells, and that of the “intact cells only” all follow the Weibullian model. Whether application of such a method, especially routinely, would be feasible in future PEF inactivation research is uncertain. However, the existence of such an option should be kept in mind, especially if unexpected survival is discovered long after the treatment. Apart from obvious logistic considerations, the main issue concerning the multiple recovery conditions incorporation in PEF studies would be the proper choice of recovery media and time-temperature conditions for the incubation. Since there is little known about the relationship between an organism’s degree of injury and its ability to recover in different media and incubation conditions, one would have to establish it experimentally, which might require a considerable time and effort. Again however, the concept should not be dismissed outright, not the least because it could be used to examining hypothetical situations using computer simulations based on assumed recovery efficiencies.

Stochastic Models of Mortality and Injury Almost universally, reported microbial survival curves have been determined by exposing a large population of the targeted vegetative cells (or spores) to the lethal agent under investigation and counting the survivors after different times. The same

Modeling Microbial Inactivation by Pulsed Electric Field

9

Fig. 4 Screen display of the Wolfram Demonstration that calculates the fractions of cells having three levels of injury

is true for PEF except that the independent variable has been the numbers of pulses at constant electric field intensity, electric field intensity at constant number of pulses, or energy at a constant numbers of pulses. The rationale for starting with a large inoculum is that it enables the detection and counting of survivors after the targeted population size has been reduced by several orders of magnitude. Obviously, this makes perfect sense and there is no reason to do otherwise. Yet, the number of a virulent pathogen’s cells actually encountered in foods, unless intentionally contaminated, rarely if ever matches the huge numbers used in microbial inactivation studies. It would therefore be interesting to investigate at least some of the potential implications of the relation between the inactivation (and growth) patterns of individual or very small groups of cells and those of large cell populations of the same organism in the same food or a surrogate medium.

10 Fig. 5 The Markov tree of microbial cells mortality by PEF. Notice that the mortality probability Pmi and hence that of the survival 1- Pmi can vary with the number of pulses

M. Peleg

Markov tree of cell mortality No. of Pulses (i)

Dead [Pm(i)]

Intact [1- Pm(i)]

I =0 i=1 i=2 i=3 i=4

A Stochastic Model of Microbial Cells Mortality Consider a scenario where a microbial cell exposed to PEF can be in one of only two states, alive (countable) and dead (uncountable). After the first pulse it can be dead with a mortality probability rate Pm1 and alive at a probability 1- Pm1 as shown in the Markov tree presented in Fig. 5. If the cell has survived the first pulse, then after the second pulse it can be dead at a probability Pm2 and alive (1- Pm2), and so on after pulses i = 3, 4, . . ., n. The probabilities, Pm1, Pm2, Pm3, Pm4, . . ., Pmn, need not be the same and one would expect that they will rise as the cell is weakened by the repeated assault. This scenario can be simulated by drawing a random number between zero and one, which will determine whether after a given pulse i the cell dies and leaves the game, or survive for the next pulse, depending on the mortality probability Pmi. The process can be repeated for a number of cells, each tested with a new set of drawn random numbers, and the total number of survivors after each pulse recorded. A plot of these survivors’ numbers divided by the initial number vs. the pulse’s number will be the corresponding survival curve. Figure 6 shows that, at least qualitatively, almost all the common microbial inactivation patterns encountered in food preservation can be expressed in terms of an underlying Pm(t), a function that describes how the mortality probability rate changes with time (Horowitz et al. 2010; Corradini et al. 2010). The procedure to generate such survival curves with the stochastic model, which allows the user to choose the initial number of cells, has been automated and posted on the Internet as a Wolfram Demonstration, open: http://demonstrations.wolfram.com/ProbabilisticMo delForMicrobialMortality/. Its screen displays for a large initial population is shown in Fig. 7 and for a small population in Fig. 8. Although the demonstration was written for arbitrary time units instead of the number of pulses, it can still be used to highlight the difference between the survival patterns of large and small microbial populations. As expected, the simulated survival curve of a large population shown

Modeling Microbial Inactivation by Pulsed Electric Field Log-linear survival curves without and with a flat “shoulder” “First order kinetics”

“First order kinetics” with “shoulder” 1

“Tailing” with finite number of survivors

1

0.8

0.8

0.8

0.6

0.6

0.6

0.6

0.4

0.4

0.4

0.4

0.2

0.2

0.2

0.2

0

0

0

0

5

10

15

20

0

5

10

15

20

10

20

30

0

0

10

20

30

0

10

20

30

1

1 Survival Ratio S(t)

1

Survival Ratio S(t) 0

5

10

15

20

0

0

0

–2

–2

–4

0

5

10

15

0

20

0

10

20

30

0

0 0 –2

LogS(t)

0

0

Complete inactivation (no survivors after td)

1

0.8

1

LogS(t)

“Absolute death time”

Tailing with residual survival

Pm(t)

Pm(t)

1

11

–4

–1 –4 –2

–6 –8

–6

–6

0

5 10 15 20 Time (arbitrary scale)

–8

–3 0

5 10 15 20 Time (arbitrary scale)

0

10 20 30 Time (arbitrary scale)

–8 0

td1 td2 td3 30 10 Time (arbitrary scale)

Fig. 6 Simulated microbial inactivation patterns produced with different mortality rate functions

in Fig. 7 (n = 1000) is smooth and deterministic until there is only a score of survivors left. In contrast, the curves of small groups of cells shown in Fig. 8 (n = 10) – compare plots a and b – are visibly jugged and more importantly irreproducible – compare the left and right displays generated with exactly the same model parameters. Notice that if Pm(t) or Pmi) = constant, the result is log-linear inactivation (Fig. 7). If Pm(t) or Pmi) varies with time as shown in Fig. 8, the survival curve has a “shoulder” regardless of whether the population is large or small. What is important is that one can retrieve the probability rate function Pm(t) or Pmi) from experimental survival data obtained with a large cell population and use it to generate realistic scenarios involving individual or small groups of cells [ibid].

A Stochastic Model of Injury and Mortality The described approach can be extended to situations where the inactivation is manifested in both injury and mortality. In such cases, the Markov tree would be based on two underlying probability rate functions, Pm(t) and Pinjury(t), or for PEF Pmi) and Pinjuryi). These probabilities are allowed to vary independently with time or

12

M. Peleg

Fig. 7 Screen display of the Wolfram Demonstration showing the mortality curve of a large cell population (n = 1000) that follows first order inactivation kinetics and no injury, generated with the stochastic model. Notice the log-linear curve’s smoothness until only a few survivors are left

the number of pulses – see Fig. 9, which also describes the resulting deterministic survival curve of a relative large population (n = 1000), except when it reaches only a score of survivors. Examples of the concept’s use to simulate microbial injury and mortality of a small microbial population (n = 10) are given in Fig. 10. This figure provides another demonstration of the irregularity and irreproducibility of the survival pattern of very small microbial populations when exposed to a lethal agent – compare plots a and b. Figures 9 and 10 describe a grossly oversimplified and perhaps even unrealistic scenario where there is only a single level of injury, and where the mortality and injury probability rate Pm(t) and Pinjury(t), respectively, or in the case of PEF the actual probabilities Pmi) and Pinjuryi) are only functions of time, or number of pulses, regardless of position on the Markov tree’s branch. In reality, one would expect both the existence of different degrees of injury, and that cells which have already sustained damaged by a previous pulse or pulses could become more vulnerable to the treatment.

Modeling Microbial Inactivation by Pulsed Electric Field

13

Fig. 8 Two screen displays of the Wolfram Demonstration showing the mortality curves of a very small cell population (n = 10) generated with the stochastic model. Notice that the curves are both irregular and irreproducible

In principle, one can introduce several injury levels into the PEF mortality/injury model as shown schematically in Fig. 11, with corresponding new probabilities Pmi) and Pinjuryi) at the beginning of each branch of the Markov tree. Either way, the resulting stochastic model would be so cumbersome mathematically, and the assignment of the different probabilities so difficult, that the model might be rendered impractical even for computer simulations. It also remains to be seen whether the simplified version of the stochastic mortality/injury model that was used to generate the survival patterns shown in Figs. 9 and 10 can capture the essence of microbial inactivation by PEF at least qualitatively. But even without such experimental demonstration, simulations of the kind that the Wolfram Demonstration offers can still provide some insight into how the treatment affects small and large microbial cell populations and how different survival patterns emerge.

Recovery from Injury and Delayed Growth In principle, the stochastic approach can be extended to cells’ recovery after being damaged by a PEF treatment. It can also be extended to cells division in the

14

M. Peleg

Fig. 9 Screen display of the Wolfram Demonstration showing survival curve of a large cell population (n = 120) that suffers mortality and injury, generated with the stochastic model. Notice the logarithmic curve’s smoothness until only a few countable survivors are left

survivors, be they those who succeeded to remain intact or those who managed to repair the damage. Schematic view of the Markov trees of such scenarios is given in Fig. 12. If the approach is adopted in future research, how to assign the corresponding probabilities and time dependencies will be a serious challenge. But regardless of the model chosen, i.e., with or without the distinction between damage severity levels and whether cell division commences or not, time would have to replace the number of pulses as the independent variable. In other words, the steps in the stochastic process, the Δt’s, would have to be expressed in pertinent time units such as hours or days which are orders of magnitudes larger than the treatment duration let alone that of an individual pulse. The probabilities of mortality, recovery, and cell division would then become probability rates, i.e., the probability that a cell’s state would change during the chosen unit time. Again, the practicality of the probabilistic approach to such scenarios is yet to be demonstrated. But, again, even if it would never serve as a starting point for any quantitative model development, it could still highlight the theoretical possibilities that certain unexpected patterns of

Modeling Microbial Inactivation by Pulsed Electric Field

15

Fig. 10 Two screen displays of the Wolfram Demonstration showing the survival curves of a very small cell population (n = 10) generated with the stochastic model. Notice that the curves are both irregular and irreproducible

Pulse number

Markov tree of cell mortality and two injury levels

i=1 i=2 i=3 i=4 intact cell

slightly injured cell

gravely injured cell

dead cell

Fig. 11 The Markov tree of an intact cell that can have two levels of injury. Notice that the mortality and injury probabilities are likely to be functions of the number of pulses

16

M. Peleg

Markov trees of cell mortality, injury, recovery and division a Recovery of a slightly

b Recovery of a severely

c Growth of an intact or

injured cell

injured cell

recovered cell

Fig. 12 The Markov trees of three scenarios: the recovery of a slightly injured cell (left), the recovery of a severely injured cell (middle), and the division of an intact surviving cells (right). Notice that the dividing cell can be a recovered injured cell

recovery and growth can emerge. This is probably true not only for assessing the microbial safety of foods but also for at least certain medical applications of the PEF technology where the possibility of damage repair and renewed cell division cannot be totally ruled out (Garcia et al. 2014; Dermol and Miklavčič 2015).

Conclusions Traditional survival models used in quantitative microbiology can be adapted for the dose–response curves encountered in food preservation by PEF. Two prominent models, the Weibullian and logistic (or Fermi’s distribution), are particularly suitable for the purpose, and in many situations can be used interchangeably. For reference or quantitative comparison, a dose–response (or survival) curve observed in a PEF treatment can be converted into an equivalent isothermal survival curve at a chosen lethal reference temperature by a simple graphical procedure. Unlike the traditional F0 which is used as measure of a thermal process’s lethality, the procedure is based on the number of decades reduction in the targeted organism and it is model independent, i.e., neither the thermal inactivation nor the lethality caused by the PEF treatment has to follow first order or any other particular kind of kinetics. The kinetics of microbial inactivation by PEF treatments, as by other lethal processes including heat, can be described by stochastic (probabilistic) models where the underlying probability of a living cell dying (or become uncountable) is a function of the pulse number. For large populations, such models predict smooth and deterministic survival patterns, except when the targeted population is reduced to the last few individual cells. Simulations of small groups of cells, such as of pathogenic bacteria when present in foods, show that the survival curve is inherently

Modeling Microbial Inactivation by Pulsed Electric Field

17

irregular and irreproducible which is consistent with many observed patterns of food spoilage and poisoning episodes. Both the deterministic and stochastic models can be extended to include injury, recovery from injury, and even delayed cell division. The practicality of such models is in doubt, but they can be used for simulations with assumed parameters. Although such simulations are unlikely to provide quantitative predictions, they might provide insight into how events at the cellular level translate into inactivation and survival patterns at the population level.

Cross-References ▶ Application of Numerical Simulation Techniques for Modeling Pulsed Electric Field Processing ▶ Application of Pulsed Electric Field for Treatment of Fish and Seafood ▶ Atomic Force Microscopy for Electroporation Mechanisms Studies in Bacteria ▶ Bacteria Cell Wall: Description, Role in Transport, and Effect of Electroporation ▶ Cell Death Due to Electroporation ▶ Detection of Electroporation in Microbial Cells: Techniques and Procedures ▶ Different Cell Viability Assays Following Electroporation in Vitro ▶ Effect of Pulsed Electric Fields on Food Constituents ▶ Effect of Pulsed Electric Fields on Food Proteins ▶ Electroporation of Biofilms ▶ Electrotransfection of Bacteria: Electrotransformation ▶ Food Structure Changes Affected by Pulsed Electric Field Treatment ▶ History of Pulsed Electric Field Research and Applications ▶ Inactivation of Bacteria Spores by Pulsed Electric Fields ▶ Mathematical Models Describing Cell Death Due to Electroporation in Vitro ▶ Mathematical Models of Pulsed Electric Field Treatment of Plant tissues and Simulation of Related Phenomena ▶ Parameters Affecting Cell Viability Following Electroporation in Vitro ▶ Pore Lifetime and Permeabilization Lifetime in Models ▶ Process Design, Improvement and Integration of Pulsed Electric Field processes ▶ Pulsed Electric Field Treatment for Beverage Production and Preservation ▶ Pulsed Electric Field Treatment for Fruit and Vegetable Processing ▶ Pulsed Electric Fields Assisted Extraction of Proteins from Bacterial Cells ▶ Pulsed Electric Fields Effects on Lycopene, Vitamin C, and Antioxidant Capacity of Tomato Juice ▶ Pulsed Electric Fields for Inactivation of Endogenous Enzymes in Foods ▶ Pulsed Electric Fields for Pasteurization: Defining Processing Conditions ▶ Responses of Plant Cells and Tissues to Pulsed Electric Field Treatments ▶ Stress Induction and Response, Inactivations and Recovery of Vegetative Microorganisms by Pulsed Electric Field Treatment ▶ Techniques to Detect Electroporation in Food Tissues ▶ Temperature Increase and Thermal Effects Due to Electroporation in Tissues

18

M. Peleg

References Alvarez I, Raso J, Sala FJ, Condon S (2003) Inactivation of Yersinia enterocolitica by pulsed electric fields. Food Microbiol 20:691–700 Corradini MG, Peleg M (2007) A Weibullian model of microbial injury and mortality. Int J Food Microbiol 119:319–328 Corradini MG, Peleg M (2010) Comparing the effectiveness of thermal and non-thermal preservation processes: the concept of equivalent efficacy. In: Doona CJ, Kustin K, Feeherry FE (eds) Case studies in novel food processing technologies. Woodhead, Cambridge, UK, pp 464–488 Corradini MG, Normand MD, Peleg M (2010) A stochastic and deterministic model of microbial heat inactivation. J Food Sci 75:R59–R70 Dermol J, Miklavčič D (2015) Mathematical models describing Chinese hamster ovary cell death due to electroporation in vitro. J Memb Biol 248:865–881 Garcia PA, Davalos RV, Miklavčič D (2014) A numerical investigation of the electric and thermal cell kill distributions in electroporation-based therapies in tissue. PLoS One 9(8), e103083 Heinz V, Alvarez I, Angersbacha A, Knorr D (2002) Preservation of liquid foods by high intensity pulsed electric fields – basic concepts for process design. Trends Food Sci Technol 12:103–111 Horowitz J, Normand MD, Corradini MG, Peleg M (2010) A probabilistic model of growth, division and mortality of microbial cells. Appl Environ Microbiol 76:230–242 Hulsheger H, Potel J, Niemann E-G (1983) Electric field effects on bacteria and yeast cells. Radiat Environ Biophys 22:149–192 Kotnik TW, Sack M, Haberl Meglič S, Peterka M, Miklavčič D (2015) Electroporation-based applications in biotechnology. Trends Biotechnol 33:480–488 Pataro G, Barca GMJ, Donsi G, Ferrari G (2015) On the modeling of electrochemical phenomena at the electrode- solution interface in a PEF treatment chamber: methodological approach to describe the phenomenon of metal release. J Food Eng 165:34–44 Peleg M (1995) A model of microbial survival after exposure to pulsed electric fields. J Sci Food Agric 67:93–99 Peleg M (1996) Evaluation of the Fermi equation as a model of dose–response curves. Appl Microbiol Biotechnol 46:303–306 Peleg M, Cole MB (1998) Reinterpretation of microbial survival curves. Crit Rev Food Sci Nutr 38:353–380 Peleg M, Normand MD, Damrau E (1997) Mathematical interpretation of dose–response curves. Bull Math Biol 59:747–761 Rodrigo D, Barbosa-Canovas GV, Martinez A, Rodrigo M (2003) Weibull distribution function based on an empirical mathematical model fro inactivation of Escherichia coli by pulsed electric fields. J Food Prot 6:911–1099 Toepfl S, Heinz V, Knorr D (2007) High intensity pulsed electric fields applied for food preservation. Chem Eng Process 46:537–546 van Boekel MAJS (2002) On the use of the Weibull model to describe thermal inactivation of microbial vegetative cells. Int J Food Microbiol 74:139–159 Vega-Mercado H, Martin-Belloso O, Qin B-L, Chang FJ, Gongora-Nieto MM, Barbosa-Canovas GV, Swan BG (1977) Non-thermal food preservation: pulsed electric fields. Trends Food Sci Technol 8:151–157 Wouters PC, Alvarez I, Raso J (2001) Critical factors determining inactivation kinetic by pulsed electric field food processing. Trends Food Sci Technol 12:112–121 Zhang Q, Barbosa-Canovas GV, Swanson BG (1995) Engineering aspects of pulsed electric fields pasteurization. J Food Eng 25:261–281 Zhao W, Yang R, Shen X, Zhang SH, Chen X (2013) Lethal and sublethal injury and kinetics of Escherichia coli, Listeria monocytogenes and Staphylococcus aureus in milk by pulsed electric fields. Food Control 32:6–12

Preclinical Studies on Electrochemotherapy Gregor Sersa, Masa Bosnjak, Maja Cemazar, and Richard Heller

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preclinical Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In Vitro Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In Vivo Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanisms of Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroporation Mediated Drug Uptake in the Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Antivascular Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Induction of Immune Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 3 5 9 9 10 10 12 13 13 14

Abstract

Electrochemotherapy is the combination of chemotherapy and local application of electric pulses. Application of electric pulses, i.e., electroporation, is used as a means to increase drug delivery to cells. Electrochemotherapy has the potential to

G. Sersa (*) • M. Bosnjak (*) Department of Experimental Oncology, Institute of Oncology Ljubljana, Ljubljana, Slovenia e-mail: [email protected]; [email protected] M. Cemazar (*) Department of Experimental Oncology, Institute of Oncology Ljubljana, Ljubljana, Slovenia University of Primorska, Faculty of Health Sciences, Izola, Slovenia e-mail: [email protected]

# Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_45-1

1

2

G. Sersa et al.

improve the efficacy of drugs that have impeded transport through the plasma membrane. While enhancement has been demonstrated for some chemotherapeutic drugs that were tested, only cisplatin and bleomycin have been evaluated clinically. Recently, calcium electroporation has also entered clinical testing. The increased cytotoxicity of bleomycin and cisplatin was demonstrated by exposure of cells to electric pulses in vitro, and potentiation of their antitumor effectiveness in vivo was also demonstrated. Preclinical studies have explored several different parameters that could affect effectiveness of electrochemotherapy, such as the dose and route of drug administration, timing of application of electric pulse, and coverage of the tumor with electric field. These aspects were all evaluated on different primary or transplanted tumors models in mice, rats, and hamsters. Most importantly, the underlying mechanisms of electrochemotherapy were explored, from increased drug uptake to vascular effects and involvement of immune system. These findings enabled swift translation of electrochemotherapy into human and veterinary oncology.

Keywords

Electrochemotherapy • Bleomycin • Cisplatin • Cells • Experimental tumors • Blood flow • Immune response

Introduction Treatment of cancer requires a multidisciplinary approach since several different types of treatment are used in combination, simultaneously or sequentially. In general, there are three major treatment modalities available: surgery, radiotherapy, and chemotherapy. While surgery and radiotherapy represent good local ablative treatments, chemotherapy is a systemic treatment modality. Drugs used for chemotherapy in general easily cross the plasma membrane and are cytotoxic once they reach their intracellular targets. However, there are still a few chemotherapeutic drugs that have hampered transport through the plasma membrane. Those drugs may be good candidates for electrochemotherapy. Electrochemotherapy is a local ablative treatment modality combining chemotherapy and application of electric pulses (i.e., electroporation) to the tumor tissue. Electroporation performed at the time of the highest extracellular drug concentration facilitates the uptake of hydrophilic chemotherapeutic drugs into the cells, which leads to good antitumor effectiveness (Sersa et al. 2008a). Therefore, electrochemotherapy possesses the qualities of an ablative technique, although it uses chemotherapeutics that can be given systemically, but are in such low doses that they act only at the site of electroporation and have no or minimal systemic effectiveness and negligible or no side effects.

Preclinical Studies on Electrochemotherapy

3

Preclinical Data In Vitro Studies Electroporation is used to facilitate transport of different molecules across the plasma membrane with an aim to study different biochemical and pharmacological phenomena. The application of external electric field changes the transmembrane potential of plasma membrane, and when the threshold value is exceeded, structural changes of the plasma membrane occur. Thus, the plasma membrane permeability is increased (Yarmush et al. 2014). In the case of chemotherapeutic drugs, intracellular drug accumulation is increased and it results in increased drug cytotoxicity. Since electroporation is used to facilitate drug transport through the cell membrane only for poorly permeant or nonpermeant molecules, electrochemotherapy is limited to hydrophilic drugs and/or drugs that lack a transport system to enter the cell. Among chemotherapeutic drugs several candidates were tested in vitro for potential application in electrochemotherapy (Jaroszeski et al. 2000). Electroporation proved to be an effective delivery method to introduce and increase intracellular concentration of several chemotherapeutic drugs, such as bleomycin, cisplatin, melphalan, methotrexate, doxorubicin, etc. (Table 1). Among these, only bleomycin and cisplatin were identified as potential candidates for electrochemotherapy of cancer patients (Miklavcic et al. 2014). Bleomycin is a hydrophilic molecule with very restricted transport through the cell membrane, limited almost exclusively to endocytosis; its cytotoxicity after electroporation can be potentiated up to 700 times (Orlowski et al. 1988). Several other anticancer drugs were also tested, among them also melphalan, netropsin, methotrexate, and Actinomycin D, but neither of them was as potentiated as bleomycin after electroporation. Survival of different tumor cell lines was evaluated in in vitro studies on electroporation combined with bleomycin or cisplatin. Increased uptake of the molecules after exposure of cells to electric pulses was also demonstrated. This extraordinary potentiation of bleomycin cytotoxicity was systematically demonstrated in an in vitro study (Orlowski et al. 1988). In in vivo studies the effect was documented on electrochemotherapy (Okino and Mohri 1987; Mir et al. 1991), including clinical application (Sersa et al. 2008a). Cisplatin similarly possesses hampered transport through the cell membrane, which is mainly regulated by transporters controlling intracellular copper homeostasis, Ctr1 for influx and ATP7A and ATP7B for efflux, and partially also by passive diffusion. Electroporation increases cisplatin cytotoxicity by up to 70-fold (Sersa et al. 1995). Therefore, extensive preclinical and clinical research focused on electroporation predominately with bleomycin and cisplatin. Studies of electrochemotherapy with bleomycin, cisplatin, and doxorubicin have also been conducted on 3D multicellular tumor spheroids, to visually and molecularly assess the effect of electrochemotherapy. Electrochemotherapy of all three tested drugs led to macroscopic morphological changes of spheroids, tumor spheroid

4

G. Sersa et al.

Table 1 Drugs tested for in vitro and in vivo potentiation in combination with electroporation (Adapted with permission from Miklavcic et al. (2014)) Drug tested Bleomycin Cisplatin Calcium Netropsin Carboplatin 2-N-methyl-9-hydroxy-ellipticinium (NMHE) Vincristine Actinomycin D Cytarabine Oxaliplatin Platinum (II) complex 3P-SK Platinum (II) complex PtAMP Mitomycin C Vinblastine 5-fluorouracil Paclitaxel Doxorubicin Ruthenium-based anticancer agent KP1339 Nimustine hydrochloride (ACNU) Methotrexate 9-OH-ellipticine Didemnin B Melphalan Mithramycin Taxotere Daunorubicin Adriamycin Etoposide Ancitabine Gemcitabine

In vitro potentiation Yes; 100–5000-fold Yes; 1.8–70-fold Yes; more hundred-fold Yes; 200-fold Yes; 1.6–13-fold Yes; 4-fold

In vivo potentiation Yes Yes Yes – – –

Yes; 1.3–3.4-fold Yes; 2–3-fold Yes; 2-fold Yes No No Yes but low; 1.3–1.4-fold Yes but low; 1.1–1.3-fold No or low; 1.25-fold No Yes but low; 0.67-fold No No No No No No No No No No No or ND ND ND

– – – – Yes – – – – – – Yes – – – – – – – – No – – –

ND – drug was tested but the potentiation could not be determined due to methodological limitations

growth arrest, and finally to its complete apoptosis-mediated dissociation. Morphology of spheroids was not changed after drug treatment alone. After electrochemotherapy individual cells or clusters separated from spheroid surface. Electrochemotherapy first led to a decrease in size of spheroid and then to its total destruction, regardless of the drug used. The study also showed that underlying mechanism of cell death and spheroid growth arrest involves apoptosis. The results

Preclinical Studies on Electrochemotherapy

5

on spheroids are very valuable, because the spheroid model mimics the behavior of tumors very well (Gibot et al. 2013). In recent years in vitro electroporation studies focused on calcium electroporation. Dose-dependent decrease in cell viability was illustrated for all tested cell lines. Cell viability after calcium electroporation was decreased to a similar level as cell viability after 0.1 μmol/L bleomycin electroporation. Further studies with calcium electroporation on spheroids demonstrated reduced size of three different cancer cell spheroids: human colon, breast, and bladder cancer cell spheroids. In contrary, calcium electroporation had no effect on the size of human fibroblast cell spheroids, thus indicating a different effect on normal and malignant cells (Frandsen et al. 2015). The effectiveness of electrochemotherapy with bleomycin was tested also on BRAF V600E mutated melanoma cells (SK-MEL-28) and nonmutated melanoma cells (CHL-1). The survival of both cell lines was reduced, with more pronounced effect of electrochemotherapy on BRAF V600E mutated SK-MEL-28 cells, which required half lower concentration of BLM at IC90 compared to nonmutated cells CHL-1. Furthermore, concomitant treatment with electrochemotherapy with bleomycin and 0.5 μM vemurafenib was also tested in vitro. The potentiation of combined treatment was more than additive; it was calculated to be synergistic (Dolinesk et al. 2016).

In Vivo Studies Based on encouraging antiproliferative results of bleomycin and cisplatin after cell electroporation in vitro, these two drugs were the first tested on murine tumor models and later translated to veterinary oncology. The antitumor effectiveness of electrochemotherapy was demonstrated on transplantable or spontaneous tumors in mice, rats, rabbits, hamsters, cats, dogs, and horses. Extensive studies on tumors in different locations, either cutaneous or subcutaneous or located in the muscle, brain, or liver, were performed. The treatment was tested on tumors of different histology: melanomas, carcinomas, gliomas, and sarcomas (Gothelf et al. 2003; Sersa et al. 2008b). The effectiveness of electrochemotherapy relies on the chemotherapeutic drug administration that by itself does not have any antitumor effectiveness, and application of electric pulses to the tumor that do not exert any antitumor effectiveness as well, but when combined together excellent antitumor effectiveness is obtained (Fig. 1). Therefore, the effectiveness of electrochemotherapy relies on two factors: the drug and the electric field. The appropriate drug dose and administration, timing of the application of electric pulses is needed, as well as the parameters of electroporation and coverage of the tumor with electric field (Fig. 2). Based on the extensive preclinical studies, several important factors were identified that need to be known and taken into consideration when performing electrochemotherapy.

6

G. Sersa et al.

Fig. 1 Example of effective electrochemotherapy in B16F10 tumors. Cisplatin was given intratumoraly (4 mg/kg) and 8 electric pulses were applied to the tumor with plate electrodes. Electric pulses were applied in two directions: 4 pulses in one and the other 4 in the perpendicular direction. Six days after treatment antitumor effectiveness of electrochemotherapy with cisplatin is evident, compared to cisplatin only group

• Two different administration routes are used in electrochemotherapy: intravenous or intratumoral injection of chemotherapeutic drug. Each of the routes has its advantages and disadvantages. The intravenous route provides better drug distribution in the tumors, when they are well vascularized, whereas the intratumoral route can provide higher drug concentration in the tumors and can be used in small and poorly vascularized tumors, such as recurrent tumors in pre-irradiated area. • At the time of application of electric pulses sufficient amount of the drug at the tumor site is mandatory. After intravenous drug administration (5 mg/kg of bleomycin or 4 mg/kg of cisplatin), the maximal drug concentration in the tumors is reached in small laboratory animals after 3 min. Even shorter interval is needed

Preclinical Studies on Electrochemotherapy

7

Fig. 2 Schematic presentation of electrochemotherapy. First the drug (yellow circles) is injected intravenously or intratumorally. At the time of maximal chemotherapeutic drug concentration at the tumor site, electric pulses are applied either by plate or needle electrodes. The whole tumor and appropriate safety margins should be covered with a sufficient electric field for optimal electroporation of cells in the tumors

after intratumoral administration of cisplatin, so the electric pulses have to follow the administration of the drug within a minute, while for intratumoral administration of bleomycin the optimal time of application of electric pulses after intratumoral injection was 10–15 min (Heller et al. 1997; Miklavcic et al. 2014; Sersa et al. 2008a). • Electrode selection is crucial for achieving good tissue electroporation and thus good antitumor effectiveness. Electrode geometry and tissue composition control electric field distribution in the tissue, which is responsible for plasma membrane permeabilization. The electric field distribution in the tissue can be improved by rotating the electrodes so that higher fraction of cells is being permeabilized. Cutaneous and subcutaneous tumors can be effectively treated by plate electrodes, whereas appropriate electric field distribution in deep-seated tumors is assured by using needle electrodes (Miklavcic et al. 2014). The advantage of the plate electrodes over needle electrodes is that they are noninvasive. • Electric pulse parameters, amplitude, number, duration, and frequency of the electric pulses are very important for antitumor effectiveness. Several studies showed that for surface tumors appropriate amplitude over distance ratio for safe and effective electrochemotherapy, when using plate electrodes, is between 1000 and 1500 V/cm. Amplitude over distance ratio below 1000 V/cm does not cause

8

G. Sersa et al.

tumor electroporation, whereas above 1500 V/cm normal tissue damage around tumor occurs due to irreversible cell permeabilization (Sersa et al. 1995; Miklavcic et al. 2014). Most of the studies demonstrated good antitumor effectiveness without tissue damage at amplitude over distance ratio of 1300 V/cm. Electric pulse parameters for other types of electrodes need to be numerically calculated to ensure appropriate electric field distribution in the tumor tissue at coverage of 400 V/cm or higher for 8  100 μs. Normally 8 electric pulses of 100 μs are used for electrochemotherapy, but minimum 4 pulses should be used (Gothelf et al. 2003; Miklavcic et al. 2014). Repetition frequencies of the pulses are either 1 Hz or 5 kHz. If the drug concentration in the tumor is sufficient, both frequencies cause equal antitumor effect (Sersa et al. 2010). In vivo experiments in animal models provided necessary data to prove that electrochemotherapy with bleomycin or cisplatin is effective in the treatment of solid tumors. Good antitumor effectiveness was obtained in all tumor models tested, with drug concentrations that had no antitumor effectiveness by itself. Furthermore, the application of electric pulses alone does not exert antitumor effectiveness but is only the means for drug delivery to cells. Antitumor effectiveness can be achieved with irreversible electroporation, where pulses only are responsible for the ablative effect, but at higher amplitude and more pulses need to be delivered. Complete or partial response is achieved after a single treatment, although repetitive treatment was possible as well. The therapy was demonstrated to be effective on a variety of tumor histotypes, based on the simple principle that with electroporation, as a physical method, cells in tumors can be electroporated and that bleomycin or cisplatin can thus exert cytotoxicity on any kind of the tumor type. However, since early studies variable effectiveness of electrochemotherapy on different tumor models has been observed. Though electrochemotherapy is effective on most tumor types, underlying mechanisms of differential susceptibility of the tumors has not been explored. The most plausible option is that different tumor cells have different intrinsic susceptibility to electrochemotherapy as observed also in in vitro studies. Cemazar et al. showed different IC50 for cisplatin and bleomycin in several different tumor cells and also in endothelial cells (Cemazar et al. 2001a). The difference in in vitro cell sensitivity could be due to intrinsic sensitivity of the cells to different antitumor drugs. Mechanisms responsible for the tumor cell’s resistance to bleomycin could be related to higher effectiveness of electrochemotherapy in sarcoma cells SA-1 than in carcinoma cells EAT were already identified. Difference in cell susceptibility to bleomycin could be due to increased DNA repair, decreased drug accumulation, metabolic inactivation with hydrolase, or partially also due to defects in membrane protein involved in bleomycin endocytosis. Furthermore, the difference in susceptibility was observed also between the parental and chemoresistant cell line (Cemazar et al. 2001b). In vivo studies on murine models confirmed these in vitro observations. Beside mechanisms already discussed in vitro in tumor cell models in vivo have also additional mechanisms contributing to overall effectiveness. Thus, the histological characteristics of the tumors have to be considered as well, predominantly the

Preclinical Studies on Electrochemotherapy

9

vascularization of tumors and the composition of extracellular matrix. Further studies are needed in the future also to identify mechanisms involved in differential effect on tumor versus normal cells or tissue.

Mechanisms of Action Electroporation of targeted cells is the principal mechanism of electrochemotherapy. Drug concentration in the cells is significantly increased and thus can efficiently reach the intracellular target. However, due to the permeabilization of other stromal cells, like endothelial cells, also the vascular effects of electrochemotherapy are present (Jarm et al. 2010). Furthermore, the nonthermal effects of electroporation enable the preservation of tumor antigens, which are shed in the tumor environment, and can elicit local immune response. All three mechanisms of action are involved in overall antitumor effectiveness of electrochemotherapy.

Electroporation Mediated Drug Uptake in the Cells Although there were several different chemotherapeutic drugs tested in vitro for combination with electrochemotherapy, two most prominent and commonly used drugs, bleomycin and cisplatin, will be discussed here. Bleomycin is a nonpermeant drug that when internalized exerts high cytotoxicity. It was demonstrated that on average, 200 internalized molecules of bleomycin is sufficient to kill cells. Bleomycin induces single and double DNA strand breaks. There are many theories, but the exact mechanism of DNA strand scission still remains unresolved. The reaction between DNA and bleomycin is universal and will occur in all cells regardless of tumor type, even in normal cells. Two types of cell death were observed: mitotic cell death and pseudoapoptosis. In vivo slow death, mimicking in vitro mitotic cell death, was observed. Mitotic cell death is specific for dividing cells and in the context of higher dividing rate of tumor cells compared to normal cells the differential susceptibility of tumor versus normal tissues is postulated. In vivo studies also clearly demonstrated that exposure of tumors to electroporation significantly increases bleomycin accumulation in the tumors, with prolonged retention in them (Mir et al. 1996). Recently, a new analytical method for the determination of bleomycin in serum and tissue samples was developed that may in further studies be used and help to explore the pharmacokinetics and pharmacodynamics of bleomycin and provide further data that could be usefully exploited in the clinics (Kosjek et al. 2016). Unlike bleomycin cisplatin is not classified as nonpermeant, but its complexes seem to be low permeant. Normally cisplatin enters the cells by passive diffusion and through copper transporters (Dilruba and Kalayda 2016). After electropermeabilization cisplatin is up to 70-fold more cytotoxic for the cells, proportional to the higher uptake of cisplatin (Sersa et al. 1995). It interferes with DNA replication; forms inter- and intrastrand DNA adducts; and thus halts proliferation of the cells

10

G. Sersa et al.

and induces apoptosis. Since enhanced proliferation is typical for tumor cells, cisplatin enters tumor cells to a greater extent. Also increased cisplatin accumulation in tumors after electroporation was demonstrated in experimental tumors. Furthermore, the increased DNA binding of cisplatin in tumors was demonstrated in the range of a factor of 2–3 (Cemazar et al. 1999). All these data indicate that electroporation of tumors increases drug accumulation in tumors and undoubtedly in tumor cells. The drugs, bleomycin or cisplatin by their DNA damaging effect, therefore exert the antitumor action and induce slow, mitotic or apoptotic cell death of tumor cells.

Antivascular Effects Early studies demonstrated that exposure of tumors to electroporation had tumor blood flow modifying effect. The electroporated area becomes whitish, for a certain period of time with restoration of blood flow after 10–15 min (Sersa et al. 1999). These early studies demonstrated that the effect is not confined to tumor vasculature but is expressed also in normal tissue, where it is however less pronounced and shorter lived (Gehl et al. 2002). Later studies demonstrated also increased tumor vascular permeabilization. Both phenomena contribute to better drug distribution and longer retention of the drug in the tumors, known also as vascular lock. Since all stromal cells in the electric field induced by electroporation can be permeabilized, several studies have explored the vascular disrupting action of electrochemotherapy. A body of evidence demonstrates that endothelial cells are permeabilized and undergo apoptosis, thus hampering the perfusion of the tumor vessels (Sersa et al. 2008b). The effect is confined predominantly to smaller vessels, while bigger ones, due to the more robust scaffold, remain intact (Markelc et al. 2013). The size of the vessels that are not affected by electroporation or electrochemotherapy are not determined yet but need to be in further studies. Similar phenomena are observed also after irreversible electroporation. This abrogation of tumor perfusion then results in ischemia and induces a secondary cascade of tumor cell death. Based on these findings, a model of vascular effects of electroporation and electrochemotherapy was proposed (Jarm et al. 2010). Later findings also pointed to differential susceptibility of the vasculature in the tumors to the normal vasculature in the surrounding normal tissue (Markelc et al. 2013). This finding is specifically of interest, since it supports the observations of normal tissue sparing effect of electrochemotherapy.

Induction of Immune Response Several ablative techniques such as electrochemotherapy induce a local immune response. The induced immunologic cell death activates the immune response that can contribute to the overall effectiveness of electrochemotherapy. This statement is supported by several lines of evidence. First the response of tumors was less

Preclinical Studies on Electrochemotherapy

11

Fig. 3 The three mechanisms involved in antitumor effectiveness of electrochemotherapy. The direct effect on tumor cells is due to the increased drug uptake. The indirect effect on tumor cells is due to the vascular disrupting effect of electrochemotherapy. The involvement of the immune response is due to the shedding of the tumor antigens and elicitation of local immune response

pronounced in immunodeficient mice, and lower degree of complete responses was observed (Sersa et al. 1997). Immunological cell death was observed after electrochemotherapy with bleomycin (Calvet et al. 2014). The response of tumors was to some degree dependent on the immunogenicity of tumors (Heller et al. 2000). And, the response of tumors was boosted by adjuvant immunotherapy (Sedlar et al. 2012; Cemazar et al. 2015). Detailed discussion about this phenomenon and its implications on the development of electrochemotherapy are discussed in other chapters in this book. In summary, the induced immune response is important for the eradication of the last remaining tumor cells that are not directly or indirectly killed by electrochemotherapy. The extent of its contribution to the overall response, though, still needs to be determined. These findings can be exploited for the “in situ” vaccination by electrochemotherapy, both using bleomycin or possibly cisplatin and boosted in different ways. One of the approaches would be by immunostimulation, with cytokines, and the other one by inhibition of immune checkpoint inhibitors. These approaches are currently under extensive investigation and are being tested also in clinical trials (“▶ Adjuvant Immunotherapy as a Tool to Boost Effectiveness of Electrochemotherapy” by Kamensek, U., Kos, S., Sersa, G).

12

G. Sersa et al.

In summary, so far three mechanisms of electrochemotherapy have been determined: increased drug delivery to tumor cells, vascular disrupting effect, and the involvement of immune response (Fig. 3). The degree of the contribution of each of the mechanisms has not been determined but may vary according to the tumor type. For example, better vascularized tumors may have greater contribution of the vascular disrupting effect, while more immunogenic tumors induce a greater immune response.

Perspectives Based on the preclinical data and the knowledge of the basic mechanisms of action of electrochemotherapy, it has been swiftly translated into the clinics. Electrochemotherapy is being widely used as a treatment modality for local tumor control but could also be used as a concomitant treatment to standard and newly emerging therapies. The mechanistic studies that would shed some light on the subtle variations of the cell sensitivity are needed, as well as studies on immune aspects and the physiology of the tumors, all, that could affect the response rate of the tumors, and help to refine electrochemotherapy in clinical setting. One of the possibilities that still awaits clinical application is radiosensitizing effectiveness of electrochemotherapy. Electroporation is a site-specific delivery method and thus it could be used for effective tumor-specific delivery of radiosensitizing drugs. By more precise targeting of tumor tissue and less involvement of surrounding normal tissue, the therapeutic index of tumor irradiation may be increased. Previous studies of electrochemotherapy with bleomycin or cisplatin combined with radiotherapy demonstrated potentiation of the radiation response in sarcoma tumor: 1.9-fold for electrochemotherapy with bleomycin and 1.6-fold for electrochemotherapy with cisplatin. The radiosensitizing effect of electrochemotherapy with bleomycin was demonstrated in fractionated radiation regime in murine sarcoma model and with cisplatin in breast cancer tumor model. Preclinical studies on the fractionated radiation regime are good indicator for potential of the combined therapy also in the clinic (“▶ Combined Treatment of Electrochemotherapy with Irradiation” by Kranjc, S., Kamensek, U., Sersa, G.). Since electrochemotherapy reduces blood flow and lowers partial oxygen pressure (pO2) in the tumors, it can activate bioreductive drugs to exhibit a cytotoxic effect on hypoxic cells (Cemazar et al. 2001c). Tumor cells in suboptimal physiological conditions, like hypoxia, are more sensitive to heat, thus electrochemotherapy could also improve therapeutic conditions for the use of hyperthermia. A more pronounced effect of electrochemotherapy could be gained by combining it with gene electrotransfer of different plasmids coding for therapeutic genes, acting either locally or systemically. For local effect electrochemotherapy could be combined with antiangiogenic plasmids such as plasmid AMEP (antiangiogenic metargidin peptide) or plasmids coding for shRNA against MCAM (melanoma cell adhesion molecule) or endoglin. Furthermore, a systemic effect could be added to electrochemotherapy by combining it with IL-2, IL-12, p53, or granulocyte

Preclinical Studies on Electrochemotherapy

13

macrophage colony-stimulating factor (GM-CSF). Some of the combinations have already been tested and results support this concept (“▶ Adjuvant Immunotherapy as a Tool to Boost Effectiveness of Electrochemotherapy” by Kamensek, U., Kos, S., Sersa, G). The emerging therapies with immune checkpoint inhibitors are gaining in clinical importance, and electrochemotherapy could be an adjunct to them. In the clinical setting, antitumor effectiveness after concomitant treatment with electrochemotherapy and CTLA-4 (cytotoxic T-lymphocyte-associated protein 4) inhibitor has already been tested. The first clinical data are encouraging, but the preclinical data on interaction of these treatment combinations are lacking. The same holds true for targeted therapies. Electrochemotherapy could also be combined with other targeted therapies such as BRAF inhibitors targeting serine/threonine-protein kinase B-Raf (BRAF). First preclinical study has already been conducted and showed synergistic action. Case report of combining dabrafenib with electrochemotherapy also confirmed safety and feasibility of treatment. All the accumulated preclinical data on these new combinations could be further used in human clinics. The combined treatments could strengthen the effect of electrochemotherapy, with additive or synergistic action. Beside its great potential in human clinical use it can also be used in veterinary medicine. Studies on dogs, cats, hamsters, rabbits, and horses demonstrated good anticancer effectiveness for different tumors, such as fibrosarcoma, neurofibroma, mast cell tumor, perianal tumors, mammary adenocarcinoma, hemangioma, hemangiosarcoma, and sarcoids. Electrochemotherapy has also been demonstrated to be effective in combination with other treatment, especially surgery (“▶ Electrochemotherapy and Gene Electrotransfer in Veterinary Oncology”: Tamzali, Y., Cemazar, M., Tozon, N.).

Conclusions In summary, after the important contribution of preclinical research for the development of electrochemotherapy and for translation to the clinics, nowadays the research is focused on the emerging questions that arise from the clinical use. The main question is how to optimally combine electrochemotherapy with targeted therapies, immunostimulators and immune checkpoint inhibitors. In this way electrochemotherapy can be extended from local to systemic treatment or, from the other point of view, use electrochemotherapy as adjuvant or in situ vaccination to new immune therapies.

Cross-References ▶ Adjuvant Immunotherapy as a Tool to Boost Effectiveness Electrochemotherapy ▶ Combined Treatment of Electrochemotherapy with Immunomodulators

of

14

G. Sersa et al.

▶ Combined Treatment of Electrochemotherapy with Irradiation ▶ Effects of Electroporation on Blood Flow ▶ Electrochemotherapy and Gene Electrotransfer in Veterinary Oncology ▶ Electrochemotherapy and Its Clinical Applications ▶ Electroporation and Electropermeabilization ▶ Immune Response after Electroporation and Electrochemotherapy

References Calvet CY, Famin D, André FM et al (2014) Electrochemotherapy with bleomycin induces hallmarks of immunogenic cell death in murine colon cancer cells. Oncoimmunology 3: e28131. doi:10.4161/onci.28131 Cemazar M, Miklavcic D, Scancar J et al (1999) Increased platinum accumulation in SA-1 tumour cells after in vivo electrochemotherapy with cisplatin. Brit J Cancer 79:1386–1391. doi:10.1038/sj.bjc.6690222 Cemazar M, Parkins CS, Holder AL et al (2001a) Electroporation of human microvascular endothelial cells: evidence for anti-vascular mechanism of electrochemotherapy. Brit J Cancer 84:556–570. doi:10.1054/bjoc.2000.1625 Cemazar M, Miklavcic D, Mir LM et al (2001b) Electrochemotherapy of tumours resistant to cisplatin: a study in a murine tumour model. Eur J Cancer 37:1166–1172 Cemazar M, Parkins CS, Holder AL et al (2001c) Cytotoxicity of bioreductive drug tirapazamine is increased by application of electric pulses in SA-1 tumours in mice. Anticancer Res 21:1151–1156 Cemazar M, Todorovic V, Scancar J et al (2015) Adjuvant TNF-α therapy to electrochemotherapy with intravenous cisplatin in murine sarcoma exerts synergistic antitumor effectiveness. Radiol Oncol 49:32–40. doi:10.1515/raon-2015-0005 Dilruba S, Kalayda GV (2016) Platinum-based drugs: past, present and future. Cancer Chemother Pharmacol 77:1103–1124. doi:10.1007/s00280-016-2976-z Dolinesk T, Prosen L, Cemazar M et al (2016) Electrochemotherapy with bleomycin is effective in BRAF mutated melanoma cells and interacts with BRAF inhibitors. Radiol Oncol 50:274–279. doi:10.1515/raon-2016-0042 Frandsen SK, Gibot L, Madi M et al (2015) Calcium electroporation: evidence for differential effects in normal and malignant cell lines, evaluated in a 3D spheroid model. PLoS One 10: e0144028. doi:10.1371/journal.pone.0144028 Gehl J, Skovsgaard T, Mir LM (2002) Vascular reactions to in vivo electroporation: characterization and consequences for drug and gene delivery. Biochim Biophys Acta 1569:51–58 Gibot L, Wasungu L, Teissié J, Rols MP (2013) Antitumor drug delivery in multicellular spheroids by electropermeabilization. J Control Release 167:138–147. doi:10.1016/j.jconrel.2013.01.021 Gothelf A, Mir LM, Gehl J (2003) Electrochemotherapy: results of cancer treatment using enhanced delivery of bleomycin by electroporation. Cancer Treat Rev 29:371–387 Heller R, Jaroszeski M, Perrott R et al (1997) Effective treatment of B16 melanoma by direct delivery of bleomycin using electrochemotherapy. Melanoma Res 7:10–18 Heller L, Pottinger C, Jaroszeski MJ et al (2000) In vivo electroporation of plasmids encoding GM-CSF or interleukin-2 into existing B16 melanoma combined with electrochemotherapy inducing long-term antitumour immunity. Melanoma Res 10:577–583 Jarm T, Cemazar M, Miklavcic D et al (2010) Antivascular effects of electrochemotherapy: implications in treatment of bleeding metastases. Expert Rev Anticancer Ther 10:729–746. doi:10.1586/era.10.43 Jaroszeski MJ, Dang V, Pottinger C et al (2000) Toxicity of anticancer agents mediated by electroporation in vitro. Anticancer Drugs 11:201–208

Preclinical Studies on Electrochemotherapy

15

Kosjek T, Krajnc A, Gornik T et al (2016) Identification and quantification of bleomycin in serum and tumor tissue by liquid chromatography coupled to high resolution mass spectrometry. Talanta. doi:10.1016/j.talanta.2016.06.062 Markelc B, Sersa G, Cemazar M (2013) Differential mechanisms associated with vascular disrupting action of electrochemotherapy: intravital microscopy on the level of single normal and tumor blood vessels. PLoS One 8:e59557. doi:10.1371/journal.pone.0059557 Miklavcic D, Mali B, Kos B et al (2014) Electrochemotherapy: from the drawing board into medical practice. Biomed Eng Online 13:29. doi:10.1186/1475-925X-13-29 Mir LM, Orlowski S, Belehradek J Jr et al (1991) Electrochemotherapy potentiation of antitumour effect of bleomycin by local electric pulses. Eur J Cancer 27:68–72 Mir LM, Tounekti O, Orlowski S (1996) Bleomycin: revival of an old drug. Gen Pharmacol 27:745–748 Okino M, Mohri H (1987) Effects of a high-voltage electrical impulse and an anticancer drug on in vivo growing tumors. Jpn J Cancer Res 78:1319–1321 Orlowski S, Belehradek J Jr, Paoletti C et al (1988) Transient electropermeabilization of cells in culture. Increase of the cytotoxicity of anticancer drugs. Biochem Pharmacol 37:4727–4733 Sedlar A, Dolinsek T, Markelc B et al (2012) Potentiation of electrochemotherapy by intramuscular IL-12 gene electrotransfer in murine sarcoma and carcinoma with different immunogenicity. Radiol Oncol 4:302–311. doi:10.2478/v10019-012-0044-9 Sersa G, Cemazar M, Miklavcic D (1995) Antitumor effectiveness of electrochemotherapy with cisdiamminedichloroplatinum(II) in mice. Cancer Res 55:3450–3455 Sersa G, Miklavcic D, Cemazar M et al (1997) Electrochemotherapy with CDDP on LPB sarcoma: comparison of the anti-tumor effectiveness in immunocompetent and immunodeficient mice. Bioelectroch Bioener 43:279–283 Sersa G, Cemazar M, Miklavcic D, Chaplin DJ (1999) Tumor blood flow modifying effect of electrochemotherapy with bleomycin. Anticancer Res 19:4017–4022 Sersa G, Jarm T, Kotnik T et al (2008a) Vascular disrupting action of electroporation and electrochemotherapy with bleomycin in murine sarcoma. Brit J Cancer 98:388–398. doi:10.1038/sj.bjc.6604168 Sersa G, Miklavcic D, Cemazar M et al (2008b) Electrochemotherapy in treatment of tumours. EJSO 34:232–240 Sersa G, Kranjc S, Scancar J et al (2010) Electrochemotherapy of mouse sarcoma tumors using electric pulse trains with repetition frequencies of 1 Hz and 5 kHz. J Membr Biol 236:155–162. doi:10.1007/s00232-010-9268-z Yarmush ML, Golberg A, Sersa G et al (2014) Electroporation-based technologies for medicine: principles, applications, and challenges. Annu Rev Biomed Eng 16:295–320. doi:10.1146/ annurev-bioeng-071813-104622

Preclinical Studies on Irreversible Electroporation Suyashree Bhonsle, Robert E. Neal II, and Rafael V. Davalos

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Considerations for Reducing Thermal Damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Organ-Specific Ablations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pancreas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kidney . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Liver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Brain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prostate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lung . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preserved Patency of Sensitive Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Blood Vessels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nerves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ductal Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Advances in Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Immune Response and Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adjuvant Augmentation of IRE Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . High-Frequency Irreversible Electroporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 4 5 6 7 8 8 9 10 10 10 11 11 11 12 13 14

S. Bhonsle Virginia Tech Wake Forest School of Biomedical Engineering and Sciences, Virginia Tech, Blacksburg, VA, USA The Bradley Department of Electrical and Computer Engineering, Virginia Tech, Blacksburg, VA, USA R.E. Neal II AngioDynamics, Inc, Latham, NY, USA R.V. Davalos (*) Virginia Tech Wake Forest School of Biomedical Engineering and Sciences, Virginia Tech, Blacksburg, VA, USA Department of Mechanical Engineering, Virginia Tech, Blacksburg, VA, USA e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_46-1

1

2

S. Bhonsle et al.

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Related Chapters that May Be of Further Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Abstract

Irreversible electroporation (IRE) employs a series of brief electric pulses to destabilize cell membranes by altering the transmembrane voltage to create nanoscale defects that induce cell death. Its nonthermal mechanism makes it an ideal treatment modality for treatment of tumors near critical structures and vasculature, which are contraindications for other thermal ablation modalities. Since its conception, IRE has been investigated for applications in several organs to determine its safety and efficacy. IRE ablation zones, thermal effects, potential for real-time imaging, and cell death mechanism have been thoroughly investigated in vivo, leading to its successful transition to the clinical environment. This chapter describes IRE findings in the preclinical setting, with a focus on the pancreas, kidney, liver, bone, brain, prostate, and lung, and general implications with respect to successful therapy outcomes. These include consistency of ablation zones with different treatment parameters, histology of ablation zones showing tissue and tumor destruction, in vivo safety during and after treatment, and efficacy in terms of tumor regression. In addition, the patency of critical structures such as blood vessels, nerves, and ductal systems is briefly discussed. Newer preclinical findings such as immune response activation, adjuvant therapies, and modulating pulse regimes to improve treatment outcome and ease treatment delivery are outlined. Keywords

Irreversible electroporation • Pulsed electric fields • Preclinical studies • Nonthermal mechanism • Critical vasculature

Introduction Electroporation is a process in which short, high voltage electric pulses cause an increase in the biological cell membrane permeability, due to the formation of nanoscale defects, referred to as “pores.” For applications that involve cell drug delivery, the goal is to limit the pulse amplitude and number so that the pores are temporary, allowing transmission of small molecules across the membrane, and the cell remains viable (Overview and History of Electrochemotherapy). Alternatively, irreversible electroporation (IRE) utilizes a higher pulse amplitude and number to induce cell death as a tissue ablation modality. With appropriately designed pulse protocols, IRE creates lesions without causing significant thermal damage to the tissue (Davalos et al. 2005). The mechanism of IRE is generally characterized by irreversible structural defects, chemical imbalances due to the influx and efflux of

Preclinical Studies on Irreversible Electroporation

3

ions, and subsequent cell death (Electropore Energy and Thermodynamics) (Lee et al. 1993). The use of IRE as a therapy for ablating tumors was first demonstrated in an animal tumor model in 2007 (Al-Sakere et al. 2007). IREs clinical efficacy has since been demonstrated in numerous organs for soft tissue targets, particularly tumors. IREs ability to generate regions of dead cells while maintaining the critical extracellular architecture allows IRE to be used in regions containing sensitive structures that make such targets contraindicated for other focal therapies or surgical resection. In this chapter, the wide array of preclinical IRE studies in numerous animals (including animal tumor models) and organs are presented. These studies characterize IRE ablations in response to different treatment parameters and show its efficacy in relation to preservation of extracellular components.

Considerations for Reducing Thermal Damage IRE employs brief (~100 μs) but intense square wave electric pulses (50–400) delivered via electrodes inserted into or adjacent to the targeted tissue. These pulses induce an electric field distribution in the target and surrounding tissue, depending on the electrode configuration, pulse amplitude, and tissue type. For a given tissue and set of pulse parameters (pulse length, number of pulses, delivery rate), an electric field threshold can be determined that induces cell death and forms a lesion. Numerical models are often employed to predict the induced electric field distribution and ablation zone, using a priori knowledge of the tissue electrical properties. The electric pulses used in IRE treatment delivery induce collateral heating due to the resistive properties of biological tissue, known as Joule heating. In order to maintain IRE’s distinction from thermally based therapies, pulse protocols must be designed to limit Joule heating below thresholds that cause structural thermal damage to the tissue and surrounding sensitive structures. Numerous publications in the literature use numerical and experimental studies to address consideration of sufficient IRE without the clinically limiting thermal damage. These studies aim to delineate IRE cell death from the thermal damage that occurs when tissues are exposed to temperatures higher than their physiological norm for extended periods of time. While 50  C is typically chosen as the threshold to begin introducing thermal damage at the cellular scale, and it has been noted that prolonged exposure to mildly elevated temperatures as low as 43  C can lead to cell death, what is considered instantaneous thermal damage occurs as high as 83.6  C (prostate) or 74.7  C (liver) (Thomsen and Pearce 2010). From an experimental perspective, in vivo studies have measured temperature during pulse delivery to examine the potential for thermal damage. The increase in temperature during treatment is proportional to the applied electric field and is therefore dependent on the electrode configuration utilized in the study. Two configurations, namely, the monopolar and bipolar electrode configurations, are commonly employed to deliver IRE treatment. The monopolar electrode configuration contains the source and sink electrodes on two different needles and spacers need to

4

S. Bhonsle et al.

be used between the needles to maintain electrode distance. On the other hand, the bipolar electrode configuration has the source and sink on the same needle. This is accomplished using concentric cylinders inside the needle, with insulating material between the two cylinders. For monopolar and bipolar needle electrodes, the electric field is highest closest to the electrodes and decays away from the electrodes, and therefore, the temperature is always the highest near the electrodes. For flat plate electrodes, the electric field is uniform between the electrodes and so is the temperature distribution. In a clinical canine case using IRE to treat a soft tissue sarcoma, the temperature immediately adjacent to the electrode was measured using a fiber optic temperature probe. This study employed monopolar or bipolar electrodes of spacing 8–15 mm to deliver 80 pulses of 70–100 μs width and 800–1250 V at a rate of 90 pulses/min or immediately following the R-wave peak on the electrocardiogram. The maximum temperature rise at this location during the IRE procedure was 2.4  C (Neal et al. 2011) with the study showing successful tumor remission. Furthermore, an in vivo study in brain with thermal probes attached to the electrodes showed that it was possible to generate substantial volumes of IRE tissue while max tissue temperature rise was 1.15  C (Garcia et al. 2010). An additional in vivo study on porcine kidneys employed a more aggressive IRE protocol of three to four monopolar electrodes with electrode spacing and exposure of 15 mm to deliver three sets of 70 pulses, each 90 μs long, at a rate of 90 pulses/min. The peak temperature recorded at the center of the three- and four-electrode configuration was 57 and 79  C, respectively (Wagstaff et al. 2014). Consistent with the rapid decay in thermal effects with increasing distance from the electrodes, max temperatures of 40 and 42  C were measured 1 cm outside the electrode geometries, which was still within the IRE ablation zone. These studies demonstrate that while IRE ablation occurs primarily due to a nonthermal mechanism, careful attention must be paid to heating effects to ensure they remain below those to cause thermal damage to the critical structures and surrounding vasculature. In current IRE protocols, the temperature rise close to electrodes is inevitable and may lead to thermal damage near the electrodes (Treatment Planning for Electrochemotherapy and Irreversible Electroporation of DeepSeated Tumors); however, majority of the ablation zone away from electrodes is nonthermal, caused by electroporation-induced effects. Several studies have suggested techniques to reduce temperature rise during IRE treatment, including shorter pulse durations, reducing the pulse delivery to allow convective and conductive cooling between pulses, as well as proactive interventions to cool the electrode and/or tissue.

Organ-Specific Ablations IRE’s potential for ablation has been studied in several organs, each showing different implications in regard to successful therapy outcomes. This section addresses some of the more common IRE therapeutic applications and the preclinical data regarding the use of IRE for these targets.

Preclinical Studies on Irreversible Electroporation

5

Pancreas Locally advanced pancreatic cancers comprise approximately 40 % of the pancreatic tumors. Although not metastasized, these tumors have innervated and surrounded critical structures such as the pancreatic duct and the superior mesenteric artery. Therefore, the tumors around this region cannot be safely resected without damage to these critical structures, thereby, risking patient morbidity and mortality. Furthermore, the use of thermally ablative therapies in such environments has largely been avoided due to the possibility of thermal injury-induced pancreatitis. This leaves a large number of pancreatic cancers inoperable by the time of diagnosis. IRE’s ability to spare the major vasculature and potentially reduce risk of pancreatitis offers an approach to addressing the tumor in these regions. It could be used as either a standalone treatment modality, in conjunction with drugs, or to augment the treatment margin of surgical resection. Preclinical data in four female swine showed IRE can be safely administered in the pancreas (Charpentier et al. 2010). Treatment parameters involved the use of monopolar electrodes at a spacing of 9–15 mm; 90 pulses of 1350–2250 V were delivered to the healthy pancreas. All animals survived their designated times of 2 h, 2 days, and 2 weeks with no treatment-related complications. The tissues in and around the ablation zones were sliced into sections of thickness 5 mm, perpendicular to the electrode placement. The 5 mm section in the center of the ablation cavity had heights ranging from 10 mm to 21 mm and widths ranging from 10 mm to 16 mm. Pancreatic tissue at 2 weeks after IRE showed scarring in the ablation zone with preservation of pancreatic ducts. This study was also the first to show that triphenyltetrazolium chloride (TTC) can predict the IRE zone of ablation within 2 h of treatment. TTC stains the live mitochondrial activity in cells. While IRE ablation zones are detectable visually without staining, TTC staining enhances the contrast between the ablated and non-ablated areas. Another study in swine pancreas was performed with two electrode configurations – monopolar (19 gauge, i.e., diameter 0.9 mm) and bipolar (16 gauge, i.e., diameter 1.2 mm) – and the animals were euthanized after 3, 7, and 14 days (Bower et al. 2011). The monopolar electrode configuration consists of two needle electrodes, with source on one needle and sink on the other. Alternately, the bipolar probe contains both the source and sink on the same probe at a fixed spacing of 8 mm. Treatment parameters included electrode spacing from 8 to 20 mm, pulse numbers of 50–100, and applied voltage ranging from 2300 to 3000 V. All animals survived the study, with transient increases in amylase and lipase that normalized on third day. Amylase and lipase are key digestive enzymes, produced by the pancreas, that help break down starch and digest fats. High levels of these enzymes are indicative of pancreatitis. Gross analysis of samples revealed ablation zones of median size of height 3 cm and width 2.8 cm. Histological analysis revealed significant destruction of the pancreatic tissue with patent vascular structures and no significant difference between the two types of probes. Additionally, terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assays’ staining identified apoptotic markers in the IRE ablation zone.

6

S. Bhonsle et al.

Furthermore, IRE was shown to be effective in treating pancreatic ductal adenocarcinoma, an aggressive form of pancreatic cancer, in orthotopic mouse models (José et al. 2012). IRE treatment exhibited significant antitumor effects and extensive tumor necrosis, reduced tumor cell proliferation, and disruption of microvessels. Data in this study indicated that IRE leads to an increase in survival (from 42 days in untreated mice to 88 days in the IRE-treated group) with 25 % of mice showing complete tumor regression.

Kidney The increase of high-resolution diagnostic imaging is partially responsible for a rapid rise in the number of diagnoses of asymptomatic small renal masses, with over half of new cases detected incidentally. Frequent detection of low-risk renal tumors has encouraged care toward less aggressive, nephron-sparing approaches. Preservation of renal function in the affected kidney is vital for patients with solitary kidneys, comorbid conditions, multiple tumor sites, and genetic predisposition for recurrent bilateral renal tumors (von Hippel-Lindau disease). Partial nephrectomy still requires considerable invasiveness, while sensitivity of the ureter and high blood perfusion rate mitigate the efficacy of thermal therapies. These factors provide considerable value for IRE treatment of small renal masses. Preclinical investigations into the ablation of kidneys investigating IRE effects on macro- and microscopic healthy renal tissue show complete cell death of glomerular and tubule cortex structures while sparing major vasculature. In an IRE study on eight female Yorkshire pigs, laparoscopic ablations were performed with either monopolar or bipolar electrodes (Tracy et al. 2011). The pigs were euthanized between 10 min and 14 days after IRE, and the kidneys were harvested for gross and histological analysis for cellular viability. Monopolar ablations were performed at 2300 Vand 90 pulses at a pulse length of 100 μs and 1.5 cm spacing, while bipolar ablation was performed at 2700 V and 90 pulses at a pulse length of 70 μs. The spacing between electrodes was fixed at 0.8 cm. Overall lesion volume was smaller when the bipolar probe was used than when the monopolar was used. In the bipolar group, the mean gross lesion size, immediately and an hour after death, showed a volume of 1756 mm3 (assuming ablation shape of a prolate spheroid). In the monopolar group, the mean volume was 4440 mm3. By 7 days, the ablations in both groups remained similar to their immediately post-IRE size (10 min). However, by 14 days, there was substantial lesion contraction, regardless of the modality applied. There was heterogeneous and unpredictable involvement of the urothelium, ranging from no damage to complete ulceration and necrosis. However, it was more common with the bipolar probe, which was placed to traverse the collecting system as opposed to the monopolar probe technique where the ablation zone traversed the collecting system without direct puncture. The urothelium was partially spared in 60 % of the cases, with further evidence of repair/regrowth by 14 days despite direct involvement during ablation. Another swine study evaluated the effect of IRE on the renal urine-collecting system using intravenous urology and urinary cytology, which are routine urological

Preclinical Studies on Irreversible Electroporation

7

examinations, magnetic resonance imaging (MRI), and histology (Wendler et al. 2012). Each IRE procedure consisted of nine sets of ten IRE pulses, pulse length 70–100 μs, voltage 2300–2700 V. All animals were euthanized after a 28-day period. MRI scans did not show substantial lesion immediately after IRE; however, the sizes increased to an average diameter size of 16.5 mm at day 7 and shrunk to ~8 mm at day 28. These sizes showed good correlation with the ones from gross macroscopic analysis at day 28. Initial urothelial injury was noted, but the urinary system was completely regenerated 28 days post-IRE with the preservation of urothelial basement membranes and no urinoma evidence. Similarly, other IRE studies showed that renal ablations showed a demarcation between necrotic and normal tissue, 24 and 36 h after IRE treatment with evidence of tubule degeneration (Deodhar et al. 2011). Ablations showed cortical fibrosis, regenerating renal pelvic epithelium, and intact pelvic extracellular matrix. No thermal injury, renal pelvis, or blood vessel injury was seen. Two to four weeks later, there was evidence of regenerating tubules and pelvic epithelium with intact extracellular matrix and ablations turning to contracted scars. The urine-collecting system was essentially preserved with regenerated urothelial tissue. This study therefore demonstrated the nonthermal and connective tissue-sparing mechanism of action of IRE.

Liver The size of the liver relative to typical tumors, coupled with its regenerative capacity, makes the liver a common target for focal ablation of tumors. IRE has been clinically applied in this environment, particularly for targeted regions near the inferior vena cava, gallbladder, and external sensitive organs such as the bowel, pancreas, heart, and spleen. Further, the relative homogeneity and isotropy of hepatic organization make the liver a suitable model for investigation of various aspects for IRE therapy. The greatest collection of preclinical investigations regarding the effects from IRE therapy exists for liver ablations. Many of these studies are generalized to effects of IRE as a whole, such as monitoring IRE ablation zones using different imaging techniques, as well as studies into different pulse parameters and electrode designs. In a study conducted on 16 pigs, IRE ablations were imaged with ultrasound (US), magnetic resonance (MR), and computed tomography (CT) (Lee et al. 2010). Treatments were delivered with two electrode configurations – a single 16-gauge bipolar probe and two to three 18-gauge monopolar probes. Treatment parameters include inter-electrode spacing between 1.5 and 3.0 cm and 90 IRE pulses, each lasting 100 μsec at 2000–3000 V. At gross section examination, the mean diameter of the ablation zones was 33.53.0 mm which was achieved in 6.9 min (mean total procedure time per ablation), with a mean difference of 2.53.6 mm between US and gross section measurements. No complications were seen in any of the 16 animals. IRE ablation zones were well characterized with US, CT, and MR imaging, and real-time monitoring was feasible with US. Bile ducts and vessels were completely preserved. Areas of complete cell death were stained positive for

8

S. Bhonsle et al.

apoptotic markers, suggesting involvement of the apoptotic process in the pathophysiology of cell death caused by IRE. Additional studies on the liver analyzed the effects of different treatment parameters on IRE ablation sizes (Ben-David et al. 2012). An IRE study performed in 25 pig livers using two 18-gauge (diameter 1.2 mm) monopolar electrodes evaluated the effect of five treatment parameters including pulse number (20–90), pulse length (20–100 μs), voltage (2250–3000 V), inter-electrode spacing (1.5–2.5 cm), and length of active electrode exposure (1.0–3.0 cm). For all combinations of pulse parameters, the study reported average ablation widths between 2.6 and 3.9 cm and depths between 1.3 and 5.0 cm. In the range tested, the study found that increasing voltage amplitude had a statistically significant increase in ablation size. However, increasing pulse numbers or pulse duration did not have a significant increase. Therefore, IRE is not strictly energy dose dependent, i.e., increasing energy through pulse number or duration does not have the same effect as increasing energy through voltage amplitude. In addition, preclinical studies have shown that IRE can be used to effectively treat liver tumors in small animal models (Guo et al. 2010). In this study, hepatomas were established in an N1-S1 rodent model. Thirty rats were divided into treatment and control groups. For treatment groups, IRE electrodes were inserted and eight 100 μs and 2500 V pulses were applied to ablate the targeted tumor tissues. Magnetic resonance images showed statistically significant tumor size reductions (p  0.05) compared to the untreated tumors within 15 days post-therapy, with reduction in tumor diameters of 3231 %. Pathology correlation studies documented extensive tumor necrosis and full regression in nine of ten treated rats in 7–15 days after treatment.

Bone While IRE has primarily been evaluated for implications in soft tissue ablation, several studies have examined its effects on regions of the bone. In a study performing IRE in porcine lumbar vertebrae, it was shown that IRE produces localized regions of welldemarcated necrosis with no detectable change in bone texture and limited neural toxicity (Tam et al. 2014). This study was performed in the lumbar vertebrae of ten pigs using bipolar electrodes. Treatment parameters utilized 20–90, 2700 V IRE pulses of pulse width 70 μs. Well-delineated areas of necrosis of the bone, bone marrow, and skeletal muscle adjacent to the vertebral body were found with gross ablation average widths of 9 mm and lengths of 33 mm. This study showed IRE to be a safe and feasible technique to be utilized in the spine, with preservation of nerves.

Brain IRE utility in brain has been characterized with preclinical and clinical canine models. The high vascularization and extreme sensitivity of adjacent neural tissue require precise treatment plans and blunt-tip electrodes to deliver the IRE pulses. Preclinical studies in canine patients have shown IRE to safely ablate pathologically

Preclinical Studies on Irreversible Electroporation

9

heterogeneous brain tissue while preserving vascular integrity and patient neurological functions. In a preliminary study on five dogs, focal IRE ablations were made in the ectosylvian gyrus (Ellis et al. 2011). Three dogs were treated with recommended IRE parameters, one dog was a control, and the other was treated with a higher voltage to determine the upper safety limits of the procedure. For the three dogs that were noncontrols, treatment parameters included electrode separations of 5–8 mm, electrode exposures 5–7 mm, voltages from 500 to 1600 V, 50 μs pulse duration, and nine sets of ten pulses. These resulted in average ablation volumes between 0.258 and 1.655 cm3, with smaller volumes for lower voltage. Histological analysis showed submillimeter boundary between the necrotic and normal brain. The animals tolerated the procedure with no apparent complications except for the animal that was treated at the upper voltage limit. Canine malignant gliomas exhibit similar clinical, biologic, pathologic, molecular, and genetic properties as their human counterparts providing a good translational model for clinical investigations. In an early case study, IRE attained a 75 % reduction in tumor volume within the first week post-IRE (Garcia et al. 2011). Treatment parameters utilized 40–80, 500 V IRE pulses of pulse width 50 μs. The electrode spacing and exposure was 5 mm. Following adjuvant fractionated radiotherapy, the tumor was determined to be in complete remission prior to the suggestive onset of early-delayed radiation encephalopathy, though recurrent glioma could not be excluded. In a recent medium-term examination of the treatment cohort, it was found that for the patients that survived and were discharged after the procedure, Karnofsky performance scores were improved in 6/6 patients over pretreatment values, while seizure control improved in 5/6 (Rossmeisl et al. 2015).

Prostate Optimal treatment strategies remain to be determined for low- to medium-risk organconfined prostate cancer. While radical prostatectomy offers strong efficacy to tumor control of tumors that have not metastasized, it carries high rates of morbidity in regard to incontinence and impotence post-prostatectomy, including with robot-assisted procedures. This morbidity results from damage to the urethra and sensitive neurovascular bundles at the perimeter of the prostate. Thermal and other focal therapies also risk such adverse effects, and efforts to mitigate these risks may jeopardize the efficacy of the treatment. Where it remains to be well defined which prostate tumors require intervention, IRE may serve as an ideal focal or regional ablation approach to address identified tumor regions with significantly lower risk to potency and continence. Preclinical studies, primarily in vivo canine prostate studies, have shown the ability to safely create ablation lesions while preserving integrity and regular system function. In an early study, IRE was shown to produce significant ablations in prostate while preserving the urethra and neurovascular bundle (Onik et al. 2007). A further study investigated the implications of metallic seed implants within the prostate, identical to those used for brachytherapy (Scheffer et al. 2003). This study

10

S. Bhonsle et al.

found that the presence of the seeds did not significantly affect the ablation zone or thermal effects from the electric pulses, though it does not consider tissue changes in response to prior radiotherapy, such as fibrosis.

Lung Lung tumors have been treated with thermal ablation; however, their application is limited due to thermal sinks from blood vessels, air spaces, and potential thermal injury to critical tissues, including the heart, making it less safe and efficacious. IRE’s nonthermal mechanism of cell death and sparing of critical structures make it a potential candidate to treat lung tumors. A preliminary study on swine lung evaluated the safety of IRE in the lung (Dupuy et al. 2011). This study performed percutaneous IRE on nine domestic swine using bipolar as well as monopolar electrodes delivering 90 and 70 μs pulses of 1700–3000 V with electrocardiogram synchronization. The results showed that all swine successfully completed IRE treatment without any cardiac arrhythmias. CT showed focal areas of spiculated high density ranging from 1.1 to 2.2 cm. Histological analysis revealed focal areas of diffuse alveolar damage with fibrosis and inflammatory infiltration with sharp boundaries of the interlobular septa. The bronchioles and blood vessels within the areas of IRE were intact and did not show signs of tissue injury. However, optimization of IRE procedures for pulmonary targets remains to be performed before this organ can be successfully targeted for IRE treatments, due to the unique electrical and thermal characteristics caused by dynamic air spaces within the lung and its unique parenchymal structure.

Preserved Patency of Sensitive Structures There are numerous relevant critical structures that may be implicated within or around targeted volumes for IRE therapy. The relative importance of each will vary with the targeted organ and cancer variety.

Blood Vessels Extensive work has examined the effects of IRE on different organ blood vessels, some of which have been mentioned above. While IRE is shown to kill the endothelial cells and disrupt capillary-level vasculature, preservation of the collagen framework facilitates continued blood flow in large blood vessels. In a pilot study, the long-term effects of IRE on a large blood vessel were studied (Maor et al. 2007). Treatment parameters included a sequence of ten IRE pulses of 3800 V/cm voltage-to-distance ratio, 100 μs each, applied at a frequency of ten pulses per second directly to the carotid artery in six rats. All the animals survived the procedure and showed no side effects. Histology performed 28 days after the procedure showed that the connective matrix of the blood vessels remained intact,

Preclinical Studies on Irreversible Electroporation

11

and the number of vascular smooth muscle cells (VSMC) in the arterial wall decreased with no evidence of aneurysm, thrombus formation, or necrosis. In another IRE study (Appelbaum et al. 2012), Doppler ultrasound showed continued blood perfusion through a major vessel contained within a tissue region immediately after IRE treatment. Further, longitudinal studies have demonstrated the regeneration of endothelial cells within the affected regions of the blood vessels within 7 days, permitting full recovery and long-term function of the vessel.

Nerves In an early in vivo study, it was shown that canine prostate tissue could be ablated while preserving the neurovascular bundles necessary for potency and continence (Onik et al. 2007). Following this study, the possibility of preserved neural function has since been explicitly investigated. In addition to the peripheral nerve effects from the vertebral bone IRE studies, a long-term in vivo IRE study was performed on a rat sciatic nerve (Li et al. 2011). A sequence of ten pulses with voltage-to-distance ratio of 3800 V/cm, each 100 μs long, was only applied directly on the sciatic nerve to produce a treated length of about 10 mm. Electrophysiological, histological, and functional studies performed immediately and up to 10 weeks following surgery showed that, despite an initial decrease in functionality, the nerve attained full recovery approximately 2 months later.

Ductal Systems In addition to vascularity and neurological implications that may be present for many organs targeted for IRE ablation, there are also many organ-specific ductal systems which also seem to have preserved function and patency while being contained within or adjacent to ablation zones. These include bile duct preservation in pancreatic and liver ablations (Bower et al. 2011), collecting system and ureter in renal ablations (Deodhar et al. 2011), and urethra in prostate ablations (Onik et al. 2007). The action behind this preservation likely relates to the structural organization of many ductal systems, which are composed of relatively low-permeability connective tissue innervated with endothelial and epithelial cellularity. Although IRE will initially kill the cells within the ductal system architecture, preservation of the extracellular constituents permits continued function of these tissues and supports the recellularization of the systems over time.

Advances in Technology The main focus of the early preclinical studies described here is targeted toward describing IRE’s ability to destroy a targeted volume of tissue, while preserving critical structures in the vicinity, and describing the variation of observed effects in

12

S. Bhonsle et al.

numerous tissue varieties. Following the successful translation of IRE to the clinic, newer advances to the technology are being explored at the benchtop and preclinical settings. These studies explore secondary effects of IRE that may further expand the treatment zone, improve selectivity, as well as improve ease of application. Some preclinical investigations into a selected number of these effects are presented below.

Immune Response and Activation Where IRE is typically implemented by delivering electric pulses through electrodes inserted directly into the targeted tissue to induce the cell death, the affected tissue remains in situ following treatment. This permits the release of tumor-specific and tumor-associated antigens, as well as the various signals of cell distress into the treated and peripheral volumes. These molecules may have implications in the promotion of local and systemic immune responses, similar to those described from other ablation modalities. One study explored the immunologic response to tumor ablation with IRE using a subcutaneous xenotransplanted osteosarcoma rat model (Li et al. 2012). The animals were randomized into four groups: the control, sham operation, surgical resection, and IRE group. Another set of rats without tumor cell implantation served as the normal non-tumor-bearing group. In each of the groups, anticoagulated venous whole blood samples were acquired before and days after the operation to monitor for T lymphocyte activity, the major source of cellular antitumor immunity in cancer patients. Specifically, the CD3+ T lymphocytes that represent the major lymphocyte subset in peripheral blood, the CD4+ (T helper cells) and the ratio of CD4+/CD8+ (T suppressor/cytotoxic cells), are generally used as an indicator of antitumor immunity. Additionally, rats were killed independently in each treatment group, and splenocytes were assayed for IFN-γ and IL-4 production using intracellular cytokine staining. T-cells exert their effector functions partly by producing and releasing cytokines and are characterized by their distinct cytokine expression patterns. Th1 cells secrete IFN-γ and IL-2, whereas Th2 cells produce IL-4, IL-5, and IL-10. The results of the study showed that the application of 1500 V/cm voltage-to-distance ratio in nine trains of ten direct current IRE pulses, each 100 μs long, could produce complete osteosarcoma cell ablation, while also providing a substantial immune response. Seven days after treatment, the IRE group had a significant increase in CD3+ and CD4+ cells and increased ratio of CD4+/CD8+. Additionally, IRE group showed an increased percentage of IFN-γ splenocytes. A comparative in vivo murine study was conducted on immunocompetent (IC) versus immunodeficient (ID) mice implanted with RENCA murine kidney cancer cells to gauge if IRE invoked a systemic immune response in the tumor environment (Neal et al. 2013). Two flat plate electrodes were utilized to apply IRE treatment on the subcutaneous murine tumors with a highly conductive gel facilitating improved current delivery into the tumor. Treatment parameters included voltage-to-distance ratio of 1500 V/cm, each 100 μs long, delivered at a rate of 1 pulse per second. A total of 100 pulses were delivered, reversing polarity after the

Preclinical Studies on Irreversible Electroporation

13

first 50. Following pulse delivery, the electrodes were reoriented 90 , and the pulsing process was repeated, delivering a total of 200 pulses to the tumor. This protocol was selected due to its ability to produce an observable treatment response relative, but not strong enough to cause complete regressions in both strains of mice, which would make it difficult to discriminate any differences in treatment outcome. The treated IC group responded significantly better than the treated ID group, despite no inherent difference in initial tumor susceptibility. Similar susceptibility was indicated by tumors reaching a treatable size for both groups within the same number of days and similar tumor growth response to sham treatments. This suggests that although an immune response is not required for complete tumor regression, therapeutic response in immunocompetent patients may be better than predictions based on experimental studies using ID cancer models. Overall, the exploitation and encouragement of the immune response demonstrated from IRE treatment of tumors remains a promising yet relatively unexplored field for improving IRE therapy outcomes.

Adjuvant Augmentation of IRE Treatments In addition to encouraging a more robust immune response to further encourage effective oncologic outcomes from IRE therapy, numerous other promising adjuvant approaches exist to increase ablation zone and oncologic outcome from a given pulse protocol for IRE therapy. The most obvious of these is the inclusion of targeted chemotherapeutics already shown for electrochemotherapy to selectively kill reversibly electroporated cells in the zone surrounding the irreversibly electroporated cells. In addition, the use of conventional therapies as adjuvants to IRE may also improve patient outcomes, such as the inclusion of standard chemotherapy regimens to kill any distant micro-metastases, while IRE kills the primary tumor, or the use of transarterial chemoembolization to further stress and promote death of tumor cells. Several studies have examined potential IRE and reversible electroporation therapy augmentation by locally manipulating the properties of the cell membranes to increase their susceptibility to electroporation and IRE. One approach that has been suggested, with implications primarily likely for in vitro applications, is the addition of dimethyl sulfoxide (DMSO) in the cell membrane to alter the structural properties of the lipid bilayer (Jiang et al. 2014). This study along with the first IRE tumor study (Al-Sakere et al. 2007) also explored the potential of modifying the temporal mode of delivery of pulses on IRE outcome, where it was shown that adding delays between trains of pulses can decrease the IRE electric field ablation threshold. An additional study examined the sensitization of cells to electroporation within a given electric field by altering the electrochemical environment of cells in tissue in vitro with cationic and anionic substances. This study found markedly promising results with the inclusion of cationic anesthetics in the cellular environment, with particular benefits exhibited from procaine and lidocaine, where a 50 % reduction in the strength of the electric field is required to induce cell death with IRE. While it

14

S. Bhonsle et al.

remains to be determined whether such a pronounced effect will translate to in vivo environments, it offers a valuable potential adjuvant to safely augment the treatment zone of IRE and other electroporation-based therapies (Combining Electrolysis and Electroporation for Tissue Ablation).

High-Frequency Irreversible Electroporation Apart from improving ablation sizes, preclinical studies have shown that the ease of application of IRE treatment can be improved by manipulating pulse parameters, such as pulse duration and polarity. Clinically, IRE requires the administration of paralytic agents to prevent muscle contractions during treatment that are associated with the delivery of unipolar pulses of longer durations (~>5 μs). By reducing the pulse duration and applying bipolar IRE-inducing pulses, the threshold for nerve stimulation is increased and the muscle contractions are eliminated, removing the need for paralytic agents. The first preclinical study on high-frequency irreversible electroporation (H-FIRE) was conducted on rat brain (Arena et al. 2011), and muscle contractions were quantified via accelerometers placed at the cervicothoracic junction. MRI and histological evaluation was performed postoperatively to assess ablation characteristics. Treatment parameters consisted of bipolar pulses of 1–2 μs duration and 100–400 V amplitude delivered in 200 μs long bursts at a frequency of 1 burst per second. Blunt electrodes of 1 mm spacing and exposure were used for the treatment. Results showed that, even in the highest energy H-FIRE protocol, no detectable peaks in acceleration above the inherent noise of the system were observed. Histopathologic examination of brain sections from all treatments demonstrated clear areas of ablation and sharp delineation from adjacent normal brain. This study showed H-FIRE to be a feasible technique for nonthermal tissue ablation that eliminates muscle contractions.

Conclusion IRE’s unique nonthermal modality of ablation without denaturing extracellular proteins permits IRE treatment in sites containing critical structures and vasculature that contraindicate them for thermal ablation or surgical resection. The advantages of IRE have resulted in the treatment being performed in an array of applications and anatomical target regions. This chapter discusses the preclinical findings of IRE in different sites and its treatment outcome in terms of ablation size and preservation of critical structures. Additionally, secondary effects that may foster exploitation for enhancement of IRE treatment, such as anti-immune response, adjuvant therapies, and advances to IRE pulse protocols, offer evidence suggesting improved clinical outcomes for the technology and its clinical implications as the base of knowledge continues to grow.

Preclinical Studies on Irreversible Electroporation

15

Related Chapters that May Be of Further Interest ▶ Immune Response After Electroporation and Electrochemotherapy ▶ Irreversible Electroporation and Its Clinical Applications ▶ Tissue Ablation by Irreversible Electroporation

References Al-Sakere B, Andre F, Bernat C, Connault E, Opolon P, Davalos RV, Rubinsky B, Mir LM (2007) Tumor ablation with irreversible electroporation. PLoS One 2(11):e1135. doi:10.1371/journal. pone.0001135 Appelbaum L, Ben-David E, Sosna J, Nissenbaum Y, Goldberg SN (2012) US findings after irreversible electroporation ablation: radiologic-pathologic correlation. Radiology 262 (1):117–125 Arena CB, Sano MB, Rossmeisl JH, Caldwell JL, Garcia PA, Rylander MN, Davalos RV (2011) High-frequency irreversible electroporation (H-FIRE) for non-thermal ablation without muscle contraction. Biomed Eng Online 10(1):1 Ben-David E, Appelbaum L, Sosna J, Nissenbaum I, Goldberg SN (2012) Characterization of irreversible electroporation ablation in in vivo porcine liver. AJR Am J Roentgenol 198(1): W62–W68 Bower M, Sherwood L, Li Y, Martin R (2011) Irreversible electroporation of the pancreas: definitive local therapy without systemic effects. J Surg Oncol 104(1):22–28. doi:10.1002/ jso.21899 Charpentier KP, Wolf F, Noble L, Winn B, Resnick M, Dupuy DE (2010) Irreversible electroporation of the pancreas in swine: a pilot study. HPB 12(5):348–351 Davalos RV, Mir IL, Rubinsky B (2005) Tissue ablation with irreversible electroporation. Ann Biomed Eng 33(2):223–231 Deodhar A, Monette S, Single GW Jr, Hamilton WC Jr, Thornton R, Maybody M, Coleman JA, Solomon SB (2011) Renal tissue ablation with irreversible electroporation: preliminary results in a porcine model. Urology 77(3):754–760 Dupuy DE, Aswad B, Ng T (2011) Irreversible electroporation in a Swine lung model. Cardiovasc Intervent Radiol 34(2):391–395 Ellis TL, Garcia PA, Rossmeisl JH Jr, Henao-Guerrero N, Robertson J, Davalos RV (2011) Nonthermal irreversible electroporation for intracranial surgical applications: laboratory investigation. J Neurosurg 114(3):681–688 Garcia PA, Rossmeisl JH Jr, Neal RE II, Ellis TL, Olson JD, Henao-Guerrero N, Robertson J, Davalos RV (2010) Intracranial nonthermal irreversible electroporation: in vivo analysis. J Membr Biol 236(1):127–136 Garcia P, Pancotto T, Rossmeisl J, Henao-Guerrero N, Gustafson N, Daniel G, Robertson J, Ellis T, Davalos R (2011) Non-thermal irreversible electroporation and adjuvant fractionated radiotherapeutic multimodal therapy for intracranial malignant glioma in a canine patient. Technol Cancer Res Treat 10(1):73–83 Guo Y, Zhang Y, Klein R, Nijm GM, Sahakian AV, Omary RA, Yang GY, Larson AC (2010) Irreversible electroporation therapy in the liver: longitudinal efficacy studies in a rat model of hepatocellular carcinoma. Cancer Res 70(4):1555–1563. doi:10.1158/0008-5472.CAN-093067 Jiang C, Qin Z, Bischof J (2014) Membrane-targeting approaches for enhanced cancer cell destruction with irreversible electroporation. Ann Biomed Eng 42(1):193–204 José A, Sobrevals L, Ivorra A, Fillat C (2012) Irreversible electroporation shows efficacy against pancreatic carcinoma without systemic toxicity in mouse models. Cancer Lett 317(1):16–23

16

S. Bhonsle et al.

Lee RC, Canaday DJ, Hammer SM (1993) Transient and stable ionic permeabilization of isolated skeletal muscle cells after electrical shock. J Burn Care Res 14(5):528–540 Lee EW, Chen C, Prieto VE, Dry SM, Loh CT, Kee ST (2010) Advanced hepatic ablation technique for creating complete cell death: irreversible electroporation. Radiology 255(2):426–433 Li W, Fan Q, Ji Z, Qiu X, Li Z (2011) The effects of irreversible electroporation (IRE) on nerves. PLoS One 6(4):e18831. doi:10.1371/journal.pone.0018831 Li X, Xu K, Li W, Qiu X, Ma B, Fan Q, Li Z (2012) Immunologic response to tumor ablation with irreversible electroporation. PLoS One 7(11):e48749 Maor E, Ivorra A, Leor J, Rubinsky B (2007) The effect of irreversible electroporation on blood vessels. Technol Cancer Res Treat 6(4):307–312 Neal RE II, Rossmeisl JH, Garcia PA, Lanz OI, Henao-Guerrero N, Davalos RV (2011) Successful treatment of a large soft tissue sarcoma with irreversible electroporation. J Clin Oncol 29(13): e372–e377 Neal RE II, Rossmeisl JH Jr, Robertson JL, Arena CB, Davis EM, Singh RN, Stallings J, Davalos RV (2013) Improved local and systemic anti-tumor efficacy for irreversible electroporation in immunocompetent versus immunodeficient mice. PLoS One 8(5):e64559 Onik G, Mikus P, Rubinsky B (2007) Irreversible electroporation: implications for prostate ablation. Technol Cancer Res Treat 6(4):295–300 Rossmeisl JH Jr, Garcia PA, Pancotto TE, Robertson JL, Henao-Guerrero N, Neal RE, Ellis TL, Davalos RV (2015) Safety and feasibility of the NanoKnife system for irreversible electroporation ablative treatment of canine spontaneous intracranial gliomas. J Neurosurg 123 (4):1008–1025 Scheffer SR, Nave H, Korangy F, Schlote K, Pabst R, Jaffee EM, Manns MP, Greten TF (2003) Apoptotic, but not necrotic, tumor cell vaccines induce a potent immune response in vivo. Int J Cancer 103(2):205–211 Tam AL, Abdelsalam ME, Gagea M, Ensor JE, Moussa M, Ahmed M, Goldberg SN, Dixon K, McWatters A, Miller JJ (2014) Irreversible electroporation of the lumbar vertebrae in a porcine model: is there clinical-pathologic evidence of neural toxicity? Radiology 272(3):709–719 Thomsen S, Pearce JA (2010) Thermal damage and rate processes in biologic tissues. In: Opticalthermal response of laser-irradiated tissue. Springer, Dordrecht, pp 487–549 Tracy CR, Kabbani W, Cadeddu JA (2011) Irreversible electroporation (IRE): a novel method for renal tissue ablation. BJU Int 107(12):1982–1987 Wagstaff PGK, de Bruin DM, van den Bos W, Ingels A, van Gemert MJC, Zondervan PJ, Verdaasdonk RM, van Lienden KP, van Leeuwen TG, de la Rosette JJ (2014) Irreversible electroporation of the porcine kidney: temperature development and distribution. In: Urologic oncology: seminars and original investigations. Elsevier 33(4) Wendler JJ, Pech M, Porsch M, Janitzky A, Fischbach F, Buhtz P, Vogler K, Hühne S, Borucki K, Strang C (2012) Urinary tract effects after multifocal nonthermal irreversible electroporation of the kidney: acute and chronic monitoring by magnetic resonance imaging, intravenous urography and urinary cytology. Cardiovasc Intervent Radiol 35(4):921–926

Preclinical Studies on Nanosecond Pulses Stephen J. Beebe

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Paradigm Shifts in Cancer Treatment from Chemistry to Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nomenclature for nsPEF-Treated Tumor Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . nsPEF Conditions for Effectively Treating Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Considering nsPEF Conditions When Establishing a Tumor Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cautions Needed Using Nanosecond High EFs on Animal Skin and Tissues. . . . . . . . . . . . . . . . . . Effect of Nanosecond Pulsed EFs on the Skin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effect of Nanosecond Pulsed EFs on Blood Flow to the Skin and Liver . . . . . . . . . . . . . . . . . . . . . . . Mechanisms for nsPEF-Induced Tumor Elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Apoptosis and Regulated Cell Death (CD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Do nsPEFs Induce DNA Damage In Vivo? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Do nsPEFs Induce Immune Responses? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 4 4 9 10 10 11 12 12 15 15 17 18 18

Abstract

Nanosecond pulses have been used for over a decade to treat tumors in preclinical studies in mice and rats. The technology based on pulsed power refocuses cancer treatment from chemistry and drugs to physics and high-intensity, nonthermal electric fields (EFs) to tumors by administering pulses with durations in the nanosecond domain. This constitutes a high-power, low-energy impact that is unique and distinct from other physical medical technologies, such as radiofrequency ablation and ionizing radiation. Nanosecond pulsed electric fields (nsPEFs) have successfully eliminated a variety of cancers in subcutaneous ectopic models or allographs, orthotopic models in liver, induced S.J. Beebe (*) Frank Reidy Research Center for Bioelectrics, Old Dominion University, Norfolk, VA, USA e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_47-1

1

2

S.J. Beebe

skin cancer models and human cancer xenographs in immunodeficient mice. Recent data suggest, but do not yet prove, that nsPEFs induce immune responses in addition to ablating primary tumors. This chapter will review what has been done with nsPEF technology in preclinical cancer models, how these studies have been conducted, what should be considered when establishing new cancer models, what precautions need to be taken with high EFs, what effects nsPEFs have on the skin and blood flow, and what is known about how nsPEFs cause tumor cells to die. Finally, what is known about the possible immune response induced by nsPEFs will be discussed. Overall, this chapter will present what is presently known about nanosecond pulses for the treatment of cancer in animal models. Keywords

Pulsed power • Electropermeabilization • Ectopic/orthotopic • Allograph/ xenograph • Electric field breakdown • Regulated cell death (RCD) • Caspases • TUNEL

Introduction Pulsed power is the fundamental technology behind high-intensity electric field (EF) pulses with sub-microsecond durations. Pulsed power has been used for decades in high-powered physics and for military applications such as radar. Now with the advent of nanosecond pulsed electric fields (nsPEFs) for modulating cell functions and for eliminating cancer, pulsed power physical principles have been intertwined with biology and medicine. Pulsed power technology accumulates and compresses energy and immediately releases high power in nanosecond durations. An example of pulsed power principles inherent in nsPEFs is to compare storage of 1 J of energy released in 1 s (1 W) versus when released in 1 μs (1 MW, 106 W) or 1 ns (1 GW, 109 W). By using high-voltage capacitors with fast-switching and discharge capabilities, nanosecond pulses are delivered with high-intensity EFs into cells and tissues. At lower EFs, this mimics hormones and neurotransmitters, and at higher EFs, this mimics chemotherapeutic drugs and internal stress without molecules – the low-energy, high-power electric input affects plasma membranes (PMs) as well as intracellular membranes on organelles and maybe proteins. Biological cells have not seen any impact like this in the history of life on earth. Therefore, how cells and tumor tissues respond to them as either nonlethal or lethal stimuli is of considerable interest. This chapter will discuss how tumors respond to nsPEFs under conditions that are intended to eliminate them. Interestingly, because energy is relatively low, yet power is high, normal cells and cancer cells do not immediately die by abrupt membrane rupture or primary necrosis. They die by regulated cell death (RCD). More about that later. Pulsed power can produce pulses with durations in a billionth of a second (10
9 s, nanosecond) and even trillionth of a second (10
12 s, picosecond). Considered

Preclinical Studies on Nanosecond Pulses

3

another way, in a nanosecond, light travels (1.86  105 mi/s or 2.99  105 km/s) about 12 in or 30 cm; in a picosecond, light travels about 0.01 in or 0.3 mm. In contrast, conventional electroporation or electropermeabilization (EP) used in gene electro-transfer (GET), electrochemotherapy (ECT), and irreversible electroporation (IRE) generates microsecond (1  10
6 s) and millisecond (1  10
3 s) pulses. In a microsecond and a millisecond, light travels about 1,000 ft or 300 m and about 200 mi or nearly 300 km, respectively. Millisecond and microsecond (>100 μs) pulses are generally considered to be conventional EP conditions (see ▶ Mir – Electroporation and Electropermeabilization). These pulses produce defects in PMs, called aqueous pores (Weaver and Chizmadzhev 1996) (also see ▶ Weaver, Pore Lifetime and Permeabilization Lifetime in Models). While PMs spontaneously form very small, unstable, rapidly resealing pores, EFs lower the energy required for pore formation, making greater numbers of more stable pores. For both GET and ECT, EFs are generally low (0.1–1.0 kV/cm), so these aqueous pores are transient, allowing resealing and maintenance of cell viability. GET allows DNA to enter cells for transcription to accomplish an intended function (see ▶ Golzio and Rols – Nucleic Acid Electrotransfer in Mammalian Cells: Mechanistic Description). In ECT, transient pores allow poorly permeable chemotherapeutic agents (bleomycin) to enter cells causing cell death (CD) (see ▶ Serša et al. Preclinical Studies on Electrochemotherapy). In contrast, IRE uses much higher EFs (1 kV/cm) such that permanent pores are formed and cells lose PM integrity and die by primarily by primary necrosis (see ▶ Rolong et al. – Tissue Ablation by Irreversible Electroporation).

Paradigm Shifts in Cancer Treatment from Chemistry to Physics Considerations for nsPEFs to eliminate cancer evolved from experiments showing that they could lyse bacteria and stun brine shrimp (Schoenbach et al. 1997). Soon, attention shifted from environmental decontamination to applications in biology and medicine (Schoenbach et al. 2001; Beebe et al. 2002, 2003; Vernier et al. 2003). Applications for pulsed power in cancer presented a new paradigm in cancer treatment and has demonstrated significant efficacy in many preclinical studies as will be discussed in detail in this chapter. In addition to eliminating tumors by RCD mechanisms, including apoptosis (Schoenbach et al. 2001; Chen et al. 2010, 2012, 2014a; Nuccitelli et al. 2015), nsPEFs have been shown to induce a vaccine-like effect in mice (Beebe et al. 2011; Chen et al. 2013) and rats (Chen et al. 2014a). Other evidence also suggests nsPEF-induced immune responses (Nuccitelli et al. 2012a, 2015) (see section “Do nsPEFs Induce Immune Responses?”). During tumor RCD, dying cells activate immune mechanisms that can prevent cancer recurrence (Kroemer et al. 2013). Consequently, nsPEF treatment may become a unique approach to immunotherapy that does not involve delivery of DNA or virus, checkpoint inhibitors, engineered patient immune cells, or molecules of any kind. Furthermore, this approach provides a shift from cancer treatments with chemistry to cancer treatments with physics.

4

S.J. Beebe

Nomenclature for nsPEF-Treated Tumor Models In this section, nomenclature for various tumor models will be presented without specific references to them, until discussed later (Table 1, section “NsPEF Conditions for Effectively Treating Tumors”). Most preclinical studies with nsPEFs have been carried out in mice with subcutaneously injected tumor cells. These formed tumors are ectopic or in an abnormal position. Some studies have used “allographs” with cells or tissues transplanted in a recipient from a genetically nonidentical donor of the same species; they are non-syngeneic. Syngeneic models have also been used when donor tumor cells and recipient are genetically identical. So, tumors can be ectopic and non-syngeneic or syngeneic. Tumors may be more readily eliminated in non-syngeneic models because animals may have some immune resistance to the nongenetically identical tumor cells. In orthotopic models, tumors are implanted in their natural tissues. All orthotopic models have been syngeneic in nsPEF studies. Other nsPEF studies have been done with subcutaneous xenographs, when human tumors or cells are implanted into immunodeficient mice. Other tumors have been induced in mice that are susceptible to UV light or ionizing radiation. Finally, nsPEFs have been used to treat tumors in humans. Thus, nsPEFs have demonstrated efficacy as therapy for local cancers in melanoma, basal cell carcinoma (BCC), hepatocellular carcinoma (HCC), pancreatic and breast cancer in mice and rats, and BCC in humans.

nsPEF Conditions for Effectively Treating Tumors In all, studies using nsPEFs to treat tumors have used pulse durations of 7, 14, 20, 30, 100, or 300 ns with EF amplitudes between 20 and 68 kV/cm and pulse repetition rates (pulses/second) of 0.5–7 Hz. All pulse durations demonstrated EF-dependent and/or pulse number-dependent decreases in tumor size without significant increases in temperature. However, repetition rates of 5–7 Hz caused increases in temperature but apparently did not produce thermal effects (30 kb. However, the primary limitation of adenoviral vectors is the strong immune response in the presence of preexisting immunity. Lentiviral vectors based on human immunodeficiency virus or nonhuman retroviruses produce long-lasting expression after integration into the host genome. However, genomic integration is associated with insertional mutagenesis, which can be tumorigenic. Wild-type adeno-associated viral vectors also integrate into the genome in a site-specific manner. Due to the observation of tumorigenesis in preclinical models, vectors for adeno-associated viral gene therapy are designed to remain within cells episomally. The gene therapies achieving regulatory approval to date are viral gene therapies. In 2004, the State Food and Drug Administration of China approved the first gene therapy, Gendicine, a recombinant p53 therapy delivered by adenovirus for head and neck squamous cell carcinoma. The European Medicines Agency (EMA) has now approved two gene therapies, Glybera, an adeno-associated vector engineered to produce lipoprotein lipase approved in 2012 for a rare inherited deficiency, and Strimvelis, an ex vivo gene therapy delivered by gamma retrovirus approved in 2016 for the treatment of severe combined immunodeficiency due to adenosine deaminase deficiency. In 2015, the United States Food and Drug Administration (FDA) approved talimogene laherparepvec, an oncolytic herpesvirus expressing granulocyte macrophage colony stimulating factor, for the treatment of unresectable melanoma recurrence. Table 1 Common viral gene therapy vectors Capsid size (nm) Packaging capacity (kb) Genomic integration Immune response Transgene expression

Adenovirus 70–120 7.5–10 No Yes (preexisting) Transient

nm nanometers, kb kilobases

Lentivirus 80–100 6–10 Yes Yes (transfected cells) Long-lasting

Adeno-associated virus 20–25 4.5–5 Low Yes (preexisting) Long-lasting

Principles of Electroporation for Gene Therapy

5

Gene transfer using plasmid DNA is an attractive option because it removes the need for a biological vector, which improves the safety profile compared to viral delivery. Plasmid DNA is more easily prepared and more economical to produce than viral vectors. Simple intramuscular plasmid injection produces gene expression (Wolff et al. 1990). However, this expression is limited due in part to low transfection efficiency. Extracellularly, the presence of nucleases and the difficulty of transit across the interstitial space between cells are barriers to delivery. Depending on the delivery method, cell-related barriers may include the plasma membrane, cytoplasmic nucleases, transport through the cytoplasm, and transport across the nuclear membrane. Several nonbiological chemical and physical delivery methods have been developed to enhance the delivery of naked nucleic acids. Chemical delivery involves packaging of the nucleic acid into particles such as nanoparticles formed primarily from cationic lipids, polymers, or peptides. Physical methods include ultrasoundinduced sonoporation, hydrodynamic delivery, laser-induced stress waves, magnetofection, jet injection with highly pressurized gas, jet injection with heavy metal particles, and electroporation. Synthetic and biologically based transposase/ integrase systems including phage phiC31, PiggyBac, and Sleeping Beauty offer nonviral integration into the genome for long-term expression. Although integration is reasonably targeted, these delivery systems may have the potential to produce insertional mutagenesis. Several chemical and physical gene delivery methods have reached clinical trials. The first nonviral gene therapy trial was initiated in 1992 (Nabel et al. 1993). A DNA-liposome complex encoding a foreign major histocompatibility complex protein was injected into the patient’s melanoma nodules. Since then, nonviral DNA and RNA delivery directly to patients by methods including simple injection, lipofection, particle bombardment, and electroporation have averaged approximately 20 per year over the last 15 years (Fig. 2). Clinical trials utilizing in vitro applications, in which cells such as T lymphocytes or dendritic cells are modified nonvirally then injected into the patient, have averaged 9 per year in that time. The concept of gene therapy is not without controversy. Recently, the discovery of the potential for biased inheritance of genes (“gene drive”) using the Clustered Regularly Interspaced Short Palindromic Repeats/CRISPR associated (CRISPR/ Cas) technique has raised ethical concerns. These considerations, although they constitute very important questions, are beyond the scope of this chapter.

Electroporation/Electropermeabilization Controlled electric pulses produce temporary permeabilized areas in a cell’s membrane that can be used to accomplish several functions (Fig. 3), which is referred to as electroporation or electropermeablization. Irreversible electroporation, the most extreme application of electroporation, is defined as cell destruction by electric pulses to ablate undesirable tissues (Jiang et al. 2015). This is accomplished with

6

L.C. Heller

Fig. 2 Number of nonviral gene therapy clinical trials worldwide by year. In vivo, therapeutic DNA or RNA is introduced directly into the patient. In vitro, autologous or allogeneic cells are modified in vitro and introduced into the patient (Source: Gene Therapy Clinical Trials Worldwide provided by the Journal of Gene Medicine)

Fig. 3 Cell modification after electroporation. Depending on the pulse regimen, electroporation can have several effects on a cell. Temporary pores are induced by reversible electroporation. The cell survives but is intentionally modified in some way, i.e., by molecular delivery, fusion between cells, or insertion of proteins or other molecules into the plasma membrane. Application of similar but more numerous pulses produces irreversible pores and cell death

Principles of Electroporation for Gene Therapy

7

extended delivery of short, high-voltage pulses. Ablation with this technique for multiple tumor types has reached clinical trials. However, pulse application does not necessarily destroy the cell. In reversible electroporation, the cell survives and several therapeutic applications are possible. Membrane fusion can be induced by electric fields (Zimmermann 1982) and has been used for the formation of hybridomas (Lo et al. 1984) and for in vivo cell-tissue electrofusion (Grasso et al. 1989). Famously, Dolly the sheep, the first mammal cloned from an adult somatic cell, was formed after electrically induced cell fusion between a nuclear donor cell and an enucleated egg (Wilmut et al. 1997). Electric pulses can also be used to insert proteins into a cell’s plasma membrane (Mouneimne et al. 1989) although protein insertion may be short-term, limiting its therapeutic value. The third reversible application, molecular delivery or gene electrotransfer, will be described here. In 1982, plasmid delivery using electric pulses was first demonstrated (see “First Demonstration DNA Electrotransfer”) to cells in culture by the formation of stable transformants (Neumann et al. 1982). In the first application of pulses to living animal (Okino and Mohri 1987), tumor reduction was observed after electroporation of bleomycin into hepatocellular carcinomas in a rat model. After additional preclinical development, this therapy, electrochemotherapy, was tested in clinical trials in the first application of pulses to human patients and is now included in several guidelines for the treatment of cutaneous tumors in the European Union (Mir et al. 2006). The demonstration of plasmid delivery and the clinical success of drug delivery enabled the development of gene therapy by electrotransfer. The first confirmation of gene delivery in vivo was the demonstration of gene transfer to mouse skin (Titomirov et al. 1991). In this case, antibiotic resistance after in vivo delivery was demonstrated in vitro. In vivo expression of the reporter genes β-galactosidase in rat brain tumors (Nishi et al. 1996) and luciferase in healthy rat liver (Heller et al. 1996) was subsequently demonstrated. In the first exploration of therapeutic gene delivery, Nishi et al. demonstrated monocyte invasion into tumors after delivery of the gene encoding monocyte chemoattractant protein-1 (Nishi et al. 1996). Within a few years, several preclinical studies demonstrated therapeutic efficacy. Intramuscular electrotransfer of a plasmid encoding erythropoietin achieved therapeutic levels and increased the hematocrit (Kreiss et al. 1999; Rizzuto et al. 1999). Intratumor delivery of a plasmid encoding dominant-negative Stat3 induced apoptosis in melanomas (Niu et al. 1999). Eventually, in the first therapeutic gene electrotransfer clinical trial, the p35 and p40 cDNAs of interleukin 12 were delivery intratumorally, inducing tumor regression and other antitumor effects in both treated and untreated tumors (Daud et al. 2008). Gene therapy by electrotransfer has reached clinical trials for several applications including cancer vaccines and therapies and infectious disease vaccines. Overall, more than 80 in vivo clinical trials have utilized electroporation for delivery, with seven of these trials reaching Phase II. Since it was first described, electroporation has been used to transport a variety of molecules ranging from ions to drugs to nucleic acids across the plasma membrane and into the cell. Small molecules such as chemotherapeutic agents diffuse freely

8

L.C. Heller

through the plasma membrane during or after electrotransfer. Negatively charged oligonucleotides such as small interfering RNAs are approximately 6 kDa in size and freely enter a cell’s cytoplasm when present during electrotransfer, but do not enter when added after pulse application. An average plasmid weighs >3000 kDa and the mechanism by which it enters the cell is significantly more complex (Rosazza et al. 2016), requiring a combination of membrane electroporation, DNA electrophoresis, endocytosis, and intracellular transport via the cytoskeleton.

