Principles and Practices of Lyophilization in Product Development and Manufacturing 3031126335, 9783031126338

Table of contents :
Preface
Contents
Overview of Freeze Drying
1 Why Freeze Drying?
2 Equipment
2.1 Chamber
2.2 Condenser
2.3 Vacuum System
2.3.1 Chamber Pressure Control
2.4 Control System
3 Formulation
4 Process
4.1 Freezing
4.1.1 Physics of the Freezing and Crystallization Process (Adapted from Book Chapter in ``Development of Biopharmaceutical Dru...
4.1.2 Impact of Freezing Process on Protein Solutions and Modes of Denaturation
4.1.3 Freezing Rate
4.1.4 Annealing
4.1.5 Effect of Thermocouples on Freezing Rates
4.2 Primary Drying
4.3 Secondary Drying
4.4 Process Control
4.4.1 Chamber Pressure Control and Monitoring
4.4.2 Condenser Pressure Control and Monitoring
4.4.3 Determination of the End Point of Primary Drying
5 Stability
References
Characterization and Determination of Freeze-Drying Properties of Frozen Formulations: Case Studies
1 Introduction
2 Definition of Freeze-Drying Properties
2.1 Collapse Temperature
2.2 Eutectic Melting Temperature (Te)
3 Characterization Techniques
3.1 Differential Scanning Calorimetry (DSC)
3.1.1 Method Design Considerations
3.2 Freeze Drying Microscopy (FDM)
3.2.1 Equipment and Experimental Procedure
3.2.2 Applications of Freeze-Drying Microscopy
3.2.2.1 Crystalline Systems
3.2.2.2 Case Studies
3.2.2.3 Crystalline Systems
3.3 Estimation of Tg´
References
Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals
1 Introduction
2 Hammett Acidity Function and pHeq
3 Factors Which Impact Apparent Solid-State Acidity in Lyophiles
4 Correlations Between Solid-State Acidity and Stability of Freeze-Dried Formulations
5 Freezing Fundamentals: Quasi-Liquid Layer, Freeze-Concentration, Polarity, and the Workman-Reynolds Potential
6 Conclusions
References
Concepts and Strategies in the Design of Formulations for Freeze Drying
1 Introduction
2 Requirements and Expectations of a Lyophilized Product
3 Pre-formulation
3.1 Forced Degradation/Stress Studies
3.2 Chemical Instability
3.3 Physical Instability
3.3.1 Colloidal Instability
3.3.2 Conformational Instability
4 Formulation Development
4.1 Rational Design of the Formulation
4.2 Strategies to Mitigate Liabilities and Protein Aggregation
5 Stability Testing
5.1 Kinetics of Degradation in the Amorphous Solid State
5.2 Selection of Accelerated Testing Conditions, Temperature, and Time
References
Formulation Design for Freeze-Drying: Case Studies of Stabilization of Proteins
1 Excipient Selection
1.1 Role of Excipients
1.1.1 Protein Stabilization
1.1.2 Manufacturability
1.1.3 Tolerability
1.1.4 Drug Administration
1.2 Examples of Commercial Lyophilized Drug Products
2 In-Process and Storage Stability of Proteins: Stabilization During Freezing, Drying, and Storage
2.1 Theoretical Considerations
2.2 Practical Considerations
2.3 Excipient Properties and their Relevance for the Lyophilization Process
2.3.1 Analysis of Excipient Properties
2.3.2 The Link Between Excipient Properties and Lyophilization Process Parameters
2.3.2.1 Freezing
2.3.2.2 Annealing Prior to Primary Drying
2.3.2.3 Primary Drying
2.3.2.4 Secondary Drying
2.4 General Formulation ``Rules´´ for Freeze Drying
2.5 Solid State Dynamics
3 Product Quality Attributes
3.1 Residual Moisture
3.1.1 Impact of Water on Protein Stability
3.1.2 Water Content in Relation to Process Performance
3.2 Specific Surface Area
3.3 Cake Appearance
3.4 Reconstitution Time
4 Conclusions
References
Challenges and Considerations in the Development of a High Protein Concentration Lyophilized Drug Product
1 Introduction
2 Characteristics of High Protein Concentration Lyophilized Drug Product
2.1 Physical Appearance and Cake Structure
2.2 Prolonged Reconstitution Time
2.3 High Viscosity
2.4 Longer Primary Drying Time
2.5 Osmolality
3 Stability Considerations for the Development of High Protein Concentration Lyophilized Drug Products
3.1 Protein Cold Denaturation
3.2 Freezing Stress
3.2.1 Freeze-Concentration
3.2.2 Ice-Water Interface
3.2.3 pH Change
3.3 Phase Separation
3.4 Dehydration Stress
4 Stability Strategy to Minimize Protein Degradation During Lyophilization Process and Shelf Life Storage
4.1 Buffer Species, Surfactants, and Viscosity Reducer Selection
4.2 Stabilizers Selection
4.3 Stability Strategy for a High Concentration Lyophilized Protein Drug Product Stable for Room Temperature Storage
4.4 Coformulation of High Concentration Lyophilized Protein Drug Product
5 Formulation and Process Strategy to Minimize the Reconstitution Time
5.1 Increase Cake Wettability and Crystallinity
5.2 Controlled Nucleation
6 Lyophilization Cycle Development for High Protein Concentration Drug Product
6.1 Freezing
6.2 Primary Drying
6.3 Secondary Drying
6.4 Lyophilization Cycle Scale-Up Considerations
7 Summary
References
Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water Formulations
1 Introduction
2 Preparation of Solution
3 Freezing Parameters
3.1 Nucleation Temperature
3.2 Solvent Crystals Morphology
4 Thermodynamical Properties
4.1 Phase Diagram
4.2 Equilibrium Vapour Pressures
4.3 Sublimation Enthalpies
5 Sublimation Kinetics
6 Residual TBA and Water Contents
7 Annealing Treatment
8 Design Space for Freeze-Drying of an Ibuprofen Formulation
8.1 Overall Heat Transfer Characterization
8.2 Mass Transfer Resistance of the Dried Layer
8.3 Determination of Sublimation Rates of Real Ibuprofen Formulation
8.4 Mean Product Temperature
8.5 Limits of Critical Quality Attributes
8.6 Reconstitution Properties
9 Conclusions
References
Primary Container Closure System Selection for Lyophilized Drug Products
1 Introduction
2 Primary Containers
2.1 Overall Considerations
2.2 Heat Transfer (Glass Vials)
2.3 Glass Versus Polymer
2.3.1 Heat Transfer
2.3.2 Breakage
2.3.2.1 Lyophilization
2.3.2.2 Storage at Low Temperatures
2.3.3 Delamination and Chemical Durability
2.3.4 Other Important Practical Consideration
2.4 Extractables and Leachables
2.5 Fogging
2.6 Coated Vials
3 Elastomeric Closures
3.1 Closure Geometry
3.1.1 Flange Thickness and Flange Design
3.1.2 Penetration Thickness
3.1.3 Plug Design
3.1.4 Standardization of Lyophilization Stoppers Geometry
3.2 Moisture Absorption/Desorption and Permeability Behavior
3.3 Fluoropolymer Coatings
4 Container Closure Fit and Integrity
4.1 Strategies and Methods for Assessing Component Fit
4.1.1 Visualization Techniques
4.1.2 Interference Fit Calculation: Assessment of the Plug Seal
4.1.3 Dimensional Stack-Up Analysis: Assessment of Flange Seal and Skirt Length
4.1.4 Container Closure Integrity and Seal Quality Test Methods
4.1.4.1 Helium Leakage
4.1.4.2 Laser-Based Headspace Analysis
4.1.4.3 Residual Seal Force
4.2 Common Causes of Loss of Vacuum and/or CCI: Raised Stoppers and Dried Product
5 Future Trends
5.1 Dual-Chamber Containers
5.2 Nested Vial Configurations and Press-Fit Caps
6 Conclusion
References
Vial Breakage During Lyophilization
1 Introduction
2 Use of Strain Gauges to Understand Vial Breakage
3 Parameters That Affect Vial Breakage
3.1 Vial Handling, Washing, Depyrogenation, and Other Handling in a Production Setting
3.2 Product Fill Volume
3.2.1 Interior Vial Coating
3.3 The ``Break-Free´´ or ``Plugging-Off´´ Event During Freezing
3.4 Warming of the Frozen Plug
3.5 Crystallizing Solutes
4 Conclusions and Recommendations
References
The Nucleation of Ice
1 Introduction
2 Definitions and the Formation of Ice
3 Ice Nucleation in Conventional Lyophilizers
3.1 The Effect of Cooling Rate on the Degree of Supercooling
3.2 The Effect of Vial Size and Fill Volume on the Degree of Supercooling
4 Controlled Ice Nucleation
4.1 Controlled Ice Nucleation Using a Reduced Pressure Ice Fog
4.2 Controlled Nucleation Using Rapid Depressurization
4.3 The Effect of Controlled Ice Nucleation on Process Time and Product Appearance
5 Summary
References
Stresses, Stabilization, and Recent Insights in Freezing of Biologics
1 Stresses During Freezing
1.1 Current Understanding of Freezing-Induced Stresses
1.2 Recent Insights in Freezing-Induced Stresses
1.2.1 Dissolved Gas
1.2.2 Mechanical Stress
2 Novel Modalities
2.1 Vaccines
2.2 DNA/RNA
2.3 Gene Therapy
2.4 Cells
3 Advances in Modeling Approaches
3.1 Practical Consideration for Frozen Products
4 Concluding Remarks
References
Lyophilization Process Understanding and Scaleup Using Ab Initio Vial Heat Transfer Modeling
1 Introduction
2 Materials and Methods
2.1 Theoretical Analysis of Heat Transfer to Product in Vials During Lyophilization
2.1.1 Definition of the Heat Transfer Coefficient, Kv
2.1.2 Phenomenological Parameters
2.1.3 Ab Initio Parameters and Their Dependence on the Equipment/Product/Process
2.2 Experimental Setup for Full-Shelf Gravimetric Measurements of Kv for Various Vials and Lyophilizers
2.3 Determining Ab Initio Kv Parameters for Various Vials and Lyophilizer Combinations
2.3.1 Shelf Heat Transfer Coefficient, Ks
2.3.2 Emissivity of the Shelf
2.3.3 Other Ab Initio Parameters - Kcs, kcs, Acs, es, ev, and lbot
2.4 Lyophilization Cycle Scaleup Using Ab Initio Vial Heat Transfer Modeling
2.4.1 Definition of a ``Design Condition/Case/Criteria (DC)´´
2.4.2 Placement Test of Data Sufficiency for Heat Transfer Analysis of a ``Target Design Condition´´
3 Results and Discussions
3.1 Experimental Results of Center Vial Kv for Different Lyophilizers and Vial Size
3.2 Example Case Studies to Scaleup Kv from a ``Known´´ to a ``Target´´ Design Condition (Lab-->Pilot) Using Ab Initio Paramet...
3.3 Discussion on the Uncertainty in the Estimated Kv
3.3.1 The Effect of Ks
3.3.2 Compilation of Literature Ab Initio Parameters for Kv
4 Conclusions
References
Secondary Drying: Challenges and Considerations
1 Introduction
2 Modeling of Secondary Drying
2.1 Desorption Kinetics Model
2.2 Heat Transfer Model
2.2.1 Empirical/Simplified Models
2.2.2 Theoretical Models
2.2.2.1 High Fidelity Model: Full 3D Simulation
2.2.2.2 Intermediate Fidelity Model: 1D Averaged Equations
2.2.2.3 Low Fidelity Model: 0D Lumped Capacitance Model
3 Characterization of Secondary Drying Process and the Lyophilized Cake
3.1 Temperature Measurement
3.2 Heat Flux Measurement
3.3 Moisture Content
3.3.1 Pressure Rise Test (PRT)
3.3.2 Tunable Diode Laser Absorption Spectroscopy
3.4 Properties of Lyophilized Cake
3.4.1 Structure of the Lyophilized Cake
3.4.2 Glass Transition Temperature and Collapse Temperature
4 Critical Process Variables During Secondary Drying
4.1 General Operational Conditions
4.2 Effect of Temperature
4.3 Effect of the Vial
4.4 Effect of Specific Surface Area
4.5 Effect of Excipients
4.6 Process Parameters Not Affecting Secondary Drying
5 Challenges in Secondary Drying
5.1 Moisture vs. Stability
5.2 Uncertainties in Heat Transfer Coefficient
5.3 Inefficient Heat Transfer
5.4 Defects in Dried Cake
5.5 Scaleup
5.6 Temperature and Moisture Uniformity
6 Concluding Remarks
References
Design and Process Considerations in Spray Freeze Drying
1 Introduction
1.1 Challenges for Traditional Freeze Drying
1.2 Desirable Attributes for an Improved Process
1.3 Spray Freeze Drying Developments
1.4 General Product and Process Considerations Related to Lyophilized Bulkware
2 Spray Freeze Drying
2.1 General Design Features and Process Characteristics
2.1.1 Spray Freezing
2.1.2 Rotary Freeze Drying
2.1.3 Equipment Configuration Options
2.2 Spray Freezing: Specific Design and Process Considerations in Spray Freezing
2.2.1 Introductory Remarks
2.2.2 Spray Freezing Equipment Design, Process, and Product Parameters
2.2.2.1 Tower
2.2.2.2 Spraying
2.2.3 Scale-up in Spray Freezing
2.3 Dynamic Rotary Bulk Freeze Drying
2.3.1 General Equipment Design, Process, and Product Parameters
2.3.2 Specific Aspects of Rotary Freeze Drying: Similarities and Differences When Comparing to Shelf Freeze Drying
2.3.2.1 Tg´
2.3.2.2 Pressure
2.3.2.3 Shelf and Drum Surface Temperature
2.3.2.4 IR-Power in Rotary Freeze Drying
2.3.2.5 Drum Rotation Speed in Rotary Freeze Drying
2.3.2.6 End Point Detection (By Comparative Pressure Measurement or Pressure Rise)
2.3.2.7 Formulation Aspects
2.3.2.8 Flow Properties of Spray Frozen Bulk in Dynamic Freeze Drying
Presence of Ice
Electrostatic Phenomena in Dynamic Bulk Freeze Drying
Formulation Aspects Relevant for Electrostatics
2.3.3 Annealing
2.3.3.1 Targets for Annealing
2.3.3.2 Annealing Procedure
2.3.3.3 Relevance of Annealing in Pharmaceutical Applications
2.3.3.4 Relevance of Annealing in Dynamic Spray Freeze Drying
2.3.3.5 Annealing Procedure in Dynamic Spray Freeze Drying
2.3.3.6 Further Annealing Effects in Dynamic Bulk Freeze Drying
2.3.4 Scale-up in Rotary Freeze Drying
2.3.4.1 Comparison of Conventional and Rotary Freeze Drying Scale-up
2.3.4.2 Changing Product Load in the Same Equipment
2.3.4.3 Changing Product Load Utilizing Different Equipment Sizes
3 Continuous Processing in Spray Freeze Drying: Aspects to Consider
4 Industrial Application of Spray Freeze Drying: Integration of Steps to an Aseptic Process Line
5 Summary
References
LyoPRONTO: Deterministic and Probabilistic Modeling - Tutorial and Case Study
1 Introduction
2 Numerical Modeling
2.1 Freezing Calculator
2.2 Primary Drying Calculator
2.3 Design Space Generator
2.4 Optimizer
3 Experimental Methodology
4 Case Study and Tutorial
4.1 Step-by-Step Tutorial
4.1.1 Freezing Calculator
4.1.2 Primary Drying Calculator: Deterministic Approach
4.1.3 Design Space Generator
4.1.4 Optimizer
4.2 Monte-Carlo-Based Probabilistic Approach and Comparison with Deterministic Methodology
4.2.1 Monte-Carlo Simulation Methodology and Input Parameters´ Variation
4.2.2 Probabilistic Approach: Case Study
5 Experimental Validation
5.1 Primary Drying Calculator: Probabilistic Approach
6 Conclusion
References
Utilizing Solid-State NMR Spectroscopy to Assess Properties of Lyophilized Formulations
1 Introduction
2 Analytical Challenges for Lyophilized Formulations
3 Basics of Solid-State NMR Spectroscopy
3.1 Relaxation
4 Characterization of Lyophilized Formulations by SSNMR
4.1 Identification of Formulation Components
4.2 Ionization Changes
4.3 Homogeneity of Formulation Components
4.4 Case Study: Phase Separation in Bovine Serum Albumin (BSA) and Lysozyme-Containing Formulations
4.5 Mobility
5 Comparison with Other Analytical Techniques
6 Conclusion
References
Design of Moisture Specification Studies for Lyophilized Product
1 Introduction
2 Stopper Moisture Studies
2.1 Stoppers Equilibrium Moisture Studies
2.2 Sterilization and Drying Studies
3 Lyophilized Product Moisture Studies
3.1 Moisture Equilibration
3.2 Sample Preparation
3.3 Moisture Analysis by Karl Fischer
4 Specifications
References
Laser-Based Headspace Moisture Analysis for Rapid Nondestructive Moisture Determination of Lyophilized Products
1 Introduction
1.1 Importance of Residual Moisture Content Determination
1.2 Traditional Methods to Determine Residual Moisture
1.3 The Need for a Better Residual Moisture Determination Method
2 Laser-Based Headspace Moisture Analysis
3 Case Study: Correlation of Headspace Moisture to Total Moisture Content
4 Case Study: Headspace Moisture as a Tool for Freeze-Drying Cycle Optimization
4.1 Conclusions
5 Case Study: Headspace Moisture as a Tool for Lyo Shelf Moisture Mapping
5.1 Conclusions
6 Case Study: Lyo Shelf Moisture Mapping to Demonstrate Freeze-Dryer Equivalence
6.1 Conclusions
7 Headspace Moisture Determination as a Water Activity Measurement
7.1 Thermodynamic Activity of Water
7.2 Moisture Sorption Isotherms
7.3 Product Decay Rate
8 Case Study: Correlation of Product Stability to Water Vapor Partial Pressure
8.1 Pharmaceutical Formulation
8.2 Sample Preparation and Storage
8.3 Results and Discussion
8.4 Conclusions
9 Chapter Summary
References
Application of PAT in Real-Time Monitoring and Controlling of Lyophilization Process
1 Introduction
2 PAT for Freeze-Drying Process Monitoring and Control
2.1 Dependent Variables/Critical Process Parameters of Freeze-Drying
2.2 Product Temperature
2.3 Product Resistance
2.4 Sublimation Rate
2.5 Nucleation Temperature
2.6 End Point Determination of Primary Drying Phase
2.7 End Point Determination of Secondary Drying Phase
3 PAT for Freeze-Drying Process Monitoring and Control
4 Single Vial Methods
4.1 Thermocouples and RTDs
4.2 TEMPRIS
4.3 Near Infrared Spectroscopy (NIR)
4.4 Microbalance Technique
5 Batch PAT Methods
5.1 Pressure Measurements: Capacitance Manometer and Pirani Gauge
5.2 Dew-Point Monitor
5.3 Manometric Temperature Measurement (MTM)
5.4 Thermodynamic Lyophilization Control (TLC)
5.4.1 Gas Plasma Spectroscopy (Lyotrack)
5.4.2 Residual Gas Analyzer Mass Spectrometer (LYOPLUS)
5.5 Tunable Diode Laser Absorption Spectroscopy
5.6 TDLAS Measurements of Vapor Mass Flow
5.7 Instrument Requirements
5.8 Sensor Validation
5.9 Sensor Applications
5.10 Lyophilizer OQ
5.11 Determination of Primary and Secondary Drying Endpoints
5.12 Determination of Vial Heat Transfer Coefficients and Product Temperature
5.13 Determination of Product Resistance
6 TDLAS Summary
References
Process Analytical Technology (PAT) for Lyophilization Process Monitoring and End Point Detection
1 Introduction
2 Product Temperature Measurement
2.1 Resistance Temperature Detector
2.2 Thermocouple
2.3 Wireless Temperature Sensor
2.4 Manometric Temperature Measurement (MTM)
3 Pressure Measurement
3.1 Pirani Gauge
3.2 Capacitance Manometer
4 Heat and Mass Transfer during Primary Drying and Secondary Drying
4.1 Primary Drying Heat Transfer and Vial Heat Transfer Coefficient
4.2 Mass Transfer Determination
4.2.1 Mass Loss by Gravimetric Measurement
4.2.2 Manometric Temperature Measurement (MTM)
4.2.3 TDLAS
5 Cycle Endpoint Detection
5.1 Limitations of Product Temperature Measurement
5.2 Alternatives to Product Temperature Measurement
5.2.1 Comparative Pressure Measurement (Pirani vs CM)
5.2.2 Electronic Hygrometer
5.2.3 Residual Gas Analysis
5.2.4 Pressure Rise Test (PRT)
5.2.5 Manometric Temperature Measurement (MTM)
5.2.6 TDLAS
6 Summary and Future Trend
Bibliography
Advances in Process Analytical Technology: A Small-Scale Freeze-Dryer for Process Analysis, Optimization, and Transfer
1 Introduction
2 Development of the MicroFD
3 Lyophilization Design Space
4 Vial Thermal Conductivity (Kv)
5 Product Cake Resistance (Rp)
6 PAT Technologies
6.1 AccuFlux Heat Flux Sensor
7 AutoDry Product Temperature Control
8 End of Primary Drying: Pirani Versus Capacitance Manometer Pressure Differential
9 FreezeBooster Controlled Nucleation
9.1 Freezing Process PAT
10 Nucleation (Intra-vial)
11 Nucleation (Inter-vial)
11.1 Freezing of Freeze Concentrate
11.2 Measurement and Control of Freezing Heat Flow Post-controlled Nucleation
12 Freezing Events
12.1 The Effect of Freezing and Heat Flow Control on Primary Drying
12.2 Summary of Experiments on the Different Freezing Methods on Primary Drying Times
13 Developing Transferrable Protocols
14 Summary
References
Overview of Heat and Mass Transfer Modeling in Lyophilization to Create Design Spaces and Improve Process Analytical Technolog...
1 Introduction
2 Heat and Mass Transfer Modeling for Process Development
2.1 Vial Heat and Mass Transfer Studied by Quasi-Steady One-Dimensional Modeling
2.2 Equipment Capability of Lyophilizers Studied by Quasi-Steady 3-D CFD Modeling of Choked Flow Conditions
2.3 Creation of the Design Spaces of Various Types
3 Heat and Mass Transfer Modeling for Process Monitoring and Optimization
3.1 Experimental Data of Chamber and Condenser Pressures, Pch and Pcd
3.2 Sublimation Flow Rate Studied by Quasi-Steady 3D CFD Modeling of Subsonic Non-choked Flow Conditions
3.3 Model-Based Process Monitoring Using the Chamber to Condenser Pressures Drops
4 Summary
References
Application of QbD Elements in the Development and Manufacturing of a Lyophilized Product
1 Introduction
2 Target Product Profile (TPP)
3 Formulation Development and Selection
4 Use of Prior Knowledge
4.1 Justification for Lyophilized Dosage Form: Why Lyophilization?
4.2 Preformulation Studies/Preliminary Work/Formulation Screening
5 Lyophilization Process Development
5.1 Justification of Commercial Manufacturing Lyophilization Cycle
6 Risk Assessment
6.1 Risk Assessment (Formulation)
6.2 Risk Assessment (Lyophilization)
7 Design of Experiments (DOE)
7.1 Combined Approach
7.1.1 Initial Screening Studies
7.1.2 Optimization Phase
7.1.3 Data Analysis to Build Response Surface Model (RSM) for Each Response
7.1.4 Identify Design Space Boundaries
7.2 Individual Approach: Formulation
7.2.1 pH Characterization Studies
7.2.2 Protein Concentration Characterization Studies
7.2.3 Characterization Mannitol and Sucrose
7.2.4 Effect of Critical Formulation Parameters on the Freeze-Drying Properties of the Formulation
7.2.5 Characterization of Polysorbate 20 Concentration
7.3 Individual Approach: Lyophilization
7.3.1 Robustness Studies and Construction of a Design Space
7.3.2 Design of Experiments (DOE)
7.3.2.1 Prior Knowledge on Freezing Ramp Rates
7.3.2.2 Annealing Temperature
7.3.2.3 Primary Drying
References
Characterization of Freeze Dryers
1 Introduction
2 Critical Characteristics of Freeze Dryer
2.1 Shelf Temperature Mapping and Ramp Rates
2.2 Pressure Control
2.3 Condenser Capacity
2.4 Vacuum Integrity Testing
2.4.1 Measurement Techniques
2.5 Characterization with Thermal Load
2.5.1 Upfront Manufacturability Assessment
2.5.2 Optimization of Process Conditions
2.5.2.1 Sublimation Tests
2.5.2.2 Experimental Methodology (Vial Kv Measurement Methodology)
Vials Loading and Product Probe Placement
2.5.2.3 Measurement of Heat Transfer Coefficient
2.5.2.4 Actual Shelf Surface Temperature
2.5.2.5 Minimum Controllable Chamber Pressure as a Function of Sublimation Rate (Experimental)
Methodology
Calculation of Minimum Controllable Pressure as a Function of Sublimation Rate (Pmin Test)
Estimation of Maximum Sublimation Rate
2.5.2.6 Minimum Controllable Chamber Pressure as a Function of Sublimation Rate (Computational Modelling)
2.5.2.7 Manufacturing Environment (Nucleation Temperature)
3 Engineering Run
References
Principles and Practice of Lyophilization Process and Product Development: Scale-Up and Technology Transfer
1 Introduction
2 Lyophilization Process Overview
3 Considerations for Scale-Up and Technology Transfer
3.1 Freezing (Ice-Nucleation) Differences
3.2 Heat and Mass Transfer Differences Due to Equipment Design Differences
3.2.1 Mass Transfer Differences
3.2.2 Heat Transfer Differences
3.3 Measurement Differences
4 Considerations for Successful Technology Transfer from Lab-Scale to Pilot-Scale to Commercial
4.1 Process Design Considerations for Scale-Up and Technology Transfer
4.1.1 Equipment Factors Affecting Scale-Up Process
4.1.1.1 Condenser Design Differences
4.1.1.2 Temperature Uniformity of Condenser Coil
4.1.1.3 Cooling Rate Uniformity
4.1.1.4 Shelf Temperature Uniformity
4.1.2 Equipment Capability Testing for Scale-Up Transfer
4.1.2.1 Minimum Controllable Chamber Pressure Test
4.1.2.2 Maximum Sublimation Rate Test
4.2 PPQ Strategies with Process Verification and Sampling Plan
4.2.1 Assessment for Homogeneity
4.3 Continuous Process Verification
4.4 Drug Product Quality Attributes (for Lyo Product)
5 Process Analytical Technology (PAT) and Next-Generation Advancements
6 Summary
References
Lyophilization Validation: Process Design and Modeling
1 Introduction
2 Lyophilization Process Validation
2.1 Stage 1: Process Design
2.2 Stage 2: Process Qualification
2.3 Stage 3: Continued Process Verification
3 Stage 1: Process Design
3.1 Generation and Use of Design Space
3.1.1 Introduction to the Driving Forces and Resistances During Primary Drying
3.1.2 Equations for the First Principles of Heat and Mass Transfer
3.1.3 Determination of Primary Drying Conditions and Construction of Design Space
3.2 Engineering/Development Runs at Commercial Scale
4 Power of Simple Modeling for Process Optimization and Scale-Up
4.1 Development and Optimization of a Lyophilization Process
4.2 Scale-Up and Transfer
4.2.1 Class 100/Particle-Free Environment (Tn)
4.2.2 Equipment Capability (Minimum Controllable Chamber Pressure as a Function of Sublimation Rate)
4.2.3 Heat Transfer Coefficient (Kv)
5 Case Studies
5.1 Construction of Design Space
5.2 Effect of Batch Sizes (Product Load), Fill Volume, and Dose Strength
5.3 Controlled Ice Nucleation Technology (CIN) and Its Effect on Product Resistance
6 Summary
References
Lyophilization Validation: Process Qualification and Continued Process Verification
1 Introduction
2 Lyophilization Process Validation
2.1 Stage 1: Process Design
2.2 Stage 2: Process Qualification
2.2.1 Qualification of the Lyophilization Equipment
2.2.2 Lyophilization Process Performance Qualification
2.3 Stage 3: Continued Process Verification
3 Current Practices in Lyophilization Process Validation
4 Stage 2: Process Performance Qualification
4.1 PPQ Protocol
4.1.1 Batch Size
4.1.2 Number of PPQ Lots
4.1.3 Potential CPPs to Be Monitored
4.1.4 Potential CQAs to Be Tested
5 Stage 3: Continued Process Verification
5.1 Use of Run or Control Chart
5.2 Selection of CQAs and CPPs to Use in the Plot
5.2.1 Lyophilized Products in Vials
5.3 Options for Plotting Variations in Control Chart
5.3.1 Sample
5.4 Options for Plotting Data in Run Chart
5.4.1 Sample
6 Special Cases of Lyophilization Process Validation
6.1 Validation Approaches to Freeze-Drying of Pharmaceuticals in Alternative Containers
6.1.1 Emerging and Existing Container Closure Systems
6.1.1.1 Dual Chamber Vials
6.1.1.2 Dual Chamber Syringes and Cartridges
6.1.1.3 Trays
6.1.2 Specifics of Heat and Mass Transfer in Dual Chamber Devices
6.1.3 Specifics of Heat and Mass Transfer in Tray Drying
6.1.4 Validation Approaches for Dual Chamber Devices and Tray Drying
7 Case Studies
7.1 Monoclonal Antibody Case: Fill Volumes and Equipment Validation Strategy
7.2 Impact of Lyophilization Chamber Loading Process on Product Appearance and Product Rejection Rates
7.3 Lyophilization Cycle for a Product in a Dual-Chamber Cartridge
8 Summary and Future Outlook
References
Homogeneity Assessment of Lyophilized Biological Drug Products During Process Performance Qualification
1 Introduction
2 Methodology
2.1 Experimental Design
2.2 Statistical Model
2.3 Equivalence Testing and the Homogeneity Acceptance Criterion (HAC)
3 Results and Discussion
3.1 Protocol Phase
3.2 Protocol Execution Phase
3.3 Report Phase
4 Summary
References
Informed Manufacturing Through the Use of Big Data Analytics for Freeze Drying Process and Equipment
1 Introduction
1.1 Continued Process Verification and Freeze Dryer System Monitoring
2 Components of a Lyophilizer
2.1 Chamber and Condenser
2.2 Refrigeration
2.3 Heat Transfer
2.4 Vacuum
2.5 Hydraulics
3 Maintenance Practices and Data Analytics
3.1 Preventive Maintenance Techniques and Challenges in Preventive Maintenance
3.2 Infrastructure and Need for Data Analytics
3.3 The Role of Process Monitoring and Process Control
3.4 Sensor Network to Support Data-Driven Manufacturing
4 Case Studies for Data Analytics on Production Freeze Dryers
4.1 Condenser Temperature Excursions
4.2 Vacuum Pump Performance
4.3 Equipment Utilization
4.4 Leak Rate Tests
4.5 Refrigeration System Performance
4.6 Pressure Fluctuations
4.7 Stage Identification
5 Conclusions
References
Multivariate Analysis for Process Understanding, Continuous Process Verification, and Condition Monitoring of Lyophilization P...
1 Introduction
2 Data Gathering, Storing, and Analysis Infrastructure
3 Multivariate Data Analysis Theory
3.1 Univariate vs. Multivariate Analysis
3.2 Principal Component Analysis
3.2.1 Latent Variables, Loadings, and Scores
3.2.2 PCA Model Tuning and Performance Indexes
3.2.3 Algorithms for Building PCA Models
3.3 Principal Component Regression (PCR)
3.4 Projection to Latent Structure Regression (PLSR)
3.4.1 Algorithms for Building PLS Models
3.5 Data Unfolding for Applying PCA/PLS on Process Trajectory Data
3.5.1 Batchwise Unfolding
3.5.2 Variable-Wise Unfolding
4 Continuous Process Verification
4.1 Example Using PLS for Lyophilization and Applying on CPV
5 Condition Monitoring of Lyophilization
6 Future Directions of MVDA
6.1 Deep Learning Techniques for Lyophilization Condition Monitoring
References
Lyophilized Drug Product Cake Appearance: What Is Acceptable?
1 Introduction
2 Current Status of Lyophilized Product Cake Appearance: General and Regulatory Expectations
3 Terminologies Used to Define Variations in Cake Appearance from Ideal Expectations of ``Uniform and Elegant´´
3.1 Nonconformity or Defect
3.2 Irregularities/Nonuniform Cake Appearance
3.3 Visual Attributes of Freeze-Dried Products
3.3.1 Collapsed Cake
3.3.2 Meltback
3.3.3 Product Ejection
3.3.4 Puffing
3.3.5 Lifted Cake
3.3.6 Cake Shrinkage and Cracked Cake
3.3.7 Dusting, Chipping, and Broken Cake
3.3.7.1 Fogging
3.3.8 Lyo Ring, Minor Splashing, and Major Splashing
3.3.9 Bubble or Foam Formation
3.3.10 Volcano
3.3.11 Cake Texture
3.3.12 Droplets and Product on Inside Walls of the Vial
4 Potential Clinical Relevance of Cake Appearance
4.1 What Cake Appearances Are Acceptable?
4.2 Additional Considerations for Accepting Variations in Cake Appearance - Example: Collapsed Cake
4.3 Summary
References
Index

Citation preview

AAPS  Advances in the Pharmaceutical Sciences Series  59

Feroz Jameel  Editor

Principles and Practices of Lyophilization in Product Development and Manufacturing

AAPS Advances in the Pharmaceutical Sciences Series

Series Editor Yvonne Perrie Strathclyde Institute of Pharmacy and Biomedical Sciences University of Strathclyde, Glasgow, UK

The AAPS Advances in the Pharmaceutical Sciences Series, published in partnership with the American Association of Pharmaceutical Scientists, is designed to deliver volumes authored by opinion leaders and authorities from around the globe, addressing innovations in drug research and development, and best practice for scientists and industry professionals in the pharma and biotech industries. Indexed in Reaxys SCOPUS Chemical Abstracts Service (CAS) SCImago EMBASE

Feroz Jameel Editor

Principles and Practices of Lyophilization in Product Development and Manufacturing

Editor Feroz Jameel Nimble BioSolutions Gurnee, IL, USA

ISSN 2210-7371 ISSN 2210-738X (electronic) AAPS Advances in the Pharmaceutical Sciences Series ISBN 978-3-031-12633-8 ISBN 978-3-031-12634-5 (eBook) https://doi.org/10.1007/978-3-031-12634-5 # America Association of Pharmaceutical Scientists 2023 This work is subject to copyright. All rights are reserved by the Publishers, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publishers, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publishers nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publishers remain neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

In memory of Mike J. Pikal Our team always looked forward to Mike joining our group during the summer. The time was affectionately known as his “Summer Camp” and was the perfect time for technical discussions as well as boat rides on Lake Monroe. Gregory A. Sacha, Ph.D. As one of the graduate students in Dr. Pikal’s lab, I have received his endless support on the academics. He was also a great mentor in life, and was always there to provide guidance even after my graduation. His dedication, diligence and patience has set a great example for me, and it was an honor to learn from a famous scientist like him. Bingquan (Stuart) Wang, Ph.D. The decade also witnessed the publication of numerous studies by M.J. Pikal (sometimes referred to, and rightly so, as “the king of freeze-drying”), dealing with more advanced and realistic models of heat and mass transfer. A particular strength of Pikal’s contributions derives from his often-expressed philosophy that experimental tests need to be applied to validate all theoretical results. Evgenyi Prof. Pikal was a gem of a person!! It was much fun to be around him, and an enriching experience, both personally and professionally. As one of his former students, we always looked forward to his words of wisdom. A highly regarded leader in his field, yet very humble and treated everyone with utmost respect. Lokesh Kumar, Ph.D. Dr. Pikal had a pure heart and was more than an adviser to his students. He cared about everyone worked in his lab and treated us as his extended family. He not only gave us the opportunities to attend many scientific conferences and introduced us to other professors and scientists from pharmaceutical industries, but also brought us to his house/family, and all different outdoor funs such as Killington ski trips, hiking, kayaking. He was always supportive to his students and happy about any little achievements his students made. Will never forget that Dr. Pikal would grand me “one dollar” as a Reward for each “A” I earned with his exaggerated gesture and giant smile on his face. Xiaolin (Charlie) Tang, Ph.D. I began working with Professor Pikal as a Post-Doctoral Fellow in 1999. In terms of personal development, it was a scientific heaven. Every day you are not only learning new things, but you are continuously growing as a scientist in very complex, multidisciplinary field such as freeze-drying. Prior to joining Pikal’s laboratory I was aware about his groundbreaking work in lyophilization process modeling and lyo product characterization. Being in a middle of hurricane of new ideas was one of the most exciting moments in my life. I learned a lot about formulation, thermal analysis, and freeze-dryers characterization which completely turned and enhanced my career. We continue the collaboration with Dr. Pikal when I left UCONN (exploration of impact of protein to sugar ratio, secondary drying modeling, introduction of new PAT tools, etc.) and he was always ready to help. Besides science, I learned one important thing from Pikal – there is no impossible task, limits are only within us. Pikal was not only a great scientist (“king of freeze-drying”, “father of rational freeze-drying” as people called him), he was also a great teacher. He had a rare talent of explaining very complex things in easy and understandable way. But above all of it, Pikal was a great human being, caring about his students and people around him. He will remain as a role model for the rest of our lives. Serguei Tchessalov, Ph.D.

Words fail me when asked to put into a few sentences my thoughts about Dr. Pikal. It was a blessing to be co-advised by him when I was a student at UConn and even long after I had graduated. Dr. Pikal’s contributions to freeze-drying and beyond will remain an inspiration for his mentees, collaborators, and scientific fans! His ability to think and articulate clearly coupled with his passion for problem solving is greatly missed and will continue to motivate us to find practical solutions for challenging problems in the field. Bakul Bhatnagar, Ph.D. Dr. Pikal has been a great leader through LyoHub/CPPR for our team at Purdue, and we will always remember his warm encouragement and guidance along our journey. Tong Zhu, Ph.D. I got to know Mike in person when we began to develop the spray freeze-drying technology for industrial use starting early 2011. We were lucky to host him in our facility in 2012 on the occasion of a visit he paid to one of his former Ph.D. students, then a postdoc in the Basel pharma industry. As Basel is close to our facility, he spontaneously agreed to come to our site. During one of my visits to UConn, besides spray freeze-drying we also exchanged on his “Westfalia” Camper from Germany and on his favorites among the varieties of Bavarian “Weissbier”. Remarkable were his open-mindedness to look at new developments and options, his very fast also technical understanding and his perception of important aspects – all this combined with a personality full of positive energy, warmth and dedication. Bernhard Luy, Ph.D. Dr. Pikal truly defined the field of lyophilization with his fundamental and seminal work. He proved to be an inspiring role model for young scientists and established researchers alike. Andrea Allmendinger, Ph.D. Dr. Pikal was the co-supervisor for my Ph.D. thesis, and I had the pleasure to perform my research for 1 year in his lab at UCONN. His knowledge and experience was not only on freeze-drying but in a much wider scientific field, and his enthusiasm for science have been truly inspirational for me. Our discussions have helped me to look at the results from different angles, and to see the bigger picture. The combination of academic excellence with strong focus on application and provision of value to the practitioners that he embodied really made him stand out. In addition, he was a very generous and good-humored person with a genuine interest in his colleagues and co-workers. Stefan Schneid, Ph.D. Powerhouse of energy and a charismatic personality, that was Dr. Pikal. I have been so fortunate to get an opportunity to work with Dr. Pikal as a post-doctoral fellow. With his unparalleled experience he shared illuminating stories from his work which were some of the great learnings. The biggest lesson I learnt is to always do what is right and utilize systematic approaches to solve formulation and processing challenges instead of “throwing a patch on them and hoping they’ll go away.” The scientific community wholeheartedly celebrates his contributions and accomplishments to freeze-drying. As an advisor and a teacher, he has left an indelible mark on our lives. Ekneet Sahni, Ph.D. Conversations with Dr. Pikal were always lively, both in absence and presence of cognac, and what made them so special for me were that he was always humble, energetic, intellectually involved and genuinely compassionate about bringing the best in science and education for everyone that shared air with him. His impact and legacy continue to live on through his teachings, his student, non-profits like ISLFD-East coast wherein he served as Academic Advisor, various forums/publications. I continue to miss him and sometimes just thinking of him is enough to make me smile. This book is one small compilation of many things he taught us and brings together his mentee together. Thank You to editor for championing this effort. Akhilesh Bhambhani, Ph.D. I’d like to share a touching experience that reflects Prof Pikal’s legacy as a devoted family person besides an excellent mentor. During my postdoctoral fellowship in his lab, my family encountered a major health issue and Prof Pikal insisted that I spend at least a month and half in India to support my family in difficult times. The unique combination of a lively, passionate researcher and a caring advisor puts Prof Pikal in a unique league and he will continue to be a role model for all who had the privilege of being part of his lyo galaxy. Paritosh Pande, Ph.D.

Dr. Pikal had complete confidence in his students. He kept telling me how proud he was even for my small achievements. When we made mistakes, he would share his own mistakes in his graduate study so that we were not discouraged. I’ll never forget his confidence in me presenting as a Keynote speaker in place of him in my second-year grad study. Dr. Pikal is a role model in many ways, both passionate and committed as a professor, as well as generous and kind as a person. Rui Fang, Ph.D. I did not get a chance to work directly with Dr. Pikal, but I had several interactions with him in conferences and during his visits to Purdue University in the early days of LyoHub. He was a brilliant yet down-to-earth scientist who had a great love for the science of lyophilization as well as an unmatched sense of humor. He will always be a source of inspiration for all of us working in this field. Ehab Moussa, Ph.D. I had not yet met Dr. Pikal back in 1982 – he was then a scientist at Eli Lilly – when I received a manuscript for review from Journal of Pharmaceutical Sciences. The manuscript was entitled “Physical Chemistry of Freeze-Drying: Measurement of Sublimation Rates for Frozen Aqueous Solutions by a Microbalance Technique.” I was very familiar with the body of published research in pharmaceutical freeze-drying at the time, and I immediately saw that his paper was something original, innovative, and scientifically rigorous. He introduced several new ideas, including measuring the resistance of the partially dried solids to flow of water vapor. I immediately became a fan of Mike Pikal. Shortly after this paper – his first paper in the freeze-drying arena – was published, he asked me to help him teach a two-day short course in freeze-drying presented by the Parenteral Drug Association. The course was offered at a variety of locations across the US, and eventually in Europe. It gave me the opportunity to become personal friends with Mike as we skied together, hiked together, and drank more than a few beers together. He was always a pleasure to spend time with. Both personally and professionally, I’ll miss Michael Pikal for the rest of my own life. Steven L. Nail, Ph.D. I echo what is said above about Dr. Pikal, a true mentor, friend, guide and an inspiration all molded in one person. I was one of his early students and fortunate enough to join him at the time when he was setting up a lab and unpacking the boxes of books, literature and instruments that he had collected over 25 years at Eli Lilly, especially assembling pieces of freeze-drying microscopy that was built by him during his days at Eli Lilly and Dura dry-Dura stop from FTS. Having had the opportunity to work closely with him during student and professional life, few of his traits that impressed me and would remain inspiring in my life are his openness in taking/accepting students and had no borders when it comes to educating, training and developing – a true educationist. As I recollect and reflect on my days in his lab, he would introduce me to his family members and others as saying this is Feroz, the “Macho man” who wants to learn freeze-drying. There were several occasions I was out of step but I never saw a wrinkle on his forehead, he always smiled with forgiveness at my shortcomings and demonstrated patience and tolerance, a rare quality/commodity to be found in people of this status – a true angel. On several difficult occasions in my professional life, I reached out to him and seeked his help. He would always provide me with wise counseling, guidance, support and end up the conversation reminding me to feel free to call him on his mobile phone should there be any questions or need help. What a great feeling of comfort to have when you know there is someone of this status is there all the time to help you – a friend in need is a friend indeed. A rare quality of willingness to spend one’s own time and expertise to guide the development of another and feel happy and great about it. I can never thank Dr. Pikal enough for giving me space in his lab, spending time with me in training/discussions, imparting skills, knowledge and mindset that have shaped my career and life. I end by saying that a great mentor is hard to find, difficult to part with and impossible to forget. I feel truly blessed to have had him in my life. Feroz Jameel Ph.D.

Preface

Over the decades, the science and technology of lyophilization has grown by many folds mostly due to its widely applicability. Besides food industry, the high potential utility of lyophilization in the Pharmaceuticals and Biopharmaceuticals has triggered interest and led the industry and academic gurus of modern era freeze-drying, Felix Franks, Michael Pikal, Steve Nail and Alina Alexeenko to name a few, to pioneer the groundbreaking work, decipher and establish concepts in a simple understandable language to be taken forward to the next heights by their students and colleagues. To date tremendous amount of work has been done and vast knowledge and information has been created. The very purpose of this book was to bring the experts in various aspects of lyophilization together and compile the latest and the greatest knowledge and information to date, disseminate and make it accessible in an easiest way as a single stopshopping, a best way of “giving-back” to scientific community. The book covers end-to-end principles and practices of lyophilization in the drug product development and manufacturing with case studies and primarily aimed at product and process developers, and manufacturing engineers in the biopharmaceutical industry and academia. It starts of with overview of the lyophilization process and equipment followed by materials characterization and determination of freeze-drying properties before dwelling into aspects of formulation development, physics of freezing, development and optimization of process. It also discusses the characterization of the product, freeze dryers and engineering of scaling up. It concludes with strategies in validation, process controls/monitoring and an extensive bibliography. It is the hope of the editor that each scientist and engineer will find it useful and informative to trigger new ideas for approaching current challenges in Pharmaceutical/Biopharmaceutical lyophilization activities. Feroz Jameel

ix

Contents

Overview of Freeze Drying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Feroz Jameel Characterization and Determination of Freeze-Drying Properties of Frozen Formulations: Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Feroz Jameel Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dominik Heger, Ramprakash Govindarajan, Enxian Lu, Susan Ewing, Ashley Lay-Fortenbery, Xiaoda Yuan, Lukáš Veselý, Eric Munson, Larry Gatlin, Bruno Hancock, Raj Suryanarayanan, and Evgenyi Shalaev Concepts and Strategies in the Design of Formulations for Freeze Drying . . . . Feroz Jameel

1

21

39

63

Formulation Design for Freeze-Drying: Case Studies of Stabilization of Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andrea Allmendinger, Christina Häuser, Lokesh Kumar, and Ilona Vollrath

83

Challenges and Considerations in the Development of a High Protein Concentration Lyophilized Drug Product . . . . . . . . . . . . . . . . . . . . . . Xiaolin (Charlie) Tang, Yuan Cheng, and Mohammed Shameem

103

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water Formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Eni Bogdani, Séverine Vessot-Crastes, and Julien Andrieu

123

Primary Container Closure System Selection for Lyophilized Drug Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Robert Ovadia, Phillippe Lam, Holger Roehl, Renaud Janssen, and Roger Asselta

143

Vial Breakage During Lyophilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jim A. Searles and Ekneet K. Sahni

171

The Nucleation of Ice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gregory A. Sacha

179

Stresses, Stabilization, and Recent Insights in Freezing of Biologics . . . . . . . . . Rui Fang, Pooja Sane, Israel Borges-Sebastiao, and Bakul Bhatnagar

189

Lyophilization Process Understanding and Scaleup Using Ab Initio Vial Heat Transfer Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tong Zhu, Ehab M. Moussa, Feroz Jameel, Madeleine Witting, Sarah Ehlers, and Alina Alexeenko

199

xi

xii

Contents

Secondary Drying: Challenges and Considerations . . . . . . . . . . . . . . . . . . . . . . Kyu Yoon and Vivek Narsimhan

219

Design and Process Considerations in Spray Freeze Drying . . . . . . . . . . . . . . . Bernhard Luy, Matthias Plitzko, and Howard Stamato

243

LyoPRONTO: Deterministic and Probabilistic Modeling – Tutorial and Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Petr Kazarin and Alina Alexeenko

269

Utilizing Solid-State NMR Spectroscopy to Assess Properties of Lyophilized Formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ashley Lay-Fortenbery, Yongchao Su, and Eric J. Munson

291

Design of Moisture Specification Studies for Lyophilized Product . . . . . . . . . . Feroz Jameel

307

Laser-Based Headspace Moisture Analysis for Rapid Nondestructive Moisture Determination of Lyophilized Products . . . . . . . . . . . Derek Duncan, James R. Veale, Ken Victor, and Adriaan H. de Goeij

315

Application of PAT in Real-Time Monitoring and Controlling of Lyophilization Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Feroz Jameel, William J. Kessler, and Stefan Schneid

333

Process Analytical Technology (PAT) for Lyophilization Process Monitoring and End Point Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bingquan (Stuart) Wang and Xiaolin (Charlie) Tang

363

Advances in Process Analytical Technology: A Small-Scale Freeze-Dryer for Process Analysis, Optimization, and Transfer . . . . . . . . . . . . T. N. Thompson and Spencer Holmes

379

Overview of Heat and Mass Transfer Modeling in Lyophilization to Create Design Spaces and Improve Process Analytical Technology (PAT) Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tong Zhu, Feroz Jameel, Pasita Pibulchinda, Vaibhav Kshirsagar, and Alina Alexeenko

405

Application of QbD Elements in the Development and Manufacturing of a Lyophilized Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Feroz Jameel

423

Characterization of Freeze Dryers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Feroz Jameel and Serguei Tchessalov Principles and Practice of Lyophilization Process and Product Development: Scale-Up and Technology Transfer . . . . . . . . . . . . . . . . . . . . . . A. Bhambhani, J. Stanbro, A. Sethuraman, and P. Pande Lyophilization Validation: Process Design and Modeling . . . . . . . . . . . . . . . . . Feroz Jameel, Alina Alexeenko, Akhilesh Bhambani, Gregory Sacha, Tong Zhu, Serguei Tchessalov, Puneet Sharma, Ehab Moussa, Lavanya Iyer, Sumit Luthra, Jayasree Srinivasan, Ted Tharp, Joseph Azzarella, Petr Kazarin, and Mehfouz Jalal

451

465 489

Contents

xiii

Lyophilization Validation: Process Qualification and Continued Process Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Feroz Jameel, Alina Alexeenko, Akhilesh Bhambani, Gregory Sacha, Tong Zhu, Serguei Tchessalov, Puneet Sharma, Ehab Moussa, Lavanya Iyer, Sumit Luthra, Jayasree Srinivasan, Ted Tharp, Joseph Azzarella, Petr Kazarin, and Mehfouz Jalal

513

Homogeneity Assessment of Lyophilized Biological Drug Products During Process Performance Qualification . . . . . . . . . . . . . . . . . . . . Fuat Doymaz, Brenda S. Ramirez, and Chris Cherry

541

Informed Manufacturing Through the Use of Big Data Analytics for Freeze Drying Process and Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vaibhav Kshirsagar, Arnab Ganguly, and Andrew Reese

555

Multivariate Analysis for Process Understanding, Continuous Process Verification, and Condition Monitoring of Lyophilization Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pierre-Philippe Lapointe-Garant, Reza Kamyar, Ramezan Paravitorghabeh, and Zilong Wang

577

Lyophilized Drug Product Cake Appearance: What Is Acceptable? . . . . . . . . . Sajal Manubhai Patel, Steven L. Nail, Michael J. Pikal, Raimund Geidobler, Gerhard Winter, Andrea Hawe, Juan Davagnino, and Shailaja Rambhatla Gupta

595

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

619

Overview of Freeze Drying Feroz Jameel

Abstract

Freeze drying which is also termed as lyophilization is one of the drying technologies that is commonly used to enhance the storage stability of products that have marginal stability in liquid state. The chapter starts with discussing the advantages and disadvantages of lyophilization, and situations where its use can be beneficial. It describes in detail the equipment involved in freeze drying, the various phases of freeze drying process, the design and composition of the formulation for freeze drying and various techniques to monitor and control the process. Finally, it also briefly describes the kinetics in the lyophilized solid state and model to predict the stability and shelf-life of the product. Keywords

Lyophilization · Freezing · Primary drying · Secondary drying · Stabilizers · Bulking agents · Collapse temperature · Glass transition temperature · Shelf temperature · Chamber pressure · Sublimation rate

1

Why Freeze Drying?

It is central to the discovery and development of a new therapeutic entity that it is filled and finished in the right dosage form that has adequate shelf-life and meets patient compliance, otherwise all the years of efforts will not bring any benefit to the pharmaceutical/biopharmaceutical industry. Therapeutic molecules that have marginal stability in the aqueous systems are often dried to improve the stability and enhance the shelf-life as it is known that the stability of most (small) molecules normally increases in the order of solution < glassy solid < crystalline solid, due to the increasingly restricted mobility of the reacting species in these phases [1–3]. Several drying technologies are currently available to the formulation scientist, such as lyophilization, spray drying, spray-freeze drying, supercritical fluid technology, foam drying, and vacuum drying microwave drying, and every technology has merits and demerits that limit its use [4]. Although process yields with spray drying continues to be a challenging disadvantage, spray drying, which is increasingly becoming the most important method for dehydration in the food industry, is seeing success in producing pharmaceutical and also biopharmaceutical powders where the size and morphology are central to the delivery and performance of the protein powder [5]. However, hot air, control of residual moisture content, and stresses imposed by atomization and interfaces continue to be other challenging factors in meeting stability requirements specifically for biopharmaceuticals. Given the limitations of other drying technologies, historically lyophilization remained the drying method of choice for both pharmaceuticals and biopharmaceutical industry due to its overwhelming advantages as noted below [6]. Lyophilization which is also termed as freeze drying is a dehydration process that converts water in to ice and removes it through sublimation, which is termed as primary drying phase. The unfrozen water is removed through desorption by the application of heat which is called secondary drying. Thus, it converts solutions into solids thereby improving/enhancing the F. Jameel (✉) Nimble BioSolutions, Gurnee, IL, USA # America Association of Pharmaceutical Scientists 2023 F. Jameel (ed.), Principles and Practices of Lyophilization in Product Development and Manufacturing, AAPS Advances in the Pharmaceutical Sciences Series 59, https://doi.org/10.1007/978-3-031-12634-5_1

1

2

F. Jameel

storage stability of materials that otherwise have marginal stability in solutions. The removed water from the product is reconverted into ice on the chilled coils in the condenser.

Advantages: 1. It is a low-temperature process, and hence is expected to cause less thermal degradation compared to a “high temperature” process, such as spray drying; 2. It does not involve a terminal sterilization step and maintains sterility and “particle-free” characteristics of the product much more easily than do other processes; 3. It offers a method for controlling residual moisture content and headspace gas composition in the vial for those products whose storage stability is influenced by residual moisture content and vial headspace gas composition such as oxygen; 4. Its scale-up is easy and it provides reasonable process yields.

2

Equipment

The Main Function of the integrated Freeze Dryers and Automated System is to facilitate lyophilization (freeze drying) of aseptically filled drug product while preserving its aseptic integrity. In order to preserve the aseptic integrity, the unit must be capable of executing Clean In Place and Sterilize in Place cycles, to defined Acceptance Criteria. Upon completion of the freeze drying process, the unit must be capable of removing ice buildup on the condenser surfaces through a Defrost cycle. The unit will also be capable of performing recipe defined major sub-system tests as part of a Plant Test cycle to verify system integrity, evaluate performance, or for maintenance troubleshooting purposes. The integrated Freeze Dryer and Automated system will feature automatic loading, unloading, and processing. The specific process, along with its associated parameters and alarm limits, will be recipe selectable by the user. Additionally, at the start of a process, a Cycle Run Report will initiate automatically to collect critical data and events for that batch. The Cycle Run Report will become an integral part of the batch record for that process. The Automated System will continuously track a user defined set of data which will be available for real-time or future access. A lyophilizer consists of two major components, the drying chamber and a condenser, in addition to refrigeration system, vacuum system, and control system [7].

2.1

Chamber

The drying chamber consists of several shelves made up of AISI type 316L stainless steel on which the vials or containers partially stoppered containing the solution are placed and these shelves are temperature-controlled ranging from 60 °C to -50 °C through the flow of the heat transfer fluid. The shelves are moved via a hydraulic ram of sufficient diameter to generate stoppering forces. The shelf movement facilitates the loading/unloading of vials and the cleaning process. The common refrigeration systems provide cooling to the shelves through heat exchangers within the shelf heat transfer fluid systems. Depending upon the purpose the number of shelves and vials varies, a research grade typically has three shelves accommodating a few hundred vials, while a typical production scale freeze dryer may have 10–20 shelves (Fig. 1) accommodating 50,000 to 100,000 vials depending upon size.

2.2

Condenser

The condenser chamber consists of coils capable of maintaining very low temperature, -70 °C to -100 °C, and connected to the drying chamber through a tube/spool. The water vapors generated from sublimation of ice travel through the tube into the condenser chamber and condense on the surface of the cooled coils. Various methods are used to chill the condenser such as by direct expansion of refrigerant which can provide operating temperature ranging from -45 °C to -100 °C, liquified gases such as liquid nitrogen which provide an operating temperature of -70 °C and below, and a circulated brine which can provide an operating temperature range of -50 °C to -70 °C. Liquid nitrogen is believed to be more cost-effective from an

Overview of Freeze Drying

3

Fig. 1 Freeze Dryer Chamber with shelves. (Courtesy SP Scientific)

energy consumption point of view. Condensers come in two designs: internal and external. External condensers are preferred as they offer a few advantages: (1) barometric control possible (iso valve), (2) faster turnaround, (3) less oil back streaming, (4) temperatures more uniform, and (5) higher ice capacities.

2.3

Vacuum System

A set of redundant dry vacuum pumps with blowers, operating in tandem or individually, are incorporated for reducing pressure in the chamber, condenser, and associated piping. The desired pressure of 50–150 mTorr is maintained through vacuum pump and bleeding of nitrogen or other gas.

2.3.1 Chamber Pressure Control The chamber pressure control system consists of the sterile vacuum vessel containing the shelves onto which the product is placed, sterile nitrogen supply piping, the gas bleed control valve, the dual serial nitrogen gas filters, the redundant booster pumps, the redundant dry vacuum pumps, the redundant chamber MKS vacuum gauges, and the control system. The gas bleed control valve is an element requiring PID control logic. The system uses the chamber MKS gauge as the basis for the PID-controlled chamber pressure loop. Chamber pressure is controlled in the freeze drying process by specifying the chamber pressure setpoint. After the recipe download and cycle initialization is completed, a recipe variable indicator is used to determine if one of the vacuum systems will be shut down when the chamber reaches the evacuation pressure. The system selects and maintains activation of one of the redundant dry vacuum pumps based on whichever was inactive in the previous process. Pressure control at the specified setpoint is accomplished by modulating the gas bleed control valve based on the difference between the actual chamber pressure measured in the chamber above the shelves and the setpoint pressure in the logic. If the actual pressure is greater than setpoint, then the sterile nitrogen gas flow is reduced according to the PID control algorithm. Similarly, if the actual pressure is less than setpoint, then additional nitrogen is requested according to the control loop tuning.

4

F. Jameel

The pressures in the drying and condenser chambers are controlled through pressure gauge capacitance manometer and monitored by the pirani gauge installed in both chamber and condenser.

2.4

Control System

The Control System Specifications constitutes the sophisticated controls including full HMI and PL. Global functions of the control system include system security, general batch management, data storage/archiving, cycle Initiation, recipe download, process monitoring, equipment monitoring, alarm monitoring, alarm responses, trend monitoring, cycle run report generation, E-Stop (emergency stop), process modification, and equipment maintenance.

3

Formulation

In addition to drug active, a typical formulation for freeze drying consists of several ingredients often referred to as excipients are included in the formulation with intended role. Although the rule of thumb is to keep the formulation simple, a dilute solution consisting of ≤1% solids is practically not feasible to freeze dry into a cake as there will not be enough solids to form a structure. In those cases where the drug active is potent and required in small quantities, a bulking agent/filler is added to provide both mechanical strength and elegance to the cake. The bulking agents could be crystallizable (mannitol, glycine) or amorphous (sucrose, trehalose) in nature. In the case of protein products, stabilizers are included in the formulation to stabilize the protein against the stresses encountered during freezing, drying, and upon storage. Surfactants (Polysorbates 20 or 80) are used to stabilize the protein against ice–air or/and ice–solution interfaces during freezing/thawing [8–12]. Cryoprotectants such as disaccharides (sucrose, trehalose) and polyols (sorbitol, PEG) are added to the formulation to protect the protein against freezing-induced denaturation, while lyoprotectants are included in the formulation to protect the protein against drying stresses during freeze drying and upon storage [13]. In situations where the collapse temperature of the formulation is low and in order to improve the collapse temperature to be able dry warmer and faster and again process efficiency, collapse temperature enhancers (cyclodextrins, hydroxy ethylstarch) are added to the formulation [6]. Some protein products may require stabilizers to address the chemical instability associated with protein such as buffers to control pH, antioxidants to address oxidation, and chelators to address metal-induced reactions. The formulation and the freeze drying process are interrelated, freeze drying properties of the excipients affects the overall collapse temperature/Tg’ of the formulation which in turn dictates the selection of lyophilization process conditions and the efficiency or ease with which it can be dried, hence, selection of excipients is central to design of the formulation and process. The selection of these excipients, their weight ratios, and processing conditions are critical to final product quality attributes and is described in detail in chapter “Concepts and Strategies in the Design of Formulation for Freeze Drying”. The collapse temperature and the Tg’ can be measured directly using freeze-drying microscopy (FDM) and modulated differential scanning calorimetry (MDSC) respectively. The Tg’ of a monophasic multicomponent system can also be estimated from the Tg’ values of the individual components using the Fox equation [14]. W W 1 = iþ 2 T g T gi T g2

ð1Þ

where Wi is the weight fraction of component “i” and Tgi is the glass transition temperature of pure component “i”. Equation (1) can be applied to determine the Tg’ of systems containing two or more amorphous components wherein Tgi is the Tg’ of aqueous component “i” and Wi is the weight fractions of the solute relative to the total mass of solutes. The value of Tg’ determined by DSC is approximately 2–3 degrees lower than the actual Tc measured using FDM. A eutectic system is a mixture of two or more crystalline compounds that melt together at the lowest freezing temperature. In a mixed formulation system where crystalline phase constitutes the major weight fraction of the matrix, the Te will be the critical temperature of the formulation. Freeze drying with the product temperature above the Tg’ of the amorphous phase of the formulation but below Te of the crystalline component will result in the collapse of the amorphous component; however, the crystalline phase will provide the necessary mechanical support to maintain the cake structure and elegance. This is an effective strategy to enable fast and robust freeze drying cycles, but the impact of drying above the collapse temperature on the product stability needs to be evaluated [6].

