Neurohistology and Imaging Techniques [1st ed.] 9781071604267, 9781071604281

Table of contents :
Front Matter ....Pages i-xiv
Neurohistology with a Touch of History: Technology-Driven Research (Floris G. Wouterlood, Thomas P. Langer)....Pages 1-48
Fixation Protocols for Neurohistology: Neurons to Genes (Elliott J. Mufson, Sylvia E. Perez, Christy M. Kelley, Melissa J. Alldred, Stephen D. Ginsberg)....Pages 49-71
Three-Dimensional Atlases of Insect Brains (Basil el Jundi, Stanley Heinze)....Pages 73-124
Neuroanatomical Tracing Based on Selective Fluorochrome Expression in Transgenic Animals (Floris G. Wouterlood)....Pages 125-156
Optical Imaging Probes for Amyloid Diseases in Brain (Pratyush Kumar Mishra, Myeong-Gyun Kang, Hyun-Woo Rhee)....Pages 157-182
Chemical Clearing of GFP-Expressing Neural Tissues (Klaus Becker, Saiedeh Saghafi, Christian Hahn, Nina Jährling, Hans-Ulrich Dodt)....Pages 183-199
The Properties of Light Governing Biological Microscopy (Pina Colarusso)....Pages 201-223
Beyond Brightfield: “Forgotten” Microscopic Modalities (Radek Pelc)....Pages 225-244
Stereomicroscopy in Neuroanatomy (Erin E. Wilson, William Chambers, Radek Pelc, Paul Nothnagle, Michael W. Davidson)....Pages 245-274
Conventional, Apodized, and Relief Phase-Contrast Microscopy (Radek Pelc, Zdeněk Hostounský, Tatsuro Otaki, Kaoru Katoh)....Pages 275-323
Ultramicroscopy of Nerve Fibers and Neurons: Fine-Tuning the Light Sheets (Saiedeh Saghafi, Klaus Becker, Nina Jährling, Christian Hahn, Hans-Ulrich Dodt)....Pages 325-339
Imaging and Electrophysiology of Individual Neurites Functionally Isolated in Microchannels (Heinz D. Wanzenboeck, Petra Scholze, Johann K. Mika)....Pages 341-377
Consumer Versus Dedicated Digital Cameras in Photomicrography (Jörg Piper, Radek Pelc)....Pages 379-401
Digital Micrographs in Pathology (Roger S. Riley, Jorge Almenara, Christine E. Fuller)....Pages 403-458
Back Matter ....Pages 459-472

Citation preview

Neuromethods 153

Radek Pelc Wolfgang Walz J. Ronald Doucette Editors

Neurohistology and Imaging Techniques

NEUROMETHODS

Series Editor Wolfgang Walz University of Saskatchewan Saskatoon, SK, Canada

For further volumes: http://www.springer.com/series/7657

Neuromethods publishes cutting-edge methods and protocols in all areas of neuroscience as well as translational neurological and mental research. Each volume in the series offers tested laboratory protocols, step-by-step methods for reproducible lab experiments and addresses methodological controversies and pitfalls in order to aid neuroscientists in experimentation. Neuromethods focuses on traditional and emerging topics with wide-ranging implications to brain function, such as electrophysiology, neuroimaging, behavioral analysis, genomics, neurodegeneration, translational research and clinical trials. Neuromethods provides investigators and trainees with highly useful compendiums of key strategies and approaches for successful research in animal and human brain function including translational “bench to bedside” approaches to mental and neurological diseases.

Neurohistology and Imaging Techniques Edited by

Radek Pelc Czech Academy of Sciences and Charles University, Prague, Czech Republic

Wolfgang Walz Department of Psychiatry, University of Saskatchewan, Saskatoon, SK, Canada

J. Ronald Doucette Department of Anatomy and Cell Biology University of Saskatchewan, Saskatoon, SK, Canada

Editors Radek Pelc Czech Academy of Sciences and Charles University Prague, Czech Republic

Wolfgang Walz Department of Psychiatry University of Saskatchewan Saskatoon, SK, Canada

J. Ronald Doucette Department of Anatomy and Cell Biology University of Saskatchewan Saskatoon, SK, Canada

ISSN 0893-2336 ISSN 1940-6045 (electronic) Neuromethods ISBN 978-1-0716-0426-7 ISBN 978-1-0716-0428-1 (eBook) https://doi.org/10.1007/978-1-0716-0428-1 © Springer Science+Business Media LLC 2020 This work is subject to copyright. All rights are reserved by the Publisher, 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 publisher, 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 publisher 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 publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. The cover photo is a color version of Fig.1E shown in Chapter 3 (by El Jundi and Heinze) - brain of solitary sweat bee (Megalopta genalis) This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer Nature. The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A.

Preface to the Series Experimental life sciences have two basic foundations: concepts and tools. The Neuromethods series focuses on the tools and techniques unique to the investigation of the nervous system and excitable cells. It will not, however, shortchange the concept side of things as care has been taken to integrate these tools within the context of the concepts and questions under investigation. In this way, the series is unique in that it not only collects protocols but also includes theoretical background information and critiques which led to the methods and their development. Thus, it gives the reader a better understanding of the origin of the techniques and their potential future development. The Neuromethods publishing program strikes a balance between recent and exciting developments like those concerning new animal models of disease, imaging, in vivo methods, and more established techniques, including, for example, immunocytochemistry and electrophysiological technologies. New trainees in neurosciences still need a sound footing in these older methods in order to apply a critical approach to their results. Under the guidance of its founders, Alan Boulton and Glen Baker, the Neuromethods series has been a success since its first volume published through Humana Press in 1985. The series continues to flourish through many changes over the years. It is now published under the umbrella of Springer Protocols. While methods involving brain research have changed a lot since the series started, the publishing environment and technology have changed even more radically. Neuromethods has the distinct layout and style of the Springer Protocols program, designed specifically for readability and ease of reference in a laboratory setting. The careful application of methods is potentially the most important step in the process of scientific inquiry. In the past, new methodologies led the way in developing new disciplines in the biological and medical sciences. For example, physiology emerged out of anatomy in the nineteenth century by harnessing new methods based on the newly discovered phenomenon of electricity. Nowadays, the relationships between disciplines and methods are more complex. Methods are now widely shared between disciplines and research areas. New developments in electronic publishing make it possible for scientists that encounter new methods to quickly find sources of information electronically. The design of individual volumes and chapters in this series takes this new access technology into account. Springer Protocols makes it possible to download single protocols separately. In addition, Springer makes its print-on-demand technology available globally. A print copy can therefore be acquired quickly and for a competitive price anywhere in the world. Saskatoon, SK, Canada

Wolfgang Walz

v

Preface Two interrelated tools that have been in use for more than a century for studying the normal and pathophysiological functioning of nervous tissue are neurohistology and the microscopic techniques used to image the histological structure of nervous tissue. Functional genomics, proteomics, systems biology, and pharmacogenomics are some of the postgenome project tools being used to study the structure and function of nervous tissue, relying heavily on imaging and neurohistology for data collection and analysis. Interpretation of imaging data often depends crucially on a correct description of the histological structure of the tissue being studied, thus giving meaning to the data in the context of the intricate interplay between structure and function. The goal of this book is to introduce the reader to the following: (a) Major light microscopic imaging techniques that can be used to view nervous tissue (b) Types of questions that can be asked using each imaging technique (c) Main steps involved in using each imaging technique and in interpreting the data. This book provides the reader with an introduction to the basics of some of the main imaging techniques applicable for use on nervous tissue. Armed with this background, neuroscientists can more effectively select the technique(s) most suitable for their experimental needs. Since fixation, histochemical/immunohistochemical staining protocols, the imaging technique, and the digital nature of the data being collected are so intimately interconnected, we have included chapters on each of these topics. To assist the reader in understanding the mechanics of imaging, chapters on the basics of microscopic optics and on the properties of light are also included. Our intent is to provide a basic level of proficiency in understanding the basics of the major light microscopic imaging techniques to both novice users and those wishing to explore alternate imaging tools. As users become more proficient in the use of the imaging tools, it can only lead to improved quality in terms of experimental design and data interpretation. In this book, particular emphasis is given to which imaging technique is most suitable for answering specific questions and on which histochemical and/or immunohistochemical techniques are the most appropriate for addressing the questions being asked. Understanding the strengths and weaknesses of each imaging technique will enable neuroscientists to more effectively study how histological variability from one brain region to another influences function and how pathophysiological changes in the histological structure of a brain region give rise to neurological disorders. Prague, Czech Republic Saskatoon, SK, Canada Saskatoon, SK, Canada

Radek Pelc Wolfgang Walz J. Ronald Doucette

vii

Joseph Ronald Doucette 1953–2016

Professor Ron Doucette was born on January 29th, 1953 at Summerside in Prince Edward Island, Canada. Ron died on May 15th, 2016 after a long and courageous battle with colon cancer. Ron began his academic career by obtaining a Bachelor of Science with Honors in Psychology in 1976 from Acadia University. He then obtained a Master of Arts degree in 1977 from the University of Guelph. In 1982, he received a PhD from the University of Western Ontario. Following this, Ron took up a Post-Doctoral Research posting at McMaster University in Hamilton. Ron then took up a second Post-Doctoral Fellowship in 1983, this time back in London, Ontario. In 1985, Ron took up an Assistant Professorship in the Department of Anatomy and Cell Biology at the University of Saskatchewan. He moved through the ranks to Full Professor and for a time he served as Department Head. Ron’s initial main field of research interest was investigating a small cell in the nasal mucosa called an olfactory ensheathing cell. This cell type directs the migration of transplanted glial cells to areas of demyelination. Subsequently, his research interest evolved to include oligodendrocytes, and in collaboration with Dr. Adil Nazarali studied early transcriptional events in the formation of myelin. During his career, Ron was awarded 25 grants with a total of over 2.2 million dollars, some in collaboration with Dr. Nazarali. Ron published over 50 peer-reviewed full-length papers in his field and supervised 12 graduate students. However, Ron was also a successful educator and received several College of Medicine Teacher awards. Ron was passionate about his work—he was a skilled teacher and researcher who loved to educate others about the brain and the anatomy of the human body. Ron loved to learn and looked at each day as an opportunity for new knowledge. He was a wonderful colleague and close friend. He leaves his wife of 42 years Daphne and his daughter Amanda, who became a successful lawyer. Richard M. Devon Department of Anatomy and Cell Biology University of Saskatchewan Saskatoon, SK, Canada

ix

Contents Preface to the Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Joseph Ronald Doucette 1953–2016 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v vii ix xiii

1 Neurohistology with a Touch of History: Technology-Driven Research . . . . . . . Floris G. Wouterlood and Thomas P. Langer 2 Fixation Protocols for Neurohistology: Neurons to Genes . . . . . . . . . . . . . . . . . . . Elliott J. Mufson, Sylvia E. Perez, Christy M. Kelley, Melissa J. Alldred, and Stephen D. Ginsberg 3 Three-Dimensional Atlases of Insect Brains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Basil el Jundi and Stanley Heinze 4 Neuroanatomical Tracing Based on Selective Fluorochrome Expression in Transgenic Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Floris G. Wouterlood 5 Optical Imaging Probes for Amyloid Diseases in Brain. . . . . . . . . . . . . . . . . . . . . . . Pratyush Kumar Mishra, Myeong-Gyun Kang, and Hyun-Woo Rhee 6 Chemical Clearing of GFP-Expressing Neural Tissues . . . . . . . . . . . . . . . . . . . . . . . Klaus Becker, Saiedeh Saghafi, Christian Hahn, Nina J€ a hrling, and Hans-Ulrich Dodt 7 The Properties of Light Governing Biological Microscopy . . . . . . . . . . . . . . . . . . . Pina Colarusso 8 Beyond Brightfield: “Forgotten” Microscopic Modalities . . . . . . . . . . . . . . . . . . . . Radek Pelc 9 Stereomicroscopy in Neuroanatomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Erin E. Wilson, William Chambers, Radek Pelc, Paul Nothnagle, and Michael W. Davidson 10 Conventional, Apodized, and Relief Phase-Contrast Microscopy. . . . . . . . . . . . . . Radek Pelc, Zdeneˇk Hostounsky´, Tatsuro Otaki, and Kaoru Katoh 11 Ultramicroscopy of Nerve Fibers and Neurons: Fine-Tuning the Light Sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Saiedeh Saghafi, Klaus Becker, Nina J€ a hrling, Christian Hahn, and Hans-Ulrich Dodt 12 Imaging and Electrophysiology of Individual Neurites Functionally Isolated in Microchannels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Heinz D. Wanzenboeck, Petra Scholze, and Johann K. Mika

1

xi

49

73

125 157 183

201 225 245

275

325

341

xii

13 14

Contents

Consumer Versus Dedicated Digital Cameras in Photomicrography . . . . . . . . . . . 379 Jo¨rg Piper and Radek Pelc Digital Micrographs in Pathology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403 Roger S. Riley, Jorge Almenara, and Christine E. Fuller

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

459

Contributors MELISSA J. ALLDRED • Center for Dementia Research, Nathan S. Kline Institute for Psychiatric Research, Orangeburg, NY, USA; Department of Psychiatry, New York University Langone Medical Center, New York, NY, USA JORGE ALMENARA • Department of Pathology, School of Medicine, Virginia Commonwealth University, Richmond, VA, USA KLAUS BECKER • Department of Bioelectronics, Institute of Solid State Electronics, Vienna University of Technology, Vienna, Austria; Center for Brain Research, Medical University of Vienna, Vienna, Austria WILLIAM CHAMBERS • Nikon Instruments Inc., Melville, NY, USA PINA COLARUSSO • Live Cell Imaging Laboratory, Department of Physiology and Pharmacology, Snyder Institute for Chronic Diseases, University of Calgary, Calgary, AB, Canada MICHAEL W. DAVIDSON • National High Magnetic Field Laboratory, Department of Biological Science, The Florida State University, Tallahassee, FL, USA HANS-ULRICH DODT • Department of Bioelectronics, Institute of Solid State Electronics, Vienna University of Technology, Vienna, Austria; Center for Brain Research, Medical University of Vienna, Vienna, Austria BASIL EL JUNDI • Biozentrum, University of Wu¨rzburg, Wu¨rzburg, Germany CHRISTINE E. FULLER • Department of Pathology, School of Medicine, Virginia Commonwealth University, Richmond, VA, USA STEPHEN D. GINSBERG • Center for Dementia Research, Nathan S. Kline Institute for Psychiatric Research, Orangeburg, NY, USA; Department of Psychiatry, New York University Langone Medical Center, New York, NY, USA; Department of Neuroscience and Physiology, New York University Langone Medical Center, New York, NY, USA; Neuroscience Institute, New York University Langone Medical Center, New York, NY, USA CHRISTIAN HAHN • Department of Bioeletronics, Institute of Solid State Electronics, Vienna University of Technology, Vienna, Austria; Center for Brain Research, Medical University of Vienna, Vienna, Austria STANLEY HEINZE • Vision Group, Department of Biology, Lund University, Lund, Sweden ZDENEˇK HOSTOUNSKY´ • The Stentor Institute, Great Moravian Academy of Arts & Sciences, Prague-Hostivice, Czech Republic NINA JA€ HRLING • Department of Bioelectronics, Institute of Solid State Electronics, Vienna University of Technology, Vienna, Austria; Center for Brain Research, Medical University of Vienna, Vienna, Austria MYEONG-GYUN KANG • Department of Chemistry, Ulsan National Institute of Science and Technology (UNIST), Ulsan, Republic of Korea KAORU KATOH • Molecular Neurobiology Research Group, Biomedical Research Institute AIST, Tsukuba, Japan CHRISTY M. KELLEY • Department of Neurobiology, Barrow Neurological Institute, Phoenix, AZ, USA THOMAS P. LANGER • Department of Anatomy and Cell Biology, University of Saskatchewan, Saskatoon, SK, Canada

xiii

xiv

Contributors

JOHANN K. MIKA • Institute of Solid State Electronics, TU Wien, Vienna, Austria PRATYUSH KUMAR MISHRA • Department of Chemistry, Ulsan National Institute of Science and Technology (UNIST), Ulsan, Republic of Korea ELLIOTT J. MUFSON • Department of Neurobiology and Neurology, Barrow Neurological Institute, Phoenix, AZ, USA PAUL NOTHNAGLE • Avimo Precision Instruments, Optem International, Fairport, NY, USA TATSURO OTAKI • Optical Research Laboratory, Research & Development Division, Nikon Co., Yokohama, Japan; School of Science and Institute of Industrial Science, The University of Tokyo, Tokyo, Japan RADEK PELC • Institute of Physiology, Czech Academy of Sciences, Prague, Czech Republic; Institute of Biochemistry and Organic Chemistry, Czech Academy of Sciences, Prague, Czech Republic; Third Faculty of Medicine, Charles University, Prague, Czech Republic SYLVIA E. PEREZ • Department of Neurobiology, Barrow Neurological Institute, Phoenix, AZ, USA JO¨RG PIPER • Laboratory for Applied Microscopy Research, Bullay, Rheinland-Pfalz, Germany HYUN-WOO RHEE • Department of Chemistry, Seoul National University, Seoul, Republic of Korea ROGER S. RILEY • Department of Pathology, School of Medicine, Virginia Commonwealth University, Richmond, VA, USA SAIEDEH SAGHAFI • Department of Bioelectronics, Institute of Solid State Electronics, Vienna University of Technology, Vienna, Austria; Center for Brain Research, Medical University of Vienna, Vienna, Austria PETRA SCHOLZE • Department of Pathobiology of the Nervous System, Center for Brain Research, Medical University of Vienna, Vienna, Austria HEINZ D. WANZENBOECK • Institute of Solid State Electronics, TU Wien, Vienna, Austria ERIN E. WILSON • National High Magnetic Field Laboratory, Department of Biological Science, The Florida State University, Tallahassee, FL, USA FLORIS G. WOUTERLOOD • Department of Anatomy and Neurosciences, Amsterdam University Medical Centers, Amsterdam, The Netherlands

Chapter 1 Neurohistology with a Touch of History: Technology-Driven Research Floris G. Wouterlood and Thomas P. Langer Abstract The light microscopic anatomy of the nervous system poses a number of technical challenges. A historical approach shows how compelling technological progress has advanced and directed our understanding of nervous tissue. It is argued that in neurohistology, what one happens to “see” when examining nervous tissues strongly depends on the available technological tools as well as the conceptual framework in which the observations are interpreted. In the present chapter the basic light microscopic anatomy of nervous tissues is reviewed along with methods that have been developed or adopted for studying them. These methods increasingly include neurophysiological, molecular, and genetic manipulation of living nervous tissue. Key words Histology, Nervous system, Stains, Glia, Neurons, Nerves, Staining methods, Genetic constructs, GFP, Gap junctions

Abbreviations CLEM CLSM CNS CSF EM FIF GFAP GFP GRASP HRP MRI MS NT OTC PALM PNS STED

Correlative light and electron microscopy Confocal laser scanning microscopy Central nervous system Cerebrospinal fluid Electron microscopy Formaldehyde-induced fluorescence Glial fibrillary acidic protein Green fluorescent protein GFP fluorescence reconstitution across synapses Horseradish peroxidase Magnetic resonance imaging Multiple sclerosis Neurotransmitter Optical tissue clearing Photoactivated localization microscopy Peripheral nervous system Stimulated emission depletion (microscopy)

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_1, © Springer Science+Business Media LLC 2020

1

2

Floris G. Wouterlood and Thomas P. Langer

STORM TIRF VDIC VT YFP

1

Stochastic optical reconstruction microscopy Total internal reflection fluorescence (microscopy) Voltage-dependent ion channel Vesicular transporter Yellow fluorescent protein

Introduction The venerable discipline of anatomy developed about 250 years ago an offshoot that progressed into what we nowadays call “neurohistology.” Anatomy itself has its roots long ago in the antique world’s medical rituals and beliefs. The new offshoot can be regarded as a natural result of the coalescence in the seventeenth and eighteenth century of radically new ideas as to how to perceive and interpret the natural world that surrounds us. This fresh mindset keeps underlying the modern concept of the central nervous system (CNS) as an extremely complex assembly of electrochemically interacting cells, that collects, processes, exchanges and controls storage of information from all parts of the body, that is, acting like a central processing unit (CPU) in a computer. One of the breakthroughs providing an early support for the “information and communication system” concept was the demonstration of “animal electricity” in the late eighteenth century by Luigi Galvani through his classical—“high-tech,” by then standards—frog leg experiment [1]. Major discoveries in other fields of science and industry of the seventeenth to eighteenth century provided a vital momentum driving a spectacular progress in neurohistological knowledge. The remarkable achievement by Galvani was not sufficient, however, as a major technical hurdle had yet to be cleared: handling the precious organ itself. Position yourself in the seventeenth century and imagine the challenge. A living brain appears remarkably unremarkable to the untrained eye, even when it is operating full blast. Brain and spinal cord are largely colorless, bearing a fragile gelatinous consistency. Moreover, these organs quickly autolyse after death so that without proper fixation it is impossible to study them in earnest. Apart from fixation they need to be embedded, sectioned and stained to reveal their fine structure. Another major obstacle was essentially of a conceptual nature but it nevertheless had a tremendous influence on the direction of progress in neurohistology: the long-lasting dispute between advocates of the Neuron and Reticulum doctrines, that is, whether nervous systems are made up of discrete cellular building blocks (neurons), or exist as one big network (syncytia of cells, i.e., a reticulum of sorts). This dispute lingered on for decades until the introduction of advanced electron microscopy techniques in the second half of the twentieth century brought microscopy

Neurohistology

3

resolution to such a level that the anatomy of the synapse became fully appreciable. The Neuron doctrine (outlined in Subheading 2.5) prevailed and neurohistology thus belongs to the realm of cellular neuroscience. When one looks at fresh brain or spinal cord, some parts may be described as translucent pinkish gray (i.e., gray matter) whereas other parts appear to be bright glistening white (i.e., white matter). The gray matter consists mainly of neuronal cell bodies, dendritic processes and unmyelinated axons. Many myelinated axons jointly make the white matter brilliantly white in much the same way that light scattering makes snow white. Both gray matter and white matter contain glial cells, with astroglia being the dominant cell type in the gray matter, and oligodendroglia in the white matter. Brain’s gelatinous consistency is due to its lack of connective tissue. Unlike most tissues, the CNS has very little extracellular space and almost no connective tissue. It is enclosed in three layers of connective tissue, called meninges, that restrain its movement and jointly with the cerebrospinal fluid (CSF) cushion it against mechanical injury. Additionally, the brain contains a dense mesh of capillary blood vessels providing a soft internal skeleton of sorts. Connective tissue within the brain parenchyma is limited to a thin sheet of basement membrane surrounding the capillaries while some sparse collagen fibers attach the third meningeal layer to the brain surface. On the other hand, the nerves of the peripheral nervous system (PNS) do have connective tissue, wrapping the individual axons (endoneurium), small bundles of axons (perineurium), and the entire nerves (epineurium), which makes them tough and easy to dissect in a cadaver. Until comparatively recently, the description of fiber connectivity in the CNS was far behind comparable descriptions in the PNS. It is not feasible to cut thin slices of fresh brain by hand. When the ancients did cut into a brain, they saw a gelatinous matrix with glistening bands running through it and a chain of hollow, fluidfilled chambers in the depths of the brain, called ventricles. The ventricles were thought to house the humors responsible for sensation and thought, and the brain’s role was believed to be pumping the humors out through the nerves [2–4].