Nonviral Nucleic Acid Formats for In Vivo Delivery Plasmids are the most common format for electrotransfer-based gene therapies. Sequences as large as 15 kilobases can be inserted into plasmid DNA although manipulating the plasmid becomes more difficult with larger inserts. Along with promoters and enhancers, a translation start site, or Kozak sequence, is necessary for eukaryotic protein expression. A secretory sequence allows protein secretion if necessary. An internal ribosome entry site allows polycistronic co-expression in the same transfected cell. This is useful for therapeutic molecules encoded by multiple mRNAs such as interleukin 12 and for co-expression of complimenting therapeutic genes such as angiogenic factors. Finally, the sequence must a have a polyadenylation signal to produce readable mRNA. Gene expression is tightly regulated in nature. For example, housekeeping genes are transcribed and translated at low levels to maintain basic cellular processes. The expression of many genes is regulated in response to environmental or developmental stimuli. Mammalian regulatory elements can be located proximal to a gene or hundreds of thousands of base pairs distant. Similar to the coding sequences, these sections must be minimized in gene therapy applications. In many clinical trials, gene expression is driven by viral promoters such as the cytomegalovirus immediate early enhancer and promoter or simian virus 40 promoter. These promoters are short sequences that produce high transgene expression in most mammalian cell types but are inactivated over time, limiting expression duration. Tissue-specific promoters, which theoretically restrict expression to a specific tissue and therefore reduce off-target effects, are also commonly used in clinical trials although expression is generally much lower than that produced by viral promoters. Cell-specific targeting is also possible via sequences controlling nuclear import (Dean 2013). Promoters that can be controlled by an exogenous factor, referred to as regulatable promoters, also have potential but should be designed for inducible rather than repressible expression to minimize patient exposure to the inducing agent. Along with the sequences necessary for expression of the therapeutic gene or cDNA in mammalian cells, plasmids require prokaryotic sequences, including an antibiotic resistance gene and origin of replication, for propagation in E. coli. Major regulatory agencies have found these prokaryotic sequences to be problematic, but disagree on the limitations that should be placed on the plasmids used for gene therapy. The EMA has recommended the use of vectors without antibiotic resistance gene or other selection element to avoid horizontal gene transfer to commensal

Principles of Electroporation for Gene Therapy

9

bacteria. In this case, these bacterial sequences can be removed by several methods to produce minicircles, although low efficiency hampers their production and purification. Linear expression cassettes generated using PCR completely avoid the presence of bacterial sequences. The FDA states that beta lactam antibiotics should be avoided due to the risk of patient hypersensitivity to these antibiotics and that antibiotics in clinical use should be avoided. The policy of the State Food and Drug Administration of China corresponds with the FDA in that penicillin should be avoided to prevent allergic reactions in some patient populations. Messenger RNA does not require nuclear transport, so transgene expression after delivery may be more reproducible than DNA delivery. The most common application of mRNA electrotransfer is to engineer antigen presenting cells in vitro to induce antitumor activity when these cells are injected into the patient. Gene expression can be downregulated via RNA interference following direct delivery of sequence targeted small RNAs. Double-stranded small interfering RNA and stemloop containing microRNA are 20–30 base pair oligonucleotides based on endogenous RNAs. These RNAs are processed in the cytoplasm via similar pathways to induce cleavage and degradation of specific target mRNAs. RNA-based gene therapies may be difficult to administer because RNA is degraded by ubiquitous and resilient ribonucleases. One alternative option for RNA interference is the delivery of plasmid DNA encoding small hairpin RNAs driven by the RNA polymerase III promoters U6 or H1. Another alternative is to stabilize the molecule by chemical modification of the nucleic acid backbone, producing nuclease resistance. Each individual mammalian cell possesses the ability to detect pathogen invasion by binding of pattern recognition receptors. These receptors detect specific pathogen components such as bacterial endotoxin as well as shared pathogen components such as nucleic acids, producing an inflammatory response and/or cell death. Any gene therapy, including viruses, plasmids, and oligonucleotides, can be detected by these receptors and misinterpreted as pathogen invasion. For example, the endosomal tolllike receptor 9, found in immune cells, detects a specific sequence motif, an unmethylated CpG motif, found in artificially and bacterially synthesized. Multiple cytosolic DNA-specific pattern recognition receptors are found in all cell types (Desmet and Ishii 2012). Since plasmid DNA is also found in the cell’s cytosol after electrotransfer (Rosazza et al. 2016), these receptors may also be bound and activated (Znidar et al. 2016), producing unanticipated endogenous responses.

Electric Pulses Used for In Vivo Electrotransfer Two basic pulse types can be used for nucleic acid delivery in vivo. Exponentially decaying pulses are produced when a charged capacitor discharges into the tissue (Fig. 4a). These are commonly used in bacterial transfection and were used in the initial demonstrations of the in vivo delivery of chemotherapeutic agents (Okino and Mohri 1987) and plasmid DNA (Titomirov et al. 1991). Square wave pulses allow better control of pulse parameters necessary for the clinical development of in vivo delivery. To create square wave pulses, the exponential pulse is truncated by fast

10

L.C. Heller

Fig. 4 Electroporation pulse types. (a) Exponential decay or capacitor discharge pulse, (b) square wave pulse produced by limiting the length of an exponential pulse. Pulse types used in gene therapies include (c) short, high voltage pulses, (d) long, low voltage pulses, and (e) pulse combinations

switching. Although these pulses are referred to as “square wave,” a small voltage drop still occurs over the time of the pulse (Fig. 4b). Pulse variables can be used to deliver nucleic acids and to tailor the consequent protein expression levels and kinetics of a therapeutic protein. These variables include pulse applied voltage, length, number, and frequency. For in vivo applications, the pulse voltage-to-distance ratio is generally 20) and longer pulse durations (>100 ms); inter-pulse interval had no effect on numbers of transduced cells (separation 50 ms–1 s); using 100 ms pulses enabled translation of HEK293 cell monolayer findings to reduced voltage dependence (to 10 V) for in vivo guinea pig cochlea gene electrotransfer (Browne et al. 2016).

Electric Field Focusing Through Variable Configuration of Cochlear Implant Electrode Arrays Sampling of the voltage around the eight-node cochlear implant array as voltage pulses were delivered to it, using platinum probe and reference electrodes isolated from the voltage pulses, enabled reconstruction of the electric fields for the range of wiring configurations used for GFP readout of gene electrotransfer in HEK293 cell monolayers (Browne et al. 2016). This revealed, as predicted, that in the tandem

14

G.D. Housley et al.

Fig. 10 Point sampling of voltage within the electric field alongside a 5 mm eight electrode cochlear implant array wired in the tandem configuration (+/+/+/+/ / / / ). Voltage in the field (Vf) is shown in response to 5 V applied to the array as shown in inset. The steepest change in voltage over distance (electric field strength) occurs in the mid-region of the array. Upper image is a video montage showing the sequential placement of the probe used to measure Vf relative to the eight-node cochlear implant array

configuration, a null extended orthogonally from the point between electrodes 4 and 5, which is the boundary between the four ganged anodes and the four ganged cathodes. The highest magnitude-sampled voltages were at the ends of the array, but the region of greatest field strength (change in voltage over distance) was in the area surrounding the mid-position of the array, which closely matched the distribution of the GFP reporter cells. Thus the CFE-based gene electrotransfer is not dependent upon the magnitude of the voltage transient for the pulses but rather to the local transcellular field strength in a place-specific manner. Figure 10 shows an example of the measurement of the electric field along the length of the cochlear implant electrode array using an isolated Pt electrode sensor on a translational stage with 5 V applied to the array. This generated a maximum field strength for the tandem

Cochlear Implant Close-Field Electroporation

15

array configuration of ~2.6 V/cm. Allowing for the polarizing potential of the platinum electrodes (~2.5 V), an estimate of the maximum field strength adjacent to the cochlear implant array at 20 V applied is ~21 V/cm, hence at a cellular level, pancellular field gradients providing efficient gene electrotransfer are ~2.1 mV/μm, generated in a highly nonlinear state within millimeters of the bionic array. In comparison, conventional open-field electroporation typically requires hundreds of volts to create a broadly linear field of equivalent intensity, with consequent toxicity and limited spatial control of the gene electrotransfer, even when using arrays of needle electrodes in applications such as gene therapy based on stimulation of immune responses in cancers (Heller 2016; Heller and Heller 2010).

Extension of Close-Field Electroporation Measurement of the electric fields around the cochlear implant array across various wiring configurations of the electrodes facilitated the understanding of minimum parameters to achieve efficient electric field focusing. To test the concept that the tandem configuration provides an optimum compression of the current path along the axis of the array, this was modeled using two electrodes, each equivalent in dimensions to the four ganged electrodes forming either the anode or cathode in the tandem eight-node array. In this prototype, the gap between the electrodes was insulated with epoxy, and a heat shrunk plastic sheath enabled delivery of the DNA through the initial electrode port (Fig. 11). This “gene delivery probe” was

Fig. 11 Prototype gene delivery probe developed using the submillimeter dimensions of the cochlear implant array but comprising two electrodes, each equivalent to the four ganged electrodes used as an anode or cathode for the tandem configuration of the cochlear implant array that has this highest gene electrotransfer efficiency with CFE. Inset shows the emission of the plasmid DNA solution from a port located between the two electrodes. This ensures control of the DNA concentration in the vicinity of the electrodes, which is a critical feature of close-field electroporation

16

G.D. Housley et al.

Fig. 12 HEK293 cell monolayer showing nuclear-localized green fluorescence protein (nlsGFP) following close-field electroporation (CFE) with a prototype two-electrode configuration of CFE gene delivery probe (see Fig. 11). (a, b) Show spherical fields of transformed cells, as anticipated based on field focusing akin to that achieved using the cochlear implant array wired in the tandem configuration. (c) control where the DNA solution was placed onto the coverslip of HEK293 cells on the microscope stage, and the gene delivery probe was lowered onto the cells, but no electrical pulses were delivered. Arrows define the perimeter of the field of cells expressing GFP

tested using the HEK293 cell monolayer, with standard CFE electroporation parameters (20 V, 5  50 ms pulses) and produced regions of GFP-positive cells equivalent to that achieved using the bionic array-based CFE (Fig. 12). Modeling of the electric fields generated by CFE-based gene electrotransfer is a valuable tool in refining the design of both electrode geometries as well as electrode return path and stimulus configurations. Figure 13 demonstrates a basic model of the cochlear array placed linearly in a saline bath with electroporation waveforms

Cochlear Implant Close-Field Electroporation

17

a

0.6 0.4 0.2 0 5

5

0 0

-5

b

-5

Multislice : log10(ec.normE) 5 4.5 4 3.5 3 2.5 2 1.5 1 y

0.5 x y

C

0 Multislice : log10(ec.normE) 5 4.5 4 3.5 3 2.5 2 1.5 1

y

0.5 x

0

Fig. 13 A finite element model (a) of the eight electrode cochlear implant array placed in a saline bath. The electric field generated by the model for a tandem CFE array configuration (b) (+/+/+/+/ / / / ) has a spherical distribution about the center. The same model but for an alternating CFE array configuration (c) (+/ /+/ /+/ /+/ ) shows a reduced field intensity that has a linear dispersion. Models were generated and solved using the COMSOL software (COMSOL Multiphysics, Burlington USA)

18

G.D. Housley et al.

applied in both alternating and tandem electrode configurations. This modeling approach has the potential to be extended to incorporate a distant monopolar ground return configuration along with the alternating and tandem configurations to create more refined field shapes (Khalili Moghadam et al. 2013).

Conclusion Close-field electroporation, which is defined as the application of a nonlinear electric field projected into a target tissue, arose from the application of the cochlear implant electrode arrays for delivery of voltage pulses into the small cylindrical chambers of the cochlea. This perilymphatic chamber has dimensions closely matched to the submillimeter diameter of the linear bionic array. The electrical pulses which were applied were outside the operational capacity of the cochlear implant controller used for hearing and were driven using a conventional isolated constant voltage electroporator which could be wired to drive the array of electrodes in varying configurations. Delivery of a neurotrophin gene therapy cassette in a naked plasmid by CFE resulted in local transduction of the mesenchymal cells lining the perilymphatic compartment of the cochlea, but only in the region where the cochlear implant array was closely positioned to the target cells, as evident from GFP reporter fluorescence. The cells produced the BDNF, and in a deafened guinea pig model, this recombinant BDNF was sufficient to locally drive regrowth of the peripheral neurite processes of the primary auditory neurons into the vicinity of the cochlear implant array. This improved the neural interface by closing the gap between the SGNs and electrodes, which was supported functionally by significant reductions in the stimulus thresholds and expansion of the recruitment bandwidth, measured by electrically evoked ABRs. These animal studies determined that a particular configuration referred to as the tandem array had unanticipated efficiency in gene electrotransfer, which was found, though CFE studies with HEK293 cell monolayers, to be due to nonlinear compression of the electric field with respect to the linear electrode array. Control of the electrode configuration enabled the shape of the region of cells that were transduced to be manipulated, while selection of voltage pulse parameters enabled dial-up precision in gene delivery in an HEK293 cell monolayer model. These factors combine to make CFE a highly flexible gene electrotransfer platform which shows promise for use as a gene therapy adjuvant to improve hearing with cochlear implants. The CFE platform also has the potential for extension of application to other tissues and experimental applications, where it could provide robust and highly controlled delivery of naked DNA payloads. Acknowledgments Dr. Amr Al Abed is thanked for the electric field modeling shown in Fig. 12. Funded by an Australian Research Council Discovery Grant (DP151014754). All data derived from animal experiments followed protocols approved by the UNSW Animal Care and Ethics Committee.

Cochlear Implant Close-Field Electroporation

19

Cross-References ▶ Basic Description of Cell Electroporation in Vitro - Gene Delivery by Electroporation in Vitro: Mechanisms ▶ Basic Modeling of Electroporation and Related Phenomena - Electric Field Distribution and Electroporation Threshold ▶ Biomedical Applications of Electroporation - Clinical Applications of Gene Therapy: Principles of Gene Electrotransfer ▶ Fundamental Mechanisms of Electroporation: Models and Experiments - Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Molecular Models ▶ Fundamental Mechanisms of Electroporation: Models and Experiments - Effect of Pulsed Electric Fields on Nucleic Acid Salt Structures ▶ Fundamental Mechanisms of Electroporation: Models and Experiments - Electroporation and Electropermeabilization ▶ Preclinical Development for Biomedical Application of Electroporation - Principles of Electroporation for Gene Therapy ▶ Preclinical Development for Biomedical Application of Electroporation - Tissue Engineering with Electroporation

References Browne CJ, Pinyon JL, Housley DM, Crawford EN, Lovell NH, Klugmann M, Housley GD (2016) Mapping of bionic array electric field focusing in plasmid DNA-based gene electrotransfer. Gene Ther 23:369–379. doi:10.1038/gt.2016.8 Burgain-Chain A, Scherman D (2013) DNA electrotransfer: an effective tool for gene therapy, gene therapy – tools and potential applications. In: Molina FM (ed) Gene therapy – tools and potential applications. InTech, Croatia doi:10.5772/52528 Clark GM, Clark JC, Furness JB (2013) The evolving science of cochlear implants. JAMA 310:1225–1226. doi:10.1001/jama.2013.278142 Golzio M, Rols M (2016) Nucleic acid electrotransfer in mammalian cells: mechanistic description. In: Miklavčič D (ed) Handbook of electroporation. Springer International Publishing, Switzerland doi:10.1007/978-3-319-26779-1 Greenwood D et al (2007) P2X receptor signaling inhibits BDNF-mediated spiral ganglion neuron development in the neonatal rat cochlea. Development 134:1407–1417 Heller LC (2016) Principles of electroporation for gene therapy. In: Miklavčič D (ed) Handbook of electroporation. Springer International Publishing, Switzerland doi:10.1007/978-3-319-26779-1 Heller LC, Heller R (2010) Electroporation gene therapy preclinical and clinical trials for melanoma. Curr Gene Ther 10:312–317 Heller R, Jaroszeski MJ, Gilbert R (2011) Electroporation system and method for facilitating entry of molecules into cells in vivo. US Patent no. 7,879,610. U.S.A. Patent Housley GD (2016) Method of providing agents to the cochlea. US Patent application no. 14/145,673 (approved). United States of America Patent Khalili Moghadam G, Wilke R, Suaning GJ, Lovell NH, Dokos S (2013) Quasi-monopolar stimulation: a novel electrode design configuration for performance optimization of a retinal neuroprosthesis. PLoS One 8:e73130. doi:10.1371/journal.pone.0073130

20

G.D. Housley et al.

O’Leary SJ, Richardson RR, McDermott HJ (2009) Principles of design and biological approaches for improving the selectivity of cochlear implant electrodes. J Neural Eng 6:055002. doi:S17412560(09)05488-3 [pii] Pinyon JL (2016) Enhancing the bionic ear through close-field electroporation gene therapy UNSW Australia Ph.D. thesis Pinyon JL et al (2014a) Close-field electroporation gene delivery using the cochlear implant electrode array enhances the bionic ear. Sci Transl Med 6:233ra254. doi:10.1126/ scitranslmed.3008177 Pinyon JL et al (2014b) Close-field electroporation gene delivery using the cochlear implant electrode array enhances the bionic ear (supplementary information). Sci Transl Med 6 (233):233ra54. doi:10.1126/scitranslmed.3008177 Rebersek M, Faurie C, Kanduser M, Corovic S, Teissie J, Rols MP, Miklavcic D (2007) Electroporator with automatic change of electric field direction improves gene electrotransfer in-vitro. Biomed Eng Online 6:25, 1475-925X-6-25 [pii] Richardson RT et al (2009) Polypyrrole-coated electrodes for the delivery of charge and neurotrophins to cochlear neurons. Biomaterials 30:2614–2624 Satkauskas S, Bureau MF, Puc M, Mahfoudi A, Scherman D, Miklavcic D, Mir LM (2002) Mechanisms of in vivo DNA electrotransfer: respective contributions of cell electropermeabilization and DNA electrophoresis. Mol Ther 5:133–140. doi:10.1006/mthe.2002.0526 Wise AK et al (2010) Effects of localized neurotrophin gene expression on spiral ganglion neuron resprouting in the deafened cochlea. Mol Ther 18:1111–1122. doi:10.1038/mt.2010.28 Yandell K (2014) Hearing help. Scientist 29(9):43–49, http://www.the-scientist.com/?articles.view/ articleNo/43819/title/Hearing-Help/ Zeng FG, Tang Q, Lu T (2014) Abnormal pitch perception produced by cochlear implant stimulation. PLoS One 9:e88662. doi:10.1371/journal.pone.0088662

Combined Treatment of Electrochemotherapy with Irradiation Simona Kranjc, Urska Kamensek, Maja Cemazar, and Gregor Sersa

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrochemotherapy and Why Combine It with Irradiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In Vitro Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . In Vivo and Clinical Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Side Effects of Electrochemotherapy and Irradiation Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 4 5 7 14 15 15 15

Abstract

Radiotherapy is the use of ionizing radiation in the treatment of malignant tumors. It is one of the main treatment modalities for many forms of cancer, with more than half of all cancer patients receiving radiation therapy at some point in their treatment. Despite the technical progress in targeting ability of radiotherapy, poor therapeutic window remains a problem; therefore in clinical practice radiotherapy is commonly combined with chemotherapy. By using electrochemotherapy, the uptake of the non or poorly permeable chemotherapeutic drugs, bleomycin and cisplatin, into the tumors can be increased and thus also their radiosensitizing effect. In combined treatment of electrochemotherapy preceding irradiation, an increased radio-response was demonstrated with an enhancement factor of up to

S. Kranjc (*) • U. Kamensek • G. Sersa Department of Experimental Oncology, Institute of Oncology Ljubljana, Ljubljana, Slovenia e-mail: [email protected]; [email protected]; [email protected] M. Cemazar Department of Experimental Oncology, Institute of Oncology Ljubljana, Ljubljana, Slovenia Faculty of Health Sciences, University of Primorska, Izola, Slovenia e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_61-1

1

2

S. Kranjc et al.

4.6, regardless of using radiomimetic (bleomycin) or radiosensitizing (cisplatin) drug. The improved antitumor effectiveness is mainly enabled by increased drug accumulation in the tumors. Radiosensitization was demonstrated in vitro and in vivo in variable tumor models irradiated in a single dose and a fractionated regime. Low and acceptable radiation damage of normal tissue in the irradiation field was observed. All the data provided a starting point for translation of combined electrochemotherapy and tumor irradiation into the clinic, which has already begun. Keywords

Electrochemotherapy • Cisplatin • Bleomycin • Irradiation • Radiosensitization • Radio-response

Introduction Radiation is one of the oldest and most common treatments of cancer and over 50 % of patients receive it at some time during their course of treatment (Begg et al. 2011). It is a localized treatment, valuable and successful in confining cancer locally. Radiation can be used with the purpose of cure or as palliative treatment to alleviate the symptoms of cancer. Although the technological improvements in the delivery of radiation that reduce side effects on normal tissues in irradiated field, while enhancing the effect on the tumor, were made in the last decade, new approaches for widening of the therapeutic index are still being sought (Kamrava et al. 2009). Understanding of cancer as a multigenetic and multicellular disease led to the development of multimodal treatment approaches of cancer. Radiation can be used in combination with surgery, immunotherapy, hormonal therapy, chemotherapy, or targeted therapies (Begg et al. 2011; Joiner and Kogel 2009; Kamrava et al. 2009). Treatment that combines chemotherapy with radiation therapy has been termed chemoradiotherapy. Systemic treatments with chemotherapeutic drugs given in different regimens in combination with radiotherapy have been shown to greatly improve the overall survival of cancer patients. In fact, certain chemotherapeutic drugs can sensitize tumors to irradiation and provide a systemic effect in addition to the local effect of the radiation; thus an additive or even synergistic effect in local tumor control has been demonstrated (Begg et al. 1994; Seiwert et al. 2007; Strojan et al. 2016). Enhanced radiation sensitivity occurs through different mechanisms, i.e., increased DNA damage and decreased DNA repair, interference with the cell cycle, hypoxic sensitization and damage of tumor vasculature, immunological effects, and others (Joiner and Kogel 2009). In combined modality regimen in the clinic, the radiosensitization of tumor cells was demonstrated when systemic drug administration preceded tumor irradiation (Joiner and Kogel 2009; Seiwert et al. 2007). However, besides potentiation of the radiation response, systemic chemotherapy also contributes to local or systemic toxicity (Joiner and Kogel 2009).

Combined Treatment of Electrochemotherapy with Irradiation

3

Among chemotherapeutic drugs in combined modality, mostly cisplatin and other platinum-based agents are used in chemotherapeutic schemes. Cisplatin binding to DNA causes single and double DNA strand cross-links that differ from single and double DNA strand breaks caused by radiation, which are reflected in decreasing efficiency of repair mechanisms in sublethal and potentially lethal DNA damages after combined treatment modality (Joiner and Kogel 2009). Furthermore, cisplatin exhibits radiosensitization through cell cycle arrest, suppression of tumor neovascularization, and especially in hypoxic cells through the formation of OH (hydroxil) radicals by scavenging of hydrated electrons (Joiner and Kogel 2009; Seiwert et al. 2007). That cisplatin radiosensitizes tumor cells by complex interaction, which was shown in several in vitro and in vivo studies (Begg et al. 1994; Gorodetsky et al. 1998). Results of preclinical studies using different schedules and dosages of each therapeutic ranged from not significant interaction to pure additive effect and also synergistic interaction. Synergistic effect was exhibited if cisplatin was given shortly before or immediately after irradiation in a single dose or fractionated regime, confined mostly to hypoxic cell population (Begg et al. 1994; Gorodetsky et al. 1998; Lagrange et al. 1993). On the contrary, when the time interval between cisplatin administration and irradiation was longer than 2 h, only an additive effect was observed (Joiner and Kogel 2009; Rantanen et al. 1995; Seiwert et al. 2007). Several clinical studies have shown synergy of combined treatment, which leads to significant improvement of radio-response, resulting in better local tumor control and overall survival. Clinical data of concurrent chemoradiotherapy confirmed the radiosensitization with cisplatin in various cancers (head and neck, gynecological, lung, breast, glioma) (Joiner and Kogel 2009; Strojan et al. 2016), and chemoradiotherapy with cisplatin is nowadays standard therapy in management of head and neck cancer, significantly improving locoregional failure-free survival at 5 years (Strojan et al. 2016). Bleomycin is another chemotherapeutic drug frequently used in chemotherapeutic schemes in combination with radiotherapy, mainly in the treatment of head and neck tumors (Zakotnik et al. 2007). Bleomycin belongs to a group of antitumor glycopeptide antibiotics and can induce single and double DNA breaks (Dorr 1992). The major mechanism of bleomycin action is the formation of pseudoenzyme due to the chelation of metal ions, which by reacting with oxygen free radicals causes the DNA breaks. Since bleomycin predominantly acts well on oxygenated cells, it is necessary to combine it with other chemotherapeutic drugs to hit the wide specter of biological targets in tumors (Joiner and Kogel 2009). By mode of action, bleomycin is classified as radiomimetic and not radiosensitizer (Joiner and Kogel 2009). In combined modality of bleomycin and irradiation, a potentiation of radiation response was demonstrated, mainly due to pronounced enhancement of bleomycin-induced DNA damages by radiation (Joiner and Kogel 2009). Preclinical studies demonstrated that bleomycin has to be administered up to 20 min prior to a single-dose or fractionated irradiation schedule in order to achieve the best potentiation of radiotherapy (Jiang et al. 1989; Joiner and Kogel 2009). Chemoradiotherapy with bleomycin alone or in chemotherapy regime

4

S. Kranjc et al.

with other drugs concurrently with a fractionated irradiation regime was also proven effective in the clinical setting, resulting in improvement of locoregional control and survival of patients. Furthermore, radiosensitization of head and neck tumors with bleomycin has become an established treatment approach (Zakotnik et al. 2007). Taken together, combined modality treatment of cancer by radiation and chemotherapy, i.e., chemoradiotherapy, improves locoregional control, disease free and overall survival, and is now standard of care for several types of tumors. However, unwanted side effects and high percentage of mortality still remain a great challenge in disease control. Therefore, better understanding of biological properties of cancer cells, accurate planning, and technological development of combined treatment is needed.

Electrochemotherapy and Why Combine It with Irradiation The major limitation of chemoradiotherapy is the cumulative normal tissue toxicity and insufficient effectiveness of the available drug in most of the tumors. Therefore, research is aimed towards the innovative approaches to target cancer cells more specifically and locally, in order to improve treatment and prevent side effects. A promising approach to increased delivery of chemotherapeutic drugs to cells and tumors is by various delivery systems, using tumor-specific antibodies, magnetic particles, liposomes, or other vehicles, or by selectively increasing permeability of the plasma membrane by chemical or physical methods (Tiwari et al. 2012). Among them electroporation (see “▶ Electroporation and Electropermeabilization”) could be used to increase intracellular accumulation of drugs, which consequently increase radiosensitization of tumor cells (Yarmush et al. 2014). The phenomenon of electroporation by using short high-voltage electric pulses, which induces membrane permeabilization (see “▶ Electroporation and Electropermeabilization”), allowing drug uptake into the cells, was discovered in the early 1980s of the last century (Yarmush et al. 2014). Local application of electric pulses (electroporation) to the tumor combined with local or systemic administration of poorly or nonpermeable chemotherapeutic drugs was named electrochemotherapy (see “▶ Electrochemotherapy and Its Clinical Applications”) (Mir et al. 1991; Yarmush et al. 2014). Among chemotherapeutic drugs used in the treatment of cancer, bleomycin and cisplatin were demonstrated as the most potent in electrochemotherapy. By using electrochemotherapy, the potentiation (several fold) of cytotoxicity and antitumor effectiveness of bleomycin and cisplatin has been proven in several preclinical studies. An increase in drug delivery to the cells and tumors was demonstrated to be the main underlying mechanism for improved cytotoxicity in vitro and antitumor effect in vivo (Yarmush et al. 2014). Furthermore, human (see “▶ Electrochemotherapy and Its Clinical Applications,” “▶ Electrochemotherapy of Cutaneous Metastases,” and “▶ Electrochemotherapy of Head and Neck Cancer”) and veterinary (see “▶ Electrochemotherapy in Veterinary Oncology”) clinical studies dealing with cutaneous and subcutaneous tumor nodules of different histology have shown

Combined Treatment of Electrochemotherapy with Irradiation

5

similar response rate, approximately 80 % objective responses and approximately 70 % long-lasting complete responses (Mali et al. 2013). Electrochemotherapy is mainly used as adjuvant or palliative therapy for solid cutaneous melanoma, carcinoma, and sarcoma metastases (see “▶ Electrochemotherapy and Its Clinical Applications,” “▶ Electrochemotherapy of Cutaneous Metastases,” and “▶ Electrochemotherapy of Head and Neck Cancer”) and with technological improvement became feasible also for deep-seated liver metastases (see “▶ Electrochemotherapy and Its Clinical Applications”) (Yarmush et al. 2014). The effectiveness of combined electrochemotherapy with cisplatin or bleomycin and radiotherapy was tested in several preclinical studies, which demonstrated that electrochemotherapy acts synergistically with radiotherapy, exerting the radiosensitizing effects on different types of cells in vitro and tumors in vivo. The encouraging results led to a few clinical reports.

In Vitro Data Only a few in vitro studies evaluated combined treatment of electrochemotherapy with irradiation. In the first study, cisplatin-sensitive SCK mammary carcinoma cells turned out to be more radiosensitive than cisplatin-resistant EAT-E (Ehrlich-Lettre ascites carcinoma) cells (Kranjc et al. 2003a). In fact, intrinsic radiosensitivity determined at D0 (D0- a dose that reduces survival from 0.1 to 0.037) was 1.2 Gy for SCK cells and 2.0 Gy for EAT-E cells. Exposure of cells to cisplatin, prior to irradiation with 4 Gy, significantly increased radiation response of cells. Moreover, concomitant exposure of cells to cisplatin and application of electric pulses, i.e., electrochemotherapy, before irradiation additionally decreased cell survival in both cell lines to the same level (IC50 = 0.9  0.2 μg/ml for SCK cells and IC50 = 0.9  0.3 μg/ml for EAT-E cells). Thus electrochemotherapy of cells with cisplatin combined with irradiation was proven to radiosensitize the cells with different intrinsic chemo- and radiosensitivity, resulting in decreased cell survival by 3.7fold in SCK cells and by 2.4-fold in EAT-E cells compared to electrochemotherapytreated cells only. Similar results were obtained on moderately radiosensitive LPB sarcoma cells (D0 = 1.6), demonstrated by two fold decrease of cell survival (Fig. 1) (Kranjc et al. 2003b). Moreover, radiosensitizing effect (1.4-fold) of cisplatin electrochemotherapy in combination with graded doses of irradiation from 2 to 8 Gy was shown. However, radiation response was enhanced to some extent also by electroporation only (1.1-fold), while cisplatin alone treatment at low concentration had no effect. Another drug used in electrochemotherapy is bleomycin (see “▶ Electroporation and Electropermeabilization” and “▶ Electrochemotherapy and Its Clinical Applications”), which is cytotoxic if only few hundred molecules are present inside the cells (Yarmush et al. 2014). Thus, by increasing its uptake after electropermeabilization of cells, the radio-response of cells could be potentiated. In fact, the electroporation of cells with low doses of bleomycin was shown to have a good radiosensitizing effect (Kranjc et al. 2005). Treatment of cells either with bleomycin

6

S. Kranjc et al.

Fig. 1 Survival curves of LPB sarcoma cells irradiated with graded doses only (IR) or irradiated after combined treatment with cisplatin (CDDP) or bleomycin (BLM) and application of electric pulses (electrochemotherapy, ECT) (ECT BLM + IR, ECT CDDP + IR) (Adapted from Kranjc et al. (2003b, 2005))

or electroporation alone had some radiosensitizing effect (around 1.2-fold), whereas after bleomycin electrochemotherapy the radiation response was increased for a factor of 1.5 (Fig. 1). In summary, it was determined from in vitro experiments, radiosensitization by electrochemotherapy with bleomycin was more pronounced compared to electrochemotherapy with cisplatin, (1.4-fold for cisplatin and 1.5-fold for bleomycin) (Kranjc et al. 2003b). However, the exposure of cells to electric pulses alone prior to irradiation potentiated radiation response as well (up to 1.3-fold). Therefore, obtained radiosensitization of cells exposed to electrochemotherapy might not be ascribed just to the increased drug concentration after electroporation but also to the effect of exposure of cells to short high-voltage electric pulses (see “▶ Involvement of Reactive Oxygen Species in Membrane Electropermeabilization”) (up to 1.2-fold) (Kranjc et al. 2003b) (Fig. 1). Namely, reversible electroporation (see “▶ Involvement of Reactive Oxygen Species in Membrane Electropermeabilization” and “▶ Preclinical Studies on Reversible Electroporation”) induces the generation of free reactive oxygen species, which are present only for a short period of time after electroporation and can damage proteins, lipids, and DNA (Shil et al. 2005). Reactive oxygen species are known to contribute to radiation damage of the cells. Therefore, predisposition of cells to irradiation by electroporation is also expected and was demonstrated (West 1992; Shil et al. 2005). In fact, a significant increase in production of free reactive oxygen species after combined treatment of Ehrlich ascites carcinoma cells with radiation and electroporation was obtained, resulting

Combined Treatment of Electrochemotherapy with Irradiation

7

in significant enhancement of radio-response after electroporation (Shil et al. 2005). Thus, presumably DNA damages due to bleomycin and free reactive oxygen species sensitize cells to radiation, and, consequently, decrease cell survival. Taken together, these data indicate that electroporation of tumor cells with cisplatin or bleomycin, irrespective of the varying degrees of chemo- or radiosensitivity of cells, potentiates radiosensitizing effect of either drug used.

In Vivo and Clinical Data Until now, there have only been a few preclinical in vivo studies evaluating electrochemotherapy and tumor irradiation. Combination of electrochemotherapy with cisplatin or bleomycin and single or fractionated irradiation regime were tested in different mouse experimental tumor models (EAT carcinoma; LPB and SA-1 sarcoma; and CaNT mammary adenocarcinoma) and one spontaneous mammary carcinoma model (Sersa et al. 2000; Kranjc et al. 2003a, 2005; Raeisi et al. 2012). A proof of principle, that electroporation can increase radiosensitizing effect of cisplatin, was first reported by Sersa et al. 2000. Single treatment of induced EAT subcutaneous tumors, either with intravenous injection of cisplatin (4 mg/kg) or application of electric pulses only, delayed tumor growth moderately, without any tumor cure, while electrochemotherapy alone resulted in 12 % tumor cures and irradiation with 15 Gy alone in 27 % tumor cures. Administration of cisplatin alone or application of electric pulses alone 20 min prior to irradiation increased the radio-response of tumors to 73 % and 54 % tumor cures, respectively. Consistently, electrochemotherapy given 20 min prior to tumor irradiation increased the radio-response of tumors even more, resulting in 92 % tumor cures (Sersa et al. 2000). Therefore, due to increased delivery of cisplatin into cells by electroporation, the radiosensitization of tumors with cisplatin was demonstrated. The observed radiosensitization of electroporation was hypothesized to be, on one hand, due to the effect of electroporation on intracellular targets or to the perturbation of the plasma membrane, and, on the other hand, possible mechanisms could also involve the effects of electric pulses on modification of tumor blood flow and oxygenation (Fig. 2). Next, combined electrochemotherapy treatment with cisplatin and tumor irradiation with graded doses was evaluated on LPB sarcoma tumors in order to achieve tumor curability dose of radiation (TCD50) (Fig. 3) (Kranjc et al. 2003b). In this study attempts were made to elucidate underlying mechanism of the combined modality treatment, i.e., tumor blood flow changes and platinum uptake determination. Local treatment of animals by application of electric pulses to the tumors 20 min prior to tumor irradiation did not affect the response of tumors to radiation, while the intravenous injection of cisplatin (4 mg/kg) 20 min prior to local tumor irradiation resulted in a statistically significant enhanced radio-response of tumors. TCD50 of combined modality treatment in relation to irradiated tumors only was lowered from 22.1 to 19.6 Gy. However, by electrochemotherapy 20 min prior to irradiation, a significant enhancement of tumor radio-response was achieved

8

S. Kranjc et al.

Fig. 2 Scheme of a tumor nodule treated with electrochemotherapy and irradiation. (a) Three minutes after systemic administration of cisplatin or bleomycin tumor nodule was treated with electric pulses (electrochemotherapy) and 20 min after irradiated with single dose. (b) Scheme of

Combined Treatment of Electrochemotherapy with Irradiation

9

(TCD50 = 14.2 Gy). These data proved that the radio-response of tumors after electrochemotherapy was significantly increased in relation to only irradiated tumors (enhancement factor; EF = 1.6) and to tumors treated with cisplatin and irradiation (EF = 1.4) (Fig. 3) (Kranjc et al. 2003b). Furthermore, based on the promising radiosensitization results obtained in these two studies (Sersa et al. 2000; Kranjc et al. 2003b), it was necessary to clarify the underlying mechanisms of combined treatment modality. Firstly, the increase in platinum content in the tumors due to electroporation was evaluated as the underlying mechanism of enhanced tumor radio-response (Kranjc et al. 2003b). The platinum content in tumors treated with electroporation with or without tumor irradiation was approximately twice as high as in tumors treated with cisplatin or cisplatin and irradiation. The observed difference in platinum content occurred already a few minutes after the treatment and lasted up to 24 h after application of electric pulses to the tumors (Kranjc et al. 2003b). These results indicated that increased delivery of cisplatin to tumors by application of electric pulses is probably the main mechanism of the observed increase in cisplatin radiosensitization of tumors. Worth mentioning

Fig. 3 Local tumor control curves of LPB sarcoma tumors treated with combination of systemic administration of cisplatin (CDDP) or bleomycin (BLM) and application of electric pulses (electrochemotherapy, ECT) 20 min prior to irradiation (IR) (ECT BLM + IR or ECT CDDP + IR); enhancement factor (EF); irradiation dose (D) (Adapted from Kranjc et al. (2003b, 2005)).

ä Fig. 2 (continued) tumor nodule after treatment with electrochemotherapy and radiation. Formation of reactive oxygen species and toxic effect of chemotherapeutic drug due to increased drug uptake in tumor cells overcame the depletion in oxygen (hypoxia) caused by reduced blood flow after application of electric pulses. Tumor cells are dying of apoptotic and necrotic cell death, resulting in good local tumor response.

10

S. Kranjc et al.

is also that irradiation of tumors had no influence on the electroporation-mediated increase in platinum content and the kinetics of platinum washout from the tumors. As a proposed second mechanism involved in tumor radio-response after electrochemotherapy are the effects of electrochemotherapy on tumor blood flow. Namely, the blood flow is strongly associated with oxygen levels in tumors that play an important role in radiation response (Fig. 2b). Measurements of tumor blood flow (see “▶ Effects of Electroporation on Blood Flow,” “▶ Tumor Blood Flow Vs Normal Blood Flow and Electroporation,” and “▶ Vascular Effects of Electroporation and Electrochemotherapy”) demonstrated that cisplatin or irradiation monotherapy, as well as combination of both did not affect perfusion level in the first 24 h after treatment (Kranjc et al. 2003b). However, the application of electric pulses to the tumors alone, and all treatment combinations with electroporation, significantly reduced perfusion in tumors immediately after electric pulses were applied, returning to 50 % of the pretreatment value at 24 h posttreatment. The reduced blood flow at the time of irradiation could lead to radiobiologically relevant hypoxia. Thus, the increased radiation response of LPB tumors treated with electrochemotherapy was the result of higher platinum content and cisplatin radiosensitization of hypoxic cells, which has exceeded the reduction in tumor blood flow and oxygenation. The increased radio-response of solid tumors treated by application of electric pulses only seems to be tumor type dependent. Namely, the application of electric pulses potentiated the radio-response in EAT carcinoma (Sersa et al. 2000; Shil et al. 2005), but not in LPB sarcoma. Proposed mechanisms could be that EAT carcinoma tumors are more sensitive to radiation because application of electric pulses generates more free reactive oxygen species (see “▶ Involvement of Reactive Oxygen Species in Membrane Electropermeabilization”), as shown in vitro (Shil et al. 2005), which prevail over the effect of reduced tumor blood flow and oxygenation, also caused by application of electric pulses. Nevertheless, the radioresponse of LPB sarcoma tumors could presumably be ascribed to outweigh the oxygen reduction over the other effects (Fig. 3). Further efforts were focused on electrochemotherapy with bleomycin as radiosensitizing therapy. To clarify the underlying mechanisms for radiosensitization by bleomycin electrochemotherapy, tumor oxygenation after combined treatment was measured in LPB sarcoma tumors (Kranjc et al. 2005). Tumors were treated by electrochemotherapy with bleomycin (0.5 mg/kg; intravenous injection) 20 min prior to single-graded doses of tumor irradiation (5–50 Gy). Treatment of animals with intravenous injection of bleomycin alone or application of electric pulses to the tumors prior to irradiation of tumors had no effect on local tumor control. Electrochemotherapy with bleomycin significantly increased radio-response of tumors compared to irradiation treatment only. The TCD50 value was lowered from 23.1 Gy, obtained after irradiation alone, to 12.4 Gy if electrochemotherapy preceded irradiation of tumors (Fig. 3). A great potentiation of radio-response (EF = 1.9) indicated that electroporation of tumors significantly contributed to radiosensitization of tumors with bleomycin, specifically because combined treatment of electrochemotherapy with irradiation was more effective than irradiation in combination with electroporation or bleomycin alone (Fig. 3). Obtained potentiation

Combined Treatment of Electrochemotherapy with Irradiation

11

of radio-response by using bleomycin in electrochemotherapy protocol was even higher compared to cisplatin, presumably due to lack of transport mechanisms for bleomycin (Kranjc et al. 2003b, 2005). In order to further support the presumption of correlation between immediate effects of electroporation or electrochemotherapy on tumor blood flow (see “▶ Effects of Electroporation on Blood Flow,” “▶ Tumor Blood Flow Vs Normal Blood Flow and Electroporation,” and “▶ Vascular Effects of Electroporation and Electrochemotherapy”) and oxygen level in tumors, in this study the partial tumor oxygen tension (pO2) in the LPB tumors was evaluated (Kranjc et al. 2005). As expected, decreased partial tumor oxygen tension (pO2) in the LPB tumors (Kranjc et al. 2005) correlated with changes in tumor perfusion. Five minutes after reversible electroporation of the tumors, pO2 was significantly reduced, in the center by 75 % and in the periphery by 50 %. Afterwards tumors started to reoxygenate, but the reoxygenation was not complete; even 24 h after electroporation oxygen levels were ~90 % of the pretreatment level. Specifically, at the time of tumor irradiation, i.e., 20 min after electroporation of tumors, oxygen levels in the tumor were still reduced to the level of radiobiological relevant hypoxia, which could have an influence on the tumor radio-response (Kranjc et al. 2005). However, radio-response of tumors after electroporation was the same as by irradiated tumors only, as demonstrated in the previous study (Kranjc et al. 2003b). Again, due to achieved enhanced radioresponse of electroporated tumor cells in vitro, some other mechanisms were assumingly involved in radio-response in vivo, i.e., reactive oxygen species formed after electroporation of cells (Shil et al. 2005; West 1992). Overall, both effects of electroporation, i.e., induced radiobiological relevant hypoxia and the formation of free reactive oxygen species, are counteracted. However, it seems that tumor type plays an important role in response of tumors to combined treatment modality of electroporation and irradiation (Sersa et al. 2000; Shil et al. 2005). Another chemotherapeutic that was used in combination with electroporation and irradiation was doxorubicin. Similar to cisplatin and bleomycin, a study in mouse EAT carcinoma tumors demonstrated that electroporation significantly enhanced antitumor effectiveness of a moderate irradiation dose and low concentration of doxorubicin; a significant tumor growth delay was obtained, 1.2 days when combined irradiation and electroporation, 1.5 days when combined doxorubicin and electroporation, and by all treatment combinations 1.7 days (Shil et al. 2006). Promising results of tumor radiosensitization with electrochemotherapy using bleomycin have driven further studies to evaluate the interaction of this treatment with single-dose or fractionated irradiation regime and to make this treatment a feasible approach in the clinic (Kranjc et al. 2009). The interaction was evaluated in two tumor models with different histology and radiosensitivity: radiosensitive SA-1 sarcoma and radioresistant CaNT adenocarcinoma tumors. In both tumor models, low irradiation doses and low dose (0.5 mg/kg) of bleomycin with a minor antitumor effect in electrochemotherapy were used in order to evaluate interaction in combined treatment. Radiosensitive SA-1 tumor with TCD50 value 25 Gy was irradiated with a cumulative dose of 10 Gy in a single or fractionated regime (2 Gy/day, 5 days/week). Radioresistant CaNT tumors with TCD50 value 70 Gy were irradiated with a

12

S. Kranjc et al.

Fig. 4 Treatment schedules of tumors with electrochemotherapy and tumor irradiation with single dose or fractionated regime

cumulative dose of 20 Gy in a single or fractionated regime (2 Gy/day, 5 days/week, 2 weeks). Electrochemotherapy was performed once, 20 min prior to the first fraction in the fractionated regime or tumor irradiation with a single dose (Fig. 4) (Kranjc et al. 2009). Radio-response of radiosensitive SA-1 and radioresistant CaNT tumors after electrochemotherapy with bleomycin was pronounced in both single-dose or fractionated irradiation regimen. The resulting tumor growth delay was synergistic, since it was more than the sum of the effects of bleomycin or application of electric pulses to the tumors combined either with single-dose or fractionated tumor irradiation. On the contrary, the radiosensitization of both tumor models was not significantly improved when combined with intravenous injection of bleomycin alone or application of electric pulses alone. In general, tumor radiosensitization effect of electrochemotherapy proved to be more pronounced in fractionated irradiation regime than in single-dose irradiation. In fact, in SA-1 tumors the radio-response after electrochemotherapy was enhanced by 2.7-fold in single-dose irradiation and by 4.6-fold in the fractionated regime (Fig. 5). In CaNT tumors, a 2.3-fold enhancement of radio-response in single-dose irradiation and a 3.1-fold in the fractionated regime was obtained. Despite obvious radio-response obtained in both tumors, the potentiation of radio-response was more evident in the radiosensitive sarcoma SA-1 tumors than in the radioresistant adenocarcinoma CaNT tumors (Kranjc et al. 2009). Furthermore, to make radiosensitization of experimental tumors with electrochemotherapy more comparable and valuable for use in clinical setting, large spontaneous mouse invasive ductal carcinoma tumors were used in the study by Raeisi et al. (2012). Tumors were irradiated in suboptimal doses of 3 Gy or 5 Gy 30 min after treatment with electrochemotherapy. Compared to other studies (Sersa et al. 2000 and Kranjc et al. 2003b), a high dose of cisplatin, 8 mg/kg, injected intratumorally, was used in electrochemotherapy protocol. Electrochemotherapy alone significantly prolonged tumor growth delay (up to 15.5 days). Local irradiation of tumors with single suboptimal doses alone or combined with cisplatin administration prior to irradiation delayed tumor growth approximately to the same extent (up to 21 days). However, electrochemotherapy with cisplatin significantly increased radio-response of tumors at both irradiation doses used. The antitumor effectiveness in combined treatment was dose dependent, resulting in tumor growth delay of 25.7 days at a dose of 3 Gy and 38.6 days at a dose of 5 Gy, with two complete responses. Compared to previously obtained radio-response of variable tumors at

Combined Treatment of Electrochemotherapy with Irradiation

13

Fig. 5 Antitumor effectiveness of electrochemotherapy (ECT) with bleomycin (BLM) combined with tumor irradiation (IR), either with the single dose or fractionated regime, in a SA-1 sarcoma and EAT carcinoma tumors. Data represent the mean of tumor doubling time and their standard error, which were pooled from at least eight animals per treatment group. Enhancement factor (EF) of radio-response was calculated based on the tumor doubling time of the compared experimental groups (Adapted from Kranjc et al. (2009))

higher irradiation doses (Sersa et al. 2000 and Kranjc et al. 2003b), these data indicate that the chosen tumor model is quite radiosensitive and not chemosensitive. With respect to clinical use, the first report of radiosensitization by electrochemotherapy with cisplatin was in a patient with tubal dedifferentiated papillary adenocarcinoma skin metastases (Sersa et al. 1999). Combined treatment, electrochemotherapy with cisplatin injected intratumorally and single-dose irradiation, was used for palliative treatment of skin metastases. In a very short observation time (11 days) after electrochemotherapy a complete reduction of tumor size was achieved, with no palpable viable tumor tissue under a scab. On contrary, no reduction in the size of skin metastases was obtained after single-dose irradiation alone or its combination with cisplatin treatment alone and application of electric pulses alone. However, treatment of skin metastases with electrochemotherapy prior to irradiation increased radio-response of metastases, resulting in a complete decrease in the size of the treated metastases, which occurred much faster than by treatment with electrochemotherapy alone (Sersa et al. 1999). In addition, a few years ago combined treatment of electrochemotherapy with bleomycin and external

14

S. Kranjc et al.

beam radiotherapy or brachytherapy for treatment of patients with very large, previously multitreated or in advanced stage tumors of different type was used (Skarlatos et al. 2011). Again, a great enhancement of radio-response by electrochemotherapy with bleomycin was observed, resulting in complete response up to 2 years. Furthermore, electrochemotherapy with bleomycin and irradiation with fractionated regime was successfully used on aggressive hidradenocarcinoma, a complete remission of this malignant lesion was achieved, and lasted up to 7 months (Kyrgias et al. 2013).

Side Effects of Electrochemotherapy and Irradiation Treatment Ongoing advances in the techniques of radiation, finding of new approaches to treat cancer in combination with irradiation, and progress made in understanding the biology of cancer cell responses to combined treatments endeavor the improvement in the survival and reduce treatment side effects for cancer patients. Thus, determination of side effects of new treatment combinations is necessary, especially to provide final therapeutic index (normal tissue damage compared to local tumor control at the dose curing 50 % of tumors) as indicator of successful therapy. Across the studies, combining electrochemotherapy and irradiation, done so far, the low doses of irradiation and chemotherapeutics (cisplatin, bleomycin) were chosen for the proper evaluation of the effects of combined treatment. As expected, treatment with electrochemotherapy prior to irradiation showed minimal or no local side effects in the irradiation field as well as no systemic side effects. Precisely, animal body weight changes during and after the treatment as well as the skin reactions around the tumor in the irradiation field after single dose or fractionated regimen were assessed (Kranjc et al. 2009). In general, single-dose irradiation alone or combined with bleomycin, application of electric pulses, and electrochemotherapy induced more serious normal skin damage and animal weight loss compared to irradiation in fractionated regime. Animals mostly lost weight at the beginning of the treatment and during the fractionated irradiation regime. However, the animal weight and physical condition were stabilized approximately 2 weeks after the treatment. Single dose irradiation alone or in combination with electrochemotherapy provoked much more serious normal skin damage a week earlier, compared to the fractionated radiotherapy regime, inducing edema, erythema, and moist desquamation with moderate ulceration. On contrary, irradiation in fractionated regime, alone or combined with electrochemotherapy, induced skin damage only after 30–40 days after the first tumor irradiation and was around four times lower, compared to single-dose irradiation, irrespective of the treatment combinations (Kranjc et al. 2009). In addition, special attention to side effects should be paid when using electrochemotherapy as adjuvant therapy in the previously irradiated field, because electrochemotherapy can induce radiation recall (Spugnini et al. 2008). However, side effects after treatment with electrochemotherapy prior to irradiation indicate this treatment as safe and feasible to provide an appropriate therapeutic index for further development of this combined treatment approach.

Combined Treatment of Electrochemotherapy with Irradiation

15

Conclusion Potentiation of the radiation response is needed in clinical situations when patients present accessible tumor masses, and good therapeutic effect is not expected in conventional irradiation treatment. In such cases electrochemotherapy can be used as radiosensitizer, using either cisplatin or bleomycin. The preclinical data provide solid evidence that either with cisplatin or bleomycin good potentiation of radioresponse can be expected in the range of enhancement factor of 1.7–4.6. The enhancement is tumor type dependent; however, radioresistant tumors are also well responsive. Only one electrochemotherapy session is needed for the response preceding the single dose or fractionated regimen. The first clinical data also indicated the potential use of combined electrochemotherapy and radiation, providing evidence for broader clinical application of this combined treatment.

Cross-References ▶ Effects of Electroporation on Blood Flow ▶ Electrochemotherapy and Its Clinical Applications ▶ Electrochemotherapy in Veterinary Oncology ▶ Electrochemotherapy of Cutaneous Metastases ▶ Electrochemotherapy of Head and Neck Cancer ▶ Electroporation and Electropermeabilization ▶ Involvement of Reactive Oxygen Species in Membrane Electropermeabilization ▶ Preclinical Studies on Reversible Electroporation ▶ Tumor Blood Flow Vs. Normal Blood Flow and Electroporation ▶ Vascular Effects of Electroporation and Electrochemotherapy

References Begg AC, Deurloo MJ, Kop W, Bartelink H (1994) Improvement of combined modality therapy with cisplatin and radiation using intratumoral drug administration in murine tumors. Radiother Oncol 31(2):129–137 Begg AC, Stewart FA, Vens C (2011) Strategies to improve radiotherapy with targeted drugs. Nat Rev Cancer 11(4):239–253. doi:10.1038/nrc3007 Dorr RT (1992) Bleomycin pharmacology: mechanism of action and resistance, and clinical pharmacokinetics. Semin Oncol 19(2 Suppl 5):3–8 Gorodetsky R, Levy-Agababa F, Mou X, Vexler AM (1998) Combination of cisplatin and radiation in cell culture: effect of duration of exposure to drug and timing of irradiation. Int J Cancer 75 (4):635–642 Jiang GL, Ang KK, Thames HD, Wong CS, Wendt CD (1989) Response of plateau-phase C3H 10T1/2 cells to radiation and concurrent administration of bleomycin. Radiat Res 120 (2):306–312 Joiner M, van der Kogel A (2009) Basic clinical radiobiology, 4th edn. Taylor & Francis Group, Boca Raton, pp 1–361

16

S. Kranjc et al.

Kamrava M, Bernstein MB, Camphausen K, Hodge JW (2009) Combining radiation, immunotherapy, and antiangiogenesis agents in the management of cancer: the three musketeers or just another quixotic combination? Mol Biosyst 5(11):1262–1270 Kranjc S, Cemazar M, Grosel A, Pipan Z, Sersa G (2003a) Effect of electroporation on radiosensitization with cisplatin in two cell lines with different chemo- and radiosensitivity. Radiol Oncol 37:101–107 Kranjc S, Cemazar M, Grosel A, Scancar J, Sersa G (2003b) Electroporation of LPB sarcoma cells in vitro and tumors in vivo increases the radiosensitizing effect of cisplatin. Anticancer Res 23 (1A):275–281 Kranjc S, Cemazar M, Grosel A, Sentjurc M, Sersa G (2005) Radiosensitizing effect of electrochemotherapy with bleomycin in LPB sarcoma cells and tumors in mice. BMC Cancer 5:115 Kranjc S, Tevz G, Kamensek U, Vidic S, Cemazar M, Sersa G (2009) Radiosensitizing effect of electrochemotherapy in a fractionated radiation regime in radiosensitive murine sarcoma and radioresistant adenocarcinoma tumor model. Radiat Res 172(6):677–685 Kyrgias G, Kostopoulou E, Zafiriou E, Zygogianni A, Skarlatos J, Roussaki-Schulze AV, Theodorou K (2013) Hidradenocarcinoma of the temporal area successfully treated with concomitant electrochemotherapy and radiotherapy. Head Neck Oncol 5(2):14 Lagrange JL, Fischel JL, Galliani S, Formento P, Guillot T, Bardon M, Milano G (1993) Importance of the irradiation timing within a chemoradiotherapy sequence including cisplatin and 5-FU-folinic acid. Experimental results. Eur J Cancer 29A(11):1531–1535 Mali B, Jarm T, Snoj M, Sersa G, Miklavcic D (2013) Antitumor effectiveness of electrochemotherapy: a systematic review and meta-analysis. Eur J Surg Oncol 39(1):4–16 Mir LM, Orlowski S, Belehradek J Jr, Paoletti C (1991) Electrochemotherapy potentiation of antitumour effect of bleomycin by local electric pulses. Eur J Cancer 27(1):68–72 Raeisi E, Aghamiri SM, Bandi A, Rahmatpour N, Firoozabadi SM, Kafi-Abad SA, Mir LM (2012) The antitumor efficiency of combined electrochemotherapy and single dose irradiation on a breast cancer tumor model. Radiol Oncol 46(3):226–232 Rantanen V, Grénman S, Kulmala J, Grénman R (1995) Simultaneous cisplatin and radiation in endometrial adenocarcinoma cell lines. Acta Oncol 34(1):93–98 Seiwert TY, Salama JK, Vokes EE (2007) The concurrent chemoradiation paradigm-general principles. Nat Clin Pract Oncol 4(2):86–100 Sersa G, Cemazar M, Rudolf Z, Fras P (1999) Adenocarcinoma skin metastases treated by electrochemotherapy with cisplatin combined with radiation. Radiol Oncol 33:291–296 Sersa G, Kranjc S, Cemazar M (2000) Improvement of combined modality therapy with cisplatin and radiation using electroporation of tumors. Int J Radiat Oncol Biol Phys 46(4):1037–1041 Shil P, Sanghvi SH, Vidyasagar PB, Mishra KP (2005) Enhancement of radiation cytotoxicity in murine cancer cells by electroporation: in vitro and in vivo studies. J Environ Pathol Toxicol Oncol 24(4):291–298 Shil P, Kumar A, Vidyasagar PB, Mishra KP (2006) Electroporation enhances radiation and doxorubicin-induced toxicity in solid tumor in vivo. J Environ Pathol Toxicol Oncol 25 (4):625–632 Skarlatos I, Kyrgias G, Mosa E, Provatopoulou X, Spyrou M, Theodorou K, Lepouras A, Gounaris A, Koukourakis M (2011) Electrochemotherapy in cancer patients: first clinical trial in Greece. Vivo 25(2):265–274 Spugnini EP1, Dotsinsky I, Mudrov N, Citro G, Caruso G, Cardelli P, Baldi A (2008) Electrochemotherapy-induced radiation recall in a cat. Vivo 22(6):751–753 Strojan P, Vermorken JB, Beitler JJ, Saba NF, Haigentz M Jr, Bossi P, Worden FP, Langendijk JA, Eisbruch A, Mendenhall WM, Lee AW, Harrison LB, Bradford CR, Smee R, Silver CE, Rinaldo A, Ferlito A (2016) Cumulative cisplatin dose in concurrent chemoradiotherapy for head and neck cancer: A systematic review. Head Neck Suppl 1:E2151–E2158 Tiwari G, Tiwari R, Sriwastawa B, Bhati L, Pandey S, Pandey P, Bannerjee SK (2012) Drug delivery systems: An updated review. Int J Pharm Investig 2(1):2–11

Combined Treatment of Electrochemotherapy with Irradiation

17

West CML (1992) A potential pitfall in the use of electroporation – cellular radiosensitization by pulsed high-voltage electric-fields. Int J Radiat Biol 61(3):329–334 Yarmush ML, Golberg A, Sersa G, Kotnik T, Miklavcic D (2014) Electroporation-based technologies for medicine: principles, applications, and challenges. Annu Rev Biomed Eng 16:295–320 Zakotnik B, Budihna M, Smid L, Soba E, Strojan P, Fajdiga I, Zargi M, Oblak I, Lesnicar H (2007) Patterns of failure in patients with locally advanced head and neck cancer treated postoperatively with irradiation or concomitant irradiation with Mitomycin C and Bleomycin. Int J Radiat Oncol Biol Phys 67(3):685–690

Combined Treatment of Electrochemotherapy with Immunomodulators Vesna Todorovic and Maja Cemazar

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ECT-Induced Cell Death and Immune Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Immunomodulatory Molecules, Networks, and Pathways of the Immune Response . . . . . . . . . . . Immunomodulation in Cancer Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Immunomodulators Used for Stimulation of Immune Response with ECT . . . . . . . . . . . . . . . . . . . . Interleukin 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tumor Necrosis Factor α . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interferon α . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CpG Oligodeoxynucleotides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ipilimumab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gene Electrotransfer of Plasmids Coding Immunomodulatory Molecules . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 4 5 6 6 8 9 9 10 12 13 14 14

Abstract

Electrochemotherapy is a nonthermal tumor ablation modality. The local delivery of electric pulses directly to the tumor transiently increases permeabilization of the cell membrane and significantly increases the uptake of chemotherapeutic drugs into cells. Several mechanisms, including active immune response, contribute to the antitumor effectiveness of electrochemotherapy. After

V. Todorovic (*) Department of Experimental Oncology, Institute of Oncology Ljubljana, Ljubljana, Slovenia e-mail: [email protected] M. Cemazar (*) Department of Experimental Oncology, Institute of Oncology Ljubljana, Ljubljana, Slovenia Faculty of Health Sciences, University of Primorska, Izola, Slovenia e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_62-1

1

2

V. Todorovic and M. Cemazar

electrochemotherapy, the tumor-associated antigens are released in situ; however, they are not presented to the cells of the immune system efficiently to stimulate also a distant antitumor response. Although electrochemotherapy alone can stimulate the immune system, the induced immune response is not sufficient to also obtain a systemic effect. Combining electrochemotherapy with adjuvant immunomodulatory molecules is a promising approach to stimulate the immune system more effectively and thus increase antitumor effectiveness of electrochemotherapy. This approach has been evaluated in preclinical studies on experimental and spontaneous tumor models. Electrochemotherapy can be combined with different immunomodulatory molecules, including cytokines, such as interleukin-2 and interleukin-12, CpG oligodeoxynucleotides, monoclonal antibodies, and plasmids coding immunomodulatory molecules. Results of the preclinical studies combining electrochemotherapy with immunomodulatory molecules are promising. In addition to potentiation of local antitumor effectiveness, a systemic response can be obtained which affects also distant untreated tumors, namely, reducing the number of metastasis and cures of untreated tumors. The induced antitumor response is tumor specific, generating tumor-specific cytotoxic T cells and protection against the development of the same tumor type. Based on these results, combination of electrochemotherapy and immunomodulatory molecules bears a great potential for cancer therapy, specifically by adding a systemic component to localized electrochemotherapy treatment. Keywords

Cytokines • Electrochemotherapy • Electroporation • Immune response • Immunomodulation • Immunomodulators

Introduction Electrochemotherapy (ECT) is a nonthermal tumor ablation modality. It is based on the local delivery of electric pulses combined with specific chemotherapeutic drugs directly to the target tissue. Application of electric pulses affects permeability of the cell membrane. This transient change in cell membrane permeability allows non-permeant or low-permeant anticancer drugs to enter the cell. In this way, more anticancer drug molecules enter cells, resulting in higher intracellular anticancer drug concentration and augmented cytotoxicity of anticancer drugs. The most suitable candidates for use in a combination with electric pulses are hydrophilic drugs which lack or have a limited transport system in the cell membrane. Many anticancer drugs were evaluated for potential use in combination with electroporation, including 5-fluorouracil, actinomycin D, bleomycin, carboplatin, cisplatin, cyclophosphamide, daunorubicin, doxorubicin, etoposide, gemcitabine, mitomycin C, paclitaxel, vinblastine, and vincristine (Miklavcic et al. 2014). However, a significant cytotoxicity potentiation up to several 1000-fold was observed only for bleomycin (Orlowski et al. 1988; Gehl et al. 1998; Jaroszeski et al. 2000)

Combined Treatment of Electrochemotherapy with Immunomodulators

3

and approximately 70-fold for cisplatin (Sersa et al. 1995); therefore, they were designated as the most suitable anticancer drugs for clinical use of ECT. Several mechanisms contributed to the effectiveness of ECT. Application of controlled electric pulses transiently increases cell membrane permeability and results in increased intracellular drug concentrations and cytotoxicity. Application of controlled electric pulses transiently reduces blood perfusion in tumors and causes a vascular lock; therefore, tumor cells are exposed to cytotoxic drug for a prolonged period of time. Electrochemotherapy causes cell death of not only tumor but also endothelial cells and thus generates a vascular disruption effect. In addition, the immune system is activated by tumor antigen shedding into the bloodstream after ECT and is also involved in the antitumor effectiveness of ECT. The mechanisms associated with ECT antitumor effectiveness include increased drug cytotoxicity, vascular disruptive effect, and vasoconstriction and immune response which are discussed in more details in chapter “▶ Preclinical Studies on Reversible Electroporation.”

ECT-Induced Cell Death and Immune Response ECT with bleomycin or cisplatin induces massive cell death in situ within 24 h after treatment. Different types of cell death, such as apoptosis, necrosis, mitotic cell death, or immunogenic cell death, can be observed after ECT, depending on the anticancer drug used, its concentrations, the parameters of electric pulses, and the overall metabolic condition of the cell. After ECT, dying cells and tumor-associated antigens remain in situ, where they can be presented to the cells of the immune system. After ECT-induced cell death, different cells of the immune system are recruited to the tumor site, including potential antigen-presenting cells, different subsets of dendritic cells, and mononuclear leukocytes (Mekid et al. 2003; Roux et al. 2008; Gerlini et al. 2013). The immune response after ECT is discussed in more detail in chapter “▶ Immune Response After Electroporation and Electrochemotherapy.” The extent of involvement of the host’s immune system in response to antitumor therapy depends on tumor immunogenicity and immunocompetence of the host. Tumors differ in their ability to elicit immune response, which is essentially mediated by functional T cells. This was confirmed by evaluation of antitumor response on immunocompetent and immunodeficient mice bearing the same tumor type, where tumor cures were obtained only in immunocompetent mice. The effect of ECT on the treated tumors is usually manifested as a local delay in tumor growth and tumor cures; however, there is no effective systemic response. In addition, the tumor-associated antigens released after ECT do not seem to be presented efficiently to the cells of the immune system as no significant distant antitumor effect can be observed. In an attempt to further increase local responses and obtain also a systemic antitumor effectiveness, ECT can be combined with immunomodulatory molecules as adjuvant treatment.

4

V. Todorovic and M. Cemazar

Immunomodulatory Molecules, Networks, and Pathways of the Immune Response The immune system is a complex and dynamic network of organs, cells, molecules, and pathways working together in a tightly regulated and interconnected fashion to provide an efficient defense mechanism against harmful or possibly harmful substances and microorganisms. Organs of the immune system, such as the bone marrow and thymus, regulate the development of immune cells, while other organs, such as the spleen and lymph nodes, coordinate the encounter of antigens with antigen-specific lymphocytes and their development into effector and memory cells. Cells of the immune system originate from a single cell type, the hematopoietic stem cell, which differentiates into two precursor cell types: myeloid progenitor cell, which in the process of hematopoiesis gives rise to different types of mature myeloid cells (i.e., mast cells, phagocytes, granulocytes, natural killer cells), and lymphoid progenitor cell, which differentiates into lymphoid cells (i.e., B cells, T cells, and natural killer T cells). These different cell types are involved in innate and adaptive immune response. Innate response is a nonspecific defense system and responds very quickly to the presence of an antigen. It forms the first line of defense and includes specific cell types, such as phagocytes, dendritic cells, and natural killer cells. Pathogens are recognized by pattern recognition receptors through specific molecular patterns, such as unmethylated CpG motifs. On the other hand, adaptive immune response is normally quiescent and gets activated by the presence of specific antigens that are presented to the cells of adaptive response by antigen-presenting cells, such as dendritic cells, which are a member of phagocytes of the innate immune response. Cells of the adaptive immune response adapt to the specific antigen through activation of potent mechanisms for eliminating the specific antigen. The adaptive immune response can be mediated by T cells (cell-mediated immunity composed of three types of cells: helper, cytotoxic, and suppressor T cells) or by antibodies produced by B cells (humoral immunity). In comparison to innate response, adaptive response takes more time to develop. The combined action of innate and adaptive response is needed to provide efficient defense mechanism. Components of the immune system communicate through elaborate networks of chemical messengers, i.e., cytokines. Cytokines are principally produced by helper T cells and macrophages, but can also be transiently induced and secreted by all nucleated cells. Cytokines regulate the immune response by binding to their corresponding receptors on target cells and initiation of a cascade of intracellular events, stimulating or inhibiting activation, proliferation, and differentiation of various cells. Cytokines act in an autocrine, paracrine, or endocrine fashion. They affect different cells in a pleiotropic, redundant, synergistic, antagonistic, or cascade action. An effective and functional immune system is maintained by a delicate balance of cytokines, which is often disrupted in cancer.

Combined Treatment of Electrochemotherapy with Immunomodulators

5

Immunomodulation in Cancer Therapy Immunomodulation is a treatment aimed at modulating the activity of the immune system. By immunomodulation, the immune system activity can be either induced or enhanced (immunostimulation) or inhibited (immunosuppression). The objective of immunomodulation in cancer therapy is to modulate the existing immune response toward the desired antitumor effect that is clearance of tumor cells by the immune system. The immune system has a dual role in cancer development and progression. It can protect against and promote tumor growth through a combination of tumorinhibiting and tumor-enhancing processes mediated by the immune system. The so-called cancer immunoediting consists of three distinct and sequential phases, elimination, equilibrium, and escape (3Es of cancer immunology). During protection against tumor cells, in the elimination phase, both innate and adaptive immunities work together and are responsible for recognizing and eliminating the vast majority of mutated and possibly harmful cells, including tumor cells. The mechanism involved in the identification and eradication of tumor cells is called immunosurveillance. In the second phase, equilibrium, the rare tumor cell clone, which is not destroyed by the immune system, persists; however, the outgrowth of the tumors is prevented by the action of adaptive immunity. T cells, IL-12, and IFN-γ are required to maintain tumor dormancy. In this phase, editing of tumor immunogenicity occurs, which can result in control of the outgrowth of the cancer for the lifetime, or due to the constant pressure of the immune system, other clones of tumor cells may evolve. These evolved (selected) tumor cell clones are not visible to the immune system due to the loss of antigens, or they become insensitive to immune effector mechanism or they are able to induce a tumor-suppressive microenvironment. In each of these cases, the tumor cells enter into the escape phase, in which the progression of cancer is not controlled (blocked) by the immunity. The onset of these clinically apparent tumors is clearly an indication that individual tumor cells managed to avoid detection by the immune system or were able to limit the extent of immunological killing, thereby evading eradication by the immune system (Schreiber et al. 2011). The ability of cancer cells to evade elimination by the immune system is now also accepted as one of the emerging hallmarks of cancer. Overcoming regulation of the immune system by tumor cells and restoring immune system activity to normal can improve the management of cancer disease. Immunomodulators are a group of diverse agents modulating the activity of the immune system. They can be either immunostimulatory or immunosuppressive. In cancer treatment, the general approach for immunomodulating treatment is to enhance the antitumor response through stimulation of different cells of the immune system. This can be achieved through the addition of different stimulatory cytokines (i.e., interferon α, tumor necrosis factor α, interleukin 2, interleukin 12) and activation of T cells and/or antigen-presenting dendritic cells. In addition, stimulation of antitumor immunity can be achieved through the promotion of antigen presentation function of dendritic cells or production of protective T cell response. Recently, immune therapies overcoming immunosuppression that exists in the tumor showed great antitumor effectiveness and are currently under rapid development.

6

V. Todorovic and M. Cemazar

Immunomodulators Used for Stimulation of Immune Response with ECT The concept of combining ECT with immunomodulatory molecules is discussed in more detail in chapter “▶ Adjuvant Immunotherapy as a Tool to Boost Effectiveness of Electrochemotherapy.” Briefly, combining ECT with immunomodulatory molecules takes advantage of tumor destruction by ECT and exposure of different tumorassociated antigens in order to increase tumor immunogenicity. Albeit ECT alone can stimulate the immune system and at least in part induce an immune response, it is not sufficient to achieve a strong immunogenic environment allowing efficient priming and cross-presentation of tumor-associated antigens to the immune system. Introduction of immunomodulatory molecules in the form of recombinant proteins, oligonucleotides, monoclonal antibodies, or plasmids coding immunomodulatory molecules can be applied in combination with ECT to stimulate the immune system more actively and possibly achieve a systemic response. The high local effectiveness of ECT can be mostly contributed to the direct cell killing due to increased cytotoxicity of chemotherapeutic agents. However, cell killing mediated by the cells of the activated immune system is also involved and could be further exploited to increase the antitumor effectiveness of ECT not just locally but also systemically. Namely, ECT treatment in solid tumors can be limited by insufficient drug distribution due to irregular vascular organization in tumors, high interstitial pressure, and composition of the extracellular matrix confining drug diffusion. In addition, the inhomogeneity of electric field distribution, leading to areas of tumors covered with electric field below the threshold value for permeabilization, can also limit the effectiveness of ECT. Combining local ECT with different immunomodulatory molecules in a tightly regulated and controlled fashion can improve the overall antitumor effectiveness and obtain also a noticeable therapeutic effect on distant untreated metastasis.

Interleukin 2 Interleukin 2 (IL-2) is a well-characterized cytokine produced by and secreted from activated helper and cytotoxic T cells, natural killer cells, and natural killer T cells. IL-2 is the most widely used drug in immunotherapy of cancer (Andersen et al. 2003). IL-2 has no direct effects on cancer cells but rather mediates its antitumor activity by altering host immune reactions. IL-2 exerts a myriad of pleiotropic effects on different target cells. IL-2 acts as a growth factor and stimulates proliferation and activation of natural killer cells and increases their cytolytic capacity, induces proliferation and differentiation of T cells, and is involved in the expansion of immunosuppressive regulatory T cells. In addition, IL-2 can also affect cellular metabolism and glycolysis. IL-2 is a powerful biological response modifier and has been extensively used for modulation of antitumor immune responses. IL-2 has been combined with ECT either indirectly through the addition of IL-2 secreting cells or directly as a

Combined Treatment of Electrochemotherapy with Immunomodulators

7

recombinant protein. The first reports of ECT in a combination with IL-2 immunomodulatory therapy utilized specific histoincompatible cells engineered to produce and secrete IL-2 (Mir et al. 1995; Ramirez et al. 1998). Preestablished murine LPB sarcoma were treated with ECT with intravenous bleomycin application followed by peritumoral injection of histoincompatible IL-2-secreting cells. As a result of this combined therapy, 90 % of the tumors were cured. Furthermore, this combined therapy also led to the cures of more than half of the contralateral untreated tumors. A systemic antitumor immune response was established through the generation of helper and cytotoxic T cells and provided protection also against new tumor development (Mir et al. 1995). Tumors treated with IL-2-secreting cells alone demonstrated a slower growth and increased generation of cytotoxic T cells; however, no cures were observed. The induced antitumor response is tumor specific since there was no effect on distant untreated tumors of different origin indicating a generation of tumor antigen-specific cytotoxic T cells. This approach was evaluated also on larger VX2 carcinoma in rabbit livers. IL-2secreting cells were injected intratumorally after ECT with intravenous bleomycin application. A combination of ECT and IL-2-secreting cells resulted in increased regression of the primary liver tumor and largely reduced the number of visceral metastases with no massive metastatic spreading in the lungs (Ramirez et al. 1998). Tumors injected with IL-2-secreting cells only demonstrated a slow tumor evolution, indicating that IL-2 alone was able to activate an immune response, however not sufficiently to completely inhibit tumor growth. Taken together, the addition of IL-2secreting cells after ECT resulted in increased local antitumor effectiveness and exerted also a systemic effect as manifested by the reduced number of visceral and lung metastases (Ramirez et al. 1998) and protection from rechallenge with tumor cells (Mir et al. 1995). Another approach combining ECT and IL-2 immunomodulatory therapy was based on the addition of low-dose recombinant IL-2. Nodules of advanced malignant melanoma were first treated with ECT with intratumoral bleomycin application followed by peritumoral injections of recombinant IL-2. During and after the combined ECT and IL-2 treatment, specific tumor-associated antigen cytotoxic T lymphocyte responses were monitored. In all patients, a decreased amount of tumorspecific cytotoxic T cells in the blood was observed during treatment. When IL-2 administration was interrupted, the cytotoxic T cells reappeared in the blood. Further studies revealed that the recurring cytotoxic T cell responses were due to the induction of new reactive cytotoxic T cells and not to the return of the same cytotoxic T cells into the circulation (Andersen et al. 2003). In addition, the decreased amount of antitumor cytotoxic T cells in the blood can be attributed to the migration of antitumor cytotoxic T cells to the tumor area where these cytotoxic T cells exert their action. Taken together, this data supports the hypothesis that the extensive tumor cell death elicited by ECT may have attracted dendritic cells that captured the newly exposed tumor antigens and initiated cytotoxic T cell responses against them.

8

V. Todorovic and M. Cemazar

Tumor Necrosis Factor a Tumor necrosis factor α (TNF-α) is a well-known proinflammatory cytokine with a wide range of biological functions. TNF-α is produced and secreted by activated macrophages, monocytes, neutrophils, natural killer cells, and T cells. It affects growth, differentiation, survival, and function of different cells, including cells outside the immune system. TNF-α is an interesting immunomodulatory agent for adjuvant cancer therapy. It is weakly cytotoxic to most tumor cells but affects cells of the immune system and has an antivascular action in tumors due to its specificity for endothelial cells of tumor blood vessels. In preclinical studies, TNF-α has been combined with ECT with suboptimal bleomycin or cisplatin concentration to evaluate its potential to improve antitumor effectiveness of ECT. After combining TNF-α injected either intratumorally or peritumorally with ECT with bleomycin, increased antitumor effectiveness on SA-1 sarcoma was observed (Sersa et al. 1997). Similarly, significant tumor growth delays and tumor cures were observed also when SA-1 sarcoma was treated with combination of ECT with intravenous cisplatin application and adjuvant intratumoral TNF-α therapy (Cemazar et al. 2015). Since TNF-α is not directly cytotoxic to tumor cells, the observed tumor growth delays, prolonged survival times, and tumor cures are attributed to the vascular action of TNF-α. When TNF-α is applied in a combination with ECT, the observed synergistic effects are attributed to the dual role of TNF-α in the potentiation of antitumor effectiveness of ECT. On the one hand, early TNF-α effects are manifested in the first 24 h posttreatment and are aimed directly at the tumor vasculature, affecting the number of functional tumor blood vessels, reducing interstitial pressure, and increasing intratumoral concentration of chemotherapeutic agents (Fig. 1). On the other hand, TNF-α-related late effects are immunomodulatory and add a systemic component to

Fig. 1 Immunomodulation of ECT-treated tumor with TNF-α. Addition of TNF-α to ECT (c) potentiates antitumor effectiveness of ECT and results in increased vascular permeability and activates different immune cells in comparison to untreated (a) or ECT-treated tumor (b)

Combined Treatment of Electrochemotherapy with Immunomodulators

9

the local ECT treatment. Namely, TNF-α has a relevant and critical role at multiple levels of T cell immune response, such as activation of antigen-presenting cells, T cell priming, proliferation, function, and recruitment. Most importantly, TNF-α is involved in promotion of T cell proliferation to cross-presented tumor antigens. The use of TNF-α is specifically significant for ECT of large tumors where a limited drug delivery and/or penetration of chemotherapeutic drug within tumors is expected. It is of key importance to apply TNF-α intra- or peritumorally to evade its systemic toxicity. The potential of combined ECT and TNF-α immunomodulatory therapy lies also in its role in T cell response, specifically adding a systemic component to the localized ECT.

Interferon a Interferon α (IFN-α) is an immunostimulatory cytokine. IFN-α stimulates natural killer cells and induces expression of major histocompatibility complex class I antigens. IFN-α is approved by the FDA as adjuvant treatment for patients with high-risk melanoma. IFN-α polarizes immune response of helper T cells toward Th1, enhances cytotoxicity and survival of natural killer cells, induces generation and survival of cytotoxic T cells and memory CD8-positive T cells, positively regulates antibody production, and promotes dendritic cell maturation, chemotaxis, and CD8-positive T cell priming against tumor antigens. In addition, IFN-α has direct antitumor activity. It upregulates major histocompatibility complex class I surface molecules, promotes caspase-dependent apoptosis, and has antiangiogenic effects on tumor vasculature. Preclinical studies combining ECT with IFN-α have not been performed; however, the effect of adjuvant IFN-α prior to ECT was evaluated in a retrospective study on 5 patients, eligible for inclusion in the study (Hribernik et al. 2016). All patients had a primary melanoma surgically removed and had a high risk of recurrence; therefore, they received adjuvant IFN-α therapy postsurgery. Recurrent metastases were effectively treated with ECT. Taken together, the observed response rate of ECT after adjuvant IFN-α therapy on melanoma metastases was higher than reported for malignant melanoma treatment with ECT alone (Mali et al. 2013). This difference is possible due to the application of IFN-α prior to ECT and its stimulation of the immune system. However, a progressed disease was observed at different time intervals after ECT treatment. Although the study of adjuvant IFN-α therapy prior to ECT was small, a potential for combining ECT with immunostimulatory therapies was shown.

CpG Oligodeoxynucleotides Oligodeoxynucleotides containing unmethylated CpG motifs (CpG ODNs) are wellcharacterized immunostimulatory molecules. The unmethylated CpG motifs can be found in high amounts in bacterial and viral DNA. These CpG motifs are recognized

10

V. Todorovic and M. Cemazar

by the vertebrate immune system through pattern recognition receptors as foreign, and a protective immune response is triggered. Specifically, CpG ODNs are recognized by endosomal toll-like receptor 9 (TLR-9) which is in turn activated and triggers a signaling cascade which subsequently leads to the transcription of proinflammatory chemokines and cytokines. This response can be achieved also with synthetic CpG ODNs which can be exploited as vaccine adjuvants and immunotherapeutics for cancer treatment. ECT in combination with CpG ODNs has been evaluated in three different murine tumor models, namely, LPB sarcoma, B16F10 melanoma, and B16-OVA melanoma (Roux et al. 2008). CpG ODNs have been selected as immunomodulatory agents based on the increased mRNA expression level of its corresponding receptor, TLR-9, after ECT alone. In all three tumor models, the combination of ECT with immunomodulatory CpG ODNs efficiently reduced local tumor growth of the treated tumors and resulted also in a systemic antitumor effect in the distant untreated contralateral tumors. Specific immune memory was induced, as the mice were protected against a rechallenge with the same tumor cells, but not with unrelated tumor cells. The combined ECT and CpG ODNs treatment induced tumor-specific T-cell effectors in the tumor-draining lymph nodes and in the spleen. Significantly more IFN-γ was secreted upon activation than with ECT or CPG ODN alone (Roux et al. 2008). A synergistic interaction between ECT and CpG ODNs was observed.

Ipilimumab Ipilimumab is a fully humanized monoclonal antibody, approved by the FDA for the treatment of advanced melanoma as monotherapy or as adjuvant therapy and is directed against cytotoxic T lymphocyte antigen-4 (CTLA-4). CTLA-4 is one of the immune checkpoint molecules and is a negative regulator of T lymphocyte activation and proliferation. CTLA-4 is present on activated T cells and binds to costimulatory molecules CD80 or CD86 expressed on antigen-presenting cells. This CTLA-4 ligation to CD80 or CD86 receptors inhibits activation and expansion of T cells, controls immune response, and attenuates chances of autoimmune inflammation. Inhibiting the interaction between CTLA-4 and its ligands by blocking antibodies terminates negative regulation of T cell activation (Fig. 2). The role of CTLA-4 in cancer development, specifically in melanoma, is to exploit this inhibitory mechanism to prevent clearance of cancer cells by the immune system. The rationale for using ipilimumab as immunomodulatory agent lies in blocking the inhibitory interaction of CTLA-4 with its B7 receptor complex on antigen-presenting cells and thus augmenting T cell activation and proliferation. Combination of ECT and ipilimumab has not been tested in preclinical studies; however, the combination of ipilimumab as immunomodulatory agent and ECT has been evaluated in a pilot study involving ten patients with advanced melanoma (Mozzillo et al. 2015). An improved overall survival was observed after a combination of ipilimumab followed by ECT with systemic bleomycin in comparison to ECT treatment alone in patients with advanced melanoma. Further immunological

Combined Treatment of Electrochemotherapy with Immunomodulators

11

Fig. 2 Immunomodulation of ECT-treated tumor with addition of ipilimumab. Ligation of CTLA4 to its corresponding receptors CD80 or CD86 inhibits activation and expansion of T cells (a). Ipilimumab binds to CTLA-4 and blocks its interaction with CD80 or CD86 (b). Interaction of CD28 with receptors CD80 or CD86 leads to activation and proliferation of T cells

analysis revealed that the amount of regulatory T cells was reduced in patients responding to the treatment and was significantly lower than in the non-responding patients (Mozzillo et al. 2015). In addition to the abovementioned pilot study, a case of advanced melanoma treated with a combination of ECT and

12

V. Todorovic and M. Cemazar

ipilimumab was reported. A patient with advanced melanoma in-transit metastasis was first treated with ECT with systemic bleomycin application and had a partial response. After the last ECT session, multiple visceral metastases were discovered and ipilimumab treatment was administered. After ipilimumab treatment, a complete regression of visceral and cutaneous metastases was observed with loss of pigmentation at the site of cutaneous metastatic lesions (Brizio et al. 2015).

Gene Electrotransfer of Plasmids Coding Immunomodulatory Molecules One of the challenges when using recombinant immunomodulatory proteins is to achieve safe concentrations for effective immunotherapy. Often multiple applications of recombinant immunomodulatory proteins are needed to achieve long-term stimulation of the immune system. However, repetitive applications of immunomodulatory molecules can also cause systemic toxicity. With the progress of gene therapy, this challenge was successfully overcome, and long-term levels of immunomodulatory molecules can now be sustained in the nontoxic concentration range. Gene electrotransfer (GET) is a nonviral method for the delivery of genetic material into cells based on the application of electric pulses. The concept of GET is discussed in more detail in chapters “▶ Clinical Applications of Gene Therapy” and “▶ Principles of Gene Electrotransfer.” GET of plasmids constructed to carry a code for different immunostimulatory molecules, such as interleukin 12 (IL-12), IL-2, and granulocyte-monocyte colony-stimulating factor, has been used for immunomodulation in a combination with ECT in preclinical studies. IL-12 is a promising immunomodulatory molecule for use in a combination with ECT mostly due to its antitumor and antimetastatic activity. It is produced by macrophages and dendritic cells and acts as a stimulatory factor for natural killer cells and as maturation factor for cytotoxic T cells. Through induced proliferation of natural killer cells and cytotoxic T cells, IL-12 induces production of IFN-γ, increases activity of cytotoxic T cells, and drives the differentiation of naïve helper T cells to Th1 cells. GET of IL-12 plasmids has been tested in combination with ECT with either bleomycin or cisplatin on different murine tumors, such as head and neck carcinomas, fibrosarcoma, and mammary carcinomas. Intratumoral coapplication of bleomycin and IL-12 plasmid followed by electroporation effectively reduced the growth of preestablished tumors, resulted in tumor cures, prevented metastatic progression and tumor redevelopment through increased IL-12 and IFN-γ expression in the tumors, and amplified cytotoxic T lymphocyte activity (Torrero et al. 2006). Intramuscular GET of IL-12 plasmid in combination with ECT with cisplatin potentiated specific tumor growth delay and increased tumor cure rate in immunogenic SA-1 sarcoma. In less immunogenic TS/A carcinoma, the observed tumor growth delay was similar in immunocompetent and immunodeficient mice; however, tumor cures were obtained only in immunocompetent mice. Clearly, an involvement of functional immune system is required to achieve the desired

Combined Treatment of Electrochemotherapy with Immunomodulators

13

antitumor effect (Sedlar et al. 2012). The observed antitumor effect of adjuvant IL-12 GET can be attributed to the observed systemic release of IL-12 and the concomitant increase of IFN-γ after intramuscular (Tevz et al. 2009), intratumoral (Pavlin et al. 2011), or peritumoral GET (Pavlin et al. 2009). In addition, combination of ECT and GET of IL-12 plasmid has been successfully used in veterinary medicine for the treatment of spontaneous canine tumors of different histologies, including head and neck tumors, mast cell tumors, melanoma, fibrosarcoma, and ameloblastoma (Cutrera et al. 2008, 2015; Reed et al. 2010; Cemazar et al. 2016). Similar to murine experimental tumor models, intratumoral or peritumoral GET of IL-12 plasmid in combination with ECT had a high antitumor efficacy on histologically different spontaneous tumors. In addition, IL-12 contributed also to the prevention of tumor recurrence or distant metastases through the activation of antitumor immune response (Reed et al. 2010; Cutrera et al. 2015; Cemazar et al. 2016). In conclusion, the combination of ECT and GET of IL-12 plasmid can be used on histologically different tumors and can induce also a systemic response, and the treatment is safe and effective. Granulocyte-monocyte colony-stimulating factor (GM-CSF) is an important hematopoietic cytokine. It can be secreted by different cells of the immune system, such as T cells and macrophages. GM-CSF is involved in the production of granulocytes and monocytes and in the maturation of macrophages and dendritic cells. GM-CSF is used in immunotherapy of melanoma as monotherapy or in combination with other approaches. It is specifically an interesting immunomodulatory cytokine for cancer therapy as it is able to increase recruitment of dendritic cells at the tumor site and enhance their ability to capture tumor-associated antigens. GET of plasmid-coding GM-CSF has been tested in experimental murine B16 melanoma alone or in combination with ECT. GET of GM-CSF plasmid alone had no significant effect on tumor growth. However, in combination with ECT, in addition to improved survival by the combined therapy, long-term immunity to recurrence and resistance to challenge were induced in up to 12.5 % of mice (Heller et al. 2000). Similarly, peritumoral GET of IL-2 plasmid in combination with ECT significantly increased tumor doubling time. Tumor-free mice after peritumoral GET of IL-2 plasmid in combination with ECT were also protected from rechallenge (Heller et al. 2000).

Conclusions Based on the abovementioned preclinical and pilot clinical studies evaluating the combination of ECT with different immunomodulatory molecules, such as cytokines (IL-2, IL-12, TNF-α, IFN-α, GM-CSF), oligonucleotides with CpG motifs, and antibodies targeting immune system’s cells (ipilimumab), the local antitumor effect of ECT can be potentiated by the addition of immunomodulatory molecules to a different extent. However, it is evident that in order to achieve a substantial and longterm local and systemic effect, the application of immunomodulatory molecules in combination with ECT has to be accurately planned. Specifically, preclinical studies

14

V. Todorovic and M. Cemazar

on temporal and spatial application of immunomodulatory molecules in combination with ECT, which will lead to an optimized immunomodulatory and long-term antitumor effect, are needed. Based on the knowledge gathered from preclinical studies, clinical studies should be carefully planned and performed to achieve the desired fine-tuned immunomodulatory response and long-term cancer control in cancer patients.

Cross-References ▶ Adjuvant Immunotherapy as a Tool to Boost Effectiveness Electrochemotherapy ▶ Clinical Applications of Gene Therapy ▶ Immune Response After Electroporation and Electrochemotherapy ▶ Preclinical Studies on Reversible Electroporation ▶ Principles of Gene Electrotransfer

of

References Andersen MH, Gehl J, Reker S et al (2003) Dynamic changes of specific T cell responses to melanoma correlate with IL-2 administration. Semin Cancer Biol 13:449–459. doi:10.1016/j. semcancer.2003.09.009 Brizio M, Fava P, Astrua C et al (2015) Complete regression of melanoma skin metastases after electrochemotherapy plus ipilimumab treatment: an unusual clinical presentation. Eur J Dermatol 25:271–272. doi:10.1684/ejd.2015.2522 Cemazar M, Todorovic V, Scancar J et al (2015) Adjuvant TNF-α therapy to electrochemotherapy with intravenous cisplatin in murine sarcoma exerts synergistic antitumor effectiveness. Radiol Oncol 49:32–40. doi:10.1515/raon-2015-0005 Cemazar M, Ambrozic Avgustin J, Pavlin D, et al (2016) Efficacy and safety of electrochemotherapy combined with peritumoral IL-12 gene electrotransfer of canine mast cell tumours. Vet Comp Oncol 1–14. doi: 10.1111/vco.12208 Cutrera J, King G, Jones P et al (2015) Safe and effective treatment of spontaneous neoplasms with interleukin 12 electro-chemo-gene therapy. J Cell Mol Med 19:664–675. doi:10.1111/ jcmm.12382 Cutrera J, Torrero M, Shiomitsu K et al (2008) Intratumoral bleomycin and IL-12 electrochemogenetherapy for treating head and neck tumors in dogs. Methods Mol Biol 423:319–325. doi:10.1007/978-1-59745-194-9_24 Gehl J, Skovsgaard T, Mir LM (1998) Enhancement of cytotoxicity by electropermeabilization: an improved method for screening drugs. Anticancer Drugs 9:319–325. doi:10.1097/00001813199804000-00005 Gerlini G, Sestini S, Di Gennaro P et al (2013) Dendritic cells recruitment in melanoma metastasis treated by electrochemotherapy. Clin Exp Metastasis 30:37–45. doi:10.1007/s10585-012-9505-1 Heller L, Pottinger C, Jaroszeski MJ et al (2000) In vivo electroporation of plasmids encoding GM-CSF or interleukin-2 into existing B16 melanomas combined with electrochemotherapy induces long-term antitumour immunity. Melanoma Res 10:577–583. doi:10.1097/00008390200012000-00010 Hribernik A, Cemazar M, Sersa G et al (2016) Effectiveness of electrochemotherapy after IFN-α adjuvant therapy of melanoma patients. Radiol Oncol 50:21–27. doi:10.1515/raon-2015-0048

Combined Treatment of Electrochemotherapy with Immunomodulators

15

Jaroszeski MJ, Dang V, Pottinger C et al (2000) Toxicity of anticancer agents mediated by electroporation in vitro. Anticancer Drugs 11:201–208. doi:10.1097/00001813-20000300000008 Mali B, Miklavcic D, Campana LG et al (2013) Tumor size and effectiveness of electrochemotherapy. Radiol Oncol 47:32–41. doi:10.2478/raon-2013-0002 Mekid H, Tounekti O, Spatz A et al (2003) In vivo evolution of tumour cells after the generation of double-strand DNA breaks. Br J Cancer 88:1763–1771. doi:10.1038/sj.bjc.6600959 Miklavcic D, Mali B, Kos B et al (2014) Electrochemotherapy: from the drawing board into medical practice. Biomed Eng Online 13:29. doi:10.1186/1475-925X-13-29 Mir LM, Roth C, Orlowski S et al (1995) Systemic antitumor effects of electrochemotherapy combined with histoincompatible cells secreting interleukin-2. J Immunother 17:30–38 Mozzillo N, Simeone E, Benedetto L et al (2015) Assessing a novel immuno-oncology-based combination therapy: Ipilimumab plus electrochemotherapy. Oncoimmunology 4:e1008842. doi:10.1080/2162402X.2015.1008842 Orlowski S, Belehradek J, Paoletti C, Mir LM (1988) Transient electropermeabilization of cells in culture. Increase of the cytotoxicity of anticancer drugs. Biochem Pharmacol 37:4727–4733. doi:10.1016/0006-2952(88)90344-9 Pavlin D, Cemazar M, Kamensek U et al (2009) Local and systemic antitumor effect of intratumoral and peritumoral IL-12 electrogene therapy on murine sarcoma. Cancer Biol Ther 8:2114–2122. doi:10.4161/cbt.8.22.9734 Pavlin D, Cemazar M, Coer A et al (2011) Electrogene therapy with interleukin-12 in canine mast cell tumors. Radiol Oncol 45:30–39. doi:10.2478/v10019-010-0041-9 Ramirez LH, Orlowski S, An D et al (1998) Electrochemotherapy on liver tumours in rabbits. Br J Cancer 77:2104–2111 Reed SD, Fulmer A, Buckholz J et al (2010) Bleomycin/interleukin-12 electrochemogene therapy for treating naturally occurring spontaneous neoplasms in dogs. Cancer Gene Ther 17:457–464. doi:10.1126/scisignal.2001449.Engineering Roux S, Bernat C, Al-Sakere B et al (2008) Tumor destruction using electrochemotherapy followed by CpG oligodeoxynucleotide injection induces distant tumor responses. Cancer Immunol Immunother 57:1291–1300. doi:10.1007/s00262-008-0462-0 Schreiber RD, Old LJ, Smyth MJ (2011) Cancer Immunoediting: Integrating Immunity’s Roles in Cancer Suppression and Promotion. Science 331:1565–1570. doi:10.1126/science.1203486 Sedlar A, Dolinsek T, Markelc B et al (2012) Potentiation of electrochemotherapy by intramuscular IL-12 gene electrotransfer in murine sarcoma and carcinoma with different immunogenicity. Radiol Oncol 46:302–311. doi:10.2478/v10019-012-0044-9 Sersa G, Cemazar M, Miklavcic D (1995) Antitumor effectiveness of electrochemotherapy with cis-diamminedichloroplatinum(II) in mice. Cancer Res 55:3450–3455 Sersa G, Cemazar M, Menart V et al (1997) Antitumor effectiveness of electrochemotherapy is increased by TNF-a on SA-1 tumors in mice. Cancer Lett 116:85–92 Tevz G, Kranjc S, Cemazar M et al (2009) Controlled systemic release of interleukin-12 after gene electrotransfer to muscle for cancer gene therapy alone or in combination with ionizing radiation in murine sarcomas. J Gene Med 11:1125–1137. doi:10.1002/jgm Torrero MN, Henk WG, Li S (2006) Regression of high-grade malignancy in mice by bleomycin and interleukin-12 electrochemogenetherapy. Clin Cancer Res 12:257–263. doi:10.1158/10780432.CCR-05-1514

Combining Electrolysis and Electroporation for Tissue Ablation Mary Phillips Ho

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fundamental Principles of Electrolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanism of Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pathology Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Advantages and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fundamental Principles of Electroporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanism of Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pathology Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Advantages and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Combination of Electrolysis with Electroporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Presence of Electrolytic Phenomenon During Traditional Electroporation . . . . . . . . . . . . . . . . . . Combining Electroporation and Electrolysis: Proposed Mechanism of Increased Tissue Ablation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods for Applying Synergistic Electroporation and Electrolysis (SEE) for Tissue Ablation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pathology Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 4 4 6 6 6 6 8 8 9 9 11 11 17 19 19 20 20

Abstract

Electrolysis and electroporation technologies have been utilized to provide several valuable nonthermal tissue ablation modalities. The synergistic combination of electroporation and electrolysis (SEE) has produced a new method of tissue ablation that has distinct advantages over electrolysis or electroporation alone. Electrolysis uses a low-magnitude direct electric current to create chemical species at the electrode-tissue interface which then diffuse through the tissue, M.P. Ho (*) Department of Engineering, Quinnipiac University, Hamden, CT, USA e-mail: [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_63-1

1

2

M.P. Ho

resulting in extreme pH changes and cell death. Electroporation, on the other hand, is used to create permeabilizations in the cell membrane that can be used to induce cell death by several different mechanisms: through electrochemotherapy, cytotoxic drugs are introduced to the cell interior, and irreversible electroporation results in cell death by loss of cell homeostasis. When electrolysis is combined with electroporation, a new mode of tissue ablation is achieved that results in a very effective method of cell death. This mechanism of action is likely due to the ability of the electrolytic products to penetrate the cell membrane through the permeabilizations created by electroporation. Here, fundamental principles of electrolysis and electroporation are presented, the mechanism of ablation by SEE is discussed, and different types of SEE protocols are examined with respect to their effect on the tissue. This chapter hopes to serve as a foundation and starting point for further research into this ablation modality as well as for the development of new types of SEE protocols that may be used to address specific clinical needs.

Keywords

Synergistic combination of electroporation and electrolysis • Electroporation • Reversible electroporation • Irreversible electroporation • Electrolytic ablation • Tissue ablation

Introduction Electroporation is a technique which employs pulsed electric fields to create permeabilized cell membranes. This phenomenon occurs when an electric field is applied across the cell, destabilizing the electric potential maintained by the cell membrane and resulting in the formation of nanoscale defects in the lipid bilayer. Reversible pore formation caused by reversible electroporation (RE) has been utilized as an effective method for introducing molecules such as genes and drugs into the cell while maintaining the cell viability. Changing the electrical parameters can, however, result in cell death by irreversible electroporation (IRE) due to extensive loss of metabolites and loss of homeostasis. Electrical parameters for IRE can be employed such that any Joule heating effects are minimized, resulting in a nonthermal tissue ablation modality that spares important tissue components such as major blood vessels, nerves, and the extracellular matrix. Electrochemotherapy (ECT) uses reversible electroporation protocols to produce temporary pores in the cell membrane, allowing cancer drugs such as bleomycin to be introduced directly into the tumor cells. Protocols have been developed for treatment of metastases to liver, lung, brain, bone, and head and neck cancers. ECT is increasingly becoming integrated with other treatments, guidelines for treatment are beginning to list ECT as an option, and in some locations, ECT is becoming available for treatment on a hospital level. Nonthermal irreversible electroporation (NTIRE) for

Combining Electrolysis and Electroporation for Tissue Ablation

3

tumor ablation has also made great strides and is currently seen as a promising nonthermal, minimally invasive tissue ablation modality. The high voltages employed with irreversible electroporation may induce muscle contractions, requiring general anesthesia and neuroblocking agents to mitigate these effects during the procedure. Such spasms may in turn cause treatment electrodes to shift in location, resulting in incomplete ablation in the target site or damage to nearby vital structures. Managing these effects adds a significant level of complexity to the surgical procedure, resulting in increased surgical time and cost of implementing this type of treatment method. Reversible electroporation does not require as high of voltages and/or number of pulses as irreversible electroporation and thus does not have the same issues with muscle contractions seen in irreversible electroporation. RE has been used with success by combining it with cytotoxic drugs, resulting in tissue ablation by electrochemotherapy. Nonetheless, there would be a significant advantage in being able to obtain tissue ablation without the requirement of introducing the cytotoxic drugs into the body as required for electrochemotherapy. Tissue ablation by electrolysis (E) (also known as electrochemical therapy) uses a direct electric current to create tumor ablation due to the creation of chemical species at the electrode-tissue interface as well as extreme pH changes throughout the ablation volume. Electrolytic treatment has been reported to be a simple, safe, and low-cost technique for ablating solid tumors. At the electrode-anode interface with the tissue, electrochemical reactions consist mainly of water decomposition and chloride oxidation. At the cathode surface, hydroxide ions and hydrogen are produced from water decomposition. Chemical species generated at the interfaces of the electrodes diffuse away into the tissue in a process driven by differences in electrochemical potential. These electrolytic reactions yield changes in pH, resulting in an acidic region near the anode and a basic region near the cathode. The changes in pH as well as the presence of the chemical species formed during the electrolytic reactions result in a cytotoxic environment. Electrolytic ablation requires very low direct currents (tens of mA) and very low voltages (single to low tens of volts), thus requiring less expensive instrumentation than needed for tissue ablation by irreversible electroporation. However, for traditional electrolytic ablation, a high level of electrolytic products must diffuse through the tissue to create a cellular environment that is cytotoxic enough to produce cell death. This requires a longer application time (tens of minutes to hours) for the electrical current, increasing the length of the surgical procedure. When electrolysis is combined with electroporation, a new method of tissue ablation is achieved. It is hypothesized that the pores formed during electroporation allow the electrolytic products to pass through the permeabilized cell membrane and contribute to cell death at low levels of exposure to electrolytic chemical species. This technique combines several advantages of electroporation and electrolysis alone and can enable different modes of tissue ablation. For example, when reversible electroporation is combined with electrolysis, it applies the low current and low voltage required to produce electrolysis along with the high speed of treatment time associated with electroporation. In addition, since the electrolytic products are at low

4

M.P. Ho

concentrations and only penetrate the electroporated cells, this new method allows for accurate treatment planning and tight margins between the ablated and untreated tissue zones. The primary goal of this chapter is to supply readers with the understanding necessary to develop synergistic electroporation and electrolysis (SEE) protocols for use in tissue ablation. Fundamental principles of electrolysis and electroporation when applied separately are first presented. The proposed mechanism of ablation by combining electrolysis with electroporation is discussed, and different types of SEE protocols are examined with respect to their effect on the tissue. With this background and knowledge of how SEE can be applied toward tissue ablation, it is hoped that the reader can gain a basic understanding of this new field and apply this foundation toward investigating and developing new SEE protocols.

Fundamental Principles of Electrolysis Mechanism of Action Electrolysis is an electrochemical process that takes place at the electrode surface and occurs during the passage of an electric current between two electrodes immersed in an ionic solution. New chemical species generated at the interface of the electrodes diffuse away from the electrodes into the tissue. If the anode is electrically soluble, metal dissolution will be a major part of the resulting chemical reaction occurring at the electrodes. The use of passivated metal electrodes can make this effect negligible. During electrolysis, protons are produced at the anode, according to the following reaction: 2H2 O Ð O2 þ 4Hþ þ 4e

(1)

The major chemical reaction occurring at the cathode is the decomposition of water into molecular hydrogen and hydroxyl ions: 2H2 O þ 2e Ð H2 þ 2OH

(2)

Sodium chloride is ionized into Na+ and Cl ions, which move toward the cathode and the anode, respectively. The concentration of Cl ions is higher at the anode, and additional reactions result in the production of chlorine gas: 2Cl Ð Cl2 þ 2e

(3)

The chlorine gas produced can go on to react with water to produce additional chemical substances: Cl2 þ H2 O Ð HClO þ Hþ þ Cl These chemical reactions are illustrated in Fig. 1.

(4)

Combining Electrolysis and Electroporation for Tissue Ablation

5

Fig. 1 Essential chemical reactions occurring at the anode and cathode during the electrolytic process. NaCl sodium chloride, NaOH sodium hydroxide, Cl2 chlorine (gas), HCl hydrochloric acid, HClO hypochlorous acid, H2O water, OH hydroxide ion, O2 oxygen (gas), H2 hydrogen (gas) (Reprinted from Gravante et al. (2011), copyright 2009, with permission from Elsevier)

Species produced at the anode and cathode are transported to the surrounding tissue due to diffusion from a concentration gradient and migration due to the potential gradient. Negatively charged ions become attracted to the anode, and positively charged ions move toward the cathode. At the anode, destructive species may include metal ions, hydrogen ions, and various chemical species containing oxygen and chlorine (Nilsson et al. 2000). Smaller, more mobile, and abundant ions initially alter the tissue’s local chemistry. Reactions at the anode attract protons (H+), resulting in the formation of an acidic region. At the cathode, hydroxyl ions and molecular hydrogen act as destructive reaction products (Nilsson et al. 2000). Hydrogen is given off as a gas, and a more alkaline region forms at the cathode due to electrochemical reactions and a repulsion of protons (Finch et al. 2002a). Electrolytic cell death is driven by the presence of a cytotoxic environment that develops due to local changes in pH as well as the presence of the new chemical species formed during electrolysis. Cell necrosis occurs when a pH of less than 6 or greater than 9 is reached (Finch et al. 2002b). The lethal effects of electrolysis on the tissue is a function of the concentration of electrochemical products produced and the time of exposure to these products. The area of necrosis has been shown to depend on the coulomb dosage applied (Colombo et al. 2007; Czymek et al. 2011), although this dependency may be contingent upon the experimental conditions.