Overview of Freeze Drying

4

5

Process

Freeze drying process is classified into three phases: freezing, primary drying, and secondary drying (Fig. 2). The liquid solution that contains drug active along with other excipients (buffer, stabilizers, and bulking agent) to be freeze dried is filled in the glass vials and placed on the shelves of the freeze dryer that are temperature-controlled. The freezing phase starts with the lowering of the shelf temperature typically from -40 °C to -45 °C and held there for a few hours depending upon the fill volume to ensure complete conversion of water into ice and ensure complete solidification of ice. Once the all the vials are frozen the next step is to carry out primary drying which is a sublimation phase where the pressure is reduced to deep vacuum typically in the range of 60–150 mTorrs. As it can be seen from the phase diagram below (Fig. 3), lowering of the pressure enables direct conversion of the ice into phase through sublimation. Additionally, heat is applied to expedite the drying process. Primary drying is the longest phase in the freeze drying process and can take from few days up to a week depending upon the components in the solution and selection of chamber pressure and the shelf temperature. Not all vials dry at the same time due to differences in heat transfer across the shelf and between the

Freezing stage Atmospheric pressure

Primary drying stage

Secondary drying stage

High vacuum Recovery to atmospheric pressure/fully stopper

Shelf temperature

Temperature

-40°C Product temperature Semistoppering

Formation of ice crystals

ials

ch Bat

Void formation

Sublimation of ice crystals

Desorption of unfrozen water

Time

Fig. 2 Depiction of events in the various phase of freeze drying

critical point

218

Water ice 1 P atm triple point 0.006

water vapor

The large drawing is not too scale. A scale drawing looks more like the one above.

sublimation

0

0.01

Fig. 3 Phase diagram of water - phase curves and triple point of water

T °C

100

374

of v

6

F. Jameel

shelves, care is taken through a soak period to ensure all vials have completed primary drying prior to advancing to secondary drying. During the freezing phase not all the water is converted to ice due to depression of freezing point with increase in concentration of solutes as the freezing progresses. Depending upon the composition of the solution, there will be still 10–15% of water remaining unfrozen and it can remain free or bound to excipients. This water is removed during secondary drying through the application of heat by carrying out secondary drying at an elevated of temperature 25–40 °C.

4.1

Freezing

Physics of the Freezing and Crystallization Process (Adapted from Book Chapter in “Development of Biopharmaceutical Drug-Device Product”) The physicochemical changes and the thermal events that take place during the process of freezing and crystallization depend on the composition of the solution. Figure 4 depicts the freezing profile of two solutions: pure water (ABCDE) and a sucrose aqueous solution (A’B’C’D’). In pure water, upon lowering the temperature from point A, the nucleation or critical mass of nuclei is not formed until point B. Once the nuclei are formed, the crystallization process starts. Since crystallization is an exothermic reaction, the latent heat of fusion is given out, and the temperature rises from B to C. The event shown as the solution progresses from point A to B is described as the degree of supercooling the water undergoes prior to nucleation. This degree of supercooling is dependent upon the cooling rates employed, and the purity of water (free of particles which serve as nucleation sites), with the lower number of free particles leading to a higher degree of supercooling. Point C, which corresponds to 0 °C, is the equilibrium freezing point of pure water, and at this point, the water continues to crystallize until point D. Once point D is achieved, all of the water is converted into ice, and because the crystallization process is complete and no heat is given out, the temperature starts dropping to the set point, E. The freezing time is usually defined as the time from the onset of nucleation to the end of the crystal growth phase. The size of ice crystals formed during crystallization (from C to D) is dependent upon the degree of supercooling: faster cooling rates lead to higher degrees of supercooling and smaller ice crystal size (Fig. 5) [14]. A different freezing behavior is expected once a solute is added to pure water. A solute-containing solution is governed by Roult’s law which relates the vapor pressure of the solution to that of the pure solvent based upon solute concentration. Figure 4 shows the key differences that exist between pure water and a sucrose solution. First, B’ is not the same as B in terms of temperature; a sucrose solution nucleates earlier than B because of the presence of sucrose molecules, which act as nuclei. Secondly, C’, the freezing point temperature, is not as high as C, due to the 4.1.1

Fig. 4 Freezing profiles of pure water (ABCDE) and a sucrose solution (A’B’C’D’)

Overview of Freeze Drying

7

Fig. 5 Effect of freezing rate on the morphology of ice

Fig. 6 Freezing curves of mixtures. (Adapted from Ref. [18])

initial freezing point depression caused by the presence of sucrose. Both of these events are dependent on the concentration of solutes in the solution. Additionally, in aqueous solutions containing solutes, a phenomenon called cryoconcentration is observed [15, 16]. As the water starts converting to ice upon cooling, the freezing front moves forward leaving behind the solutes to concentrate, resulting in pockets rich of solutes. These pockets rich in solutes further depress the freezing point of free water, and this phenomenon continues (C’D’) with the cooling, leaving some residual unfrozen water, regardless of how low the cooling temperature is set. Also, as cryoconcentration increases, the viscosity of the free water increases, which decreases the mobility and diffusion properties of the system and inhibits the crystallization process. The cryoconcentration process establishes the freezing curve as shown in Fig. 6. This curve can be used to predict the amount of ice at any given temperature, which in turn, is a function of the freezing point depression caused by the concentration of solutes in the solution. Recrystallization of Ice/Ostwald Ripening Ostwald ripening is a phenomenon where the bigger ice crystals become larger at the expense of the smaller ice crystals during warming and cooling. Smaller ice crystals are unstable and tend to melt upon temperature fluctuations due to the cycling of the freezers and/or the automatic defrosts. As a result of the smaller ice crystals melting, the amount of unfrozen water in the freeze concentrate phase increases, which will refreeze upon a decrease in temperature, but does not renucleate.

8

F. Jameel

Instead of forming new ice crystals, they get deposited on the surface of existing larger crystals so the net result is that the total number of crystals diminishes and the mean crystal size increases (Fig. 7). This advantage of this phenomenon is utilized in freeze drying to achieve homogeneity in ice crystal size and favor larger ice crystals to facilitate faster drying.

Fig. 7 Cryo-scanning electron micrograph images illustrating the effect of temperature fluctuations on crystal size. The top panel shows images before the temperature fluctuation, the middle panel illustrates the tremendous increase in crystal size that has occurred after heat shock, and the bottom panel shows an example of accretion, where crystals fuse as they grow. (Adapted from the work of A. Flores and H. D. Goff. https:// www.uoguelph.ca/foodscience/ book-page/temperaturefluctuations-and-icerecrystallization)

Overview of Freeze Drying

9

Formation of the Glassy Phase in Frozen Systems Upon lowering the cooling temperature, the water starts to form ice through a two-step crystallization process: nucleation followed by propagation. As the temperature continues to decrease, water is converted into ice, resulting in the concentration of the solutes in the free, unfrozen water. An equilibrium freezing temperature exists for each ice/unfrozen phase ratio, which is a function of the solute concentration. Figure 8 depicts the equilibrium thermodynamic process modelled on a phase diagram as an equilibrium freezing (liquidus) curve, which goes from the melting temperature (Tm) of pure water (0 °C) to the eutectic temperature (Te) of the solute. Te is the point at which the solute has been freeze-concentrated to its saturation concentration. If the solutes reach supersaturation, then crystallizable excipients such as mannitol or glycine will crystallize and precipitate. The other solutes will remain amorphous, and when the critical solute-dependent concentration is reached, the unfrozen amorphous freeze concentrate exhibits restricted mobility. At this point, the physical state of the system changes from viscoelastic liquid to an amorphous solid phase called “glass” [16]. The temperature at which this occurs is called the glass transition temperature of maximally freeze-concentrated systems (Tg’), and the corresponding unfrozen water and amorphous solutes concentrations are termed Wg’ and Cg’, respectively (Fig. 8). A glass is defined as a non-equilibrium, metastable, amorphous, disordered solid of extremely high viscosity (e.g., viscosity coefficient ranging from 1010 to 1014 Pa. s.) as a function of temperature and concentration. The glass transition curve extends from the glass transition temperature (Tg) of pure water (-134 °C) to the Tg of pure solute. The equilibrium phase diagram and the kinetically derived state diagram can be modelled together to form a supplemented state diagram. The supplemented state diagram illustrating the solid/liquid coexistence boundaries and glass transition profile for a binary sucrose/water system is shown in Fig. 8. Below and to the right of the glass transition line, the solution exists in the amorphous glass state, with or without ice present, depending on the temperature and freezing path followed. On the other hand, above and to the left of the glass transition line, the solution is in the liquid state, with or without ice, depending on the temperature. Point A in Fig. 8 depicts the initial concentration of 20% of sucrose at room temperature, and point B depicts the initial glass transition temperature (Tg) of the 20% sucrose solution (if the solution could be undercooled to this temperature without ice formation). Upon slowly cooling the sucrose solution, nucleation and subsequent crystallization begin at point C. This occurs after some degree of supercooling due to the presence of sucrose, which initiates the freeze concentration process following the water removal as ice. As ice crystallization proceeds, the continual increase in solute concentration (removal of water) further depresses the equilibrium freezing point of the unfrozen water phase in a manner which follows the liquidus curve (shown as path C). The increased concentration results in the glass transition line being moved up with a rapid increase in viscosity (path B), thus improving the Tg of the unfrozen water phase. Co-crystallization of solute at the Te is unlikely to happen as sucrose is not a crystallizable excipient, and thus freeze concentration continues past Te into a nonequilibrium state because the solute becomes supersaturated. When a critical solutedependent concentration is reached, the unfrozen liquid exhibits very restricted mobility, and the physical state of the unfrozen water phase changes from a viscoelastic liquid to a brittle, amorphous solid glass. At the Tg’, the supersaturated solute takes on Fig. 8 Phase diagram of equilibrium freezing for binary sucrose-water system. (Adapted from Ref. [18])

Tg sucrose (52°C)

Liquid Phase

Eutectic Line

A

Temperature

Tf (0°C) Te (-9.5°C) Tm' (-32°C) Tg' (-40°C)

Equilibrium freezing (liquidus) curve C NonE equilibrium freezing

D

Tg Ice Phase

Glass Phase

B Tg H2O (-134°C) 0% Solute

Glass transition line Cg/Wg Cg'/Wg' Concentration

100% Solute

10

F. Jameel

solid properties because of reduced molecular motion, which is responsible for the tremendous reduction in translational, and not rotational, mobility. It is this intrinsically low mobility below Tg’ that dictates that protein products to be stored frozen below their Tg’. Warming from the glassy state to temperatures above the Tg’ provides tremendous increases in mobility and diffusion, not only from the effects of the amorphous to viscous liquid transition but also from increased dilution due to the melting of small ice crystals that occurs almost simultaneously (Tg’ = Tm’). The time scale of molecular rearrangement continually changes as the Tg is approached. Therefore, some enhanced stability at temperatures above Tg’ can be gained by minimizing the delta T between the storage temperature and Tg’, which can be achieved either by reducing the storage temperatures or enhancing the Tg’ through freezing methods or formulation. Hence, knowledge of the glass transition temperature provides a clear indication of molecular diffusion and reactivity and, therefore, shelf-stability.

4.1.2 Impact of Freezing Process on Protein Solutions and Modes of Denaturation The freezing process can denature the protein through three mechanisms: (1) cryoconcentration, (2) ice surface denaturation, and (3) cold denaturation. 1. Cryoconcentration The objective of freezing is to lower the temperature to a point that the solution is completely solidified, thereby enabling sublimation and also arresting reactions that lead to degradation of the protein in the liquid state. As the solution is cooled, the liquid may supercool to a temperature well below the equilibrium freezing temperature, particularly in the case of vials and small containers. With sufficient supercooling, nucleation of ice proceeds rapidly, and the system freezes quickly. During the freezing of bulk solution in large-scale containers, freezing occurs slowly, and as the liquid water converts to ice, the protein and formulation excipients are progressively concentrated in the regions between the ice crystals. After the initial ice nucleation and crystallization, the product cools with continuous conversion of water to ice. As this occurs, the amount of water in the remaining liquid phase decreases, and the concentration of the solute in the remaining solution increases. This freeze concentration effect results in an increase in protein concentration, which dramatically increases the probability of molecular collisions. The bimolecular collisions between protein molecules can lead to denaturation of the protein through aggregation. For example, although a reduction in temperature from 5 °C to -40 °C would reduce the rate constant significantly, the increase in the concentration factor due to the increase in concentration has a more significant impact, thus resulting in a net increase in reaction rate. If excipients such as ionic salts and buffer species are present in the formulation, they will also concentrate during the freezing process. For example, during the freezing process, a formulation containing 0.15 M NaCl will increase to 6 M NaCl before it forms eutectic with ice. Exposure of protein to high ionic strengths could contribute to the instability of the native conformation [19]. In addition, the effect of freezing on buffer choice must be considered. Buffers are included in the formulations to help maintain a stable pH. However, during the freezing process, decreases in solubility with a simultaneous increase in concentration can cause selective crystallization of the buffer component and result in dramatic pH shifts. The classic example is the sodium phosphate buffer system. It shows a dramatic decrease in pH of about four units due *to the crystallization of the basic component. On the other hand, the potassium phosphate system shows an increase in pH upon freezing [20]. Mitigation Strategies for Cryoconcentration Effects 1. The ice front velocity should be higher than the diffusion rate of solutes so that the protein molecules/solutes become entrapped by the freezing front. This can be achieved through the combination of shorter freezing path lengths and efficient external heat transfer. 2. Increase the temperature differential between the heat transfer fluid and the product, which shortens freezing path lengths. 3. Minimize the product residence time within the cryoconcentrated stage. 4. Control the freezing rate within known limits. 5. Use small-scale containers for efficient heat transfer and rapid liquid-to-solid phase transition. 6. Do not mix while freezing. Mix during thawing and aim for uniform melting with mixing. 2. Ice–Liquid interface Denaturation Through phosphorescence lifetime decay of tryptophan residues, it was demonstrated that freezing of aqueous solutions of proteins causes perturbation or loosening of the native fold due to denaturation at the ice-liquid interface, which often results in the loss of secondary and tertiary structure [21, 22]. In some cases, this denaturation is largely reversible upon melting of the ice, and in other cases, substantial loss of activity is observed. This variation is believed to be due to its dependence on the residual volume of liquid water in equilibrium with ice and on the morphology of the ice.

Overview of Freeze Drying

11

Strategies to Minimize Ice–Liquid Interface 1. Avoid extensive undercooling which leads to flash nucleation and smaller ice crystals. 2. Optimize freezing rate to achieve low ice surface area. 3. Investigate the use of formulation components to avoid surface interaction. The addition of cryoprotectants such as polyols and disaccharides (e.g., sorbitol, glycerol, sucrose) and surfactants profoundly attenuates or even eliminates the perturbation. 3. Cold Denaturation While some proteins survive freezing with little or no measurable loss in activity, the freezing process irreversibly inactivates others. Just as proteins undergo thermal denaturation at elevated temperatures, proteins also undergo spontaneous unfolding at very low temperatures, denoted “cold denaturation” [23]. This is partly because of the unsuitable environment created during freezing. As discussed above, as solute species are concentrated, the ionic strength increases, the pH may shift, and most importantly, the “hydrophobic interactions” that stabilize the native conformation of the protein in water are reduced or eliminated as bulk water is removed from the protein phase. The transition between the denatured and native state is described by changes in enthalpy (ΔH ), entropy (ΔS), and Gibbs free energy (ΔG) through the following equation: ΔG = ΔH - TΔS Gibbs free energy relates to the amount of work required to disrupt the structure of a protein molecule and is used to describe the protein stability. The Gibbs free energy equation has a parabolic shape (Fig. 9), which suggests that both high and cold denaturation is thermodynamically possible. The maximum stability of the protein at its native state temperature (Ts) occurs when the entropy difference between the native and denatured state is zero. This means that the stability depends mainly on the enthalpy differences between the native and denatured states. The enthalpy of transition can be determined as a function of temperature using either microcalorimetry or modulated DSC. Cold denaturation is not easy to determine experimentally since the declining part of the Gibbs free energy curve below Ts may be below 0 °C. Although cold denaturation is not widely reported for protein drugs, it remains a possibility. The denaturation of protein resulting from cryoconcentration effects and ice–liquid interface adsorption can be eliminated or attenuated through the optimization of critical freezing parameters. If the cause for the protein denaturation during freezing is due to cold denaturation, then addition of small amounts of one of the “excluded solutes,” termed cryoprotectants (amino acids, polyols, sugars, and poly(ethylene) glycols), in molar concentrations will increase the free energy of denaturation. Therefore, the protein is protected against cold denaturation through the preferential exclusion of solutes from the surface of the protein [25–34].

Fig. 9 Schematic representation of the protein stability curve illustrating the temperature dependence of the free energy of unfolding, ΔG. (Replotted from Ref. [24])

12

F. Jameel

Strategies to Minimize Cold Denaturation 1. Formulation additives to increase freeze-thaw stability: (a) thermodynamic stabilizers, (b) cryoprotectants, (c) glass forming substances 2. Rapid liquid-to-solid phase transition

4.1.3 Freezing Rate The main purpose of the freezing phase is to convert water into ice and achieve complete solidification of ice. In order to achieve complete solidification, the formulation solution needs to be frozen to at least 5–10 °C below the Tg’ and to be held at it for at least 1–2 h depending on the fill volume to ensure the complete crystallization of water. Failure to achieve this will potentially result in the upliftment of the cake when vacuum is applied during primary drying as indicated in Fig. 1. The morphology and size of the ice crystals formed during the freezing phase influence the performance of the subsequent phases of freeze drying. Both features are dependent on the freezing protocol such as the cooling rates and the annealing time and temperature. The freezing process is characterized by two parameters, the degree of undercooling and ice crystallization rate. The degree of undercooling which is also loosely called degree of supercooling is the difference between the equilibrium freezing point and the temperature at which ice crystals are first formed in the solution. It is dependent on cooling rates and particle environment. The faster the cooling rate the higher would be the degree of undercooling. It is an important parameter for freeze drying because it determines the number of nuclei which determines the number of ice crystals which in turn determines the size of the ice crystals as the total amount of water that freezes is fixed. Ice crystals after sublimation leaves behind the pores, the size of the ice crystals determines the size of the pores which impacts/influences the mass transfer rate/drying rate. Hence, higher the degree of undercooling the smaller will be the size of ice crystals more resistance to mass transfer [35], longer will be the primary drying time and shorter will be the secondary drying time. Since secondary drying is a desorption phenomenon and dependent upon specific surface area, smaller the size of ice crystals bigger will the surface area and faster will be the secondary drying. The degree of undercooling and the size of the ice crystals will also impact the product quality as protein tends to denature at the ice/air interface, bigger the number of ice crystals greater will be the potential of interfacial denaturation of labile proteins [15, 37, 38]. Additionally, the freezing protocol influences the physical state of the excipients and their intended role in the formulation. Certain excipients either remain amorphous or crystallize depending upon several factors like molecular structure, solubility, concentration, and the presence of other formulation components [39]. The second freezing parameter, ice crystal growth rate determines the time the product spends in a freezeconcentrated fluid state. As indicated above it is in the freezeconcentrated state where all the deleterious reactions occur, hence, it is prudent to minimize the residence time of the product in that stage by having a rapid rate of ice crystal growth. Generally speaking, in freezing process, the heat removal rate dictates the rate of ice crystal growth. In vial freeze drying, since heat removal occurs through the bottom of the vial, three things can be done to expedite the ice crystal growth: (1) low shelf temperature, (2) small fill volume-to-container area ratio (i.e., small fill depth), and (3) good contact between the container bottom and the freeze dryer shelf. Another important aspect of freezing protocol is not to place warm vials on the cold shelf as it will produce a product with two distinct cakes as the solution at the bottom will freeze with faster freezing rate and the upper part of the vial will freeze with a slower rate as the freezing front travels from bottom to top especially in cases with high fill height. Hence, it is recommended that the shelves be pre-cooled to -5 °C and the vials are allowed to equilibrate to -5 °C before cooling down to low shelf temperatures to produce uniform degree of undercoolings and uniform cake structure across the vial. 4.1.4 Annealing Annealing is included in the freezing protocol as a thermal treatment step with one of the two reasons either to achieve bigger size of ice crystals through a phenomenon called Ostwald ripening to facilitate faster homogenous/uniform primary drying or to achieve complete crystallization of crystallizable excipients such as mannitol and glycine sodium chloride. Ostwald ripening is a phenomenon where large ice crystals grow at the expense of the smaller ice crystals. Whatever the intention might be in both cases, the solution needs to be frozen 10 degrees below the Tg’ and warmed up and annealed at least 10 degrees above the Tg’ to provide mobility and nuclei to grow into crystal. The optimal temperature and time can be determined through DSC, XRD [40], and/or FDM [39]. Complete crystallization can then be confirmed using several techniques including (1) Tg’ annealing temperature curves, (2) the area under the eutectic melting endotherm in a frozen system, (3) the area under the bulking agent melting endotherm in dry powder system, and/or (4) the absence of an exotherm upon heating the dry pow der on DSC.

Overview of Freeze Drying

13 Product TC 0 non-TC

–10

C –20

Shelf

–30 –40 0

1 Time, hours

2

3

Fig. 10 Experimental observation of freezing bias. (From Ref [6])

4.1.5 Effect of Thermocouples on Freezing Rates Typically, during development stages temperature probes/thermocouples are placed inside the vials to monitor the product temperature. Vials containing thermocouples/sensors undergo less degree of undercooling causing nucleation of ice at a lower temperature, freeze faster than vials without temperature sensors [41] resulting in larger ice size crystals, leaving larger pores after sublimation, less resistance to mass transfer, freeze dry at lower temperature consequently less time to dry than the rest of the vials. Hence, vials with thermocouples are not considered representative of the whole batch, hence, a soak period of 10–15% are allocated for rest of the vials to catch up with the non-thermocouples vials and to compensate for the bias in drying time. This effect is particularly significant in manufacturing where the particle-free environment is due to class 100 area. The placement of temperature sensors in the vials introduces a significantly higher level of heterogeneous nucleation sites, thereby causing nucleation of ice at a lower temperature than in vials without temperature sensors (see Fig. 10). Additionally, location of the monitored vials in the vial array on the shelf is also an important factor in recording representative data.

4.2

Primary Drying

The primary drying phase in freeze drying process is considered to be the longest phase. Its understanding and optimization through right selection of shelf temperature and chamber pressure is central to the design of an efficient process. A right balance between the fast-drying rate and product elegance, and in some cases degradation (avoidance of collapse of the product is of essence) is important. The drying rate or the sublimation rate is governed by a simple equation, illustrated below, dm P0 - Pc = Rp þ Rs dt Where dm/dt is the rate of change in mass, and is proportional to the driving force, the difference between the vapor pressure of ice at the temperature of the frozen product P0 and the chamber pressure Pc and inversely proportional to the mass transfer represented as sum of resistance due the product Rp and stopper Rs. The sublimation rate during primary drying depends on the product resistance Rp, which in turn depends on the crosssectional area of the product (i.e., internal diameter of the container used, Ap), nature of the product (solid content and nature of excipients), and the thickness of the dried product. Thus, the product resistance increases during primary drying due to the increase in the dried layer thickness (Fig. 11) [35, 42]. At each temperature, drying time is roughly proportional to the square of the fill depth, and drying a product at target temperatures below 40 °C with 2-cm fill depths becomes challenging. Thus, fast freeze drying requires both high target product temperature and a small fill depth.

14

F. Jameel 4

C a le ne m

Rp*10–5 (m s–1)

oo rR

3

2 LabLyo1(MTM) LabLyo1(fitted) LabLyo2(fitted) PilotLyo1(fitted) PilotLyo2(fitted)

1

0 0

0.1 0.2 0.3 Dried Cake Length (cm)

0.4

Fig. 11 Product resistance as a function of dried layer thickness

80 70 60 Glass transition temperature of product, Tg

Temperature, °C

50 40 30 20 10 0

Product temperature, T

–10 –20 –30 0

5

10

15

Secondary drying time, hr

Fig. 12 Variation of glass transition temperature and product temperature for moxalactam di-sodium during secondary drying. (Calculated from data Refs. [43, 44])

4.3

Secondary Drying

Similar to primary drying, the product needs to be dried below the collapse temperature to be able to dry with retention of cake structure. The collapse temperature during secondary drying is mostly the Tg of the product. Exposure of the product to elevated temperature of secondary drying immediately after the end of primary drying phase carries risk of structural collapse as there will be still 5–15% residual moisture depending upon formulation components. If the system is amorphous in nature, then the residual moisture content would be expected towards higher end otherwise it is expected to be towards lower end in case of crystalline system. Since water acts as a plasticizer and lowers the Tg, the Tg of the product at the end of primary drying will be quite low and it sharply improves with the removal of water as shown in Fig. 12, hence, a gradual ramping from primary drying to secondary drying shelf temperature at a ramp rate of 0.22 °C/min is recommended.

Overview of Freeze Drying

15

The concept of using low chamber pressures during primary drying is quite reasonable as the sublimation rate is directly proportional to P0 – Pc. However, it is reported that [44] the secondary drying is insensitive to chamber pressures in the range of 0–0.2 Torr, but either diffusion in the solid or evaporation at the solid–vapor boundary is the rate-limiting step to mass transfer process for drying an amorphous solid, suggesting that relatively high chamber pressures (0.1–0.2 Torr) will be useful. One of the main objectives of freeze drying is to enhance the storage stability of the products that has marginal stability in liquid state and design of secondary drying conditions are central to it. Stability of protein or non-protein product is dependent on the mobility as it supports reactivity and the mobility is dependent on the residual moisture of the product. Since the moisture affects the glass transition temperature (Tg) and mobility is related to Tg, it is imperative that the residual moisture is targeted in such a way that Tg value is significantly much higher than the temperature the product will be exposed to during transportation and storage. This targeted residual moisture value that may remain in the product without impacting product stability needs to be determined. Empirical moisture studies are performed by exposing the lyophilized product to saturated salt solutions corresponding to different relative humidities and creation of water sorption isotherm [for details on moisture studies, please refer to Chapter “Design of moisture studies for a lyophilized product”]. The optimal residual moisture level is quite specific and varies from product to product, usually it is less than 1% w/w for proteins and varies between 2% and 3% for vaccines. The optimization of secondary drying involves two process parameters: the temperature and time. The maximum temperature the product that can be exposed during secondary drying varies from formulation to formulation, and molecule stability as a function of the temperature. Once the optimal temperature is determined, the optimum secondary drying time can be determined by extracting samples from the freeze dryer at various time intervals using a “sample thief/extractor” without interrupting the freeze drying cycle and measuring the moisture content using either Karl Fischer titrimetry (KFT), thermal gravimetric analysis (TGA), or near-IR spectroscopy (NIR). In our experience we have found that drying at low temperatures for a longer time than at high temperatures for a shorter time provides better chances of uniformity of moisture across the cake and between the vials. Generally, for heat labile molecules and formulations secondary drying at 25–30 °C for 5–6 h is adequate. For some formulations such as mannitolbased where mannitol hemihydrates are formed during freezing and/or primary drying phase, secondary drying is used as an opportunity to desolvate the mannitol hemihydrates. Secondary drying at elevated temperatures of 50–55 °C for 2–3 h helps to eliminate mannitol hemihydrates [45]. One has to be cognizant of the fact that the moisture content increases during storage and is most often related to moisture release by the stopper. It is often observed after few months of storage at elevated temperatures, and this moisture exchange between stopper and product may have important stability impact. The moisture transfer from stoppers can be addressed through extensive high temperature vacuum drying of the stoppers after steam sterilization.

4.4

Process Control

Primary drying is controlled through the control of the critical process parameters, the product temperature and the sublimation rate which are dependent on the independent process variables shelf temperature and chamber pressure. Product temperature is the balance between the heat transfer and mass transfer, and heat is transferred to the vial and to the product through conduction (shelf temperature), convection (through the collisions of gas molecules with the hot shelf surface and the cold vial bottom), and radiation. The mass transfer is again dependent upon the chamber pressure as the driving force for mass transfer is the difference between vapor pressure of ice and chamber pressure. Thus, the product temperature is determined by shelf temperature, chamber pressure, the heat transfer characteristics of the vials, and the mass transfer characteristics of the product and semi-stoppered vials. The product temperature needs to be monitored and controlled during the freeze drying as it is the critical process parameter that impacts the product quality attributes. Typically, during the development of the lyophilization process, thermocouples are utilized to determine the product temperature in order to optimize/adjust the input parameters shelf temperature and chamber pressure so that the product temperature is below the target product temperature to avoid collapse. During the validation of the process, thermocouples are also placed in the vials at various locations of the shelf and between the shelves to verify and demonstrate the uniformity of the product temperature and the drying rate. However, during commercial manufacturing due to sterility concerns it is not feasible to place thermocouples in the vials, and one has to rely on the robustness of the process and the validation data.

16

F. Jameel

Thus, vials containing thermocouples/sensors undergo less degree of undercooling causing nucleation of ice at a lower temperature, freeze faster than vials without temperature sensors [41] resulting in larger ice size crystals, leaving larger pores after sublimation, less resistance to mass transfer, freeze dry at lower temperature consequently less time to dry than the rest of the vials. Secondly, due to the variations in the heat transfer across the shelf and between the shelves, thermocouple data at the edges will be warmer than the center. Hence, vials with thermocouples are not considered representative of the whole batch. Due to the above cited limitations of using thermocouples to directly measure and monitor the product temperature and the process, indirect ways of measuring product temperature have emerged lately. Manometric temperature measurement (MTM) is one of the techniques that has been developed to determine the product temperature at the sublimation front in addition to determination of sublimation rate and end of primary drying phase. It is based on pressure rise data when the valve between the chamber and the condenser is swiftly closed for a brief period of 15 s. The pressure rise data is fitted to bunch of differential equations relating to heat and mass transfer to obtain the average product temperature at the ice–vapor interface. It gives a representative temperature of the product vials of the whole batch without the risk of sterility compromise. Its applicability has been proven, demonstrated, and in use at lab scale; however, its applicability at commercial scale is far from reach [46]. LyoPAT is another technology that is emerging again limited to lab studies where they use heat flux sensors to measure and control the heat transfer dynamics to be able measure Critical Process Parameters, mass flow, product temperature, cake resistance and additionally, vial thermal conductivity (Kv), and heat transfer parameters (heat flux). Another technique that is currently being used at lab scale and proving to be promising at commercial scale is the Tunable Diode Laser Spectroscopy (TDLAS) [47]. TDLAS-based sensor, LyoFlux, measures water vapor concentration and gas flow velocity in the duct connecting a freeze-dryer chamber and condenser. The near-IR spectrometer provides real-time measurements of water concentration and gas flow velocity that are used to determine the water mass flux (grams/second/ cm2) in the duct. The flux measurements are combined with the knowledge of the duct’s cross-sectional area to provide a determination of the water vapor flow (grams/second) exiting the product drying chamber. The flow measurements are integrated during the product drying cycle to provide a determination of the total water removed (grams). Through the combination of gravimetric determinations of mass flow and the well-accepted steady state model of heat and mass transfer of vial-based freeze drying, now it is feasible to determine the Product temperature at the sublimation interface, Product temperature at the bottom center of the vial, Product resistance to drying, Residual moisture content, Primary and secondary drying endpoints, Vial heat transfer coefficients, Continuous determination of the ice and dryer layer thickness, and the maximum lyophilizer equipment capability: mass flow as a function of pressure.

4.4.1 Chamber Pressure Control and Monitoring The chamber and the condenser pressure is controlled through the capacitance manometer, which is installed both in the chamber and pressure. The capacitance manometer determines the absolute pressure and controls the chamber pressure. The chamber pressure could be controlled in three ways: (1) nitrogen leak into the drying chamber through opening and closing of PTD valve, (2) conductance control, and (3) control of the condenser temperature. Controlled Nitrogen Leak The commonly used and preferred technique is the controlled nitrogen leak where the PTD valve is connected to a nitrogen source at atmospheric pressure and sterile nitrogen is leaked into the drying chamber in response to the deviation from the set point as measured by capacitance manometer. Conductance Control It is based upon opening and closing of the valve at the entrance of the duct connecting the drying chamber from the condenser [48]. It works fine during primary drying where the pressure could be controlled through the flow of the water vapors into the condenser; however, during secondary drying when the sublimation rate tapers off, and no water vapors, it will be unable to control the pressure. The chamber pressure will reduce to whatever ultimate vacuum the system will produce and potential for product contamination by adsorption of volatile stopper impurities or any other foreign vapors in the freeze dryer [49] and deprive one from carrying out secondary drying at a somewhat elevated level of chamber pressure of for example 200 mTorr. Control of Condenser Temperature The other option of controlling chamber pressure is through the control of condenser temperature, provided the lyophilizer offers the option of fine control of condenser temperature [49], where control of condenser temperature controls the vapor pressure of ice on the condenser, thereby controlling the partial pressure of water in the drying chamber.

Overview of Freeze Drying

17

A process where the chamber pressure is controlled through a nitrogen leak to provide pressure control is not necessarily the same as the corresponding process run with pressure control via control of condenser temperature, however, the difference is not expected to be of practical significance. One disadvantage of this process would be the inability to provide process control options for determining the end point of primary drying as it will not produce the change in gas composition as the process moves from primary drying to secondary drying.

4.4.2 Condenser Pressure Control and Monitoring Controlling the condenser pressure is critical to the control of chamber pressure as the driving force for mass transfer is the pressure gradient between chamber and condenser. The condenser temperature should be low enough, -50 °C, to allow control of the chamber pressure at the desired set point. It should be noted that an aggressive process with very high sublimation rate may overload the condenser causing a loss of chamber pressure control leading to high product temperature and ultimately loss of the batch due to product collapse or ice melt. This phenomenon is referred to as choke flow or when it hits Mach 1, i.e., when the velocity of water vapor hits the velocity of sound. An overloaded condenser may also be manifested by nonuniform buildup of the ice on the coils of the condenser due to the inability of the refrigerating system to remove heat from rapidly condensing water vapor (i.e., from very high sublimation rate) and yet maintain the condenser plate temperature low. 4.4.3 Determination of the End Point of Primary Drying One of the important determinations that needs to be made during primary drying as part of process control is when to stop the primary drying phase and advance to secondary drying. An inaccurate measurement carries a risk of melt back or collapse, as advancing to secondary drying with still ice left will depress the Tg of the product and collapses the product structure. Hence, some indicator of the end of primary drying is required for optimum process control. Traditionally, product temperature response and Pirani gauge are the most commonly used indicators of the end of primary drying and can be used at both lab scale and commercial scale. When product temperature is used as the indicator then the end of primary drying is determined when the coldest running vials (usually center or back side vials) product temperature approaches the shelf temperature. Since the fact that the vials containing temperature sensors run warmer and are not typical of the batch as a whole, a soak period of 10–15% is given for cold vials to catch up with warm vials. Since the determination of 10–15% is arbitrary and what is determined at lab scale may not translate and hold good for manufacturing due to freezing bias, using product temperature sensors to determine the end point of primary drying is far from being accurate. The vacuum data obtained from the Pirani gauge is widely used as a more reliable indicator at both development and manufacturing level. As it is based on the thermal conductivity of nitrogen gas which is 1.6 times less than the thermal conductivity of water vapors, the inflection point or when the value drops and equals to capacitance manometer indicates a change in the gas composition from water to nitrogen suggesting end of primary drying. Another sensitive method that can be used for measurement of vapor composition is through an electronic moisture sensor with output in dew point or partial pressure of water [41, 50]. An electronic moisture sensor has the sensitivity to determine the presence of residual ice in less than 1% of the vials [41].

5

Stability

Generally speaking stability increases in the following order: solution < glassy solid < crystalline solid [1–3]; this is likely due to restricted mobility in solids with the high degree of order in the crystalline solid retarding reactivity even further. With freeze drying it is often easy to achieve residual moisture contents in the cake ≤1%, but in some cases that may not be enough to attain a shelf life of ≥2 years at ambient temperature and may have to be refrigerated. In other cases either freezing or drying or both processes may inflict some damages to the drug active especially biological molecules that have fragile native conformation that could be perturbed due to removal of hydration layer needed to maintain the native conformation and may require stabilizers to protect them during freeze drying and/or upon storage. In-process stability and storage stability are often addressed through a combination of formulation and process optimization. Excipients or stabilizers that protect the drug active during freezing are called cryoprotectants and are often added to the formulation to protect the drug active against degradation or denaturation. If the drug active is sensitive to freeze drying process and exhibits instability during freeze drying and upon storage, stabilizers called lyoprotectants are often added to the formulation to protect the drug active against degradation or denaturation.

18

F. Jameel

Optimization of process conditions is equally critical to the stability of the product both during process as well as upon storage. Freeze drying should be carried out below the critical temperature called collapse temperature. It is the maximal allowable product temperature below which the product will dry with the retention of structure of the cake and drying above which will result in the collapse of the cake structure. The critical process parameters, shelf temperature and chamber pressure, should be selected and optimized in such a way that the resulting product temperature remains below the collapse temperature during drying, yielding a product that looks pharmaceutically elegant with short reconstitution time and desired low residual moisture content, while, from process efficiency point of view the freeze drying process should not be long. A collapsed product will not only cosmetically look inelegant but also will have repercussions on the integrity of the drug active and other product quality attributes, often accompanied by high residual moisture and reconstitution time. Formulation design without process considerations and process design without knowledge of manufacturing capabilities and limitations are the main causes of flawed/unsuccessful scale-up and technology transfers to manufacturing. In freeze drying, the formulation and process are interrelated. What is in the formulation dictates the process and vice versa, hence, selection of the components of the formulation and their characterization to understand their physical state and behavior during freezing and drying is key to efficient freeze drying process. The details of materials characterization and the techniques are covered or summarized in chapter “Concepts and Strategies in the Design of Formulation for Freeze Drying”.

References 1. Pikal MJ, Lukes AL, Lang JE. Thermal decomposition of amorphous β-lactam antibacterials. J Pharm Sci. 1977;66:1312. 2. Pikal MJ, Lukes AL, Lang JE, Gaines K. Quantitative crystallinity determinations for beta-lactam antibiotics by solution calorimetry: correlations with stability. J Pharm Sci. 1978;67:767. 3. Pikal MJ, Dellerman M. International stability testing of pharmaceuticals by high-sensitivity isothermal calorimetry at 25°C: cephalosporins in the solid and aqueous solution states. J Pharm. 1989;50:233–52. 4. Abdul-Fattah AM, Truong VL. Drying process methods for biopharmaceutical products: an overview. In: Jameel F, Hershenson S, editors. Formulation and process development strategies for manufacturing biopharmaceuticals. Wiley; First published: 26 July 2010. p. 705–38. 5. Searles J, Mohan G. Spray drying of biopharmaceuticals and vaccines. In: Jameel F, Hershenson S, editors. Formulation and process development strategies for manufacturing biopharmaceuticals. Wiley; First published: 26 July 2010. p. 705–38. 6. Pikal MJ. Lyophilization. In: Swarbrick J, Boylan J, editors. Encyclopedia of pharmaceutical technology. New York: Marcel Dekker; 2002. p. 1299–326. 7. Trappler E. Validation of lyophilization: equipment and process. In: Costantino HR, Pikal MJ, editors. Lyophilization of biopharmaceuticals. Arlington: AAPS press; 2004. p. 43. 8. Chang BS, Kendrick BS, Carpenter JF. Surface-induced denaturation of proteins during freezing and its inhibition by surfactants. J Pharm Sci. 1996;85:1325–30. 9. Mumenthaler H, Hsu C, Pearlman R. Feasibility study on spray-drying protein pharmaceuticals; recombinant growth harmone and tissue-type plasminogen activator. Pharm Res. 1994;11:12–20. 10. Barn NB, Cleland JL, Yang J, Manning MC, Carpenter JF, Kelley RF, Randolph TW. Tween protects recombinant human growth hormone against agitation-induced damage via hydrophobic interactions. J Pharm Sci. 1998;87:1554–9. 11. Barn NB, Randolph TW, Cleland JL. Stability of protein formulations: investigation of surfactant effects by a novel EPR spectroscopic technique. Pharm Res. 1995;12:2–11. 12. Kerwin BA, Heller MC, Levin SH, Randolph TW. Effects of Tween 80 and sucrose on acute short-term stability and long-term storage at-20 degrees of a recombinant haemoglobin. J Pharm Sci. 1998;87:1062–8. 13. Townsend MW, DeLuca PP. Use of lyoprotectants in the freeze-drying of a model protein, ribonuclease A. J Parenter Sci Technol. 1988;42:190– 9. 14. Hancock BC, Zografi G. The relationship between the glass transition temperature and the water content of amorphous pharmaceutical solids. Pharm Res. 1994;11:471–7. 15. Franks F. Protein destabilization at low temperatures. Adv Protein Chem. 1995;46:105–39. 16. Franks F. Freeze-drying: from empiricism to predictability. The significance of glass transitions. Dev Biol Stand. 1992;74:9–18; discussion 19 17. Bradley R. Plotting freezing curves for frozen desserts. Dairy Record. 1984;85:114–5. 18. Jameel F, Searles J. Development and optimization of the freeze-drying processes. In: Formulation and process development strategies for manufacturing biopharmaceuticals. Hoboken: Wiley; 2010. 19. Pikal MJ. Freeze-drying of proteins. Part II: Formulation selection. BioPharm. 1990;3(9):26–30. 20. Gomez G, Pikal MJ, Rodriguez-Hornedo N. Effect of initial buffer composition on pH changes during far-from-equilibrium freezing of sodium phosphate buffer solutions. Pharm Res. 2001;18(1):90–7. 21. Gabellieri E, Strambini GB. Perturbation of protein tertiary structure in frozen solutions revealed by 1-anilino-8-naphthalene sulfonate fluorescence. Biophys J. 2003;85(5):3214–20. 22. Gabellieri E, Strambini GB. ANS fluorescence detects widespread perturbations of protein tertiary structure in ice. Biophys J. 2006;90(9): 3239–45. 23. Privalov PL. Cold denaturation of protein. Crit Rev Biochem Mol Biol. 1990;25(4):281–306. 24. Becktel WJ, Schellman JA. Protein stability curves. Biopolymers. 1987;26:1859–77. https://doi.org/10.1002/bip.360261104. 25. Carpenter JF, Crowe JH. The mechanism of cryoprotection of proteins by solutes. Cryobiology. 1988;25(3):244–55.

Overview of Freeze Drying

19

26. Lee JC, Timasheff SN. The stabilization of proteins by sucrose. J Biol Chem. 1981;256(14):7193–201. 27. Arakawa T, Timasheff SN. Stabilization of protein structure by sugars. Biochemistry. 1982;21(25):6536–44. 28. Arakawa T, Timasheff SN. Preferential interactions of proteins with salts in concentrated solutions. Biochemistry. 1982;21(25):6545–52. 29. Arakawa T, Timasheff SN. Mechanism of poly(ethylene glycol) interaction with proteins. Biochemistry. 1985;24(24):6756–62. 30. Arakawa T, Timasheff SN. Protein stabilization and destabilization by guanidinium salts. Biochemistry. 1984;23(25):5924–9. 31. Arakawa T, Kita Y, Carpenter JF. Protein – solvent interactions in pharmaceutical formulations. Pharm Res. 1991;8(3):285–91. 32. Arakawa T, et al. Factors affecting short-term and long-term stabilities of proteins. Adv Drug Deliv Rev. 2001;46(1–3):307–26. 33. Carpenter JF, Izutsu K, Randolph TW. Freezing and drying-induced perturbations of protein structure and mechanisms of protein protection by stabilizing additives. In: Rey L, May JC, editors. Freeze-drying/lyophilization of pharmaceutical and biological products, vol. 96. New York: Marcel Dekker, Inc; 1999. 34. Carpenter JF, et al. Rational design of stable lyophilized protein formulations: theory and practice. In: Carpenter JF, Manning MC, editors. Rational design of stable protein formulations, Pharmaceutical biotechnology, vol. 13. Springer; 2002. 35. Pikal MJ, Shah S, Senior D, Lang JE. Physical chemistry of freeze drying: measurement of sublimation rates for frozen aqueous solutions by a micro balance technique. J Pharm Sci. 1983;72:635–50. 36. Cao E, et al. Effect of freezing and thawing rates on denaturation of proteins in aqueous solutions. Biotechnol Bioeng. 2003;82(6):684–90. 37. Pikal MJ. Mechanisms of protein stabilization during freeze drying and storage: the relative importance of thermodynamic stabilization and glassy state relaxation dynamics. In: Rey L, May J, editors. Freeze drying/lyophilization of pharmaceutical and biological products. Marcel Dekker, Inc; 1999. 38. Chang BS, Kendrick BS, Carptenter JF. Surface-induced denaturation of proteins during freezing and its inhibition by surfactants. J Pharm Sci. 1996;85:1325–30. 39. Kim AI, Akers MJ, Nail SL. The physical state of mannitol after freeze-drying: effects of mannitol concentration, freezing rate, and a noncrystallizing cosolute. J Pharm Sci. 1998;87(8):931–5. 40. Cavatur RK, Suryanarayanan R. Characterization of phase transitions during freeze-drying by in situ X-ray powder diffractometry. Pharm Dev Technol. 1998;3(4):579–86. 41. Roy ML, Pikal MJ. Process control in freeze drying: determination of the end point of sublimation drying by an electronic moisture sensor. J Parenter Sci Technol. 1989;43:60–6. 42. Pikal MJ, Roy ML, Shah S. Mass and heat transfer in vial freeze drying of pharmaceuticals: role of the vial. J Pharm Sci. 1984;73:1224–37. 43. Pikal MJ, Shah S. The collapse temperature in freeze drying: dependence on measurement methodology and rate of water removal from the glassy phase. Int J Pharm. 1990;62:165–86. 44. Pikal MJ, Shah S, Roy ML, Putman R. The secondary drying stage of freeze drying: drying kinetics as a function of temperature and chamber pressure. Int J Pharm. 1990;60:203–17. 45. Jameel F, Bjornson E, Hu L, Kabingue K, Besman M, Pikal M. Effects of formulation and process variations on the stability of lyophilized recombinant human factor VIII. Haemophilia. 2000;6(4):343–68. 46. Milton N, Nail SL, Roy ML, Pikal MJ. Evaluation of manometric temperature measurement as a method of monitoring product temperature during lyophilization. PDA J Pharm Sci Technol. 1997;51:7–16. 47. Sharma P, Kessler WJ, Bogner R, Thakur M, Pikal MJ. Applications of the tunable diode laser absorption spectroscopy: in-process estimation of primary drying heterogeneity and product temperature during lyophilization. J Pharm Sci. 2019;108:416–30. 48. Kobayashi M. Development of new refrigeration system and optimum geometry of the vapor condenser for pharmaceutical freeze dryers proceedings of the 4th International drying symposium, Kyoto, Japan, July 9–12, 1984; Toei RR, Mujumdar A, editors; 2, 464–71. 49. Pikal MJ, Lang JE. Rubber closures as a source of haze in freeze dried parenterals: test methodology for closure evaluation. J Parenter Drug Assoc. 1978;32:162. 50. Bardat A, Biguet J, Chatenet J, Courteille F. Moisture measurement: a new method for monitoring freeze drying cycles. J Parenter Sci Technol. 1993;47:293–9.

Characterization and Determination of Freeze-Drying Properties of Frozen Formulations: Case Studies Feroz Jameel

Abstract

It is quite critical prior to design of the lyophilization process to completely characterize the formulation and understand the thermal events and physical state changes that occurs as a function of cooling and warming. Definitions of various freeze drying properties are described and illustrated with examples/figures. Various techniques such as MDSC and FDM that can be used to characterize under various conditions and determine the freeze drying properties/values of the formulation along with the science behind the design of the experiments, analyze and interpret the data is described and Illustrated with case studies. Finally, how some freeze drying properties can be predicted theoretically using Gordon & Taylor equations is discussed with example. Keywords

Lyophilization · Freeze drying properties · Frozen formulation · Modulated DSC · Freeze drying microscopy · Collapse temperature · Eutectic melting temperature · Glass transition temperature of freeze concentrate · Melt-back · Crystallization · Amorphous state

1

Introduction

It is quite common to see lyophilized products in the market that the freeze drying of which takes several days to weeks and yet produce a product that lacks pharmaceutical elegance, reconstitutes with difficulty with high reconstitution time, and lot-to-lot and vial-to-vial variability/heterogeneity in quality attributes; and most important is failure to achieve ambient storage stability and requires cold chain for shipment and storage. These issues arise from the lack of clear understanding of thermal behavior and freeze drying properties of the excipients, their selection, and their weight ratios while designing the formulation and process. Additionally, it is important to consider during the design of formulation and process that the commercial manufacturing requires that the process should be short (i.e., economically viable), operative within the capabilities of the equipment with appropriate safety margins and efficient plant utilization. Thus, the above expectations require the design of formulation and lyophilization cycles to be such that the collapse temperatures are maximized and drying rates are as high as possible and robust enough to be implementable on typical production freeze dryers [1]. As Freeze drying is a cold process, it is time-consuming process and accurate determination of freeze drying properties of the formulation such as eutectic melting temperature (Te), the glass transition temperature of the maximum freeze concentrated solution (Tg’), crystallization and annealing temperatures and times, residual unfrozen water content at the Tg’ temperature (Wg’) and the glass transition temperature of the dry powder (Tg) is not only the first step but also central to the design of the freeze drying process. The onset of crystallization and the annealing temperature guide the freezing parameters and the protocol, with Tg’ and/or Te values inform the selection of primary drying conditions (shelf temperature, chamber pressure, F. Jameel (✉) Nimble BioSolutions, Gurnee, IL, USA # America Association of Pharmaceutical Scientists 2023 F. Jameel (ed.), Principles and Practices of Lyophilization in Product Development and Manufacturing, AAPS Advances in the Pharmaceutical Sciences Series 59, https://doi.org/10.1007/978-3-031-12634-5_2

21

22

F. Jameel

and time). The Wg’ value also forms the basis for selection of secondary drying conditions, ramping rate advancing from primary to secondary drying, temperature and time of secondary drying. The Tg of the freeze dried cake informs the selection of accelerated stability studies conditions and forms the basis for the recommended transportation and storage conditions of the product [2]. The overall values of the abovementioned freeze drying properties for a given formulation depend heavily on the values of these properties for individual excipients/ingredients present in the formulation and their weight ratios in the mixture. Biopharmaceutical formulations typically end up in a multicomponent salt system in an effort to stabilize the labile biologic coupled with end-use requirements. Thus, it becomes critical to characterize the thermal behavior of these components both individually and in a mixture and determine the values of these properties before selecting the excipients/composition of the formulation and process conditions. Most of the excipients that are used in the biopharmaceutical formulations includes buffers, surfactants, stabilizers, bulking agents, and tonicity modifiers and they behave differently because of the variations in their concentrations, temperature (cooling and warming), and presence and absence of other excipients [3]. Formulation and process development are interdependent and the formulation candidate screening should consider the freeze drying properties of the prototypes besides the stability aspects. There are two techniques: the freeze drying microscopy (FDM) and modulated differential scanning calorimetry (MDSC), that are commonly employed to understand the thermal behavior/thermal events, physical state of the excipients individually and in mixture in the frozen state and determine the abovementioned freeze drying properties, see Tables 1 and 2.

Table 1 Collapse Temperature, Tc (°C) and Glass Transition Temperature, Tg’ (°C) data for selected excipients

Material BSA Dextran Ficol Gelatin PVP (40k) Dextrose Hydroxypropyl β cyclodextrin Lactose Mannitol Raffinose Sorbitol Sucrose Trehalose β Alanine Glycine Histidine Acetate, potassium Acetate, sodium CaCl2 Citric acid Citrate, potassium Citrate, sodium HEPES NaHCO3 Phosphate, KH2PO4 Phosphate K2HPO4 Phosphate, NaH2PO4 Tris base Tris HCl Tris acetate ZnCl2

Tg’(°C) -11 -10 -19 -9 -20 -44 -10.3 -28 -35 -27 -46 -33 -29.5 -65 -62 -33 -76 -64 -95 -54 -62 -41 -63 -52 -55 -65 -45 -51 -65 -54

Reference 6 6, 5 5 5 5 5 5,10 6, 6, 5 7 5 5 5, 7 6 8 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6

Tc (°C)

Reference

-10 -20 -8 -23

[16] [17] [17] [17]

-32

[7]

-26 -27 -32

[7] [7] [7]

Characterization and Determination of Freeze-Drying Properties of Frozen. . . Table 2 Glass Transition Temperatures, Tg, of selected excipients measured by DSC

Compound Citric acid Glycine Lactose Maltose Mannitol Raffinose Sorbitol Sucrose Trehalose Maltodextrin 860 PVP k90

23 Tg, °C 11 ~30 114 100 13 114 -1.6 75 118 169 176

Reference [9] [11, 12] [12–14] [12, 14, 15] [12, 14, 15] [12, 14, 15] [12, 14, 15] [12, 14, 15] [12, 14, 15] [12, 14, 15] [12, 14, 15]

Consult the references for details of the techniques, value in the parenthesis is extrapolated from mixtures using Fox equation and is highly approximate

Collapse temperature data were obtained with freeze drying microscopy and Tg’ data were obtained using DSC at roughly 10 °C/min heating rates and represent mid-points of the glass transition region. Values in parenthesis were estimated by extrapolation from non-crystallizing mixtures to the pure compound

2

Definition of Freeze-Drying Properties

2.1

Collapse Temperature

In a system where all the excipients form a single amorphous system post-freezing, the maximum allowable product temperature during primary drying without the loss of porous “cake-like” structure with the dimensions equivalent to those of the frozen solid [4, 18] is termed as collapse temperature. As illustrated in Fig. 1a there are various degrees of collapse, onset of collapse, partial collapse, and full collapse, depending upon how much the product temperature is away from collapse temperature. Drying a product above the collapse temperature results in the material becoming rubbery and is subject to viscous flow causing porous structure to collapse. Advancing to secondary drying without the full completion of sublimation of ice in the primary drying phase results in the melting of the remaining ice causing melt back. Figures 1b illustrates various degrees of melt back. Collapse temperature can be measured through direct microscopic observation of collapse/loss of structure during freeze drying using freeze drying microscopy. Pikal et al. [4, 18] investigated the differences between collapse temperatures determined by laboratory procedures and the observation of collapse in production processes. They observed that the collapse temperature increased as the sublimation rate increased (i.e., as the solute concentration decreases), and at constant sublimation rate, the collapse temperature may increase as the surface area of the solid increases. They also noted that in general, product freeze drying in a vial will collapse at a slightly higher temperature than collapse measured by the microscopic method. Tg’ is the glass transition temperature of the maximally freeze concentrated solute (Tg’), and collapse temperature and glass transition temperature, Tg’, are not identical. The collapse temperature when measured using DSC or MDSC will be 2–3 °C higher than the Tg’ value because the system will not undergo viscous flow, and loss of structure will not be observed until the product temperature exceeds the Tg’ value by a 2–3 °C when measured at low rates of temperature increase.

2.2

Eutectic Melting Temperature (Te)

A eutectic mixture is a physical mixture of two or more crystalline compounds that melt together at the same temperature as one compound and that temperature is referred as eutectic temperature. As illustrated in Fig. 2 there are various degrees of melt depending upon the composition of the solution and how far the product temperature is from the eutectic melting temperature. In a system where all the excipients crystallize upon freezing then Te would be the collapse temperature/critical temperature. In a binary system where both amorphous and crystalline systems present and if crystalline constitutes the major component then performing primary drying with the product temperature above Tg’ but below Te will dry the product with the collapse of the amorphous component on the surface of the crystalline phase, and the crystalline phase will render the

24

F. Jameel

Fig. 1 (a) Various degrees of collapse: (A) micro or onset of collapse, (B) partial collapse, (C) full collapse. (b) Various degrees of melt back. (Courtesy lyophilizationworld)

Fig. 2 Examples of various degrees of eutectic melt: (a) full eutectic melt, (b) partial eutectic melt, (c) elegant crystalline cake

necessary mechanical support to the cake structure [3]. On the other side if the amorphous phase constitutes the major component and crystalline phase the minor component, then, under those situations drying above Tg’ but below Te will be risky and will depend on how far the product temperature is from the Tg’ as the crystalline structure will not provide the sufficient mechanical support needed.

3

Characterization Techniques

The characterization techniques that are commonly employed to determine the freeze drying properties are (1) freeze-drying microscopy, (2) modulated differential scanning calorimetry (MDSC), and (3) electrical resistance.

Characterization and Determination of Freeze-Drying Properties of Frozen. . .

3.1

25

Differential Scanning Calorimetry (DSC)

This technique has been around since many years, and since thermal events involve heat flow, this technique has been successfully used to quantitatively measure changes in heat flow and heat capacity resulting from first-order irreversible/ kinetic thermal events such as crystallization and eutectic melt (exotherms or endotherms), and from second-order reversible events such as glass transitions a function of time and temperature. Principle The underlying principle behind the standard or conventional DSC is based on the Ohm’s law that governs the heat flow, Eq. 1 dQ ΔT = dt RD

ð1Þ

Where dQ dt is the heat flow and ΔT is the temperature difference between reference and sample, and RD is the thermal resistance of the constantan disc. An aluminum pan containing the sample and the empty pan as a reference are placed on the discs made up of constantan through which the heat transfers from the sample and the reference to the thermocouples made up of CHROMEL that are placed beneath the discs. These thermocouples measure the differential heat flow. Although in the standard DSC the chamber is purged with nitrogen or argonon gas through an orifice in the heating block to ensure uniform and stable thermal environment resulting in a controlled flat baseline, sometimes a crooked baseline can be observed due to some impurities on the discs which makes some real transitions questionable. To confirm whether these transitions are real in nature it is often repeated. To overcome such difficulties and resolve some complex transitions Mike Reading came up with an idea of modulated differential scanning calorimetry (MDSC) where a sinusoidal modulation is imposed on the conventional underlying linear heating or cooling ramp, thus the guiding principle of MDSC is to apply two heating rates simultaneously and measure how they affect the rate of heat flow [5, 19]. As a result of which the sample experiences two heating rates or cooling profiles or as if two experiments, one conventional linear (average) heating rate [dashed line in Fig. 3] and the other at a sinusoidal (instantaneous) heating rate [dashed-dot line in Fig. 3], are being run simultaneously. The average heating rate provides total heat flow information while the sinusoidal heating rate provides heat capacity information from the heat flow that responds to the rate of temperature change. The advantage of doing this is that it will enable separation of the reversible and irreversible heat flows from the total heat flow as indicated in the following Eq. 2. dQ Cp ΔT þ f ðT, t Þ = t dt

ð2Þ

dQ The standard or conventional DSC measures only the total heat flow (dQ dt ), whereas MDSC measures total heat flow ( dt ), reversible heat flow which is a function of the sample’s heat capacity and rate of temperature change, and irreversible heat

Fig. 3 Heating rate profiles in MDSC. (Adapted from Refs. [6, 20])

26

F. Jameel

Fig. 4 Depicting three heat flows for a quenched material. (Adapted from Ref. [20])

flow which is a function of absolute temperature and time. Reversible heat flow and heat capacity are associated with glass transitions which are reversible in nature while the crystallization melts and heat of enthalpy are associated with irreversible heat flow and kinetic in nature, see Fig. 4. Advantages of MDSC over standard DSC: 1. Ability to resolve complex transitions such as enthalpy of relaxation from glass transition temperature which appear together and can make the glass transition appear to be a melting transition. Similarly, crystallization of an excipient prior to or during melting make it difficult to determine the real crystallinity of the sample 2. Ability to detect weak transitions which are often overshadowed by drifting of baseline in standard DSC 3. One of the disadvantages of standard DSC is that some measurements such as heat capacity and thermal conductivity require more than a single experiment 4. Conventional DSC results are always a compromise between sensitivity and resolution. Resolution or separating transitions that are only few degrees apart require the use of small samples and low heating rates but the size of the heat flow signal decreases with reduced sample size and heating rate. This suggests that any improvement in resolution comes with reduction in sensitivity and vice versa. The sensitivity of the weak transitions can be improved by (1) increasing the sample size, which enhances the amplitude of the transition; (2) increasing the scanning rates; and (3) increasing the concentration but keeping the mass ratio of the components in the mixture constant. The downside with these increasing sample size and scanning rates are that they move the transitions to higher temperatures and can also cause merging of the nearby thermal events. The dependence of glass transitions on heating rates needs to be corrected as heating rates employed in freeze drying are much lower and the real values will be lower by a few degrees, especially in cases where drying is carried out at product temperatures close to Tg’. Further resolution and sensitivity in the determination of Tg’ can be obtained using the derivative of the power–time curve [21, 22].

3.1.1 Method Design Considerations [23] As indicated above, in MDSC as opposed to standard DSC the sample is subjected to two heating rates simultaneously to obtain all the information relating to thermal events, and optimization of the values of these heating rates is key to obtaining quality data. Prior to selecting and optimizing the values, three key parameters, average heating rate, temperature modulation period and amplitude, should be considered: average heating rate, temperature modulation period, and amplitude. 1. Keep the average heating rate (°C/min) slow enough to get sufficient number of modulation cycles over transitions of interest. 2. The modulation period should be selected in such a way that there is enough time for the heat to flow between the sensor and the sample 3. Modulation amplitude (± °C) affects both the sensitivity and the resolution. Keep the amplitude larger to achieve greater sensitivity but not so large that it will reduce the resolution

Characterization and Determination of Freeze-Drying Properties of Frozen. . .