2

Early Efforts to Describe the Nervous System

2.1 The Ancients Take the First Steps

Investigators in the ancient world were in doubt about the brain. The Egyptians did not bother to preserve the brain for the afterlife, choosing to hook it out through the nose prior to mummification. However, a medical treatise of ancient Egypt, the Edwin Smith Surgical Papyrus (written about 1700 B.C. but based upon medical texts dating back to ~3000 B.C.), describes changes in consciousness and behavior following head injuries [2, 5]. Therefore, the

4

Floris G. Wouterlood and Thomas P. Langer

control of movements, thoughts, and consciousness was ascribed to the brain. The brain and spinal cord were studied largely at the time of accidents causing a breach in the skull, or when butchering animals. Some scholars, like Aristotle and Galen, thought that intellect and emotions resided in the heart, a way of thinking still deeply embedded in our language. On the other hand, Plato (Pla´to¯n) considered that perception and intellect were functions of the brain [2]. Hippocrates (in his treatise, “On the Sacred Disease” [6], ~400 B.C.) argued that intellect was an attribute of the brain, because injury to the head resulted in the loss of consciousness. Epilepsy also seemed to be associated with brain injury. The Greek philosopher Alcmaeon (~500 B.C.) conjectured that all the senses are linked to the brain. In 322 B.C., two physicians in Alexandria, Herophilus and Erasistratus, actually dissected human bodies and described many features of the nervous system, including nerves [2]. On the other hand, Galen (~177 A.D.), who also studied in Alexandria, felt that most of the intellect was a function of the heart [2, 7]. Sensation and memory were attributed to the brain’s substance and its ventricles, with “spiritus animalis,” that is, invisible, weightless stuff flowing from the brain through hollow nerves to their destination (muscles) [4]. The notion of spiritus animalis was still alive in the 1600s and even supported by Rene´ Descartes in his 1664 book “l’Homme” [8]. Later followers placed the intellect entirely in the cerebral ventricles. The latter line of thought was dominant in the West until the 1700s [4]. 2.2 The Renaissance Moves the Mind to the Brain

Andreas Vesalius produced credibly accurate images of the brain and its ventricles in his magnum opus “On the Workings of the Human Body” (1543; annotated English translation 2014) [9]. However, it was Thomas Willis, an English royalist during the English Restoration who led in 1664 with his book “Cerebri anatome” [10] to the modern era of neurosciences, much as William Harvey had revolutionized thinking about the heart and circulation a generation earlier. Harvey had shown that the heart was just a pump for the blood, orphaning the intellect and emotions to find a new domicile. Willis and his contemporaries (e.g., Raymond Vieussens in France [11] and Franc¸ois Boe¨ (latinized into Sylvius) in Holland [12]) worked with fresh brains, as it can be deduced from their illustrations showing the brain much flattened by gravity. Around this time, it was discovered that the brain could be hardened by placing it in brandy or port wine. This treatment had the additional advantage of staining the brain’s substance to some extent. As a hardened brain can be sliced some of the brain’s internal structure could now be subjected to further study. One of the new discoveries was Gennari’s stria [13], the fibrous lamina in primary visual cortex formed by afferent myelinated fibers belonging to the optic radiation. In these early preparations, there were regions that took up the stain, which, as we nowadays know are collections of

Neurohistology

5

neuronal cell bodies called nuclei or cortices, and regions that resisted staining, which is due to the high-fat content in myelinated tracts. Also, one could bluntly dissect along tracts and fascicles of myelinated fibers and show that they connect different regions of the nervous system. With the technology available at that time it was possible to start describing the gross anatomy of the nervous system. Our nomenclature still retains many terms introduced by these early investigators, such as “hippocampus,” “substantia nigra,” “fornix,” and “red nucleus.” 2.3 Anatomy Aided by Industrial Technology

Early medical illustrations of bones and muscles are fairly clear as to where the ancients already had a reasonable understanding of human anatomy. The first fairly realistic images of brain and spinal cord appear in the Renaissance period [9, 14] while truly accurate illustrations start to be encountered in eighteenth and nineteenth century publications, coinciding with the Industrial Revolution (mid-eighteenth to mid-nineteenth century). The critical technological innovation at this point was proper fixation of the tissue. The good and versatile fixatives still used today became available relatively recently, such as formaldehyde (introduced by Blum in 1893 [15]; reviews in [16, 17]) and, much later, glutaraldehyde [18]. These are chemicals that crosslink polypeptides and hence inactivate proteins while preserving their structure. Formaldehyde is a white powder that when depolymerized and dissolved in water at 4% (the so-called formalin) mildly fixes the brain to the extent that the tissue is still available for histological staining, histochemical reactions, and immunohistochemical incubations. The introduction of this particular fixative at the turn of the nineteenth and twentieth centuries can be considered as one of the milestones in neurohistology. Fixation also prepares the nervous tissue for a second innovation: differential staining with dyes capable of visualizing cellular constituents. Some fixatives act as mordants that change the chemical properties of the tissue so that it binds dyes more readily. The fixative composition and conditions of its application (pH, temperature, osmolarity, whether applied by immersion or perfusion) is important in many light microscopic stains, and we consider the careful application of a proper fixative and control of the fixation conditions as a key to success. As a by-product of the spectacular rise of the chemical and mechanical industries in Germany in the nineteenth century, another breakthrough occurred in neurohistology: the introduction of organic dyes. A vast array of textile dyes were discovered, some of which happened to stain also brain tissue. The first such stain was thymidine, but that dye was rapidly superseded by others [19, 20]. Cresyl echt violet became the basis of the Nissl stain (the latter is a generic name honoring the German psychiatrist and neuropathologist, Franz Nissl [21], who introduced this class of

6

Floris G. Wouterlood and Thomas P. Langer

staining; other cationic dyes such as thionin, also produce “Nissl” staining) while hematoxylin and eosin became the mainstays of pathological anatomy. Cresyl echt violet and thionin bind to nucleoproteins associated with certain organelles in the cell body of neurons, most notably the rough endoplasmic reticulum. These organelles are hence called “Nissl bodies,” collectively referred to as “tigroid” (i.e., spotted) substance. The nucleolus is also readily stained by “Nissl” stains. Lipophilic dyes bind to fats in the myelin surrounding axons or to degenerating myelin in damaged tracts. Stains developed in this era are based on metallic (especially silver) salts, and are central to our understanding of the structure of the nervous system, because several of them deposit silver salts in intracellular organelles (e.g., the Golgi apparatus), or even produce completely filled (Golgisilver–impregnated) neurons wherein the cytoplasmic matrix contains the silver deposit, leaving organelles such as mitochondria “clean.” The initial paper on silver salts deposition is Camillo Golgi’s description in 1873 of the “reazione nera” [22], but it was the supreme Spanish master, Santiago Ramo´n y Cajal, who refined this impregnation technique to perfection [23]. Dendritic trees and axonal networks, complete with their delicate details thus became prominent while long axonal projections could now be traced through major portions of the brain. It is remarkable that almost all of the histological stains commonly used today were developed about or shortly after 1890. In order to properly see the microscopic structure of the brain it is necessary to cut the brain into sections thin enough to allow uniform permeation of the tissue by the stains and to ensure sufficient transparency. This delicate work requires prior embedding of the brain in a matrix holding it together under the tearing forces of the knife which at the same time must be allowed to cut through the tissue. A third innovation turned up: embedding materials such as paraffin wax and celloidin. The sectioning of nervous tissue is facilitated by dehydrating the tissue through a series of alcohols, replacing the final alcohol with aromatic hydrocarbons such as xylene or toluene, and next by replacing the hydrocarbons by paraffin or plastic, while embedding the tissue in a block that can be cut into sections only a few micrometers thick. A technical development parallel to fixation and embedding was to freeze the brain and then section the frozen tissue block on a microtome. Mechanical microtomes as such were available around 1800 [24]. William Rutherford constructed in 1873 one of the first freezing microtomes [25]. 2.4 Microscopes as Part of the Critical Mass of Innovation

Once a block of fixed brain tissue has been embedded, sectioned, and stained with dyes binding to specific structures in the tissue, one needs an instrument to view the sections at high magnification, that is, a microscope. Although the first primitive microscopes were

Neurohistology

7

constructed as early as ~1600, it was major advances in glass manufacturing and optics design in nineteenth-century Germany that made it possible to construct increasingly better, and at the same time sufficiently user-friendly (i.e., compound) optical instruments until the optical resolution hit a theoretical barrier first recognized in 1873 by Ernst Abbe (“Abbe diffraction limit”), that is, a physical limit equal to approximately half the wavelength of the light used to view the preparations [26]. Any two points in the specimen that are closer to each other than this limit merge into one, and blurriness degrades images of such small details. Precision lens grinding, application of lens coatings and combining lenses made from special (e.g., lithium) glass produced superior microscopes that achieved maximum resolution at minimal chromatic and other aberrations. The stage was thus set to begin studying nervous system structures in earnest. Nervous tissue could now be preserved in a state close to in vivo, the structural components of tissue could be distinguished with selective dyes, and material could be sectioned thin enough to enable its examination under a microscope fitted with achromatic optics powerful enough to discern individual cellular organelles. It can be argued that modern neurohistology only started around 1890 when the innovative forces had reached critical mass [27]. In the twentieth century, a steady stream of technological innovations has expanded neurohistology into a wider discipline where histology, histochemistry, molecular and genetic techniques contribute to a large biomedical field of research: cellular and molecular neuroscience, with overlaps in neurophysiology, behavioral neuroscience, and even medical brain imaging. The optical barrier described nearly a century and a half ago by Abbe [26] appeared not insurmountable. Several solutions to achieve higher resolution by optical means have been proposed [28] of which photoactivated localization microscopy (PALM, also referred to as stochastic optical reconstruction microscopy (STORM)) and microscopy using illumination spot reduction by a second laser (“stimulated depletion”), that is, quenching of emission (STED) are most widely applied [29, 30]. The inventors of PALM/STORM and STED microscopy were awarded the 2014 Nobel Prize in Chemistry for their achievements. Their microscope designs are capable of achieving a lateral resolution better than 50 nm. A completely different approach toward higher optical resolution is to use structured illumination in combination with computer image reconstruction [31]. Confocal microscopy techniques (dealt with in more detail in Subheading 6.2) of course also reshaped the microscopic anatomy landscape a great deal. Two-photon microscopy provides high penetration into tissue [32, 33], with low phototoxicity and photobleaching, and for that matter is being favored for studying living cells, making it possible to visualize almost the entire depth of

8

Floris G. Wouterlood and Thomas P. Langer

(mouse) neocortex. Developments in light sheet microscopy and tissue clearing [34] enable seeing even deeper into the tissue (cf. Subheading 6.1). Microscope adaptive optics (a by-product of telescope technology) further facilitates such efforts [35]. 2.5 The Neuron Doctrine

In 1838, Theodor Schwann described myelinated nerves, and in 1839, he proposed the cell theory [36]. By the 1850s descriptions appeared of a great many structures belonging to the central, peripheral, and enteric nervous systems. In the 1870s and 1880s there was a flourishing of neurohistology and by 1889 Ramo´n y Cajal was beginning to formulate the Neuron doctrine declaring neurons to be the discrete fundamental elements of the nervous system. Von Waldeyer formally introduced the Neuron doctrine in 1891 [37]. In 1906, Golgi and Cajal were jointly awarded a Nobel Prize for their histological studies of the nervous system, largely based on the silver stain method developed by Golgi about 25 years earlier [38–41]. The Neuron doctrine has four tenets: (1) the fundamental structural and functional units of the nervous system are neurons; (2) neurons are discrete cells, not in cytoplasmic continuity with other cells; (3) neurons have three functional and structural regions, the cell body, dendrites, and an axon; (4) information flows from the dendrites and cell body to the axon and travels along the axon to influence other neurons or muscle fibers at synapses. Synapses were proposed for the first time in a physiological context by Sherrington [42] and later included in the Neuron doctrine.

2.6 Cataloging and Comparative Anatomy in Early Neurohistology

The mid to late 1800s was a time of major progress in understanding the microscopic anatomy of nervous tissues. Descriptions appeared of all parts of the nervous system in a wide range of animals. By comparing structures in different species and studying the consequences of deliberate lesions in animals and fortuitous lesions in humans, hypotheses on the roles of gray and white matter areas in brain function could be proposed and verified. One of the big innovations was the development of myelin (Weigert) stains [19] that allowed one to trace axons throughout the nervous system. Damaged myelinated axons lose their myelin, so it is possible to follow connections away from a cell body by looking for the absence of a myelinated tract, or the presence of degenerated myelin (Marchi stain) [43]. It was also observed that the axonal processes of certain collections of neuronal cell bodies converge to form peripheral nerves. Severing such a nerve would cause a characteristic series of changes in the cell bodies of the contributing neurons. Most conspicuous in this respect are the loss of Nissl bodies as reflected in reduced Nissl staining, and swelling of the cell bodies (a condition called “chromatolysis” [44]). This reaction of neurons to lesion of their axons

Neurohistology

9

could be used to retrogradely trace peripheral connections of neurons. The combination of lesion and chromatolysis works less well for central connections. Central connectivity is thus much more difficult to study with that method than peripheral connectivity. The first half of the twentieth century was occupied by a systematic elucidation of the structure and connectivity of the nervous system. It was an important time for laying foundations of modern neuroanatomy and neurophysiology. We feel that the ultimate coalescence of descriptive and comparative neurohistology of that era is the book published in 1936 by Arie¨ns-Kappers, Huber, and Crosby [45]. 2.7 Single Units, Electron Microscopy and Tracing of Neural Connectivity

3

In the 1950s, several innovations ushered in a new period of intense activity that we are still fully engaged in. The perfection of the microelectrode in the 1960s to the extent that individual, living CNS neurons could be impaled [46] made it possible to study single neurons (“units”). The introduction in biomedicine of the electron microscope (EM) in biomedical research in the 1940s [47] paved the way for looking at the details of cell organelles and, ultimately, those of synapses [48–50]. Nauta’s method to selectively silver stain degenerating axoplasm made it possible to trace axonal connections to their terminal arborizations at the light microscope level [51, 52]. The silver degeneration staining methods are quite labor-intensive, albeit refined and made more reliable by Fink and Heimer [53]. This generation of tracing techniques was outperformed in the second half of the twentieth by three new successive generations: radioautographic tracing, macromolecule transport tracing and, in the twenty-first century, tracing based on metabolically produced fluorescent molecules in specific neurons in genetically engineered animals (review in [54]). In the final decades of the twentieth century, molecular biological methods became strongly entrenched in neurohistology: in situ detection of receptors, neuroactive molecules and mRNA, and attempts to unravel the molecular structure and function of the synapse. In recent decades we are increasingly able to study intact, living nervous systems (including in animals performing specific behavioral tasks), and to directly visualize functioning neurons on a microscale, both in tissue culture and in vivo.

The Structure of the Nervous System: Neurohistology Nervous systems contain two broad cell categories other than the support cells in the meninges and blood vessels. These are neurons and glial cells. Neurons actually constitute a minority but, as they are generally larger than glial cells, they occupy about half of the volume of the CNS. Neurons are cells specialized for communication, the core business of the nervous system. Glial cells (named

10

Floris G. Wouterlood and Thomas P. Langer

“Nervenkitt” [nerve glue] by Rudolf Virchow [55, 56]) support the activity in the neurons, by electrically insulating them, modulating their nutrition, and by conditioning the extracellular environment in which they reside. A detailed introduction to neurohistology may be found in Greenfield’s Neuropathology [57] while excellent overviews are provided in Gray’s Anatomy [58] and in Kandel et al.’s textbook [59]. Brodal’s introductory chapter [26] is also an excellent survey up to about 1980. Glia is comprehensively dealt with by Verkhratsky and Butt [60]. In 1979, Radivoj Krstic´ published a classical book with a large amount of artistic 3D interpretations (artist’s impressions) of EM micrographs of mammalian cells, including glial cells and neurons [61]. 3.1

Glia

3.1.1 Schwann Cells and Oligodendroglia

Processes of glial cells fill the interstices between the neurons so that there is very little extracellular space in the CNS, which is very unusual indeed. It is estimated that neurons and glia jointly occupy about 99% of the parenchyma of the mammalian CNS. In the peripheral nervous system all glial cells are Schwann cells, but in the CNS there are multiple types of glia, of which astrocytes, oligodendrocytes, and microglia are the most common in mature brain and spinal cord. Glial cells occur in a number of morphological and functional types. Schwann cells in the PNS and oligodendrocytes in the CNS myelinate axons of neurons and encapsulate neuronal cell bodies. A Schwann cell wraps around a segment of axon in multiple layers and thus consolidates its membrane into an array enriched in certain lipids and proteins, called a myelin sheath. Myelin acts similarly to the insulation of an electrical wire in that it reduces the loss of current from the axon and decreases its capacitance so that more current may travel along the axon rather than out of it [62]. The net effect is that in myelinated axons membrane potentials propagate more efficiently and more rapidly than in “naked” axons. Schwann cells also form a capsule around the cell bodies of ganglion cells, in which case they are called satellite cells. Bundles of unmyelinated axons in the PNS may be enclosed by a membrane derived from a Schwann cell. In the CNS, oligodendrocytes do the same job as the Schwann cells in the periphery. Some of the details are different but the general layout is similar. A major difference is that oligodendrocytes myelinate multiple axons (review in [63]), whereas Schwann cells myelinate single axons. Unmyelinated axons in the CNS are not ensheathed in glial cells but lie apposed to other glial and neuronal processes.