6

M.P. Ho

Pathology Findings Electrolytic ablation is characterized by a lesion that begins at the electrodes and propagates outward with increased delivered charge, as expected for a process that is driven by diffusion mechanisms. The resulting lesion has an irregular shape, with a greater volume affected near the tissue surfaces that were in contact with the electrodes. The cell morphology is different at the cathode side of the lesion in comparison to the anode side. Around the anode, the lobular architecture tends to be retained, and the cells become more amorphous and eosinophilic, whereas, at the cathode side, the tissue is more disrupted (Finch et al. 2002a). This dissimilar appearance is caused by different products of electrolysis near the cathode and the anode, resulting in different ablative reactions. Near the anode, the pH is low, and many of the toxic chemical species are related to the various pH-dependent components of Cl. Near the cathode, on the other hand, the pH is high and the toxicity is primarily caused by various components of OH. In addition, electroosmotic migration of water from the anode to the cathode results in desiccation near the anode and edema at the anode. It has been observed during electrolysis that the resulting lesion near the anode surface is larger than that near the cathode. This may be attributed to the formation of chlorine at the anode, which acts as a more destructive chemical than those produced at the cathode (Finch et al. 2002a).

Advantages and Limitations Electrolysis has been harnessed for tissue ablation in medicine since the early 1800s (Amory 1886). During the last two decades, this field has seen substantial research advances. From an operational standpoint, electrolysis requires very low voltages and currents, providing advantages relative to other ablation techniques. Low voltages can help avoid muscle contractions during application, and these low voltage and current parameters reduce the need for instrument complexity. One major limitation of electrolytic ablation, however, is that it is a lengthy procedure (requiring current applications for tens of minutes to hours). The time scale for application is controlled by the process of diffusion of the toxic chemical species from the electrodes into the tissue. High concentrations of electrically produced ablative chemical species are required, thus limiting the speed of the procedure.

Fundamental Principles of Electroporation Mechanism of Action Electroporation is characterized by the permeabilization of the cell membrane’s lipid bilayer through the application of very brief (nanosecond to millisecond), high-field (in the range of MV/m) electric pulses across the cell. The effects of electroporation depend on the magnitude and duration of the pulsed electric field as well as other

Combining Electrolysis and Electroporation for Tissue Ablation

7

factors such as cell size and shape and number of electrical pulses applied. The electric field magnitude triggers pore formation, whereas electric field strength and/or pulses with longer duration and/or number affect the pore expansion process (Rems and Miklavcic 2016). Though a comprehensive theory has not yet been developed to fully explain the mechanism of electroporation, extensive experimental work and proposed models have developed a strong foundation that has allowed for the development of electroporation for a wide array of applications and is currently being built upon to produce a more detailed understanding of the phenomenon. The essential features of the electroporation process are known and can be briefly summed up as follows: 1. Electroporation utilizes short (on the order of ns to ms) electrical pulses, applying an elevated transmembrane potential. The cell membrane requires a certain charging time before the transmembrane potential reaches a critical threshold. This charging time depends on electrical properties of the cell and solution as well as cell size and shape (Rems and Miklavcic 2016) and may vary from an order of 100 ns to several microseconds. 2. The membrane conductivity increases immediately, and a time-dependent membrane transition occurs as long as the externally applied electric field is held at over above the critical value. 3. Depending on the electrical parameters utilized, once the external electric field is removed, either membrane stabilization and resealing occur for reversible electroporation or loss of cell homeostasis leads to cell death by irreversible electroporation. For reversible electroporation, once the electric field is lowered below the critical value, a stabilization process occurs over a few microseconds. The transmembrane potential drops quickly to near zero, and the membrane dramatically recovers to a level in which it is permeable to only small molecules. The membrane reseals slowly over seconds or even minutes. For irreversible electroporation, cell death occurs via a number of potential mechanisms such as continued pore growth and membrane rupture, membrane rupture due to colloidal-osmotic swelling, changes in ionic concentrations, and loss of cellular content. This process of pore formation is illustrated in Fig. 2, based on the theory of aqueous pore formation. The family of electrical pulses that cause electroporation are divided into three types: in reversible electroporation (RE), the cells survive the permeabilization process, nonthermal irreversible electroporation (NTIRE) results in cell death due to the lipid bilayer destabilization and permeabilization, and irreversible electroporation coupled with thermal effects results when sufficiently strong fields are applied, causing a temperature increase sufficiently high enough for thermal damage to occur. Electroporation is becoming extensively utilized in biotechnology and medicine. In the reversible mode, electroporation has become a central technology for cell manipulation, and, in combination with chemicals, fundamental research and clinical trials have demonstrated its promise for gene therapy (Lambricht et al. 2016), and it is used clinically to target cancerous tissue through electrochemotherapy.

8

M.P. Ho

Fig. 2 Membrane pore formation during electroporation. An idealized schematic of the lipid bilayer membrane of electroporation based on the aqueous pore formation theory. (a) Thermal fluctuations occur in the lipid bilayer membrane. (b) After the membrane has charged to threshold value, a hydrophobic pore is created, allowing water molecules to start penetrating the bilayer. (c) The lipids start reorienting to form a more stable hydrophilic pore, enabling more water and other polar molecules and ions to pass through the membrane. (d) The pore will shrink and eventually reseal after the external electric field has been removed. For irreversible electroporation, extensive transport through the cell membrane and slow pore resealing result in cell death

Irreversible electroporation has been harnessed to ablate tissue while retaining the structural integrity of blood vessels, nerves, and extracellular matrix (Jourabchi et al. 2014). The ablative modality of NTIRE has been shown to result in a quicker recovery of the biological tissues (Phillips et al. 2012). The ability to apply NTIRE in a minimally invasive manner and the safety of this procedure have led to a recent surge in its clinical use.

Pathology Findings NTIRE results in focused apoptotic cell death and the creation of sharp boundaries between the ablated cells and the adjacent, normal tissue. NTIRE has been shown to spare the surrounding extracellular tissue structure. Blood vessels, bile ducts, the urethra, and nerves are left intact and continue to function normally (Jourabchi et al. 2014). It is believed that these structures remain undamaged because their high collagenous connective tissue and elastic fiber content lack a cellular membrane that could be targeted by NTIRE. This preserving feature could also be due to the gap junction found within the cellular structure, allowing the electric current produced by NTIRE to travel through the gap junctions without affecting the cell membrane (Jourabchi et al. 2014).

Advantages and Limitations A major advantage of tissue ablation by electroporation is the relative speed of the procedure in comparison to any other ablation technique. In addition, because the procedure targets the cell membrane, critical features of the extracellular matrix are

Combining Electrolysis and Electroporation for Tissue Ablation

9

spared. However, irreversible electroporation usually employs up to one hundred pulses of microsecond length and electric fields up to the several kV/cm range which produce muscle contractions and require the use of paralyzing drugs and may affect the electrical function of the heart. Reversible electroporation for electrochemotherapy employs lower electric fields in the range of 300 V/cm to 1 kV/cm, typically applied over eight pulses. Tissue ablation by RE, however, requires the injection of chemo toxic drugs, such as bleomycin or cisplatin. Finding a way to harness electroporation for tissue ablation without the need for additional drugs, including total-body paralysis agents, would provide great strides in simplifying the use of electroporation for medical applications and increase its ability for use within developing countries.

The Combination of Electrolysis with Electroporation Presence of Electrolytic Phenomenon During Traditional Electroporation Electroporation pulses in tissue have been shown to produce some level of electrolysis at the electrodes, generating products of electrolysis that could ablate cells. This phenomenon has been demonstrated through agar gel models (Turjanski et al. 2011), ex vivo experimental work on excised tissue (Maglietti et al. 2013), and in vivo and mathematical models (Olaiz et al. 2014). The resulting cytotoxic environment could prove detrimental to some reversible electroporation processes that depend on the cell to stay alive. For example, cell damage during gene electrotransfer would decrease the overall efficiency of the procedure. Thus, initial efforts sought to reduce and minimize this effect (Olaiz et al. 2014). Electrolytic effects have been demonstrated to contribute to cell death in applications that were previously attributed solely to irreversible electroporation. Rubinsky et al. (2015) demonstrated that the cytotoxic effects previously attributed to IRE could actually be caused by at least three different mechanisms: reversible electroporation combined with electrolysis, irreversible electroporation combined with electrolysis, or irreversible electroporation without electrolysis. These different mechanisms of cell death may result in distinct effects on the cell and tissue ablation. A schematic to summarize which mode may become relevant, based on electric field magnitude and electric current delivery time, is illustrated in Fig. 3. The presence of electrolysis during electroporation procedures may also explain a phenomenon that has been observed clinically. During clinical use of NTIRE as well as occasionally during the clinical use of reversible electroporation, loud explosionlike sounds may be produced. These noises may be caused by electric breakdown across gases near the electrodes produced by electrolysis. During NTIRE studies, a bright spot is often produced near the electrodes on ultrasound that increases in size with the number of applied pulses. A bright hyperechoic appearance has also been noted in the region adjacent and between the electrodes when performing electrolysis experiments under the guide of ultrasound (Stehling et al. 2016). As noted by

10

M.P. Ho

Fig. 3 Tissue ablation domains as a function of electric field and exposure time. E indicates electrolysis ablation; R + E represents a combination of reversible electroporation and electrolysis ablation; IRE + E represents a combination of irreversible electroporation and electrolysis (Reprinted from Rubinsky et al. (2015) # the Author(s) 2015, permission from Sage Publishing)

Stehling et al. (2016), it is likely that this bright spot in both cases is caused by the presence of gases produced due to electrolytic reactions at the electrodes. The ultrasonic waves are reflected from the interface between the tissue and gases, resulting in the observed hyperechoic appearance. Guenther et al. (2015) demonstrated that the increase in electrolytic gases around the electrode, as observed on ultrasound, corresponds to an increased electric discharge. This electric discharge occurs primarily at the cathode and follows a pattern typical of electrical breakdown across ionized gases at atmospheric pressures. The electrolytic products produced at the cathode contribute toward the electrons at discharge. When the electric field across the gas layer increases above the breakdown voltage, electric discharge occurs, accompanied by light from the ionized gas and producing the violent sound often observed (Guenther et al. 2015). The electric discharge phenomenon generates high-pressure waves and could be detrimental to the organ being treated (Stehling et al. 2016). A further understanding of the presence of electrolysis during electroporation may help to devise electroporation protocols that minimize this electric breakdown effect.

Combining Electrolysis and Electroporation for Tissue Ablation

11

Overall, these studies indicate that some level of electrolysis may occur in all electroporation protocols and must be considered and taken into account. However, the combination of electrolysis and electroporation is not a detrimental in all cases and indeed may be harnessed as new modality for tissue ablation that is seen to meet many of the shortcomings of electroporation or electrolysis alone.

Combining Electroporation and Electrolysis: Proposed Mechanism of Increased Tissue Ablation It has been proposed that combining the cell permeabilization effect of electroporation with the effects of the electrolytic products generated from electrostimulation could result in a new ablation modality with an efficiency that is dramatically increased compared to electrolysis alone. Though additional studies are warranted to study this phenomenon, one explanation for this effect is that the electrolytically produced chemicals can pass through the pores formed in the cell membrane during electroporation, as illustrated in Fig. 4. This interaction could be leveraged to induce cell death using a lower applied electric charge than required for traditional electrolytic ablation. This mechanism of tissue ablation could be harnessed through a variety of methods. New methods that provide support for the proposed mechanism of action and take advantage of this combined effect are being developed. Recent investigation into new synergistic electroporation and electrolysis (SEE) protocols is described in the next section.

Methods for Applying Synergistic Electroporation and Electrolysis (SEE) for Tissue Ablation New methods and protocols are currently being developed to investigate the effect of SEE and develop methods for tissue ablation that may be used clinically. Though this field is new and much more work is needed for understanding the mechanisms at play and optimizing treatment parameters for clinical use, it is beneficial to look at different methods that have been used thus far to illustrate the SEE effect. These experimental results give new insight into the technology and help develop a foundation for future work. The types of SEE parameters that have been applied experimentally can be divided into four different groups, each of which illustrates a unique way to achieve tissue ablation by enabling electrolytically produced chemical products’ access to the cell interior through electroporation. Each type of SEE protocols may be more applicable for specific applications. 1. 2. 3. 4.

SEE by reversible electroporation and low-charge electrolysis in series SEE by irreversible electroporation and low-charge electrolysis in series SEE applied through multiple reversible electroporation-type pulses SEE achieved through the use of a single decay pulse

12

M.P. Ho

Fig. 4 Proposed mechanisms of combining electrolysis with electroporation. Low levels of electrolysis may not be strong enough to cause cell death on their own, and reversible electroporation on its own causes cell to become temporarily permeabilized followed by recovery. Combining these two phenomena, however, would allow the electrolytic products to pass through the permeabilized cell and cause cell death at much lower concentrations

Each of these methods illustrates a unique way to achieve tissue ablation by enabling electrolytically produced chemical products’ access to the cell interior through electroporation. Overall, they help to gain a stronger foundation of understanding around the technology that can be used for further development in this field.

SEE by Reversible Electroporation and Low-Dose Electrolysis in Series An ablation protocol that combines reversible electroporation with low-charge electrolysis would allow electrolytic products’ direct access to the cell interior, requiring a much lower electric field than irreversible electroporation, reducing the treatment time and level of electrolytic products required during electrolysis, and avoiding the need for drug injection as practiced with electrochemotherapy. The synergistic electroporation and electrolysis effect has been demonstrated as a new tissue ablation modality by combing reversible electroporation and low-dose electrolysis together in series. This was first proven using a simple parallel-plate electrode configuration to apply electrical parameters across the rat liver in vivo (Fig. 5f) (Phillips et al. 2015a). Though the experimental setup is less clinically relevant, the simple experimental design allows for results that effectively illustrate the result of SEE. Hematoxylin and eosin (H&E) cross sections of the treated livers are shown in Fig. 5. Parameters of typical electrolytic ablation (8 mA applied for 60 s) resulted in transverse ablation across the liver thickness, as expected (Fig. 5a). A set of low-charge electrolysis parameters were developed by reducing the current and applied time (4 mA, 30 s), resulting in a decreased charge delivered to the tissue and a much smaller ablation volume, centered around the electrode surfaces (Fig. 5b). Reversible electroporation parameters (electric field of 250 V/cm, 100 μs pulse length, and 8 pulses applied at a 1 Hz frequency) were also shown to result in minimal tissue ablation, with the volume of ablation located near the electrode surfaces only (Fig. 5c). When the low-charge electrolysis was combined with

Combining Electrolysis and Electroporation for Tissue Ablation

13

Fig. 5 Demonstration of the effectiveness of combining electrolysis and electroporation for increased ablation volume. Electrical parameters are applied across the liver lobe thickness using plate electrodes (f). The anode is adjacent to the top surface of the liver and the cathode is adjacent to the bottom surface. H&E staining is used to visualize the volume of ablation, and the region of ablation is outlined using dashed lines. Typical parameters used for electrolysis (E) result in transverse ablation across the liver cross section (a), delivering a total charge per unit area of 1.24 C/cm2. Decreasing the delivered charge to 0.31 C/cm2 results in non-transverse electrolytic ablation located near the electrode surfaces (b). Typical parameters used for reversible electroporation (RE) deliver a charge per unit area of 0.0001 C/cm2 and result in some minimal tissue ablation near the electrode surfaces due to electrolytic effects (c). When electrolysis and reversible electroporation are combined in series, however, transverse ablation is achieved (d, e)

reversible electroporation, however, transverse tissue ablation was achieved (Fig. 5d, e). During electrolytic ablation, the delivered charge can be used as a quantitative measure of the amount of products of electrolysis generated at the electrodes. Since these products of electrolysis propagate inward into the tissue through diffusion, the extent of tissue ablation depends on the amount of charge delivered at the electrodes as well as the distance of that point from the electrodes. Thus, when comparing SEE protocol results with those of electrolysis and reversible electroporation alone, the delivered charge can be used as a measure of the electrolytic products produced. The delivered charge by reversible electroporation alone was orders of magnitude lower than either electrolysis protocol and thus should have an infinitesimal effect on pure electrolysis. Nonetheless, the results shown in Fig. 5 show that a substantial increase in tissue volume when electrolysis and reversible electroporation are combined. This hypothesis has been further supported by the use of a more clinically relevant experimental setup, using titanium needle electrodes to apply the electrical parameters to an in vivo porcine liver model (Stehling et al. 2016). When the reversible electroporation parameters were followed by the low-charge electrolysis protocol, the ablated area significantly increased. The treated volume was increased

14

M.P. Ho

even further by applying reversible electroporation parameters, followed by the low-charge electrolysis protocol, followed by another set of reversible electroporation parameters. The mechanisms at play likely involve the use of conventional electrolysis as the central tissue ablation modality while utilizing electroporation pulses to permeabilize the cell membrane, allowing for cell death to occur at lower amounts of electrolytic products.

SEE Applied as a Combination of Electrolysis and Irreversible Electroporation The combined use of electrolysis and electroporation has the potential application in extending the ablated tissue volume in conventional NTIRE without increasing the voltage on the electrodes. This would take advantage of the region that develops outside of the irreversible electric field zone in which a lower electric field, typical of reversible electroporation protocols, is present. The presence of even low-level electrolytic products in this region could enable the ablation zone to be extended to the edge of the reversible electroporation threshold of the electric field. Survival of the tissue between electrodes is a critical concern during the use of NTIRE clinically, and developing methods to increase the ablation zone would be greatly beneficial to the field. SEE has been shown to increase this ablation zone (Stehling et al. 2016), using an NTIRE-electrolysis-NTIRE treatment sequence. These results fit the overall hypothesis of SEE. When NTIRE treatment parameters are applied by needle electrodes, the strength of the electric field decreases with distance from the electrodes. Beyond the volume of tissue treated by irreversible electroporation lies a region where electric fields are still strong enough to produce reversible electroporation effects (Fig. 6). It is expected that by combining electrolysis with NTIRE, ablation can now be achieved within the reversible electroporation region via electrolytically produced products. SEE Applied Through Multiple Reversible Electroporation-Type Pulses Reversible electroporation on its own can produce a limited electrolytic ablation effect, typically confined to the tissue volume immediately adjacent to the electrode surfaces. A third type of SEE parameters uses a single protocol of multiple reversible electroporation-type pulses to increase the electrolytic effect, enabling tissue ablation by applied electric fields below the threshold for irreversible electroporation (Phillips et al. 2015b). By using delivered charge as a comparative measure, the multiple-pulse SEE protocol can be designed to produce a much more effective method of electrolytic ablation. The SEE protocol is able to produce damage comparable to that of conventional electrolytic protocols while delivering orders of magnitude less charge to the target tissue over much shorter periods of time. This effect is illustrated in Fig. 7, using a simple parallel-plate electrode configuration (same as shown in Fig. 5f) to apply the electrical parameters to liver tissue. The ablation zone achieved by applying SEE through multiple low-voltage electroporation pulses extends throughout the reversible electroporation zone surrounding electrodes when using a clinically relevant two-needle electrode setup

Combining Electrolysis and Electroporation for Tissue Ablation

15

Fig. 6 A mathematical model of two 1-mm-diameter needle electrodes in a volume of tissue. Regions that experience an electric field magnitude over a given threshold (region within inner contour line) experience irreversible electroporation, resulting in tissue ablation. Reversible electroporation is achieved for lower electric fields (region between outer and inner contour line). Results were obtained for an 800-μs, 1331-V pulse. By inducing cell death within the reversible electroporation zone, the volume of treated tissue can be increased (Reprinted from Davalos et al. (2005), # 2005, with permission of Springer)

(Stehling et al. 2016). This further supports the mechanistic explanation for how SEE increases ablation efficiency and can be used to increase the treated volume. In Fig. 8, calculated isoelectric field lines are superimposed on histology images of the ablated zone. When a low number of pulses are applied, the ablation zone extends to electric field lines greater than 500 V/cm corresponding to ablation by IRE and suggesting that the combined effect of electrolysis and electroporation did not occur. However, when the number of pulses is increased, increasing the total charge applied to the tissue, the region of ablation extends to electric field lines of approximately 200 V/cm, indicating that cell death is occurring in the surrounding reversible electroporation zone due to the SEE effect. A possible application of this type of SEE parameters is a method of tissue ablation that uses reversible electroporation strength fields to permeabilize the cell membrane to products of electrolysis, thus inducing cell death. These SEE parameters result in cell death without requiring the high voltages of irreversible electroporation while relying on a drastically decreased level of electrolytic product production than required for traditional electrolytic ablation.

16

M.P. Ho

Fig. 7 SEE delivered by multiple reversible electroporation-type pulses. Electrical parameters are applied across the liver lobe thickness using plate electrodes (Fig. 5f). The anode is adjacent to the top surface of the liver and the cathode is adjacent to the bottom surface. Masson trichrome staining is used to visualize the volume of ablation, and the region of ablation is outlined using dashed lines. Typical parameters used for electrolysis (E) illustrate the pattern of tissue ablation with increasing charge delivered (a–c). Reversible electroporation-type parameters were chosen to increase the charge delivered (d–f), resulting in a pattern of ablation that is markedly similar to that obtained by typical electrolysis. The volume of ablation achieved with SEE, however, occurs at orders of magnitude less delivered charge than required for typical electrolysis

A Single Exponential Decay Pulse for Achieving SEE Ablation A single exponential decay pulse is another type of SEE parameter that can produce ablation with the tissue. A discharge capacitor can deliver a pulse that decays exponentially and is defined by the initial voltage and the time constant. The applied voltage for the exponential decay pulse can be given as: t

V ðt Þ ¼ V o e τ

(5)

where Vo is the initial voltage, t is time, and τ is the time constant. A set of capacitor discharge parameters can be chosen such that the initial high voltage at the beginning of the pulse serves to generate electroporation, while the rapid decay toward a trailing low voltage generates sufficient charge for the generation of electrolytic products. When the exponential decay pulse delivers a relatively low level of charge, no tissue ablation is observed. Increasing the delivered charge by applying a higher initial voltage and/or longer time constant can be used to increase the ablation volume (Phillips et al. 2016). It should be noted, however, that continued increase in applied voltage will eventually lead to tissue ablation by irreversible electroporation effects rather than SEE. These single exponential decay SEE parameters indicate a method for tissue ablation that allows for reduced instrument complexity in comparison to that used by irreversible electroporation while requiring shorter application time than

Combining Electrolysis and Electroporation for Tissue Ablation

17

Fig. 8 Extent of tissue ablation achieved by eight electrochemotherapy-type pulses compared to a high number of electrochemotherapy-type pulses. Eight typical electrochemotherapy magnitude-type pulses (1000 V, 100 μs, 1 Hz) result in a minimal level of ablation, as shown by gross macroscopic section (a) and trichromatic staining (b). This is compared to increased level of ablation achieved by increasing the number of pulses to 297 (c) and (d). Calculated isoelectric field lines are superimposed on the bottom row, showing ablation extents to an electric field of approximately 200 V/cm. These isoelectric field lines are also applicable for the top row as well and would indicate that ablation extents to an electric field of greater than 500 V/cm for the 8-pulse protocol (Reprinted from Stehling et al. (2016) (# Stehling et al. 2016) licensed under CC BY 2.0)

traditional electrolysis. In addition, the single exponential decay pulse causes a substantial reduction in muscle contractions in comparison to the same level of ablation achieved by a typical multiple-pulse electroporation protocol. This may pave the way to the development of protocols that do not require the use of general anesthesia and neuroblocking agents during clinical use, an area which is currently seen as a major limitation of current irreversible electroporation protocols.

Pathology Findings Histological analysis of the lesion produced by SEE ablation reveals that multiple modes of cell death may be present, depending on the type of SEE parameters applied, the electrode configuration, and the location within the tissue relative to the electrodes. The mechanisms of damage can be divided into at least five different categories (Stehling et al. 2016):

18

1. 2. 3. 4. 5.

M.P. Ho

Dominant anodic ablation Dominant cathodic ablation Combination reversible electroporation and anodic compounds Combination reversible electroporation and cathodic compounds Irreversible electroporation

SEE ablation protocols result in a sharp transition between the untreated cells and the ablated region, marked by the presence of congestion of the sinusoids. The appearance of the ablated cells differs throughout the ablated region, and marked differences are apparent when comparing cell morphology near the cathode to that near the anode. The treated zone is larger near the electrodes and grows outward toward the center tissue, resulting in either two separate lesions or a single dumbbellshaped lesion. Large blood vessels remain intact and patent (Stehling et al. 2016), a feature that has been also observed for both electrolysis and electroporation when applied in separate protocols. When SEE was applied across the liver tissue using parallel-plate electrodes both by the method of reversible electroporation and low-charge electrolysis in series (Phillips et al. 2015a) and by the method of using multiple reversible electroporation-type pulses (Phillips et al. 2015b), histological analysis indicated that the mode of tissue ablation maintained key characteristics of electrolytic ablation. The appearance of the treated tissue at the anode, at the cathode, and in the core of the treated tissue was markedly different, caused by different products of electrolysis near the anode versus the cathode that result in different ablative reactions. The general mode of damage is consistent with an electrolysis mechanism of damage; the tissue ablation begins from the vicinity of the electrodes, is different at the anode side in comparison to the cathode side, and increases to encompass the entire tissue between the electrodes with an increased delivered charge. Thus, it is likely that dominant anodic ablation and cathodic ablation occurred immediately adjacent to the electrode surfaces. However, the ablation seen throughout the majority of the liver thickness in Figs. 5 and 6 is likely due to the third and fourth mechanisms of damage, combining reversible electroporation with either electrolysis products produced from the anode or from the cathode. These electrolytic products are able to penetrate the electroporated cells, resulting in electrolytic cell death even when only low levels of electrolytic product are present. An irreversible electroporation zone, however, can also be present, depending on the parameters used. When applying irreversible electroporation with electrolysis to increase the ablation volume, a volume of the tissue will experience histological effects that can be primarily attributed to cell death by NTIRE. An NTIREdominated ablation zone has also been shown to be present in histology analyzed from multiple low-voltage electroporation pulses applied through needle electrodes (Stehling et al. 2016). Histological analysis for a single decay pulse also demonstrates an ablation pattern similar to that achieved by electrolytic ablation. By increasing the applied charge, the volume of ablation also increases, starting at the electrode surfaces and moving in toward the center of the tissue volume. For the parallel-plate electrode

Combining Electrolysis and Electroporation for Tissue Ablation

19

experimental model described by Phillips et al. (2016), initial electric field of 1000 V/cm and 500 V/cm appears to apply a mechanism of tissue ablation that combines electroporation effects with electrolysis (damage modes 1–4). Increasing the electric field to 1500 V/cm, however, results in a mode of ablation is likely dominated by irreversible electroporation effects rather than the combination of electrolysis and electroporation.

Imaging Imaging methods for combined electrolysis and electroporation protocols may be developed based on imaging success seen when each ablation modality is applied separately. For example, ultrasound monitoring, magnetic resonance imaging, and non-contrast and contrast-enhanced CT have been used to assess the volume of ablation zone after electroporation (Yarmush et al. 2014). In addition, electroporation leads to an immediately detectable impedance decrease, and electrical impedance tomography is a method that is under development, providing a two-dimensional reconstruction of tissue impedance that allows for near-real-time monitoring of tissue electroporation (Yarmush et al. 2014). Local changes in pH can be used to reliably monitor the extent of tissue ablation achieved by electrolytic ablation (Finch et al. 2002b). Discrete pH-measuring devices, however, cannot provide a continuous, spatial image of the pH distribution, and thus an alternative method for real-time monitoring of lesion size is desired. Bubbles of gas produced during electrolysis have made the use of ultrasound monitoring of lesions difficult (Finch et al. 2002a). A technology that can detect pH changes in both time and space is directly applicable for research and clinical applications of electrolysis. Electrical impendence tomography has been shown as one possible mechanism for monitoring lesions created during electrolysis in real time (Meir and Rubinsky 2015). This method takes advantage of the relatively small size of proton ions and hydroxyl ions compared to other physiological ions produced during the electrochemical reaction. These smaller ions not only result in local pH changes but also lead to increased local mobility, which in turn results in increased conductivity around the anode and the cathode. This local change in tissue conductivity can be monitored in real time using principles of electrical impedance tomography. Magnetic resonance imaging (MRI) can also be used to image the process of electrolysis by detecting pH fronts (Meir et al. 2015), resulting in another promising method for monitoring and advancing the use of this technology. Future work will develop these imaging techniques for real-time monitoring of tissue ablation by SEE.

Conclusions The synergistic combination of electroporation and electrolysis can be used as a new tissue ablation modality, resulting in cell death at much lower levels of electrolytic product than required by traditional electrolysis alone. By combining electrolysis

20

M.P. Ho

and electroporation, a new method may be achieved that addresses some of the clinical drawbacks of each technology when applied individually. SEE may be applied through a variety of different types of ablation protocols. This new type of treatment method is still in the early stages of development, and substantial additional research and quantitative analysis are necessary to advance this field. Though initial studies appear to support the proposed mechanism of SEE ablation, further research is necessary to gain an in-depth knowledge of this phenomenon. There are various parameter configurations that can be chosen to apply SEE, and the method of choice may depend strongly on the individual application. These different types of SEE protocols need to be further examined and refined in order to choose protocols that are most relevant clinically. In addition, new types of SEE protocols may be developed that allow for increased ablative efficiency, reduced instrument complexity, or serve to meet other specific clinical needs.

Cross-References ▶ Current Density Imaging as Means to Follow Tissue Electroporation ▶ Electrolysis During Pulsed Electric Field Treatment ▶ Mass Transfer of Electrolytic Species During Electroporation ▶ Medical Imaging of Electroporation ▶ Modeling of Electrochemical Reactions During Pulsed Electric Field Treatment ▶ Principles and Use of Magnetic Resonance Electrical Impedance Tomography in Tissue Electroporation

References Amory R (1886) A treatise on electrolysis and its therapeutical and surgical treatment in disease. William Woof & Co., New York Colombo L, Gonzalez G, Marshall G, Molina FV, Soba A, Suarez C, Turjanski P (2007) Ion transport in tumors under electrochemical treatment: in vivo, in vitro and in silico modeling. Bioelectrochemistry 71:223–232 Czymek R, Dinter D, Loffler S, Gebhard M, Laubert T, Lubienski A, Bruch HP, Schmidt A (2011) Electrochemical treatment: an investigation of dose–response relationships using an isolated liver perfusion model. Saudi J Gastroenterol 17(5):335–342 Davalos RV, Mir LM, Rubinsky B (2005) Tissue ablation with irreversible electroporation. Ann Biomed Eng 33(2):223–231 Finch JG, Fosh BG, Maddern GJ (2002a) Direct current electrolysis for local ablation of liver metastases. In: Habid NA (ed) Multi-treatment modalities of liver tumours. Springer, Boston, pp 269–292. doi:10.1007/978-1-4615-0647-1_22 Finch JG, Fosh B, Anthony A, Slimani E, Texler M, Berry D, Dennison A, Maddern G (2002b) Liver electrolysis: pH can reliably monitor the extent of hepatic ablation in pigs. Clin Sci 102:389–395 Gravante G, Ong SL, Metcalfe MS, Bhardwaj N, Maddern GJ, Lloyd DM, Dennison AR (2011) Experimental application of electrolysis in treatment of liver and pancreatic tumours: principles, preclinical and clinical observations and future perspectives. Surg Oncol 20:106–120. doi:10.1016/j.suronc.2009.12.002

Combining Electrolysis and Electroporation for Tissue Ablation

21

Guenther E, Klein N, Mikus P, Stehling M, Rubinsky B (2015) Electrical breakdown in tissue electroporation. Biochem Biophys Res Commun 467:736–741 Jourabchi N, Beroukhim K, Tafti BA, Kee ST, Lee EW (2014) Irreversible electroporation (NanoKnife) in cancer treatment. Gastrointest Interv 3:8–18 Lambricht L, Lopes A, Kos S, Sersa G, Preat V, Vandermeulen G (2016) Clinical potential of electroporation for gene therapy and DNA vaccine delivery. Expert Opin Drug Deliv 13 (2):295–310. doi:10.1517/17425247.2016.1121990 Maglietti F, Michinski S, Olaiz N, Castro M, Suarez C, Marshall G (2013) The role of pH fronts in tissue electroporation based treatments. PLoS One 8(11):e80167 Meir A, Rubinsky B (2015) Electrical impedance tomography of electrolysis. PLoS One 1–16. doi:10.1371/journal.pone.0126332 Meir A, Hjouj M, Rubinsky L, Rubinsky B (2015) Magnetic resonance imaging of electrolysis. Sci Rep 5(8095):1–9. doi:10.1038/srep08095 Nilsson E, von Euler H, Berendson J, Thorne A, Wersall P, Naslund I, Lagerstedt AS, Narfstrom K, Olsson J (2000) Electrochemical treatment of tumours. Bioelectrochemistry 51:1–11 Olaiz N, Signori E, Maglietti F, Soba A, Suarez C, Turjanski P, Michinski S, Marshall G (2014) Tissue damage modeling in gene electrotransfer: the role of pH. Bioelectrochemistry 100:105–111 Phillips M, Raju N, Padath T, Rubinsky B (2012) Irreversible electroporation on the small intestine. Br J Cancer 106(3):490–495 Phillips M, Raju N, Rubinsky L, Rubinsky B (2015a) Modulating electrolytic tissue ablation with reversible electroporation pulses. Technology 3(1):1–9 Phillips M, Rubinsky L, Meir A, Raju N, Rubinsky B (2015b) Combining electrolysis and electroporation for tissue ablation. Technol Cancer Res Treat 14(4):395–410. doi:10.1177/ 1533034614560102 Phillips M, Krishnan H, Raju N, Rubinsky B (2016) Tissue ablation by a synergistic combination of electroporation and electrolysis delivered by a single pulse. Ann Biomed Eng. doi:10.1007/ s10439-106-1624-4, Published online May 4, 2016 Rems L, Miklavcic D (2016) Tutorial: electroporation of cells in complex materials and tissue. J Appl Phys 119:201101. doi:10.1063/1.4949264 (21 pages) Rubinsky L, Guenther E, Mikus P, Stehling M, Rubinsky B (2015) Electrolytic effects during tissue ablation by electroporation. Technol Cancer Res Treat 1–9. doi:10.1177/1533034615601549 Stehling MK, Guenther E, Mikus P, Klein N, Rubinsky L, Rubinsky B (2016) Synergistic combination of electrolysis and electroporation for tissue ablation. PLoS One 11(2): E0148317. doi:10.1371/JOURNAL.PONE.0148317 Turjanski P, Olaiz N, Maglietti F, Michinski S, Suarez C, Molina FV, Marshall G (2011) The role of pH fronts in reversible electroporation. PLoS One 6(4):e17303. doi:10.1371/journal. pone.0017303.t001 Yarmush ML, Goldberg A, Sersa G, Kotnik T, Miklavcic D (2014) Electroporation-based technologies for medicine: principles, applications, and challenges. Annu Rev Biomed Eng 16:295–320. doi:10.1146/annurev-bioeng-071813-104622

Electroporation in Scars/Wound Healing and Skin Response Laure Gibot and Alexander Golberg

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fundamentals of Wound Healing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Why Electroporation? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Skin Response to the Externally Delivered Pulsed Electric Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Skin Rejuvenation with Pulsed Electric Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gene Electrotransfer (GET) as a Tool in Wound-Healing Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrostimulation of Healing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrochemotherapy Stimulates Aesthetic and Functional Wound-Healing Process . . . . . . . . . . . Burn Wound Disinfection with Pulsed Electric Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications of Electroporation in Surgery: Mesh Disinfection and Scar Treatment . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 5 5 6 11 12 12 14 15 17 18 18

Abstract

Wounds are an essential part of the life cycle of organisms due to their interaction with environment, which can sometimes be very harsh. Although wound healing has puzzled humanity from early days and has gradually emerged from art and witchcraft to modern medical procedures, the detailed mechanisms underlying what results in normal and abnormal wound healing are not well established. This gap in knowledge exists because of the tremendous complexity of the overall wound-healing process. Despite this gap in our understanding of detailed L. Gibot (*) Institut de Pharmacologie et de Biologie Structurale, Université de Toulouse, CNRS, UPS, Toulouse, France e-mail: [email protected] A. Golberg (*) Porter School of Environmental Studies, Tel Aviv University, Tel Aviv, Israel e-mail: [email protected]; [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_64-1

1

2

L. Gibot and A. Golberg

mechanisms, new methods to treat wounds continue to appear, bringing promising new solutions to patients suffering from acute or chronic wounds. Because of its specific action on cell membrane without extensive heating and with no observed effects on the extracellular matrix, electroporation found application in multiple aspects of treating wounds. In this chapter, the authors explain the basics of what the wounds are and what is known about wound-healing process. The use of electroporation-based technologies for various applications for treating wounds, wound disinfection, gene electrotransfer for the treatment of diabetic chronic wounds, and electrostimulation of wound healing, is discussed in this chapter. In addition, the skin response to electroporation, especially with emerging applications of electroporation in skin rejuvenation and scars and keloid treatments, is discussed. Keywords

Electroporation • Wound healing • Human skin • Wound disinfection • Scar treatment • Rejuvenation

Introduction Wounds are an essential part of the life cycle of organisms due to their interaction with the harsh environment. The development of new tissues in place of a wound is a complex set of processes controlled by a myriad of internal and external factors. Understanding and exercising control over this overall set of processes could transform a huge segment of patient care. Although wound healing has puzzled humanity from early days, the detailed mechanisms underlying what results in normal and abnormal wound healing are not well established. This gap in knowledge exists because of the tremendous complexity of the overall wound-healing process with results from interactions among the fields of biology, chemistry, mechanics, and electrics all occurring in space and time. Despite this gap, new methods to treat wounds continue to appear, bringing promising new solutions to patients. In recent years, electroporation-based therapies have been suggested to improve various aspects of wound-healing process. Electroporation of skin is a unique process, when short, high-voltage pulsed electric fields induce temporary or permanent permeabilization of the tissue. This is a nonthermal, chemical-free method to increase permeability of skin tissues and skin individual cells to various molecules, through reversible electroporation pathway. It is also a method to control cell viability, through irreversible electroporation pathway. To achieve electroporation, a specific electric field strength is required. A precise control over electric field distribution in the skin through electrode positioning or conductivity manipulations provides a convenient tool for procedure management. This control, possible in time and space, is rarely achievable by other physical therapies. This chapter discusses the fundamentals of wound healing; the current challenges in the field and insights on how electroporation-based technologies can address these challenges. Electroporation-based applications described here

Electroporation in Scars/Wound Healing and Skin Response

3

could potentially lead to new devices and methods to improve patient care in this challenging field of medicine.

Fundamentals of Wound Healing A living system at any level, cell, tissue, organ, or organism, is defined by the boundary with the external environment. The disruption of this boundary by environmental hazards is a wound (Fig. 1). Once the injury takes place, all organisms are mobilized to close the wound and restore the boundary; otherwise, the bacterial contamination or the leakage of water and nutrients and imbalance of energetic expenditures will lead to organism death. A process of this boundary restoration between the cell, organ, or organism and its environment is known as a wound-healing process (Fig. 1). In mammals, wound healing is a dynamic, chronic process that is divided into four overlapping phases: hemostasis, inflammation, proliferation, and remodeling (Robson et al. 2001). During hemostasis, constriction of the damaged vessels and clot formation physically limit blood loss. During the inflammatory phase, leukocytes and then monocytes accumulate to combat infection in the wounded tissue. In this phase, multiple cytokines and growth factors are released to the wound bed and contribute to the fibroblast migration, differentiation, and activity. During the proliferative phase, fibroblasts deposit a new and specific extracellular matrix and differentiate into myofibroblasts. In the final, i.e., remodeling phase, reorganization of the closed wound environment occurs until repair is completed. Inflammation plays an important role in the organism response to the injury and in the control of all stages of the wound healing (Koh and DiPietro 2011; Fig. 2). To describe this complex dynamic process in tissues, Robson et al. (2001) introduced the concept of wound-healing trajectory (Fig. 3), which demonstrates the time-dependent cumulative effects of these multiple processes that occur from injury through healing. According to the healing trajectory curve, normally healed tissues are characterized by complete restoration of functions and structure (Fig. 3). In contrast, chronic wounds are characterized by incomplete restoration of structure and functions. In proliferative, hypertrophic scarring, however, the healing process does not stop as it should, and the tissue fails to reach a normal cell density and a balance between collagen deposition and degradation (Robson et al. 2001). Proliferative scarring and chronic wounds in human have adverse physical, aesthetic,

Fig. 1 Organism injuring pathways

4

L. Gibot and A. Golberg

Fig. 2 The pattern of leukocyte infiltration into the wound. In hemostasis the tissue is populated with rare mast cells and macrophages. After the injury, at the early inflammation period, neutrophils are rapidly recruited from the circulation. They are followed by macrophages and then lymphocytes and mast cells at the late inflammation period. At the resolution and remodeling phase, the wound is mostly populated by mast cells and rare macrophages again (Duffield et al. 2012) Fig. 3 Wound-healing trajectories (Figure adapted with permission from (Golberg et al. 2013). Copyright # 2013 Wiley Periodicals, Inc)

functional, psychological, and social consequences. The major aim of all woundhealing technologies and treatments is to restore completely the functions and structure of the wounded site to its initial condition.

Why Electroporation? In cases of abnormal healing, to reverse the healing trajectory to normal, there is an obvious need for intervention in cell metabolism, life cycle, and density. Electric fields provide a convenient method for intervention as there is no need in external

Electroporation in Scars/Wound Healing and Skin Response

5

chemicals, which by themselves can induce milliards of unknown by-reactions. Physical methods such as light and ultrasound have limited depth penetration properties in the skin; in contrast, electric field distribution is mostly predicated by the electrical properties of an organ. Once delivered to the cell membrane, externally applied electric fields can modulate the natural transmembrane potential and thus modify cell behavior. By modulating transmembrane potential of cells, calibrated pulsed electric field application leads to a local and transient cell membrane permeabilization, i.e., electroporation phenomenon (Neumann and Rosenheck 1972). The main advantage of this physical approach is that cell perturbation is perfectly controlled in time since the membrane electropermeabilization is transient and space since solely the tissue located in between electrodes is affected. The developments of this methodology are multiple both in food, environmental, and human health areas (Yarmush et al. 2014).

Skin Response to the Externally Delivered Pulsed Electric Fields When external electric fields are applied to the normal skin, several phenomena take place at the same time. Electric fields affect skin on tissue and cellular levels. On a tissue level, it was shown in multiple studies that electric fields increase the permeability of small molecules, DNA, and RNA into the skin (Yarmush et al. 2014). This long investigated property of increasing skin permeability suggested in the early studies the use of electroporation for needless drug delivery and later for DNA vaccination applications. An additional tissue-level effect of electric fields on the skin is the modulation of the blood flow. Studies on electrochemotherapy of tumors and normal skin (Fig. 4) showed that pulsed electric fields cause temporary vasoconstriction and then vasodilation at the areas where electrodes were applied (Jarm et al. 2010). In addition, recent studies showed that pulsed electric field applied in electroporation mode could kill cells, but did not affect the structure of extracellular matrix (Jiang et al. 2015). On the cellular and molecular level, the application of pulsed electric fields leads to cell reversible electroporation, cell irreversible electroporation, and death most probably through both necrosis and apoptosis, degranulation of mast cells, and release of multiple molecular factors to the treated areas at the time of treatment and up to days after the electric field was removed (Fig. 5). Proteins and DNA synthesis were showed to be promoted by non-permeabilizing pulsed electric field, but its fine characterization needs to be performed in skin electroporation context to use this effect in electro-induced healing strategy. The threshold for each phenomenon depends on the cell type and its specific location –“niche” in the skin. This “niche” is important as it defines the local conductivity and, thus, electric field distribution near and at the cell membrane. These complex responses of the skin to the applied pulsed electric field are only partially described. Precise control of the delivered electric fields in time and space to the skin opens new opportunities to treat wounds. In the following sections, the

6

L. Gibot and A. Golberg 1 min

Pulse generator

6h

12h

Blood Flow Change (%) with baseline

Current probe

80 60 40 20 0 −20 −40 −60 −80 −100

0 1 2 3 4 5 6 7 8 9 10 11 12 Time (h)

Fig. 4 Pulsed electric field effects on the normal rat skin. Schematic representation of the experimental setup and digital image of electrodes applied on the dorsal rat skin (left panel). Electroporation caused temporary vasoconstriction at the treated site (right panel). The blood flow returned to the basal levels around 6 h after pulsed electric field (PEF) application. In the following hours, the increase of the flow at the PEF-treated areas was observed. Figures show the images taken using Moor LDI instrument. The plot shows the ratio between the flow at the center of the PEF-treated area and the flow measured at the same area before PEF. Black squares on the images show the regions of interest at which the flow was measured before and after PEF. PEF protocol applied: 500 V, 200 pulses, 70 μs pulse duration, 3Hz, 2 mm gap between the electrodes (Figure adapted with permission from (Golberg et al. 2015). Copyright # 2015 Rights Managed by Nature Publishing Group. Distributed under a Creative Commons CC-BY license)

current state of the art of wound healing by pulsed electric fields will be described, and further studies in this field will be suggested.

Skin Rejuvenation with Pulsed Electric Fields Aging, trauma, and chronic metabolic diseases, including diabetes, often lead to alterations in skin color, texture, and barrier function. Loss of skin function leads to atrophy, infection, chronic wounds, laxity, and rhytides. Epidermal atrophy – characterized by a thinning of the epidermis and an increase in fragility – is observed in ~32 % of individuals over 60 years of age (Mengeaud et al. 2012). Thus, it is not surprising that 2,156,075 skin rejuvenation procedures were performed in 2013 in the United States alone. Because these skin diseases are often associated with poor reepithelialization, poor blood supply, reduced collagenesis, and a loss of collagen functional properties, current rejuvenation therapies are focused on the removal of nonfunctional tissue and the induction of de novo growth of healthy dermal cells, blood vessels, and extracellular matrix. Various physical and chemical methods are used for skin rejuvenation. However, the major disadvantage of current physical rejuvenation

Electroporation in Scars/Wound Healing and Skin Response

7

MCP-1

*

VEGF

* Time after PEF

IP-10

5weeks

*

2weeks

IL-10

*

2h

EGF

*

IL-1b

* 0

2

4

6

8

10

12

14

Fold change in comparison with normal skin

Fig. 5 Secreted molecules in the normal rat skin after electroporation. Factors secreted to the area of normal rat skin exposed to PEF protocol of 500 V, 200 pulses, 70 μs pulse duration, 2Hz, 2 mm gap between the electrodes. The bars show fold increase in the normalized total protein concentration at the treated areas in comparison to controls. The bars show the average from measurements between six treated areas in three different animals (Figure adapted with permission from (Golberg et al. 2015). Copyright # 2015 Rights Managed by Nature Publishing Group. distributed under a Creative Commons CC-BY license)

methods is that they deliver external energy to the whole tissue bulk, affecting both cells and extracellular matrix; this changes the function and architecture of treated tissue. The major disadvantage of chemical rejuvenation therapies is that even though they target only cells, they involve the delivery of external molecules that can cause an off-target tissue response. This uncontrollable subsequent reaction might result in clinical complications such as burns, skin vascular malformation, tumors, keloids, hypertrophic scarring, skin contraction, paralysis of facial muscles, necrosis, intravascular penetration, and infection. As electroporation affects precisely cell membrane and does not affect extracellular matrix, it was suggested that it could lead to cell-based therapy for skin rejuvenation. The cell-based therapy concept is that the energy is delivered to the skin cells and not to the bulk tissues as done by other physical methods. The application of pulsed electric field on the normal rat skin led to >72 % increase of total collagen content 3 weeks after the treatment in comparison with untreated skin (Table 1). This amount reduced to ~50 % increase of total collagen content 2 months after the treatment in comparison with untreated skin (Table 1; Golberg et al. 2015). In addition, the same study showed that the PEF protocol consisting of applied voltage, 500 V; pulse duration, 70 μs; number of pulses, 200; and frequency of pulses delivery, 3 Hz, leads to the increased epidermis thickness (up to 3 weeks after treatment) – important for infection prevention and angiogenesis, important for nutrient supply and waste removal, and important for increased metabolism, measured by the rate of glucose uptake (Fig. 6).

L. Gibot and A. Golberg

80 70 60 50 40 30 20 10 0

25

* *

20

* *

15 *

10 5 0

C on

tro 1 l 1 day da 1w ys e 3w es 2 ee m s on th s

Epidermis thickness (µm)

a

Stratum corneum Thickness (µm)

8

6h

24h

3w

2m

200 *

150

*

100 *

50

*

0 −50 -100

*

c

% Increase of 18FDG uptake

tim 6h e 0 o 12 urs ho 24 urs ho 72 urs ho 1w urs e 3w eks e 8w eks ee ks

Blood Flow change (%) with base line

b 1 min

200 180 160 140 120 100 80 60 40 20 0

*

* * *

12 hours

1 day

3 days

1 week

Fig. 6 Skin rejuvenation with pulsed electric fields. (a) Dynamics of thickened skin epidermis and resolution to the baseline levels after PEF. Images show H&E staining of the PEF-treated epidermis. The plots show the average thickness of the epidermis and stratum corneum. Yellow arrows show occasional apoptotic basal cells (*) p-val < 0.001. (b) Angiogenesis in the skin detected with Doppler blood flow imaging (left panel) and immunohistochemistry of the newly formed capillaries (right panel) with detection of angiogenesis marker Nestin. Increase in the Nestin expression in the papillary dermis capillaries (black arrows) in comparison to untreated skin was observed from 1 day to 3 weeks after PEF treatment. The intensity of staining 2 months after PEF was very similar to control, suggesting the maturation of the vessels. (c) The increased metabolism of skin areas treated by PEF as detected using 18FDG uptake (Figure adapted with permission from (Golberg et al. 2015). Copyright # 2015 Rights Managed by Nature Publishing Group. Distributed under a Creative Commons CC-BY license)

Electroporation in Scars/Wound Healing and Skin Response

9

Table 1 The impact of PEF on total collagen content in the normal rat skin. Based on data from (Golberg et al. 2015) Voltage (V) 100 200 300 400 500 100 200 300 400 500 100 200 300 400 500 100 200 300 400 500 100 200 300 400 500 Control

Duration (μs) 10 10 10 10 10 30 30 30 30 30 50 50 50 50 50 70 70 70 70 70 90 90 90 90 90

Number of pulses N 200 400 600 800 1000 400 600 800 1000 200 600 800 1000 200 400 800 1000 200 400 600 1000 200 400 600 800

3 weeks (mg/100 mg tissue) (n = 3) 136 217 197 205 210 153 152 180 213 194 143 177 192 163 171 143 137 223 197 193 165 183 222 185 205 129

STDEV 22 36 25 15 9 26 22 43 25 25 28 12 23 4 12 26 21 11 35 41 18 14 14 7 5 9

2 months (mg/100 mg tissue) (n = 3) 178 147 143 144 178 95 151 206 147 174 150 154 156 170 122 145 183 186 164 153 111 136 153 143 167 129

STDEV 18 29 3 14 2 14 13 7 2 1 21 7 20 16 9 11 6 23 18 17 14 21 13 13 15 9

Gene Electrotransfer (GET) as a Tool in Wound-Healing Therapy Electroporation is a reliable technique for nucleic acid delivery in targeted organs and tissues: it is the gene electrotransfer, or GET. At skin level, for example, this technique allows to safely and locally deliver therapeutic genes to stimulate wound healing (Gibot and Rols 2016). Gene electrotransfer was first developed for DNA vaccination and gene therapy, but it also shows great promise in the field of regenerative medicine since it allows the continuous and local production of therapeutic factors involved in woundhealing process.

10

L. Gibot and A. Golberg

While acute wounds occur suddenly and heal at a predictable rate according to the classical wound-healing process, chronic wounds are associated with impaired/ delayed healing and include pressure ulcer, diabetic foot ulcer, and venous stasis ulcer. Numerous changes associated with diabetes at the cellular, molecular, and genetic levels have been identified and promote this delayed wound healing (Blakytny and Jude 2006). For these situations of nonhealing wounds, the development of efficient therapeutic strategies is urgent. Indeed, chronic wounds are becoming a major societal issue with the aging of the population and represent a significant burden to patient and health-care system. While the topical application of specific recombinant growth factors was shown to benefit cutaneous healing, it requires large amounts of peptide because of multiple administrations due to their short half-life which make this approach complex and expansive. A promising and alternative approach consists in the local and transient administration of a gene encoding the growth factor of interest. This is where gene electrotransfer (GET) is highly valuable. Indeed, GET is a safe and efficient way to deliver therapeutic genes which make it an attractive approach for the treatment of wound healing and regenerative medicine in general. Several recent reviews compile in vitro and in vivo studies on gene electrotransfer in wound healing (Gibot and Rols 2016). The vast majority of studies on GET in wound healing were performed in vivo. The pig is the most relevant animal model for studying healing process because its mechanisms are comparable to those found in humans. However, for convenient reasons, it is the murine model which is the most widely used. It offers the advantage of working with available mutants, who allows to study GET for wound healing in both normal context and within diabetic models. Indeed, leptin receptor-deficient diabetic mice Leprdb/Leprdb are a relevant model of delayed wound healing since they recreate a biochemical environment found in diabetic patients. The choice of a model is based on many factors such as availability, cost, and ease of use and on its anatomic/functional similarity to human tissue. Small mammals such as rodents are conventionally used for studies on cutaneous physiopathology. However, murine skin differs from the human skin from the viewpoint of its architecture and its cellular composition. In addition, a transcriptomic study recently showed that only 30 % of genes associated with the skin are identical between human and mouse (Gerber et al. 2014). Thus, it is of utmost importance for the next years to develop human tissue models, especially by tissue engineering approaches, in order to study gene electrotransfer in human tissue context. In order to stimulate wound-healing process, therapeutic genes have been successfully electrotransferred to injured skin. Plasmids encoding growth factors known to promote cell proliferation, migration, and differentiation and to stimulate angiogenesis were applied. Vascular endothelial growth factor (VEGF), keratinocyte growth factor (KGF), and transforming growth factor (TGF) were nowadays the most used in experiments. Interestingly, using electroporation type of pulses to deliver therapeutic gene in diabetic wound murine model, Lee et al. demonstrated that the control group “electric field alone” significantly improves cutaneous wound closure (Lee

Electroporation in Scars/Wound Healing and Skin Response

11

et al. 2004). This observation underlines the electrostimulation potential of electrical parameters used in electroporation context.

Electrostimulation of Healing Normal human epidermis possesses a transepithelial potential which is established via an asymmetrical distribution of Na + ions in the different compartment of the epidermis. Skin injury disrupts this transepithelial potential and induces an endogenous electric field directed toward the center of the wound. This electric field seems to facilitate reepithelialization of the wound. Indeed, cutaneous cells are sensitive to this electric field that directs their migration during wound healing. This phenomenon is named galvanotaxis. In vitro studies led on monolayers showed that cell exposure to external electric field affects protein distribution and synthesis and cell orientation and migration through phosphatidylinositol-3-OH kinase-gamma and PTEN pathways (Zhao et al. 2006). When standard wound care alone fails to heal chronic wounds, electric field application has been shown to improve healing and wound closure: It is electrostimulation (Kloth 2014). A meta-analysis recently published demonstrated that electrostimulation is efficient in improving wound healing, even if electrical parameters applied are empirically chosen (Koel and Houghton 2014). Evidences obtained from experimental data and clinical trials efficiently support the benefic role of electrostimulation for promoting wound-healing process, especially in a context of pathologic chronic wounds. Thus, American Centers for Medicare and Medicaid Services (CMMS) have officially recommended the use of electric fields for the treatment of chronic wounds (Kloth 2014).

Electrochemotherapy Stimulates Aesthetic and Functional Wound-Healing Process In oncology, electroporation ensures targeted and massive influx of cytotoxic molecules in tumor cells. This clinical application, called electrochemotherapy (ECT), is nowadays used in several European countries for the treatment of skin and subcutaneous cancers. By potentiating the cytotoxicity of drugs, ECT reduces the doses injected into patients and thus limits their side effects. Furthermore, clinicians commonly observed, but cannot explain, aesthetic and functional wound healing of the sites treated with ECT. Molecular assumptions regarding the mechanisms underlying this phenomenon will be proposed in the section devoted to ECT and wound healing. When tumors are located in head and neck sites, it is of utmost importance that the treatment promotes an aesthetic and functional healing, which is always observed by the clinicians with ECT (Fig. 7). The common assumption to explain this observation is that normal cells located at the margin of the tumor are spared because of their quiescence, which makes them less sensitive to genotoxic effect of drugs used in ECT such as bleomycin and

L. Gibot and A. Golberg

Melanoma

Basocellula carcinoma

12

Temple

Scalp

Fig. 7 Aesthetic and functional wound healing observed after tumor treatment by ECT (Adapted from (Glass et al. 1997; Marty et al. 2006). Copyright # 2016 Elsevier with permissions)

cisplatin. New data obtained recently allow proposing a novel molecular hypothesis that needs to be experimentally confirmed. Calvet et al. demonstrated that both ECT with bleomycin and electric field alone induced the externalization of calreticulin (Calvet et al. 2014). Calreticulin is usually an endoplasmic reticulum-resident protein. When externalized, it becomes a danger-associated molecular pattern molecules (DAMP) associated with the induction of immunogenic cell death. However, interesting studies showed that calreticulin externalization also plays a major role in cutaneous wound healing by stimulating fibroblast and keratinocyte proliferation, migration, and differentiation (Nanney et al. 2008). These effects were shown to pass through TGF-β3 secretion, a growth factor well known to induce collagen and hyaluronan remodeling, to accelerate wound-healing process, and to improve cell migration.

Burn Wound Disinfection with Pulsed Electric Fields Approximately 500,000 people seek medical treatment for burns every year in the United States; infection remains a major cause of morbidity and mortality in these patients. In addition to the extent and nature of the thermal injury affecting the susceptibility to infection, the type and amount of the microbial burden colonizing the wound appear to influence the risk of morbidity and mortality. Pathogens that infect burn wounds are primarily Acinetobacter baumannii, methicillin-resistant Staphylococcus aureus, Pseudomonas, and Klebsiella – pathogens that are increasingly resistant to various antimicrobial agents. To address this problem, based on

Electroporation in Scars/Wound Healing and Skin Response

13

Fig. 8 IRE disinfection in vivo: C57BL/6 black mice model. (a) Schematic illustration for procedures used in the study. The 1 cm2 burn injury was followed by dispersion of A. baumannii on part of the wound. Next pulsed electric field was applied using two plate electrodes. A. baumannii infection load was quantified using bioluminescent imaging. (b) Left panel shows digital photography of the burned (white frame) and infected (orange frame) areas of the skin. Central panel shows digital photography of the applied electrodes. Right panel shows the images of the mice as observed inside the dark imaging box. Orange frame shows the infected area as detected by a strong bioluminescent signal emitted from bacteria. (c) The performed experiment timeline (Figure adapted with permission from (Golberg et al. 2014). Copyright # 2014 World Scientific and Imperial College Press Co Pte Ltd. Illustration schemes by Inna Desyatnik)

successful disinfection of food, IRE was applied to treat the bacterial wound infections (Fig. 8). Using 80 pulses of 5000 Vcm 1, stable disinfection with 4.91  0.71 Log10 reduction of Acinetobacter baumannii, 3 h after treatment, was achieved (Figs. 8, 9, and 10) (Golberg et al. 2014). The magnitude of disinfection was correlated with both the electric field strength and the number of delivered pulses. The increase in pulse number led to a larger reduction in bacterial load and bioluminescent signal immediately after treatment, as compared to the increase in the field strength. Increasing the pulse number from 40 to 80 led to a 255 % increase in the reduction of bacterial load in the wound, from 1.49  0.07 Log10 to 5.30  0.85 Log10. Increasing the field, however, from 2500 V cm 1 to 5000 V cm 1, while keeping the number of pulses at 40, led to only a ~37 % increase in the log reduction of bacterial load in the wound, from 1.49  0.07 Log10 to 2.04  0.29 Log10.

14

L. Gibot and A. Golberg

b

1.E−04

500V, 80pulse 35min

55min

=3 s ul

se

s in m 55

1000V, 80pulse 35min

55min

240min

p=0.179 1.E−07 Survival Fraction

Survival Fraction

e

240min

p=0.055

1.E−07

1.E−01

se ,4 in

m 45

240min

p=0.0237

(N

=3 ul

10

(N

=3 (N in m

se ul

d

Control

1.E−02 Survival Fraction

)

) =3 s

(N 0p ,8 in m 55

45

m

in

,4

0p

35

ul

m

se

in

s

(N

=3

) =3 (N

=3 (N in m 10

)

1.E+00 )

1.E−01

1.E+00

)

1.E−02

1.E−01

35min

p=0.0052

1.E−03

0p

1.E−02

p=0.095

1.E−05

,8

1.E−03

10min

1000V, 70µs, 1Hz p=0.0052

)

Survival Fraction

p=0.0097

c

55min

1.E−06

1.E−05 1.E−04

45min

1.E−07

p=0.03

1.E−06

35min

)

500V, 70µs, 1Hz p=0.0096

1.E−07 Survival Fraction

10min

=3

55min

0p

45min

in

35min

m

10min

1000V

(N

500V

35

a

1.E−05 1.E−03 1.E−01

1.E−05 1.E−03 1.E−01

35 m

in

55 (N= m 3) in (N 24 0m =3 ) in (N =3 )

55 m =3) in (N 24 0m =3 ) in (N =3 )

(N in 35 m

(N 24 0m =2 ) in (N =2 )

in

35 m

10 m

in

(N

=2

)

1.E+00

Fig. 9 The effect of pulse number and electric field strength on A. baumannii infection load reduction. (a) Applied voltage 500 V, 2 mm gap between electrodes. (b) Applied voltage 1000 V, 2 mm gap between electrodes. (c) Control: not treated, burned, and infected skin. (d) Applied voltage 500 V, 2 mm gap between electrodes. (e) Applied voltage 1000 V, 2 mm gap between electrodes. Top panel shows the post-burn time when the images were taken. The bottom panel shows the survival fraction of microorganisms as detected by the top panel images (logarithmic, inversed scale). N shows the number of animals per group. Error bar  standard deviation of the mean (Figure adapted with permission from (Golberg et al. 2014). Copyright # 2014 World Scientific and Imperial College Press Co Pte Ltd)

Electroporation in Scars/Wound Healing and Skin Response

15

Fig. 10 Summary of A. baumannii survival 3 h after IRE, the effect of electric field intensity. N shows the number of animals per group. Error bar  standard deviation of the mean. Gap between electrodes 2 mm. Pulse duration: 70 μs (Figure adapted with permission from (Golberg et al. 2014). Copyright # 2014 World Scientific and Imperial College Press Co Pte Ltd)

Applications of Electroporation in Surgery: Mesh Disinfection and Scar Treatment Two million Americans undergo abdominal surgery annually with a postoperative incisional hernia rate of 10–23 % (Khan, Blumrosen et al. 2016). About 400,000 ventral hernia repairs are performed each year in the United States alone, with reported hernia recurrences in 40–50 % of cases (Khan, Blumrosen et al. 2016). Synthetic mesh reinforces hernia repairs and has been shown to decrease recurrences compared to primary repair alone. However, morbidities related to mesh infection can limit efficacy. Reported mesh infection rates range from 4 % to 16 %. Antibiotics alone are not an effective treatment for mesh infections. Khan et al. (2016) investigated the effect of pulsed electric fields (PEF) on biofilm-infected mesh (Khan, Blumrosen et al.). In that study, Prolene mesh was infected with bioluminescent Pseudomonas aeruginosa and treated with PEF using a concentric electrode system to derive, in a single experiment, the critical electric field strength needed to kill bacteria. The effect of the electric field strength and the number of pulses on bacterial eradication was investigated. For all experiments, biofilm formation and disruption were confirmed with bioluminescent imaging and scanning electron microscopy (SEM). It was found that the critical electric field strength needed to eradicate 100–80 % of bacteria in the treated area was 121  14 V/mm when 300 pulses (50 μs, delivered at 2 Hz) were applied and 235  6.1 V/mm when 150 pulses were applied. The area at which 100–80 % of bacteria were eradicated

16

L. Gibot and A. Golberg

was 50.5  9.9 mm2 for 300 pulses and 13.4  0.65 mm2 for 150 pulses. Eighty percent threshold eradication was not achieved with 100 pulses. Hypertrophic scarring (HTS) after trauma and burn injury remains a major clinical challenge that leads to physical, aesthetic, functional, psychological, and social stresses in thousands of patients. Current data show that alterations in coagulation, inflammation, angiogenesis, fibroplasia, contraction, remodeling, and mechanical tension correlate with the formation of HTS. But as of yet, the mechanisms that induce HTS are not well understood. This gap in knowledge leads to limited clinical success in therapeutic procedures. Various techniques such as surgical excision, intralesional steroid or interferon injection, cryotherapy, laser therapy, electron-beam irradiation, mechanical compression dressing, silicone sheet application, and combinations thereof have been tested to treat scars over the years. Despite all of these efforts, a recent metareview shows that there are only modest improvements in the healing outcomes among all these treatments (Leventhal et al. 2006). Recent study used electrochemotherapy for keloid and hypertrophic scar treatment in the clinical trials (Manca et al. 2013). Twenty patients with keloids or hypertrophic scars were treated with one or more sessions of electrochemotherapy with bleomycin. Treatment was well tolerated by patients, and no serious adverse events were observed with one recurrence observed after 18 months. A median reduction of 87 % was observed in volume size, and 94 % of lesions showed a volume reduction of more than 50 %. Scar pliability and erythema scores were also significantly reduced. A reduction in hitching was observed in 89 % of patients, and a reduction in pain was observed in 94 % (Manca et al. 2013).

Conclusion In the case of the normal wound healing, no intervention might be needed. However, in cases of abnormal healing, both chronic wounds and hypertrophic scarring, there is an obvious need for intervention. For many decades, wound care methods have been developed to improve healing by addressing cutaneous cell stimulation, reduction of bacterial contamination, and regulation of cytokines and proteases or by promoting exogenous growth factors expression. It is important to remember that wound healing is an extremely complex processing, and interventions that could achieve the system effects are needed to advance this field. Manipulations of cell metabolism, life cycle, and density could provide partial solution to improve the patient care. Electroporation-based technologies can obviously help achieve all these goals. It has been already shown that electroporation-based procedures can manipulate cell metabolism and gene expression in wounds; they can control cell density and infection and can directly be applied to the skin to both induce skin rejuvenation and reduce hypertrophic scarring. Each of these applications requires its own optimization for the applied parameters as sometimes the required effect is on a single cell-type level and sometimes overall skin organ level. Future work should address both specificity of electric field parameters for various applications and

Electroporation in Scars/Wound Healing and Skin Response

17

integration of electroporation-based devices and procedures to the different existing protocols and workflows of clinical wound care.

Cross-References ▶ Clinical Applications of Gene Therapy: Principles of Gene Electrotransfer ▶ Electrochemotherapy and Its Clinical Applications ▶ Electroporation and Electropermeabilization ▶ Electroporation-Based Applications in Biotechnology ▶ Gene Electrotransfer for Ischemic Tissue ▶ Irreversible Electroporation and Its Clinical Applications

References Blakytny R, Jude E (2006) The molecular biology of chronic wounds and delayed healing in diabetes. Diabet Med 23(6):594–608 Calvet CY, Famin D, Andre FM, Mir LM (2014) Electrochemotherapy with bleomycin induces hallmarks of immunogenic cell death in murine colon cancer cells. Oncoimmunology 3:e28131 Duffield JS, Lupher M, Thannickal VJ, Wynn TA (2012) Host responses in tissue repair and fibrosis. Annu Rev Pathol 8:241–276 Gerber PA, Buhren BA, Schrumpf H, Homey B, Zlotnik A, Hevezi P (2014) The top skinassociated genes: a comparative analysis of human and mouse skin transcriptomes. Biol Chem 395(6):577–591 Gibot L, Rols MP (2016) Gene transfer by pulsed electric field is highly promising in cutaneous wound healing. Expert Opin Biol Ther 16(1):67–77 Glass LF, Jaroszeski M, Gilbert R, Reintgen DS, Heller R (1997) Intralesional bleomycin-mediated electrochemotherapy in 20 patients with basal cell carcinoma. J Am Acad Dermatol 37 (4):596–599 Golberg A, Bei M, Sheridan RL, Yarmush ML (2013) Regeneration and control of human fibroblast cell density by intermittently delivered pulsed electric fields. Biotechnol Bioeng 110 (6):1759–1768 Golberg A, Broelsch GF et al (2014) Eradication of multidrug-resistant in burn wounds by antiseptic pulsed electric field. Technology 2(2):153–160 Golberg A, Khan S et al (2015) Skin rejuvenation with non-invasive pulsed electric fields. Sci Rep 5:10187 Jarm T, Cemazar M, Miklavcic D, Sersa G (2010) Antivascular effects of electrochemotherapy: implications in treatment of bleeding metastases. Expert Rev Anticancer Ther 10(5):729–746 Jiang C, Davalos RV, Bischof JC (2015) A review of basic to clinical studies of irreversible electroporation therapy. IEEE Trans Biomed Eng 62(1):4–20 Khan SI, Blumrosen G et al (2016) Eradication of multidrug-resistant pseudomonas biofilm with pulsed electric fields. Biotechnol Bioeng 113(3):643–650 Kloth LC (2014) Electrical stimulation technologies for wound healing. Adv Wound Care 3 (2):81–90 Koel G, Houghton PE (2014) Electrostimulation: current status, strength of evidence guidelines, and meta-analysis. Adv Wound Care 3(2):118–126 Koh TJ, DiPietro LA (2011) Inflammation and wound healing: the role of the macrophage. Expert Rev Mol Med 13:e23

18

L. Gibot and A. Golberg

Lee PY, Chesnoy S, Huang L (2004) Electroporatic delivery of TGF-beta1 gene works synergistically with electric therapy to enhance diabetic wound healing in db/db mice. J Invest Dermatol 123(4):791–798 Leventhal D, Furr M, Reiter D (2006) Treatment of keloids and hypertrophic scars: a meta-analysis and review of the literature. Arch Facial Plast Surg 8(6):362–368 Manca G, Pandolfi P, Gregorelli C, Cadossi M, de Terlizzi F (2013) Treatment of keloids and hypertrophic scars with bleomycin and electroporation. Plast Reconstr Surg 132(4):621e–630e Marty M, Sersa G et al (2006) Electrochemotherapy – an easy, highly effective and safe treatment of cutaneous and subcutaneous metastases: results of ESOPE (European Standard Operating Procedures of Electrochemotherapy) study. Eur J Cancer Suppl 4(11):3–13 Mengeaud V, Dautezac-Vieu C, Josse G, Vellas B, Schmitt AM (2012) Prevalence of dermatoporosis in elderly French hospital in-patients: a cross-sectional study. Br J Dermatol 166(2):442–443 Nanney LB, Woodrell CD et al (2008) Calreticulin enhances porcine wound repair by diverse biological effects. Am J Pathol 173(3):610–630 Neumann E, Rosenheck K (1972) Permeability changes induced by electric impulses in vesicular membranes. J Membr Biol 10(3):279–290 Robson MC, Steed DL, Franz MG (2001) Wound healing: biologic features and approaches to maximize healing trajectories. Curr Probl Surg 38(2):72–140 Yarmush ML, Golberg A, Sersa G, Kotnik T, Miklavcic D (2014) Electroporation-based technologies for medicine: principles, applications, and challenges. Annu Rev Biomed Eng 16:295–320 Zhao M, Song B et al (2006) Electrical signals control wound healing through phosphatidylinositol3-OH kinase-gamma and PTEN. Nature 442(7101):457–460

Principles and Use of Magnetic Resonance Electrical Impedance Tomography in Tissue Electroporation Eung Je Woo and Matej Kranjc

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrical Behavior of Biological Tissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Assessing Electrical Conductivity of Tissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electrical Impedance Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Magnetic Resonance Electrical Impedance Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MREIT in Tissue Electroporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 4 4 4 11 15 16 17

Abstract

Electroporation is a phenomenon caused by externally applied electric field of an adequate strength and duration to cells that results in increase of cell membrane permeability to various molecules which otherwise are deprived of transmembrane transport mechanism. As the membrane electroporation is a consequence of an induced transmembrane potential which is directly proportional to the local electric field, magnetic resonance electrical impedance tomography (MREIT) was proposed for reconstruction of electric field distribution during electroporation. MREIT enables determination of electric field distribution by measuring the electric current density distribution and electric conductivity of the treated subject during application of electric pulses by using MRI scanner and E.J. Woo (*) Department of Biomedical Engineering, College of Medicine, Kyung Hee University, Seoul, Republic of Korea e-mail: [email protected] M. Kranjc (*) Faculty of Electrical Engineering, University of Ljubljana, Ljubljana, Slovenia e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_65-1

1

2

E.J. Woo and M. Kranjc

numeric algorithms. MREIT yields an electric field distribution, which is a time average of its altering time course so that all the consequences of conductivity alteration of the treated tissue due to electroporation are not neglected within obtained electric field distribution. Feasibility of this method has been demonstrated by determining electric field distribution during electroporation in silico, in agar phantoms, plant tissues, and animal tissues ex vivo and in vivo. In this chapter, mathematical framework of MREIT and the concept of monitoring electric field distribution are provided together with fundamentals of electrical conductivity imaging. Maps of electric field distribution established during electroporation of various tissues that were obtained by means of MREIT are also included. Keywords

Electric field • Magnetic resonance imaging • Electrical impedance • Current density

Introduction An accurate coverage of the cell with sufficiently large electric field (see “▶ Electric Field Distribution and Electroporation Threshold”) presents one of the most important conditions for successful electroporation. A method that would allow determination of the electric field strength within tissue would thus be of great importance for electrochemotherapy as well as for other electroporation based applications. A method capable of determining electric field distribution during the pulse delivery seems to be useful as electroporation depends on local electric field. This would enable detection of insufficient electric field coverage before the end of the treatment, thus increasing and assuring its effectiveness. As the electric field distribution inside the observed tissue cannot be measured directly, an indirect approach of obtaining it needs to be evaluated. By means of Ohm’s law, the electric field distribution can be determined when an electric current density and an electrical conductivity of the tissue are obtained. The electric current density inside a conductive tissue can be obtained by magnetic resonance imaging (MRI) using current density imaging technique (CDI) (see “▶ Current Density Imaging as Means to Follow Tissue Electroporation”) by measuring magnetic field changes caused by applied current. Whereas tissue conductivity can be obtained by magnetic resonance electrical impedance tomography (MREIT), a technique used for reconstruction of electrical conductivity inside a tissue by means of current density. The use of MREIT for conductivity imaging has advanced rapidly in the last decade and has now reached the stage of in vivo animal and human imaging experiments. As the method does not present additional cost to the conventional MRI procedure, it could become a part of an MRI system and provide additional and valuable contrast information. This chapter is divided in two parts. In the first part, mathematical framework of MREIT is provided together with fundamentals of electrical conductivity imaging.

Principles and Use of Magnetic Resonance Electrical Impedance Tomography in. . .

3

The second part of the chapter is dedicated to results of monitoring electric field distribution during application of electric pulses in agar phantom, tissue ex vivo, and mouse tumor in vivo and in silico.

Electrical Behavior of Biological Tissue In this chapter, we deal with bioelectromagnetic phenomena at dc or low frequency so that we can ignore the effects of electric polarization and magnetic induction. When exposed to such an electric field, a biological tissue conducts electrical current. Given the electric field, the amount of the current depends on a passive material property called the electrical conductivity. In Ohm’s law, we can express the relation as JðrÞ ¼ σ ðrÞEðrÞ ¼ σ ðrÞ∇vðrÞ

(1)

where J is the current density, σ the conductivity, E the electric field intensity, v the electric potential or voltage, and r ¼ ðx, y, zÞ a position vector. In an electrolyte, concentrations and mobility of charge carriers, predominantly ions, determine the conductivity σ. A biological tissue consists of cells, extracellular matrix, and extracellular fluid. Inside the cell, there exist numerous organelles in intracellular fluid. The complicated structures and compositions of various tissues influence their electrical conductivity values producing an inhomogeneous conductivity distribution inside the human body (Grimnes and Martinsen 2014). The cell has a membrane which behaves electrically as a thin insulator. At low frequency, the membrane blocks electrical current from flowing into the cell and the entire cell appears as a solid insulator when neglecting ion movements through ion channels. At high frequency, a displacement current may flow through the cells since the thin cellular membranes behave as capacitors. The conductivity at a macroscopic scale may exhibit anisotropy at low frequency depending on the structure of the cells and extracellular matrix materials. When the conductivity is anisotropic, it can be represented as a tensor. Note that in this case, the direction of the current density differs from the direction of the applied electric field. The electric field can stem from an endogenous neuronal current or externally injected current (Malmivuo and Plonsey 1995). In this chapter, we study the effects of the electric field subject to an externally injected current. For the case where we are interested only in the internal current density produced by the externally injected current, the current density J satisfies ∇  Jð r Þ ¼ 0

(2)

since there is no such current source inside the body. From Eqs. 1 and 2, the externally injected current produces a voltage distribution v inside the body satisfying the following boundary value problem:

4

E.J. Woo and M. Kranjc




∇  σ ðrÞ∇vðrÞ ¼ 0 in Ω σ∇v  n ¼ g on @Ω

(3)

where Ω is the three-dimensional domain of the body, @Ω its boundary, n the outward unit normal vector on @Ω and g the normal component of the current density on @Ω (Neumann boundary condition). Note that on the portion of the boundary @Ω excluding the current-injection electrodes, g ¼ 0 since the air outside Ω is an insulator. Choosing a voltage reference position r0 where vðr0 Þ ¼ 0, the boundary value problem in Eq. 3 has a unique solution v in Ω.

Assessing Electrical Conductivity of Tissue Electrical Impedance Tomography Electrical impedance tomography (EIT) is an imaging modality to produce crosssectional images of the conductivity distribution inside an electrically conducting object such as the human body (Holder 2004). An EIT system may consist of a single or multiple current sources that can produce sinusoidal currents of a safe amplitude of less than a few mA at a single or multiple frequencies in the range of 1–500 kHz. Such currents are injected into the human body through a chosen pair or plurality of surface electrodes attached around an imaging plane of interest. Typical EIT systems use 8–32 electrodes. From those electrodes, the EIT system measures the induced boundary voltages using a single or multiple voltmeters. The measured boundary current–voltage data, so called the Neumann-to-Dirichlet map (NtD map), is determined by the conductivity distribution inside the human body and also its boundary shape and electrode positions. Using the measured boundary data, an image reconstruction algorithm, which is basically based on Eq. 3, produces conductivity images (Seo and Woo 2013). There have been numerous studies of EIT in terms of its theories, algorithms, system developments, imaging experiments using phantoms, animals and human subjects, and also clinical applications (Holder 2004). EIT produces images with a low spatial resolution of about 5–10 mm pixel size in most cases. Its temporal resolution is superb with more than 50 frame/s using a fast parallel EIT system. Since the system is portable and can be less expensive than other medical imaging modalities, EIT is finding clinical applications where image-based real-time monitoring is desired.

Magnetic Resonance Electrical Impedance Tomography Noting that EIT using only boundary measurements has a technical limitation to achieve a high spatial resolution, magnetic resonance electrical impedance tomography (MREIT) has been suggested (Zhang 1992; Woo et al. 1994; Seo et al. 2014). The basic idea is to utilize the internal information of the magnetic field induced by

Principles and Use of Magnetic Resonance Electrical Impedance Tomography in. . .

5

an externally injected current. From the Biot-Savart law, the induced magnetic flux density B is ð 0 μ0  0  r  r 0 J r dr þ BX ðrÞ BðrÞ ¼ 0 3 4π jr  r j

(4)

Ω

where J is the current density inside the domain Ω, and BX is the magnetic flux density produced by the currents outside the domain such as lead wires. Note that Laplacian of BX inside the domain Ω is zero (Lee et al. 2003). Ampere’s law provides the relation between the current density and the magnetic flux density inside the domain Ω as Jð r Þ ¼

1 ∇  BðrÞ: μ0

(5)

MREIT utilizes a MRI scanner to noninvasively measure the induced magnetic flux density B inside the human body. MR imaging is based on the magnetization of the protons inside the human body by a strong dc magnetic field of 3 T, for example. Aligned magnetic dipole moments of the protons precess at the Larmor frequency that is 127.74 MHz at 3 T. Application of an electromagnetic wave at the Larmor frequency flips the protons to an unstable state. When this radio frequency (RF) pulse disappears, the protons return to their original magnetized state and emit an electromagnetic wave, which can be detected using RF coils. Adding spatially varying magnetic fields called the gradient fields, the resonance frequency of the protons varies depending on their positions inside the human body. This kind of position encoding allows formation of a cross-sectional image. A pulse sequence includes such RF and gradient pulses to affect the protons with position encoding. Different pulse sequences may produce MR images with various contrast information. The magnetic flux density in Eq. 4 perturbs the main magnetic field of the MRI scanner (Joy et al. 1989; Seo and Woo 2014). The magnetization of the protons occurs in the direction of the main magnetic field, which is the z direction in this  chapter. This means that only the z-component Bz of B ¼ Bx , By , Bz influences the magnetization. Note that Bz varies with position, and its spatial variation is determined by the conductivity distribution σ. Figure 1 is a typical pulse sequence used in MREIT. It is a basic spin-echo (SE) pulse sequence with added current pulses, which are synchronized with the RF pulses. The current injection is repeated twice using the current pulses with the same amplitude and opposite polarities denoted as I . This kind of MREIT pulse sequence produces two sets of complex MR data S , so-called the k-space data, as follows: 

S ðm, nÞ ¼

ðð

Mðx, yÞeiδðx, yÞ eiγBz ðx, yÞT c eiðxmΔkx þynΔky Þ dxdy

(6)

6

E.J. Woo and M. Kranjc

Fig. 1 Typical MREIT pulse sequence based on the spin-echo pulse sequence. Injection current pulses are inserted between RF pulses and reading gradient

where  are for the current injections of I  , respectively, M the MR magnitude image, δ the systematic phase artifact, γ the gyromagnetic ratio of hydrogen, Bz the z-component of the induced magnetic flux density, and Tc the width of the injection current pulses (Joy et al. 1989; Seo and Woo 2014). Applications of the Fourier transform result in two complex MR images Ψ as Ψ ðx, yÞ ¼ Mðx, yÞeiδðx, yÞ eiγBz ðx, yÞT c :

(7)

Note that Bz appears in the phase part of the complex MR images. To remove the systematic phase artifact δ due to the main field inhomogeneity, these two complex images are divided to get the phase difference of  Φðx, yÞ ¼ arg

Ψþ ðx, yÞ Ψ ðx, yÞ

 ¼ 2γBz ðx, yÞT c :

(8)

After applying proper preprocessing methods including phase unwrapping and scaling, the induced magnetic flux density Bz is obtained as Bz ðx, yÞ ¼

Φðx, yÞ : 2γT c

(9)

Figure 2 shows a typical example of a phantom imaging experiment in MREIT. Note that the conductivity anomaly at the center of the phantom refracts the Bz signal. To compute the internal current density J using Ampere’s law in Eq. 5, all  three components of B ¼ Bx , By , Bz should be measured. This can be done by

Principles and Use of Magnetic Resonance Electrical Impedance Tomography in. . .

a

7

b

-8

x 10 [T]

c

d

6 4 2 0 -2 -4 -6

Fig. 2 (a) Conductivity phantom including an agar object with 0.5 S/m conductivity at the center of the phantom with a background saline of 2 S/m conductivity. (b) Its MR magnitude image at the middle of the phantom. (c) Phase image at the same imaging slice. (d) Bz image computed by using Eq. 9

repeating the data collection process twice more while rotating the imaging object inside the MRI scanner (Joy et al. 1989). Once B is available, J is computed using Eq. 5, and images of the conductivity distribution can be obtained by using a method such as the J-substitution algorithm (Kwon et al. 2002). Though there exist numerous studies of utilizing B in MREIT and also in magnetic resonance current density imaging (MRCDI) as reviewed in Woo and Seo (2008), this object rotation is difficult or impractical using a currently available clinical MRI scanner. To avoid the impractical object rotations, it is highly desirable to reconstruct current density and conductivity images using the measured data of Bz. Using only Bz data, computation of only an approximate J is possible since Bz does not contain any information about Jz. Noting that J is strongly influenced by the threedimensional geometry of the imaging object as well as its internal conductivity distribution, a model-based approach may produce a best achievable current density JP, so-called the projected current density (Park et al. 2007). Using a three-

8

E.J. Woo and M. Kranjc

dimensional model of the imaging object, numerical solutions of Eq. 3 and Eq. 1 assuming a homogeneous conductivity distribution of 1 S/m inside Ω provide the internal current density JH (Lee et al.  2003). Assuming that the z-component Jz of the true current density J ¼ J x , J y , J z is a perturbation of the z-component JH z of the   H H H H computed current density J ¼ J x , J y , J z , the projected current density can be computed as a best approximation of the true current density one can compute by using the measured Bz data. In addition to the estimation of the internal current density from the measured Bz data, the projected current density can be utilized in conductivity image reconstructions using a similar approach as the J-substitution algorithm (Kwon et al. 2002). Figure 3 shows an example of the projected current density experimentally obtained from an in vivo canine brain. For conductivity image reconstructions using only Bz, the induced magnetic flux density Bz can be expressed from Eqs. 1 and 4 as
 0   dv  0   0 0  dv ð σ ðrÞ x  x r  yy r μ 0 dy dx Bz ðrÞ ¼ 0 dr 0 3 4π jr  r j

(10)

Ω

to emphasize the fact that it includes the information about the conductivity σ. Here, we ignore the magnetic flux density originated from the external lead wires since its Laplacian is zero in the domain Ω. Direct reconstruction of a conductivity image from the measured Bz data requires at least two current injections in orthogonal directions as much as possible (Seo and Woo 2011). For isotropic or equivalent isotropic conductivity image reconstructions, two current injections are commonly used. In this kind of Bz-based MREIT, the most widely used conductivity image reconstruction method called the harmonic Bz algorithm is based on the following observation (Seo et al. 2003):



 1 2 dσ dσ dv dv dσ dv dσ dv , ,  ∇ Bz ðx, yÞ ¼  ¼ μ0 dx dy dy dx dx dy dy dx

(11)

at the position (x, y) in a chosen imaging slice. Since Eq. 11 is satisfied for both current injections I  with the same conductivity distribution σ, we can get 2

dv1 6 dy 6 4 dv2 dy

3 dv1 2 dσ 3

  1 ∇2 Bz1 7 dx 7 dx 76 dv2 54 dσ 5 ¼ μ0 ∇2 Bz2 :  dy dx

(12)

This provides a pixel-by-pixel relation between the two-dimensional gradient of the conductivity and the Laplacian h iof the measured Bz data subject to two current injections. After computing

dσ dσ dx , dy

by solving the linear system of equations in

Eq. 12, applications of numerous integration techniques can produce a conductivity

Principles and Use of Magnetic Resonance Electrical Impedance Tomography in. . .

a

ε2−

ε1+

9

400

ε+2

ε1− 0

b

17.5 [nT]

-0.5

c

0.55 [A/m2]

0

Fig. 3 (a) MR magnitude image of a canine head with two pairs of current injection electrodes. (b) Two images of the measured Bz data subject to two current injections. (c) Magnitude of the projected current densities subject to the current injections at two different directions

image σ (Oh et al. 2003). At least one boundary voltage data or at least one known conductivity value preferably on a position at the boundary should be incorporated into the image reconstruction process to produce a unique conductivity image (Seo and Woo 2011; Seo and Woo 2013).

10

E.J. Woo and M. Kranjc

Fig. 4 MR magnitude images versus MREIT conductivity images of animals and human subjects: (a) canine brain, (b) canine chest, (c) canine abdomen, (d) canine pelvis, (e) human leg, and (f) human knee

There have been numerous experimental studies in MREIT using conductivity phantoms, animals, and human subjects (Woo and Seo 2008; Seo and Woo 2014). These results indicate that MREIT is capable of producing multi-slice cross-sectional images of a conductivity distribution inside an electrically conducting object such as the human body with a spatial resolution of an adopted MRI scanner. Conductivity image reconstructions of animals and human subjects with about 1 mm pixel size have been demonstrated using clinical 3 T MRI scanners. Figure 4 shows numerous conductivity images of animals and human subjects reconstructed by using the harmonic Bz algorithm. As mentioned earlier, some biological tissues such as the muscle and white matter exhibit anisotropy in their conductivity values. Though MREIT can reconstruct anisotropic conductivity tensor images in theory, its experimental demonstration has been deferred primarily due to the requirement of the measured Bz data with a very high signal-noise ratio (SNR) (Seo et al. 2004). To overcome this practical technical difficulty, Kwon et al. (2014) lately proposed a method called DT-MREIT to combine diffusion tensor imaging (DTI) with MREIT. The major technical hurdle remained in MREIT as a diagnostic imaging tool is the weak measurement sensitivity of Bz using a clinical MRI scanner for a given local conductivity change. The induced Bz is in the range of tens of nT when the injection current amplitude is a few mA. From Eqs. 7 and 8, the amount of the phase change

Principles and Use of Magnetic Resonance Electrical Impedance Tomography in. . .

11

subject to the externally injected current is proportional to the product of Bz and Tc that is the time duration of the current injection. Since Bz is directly proportional to the amplitude of the injection current, the phase change is basically proportional to the product of the amplitude and duration of the injection current pulse. Though there is no safety standard yet for this kind of imaging current pulses, the pulse amplitude and duration should be kept within a certain limit such as the threshold of perception or the threshold of muscle contraction. To improve the image quality in MREIT, multi-channel RF coils with better sensitivity, innovative MREIT pulse sequences, pre- and post-processing methods, and better image reconstruction algorithms without noise amplification are needed in addition to the performance improvement of the adopted MRI scanner itself. If a three-dimensional conductivity distribution is available for an electrically conducting object, one may numerically solve Eq. 3 to compute the voltage distribution v for any given current injection or voltage application. This leads to computations of the internal current density J as well as the electric field intensity E in Eq. 1. For those cases of electric stimulations, such as electroporation, cardiac defibrillation, transcranial dc stimulation (tDCS), deep brain stimulation (DBS), RF ablation, and so on, the knowledge of the internal conductivity distribution can be utilized in the treatment planning (see “▶ Treatment Planning for Electrochemotherapy and Irreversible Electroporation of Deep-Seated Tumors”) (Chauhan et al. 2013; Jeong et al. 2016).

MREIT in Tissue Electroporation Electroporation is used for therapeutic purposes and electric pulses can reach up to 3000 V, and they can establish electric field distribution with strength up to 150 kV/m depending on electrodes geometry and distance between them. As CDI and MREIT have been developed for diagnostic purpose and there used to be a lack of reports where electric pulses that are normally used in electroporation applications would be used. Still, in the last 5 years it was demonstrated that both, CDI and MREIT, can be applied for obtaining electric field with such high field strength. Determination of electric field distribution during tissue electroporation was at first demonstrated both experimentally and numerically on homogeneous and heterogeneous agar phantom with electrical properties similar to human tumor and surrounding tissue. A good agreement between experimental and numerical results was obtained in homogeneous and heterogeneous agar (Fig. 5) and for different pulse sequences, i.e., for different number and amplitude of pulses (Kranjc et al. 2011). Determination of electric field distribution during tissue electroporation was also successfully demonstrated on ex vivo chicken liver tissue (Kranjc et al. 2012) and in porcine muscle as shown in Fig. 6. Furthermore, experimental and numerical investigation on the anisotropy ratio of ex vivo tissue was also performed (Essone Mezeme et al. 2012). Alteration of anisotropy ratio of the conductivity tensor was detected when reversible electroporation threshold was exceeded. Experimental

12

E.J. Woo and M. Kranjc

a

c

b

|EMREIT| [V/m]

+ σ

× 104 18 14

CDI

MREIT

10

Uapp

6 2

d

|Enum| [V/m]

× 104 18 14 10 6 2

e

f

σ2

Uapp

CDI

g

|EMREIT| [V/m]

× 106 5.5 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5

|Enum| [V/m]

× 106 5.5 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5

MREIT

σ1 +

h + σ2 σ1

CDI Uapp

Fig. 5 Homogeneous (a) and heterogeneous phantom (e) exposed to four 100 μs long high voltage pulses of 1000 V delivered between diagonal electrodes. The initial phase images (b, f) were acquired by the two-shot rapid acquisition with relaxation enhancement (RARE) RF based CDI sequence. The electric field distributions (c, g) were calculated using MREIT J-substitution algorithm from the current density distribution. Numerical simulation of electric field distribution (d, h) in the phantom are shown for comparison below the corresponding experimental results (The figure is adapted from Kranjc et al. (2011))

Principles and Use of Magnetic Resonance Electrical Impedance Tomography in. . .

13

Fig. 6 Electric field mapping during current injection between a pair of needle electrodes inserted into a chunk of porcine muscle: (a) MR magnitude image where the inserted needle electrodes appear as two black dots, (b) measured Bz image subject to injection current between the pair of needle electrodes, (c) computed image of the projected current density, and (d) image of the magnitude of the electric field intensity

Fig. 7 Electric field distribution in a potato tuber (a), in an agar phantom (b), and in a tissue ex vivo (c). Pulses were delivered between two needle electrodes (marked with + and ˗)

results agreed with numerical and were also consistent with experimental investigations performed by other research groups. Electric field distribution was also successfully applied in monitoring of electric field distribution during application of electric pulses in potato tubers (Kranjc et al. 2016). For comparison, electric field distribution in an agar phantom and in ex vivo tissue are shown together with the distribution in potato tuber on Fig. 7. The study was performed on potato tubers cultivar “Agata” using eight electric pulses with an amplitude of 1000 V, pulse duration of 100 μs, and repetition frequency of 5 kHz. Electric pulses were delivered between two needle electrodes. Electric field in the potato tuber was not distributed symmetrically (A) as in the agar phantom (B), presumably due to the heterogeneous potato structure and consequent heterogeneous electrical conductivity, which resulted in observed asymmetric distribution. More inhomogeneous distribution than in agar was also obtained in

14

E.J. Woo and M. Kranjc

Fig. 8 (a) The electric field distribution in the tumor obtained by MREIT superimposed to the T1-weighted image acquired before the application of electric pulses. A white contour line encloses an area exposed to an electric field strength between reversible (400 V/cm) and irreversible (900 V/cm) electroporation threshold values. Tumor cells located outside the area are either irreversibly electroporated (the area close to the electrodes) or remain unelectroporated (the area toward the tumor boundary). (b) Scatterplot of the coverage of five tumors with the electric field of reversible electroporation CMREIT and Gd-DOTA cell entrapment FGd-DOTA

an ex vivo tissue (C), where heterogeneous distribution can also be observed, although not as apparent as in the potato tuber. Since MREIT can be used in reconstruction of electric field in tissues, there are no limitations in applying it on food samples undergoing the PEF treatment. Monitoring of electric field can be used either for general investigations on distribution of electric field in various food samples or for fine adjustments of PEF treatment parameters, such as amplitudes of electric pulses or changing their number, during the PEF treatment. Investigation of the feasibility of MREIT for in situ monitoring of the electric field distribution during in vivo reversible electroporation was performed on mouse tumors (Kranjc et al. 2015). Animal experiments were conducted in accordance with the guidelines for animal experiments of the European Union directives, and permission was obtained from the Ministry of Agriculture and the Environment of the Republic of Slovenia (permission no. 34401-43/2011/5). Electroporation was performed by applying two sequences of four high voltage electric pulses with an amplitude of 500 V, duration of 100 μs, and at a pulse repetition rate of 5 kHz via two needle platinum-iridium electrodes inserted into the tumor. Electric field distribution in the tumor during pulse delivery was obtained by means of MREIT J-substitution algorithm. For post-treatment assessment of reversibly electroporated areas in the tumor, the contrast agent gadoliniumtetraazacyclododecane tetraacetic acid (Gd-DOTA) was injected before the application of pulses. After 24 h, T1-weighted images were acquired for observation of tumor areas where Gd-DOTA molecules were entrapped inside the reversibly electroporated tumor cells. An example of electric field distribution obtained by means of MREIT is shown in Fig. 8a where it is overlaid to T1-weighted image acquired just before the application of electric pulses. As expected, the electric field was the highest around

Principles and Use of Magnetic Resonance Electrical Impedance Tomography in. . .

15

the electrodes where it exceeded irreversible electroporation threshold (900 V/cm), while it remained under the reversible electroporation threshold (400 V/cm) in the areas away from electrodes. The average coverage of tumors with electric field leading to reversible electroporation of tumor cells CMREIT was calculated by dividing the predicted surface area of the reversibly electroporated tumor cells with the surface area of the entire tumor. Whereas tumor fraction FGd-DOTA was calculated by dividing the surface area of reversibly electroporated tumor cells as obtained by entrapped contrast agent with the surface area of the entire tumor. CMREIT  SD and FGd-DOTA  SD were 38  9 % and 41  13 %, respectively. Correlation was evaluated between CMREIT and tumor fractions with FGd-DOTA with linear Pearson correlation analysis. Coverage of tumors with electric field in the range of 400–900 V/cm, where reversible electroporation is expected, had good correlation with Gd-DOTA cell entrapment (r = 0.956, P = 0.005) as shown in Fig. 8b. A concern whether proposed method for determination of electric field distribution can be implemented in electroporation applications was addressed by a simulation in the case of a three-dimensional (3-D) numerical model designed for the purpose of electrochemotherapy treatment of deep-seated liver tumors. The treatment was done as part of an on-going Phase I/II clinical study (EudraCT number 2008-008290-54; clinicaltrials.org – NCT01264952). The study was approved by Institutional Medical Board and Ethical Committee of the Republic of Slovenia. Briefly, a model of a patient with a metastasis located between the inferior vena cava and the main hepatic veins was studied. The model included a 3-D geometry of the metastasis that was built by means of segmented MRI images of the patient. As shown on Fig. 9, it was demonstrated that it is possible to obtain sufficiently accurate information on electric field distribution in the targeted and surrounding tissue by measuring only one component of magnetic flux density and thus enable detection of areas with insufficient electric field coverage before the end.

Conclusion Exposure of the treated tissue to a sufficiently large electric field presents one of the most important conditions for successful electroporation. A monitoring method that would allow determination of the electric field would be of great importance for electroporation based applications such as electrochemotherapy, irreversible electroporation tissue ablation, and pulsed electric field processing. In recent studies, it was demonstrated by means of experimental and numerical approaches that magnetic resonance electrical impedance tomography together with current density imaging indeed can be used for determination of electric field distribution during electroporation pulse delivery. As there is a lack of tissue-specific experimental data on tissue properties for reliable numerical treatment planning of electroporation based clinical applications, MREIT could be of significant help in obtaining more accurate electrical conductivity values.

E.J. Woo and M. Kranjc

b a

E [kV/cm]

16

2 1.8 1.6

+

1.4

5

4

0.6

2

5 mm

10 mm

1 0.8

3

0 mm

–

1.2

1

6

0.4 0.2

y x z

0 0

1 2 3 4 tumor-liver boundary

5 x [mm]

Fig. 9 A 3-D numerical model of a deep-seated tumor in a liver (a). The tumor (in red) was located between the inferior vena cava and main hepatic veins (all in blue). The liver tissue, surrounding the tumor and veins, is not shown for the purpose of visualization. Electrodes (in grey) are labeled with numbers from 1 to 6. An electric field distribution across the tumor-liver region (between electrodes no. 4 and 6) obtained by means of the MREIT algorithm using only one magnetic field component (solid line) and the corresponding true electric field calculated by the numerical model (dashed line) (b). Tumor-liver boundary is marked with x. (The figure is adapted from Kranjc et al. (2012))

Still, the main difficulty of using MREIT to determine electric field distribution during electroporation is associated with the limited capability of MRI scanners for their use during application of electroporation such as electrochemotherapy and irreversible electroporation tissue ablation. Hopefully, this could be surpassed in the near future using open MRI scanners. Conductivity changes that occur during the pulse are at the moment also too fast to be estimated with MREIT as a function of time. Nevertheless, it is important to be aware that the cumulative effect of electric current on the MRI signal phase is measured. Therefore, this technique yields an electric field distribution, which is a time average of its altering time course so that all the consequences of conductivity alteration (see “▶ Electric Field Distribution Modelling in Tissue Considering Tissue Conductivity Increase Due to Electroporation”), which affect electric current, are not neglected within obtained electric field distribution.

Cross-References ▶ Current Density Imaging as Means to Follow Tissue Electroporation ▶ Electric Field Distribution and Electroporation Threshold ▶ Electric Field Distribution Modelling in Tissue Considering Tissue Conductivity Increase Due to Electroporation ▶ Treatment Planning for Electrochemotherapy and Irreversible Electroporation of Deep-Seated Tumors

Principles and Use of Magnetic Resonance Electrical Impedance Tomography in. . .

17

References Chauhan M, Jeong WC, Kim HJ et al (2013) Radiofrequency ablation lesion detection using MR-based electrical conductivity imaging: a feasibility study of ex vivo liver experiments. Int J Hyperth 29:643–652. doi:10.3109/02656736.2013.842265 Essone Mezeme M, Kranjc M, Bajd F et al (2012) Assessing how electroporation affects the effective conductivity tensor of biological tissues. Appl Phys Lett 101:213702–213704. doi:10.1063/1.4767450 Grimnes S, Martinsen OG (2014) Bioimpedance and bioelectricity basics, 3rd edn. Academic, London Holder DS (ed) (2004) Electrical impedance tomography: methods, history and applications, 1st edn. Taylor & Francis, London Jeong WC, Sajib SZK, Oh TI et al (2016) Electric field mapping in ex vivo anisotropic muscle tissue using DT-MREIT. Proc. 1st World Congress on Electroporation and Pulsed Electric Fields in Biology, Medicine and Food & Environmental Technologies 71–74 Joy M, Scott G, Henkelman M (1989) In vivo detection of applied electric currents by magneticresonance imaging. Magn Reson Imaging 7:89–94 Kranjc M, Bajd F, Sersa I, Miklavcic D (2011) Magnetic resonance electrical impedance tomography for monitoring electric field distribution during tissue electroporation. IEEE Trans Med Imaging 30:1771–1778. doi:10.1109/TMI.2011.2147328 Kranjc M, Bajd F, Sersa I et al (2012) Ex vivo and in silico feasibility study of monitoring electric field distribution in tissue during electroporation based treatments. PLoS One 7:e45737. doi:10.1371/journal.pone.0045737 Kranjc M, Markelc B, Bajd F et al (2015) In situ monitoring of electric field distribution in mouse tumor during electroporation. Radiology 274:115–123. doi:10.1148/ radiol.14140311 Kranjc M, Bajd F, Serša I et al (2016) Electric field distribution in relation to cell membrane electroporation in potato tuber tissue studied by magnetic resonance techniques. Innovative Food Sci Emerg Technol. doi:10.1016/j.ifset.2016.03.002 Kwon O, Woo EJ, Yoon J-R, Seo JK (2002) Magnetic resonance electrical impedance tomography (MREIT): simulation study of J-substitution algorithm. IEEE Trans Biomed Eng 49:160–167. doi:10.1109/10.979355 Kwon OI, Jeong WC, Sajib SZ et al (2014) Anisotropic conductivity tensor imaging in MREIT using directional diffusion rate of water molecules. Phys Med Biol 59:2955–2974. doi:10.1088/ 0031-9155/59/12/2955 Lee BIL, Oh SH, Woo EJ et al (2003) Three-dimensional forward solver and its performance analysis for magnetic resonance electrical impedance tomography (MREIT) using recessed electrodes. Phys Med Biol 48:1971–1986 Malmivuo J, Plonsey R (1995) Bioelectromagnetism: principles and applications of bioelectric and biomagnetic fields. Oxford University Press, Oxford Oh SH, Lee BIL, Woo EJ et al (2003) Conductivity and current density image reconstruction using harmonic Bz algorithm in magnetic resonance electrical impedance tomography. Phys Med Biol 48:3101–3116 Park C, Lee BIL, Kwon OI (2007) Analysis of recoverable current from one component of magnetic flux density in MREIT and MRCDI. Phys Med Biol 52:3001–3013. doi:10.1088/0031-9155/52/ 11/005 Seo JK, Woo EJ (2011) Magnetic resonance electrical impedance tomography (MREIT). SIAM Rev 53:40–68. doi:10.1137/080742932 Seo JK, Woo EJ (2013) Nonlinear inverse problems in imaging. Wiley, Chichester Seo JK, Woo EJ (2014) Electrical tissue property imaging at low frequency using MREIT. IEEE Trans Biomed Eng 61:1390–1399. doi:10.1109/TBME.2014.2298859

18

E.J. Woo and M. Kranjc

Seo JK, Yoon J-R, Woo EJ, Kwon O (2003) Reconstruction of conductivity and current density images using only one component of magnetic field measurements. IEEE Trans Biomed Eng 50:1121–1124. doi:10.1109/TBME.2003.816080 Seo JK, Pyo HC, Park C et al (2004) Image reconstruction of anisotropic conductivity tensor distribution in MREIT: computer simulation study. Phys Med Biol 49:4371–4382 Seo JK, Woo EJ, Ulrich K, Wang Y (2014) Electromagnetic tissue properties MRI. Imperial College Press, London Woo EJ, Seo JK (2008) Magnetic resonance electrical impedance tomography (MREIT) for highresolution conductivity imaging. Physiol Meas 29:R1–R26. doi:10.1088/0967-3334/29/10/R01 Woo EJ, Lee SY, Mun CW (1994) Impedance tomography using internal current density distribution measured by nuclear magnetic resonance. Proc SPIE 2299 Math Methods Med Imaging III 2299:377–385 Zhang N (1992) Electrical impedance tomography based on current density imaging. University of Toronto, Toronto

Current Density Imaging as Means to Follow Tissue Electroporation Igor Serša and Franci Bajd

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . From Electric Current Pulses to CD Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DC-CDI Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CDI Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroporation CDI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroporation CDI Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 5 5 8 10 12 14 14 15 15 16 19 20 20

Abstract

An expanding use of electroporation in everyday clinical practice necessitates development of accurate and preferably noninvasive methods for assessment of electroporation outcome. Electroporation outcome is highly dependent on tissue coverage with sufficiently high electric fields that can be reconstructed by the magnetic resonance electric impedance tomography, which uses current density distribution data and electric potentials at the electrodes for reconstruction of electric field in the sample. Current density imaging (CDI) is a magnetic resonance imaging (MRI) modality for noninvasive assessment of current density

I. Serša (*) • F. Bajd (*) Jožef Stefan Institute, Ljubljana, Slovenia Faculty of Mathematics and Physics, University of Ljubljana, Ljubljana, Slovenia e-mail: [email protected]; [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_66-1

1

2

I. Serša and F. Bajd

through phase information and is therefore a method of choice for monitoring of electroporation efficacy. CDI consists of the current encoding part in which electric current due to a set of delivered electric pulses is encoded in a signal phase and of the imaging part in which the encoded signal is acquired. In the overview, CDI is presented as an efficient method for a use in electroporation applications. With respect to these applications, different k-space traversing strategies employed in the imaging part of CDI are discussed. While the most common strategies are based on standard spin-echo and gradient-echo pulse sequences in which each k-space line traversal is preceded by a new set of electric pulses, advanced strategies employing single-shot rapid acquisition with relaxation enhancement and echo planar imaging include minimal number delivered electric pulses for entire k-space traversal. The strategies are discussed also with respect to parameters, such as an effect of cumulative duration of electric pulses on CDI sensitivity, SAR, SNR performance, and susceptibility of CDI to artifacts. In addition, a reconstruction algorithm for current density distribution using Ampere’s law and the corresponding limitations associated with the need for sample rotations for a determination of three current density spatial components are discussed as well.