27

Selection of Modulation Temperature Period The temperature modulation period is the time, in seconds, required to complete one modulation cycle. The selection of the period varies between 10 and 200 s and depends on the sample weight, its thickness, and conductivity in addition to sample pan and objective of the experiment. The following are the recommendations or guidelines that can be used as starting points and can be further optimized if needed. One key point that needs to be considered during the selection of period is that it should not be longer than necessary for quantitative heat flow, otherwise it will require a reduction in the average heating rate. • A period of 45 s is recommended for samples containing 15 mg in crimped aluminum pans and 60 s in samples up to 15 mg contained in hermetic aluminum pans. • For heat capacity measurement of materials 100 s is recommended • For large volumes in stainless steel hermetic pans a period of 200 s is recommended

Selection of Modulation Temperature Amplitude Similar to standard DSC, MDSC also provides increased sensitivity for transitions involving change in heat capacity with the increased modulation temperature amplitude due to larger changes in the heating rate. The instrument allows to select amplitude values from ±0.001 °C to 10 °C; however, generally amplitudes greater than ±2.0 °C have been found to decrease the resolution and amplitudes less than ±0.1 °C are not recommended for use because they give poor sensitivity. Amplitudes can be selected in a way to provide cooling or not to provide cooling. For determination of glass transition (Tg) temperatures which are function of changes in heat capacity both heating and cooling during the temperature modulation are found to be best. For example, in Fig. 5, where the average heating rate is 1.0 °C/min with a modulation period of 60 s and modulation amplitude of ±0.319 °C this condition provides modulated temperature to increase and decrease as the average temperature increases. The time-based derivative of this signal indicates that the average heating rate is 1.0 °C/min, while the modulated heating rate ranges from approximately -1.0 to 3.0 °C/min. This overall range of 4.0 °C/min is the result of the selected period and amplitude. Such conditions that provide both heating and cooling during the temperature modulation are found to be best for measurement of glass transitions temperatures. 8

56 Period: 60 seconds Amplitude: 0.319 (Heat-Cool) Heating Rate: 1°C/min

Measured Temperature 6

Modulated Temp (°C)

52

50 Modulated Heating Rate (Heat-Cool Conditions)

4

48

46

2

44

[——— ·] Deriv. Modulated Temp (°C/min)

54

0

42

40 70

72

74

76 Time (min)

78

80 Universal V2.7C TA Instruments

Fig. 5 The modulated temperature is seen to increase and decrease as the average temperature increases. The time-based derivative of this signal shows that the average heating rate is 1.0 °C/min, while the modulated heating rate ranges from approximately -1.0 to 3.0 °C/min [23]

28

F. Jameel 95

90

Period: 60 seconds Amplitude: 0.319 (Heat-Cool) Heating Rate: 2°C/min

Measured Temperature 6

Modulated Temp (°C)

85

Modulated Heating Rate (Heat-Cool Conditions)

80

4

75

70 2 65

[——— ·] Deriv. Modulated Temp (°C/min)

8

60 0 55 70

72

74

76

78

Time (min)

80 Universal V2.7C TA Instruments

Fig. 6 Results from selection of conditions involving the same modulation period (60 s) and same modulation amplitude (± 0.319 °C) as for that shown in Fig. 3. The difference is that the average heating rate has been increased to 2 °C/min [23]

Table 3 This table is additive, i.e., the heat only amplitude for a period of 40 s and heating rate of 2.5 °C/min is sum of the values for 2.0 °C/min and 0.5 °C/min Period (s) Heating rate 0.1 0.2 0.5 1.0 2.0 5.0

40 0.011 0.021 0.053 0.106 0.212 0.531

50 0.013 0.027 0.066 0.133 0.265 0.663

60 0.016 0.032 0.080 0.159 0.318 0.796

70 0.019 0.037 0.093 0.186 0.371 0.928

80 0.021 0.042 0.106 0.212 0.424 1.061

90 0.024 0.048 0.119 0.239 0.477 1.194

100 0.027 0.053 0.133 0.265 0.531 1.326

Amplitude (40s, 2.5 °C/min) = 0.212 ± 0.053 = ±0.265 °C, [23]

For determination of crystallization temperatures, the “heat-iso” conditions would be ideal where the heating rate is designed to go to zero (isothermal) with no cooling. This is well illustrated in Fig. 6 where the modulation period (60 s) and modulation amplitude (± 0.319 °C) are the same as for that shown in Fig. 5, but the average heating rate has been increased to 2 °C/min. As can be noticed these conditions still provide a range of 4 °C/min involved (0–4 °C/min), but now there is no cooling during the modulation. Based on experience the TA instruments has come up with a table, Table 3, that eases the selection of the modulation temperature amplitude, following is the guidelines 1. First select the modulated period and average heating rate that will be used for the experiment 2. To run experiment under heat-iso conditions for a given period and heating rate, the corresponding amplitude values can be taken directly from Table 3 3. For heat-cool conditions any amplitude value greater than indicated will produce cooling during modulation. The amplitude values should be increased by factors as suggested below as starting conditions and can be further optimized based on the thermal event of interest

Characterization and Determination of Freeze-Drying Properties of Frozen. . .

29

Glass Transitions (Tg) • For standard Tg”: Sample size: 10–15 mg Period: 40 s

Amplitude: 2× Table 1 Heating Rate: 3 °C/min

• Tg is Hard to Detect Sample size: 10–20 mg Period: 40 s

Amplitude: 4× Table 1 Heating Rate: 2 °C/min

• Tg has Large Enthalpic Relaxation Sample size: 5–10 mg Period: 40 s

Amplitude: 1.5× Table 1 Heating Rate: 1 °C/min

In case of standard Tg determinations with sample sizes in the range of 10–15 mg, use heating rate of 3 °C/min, a period of 40 s and use amplitudes 2× the value indicated in the Table 3 In cases where the Tg values are difficult to determine as shown in Fig. 6, sample sizes in the weight range of 10–20 mg, use heating rate of 2 °C/min, period of 60 s, and an amplitude of 4× the value indicated in the Table 3. In cases where the Tg are associated with large enthalpic relaxations use sample sizes in the range of 5–10 mg, slower heating rates of 1 °C/min, period of 40 s, and an amplitude of 1.5× the value indicated in the Table 3. In general, for most experiments, a starting weight of 10 mg is advised, and it can be increased for improved sensitivity or decreased for improved resolution. Selection of Average Heating Rate The average heating rate should be selected in such a way that at least 3–4 modulation cycles should fit over the temperature range of the transition either glass transition or other thermal events involving a step-change in heat capacity in order to call the transition/thermal event under consideration with confidence and with no ambiguity. The number of modulation cycles should be determined between the extrapolated on-set and extrapolated end-set of the step and should be 3–4 cycles. For transitions involving peaks, the minimum is determined at half-height of the peak as illustrated in Fig. 7. At temperature near 250 °C, at half-height of the melting peak near 250 °C, there are roughly 6–7 cycles, which meets the requirement of four (4) or more. However, at half-height of the cold-crystallization peak near 150 °C there are only 3 cycles suggesting that the average heating rate is slightly too fast. It is to be noted that this experiment was performed at 4 °C/min and better results would be obtained if it is performed at 3 °C/min as all transitions involving peaks would have four or more cycles at half-height of the peak and meets the requirement.

3.2

Freeze Drying Microscopy (FDM)

Freeze drying microscopy is another technique that enables direct observation of physical behavior of the components of the formulation during freezing and freeze drying under a polarized light microscope. It allows the determination of freeze-drying properties such as collapse temperature, crystallization temperature, optimal annealing temperature and time, and eutectic melting temperature. It involves placing a sample between the cover slips and cryostage chamber which is placed under the microscope, the whole set up acts like a micro freeze-dryer. The sample is then frozen down to a pre-selected temperature and cooling rate, after which a vacuum is pulled down in the chamber to begin the drying process. The structure behind the drying front is watched under the microscope to see whether it is drying with the retention of structure or not. If it is drying with the retention of the structure the temperature is raised slowly in increments until the structure starts breaking. The level of severeness of the breakage of the structure is defined in terms of onset, partial, or full collapse. The temperature of the stage can then be manipulated in order to identify the temperature at which the product will collapse or melt.

F. Jameel

30 24 MDSC Raw Data Signals: Note That All Transitions Are Visible In MHF Signal

Modulated Heat Flow (mW)

16

12

–4

Modulated Heating Rate

8

–8 4

Deriv. Modulated Temperature (°C/min)

20

Modulated Heat Flow 0

0 –12 50 Ex0 Up

100

150

200 Temperature (°C)

250

300 Universal V3.4C TA Instruments

Fig. 7 Demonstration of number of modulation cycles within the transitions dependence on the heating rate [23]

3.2.1 Equipment and Experimental Procedure It is composed of three major components: a (1) temperature-controlled freeze-drying stage (cryostage), (2) an optical window through which the progress of freezing or/and drying in sample can be observed via a microscope, and (3) a programmable temperature controller. Procedure Insert the sample holder on the Cryostage and align the XY manipulator so that the sample holder is in the middle of the silver block. Place the quartz crucible within the sample holder and to get a perfect thermal seal between Cryostage and the crucible a small drop of silicon oil on the Cryostage will be helpful. Load approximately 3–5 μL of sample in the center of the quartz crucible and cover it up through a 9 mm glass cover slip using vacuum tweezers. Using XY manipulators move the sample holder so that the edge of the sample sits across the aperture hole of the Cryostage which helps in locating the sample boundary once the analysis starts. Adjust the microscope and focus on edge of sample first with the 10× objective and then with the 20×. Open the software and enter the temperature profile for the experiment. Before loading and analyzing the product sample it is always a good idea to run some standards such as potassium chloride or sodium chloride to ensure the temperature sensors are calibrated and the findings are reliable.

3.2.2

Applications of Freeze-Drying Microscopy

3.2.2.1 Crystalline Systems In the case of crystalline systems where a crystallizable excipient is included in the formulation and the intended role of it is to provide crystalline matrix besides improving the overall collapse temperature, it is imperative to ensure its complete crystallization during the freezing phase. It is important to note that the determination of collapse temperature of the crystalline system should be made only after complete annealing of the crystallizable excipient (example, Mannitol). The collapse temperature will be the eutectic melting temperature if the system is predominantly crystalline in nature. In order to achieve complete crystallization of the crystallizable excipient, the temperature and time at which the kinetics of maximum crystallization occurs be identified, and FDM helps to accomplish that.

Characterization and Determination of Freeze-Drying Properties of Frozen. . .

31

Ice crystallization Cooling Annealing scaffold Te (onset) Eutectic melting

Tg’

Mannitol crystallization Exo 

Fig. 8 An illustration of determination of thermal events of a crystalline system on MDSC

The experimental procedure should involve first freezing the sample below the Tg’ of the formulation, at least 6–10 °C below the Tg’, to create nuclei, prior to annealing. To pinpoint the optimal temperature, bring the temperature to approximately 10 °C above the Tg’ value to provide mobility and growth of nuclei into crystals and slowly raise the temperature in increments of 2 °C while holding for 15–20 mins at each temperature increment to observe rate of the crystallization growth. Once the annealing temperature is determined at which maximum growth occurs then determine the time needed for complete crystallization as evidenced by a higher collapse temperature in the frozen solution. This identified annealing time and temperature on FDM translates well in the vial and can further be optimized. An example of the cooling and heating profiles typically used for the determination of collapse temperature of a crystalline system is provided below 1. Ramp down to -45 °C at 10 °C/min 2. Hold at -45 °C for 5 min 3. Ramp up to -15 °C at 0.25 °C/min 4. Hold at -15 °C for 180 min (Annealing Step) 5. Ramp down to -45 °C at 10 °C/min 6. Hold at -45 °C 7. Apply Vacuum 8. Heat at 5 °C/min till -35 °C 9. Heat at 2 °C/min increments and hold for 5 mins at each temperature and observe for drying with loss of structure/when the structure starts breaking either partially called partial collapse and fully called full collapse/eutectic melt till 0 °C. Similar freezing protocol can be used on MDSC to obtain thermal behavior of mannitol or glycine-based formulations as schematically illustrated in Fig. 8. 3.2.2.2

Case Studies

Amorphous Systems Determination of collapse temperature of pure excipient such as sucrose or trehalose, which are the commonly used disaccharides in pharmaceutical/biopharmaceutical, is not difficult as their values are known in the literature and just requires careful attention of the drying front/sublimation front as it gets closer to the known literature values. It would serve as a good amorphous system standard for checking the temperature probe/sensor calibration prior to start of any experiment involving determination of collapse temperature of an amorphous system. Sometimes if the coverslip is not placed properly on the quartz crucible, the solution squeezes out of the edges and a very thin layer is left in the center and makes observation difficult. Additionally, the squeezed-out sucrose or trehalose dries/ crystallizes at the edges making the vacuum penetrate the coverslip difficult and inhibits drying and will not see sublimation front. In Fig. 9a you can see clearly the start of breaking of the structure of sucrose around -31 °C and a full collapse at any temperatures above it, and in Fig. 9b you see for a 2% Raffinose an onset of collapse at -24 °C and full collapse at -21 °C.

32

F. Jameel

Fig. 9 (a) Example of 10% sucrose. (b) Freeze drying microscopy images of 2% Raffinose (A) showing onset of collapse at –24 °C (B) showing full collapse at -21 °C [18]

Characterization and Determination of Freeze-Drying Properties of Frozen. . .

Frozen layer

Dry layer

33

Frozen layer

Dry layer

Sublimation front Collapse Sublimation Front

Photograph of a Sucrose:Tris formulation depicting drying with retention of structure (no collapse) at -39°C and 100 mTorr

Photograph of a Sucrose:Tris formulation depicting drying with loss of structure (collapse) at -37°C and 100 mTorr

Fig. 10 Freeze drying microscopy of a formulation solution containing Sucrose:Tris [24]

Fig. 11 Thermograms of pure Trehalose and a formulation solution containing Tris, phosphate, PEG, and Arginine

Addition of buffer salts or other excipients affects positively or negatively to the overall collapse temperature and Tg’ of the formulation depending upon the nature of the excipient. For example, Fig. 10 below depicts the collapse temperature depression of sucrose to -37 °C by the addition or presence of Tris. Similarly, it can be seen on MDSC thermograms in Fig. 11 the depression of Tg’ value of pure trehalose from -30 °C to 48 °C by the addition of buffer salts (in this example the solution contains Tris, phosphate, PEG, and Arginine).

34

F. Jameel

Fig. 12 Freeze drying microscopy photographs depicting effect of protein concentration on the collapse temperature of (a) 20 mg/mL protein in trehalose formulation (b) 100 mg/mL protein in sucrose formulation

Table 4 Comparison of DSC and FDM results obtained for mAb formulated at various protein concentrations in a formulation buffer mAB concentration (mg/mL) 0 10 20 40 60 80 100

DSC Tg’ ± SD (°C) -32.6 ± 0.7 -31.4 ± 0.4 -30.5 ± 0.4 -28.7 ± 0.2 -26.9 ± 0.1 -25.9 ± 0.2 -25.9 ± 0.3

Freeze drying microscopy Tc, on (°C) -33 -32 -29 -27 -27 -19 -21

Tc, com (°C) -31 -30 -24 -23 -17 -12 -13

Data from Ref. [26]

However, there are excipients such as cyclodextrin and protein the addition of which will positively improve the overall collapse of the formulation. There are some minimum concentrations required to start seeing positive effects. For example, inclusion of protein will show positive effect only above 20 mg/mL as shown below [25, 26]. This is illustrated in Fig. 12 and Table 4. Additionally, a continuous transition from onset to full collapse over a few degrees in temperature instead of a distinct collapse behavior is generally observed, mostly due to the high protein concentration. The temperature at which initial changes in the structure are observed is referred to as Tc, on (onset of collapse or partial collapse) while when complete loss of structure (collapse) is observed it is referred to as Tc, com. It is interesting and important to note from Fig. 11 and Table 4 that not only the difference or width between the onset temperature and end of either collapse or end of glass transition increases quite significantly with increase in protein

Characterization and Determination of Freeze-Drying Properties of Frozen. . .

35

concentration but also the difference between Tg’ value and the collapse temperature increases. In some cases, especially high protein concentration formulations, the viscous flow is not observed until 13–14 °C away from onset temperature. Hence, it is always prudent/recommended to use FDM to determine the upper temperature limit for a given formulation during primary drying rather than just relying on DSC data. Additionally, for high protein concentration formulations it is possible, and advantage can be taken to carry out primary drying above the Tg’ and at or slightly above the Tc, on without evidence of lack of elegance and stability. This is based on the hypothesis that at high protein concentrations the solids are thickly/densely populated, thereby increasing the viscosity and restricting the mobility resulting in the prevention of macroscopic collapse under the time scale of freeze drying [26]. 3.2.2.3 Crystalline Systems As indicated above it is always prudent to calibrate the systems with a standard, and sodium chloride and potassium chloride serve as good standards for determination of eutectic melting temperatures as their eutectic melting temperatures are known to be -21 °C and -11 °C, respectively. They are isotropic in nature, has cubic crystalline lattice structure, where all of the sodium or potassium and chloride ions are arranged with uniform spacing along three mutually perpendicular axes which means you will not see birefringence as you can see with the water or mannitol or glycine or calcite as they are anisotropic. In a formulation, if all the excipients are crystallizable and crystallize upon freezing, then the collapse temperature would be the eutectic melting temperature (Te). Since Te are much higher than the Tc, advantage is taken by creating formulations with the crystalline component being a major component that would enable to perform primary dying with the product temperature above Tg’ but below Te resulting in a significantly shorter lyophilization process. In this situation the product will dry with the collapse of the amorphous component on the surface of the crystalline phase, and the crystalline phase will render the necessary mechanical support to the cake structure, but the impact on the product stability must evaluated [1, 3]. On the other hand, if the amorphous phase constitutes the major component and crystalline phase the minor component, then, under those situations Te would not be the collapse temperature. The crystallizable excipient will not crystallize as the predominant amorphous component will inhibit its crystallization process and the crystallizable excipient will remain mostly amorphous and contribute to the overall collapse temperature of the formulation. Pure mannitol has a Te in the range of -3 °C to -5 °C. When a sugar like sucrose is added, depending upon the weight ratio, the crystallization of mannitol is inhibited and will remain amorphous and one has to determine the collapse temperature using FDM. Annealing or thermal treatment of mannitol will impact the collapse temperature. For example, in a formulation containing mannitol, sucrose, and protein where mannitol was the predominant component, the collapse temperature was determined to be -20 °C to -23 °C without annealing. However, optimization of annealing conditions (time & temp) led to identification of optimal annealing conditions at which the collapse temperature improved from -23 °C to -9 °C as depicted in Figs. 13 and 14, and Table 5.

Fig. 13 Freeze drying microscopy of an mAb solution formulated in mannitol and sucrose

36

F. Jameel -5

-10 -20 -15.1

-15 -14

-12 -10 -13.3-12.9

Tg' (deg C)

-15

-7 -18.9

-20

-30 -23.5

-25 -21.7

-5 -21

-25

-30 -40 -35.5

-35

-40 -45

-40

-35

-30

-25

-20

-15

-10

-5

0

Annealing Temperature (deg C) Fig. 14 Optimization of annealing temperature and time to improve the Tg’ value of an mAb solution formulated in mannitol and sucrose

Table 5 Freeze drying microscopy and DSC results of optimization of annealing temperature and time for an mAB formulated in mannitol and sucrose

S. No. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Annealing temperature (°C) No annealing -40 -30 -25 -20 -15 -12 -10 -7 -5

Tg’ (°C) -34.9 -35.5 -23.5 -21.7 -15.1 -14.0 -13.3 -12.9 -18.9 -21.0

Partial collapse(°C) -23 -10 -9 -6 -6 -8 -

Effect of NaCl on Trehalose and Mannitol Freeze drying microscopy was performed on formulation containing trehalose and mannitol in the presence and absence of NaCl to compare the effect of the salt on critical temperatures of the freeze-drying process, especially on the collapse event [27]. This analysis showed a significant difference between the collapse temperature (Tcol) of these two conditions as expected. As shown in Fig. 15, Tcol was -24.1 °C for the sample without NaCl, and -35.4 °C for the sample with NaCl. Since salts exhibit low Tg’ values, even low NaCl concentrations (0.2% m/v or 34 mM) can significantly depress the Tg’ values [27]. When the product temperature exceeds the Tg’ value during the lyophilization process, the rigid glass softens to become a highly viscous rubbery material and collapses.

3.3

Estimation of Tg’

The glass transition temperature of a maximal freeze concentrate (Tg’) of a monophasic multicomponent system can be estimated from the Tg’ values of the individual components using the Fox equation [28, 29] as indicated in Eq. 3.

Characterization and Determination of Freeze-Drying Properties of Frozen. . .

37

Fig. 15 Images of freeze-drying microscopy of recombinant human factor IX formulation solution containing Trehalose and mannitol (a) with NaCl (b) and without NaCl. The temperature at which the collapse began to be observed was -35.4 and -24.1 °C, respectively. The annealing conditions used were -10 °C for 10 min. (Adapted from Ref. [27])

Fig. 16 Effect of glycine on the Tg’ of amorphous sucrose:glycine frozen systems. (Data adapted from Shalaev and Kanev [30]

W W 1 = 1 þ 2 Tg 1 Tg 2 Tg

ð3Þ

Where W1 is the weight fraction component 1 and Tg1 is the glass temperature of pure component 1. Since the Fox equation applies to multicomponent systems, the impact of a second solute component on Tg’ of a formulation can be estimated from the equation if Tgi in Eq. 3 is identified as Tg’ of aqueous component i, and Wi are weight fractions of solute relative to the total mass of solute. As illustrated in Fig. 16, the effect of glycine on Tg’ of aqueous sucrose systems [30] is in agreement with Eq. 3, where the Tg’ value of aqueous glycine is obtained by fitting the data to the Fox equation, as it is difficult to determine the value of Tg’ of glycine as it readily crystallizes upon freezing.

38

F. Jameel

References 1. Pikal M. Lyophilization. In: Swarbrick J, Boylan JC, editors. Encyclopedia of pharmaceutical technology, vol. 6. New York: Marcel Dekker; 2002. p. 1299–326. 2. Pikal MJ. Mechanisms of protein stabilization during freeze-drying and storage: the relative importance of thermodynamic stabilization and glassy state relaxation dynamics. In: Louis Rey L, May J, editors. Freeze-drying/lyophilization of pharmaceutical and biological products. New York: Marcel Dekker; 1999, Chapter 6. 3. Franks F. Freeze-drying: from empiricism to predictability. Cryo-Letters. 1990;11:93–110. 4. Pikal MJ, Shah S. The collapse temperature in freeze drying: dependence on measurement methodology and rate of water removal from the glassy phase. Int J Pharm. 1990;62:165–86. 5. Her LM, Nail SL. Measurement of glass transitions of the maximally freeze-concentrated solutes by differential scanning calorimetry. Pharm Res. 1994;11:54–9. 6. Chang BS, Randall CS. Use of subambient thermal analysis to optimize protein lyophilization. Cryobiology. 1992;29:632–56. 7. Franzé S, Selmin F, Samaritani E, Minghetti P, Cilurzo F. Lyophilization of liposomal formulations: still necessary, still challenging. Pharmaceutics. 2018;10(3):139. 8. Chongprasert S, Knopp SA, Nail SL. Characterization of frozen solutions of glycine. J Pharm Sci. 2001;90(11):1720–8. 9. Qun L, Zografi G. Properties of citric acid at the glass transition. J Pharm Sci. 1997;86(12):1374–8. 10. Siow C. Pharmaceutical application scientist, Roquette, M0930-05-34 – insights into the thermal properties and freeze-drying of Hydroxypropyl-beta-Cyclodextrin formulations. AAPS PharmSci. 2019:360. 11. Toru SUZUKI, Rikuo TAKAI, Tsuneo KOZIMA, Felix FRANKS. Effect of glycine on glass transition temperature of sucrose solution (Papers presented at the 39th Annual Meeting). Japan J Freez Dry. 1993;39:31–5. 12. Hancock BC, Zografi G. The relationship between glass transition temperature and the water content of amorphous pharmaceutical solids. Pharm Res. 1994;11:471–7. 13. Jouppila K, Roos YH. Glass transition and crystallization in milk powders. J Dairy Sci. 1994;77:2907–15. 14. Saleki-Gerhardt A. Role of water in the solid-state properties of crystalline and amorphous sugars, Ph.D. thesis, Univ. Wisconsin— Madison; 1993. 15. Saleki-Gerhardt A, Zografi G. Non-isothermal and isothermal crystallization of sucrose from the amorphous state. Pharm Res. 1994;11:1166– 73. 16. Pikal MJ, Shah S. unpublished data. Eli Lilly & Co; n.d. 17. Mackenzie AP. Basic principles of freeze drying for pharmaceuticals. Bull Parenter Drug Assoc. 1966;26:101–29. 18. Kett V. Development of freeze-dried formulations using thermal analysis and microscopy. Am Pharm Rev. 2010; 19. Reading M, Hourston DJ. Modulated-temperature differential scanning calorimetry. New York: Springer; 2006. 20. Modulated DSC® Basics; Calculation and Calibration of MDSC® Signals, Leonard C. Thomas TA Instruments, New Castle. 21. Slade L, Levine H. Glass transitions and water-food. Structure interactions. Crit Rev Food Sci Nutr. 1995;30:115–360. 22. Levine H, Slade L. Principles of cryostabilization technology from structure/property relationships of carbohydrate/water systems—a review. Cryo-Letters. 1988;9:21–63. 23. Thomas LC. Modulated DSC® paper #3, modulated DSC® basics; optimization of MDSC® experimental conditions, TA Instruments. 24. Baxter presentation. Courtesy of Gregory Sacha, Senior scientist, Baxter, Bloomington. 25. Liao X, Krishnamurthy R, Suryanarayanan R. Influence of the active pharmaceutical ingredient concentration on the physical state of mannitol—implications in freeze-drying. Pharm Res. 2005;22(11):1978–85. 26. Colandene JD, Maldonado LM, Creagh AT, Vrettos JS, Goad KG, Spitznagel TM. Lyophilization cycle development for a high-concentration monoclonal antibody formulation lacking a crystalline bulking agent. J Pharm Sci. 2007;96:1598–608. 27. Almeida AG, Pinto RCV, Smales CM, et al. Investigations into, and development of, a lyophilized and formulated recombinant human factor IX produced from CHO cells. Biotechnol Lett. 2017;39:1109–20. https://doi.org/10.1007/S10529-017-2353-Y, https://link.springer.com/article/10. 1007/s10529-017-2353-y 28. Fox TG. Influence of diluent and of copolymer composition on the glass temperature of a poly-mer system. Bull Am Phys Soc. 1956;1:123. 29. Gordon M, Taylor JS. Ideal co-polymers and the second order transitions of synthetic rubbers. J Appl Chem. 1952;2:493–500. 30. Shalaev EY, Kanev AN. Study of the solid-liquid state diagram of the water-glycine-sucrose system. Cryobiology. 1994;31:374–82.

Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals Dominik Heger, Ramprakash Govindarajan, Enxian Lu, Susan Ewing, Ashley Lay-Fortenbery, Xiaoda Yuan, Lukáš Veselý, Eric Munson, Larry Gatlin, Bruno Hancock, Raj Suryanarayanan, and Evgenyi Shalaev

Abstract

Stability of the majority of pharmaceuticals and biopharmaceuticals depends on acidity/basicity of the environment. In aqueous solutions, acidity/basicity is commonly expressed using proton activity scale, pH, while definition and experimental measurements of acid-base relationships in frozen and freeze-dried materials are less straightforward. The chapter starts with a brief summary of the current understanding of several critical aspects of pH and apparent acidity/basicity as related to freezing and freeze-drying, whereas the main part of the chapter is focused on areas which are underrepresented in the pharmaceutical literature, with both overview of the literature and previously unpublished data presented. In particular, the following topics are covered: (i) Hammett acidity function and pH-equivalent (pHeq); (ii) Factors which impact apparent solid-state acidity in lyophiles; (iii) Solid-state acidity and chemical instability of lyophiles; (iv) Freezing fundamentals: quasi-liquid layer, polarity of the freeze-concentrated solution, and the Workman-Reynolds potential. Potential directions for future studies in this field are also outlined.

D. Heger · L. Veselý Masaryk University, Brno, Czech Republic R. Govindarajan Department of Pharmaceutics, University of Minnesota, Minneapolis, MN, USA College of Pharmacy, The University of Iowa, Iowa City, IA, USA E. Lu Department of Pharmaceutics, University of Minnesota, Minneapolis, MN, USA Auson Pharmaceuticals Inc., Bridgewater, NJ, USA S. Ewing · L. Gatlin · B. Hancock Pharmaceutical R & D, Pfizer Inc, Groton, CT, USA A. Lay-Fortenbery College of Pharmacy, University of Kentucky, Lexington, KY, USA Merck & Co, Inc, Kenilworth, NJ, USA X. Yuan · E. Shalaev (✉) Pharmaceutical Sciences R & D, Abbvie, Irvine, CA, USA e-mail: [email protected] E. Munson College of Pharmacy, Purdue University, West Lafayette, IN, USA R. Suryanarayanan Department of Pharmaceutics, University of Minnesota, Minneapolis, MN, USA # America Association of Pharmaceutical Scientists 2023 F. Jameel (ed.), Principles and Practices of Lyophilization in Product Development and Manufacturing, AAPS Advances in the Pharmaceutical Sciences Series 59, https://doi.org/10.1007/978-3-031-12634-5_3

39

40

D. Heger et al.

Keywords

Hammett acidity function · Freezing · Freeze-drying · Acid-base relationships · Stability · pH indicators · Freezing potential

1

Introduction

Stability of the majority of active pharmaceutical ingredients depends on proton activity (pH). For example, reactions under specific acid- and base-catalysis, such as many hydrolysis reactions, commonly have a minimum in the reaction rate (i.e., maximal stability) near neutral pH [1]. In other cases, maximal stability is observed in the moderately acidic pH region [2]. The pH region of maximal stability for a particular drug molecule is usually selected as the target pH, with pH specification covering ±0.5 pH units around the target value. To maintain pH during manufacture, shelf life, and administration, many formulations include a buffer. There are multiple reviews which cover various aspects of acid-base relationships, pH, and buffers for pharmaceuticals and biopharmaceuticals. An excellent overview of the topic for parenteral dosage forms was provided by Flynn [3] (note although that there is a typographic error for pKa of citrate in that paper), whereas use of buffers in protein formulations is reviewed recently [4]. The latter paper [4] combines description of general properties of buffers (e.g., buffering capacity, pKa, and temperature dependence of pH for typical pharmaceutical buffers) with multiple examples of both stabilization and destabilization of various proteins by buffers. Buffering capacity in the target pH range and a potential for buffer catalysis are the main considerations in buffer selection. In cases of general buffer catalysis, both buffer type and concentration could have a significant impact on degradation rate. The buffering capacity requirement is usually satisfied by selecting a buffer with pKa within one unit of the target pH. For frozen and freeze-dried forms, there are additional constraints for selection of target pH and buffers. Based on multiple original studies and reviews [5–15], current understanding of the buffers and pH subject for freeze-drying can be summarized as follows: • Freeze-dried products introduce extra considerations for buffer selection, including buffer crystallization behavior, collapse temperature of the freeze concentrated solution, Tg (glass transition temperature) of the dried formulation, and the volatility of the buffer components under vacuum [5]. • Main mechanisms for pH changes during freezing include temperature dependence of buffer ionization constant, pKa (e.g., for tris and histidine buffers [7]) and crystallization of buffer components. Sodium phosphate is the most known example of crystallizable buffer, while crystallization was also reported for histidine [13] and succinate [14] buffers. Other buffers, e.g., acetate and tris, are usually regarded as non-crystallizing under representative freeze/thaw conditions. Sodium citrate buffer has also been considered to be non-crystallizing [5], although a different opinion was also expressed [7]. Initial pH and buffer concentration could impact buffer crystallization and pH changes during freezing, with higher concentration usually associated with a higher crystallization risk. • Crystallization of buffer components and corresponding pH changes can be reduced/prevented by other solutes present. Both amorphous (e.g., sucrose) and crystalline (e.g., mannitol) solutes can interfere with buffer crystallization [8]. Higher weight ratio solute/buffer is usually associated with more efficient crystallization inhibition. Another successful approach to mitigate the freeze-induced acidity changes was realized by mixing buffers having opposing freeze-induced acidification tendency, e.g., a combination of one of the Goods buffers (HEPES, MOPS, and MES) with sodium or potassium phosphates buffers [15, 16]. • Freeze-induced pH changes can destabilize proteins during freeze-thaw and freeze-drying/reconstitution, as reported, e.g., in [9–11]. In addition, the freeze-induced acidity changes could promote chemical reactions of small molecules, such as N-nitrosation reaction of dimethylamine [17], molecular iodine production from various precursors [18, 19], and reduction of Cr(VI) [20]. • Glass transition temperature, Tg, of amorphous buffers depends both on the buffer type and the counterion [12]. For example, Tg of lactic acid was reported to be -60 °C [21], while another carboxylic acid, tartaric acid, has a higher Tg of 16 °C [22]. An increase in the counterion concentration can also have a major impact on Tg, as observed for citrate buffer, with the Tg values of 11 °C, 69 °C, and 115 °C reported for citric acid, monosodium citrate, and di-sodium citrate, respectively [23]. Therefore, buffer could either plasticize or antiplasticize a formulation, depending on the buffer type and the ratio of acidic/basic species. • In the frozen solutions, the Tg’ event, which is commonly regarded to be the glass transition temperature of the maximally freeze-concentrated solution, depends on buffer acid/base ratio and therefore pH; for sodium citrate buffer, e.g., Tg’ increased with pH from -50 °C (pH 2.5) to -35 °C (pH 4) and then decreased to below -40 °C as pH exceeded 5.5 [24].

Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals

41

• Buffers containing acetate, carbonate, ammonia, and HCl are considered to be volatile [25], and components of these buffers may be removed (at least partially) during freeze-drying, resulting in pH shift as measured in solution after reconstitution. • Proteins can serve as buffering agents, if they are present in a sufficiently high concentration [26, 27]. In such cases, formulation and manufacturing process can be simplified by not including buffer in the formulation. In the present chapter, we focus on several areas which are underrepresented in the pharmaceutical literature, using both published studies and previously presented (but unpublished) data. In particular, the following topics are covered: • • • •

Section 2. Hammett acidity function and pHeq. Section 3. Factors which impact apparent solid-state acidity in lyophiles Section 4. Solid-state acidity and chemical instability of lyophiles. Section 5. Freezing fundamentals: quasi-liquid layer, freeze-concentration, polarity, and the Workman-Reynolds potential

Main observations are summarized in the Conclusion (Sect. 6), which also provides a brief outline for future studies in this area.

2

Hammett Acidity Function and pHeq

Thermodynamic proton activity (aH+) scale, –logaH+ = pH, is easily accessible for dilute aqueous solutions. In order to provide quantitative description of acid–base relationships in other systems, Hammett proposed an acidity function based on ionization of probe molecules (pH indicators) [28] (Eq. 1) F ðRÞd ε0p C H x = pKa þ log 10 d = pKa þ log 10 Cp F ðRÞp ε0d

! ð1Þ

where C is the indicator concentration, the subscripts “p” and “d” refer to the protonated and deprotonated indicator species, respectively, and Ka is the ionization constant of the indicator in dilute aqueous solutions. The subscript “x” corresponds to the charge of the basic form of the given probe molecule. The original Hammett indicators have zero charge in the basic form, with the corresponding Hammett acidity function expressed as H0 (x = 0). For sulfonephthalein pH indicators, which are usually used for pharmaceutical materials, x is typically equal to -2 (therefore H2-). To calculate the Hx, one would need to add a probe molecule to a system of interest and experimentally measure ratio of the concentrations of the protonated and deprotonated species of the probe molecule. The Hammett acidity function was initially introduced to express acidity of concentrated acid solutions, and was extended to other systems including aqueous solutions of strong bases [29], nonaqueous solutions [30], and solid surfaces [31]. Note that the Hammett acidity does not represent proton activity, and its numerical value can depend on the type of probe molecule used for measurements. To illustrate this point, Fig. 1 provides Hammett acidity functions as measured with two types of probe molecules, N,N′-diethyl-toluidine (H0) and p-nitrophenol (H-), in water/ethanol mixtures. For 80% ethanol, e.g., H0 is equal to approx. 8.5, while the value for H- is approx. 5.5 for the same solution. This graph illustrates two additional points. First, the apparent acidity depends strongly on the water concentration. Therefore, a major decrease in water content during freeze-concentration could lead to large changes in the acidity, even in the absence of buffer crystallization and temperature-related pKa changes. Second, Fig. 1 includes four acidity scales, pH, p(aHγ HCl), and two Hammett acidity scales, H0 and H-, which show different trends as water concentration decreases. To emphasize the latter point, Paul and Long wrote: “. . .there is no single, unique good measure of acidity. There are a variety of them, and any preference depends on such things as ease of measurement and ultimate application” [32]. First application of the Hammett acidity function concept in solid pharmaceutical systems goes back to the 1990s, where surface acidity of common pharmaceutical excipients was measured using common pH indicators and UV/vis diffuse reflectance spectroscopy (DRS) [34, 35]. In these publications, the surface acidity was expressed as pH-equivalent (pHeq), which is defined as the pH of the solution which has the same ionization of the indicator as in the solid sample. Applications of the common pH indicators and UV/vis diffuse reflectance spectroscopy (DRS) for frozen solutions and freeze-dried formulations were first reported in [23, 36], respectively. Note that pH indicators were used previously to evaluate freezeinduced pH changes based on visual observations of color changes during freezing [37]. Such visual (semi-quantitative) assessment of acidity changes can also be performed during freeze-drying using either individual pH indicators or their mixture.

42

D. Heger et al.

Fig. 1 Acidity functions in ethanol–water solutions with 4 mM triethanolammonium chloride and 2 mM triethanolamine. pH = -log aH; aH is proton activity. γCl is the activity coefficient of Cl-. H0 and H- refer to the Hammett acidity function with different indicators. H0 indicator: N, N′-diethyl-toluidine (yellow); H- indicator: p-nitrophenol (green). (Graph is prepared using data from Refs. [30, 33])

Fig. 2 (a) Visible spectra of bromophenol blue solutions (25 mg/mL) in citrate buffer at pH values ranging from 2.57 to 5.14. (b) Diffuse reflectance spectra of trehalose–citrate lyophiles containing bromophenol blue, obtained by freeze-drying solutions buffered to pH values ranging from 2.27 to 6.66. The signals were normalized to enable a direct comparison of the spectra. (Reproduced with permission from: Govindarajan et al. [38])

Determination of the Hammett acidity function for freeze-dried formulations was described in [38]. In this approach, sulfonephtalein dye molecule (pH indicator) is co-lyophilized with a formulation of interest, and the visible absorbance spectra of the freeze-dried sample is measured using diffuse reflectance spectroscopy. Examples of the visible absorption spectra for both pre-lyophilization solutions and resulting lyophiles are provided in Fig. 2. The solution spectra have two peaks corresponding to the deprotonated and protonated species of the indicator; the peak of the deprotonated indicator grows as pH increases at the expense of the protonated species. The same trend is observed in freeze-dried samples prepared from solutions with different pH. The DRS results are used to calculate the Hammett acidity function using Eq. 1, with the Cd/Cp value in Eq. 1 determined using Eq. 2

Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals 0 cd F ðRÞd εp = cp F ðRÞp ε0d

43

ð2Þ

where F(R) is the Kubelka-Munk function, and ?′p∕?′d is the ratio of the extinction coefficients of the two species of a pH indicator (protonated and deprotonated). When using pH indicators to measure pHeq or Hammett acidity function in frozen solutions and in lyophiles, the following points should be considered: • Any particular probe molecule would cover only a narrow acidity range, and several indicators with different pKa values may be needed to measure acid–base properties of different formulations. • The Hammett acidity is numerically identical to pHeq values, provided that the ratio of the extinction coefficients of protonated and unprotonated indicator species in the solid state is similar to that in solution [38]. pHeq/Hx values can depend on the probe molecule used. e.g., the H2- values of the same material differed by 0.1–0.5 units when measured using different sulfonephtalein indicators [38]. Even larger difference in H2- of up to 0.85 was observed between bromocresol green and chlorophenol red in another study [39]. • pKa of both the indicators and buffers can change as a function of water content, because of the change in the dielectric constant of the media. Water crystallization (ice formation) results in significant decrease of water concentration in the solution coexisting with ice crystals, with a corresponding decrease in polarity due to increase in solutes concentration. In sucrose-water system, for example, dielectric constant of the 10 wt% solution is 76.2 at 25 °C [40] while the maximally freeze-concentrated solution (sucrose concentration approximately 80 wt%) is estimated to have dielectric constant of approximately 51 at 25 °C (graphical extrapolation using data from [40]). Reduction in the polarity of the freezeconcentrated solution is discussed also in Sect. 5 below. The decrease in the dielectric constant may result in major pKa changes; e.g., the apparent pKa of acetic acid increases significantly with decrease in polarity, from 4.76 in water (dielectric constant 78.3) to 10.32 in ethanol (dielectric constant 24.3) [41]. Accordingly, the pH shift observed for citrate buffer [7] could be attributed to freeze-concentration effect rather than the buffer crystallization; the freeze-concentration could lead in a significant change in pKa of a COOH group resulting in freeze-induced pH change. Such media-related pKa changes are different between different functional groups; for example, the pKa of an amino group (-NH3+/-NH2 equilibria) is much less sensitive to changes in the dielectric constant of the solvent than the carboxylic group [42]. • Significant increase in apparent pKa of sulfonephthalein indicators was reported in mixed solvents with increased concentration of organic solvent. For example, pKa of bromocresol green increased by approximately 4 units, from 5.4 to 9.5, with a decrease in water content from 84 to 10 wt% in acetone–water mixtures [41]. In order to compensate for the change of pKa of the sulfonephthalein probe molecules due to decrease in water content, use of pKa values in a mixed solvent (water with 25% IPA) was suggested for calculations of the apparent acidity with the pH indicators [43]. While this might be a sensible approach from the general perspectives, practical application of this method is complicated because it would require different mixed solvents to match dielectric properties and polarities of different formulations. • Reliance on a specific type of the indicators (i.e., sulfonephtaliens) of the 2-/- (base/acid) charge type may not be sufficient; e.g., see Fig. 1 which shows opposite trends in the apparent acidity with water content for H0 vs H- Hammett acidity functions. As alternative probe molecules, carboxylic acids have been evaluated, with the charge type -/0 (base/ acid). The extent of protonation of these molecules can be measured using 13C NMR, provided that 13C -enriched carboxylic acid is used [44]. Preliminary results of the solid-state 13C NMR study with a carboxylic acid, succinic acid (SA), as a probe molecule are presented in Figs. 3, 4 and 5 and Table 1, as described below.

Figure 3 shows the 13C SSNMR spectrum in the carbonyl region for succinic acid and sodium succinate salts as received (non-lyophilized). There is a difference in peak location for the protonated and deprotonated carbonyl carbon, which means that ionization can be measured in the solid state. Figure 3 shows also that monosodium succinate salt has two peaks; therefore, there are total of four peaks which are associated with three succinate species. Lyophilized formulations containing trehalose (10%) and 13C-SA (0.01%) were prepared with three different buffers (175 mM): sodium citrate, sodium phosphate, or potassium phosphate. Formulations were made at pH 5.5 and then adjusted to the desired pH value (5.75, 6.00, 6.25 and 6.50) using NaOH. The solutions also contained bromocresol purple, for visual evaluation of color changes as an orthogonal semi-quantitative test for apparent acidity. Figure 4 shows the 13C SSNMR spectra for lyophilized citrate formulations, with the experimental spectra fitted to two peaks in the carbonyl region. The simplified (i.e., using two peaks) fitting method

44

Fig. 3

D. Heger et al.

13

C SSNMR spectra of the three forms of succinic acid, as-received forms

Fig. 4 (Left) 13C SSNMR spectra of the trehalose/citrate lyophiles in the carbonyl region, with fitting performed using the 2-peaks method; the residuals are in blue. (Right) photos of vials with pH indicator with pre-lyophilization solutions (top) and the lyophiles (bottom); the pH increased from left (pH 5.5) to right at 0.25 intervals

Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals

45

Monosodium succinate

Succinic acid Disodium succinate

200

195 190 13C

Fig. 5

185

180 175

170 165

160

Chemical Shift (ppm)

13

C SSNMR spectra of a trehalose/phosphate pH 5.75 lyophile in the carbonyl region. Fitting results with the 4-peaks method are also shown

Table 1 pHeq for trehalose/citrate formulations based on solid-state 13C NMR measurements with 13C-enriched succinic acid probe molecule pH before lyophilization, solution 5.5 5.75 6.0 6.25 6.5

pHeq, lyophile 5.38 5.65 5.88 6.12 6.33

appears to be adequate for citrate samples. The fitting results were used to calculate pHeq, and the results are presented in Table 1. The pHeq for the freeze-dried formulations trend with the pH of pre-lyophilization solutions. The pHeq values are slightly lower than pre-lyophilization solution pH, indicative of a minor acidic shift during freeze-drying. The same trend is observed using the pH indicator, with change in color from solution to lyophile samples showing acidic shifts. For trehalose/ phosphate formulations, fitting using 2-peaks approach was not sufficient, and all 4 peaks were necessary to achieve a reasonable fit (Fig. 5); such complications could increase uncertainty in calculation of pHeq. In order to resolve uncertainties associated with the presence of multiple species, a carboxylic acid probe with a single ionizable group, butyric acid, can be used instead [44]. Use of different probe molecules and different methods to detect extent of protonation would increase confidence in the empirical measurements of apparent acidity for freeze-dried formulations.

3

Factors Which Impact Apparent Solid-State Acidity in Lyophiles

As discussed in Sect. 2, solid-state acidity has been quantified by measuring ionization of probe molecules, and can be presented as either pHeq or Hammett acidity function. Main factors which determine solid-state acidity in freeze-dried materials are as follows: • Pre-lyophilization solution pH has a major impact on apparent acidity of freeze-dried material. Linear correlations between pre-lyo solution pH and the H2- were reported, when other factors (buffer type, type and concentration of other solutes present) are kept constant [38]. • In addition to pre-lyo solution pH, excipients (e.g., lyoprotectors and bulking agents) could also have an impact on the apparent acidity of lyophiles. Some excipients are associated with larger shifts in the apparent acidity. One prominent example is dextran, which demonstrated a significantly higher acidity (i.e., lower H2-) in the freeze-dried state as compared with disaccharides sucrose and lactose. PVP, on the other hand, has relatively higher basicity than sucrose, lactose, and dextran [45]. • Formulations with sodium phosphate have larger acidic shifts during freeze-drying compared to citrate, even in cases of relatively low buffer concentrations and the presence of amorphous solutes when buffer crystallization is unlikely [46]. Solid-state acidity can also depend on the buffer concentration, as discussed below.

46

D. Heger et al.

• Sugars, such as trehalose, can reduce change in the ionization during freeze-drying [38]. This observation has led to a hypothesis of an additional lyoprotection mechanism for stabilization of proteins by sugars, that is, that sugars can stabilize proteins by minimizing disruption of ionization of ionizable groups of proteins during drying. • Determination of ‘acidity’ of a lyophile may not be sufficient for a comprehensive evaluation of acid-base stresses during freeze-drying of a particular formulation. It was observed that the acidity of freeze-dried sodium and potassium phosphate formulations were higher (i.e., the Hx lower, by 1–3 units) than that of the starting solutions, when the initial (pre-lyophilization) solution pH values were 5–9 [47]. However, the decrease in the Hx in the frozen state was much more pronounced (from 3 to 7 units). Therefore, it would be essential to monitor acidity changes both in frozen and freezedried formulation to have a comprehensive understanding of potential destabilization stresses during freeze-drying. • While lyophilization could result in major change in the ionization of a probe molecule, ionization of a buffer (e.g., histidine or citrate) in the dried state is usually found similar to that in the initial solution before drying [23, 48–51]. Based on such observations, it has been suggested that neither pKa nor pH change as a solution is converted into a dried material [51]. It was pointed out, however, that apparently unchanged ionization of buffer does not necessarily mean that acidity didn’t change [23]. Specifically, if the concentration of counterions is significantly higher than that of protons, the extent of ionization would mainly be dictated by the counterions (e.g., sodium cations); the ratio of acidic/basic forms of a buffer would not change during freezing and freeze-drying. • Ionization state of the probe molecules (and therefore solid-state acidity) is not fixed during the freezing stage. In a study of an amorphous freeze-dried trehalose/citrate mixture, an increase in the ionization extent of a probe molecule, and the corresponding decrease in the acidity, was observed as water content increased by equilibrating the freeze-dried material at different relative humidity (RH) environments [38]. This report was corroborated and extended in a follow-up study in which the impact of water sorption and desorption on the Hammett acidity function was investigated. The study, which also addressed relationships between buffer concentration and solid-state acidity, was presented elsewhere [52], and described below in some details. Aqueous solutions containing trehalose (10% w/v of trehalose dihydrate), sodium citrate buffer (0–100 mM), and bromocresol green (BG) indicator (20μg/ml corresponding to 0.028 mM) were freeze-dried in a Virtis Advantage freezedryer. The [Na+]: [total citrate] molar ratio was fixed at 2.13. Lyophiles were confirmed to be amorphous, and the Tg of anhydrous lyophiles was determined by DSC to decrease from 394 K (no buffer) to 387 K (0.18 weight fraction of citrate buffer). Ionization state of BG in both solutions and lyophiles was measured using visible DRS. DRS spectra of solutions are compared with the spectra of lyophilized samples in Fig. 6, while Fig. 7 shows indicator ionization, solution pH and Hammett acidity of the freeze-dried samples as a function of citrate buffer concentration. Figures 6 and 7 demonstrate that lyophilization-induced changes in ionization and acidity were more pronounced in the presence of a higher buffer concentrations. While the physical mechanism for this observation has not been established, this data could point to the acidic component of the citrate buffer (-COOH groups) as a major source of protons: higher buffer concentration = greater extent of proton transfer to indicator molecules, resulting in higher protonation of the indicator molecules and correspondingly lower Hammett acidity function (i.e., higher acidity of the matrix). This description is similar to general buffer catalysis in chemical reactions in solution. A practical implication of these results is that the presence of a carboxylic buffer, even when buffer remains amorphous, could result in a higher apparent acidity as compared to a formulation without buffer; therefore, inclusion of buffer might be undesirable if the active ingredient is acid sensitive. To study impact of water sorption/desorption on the solid-state acidity, trehalose-citrate lyophiles with BG indicator were equilibrated at different RH (11–33%), and the equilibrated samples were tested by DRS. Water content was determined by Karl Fischer titration; powder X-ray diffraction (PXRD) analysis of the hydrated samples confirmed that the materials remained in amorphous state. After the equilibration of the samples at different RH, they were dehydrated again, either over anhydrous calcium sulphate or under vacuum; water content after dehydration was also determined. Examples of diffuse reflectance spectra after rehydration are provided in Fig. 8, and the data on % ionization of BG from rehydration/dehydration experiments are presented in Figs. 8 and 9. Sorption of water by the lyophiles resulted in an increase in the ionization of the probe molecule (i.e., decrease in the solidstate acidity, increase in H2-) in all samples. Subsequent desorption of water was found to cause partial reversal of the changes produced by sorption in the majority of cases. The hydration/dehydration data show hysteresis; this hysteretic behavior could reflect the nonequilibrium nature of the glassy state, and might indicate that kinetics of water removal is faster than the rate of protonation/deprotonation process. This is an intriguing possibility, considering that proton transfer reactions in solution are among the fastest [53]. Follow-up studies are therefore warranted to study kinetics of changes in ionization during hydration/ rehydration.

Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals

47

In2(deprotonated)

b Kubelka Munk function, F(R)

a

Absorbance

1.0 0.8

In(protonated)

0.6 0.4 0.2 0.0 400

500

600

0.40

0.30

0.20

0.10

0.00 400

700

c 1.0

10 mM cit

700

0.3

10 mM cit 50 mM cit

20 mM cit F(R), normalized

Absorbance

600

Unbuffered 20 mM cit 100 mM cit

d

Unbuffered

0.8

500

Wavelength (nm)

Wavelength (nm)

50 mM cit 0.6

100 mM cit

0.4

0.2

0.1

0.2 0.0

0.0 400

500

600

Wavelength (nm)

700

400

600 500 Wavelength (nm)

700

Fig. 6 Visible spectra of pre-lyophilization trehalose/citrate solutions and corresponding freeze-dried amorphous solids. (a) Pre-lyophilization solutions with bromocresol green; pH was adjusted to obtain bromocresol green in the completely unionized (HIn-) or completely ionized (In2-) states. (b) Lyophiles containing bromocresol green; the lyophiles were obtained from solutions pH-adjusted to pH 2 and 8, to generate completely protonated and deprotonated forms, respectively. The solutions contained 20 mg/mL of BG, trehalose (10% w/v as dihydrate) and 20 mM citrate. (c) Pre-lyophilization solutions containing varying concentrations of citrate buffer, [Na+]: [total citrate] ratio 2.13. and (d) lyophiles obtained from the solutions in Fig. 6c

4

Correlations Between Solid-State Acidity and Stability of Freeze-Dried Formulations

The pH of solution before freeze-drying is an important formulation variable as it may have a significant impact on stability of the resulting material [54–57]. However, stability in the dehydrated state does not always parallel the pH dependence in solution when solid-state degradation is presented as a function of pre-lyophilization solution pH [58–60], suggesting that knowledge of the pH-stability relationships in solution may not be sufficient to predict solution pH that would yield a freezedried formulation with maximal solid-state stability. It was shown, for example, that freeze-dried formulations produced from solutions with the same pH & buffer but different bulking agents can have significant difference in the rate of acid-catalyzed chemical degradation [45, 61]. In these studies, common stability-relevant factors, such as residual water content, crystallinity, and molecular mobility were controlled, and it was shown that these properties would not explain difference in chemical stability between different formulations. In these and several other studies, the stability correlated with the apparent solid-state acidity (expressed as either pHeq or Hammett acidity function) was demonstrated, as summarized in Table 2. There are also studies on relationships between surface acidity and chemical reactivity in physical mixtures of powders; these studies are described elsewhere [35, 66–69] and are not included in Table 2.

48

D. Heger et al.

Fig. 7 (Top) Ionization of bromocresol green indicator in pre-lyophilization solution, in lyophile, and in lyophile after rehydration. (Bottom) Pre-lyophilization solution pH and Hammett acidity function of freeze-dried samples as a function of citrate buffer concentration. Note that the ionized form exhibited an absorption peak with an lmax at ~400 nm (Fig. 6a, b). Therefore, for samples with intermediate extents of ionization, signals measured at the lmax of the unionized form would have contributions from both the ionized and the unionized species and the lower wavelength peak would shift closer to 400 nm as ionization extent increased. To calculate the ionization extents and H2- values, the signal at 443 nm was corrected to subtract the contribution of In2- at this wavelength, to yield the signal attributable to HIn-

In discussions of stability of acid/base sensitive compounds in freeze-dried formulations, there are several points to consider, as follows: (i) Table 2 provides examples of correlations between the solid-state Hammett acidity function and chemical instability for several compounds in various freeze-dried formulations. As shown in the Table, the signs (+ or –) of the coefficients in dependences between acidity and reaction rate are the same in solution and the solid-state. Therefore, knowledge of the pH stability profile in solution and determination of the solid-state Hammett acidity functions could be used to predict stability rank-order between different formulations for an acid/base sensitive active ingredient.

Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals

49

0.5

F(R), normalized

0.4

Unbuffered lyophile

Water content 1.3%

0.3

4.0% 5.9% 8.1%

0.2

0.1

0.0 400

500

600

700

800

% lonization of bromocresol green

Wavelength (nm) 82

Unbuffered lyophile 77 72 Water sorption

67

Desorbed over anhydous Ca sulphate Desorbed under high vaccum

62 57

0.0

2.0

4.0 6.0 Water content (% w/w)

8.0

Fig. 8 Top panel: Diffuse reflectance visible spectra of trehalose-bromocresol green lyophiles obtained from an unbuffered pre-lyophilization solution, at different water contents. Bottom panel: Effect of water sorption and desorption on the ionization of bromocresol green in an amorphous trehalose matrix. The diamonds represent freeze-dried amorphous samples stored under RH conditions ranging from 0% to 33% at room temperature. The squares represent samples desorbed by storage over anhydrous calcium sulphate (0% RH) following sorption. The triangle represents the sample dried under reduced pressure in a freeze-dryer. The broken arrow represents the desorption pathway. Each data point is a mean of multiple measurements and the standard deviations are smaller than the size of the legends

(ii) The slopes of the log k versus H2- are lower than that for the same reaction in solution when determined for log k vs pH. For the sucrose inversion reaction, for example, the slope of ~0.5 was observed, which is significantly lower than the theoretical slope of 1. For a reaction with a unimolecular decomposition of the conjugate acid as the rate-limiting step (protonated sucrose molecule as in case of sucrose inversion), the slope should be equal to 1, which was indeed observed for sucrose inversion in various liquid matrixes [62]. The Hammett acidity function does not represent universal acidity scale because it also depends on the indicator type (see previous section). Therefore, the Hammett acidity function should be considered as an empirical solid- state acidity scale for lyophiles, and it not a true measure of proton activity. From the practical perspectives, solution pH stability profiles, while being relevant to understand trends and stability rank-order for freeze-dried formulations (see previous point), could not be used for quantitative extrapolation of relationships between the Hammett acidity function and solid-state stability. (iii) There are multiple acidity scales, in addition to Hammett acidity function. Evaluation of relative acidity/basicity using solvatochromic probe molecules is described briefly in the next section. As another example, using the rate of sucrose inversion to study acid–base relationships in solution has been reported [63]. Similarly, solid-state acidity of different freeze-dried formulations could also be represented using a rate of sucrose inversion, or another well-characterized acidcatalyzed chemical reaction, as proposed in [45], at least for freeze-dried formulations with the H2- < 4. (iv) Removal of water during freeze-drying (by freeze-concentration and subsequent drying) and corresponding decrease in polarity of the environment can result in significant pKa changes of ionizable groups, which would lead to change in the extent of protonation. For protein drugs, such changes could disrupt electrostatic interactions and destabilize higher-order

50

D. Heger et al.

Fig. 9 Effect of water sorption and desorption on the ionization of bromocresol green in amorphous trehalose–citrate matrices obtained from pre-lyophilization solutions with different citrate concentrations. Each data point is a mean of multiple measurements and the vertical error bars (standard deviations) are smaller than the size of the legends. The diamonds represent freeze-dried amorphous samples stored under RH conditions ranging from 0% to 33% at room temperature. The squares represent samples desorbed by storage over anhydrous calcium sulphate (0% RH) following sorption. The triangle represents the sample dried under reduced pressure in a freeze-dryer. Lines are provided as visual guides. The broken arrow represents the desorption pathways. In four samples (desorbed over calcium sulfate), minimal change or a minor increase in ionization was observed during desorption, in contrast to the majority of cases; it is possible that these four samples represent artifacts in the DRS measurements, during which samples were exposed to ambient RH

structure. Stability of small molecular weight drugs can also be sensitive to pKa changes and corresponding changes in the ionization extent of the molecule, because rate of many chemical reactions depend on the ionization state; e.g., unionized form of sertraline underwent oxidation whereas the charged form of the same molecule was stable in both solution and solid state [70]. (v) There could be a change in predominant degradation mechanism from solution to solid state. For example, general buffer catalysis could be enhanced in freeze-dried materials, because of a major increase in buffer concentration. A potential example of general acid catalysis for a freeze-dried material was described in [57], where increased rate of sucrose inversion in freeze-dried citrate/sucrose mixtures at higher citrate/sucrose ratio was directly related to higher concentration of proton-donating -COOH groups of citric acid. (vi) Because of an increased concentration of salts as the result of water removal, salt effects could play a significant role in stability of freeze-dried formulations. Both stabilizing and destabilizing actions of neutral salts for freeze-dried protein formulations were reported in the literature. For example, NaCl was observed to reduce aggregation of lyophilized bovine serum albumin (BSA) [71], recombinant humanized albumin (rHA) [72], tetanus and diphtheria toxoids [73], as well as to improve stability of freeze-dried recombinant Factor VIII [74]. On the other hand, NaCl destabilized both human growth hormone and Factor VIII by accelerating chemical degradation [75] and loss of activity [76] in freeze-dried state. It should also be mentioned that stabilization/destabilization mechanisms of neutral salts for freeze-dried proteins has not been established, as noted, e.g., in [76]: “... it is not obvious why stability is better (without NaCl)”. Moreover, role of

[46] [65]

Predominantly hydrolysis

Hydrolysis

Small molecular weight drug Cefotaxime sodium

7.89 – 8.93

6.4 – 7.35

2.64 – 3.17

2.62 – 3.02

2.51 – 3.20

Pre-lyo solution pH 2.59 – 3.35

8.12 – 8.61a

Solid-state H22.01 – 2.85a; 2.16 – 2.73b 1.78 – 3.09b (1 point at 3.49 out-of-trend) 7), while the negative FP corresponds the increased acidity (i.e., lower values of the Hammett acidity function). It should be noted that this effect differs from the partial buffer crystallization as the salts under the current discussion are derived from the strong acids and bases and thus do not allow for salts other than neutral. Furthermore, the detailed study of the acidity changes in frozen

56

D. Heger et al.

Fig. 13 Schematics of freezing of NaCl and NH4Cl aqueous solution, where buildup of the freezing potential is followed by neutralization by the flow of protons. Blue shapes represent solution, and white shapes correspond to ice crystals Table 4 The signs of freezing potential (FP) between FCS and ice together with the values of the Hammett acidity function [98]

Salt NH4Cl No salt Na2SO4 NaI NaBr NaCl KCl NaF CaCl2 BaCl2 TMACl MgCl2

FP -

+ + + + + + + + +

Hammett acidity function 1.1 1.3 1.5 7.1 7.1 7.6 7.6 7.7 7.8 7.9 7.9 8.0

NaCl solutions as a function of the cooling rates and the sample temperature revealed complex behavior [99]. Fast cooling resulted in the crystallization of the majority of the NaCl solution, whereas amorphous FCS was also formed. Upon heating, the FCS underwent glass-to-liquid transformation at 180 K, followed by the “cold crystallization” as observed by DSC. Most importantly, the individual crystallization events were accompanied by acidity shifts. The example demonstrates the complexity of the freezing process even for a simple neutral salt and demonstrates a pressing need for the experimental determination of the actual acidity in various frozen systems. It should be stressed that formation of FP in the frozen solution depends on a number of factors including freezing rate, crystals orientation [102], and concentration of both ions and nonionic solutes. Impact of nonionic compounds on FP was studied for dimethylsulfoxide (DMSO), glycerol, sucrose, urea, and tetrahydrofuran (THF) [101–103]. While DMSO, THF, and glycerol reduce the absolute values of FPs, urea enhances it. Interested reader is referred for details to a good review article [103] and book [86]. The theoretical models of various complexities were constructed to describe the phenomena [104, 105, 106]. Knowledge of FP-induced acidity change has been successfully applied to stabilize an enzyme during freeze-thaw, by neutralization of the freeze-induced acidity changes [98]. Activity of haloalkane dehalogenase was studied after the freezethaw cycles for solutions containing sodium phosphate (Na-P) and potassium phosphate (K-P) buffers. The solutions were characterized for their acidity in the frozen states. From the original pH 7.5 the sodium phosphate dropped to H2-1.4, whereas the potassium phosphate only to 6.1 (Fig. 14); such finding is in the agreement with the acidity changes due to the crystallization of buffer components [107]. The examined haloalkane dehalogenase was shown to be sensitive to the acidity changes as the recovery activity in K-P was 66 ± 6%, whereas only 16 ± 9% in Na-P. Furthermore, the freezing-induced acidification of phosphate buffers was neutralized by the presence of tetramethylammonium chloride (TMAC). This salt consists of bulky cation, which is impossible to fit into the ice lattice and chloride anion, which can substitute the water molecules in the ice. TMAC has, therefore, positive FP (solution to ice), whose neutralization results in more basic FCS. Various concentrations of TMAC were experimentally applied to find sufficient concentration for the neutralization of the acidification of 50 mM phosphate buffer; it was found that 0.1 M TMAC was required. The resulting frozen solution did not deviate from the initial solution acidity by more than 0.4 units. The major reduction in freeze-induced change in the acidity resulted in preservation of the activity of the enzyme within the three freeze-thaw cycles. The successful method of activity

Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals

120

57

C

80 60 40

+ 0.1 M TMAC

+ 0.1 M TMAC

Recovery activity / %

100

20 H0

0

1.4

7.1

50 mM Na-P

6.1

7.5

50 mM K-P

Fig. 14 Activity recovery for haloalkane dehalogenase after freeze-thaw in sodium phosphate (Na-P) and potassium phosphate (K-P) buffers, and in these buffers with the addition of TMAC. (Reproduced with permission from: Krausková et al. [98]) Fig. 15 Representation of the “ionic cryoprotection” mechanism. The acidity of particular buffer is first determined in the frozen state and the salt with the opposite freezeinduced acidity change is applied to preserve the solution acidity in the freezing process. (Reproduced with permission from: Krausková et al. [98])

preservation in the freezing was named ionic cryoprotection and has the potential to be generalized for other systems. The ionic cryoprotection concept is schematically presented in Fig. 15. It should be stressed that, while elevated FP would be indicative of a risk of significant acidity changes, the absence of the FP does not necessarily mean that the acidity would be unchanged. Further studies are needed to establish the relationship between FP and acidity changes at various conditions.