Neurohistology

11

3.1.2 Astroglia

Astrocytes are involved in nurturing neurons by conveying nutrients from the capillaries to the vicinity of neurons, absorbing ions and neurotransmitter molecules from the extracellular space. They also indirectly control the microenvironment of neurons by regulating blood flow through local capillaries to meet the energy needs of neurons in their vicinity [64]. Astrocytic processes fill most of the space between neurons, and endfeet of these processes line capillary walls. The subpial end feet of innumerable astrocytes jointly form one continuous, thin lamina separating the brain parenchyma from the innermost meningeal layer [65]. At sites where processes of different astrocytes appose each other they often form specialized junctions called “gap junctions” containing channels where the astrocytes exchange small molecules (up to ~1.2 kDa) (gap junctions also exist between neurons, see Subheading 9.2). It has thus been speculated that the astrocytes in brain compartments form a kind of syncytium (review in [66]). The main job of astroglia is to maintain the ionic environment in the CNS by sequestering and releasing ions. Additionally, they process (recycle) some neurotransmitters so that they can be reused later. If there is neuronal degeneration in any brain area, the astrocytes locally proliferate, engulf and remove the debris, while forming a glial scar. However, this is not a scar in the usual sense as no connective tissue is created. Astroglia come in a number of morphological types, but the major difference is between those cells that reside in gray matter, the protoplasmic astrocytes, and those that lie in white matter, the fibrous astrocytes. Protoplasmic astrocytes have plumper, more irregularly shaped processes while fibrous astrocytes have straighter, thinner processes. A general astrocyte marker is glial fibrillary acidic protein (GFAP) (Fig. 1).

3.1.3 Microglia and Ependyma

In healthy brain, microglia consist of inconspicuous, small, rather shriveled cells with a relatively large nucleus, that occur scattered in brain parenchyma. They differ in embryonic origin from neurons and astroglia in the sense that they stem from mesodermal progenitors. Microglial cells can be considered as resident brain macrophages whose role is to act as immunological watchdogs monitoring the brain milieu and mobilizing an immunological attack on invasive organisms and/or molecules. If a microglial cell encounters a molecule that it recognizes as a potential problem, its increased activity goes hand in hand with a change in morphology: it becomes larger while it signals the systemic immune system to transfer macrophages into the area in response to potential damage. This signaling may backfire, though: microglia cells are suspected to exacerbate pathological events in multiple sclerosis (MS) [67]. As microglia and macrophages share the same embryonic cell heritage it is often difficult using immunohistochemical markers to distinguish active microglia from infiltrated blood macrophages [68].

12

Floris G. Wouterlood and Thomas P. Langer

Fig. 1 Glial cells. (A) Golgi-silver–stained astrocyte with many thin processes radiating away from the cell body. (B) Detail of glial fibrillary acidic protein (GFAP) immunostaining in rat hippocampus: astrocytic processes. (C) Golgisilver–stained astrocyte with lamellar processes and endfeet (arrows) on a capillary (cap). (D) Silver-stained section showing cell bodies (arrows) and proximal processes of several oligodendrocytes

Ependymal cells, another type of glia, line the cerebral ventricles, regulating the interface between brain and CSF. In lizards, ascending processes of such cells may reach the surface of the brain and form endfeet here [69]. In many animal species, ependymal cells carry cilia [69, 70]. During the development of the brain, glial cells called “radial glia” are indispensable as they guide newborn neurons to their final location. The cerebellum hosts a specialized glial cell type called Bergmann radial glia (review in [71]), while tanycytes are a kind of radial glial cells associated with the third ventricle and the hypothalamus (review in [72]). Although it appears that glial cells do not directly participate in neural circuitry they are very important indeed as they regulate neural activity by conditioning (within very strict specifications) the environment in which the neurons themselves operate. Neurons are extremely sensitive to changes in their chemical environment and probably cannot properly function without being supported by glial cells.

Neurohistology

3.2

Neurons

3.2.1 Neurons Are Extremely Specialized Cells

13

Neurons generally possess several morphologically and functionally different regions closely related to function. Neurons, assisted by their own specialized peripheral or internal receptors, detect events, establish communication among themselves, adapt to information and store/retrieve information (learning and memorizing), and finally modulate activity in muscle fibers and secretory cells. Their dendrites are normally the site of most input from other neurons via synaptic contacts, although synapses may occur anywhere on a neuron, for example, on the cell body. The cell body is the locus of most of the genetic and metabolic activity of the neuron. The axon conveys information to other neurons, to muscle fibers, and to secretory cells. Communication between neurons generally occurs at synaptic terminals, which usually mediate contacts between axons and dendrites [50]. Synapses are specialized regions of contact between neurons where information is conveyed from one cell to another. The actual morphology of each of these functional regions varies considerably with neuronal type and many neuronal populations can be differentiated on the basis of their morphology (Fig. 2). Some specialized neurons are devoid of dendrites (e.g., sensory dorsal root ganglia neurons and olfactory receptor neurons) or axons (e.g., amacrine cells in the retina). Neurons are cells that contain the usual set of organelles commonly found in eukaryotic cells, such as plasma membrane, nucleus with chromatin and nucleolus, mitochondria, rough and smooth endoplasmic reticulum, Golgi apparatus, polysomes, microfilaments and microtubules, lysosomes, and storage and secretion vesicles [50]. In spite of this, neurons come in great many morphological and physiological varieties that are very specific in their morphology, relation to other cells, and activity. A major aspect of neuroscience is the study of interactions between specific groups of neurons. Neurons take many forms, depending upon their location and function, but there are certain features that are common to most types. An “archetypal” (model) neuron is schematically depicted in Fig. 2A. Typically, there is a centrally placed cell body hosting the cell nucleus and most of the metabolic machinery. The nucleus usually stains pale, because the chromosomes are largely dispersed to facilitate gene transcription. There is usually a prominent nucleolus because of the considerable amount of transcription activity in the nucleoplasm to produce mRNA necessary for regulating the production of proteins. The cytoplasm contains numerous aggregates of rough endoplasmic reticulum (RER), smooth endoplasmic reticulum (SER), and Golgi apparatus for the production and packaging of proteins. The aggregates of RER take characteristic forms in the light microscope in neurons stained to reveal nucleoprotein complexes. These light microscopically discernible clumps are called the Nissl substance (Fig. 2B). Not usually stained at the light microscopic level, but very important to the cell’s energy

14

Floris G. Wouterlood and Thomas P. Langer

Fig. 2 Neurons. (A) Scheme illustrating the functional regions of a neuron. The centrally located cell body hosts the metabolic molecular machinery; dendrites receive information via synapses with axon terminals of afferent axons (colored red), whereas the axon transmits information to other neurons, muscles or glands. (B) Nissl-stained motoneuron of the spinal cord, illustrating the main features of a neuronal cell body under a light microscope. A large nucleus contains a prominent nucleolus; stained material in the cytoplasm is clumped into the so-called Nissl substance extending a short distance into proximal dendrites. (C) Neurofibrils in silver-stained neurons of a kitten, as drawn by Ramo´n y Cajal ([23]—Fig. 54 on p. 180). The nucleus and the numerous intracytoplasmic organelles are suspended in a dense fibrillar mesh that may be microfilaments (structural elements) and microtubules, which (as we nowadays know) serve to transport vesicles containing material required in the neural processes and synapses

Neurohistology

15

supply are great numbers of mitochondria. In appropriately stained material, one can see that neurons have dense neurofilamentous webs, within the cell body and extending out into its processes, and arrays of microtubules filling the core of the axon and all of its branches (Fig. 2C). In early development neurons carry cilia, and in electron micrographs of adult brain, neurons with a cilium emerging from them can occasionally be encountered. Dendrites are tapering cytoplasmic processes that arise from the cell body and extend into the surrounding cellular matrix. Dendrites often extend radially from the cell body, for example, the spiny stellate cell type of the striatum, but in many neurons they form a special arbor intersecting with arrays of afferent axons. A prominent example of this category is the Purkynĕ cell of the cerebellar cortex, whose dendrites show extensive branching in “candelabra” style. The entire dendritic tree of a Purkynĕ cell is oriented strictly in one, sagittal, plane [73, 74] (see also Ref. [71]), and thus represents a more-or-less two-dimensional object [75]. The dendrites represent the part of the neuron that is usually specialized to receive information from many other neurons, through synapses between the afferents and the dendritic processes. In some neurons, such as pyramidal cells in the cerebral cortex, there are numerous small, lollypop-like appendages extending from the dendritic shafts, called dendritic spines (Fig. 4B). These structures are believed to be important in learning as animals raised in a cognitively enriched environment have more spines on the dendrites of their cortical pyramidal cells (review in [77]) while spine loss has been reported in brains of patients suffering from dementia (review in [78]). Most neurons emanate an axon, a process that differs from the dendrites by being of uniform caliber once it tapers rapidly from its attachment to the cell body or a proximal dendrite. The first, tapering, unmyelinated segment of any axon, that is, the initial axon segment or the axon hillock (Fig. 2A), is the site where the graded potentials in the dendritic and cell body plasma membranes are converted into a series of action potentials propagating down the axon [59]. In neurophysiology, this region is thus often referred to as the spike initiation zone. The axon tends to branch at approximately right angles, with both daughter branches being of the same size as the mother branch. Axons may ramify in the vicinity of the source neuron, elsewhere in the same nucleus or cortical region, and/or at a considerable distance from the cell body. The axon is specialized to support the propagation of action potentials, which are regenerative, all-or-none membrane potentials carrying the same information to all of the axonal terminals. Axons may show swellings on their shafts where they form synapses (boutons en passant or “buttons in passing”), or may end in terminal rosettes of synaptic boutons. Axons may specialize in several ways. They may spiral around a receiving (recipient) dendrite (pallidal “woolly”

16

Floris G. Wouterlood and Thomas P. Langer

fibers [79]), form climbing fibers as in the cerebellum [73–75, 80], clinch around target cell bodies (basket neurons [80]), form candlelike endings as in chandelier interneurons in cerebral cortex [81, 82], and of course there is the motor end plate at the periphery. As their name suggests, “interneurons” have axons that connect to neurons within the same nucleus, cortical layer or cortical area, while in laminated brain areas (e.g., cortex), they may connect as well to neurons in more superficial or deeper layers.

4

Classical Neurohistological Stains With Nissl or cell body stains it is possible to distinguish and subsequently categorize neuronal and glial cell bodies in white and gray matter. For instance, motor neurons tend to have large, densely stained, cell bodies, with large amounts of the Nissl substance and multiple cytoplasmic processes radiating away from the cell body. However, one can follow those processes for only a short distance from the cell body, because the Nissl substance does not extend far into dendrites, and not at all into axons. Glial cells usually stain intensely with the Nissl stain, yet can be easily distinguished from neurons and interneurons as their Nissl substance is very compact indeed (compare oligodendrocytes versus Nissl-stained Purkynĕ cells, interneurons and the tiny and extremely densely packed “granule” cells in Fig. 3B). Clusters of neuronal cell bodies in subcortical areas of the brain are called nuclei. Distributions of neuronal cell bodies in three to six laminae at the surface of the brain are called “cortex” (for instance hippocampal cortex, cerebellar cortex and cerebral cortex) while mixtures of neuronal cell bodies and loosely organized fiber bundles are referred to as reticular formation. A general modern distinction of cerebral cortex based on lamination is neocortex (six-layered), paleocortex (three-layered), and archicortex (hippocampus; three-layered). Subcategories of neocortex distinguished by cytoarchitectonic features are agranular cortex, dysgranular cortex and granular cortices [83]. During the heyday of descriptive cerebral cortical cytoarchitectonics, von Economo and Koskinas [84, 85] distinguished five types of cortex as well as 107 modifications thereof.

4.1 Two Protocols to Stain Neurons 4.1.1 Fast Nissl Stain

Here follows a recipe for a fast and effective Nissl stain on 40 μm thick sections (cut with a freezing-sliding microtome) of 4% phosphate-buffered formalin-fixed brain. The procedure is used in our laboratory to stain normal series or to counterstain immunohistochemically stained sections. Prior to the staining procedure the sections should be mounted on gelatinized glass slides and dried.

Neurohistology

17

Fig. 3 Nissl-stained human and rat cerebellum. Sections cut in the sagittal plane of frozen, formalin-fixed brain. GL granular cell layer, ML molecular layer, WM white matter, in interneuron, P Purkynĕ cell, ∗ Oligodendrocyte. (A) Low magnification micrograph of a human cerebellar lobule. The cerebellum is typically three layered (ML, GL and WM, from the bottom of the fissure). The Purkynĕ cells (P) lie deep in the molecular layer and are regularly spaced (like pearls) at the ML-GL interface. The granule cells are small, densely packed and stain overwhelmingly in Nissl, giving the cerebellum its characteristic intensely blue layered appearance with this stain. (B) High magnification of part of the boxed area in a, showing the Purkynĕ cells (P) at the ML-GL interface and the granule cells in GL. Oligodendrocytes (∗) are compact, intensely stained cells that, with some experience, can be distinguished from less intensely stained interneurons (some of them marked “in”). Inset: Single, completely Golgi-silver–impregnated Purkynĕ cell in rat cerebellum. (C) Human cerebellum, immunostained with antibodies against parvalbumin (brown) and Nissl counterstaining (blue). Note that the Purkynĕ cells (P) and their dendrites are strongly parvalbuminimmunopositive

18

Floris G. Wouterlood and Thomas P. Langer

Staining solution: 90 ml Acetic acid (15 drops of 35% acetic acid in 90 ml aqua dest.) 10 ml Sodium acetate (dissolve 0.14 g CH3COONa in 10 ml aqua dest.) 15 ml Cresyl violet (1% in aqua dest.)

The staining solution is acidic (pH 3.7). Procedure (some experimentation is recommended to optimize for specific brain tissue): 1. Rinse slides in aqua dest. 2. Place slides in the staining solution (room temp.) for 60 s. 3. Rinse rapidly in three changes of aqua dest. 4. Differentiate in acidified aqua dest. (few drops of glacial acetic acid). 5. Continue differentiation in 30% ethanol in aqua dest. 6. Dehydrate through 50–70–96% ethanol to 100% ethanol (in the 96% ethanol step, a supplemental dehydration can be performed in 96% ethanol, with a few drops of 35% acetic acid added). 7. Rinse in several changes of xylene, embed and coverslip. An example is shown in Fig. 3C: human cerebellum, frozen section, immunostained with antibodies against the calcium binding protein, parvalbumin, sections counterstained with fast Nissl staining. 4.1.2 Golgi-Kopsch Silver Impregnation of Individual Neurons

Before the advent of intracellular injection techniques and single neuron visualization in brains of genetically engineered animals, the only way to make individual neurons fully visible was to stain with one of the variations of Golgi-silver staining. Problematic with all Golgi-silver stains is their unpredictability with respect to the types of neuron that become completely filled with the black precipitate (identified as silver chromate [86]), and the number of neurons thus filled. Most frequently used are variations of the rapid Golgi staining, using immersion first in potassium dichromate and next in silver nitrate. One of the relatively predictable types of Golgi-silver staining, the Golgi-Kopsch variation, was introduced by del Rio-Hortega in 1928 [87]. The modification of the Golgi-Kopsch procedure presented here was designed to stain complete cerebral hemispheres of rat brain; it can also be applied to blocks of brain tissue with a volume of about 10
10
10 mm (e.g., small blocks of human cerebral cortex). Material needs to be fixed in 4% phosphate-buffered formalin. For EM an intermediate treatment of 5 h in a refrigerator in 2%

Neurohistology

19

cacodylate-buffered OsO4 is recommended. Blocks may be coated prior to chromation with 2% agar to prevent buildup of a crust of silver chromate crystals at the surface of the blocks or hemispheres. Staining procedure: 1. Chromation Stay in 3% potassium dichromate for 3–4 days (dark, room temp). Some authors prefer to add 2% chloral hydrate to the chromation solution [88]. 2. Impregnation Stay in 0.75% silver nitrate for 24 h to 3 days (dark, room temp.). After impregnation the blocks are cut into slabs 80–100 μm thick with a vibrating microtome. To prevent loss of precipitate the bathing fluid (50% ethanol) should be saturated with silver chromate. After cutting and inspection slabs are post-fixed in 2% aqueous OsO4 (1 h, refrigerator, dark), then infiltrated with plastic (e.g., Araldite, an EM procedure) and flat-embedded between two pieces of polyethylene foil, with a small weight placed during curing on top of the sandwich to prevent curling. As Golgi-silver preparations show the tendency to deteriorate in time and under exposure to daylight, several procedures have been developed to stabilize the silver chromate precipitate, for example, via gold toning [89] or treatment with a photographic developer [90]. Counterstaining with a Nissl staining procedure is possible [76] (Fig. 4D). The Purkynĕ cell in Fig. 9 illustrating correlative light-electron microscopy (CLEM) was Golgi-silver– impregnated according to the above procedure. 4.2

Staining Myelin

4.2.1 Luxol Fast Blue Stain (LFB)

Myelin sheaths are very rich in fats. In the second half of the nineteenth century, stains were discovered that selectively stain myelin. The first of these is attributed to the brilliant pioneer Carl Weigert [19] (see also Ref. [91]), who recognized the magic of the new industrial textile stains that became available in nineteenth century Germany. Weigert’s procedure required weeks to months of preprocessing in a mordant to degrade the lipids, and was quite tedious [92]. Compatibility of the Weigert stain with other neurohistological stains was rather limited. Heidenhain [20] is acknowledged for his iron hematoxylin staining that, with some modifications, appeared to be useful for staining myelinated tracts [92, 93]. This stain also requires pretreatment in mordant solution. In either case, the tissue gets severely etched. The Luxol Fast Blue stain (LFB) developed by Klu¨ver and Barrera [94] is still popular because it is a nonetching, rapid procedure that brilliantly stains both CNS neurons and myelinated fibers. The procedure listed below is designed to be used with 40 μm thick

20

Floris G. Wouterlood and Thomas P. Langer

Fig. 4 Golgi-silver–stained rat and human neurons. (A) Rat hippocampal CA1 pyramidal neuron. The cell bodies of CA1 pyramidals are located in stratum pyramidale (SP). An extremely long apical dendrite ascends from the cell body approximately 200 μm through stratum radiatum (SR) into stratum lacunosum moleculare (SLM). (B) Dendritic spines (arrows) are visible at high magnification on the dendrites of this neuron. (C) Golgi-silver–impregnated pyramidal and stellate neurons intermingle in human prefrontal cerebral cortex (material courtesy of Dr. Harry Uylings). (D) Pyramidal neuron in rat parietal cerebral cortex, section Nissl-counterstained (procedure in Ref. [76]). (E, F) Multipolar stellate cells in the deep layers of rat superior colliculus

sections of 4% phosphate-buffered formalin–fixed brain cut with a freezing-sliding microtome. After curing, sections are mounted on gelatinized glass slides and thoroughly dried (~60 h in an oven at 60  C). Staining solutions: Luxol Fast Blue (LFB): 0.5 g in 500 ml ethanol (96%), add 2.5 ml of acetic acid (35%). Lithium carbonate (Li2CO3): 0.5 g in 500 ml aqua dest. Cresyl violet: 0.5 g in 500 ml aqua dest; add 2.5 ml of acetic acid (35%).

Neurohistology

21

Procedure: 1. Stain 24 h in the LFB solution. 2. Rinse in 96% ethanol in aqua dest. 3. Differentiate in two changes of Li2CO3 solution 4. Rinse in two changes of ethanol (70%) in aqua dest. 5. Rinse twice in aqua dest. 6. Stain 1–10 min in warm (60  C) cresyl violet solution 7. Rinse in 96% ethanol in aqua dest. 8. Transfer to fresh ethanol (100%, should be water free). 9. Transfer to xylene, mount and coverslip. 4.2.2 Degenerating Myelin

5

Myelin and its staining properties change as axons start to degenerate once severed from their parent cell body. Marchi and Algeri [43] exploited in 1893 this particular feature to develop a specific stain. A different set of stains introduced 30 years later, of which the most famous is the Weil stain [92, 93], became very useful to reveal the distribution of degenerating myelin following CNS lesions. These stains were the main tools for tracing connections in the CNS until the 1960s. The classical silver stains were modified in the 1950s and 1960s to make them more specific to degenerating axoplasm. The best known examples are the Nauta and Fink-Heimer silver stains [51– 53]. These stains made it possible to follow a degenerating axon all the way to its terminus. As they can be used for unmyelinated axons as well they represented a substantial improvement of the myelin stains, because myelin usually ends some distance before the synaptic terminal, and unmyelinated parts of axons are thus invisible in myelin-stained sections. With the Fink-Heimer method [53] a punctate staining occurs at sites where axons terminate. These punctae were identified by EM as degenerating axon terminals [95].