Keywords

Magnetic resonance imaging • Current density imaging • Electric currents • Tissue conductivity • Electroporation

Introduction In electrochemotherapy (ECT) (Marty et al. 2006, Miklavcic et al. 2014) and irreversible electroporation (IRE) tissue ablation (Lee et al. 2010, Neal et al. 2011), which are two widely used clinical applications of electroporation for solid tumor treatment, temporary cell membrane permeabilization, induced by applying short high-intensity electric field pulses, is followed by the cell death due to the cell uptake of an anticancer drug (ECT) or due to the cell exposure to an excessively high electric field (IRE) (Yarmush et al. 2014). The efficacy and duration of permeabilization are dependent on parameters such as duration, frequency, and number of the pulses as well as on electrode geometry, number of the electrodes, and their placement. Accurate coverage of the treated tissue with sufficiently large electric field that is established during an application of electroporation pulses plays an important role for a successful electroporation outcome (Miklavcic et al. 1998). Therefore, effective and preferably noninvasive methods are needed for assessment of the parameters maximizing tissue coverage with electric field. For a reconstruction of electric field distribution, magnetic resonance electrical impedance tomography (MREIT) was suggested and recently demonstrated in an agar phantom (Kranjc et al. 2011), in liver tissue ex vivo (Kranjc et al. 2014), in silico (Kranjc et al. 2012), as well as during electroporation of mouse tumor in vivo

Current Density Imaging as Means to Follow Tissue Electroporation

3

(Kranjc et al. 2015). The MREIT method is based on current density imaging (CDI), which is an MRI modality developed for a detection of current density distribution inside a conductive sample. The detection is enabled through temporary currentinduced magnetic field spatial variations that are manifested in MR phase alterations. From the alterations, current density is calculated by using Ampere’s law. In the last two decades, CDI has become an increasingly important tool for studying electrical conductivity properties of different biological tissues in various frequency ranges (DC, AC, and RF). Recently, the CDI method along with the MREIT algorithms was successfully applied also for measurements of tissue conductivity tensor (Mezeme et al. 2012, Kwon et al. 2014b). Moreover, the combined approach was successfully applied as a monitoring tool and for guidance in treatment by RF ablation (Kwon et al. 2014a). CDI is enabled by the use of electric pulses that are synchronized with an imaging sequence. During the electric pulses, electric current flows through the sample between the electrodes that are attached to the sample or inserted in it. The CDI sequence can in principle be of any type of a MR sequence with a characteristic that it can hold phase information accumulated during application of electric pulses. In the simplest form, the CDI sequence can be a modified spin-echo sequence (Joy et al. 1989), or it can be a sequence based on gradient-echo imaging (Kwon et al. 2014a). In these two forms of the sequences, the k-space (Callaghan 1991) is traversed in consecutive lines, and for each line or a few lines, a new signal excitation is needed with a new electric pulse or a set of electric pulses. More sophisticated sequences include a spiral acquisition (Yan et al. 1997), echo planar (EPI) acquisition (Serša et al. 2015), or rapid acquisition with relaxation (RARE) (Sersa 2008). Last three forms of the CDI sequence are of a fast acquisition type and can be run also in a single-shot mode (spiral and EPI) or in a two-shot mode (RARE). Therefore, only one signal excitation or at most two signal excitations followed by an application of an electric pulse or a set of pulses is needed. In CDI, the phase information is related to the magnetic field created by currents flowing through the sample during the application of electric pulses. The information is then used to calculate magnetic field change induced by the currents and finally to calculate current density from the magnetic field change by using Ampere’s law. The outcome of CDI depends highly on the conductivity properties of the sample. With highly conductive samples, sufficient MR phase differences are obtained with relatively low voltages in the range from few volts up to few tens of volts and with moderated pulse durations of typically few tens of milliseconds. However, such conductivities are obtained mostly only in electrolytes, where conductivity of a sample can be adjusted simply by changing concentration of salts or by changing pH of the sample. Unfortunately, most tissues are not that conductive, so voltages of the pulses must be increased or if that is not possible their duration must be increased. The first can damage the sample and cause pain if applied in vivo, while the second results in poor SNR of the obtained images due to the spin-spin NMR relaxation of the sample. The theory of CDI sensitivity presented in Scott et al. (1992) and Sersa et al. (1994) reveals that the optimal sensitivity of the CDI experiment is obtained when the current application time is equal to T2 relaxation

4

I. Serša and F. Bajd

Fig. 1 The sensitivity curve, i.e., sensitivity of a generic CDI pulse sequence as a function of cumulative electric pulse duration tc. Relatively low sensitivities are obtained with short and highvoltage pulses (tc < T2) that are typically used in electroporation applications (ECT and IRE) as well as with long and low-voltage pulses (tc > T2) that are exploited for noninvasive determination of electrical conductivity properties of tissues in vivo. In CDI, the optimal sensitivity is obtained with cumulative duration of electric pulses equal to spin-spin relaxation time (tc = T2)

time of the sample. In reality, for most tissues, T2 is of the order of 100 ms. This dictates the use of voltages of one order of magnitude higher than in electrolytes (from few tens of volts to few hundred volts) for a sufficient effect of the currents on the signal phase shift (Beravs et al. 1997, Sersa et al. 1997) and consequently on a sufficient CDI quality. Another extreme of the CDI use is electroporation. In electroporation, electric pulses are short, and they are usually used in series of few pulses. For example, a common configuration is eight pulses of 100 μs with voltages from 500 V to 1500 V separated by 100 μs long pauses in between the pulses (Yarmush et al. 2014). Obviously, in electroporation applications, electric pulses are cumulatively about two orders of magnitude shorter than when used in a non-damaging configuration for tissue imaging, and their voltage is approximately one order of magnitude higher. A use of CDI in electroporation is therefore in a low sensitivity regime, however on the other side of the sensitivity curve as in standard in vivo CDI applications (Fig. 1). An advantage of CDI in electroporation is that there is practically no effect of signal relaxation during application of electric pulses, which reduces imaging noise. However, electroporation pulses cannot be applied during image acquisition many times as they significantly alter electrical properties of the sample during imaging. For that reason, spin-echo- and gradient-echo-based CDI sequences are not appropriate for electroporation CDI experiments. However, RARE, EPI, or spiral sequences are due to their single- or two-shot imaging ability appropriate candidates for the use. There is also another limitation that could prevent the use of the fast CDI sequences and that is NMR relaxation. All three sequences (RARE, EPI, and spiral) require a long T2 relaxation time of the sample for their successful performance. This

Current Density Imaging as Means to Follow Tissue Electroporation

a

5

b

Fig. 2 (a) Schematic presentation of an electric current-carrying straight wire and magnetic field around it. Magnetic field Bc makes a circular path around the wire, and its amplitude decreases inversely proportional with a distance from the wire. In CDI, similarly, currents between the electrodes create magnetic field variations that are encoded into phase information. (b) In CDI, only magnetic field variations in the direction of the static magnetic field B0∙ez contribute to currentinduced phase variations. By using Ampere’s law, 2D current density j = ( jx, jy) within a slice perpendicular to the static magnetic field can be calculated from these phase variations. Information on electric current distribution is therefore most efficiently stored into phase variations when the electroporation electrodes are inserted in parallel with the static magnetic field and the imaging slice is perpendicular to the static filed and the electrodes

is because they all need long signal to enable entire images signal acquisition after just one excitation. In this overview, typical CDI pulse sequences for determination of electric current distribution in tissues during delivery of electroporation pulses are presented. The sequences are classified according to their SNR performance, which highly determine the quality of calculated electric fields and electric conductivity maps through their reconstruction by MREIT algorithms, as well as according to their specific absorption rate (SAR) performance and adequacy for monitoring of different electroporation applications (ECT vs. IRE).

Theory From Electric Current Pulses to CD Image One of the effects caused by electric currents is also a change of the magnetic field around a current-carrying sample (Fig. 2a). Current I that flows along a straight wire produces a magnetic field Bc that makes a circular path around the wire. Its amplitude is inversely proportionally to the radial distance from the wire Bc ð r Þ ¼

μ0 I : 2πr

(1)

6

I. Serša and F. Bajd

In a CDI experiment, a sample is in the MRI magnet, and it has electrodes attached to it or inserted in it. Then, voltage is applied to the electrodes which results in currents flowing through the sample. The pattern of the currents in the sample can be of course much more complex than in Fig. 1, but the end effect is similar; in the sample, additional magnetic field Bc is created. The additional field is superimposed to the statistic magnetic field of the magnet B0. As B0 » Bc, other components of Bc than the component along B0 have practically no effect on the precession frequency of the nuclei, while the z-component changes their precession (Larmor) frequency by Δω ¼ γBcˍ B0 :

(2)

If the currents are constant, then the corresponding magnetic field shift is constant too. Therefore, the currents applied in a pulse of duration tc will induce a precession phase shift of φB0 ¼ γBcˍ B0 tc :

(3)

The phase shift can then be used to calculate the corresponding magnetic field change. As the magnetic field change is a vector, it has therefore three components. All three can be measured in separate experiments by rotating the sample along with the electrodes to three perpendicular orientations so that in a coordinate system attached to the sample in each of the orientations, one of the three coordinates is aligned with the static magnetic field Bcˍ x ¼

φx γtc

Bcˍ y ¼

φy γtc

Bcˍ z ¼

φz : γtc

(4)

Once all three components of the magnetic field change are measured, current density in the sample can be calculated by using Ampere’s law ⇀

j¼

⇀ 1 ∇  Bc: μ0

(5)

Rotating the sample to the perpendicular orientations to obtain all three components of the magnetic field change is very impractical and also often not possible. Therefore, it is desirable to use in CDI such arrangement of the electrodes with respect to the image orientation that there is no need for the rotations and repeated experiments. One such arrangement is obtained with a sample where currents flow in a relatively thin slice which is also the imaging slice (Fig. 2b), and the electrodes are positioned perpendicularly to the slice. With the arrangement, Bc_x  0 and Bc_y  0 so that current density in the slice is calculated only from the component Bc_z  
 1 @Bcˍ z @Bcˍ z , jx , jy  : μ0 @y @x

(6)

Current Density Imaging as Means to Follow Tissue Electroporation

7

The procedure of CDI experiment, measurement and data processing, is presented in Fig. 3.

DC-CDI Sequence As already mentioned in the introduction section, CDI is enabled by the use of an imaging sequence that during the CD encoding part triggers a current pulse to the sample and then enables registration of the phase shift produced by the currents (Eq. 3). The simplest sequence which meets these criteria is a spin-echo sequence with superimposed two electric pulses (Fig. 4a). The first electric pulse is applied before the refocusing RF pulse, and the second electric pulse of the same duration and amplitude but of an opposite polarity is applied after it. The second pulse must have the opposite polarity because the refocusing RF pulse in the sequence reverses the phase shift of the first electric pulse. The reversed phase shift of the first pulse can then co-add with the phase shift of the second pulse. The spin-echo CDI sequence in Fig. 4a belongs to a groups of DC-CDI (direct current) or LF-CDI (low frequency) sequences due to relatively long application time of currents tc in only two electric pulses. In Fig. 4b another version of a DC-CDI sequence is shown, namely, the two-shot RARE CDI sequence. With the two-shot RARE CDI sequence, equal electric pulses can be used as with the spin-echo CDI sequence; however, signal acquisition strategy with this sequence is significantly different. In the spin-echo CDI sequence, image signal is scanned in the k-space line by line. For each line, a signal is excited, and both current pulses are applied. For a standard image with N = 256 lines, 256 pairs of electric pulses are delivered to the sample. This is not only timeconsuming but could also result in alteration of sample conductivity if the applied voltage of current pulses is high. To avoid this problem, the number of delivered electric pulses must be reduced. Ideally, only one electric pulse delivered to the sample is sufficient to obtain one CD image. In the single-shot RARE sequence, all image signal is acquired only after one signal excitation, which is enabled by repetitive scanning of k-space lines in consecutive echoes. Unfortunately, as shown in (Sersa 2008), the sequence becomes unstable when current-induced phase shift is encoded at the beginning of the sequence. However, the problem of instability can be solved by running the sequence twice, which is followed by a co-addition of the corresponding signals. The sequence is therefore called the two-shot RARE CDI sequence. In the first run, phases of the refocusing RF pulses are identical to the phase of the excitation RF pulse, while in the second run, the phases of the refocusing RF pulses are increased by 90 .

CDI Sensitivity Sensitivity of a CDI sequence can be considered as the reciprocal noise of the measured current density (Scott et al. 1992; Sersa et al. 1994). The noise has an

8

I. Serša and F. Bajd

Fig. 3 Magnitude MR image of a conductive phantom shows no indication for electric current flowing between the electrodes that are inserted close to the phantom boundary. However, the effect can be clearly seen in the corresponding MR phase image (with modulus 2π) by a characteristic nonuniform pattern of phase variation. The pattern is a consequence of encoded magnetic field variations due to electric current flowing through a conductive material under a potential difference (voltage V and duration tc) between the inserted electrodes. From the phase pattern, experimentally obtained current density distribution is calculated in a pixelwise manner by using Ampere’s law. Finally, the experimentally obtained current density is compared with the simulated electric current density

Current Density Imaging as Means to Follow Tissue Electroporation

9

a

b

Fig. 4 (a) Standard spin-echo-based CDI pulse sequence with electric pulses inserted between the excitation and refocusing RF pulses. Each electric pulse is of duration of tc/2, thus resulting to cumulative duration of tc. Polarity of the second electric pulse is reversed due to a sign-reversing character of the refocusing RF pulse. A drawback of the sequence is that a new pair of electric pulses is needed prior to acquisition of each k-space line. (b) Two-shot RARE CDI pulse sequence, in which the current encoding part is followed by multiple signal acquisitions traversing the entire k-space. In order to eliminate image artifacts, the final MR phase image is obtained by co-adding signals of two single-shot RARE CDI sequences, each with different phases of the refocusing pulses ([x] and [y])

origin in noise of the conventional image, but also depends on the signal magnitude (Signal). In the complex plane, the MR signal can be represented by a point. However, due to noise, the point is smeared in both directions, in real and in imaginary direction (Fig. 5). Distribution profile of the signal along the two directions is a normal one with a standard deviation (RMS noise) equal to σ. The ratio between the RMS noise and the signal magnitude therefore represents uncertainty in the signal phase and can therefore be considered as the phase noise

10

I. Serša and F. Bajd

Fig. 5 In the complex plane, MR signal can be represented by a point of which probability is due to RMS noise of its real and imaginary signal components smeared in two-dimensional normal distribution with a standard error σ. The uncertainty of the point’s angle in the complex plane, i.e., the phase noise, is therefore equal to σ φ ¼ σ=Signal, where Signal is the signal magnitude

pffiffiffi 2 σ 1 ¼ ; σφ ¼ f Signal f SNR

(7)

where SNR is a signal-to-noise ratio of the conventional magnitude image and factor f comes into the formula due to possible different signal manipulations needed to eliminate the background phase. This can be eliminated by subtracting a phase image without application of currents ( f = 1) or with reversed currents ( f = 2) from the phase image obtained with currents; if no subtraction is used, f = √2. The phase noise determines also the magnetic field noise σ Bc ¼

1 γ tc f SNR

(8)

and ultimately, via the relation for calculation of a current density component in discrete space jzˍ i, j ¼

 N  Byˍ iþ1, j  Byˍ i, j  Bxˍ i, jþ1 þ Bxˍ i, j μ0 FOV

(9)

also the current density noise σj ¼

2N : μ0 FOV f γ tc SNR

(10)

Here FOV is image field of view and N by N is image matrix. From Eq. 10 can clearly be seen that CDI sensitivity, i.e., reciprocal σ j, is proportional to a product of the

Current Density Imaging as Means to Follow Tissue Electroporation

11

current application time (cumulative time of all current pulse durations) tc and signalto-noise ratio of the conventional magnitude image Sensitivity 

1 / tc SNR: σj

(11)

For the DC-CDI pulse sequences (Figs. 4a, b), SNR is proportional to exp(tc/T2) so that CDI sensitivity is therefore proportional to Sensitivity / tc expðtc =T 2 Þ:

(12)

Equation 12 reaches maximum when tc = T2, which is also the condition for the optimal current application time in DC-CDI experiment as well as in all other CDI experiments where signal magnitude decays exponentially with a decay rate equal to T2. Dependence on CDI sensitivity on the current application time is presented in Fig. 1. The graph in the figure clearly shows how CDI sensitivity decreases when tc is either longer or shorter than the optimal value of tc = T2. The first case is often found in in vivo CDI experiments where current amplitudes are limited to avoid tissue damage and pain, while electroporation CDI, where current application times are very short, is a good example of the second case. Current density signal-to-noise ratio SNRj can be considered as the ration between the current density in the samples and current density noise. According to Eq. 11, this is equal to SNRj 

j f γ μ0 FOV j tc SNR : ¼ σj 2N

(13)

From Eq. 13 can be seen that SNRj is proportional to pixel size FOV/N (inversely proportional to image resolution) to a product of current density and current application time and to the signal-to-noise ratio of the conventional magnitude image.

Electroporation CDI As can be seen from Eq. 13, to keep reasonable level of SNRj in CD images, a product of current density and current application time must be kept at least constant or better increasing with tc to compensate T2 relaxation signal losses. This means that in in vivo low-current CDI applications, tc needs to be long. Best SNRj is then obtained when tc = T2; however, in electroporation experiments, this is not possible as tc is fixed by the electroporation protocol, usually to time below 1 ms. With current application times that short, T2 relaxation can be neglected so that SNRj can be kept constant when a product of current density and current application time is constant. The effect of current application time shortening can then be compensated by increasing current density proportionally.

12

I. Serša and F. Bajd

As application of each electroporation pulse can change tissue properties additionally, progress of electroporation treatment can be followed only with those methods that enable immediate analysis of electroporation effects on the treated tissue. In case of CDI, this implies that a CD image must be obtained after only a few electroporation pulses delivered to the sample. This condition can be met by the use of the two-shot RARE CDI sequence (Fig. 6a). Note that Figs. 4a and 6a, which both show the two-shot RARE CDI sequence, differ only on different electric pulses used in the sequence. The sequence in Fig. 6a uses electroporation pulses which are all of the same polarity, and therefore all need to be included in the same interval, between the excitation and refocusing RF pulse of after the refocusing RF pulse and before formation of the first echo. Another important aspect of the two-shot RARE CDI sequence is the k-space traversal order. The lines can be ordered very differently (sequentially, centric, reversed centric, etc.), which in conventional MR image leads to different image contrasts as, for example, proton-density-weighting vs. T2weighting. In case of CDI, the contrast is not important; however, accuracy of the measured phase shift is important. The latter can be maximized by using protondensity-weighting, which corresponds to centric ordering of the scanned k-space lines (Fig. 7). With the centric ordering, lines from the k-space center that contribute the most signal to the image are scanned first, i.e., immediately after the excitation RF pulse, and are therefore not affected much by the T2 relaxation. Therefore, the obtained image has more signal, and phase error is the lowest. The centric k-space ordering is especially important with samples that have short T2 relaxation times. Another sequence that can be used in electroporation CDI is the EPI CDI sequence (Fig. 6b). In an EPI sequence, k-space is traversed in a “zigzag” trajectory, thus enabling acquisition of the entire image signal after just one signal excitation. The sequence with such signal sampling is known as a single-shot EPI sequence. Electroporation pulses in the EPI CDI sequence can be identical to those used in the two-shot RARE CDI sequence and are added between the excitation and refocusing RF pulse. Unlike in the RARE sequence, the additional phase shift produced by the electroporation pulses does not destabilize the single-shot EPI sequence so that the EPI CDI sequence, i.e., a standard single-shot EPI sequence with superimposed electroporation pulses, is a single-shot sequence. Therefore, with the EPI CDI sequence, double number of images can be acquired with the same number of electroporation pulses as with the two-shot RARE CDI sequence. This is an important advantage when high temporal resolution of electroporation monitoring is of a high priority. In addition, the EPI CDI sequence uses only two RF pulses: an excitation RF pulse and a refocusing RF pulse. This is convenient for in vivo applications where SAR is an important issue. A standard EPI sequence can be used in two versions: with the refocusing RF pulse or without it. The first version yields a T2-weighted image due to the spin-echo formation during reading the signal from the k-space center, while the second version yields a T2*-weighted image. For the CDI use, the spin-echo EPI version is better as it gives more signal and therefore produces results with lower current density noise. Despite all undeniable advantages of the EPI CDI sequence in comparison to the two-shot RARE sequence, it has also an important drawback, namely, it is highly susceptible to different imaging artifacts.

Current Density Imaging as Means to Follow Tissue Electroporation

13

a

b

Fig. 6 (a) Modified RARE CDI pulse sequence with the current encoding part containing a set of eight high-voltage electric pulses that are typically used in electroporation applications. (b) EPI CDI pulse sequence with a set of eight high-voltage electric pulses included in the current encoding part of the sequence that is followed by the EPI imaging part. In comparison to the two-shot RARE CDI sequence, this sequence exhibits comparatively lower SAR, but is more susceptible to artifacts

Methods Sample Preparation The performance of the two-shot RARE CDI and EPI CDI sequences was tested on a beef liver tissue. The liver tissue was obtained from a slaughterhouse, as meat product for human consumption. The slaughterhouse operates in accordance to Slovenian law, and the process of slaughtering is regulated by rules on animal protection and welfare at slaughter (Ur. l. RS, N. 5/2006), which ensures ethical standards of slaughtering procedure and is in compliance with European Union

14

I. Serša and F. Bajd

Fig. 7 A graphic presentation of two different k-space traversing strategies of the two-shot RARE CDI sequence: sequential k-space traversal with maximal/minimal MR signal at the ky edges and centeric k-space traversal with maximal MR signal in the center of the k-space and minimal MR signal at the ky edges. A decay of the MR signal due to spin-spin relaxation is indicated by a gradual color transition from black to light gray

Council directive on the protection of animals at the time of slaughter or killing (93/119/EC). Temperature of the liver tissue was maintained at 4  C before beginning of the experiment and then allowed to warm to the room temperature. The liver tissue was sectioned to flat cylindrically shaped samples with a diameter of 21 mm and height of 10 mm and then placed in a plastic container (Fig. 8). Two needle platinum-iridium electrodes with a 1 mm diameter were inserted at an inter-electrode distance of 14 mm in the sample in orientation perpendicular to the sample slice and in parallel to each other. The sample was exposed to high-voltage electric pulses that are typically used in irreversible electroporation (IRE). Altogether 90 pulses of 3000 V amplitude, 1 Hz repetition rate, and 100 μs (RARE-type of sequence) or 300 μs (EPI-type of sequence) duration and were delivered to the sample. Electric pulses were delivered using customized Cliniporator Vitae (IGEA, Carpi, Italy) pulse generator. The current of electric pulses was measured with an oscilloscope (WavePro 7300A, LeCroy, USA) using current probe (AP015, LeCroy, USA). All experiments were repeated three times. The sample was replaced with a fresh one after each electroporation pulse delivery to ensure identical initial conditions in all electroporation experiments.

Electroporation CDI Implementation Electroporation CDI experiments for IRE monitoring were performed on a 2.35 T horizontal bore small animal MRI scanner. The scanner was based on an Oxford superconducting magnet (Oxford Instruments, Abingdon, UK), an Apollo NMR/MRI spectrometer (Tecmag Inc., Houston TX, USA), and MRI probes for MR microscopy (Bruker, Ettlingen, Germany). The sequences were run with identical geometric and resolution parameters: field of view 30 mm, imaging matrix 64 by 64, slice thickness 4 mm, and repetition time 1 s, while echo-time parameters were different: inter-echo time in the two-shot RARE CDI sequence was 2.64 ms and spin-echo time in the EPI CDI sequence was 35 ms. To maximize CDI

Current Density Imaging as Means to Follow Tissue Electroporation

15

Fig. 8 A schematic presentation of a CDI sample comprised of a chamber with inserted conductive material through which two electroporation electrodes, parallel to the static magnetic field B0, are inserted. The sample is placed inside a radio-frequency MRI probe (segmented wire around the sample)

sensitivity, k-space traversal ordering was centric for the RARE-type of sequence and was sequential for the EPI CDI sequence.

Results CDI results of the two-shot RARE CDI sequence and EPI CDI sequence on the liver sample are shown in Fig. 9 by vector field maps of CD distributions (left column images) as well as by CD magnitude images (right column images). In the RARE CDI experiment, 90 electroporation pulses were delivered to the sample, and 45 CD images were calculated as delivery of two electroporation pulses was needed for calculation of one CD image. Results obtained with the two-shot RARE CDI sequence are presented with CD images corresponding to the first two electroporation pulses and to the last two (89th and 90th) electroporation pulses. As can be seen from the vector field images (first vs. second row in Fig. 9), an area with higher current density (larger than 10 kA/m2) was established around the electrodes during the application of first two pulses, whereas during the last two pulses, this area has expanded toward the area between the electrodes. The direction of current path did not change with repeated electric pulses significantly; it only got more focused to the direct path between the electrodes. Thus, the current density in the middle of the sample increased from first to last electric pulses by 60 %, i.e., from 8 kA/m2 to 13 kA/m2. These results were confirmed also by measurements of electric current during delivery of electroporation pulses using an oscilloscope shown in Fig. 10a. Currents calculated from CD images (red circles) are well in the range of the oscilloscope readings (blue crosses). Figure 10b shows a typical current profile during application of an electroporation pulse; the profile was measured with the last electroporation pulse. From the profile can be seen that current was increasing during the pulse. CDI results shown in the bottom row of Fig. 9 were obtained by the EPI CDI sequence on a parallel liver sample to the one used in the RARE CDI experiment. In the experiment, 90 electroporation pulses were delivered, and the same number of CD images was obtained. The presented images correspond to CD

16

I. Serša and F. Bajd

distribution at the end of the EPI CDI experiment. The results confirm that a CD distribution can be imaged also by the EPI CDI sequence. The magnitude CD image has the highest current density next to the electrodes and then current density decreased as distance from the electrodes increased. However, streamline of the highest current was not a straight line connecting the electrodes, but was shifted sideways.

Discussion Current density imaging is an efficient tool for monitoring electroporation. Electroporation CDI is enabled by fast MRI sequences modified by addition of electroporation pulses. All the sequences have in principle two parts, the first current encoding part in which a magnetic field change induced by electroporation currents is encoded in the magnetization precession phase and the second part where image signal is acquired preferably from the entire k-space. The obtained magnetic field change maps are then converted to the corresponding current density maps using Ampere’s law. The current density maps along with electric potentials at the electrodes are inputs to the MREIT algorithm (MREIT, “▶ Principles and Use of Magnetic Resonance Electrical Impedance Tomography in Tissue Electroporation”) for reconstruction of the electric field between the electrodes and for calculation of the sample conductivity maps. In the overview, two CDI sequences designed for electroporation monitoring are presented, the two-shot RARE CDI sequence and the EPI CDI sequence. Experiments performed on a liver test sample confirmed that with both sequences, images of current distribution can be obtained, however, with different limitations. The two-shot RARE CDI sequence produces results that have fewer artifacts than the results obtained by the EPI CDI sequence. However, the two-shot RARE CDI sequence requires application of many refocusing RF pulses and has therefore a considerable SAR value, which can prevent its use in clinical or other in vivo applications. SAR is a measure of the rate at which energy is absorbed by the human body when exposed to an RF electromagnetic radiation. As there are strict regulations regarding SAR in clinical MRI, the performance of the two-shot RARE CDI sequence could be impeded when used in human MRI. For comparison, in the presented study, for acquisition of one CD image, 130 refocusing RF pulses were delivered to the sample with the two-shot RARE CDI sequence, while only one refocusing RF pulse was delivered to the sample with the EPI CDI sequence. Another disadvantage of the two-shot RARE CDI sequence is that it requires double the number of electroporation pulses for obtaining the same number of images as the EPI CDI sequence. Based on these limitations, it can be concluded that the two-shot RARE CDI sequence is a more appropriate technique for electroporation of tissue samples ex vivo, while for in vivo electroporation studies, EPI CDI or perhaps also spiral CDI (Yan et al. 1997) is better suited. Artifacts in EPI CDI have often origin in poor magnetic field inhomogeneity of the scanned samples. These can be to some extent reduced by using parallel imaging

Current Density Imaging as Means to Follow Tissue Electroporation

17

Fig. 9 Cumulative effect of repetitive electroporation pulses on current density distribution in the electroporated beef liver tissue sample measured by the two-shot RARE CDI pulse sequence and EPI CDI sequence. A clear difference in both in-plane current density vector field j = ( jx, jy) calculated by using Ampere’s low and the corresponding current density magnitude field can be seen between the first two delivered electric pulses (relatively smaller values) and the last two delivered pulses (relatively higher values) of totally 90 electric pulses of 3000 V. For comparison, EPI CDI results of a parallel beef liver tissue sample are also shown

technology, which allows a reduction in number of scanned k-space lines and therefore allows a reduction of the EPI signal acquisition train length on account of the use of multiple receiver coils (Pruessmann et al. 1999). The technology could also contribute to better results of the EPI CDI sequence with fewer artifacts. Another artifact contributing problem that is common to all CDI experiments are electrodes. These need to be conductive and therefore inevitably produce MR signal

18

I. Serša and F. Bajd

Fig. 10 Measured current increase during the delivery of 90 electroporation pulses (a) and current profile during the delivery of the last one electroporation pulse (b)

voids in their vicinity. In the study, platinum-iridium electrodes were used as they are biocompatible and produce relatively low artifacts. The metal electrodes were needed because the electrodes were implanted into the liver sample. If the electrodes could be attached to the sample, then perhaps a better solution would be a use of hydrogel (nonmetal) electrodes. These have a conductivity closer to that of the tissue and therefore produce very limited or almost no artifacts. Electroporation monitoring was enabled in the study not only by the use of fast imaging methods but also by the use of a specific electrode/slice arrangement that produces reasonable current distribution maps without any sample rotations to orthogonal orientations. A need of these would at least impede the electroporation monitoring by CDI if not even make it impossible. For now, no easy solution is known that would overcome the problem associated with CDI in a general slice orientation. As shown in Fig. 1, CDI has its peak sensitivity at tc = T2, and, unfortunately, in practice most biomedical applications of CDI are such that there is a demand for either application of very short- and high-current pulses or long- and low-current pulses. The first case is associated with CDI applications in monitoring electroporation, while the second case is associated with attempts to image physiological currents, which are low due to low conductivity of biological tissues and also due to low voltages that living systems can bear (Joy et al. 1989). The most challenging of these is undoubtedly CDI of neuronal currents, which is still mostly at the level of theoretical concepts (Hagberg et al. 2006, Cassara et al. 2009). The CDI sensitivity may become problematic in IRE experiments, where the standard IRE protocol uses only 100 μs long electric pulses repeated 90 times at the frequency of 1 Hz. The frequency of 1 Hz is too low to enable execution of more than one electric pulse in one CDI sequence. This limit is posed by the T2 relaxation time of biological tissues, which is of the order of 100 ms. Only one 100 μs pulse is four times shorter than what was used in a standard electroporation experiment on the same test sample,

Current Density Imaging as Means to Follow Tissue Electroporation

19

where four 100 μs were delivered in intervals of 100 μs (Kranjc et al. 2011). A way to compensate the fourfold loss of sensitivity loss due to shortening the electric pulses is to increase the signal-to-noise ratio of the image, which can be done by increasing imaging field of view and slice thickness or to decrease image resolution. Alternatively, instead of one 100 μs electroporation pulse, a packet of up to four 100 μs electroporation pulses delivered in short intervals or one 400 μs electroporation pulse can be used. If none of these is an option, then the only solution would be to perform experiments in a higher field magnet. Namely, image signal in theory increases approximately proportionally with the magnetic field squared. Signal averaging is not an option as the temporal resolution is a priority in electroporation monitoring by CDI. In the presented study, the highest obtained electric current was around 5.5 A, while in clinical cases of in vivo IRE, the current can reach up to 50 A (Bertacchini et al. 2007) due to electrode depth of insertion. Application of CDI during IRE in vivo would therefore enable measurement of current density distributions with a better SNR than was obtained in this study.

Conclusion In the overview, two sequences designed for electroporation monitoring by CDI, namely, the two-shot RARE CDI and EPI CDI sequence, are presented. With both sequences, it was confirmed that current distribution imaging during delivery of short high-voltage electric pulses is feasible. The sequences require only one (EPI) or two (RARE) short electric pulses repeated in an interval of approximately 1 s for acquisition of one current density image. This enables sequential imaging of current distribution during delivery of the electric pulse train, which could potentially be used in MREIT-enabled monitoring of tumor coverage by electric field during IRE tissue ablation.

Cross-References ▶ Diffusion Weighted Magnetic Resonance Imaging for Detection of Tissue Electroporation In Vivo ▶ Medical Imaging of Electroporation ▶ Principles and Use of Magnetic Resonance Electrical Impedance Tomography in Tissue Electroporation

References Beravs K, White D, Sersa I, Demsar F (1997) Electric current density imaging of bone by MRI. Magn Reson Imaging 15(8):909–915

20

I. Serša and F. Bajd

Bertacchini C, Margotti PM, Bergamini E, Lodi A, Ronchetti M, Cadossi R (2007) Design of an irreversible electroporation system for clinical use. Technol Cancer Res Treat 6(4):313–320 Callaghan PT (1991) Principles of nuclear magnetic resonance microscopy. Oxford University Press, Oxford, UK/New York Cassara AM, Maraviglia B, Hartwig S, Trahms L, Burghoff M (2009) Neuronal current detection with low-field magnetic resonance: simulations and methods. Magn Reson Imaging 27 (8):1131–1139 Hagberg GE, Bianciardi M, Maraviglia B (2006) Challenges for detection of neuronal currents by MRI. Magn Reson Imaging 24(4):483–493 Joy M, Scott G, Henkelman M (1989) In vivo detection of applied electric currents by magnetic resonance imaging. Magn Reson Imaging 7(1):89–94 Kranjc M, Bajd F, Sersa I, Miklavcic D (2011) Magnetic resonance electrical impedance tomography for monitoring electric field distribution during tissue electroporation. IEEE Trans Med Imaging 30(10):1771–1778 Kranjc M, Bajd F, Sersa I, Woo EJ, Miklavcic D (2012) Ex vivo and in silico feasibility study of monitoring electric field distribution in tissue during electroporation based treatments. PLoS One 7(9):e45737 Kranjc M, Bajd F, Sersa I, Miklavcic D (2014) Magnetic resonance electrical impedance tomography for measuring electrical conductivity during electroporation. Physiol Meas 35 (6):985–996 Kranjc M, Markelc B, Bajd F, Cemazar M, Sersa I, Blagus T, Miklavcic D (2015) In situ monitoring of electric field distribution in mouse tumor during electroporation. Radiology 274(1):115–123 Kwon OI, Chauhan M, Kim HJ, Jeong WC, Wi H, Oh TI, Woo EJ (2014a) Fast conductivity imaging in magnetic resonance electrical impedance tomography (MREIT) for RF ablation monitoring. Int J Hyperthermia 30(7):447–455 Kwon OI, Jeong WC, Sajib SZ, Kim HJ, Woo EJ (2014b) Anisotropic conductivity tensor imaging in MREIT using directional diffusion rate of water molecules. Phys Med Biol 59 (12):2955–2974 Lee EW, Chen C, Prieto VE, Dry SM, Loh CT, Kee ST (2010) Advanced hepatic ablation technique for creating complete cell death: irreversible electroporation. Radiology 255(2):426–433 Marty M, Sersa G, Garbay JR, Gehl J, Collins CG, Snoj M, Billard V, Geertsen PF, Larkin JO, Miklavcic D, Pavlovic I, Paulin-Kosir SM, Cemazar M, Morsli N, Rudolf Z, Robert C, O'Sullivan GC, Mir LM (2006) Electrochemotherapy – an easy, highly effective and safe treatment of cutaneous and subcutaneous metastases: results of ESOPE (European Standard Operating Procedures of Electrochemotherapy) study. EJC Suppl 4(11):3–13 Mezeme ME, Pucihar G, Pavlin M, Brosseau C, Miklavcic D (2012) A numerical analysis of multicellular environment for modeling tissue electroporation. Appl Phys Lett 100(14):1437011–143701-4. Article number is 143701 Miklavcic D, Beravs K, Semrov D, Cemazar M, Demsar F, Sersa G (1998) The importance of electric field distribution for effective in vivo electroporation of tissues. Biophys J 74 (5):2152–2158 Miklavcic D, Mali B, Kos B, Heller R, Sersa G (2014) Electrochemotherapy: from the drawing board into medical practice. Biomed Eng Online 13(1):29 Neal RE 2nd, Rossmeisl JH Jr, Garcia PA, Lanz OI, Henao-Guerrero N, Davalos RV (2011) Successful treatment of a large soft tissue sarcoma with irreversible electroporation. J Clin Oncol 29(13):e372–e377 Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P (1999) SENSE: sensitivity encoding for fast MRI. Magn Reson Med 42(5):952–962 Scott GC, Joy MLG, Armstrong RL, Henkelman RM (1992) Sensitivity of magnetic-resonance current-density imaging. J Magn Reson 97(2):235–254 Sersa I (2008) Auxiliary phase encoding in multi spin-echo sequences: application to rapid current density imaging. J Magn Reson 190(1):86–94

Current Density Imaging as Means to Follow Tissue Electroporation

21

Sersa I, Jarh O, Demsar F (1994) Magnetic resonance microscopy of electric currents. J Magn Reson A111(1):93–99 Sersa I, Beravs K, Dodd NJF, Zhao S, Miklavcic D, Demsar F (1997) Electric current density imaging of mice tumors. Magn Reson Med 37(3):404–409 Serša I, Bajd F, Kranjc M, Busse H, Garnov N, Trampel R, Miklavčič D (2015) Comparison of single-shot rapid acquisition with relaxation enhancement and echo planar current density MRI sequences for monitoring of electric pulse delivery in irreversible electroporation. 1st World Congress on Electroporation and Pulsed Electric Fields in Biology, Medicine and Food & Environmental Technologies (WC 2015) Yan R, Yoon R, Joy M (1997) Fast current density imaging with spiral acquisition. Proceedings of the 5th Annual Meeting of ISMRM, Vancouver, Canada, 1997. p 1815 Yarmush ML, Golberg A, Sersa G, Kotnik T, Miklavcic D (2014) Electroporation-based technologies for medicine: principles, applications, and challenges. Annu Rev Biomed Eng 16:295–320

Mass Transfer of Electrolytic Species During Electric Field-Based Tumor Treatments Guillermo Marshall

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Electric Field-Based Tumor Treatments: Electrolytic Ablation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Electric Field-Based Tumor Therapies: Electroporation-Based Treatments . . . . . . . . . . . . . . . . . . . . . 9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Abstract

To attain a reliable outcome in electric field-based tumor treatments, dose planning is a must, and dose planning, in turn, requires establishing the dose–response relationship. But finding reliable dose and response parameters implies analyzing the electric field–tissue interaction, in particular, the inevitable appearance of complex electrolytic mass transfer processes and the inherent tissue damage. This review of electric field-based tumor treatments highlights the fundamental role played by the mass transfer of electrolytic species in the dose–response relationship. During the electrolysis process in electric field-based tumor treatments, electrochemical reactions take place at the electrodes, producing at the anode oxygen, chlorine, and protons as the main by-products, while hydrogen and hydroxide ions are released at the cathode. Proton and hydroxyls generation yields strong pH fronts. Since these fronts are the main product of electrolytic reactions, it is reasonable to assume that they are the main cause of tissue damage.

G. Marshall (*) Laboratorio de Sistemas Complejos, Departamento de Computación, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Instituto de Física del Plasma, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires and Consejo Nacional de Investigaciones Científicas y Técnicas, Buenos Aires, Argentina e-mail: [email protected]; [email protected]; [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_68-1

1

2

G. Marshall

The amount of electrolytic products emerging from chemical reactions is proportional to the amount of electric charges or Coulomb dose passing through the tissue; thus, Coulomb dose is proportional to tissue damage. Theory shows that in a constant electric field such as in electrolytic ablation, an optimal dose–response relationship is the minimum Coulomb dose necessary to achieve total tumor destruction while minimizing healthy tissue damage. In a pulsed electric field such as in gene electrotransfer, unwanted damage due to electrolysis is non-negligible; here, an optimal gene electrotransfer treatment is predicted as the critical Coulomb dose yielding maximum electroporated area with minimum damage. Theory shows that when tissue natural buffer is taken into account, damage is attenuated (though still remaining non-negligible), and the critical Coulomb dose for optimal gene electrotransfer increases. Keywords

Electrolytic ablation • Electrochemotherapy • Gene electrotransfer • Irreversible electroporation • Dose planning methodology • Dose–response relationship • Buffer natural system • Mathematical modeling

Introduction The use of electric field (EF)-based tumor treatments triggers complex electrolytic mass transfer processes underlining these treatments, thus constituting a paradigmatic example of biomedical applications of electrochemistry (Marshall 2014). Depending on the type of EF being applied, direct or pulsed, these treatments can be classified into electrolytic ablation (EA) and electroporation (EP)-based tumor treatments. Electrolytic ablation (EA) of tumors, also called electrochemical treatment of tumors (EChT), is a nonthermal ablative method consisting in the application of a low constant electric field (long pulse) through two or more electrodes inserted in the tissue generating electrolytic products that induce tumor necrosis. This treatment was pioneered by Nordenström (1983) and has been widely used in China with good clinical results (Li et al. 2006). Some of the characteristics of EA are its simplicity, effectiveness, low cost, and negligible side effects. In EA typical clinical applications, 6–8 V/m are used during periods from minutes to hours. There are several groups working in Australia, China, Cuba, Japan, Sweden, and the USA; a review can be found in Nilsson (2000). Electroporation (EP)-based tumor treatments emerged during the last years as a new cancer treatment. EP perturbs cell membrane integrity by the application of pulsed electric fields. EP-based techniques, such as electrochemotherapy (ECT), gene electrotherapy (GET), and irreversible electroporation (IRE), were implemented for medical purposes. ECT combines reversible EP with poorly permeant anticancer drugs, such as bleomycin, to potentiate their entry to the cell, thus their intrinsic cytotoxicity. Since its beginnings in the late 1980s, ECT has evolved into a clinically verified treatment for tumors of different origin in Europe

Mass Transfer of Electrolytic Species During Electric Field-Based Tumor. . .

3

(▶ Electrochemotherapy and Its Clinical Applications). GET uses the same EP mechanism as in ECT to transfer plasmidic DNA instead of a drug. GET has a great potential as a nonviral gene delivery system as it achieves the introduction of plasmids or oligonucleotides into the cells (▶ Standard Operating Procedures for Electrochemotherapy). After the development of EP devices, this technique became widespread for delivering molecules inside the cell. If GET is applied to muscle cells, it works in favoring the antigen production within the skeletal muscle and in the activation of pro-inflammatory pathways and in the recruitment of cells involved in antigen presentation. This mechanism turns GET into a useful strategy in DNA vaccination protocols and a promising approach for the introduction of foreign antigens into the host for inducing an immune response not only against infectious diseases but also against malignant tumors (Chiarella et al. 2010). A recent derivation of ECT is IRE, introduced by Davalos et al. (2005), an irreversible EP (electric pulses above the irreversible threshold), without thermal effects, that leaves intact main tissue structures (▶ Irreversible Electroporation and Its Clinical Applications). A combination of EA with reversible EP has been shown recently to significantly increase the extent of tissue ablation in comparison to that obtained with EA alone (▶ Combining Electrolysis and Electroporation for Tissue Ablation). A possible explanation is that reversible EP potentiates the entrance of toxic electrolytic products into the cell increasing tissue ablation. The use of a high-intensity nanosecond pulsed electrical fields (nsPEF) or nanosecond EF is another type of tissue electroporation gaining momentum (▶ Tissue Ablation Using Nanosecond Electric Pulses; ▶ Preclinical Studies on Nanoseconds Pulses). It has the effect of reaching the nucleus of the cell, thus allowing a direct manipulation of the nucleoplasm apparently leaving intact the membrane cell. The effects on plasma membrane are to permeabilize it after a few nanoseconds, while the effects in subcellular membranes are to permeabilize intracellular organelles producing a massive calcium release to the cytoplasm which triggers cell death mechanisms. Although having less undesired effects than traditional cancer therapies like radiotherapy or chemotherapy, EF therapies still have some side effects (pain during pulse delivery, swelling of the treated area, necrotic ulcers when over treating the area, plasmid damage) that it is necessary to minimize. In silico modeling validated with in vivo and in vitro measurements can greatly contribute to elucidate fundamental aspects of electric field–tissue interaction, thus contributing to minimize side effects through dose planning. In this context, it is worth observing that in spite of the wide differing scales of the electric parameters involved in EF therapies, there is a close electrochemical analogy between EA- and EP-based treatments. This becomes evident imagining an EA treatment with a very short pulse similar to one pulse of an EP-based treatment. To a greater or lesser extent, both are affected by the underlying electrolytic process, a process that is a function of the applied electric charge, i.e., Coulomb dose. Knowledge of the dose–response relationship leading to an optimal treatment implies the analysis of the electrolytic mass transfer processes and tissue damage. This analysis is firstly introduced for electrolytic ablation, whose aim is tissue

4

G. Marshall

destruction, and then extended to EP-based tumor treatments, in particular to GET, in which tissue destruction is an unwanted effect. In EA it is shown that an optimal dose–response relationship is the minimum Coulomb dose necessary to achieve total tumor destruction while minimizing healthy tissue damage. An optimal GET is the critical Coulomb dose that yields maximum electroporated volume (assuming uptake proportional to electroporated volume) while minimizing healthy tissue damage. The plan of the chapter follows: section “Electric Field-Based Tumor Treatments: Electrolytic Ablation” presents a brief review of constant EF tumor treatments, the mass transfer processes underlying this treatment, and its dose planning; section “Electric Field-Based Tumor Therapies: Electroporation-Based Treatments” describes a brief review of pulsed EF therapies whether combined or not with anticancer drugs or plasmids, the associated mass transfer underlying these therapies, and its dose planning; and finally, section “Conclusions” draws some general conclusions.

Electric Field-Based Tumor Treatments: Electrolytic Ablation In the following, a brief review of experimental and theoretical modeling of EA protocols with a differing degree of complexity is presented. It emphasizes the underlying ion transport process leading to tissue damage and its role in obtaining the dose–response relationship. Tissue damage in EA is mainly produced by necrosis. This is because, during the electrolysis process, electrochemical reactions take place at the electrodes, producing at the anode oxygen, chlorine, and protons as the main by-products, while hydrogen and hydroxide ions are released at the cathode. Proton and hydroxyls generation yields strong pH variations that are named pH fronts. Since pH fronts are the main product of electrolytic reactions, it is reasonable to assume that pH fronts are the main cause of tissue damage. According to Faraday’s law, the amount of electrolytic products emerging from chemical reactions is proportional to the amount of electric charges or Coulomb dose passing through the tissue. As a consequence, Coulomb dose is proportional to tissue damage. In a pioneering series of papers discussed in Nilsson (2000), in silico modeling of an EA protocol applied to a tumor tissue was presented. The tissue matrix was seen as an electrolyte and the EA protocol as an electrolysis process. The ion transport in a zone near one of the electrodes (cathode or anode) was described by a quasi-threedimensional model (spherical symmetry model) using the Nernst-Planck equations for ion transport. Results show pH profiles at the anode to be strongly correlated with the size of in vitro measured necrosis, thus confirming the effect of pH (the spreading of hydroxyl ions) in tissue necrosis and suggesting that the model could be used for predicting EA tumor treatment. More realistic models were presented in Colombo et al. (2007). The EA modeling, based on first principles and taking into account the most relevant variables of the phenomenon, consists in the one-dimensional Nernst–Planck equations for ion transport in a four-component electrolyte (Na+, OH , Cl , and H+) and the Poisson

Mass Transfer of Electrolytic Species During Electric Field-Based Tumor. . .

5

equation for the electrostatic potential under galvanostatic conditions in which the full cathode–anode ion transport interaction is described; the model was solved numerically and validated with in vivo and in vitro modeling using two electrodes with a separation of 3 cm between them. The main finding was that, whether in vivo, in vitro, or in silico, an initial condition with almost neutral pH evolves between electrodes into extreme cathodic alkaline and anodic acidic fronts moving toward each other, leaving the possible existence of a biological pH region between them (if sufficiently long time has not been elapsed); toward the periphery, the pH falls to its neutral values. In particular, in vivo measurements of necrotized tissue as a function of the Coulomb dose were presented. Here a point platinum electrode (anode) with a diameter of 1 mm was inserted in the center of the tumor, while the cathode was placed subcutaneously far away from it. The macroscopic necrotic area, defined by a dark, almost spherical colored zone around the anode, was measured for different Coulomb doses (10, 30, and 50 C) scaling linearly with the Coulomb dose. Toward a more realistic in vitro modeling for studying pH effects, in Olaiz et al. (2010a) a new gel model composed of a matrix of collagen into which sodium chloride (main salt present in tissues), NaHCO3 (incorporating, together with CO2, buffering capacity), and egg yolk (organic matter similar to tissue composition) were added was presented. The model was called the collagen–macronutrient gel model, and as previous agar–agar models, it is basically a porous and hydrated elastic gel structure. In this work, the concept of electrodenaturation front was introduced, showing that it can be easily followed by a color virage, thus, used for mimicking tumor tissue destruction by an EA. Since the main objective of EA is tumor tissue destruction, electrodenaturation front tracking turns out to be an appropriate measure of its determination. In particular, front tracking reveals that the electrodenaturation front grows under a diffusion-controlled regime, a result allowing in principle, to predict the time needed for tumor destruction without compromising healthy tissue. To show the effects of pH fronts in a different electrode configuration in Olaiz et al. (2010b), a one-probe two-electrode device (OPTED) containing the cathode and the anode very close to each other (1 mm) was introduced. Main advantages of the OPTED are the insertion of one applicator rather than two or more (thus minimizing tissue intrusion, for instance, in the nervous system), the ability to reach tumors beyond capabilities of conventional surgery, and the minimization of electric current circulation through the treated organ. Since the effects of the exposure of a 3D gel model during OPTED-EA can be assessed by the volume covered by pH fronts, in the abovementioned work, a pH front tracking by means of pH indicators’ color change was performed. Results of the OPTED-EA 3D gel modeling show that starting from an initial uniform condition, two half-spherical pH fronts, one basic and the other acid (from cathode and anode, respectively), expand toward the periphery configuring a distorted full sphere. Between electrodes, the two fronts collide and are neutralized. A 2D space and time representation of the OPTED-EA experiment showing the evolution of the acid and basic pH fronts and the measurement of the damage area are depicted in Fig. 1. The left image in Fig. 1 is a sequence of four snapshots at different times extracted from a video of 1200-s duration in which the camera is

6

G. Marshall

Fig. 1 Time (s) sequence of digital snapshots (taken from above) showing the evolution of the acid and basic pH fronts in a 3D gel model during OPTED-EA with a constant current of 4 mA. Left image, from top to down: (a) 100 s, (b) 300 s, (c) 700 s, and (d) 1200 s. Center image: reslice from the sequence of snapshots. Lines joining left and center panes indicate where the left snapshot is located in the stack. Right image: calculated trajectory of both pH fronts (From Olaiz et al. 2010b)

Mass Transfer of Electrolytic Species During Electric Field-Based Tumor. . .

7

viewing the experimental process from above. Acid and basic pH fronts are represented by bright and dark pixels, respectively. The central image named a reslice is constructed from digital gray scales made from a sequence of snapshots such as those shown in the left image (using imaje from the National Institutes of Health, NIH) as follows. Snapshots made each second are averaged over 1% of the cell width (direction parallel to the y-axis) to reduce each image to a line. Then these lines are stacked to yield the space-time image spanning the duration of the experiment (1200 s). The right image depicts the trajectory of both fronts and is the result of a border detection algorithm applied to the central image. Dose planning methodology is critical for a reliable treatment outcome. Since the late 1970s, Coulomb dose according to tumor size was the guideline for the optimal choice of electric parameters (Nilsson 2000). Later on, searching for optimal dose planning, several authors investigated the dose–response relationship between the applied current, treatment time, Coulomb dose, and necrotized tissue volume (NVT). Some authors (Nilsson 2000) used different Coulomb doses and found a linear relationship between damage and Coulomb dose. This result was also experimentally corroborated in Colombo et al. (2007) though it was not checked with a theoretical model. In Luján et al. (2015) dose planning in EA was addressed for the first time through in silico modeling using the Nernst–Planck equations for ion transport validated with data taken from Olaiz et al. (2010b). Here, the dose planning methodology is presented in detail. It is recalled that in EA, Coulomb dose and tissue damage are reliable dose and response parameters, respectively. For dose planning, there is a need to obtain the dose–response relationship, i.e., the Coulomb dose–tissue damage. The Coulomb dose is obtained for a fixed electric current as a function of the time elapsed during the EA experiment. Damage is indirectly obtained, tracking the pH front in space and time during the EA experiment (this front is put into evidence with the tracking technique previously shown). Clearly, the volume covered by pH fronts (phenolphthalein and methyl red with transition pH ranges 8.0–9.6 and 4.8–6.2, respectively) determines tissue damage. Here, damage is obtained with the necrotized tissue volume (NTV) defined as the volume of the tumor necrotized by EA. With these two parameters, it is possible to construct the NTV vs. Coulomb dose graph thus experimentally establishing the functional relation between them. This graph is presented in Fig. 2 (marker lines) showing the experimentally obtained linear relation between dose and response. How the dose–response relationship is predicted and their functional relationship theoretically established? The Coulomb dose is obtained for a fixed electric current as a function of the time elapsed through the numerical simulation of the experimental EA model previously described. The numerical simulation uses the Poisson–Nernst–Planck equations describing electric field and ion transport in an electrolyte with four components; see details in Luján et al. (2015). Damage is indirectly obtained numerically tracking the pH front in space and time during the EA simulation. The volume covered by pH fronts determines the predicted damage. Figure 3 shows predicted acid (blue circle) and basic (red circle) pH fronts trajectories in space (horizontal) and time (vertical). The light green background

8 2mA, in silico 4mA, in silico 8mA, in silico 2mA, in vitro 4mA, in vitro 8mA, in vitro

NTV/(10–3 m)3

Fig. 2 NTV experimental (marker lines) and predicted (continuous lines) vs. Coulomb doses for different electric currents (From Luján et al. 2015)

G. Marshall

1000

1

10 Q/C

200

400

600

t/s

Fig. 3 Predicted acid (blue area) and basic (red area) determined by pH fronts trajectories in space (horizontal axis) and time (vertical axis downwards). The light green background corresponds to the gel matrix. Shadowed spheres indicate two ideal tumors

800

1000

1200

corresponds to the gel matrix. Shadowed spheres indicate two ideal tumors whose volume is covered by the equivalent red blue pH front. With these two parameters, it is possible to construct the necrotized tissue volume vs. Coulomb dose graph, thus theoretically establishing the functional relation between them. This is presented in Fig. 2 (continuous lines) for different electric currents. Results show that predictions follow experimental trends and there is a nonlinear relationship between NTV and Coulomb dose, that is, NTV scales as Q 1.4. The question arises whether this predicted dose–response relationship between the Coulomb dose and NTV is an optimum for clinical use. The answer requires some basic assumptions: the ideal generic tumor to be ablated by the physician (plus a security margin) is contained in a sphere of radius R, thus, a fixed volume size, and

Mass Transfer of Electrolytic Species During Electric Field-Based Tumor. . .

9

Fig. 4 Evolution in space (horizontal axis, anode at the left, cathode at the right) and time (vertical axis up–down) of the anodic acid (pink) and cathodic basic (red) pH fronts in an ECT gel model (8 pulses, 400 V, 300 μs, 1 Hz). Total time length: 8 s (From Turjanski et al. 2011)

the OPTED electrode is located at the center of the tumor. Therefore, when this generic tumor is subjected to an OPTED-EA treatment, its volume is gradually covered in time by a spherical pH front of radius R (I, t) that determines NTV as illustrated in Fig. 3 (for two tumors in different positions). Clearly, when NTV reaches the sphere of radius R, the tumor is totally destroyed with a minimum Coulomb dose.

Electric Field-Based Tumor Therapies: Electroporation-Based Treatments While in EA tumor necrosis induced by pH fronts is the main goal of the treatment, in reversible EP-based treatments such as in ECT, IRE, or GET, it must be minimized because of its collateral effects. With the aim of illustrating the role of pH fronts in EP-based tumor treatments, a brief review of experimental and theoretical models of EP treatments follows. The effect of electric pulses in pH changes in a cell suspension was analyzed in Saulis et al. (2005). Using NaCl solutions buffered with different amount of sodium phosphate under high-voltage electric pulses and stainless steel anode and different materials in the cathode, authors found that although variations in the whole volume of the electrolytic chamber can exceed 1–2 pH units in average, near the cathode this variation was significantly greater. It was conjectured that the change of pH, in some cases, might be one of the factors causing cell death. In Turjanski et al. (2011), the EP process was seen from a new angle apparently overseen in the literature, the role of pH in ECT modeling based on ion transport during the treatment. The analysis was developed through in silico modeling validated with in vitro gel measurements drawing from previous experience of the authors in electrolytic ablation. For conditions typical to many ECT clinical studies found in the literature, experimental results, as shown in Fig. 4, unveil the presence of strong pH fronts emerging from both electrodes. Moreover, theoretical predictions of the comparison of ECT pH fronts with those arising in EA showed a striking result: anodic acidification is larger in ECT than in electrolytic ablation, suggesting that tissue necrosis could also be greater. Since they might give rise to tissue

10

G. Marshall

necrosis, an unwanted condition in clinical applications of ECT as well as in GET, it was deemed relevant to quantify their extent and evolution. Experimental results from literature disclose that in EP protocols consecutive pulses increase tissue electric conductivity and the electroporated area, effects not taken into account in previous theoretical modeling. With the aim of including these effects, in Suárez et al. (2014), a theoretical model describing the electric field with a new formulation of the electric conductivity and its validation with measurements in a vegetable tissue model (potato slice) was presented. The theoretical model solves the nonlinear Laplace equation with a variable conductivity depending on the electric field, the temperature and the number of pulses, and the Penne’s bioheat equation for temperature. Theory predicts and experiments confirm that, increasing pulse number, the current density and the electroporated area (potato-blackened area) increase, a fact that can only be explained by an increase in conductivity due to higher pulse numbers. In the context of this review, the unaccounted conductivity increase due to consecutive pulses can be described in terms of pH effects as follows. Pulse number is proportional to Coulomb dose; therefore, at a constant applied electric field, as pulse number increases Coulomb dose increases, leading to higher pH effects. Naturally, higher pH effects increase conductivity and therefore electric current and electroporated area. Toward more realistic models of EP-based tumor treatments, in Maglietti et al. (2013), the role of pH fronts in EP-based tumor treatments was experimentally measured in an ex vivo model. The aim was to show that the pH fronts generated produce non-negligible pH changes in a tissue regardless of the presence of natural buffers. Experimental measurements are based on the application of EP for different sets of pulse parameters (corresponding to IRE, ECT, or GET modalities) to an ex vivo model (chicken muscular tissue) for conditions found in many studies. Results show the existence of pH fronts emerging from both electrodes and that these fronts are immediate and substantial. pH fronts are indirectly measured through the evanescence time (ET) of a pH indicator. The evanescence time is defined as the time required for the tissue buffer to neutralize them. Muscular tissue samples were sliced from fresh chicken. To estimate the hydroxyl production around the cathode inside the tissue, a pH indicator dye (phenolphthalein, C20H14O4, transition pH range of 8.0–9.6, from colorless to red) was used. One drop of phenolphthalein was applied over the surface of the samples, and two electrodes (parallel surgical steel needles 0.8 mm thick, 2.5 cm long, their surfaces separated from each other by 0.4 cm) were laid horizontally over the region of interest. Only the hydroxyl production at the cathode was studied as it is equivalent to the proton production at the anode. Immediately after the onset of pulse delivery, the pH indicator dye in the sample changes from colorless to red. After the end of the pulse delivery and as hydroxyls are neutralized by the tissue buffer, the sample turns to colorless again, indicating a neutralizing phenomenon. This evanescence time, which is an indirect measure of the amount of hydroxyls produced, is estimated by the time needed for the tissue buffer to turn the pH value to less than 8 (nearly neutral). Figure 5 presents a chart of the evanescence time for each experimental series. Here, standard pulse parameters

Mass Transfer of Electrolytic Species During Electric Field-Based Tumor. . .

ET/s

11

180,00 160,00 140,00 120,00 100,00 80,00 60,00 40,00 20,00 P G

ET

x3

x3

T

V

ET G

x3

P

ET

x2 ET G

G

T x2 ET G

x2 ET G

ST

V

D

P ET G

ET

x1

x1 G

x1

ET G

ET

/2

/2

T

V /2

D ST G

ST IR E

EC T

D

0,00

Experimental Series

Fig. 5 Chart of the mean evanescence time ET (s) for each experimental series. The standard pulse parameters for IRE, ECT, and GET are depicted in blue. Pulse parameters corresponding to half, double, and triple standard GET pulse amplitudes are shown in red. Pulse parameters corresponding to half, double, and triple standard GET pulse lengths are depicted in green. Pulse parameters corresponding to half, double, and triple standard GET pulse numbers are shown in orange. Standard IRE, standard ECT, and all GET parameter variations are significantly different from standard GET (p τm). If the plasma membrane is considered to be a pure spherical dielectric (λm = 0), we obtain g(λ) = 1. In these conditions and when the steady state is reached (t> > τm) as for millisecond pulses, Eq. 1 simplifies to: ΔViðM, EÞ ¼ 1:5rEcosθðMÞ

(3)

The highest amplitude of ΔVi (M, Ε, t) is then 1.5Er. The consequence is that it affects more dramatically the largest cell in a population with cells with different sizes but biologically identical. In cells of all types, there is an electrical potential difference between the inside of the cell and the surrounding extracellular fluid. This is termed the resting transmembrane voltage of the cell (transmembrane voltage). When living cells are electropulsed, ΔVi adds to the resting one ΔV0. An electrical potential difference results from a net separation of electric charge by the thin nonconductive layer, i.e., the membrane in a cell. TMVs exist across the membranes of organelles. Electrical potentials (voltage) are measured in units of volts. (A volt is defined in terms of energy per unit charge; that is, 1 V is equal to 1 J/C). In a mammalian cell, its resting transmembrane voltage is about 70 mV. The minus sign indicates that the inside of the cell is negative with respect to the

4

J. Teissie

extracellular buffer by definition, and when the ground electrode in measurements is placed outside of the cell. ΔV0 is considered as not dependent on the cell size in a homogeneous population. Due to this ionic difference between the extracellular buffer and the cytoplasmic volume, two driving forces influence the movement of an ion across a plasma membrane. • The concentration gradient • The transmembrane voltage The concentration gradient is valid for all molecules uncharged or charged (including ions). But for ions, the flux, i.e., the movement of an ion across the membrane, should consider the sum of the flux due to concentration gradient as well as due to force exerted by electric field acting on charged particle, i.e., due to TMV (electrophoresis). These two factors may act in the same direction or in opposite directions. The TMV is associated to a transmembrane field (V/cm). E ¼ ΔV=d

(4)

As a membrane is thin (d is about a few nm), even if the TMV is of the order of 100 mV, the associated field can be as large as 10 MV/m. If a membrane is porous, charged molecules are submitted to an electric drift by the transmembrane field as well as a diffusion driven by the concentration gradient. Due to its intrinsic ion leakage, the membrane need to have active transporters to maintain the concentration gradient. The consequence is that metabolic energy is used. The resting TMV is maintained at the expense of this energy. If the metabolic level of the cell is affected (starved cells for example), the gradient is not maintained and the resting TMV is reduced. When an external field is present, then the resulting TMV is ΔVðEÞ ¼ ΔV0 þ ΔViðM, EÞ

(5)

On Fig.1, the TMV point from negative to positive and their lengths is proportional to the voltage. The large arrow on the top represents the electric field lines. On the left, a cell is pictured with its homogeneous resting voltage, the middle is indicative of the field-induced TMV that is highly position dependent along the cell surface. On the right, the resulting modulation of the TMV when the field is present, a highly complex position-dependent TMV is observed. This gives an additive effect on one side of the cell and a subtractive one on the opposite. The electric field effect is strongly position dependent (Fig.1). When a 5103 V/M is delivered on a cell with a 10 μm radius, Eq. 3 predicts that the TMV can locally be doubled. One limit in the use of the Laplace equation to obtain the ΔV from the external field E is that it is limited to spherical cells (Laplace equation). Biology is handling

Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation. . .

5

Fig. 1 Modulation of the TMV of a spherical cell by an applied external field

irregularly shaped cells. In an approach making an analytical approach, the finitethickness, nonzero-conductivity membrane was replaced by a boundary condition in which a specific surface conductivity is assigned to the interface between the cell interior (the cytoplasm) and the exterior (Pucihar et al. 2006). The simulation was validated on spherical cells (using the Laplace equation) and by a direct observation of cells with irregular shapes where the field effect was monitored with a potentiometric fluorescent dye.

The Resulting TMV Induces Membrane Permeabilization (Electropermeabilization) When the new membrane electric potential difference ΔVm (ΔV0 + ΔVi) locally reaches a critical value (ΔVc) (Rems and Miklavcic 2016), a local alteration of the membrane structure leads to membrane permeabilization (electropermeabilization). ΔVc ¼ ΔV0  1:5rEc

(6)

Ec is the lowest field strength that is supposed to trigger permeabilizing events for a spherical cell with a radius r. As is ΔV0 = 70 mV, this is on the pole of the cell facing one electrode (θ(M) = 0). Electropermeabilization initial event is described as a local event on the cell surface. Molecular processes supporting classical electropermeabilization remain poorly understood (Teissie et al. 2005). The membrane is called “porous.” The critical field Ec can be experimentally approached. It is the lowest field strength that induces an enhanced transport or an increase in membrane conductance. If ΔVc is assumed to be the same whatever the cell, that is, considered in an homogeneous population, then it is controlled by the metabolic state (through ΔV0). The critical field will be affecting the largest cell in the population (through r).

6

J. Teissie

As it is an experimental detection of transport or membrane conductance, the sensitivity of the assay plays a capital role.

Methods to Detect Permeabilization Conductance If at any point, ΔVm gets high enough (ΔVc) to induce a sufficient number of conductive defects, the defects allow enhanced ionic transport through the membrane and increase the membrane conductivity by several orders of magnitude (transport, conductivity).

Direct Assays by Electrical Measurements By measuring the voltage and the current flowing between the electrodes, one gets access to the conductivity of the suspension. Under steady state condition, the conductance of a diluted intact cell suspension can be considered controlled by the conductance of the external solution. An ohmic behavior is observed when a voltage pulse is delivered on the cell suspension (as long as the Joule heating is small). When electropermeabilization appears, two events are present: (i) The dielectric shell has conductive defects that increase during the pulse. (ii) There is leak of the internal (cytoplasmic) solution. Both events result in an increase in the cell suspension conductivity. To obtain the critical field Ec, cell suspensions are pulsed by strong fields with increasing strengths and the change in conductivity of the suspension is monitored after each step (by applying low voltage AC train). The limit is indeed the sensibility in the current detection. At the onset of permeabilization, only a very limited cap is affected (θ(M) is close to zero). Therefore, few defects are present on the membrane and the ionic leakage is limited (Kinosita and Tsong 1977; Pavlin et al. 2005). A linear ramp in E with a small AC modulation can be used to detect the dynamic change in current. When the permeabilization occurs, the amplitude of the AC current shows fluctuations. This is described in details on planar lipid membrane (Kramar et al. 2012). The steadily increasing voltage across the bilayer imposed by linearly increasing current (capacitive loading) led to electropermeabilization of the membrane when voltages above a few hundred millivolts were obtained. Small voltage drops were detected preceding the destructive burst of the bilayer due to “irreversible electroporation.” These voltage drops were often followed by a voltage re-rise within tens of milliseconds. The fluctuations are associated to opening and closing of conducting defects through the bilayer. With a coulter counter, cells, due to their poor conductivity when intact, alter the effective cross-section of the conductive channel (conductivity). The electrical

Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation. . .

7

resistance across the channel increases, and under voltage clamp, the electric current passing across the channel decreases. The size of the electric current drop is a function of the size of the cell and of the leakiness of the membrane. Therefore, when the voltage across the channel induces a field high enough to trigger electropermeabilization, the associated current drop is smaller. The limit is that the definition of the field distribution is poor (Zimmermann et al. 1974).

Patch Clamp The patch-clamp technique allows the investigation of currents through a small set or even single ion (conducting) channels and therefore about the membrane conductance. A small pipette is brought in contact with the cell membrane and a tight seal between the pipette and the membrane is obtained by a mild suction. A brief but strong suction ruptured the cell membrane. The pipette has a direct access to the cytoplasm (Ryttsen et al. 2000). The whole-cell patch-clamp technique was employed to investigate the effect of millisecond pulsed electric fields on plant protoplast (wall free plant cell) and DC-3F cells (Wegner et al. 2013). Cellular membrane permeabilization was monitored by a conductance increase. It could be established experimentally that electropermeabilization is associated to a rapid membrane conductance change (due to the creation of conducting defects in the cell membrane). This critical event occurs while the field is present to depolarize or hyperpolarize the membrane to voltages beyond the critical potentials (evaluated to be +201 mV and 231 mV, respectively, in the study).

Transport Membrane permeabilization is associated with an enhanced transport of watersoluble small molecules (that are with a low molecular weight, say less than 2 kD) and ions (transport). It can be assayed by the inflow of dyes (Trypan blue, Propidium iodide (PI), and similar) by observation under the microscope or with a flow cytometer. Permeabilization can then be quantified by two parameters: the number of cells in the population where the uptake can be detected and the amount of uploaded dyes S (i.e., the distribution of fluorescence intensity in each single cell in the population). QðtÞ¼ Pperm Aperm ΔSðtÞ ð1  exp:ðt=TresÞÞ

(7)

Q(t) is flow of S across the membrane at time t after the pulse. Pperm is the permeability coefficient for the reporter molecule. Aperm is the part of the cell surface that is permeabilized, proportional to (1Ep/E). ΔS(t) is the concentration difference of S between the bulk and the cytoplasm.

8

J. Teissie

Tres is the time constant of the membrane resealing that is controlled by the pulse duration (resealing). Permeabilization for S is detected only when the amount of S that is accumulated through the flow is detectable by the assay. The critical field is the one that induces permeabilization on the tiniest cap (A perm must be close to zero). ΔS(0) should be large to increase Q(0), but the problem is the intrinsic leakage to S of the target cell membrane. To keep the flow long lived, Tres must be long (long pulse duration), but cell viability must be preserved. One problem is that there are some concerns on the mechanisms affecting the cell with long pulses. The biomechanics of a cell should consider it as a complex visco-elastic body, with a nonlinear response on the duration of the stress (here the electrical forces associated to the field pulse). A key practical information is that the inflow and the resulting accumulation (that controls the sensitivity of detection of permeabilization) is controlled by Pperm that is specific of the reporter molecule S. Experimental data are obtained with fluorescent dyes either by observing the cells under a digitized fluorescence microscope or with a flow cytometer. Another approach is to follow the leakage of cytoplasmic compounds (endogenous such as Adenosine triphosphate (ATP) or trapped by using loading with acetomethoxy derivatives). Either a single cell approach is possible by following the decrease in emission in each cell or on population by observing the increase of the signal in the bulk phase (as in the case of ATP with the luciferin luciferase assay on a luminometer).

Lipidic Models A lot of studies have been developed on pure lipid assemblies: (Black (Planar) lipid membrane (BLM), liposomes (small unilamellar vesicles (SUV), large unilamellar vesicles (LUV), giant unilamellar vesicles (GUV), and multilamellar vesicles (MLV)). Processes on such physical models are relevant of soft matter physics. Stretching under the applied electrical forces is present and affects the membrane curvature and the layer packing. Modifications of bilayer responses would only be considered in the present chapter. Only definitions in terms of TMVs of critical events occurring in living cell membranes will be approached.

Does the Critical Electropermeabilization TMV Exist? Electrooptical and conductometrical relaxation methods allow a direct monitoring of the structural events affecting lipid nanoassemblies during the electric field discharge. The conclusions from such observations have given an accurate physical insight in the molecular mechanisms of the delivery of drug-like dyes by electropermeabilization to cells and tissues (electropermeabilization). The degree of defect formation as the primary field response increases continuously without a threshold field strength, whereas secondary phenomena, such as a dramatic increase in the

Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation. . .

9

membrane permeability to drug-like dyes (electropermeabilization), indicate threshold field strength ranges. Membrane electropermeabilization is facilitated by an increase in local curvature of the membrane as well as by a gradient of the ionic strength across charged membranes, known to affect the spontaneous curvature. Highly curved systems are affected by heterogeneities in packing on the outer layer that appear as weaknesses facilitating the formation of defects (Neumann et al. 1999). When unmodified planar bilayers are clamped at 150 mV, transient single defects are observed for a long period of time (Melikov et al. 2001). Fast transitions between different conductance levels are recorded due to opening and closing of metastable lipid conducting defects (conductivity). The mean amplitude of conductance fluctuations (approximately 500 pS) was independent of the applied voltage indicating the local nature of the conductive defects. The distribution of defect conductance was wide (dispersion of approximately 250 pS). Short bursts of conductance spikes (opening and closing of defects) were followed by periods of background low conductance. The on-off appearance of conducting defects proceeds through some electrically invisible (silent) mismatches in the lipid matrix. Similar predefects resulting in metastable conductive pathways are intrinsic features of cell membranes and may be involved in different biologically relevant processes (defects). The electropermeabilization process was directly assayed on the minimum model of giant lipid vesicles, by using a highly sensitive fluorescence method based on manganese ion transport (Mauroy et al. 2015). The approach at the single-lipid selfassembly level detects divalent ion permeabilization on those cell sized giant unilamellar lipid vesicles. Critical field and associated TMV are obtained by pulsing with trains of increasing amplitudes. Short (0.1 ms long) pulses were delivered to avoid electromechanical stretching of the vesicles. The critical values are observed to vary from 10 to 150 mV. These values appear to be much lower than those classically reported in the literature for cells and vesicles. The detection method appears to be a decisive parameter as it is controlled by the transport of the reporter dye. The critical TMVs are strongly dependent on the lipid composition. The electropermeabilization process is a transient transition of the lipid self-organization due to the dramatic loss of assembly cohesion that depends on the nature of the lipid mixture building the bilayer. The role of the applied membrane tension on the critical electric field strength for membrane breakdown (rupture due to electropermeabilization) was approached by a simple micropipette technique on lipid vesicles (Needham and Hochmuth 1989). An increase in membrane tension is associated with an increase in the critical membrane voltage, ΔVc (membrane tension). More cohesive membranes are more “resistant” to the TMV effect. This is what is observed when the percentage of cholesterol in a lipid mixture is increased. A theoretical model predicts that membrane rupture occurs at a critical area increase. This dilatation is obtained either by the electrical effect or through the applied membrane tension. Both stresses increase the elastic energy up to this critical level. The model predicts a feedback between tension and critical TMV to obtain vesicle bursting as a result of the electropermeabilization. As a conclusion, under critical tension, the critical TMV can be zero.

10

J. Teissie

Critical Electropermeabilizing TMV on Red Blood Cells In the 1970s, electropermeabilization studies were routinely performed on human red blood cells (RBC, erythrocytes). Samples were obtained either from outdated samples from the blood banks or freshly drawn from volunteers. Aging of the sample was unfortunately never reported. This is an experimental problem as aging erythrocyte undergoes alterations in their plasma membrane. Red blood cells are prepared from whole blood by a simple centrifugation, which pellets the cells while the blood plasma remains in the supernatant. The process is known as blood fractionation and is a routine practice. Packed red blood cells are then kept in a saline buffer at low temperature. Red blood cells are fairly different from other mammalian cells (red blood cells). A human erythrocyte has a donut shape. The diameter is approximately 6.2–8.2 μm and a thickness at the thickest point of 2–2.5 μm and a minimum thickness in the center of 0.8–1 μm. Therefore, RBCs are much smaller than most other human cells. In a suspension, RBC have random orientation meaning that an homogeneous field (achieved with parallel plate electrodes) induces a wide distribution of TMV even when assuming that the population is highly homogeneous in size and physiology. The red blood cell membrane complex is composed of three subunits: the carbohydrate-rich glycocalyx, the lipid bilayer which contains many transmembrane proteins and the membrane skeleton. Half of the membrane mass in human and most mammalian erythrocytes are proteins. The plasma membrane is rich in proteins such as band 3 and active transporter (such as the Na+/K+ ATPase). The proteins of the membrane skeleton such as spectrin are responsible for the deformability, flexibility, and durability of the red blood cell. Specific interactions with lipids such as Phosphatidylserine (PS) are involved in the control of this flexibility. They enable recovering the discoid shape as soon as these cells are not submitted to compressive forces present in the capillaries. They are nonnucleated cells in their mature form. The mature (enucleated) erythrocytes lose all other cellular organelles. Due to the lack of mitochondria, they produce the energy carrier ATP by the glycolysis of glucose and lactic acid fermentation. Hemoglobin is concentrated within the RBC cytoplasm. Hypo-osmolar shocks can rupture the plasma membrane and hemolysis occurs through hemolytic pores. This is a result of the mechanical tensions associated to the induced osmotic swelling. Permeabilization of the RBC membrane results in an osmotic swelling, which if not prevented by osmotic agents (sucrose or larger carbohydrates, polymers such as dextran or PEG) is followed by hemolysis (swelling). With human and bovine red blood cells, electric field induced so-called dielectric breakdown of cell membranes could be detected using a Coulter Counter with a hydrodynamic focusing orifice (Zimmermann et al. 1974). In making measurements on red blood cells versus increasing electric field strength, the plot of the pulse current heights versus the estimated electric field strength (obtained from the applied voltage on the hydrodynamic flow chamber), a sharp bend in the classical linear curve (ohmic behavior), is observed associated to what was called the dielectric breakdown of the membranes. Solution of Laplace’s equation (assuming RBC being

Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation. . .

11

spherical) for the electric field generated in the flow chamber brings a value of about 1.6 V for the critical permeabilizing TMV. This phenomenon was used to study dielectric breakdown of red blood cells and the associated post pulse hemolysis in a homogeneous electric field between two flat platinum electrodes. The electric field was applied by a capacitor discharge generator (with flat parallel electrodes) on a batch chamber. The calculated value (from the Laplace equation) of the critical TMV in the homogeneous field is of the same order as when using the Coulter Counter. This supports the conclusion that the hydrodynamic forces in the orifice of the Coulter Counter do not play a role in inducing mechanical rupture of the red blood cells. The dielectric breakdown (membrane conductance increase) TMV of human red blood cells was measured to be at approximately 1 V (with field delivered with a capacitive discharge generator) (Zimmermann et al. 1976). It is not affected by the addition to the buffer of agents protecting against hemolysis (phosphate, sulfate, sucrose, inulin, and EDTA). Indeed, hemolysis is detected for field pulses with a larger amplitude. Electrical hemolysis is a secondary process of osmotic nature induced by the permeability change of the membrane (dielectric breakdown/creation of defects in the membrane) in response to an electric field pulse. Square wave electric pulses applied to the erythrocyte suspension is a more reliable approach where both the applied voltage and the delivered current are recorded on line with a μs time resolution. An electrochemical coating of the electrodes (platinum black) prevents the interfacial artifacts. Electric field induces a transmembrane potential that, at a critical point, induces a sharp increase in the current (conductivity). This results from a membrane structural alteration that either opens up preexisting channels or transporters or creates defects (Kinosita and Tsong 1977) (defects). At least two steps are involved in the formation of the conducting pathways: the initiation and the subsequent growth of the pore size. The induced TMV is not measured but again obtained from the Laplace equation under the simplification that a RBC is a sphere. The initiation step requires a transmembrane potential greater than a threshold (approximately 1 V), whereas the latter process is controlled by many factors such as the ionic strength, the field, or the pulse duration. Leakage of ions leads to an osmotic imbalance which results in a colloidal hemolysis of the red cells. Hemoglobin leakage and the associated craters are not mechanisms but results from red blood cell electropermeabilization. The critical TMV is associated to the initiation step.

Muscle Fibers A transmembrane voltage can be imposed on these giant cell plasma membrane by a simple chemical treatment and the associated current is measured (Chen and Lee 1994) (transmembrane voltage ). Transmembrane potentials of up to 120 mV are physiologic and well tolerated, but when the potential is more than 300 mV, large currents are observed due to the electropermeabilization. Currents through frog muscle cell membranes are observed under 4-ms pulses of up to 440 mV. The relaxation of transmembrane currents under this voltage clamp approach back to

12

J. Teissie

zero after the pulse is observed by using low transmembrane potential pulses and described in terms of membrane physical parameters. Electropermeabilization by supra-critical clamped TMVs results in damaging the K+ channels, through an electroconformational change in the transmembrane proteins. The membrane threshold of 300 mV was detected as a result of this protein conformational change that results in huge transmembrane current.

Patch Clamp Experiments Threshold transmembrane potential for the induction of conductive defects in plated NG108-15 neuroblastoma cells was obtained by treating small regions on the cell surface (Ryttsen et al. 2000). Tiny noncontact electrodes were used to deliver highly localized electric field on the cell membrane; 1 ms pulses were delivered with increasing voltages. From patch-clamp measurements, the transmembrane threshold voltage for pore formation of NG108-15 cells was determined to be 250 mV when corrected for the voltage drop at the double layer. A cell-attached patch clamp method was used to impose voltage waveforms of desired amplitude to the cell membrane, on single ventricular cardiac myocytes isolated from Sprague-Dawley rats using a collagenase and protease treatment procedure (Tung and O’Neill 1995). Pulse waveform with amplitudes up to 1 V (rectangular) were delivered using the patch pipette, as well as a low amplitude, rectangular “conductance monitoring” pulse train. The onset of electropermeabilization was characterized as the appearance of small fluctuations or jumps in current. The critical TMV was between 250 and 300 mV. Those transient bursts in current do not occur instantaneously with the onset of the voItage pulse, but rather with a variable delay that can occur over milliseconds. This reflects the complex mechanisms supporting membrane electropermeabilization. The patch clamp technique in the whole cell and the outside out configuration was used to monitor current-voltage relations of protoplasts obtained from a tobacco culture cell line (Wegner et al. 2013). Protoplasts were exposed to voltage pulses with amplitudes up to supra-physiological values. A transition from a low-conductance (0.1 nS/pF) to a high-conductance state (5 nS/pF) was observed when the membrane was either hyperpolarized or depolarized beyond threshold values of around 250 to 300 mV and +200 to +250 mV, respectively. K+ channel activity was affected through an increase of cytosolic Ca2+ that is known to inhibit outward-rectifying K+ channels in tobacco cells. But the high conductance is to be associated with the formation of aqueous membrane defects at those supraphysiological voltages. A fluorescence method confirmed the conductance experiments under the patch clamp approach. Plotting voltage-sensitive dye ANNINE-6 fluorescence intensity against command voltage steps revealed saturation of the curve at values +300 mV. This saturation is indicative of the effect of an enhanced conductance on the induced transmembrane voltage (conductivity).

Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation. . .

13

Transport Assay on a Cell Population Electropermeabilization is routinely detected by the transport assay (transport). Cells are pulsed in a dye-containing buffer and permeabilization (uploading of the dye) is investigated under increasing field conditions (at given pulse duration and number of delivered pulses at a 1 Hz frequency). Transport resulting from electropermeabilization occurs at a critical field strength Ep. Indeed experimentally this value is difficult to estimate. When cells are pulsed with fields of larger strengths, it is observed that more cells are detected as positive (dye is detected within the cytoplasm). Mathematical fittings of the electropermeabilization versus field strength can be used such as a two-parameter sigmoidal curve (Kotnik et al. 2001), yðxÞ ¼ 100%=ð1 þ exp:½ðxc  xÞ=bÞ

(8)

where x is the field pulse amplitude, y is the percentage of permeabilized cells, xc is the x value corresponding to y = 50%, and b determines the slope of the sigmoid curve, i.e., a fitting parameter. For dye uptake, the intensity of fluorescence can be fitted by a three-parameter Gaussian peak, h i yðxÞ ¼ ymax exp: ðxc  xÞ2 =2b2

(9)

where x is again the field pulse amplitude, y is the intracellular concentration of lucifer yellow, ymax is the maximum intracellular concentration of lucifer yellow in a given experiment, xc is the x-value corresponding to y = ymax, and b determines the width of the peak. But due to the sigmoidal character of the curve, it does not provide access to the critical field needed for electropermeabilization (i.e., conditions where the transport is detected.)

The Pulsing Protocol Affects the “Experimental” Critical TMV Interestingly, the critical parameter Ec is observed to decrease with an increase in the pulse duration and the number of repetitive pulses in a train. A linear relationship is observed between Ec(T) and the reciprocal of the pulse duration (T) or number of pulses (N). This gives access by extrapolation to a limiting value of Ec (for an infinite duration). As the pulse duration may affect the mechanism of electropermeabilization due to the nonlinear electrostretching, it may be safer to increase the number of pulses rather a very long single pulse duration. Using a delay between each pulse long enough to provide a significant cooling allows to avoid any Joule heating artifact.

14

J. Teissie

Imaging electropermeabilization with the uptake of fluorescence dyes such as PI showed that the critical Ec was different on the two cell hemisphere facing the electrodes (imaging). This was explained by the contribution of the resting TMV ΔV0 in the amplitude of the TMV present during the pulse delivery (see Eq. 5 and Fig.1). The rate of transport was further affected by the electrophoretic force acting on the charged PI. Changing the temperature of the pulsing buffer was not shown to affect the initiation step (i.e., the critical field needed to induce permeabilization). The membrane translational and rotational dynamics are not affecting the initial event in permeabilization (trigger of conducting defects), while they strongly affect the expansion step of those defects and as a result the level of transport. Osmotic pressure of the pulsing buffer has no effect on the induction step of permeabilization. Increasing the ionic strength of the pulsing medium only affects the expansion step. Molecules that modify membrane order (lysolecithin, ethanol), when used in concentrations compatible with cell viability, are shown to affect only the expansion step, but not the critical field that triggers electropermeabilization. Cell membrane fluidity does not have significant effect on reversible electroporation (in agreement with the effect of temperature). From the experimental values of Ec obtained with Propidium iodide uptake, taking into account the size distribution of the cell population, most studies used the Laplace equation to get access to the critical TMV taking for r the mean value. Using pulse duration of less than 1 ms within a train of 8 pulses, most published data were associated with critical TMV in the range of 0.4 to 0.5 V (transmembrane voltage). Pulse electric field effects on phagocytes (human blood neutrophils and rat peritoneal macrophages) have been investigated with an exponentially decaying pulse (t = 90 μs, E = 1 to 7 kV/cm) (Malinin et al. 1989). The biological response to electropermeabilization was to induce a chemiluminescent response due to a respiratory burst in the phagocytes. When the mean diameter of the neutrophils and macrophages is considered as 10 and 20 μm, respectively, the Laplace equation (Eq. 3) yields the same threshold TMV, 0.75 V, for the two types of cell. By monitoring the enhanced influx of Ca2+, the variability of the critical TMV inducing electropermeabilization of the plasma membrane of CHO cells (ΔVc) was observed to be very high in a population (Towhidi et al. 2008). It ranged from 512 to 1028 mV in those irregularly shaped cells. The TMV was computed on a 3D finiteelements model of each cell as the Laplace equation was not relevant. Cells of the same type and exposed to the same number of pulses with the same duration are affected by cell specific ΔVc. This cannot be explained by fluctuations in the resting TMV ΔV0. The description of a cell provides some explanations. A cell is not a smooth sphere. A lot of invaginations are present where the geometry and the biomechanical properties of the membrane are far from what is expected with a spherical shape. Enhanced local membrane curvature facilitates field-induced membrane defects (membrane curvature). Packing is more heterogeneous in this part of the membrane facilitating the creation of defects.

Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation. . .

15

Fluorescence Imaging Observations The use of a very fast fluorescence imaging microscopy connected to a pulse generator gives access to the conductance changes (Hibino et al. 1991) (imaging). A voltage-sensitive fluorescent dye was used as an indicator of the time-dependent transmembrane change induced by the external field on sea urchin eggs. But when the electric field was over a critical value, the transmembrane voltage kept locally a constant value on a cap facing the electrodes. The size of the conductive cap is under the control of the strength of the electric field. A high membrane conductance is therefore present within 2 μs after the onset of the external field and affects the fieldinduced TMV as predicted by Eq. 1. A steady value is observed in the cap where the transmembrane conductance is obtained (Hibino et al. 1993). The critical field is the one where the formation of this high conductance cap is formed. Nanosecond pulsed electric fields pulses with a duration of 60 ns, and maximum field strengths of approximately 100 kV/cm (100 V/cell diameter) are known to induce membrane electropermeabilization (Frey et al. 2006). Membranes of Jurkat cells were stained with a fast voltage-sensitive dye, ANNINE-6. A temporal resolution of 5 ns was achieved by the excitation with a ultrashort laser pulse of this dye due to its subnanosecond voltage response time. The laser pulse was synchronized with the applied electric field to record images at times before, during, and after exposure. Along the pulse, the TMV increase at the anodic pole of the cell was limited at 1.6 V after 15 ns and then decreased, indicating that electropermeabilization was present (transmembrane voltage). TMV on the side facing the cathode reached values of only 0.6 V in the same time period. The strong asymmetry in the induced TMV difference cannot be linked to the resting TMV ΔV0. Different conduction mechanisms seem to be present, which are associated to different critical potentials.

Electropermeabilization Is Inducing Membrane Fusogenicity. The field effect is not limited to permeabilization (Rems and Miklavcic 2016. When applied on intact cell suspension, electric field pulses are known to induce membrane permeabilization (electropermeabilization) and fusion (electrofusion) (membrane fusion). These effects are triggered through a modulation of the membrane potential difference. Due to the vectorial character of the electric field effects, this modulation is position-dependent on the cell surface. Fusion results from the contact between electropermeabilized cell membrane (Teissié and Rols 1993). When plated cells are at confluence, delivery of pulses can induce electropermeabilization and fusion (formation of polynucleated cells). The critical field to obtain permeabilization (dye uptake) at given pulse duration (0.1 ms) and number of pulses is lower than what is needed to get fusion (induction of polynucleated cells). This is the result of the contribution of the resting TMV as shown in Eq. 5. ΔV0 facilitated the induction of permeabilization on one side of the cell (low threshold in Ep for permeabilization) but must be overcome by a higher field Ef on the other side to get its

16

J. Teissie

permeabilization and the resulting membrane fusion. The critical membrane potential difference which induces membrane permeabilization can be calculated from the experimental observations of Ep and Ef and the value of ΔV0. It is observed that its value is always about 200 mV for many different cell systems. This is much less than assumed in most other studies on cell electropermeabilization.

What Lipid Systems Can Tell Us A study of the voltage induction of transient transport defects in phospholipid bilayer vesicles was obtained on Unilamellar vesicles (dipalmitoylphosphatidylcholine) (Teissie and Tsong 1981. The 100 nm vesicles loaded with [14C]sucrose were exposed to an intense electric field in the range of 20–40 kV/cm, with a field decay time of 5–15 ms. A transient leakage of labeled sucrose was detected when the field strength exceeded 30 kV/cm. When the TMV exceeded a threshold value of 200 mV, corresponding to an applied field strength of 30 kV/cm, there was a transient permeabilization of the bilayer structure. No global and permanent damages to the vesicle bilayer were detected. These results showed the implication of a membrane potential triggered conducting state of lipid bilayers in cell membrane electropermeabilization. More recently, similar results were obtained on giant vesicles (GUV) either by a loss of contrast showing the leakage of the internal content and the uptake of the buffer by a free exchange or by a fluorescence assay (Mauroy et al. 2015 (lipid vesicles). The experiments on lipid vesicles can provide a link with observations on planar lipid membranes (BLM). Palmitoy-oleoyl-phosphatidylcholine (POPC) bilayers were modified with incorporation of nonionic surfactant polyoxyethylene glycol (C12E8) surfactants (Troiano et al. 1998). A 22% decrease of the destabilizing TMV was observed upon addition of C12E8 to pure POPC bilayers. The addition of C12E8 at concentrations of 0.1, 1, and 10 μM to the bath surrounding the membranes decreased the electropermeabilization threshold monotonically with concentration. These results suggested that the polyoxyethylene glycol molecules play a major role in the formation of hydrophilic defects in the bilayers above the critical TMV. The C12E8 head groups are supporting the defect formation without the need of major lipid head-group rearrangements. Surfactants can thus be used to manipulate the electropermeabilization threshold of lipid bilayers and cell membranes. The addition of the channel-forming peptide gramicidin D (gD) on the conductance and electropermeabilization thresholds of planar bilayer POPC membranes in a ratio of 1:10,000 had no effect, but ratios of 1:500 and 1:15 significantly increased the thresholds by 16% (p < 0.0001) and 40% (p < 0.0001), respectively (Troiano et al. 1999). The effect of gD (1/15 ratio) on the membrane area expansivity modulus (K) of GUVs was an increase by 110%. The effect is consistent with the prediction of electromechanical model of electropermeabilization. The presence of membrane

Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation. . .

17

proteins may affect the electropermeabilization of cell membrane by a local change in their mechanical properties.

Conclusions Critical electric field and transmembrane voltage for membrane defect formation is a complex experimental problem. Very few direct evaluations of the critical TMV are described, and most results are obtained from the effect of external field pulses. Obtaining the TMV from the external field value is mostly through the use of the Laplace equation considering the cell as a regular sphere (Laplace equation). This is clearly a crude approximation. In a few cases, new numerical approaches should give us more accurate insights on more complex geometries (Kotnik and Miklavčič 2000). The other experimental limit is that the detection of permeabilization is under the control of the sensibility of the assay. This may explain why there is no consensus between the different experiments. One open question is indeed the physical definition of the critical TMV. The clear conclusion is that Electropermeabilization still needs more basic investigations (Teissie et al. 2005; Rems and Miklavcic 2016). Acknowledgments Research was conducted in the scope of the EBAM European Associated Laboratory (LEA) and resulted from the networking efforts of the COST Action TD1104 (http:// www.electroporation.net).

Cross-References ▶ Different Approaches Used in Modeling of Cell Membrane Electroporation ▶ Electric Pulse Parameters Affecting Electroporation Treatment Outcome ▶ Electrical Conductance of Lipid Pores ▶ Electroporation and Electropermeabilization ▶ Experimental Electroporation of Planar Lipid Bilayers ▶ Fluorescent Indicators of Membrane Permeabilization Due to Electroporation ▶ Measurement of Molecular Transport into Electropermeabilized Cells ▶ Mechanistic Description of Membrane Electropermeabilization ▶ Membrane Interfacial Water Dipoles and Electropore Formation ▶ Membrane Permeabilization Lifetime in Experiments ▶ Modeling Transport Across the Electroporated Membrane ▶ Parameters Affecting Cell Viability Following Electroporation In Vitro ▶ Patch Clamp in Use of Electroporation Mechanisms Studies ▶ Phospholipid Head Group Dipoles and Electropore Formation ▶ Pulsed Electric Fields Treatment of Biological Suspensions ▶ Single-Cell Electrical Characterization Techniques ▶ Transmembrane Voltage Induced by Applied Electric Fields ▶ Water Defects in Phospholipid Bilayers

18

J. Teissie

References Chen W, Lee RC (1994) Altered ion channel conductance and ionic selectivity induced by large imposed membrane potential pulse. Biophys J 67(2):603–612 Frey W, White JA, Price RO, Blackmore PF, Joshi RP, Nuccitelli R, Beebe SJ, Schoenbach KH, Kolb JF (2006) Plasma membrane voltage changes during nanosecond pulsed electric field exposure. Biophys J 90(10):3608–3615 Hibino M, Shigemori M, Itoh H, Nagayama K, Kinosita K Jr (1991) Membrane conductance of an electroporated cell analyzed by submicrosecond imaging of transmembrane potential. Biophys J 59(1):209–220 Hibino M, Itoh H, Kinosita K Jr (1993) Time courses of cell electroporation as revealed by submicrosecond imaging of transmembrane potential. Biophys J 64(6):1789–1800 Kinosita K Jr, Tsong TT (1977) Hemolysis of human erythrocytes by transient electric field. Proc Natl Acad Sci USA 74(5):1923–1927 Kotnik T, Miklavčič D (2000) Analytical description of transmembrane voltage induced by electric fields on spheroidal cells. Biophys J 79:670–679 Kotnik T, Mir LM, Flisar K, Puc M, Miklavcic D (2001) Cell membrane electropermeabilization by symmetrical bipolar rectangular pulses. Part I Increased efficiency of permeabilization. Bioelectrochemistry 54(1):83–90 Kramar P, Delemotte L, Maček Lebar A, Kotulska M, Tarek M, Miklavčič D (2012) Molecularlevel characterization of lipid membrane electroporation using linearly rising current. J Membr Biol 245(10):651–659. doi:10.1007/s00232-012-9487-6 Malinin VS, Sharov VS, Putvinsky AV, Osipov AN, Vladimirov YA (1989) Chemiluminescent reactions of phagocytes induced by electroporation: the role of Ca*+ and Mg *+ ions. Bioelectrochem Bioenergetics 22:37–44 Mauroy C, Rico-Lattes I, Teissié J, Rols MP (2015) Electric destabilization of supramolecular lipid vesicles subjected to fast electric pulses. Langmuir 31(44):12215–12222. doi:10.1021/acs. langmuir.5b03090 Melikov KC, Frolov VA, Shcherbakov A, Samsonov AV, Chizmadzhev YA, Chernomordik LV (2001) Voltage-induced nonconductive pre-pores and metastable single pores in unmodified planar lipid bilayer. Biophys J 80(4):1829–1836 Needham D, Hochmuth RM (1989) Electro-mechanical permeabilization of lipid vesicles. Role of membrane tension and compressibility. Biophys J 55(5):1001–1009 Neumann E, Kakorin S, Toensing K (1999) Fundamentals of electroporative delivery of drugs and genes. Bioelectrochem Bioenerg 48(1):3–16 Pavlin M, Kandušer M, Reberšek M, Pucihar G, Hart FX, Magjarević R, Miklavčič D (2005) Effect of cell electroporation on the conductivity of a cell suspension. Biophys J 88:4378–4390 Pucihar G, Kotnik T, Valic B, Miklavcic D (2006) Numerical determination of transmembrane voltage induced on irregularly shaped cells. Ann Biomed Eng 34(4):642–652 Rems L, Miklavcic D (2016) Tutorial: electroporation of cells in complex materials and tissue. J Appl Phys 119:201101. doi:10.1063/1.4949264 Ryttsen F, Farre C, Brennan C, Weber SG, Nolkrantz K, Jardemark K, Chiu DT, Orwar O (2000) Characterization of single-cell electroporation by using patch-clamp and fluorescence microscopy. Biophys J 79:1993–2001 Teissie J, Tsong TY (1981) Electric field induced transient pores in phospholipid bilayer vesicles. Biochemistry 20(6):1548–1554 Teissié J, Rols MP (1993) An experimental evaluation of the critical potential difference inducing cell membrane electropermeabilization. Biophys J 65(1):409–413 Teissie J, Golzio M, Rols MP (2005) Mechanisms of cell membrane electropermeabilization: a minireview of our present (lack of ?) knowledge. Biochim Biophys Acta 1724(3):270–280 Towhidi L, Kotnik T, Pucihar G, Firoozabadi SM, Mozdarani H, Miklavcic D (2008) Variability of the minimal transmembrane voltage resulting in detectable membrane electroporation. Electromagn Biol Med 27(4):372–385. doi:10.1080/15368370802394644

Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation. . .

19

Troiano GC, Tung L, Sharma V, Stebe KJ (1998) The reduction in electroporation voltages by the addition of a surfactant to planar lipid bilayers. Biophys J 75(2):880–888 Troiano GC, Stebe KJ, Raphael RM, Tung L (1999) The effects of gramicidin on electroporation of lipid bilayers. Biophys J 76(6):3150–3157 Tung L, O’Neill RJ (1995) Comparison of electroporation thresholds of cardiac cell membranes by rectangular and exponential pulses IEEE-EMBC and CMBEC Theme 1: Cardiovascular System Engineering in Medicine and Biology Society, IEEE 17th Annual Conference 251–252. doi: 10.1109/IEMBS.1995.575095 Wegner LH, Frey W, Schönwälder S (2013) A critical evaluation of whole cell patch clamp studies on electroporation using the voltage sensitive dye ANNINE-6. Bioelectrochemistry 92:42–46. doi:10.1016/j.bioelechem.2013.03.002 Zimmermann U, Pilwat G, Riemann F (1974) Dielectric breakdown of cell membranes. Biophys J 14(11):881–899 Zimmermann U, Pilwat G, Holzapfel C, Rosenheck K (1976) Electrical hemolysis of human and bovine red blood cells. J Membr Biol 30(2):135–152

Phospholipid Head Group Dipoles and Electropore Formation Justin Teissie

Abstract

Classical electropermeabilization is the result of the delivery of electric field pulses on cells. The field pulse lasts from submicro to several milliseconds. The field intensity is large enough to induce a dramatic structural local alteration of the cell membrane organization. This results in an enhanced permeabilization of the target cell membrane. This is indeed a complex process and its molecular characterization remains an intense field of investigations. A membrane is a complex assembly where the polar head region is a charged system. It appears as a target of the field effect. A description of the lipid bilayer organization points out its role in the function of membrane. Interfacial water appears as a key component in the structural organization of the phospholipid head groups. A feedback between polar head dipoles and water dipoles results in a specific organization. Different biophysical approaches were developed to investigate the behavior of the phospholipid head group dipoles during and after the pulse delivery. The experiments were performed on membrane lipid models as well as on cell membranes under reversible electropermeabilization, where the cell viability is preserved through a resealing active process. The external field pulse affects the mean orientation of the head group dipoles and alters the local water content and self-organization. This is a key step in the induction of defects in electropermeabilization and their associated properties. Keywords

Membrane • Transmembrane voltage • Membrane structure • Interface • Polar head group • Dipole • Interfacial water J. Teissie (*) Institut de Pharmacologie et de Biologie Structurale, Université de Toulouse, CNRS, UPS, Toulouse, France e-mail: [email protected]; [email protected] # Springer International Publishing AG 2017 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_78-1

1

2

J. Teissie

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Classical Description of Electropermeabilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The External Field Induces a Transmembrane Voltage (TMV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Resulting TMV Induces Membrane Permeabilization (Electropermeabilization) . . . . . . A Cell Membrane Is Not Just a Thin Dielectric Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electropermeabilization on Lipid Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vesicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multilayer Lipid Stacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Planar Single Bilayer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Polar Head Group Reorientation in Pulsed Supported Lipid Multilayers . . . . . . . . . . . . . . . . . . . Head Group Reorientation in “In Silico” Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electropermeabilization on Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interfacial Modifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Field Pulse Delivery Results in Asymmetric Event on the Cell Surface . . . . . . . . . . . . . . . . . . . . Structural Events at the Cell Membrane Interphase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 2 2 3 4 6 6 7 8 9 10 10 10 11 13 14 14 14

Introduction Short high-field calibrated electric pulses when delivered to a cell suspension or a tissue result in a membrane permeabilization (electropermeabilization). Classical electropermeabilization results in an enhanced transport across the membrane and in a change in its electrical conductivity. The general consensus is that the structural transition in the membrane is due to the transmembrane voltage resulting for the external field. The membrane is considered as a dielectric thin sheet. This is clearly an oversimplification. Structural investigations of lipid bilayers pointed out the complexity of the interfacial layer between the bulk water and the hydrocarbon region. This chapter gives a survey of the recent knowledge on this organization. Charge and dipole distribution in this 1 nm layer make it a target to the field effect (forces and torques). In the second part of the chapter, the results obtained on the field-induced effects on phospholipid head group dipoles by in silico simulations and experiments on lipid models and on cells are presented with their consequences on the electropermeabilization process.

The Classical Description of Electropermeabilization The External Field Induces a Transmembrane Voltage (TMV) When a cell is submitted to an external electric field, the cell (where the plasma membrane can be considered as a dielectric shell) should be considered as a spherical capacitor, where the redistribution of electrophoretically driven charged ions in the electrolytes surrounding the membrane, i.e., electric current, leads to an induced

Phospholipid Head Group Dipoles and Electropore Formation

3

transmembrane voltage (TMV). During the electric field application (electropulsation) on the cell, an electro-induced membrane electric potential difference (ΔVi) is created which is locally associated with the dielectric properties of the plasma membrane. Using a physical model based on a thin, weakly conductive shell (the membrane conductivity λm) full of an internal conductive medium (the cytoplasm conductivity λi) and bathed in an external conductive medium (conductivity λe), solution of the Laplace’s differential equation gives ΔVi as
ΔViðM, E, tÞ ¼ f gðλÞrE cos θðMÞ 1  expðt=τmÞ

(1)

where M is the point on the cell that is considered, t is the time after electropulsation is turned on, f is a factor depending on the cell geometry (for a sphere, f = 1.5), r is the radius of the pulsed cell, E is the electric field strength, and (M) is the angle between the direction of the field and the normal of the cell surface in M. g(λ) is related to the different conductivities as (Zimmermann et al. 1974) gðλÞ ¼



h i 2λe 2λm þ λi þ ðλm  λiÞðr  d=r Þ3  3λmðr  d=r Þ h i = ð2λe þ λmÞð2λm þ λiÞ þ 2ðr  d=r Þ3 ðλi  λmÞðλm  λeÞ

(2)

where d is the thickness of the membrane (a few nm) and τm is the characteristic time constant of the membrane charging and can be written as (Kinosita and Tsong 1977) τm ¼ rCmð2λe þ λiÞ=ð2λeλiÞ

(3)

where Cm (0.5–1.0 μF/cm2) is the specific membrane capacitance. τm is calculated to be in the submicrosecond time range for mammalian cells. It is strongly dependent on the buffer composition as the internal composition is fixed by the cell metabolism.

The Resulting TMV Induces Membrane Permeabilization (Electropermeabilization) When the new membrane electric potential difference ΔVm(ΔV0 + ΔVi) locally reaches a critical value (ΔVc) (nowadays estimated between 0.25 and 0.6 V for living cells) (Rems and Miklavcic 2016), a local alteration of the membrane structure leads to membrane permeabilization. Molecular processes supporting classical electropermeabilization remain poorly understood (Teissie et al. 2005) in spite of more than 40 years of effort (Neumann and Rosenheck 1972). The membrane is called “porous.” There is a general agreement that it cannot be adequately described simply by the occurrence of reversible hydrophilic holes, so-called pores (“electroporation”), in the lipid bilayer. An accurate biophysical description of a cell membrane is far from the simplifying description of using a physical model based on a thin, weakly conductive shell (the membrane conductivity λm).