6

Conclusions

In both frozen and freeze-dried materials, acidity can be measured with the help of probe molecules, and expressed as either Hammett acidity function or pHeq. Sulfonephtalein dyes (pH indicators) represent the most common type of such probe molecules, while C13-enriched carboxylic acid probes with C13 SSNMR detection have been also introduced more recently. Correlations between the H2-/pHeq and stability have been demonstrated for several small molecular weight compounds in freeze-dried formulations. The relationships between the solid-state acidity and degradation of acid-sensitive proteins were already found in the frozen state [7, 11, 98] and we expect the similar relation to hold for freeze-dried protein formulations as well. Solid state acidity in freeze-dried formulations strongly depends upon initial solution pH, whereas other defining factors include the type and concentration of buffer, bulking agent, and neutral salts. In addition, solid-state acidity depends, albeit in a lesser extent, on the water content and hydration history of an amorphous freeze-dried material. Interestingly, a hysteresis in the ionization of pH indicators was observed in hydration/rehydration experiments, which is consistent with a nonequilibrium

58

D. Heger et al.

nature of glasses, but also raises an intriguing question about the rate of proton transfer in carbohydrate-based amorphous systems. Freezing and drying can result in significant changes in acidity of frozen solutions and freeze-dried formulations. There are a several mechanisms, which could be responsible for the acidity changes occurring in the freezing and lyophilization process, with buffer crystallization and temperature-dependent pKa being the most acknowledged at this point. The chapter introduced a novel (for pharmaceutical science) concept of Workman-Reynolds freezing potential, allowing to develop an additional approach to minimize freeze-induced pH changes, utilizing neutral salts. Neutral salts were shown to reduce freezing potential between ice and liquid layer, therefore minimizing freeze-induced acidification of the freeze-concentrated solution and stabilizing proteins against freeze-induced stresses. In addition to the Workman-Reynolds potential, a possible role of a medium effect in both apparent acidity changes and the chemical instability of freeze-dried formulations is highlighted, using a study of salt effect in amorphous solid-state sucrose inversion. In this particular case, neutral salts accelerated chemical degradation of the acid-catalyzed reaction via both primary and secondary salt effects, while it is also noted that the magnitude and direction of the salt effect (i.e., either stabilization or destabilization by salts) could depend on the charge type of the activated complex of a particular chemical process. To summarize, fundamental understanding of acid–base relationships during freezing and drying would be essential for both rationale formulation design and development of manufacturing process. For a further progress in the field of acid–base relationships for frozen solutions and freeze-dried formulations, several areas would warrant further investigations, as follows: (i) In the majority of cases, Hammett acidity function/pHeq measurements were performed using sulfonephtaleins with the charge type 2-/- (basic/acidic form). In addition, pKa of these molecules exhibit strong dependence on the polarity of the media. It would be essential to introduce alternative types of probe molecules with other charge types, e.g., 0/+, and with weaker changes in pKa with polarity. (ii) While there are examples of relationships between solid-state acidity and stability of small molecules in freeze-dried formulations, similar studies for proteins are lacking. This is an obvious gap in mechanistic understanding of stability of freeze-dried biologicals, which should be addressed in future studies. (iii) Changes in ionization state of probe molecules were observed during partial rehydration/dehydration of freeze-dried formulations; the rehydration/dehydration changes in the ionization were partially reversible, showing a hysteresis in proton transfer. Extension of these initial studies could provide an important fundamental insight into relationships between proton transfer and rates of degradation reactions in amorphous solids. (iv) Solvatochromic dye molecules, which are used extensively in organic physical chemistry to describe chemical reactions in different media, have been introduced recently to measure polarity and acidity/basicity in frozen aqueous systems. Similar approach, if applied to pharmaceutically relevant frozen solutions and freeze-dried materials (e.g., to protein- and sugar- containing systems), could improve the understanding of the main destabilization factors for frozen and freezedried pharmaceuticals.

Acknowledgements We thank Dr. Mark Berry for his expert help in statistical analysis of the data reported in Sect. 4. Dominik Heger is thankful for the support by Czech Science Foundation via project GA 19-08239S.

References 1. Waterman KC, Adami RC, Alsante KM, Antipas A, Arenson D, Carrier R, Hong J, Landis MS, Lombardo F, Shah J, Shalaev E, Smith SW, Wang H. Hydrolysis in pharmaceutical formulations. Pharm Dev Technol. 2002;7:113–46. 2. Connors KA, Amidon GL, Stella AJ. Chemical stability of pharmaceuticals. A handbook for pharmacists. 2nd ed. New York/Chichester/ Brisbane/Toronto/Singapore: Wiley; 1986. p. 847. 3. Flynn GL. Buffers—pH control within pharmaceutical systems. PDA J Pharm Sci Technol. 1980;34:139–62. 4. Zbacnik TJ, Holcomb RE, Katayama DS, Murphy BM, Payne RW, Coccaro RC, Evans GJ, Matsuura JE, Henry CS, Manning MC. Role of buffers in protein formulations. J Pharm Sci. 2017;106:713–33. 5. Shalaev EY. The impact of buffer on processing and stability of freeze-dried dosage forms. Part. 1 solution freezing behavior. Am Pharm Rev. 2005;8:80–7. 6. Thakral S, Sonje J, Munjal B, Suryanarayanan R. Stabilizers and their interaction with formulation components in frozen and freeze-dried protein formulations. Adv Drug Deliv Rev. 2021;173:1–19. 7. Kolhe P, Amend E, Singh SK. Impact of freezing on pH of buffered solutions and consequences for monoclonal antibody aggregation. Biotechnol Prog. 2010;26:727–33.

Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals

59

8. Sundaramurthi P, Suryanarayanan R. The effect of crystallizing and noncrystallizing cosolutes on succinate buffer crystallization and the consequent pH shift in frozen solutions. Pharm Res. 2011;28:374–85. 9. Thorat AA, Munjal B, Geders TW, Suryanarayanan R. Freezing-induced protein aggregation - role of pH shift and potential mitigation strategies. J Control Release. 2020; https://doi.org/10.1016/j.jconrel.2020.04.033. 10. Kohle P, Pikal-Cleland KA, Carpenter JF. Lyophilization-induced protein denaturation in phosphate buffer systems: monomeric and tetrameric bgalactosidase. J Pharm Sci. 2001;90:1255–68. 11. Pikal-Cleland KA, Rodríguez-Hornedo N, Amidon GL, Carpenter JF. Protein denaturation during freezing and thawing in phosphate buffer systems: monomeric and tetrameric b-galactosidase. Arch Biochem Biophys. 2000;384:398–406. 12. Shalaev EY, Gatlin LA. The impact of buffer on solid-state properties and stability of freeze-dried dosage forms. In: Jameel F, Hershenson S, editors. Formulation and process development strategies for manufacturing biopharmaceuticals. Willey; 2010. p. 507–20. 13. Osterberg T, Wadsten T. Physical state of L-histidine after freeze-drying and long-term storage. Eur J Pharm. 1999;273:85–93. 14. Sundaramurthi P, Shalaev E, Suryanarayanan R. Calorimetric and diffractometric evidence for the sequential crystallization of buffer components and the consequential pH swing in frozen solutions. J Phys Chem B. 2010;114:4915–23. 15. Sieracki NA, Hwang HJ, Lee MK, Garner DK, Lu Y. A temperature independent pH (TIP) buffer for biomedical biophysical applications at low temperatures. Chem Comm. 2008;7:823–5. https://doi.org/10.1039/b714446f. 16. Vesely L, Susrisweta B, Heger D. Making good’s buffers good for freezing: the acidity changes and their elimination via mixing with sodium phosphate. Int J Pharm. 2021;593:120128. https://doi.org/10.1016/j.ijpharm.2020.120128. 17. Kitada K, Suda Y, Takenaka N. Acceleration and reaction mechanism of the N-Nitrosation reaction of dimethylamine with nitrite in ice. J Phys Chem A. 2017;121:5383–8. https://doi.org/10.1021/acs.jpca.7b03246. 18. Choi Y, Yoon H-I, Lee C, Vetráková LU, Heger D, Kim K, Kim J. Activation of periodate by freezing for the degradation of aqueous organic pollutants. Envir Sci Tech. 2018;52:5378–85. https://doi.org/10.1021/acs.est.8b00281. 19. Kim K, Ju J, Kim B, Chung HY, Vetrakova L, Heger D, Saiz-Lopez A, Choi W, Kim J. Nitrite-induced activation of iodate into molecular iodine in frozen solution. Environ Sci Technol. 2019;53:4892–900. https://doi.org/10.1021/acs.est.8b06638. 20. Ju J, Kim J, Vetráková Ľ, Seo J, Heger D, Lee C, Yoon H, Kim K, Kim J. Accelerated redox reaction between chromate and phenolic pollutants during freezing. J Hazard Mater. 2017;329:330–8. https://doi.org/10.1016/j.jhazmat.2017.01.031. 21. Maltini E, Anese M, Shtylla I. State diagrams of some organic acid-water systems of interest in food. Cryo-Lett. 1997;18:263–8. 22. Fukuoka E, Makita M, Yamamura S. Glassy state of pharmaceuticals. III. Thermal properties and stability of glassy pharmaceuticals and their binary glass mixtures. Chem Pharm Bull. 1989;37:1047–50. 23. Li J, Chatterjee K, Medek A, Shalaev E, Zografi G. Acid-base characteristics of bromophenol blue-citrate buffer systems in the amorphous state. J Pharm Sci. 2004;93:697–712. 24. Shalaev E, Johnson-Elton T, Chang L, Pikal MJ. Thermophysical properties of pharmaceutically compatible buffers at sub-zero temperatures: implications for freeze-drying. Pharm Res. 2002;19:195–201. 25. Stoll VS, Blanchard JS. Buffers: principles and practice. Methods Enzymol. 1990;182:24–38. 26. Gokarn YR, Kras E, Nodgaard C, Dharmavaram V, Fesinmeyer RM, Hultgen H, Brych S, Remmele RL Jr, Brems DN, Hershenson S. Selfbuffering antibody formulations. J Pharm Sci. 2008;97:3051–66. 27. Garidel P, Pevestorf B, Bahrenburg S. Stability of buffer-free freeze-dried formulations: a feasibility study of a monoclonal antibody at high protein concentrations. Eur J Pharm Biopharm. 2015;97:125–39. 28. Hammett LP, Deyrup AJ. A series of simple basic indicators. I. the acidity functions of mixtures of sulfuric and perchloric acids with water. J Am Chem Soc. 1932;54:2721–39. 29. Rochester CH. The application of the acidity function concept to strongly basic media. Q Rev Chem Soc. 1966;20:511–25. 30. Bates RG. Determination of pH. Theory and practice. 2nd ed. New York/London/Sydney/Toronto: Wiley; 1973. p. 479. 31. Walling C. The acid strength of surfaces. J Am Chem Soc. 1950;72:1164–8. 32. Paul MA, Long FA. H0 and related indicator acidity functions. Chem Rev. 1957;57:1–45. 33. Bates RG, Schwarzenbach G. Uber pH-werte nichtwasseriger losungen. Helv Chim Acta. 1955;38:699–716. 34. Glombitza BW, Oelkrug D, Schmidt PC. Surface acidity of solid pharmaceutical excipients. Part 1. Determination of the surface acidity. Eur J Pharm Biopharm. 1994;40:289–93. 35. Scheef CA, Schmidt PC. Influence of surface acidity of excipients on the solid state stability of pirenzepine. STP Pharm Sci. 1998;8:91–7. 36. Heger D, Klanova J, Klan P. Enhanced protonation of cresol red in acidic aqueous solutions caused by freezing. J Phys Chem B. 2006;110: 1277–87. https://doi.org/10.1021/jp0553683. 37. Orii Y, Morita M. Measurement of the pH of frozen buffer solutions by using pH indicators. J Biochem. 1977;81:163–8. https://doi.org/10. 1093/oxfordjournals.jbchem.a131431. 38. Govindarajan R, Chatterjee K, Gatlin L, Suryanarayanan R, Shalaev EY. Impact of freeze-drying on ionization of sulfonephthalein probe molecules in trehalose-citrate systems. J Pharm Sci. 2006;95:1498–510. 39. Luthra SA, Shalaev EY, Medek A, Hong J, Pikal MJ. Chemical stability of amorphous materials: specific and general media effects in the role of water in the degradation of freeze-dried zoniporide. J Pharm Sci. 2012;101:3110–23. 40. Mathlouthi M, Reiser P, editors. Sucrose: properties and applications. London/Glasgow/Weinheim/Tokyo/Melbourne/Madras: Blackie Academic & Professional; 1995. p. 294. 41. Mchedlov-Petrosyan NO, Lyubchenko IN. Ionization of sulfonephthalein dyes in organic solvents. Zhurnal Obshchei Khimii. 1987;57:1371– 8. 42. Rossini E, Bochevarov AD, Knapp EW. Empirical conversion of pKa values between different solvents and interpretation of the parameters: application to water, acetonitrile, dimethyl sulfoxide, and methanol. ACS Omega. 2018;3:1653–62. 43. Alkhamis KA. Influence of solid-state acidity on the decomposition of sucrose in amorphous systems (I). Int J Pharm. 2008;362:74–80. 44. Lay-Fortenbery A. Investigation of factors influencing protein stability in lyophilized formulations using solid-state NMR spectroscopy. University of Kentucky; 2019, https://uknowledge.uky.edu/pharmacy_etds/105/ 45. Chatterjee K, Shalaev EY, Suryanarayanan R, Govindarajan R. Correlation between chemical reactivity and the Hammett acidity function in amorphous solids using inversion of sucrose as a model reaction. J Pharm Sci. 2008;97:274–86.

60

D. Heger et al.

46. Wu C, Shamblin S, Varshney D, Shalaev E. Advance understanding of buffer behavior during lyophilization. In: Varshney D, Singh M, editors. Lyophilized biologics and vaccines: modality-based approaches. Springer; 2015. p. 25–42. 47. Vetráková Ľ, Vykoukal V, Heger D. Comparing the acidities of aqueous, frozen, and freeze-dried phosphate buffers: is there a “pH memory” effect? Int J Pharm. 2017;530:316–25. https://doi.org/10.1016/j.ijpharm.2017.08.005. 48. Costantino HR, Griebenow K, Langer R, Klibnov AM. On the pH memory of lyophilized compounds containing protein functional groups. Biotechnol Bioeng. 1997;53:345–8. 49. Frey MH, Opella SJ. The effect of pH on the solid-state 13C NMR spectra of histidine. J Magn Reson. 1986;66:144–7. 50. Munowitz M, Bachovchin WW, Herzfeld J, Dobson CM, Griffin RG. Acid-base and tautomeric equilibria in the solid state: 15N NMR spectroscopy of histidine and imidazole. J Am Chem Soc. 1982;104:1192–6. 51. Henry B, Tekely P, Delpuech JJ. pH and pK determinations by high-resolution solid-state 13C NMR: acid-base and tautomeric equilibria of lyophilized 1-histidine. J Am Chem Soc. 2002;124:2025–34. 52. Govindarajan R, Shalaev E, Gatlin L, Suryanarayanan R. Effect of water and buffer on the Hammett acidity function in amorphous lyophiles. AMORPH 2014. The Felix Franks symposium: a celebration. Cambridge: Girton College; 2014. 53. Donten ML, Vande Vondele J, Hamm P. Speed limits for acid–base chemistry in aqueous solutions. Chimia. 2012;66:182–6. 54. Liu WR, Langer R, Klibanov MA. Moisture-induced aggregation of lyophilized proteins in the solid state. Biotechnol Bioeng. 1991;37:177– 84. 55. Lai MC, Topp EM. Solid-state chemical stability of proteins and peptides. J Pharm Sci. 1999;88:489–500. 56. Pikal MJ, Dellerman KM, Roi MI, Riggin RM. The effects of formulation variables on the stability of freeze-dried human growth hormone. Pharm Res. 1991;8:427–36. 57. Shalaev EY, Lu Q, Shalaeva M, Zografi G. Acid-catalyzed inversion of sucrose in the amorphous state at very low levels of residual water. Pharm Res. 2000;17:366–70. 58. Strickley RG, Visor GC, Lin L-H, Gu L. An unexpected pH effect on the stability of moexipril lyophilized powder. Pharm Res. 1989;6:971–5. 59. Bell LN, Labuza TP. Potential pH implications in the freeze-dried state. Cryo-Letters. 1991;12:235–44. 60. Song Y, Schowen R, Borchardt RT, Topp EM. Effect of “pH” on the rate of asparagine deamidation in polymeric formulations: “pH”-rate profiles. J Pharm Sci. 2001;90:141–56. 61. Lu E, Ewing S, Gatlin L, Suryanarayanan R, Shalaev E. The effect of bulking agents on the chemical stability of acid-sensitive compounds in freeze-dried formulations: sucrose inversion study. J Pharm Sci. 2009;98:3387–96. 62. Alkhamis KA. Influence of solid-state acidity on the decomposition of sucrose in amorphous systems II (effect of buffer) solid-state acidity and sucrose inversion. Drug Dev Ind Pharm. 2009;35:408–16. 63. Altaani BM, Alkhamis KA, Abu Baker S, Haddad R. The relationship between the Hammett acidity and the decomposition of cefotaxime sodium in the solid state. Drug Dev Ind Pharm. 46:1632–8. https://doi.org/10.1080/03639045.2020.1813754. 64. Glombitza BW, Schmidt PC. Surface acidity of solid pharmaceutical excipients II. Effect of the surface acidity on the decomposition rate of acetylsalicylic acid. Eur J Pharm Biopharm. 1995;41:114–9. 65. Gana FZ, Rashid I, Badwan A, Alkhamis KA. Determination of solid-state acidity of chitin-metal silicates and their effect on the degradation of cephalosporin antibiotics. J Pharm Sci. 2012;101:2398–407. 66. Govindarajan R, Landis M, Hancock B, Gatlin LA, Suryanarayanan R, Shalaev EY. Surface acidity and solid-state compatibility of excipients with an acid-sensitive API: case study of atorvastatin calcium. AAPS PharmSciTech. 2015; https://doi.org/10.1208/s12249-014-0231-7. 67. Hailu S, Bogner RH. Solid-state surface acidity and pH-stability profiles of amorphous quinapril hydrochloride and silicate formulations. J Pharm Sci. 2010; https://doi.org/10.1002/jps. 68. Hammett LP. Reaction rates and indicator acidities. Chem Rev. 1935;16:67–79. 69. Indelli A, Mantovani G. The dissociation constants of tri- and tetrametaphosphoric acids by the rate of inversion of sucrose. Trans Faraday Soc. 1965;61:909–13. 70. Hsieh Y-L, Yu W, Xiang Y, Pan W, Waterman KC, Shalaev EY, Shamblin SL, Taylor LS. Impact of sertraline salt form on the oxidative stability in powder blends. Int J Pharm. 2014;461:322–30. 71. Liu WR, Langer R, Klibanov AM. Moisture-induced aggregation of lyophilized proteins in the solid state. Biotechnol Bioeng. 1990;37:177– 84. 72. Costantino HR, Langer R, Klibanov AM. Aggregation of a lyophilized pharmaceutical protein, recombinant human albumin: effect of moisture and stabilization by excipients. Biotechnology (NY). 1995;13:493–6. 73. Schwendeman SP, Costantino HR, Gupta RK, Siber GR, Klibanov AM, Langer R. Stabilization of tetanus and diphtheria toxoids against moisture-induced aggregation. Proc Natl Acad Sci U S A. 1995;92:11234–8. 74. Osterberg T, Fatouros A, Mikaelsson M. Development of a freeze-dried albumin-free formulation of recombinant factor VIII SQ. Pharm Res. 1997;14:892–8. 75. Pikal MJ, Dellerman KM, Roy MI, Riggin RM. The effects of formulation variables on the stability of freeze-dried human growth hormone. Pharm Res. 1991;8:427–36. 76. Jameel F, Tchessalov S, Bjornson E, Lu X, Besman M, Pikal MJ. Development of freeze-dried biosynthetic factor VIII: I. a case study in the optimization of formulation. Pharm Dev Technol. 2009;14:687–97. 77. Shalaev EY, Wang W, Gatlin LA. Rational choice of excipients for use in lyophilized formulations. In: Drugs and the pharmaceutical sciences, 175 (Protein formulation and delivery (2nd edition)). Boca Raton: CRC Press; 2008. p. 197–217. 78. Connors K. Chemical kinetics. The study of reaction rates in solution. New York: VCH Publishers, Inc; 1990. p. 386. 79. Guggenheim E, Wiseman L. Kinetic salt effects on the inversion of sucrose. Proceedings of the royal society of London, Series A: mathematical, physical and engineering sciences; 1950. p. 17–32. 80. Dordick RS, Clarke GA. Salt effects on the hydrolysis of sucrose. J Chem Educ. 1979;56:352. 81. Shamblin SL, Taylor LS, Zografi G. Mixing behavior of colyophilized binary systems. J Pharm Sci. 1998;87:694–701. 82. Pethybridge A, Prue J. Kinetics salt effects and the specific influence of ions on rate constants. Prog Inorg Chem. 1972;17:327–90. 83. Rice J. Mathematical statistics and data analysis. 2nd ed. Belmont: Duxbury press; 1995. 84. Franks F, editor. Water and aqueous solutions at subzero temperatures. Water: a comprehensive treaties, vol. 7. New York: Plenum Press; 1982.

Beyond pH: Acid/Base Relationships in Frozen and Freeze-Dried Pharmaceuticals

61

85. Bartels-Rausch T, Jacobi HW, Kahan TF, Thomas JL, Thomson ES, Abbatt JPD, Ammann M, Blackford JR, Bluhm H, Boxe C, Domine F, Frey MM, Gladich I, Guzmán MI, Heger D, Huthwelker T, Klán P, Kuhs WF, Kuo MH, Maus S, Moussa SG, McNeill VF, Newberg JT, Pettersson JBC, Roeselová M, Sodeau JR. A review of air–ice chemical and physical interactions (AICI): liquids, quasi-liquids, and solids in snow. Atmos Chem Phys. 2014;14:1587–633. https://doi.org/10.5194/acp-14-1587-2014. 86. Hobbs PV. Ice physics. Oxford University Press; 1975. 87. Petrenko VF, Whitworth RW. Physics of ice. Oxford: Oxford University Press; 1999. 88. Elbaum M, Lipson SG, Dash JG. Optical study of surface melting on ice. J Cryst Growth. 1993;129:491–505. 89. Bronshteyn VL, Steponkus PL. Calorimetric studies of freeze-induced dehydration of phospholipids. Biophys J. 1993;65:1853–65. 90. Bhatnagar B, Zakharov B, Fisyuk A, Wen X, Karim F, Lee K, Seryotkin Y, Mogodi M, Fitch A, Boldyreva E, Kostyuchenko A, Shalaev E. Protein/ice interaction: high-resolution synchrotron X-ray diffraction differentiates pharmaceutical proteins from lysozyme. J Phys Chem B. 2019; https://doi.org/10.1021/acs.jpcb.9b02443. 91. Heger D, Jirkovsky J, Klan P. Aggregation of methylene blue in frozen aqueous solutions studied by absorption spectroscopy. J Phys Chem A. 2005;109:6702–9. https://doi.org/10.1021/jp050439j. 92. Heger D, Klan P. Interactions of organic molecules at grain boundaries in ice: a solvatochromic analysis. J Photochem Photobiol A: Chem. 2007;187:275–84. 93. Hauptmann A, Hoelzl G, Loerting T. Distribution of protein content and number of aggregates in monoclonal antibody formulation after largescale freezing. AAPS PharmSciTech. 2019;20:72. 94. Authelin J-R, Rodrigues MA, Tchessalov S, Singh S, McCoy T, Wang S, Shalaev E. Freezing of biologicals revisited: scale, stability, excipients, and degradation stresses. J Pharm Sci. 2020;109:44–61. 95. Workman EJ, Reynolds SE. A suggested mechanism for the generation of thunderstorm electricity. Phys Rev. 1948;74:709–709. https://doi. org/10.1103/PhysRev.74.709. 96. Workman EJ, Reynolds SE. Electrical phenomena occurring during the freezing of dilute aqueous solutions and their possible relationship to thunderstorm electricity. Phys Rev. 1950;78:254. 97. Conde MM, Rovere M, Gallo P. Spontaneous NaCl-doped ice at seawater conditions: focus on the mechanisms of ion inclusion. Phys Chem Chem Phys. 2017;19:9566–74. https://doi.org/10.1039/c7cp00665a. 98. Krausková Ľ, Procházková J, Klašková M, Filipová L, Chaloupková R, Malý S, Damborský J, Heger D. Suppression of protein inactivation during freezing by minimizing pH changes using ionic cryoprotectants. Int J Pharm. 2016;509:41–9. https://doi.org/10.1016/j.ijpharm.2016. 05.031. 99. Imrichova K, Vesely L, Gasser TM, Loerting T, Nedela V, Heger D. Vitrification and increase of basicity in between ice Ih crystals in rapidly frozen dilute NaCl aqueous solutions. J Chem Phys. 2019;151:014503. https://doi.org/10.1063/1.5100852. 100. Robinson C, Boxe CS, Guzman MI, Colussi AJ, Hoffmann MR. Acidity of frozen electrolyte solutions. J Phys Chem B. 2006;110:7613–6. 101. Sola MI, Corti HR. Freezing induced electric potentials and Ph changes in aqueous-solutions of electrolytes. Anales De La Asociacion Quimica Argentina. 1993;81:483–98. 102. Wilson PW, Haymet ADJ. Workman-Reynolds freezing potential measurements between ice and dilute salt solutions for single ice crystal faces. J Phys Chem B. 2008;112:11750–5. 103. Rastogi RP, Tripathi AK. Effect of nonionic solutes on the freezing potential of dilute ionic aqueous solutions. J Chem Phys. 1985;83:1404–5. 104. Gross GW. Some effects of trace inorganics on the ice/water system. In: Trace inorganics in water, vol. 73. American Chemical Society; 1968. p. 27–97 105. Gross GW, Wu C, Bryant L, McKee C. Concentration dependent solute redistribution at the ice/water phase boundary. II Experimental investigation. J Chem Phys. 1975;62:3085–92. https://doi.org/10.1063/1.430909. 106. Bronshteyn VL, Chernov AA. Freezing potentials arising on solidification of dilute aqueous-solutions of electrolytes. J Cryst Growth. 1991;112:129–45. 107. Murase N, Franks F. Salt precipitation during the freeze-concentration of phosphate buffer solutions. Biophys Chem. 1989;34:293–300. https:// doi.org/10.1016/0301-4622(89)80066-3.

Concepts and Strategies in the Design of Formulations for Freeze Drying Feroz Jameel

Abstract

In order to enhance the storage stability of water labile molecules the common practice is to employ lyophilization (freeze drying) process to create dry glassy state dosage form. The lyophilization process is not free of imparting stresses during the processing, and since what is in the formulation dictates the design of the processing conditions, both the formulation and process should be designed in such a way that the drug entity, especially protein molecule, behaves well against the stresses during the various steps or phases of freeze drying, namely, in-process stability, long-term storage stability, and sufficient stability post-reconstitution until it is administered into the patient body. Understanding of key attributes/ liabilities of the protein molecules and the selection of excipients and design of process parameters is key to robust formulation and lyophilization process that can produce a product that has adequate stability, quality, and commercial manufacturing viability. The objective of this chapter is to discuss briefly the concepts and strategies that would help to design a robust formulation and a lyophilization process. Keywords

Pre-formulation · Formulation · Protein · Physical instability · Chemical Instability · Physico-chemical characterization · Stresses · Forced degradation · Colloidal instability · Conformational instability · pH · Surfactant · Stabilizers · TPP · CQAs · Design space · Kinetics

1

Introduction

It is well recognized that the years of efforts in discovery and research in identifying a therapeutic entity will not bring full revenue to the efforts and investments until it is delivered in the right dosage form that ensures a minimum of 2 years of shelflife and meets patient compliance. With the burst of novel formats of both small and large molecules, scientists are increasingly facing challenges in formulating them to address the marginal stability of these entities in the liquid state. The dogma suggests that the stability of most molecules normally increases in the order of solution < glassy solid < crystalline solid due to the increasingly restricted mobility of the reacting species in solid phases [1–3]. Lyophilization (freeze drying) stands out as the choice of drying technology to provide a solid-state product that can enhance the storage stability compared to other drying technologies such as spray drying, spray freeze drying, vacuum drying, and supercritical fluid technology. However, all of these techniques have pros and cons associated with them either related to processing conditions or scalability or yield or cost of goods manufactured (COGM). Although lyophilization in the biopharmaceutical and pharmaceutical industry is recognized as a technique of choice for stabilization of water labile molecules, it presents several challenges and requirements that the formulation scientist must be aware of before embarking on formulation design. The objective of this chapter is to discuss briefly the concepts and strategies that would help to overcome those challenges and design a robust formulation. F. Jameel (✉) Nimble BioSolutions, Gurnee, IL, USA # America Association of Pharmaceutical Scientists 2023 F. Jameel (ed.), Principles and Practices of Lyophilization in Product Development and Manufacturing, AAPS Advances in the Pharmaceutical Sciences Series 59, https://doi.org/10.1007/978-3-031-12634-5_4

63

64

2

F. Jameel

Requirements and Expectations of a Lyophilized Product

The general requirements and expectations of a freeze drying process is to produce a product that has adequate stability, quality, and commercial manufacturing viability, and the formulation design influences all these attributes directly or indirectly. As noted above the lyophilization process is not free of imparting stresses during the processing, and since what is in the formulation dictates the design of the processing conditions, both the formulation and process should be designed in such a way that the drug entity especially protein molecule behave well against the stresses during the various steps or phases of freeze drying, namely, in-process stability, long-term storage stability, and sufficient stability post-reconstitution until it is administered into the patient body. It should also have low residual moisture content, rapid/easy reconstitution, and ease of administration; and should be compatible with the device (reconstitution kit). The aesthetic aspect of the lyophilized product is also equally important from patient and marketing compliance perspective, and hence should appear pharmaceutically elegant without collapse or meltback; The process should be efficient, i.e., the total cycle time should be within few days, typically not more than 4–5 days, easily scalable without “major equipment restrictions/modifications,” and be robust enough to be implemented on any typical production freeze dryer at any site.

3

Pre-formulation

Typically, the formulation development begins with the pre-formulation evaluation that involves forced degradation of the drug entity coupled with physicochemical characterization, the findings of which are intended to guide the formulation design. The drug entity, the protein in the case of biologicals, is subjected to appropriate levels of stress conditions such as high temperature, extreme pH, light exposure, oxidizing agents, freezing and thawing, and/or mechanical stress and the objective of this study is twofold. Firstly, to understand and elucidate its pathway(s) of degradation, chemical and/or physical degradation/ instability. Secondly, to develop/validate stability-indicating assays needed for future formulation development work. This opportunity can also be used to identify preliminary critical quality attributes (CQAs), which is subject to changes as the development work moves from early to late stage [4].

3.1

Forced Degradation/Stress Studies

The conditions that are generally recommended for forced degradation/stress studies are summarized below • Temperature Stress: Elevated temperature (at least 10 °C increment above the accelerated testing temperature as recommended in ICH Q1A and below the lowest melting temperature Tm). • Freeze Thaw Stress: Relevant freezing and thawing cycles (3×) & conditions (-40 °C, -20 °C & 2–8 °C or room temperature (RT)) • Low and High pH Stress: (For instance, below pH 4 and above pH 8 also combined with elevated temperature if needed, to facilitate deamidation) • Oxidation Stress: Oxidizing conditions (chemical agents such as hydrogen peroxide or relevant metals) For Methionine – tert-butyl hydroperoxide (t-BHP) For Tryptophan – 2,2-azobis(2-amidinopropane) dihydrochloride (AAPH) • Light Stress: UV and visible light (for example, using conditions prescribed in ICH Q1B or milder conditions depending on the molecule) • Agitation Stress: Relevant agitation conditions To be “smart from the start” and to avoid surprises and failures in late-stage development, it is becoming increasingly important to perform some of these pre-formulation activities early on during candidate screening and selection so that the liabilities are identified upfront and resolved through mutagenesis if possible while ensuring that the molecule has suitable drug-like properties (e.g., acceptable degradation profile, high solubility, and low viscosity) [5–7]

Concepts and Strategies in the Design of Formulations for Freeze Drying

3.2

65

Chemical Instability

As noted above before embarking on formulation development it is imperative that the molecule is subjected to forced degradation to understand what chemical and physical liabilities that the molecule has. The chemical instabilities which are referred to as those reactions that involve cleavage of the double bond leading to chemical modifications are briefly described below Oxidation Depending upon the location, the residues that undergo oxidation are mostly methionine and tryptophan, but occasionally histidine, cysteine, phenylalanine, or tyrosine are also quite susceptible to oxidation. Oxidation occurs when the molecule is exposed to either light or peroxide or atmospheric oxygen or Redox active metal ions (such as iron, copper, or nickel) [8]. The most powerful oxidant is the hydroxyl radical and/or reactive oxygen species (ROS) that readily reacts with the susceptible amino acids. Oxidation of methionine results in the formation of chemically modified methionine sulfone while the oxidation of tryptophan forms kynureine vial hydroxytryptophan intermediate. It is important to characterize the drug active (protein molecule) and determine if there exists any oxidation propensity and address the liability accordingly, as oxidation of amino acid side chains can lead to changes in the protein conformation, affecting stability, efficacy (effector function), and decrease plasma half-life time [9–15]. The oxidation study is recommended to be conducted in the commercial primary package using the formulation components identified, and testing should be conducted after exposure to relevant oxidation stress conditions such as hydrogen peroxide, UV light, relevant LUX hours, and relevant amounts of appropriate transition metals. Deamidation Asparagine (Asn) deamidation is a chemical modification that occurs in protein molecules via a succinimide intermediate, resulting in either aspartic acid (Asp) or iso-aspartic acid (iso-Asp) residue formations. It results in the creation of acidic charge variant through the introduction of the negatively charged carboxylic acid in place of the neutral amide. Certain sequences, such as Asn-Gly or Gly-Asn, are more susceptible to deamidation and are pH sensitive. Asparagine deamidation is more likely to occur at high pH and ion exchange chromatography, isoelectric focusing, mass spectrometry, and if site identification is needed, tryptic peptide mapping (TPM) may be required. Asp isomerization is more likely to occur at low pH, and high-throughput bioluminescent assays have been used to quantify iso-Asp formation [16, 17]. If the susceptible asparagine is located in a structurally and/or functionally impactful location, efficacy and/or stability can be affected [18]. Disulfide Exchange (Scrambling) - Unpaired Cysteine Residues Disulfide bonds are responsible for maintaining the stability of the protein in a compact three-dimensional structure through crosslinking distant regions. They are also susceptible to hydrolytic and oxidative degradative pathways to form non-native disulfide bonds and other reactive species. Disulfide bonds can form by oxidation, between the thiol (-SH) groups on two adjacent cysteine residues. The free cysteine thiols of proteins can alter existing disulfide bonds, resulting in disulfide bond exchange [19]. This phenomenon is known as disulfide scrambling that can take place intra- and intermolecularly causing conformational changes leading to formation of aggregates. Glycation Glycation is a reaction in which free amino groups of proteins, lipids, and nucleotides are modified by monosaccharides. During the reaction, a Schiff base, Amadori products, various intermediate compounds, and eventually advanced glycation end products (AGEs) are formed. This sequence is also referred to as “nonenzymatic glycosylation” or the “Maillard reaction” [20]. Glycation is one of the most important unwanted posttranslational modifications (PTM), which modifies protein threedimensional decoration and triggers its abnormalities. The large hydrophilic carbohydrate moiety attached to proteins has been implicated in various biological processes, including modification on protein folding [22], modulation of protein stability, oligomerization and aggregation [21, 23], and modulation of enzyme activity [24]. N-terminal Pyroglutamate Cyclization of N-terminus glutamine into pyroglutamic acid is one of the posttranslational or co-translational event and is greatly facilitated by the enzyme glutaminyl cyclase where the N-terminal glutamic acid (Glu) is cyclized to form pyroglutamate (pGlu). As a formulation scientist, one need to be aware of the fact that it is favored at pH 4 and 8, but is

66

F. Jameel

less common at the neutral pH; however, it is believed that this type of N-terminal modification is not expected to have a substantial effect on structural stability, efficacy, and safety of the product [25].

3.3

Physical Instability

Physical instability as opposed to chemical instability does not involve chemical bond but manifest as aggregates (soluble and insoluble) or fragmentation or adsorption. Two most common forms of physical instabilities are fragmentation (clipping) and aggregation. Fragmentation most of the time is not a major concern as long as it is not part of CDR region affecting the potency/biological activity; however, aggregation regardless would be highly undesirable. Aggregation compromises (1) biological functions (effector functions) (2) Potential immunogenicity implications as it can induce immune responses by breaking B-cell tolerance, evoke antibody clearance machinery in vivo, formation of anti-drug antibodies (Neutralization of Drug) and conceptual parallels drawn between ‘aggregates displaying repeats of monomers’ and ‘repetitive nature of the antigens that trigger immune responses with higher efficiency’ – Cytokine release syndrome (CRS) caused by an overactive immune response [26]. Thus, the above-mentioned disadvantages make the control of protein/antibody aggregation imperative in the route to developing successful therapeutics. The formulation scientist should be aware of that the chemical modifications such as methionine oxidation in an IgG1 Fc region lead to an altered secondary and tertiary structure, which can be assessed by Mass spec, circular dichroism, and NMR, higher aggregation propensity upon heat stress, histidine of an antibody may be oxidized and form a covalently linked aggregate, antibody containing unpaired cysteine may form reducible aggregate through intermolecular disulfide bond upon agitation, and deamidation and glycation promote aggregation. Most recently, protein carbonylation was reported to positively correlate to higher antibody aggregation burst rate. Generally speaking, there are two forms or pathways of aggregation, the colloidal instability and the conformational instability.

3.3.1 Colloidal Instability In this form of instability the protein remains in the native state or conformation but tends to aggregate due to intrinsic properties, specifically the charge distributions on the surfaces of the protein. It depends on the attractive and repulsive interactions between the molecules in the solution where non-covalent forces like electrostatic forces, van-der-Waals forces, and hydrophobic and hydration forces are involved resulting in self-association/loosely associated. These weak solute–solute interactions resulting in self-association is protein concentration dependent, which means it self-associates only above a critical protein concentration and is reversible in nature, which means it goes back to monomer upon dilution. The colloidal instability or the aggregate formation is thermodynamically unstable but kinetically stable (see Fig. 1) and often accompanies

Fig. 1 Free energy and colloidal instability

Concepts and Strategies in the Design of Formulations for Freeze Drying

67

Fig. 2 Mutual diffusion coefficient as a function of protein concentration

the increase in viscosity & opalescence [27]. It can manifest as submicron (soluble) aggregates, subvisible and visible particles, or as other phase behaviors. Metrics of Colloidal Stability The osmotic second virial coefficient (B22), a fundamental physiochemical property that describes the molecular interactions between proteins in solution which could result in aggregation, is commonly used as an indicator of colloidal stability where attractive protein–protein colloidal interactions are dominant [28]. Static light scattering (SLS) is the most commonly used and established technique employed to determine the B22 experimentally [29]. The other techniques such as self-interaction chromatography (SIC) [30], membrane osmometry (MO) [31], and analytical ultracentrifugation (AUC) [32, 33] can also be used and they provide equally comparable results. Negative B22 values denote net attractive protein–protein interactions, which means potential/propensity for aggregate formation whilst positive B22 values represent overall repulsive interactions which means stable. The other parameter that is widely used to asses/gauge the propensity for colloidal stability is the diffusion interaction parameter, Kd. It is expressed by the slope of the linear relationship of diffusion coefficient vs. concentration, and can be determined by dynamic light scattering in a high-throughput manner. If the diffusion coefficient increases with the increasing protein concentration due to decreased protein–protein distance it is indicative of repulsive interactions. Attractive interactions or aggregation propensity can be suspected when the diffusion coefficient decreases with the increasing protein concentration due to decreased protein–protein distance [34, 35]. The concept is schematically illustrated below (Fig. 2) They are related through Harding and Johnson eqn. kD = 2B22 M–ks - v where M is the molar mass, ks the first order concentration coefficient of sedimentation velocity and, ν is the partial specific volume Other techniques such as opalescence, small angle-X-ray scattering (SAXS), and measurement of net charge utilizing the principle of electrophoretic mobility, e.g., capillary zone electrophoresis (CZE), membrane confined electrophoresis (MCE), and a new technique called massively parallel phase analysis light scattering (MP-PALS), Mobius mobility instrument (Wyatt), and effective surface charge potential (zeta potential) using electrophoretic light scattering help to understand the charge distribution/colloidal stability and eventually the propensity to aggregate.

3.3.2 Conformational Instability Conformational instability, which is defined as the difference in free energy between the folded and unfolded states of a protein molecule, is due to interfacial denaturation, shear stresses, and solution conditions resulting in hydrophobic interactions leading to non-covalent irreversible aggregates. As opposed to colloidal, it is a non-native aggregation which means perturbation of conformation causing partial or full unfolding is a pre-requisite and is triggered by external conditions.

68

F. Jameel

It is thermodynamically stable but kinetically unstable which means it grows over the period of time [28]. It is not dependent on protein concentration, irreversible in nature, and not necessarily accompanied with increase in viscosity. It is extrinsic in nature and is caused by external factors solution conditions (pH, ionic strength, temperature, etc.) and processing conditions which includes interfaces (i.e., air–water, ice crystal–bulk solution, intermediate/product contact surfaces), freeze-thaw, mechanical stresses (cavitation, shear), ambient light, temperature, and time Conformational stability can be assessed and evaluated as a function of various solution and process conditions using various biophysical techniques such as circular dichroism (Far-UV (190–250 nm for secondary structure) and near-UV (250–350 nm for tertiary structure), Analytical ultracentrifugation/Sedimentation velocity experiments for higher order aggregates, Differential scanning calorimetry (DSC), MicroCal for melting temperatures/thermal transition midpoint (Tm) [36]. The intrinsic fluorescence of aromatic rings containing residues such as tryptophan is dependent upon the overall 3D structure and the surroundings. When a protein is perturbed due to chemical or physical stresses leading to unfolding, the hydrophobic groups are exposed which always results in changes in the intensities of fluorescence and shifts in emission maxima. Differential Scanning Fluorimetry (DSF) which utilizes this principle has become a high-throughput method of choice for the analysis of protein stability, thermal protein unfolding, and melting temperature analysis. The unfolding transition midpoint Tm (°C), which is the point where half of the protein is unfolded is used as an indicator of conformational stability of protein.

4

Formulation Development

During the early stages of molecular assessments for developability, in silico software tools are used to identify regions that are potentially susceptible to chemical (example: oxidation, deamidation sites) and physical instabilities such as unfolding (hydrophobic sites), self-association (charge distribution), aggregation, and higher viscosities associated with high protein concentrations, and the molecules are engineered to mitigate them. Since this strategy helps to moderate the problem but not always possible to completely eliminate the liabilities, as some of the residues may be part of CDR regions and requires biological activity, formulation development is required.

4.1

Rational Design of the Formulation

Formulation development is the interplay of selection and optimization of solution conditions and process conditions with the aid of excipients. Thus, the overall objective of the formulation development effort is to identify the optimal conditions of the final drug product composition, primary container configuration that can maintain the stability, safety, and efficacy of the drug as it is processed through the various unit operations of drug substance, drug product through administration to the patients. Since the posttranslational changes, chemical modifications/instabilities, and the physical instabilities eventually end in the formation of aggregates [37], it is central to the formulation development to understand the mechanism or how does a protein aggregate. The following schematic put together by Christ Roberts et al. [28] provides insight into the various pathways that lead to aggregate formation (Fig. 3).

4.2

Strategies to Mitigate Liabilities and Protein Aggregation

Based on the above scheme the overall strategy would be to target the aggregation Stages 1, 2, and 3 for conformational stability/non-native aggregation where protein unfolding and the nucleation are the key steps for aggregation and stage 1 for colloidal stability/native state aggregation. The corresponding approaches for mitigating aggregation would be • Stabilizing the native monomer which means decreasing free energy for folding (GF) or destabilizing the partially unfolded monomer which means increasing free energy for unfolding (GU) to reduce the potential of protein unfolding at Stage I; • Altering the protein surface charge distributions to increase the electronic repulsion between the unfolded monomers at Stage II & native monomers • Disturbing the structural rearrangements of unfolded monomers in Stage III to disfavor hydrophobic contacts and the packing of β strands

Concepts and Strategies in the Design of Formulations for Freeze Drying

69

Fig. 3 Schematic of aggregation pathways, the red arrows represent the non-native aggregation, while the dark blue arrows denote the native aggregation. The bidirectional arrows show the reversible steps, and the mono-directional arrows account for the irreversible process. (Adapted from Ref. [28])

A stepwise approach to the design of the formulation is proposed below. Step 1: Define the commercial target product profile Compiling the target product profile (TPP) is the first step in the development of the product, as it defines/summarizes the product attributes or specifications and describes how the product will be utilized by the end user. It serves as an essential tool in the development and strategic management of a new or modified Drugs/Biologics/Device and should be prepared jointly by the all departments of the company involved in the development of the product. It is a “living document” evolving and maturing with increasing knowledge and experience and FDA strongly advocates the use of a TPP although it does not mandate it An example of a common template that can be used to assemble TPPs for each product entering development and each new indication for an existing drug/biologic is illustrated in Table 1. Step 2: Selection of pH and Buffer Species Screening and identification of optimal pH, buffer salts, and ionic strength is an important study after defining the TPP, as they influence solubility, stability, and viscosity of a protein solution. Solubility increases as the solution pH gets further away from the isoelectric point (pI). Identify the pI of the mAb from the primary sequence, typically using a commonly available software such as ExPASy or Sednterp or through zetapotential measurements using instruments such as Zetasizer Nano ZSP [38]. The solution pH changes the protein surface charge and may interfere with the favorable electrostatic interactions required for maintenance of the native folded structure. The solution pH also affects the various pathways of chemical degradation as depicted in Fig. 4. It is suggested that a pH study in the range of 4–7.5 with increments of 0.2–0.5 units be carried out to identify the optimal pH while keeping other formulation components constant. While selecting the buffer species/salts the formulation scientist needs to consider the following, • Selection of a buffer species with pKa value close to the desired pH will provide higher buffer capacity and will require lower quantities, and a range of 15–25 mM may be sufficient. Low quantities of buffer salts prevent from depressing the collapse temperature and also prevents from increasing the total solid content both of which helps lyophilization process • Caution must be exercised while selecting buffer salts, as selective crystallization of the less soluble buffer component (acid or base) during freezing (Bulk freeze thaw in a container or vial freezing during lyophilization) can result in massive pH shifts in the freeze concentrate than would have been obtained in an unbuffered system [40–44]. Figure 5 depicts the physical state (amorphous or crystalline) of the buffer salts as a function of temperature and Fig. 6 shows the degree of pH shifts as a function of temperature.

70

F. Jameel

Table 1 Target product profile (TPP) for a lyophilized dosage form Product attribute Target indication Dosage form Final presentation Route of administration Dose (protein) Dose range Dose frequency Setting for use Single or multidose Volume per dose IV bag

Acceptable profile Oncology treatment of leukemia Sterile lyophilized powder for reconstitution with WFI Vial configuration, 20 cc IV infusion 125 mg per vial 50–250 mg Every 2 weeks (preferred) Clinic-based administration Multidose vial with elastomeric closure Based on patient weight 50 mL or 100 mL bag

Target recon volume Diluent (reconstitution medium) Target recon time pH Osmolality Recommended storage conditions Shelf-life

5.3 mL Commercially available WFI 7. Since pH also has a significant impact on solubility and aggregation, it is recommended that deamidation is monitored and that a sweet spot for the pH value is determined early on during development Oxidation Although several amino acids such as methionine, cysteine, histidine, tryptophan, and tyrosine are susceptible to oxidation; however, the oxidation of methionine is fastest and more common than others. Oxidation of conserved heavy chain methionine residues located at the interface of the CH2 and CH3 domains decreases the thermal stability, Protein A binding, FcRn binding, and circulation half-life. Oxidation of methionine residues in the CDRs could potentially impact antigen binding and hence the efficacy of the molecule. Accordingly, it is essential to characterize the effect of methionine oxidation (if any) on the structure, stability, and biological activity of the antibody in the drug product. Once it is determined that it is critical for stability and efficacy, investigate the root cause of oxidation and determine whether • • • • •

It occurs only during quiescent storage It is caused by exposure to metal surfaces It is a temperature-induced formation of free radicals It is induced by oxidative degradants of the excipients It occurs due to exposure to light.

Some of the remedies or measures that can be taken to address or prevent the oxidative degradations are briefly discussed below. Antioxidants, such as methionine, sodium thiosulfate, ascorbic acid, BHT, BHA, sodium bisulfite, glutathione, and propyl gallate are believed to serve as free radicals or oxygen scavengers and are used to decrease oxidation propensity in parenteral products [55, 56]. In protein therapeutic products, sacrificial antioxidants, such as methionine, have been the most commonly used excipients. The minimum effective levels (molar ratios of protein to antioxidant) required to inhibit temperature-induced oxidation is generally observed to be 1:5 and 1:25 for methionine and thiosulfate, respectively. It has been reported by Genentech group that stoichiometric amounts of methionine and thiosulfate are sufficient to eliminate temperature-induced oxidation of rhuMAb HER2 caused by free radicals that were generated by the presence of metal ions and peroxide impurities in the formulation, other examples being Actemra™ and Cosentyx™. Some products like Campath® and Ajoyv® have seen success in using EDTA as a chelator to avoid hydroxyl radical formation. In addition to using antioxidants, overlaying/filling the vial head space with nitrogen and implementing tight control strategies on the impurities in raw materials, especially polysorbate, are also commonly used strategies. Physical Instability As indicated earlier, physical instability can be categorized into conformational instability (partial or full unfolding) and colloidal instability:

74

F. Jameel

Fig. 8 Preferential exclusion of solutes

Conformational If the conformational instability is identified due to interfacial denaturation or shear stresses, or due to solution conditions resulting in hydrophobic interactions leading to non-covalent irreversible aggregates, then the following excipients are typically screened and they have historically shown success with several products. Screen exclusion solutes such as disaccharides – sucrose or trehalose, PEG 400, glycerol, sorbitol etc., at various concentrations or weight ratios, a weight ratio of 1:1, protein:stabilizer serves the purpose and is generally recommended. Its effectiveness should be assessed by its ability to induce a change in the tertiary structure through circular dichroism spectrum and/or the unfolding transition midpoint Tu/Tm (°C) through differential scanning fluorimetry (DSF). The mechanism of protection behind it is believed to be due to be solutes are preferentially excluded from the protein domain, increasing the free energy of the unfolded system, thermodynamically this leads to folded form as it will have decreased free energy. This is graphically illustrated in the Fig. 8. From lyophilization point of view polyols, sorbitol, glycerol, and PEGs are not desirable, as they possess very low Tg’ values (see Table 1, Chapter “Characterization and Determination of Freeze-Drying Properties of Frozen Formulations: Case Studies”) and they significantly depress the overall Tg’ or collapse temperature of the formulation. Formulations containing these excipients will require long drying times as they need to be freeze dried at very low temperatures and in some cases not practically feasible in commercial setting. Disaccharides such as sucrose and trehalose are preferable as they relatively possess higher Tg’ values and practically feasible to freeze dry at commercial setting. Trehalose is preferred over sucrose from freeze drying point view, as it has few advantages over sucrose 1. It has higher Tg’ and Tg values compared to sucrose 2. It works as an effective stabilizer and does not hydrolyze to glucose and fructose over a wide range of pH, especially in the acidic range while sucrose hydrolyses to reducing sugars, glucose, and fructose in the acidic range that can form adduct with protein leading to glycation/Maillard reaction or browning of sugar, a discoloration in the cake appearance. However, one disadvantage of trehalose use is, it is 10 times more expensive that sucrose. As indicated above, while the main objective of the formulation design is to achieve both liquid stability (at least for few weeks to months at both room temperature and 2–8 °C) and lyophilized stability (≥ 2 years), due diligence needs to be paid while selecting the excipients so that both stability and ease of processing/manufacturability are achieved. Colloidal Instability (Native protein–protein interactions resulting in self-association) Colloidal instability can be due to hydrophobic and electrostatic interactions, which is briefly discussed below. Hydrophobic Interactions Leading to Non-covalent Aggregates If the colloidal instability is identified and the cause is believed to be due to hydrophobic interactions between the hydrophobic groups present on the surface of the protein leading to non-covalent aggregates, generally it is recommended to

Concepts and Strategies in the Design of Formulations for Freeze Drying

75

Fig. 9 Interaction parameter, kD, for mAb-G and mAb-R as a function of buffer pH (pH 5 to pH 8) at ionic strength of (a) 15 mM and (b) 150 mM. (Printed with permission, Ref. [57])

screen cyclodextrins. The two common forms of cyclodextrins that have demonstrated protective effect through solubilization are 2-hydroxypropyl-beta-cyclodextrin (HPCD) and the Sulfobutylether-β-Cyclodextrin (SBE-β-CD). Their solubilization and stabilization effect are believed to be due the presence of hydrophilic and hydrophobic groups and their ability to form inclusion bodies. Electrostatic Interactions When the attractive forces exceed the repulsive forces due to imbalance of charge distribution on the surface of the protein, self-association occurs resulting in the formation of reversible aggregates. In situations when there is a single pI (isoelectric point) and the pI is in the basic region, it is recommended to adjust pH 2–3 units away from pI. The further pH away from pI, i.e., towards low pH, the more net charge mAb carries, leading to stronger repulsive electrostatic interaction between molecules and higher would be the solubility/stability. Adding any salts or increasing the ionic strength in this situation will make the situation worse, as it is known to screen charge–charge interactions, and thus decrease the proximity energy [57]. Decrease in proximity energy has been proposed to be the cause of colloidal instability, and thus increase in aggregation [57, 60]. So, keeping the ionic strengths low ≤15–20 mM with pH 5 ish would be helpful as illustrated in Fig. 9. Thus, use of salt as a potential stabilizer of biopharmaceuticals requires careful assessment both from liquid stability and its lyophilization. Solutions containing salts especially sodium chloride (NaCl) are very difficult to lyophilize, as they depress the Tg’ of the formulation significantly and tend to crystallize over storage In situations where you have pI in the acidic range or more than one pI and spread over in the acidic and basic range, one needs to screen both pH and salts. On addition of standard salts like arginine HCl, phosphate, sulfate, and citrate, CaCL2, Ca (C2H3O2)2, MgCl2, and Hofmeister ions will increase electrostatic repulsion between protein molecules, suppress colloidal aggregation, and also help in solubilization. At high concentrations, the effect on solubility and conformational stability has

76

F. Jameel

Fig. 10 Hoffmeister series

been proposed to follow the Hoffmeister series [58, 59]. For anions, the stabilizing effect has been reported to increase as one moves from right to left in the series and for cations right to left as shown in Fig. 10. Combination of arginine hydrochloride with sodium glutamate is receiving much attention due to a seemingly synergetic effect when the two solutes are combined in an equimolar ratio [61]. Step 5: Screening of Bulking Agents Generally, the intended role of bulking agents in pharmaceuticals is to simply function as fillers to increase the density of the product cake, prevent product blow-out, enhance product elegance, and act as inert and not intended to provide enhanced chemical or physical stability of the drug substance. There are two situations where the inclusion of bulking agents in the formulation makes lyophilization easy and more efficient. Situation 1: When the quantity of drug per vial is extremely small With the advent of more potent drugs both in the small molecule world like peptides and large molecule world such as engineered antibodies (ADCs, DVDs, Diabodies, Bispecific, fabs, etc.) it is becoming increasingly important to consider addition of bulking agents to mitigate risks/challenges associated with processing, testing, and delivery systems. Dilute solutions containing 1% or less solid content is practically difficult to lyophilize with the retention of cake structure as during primary drying the flow of water vapor may create enough force on the fragile cake to break the cake structure and potentially carry some of the product out of the vial with the water vapor stream and this phenomenon is referred to as product blow-out [62]. In such situations inclusion of bulking agents becomes indispensable. Situation 2: Improve/enhance collapse temperature One of the main objectives of formulation development is to achieve the desirable stability of the product in the liquid state as well in the lyophilized state through the screening of excipients (buffer salts and stabilizers). At times it is not possible/ feasible to achieve the required stability with lyophilization-friendly excipients that have high collapse temperature but end up in choosing an excipient that has very low collapse temperature, which will result in long freeze drying cycle and may not be practically feasible in commercial setting. In such situations, although the total solid content may be adequate enough to freeze dry without product “blow-out,” addition of collapse temperature enhancers to the formulation composition in bulk quantities becomes very desirable from lyophilization point of view, as it will serve two purposes: (1) shorten freeze drying cycle significantly and (2) enhance product elegance. There are several excipients that serve as bulking agents and can be categorized into amorphous and crystalline. Amorphous Bulking Agents Some of the amorphous bulking agents such as sucrose (approx. –34 °C), trehalose (approx. –28 °C), lactose (approx –30 ° C), and PEGs (approx -50 °C) are not suitable as bulking agents, as they have low collapse temperature and therefore require low drying temperatures and long processing times. Although lactose is commonly used as a bulking agent in small molecules, besides its low collapse temperature, it is a reducing sugar and can form adducts with amine in protein products and must be questioned. Hydroxyethyl starch, which is used as a plasma expander in clinicals, is an inert amorphous excipient, which has high collapse temperature (-10 °C) and could function as an amorphous bulking agent without requiring long processing times. However, it is believed to undergo some cake shrinkage and cracking during drying and therefore and may provide less elegance to the cake.