Living Tissue as a Tool: Transport Tracing In the 1970s, neurohistology changed in the sense that experimental biomedical techniques became further integrated in neurohistology. Tract degeneration following an experimental lesion had been studied for over 80 years [43, 44], but the really big innovation was the introduction in the second half of the twentieth century of much more sensitive methods relying on the uptake and transport of molecules by living neurons, and subsequent metabolic processing/packaging of these molecules, followed by intracellular transport. Two broad types of transport are distinguished in neurons, based on the direction the material is transported: anterograde transport (centrifugal, or uptake at the cell

22

Floris G. Wouterlood and Thomas P. Langer

body and transport into the axon) and retrograde transport (centripetal, or uptake at the axon terminal and transport back to the cell body). Functionally, there are at least three kinds of intra-axonal transport, driven by so-called molecular motors such as kinesin (involved in anterograde transport [96]) and dynein (involved in retrograde transport [97]) stepping along microtubules. Fast anterograde axoplasmic transport moves material at millimeters to centimeters a day. Slow anterograde axoplasmic transport is more of a bulk flow of cargo necessary for the structural renewal of the neuron’s organelles. Fast retrograde axoplasmic transport is directed toward the cell body and mediated by vesicles being dragged along microtubules like tiny funicular cabins. It has a speed comparable to that of fast anterograde transport. Following neurosurgical procedures a small volume of tracer substance is deposited in a stereotaxically determined position in the CNS. In the PNS, the tracer substance can be applied to a cut nerve or simply injected into an organ innervated by a peripheral nerve [98]. After a certain period allowing for uptake and transport to occur, the experimental animal is sacrificed, the brain is recovered, sections are cut, and neurohistological procedures are followed. Since the 1970s, a plethora of macromolecules and fluorescent dyes that are subject to retrograde transport has been documented, starting with horseradish peroxidase (HRP), introduced as a tracer by Kristensson and Olsson [98]. As HRP is not destroyed en route and catalyzes a colorimetric reaction, it can be readily visualized by soaking sections of the brain in a solution of peroxidase substrate, yielding a precipitating product [99]. As the sensitivity of HRP detection procedures increased over the years, it became apparent that the enzyme is transported anterogradely as well and that it is taken up at the sites of injection by damaged fibers of passage (review in [100]). Pioneering work in anterograde tracing was done by Droz and Leblond [101] who described uptake of radioactive amino acids by neurons in the CNS, incorporation in proteins, subsequent transport and, finally, detection by radioautography. Singer and Salpeter [102] reported the use of [3H]-labeled L-histidine as a tracer in the PNS. However, it was Cowan et al. [103] who forged Droz and Leblond’s procedures into a robust and reliable neuroanatomical tracing technique for use in the CNS. Briefly, the experimental animal receives an intracerebral injection of radioactive amino acids. The amino acids are taken up by neurons and metabolized in their cell bodies into structural proteins that are then anterogradely transported along the axons, ultimately to become incorporated in their membranes and accumulated at synaptic terminals. Few days later, the animal is sacrificed, its brain recovered and sectioned. The sections are mounted on slides, subsequently coated with a thin layer of photographic emulsion, and stored in the

Neurohistology

23

dark for a few weeks to allow β particles (electrons) produced by radioactive isotope decay in the labeled proteins to register in the emulsion. The photographic emulsion is developed like a photographic film, and the labeled proteins “colocalize” with the silver grains in the emulsion overlying the section. The grains are visible under a microscope, ideally fitted with a darkfield condenser. In this way, one can label the entire axon arbor from cell body to axon terminals (Fig. 5A). The resolution of radioautography is so high that it has been also applied to electron microscopic preparations, to study intracellular transport [104, 105]. Reviews of these second-generation neuroanatomical tracing methods can be found in [100, 106]. Modern anterograde tracing methods employ lectins or biotinylated compounds as a tracer rather than radioactive amino acids because the nonradioactive techniques are far more simple, compatible with immunohistochemistry (Fig. 5B) and neurophysiology, and do not require special containment and waste disposal measures [54, 100, 103]. Modern retrograde tracing employs fluorescent compounds (Fig. 5D), fluorescent nanoparticles, viruses or bacterial toxins [100].

6

Three-Dimensional Reconstruction Versus Whole-Mount Neurohistology The intrinsically three-dimensional (3D) architecture of neurons and their networks enticed scientists already in Ramo´n y Cajal’s time to cut thick sections in order to render dendritic tree configurations visible as completely as possible. There are two approaches to study 3D aspects of neurons: top down or bottom-up.

6.1 Top-Down Approach

Whole-mount preparations reveal single organs or structures in a preparation made transparent with an optical tissue clearing (OTC) technique. Because Golgi-silver staining works best with tissue blocs or complete brains of small animals, and the 3D organization of the neurons is best visible in thick sections, this staining can be considered as a whole-mount technique in its own right. A wholemount staining technique designed to selectively stain PNS nerves while rendering the surrounding tissue transparent is based on Sihler’s stain [107–109]. In this procedure, the sacrificed animal is stripped of all its skin and immersed in potassium hydroxide to provide a maceration and depigmentation environment. Subsequently, the specimen is decalcified, stained with Ehrlich’s hematoxylin, differentiated, neutralized and cleared. Clearing is achieved in glycerin, and with a biocide agent added, the preparations can be indefinitely stored in glycerin in a glass jar. Recently, embedding in polyester resin has been proposed [110]. The OTC methods were pioneered by a German anatomist named Werner Spalteholz [111] who cleared biological material in methyl salicylate and benzyl

24

Floris G. Wouterlood and Thomas P. Langer

Fig. 5 Connectivity studies. (A) Anterograde tracing: radioautography and darkfield microscopy. Labeled axons in a tract identified by elongated clusters (arrows) of silver grains. Terminal labeling is evident by the individual white, scattered silver grains produced by reduction in photographic emulsion of silver bromide by β-particles (electrons) emanating from radioactive isotopes (see Ref. [103] for details) and visualized by photographic developer. (B) Anterograde tracing: dense network of labeled fibers and terminals (arrows) in nucleus accumbens (Acb) in a rat after injection of a lectin (leucoagglutinin) from Phaseolus vulgaris in the amygdala. Labeled fibers detected by immunohistochemistry with a fluorochromated secondary antibody. AC anterior commissure, CPu caudate-putamen, LV lateral ventricle, VP ventral pallidum. Inset: High magnification image obtained by confocal laser scanning microscopy of a single lectin-labeled fiber forming multiple boutons en passant (a) and boutons terminaux (b). (C) Retrograde tracing with horseradish peroxidase (HRP). A retrogradely labeled neuron containing HRP histochemistry reaction product [98, 99] in its soma (S) and dendrites (D), and even in the axon (Ax). (D) Retrograde tracing in the rat with a fluorescent dye, Fluoro-Gold injected in the caudateputamen. Labeled cell bodies are located in substantia nigra pars compacta (SNc) and adjacent ventral tegmental area (VTA). Fluoro-Gold fills cell bodies and main dendrites. For details of this type of tracing, see Ref. [100]

Neurohistology

25

benzoate. For the theory behind OTC and the comparison of various popular OTC methods we refer the reader to reviews [112, 113]. Quite another line of the whole-mount approach is to combine molecular phenotyping (i.e., forcing specific neurons by genetic engineering to express green fluorescent protein [GFP]) while rendering parts or complete brains fully transparent (e.g., in embryos or in mouse brain). While classical OTC techniques such as Sihler’s [107] glycerin and Spalteholz’s [111] methyl salicylate work quite well, modern optical imaging methods typically require clearing of tissue by refractive index matching to ca. 1.4–1.5. This is further complicated if the objective lens has to be in contact with the immersion fluid as that can seriously impair image quality unless special adapters [33] or costly lenses such as Clarity (Leica, n ¼ 1.46) or Scaleview (Olympus, n ¼ 1.38) are used. Such lenses are tuned to the refractive index of the clearing solution. A recently developed OTC method specifically designed to image neurons in genetically engineered mouse brain is CLARITY designed by the Deisseroth group at Stanford [114]. “Scale” is yet another modern technique for making complete brains optically transparent [115, 116]. Although the specimen becomes fully transparent through OTC treatment the imaging remains problematic not so much for the refractive properties of the brain made transparent, but because of the limited working distance even of special (extremely long working distance) objectives. A special rod-shaped gradient refractive index (GRIN) lens inserted into the living brain represents a novel microendoscopy solution circumventing this problem [117]. The consequence of lipid removal (inherent to the OTC techniques) is that dyes that bind to lipid components of myelin are extracted as well, among them DiI, DiO (these are very strongly lipophilic compounds used to “trace” myelinated tracts in fixed brain) and similar compounds. Jensen and Berg [118] introduced fixable, CLARITY compatible, lipid extraction resistant dyes that can be used in neurophysiological electrode marking and also in passive neuroanatomical tracing. 6.2 Bottom-Up Approach

3D reconstruction of organs from serial sections flourished already at the onset of the twentieth century when embryologists reconstructed developing organs in embryos [119, 120]. All the work was done by hand and consisted of drawing on cardboard the outlines of embryonic organs with the aid of a drawing tube (“camera lucida”) attached to the microscope, cutting the cardboard along the outlines, and stacking the resulting cardboard “sections” on top of each other by pins, while keeping orientation and outline matching (registration) as good as possible. Bee wax was smeared onto this heap to produce a smooth surface of the 3D reconstructed organ and voila`! Beautiful examples can be admired in

26

Floris G. Wouterlood and Thomas P. Langer

science museums all over the world. An alternative to bee wax was modeling clay [121]. The advent of the computer and integration of computers and microscope systems in the final decades of the twentieth century changed all this, of course. Neurohistological image processing emerged. One early example is a paper published in 1983 proposing a software package for the Apple II plus microcomputer, to enable 3D reconstruction of serial sections [122]. A typical example of the bottom-up approach is confocal microscopy. The confocal principle is quite old: Hiroto Naora published in 1951 a cytometric research paper conducted with a microdensitometer that has all characteristics of a confocal microscope: spot illumination, pinhole, and photomultiplier detector [123]. An American mathematician, Marvin Minsky, applied in 1957 for a patent on the principle of confocal microscopy which he was awarded in 1961 (US Patent Number 3.013.467) [124]. However, microscopes exploiting this principle were hard to build with 1950s technology. After a period with experimental devices [125, 126] commercial confocal instruments became available in the 1980s, mostly equipped with lasers (confocal laser scanning microscopy [CLSM], or “Nipkow” spinning-disks that can operate without laser). By far most imaging and 3D reconstruction in CLSM is conducted with tissue including GFP expressing cells or with brain sections stained either through immunofluorescent procedures or directly with fluorescent dyes. 3D reconstruction software is nowadays commonplace in CLSM. A raw confocal image is inherently strictly two dimensional, in the sense that it represents (more-or-less) an “optical section” of the specimen. Confocal imaging can be carried out on tissue sections, organs or even in vivo on living brain via a cranial window [127]. By moving the objective or the stage stepwise in the axial direction a series (the so-called stack) of “optical sections” is acquired that, with the aid of proper 3D reconstruction software will reveal complete, spatially complex structures with all their processes and appendages. In fact, a similar image acquisition scheme as in CLSM is followed in magnetic resonance imaging (MRI) (e.g., high-resolution MRI scanning of mouse brain) [128]. In order to function perfectly, that is, to excite fluorescent molecules located deep within the tissue and to detect emitted photons that have to travel unabsorbed by tissue all the way “back” to the detector, image formation in CLSM heavily relies on highly optically transparent tissue. The OTC procedures are thus of utmost importance in CLSM [129]. High-numerical aperture, immersion objective lenses, two-photon microscopy, and adaptive optics are instrumental in addressing this issue, as briefly outlined in Subheading 2.4. Aside from image format issues, the software that renders on computer screen 3D structures from MRI data sets works equally well with CLSM data sets. An example of multipurpose, open-

Neurohistology

27

source image processing, 3D reconstruction software is ImageJ [130]. In other words and not surprisingly, confocal microscopy is suited to serve the top-down approach with a bottom-up instrument design, and owing to computer software. Examples of CLSM-3D reconstruction imaging are presented in Figs. 6, 7 and 8A. Figure 6 shows a mouse hippocampal CA1 neuron expressing GFP. Panel a in Fig. 6 shows an image obtained

Fig. 6 Bottom-up approach using 3D computer reconstruction from images acquired by confocal laser scanning microscopy. SLM stratum lacunosum moleculare, SR stratum radiatum, SP stratum pyramidale, SO stratum oriens. (A) Hippocampal GFP-expressing CA1 pyramidal neuron (GFP knock-in mouse 6296; material courtesy of Dr. Christiaan Levelt). The neuron was imaged with a water immersion
40/1.2 objective. Summed Z-projection (all images in the Z stack “flattened” into a single image). (B) The same neuron scanned at high magnification (
63/1.3 glycerin immersion objective) in four color-coded stacks, and then subjected to computer 3D reconstruction. On screen, the composite 3D reconstruction can be rotated and viewed from any direction. The white box denotes the boundaries of the SP region (“green” image stack). Compare with the Golgi-silver–impregnated CA1 neuron in Fig. 4A, and AF555injected CA1 pyramidal neuron in Fig. 8A

28

Floris G. Wouterlood and Thomas P. Langer

by “flattening,” or “summed” Z-projection of an image stack acquired at low magnification (20
) by CLSM. This is a modern solution to manually drawing neurons using “camera lucida,” while refocusing (also manually) throughout the specimen thickness. Panel b is a screen capture of one out of many frames during on-screen rotation of a 3D image (reconstruction) of the same neuron. To obtain the fine details shown in Fig. 6B, the cell body, dendritic branches and axon collaterals were scanned tile-wise with a high magnification (
63) immersion objective, yielding a total of four image stacks (color-coded in Fig. 6B) that were subsequently combined into the final 3D montage. Because the Golgi-silver

Fig. 7 Golgi-silver–impregnated human prefrontal cortex in a reflectance mode of confocal laser scanning microscopy (a type of dark-field imaging detecting light that is backscattered by the silver [chromate] particles). Material is courtesy of Dr. Harry Uylings. (A) Summed Z-projection of the image stack (orientation 0 ). cap capillary, Pyr pyramidal neuron, Stell stellate neuron. The cortical surface is located to the left. (B–D) Individual frames (slightly squeezed vertically) of a 360 rotation movie made after 3D computer rendering of the image stack that also gave rise to frame A. Orientation of B and D is 0 and 180 , respectively

Neurohistology

29

Fig. 8 Bottom-up 3D reconstruction of a rat hippocampal CA1 neuron intracellularly injected with the fluorescent dye AF555 in a slice of lightly fixed brain (see Ref. [131] for details). (A) Confocal laser scanning microscopy at high magnification followed up with 3D computer reconstruction. Inset: the main dendrite scaled 200% to show the spines (arrowheads). Material is courtesy Dr. Riichi Kajiwara. SLM stratum lacunosum-moleculare, SR stratum radiatum, SP stratum pyramidale, SO stratum oriens. (B) Example of “neurons themselves as tools.” Cholinergic neurons in a genetically engineered ChAT-IRES-Cre mouse basal forebrain, expressing GFP (green) after a focal transfection with an AAV2 adenovirus carrying a plasmid coding for GFP. NeuN-expressing neurons (red) are stained with an antibody against NeuN and a secondary antibody with a red fluorophore (see Ref. [54] for details). (C) Gap junctions: Electron micrograph of a mixed axon terminal (AT) in the accessory abducens nucleus of the lizard Varanus exanthematicus: synaptic vesicles converge onto a synapse (arrow) while there is also a gap junction (GJ). The lower panel shows a typical close apposition of the gap junction’s membranes of the pre- and postsynaptic (Post) neuron, with a 2 nm cleft separating them. Mit mitochondrion. See Ref. [132] for details

30

Floris G. Wouterlood and Thomas P. Langer

staining technique has a prominent place in this chapter we include a 3D reconstruction of a small piece of Golgi-silver–stained human prefrontal cortex imaged at low magnification in the so-called reflectance mode (Fig. 7). Modern technology applied to century-old Golgi-silver–stained material! Figure 8A depicts a flattened 3D reconstruction of a rat hippocampal CA1 pyramidal neuron intracellularly injected in a slice of lightly fixed brain with the fluorescent dye AF555, and scanned at high magnification in 65 overlapping optical sections. Unlike in Fig. 6A, only a single image stack was acquired. Parallel to the confocal (bottom-up) approach, technical improvements in EM neurohistology and software and image processing make it nowadays possible to reconstruct neurons in 3D at the EM level in a semi-automated way. Pivotal in this field was the introduction in biomedical research of instruments combining cyclic focused ion beam milling of plastic-embedded CNS material followed by scanning EM of the abraded block surface (FIB-SEM) [133]. The milling-scanning cycle can be fully automated, and acquired digital images can be processed with advanced 3D reconstruction software. Instruments like these and the mechanical milling variant, the automatic tape-collecting ultramicrotome scanning electron microscope (ATUM-SEM) are enormously accelerating 3D reconstruction at the ultrastructural level [134]. FIB-SEM and ATUM-SEM studies can be considered as typical large-scale bottom-up approaches. 6.3 Morphological Variability of Neurons

Golgi, GFP expression, and peri-/intracellular neurophysiology studies show that the morphology of neurons varies in characteristic ways in the nervous system. With a little practice one can readily distinguish mitral cells in the olfactory bulb, pyramidal cells in the cerebral cortex, granule cells in the dentate gyrus of the hippocampus, cerebellar Purkynĕ cells, and motoneurons in brainstem cranial nerve nuclei. Cerebellar cortex is a wonderful example as it contains a rich variety of neuron types, each restricted to a particular lamina (Fig. 3) and connected in a consistent way to the other neurons [73–75]. Space in this chapter is too limited to describe the morphological details of all types of neuron that have been hitherto distinguished. Neuronal type classification is nowadays a topical issue as neuron types distinguished purely on morphological criteria only partially match neuron types distinguished with neurophysiological or chemical tools [135] (see also Ref. [136]). Especially interneurons exhibit an enormous variability in neurochemical and neurophysiological features [82].

Neurohistology

7

31

Living Neurons as Tools to Study Themselves Traditionally, neurohistology deals with the morphology of dead neural tissue. The point here is that although fixation preserves the tissue in a state that is morphologically only slightly different from that in vivo, the cells do not express DNA any more, nor do they metabolize or exert any (electrical) activity. Molecular machinery and activity have come to a standstill. The unique functions of neurons in information processing are no longer available for examination, and the situation resembles inspecting a piece of complex electronic equipment after being disconnected from its power supply, or at least having software instructions removed. The transport methods of the 1970s can be perceived in retrospect as a prelude to the situation today where neurohistology increasingly blends with physiology, molecular biology and genetics to the extent that we are optogenetically triggering receptors in action (reviews in [137–139]), tagging molecules with fluorescent markers, staining specific organelles with antibodies that recognize specific components, and analyzing changes that occur when a neuron is damaged. Thus, in the twenty-first century the neuron itself has become the neurohistologist’s tool. All these actions provide markers for tracing connections and information processing. Exciting new technologies have been developed that allow us to use the genetic machinery of living neurons. The immediate early gene expression technique monitoring mRNA levels is not listed here, while selective fluorochrome expression (“cre-lox”) in genetically engineered mice is amply described elsewhere in the present book, and is illustrated here with the aid of Fig. 8B). One very recent approach to identify synaptic contacts between neurons is the GRASP technique (GFP fluorescence reconstitution across synapses) [140]. This method was originally brought to fruition in invertebrate species, Caenorhabditis elegans [141] and Drosophila melanogaster [142], and relies on the production in preand postsynaptic neurons of nonfluorescent protein fragments that, when they meet at synapses, assemble into YFP (yellow fluorescent protein) or other color-shifted variants of GFP [143] that fluoresce under UV illumination. Yamagata and Sanes [144] succeeded in adapting the method to transgenic mice. They successfully produced a double-transgenic mouse (mGRASP1-rhodopsin-C) and reported rod photoreceptors and horizontal cells generating the GRASP signal in retina, at sites where these cell types closely interact. Lee et al. [134] extensively reviewed variants of fragment-fusion synaptic connectivity identification methods.