4

J. Teissie

A Cell Membrane Is Not Just a Thin Dielectric Layer Lipid bilayers have been studied for more than 50 years that it is missed by nonexperts in the field just how uncertain are structural quantities for the fully hydrated, fluid (Lα or liquid crystalline) phase, that is relevant to biology (Nagle and Tristan Nagle 2000). While lipid crystallography has been pursued and has provided many relevant informations, it is important to recognize that fully hydrated lipid bilayers are not even close to being considered as a crystalline state. This is mostly valid for bilayers that are in the Lα phase because the hydrocarbon chains are conformationally disordered in contrast to the all-trans chains present in lipid crystals. Fluctuations, inherent in flexible and biologically relevant lipid bilayers, make quantitative structure determination challenging. It makes no sense to describe a lipid bilayer as if it had a regular crystalline structure because of the fluctuations. Such crystalline structures simply do not exist in the biologically relevant state. The appropriate description for the positions of atoms in the lipid molecule is that of broad statistical distribution functions (that are non-Gaussian). Disorders present on a long range are not inherent in the structure of the bilayer but intimately involve bilayer undulations. Fluctuations are a central feature in membrane biology. Local fluctuations and the induced transient mismatches support passive diffusion of solutes and facilitate the conformational changes of transmembrane proteins. Fluctuations mean transient random packing defects. They support the lateral diffusion of membrane molecules (fast for lipids). They are locally enhanced due to the membrane curvature, where peptide insertion is facilitated. A description of the bilayer structure should take into account two regions. The central part is the hydrocarbon chain region, but the end region is a mix between the polar head group (carbonyls, glycerol, phosphate, and end groups such as choline) and water molecules. There is indeed no sharp boundary between the regions as shown by the overlap of the distribution functions. By the use of nuclear Overhauser enhancement spectroscopy (NOESY) and spinning under the magic angle of lipid multilayers, it was shown that lipid molecules are flexible, and the correlation times between their different conformations are shorter than nanoseconds. Indeed, proximity between polar head groups and the end of hydrocarbon chains is not a rare event. Shifts in the position of lipid segments between neighboring molecules are observed. Water is strongly interacting with polar head groups. It is well accepted that at least 12 water molecules provide the first strong hydration shell for dioleoylphosphatidylcholine (DOPC). It is not surprising that stripping off essential water involves strong forces that should result in drastic structural changes. The polar region in lipid bilayers, where the polar head groups of lipids are present, built an interphase, between the bulk of the aqueous phase and the hydrophobic core of the fatty acid chains. In this region, ionization of phospholipids as well as adsorption of ions to lipid head groups generates electrostatic potentials (Tocanne and Teissie 1990). They are added to polarization potentials associated to the chemical structure of the polar head groups and to their hydration shell. These

Phospholipid Head Group Dipoles and Electropore Formation

5

complex potentials modulate the interphase concentration of protons, cations, and anions with respect to those in the bulk aqueous phase. In the bilayer, lipids are packed with their polar head preferentially oriented close to a perpendicular direction to the interface normal. As already mentioned, due to fluctuations, lipids are also free to diffuse laterally, and polar heads undergo fast vibrational and rotational motions. This reflects the complexity of this interphase region. Strong interactions exist between lipid polar head groups and make feasible the exchange of ions such as protons between adjacent lipid molecules. The dielectric constants which are measured at the surface of two phospholipid species, PC (phosphatidylcholines) (zwitterionic) and PS (phosphatidylserine) (charged), stand around 40–10. For these lipids, the larger dielectric constants are found located at the interface, near the bulk water. The lower values are observed located deeper in the lipid matrix, at the polar head/hydrocarbon core, in the region of high probability of the ester carbonyl groups. A very steep gradient for the dielectric constants in the interphase region is present with very local time-dependent fluctuations. The dielectric constants on average decrease from 80 in the bulk phase to a value of about 2 in the central hydrocarbon domain of the lipid bilayer. This dramatic change takes place over a distance which appears to be less than 1 nm and is affected by the bilayer undulations. From an electrical point of view, a membranebulk phase interface cannot be considered as a sharp boundary (Nymeyer and Zhou 2008). Of the individual potentials which comprise the potential profile of a membrane, the least well understood is the dipole potential (Brockman 1994). In general, the dipole potential is manifested between the hydrocarbon interior of the membrane and the first few water layers adjacent to the lipid head groups. For a typical zwitterionic phospholipid, like phosphatidylcholine, its measured value is approximately 280 mV in bilayer membranes, the hydrocarbon region appearing positive relative to the bulk phase. Dipole potentials result from the ability of lipid head groups to globally reorganize water structure at the interface. The functional group dipoles of the terminal methyl groups of aliphatic chains, the glycerol-ester backbone, and the hydrated polar head groups all contribute to the dipole potential. Any physical stress affecting the organization (change in dipole orientation due to a torque induced by an external field) results in a change in the dipole potential. Interphase water is involved in its value, but the precise description of the microscopic nature of the interactions from which it arises is far from being known. One should take into account that the dipole potential is built in a thin region over which the dielectric constant of the medium is changing from 2 to 80. Even if poorly characterized, the dipole potential appears as an important regulator of membrane dynamical structure and associated function. The hydration force is controlled by the dipole potentials of the membranes. It is the leading factor in the close contacts in membrane-membrane and membrane-ligand interactions. For multicomponent membranes (the case in cells), the size and shape of lipid domains can be in a first approach considered due to the balance between the line tension at the domain borders and the difference in dipole density between the domains. Line tension tends to make the domains regular,

6

J. Teissie

whereas dipole repulsion promotes transitions to complex shaped larger domains (control of so-called rafts). Rotation of dipoles would affect the size and shapes of domains! Cell membranes cannot be described as a lipid bilayer with embedded transmembrane proteins. The membrane-bound protein skeleton, a network of actin filaments and associated proteins, covers the entire cytoplasmic surface and is closely linked to clathrin-coated pits and caveolae, key factors in membrane traffic (Morone et al. 2006). The actin filaments are likely to form the limit of the membrane compartments where membrane molecules are confined (making a “fence” preventing large scale diffusion). The plasma membrane is a patchwork with regard to their lateral diffusion. It is not a fluid bidimensional mosaic. A membrane in a cell is a more complex system than a homogeneous bilayer. Membrane traffic is continuously present. The plasma membrane is an out of equilibrium complex from a thermodynamic point of view. It forms a complex with the extracellular matrix and the cytoskeleton. Its transmembrane potential is dependent on the cellular energy reserves. As a conclusion, a cell membrane or even the “simple” single lipid bilayer is not a thin dielectric sheet. Fluctuations in the relative position of the different partners are affecting its organization. Definition of the “structure” should take into account all the internal movements. The interphase with the bulk phase remains poorly understood but is the region where a sharp gradient in the dielectric constant is present and where the dipolar character of polar heads makes them sensitive to external electric fields, but the associated torques are in opposite direction on the different water membrane interphases.

Electropermeabilization on Lipid Models Vesicles Light Scattering Unilamellar small lipid vesicles (SUVs, 30 nm in diameter), doped with the diphenylhexatrienyl-phosphatidylcholine (beta-DPH pPC) lipids, were observed by turbidimetry and light absorption when pulsed with short high-field rectangular electrical field pulses (Neumann et al. 1998). The major changes of the turbidity and absorbance dichroism result from vesicle elongation under electric Maxwell stress. The kinetics of this electro-chemo-mechanical shape deformation are illustrative of the very fast events of membrane stretching resulting in smoothing of thermal undulations. Local displacements of the chromophore (beta-DPH pPC) relative to the membrane normal are triggered by the pulsed field. Slower changes of the turbidity and absorbance signals reflect the entrance of solvent into the membrane/ medium (change in the local refractive index), resulting from the alignment of the bipolar lipid head groups in one of the leaflets at the pole caps of the vesicle bilayer.

Phospholipid Head Group Dipoles and Electropore Formation

7

One limit of this membrane system is that due to their small size, a high curvature is present on the external layer, and packing defects are abundant on the external layer.

Imaging An externally applied electric field across giant unilamellar vesicles (GUVs) (20–30 μm in diameter) leads to transient permeabilization of the membrane as assayed by a loss of optical contrast due to a transmembrane exchange. The distribution and lifetime of these permeabilizing defects on DOPC phospholipid vesicles were assayed under a fluorescent microscope as well as from negatively stained electron micrographs (after fixation) (Tekle et al. 2001). An asymmetrical behavior was observed. During the pulsed field application, a single membrane hole up to 7 μm in diameter (i.e., huge on the vesicle surface) was observed on the vesicle side facing the negative electrode. Nothing was detected at the anode-facing hemisphere where Ca2+ influx was nevertheless present supporting the formation of many defects supporting its transport. A clear difference in the permeabilizing defects is present associated with the direction of the transmembrane electric field (parallel to the external field). Coherent Anti-Stokes Raman Scattering (CARS) Microscopy Spontaneous Raman spectroscopy (SRS) is a powerful probe in the analysis of the chemical bounds of samples. Electropermeabilization-linked molecular events were monitored by comparing the vibrational spectra of biological sample before and after pulsed electric field through the detection of critical bandwidth. Coherent antiStokes Raman scattering (CARS) microscopy acquires images of a sample at a specific vibrational frequency. Based on the critical bandwidths previously determined with Raman spectroscopy, it is possible to follow the evolution of the biochemical composition of the membrane and associated water molecules during a pulsed electric field. The behavior of interfacial water molecules was observed on lipid GUVs. Two different spectral signatures were obtained associated to two different hydrogen bonds: one for lipid associated and the other for the water clusters. A dramatic drop of the water-associated signal is induced by the pulsed electropermeabilizing field. This reflects that interfacial water organization due to the lipid polar heads is dramatically affected in the electropermeabilized state.

Multilayer Lipid Stacks Polar Head Group Dipoles Are Sensitive to Their Local Environment Phosphorus-31 and deuterium nuclear magnetic resonance (P31 NMR and 2H NMR) measurements on PC, where the phosphocholine head group that was deuterated at the two methylene segments monitored the head group dipole orientation on lipid multilayers. The electric surface charge is the essential driving force in changing the phospholipid head group orientation and conformation (Seelig et al. 1987). While

8

J. Teissie

the P-N dipole is rotating nearly parallel to the membrane surface in the pure phospholipid membrane, the addition of a positively charged amphiphile or the binding of cationic molecules moves the N+ end of the dipole toward the water phase, changing the orientation of the phosphate segment by more than 30 at the highest amphiphile concentration. The effect is present only on the PC head groups with no perturbations on the hydrocarbon region and the glycerol backbone. Hydration of the lipid stacks can be controlled. When it is changed in the range of 10–70 wt.% H2O, a distinct change in the field-induced alignment of the PC head group is present. At low hydration, the N+ end of the PC head group dipole moves closer to the hydrocarbon layer. A similar conformational change is obtained upon addition of polyhydroxyl compounds. Perturbations in the polar head environment (local electric charge, level of hydration) result in a change in the orientation of their dipole. Such an effect affects the dipolar potential and its associated control on the membrane functions.

Planar Single Bilayer Bilayer lipid membranes (black lipid membranes, BLM) are a convenient model to observe transmembrane potential effects on lipid bilayers. The transition at a critical value of the applied electric field (transmembrane electric field) is characterized by a sharp increase in conductance and represents electropermeabilization. Under current-clamp conditions, the classical rupture of the membrane is avoided, and the process is observed on a long term at the critical transmembrane voltage (Genco et al. 1993). It depends dramatically on pH and ionic concentration with zwitterionic lipids, indicating that the surface electrical properties of the interphase determine the characteristics of such a transition (electropermeabilization). Asymmetric bilayers consisting of a charged and an uncharged monolayer mimic natural membranes. The instability inducing the increase in conductance starts in the uncharged monolayer, and the final electropermeabilization process results from the coupling between the two monolayers. The stability and rupture kinetics of bilayer negatively charged lipid membranes covered with electrostatically adsorbed polyelectrolytes, polylysines (PLs) of different molecular weights (MW), were analyzed under the short-voltage clamp conditions (Diederich et al. 1998). Adsorption of PL on one side of the bilayer results in an asymmetric transmembrane potential, which adds to the externally applied voltage. High MW PL decreases the critical breakdown voltage of the membrane significantly but also increases the delay time between the onset of the voltage pulse and conductance increase. The time course of conductance increase (defect expansion) is PL molecular weight dependent. Interfacial adsorption of high MW PLs causes a dramatic decrease of the velocity of defect expansion. 2D (surface) viscosity is increased by the interphase modification and affects the electropermeabilization process.

Phospholipid Head Group Dipoles and Electropore Formation

9

Polar Head Group Reorientation in Pulsed Supported Lipid Multilayers Stacks of lipid bilayers can be formed and spread on the surface of conductive electrodes. The polar head region can be directly observed by different spectroscopic methods when the field is present and induce the conductive state of the assembly. The influence of the field pulse on the molecular structure of the phosphatidylcholine head group was monitored by NMR of the phosphorus atom. The shape and the position of the peak in the P31 NMR signal reflect the average configurational position of the phosphate group from the polar head region. Multiple bilayers of dimyristoyl phosphatidylcholine and potassium oleate were macroscopically oriented between silver-coated slides, and a voltage bringing an electric field of up to 105 V.cm1 was delivered between the parallel slides (Stulen 1981). At electric fields above 104 V.cm1, spontaneous current fluctuations associated with the formation of the conducting defects are observed; structural effects are transient and reversible. The conformation of the head group is greatly affected as shown by the drastic change in the P31 NMR signature. A broadening and a shift of the peak are obtained. This reflects a dramatic average configurational change of the phosphate group region when the permeabilizing pulse is present but no definitive detail can be obtained. Polarized attenuated total reflection (ATR) experiments were designed to study the effect of electric fields on membrane molecule structure and orientation by Fourier transform-infrared (FTIR) spectroscopy (Le Saux et al. 2001; Miller 2002). Fully reversible orientational changes in the lipid head groups are induced by an electric potential difference field delivered on an oriented stack of lipid bilayers made of dioleoylphosphatidylcholine (DOPC) or of dipalmitoyl phosphatidylcholine (DPPC). Potentials above 600–700 V on 1-μm-thick multibilayer are needed to trigger the direct reorientation of the dipoles of the different polar residues by the electric field. The electric fields varied up to 5.5  106 V/cm in the hydrocarbon layer and up to 1.1  106 V/cm in the polar head region (being considered as a homogeneous dielectric!) due to the change in the dielectric constant between the interphase and the hydrocarbon regions. The inhomogeneous field between the polar and the hydrocarbon layers drives the polar groups into the hydrocarbon layer, exerting a pressure and penetrating them. Polarization-modulation infrared reflection-absorption spectroscopy (PM-IRRAS) experiments provided IR spectra for 1,2-dimyristoyl-sn-glycero-3phosphocholine (DMPC) bilayer formed on or close to gold electrodes (Zawisza et al. 2003). Changing the electrodes surface charge density induces a modulation of the electrode potential and of the local static electric field present on the bilayer. The quantitative analysis of the IR bands provided information on both the conformation and orientation of the polar head region of the DMPC molecule. The hydration and conformation of the head group of the DMPC molecule strongly depend on the electrode potential.

10

J. Teissie

The different spectroscopic approaches (P31 NMR, ATR-FTIR, PM-IRRAS) all brought the same information: the polar head region mean orientation is affected when a permeabilizing electric field is present. As a result, the distribution of water in the interphase region is changed.

Head Group Reorientation in “In Silico” Simulation New insights on the molecular process are given by computer-assisted simulations (Vernier and Ziegler 2007). Conditions for the simulation are very high field (300 MV/m) (higher than under experimental conditions where the voltage to electrode width is about 0.1 MV/m) and short pulse duration (a few ns at the best) due to the computing time limit. Using molecular dynamic simulations on one single phospholipid bilayer, defect formation results from a dramatic change in the organization of the interfacial water molecules. The process of conductive defect formation is driven by the transition to a more energetically favorable configurations of water dipoles at the water-lipid interfaces. The water molecular dipoles are oriented in the external field (torque on the dipole). Interfacial water molecules are the only players in the process, being its seed and its driving force. The field-generated torque nevertheless tilts the mean head group dipole orientation by a few degrees away from its equilibrium (at no field) position relative to the bilayer normal. Field-directed rotation of the head group dipoles, in combination with water dipole reorientation and solvation interactions at the aqueous-lipid interface, participates to the coordinated ensemble leading to the defect formation.

Electropermeabilization on Cells Interfacial Modifications The organization of the membrane water interphase (dipolar potential) gives rise to a set of different forces that are called hydration forces (Leikin et al. 1993). Due to their repulsive nature, they prevent the spontaneous fusion of two cells when their respective membranes are brought in close contact. A strong repulsion is detected when the intermembrane space is reduced to a few nanometers. The associated forces are described resulting from a local organization of the interfacial water. As shown on lipid layers, water dipoles are sensitive to the local field, and as a result a regular organization of the water takes place, bringing a collective regular assembly of dipoles. The associated dipolar fields on the two membranes are in opposite directions during the close approach. A strong repulsion is observed when a close contact is attempted. When cells are electropermeabilized, a spontaneous fusion is detected (Sowers 1986). The repulsive forces vanished during the pulse delivery, but the ability of the membranes to coalesce remains present in the few seconds following the pulse delivery. This is indicative that the structures supporting these repulsive forces, i.e., the regular organization of interfacial water molecules, have

Phospholipid Head Group Dipoles and Electropore Formation

11

been altered under the electropermeabilization processes. This interfacial modification needed to obtain cell membrane coalescence is restricted to the electropermeabilized part during and in the few seconds following the pulse delivery. This local perturbation is present only where the critical electropermeabilization transvoltage has been induced by the external field. Spontaneous Raman spectroscopy (SRS) is a powerful probe in the analysis of the chemical bounds not only on simple lipid assemblies as reported above but on cells of samples. Cells were pulsed, and the CARS signature of the water molecules was monitored before and after the pulse. Again a dramatic drop of the water-associated signal is induced by the pulsed electropermeabilizing field. This reflects that interfacial water organization due to the interphase is dramatically affected in the electropermeabilized state, as predicted by the spontaneous fusion of electropermeabilized cells.

Field Pulse Delivery Results in Asymmetric Event on the Cell Surface Asymmetry in the Critical TMV Nanosecond pulsed electric fields (nsPEF) with a “long” duration of 60 ns and high field strengths up to 100 kV/cm (100 V on a 10-μm cell diameter) are shown to trigger membrane permeabilization by a process rather similar to the one in classical electropermeabilization. A fast voltage-sensitive dye, ANNINE-6, has a subnanosecond voltage response time and was used to follow the change in the transmembrane voltage during the pulse delivery (Frey et al. 2006). A temporal resolution of 5 ns of the fluorescence signal (i.e., of the transmembrane voltage (TMV)) was obtained by the excitation of this dye with a tunable laser pulse. The very short dye excitation by the flash was synchronized with the longer applied electric field pulse to observe cell imaging (spatial resolution) at times before, during, and after the field delivery. The TMV reached values of 1.6 V after 15 ns at the anodic pole of the cell, several times larger than the TMV level supposed to be required for inducing a conductive state of the membrane. The TMV then decreases illustrative of a high local membrane conductance. TMV on the side facing the cathode reached values of only 0.6 V in the same time period, indicating a strong asymmetry in the field-dependent conduction mechanisms created in the membranes of the two opposite cell hemispheres. The difference in behavior of the caps of a cell facing the two electrodes was present in classical electropermeabilization. The membrane conductance of sea urchin eggs was investigated over the time range of 0.5–1 ms by directly imaging the transmembrane potential at a submicrosecond resolution with the voltagesensitive fluorescent dye RH292 (Hibino et al. 1993). When a rectangular electric pulse of a non-permeabilizing intensity was delivered on a cell, a position-dependent fluorescence change illustrative of the TMV developed synchronously with the pulse. Under an electric pulse of higher intensity, the rise of the induced potential stopped at a certain level and then slowly decreased on the microsecond time scale. Again this is the result of the creation of finite membrane conductance or

12

J. Teissie

permeabilization of the membrane. The two opposite sides of the egg facing the positive and negative electrodes were affected; a slight asymmetry was observed as the conductance on the positive side appeared higher. During the pulse duration, the conductance increased steadily up to ten times during the next 1 ms. Again asymmetry in the processes is detected. This conductance increase was larger on the negative electrode side; by 1 ms the conductance on the negative side was more than twice that on the positive side, while it was lower at the 0.5-μs image. The defects induced by the critical TMV are structurally different on both sides of the cell membrane facing the electrodes.

Asymmetry in the Transport by Diffusion Fluorescence imaging under the microscope provides an approach of the transmembrane transport in cells with a good spatial resolution. Transport of divalent cation (Ca2+) and three probes with a high affinity for DNA (ethidium bromide (EB), propidium iodide (PI), and ethidium homodimer (EthD-1)) across electropermeabilized membranes of several mammalian cell lines was found to be selective and asymmetrical (Tekle et al. 1994). The conductivity of the pulsing buffer affects the ionization of the cell surface (charged groups are present on the external layer of a cell membrane), and it is observed to modulate the transport. In low-salt medium (where electrostatic forces are strong), Ca2+ and EB were mostly crossing the anode facing cell membrane, while PI and EthD-1 predominately entered at the site facing the cathode. In high-salt medium (low electrostatic repulsion), the entry site for Ca2+ and EB was reversed to the cathode-facing hemisphere while it remained unchanged for PI and EthD-1. In all these experiments, the observed transport patterns result from a diffusion-driven process where the flow was controlled by the permeability coefficient of the molecule that was assayed. These observations support the conclusion that the defects induced by the critical TMV are structural different on both sides of the cell membrane facing the electrodes. Biphotonic Microscopy Molecular center asymmetry is required for the creation of second-harmonic generation (SHG) signals. It makes this spectroscopy a powerful technique for visualizing changes in interfacial layers in the plasma membrane of biological cells. Lipophilic SHG probes such as di-4-ANEPPDHQ (Di-4) can detect minute perturbations in the plasma membrane. Rapid changes in membrane symmetry could be detected using SHG (Moen et al. 2014). Following pulsed permeabilizing electric field exposure, an instantaneous drop of ~50 % in SHG signal was detected only from the anodic pole of the cell. This dramatic loss in SHG signal reflects a local dramatic perturbation in the interfacial nature of the membrane but with a vectoriality of the effect. Fluorescence Investigations with FM1-43 Millisecond-long permeabilizing pulses induced membrane disorganization supporting the formation of long-lived permeant structures that were present during membrane resealing. The insertion of FM1-43, a fluorescent dye, into the membrane

Phospholipid Head Group Dipoles and Electropore Formation

13

from the bulk is a fast process controlled by a decrease of lipid packing and/or an electrostatic process as FM1-43 is positively charged. The amphiphilic fluorescent probe FM1-43 was used to detect those post pulse. A fluorescence post-pulse overshoot is associated with the membrane assembly changes (Escoffre et al. 2014). FM1-43 fluorescence gave a local response in the membrane as shown by fluorescence imaging. This increased insertion reflects either an increase in the lipid assembly fluctuations or a change in the membrane surface charge. Moreover, the fluorescence increase facing the anode was twice that of the one facing the cathode showing that the perturbation of the interface was not the same on both sides of the cell. These observations on the induced TMV and the associated transport when electropermeabilization was triggered support the conclusion of asymmetric events on the cell membrane following the direction of the transmembrane field. The dynamical structure of the conducting defects is dependent on the direction of the electrical constraints acting on the interfacial regions.

Structural Events at the Cell Membrane Interphase 31P NMR spectroscopy gives structural informations on the lipid polar head region. Chinese hamster ovary (CHO) cells were reversibly permeabilized by submitting them to short, high-intensity, square wave pulses (1.8 kV/cm, 0.1 ms) (Lopez et al. 1988). Due to these pulsing conditions (high field, short duration), a large cap was electropermeabilized covering the majority of the cell surface, but the population viability was preserved. It was taken advantage for the NMR investigation (which needed averaging of the collected signal) from the observation that cells remained in a permeable state without loss of viability for several hours at 4 C. This gave a good signal-to-noise ratio in the definition of the spectroscopic signatures. A differential method in the NMR pulse sequence was used to analyze the phospholipid head groups and to get rid of the signals coming from other phosphorus groups. A new anisotropic peak with respect to control cells was observed on 31P NMR spectroscopic analysis of the phospholipid components. This peak was only present when the cells were permeable, and normal anisotropy was recovered after resealing. Taking into account the fusogenicity of electropermeabilized cells, comparative studies were performed on 5% poly(ethylene glycol) treated cells. The 31P NMR spectra of the phospholipids displayed the same anisotropic peak as in the case of the electropermeabilized cells. In the two cases, this anisotropic peak was located downfield from the main peak associated with the phospholipids when organized in bilayers. The localization of this anisotropic peak was very different from the one of a hexagonal phase. As a conclusion it was proposed that cell membrane reversible electropermeabilization is associated with a reorganization of the polar head group region (new mean orientation or randomization of the orientations) leading to a weakening of the hydration forces to account for these observations.

14

J. Teissie

Conclusion The complexity in the structural organization of membrane models (lipid bilayers) or of cell membranes is most of the time neglected in the description of electropermeabilization. Biophysical analysis of lipid assemblies pointed out that the interfacial region between the core of the membrane (hydrocarbon chains) and the bulk aqueous phase is a complex assembly that must be called an interphase. This interphase is where a sharp change in the dielectric constant is present. The presence of charged groups with opposite charges built dipoles that affect the local organization of water (dipolar orientation). This interphase is therefore highly sensitive to the external field, which applies a torque on the dipoles. Sophisticated spectroscopic investigations bring converging conclusions: (i) structural transitions affect the interphase region during the pulse delivery on lipid models as well as on cells, (ii) interfacial water is modified, (iii) the effects remain present during resealing of cell membrane, and (iv) the field effect is vectorial as different effects are observed on the two sides of the cell facing the electrodes. Acknowledgments Research was conducted in the scope of the EBAM European Associated Laboratory (LEA) and resulted from the networking efforts of the COST Action TD1104 (http:// www.electroporation.net).

Cross-References ▶ Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Experiments ▶ Electric Pulse Parameters Affecting Electroporation Treatment Outcome ▶ Experimental Electroporation of Planar Lipid Bilayers ▶ Fluorescent Indicators of Membrane Permeabilization Due to Electroporation ▶ Mechanistic Description of Membrane Electropermeabilization ▶ Membrane Interfacial Water Dipoles and Electropore Formation ▶ Membrane Permeabilization Lifetime in Experiments ▶ Water Defects in Phospholipid Bilayers

References Brockman H (1994) Dipole potential of lipid membranes. Chem Phys Lipids 73(1–2):57–79 Diederich A, Bähr G, Winterhalter M (1998) Influence of polylysine on the rupture of negatively charged membranes. Langmuir 14:4597–4605 Escoffre JM, Bellard E, Faurie C, Sébaï SC, Golzio M, Teissié J, Rols MP (2014) Membrane disorder and phospholipid scrambling in electropermeabilized and viable cells. Biochim Biophys Acta 1838(7):1701–1709. doi:10.1016/j.bbamem.2014.02.013

Phospholipid Head Group Dipoles and Electropore Formation

15

Frey W, White JA, Price RO, Blackmore PF, Joshi RP, Nuccitelli R, Beebe SJ, Schoenbach H, Kolb JF (2006) Plasma membrane voltage changes during nanosecond pulsed electric field exposure. Biophys J 90(10):3608–3615. doi:10.1529/biophysj.105.072777 Genco I, Gliozzi A, Relini A, Robello M, Scalas E (1993) Electroporation in symmetric and asymmetric membranes. Biochim Biophys Acta 1149(1):10–18 Hibino M, Itoh H, Kinosita K Jr (1993) Time courses of cell electroporation as revealed by submicrosecond imaging of transmembrane potential. Biophys J 64(6):1789–1800 Kinosita K Jr, Tsong TY (1977) Voltage-induced pore formation and hemolysis of human erythrocytes. Biochim Biophys Acta 471:227–242 Le Saux A, Ruysschaert JM, Goormaghtigh E (2001) Membrane molecule reorientation in an electric field recorded by attenuated total reflection Fourier-transform infrared spectroscopy. Biophys J 80(1):324–330 Leikin S, Parsegian VA, Rau DC, Rand RP (1993) Hydration forces. Annu Rev Phys Chem 44:369–395 Lopez A, Rols MP, Teissie J (1988) 31P NMR analysis of membrane phospholipid organization in viable, reversibly electropermeabilized Chinese hamster ovary cells. Biochemistry 27 (4):1222–1228 Miller IR (2002) Effect of electric fields on the structure of phosphatidylcholine in a multibilayer system. Bioelectrochemistry 57(2):145–148 Moen EK, Ibey BL, Beier HT (2014) Detecting subtle plasma membrane perturbation in living cells using second harmonic generation imaging. Biophys J 106(10):L37–L40. doi:10.1016/j. bpj.2014.04.008 Morone N, Fujiwara T, Murase K, Kasai RS, Ike H, Yuasa S, Usukura J, Kusumi A (2006) Threedimensional reconstruction of the membrane skeleton at the plasma membrane interface by electron tomography. J Cell Biol 174:851–862. doi:10.1083/jcb.200606007 Nagle JF, Tristram-Nagle S (2000) Lipid bilayer structure. Curr Opin Struct Biol 10(4):474–480 Neumann E, Kakorin S, Toensing K (1998) Membrane electroporation and electromechanical deformation of vesicles and cells. Faraday Discuss 111:111–125. discussion 137–57 Neumann E, Rosenheck K (1972) Permeability changes induced by electric impulses in vesicular membranes. J Membr Biol 10(3):279–290 Nymeyer H, Zhou HX (2008) A method to determine dielectric constants in nonhomogeneous systems: application to biological membranes. Biophys J 94:1185–1193 Rems L, Miklavcic D (2016) Tutorial: electroporation of cells in complex materials and tissue. J Appl Phys 119:201101. doi:10.1063/1.4949264 Seelig J, Macdonald PM, Scherer PG (1987) Phospholipid head groups as sensors of electric charge in membranes. Biochemistry 26(24):7535–7541 Sowers AE (1986) A long-lived fusogenic state is induced in erythrocyte ghosts by electric pulses. J Cell Biol 102(4):1358–1362 Stulen G (1981) Electric field effects on lipid membrane structure. Biochim Biophys Acta 640 (3):621–627 Teissie J, Golzio M, Rols MP (2005) Mechanisms of cell membrane electropermeabilization: a minireview of our present (lack of ?) knowledge. Biochim Biophys Acta 1724(3):270–280 Tekle E, Astumian RD, Chock PB (1994) Selective and asymmetric molecular transport across electroporated cell membranes. Proc Natl Acad Sci U S A 91(24):11512–11516 Tekle E, Astumian RD, Friauf WA, Chock PB (2001) Asymmetric pore distribution and loss of membrane lipid in electroporated DOPC vesicles. Biophys J 81(2):960–968 Tocanne JF, Teissié J (1990) Ionization of phospholipids and phospholipid-supported interfacial lateral diffusion of protons in membrane model systems. Biochim Biophys Acta 1031 (1):111–142 Vernier PT, Ziegler MJ (2007) Nanosecond field alignment of head group and water dipoles in electroporating phospholipid bilayers. J Phys Chem B 111(45):12993–12996 Zawisza I, Lachenwitzer A, Zamlynny V, Horswell SL, Goddard JD, Lipkowski J (2003) Electrochemical and photon polarization modulation infrared reflection absorption spectroscopy study

16

J. Teissie

of the electric field driven transformations of a phospholipid bilayer supported at a gold electrode surface. Biophys J 85(6):4055–4075 Zimmermann U, Pilwat G, Riemann F (1974) Dielectric breakdown of cell membranes. Biophys J 14:881–899

Membrane Permeabilization Lifetime in Experiments Justin Teissie

Abstract

Short high field electric pulses when delivered to a cell suspension or a tissue results in a membrane permeabilization (electropermeabilization). Classical electropermeabilization (pulse longer than 1 μs) results in an enhanced transport across the membrane and in a change in its electrical conductivity. Electropermeabilization can be reversible if the pulsing parameters are selected in a proper way. This reversibility is due to set of reactions called resealing. This means that an enhanced membrane permeabilization is transiently induced while preserving the cell viability. More drastic pulsing conditions can make this permeability irreversible (cell death would result). Molecular mechanisms supporting electropermeabilization of cell membrane are complex. They are not limited to the lipid matrix (punching holes in the lipid bilayer as figured out in “electroporation”) but involved the cellular machinery. Keywords

Membrane • Transmembrane voltage • Transport • Membrane structure • Repair

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resealing Observed from Electrical Conductance Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Online Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Patch Clamp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Muscle Fibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electric Cell-Substrate Impedance Sensing (ECIS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resealing as Observed by the Time-Dependent Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resealing Is Observed by a Post Addition of the Reporter Dye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 3 4 4 5 5 6

J. Teissie (*) Institut de Pharmacologie et de Biologie Structurale, Université de Toulouse, CNRS, UPS, Toulouse, France e-mail: [email protected]; [email protected] # Springer International Publishing AG 2017 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_79-1

1

2

J. Teissie

Dependence of the Resealing on the Pulsing Electrical Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resealing Is Strongly Controlled by the Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Buffer Osmotic Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Buffer Ionic Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Membrane Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resealing Is an Active Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cytoskeleton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Exocytosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7 8 8 9 10 10 11 12 13 13 13

Introduction Electropermeabilization of the cell membrane is a phenomenon caused by the direct exposure of the cell to calibrated electric pulses. Permeabilization depends on pulse duration, pulse amplitude, and number of pulses delivered and also on other experimental conditions. With these parameters properly chosen, the process of permeabilization is reversible, and cells return to their normal physiological state. The process of permeabilization is divided into a short permeabilizing phase that is initiated during the pulse, and a longer complex resealing phase that begins after the end of the pulse (Puc et al. 2003). Membrane permeabilization is associated with an enhanced transport of soluble small molecules (that are with a low molecular weight, say less than 2 kD) and ions. It is a time-dependent process, and its kinetics can be observed by monitoring the electrical conductivity change in the membrane or the change in the transport parameters. The transport after the pulse is diffusion driven. This means that its amplitude is dependent on the molecule that is detected. No postpulse transport is detected for macromolecules (siRNA, pDNA) (Paganin-Gioanni et al. 2011). It can be assayed by the inflow of dyes (trypan blue, propidium iodide) by observation under the microscope or with a flow cytometer. Permeabilization can then be quantified by two parameters: the number of cells in the population where the uptake can be detected and the amount of uploaded dyes (i.e., the distribution of fluorescence intensity in each single cell in the population). As electrically induced permeabilization is a change in the state of the membrane, it can also be detected by the leakage of endogenous metabolites (such ATP by a luciferin luciferase chemiluminescent assay) or of a prepulse loaded dye (using the same methodologies as for the uptake). Resealing is then measured by the post-pulse time dependence of one of these assays. Molecular processes supporting classical electropermeabilization remain poorly understood (Teissie et al. 2005) in spite of more than 40 years of effort (Neumann and Rosenheck 1972). The membrane is called “porous.” There is general agreement that it cannot be adequately described simply by the occurrence of reversible hydrophilic holes, so-called pores (“electroporation”), in the lipid bilayer. A membrane in a cell is

Membrane Permeabilization Lifetime in Experiments

3

a more complex system. It is an out of equilibrium complex from a thermodynamic point of view. It forms a complex with the extracellular matrix and the cytoskeleton. Its transmembrane potential is dependent on the cellular energy reserves. Therefore the resealing of the structural defects supporting the transmembrane transport can be reported only in an empiric way: i.e., an accurate description of its control by the different physical and biological parameters but a speculative approach of the molecular mechanisms. A lot of studies have been developed on pure lipid assemblies (planar lipid membrane, liposomes). Processes on such physical models are relevant of soft matter physics and would not be considered in the present chapter. Only events occurring in membranes within cells will be approached. Resealing is assayed by different biophysical approaches. As mentioned above they are shortly described in a first part of this chapter. They are used to analyze the processes and their control by biophysical and physiological parameters.

Resealing Observed from Electrical Conductance Changes Online Assay By measuring the voltage and the current flowing between the electrodes, one gets access to the conductance of the suspension. Under steady-state condition, the conductance of a diluted cell suspension can be considered controlled by the conductance of the external solution. An ohmic behavior is observed when a voltage pulse is delivered on the cell suspension (as long as the Joule heating is small). When electropermeabilization appears, two events are present: i. The dielectric shell has conductive defects that increase during the pulse. ii. There is leak of the internal (cytoplasmic) solution. Both events result in an increase in the cell suspension conductivity, when it is initially low. This is a very fast process that can be detected on the submicrosecond time scale after correcting from the interfacial electrochemical reactions on the electrodes. The kinetic is biphasic: a very sharp increase during the first microsecond followed by a slower continuous increase. After the pulse, the conductivity of the cell suspension is due to the contribution of the leaked ions and the defects under resealing. If a second permeabilizing pulse is applied, the conducting defects will again increase. If the delay between the two pulses is short (a few microseconds), the conductivity of the suspension at the onset of the second pulse is the same as at the end of the first one, and no resealing of the defects will be detected. With longer delay (10 ms), almost complete resealing occurs, and the observed initial increase in conductivity observed at the onset of the second pulse is due to the contribution of the leaked ions. The resealing of the

4

J. Teissie

dielectric defects is biphasic, a large but partial resealing is detected within a few milliseconds, but the remaining contribution is long lived. The ion leakage from the cytoplasm is long lived and appears as the major contribution in the long-lived cell suspension conductance change, when the ionic content of the pulsing buffer is low (Kinosita and Tsong 1979; Hibino et al. 1991). Increased permeability of a cell membrane is accompanied by increased membrane conductivity; by measuring electric conductivity of a cell suspension, the mean extent of permeabilized tissue could be monitored in real time. Total conductivity changes and impedance measurements showed easy to detect significant changes in conductivity due to this increased membrane conductivity, the ion efflux in low-conductive medium and colloid-osmotic swelling in low- and high-conductive medium. By measuring electric conductivity online during the pulses permeabilization, threshold can be detected, but permeabilization levels are not obtained as the signature of the electric changes is too complex (Pavlin et al. 2005).

Patch Clamp The patch-clamp technique allows the investigation of currents through a small set or even single ion (conducting) channels and therefore about the membrane conductance. The whole-cell patch-clamp technique was employed to investigate the effect of millisecond pulsed electric fields on mammalian cells (Wegner et al. 2015). Cellular membrane permeabilization was monitored by a conductance increase. It could be established experimentally that electropermeabilization consists of two clearly separate processes: a rapid membrane conductance change (due to the creation of conducting defects in the cell membrane or transient “electroporation”) that occurs while the field is present to depolarize or hyperpolarize the membrane to voltages beyond so-called threshold potentials (evaluated to be +201 mV and
231 mV, respectively in the study) and is reversible within 100 ms after the pulse, and a long-term, or persistent, permeabilization covering the whole voltage range. The latter remained after the pulse for at least 40 min at room temperature.

Muscle Fibers A transmembrane voltage can be delivered directly on isolated muscle fibers that are giant cells, and the associated current is measured (Bier et al. 2002). Transmembrane potentials of up to 120 mV are physiologic and well tolerated, but when the potential is more than 300 mV, large currents are observed due to the electropermeabilization. Currents through frog muscle cell membranes are observed under 4 ms pulses of up to 440 mV. The relaxation of transmembrane currents under this voltage clamp approaches back to zero after the pulse is observed by using low transmembrane potential test pulses and described in terms of membrane physical parameters.

Membrane Permeabilization Lifetime in Experiments

5

Electric Cell-Substrate Impedance Sensing (ECIS) An approach referred to as (ECIS) studied the behavior of adherent animal cells grown to confluence on circular gold-film electrodes of 250 μm in diameter within a culture dish (Stolwijk et al. 2011). A large coplanar counter electrode was used. The impedance of the cell-covered electrode was measured by applying a low AC voltage with designated frequencies to monitor the time-dependent normalized impedance behavior of the cells. The gold-film electrodes were used to deliver well-defined electropermeabilizing voltage pulses (several volts amplitude and several hundred milliseconds duration) across the adherent cells in order to achieve reversible membrane electropermeabilization. Time-resolved impedance measurements before and immediately after the pulse pictured the kinetics of membrane resealing as well as associated changes in cell morphology. Membrane permeabilization and resealing were only clearly visible on rather short time scales. A fast and sharp drop of the normalized impedance to less than 0.55 was followed by its quick recovery within a few seconds to 0.9. Full recovery to pre-pulse relative impedance values (1 by definition) resulted from a second process with a much slower time constant assigned in part to the induced fluctuations in cell shape. Full cell membrane (electrical and morphological) recovery was slow and obtained within less than 90 min.

Resealing as Observed by the Time-Dependent Transport Transport is present across electropermeabilized membrane. After the pulse, when no external field is present, the transport is diffusion driven and is controlled by the gradient of concentration of the reporter molecules between the two sides of the plasma membrane. It depends on the permeability of the electropermeabilized membrane to the reporter molecule (in many cases a water soluble fluorescent dye such as propidium iodide). At the single cell level or on a population, this can be monitored by the post-pulse inflow, i.e., the fluorescence time-dependent change in the cell cytoplasm. A competition is present between the diffusion process and the resealing that decreases the permeabilization. Deconvolution of the kinetics of the fluorescence change gives the parameters of the resealing. A good fit is obtained under the assumption that the resealing processes are all first order. A very fast approach under a microscope showed that three steps were present (Pucihar et al. 2008). The pulsing chamber was set on a microscope stage, and the fluorescence signal was observed by a fast photomultiplier tube. Its output voltage U was proportional to the fluorescence signal emitted by the cell. The fluorescence increase was due to the inflow of PI. When a permeabilizing field pulse was delivered with different pulse durations, a long-lived propidium iodide (PI) uptake (fluorescence increase) was observed with amplitudes that depended on the pulse duration.

6

J. Teissie

The complex kinetic of the fluorescence increase (I(t)) was deconvoluted as a three-step process: I ðtÞ ¼ I1ð1
expð
t=T1ÞÞ þ I2ð1
expð
t=T2ÞÞ þ I3ð1
expð
t=T3ÞÞ

(1)

The faster ones were all with time constant (T1, T2) shorter than 100 ms (for a square waved pulse of 1 ms), and their detection was achievable only due to the advanced technology of the study. With classical equipment, only the slow step (T3) could be detected and was on the scale of several seconds or minutes (Teissie et al. 2005; Prausnitz et al. 1995). Its amplitude I3 was large meaning that uploading occurred mostly during this slow resealing step by a diffusion-driven process. This is a key information for practical applications, a slow resealing will increase the amount of molecules that are uploaded (or released): QðtÞ ¼ Pperm Aperm ΔSðtÞð1
expð
t=T3ÞÞ

(2)

where Q(t) is the flow of S across the membrane at time t after the pulse, Pperm is the permeability coefficient for the reporter molecule, Aperm is the part of the cell surface that is permeabilized, and ΔS(t) is the concentration difference of S between the bulk and the cytoplasm. A key practical information is that the inflow and the resulting accumulation (that controls the sensitivity of detection of permeabilization) are controlled by Pperm that is specific of the reporter molecule.

Resealing Is Observed by a Post Addition of the Reporter Dye The change in the transport properties of the electropermeabilized membrane can be easily approached by adding the reporter dye at different delays after the pulse delivery (inducing the membrane permeabilization) and counting the number of cells where the dye uptake can be detected (i.e., that are still permeabilized) or the amount of dye uptake: QdelðtÞ ¼ PpermAperm ΔSðtÞð1
expð
t
Tadd=T3ÞÞ

(3)

Qdel(t), flow of S across the membrane at time t after the pulse. Tadd, delay for the addition after the pulse delivery. Qdel(t) decreases with an increase in Tadd. The first approach (relative number of cells where the transport is detected) is the most popular, and permeabilization is assayed as the percentage of positive cells (cells where uptake is detected). The second approach gives more informations. It can be approached either by digitized fluorescence microscopy by single cell analysis, but of course the number of cells that are analyzed remains limited, or by flow cytometry. In this latter case, by a proper setting of the intensity windows, three populations were detected: cells with no permeabilization, where no fluorescence

Membrane Permeabilization Lifetime in Experiments

7

could be detected (population 1); cells that remained permeabilized with a significant uptake (population 2), i.e., fluorescence signal; and cells that were irreversibly permeabilized with a high fluorescence signal (population 3). Resealing was observed by a shift over time of cells from population 2 to population 1. The mean intensity in population 2 was indicative of the uptake along the delayed dye addition and was observed to decrease upon time. The increase in population 3 was reflecting irreversible permeabilization and the associated cell death (Teissie et al. 2005). Of course the assay is strongly dependent on the sensibility of the detection assay. The assay is slow (time needed to add the dye to the pulsed cell suspension) and cannot detect the fast steps in resealing. Only the slow step (on the scale of several seconds) can be observed. This approach confirms that the “slow” step in resealing is a first-order process. The number of permeabilized cells in the population is then PðE, tÞ ¼ LðEÞ þ ½ðPðE, 0Þ
LðEÞÞexpð
KtÞ

(4)

P(E,t) is the percentage of cells detected as fluorescent (populations 2 and 3 in the FACS plot) with dye addition at time t after a pulse with an E amplitude that triggers the lysis of a subpopulation L(E)(population 3); P(E,0) is the percentage of cells detected as permeabilized just after the pulse (indeed where the membrane is leaky for physiological reasons). K is the resealing constant (the reciprocal of T3).

Dependence of the Resealing on the Pulsing Electrical Parameters Electropermeabilization is known to be under the control of the field strength E, on the single pulse duration T, on the number of successive pulses N, and on the delay between the pulses. Resealing lifetime (the reciprocal of K) is observed to be controlled by these parameters in a specific way. K is not dependent on the field strength (Teissie et al. 2005). The field strength controls the number of cells that are permeabilized and the occurrence of the cell lysis. When the reversible permeabilization is observed (correcting the observed permeabilization P from the percentage of lysed cells L), then the kinetic time constant is similar to the one observed under mild field conditions (where electropermeabilization is reversible, no cell lysis). The resealing lifetime is observed to be linearly dependent on T as long as T is less than 1 ms. The process is more complex when longer pulses (larger than 5 ms) are delivered on the population. This supports the hypothesis that the biophysical mechanisms affecting the cell long-lived electropermeabilization may be different when long pulses (lasting several milliseconds each) are applied as compared with the events triggered by much shorter pulses (in the microsecond range). When T is less than 1 ms, the resealing lifetime is a linear function of the number of successive pulses (delivered with a delay shorter than 1 s) as long as a limited number is delivered (less than 10).

8

J. Teissie

As long as the delay between successive pulses is less than 1 s, there is no obvious dependence of the resealing lifetime on the delay. The use of longer delays (several seconds) is still the field of controversy on its consequence on electropermeabilization (memory effects?) (Demiryurek et al. 2015). This may result from an excessive leakage of cytoplasmic molecules such as ATP and GTP or an exchange of ions affecting the resting potential difference by altering the Nernst potential. One main consequence of the dependence of the resealing lifetime on the pulse duration and on the number of pulses is the associated control on the transport induced by electropermeabilization. Transport after the pulse is diffusion driven and follows Fick’s law. Resealing means a decrease in the density of defects supporting the transmembrane transfer. This density (within the cell cap that is affected by electropermeabilization) is known to be controlled by the pulse duration and the number of successive pulses (when the delay between the pulses remains short), during the so-called expansion step. Increasing the pulse duration brings an increase in the density of defects and an associated increase in transport. This increase is further enhanced along the resealing, as defects remain present during a longer time as it was just described. Long pulses (up to 1 ms) give a higher cumulative transport by acting on the density of defects and on their lifetime. A similar role is played by the number of pulses (as long as they do not affect the viability).

Resealing Is Strongly Controlled by the Temperature The post-pulse temperature is strongly affecting the resealing process (Teissie et al. 2005). Most experiments are run at room temperature (20–25  C). If the pulsed sample is brought to its “physiological” temperature (37  C), a faster resealing is observed, and the membrane enhanced permeability vanished in a short time. But if the sample is kept on ice after the pulse train delivery, the membrane remains permeabilized over long periods (up to several hours). Cells remain viable as membrane resealing can be obtained by bringing the sample temperature back to 37  C. Adding the reporter dye after the “heat” jump does not bring any upload; the membrane has lost its enhanced permeability. Cells are able to grow when brought in the culture medium. This means that at low temperature where the resealing is not working, cell membrane is permeabilized, but the cell viability remains preserved. The main conclusion is that the resealing can be easily controlled by playing with the post-pulse temperature of the sample.

Buffer Osmotic Pressure The effects of pulsing buffer osmotic pressure were observed on short electric field pulses (0.1 ms) induced permeabilization on Chinese hamster ovary cells (CHO) growing either in monolayers or in suspension (Teissie et al. 2005). No osmotic swelling was observed on mammalian nucleated cells under these experimental conditions (this is not valid for red blood cells where electropermeabilization under

Membrane Permeabilization Lifetime in Experiments

9

hypoosmotic conditions induced osmotic swelling and dramatic damages in the membrane, so called craters). Osmotic pressure (increases hyperosmolarity) (decreases (hypoosmolarity)) was shown to decrease (to increase) the efficiency of the resealing phase. The observation was that the resealing for cells pulsed in isotonic buffer is faster than those obtained in a hypertonic one. These results were tentatively explained in terms of the effect of the osmotic pressure on the membrane organization (interfacial forces and undulation forces). A more practical conclusion is that a larger transport across the membrane is present under hyperosmolar conditions. It must be emphasized the consequence of the observations that CHO size is not affected by osmotic pressure. The applied pressure on the membrane keeps the same magnitude in the hypoosmotic and in the hyperosmotic buffers. It is only the direction of application of the pressure which is inverted. As a consequence, it is difficult to explain the observations on cells by an effect on the stretching properties of the membrane as it is done by others with pure lipid vesicles (Needham and Hochmuth 1989). Osmotic effects are different when long pulses are delivered. As PI inflow occurred mostly after the pulse (Teissie et al. 2005; Prausnitz et al. 1995; Pucihar et al. 2008), the effect of the osmotic pressure during the 10-min incubation after 5 ms pulses (long pulse) was investigated by changing the buffer immediately after pulsing (Teissie et al. 2005). A large transport was observed. Under this experimental condition, cell swelling along the pulse delivery was observed. The percentage of permeabilized cells and the associated fluorescence intensity were not affected by this treatment. The effect of osmolarity only existed during the application of the pulses and not during post-pulse incubation (resealing step), when the inflow of molecules of PI mostly took place. Resealing experiments were performed. Resealing curves were observed to fit a first-order kinetic well. The rate constant, K, was found to be constant whatever the osmolarity. This is a strong support for the conclusion that the electropermeabilization process is different for short pulse at a high field and long pulse at a lower field.

Buffer Ionic Strength The effects of the ionic strength of the pulsing buffer on the electropermeabilization of Chinese hamster ovary cells (CHO) were investigated (Teissie et al. 2005). Cells were pulsed and kept at room temperature before addition of trypan blue (0.4% mass/ vol.) at various times after pulsation. The permeated state of electropulsed CHO cell membranes (the permeabilized population) could be maintained for several minutes at 21  C under electric field strengths larger than the critical threshold. The cells then progressively recovered their original impermeable plasma membranes (decrease in the permeabilized population P). The initial rate of the resealing process, i.e., the ratio dP/dt (P = permeabilization, t = time), i.e., the kinetics of defects annihilation (controlling the dye uptake) was followed at room temperature for CHO cells electropulsed under the same electrical conditions (with different field strengths but same pulse duration) but in isosmotic pulsing buffers of different ionic strengths.

10

J. Teissie

Similar curves were obtained for the different buffers: cells were maintained permeable for 5 min, and resealing was completed during the following 60 min. The effect of buffer ionic concentration during the 5-min incubation period following the pulses was also investigated by changing the buffer immediately after application of pulses. The ion concentration during the post-pulse incubation step did not affect the detected permeabilization. Permeabilization was only dependent on the ionic content during pulsation.

Membrane Order The energy barrier for the structural transition in the membrane bringing conducting defects was checked for its entropic character controlled by the membrane order. A systematic attempt tried to relate the intrinsic plasma membrane fluidity of three different cell lines to their reversible electroporation (Kanduser et al. 2006). Membrane resealing was related to cell membrane fluidity as determined by electron paramagnetic resonance spectroscopy and computer characterization of membrane domains. Nevertheless, this effect, if present, was masked with different time courses of membrane resealing found for the different cell lines studied. A more chemical approach was obtained by adding molecules known to affect the membrane order. Cell electropermeabilization is detected only under more stringent conditions when cells have been treated by ethanol. On the other hand, lysolecithin was observed to facilitate cell electropermeabilization and to negatively affect the resealing (Teissie et al. 2005). 0.4 M ethanol incubation did not inhibit the reversibility of electropermeabilization. Nevertheless, these experimental results suggested that, apparently, the loss of the defects (decrease in P) was faster when cells were pretreated with ethanol. When the membrane order is increased, the temporarily permeable structures (defects) are apparently stabilized, and the lifetime of the permeable state is longer when the cells are treated with lysolecithin before pulsing. Electropermeabilization is inducing a transition in the membrane organization. Membrane order is modulating the energy barrier needed to evoke this membrane transition which occurs when cells are submitted to a field larger than a characteristic threshold (expansion step). Less order would decrease the magnitude of this energy barrier and facilitate the resealing (the decrease in P is faster as observed on ethanol treated cells); more order would increase it (P remains high on a longer period after lysolecithin addition than with untreated pulsed cells).

Resealing Is an Active Process Electropermeabilization is affecting not only the plasma membrane, but other organelles are also targets of the process. This can be a direct effect but only when very strong fields are delivered. The field inside the cytoplasm is weaker than the external one as the membrane, even when electropermeabilized, induces an electrical shield

Membrane Permeabilization Lifetime in Experiments

11

(Esser et al. 2010; Weaver et al. 2012). But due to the enhanced transport, the cytosolic composition is altered. This shown by the leak of secondary metabolites and by the dramatic changes in Calcium content (increase). This results in an indirect effect on organelles. The second metabolite ATP is playing a key role in the resealing process. Its cytoplasmic level should be high. Keeping CHO cells in PBS for 30 min before pulsing prevented the resealing. But this was observed to starve the cells; the ATP level dropped by 75% before the pulse delivery. Same effect on resealing was obtained after a deoxyglucose treatment. Again the ATP level was only 25% of the basal one (or less) and was further decreased due to the ATP leak induced by electropermeabilization. This could not be repaired by adding external ATP to balance the electroinduced leak.

Cytoskeleton The membrane-bound protein skeleton, a network of actin filaments and associated proteins, covers the entire plasma membrane cytoplasmic surface. It is closely linked to clathrin-coated pits and caveolae, key factors in membrane traffic (Morone et al. 2006). The actin filaments are likely to form the limit of the membrane compartments where membrane molecules are confined. Microtubules are filamentous intracellular structures. Microtubules are involved in the organization of intracellular structure and intracellular transport. GTP has an active role in the polymerization of microtubules (Carlier et al. 1987) which is prevented by colchicine (Howard and Hyman 2003). ATP plays a similar role on the polymerization of actin microfilaments (Chhabra and Higgs 2007). Addition of I mM GTP or ATP in the pulsing buffer, to maintain their concentration in the cytoplasm at a constant level, did not result in any change in the permeabilization, but resealing was strongly affected when GTP was added (Teissie et al. 2005). But, the resealing rate was greatly affected when cells were treated with colchicine that prevented the tubulin polymerization: half of the cell population became unpermeable to trypan blue in less than 2 min, with 6 min being the required time in the case of control cells (pulsed under the same electrical conditions). In the case of cytochalasin treatment that blocks polymerization and the elongation of actin filaments, no difference was observed at the Iower concentration. But at 20 μM and above, most cells were still dye permeable after 1 h incubation at room temperature, but this could be associated with the drug lethal effect (Teissie et al. 2005). Permeabilization appears as a structural transition in the membrane organization with an energy cost for rupturing chemical interaction between the membrane partners. This is strongly supported by the strong temperature dependence of the resealing. This energy barrier could be greatly increased if interconnections between membrane components and cytoskeleton proteins existed, mainly for microtubules. Such a change in the activation energy would affect the kinetics of resealing by bringing a decrease of the rate constant of the repair process. This induces an increase

12

J. Teissie

in the lifetime of permeabilization. This increase in stabilization of permeated structures could be explained by the low lateral mobility of membrane proteins which are linked to cytoskeleton but can be affected by the electrophoretic stress along the membrane associated with the pulse delivery. All these experimental observations strongly suggest that the cytoskeleton is involved in the electropermeabilization phenomenon. Firstly, it appears to be altered as a consequence of permeabilization (maybe not directly by the field that is weak in the cytoplasm (Esser et al. 2010)) at a morphological level and secondly, it seems to play a regulatory role in the stabilizating process of the defects supporting electropermeabilization. By scanning electron microscopy, numerous microvilli and blebs were observed almost immediately after application of short field pulses in a low conductivity buffer (Escande-Géraud et al. 1988). No other membrane changes were observed. The appearance of osmotic pressure-dependent “blebs” was indicative of local weakening of the plasma membrane. The application of 100-μs rectangular pulses of 1.3 kV/cm electric field to different types of cells (FBT, MEF, RAT-1, L-cells) in the physiological medium (where a high conductivity was present) led to the formation and growth of spherical and hemispherical protuberances of the cell membrane (Gass and Chernomordik 1990). Most of these effects were fully reversible and disappeared within 30 min at 37  C. The major effect which was induced during or in the very first seconds following the pulse was the eruption of numerous microvilli (EscandeGéraud et al. 1988). This increase in the density of villi was transient, and the same density as before pulsation was observed after the 20 min post-field incubation at 37  C. The means of the size of the cell surface and of the number of filopodia are not affected by the pulsation in a statistical significant way. The density of microvilli increased by 40%. When cells were treated in low conductivity pulsing buffers where nucleotides were added (cAMP, GTP, ATP all at 1 mM), a dramatic effect was detected only when ATP was added. Both the length and the density of microvilli increased significantly, but this effect is transient and disappeared when the cells completely resealed (Rols and Teissié 1992).

Exocytosis Exocytosis is described as a key process for repairing wounded membranes (McNeil and Steinhardt 2003). One should consider cell membrane electropermeabilization as a membrane wounding. Therefore resealing should be supported by exocytosis. Along resealing of membrane electropermeabilization, lysosomal glycoproteins were detected by immunofluorescence at the level of the plasma membrane. Their density was related to the density of electrically triggered defects (by changing the pulse duration at a given field strength) (Huynh et al. 2004). One consequence was therefore Calcium that played a role in exocytosis should play a regulatory role in resealing. It should be mentioned that a calcium transient increase is always present along electropermeabilization due to the low resting cytoplasmic free concentration of Calcium (Rems and Miklavcic 2016).

Membrane Permeabilization Lifetime in Experiments

13

Conclusions The main conclusion of all these experimental investigation of the lifetime of electropermeabilization (resealing) is that it is indeed a multistep process. Structures present as long as the field is delivered vanished in a few milliseconds as soon as the field is switched off. New structures remain present supporting a diffusion-driven transport of small non-permeant species. These long-lived defects are controlled by the single pulse duration. They support the major part of the uploading (or leakage from) of cells. Their lifetime is dependent on complex mechanisms involving the cell metabolism with a strong implication of exocytosis and the organization of the cytoskeleton. An important role is played by the interfacial organization of the membrane. Acknowledgments Research was conducted in the scope of the EBAM European Associated Laboratory (LEA) and resulted from the networking efforts of the COST Action TD1104 (http:// www.electroporation.net).

Cross-References ▶ Cell Stress Responses to Pulsed Electric Fields ▶ Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Experiments ▶ Different Cell Viability Assays Following Electroporation In Vitro ▶ Electric Pulse Parameters Affecting Electroporation Treatment Outcome ▶ Fluorescent Indicators of Membrane Permeabilization Due to Electroporation ▶ Gene Delivery by Electroporation In Vitro: Mechanisms ▶ Measurement of Molecular Transport into Electropermeabilized Cells ▶ Mechanistic Description of Membrane Electropermeabilization ▶ Nucleic Acid Electrotransfer in Mammalian Cells: Mechanistic Description ▶ Parameters Affecting Cell Viability Following Electroporation In Vitro ▶ Pulsed Electric Fields Treatment of Biological Suspensions ▶ Single-Cell Electrical Characterization Techniques

References Bier M, Chen W, Gowrishankar TR, Astumian RD, Lee RC (2002) Resealing dynamics of a cell membrane after electroporation. Phys Rev E Stat Nonlinear Soft Matter Phys 66(6 Pt 1):062905 Carlier HF, Didry D, Pantaloni D (1987) Microtubule elongation and guanosine 50 -triphosphate hydrolysis. Role of guanine nucleotides in microtubule dynamics. Biochemistry 26:4428–4437 Chhabra ES, Higgs HN (2007) The many faces of actin: matching assembly factors with cellular structures. Nat Cell Biol 95:1110–1121 Demiryurek Y, Nickaeen M, Zheng M, Yu M, Zahn JD, Shreiber DI, Lin H, Shan JW (2015) Transport, resealing, and re-poration dynamics of two-pulse electroporation-mediated molecular delivery. Biochim Biophys Acta 1848(8):1706–1714. doi:10.1016/j.bbamem.2015.04.007

14

J. Teissie

Escande-Géraud ML, Rols MP, Dupont MA, Gas N, Teissié J (1988) Reversible plasma membrane ultrastructural changes correlated with electropermeabilization in Chinese hamster ovary cells. Biochim Biophys Acta Biomembr 939:247–259 Esser AT, Smith KC, Gowrishankar TR, Vasilkoski Z, Weaver JC (2010) Mechanisms for the intracellular manipulation of organelles by conventional electroporation. Biophys J 98 (11):2506–2514. doi:10.1016/j.bpj.2010.02.035 Gass GV, Chernomordik LV (1990) Reversible large-scale deformations in the membranes of electrically-treated cells: electroinduced bleb formation. Biochim Biophys Acta Biomembr 1023:1–11 Hibino M, Sighemori H, Itoh H, Nagayama K, Kinosita K (1991) Membrane conductance of an electroporated cell analyzed by submicrosecond imaging of transmembrane potential. Biophys J 53:209–220 Huynh C, Roth D, Ward DM, Kaplan J, Andrews NW (2004) Defective lysosomal exocytosis and plasma membrane repair in Chediak-Higashi/beige cells. Proc Natl Acad Sci U S A 101 (48):16795–16800 Howard J, Hyman AA (2003) Dynamics and mechanics of the microtubule plus end. Nature 422:753–758. doi:10.1038/nature01600] Kanduser M, Sentjurc M, Miklavcic D (2006) Cell membrane fluidity related to electroporation and resealing. Eur Biophys J 35(3):196–204 Kinosita K Jr, Tsong TY (1979) Voltage-induced conductance in human erythrocyte membranes. Biochim Biophys Acta 554:479–497 McNeil PL, Steinhardt RA (2003) Plasma membrane disruption: repair, prevention, adaptation. Annu Rev Cell Dev Biol 19:697–731 Morone N, Fujiwara T, Murase K, Kasai RS, Ike H, Yuasa S, Usukura J, Kusumi A (2006) Threedimensional reconstruction of the membrane skeleton at the plasma membrane interface by electron tomography. J Cell Biol 174:851–862. doi:10.1083/jcb.200606007 Needham D, Hochmuth RM (1989) Electro-mechanical permeabilization of lipid vesicles. Role of membrane tension and compressibility. Biophys J 55(5):1001–1009 Neumann E, Rosenheck K (1972) Permeability changes induced by electric impulses in vesicular membranes. J Membr Biol 10(3):279–290 Paganin-Gioanni A, Bellard E, Escoffre JM, Rols MP, Teissié J, Golzio M (2011) Direct visualization at the single-cell level of siRNA electrotransfer into cancer cells. Proc Natl Acad Sci U S A 108(26):10443–10447. doi:10.1073/pnas.1103519108] Pavlin M, Kanduser M, Rebersek M, Pucihar G, Hart FX, Magjarevic R, Miklavcic D (2005) Effect of cell electroporation on the conductivity of a cell suspension. Biophys J 88(6):4378–4390 Prausnitz MR, Corbet JD, Gimm JA, Golan DE, Langer R, Weaver JC (1995) Millisecond measurement of transport during and after an electroporation pulse. Biophys J 98:1864–1870 Puc M, Kotnik T, Mir LM, Miklavcic D (2003) Quantitative model of small molecules uptake after in vitro cell electropermeabilization. Bioelectrochemistry 60(1–2):1–10 Pucihar G, Kotnik T, Miklavcic D, Teissié J (2008) Kinetics of transmembrane transport of small molecules into electropermeabilized cells. Biophys J 95(6):2837–2848. doi:10.1529/ biophysj.108.135541] Rems L, Miklavcic D (2016) Tutorial: electroporation of cells in complex materials and tissue. J Appl Phys 119:201101. doi:10.1063/1.4949264 Rols MP, Teissié J (1992) Experimental evidence for the involvement of the cytoskeleton in mammalian cell electropermeabilization. Biochim Biophys Acta 1111(1):45–50 Stolwijk JA, Hartmann C, Balani P, Albermann S, Keese CR, Giaever I, Wegener J (2011) Impedance analysis of adherent cells after in situ electroporation: non-invasive monitoring during intracellular manipulations. Biosens Bioelectron 26(12):4720–4727. doi:10.1016/j. bios.2011.05.033 Teissie J, Golzio M, Rols MP (2005) Mechanisms of cell membrane electropermeabilization: a minireview of our present (lack of ?) knowledge. Biochim Biophys Acta 1724(3):270–280

Membrane Permeabilization Lifetime in Experiments

15

Weaver JC, Smith KC, Esser AT, Son RS, Gowrishankar TR (2012) A brief overview of electroporation pulse strength-duration space: a region where additional intracellular effects are expected. Bioelectrochemistry 87:236–243. doi:10.1016/j.bioelechem.2012.02.007] Wegner LH, Frey W, Silve A (2015) Electroporation of DC-3F cells is a dual process. Biophys J 108 (7):1660–1671. doi:10.1016/j.bpj.2015.01.038]

Lipid Electropore Stabilization M. Laura Fernández and Marcelo R. Risk

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electropermeabilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Molecular Dynamics and Electropermeabilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pore Stability by Molecular Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stable Pores Obtained by Electric Field Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stable Pores Obtained by Charge Imbalance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stable Pores Obtained by Mechanical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pore Stability and Chemical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pore Stability by Continuum Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pore Stability in Planar Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Permeabilized State in Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 4 4 6 6 10 11 12 13 14 14 15 16 16

M.L. Fernández (*) Consejo Nacional de Investigaciones Científicas y Técnicas, Instituto de Física del Plasma, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Departamento de Física, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Buenos Aires, Argentina e-mail: [email protected] M.R. Risk Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Buenos Aires, Argentina Instituto Tecnológico de Buenos Aires (ITBA), Buenos Aires, Argentina e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_83-1

1

2

M.L. Fernández and M.R. Risk

Abstract

The stabilization of pores can be studied by different approaches such as simulations in silico or experimental procedures in vivo or in vitro. The energy to open a pore in a lipid membrane can be delivered by different external stimuli. To disrupt the membrane and initiate the pore opening, this energy has to reach a threshold. Then, once the pore is open, the external stimulus can be modulated to maintain the pore stable in time. This chapter first describes the basics of electropermeabilization, a process also called electroporation, and the basics of molecular dynamics in electropermeabilization. The chapter then describes in detail the molecular changes that lead to the pore opening and evolution by molecular dynamics. The chapter focuses on molecular dynamics because this technique allows the study of pore stabilization at molecular level, the interpretation of the lipid and water molecule rearrangements that are behind this phenomenon, and the visualization of the pore at the scale of size and time, in the order of nanometers and nanoseconds, respectively. Finally, the chapter also describes other approaches where pores remain open or the permeabilized state remains stable for a period of time, such as continuum modeling, experiments in planar membranes, and experiments in cells. The objective of this selection is to relate the results obtained by molecular dynamics with those obtained experimentally, or by other types of modeling, aiming to connect the mechanisms of pore stabilization by molecular dynamics at different scales. Keywords

Electropermeabilization • Electroporation • Pore stabilization • Molecular dynamics

Introduction The plasma membrane defines the boundaries of all living cells. Cell membranes are composed of lipid bilayers and proteins. The lipid bilayers of animal cell membranes are mainly composed of phospholipids and sterols, such as cholesterol. Both of these molecules are amphiphilic and have two defined parts: a hydrophilic one, consisting of a polar head group, and a hydrophobic one, composed of two hydrocarbon tails in phospholipids or four carbon rings and a hydrocarbon tail shorter than that of phospholipids in cholesterol. Due to their amphiphilicity and shape, when phospholipids are placed in water, they are spontaneously organized to form a bilayer with the polar head interacting with the water molecules, and the nonpolar tails of each layer facing each other in the inner part of the bilayer, preventing the contact of this hydrophobic region with the water molecules. This lipid bilayer separates the inner medium of the cell, called cytosol, from the extracellular fluid. The bilayer thus acts as a barrier with a selective permeability to small polar molecules (Fig. 1a). Another characteristic of cell membranes is the asymmetry in lipid composition of both leaflets. Under physiological conditions, the membranes are flexible and can break

Lipid Electropore Stabilization

3

Fig. 1 Evolution of the pore: Schematic three-dimensional representation of a bilayer in a simulation box. The red beads are the phospholipid polar head groups, the green sticks are the hydrocarbon lipid tails, and the light blue beads are the water molecules. The sizes of the components are not in scale. (a) Intact bilayer before the application of the electric field; (b) Beginning of the pore formation with a hydrophobic pore after the application of an external electric field with an intensity equal to or greater than the minimum porating electric field (E0); (c) Stage of expanded pore; (d) Stabilized pore at a value of electric field lower than E0. If the electric field applied is equal to or greater than E0, the pore begins to be formed and evolves to an expanded pore, from a to c. If the electric field applied is removed, the pore disappears, from c to a. If the electric field applied decreases to a low sustaining value (Es) lower than E0, the pore shrinks, from c to d. If the electric field applied increases to a value higher than the selected Es, the pore expands, from d to c

and spontaneously reseal. The disruption of the membrane exposes the hydrophobic surface formed by the lipid tails, hidden in the core of the bilayer, to the water molecules, creating an energetically unfavorable condition. This induces a rapid rearrangement of the lipids, aimed to repair the rupture in the membrane. The disruption in the continuity of the membrane can be artificially induced by different stimuli such as electrical or mechanical stress or chemical agents. If the intensity of the external electric field or the force applied to the membrane is high enough to exceed the respective membrane thresholds, the bilayer starts to break

4

M.L. Fernández and M.R. Risk

down due to the formation of artificial aqueous pores, which increase its permeability and even allow the diffusion of ions and molecules to which the membrane is normally impermeable. Thus, once the pore is open, water molecules, ions, metabolites, and other molecules can cross the membrane either from the inside to the outside of the cell or in the opposite direction, depending on where they were confined, following their own gradient. The molecules that are allowed to go through the membrane can be normal cellular metabolites as exogenous molecules, such as drugs, which can be placed outside the cell during an experimental procedure.

Electropermeabilization Electropermeabilization, also called electroporation (“▶ Electroporation and Electropermeabilization”), is a methodology of transient disruption of the membrane induced by pores due to the exposure of artificial bilayers or cell membranes to an external electric field in vitro or in vivo (Tsong 1991). The reversible or irreversible effects of this process on cell membranes are linked to the intensity (which can be high or low), the repetition, the duration (which can range from nanoseconds to microseconds), and the number of electric pulses applied. Artificial bilayers, cells or tissues, can be successfully electroporated by means of several protocols, which depend on the nature of the object of study selected. For instance, to electroporate cells in a reversible way, a field magnitude in the order of kV/cm can be applied in microsecond pulses, while higher field magnitudes in the order of MV/m can be applied in nanosecond pulses (Silve et al. 2014; Pakhomova et al. 2011; Vernier et al. 2006).

Molecular Dynamics and Electropermeabilization The molecular events produced at atomic or molecular level that lead to pore formation cannot be visualized experimentally. So, the molecular structures of the induced pores during their formation, maturation, and closure can only be studied and described by molecular dynamics simulations. Molecular dynamics is a method that allows the modeling of a biological structure such as the lipid bilayer at atomic level. In this method, the structure of interest (lipid bilayers, proteins, nucleic acids, etc.) is placed in a tridimensional box where the initial position of each atom is known. The structure in the box can be solvated by different approaches, such as adding implicit or explicit solvents, thus giving the biomolecule, or the biological structure, an adequate aqueous environment, including ions. Each atom in the simulation box is related to and interacts with the others and has a defined initial position and an assigned velocity. Then, by solving the Newton’s equation of motion, the forces on each particle due to the interactions are calculated and the movement of the atoms is simulated at each time step, in the order of femtoseconds. At each time step, the new position or velocity of each atom is generated. Following the movement of each atom at each time step, a trajectory is

Lipid Electropore Stabilization

5

generated and the changes of the whole simulated system can be evaluated along the time. The molecular dynamics method works with force fields, which allow the calculation of the potential energy at each step. The force field is constituted by a functional form to calculate the potential energy and all the parameters needed for this calculation. With the functional form and the parameters, the potential energies and the resulting forces can be calculated as a function of the position of each interacting atom. In the simplest force field, the calculation of the energy is the sum of terms that model the interactions. The following methods have been developed for molecular dynamics simulations to induce the formation of pores and study their further evolution. External Electric Field Molecular dynamics procedures allow simulating the application of an external electric field in the z-direction (transversal to the membrane) as a new force in the system, exerted over all atoms with partial charges, in the calculations of the potential energy. By means of this simulation, the transmembrane potential rises until the membrane threshold is reached, causing the formation of the pore (“▶ Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Experiments”). This method allows the comparison of results obtained by molecular dynamics with those obtained experimentally, where high-intensity nanosecond pulsed electric fields are applied to a system (Tieleman 2004; Tarek 2005). Ion Charge Imbalance Another approach to describe the generation of electropores is the induction of a transmembrane potential by the method of ion charge imbalance, where the transmembrane potential is generated by imposing an ion charge imbalance across the bilayer. This method simulates the effects on the transmembrane potential due to the charge imbalance in the membrane, which acts as a capacitor. The results obtained from these molecular dynamics experiments have been previously compared with those obtained experimentally (Sachs et al. 2004; Gurtovenko and Vattulainen 2005; Delemotte et al. 2008). Mechanical Stress Another way to induce pores in membranes, in the absence of an electric field, is by mechanical stress. This method is based on the application of a surface tension, which exceeds the bilayer threshold tension needed to maintain the continuity of the membrane and initiates the process of pore opening (Leontiadou et al. 2004). Chemical Agents The addition of amphiphilic molecules to the simulated system can affect the membrane stability and induce the pore formation. The presence of these molecules can also lower the electroporation threshold (Fernández and Reigada 2014; Polak et al. 2015). The following sections in this chapter describe the principal methods to obtain and maintain stable electropores.

6

M.L. Fernández and M.R. Risk

Pore Stability by Molecular Dynamics As mentioned above, the formation of pores in bilayers may be induced by several methods by molecular dynamics. This chapter describes some of these methods, including the application of an external electric field, the modification of the transmembrane potential due to charge imbalance, and the poration process by mechanical stress or by chemical agents, in order to analyze the effects of these methods on the stability of pores.

Stable Pores Obtained by Electric Field Application When an external electric field is applied perpendicularly to the membrane (z-direction) and the intensity of the field reaches the electroporation threshold, the pore begins to be formed. The only way to analyze the molecular events involved in the initiation of the formation and evolution of an induced pore in the membrane during electroporation is through in silico simulations. The method of molecular dynamics simulation has revealed that the application of an external electric field that reaches a threshold, the minimum porating electric field (E0), in an intact membrane (Fig. 1a) induces and then stabilizes water defects. These defects, which are also called fingers, are formed in the interface between the water molecules and the hydrophobic part of the bilayer membrane. These water molecules forming the defects are reorganized by the alignment of their dipole moments with the applied electric field starting the pathway to the formation of the pore. These fingers are extended across the hydrophobic core of the bilayer membrane, connecting to another defect formed in the opposite leaflet, finally resulting in a continuous path of water wires. These structures are known as hydrophobic pores because the water molecules are in contact with the hydrophobic part of the lipids, which constitutes the inner part of the bilayer (Fig. 1b). The new contacts between the water molecules and the hydrophobic tails of the lipids are energetically unfavorable. As mentioned previously, in a physiological situation, the disruption of the membrane continuity is immediately resealed, but, in this situation, the stimulus that causes the disruption continues exerting its effect over the membrane, preventing the resealing and inducing more water molecules to penetrate into the hydrophobic core, thus thickening the water wires. Immediately after, the lipids start a structural rearrangement to surround the water wires. This rearrangement begins with the movement of the polar lipid head group, in order to reach the inner part of the bilayer, and therefore ensuring the complete lining of the water molecules. This stage of the pore formation process is called hydrophilic pore. Then, this process restores the prepore surface contact, where the water molecules are located only around the polar lipid head groups and are not allowed to be in contact with the hydrophobic region of the bilayer (Fig. 1c). Now, at this point of the formation of the pore, and under the influence of the external electric field, a new structure of an aqueous pore in the bilayer membrane is created, where there is a sort of continuity between the lipids of the upper leaflet and

Lipid Electropore Stabilization

7

those of the lower leaflet, which allows the movement of lipids from one layer to the other. Prior to the formation of the pore, this arrangement was nonexistent because the migration of the lipids from one leaflet to the other, known as flip-flop process, is a very infrequent event in normal conditions, so the external electric field stabilizes this pore configuration. If the stimulus of the external electric field applied continues in time, the hydrophilic pore keeps growing to form an expanded pore (Fig. 1c), but if the stimulus is not removed, the growth of the pore ends up in the destabilization of the lipid bilayer and thus leads to the rupture of the membrane. Finally, removal of the field during the stage of expanded pore leads to the destabilization of the pore and later to pore annihilation, returning the membrane to the ground state (Tieleman 2004; Tarek 2005; Levine and Vernier 2010; Tokman et al. 2013) (“▶ Lipid Electropore Lifetime in Molecular Models”). Molecular dynamics studies show that electropores can be stabilized in two steps. In the first step, high electric field strength is applied in the z-direction. This has been tested, for instance, in the simulation of a bilayer formed by only one lipid, such as palmitoyl-oleoyl-phosphatidylcholine, and of bilayers with a mixture of lipids mimicking the membranes of bacteria, such as Staphylococcus aureus and Escherichia coli. In all the simulations, the pore evolution is the same as that depicted in Fig. 1a–c. The simulations are stopped when an expanded pore is obtained (Fig. 1c). In the second step, the structure of the expanded pore obtained in the first step is used as the initial configuration in a second simulation, where the field magnitude is reduced to obtain low field stabilized electropores for several nanoseconds, as depicted in Fig. 1d (Böckmann et al. 2008; Piggot et al. 2011). This two-step protocol for the development and stabilization of the pore may allow understanding the effect on a preformed expanded pore by modulating the external electric field to several sustaining values (ES) lower than the minimum porating electric field (E0) (Fernández et al. 2012). The initial configuration for an expanded pore can be obtained in a first-step simulation performed at the minimum porating electric field for a bilayer composed of dioleoylphosphatidylcholine and cholesterol. Then, in the second step, several simulations can be performed by applying different field values lower than the minimum porating electric field to analyze the temporal evolution of the pore volume. Three different ES values have been found to stabilize the pore indefinitely at different final volumes. One of these results is shown in Fig. 1d, where the sustaining value for the field maintains the pore open in a volume lower than that obtained for the expanded pore. Figure 2 represents the Es values obtained in the above-described simulation and shows that the higher the Es, the higher the volume of the pore (Fig. 2a–c). The whole volume of the pore is estimated dividing the inner part of the pore, located into the hydrophobic core of the bilayer, into bins. In each bin, a cylindroid that includes all the water molecules present in the pore is constructed. So, the whole volume of the pore is calculated as the sum of the volumes of stacked cylindroids inside the pore (Fig. 3 and Fig. 2d–g). These results show the three field sustaining values and determine two nonequilibrium conditions for the stability of the pore:

Fig. 2 Snapshots for the stabilized pore structures after 100 ns at (a) Es = 0.125 MV/m; (b) Es = 0.150 MV/m; and (c) Es = 0.175 MV/m. (d–f) Plots of the (x, z) view of the cylindroids that form the stabilized pores (Figure from Fernández et al. 2012, taken with permission from Elsevier). Blue spheres are water oxygen atoms, red spheres are the phosphorus atoms of dioleoylphosphatidylcholine, yellow spheres are the hydroxyl oxygen atoms of cholesterol, green spheres are the first sn-1 carbon atoms of dioleoylphosphatidylcholine, and turquoise stick structures are the lipid hydrocarbon tails. All the snapshots were performed with Visual Molecular Dynamics (Humphrey et al. 1996)

8 M.L. Fernández and M.R. Risk

Lipid Electropore Stabilization

9

Fig. 3 Schematic representation of (a) the simulation box, indicating the water region, the upper and the lower polar regions of the bilayer and the internal nonpolar region of the bilayer and (b) the cylindroids that constitute the pore (lateral and top view). (c) Snapshot of the initial pore structure. (d) Plot of the (x, z) view of the cylindroids that form the starting pore. The blue spheres are the water oxygen atoms, the red spheres are the phosphorus atoms of dioleoylphosphatidylcholine, the yellow spheres are the hydroxyl oxygen atoms of cholesterol, and the green spheres are the first sn-1 carbon atoms of dioleoylphosphatidylcholine. (e) Plot of the (x, y) view of the cylindroids that characterize the water pore. Crosses correspond to the center of the cylindroids. The cylindroids of the nonpolar region that contribute to the pore volume are in red; the “external” cylindroids excluded from the calculation are in blue (Figure from Fernández et al. 2012, taken with permission from Elsevier)

E < ES: If the field applied is less than the minimum sustaining value, it leads to a decrease in the pore volume and later to the closure of the pore, as it occurs when the field is removed (E = 0), thus allowing the complete resealing of the membrane (Fig. 1 from 1c to 1a). So, fields lower than the minimum Es are not high enough to maintain the pore open in a constant volume. E > ES: If the field applied is higher than the maximum sustaining value, it leads to a continued growth of the pore volume. These results show a direct association between the magnitude of ES and the volume of the stable pore, indicating that the pore size can be controlled and sustained in time (Fernández et al. 2012). The method of controlling the pore size by adjusting the value of the sustaining field can also be used to measure the ion conductance through nanopores (Ho et al. 2013). The simulations performed using this method have been developed in two systems: one in the presence of sodium chloride and the other in the presence of potassium chloride, showing that the high interaction of the sodium ions with the

10

M.L. Fernández and M.R. Risk

interface between water and phospholipids induces a slower conductance than potassium ions. This slowing effect can be compensated by increasing the field intensity and as a consequence increasing the pore radius, since at higher pore radii, the interaction of the sodium ions with the pore wall diminishes as the bulk water predominates inside the pore. Another way to modify the pore size is by periodically modulating the intensity of the field (Kohler et al. 2015). This can be simulated by means of a time-varying external electric field. Then, using a unipolar sinusoidal function, the variations in the intensity of the field are associated with variations in the same direction of the pore size. Therefore, a modulation in the expansion and shrinkage of the pore can be achieved in a single simulation following the same frequency applied to the electric field. This is schematically shown in Fig. 1 (Fig. 1c–d and d–c).

Stable Pores Obtained by Charge Imbalance This method generates an artificial charge imbalance by simulating bilayers that separate two water regions where different ion concentrations are imposed. This ion imbalance generates a transmembrane potential able to induce pores as in simulations where an external electric field is applied. There are two proposed setups: 1. In the first one, two bilayers are modeled in the same simulation box, separated by a water region. So, one leaflet of each bilayer is in contact with the middle water slab, and the other faces a second aqueous region. This second water slab is connected by the upper and lower borders of the box in the z-direction due to the periodic boundary conditions. 2. In the second one, a single bilayer separating two water slabs is modeled, but these aqueous regions are not connected because two additional vacuum/air slabs are inserted in the upper and the lower borders of the box in the z-direction, generating a water/vacuum interface that impedes the ionic movements between both water reservoirs (Kirsch and Böckmann 2015). In both methods, the ion diffusion through the pore dissipates the charge imbalance, leading to a drop in the transmembrane potential. Molecular dynamics simulations have shown that in a double bilayer of dimyristoylphosphatidylcholine system with a sodium charge imbalance, a pore begins to be formed within the first nanoseconds in one of the two bilayers. As the ions cross the bilayer, the field decreases and the permeation of the sodium ions and the pore radius decrease, leading to a stable pore for several nanoseconds, with a remaining charge imbalance of only one sodium ion. If this residual charge imbalance is removed while the pore is still stable, the pore remains open for a few additional nanoseconds and then closes in less than 10 ns. This indicates that if the transmembrane voltage is lower than a certain threshold, the pore becomes metastable and independent of the fall of the transmembrane potential. In this condition

Lipid Electropore Stabilization

11

where the transmembrane voltage is below the threshold, the closure is associated with the pore size fluctuations, rather than with the disappearance of the charge imbalance (Gurtovenko and Vattulainen 2005). This is comparable with the simulations performed in size-controlled pores by electric field, where if the field applied is lower than the Es threshold, the pore shrinks and finally closes even in the presence of an external electric field (Fernández et al. 2012). In the simulations performed to evaluate the charge imbalance effect on the transmembrane potential and the pore induction using either the method of the double bilayer or the vacuum method, as described in the previous example, when the pore opens, the ion gradient dissipates. When this happens, the transmembrane potential decreases and the pore cannot be longer sustained in time. This drop in the transmembrane potential can be avoided by maintaining the initial imposed charge imbalance, by the swapping method (Kutzner et al. 2011). In this approach, if one ion crosses the membrane from its initial water reservoir to the other, this ion and a water molecule from the initial reservoir are swapped and the charge imbalance is restored. This method allows obtaining stable pores at different transmembrane potentials and a pore radius that is linearly associated with the transmembrane potential (Casciola et al. 2016).

Stable Pores Obtained by Mechanical Methods The previous two sections of this chapter described the methods to open a pore and maintain it open in a stable way by molecular dynamics, which is by using external electric fields or the charge imbalance method. As previously mentioned, another method that can be used to form pores is that based on mechanical stress. This mechanical stress may be induced experimentally by the aspiration of the lipid structures of a membrane, a vesicle or cells into a micropipette. This mechanical stress produced by the aspiration, in other words, by a negative pressure, makes possible the rupture of the membrane in the form of a pore, equivalent to the electropores obtained by exposure to electric fields (Kirsch and Böckmann 2015). Both the opening and closure of a pore can be considered to be due to the interaction of two tensions: one aiming to maintain the pore closed (the so-called line tension) and the other aiming to open the pore (the so-called surface tension). The latter causes a stretching of the membrane. Then, the aspiration of the micropipette causes a negative pressure, causing a surface tension at the level of the membrane. The energy of formation of the pore (Epf), according to the theory of pore formation, is defined by the following equation: Epf = 2πRTL πR2TS where R is the radius of the pore, TL is the line tension, and TS is the surface tension (Barnett and Weaver 1991). This equation of pore formation shows the energy necessary to open a pore, which is the difference between the energy produced by the line tension and the surface tension components, in a way that, if the line tension is maintained, an

12

M.L. Fernández and M.R. Risk

Fig. 4 Three scenarios where the radius R of the pore increases as the surface tension increases: R 1 when the surface tension is small and increasing, R0 > R 1 when the surface tension reaches the critical threshold for tension, and finally R1 when the surface tension is beyond the critical threshold for tension

increase in surface tension will result in a decrease in the energy needed to open or maintain a pore open. In molecular dynamics simulations of a dipalmitoylphosphatidylcholine lipid bilayer membrane, a surface tension is applied, causing the formation of stable hydrophilic pores (Leontiadou et al. 2004). This mechanical stress on the membrane, under the form of a surface tension, was the result of the application of a lateral pressure, in this way simulating a pressure produced by the aspiration of a micropipette. Molecular dynamics simulations were performed in two different initial states: one using a bilayer membrane with a metastable pore, and the other using an equilibrated bilayer membrane without a pore. In both states, a threshold of critical surface tension was reported. Below this threshold, the pores remain stable, reaching a minimum radius in the order of 1 nm, at the center of a toroidal-shaped pore. If the critical threshold is exceeded, the pore results unstable, thus expanding, and finally causing the rupture of the membrane. Figure 4 shows three scenarios of the opening of the pore due to mechanical stress, in a transversal plane view, where the radius R of the pore is increased as the surface tension increases: R 1 when the surface tension is small and increasing, R0 > R 1 when the surface tension reaches the critical threshold for tension, and finally R1 when the surface tension is beyond the critical threshold for tension. In another work, the study of the transport of ions across pores in the bilayer is combined with mechanical stress to stabilize pores with different radii. The presence of ions seems to induce a reduction of the stability of the pores, mainly because of the increase in the line tension. As stated before, this tension maintains the pore closed, and therefore, an extra mechanical stress is needed to stabilize the pore. This can be attributed to the binding of sodium ions at the interface between the lipids and water, in other words, the surface of the bilayer membrane (Leontiadou et al. 2007).

Pore Stability and Chemical Methods The addition of amphiphilic agents to the system can affect the integrity of the membrane due to the chemical nature of these compounds. A small amphiphilic molecule such as dimethyl sulfoxide added to a dipalmitoylphosphatidylcholine bilayer can alter the normal structure of the bilayer. Three dimethyl sulfoxide

Lipid Electropore Stabilization

13

concentration regimes are described to have different effects on the bilayer (Gurtovenko and Anwar 2007): • At a low concentration, the molecules of dimethyl sulfoxide locate in the interface between the lipids and water molecules, inducing a lateral expansion of the bilayer and a decrease in the bilayer thickness. • An intermediate concentration induces a decrease in the bilayer thickness and the dimethyl sulfoxide molecules locate in the interface as well as in the hydrophobic core of the bilayer, favoring the intrusion of water molecules to the inner part of the bilayer and leading to the stabilization of hydrophobic pores and then allowing the formation of hydrophilic pores. • A high concentration leads to the rupture of the bilayer. A molecular dynamics simulation work performed in dioleoylphosphatidylcholine bilayers in the presence of different amounts of cholesterol combined the effect of different concentrations of dimethyl sulfoxide with the presence of an external electric field. The analysis of the results showed that the dimethyl sulfoxide can facilitate the stability of the hydrophobic pores and diminishes the electroporation threshold at very low concentrations of dimethyl sulfoxide even in the presence of cholesterol (Fernández and Reigada 2014). The effects of another amphiphilic molecule, the polyoxyethylene glycol, on membranes can be analyzed by molecular dynamics and compared with experimental data. The results obtained using the method of charge imbalance to induce pores showed that polyoxyethylene glycol seems to favor the formation and stabilization of the water columns, even allowing the ion transport across the bilayer through these stabilized pores. In a later stage, a few lipid molecules migrate to increase the pore stabilization. The presence of this molecule lowers the electroporation threshold (Polak et al. 2015). This finding is in line with the results described in previous experiments performed in planar lipid bilayers electroporated in the presence of polyoxyethylene glycol (Troiano et al. 1998).

Pore Stability by Continuum Modeling An approach different from molecular dynamics to simulate the effects of electroporation is the continuum modeling of a cell exposed to an electric field (Son et al. 2014). Two different pulses are included in the model: a low-intensity long pulse, in the order of microseconds, and a high-intensity nanosecond pulse. In this model, the lifetime of the pore is a parameter in the order of seconds. The results obtained using this approach show that the long pulses lead to populations of pores with a wide range of sizes distributed in two major subpopulations with radius of approximately 1–10 nm and that the short pulses produce only one population of small size pores but in high number. This model indicates that the differences in the size of the radii of the pores obtained for each pulse type are related to the association of the radius with the duration of the pulse, because the pores can only grow during the application of

14

M.L. Fernández and M.R. Risk

the pulse. This is consistent with the results obtained in molecular dynamics simulations where, if the electric field applied exceeds the threshold of the sustaining electric field, the pore keeps growing. On the other hand, the model also explains the differences in the number of pores formed by each pulse, associating the dependence of the pore creation on the transmembrane voltage. So, in this model, high-intensity pulses induce a greater number of pores (Son et al. 2014). Another model simulating cells in 3-dimensions takes into account the conducting and the permeable states separately in the calculation to simulate the effects on the membrane and the transport due to the application of an external electric field. These states depend, among other parameters, on poration characteristic time, permeabilization dynamic, and membrane recovery time, all of them inputs of the model. In this model also, the pores open when a voltage threshold is exceeded (Leguèbe et al. 2014).

Pore Stability in Planar Membranes Pore stability can also be studied in planar membranes. In these experiments, a planar bilayer lipid membrane is formed, separating two water reservoirs, and then the effects of membrane poration are analyzed under current-constant conditions. In a work performed by Koronkiewicz et al., the constant current causes an increase in the transmembrane potential and the following membrane poration. Once the current applied reaches the membrane threshold, the potential increases up to 250 mV and then keeps varying around 150 mV for several nanoseconds. These potential variations can be attributable to the generation of a pore with a fluctuating size in the order of nanometers. In a different approach, the current is maintained for 10 ns and then interrupted for few nanoseconds during several on-off current cycles. In this approach, the pore keeps its radius stable for the whole experiment. If the interval of interruption of the current is increased up to 100 s, the pore still remains stable in size, but at higher interruption periods (300 s), the pore disappears and the continuity of the membrane is fully restored (Koronkiewicz et al. 2002). This indicates that the stability of the pore is sustained even in the absence of the stimulus for a period of time, but in the end the membrane restores its initial continuity, as it occurs in molecular dynamics pore simulations (“▶ Experimental Electroporation of Planar Lipid Bilayers”).

Permeabilized State in Cells This section describes different electroporation protocols for pore stabilization in experimental procedures in cells (“▶ Different Cell Sensitivity to Pulsed Electric Field”). Early works performed in unfertilized sea urchin eggs in a low-calcium medium, similar to the calcium concentration of the intracellular medium, showed that cells can be loaded with radiolabeled low molecular weight metabolites for

Lipid Electropore Stabilization

15

several minutes. These works indicated that, in these conditions, the permeabilized state remains stable for this period of time (Swezey and Epel 1989). In experiments performed by Pakhomov et al. in CHO cells preloaded with a thallium-sensitive fluorophore and exposed to nanopulses, thallium was added at different times after the application of the pulse. The measure of thallium uptake due to electroporation indicates that the nanopores remain open for several minutes after the application of the electroporating stimulus (Pakhomov et al. 2009).

Summary Molecular dynamics simulations show that pores in a lipid bilayer membrane can be obtained and stabilized either by electrical or mechanical stress, or by addition of chemical agents, such as amphiphilic molecules. In either electrical or mechanical stress, the simulations show that the external stimulus has to reach a minimum electric field or mechanical stress, in order to initiate the pore opening, and then the stimulus has to be maintained lower than a critical value to stabilize the pore. The addition of amphiphilic molecules to the simulated system facilitates the stabilization of the water columns, which initiate the pore formation and decrease the electroporation threshold. The approaches using molecular dynamics establish that the modulation of the external stimulus in values below the threshold modifies the size and keeps the pores stable, while the stimulus is still present. In the case that the stimulus is ceased or the intensity is lower than a sustaining value, the pore closure begins. The approach to simulate the effects of an electric field on cells with the continuum modeling is less detailed than molecular dynamics but more versatile in terms of incorporating parameters as inputs of the model to describe the electroporation process. The experiments made on planar membranes show that the pore stability can also be sustained in time. The experiments of electroporation in cells show that the permeabilized state can be detected for several minutes, thus indicating higher stability of this state when the experiments are performed in cells. The stability of the pores or the permeabilized state is related to the complexity of the system and may range from nanoseconds in molecular dynamics simulations to seconds in the continuum modeling and the planar membrane experiments under constant current conditions, and to minutes in cell experiments with pulses of different intensity and duration. Finally, the stability of the pores is due to the modulation of the stimulus, for both molecular dynamics and planar membrane experiments. In cells, the mechanisms that stabilize the permeabilized state after electroporation remain unknown. This could be attributed to the complexity of the plasma membrane and to the interactions of the plasma membrane with other components either inside or outside the cell.

16

M.L. Fernández and M.R. Risk

Acknowledgments This work was supported in part by grants from Universidad de Buenos Aires (UBACyT GC 20620130100027BA), CONICET (PIP GI 11220110100379) and ITBA (ITBACyT 2015).

Cross-References ▶ Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Experiments ▶ Different Cell Sensitivity to Pulsed Electric Field ▶ Electroporation and Electropermeabilization ▶ Experimental Electroporation of Planar Lipid Bilayers ▶ Lipid Electropore Lifetime in Molecular Models

References Barnett A, Weaver JC (1991) Electroporation: a unified, quantitative theory of reversible electrical breakdown and mechanical rupture in artificial planar bilayer membranes. Bioelectrochem Bioenerg 25:163–182 Böckmann RA, de Groot BL, Kakorin S, Neumann E, Grubmüller H (2008) Kinetics, statistics, and energetics of lipid membrane electroporation studied by molecular dynamics simulations. Biophys J 95:1837–1850 Casciola M, Kasimova MA, Rems L, Zullino S, Apollonio F, Tarek M (2016) Properties of lipid electropores I: Molecular dynamics simulations of stabilized pores by constant charge imbalance. Bioelectrochemistry 109:108–116 Delemotte L, Dehez F, Treptow W, Tarek M (2008) Modeling membranes under a transmembrane potential. J Phys Chem B 112:5547–5550 Fernández ML, Reigada R (2014) Effects of dimethyl sulfoxide on lipid membrane electroporation. J Phys Chem B 118:9306–9312 Fernández ML, Risk M, Reigada R, Vernier PT (2012) Size-controlled nanopores in lipid membranes with stabilizing electric fields. Biochem Biophys Res Commun 423:325–330 Gurtovenko AA, Anwar J (2007) Modulating the structure and properties of cell membranes: the molecular mechanism of action of dimethyl sulfoxide. J Phys Chem B 111:10453–10460 Gurtovenko AA, Vattulainen I (2005) Pore formation coupled to ion transport through lipid membranes as induced by transmembrane ionic charge imbalance: atomistic molecular dynamics study. J Am Chem Soc 127:17570–17571 Ho M-C, Casciola M, Levine ZA, Vernier PT (2013) Molecular dynamics simulations of ion conductance in field-stabilized nanoscale lipid electropores. J Phys Chem B 117:11633–11640 Humphrey W, Dalke A, Schulten K (1996) VMD – visual molecular dynamics. J Mol Graph 14:33–38 . http://www.ks.uiuc.edu/Research/vmoleculardynamics Kirsch SA, Böckmann RA (2015) Membrane pore formation in atomistic and coarse-grained simulations. Biochim Biophys Acta. doi:10.1016/j.bbamem.2015.12.31 Kohler S, Levine ZA, García-Fernández MA, Ho M-C, Vernier PT (2015) Electrical analysis of cell membrane poration by an intense nanosecond pulsed electric field using an atomistic-tocontinuum method. IEEE Trans Microwave Theory Techn 63:2032–2040 Koronkiewicz S, Kalinowski S, Bryl K (2002) Programmable chronopotentiometry as a tool for the study of electroporation and resealing of pores in bilayer lipid membranes. Biochim Biophys Acta 1561:222–229

Lipid Electropore Stabilization

17

Kutzner C, Grubmüller H, de Groot BL, Zachariae U (2011) Computational electrophysiology: the molecular dynamics of ion channel permeation and selectivity in atomistic detail. Biophys J 101:809–817 Leguèbe M, Silve A, Mir LM, Poignard C (2014) Conducting and permeable states of cell memabrane submitted to high voltage pulses: mathematichal and numerical studies validated by the experiments. J Theor Biol 360:83–94 Leontiadou H, Mark AE, Marrink SJ (2004) Molecular dynamics simulations of hydrophilic pores in lipid bilayers. Biophys J 86:2156–2164 Leontiadou H, Mark AE, Marrink SJ (2007) Ions transport across transmembrane pores. Biophys J 86:2156–2164 Levine ZA, Vernier PT (2010) Life cycle of an electropore: field-dependent and field-independent steps in pore creation and annihilation. J Membr Biol 236:27–36 Pakhomov AG, Bowman AM, Ibey BL, Andre FM, Pakhomova ON, Schoenbach KH (2009) Lipid nanopores can form a stable, ion channel-like conduction pathway in cell membrane. Biochem Biophys Res Commun 385:181–186 Pakhomova ON, Gregory BW, Khorokhorina VA, Bowman AM, Xiao S, Pakhomov AG (2011) Electroporation-induced electrosensitization. PLOS One 6:e17100 Piggot TJ, Holdbrook DA, Khalid S (2011) Electroporation of the E. coli and S. aureus membranes: molecular dynamics simulations of complex bacterial membranes. J Phys Chem B 115:13381–13388 Polak A, Velikonja A, Kramar P, Tarek M, Miklavčič D (2015) Electroporation threshold of POPC lipid bilayers with incorporated polyoxyethylene glycol (C12E8). J Phys Chem B 119:192–200 Sachs JN, Crozier PS, Woolf TB (2004) Atomistic simulations of biologically realistic transmembrane potential gradients. J CHem Phys 121:10847–10851 Silve A, Brunet AG, Al-Sakere B, Ivorra A, Mir LM (2014) Comparison of the effects of the repetition rate between microsecond and nanosecond pulses: electropermeabilization-induced electro-desensitization? Biochim Biophys Acta 1840:2139–2215 Son RS, Smith KC, Gowrishankar TR, Vernier PT, Weaver JC (2014) Basic features of a cell electroporation model: illustrative behavior for two very different pulses. J Membr Biol 247:1209–1228 Swezey RR, Epel D (1989) Stable, resealable pores formed in sea urchin eggs by electric discharge (electroporation) permit substrate loading for assay of enzymes in vivo. Cell Regul 1:65–74 Tarek M (2005) Membrane electroporation: a molecular dynamics simulation. Biophys J 88:4045–4053 Tieleman DP (2004) The molecular basis of electroporation. BMC Biochem 5:10 Tokman M, Lee JH, Levine ZA, Ho M-C, Colvin ME, Vernier PT (2013) Electric field-driven water dipoles: nanoscale architecture of electroporation. PLOS One 8:e61111 Troiano GC, Tung L, Sharma V, Stebe KJ (1998) The Reduction in Electroporation Voltages by the Addition of a Surfactant to Planar Lipid Bilayers. Biophys J 75:880–888 Tsong TY (1991) Electroporation of cell membranes. Biophys J 60:297–306 Vernier PT, Ziegler MJ, Sun Y, Gundersen MA, Tieleman DP (2006) Nanopore-facilitated, voltagedriven phosphatidylserine translocation in lipid bilayers–in cells and in silico. Phys Biol 3:233–247

Molecular Transmembrane Transport with Giant Unilamellar Vesicles (GUVs) Marie-Pierre Rols

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Giant Lipid Vesicles Formation and Electropulsation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background on Lipid Vesicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electroformation of GUV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electropulsation of GUV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effects of Electric Fields on GUV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Different Approaches to Assayed Molecular Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quantitative Optical Microscopy and Micromanipulation Studies . . . . . . . . . . . . . . . . . . . . . . . . . . Visualization and Quantification of Transmembrane Ion Transport . . . . . . . . . . . . . . . . . . . . . . . . . Monitoring and Quantifying the Passive Transport of Molecules Through Patch-Clamp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Imaging Molecular Transport Across Lipid Bilayers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Imaging Electropermeabilization Process by CARS Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . Imaging Lipid Loss Associated to Electropermeabilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 3 4 6 8 10 10 11 11 12 12 13 15 16 16

Abstract

Giant vesicles provide a useful model system for measuring a variety of physical properties of lipid membranes and for improving our understanding of the electropermeabilization and electrofusion phenomena. The purpose of this chapter is to present the ways GUV can be formed and submitted to electric pulses. Moreover, it summarizes the presently known effects of electric fields on giant vesicles and some of their practical applications. Subjecting GUVs to DC pulses can destabilize the lipid bilayer, inducing different behavior such as deformation M.-P. Rols (*) IPBS-CNRS (UMR 5089), University of Toulouse, Toulouse, France e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_84-1

1

2

M.-P. Rols

of the vesicles, permeabilization, and motion of domains or fusion that can be directly observed under a microscope. In addition to usual optical microscopies, a number of approaches can be used to assayed molecular transport are described, including micromanipulation, patch-clamp, and CARS spectroscopy. Electric pulses can have dramatic consequences on membrane structure and function. With suitable electric field parameters, it is possible to electropermeabilize the membrane leading to exchange of molecules between the inside and the outside of the vesicle. If the electropermeabilization is strong enough, membrane vesicles, tubules, and macropores can be visualized and are in general accompanied with loss of lipid membrane material and a subsequent GUV size decrease. GUV can therefore be considered as useful tools for resolving the effect of electric fields on cells, even if they present some clear limits in particular for the transport of plasmid DNA.

Keywords

Giant unilamellar vesicles • Model membranes • Electrodeformation • Electroporation • Electrofusion • Tubules • Microscopy

Introduction Membranes are key constituents of cells, acting as barriers which hinder the free passage of ions and hydrophilic molecules. If this impermeability is crucial for life to develop, it can represent a hurdle in medicine and biotechnologies where molecule delivery and/or extraction are required. Molecule transport across biological membranes can be obtained by several methods. Virus bioinspired vectors are quite efficient, but their safety has been questioned. Therefore, alternative chemical and physical methods have been developed in parallel. Among the physical methods, electropermeabilization is probably the most promising one mainly due to its efficiency and safety. Indeed, cell membranes can efficiently be permeabilized by applying electric pulses. Provided that the electric pulses are of a sufficient and welldefined amplitude and duration, impermeant molecules can enter or leave the cytoplasm of cells. Because of its efficiency, this method, also referred as electroporation, has rapidly becoming an established approach in medicine for the local treatment of cancer (Yarmush et al. 2014) and it also holds great promise for gene therapy. Electropermeabilization is also of high interest in biotechnology in particular for molecule extraction in food industry (Kotnik et al. 2015). If the exact mechanisms of molecular uptake are still not fully understood for the moment, it is known that the transport of molecules depends strongly on the size of the molecules. Apart from pulse strength that has to be higher than a threshold, pulse duration is another key parameter for the success of permeabilization, especially in the case of large molecules such as plasmid DNA which are driven by electrophoresis toward the permeabilized membrane (Rosazza et al. 2016). A precise description of the events leading to membrane permeabilization, and its consequences on lipid

Molecular Transmembrane Transport with Giant Unilamellar Vesicles (GUVs)

3

organization, still is missing and would help design safer and more efficient protocols particularly in the case of gene transfer. However, studying such phenomena at the molecular level in cells and tissues is unfortunately out of reach due to the high complexity of cells. This is a major motivation for investigating the behavior of simpler model systems, namely lipid vesicles. The conception of the role of lipid molecules in biological membranes has dramatically evolved over the past few decades. The lipid membrane was first considered as a purely passive barrier and a substrate to membrane proteins which carried out biological functions. It is now widely accepted that lipids play an active role in a number of biological processes. Their physicochemical properties, conferred by their structure and their amphiphilic nature, have consequences well beyond their self-assembly and ability to form a barrier. It is now possible to form cell-sized artificial membranes with well-controlled compositions. These objects are called Giant Unilamellar Vesicles (GUVs) and are also referred to as giant vesicles or giant liposomes. Their study has become increasingly popular in chemistry, biology, and physics laboratories. Their primary interest is probably their size which, being of the order of several micrometers, allows their direct observation via optical microscopy techniques. In addition, this system presents the advantage of having no cytoskeleton. Dramatic membrane consequences can therefore be observed such as deformation and lipid loss. During the 1980s, Angelova and Dimitrov developed an efficient protocol for giant vesicles production known as electroformation (Angelova and Dimitrov 1986), which was later shown to indeed produce unilamellar vesicles (Rodriguez et al. 2005). Due to the ease of their fabrication and their rich phenomenology, artificial vesicles have received an ever growing interest from the scientific community. This chapter aims to present what is known about the various phenomena that occur when GUVs are subjected to electric fields. Reviews have been already published about the effects of electric fields on GUVs (Haluska et al. 2006) and about the physics of GUVs in general (Dimova et al. 2006). The present paper is therefore obviously not exhaustive. However, it should be of use to newcomers in the field or to people familiar with previous reviews and who wish to know about recent developments. It is organized in two main sections: (Yarmush et al. 2014) description of synthetic lipid vesicles and tips to synthetize and submit them to electric pulses, (Kotnik et al. 2015) presentation of different approaches to assayed molecular transport on giant liposomes, along with their perspectives in biology and medicine for molecule delivery when they exist.

Giant Lipid Vesicles Formation and Electropulsation Background on Lipid Vesicles Synthetic lipid vesicles provide membrane models suitable for systematic investigations of the effects of electric fields on lipid bilayers such as electrodeformation, electropermeabilization, and electrofusion. Membrane electropermeabilization has

4

M.-P. Rols

been first assessed by conductance measurements or transfer of radioactive or fluorescent molecules. Different types of vesicles have therefore been used to study the underlying process, but only the largest vesicles (above 10 μm) can offer the direct visualization of the processes. First experiments have been performed on small unilamellar vesicles (SUV) (15–30 nm) more than 30 years ago. A transient leakage of radioactive sucrose was detected and was attributed to the formation of pores, which allowed passage of sucrose. However, there was no direct proof of their existence, or other related phenomena such as membrane deformations. DNA electrotransfer into liposomes was first studied 20 years ago on large unilamellar vesicles (LUV) (100–200 nm), suggesting endocytosis-like vesicles which shielded the electrotransferred DNA from the internal medium. Yet, also in this case, the liposomes could not be observed because of their small size. The results were subsequently questioned by data obtained on giant liposomes supporting a mechanism involving electropores and a direct entrance of fluorescent DNA into the liposomes. GUVs containing fluorescent phospholipid analogs allowed the visualization of membrane associated perturbations under fluorescence microscope. As will be later detailed in this chapter, lipid loss was observed in electropermeabilized membranes and was associated with pores, vesicles, and tubules formation, structures that can be involved in transmembrane transport of molecules.