Concepts and Strategies in the Design of Formulations for Freeze Drying

77

Fig. 11 Design space created using CFD and quasi-steady-state models to predict the equipment and process performance and guide operation

Disaccharides, sucrose, and trehalose are commonly used as stabilizers for protein products and are also used as amorphous bulking agents in situations when the drug quantity is less and/or need to adjust the osmolality for subcutaneous administration. However, they are not suitable for enhancing the collapse temperature of the formulations as they possess low collapse temperatures. Addition of any quantities of sucrose or trehalose more than minimum needed for stabilization of protein should be questioned. Since protein has a Tg’ of -10 °C while sucrose has a Tg’ of -34 °C, it would be prudent to have formulations where the protein concentration/weight ratio is higher than sucrose or trehalose as it is known to start contributing positively or enhance the overall collapse temperature in concentrations >20 mg/mL. Cyclodextrins like 2-hydroxypropyl-beta-cyclodextrin (HPCD) and the Sulfobutylether-β-Cyclodextrin (SBE-β-CD) serve as good amorphous bulking agents in combination with sucrose or trehalose as stabilizers. They possess high collapse temperature of approximately -9 °C and serve as collapse temperature enhancers in bulk quantities enhancing the overall collapse of the formulation [63]. For protein products a formulation with a combination of sucrose or trehalose as a stabilizer and either 2-hydroxypropyl-beta-cyclodextrin (HPCD) or Sulfobutylether-β-Cyclodextrin (SBE-β-CD) as a bulking agent will serve as a good amorphous system for lyophilization and elegant-looking cake structure. It must be noted that for a collapse temperature enhancer to work as a collapse temperature enhancer it should form a predominant component of the formulation, i.e., should be minimum 2× the other total amorphous components in the formulation. Crystallizable Bulking Agents The two most commonly used crystallizable bulking agents are mannitol and glycine, and they provide crystalline matrix to the cake. For their crystallization they should be the major solute component at least 2× greater than the sum of the concentrations of all other solute components They perform superior to amorphous bulking agents in many aspects as indicated below: • They possess high collapse temperature (eutectic temperature) in the range of –3 °C to –5 °C, which enables drying carried out at higher shelf temperature/pressure shortening the freeze drying cycle significantly. It does not collapse. • Provides mechanical strength to the cake and prevents cake from reducing to powder during transportation. • Provides elegant looking crystalline cake that reconstitutes easily. • Enables design of robust freeze drying cycles with larger design space offering wide flexibility in tech transfer and operations., see Fig. 11 Mannitol-based formulations besides offering several advantages over amorphous systems presents a couple of potential challenges; one relating to vial breakage and the other one is relating to formation of hydrate form that the process engineer should be aware of. The potential of vial breakage occurs if the freezing protocol is not designed properly to ensure its

78

F. Jameel

complete crystallization during the freezing phase. If the mannitol is not completely crystallized during freezing phase it will crystallize during primary drying phase as the product temperature warms up but still below 0 °C. Since crystallization process is an exothermic reaction and always accompanied by release of water which converts it into ice, both factors contribute to unequal expansion of glass causing the vial to break from the sides and/or from bottom (lensing) [64–66]. In order to avoid vial breakage the following steps are recommended: • Ensure first freezing is 10 °C below the Tg’ of the formulation and hold it for adequate time based on fill volume to ensure formation of nuclei followed by an annealing step (heat treatment). The optimal time and temperature for annealing step should be determined through freeze drying microscopy to ensure complete crystallization of mannitol • Avoid high fill depths and higher concentrations of mannitol Mannitol tends to form hydrate forms during freezing phase and hydrate crystals desolvate at temperatures ≥50 °C. In the author’s view it should be of little concern, as there are no practical implications on the product quality, since it is bound to the crystal lattice and not a free water unless the product is exposed to elevated temperatures of around 50 °C when it desolvates and release water compromising the stability of the product. Glycine functions equally well compared to mannitol and crystallizes easily to form an elegant product that reconstitutes quickly. One of the advantages of glycine is that it does not induce vial breakage. However, a glycine cake is believed to be slightly fragile than a mannitol cake, and is generally perceived as being relatively less elegant than a mannitol cake. Mannitol or glycine-based formulations are binary systems where both amorphous and crystalline components coexist. Stabilizers and other buffer species constitute minor amorphous component while mannitol or glycine will constitute the major crystalline system. Since mannitol or glycine constitutes the major component of the formulation and represents the major collapse of the system, the product can be dried above the collapse temperature of minor amorphous component and below the eutectic temperature of major crystalline component. Thus, freeze drying such a mannitol or glycine-based formulation amounts to freeze drying with microscopic or partial collapse (i.e., complete collapse of the amorphous phase) but cake structure is maintained by the crystalline component. Here, the drug form is that of an amorphous coating on the crystalline mannitol, with stability properties normally close to that of a system freeze dried without the mannitol [62]. This enables one to achieve both elegance and efficient drying with fast reconstitution time and low residual moisture content. Step 6: Upfront Manufacturability Assessment The two main causes for unsuccessful tech-transfer and manufacturing issues are designing a formulation without process considerations and a process design without the knowledge of manufacturing capabilities and limitations. They need to go hand-in-hand, sometimes the liability may be either better addressed through formulation design or may be through process conditions or may be combination of both. Hence, it is imperative that both formulation scientist and process engineer work together performing upfront manufacturability assessment of formulation candidates and this is especially true in the case of development of a lyophilized product. The objective of this study is to assess the compatibility of the formulation candidates and the manufacturing process to meet the TPP, and to identify and mitigate any liability by fine tuning formulation and/or process parameters as needed. The top 2–3 formulation candidates with optimal solution conditions should be subjected to unit operation of manufacturing (freezing and thawing, mixing, filtration, holding in stainless steel vessels or disposable bags, filling, lyophilization, and inspection) using scale-down models that are representative of manufacturing conditions to assess their manufacturability. Typically, the effects of processing conditions on the integrity of the molecule are studied at the lab scale, and these studies can potentially be misleading as they are not necessarily representative of or mimic the large-scale process. It is thus suggested here that scaled-down models be built that would be representative of large-scale unit operations and that these models be used to screen formulation candidates to identify the best ones and backup units. The lyophilized drug product vials should then finally be subjected to shipping conditions using the transportation simulation testing system (TSTS) before being placed on accelerated and real-time storage stability. For each formulation, both in-process and drug product samples from the scale-down studies be analyzed for relevant pCQAs. The final best and back-up commercial formulation recommendation should be made based on manufacturability and 6 months stability data from the manufacturability assessment study.

Concepts and Strategies in the Design of Formulations for Freeze Drying

79

5

Stability Testing

5.1

Kinetics of Degradation in the Amorphous Solid State

Generally, one thinks kinetics in terms of reaction order, which is true but applies only in liquid or gaseous solutions [67]; with solids, the elucidation of kinetics refers to the evaluation of the time dependence of the reaction rate, or alternatively, the time dependence of the loss of parent or appearance of degradation products. The degradation kinetics in the amorphous solid state, or glassy state, is often “stretched time” kinetics, meaning that the relationship between parent (or purity, P) and time is governed by the following equation, Pðt Þ = Pðt= 0Þ: exp - k:tβ




where k is the rate constant on the stretched timescale tβ and β is a constant between zero and unity. Amorphous solids are believed to follow stretched time kinetics as they are composed of distribution of microstates with each state having a distinct degradation rate, and these microstates are not in structural equilibrium [67–71]. Normally, one can find empirically, that β ≈ 1/2, and the kinetics is “square-root of time kinetics.”

5.2

Selection of Accelerated Testing Conditions, Temperature, and Time

When designing the accelerated stability studies, the selection of temperature becomes critical to the successful outcome otherwise it can be misleading. The highest test temperature in the study should be at least 10 °C below the glass transition temperature (Tg) of the dried solid, at the same time should be sufficiently high such that the level of degradation characteristic of that produced during the shelf life of the product is produced in a relatively short period of time. The difference between the test temperature and the anticipated storage temperature should not be so great that nonrepresentative results are obtained. If one has a system where two (or more) degradation pathways are significant, but these pathways are characterized by very different activation energies, for an example a system that undergoes both a hypothetical oxidation, with an activation energy of 15 kcal/mol, and a hypothetical deamidation reaction, with an activation energy of 25 kcal/mol one may find that the dominant reaction at the accelerated test condition is not even important at the actual temperature of interest. At high temperature, particularly above 50 °C, the reaction is predominantly a higher-activation-energy reaction, but in the range of refrigerated storage, deamidation is insignificant, and oxidation is the dominant reaction. Thus, if a formulation is being optimized using high-temperature stability data, optimization is being carried out for a reaction that does not occur to an appreciable extent at the intended storage temperature of 5 °C, and the test results may well be meaningless. Clearly, one needs to verify that the degradation products that are produced at the accelerated test conditions are essentially those, in roughly the same proportions, as produced at the intended storage temperature. Early in a project, this is difficult at best, and thus there is significant risk in accepting predictions based on accelerated test conditions run at very high temperatures. The time required for an accelerated test to produce the same level of a particular degradation product as would be found at the end of the shelf life at the intended product storage conditions may be estimated from the Arrhenius equation and assumed time dependence of the degradation process. Such estimates are provided in Fig. 9 where calculations were performed for both reactions with a low activation energy (15 kcal/mol) and a high activation energy (25 kcal/mol). Of course, the times are shorter with the higher activation energy, and are also much shorter with the square root of time kinetics than for exponential (effective First-order) degradation kinetics. Note that the equivalent times are relatively short, even when the activation energy is low and the difference between accelerated test temperature and intended storage temperature is only 10–15 °C, particularly for the square root of time kinetics. This is a significant point since degradation in glassy systems commonly follows the square root of time kinetics. The implication of this observation is that one does not necessarily need to employ accelerated test temperatures enormously higher than the intended storage temperature to ensure rapid turnaround of data. However, it must also be recognized that if the objective is to produce sufficient degradation to allow evaluation of a precise rate constant that, in turn, allows a quantitative assessment of differences in stability of trial formulations, one may need to produce significantly more degradation during the accelerated test than would result during the shelf life of a product. In such cases, the times would obviously need to be longer than indicated by Fig. 12.

80

F. Jameel A. Product Storage at 5°C Days Equivalent to 2 yrs at storage Temperature

1000

100

10

1

0.1

Days equivalent to 2 yrs at storage Temperature

0

10

20

30

40

50

B. Product Storage Temperature, 25°C 1000

100

10

1

0.1 20

30

40

50

60

70

Accelerated Test Temperature, °C

Fig. 12 Equivalent Times for Accelerated Stability Testing for Refrigerated (5 °C) and Room Temperature (25 °C) Storage. Symbols Key: Diamonds = 15 kcal/mol, Circles = 25 kcal/mol; Filled symbols = exponential kinetics (1st order), Open symbols = square root of time kinetics. (a) Product storage at 5 °C, (b) Product storage temperature, °C

References 1. Pikal MJ, Lukes AL, Lang JE. Thermal decomposition of amorphous beta-lactam antibacterials. J Pharm Sci. 1977;66:1312–6. 2. Pikal MJ, Lukes AL, Lang JE, Gaines K. Quantitative crystallinity determinations of beta-lactam antibiotics by solution calorimetry: correlations with stability. J Pharm Sci. 1978;67:767–73. 3. Pikal MJ, Dellerman KM. Stability testing of pharmaceuticals by high-sensitivity isothermal calorimetry at 25C: cephalosporins in the solid and aqueous solution states. Int J Pharm. 1989;50:233–52. 4. Gervais A, Angiuoni G, O’Hara J, Juul Jensen K, Sewerin K, Rossi M, Nedved M, Dillon P, Uddin S, Cornen S, Schnaible V. Forced degradation studies for therapeutic proteins. Brussels: European Biopharmaceutical Enterprises; 2015. 5. Siedler M, Kumar V, Chari R, Saluja S, Fraunhofer W. Development of drug product formulations: molecular design and early candidates screening. In: Jameel F, Hershenson S, Khan MA, Martin-Moe S, editors. Quality by design for biopharmaceutical drug product development. New York: Springer New York; 2015. p. 61–85. 6. Perez-Ramírez B, Guziewicz N, Simler R, Sreedhara A. Approaches for early developability assessment of proteins to guide quality by design of liquid formulations. In: Jameel F, Hershenson S, Khan MA, Martin-Moe S, editors. Quality by design for biopharmaceutical drug product development. New York: Springer New York; 2015. p. 87–114. 7. Jarasch A, Koll H, Regula JT, Bader M, Papadimitriou A, Kettenberger H. Developability assessment during the selection of novel therapeutic antibodies. J Pharm Sci. 2015;104(6):1885–98. 8. Folzer E, Diepold K, Bomans K, Huwyler J, Mahler H-C, Koulov AV. Selective oxidation of methionine and tryptophan residues in a therapeutic IgG1 molecule. J Pharm Sci. 2015;104(9):2824–31. 9. Hermeling S, Crommelin DJ, Schellekens H, Jiskoot W. Structure-immunogenicity relationships of therapeutic proteins. Pharm Res. 2004;21 (6):897–903. 10. Liu D, Ren D, Huang H, Dankberg J, Rosenfeld R, Cocco MJ, Li L, Brems DN, Remmele RL Jr. Structure and stability changes of human IgG1 Fc as a consequence of methionine oxidation. Biochemistry. 2008;47(18):5088–100. 11. Houde D, Peng Y, Berkowitz SA, Engen JR. Post-translational modifications differentially affect IgG1 conformation and receptor binding. Mol Cell Proteomics. 2010;9(8):1716–28.

Concepts and Strategies in the Design of Formulations for Freeze Drying

81

12. Bertolotti-Ciarlet A, Wang W, Lownes R, Pristatsky P, Fang Y, McKelvey T, Li Y, Li Y, Drummond J, Prueksaritanont T, Vlasak J. Impact of methionine oxidation on the binding of human IgG1 to Fc Rn and Fc gamma receptors. Mol Immunol. 2009;46(8–9):1878–82. 13. Pan H, Chen K, Chu L, Kinderman F, Apostol I, Huang G. Methionine oxidation in human IgG2 Fc decreases binding affinities to protein A and FcRn. Protein Sci. 2009;18(2):424–33. 14. Zhang A, Hu P, MacGregor P, Xue Y, Fan H, Suchecki P, Olszewski L, Liu A. Understanding the conformational impact of chemical modifications on monoclonal antibodies with diverse sequence variation using hydrogen/deuterium exchange mass spectrometry and structural modeling. Anal Chem. 2014;86(7):3468–75. 15. Wang W, Vlasak J, Li Y, Pristatsky P, Fang Y, Pittman T, Roman J, Wang Y, Prueksaritanont T, Ionescu R. Impact of methionine oxidation in human IgG1 Fc on serum half-life of monoclonal antibodies. Mol Immunol. 2011;48(6–7):860–6. 16. Pace AL, Wong RL, Zhang YT, Kao Y-H, Wang YJ. Asparagine deamidation dependence on buffer type, pH, and temperature. J Pharm Sci. 2013;102(6):1712–23. 17. Wakankar AA, Borchardt RT. Formulation considerations for proteins susceptible to asparagine deamidation and aspartate isomerization. J Pharm Sci. 2006;95(11):2321–36. 18. Vlasak J, Bussat MC, Wang S, Wagner-Rousset E, Schaefer M, Klinguer-Hamour C, Kirchmeier M, Corvaïa N, Ionescu R, Beck A. Identification and characterization of asparagine deamidation in the light chain CDR1 of a humanized IgG1 antibody. Anal Biochem. 2009;392(2):145–54. 19. Wedemeyer WJ, Welker E, Narayan M, Scheraga HA. Disulfide bonds and protein folding. Biochemistry. 2000;39(15):4207–16. 20. Wei B, Berning K, Quan C, Zhang YT. Glycation of antibodies: modification, methods and potential effects on biological functions. MAbs. 2017;9(4):586–94. 21. Mitra N, Sinha S, Ramya TN, Surolia A. N-linked oligosaccharides as outfitters for glycoprotein folding, form and function. Trends Biochem Sci. 2006;31:156–63. [PubMed] [Google Scholar]. 22. Hanson SR, et al. The core trisaccharide of an N-linked glycoprotein intrinsically accelerates folding and enhances stability. Proc Natl Acad Sci U S A. 2009;106:3131–6. 23. Bosques CJ, Imperiali B. The interplay of glycosylation and disulfide formation influences fibrillization in a prion protein fragment. Proc Natl Acad Sci U S A. 2003;100:7593–8. 24. Skropeta D. The effect of individual N-glycans on enzyme activity. Bioorg Med Chem. 2009;17:2645–53. 25. Liu H, Ponniah G, Zhang H-M, Nowak C, Neill A, Gonzalez-Lopez N, Patel R, Cheng G, Kita AZ, Andrien B. In vitro and in vivo modifications of recombinant and human IgG antibodies. MAbs. 2014;6(5):1145–54. https://doi.org/10.4161/mabs.29883. 26. Ratanji KD, Derrick JP, Dearman RJ, Kimber I. Immunogenicity of therapeutic proteins: influence of aggregation. J Immunotoxicol. 2014;11(2): 99–109. 27. Salinas BA, Sathish HA, Bishop SM, Harn N, Carpenter JF, Randolph TW. Understanding and modulating opalescence and viscosity in a monoclonal antibody formulation. J Pharm Sci. 2010;99(1):82–93. 28. Roberts CJ, Das TK, Sahin E. Predicting solution aggregation rates for therapeutic proteins: approaches and challenges. Int J Pharm. 2011;418 (2):318–33. 29. Velev OD, Kaler EW, et al. Protein interactions in solution characterized by light and neutron scattering: comparison of lysozyme and chymotrypsinogen. Biophys J. 1998;75(6):2682–97. 30. Tessier PM, Lenhoff AM, et al. Rapid measurement of protein osmotic second virial coefficients by self-interaction chromatography. Biophys J. 2002;82(3):1620–31. 31. Moon YU, Curtis RA, et al. Protein-protein interactions in aqueous ammonium sulfate solutions. Lysozyme and bovine serum albumin (BSA). J Solut Chem. 2000;29(8):699–717. 32. Alford JR, Kendrick BS, et al. Measurement of the second osmotic virial coefficient for protein solutions exhibiting monomer-dimer equilibrium. Anal Biochem. 2008;377(2):128–33. 33. Atul Saluja R, Fesinmeyer M, Hogan S, Brems DN, Gokarn YR. Diffusion and sedimentation interaction parameters for measuring the second virial coefficient and their utility as predictors of protein aggregation. Biophys J. 2010;99(8):2657–65. 34. Thiagarajan G, Semple A, James JK, Cheung JK, Shameem M. A comparison of biophysical characterization techniques in predicting monoclonal antibody stability. MAbs. 2016;8(6):1088–97. 35. Connolly BD, Petry C, Yadav S, Demeule B, Ciaccio N, Moore JMR, Shire SJ, Gokarn YR. Weak interactions govern the viscosity of concentrated antibody solutions: high-throughput analysis using the diffusion interaction parameter. Biophys J. 2012;103(1):69–78. 36. Pace CN. Conformational stability of globular-proteins. Trends Biochem Sci. 1990;15(1):14–7. 37. Li W, Prabakaran P, Chen W, Zhu Z, Feng Y, Dimitrov DS. Antibody aggregation: insights from sequence and structure. Antibodies. 2016;5(3): 19. 38. https://www.americanpharmaceuticalreview.com/25604-Pharmaceutical-Particle-Size-Analyzers/12040038-Zetasizer-Nano-ZSP-System/ 39. Nema S. Key formulation challenges of protein (mAb) drugs, Pfizer, http://users.unimi.it/gazzalab/wordpress/wp-content/uploads/2011/12/9Key-formulation-challenges-of-protein-drugs 40. Larsen SS. Studies on stability of drugs in frozen systems. Arch Pharm Chem Sci Ed. 1973;1:41–53. 41. Murase N, Franks F. Salt precipitation during the freeze concentration of phosphate buffer solutions. Biophys Chem. 1989;34:293–300. 42. Gomez G, Rodriguez-Hornedo N, Pikal MJ. Effect of freezing on the pH of sodium phosphate buffer solutions. Pharm Res. 1994;11:S-265, PPD 7364. 43. Szkudlarek BA, Rodriguez-Hornedo N, Pikal MJ. Analysis of pH changes of potassium phosphate buffer salt solutions during freezing. Pharm Res. 1994;11:S-228, PPD 7215. 44. Pikal MJ, Dellerman KM, Roy ML, Riggin RM. The effects of formulation variables on the stability of freeze dried human growth hormone. Pharm Res. 1991;8:427–36. 45. Gomez G, Pikal MJ, Rodriguez-Horned N. Effect of initial buffer composition on pH changes during far-from equilibrium freezing of sodium phosphate buffer solutions. Pharm Res. 2001;18:90–7. 46. Milton N. Eli Lilly, presentation in IIR, Sept, 2005. 47. Mahler H-C, Friess W, Grauschopf U, Kiese S. Protein aggregation: pathways, induction factors and analysis. J Pharm Sci. 2009;98(9):2909–34.

82

F. Jameel

48. Narhi L. Characterization and biological relevance of protein aggregates and other particles 100–200,000 nm in size (sub micron and subvisible), https://www.casss.org/resource/resmgr/hos_speaker_slides/2019_narhi_linda_slides 49. Das TK. Protein particulate detection issues in biotherapeutics development--current status. AAPS PharmSciTech. 2012;13(2):732–46. 50. Thirumangalathu R, Krishnan S, Ricci MS, Brems DN, Randolph TW, Carpenter JF. Silicone oil- and agitation-induced aggregation of a monoclonal antibody in aqueous solution. J Pharm Sci. 2009;98(9):3167–81. 51. Lee HJ, McAuley A, Schilke KF, McGuire J. Molecular origins of surfactant-mediated stabilization of protein drugs. Adv Drug Deliv Rev. 2011;63:1160–71. 52. Barn NB, Cleland JL, Yang J, Manning MC, Carpenter JF, Kelley RF, Randolph TW. Tween protects recombinant human growth hormone against agitation-induced damage via hydrophobic interactions. J Pharm Sci. 1998;87:1554–9. 53. Barn NB, Randolph TW, Cleland JL. Stability of protein formulations: investigation of surfactant effects by a novel EPR spectroscopic technique. Pharm Res. 1995;12:2–11. 54. Kerwin BA, Heller MC, Levin SH, Randolph TW. Effects of tween 80 and sucrose on acute short-term stability and long-term storage at-20 degrees of a recombinant haemoglobin. J Pharm Sci. 1998;87:1062–8. 55. Narang AS, Rao VM, Desai DS. Effect of antioxidants and silicates on peroxides in povidone. J Pharm Sci. 2012;101(1):127–39. 56. Akers MJ. Excipient-drug interactions in parenteral formulations. J Pharm Sci. 2002;91(11):2283–300. 57. Pindrus M, Shire SJ, Kelley RF, Demeule B, Wong R, Yiren X, Yadav S. Solubility challenges in high concentration monoclonal antibody formulations: relationship with amino acid sequence and intermolecular interactions. Mol Pharm. 2015;12(11):3896–907. 58. Gokarn YR, Matthew Fesinmeyer R, Saluja A, Razinkov V, Chase SF, Laue TM, Brems DN. Effective charge measurements reveal selective and preferential accumulation of anions, but not cations, at the protein surface in dilute salt solutions. Protein Sci. 2011;20:580–7. 59. Fesinmeyer RM, Hogan S, Saluja A, Brych SR, Kras E, Narhi LO, Brems DN, Gokarn YR. Effect of ions on agitation- and temperature-induced aggregation reactions of antibodies. Pharm Res. 2009;26:903–13. 60. Laue T. Proximity energies: a framework for understanding concentrated solutions. J Mol Recognit. 2012;25:165–73. 61. Schneider CP, Shukla D, Trout BL. Arginine and the Hofmeister series: the role of ion-ion interactions in protein aggregation suppression. J Phys Chem B. Author manuscript; available in PMC 2012 Jun 9. Published in final edited form as: J Phys Chem B. 2011;115(22):7447–7458. 62. Pikal MJ. Lyophilization. In: Swarbrick J, Boylan J, editors. Encyclopedia of pharmaceutical technology. New York: Marcel Dekker; 2002. p. 1299–326. 63. Haeuser C, Goldbach P, Huwyler J, Friess W, Allmendinger A. Be aggressive! Amorphous excipients enabling single-step freeze-drying of monoclonal antibody formulations. Pharmaceutics. 2019;11:616. 64. Williams NA, Dean T. Vial breakage by frozen mannitol solutions: correlation with thermal characteristics and effect of sterioisomerism, additives, and vial configuration. J Parenter Sci Technol. 1991;45:94–100. 65. Williams NA, Lee Y, Polli GP, Jennings TA. The effects of cooling rate on solid phase transitions and associated vial breakage occurring in frozen mannitol solutions. J Parenter Sci Technol. 1986;40(135):71. 66. Williams NA, Guglielmo J. Thermal mechanical analysis of frozen solutions of mannitol and some related steroisomers: evidence of expansion during warming and correlation with vial breakage during lyophilization. J Parenter Sci Technol. 1993;47:119–23. 67. Pikal MJ, Rigsbee D, Roy ML, Galreath D, Kovach KJ, Wang W, Carpenter JF, Cicerone MT. Solid state chemistry of proteins: II. The correlation of storage stability of freeze-dried human growth hormone (hGH) with structure and dynamics in the glassy solid. J Pharm Sci. 200;97(12):5106–21. 68. Pikal MJ, Rigsbee DR. The stability of insulin in crystalline and amorphous solids: observation of greater stability for the amorphous form. Pharm Res. 1997;14:1379–87. 69. Yoshioka S, Aso Y, Kojima S. Usefulness of the Kohlrausch-Williams-Watts stretched exponential function to describe protein aggregation in lyophilized formulations and the temperature dependence near the glass transition temperature. Pharm Res. 2001;18:256–60. 70. Abdul-Fattah AM, Dellerman K, Bogner RH, Pikal MJ. The effect of annealing on the stability of amorphous solids: chemical stability of freezedried moxalactam. J Pharm Sci. 2007;96:1237–50. 71. Cicerone MT, Soles CL, Chowdhuri Z, Pikal MJ, Chang LL. Fast dynamics as a diagnostic for excipients in preservation of dried proteins. Am Pharm Rev. 2005;8(6):24–7.

Formulation Design for Freeze-Drying: Case Studies of Stabilization of Proteins Andrea Allmendinger, Christina Häuser, Lokesh Kumar, and Ilona Vollrath

Abstract

Formulation design is an integral part of drug product development for parenterally applied dosage forms. In particular for freeze-dried formulations, the choice of excipients is crucial and directly influences lyophilization cycle performance and ultimately the product quality. The current chapter focuses on the development of freeze-dried protein formulations, such as monoclonal antibody formulations. Vaccines, oligonucleotides, and gene and cell therapeutic products are not within the scope of this chapter. Typical excipients for parenteral freeze-dried biopharmaceuticals and their functions are summarized and examples of recent commercial formulations are provided. The stabilization mechanism during freezing and thawing, molecular dynamics in the solid state, critical quality attributes and the link between product attributes and lyophilization cycle performance are discussed, and general rules for lyophilization dependent on formulation parameters are exemplified in case studies. Keywords

Monoclonal antibody · Excipients · Lyophilization · Freeze-drying · Residual moisture · Reconstitution time · Cake appearance · Solid state dynamics · Protein to sugar ratio

1

Excipient Selection

Excipients are therapeutically nonactive formulation components, which stabilize the protein in the formulation matrix. The choice of excipients is crucial, as it directly influences the critical quality attributes of the final drug product. Excipients need to be nontoxic, nonimmunogenic, and safe in the applied clinical doses. Excipients must be approved for the intended route of administration and must comply with compendial requirements. This is particularly relevant for biopharmaceuticals, which are typically administered through parenteral routes, which limits the choice of excipients. In addition, excipients should be stable during manufacturing, transport, and throughout the drug product’s shelf-life, ideally with them being easy to handle [1].

Andrea Allmendinger, Christina Häuser, Lokesh Kumar and Ilona Vollrath contributed equally with all other contributors. A. Allmendinger (✉) · C. Häuser · I. Vollrath Pharmaceutical Development, Pharmaceutical Technical Development Biologics Europe, F. Hoffmann-La Roche Ltd, Basel, Switzerland L. Kumar Pharmaceutical Development, Genentech-Roche, South San Francisco, CA, USA # America Association of Pharmaceutical Scientists 2023 F. Jameel (ed.), Principles and Practices of Lyophilization in Product Development and Manufacturing, AAPS Advances in the Pharmaceutical Sciences Series 59, https://doi.org/10.1007/978-3-031-12634-5_5

83

84

A. Allmendinger et al.

Table 1 Summary of typical components in lyophilized monoclonal antibody formulations Function Stabilizer Buffer system Surfactant Antioxidant Cryoprotectant Lyo-protectant Collapse temperature modifier Bulking agenta

Type of stress Shift in pH Interfacial stress Oxidation Freezing (cryoconcentration) Drying Drying, Storage

Tonicifiera Viscosity reducing agent Preservative Acceleration of reconstitution a

Examples

Caveats specific to lyophilization process

Histidine buffer, citrate buffer Polysorbate 20, Polysorbate 80, Solutol, Poloxamer 188 Methionine, chelators (e.g., EDTA, DPTA) Sucrose, Trehalose

Crystallization of buffer components: e.g., phosphate buffer –

2-Hydroxypropyl-betacyclodextrin, Polyvinylpyrroldione, Dextran Mannitol, Sucrose, Trehalose, Dextran Sucrose, Trehalose, NaCl Arginine derivatives Benzyl alcohol Tert-butyl alcohol

– Excipient crystallization: e.g., trehalose; Maillard reaction with reducing sugars: e.g., lactose Cake collapse due to low glass transition temperature (related to process parameters) Maillard reaction of reducing sugars or terminal glucose ends of macromolecules (e.g., Dextran) Crystalline polymorph (detrimental to protein stability): e.g., mannitol; Different solid forms requiring annealing: e.g., mannitol; Vial breakage: e.g., mannitol Refer to cryoprotectants – Adsorption to interfaces: e.g., benzyl alcohol –

Excipients may have several functions

1.1

Role of Excipients

Table 1 provides an overview of excipients typically used in recent freeze-dried formulations of monoclonal antibodies, including their function, concrete examples, and caveats. Additional examples can be found in the mini-compendium of excipients in biopharmaceutical products published by Constantino and Pikal [2]. In general, the role of excipients comprises protein stabilization, manufacturability of the product (bulking agent, viscosity reducing agent), assurance of tolerability in humans (buffer system, tonicifier), assurance of correct and safe drug administration, as well as providing in-use stability (preservatives, acceleration of reconstitution, surfactants, viscosity reducing agent). Importantly, excipients may perform several functions simultaneously. In particular for freeze-dried formulations, the choice of excipients is also linked to lyophilization process performance, as discussed in the following sections. The quantitative composition of a formulation, such as the total solid content or the protein to sugar ratio, in combination with the fill volume, significantly influences the lyophilized cake resistance and ultimately lyophilization process performance and resulting critical quality attributes, such as cake appearance, residual moisture, and reconstitution time.

1.1.1 Protein Stabilization For freeze-dried formulations, excipients first need to stabilize the protein in the solution before lyophilization, then during the freeze-drying process, and finally during the drug product transport and storage in the dried state. Buffers are utilized to protect against significant pH change, which might otherwise compromise protein folding and thus bioactivity. Protein formulations are typically prepared several pH units away from their isoelectric point to avoid protein aggregation based on electric repulsion. In many cases, the selection of pH is a compromise between physical and chemical stability as screened during formulation development activities, along with physiological tolerability, as outlined further below. For monoclonal antibodies, the pH of a typical formulation ranges between 5 and 7. Additionally, for freeze-dried formulations, buffer systems should not crystallize during freezing, thereby losing their function, as in the case of phosphate or succinate buffers [2–4]. Surfactants are amphiphilic molecules and an important component of protein formulations. They protect the protein against various types of interfacial stress by preferentially occupying the same competing sites, thus protecting the protein. Surfactants prevent protein adsorption during the fill-finish process in the liquid state, for example to sterile filters or tubing, and protect against perturbation at the liquid–air interface. During the freezing and drying process, as well as during subsequent storage, surfactants are of minor relevance for freeze-dried formulations, and can thus be used in lower

Formulation Design for Freeze-Drying: Case Studies of Stabilization of Proteins

85

concentrations compared to equivalent liquid formulations. For parenteral dosage forms, there are only a small number of approved surfactants (e.g., polysorbate 20, polysorbate 80, solutol, and poloxamer 188). Each of them presents its particular challenges. Polysorbates are prone to oxidative and hydrolytic degradation, the latter likely caused by enzymes, which are residual host cell proteins from the drug substance manufacturing process [5–7] and poloxamer may form visible particles with protein aggregates and silicone, as has been recently reported [8]. As these are long-term events and typically occur over the product’s shelf-life in the liquid state, surfactant degradation is of minor relevance for freeze-dried formulations. However, surfactants are important during drug administration in the case of further dilution of the product into infusion bags, in order to prevent protein adsorption; this scenario is elaborated upon further below. Oxidation events are accelerated in liquid solution compared to freeze-dried formulations. Nevertheless, antioxidants such as methionine may be added to the formulation to protect the protein against oxidative stress, including during the fill-finish process as a result of potential residual hydrogen peroxide from decontamination cycles in the manufacturing area. Antioxidants may also be added to protect other excipients from oxidation. In freeze-dried formulations, excipients need to stabilize the protein during both the freezing and drying processes. While cryo-concentration to the maximum freeze concentrate leads to both molecular crowding and interfacial stress through the generation of new ice/formulation interfaces, the drying step creates stress on the protein by sublimation and desorption of the ice layer or remaining adsorbed water. In particular, the disruption of hydrogen bonds between the water/ice molecules and the protein may lead to protein unfolding. The most common excipients established over the past decades for protein stabilization are sugars, such as sucrose and trehalose. These excipients are thought to stabilize the protein by preferential exclusion, vitrification, and water replacement, as elaborated upon in subsequent sections. The use of reducing sugars is discouraged due to their ability to glycate the protein [9]. Excipients in lyophilized formulations influence the glass transition temperature of the maximally freeze-concentrated liquid (Tg’), the collapse temperature (Tc) during the lyophilization cycle, and the glass transition temperature of the freeze-dried product after lyophilization (Tg) and are thus relevant also from a processing perspective. Tg’ and Tc help optimize lyophilization cycle time obtaining short lyophilization cycles while maintaining lyophilized cake integrity [10]. Tg is relevant for product storage temperature, as outlined in more detail in Sect. 3. While trehalose is preferred over sucrose in terms of its higher Tg’/Tg, it poses challenges due to its potential for excipient crystallization in the frozen state, e.g. during drug substance storage [11]. To further reduce lyophilization cycle time and to allow a higher product storage temperature while maintaining protein stability and elegant cake appearance, excipients such as 2-Hydroxypropyl-beta-cyclodextrin, polyvinylpyrrolidone, or dextran in mixtures with cro-/lyoprotectants such as sucrose have been explored for the purpose of increasing the glass transition and collapse temperatures (referred to as Collapse Temperature Modifiers in Table 1) [10, 12, 13]. Although dextran is a promising excipient in terms of maintaining cake appearance, it has been shown to lead to the glycation of proteins, particularly during storage at elevated temperatures [13].

1.1.2 Manufacturability Cryo- and lyoprotectants are used in sufficiently high concentrations to stabilize proteins while yielding an intact lyophilized cake. However, in case of a low drug load, bulking agents may be required to produce an elegant product. Bulking agents can be sugars such as sucrose or trehalose, but other excipients may be preferred due to favorable process performance. For example, mannitol is often used, which produces elegant cakes and enables fast and easy freeze-drying cycles. However, freezs-drying of mannitol can lead to different polymorphs including amorphous and crystalline forms, with the latter impacting protein stability. An annealing step is thus recommended when using mannitol-based formulations to allow for conversion of one in the other. Furthermore, it has been recently shown that high fill volumes, fast freezing rates, and high mannitol concentrations dependent on the target freezing temperature of the mannitol-based formulations may lead to vial breakage, which can be avoided through process parameter optimization, such as annealing [14, 15]. For concentrated protein formulations, the addition of viscosity lowering agents may be required to avoid challenges during sterile filtration, filling, or subsequent drug administration, but also during the lyophilization process due to increased viscous flow and cake resistance. High viscosity is generated by molecular crowding and particularly by non-covalent protein–protein interactions. These are typically dominated by long-term, and in many cases electrostatic interactions at moderate protein concentrations (e.g., ~50 mg/mL), transitioning to the domination of short-term interactions, such as hydrophobic interactions at higher protein concentrations (e.g., ~200 mg/mL). The former can be screened through the addition of different counter ions, and the latter by the addition of amino acids. Arginine salts and derivatives have been demonstrated to be most effective [16–18].

86

A. Allmendinger et al.

1.1.3 Tolerability Parenteral products are typically formulated at isotonic osmolality and euhydric pH for best tolerability in patients. Tolerability, including from pain during injection, depends on the route of administration, and exemptions have been recently reviewed by Roethlisberger and colleagues [19]. As outlined above, euhydric conditions are adjusted by the buffer system in compromise with physical and chemical stability. In order to maintain an acid-base balance, buffer systems are abandoned for larger volumes of intravenously infused formulations due to the self-buffering capacity of the blood. Isotonicity can be obtained by different excipients. However, they are in many cases selected based on one or several additional functions. For freeze-dried formulations in particular, cryo- and lyoprotectants or bulking agents do thus qualify as tonicifiers. 1.1.4 Drug Administration Parenteral products are typically supplied as single-use dosage forms due to microbial growth considerations. Nevertheless, drug products may also be supplied as multidose vials if potential microbial contamination during multiple withdrawl from the primary packaging container is prevented by either the addition of preservatives [20], by the design of the primary packaging containe, or by immediate intended user. The preparation of freeze-dried products comprises reconstitution of the product under aseptic conditions, followed by either direct injection or further dilution into infusion bags. The reconstitution time should be fast, i.e. immediate or within a few minutes, and thus be convenient for the end-user, ideally without the requirement of additional devices. In case of challenges with achieving a fast reconstitution time, an initial thorough characterization of cake structure and a close look at lyophilization process parameters and cycle performance is recommended. In addition, reconstitution can be accelerated by a reduction of headspace pressure after lyophilization. If required, reconstitution can be further expedited by the addition of excipients such as tert-butyl alcohol [21]. In case of injection, the reconstituted formulation needs to be syringeable through typically 25–27G (subcutaneous) or 31G (ocular) needles, which may require the viscosity reducing agents as detailed above. The addition of surfactants is essential to prevent protein adsorption to the administration material during infusion.

1.2

Examples of Commercial Lyophilized Drug Products

To exemplify the above-mentioned choice of excipients and their roles, Table 2 provides a non-exhaustive list of marketed formulations of freeze-dried biopharmaceuticals within the scope of this chapter. A complete list of parenteral, lyophilized products, including biological molecules different from monoclonal antibodies and prior to 2004 can be found in a previous publication by Constantino and Pikal [2]. In particular, formulation components are highlighted in Table 2, providing examples for excipients as listed in Table 1. In addition, formulation of the first biosimilars entering the European and US markets for two references products, Herceptin® and Remicade®, are summarized. Table 2 shows that the majority of commercial antibody-based formulations comprise a histidine or phosphate buffer at pH values between 5.0 and 7.2 in combination with sucrose or trehalose as a cryo/lyoprotectant/tonicifer, and a polysorbate as a stabilizer. Some formulations are composed of mannitol, glycine, or dextran as a bulking agent. Interestingly, biosimilar formulations of the same reference product are not identical, and for example include stabilizers such as sorbitol and glycine as in the case of Infliximab Biosimilars Zessly® and Ixifi®.

2

In-Process and Storage Stability of Proteins: Stabilization During Freezing, Drying, and Storage

2.1

Theoretical Considerations

The stabilizing mechanism of proteins during lyophilization and subsequent storage in the dried state can generally be described by following three concepts: • Water replacement theory describes the thermodynamic stabilization of a protein’s native structure. According to this theory, excipients interact with the protein to form hydrogen bonds, thereby acting as a substitute for water molecules being removed during drying. A monolayer of the excipient around the protein’s surface is required to replace water at all hydrogen bonding sites in order to sufficiently stabilize the protein. A minimum molar ratio of excipient to protein of 360:1 is generally required to ensure protein stability in freeze-dried products [49, 50].

Cimizia [27]

Ilaris [28]

Keytruda [29]

Cosentyx [30]

Nucala [31]

Certolizumab pegola Canakinumab

Pembrolizumab

Secukinumab

Mepolizumab

Samsung Bioepis

Celltrion

Ontruzant [36]

Herzuma [37]

Full set or subset of indications of Herceptin (reference product)

21

Cancer

Biosimilar of Trastuzumabc

20

Cancer

Enhertu [34] Daiichi Sankyo famTrastuzumab Deruxtecannxki (C) Biosimilars and reference products Herceptin Genentech/ Trastuzumabc [35] Roche

21

21

1

Cancer

20

100

150

Pfizer (EU) /[5– 7]Wyeth (US)

Asthma

Inflammatory disease

25

150

Cancer

GlaxoSmithKline

Novartis

Cancer

Inflammatory disease

200

125

Asthma

Inflammatory disease

100

Respiratory tract disease

4

Genentech/ Roche

(B) Antibody-Drug conjugates Kadcyla [32] adoTrastuzumab Emtansine Gemtuzumab Mylotarg [33] ozogamicin

Novartis

Xolair [26]

Omalizumab

Merck Sharpe Dohme

Abbvie (EU), MedImmune (US) Novartis (EU)/ Genentech/ Roche (US) UCN

Acute organ rejection

Novartis

Synagis [25]

Indication class

Palivizumab

Active Product name (A) Monoclonal antibodies Basiliximab Simulect [24]

Authorization holder US/EU

L-Histidine-HCl
H2O (0.48), L-Histidine (0.31)

6.0

6.0

6.0

L-Histidine-HCl
H2O (0.31), L-Histidine (0.47)

α,α-Trehalose
2H2O (19), Polysorbate 20 (0.09) α-α-Trehalose
2H2O (19), Polysorbate 20 (0.08)

Dextran 40 (9.1), Sucrose (16), NaCl (5.8) Sucrose (90), Polysorbate 80 (0.38)

–d

5.5

Sucrose (60), Polysorbate 20 (0.2)

Sucrose (92), Polysorbate 80 (0.60) Sucrose (160), Polysorbate 80 (0.67), 5.0

7.0

5.8

5.5

–d

5.2

Sucrose (100), Polysorbate (0.1) Sucrose (92), Polysorbate 80 (0.60) Sucrose (70), Polysorbate 80 (0.20)

Sucrose (108), Polysorbate 20 (0.37)

–d

–d

NaCl (0.32), Sucrose (4), Mannitol (16), Glycine (8) Mannitol (56), Glycine (0.23)

Stabilizer (mg/mL)

–d

pH

L-Histidine (0.89), L-Histidine-HCl
H2O (4.0)

NaH2PO4 (0.6), Na2HPO4
H2O (0.1)

Sodium succinate (1.6)

L-Histidine
HCl
H2O/ L-Histidine (4.7) Na2HPO4
7H2O (7.1)

L-Histidine
HCl
H2O (1.7), L-Histidine (2.8) L-Histidine (1.6), HCl q. s., NaOH q.s.

L-Histidine-HCl
H2O (2.1), L-Histidine (1.3) Lactic acid (0.9)

L-Histidine (7.3)

KH2PO4 (1.4), Na2HPO4 (0.2)

Formulation composition Protein concentration Buffer system (mg/mL) (mg/mL)b

Table 2 Formulation examples of lyophilized monoclonal antibody products including biosimilars derived from FDA and EMA drug search databases [22, 23]

15.11.2017 EMEA/ H/C/004323

28.08.2000 EMEA/ H/C/000278

18.01.2021 EMEA/ H/C/005124

19.04.2018 EMEA/ H/C/004204

15.11.2013 EMEA/ H/C/002389

01.10.2009EMEA/H/ C/001037 23.10.2009 EMEA/ H/C/001109 (positive opinion) 28.02.2021 EMEA/ H/C/003820 14.01.2015 EMEA/ H/C/003729 01.12.2015 EMEA/ H/C/003860

25.10.2005 EMEA/ H/C/000606

13.8.1999 EMEA/H/C/000257

09.10.1998 EMEA/H/C/000207

EMA approval date, Product number

(continued)

18.01.2019 761100

25.09.1998 103792

20.12.2019 761139

01.09.2017 761060

22.02.2013 125427

21.01.2015 125504 04.11.2015 125526

22.04.2008 125160 17.06.2009 125319 04.09.2014 125514

20.06.2003 103976

19.6.1998 103770

12.5. 1998 103764

FDA approval date, BLA number

Formulation Design for Freeze-Drying: Case Studies of Stabilization of Proteins 87

10

Sandoz

Pfizer

Amgen

Ixifi [47]

Avsola [48]

Sucrose (25), Polysorbate 80 (0.05)

6.0

7.2

Sucrose (50), Polysorbate 80 (0.05)

Sucrose, Polysorbate 80

–d

6.2

–d

Sucrose (50), Polysorbate 80 (0.05) Sucrose, Polysorbate 80 Sucrose (50), Polysorbate 80 (0.05)

7.2

7.2

–d

6.0

–d

6.0

pH

Stabilizer (mg/mL) α-α-Trehalose
2H2O (42), Polysorbate 20 (0.09) α-α-Trehalose
2H2O (19), Polysorbate 20 (0.09) α-α-Trehalose
2H2O (19), Polysorbate 20 (0.09) D-Sorbitol (16), PEG 3350 (4.7) α-α-Trehalose
2H2O, Polysorbate 20 Sucrose (50), Polysorbate 80 (0.05)

–

13.12.2017 761072 06.12.2019 731086

– –

21.04.2017 761054

05.04.2016 125544 –

24.08.1998 103772

01.12.2017 761074 –

18.05.2018 EMEA/ H/C/004647

09.09.2013 EMEA/ H/C/002778 10.09.2013 EMEA/ H/C/002576 26.05.2016 EMEA/ H/C/004020

12.12.2018 EMEA/ H/C/004916 27.07.2020 EMEA/ H/C/005209 13.08.1999 EMEA/ H/C/000240

11.03.2019

13.06.2019 761073

16.05.2018 EMEA/ H/C/004361 26.07.2018 EMEA/ H/C/004463

FDA approval date, BLA number 14.12.2018 761091

EMA approval date, Product number 08.02.2018 EMEA/ H/C/002575

(A) Monoclonal antibodies. Monoclonal antibodies with existing biosimilars are listed under C. (B) Antibody-Drug Conjugates. (C) Biosimilars and reference products. For certain products, the quantitative composition is not disclosed, as indicated by a dash q.s. quantity sufficient a Pegylated Fab fragment b Protein concentration after reconstitution c Trastuzumab is either supplied as a single-dose 150 mg vial or a multidose 440 mg dose strength vial. The multidose vial is co-packaged with Bacteriostatic Water for Injection, USP, containing 1.1% Benzyl alcohol as a preservative. If reconstituted with Sterile Water for Injection without a preservative, the reconstituted solution is considered single-dose d Data not available

10

10

10

Disodium succinate
6H2O, Succinic acid Disodium succinate hexah ydrate (1.2), succinic acid (0.06) NaH2PO4 (0.49), Na2HPO4
H2O (0.22)

NaH2PO4
H2O (0.22), Na2HPO4
2H2O (0.61) NaH2PO4
H2O, Na2HPO4 2H2O NaH2PO4
7H2O (0.26), Na2HPO4 H2O (0.56)

10 10

L-Histidine
HCl
H2O (0.47), L-Histidine (0.3) L-Histidine
HCl
H2O, L-Histidine NaH2PO4 H2O (0.22), Na2HPO4
2H2O (0.61)

L-Histidine-HCl
H2O (0.48), L-Histidine (0.40)

L-Histidine-HCl
H2O (0.48), L-Histidine (0.31)

21

21

21

Samsung Bioepis

Full set or subset of indications of Remicade (reference product)

Inflammatory disease

Indication class

Formulation composition Protein concentration (mg/mL)b Buffer system (mg/mL) L-Histidine
HCl (0.48), L-Histidine (0.31)

Flixabi (EU)/ Renflexis [45] (US) Zessly [46]

Remsima [44]

Inflectra [43]

Ogivri [39, 40] Zercepac [41]

Biosimilar of Infliximab

Mylan

Trazimera [39]

Remicade [42]

Pfizer

Kanjinti [38]

Accord Healthcare Janssen Biologics (EU)/ Centocor (US) Pfizer (EU)/ Celltrion (US) Celltrion

Amgen

Product name

Infliximab

Active

Authorization holder US/EU

Table 2 (continued)

88 A. Allmendinger et al.

Formulation Design for Freeze-Drying: Case Studies of Stabilization of Proteins

89

• Glass dynamics and vitrification theory describes the kinetic stabilization, in which the protein is immobilized in a glassy matrix of amorphous excipients. Vitrification has been suggested to contribute to protein stability during freezing and drying, as well as storage in the dried state [51–53]. More precisely, this mechanism describes stabilization by reduced mobility as a result of an increase in viscosity (through freeze-concentration and temperature reduction during freezing and through water removal during drying), thus slowing down the dynamic degradation processes. This immobilization due to increased viscosity is related to the glass transition, which for a maximally freeze-concentrated solution is referred to as Tg’. The glass transition temperature in the dried state is referred to as Tg and represents global mobility (α relaxations or slow motions). More recently, multiple studies have corroborated the importance of local mobility (β relaxation or fast motions) in the solid state [54, 55]. The roles of local and global mobility with regard to protein stability in the dried state will be discussed more thoroughly in Sect. 3.5 below. • Preferential exclusion theory developed by Timasheff et al. describes thermodynamic stabilization of proteins, where solutes are excluded from the protein surface, thereby keeping the protein preferentially in its hydrated form [56]. While this mechanism primarily describes protein stabilization by co-solutes in the liquid state, it also contributes to protein stability at the beginning of the freezing step, when the solution has not yet been vitrified [57].

2.2

Practical Considerations

Vitrification and water replacement theory are both relevant for protein stability in the dried state. This paramount importance of both theories has been demonstrated in a study by Tonnis et al. [58]. They investigated an antibody lyophilized with different types of sugars: the disaccharide trehalose, inulin 4 kDa as a flexible oligosaccharide, and with dextrans of different molecular weights. All formulations were stored at 60 °C, which was well below their glass transition temperatures (Tg), thus stabilizing through vitrification. Still, a difference in stability was observed among the formulations, which could not be explained by vitrification. Each excipient had hydroxyl groups and is thus expected to stabilize via water replacement. This study showed that vitrification alone is not enough to ensure protein stability. Additionally, the study provided a further refinement of the water replacement theory, highlighting the importance of steric hindrance. Steric properties of the excipient are an important parameter to determine if a sugar is able to stabilize via water replacement. Tonnis et al. found that flexible sugars are better stabilizers compared to rigid sugars, which lack the ability to form a close monolayer around the protein’s surface. The relevance of steric hindrance with regards to protein interaction becomes increasingly important as the size of the sugar molecule increases. Finally, the study by Tonnis et al. suggested that protein stability is determined by water replacement rather than vitrification. This is consistent with the literature, which states that protein stability at storage temperatures 10–20 °C below the glass transition temperature is determined by vitrification, whereas at temperatures far below the Tg, water replacement is the predominant mechanism [52]. This interplay of water replacement, steric hindrance, and vitrification is further demonstrated by the data shown in Fig. 1. Upon lyophilization of an antibody formulated either with the disaccharide sucrose or the flexible oligosaccharide 2-hydroxypropyl-betacyclodextrin (HPβCD) at molar protein to sugar ratios of 1:3485 and 1:794, respectively, the antibody formulation is well stabilized. In contrast, when formulated with the rigid polysaccharide dextran 500 kDa at a molar ratio of 1:3068 (calculated based on the glucose units), an increase in soluble aggregates was observed, suggesting an impact of steric hindrance. However, when a mixture of dextran 500 kDa and sucrose was used, the small disaccharide (molar sucrose to protein ratio >360:1) was able to closely interact with the protein via hydrogen bonds, whereas dextran can stabilize via vitrification. Stability data after storage for 9 months at 40 °C demonstrated the importance of vitrification. Tg of the sucrose formulation was ~43 °C (residual moisture ~4%), whereas Tg of the formulation with HPβCD was ~200 °C. Thus, when stored at 40 °C, the antibody in the sucrose formulation was not sufficiently immobilized and displayed high aggregation, whereas much less aggregation was observed with the HPβCD formulation stored well below its Tg.

2.3

Excipient Properties and their Relevance for the Lyophilization Process

Lyophilization is a costly and time-consuming drug product manufacturing process. Therefore, an optimal lyophilization development process aims to develop short lyophilization cycles, rendering elegant lyophilizates while ensuring the protein’s

90

35

pre-FD post-FD 9months_40°C

30

Soluble aggregates (%)

Fig. 1 Soluble aggregates by size-exclusion chromatography before lyophilization (pre-FD) and after lyophilization (post-FD) shown for 10 mg/mL monoclonal antibody formulated with a total of 80 mg/mL sugars. (Disclaimer: Data was collected in different experiments as part of a PhD thesis [59])

A. Allmendinger et al.

25 20 15 10 5

ro x yd 2-H

De

xtr

an 5

00

kD

a+

yp rop

yl -

Su

De xt r

c ro se

be tac yc

(1 : 1,

lod ex

trin

w/w )

Da 50 0k an

Su cr o se

0

stability. To do so, one prerequisite is the knowledge of the formulation’s Tg’, which is an important characteristic for the design of the lyophilization process.

2.3.1 Analysis of Excipient Properties Tg’ is typically determined using differential scanning calorimetry (DSC), by comparative measurement of heat capacities of a sample and a reference pan during cooling and heating cycles. In contrast to other endothermic (e.g., melting) or exothermic (e.g., crystallization) events, the glass transition is not a phase change but a kinetic phenomenon that can be detected as a step in the baseline due to a change in heat capacity. In general, the higher the heating or cooling rate, the higher the sensitivity to detect the soft step during a DSC measurement. However, a faster rate involves a broadening of the step and typically shifts the glass transition temperature to higher temperatures [60]. It might thus be helpful to first perform a screen at higher rates of e.g., 5 °C/min or 10 °C/min, followed by a second run at lower heating and cooling rates (e.g., 1–3 °C/min) for accurate determination of Tg’. For lyophilization process development, in addition to DSC measurements, freeze-drying microscopy is often used to determine the collapse temperature (Tc) of the formulation. As mentioned above, DSC can also determine crystallization events such as eutectic temperature (Teu), which becomes important in the case of crystalline bulking agents such as mannitol. After lyophilization, X-ray powder diffraction (XRPD) may be needed as an additional technique to characterize the lyophilizate’s morphology when crystalline excipients such as mannitol are used, which can form different polymorphs that can also be determined using XRPD. Crystalline excipients are often used to overcome challenges related to low glass transition temperatures of sucrose- or trehalose-based formulations, particularly for low concentration protein formulations, but do add additional complexity, not only by requiring additional analytics, but also to the lyophilization process.

Formulation Design for Freeze-Drying: Case Studies of Stabilization of Proteins

2.3.2

91

The Link Between Excipient Properties and Lyophilization Process Parameters

2.3.2.1 Freezing Tg’ of the formulation significantly influences the freezing and primarily the primary drying step of a lyophilization cycle. For the freezing step, it is important to select a final freezing temperature that is below Tg’ for an amorphous formulation and/or below Teu for a eutectic formulation, including an adequate hold time to ensure complete solidification of the protein before moving into the primary drying step. Therefore, the lower the Tg’ of a formulation, the lower the final freezing temperature should be. 2.3.2.2 Annealing Prior to Primary Drying Annealing describes a thermal treatment introduced into the freezing protocol where samples are held above the Tg’ and below the ice melt temperature for a defined period of time [61]. This step is often applied in cases where a crystalline bulking agent such as mannitol is used, since its crystallization is promoted through annealing. When used as a bulking agent, full crystallization of mannitol must be ensured to allow for drying below its eutectic temperature and to ensure that protein stability during subsequent storage is not compromised. If mannitol remains in its amorphous form during lyophilization, it interacts and stabilizes the protein. Mannitol crystallizing upon storage can no longer interact with the amorphous protein and thus compromises protein stability. 2.3.2.3 Primary Drying The temperature at which collapse (loss of structure) during primary drying is observed is referred to as the collapse temperature (Tc). During primary drying, where water is removed via sublimation, the product temperature in the steady state of primary drying should be maintained below Tg’ or at least below Tc. If primary drying is carried out above Tg’ and/or Tc, viscous flow is induced that may lead to collapse of the lyophilized product. Collapse is an undesirable quality attribute with regards to cake appearance and should thus be avoided [62]. Collapse may [12] or may not [63] impact protein stability, depending on the extent of collapse. It is therefore recommended to select primary drying conditions that result in a product temperature below Tc. However, for commonly used cryo-/lyoprotectants such as sucrose and trehalose, this leads to lengthy and thus costly lyophilization processes due to low Tg’ of these disaccharides. For highly concentrated antibody formulations, the low Tg’ of the sugar is compensated by the high Tg’ of the protein itself. Moreover, due to increasing viscosity with higher antibody concentrations, the delta between Tg’ and Tc increases [64]. Colandene et al. found complete collapse at around 13 °C above the Tg’ for a 100 mg/mL protein formulation [65]. While for highly concentrated protein formulations, lyophilization above Tg’ might thus be possible, for low viscous formulations, Tg’ and Tc are typically very similar (Tc 1–3 °C above Tg’). In order to enable fast lyophilization processes for low concentration protein formulations, bulking agents such as mannitol are often added to the formulation to compensate for the low Tg’ of the amorphous formulation. Bulking agents are often crystalline compounds such as mannitol, which do not contribute to protein stability, but do render elegant lyophilizates. Recently, Horn et al. investigated crystalline amino acids as alternative bulking agents to mannitol. Such bulking agents could be superior to mannitol, as they can be added at much lower concentrations compared to mannitol in exchange for a higher stabilizer content, while still showing full crystallization [66]. However, such excipients often require annealing to ensure full crystallization, thus adding additional complexity to the lyophilization process. 2.3.2.4 Secondary Drying Secondary drying is performed to remove the unfrozen, bound water by desorption, and achieve low residual moisture, typically 200 mg). High concentration formulation usually contains ≥100 mg/mL of protein in a formulation buffer stabilized with excipients. In addition to high viscosity, high protein concentration can also lead to increased protein instability due to increased protein–protein and protein–excipient intermolecular interactions. Impurities such as host cell proteins, which are often controlled at very low levels in low concentration X. (C). Tang (✉) · Y. Cheng · M. Shameem Formulation Development, Regeneron Pharmaceuticals Inc, Tarrytown, NY, USA e-mail: [email protected] # America Association of Pharmaceutical Scientists 2023 F. Jameel (ed.), Principles and Practices of Lyophilization in Product Development and Manufacturing, AAPS Advances in the Pharmaceutical Sciences Series 59, https://doi.org/10.1007/978-3-031-12634-5_6

103

104

X. (C). Tang et al.

protein formulations, can be co-concentrated with proteins during the production and affect the stability and safety of the drug product [5]. In comparison to liquid formulations, lyophilized formulations often feature the advantages of better stability, easy storage and shipping, and shorter development time [6, 7]. This is particularly true for high concentration formulations when the development of stable liquid formulation tends to be more challenging. Lyophilization formulations also offer greater dosing flexibility. By adjusting the reconstitution volume, one lyophilized drug product can be reconstituted to various concentrations and support the need of both IV and SC administration. Due to all these advantages, lyophilized protein drug products are particularly suitable for early stage clinical development when the speed to first-in-human (FIH) is often the priority while limited time and resources are available for formulation development. In the case that the development of a stable high protein concentration liquid formulation is not feasible, the lyophilized high concentration protein drug product would also be a viable strategy for commercial drug product development. It is worth mentioning that lyophilized protein drug products have the potential to be developed as room temperature stable drug products for more convenient storage and shipment [8]. Development of room temperature stable lyophilized protein drug products can make the drug available to the remote areas where the product cold chain is not available. Despite the advantages, the potential challenges encountered during the development of high concentration lyophilized protein drug products such as suboptimal drying process, limited long-term storage and slow reconstitution time could influence the choice between a liquid or lyophilized formulation. A significant advantage of using a lyophilized drug product, specifically early in the drug development, for first-in-human clinical trials (e.g. safety assessment, phase I) is to allow enough time and resources towards optimizing the final formulation (i.e. liquid or lyophilized) for a commercially desirable product. This chapter discusses the scientific foundation and the critical issues related to the lyophilization of the high protein concentration drug product. Specifically, the interrelated process of lyophilization, the associated denaturation stresses and, more importantly, its impact on the lyophilized protein formulations are discussed. Additionally, guidelines for successful development of a lyophilized high concentration antibody formulation are presented.

2

Characteristics of High Protein Concentration Lyophilized Drug Product

Increase in protein concentration could have a broad impact on the critical quality attributes (e.g. physical appearance, cake structure, viscosity, osmolality, reconstitution time) and lyophilization process of a lyophilized protein drug product (e.g. primary drying time). In this section, we will discuss the key characteristics of high protein concentration lyophilized drug products.

2.1

Physical Appearance and Cake Structure

High protein concentration lyophilized drug products often exhibit off-white to pale yellow color that typically arises from the formulated drug substance and intensifies with the increase in protein concentration. After reconstitution, the drug product often appears opalescent. Opalescence of protein formulation typically results from the Rayleigh scattering of protein molecules and is generally considered as an undesirable characteristic for a protein drug product. Opalescence is inconsistent with the general perception of pharmaceutical elegancy and could also be easily confused with turbidity arising from protein aggregation and particulate formation, causing concerns about the quality of the drug product. In addition, it is generally more difficult to develop a placebo product that matches the appearance of an opalescent drug product. Studies have shown that opalescence in high protein concentration formulations is often correlated with the reversible self-association of protein molecules and formation of transient protein network [9]. Mitigation of protein reversible self-association via modifying formulation conditions (e.g. pH and salt concentration) has been demonstrated to be able to reduce the opalescence in high protein concentration formulations [10]. Protein concentration can also have an impact on cake structure. As shown in Fig. 1, the cake of lyophilized protein drug product becomes denser and less porous with the increase in protein concentration [11]. Although this characteristic makes it generally easier to obtain robust cake for high protein concentration drug products, it also makes the reconstitution of high protein concentration lyophilized drug products more difficult.

Challenges and Considerations in the Development of a High Protein Concentration. . .

105

Fig. 1 Lyophilization of a monoclonal antibody as a function of loading concentration. Upper left panel: Loading concentration from left to right was 40, 60, 80, 100, and 110 mg/mL, respectively, while maintaining the same total mass of MAb and excipients. Lower left panels: Scanning electron microscopy of lyophilized solid for the 40 and 110 mg/mLMAb loading concentrations [11]

2.2

Prolonged Reconstitution Time

High protein concentration lyophilized drug products often exhibit prolonged reconstitution time, ranging from minutes to several hours [12, 13]. As shown in Fig. 2, the reconstitution time is typically positively correlated with protein concentration. Prolonged reconstitution is highly undesirable from a dose preparation and administration perspective and can significantly limit the practical usage of a lyophilized drug product. Prolonged reconstitution can also increase the likelihood of incomplete reconstitution prior to administration, particularly in the case of patient self-administration, which can lead to the presence of undissolved proteinaceous particles in the dosing solution and increase the risk of immunogenic reactions, compromising the safety and efficacy of the lyophilized drug product. The prolonged reconstitution of high protein concentration lyophilized drug products can often be attributed to their undesirable cake properties such as poor wettability, low tendency for rehydration, and disintegration, and low porosity [15– 17]; a recent study indicates that the prolonged reconstitution of high protein concentration lyophilized drug products can also result from the high local viscosity at the cake dissolving surface [18]. Kulkarni et al. showed that for amorphous formulations of formulation >50 mg/mL protein, instead of the cake properties, the reconstitution time was primarily correlated with the “concentrated formulation viscosity” (defined here as the viscosity of the solution obtained by reconstitution with 1/3 volume of the reconstitution fluid). The reconstitution time, however, was poorly correlated with the viscosity of fully reconstituted solution. The author proposed that the better correlation between “concentrated formulation viscosity” and reconstitution time was observed because the “concentrated formulation viscosity” better reflects the viscosity at the cake dissolving surface that is expected to have a major impact on the erosion rate, wetting, and hydration of the cake core. The authors also suggested that viscosity lowering excipients, such as arginine hydrochloride and sodium chloride can potentially be used to shorten reconstitution time by reducing the “concentrated formulation viscosity” [18]. Considering the relationship between viscosity and protein concentration, it is logical to believe that the prolonged reconstitution time of high protein concentration lyophilized drug products can result from the high local viscosity at the dissolving surfaces.