32

8

Floris G. Wouterlood and Thomas P. Langer

High Energy Requirements of Neurons: Structural Maintenance Neurons have very high-energy requirements causing the CNS to consume about an order of magnitude more energy carriers (ATP) and oxygen per kilogram than the body as a whole [59]. Much of that energy goes into the maintenance of ion pumps and channels necessary to conduct neurotransmission (“firing and reloading”), whereas the rest is spent on the maintenance of large dendritic and axonal arbors, and the continuous secretion and reuptake of neurotransmitters at neurons’ many synapses (shipping provisions and waste). The high metabolic rate in neurons is manifested by dispersed chromosomes and a prominent nucleolus in the nucleus, extensive rough and smooth endoplasmic reticulum, and in large numbers of mitochondria. While one can encounter occasional clumps of rough endoplasmic reticulum in dendrites it is virtually absent from axons. Consequently, neurons require constant synthesis and transport of proteins and ATP from their cell body into the axon to maintain synapses structurally and in terms of their membrane potentials and synaptic activity. Even a brief burst of activity, in the order of seconds, will evoke a quick increase in metabolic activity and an increased demand for energy carriers. The high metabolic requirement of active neurons has been used in a number of techniques that detect the flow of glucose and/or oxygen to brain tissue. Indeed, diagnostic applications of functional MRI (fMRI) and positron emission tomography (PET) are based on this particular feature of the nervous system [145]. Before this technology became available, 2-deoxy-D[1-14C]glucose (2DG), a glucose analog that is taken up by neurons but not metabolized, was given to an animal that was performing a specific behavioral task [146]. 2DG accumulates most densely in neurons that are most active. Recently, it has been shown that more active neurons trigger activation of the surrounding astrocytes, which changes the caliber of local capillaries and increases blood flow to the activated region [147]. We can also recognize the more active neurons because they have increased ionic flows that can be visualized with fluorescent dyes sensitive to specific ions (voltage-sensitive dyes) whose biological applications were first described by Hartline [148]). The constant rebuilding of the neuron involves the integration of proteins into its membranes and organelles, and was explored as early as in the 1960s [101]. In the hands of Singer and Salpeter [102], and Cowan et al. [103], the technique matured into a standard neuroanatomical tracing tool in the PNS and CNS, respectively (see Subheading 5). An alternative to structural labeling is to inject radioactive nucleic acids into a pregnant animal (mother) so that the

Neurohistology

33

developing embryo or fetus incorporates radioactivity into its replicating chromosomes [149]. If the labeled cell is at its terminal mitosis, before differentiating into a neuron or glial cell, then the amount of incorporated label is maximal and the label remains in the chromosomes. Preparation of the brains of the labeled animals is the same as in radioautographic pathway tracing, and eventually makes it possible to identify the cells that stopped dividing (and started to differentiate into neurons or glia) at the time of the injection, called the cell’s “birthday” (hence “birthday studies”). In this way, one can determine the chronological order in which individual neurons or glia are “born,” and how they migrate to their final location in the brain. Today, cell migration studies in living brain are carried out using genetic techniques; for instance with the aid of a genetically engineered, photoconvertible protein marker Kaede [150], or with gene plasmids (introduced by electroporation) that force cells in early embryos to express GFP [151, 152].

9

Electrical Activity, Channels, Receptors, and Neurotransmitters Neuronal activity strongly depends on a continuous flow of ions through plasma membranes, especially at synaptic contacts, spike initiation zones and in the nodes of Ranvier. In regions with high ionic flows one can usually see in the EM a membrane thickening due to dense accumulations of receptors and/or channels in and near the membrane.

9.1

Ion Channels

Voltage-dependent ion channels (VDICs) dictate which ionic species is allowed to pass the neuron’s limiting membrane, how often such an event may occur, and under what conditions. At the presynaptic terminal, VDICs play a critical role in regulating nerve terminal excitability [153]. As sodium-, potassium-, and calcium-dependent VDICs exist, each composed of different subunits and auxiliary units, an enormous variety of VDIC-associated proteins and mRNAs has been documented [153]. VDICs per se are structures too small to be seen with an ordinary light microscope, but total internal reflection fluorescence (TIRF) microscopy capable of optical “dissection” has localized with 1000 nm precision single YFP-tagged VDICs, with the aid of monoclonal antibodies raised against their subunits [154]. A similar approach is employed in the localization-based super-resolution microscopy techniques [28–30]. Receptor proteins have been pharmacologically localized in light and electron microscopy using irreversible binding with a labeled ligand, be it an enzymatic catalyst or a radioactive compound that can be used in radioautography. Physiological manipulation of neurons is conducted through optogenetics, that is, neurons in genetically engineered organisms

34

Floris G. Wouterlood and Thomas P. Langer

express light-sensitive ion channels that can be manipulated (e.g., “switched-off” by short-circuiting them) with light [137, 155, 156]. 9.2 Electrotonic Coupling

A peculiar type of a channel aggregate is the gap junction. Gap junctions appear in EM images as sites where pre-and postjunctional membranes get very close to each other, leaving a gap of about 2 nm (Fig. 8C) [157, 158]; see also Ref. [50]. Functionally, gap junctions enable exchange of ions between neurons and are thus regarded as sites where electrotonic coupling of neurons takes place. Originally discovered in crayfish and in electromotor neurons of mormyrid fish [50] (hence called “electrotonic synapses”), gap junctions have been reported in various parts of mammalian brains, (e.g., in the inferior olive [159], dorsal cochlear nucleus [160], retina [161], and even cerebral cortex [162, 163]). Astrocytes have also been shown to be extensively coupled to each other through gap junctions [66]. The channels in gap junctions are made up of subunits of a specific class of protein called connexins. Today, about 20 different connexins have been identified [164], of which connexin 36 is present in mammalian neurons [165]. Gap junctions may occur isolated, for example, as dendrodendritic junctions, but they are also found in mixed axon terminals, adjacent to chemical synapses (Fig. 8C) [132, 166].

9.3 Neurotransmitters

Localizing neurotransmitters (NTs) is a challenge running parallel with research into synapses. As NTs are molecules they cannot be defined as “structures,” so strictly speaking they fall outside the realm of neurohistology. Yet NTs are so important for neuronal action and functionally so intimately related with particular neuronal types that they need to be mentioned here. Small amino acid NTs (e.g., glutamate or glycine) have a strong tendency to diffuse out of brain sections during or after neurohistological fixation. Some of the larger neurotransmitters and especially neuromodulators and neurohormones are oligopeptides that may be used to raise monoclonal antibodies. If they are fixed in situ, one may be able to identify the synapses releasing a particular neurotransmitter. The first antibodies to localize NTs immunohistochemically were raised around 1978 against a neurotransmitter fixed to a substrate [165]. During a period of 20 years before that the location of catecholamine neurotransmitters (dopamine, noradrenaline, adrenaline, serotonin, and histamine) was demonstrated by exposing brain sections to formaldehyde vapors that induced fluorescence (the so-called FIF technique) [167, 168]. The FIF technique thus revealed detailed information about the distribution of neurons expressing this class of NT. However, the compatibility of FIF with neurohistological stains was rather poor. The introduction of antibodies against NT-substrate complexes [169], at approximately the same time

Neurohistology

35

when antibodies became available against cytosolic enzymes metabolizing NTs, brought about the demise of the FIF technique. A way around the difficulty to immunohistochemically pinpoint putative NTs in sections is to apply antibodies against transporter proteins located in the walls of synaptic vesicles. These proteins, aptly named “vesicular transporters” (VTs), are carriers that bind neurotransmitter molecules from the neurotransmitter pool and carry them across the synaptic vesicle membrane into the interior of the synaptic vesicle [170]. VTs have been immunohistochemically identified to be involved in glutamatergic [171, 172], GABAergic [173], cholinergic [174], and monoaminergic [175] axon terminals. As synaptic vesicles are limited to axon terminals and tend to cluster there, immunofluorescent labeling using antibodies to VTs provides in the light microscope a nice punctate immunosignal indicating the presence of axon terminals utilizing the specific neurotransmitter associated with the VTs [170]. Dale’s postulate One Neuron, One Neurotransmitter, however, does not hold for all VTs [176]. By combining fluorescence neuroanatomical tracing and highresolution confocal microscopy the presence of particular VTs inside axon terminals can be determined, thus providing information about the NT involved in the identified connectivity (reviews in [54, 177]). The advent of super-resolution fluorescence microscopy [28–30] opened up the era of real-time imaging of single synaptic vesicles in live neurons [178]. Again, this underscores our view that neurohistology, like many other fields, is by and large technology driven.

10

Integrative Approach: Correlative Light and Electron Microscopy Attempts to pinpoint single synaptic vesicles in live neurons can certainly be considered a tour de force with currently available optical microscopy techniques. In fact we are now witnessing a revolution in optical imaging owing to super-resolution fluorescence microscopy (see Subheading 2.4). The bridging of the wide gap in resolution between light microscopy—including superresolution microscopy (tens of nanometers) and electron microscopy (ca. 1 A˚ngstro¨m ¼ 0.1 nm) especially in living neurons has been a holy grail for neurohistologists from the point when the first EM studies of the CNS were published. Initially, light microscopic studies of particular types of neuron were conducted independently from electron microscopic studies of the same neurons. However, if one studies a particular type of neuron through Golgi-silver staining, for instance the cerebellar Purkynĕ cells, at a certain point it becomes necessary to study the ultrastructural details of processes of the same neuron also in the EM. Thus, correlative light-electron microscopy (CLEM) was born (Fig. 9).

36

Floris G. Wouterlood and Thomas P. Langer

Fig. 9 Correlative light and electron microscopy of Purkynĕ cell in rat cerebellum. The neuron has been Golgi-silver–impregnated (Golgi-Kopsch, see Subheading 4.1.2) and processed for EM [89]. D dendrite, GL granular layer, ML molecular layer, P cell body of the Purkynĕ cell. (A) Light microscopy image of the complete section embedded in plastic, prior to ultrasectioning. (B) One micron thick (semithin) control section for orientation. The silver chromate contents of the Purkynĕ cell (P) provides the primary light and electron dense label while capillaries (cap) act as secondary fiducial markers. (C) Montage of low magnification electron micrographs of the Purkynĕ cell’s dendrite (D). (D) High magnification electron micrograph of a distal dendrite (D) of the Golgisilver–impregnated Purkynĕ cell, showing three dendritic spines (sp1 to sp3) of which one (sp2) is in a synaptic contact with an axon terminal (at). Note the oligodendrocyte (Oligo) with its characteristic electron-dense chromatin in the nucleus (Nc). My myelinated axon, Nissl aggregate of rough endoplasmic reticulum, the EM equivalent of the Nissl substance

The first CLEM studies included procedures such as embedding sections in plastic, photographing areas of interest under a microscope equipped with a camera, recording significant landmarks, cutting out blocks from the sections, reembedding them for ultrasectioning, and subsequent investigation of the ultrathin

Neurohistology

37

sections in transmission EM, while navigating by the light microscopic landmarks and documentation. In 1965, a Norwegian scholar, Theodor Blackstad [179] pioneered studying ultrastructural details of Golgi-silver preparations of rat hippocampal neurons in this way, but it was Faire´n et al. [89] who introduced gold substitution of the silver chromate precipitate inside Golgi-silver– impregnated neurons. This innovation eliminated a major technical hurdle, notably that the solid silver chromate tended to ruin the glass and diamond knives of ultramicrotomes, and tear ultrathin sections apart. CLEM procedures were later extended to neurons subjected to intracellular injections of fluorescent dyes [131, 180] and DiI labeling [181]. The key to success with fluorescent dyes, themselves being not electron opaque and hence not showing up in the EM, is photoconversion, that is, irradiation with ultraviolet light of the intracellular dye in the presence of diaminobenzidine to obtain an electron dense product (a technique originally developed by Maranto [182]). With living neurons, the step from light microscopic observations to the electron microscopy level is much more challenging than with fixed nervous tissue, due to two major obstacles. The first one is the simple fact that a fluorescent molecule is typically not electron dense, so that photoconversion is again required [183]. The second hurdle is that GFP is no longer fluorescent when tissue is dehydrated. Raising antibodies against GFP and subsequent immunostaining has resolved this issue. Cryosectioning after live imaging and subsequent cryo-transmission EM has been proposed as a step further [184]. De Boer and his collaborators [185] recently reviewed the state of the art of CLEM approaches in cell biology in general. In 2017, the Nobel Prize in Chemistry was awarded to Jacques Dubochet, Joachim Frank, and Richard Henderson who developed the cryo-electron microscopy technique.

11

Outlook It should be clear from this brief survey of neurohistology that what one “sees” when examining nervous tissue strongly depends upon the technological tools available to visualize the structures of interest. Progress in neurohistology has advanced hand in hand with improvements in technology, to the point that the traditional approach, “fixing and staining” has matured into “manipulating and observing” (Fig. 10). Essentially, this is an integrative approach combining many disciplines engaged in studying living organisms, to the extent that the neurons themselves start to act as tools in revealing the intricacies and functionalities of this remarkable kind of tissue that we recognize as the nervous system.

38

Floris G. Wouterlood and Thomas P. Langer

Fig. 10 “Neurohistologists at work: scientists tinkering with neurons.” Can we still lure highly motivated researchers into the ever-exciting field of neurohistology?

As the tools are becoming more diverse and powerful, our understanding of the organization of the nervous system has been broadening and deepening. Conversely, this very process serves as a vital feedback to improve the methodology itself, This has been particularly so in the last few decades as the physiology and metabolism of neurons is increasingly employed as the means of tracing pathways and identifying specific populations of neurons lending the nervous system particular functional capabilities. As human ingenuity continues to invent new exciting techniques, neurohistology has never been as vital and exciting as it is nowadays.

Glossary Abbe diffraction limit

Counterstaining

Granular cortex Granule cell layer

Mathematical expression by Ernst Abbe describing a physical limit in the resolution of an optical lens system (holds equally for microscopes, telescopes, binoculars and camera lenses). Essentially controlled staining of the background, aimed to make previously stained cells in the same specimen better visible, and to appreciate the surrounding tissue architecture. Part of neocortex with a well-developed layer IV, the so-called internal granular layer. Discrete layer in a brain area densely packed with cell bodies of small compact neurons. Examples: granule cell layer of the dentate

Neurohistology

Molecular layer

Neuron doctrine

Nuclei

Reticular doctrine

Reticular formation

39

gyrus (part of the hippocampus), and in cerebellar cortex. Layer in one of the layered brain areas with relatively few neuronal cell bodies (e.g., layer I of cerebral cortex, outer layer of the hippocampus). A famous set of principles defining neurons as cytoplasmically discontinuous “elementary” units jointly forming the brain. Clusters of neuronal cell bodies in subcortical areas of the brain. Not to be confused with cell nuclei. Early interpretation of the brain, claiming that the brain is a syncytium with neurons in cytoplasmic continuum, thus forming a “reticulum.” Two connotations: (1) generic: brain area containing a mixture of neuronal cell bodies and loosely organized fiber bundles, (2) specific name of an extensive area in the brain stem controlling, among others, the level of consciousness, heart rate, and respiration.

References 1. Johnson JP (2011) Animal electricity, circa 1781. The Scientist 25 (Sept issue). https:// www.the-scientist.com/foundations/animalelectricity-circa-1781-41872 2. Knight B (1992) Discovering the human body. How pioneers of medicine solved the mysteries of the body’s structure and function, 2nd edn. Lippincott & Crowell, New York. ISBN:1854710842 3. Zimmer C (2005) Soul made flesh. The discovery of the brain - and how it changed the world. Free Press, New York. ISBN: 0743272056 4. Manzoni T (1998) The cerebral ventricles, the animal spirits and the dawn of brain localization of function. Arch Ital Biol 136:103–152. https://www.ncbi.nlm.nih. gov/pubmed/9492949 5. Breasted JH (1930) The Edwin Smith surgical papyrus. Kessinger Publishing LLC (facsimile print), Whitefish, MT. ISBN: 1162924330 6. Hippocrates on the sacred disease. http:// etext.library.adelaide.edu.au/mirror/classics. mit.edu/Hippocrates/sacred.html

7. Freemon FR (1994) Galen’s ideas on neurological function. J Hist Neurosci 3:263–271. https://doi.org/10.1080/ 09647049409525619 8. Sepper DL (2012) Animal spirits. In: Smith CUM, Frixione E, Finger S, Clower W (eds) The animal spirit doctrine and the origins of neurophysiology. Oxford University Press, Oxford, pp 26–28. https://doi.org/10. 1017/CBO9780511894695.011 9. Vesalius A (2014) The fabric of the human body. In: Garrison DH, Hast MH (eds) An annotated translation of the 1543 and 1555 editions of “De Humani Corporis Fabrica Libri Septem”, 2 Vols. Karger, Basel. ISBN: 3318022462 10. Willis T (1664) Cerebri anatome. https:// archive.org/details/cerebrianatomecu00will 11. Vieussens R (1684) Neurographia universalis. Hoc Est Omnium Corporis Humani Nervorum. Certe, Lyon. https://books.google. nl/books?id¼c681bMXHNyIC& printsec¼frontcover&hl¼nl&source¼gbs_ ge_summary_r&cad¼0#v¼onepage&q& f¼false

40

Floris G. Wouterlood and Thomas P. Langer

12. Baumann ED (1949) Franc¸ois dele Boe¨ Sylvius. E J Brill, Leiden 13. Gennari F (1782) De Peculiari Structura Cerebri Nonnulisque Eius Morbis. Parma Ex Regio Typographeo, Parma 14. Anatomia (1522–1867) Anatomical plates from the Thomas Fisher Rare Book Library. https://anatomia.library.utoronto.ca/ 15. Blum F (1893) Der Formaldehyd als H€artungsmittel. Z Wiss Mikrosk 10:314–315. https://biodiversitylibrary.org/ page/4637180 16. Fox CH, Johnson FB, Whiting J, Roller PP (1985) Formaldehyde fixation. J Histochem Cytochem 33:845–853. https://doi.org/10. 1177/33.8.3894502 17. Musiał A, Gryglewski RW, Kielczewski S, Loukas M, Wajda J (2016) Formalin use in anatomical and histological science in the 19th and 20th centuries. Folia Med Cracov 56:31–40. https://www.ncbi.nlm.nih.gov/ pubmed/28275269 18. Sabatini DD, Bensch K, Barrnett RJ (1963) Cytochemistry and electron microscopy. The preservation of cellular ultrastructure and enzymatic activity by aldehyde fixation. J Cell Biol 17:19–58. https://www.ncbi.nlm. nih.gov/pubmed/13975866 € 19. Weigert C (1882) Uber eine neue Untersuchungsmethode des Zentralnervensystems. In: Gesammelte Abhandlungen (1906), vol 2, pp 533–538 € 20. Heidenhain M (1892) Uber Kern und Protoplasma. Festschrift fu¨r A. von Ko¨lliker. W. Engelmann, Leipzig, pp 109–166 21. Nissl F (1894) Ueber eine neue Untersuchungsmethode des Centralorgans zur Feststellung der Localisation der Nervenzellen. Neurologisches Zentralblatt (Leipzig) 13:507–508 22. Golgi C (1873) Sulla struttura della sostanza grigia del cervello (Comunicazione preventiva). Gazzetta Medica Italiana, Lombardia 33:244–246 23. Ramo´n y Cajal S (1909) Histologie du Syste`me Nerveux de l’Homme et des Verte´bre´s, 2 Vols. Maloine, Paris. https://bio diversitylibrary.org/page/33072081 24. Quekett J (1848) A practical treatise on the use of the microscope, Chapter XII. In: Microtomes and microtome knives, 1st edn. Hippolyte Bailliere, London, p 306 25. Rutherford W (1873) A new freezing microtome. J Microsc (Oxf) 10:185–189. https:// doi.org/10.1111/j.1365-2818.1873. tb04666.x

26. Abbe E (1873) Beitr€age zur Theorie des Mikroskops und der mikroskopischen Wahrnehmung. Arch Mikrosk Anat (Bonn) 9 (1):413–468. https://doi.org/10.1007/ BF02956173. BHL:13290331. https://bio diversitylibrary.org/page/13290331 27. Brodal A (1980) Neurological anatomy in relation to clinical medicine. Oxford University Press, New York. ISBN: 0195026942 28. Saka SK (2014) Super-resolution microscopy: principles, techniques, and applications. In: Formsiero EF, Rizzoli SO (eds) Superresolution microscopy techniques in the neurosciences, Neuromethods, vol 86. Springer, New York, pp 1–11. https://doi.org/10. 1007/978-1-62703-983-3_1 29. Betzig E, Patterson GH, Sougrat R, Lindwasser OW, Olenych S, Bonifacino JS, Davidson MW, Lippincott-Schwartz J, Hess HF (2006) Imaging intracellular fluorescent proteins at nanometer resolution. Science 313 (5793):1642–1645. https://doi.org/10. 1126/science.1127344 30. Hell SW, Wichmann J (1994) Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy. Opt Lett 19:780–782. https://doi.org/10.1364/OL.19.000780 31. Gustafsson MGL (2000) Surpassing the lateral resolution limit by a factor of two using structured illumination microscopy. J Microsc (Oxf) 198:82–87. https://doi.org/10.1046/ j.1365-2818.2000.00710.x 32. Helmchen F, Denk W (2005) Deep tissue two-photon microscopy. Nat Methods 2:932–940. https://doi.org/10.1038/ NMETH818 33. Kobat D, Horton NG, Xu C (2011) In vivo two-photon microscopy to 1.6-mm depth in mouse cortex. J Biomed Opt 16(10):106014. https://doi.org/10.1117/1.3646209 34. Dodt HU, Saghafi S, Becker K, J€ahrling N, Hahn C, Pende M, Wanis M, Niendorf A (2015) Ultramicroscopy: development and outlook. Neurophotonics 2:041407. https://doi.org/10.1117/1.NPh.2.4. 041407 35. Park J-H, Kong L, Zhou Y, Cui M (2017) Large-field-of-view imaging by multi-pupil adaptive optics. Nat Methods 14:581–581. https://doi.org/10.1038/NMETH.4290 36. Schwann T (1847) Microscopical research into the accordance in the structure and growth of animals and plants. (“Mikroskopische Untersuchungen ueber die Uebereinstimmung in der Struktur und dem Wachstum der Thiere und Pflanzen”, 1839).