Electroformation of GUV Electroformation is probably the most common method for vesicle production. As already mentioned in the introduction, the process was first reported in 1986 (Angelova and Dimitrov 1986). Later refinements of the method have allowed for unilamellar vesicles to be consistently produced (Angelova et al. 1999). The process of electroformation involves different steps including (Yarmush et al. 2014) the spreading of phospholipids on the surface of a conductive electrode, (Kotnik et al. 2015) the evaporation of the solvent to form a dry phospholipid film, (Rosazza et al. 2016) the immersing of the conductive electrode in an aqueous solution, and finally (Angelova and Dimitrov 1986) the application of Alternative Current (AC) electric pulses across the lipid film (Fig. 1). By this way, it is possible to easily obtained GUV which sizes wary from a few μm to tens of μm. Electroformation devices can be purchased from companies, but it is possible to build its one setup that is an easy and cheap way to obtain GUV. Electroformation chamber can be made by two glass layers covered with indium tin oxide to ensure the electrical conductivity of the surface (Fig. 2). The two layers are separated by an adhesive silicone joint of 1 mm width. The connection with an AC generator is maintained by two wires, each one soldered on a small copper strip stuck on the ITO slide. Phospholipids are diluted in chloroform, at a mass concentration of 0.5 mg/mL. Fluorescent probe can be eventually added at 0.005 mg/mL in order to help GUVs visualization under fluorescent microscopes. 15 μL of lipid solution is deposited on the conducting sides of the glass slides. The deposition is carried out slowly and at constant rate in a chamber held at 4
C to

Molecular Transmembrane Transport with Giant Unilamellar Vesicles (GUVs)

5

Fig. 1: Electroformation of GUV. Left: Schematic representation of the process; Right: GUV observed under a microscope Fig. 2 Experimental homemade setups for GUV formation and electropulsation. Left: electroformation chamber; Right: electropulsation device

slowly evaporate the chloroform and then the slides are dried under vacuum for a couple of hours to entirely remove the remaining solvent molecules. Finally, the slides are sealed together, and the chamber is filled with the formation medium which is a 240 mM sucrose solution in water. Sinusoidal voltage of 25 mV peak to peak is then applied at 8 Hz. The voltage is increased by 100 mV steps every 5 min, up to a value of 1225 mV. It is maintained under these conditions overnight. Next, a square wave of same amplitude at 4 Hz has to be applied for 1 h to detach the liposomes from the slides. The lipids indeed interact with the aqueous solution and electric field by “peeling off” the electrode surface in layers and self-assembling into giant, sometimes multilamellar, vesicles. One drawback of the first implementations of the electroformation method was the requirement for low salt concentrations in the buffer. Since most proteins require high salt concentrations in order to function properly, this made it very difficult to encapsulate active proteins using electroformation. However, recent refinements to the technique have overcome this limitation and produced liposomes by electroformation using physiologically relevant salt concentrations. Bagatolli and collaborators reported that GUVs can be prepared using a protocol based on the electroformation method from either native membranes or organic lipid mixtures at physiological ionic strength (Montes et al. 2007). They showed that membrane proteins and glycosphingolipids preserve their natural orientation after electroformation.

6

M.-P. Rols

One has to add that in the last ten years there have been many new developments in liposome formation technology. Most notably, macroscale methods have been transposed to the microscale while entirely new methods that are only possible with microfluidic technologies were also developed. A recent review by van Swaay and deMello provides a critical comparison of existing microfluidic technologies for forming liposomes (van Swaay and deMello 2013). The properties of the generated liposomes, including size, size distribution, lamellarity, membrane composition, and encapsulation efficiency, form the basis for comparison. According to their qualities and drawbacks, these different methods can be useful in different applications.

Electropulsation of GUV Once obtained by electroformation, GUV can then be easily submitted to electric field pulses by homemade pulsation chamber (Portet et al. 2009). Homemade chamber can be composed as shown in Fig. 2 of a glass slide and a coverslip. In the present device, two parallel copper strips of thickness 70 μm are stuck on the slide at a distance of 1-cm apart. The coverslip is then stuck onto the slide and strips with heated parafilm. The chamber is 1-cm long (between electrodes), 2.6-cm wide (width of the coverslip), and 250-μm high (value estimated via measurements with a microscope). 60 μL of pulsation buffer are introduced between the slide and the coverslip, while taking care of filling the whole chamber to ensure the conductivity of the medium. Next, 5 μL of GUV preparation are added. Capillarity phenomena prevent the solution from leaking out of the chamber. The electrode thickness is about the size of the biggest liposomes, which represents only a quarter of the chamber height. It is not a priori possible to be certain of the homogeneity of the field. However, solving numerically Laplace’s equation with finite element software for such a geometry reveals that the field is almost homogeneous in the bottom part of the chamber between the electrodes, and that the size and shape of the permeabilized area are not significantly different from that computed for a geometry with much bigger electrodes. Electropulsation can be carried out using different kind of electropulsator which delivered square-wave electric pulses. An oscilloscope has to be used to monitor the pulse shape and amplitude. The process of electropulsation can then be performed directly under the microscope. For the phase contrast visualization, one can use an inverted epifluorescence microscope equipped with a camera and phase contrast objective. Visualization of the liposomes is possible thanks to the resuspension of the GUV in a 260 mM glucose containing solution. For the fluorescence visualization, one can use an inverted confocal microscope with an objective for fluorescence imaging. The pulse duration has to be set from a few hundreds of microseconds to several ms, according to the effects that have to be studied. Short pulses are used to simply permeabilize the cell membranes and allow the transfer of small molecule such as cytotoxic drugs. Long pulses, at the ms time range of order, are commonly used for gene transfer protocols in mammalian cells.

Molecular Transmembrane Transport with Giant Unilamellar Vesicles (GUVs)

7

Fig. 3 Example of GUV submitted to electric pulses. Left: control, before EP; Right: 10 s after delivery electric pulses at two electric field intensities E1 and E2 (E2 > E1). Scale bar is 10 μm

It is sometimes necessary to interrupt the pulse train for a few seconds to recenter the image on the liposome of interest. Indeed, the observed vesicle does not always stay immobile. It often experiences a translational motion toward the positive electrode. As shown in Fig. 3, direct observation of GUV submitted to electric pulses can be performed under an inverted microscope, allowing an easy way to visualize on real time the process of electropermeabilization. The different refractive indexes of the internal and external media yield a contrast which enables the vesicles to be visualized using a microscope, and the density difference allows the sedimentation of the vesicles on the bottom of the chamber, thus reducing their distance from the objective. Giant liposomes have different sizes ranging from a few μm to several tens of μm. Such a heterogeneity is however very useful to study electropermeabilization and in fact reflects the heterogeneity of cells present in tissues. Electropermeabilization is, in this example where GUV are formed in a sucrose containing solution and pulsed in the presence of glucose, quite easy to assess. Any decrease of phase contrast indeed reflects molecule exchange between the interior and the exterior of the vesicles. It is well known, for many years, that the trigger of membrane electropermeabilization is the transmembrane potential which must reach critical values around 200 mV. The application of the electric field indeed superimposes an electroinduced transmembrane potential, Δψi, to the resting transmembrane potential Δψo. The value of Δψi is given by the Schwan equation: Δψi ¼ f  gðλÞ  r  E  cos ðθÞ,

(1)

where f is a factor depending on the biological object shape, g(λ) a parameter depending on the conductivities λ of the membrane, of the external medium, and of the internal one, r is the radius of the biological object, E the electric field strength, and θ the angle between the electric field direction and the normal to the membrane at the considered point of the surface. Being dependent on the angle θ, the electric field effect is not uniform along the membrane and its maximum effects are present at the poles facing the electrodes.

8

M.-P. Rols

The basics for electropermeabilization imply that the transmembrane potential must reach values closed to 200 mV. According to Eq. 1, electric field strength values E higher than a threshold value Ec must be applied to induce membrane permeabilization. In the case of two vesicles 1 and 2 with radii r1 and r2, respectively, permeabilization should be induced for a constant Er value. This implies that if r1 > r2, then E1 < E2. As shown in Fig. 3, this is clearly the case. Large vesicles can be permeabilized by electric field values which do not affect the smaller vesicles. There is a direct linear relation between Ec and vesicles radius (Fig. 3, right). Moreover, the larger the vesicles are, the lower the electric field threshold value is. By applying another train of pulses at a higher electric fields strength (E2 > E2), it becomes then possible to permeabilize GUV with lower sizes that were not previously affected by applying the train of pulses (E = E1). Such experimental results are in complete agreement with the above theory of electroporation and gives a piece of evidence that GUV are a suitable model to assess electropermeabilization related process. One however has to notice that the determination of the threshold values Ec by the method used here, contrast decrease, is not a very sensitive one. Other works indeed show that the detection method is a very important parameter for the precise determination of the critical transmembrane potential. Fluorescence is more sensitive. The use of quenching ions as probes for lipid bilayer permeabilization allows for very high resolution detection because of their small size which implies the need for a lower destabilization threshold than the one needed with larger molecules such as fluorescent dyes or sugar (Mauroy et al. 2015).

Effects of Electric Fields on GUV Both DC pulses and AC fields can strongly destabilize giant vesicles and influence their behavior. As shown in Fig. 4, depending on the parameters of the applied field, giant vesicles can deform, porate, fuse, or even exhibit domain motion in the case of multicomponent liposomes (Portet et al. 2012). Deformation of giant vesicles by electric fields was first extensively studied in the early nineties showing that spherical liposomes deformed into prolate ellipsoids oriented in the field direction, the degree of deformation increasing with the magnitude of the applied field. However, DC pulses usually have durations of the order of several tens or hundreds of microseconds, or at most of some milliseconds. Imaging at such high rates is impossible with classical video cameras, so if one wants to get insight at what happens during an electric pulse application, one has to work with a fast imaging setup allowing image acquisition at several thousand frames per second. This approach was first applied by Riske and Dimova, and they indeed found that GUVs deformed into prolate ellipsoids upon the application of electric pulses of 50–300 μs duration and 1–3 kV/cm amplitude (Riske and Dimova 2005). Later, the same authors worked in conditions closer to physiological ones and found that vesicles subjected to electric pulses in salt solutions always adopted cylindrical shapes, irrespective of their ionic content (Riske and Dimova 2006).

Molecular Transmembrane Transport with Giant Unilamellar Vesicles (GUVs)

9

AC field Electrodeformation

Motion of domains

GUV

DC pulse

Electrodeformation

Size decrease

Electrofusion

Electroporation

Fig. 4 The influence of electric fields on giant lipid vesicles. GUV submitted to AC or DC pulses can be deformed; DC pulses can induce permeabilization (which eventually leads to a decrease in size and an exchange of molecules between interior and exterior of the vesicles) and fusion in the case of GUV in contact (Adapted with permission from Portet et al. (2012))

Electric pulses may also have much more dramatic effects on lipid bilayers than simple deformation. They can sufficiently disrupt the membrane as to allow the uptake of nonpermeant molecules. No consensus exists on the way the membrane reorganizes at the molecular level and this question still remains a challenge. Here again, GUVs are convenient tools to investigate these questions. Some studies have been performed, and have shown that giant vesicles can indeed become permeabilized by DC electric pulses. It was shown by Tekle et al that the mechanism of pore formation is asymmetric (Tekle et al. 2001). A macropore is formed on the cathode-facing side of the vesicle, whereas the presence of many smaller pores on the other hemisphere can be inferred from the size decrease of the liposome following pulse application. This phenomenon of GUVs shrinkage during electropulsation was then further extensively studied (Portet et al. 2009). By applying a sequence of long (5 ms) electric pulses, it was found that vesicles shrank, down to a critical radius beyond which their size no longer changed. Three mechanisms for the lipid loss were identified: formation of macropores on the cathode-facing side as already reported, formation of tubular structures on the anode facing side, and formation of small vesicles at both poles. These three features should probably not be considered as distinct mechanisms. Macropore and tubules formation probably reflect the same phenomena, as they were found to occur together. Small vesicles expulsion can be understood as another way of expelling lipids, and may under certain conditions be more energetically favorable than tubule formation.

10

M.-P. Rols

Increasing the membrane complexity by addition of different phospholipids and cholesterol leads to phenomena that cannot be observed in single component GUVs. Staykova et al managed to monitor the motion of liquid ordered domains in giant vesicles made of DOPC/DPPC/Cholesterol (different molar ratios were explored) subjected to AC electric fields of 500 V/cm amplitude in the kHz frequency range (Staykova et al. 2008). They report that this movement had characteristic features depending on the field parameters, and that it was caused by the inhomogeneous surface tension induced by the field because of the chamber geometry. It was the first time this phenomenon of induced charge electro-osmosis was observed and studied on a lipid membrane. Membrane fusion is a key process of life. However, membrane fusion does not occur spontaneously because of large energetic barriers in biological membranes. To promote membrane fusion, it is essential to overcome these barriers. Electrofusion is a very convenient way to control, both spatially and temporally, fusion events. It is thus possible to trigger and observe the whole fusion process to study the underlying mechanisms. To occur, electrofusion requires two conditions: (a) electropermeabilization and (b) contact between lipid membranes. When the two membranes are close enough and lipid perturbation is high enough, fusion occurs; for cells. Electropermeabilized membranes are thus fusogenic. For a better understanding of the fundamental processes involved in membrane fusion, lipid vesicles are often used as model systems. An AC field can be used to align vesicles in the field direction and bring two vesicles into contact. A subsequent application of a DC pulse induces the permeabilization of the two vesicles and if permeabilization is induced in the contact area fusion is induced (Haluska et al. 2006).

Different Approaches to Assayed Molecular Transport Quantitative Optical Microscopy and Micromanipulation Studies Very few publications have been reported merging micropipet manipulation technique with fluorescence microscopy experiments, and so the combination of these two techniques still remained to be fully exploited. The possibility to simultaneously perform both types of experiments in the same microscope offers a powerful tool to relate supramolecular membrane events at different lengths and time in the case of lipid vesicles. Because their size that is well above the resolution limit of regular light microscopes, GUVs are indeed suitable membrane models both for optical microscopy and micromanipulation experimentation. In their recent review, Bagatolli and Needham discuss the basic methodological aspects of optical microscopy and micromanipulation methods to study membranes (Bagatolli and Needham 2014). They focus on the use of fluorescence microscopy and micropipet manipulation techniques to study composition–structure–property materials relationships of free-standing lipid bilayer membranes. For instance, using different fluorescent reporters, fluorescence microscopy allows strategies to study membrane lateral structure/dynamics at the level of single vesicles of diverse

Molecular Transmembrane Transport with Giant Unilamellar Vesicles (GUVs)

11

compositions. The micropipet manipulation technique allows studies on the mechanical, thermal, molecular exchange and adhesive-interactive properties of compositionally different membranes under controlled environmental conditions. By utilization both techniques simultaneously on the same vesicle, they bring the ability to characterize structure and strain responses together with the direct application of well-defined stresses to a single membrane. Knowledge acquired by these studies has informed several applications of lipid membranes including their use as drug delivery systems.

Visualization and Quantification of Transmembrane Ion Transport Transmembrane ion transporters are widely investigated as supramolecular agents with potential for biological activity. Tests are usually performed in large unilamellar vesicles (LUVs). However, transport must be followed through bulk properties of the vesicle suspension, because LUVs are too small for individual study. Ion transport can be revealed and quantified through direct observation employing GUVs. This allows characterization of individual GUVs containing transporter molecules, followed by studies of transport through fluorescence emission from encapsulated indicators. In their publication, Valkenier and collaborators have devised a new method whereby ion transport by small molecules into individual giant unilamellar vesicles can be observed and quantified (Valkenier et al. 2015). By directly visualizing transport into GUVs, their approach offers a high level of certainty and integrity compared to experiments on bulk suspensions of smaller vesicles. Instead of quantifying transport into a population of vesicles with a distribution of sizes, they are able to analyze the transport into individual GUVs. The method provides new levels of certainty and relevance, given that the GUVs are similar in size to living cells. Indeed, positive results in this test provide clear encouragement that a transporter has potential for biological activity. It has been demonstrated using a highly active anion carrier, and should aid the development of compounds for treating channelopathies such as cystic fibrosis.

Monitoring and Quantifying the Passive Transport of Molecules Through Patch-Clamp Transport of active molecules across biological membranes is a central issue for the success of many pharmaceutical strategies. In their work, Amatore and collaborators combined the patch-clamp principle with amperometric detection for monitoring fluxes of redox-tagged molecular species across a suspended membrane patched from single-wall GUV and the corresponding fluxes measurements quantified (Messina et al. 2014). The quality of the patches and their proper sealing were successfully characterized by electrochemical impedance spectroscopy. Cholesterol was added to improve the patches stability This procedure appears versatile and perfectly adequate to allow the investigation of transport and quantification of the

12

M.-P. Rols

transport properties through direct measurement of the coefficients of partition and diffusion of the compound in the membrane, thus offering insight on such important biological and pharmacological issues.

Imaging Molecular Transport Across Lipid Bilayers In their publication, Li and collaborators used confocal microscopy to image the transport of carboxylic acids with different lengths of carbon chains into a single GUV (Li et al. 2011). Fluorescein-dextran, which acts as a pH-sensitive dye, was encapsulated in the GUV to trace the transport of acid. The GUV was immobilized on the surface of a microfluidic channel by biotin-avidin binding. This microchannel allows the rapid and uniform exchange of the solution surrounding the GUV. Using a spinning-disk confocal microscope, the entire concentration field is captured in a very short exposure. This technique combines rapid buffer exchange in a microfluidic device with high-speed confocal microscopy to accurately observe rapid transmembrane transport processes. Results showed that more lipophilic acids cross the bilayer more quickly. A finite difference model was developed to simulate the experimental process and derive permeability. This technique promises to be useful in understanding with greater detail how the molecular properties of drug like molecules determine their transport behavior into cells.

Imaging Electropermeabilization Process by CARS Spectroscopy In the high majority of studies, visualization of electropermeabilization is performed by using fluorescent molecules (lipid analogue, fluorescent dye) and not pure lipid systems thus including possible artefacts linked to these molecules. Fluorescence microscopy, while extremely sensitive and widely used, indeed requires the introduction of extrinsic fluorophores but often causes unwanted perturbations. In their paper, Mauroy and collaborators performed experiments on pure lipid GUVs, i.e. without any addition of exogenous molecules (lipid analogue, fluorescent dye) and used two different methods, phase contrast and CARS microscopies (Mauroy et al. 2012). Optical microscopy is unique in its ability to probe living specimens with subcellular resolution. Strong vibrational signals can be obtained with coherent anti-Stokes Raman scattering (CARS) microscopy, a nonlinear Raman technique. GUVs were prepared in a medium containing sucrose and diluted in a medium containing glucose. The difference in refractivity between the internal and the external media was enhanced by phase-contrast optics. CARS has the advantage of offering molecular specificity, without necessitating the application of external labels. As reported by the group of Kinosita, GUVs were prepared in an internal medium containing sucrose and diluted in the external medium containing glucose. The insides of the liposomes looked darker than the outsides because the difference in refractivity between the internal and the external media was enhanced by the phase-contrast optics (Akashi et al. 1996). They present for the first time results of

Molecular Transmembrane Transport with Giant Unilamellar Vesicles (GUVs)

13

electropermeabilization of pure lipid systems obtained by CARS microscopy. Their results indicate that tubule formation indeed is not an artifact due to fluorescence labeling. Similar results were obtained for both fluorescence and CARS microscopy. The results underline the high potential for use of CARS microscopy as a tool to study electropermeabilization processes in living systems.

Imaging Lipid Loss Associated to Electropermeabilization In their work, Portet and collaborators studied the effect of permeabilizing electric fields applied to two different types of giant unilamellar vesicles (Portet et al. 2009). Experiments on vesicles show a decrease in vesicle radius, which is interpreted as being due to lipid loss during the permeabilization process. They show that the decrease in size can be qualitatively explained as a loss of lipid area, which is proportional to the area of the vesicle that is permeabilized. The phenomenon of electropermeabilization can be defined as a two-step process: (1) a physical change induced in the membrane by the field (in the absence of molecules to be transported) and (2) an interaction of the molecules that are to be transported with the modified membrane. At the simplest level, combination of Steps 1 and 2 can be observed experimentally as a transport phenomenon using marked molecules or via conductivity experiments. In their article, they demonstrate that Step 1 can be indirectly detected via a change in the size of giant liposomes under electropulsation and an associated direct visualization of the expulsion of lipids from the liposomes. Concretely they study the effect of a series of permeabilizing pulses, well separated in time, on the size of GUV. In the experiments, the radius of the GUV is measured after each pulse and each GUV studied shows, on average, a decrease in its radius down to a critical radius beyond which its size no longer changes. This decrease in size points to the fact that, during the physical processes leading to electropermeabilization, lipids are lost from the vesicle—thus leading to a reduction in their size. One of the most fascinating aspects of the experiments is the wide variety of mechanisms of lipid loss that can be observed: pore formation, vesicle formation, and tubule formation (Fig. 5). Three different mechanisms of lipid loss are observed when the lipids are fluorescently marked, the term “lipid loss” implies loss of lipid from the bulk spherical part of the vesicle; the lipid ejected appears, in most cases, to remain attached to or close to the parent vesicle. The first and most frequent mechanism is the formation of small vesicles at both the anode-and cathode-facing poles. Those vesicles are mainly thrown out of the GUVs, but some of them were also driven inside the GUVs. The second phenomenon was the creation of lipid tubules on the exterior of the anode-facing hemisphere. DOPC molecules expelled from the membrane rearranged in the form of tubular structures, whose lengths grew with the number of applied pulses. These structures initiated from the pole facing the positive electrode and remained attached to the vesicle. However, they then appeared to diffuse away from the pole toward the equator (while remaining attached to the membrane) and appeared to cover most of the anode-facing hemisphere. They also

14

M.-P. Rols

Fig. 5 The effects of electric fields on giant lipid vesicles. GUV submitted to DC pulses can be permeabilized leading to the formation of vesicles and tubules which number and or size increase with the number of pulses. Macropores can also be observed (Adapted with permission from Portet et al. (2009))

saw on the cathode-facing side that tubules can grow on the interior surface of the liposome. These structures also diffuse toward the equatorial regions, the number and size of tubules, however, being smaller. This mechanism of tubule formation appears to be stronger on the anode-facing hemisphere. Finally, they also noticed the presence of pores on the cathode-facing hemisphere. This was a quite rare observation, but it is normal because acquisition times were of a few hundreds of milliseconds, the same order of magnitude as the lifetimes of such pores. The eventual long-term evolution of the structures described above (after pulsation has been stopped) varied from one experiment to another. The small vesicles, in most cases, diffused away from the liposome and the vesicle radius stayed constant.

Molecular Transmembrane Transport with Giant Unilamellar Vesicles (GUVs)

15

However, the behavior of the tubular structures exhibited wide variation. Some of the tubules broke away from the GUV and diffused away, sometimes forming vesicles and sometimes not. Other tubules remained attached to the vesicles, exhibiting polymerlike fluctuations. In some cases, they were reabsorbed into the GUV membrane after a time of approximately minutes. In fact, the eventual fate of tubules was strongly dependent on their environment, notably on whether other vesicles came in contact with them or not. In the cases where tubules were reabsorbed, the volume of the vesicle they were attached to increased, and the final state of the vesicle was often nonspherical, and appeared to be under little tension. Such transient membrane alterations may be involved in the transport of molecules across membranes and explain the different pathways molecules can enter electropermeabilized cells. However, experimental results regarding the electrotransfer of plasmid DNA into phosphatidylcholine GUVs indicate that a direct entry is the predominant mechanism of electrotransfer (Portet et al. 2011). This is not the case for cells, where DNA electrotransfer is a multi-step process including a key step of interaction between the plasmid and the electropermeabilized membrane. A quantitative analysis of the DNA concentration increments inside the GUVs can be very well described by a simple theoretical model in which DNA entry is mostly driven by electrophoresis. Such results pave the way towards a novel method for encapsulating with high efficiency not only DNA, but any negatively charged macromolecules into GUVs. They however show the limits of GUV as a model.

Conclusions This chapter has present main data reporting that giant vesicles can be considered as relevant biological models, for resolving the effect of electric fields on cells. GUV have been shown as a powerful model, easy to form and observe under a microscope, that fairly reproduce electropermeabilization associated phenomena such as deformation, fusion, molecule transport. As pure lipid membranes, they however present some limits people have to know about. According to the dictionary, a model is a person or thing that serves as a subject for an artist, sculptor, writer. For sciences, it is in only fact a simplified representation of a system or phenomenon. Richard Feynmann, Nobel laureate in Physics, famously wrote: “what I cannot create, I do not understand”. Biomimetic approach bridges the gap between the simplest GUV consisting only of a spherical lipid bilayer enclosing a buffer, and more complex GUVs made from lipid mixtures, carrying adhesion proteins and/or filled with artificial cytoskeleton providing more advanced test cells. A multitude of different modifications to the basic GUV are now available, each capturing the essence of a different aspect of the cell (Fenz and Sengupta 2012). Such more realistic GUV are of great interest to go deeper in the elucidation of the mechanisms underlying transport of molecules in cells and tissues.

16

M.-P. Rols

Cross-References ▶ Critical Electric Field and Transmembrane Voltage for Lipid Pore Formation in Experiments ▶ Electro-Deformation, -Poration and -Fusion of Giant Unilamellar Vesicles ▶ Electroporation and Electropermeabilization ▶ Experimental Determination of Lipid Electropore Size ▶ Measurement of Molecular Transport into Electropermeabilized Cells ▶ Molecular Models of Lipid Bilayers and Electropore Formation ▶ Nucleic acid electrotransfer in mammalian cells: mechanistic description ▶ Transmembrane Voltage Induced by Applied Electric Fields

References Akashi K, Miyata H, Itoh H, Kinosita K Jr (1996) Preparation of giant liposomes in physiological conditions and their characterization under an optical microscope. Biophys J 71:3242–3250 Angelova MI, Dimitrov DS (1986) Liposome electroformation. Faraday Discuss Chem Soc 81:303–311 Angelova MI, Hristova N, Tsoneva I (1999) DNA-induced endocytosis upon local microinjection to giant unilamellar cationic vesicles. Eur Biophys J 28:142–150 Bagatolli LA, Needham D (2014) Quantitative optical microscopy and micromanipulation studies on the lipid bilayer membranes of giant unilamellar vesicles. Chem Phys Lipids 181:99–120 Dimova R, Aranda S, Bezlyepkina N, Nikolov V, Riske KA, Lipowsky R (2006) A practical guide to giant vesicles. Probing the membrane nanoregime via optical microscopy. J Phys Condens Matter 18:S1151–S1176 Fenz SF, Sengupta K (2012) Giant vesicles as cell models. Integr Biol 4:982–995 Haluska CK, Riske KA, Marchi-Artzner V, Lehn JM, Lipowsky R, Dimova R (2006) Time scales of membrane fusion revealed by direct imaging of vesicle fusion with high temporal resolution. Proc Natl Acad Sci U S A 103:15841–15846 Kotnik T, Frey W, Sack M, Haberl Meglic S, Peterka M, Miklavcic D (2015) Electroporation-based applications in biotechnology. Trends Biotechnol 33:480–488 Li S, Hu PC, Malmstadt N (2011) Imaging molecular transport across lipid bilayers. Biophys J 101:700–708 Mauroy C, Portet T, Winterhalder M, Bellard E, Blache MC, Teissie J, Zumbusch A, Rols MP (2012) Giant lipid vesicles under electric field pulses assessed by non invasive imaging. Bioelectrochemistry 87:253–259 Mauroy C, Rico-Lattes I, Teissie J, Rols MP (2015) Electric destabilization of supramolecular lipid vesicles subjected to fast electric pulses. Langmuir 31:12215–12222 Messina P, Lemaitre F, Huet F, Ngo KA, Vivier V, Labbe E, Buriez O, Amatore C (2014) Monitoring and quantifying the passive transport of molecules through patch-clamp suspended real and model cell membranes. Angew Chem Int Ed Engl 53:3192–3196 Montes LR, Alonso A, Goni FM, Bagatolli LA (2007) Giant unilamellar vesicles electroformed from native membranes and organic lipid mixtures under physiological conditions. Biophys J 93:3548–3554 Portet T, Febrer FCI, Escoffre JM, Favard C, Rols MP, Dean DS (2009) Visualization of membrane loss during the shrinkage of giant vesicles under electropulsation. Biophy J 96:4109–4121 Portet T, Favard C, Teissie J, Dean DS, Rols MP (2011) Insights into the mechanisms of electromediated gene delivery and application to the loading of giant vesicles with negatively charged macromolecules. Soft Matter 7:3872–3881

Molecular Transmembrane Transport with Giant Unilamellar Vesicles (GUVs)

17

Portet T, Mauroy C, Demery V, Houles T, Escoffre JM, Dean DS, Rols MP (2012) Destabilizing giant vesicles with electric fields: an overview of current applications. J Membr Biol 245:555–564 Riske KA, Dimova R (2005) Electro-deformation and poration of giant vesicles viewed with high temporal resolution. Biophys J 88:1143–1155 Riske KA, Dimova R (2006) Electric pulses induce cylindrical deformations on giant vesicles in salt solutions. Biophys J 91:1778–1786 Rodriguez N, Pincet F, Cribier S (2005) Giant vesicles formed by gentle hydration and electroformation: a comparison by fluorescence microscopy. Colloids Surf B Biointerfaces 42:125–130 Rosazza C, Meglic SH, Zumbusch A, Rols MP, Miklavcic D (2016) Gene electrotransfer: a mechanistic perspective. Curr Gene Ther 16:98–129 Staykova M, Lipowsky R, Dimova R (2008) Membrane flow patterns in multicomponent giant vesicles induced by alternating electric fieldsdaggerElectronic supplementary information (ESI) available: Vesicle preparation procedure, numerical calculations and confocal microscopy movies of domain motion. Soft Matter 4:2168–2171. doi:10.1039/b811876k Tekle E, Astumian RD, Friauf WA, Chock PB (2001) Asymmetric pore distribution and loss of membrane lipid in electroporated DOPC vesicles. Biophys J 81:960–968 Valkenier H, Lopez Mora N, Kros A, Davis AP (2015) Visualization and quantification of transmembrane ion transport into giant unilamellar vesicles. Angew Chem Int Ed Engl 54:2137–2141 van Swaay D, deMello A (2013) Microfluidic methods for forming liposomes. Lab Chip 13:752–767 Yarmush ML, Golberg A, Sersa G, Kotnik T, Miklavcic D (2014) Electroporation-based technologies for medicine: principles, applications, and challenges. Annu Rev Biomed Eng 16:295–320

Lipid Electropore Lifetime in Molecular Models Zachary A. Levine

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Electropore Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Molecular Dynamics Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Electropore Lifetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Abstract

Membrane electropermeabilization describes the electric field-mediated depolarization, and subsequent breakdown of cellular membranes, and is widely used in clinical and academic environments to deliver extracellular materials into the cell interior. Recently, these methods have contributed to the optimization of food sterilization and next-generation therapeutics that aim to enhance the susceptibility of tumor cells to traditional chemotherapies, signaling a significant shift in biomedical and medical modalities. However, the relationship between macroscale membrane electropermeabilization and the influx of individual materials through discrete electropores (often termed electroporation) is often unclear, at best. Because detection and characterization of discrete electropores in experiments containing cells or vesicles are difficult, if not impossible, this section will describe discrete electroporation models based on theoretical molecular dynamics simulations, with the intent of reconciling theoretical models with observations Z.A. Levine (*) Department of Physics, University of California, Santa Barbara, Santa Barbara, CA, USA Department of Chemistry and Biochemistry, University of California Santa Barbara, Santa Barbara, CA, USA Materials Research Laboratory, University of California Santa Barbara, Santa Barbara, CA, USA e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_86-1

1

2

Z.A. Levine

from macroscopic experiments. Access to massively parallel supercomputing resources have greatly enhanced the timescales and typical sample sizes modeled in simulations, and these trends are expected to grow as computing hardware and software become more integrated and affordable. Therefore, discussion will focus primarily on all-atom molecular models rather than larger, coarse-grained models that lack contributions from individual water molecules. Taken together, this section will broadly cover studies that aim to codify the discrete biophysical processes that mediate membrane electropermeabilization, in an effort to better understand its biological basis. Keywords

Electroporation • Electropermeabilization • Electroevaporation • Computer simulations • Electropore life cycle • Phospholipid bilayers

Introduction The mechanical structure of cellular membranes, which compartmentalizes and partitions intra- and extracellular materials from one another, can be transiently or irreversibly permeabilized with the application of sufficiently high electric fields (Hamilton and Sale 1967). Such electric fields, typically of MV/m amplitude and nano- to microsecond duration, increase the electrical conductance of plasma membranes and allow influx and efflux of small, normally impermeant molecules across cellular barriers (Neumann et al. 1982). These changes in conductance, while small, can be observed after only a few nanoseconds following the delivery of pulsed electric fields (Benz and Zimmermann 1980). Similarly, small nanometer-sized fluorescent dyes can be utilized as molecular markers to measure the amount of membrane electropermeabilization and disruption, but how and where small molecules enter membranes and what atomic factors mediate their interaction with membranes under high electric fields is poorly understood and difficult to observe directly with microscopy (Teissie et al. 2005). As a result, theoretical methods including atomistic simulations are currently better suited for characterizing properties of discrete electropores at the atomic level, where Coulombic and Van der Waals forces dominate. Structurally, membranes can consist of hundreds of distinct lipid species, transmembrane proteins, cytoskeletal scaffolds, and a virtual zoo of interacting molecules; however, the most dominant structural unit of cellular membranes is phospholipids (Fig. 1). Phospholipids are amphiphilic molecules containing a dipolar headgroup, two aliphatic (fatty acid-derived) tails, and a glycerol backbone linking the polar headgroup to the tails. The two tails are attached to the C1 and C2 glycerol carbon atoms, thereby denoting the tails as sn-1 and sn-2, respectively. Depending on the fatty acid chain length, the length of lipid tails can vary considerably from 12–22 carbon atoms and typically contain a handful of unsaturated carbon atoms. Lipid names are then constructed by combining, in

Lipid Electropore Lifetime in Molecular Models

3

Fig. 1 Diagram of a typical phospholipid from molecular simulations

order, the sn-1, sn-2, backbone, and headgroup names and are often referred to by acronyms containing the first letters of each group. In this section, a summary of electroporation models in homogeneous phospholipid membranes (i.e., membranes containing only one lipid type) will be presented; however, in the next section (titled “Effects of Heterogeneous Membranes and Electrolytes on Electropore Formation”), additional discussion will be given to membranes containing multiple lipid types or electrolytes, which exhibit somewhat different permeabilizing behaviors from homogeneous membranes. The mechanical and electrical properties of simple lipid bilayers closely resemble the properties of real cell membranes. Lipid bilayers in cellular membranes typically span 4–5 nm in thickness and hold a resting potential of roughly 70 mV. This is due to a number of factors, such as surface or zeta potentials which stem from differences between bulk electrolyte and the membrane surface charge, dipolar potentials which inhabit the lipid headgroup region, and other various transmembrane potentials based on differences in electrolyte concentration from one side of the bilayer to the other (Fig. 2). This implies that a single charge would only need 0.07 eV of energy to cross a lipid bilayer; however, this is not the case. The lipid headgroup

4

Z.A. Levine

Fig. 2 Electrical properties of a phospholipid bilayer. ΔΨm, ΔΨs, and ΔΨd stand for the transmembrane, surface, and membrane dipole potentials, respectively, whereas Εm, Εs, and Εd stand for the corresponding electric field values

region contains a very large 280 mV potential difference across it, although that potential is often neutralized because of the equal and opposite headgroup dipoles located on opposing leaflets. A single charge actually requires about 0.3 eVof energy just to cross the headgroup region (consider a particle with charge 1e traveling 1 nm against a ~300 MV/m electric field). However, a single charge needs only to traverse halfway across the bilayer to make it through, because the hydrophobic lipid interior will push the particle outward with little to no additional energy. For symmetric bilayers one can imagine the headgroup polarization spanning, at most, 2.5 nm (halfway across the bilayer) in which case the translocation energy approaches 1 eV, though this can vary if there are asymmetries present.

Electropore Models Early efforts to model membrane reorganization under the influence of electric fields began in the 1950s and were rooted in cellular experiments where conductive and dielectric properties of the membrane were modulated as a function of the applied electric field. During this time, researchers discovered that the electrical impedance of membranes would significantly drop when large voltages generated from ion gradients across the membrane were applied. Later in the 1970s, Neumann and Rosenheck (1972) demonstrated that both membrane permeability and conductance could be reversibly increased if the applied electric field was sufficiently short. Transient electrical perturbations have since been termed “reversible

Lipid Electropore Lifetime in Molecular Models

5

electroporation” because increases in membrane permeability and electrical conductance can be completely reversible if the applied field is removed after a sufficiently short period of time. This is in contrast to “irreversible electroporation,” which describes plasma membranes that are permanently disrupted from prolonged exposure to electric fields, leading to eventual cell death or necrosis. Irreversible electropermeabilization of membranes was initially described by Hamilton and Sale (1967) in a triple set of detailed publications that describe the nonthermal destruction of bacteria, yeast, and erythrocytes under kV/cm electric field amplitudes. Membrane lysis was found to be present at externally applied transmembrane voltages of 0.30–1.15 volts. Reversible electropermeabilization of cellular membranes, however, appeared to be more closely correlated with earlier studies on changes in membrane conductance (Abidor et al. 1979). Because reversible electroporation experiments were reversible, and generally nonlethal to cells, it became an attractive system for further study and subsequently laid the groundwork for modern continuum models to build upon. Observing electroporation at the nanoscale is still not possible experimentally, but a plethora of analytical models (DeBruin and Krassowska 1999; Gowrishankar and Weaver 2003) and indirect experimental observations (Gabriel and Teissie 1997) have been proposed in an attempt to explain the molecular mechanisms of electroporation using continuum electrodynamic equations. The simplest continuum models interpret biomembranes as smooth elastic surfaces, where potentials at the surface can be extracted using finite element models and Poisson’s equation. Electric double-layer models that incorporate adsorbed interfacial electrolytes (built up at the water-lipid interface) include the Gouy-Chapman-Stern model; however, individual water molecules and phospholipid headgroups in these models are coarsely defined and do not incorporate individual effects of water hydration or steric repulsion between individual lipids. Analytical electroporation models can be subdivided into two broad categories that include (Tokman et al. 2013) “surface potential models,” or models which calculate static potential profiles for lipid bilayers and combine them with stability conditions (e.g., setting the surface tension equal to zero), and (Ziegler and Vernier 2008) “stochastic pore models,” which claim that there exists a complex distribution of transient pores in membranes upon the application of external electric fields, and that these distributions depend on macroscopic electrical properties such as resistance and capacitance. Chizmadzhev and colleagues pioneered models of the first type (Abidor et al. 1979; Chizmadzhev and Abidor 1980) using circuit analogs to analyze the electrical permeation of lipid bilayers (e.g., that the pore energy is proportional to the edge energy minus the surface energy minus some function of membrane capacitance, resistance, and other electrical properties). Conversely, Sugar and Neumann developed models of the second type (Sugar 1981) using bilayer transition probabilities. Modern models use combinations of these such as the asymptotic model of electroporation (Neu and Krassowska 1999) that incorporates both the electrical properties of the membrane and probabilistic estimates for the number of open pores (N) per unit time (Eq. 1). Here, α and β are normalization constants and Neq is the equilibrium number of pores. By taking the inverse

6

Z.A. Levine

relationship of Eq. 1, one can deduce the characteristic time that one might expect a single electropore to form in a viscoelastic membrane. This yields a pore creation time per unit pore, or dt/dN, that depends on exp.(
V2), or on the inverse of the voltage squared. This is not unexpected, as it resembles a Boltzmann factor with a Hamiltonian containing a capacitive energy storage term, or U = ½ CV2, where C is the magnitude of capacitance. Monte Carlo models can also be used to study lipid bilayers, but because reversible electroporation is a transient, nonequilibrium event, molecular dynamics (MD) simulations are much better suited to study these phenomena.   dN 2 N ¼ αeβV 1
dt N eq

(1)

Recent observations have also indicated that short, nanosecond electric pulses induce different types of cellular responses compared to microsecond pulses (Vernier et al. 2003). These studies found that ultrashort nanosecond MV/m electric fields can trigger intracellular calcium bursts and the externalization of anionic lipids from the plasma membrane’s inner leaflet to the outer leaflet. This suggests that the electric field duration is an important factor in determining electropore expansion and function, in addition to highlighting how pulse frequency affects physical properties of electropores and subsequent uptake of ions, dyes, and other cytosolic molecules.

Molecular Dynamics Simulations The first molecular dynamics (MD) simulations of a lipid bilayer were performed in 1982 by Herman J.C. Berendsen using two monolayers of 16 decane molecules held together by a simple dihedral potential (Van der Ploeg and Berendsen 1982), where he achieved 80 picoseconds of simulation time for every 10 h of computation time. From his early simulations, he was able to extract carbon-deuterium order parameters relative to the bilayer normal using a second-order Legendre polynomial as a geometric measurement of hydrocarbon orientation and rigidity. Following this, Seelig et al. (1974) confirmed through simulations that lipid tail rigidity decreased as the size of the lipid tails grew, a result which was consistent with Berendsen’s work and with experimental observations at the time. These studies showed, for the first time, that it was possible to use computer simulations as a research tool to explain complex biological processes using simple, classical physical models as a basis. However, because only small systems and short timescales were available for computation at the time, only a limited number of measurements were possible. As computation became cheaper and more affordable, the timescales and system sizes that could be easily simulated grew as well, enabling larger and more complex studies to take place. Exploratory molecular dynamics (MD) simulations of electroporation (Tieleman 2004; Tarek 2005) showed that when large, MV/m electric fields were applied in simulations of phospholipid membranes, a water column would form

Lipid Electropore Lifetime in Molecular Models

7

across the bilayer interior, followed by the construction of a bridge of hydrophilic lipid headgroups (and their hydrating water molecules). It has also been shown that the pore creation time decreases as the pore-initiating electric field increases (Ziegler and Vernier 2008); however, the detailed time course of electropore construction and annihilation in MD simulations has only recently been studied. While several pioneering studies have described key steps in the electroporation process, only a handful of researchers have used simulations to systematically quantify and compare objective stages in the life cycle of lipid electropores, thereby delineating a number of field-dependent and field-independent steps that must occur in order for electroporation to commence. In an early study by Böckmann (2008), the authors found a considerable increase in the bulk water dipole moment surrounding a POPC membrane as a function of the applied electric field. By visual inspection, they also measured the timescales of membrane electroporation and extracted rates of permeabilization across 48 independent trials. In each of the trials, all containing POPC lipids, they observed similar intrusion of water molecules into the membrane interior, followed by a rearrangement of lipids (Fig. 3). The extracted electroporation constants were then logarithmically fit to simulation times, and extrapolated to experimental values, which agreed with one another. This represents the enormous power of atomistic simulations to characterize small membrane patches and extrapolate the results to larger macroscale experiments. Finally, the authors constructed an electric field normalized two-state system (corresponding to the intact bilayer and the electroporated bilayer), where they show the same exponential distribution in pore creation times across all 48 simulations. They also note that ~8 lipids border the interior of the hydrophilic pore and that there are about 57 water molecules present in the pore interior, with an effective pore area of 0.7 nm2. Ziegler and Vernier (2008) utilized molecular dynamics (MD) simulations to identify a minimum porating electric field for multiple homogeneous membranes, where there was a 33 % chance to form electropores over 25 ns (i.e., in one trial out of three). They found that, in agreement with the asymptotic model of electroporation, the pore formation time depended exponentially on the inverse square of the applied voltage (Fig. 4). Minimum porating electric fields for homogeneous DLPC, DPPC, POPC, and DOPC membranes were found to be on the order of 260 MV/m, 280 MV/m, 320 MV/m, and 380 MV/m, respectively, which increased as a function of lipid tail length or bilayer thickness. No preference was observed for the nucleation site of discrete electropores, but characteristic water pedestals intruding into the lipid interface were observed that preceded the formation of hydrophobic pores. Further analysis by Ziegler concluded that water was the primary contributor to electropore formation in lipid membranes, as was hypothesized by Tieleman (2004). Large electric field gradients present at the lipid/water interface, stemming from the alignment of lipid headgroup dipoles during the application of an electric field, are thought to be the initiator of these water pedestals, determining the rate of water entry and subsequent pedestal growth. And, while externally applied electric fields enhance the electric field gradients found at lipid interfaces, changes in the average lipid headgroup tilt are small and do not significantly diverge on opposing sides of

8

Z.A. Levine

Fig. 3 Molecular dynamics simulation of a 512 POPC membrane shown after a permeabilizing electric field of 500 MV/m is applied. Times shown are (a) 900 ps following exposure and (b) 50 ns after exposure, where a smaller pore-stabilizing field was substituted (Reprinted from Bockmann et al. (2008), Copyright (2008), with permission from Elsevier)

the membrane. Therefore, electropores can form from either the cathode or anode end of a bilayer, with little affinity to one side or the other.

Electropore Lifetime Capitalizing on previous work, Levine and Vernier (2010) proposed a number of codified stages of electroporation that were consistent across multiple simulations, allowing for the construction of an objective metric known as electropore “lifetime” that could be used to characterize and compare individual electropores. After applying 2–6 V across POPC and DOPC membranes (corresponding to 300–600 MV/m electric fields), the authors noted that certain processes always preceded the formation of individual electropores. Electropore life cycles were

Lipid Electropore Lifetime in Molecular Models 25 DLPC DPPC POPC

20 Time to Hydrophilic Pore (ns)

Fig. 4 Average time to the formation of a hydrophilic pore based on three independent simulations is plotted against the applied external field. A hydrophilic pore is a quasi-stable structure defined as a continuous water column surrounded by solvated lipid headgroups. Note that 1 mV/nm is equivalent to 1 MV/m (Adapted with permission from Ziegler and Vernier (2008). Copyright (2008) American Chemical Society)

9

DOPC

15

10

5

0 200

300

400

500

600

Field (mV/nm)

broadly divided into a number of pore creation steps and a number of pore annihilation steps. Pore creation was comprised of three stages: initiation, construction, and maturation (later referred to as expansion). Pore annihilation begins when the external electric field is removed from a mature pore structure and proceeds through destabilization, degradation, deconstruction, and dissolution (Fig. 5). The destabilization and degradation steps were later renamed to settling and stabilization, while the deconstruction and dissolution names remained the same (Levine and Vernier 2012). Pore creation. Pore initiation begins with the application of an external electric field and ends when the two groups of water molecules, initially separated by the bilayer, merged to become a single group. Pore construction begins with the formation of the membrane-spanning water column [hydrophobic pore (Abidor et al. 1979)] that marks the merger of the water groups and ends when the phosphorus groups that are initially found on the two leaflets of the bilayer follow the water into the membrane interior and merge into a single phosphorus group. A phosphorus group is defined as a set of atoms, each separated by a maximum distance of 1.2 nm. Because water and the charged phospholipid headgroups now bridge the membrane interior, this structure is comparable to what has been called in several contexts a hydrophilic pore (Abidor et al. 1979). Continued application of the porating electric field resulted in an evolution or maturation/expansion of the hydrophilic pore. In this study, a mature pore was defined as a hydrophilic pore where at least ten phosphorus atoms from the initial anodic leaflet were found within 1.2 nm of phosphorus atoms from the cathodic leaflet.

10

Z.A. Levine

Fig. 5 Phospholipid electropore life cycle. Structurally distinct steps in pore creation and annihilation as observed in molecular dynamics simulations

Pore annihilation. Pore destabilization/settling, the first stage of pore annihilation, is defined as the quasi-stable period after the field is removed during which the number of anode to cathode phosphorus connections fluctuates around the mature pore criterion (ten connections). Pore degradation/stabilization begins when the number of anode-to-cathode phosphorus connections drops below ten and ends when there is only one anode to cathode phosphorus connection. The pore diameter decreases during this time to a minimum, about 0.4–0.6 nm. Pore deconstruction, another quasi-stable period, terminates when the single phosphorus group of the porated bilayer splits into two groups, which remain separate for the remainder of the simulation. At the end of pore deconstruction, only the water column remains. The disassembly of the water column is thereby referred to as pore dissolution. Similar to previous models (Weaver and Chizmadzhev 1996; Ziegler and Vernier 2008), Levine and Vernier found that the pore creation times for POPC bilayers decreased with increasing electric fields. This was almost entirely dependent on the

Lipid Electropore Lifetime in Molecular Models

11

time required for pore initiation, the first stage of pore creation. Pore construction and maturation/expansion times were only slightly dependent on the electric field strength. Pore annihilation times, however, were significantly longer than pore creation times, by an order of magnitude. In pore creation, once the initial membrane-spanning water column had formed, it took less than 1 ns for the headgroups to follow the water into the bilayer interior. When the electric field was removed, however, tens of nanoseconds elapsed before the headgroups migrated back to the bilayer leaflets. During pore destabilization, the period immediately following removal of the electric field, the number of water molecules in the pore (within 0.5 nm of the bilayer midplane) decreased substantially in about a nanosecond, and the pore diameters decreased to about 0.6 nm. In a small percentage of cases, pore annihilation continued to the next stage without delay, but more often the pores remained intact, with the number of inter-leaflet phosphorus connections fluctuating around ten, for a time that ranged from less than one to over 20 ns. At the end of the destabilization stage, the number of anode-to-cathode phosphorus connections permanently declined from ten to one – the reverse evolution of the developed pore to a minimum, single phosphorus group structure. The time required for this decline, the pore degradation/stabilization time, was typically several nanoseconds, significantly longer than pore maturation/expansion, the corresponding pore creation stage. Destabilization/settling and degradation/stabilization times exhibited large standard deviations, indicating the stochastic nature of these processes. At the highest porating field, however, there was a significant reduction in the destabilization time, suggesting that pore destabilization/settling has a fielddependent component. Pore deconstruction (separation into two phosphorus groups) and dissolution (expulsion of water from the membrane interior) appeared independent of field strength. Deconstruction times were tens of nanoseconds with large standard deviations, while the minimal hydrophilic pores were quasi-stable. Pore dissolution was rapid, though, since the hydrophobic pores are known to be highly unstable. Levine and Vernier found no major differences in electropore lifetime between homogeneous POPC and DOPC bilayers, though this is perhaps not surprising since each system shares similar minimum porating electric fields (Ziegler and Vernier 2008). DOPC pore initiation was also strongly field-dependent, similar to POPC. DOPC pore construction and maturation/expansion times, similar to POPC, exhibited only a small dependence on electric field strength. Each of the DOPC life cycles was comparable to the corresponding POPC times. However, pore creation times in all of systems remained exponentially proportional to the inverse square magnitude of the bilayer internal electric field (POPC results shown in Fig. 6 with two popular water models, SPC and SPC/E). Membrane internal electric fields (and therefore the field gradients present at each of their interfaces) in DOPC and POPC systems were of similar magnitude. Electropore lifetimes, while useful, can also have significant deficiencies. First, definitions based on arbitrary expansion cutoffs (e.g., as defined by an integer number of phosphorus connections) may not capture more intrinsic interactions that do not depend on, say, phosphorus. These definitions are not based on physical

12

Z.A. Levine

Fig. 6 Electropore creation times across POPC membranes as a function of the membrane internal electric field. Note that Einternal*Lmembrane = Ebox*Lbox. Results are shown for two popular water models – SPC and SPC/E

properties of pores in real systems, because the discrete structures are not well understood on the experimental cellular level. It would be advantageous to define a “mature” pore corresponding to quasi-stable structures that finish expanding in both experiments and simulations, but at present no such structures have been observed in experiments or simulations (Teissie et al. 2005). Second, atomistic simulation times are (compared to experiments) very short and are less than a microsecond. Although this molecular dynamics regime is well suited for comparisons to data from experiments using nanosecond pulsed electric fields, it is possible that longer rise times, lower electric fields, and longer pulses (microseconds to milliseconds) associated with conventional electroporation technology result in different energy landscapes for pore formation and porated membrane structures that are significantly different from those observed in simulations. Finally, important stages may be missing from these definitions. It is possible that pore initiation actually includes two steps, similar to the first few steps in pore annihilation. In the first of these two steps, the true initiating step, thermal jostling of water dipoles and headgroups at the interface, results in the water pedestals described previously (Ziegler and Vernier 2008). A hypothetical structure, which has not yet been identified, may develop in a small percentage of these bumps, either as the bump is formed or as a consequence of the continued interaction of the bump with the electric field. This structure would likely facilitate a low energy path to the formation of a membrane-spanning water column, which would be a clear signature of commitment to pore formation. If there were additional missing stages, they would need to connect the conversion of a pre-pore bump to a structure where water is launched into the membrane interior. Simulations can also model electroporation by maintaining charge imbalances across the membrane, where the formation of pores can mediate the flow of ions

Lipid Electropore Lifetime in Molecular Models

13

across the membrane, thereby reducing the transmembrane potential difference and causing the pore to collapse (Tarek 2005). Other simulations have shown that electroporation can result in permanent changes to the structure of membranes. Fernandez et al. (2012) have also presented a formal mechanism for applying very strong permeabilizing fields, followed by smaller stabilizing electric fields to maintain electropores in simulation for hundreds of nanoseconds, similar to the studies of Böckmann et al. (2008). In larger membrane systems, Tieleman (2004) found that lipid bilayers significantly deform in response to large electric fields. Others have shown that electroporation in lipid bilayers precedes a wider phase transition of lyotropic liquid crystals comprised of membrane and water, stabilized by the electric field. Such studies state that the water hydration of lipids determines the convergent geometry of the liquid-crystal phase (for instance, lamellar or inverted columnar), giving greater insight into the idea that electric fields mediate phase transitions between amphiphilic solutions and fusing cellular membranes. Taking a step further, Tokman et al. (2013) compares the electropore lifetimes between water/lipid systems (e.g., lipid membranes) and water/vacuum systems. When large electric fields are applied to water/vacuum interfaces, a process known as electroevaporation occurs where liberated water molecules form interfacial, energy-minimized pedestals similar to those observed in membrane electropores immediately before electroporation occurs. Therefore, the authors explored whether lipids themselves are necessary for water permeabilization or rather a passive barrier to entry that inhibits water entry into membranes. Interestingly, for similar external electric fields, water electroevaporation occurs significantly faster than membrane electroporation and yields similar water pedestal profiles (Fig. 7). This further supports the idea that lipids inhibit electric field-mediated water entry in membranes and do not cause the electric field gradients that ultimately displace interfacial water into the membrane interior, as some have suggested (Bockmann et al. 2008). Tokman and colleagues found that the dynamic progression of water columns in both water-lipid-water (WLW) and water-vacuum-water (WVW) systems was strikingly similar to one another. The dynamics of pore (water column) formation and the similarity between WLW and WVW simulations were invariant across simulations over a wide range of parameters (external electric field magnitude, vacuum gap width, etc.). WLW and WVW simulations differ mainly in the timescales over which the formation of the water bridge occurs, i.e., the pore initiation time (Levine and Vernier 2010). To compare initiation times between WLW and WVW systems, the authors selected a gap size for WVW systems that resulted in similar magnitudes of the external and internal (in the lipid bilayer interior and the vacuum gap) electric fields. This ensured that the interfacial water molecules were exposed to similar electric fields in the WLW and WVW systems, permitting a fair comparison of pore initiation times. After normalizing equivalent external and internal electric fields, WVW systems formed water channels significantly faster than WLW systems (Fig. 8). With regard to the energetics of water during electroporation, Tokman demonstrated that the pore formation process described above is driven by a collective tendency of interfacial water dipoles to minimize their electrostatic interactions

14

Z.A. Levine

Fig. 7 Snapshots of the time evolution of water-lipid-water (WLW) and water-vacuum-water (WVW) configurations under an external electric field of 500 MV/m. (a) WLW configuration at times 5.8, 6.7, and 7.3 ns from the start of the simulation with both water molecules (oxygen, red; hydrogen, gray) and lipid molecules (phosphorus, yellow; nitrogen, blue; lipid tail groups, silver) displayed. (b) Same WLW data as in (a) but with only water molecules shown. (c) WVW configuration at times 1.157, 1.160, and 1.194 ns (Figure adapted from Tokman et al. (2013))

Lipid Electropore Lifetime in Molecular Models

15

Fig. 8 Average pore initiation times for waterlipid-water (WLW) and watervacuum-water (WVW) systems calculated with three sets of simulations for each configuration (Figure adapted from Tokman et al. (2013))

while adopting an orientation that minimizes the energy of the water dipole in the external electric field. This is reflected in a steady drop in the per-molecule energies of water in the nascent pore as the protrusion develops. This behavior is present in both WVW and WLW simulations. To do this, Tokman defined an interface region by identifying the average density of water molecules in each (1 angstrom thick) slice of the simulation box, selecting the highest value. Slices with water density not exceeding 50 % of the bulk value were considered interfacial, while those with water densities equal to or greater than 50 % of the bulk value were labeled as bulk water. Water molecules in the pore were selected visually and traced back into the bulk at earlier timesteps so that their subsequent trajectories could be analyzed. Once the water molecules comprising the protrusion were identified, Tokman computed the average potential energy per water molecule and its constituent terms that consisted of (i) the electrostatic interaction energy between water molecules in the protrusion and all other water molecules, (ii) the van der Waals interaction between protrusion and bulk water, and (iii) the interaction energy between the protrusion water dipoles and the external electric field. For WLW simulations, Tokman also computed the electrostatic and Lennard-Jones energies between the protrusion water molecules and the lipids. At each time frame, energies are expressed per unit water in the protrusion box, which varied over time. In addition to energies, Tokman calculated the height (H) of water protrusions as the distance between the interface region boundary and the furthest protrusion atom. These measurements revealed a notable drop in the per-molecule energies of waters in the protrusion as the height of the protrusion grows for both WLW and WVW. Correlations between a drop in potential energy and the protrusion growth for each simulation were quantified with Pearson correlation coefficients that revealed a mean correlation of
0.65 for WVW systems and
0.5 for WLW systems. In both

16

Z.A. Levine

systems, protrusion growth is anticorrelated with potential energy or can be thought of as correlated with a drop in potential energy. On average, Tokman found that protruding water molecules in WLW simulations exhibited a 2.8 kJ/mol potential energy drop during pedestal formation, while WVW protruding waters exhibited a 1.4 kJ/mol potential energy drop. In other words, energy minimization is quicker in WVW systems because the energies needed to electroevaporate are smaller than the energies needed to electropermeabilize. Tokman decomposes the potential energies of water into the three components discussed earlier, (i), (ii), and (iii), to illustrate which components are most sensitive to energy minimization. Drops in the potential energies of water during electroporation appear to be dominated by minimization of Coulomb forces. Contributions from dipole-field interactions only decrease slightly, while changes in Lennard-Jones potentials remain largely unchanged. Interestingly, the decrease in the dipole-electric field interaction corresponds with the alignment of the protrusion molecules with the external field. Finally, Tokman attempts to reconcile the minimum porating electric fields in membranes described earlier (Ziegler and Vernier 2008) with simple physical models of idealized dipoles in external electric fields. The authors construct two model systems containing seven dipoles in either a (Tokman et al. 2013) random bulk aggregate geometry or (Ziegler and Vernier 2008) an aligned geometry where each dipole is stacked on top of one another and is oriented in the direction of an electric field. When both systems are exposed to an external electric field, Tokman tabulated contributions from Coulomb interactions, van der Waals interactions, dipole-dipole coupling, and dipole-field coupling and ensured that each dipole followed a Langevin function that orients each dipole in the direction of the field based on the ratio of the electric field energy (μE) to that of the ambient thermal energy (kT). Tokman found that as the external electric field increased, the energy required to maintain the stacked configuration decreased, while the energy required to maintain the bulk aggregate increased. At some threshold electric field, the energies required to maintain these states were the same, eventually allowing one system the opportunity to transition into the other system. This supports the idea that at some critical electric field, interfacial water pedestals have sufficient energy to transition from an aggregate morphology to a membrane-spanning water channel, given enough time to sample each of the available water states. Water translocation, even in the absence of external electric fields, can occur and be measured in simulations, though inward water flux matches the outward water flux, resulting in no net water transport. These events can capture how leaky membranes are (Fig. 9) and can be used to better understand osmotic effects typically encountered in experiments with live cells. The application of a porating electric field has little effect on either the frequency or the direction of these random crossing events. The primary determinant of the frequency (expressed experimentally as the permeability coefficient) is the area per lipid. Systems with a smaller area per lipid have smaller permeability coefficients. Consideration of this phenomenon can enhance our understanding of the mechanism of water bridge construction, from pre-pore pedestal to extension of the water chain across the bilayer. In Fig. 9, thermal rotation of the yellow water molecule brings its hydrogens into close and repulsive

Lipid Electropore Lifetime in Molecular Models

17

Fig. 9 Random permeation event in a phospholipid bilayer. Single H2O molecule (yellow) crosses the phospholipid bilayer interface without an external electric field present, presumably due to thermal jostling. (a) 0 ps, (b) 50 ps, (c) 105 ps, (d) 110 ps

proximity to the hydrogen on the soon-to-be crossing water molecule, which is already poised on the edge of the low-permittivity hydrocarbon region of the lipid tails that make up the membrane interior. Note that there are acyl oxygens surrounding this event, an important component of three-dimensional electrophysiological contours of the site. Similar but less energetic interactions lead to the assembly of intruding water pedestals, with their associated acyl oxygen and other lipid atom neighbors, and to the molecule-by-molecule (sometimes plus one, minus two, etc.)

18

Z.A. Levine

growth of bilayer-spanning water columns. On a macroscale, this is statistical mechanics. On the atomic scale in a phospholipid bilayer, the sum of these events is a new structure, the chain of waters through the membrane interior that results in the formation of a hydrophilic pore, a conduction pathway through a normally impermeant barrier.

Conclusions Each of these studies represents a small yet important piece of what constitutes membrane electropermeabilization. Simulations of lipid bilayers suggest that electropermeabilization is actually comprised of multiple electroporative events, where a large number of discrete, water-filled pores that are capable of facilitating the crossing of extra- and intracellular molecules through the cellular barrier work in tandem. These discrete structures have similar morphologies, free energy barriers, and constituent components that mediate how easily they form and determine their resulting life cycles. Modern simulations, while often idealized or sterilized, offer important insights that can help guide experiments into new and previously unexplored territories. For instance, by characterizing the many field-dependent and field-independent stages of electroporation in simulations, experiments can better test how various perturbations (such as changes in pH, electrolyte concentrations, addition of anionic or zwitterionic lipids) affect pore creation and annihilation at various external voltages and how changes to these properties affect membrane conductance. Simulations also elucidate how multiple factors such as ion imbalances or fixed charged interfaces (e.g., electrodes) contribute to membrane electropermeabilization on both the nanoand macrolevels. In both of these situations, water appears to be the primary factor that determines how readily membrane electropores form or, at water/vacuum interfaces, how quickly electroevaporation occurs. Potential energy drops correlated with the construction of pre-pore pedestals appear to precede electropore formation, and simple toy models can be constructed that help explain how pore pedestals, essentially aggregates of bulk water, transition at critical electric fields to stacked water channels capable of spanning the length of a phospholipid membranes. Simulations, then, offer a ground-up approach to incorporate what is believed to be the most important contributing factors in biology and provide valuable first-order estimates on the inner workings of nature.

References Abidor IG, Arakelyan VB, Chernomordik LV, Chizmadzhev YA, Pastushenko VF, Tarasevich MR (1979) Electric breakdown of bilayer lipid-membranes. 1. main experimental facts and their qualitative discussion. Bioelectrochem Bioenerg 6(1):37–52 Benz R, Zimmermann U (1980) Pulse-length dependence of the electrical breakdown in lipid bilayer-membranes. Biochim Biophys Acta 597(3):637–642

Lipid Electropore Lifetime in Molecular Models

19

Bockmann RA, de Groot BL, Kakorin S, Neumann E, Grubmuller H (2008) Kinetics, statistics, and energetics of lipid membrane electroporation studied by molecular dynamics simulations. Biophys J 95(4):1837–1850. doi:10.1529/biophysj.108.129437 Chizmadzhev YA, Abidor IG (1980) Bilayer lipid-membranes in strong electric-fields. Bioelectrochem Bioenerg 7(1):83–100 DeBruin KA, Krassowska W (1999) Modeling electroporation in a single cell. I. Effects of field strength and rest potential. Biophys J 77(3):1213–1224 Fernandez ML, Risk M, Reigada R, Vernier PT (2012) Size-controlled nanopores in lipid membranes with stabilizing electric fields. Biochem Biophys Res Commun 423(2):325–330. doi:10.1016/j.bbrc.2012.05.122 Gabriel B, Teissie J (1997) Direct observation in the millisecond time range of fluorescent molecule asymmetrical interaction with the electropermeabilized cell membrane. Biophys J 73(5):2630 Gowrishankar TR, Weaver JC (2003) An approach to electrical modeling of single and multiple cells. Proc Natl Acad Sci 100(6):3203–3208 Hamilton WA, Sale AJH (1967) Effects of high electric fields on microorganisms. 2. Mechanism of action of lethal effect. Biochim Biophys Acta 148(3):789–800 Levine ZA, Vernier PT (2010) Life cycle of an electropore: field-dependent and field-independent steps in pore creation and annihilation. J Membr Biol 236(1):27–36. doi:10.1007/s00232-0109277-y Levine ZA, Vernier PT (2012) Calcium and phosphatidylserine inhibit lipid electropore formation and reduce pore lifetime. J Membr Biol 245(10):599–610. doi:10.1007/s00232-012-9471-1 Neu JC, Krassowska W (1999) Asymptotic model of electroporation. Phys Rev E 59(3):3471–3482 Neumann E, Rosenheck K (1972) Permeability changes induced by electric impulses in vesicular membranes. J Membr Biol 10(1):279–290 Neumann E, Schaeferridder M, Wang Y, Hofschneider PH (1982) Gene-transfer into mouse lyoma cells by electroporation in high electric-fields. EMBO J 1(7):841–845 Sale AJH, Hamilton WA (1967) Effects of high electric fields on microorganisms. I. Killing of bacteria and yeasts. Biochim Biophys Acta 148(3):781–788 Seelig A, Seelig J (1974) Dynamic structure of fatty acyl chains in a phospholipid bilayer measured by deuterium magnetic resonance. Biochemistry 13(23):4839–4845 Sugar IP (1981) The effects of external fields on the structure of lipid bilayers. J Physiol Paris 77 (9):1035–1042 Tarek M (2005) Membrane electroporation: a molecular dynamics simulation. Biophys J 88 (6):4045–4053. doi:10.1529/biophysj.104.050617 Teissie J, Golzio M, Rols MP (2005) Mechanisms of cell membrane electropermeabilization: a minireview of our present (lack of?) knowledge. BBA-Gen Subjects 1724(3):270–280. doi:10.1016/j.bbagen.2005.05.006 Tieleman DP (2004) The molecular basis of electroporation. Biophys J 86(1):371a–372a Tokman M, Lee JH, Levine ZA, Ho MC, Colvin ME, Vernier PT (2013) Electric field-driven water dipoles: nanoscale architecture of electroporation. PLoS One 8(4):e61111. doi:10.1371/journal. pone.0061111 Van der Ploeg P, Berendsen H (1982) Molecular dynamics simulation of a bilayer membrane. J Chem Phys 76(6):3271–3276 Vernier PT, Sun Y, Marcu L, Salemi S, Craft CM, Gundersen MA (2003) Calcium bursts induced by nanosecond electric pulses. Biochem Biophys Res Commun 310(2):286–295 Weaver JC, Chizmadzhev YA (1996) Theory of electroporation: a review. Bioelectrochem Bioenerg 41(2):135–160 Ziegler MJ, Vernier PT (2008) Interface water dynamics and porating electric fields for phospholipid bilayers. J Phys Chem B 112(43):13588–13596. doi:10.1021/Jp8027726

Effects of Heterogeneous Membranes and Electrolytes on Electropore Formation Zachary A. Levine

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Salt and Lipid Heterogeneity Affect Membrane Electrostatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Pore Lifetime in Heterogeneous Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Abstract

Atomistic simulations such as molecular dynamics (MD) simulations have revealed much about the fundamental biophysics of electroporation in homogeneous phospholipid bilayers; however, the structures and behaviors of live cellular membranes differ considerably from idealized zwitterionic lipid bilayers. Biological membranes contain both neutral and charged lipid types and interact with a large number of bulk and interfacial electrolytes that form complexes with individual lipids, thereby modulating their local surface tensions and creating domains and rafts regions. Even without considering the effects of transmembrane proteins, some of which are voltage-gated and are likely to perturb electropore formation and annihilation, the differences between electroporation in heterogeneous membranes, especially those containing salts, and homogeneous membranes described in the last section, are significant. This section will focus on how local perturbations to membranes such as the inclusion of anionic lipids, divalent cations such as calcium, oxidized lipids, and other additions can Z.A. Levine (*) Department of Physics, University of California Santa Barbara, Santa Barbara, CA, USA Department of Chemistry and Biochemistry, University of California Santa Barbara, Santa Barbara, CA, USA Materials Research Laboratory, University of California Santa Barbara, Santa Barbara, CA, USA e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_87-1

1

2

Z.A. Levine

significantly change the behaviors of membrane electropermeabilization. Similarly, additional metrics such as calcium binding isotherms will be presented to assess the validity of these simulations and how well they relate to experiments. Finally, some additional studies will be discussed to deduce whether heterogeneous systems (more representative of live cellular membranes) form electropores with an exponential inverse dependence on applied voltage and electric field, as is observed for homogeneous systems. Keywords

Electroporation • Electropermeabilization • Heterogeneous Membranes • Computer Simulations • Electropore Life Cycle • Anionic Lipids • Electrolytes

Introduction Live cells exposed to short-, micro- to nano-second external electric fields develop a time-dependent potential across the cellular membrane, which strongly depends on the magnitude and rise time of the applied potential, and the size and diameter of the cell, in addition to the unique dielectric solvent surrounding the cell. When cells are exposed to micro- to milli-second, low-frequency (Hz to kHz) electric fields, the energy is primarily absorbed in the external cellular membrane, which exhibits an electrical charging constant (in standard physiological media) of tens to hundreds of nanoseconds, depending on the cell shape and type. However, the application of microsecond kilovolt-per-meter electric pulses on cells yields conductive pores in the plasma membrane that allows for the migration of charged small and large molecules across the mechanical and dielectric barrier presented by the lipid bilayer. Electroporation technology, which operates in the low-field, long-pulse regime, is widely used to facilitate transport of nucleic acids, pharmaceutical compounds, and other materials into the cytoplasm of living cells (Neumann et al. 1982). Quickly rising electric pulses (i.e., pulse periods that are shorter than the charging time of the external membrane) and high frequency electric fields and voltages are better able to perturb all parts of the cell. For instance, ns electric pulses and GHz frequencies can bypass the plasma membrane entirely and reach into the membranebound structures and biomolecular assemblies found deep in the cytosol and nucleus (Schoenbach et al. 2001). Intracellular responses to these ultra short (ns), highamplitude (MV/m) electric pulses (which, while delivering megawatts of power, do not produce significant thermal effects due to the small energy densities involved, e.g. joules per gram) include calcium bursts and, in some situations, cell apoptosis (Vernier et al. 2003). Thus, the potential for utilizing electroporation to kill cancer cells in vivo, or at least to stunt tumor growth, depends heavily on the effects of large calcium or other salt concentrations on electropore formation and on various apoptosis- or phagocytosis-triggering molecules, such as the anionic lipid phosphatidylserine (PS) (Vernier et al. 2004).

Effects of Heterogeneous Membranes and Electrolytes on Electropore. . .

3

While an attractive target of short, pulsed electroporation has historically been the cell interior, PS externalization, or the translocation of PS from the inner leaflet of the lipid bilayer to the exterior, representing the outside of a cell, can be triggered by ns, MV/m electric fields (Vernier et al. 2004). PS is typically confined to the interior of cells and is regulated by a number of proteins and energy barriers (representing over 100 kJ/mol) that impede native externalization. These barriers include phospholipid head groups, since externalized PS can result in the recognition, binding, and signaling of platelet aggregation and catalysis of clot formation during blood coagulation, thereby setting the stage for phagocytosis and mediating intramembrane signal transduction in lymphocytes. Because PS externalization is a well-defined event that is easily observed in simulations and is associated with nano-electropulse treatment, it provides a good starting place for efforts to characterize the interactions of pulsed electric fields with biological systems. The need to establish objective criteria for the analysis of both the formation and dissipation of membrane electropores in applied electric voltages and fields requires careful measurement and comparisons of the fundamental molecular mechanisms that are explicitly dependent on, and independent of, external electric field strengths. Additionally, such criteria and metrics need to be evaluated across multiple model systems containing a large variety of lipids, solvents, and proteins and under multiple thermodynamic ensembles. This includes incorporating PS into existing frameworks for electropore lifecycles and observing how electroporation affects PS externalization. Many studies have already shown a large association between electroporation and the interactions between interfacial water molecules and their hydrated lipid headgroups (Tieleman 2004; Tarek 2005; Ziegler and Vernier 2008); therefore, one might expect that small differences in lipid type (and other simulation parameters) may result in large downstream effects on electropore creation and annihilation dynamics, in addition to changes in cell viability.

Salt and Lipid Heterogeneity Affect Membrane Electrostatics While experiments have long indicated that calcium and other cations affect the conformation of polar lipid headgroups, simulations have also suggested that ions such as sodium and calcium can induce changes deeper in the membrane interior, near the lipid carbonyl region (Bockmann and Grubmuller 2004). More advanced experiments utilizing IR spectroscopy have since confirmed these observations, where the presence of heterogeneous lipid-ion complexes presents a huge perturbation to idealized electropermeabilization processes in homogeneous (ion-free) membranes. These complexes reduce the lateral mobility of lipids within membranes and increase their surface tensions to the extent that electroporation timescales are affected. Stoichiometrically, monovalent cations such as sodium typically bind to one (often anionic) phospholipid headgroup, while divalent cations like calcium typically bind to two phospholipid headgroups. Alternatively, calcium can bind up to four lipid carbonyl oxygen atoms when it more deeply penetrates the phospholipid bilayer (Bockmann and Grubmuller 2004). Upon binding to lipid headgroups, ions

4

Z.A. Levine

can affect both the net tilt and dipole moment of headgroups (Sachs et al. 2004a), to the extent that the magnitude and direction of these shifts depends on the ion’s charge and size. Additionally, the author observed ordering of water molecules well past the bilayer interface, though such changes are difficult to confirm experimentally. When ion concentrations are equal on both sides of a membrane, no external transmembrane potential exists. However, when ion concentrations differ on each side of a membrane (as can often be the case across cellular membranes), external voltages can result in electrified membrane interfaces (Sachs et al. 2004b; Gurtovenko and Vattulainen 2005), sometimes resulting in permeabilization of the partitioning membrane until the ion imbalance is neutralized (Gurtovenko and Vattulainen 2005). Molecular dynamics (MD) simulations can model these phenomena by implementing double bilayer simulations, explicitly representing an exterior and interior aqueous region containing different salt concentrations. Sachs (Sachs et al. 2004b) measured a potential difference of 170 mV across a dimyristoylphosphatidylcholine (DMPC) membrane with an ion imbalance of only a couple of ions, and while this is likely below the electroporative threshold required to permeabilize DMPC, local anisotropies in ion concentrations near the membrane surface could produce external electric fields that far surpass that threshold, though the timescales of these anisotropies are likely small. However, it is important to note that while external ionic potentials may permeabilize idealized, homogeneous membranes, cellular walls in living tissue often require transmembrane potentials to be present for normal physiological processes to occur, e.g., ATP synthesis, and may be resistant to electropermeabilization due to additional compensating ions and transmembrane proteins. Simulations of heterogeneous membranes have also indicated that even in the absence of ions or transmembrane proteins, small resting potentials across membranes are observed due to asymmetries in lipid composition, often on the order of 100 mV (Gurtovenko and Vattulainen 2007). Further complicating this picture, studies by Nathan Baker (Lee et al. 2008) have shown that when membranes are separated by different salts like NaCl and KCl, but at similar ionic strengths, there still exists a nonzero net potential across the membrane, on the order of 70 mV. The authors attribute this behavior to asymmetric binding behaviors between Na and K+, where some ions bind to lipid carbonyl oxygens, while others do not. This results in a net potential across the membrane that tends to be localized near asymmetric binding sites. Similar results were observed for alkali cations and halide anions (Vácha et al. 2009), where multiple membrane binding sites were identified that uniquely perturbed electron density profiles, resting electrostatic potentials, and area per lipid values. However, the authors claim that such effects are largely underestimated in MD simulations due to the lack of polarizable force fields. To test how sensitive these interactions are to molecular force fields, Gurtovenko and Vattulainen (2008) compared ion:lipid interactions between the GROMACS and CHARMM force-fields and found that in both models the strongest interactions were between phosphatidylcholine (PC) lipids and sodium ions, which resulted in eventual membrane compression. Potassium ions, which are larger than sodium ions,

Effects of Heterogeneous Membranes and Electrolytes on Electropore. . .

5

exhibited relatively weak interactions with PC in the GROMACS force field; however, it is known that in GROMACS, the size of potassium is over exaggerated. Phosphatidylethanolamine (PE) lipids were much less affected by monovalent cations, likely due to their superior ability to form both intra- and interlipid hydrogen bonds. Thus, salt is expected to affect the inner and outer cell membrane leaflets differently, where PE concentrations differ, asymmetrically affecting electropermeabilizion dynamics. To test this hypothesis, Gurtovenko (Gurtovenko and Lyulina 2014) applied permeabilizing electric fields to mixed PC:PE bilayers, and found that water-filled electropores predominantly stem from the PC leaflet, and not the PE leaflet. This resulted in hydrophilic electropores where only PC lipids lined the pore exterior, not both of the lipids present in the membrane. This also yields asymmetric lateral pressures in the membrane from one leaflet to the other, decreasing membrane cohesion. Pore formation times for heterogeneous membranes were typically in between the timescales expected for each of the corresponding homogeneous membranes. Similar studies by Sinn and Dimova (Sinn et al. 2006) found that when the anionic lipid phosphatidylserine (PS) is inserted into PC large unilamellar vesicles (LUVs), calcium binds preferentially to PS lipids over PC. However, when a salt solution is mixed with lipid vesicles, there is initial osmotic shock. To compensate for this, Sinn extruded their samples with NaCl solutions and titrated with iso-osmolar CaCl2 solutions. Under physiological conditions, calcium interacted with cellular membranes in the presence of NaCl, typically at concentrations of 15 mM. Therefore, it remained difficult to interpret the binding behaviors of calcium to PS since Na+ may complete with Ca2+ for available lipid headgroups. It is thought that the presence of NaCl actively screens the membrane from Ca2+, resulting in low binding behaviors. In fact, Sinn found that increasing NaCl concentrations by a factor of 10 reduced CaCl2 concentrations by a factor of 100. Despite these complications, the authors found that when 7 mM of CaCl2 was added to dioleoylphosphatidylcholine (DOPC) LUVs (containing 20 % DOPS and equilibrated with 10 mM NaCl), vesicle fluctuations were dramatically reduced, indicating a large increase in membrane surface tension. As a result, the vesicles began to bleb, and water began to leak out from the interior, eventually preceding membrane rupture (Fig. 1). In addition, Sinn and Dimova show well-behaved Langmuir binding isotherms for calcium bound to heterogeneous PC:PS LUVs using titration of added calcium (Fig. 2). For small calcium concentrations, nearly all of the added calcium binds to the lipids. Interestingly, calcium binds to the vesicles far beyond stoichiometric values, assuming that a single divalent calcium binds to two PS lipids. Binding of calcium appears to exceed ratios where each PS headgroup is accounted for, indicating that calcium binds to both PS and PC headgroups alike, though PS is electrostatically preferred and appears to be saturated by calcium first. Fitting of the Langmuir isotherm with a 1:2 binding model yields a binding constant of 650 M 1. Considering the electrostatic enrichment within the double layer, this value is notably similar to the value for pure DOPC vesicles. Therefore, it appears that the presence of PS encourages additional calcium ions to bind electrostatically but that

6

Z.A. Levine

Fig. 1 A sequence of pictures of a giant DOPS:DOPC (γPS = 0.2) vesicle in 10 mM NaCl before and after injection of iso-osmolar 7 mM CaCl2. Time is relative and set to zero before injection as shown on each snapshot (min:s). (a) No CaCl2 is present; the vesicle strongly fluctuates. (b) CaCl2 solution has been injected and a significant decrease of fluctuations is observed. (c) The membrane appears thicker and the vesicle interior fades indicating leakage; the vesicle volume has also decreased. (d) The vesicle collapses and vanishes (Adapted from Sinn et al. 2006)

the binding strength to heterogeneous PC:PS bilayers is comparable to homogeneous PC bilayers. Follow-up simulations by Vernier and Ziegler (Vernier et al. 2009b) were intended to better understand these results using MD simulations, since calcium binding to anionic lipids in simulations must agree with experimental observations if further studies on heterogeneous membrane electroporation are to be trusted. The authors found that simulations of mixed, zwitterionic-anionic, asymmetrically arranged DOPC:DOPS bilayers with monovalent and divalent cations are consistent with experimental observations. Simulations found that in DOPC:DOPS mixtures, the area-per-lipid decreased when either an anionic PS or a calcium ion was inserted into the system, while membrane thickness increased after the addition of each. Calcium also appeared to dehydrate the membrane interface, and a large affinity between phospholipids and calcium was observed relative to sodium, especially at PS carboxyl sites compared to PS and PC phosphate and carbonyl sites. Calcium also induced the rotation of PC headgroups outward toward the solvent, representing new features observed in simulations. The authors also observe that while heterogeneous membranes exhibit convergent area per lipid values at 50 ns, binding between calcium and PS/PC lipids continues up to 100 ns, indicating that a reliance on experimental observations of

Effects of Heterogeneous Membranes and Electrolytes on Electropore. . .

7

Fig. 2 Binding isotherms of the titration of DOPS/DOPC (γPS = 0.2) and DOPC LUVs solution in 10 mM NaCl with 7 mM CaCl2. The line indicated with “slope 1” represents the case of 100 % binding. The concentration of bound ion [CaB2+] was determined using a calibration curve before the measurement. The solid curve represents a fit with a 1:2 binding model (Adapted from Sinn et al. 2006)

vesicle area as a marker of equilibrium can be misleading. For a 128-lipid DOPC: DOPS system, calcium saturation of the bilayer occurs between 10 and 100 calcium ions. The PC headgroup dipole angle distribution in PC:PS bilayers is bimodal, an effect of the interaction between the anionic PS and the zwitterionic PC, and this distribution can be further modulated by varying the number of calcium ions in the system. Atomically local binding profiles for anionic versus zwitterionic phospholipids and their combinations, with monovalent and divalent counter-ions, may have significant consequences for membrane lipid and protein aggregations. In homogeneous and heterogeneous membranes that were calcium-saturated, the calcium concentrations in the phosphate or carboxyl region of the bilayer reached a maximum, but then quickly dropped to a minimum at the interfacial water boundary. However, calcium concentrations rose again in the bulk medium to their equilibrium bulk concentrations (Fig. 3). The construction of a Gouy-Chapman double electric layer can be seen in more detail in Fig. 3, which highlights the differences between homogeneous PC membranes and heterogeneous PC:PS membranes. At the DOPConly interface (Fig. 3a), the phosphate-localized calcium peak, the counteracting calcium and chloride profiles, and the exclusion of sodium from the headgroup region are straightforward. At the aqueous boundary of the mixed DOPC:DOPS leaflet on the other side of the bilayer (Fig. 3b), the calcium and chloride peaks are distinct, but the calcium profile in the headgroup region traces a landscape complicated by the presence of the serine carboxyl groups, and the calcium is accompanied by a significantly greater amount of sodium than in the DOPC-only leaflet. Vernier and Ziegler also observed the formation of 1:2:4 calcium:lipid:water complexes that formed in mixed bilayers, especially between calcium and PS carboxyl groups (Fig. 4). Calcium binding to lipid backbone carbonyl oxygens

8

a

Cl-

Water/2 DOPC PO4

15

Ca+10 DOPS COO-2

10

Cl-10 Na+10

Ca2+ Number Density (nm–3)

Fig. 3 Number density (nm 3) profiles for the interface regions of the DOPC-only (a) and DOPC: DOPS (b) leaflets of a DOPC: DOPS (128:20) asymmetric bilayer. The DOPC-only leaflet contains 64 DOPC. The DOPC:DOPS leaflet contains 44 DOPC and 20 DOPS. In addition to the other scaling constants, the chloride and sodium densities are each multiplied by 10 (Adapted from Vernier et al. 2009b)

Z.A. Levine

5 0

–4

–3

b

–2

–1

ClPO4-

15 Ca2+

10

COO-

Ca2+

5 0

1

2 3 4 Distance From Bilayer Center (nm)

and headgroup phosphate oxygens remained constant between homogeneous PC membranes and heterogeneous PC:PS membranes; however, calcium’s affinity for PS carboxyl groups was an order of magnitude higher than its affinity to carbonyl and phosphate oxygens, according to radial distribution functions produced by the authors. These results paint a picture where PS carboxyls are occupied by calcium first, followed by binding to glycerol carbonyl oxygens, and to a lesser extent, phosphate oxygens. Additionally, since PS lipids contain one additional binding site compared to PC lipids (e.g., at the serine carboxyl group), PS lipids are the most likely candidates that contribute to the reduction in membrane area per lipid, and likely participate in raft or domain formation, due to their ability to preferentially bind to calcium.

Pore Lifetime in Heterogeneous Membranes Building on previous studies, Levine utilized MD simulations to study the life cycles of electropores in the presence of anionic lipids and calcium (Levine and Vernier 2012), and compared their results to existing continuum theories (Weaver and Chizmadzhev 1996) and experiments (Sinn et al. 2006). The authors utilized a similar lifecycle measurement to characterize these systems, which was discussed in an earlier section (Lipid Electropore Lifetime in Molecular Models). At smaller

Effects of Heterogeneous Membranes and Electrolytes on Electropore. . .

9

Fig. 4 A representative calcium-DOPS-water complex from an equilibrated DOPC: DOPS (108:20) system. Calcium binds to DOPS and water with a stoichiometric ratio of 1:2:4. Blue lines connect calcium and atoms within 0.3 nm (oxygens from water and the DOPS carboxyl groups). Violet: Ca; red: O; white: H; teal: C; blue: N; gold: P. (Adapted from Vernier et al. 2009b)

external electric fields, heterogeneous palmitoyloleoylphosphatidylcholine: palmitoyloleoylphosphatidylserine (POPC:POPS) bilayers had pore creation times that were, on average, slightly longer than homogeneous POPC bilayers (Fig. 5), primarily because of an increase in the pore initiation time, the time it takes water to bridge the membrane interior. The area per lipid – which is strongly correlated with membrane permeability – in POPC:POPS systems was about 0.60 nm2 compared to 0.66 nm2 for POPC bilayers. At higher applied electric fields, the differences in pore creation time became minimal, and mixed bilayers had pore creation times that were not significantly different from those for homogenous bilayers. Pore initiation times for POPC:POPS were inversely related to the externally applied electric field, and pore construction times remained constant over all the electric fields sampled, similar to previously reported studies on homogeneous POPC bilayers (Levine and Vernier 2010). Pore expansion times decreased slightly as higher external electric fields were applied. POPC bilayer systems containing calcium ions had exhibited, at small external electric fields, pore creation times which were about twice as long as systems without calcium, and about one and a half times as long as those in mixed (POPC:POPS) bilayers (Fig. 5). As with systems containing PS, when the external electric field was increased, pore creation times of pure POPC systems with calcium and those without calcium did not appear to be significantly different. The convergent area per lipid for POPC bilayers with calcium was about 0.56 nm2, a value significantly smaller than

10

Z.A. Levine

Fig. 5 Pore creation times for three different porating electric fields (400, 500, 600 MV/m) for bilayers consisting of 128 POPC (0PS:0Ca), 128 POPC saturated with calcium (0PS:100Ca), 108 POPC and 20 POPS on a single leaflet without calcium present (20PS:0Ca), and bilayers containing both PS and Ca2+ (20PS:100Ca). Systems containing 20 PS also contain 20 Na+ as counter ions, and systems containing Ca2+ contain two chloride counter ions for every calcium ion. Systems that contain both PS and Ca2+ have both sodium and chloride counter ions present. Ca2+ and POPS in the bilayer increase the pore initiation time

the area per lipid without calcium. Similar to the systems containing PS, the authors observed an inverse relationship between the externally applied electric field and the pore initiation time and pore expansion time for POPC bilayer systems containing calcium. Pore construction times remained constant at all the electric field values reported. POPC:POPS bilayers containing calcium had pore creation times which were significantly longer than for pure POPC systems, with or without calcium, and were also longer than POPC:POPS systems without calcium. This appeared true for all electric field values the authors applied, even at very high fields where the pore creation times were not significantly different. The average area per lipid of POPC: POPS bilayers with calcium was about 0.55 nm2 before a field was applied, somewhat smaller than the area per lipid of POPC bilayers with calcium present.

Effects of Heterogeneous Membranes and Electrolytes on Electropore. . .

11

Fig. 6 Pore annihilation times after pore formation in the same systems shown in Fig. 5. Ca2+ and POPS in the bilayer decrease the average pore annihilation time

Similar to the other systems reported, both pore initiation times and pore expansion times decreased as the external electric field increased, while pore construction times were similar for all the values of applied electric field. With regard to pore annihilation, Levine and Vernier found that POPC:POPS bilayers had pore annihilation times about three times smaller than for pure POPC bilayers (Fig. 6). All of the pore annihilation stages (except stabilization) were shortened relative to pure POPC, and this was true regardless of the values of the electric field used to create the pore. The average pore radius at the start of the annihilation step appeared to be about 2.3 nm for POPC:POPS systems, similar to the average pore radius for pure PC bilayers, 2.2 nm. Pure POPC bilayer pore annihilation times were heavily dominated by the pore settling and pore deconstruction phases. POPC:POPS bilayers exhibited very short pore settling times and

12

Z.A. Levine

significantly reduced pore deconstruction times compared to POPC systems, reducing the overall time required to annihilate POPC:POPS electropores. An initial decrease in pore radius occurred immediately after the field was removed in both POPC and POPC:POPS bilayers, from about 2.3 nm to about 0.5 nm over the first few nanoseconds. This initial reduction of the pore radius does not appear to be correlated with the pore settling time. Pure POPC systems containing calcium had dramatically reduced pore annihilation times compared to POPC systems without calcium (Fig. 6). For pores created at 400 MV/m and 500 MV/m, POPC–Ca2+ systems exhibited virtually no pore settling, while the remaining stages took no more than a few nanoseconds to complete. For pores created at 600 MV/m, calcium still significantly reduced the pore annihilation time, but the variation from simulation to simulation was large. As with POPC:POPS and pure POPC bilayers without calcium, the pore radius was reduced in POPC systems containing calcium to 0.5 nm after only a few nanoseconds, but with calcium present the initial pore radius immediately after the external field is removed is 1.7 nm, about 0.6 nm smaller than the pores in the annihilation simulations without calcium. POPC:POPS systems with calcium present had pore annihilation times which were also significantly shorter than those for pure POPC bilayers and POPC:POPS systems without calcium. POPC:POPS–Ca2+ annihilation times were comparable to those for POPC–Ca2+ and POPC:POPS (without calcium) systems. For pores created at the highest electric fields, POPC:POPS–Ca2+ systems exhibited the shortest pore annihilation times of all the systems reported. Pore settling time for these systems was too short to measure in about half of the trials; though the longest settling time measured for POPC:POPS–Ca2+ was 200 ps. Stabilization was the dominant step in pore annihilation for these systems, and the variance in pore annihilation time appeared to be much smaller than for pure POPC systems without calcium. As with POPC–Ca2+ systems, the initial pore radius for POPC:POPS–Ca2+ systems after the porating field was removed was about 1.8 nm. As soon as the pore radius decreased to about 0.5 nm in the POPC:POPS–Ca2+ systems, the pores dissipated quickly, in stark contrast to the pure POPC bilayers without calcium, where pores remained open with a radius around 0.5 nm for many tens of nanoseconds. To assess the validity of the calcium ion models utilized with POPC:POPS systems, Levine and Vernier constructed a binding isotherm (Fig. 7), which shows the amount of bound calcium ions against the bulk calcium ion concentration. These data were fit to a 1:2 Langmuir binding isotherm between calcium and phospholipid and with binding isotherms taken from experiments with pure and mixed vesicles (Sinn et al. 2006). For small calcium concentrations, the authors saw linear 1:1 binding, where calcium binded preferentially to PS carboxyl and PC:PS phosphoryl oxygen atoms. From systems equilibrated for 150 ns (Vernier et al. 2009b), the authors show a calcium binding coefficient of K = 2.56 M 1. In summary, Levine and Vernier show that the incorporation of anionic phospholipids into a zwitterionic homogeneous bilayer slightly increases the time required for pore creation at lower external electric fields and drastically decreases the time

Effects of Heterogeneous Membranes and Electrolytes on Electropore. . .

13

Fig. 7 Calcium binding curves show rough correspondence between experimental and simulated results. The data can be fit by a 1:2 Langmuir binding isotherm, consistent with the formation of Ca:PS2 complexes

required for pore annihilation for electropores created at all external electric fields. PS bilayers have a smaller area per lipid compared to PC bilayers, and a decrease in surface tension has been shown to affect pore formation (Tieleman 2004). At higher electric fields, pore creation times for heterogeneous bilayers were similar to those for homogeneous bilayers. As previous studies had shown, interfacial water was the dominant component of pore formation (Tieleman 2004; Tarek 2005; Ziegler and Vernier 2008), and one can speculate that the formation of bilayer-spanning water bridges in the strong interfacial electric field gradients that result from the application of large external electric fields is only weakly influenced by PS or PC interactions. This is consistent with the stochastic pore hypothesis for electropermeabilization (Neu and Krassowska 1999) and with previous observations (Levine and Vernier 2010), where pore creation time is exponentially correlated with the inverse magnitude of the applied electric field squared. Similarly, calcium in homogeneous PC bilayers delays pore creation at lower external electric fields and greatly reduces the time required for pore annihilation. Heterogeneous PC:PS systems containing calcium had even longer pore creation times at lower electric fields. The pore radii after creation was lower in both POPC and POPC:POPS systems containing calcium compared to systems without calcium. This may also be associated with significantly smaller area per lipid values found both experimentally and in simulations (Bockmann and Grubmuller 2004) when calcium is present. Even though the area per lipid was similar for POPC–Ca2+ and POPC:POPS–Ca2+ systems, it has been shown experimentally (and in MD simulations) that calcium binds with more affinity to PC:PS vesicles than to pure PC vesicles (Sinn et al. 2006; Vernier et al. 2009b), forming Ca2+–PS complexes involving PS carboxyl and phosphoryl oxygens. By increasing the total number of

14

Z.A. Levine

calcium-lipid complexes, one effectively decreases the area per lipid of each system and thus increases the surface tension of the membrane tension. Alternatively, when the external field is removed a change in surface tension would result in a change in the pore edge tension since the two quantities are related; thus, one might expect to see modified pore annihilation times for systems with PS and/or Ca2+ present, compared to systems without PS and Ca2+. POPC:POPS–Ca2+ systems exhibit very short pore annihilation times (a few nanoseconds), similar to the timescales of pore creation. However, this is an order of magnitude faster than what was observed for pure POPC systems, where pore annihilation times were tens of (or, in some cases, over 100) nanoseconds in simulations (Levine and Vernier 2010). These timescales are similar to the resealing timescales reported in other molecular dynamics simulations (Tarek 2005); however, a large discrepancy remains between simulated pore annihilation times and experimental resealing times which occur on timescales of milliseconds to hundreds of seconds. These long-lasting pores occur in living cell membranes, not simplified phospholipid bilayers, suggesting that some membrane restructuring beyond lipid nanopore formation occurs, and that the permeabilizing structures in cell membranes have significantly stabilizing features. Studies by Ho et al. (2013) have also highlighted how electropores that form in the presence of electrolytes facilitate ionic current through permeabilized membranes, allowing comparisons to be made between simulations and experiment. Using the method of Fernandez et al. (2012) for stabilizing open electropores in MD simulations, the authors characterized the electrical properties of open electropores under controlled, steady-state conditions. They found that pores could be sustained for long periods of time (hundreds of nanoseconds or longer) by substituting a smaller, pore-stabilizing electric field in place of the larger, porecreating electric field. The diameters of these pores could be fine tuned by varying the external field, enabling a systematic examination of the ion conductance of lipid electropores (Fig. 8). Historically, increases in electric conductance were the first indicators of the electric field-driven breakdown of the membrane barrier function that has come to be known as electroporation; thus, conductance remains a useful metric for characterizing open membrane pores. Ho compares conductance values from chronopotentiometric studies of planar lipid bilayers, allowing for the extraction of conductances from single electropores, to those obtained from MD simulations of field-stabilized electropores in homogeneous POPC bilayers containing NaCl and KCl. The authors find that Na+ binds more strongly to the phospholipid interface compared to K+ or Cl , with bound Na+ located deep in the interface at the acyl oxygens of the glycerol backbone. Conversely, K+ and Cl show little affinity for the bilayer, consistent with previous MD studies. As mentioned previously, the maintenance of intracellular and extracellular concentrations of these ions ([Na+] is much higher outside the cell; [K+] is much higher inside the cell) is critical for the electrophysiology of living cells. Disruption of intra- and extra-cellular ion gradients, normally maintained by ion channels and pumps at considerable energy expense, by electroporation allows for diffusive transport of ions and small molecules through the permeabilized membrane. This

Effects of Heterogeneous Membranes and Electrolytes on Electropore. . .

15

Fig. 8 A snapshot from an MD simulation that shows ionic current flowing through an open electropore. Orange ions represent sodium while green ions represent chloride (Adapted from Ho et al. 2013)

effect subsequently has profound consequences for the cell, which must restore osmotic and ionic balance or perish. During the application of a porating electric field (Ep = 400 MV/m), an electropore is formed in both NaCl-containing and KCl-containing POPC membranes within 2 ns. During the application of such high electric fields, electropore will continue to expand until finite size effects (artifacts of simulated unit-cells) dominate the simulation and the bilayer becomes disorganized. Pores are stabilized with a sustaining electric field Es, until pore radius reaches approximately 1 nm. While Ho’s choice of Ep, and the target pore radius are arbitrary values, they should not significantly influence the steady-state properties of the electropore. Sustaining fields included Es = 30, 40, 50, 75, 100, and 120 MV/m for both NaCl and KCl systems. For Es = 30 MV/m, the external field failed to maintain an open electropore in the NaCl system (the pore annihilated within 10 ns in 2 independent trials). For Es = 40 MV/m, a pore was sustained for more than 15 ns in the NaCl system, but no significant Na+ transport was observed. In KCl systems, the electropores were sustained for over 15 ns at Es = 30 MV/m (and also at 40, 50, 75, and 100 MV/m) in three independent trials. At Es = 120 MV/m, the bilayer became disorganized within 100 ns in both NaCl and KCl systems, a result of finite size effects and the periodic boundary conditions. For stabilizing fields around 50, 75, and 100 MV/m, electropores were stabilized for at least 200 ns in both NaCl and KCl systems. Pore radii (which were, on average, 1 nm) reached a field-stabilized value within 30 ns. The stabilized, time-averaged pore radius for each stabilized electropore, taken from the last 50 ns of each simulation, increased linearly with Es over the full range of sustaining electric fields. For each stabilizing field, the pore radius for the NaCl system was 140 pm smaller

16

Z.A. Levine

than for the KCl system, likely a result of the more extensive Na+ binding to POPC, which shrank the area per lipid and increased electropore line tension. The mobilities of Na+ in MD simulations were less than that of K+ and Cl , a consequence of the larger hydrodynamic radius of Na+. At infinite dilution, the ratio between the known values of K+ molar conductivity and Na+ molar conductivity was approximately 1.5. Approximately the same ratio (1.6) was extracted from MD simulations. At 0.15 M, the molar conductivity of ions decreased according to Kohlrausch’s law. As a result of cation anion interactions, the ratio between the molar conductivities in simulations increased to approximately 1.8. The conductance of an electrically permeabilized lipid bilayer (for a given ion species) was determined by Ho using a combination of the native ion conductance, ion-pore interactions, and by the radius of the pore, which was fine tuned with the sustaining field Es mentioned earlier. Increases in ion conduction scale with pore radius, as one might expect, but Na+ conductance was significantly less than K+ conductance for a given radius. This was a consequence of stronger Na+ binding to the interface, which resulted in a nonlinear relationship between pore radius and Na+ conductance at small pore radii. At pore radii around 0.77 nm (the smallest pore radius measured in the KCl system), the ratio of the cation conductances was approximately 5.6, a reflection of the strong interaction between Na+ and the phospholipid head groups in the pore wall (relative to K+). For larger pore radii (e.g., 1.5 nm, the largest pore radius measured in NaClcontaining systems), the ratio of conductivities was only 2.3. Clearly, as the pore size increased, a larger fraction of bulk sodium ions were free to traverse the pore, not hindered by interactions with the electropore wall. Confinement in these simulations increases the likelihood for cation anion electrostatic interaction, which subsequently decreases the net velocity of the ions. The snapshots in Fig. 9 capture this behavior. Ho examined this effect by measuring electrophoretic mobilities in bulk water systems at varying ion concentrations and found that for NaCl systems, the ion mobility decreased by nearly half when the concentration increased from about 0.1 to 1 M. In KCl systems, the decrease in ion mobility was approximately 35 %. Ho also compares conductance values from MD simulations with those reported from chronopotentiometry experiments for POPC electropores in KCl systems. Although the two sets of conductance values lie within the same order of magnitude, the pore radii in simulations was smaller than those reported in the chronopotentiometry data. There is, unfortunately, no direct method for measuring electropore radius experimentally; therefore, pore radius values from chronopotentiometric results were extracted from the measured conductances using a simple model where ions are assumed to travel through the pore as they would through a cylinder of bulk water. In this model, the pore diameter can be estimated by d = 2(GpL/kπ)1/2, where Gp is the pore conductance, L is the length of the pore, and k is the conductivity of the electrolyte. Since the accuracy of the extracted pore diameter depends on the applicability of this simple geometric model, Ho placed large bands of uncertainty around the chronopotentiometric results. Additional assumptions in the chronopotentiometric data were that the ion concentrations inside the pore are the same as in the electrolyte, though Ho later finds that the concentration of KCl in

Effects of Heterogeneous Membranes and Electrolytes on Electropore. . .

17

Fig. 9 Ion interactions during migration through an electropore. External electric field of 75 MV/m in the +z (upward) direction was applied in the snapshots. Na+ and Cl ions are represented as magenta and green spheres, respectively. Phosphorus atoms and nitrogen atoms of POPC are shown as brown spheres and blue spheres to delineate the pore. Lipids tails are represented as blue lines. Water molecules and other atoms are omitted for clarity (Adapted from Ho et al. 2013)

electropores was twice the bulk value. This makes direct comparisons to experiments difficult, if not unreliable at this time. Overall, simulations like those performed by Ho exhibit that the electropore radius can be globally modulated by sustaining external fields. Ion conductance was measured by only two factors: pore radius and the extent of ion interaction with the phospholipid interface along the pore walls. The stronger binding behaviors of Na+ resulted in significantly lower conductance for NaCl systems (compared to KCl systems). This difference decreases as the pore radius was increased, and the pore cross-section became more dominated by bulk water, reducing the probability of pore wall interactions. The electric conductance of simulated electropores in KCl-POPC systems was roughly comparable to the single-pore conductance extracted from chronopotentiometric measurements, but different assumptions underlying the values used for pore radius and conductance prevented precise comparison between the two datasets. In addition, the choice of ion and water models in simulations may also affect the ion conductance values extracted. Indeed, better models are needed to enable quantitative comparisons between atomistic simulations and experiment. Vernier et al. (2009a) also considered the effects of oxidized lipids on electropore formation after earlier studies found that membranes containing a number of oxidized lipids became leakier and allowed a large number of water molecules to pass

18

Z.A. Levine

Fig. 10 Oxidized and unoxidized phospholipid conformations change over time. Composite snapshots (21 images captured at 0.5 ns intervals over a 10 ns period) of PLPC and two oxidized variants, oxPLPC (12-al) and oxPLPC (13-tc), showing their conformations in molecular dynamics simulations of PLPC with 11 % oxPLPC bilayers in a 360 mV/nm field. The spheres near the end of the lipid tails mark the location of the introduced oxygens or C13 of PLPC. Structures of the individual lipid molecules are shown below the corresponding composite. Teal – C, red – O, gold – P, blue – N, gray – C-13 (Adapted from Vernier et al. 2009a)

through the membrane with ease. Since reversible electropermeabilization is widely used to facilitate the introduction of genetic material and pharmaceutical agents into living cells, Vernier shows (using molecular dynamics simulations) that oxidation of membrane components enhances the susceptibility of the membrane to electropermeabilization. This occurs because additional oxygen atoms are present in the membrane interior, which facilitates entry of water molecules across the membrane (Fig. 10). Manipulation of the level of oxidative stress in cell suspensions and in tissues may lead to more efficient permeabilization procedures in the laboratory and in clinical applications such as electrochemotherapy and electrotransfectionmediated gene therapy. Oxidative stress is readily encountered in cultured cells and in live organisms under a variety of conditions and significantly impacts cellular activities across the metabolic spectrum. Because it directly affects the physical properties of the cellular

Effects of Heterogeneous Membranes and Electrolytes on Electropore. . .

19

membrane, it is an important factor to consider in the lifecycle of electropores, which is sensitive to the presence of oxygen atoms and other charges present in the membrane interior. Studies documenting the peroxidation of membrane lipids after electroporation have been reported, but the effects of pretreatment oxidative stress have received little experimental attention. Since molecular dynamics simulations have recently shown that incorporating oxidized lipids into phospholipid bilayers increases the water permeability of these membranes, it may be expected that bilayers containing oxidized lipids will also electroporate more readily (especially since the formation of membrane-spanning water defects is one of the initial steps of electroporation). Molecular dynamics simulations and validating laboratory studies of model membranes and cells in electric fields are consistent with the stochastic pore hypothesis for electropermeabilization (Neu and Krassowska 1999). Additionally, studies such as these show that it is possible to hasten pore formation at the molecular level by oxidatively modifying the properties of the membrane in situ, thereby lowering the energy barrier to poration. In Fig. 11, oxidized membrane patches are shown to be preferential nucleation sites for membrane electroporations, while unoxidized patches remain unaffected. Such a mechanism that might be used to localize where in the membrane poration occurs. This demonstration of the increased sensitivity of oxidatively damaged cells to electropermeabilization has practical implications for the laboratory and the clinic. For instance, an appropriately controlled peroxidizing regimen, that is, one which enhances permeabilization without significantly affecting viability, might increase the efficiency of electrotransfection protocols either indirectly by enabling the use of lower porating voltages (higher voltages are associated with lower cell survival rates) or directly by increasing the amount of genetic material that enters the cells for a given series of electrical pulses. Alternatively, a reducing environment would be expected to protect cells against electropermeabilization. Extending this line of thinking, electrochemotherapy and direct ablation and killing of tumor cells using electrical pulse therapy may likewise be enhanced by procedures which promote oxidative stress in the tumor tissue before pulse delivery – and inhibited when the environment in or around the tumor is reducing. In any mixed population of cells, the different native sensitivities of various cell types to permeabilizing electric fields and oxidative stress might be exploited by adjusting electrical, physical, and chemical parameters to selectively transfect subsets of cells, in vitro and in vivo.

Conclusions Although simple, largely intuitive characterizations of the stages in the lipid electropore life cycles provide a useful scheme for analysis of pore creation and pore annihilation, a more sophisticated approach must build on this foundation to incorporate systematic measurements of the pore radius and pore energies and must include the atomic-scale electric field landscapes and interactions of water-oxygen

20

Z.A. Levine

Fig. 11 Quilted PLPC: oxPLPC bilayer. The simulated system, bounded by the black square in panel A, is divided (dashed lines) into two regions of approximately 100 % PLPC (light gray) and two regions of approximately 50 % oxPLPC (12-al) (dark blue) and 50 % PLPC, as described in the text. To show more clearly where poration occurs after application of an external electric field, copies of the simulated system are tiled to make a periodic system. Preferential electroporation of the PLPC bilayer in regions of high oxidized lipid content is demonstrated in panel B, which shows the system 1 ns after applying a 360 mV/nm field normal to the bilayer. The bilayer is in the plane of the page (Adapted from Vernier et al. 2009a)

and water-hydrogen with the electron-dense acyl oxygens deep in the phospholipid bilayer interface. Once these key components of the molecular structure of electropores are accurately represented, the additional complexities of lipid heterogeneity, membrane proteins, damaged or oxidized lipids, and cytoskeletal and glycocalyx attachments can be added to the model one by one, until one approaches a useful representation of the living cell membrane in porating electric fields.

References Bockmann RA, Grubmuller H (2004) Multistep binding of divalent cations to phospholipid bilayers: a molecular dynamics study. Angew Chem Inter Ed 43(8):1021–1024. doi:10.1002/ anie.200352784

Effects of Heterogeneous Membranes and Electrolytes on Electropore. . .