106

X. (C). Tang et al.

Fig. 2 Reconstitution time of lyophilized monoclonal antibody formulations for protein concentrations ranging from 40 to 210 mg/m [14]

It is also reported that high protein concentration can affect the reconstitution time by affecting the crystallinity of the lyophilized cake [13, 19]. When studying a formulation containing a Fc fusion protein, sucrose, mannitol, and polysorbate 20, Cao et al. revealed that high protein concentration can prolong reconstitution time by inhibiting the crystallization of mannitol and thereby reducing the crystallinity of the lyophilized cake. The authors also showed that, at 100 mg/mL protein concentration, a higher mannitol to sucrose ratio could lead to the formation of cakes with higher degree of crystallinity and shorter reconstitution time [13]. Another interesting discovery from this study is that the specific surface area of the cakes remain at around 1 m2/g in the protein concentration range of 30–150 mg/mL, indicating that the specific surface area may not always be a critical factor to the reconstitution time in this case [13].

2.3

High Viscosity

The relationship between protein concentration and viscosity has been studied extensively [10, 20, 21]. When the protein concentration is relatively low, the viscosity of a protein formulation typically increases moderately as a function of protein concentration. For highly concentrated protein formulation, however, an approximate exponential relationship is often observed between the viscosity and protein concentration. It has been shown that the dramatic increase in viscosity at high protein concentration is often caused by reversible self-association of protein molecules and formation of transient protein network [10, 21]. In the case that a lyophilized product is reconstituted to a smaller volume to achieve a higher protein concentration (a common strategy in the use of lyophilized drug product), the viscosity of the reconstituted solution can be significantly higher than that of the pre-lyophilization liquid formulation depending on the protein concentration post reconstitution and the protein itself (i.e. the tendency for reversible self-association and formation of transient protein network). Since high protein concentration lyophilized drug products are typically targeted for subcutaneous administration, the viscosity of the reconstituted resolution needs to be controlled within the acceptable range of the administration device, which is typically below 20–30 cp for auto-injectors. Viscosity reducers, such as arginine hydrochloride and sodium chloride, can be included in the formulation to mitigate the viscosity post reconstitution.

2.4

Longer Primary Drying Time

The total solute weight of high protein concentration formulations often exceeds 10% of the total weight of the formulation. High solute content can lead to higher resistance (or lower flux of water vapor through the dry layer) during primary drying and therefore longer primary drying time which makes the lyophilization process more time-consuming and costly. To avoid cake collapse, primary drying is typically performed at a temperature 2–3 °C below the glass transition temperature of

Challenges and Considerations in the Development of a High Protein Concentration. . .

107

Fig. 3 Glass transition temperature of the maximally freeze-concentrated solution and Tc as a function of protein concentration. Values for mAb A are denoted by triangles, mAb B by circles, mAb C by squares, and Pro X by diamonds. Closed symbols: Tg’; open symbols: Tc [22]

maximally freeze-concentrated solution (Tg’) or the collapse temperature (Tc) [6]. As is shown in Fig. 3, Tg’ and Tc increase as a function of protein concentration. At lower protein concentration, Tg’ and Tc are typically within the range of 1–2 °C and used interchangeably. At higher protein concentration, however, Tc can be significantly higher than Tg’ [23]. It has been reported that, for high protein concentration amorphous formulations, primary drying conducted at a temperature above Tg’ but below Tc could significantly reduce the primary drying time, but still yield pharmaceutically acceptable cakes [22]. The same primary drying strategy, however, could cause cake collapse when the protein concentration is relatively low [22]. Another reason for conducting primary drying at a product temperature below Tg’ is to avoid protein unfolding because proteins are thermodynamically unstable at the temperature above Tg’. Research from the author et al. showed that, even at a product temperature much higher than Tg’, protein unfolding may still be too slow to occur on the timescale of lyophilization if the formulation is highly viscous [24]. High protein concentration formulations are often associated with high viscosity and therefore can be kinetically more resistant to protein unfolding during lyophilization. This characteristic of high protein concentration formulation facilitates the operation of primary drying at a relative high shelf temperature during primary drying in order to reduce the time and cost of the lyophilization process.

2.5

Osmolality

Osmolality is another critical quality attribute that needs to be particularly considered during the development of high protein concentration lyophilized drug products because high protein concentration formulations are typically designed for subcutaneous administration that requires an osmolality of 600 mOsm/kg or less to minimize injection pain. Most proteins are vulnerable to the stresses incurred by the lyophilization process and long-term storage and require the addition of stabilizing excipients to maintain their stability. Since a certain molar ratio of stabilizer to protein is often required to achieve sufficient stabilizing effect [25], more stabilizing excipients are often required for high protein concentration formulations, making the formulation hypertonic and undesirable for patient comfort [26]. To minimize the osmolality of high protein concentration formulations, it is often necessary to screen the pH and excipients to identify the stabilizing excipients that exhibit sufficient stabilizing effect at a lower molar concentration. Sometimes the formulation scientists may need to make a balance between formulation stability and osmolality when developing a formulation.

108

3

X. (C). Tang et al.

Stability Considerations for the Development of High Protein Concentration Lyophilized Drug Products

A variety of protein-denaturing stresses can develop during the lyophilization process, including cold denaturation, freezeconcentration, ice–water interface, pH shift, phase separation, and dehydration. These stresses can directly compromise the stability of drug products during the lyophilization process and/or long-term storage. In this section, we will discuss the main stresses that proteins will be exposed to during the lyophilization process.

3.1

Protein Cold Denaturation

Cold denaturation refers to the spontaneous unfolding of proteins under low temperatures [27–29]. Compared to thermal denaturation, cold denaturation has been much less studied. In current theory, the free energy of protein unfolding has a parabolic relationship with temperature and there exist two unfolding transition points where the free energy of unfolding equals zero. The transition point that is above the room temperature is commonly referred to as the thermal denaturation point, while the other transition point that is typically below the ice freezing point is called cold denaturation point. The unfolding of proteins becomes energetically favorable when the temperature is above the thermal denaturation point or below the cold denaturation point [29]. Unlike the entropy-driven thermal denaturation, cold denaturation is an enthalpy-driven process that is largely caused by the weakening of hydrophobic interaction in proteins as a result of the increased solubility of nonpolar groups in water at decreasing temperatures [30, 31]. Cold denaturation could occur in the freezing process of lyophilization that operates at low temperatures (-20 °C to -50 °C) and compromise the in-process stability as well as long-term storage stability [32]. Using β-lactoglobulin and phosphoglycerate kinase (PGK) as model proteins, the author et al. showed that the cold stability of a protein formulation was dependent on pH, protein concentration, and additives, and the presence of stabilizers (e.g. sugars and/or polyols) could significantly enhance protein cold stability [32]. As shown in Fig. 4, the study also reported the observation of significantly increased cold stability at higher protein concentration, presumably resulting from the macromolecular crowding effect arising from nonspecific steric repulsion between molecules [32]. This result indicates that high protein concentration formulations may generally have lower risk in cold denaturation than low protein concentration formulations. In another study, the author et al. investigated the correlation between protein unfolding rate and system viscosity at low temperatures. The results indicated that the freeze-concentration effect and low temperature during the freezing step of lyophilization can lead to a highly viscous system in which the rate of cold denaturation is remarkably reduced so that protein cold denaturation may not occur at a temperature that is well below the cold denaturation temperature [24]. This result indicates that high protein concentration formulations can be kinetically more resistant to cold denaturation due to their higher viscosity.

Fig. 4 The effect of protein concentration on β-lactoglobulin cold denaturation temperature as determined by DSC [32]

Challenges and Considerations in the Development of a High Protein Concentration. . .

3.2

109

Freezing Stress

3.2.1 Freeze-Concentration The exclusion of proteins and other solutes from ice crystals during the freezing process of lyophilization can lead to a drastic increase in the concentrations of all solutes remaining in solution. This phenomenon is commonly referred as freezeconcentration. For example, the concentration of a 150 mM NaCl solution can increase 24-fold to 3.5 M when the solution is frozen to its eutectic temperature of -21 °C [33]. Freeze-concentration could also result in a sevenfold increase in protein concentration [34] and elevate the concentration of low molecular weight carbohydrates to as high as 80% [35]. Any properties that are associated with solute concentrations can be significantly altered by freeze-concentration and such alteration may cause protein destabilization. For example, a drastic increase in ionic strength due to freeze-concentration and corresponding increase in salt concentration could result in protein denaturation and aggregation. Freeze-concentration can also increase the concentrations of impurities, such as reactive oxidative species and proteases in solution. Wisniewski et al. reported that the oxygen concentration in solution could increase 1150-fold due to freeze-concentration [36]. Although the rate of chemical reaction typically decreases as temperature drops, a drastic increase in the concentration of impurities as a result of freeze-concentration can still accelerate the degradation of proteins [37]. It is also reported that residence time in freeze-concentrate could have a profound impact on protein stability and longer residence time in freeze-concentrate could lead to significant increase in protein aggregation [36]. 3.2.2 Ice–Water Interface As ice crystals form during freezing, proteins are exposed to ice–water interface that widely exists in partially frozen state. Direct adsorption of proteins to ice–water interface can induce protein unfolding and aggregation [38, 39]. Ice–water interface-induced protein denaturation can also occur without the direct adsorption of proteins to the interface [40]. Using molecular dynamics simulation, Arsiccio et al. demonstrated that ice–water interface could enhance the cold denaturation of proteins near the interface by increasing the activity of water molecules near the ice–water interface in hydrating the nonpolar groups of protein and thereby reducing the free-energy barrier for protein unfolding. This study also showed that ice–water interface-mediated enhancement in protein cold denaturation can be mitigated by the presence of cryoprotectants, such as sugars [40]. The area of ice–water interface depends on the size and morphology of ice crystals. Larger ice crystals are typically correlated with smaller area of ice–water interface. The freezing rate can have a direct impact on the size and morphology of ice crystals. Larger ice crystals or smaller area of ice–water interface can be obtained at a slower cooling rate [41]. The temperature of ice nucleation and thermal history of the formulation post ice nucleation can also affect the size and morphology of ice crystals and therefore the area of ice–water interface. Controlled nucleation at a higher temperature has shown benefit in mitigating the risk of ice–water interface by facilitating the formation of larger ice crystals with smaller area of ice–water interface [42]. 3.2.3 pH Change Significant pH change can occur during the freezing process of lyophilization due to the differential crystallization of buffer salts. For example, the dibasic form of sodium phosphate has a higher tendency to crystallize than the mono-basic form during freezing, which can result in the shift of pH to 4.0 or lower [43]. Differential crystallization of the dibasic and mono-basic form of potassium phosphate could lead to a final pH of 9 [43]. Drastic change in pH due to the differential crystallization of buffer salts during freezing can compromise in-process protein stability and stability of dried products during the long-term storage. Change in pH during freezing is of a particular concern for the proteins that are sensitive to pH and only stable in a relatively narrow pH range. Therefore, for the development of lyophilized drug products, it is generally preferable to choose the buffer species that exhibit minimal pH change during freezing, such as histidine, Tris, and citrate, over those that could lead to a drastic pH change, such as sodium phosphate and potassium phosphate. Crystallization of buffer salts during freezing can be mitigated by optimization of the freezing process and formulation composition. It was reported that rapid freezing and shortening of the duration of annealing can help minimize the crystallization of buffer salts [43]. In another study, several excipients, including sucrose, sorbitol, lactose, trehalose, and dextran, could mitigate or prevent the crystallization of the dibasic form of sodium phosphate [44].

110

3.3

X. (C). Tang et al.

Phase Separation

During the freezing process of lyophilization, freeze-concentration and reduced solubility at lower temperature can make it thermodynamically unfavorable for all the solutes in the freeze-concentrates to remain in a single phase. As a result, the formulation matrix may be separated into several distinct phases. Formulation components such as buffer salts, bulking agents, and stabilizing excipients can crystallize during freezing and be separated from the protein-rich phase(s). Liquid– liquid phase separation can also occur during freezing, which can lead to the formation of protein-rich phase(s) and stabilizerrich phase(s) [45] and a large excess of protein-denaturing liquid–liquid interface [46]. Drastic change in the solution properties (such as pH and ionic strength) and separation of proteins from stabilizers caused by phase separation can potentially compromise the stability of proteins during the lyophilization process and long-term storage. Considering the remarkable impact of phase separation on the stability of lyophilized drug products, it is critical to detect and control phase separation in both freeze-concentrates and dried states. Detection of the formation of crystalline phases is relatively straightforward. Several techniques, including powder X-ray diffraction, polarized light microscopy, and vibrational spectroscopy, have been applied to the detection of crystalline phases [47]. Detection of the formation of amorphous phases, however, is generally more challenging. Differential scanning calorimetry (DSC), scanning electron microscopy (SEM), and Raman mapping have been used to detect amorphous phase separation in freeze-concentrates and solid state [47, 48]. The detection of multiple glass transition temperatures by DSC typically indicates the presence of multiple amorphous phases. DSC is generally not very sensitive to protein-rich phases due to their low heat capacity change (during glass transition) and the broad glass transition temperature of proteins. Scanning electron microscopy (SEM) relies on the visual identification of features that represent amorphous phase separation. Raman mapping analysis has been reported to detect the separation of amorphous phases in several protein-polymer formulations and protein-disaccharide formulations [47, 48].

3.4

Dehydration Stress

Protein molecules in aqueous solution are covered with a hydration shell (or a monolayer of water molecules) that weighs approximately 0.35 gram per gram of proteins [49]. Water molecules in the hydration shell can form hydrogen bonds with the polar groups on protein surface, and these hydrogen bonds are critical for maintaining the native structure and conformation of proteins [49, 50]. During the drying process, a large part of the hydration shell is removed from protein surface, leading to the disruption of stabilizing hydrogen bonding and protein unfolding. Dehydration-induced protein structural and conformational change has been well documented in the literatures [51–53]. Such changes are typically irreversible and can compromise both in-process stability and the stability of dried product over long-term storage. It is also reported that removal of the hydration cell can cause the loss of protein surface charge and thereby facilitate protein aggregation [49]. If waters are essential for the function of proteins (e.g. enzymes that relies on water in their active site for catalysis), dehydration can directly cause the loss of protein function without disturbing protein structure [54]. To protect proteins from the dehydration stress during the drying process, stabilizers with functional hydroxyl groups such as disaccharides (e.g. sucrose and trehalose) and polyols (e.g. sorbitol and mannitol) are often included in the formulation [51]. These excipients can substitute for water molecules and form hydrogen bonds with proteins, which are essential for the maintaining of protein structure in the solid state.

4

Stability Strategy to Minimize Protein Degradation During Lyophilization Process and Shelf Life Storage

The protein stability of the lyophilized drug product can be improved with addition of stabilizers, such as sugars (e.g., sucrose or trehalose), and amino acids in addition to the pH buffer system with a trace amount of surfactant [55].

4.1

Buffer Species, Surfactants, and Viscosity Reducer Selection

To minimize pH shift, it is preferable to choose the organic buffer species that are not susceptible to crystalize during freezing process, such as histidine, citrate, and Tris. In case, phosphate buffer has to be used due to protein stability, a relatively low

Challenges and Considerations in the Development of a High Protein Concentration. . .

111

concentration of the phosphate buffer (e.g., ≤20 mM) should be used to minimize the crystallization of the dibasic form of sodium phosphate thus reducing the potential pH shift during freezing [56]. The surfactant (usually at a level of ≥ the critical micelle concentration) is included in the lyophilized protein formulation to protect protein from interfacial stresses upon ice formation during freezing process and air–liquid interface stress during reconstitution. As a general rule, it is recommended to use the least amount of surfactant that is needed to effectively protect protein molecules from protein–ice crystal interface stress during freezing and the air–liquid interface stress during reconstitution [57, 58]. A viscosity reducer such as arginine hydrochloride or sodium hydrochloride can be included in the formulation to manage the viscosity of the high concentration lyophilized protein drug products [18, 59]. Sodium chloride tends to partially crystalize out during lyophilization process or shelf life storage which may cause protein degradation [60]. Arginine hydrochloride is a preferred choice of viscosity reducing agent since it stays in an amorphous state and stabilizes protein.

4.2

Stabilizers Selection

The stabilizers can serve as cryoprotectants and lyoprotectants to protect protein molecules from freezing and dehydration stresses during lyophilization process. The stabilizers also protect protein from degradation (predominately aggregation formation for lyophilized protein drug product) duration the product long-term shelf life storage. Among the commonly used stabilizers, disaccharides such as sucrose (or trehalose) have been demonstrated as effective lyoprotectants and cryoprotectants and have been widely used in the lyophilized protein drug products [55]. Polysaccharides (such as dextran and HES) and other polymers (such as PVG) are not recommended for high concentration lyophilized protein formulations since there is potential of phase separation of the stabilizer and protein during the lyophilization process [61, 62]. A minimal specific molar ratio of sugar (such as, sucrose or trehalose) to protein is required to provide sufficient stabilization effect for protein molecules against aggregation during storage of lyophilized protein formulations. Protein stability was optimized at molar ratios of 360:1 or greater [25]. For a high concentration monoclonal antibody formulation (e.g., 150 mg/mL mAb), the sucrose or trehalose concentration would be 360 mM or above (≥12.3% w/v sucrose or trehalose in the reconstituted drug product) in order to achieve a stabilizer to protein molar ratio of 360:1. In general, the mAb is more stable when higher molar ratio of stabilizer to protein is contained in the formulation. As shown in Fig. 5, the stability of the 150 mg/mL lyophilized mAb (aggregation formation rate as a function of square root of storage time) was investigated at different storage conditions (i.e., 5 °C, 25 °C and 30 °C) [8]. The results showed that protein stability improved with increasing concentrations of sucrose as the stabilizer. The best stability was observed when the 150 mg/mL lyophilized mAb was formulated with 18% (w/v) of sucrose which represents the highest stabilizer concentration assessed in this study. The high protein concentration formulation (e.g., 150 mg/mL mAb) that contains approximately 9% sucrose (or trehalose) is an isotonic formulation and increasing sucrose (or trehalose) concentration significantly increases the formulation osmolality and viscosity and the formulation with

Fig. 5 Stability of 150 mg/mL lyophilized mAb (aggregation formation rate as a function of square root of storage time) with different concentrations of sucrose as stabilizer

112

X. (C). Tang et al.

too high viscosity and osmolality makes it difficult or impossible for manufacture production and subcutaneous administration [63]. Considering the contribution of osmolality from the viscosity reducer (such as arginine hydrochloride) in the high concentration protein formulation, the concentration of stabilizer would be further limited in order to achieve an isotonic drug product. Therefore, for a high concentration lyophilized mAb formulation intended for subcutaneous administration, there is an obvious limitation to enhance the protein stability just by increasing the concentration of the stabilizers.

4.3

Stability Strategy for a High Concentration Lyophilized Protein Drug Product Stable for Room Temperature Storage

In order to overcome the limitation of protein stability just by increasing the stabilizer concentrations, an alternative strategy is needed to improve the stability of high concentration lyophilized mAb without adversely increasing the product viscosity and osmolality. In general, the lyophilization process is designed to remove the residual water from the drug product to achieve a dry product with low moisture content for long-term storage stability of the drug products. The lyophilized drug product provides a dry matrix that immobilizes the protein molecules and stabilizes the protein molecules from degradation [63, 64]. However, the plasticizers may also be included to decrease the local mobility of the protein molecules in the dry amorphous matrix and in some cases may help to preserve the native structure and the stability of proteins. Plasticizers include small molecular weight sugar alcohols like sorbitol and glycerol, and other polyols, or small amounts of moisture (water molecules) that is intentionally kept in the lyophilized protein drug products [65]. Use of moisture in the lyophilized protein drug product as a stabilizer would not require additional excipients in the formulation and thus no increase to the reconstituted product viscosity or osmolality [8]. In Fig. 6, the stability of 150 mg/mL lyophilized mAb formulations at different moisture contents is summarized. The stability of the lyophilized drug product as indicated by aggregation formation rate as a function of square root of storage time showed that the lyophilized DP with the lowest moisture content (~0% moisture content or undetectable moisture content) didn’t yield the best stability at any storage conditions ranging from 5 °C to as high as 50 °C. A moisture content of approximately 4% (3–5%) is optimal for refrigeration (5 °C) and room temperature storage (25 °C and/or 30 °C). The optimal moisture content for best protein stability is lower at higher storage temperature (~2.5% at 50 °C). The advantage of this stability strategy is that the moisture content in the drug product does not increase formulation viscosity or osmolality.

Fig. 6 Stability of 150 mg/mL lyophilized mAb (aggregation formation rate as a function of square root of storage time) at different moisture contents. The 0% of moisture content of the product indicated a moisture content undetectable by the moisture content instrument

Challenges and Considerations in the Development of a High Protein Concentration. . .

113

Fig. 7 Room temperature stable lyophilized high concentration mAb formulation (150 mg/mL mAb) with optimal moisture content (~4% H2O, red trace) and low moisture content (10 °C) and fast ice sublimation rate. Since the product temperature in primary drying is relatively high, a moderate chamber pressure (e.g. 100–150 mTorr) is recommended for the primary drying for optimal homogeneity of the heat transfer among a set of glass vials [71]. However, if a large amount of arginine hydrochloride (e.g., ≥100 mM) is included in the formulation to manage the solution viscosity, the negative impact of arginine hydrochloride on the product collapse temperature should be considered and the primary drying conditions need to be adjusted accordingly [72]. For a high concentration protein formulation (with a high content of amorphous stabilizers, such as, sucrose or trehalose), the lyophilized product would be in amorphous state with a moderate cake shrinkage appearance. The expected moderate cake shrinkage, which is commonly considered as cosmetic appearance characteristic of the amorphous cake, would not adversely impact drug product critical quality attributes including protein stability, moisture content, and reconstitution time. The cake shrinkage can be mitigated by lyophilizing the product at a lower shelf temperature during primary drying and at a slower ramp rate during secondary drying leading to a long lyophilization cycle [73]. However, exceeded effort to develop a long lyophilization cycle in order to mitigate the cake shrinkage would not be technically or economically beneficial. In general, the high concentration protein formulation with a high solute content (e.g., 15–20% or above) would result in a great cake resistance to ice sublimation during primary leading to a relatively long primary drying time for a given combination of shelf temperature and chamber pressure. Sufficient primary drying time is necessary to ensure completion of ice sublimation and no product melt-back (or cake collapse). Therefore, it is important to determine the end point of primary drying and the duration of the primary drying for complete ice sublimation. The end point of primary drying can be detected by several different methods including thermocouple in the product vials, the measured chamber pressure differences between Pirani gauge and capacitance gauge, and manometric temperature measurement (MTM). At the end of primary drying, the product temperature increases to the shelf temperature, and the vapor composition in the freeze-drying chamber changes from essentially all water vapor during primary drying to mostly air or nitrogen [71]. Thus, the product temperature data can be used as the indication of the end point of primary drying when the product temperature approaches the shelf temperature. Since the thermocouple vials complete ice sublimation earlier than the rest of the vials because of freezing-bias-induced mass transfer effects, a safety margin is usually added to ensure completion of primary drying all the vials. Another commonly used method uses the pressure difference between the Pirani pressure gauge and capacitance pressure gauge which indicates the end of primary drying when the pressure difference decreases and approaches zero [74]. The MTM method [70] is another very sensitive method to indicate the end point of primary drying. The vapor pressure of ice determined by a fit of pressure rise data to the MTM equations approaches the chamber pressure when no ice remains which can be used to determine the end point of primary drying [61].

118

6.3

X. (C). Tang et al.

Secondary Drying

The last stage of lyophilization is the secondary drying. In this stage, the absorbed water in the lyophilized cake is removed by desorption from the solute phase. The objective of secondary drying is to reduce the residual moisture content to a level optimal for stability. The shelf temperature should be increased slowly for secondary drying to ensure cake integrity since a fast temperature ramp might cause collapse of amorphous products with fairly high residual moisture content early in secondary drying and, thus, low glass transition temperature. A moderate ramp rate of 0.2–0.3 °C/min is generally a safe and appropriate procedure considering the fairly high glass transition temperature for high concentration protein formulations. The shelf temperature in secondary drying is typically higher than that used for primary drying to facilitate desorption of water from the amorphous matrix with the same primary drying chamber pressure [62]. At higher solute concentration including the high concentration protein and the stabilizers in the formulation (i.e., >15% solids in solution), the dry product has smaller specific area, and it is more difficult to remove the absorbed water in the product. To achieve the desired moisture content in the lyophilized drug product, a longer secondary drying time (e.g., ≥6 h) at a relatively higher shelf temperature (e.g., ≥35 °C) may be needed to complete the secondary drying [61].

6.4

Lyophilization Cycle Scale-Up Considerations

A lyophilization process developed using a laboratory scale lyophilizer needs to be scaled up to manufacture scale lyophilizers for drug production. Performance of laboratory and production lyophilizers should be comparable in terms of shelf temperature control and shelf cooling and heating rate, chamber vacuum control, and shelf temperature uniformity within shelf and between shelves. In addition, the limitation of the manufacture lyophilizer condenser capacity and the ice capture rate also need to be considered to ensure no overloading during primary drying. When a lyophilized drug product is developed for early phase clinical trials, the lyophilization cycle can be relatively conservative with greater safety margin to ensure success scaleup to manufacture lyophilizer considering the fast path to clinic and the limited manufacture productions. A conservative lyophilization cycle has the advantage for drug product manufacturing at different sites, therefore, with different manufacture production units. However, an optimized robust lyophilization process is required for late phase or commercial productions, and the process scalability as well as the differences between the laboratory scale and product scale units must be considered. The most important lyophilization scalability factors are the differences in heat transfer between laboratory and commercial lyophilizers, differences in dry cake resistance to ice sublimation during primary drying between laboratory, and manufacture lyophilized products and variability in process parameter control. The atypical radiation and edge effect are known to be more pronounced in the laboratory lyophilizers since the commercial units are usually better insulated and installed in controlled environment [75]. The primary drying conditions (e.g., shelf temperature set point and the primary drying time) should be adjusted based on the vial heat transfer coefficients differences between the laboratory and production scale lyophilizers [76]. In addition, the clean GMP operational working environment (i.e., with low particulates) tends to induce higher degree of supper-cooling during the freezing stage of lyophilization process especially for the high protein concentration and high viscosity drug solutions and lead to formation of small ice crystals and large dry cake resistance to ice sublimation during primary drying which needs to be accounted for during lyophilization process scale-up. In this case, an extension of primary drying time is required to ensure completion of ice sublimation before the process advances to the secondary drying step.

7

Summary

In addition to the typical advantages of lyophilized drug products including better stability (particularly for ADC and other stability sensitive molecules), ease of storage and shipping, and shorter development time (suitable for early stage clinical trials), the fast growing demand for low-volume high-dose subcutaneous delivery of protein therapeutics has greatly raised the interest in the development of high protein concentration lyophilized drug products. Increase in protein concentration can have a variety of direct or indirect impacts on the critical quality attributes (e.g. stability, viscosity, cake structure, reconstitution time) and lyophilization process (e.g. primary drying time) of a lyophilized drug product, and therefore the ability to manufacture, store, and administer the drug product. A desirable lyophilized high concentration protein drug product will have one or more of the following product quality attributes:

Challenges and Considerations in the Development of a High Protein Concentration. . .

119

• Suitable viscosity and osmolality for intravenous and subcutaneous administration • Short reconstitution time for clinical use • Convenience for at-home administration by patient with devices such as dual chamber syringe system (with front chamber for dry powder and back chamber for reconstitution solution) • Stable for long-term storage and distribution at 2–8 °C or • Ideally stable for room temperature storage and distribution

Successful development of a pharmaceutically and economically suitable high protein concentration lyophilized drug product requires a solid understanding of the impacts of high protein concentration on the critical quality attributes and lyophilization process, as well as the implementation of an integrated approach for formulation and process development. In this chapter, we have discussed the scientific basis and unique challenges associated with the development of high protein concentration lyophilized drug products. Practical guidance for the successful development of high protein concentration lyophilized drug products was also provided.

References 1. Ecker DM, Jones SD, Levine HL. The therapeutic monoclonal antibody market. Taylor & Francis; 2015. 2. Antibody S. Antibody therapeutics approved or in regulatory review in the EU or US; 2020. 3. Lu R-M, Hwang Y-C, Liu IJ, Lee C-C, Tsai H-Z, Li H-J, et al. Development of therapeutic antibodies for the treatment of diseases. J Biomed Sci. 2020;27(1):1. 4. Mathaes R, Koulov A, Joerg S, Mahler H-C. Subcutaneous injection volume of biopharmaceuticals—pushing the boundaries. J Pharm Sci. 2016;105(8):2255–9. 5. Wang X, Hunter AK, Mozier NM. Host cell proteins in biologics development: identification, quantitation and risk assessment. Biotechnol Bioeng. 2009;103(3):446–58. 6. Pikal MJ. Freeze-drying of proteins. Part I: process design. BioPharm. 1990;3(8):18–27. 7. Remmele RL, Krishnan S, Callahan WJ. Development of stable lyophilized protein drug products. Curr Pharm Biotechnol. 2012;13(3):471–96. 8. Tang XC, Ludwig DB. Room temperature stable lyophilized protein. US Patent No: 11,241,498 B2; Feb. 8, 2022. 9. Sukumar M, Doyle BL, Combs JL, Pekar AH. Opalescent appearance of an IgG1 antibody at high concentrations and its relationship to noncovalent association. Pharm Res. 2004;21(7):1087–93. 10. Salinas BA, Sathish HA, Bishop SM, Harn N, Carpenter JF, Randolph TW. Understanding and modulating opalescence and viscosity in a monoclonal antibody formulation. J Pharm Sci. 2010;99(1):82–93. 11. Shire SJ, Shahrokh Z, Liu JUN. Challenges in the development of high protein concentration formulations. J Pharm Sci. 2004;93(6):1390–402. 12. Bhambhani A, Blue JT. Lyophilization strategies for development of a high-concentration monoclonal antibody formulation: benefits and pitfalls. Am Pharm Rev. 2010;13(1):31–8. 13. Cao W, Krishnan S, Ricci MS, Shih L-Y, Liu D, Gu JH, et al. Rational design of lyophilized high concentration protein formulations-mitigating the challenge of slow reconstitution with multidisciplinary strategies. Eur J Pharm Biopharm. 2013;85(2):287–93. 14. Garidel P, Presser I. Lyophilization of high-concentration protein formulations. In: Lyophilization of pharmaceuticals and biologicals. Springer; 2019. p. 291–325. 15. Sane P, Bogner RH, Bhatnagar B, Tchessalov S. Reconstitution of highly concentrated lyophilized proteins: part 1 amorphous formulations. J Pharm Sci. 2020;109(5):1681–91. 16. Kulkarni SS, Suryanarayanan R, Rinella JV Jr, Bogner RH. Mechanisms by which crystalline mannitol improves the reconstitution time of high concentration lyophilized protein formulations. Eur J Pharm Biopharm. 2018;131:70–81. 17. Beech KE, Biddlecombe JG, Van Der Walle CF, Stevens LA, Rigby SP, Burley JC, et al. Insights into the influence of the cooling profile on the reconstitution times of amorphous lyophilized protein formulations. Eur J Pharm Biopharm. 2015;96:247–54. 18. Kulkarni SS, Patel SM, Bogner RH. Reconstitution time for highly concentrated lyophilized proteins: role of formulation and protein. J Pharm Sci. 2020;109(10):2975–85. 19. Liao X, Krishnamurthy R, Suryanarayanan R. Influence of the active pharmaceutical ingredient concentration on the physical state of mannitol—implications in freeze-drying. Pharm Res. 2005;22(11):1978–85. 20. Neergaard MS, Kalonia DS, Parshad H, Nielsen AD, Møller EH, van de Weert M. Viscosity of high concentration protein formulations of monoclonal antibodies of the IgG1 and IgG4 subclass–prediction of viscosity through protein–protein interaction measurements. Eur J Pharm Sci. 2013;49(3):400–10. 21. Yearley EJ, Godfrin PD, Perevozchikova T, Zhang H, Falus P, Porcar L, et al. Observation of small cluster formation in concentrated monoclonal antibody solutions and its implications to solution viscosity. Biophys J. 2014;106(8):1763–70. 22. Depaz RA, Pansare S, Patel SM. Freeze-drying above the glass transition temperature in amorphous protein formulations while maintaining product quality and improving process efficiency. J Pharm Sci. 2016;105(1):40–9. 23. Pikal MJ, Shah S. The collapse temperature in freeze drying: dependence on measurement methodology and rate of water removal from the glassy phase. Int J Pharm. 1990;62(2–3):165–86. 24. Tang XC, Pikal MJ. Measurement of the kinetics of protein unfolding in viscous systems and implications for protein stability in freeze-drying. Pharm Res. 2005;22(7):1176–85.

120

X. (C). Tang et al.

25. Cleland JL, Lam X, Kendrick B, Yang J, Yang T, Overcashier D, et al. A specific molar ratio of stabilizer to protein is required for storage stability of a lyophilized monoclonal antibody. J Pharm Sci. 2001;90(3):310–21. 26. Weinbuch D, Hawe A, Jiskoot W, Friess W. Introduction into formulation development of biologics. In: Challenges in protein product development. Springer; 2018. p. 3–22. 27. Franks F, Hatley RHM, Friedman HL. The thermodynamics of protein stability: cold destabilization as a general phenomenon. Biophys Chem. 1988;31(3):307–15. 28. Marqués MI, Borreguero JM, Stanley HE, Dokholyan NV. Possible mechanism for cold denaturation of proteins at high pressure. Phys Rev Lett. 2003;91(13):138103. 29. Privalov PL. Cold denaturation of protein. Crit Rev Biochem Mol Biol. 1990;25(4):281–306. 30. Dill KA, Alonso DOV, Hutchinson K. Thermal stabilities of globular proteins. Biochemistry. 1989;28(13):5439–49. 31. Shortle D. The denatured state (the other half of the folding equation) and its role in protein stability. FASEB J. 1996;10(1):27–34. 32. Tang XC, Pikal MJ. The effect of stabilizers and denaturants on the cold denaturation temperatures of proteins and implications for freezedrying. Pharm Res. 2005;22(7):1167–75. 33. Franks F. Freeze-drying: from empiricism to predictability. The significance of glass transitions. Dev Biol Stand. 1992;74:9–18. 34. Dong J, Hubel A, Bischof JC, Aksan A. Freezing-induced phase separation and spatial microheterogeneity in protein solutions. J Phys Chem B. 2009;113(30):10081–7. 35. Roos Y. Melting and glass transitions of low molecular weight carbohydrates. Carbohydr Res. 1993;238:39–48. 36. Fang R, Bogner RH, Nail SL, Pikal MJ. Stability of freeze-dried protein formulations: contributions of ice nucleation temperature and residence time in the freeze-concentrate. J Pharm Sci. 2020;109(6):1896–904. 37. Pikal MJ. Mechanisms of protein stabilization during freeze-drying and storage: the relative importance of thermodynamic stabilization and glassy state relaxation dynamics. In: Freeze-drying/lyophilization of pharmaceutical and biological products, vol. 213. CRC Press; 2004. p. 210–44. 38. Strambini GB, Gabellieri E. Proteins in frozen solutions: evidence of ice-induced partial unfolding. Biophys J. 1996;70(2):971–6. 39. Chang BS, Kendrick BS, Carpenter JF. Surface-induced denaturation of proteins during freezing and its inhibition by surfactants. J Pharm Sci. 1996;85(12):1325–30. 40. Arsiccio A, McCarty J, Pisano R, Shea J-E. Heightened cold-denaturation of proteins at the ice–water interface. J Am Chem Soc. 2020;142(12): 5722–30. 41. Hottot A, Vessot S, Andrieu J. Freeze drying of pharmaceuticals in vials: influence of freezing protocol and sample configuration on ice morphology and freeze-dried cake texture. Chem Eng Process Process Intensif. 2007;46(7):666–74. 42. Nakagawa K, Hottot A, Vessot S, Andrieu J. Influence of controlled nucleation by ultrasounds on ice morphology of frozen formulations for pharmaceutical proteins freeze-drying. Chem Eng Process Process Intensif. 2006;45(9):783–91. 43. Carpenter JF, Chang BS, Garzon-Rodriguez W, Randolph TW. Rational design of stable lyophilized protein formulations: theory and practice. Pharm Biotechnol. 2002:109–33. 44. Chang BS, Randall CS. Use of subambient thermal analysis to optimize protein lyophilization. Cryobiology. 1992;29(5):632–56. 45. Heller MC, Carpenter JF, Randolph TW. Application of a thermodynamic model to the prediction of phase separations in freeze-concentrated formulations for protein lyophilization. Arch Biochem Biophys. 1999;363(2):191–201. 46. Heller MC, Carpenter JF, Randolph TW. Manipulation of lyophilization-induced phase separation: implications for pharmaceutical proteins. Biotechnol Prog. 1997;13(5):590–6. 47. Forney-Stevens KM, Pelletier MJ, Shalaev EY, Pikal MJ, Bogner RH. Optimization of a Raman microscopy technique to efficiently detect amorphous–amorphous phase separation in freeze-dried protein formulations. J Pharm Sci. 2014;103(9):2749–58. 48. Padilla AM, Ivanisevic I, Yang Y, Engers D, Bogner RH, Pikal MJ. The study of phase separation in amorphous freeze-dried systems. Part I: Raman mapping and computational analysis of XRPD data in model polymer systems. J Pharm Sci. 2011;100(1):206–22. 49. Rupley JA, Careri G. Protein hydration and function. Adv Protein Chem. 1991;41:37–172. 50. Kuhlman B, Yang HY, Boice JA, Fairman R, Raleigh DP. An exceptionally stable helix from the ribosomal protein L9: implications for protein folding and stability. Elsevier; 1997. 51. Prestrelski SJ, Tedeschi N, Arakawa T, Carpenter JF. Dehydration-induced conformational transitions in proteins and their inhibition by stabilizers. Biophys J. 1993;65(2):661–71. 52. Liao Y-H, Brown MB, Quader A, Martin GP. Protective mechanism of stabilizing excipients against dehydration in the freeze-drying of proteins. Pharm Res. 2002;19(12):1854–61. 53. Hédoux A, Paccou L, Achir S, Guinet Y. In situ monitoring of proteins during lyophilization using micro-Raman spectroscopy: a description of structural changes induced by dehydration. J Pharm Sci. 2012;101(7):2316–26. 54. Nagendra HG, Sukumar N, Vijayan M. Role of water in plasticity, stability, and action of proteins: the crystal structures of lysozyme at very low levels of hydration. Proteins: Struct Funct Bioinforma. 1998;32(2):229–40. 55. Pikal MJ. Freeze-drying of proteins. Part II: Formulation selection. BioPharm. 1990;3(9):26–30. 56. Van den Berg L, Rose D. Effect of freezing on the pH and composition of sodium and potassium phosphate solutions: the reciprocal system KH2PO4-Na2HPO4-H2O. Arch Biochem Biophys. 1959;81(2):319–29. 57. Kerwin BA. Polysorbates 20 and 80 used in the formulation of protein biotherapeutics: structure and degradation pathways. J Pharm Sci. 2008;97(8):2924–35. 58. Zhang MZ, Pikal K, Nguyen T, Arakawa T, Prestrelski SJ. The effect of the reconstitution medium on aggregation of lyophilized recombinant interleukin-2 and ribonuclease A. Pharm Res. 1996;13(4):643–6. 59. Liu J, Nguyen MDH, Andya JD, Shire SJ. Reversible self-association increases the viscosity of a concentrated monoclonal antibody in aqueous solution. J Pharm Sci. 2005;94(9):1928–40. 60. Dixon D, Tchessalov S, Barry A, Warne N. The impact of protein concentration on mannitol and sodium chloride crystallinity and polymorphism upon lyophilization. J Pharm Sci. 2009;98(9):3419–29. 61. Tang XC, Nail SL, Pikal MJ. Freeze-drying process design by manometric temperature measurement: design of a smart freeze-dryer. Pharm Res. 2005;22(4):685–700.

Challenges and Considerations in the Development of a High Protein Concentration. . .

121

62. Tang XC, Pikal MJ. Design of freeze-drying processes for pharmaceuticals: practical advice. Pharm Res. 2004;21(2):191–200. 63. Perez-Ramirez B, Geiger C, Mezhebovsky T, Bussemer T. Approaches in subcutaneous delivery of monoclonal antibodies. Eur Pharm Rev. 2016;4 64. Shamblin SL, Tang X, Chang L, Hancock BC, Pikal MJ. Characterization of the time scales of molecular motion in pharmaceutically important glasses. J Phys Chem B. 1999;103(20):4113–21. 65. Chang LL, Shepherd D, Sun J, Tang XC, Pikal MJ. Effect of sorbitol and residual moisture on the stability of lyophilized antibodies: implications for the mechanism of protein stabilization in the solid state. J Pharm Sci. 2005;94(7):1445–55. 66. Chauhan VM, Zhang H, Dalby PA, Aylott JW. Advancements in the co-formulation of biologic therapeutics. J Control Release. 2020;327:397– 405. 67. Du J, Shah A. Compositions comprising coformulation of anti-PD-L1 and anti-CTLA-4 antibodies. Google Patents; 2019. 68. Adamson AW. Contact angle physical chemistry of surfaces; 1990. p. 385–8. 69. Geidobler R, Konrad I, Winter G. Can controlled ice nucleation improve freeze-drying of highly-concentrated protein formulations? J Pharm Sci. 2013;102(11):3915–9. 70. Tang X, Nail SL, Pikal MJ. Evaluation of manometric temperature measurement, a process analytical technology tool for freeze-drying: part I, product temperature measurement. AAPS PharmSciTech. 2006;7(1):E95–E103. 71. Pikal MJ, Roy ML, Shah S. Mass and heat transfer in vial freeze-drying of pharmaceuticals: role of the vial. J Pharm Sci. 1984;73(9):1224–37. 72. Stärtzel P, Gieseler H, Gieseler M, Abdul-Fattah AM, Adler M, Mahler H-C, et al. Freeze drying of l-arginine/sucrose-based protein formulations, part I: influence of formulation and arginine counter ion on the critical formulation temperature, product performance and protein stability. J Pharm Sci. 2015;104(7):2345–58. 73. Rambhatla S, Obert JP, Luthra S, Bhugra C, Pikal MJ. Cake shrinkage during freeze drying: a combined experimental and theoretical study. Pharm Dev Technol. 2005;10(1):33–40. 74. Roy ML, Pikal MJ. Process control in freeze drying: determination of the end point of sublimation drying by an electronic moisture sensor. PDA J Pharm Sci Technol. 1989;43(2):60–6. 75. Rambhatla S, Pikal MJ. Heat and mass transfer scale-up issues during freeze-drying, I: atypical radiation and the edge vial effect. AAPS PharmSciTech. 2003;4(2):22–31. 76. Pikal MJ. Use of laboratory data in freeze drying process design: heat and mass transfer coefficients and the computer simulation of freeze drying. PDA J Pharm Sci Technol. 1985;39(3):115–39.

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water Formulations Eni Bogdani, Séverine Vessot-Crastes, and Julien Andrieu

Abstracts

The organic co-solvent + water formulations present many interesting advantages for the freeze-drying of thermosensible drugs to produce stabilized and elegant powders of marketed pharmaceuticals products: high freezing temperatures, very short sublimation times, low sublimation enthalpies, high equilibrium vapour/frozen solid pressures and low toxicity of the residual solvent contents. The formulations that have been the most investigated have concerned certainly tert-butanol (TBA) + water mixtures. Thus, some main characteristics of the water + TBA formulations have been reviewed, especially its interesting thermodynamical properties (sublimation enthalpies; equilibrium vapour pressures), the impact of freezing conditions on morphological properties of frozen formulations (nucleation, crystals size and shape), the influence of operating parameters (total pressure, temperature) on sublimation times and finally on organic co-solvent and water residual contents. The crystals morphology of frozen formulations prepared with tert-butanol (TBA) revealed unexpected results compared with the results reported in the literature for water-based formulations, pointing the complex relationships between freezing rates, supercooling, nucleation temperatures and solvent crystals morphology (size and shape). To illustrate the functional relationships between the sublimation rates, the mean product temperatures and the two principal independent process variables, namely the shelf heat transfer fluid temperature and the total gas pressure, we used the Design Space concept presented as an envelope in a graph with sublimation rates and total gas chamber pressures on main (x, y) axis. The limits of this Design Space are determined by the influence of product and process variables on main quality attributes of the freeze-dried drug, more precisely by the failure of these attributes under aggressive cycle conditions. In the case of large industrial freeze-dryers, other limits are also imposed by freeze-dryer performances. Next, as an illustration of this concept, we have presented a case study concerning methodology of construction of the Design Space for an ibuprofen organic co-solvent (TBA)-based formulation determined with a pilot laboratory freezedryer. It proved that the setting up of optimum freeze-drying cycles has to be realized by taking into account simultaneously the impact of formulation variables, especially the tert-butanol content and the classical freeze-drying variables at the freezing step (nucleation temperatures, freezing rates) and at the two drying steps (shelf temperature, total gas pressure) to maximize the drying rates and to minimize the residual solvent levels while preserving the main quality attributes of the freeze-dried powder. Keywords

Freeze-Drying · tert-butanol · Vapour pressure · Sublimation enthalpies · Crystals morphology · Sublimation rates · Optimization · Annealing · Design space · Residual solvent

E. Bogdani · S. Vessot-Crastes · J. Andrieu (✉) University of Lyon, University of Lyon 1, Laboratoire Automatique et de Génie des Procédés. LAGEP, Lyon, France e-mail: [email protected]; [email protected] # America Association of Pharmaceutical Scientists 2023 F. Jameel (ed.), Principles and Practices of Lyophilization in Product Development and Manufacturing, AAPS Advances in the Pharmaceutical Sciences Series 59, https://doi.org/10.1007/978-3-031-12634-5_7

123

124

1

E. Bogdani et al.

Introduction

Freeze-drying is the process by which the solvent is removed from a frozen solution by sublimation [1, 2]. The freeze-drying process may be divided into three steps: freezing, primary drying (or sublimation) and secondary drying (or desorption). A large majority of marketed pharmaceutical products are freeze-dried with simple aqueous solutions. In this case, water is typically the only solvent component which is removed from the solid phase, firstly by sublimation and next by desorption. However, some hydrophobic and insoluble active principle ingredient (API) cannot be freeze-dried adequately with waterbased formulations, so that pure organic solvent or organic co-solvent + water formulations have been also investigated in the recent years. The main advantages of using non-aqueous solvents are: increased drug solubility, great acceleration of sublimation rates, decreased reconstitution times, improvement of freeze-dried product stability. On the other hand, the possible disadvantages concern: the required freeze-dryer low condenser temperature, the safe handling or storage of flammable components, the possible toxicity and the control of residual solvent contents, the overall costs benefits and the increase of regulatory policies for using organic solvents. An exhaustive review concerning freeze-drying with various organic co-solvents was published by Teagarden et al. [3], Vessot et al. [4] and other data are reported by Rey et al. [5] These authors compiled the different published studies and papers on these topics and they listed the case studies with these formulations especially by using tert-butanol (TBA) and water mixtures or pure TBA (Table 1). In a previous paper, Daoussi et al. [6] reported sublimation kinetics data of formulations with an active principle ingredient (API) by using organic co-solvent + water mixtures. These authors observed that the use of organic co-solvents like tertbutanol (TBA) reduced considerably the sublimation times – divided by about a factor 10. Next, in the same laboratory, in the purpose of physically explaining these data, Bogdani et al. [7] investigated extensively some thermodynamical properties of these systems at frozen state. The sublimation enthalpies with organic systems were much lower than the corresponding values observed with formulations using pure water. For example, at the interesting eutectic composition at 90% (w/w) of TBA, the equilibrium vapour pressure values at solid state were found to be about twice as high as the data obtained with pure water. Thus, these new thermodynamical data sets explained physically why the organic formulations present very interesting sublimation times and the respective position of the sublimation kinetics curves corresponding respectively to the eutectic composition- formulation at 90% (w/w) of TBA – with respect to the pure TBA or to 80% (w/w) of TBA formulations. Moreover, by using the direct observation method by optical microscopy with episcopic coaxial lighting in cold chamber, Nakagawa et al. [8] and Hottot et al. [9] observed that the nature of formulation itself (organic co-solvent concentration) and the freezing conditions—namely the nucleation temperature and the freezing rates—had a key role in controlling the morphology, and incidentally, the stability and the reconstitution properties of the final freeze-dried cake. Moreover, Cui et al. [10] used TBA + water formulations to prepare dehydrated liposomes by freeze-drying. These authors also observed that the addition of TBA co-solvent to water-based solutions significantly enhanced the sublimation kinetics leading to short times of freeze-drying cycles. More recently, Zhang et al. [11] investigated the freeze-drying of water + TBA Table 1 Some athermophysical properties of solvents used for drugs freeze-drying formulations

Solvent Ter-butanol Ethanol n-Propanol n-Butanol Isopropanol Ethyl acetate Dimethyl carbonate Acetonitrile Dichloromethane Methyl ethyl ketone Methyl isobutyl ketone

Solubility in water (%)a 100 100 100 7.7 100 8.7 9.5

Vapor pressure (mm Hg at 20°C 26.8 41.0 14.5 5.6 31.0 64.7 72

Freezing point (° C) 24.0 -114 -127 -90 -89.5 -84 2

Boiling point (° C) 82 78.5 97.1 117.5 81 77.1 90

Flash point (OFrC) 52/11 62/16 59/15 95/35 53.6/12 24/-4 65/18

Autoignition temperature (OFrC) 892/478 793/423 760/404 689/365 750/398 800/426 –

Lower flamm limit (in air vol, % 2.4 3.3 2.1 IA 2.5 2.2 4.2

Upper flamm limit (in air vol, % 8.0 19 13.5 11.2 12 11.5 12.9

100 13 27

69.8 343.9 76.2

-48 -97 -87

80.1 40 79.6

45/8 none 26/-3

975/524 1033/556 885/474

4.4 14 1.7

16.0 22 10.1

2.0

5.1

-80

117

56/13

860/460

1.2

8

Teagarden and Baker [3]

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water. . .

125

Table 2 Pharmaceutical drugs freeze-dried with organic co-solvents + water formulations Drug Alprostadil Aplidine Amoxicillin sodium Tobramycin sulfate Gentamycin sulfate N-cyhclodexyl-N-Methy-4-(2-oxo-I, 2,3,5-tetrahydroimidazo-[2,1-b] quinazolin-7-yl) oxybutyramide with ascorbic acid Cyclohexane-1,2-diamine Pt (II) complex Annamycin Cephalothin sodium Cephalothin sodium Prednisolone acetate/polyglycolic acid Gabexate mesylate Piraubidin hydrochloride Progesterone, Coronene, Fluasterone, Phenytoin Fructose-1,6-diphosphate Poly(lactide-co-glycolide) Dioleoylphosphatidylcholine and dioleoylphosphatidylglycerol Vecuroniumbromide Bovine pancreatic trypsin inhibitor

Co-solvent system 20% v/v tert-butanol/water 40% v/v tert-butanol/water 20% v/v tert-butanol/water Tert-butanol/water Tert-butanol/water 50% v/v tert-butanol/water Tert-butanol/water Tert-butanol/dimethyl 1 sulfoxide/water 5% w/w isopropyl alcohol/water 4% ethanol, 4% methanol, or 4% acetone/ water Carbon tetrachloride/hexafluoroacetone sesquihydrate Ethanol/water Ethanol/water Chlorobutanol hemihydrate/dimethyl sulfone Tert-butanol/water Acetic acid Cyclohexane Acetronitrile Dimethyl sulfoxide/1% water

Teagarden and Baker [3]

organic formulations to stabilize insulin and they evaluated how the process and the formulation variables affected the structural stability and the activity of this important pharmaceutical protein largely commercialized. Elgindy et al. [12] studied the freeze-drying process of flutamide—an anticancer drug for prostatic carcinoma—with tert-butanol (TBA) + water mixtures to increase the solubility and to improve some important physico-chemical properties (amorphization, dissolution rate, surface area, etc.) impacting this drug delivery. Other co-solvent systems like different types of alcohols such as methanol, n-propanol and n-butanol have been also investigated in the past [3, 5]. They do not freeze in standard pilot or laboratory freeze-dryers due to their very low freezing temperature: this represents important drawbacks with respect to the operating costs of the whole freeze-drying process. Besides, the literature reports that most of these formulations generally led to unacceptable freeze-dried cakes (poor appearance, collapse, reconstitution, etc.) with a drastic degradation of the drug therapeutical activity [3, 4] (Table 2). Thus, the objective of this chapter is to present a methodology of optimization of the whole freeze-drying cycle with numerous and accurate physico-chemical data concerning the TBA + water system showing the influence of freezing and sublimation parameters with a model formulation usually selected with thermosensible pharmaceutical drugs contained in glass vials. The freezing and sublimation operating conditions have been chosen in the same range as encountered for developing freeze-drying cycles at commercial scale with these types of products. Thus, with this whole set of data, by using the concept of Design Space, Bogdani et al. [13] presented a rational optimization for the whole freeze-drying cycle.

2

Preparation of Solution

The first step to set-up a freeze-drying process is to be able to realize a more or less homogeneous solution containing the active principle ingredient (API)—with its additives or excipients—that has to be stabilized during storage before its reconstitution and its final delivery in the human body. As many API and/or excipients are hydrophobic and, consequently, difficult to dissolve, the use of organic co-solvent as tert-Butanol highly facilitates the solubilization, decreases the time to achieve the dissolution and reduces the amount of solvent that must be eliminated during the drying steps. Several examples of increased drug solubility with TBA have been extensively listed by Teagarden et al. [3] and Vessot et al. [4] and Daoussi et al. [6] Another important challenge in developing drug manufacturing process concerns its instability in solution during the steps

126

E. Bogdani et al.

of filtering (for sterilizing effect) and of vial filling that precedes the proper freeze-drying steps and during which some degradations can occur depending on the type of degradation kinetics. Thus, Teagarden et al. [3] in their literature review, present some data for different types of drugs showing that the presence of TBA can have a significant positive effect on the chemical solution stability by decreasing the degradation reaction kinetics.

3

Freezing Parameters

The first stage of a freeze-drying process is always a freezing step, during which the crystals of solvent—more often ice crystals for water-based solutions or mixed ice and solvent (TBA) crystals in the case of organic co-solvent formulations— separate from a cryo-concentrated phase containing the API and its excipients. The cooling process that involves complex supercooling, nucleation (homogeneous and/or heterogeneous) and crystals growth phenomena is maintained up to the transformation of all the liquid formulations into a frozen solid. This freezing step results in a complex mixture of crystalline, amorphous or metastable phases that are deeply affected by the concentrations of TBA and of the other additives as reported in details by Seager et al. [14] Owing to these authors, the solvent type and its concentration affect the structure and the morphology of the frozen material phases (crystallinity or amorphous state). In some cases, the use of organic solvents can result in drug precipitation due to some solvent vaporization. In the case of high melting point solvents as tert-Butanol, it probably happens that the solvent (TBA) crystallizes between the ice matrix as the temperature is decreased so that the presence of tert-Butanol alters the crystal habit of the regular ice matrix. The solvent’s crystals morphology of TBA + water systems has been extensively investigated by microscopy in cold chamber by Daoussi et al. [15]

3.1

Nucleation Temperature

Nakagawa et al. [8], Daoussi et al. [15] Pikal et al. [16] and Kasper et al. [17] confirmed the stochastic nature of nucleation phenomena which are certainly affected randomly by impurities or other heterogeneities. These authors concluded that the nucleation temperature, which is a non-deterministic parameter, was nevertheless the more relevant parameter that determines the ice crystals’ overall structure and the primary drying times. Thus, Daoussi et al. [15] determined from the temperature profiles with samples contained in glass vials, the nucleation temperatures, noted TN, for TBA + water mixtures prepared and cooled in the same rigorous conditions and submitted to the same standard cooling rate equal to 1 °C/min for three formulations at 80% (w/w), 90% (w/w) to 100% (pure) TBA contents. The nucleation probability was defined by the ratio of the number of observed TN values to the number of total repetitive runs; the experimental runs were repeated about 20 times for each of the three above-mentioned formulations. Experimental results showed that about 60% of the samples prepared from mixtures with 80% (w/w) of TBA nucleated at -15 °C (Fig. 1) while samples prepared from mixtures with 90% (w/w) of TBA nucleated mainly at -23 °C (Fig. 2). For pure TBA solvent formulations, these authors found that the probability distribution curve presented two nucleation peaks, the first one occurring at about +15 °C and the second one at about -30 °C. Thus, it proved that all these formulations presented different nucleation temperatures probabilities, so that the initial concentration of TBA was an important factor that governs the nucleation phenomenon of these systems. Moreover, previous studies of ice crystals growth phenomena in our laboratory reported by Nakagawa et al. [8] have shown that the size and the shape of ice crystals are strongly related to nucleation temperatures. Consequently, some quality attributes of the freeze-dried cake (permeability, morphology, reconstitution times) are strongly related to the initial TBA content and to the nucleation rates which govern the number of ice crystals and also to the rate of heat transfer which governs the growth rate of solvent crystals after the nuclei formation.

3.2

Solvent Crystals Morphology

Another reason that could explain the great advantages of using organic formulations relies on the differences in frozen samples’ morphologies observed with organic solvents due to the different nucleation and crystal growth mechanisms in comparison with water-based formulations. Daoussi et al. [15] investigated these morphological characteristics with frozen organic formulations by direct photonic microscopy in cold chamber (direct observations at -25 °C). Indeed, it is well known that for aqueous-based systems, the frozen material morphology is directly related to the morphology of the freeze-dried matrix containing the API to be stabilized. This study showed that the crystals were heteregeneous in size and in shape. With

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water. . .

70 60 Frequency distribution (%)

Fig. 1 Nucleation probability of 20% (w/w) water + 80% (w/w) TBA mixture inside 3 mL vials cooled at 1 °C/min. (Daoussi et al. [15])

127

50 40 30 80% TBA 1°C/m in

20 10 0 -40

-30

-20

-10

0

10

20

Nucleation temperature (°C)

Fig. 2 Temperature evolution during the cooling of a formulation at 90% (w/w) of TBA. Cooling rate at 1 °C/min. Nucleation at about -23 °C. (Daoussi et al. [15])

20

Temperature (°C)

10 0 -10 -20 -30 -40 -50 0

20

40

60

80

100

120

140

160

Time (min)

samples corresponding to 90% (w/w) of TBA, crystals were pretty small and rather spherical. With samples prepared from 100% (pure) TBA the crystals had a lamellar shape. For 80% (w/w) of TBA mixture, the solvent crystals presented an intermediate shape between lamellar and dendritic shapes. Moreover, the cumulative distributions of the mean solvent crystal sizes presented on Fig. 3 (formulation at 80% of TBA by mass) show that the mean crystals sizes increase very significantly with the freezing rates. These results are consistent with those of a previous study of the same authors with the same formulations for which they observed that the degree of super cooling decreased as the freezing rate increased. However, these observations are in strong contrast with previous results reported for the nucleation of purely aqueous system for which the degrees of super cooling increased when the cooling rates increased leading to smaller crystal sizes when the freezing rates increased [1, 2, 4, 5]. These tendencies were confirmed with pure TBA formulations at the two other freezing rates investigated, namely at 0.1 °C/min and at rapid freezing (liquid nitrogen).

128

E. Bogdani et al.

Fig. 3 Solvent crystal sizes cumulative distribution curves for different freezing rates. TBA initial concentration at 80% (w/w). (Daoussi et al. [15])

1

Relative cumulative distribution

0.8

0.6

0.4 80%TBA 0.1°C/m in 80%TBA 1°C/m in

0.2

80%TBA Rapid freezing

0 0

50

100

150

200

Crystal diameter (µm)

Fig. 4 Solvent crystal sizes cumulative distribution curves for different freezing rates. TBA initial concentration at 90% (w/w). (Daoussi et al. [15])

1

Relative cumulative distribution

0.8

0.6 90%TBA 0.1°C/m in

0.4

90%TBA 1°C/m in 90%TBA Rapid freezing

0.2

0 0

50

100

150

200

Crystal diameter (µm)

Contrary to the results observed with pure TBA formulations, the crystal sizes obtained with formulations at 90% of TBA (Fig. 4) did not vary significantly with the freezing rates, 50% of crystals having average sizes lying between 17 μm and 35 μm. Thus, the crystal sizes obtained with 90% TBA mixtures are much smaller—with sizes about seven times smaller— than the sizes observed with samples prepared from formulations with 80% (by w/w) (Fig. 3). Furthermore, with the formulation at 90% (by w/w) of TBA, the width of the dispersion was much smaller which means that the crystal sizes were found more homogeneous. These last results showed again that the eutectic mixture at 90% of TBA presented significant different behaviours during freezing and also different physical properties compared to the two other formulations investigated. Daoussi et al. [15] concluded that the solvent crystals morphology of frozen formulations prepared with tertbutanol (TBA) revealed unexpected results compared with the results reported in the literature for water-based formulations. Consequently, the TBA co-solvent concentrations and the freezing rates had a strong influence on the solvent crystals morphology, and, consequently, on one side on some quality factors related to the freeze-dried cake morphology and on the other side, on the operating costs of the whole process (sublimation and desorption times).

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water. . .

4

Thermodynamical Properties

4.1

Phase Diagram

129

Many experimental studies have shown that with systems that contain one solute that gives a crystalline phase at solid state, the primary drying step must be conducted below the eutectic temperature [1, 2, 4, 5]. Nevertheless, most drug formulations do not crystallize or do not have high eutectic temperatures so that eutectic temperature limit is not very often an issue in freeze-drying cycles optimization if we except the systems containing high concentrations of NaCl, this salt having an eutectic temperature around -20 °C. Otherwise, with amorphous or glassy systems, the sublimation step must be carried out below the collapse temperature, which is generally a few degrees above the vitreous transition temperature of the solute phase. If the product is sublimated above the vitreous transition temperature, the solid phase collapses and has enough fluidity to induce a viscous flow when the ice is removed from the frozen sample. This collapse will lead to a loss of cake mechanical structure and will result in higher moisture contents and into a loss of biological activity, notably with some pharmaceutical protein formulations (vaccines). Some other defects like an increase of the rehydration times or some loss of pharmaceutical elegance can also occur. This is why the state diagram, which is greatly characteristic of the formulation composition, represents the basic thermodynamical data absolutely necessary for setting up rationally the optimal freeze-drying cycle. The phase diagram of TBA + water system has been published by Kasraïan et al. [18] and is plotted on Fig. 5. This system presents an eutectic point, noted B, at -3.3 °C for a TBA concentration at 90% (w/w). Below this eutectic point temperature, solid phases of pure frozen TBA and of TBA dehydrate probably co-exist. This mixture composition presents some singular morphological characteristics at frozen state and also some interesting water pressure values for sublimation applications as reported by Daoussi et al. [6] Another eutectic point at high water concentration exists at – 8.2 °C with TBA concentration near 15% (w/w). Thus, Daoussi et al. [6] selected this eutectic mixture – at 90% (w/w) of TBA – as the most interesting mixture for achieving the stabilization and the amorphization of their thermosensible API. The solvent crystals that sublimated during their extensive study of their organic co-solvent freeze-drying process were probably pure TBA and pure TBA dihydrate crystals with separated or mixed crystalline lattices.