Neurohistology English translation of the original book by H. Smith. Sydenham Society, London 37. von Waldeyer-Hartz HWG (1891) Ueber einige neuere Forschungen im Gebiete der Anatomie des Centralnervensystems. Deutsche Medizinische Wochenschrift (Berlin) 17:1213–1218, 1244–1246, 1287–1289, 1331–1332, 1352–1356. https://doi.org/10.1055/s-0029-1206824 38. Santiago Ramo´n y Cajal. The Nobel Prize in Physiology or Medicine 1906. http://www. nobelprize.org/nobel_prizes/medicine/ laureates/1906/cajal-bio.html 39. Santiago Ramo´n y Cajal (1906) The structure and connexions of neurons (Nobel lecture). http://nobelprize.org/nobel_prizes/medi cine/laureates/1906/cajal-lecture.html 40. Santiago Ramo´n y Cajal (1852–1934) Life and discoveries. http://www.nobelprize.org/ nobel_prizes/medicine/laureates/1906/ cajal-article.html 41. Golgi C (1906) The neuron doctrine – theory and facts (Nobel lecture). http://nobelprize. org/nobel_prizes/medicine/laureates/ 1906/golgi-lecture.html 42. Foster M, Sherrington CS (1897) A textbook of physiology. Macmillan, New York. www. archive.org/stream/ textbookofphysio1897fost#page/n5/mode/ 2up 43. Marchi V, Algeri G (1885) Sulle degenerazioni discendenti consecutivea lesioni sperimen-tali in diverse zone della corteccia cerebrale. Riv Sper Freniatria Med Legal 11:492–494 € 44. Nissl F (1892) Uber experimental erzeugte Ver€anderungen an den Vorderhornzellen des Ru¨ckenmarks bei Kaninchen mit Demonstrationen mikroskopischer Pr€aparate. Allgemeine Zoologie und Psychiatrie 50:370–378 45. Arie¨ns-Kappers CU, Huber GC, Crosby EC (1936) The comparative anatomy of the nervous system of vertebrates, including man, 1st edn. Macmillan, New York 46. Stretton AO, Kravitz EA (1968) Neuronal geometry: determination with a technique of intracellular dye injection. Science 162:132–134. https://www.ncbi.nlm.nih. gov/pubmed/4175300 47. Porter KR, Claude A, Fullam EF (1945) Study of tissue culture cells by electron microscopy. Methods and preliminary observations. J Exp Med 81:233–246. https://doi. org/10.1084/jem.81.3.233 48. Palade GE, Palay SL (1954) Electron microscope observations on interneuronal and neuromuscular synapses. Anat Rec 118

41

(2):335–336. https://doi.org/10.1002/ar. 1091180211 49. Palay SL, Palade GE (1955) The fine structure of neurons. J Biophys Biochem Cytol 1:68–88. https://doi.org/10.1083/jcb.1.1.69 50. Peters A, Palay SL, De Webster FH (1991) The fine structure of the nervous system: neurons and their supporting cells, 3rd edn. Oxford University Press, Oxford. ISBN: 0195065719 51. Nauta WJ (1952) Selective silver impregnation of degenerating axons in the central nervous system. Stain Technol 27:175–179. https://www.ncbi.nlm.nih.gov/pubmed/ 14958510 52. Ebbesson SOE (1970) The selective silver impregnation of degenerating axons and their synaptic endings in nonmammalian species. In: Nauta WJH, Ebbesson SOE (eds) Contemporary research methods in neuroanatomy. Springer, New York, pp 132–161. https://doi.org/10.1007/978-3-64285986-1_7 53. Fink RP, Heimer L (1967) Two methods for selective silver impregnation of degenerating axons and their synaptic endings in the central nervous system. Brain Res 4:369–374. https://www.ncbi.nlm.nih.gov/pubmed/ 4166480 54. Wouterlood FG, Bloem B, Mansvelder HD, Luchicci A, Deisseroth K (2014) The fourth generation of neuroanatomical tracing techniques: exploiting the offspring of genetic engineering. J Neurosci Methods 236:331–348. https://doi.org/10.1016/j.jneumeth.2014. 07.021 55. Virchow R (1856) Gesammelte Abhandlungen zur Wissenschaftlichen Medizin. Hamm, Frankfurt am Main 56. Somjen GG (1988) Nervenkitt: notes on the history of the concept of neuroglia. Glia 1:2–9. https://www.ncbi.nlm.nih.gov/ pubmed/2976736 57. Love S, Perry A, Ironside J, Budka H (2015) Greenfield’s neuropathology, 9th edn. CRC Press, Boca Raton. ISBN: 9781498721288 58. Standring S (ed) (2008) Gray’s anatomy. The anatomical basis of clinical practice, 41th edn. Churchill Livingstone Elsevier, London. ISBN: 0702052302 59. Kandel ER, Schwartz JH, Jessell TM (eds) (2012) Principles of neural science, 5th edn. McGraw-Hill, New York. ISBN: 0071390111 60. Verkhratsky A, Butt A (2007) Glial neurobiology: a textbook. Wiley, Hoboken. https:// doi.org/10.1002/9780470517796

42

Floris G. Wouterlood and Thomas P. Langer

61. Krstic´ RV (1979) Ultrastruktur der S€augetierzelle. Springer, Berlin. ISBN: 9783662022825 62. Morell P, Quarles RJ (1999) Myelin formation, structure and biochemistry. In: Siegel GJ, Agranoff BW, Albers RW, Fisher SK, Uhler MD (eds) Basic neurochemistry: molecular, cellular and medical aspects, 6th edn. Lippincott-Raven, Philadelphia. ISBN: 039751820X 63. Simons M, Nave KA (2015) Oligodendrocytes: myelination and axonal support. Cold Spring Harb Perspect Biol 8:a020479. https://doi.org/10.1101/cshperspect. a020479 64. Sofroniew MV, Vinters HV (2010) Astrocytes: biology and pathology. Acta Neuropathol 119:7–35. https://doi.org/10.1007/ s00401-009-0619-8 65. Braak E (1975) On the fine structure of the external glial layer in the isocortex of man. Cell Tissue Res 157:367–390. https://www. ncbi.nlm.nih.gov/pubmed/1122547 66. Ma B, Buckalew R, Du Y, Kiyoshi CM, Alford CC, Wang W, McTigue DM, Enyeart JJ, Terman D, Zhou M (2016) Gap junction coupling confers isopotentiality on astrocyte syncytium. Glia 64:214–226. https://doi. org/10.1002/glia.22924 67. Bogie JF, Stinissen P, Hendriks JJ (2014) Macrophage subsets and microglia in multiple sclerosis. Acta Neuropathol 128:191–213. https://doi.org/10.1007/s00401-0141310-2 68. Bennett ML, Bennett FC, Liddelow SA, Ajami B, Zamanian JL, Fernhoff NB, Mulinyawe SB, Bohlen CJ, Adil A, Tucker A, Weissman IL, Chang EF, Li G, Grant GA, Hayden Gephart MG, Barres BA (2016) New tools for studying microglia in the mouse and human CNS. Proc Natl Acad Sci U S A 113:E1738–E1746. https://doi.org/ 10.1073/pnas.1525528113 69. Wouterlood FG, Alones VE, Lohman AHM (1981) An electron microscopic study of the mediodorsal cortex in the lizard Agama agama. Anat Embryol 163:185–200. https://doi.org/10.1007/BF00320675 70. Cifuentes M, Baeza V, Arrabal PM, Visser R, Grondona JM, Saldivia N, Martı´nez F, Nualart F, Salazar K (2017) Expression of a novel ciliary protein, IIIG9, during the differentiation and maturation of ependymal cells. Mol Neurobiol 55(2):1652–1664. https:// doi.org/10.1007/s12035-017-0434-5 71. de Zeeuw CI, Hoogland TM (2015) Reappraisal of Bergmann glial cells as modulators

of cerebellar circuit function. Front Cell Neurosci 9:246. https://doi.org/10.3389/fncel. 2015.00246 72. Rodriguez EM, Bla´zquez JL, Pastor FE, ˜ a P, Peruzzo B, Amat P (2005) Pela´ez B, Pen Hypothalamic tanycytes: a key component of the brain-endocrine interaction. Int Rev Cytol 247:89–164. https://www.ncbi.nlm.nih. gov/pubmed/16344112 73. Eccles JC, Ito M, Szenta´gothai J (1967) The cerebellum as a neuronal machine. Springer, Frankfurt. https://doi.org/10.1007/978-3662-13147-3 74. Ito M (2006) Cerebellar circuitry as a neuronal machine. Prog Neurobiol 78:272–303. https://doi.org/10.1016/j.pneurobio.2006. 02.006 75. Voogd J, Ruigrok TJH (2012) Cerebellum and precerebellar nuclei. In: Mai JK, Paxinos G (eds) The human nervous system, 3rd edn. Elsevier, London, pp 471–545. https:// doi.org/10.1016/B978-0-12-374236-0. 10015-X 76. Wouterlood FG (1986) Stabilization of silver chromate Golgi impregnation in rat central nervous system neurons using photographic developers. I. Light microscopy. Stain Technol 61:337–352. https://www.ncbi.nlm.nih. gov/pubmed/3798487 77. Nimchinsky EA, Sabatini BL, Svoboda K (2002) Structure and function of dendritic spines. Annu Rev Physiol 64:313–353. https://doi.org/10.1146/annurev.physiol. 64.081501.160008 78. Dorostkar MM, Zou C, Blazquez-Llorca L, Herms J (2015) Analyzing dendritic spine pathology in Alzheimer’s disease: problems and opportunities. Acta Neuropathol 130:1–19. https://doi.org/10.1007/ s00401-015-1449-5 79. Haber SN, Wolfe DP, Groenewegen HJ (1990) The relationship between ventral striatal efferent fibers and the distribution of peptide-positive woolly fibers in the forebrain of the rhesus monkey. Neuroscience 39:323–338. https://www.ncbi.nlm.nih. gov/pubmed/1708114 80. Purves D, Augustine GJ, Fitzpatrick D, Hall WC, LaMantia A-S, White LE (eds) (2011) Neuroscience, 5th edn. Oxford University Press, Oxford. ISBN: 9780878936953 81. Somogyi P, Klausberger T (2005) Defined types of cortical interneurone structure space and spike timing in the hippocampus. J Physiol Lond 562:9–26. https://doi.org/10. 1113/jphysiol.2004.078915

Neurohistology 82. Ascoli GA, Alonso-Nanclares L, Anderson SA, Barrionuevo G, Benavides-Piccione R, Burkhalter A, Buzsa´ki G, Cauli B, Defelipe J, Faire´n A, Feldmeyer D, Fishell G, Fregnac Y, Freund TF, Gardner D, Gardner EP, Goldberg JH, Helmstaedter M, Hestrin S, Karube F, Kisva´rday ZF, Lambolez B, Lewis ˜ oz A, DA, Marin O, Markram H, Mun Packer A, Petersen CC, Rockland KS, Rossier J, Rudy B, Somogyi P, Staiger JF, Tamas G, Thomson AM, Toledo-RodriguezM, Wang Y, West DC, Yuste R (2008) Petilla terminology: nomenclature of features of GABAergic interneurons of the cerebral cortex. Nat Rev Neurosci 9:557–568. https:// doi.org/10.1038/nrn2402 83. Barrett LF, Simmons WK (2015) Interoceptive predictions in the brain. Nat Rev Neurosci 16:419–429. https://doi.org/10. 1038/nrn3950 84. von Economo C (1925) Die fu¨nf Bautypen der Grosshirnrinde. Schweiz Arch Neurol Psychiatr 16:260–269 85. von Economo C, Koskinas GN (2008) Atlas of cytoarchitectonics of the adult human cerebral cortex. In: Triarhou LC (ed) Translation and revision of “Die Cytoarchitektonik der Hirnrinde des erwachsenen Menschen” (1925). Karger, Basel. ISBN: 9783805582896 86. Fregerslev S, Blackstad TW, Fredens K, Holm MJ (1971) Golgi potassium-dichromate silver-nitrate impregnation. Nature of the precipitate studied by x-ray powder diffraction methods. Histochemie 25:63–71. https:// www.ncbi.nlm.nih.gov/pubmed/4929087 87. Rio-Hortega P (1928) Tercera aportacio´n al conocimiento morfolo´gico e interpretacio´n funcional de la oligodendroglia. Mem Real Soc Esp Nat 14:1–128 88. Mugnaini E, Osen KK, Dahl AL, Friedrich VL Jr, Korte G (1980) Fine structure of granule cells and related interneurons (termed Golgi cells) in the cochlear nuclear complex of cat, rat and mouse. J Neurocytol 9:537–570. https://www.ncbi.nlm.nih.gov/pubmed/ 7441303 89. Faire´n A, Peters A, Saldanha J (1977) A new procedure for examining Golgi impregnated neurons by light and electron microscopy. J Neurocytol 6:311–337. https://www.ncbi. nlm.nih.gov/pubmed/71343 90. Wouterlood FG, Nederlof J, Paniry S (1983) Chemical reduction of silver chromate: a procedure for electron microscopical analysis of Golgi impregnated neurons. J Neurosci Methods 7:295–308. https://www.ncbi.nlm. nih.gov/pubmed/6865470

43

91. Ehrlich P, Krause R, Mosse M, Rosin H, Weigert C (eds) (1910) Enzyklop€adie der mikroskopischen Technik, 2nd edn. Urban und Schwarzenberg, Berlin-Wien. Reprinted in 2011 by Nabu Press, ISBN: 1248177053 92. Smith WK, Quigley B (1937) A new method for the rapid staining of myelin sheaths. Am J Pathol 13:491–496. PMID:19970339 93. Weil A (1928) A rapid method for staining myelin sheaths. Arch Neurol Psychiatr 20:392–393. https://doi.org/10.1001/ archneurpsyc.1928.02210140160013 94. Klu¨ver H, Barrera EA (1953) Method for the combined staining of cells and fibers in the nervous system. J Neuropathol Exp Neurol 12:400–403. https://www.ncbi.nlm.nih. gov/pubmed/13097193 95. Wouterlood FG, Alones VE, Elprana D, Lohman AHM (1982) Terminal degeneration in the mediodorsal cortex of the lizard Agama agama: light and electron microscopy. J Morphol 172:45–58. https://doi.org/10.1002/ jmor.1051720105 96. Hirokawa N, Noda Y (2008) Intracellular transport and kinesin superfamily proteins, KIF’s: structure, function, and dynamics. Physiol Rev 88(3):1089–1118. https://doi. org/10.1152/physrev.00023.2007 97. Lodish H, Berk A, Zipursky SL, Matsudaira M, Baltimore D, Darnell J (2000) Chapter 19: Cell motility and shape II: Microtubules and intermediate filament. In: Molecular cell biology, 4th edn. W.H. Freeman, New York. ISBN: 0716731363 98. Kristensson K, Olsson Y (1971) Uptake and retrograde axonal transport of peroxidase in hypoglossal neurons. Electron microscopical localization in the neuronal perikaryon. Acta Neuropathol 19(1):1–9. https://www.ncbi. nlm.nih.gov/pubmed/4107378 99. Graham RC Jr, Karnovsky MJ (1965) The histochemical demonstration of monoamine oxidase activity by coupled peroxidatic oxidation. J Histochem Cytochem 13:604–605. https://www.ncbi.nlm.nih.gov/pubmed/ 5861361 100. Lanciego JL, Wouterlood FG (2011) A half century of experimental neuroanatomical tracing. J Chem Neuroanat 42:157–183. https://doi.org/10.1016/j.jchemneu.2011. 07.001 101. Droz B, Leblond CP (1962) Migration of proteins along the axons of the sciatic nerve. Science 137(3535):1047–1048. https:// www.ncbi.nlm.nih.gov/pubmed/13887927

44

Floris G. Wouterlood and Thomas P. Langer

102. Singer M, Salpeter MM (1966) Transport of tritium-labelled l-histidine through the Schwann and myelin sheaths into the axon of peripheral nerves. Nature 210 (5042):1225–1227. https://www.ncbi.nlm. nih.gov/pubmed/5967796 103. Cowan WM, Gottlieb DI, Hendrickson AE, Price JL, Woolsey TA (1972) The autoradiographic demonstration of axonal connections in the central nervous system. Brain Res 37:21–51. https://www.ncbi.nlm.nih.gov/ pubmed/4110604 104. Bachmann L, Salpeter MM (1967) Absolute sensitivity of electron microscope radioautography. J Cell Biol 33:299–305. https:// www.ncbi.nlm.nih.gov/pubmed/6039374 105. Bachmann L, Salpeter MM, Salpeter EE (1968) Resolving power of electron microscope autoradiography. Histochemie 15:234–250. https://www.ncbi.nlm.nih. gov/pubmed/5742134 106. Cowan WM, Cuenod M (1975) The use of axoplasmic transport for the study of neural connections: a retrospective survey. In: Cowan WM, Cuenod M (eds) The use of axonal transport for studies of neural connectivity. Proceedings of a symposium (GwattThun, Switzerland, July 1974), vol 99. Elsevier, Amsterdam, pp 1–24. https://doi.org/ 10.1093/brain/99.2.391 € 107. Sihler C (1895) Uber Muskelspindeln und intramuskulare Nervenendigungen bei Schlangen und Froschen. Archiv fu¨r Mikroskopische Anatomie und Entwicklungsgeschichte 46:709–723. https://bio diversitylibrary.org/page/13989376 108. Mu L, Sanders I (2010) Sihler’s whole mount nerve staining technique: a review. Biotech Histochem 85:19–65. https://doi.org/10. 1080/10520290903048384 109. Gu¨lekon N, Peker T, Turgut HB, Anil A, Karako¨se M (2007) Qualitative comparison of anatomical microdissection, Sihler’s staining and computerized reconstruction methods for visualizing intramuscular nerve branches. Surg Radiol Anat 29:373–378. PMID:17568985 110. Peker T, Gu¨lekon N, Anil A, Turgut HB, Liman F (2003) Clearing and embedding in polyester resin for demonstrating the nerve distribution pattern of skeletal muscles. Biotech Histochem 78:187–190. https://www. ncbi.nlm.nih.gov/pubmed/14714882 € 111. Spalteholz W (1911) Uber das Durchsichtigmachen von menschlichen und tierischen Pr€aparaten und seine theoretischen Bedingungen. S. Hirzel, Leipzig. http://

s2w.hbz-nrw.de/zbmed/content/titleinfo/ 555354 112. Azaripour A, Lagerweij T, Scharfbillig C, Jadczak AE, Willershausen B, Van Noorden CJF (2016) A survey of clearing techniques for 3D imaging of tissues with special reference to connective tissue. Prog Histochem Cytochem 51:9–23. https://doi.org/10. 1016/j.proghi.2016.04.001 113. Richardson DS, Lichtman JW (2016) Clarifying tissue clearing. Cell 162:246–257. https://doi.org/10.1016/j.cell.2015.06. 067 114. Chung K, Wallace J, Kim SY, Kalyanasundaram S, Andalman AS, Davidson TJ, Mirzabekov JJ, Zalocusky KA, Mattis J, Denisin AK, Pak S, Bernstein H, Ramakrishnan C, Grosenick L, Gradinaru V, Deisseroth K (2013) Structural and molecular interrogation of intact biological systems. Nature 497(7449):332–337. https://doi. org/10.1038/nature12107 115. Hama H, Kurokawa H, Kawano H, Ando R, Shimogori T, Noda H, Fukami K, SakaueSawano A, Miyawaki A (2011) Scale: a chemical approach for fluorescence imaging and reconstruction of transparent mouse brain. Nat Neurosci 14:1481–1488. https://doi. org/10.1038/nn.2928 116. Hama H, Hioki H, Namiki K, Hoshida T, Kurokawa H, Ishidate F, Kaneko T, Akagi S, Saido T, Miyawaki A (2015) ScaleS: an optical clearing palette for biological imaging. Nat Neurosci 18:1518–1529. https://doi.org/ 10.1038/nn.4107 117. Bocarsly ME, Jiang WC, Wang C, Dudman JT, Ji N, Aponte Y (2015) Minimally invasive microendoscopy system for in vivo functional imaging of deep nuclei in the mouse brain. Biomed Opt Express 6:4546–4556. https:// doi.org/10.1364/BOE.6.004546 118. Jensen KH, Berg RW (2016) CLARITYcompatible lipophilic dyes for electrode marking and neuronal tracing. Sci Rep 6:32674. https://doi.org/10.1038/srep32674 119. Peter K (1906) Die Methoden der Rekonstruktion. Gustav Fischer Verlag, Jena 120. Reese AM (1910) Development of the digestive canal of the American alligator. Smithsonian Miscellaneous Collections 56(11):1–25. www.gutenberg.org/files/22327/22327-h/ 22327-h.htm 121. Waterston D (1913) Reconstruction in modelling clay: a rapid method of plastic reconstruction from serial sections. J Anat Physiol 48(Pt 1):19–23. https://www.ncbi.nlm.nih. gov/pubmed/17232979