21

Fernandez ML, Risk M, Reigada R, Vernier PT (2012) Size-controlled nanopores in lipid membranes with stabilizing electric fields. Biochem Biophys Res Commun 423(2):325–330. doi:10.1016/j.bbrc.2012.05.122 Gurtovenko AA, Lyulina AS (2014) Electroporation of asymmetric phospholipid membranes. J Phys Chem B 118(33):9909–9918 Gurtovenko AA, Vattulainen I (2005) Pore formation coupled to ion transport through lipid membranes as induced by transmembrane ionic charge imbalance: atomistic molecular dynamics study. J Am Chem Soc 127(50):17570–17571. doi:10.1021/ja053129n Gurtovenko AA, Vattulainen I (2007) Lipid transmembrane asymmetry and intrinsic membrane potential: two sides of the same coin. J Am Chem Soc 129(17):5358–5359 Gurtovenko AA, Vattulainen I (2008) Effect of NaCl and KCl on phosphatidylcholine and phosphatidylethanolamine lipid membranes: insight from atomic-scale simulations for understanding salt-induced effects in the plasma membrane. J Phys Chem B 112(7):1953–1962 Ho MC, Casciola M, Levine ZA, Vernier PT (2013) Molecular dynamics simulations of ion conductance in field-stabilized nanoscale lipid electropores. J Phys Chem B 117 (39):11633–11640. doi:10.1021/jp401722g Lee S-J, Song Y, Baker NA (2008) Molecular dynamics simulations of asymmetric NaCl and KCl solutions separated by phosphatidylcholine bilayers: potential drops and structural changes induced by strong Na+-lipid interactions and finite size effects. Biophys J 94(9):3565–3576 Levine ZA, Vernier PT (2010) Life cycle of an electropore: field-dependent and field-independent steps in pore creation and annihilation. J Membr Biol 236(1):27–36. doi:10.1007/s00232-0109277-y Levine ZA, Vernier PT (2012) Calcium and phosphatidylserine inhibit lipid electropore formation and reduce pore lifetime. J Membr Biol 245(10):599–610. doi:10.1007/s00232-012-9471-1 Neu JC, Krassowska W (1999) Asymptotic model of electroporation. Phys Rev E 59(3):3471–3482 Neumann E, Schaeferridder M, Wang Y, Hofschneider PH (1982) Gene-transfer into mouse lyoma cells by electroporation in high electric-fields. EMBO J 1(7):841–845 Sachs JN, Nanda H, Petrache HI, Woolf TB (2004a) Changes in phosphatidylcholine headgroup tilt and water order induced by monovalent salts: molecular dynamics simulations. Biophys J 86 (6):3772–3782 Sachs JN, Crozier PS, Woolf TB (2004b) Atomistic simulations of biologically realistic transmembrane potential gradients. J Chem Phys 121(22):10847–10851. doi:10.1063/1.1826056 Schoenbach KH, Beebe SJ, Buescher ES (2001) Intracellular effect of ultrashort electrical pulses. Bioelectromagnetics 22(6):440–448 Sinn CG, Antonietti M, Dimova R (2006) Binding of calcium to phosphatidylcholinephosphatidylserine membranes. Colloids Surf APhysicochem Eng Asp 282:410–419. doi:10.1016/j.colsurfa.2005.10.014 Tarek M (2005) Membrane electroporation: a molecular dynamics simulation. Biophys J 88 (6):4045–4053. doi:10.1529/biophysj.104.050617 Tieleman DP (2004) The molecular basis of electroporation. Biophys J 86 (1):371a-372a Vácha R, Siu SW, Petrov M, Böckmann RA, Barucha-Kraszewska J, Jurkiewicz P, Hof M, Berkowitz ML, Jungwirth P (2009) Effects of alkali cations and halide anions on the DOPC lipid membrane. J Phys Chem A 113(26):7235–7243 Vernier PT, Li AM, Marcu L, Craft CM, Gundersen MA (2003) Ultrashort pulsed electric fields induce membrane phospholipid translocation and caspase activation: differential sensitivities of Jurkat T lymphoblasts and rat glioma C6 cells. IEEE Transactions on Dielectrics and Electrical Insulation 10(5):795–809 Vernier PT, Sun Y, Marcu L, Craft CM, Gundersen MA (2004) Nanosecond pulsed electric fields perturb membrane phospholipids in T lymphoblasts. FEBS letters 572(1):103–108 Vernier PT, Levine ZA, YH W, Joubert V, Ziegler MJ, Mir LM, Tieleman DP (2009a) Electroporating fields target oxidatively damaged areas in the cell membrane. PLoS One 4 (11). doi:10.1371/Journal.Pone.0007966

22

Z.A. Levine

Vernier PT, Ziegler MJ, Dimova R (2009b) Calcium binding and head group dipole angle in phosphatidylserine-phosphatidylcholine bilayers. Langmuir 25(2):1020–1027. doi:10.1021/ La8025057 Weaver JC, Chizmadzhev YA (1996) Theory of electroporation: a review. Bioelectrochem Bioenerg 41(2):135–160 Ziegler MJ, Vernier PT (2008) Interface water dynamics and porating electric fields for phospholipid bilayers. J Phys Chem B 112(43):13588–13596. doi:10.1021/Jp8027726

Lipid Electropore Geometry in Molecular Models P. Marracino, P. T. Vernier, M. Liberti, and F. Apollonio

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 The “Hole” Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 The “Water Cylindrical Slabs” Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 The “Statistical/Analytical” Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Abstract

Molecular dynamics (MD) simulations recently proved to be a useful tool for unveiling many aspects of pore formation in lipid membranes under the influence of external electric fields. In particular, the study of the structure and transport properties of electropores must definitely take advantage of a rigorous characterization of pore geometry and its evolution in time. In order to compare sizerelated properties of pores in bilayers of various compositions, generated and maintained under different physical and chemical conditions, reference metrics are needed. In the present chapter three different methodologies to evaluate electropore geometrical behavior will be presented: (i) the first developed method which allows the analysis of the dimensions of the pore through an algorithm that uses a Monte Carlo simulated annealing procedure to find the best route for a sphere P. Marracino (*) • M. Liberti • F. Apollonio Department of Information Engineering, Electronics, and Telecommunications, Sapienza University of Rome, Rome, Italy e-mail: [email protected]; [email protected]; [email protected] P.T. Vernier Frank Reidy Research Center for Bioelectrics, Old Dominion University, Norfolk, VA, USA e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_88-1

1

2

P. Marracino et al.

with variable radius to squeeze through the pore channel; (ii) a more recent one allowing the three-dimensional modeling of the irregular shape of the pores obtained as the quantification of its volume; and (iii) a new method based on a statistical approach (following essential dynamics principles), able to describe pore geometrical fluctuations in a robust and reproducible way. The three approaches described here are not system specific, i.e., the methods can be generalized for any kind of pore for which appropriate structural information is available.

Keywords

Electroporation • Molecular dynamics • Pore radius • Electropore geometry

Introduction The complex mechanisms underlying electroporation processes in living cells are known only in broad outline. The use of molecular dynamics (MD) simulations provide atomic-scale information on energetic and dynamic contributions of complex biomolecular structure and, specifically for membrane systems, it indicates that application of a strong enough electric field across the membrane bilayers leads to the intrusion of water defects and the formation of aqueous, conductive pores (Tieleman 2004; Tarek 2005; Ziegler and Vernier 2008; Casciola et al. 2014). The capabilities of computer simulations are limited by the accuracy of the underlying models describing atomic interactions and also by the computational expense of adequately exploring all the relevant conformations of the membrane and surrounding water and ions. In order to compute forces, MD simulations require a potential energy function, which describes the interactions between the atoms in the system. If the potential energy of the system is known, given the coordinates of a starting structure and a set of velocities, the force acting on each atom can be calculated. The accuracy of the simulations is directly related to the potential energy function used to describe the interactions between particles. Microscopic all-atom models use detailed all-atom representations of both the membrane and the solution, describing the system as a collection of particles that interact via molecular mechanics force fields, using classical approximations of quantum mechanical energies to describe the Coulombic (electrostatic), van der Waals, and covalent (bond, angle) interactions. In addition to the standard potential interactions, there is the possibility to include an additive force as the one given by an external electromagnetic field. The field is an external/exogenous perturbation which could, in principle, alter the charge distribution, the chemical state, and the energy of the biological structures. In particular, by simulating a “virtual” experiment, it could be possible to observe the field action on the microscopic structure in an accurate and rigorous way, providing a realistic description of the interaction occurring at the atomic level.

Lipid Electropore Geometry in Molecular Models

3

The output of a MD simulation is the trajectory that is the series of molecular configurations evolving in time, generated by solving the classical equations of motion at each time step. According to the ergodic hypothesis, one can simulate a membrane system with its surroundings for a period of time and get time-averaged molecular properties that approach the experimentally measurable ensemble averages. In other words, if the MD simulation is long enough to cover the entire configuration phase-space, the system reaches an equilibrium condition and statistical-mechanics is applicable to calculate macroscopic properties as, for example, the geometrical properties of a stable electropore. Recently, molecular dynamics (MD) simulations have shown that a transverse electric field can produce pores in phospholipid bilayers (Tieleman 2004), and MD has been adopted as a useful method for studying electroporation. The molecular mechanism has been unveiled (Tieleman 2004; Tarek 2005; Ziegler and Vernier 2008), at least for phospholipid bilayers, and the different stages of pore formation have been characterized (Levine and Vernier 2010). However, despite its doubtless utility, MD has a major drawback. The value of the applied field, E, used in the simulations has to be high enough in order to electroporate (Ziegler and Vernier 2008; Fernández et al. 2010) to overcome the large activation energy required to initiate pore formation in a reasonable time window ( 106 cells, up to a few mL) is introduced in a cuvette equipped with two facing electrodes, on which a high voltage is applied in the form of short pulses (exponentially decaying or square pulses, in the

Electroporation in Microfluidic Devices

3

Fig. 1 Principle of cell electroporation. When a cell is exposed to short pulses of a high electric field, the cell transmembrane potential increases. If the latter reaches a critical value of 0.2–1 V, pores are created in the cell membrane. These pores allow molecular exchange between the extraand the intracellular media and, for instance, loading into the cell of foreign substances for which the cell membrane is normally impermeable. In case of reversible electroporation, the pores in the cell membrane reseal (Single Cell Analysis, Single Cell Electroporation using Microfluidic Devices, 2012, page 66, S. Le Gac and A. van den Berg, “with permission of Springer”)

low ms range). In such a cuvette, voltages in the kV range are applied to reach the required threshold for poration. Furthermore, in bulk electroporation, the viable poration rate, corresponding to cells being porated but remaining viable after the treatment, is low (40–70%) (Chen et al. 2008). In general, three scenarios concomitantly take place, which explains this overall low success rate. In a first scenario, there is no membrane poration. In a second scenario, too many pores or too large pores are created so that the cell is not able to recover, which is known as irreversible poration. Finally, in the most favorable case, the membrane is permeabilized and reseals. Bulk electroporation suffers from a number of issues, which explains this heterogeneity in the response of the cells to the electric field and the overall low success rate. First, the electric field is not homogeneous in the cuvette, and cells are randomly oriented with respect to the electric field direction, so that they experience different treatments. Next, the use of high voltages brings about a number of problems. Upon application of high voltages, reactive chemicals, which are toxic to the cells, are produced. Furthermore, Joule heating can induce a significant change of temperature in the cell suspension. Similarly, water electrolysis can happen, with the creation of bubbles, which can also result in a loss of cell viability. Finally, toxic metal ions released from the large plate electrodes also endanger cell viability, depending on the nature of the electrode material. Moreover, this cuvette-based approach is not optimal for the treatment of all kinds of cells. Adherent cells must undergo a (bio) chemical treatment to be placed in solution, which may damage the cell membrane and sensitize the cells to the electric field. Furthermore, rare cells, such as stem cells or induced pluripotent stem cells (iPSCs), are available in too limited amounts to be handled in an mL-sized cuvette. A promising approach to solve some of the

4

S. Le Gac and I. van Uitert

aforementioned issues while bringing better control on the treatment to eventually enhance its success rate is to use microfluidic technology and miniaturized devices. In this chapter, microfluidics is shortly introduced, and the unique features this technology has to offer for cell electroporation in comparison with a bulk process are discussed. Thereafter, different microfluidic-based strategies for cell electroporation are presented as two distinct categories of devices, which are especially suitable for single-cell electroporation and for flow-through treatments, respectively, and the specific advantages and limitations of these two types of devices are discussed. Finally, different fields of applications of cellular electroporation that can especially benefit from the use of microfluidic technology are briefly reviewed.

Microfluidic Technology Microfluidics, as pointed out by Georges Whitesides, can be defined as “the science and technology of systems that process or manipulate small (10 9 to 10 18 l) amounts of fluids using channels with dimensions of tens to hundreds of micrometers” (Whitesides 2006). Microfluidic devices – also known as lab-on-a-chip devices – are miniaturized devices, which have a footprint of a few square centimeters. These devices include structures with dimensions in the 1–100 μm range such as channels or chambers, in which such small volumes of liquids can accurately be transported and processed. The field of microfluidics initially focused on the miniaturization of analytical devices. The motivations to use miniaturized devices for biomolecule and medical analysis were manifold, e.g., the possibility to work with small-sized samples (in the sub-microliter range), the rapidity and sensitivity of the analysis, and the portability of the systems to eventually yield autonomous point-of-care devices. Since their introduction in the 1990s, microfluidic devices have become more and more popular, and their applications have greatly diversified toward, for instance, the field of chemistry and synthesis, and also, more and more, for cell biology-related and biological research. Since the field of microfluidics directly originated from those of microelectronics and microelectromechanical systems (MEMS), microfluidic devices have first been fabricated from silicon and silicon-based materials using processes from the microelectronics industry (i.e., photolithography, wet and dry etching, etc.) in a clean room environment. Glass quickly came in the picture and was preferred to silicon for its transparency and electrical insulating properties. As the field has been further evolving and expanding, materials and associated fabrication processes have been changing. Traditional materials now have to compete with polymers, which are much cheaper and already ubiquitous in laboratories. Polymer processing relies on alternative approaches such as replication or ablation techniques (Becker and Gartner 2008), which do not require having access to dedicate and expensive clean room facilities, but which can readily be implemented in a standard wet lab.

Electroporation in Microfluidic Devices

5

The popularity of microfluidic devices is explained not only by the numerous advantages they present compared to their lab-scale counterparts but also by the fact that they offer new experimentation opportunities, especially in the fields of chemistry and cell biology. Among commonly cited advantages linked to their small dimensions are reduced consumptions of samples and reagents and associated lower operational costs. Similarly, all processes that scale with the dimensions of the device become more efficient and faster like mass transport, heat transfer, gas exchange, molecular diffusion, etc. For experimentation involving cells, the micrometer-sized structures make these devices ideal tools for the manipulation and isolation of individual cells and to conduct experimentation at the single-cell level (Le Gac and van den Berg 2010). Next to this, microdevices enable to emulate the confined environment cells experience in vivo. At the micrometer scale, flows also behave in a very different way than at the macrometer scale, which can be evaluated by examining the dimensionless Reynolds number (Re). The Reynolds number compares inertial and viscous forces and indicates the transition from a turbulent to a laminar regime. In miniaturized devices, as a result of the increased surface-to-volume ratio, surface forces dominate bulk phenomena, and the flow is mostly in the laminar regime. Consequently, flows are predictable and highly controlled in microfluidic devices. A downside of a laminar flow regime however is that mixing only proceeds via molecular diffusion and is therefore not as efficient as when turbulences are involved. Altogether, microfluidic devices offer the possibility to accurately control any physical and chemical parameter both spatially and temporally. This ability is essential in biology to fine-tune the cellular microenvironment, as well as in chemistry to yield more homogeneous reaction mixtures. Finally, microfluidic devices present a high level of integration. One device can, for instance, comprise a series of identical systems (horizontal integration) for assay parallelization. Alternatively, a series of distinct operations can be implemented in one single device (vertical integration) for the accomplishment of a complete (analytical) process. Furthermore, microfluidic structures can be combined with add-on capabilities such as electrodes or sensors (smart integration), e.g., for electrical detection of cells or biomarkers, for controlling the temperature, or for fluid actuation.

Microfluidics for Electroporation For cellular electroporation, microfluidic technology brings a number of attractive features. First, distances in microfluidic devices are in the micrometer range, and since the required electroporation voltage scales with the device dimensions, a few volts are typically sufficient to reach the kV/cm electroporation threshold. Furthermore, microstructures can be designed to locally enhance and/or shape the electric field so that even lower voltages are required. For instance, channels can include constrictions in which the electric field is enhanced, since the E-field strength scales with the cross section of the channels. Alternatively, micrometer-sized structures can

6

S. Le Gac and I. van Uitert

be added to isolate cells, which allows locally creating hot spots in electric field across the trapped cell. Similarly, electrodes can be directly integrated in the device and their shape and geometry optimized toward the same goal. For instance, saw-tooth electrodes in a microchannel have been reported to locally enhance the field strength (Lu et al. 2005). Furthermore, 3D electrode structures provide a more homogeneous electric field distribution across a channel compared with planar electrodes. Altogether, using microfluidics, the electroporation treatment becomes much milder, and all adverse effects originating from the use of high voltage, as normally encountered in bulk electroporation, are drastically reduced or entirely alleviated using such miniaturized devices. As mentioned earlier, microfluidics lends itself well to single-cell experimentation, so that the electroporation treatment can be downscaled to the level of a single cell – or even to the subcellular level – by isolating individual cells at specific positions between two electrodes. This configuration brings about a high control on the electroporation treatment, and all parameters (electric field strength, duration of the treatment, etc.) can be adjusted for every single cell. Arguably, this single-cell approach opens the doors for a 100% theoretical success rate of the treatment. Furthermore, single-cell electroporation is particularly attractive for rare cells, which can be treated individually in a highly controlled and customized fashion, with a high success rate and no loss in viability. Finally, this single-cell approach is interesting to control molecular delivery and/or extraction following cell membrane poration, as discussed in more detail later in this chapter. The high level of integration offered by microfluidics allows combining the electroporation treatment with another process. Cells can be tracked in situ to monitor the process of pore formation in real time and the outcome of the treatment, on a longer time scale. To that end, fluorescent probes are typically employed to instantaneously follow molecular transport across the membrane resulting from its successful permeabilization. Alternatively, the process of pore formation can be detected electrically, by monitoring the membrane conductance or the impedance (▶ Impedance measurements as electroporation measure) of individual cells, using either the same electrodes as for the electroporation or a second pair of electrodes. For longer-term examination, cells can be cultured in the same device and their state evaluated in situ after one or more days using bright-field and/or fluorescence microscopy. A common approach to validate a cell electroporation protocol consists of transfecting a plasmid coding for the green fluorescent protein (GFP); this strategy however implies keeping the cells in culture for about 1 day to verify not only that they have survived the treatment but also that they are functioning properly and able to produce proteins from the exogenous and transfected plasmids. This whole sequence can be implemented in a microfluidic device, with cells being tracked to examine the success of the electroporation/transfection treatment at the single-cell level. The electroporation treatment can finally be coupled to a cell-sorting step, e.g., using dielectrophoresis, either prior to electroporation to select cells to be exposed to the electric field or after, to isolate successfully porated cells.

Electroporation in Microfluidic Devices

7

Classification of Microfluidic Devices for Electroporation In this section, different microfluidic-based cell electroporation strategies are discussed to illustrate the potential of this technology for cell electroporation. Microfluidic devices for cell electroporation are classified in two distinct categories depending on how the treatment is performed: either at the single-cell level or using a flow-through configuration on cell populations. For each of the two approaches, the concept underlying the strategy is explained, and possible variations around the strategies are presented, together with a short discussion on the specific advantages of these respective devices and their possible limitations.

Single-Cell Electroporation Devices Principle Single-cell electroporation devices all rely on the isolation of individual cells, which are typically secured in trapping microstructures. These mechanical trapping sites consist of either microholes, which are located underneath the cells (Huang and Rubinsky 2001, 2003), or lateral trapping slits connected to another microchannel (Khine et al. 2005; Valero et al. 2008; van den Brink et al. 2011). In both cases, the trapping structures present a characteristic dimension of 2–3 μm, which is several times smaller than the cell diameter. Importantly, these low micrometer-sized structures also allow shaping the electric field and creating hot spots of electric field across the isolated cells. Moreover, a microsystem can easily include arrays of these structures for the electroporation of a series of individual cells. Trapping of the cells is a key step in this single-cell electroporation protocol. Typically, a negative pressure is applied across the trap after injection of the cell suspension in the device to attract and immobilize individual cells in the trapping sites. A too high negative pressure may damage the cells and increase their vulnerability and associated risks for cell death. On the contrary, if this force is too low, the cell may not be tightly immobilized in the trapping site and can easily escape. Furthermore, if a cell is loosely trapped, a leakage pathway is created so that a higher potential must be applied for cell poration. This leakage pathway also obviously prevents the electrical detection of pore formation. Once the trap(s) is (are) filled, the cell(s) can be electroporated. To apply the electroporation voltage, electrodes are either integrated in the device or inserted in the access reservoirs. In the former case, electrodes can be placed in the close vicinity of each trapping site, so that a voltage of a few volts is typically needed to achieve cell electroporation. Furthermore, in this configuration, the electric field parameters can be adjusted for each individual cell, would the device include an array of trapping sites and individually addressed electrodes for each trapping site. Therefore, this strategy is particularly attractive to optimize the treatment for every single cell to eventually give a 100% (viable) poration yield.

8

S. Le Gac and I. van Uitert

Bottom Trapping of Cells A first type of single-cell electroporation devices includes a horizontal substrate with an array of micrometer-sized apertures in which cells are trapped. Two channels or chambers are placed above and underneath this perforated substrate to, respectively, introduce the cell suspension and apply a negative pressure for cell trapping. The first microfluidic cell electroporation device worked along this principle of single-cell trapping in a horizontal microhole (Huang and Rubinsky 2001, 1999). This device, reported by Huang and Rubinsky, was composed of two silicon chambers separated by a horizontal 1-μm-thick silicon nitride membrane containing a 2–10-μm diameter orifice (Fig. 2a). The electrodes, which were integrated in the device, consisted of n+-polysilicon membranes placed in the two silicon chambers, and they were separated by 900 μm. The semiconductive and translucent properties of this material allowed both electrical and optical monitoring of the trapping and poration processes. Trapping of a cell and sealing of the cell in the orifice resulted in a great reduction of the current through the trap – so-called leakage current – while cell electroporation was accompanied by an increase in the conductance across the orifice. Using this device, the electroporation voltage was successfully lowered to a few 10s V (single square pulse of 60 ms). Adaptations have been brought to this first device, which proved the concept of single-cell microfluidic electroporation while being limited to one single cell. Addition of another pair of electrodes allowed detecting cells arriving on the micro-orifice to trap them. After electroporation, reversing the pressure released the cell, so that another cell could be attracted in the orifice and subsequently treated (Huang and Rubinsky 2003). This planar configuration of orifices is easily amenable to parallelization to increase the poration throughput through the use of 2D arrays of orifices, in which populations of cells are immobilized and submitted to an electric field. However, in most of these planar devices, common electrodes are used for all orifices in the array so that the electric field treatment cannot be customized per cell, and similarly, the poration process cannot be monitored electrically. Lateral Trapping of Cells In a second configuration, trapping is achieved using lateral structures separating the channel, in which the cell suspension is injected, and another channel (Valero et al. 2008) or a series of individual channels (Khine et al. 2005; van den Brink et al. 2011), from which the negative pressure is applied (Fig. 2b, c). This lateral design lends itself better to electrical decoupling of the traps and to multiplexed single-cell electroporation. Furthermore, provided the different trapping sites are connected to individual channels, the lateral approach is also compatible with parallel single-cell analysis (van den Brink et al. 2011). A second main advantage of this lateral approach, when individual channels and electrodes are used, is that the electroporation process can be monitored electrically; thereby, as soon as pore formation is detected through a change in the measured membrane conductance, the electroporation voltage can be stopped. Similarly, individual channels and individually addressed electrodes can be employed to assist and enhance the introduction of

Electroporation in Microfluidic Devices

a

Bottom electrode

9

Micro hole

Top electrode

Al wires

Top layer Middle layer

Circuit board

Bottom layer

Acrylic substrate

Fluids inlets/outlets

c

b

200µm

50

Ag/AgCl electrode

PDMS Trapping channel

PDMS

Conductive media

Cell 4µm

Fig. 2. Cell trapping devices. (a) Three-layer single-cell electroporation device consisting of two microfluidic chambers in silicon, separated by a 1-μm-thick silicon nitride membrane containing an aperture (2–10 μm in diameter) on which the cell is trapped before electroporation. The top and bottom electrodes are formed by a transparent n+ polysilicon membrane (Courtesy Prof. Rubinsky). (b) Top: Top view of a device for lateral cell trapping. Two wide inlets on either side of the large chamber in the middle are employed to fill it with the solution containing the cells. By applying a suction force via the small radiating side channels, cells are trapped at the channel entrances. The electrodes employed for electroporating the cells and detecting the increase in membrane conductance are connected via the large chamber and to the small cell trapping channels, enabling singlecell electroporation. Bottom left: Schematic cross section of the chip displaying a trapped cell and the two electrodes employed to generate the required field. The trapped cell is pulled into the small opening due to the applied suction force. Bottom right: Photograph of a trapped cell (Reproduced from (Khine et al. 2005) with permission of The Royal Society of Chemistry). (c) Top: Top view of a

10

S. Le Gac and I. van Uitert

foreign substances into the cells, after pore formation, using electrophoresis (Ionescu-Zanetti et al. 2008).

Advantages and Limitations The main advantages of these single-cell miniaturized electroporation devices are the use of milder electrical parameters and the higher control on the poration process, which altogether enhances the success yield in cellular treatment. Since the electrodes are very close to each other and the cells can be tightly trapped, the required potentials become very low compared to conventional systems, ranging from 20 V (Huang and Rubinsky 2001) down to ca. 1 V (Khine et al. 2005), while the success rate can go up to 100% (Ionescu-Zanetti et al. 2008). Moreover, plasmids can be actively brought into the cells by electrophoresis (Ionescu-Zanetti et al. 2008), which is not possible in bulk electroporation. Furthermore, the electric field can be tuned to match the cell properties, which further increases the poration yield. Finally, the electric field distribution across the trapped cell can easily be modeled; the cells are considered as spherical objects (Khine et al. 2005; Valero et al. 2008) (Figs. 2b, c), and their position in the electric field is well defined. These modeling aspects also ultimately contribute to the improvement of the poration treatment. All these advantages make these (single)-cell trapping devices the best suited for the treatment of rare cells. An important drawback of these devices is their level of complexity: they include multiple channels or chambers with low micrometer-sized trapping structures and, sometimes, integrated electrodes. Therefore, their realization involves sophisticated fabrication processes. However, the main limitation of (single)-cell electroporation devices is their treatment capacity; they mostly include a small number of traps, and this, in turn, gives a low cell treatment throughput. It is still worth noticing that a few recent devices including 2D arrays of micro- and nanopores have scaled up the process to tens of thousands of cells in parallel (Chang et al. 2016 ).

Flow-Through Electroporation Devices Principle A second category of microfluidic devices for cellular electroporation, which addresses this issue of throughput, consists of flow-through systems (Bao et al. ä Fig. 2. (continued) silicon-glass microfluidic single-cell electroporation device. Two microfluidic channels are connected by nine cell traps where cells are captured with the help of a suction force. The electroporation signal is individually applied to each cell with nine stimulation electrodes located above the traps and the common ground electrode placed below the traps. The inset shows a picture of trapped single cells between the electrodes. Bottom: Bright-field and fluorescent microscopy images of four trapped cells (C2C12) before and 24 h after transfection with eGFP (enhanced green fluorescent protein), respectively. The transfection resulted in a 100% success rate (Reproduced from (Valero et al. 2008) with permission of The Royal Society of Chemistry)

Electroporation in Microfluidic Devices

11

2008; Kim et al. 2007; Wang and Lu 2008). In these devices, a cell suspension is continuously perfused in a microchannel, and the cells are exposed one by one to the electrical treatment, the voltage being applied again using integrated or external electrodes. With this approach, cells are still treated individually, but larger cell populations can be processed compared to the single-cell electroporation devices. Cells can also be collected at the outlet of the device for further and off-line studies. The channels or structures in which cells are flowing comprise most of the time geometric constrictions, in which the electric field is locally enhanced, so that electroporation takes place in those areas of higher electric field. Therefore, the flow rate at which the cell suspension is perfused in the system is essential, since it directly relates to the time the cells are exposed to the electric field and the electroporation treatment strength and thus, in turn, to the success yield and cell viability rate (Kim et al. 2007). Specifically, increasing the flow rate translates into shorter exposure of the cells to the electric field, which is equivalent to the use of shorter pulses. Flow-through devices can be further divided into two subcategories, depending on how the cell suspension is flown: (i) using a continuous column of liquid or (ii) using a discrete flow based on droplet microfluidics, where (single) cells are encapsulated in sub-nanoliter aqueous droplets in a continuous flow of an immiscible liquid (i.e., an oil phase).

Continuous Flow Approach A first option relies on the use of a simple microchannel with electrodes placed at a short distance from each other, to locally create – between the electrodes – a high electric field. Most of the flow-through devices include in their design constricted areas, where the electric field is concentrated (Bao et al. 2008; Wang and Lu 2006) (Fig. 3a). The electric signal in this configuration is mostly applied using external electrodes, which are inserted in the reservoirs of the device. The width ratio between the channel and the constricted areas determines the electric field enhancement at the poration area (s). The length of the constricted area and the flow rate define the pulse length for the electrical treatment cells are exposed to. Furthermore, multiple pulses can be applied by implementing a series of constrictions in a single channel. In case electrodes are integrated, a design with both electrodes on the same substrate is preferred from a fabrication point of view, even if it gives an inhomogeneous distribution in the electric field. If electrodes are placed on the bottom and top substrates, a more homogeneous electric field is created, but their fabrication is more delicate, especially for the alignment of the electrodes. Furthermore, in this configuration, electrodes are obstructing the view so that the electroporation process cannot be monitored in real time. Another configuration of integrated electrodes has been proposed to solve both issues, where electrodes were placed at the half height of the channel (Sukas et al. 2014) (Fig. 3b). A last alternative, which yields better results in terms of uniformity and strength of the electric field while providing optical access, relies on the use of 3D electrode structures. Liquid-based coupling strategies have also been reported to apply a homogeneous electric field through the

12

S. Le Gac and I. van Uitert

cell suspension while using external electrodes and avoiding any direct cellelectrode contact. In those cases, the electric field is enhanced at the intersection between the main channel, in which the cell suspension is perfused, and a side channel in which the electrodes are introduced. In a first liquid-based strategy, the cell suspension is hydrodynamically focused as a single-cell line in the main channel using lateral co-flows of a high ionic strength solution (Zhu et al. 2010); cells are exposed to the electric field when passing the channel intersection, and the voltage drop across the cells is determined by the width of the focused flow. Alternatively, a fluidic junction using a polyelectric salt bridge was created between the main channel where cells flow and the side channels which were filled with a high ionic strength buffer (Kim et al. 2007); here the main channel was made narrower at the channel intersection to further enhance the electric field at the place of cell electroporation (Fig. 3c).

Droplet Microfluidic Approach Another class of flow-through microfluidic devices uses a discrete flow, which is also known as droplet microfluidics (Baroud et al. 2010). In droplet microfluidics, an emulsion is created between two immiscible liquids – an aqueous phase and an oil phase – to yield water-in-oil or oil-in-water emulsions, with both an exquisite control on the droplet size and a very high production frequency of more than 1000 droplets/s (Baroud et al. 2010). Thereby, an extremely large number of identical picoliter-tonanoliter-sized droplets reactors are created in a very short period of time. Water-inoil emulsions have found a variety of applications in biology (Baroud et al. 2010). They particularly present a number of advantages for single-cell experimentation such as (i) the ability to confine one cell and its environment in a volume 100–1000 times larger than the volume of a cell and (ii) the massive parallelization of the assays. As such, droplet microfluidics allows high-throughput electroporation of individual cells, which are encapsulated in a small aqueous volume in a continuous nonconductive oil phase (Fig. 3d). In droplet microfluidic devices, electrodes are mostly integrated. As before, the electrical treatment depends on the number of electrode pairs (which correlates with the number of applied electric field pulses on the cells), the distance between the electrodes, the strength of the applied signal and the flow velocity, as well as on the droplet volume (Zhan et al. 2009). The confinement offered by droplet microfluidics brings specific advantages compared to other flow-through approaches: a high control on the number of cells and substances to be delivered in the cell, which are encapsulated per droplet, which gives additional control on the molecular delivery process; the possibility to incubate the cells and, for instance, DNA to be transfected in the cell prior to electroporation, which can enhance the transfection efficiency (Qu et al. 2012); the possibility to monitor the cell fate and conduct post-electroporation functional assays at the singlecell level in the droplets; and, for molecular analysis, limited dilution of the molecules extracted from the cell, which is essential for the detection of low-concentration species.

L1

50 µm

L2

W2

e-

Cell movement

Hypertonic solution

Cell injection

Receiving Reservoir

GND

Electroporated cell

Ionic flux Hydrogel plug

Ag/AgCl electrode

L1

W1

Sample Reservoir

Isometric view

100 µm

Buffer with Cells and DNA

d

b

cell

cell

cell

60 µm

Gold Electrode

Power supply

electrode

A9 Cross section:AA9

electrode

Top view A

cell

fused silica

fused silica

Fig. 3. Flow-through devices. (a) Right: Schematic representation of a channel comprising a constricted area to create a region of high electric field to electroporate the cells. This device employs the principle that the electric field scales inversely with the channel cross section, i.e., when a region in the microfluidic channel is smaller, the electric field across it is higher. Inset: Picture of the microfluidic channel at the constricted area (Reprinted with permission from (Wang and Lu 2006). Copyright 2006 American Chemical Society). (b) Schematic representation of a flow-through microfluidic device consisting of a

c

a

Electroporation in Microfluidic Devices 13

14

S. Le Gac and I. van Uitert

Advantages and Limitations The main interest of flow-through devices is the throughput in cell treatment, which can reach values of up to 104–108 cells/min. In flow-through devices, cells are treated individually, and the electroporation efficiency rate is as high as for (single)-cell devices. Similarly, voltages as low as a few volts are required, depending on the exact device geometry and dimensions and on how the electric field is applied. However, parameters of the electrical treatment cannot really be optimized for each individual cell so that the whole population is exposed to the same treatment. Still, this configuration can be further developed to couple cell characterization to cell electroporation, prior or after application of the electrical treatment. For instance, the cell size and morphology can first be evaluated, e.g., using impedance spectroscopy and another pair of integrated electrodes, to tune the electroporation parameters. Most of the flow-through devices do not provide any active control on molecular delivery, and molecular exchange after cell poration only relies on diffusion phenomena. Another limitation associated with these devices is the difficulty to precisely model the electric field distribution across the cells, as the latter keep a certain degree of freedom in their position and orientation with respect to the electrodes, and most of the time the electric field is not homogeneous through the entire channel cross section. Lastly, this flow-through approach is still not suitable for the treatment of both adherent and rare cells, but it is more appropriate for largescale cellular poration.

Promising Applications of Microfluidics for Cell Electroporation Implementing the process of electroporation in microfluidic or miniaturized devices is particularly attractive for certain fields of application, as discussed in the following. Furthermore, depending on the application and the required throughput, one specific category of devices, single cell or flow through, may be preferred. ä Fig. 3. (continued) fused silica microchannel equipped with integrated electrodes placed at half height of the channel to provide an homogeneous electric field across the channel while allowing optical inspection of the cells during their exposure to the electric field (Reproduced from (Sukas et al. 2014) with permission of The Royal Society of Chemistry). (c) Schematic drawing of the electroporation chip including hydrogel plugs that function as salt bridges to form an integrated electrode junction in the channel. Cells are exposed to the electric field and porated in the region between the salt bridges (Reprinted with permission from (Kim et al. 2007). Copyright 2007 American Chemical Society). (d) Droplet microfluidics platform for cell electroporation. Using a T configuration, a water-in-oil emulsion is created with high control on the aqueous droplet size. Each droplet contains one single cell – or no cell – as well as a well-defined amount of DNA to be loaded into the cells. The device includes two electrodes for cell electroporation in the droplets and cell transfection (Reprinted with permission from (Zhan et al. 2009). Copyright 2009 American Chemical Society)

Electroporation in Microfluidic Devices

15

Fundamental Research Since microfluidic devices bring excellent control on the poration parameters and allow varying them in a precise way, they are ideally suited to conduct fundamental studies on the process of electroporation. Furthermore, cells can be monitored in situ to evaluate the effect of the electrical treatment, both in real time or in a longer term, using either electrical or optical means. This set of assets is instrumental, for instance, in gaining new insight into the processes of pore formation and membrane recovery. Similarly, parameters for the electroporation-based treatment can be screened in one device on multiple cells exposed to different conditions and subsequently optimized for various cell types and applications. For such screening, geometrical variations can be implemented in simple channels (e.g., a tapered shape, different constriction dimensions) to locally or gradually alter the electrical treatment and to ultimately study in situ the influence of specific parameters on the poration outcome. The combination of cell electroporation and microfluidics can also benefit fundamental research in the field of cell biology, for instance, to elucidate targeted signaling pathways, and the activation of specific signaling pathways has notably been studied using reversible electroporation. For instance, the translocation of specific kinases in a cell has been followed using “electroporative flow cytometry,” which involves reversible cell poration and analysis of molecular species that diffuse out of the cells through the created pores (Wang et al. 2008). In another approach, cells were transfected with plasmids coding for a protein of interest extracellular signal-regulated kinase 1 (ERK1) coupled to enhanced green fluorescent protein (eGFP), and the accumulation of the resulting fusion protein in the nucleus was tracked in situ in the device using time-lapse imaging after stimulation of the cells with specific growth factors (Valero et al. 2008). The possibility to porate and transfect cells locally in multicellular constructs such as an embryo is of particular interest in the field of developmental biology to unravel processes involved in the development of embryos (▶ Electroporation to study embryology morphogenesis and organogenesis) and to follow cell migration (Mazari et al. 2014).

Intracellular Delivery The combination of microfluidics and electroporation has been explored for loading cells with very different substances, ranging from small molecules such as drugs and siRNA to larger substrates such as DNA, proteins, or even nanoparticles and quantum dots (Sun et al. 2014). A clear benefit brought by miniaturized strategies for the introduction of foreign substances in cells is the possibility to better control the delivery process and to dose the amount of foreign material loaded in the cells. Enhanced DNA delivery can be achieved, for instance, using electrophoresisassisted molecular loading after the cell membrane electroporation (Ionescu-Zanetti et al. 2008). Specifically, a single cell can be trapped in a micro- or even nanometersized constriction (▶ Nanochannel electroporation: delivery of precise amounts of

16

S. Le Gac and I. van Uitert

biomolecules into living cells) connected to a microchannel, in which the substances to be loaded into the cell are introduced. Upon application of a second low-voltage electrical signal across this channel and constriction and after poration of the cell membrane, the substances are pushed into the cell in a highly controlled manner. Furthermore, using electrophoresis-assisted delivery, the transfected DNA is in a free state in the cytoplasm, so that it is more available for its translocation to the nucleus. Similarly, preincubation of individual cells with DNA in confined volumes created using droplet microfluidics proved not only to enhance the transfer efficiency but also to yield more homogeneous transfection in a cell population (Qu et al. 2012). Furthermore, the need in DNA or other substances to load in the cell is drastically reduced since the volumes involved in microfluidic devices are more than three orders of magnitude smaller than in the traditional electroporation cuvettes. For such cellular delivery applications, both types of devices, single-cell electroporation and flow-through devices, can be used depending on the throughput required and number of cells to be treated. The possibility to use milder conditions for cell electroporation, together with the higher success rate, and the ability to manipulate small populations of cells or even single cells make a microfluidic format particularly attractive for the treatment of rare and fragile cells such as stem cells, primary cells, or induced pluripotent stem cells, with plethora of applications in various fields such as regenerative medicine (▶ Tissue engineering with electroporation) and gene therapy (▶ Principles of electroporation for gene therapy). Similarly, the excellent control on the electroporation parameters brought by the micrometer scale has proven to benefit the electroporation of hard-to-transfect cells (Qu et al. 2012).

Food and Biotechnology Applications Flow-through microfluidic devices find multiple applications in the food industry for pulsed electric field (PEF) treatment (▶ Pulsed electric fields treatment of biological suspensions) for the pasteurization of liquid food samples (▶ Pulsed electric field treatment for beverage production and preservation) like dairy products and fruit juices. There, all bacteria and microorganisms must be inactivated in the sample, which is achieved through irreversible electroporation. PEF also consists of an attractive low energy-consuming treatment for the permeabilization of microalgae for the extraction (▶ Selective extraction of molecules by pulsed electric field treatment, Extraction of valuable compounds from microalgae using pulsed electric fields) of various cellular components such as the lipid biomass. As microorganisms and algae are much smaller than mammalian cells, higher voltages must be applied than for other applications, which are accompanied by a significant production of heat through Joule effects. In this context, the advantage of using microfluidic devices, which are characterized by higher surface-to-volume ratios, is linked to their ability to dissipate heat in a more efficient way. A shortcoming however of using microscale systems is their relatively low throughput since these applications require processing of large volumes.

Electroporation in Microfluidic Devices

17

Single-Cell Analysis Microfluidic technology shows great promises for the molecular analysis of single cells due to its ability to accurately manipulate individual cells and small-sized samples down to the low picoliter range (Le Gac and van den Berg 2010). To access a cell molecular content, the plasma membrane must be ruptured. To that end, electroporation, whether it is reversible or irreversible, offers specific advantages compared to conventional chemical approaches: it suppresses issues associated with sample dilution and contamination, which can hamper molecular analysis; it allows on-demand and targeted lysis of certain cells if electrodes are individually addressed; it is ideally suited to look at transient species, since electric pulses can rupture the plasma membrane within milliseconds against seconds to minutes for chemical processes; and the membrane poration process can be tuned to be either reversible and noninvasive or irreversible. When electroporation is combined with microfluidics, the retrieved cellular content can be confined in small volumes for further processing or analysis, in situ or off-line, without any extensive dilution and sample loss. For instance, capillary electrophoresis has been employed, in a proof-of-concept experiment, to separate two dyes previously loaded into the cells and that were detected by fluorescence (Han et al. 2003). Using droplet microfluidics, the entire content of individual cells can be retrieved in individual micrometer-sized reactors with limited dilution of the cell content. Using this approach, de Lange et al. successfully performed enzymatic assays on the content of individual E. coli cells to assess the activity of β-galactosidase at the single-cell level (de Lange et al. 2016). Finally, of particular interest for single-cell analysis is the so-called technique of “electroporative flow cytometry” (Bao et al. 2008), which relies on the selective extraction of molecules located at the periphery of the cytoplasm after reversible cell poration. Using this technique, controlled extraction of small molecules as well as proteins was reported without compromising the viability of the cells (Zhan et al. 2012).

Conclusion Cell electroporation can benefit considerably from its implementation in miniaturized and/or microfluidic devices. The whole process is better controlled, down to the single-cell level or even subcellular level, and the success yield is greatly enhanced to virtually reach 100%. Moreover, the risks associated with the treatment are diminished. Indeed, when the dimensions of the devices are reduced to the micrometer range, much lower voltages are required, and signals as low as < 1 V have led to successful permeabilization of the cell membrane. These low potential values make the devices not only safer to work with but also less energy-consuming while avoiding significant temperature changes induced by Joule heating. Furthermore, miniaturized devices bring enhanced protocols for molecular exchange, i.e., by assisting intracellular delivery using electrophoresis, for instance, and dosing the

18

S. Le Gac and I. van Uitert

amount of substances loaded in the cell or by confining the extracted cellular content into sub-nanoliter volumes for their analysis or further processing. Microfluidic electroporation devices can be classified into different categories depending on the way the cells are manipulated: whether they are individually trapped in dedicated microstructures, flown in a channel either as a single-cell file between two electrodes or as individual cells encapsulated in sub-nanoliter droplets. Interestingly, these different types of devices are complementary to each other. Devices from the first category are the best suited for treating single or rare cells and for following the response of each individual cell to the electric field, which is particularly interesting for fundamental studies on the process of electroporation. These devices are also particularly attractive for single-cell study using either imaging of intact and living cells or molecular analysis approaches after cell lysis and extraction of the cellular content. Flow-through devices better apply for middleto-high-throughput treatment of populations of cells. In these devices, cells can be characterized online before and after exposure to the electric field, and they can easily be retrieved from the device for off-line analysis or utilization. Promising applications of flow-through devices are the inactivation of bacteria and microorganisms for the food industry, the transformation of plant cells, or the treatment of algae.

Cross References: Related Chapters that May Be of Further Interest ▶ Critical electric field and transmembrane voltage for lipid pore formation in experiments ▶ Electroporation and Electropermeabilization ▶ Electroporation to study embryology morphogenesis and organogenesis ▶ Extraction of valuable compounds from microalgae using pulsed electric fields ▶ Gene delivery by electroporation in vitro: mechanisms ▶ Impedance measurements as electroporation measure ▶ Nanochannel electroporation: delivery of precise amounts of biomolecules into living cells ▶ Pore lifetime and permeabilization lifetime in models ▶ Principles of electroporation for gene therapy ▶ Pulsed electric field treatment for beverage production and preservation ▶ Pulsed electric fields treatment of biological suspensions ▶ Selective extraction of molecules by pulsed electric field treatment Tissue engineering with electroporation

Electroporation in Microfluidic Devices

19

References Bao N, Wang J, Lu C (2008) Microfluidic electroporation for selective release of intracellular molecules at the single-cell level. Electrophoresis 29(14):2939–2944. doi:10.1002/ elps.200700856 Baroud CN, Gallaire F, Dangla R (2010) Dynamics of microfluidic droplets. Lab Chip 10 (16):2032–2045. doi:10.1039/c001191f Becker H, Gartner C (2008) Polymer microfabrication technologies for microfluidic systems. Anal Bioanal Chem 390(1):89–111. doi:10.1007/s00216-007-1692-2 Chang LQ, Bertani P, Gallego-Perez D, Yang ZG, Chen F, Chiang CL, Malkoc V, Kuang TR, Gao KL, Lee LJ, Lu W (2016) 3D nanochannel electroporation for high-throughput cell transfection with high uniformity and dosage control. Nanoscale 8(1):243–252. doi:10.1039/c5nr03187g Chen JJ, Luo XQ, Wang J, Liao HH, Zhou WL, Zhang L, Han HY, Yu J (2008) Theoretical analysis of T-lymphocytes electroporation model. International Conference on Biomedical Engineering and Informatics. pp 688–690. doi:10.1109/Bmei.2008.149 de Lange N, Tran TM, Abate AR (2016) Electrical lysis of cells for detergent-free droplet assays. Biomicrofluidics 10(2):024114. doi:10.1063/1.4944742 Han FT, Wang Y, Sims CE, Bachman M, Chang RS, Li GP, Allbritton NL (2003) Fast electrical lysis of cells for capillary electrophoresis. Anal Chem 75(15):3688–3696. doi:10.1021/ ac0341970 Huang Y, Rubinsky B (1999) Micro-electroporation: improving the efficiency and understanding of electrical permeabilization of cells. Biomed Microdevices 2(2):145–150. doi:10.1023/ A:1009901821588 Huang Y, Rubinsky B (2001) Microfabricated electroporation chip for single cell membrane permeabilization. Senr Actuators A 89(3):242–249. doi:10.1016/S0924-4247(00)00557-4 Huang Y, Rubinsky B (2003) Flow-through micro-electroporation chip for high efficiency singlecell genetic manipulation. Sens Actuators A 104(3):205–212. doi:10.1016/S0924-4247(03) 00050-5 Ionescu-Zanetti C, Blatz A, Khine M (2008) Electrophoresis-assisted single-cell electroporation for efficient intracellular delivery. Biomed Microdevices 10(1):113–116. doi:10.1007/s10544-0079115-x Khine M, Lau A, Ionescu-Zanetti C, Seo J, Lee LP (2005) A single cell electroporation chip. Lab Chip 5(1):38–43. doi:10.1039/b408352k Kim SK, Kim JY, Kim KP, Chung TD (2007) Continuous low-voltage dc electroporation on a microfluidic chip with polyelectrolytic salt bridges. Anal Chem 79(20):7761–7766. doi:10.1021/ac071197h Le Gac S, van den Berg A (2010) Single cells as experimentation units in lab-on-a-chip devices. Trends Biotechnol 28(2):55–62. doi:10.1016/j.tibtech.2009.10.005 Lu H, Schmidt MA, Jensen KF (2005) A microfluidic electroporation device for cell lysis. Lab Chip 5(1):23–29. doi:10.1039/b406205a Mazari E, Zhao X, Migeotte I, Collignon J, Gosse C, Perea-Gomez A (2014) A microdevice to locally electroporate embryos with high efficiency and reduced cell damage. Development 141 (11):2349–2359. doi:10.1242/dev.106633 Qu B, Eu YJ, Jeong WJ, Kim DP (2012) Droplet electroporation in microfluidics for efficient cell transformation with or without cell wall removal. Lab Chip 12(21):4483–4488. doi:10.1039/ c2lc40360a Sukas S, Schreuder E, de Wagenaar B, Swennenhuis J, van den Berg A, Terstappen L, Le Gac S (2014) A novel side electrode configuration integrated in fused silica microsystems for

20

S. Le Gac and I. van Uitert

synchronous optical and electrical spectroscopy. Lab Chip 14(11):1821–1825. doi:10.1039/ c3lc51433a Sun C, Cao ZN, Wu M, Lu C (2014) Intracellular tracking of single native molecules with electroporation-delivered quantum dots. Anal Chem 86(22):11403–11409. doi:10.1021/ ac503363m Valero A, Post JN, van Nieuwkasteele JW, ter Braak PM, Kruijer W, van den Berg A (2008) Gene transfer and protein dynamics in stem cells using single cell electroporation in a microfluidic device. Lab Chip 8(1):62–67. doi:10.1039/b713420g van den Brink FTG, Gool E, Frimat JP, Bomer J, van den Berg A, Le Gac S (2011) Parallel singlecell analysis microfluidic platform. Electrophoresis 32(22):3094–3100. doi:10.1002/ elps.201100413 Wang HY, Lu C (2006) Electroporation of mammalian cells in a microfluidic channel with geometric variation. Anal Chem 78(14):5158–5164. doi:10.1021/ac060733n Wang HY, Lu C (2008) Microfluidic electroporation for delivery of small molecules and genes into cells using a common DC power supply. Biotechnol Bioeng 100(3):579–586. doi:10.1002/ bit.21784 Wang J, Bao N, Paris LL, Wang HY, Geahlen RL, Lu C (2008) Detection of kinase translocation using microfluidic electroporative flow cytometry. Anal Chem 80(4):1087–1093. doi:10.1021/ ac702065e Whitesides GM (2006) The origins and the future of microfluidics. Nature 442(7101):368–373. doi:10.1038/nature05058 Zhan YH, Wang J, Bao N, Lu C (2009) Electroporation of cells in microfluidic droplets. Anal Chem 81(5):2027–2031. doi:10.1021/ac9001172 Zhan YH, Sun C, Cao ZN, Bao N, Xing JH, Lu C (2012) Release of intracellular proteins by electroporation with preserved cell viability. Anal Chem 84(19):8102–8105. doi:10.1021/ ac302462s Zhu T, Luo CX, Huang JY, Xiong CY, Ouyang Q, Fang J (2010) Electroporation based on hydrodynamic focusing of microfluidics with low dc voltage. Biomed Microdevices 12 (1):35–40. doi:10.1007/s10544-009-9355-z

Detection of Electroporation in Microbial Cells: Techniques and Procedures Diego García-Gonzalo and Rafael Pagán

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Entrance of External Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fluorescent Probes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Antimicrobial Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Leakage of Intracellular Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electron and Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selective Medium Plating Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Measurement of Osmotic Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fourier Transform Infrared Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Molecular Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 4 5 7 8 9 9 11 11 12 13 14 14

Abstract

Application of an electric field to microbial cells can cause electroporation of their membranes, leading to permeabilization and formation of pores in these structures. Assessment of electroporation in microbial cells, such as fungi and bacteria, could be hindered by the presence of additional structures located externally to the cytoplasmic membrane, such as the cell wall and outer membrane. The most common methods to assess the electroporation of these membranes can be divided between those based in the entrance of external compounds into the bacterial cytoplasm (e.g., a fluorescent probe) and those which detect the leakage of intracellular compounds to the environment (e.g., nucleic acids, D. García-Gonzalo (*) • R. Pagán Tecnología de los Alimentos, Departamento de Producción Animal y Ciencia de los Alimentos, Facultad de Veterinaria, Instituto Agroalimentario de Aragón– IA2 (Universidad de ZaragozaCITA), Zaragoza, Spain e-mail: [email protected]; [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_137-1

1

2

D. García-Gonzalo and R. Pagán

proteins or ATP). Moreover, other techniques could be used to understand the mechanism of electroporation in microbial membranes, such as direct visualization of damages in cell membranes by electron and atomic force microscopy, evaluation of membrane damages by a selective medium plating technique, measurement of osmotic response, Fourier transform infrared spectroscopy, and molecular tools (including transcriptomics and site-directed mutagenesis). In addition, several methods are proposed in order to avoid interference of the cell wall and/or outer membrane in the evaluation of cytoplasmic membrane permeabilization: (pre)treatment of cells with cationic agents or chelators and preparation of protoplasts or spheroplasts. The described methods to detect electroporation of bacterial cells have their own advantages and disadvantages. Therefore, analysis of results obtained with different methods is recommended to provide a more accurate knowledge in the mechanism of microbial permeabilization by electroporation.

Keywords

Outer membrane • Peptidoglycan • Irreversible and reversible permeabilization • Fluorescent probes • Leakage of intracellular material • Sublethal injuries • DNA microarrays • Site-directed mutagenesis

Introduction The effects of electric fields in cell membranes have been extensively documented, due to their extended use in cell biology, biotechnology, or medicine (Prasanna and Panda 1997). Although the exact molecular mechanisms are not fully understood, it is known that application of an electric field to cell membranes can cause electroporation, involving the formation of pores on the membranes of cells and organelles. Depending on the intensity of the applied electric field, electroporation can induce the formation of transitory and/or permanent pores leading to reversible or irreversible electroporation, respectively (Teissie et al. 2005). Reversible pores are resealed when the electrical field ceases, and the electrically stimulated cells survive. However, at high intensities, permeabilization of cell membranes might be irreversible and even cell membranes and other structures might break down. Under these circumstances, electric fields cause permanent membrane damage and cell death which is the basis for the success of pulsed electric fields (PEF) as a novel processing method for food preservation. In addition, application of electric fields has also been proposed as an excellent method to extract compounds of interest from the cytoplasm of bacteria and unicellular organisms (Luengo et al. 2015). The promising use of PEF as a nonthermal process for food preservation resides in its potential to inactivate microbial cells without altering sensorial and nutritional food properties. The design of effective PEF treatments which provide safe food with a prolonged shelf life requires, among others, knowledge of how PEF kill microorganisms (i.e., the mechanism of microbial inactivation by PEF). A better

Detection of Electroporation in Microbial Cells: Techniques and Procedures

3

understanding of this mechanism would help to define PEF treatments as alternatives to traditional heat preservation. Knowledge of the mechanisms of electroporation has been mostly obtained using artificial membranes, theoretical models, and eukaryotic or mammalian cells. However, its occurrence and relation to microbial inactivation by PEF is less understood. In contrast to ideal systems, such as artificial membranes or mathematical models, biological membranes are complex and dynamic systems: their composition and structure could be modified as a response to stressing conditions, such as temperature variations or exposure to chemical compounds. In addition, microorganisms, such as fungi and bacteria, have additional structures located externally to the cytoplasmic membrane (Neidhardt et al. 1990). For the sake of simplicity, this chapter will focus on the techniques used to detect electroporation in bacteria, although most of the techniques could be easily adapted to fungal cells. Composition of bacterial membranes differs between species and strains, being highly noticeable between Gram groups. Gram staining technique is based on different chemical and physical properties of peptidoglycan present in cell walls: after a washing step, Gram-positive cells retain crystal violet dye, but the dye is washed out of Gram-negative cells. In the next counterstaining step, safranin only stains the washed Gram-negative cells. This differential staining is due to the presence of the outer membrane in Gram-negative cells. In Gram positives, outside the cytoplasmic membrane, there is a cell wall formed by several peptidoglycan layers and teichoic acids (Fig. 1), which confers rigidity and physical resistance to the bacterial cell. In Gram negatives, the cell wall is thinner, but surrounded by a protective outer membrane (Fig. 1). The outer membrane differs from the cytoplasmic membrane since its external leaflet is made up of lipopolysaccharides (LPS) instead of phospholipids, and it confers special resistance against the entrance of some molecules such as antimicrobial compounds, bile salts, and some fluorescent probes. In both Gram groups, but more especially in Gram negatives, these external structures create a zone outside the

Fig. 1 Structure of cell membranes of Gram-positive and Gram-negative bacteria: (1) the outer membrane with lipopolysaccharides, (2) cell wall with peptidoglycan, (3) cytoplasmic membrane

4

D. García-Gonzalo and R. Pagán

cytoplasmic membrane with a different chemical composition from the environment, known as periplasmic space. This particular disposition of bacterial membranes has to be taken into account for the evaluation of electroporation. For example, PEF could cause electroporation of the cytoplasmic membrane, but not of the outer membrane, and consequently access of some fluorescent probes to the cytoplasm could be hindered, as described for Gram-negative cells at acid pH (Garcia et al. 2007). Maintenance of the integrity and functionality of cell membranes is of key importance for bacterial cells. First of all, these structures protect microorganisms from the external conditions acting as a semipermeable selective filter. In addition the cytoplasmic membrane is involved in other cellular processes such as the control of the passage in both directions of small molecules, nutrients, and end products of metabolic activities; synthesis of the RNA, protein, and cell wall; control of DNA synthesis; and electron transport and oxidative phosphorylation (Neidhardt et al. 1990). Thus, the cytoplasmic membrane controls the cellular metabolic activities by maintaining an effective osmotic boundary between the cell interior and the surrounding environment. Any damage in the integrity or functionality of this membrane could impair one or more of these processes, resulting in bacterial death. This chapter presents an overview of the techniques and procedures used to provide direct and indirect evidences of electroporation in microbial cells. The presented methods could detect irreversible and/or reversible permeabilization as well as modifications in cell envelopes as a consequence of PEF treatments. Considering that cell membranes separate the cytoplasm from the outside environment, the methods used to assess the electroporation of these membranes can be divided into (1) those based in the entrance of external compounds into the bacterial cytoplasm (i.e., a fluorescent probe) and (2) those which detect the leakage of intracellular compounds to the environment (i.e., nucleic acids). In addition, there are other excellent methods, highlighting (3) direct visualization of damages in cell membranes by electron microscopy, (4) evaluation of membrane damages by a selective medium plating technique, and (5) measurement of osmotic response of PEF-treated cells or (6) Fourier transform infrared spectroscopy. These techniques could be complemented by (7) molecular tools, such as directed mutagenesis or transcriptomics, to understand the mechanism of electroporation in microbial membranes.

Entrance of External Compounds As compared to evaluation of leakage of intracellular compounds, the determination of the access of external compounds to the cytoplasm is easier because of the possibility to modify the concentration of the latter compounds. Consequently, techniques based on the detection of foreign substances in the bacterial cytoplasm have been widely used to measure electroporation efficiency and quantify electroporated cells. Two main classes of substances have been used with this objective: fluorescent probes and antimicrobial agents.

Detection of Electroporation in Microbial Cells: Techniques and Procedures

5

Fluorescent Probes The measurement of the increased uptake of a fluorescent dye, which normally does not penetrate the membrane of healthy cells, is a direct and sensitive technique to evaluate membrane integrity and/or functionality. In addition, this methodology could allow for single-cell analysis and identification of permeabilized cell fraction by the use of fluorescence microscopy or flow cytometry (FCM). FCM assists in the study of microbial behavior and responses to PEF with single-cell resolution, since it quantifies small differences between cell populations analyzing thousands of cells in seconds. A great advantage of fluorescent probes is the delivery of rapid and sensitive results. Even more, in comparison to radioactive probes, these probes present a lower toxicity for lab operators. The most used fluorescence-based methods to evaluate electroporation are based on the addition to the treatment medium of cell-impermeant nucleic acid stains, such as propidium iodide (PI), cyanine monomers, or SYTOX ® dyes. PI is a small (660 Da), strongly hydrophilic molecule, showing a high sensitivity and potential to provide quantitative results by using a spectrofluorometer, to identify stained cells by using an fluorescence microscope and to analyze individual cells in combination with FCM or image analysis. PI has been commonly used to assess membrane electroporation of PEF-treated microorganisms (Wouters et al. 2001; Aronsson et al. 2005; Garcia et al. 2007; Jaeger et al. 2009; Sagarzazu et al. 2013). Moreover, PI could be added to the treatment medium before and after PEF treatments (Fig. 2) in parallel experiments to detect reversible or irreversible permeabilization (Garcia et al. 2007). PI has also been used in combination with other probes to successfully identify sublethally injured populations in Escherichia coli and Lactobacillus plantarum PEF-treated cells (Jaeger et al. 2009; Zhao et al. 2011). For future research, FCM might also be used to assess heterogeneity in stress survival, including damage repair capacity and PEF targets. Ethidium bromide (EtBr) can also be used to evaluate electroporation, although it is less reliable than PI, because PI binds to naked DNA with a higher affinity than EtBr. This is because PI is doubly charged as compared to EtBr. However, this disadvantage could be overcome by the use of ethidium homodimer-1, with high affinity for DNA (it binds to nucleic acids 1,000 times more tightly than does EtBr, and its fluorescence increases 40-fold upon binding) and low membrane permeability in intact cells (Johnson and Spence 2010). Due to its high affinity, low concentrations needed for fluorescence experiments lead to the use of small quantities of potentially hazardous compounds. Cyanine monomers, such as TO-PRO ® dyes, are among the best fluorescent probes available for nucleic acid staining because of their high affinity for nucleic acids and high fluorescence enhancement and quantum yield upon binding (Johnson and Spence 2010). In addition, cyanine monomers have narrow emission bandwidths, making them a good choice for multicolor experiments. For example, in FCM applications, the complex formed by nucleic acids with TO-PRO ®-3 can be directly excited by the He-Ne red laser. The binding affinity to dsDNA, RNA, and

6

D. García-Gonzalo and R. Pagán

Fig. 2 Protocol for propidium iodide (PI, ) staining of microorganisms to detect reversible and irreversible electroporation. Addition of PI after electroporation detects only cells with irreversible permeabilization of their membranes. However, the presence of PI during the treatment allows for detecting not only irreversible cell permeabilization but also the presence of reversible pores that are repaired immediately after the treatment. Addition of PI at different times after electroporation (e.g., immediately, 1 min, 5 min, etc.) would also allow for the determination of cells that are repairing their pores after the treatment, i.e., the dynamics of reversibility of pores. It should be noted that this protocol could also be used with other membrane-impermeant probes

ssDNA of TO-PRO ® dyes is lower than that of other dyes, with lower fluorescence quantum yields. SYTOX® nucleic acid stains are also good cell-impermeant dyes to assess electroporation because they show high affinity to nucleic acids and little base selectivity. They are especially useful for staining both Gram-negative and Gram-positive bacteria, providing a bright green, blue, or orange signal (Johnson and Spence 2010). These dyes could be used simultaneously with cell-permeant SYTO® stains (blue, green, orange, or red), which passively diffuse through membranes and show a lower affinity to nucleic acids than SYTOX ® dyes. The aforementioned probes have been widely used to assess membrane permeabilization caused by antimicrobial compounds, such as antibiotics. However, it should be taken into account that the action of most antimicrobials, such as disinfectants, is inward to the cytoplasm, i.e., they should pass through the outer membrane and/or cell wall to access the cytoplasmic membrane. However, PEF might cause the electroporation of the cytoplasmic membrane without altering the outer membrane, or the size of pores caused by PEF in the outer membrane might be smaller than the size of the probes and thus their pass through the outer membrane would be prevented. Under these circumstances, it is difficult to evaluate electroporation of the cytoplasmic membrane by these fluorescent techniques: it can be concluded that all stained cells are permeabilized, but not that all permeabilized cells are stained and can be detected. A possible solution could be to permeabilize the outer membrane by cationic agents (e.g., polymyxin derivatives, lysine polymers, or benzalkonium

Detection of Electroporation in Microbial Cells: Techniques and Procedures

7

chloride) or chelators (e.g., ethylenediaminetetraacetic acid (EDTA), sodium hexametaphosphate, or acetylsalicylate), although these compounds could also alter the cytoplasmic membrane. Alternatively, bacteria with mutations disrupting the outer membrane permeability barrier, such as E. coli lptD4213 (Ruiz et al. 2006), could be used to prevent interferences of the outer membrane in the assessment of cytoplasmic membrane permeabilization by fluorescent techniques. Moreover, prevention from protective role of the cell wall in the assessment of cytoplasmic membrane permeabilization of Gram-positive cells could be achieved by preparation of protoplasts. Treatment of cells with lysozyme degrades the peptidoglycan component of cell walls, resulting in the formation of protoplasts. Since protoplasts are osmotically sensitive, cell wall digestion, storage, and further experiments must be done in an isotonic solution. The method to obtain protoplasts from Gram-negative cells is complicated by the presence of the outer membrane which prevents access of lysozyme to the cell wall. As a consequence, EDTA should be added to the isolation medium as a destabilizer of the outer membrane. This technique is not capable to totally remove the outer membrane and the cell wall is only partially degraded, and so the usual products obtained from Gram-negative cells are spheroplasts (with spherical shape) rather than protoplasts. On the other hand, permeabilization of the outer membrane could be evaluated using probes that are usually excluded by this membrane, but once the integrity of the outer membrane is compromised, they can access the cytoplasmic membrane or the bacterial cytoplasm. Most used probes for this purpose are 1-N-phenyl-naphtalene (NPN) and hexidium iodide (HI). NPN (currently found as N-phenyl-1-naphthylamine) is a hydrophobic fluorescent probe which binds to hydrophobic regions of the bacterial cytoplasmic membrane: it fluoresces weakly in aqueous environments, but becomes very strongly fluorescent in nonpolar or hydrophobic environments. HI is a cell-permeant nucleic acid dye which selectively stains Gram-positive bacteria because it is blocked by the LPS layer of Gram-negative bacteria and thus only permeable to Gram-positive bacteria and Gram-negative bacteria with an altered LPS layer. Therefore, the increase of fluorescence upon addition of NPN or HI to PEF-treated Gramnegative cells would indicate the electroporation of the outer membrane.

Antimicrobial Agents This method is based on the uptake of antimicrobials into the cytoplasm or the cytoplasmic membrane which normally would not access to these structures. The increase of sensitivity of the target cell to lysozyme or to hydrophobic antibiotics, such as actinomycin, is an indirect assay to determine electroporation of the outer membrane (Sánchez-Gómez et al. 2008). In this context bile salts could be included, but the use of these compounds will be commented in section “Selective Medium Plating Technique.” Lysozyme damages peptidoglycan layers of the cell wall, and actinomycin inhibits transcription by binding the DNA at the transcription initiation complex and preventing elongation of RNA chain by RNA polymerase in Gram-

8

D. García-Gonzalo and R. Pagán

positive bacteria. However, their access to cell targets is hindered by the outer membrane in Gram-negative cells. Consequently, sensitization of PEF-treated Gram-negative cells to these antimicrobial agents would demonstrate the permeabilization of the outer membrane.

Leakage of Intracellular Compounds The most commonly used techniques to evaluate the presence of intracellular material outside the cell are the measurement of UV-absorbing material such as nucleic acids and proteins at 260 and 280 nm, respectively, and the presence of adenosine triphosphate (ATP) by luciferin–luciferase assay. The sensitivity of these methods could be increased by the use of fluorescent probes (such as nucleic acid stains mentioned above) which bind to leaked compounds. However, the occurrence of various degrees of cell permeabilization complicates to find a relationship between the amount of intracellular components outside the cell and the number of inactivated cells (Wouters et al. 2001; Aronsson et al. 2005). Furthermore, this technique would not allow for quantification of electroporated cells. An alternative method with a higher sensitivity than detecting cell components in the environment is the evaluation of leaked fluorescent probes that have been previously introduced in the bacterial cytoplasm. For this purpose, esterase substrates could be used, such as calcein acetoxymethyl ester (calcein AM), 20 ,70 -bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein acetoxymethyl ester (BCECF AM), and carboxyfluorescein diacetate (CFDA). These nonfluorescent substrates freely diffuse into the cytoplasm (cell permeants) where they are converted by intracellular esterases into fluorescent compounds which are retained by cells with intact cytoplasmic membranes (Johnson and Spence 2010). Therefore, these probes assess enzymatic activity (for hydrolysis of incorporated substrates) and integrity of the cytoplasmic membrane (for intracellular retention of fluorescent products). As explained in the previous section, these probes may also be combined with other probes (e.g., with PI) for multicolor experiments. Calcein AM is usually the first choice because of its notable cell retention and its stability at physiological pH. Calcein is well retained in viable cells due to its six negative charges and two positive charges at pH 7. Other chemical variations with different emission wavelengths can be found in the market: calcein blue AM, calcein violet AM, and calcein red-orange AM. Hydrolysis of BCECFAM leads to formation of BCECF with four to five negative charges. These charges improve its cell retention, but the emission intensity is reduced at acid pH, making it less optimal for some applications. CFDA, originally conceived as a probe to measure intracellular pH but promptly adapted as an indicator of cell viability and membrane integrity, forms carboxyfluorescein upon hydrolysis. Extra negative charges as compared with fluorescein improve cell retention of carboxyfluorescein, making it preferred for these assays than fluorescein diacetate.

Detection of Electroporation in Microbial Cells: Techniques and Procedures

9

Electron and Atomic Force Microscopy Because of its higher resolving power, electron microscopy has been used to examine morphological changes in cells, either at cytoplasmic organelles or cell envelopes, after PEF treatments. With regard to membrane damage, the technique allows detecting pores from 1 nm in diameter. However, it should be noted that it would not always allow observing lysed cells since their debris might be removed during the washing and cell concentration steps. Furthermore, PEF-induced pores smaller than 1 nm would be undetected. The use of this technique has not confirmed any relationship between membrane damage and microbial inactivation by PEF because frequency of morphologically damaged cells did not correspond to the log reductions of viability. Nevertheless, electron microscopy could be useful as a complementary technique to gain insight into PEF action, by revealing effects not only in cell membranes but also in intracellular modifications. Atomic force microscopy could also be used to visualize cell changes. This technique provides a three-dimensional image in real time of living cells with a resolution of fractions of a nanometer, reflecting modifications in the ultrastructure of the electroporated cells (Pillet et al. 2016). Apart from surface topography, atomic force microscopy provides information related to electrical and mechanical properties, such as surface potential or adhesion strength.

Selective Medium Plating Technique Membrane damage measured using a selective medium plating technique includes the loss of both membrane integrity and functionality. The method consists of plating survivors after treatments into two culture media: a nonselective one, which allows cells to repair sublethal damages and recover, and a selective one, in which survivors are not capable of repairing their damages and finally die (Mackey 2000). Those membrane damages that are repairable under suitable conditions and detected following this methodology are commonly called sublethal injuries. As illustrated in Fig. 3, sublethally injured cells in their membrane(s) are estimated by the difference in the number of survivors obtained after plating treated cells in both nonselective and selective media. The most common selective media used to detect damage to the cytoplasmic membrane is agar medium with sodium chloride added (Mackey 2000). In addition, bile salts can be used in order to detect sublethal damages in the outer membrane of Gram-negative bacteria. As noted for hydrophobic antibiotics, the outer membrane is also an excellent barrier to membrane active agents such as bile salts, with LPS playing a major role in resistance to bile salts. The loss of resistance to sodium chloride and bile salts is likely to be multifactorial. However, sensitivity to sodium chloride implies a loss of osmotic functionality of the cytoplasmic membrane, whereas sensitivity to bile salts reflects the loss of outer membrane integrity and/or impairment of multidrug efflux systems by the loss of proton motive force (Thanassi et al. 1997; Mackey 2000; Begley et al. 2005).

10

D. García-Gonzalo and R. Pagán

Fig. 3 Fundamentals of selective medium plating technique for detecting sublethal damages in microbial membranes. After an electroporating treatment, there could be alive (intact), dead (inactivated), and cells with sublethal injures in their membranes (damaged). Plating these treated cells in a nonselective medium allows the recovery of both alive and sublethally injured cells. However, if these cells are plated in a selective recovery medium, only alive cells will produce visible colonies. For example, if 1 mL of a microbial suspension at 109 CFU/mL is PEF treated and survival counts in nonselective medium are 108 CFU/mL, the treatment caused 1 log10 of inactivation. If survival counts of the same treated sample in a selective medium with sodium chloride are 106 CFU/mL would indicate that 2 log10 cycles of the survivors to PEF are sublethally damaged in their cytoplasmic membranes. This methodology could be performed with higher or lower cell concentrations

Sublethally injured bacteria in their cytoplasmic membrane after PEF treatments were firstly detected by the use of this technique. Moreover, the growth of sublethally injured cells in nonselective media indicated that, in some cases, damages

Detection of Electroporation in Microbial Cells: Techniques and Procedures

11

inflicted in cell membranes by PEF are repairable (Garcia et al. 2006). The addition of inhibitors such as chloramphenicol, cerulenin, penicillin G, rifampicin, and sodium azide to the repair medium would indicate the biosynthetic requirements to repair these damages in the cell envelopes. In comparison to the previous mentioned techniques, it is more time consuming and requires at least 48 h to get the final results. Nevertheless, this technique shows an extraordinary sensitivity because it can be used to evaluate the damages in treated suspensions with a high number of bacterial cells (e.g., 109 CFU/mL), allowing the detection of several log cycles of cells with sublethal damages in their membranes. Moreover, the recovery of cells in media with different proportions of selective agents would allow the identification of several degrees of membrane damages.

Measurement of Osmotic Response This rapid and inexpensive technique evaluates alterations in membrane integrity through the loss of the ability of the cell to plasmolize. When a cell with an undamaged membrane is suspended in a hypertonic medium, such as a strong salt or sucrose solution, water diffuses from the cytoplasm to the external medium. This initial passive response causes gross morphological changes with the cytoplasm shrinking and the plasma membrane contracting from the cell wall, leading to a strong condensation of the cell material, which can be measured through the increase in optical density [usually at 600 nm (OD600)] of cell suspensions (Korber et al. 1996). On the contrary, in cells whose membrane is altered, salt and water diffuse freely through the membrane, and no response is detected (i.e., no changes in OD). Additionally, FCM and microscopic methods, such as low-magnification dark-field microscopy, highmagnification phase-contrast microscopy, and scanning confocal laser microscopy, used with or without image analysis are suitable to determine membrane integrity through direct visualization and/or the measurement of the cell area.

Fourier Transform Infrared Spectroscopy Fourier transform infrared (FTIR) spectrum in mid-infrared range provides bands from all the cellular components of microorganism (e.g., cell membrane and wall components, proteins, and nucleic acids). Therefore, the obtained spectral signatures or “fingerprints” can be used for detection, classification, and identification of microorganisms (Alvarez-Ordóñez et al. 2011). Thus, FTIR is a suitable tool to detect the changes occurring in bacterial cells and indicate the physiological state in response to electroporation. Actually, most of the studies assessing the effects of stressing conditions on the bacterial biochemical composition of bacterial cells by FTIR have been focused on the examination of the membrane properties, such as changes in the membrane phase behavior, the presence of longer and/or more saturated acyl chains, and structural changes of membrane phospholipids. Similarly, this method indicated that PEF treatments would mainly target LPS cell fraction or

12

D. García-Gonzalo and R. Pagán

outer membrane proteins of E. coli, as a function of the treatment pH (Gelaw et al. 2014).

Molecular Tools Molecular microbiology tools are at the vanguard of most cutting-edge scientific research because they allow investigating and manipulating microbial cells at a molecular level. Among the high-throughput techniques that are available, both DNA microarrays and site-directed mutagenesis methods could be most useful to study permeabilization of microbial cells. DNA microarrays provide information at a molecular level: transcriptome analysis of cells (i.e., all RNA molecules in the bacterial population) under different conditions may reveal differences in factors contributing to PEF resistance and damage repair systems. A DNA microarray contains thousands of microspots on a glass slide. Each spot contains identical single-stranded DNA oligonucleotide-capturing probes, whose locations are fixed during the process of hybridization and detection. This highthroughput technique can be used to measure changes in gene expression after exposure to PEF. RNA will be isolated from the samples at preset times and convert it into cDNA. Each DNA oligonucleotide spot represents a specific gene sequence to which cDNA will bind by complementarity. After labeling, a microarray reader will provide the intensity of signal per gene that can be compared between untreated and electroporated cells. Thus, the transcriptome at the moment of the RNA isolation will be inferred. By comparing the gene expression profiling before and after electroporation would reveal sets of genes upregulated or downregulated and subsequently the proteins activated and repressed by PEF (Chueca et al. 2015). This knowledge will allow to better understand the mechanism of inactivation and the mechanisms of resistance that the cell uses to cope with PEF treatments by the identification and selection of specific targets for further characterization, e.g., assessment of their role (s) in stress adaptation and damage repair. While DNA microarrays will yield a vast amount of data that will undoubtedly give insight into the electroporation of bacterial membranes, one major limitation is its inability to detect the cells within the bacterial population that are differentially expressing a gene. Therefore, the combination of FCM with transcriptional fusions using fluorescent proteins (e.g., green fluorescent protein controlled by a promoter which activity is modulated by electroporation) would show those cells affected by electroporation. Construction of deletion mutants in selected genes or gene clusters (site-directed mutagenesis) allows the creation of microbial cells with a modified composition and structure of membranes. As explained above, a mutation in lptD gene of E. coli (Ruiz et al. 2006) altered the outer membrane, avoiding the potential interferences of the outer membrane in electroporation. Moreover, candidate genes which could play a role in membrane permeabilization could be modified or deleted to assess the influence of certain proteins and/or structures in cell electroporation. Deletion mutants in E. coli can be constructed using P1 phage transduction derived from an E. coli single-gene knockout library (Baba et al. 2006). Positive P1 transductants would be confirmed by

Detection of Electroporation in Microbial Cells: Techniques and Procedures

13

Table 1 Summary chart with described procedures for detection of electroporation in microbial cells with their main advantages and disadvantages Procedure Entrance of external compounds

Fluorescent probes

Antimicrobial compounds

Leakage of intracellular compounds Electron and atomic force microscopy

Selective medium plating technique

Measurement of osmotic response Fourier transform infrared spectroscopy Molecular tools

Advantages Rapid and sensitive results Possibility of single-cell analysis Easy and fast technique

Rapid results Sensitivity could be increased with fluorescent probes Visualization of morphological changes High resolution (1 nm)

Easy and fast technique Extraordinary sensitivity (allows for discrimination between 99 and 99.99 % of damaged cells) Rapid and inexpensive technique Detection of biochemical changes in PEF-treated cells Identification of metabolic pathways for cell survival and repair Discovery of resistance mechanisms to PEF

Disadvantages Some probes are potentially hazardous Permeabilization of all cell envelopes is required Normally incubation of plates (24 h) is required Indirect evidence of permeabilization Difficult correlation between leaked components and permeabilized cells Difficult identification of all permeabilized cells Extensive training and special equipment are required Normally, it is more time consuming, because a long incubation of plates (48 h) is required Low sensitivity Extensive training and special equipment are required Extensive training and special equipment are required Indirect evidence of permeabilization

acquisition of kanamycin resistance and polymerase chain reaction (PCR). Subsequently, the kanamycin-resistance cassette would be removed using the previously described pCP20 plasmid and confirmed by PCR. Effect of mutations could be initially studied by comparing survival counts in differential plating media between wild type and mutant strains after exposure to PEF. Next, their transcriptomes could be compared by DNA microarrays after PEF treatments. These important tools will help to identify not only cellular targets but also to discover (novel) stress-resistance mechanisms.

Conclusion In summary, there is a wide variety of available techniques and methods to detect and evaluate the electroporation of microbial cells (Table 1).

14

D. García-Gonzalo and R. Pagán

Since they evaluate different membrane modifications, all of them have their limitations in assessing cell electroporation. Consequently, analysis of results obtained with different methods could provide a more accurate picture of membrane permeabilization. Moreover, it should be noted that the techniques and methods described in this chapter have already been used to confirm the occurrence and relation of the electroporation to microbial inactivation by PEF. Some techniques such as electron microscopy examinations or the measurement of leakage of intracellular material have allowed obtaining visual or direct evidence of the phenomenon; others, such as the measurement of dye exclusion/uptake and selective medium plating technique, have allowed finding out correlation between electroporation and microbial inactivation. Available molecular tools will be extremely valuable to increase our knowledge on the mechanism of electroporation and consequently complement our knowledge in the mechanism of bacterial inactivation by PEF. Acknowledgments Authors thank to CICYT (Spanish Interministerial Commission of Science and Technology, Projects AGL2012-32165 and AGL2015-69565-P), European Social Fund, and Aragonese Departamento de Ciencia, Tecnología y Universidad for their financial support.

Cross-References ▶ Atomic Force Microscopy for Electroporation Mechanisms Studies in Bacteria ▶ Bacteria Cell Wall: Description, Role in Transport, and Effect of Electroporation ▶ Electropore Formation in Heterogeneous Lipid Bilayers ▶ Experimental Determination of Lipid Electropore Size ▶ Fluorescent Indicators of Membrane Permeabilization Due to Electroporation ▶ Membrane Permeabilization Lifetime in Experiments

References Alvarez-Ordóñez A, Mouwen DJM, López M, Prieto M (2011) Fourier transform infrared spectroscopy as a tool to characterize molecular composition and stress response in foodborne pathogenic bacteria. J Microbiol Methods 84:369–378 Aronsson K, Rönner U, Borch E (2005) Inactivation of Escherichia coli, Listeria innocua and Saccharomyces cerevisiae in relation to membrane permeabilization and subsequent leakage of intracellular compounds due to pulsed electric field processing. Int J Food Microbiol 99:19–32 Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko KA, Tomita M, Wanner BL, Mori H (2006) Construction of Escherichia coli k-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol 2:2006.0008 Begley M, Gahan CGM, Hill C (2005) The interaction between bacteria and bile. FEMS Microbiol Rev 29:625–651 Chueca B, Pagán R, García-Gonzalo D (2015) Transcriptomic analysis of Escherichia coli MG1655 cells exposed proto pulsed electric fields. Innov Food Sci Emer Technol 29:78–86 Garcia D, Mañas P, Gomez N, Raso J, Pagan R (2006) Biosynthetic requirements for the repair of sublethal membrane damage in Escherichia coli cells after pulsed electric fields. J Appl Microbiol 100:428–435

Detection of Electroporation in Microbial Cells: Techniques and Procedures

15

Garcia D, Gomez N, Mañas P, Raso J, Pagan R (2007) Pulsed electric fields cause bacterial envelopes permeabilization depending on the treatment intensity, the treatment medium pH and the microorganism investigated. Int J Food Microbiol 113:219–227 Gelaw TK, Espina L, Pagán R, García-Gonzalo D, De Lamo-Castellví S (2014) Prediction of injured and dead inactivated Escherichia coli O157:H7 cells after heat and pulsed electric field treatment with attenuated total reflectance infrared microspectroscopy combined with multivariate analysis technique. Food Bioprocess Tech 7:2084–2092 Jaeger H, Schulz A, Karapetkov N, Knorr D (2009) Protective effect of milk constituents and sublethal injuries limiting process effectiveness during pef inactivation of Lb. rhamnosus. Int J Food Microbiol 134:154–161 Johnson I, Spence MTZ (2010) Molecular probes handbook: a guide to fluorescent probes and labeling technologies, 11th edn. Life Technologies, Carlsbad Korber DR, Choi A, Wolfaardt GM, Caldwell DE (1996) Bacterial plasmolysis as a physical indicator of viability. Appl Environ Microbiol 62:3939–3947 Luengo E, Martínez JM, Bordetas A, Álvarez I, Raso J (2015) Influence of the treatment medium temperature on lutein extraction assisted by pulsed electric fields from Chlorella vulgaris. Innov Food Sci Emerg Technol 29:15–22 Mackey BM (2000) Injured bacteria. In: Lund BM, Baird-Parker TC, Gould GW (eds) The microbiological safety and quality of food. Aspen Publisher Inc, Gaithersburg, pp 315–341 Neidhardt FC, Ingraham JL, Schaechter M (1990) Structure and function of bacterial cells parts. In: Neidhardt FC, Ingraham JL, Schaechter M (eds) Physiology of the bacterial cell. A molecular approach. Sinauer Associates, Inc. Publishers, Sunderland, p 30 Pillet F, Formosa-Dague C, Baaziz H, Dague E, Rols MP (2016) Cell wall as a target for bacteria inactivation by pulsed electric fields. Sci Rep 6:19778 Prasanna GL, Panda T (1997) Electroporation: basic principles, practical considerations and applications in molecular biology. Biotechnol Bioeng 16:261–264 Ruiz N, Kahne D, Silhavy TJ (2006) Advances in understanding bacterial outer-membrane biogenesis. Nat Rev Microbiol 4:57–66 Sagarzazu N, Cebrian G, Pagan R, Condon S, Mañas P (2013) Emergence of pulsed electric fields resistance in Salmonella enterica serovar Typhimurium SL1344. Int J Food Microbiol 166:219–225 Sánchez-Gómez S, Lamata M, Leiva J, Blondelle SE, Jerala R, Andrä J, Brandenburg K, Lohner K, Moriyón I, Martínez-de-Tejada G (2008) Comparative analysis of selected methods for the assessment of antimicrobial and membrane-permeabilizing activity: a case study for lactoferricin derived peptides. BMC Microbiol 8:1–9 Teissie J, Golzio M, Rols MP (2005) Mechanisms of cell membrane electropermeabilization: a minireview of our present (lack of ?) knowledge. Biochim Biophys Acta 1724:270–280 Thanassi DG, Cheng LW, Nikaido H (1997) Active efflux of bile salts by Escherichia coli. J Bacteriol 179:2512–2518 Wouters PC, Bos AP, Ueckert J (2001) Membrane permeabilization in relation to inactivation kinetics of Lactobacillus species due to pulsed electric fields. Appl Environ Microbiol 67:3092–3101 Zhao W, Yang RJ, Zhang HQ, Zhang WB, Hua XA, Tang YL (2011) Quantitative and real time detection of pulsed electric field induced damage on Escherichia coli cells and sublethally injured microbial cells using flow cytometry in combination with fluorescent techniques. Food Control 22:566-573

The Use of Electroporation in Developmental Biology C. Gosse, X. Zhao, I. Migeotte, D. Suárez-Boomgaard, I. Hue, S. Degrelle, A. Perea-Gomez, and E. Mazari

Abstract

During the formation of a complex organism, cells divide, die, migrate, and differentiate. Biologists have established tools to observe those phenomena but also to change their course, which subsequently enables to infer causal relationships between various events occurring in different cell groups. More precisely, present approaches mostly rely on modifications of gene expression. For instance, cells are labeled with fluorescent proteins and tracked within the embryo, C. Gosse (*) • X. Zhao Laboratoire de Photonique et de Nanostructures, LPN-CNRS, Marcoussis, France e-mail: [email protected]; [email protected] I. Migeotte • D. Suárez-Boomgaard Institut de Recherche Interdisciplinaire en Biologie Humaine et Moléculaire, Université Libre de Bruxelles, Bruxelles, Belgium e-mail: [email protected]; [email protected] I. Hue Biologie du Développement et Reproduction, UMR 1198 INRA, ENVA, Université Paris-Saclay, Jouy-en-Josas, France e-mail: [email protected] S. Degrelle Physiopathologie et Pharmacotoxicologie Placentaire Humaine, INSERM UMR-S1139, Université Paris Descartes, Sorbonne Paris Cité, Paris, France e-mail: [email protected] A. Perea-Gomez Institut de Biologie Valrose, CNRS UMR7277, Inserm U1091, Université Nice Sophia Antipolis, Nice, France e-mail: [email protected] E. Mazari Center for International Research on MicroMechatronics, Institute of Industrial Science, The University of Tokyo, Tokyo, Japan e-mail: [email protected] # Springer International Publishing AG 2017 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_139-1

1

2

C. Gosse et al.

molecular signals are switched on and off to perturb regulatory pathways. Importantly, in all those experiments, the exogenous genetic material must be delivered at the right place and with the appropriate timing: requirements that can both be fulfilled by electroporation. After 15 years of constant refinement, this technique has now superseded methods like viral infection, microinjection, and lipofection. Applications encompass a large number of model organisms, targeted anatomical structures, and molecular biology techniques. Keywords

Embryogenesis • Organogenesis • Cell labeling • Genetic engineering • Numerical simulations • Microfabrication

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Specificities Related to the Scientific Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Typical Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Organisms and Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Molecular Biology Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Competing Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Strategies to Spatially Restrict Transfection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Orientation of the Embryo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Containment of the Nucleic Acids Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Generation of an Electric Field Gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Combination of both Molecular Localization and Electric Field Focusing . . . . . . . . . . . . . . . . . Engineering Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Setup Conception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Material Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reagents Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Performances Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 3 5 8 10 13 13 15 16 17 18 18 21 23 24 27 29 29

Introduction The use of electroporation to study embryogenesis and organogenesis was initiated around 1995, and most of the associated technical developments took place in the following decade (Nakamura and Funahashi 2013). Therefore, the book Electroporation and Sonoporation in Developmental Biology edited by H. Nakamura in 2009 displays a quite complete overview of the possibilities offered by this technique, with detailed accounts on manipulations performed on model organisms such as chick and mouse (Nakamura 2009). Since then electroporation has become a widespread tool enabling embryologists to answer questions specific to their discipline. More than a hundred of original articles have been published on the sole methodological advances. Additionally, many reviews and focus papers have already extensively summarized and discussed this large corpus of studies (Itasaki et al.

The Use of Electroporation in Developmental Biology

3

1999; Swartz et al. 2001; Ogura 2002; Krull 2004; Odani et al. 2008; SaukaSpengler and Barembaum 2008; Takahashi et al. 2008a; Tanaka et al. 2014). Thus, the present chapter has been conceived along somewhat different lines: it aims to be a primer that presents the utilization of electroporation in developmental biology from an engineering point of view. To start with, the fundamental objectives pursued by embryologists will be described, as well as the typical experiments they carry out. How electroporation can fulfill some of the corresponding technical needs will also be explained. Next, a comprehensive summary of the realizations so far published will be provided, using the strategy devised to localize transfection as an organizing principle for the redaction. Finally, some of the practical issues related to the implementation of electroporation in this particular field of life sciences will be discussed. The bibliography will be more illustrative than exhaustive. It will mainly focus on seminal publications and on well-detailed technical papers, leaving aside reports on biological questions in which electroporation is marginally employed. Furthermore, references for many reviews will be included, to be used as introductions to more specific parts of the literature. We will also try to discuss as much as possible the advances posterior to the 2009 publication of Nakamura’s book. Noteworthily, in the last 10 years both Nature Protocols and the Journal of Visualized Experiments have released a few dozens of detailed protocols, old and new. In the case of the latter journal, each article is associated with videos that really bring a plus, because skilled manual experimentation is often a prerequisite in embryology. Finally, for complementary technical and bibliographical information, one can profitably check the websites of some of the main providers of electroporation apparatus (e.g., BTX Harvard Apparatus, NepaGene).

Specificities Related to the Scientific Context Typical Experiments The formation of a multicellular organism from a single cell, the fertilized egg (or zygote), involves fundamental processes like differentiation, which produces diverse cell types such as neurons or insulin producing cells, and morphogenesis, which allows their spatial and functional organization into organs such as the brain or the pancreas. During embryonic development, a tight regulation of cell division, cell death, and cell migration is also at play to control the global growth (Slack 2006; Gilbert 2013). Importantly, developmental processes are not restricted to the time of the embryonic life. Indeed, in human for instance, some organs like the skin or the surface of the intestine are constantly renewed. Moreover, in animals with high regenerative capacities like amphibians, an entire adult limb or a tail can be replaced after injury. Understanding the development of a complex organism requires acquiring a precise knowledge of the origins and of the dynamics of the cell populations that form the different tissues of the embryo. So-called fate maps are then drawn by tracing the descendants of a particular cell population (cell lineage analysis) or even

4

C. Gosse et al.

of a single cell (clonal analysis) (Stern and Fraser 2001; Kretzschmar and Watt 2012). Classically, marking selected cell populations relied on the injection of organic dyes (Bhattacharyya et al. 2008) or the grafting of cells possessing specific characters such as a pigmentation or a specific nuclear organization (Le Douarin et al. 2008 – Fig. 1a). In modern developmental biology, cells are labeled by modifying their gene expression so that they produce specific proteins like the enzyme β-galactosidase, which product can be detected in fixed tissues or fluorescent proteins, that allow live imaging experiments (Fig. 2a). Another important experimental paradigm in developmental biology is to modify the behavior of a given cell population in order to understand its role in the formation of a given tissue. In classic embryology it was achieved by the grafting or the ablation of specific cell groups. This approach is best illustrated by the Spemann and Mangold experiment, demonstrating that grafting some specific dorsal cell population of an amphibian embryo, the organizer, is sufficient to modify the

Fig. 1 Examples, in developmental biology, of seminal experiments relying on tissue transplantation. (A1–2) Lineage tracing performed on avian embryos to decipher the origin of the skull components (Adapted from Couly et al. 1993). (A1) A piece of the anterior part of the neural fold is dissected from a quail embryo at the three-somite stage and grafted in a chick one. (A2) The fate of the tissues issued from this isotopic and isochronic substitution is analyzed by inspection of head sections at later stages. Quail cells can indeed be distinguished from chick ones thanks to differences in the structure of their interphase nuclei. The xenograft is here at the origin of the upper part of the skull. (B1–2) Induction experiment evidencing the presence of the organizer in newt embryos (Adapted from Kretzschmar and Watt 2012) and inspired from the work presented in (Spemann and Mangold 1924). (B1) A piece of tissue from above the dorsal blastopore lip is collected on a Triton cristatus gastrula and ectopically transplanted in the ventral marginal zone of a Triton taeniatus one (the two species differ by their pigmentation). (B2) The signals produced by the graft subsequently lead to the apparition of a secondary body axis

The Use of Electroporation in Developmental Biology

5

behavior of the adjacent tissues in the receiving host, so that a secondary axis is induced (Spemann and Mangold 1924 – Fig. 1b). Later on, the molecular nature of the communication between cell groups was further evidenced by grafting polymeric beads releasing so-called morphogens, i.e., small molecules or proteins that induce the formation of anatomical structures from a distance (Eichele et al. 1985; Stern et al. 1990). Nowadays, local modification of gene expression is commonly used to change the characteristics of targeted cells (Fig. 2b). Gain- or loss-of-function strategies result in the respective up- or downregulation of the activity of the protein of interest, by modulating its synthesis or by interfering with its normal molecular behavior (section “Molecular Biology Tools”). The consequences of these changes on the modified cell population or on the adjacent tissues can then be addressed in the context of the whole developing organism (Sauka-Spengler and Barembaum 2008; Tanaka et al. 2014). The ultimate goal of such perturbative approach is to explain morphogenesis thanks to wiring diagrams that emphasize the spatiotemporal coupling, either chemical or mechanical, between intracellular genetic regulatory networks and intercellular communication pathways (Streit et al. 2013). Local modification of gene expression to trace cell populations or to change their behavior can be achieved in several ways. In some animal models like the Drosophila, powerful genetic tools allow the production of lines where clones of cells harbor stable transgenes or mutations. In contrast, some other models like the chick or the Xenopus do not offer such possibilities. Finally, in some cases like the mouse, genetically modified embryos can be obtained but it requires much time and labor (Itasaki et al. 1999; Slack 2006; Sharpe and Mason 2008). In addition, when a stable line can be produced, one also has to cope with the lethality associated with mutations affecting most developmental genes – a protein might have several roles at different stages and in different tissues but the death of the embryo during early regulatory processes prevents the study of the following ones. Therefore, conditional expression systems have been devised, relying on the response to an external chemical trigger (Danielian et al. 1998; Shin et al. 1999) and/or on the activation of tissue-specific cis-regulatory elements (Jaenisch 1988). However, these techniques are quite heavy, and alternative methods have been developed to modify gene expression with high spatial and temporal resolutions (section “Competing Techniques”). Electroporation is one of them, which enables embryologists to perform experiments that are basically the same than a century ago but in the context of functional genomics and with the incredible precision associated with modern molecular biology – compare Figs. 1 and 2.

Organisms and Tissues As stated above, the scale of embryology is the one of cell populations and tissues. Therefore, we will not consider here transfection studies that are performed on zygotes in order to produce stable lines of transgenic animals. In a similar way, researches on isolated embryonic cells will be ignored. Conversely, we will take into account manipulations realized on explants to understand organogenesis. We will

Fig. 2 Examples, in developmental biology, of contemporary experiments relying on tissue electroporation. (A1–4) Lineage tracing performed on a chick embryo to evidence somitogenesis (Adapted and reproduced from Iimura and Pourquié 2008). (A1) Schematic dorsal view of the injection of the DNA solution (in green) between the epiblast and the vitelline membrane at stage HH4. Using a fine glass needle, the anterior first third of the primitive streak groove is targeted. (A2) Transverse representation of the electroporation. (A3) Overlay of fluorescence and bright-field images evidencing the area over which the dsRed is protein expressed 4 h after electroporation with the corresponding plasmid. (A4) Same result after 24 h. (B1–4) Induction experiment demonstrating that the ectopic expression of the Hoxb-8 transcription factor can yield “mirror image” digit duplication in chick (Adapted and reproduced from Oberg et al. 2002). (B1) Schematic dorsal view of the injection of the DNA solution (in green) into the limb bud mesoderm at stage HH19. (B2) Transverse representation of the

6 C. Gosse et al.

The Use of Electroporation in Developmental Biology

7

also include experiments related to regeneration (of tail and limb in fishes and amphibians) and neonatal development (of brain in mouse and rat). The early implementations of electroporation in developmental biology were directly inspired by the protocols devised for cell suspensions. Transfections were mainly performed in cuvette equipped with parallel plate electrodes, relying on apparatuses providing exponentially decaying voltage pulses. This type of experiments was well adapted to very early stage embryos, containing from one to a few cells. Among the first scrutinized animals were fishes (Inoue 1992; Buono and Linser 1992; Powers et al. 1992; Müller et al. 1993; Murakami et al. 1994), insects (Kamdar et al. 1995; Leopold et al. 1996), and ascidians (Corbo et al. 1997). Soon after, electroporation was demonstrated on chick embryos in ovo (Muramatsu et al. 1996, 1997). This advance was made possible thanks to the use of a generator delivering square voltage pulses of low amplitude and long duration (section “Material Selection”), which enabled to significantly reduce cell death (Itasaki et al. 1999; Ogura 2002; Krull 2004; Nakamura and Funahashi 2013). Naturally, the fact that chick was characterized by a long tradition in microsurgery as well as in in ovo culture was also important for those initial trials (Momose et al. 1999). Now that this animal model had become amendable to gene expression manipulation, developmental studies including gain- and loss-of-function could be performed, especially on neural tissues (Itasaki et al. 1999; Swartz et al. 2001; Odani et al. 2008). Transfection protocols adapted to other tissues (Oberg et al. 2002; Krull 2004) or to in vitro culture (Iimura and Pourquié 2008; Voiculescu et al. 2008) were further published. Two years after the pioneering work of Muramatsu and his colleagues, mouse embryo electroporation was successfully undertaken (Akamatsu et al. 1999; Itasaki et al. 1999), once more focusing on nervous system formation. In these in vitro protocols transfection was followed by whole embryo culture (Takahashi et al. 2008a). However, not all developmental stages could be studied that way – in particular, it was impossible to obtained living pups. Consequently ex and in utero electroporation techniques were conceived (Tabata and Nakajima 2001; Saito and Nakatsuji 2001; Fukuchi-Shimogori and Grove 2001). Finally, the preimplantation (Grabarek et al. 2002; Soares et al. 2005) and the early post-implantation (Mellitzer et al. 2002; Davidson et al. 2003; Soares et al. 2008) stages were also transfected in vitro. Until now, chick and mouse embryos are the ones that have been the most often electroporated. These two species, respectively, Gallus domesticus and Mus musculus, belong to the group of the six most popular animal models in developmental biology, with the zebrafish Danio rerio, the frog Xenopus laevis, the ä Fig. 2 (continued) electroporation. (B3) Fluorescence image displaying the area over which the GFP is expressed 24 h after electroporation of a mixture of plasmids encoding for the latter fluorescent protein and for Hoxb-8. The dark spot indicates the injection site where, in this technique, some oil remains. (B4) Anatomical results after trichloroacetic acid fixation and alcian green staining at stage HH35-36

8

C. Gosse et al.

nematode worm Caenorhabditis elegans, and the fruit fly Drosophila melanogaster (Slack 2006; Sharpe and Mason 2008). Reports on electroporation of the two latter organisms can hardly be found in the literature because genetic tools have been available for long, and establishing a new technique was unnecessary (section “Typical Experiments”). Conversely, methods to control gene expression were also eagerly required for the two remaining species, and therefore electroporation was readily transferred to zebrafish (Tawk et al. 2002; Teh et al. 2003; Cerda et al. 2006; Hendricks and Jesuthasan 2007; Tawk et al. 2009; Kera et al. 2010) and Xenopus (Eide et al. 2000; Haas et al. 2001, 2002; Sasagawa et al. 2002; Falk et al. 2007; Lin et al. 2007; Chernet and Levin 2012) after it had been established in chick and mouse. Subsequently, more confidential animal models could also be transfected by electroporation. It includes the silkworm Bombyx mori (Moto et al. 1999; Thomas 2003; Ando and Fujiwara 2013), the sea squirt Ciona intestinalis (Corbo et al. 1997), the freshwater polyps Hydra vulgaris and Hydra magnipapillata (Smith et al. 2000; Bosch et al. 2002), the axolotl Ambystoma mexicanum (Echeverri and Tanaka 2003, 2005), and the newt Notophthalmus viridescens (Kumar et al. 2007). As far as in vitro organ electroporation is concerned, the proof of concept was realized on chick retinal explants (Pu and Young 1990) – in fact it was the first use of the technique in a developing tissue. Then, the transfection of a whole heart (Harrison et al. 1998) as well as pieces of gut (Fukuda et al. 2000) and slices of brain (Hashimoto-Torii et al. 2003) followed. To study organogenesis in mouse, protocols are now available for the intestine (Abud et al. 2004), retina (Donovan and Dyer 2006), kidney (Alie et al. 2007), palate (Lee et al. 2008), cochlear (Driver and Kelley 2010), genital ridge (Tanaka et al. 2014), mandibular arch, tail bud, neural plate, and lung endodermal bud (Mazari et al. 2014). Finally, in regenerative biology, electroporation has enabled to better understand the regrowth of tentacle in Hydra (Smith et al. 2000), spinal cord in Xenopus tadpole (Lin et al. 2007) and axolotl (Echeverri and Tanaka 2003), limb in newt (Kumar et al. 2007) and axolotl (Echeverri and Tanaka 2005), as well as fin in zebrafish (Tawk et al. 2002).

Molecular Biology Tools From the physicochemical point of view, there are only four kinds of molecules that are electroporated in embryos to perform gain- and loss-of-function experiments: the plasmids (pDNA) which are double-stranded circular DNA polymers (up to 14,000 base pairs long), the messenger RNA (mRNA) which are single-stranded RNA polymers (up to 1,500 nucleotides long), the small interfering RNA (siRNA) which are double-stranded RNA molecules (20–24 base pairs long), and the morpholinos (MO) which are oligomeric single-stranded analogues of DNA (around 25 morpholino units long). For the sake of simplicity, all these chemical species will be denoted as nucleic acids in the following, although morpholinos do not have a phosphate backbone. In fact, the latter oligomers are neutral, and to favor their delivery to the targeted cells by electrophoresis, they are often conjugated to a

The Use of Electroporation in Developmental Biology

9

charged dye (section “Performances Validation”). The difference in size characterizing those four reagents is also noteworthy when one considers the transfection mechanism: siRNA are small enough to first enter through the pores opened by the electric pulses and then diffuse freely in the cytoplasm before reaching their biological target (Paganin-Gioanni et al. 2011); in contrast, large plasmids form aggregates at the cell surface (Golzio et al. 2002; Mazari et al. 2014) before being actively transported to the nucleus (chapters “▶ Gene Delivery by Electroporation In Vitro: Mechanisms” and “▶ Nucleic Acid Electrotransfer in Mammalian Cells: Mechanistic Description”). Information on mRNA and MO are scarce (Chabot et al. 2013), although processes should be similar to the ones occurring in the siRNA case. From the biological point of view, the nature of the obtained biological perturbation, i.e., either an up- or a downregulation, is not unambiguously related to the nature of transfected molecules, as reviewed for chick (Krull 2004; Sauka-Spengler and Barembaum 2008; Streit et al. 2013) and for mouse (Takahashi et al. 2008a; Tanaka et al. 2014). Gain-of-function experiments consist in the overexpression of the gene of interest, either in its wild-type form or in a constitutively activated one. This is commonly achieved through the electroporation of a plasmid encoding for the corresponding protein; relying on mRNA is more expensive and thus rarer (Sasagawa et al. 2002; Cerda et al. 2006; Bansal et al. 2009; Chernet and Levin 2012). On the other hand, the ways to realize loss-of-function experiments are more diverse. First, one can use a plasmid that encodes for a dominant-negative form of the targeted protein, e.g., a transcription factor with an inactive transactivation domain or a receptor without transmembrane domain. Consequently, the latter molecule will act as a competitive inhibitor of the wild-type form and perturb the associated regulatory pathway (Akamatsu et al. 1999; Bartkowska et al. 2007, Barembaum and Bronner-Fraser 2007). Another common strategy consists in blocking the transcription of the gene of interest, which is classically achieved through morpholino electroporation (Mellitzer et al. 2002; Kos et al. 2003). These nucleic acids are in fact antisense oligonucleotides which are either directed against the proximal region of the translation initiation site, to sterically inhibit the initiation complex, or designed to encompass the intron/exon boundaries and thus interfere with RNA splicing. The last well-established loss-of-function approach involves siRNA. These nucleic acids operate within the RNA interference pathway, a succession of several enzymatic reactions catalyzed by various protein complexes, the final result being that the sequence complementarity between one of the RNA strand and a given mRNA drives the degradation of the latter molecule and thereby prevents protein synthesis. siRNA molecules can be introduced as such in the cytoplasm by electroporation (Mellitzer et al. 2002; Calegari et al. 2002; Pekarik et al. 2003) or one may have them produced by the cell from a transfected plasmid (Katahira and Nakamura 2003; Peng et al. 2012). In this case, a promoter for RNA polymerase III is used, and the sequences are chosen to yield small hairpin RNA (shRNA) that are then processed as siRNA. To conclude, we would like to mention that up to now, the CRISPR/Cas9 genome editing technology has only been scarcely used in embryos in conjunction with

10

C. Gosse et al.

electroporation, except for transgenic line production (Sato et al. 2016). Concerning developmental biology, knockout experiments have so far been realized in ascidia (Stolfi et al. 2014; Sasaki et al. 2014), chick (Véron et al. 2015), and mouse (Tsunekawa et al. 2016). The differences in mechanism of action outlined above come with differences in timing of perturbation. In experiments using plasmids encoding fluorescent proteins, expression is detectable 2–4 h after electroporation, it then reaches a maximum 4–7 days later, and finally it fades within 1–2 weeks (Momose et al. 1999; Teh et al. 2003; Echeverri and Tanaka 2003; Oberg et al. 2002; Davidson et al. 2003; Falk et al. 2007; Takahashi et al. 2008a; Kera et al. 2010; Sauka-Spengler and Barembaum 2008; Simkin et al. 2014). However, these numbers may strongly vary with the cell type: for instance, a signal still measured after 4 months in postmitotic neurons (Saito 2006). Such an observation thereby indicates that pDNA displays a fair chemical stability and that the loss of expression is likely due to the dilution of the vector in the dividing cells. To tackle the latter issue, Tol2 transposon-mediated gene transfer has been devised, which provides a stable integration into the host genome (Takahashi et al. 2008b; Simkin et al. 2014). With mRNA, expression can be observed as early as 1 h after electroporation (Sasagawa et al. 2002). Yet, this technique is more costly and more challenging due to the poor stability of mRNA and its degradation by cell nucleases. As far as siRNA are concerned, things are even worst since the dicer enzyme cut them into pieces. Therefore, to sustain the knockout activity, plasmids permanently enabling shRNA production are utilized (Katahira and Nakamura 2003; Peng et al. 2012). Finally, MO are stable but their antisense effect tends to disappear due to the dilution associated with cell division.