4.2

Equilibrium Vapour Pressures

The vapour pressures are also important data for setting up and for validating physical models of freeze-drying processes. These data for pure TBA and for the + 10% (w/w) water system (eutectic composition) were determined in our laboratory by Bogdani et al. [7] by using the new rapid thermogravimetric method. For pure TBA, these authors observed that the equilibrium solid/vapour pressures values were quite equal to the values of pure ice/water vapour system what seems quite

Fig. 5 Phase diagram of TBA + water system plotted on fraction (by mass). (From Kasraïan et al. [18])

130

E. Bogdani et al.

Fig. 6 LnP0 vs. 1/T plots for the 90% (w/w) TBA) + 10% (w/w) water system (□) and for the pure water (*). (Bogdani et al. [7])

12 Solid

Liquid

10

lnPo (Pa)

8

Eutectic B triple point (907 Pa, -5°C)

6 Ice triple point (610 Pa, 0°C)

4

Vapour

2 0 220

240

260

280 T (K)

300

320

340

surprising due to the usual high volatility of organic compounds at liquid/vapour equilibrium. These data, which are very useful for the physical modelling of the sublimation process, could be correlated by the following equation [7]:

 P0 = exp A þ B=T þ DT F =T C where P0 represents the equilibrium vapour pressure (Pa) and T the absolute temperature (K) and where the coefficient constant values (S.I. units) are equal to A = 172.3; B = -11,590; D = 1.37×10–5; F = 2; C = 22.118; On the contrary, for the eutectic composition at 90% (w/w) of TBA, the same authors observed that the equilibrium frozen solid/vapour pressures values were about twice as high as the vapour pressures of the pure ice (Cf. Fig. 6). Moreover, it is worth noting that this eutectic mixture corresponds to a positive azeotropic composition in the gas/liquid equilibrium state what could physically explain these high equilibrium vapour pressure values at this particular composition. Thus, this particular composition presents the most favourable thermodynamical data to increase the sublimation rates in case of a sublimation process controlled by the mass transfer resistance of the solvent vapour through the dry layer as discussed in the paper of Daoussi et al. [6, 15]

4.3

Sublimation Enthalpies

Indeed, the equilibrium solid–vapour pressures are not the only thermodynamic parameters that could affect directly the sublimation rates. The enthalpies of sublimation also influence directly the sublimation rates, especially when the heat transfer is the limiting mechanism as observed by Daoussi et al. [6] and by Hottot et al. [19, 20] in standard conditions for the freezedrying of very thermosensible API as pharmaceutical proteins. These values obtained by Bogdani et al. [7] (Cf. Table 3) proved to be 3.8 times lower than the corresponding value for pure ice. Quite the same factor for the latent heat of sublimation was also observed for the eutectic mixture at 90% (w/w) of TBA + 10% (w/w) of water. Thus, these low values of sublimation enthalpies represent the main factor that explains the very high increase of the sublimation rates observed by using tertbutanol (TBA)-based formulations.

5

Sublimation Kinetics

During the sublimation step, the mass transfer of the solvent vapour from the sublimation front up to the condenser has overcome three mass transfer resistances that can be considered as additive, namely the dried layer resistance—which increases with the dry layer thickness and which is the most important [1, 2, 19, 20] — the resistance of the vial stopper

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water. . .

131

Table 3 Experimental sublimation enthalpies and melting temperatures for 90% (w/w) TBA + 10% (w/w) water and for pure TBA formulations water Pure TBA 90% TBA

Tm (°C) 0 25 5

ΔHsub (kJ/kg) 2809 732

ΔH subice =ΔH subTBA 1 3.8

1027

2.7

Bogdani et al. [7]

Fig. 7 Drying rate as a function of total solvent concentration. Influence of the initial content of organic co-solvent (TBA). Ts = -20 °C and Pt = 6 Pa. Freezing rate: 1 °C/min. (Daoussi et al. [6, 15])

-3

1.5

x 10

80% (w/w) TBA

100% (w/w) TBA

90% (w/w) TBA

dX/dt (s-1)

1

0.5

0 0

100

200

300

400

500

600

700

800

Time(s)

(often negligible) and the chamber resistance that depends on the vapour flow regime and of the chamber geometry. The mass transfer in the porous dried layer takes place via two main mechanisms, namely either by viscous flow under a total pressure gradient or by molecular diffusion under gradients of concentration in Fick’s regime which generally occurs at moderate chamber total pressures—or in Knudsen regime when the mean free path of solvent vapour molecules becomes higher than the mean pores diameter. The molecular diffusion in Knudsen regime generally occurs at low shelf temperatures and very low total gas pressures, conditions which correspond to the optimum conditions of sublimation for thermosensible drugs as vaccines [1, 2, 4, 15]. In all these regimes, the morphology of the pores of the dried layer (permeability, mean pores diameter, pores surface area, etc.) has an important impact on the intensities of the solvent vapour fluxes through the dried layer. This morphology is similar to the morphology of the crystals of the frozen solvent, morphology that is mainly fixed at the nucleation and crystals growth steps, what explains the great practical importance of these steps that are generally neglected or not taken in account in freeze-drying cycle development studies. Daoussi et al. [6] carried out different freeze-drying experiments at constant shelf temperatures and constant total gas pressures—respectively equal to Ts = -20 °C and Pt = 6 Pa – for different water + TBA mixtures from 80% (w/w) to (pure) 100% TBA formulations. According to the data in Fig. 7, it is clear that, at quasi-constant drying rate period, the highest drying rates values correspond to the co-solvent mixture containing 90% (w/w) of TBA. As already pointed out in the previous paragraph, this particular composition corresponds to the formation of a positive azeotrope at the gas/liquid state with a maximum of partial pressures, which probably correspond to an increase of the vapour/solid equilibrium pressures at the frozen state and, consequently, to an increase of the solvent vapour mass transfer driving force during the sublimation [2, 4, 6, 19, 20]. Thus, these authors [15] pointed out that the use of mixtures highly concentrated with organic co-solvent (TBA) resulted in very short sublimation times – about 3 h – these values being about 10 times shorter than the values observed with pure aqueous-based formulations under the same operating conditions. With their API formulations, this great increase was

132

E. Bogdani et al.

maximum at the eutectic composition certainly due to the favourable thermodynamical properties of the frozen formulation (high equilibrium vapour pressures and low sublimation enthalpies) observed at this particular composition. Moreover, the shelf temperature had the higher influence on the sublimation times, but the total gas pressure, the vial type and the vial filling could also have significant influence on the primary drying (sublimation) times. Thus, Daoussi et al. [6] concluded that in their particular conditions – at very low total gas pressure (Pt = 10–30 Pa) and at low product temperature (between -20 °C and 40 °C) – the heat transfer was the main governing phenomenon of the sublimation process. These high acceleration rates of sublimation times during freeze-drying of lactose and sucrose solutions with TBA formulations are also reported in Teagarden’s review [3]. The same enhancement of the primary drying kinetics (sublimation) was also observed by Baldi et al. [21] during the freeze-drying of gentamicin sulphate (model drug) with water+maltose+TBA formulations and more recently by Cui et al. [10] during freeze-drying experiments of liposomes by using tert-butanol (TBA) as co-solvent with water mixtures.

6

Residual TBA and Water Contents

The retention of volatile has been largely studied during optimization of food freeze-drying processes. It is advantageous to trap these components (aroma) in the freeze-dried matrix in order to maintain at their best the original organoleptic properties of the foodstuff. In the case of drug freeze-drying processes, it is generally an opposite retention effect that is looked for; too high residue solvent contents generally have a negative impact on the storage of the freeze-dried drug and on the preservation of its therapeutical properties. In the case of freeze-drying with organic co-solvent formulations, these constraints can be increased due to possible toxicity of these solvents residues. The impact of formulation and process variables on residual solvent levels for formulations freeze-dried with tert-butanol (TBA) have been critically reviewed by Teagarden et al. [3], Chouvenc [22], Wittaya-areekul et al. [23] and more recently by Daoussi [24], Gieseler [25]. These authors investigated the effect of the physical state of solute (amorphous or crystalline), the initial TBA concentration, the freezing rate, the cake thickness and the temperature during the secondary drying (desorption step). They pointed out that the TBA residual contents were very low in the case of formulations leading to a crystalline matrix (0.01–0.03%) regardless of freezing rates or initial TBA concentrations. On the contrary, very high residual TBA contents were obtained with amorphous systems obtained with sucrose solutions. Moreover, processing conditions during the secondary drying period had a deep impact on residual levels at the end of the freeze-drying cycle. Similar results were obtained by Teagarden [3] during the freeze-drying of lactose solutions with TBA + water co-solvent mixtures. Otherwise, freezing conditions seemed to have an influence on residual solvent contents only for amorphous systems because very fast freezing (with liquid nitrogen) leads to high residual levels of TBA. Daoussi et al. [24] have evaluated, in the case of API + TBA+ water system, the impact of the formulation of the freezing conditions (freezing rates; annealing) and of the duration of the sublimation step on the co-solvent (TBA) and water residual contents. Their data have shown that the residual TBA content increased with the initial concentration of the active principle ingredient (API). As concerning the impact of the operating conditions, the TBA residual contents increased or decreased with the initial concentration of TBA.

7

Annealing Treatment

It is well known that annealing treatments homogenize the size and the shape of the solvent or the ice crystals by a water or co-solvent re-crystallization mechanism as observed by Chouvenc et al. [22] with water-based formulations. Daoussi et al. [24] observed with their active principle ingredient (API) dissolved in tert-butanol (TBA)-based formulations that the co-solvent (TBA) residual content increased with the application of efficient annealing treatments. Moreover, Daoussi et al. [24] found that the freezing rates and the annealing treatments did not have any significant impact on the residual water contents of the freeze-dried cakes, on one side, and, that the residual water contents increased with the API concentrations, on the other side. These authors reported that adequate annealing treatments reduced notably the duration of the sublimation period by about a factor 2. This annealing treatment was generally more efficient for formulations prepared with pure TBA as unique API solvent than with water + TBA mixtures containing respectively 80% or 90% (w/w) of TBA.

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water. . .

8

133

Design Space for Freeze-Drying of an Ibuprofen Formulation

In order to illustrate this quite new concept of Design Space, which is more and more used in pharmaceutical industry R & D projects, we have chosen to deeply investigate the freeze-drying process of a single formulation containing ibuprofen as active principle ingredient (API), a drug largely used as anti-inflammatory drug in many countries. With this case study, we will show the methodology of practical and experimental determination of the Design Space graph as proposed by Nail et al. [26], by Chang et al. [27], by Fissore et al. [28] and by Giordano et al. [29] for the freeze-drying process with a laboratory-type pilot freeze-dryer. For these reasons, we have chosen to experimentally study the kinetics of the sublimation during the freeze-dying process of a system constituted of ibuprofen + solvent mixture containing 20% (w/w) of organic co-solvent TBA + 80% (w/w) of water (eutectic A) and to set up the corresponding Design Space. The phase diagram of this system has been reported in the literature by Kasraïan et al. [18] and, moreover, some important and key thermodynamical properties of this system necessary for the sublimation step modeling have been determined recently by Bogdani et al. [7] in our laboratory.

8.1

Overall Heat Transfer Characterization

In order to characterize the external heat transfer between the surroundings (shelf, plate, chamber wall, etc.) and the sublimation front and the product inside each single vial, we have determined experimentally, in situ, the real values of the overall heat transfer coefficient of our system, noted Kv, values that are dependent on the operating conditions (Tshelf and Pc) and on the vial type and geometry as observed by Hottot et al. [19, 30] in our laboratory. Nevertheless, these values are quite independent of the nature of the formulation, so that, in the case of costly API it is experimentally possible to run these thermal characterization experiments by filling the vials with the only solvent (without API and excipients) Thus, in this configuration, there exists no dried layer and, consequently, no dry layer mass transfer resistance for the solvent vapor transfer. Then, the determination of Kv values was realized by Hottot et al. [30] by using vials filled with the only solvent without any API and excipients. A set of 30 vials were filled with 1 g of liquid solvent in each vial (filling height around 10-2 m) at the eutectic composition of our system referenced as eutectic A [13, 30], and then frozen by following the procedure described above. The mass of one representative vial was continuously weighed with the microbalance Christ placed inside the shelf of the sublimation chamber. From the mass loss curve it was possible to calculate the average values of the sublimation rate, _ for each set of fixed shelf temperature and total gas pressure values in the sublimation chamber. noted m, This value of m_ corresponds to the slope value of the linear part of this curve. On the bottom of two other vials, thin thermocouples were placed in order to record the mean product temperature profiles (Tb = f(t)). From Fig. 8, we observed a product temperature plateau at Tb = -35.5 °C. This plateau means that the system has reached a quasi-stationary state, during which all the energy supplied to the product was essentially used for the sublimation of the solvent. All along this quasistationary state, the values of the sublimation rates could be considered as constant. The end of the plateau corresponded to the Fig. 8 Product temperature profiles measured at the vial bottom. (Tshelf = 10 °C and Pc = 6 Pa). 4 mL glass vial. (Bogdani et al. [13])

30 20

Tshelf

10

T°C

0

0

100

200

300

400

500

600

-10 Temperature profile at vial bottom (Tb)

-20 -30 -40 -50 -60

Tb = cst Sublimation at quasi-stationary state t min

134

E. Bogdani et al.

Fig. 9 Sublimation rates for one vial measured with the microbalance Christ for different operating conditions (Tshelf and Pc). 4 mL glass vial. (Bogdani [13, 31])

end of the sublimation step with the starting of the product warming and an increase of the mean product temperature to reach progressively the shelf temperature. Then, these sublimation rates and product temperatures measured during the sublimation quasi-stationary state were used to calculate the overall heat transfer coefficient values, noted Kv (W.m-2.K-1) from the Eq. 1 below [30, 31]: _ K v = ðm:ΔH s Þ=ðAv ðT shelf- T b ÞÞ

ð1Þ

where the value of the vial right section area calculated from the outer vial diameter was equal to Av = 2.01 × 10-4m2 and the sublimation enthalpy value of the eutectic A was equal to ΔH s = 2352 kJ=kg. Then, as an example, for a shelf temperature equal to 10 °C and a total pressure of the chamber equal to 6 Pa, a value of Kv equal to 20.44 W.m-2.K-1 was obtained in agreement with previous values measured in our laboratories in these conditions (Tshelf; Pc, vial type) with the same vials [30, 31]. Following this methodology initially proposed by Nail [26], we measured the sublimation rates and the product temperature profiles during the quasi-stationary sublimation steps for four different shelf temperatures, that is to say at 0 °C; 10 °C; 20 °C and 30 °C and for five total gas pressures of the sublimation chamber respectively equal to 6 Pa, 15 Pa, 20 Pa, 30 Pa and 50 Pa. These values for a single 4 mL vial are reported in Fig. 9. From Fig. 9, we observed that the values of the sublimation rates increased with the shelf temperature. This tendency was only observed in the domain of the lower total gas pressures lying between 6 Pa and 30 Pa. On the other hand, for each shelf temperature, we observed that for total gas pressures higher than 30 Pa, the values of m_ reached a constant value corresponding to a plateau. This behaviour can be explained by the fact that the positive impact of the increase in the chamber total gas pressure on heat flux and, consequently, on the value of m_ , outweighed the negative (or inverse) impact resulting from the increase of Pc on the mass transfer driving force (P0 – Pc) that controls the mass transfer of solvent vapor through the dried layer. Then, the values of the mean product temperature (Tb) measured at quasi-stationary state were plotted in Fig. 10. It was observed that these values simultaneously increased with the shelf temperature and the total gas pressure in the lyophilization chamber. Then, the values of the overall heat transfer coefficient (Kv) could be calculated for our selected standard operating conditions for rather thermosensible drugs [1, 2, 30] from Eq. 1 and these values were reported on Fig. 11. From the same Fig. 11, for shelf temperatures in the range between 10 °C and 30 °C, we observed that for a constant value of the shelf temperature, the values of Kv increased with the total gas pressure in the sublimation chamber as already pointed out by Hottot et al. [30] in the same operating conditions. This increase results principally from the increase of the heat transfer flux by conduction at the vial bottom through the gas layer located between the curved vial bottom and the shelf due to the increase in the gas pressure of the thermal conductivity of gases at this very low gas pressure as reported by many other

Fig. 10 Mean product temperature (°C) as a function of operating parameters (Tshelf and Pc). 4 mL glass vial. (Bogdani et al. [13, 31])

Product temperature Tb at vial bottom (qC)

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water. . .

Fig. 11 Overall heat transfer coefficient values, noted Kv (W/m2.K), for different operating conditions (Tshelf, Pc). 4 mL glass vial. (Bogdani et al. [13, 31])

135

-22 -24 -26

Tshelf=0qC

-28

Tshelf=10qC

-30

Tshelf=20qC

-32

Tshelf=30qC

-34 -36 -38 0

10

20

30 Pc (Pa)

40

50

60

45

Kv Wm2 K

40 Tshelf = 0qC Tshelf = 10qC Tshelf = 20qC Tshelf = 30qC

35 30 25 20

0

10

20

30

40

50

60

P (Pa)

authors [1, 2, 19, 30]. Besides, in the same domain of the total gas pressure in the sublimation chamber, Fig. 11 shows a slight increase in Kv values as a function of temperature. On the contrary, for the lower shelf temperature equal to 0 °C, we observed that the values of Kv were equal to or larger than the values of Kv obtained with higher shelf temperatures in the range from 10 °C to 30 °C. This inverse tendency of Kv values increase, at lower temperatures for Tshelf < 0 °C, results probably from the fact that at these low or negative temperatures, the radiative component of the total heat flux reaching the sublimation front becomes more important than at higher positive temperatures [19, 30, 31].

8.2

Mass Transfer Resistance of the Dried Layer

The values of sublimation rates of the solvent mixture at eutectic A composition – namely 20% TBA + 80% water (by w/w) – reported in the preceding paragraph were determined for a solution without API (ibuprofen) or without excipients, and therefore, the sublimation process occurred without any microporous dried layer. The presence of this dried layer generally results in a decrease of the sublimation rates due to some mass transfer resistance encountered by the solvent vapor flow. Nevertheless, the values of sublimation rates that must be entered in the Design Space are the actual values that take into account this possible dried layer mass transfer resistance that could be significant with real solvent formulation containing the API and many different excipients. Thus, new sublimation experimental runs were realized by using an ibuprofen formulation (1 g of ibuprofen formulation per vial) with the organic co-solvent (TBA) + water mixture (eutectic A). These runs were carried out at constant total gas pressures for different shelf temperatures, namely 0 °C, 10 °C, 20 °C and 30 °C by using the same continuous weighing method with the microbalance already described. As an example, Fig. 12 shows typical weight loss

136

E. Bogdani et al.

Fig. 12 Variation of the total vial mass with the organic co-solvent formulation of ibuprofen. Tshelf = 30 °C and Pc = 6 Pa. 4 mL glass vial. (Bogdani et al. [13, 31])

0,0086 0,00855 0,0085 0,00845

m (kg)

0,0084 slope = -8.70E-08 kg/s 0,00835 0,0083 0,00825 0,0082 0,00815 0,0081 0

5000

Fig. 13 Shelf temperature and product temperature profiles at the vial bottom. (Tshelf = 30 °C and Pc = 6 Pa). 4 mL glass vial. (Bogdani et al. [13, 31])

10000 15000

20000

25000 t (sec)

30000

35000 40000

45000 50000

40 30 20

Tshelf

10

T°C

0 -10

0

50

100

150

200

250

300

350

400

-20 -30 -40

Tb = cst Sublimation at quasi-stationary state

-50 -60 t min

_ data obtained in the following operating conditions: Tshelf = 30 °C and Pc = 6 Pa that lead to a mean sublimation rate (m) equal to 8.70 × 10-8kg/s. This value was obviously lower than the value of m_ determined under the same operating conditions with a solution of pure eutectic A, namely m_ = 9.86 × 10-8kg/s which could indicate some presence of solvent vapor mass transfer resistance with this more realistic and commercial type formulation. During the same experiments with the ibuprofen system described in the previous paragraph, the product temperature at the vial bottom was also measured. The results presented in Fig. 13, show a product temperature plateau and a quasi-steady state around -23 °C, value 10 °C higher than the corresponding value measured with a solvent–solution mixture (at eutectic A composition) without ibuprofen. From the experimental values of the sublimation rates and the product temperature, Tb, the values of the solvent vapor mass transfer resistance, noted Rp , were calculated by using the Eq. (2) below:

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water. . .

137

Rp = AS ðP0- Pc Þ=m_

ð2Þ

where AS represents the sublimation surface assumed to be equal to the surface of the internal right cross section area of one vial, namely AS = 1.76 × 10-4m2 (for 4 mL glass vial), P0 represents the equilibrium vapor pressure of eutectic A solvent mixture at the sublimation front temperature, values calculated by the Clapeyron’s relationship determined experimentally in þ 27:909 (P0 in Pascal and Tb in Kelvin). Thus, for a shelf our laboratory by Bogdani et al. [7, 31], namely ln P0 = - 5745:9 Tb temperature equal to 30 °C and a total gas pressure equal to 6 Pa, we could estimate a mass transfer resistance value bp = 2:71 × 105 Pa:m2 :s=kg. These values of R bp depend essentially, for a given formulation, on the pores morphology of the R dried layer (size and shape) and thus, indirectly, of the morphology of the frozen solvent crystals formed during the freezing step and also, on the freezing protocol as observed by Hottot et al. [30] Due to the fact that all our experimental runs for setting up the Design Space of our system were obtained with the same freezing protocol, based on the laws of molecular diffusion in bp for a fixed Knudsen regime that prevailed during our experimental sublimation conditions, we assumed that this value of R shelf temperature remained unchanged with the variation of the total gas pressure in the sublimation chamber [13, 30]. Thus, adopting the hypothesis of sublimation at quasi-steady state—Cf. temperature profile shown in Fig. 13 — the values of the mean product temperature, noted Tb, could be calculated by solving the successive iterations with Matlab software the following implicit equation. ΔH s :AS ðexpð - 5745:9=T b þ 27:909Þ - Pc Þ = K v Av ðT shelf- T b Þ Rp

ð3Þ

The resulting values at constant Tshelf = 30 °C and for Pc values successively equal to 15 Pa, 20 Pa, 30 Pa and 50 Pa are plotted on Fig. 14 below. On the same figure, these values are compared with the experimental values obtained with the solution of eutectic A without any ibuprofen or excipients. We observed that all these Tb values obtained with the ibuprofen formulation were about 10 °C higher than the corresponding Tb values obtained with the formulation composed of the simple solvent (eutectic A) without ibuprofen and excipients.

8.3

Determination of Sublimation Rates of Real Ibuprofen Formulation

Next, by using Eq. (4), these Tb values were used to calculate the sublimation rates with the real formulation taking into account the limitations introduced by the solvent vapor flow through the dried layer:

-5 -10

Tb (°C)

-15 -20 -25 -30 -35 -40 0

20

40

60

Pc (Pa) Fig. 14 Product temperature (Tb) with ibuprofen organic-based formulation (X); without ibuprofen □); Tshelf = 30 °C. 4 mL glass vial. (Bogdani et al. [13, 31])

138

E. Bogdani et al.

Sublimation rate (kg/s)

1,88E-07 1,78E-07 1,68E-07 1,58E-07 1,48E-07 1,38E-07 1,28E-07 1,18E-07 1,08E-07 9,8E-08 8,8E-08 0

10

20

30

40

50

60

Pc (Pa)

Fig. 15 Sublimation rates of organic-based formulation with (x) and without (□) ibuprofen. Tshelf = 30 °C. 4 mL glass vial. (Bogdani et al. [13, 31])

Table 4 Mass transfer resistance values, Rp (Pa m2s/kg) for different shelf temperatures, Rp (Pa m2s/kg). 4 mL glass vial Tshelf (°C) 0 10 20 30

Pc (Pa) 20 6 6 6

Rp (Pa m2s/kg) 2.17 × 105 1.99 × 105 2.62 × 105 2.71 × 105

Bogdani et al. [13, 31]

m_ = K v Av ðT shelf- T b Þ=ΔH s

ð4Þ

These values are plotted in Fig. 15 and compared with the corresponding sublimation rates observed with the formulation without active principle (ibuprofen) or ingredients. The same figure shows that, at the constant temperature Tshelf = 30 °C, for the total gas pressure range investigated from 5 to 60 Pa, the m_ values obtained with the real formulation containing ibuprofen (API) were always below the experimental values observed without ibuprofen with a mean difference around 10%. These data certainly resulted from a significant and non-negligible mass transfer resistance of the solvent vapor flow through the dried layer located above the sublimation front even if in our sublimation conditions, the sublimation process was mostly controlled and governed by the heat transfer phenomena from the vial surroundings to the product sublimation front [19, 30]. Next, following the same methodology, we have experimentally determined the values of the mass transfer resistance of the solvent vapor flow for the other shelf temperatures, namely from 0 °C to 30 °C. The values for these experimental conditions are gathered on Table 4. In agreement with the mass transfer laws of gas diffusion encountered in our conditions of very total gas low pressure – Knudsen regime – we observed, for example, at constant total gas pressure equal to 6 Pa, that the Rp values increased significantly with the shelf temperatures, which represented the main operating parameter that determined directly the mean product temperatures and the sublimation front temperatures. Next, by using the method previously presented, these values of bp were used to calculate the values of the mean product temperature, noted Tb, by using Eq. (3), and, finally, to estimate the R sublimation rates, m_ with Eq. (4). These calculated values are plotted on Fig. 16, which gathers the basic data of the Design Space Graph that can be used for optimizing the sublimation steps of freeze-drying cycles of co-organic solvent (TBA) + water formulations of ibuprofen [13, 31].

8.4

Mean Product Temperature

The mean product temperature profiles all along the different freeze-drying steps including the sublimation steps are the key parameter that governed many quality factors of the final freeze-dried drug and also its final mechanical structure with respect to the possible collapse phenomena that are generally occurring with amorphous systems as previously outlined by many authors [1, 2]. Consequently, it is crucial to precisely control this important process parameter which results from very

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water. . .

139

Tb =T’ g =-7,5qC Tb =Tm =-8,2qC

Sublimation rate (kg/s)

2,80E-07

Forbidden zone

Tb = -10qC

2,40E-07 Safe zone 2,00E-07

Tb = -15qC Tshelf = 30qC

1,60E-07

Tshelf = 20qC 1,20E-07

Tb = -20qC

Tshelf = 10qC

Tb = -25qC Tshelf = 0qC

8,00E-08

4,00E-08 0

10

20

30

40

50

60

Pc (Pa) Fig. 16 Sublimation rates of the solvent of an ibuprofen organic co-solvent (TBA)-based formulation (eutectic A) vs. Tshelf and Pc. 4 mL glass vial. (Bogdani et al. [13, 31])

complex interdependent phenomena (nucleation, heat and mass transfer phenomena, etc.) related to the formulation and to the operating conditions during the freezing and the two drying steps (sublimation and desorption steps). This is why, for the engineers and the scientists that are developing, setting-up and running new freeze-drying cycles, it is very useful, for each set of independent operating conditions (Tshelf; Pc) to calculate and to plot these values on the Design Space Graph presented on Fig. 16 by using a representation at constant Tb values, called Tb isotherms. For drawing these isotherms, for a given shelf temperature, Tshelf we have proceeded as follows: firstly, we have chosen arbitrarily values of Tb and Pc and, next, we have calculated the corresponding values of the sublimation rates by using Eq. (5) below [13, 31]. m_ =

AS ðexpð - 5745:9=T b þ 27:909Þ - Pc Þ Rp

ð5Þ

It should be noted that each isotherm Tb = constant corresponds to a straight line so that only two calculated points are necessary to draw it in the Design Space graph. Thus, these product temperature isotherms, Tb = constant, have been added on Fig. 16 so that, at this stage of construction, this figure describes quite well the dynamics of the sublimation step as it gives the values of the sublimation rates and the average temperatures of the product for the two principal independent operating conditions (Tshelf and Pc) of sublimation that can be fixed independently by the freeze-dryer operator.

8.5

Limits of Critical Quality Attributes

Moreover, in the case of large industrial freeze-dryers, Nail [26] pointed out, that other limit conditions should be also considered for fixing the final optimal conditions of the freeze-drying cycles. These conditions depend not only on the intrinsic properties of the formulation itself or on the operating conditions of the two drying steps but also on the performance of the freeze-dryer itself like the surface and the refrigeration capacity of the condenser, the heating capacities between the shelf surface and the vial bottom in the product chamber (nature and flow rate of the heating fluid assuming sufficient heat transfer coefficient values, etc.) and on the dynamics of the solvent vapor flow inside the tube connecting the sublimation chamber and

140

E. Bogdani et al.

the condenser. For large freeze-drying equipments, according to this author [26], the main limit to be considered resulted generally from the flow of the solvent vapor between the two freeze-dryer chambers (product chamber and condenser chamber). When the sublimation rates became too high, the flow of solvent vapor reached the “choked” point – a kind of flooding point – which resulted in a sharp drop and a loss of control of the total gas pressure inside the sublimation chamber. Gieseler et al. [32] developed and implemented a new sensor and a new method based on Tunable Diode Laser Absorption Spectroscopy (TDLAS) to determine experimentally this “choked” point limit. Nevertheless, in the case of freeze-drying processes study at laboratory scale with our API systems, the limit due to the equipment cannot be generally considered. Indeed, in our experiments, the amount of solvent sublimated in our SMH 45 pilot freeze-dryer was very low – sublimation runs with 30 vials, so that we never reached these clogging or “choked” points for the solvent vapor flow through the connecting tube of the two chambers inside the freeze-dryer. To finalize and to complete the Design Space graph presented on Fig. 16, it was also necessary to introduce other limits that will ensure the main quality attributes of the final freeze-dried drug, generally shaped as a porous consolidated cake or as a fine powder. These main quality factors concern generally the dried cake appearance, its color, the absence of cracks, the facility of rehydration, its pharmaceutical activity, the storage duration, etc. Nevertheless, the more important limit to be considered in setting –up a drug freeze-drying cycle is the mean product temperature. For this purpose, the temperature of the shelves should be controlled all along the successive primary step (sublimation) and also along the secondary drying step (desorption) to avoid the collapse phenomena. This collapse phenomenon is characterized by the loss of mechanical resistance of the micro-structure of the partially dried solid layer – the morphology of which was fixed during the freezing step – due to the viscous flow induced by a temperature increase [13, 16]. After an irreversible collapse phenomenon, the freeze-dried drug was no more pharmaceutically acceptable due to its poor appearance, to its higher water content or its large reconstitution times. To avoid this very damaging phenomenon, for an amorphous system, the cryo-concentrated phase temperature should be maintained during the sublimation step below the vitreous transition temperature of the maximally cryo-concentrated phase, noted Tg’, and, during the desorption step, below the vitreous transition temperature, noted Tg. For a predominantly crystalline system, the upper temperature limit during the sublimation is the eutectic temperature, noted Tm, that corresponds usually to a sharp melting point. Overpassing this melting point during the sublimation results in product puffing inside the vial and globally in a loss of pharmaceutical acceptability as underlined by Pikal et al. [1], Andrieu et al. [2] and by Pikal et al. [16] This collapse phenomenon can be quite easily detected by laboratory DSC or freeze-drying cryo-microscopy experiments with small samples of the real formulation. It was observed by many authors that the collapse temperature, Tcoll, for the vitreous systems frequently encountered with formulations of thermosensible drugs are quite close to the vitreous transition temperature, Tg, given by the phase diagram of the system [1, 2, 13, 18]. Then, it was very useful to plot on the Design Space Graph the two corresponding isotherms at Tm = constant and Tg = constant, as shown in the final Fig. 16. Thus, this figure determines the so-called security zone within which any change in the values of the main independent sublimation parameters (Tshelf and Pc) will not affect the most important quality factors required by the final freeze-dried product. Any process conditions in the Design Space would be acceptable to obtain a final freeze-dried product with acceptable quality attributes even if, of course, it is desirable to operate near the apex of this more or less large space with respect to the sublimation rates because this maximum corresponds to the most efficient process conditions (low operating costs).

8.6

Reconstitution Properties

The capability of the freeze-dried cake or the freeze-dried powder to rapidly dissolve and reconstitute after addition of an adequate pharmaceutical solvent is not a negligible factor to take in account, particularly in the case of injectable drugs. This property depends on numerous parameters that are fixed during the freeze-drying as the freeze-dried cake morphology (pores shape and surface area homogeneity of the dried matrix), or the degree of collapse. It was observed that the freeze-drying with TBA formulations produced cakes with large surface area and high permeabilities which generally could be reconstituted very rapidly upon addition of most standard solvents [1, 2, 24, 31].

Freeze-Drying of Thermosensible Pharmaceuticals with Organic Co-solvent + Water. . .

9

141

Conclusions

Non-aqueous co-solvent systems are more and more studied and practically used in the freeze-drying of drugs and pharmaceutical products. The reasons for this increasing interest in potential industrial applications with commercial products include the increase of solubility (hydrophobic products); the high increase of sublimation rates and the consequent decrease of sublimation times and, consequently, the decrease of operating costs; the increase of pre-dried (solution) and of the freezedried product stability; the decrease of freeze-dried cake reconstitution times. Nevertheless, the practical implementation of these organic co-solvent formulations at commercial stages needs to evaluate important issues as the precise control of residual co-solvent contents; the rigorous control of their toxicity levels; the performances of the freeze-dryer (condenser temperature); the pertinent choice of sensors and control devices (security constraints); the safe handling and storage of flammable and/or explosive solvents; the evaluation of overall cost benefits by using these organic solvents. TBA + water formulations have been certainly the most largely evaluated and used systems in the manufacture of commercial pharmaceutical products. Their main advantages rely on the easy freezing in standard freeze-dryers due to TBA high freezing temperature; the high equilibrium solid–vapour pressures and the low sublimation enthalpies values; the low residual solvent contents and the low toxicity of residual TBA. When setting up the freeze-drying cycle, both formulation parameters and process parameters – including important freezing parameters that were more often neglected by the R & D engineers – necessitates an overall optimization to control the main quality factors of the final freeze-dried product and to lower as much as possible the freeze-drying operating costs. Finally, with regard to freeze-drying cycle optimization, it is well established that both formulation parameters and operating process parameters must be taken into account together; for example, through Design Space methodology. In the special case of organic co-solvent systems such as TBA+water systems, there exist a strong influence of freezing conditions on the morphology of the frozen system and, consequently, on the final freeze-dried cake. Thus, it is important to take into account most of the freezing parameters that control the nucleation temperature of the undercooled solution; these phenomena being more often neglected with regular and classical aqueous-based systems.

References 1. Pikal MJ. Freeze drying, Encyclopedia of pharmaceutical technology. New York: Marcel Dekker; 2002. p. 1299–326. 2. Andrieu J, Vessot S. Optimisation of some physical quality factors during freeze-drying of pharmaceuticals in vial configuration. Modern Dry Technol. 3. Chapter 3. Quality in drying. Editor Wiley-VCH. 3. Teagarden DL, Baker DS. Practical aspects of lyophilization using non aqueous co-solvent systems. Eur J Pharm Sci. 2001;15:115–33. 4. Vessot S, Andrieu J. Freeze-drying of drugs with tert-butanol (TBA) + water systems: characteristics, advantages, drawbacks. A review. Dry Technol. Special issue on biologicals and drugs drying. 2012;30:377–85. 5. Rey L, May JC. Freeze drying/lyophilization of pharmaceutical and biological products. 2nd ed. New York: Marcel Dekker; 2004. p. 239–78. 6. Daoussi R, Vessot S, Andrieu J, Monnier O. Sublimation kinetics and sublimation end-point times during freeze-drying of pharmaceutical active principle with organic co-solvent formulations. Chem Eng Res Des. 2009;87:899–907. 7. Bogdani E, Daoussi R, Vessot S, Jose J, Andrieu J. Implementation and validation of the thermogravimetric method for the determination of equilibrium vapour pressure and sublimation enthalpies of frozen organic formulations used in drug freeze-drying process. Chem Eng Res Des. 2011;89:2006–12. 8. Nakagawa K, Hottot A, Vessot S, Andrieu J. Influence of controlled nucleation by ultrasounds on ice morphology of frozen formulations for pharmaceutical proteins freeze-drying. Chem Eng Process. 2006;45:783–91. 9. Hottot A, Vessot S, Andrieu J. Freeze-drying of pharmaceuticals in vials: influence of freezing protocol and sample configuration on ice morphology and freeze-dried cake texture. Chem Eng Process. 2007a;46:666–74. 10. Cui JC, Li CL, Deng YJ, Wang YL, Wang W. Freeze-drying of liposomes using tertiary butyl alcohol/water cosolvent systems. Int J Pharm. 2006;312:131–6. 11. Zhang Y, Deng Y, Wang X, Xu J, Li Z. Conformational and bioactivity analysis of insulin: freeze-drying TBA/water co-solvent system in the presence of surfactant and sugar. Int J Pharm. 2009;371:71–81. 12. Elgindy N, Elkhodairy K, Molokhia A, Elzoghby A. Lyophilization monophase solution technique for improvement of the physicochemical properties of an anticancer drug, flutamide. Eur J Pharm Biopharm. 2010;74:397–405. 13. Bogdani E, Vessot S, Andrieu J. Optimisation of freeze-drying cycles of pharmaceuticals organic co-solvent-based formulations using the Design Space methodology. Dry Technol. 2012;30:1297–306. 14. Seager H, Taskis CB, Syrop M, Lee TJ. Structure of products prepared by freeze-drying solutions containing organic solvents. J Parental Sci Technol. 1985;39(4):161–79. 15. Daoussi R, Bogdani E, Vessot S, Andrieu J, Monnier O. Freeze-drying of an active principle ingredient by using co-solvent formulations. Influence of freezing conditions and formulation on solvent crystals morphology, on thermodynamical data and on sublimation kinetics. Dry Technol. 2011;29:1858–7.

142

E. Bogdani et al.

16. Pikal MJ, Rambhatla S, Ramot R. The impact of the freezing stage in lyophilisation: effects of ice nucleation temperature on process design and product quality. Am Pharm Rev. 2002;5:48–53. 17. Kasper JC, Friess W. The freezing step in lyophilisation: physicochemical fundamentals, freezing methods and consequences on process performance and quality attributes of biopharmaceuticals. Eur J Pharm Biopharm. 2011;78:248–63. 18. Kasraian K, DeLuca PP. Thermal analysis of the tertiary butyl alcohol-water system and its implications on freeze-drying. Pharm Res. 1995;12 (4):484–90. 19. Hottot A, Peczalski R, Vessot S, Andrieu J. Freeze-drying of pharmaceutical in vials. Modelling of freezing and sublimation steps. Dry Technol. 2006;60(2):95–103. 20. Hottot A, Vessot S, Andrieu J. Sublimation kinetics during freeze-drying of pharmaceutical protein formulation. Dry Technol. 2007b;25(5): 753–8. 21. Baldi G, Gasco M, Pattarino F. Statistical procedures for optimizing the freeze-drying of a model drug in tret-butanol mixtures. Eur J Pharm Biopharm. 1994;40(3):138–41. 22. Chouvenc P, Vessot S, Andrieu J. Experimental study of the impact of annealing on ice structure and freeze-drying parameters during freezedrying of pharmaceuticals model formulations. PDA J Pharm Sci Technol. 2006;60(2):95–103. 23. Wittaya-areekul S, Nail SL. Freeze-drying of ter-butanol/water systems: effects of formulation and process variables on residual solvents. J Pharm Sci. 1998;26(1–4):33–43. 24. Daoussi R, Vessot S, Andrieu J, Monnier O. Lyophilisation d’un produit pharmaceutique en milieu non aqueux. Impact des conditions de congélation et de formulation sur les taux résiduels de solvant et d’humidité et sur le temps de séchage. Comptes-rendus du Congrès SFGP, Saint-Etienne, (2007), CDROM (6 pages). 25. Kunz C, Gieseler H. Factors influencing the retention of organic solvents in products freeze-dried from co-solvent systems. J Pharm Sci. 2018;107(8):2005–12. 26. Mochus LN, Paul TW, Pease NA, Harper NJ, Basu PK, Oslos EA, Sacha GA, Kuu WY, Hardwick LM, Karty JJ, Pikal MJ, Hee E, Khan MA, Nail SL. Quality by design in formulation and process development for a freeze-dried, small molecule parenteral product: a case study. Pharm Dev Technol. 2011;16(6):549–76. 27. Chang BS, Fischer NL. Development of an efficient single-step freeze-drying cycle for protein formulations. Pharm Res. 1995;12:831–7. 28. Fissore D, Pisano R, Barrési AA. Advanced approach to build the design space for the primary drying of a pharmaceutical process. J Pharm Sci. 2011;100(11):4922–33. 29. Giordano A, Barrési AA, Fissore D. On the use of mathematical models to build the design space for the primary drying phase of a lyophilization process. J Pharm Sci. 2011;100(1):311–24. 30. Hottot A, Vessot S, Andrieu J. Determination of mass and heat transfer parameters during freeze-drying cycles of pharmaceutical products. PDA J Pharm Sci Technol. 2005;52(2):138–53. 31. Bogdani E. Etude expérimentale et optimisation du procédé de lyophilisation de l’ibuproféne en milieu organique, Thèse, Université C Bernard Lyon 1, 2011. 32. Gieseler H, Kessler WJ, Fonson M, Davis SJ, Mulhall PA, Bons V, Debo DJ, Pikal MJ. Evaluation of tunable diode laser absorption spectroscopy for in-process water vapor flux measurements during freeze-drying. J Pharm Sci. 2007;96:1776–93.

Primary Container Closure System Selection for Lyophilized Drug Products Robert Ovadia, Phillippe Lam, Holger Roehl, Renaud Janssen, and Roger Asselta

Abstract

For lyophilization applications, the selection of the primary container system (i.e. vial, stopper, seal) must be given careful consideration in order to avoid surprises in the later stages of development and potentially costly issues throughout the product’s life cycle. The system must provide a sterile barrier, enable robust fill/finish and lyophilization processes, and preserve the headspace conditions. It may not be possible to arrive at the perfect system and compromises need to be made. In this chapter, we hope to give the reader the tools and understanding needed to make the best decision for their application. Keywords

Parenteral packaging · Vial · Stopper · Seal · Container closure integrity

1

Introduction

Philippe Lam

For lyophilization applications, the selection of the primary container system (i.e., vial, stopper, seal) must be given careful consideration in order to avoid surprises in the later stages of development and potentially costly issues throughout the product’s life cycle. For example, Dr. Pikal discussed how vial type and geometry affect drying [1]. His studies revealed that while there are obvious physical differences between molded and tubing glass vials, it is the exterior shape and not the

R. Ovadia (✉) Gilead Sciences, Inc., Foster City, CA, USA e-mail: [email protected] P. Lam Amgen Inc., Cambridge, MA, USA e-mail: [email protected] H. Roehl F. Hoffmann-La Roche AG, Basel, Switzerland e-mail: [email protected] R. Janssen Datwyler Pharma Packaging, Alken, Belgium R. Asselta Auxilium Packaging Advisors LLC, South Harrison Township, NJ, USA e-mail: [email protected] # America Association of Pharmaceutical Scientists 2023 F. Jameel (ed.), Principles and Practices of Lyophilization in Product Development and Manufacturing, AAPS Advances in the Pharmaceutical Sciences Series 59, https://doi.org/10.1007/978-3-031-12634-5_8

143

144

R. Ovadia et al.

thickness of the glass at the bottom of the vial that has the greatest impact on heat transfer and, therefore, on primary drying. Consequently, even vials of the same type and size from different manufacturers may exhibit substantially different heat transfer characteristics and should not be used interchangeably without first determining their contribution to the process. The stopper is the other component that can also have a substantial impact on lyophilized product quality over its shelf life. Specifically, Dr. Pikal presented data showing the transient residual moisture profile for various lyophilized products and how the stopper elastomeric material properties and stopper processing conditions dictate the characteristic of environmental water vapor permeation and initial water release (from the stopper), respectively [2]. Selecting the appropriate stopper and allowing sufficient drying time post-steam sterilization of the processed stoppers will minimize risks of excessive moisture uptake by the lyophilizate during storage. These are a few aspects of the primary container system that have long been recognized to be significant to lyophilized drugs manufacturing. There are other factors that are less obvious and sometimes overlooked but are also of prime importance, as discussed in this chapter. The role of the primary container, other than to function as a container, is to provide a sterile barrier and to maintain controlled internal conditions so as to preserve the product it contains. For liquid products, the time between stoppering and the application of the seal (typically an aluminum crimp cap with or a without removable protective top cover) is mere seconds as the vials are moved along the line from filling to stoppering to capping. For lyophilized products, the primary container system faces additional challenges. For instance, it may be many hours between stoppering at the end to the drying cycle (where the stoppers are fully pushed into the vials by the hydraulic ram of the lyophilizer) to when the crimp caps are finally applied. As the vials await to be sequentially unloaded and transported to the capping station, container closure needs to be maintained to ensure sterility and to preserve the low humidity/reduced pressure condition within the vials. It has been argued that for product vials that are stoppered under reduced pressures, the inside-outside pressure differential results in a net force that pushes the stopper inwards and helps keep it in place until capping. Indeed, the force exerted from atmospheric pressure can be substantial. For example, at a pressure differential of 500 mbar, the vial with 20 mm finish and an opening ID of 12.6 mm, the stopper will be pushed in with about 6 N of force; for the 13 mm finish vial with an opening ID of 7 mm, the force will be about 2 N. However, it must be pointed out that this state can only be attained if the vial-stopper system is already integral immediately after full stopper insertion at the end of the drying cycle and remains integral thereafter. A vialstopper system that is “leaky” will not result in a pressure differential; as the lyophilizer chamber is equilibrated to ambient conditions, so will the vial. There are various reasons for which leaks can occur. The components may simply fit poorly together and not form a proper seal in the first place. The top of the stopper may adhere to the lyophilizer shelf above during the stoppering phase and be pulled out of the vial when the hydraulic ram is retracted. The most spectacular seal failure is that which has been anecdotally reported by many experienced practitioners as “stopper pop up.” Here, the stopper, after having been fully seated, spontaneously moves back out of the vial opening. This is caused by a combination of non-optimal stopper and vial geometry and stopper processing conditions where excessive silicone oil has been applied. Specialized equipment, “raised stopper detectors,” have been developed and installed ahead of the capper to cull out vials with unseated stoppers, which speaks to the prevalence of this type of defect. The primary container system needs to perform other functions besides providing a robust barrier. The lyophilization stopper features vent hole(s) to allow for egress of water vapor during drying. To function properly, the stopper must remain securely in the “up” or “lyophilization” position from the time the stopper is placed on the vial post-filling until it is finally pushed into the fully seated position on the vial by the hydraulic ram. A stopper that fits too loosely may be dislodged during transport and loading operations or fall too far into the vial opening, occluding the path for water vapor flow. Conversely, a too tight-fitting stopper may require high insertion force during stoppering, and, depending on the number of vials per shelf, the necessary total stoppering force may exceed the lyophilizer hydraulic ram capabilities. The stopper elastomeric material itself should provide good sealing (to the vial) and barrier properties as mentioned earlier. Furthermore, the final drug product vial (vial, stopper, crimp cap) should be evaluated for compatibility with closed system transfer devices (CSTD), as their usage have been much more prevalent in the field in recent years [3]. Many CSTDs employ a very large (diameter > 2 mm) “spike” to puncture the stopper, potentially causing excessive rubber fragmentation which can introduce particles in the reconstituted drug product solution. As alluded to previously, the vial geometry is critical not just for consistency of heat transfer during the drying operation but also for proper interaction with the stopper to obtain a robust container closure. In addition, the shape of the vial bottom impacts machinability. Most modern lyophilizers are equipped with automatic loading systems. Issues can arise during lyophilizer loading and unloading operations as the movement of a vial pack is sensitive to slight misalignment between the shelf and loading system transfer ramp. The shape of the heel on the vial has a major impact on how smoothly vials move over

Primary Container Closure System Selection for Lyophilized Drug Products

145

this transition. Additionally, vials with very flat bottoms may exhibit higher friction when moved across the shelf surface, especially if there is condensation present (due to a cold shelf surface). We pointed out some of the aspects of the primary container that should be considered while developing a lyophilized drug product process. The following sections in this chapter give more detailed discussion on these topics and offer solutions to some of the issues raised. Ultimately, it may not be possible to arrive at the perfect system, and compromises need to be made. To that end, we hope to give the reader the tools and understanding needed to make the best decision for their application.

2

Primary Containers

Holger Roehl

There are multiple options to choose from for containers used for lyophilization applications. This not only includes vials, syringes, or dual-chamber cartridges but also the container material (glass or polymer/plastics); all these may be had with or without inner-surface coatings. For glass vials, there is even another dimension of complexity, as one can either use tubular or molded containers. There are pros and cons for all the different options. For a robust process with minimal vial breakage and machinability issues and with optimal product properties, it is crucial to identify the right components. This section intends to cover the most important topics that influence the performance of the containers for lyophilization.

2.1

Overall Considerations

In general, a container for pharmaceuticals must protect the content in a way that sterility and drug product quality are maintained and that interactions between the drug and the container are minimized (drug container interaction, ion leaching, etc.) [4]. Tubular vials are normally made from Type I borosilicate glass. The other common types, Type II or III, refer to glasses that are treated (e.g., ammonium sulphate treatment) in a way to reduce the sodium content in the region near the surface. Type I glasses consist mainly of SiO2 (70–80%) and B2O3 (7–13%) which form the backbone of the glass (network formers). The addition of network modifiers like alkaline metals (Na2O and/or K2O, up to 8%), earth alkaline metals (BaO and/or CaO, 0–5%), and intermediates like Al2O3 (2–7%) results in the desired physical (e.g., lower forming temperatures, strength, etc. . . .) and chemical properties [4]. Other optional components such as Fe2O3 (approx. 1%) or TiO2 (approx. 5%) are added to obtain amber glass which exhibits additional protection from UV light. Plastic/polymer vials are injection molded and made from either COP (Cyclo Olefin Polymer) or COC (Cyclo Olefin Copolymer) [4]. Due to the material composition, inorganic metal ion leaching issues are eliminated.

2.2

Heat Transfer (Glass Vials)

In order to understand the influencing factors on the performance of the containers, it is crucial to capture the different stages of lyophilization [4–6]: • Freezing • Sublimation (primary drying) • Desorption (secondary drying) All three processes require good heat transfer either to or from the filled drug product. This is the pre-requisite to achieve an optimized lyophilization cycle as the current temperature of the drug product has the highest impact on drying time and final product quality [7]. Energy needs to be moved from the product to the heat transfer liquid in the shelves during the freezing stage; the direction of heat transfer is reversed during drying as the frozen product must be heated for the sublimation process

146

R. Ovadia et al.

Fig. 1 Heat transfer mechanisms during primary drying

[8]. The vial heat transfer coefficient during drying, Kv, is an indicator for the rate of energy transfer per area, temperature differential, and time between the shelf and the container. Kv is a useful parameter to consider when selecting containers. There are three main mechanisms for the heat transfer during primary drying, see Fig. 1 [5]: • Heat convection (through gap between shelf and vial bottom) • Heat conduction (from the shelf to the containers at contact areas) • Heat radiation

The container exhibits a specific resistance to heat transfer depending upon the shelf temperature and the chamber pressure, thereby directly impacting the duration of the lyophilization process and the quality of the final product. Studies revealed that the gas-filled gap between the vial bottom and the shelf surface contributes up to 90% to the heat resistance and, therefore, is the main culprit for the poor heat transfer [7]. Thus, the vial bottom geometry has the highest impact (in conjunction with chamber pressure) to heat transfer while the thickness of the vial wall is only of minor importance [7]. This observation is confirmed by various other studies revealing that the vial bottom curvature is limiting the heat transfer as it defines the direct surface contact between the vial and the shelf [9]. The considerations made above lead to the conclusion that the vial geometry (bottom) must be flat, thin, and uniform to achieve the highest heat transfer rates. Figure 2 displays lateral cuts of different molded vials in comparison to a standard tubular glass vial [4, 8]. The images provided above therefore would indicate a superior behavior of the tubular glass over molded vials with respect to heat transfer as molded vials traditionally have a higher bottom curvature and wall thickness (due to the manufacturing process). As a direct consequence, a much better lyophilization performance of tubular vials is expected. However, improvements in the manufacturing of molded vials (e.g., EasyLyo® from SGD) result in similar Kv values than state-ofthe-art TopLyo® vials from Schott (only 4% difference at 100 mbar chamber pressure) [10]. In addition, the absolute values and the dependence of Kv on the chamber pressure were in the same range for both the molded and the tubular vials over the pressure range of 50–200 mbar. These observations were confirmed by further studies [11]. Therefore, improved manufacturing processes for molded vials result in bottom geometries comparable to tubular glass vials and, as a consequence, in similar heat transfer properties of the respective containers. At 200 mbar, the drug product even dried 15% faster with molded glass vials. However, the differences in the structural properties of the vials also need to be taken into consideration. A comparison of the manufacturing process of the molded vials (PB = press-blow vs. BB = blow-blow) reveals that the manufacturing and the different glass composition of the respective vials has a direct impact on the heat transfer: Kv values for PB vials are slightly higher than for BB vials with identical design [8]. The BB process results in a more heterogeneous glass distribution and hence in a different cooling process. PB are supposed to have marginally better heat transfer characteristics. With regard to the glass composition, studies performed on clear and amber glass vials come to an inconsistent conclusion [8]. Cannon et al. reported significantly differences in sublimation rates between clear and amber glass vials of the same size,

Primary Container Closure System Selection for Lyophilized Drug Products

147

Fig. 2 Example lateral cuts of different vial systems (BB blowblow, PB press-blow, ST standard tubular vials)

while Hibler et al. reported identical Kv values for a different pair of geometrically identical clear and amber vials [10, 12]. Additional studies revealed that the addition of coloring agents like Ti or Fe do not affect Kv values of molded amber glass vials [10]. Such an effect could be anticipated as the amber glass exhibits a faster cooling and solidification during manufacturing of the vials.

2.3

Glass Versus Polymer

2.3.1 Heat Transfer Polymer or plastic vials are normally chosen for their superior breakage properties [4]. This is why they are commonly used for larger volumes or drugs with increased hazard potential [7]. From a geometrical perspective, tubing vials have superior heat transfer properties because of better direct contact and enhanced gas conduction (reduced bottom concavity). Thus, tubing vials dominate freeze-drying of small-volume parenterals, as discussed earlier [7]. From a thermal conductivity perspective, glass is better than polymer because of a higher thermal conductivity coefficient, i.e., it conducts heat more efficiently [4, 6]: • Glass: approx. 1.05 W∙m-1∙K-1 • Polymer: approx. 0.2 W∙m-1∙K-1 Hibler et al. report a 30% higher heat transfer for the glass TopLyo® vials from Schott at a pressure chamber of 400 mTorr, but heat transfer was reported as very homogeneous for the polymeric TopPac® vials [7, 10]. Other studies revealed a 9% longer primary drying time at 68 and 100 mTorr for polymer vials in comparison to a variation of only 3–4% in glass vials [11].

148

R. Ovadia et al.

In general, according to Schott, COC is outperforming other polymers like polypropylene (PP), polyethylene terephthalate (PET), or polyvinyl chloride (PVC). The TopLyo® glass vials, which are referred to as the “gold standard” for lyophilization, exhibit only an 8% higher Kv value than the TopPac® vials (at a 100 mTorr chamber pressure), but are still lower than the molded glass SGD EasyLyo® vials [7]. In addition, the water vapor permeation and the diffusion of oxygen and/or carbon dioxide through the polymer container walls during storage over the shelf life need to be taken into consideration.

2.3.2

Breakage

2.3.2.1 Lyophilization Though vial breakage during lyophilization is rare, the impact on the production process is very high: not only is product lost, but downtimes of the production lines are high because of the necessity of line stoppage and cleaning [13]. Increased amounts of vial breakage is in general associated with the following: • • • • • •

Higher concentrations of excipients in the drug product [14] Higher fill volumes (a max. of 35% of the full vial fill capacity is therefore recommended) [5, 14] Fast freeze-thaw cycles showed more vial breakage during thawing [14] While a slow freeze-thaw at 0.1 °C/min accounted for more vial breakage during freezing [14] Vial tend to break at the endpoint of complete freezing [15] Stresses in the glass are caused by an internal force from product expansion during freezing [15, 16]

Though the bottom wall thickness has only a minimal impact on the heat transfer, high variability of the wall thickness can lead to increased rates of breakage [5, 7]. Both the bottom inside and outside radius have a significant impact on vial breakage [5]. The greater the bottom cavity, the sharper the inside angle and the higher the breakage probability. Freezing material will expand in that region and will exert a pressure that could lead to the so-called “lens-out” effect. The shape of the outside surface is important for the contact area of the vial bottom with the lyo shelf. If the angle is too sharp, it could result in the same vial breakage behavior as described for the inside radius. 2.3.2.2 Storage at Low Temperatures Freeze-thaw temperature conditions have great impact on vial integrity [17]. Vial breakage of protein products occurred at 70 °C freeze-thaw, but was found to be negligible during -30 °C freeze-thaw. Major root cause for vial breakage during freezing was identified as the thermal contraction of protein formulation when cooled to below -30 °C, causing inward deformation of glass adhering to the frozen plug and subsequent rapid movement of glass when the sidewall separated from frozen plug (“plugging off”). Conversely, thermal expansion during thawing resulted in positive strains and explained the occasional breakage during thawing. This is why it is recommended to freeze down to -30 °C only to avoid vial breakage during freeze-thaw. If temperatures of -70 °C storage are unavoidable, there will be a risk of glass vial breakage, as the thermal contraction is an inherent property of the frozen formulations. Maurer et al. provide an in-depth overview of glass vial properties and their respective breakage behavior and the theory of glass breakage [5].

2.3.3 Delamination and Chemical Durability These topics are, under normal circumstances, not in scope for lyophilized products as the time of interaction between the liquid and the container surface is reduced to a minimum. But it needs to be mentioned that delamination issues are eliminated when using polymer vials which also exhibit in general a higher durability against extreme pH values [4]. 2.3.4 Other Important Practical Consideration It should be noted that polymeric vials cannot be depyrogenated and, therefore, are not compatible with conventional fill lines that employ a wash-heat depyrogenation process. Polymeric containers are typically supplied in a pre-sterilized format, requiring aseptic procedures for introduction onto the fill line.

Primary Container Closure System Selection for Lyophilized Drug Products

2.4

149

Extractables and Leachables

The lyophilization process itself is not eliminating the interactions of the drug product with the container [4]. After filling, before the freeze-drying, ionic species in the glass like sodium, potassium, calcium, silicone, or aluminum could migrate into the solution. But the absolute amounts are fairly low due to the short time of interaction, which is in the range of hours up to (in total) a very few days. This is why the risk can be considered as very low and the levels of leacheates from the glass are unlikely to exceed the permissible limits of metal ions as defined in the ICH Q3D elemental impurities guideline [18]. But even at a frozen stage, volatile components must be taken into consideration [4]. They originate from polymer components (rubber stopper or container) during the reduced pressure cycles of lyophilization and during the storage of the drug product at low pressure conditions inside the vial to facilitate reconstitution. These volatile components can adsorb on the product and can lead to interactions with the product after reconstitution. This is why the FDA clearly indicates that extraction studies are needed to determine if “a material of construction used in the manufacture of a packaging component is safe for its intended use.”

2.5

Fogging

Under specific circumstances, the drug product liquid can creep up the sidewall of the container (Marangoni effect). After lyophilization, this thin dried product film is visible as a haze and often considered a cosmetic defect, though, studies revealed that “fogging” is typically not compromising container closure integrity (CCI) [12, 19]. The haze exhibits irregular structures like streaks, branches, spots, or even uniform thin layers (see Fig. 3). The occurrence of these structures is inconsistent between vials of the same batch [19]. Despite the fact that fogging is classified as cosmetic defect, impacted vials are often discarded as severe cases often reach the shoulder or even vial neck region. Even if experiments show no impact on CCI, this is rated as unacceptable. Furthermore, the presence of fogging may present a significant challenge for automated visual inspection equipment and result in an excessive level of rejects. The glass vial surface properties, in combination with the pre-treatments of the container and the various processing parameters during lyophilization, were identified to have a significant impact on fogging propensity [12, 19]. However, a prediction of the fogging behavior of a single container currently proves to be an unsolvable challenge. This is why containers with a hydrophobic surface (siliconized vials, vials with a hydrophobic inner surface coating or polymer vials) are considered as the only reliable solution as they prevent the creeping of the drug product solution [12, 19].

Fig. 3 Fogging behavior of different vial types

150

R. Ovadia et al.

Fig. 4 (a) Comparison of hydrophilic (top) and hydrophobic (down) surfaces for standard type I glass and TopLyo® and (b) difference in cake appearance

2.6

Coated Vials

Coatings on the inner vial surface fall into two different categories: silicone oil or hydrophobic Si-O-C-H layers (see Fig. 4) that are applied by vapor deposition processes as PI(E)CVD (Plasma Induced (or Enhanced) Chemical Vapor Deposition). Examples include the TopLyo® glass vials from Schott and polymer vials from SiO2 Material Sciences. The container itself is used as the reaction chamber for the layer deposition. This ensures that only the inner surface is coated. TopLyo® vials can be depyrogenated at approx. 300 °C as the coating withstands these temperatures without losing its functionality, and therefore, these vials are compatible with traditional fill lines. TopLyo® vials claim to exhibit [6]: • • • •

Better cake appearance (see Fig. 4) Prevention of cake collapse Less residual volume Optimized vial bottom geometry

3

Elastomeric Closures

Renaud Janssen

By permission of Springer adapted from Springer, Janssen [6].