Neurohistology 122. Gras H, Killmann F (1983) NEUREC - a program package for 3D-reconstruction from serial sections using a microcomputer. Comput Programs Biomed 17 (1–2):145–155. https://www.ncbi.nlm.nih. gov/pubmed/6689291 123. Naora H (1951) Microspectrophotometry and cytochemical analysis of nucleic acids. Science 114(2959):279–280. https://doi. org/10.1126/science.114.2959.279 124. Minsky M (1988) Memoir on inventing the confocal scanning microscope. Scanning 10:128–138. https://doi.org/10.1002/sca. 4950100403 125. Egger MD, Petra´nˇ M (1967) New reflectedlight microscope for viewing unstained brain and ganglion cells. Science 157:305–307. https://doi.org/10.1126/science.157. 3786.305 126. Brakenhoff GJ, Blom P, Barends P (1979) Confocal scanning light microscopy with high aperture immersion lenses. J Microsc (Oxf) 117:219–232. https://doi.org/10. 1111/j.1365-2818.1979.tb01178.x 127. Chen BE, Lendvai B, Nimchinsky EA, Burbach B, Fox K, Svoboda K (2000) Imaging high-resolution structure of GFP-expressing neurons in neocortex in vivo. Learn Mem 7:433–441. https:// www.ncbi.nlm.nih.gov/pubmed/11112802 128. Dorr AE, Lerch JP, Spring S, Kabani N, Henkelman RM (2008) High resolution threedimensional brain atlas using an average magnetic resonance image of 40 adult C57Bl/6J mice. NeuroImage 42:60–69. https://doi. org/10.1016/j.neuroimage.2008.03.037 129. Pawley JB (ed) (2006) Handbook of biological confocal microscopy, 3rd edn. Springer, New York. https://doi.org/10. 1007/978-0-387-45524-2 130. Rasband WS (1997–2006) ImageJ. Natl Inst Health, Bethesda, MD. http://rsb.info.nih. gov/ij/ 131. Buhl EH (1993) Intracellular injection in fixed slices in combination with neuroanatomical tracing techniques and electron microscopy to determine multisynaptic pathways in the brain. Microsc Res Tech 24:15–30. https://doi.org/10.1002/jemt. 1070240104 132. Wouterlood FG, Barbas-Henry HA, Lohman AH (1988) Mixed-junction axon terminals on identified principal abducens motoneurons in the monitor lizard. Brain Res 463:198–203. https://www.ncbi.nlm.nih. gov/pubmed/3196908

45

133. De Winter DA, Schneijdenberg CT, Lebbink MN, Lich B, Verkleij AJ, Drury MR, Humbel BM (2009) Tomography of insulating biological and geological materials using focused ion beam (FIB) sectioning and low-kV BSE imaging. J Microsc (Oxf) 233:372–383. https://doi.org/10.1111/j. 1365-2818.2009.03139.x 134. Lee D, Huang T-H, De La Cruz A, Callejas A, Lois C (2017) Methods to investigate the structure and connectivity of the nervous system. Fly (Austin) 16:1–15. https://doi.org/ 10.1080/19336934.2017.1295189 135. Battaglia D, Karagiannis A, Gallopin T, Gutch HW, Cauli B (2013) Beyond the frontiers of neuronal types. Front Neural Circuits 7:7–13. https://doi.org/10.3389/fncir.2013.00013 136. van Aerde KI, Feldmeyer D (2015) Morphological and physiological characterization of pyramidal neuron subtypes in rat medial prefrontal cortex. Cereb Cortex 25:788–805. https://doi.org/10.1093/cercor/bht278 137. Deisseroth K (2010) Optogenetics. Nat Methods 8:26–29. https://doi.org/10. 1038/nmeth.f.324 138. Han X (2012) In vivo application of optogenetics for neural circuit analysis. ACS Chem Neurosci 3:577–584. https://doi.org/10. 1021/cn300065j 139. Sidor M, Davidson TJ, Tye KM, Warden MR, Deisseroth K, McClung CA (2015) In vivo optogenetic stimulation of the rodent central nervous system. J Vis Exp 95:e51483. https://doi.org/10.3791/51483 140. Hu C-D, Kerppola TK (2003) Simultaneous visualization of multiple protein interactions in living cells using multicolor fluorescence complementation analysis. Nat Biotechnol 21:539–545. https://doi.org/10.1038/ nbt816 141. Feinberg EH, Vanhoven MK, Bendesky A, Wang G, Fetter RD, Shen K, Bargmann CI (2008) GFP reconstitution across synaptic partners (GRASP) defines cell contacts and synapses in living nervous systems. Neuron 57:353–363. https://doi.org/10.1016/j. neuron.2007.11.030 142. Gordon MD, Scott K (2009) Motor control in a Drosophila taste circuit. Neuron 61:373–384. https://doi.org/10.1016/j. neuron.2008.12.033 143. Macpherson LJ, Zaharieva EE, Kearney PJ, Alpert MH, Lin T-Y, Turan Z, Lee C-H, Gallio M (2015) Dynamic labelling of neural connections in multiple colours by transsynaptic fluorescence complementation. Nat

46

Floris G. Wouterlood and Thomas P. Langer

Commun 6:10024. https://doi.org/10. 1038/ncomms10024 144. Yamagata M, Sanes JR (2012) Transgenic strategy for identifying synaptic connections in mice by fluorescence complementation (GRASP). Front Mol Neurosci 18:1–9. https://doi.org/10.3389/fnmol.2012. 00018 145. Kim J, Cho S-G, Song M, Kang S-R, Kwon SY, Choi K-H, Choi S-M, Kim B-C, Song H-C (2016) Usefulness of 3-dimensional stereotactic surface projection FDG PET images for the diagnosis of dementia. Medicine (Baltimore) 95:e5622. https://doi.org/10. 1097/MD.0000000000005622 146. Wree A (1990) Principles of the 2-deoxyglucose method for the determination of the local cerebral glucose utilization. Eur J Morphol 28:132–138. https://www. ncbi.nlm.nih.gov/pubmed/2245129 147. Shummers J, Yu H, Sur M (2008) Tuned responses of astrocytes and their influence on hemodynamic signals in the visual cortex. Science 320(5883):1638–1643. https://doi. org/10.1126/science.1156120 148. Hartline FF (1979) Biological applications for voltage sensitive dyes. Science 203 (4384):992–994. https://www.ncbi.nlm.nih. gov/pubmed/424731 149. Sidman RL (1970) Autoradiographic methods and principles for study of the nervous system with thymidine-H3. In: Nauta WJH, Ebbesson SOE (eds) Contemporary research methods in neuroanatomy. Springer, New York, pp 252–274. https://doi.org/10. 1007/978-3-642-85986-1_12 150. Horie T, Shinki R, Ogura Y, Kusakabe TG, Satoh N, Sasakura Y (2011) Ependymal cells of chordate larvae are stem-like cells that form the adult nervous system. Nature 469 (7331):525–528. https://doi.org/10.1038/ nature09631 151. Tabata H, Nakajima K (2001) Efficient in utero gene transfer system to the developing mouse brain using electroporation: visualization of neuronal migration in the developing cortex. Neuroscience 103:865–872. https:// doi.org/10.1016/S0306-4522(01)00016-1 152. Saito T, Nakatsuji N (2001) Efficient gene transfer into the embryonic mouse brain using in vivo electroporation. Dev Biol 240:237–246. https://doi.org/10.1006/ dbio.2001.043

153. Vacher H, Mohapatra DP, Trimmer JS (2008) Localization and targeting of voltagedependent ion channels in mammalian central neurons. Physiol Rev 88:1407–1447. https://doi.org/10.1152/physrev.00002. 2008 154. Yamamura H, Suzuki Y, Imaizumi Y (2015) New light on ion channel imaging by total internal reflection fluorescence (TIRF) microscopy. J Pharm Sci 128:1–7. https:// doi.org/10.1016/j.jphs.2015.04.004 155. Boyden ES, Zhang F, Bamberg E, Nagel G, Deisseroth K (2005) Millisecond-timescale, genetically targeted optical control of neural activity. Nat Neurosci 8:1263–1268. https:// doi.org/10.1038/nn1525 156. Yizhar O, Fenno LE, Davidson TJ, Mogri M, Deisseroth K (2011) Optogenetics in neural systems. Neuron 71:9–34. https://doi.org/ 10.1016/j.neuron.2011.06.004 157. Pappas GD, Bennett MVL (1966) Specialized junctions involved in electrical transmission between neurons. Ann N Y Acad Sci 137:495–508. https://www.ncbi.nlm.nih. gov/pubmed/5229811 158. Pappas GD, Asada Y, Bennett MVL (1971) Morphological correlates of increased coupling resistance at an electrotonic synapse. J Cell Biol 49:173–188. https://www.ncbi. nlm.nih.gov/pubmed/4995387 159. Sotelo C, Llina´s R, Baker R (1974) Structural study of inferior olivary nucleus of the cat: morphological correlates of electrotonic coupling. J Neurophysiol 37:541–559. https:// www.ncbi.nlm.nih.gov/pubmed/4827021 160. Wouterlood FG, Mugnaini E, Osen KK, Dahl A (1984) Stellate neurons in rat dorsal cochlear nucleus studies with combined Golgi impregnation and electron microscopy: synaptic connections and mutual coupling by gap junctions. J Neurocytol 13:639–664. https://www.ncbi.nlm.nih.gov/pubmed/ 6481413 161. Raviola E, Gilula NB (1973) Gap junctions between photoreceptor cells in the vertebrate retina. Proc Natl Acad Sci U S A 70:1677–1681. https://www.ncbi.nlm.nih. gov/pubmed/4198274 162. Sloper JJ, Powell TP (1978) Gap junctions between dendrites and somata of neurons in the primate sensori-motor cortex. Proc R Soc Lond B Biol Sci 203(1150):39–47. https:// www.ncbi.nlm.nih.gov/pubmed/32543

Neurohistology 163. Fukuda T, Kosaka T (2003) Ultrastructural study of gap junctions between dendrites of parvalbumin-containing GABAergic neurons in various neocortical areas of the adult rat. Neuroscience 120:5–20. https://doi.org/10. 1016/S0306-4522(03)00328-2 164. Evans WH, Martin PE (2002) Gap junctions: structure and function. Mol Membr Biol 19:121–136. https://doi.org/10.1080/ 09687680210139839 165. Moore KB, O’Brien J (2015) Connexins in neurons and glia: targets for intervention in disease and injury. Neural Regen Res 10:1013–1017. https://doi.org/10.4103/ 1673-5374.160092 166. Rash JE, Kamasawa N, Vanderpool KG, Yasumura T, O’Brien J, Nannapaneni S, Pereda AE, Nagy JI (2015) Heterotypic gap junctions at glutamatergic mixed synapses are abundant in goldfish brain. Neuroscience 29:166–193. https://doi.org/10.1016/j. neuroscience.2014.10.057 167. Carlsson A, Falck B, Hillarp NA (1962) Cellular localization of brain monoamines. Acta Physiol Scand 56(Suppl 196):1–28. https:// www.ncbi.nlm.nih.gov/pubmed/14018711 168. Ho¨kfelt T (2016) Early attempts to visualize cortical monoamine nerve terminals. Brain Res 1645:8–11. https://doi.org/10.1016/j. brainres.2016.01.024 169. Steinbusch HW, Verhofstad AA, Joosten HW (1978) Localization of serotonin in the central nervous system by immunohistochemistry: description of a specific and sensitive technique and some applications. Neuroscience 3:811–819. PMID:362232 170. Eiden LE (2000) The vesicular neurotransmitter transporters: current perspectives and future prospects. FASEB J 14:2396–2400. https://doi.org/10.1096/fj.00-0817rev 171. Blakely RD, Edwards RH (2012) Vesicular and plasma membrane transporters for neurotransmitters. Cold Spring Harb Perspect Biol 4:1–24. https://doi.org/10.1101/ cshperspect.a005595 172. Fremeau RT Jr, Kam K, Qureshi T, Johnson J, Copenhagen DR, Storm-Mathisen J, Chaudhry FA, Nicoll RA, Edwards RH (2004) Vesicular glutamate transporters 1 and 2 target to functionally distinct synaptic release sites. Science 304:1815–1819. https://doi.org/10.1126/science.1097468 173. Chaudhry FA, Reimer RJ, Bellocchio EE, Danbolt NC, Osen KK, Edwards RH,

47

Storm-Mathisen J (1998) The vesicular GABA transporter, VGAT, localizes to synaptic vesicles in sets of glycinergic as well as GABAergic neurons. J Neurosci 18:9733–9750. PMID:9822734 174. Gilmor ML, Nash NR, Roghani A, Edwards RH, Yi H, Hersch SM, Levey AI (1996) Expression of the putative vesicular acetylcholine transporter in rat brain and localization in cholinergic synaptic vesicles. J Neurosci 16:2179–2190. https://www.ncbi.nlm.nih. gov/pubmed/8601799 175. Peter D, Liu Y, Sternini C, de Giorgio R, Brecha N, Edwards RH (1995) Differential expression of two vesicular monoamine transporters. J Neurosci 15:6179–6188. https:// www.ncbi.nlm.nih.gov/pubmed/7666200 176. Boulland J-L, Jenstad M, Boekel A, Wouterlood FG, Edwards RH, Storm-Mathisen J, Chaudhry FA (2009) Vesicular glutamate and GABA transporters sort to distinct sets of vesicles at a symmetric synapse. Cereb Cortex 19:241–248. https://doi.org/10.1093/ cercor/bhn077 177. Wouterlood FG (2015) A survey of current neuroanatomical tracing techniques. In: Arenkiel BR (ed) Neural tracing methods: tracing neurons and their connections, Neuromethods, vol 92, pp 1–49. https://doi.org/ 10.1007/978-1-4939-1963-5_1 178. Yu C, Zhang M, Qin X, Yang X, Park H (2016) Real time imaging of single synaptic vesicles in live neurons. Front Biol 11:109. https://doi.org/10.1007/s11515-0161397-z 179. Blackstad TW (1965) Mapping of experimental axon degeneration by electron microscopy of Golgi preparations. Z Zellforsch Mikrosk Anat 67(6):819–834. https://doi.org/10. 1007/BF00339303 180. Meredith GE, Wouterlood FG (1993) Identification of synaptic interactions of intracellularly injected neurons in fixed brain slices by means of dual-label electron microscopy. Microsc Res Tech 24:31–42. https://doi. org/10.1002/jemt.1070240105 181. von Bartheld CS, Cunningham DE, Rubel EW (1990) Neuronal tracing with DiI: decalcification, cryosectioning, and photoconversion for light and electron microscopic analysis. J Histochem Cytochem 38 (5):725–733. https://doi.org/10.1177/38. 5.2185313

48

Floris G. Wouterlood and Thomas P. Langer

182. Maranto AR (1982) Neuronal mapping: a photooxidation reaction makes Lucifer yellow useful for electron microscopy. Science 217 (4563):953–955. https://www.ncbi.nlm.nih. gov/pubmed/7112109 183. Kukulski W, Schorb M, Welsch S, Picco A, Kaksonen M, Briggs JA (2012) Precise, correlated fluorescence microscopy and electron tomography of Lowicryl sections using fluorescent fiducial markers. Methods Cell Biol 111:235–257. https://doi.org/10.1016/ B978-0-12-416026-2.00013-3

184. van Rijnsoever C, Oorschot V, Klumperman J (2008) Correlative light-electron microscopy (CLEM) combining live-cell imaging and immunolabeling of ultrathin cryosections. Nat Methods 5:973–980. https://doi.org/ 10.1038/nmeth.1263 185. de Boer P, Hoogenboom JP, Giepmans BNG (2015) Correlated light and electron microscopy: ultrastructure lights up! Nat Methods 12:503–513. https://doi.org/10.1038/ nmeth.3400

Chapter 2 Fixation Protocols for Neurohistology: Neurons to Genes Elliott J. Mufson, Sylvia E. Perez, Christy M. Kelley, Melissa J. Alldred, and Stephen D. Ginsberg Abstract Since ancient times, tissue fixation for neurohistology has been an evolving area of research. Alcohol fixation first made possible examination of brain tissue specimens. In the late 1800s formaldehyde was introduced as a cross-linking fixative, which enabled histological advances. Various aldehyde solutions remain the staple for neurohistology in the neuroscience community including formalin, paraformaldehyde, glutaraldehyde, and combinations of these fixatives. A 4% paraformaldehyde solution is commonly used for numerous cytochemical procedures including tract tracing, immunohistochemistry, and in situ hybridization. A glutaraldehyde–paraformaldehyde solution is the fixative of choice for ultrastructural investigations using the electron microscope. With the advent of modern molecular and cellular biological techniques, it is now possible to isolate and study genomic DNA, RNA species, and proteins from microdissected tissue sources. Similar to light and electron microscopy, alcohol and aldehyde tissue fixation are compatible with preserving RNA integrity for regional and single cell gene array experiments using immersion- and perfusion-fixed tissue from a myriad of brain tissue sources including humans and relevant animal and cellular models. Key words Aldehydes, Brain, Cryoprotectant, Fixation, Formaldehyde, Glutaraldehyde, Microarray, Microscopy, Neurogenetics, Neurohistology

Abbreviations CA1 DMSO LCM NBF OCT PLP PMI qRT-PCT

Region 1 of cornu ammonis of the hippocampus Dimethylsulfoxide Laser capture microdissection Neutral buffered formalin Optimal cutting temperature (compound) Periodate lysine paraformaldehyde Postmortem interval Quantitative real-time polymerase chain reaction

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_2, © Springer Science+Business Media LLC 2020

49

50

1

Elliott J. Mufson et al.

Historical Perspective of Neurofixation Alcohol has been used as a fixative for soft tissue since the middle of the seventeenth century. The Dutch anatomist Frederik Ruysch introduced the preservative in the 1660s [1], and it became the fixative of choice for the burgeoning field of specimen collection. Like many scientific endeavors of the time, specimen obtaining and showcasing was as much art as it was science. Indeed, alcohol fixation quickly became the norm for physicians and anatomists— the possessors of the curio collections. Although alcohol made it possible to investigate fixed organs of the body, examination of the brain remained uncharted waters. Interestingly, anatomical guides for students in the eighteenth century and as late as the early nineteenth century still provided techniques for flushing out the brain so as to continue preserving the rest of the body [2]. Johann Christian Reil, like many of his neuroscientist predecessors, observed that the brain degrades over time when simply left in water or other fluids [3]. To resolve this problem, Reil expanded the prevalent fixation method of alcohol submersion by combining a strong alkaline agent such as potassium carbonate or ammonia to this fixative [4]. The brain was placed in this novel preservative, which allowed for the hardening of the brain and examination of its hidden structures. With advent of the achromatic microscope in the 1820s, anatomists, physicians, and their scientific colleagues devised histological stains for the examination of brain tissue. A few of the more notable stains were carmine dyes in 1867 by Gerlach; gold, circa 1867 by Cohnheim; silver in 1873 by Golgi; and methylene blue in 1886 by Ehrlich [5, 6]. Many of these chemicals were prepared in idiosyncratic concoctions involving mordants and buffers, which resulted in an array of histological effects. For example, a strong acid made from the dissolution of chromium trioxide in water frequently formed into salts. Prior to the use of this preservative for brain tissue, the Danish microscopist Adolph Hannover used chromic acid in 1840 as a fixative for brain tissue. Thus, with the introduction of chromic acid fixation and the development of new tissue staining preparations it was becoming possible to unravel the structure of the brain. Despite the utility of variations of chromic acid and its salts for fixation, alcohol remained a common reagent throughout the latter half of the nineteenth century by brain surveyors. In addition to the use of chromic acid and alcohol for fixation, other formulas for fixation were tried including the injection of gelatin into the brain by the French histologist LouisAntoine Ranvier in the 1860s [6]. It was not until the end of the nineteenth century that a new method of fixation gained prominence.