Competing Techniques Different methods have been developed to introduce nucleic acids into cells, so as to produce genetically modified animals or to cure hereditary diseases (Kaestner et al. 2015; Sato et al. 2016). Among the ones listed in the first column of Table 1, four are today routinely used in embryology laboratories: viral infection, lipofection, microinjection, and electroporation. Incidentally, a few articles report on sonoporation (Ohta et al. 2003; Skachkov et al. 2014) on biolistics (Muramatsu et al. 1997; Thomas et al. 2001; Lee et al. 2005). On the other hand, the Table 1 header gives some of the parameters that have to be taken into account when selecting a technique. It is here worth noting that protocols are not exclusive. For instance, lipofection has been coupled to sonoporation (Lee et al. 2005) or to magnetofection (Svingen et al. 2009); grafting experiments, as described in Fig. 1, have been performed after transfection of the donor tissues (Iba 2000; Iimura and Pourquié 2008). Gene transfer by viral infection has been important in chick embryology, as a way to cope with the lack of genetic tools characterizing this model (Iba 2000; SaukaSpengler and Barembaum 2008). With retrovirus, the transgene is integrated in the

High

Good

High

Good

Microinjection

Electroporation

Sonoporation

Physical transfection Gene biolistics

Adenoviral Variable infection Chemical transfection Lipofection

Lentiviral infection

Technique Efficacy Biological transfection Retroviral High infection

Mechanical damage Mechanical damage Electrical damage Mechanical damage

Chemical toxicity

Biological toxicity; mutagenic Biological toxicity; mutagenic

Possible risk for the tissue

Labor

Labor; cost

Labor; cost; biohazard

Labor; cost; biohazard

Drawback for the researcher

pDNA; dye

All

All

pDNA

pDNA; siRNA

Insert 1000 V/cm) or by the measurement of electrical conductivity of the freeze-thawed material. However, the strong PEF treatment or freeze-thawing procedure can also differently affect the structure of cell walls. The existing textural data evidenced the different softening of PEF-treated and freezethawed potato, apple, and carrot tissues. So, theσ d value in the low-frequency method can not be uniquely determined, and the value of electrical conductivity

14

N. Lebovka and E. Vorobiev

1

sd /si ª14

Conductivity disintegration index (ZC)

Potato 0.8

t 0.6

0.4

Orange sd /si =1.3 E, V.cm–1 400

0.2

1000 0

10–3

10–2

10–1

100

Time of PEF treatment (tPEF, S)

Fig. 9 Electrical conductivity disintegration index Zc versus the time of PEF treatment tPEF for potato and orange. PEF treatment was done with bipolar pulses duration ti =1 ms and pulse repetition time Δt =10 ms. The data were obtained using low frequency method (Ben Ammar 2011)

disintegration index Zc can be also dependent on the selected way for determination of σ d. Figure 9 presents typical Zc(tPEF) dependencies for potato and orange (Ben Ammar 2011). For potato tissue more efficient electroporation was observed as compared with orange. It is surprising because the radius of the orange cells (65 μm) is larger than of the potato cells (35 μm) and one expected the opposite behavior (see, Eq. (1)). However, the potato and orange have significantly different electrical conductivity contrasts, k = σ d/σ i  14.3 and σ d/σ i = 1.3, respectively. It was demonstrated that differences in electroporation efficiency for these tissues can be explained by the differences in the electrical conductivity contrast.

Low-High Frequency Method This method is based on the measurements of electrical conductivity σ of intact and electroporated materials at low (1–10 kHz) and high (50 MHz) frequencies; it is based on the model of bio-impedance. At low frequency, cell membranes have very high resistance and reactance, and their electroporation can result in noticeable changes in the electrical conductivity spectrum in the ranges of α-and β-dispersions. However, at high frequencies, the effect of intact membranes on the conductivity spectrum becomes negligible. That is why electroporation of material results in weakening of frequency dependence of σ.

Techniques to Detect Electroporation in Food Tissues

0.8

15

Partially electroporated

sm∞ •

a = si / sm

Measured

0.7

si ∞

Corrected

Electrical conductivity (s ,S/m)

•

0.6 sm

0.5

sc ( f) = asm ( f)

sc

0.4 0.3

si •- si Intact

asm - si

0.2 0.1 si

0 103

104

105

106

107

108

Frequency ( f, Hz)

Fig. 10 Application of low-high frequency method for determination of the electrical conductivity disintegration index. Here, the typical dependencies of electrical conductivity σ versus the frequency f for intact and partially electroporated materials are presented

Figure 10 presents typical frequency dependencies of electrical conductivity σ for intact and partially electroporated materials. For the totally electroporated material, the electrical conductivities measured in low-frequency range σ and high-frequency range σ 1 should coincide and to be equal to that for intact material σ 1 i . The measured frequency dependence σ m( f ) commonly required corrections on changes in temperature, tissue porosity, and electrolyte concentration. Such correction can be 1 easily done using correction coefficient α ¼ σ 1 i =σ m that accounts for the difference 1 1 between σ i and measured value of σ m . Electrical conductivity disintegration index ZC for low-high frequency method was defined as (Angersbach et al. 2002) Z c ¼ ðασ m  σ i Þ= σ 1 i  σi




(5)

Thus determination of ZC requires measurements of electrical conductivity of 1 intact and electroporated materials at low (σ i, σ m) and high (σ 1 i , σ m ) frequencies. This equation gives Zc = 0 for intact and Zc =1 for completely electroporated material.

Method of Phase Shift The electric impedance method and measurements, based on the frequency dependency of the phase shift φ in the range of 500 Hz–10 MHz, were also used for estimation of the electroporation effects in PEF-treated mash from the wine grapes.

N. Lebovka and E. Vorobiev

Electrical conductivity ( σ, S/m)

16

0.3

0.2 1 2

Phase shift (j °)

0.1 3 10

104

105 Frequency (f, Hz)

106

107

104

105 Frequency (f, Hz)

106

107

15

10 5 0 3 10

Fig. 11 Frequency dependency of specific conductivity σ and phase shift φ for mash from Muscatel grapes without (1) and after (2) electroporation (specific energy:42 kJ/kg) (Compiled from (Sack et al. 2009))

Figure 11 shows example of frequency dependencies of specific conductivity and phase shift for mash from Muscatel grapes without (1) and after (2) electroporation (Sack et al. 2009). The maximum phase shift was approximately located at approximately 300 kHz, and the remaining phase shift for electroporated sample indicates that not all cells have been opened. Good correlations were observed between the measurements of the complex impedance and color intensity of the must. However, the qualitative parameter for characterization of disintegration index on the base of phase shift was not proposed.

Diffusion Coefficient The level of electroporation induced by PEF treatment can be also estimated from studies of mass transport characteristics in solid–liquid extraction and convective drying processes. The diffusion coefficient disintegration index ZD was defined as (Barba et al. 2015): Z D ¼ ðD  Di Þ=ðDd  Di Þ;

(6)

Techniques to Detect Electroporation in Food Tissues

17

where D is the diffusion coefficient and the subscripts “i” and “d” refer to the values for intact and completely electroporated tissue, respectively. As for electrical conductivity low-frequency method this method also requires the knowledge of Dd. It can be estimated from the data obtained in the limit of long treatment time (tPEF  0.1–1 s) and high electric field strength (E > 1000 V/cm) or the data for the freeze-thawed material. So, the value of Dd cannot be uniquely determined. For example, drying experiments with potato tissue have shown that Dd value of a freeze-thawed tissue is noticeably higher than Dd of a PEF-treated tissue with Zc  1.

Texture The textural experiments can be rather useful for qualitative characterization of PEF-induced changes, and they were frequently used for characterization of PEF-treated tissues (Barba et al. 2015; Fincan and Dejmek 2003). The stress–deformation and relaxation tests can be applied to analyze the texture of intact and PEF-treated tissues. The differences in the pressure-displacement curves P-e of PEF-treated and intact tissues were observed in many experimental works. For example, after PEF treatment (E = 1.1 kV/cm, tPEF = 0.1 s), the tissues (carrot, potato, and apples) lose a part of their textural strength, and their elasticity modulus and fracture stress decrease with increase of the conductivity disintegration index. PEF-induced changes in elasticity modulus Gm and fracture stress PF were significantly smaller than those observed in the freeze-thawed and thermally (T = 45  C, 2 h) pretreated tissues. The tissue structures were less affected by the PEF treatment as compared to the freeze-thawing or heating. This conclusion was confirmed by the textural studies of sugar beet tissue treated by PEF (Shynkaryk 2006). Linear dependency between the fracture pressure and the value of ZC was observed for PEF-treated apple samples. The PEF impacts on compression and solid–liquid expression of different vegetable tissues were also extensively studied (Barba et al. 2015). PEF-treatment accelerated the stress relaxation of tissues. The relaxation behavior reflected membranes damage, and it was sensible to the state of the cell walls and turgor pressure. In order to quantify electroporation, the texture disintegration index was introduced as Zt ¼ ðF  Fd Þ=ðFi  Fd Þ;

(7)

where F is the force measured after sufficient time of textural relation (e.g., at t = 50 s), and subscripts i and d refer to the values of intact and completely damaged tissue, respectively. The force relaxation tests for PEF-treated and freeze-thawed tissues had shown that at t 50 s the changes in values of force are unessential. The studies evidenced that PEF-treated tissue with high level of disintegration (Zc  1) has stronger texture compared with the texture of maximally disintegrated freeze-thawed tissue. Texture disintegration index Zt is convenient for understanding

18

N. Lebovka and E. Vorobiev

of the viscoelastic properties of tissue. Unfortunately, the textural parameters reflect the PEF-induced changes in an indirect manner and the definitive relations between these parameters and fraction of damaged cells Z are still absent. Moreover, the results of textural tests can be dependent on the mode of textural experiment. For example, they can different for uniaxial (1d) and three dimensional (3d) tests, and each tissue requires careful tests and comparison with other characterization techniques. Note that atomic force microscopy (AFM) technique can be also used for characterization the texture and local mechanical properties of food tissues (Cybulska et al. 2013). Particularly, application of AFM technique demonstrated a decrease in membrane elasticity (living CHO cells) by 40 % due to electroporation (Chopinet et al. 2013).

Acoustic Tests The acoustic technique is traditionally used for characterization of the quality of different agricultural products. The acoustical index of firmness S (or stiffness coefficient) shows good correlations with the quality and maturity of fruits and vegetables. Acoustic techniques have been applied for characterization of the different fruit and vegetable PEF-treated tissues (Grimi et al. 2010). The acoustic disintegration index Za was defined in a similar manner to electrical conductivity disintegration index Zc (Eq. 4) and diffusion disintegration index, Zd (Eq. 6): Z a ¼ ðS  Si Þ=ðSd  Si Þ, ;

(8)

where subscripts i and d refer to the indices of firmness of the intact (untreated) and completely damaged tissues, respectively. Examples of acoustic disintegration index Za versus time of PEF treatment tPEF for different fruit and vegetable tissues are presented in Fig. 12. Acoustic technique allows estimations of the whole unpeeled samples. It can be important when fruits and vegetables are processed as whole unpeeled samples in existing industrial examples of PEF application to whole samples, e.g., sugar beet, potato, tomato.

Conclusions The different experimental techniques for estimation of PEF-induced electroporation in food tissues were proposed in recent years. Their advantages and disadvantages are summarized in the Table 1. Practically all techniques are destructive (invasive), and they can violate the structure of the tissues and can affect the evaluated index of disintegration by applied procedure of measurements.

Techniques to Detect Electroporation in Food Tissues

19

Fig. 12 Acoustical disintegration index Za versus the time of PEF treatment tPEF for different fruit and vegetable PEF-treated tissues. PEF treatment was applied to whole unpeeled samples at electric field strength E = 200 V/cm and pulse duration ti =100 μs in an aqueous medium (0.06 S/m, tap water) (Grimi et al. 2010)

The details of the sample preparation and individual adaptation for each selected type of the material can be also very important. Unfortunately, there is no linear dependence between fraction of damaged (electroporated) cells Z and any of proposed indexes Zc, Zd, Zt, or Za. The correlations between different indexes, for example Zc and Zd, and Zc and Za, were discussed for some foods (Barba et al. 2015), but further studies for different types of food materials are still required. In general, for characterization of electroporation, one can advice using the combination of different techniques and compare them using the same PEF-protocols and the same food materials. Some other methods to detect electroporation phenomena can be used, including different tomography techniques (NMR, impedance, microwave) (see, section Cross-References). Such techniques require more qualified service, and they were only applied in particular studies aimed for elucidation of electroporation mechanisms.

20

N. Lebovka and E. Vorobiev

Table 1 Advantages and disadvantages of different techniques for estimation of PEF-induced electroporation in food tissues Technique Optical microscopy

Advantages The direct visualization is very attractive, demonstrable and decisive; The precise mechanism and spatial distributions of different components can be revealed; It can be applied to the visual study of different processes: diffusion, osmotic impregnation, drying, and freezing

Electron microscopy

SEM/TEM technique is very illustrative and powerful for quantification the morphological properties of membranes and cell walls

Electrical characteristics

This electrical impedance technique is straightforward, relatively simple and does not require expensive equipment; It is express and can be easily applied for continuous monitoring of electroporation during PEF treatment

Diffusion coefficient

This mass transfer technique is straightforward, relatively simple and does not require expensive equipment; It is useful for the estimation of PEF impact on solute and moisture diffusivities in solid–liquid extraction and drying experiments

Texture

It is straightforward and relatively simple; It is useful for qualitative determination of PEF-induced changes in mechanical characteristics:

Disadvantages It is destructive technique; The computerized visual fractionalization for counting of the cells is not easy task; The preparation of the samples, mounting of the epidermis and staining with dyes requires special strategies, individual adaptation of pH and conductivity of used solutions for each selected type of the material It is destructive technique; It requires application of complex individual methods of specimen preparation (staining, fixation, dehydration, ultrathin sectioning etc. . .) It is destructive technique; The changes in distributions of the air and moisture content inside the tissue can affect the results; The direct proportionality between mechanical disintegration and measured electrical characteristic is questionable; The estimated values of disintegration index depend on the procedure used for the determination of electrical conductivity of completely damaged tissue, σ d, (freezing, heating or complete electroporation), and from the temperature of the medium It is destructive technique; The direct proportionality between disintegration and measured diffusivity characteristic is questionable; The estimated values of disintegration index depend on the procedure used for the determination of solute (moisture) diffusivity in completely damaged tissue, Dd, (freezing, heating or complete electroporation) and temperature of the medium It is destructive or partially technique; The textural parameters reflect the PEF-induced changes in an indirect manner and there is no direct proportionality between disintegration and measured (continued)

Techniques to Detect Electroporation in Food Tissues

21

Table 1 (continued) Technique

Acoustic test

Advantages

Disadvantages

cutting strength, stress–deformation and relaxation

characteristic; The results of the textural tests depend on the mode of textural experiment (e.g., they are different for uniaxial (1d) and three dimensional (3d) tests) Justification of direct proportionality between electroporation disintegration index and measured acoustic characteristics is still required

It is straightforward and relatively simple; It is nondestructive and applied to whole roots or fruits, e.g., sugar beet, potato, tomato, apple

Acknowledgments The authors appreciate the support from the COST Action TD1104 (EP4Bio2Med – European network for development of electroporation-based technologies and treatments).

Cross-References ▶ Atomic Force Microscopy for Electroporation Mechanisms Studies in Bacteria ▶ Biophysics and Metrology of Electroporation in Tissues ▶ Current Density Imaging as Means to Follow Tissue Electroporation ▶ Detection of Electroporation in Microbial Cells: Techniques and Procedures ▶ Diffusion Weighted Magnetic Resonance Imaging for Detection of Tissue Electroporation In Vivo ▶ Electroporation and Electropermeabilization ▶ Fluorescent Indicators of Membrane Permeabilization ▶ Impedance Measurement as Electroporation Measure ▶ Principles and Use of Magnetic Resonance Electrical Impedance Tomography in Tissue Electroporation ▶ Single-Cell Electrical Characterization Techniques

References Angersbach A, Heinz V, Knorr D (2002) Evaluation of process-induced dimensional changes in the membrane structure of biological cells using impedance measurement. Biotechnol Prog 18:597–603 Asavasanti S, Ristenpart W, Stroeve P, Barrett DM (2011) Permeabilization of plant tissues by monopolar pulsed electric fields: effect of frequency. J Food Sci 76(1):E96–E111 Barba FJ, Parniakov O, Pereira SA, Wiktor A, Grimi N, Boussetta N, Saraiva JA, Raso J, MartinBelloso O, Witrowa-Rajchert D, Lebovka N, Vorobiev E (2015) Current applications and new opportunities for the use of pulsed electric fields in food science and industry. Food Res Int 77:773–798 Ben Ammar J (2011) Etude de l’effet des champs electriques pulses sur la congelation des produits vegetaux. PhD thesis. Universite de Technologie de Compiegne, Compiegne

22

N. Lebovka and E. Vorobiev

Chalermchat Y, Malangone L, Dejmek P (2010) Electropermeabilization of apple tissue: effect of cell size, cell size distribution and cell orientation. Biosyst Eng 105(3):357–366 Cholet C, Delsart C, Petrel M, Gontier E, Grimi N, L’Hyvernay A, Ghidossi R, Vorobiev E, Mietton-Peuchot M, Geny L (2014) Structural and biochemical changes induced by pulsed electric field treatments on Cabernet Sauvignon grape berry skins: impact on cell wall total tannins and polysaccharides. J Agric Food Chem 62:2925–2934 Chopinet L, Roduit C, Rols M-P, Dague E (2013) Destabilization induced by electropermeabilization analyzed by atomic force microscopy. Biochim Biophys Acta (BBA)-Biomembr 1828:2223–2229 Condello M, Caraglia M, Castellano M, Arancia G, Meschini S (2013) Structural and functional alterations of cellular components as revealed by electron microscopy. Microsc Res Tech 76:1057–1069 Cybulska J, Zdunek A, Psonka-Antonczyk KM, Stokke BT (2013) The relation of apple texture with cell wall nanostructure studied using an atomic force microscope. Carbohydr Polym 92:128–137 El Zakhem H, Lanoiselle J-L, Lebovka NI, Nonus M, Vorobiev E (2006) The early stages of Saccharomyces cerevisiae yeast suspensions damage in moderate pulsed electric fields. Colloids Surf B Biointerfaces 47:189–197 Ersus S, Barrett DM (2010) Determination of membrane integrity in onion tissues treated by pulsed electric fields: use of microscopic images and ion leakage measurements. Innovative Food Sci Emerg Technol 11:598–603 Fazaeli M, Tahmasebi M, Djomeh EZ (2012) Characterization of food texture: application of Microscopic technology. In: Mendez-Vilas, A. (ed.) Current microscopy contributions to advances in science and technology. pp. 855–871. Formatex Research Center: Badajoz, Spain Fincan M, Dejmek P (2002) In situ visualization of the effect of a pulsed electric field on plant tissue. J Food Eng 55:223–230 Fincan M, Dejmek P (2003) Effect of osmotic pretreatment and pulsed electric field on the viscoelastic properties of potato tissue. J Food Eng 59:169–175 Grimi N, Mamouni F, Lebovka N, Vorobiev E, Vaxelaire J (2010) Acoustic impulse response in apple tissues treated by pulsed electric field. Biosyst Eng 105:266–272 Herman P, Drapalova H, Muzikova R, Vecer J (2005) Electroporative adjustment of pH in living yeast cells: ratiometric fluorescence pH imaging. J Fluoresc 15:763–768 Janositz A, Knorr D (2010) Microscopic visualization of pulsed electric field induced changes on plant cellular level. Innovative Food Sci Emerg Technol 11:592–597 Kinosita K Jr, Ashikawa I, Saita N, Yoshimura H, Itoh H, Nagayama K, Ikegami A (1988) Electroporation of cell membrane visualized under a pulsed-laser fluorescence microscope. Biophys J 53:1015 Lee EW, Wong D, Prikhodko SV, Perez A, Tran C, Loh CT, Kee ST (2012) Electron microscopic demonstration and evaluation of irreversible electroporation-induced nanopores on hepatocyte membranes. J Vasc Interv Radiol 23:107–113 Loginova SK (2011) Mise en oeuvre de champs electriques pulses pour la conception d’un procede de diffusion a froid a partir de betteraves a sucre et d’autres tubercules alimentaires (etude multiechelle). PhD thesis. Universite de Technologie de Compiegne, Compiegne Reilly JP (2012) Applied bioelectricity: from electrical stimulation to electropathology. Springer, New York Ryttsen F, Farre C, Brennan C, Weber SG, Nolkrantz K, Jardemark K, Chiu DT, Orwar O (2000) Characterization of single-cell electroporation by using patch-clamp and fluorescence microscopy. Biophys J 79(4):1993–2001 Sack M, Eing C, Stangle R, Wolf A, Muler G, Sigler J, Stukenbrock L (2009) Electric measurement of the electroporation efficiency of mash from wine grapes. IEEE Trans Dielectr Electr Insul 16 (5):1329–1337

Techniques to Detect Electroporation in Food Tissues

23

Shynkaryk M (2006) Influence de la permeabilisation membranaire par champ electrique sur la performance de sechage des vegetaux. PhD thesis. Universite de Technologie de Compiegne, Compiegne Spugnini EP, Arancia G, Porrello A, Colone M, Formisano G, Stringaro A, Citro G, Molinari A (2007) Ultrastructural modifications of cell membranes induced by electroporation on melanoma xenografts. Microsc Res Tech 70:1041–1050 Valic B, Golzio M, Pavlin M, Schatz A, Faurie C, Gabrie B, Teissie J, Rols M-P, Miklavčič D (2003) Effect of electric field induced transmembrane potential on spheroidal cells: theory and experiment. Eur Biophy J 32:519–528 Yuan X-S, Shen J-L, Wang X-L, Wu X-S, Liu D-P, Dong H-F, Jiang M-S (2005) Schistosoma japonicum: a method for transformation by electroporation. Exp Parasitol 111:244–249

Determination of Pulsed Electric Fields Effects on the Structure of Potato Tubers Indrawati Oey, Farnaz Faridnia, Sze Ying Leong, David J. Burritt, and Tingting Liu

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . An Overview of the Use of Electrical Impedance and Electrical Conductivity Measurements to Determine Cell Membrane Electroporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Determination of the Distribution of the PEF Effect Across Potato Tubers Using Cell Viability Staining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Determination of the Distribution of PEF Effect Across the Tuber Using Enzymatic Browning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microscopic Techniques Used to Assess the Microstructure of Potato Tubers After PEF Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Light Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fluorescence Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scanning Electron Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 4 8 10 13 13 14 14

I. Oey (*) Department of Food Science, University of Otago, Dunedin, New Zealand Riddet Institute, Palmerston North, New Zealand e-mail: [email protected] F. Faridnia (*) • S.Y. Leong (*) Department of Food Science, University of Otago, Dunedin, New Zealand Department of Botany, University of Otago, Dunedin, New Zealand e-mail: [email protected]; [email protected] D.J. Burritt (*) Department of Botany, University of Otago, Dunedin, New Zealand e-mail: [email protected] T. Liu (*) Department of Food Science, University of Otago, Dunedin, New Zealand Riddet Institute, Palmerston North, New Zealand Department of Botany, University of Otago, Dunedin, New Zealand e-mail: [email protected] # Springer International Publishing Switzerland 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_151-1

1

2

I. Oey et al.

Combining Scanning Electron Microscopy and Ion Migration Using Energy Dispersive Spectroscopy Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Determination of the Electric Field Distribution During electroporation Using Magnetic Resonance Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

15 16 17 18

Abstract

Potato tubers are often considered by food processors as being “homogenous” plant structures. However, tubers are complex organs made up of many tissue types, including a complex vascular system, each with different functions and properties that can affect the influence of pulsed electric fields (PEF) on cell and tissue structures. Many research investigations evaluating the effects of PEF on the microstructure of potato tuber have used samples that have been mechanically fragmented prior to PEF treatment and unfortunately the internal structure of vascular system inside the tuber has been ignored during sample preparation, PEF treatment, and microstructure analysis. This chapter discusses different methodologies that have been used to assess the effect of PEF treatment on potato tubers, including electrical impedance and conductivity measurements, cell viability staining, enzymatic browning, and microscopic techniques such as light microscopy, fluorescence microscopy, scanning electron microscopy, energy dispersive spectroscopy, and also magnetic resonance imaging. The limitations of each technique and how these techniques have been used in published studies are discussed. In addition, this chapter demonstrates how sample pretreatments such as peeling or cutting influence the assessment of PEF effect on the microstructure of the tuber and how several techniques should be used and combined in order to understand how the PEF effects are distributed throughout the whole potato tubers. Keywords

Potato • Tubers • Microstructure • Staining • Browning • Microscopy • Magnetic resonance imaging • Electroporation

Introduction Pulsed electric fields (PEF) technology offers promising opportunities to improve product quality in the potato industry. PEF technology induces electroporation, leading to increased cell membrane permeability which also affects tissue structure in a way to soften potato tissues resulting in better cutting quality and accuracy (Ignat et al. 2015). In addition, electroporated potato cells are more prone to release of intracellular compounds such as reducing sugars and low-molecular substrates involved in the Maillard reaction and hence reduce the tendency of fried potatoes to brown (Janositz et al. 2011). How PEF processing affects the potato tuber, which is a complex heterogeneous structure, is not completely understood. Important questions

Determination of Pulsed Electric Fields Effects on the Structure of Potato. . .

3

Potato tuber Stem end Outer medulla Medullary ray

Pith (Inner medulla) Skin Vascular ring

Bud end Fig. 1 Cross section of a potato tuber showing the internal structure

still to be addressed include whether the effects of PEF treatment are distributed uniformly across the whole tuber and which of the cell types that make up tubers are most affected by PEF. Being a living intact plant organ, a potato tuber (Solanum tuberosum) is made up of several tissue types with different cellular properties (Fig. 1) and functionalities. The stem tubers of S. tuberosum have a vascular system, in which the vascular bundles form a ring beneath the skin and the pith (or inner medulla) forms branched medullary rays that terminate in vegetative buds (Reeve et al. 1969). The pith is comparatively low in starch, but is high in free amino acids, resulting in the pith having a relatively high electrical conductivity. Sweet potato (Ipomoea batatas) is a distant relative of S. tuberosum that produces tubers from the root system. The root tubers of I. batatas plants are structurally different than the stem tubers of S. tuberosum and initially are structurally similar to the roots of carrots. However, during the transitional process to be a storage structure, additional cambia within the secondary xylem produce mostly storage parenchyma cells in both directions, resulting in xylem and phloem composed mostly of storage parenchyma cells, which make up most of the ground tissues (Fig. 2). From the above description of potato and sweet potato tubers, it is clear that tuber should not be considered to be “homogenous” plant structures, and this fact needs to be taken into consideration when investigating PEF-induced structural changes in plant cells, tissues, and organs. To date, previous work evaluating the effects of PEF on potato tuber microstructure has used tuber samples that have been mechanically fragmented prior to PEF treatment, e.g., slices (Mhemdi et al. 2013), cylinders (Boussetta et al. 2013), strips (Shayanfar et al. 2013), or cubes (Ignat et al. 2015).

4

I. Oey et al.

Fig. 2 Cross section of a sweet potato tuber showing the internal structure

These forms of sample preparation result in potato tuber tissues that have already experienced considerable tissue damage prior to PEF treatment, which could mask the nature of any PEF-induced changes. More importantly, these earlier works did not consider the complex nature of plant organs and neglected the fact that potato tubers are heterogeneous structures. Evaluation of cell electroporation due to PEF treatment in multilayer biological tissues such as potato tubers is challenging. Most research has been based on theories derived from experiments carried out on model systems, such as liposomes, planar bilayers, and phospholipid vesicles. These models are undoubtedly useful when explaining the exact mechanisms associated with irreversible and reversible electroporation and/or electrically induced membrane breakdown (Zimmermann 1986). However, there is a need for simple high-throughput experimental approaches to assess cell membrane permeability and integrity in multicellular plant tissues and to study how electric fields and their effects are distributed throughout complex plant tissues and organs. This chapter will discuss different experimental approaches to assess the effect of PEF treatment on the structure of potato tubers. The limitations of each technique will be discussed and examples about how these techniques have been used in the published literature are given.

An Overview of the Use of Electrical Impedance and Electrical Conductivity Measurements to Determine Cell Membrane Electroporation The measurement of electrical impedance and electrical conductivity considers plant tissue to be an electrical system (circuit) with both the resistive and capacitive properties (Zhang and Willison 1993). When plant samples are treated with PEF,

Determination of Pulsed Electric Fields Effects on the Structure of Potato. . .

5

the sample is considered to be an “electrical object,” the electrical properties of which can be changed by PEF treatment (Angersbach et al. 2002). An increase in the cell membrane permeability due to PEF would lower membrane resistance allowing the cell contents to pass through the membrane into the intercellular spaces, thus changing the impedance of the plant cell (Angersbach et al. 2002). Therefore, the impedance of a plant sample is a useful parameter to immediately assess the extent of cell membrane electroporation or cell damage due to PEF. Determining the extent of cell damage by measuring the impedance has been reported for potato tubers (Gachovska et al. 2015) and other plant samples (Donsì et al. 2010), and is a wellestablished technique used to gain an overview of the condition status of a plant sample (Zhang and Willison 1993). Cell membranes exhibit frequency-dependent electrical properties (Angersbach et al. 2002). At low frequencies (1 kHz), the cell membrane acts as a capacitor to prevent the flow of electric current in the intracellular medium, which is a classical example of ohmic-capacitive behavior. At high frequencies (10–100 MHz), the cell membrane shows no further resistance to electric current flow, resulting in a pure ohmic behavior in which the absolute value of impedance does not vary greatly between intact cells and cells with ruptured membranes. Taking the advantage of the changes in the electrical properties of plant samples at low and high frequencies, impedance measurements can be performed. At the laboratory scale, a plant sample is placed in a test vessel consisting of two parallel plate electrodes or cylindrical electrodes. The electrodes are connected to a generator that produces a sinusoidal waveform at a specified effective voltage (usually 1–5 V, peak-to-peak) with a frequency ranging between 1 kHz and 50 MHz. A frequency-impedance spectrum is then obtained, revealing the transition of the untreated and PEF-treated samples from an intact to ruptured state within the frequency range investigated (Donsì et al. 2010). The electrical impedance (unit in Ohm, Ω) is the ratio of the voltage drop across the sample and the current crossing it, when the measurement is taken (Eq. 1).  Cell impedance, Z ðΩÞ ¼ Rf

Vs Vf  1

 (1)

Rf (Ω) is the reference resistor that is connected in series to the impedance instrument generating the sinusoidal voltage, while Vs and Vf represent the peak-to-peak voltage of the sinusoidal waveform, i.e., voltage applied and the voltage measured on the reference resistor, respectively. As an alternative, measuring the electrical conductivity response of the sample at the low-to-high frequency range is also a good method to evaluate a change in cell impedance due to PEF (Angersbach et al. 2002). To allow a better interpretation of the impedance result and to quantify cellular damage caused by independent PEF treatments, a polarization coefficient termed the cell disintegration/permeabilization index (Zp) has been introduced (Lebovka et al. 2002). This coefficient is calculated based on the ratio of the measured impedance or electrical conductivity measured at low (Z1  1 kHz) and high frequencies (Z2  10 MHz) (Eq. 2).

6

I. Oey et al.

Zp ¼ ðZ1untreated  Z 1PEF-treated Þ=ðZ1untreated  Z2PEF-treated Þ

(2)

Based on this equation, Zp = 0 and Zp = 1 represent the intact and maximally damaged cells, respectively. There is strong evidence that a higher PEF treatment intensity always leads to a higher degree of cell membrane permeabilization. Another way to interpret the impedance result is to determine the degree of damage (Sd), which is the ratio of impedance before and after PEF treatment. In the case of potato tubers, the use of Sd to compare the effectiveness of increasing pulse number on potato tuber cell damage was first implemented by Gachovska et al. (2015). In general, the degree of cell damage due to PEF can be more clearly observed at low frequencies (1 kHz). This is because where PEF-induced cell electroporation occurred, the ion concentration in the extracellular spaces would be already high enough to contribute to a change in the impedance values at low frequencies. Generally, permeabilization of cell membranes causes diffusion of intracellular cell components, containing various ionic species, from the intracellular to the extracellular space. The measurement of electrical conductivity (σ, unit in S/m) considers the total ion leakage from plant cells. This is also another simple method to estimate the degree of cell damage and to generate a cell disintegration/permeabilization index (Zp) for plant tissues after PEF treatment (Eq. 3). Zp ¼

ðσPEF  σintact Þ
σdamaged  σintact

(3)

σ is the measured electrical conductivity at low frequencies (usually 0.5–1 kHz) of PEF-treated samples (σPEF), intact cells (untreated samples, σintact), and maximally damaged cells (σdamaged). Similar to Eq. 2, Zp = 0 and Zp = 1 represent the intact and totally damaged cells, respectively, on the basis of electrical conductivity disintegration. In all cases, the electrical conductivity increases with PEF treatment intensity. When the applied PEF treatment is sufficient to cause maximal cell membrane permeabilization (Zp  1), no further increase in cell damage with more intense PEF treatments can be obtained (Lebovka et al. 2002). This is helpful when defining optimal PEF treatment conditions to achieve the desired level of cell damage in the final product, without excessive energy consumption. In an attempt to improve PEF processing efficiency, previous works have employed the measurement of electrical conductivity to (i) predict the characteristic damage time of potato tubers, with the PEF treatment condition required for attaining one half of the maximum Zp value (i.e., Zp = 0.5) (Lebovka et al. 2002), (ii) recommend the combined use of mild preheating ( 60  C. PEF pretreatment considerably facilitated diffusion processes even below 50  C. For example, at 40  C the practically almost releasing of soluble matter from chicory was observed for
1.5 h of extraction. The possibility of PEF-assisted (E = 600 V/cm, ti = 100 μs, n = 100–500, tPEF = 0.01–0.05 s) aqueous extraction (T = 30–80  C) of inulin from chicory roots has been also tested at the pilot scale using the special pilot countercurrent

20

E. Vorobiev and N. Lebovka

Normalized oBrix, B

1

0.8

0.6

Chicory

0.4

T= – 60 oC – 50 oC – 40 oC

0.2

0 0

2000

4000

6000

8000

10000

Time, t, s Fig. 12 Normalized oBrix of the chicory juice, B, (Eq. 1) versus the extraction time t for untreated (open symbols, solid lines) and PEF-treated (closed symbols, dashed lines) slices at different temperatures. PEF treatment was done at E = 600 V/cm and tPEF = 1.0 s, which corresponds to the highest level of disintegration degree (Z
1)

extractor (Zhu 2014). The draft (liquid to solid mass ratio) was fixed at 140 %, similar to the industrial conditions. With PEF treatment, the inulin concentration increases from 10.68 to 12.18 g per 100 mL when diffusion temperature varies from 30  C to 80  C. At T = 60  C, the inulin concentration in juice (11.65 g per 100 mL) and the juice purity (
87.1 %) were comparable to those of conventional thermal diffusion at T = 80  C. Moreover, the diffusion juices obtained with PEF at low temperatures were less colored, had lower turbidity, lower protein content, and higher purity than the conventional juice obtained at 80  C (Zhu 2014). The dead-end ultrafiltration of obtained diffusion juices in a steering cell permits their following purification. The hydrophilic polyethersulfone membrane with molecular weight cut-off (MWCO) of 50 kDa produced the most purified filtrate, while the membrane with lower MWCO (5 kDa) produced less purified filtrate probably due to the partial retention of inulin molecules. The increase of transmembrane pressure (TMP) from 1 to 2 bar was beneficial increasing the filtration flux and the juice purity. The following increasing of TMP to 4 bar leaded to the additional membrane fouling and did not increase the filtration flux. The ultrafiltration behavior of obtained chicory juice can be successfully described by the model of intermediate blocking mechanism.

Application of Pulsed Electric Fields for Root and Tuber Crops Biorefinery

21

Potato Potato is one of the most important food crop in the world. Potato is a rich source of starch (
82 % of dry matter), protein, vitamins, minerals, and a number of health promoting phytonutrients (phenolics, flavonoids, anthocyanins, and carotenoids). Most of the vitamins in potato are located in a vicinity of skin. Moreover, the potato skins are rich in steroidal alkaloids useful for production of hormones, antibiotics, and anticancer drugs. Potato has been frequently used as a model system for testing electroporation effects and studying of reversible electroporation, transient viscoelastic behavior, different stress-induced effects, metabolic responses, and electrostimulated effects in plant tissues (Grimi 2009; Praporscic 2005; Shynkaryk 2006). Effects of PEF treatment on the textural and compressive properties of potato have been studied in details (Grimi 2009). It was shown that application of only PEF treatment was not sufficiently effective for complete elimination of the textural strength. However, mild thermal pretreatment at 45–55  C allowed increasing the PEF efficiency. Potato has been also used for investigation of temperature and PEF protocol effects on characteristic damage time, dehydration, freezing, and drying (Praporscic 2005; Shynkaryk 2006). Effects of PEF treatment (electric field strength of E = 0.2–1.1 kV/cm, ti = 20 μs, pulse frequency of f = 50 Hz, total pulse number of n = 540, specific energy of W = 1–10 kJ/kg, and temperature of T = 20  C) on a microstructure of potato tubers have been studied (Faridnia et al. 2015). It was demonstrated that the orientation of a tuber towards the electrodes and presence of a skin greatly affected the impact of PEF on cell disruption and viability. The potato cells located in the inner medulla were more electroporated as compared to cells located in the outer medulla. PEF treatment has been useful for facilitation of extraction of the steroidal alkaloids from potato peels (Hossain et al. 2015). PEF-assisted (E = 0.75 kV/cm and tPEF = 600 μs) solid-liquid extraction (methanol) resulted in extraction yield 1856.2 μg/g dried potato PEFls that was 99.9 % higher than that of the untreated peels. The attempts of extraction-oriented practical application of PEF have been also done. As an example, PEF application for facilitation of starch extraction from potato and enhancement of the extractability of an anthocyanin-rich pigment have been reported (Toepfl 2006). An industrial prototype for starch extraction from potato has been developed by Propuls GmbH, Bottrop, Germany (Toepfl 2006). The automated flow of potatoes was coming from a feeding funnel with two cross electrodes. After passing the water-filled electrode section, the electrically treated potatoes were separated from water with a screw conveyer for their further treatment.

Conclusions In recent decades, pulsed electric fields (PEF) have been applied for assistance of root and tuber crops biorefinery. The different PEF experiments with sugar beets, sugar cane, red beets, chicory, and potatoes have been performed in order to enhance

22

E. Vorobiev and N. Lebovka

extraction of sugar, starch, proteins, amino and organic acids, and other high-added values. Different PEF-assisted processes such as cold expression, diffusion, combined pressing-diffusion, and combined pressing-diffusion-liming techniques have been also tested. Efficiency of cold (nonthermal) recovery of sugars, colorants, inulin, and starch has been demonstrated. Application of PEF provides a significant reduction of the energy requirement during extraction and dehydration of the extracted cossettes. Commonly a purified cold juice is less colored, less turbid, and more pure comparatively to the purified “thermal” juice. First important steps for practical implementations of PEF-assisted root and tuber crops biorefinery on the pilot and industrial scales have been already done. It is expected that biorefinery applications of electroporation can simplify (or even eliminate) the very complicated and polluting carbonic purification process and future technology will use less energy and be less polluting. Acknowledgments The authors appreciate the support from the COST Action TD1104 (EP4Bio2Med – European network for development of electroporation-based technologies and treatments).

Cross-References A list of related chapters that may be of further interest: ▶ Effect of Pulsed Electric Fields on Food Constituents ▶ Effect of Pulsed Electric Fields on Food Proteins ▶ Extraction of Biocompounds from Forest Feedstocks (Wood, Leafs, Mushrooms) ▶ Polyphenols and Proteins Extraction from Rapeseed Stems and Leaves Assisted ▶ Pulsed Electric Fields and High-Voltage Electrical Discharge Assisted Extraction of Biocompounds from Vine Shoots ▶ Pulsed Electric Fields and High-Voltage Electrical Discharge Assisted Extraction of Valuable Compounds from Grape Seeds ▶ Pulsed Electric Fields Assisted Extraction from Exotic Fruits Residues ▶ Pulsed Electric Fields Assisted Extraction of Plant Pigments ▶ Pulsed Electric Fields Assisted Extraction of Valuable Compounds from Grape Pomace ▶ Pulsed Electric Fields Induced Selective Extraction of Biocompounds from Stevia Leaves ▶ Pulsed Electric Fields Processing for Grape Wastes Biorefinery ▶ Pulsed Electric Fields Processing for Lignocellulosic and Forest Based Biorefinery ▶ Selective Extraction of Molecules

References Almohammed F, Mhemdi H, Vorobiev E (2016) Pulsed electric field treatment of sugar beet tails as a sustainable feedstock for bioethanol production. Appl Energy 162:49–57

Application of Pulsed Electric Fields for Root and Tuber Crops Biorefinery

23

Arnold J, Frenzel S, Michelberger T, Scheuer T (2010) Extraction of constituents from sugar beet chips, US Patent 7,695,566 Barba FJ, Parniakov O, Pereira SA, Wiktor A, Grimi N, Boussetta N, Saraiva J, Raso J, MartinBelloso O, Witrowa-Rajchert D, Bals O, Vorobiev E, Lebovka N (2015) Current applications and new opportunities for the use of pulsed electric fields in food science and industry. Food Eng Rev 77:773–798 Bouzrara H (2001) Amélioration du pressage de produits végétaux par Champ Electrique Pulsé. Cas de la betterave à sucre. PhD Thesis, Universite de Technologie de Compiegne, Compiegne El-Belghiti K (2005) Amélioration de l’extraction aqueuse de solutés des produits végétaux par champs électriques pulsés. PhD Thesis, Universite de Technologie de Compiegne, Compiegne Eshtiaghi MN, Knorr D (2002) High electric field pulse pretreatment: potential for sugar beet processing. J Food Eng 52:265–272 Eshtiaghi MN, Yoswathana N (2012) Laboratory scale extraction of sugar cane using high electric field pulses. World Acad Sci Eng Technol 65:1217–1222 Faridnia F, Burritt DJ, Bremer PJ, Oey I (2015) Innovative approach to determine the effect of pulsed electric fields on the microstructure of whole potato tubers: use of cell viability, microscopic images and ionic leakage measurements. Food Res Int 77:556–564 Fincan M, DeVito F, Dejmek P (2004) Pulsed electric field treatment for solid–liquid extraction of red beetroot pigment. J Food Eng 64:381–388 Grimi N (2009) Vers l’intensification du pressage industriel des agroressources par champs electriques pulses: etude multi-echelles. PhD Thesis, Universite de Technologie de Compiegne, Compiegne Hossain MB, Aguiló-Aguayo I, Lyng JG, Brunton NP, Rai DK (2015) Effect of pulsed electric field and pulsed light pre-treatment on the extraction of steroidal alkaloids from potato peels. Innovative Food Sci Emerg Technol 29:9–14 Jemai AB, Vorobiev E (2003) Enhanced leaching from sugar beet cossettes by pulsed electric field. J Food Eng 59:405–412 Jemai AB, Vorobiev E (2006) Pulsed electric field assisted pressing of sugar beet slices: towards a novel process of cold juice extraction. Biosyst Eng 93:57–68 Kotnik T, Kramar P, Pucihar G, Miklavcic D, Tarek M (2012) Cell membrane electroporation – Part 1: the phenomenon. IEEE Electr Insul Mag 28:14–23 Lebovka NI, Shynkaryk MV, El-Belghiti K, Benjelloun H, Vorobiev E (2007) Plasmolysis of sugarbeet: pulsed electric fields and thermal treatment. J Food Eng 80:639–644 Loginova K (2011) Mise en oeuvre de champs electriques pulses pour la conception d’un procede de diffusion a froid a partir de betteraves a sucre et d’autres tubercules alimentaires (etude multiechelle). PhD Thesis, Universite de Technologie de Compiegne, Compiegne Lopez N, Puertolas E, Condon S, Alvarez I (2009a) Enhancement of the extraction of betanine from red beetroot by pulsed electric fields. J Food Eng 90:60–66 Lopez N, Puertolas E, Condon S, Raso J, Alvarez I (2009b) Enhancement of the solid–liquid extraction of sucrose from sugar beet (Beta vulgaris) by pulsed electric fields. LWT- Food Sci Technol 42:1674–1680 Mhemdi H (2013) Etude d’une technologie d’extraction froide par les champs électriques pulsés pour une utilisation économe et propre des agro-ressources. UTC, Compiègne, France Naqvi M, Yan J (2015) Biorefinery: production of biofuel, heat, and power utilizing biomass. In: Yan J (ed) Handbook of clean energy systems. Wiley Online Library, New York, pp 169–186 Praporscic I (2005) Influence du traitement combine par champ electrique pulse et chauffage modere sur les proprietes physiques et sur le comportement au pressage de produits vegetaux. PhD Thesis, Universite de Technologie de Compiegne, Compiegne Sack M, Schultheiss C, Bluhm H (2005) Triggered Marx generators for the industrial-scale electroporation of sugar beets. IEEE Trans Ind Appl 41(3):707–714 Shynkaryk M (2006) Influence de la permeabilisation membranaire par champ electrique sur la performance de sechage des vegetaux. PhD Thesis, Universite de Technologie de Compiegne, Compiegne

24

E. Vorobiev and N. Lebovka

Shynkaryk MV, Lebovka NI, Vorobiev E (2008) Pulsed electric fields and temperature effects on drying and rehydration of red beetroots. Drying Technol 26:695–704 Toepfl S (2006) Pulsed electric fields (PEF) for permeabilization of cell membranes in food- and bioprocessing. Applications, process and equipment design and cost analysis. Technical University of Berlin, Berlin, Germany Vidal OP, Vorobiev E (2011) Procédé et installation de traitement des tissus végétaux pour en extraire une substance végétale, notamment un jus. International Patent, Brevet déposé en France du 03.05.2010. N 1053413, WO2011/138248 A1 Vorobiev E, Lebovka N (2010) Enhanced extraction from solid foods and biosuspensions by pulsed electrical energy. Food Eng Rev 2:95–108 Zhu Z (2014) Study of pulsed electric fields (PEF) assisted inulin extraction from chicory root and chicory juice purification. PhD Thesis, Universite de Technologie de Compiegne, Compiegne

Selective Extraction of Molecules from Biomaterials by Pulsed Electric Field Treatment Eugene Vorobiev and Nikolai Lebovka

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Actions of Pulsed Electric Fields on Biomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 PEF-Assisted Solute Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 PEF-Assisted Extraction from Fruit and Vegetable Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 PEF-Assisted Extraction from Microorganisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 PEF-Assisted Extraction by Pressing (Expression) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Abstract

In present time, the selective extraction of molecules assisted by pulsed electric fields (PEF) has gained growing interest. The PEF can serve as an effective tool for improvement of traditional extraction processes in biotechnological and food industries. The PEF-assisted extraction can be applied in two different modes by: (a) diffusion in solvent (solvent extraction) and (b) application of pressing

E. Vorobiev (*) Laboratoire de Transformations Intégrées de la Matière Renouvelable, EA 4297, Centre de Recherches de Royallieu, Sorbonne Universités, Université de Technologie de Compiègne, Compiègne, France e-mail: [email protected] N. Lebovka Laboratoire de Transformations Intégrées de la Matière Renouvelable, EA 4297, Centre de Recherches de Royallieu, Sorbonne Universités, Université de Technologie de Compiègne, Compiègne, France Department of Physical Chemistry of Disperse Minerals, Institute of Biocolloidal Chemistry named after F. D. Ovcharenko, NAS of Ukraine, Kyiv, Ukraine e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_163-1

1

2

E. Vorobiev and N. Lebovka

procedures (expression). The PEF-assisted extraction methods are suitable for the selective recovery and extraction of sugar, inulin, starch, proteins, polysaccharides, polyphenols, pigments, flavor compounds, phytochemicals, and other highvalue components. PEF can be applied for treatment of different fruits and vegetables (apple, carrot, table beet, etc.), leaves (tea, spinach), herbs, mushrooms, and suspensions of cells (yeasts, microalgae, etc.). The PEF-assisted techniques have low-operating costs and can be applied to the thermally sensitive products in “cold” mode without deterioration of color, flavor, vitamins, and other important nutrients of foods. “Cold” refers to the fact that only mild heat is added during extraction procedure. In this chapter, the main extraction procedures assisted by PEF are defined. Different examples of PEF-assisted solvent extraction and expression of useful components from different food and biomaterials are collected. The emphasis is placed on the selectivity of extraction. Following the handbook format, the chapter is concise and covers various common features of PEF processing, review on recent efforts in the field, as well as several aspects of our own research. Keywords

Pulsed electric fields (PEF) • Selective extraction • Extraction by diffusion • Extraction by pressing (expression)

Introduction The extraction of molecules from different materials of biological origin (including foods) has a long history, and it was used for production of sugar, fruit juices, wine, pigments and colorants, polyphenols, proteins, extracts from herbs, perfumes, oils, and waxes (Bart and Stephan 2011). The conventional extraction techniques are based on diffusion in solvent, extraction by cold pressing (expression), Soxhlet extraction, maceration, and hydrodistillation (Lebovka et al. 2011). However, in majority cases, the conventional techniques have low extraction selectivity and separation and purification of extracted bioactive substances commonly requires application of costly supplementary procedures. Moreover, extraction techniques require costly and high-purity solvents and can provoke decomposition or degradation of bioactive substances. Recently, the nonconventional extraction processes assisted by enzymes, supercritical fluids, high pressure, ultrasound, microwaves, and pulsed electric fields have been developed (Azmir et al. 2013). The principal advantages of PEF-assisted extraction compared to conventional extraction techniques can be formulated as (Barba et al. 2015): • Increased mass transfer • Improved extraction yield • Decreased processing time

Selective Extraction of Molecules from Biomaterials by Pulsed Electric Field. . .

3

• Decreased severity of the conventional extraction parameters (i.e., decrease of extraction temperature, lower solvent concentration) • Reduction of heat-sensitive compounds degradation (e.g., flavors, proteins) • Facilitation of purified extract (i.e., reducing grinding) • Reduction of energy costs and environmental impact Moreover, applications of hybrid methods combining at different steps conventional and non conventional extraction can be very useful for enhancement of extraction selectivity and efficiency. This chapter concentrates on the analysis of recent examples of PEF-assisted diffusion extraction and expression techniques in application to different materials of biological origin. The special emphasis is placed on the selectivity of extraction of bioactive or useful compounds.

Actions of Pulsed Electric Fields on Biomaterials The action of PEF treatment on biomaterial is based on dielectric damage of cell membranes that is the very thin barriers (h
5 nm) separating the interior and exterior of cells. The dielectric damage requires some threshold potential across membrane, typically of order ucm = 0.2–1 V (Weaver and Chizmadzhev 1996). At this threshold potential, the breakdown field strength applied to membrane is of order of 106 V/cm that noticeably exceeds breakdown field strength of mineral oil (
105/cm). The electroporation concept is very popular for explaining the dielectric breakdown of the biological membrane. It is supposed that in the vicinity of breakdown, the external electric field forms pores (holes) inside membranes. Transmembrane potential across membrane in the external field E is proportional to the diameter of the biological cell d, um
dE, and electroporation of small microbial cells requires higher fields than electroporation of larger plant cells. For the plant tissues, efficient electroporation can be observed at the relatively moderate electric field strengths within the interval E = 200–1000 V/cm and time of PEF treatment tPEF = 104–101 s. The electroporation of yeasts or microalgae commonly requires greater electric fields (E = 20–50 kV/cm) and shorter treatment times (tPEF = 105–103 s) (Vorobiev and Lebovka 2008). Figure 1 presents the electroporation diagram in E versus tPEF coordinates. In a nondischarge mode, the PEF-technique is relatively nonthermal and nondestructive. Typically the PEF-treatment is applied without significant elevation of temperature of order ΔT  10  C. Moreover, the PEF induces preferential or selective damage of membranes and does not affect noticeably a structure of cell walls. A relative mass of membrane component in fruits or vegetables is negligibly small in compare with the total mass of tissue (
0.05 %). Application of PEF in discharge mode (high voltage electrical discharges, HVED) can induce the electroporation of membranes and damage of cell walls, mechanical fragmentation of materials, temperature elevation, and changes in other characteristics of treated materials. Commonly, the HVED technique is

4

E. Vorobiev and N. Lebovka

105

Electric field strength, E, V/cm

Thermal Electroporation & Ohmic heating (OH) 104

NonThermal

Thermal threshold, ΔT≈10°C

103

Electroporation

Electroporation threshold 102

E µ1/ tPEF No electroporation ms

101

10-6

ms 10-5

10-4

10-3

s 10-2

10-1

1

Duration of PEF treament, tPEF, s

10

100

Fig. 1 Schematic presentation of electroporation diagram. The dashed line corresponds to the electroporation threshold. The solid line corresponds to the 10  C limit of temperature elevation. The hatched area corresponds to the nonthermal electroporation (From (Maity 2015) with permission)

more injurious and nonthermal. Such technique accompanies with electrical breakdown, propagation of streamer, bubble formation and cavitations, light emission, appearing of localized regions with high pressure, and formation of shock and acoustic waves. In addition, this treatment can produce a large quantity of radical species, ozone, and provokes the fragmentation of solid particles suspended in the treated mixtures. The electroporation efficiency noticeably depends upon the details of pulse protocols (electric field strength, pulse duration, polarity, distance between pulses, etc.. . .), properties of biomaterials (size of the particles and cells, degree of maturity, electrophysical properties), temperature, conductivity, and pH of outside media (Raso and Heinz 2006; Vorobiev and Lebovka 2008). For example, the longer pulse durations (ti > 100–1000 μs) were found to be more effective for treatment of plant tissues, and their effect was particularly pronounced at the room temperature and moderate electric fields (E = 100–300 V/cm). Application of bipolar pulses allows reduction of the food electrolysis phenomena and asymmetry of membrane

Selective Extraction of Molecules from Biomaterials by Pulsed Electric Field. . .

5

1

Disintegration index, Z

E=400 V/cm ti= 1000 µs

Orange

0.8 Apple

Orange

Carrot

Potato

Courgette

Banana

Apple 0.6 100 mm

Potato 0.4

Carrot 0.2

Banane Courgette 10-3

10-2

10-1

Treatment time, tPEF ,s

100

Fig. 2 Disintegration index, Z, versus time of PEF treatment, tPEF, for different plant tissues (Ben Ammar et al. 2011)

damage in the cells. Moreover, a complex protocol with adjustable long pause between the trains allows fine regulation of tissue disintegration without noticeable temperature elevation during the PEF treatment. For the evaluation of the damage degree (i.e., ratio of damaged cells to the total number of cells), the electrical conductivity disintegration index can be used (Vorobiev and Lebovka 2008) Z ¼ ðσ  σ i Þ=ðσ d  σ i Þ

(1)

Here σ is the electrical conductivity measured at low frequency (1–5 kHz), and indexes “i” and “d” refer to the conductivities of intact and totally damaged cell, respectively. Figure 2 shows typical examples of Z(tPEF) dependencies for different plant tissues (Ben Ammar et al. 2011). The Z(tPEF) curves can be rather different in dependence of temperature T, applied electric field strength E, size of the cells, as well as the differences in the tissue electrical conductivity contrast σ d/σ i. For vegetable and fruit tissues (apple, potato, cucumber, aubergine, pear, banana, and carrot), the optimal values of power consumption are within 1–15 kJ/kg. They are noticeably lower as compared to other methods of treatment like mechanical (20–40 kJ/kg), heating, or freezing/thawing (>100 kJ/kg).

6

E. Vorobiev and N. Lebovka

PEF-Assisted Solute Extraction PEF-Assisted Extraction from Fruit and Vegetable Tissues Commonly the procedures of solute extraction require a large quantity of energy and utilize the high temperatures (hot extraction) and/or polluting solvents. Increasing the temperature enhances extraction of the bioactive compound of interest as well as other soluble (and undesirable) components (e.g., pigments, pectins, and polymers) and reduces the selectivity of extraction. Moreover, at high temperature, the degradation of thermolabile compounds (e.g., anthocyanins, flavonols, components of essential oils) can occur. A prolonged extraction time results in an increased yield of extracted compound. However, it requires more power consumption and can result in enhanced degree of thermal degradation. The use of suitable solvent may be also critical for degradation or coagulation of bioactive compounds or their recovery from the target material. Moreover, the solvent consumption, its cost and toxicity, and requirement post-extraction purification step should be also taken into account. As example, the extraction of sugar from beetroot slices can be referred (Loginov et al. 2011). The conventional process requires prolonged hot water diffusion at 70–75  C. Such aqueous diffusion results in a noticeable thermal degradation of the cell wall and extraction of undesirable components such as pectins, oligo- and polymolecular compounds into the juice. The extraction at high temperatures can also accelerate various chemical reactions between extracted components of the juice (e.g., color development due to Maillard reaction). Extraction of high molecular weight compounds and formation of undesirable products of chemical reactions result in decrease of extracted juice quality (Loginov et al. 2011). As a result, the purification of juice requires a very complex multistaged processes and large quantity of chemical reagents, e.g., lime (Loginova (Sereda) 2011). Recent efforts have shown the promising ways for different applications of PEF-assisted selective extraction, e.g., sucrose from sugar beetroot, betalain from red beet, inulin from chicory, beta-carotene from carrot, phenolics from grapes, extraction of valuable components from apple, paprika, carrot, red cabbage, etc. (Vorobiev and Lebovka 2008). The PEF treatment can considerably enhance the diffusivity of useful component using green solvents at ambient temperatures, without heating (“cold” diffusion), and application of such technology allows prevention of the thermal degradation of the cell walls and extraction of other undesirable components. The most important industrial expectation is related with PEF-assisted aqueous extraction of sugar from sugar beet. Effects of electroporation in sugar beets have been extensively investigated in laboratory scale in numerous studies (Eshtiaghi and Knorr 1999; Loginova (Sereda) 2011; Lopez et al. 2009b). The juice from the cold diffusion had the lowest quantity of pectin and the color of juice was systematically 3–4 times less intensive than the color of commercial juices (Loginova 2011). In addition, the juice purity was higher for slices treated by PEF than for untreated slices after thermal diffusion at 70  C. A scale-up study demonstrated that the combination of mild heating at 50  C and PEF-treatment can be a useful tool for

Selective Extraction of Molecules from Biomaterials by Pulsed Electric Field. . .

7

shortening the diffusion time. The use of PEF-assisted diffusion technique in the sugar industry is very promising, more environmentally friendly, efficient, and thereby reducing energy consumption and economical costs. The promising perspectives of PEF-assisted diffusion aqueous extraction of inulin from chicory (Cichorium intybus) roots (E = 100–600 V/cm) have been also demonstrated (Loginova (Sereda) 2011). The PEF technology noticeably accelerates the diffusion of inulin from chicory slices even at moderate temperatures. The very efficient extractability of betalain (red-purple pigments) from red beet (Beta vulgaris L.) by PEF-assisted aqueous diffusion has been also demonstrated (Loginova (Sereda) 2011; Lopez et al. 2009a). The “cold” extraction by diffusion (T = 30  C) attained the highest colorant yield (
95 %) with the lowest level of colorant degradation (
10 %). The temperature contribution to electroporation efficiency was similar for chicory and sugar beet tissues, and it reflected the synergistic effect of the simultaneous thermal and PEF application (Loginova (Sereda) 2011). Nowadays, there exist many interesting examples of PEF-assisted solute extraction of valuable food, medical, cosmetic, and pharmaceutical components such as podophyllotoxin (rhizomes of P. peltatum), pure collagen proteins (bovine bone), chitosan (shrimp shells), polyphenols (fresh tea leaves), crocin, safranal, picrocrocin (saffron Crocus sativus), alkaloid Guanfu base A, GFA (Chinese medicinal herb Aconitum coreanum), polysaccharides (corn silk), ginseng extract, and exopolysaccharides (Tibetan spiritual mushroom broth) (see, (Barba et al. 2015) for a review). Over the last years, the PEF-assisted diffusion has attracted considerable attention for the valorization of high–added value compounds from by-products and food wastes, which can be used for different purposes (e.g., food additives and/or nutraceuticals). The PEF-assisted recovery of valuable compounds from pomace, seeds, peels, kernels, husks of grapes, marks, and oil-cakes has been discussed (Mahnič-Kalamiza et al. 2014). The PEF-assisted diffusion methods have been also applied for biorefinery of agricultural, and forestry residues (stems, sawdust, leaves, and bark), energy crops and municipal wastes, vine shoots, leafs, debris, etc. Processed biological materials also include grape pomace (skins, pulp, seeds, and stems), vine shoots, different leaves, seeds, peels, kernels, hulls, lignocellulosic biomass from terrestrial plants, energy crops and crop residues, agricultural and forestry residues (stems, sawdust, leaves and bark), and food wastes and beer waste brewing yeasts. These materials are rich in bioactive compounds, especially polyphenols (anthocyanins, catechins, flavonol glycosides, phenolic acids etc.), with antioxidant, antibacterial, antifungal, anticancer, and antiviral properties.

PEF-Assisted Extraction from Microorganisms Examples of PEF-assisted diffusion include also applications to different biological suspensions of yeast and microalgae cells. The possibility of highly-selective and effective extraction of different intracellular components (ions, saccharides, enzymes, proteins, and nucleic acids) from yeast cells has been recently

8

E. Vorobiev and N. Lebovka

demonstrated (Liu et al. 2013). For example, the combination of PEF pretreatment (E = 40 kV/cm) and mild temperature ( 1 ms) and, therefore, dynamic changes during the pulse would be not recorded because pulse durations are typically 34  C. The different inactivation degrees of lipase observed in these studies might be due to different processing conditions and media used.

14

M.M. Poojary et al.

Table 2 Effect of PEF treatment on various enzymes Enzyme Treatment source medium Alkaline phosphatase Bovine Buffer pH 9.8 Milk

α-Amylase Bacillus licheniformis

Deionized water

Bacillus Buffer subtilis Glucose oxidase Aspergillus Buffer niger pH 5.1 Lipase Milk

Milk

Wheat germ

Distilled water

Pseudomonas fluorescens

Simulated milk ultrafiltrate

Lipoxygenase Soybean Green pea Tomato juice Soya milk

Distilled water Green pea juice Tomato juice Soya milk

% residual activity

References

95

(Ho et al. 1997)

26

(Van Loey et al. 2002)

35

(Castro et al. 2001b)

E: ~80 kV/cm; τ, 2 μs; t, 2 s; f, 0.5 Hz; N, 30; Tmax, 20  C a,c E: 15 kV/cm; τ, 2.5 μs; N, 80; f, 1, 128 Hz; Tmax, 60  C

10

(Ho et al. 1997)

122.13

(Tian et al. 2016)

a,d

E: ~63 kV/cm; τ, 2 μs; t, 2 s; f, 0.5 Hz; N, 30; Tmax, 20  C

25

(Ho et al. 1997)

a,d

E: 21.5 kV/cm; N, 20; f, 1–22 Hz; Tmax, 45–50  C a,d E: ~87 kV/cm; τ, 2 μs; t, 2 s; f, 0.5 Hz; N, 30; Tmax, 20  C a,b,d E: 27.4–37.3 kV/cm; τ, 4 μs; t, 314.5 μs; N, 80–100; Tmax, > 34  C

40

(Grahl and Markl 1996) (Ho et al. 1997)

37.9–87

(Bendicho et al. 2002)

a,c E: 10–30 kV/cm; τ, 5–40 μs; N, 1–100; f, 1 Hz a,c E: 2.5–20 kV/cm; τ, 1 μs; N, 100–400; f, 1 Hz b,c E: 35 kV/cm; τ, 1–7 μs; t, 1, 000 μs; f, 50–250 Hz a,c E: 40 kV/cm; τ, 1 μs; t, 100 μs; f, 15 Hz; Tmax, 50  C

90–100

(Van Loey et al. 2002) (Van Loey et al. 2002) (Aguiló-Aguayo et al. 2009) (Riener et al. 2008b)

Treatment conditions E: ~80 kV/cm; τ, 2 μs; t, 2 s; f, 0.5 Hz; N, 30; Tmax, 20  C a,c E: 10 kV/cm; τ, 1–40 μs; N, 200; f, 1 Hz; Tmax, 70  C a,d E: 18.8 kV/cm; τ, 400 μs; N, 70; Tmax, 43.9  C a,d

a,d

15

100 81 15.5 %

(continued)

Impact of Pulsed Electric Fields on Enzymes

15

Table 2 (continued) Enzyme source Lysozyme Egg white

Treatment medium Buffer pH 6.2

Pectin methylesterase Tomato NaCl solution Distilled water Buffer pH 7 Orange

Distilled water Orange juice Buffer pH 7 Orange juice

Red grape Banana

Grapefruit juice Buffer pH 7

Carrot

Peroxidase Soya bean

Milk

Buffer pH 6.0 Raw milk

Treatment conditions E: 15 kV/cm; τ, 2 μs; t, 2 s; f, 1, 000 Hz; Tmax, 20  C b,c E: 35 kV/cm; τ, 2 μs; t, 1, 200 μs; f, 1,000 Hz; Tmax, > 60  C b,c E: 35 kV/cm; τ, 2 μs; t, 1, 200 μs; f, 1,000 Hz; Tmax, > 20  C b,d

E: 24 kV/cm; τ, 20 μs; t, 8, 000 μs; N, 400; f, 15 Hz; Tmax, 15  C a,c E: 10–30 kV/cm; τ, 40 μs; N, 1, 000; f, 10 Hz a,c E: 16.8 kV/cm; τ, 4 μs; t, 1, 600 μs; f, 0.5 Hz; Tmax, 43.4  C a,c E: 10–30 kV/cm; τ, 40 μs; N, 1, 000; f, 10 Hz b,c E: 25 kV/cm; τ, 2 μs; t, 250 μs; Tmax, 50  C a,c E: 16.8 kV/cm; τ, 4 μs; t, 1, 600 μs; f, 0.5 Hz; Tmax, 67.2  C b,c E: 35 kV/cm; τ, 1, 500 μs; t, 1, 600 μs; f, 200 Hz; Tmax, 37.5  C b,c E, 40 kV/cm; τ, 1 μs; t, 100 μs; f, 15 Hz; Tmax, 50  C a,c E, 16.8 kV/cm; τ, 4 μs; t, 1, 600 μs; f, 0.5 Hz; Tmax, 43.4  C a,c E: 16.8 kV/cm; τ, 4 μs; t, 1, 600 μs; f, 0.5 Hz; Tmax, 67.2  C a,d

E: ~75 kV/cm; τ, 2 μs; t, 2 s; f, 0.5 Hz; N, 30; Tmax, 20  C a,c E: 6–20 kV/cm; τ, 1–40 μs; N, 1, 000; f, 1 Hz; Tmax, 70  C a,d

% residual activity

References

50

(Ho et al. 1997)

61.9

(Zhao et al. 2007; Zhao and Yang 2008b) (Zhao and Yang 2010)

62

6.2

(Giner et al. 2000)

90

(Van Loey et al. 2002) (EspachsBarroso et al. 2006) (Van Loey et al. 2002) (Yeom et al. 2002) (EspachsBarroso et al. 2006) (Elez-Martínez et al. 2007)

13

90 10 13

21.9

3.2 65

17

(Riener et al. 2009) (EspachsBarroso et al. 2006) (EspachsBarroso et al. 2006)

70

(Ho et al. 1997)

40

(Van Loey et al. 2002) (continued)

16

M.M. Poojary et al.

Table 2 (continued) Enzyme source Horseradish

Treatment medium Buffer

Orange juice

Orange juice

Protease Pseudomonas fluorescens Bacillus subtilis

Tryptic soy broth Skim milk

Polyphenol oxidase Apple Apple juice

Apple juice with ascorbic acid Buffer pH 6.5 Buffer

Treatment conditions b,d E: 5–25 kV/cm; τ, 1.5 μs; N, 207–1, 242; f, 10 Hz; Tmax, > 40  C b,c E: 35 kV/cm; τ, 4 μs; t, 1, 500 μs; f, 200 Hz; Tmax, > 40  C E: 18 kV/cm; τ, 2 μs; N, 20; f, 0.25 Hz b,c E: 35.5 kV/cm; τ, 4 and 7 μs; t, 866 μs; f, 111 Hz; Tmax, > 46  C b

E: 7 kV/cm; τ, 40 μs; N, 1, 000; f, 10 Hz; Tmax, 60  C b,c E: 40 kV/cm; τ, 1 μs; t, 100 μs; f, 15 Hz; Tmax, 50  C b E: 35 kV/cm; τ, 3 μs; t, 1, 500 μs; f, 200 Hz; Tmax, 40  C a,c

Pear

Buffer pH 6.5

Peach

Buffer pH 4.5

a,d E: 50 kV/cm; τ, 2 μs; t, 2 s; f, 0.5 Hz; N, 30; Tmax, 20  C a,d E: 25 kV/cm; τ, 0.6 μs; t, 744 μs; f, 10 Hz; Tmax, > 40  C a,d E: 22.30 kV/cm; τ, 200 μs; t, 6, 000 μs; N, 400; f, Hz; Tmax, > 25  C a,d E: 24.30 kV/cm; τ, 400 μs; t, 5, 000 μs; Tmax, > 25  C

Buffer pH 2.0

a,d E: 15 kV/cm; τ, 2 μs; t, 2 s; f, 0.5 Hz; N, 30; Tmax, 20  C

Mushroom

Pepsin Porcine stomach mucose

% residual activity 65.3–83.3

References (Zhong et al. 2005)

0

(Elez-Martínez et al. 2006)

20

(Vega-Mercado et al. 2001) (Bendicho et al. 2003)

19

78 32 20

(Van Loey et al. 2002) (Riener et al. 2008a) (Schilling et al. 2008)

60

(Ho et al. 1997)

24.8

(Zhong et al. 2007)

38

(Giner et al. 2001)

30

(Giner et al. 2002)

250 %

(Ho et al. 1997)

E, electric field strength; τ, pulse width; t, treatment duration; N, number of pulses; f, pulse frequency; and Tmax, maximum temperature employed/achieved during PEF treatment a Batch mode treatment b Continuous mode treatment c Square wave pulse d Exponential decay pulse

Impact of Pulsed Electric Fields on Enzymes

17

Lipoxygenase Lipoxygenases (LOX) are a family of nonheme iron-containing enzymes that catalyze the insertion of dioxygen into polyunsaturated fatty acids. Reports regarding the effect of PEF treatment on LOX inactivation differ widely, which may be due to the differences between the properties of the treated materials, processing conditions, or technical characteristics of the applied PEF equipment. Aguiló-Aguayo et al. (2009) investigated the effect of PEF treatment on LOX activity in tomato juice at the following conditions: square-shaped pulses of 35 kV/cm for 1, 000 μs, pulse width from 1 to 7 μs, frequencies from 50 to 250 Hz, and monopolar or bipolar modes. An increase in frequency or pulse width resulted in a decrease of residual LOX activity. The lowest residual LOX activity (81 %) was observed in tomato juice treated at 250 Hz for 7 μs in bipolar mode. The polarity of the treatment had significant impact on LOX activity. Similarly, Van Loey et al. (2002) reported a minimum activity of soybean LOX in distilled water when 1, 000 monopolar pulses were used at 1 Hz. In another study, the combination of preheating to 50  C and a PEF treatment of 100 μs at 40 kV/cm revealed high inactivation level (84.5 %) of soya milk LOX (Riener et al. 2008b).

Pectin Methylesterase Pectin methylesterase (PME) is a pectic enzyme widely distributed in higher plants and can be synthesized by different microorganisms. Generally, the reduction of PME activity is needed for cloud stability of cloudy fruit juices due to the reason that PME catalyzes the hydrolysis of the methyl ester groups from pectin and consequently generating acidic pectin with lower esterification and methanol degree (Espachs-Barroso et al. 2006). Up to now, several studies have examined the effect of PEF processing on inactivation of PME in tomato (Giner et al. 2000), orange (Elez-Martínez et al. 2007), red grape (Riener et al. 2009), banana, and carrot (Espachs-Barroso et al. 2006). In a study conducted by Espachs-Barroso et al. (2006), PEF processing consisted of 40 μs square wave pulses, treatment time of 1.6 ms, frequency of 0.5 or 5 Hz, and electric field strengths between 13.2 and 19.1 kV/cm was used. They reported that the higher electric field, total treatment time, or pulse frequency resulted in the higher degree of PME inactivation in all sources. Maximum PME inactivation was observed in orange and tomato (87 %), carrot (83 %), and banana (45 %) at the most intense conditions. It was demonstrated that higher electric field strength, a longer treatment time, and a higher pulse frequency had a higher degree of PME inactivation. PME appears to be relatively stable against PEF treatment as no complete inactivation of this enzyme has been reported.

18

M.M. Poojary et al.

Peroxidase Peroxidase (POD) is part of a large group of enzymes known as oxidoreductases. The oxidation of a wide range of natural materials present in plants and especially those having aromatic groups can be catalyzed by POD. The mode of action involves the fabrication of high-energy free radicals, which are able to abstract hydrogen radicals from such materials. POD can promote large different reactions and thus displays versatility unsurpassed by any other enzyme. This enzyme contributes to deteriorative alterations in flavor, texture, color, taste, and nutrition in processed fruits and vegetables (Elez-Martínez et al. 2006). Several studies evaluated the feasibility of using PEF for the inactivation of POD enzyme. For instance, Zhong et al. (2005) investigated the inactivation and conformational change of horseradish POD induced by PEF processing. They reported that the activity of POD decreased with the increase in applied field strength (5–25 kV/cm) and pulse numbers (207–1, 242 pulses). After PEF treatment at 25 kV/cm for 207 pulses and 22 kV/cm for 1, 214 pulses, an immediate reduction of 16.7 % and 34.7 % in POD relative activity was observed, respectively. The inactivation of PEF-treated POD was found to be related to the conformational change of α-helix and the increase of relative fluorescence intensity. Similar results were observed by Ho et al. (1997) who found an activity reduction of 30 % in soybean POD suspended in a buffer solution when it was treated at 75 kV cm/1 for 60 μs. In another study, the inactivation of orange juice POD was studied under different PEF processing conditions (Elez-Martínez et al. 2006). POD was totally inactivated by PEF and the treatment was found to be more effective than thermal processing. In general, different levels of inactivation of PODs have been reported, where processing efficiency was dependent on the enzyme source and the PEF treatment condition used (Table 2).

Polyphenol Oxidase Polyphenol oxidase (PPO) is a copper containing enzyme found in many higher plants. Its residual activity is detrimental to the quality of fruits, vegetables and processed food products, resulting in effects such as browning, off-flavor, and loss of vitamins (Zhong et al. 2007). Several studies reported the inactivation of PPO using PEF treatment. For instance, the effect of PEF processing on laboratory scale resulted in complete deactivation of PPO when PEF treatment and preheating of the apple juices to 60  C were combined (Schilling et al. 2008). Similarly, the highest level of decrease in the POD activity (68 %) of freshly prepared apple juice was observed using a combination of preheating to 50  C and a PEF treatment time of 100 μs at 40 kV/cm (Riener et al. 2008a). In another study, PPO activities were reduced up to 3.15 % and 38.0 % from the initial value in apple extract at 24.6 kV/ cm and pear extract at 22.3 kV/cm both for 6 ms total treatment time, respectively (Giner et al. 2001). Zhong et al. (2007) found a decrease in PPO activity (16.9 %) after PEF treatment at 25 kV/cm for 124 μs. In another study, when 24 kV/cm and 60 μs were used to PPO suspended in a buffer solution, a reduction in PPO

Impact of Pulsed Electric Fields on Enzymes

19

inactivation (40 %) was achieved (Ho et al. 1997). These reports clearly suggest that PPO has a stable structure and much resistance to PEF treatments; however, a combination of PEF and preheating could substantially inactivate these enzymes.

Protease Although proteases seem to be more sensitive to PEF treatment than heat, the extent of their inactivation depends on the treatment conditions (e.g., higher electric field strengths, longer treatment times, and higher treatment temperatures). Bendicho et al. (2003) examined the impact of continuous PEF equipment on milk samples inoculated with Bacillus subtilis protease. Samples were subjected to PEF treatments of up to 866 μs of square wave pulses at field strengths from 19.7 to 35.5 kV/cm. In addition they tested the effects of different parameters such as pulse width (4 and 7 μs), pulse repetition rates (67, 89, and 111 Hz), and milk composition (skim and whole milk). Application of PEF treatment was demonstrated to reduce the activity of protease in milk. Maximum inactivation (81.1 %) was observed in samples treated at field strength of 35.5 kV/cm, 866 μs, and frequency of 111.11 Hz. VegaMercado et al. (2001) observed an activity reduction of 80 % in Pseudomonas fluorescens protease in tryptic soy broth treated at electric field strengths of 18 kV/ cm with 20 pulses. However, no inactivation of Pseudomonas fluorescens protease was observed when samples were treated in a casein–Tris buffer. A detailed mechanistic investigation taking different sources, however, is required to understand the effect of PEF on proteases.

Conclusions Available data and literature suggested that PEF could significantly influence the enzyme activity. However, the effect depends on the enzyme characteristics and treatment conditions. At sufficiently high-energy inputs, PEF can partially or completely inactivate enzymes by causing structural and conformational changes. Mild treatment conditions sometimes activate enzymes or enhance their activity. The conclusions about the effects of PEF on enzyme activity from different research groups are inconsistent probably due to different experimental setups and processing parameters. Although exact mechanism of activation or inactivation of enzymes by PEF is not yet established, reports suggested that it is often associated with changes in the structure or three-dimensional conformations of enzymes. Also, some of the reports strongly concluded that both electrochemical effects and ohmic heating associated with PEF contributed to the observed activity. It is still unclear whether enzymes behave differently in a solution compared to a food matrix under PEF treatment. Thus, further investigation is required to examine the influence of PEF on enzyme structure and function in real matrices; this could possibly allow an efficient usage of PEF technology in food and pharmaceutical sectors. The synergistic efficacy of PEF and temperature on enzymes should be mechanistically studied,

20

M.M. Poojary et al.

and this could allow reducing the overall energy costs and thus could allow a sustainable PEF processing technology.

References Aguiló-Aguayo I, Soliva-Fortuny R, Martín-Belloso O (2009) Effects of high–intensity pulsed electric fields on lipoxygenase and hydroperoxide lyase activities in tomato juice. J Food Sci 74: C595–C601 Bendicho S, Estela C, Giner J, Barbosa-Cánovas GV, Martín O (2002) Effects of high intensity pulsed electric field and thermal treatments on a lipase from Pseudomonas fluorescens. J Dairy Sci 85:19–27 Bendicho S, Barbosa-Cánovas GV, Martín O (2003) Reduction of protease activity in milk by continuous flow high–intensity pulsed electric field treatments. J Dairy Sci 86:697–703 Castro A, Swanson B, Barbosa-Canovas G, Dunker A (2001a) Pulsed electric fields denaturation of bovine alkaline phosphatase. In: Barbosa-Caanovas GV, Zhang QH (eds) Pulsed electric fields in food processing, 1st edn. Technomic Publishing, Lancaster, pp 65–82 Castro A, Swanson B, Barbosa-Canovas G, Zhang Q (2001b) Pulsed electric field modification of milk alkaline phosphatase activity. In: Barbosa-Canovas GV, Zhang QH (eds) Pulsed electric fields in food processing. Fundamental aspects and applications. Technomic Publishing Company, Lancaster, pp 65–82 Elez-Martínez P, Aguiló-Aguayo I, Martín-Belloso O (2006) Inactivation of orange juice peroxidase by high–intensity pulsed electric fields as influenced by process parameters. J Sci Food Agric 86:71–81 Elez-Martínez P, Suárez-Recio M, Martín-Belloso O (2007) Modeling the reduction of pectin methyl esterase activity in orange juice by high intensity pulsed electric fields. J Food Eng 78:184–193 Espachs-Barroso A, Van Loey A, Hendrickx M, Martín-Belloso O (2006) Inactivation of plant pectin methylesterase by thermal or high intensity pulsed electric field treatments. Innovative Food Sci Emerg Technol 7:40–48 Giner J, Gimeno V, Espachs A, Elez P, Barbosa-Cánovas GV, Martín O (2000) Inhibition of tomato (Licopersicon esculentum Mill.) pectin methylesterase by pulsed electric fields. Innovative Food Sci Emerg Technol 1:57–67 Giner J, Gimeno V, Barbosa-Cánovas GV, Martín O (2001) Effects of pulsed electric field processing on apple and pear polyphenoloxidases. Food Sci Technol Int 7:339–345 Giner J, Ortega M, Mesegue M, Gimeno V, Barbosa-Cánovas GV, Martín O (2002) Inactivation of peach polyphenoloxidase by exposure to pulsed electric fields. J Food Sci 67:1467–1472 Grahl T, Markl H (1996) Killing of microorganisms by pulsed electric fields. Appl Microbiol Biotechnol 45:148–157 Ho SY, Mittal GS, Cross JD (1997) Effects of high field electric pulses on the activity of selected enzymes. J Food Eng 31:69–84 Jaeger H, Meneses N, Knorr D (2009) Impact of PEF treatment inhomogeneity such as electric field distribution, flow characteristics and temperature effects on the inactivation of E. coli and milk alkaline phosphatase. Innovative Food Sci Emerg Technol 10:470–480 Lin S, Guo Y, You Q, Yin Y, Liu J (2012) Preparation of antioxidant peptide from egg white protein and improvement of its activities assisted by high–intensity pulsed electric field. J Sci Food Agric 92:1554–1561 Luo W, Zhang RB, Wang LM, Chen J, Guan ZC (2010) Conformation changes of polyphenol oxidase and lipoxygenase induced by PEF treatment. J Appl Electrochem 40:295–301 Ohshima T, Tamura T, Sato M (2007) Influence of pulsed electric field on various enzyme activities. J Electrost 65:156–161

Impact of Pulsed Electric Fields on Enzymes

21

Riener J, Noci F, Cronin DA, Morgan DJ, Lyng JG (2008a) Combined effect of temperature and pulsed electric fields on apple juice peroxidase and polyphenoloxidase inactivation. Food Chem 109:402–407 Riener J, Noci F, Cronin DA, Morgan DJ, Lyng JG (2008b) Combined effect of temperature and pulsed electric fields on soya milk lipoxygenase inactivation. Eur Food Res Technol 227:1461–1465 Riener J, Noci F, Cronin DA, Morgan DJ, Lyng JG (2009) Combined effect of temperature and pulsed electric fields on pectin methyl esterase inactivation in red grapefruit juice (Citrus paradisi). Eur Food Res Technol 228:373–379 Schilling S, Schmid S, Jäger H, Ludwig M, Dietrich H, Toepfl S, Knorr D, Neidhart S, Schieber A, Carle R (2008) Comparative study of pulsed electric field and thermal processing of apple juice with particular consideration of juice quality and enzyme deactivation. J Agric Food Chem 56:4545–4554 Tian M-L, Fang T, Du M-Y, Zhang F-S (2016) Effects of Pulsed electric field (PEF) treatment on enhancing activity and conformation of α–amylase. Protein J 35:154–162 Van Loey A, Verachtert B, Hendrickx M (2002) Effects of high electric field pulses on enzymes. Trends Food Sci Technol 12:94–102 Vega-Mercado H, Powers JR, Barbosa-Cánovas GV, Swanson BG (1995) Plasmin inactivation with pulsed electric fields. J Food Sci 60:1143–1146 Vega-Mercado H, Powers JR, Martin-Belloso O, Luedecke L, Barbosa-Canovas GV, Swanson BG (2001) Change in susceptibility of proteins to proteolysis and the inactivation of an extracellular protease from Pseudomonas fluorescens M3/6 when exposed to pulsed electric fields. In: Barbosa-Canovas GV, Zhang QH, Tabilo-Munizaga G (eds) Pulsed electric fields in food processing fundamental aspects and applications. Technomic Publishing Company, Lankaster, pp 105–120 Wang X, Li Y, He X, Chen S, Zhang JZH (2014) Effect of strong electric field on the conformational integrity of insulin. J Phys Chem A 118:8942–8952 Yang R, Li S-Q, Zhang QH (2004) Effects of pulsed electric fields on the activity and structure of pepsin. J Agric Food Chem 52:7400–7406 Yeom HW, Zhang QH, Dunne CP (1999) Inactivation of papain by pulsed electric fields in a continuous system. Food Chem 67:53–59 Yeom HW, Zhang QH, Chism GW (2002) Inactivation of pectin methyl esterase in orange juice by pulsed electric fields. J Food Sci 67:2154–2159 Zhao W, Yang R (2008a) Comparative study of inactivation and conformational change of lysozyme induced by pulsed electric fields and heat. Eur Food Res Technol 228:47–54 Zhao W, Yang R (2008b) The effect of pulsed electric fields on the inactivation and structure of lysozyme. Food Chem 110:334–343 Zhao W, Yang R (2010) Experimental study on conformational changes of lysozyme in solution induced by pulsed electric field and thermal stresses. J Phys Chem B 114:503–510 Zhao W, Yang R, Lu R, Tang Y, Zhang W (2007) Investigation of the mechanisms of pulsed electric fields on inactivation of enzyme: Lysozyme. J Agric Food Chem 55:9850–9858 Zhong K, Hu X, Zhao G, Chen F, Liao X (2005) Inactivation and conformational change of horseradish peroxidase induced by pulsed electric field. Food Chem 92:473–479 Zhong K, Wu J, Wang Z, Chen F, Liao X, Hu X, Zhang Z (2007) Inactivation kinetics and secondary structural change of PEF–treated POD and PPO. Food Chem 100:115–123

Impact of Pulsed Electric Field Treatment on Must and Wine Quality Lucía González-Arenzana, Javier Portu Reinares, Noelia López, Pilar Santamaría, Teresa Garde-Cerdán, Ana Rosa Gutiérrez, Isabel López-Alfaro, and Rosa López

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Improving Must and Wine Quality with PEF Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 PEF for Increasing the Microbial Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Improvement of Must and Wine Composition After PEF Treatments . . . . . . . . . . . . . . . . . . . . . . 10 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Abstract

Nowadays, consumers are demanding high-quality and healthy wines. This change in the trend of consumption could be considered a challenge to wine industry. For this reason, new technologies are being studied to better know their possible implementation in the oenological industry. One of these promising technologies is the pulsed electric field (PEF). This is a nonthermal technology in which high-intensity but short micropulses are applied L. González-Arenzana • J.P. Reinares • P. Santamaría • T. Garde-Cerdán • A.R. Gutiérrez • R. López (*) La Rioja Government, ICVV, Instituto de Ciencias de la Vid y del Vino (Gobierno de La Rioja, Universidad de La Rioja, CSIC), Logroño, La Rioja, Spain e-mail: [email protected]; [email protected]; [email protected]; teresa.garde@icvv. es; [email protected]; [email protected] N. López CNTA, Centro Nacional de Tecnología y Seguridad Alimentaria, San Adrián, Navarra, Spain e-mail: [email protected] I. López-Alfaro (*) La Rioja Government, ICVV, Instituto de Ciencias de la Vid y del Vino (Gobierno de La Rioja, Universidad de La Rioja, CSIC), Logroño, La Rioja, Spain Sección de Viticultura y Enología, ICVV. Finca La Grajera, Logroño, La Rioja, Spain e-mail: [email protected] # Springer International Publishing AG 2016 D. Miklavcic, Handbook of Electroporation, DOI 10.1007/978-3-319-26779-1_174-1

1

2

L. González-Arenzana et al.

to a liquid material placed between two electrodes. The energy causes the electroporation of cells what have two direct applications in oenology. The first technique is aimed to electroporate vegetal cells to enhance the extraction of interesting must and wine compounds and the other one is focused on electroporating the microorganism cells to render them inactivate. Both applications have demonstrated to be interesting and positive for general wine quality. PEF increases the grape volatile composition and reduces the volatile acidity of wines, furthermore interesting phenolic compounds related to color and healthy molecules as resveratrol are more extracted from grape skins after PEF treatments. Additionally, PEF inactivates the natural microbiota present in wines, being especially effective for bacteria as acetic acid and lactic acid bacteria populations. In spite of these positive applications of PEF for must and wine quality, further research is still needed to solve some possible problem of the scale-up of PEF from pilot plant to oenological industrial level. Keywords

PEF • Wine • Microbial inactivation • Extraction • Phenolic compounds

Introduction Wine has been defined in the International Code of Oenological practices (2015) as “the beverage resulting exclusively from the partial or complete alcoholic fermentation of fresh grapes, whether crushed or not, or of grape must. Its actual alcohol content shall not be less than 8.5 % v/v.” This determines that the vinification process always begins with grapes and consists of different oenological practices aimed at transforming them into wine. This broad definition establishes wines as potentially different products depending on several factors, according to the management tasks carried out in the vineyard and in the winery. Furthermore, wine can be considered a food associated to human life since ancient times, and a moderate intake of this fermentative product has been traditionally recognized as a healthy habit recommended by many specialists in the Mediterranean diet. Wine intake can be thought as a hedonistic activity, because it produces pleasure when consuming, being this pleasure directly linked to the perceived quality. Describing the wine quality is not an easy task because many factors are directly involved, such as climate, grapevine, grapes health, winemaking, etc. Currently, according to the world trend of controlling the characteristics, benefits, and drawbacks of food that we consume, consumers are demanding a more natural wine, produced via organic growing methods, with no added chemical products, etc., definitely healthier wines. This change in the trend of consumption has made the scientist community to deal with the development of new technologies that in a near future could be probably applied to food industries.

Impact of Pulsed Electric Field Treatment on Must and Wine Quality

3

In particular, the development of alternative, cutting-edge technologies to the traditional ones provides the oenological industry the challenge of improving quality through generating new products and optimizing processes, while production costs can be reduced. Traditionally, several strategies aimed at preventing spoilage of musts and wines as well as accurately monitoring the winemaking process have been applied. Among these, the employment of yeasts and lactic acid bacteria starter cultures or the sulfur dioxide (SO2) addition during the different stages of the vinification process has become bold oenological practices. The relevance of the SO2 products in the winery is not only its effect of rendering certain microorganisms inactive, such as yeasts, lactic acid bacteria, and to a lesser extent acetic acid bacteria, but its antioxidant properties are actually also significant. It is commonly added to must and wine in dissolution favoring its homogenization, and generally its addition inhibits the development of non-Saccharomyces populations and spoilage bacteria. Furthermore, it reduces the effect of dissolved oxygen in wine and even makes the oxidation enzyme activity decline. Although SO2 has been considered indispensable in wineries, it has been determined that adding it in an improper way can generate complications. For instance, in excess this chemical agent produces organoleptic deviation and unpleasant flavors because of the synthesis of mercaptans and hydrogen sulfide formation (Garde-Cerdán et al. 2008). Furthermore, it has been described as hazardous for human health particularly when it is inhaled, producing allergic reactions in sensitive people (Vally and Neil 2012). Consequently, the International Organization of Vine and Wine (OIV) has established limits for the SO2 content in wines, and the use of SO2 is expected to be dramatically reduced in the near future. For this reason, there is a particular interest within the scientific community in the development of alternatives to the long-established use of SO2 in winemaking. These alternatives are brought together in several research lines that are currently being conducted in order to search for an outstanding substitute for SO2. In this sense, products such as glutathione or ascorbic acid have been tested due to their antioxidant properties. Other natural substances such as thyme, oregano, grape phenolic extracts, and chitosan are really interesting, not only because they possess antioxidant properties but also for their antimicrobial character. Indeed, different antimicrobial additives, for instance, lysozyme, dimethyl dicarbonate, sorbate, or silver molecules, are currently being studied as possible substitutes for the antimicrobial effects of the SO2. There are other traditional practices such as fining or sterile filtration that, on their own, are not effective enough to substitute the use of SO2. Finally, in order to produce additive-free wines, innovative technologies capable of making microorganisms associated with wines inactive have been studied (Santos et al. 2012). The application of technologies based on heating was completely unsuitable for inactivating microbiota present during the transformation from must into wine because thermal processing for wines will clearly spoil wine quality. Therefore, studying nonthermal technologies was imperative in order to monitor the neutralization of the microbiota associated with wine.

4

L. González-Arenzana et al.

Pulsed electric field (PEF) treatment consists in the application of micropulses of high voltage and short duration to a liquid or solid product, staying or flowing between two electrodes located inside a treatment chamber. The electric field generated can electroporate the cells, increasing permeability in a permanent or temporary way. The mechanisms that produce this permeabilization are not perfectly understood but what it is widely known is that produces the pore formation in the cell membranes. These pores could be repaired in some cases but in other cases they are totally lethal for cell survival. PEF technology has two different applications to oenology depending on the type of cell that it was aimed to. PEF can electroporate microorganisms to inactivate them or can be focused on vegetal cells from grapes to increase the extraction. The possibility of PEF for inactivating cells could be directly involved with its application to microbial inactivation; but moreover, it is really interesting because it could improve the transfer from cells to outside which might favor the extraction of polyphenolic compounds and aromatic precursors usually located on grape skins. Static PEF treatments and continuous systems at pilot plant, applied to the grape skin before the maceration–fermentation stage, increase the extraction of total phenolic compounds during different grape variety vinifications. The phenolic compounds are in a significant way responsible for the organoleptic characteristics of wine, such as color and astringency. Some phenolic compounds offer antioxidant and free radical-scavenging properties which may play a role in human health, including protection against cardiovascular diseases and cancer. Antioxidant activity of red wines has been correlated with total anthocyanin and flavonol concentration. The content of total flavanols, among them catechin and epicatechin, is strongly related to this antioxidant capacity and free radical-scavenging activity found. The possibility of employing PEF in oenology is going to be described in the current chapter (Fig. 1) based on results close to the real practice of the wineries. The microbial inactivation registered after PEF treatments and the differences in the wine composition are going to be detailed to better know the advantages and disadvantages of PEF technology applied to wine industry.

Improving Must and Wine Quality with PEF Application PEF for Increasing the Microbial Stability The Winemaking Process It is well known that winemaking is one of the most ancient microbiological transformations, arising accidentally in society 7500 years ago. The fermentative character of winemaking is determined by the two different fermentative phases which make up the process: alcoholic fermentation initiated by yeasts, in which the sugar from grapes is transformed into wine, and malolactic fermentation performed by lactic acid bacteria, in which the wine acidity slowly decreases. The two fermentative stages could occur simultaneously or successively (Ribéreau-Gayon et al.

Impact of Pulsed Electric Field Treatment on Must and Wine Quality

5

Grapevines

Grapes Skins: phenolic content (Acetic acid bacteria, lactic acid bacteria, yeast and filamentous fungi)

PEF: increasing extraction Alcoholic fermentation Yeasts:Saccharomyces cerevisisae

PEF: microbial inactivation

Malolactic fermentation Lactic acid bacteria: Oenococcus oeni

Ageing (oak barrel, bottle, etc.) (Acetic acid bacteria, lactic acid bacteria and Yeasts)

Consumer

Fig. 1 Simplify scheme of the winemaking process and possible application of PEF during its development

2007a) and are characterized by the depletion of thousands of chemical, physical, and biological reactions and interactions what make winemaking a very complicate process. By the time the grapes reach the winery, a complex microbiological population has gradually developed (Fig. 1). Grapes are not a sterile substrate because they hold a relevant environmental microorganism population consisting mainly of yeasts, lactic acid bacteria, acetic acid bacteria, and filamentous fungi. Nonetheless, only yeasts and lactic acid bacteria are directly involved in the winemaking process because of the part they actively play in the vinification of grapes. In most cases, the hygienic conditions of modern wineries have been notably improved; even so, the winery environment is not an aseptic area. In practice, wineries are inhabited by hundreds of microorganisms belonging to different families, genera, and species. Therefore, the microbiota proceeding from the vineyard converge with the microorganisms which remain in the wineries from vintage to vintage, and thus, this would be considered the starting point of vinification. Grapes are received at the winery crammed with microorganisms but also with plenty of sugars as glucose and fructose. These sugars will represent the energy source for yeasts until alcoholic fermentation depletion. This fermentative stage is an ecological competition for the substrate, and, as a consequence, the best adapted yeast will be the only one that survives in such a difficult situation.

6

L. González-Arenzana et al.

Inactivation of Yeasts Associated to Wine by PEF The yeast communities involved in winemaking could be separated into two welldefined groups by the species composition. The first would be the non-Saccharomyces species group and the other the Saccharomyces species group. The population of non-Saccharomyces in sound grapes and musts is around 103–105 colony-forming units per milliliter (CFU/mL), which is greater than the population of Saccharomyces cerevisiae at that moment. In the initial stage of vinification, the group of non-Saccharomyces yeasts can actively intervene by generating numerous secondary metabolites that affect the final organoleptic characteristics of the wine. The most usually described non-Saccharomyces species isolated in grapes and musts belong to the Kloeckera, Candida, Brettanomyces, Cryptococcus, Kluyveromyces, Pichia, and Rhodotorula genera. Some non-Saccharomyces yeasts species such as Hanseniaspora uvarum, Metschnikowia pulcherrima, and Candida stellata are usually detected in rotten grapes, being dangerous producers of ethyl acetate and acetic acid during the initial stages of alcoholic fermentation. These species are usually present in environment with low ethanol so they trend to disappear when alcoholic fermentation begins. Other yeasts, such as Saccharomycodes ludwigii, Torulaspora delbrueckii, and species of the Zygosaccharomyces genus, are considered as spoilage yeasts because of their ability to produce sediment in bottles generating cloudiness. However, the ones currently considered as the “winemakers’ nightmare” are Dekkera bruxellensis and Brettanomyces bruxellensis. Dekkera bruxellensis could be able to generate spores under difficult conditions, and Brettanomyces bruxellensis is the asexual or non-sporulating form, so they are often confused in the literature. Both spoilage yeasts survive in the winery environment even in the tiny pores of the wooden barrels and metabolize the p-coumaric acid producing ethylphenol which could be completely disgusting for the wine consumers (Malfeito-Ferreira 2014). Clearly, the presence of Saccharomyces species is indispensable for the activity of winemaking. In effect, the S. cerevisiae species is totally preadapted for surviving in must due to its osmotolerance. Curiously, S. cerevisiae prefers anaerobic fermentation rather than respiration even when an aerobic atmosphere is available, because secreting great ethanol quantities to the media facilitates the exclusion of the majority of its potential competitors. This complete adaptation allows this yeast to perform alcoholic fermentation with populations reaching around 107–108 CFU/mL compelling the non-Saccharomyces species to almost disappear. The natural adaptation of S. cerevisiae to the spontaneous alcoholic fermentation of wines does not mean that few strains were participating in the process. In fact, the older the winery, the more S. cerevisiae strains participate in alcoholic fermentation. In contrast, new wineries lacked a mature yeast ecosystem, which was shown by the presence of not numerous but predominant S. cerevisiae strains. S. cerevisiae, although considered the first domesticated yeast, also has the ability of producing re-fermentation inside bottles of wines. The PEF treatment is considered to be one of the most promising emerging technologies for liquid food preservation. In order to know the effect of PEF over

Impact of Pulsed Electric Field Treatment on Must and Wine Quality

7

microorganisms present in a food, in this case in must or wine, flowing systems of treatments with real wines containing naturally different microorganisms have been performed (▶ Pulsed Electric Field Treatment for Beverage Production and Preservation). Cells that are inoculated in culture media can be stronger or weaker than in their natural environment what would produce a bias in the interpretation of the results. Quite different results have been published about microbial inactivation, but differences in the media or buffer, in the treatments of PEF applied and of course in the treatment chamber, and in the parameters of electricity applied have made the comparison of rendering microorganisms inactivate mainly impossible. Taking this into account, results of the studies closest to the reality of a winery, specifically for this issue, should be considered. In studies performed in conditions similar to a winery with inoculated red wines, the inactivation of yeasts after PEF was moderate, although in must samples the inactivation of yeasts as Saccharomyces cerevisiae was higher than in wine samples but fewer than 104 CFU/mL. As an example, in wine samples, Saccharomyces is one of the most resistant to PEF (with high energy treatment) maybe because of its adaptation to the hard environment of wine. In practice, yeast inactivation by PEF with a continuous flow of wine (12 L/h) was always less than three logarithmic cycles (103 CFU/mL), and the application of higher energy did not mean significantly higher yeast inactivation (González-Arenzana et al. 2015). Sometimes, the different electric field strengths and different specific energy applied caused no variation on yeast population inactivation. Consequently, the effectiveness of PEF over yeast is currently not enough to produce a safe wine in terms of microbial quality. Higher reductions were described for Dekkera species but in static chambers (Puértolas et al. 2009). Some recent results performed in wines after alcoholic fermentation and after malolactic fermentation could be indicating that after PEF treatments of wines, independently on the SO2 addition, there is an inactivation of the endogenous wine microbial population. In this sense, the indigenous yeast population after the application of a concrete PEF treatment (previously optimized to oenological conditions) were lower than those described for non-PEF-treated wines, with important yeast inactivations (>70 % of the initial population) in most tested wines.

Inactivation of Lactic Acid Bacteria Associated to Wine by PEF The harvested grapes, and the consequent must, have modest (around 103 CFU/mL) yet diverse lactic acid bacteria communities (González-Arenzana et al. 2016). These lactic acid bacteria populations from grapes converge with the lactic acid bacteria population remaining in the air and on winery equipment and facilities, exactly the same as was described for yeasts. The lactic acid bacteria population shows a poorer adaptation to wine than yeasts so that, as a consequence, the beginning of alcoholic fermentation is quite immediate while lactic acid bacteria exert a secondary role during alcoholic fermentation. When sugars from grapes downsize, yeasts decrease and lactic acid bacteria increase; so spontaneous malolactic fermentation usually follows alcoholic fermentation. Strictly, malolactic fermentation is just a

8

L. González-Arenzana et al.

decarboxylation in which the L-malic acid is converted into L-lactic acid by the malolactic enzyme of lactic acid bacteria. As a result of decarboxylation, the wine’s acidity decreases, transforming the environment into a more comfortable medium for lactic acid bacteria populations. Only four lactic acid bacteria genera, Lactobacillus, Pediococcus, Leuconostoc, and Oenococcus, have been traditionally described in must and grapes although recently it has been determined that actually up to six lactic acid bacteria genera can be determined. Several lactic acid bacteria species are sometimes present, but the main bacterium involved is the Oenococcus oeni species, as this is the one of the most suitable bacteria for the malolactic fermentation to take place. Nowadays, malolactic fermentation is considered an important stage in the vinification of some red and white wines carried out in the proper conditions, being beneficial for wine quality, improving their sensory characteristics of wines and ensuring microbiological stability. In this respect, some lactic acid bacteria species involved in winemaking could, in certain conditions, be responsible for wine spoilage. This is clearly illustrated by the fact that some Lactobacillus species, even Oenococcus oeni, are able to degrade the wine quality after malolactic fermentation by producing an excessive buttery character, as a consequence of the degradation of citric acid. Despite being the most important biological agent during malolactic fermentation, Oenococcus oeni is also involved in the presence of some biogenic amines such as histamine, ornithine, and putrescine, but it is not the only one linked to the spoilage of the quality of wine. Apart from that, species such as Lactobacillus hilgardii, Lactobacillus mali, Leuconostoc mesenteroides, and Pediococcus parvulus are related to the appearance of histamine in wines after malolactic fermentation. The same is true for Lactobacillus brevis, closely associated with the presence of tyramine in wines. These biogenic amines cause the organoleptic detriment of wine quality and are related to some medical disorders such as headaches, and the winery may also face legal restrictions on the export of their wines. Furthermore, wines with a mousey off-flavor are usually associated with the metabolism of ornithine and lysine by some heterofermentative Lactobacillus and some Leuconostoc species. Sometimes the wine takes on an odd bitterness caused by the acrolein that some Lactobacillus and Pediococcus generate. In particular, the Lactobacillus species are involved in the decrease in tartaric acid and in the increase in acetic acid. Moreover, spoilage of ropy wines, although not dangerous, is caused by the Pediococcus genus, and finally, O. oeni could be related to the synthesis and, consequently, to the increase of mannitol in spoilt wines. As a general rule in synthetic media, bacteria are supposed to have less PEF sensitivity than yeasts because of their smaller size (Fox et al. 2008; Timmermans et al. 2014). Nevertheless, significantly different inactivations have been described in the literature only for bacteria and particularly between the treatments that induced the highest and the lowest energy, reaching inactivation levels greater than four logarithmic cycles (104 CFU/mL) (González-Arenzana et al. 2015; Puértolas et al. 2009). Some studies established inactivation of lactic acid bacteria by PEF in static chambers of less than 3 log units (Puértolas et al. 2009) and in continuous flow, less than 2 log units (Buckow et al. 2010; Sharma et al. 2014).

Impact of Pulsed Electric Field Treatment on Must and Wine Quality

9

Studies carried out in wines after alcoholic fermentation and after malolactic fermentation could demonstrate that there is an inactivation of the natural wine lactic acid bacteria population after PEF treatments of red wines. In effect, the indigenous lactic acid bacteria population after the application of a previously optimized PEF treatment was notoriously downsized (>74 %) after alcoholic fermentation and malolactic fermentation.

Inactivation of Acetic Acid Bacteria Associated to Wine by PEF The acetic acid bacteria are a spoilage group of microorganisms regularly encountered in grapes and wines, as well as remaining in wineries. Nowadays, 12 genera compose the Acetobacteraceae family, but some of them are more frequently detected in wines, such as, for instance, Acetobacter, Gluconobacter, and Gluconacetobacter, which are able to oxidize the ethanol produced by yeasts into acetic acid and thereby increase volatile acidity. This wine spoilage can be described as the appearance of vinegar aromas and the increase of the solvent character. It occurred frequently during aging in oak barrels or in bottles, particularly in the presence of residual sugars and in aerobic conditions (Ribéreau-Gayon et al. 2006a). The big rod-shaped lactic acid bacteria are not necessarily more sensitive than small cocci or than acetic acid bacteria, although the smallest Gram-negative coccishaped cells are considered the most resistant to PEF application. The previously described inactivation of the Gluconobacter oxydans with PEF (Marsellés-Fontanet et al. 2009) was low. Curiously, some bacteria have been described as really resistant to PEF even when the applied energy was not accompanied by an increase of the temperature (González-Arenzana et al. 2015). Recently, the resistance of natural acetic acid bacteria population in wines is being observed in some research projects. In effect, applying PEF after alcoholic and malolactic fermentations renders a 54 % of the initial acetic acid bacteria population inactivated. Temperature and Energy Synergy The specific energy applied with PEF provided an inherent increase in temperature, independently on the refrigeration of the system. In most cases, scientists try that the increase of temperature inherent to electricity was not lethal for microorganisms, but an evident synergistic effect of energy and temperature was observed (Lebovka et al. 2005). The implementation of cooling systems before and after the treatment chamber is totally necessary to prevent that the high temperature act in detriment of the nutritional and organoleptic characteristics of wines. The collaboration between temperature and energy has made some acetic acid bacteria the most PEF resistant after one treatment but the most sensitive when the PEF treatment was accompanied by the highest temperature of the essay. So two factors of inactivation could be working together in rendering cells inactivated during PEF treatments. Thus, some authors testing other nonthermal techniques have determined that temperatures up to 45  C were not dangerous for cell integrity, but higher temperatures could cause changes in the membrane fluidity of microorganisms inducing their inactivation (Condón 2012).

10

L. González-Arenzana et al.

Inactivation of Different Strains to Winemaking Most of the responses of the cells are directly modulated by the strain factor. Even two cells belonging to the same species of microorganisms, if they are not genetically equal, will probably have different resistance to a stress condition. For the first time, significant differences were established between strains when the PEF treatment associated to a high specific energy and temperature was applied to two different strains of Oenococcus oeni species (González-Arenzana et al. 2015). These differences could be associated with the lack of some resistant genetic markers in the most sensitive strain. One of these genetic markers was involved in the cell wall biogenesis, which might mean that responses to PEF treatment totally depended on the strain.

Improvement of Must and Wine Composition After PEF Treatments Must and Wine Compounds and the General Changes After PEF The wine is a very complex matrix with plenty of different compounds forming the dissolution. For instance, organic acids, alcohols, esters, carbohydrates, minerals, and nitrogen compounds can be found in wines. Among them, the phenolic acid and the varietal aromas are really interesting for quality wine profile (Ribéreau-Gayon et al. 2007b). The PEF technology induces the permeabilization of cells what increases the interchange of materials from the inner part of the cell to the outer part. In this aspect, every one of the must and wine components interesting for the wine quality is susceptible of being more accessible after the electroporation of cells, even more if they are located into the vacuoles or when they are components of the cell wall or cell membrane (Yang et al. 2016). The most important compounds related to the wine quality that are usually tested in reference studies are described below (Hornsey 2007; Ribéreau-Gayon et al. 2007b). The basic molecules of the winemaking are the sugars that are usually named carbohydrates. These molecules are produced primarily by the photosynthesis carried out in the grapevine leaves and are formed by functional molecules that are involved in several metabolic reactions. For these reasons, they can act as precursors of many molecules as organics acids. The most abundant carbohydrates in the grape vacuoles are the D-glucose and the D-fructose usually in a ratio 1:1, respectively, with a concentration between 150 mg/L and 250 mg/L in ripe and healthy grapes. The carbohydrates from grapes are consumed by the yeast metabolism so the presence of these hexoses in a dry wine is usually minor than 1 g/L, being the fructose the most frequently detected in wines because the glucose is the one favorite for yeast during the alcoholic fermentation. Other sugars are also present in grapes and wines; for instance, D-galactose can be found in small concentrations (