3.1

Closure Geometry

Lyophilization closures have to be compatible with the freeze-drying process. The first step in the process is the filling of vials with drug solution, followed by partial stoppering of the vials, meaning that the lyophilization closures are only partially inserted in the vial neck (Fig. 5). A major part of the stopper plug is still protruding above the vial neck opening. The partially stoppered vials are then transported into a freeze-drying chamber, where the freeze-drying cycle takes place. Only at the end of the cycle the vials are fully stoppered. In most cases, the stoppering takes place with a vacuum still being present in the chamber. What comes out of the freeze-dryer are vials under a certain vacuum level containing the drug product in the form of a freeze-dried cake and closed with stoppers that are not yet secured with a crimp cap.

Primary Container Closure System Selection for Lyophilized Drug Products

151

Fig. 5 Half-stoppered vial with drug solution (right) and totally stoppered with freeze-drying cake (left). In front of the vials: freeze-drying closures with their vent openings facing forward

During the primary drying phase, when the water in the vial is converted from the frozen state (ice) into the gaseous state (water vapor), the gases are evacuated from the vial via the vent opening of the lyophilization stopper. Since the pressure is so low, the volumes of gas exiting the vial are appreciable, meaning that for a good functioning the vent opening has to be of a certain size. Studies have shown however that the size of the vent opening is not rate determining for the common freezedrying process [1, 20]. The size of the vent opening is far less important than the resistance formed by the already dried product, through which additional gases have to make their way to be evacuated from the vial. As a result of the freeze-drying process and the transport stages of the vials from the filling station to the freeze-dryer in half-stoppered condition and from the freeze-dryer to the capping station in stoppered but still uncapped condition, the design and the dimensioning of lyophilization stoppers have a number of particularities that are not seen with stoppers that are used for a liquid or a dry powder fill.

3.1.1 Flange Thickness and Flange Design The flange of the lyophilization stopper is that part of the stopper that is above the rim of the vial mouth, when the stopper has been firmly brought into its final seating position. The closure manufacturer has to keep flange thickness well under control for the reason of vial capping. At the time of package development, an aluminum/plastic cap has been chosen where the inner height of the skirt of the cap is capable of gripping the flange height of the stopper plus the height of the vial neck collar, while at the same time being well folded under the collar of the vial to firmly hold the stopper and assure container closure integrity. If stopper flange thickness is excessively high, crimping the cap under the collar could become difficult and container closure integrity may be impacted, while too low a flange thickness could result in an excess of folded aluminum, which is considered a cosmetically imperfect crimp. Of special importance for a lyophilization stopper are its top flange marking. The functionality of flange markings (see Fig. 6) for all elastomeric closures can be described in terms of prevention of stopper clumping during storage, steam sterilization, and machining. This functionality is also valid for lyophilization closures; however, there is an additional functionality that is related to the final stoppering of the vials inside the lyophilization chamber at the end of the freeze-drying cycle. As explained above, the stoppers are pressed down by the shelf that is located above them. During this stage, a significant pressure is exerted on the stopper at the time it contacts the vial neck. The pressure is transferred to the stopper flange that is in contact with the underside of the shelf, in other words, the stoppers are firmly pressed with the top of their flanges against a stainless steel plate that for reasons of good cleanability has only a low degree of surface roughness. These conditions are ideal to make the stoppers stick to the underside of the shelf. After pressing the stoppers down, the shelves are separated again. At this stage, it is absolutely undesired that the stoppers adhere to the shelf above. If they do, then the shelf may pull the stopper slightly out of its seated position and thereby impair the integrity of the seal until the aluminum seal is

152

R. Ovadia et al.

Fig. 6 Lyo stoppers with one, two, and three vent openings, and with different flange markings

Fig. 7 A shelf of a pilot lyophilization chamber after unsuccessful insertion of the stoppers. Stoppers adhered to the shelf above and the stoppers/ vials were pulled out of the vial pack upon raising the shelves. This situation would present a great challenge to automatic unloading systems

applied during capping. Another consequence of stopper adhesion to the shelf is that the entire vial may be lifted out from the vial pack and then, under the influence of gravity, falls down again but out of position (Fig. 7). This phenomenon has highly undesired consequences that reach from product loss and outside contamination of neighboring vials to the inability to automatically unload the shelves. Therefore, the flange marking on the stoppers, along with other stopper properties and shelf properties, plays an important role in preventing this problem. If the flange markings are well designed, the probability of shelf sticking can be substantially reduced. Of the flange markings that are depicted in Fig. 6, the one at the right-hand side is the worst in terms of sticking to shelves because the uninterrupted circle on the flange can act as a suction cup on the underside of the shelf that presses the stopper down. With the flange design in the middle, which consists of two interrupted circles of different diameters, this is extremely less likely. The design on the left-hand side in this respect shows an intermediate behavior. As an alternative to flange markings, there is another solution to prevent hang-up to shelves. That solution consists of laminating the stopper flange top with a non-sticky coating, typically a fluoropolymer film (Fig. 8). The result is that no siliconized rubber surface contacts the shelf but the plastic film that is completely non-tacky. Yet another solution is to coat the complete stopper with a non-tacky fluoropolymer film. The presence of such a film on the top of the flange will prevent shelf hang-up, while its presence on the plug part is favorable for improved drug compatibility. It is to be noted that coating or laminating the stopper to prevent shelf sticking has a significant impact on stopper cost because, apart from the cost of coating or film, the standard stopper manufacturing process cannot be used and a more complex, costlier method must be employed. Other aspects of fluoropolymer-coated lyo stoppers will be discussed later on in this chapter.

Primary Container Closure System Selection for Lyophilized Drug Products

153

Fig. 8 A stopper with film coating on the top flange (left) and completely coated stopper (right)

3.1.2 Penetration Thickness The penetration thickness of an elastomeric closure is the thickness of the stopper in the area where it is penetrated with a needle. This thickness is an important dimension since it plays a significant role in determining the coring (fragmentation) behavior, the resealing behavior and the needle penetration behavior of the stopper. All other factors like needle quality, needle bevel design, rubber compound, etc., remaining equal, a higher penetration thickness definitely leads to a higher penetration force, and to a higher likelihood that during penetration rubber fragments are “scraped off” by the needle (named fragmentation or coring). When penetrating with larger diameter needle such as these used on Closed System Transfer Devices (CSTD), as is becoming more and more common, the coring behavior is often even more prominent. At the same time, at least in theory, a higher penetration thickness leads to a higher probability of adequate resealing upon withdrawal of the needle, which could be important for multi-use vials. Additionally, after capping of the vial, when it has been ascertained that the closure/vial interface is tightly sealed so that no gas can come in via that route, penetration thickness of the closure determines the permeability to gases of the stopper/vial/cap combination. Given a certain rubber material, higher penetration thicknesses lead to higher resistance to permeation of air (oxygen may be of concern) and moisture into the vial and thus into the drug. This aspect of the membrane in the center of the stopper acting as a gas barrier after crimping of the vial is especially significant for freeze-drying stoppers. The background of this is twofold. First of all, many freeze-dry vials are stoppered with the lyophilization chamber under vacuum. This vacuum is present mainly to facilitate lyophilizate reconstitution by assisting the transfer of reconstitution liquid, which is in a syringe at atmospheric pressure and will therefore be drawn into the vial by the pressure differential. Obviously, it is the intention that this vacuum is preserved over the lifetime of the drug product. Therefore, after capping of freeze-drying vials, the role of the penetration area of the stopper as gas barrier is of very high importance. The higher the penetration thickness, the longer ingress of air is delayed. Secondly, the reason for freeze-drying a product is that it is not stable in an aqueous environment. The purpose of the primary and secondary drying phases is to remove as much as possible water from the freeze-dried cake in order to ensure drug stability. Consequently, again after capping, the penetration area of the stopper also plays a role as a moisture barrier. (Of course, atmospheric moisture is also a gas!) In order to improve barrier property against the ingress of air and moisture into the freeze-dried vial, the penetration thickness of a lyophilization closure is therefore typically higher than for a stopper that is used for a liquid fill. This is particularly true for 13 mm parenteral lyophilization closures. 3.1.3 Plug Design The role of the plug of a freeze-drying closure is more critical from a functional point of view than that of a closure used for a liquid or dry powder fill. The plug design for the lyo stopper is more intricate in order to address the different phases of transport and of freeze-drying. The different functionalities of the plug are illustrated using the stopper drawing in Fig. 9. The discussion below holds for this particular design but may be transferable only in part to lyophilization stoppers of a different design. A first feature that is typical of a lyophilization stopper is that the plug wall does not form a full cylinder over an angle of 360°. It is interrupted for the creation of a vent opening that during freeze-drying allows the evacuation of the sublimated ice. In this example, the plug wall is interrupted twice. Thereby two “legs” are created. Next, it is easily recognizable that the above stopper has a plug with different diameters at various positions along the height of the plug. Dimension D1 is the diameter of the plug that before capping ensures closure/vial seal integrity. D1 is slightly larger than diameter D2. This D2 is the diameter that, when positioning the stopper after filling of the freeze-dry solution, is the first to enter in the vial neck. For that purpose, it is advantageous to have a smaller diameter. D2 on the other

154

R. Ovadia et al.

Fig. 9 A particular design of a freeze-drying closure and important dimensions

hand must not be too small either. It must be larger than the internal neck diameter of the vial since this oversizing will contribute to the stability of the stopper in half-way down position. Of course, during all of the transport stages on the filling line itself, on the way to the lyophilizer, and during loading of the vials into the freeze-dryer, the stopper must firmly stay on the vial without tilting or falling down, because this would jeopardize subsequent steps in the process. Stability of the stopper in half-way down position however is not only dependent on the oversizing of D2 versus the internal vial neck diameter. There is another diameter that plays a role, namely D3. D3 is larger than D2 and also larger than internal neck diameters. The transition between the two diameters D2 and D3 is made by a chamfer, so that a rib is formed that runs over the circumference of the 2 legs of the stopper. In half-way down position, this rib is resting on the rim of the vial. Thereby, it forms an obstacle against the tilting and tumbling of the stopper in a plane that is “parallel” to the two vent openings. Equally of interest is a second, this time partial rib that is located a little bit lower on the plug. The rib that is described above, together with this partial rib, over a limited angle form an annular space that can accommodate the blowback rim of the vial on which the stopper is placed. Also this increases the stability of the stopper in half-way down position since a second obstacle against tilting and tumbling is formed. This statement is valid provided that the stopper and the vial are designed to match with each other. If this is not the case, the partial rib will not exert its functionality, but on the other hand, it is not likely to negatively influence stopper stability. After complete insertion of the stopper and before capping, closure/vial seal integrity is assured by the plug diameter D1 that is oversized versus the internal vial neck diameter. The seal however will only stretch over the height H1 underneath the stopper flange and above the vent openings. The sealing height is a very important dimension for a lyophilization stopper. If the stopper after full insertion, e.g., as a result of flange/shelf sticking or of mechanical agitation during transport to the capping station would creep out of the vial neck, then part of the sealing height is lost and the quality of the seal before capping is affected, in part or in full. Raised stoppers may cause loss of the vacuum that was present before stopper insertion. This is seen as a critical quality attribute for a finished lyophilization vial. Illustration and discussion of this phenomenon can be found in publications like [21]. Additional provisions in the stopper plug design can be made to control and prevent stopper pop-up after full seating, e.g., by providing a blowback ring below the flange and/or an annular ring that is located on the plug body below the flange and that engages the blowback of the vial. These provisions are not present in the stopper that is illustrated above. They also will only have their effect if the stopper and the vial are designed to perfectly match with each other. If this is not the case, then seal integrity before capping worst case may even be harmed. Interesting illustrations of exemplary and non-exemplary stopper/vial matches can be found in [22].

3.1.4 Standardization of Lyophilization Stoppers Geometry There are ISO standards for lyophilization closures. These standards are not exhaustive when it comes to stopper dimensioning. Only a minimum number of dimensions are standardized. This is a consequence of so many lyophilization stopper geometries having been developed and being made available to the market over time so comprehensive universal standard dimensions cannot be easily defined. One feature that leads to diversity is the number of vent openings. In Fig. 9, there are two vent openings that are symmetrically located over the plug circumference (“2-leg-lyo-stopper”). Other designs have three vent openings (“3-leg lyo stopper”). Yet another frequently encountered stopper type has only one, but wider vent opening (“igloo stopper”). All of these styles of stoppers are illustrated in Fig. 6. Splitting the vent area over multiple openings also has other consequences that become apparent at the time of lyo cake reconstitution and preparation of the injection into the patient. The single large vent opening of the igloo type stopper shown

Primary Container Closure System Selection for Lyophilized Drug Products

155

on the left-hand side in Fig. 6 allows to visually follow the introduction of the needle through the stopper penetration area into the vial, at least if the turbidity of the reconstituted solution and the size of the stopper both permit this. The needle is attached to a syringe in which the drug solution is collected before injection into the patient. With the igloo type of stopper, it is possible to easily visually monitor how much of the vial drug solution is withdrawn into the syringe and eventually delivered to the patient. With a 2-leg stopper, this is already not easy anymore. With a 3-leg stopper, it is not possible at all anymore, since the cannula tip will always be hidden, if not by the crimp cap, then the legs of the rubber stopper. Moreover, with this type of stopper, there is the additional disadvantage that when the stopper is in fully seated position, the legs are forced to come together. Thereby, capillaries are formed between the legs in which reconstituted drug solution is held up. Such 3-leg designs therefore increase the residual volume in the vial. A last disadvantage of 3-leg designs is that during packaging at the closure manufacturer or during transport to the user of the closures, the legs get entangled, resulting in “twin pairs” of stoppers that do not separate again and that disturb good machineability behavior on filling lines.

3.2

Moisture Absorption/Desorption and Permeability Behavior

Lyophilization is a process that is applied if the drug product on the longer run is not stable in an aqueous medium, and where identity, strength, quality, and purity can only be guaranteed if the product is brought in the form of a lyophilized cake. Upon storage, it is therefore also of importance to keep moisture away from the freeze-dried product. In general, the lower the active dose is, the more important it is to shield the product from moisture. This shielding can be done via appropriate choices of packaging components and crimping conditions. For a lyophilized drug product there are three principal routes by which, after packing, it can be “contaminated” with water: ingress of water via the closure/vial interface (path “I” in the Fig. 10), desorption of water from the elastomeric closure (path “D”), and permeation of water through the closure (path “P”). The ingress route would point to insufficient container/closure integrity and therefore is problematic, both from the point of view of microbial ingress as from the perspective of moisture ingress and loss of underpressure in the vial. For a more extensive discussion, reference is made to other parts in this chapter that discuss container closure integrity. Desorption and permeability will be discussed further below. Elastomeric closures for parenteral use, at the end of their manufacturing process, are always subjected to a washing and drying process. The purpose of this process is to bring the closures in a controlled state of microbiological and particulate cleanliness and to satisfy regulatory requirements that are imposed by various regulations [23]. During the washing and rinsing phases of this process, the closures are for a certain period exposed to water, while also in the step before washing, the closures are for a certain time remaining in wet condition. This means that right before and during the washing and rinsing,

Fig. 10 Moisture paths into freeze-dried vials

156

R. Ovadia et al.

there is time for the closures to absorb water. The absorption of the water takes place at the surface of the stopper and will not reach into the bulk of it. Most of the water will be dried during the drying phase that follows washing and rinsing, but part of it will remain in the stopper. The percentage of moisture that is left in the stoppers will depend on parameters such as stopper compound, duration of the exposure to water before washing, duration of the drying cycle after washing, temperature at drying, storage conditions after drying, etc. Typical numbers for moisture content of elastomeric stoppers as delivered by closure manufacturers are in the order of magnitude of 0.1–0.4% [24]. Before aseptic filling, elastomeric closures are however subjected to sterilization. In most of the cases, this is done by steam sterilization of the closures, followed by drying. During the first part of this process, closures are brought in a saturated steam environment, typically for 30 min at 121 °C (2 bar). The objective of this process obviously is to bring stoppers in a sterile state, ready to be used for filling. However, the side effect of it is that stoppers again will absorb water (steam!) and that the moisture level of the stoppers increases to levels that are 4–5 times the initial level before sterilization. At longer sterilization times, the effect is even more pronounced [25, 26]. In order to reduce this level again, stoppers are dried. The moisture level after drying can be influenced by varying temperature and pressure during drying and by the length of the drying cycle [24–28]. Lately, the use of ready-to-use stoppers is becoming more and more popular. Such stoppers are sterilized either by steam or by radiation. In the latter case, the moisture uptake during sterilization obviously will not take place, but the radiation dose must not be excessive as to degrade the elastomer. Low rate water absorption during storage after sterilization however cannot be excluded and must be mitigated if needed. A number of “low residual moisture” rubber formulations have been brought to the market that are characterized by a considerably lower moisture pick-up during steam sterilization in comparison with other, traditional rubber formulations that are used for lyophilization stoppers. This feature of those formulations can be attributed to their composition in that they do not or only contain little rubber ingredients that easily absorb moisture. Furthermore, they are also characterized by drying rates that are comparable to those of other, more moisture absorbing formulations. The moisture absorption and drying behavior of such rubber formulation is illustrated here below (Fig. 11). The advantage of this type of rubber formulation is that the stoppers at the beginning of the drug life cycle contain less moisture. Since the moisture in the stopper is in the immediate vicinity of the lyophilized drug and is therefore, by desorption from the stopper, easily accessible, the use of low residual moisture stoppers is beneficial for drug stability. For this reasoning to be valid, there are however a number of conditions to be met. The first of those conditions is that the stoppers that are sterilized and dried do not pick up moisture again between the time of drying and the time of being placed on vials. Under normal practical conditions this is not an issue, since the duration of the storage and the storage conditions of the stoppers after drying are not excessive, and since stoppers that pick up less moisture during steam sterilization display the same behavior upon exposure to atmospheric moisture. A second condition however is that the moisture desorption mechanism from the stopper is not outscored by a second mechanism of moisture transfer to the drug product, namely permeation through the stopper. Desorption of moisture from the stopper and permeation of moisture through the stopper are 2 phenomena that run in parallel. However, whereas moisture absorption/desorption (“D” in Fig. 10) is determined by stopper composition and the presence of materials with a “hydrophilic” character, stopper permeation (“P” in the figure) is related to stopper permeability, which relies on a different physical principle. The moisture that is most readily transferred to a lyophilization cake is the moisture that is left in the stopper after steam sterilization and drying. This water has the shortest way to the cake, as it does not need to permeate through the stopper. Desorption of water from the stopper is a phenomenon that starts right away after complete stoppering of the vials. Permeation of water through the stopper is an effect that plays a role on the longer term, since the water that reaches the cake in this way first has to permeate through the entire thickness of the rubber stopper in its

Fig. 11 Moisture absorption/desorption of a low moisture absorption rubber compound in comparison with a typical compound (both bromobutyl)

Primary Container Closure System Selection for Lyophilized Drug Products

157

penetration zone. The ideal lyophilization stopper shall therefore have both a low residual moisture level at stoppering and a low moisture permeability, or more generally, a low gas permeability. Stopper permeability equally depends from stopper composition, but not in the same way as absorption/desorption does. Stopper permeability is primarily dictated by the permeability of the elastomer that is used as the base polymer in the rubber and, in second place, by the type of fillers that are used in the rubber. Low gas permeability is always achieved by the use of halobutyl elastomer (bromobutyl or chlorobutyl) as elastomer base in lyophilization closures. Fillers that are currently used are always silicates, but depending on the type, a lower or higher permeability will be reached. Blending of bromobutyl or chlorobutyl with elastomers that have a higher permeability such as polyisoprene in rubber formulations for lyophilization closures from a perspective of stopper permeability is not indicated. Such blending sometimes is undertaken in rubber formulations for vial stoppers for liquid fill applications or for plungers for prefilled syringes, in order to improve the mechanical performance of the material, e.g., by increase of its elasticity. A discussion of the role of water/absorption and of water permeability can be found in [28].

3.3

Fluoropolymer Coatings

Whereas the base polymer used for rubber compounds for lyophilization closures is always halobutyl, more and more lyophilization closures, especially for biologicals, are closures that at their surface are covered with a fluoropolymer coating. If a coating of this nature is applied, then it will always be present in the area of the stopper that is facing the lyophilization cake (“drug contacting area”), and it may or may not be present on the top surface of the stopper flange. The primary reason for applying such a type of coating in the drug contacting area is to achieve a better compatibility of the stopper with the freezedried cake. Improving compatibility in fact comes down to further reducing the probability of interaction between the closure and the freeze-dried drug product. Reduction of interaction is achieved by the barrier effect of the coating, meaning that a fluoropolymer coating further reduces the extractability of chemical compounds from the stopper. The extractables level from a coated stopper therefore is lower than the level from a stopper in the same rubber compound but without fluoropolymer coating. The coating in this respect acts as a barrier between the stopper and the drug product. Choosing stoppers with a fluoropolymer coating in the drug contacting area gives the advantage of achieving substantially higher success rates in drug stability studies and in shortening time to market for the drug product. The reason for applying a fluoropolymer coating on the stopper flange is different. Obviously, the barrier effect of the coating on this part of the stopper does not serve a purpose. The fluoropolymer coating however, very much unlike the halobutyl substrate for the coating, has a totally different tackiness. The fluoropolymer coating makes the stopper completely non-sticky, also without silicone being applied to it. Therefore, stoppers that are fluoropolymer coated on the top of their flanges will not stick to the underside of lyophilizer shelves when they are pressed down. Application of a fluoropolymer coating in the flange area is the most powerful, however also the most expensive measure to avoid stoppers sticking to shelves. A further benefit of a closure that is fluoropolymer coated over its entire surface, not only in the drug contacting area and in the flange area, but also present on the sidewall and on the underside of the flange, is that such a closure does not need any surface siliconization at the end of its manufacturing process and yet will not clump together. The absence of surface silicone is beneficial again from a compatibility point of view for silicone sensitive drugs, such as certain biologicals. Whereas fluoropolymer coated stoppers offer many advantages, there are also some considerations to make. The first one is that the fluoropolymer coating does not act as a water vapor barrier. The absorption/desorption behavior of a fluoropolymer coated stopper is not governed by the coating but by the behavior of the rubber compound on which the coating is placed. An illustration of the absence of impact of the fluoropolymer coating on moisture absorption/desorption behavior of a totally coated stopper in steam sterilization and drying can be found in [26]. As a result, there is also no effect on headspace moisture levels in vials that are stoppered with uncoated and with fluoropolymer coated stoppers of this type. A second consideration is that the fluoropolymer coating in the drug contacting area is not limited to just this area, but that the coating extends to the outer surface of the igloo or of the legs that form the plug of the stopper. The latter surface comes in contact with the vial neck when the stopper is in half-way down position before lyophilization and when it is fully down after complete stoppering of the vial in the freeze-drier. The fluoropolymer layer behaves differently in comparison with a typically siliconized surface of an uncoated stopper. Fluoropolymer-coated stoppers are more slippery in comparison with their siliconized versions, and, all other things such as the vial remaining the same, are more sensitive to displacement in half-way down position. They are also are seen to more easily lead to rising of stoppers after full stopper insertion and before capping. This in turn can lead to a transient container/closure integrity failure that ends when the stoppered vial is crimped, but in the meantime may have led to

158

R. Ovadia et al.

effects on the composition of the vial head space and potentially to a lack of sterility assurance. Altogether the behavior of the stopper/vial combination in transport phases right after filling, during loading and unloading of the freeze-drier and between unloading and vial crimping need to be closely monitored.

4

Container Closure Fit and Integrity

Robert Ovadia Roger Asselta

Selecting each individual primary packaging component for a lyophilized drug product requires careful consideration, as alluded to in the earlier sections. Pharmaceutical companies must also consider the compatibility and dimensional fit between the chosen vial, stopper, and cap. Health agencies require all pharmaceutical drug product primary containers systems to have demonstratable container closure integrity (CCI) after the crimping operation; however, for lyophilized drug products, stoppered but uncrimped vials must also maintain CCI and the prescribed level of vacuum before the crimping operation because they may remain uncrimped for at least hours and potentially, days. There are three main sealing surfaces between a stopper and vial: (1) the plug seal, also known as the valve seal, (2) the transition seal, and (3) the land seal, also known as the flange seal, as shown in Fig. 12 [22, 29]. It has long been established that the flange seal is the primary seal and occurs at the interface of the top of the vial finish and the bottom surface of the stopper flange [29]. It is considered the only seal which can be consistently relied upon due to the many stresses (temperature, pressure, transportation, vibration, etc. . . .) that a final CCS may experience. Prior to crimping of the aluminum seal, the vial is considered closed but not sealed. The plug seal, formed between the stopper plug and the inner neck of the vial, is considered a secondary seal. With lyophilized products, the plug seal plays a critical role in maintaining the quality characteristics of the headspace after full stopper insertion at the conclusion of the lyophilization cycle; the plug seal is responsible for maintaining the integrity of the CCS prior to the crimping operation [29]. Many companies opt to use components that have standardized dimensions and properties, such as those proposed by the International Organization for Standardization (ISO), because they are perceived to be suitable off-the-shelf options [30–32]. Yet, pairing any ISO vial with any ISO stopper does not guarantee that the system will have a robust fit due to three main reasons. First, ISO vials may have one of three different blowback options and a particular stopper type may not fit well with all three [30]. Second, ISO stoppers have varying vent configurations of “igloo,” “2-leg,” and “3-leg” (described in Sect. 3), and may or may not have a no-pop ring; a single vial type may not fit well with every stopper option. Third, two “identical” ISO components from different manufacturers, or even the same manufacturer but produced at a different manufacturing site, may look and feel different even though they are considered “equivalent.” This is not necessarily the fault of the component manufacturer since there may be multiple interpretations for what ISO has suggested or differences in raw materials (i.e.,

Fig. 12 The sealing surfaces of a vial-stopper system: (1) plug, (2) transition, and (3) flange seals

Primary Container Closure System Selection for Lyophilized Drug Products

159

different glass tubing sources for vials). Companies may also have a strong business or technical justification to select a non-ISO conforming component. Due to the many variations in components, assessing the fit of both a crimped and uncrimped container closure system (CCS) is an essential part in the development of a robust lyophilization CCS to ensure a safe drug product.

4.1

Strategies and Methods for Assessing Component Fit

The concepts of Quality by Design (QbD) should be used to mitigate issues during the packaging development phase for each product. That is building in quality with a deep understanding of the components and their processes and addressing potential risks. Proper component selection enhances manufacturability by minimizing risk and reducing loss. There are several tools that can be utilized during package development to assess the fit between components. Each method has benefits and drawbacks so pharmaceutical companies should utilize multiple complimentary techniques when selecting their primary CCS, troubleshooting existing configurations, or designing custom configurations.

4.1.1 Visualization Techniques Visualization techniques provide a means to qualitatively assess how components fit together and the contact surface areas between them, allowing a rapid elimination of obvious poor matches. Three different visualization methods have been reported that allow for direct observation of critical sealing surfaces [22, 33–35]. The first two methods, reported by the same authors, assess stoppered but uncrimped systems by (1) using novel image generation techniques and superimposing the images of each component and (2) immersing the stoppered vial in oil for photography using a camera equipped with a telephoto lens [22]. In this work, four different configurations were evaluated for their sealing surfaces and stopper retention features. The images and analysis of all four systems under both methods are reproduced in Fig. 13. The first two methods are quick, experimentally facile, and inexpensive ways to assess uncrimped systems. One potential drawback is that they cannot visualize crimped systems since a cap hinders the “optical viewing” of sealing surfaces. Microcomputed tomography (micro-CT) is an X-ray-based technique capable of generating images at micrometer level resolution for direct visual analysis of not only uncrimped but also crimped CCSs [33–35]. In addition, 3-D renderings of each individual

Fig. 13 For (a–d), the top represents superimposed images and the bottom represents oil-immersed photographs. (a) Components with adequate sealing surfaces but an inward pressure on the stopper legs can deform the stopper flange so that the sealing surface above the vent is raised slightly, (b) components with adequate sealing surfaces but lack a mechanism to keep the stopper fully inserted (i.e., raised stoppers are observed), (c) components with poor sealing surfaces and frequently lose vacuum despite having a stopper no-pop ring, and (d) components with robust sealing surfaces and stopper retention features

160

R. Ovadia et al.

Fig. 14 Top, bottom (Bot), and side views of a micro-CT scanned CCS that was rendered into 3D models; models before and after segmentation are shown. Segmentation provides a means to analyze the individual features of interest on the vial, cap, and stopper

component can be generated, and the component’s features of interest can be analyzed in Fig. 14. Micro-CT is not a high throughput method and requires costly, complex equipment and software. The benefits of generating micro-CT images and 3-D renderings should be compared against the high cost and time-consuming process. Superimposition and oil immersion methods are cheaper and faster for screening uncrimped systems whereas micro-CT could be applied for crimped systems. Finally, it’s worth noting that micro-CT is the only method capable of visualizing in-situ systems under deep cold storage conditions such as -80 °C [36]; however, this is not a concern for most lyophilization drug products since they are stored at room temperature or refrigerated conditions. Direct visual analysis of critical component contact surfaces could be evaluated using any of the three methods. All three techniques complement each other and the other approaches listed in this chapter.

4.1.2 Interference Fit Calculation: Assessment of the Plug Seal The previous section introduced qualitative and rapid ways of assessing component pairing. Interference fit calculation is a complementary method that provides quantitative comparison between different promising configurations. The plug seal is formed between the stopper plug and the vial bore. The interference fit percentage, an indirect measurement of lateral stopper compression, can be calculated to assess the effectiveness of the plug seal between a vial and a stopper. The equation is shown below: Stopper plug diameter - vial inner neck diameter × 100 = % interference fit Stopper plug diameter The strain of the compressed rubber stopper creates a surface stress that establishes the seal inside the vial. While this non-subjective calculation provides a quantitative value, the results may not always be easily interpreted. Values that approach zero percent suggest that there is not enough lateral compression on the stopper plug to form a seal with the vial bore (i.e., a loose fitting stopper). On the other hand, very high values may present other challenges such as over deformation and increased stoppering force in the lyophilizer or increased chance of stopper pop-up. In both scenarios, the CCS may lose CCI and the required headspace attributes. There are reported recommended target values of 2–10% [37]; however, different vial-stopper combinations will have different optimal values. Understanding the influence of the dimensional variation of the components’ diameters and the mating of the configured geometries of the selected components are critical to establishing a suitable CCS. The most robust interference fit percentage for a configuration depends on not only the dimensions but also

Primary Container Closure System Selection for Lyophilized Drug Products

161

other factors such as: the rubber formulation and durometer, the geometries of the vial and stopper, and the size of the components (13 mm, 20 mm, etc. . . .). Ideally, one would have access to a library of components for a given vial and stopper that have incremental variations, such as 0.1 mm, at key dimensions (i.e., stopper outer diameter and vial neck inner diameter). This would allow the user to experimentally assess multiple interference fit percentages and select the most robust value for a specific configuration by using CCIT (container closure integrity testing); however, fabricating a library is costly and time consuming for component suppliers. The interference fit percentage can be used during early stages of initial component selection to rule out any configurations that have values too low or too high. Pharmaceutical companies should assess component fit by complementing the calculation with alternate methods.

4.1.3 Dimensional Stack-Up Analysis: Assessment of Flange Seal and Skirt Length The flange seal is formed under appropriate stopper compression during application of a crimp cap. Sufficient compression, which can be measured by residual seal force testing (described later), is required to maintain robust CCI of the CCS; the integrity of the system may be compromised if the stopper is under low compression [38, 39]. On the other hand, over compressed stoppers increase the risk of cosmetic defects and while they may not always have an impact on CCI, they should be avoided [33, 39, 40]. The target level of stopper compression to maintain CCI and avoid defects must be balanced with the aluminum seal skirt length as seen in Fig. 15 [39, 40]. Seals can be supplied with different target skirt lengths in which shortest and longest offerings may vary as little as 1.0 mm; this small difference could impact CCI of a system if the seal was not correctly sized to the vial and stopper. Systems that use seals with short skirt lengths may lead to an accidental over compression of stoppers during capping to avoid assembly challenges at target compression in which the aluminum cannot be fully tucked under the vial flange (Fig. 16). Over compressing stoppers could lead to an increased rate of defects such as deformation, dimpling, or rupture of the stopper, which may impact CCI, as seen in Fig. 17 [33]. Using seals with long skirt lengths may lead to an accidental under compression of stoppers to avoid visual defects seen at target compression, such as excessive aluminum overhang, and may impact CCI (Fig. 18). Exclusively relying on crimp visual appearance exacerbates the outcome of using seals with skirt lengths that are either too short or too long. Each CCS has a performance window that should be identified and characterized to ensure robust sealing [40].

Fig. 15 Mapping a balance of container closure system performance window to avoid defects

162 Fig. 16 CCS with a short skirt length. (a) Target stopper compression cannot be achieved since the aluminum cannot be fully tucked under the vial. (b) Over stopper compression is likely in order to obtain proper crimp aesthetics

Fig. 17 Photograph (left) and micro-CT scanned image (right) of a dimpled stopper, which may not be easily detected without removing the plastic flip-off button

Fig. 18 CCS with a long skirt length. (a) Target stopper compression cannot be achieved since the aluminum wraps below the vial neck. (b) Under stopper compression is likely in order to obtain proper crimp aesthetics

R. Ovadia et al.

Primary Container Closure System Selection for Lyophilized Drug Products

163

Dimensional stack-up analysis is a calculative method capable of assessing the amount of aluminum available for crimping and can predict if the stopper can be robustly compressed to maintain CCI while minimizing any potential defects. The equation to calculate the excess skirt length (i.e., the portion that will get tucked under the vial flange) is shown below: ðSkirt length - Aluminum thicknessÞ - Vial flange thickness - ðStopper flange thickness × ð1 - %Stopper compressionÞÞ = Excess skirt length

Previous reports have suggested that the target parameters for 20 mm configurations are around 20% stopper compression and an excess skirt length between 0.75 mm and 1.5 mm [40, 41]. The target stopper compression and excess skirt length for each configuration may vary and should be correlated to a CCIT method for robust protection. The calculation may also be supplemented with statistical variability data on the critical dimensions for each component and the capping process [33, 40]. Calculating the skirt length by relying exclusively on the target dimensions may result in operating close to an edge of failure. Dimensional stack-up analysis should be applied when assessing the robustness of the flange seal and can aid in characterizing the performance window for a CCS.

4.1.4 Container Closure Integrity and Seal Quality Test Methods Container closure integrity (CCI) is the ability of the packaging system to protect and maintain the quality of the drug product though delivery to the patient or expiry. USP states: “. . .the maximum allowable leakage into and out of intact packages should be so minimal that there is no impact on product safety, and no consequential impact on the product’s physicochemical stability.” CCI or package integrity is defined as “the absence of package leakage greater than the product package maximum allowable leakage limit (MALL).” An integral package must: • • • •

Prevent microbial ingress (ensure sterility) Maintain drug product quality Limit loss of product contents Prevent transmission of debris or detrimental gasses

Per USP , MALL is the smallest gap (leak) or leak rate that puts product quality at risk (sometimes called the “critical leak”). Inherent package integrity is the leakage rate of a well-assembled (sealed) CCS using defect-free (conforming to specifications) components. Lyophilized products typically require a lower MALL as it is often important to maintain defined headspace requirements which may include partial vacuum, inert gasses, and moisture ingress prevention. CCSs for lyophilized drug products must have a higher inherent package integrity (IPI). Any evaluation of the suitability of the CCS should consider storage and transportation. CCI and seal quality test methods are commonly used to assess component fit. These methods are described in USP and have been thoroughly studied [41–44]; each has benefits and drawbacks. No single method should be exclusively applied during CCS development and method selection will depend on the drug product and primary CCS. CCI test methods are split into two categories: deterministic and probabilistic. USP , EU GMP Annex 1, and health agencies prefer deterministic methods, such as he-leak and laser headspace (both discussed later), over probabilistic methods, such as dye ingress, for improved method sensitivity, reliability, and patient safety. Residual seal force represents the only non-subjective and quantifiable seal quality test applicable to parenteral CCSs. The methods discussed in this section are not inclusive of all those that can be used during development of a lyo CCS. Instead, the focus is on a select few that can support assessing the component fit for both uncrimped and crimped CCSs. Demonstrating CCI on the crimped CCS is a regulatory requirement; however, the redundant sealing surfaces (plug, flange, and transition) of a CCS may all contribute to an integral system. For example, a system with a robust plug seal will pass CCI when crimped even in absence of a flange seal (i.e., not enough stopper compression). Understanding which surfaces contribute to the integrity of a CCS is essential to developing a robust configuration, and while there is no regulatory requirement mentioning which sealing surface must be present, it is understood that both plug (uncrimped) and flange (crimped) seals must be present for a lyophilization CCS.

164

R. Ovadia et al.

4.1.4.1 Helium Leakage Helium leakage (he-leak) is considered the most sensitive CCI test method and can be applied in combination with other techniques when assessing component fit. There are numerous publications that describe the he-leak method, application with crimped CCSs, leak rate limits, and testing CCSs at deep cold storage conditions [33, 34, 42, 43]. A recent study reported a method to independently assess the redundant sealing surfaces of a CCS using he-leak [38]. Lyo rubber stoppers were sectioned to compromise either flange or plug seals; the procedure can be seen in Fig. 19. This technique allows for direct measurement of the one sealing surface, without interference from the other, and should be used in combination with other methods when testing for CCI and/or vacuum retention prior to the capping operation. Results on the plug-only configuration demonstrated that integrity was maintained in absence of a flange seal (for that particular stopper and vial configuration), which represents an uncrimped configuration (i.e., after stoppering and before capping), while the flange-only configuration failed CCI testing at instances in which residual seal forces were < 20 N. A separate study was conducted to assess the same CCS using intact, not modified, stoppers at residual sealing forces of < 20 N and all tested vials passed CCI testing, which suggests that the plug seal is the major/only contributor at low stopper compression. Scenarios in which the stopper is not sufficiently compressed by the cap will likely go undetected during routine batch CCI testing. Patient safety and product efficacy is at risk since the flange seal is the only seal which provides robust sealing behavior under stress conditions (e.g., cold storage, air transportation etc. . . .). Most pharmaceutical companies perform CCI testing only on intact systems, yet these results highlight the importance of independently assessing the sealing surfaces during development. To mitigate the risk of insufficient stopper compression, methods could be implemented during routine manufacturing such as 100% in-line CCI testing via laser headspace, or residual seal force testing when combined with a statistical approach to estimate batch-wide stopper compression; both topics are discussed later. While there are many benefits to using he-leak as a CCI test method, companies should also consider the drawbacks as it pertains to their CCS and product-specific requirements. He-leak is not “operator friendly” since vials must be manipulated, such as cutting their heel using specialized electrically powered tools or applying epoxy to previously punctured stoppers. In addition, the lyophilized powder or cake must be reconstituted and washed out so handling potent compounds, such as antibody drug conjugates, poses other environmental, health, and safety concerns. To mitigate these concerns, a non-destructive method, such as laser headspace, could be used in combination with and correlated to he-leak.

Fig. 19 Procedure to section stoppers that enable independent assessment of the plug seal (a) or flange seal (b)

Primary Container Closure System Selection for Lyophilized Drug Products

165

4.1.4.2 Laser-Based Headspace Analysis Laser-based headspace analysis (HSA) is another technique for monitoring CCI. USP and many publications describe the technology in detail [44–46]. In short, a vial’s headspace environment, such as moisture content, oxygen content, or pressure, can be rapidly detected and quantified using modulated diode laser spectroscopy. The technique is applied non-destructively in a laboratory setting or in-line during routine manufacturing for 100% in-line monitoring (not to be confused with 100% batch inspection). Using HSA during routine manufacturing is highly desirable as it provides a means to monitor the integrity of every vial. 100% in-line monitoring should not replace the development efforts needed to qualify a robust CCS because manufacturing challenges will arise, even for a robust CCS. These challenges could be caused by many reasons such as a batch of components that are near their edge of specification for a critical dimension or an equipment operating at the edge of a critical process parameter range (e.g., capper head pressure). These scenarios, while uncommon, may increase the likelihood of a CCI failure putting product efficacy and patient safety at risk. Instructions are generally given to the end user (e.g., health care providers) to discard any vial that has lost vacuum but this safety precaution, only applicable to lyophilized products, should be used as a last resort because it cannot be consistently relied on; hence, the benefit to monitor every vial during manufacturing and reject those that do not pass the acceptance criteria. The benefit of implementing 100% in-line monitoring during routine manufacturing should be weighed against the probability of such an event occurring, especially if other in-process controls or alternate strategies are implemented to reduce the frequency of CCI failures (e.g., RSF, which is discussed later). Pharmaceutical companies could use ad-hoc 100% CCI testing to support deviations and/or investigations in which the integrity of vials within a batch is unknown. In either scenario (100% in-line monitoring or ad-hoc), development work is required before companies can apply the technology for a new drug product and CCS. Finally, the sensitivity of HSA is lower compared to he-leak; thus, there is a strong technical justification to characterize the robustness of a CCS using complementary CCIT methods during development prior to implementation of 100% in-line monitoring or ad-hoc testing [46]. 4.1.4.3 Residual Seal Force The residual seal force (RSF) is the force exerted by a compressed stopper against a vial finish [47–50]. The RSF test method provides an indirect way of measuring percent stopper compression and can be correlated to an appropriate CCIT method to facilitate CCS characterization or technology transfers. Several RSF-related studies have concluded that this non-subjective seal quality test enables consistent and robust sealing across various capping equipment [33, 35, 38, 51–54]. RSF measurements, in combination with statistical models, can support to use of acceptable ranges on test samples taken during capper setup to ensure that equipment process parameters and inherent component variations will not impact seal quality or CCI for a given drug product lot [38]. RSF acceptable ranges also lower the probability of observing samples that have under or over compressed stoppers (described earlier), which could remain undetected after visual inspection. RSF methodology is best applied after the CCS has been thoroughly characterized by other techniques. The time dependent variability of RSF should be understood for each CCS. Stoppers are viscoelastic so under constant stress (compression with an aluminum seal), they not only maintain elastic energy but also dissipate viscous energy; this is known a stress-strain relaxation. Theoretical models of the stoppers’ stress-strain relaxation have proven to be great predictors of experimental data [54]. RSF values decrease over time with the largest decrease typically observed over the first 24 h; however, the extent of this effect is dependent on the configuration and initial compression force [33, 38, 53]. A well-sealed CCS will maintain a sufficient RSF to assure CCI for the shelf-life of the product. In addition, gathering RSF values during capping development support the setting of appropriate capping parameter ranges. The statistical distribution of the RSF data is a key input to generating a representative statistical model [38]. RSF methodology can be applied during capper development and routine setup as a complementary tool to CCIT.

4.2

Common Causes of Loss of Vacuum and/or CCI: Raised Stoppers and Dried Product

Previous studies demonstrate a small but not insignificant number of vials fail to meet headspace attribute requirements [21, 44]. The failures are primarily caused by vials that had a “leaky” valve seal prior to capping and the failure rate can vary significantly from lot to lot. One explanation of these failures can be attributed to vials that had raised or displaced stoppers. Once crimped, previously failed vials may become integral and CCIT methods other than HSA will not detect the headspace changes such as loss of vacuum [44].

166

R. Ovadia et al.

EU GMP Annex 1 requires that vials with lyophilized product be 100% inspected for raised or displaced stoppers prior to capping. Raised stopper detection systems (RSDSs) are sophisticated, in-line equipment that utilize multiple camera angles to generate images of CCSs. Image processing software determines the maximum observed distance between the bottom of the stopper flange and the top of the vial finish and CCSs that exceed a safety limit are rejected. Prior to implementation of RSDS, various distances or gaps are correlated to a CCIT method to support setting of the safety limit; this is accomplished by placing shims between the vial and stopper to artificially create a gap. RSDSs can reduce the frequency of CCSs that fail to maintain their headspace requirements, but it may not eliminate their occurrence entirely. The correlation between distances and CCIT is usually based on a small sample set and is probabilistic by nature. RSDSs may also falsely reject samples that are integral. Designing and selecting components that don’t have tendency to form raised stoppers is the most effective approach at controlling the failure mode. Recent advances in component design have included geometric features, such as a no pop-ring on the stopper, and processing improvements, such as elimination of uncured/unbound silicone oil on the stopper, have mitigated, but not eliminated, the occurrence of raised stoppers. However, companies may not have the option to change the CCS for an older product so in this case, using RSDSs and/or 100% in-line HSA is the most effective way to reject vials that do not meet headspace attribute requirements. Another potential failure mode for a “leaky” valve seal prior to capping occurs during filling and/or lyophilizer loading processes and may occur regardless if raised stoppers are detected. During these operations, excessive splashing within the vial could occur resulting in liquid DP that is present between the sealing surfaces (plug, land, or both) of a vial and stopper [55]. After lyophilization, the freeze-dried DP in this region may compromise CCI if present across an entire sealing surface. Vials that are observed to have dried DP in the neck are rejected because of their potential impact on product quality attributes. 100% in-line monitoring via HSA is the only method capable to detecting this defect; therefore, improvements to the filling and/or lyophilizer loading process should be implemented if the failure rate is unacceptable.

5

Future Trends

5.1

Dual-Chamber Containers

Vials are by far the container type which is used most for freeze-drying applications [6]. The diluent is delivered in another vial or in a pre-filled syringe. A dual-chamber system (either a cartridge or a syringe) offers the advantage of having the lyophilized product and the diluent in one container. A plunger stopper in the middle of the container separates both chambers (see Fig. 20). By pressing on the second stopper at the end of the container, the liquid can flow through the bypass at the side and starts the reconstitution. Korpus et al. provide an overview for the heat-transfer during lyophilization for dual-chamber cartridges [56]. Despite their obvious in-use convenience, lyophilized products in dual chamber containers require a different, more complex and costly manufacturing process, limiting their general adoption. Furthermore, water permeation properties of the plunger separating the chambers must be thoroughly evaluated.

5.2

Nested Vial Configurations and Press-Fit Caps

The introduction of personalized healthcare for patients within various fields of applications drives the need for more flexible filling concepts, especially for smaller batch sizes. This resulted in the development of standardized robotic aseptic filling technologies like, e.g., the SA25 from Vanrx (part of Cytiva) [20]. Such isolator-based filling systems are designed to run without human intervention but require ready-to-use components in a nested configuration. Vials in a nest face a different thermal environment than in a hexagonal close-packed configuration [57]. This could impact product temperature profiles, primary drying times, and therefore product quality.

Fig. 20 Schematic image of the functionality of dual-chamber syringe (a) and actual image of the system (b)

Primary Container Closure System Selection for Lyophilized Drug Products

167

Fig. 21 Nested vial and cap/stopper configurations (left) and standard crimped vial and a vial with a Daikyo Plascap® (right)

New nested press-fit caps were developed in the past years to pair with the nested vials. The caps are supposed to be used within the lyophilization chamber or in the isolators and consist of polymer components like polypropylene or polycarbonate. These sterile caps already include the stopper and allow a single step sealing and “crimping” process by just pressing down the cap onto the vial opening allowing the cap to “snap” into place at the end of the freeze-drying cycle. Thus, the standard postlyophilization crimping process and equipment is replaced by this procedure (see Fig. 21). The capping force within the lyophilization chamber should be thoroughly considered for each configuration. For a conventional aluminum crimp sealed CCS, stoppering a shelf of vials within a lyophilization chamber relates to the force required to stopper a single vial multiplied by the number of vials on a given shelf. In reality, the force is larger than this calculation due to slight variations in component heights; taller vials will get stoppered first and in order to fully stopper the shorter vials, the stoppers on the taller vials will need to get slightly compressed thus increasing the stoppering force. The capping force for press-fit caps is far greater than the stoppering force because each stopper will need to be fully compressed in order to “snap-on” the caps. Like stoppering vials, slight variations in component heights will further increase the capping force. The capping force should be evaluated for smaller, densely nested vials, such as 2R configurations, in which there are more vials to cap within a single shelf and the capping process potentially pose equipment challenges [58]. Such caps could potentially interfere with the water vapor flow during sublimation, but studies using LyoSeals® and Plascaps® revealed no interference with the sublimation process and no systematic influence on the product temperature during primary drying [57]. The press-fit caps did not affect residual moisture, crystallinity, or the macro- and microscopic product structure. Additionally, no impact of the nest on the water vapor flow was observed. However, the nests had an influence on the macro- and microscopic appearances of the lyo cakes with a trend towards lower product temperatures for the nested vials. This needs to be taken into consideration [57]. Finally, ready-to-use components that are supplied in a nest and tub format require more consumables compared to bulk components. Sustainability should be thoroughly evaluated as they relate to each company’s mission.

6

Conclusion

During drug product development, the importance of the primary container is well recognized, though, this is typically only in the context of component compatibility with the drug product solution. Besides chemical compatibility, the selection of primary container components is often just based on availability or convenience, driven by tight timelines. Usually, for the development scientists and their management, container closure integrity is implied. Furthermore, cost is also factor not necessarily visible at that stage, particularly important for high volume, low profit margin products (e.g., vaccines) which are still subjected to the same CCI requirements as for high margin products. While the newer and more exotic components available may indeed offer superior performance and other manufacturing benefits, the cost associated with these may be 2–5 times that of the traditional glass vial and off-the-shelf rubber stopper – crimp seal closure. Many organizations lack the resources or the expertise to perform systematic component selection, and the consequences of poor choices do not surface until much later in the development phase where mitigation of the problem is lengthy and costly. For parenteral drugs, preserving container closure integrity throughout their shelf life is absolutely critical to prevent contamination by potential pathogens and to preserve product quality. In that regard, health agencies have very low tolerances for mistakes.

168

R. Ovadia et al.

In this chapter, we have presented various aspects pertaining to the function of primary containers, some of which are not widely discussed in most publications primarily focused on lyophilization. It is our hope that the material herein serves as a good introduction to primary container systems and highlights their importance in the overall drug product development and manufacturing processes.

References 1. Pikal MJ, Roy ML, Shah S. Mass and heat transfer in vial freeze-drying of pharmaceuticals: role of the vial. J Pharm Sci. 1984;73 2. Pikal MJ, Shah S. Moisture transfer from stopper to product and resulting stability implications. Dev Biol Stand. 1991;74:165–79. 3. Sreedhara A, et al. Challenges of using closed system transfer devices with biological drug products: an industry perspective. J Pharm Sci. 2020;109:22–9. 4. Ward K, Matejtschuk P, editors. Lyophilization of pharmaceuticals and biologicals: new technologies and approaches. New York: Humana Press/Springer; 2019. 5. Rey L, May JC. Freeze drying/lyophilization of pharmaceutical and biological products. 3rd ed. New York: Informa Healthcare; 2016. 6. Janssen R. Recent trends in lyophilized delivery devices and packaging. In: Varshney DB, Singh MJ, editors. Lyophilized biologics and vaccines. New York: Springer; 2015. https://doi.org/10.1007/978-1-4939-2383-0_16. 7. Hibler S, Gieseler H. Heat transfer characteristics of current primary packaging systems for pharmaceutical freeze-drying. J Pharm Sci. 2012;101:4025–31. 8. Wenzel T, Gieseler H. Molded vial manufacturing and its impact on heat transfer during freeze-drying: vial geometry considerations. AAPS PharmSciTech. 2021;22:57. 9. Scutella B, Passot S, Bourles E, Fonseca F, Trelea IC. How vial geometry variability influences heat transfer and product temperature during freeze-drying. J Pharm Sci. 2017;106:770–8. 10. Hibler S, Wagner C, Gieseler H. Vial freeze-drying, part 1: new insights into heat transfer characteristics of tubing and molded vials. J Pharm Sci. 2012;101:1189–201. 11. Hibler S, Wagner C, Gieseler H. Comparison of product drying performance in molded and serum tubing vials using gentamicin sulfate as a model system. Pharm Dev Technol. 2012;17:541–51. 12. Cannon A, Shemeley K. Statistical evaluation of vial design features that influence sublimation rates during primary drying. Pharm Res. 2004;21:536–42. 13. Machak DR, Smay GL. Failure of glass tubing vials during lyophilization. PDA J Pharm Sci Technol. 2019;73:30–8. 14. Jiang G, et al. Mechanistic studies of glass vial breakage for frozen formulations. I. Vial breakage caused by & hellip. PDA J Pharm Sci Technol. 2007;61:441–51. 15. Chow EJ, Kitaguchi B, Trier M. Effects of subzero temperature exposure and supercooling on glass vial breakage: risk management and other applications in cold chain distribution. PDA J Pharm Sci Technol. 2012;66:55–62. 16. Milton N, et al. Vial breakage during freeze-drying: crystallization of sodium chloride in sodium chloride-sucrose frozen aqueous solutions. J Pharm Sci. 2007;96:1848–53. 17. Jiang G, et al. Mechanistic studies of glass vial breakage for frozen formulations. II. Vial breakage caused by & hellip. PDA J Pharm Sci Technol. 2007;61:452–60. 18. ICH guideline Q3D (R1) on elemental impurities. European Union News. 19. Abdul-Fattah AM, et al. Investigating factors leading to fogging of glass vials in lyophilized drug products. Eur J Pharm Biopharm. 2013;85: 314–26. 20. Mungikar A, Ludzinski M, Kamat M. Effect of the design of the stopper including dimension, type, and vent area on lyophilization process. PDA J Pharm Sci Technol. 2010;64 21. Lighthouse-instruments, Application note: detecting raised stoppers in sterile freeze dried vials; 2013. 22. Lam P, Stern A. Visualization techniques for assessing design factors that affect the interaction between pharmaceutical vials and stoppers. PDA J Pharm Sci Technol. 2010;64:182–7. 23. Food and Drug Administration. Guidance for industry: sterile drug products produced by aseptic processing — current good manufacturing practice. Rockville: Center for Biologics Evaluation and Research. US Department of Health and Human Services; 2004. 24. Templeton AC, et al. Determination of the moisture content of bromobutyl rubber stoppers as a function of processing: implications for the stability of lyophilized products. PDA J Pharm Sci Technol. 2003;57 25. Donovan PD, Corvari V, Burton MD, Rajagopalan N. Effect of stopper processing conditions on moisture content and ramifications for lyophilized products: comparison of “low” and “high” moisture uptake stoppers. PDA J Pharm Sci Technol. 2007;61 26. Janssen R. Transmission of moisture into the headspace of lyophilization vials – a study on various lyophilization closures. Webinar; 2012. 27. Kruszynski M, Tempel J, Akers MJ. Drying effects on residual moisture of lyophilization rubber closures. PDA J Pharm Sci Technol. 2004;58 28. Sasaki H, Kikuchi J, Maeda T, Kuboniwa H. Impact of different elastomer formulations on moisture permeation through stoppers used for lyophilized products stored under humid conditions. PDA J Pharm Sci Technol. 2010;64 29. Morton DK. Container/closure integrity of parenteral vials. J Parenter Sci Technol. 1987;41:145–58. 30. ISO 8362-1: Injection containers and accessories - Part 1: Injection vials made of glass tubing. 31. ISO 8362-5: Injection containers and accessories - Part 5: Freeze drying closures for injection vials. 32. ISO 8362-6: Injection containers and accessories - Part 6: Caps made of aluminium-plastics combinations for injection vials. 33. Mathaes R, et al. Impact of vial capping on residual seal force and container closure integrity. PDA J Pharm Sci Technol. 2016;70:12–29. 34. Nieto A, et al. Evaluation of container closure system integrity for frozen storage drug products. PDA J Pharm Sci Technol. 2016;70:120–33. 35. Ovadia R, Lam P, Tegoulia V, Maa YF. Quantifying the vial-capping process: reexamination using micro-computed tomography. PDA J Pharm Sci Technol. 2020;74:171–84.

Primary Container Closure System Selection for Lyophilized Drug Products

169

36. Edey M, Duncan D. Developing a readily available primary packaging system for use in an ultra cold chain for COVID19 vaccine global distribution using a scientific approach. PDA Europe Parenteral Packaging Conference; 2021. 37. Curry W, et al. Reducing the risk of contamination of sterile parenteral products via ready-to-use closure components. AAPS PharmSciTech. 2010;11:1572–9. 38. Ovadia R, et al. Quantifying the vial capping process: residual seal force and container closure integrity. PDA J Pharm Sci Technol. 2019;73:2– 15. 39. Zeng Q. Considerations for achieving an optimal performance window for container closure systems. PDA Letter LV; 2019. p. 30–4. 40. Bucci A, Ho L, Orme L, Zeng Q. A vial container closure system performance optimization case study using comprehensive dimensional stackup analyses. PDA J Pharm Sci Technol. 2020;74:368–76. 41. Asselta R. Parenteral vial sealing and integrity. Webinar, (April 10, 2019). 42. Kirsch LE, Nguyen L, Moeckly CS, Gerth R. Pharmaceutical container/closure integrity II: the relationship between microbial ingress and helium leak rates in rubber-stoppered glass vials. PDA J Pharm Sci Technol. 1997;51:195–202. 43. Morrical BD, et al. Leak testing in parenteral packaging: establishment of direct correlation between helium leak rate measurements and microbial ingress for two different leak types. PDA J Pharm Sci Technol. 2007;61:226–36. 44. Duncan DI, Veale JR. White paper: an introduction to laser-based headspace inspection and the application to 100% container closure inspection of sterile pharmaceutical containers; 2015. p. 1–10. 45. Hede JO, Fosbol PL, Berg SW, Dahl S. Container closure integrity testing-method development for freeze-dried products using laser-based headspace oxygen analysis. PDA J Pharm Sci Technol. 2019;73:170–80. 46. Pelaez SS, et al. Comparing physical container closure integrity test methods and artificial leak methodologies. PDA J Pharm Sci Technol. 2019;73:220–34. 47. Morton DK, Lordi NG. Residual seal force measurement of parenteral vials. II. Elastomer evaluation. J Parenter Sci Technol. 1988;42:57–61. 48. Morton DK, Lordi NG. Residual seal force measurement of parenteral vials. I Methodology. J Parenter Sci Technol. 1988;42:23–9. 49. Ludwig JD, Nolan PD, Davis CW. Automated method for determining Instron Residual Seal Force of glass vial/rubber closure systems. J Parenter Sci Technol. 1993;47:211–53. 50. Ludwig JD, Davis CW. Automated method for determining Instron Residual Seal Force of glass vial/rubber closure systems. Part II. 13 mm vials. PDA J Pharm Sci Technol. 1995;49:253–6. 51. Mathaes R, et al. Influence of different container closure systems and capping process parameters on product quality and container closure integrity (CCI) in GMP drug product manufacturing. PDA J Pharm Sci Technol. 2016;70:109–19. 52. Mathaes R, et al. The pharmaceutical capping process-correlation between residual seal force, torque moment, and flip-off removal force. PDA J Pharm Sci Technol. 2016;70:218–29. 53. Buecheler J, Luemkemann J, Gieseler H, Mohl S, Streubel A. Residual seal force (RSF) testing: a suitable method for seal quality determination also for high potent parenterals. PDA J Pharm Sci Technol. 2019;73(2):111–20. 54. Zeng Q, Zhao X. Time-dependent testing evaluation and modeling for rubber stopper seal performance. PDA J Pharm Sci Technol. 2018;72: 134–48. 55. Patel SM, et al. Lyophilized drug product cake appearance: what is acceptable? J Pharm Sci. 2017;106:1706–21. 56. Korpus C, Haase T, Sonnichsen C, Friess W. Energy transfer during freeze-drying in dual-chamber cartridges. J Pharm Sci. 2015;104:1750–8. 57. Wenzel T, Gieseler H. Evaluation of packaging materials in freeze-drying: use of polymer caps and nested vials and their impact on process and product attributes. AAPS PharmSciTech. 2021;22:82. 58. Mossman M. Characterization of nested press-fit caps and container closure integrity. PDA Europe Parenteral Packaging Conference. 2020.

Vial Breakage During Lyophilization Jim A. Searles and Ekneet K. Sahni

Abstract

This chapter reviews causes of breakage of borosilicate glass vials during lyophilization and provides advice for troubleshooting. Such breakage can contaminate neighboring vials with glass particulate, damage equipment, and disrupt manufacturing operations. Key vial breakage mechanisms are the crystallization of solutes like mannitol; thermal expansion of the frozen plug during warming, and a rapid break-free/plugging-off event during freezing. Volume expansion of water upon freezing is not a major cause of breakage because the fill volume is usually half the vial capacity. We describe how strain gauges can be used at small scale to reveal underlying mechanisms and eliminate the problem. Remediation options include external vial coatings to protect vials from damage in washing and depyrogenation tunnel equipment; internal hydrophobic coating to reduce adherence between the vial and the frozen plug; annealing to ensure full solute crystallization before drying begins; and use of lower initial primary drying product temperature. Keywords

Lyophilization · Glass vials · Breakage · Strain gauges · Crystallization · Annealing · Depyrogenation · Mannitol

1

Introduction

Glass vials have been used as a primary container closure for many pharmaceuticals with a history of maintaining the container closure integrity, assurance of sterility, and product compatibility while offering low cost [1, 2]. However, current pharmaceutical borosilicate glass compositions are nearly the same as they were in the 1930s [3]. Limitations of traditional borosilicate glass vials include susceptibility to difficult-to-detect damage during shipping and processing, chemical erosion called delamination, and outright cracking and breakage. It is of utmost importance to ensure that pharmaceutical products are free from extraneous materials. Cracking or breakage of glass vials in aseptic product manufacturing facilities can lead to the presence of glass particulates within sealed final containers [4]. Injection of such particulates can cause life-threatening adverse consequences, including pulmonary emboli and infarction, stroke, and heart attack. In spite of layers of controls and prevention, glass particles in sterile injectable product have caused 8 recalls in the United States between Aug 2017 and May 2021, encompassing 14 lots of 7 products [5]. This review will focus upon vial breakage, specifically during the lyophilization process, and mitigations through formulation and process changes. Without considering the product filled into them, the vials that are used for sterile products are able to withstand the temperature changes during lyophilization and freeze-thaw [6]. Breakage of vials that occurs during the lyophilization and freeze-thaw processes are the result of excessive forces placed upon the vials by the product within

J. A. Searles Pfizer Biotherapeutics Pharmaceutical Research and Development, Chesterfield, MO, USA E. K. Sahni (✉) Pfizer Global Supply, Global Technology and Engineering, McPherson, KS, USA e-mail: ekneet.sahni@pfizer.com; # America Association of Pharmaceutical Scientists 2023 F. Jameel (ed.), Principles and Practices of Lyophilization in Product Development and Manufacturing, AAPS Advances in the Pharmaceutical Sciences Series 59, https://doi.org/10.1007/978-3-031-12634-5_9

171

172

J. A. Searles and E. K. Sahni

them. Even a small number of broken vials (