Fixation Protocols for Neurohistology

51

In 1893, the German physician Ferdinand Blum introduced formaldehyde as a viable fixative for tissue. While testing the chemical for antiseptic properties, Blum experienced an increased firmness in his fingers that resulted from contact with the solution. Soon after this, formaldehyde was confirmed as an advantageous fixative; its users noticed that it was less hygroscopic than alcohol with no loss in capacity to stain tissue [7]. Formaldehyde was also found to produce only marginal shrinkage and distortions of tissues compared to alcohol-fixed samples [8]. Various formaldehyde solutions remain to this day the staple for neurohistology in the neurosciences. Formaldehyde is a gas and is currently sold as a solution containing 37% by weight of the gas in water (10% neutral buffered formalin), which was once described by Blum [8] as a 40% formaldehyde working solution, or as a solid polymer, paraformaldehyde. To prevent spontaneous condensation reactions when stored in a concentrated form, small quantities of alcohols are added to the solution. In the 1930s it was determined that formaldehyde penetrates tissue rapidly but fixes slowly. Although the molecular mechanism of tissue fixation is not clear, formaldehyde is a reactive electrophilic species that cross-links with various biological macromolecules including proteins, glycoproteins, nucleic acids, and polysaccharides. In contrast to formaldehyde, when glutaraldehyde, a dialdehyde with a molecular-weight three times that of formaldehyde, is employed for preservation a much lower concentration is required. Glutaraldehyde has the advantage that most aldehyde groups in the solution are not bound into glycols, which affect penetration but not fixation [7]. This review provides a compilation of the effects of different concentration of formaldehyde fixation upon tissue reactions and shrinkage [7].

2

Fixatives for Tract Tracing, Histochemistry and Immunocytochemistry Despite extreme interest in the field of molecular biology there is still the need for proper brain fixation to generate well-preserved tissue samples for the application of tract-tracing, morphometric, histochemical, immunocytochemical and in situ hybridization neuroanatomical procedures. The literature contains several excellent reviews describing the principles underlying these methodologies [9–14]. However, the type of fixative used is somewhat variable, dependent upon the procedure or in the case of immunohistochemistry, upon the antigen under investigation. Fixation of any type preserves the ultrastructure and stabilizes protein and peptide conformation so that antibodies can bind to antigenic epitope sites. In situ perfusion limits the autofluorescence that can result from brain immersion fixation, and paraformaldehyde is superior to the other aldehydes for immunofluorescence detection, as it provides

52

Elliott J. Mufson et al.

the most brilliant fluorescence with minimal nonspecific staining. Of the fixatives available to the neuroscience community, paraformaldehyde prepared at 4% in a phosphate buffer solution is optimal for virtually all of the above neurohistological procedures. Other paraformaldehyde-based fixatives that are excellent for brain histology include Zamboni’s and Bouin’s. These fixatives are combined with picric acid, which stabilizes the paraformaldehyde, allowing for a long shelf life. In addition, acrolein (2-propenal), an unsaturated aldehyde, can be mixed with 4% paraformaldehyde for select immunocytochemical studies [15]. The following are protocols for generating paraformaldehyde fixatives: 4% Paraformaldehyde l Mix 4 g paraformaldehyde powder (available from Fisher, Electron Microscopy Science, and Sigma), 50 ml distilled H2O and 20 μl 10 N NaOH, and heat at 60  C for ~10–20 min in a fume hood. l

Do not overheat the solution past 70  C as the formaldehyde will vaporize and break down into formic acid and water, losing its optimal fixation properties.

l

Swirl to ensure all of the paraformaldehyde is in solution.

l

Add 50 ml of 0.2 M phosphate buffer in distilled water for a final volume of 100 ml.

l

Cool and filter with Whatman #4 filter paper.

l

Adjust final pH to 7.2 and store at 4  C.

Solution is good for ~1 week, although fresh paraformaldehyde is always preferred. Certain antigens require a gentler fixation for immunohistochemical visualization. In these cases a 2% paraformaldehyde solution is recommended. For the immunohistochemical staining of the neurotrophin nerve growth factor, a solution containing 2% paraformaldehyde (Fisher Sci., IL, USA) + 0.2% parabenzoquinone (Fisher Sci., IL, USA) in 0.1 M phosphate buffer is appropriate [16, 17]. Zamboni’s Fixative (according to [18]) l Mix 20 g paraformaldehyde with 150 ml double-filtered, saturated aqueous picric acid. l

Heat to 60  C in a fume hood.

l

Add 2.52% sodium hydroxide (w/v) in distilled water, drop by drop, until the solution is clear. N.B.: NaOH concentration is not critical, and for example, 10 M may also be used.

l

Filter solution with Whatman #4 filter paper and cool.

l

Make up to 1000 ml with phosphate buffer (3.31 g NaH2PO4·H2O, 33.77 g Na2HPO4·7H2O in 1000 ml distilled H2O).

Fixation Protocols for Neurohistology

53

Fixative is stable at 25  C for 12 months. Bouin’s Fixative: fixation time 6 h l 75 ml of saturated aqueous solution of picric acid l

25 ml of formalin (~40% aqueous solution of formaldehyde)

l

5 ml of glacial acetic acid Fixed tissue should be transferred to 70% alcohol.

3

Fixation Delivery Protocols

3.1 Transcardial Fixation

Transcardial perfusion under terminal anesthesia takes advantage of the subject’s circulatory system to deliver the fixative solution evenly throughout the body tissues, with optimal penetration of the brain. Transcardial perfusion consists of saline or phosphate buffer flush to remove blood followed by 4% paraformaldehyde mixed in phosphate buffer. The quantity of fixative varies depending upon the size of the animal used in the study. Saline flush can range from 30 ml for mice to a liter for nonhuman primates. Fixative amounts range from 50 ml for mice to 2 l for monkeys. Rats are usually well fixed with between 100 and 200 ml of the perfusate. Small quantities of glutaraldehyde (e.g., 0.05%) can be mixed with 4% paraformaldehyde solution to increase the hardness of the fixation. After transcardial fixation, brains (rodents and nonhuman primates) may be postfixed by immersion in the same or other fixative solution to also preserve external surfaces. Postfixation times are dependent upon specimen size (3 h for small brains [mouse], and up to 12 h for monkey brain) and type of fixative used.

3.2 Immersion Fixation

Immersion fixation of proteins follows an outside to inside gradient from the surface of the brain inward at a rate of about 1 mm/h at room temperature resulting in complete fixation after several hours for rodent brain but may take days to months for a larger brains [7, 19, 20]. However, prolonged fixation periods could compromise the antigenicity of proteins interior to the surface of the brain [21–23]. This is a problem in the area of postmortem human brain fixation where samples are stored for long periods of time in a formaldehyde solution. Moreover, prolonged storage in formalinbased fixatives may result in acidification, which further reduces both tissue quality and antigen detectability. To counter the antigen masking effects of formaldehyde, various antigen retrieval methods were developed, to fully reveal antibody reactivity particularly after long periods of fixation. The methods include pre-treatment of tissue sections with proteases [24], formic acid [25], ultrasound [26] or heating mounted sections in an ionic salt solution by microwave energy [21, 23, 27–33]. Despite these

54

Elliott J. Mufson et al.

potential caveats, immersion fixation remains a staple of neurohistology to maximize the use of selected tissue blocks obtained from highly valued brains. In the case of immersion fixation of a nonhuman subject, the animal is perfused with physiological saline, the brain removed from the calvarium and placed in the appropriate fixative solution for a specified time (e.g., 12–24 h for rodents) followed by immersion in a cryoprotectant solution as described below. Traditionally, whole human brain is stored in 10% formalin for months to years. More recently with the development of human brain banks for the study of aging and age-related neurodegenerative diseases modified fixation procedures have been introduced. For example, the human brain is hemisected or sliced into 1 cm thick slabs and stored in fixative for shorter periods of time. In our group, we store 1 cm thick human brain slabs in a 4% paraformaldehyde/phosphate buffer solution for 24–72 h and then transfer them to a cryoprotectant solution as described below. This procedure produces excellent microtome-cut tissue sections for immunocytochemistry, standard histological stains as well as single cell gene array technology [34–37].

4

Postfixation Cryoprotection Protocols An important variable in neuroanatomical studies of the CNS and, in particular, histochemical procedures that require tissue sections prepared without embedding (which is detrimental to many forms of histochemistry) is the prevention of “freezing artifact” or “freeze–thaw artifact.” This type of artifact appears as abnormal vacuoles in the tissue, most likely caused by ice crystal formation [38]. The presence of this form of artifact may leave tissues useless for many neuroanatomical methods. Freezing artifact can be obviated by postfixing tissue with various cryoprotectants, which reduce ice crystal formation.

4.1 Sucrose Cryoprotectant

Most neuroanatomical references suggest protecting against freezing artifact by infiltrating the tissue prior to freezing with a cryoprotectant of 30% sucrose in 0.1 M phosphate buffer [39].

4.2 Glycerol and Dimethylsulfoxide (DMSO) Cryoprotectant

For large blocks of brain tissue (e.g., >60 cm3), 30% sucrose infiltration provides variable results [37]. An alternative cryoprotectant is a mixture of glycerol and DMSO [38, 40, 41]. Rosene and coworkers [37] demonstrated that graded infiltration with up to 20% glycerol and 2% DMSO (in buffer or fixative) prevented freeze–thaw artifact in thick brain samples. In our laboratory, 1 cm thick human brain slabs are fixed in 4% paraformaldehyde, then transferred into a 10% glycerin (Fisher: cat. #BP229-4), 20% DMSO (Fisher: cat. #D128-1) solution for 48 h and finally into a 20% glycerin, 2% DMSO cryoprotectant solution for long-term

Fixation Protocols for Neurohistology

55

storage for at least up to 20 years and counting. After tissue blocks are sectioned on a sliding freezing microtome, they are stored in a solution consisting of 30% glycerin, 30% ethyl glycol (Fisher: cat. #E178-4) in 40% phosphate buffer. These cryoprotectants not only prevent crystallization during brain sectioning on a freezing sliding microtome but also eliminates the need to add antibacterial agents (e.g., azide) to sucrose-based cryoprotectant solutions for the longterm storage of brains and cut sections.

5

Fixation for Electron Microscopy In 1931, the German engineers Ernst Ruska and Max Knoll built the first electron microscope. Electron microscopy (EM) requires its own specimen fixative procedures to produce suitable tissue preservation. Chemical fixation for EM biological specimens stabilizes the tissues’ mobile macromolecular structure by chemical cross-linking of proteins with aldehydes such as formaldehyde and glutaraldehyde, and lipids with osmium tetroxide (OsO4). EM fixation involves a two-stage process; tissue is fixed with glutaraldehyde and/or paraformaldehyde followed by OsO4 treatment, which is highly toxic and can cause corneal damage. Extreme caution is due; it is recommended that eye protection is worn when using OsO4. This combinational fixative is important since glutaraldehyde is a poorly penetrating substance, whereas paraformaldehyde is a very effective tissue penetrant. Using them together allows the paraformaldehyde to rapidly penetrate the tissue because it is a monoaldehyde; the temporary and rapid fixation holds the tissue and membranes in place long enough to allow glutaraldehyde (a slowly penetrating dialdehyde) sufficient time to penetrate and irreversibly cross-link the proteins [42]. However, since crosslinking interferes with antigenicity of proteins, glutaraldehyde fixative should be used in low concentrations when tissue is immunohistochemically stained for EM. Osmium tetroxide is best used as a postfixative. Although OsO4 is an even slower penetrant than glutaraldehyde, its primary advantage is that it is a heavy metal salt, thus adding electron density to the tissue. It also preserves many lipids, most of which are not preserved by aldehydes. Hayat [43] states that “. . .fixation with glutaraldehyde without subsequent treatment with OsO4 is considered unsatisfactory for routine EM except for some cytochemical studies.” The primary reason that glutaraldehyde is preferred over paraformaldehyde for EM is that the protein cross-linking performed by glutaraldehyde is irreversible and involves randomly spaced primary amino acid groups, whereas that performed by paraformaldehyde is reversible and involves cross-linking of peptide chains. Several other reagents can be mixed with glutaraldehyde for EM fixation including acrolein, alcian blue, caffeine, hydrogen peroxide, lead acetate, phosphotungstic acid, and picric acid [43].

56

Elliott J. Mufson et al.

It is crucial when making up EM fixatives that the proper buffers are used. For example, using phosphate buffer is often problematic, and many investigators often include calcium chloride (CaCl2) in the buffer to preserve tissue osmolarity. When phosphate buffer is exposed to CaCl2, a fine precipitate forms, which resembles immunogold particles. Therefore, researchers often use cacodylate buffer (though it is a carcinogen). Also, phosphate buffer is incompatible with uranyl acetate (stains nucleic acids), which is another postfixative often used in combination with OsO4 (uranyl acetate treated tissue in phosphate buffer causes staining to “bleed”). Therefore, it is important to rinse tissue with a nonphosphate buffer before treatment with uranyl acetate. There has been renewed interest in performing EM on tissue harvested from human brain samples obtained at autopsy. Our group has been performing EM studies using tissue harvested from both aged normal and diseased brain including specimens from people with Alzheimer’s disease and mild cognitive impairment [44–47]. There are several important factors that should be considered in treating human tissue for EM and these are listed below in the order of importance: l

Limit air exposure (i.e., immersion in fixative immediately upon dissection).

l

Tissue thickness should be at 0.5 cm or less.

l

Limit postmortem interval (PMI) as quality tissue can be obtained up to 8 h postmortem.

l

Change solutions between 6 and 24 h.

l

Postfix in OsO4 as soon as feasible.

l

Keep time of dehydration as short as possible.

l

Embed in plastic (e.g., Epon or Araldite).

EM fixative for animal (transcardial) and human autopsy (immersion) fixation: 4% paraformaldehyde and 1% glutaraldehyde in 0.1 M phosphate buffer solution l

150 ml of 0.2 M phosphate buffer from a 2 l stock prepared as follows: – 1800 ml distilled H2O – 86.84 g phosphate dibasic heptahydrate (Na2HPO4·7H2O) – 10.48 g phosphate (NaH2PO4·H2O)

monobasic

monohydrate

– Enough distilled H2O to bring level to 2000 ml – Adjust pH to 7.4 l

150 ml of 8% paraformaldehyde in distilled H2O

Fixation Protocols for Neurohistology l

12 ml of 25% EM-grade glutaraldehyde

l

1.62 g dextrose

57

All solutions should be kept chilled (4 C) until time of use.

6

Molecular Biological Procedures Advances in molecular biology provide the tools needed to sample gene expression from specific homogeneous cell populations within defined brain regions without potential contamination of adjacent neuronal subtypes and nonneuronal cells. With the advent of modern molecular and cellular techniques, it is now possible to isolate and study genomic DNA, RNA species, and proteins from microdissected tissue sources [36, 48–51]. At present, an optimal methodology is to evaluate single cells, either identified physiologically in living preparations [52–56] or by immunocytochemical or histochemical procedures in fixed cells in vitro or in vivo [36, 49, 50, 57–62].

6.1 Gene Expression Profiling Using Fixed Tissues

Assessment of single cells and homogeneous cell populations in optimally prepared, perfused, fixed animal tissues as well as fixed postmortem human brain tissues is desirable. This is due to the abundance of animal and human brain tissues that are archived within individual laboratories and brain banks [63]. A variety of tissues and cells can be used to extract mRNA for gene profiling experiments. When employing mRNA as a starting material, one cannot overemphasize the importance of the preservation of RNA integrity. RNA species are particularly sensitive to degradation by ribonuclease (RNase). RNases are found in virtually every cell type [57, 64]. RNase-free precautions are essential for all genomicsbased tissue protocols. All biological samples require prompt handling, either through rapid RNA extraction, flash freezing, or fixation to minimize degradation. At present, no consensus protocol exists for the fixation and/or extraction of brain tissues obtained from animals or from postmortem human preparations. Several laboratories have evaluated the effects of different fixation protocols on RNA quality, ease of tissue microdissection, and success of high-throughput genomic assays (e.g., microarray and single nucleotide polymorphism [SNP] sequencing) [57, 60–62, 65–67]. Many variables, including antemortem characteristics, duration of fixation, and length of storage are relevant parameters that should be taken into consideration prior to the initiation of a study [58, 60, 68–71]. Caveats must be considered when undertaking molecular and cellular based studies in human postmortem brain tissue. One factor is postmortem interval (PMI), or the time that elapses between time of death and preservation of the tissue sample. PMI is particularly relevant when

58

Elliott J. Mufson et al.

obtaining postmortem human materials, as animal models can be fixed rapidly using perfusion techniques. Perfusion procedures are not available to accrue postmortem human tissues. Moreover, the choice of fixation for tissue preservation is an important factor affecting RNA stability. The means by which RNA is preserved is unknown but likely involves the inactivation of degradative enzymes. The choice of fixative must be balanced between optimizing tissue morphology and preserving nucleic acid integrity for evaluation. As reviewed by Van Deerlin and coworkers [57], ethanol-, neutral buffered formalin-, and 4% paraformaldehydebased fixatives provide optimal results for molecular-based studies. Another factor mainly related to human postmortem gene expression studies is the agonal state of the cases examined and the presence of overlapping neurologic conditions. Therefore, numerous variables, including antemortem characteristics, duration of fixation, and length of storage are relevant parameters that should be considered prior to the initiation of molecular studies [58, 61– 72]. 6.2 Choice of Fixative

Similar to other related biochemical procedures, there are various fixatives to choose from for preservation of tissue prior to molecular evaluation, such as 10% neutral buffered formalin (NBF), 4% paraformaldehyde, and 70% ethanol. Several investigations have examined the effects that the chemical composition of different fixatives has upon the integrity of the fixed tissue for downstream applications involving RNA analysis [57, 60–63, 65]. For example, amplification of mRNA by real-time PCR (RT-PCR) from paraffinembedded tissues was more effective when tissues were fixed using either ethanol, acetone, or OmniFix II, an alcohol-based fixative (Zymed Laboratories, South San Francisco, CA, USA), while results following formalin fixation have been equivocal [61, 65, 66, 73–75]. Overall, fixatives containing mercuric chloride and/or potassium dichromate, such as B-5 and Zenker’s, are more deleterious for nucleic acid isolation than milder fixatives such as formalin. The reason is that mercury and chromium have a high affinity for protein side chains, which leads to large metal-protein complexes that are relatively resistant to digestion and disaggregation [62, 66, 67]. Fixatives containing acid such as picric acid, acetic acid (Carnoy’s), and Bouin’s solution are also generally not compatible with nucleic acid analysis [62, 66, 67, 76]. Studies have shown that when using acridine orange histofluorescence (a fluorescent dye that intercalates into nucleic acids) brain tissue fixed in NBF and 70% ethanol buffered with 150 mM sodium chloride work optimally compared to Bouin’s fixation [76–79]. The introduction of novel methods for the assessment of gene expression in single cells poses unique challenges for optimal tissue fixation. In this regard, a study compared the effects of different fixatives on tissue morphology, ease of microdissection by laser

Fixation Protocols for Neurohistology

59

capture microdissection (LCM), and RNA integrity as determined by RT-PCR amplification [60]. This report compared the following fixatives on both frozen and paraffin-embedded mouse liver tissue: 70% ethanol, 95% ethanol, 10% NBF, and 3% paraformaldehyde. In addition, acetone and Rapid-Fix (75% methanol, 20% formaldehyde, 5% glacial acetic acid; Shandon, Pittsburgh, PA, USA) were used for frozen tissue and periodate–lysine–paraformaldehyde (PLP) fixative (2% paraformaldehyde, 7.5 mM L-lysine, and 10 mM sodium periodate) for paraffin embedded tissue [20]. Each fixative was compared for morphological quality, completeness of microdissection, and quantity of mRNA, as assessed by RT-PCR amplification of either β2-microglobulin or glyceraldehyde-3-phosphate-dehydrogenase [49]. The precipitative fixatives ethanol and acetone consistently produced more RT-PCR amplification product than formalin, although overall more amplification product was derived from frozen samples than from paraffin embedded fixed tissues. The use of 70% ethanol and PLP also produced good quality tissue morphology for the purpose of microdissection fixation for paraffin-embedded tissues [60]. Depending upon whether amplification by RT-PCR was the only criterion by which to evaluate mRNA quality, ethanol fixation would appear to be the best choice. However, when using in situ hybridization for the examination of cellular mRNA the use of 10% NBF or 4% paraformaldehyde fixation consistently provides optimal results [65, 80, 81]. In many situations more than a single method may be required to confirm the findings derived from a gene expression analysis including RT-PCR and in situ hybridization. Therefore, it is imperative to keep in mind that the use of more than one fixative should be considered when rare or irreplaceable tissue is procured for a study. 6.3 Duration of Fixation

Over fixation is detrimental to subsequent analysis of mRNA independent of the methodology employed. In most instances only relatively short target sequences, up to ~270 bases including microRNAs (miRNAs) can be amplified by RT-PCR from formalin-fixed, paraffin-embedded tissue [49, 60–62, 82–85]. Increased fixation time (greater than 18 h) decreases the size of the targets which can be amplified [86, 87]. The mechanism underlying this size limitation is most likely related either to degradation of the RNA before or during fixation or to the cross-linking effects of the fixative. Another possibility is that formalin produces a chemical modification of the RNA such as hydroxymethylation and methylene bridge formation [48]. Several studies have determined the effect of formalin on the chemical modification of RNA and its effect on amplification by RT-PCR [60, 86, 87]. For example, increased fixation time between 16 h and 7 days decreased the maximal size of RT-PCR product amplified. This decrease in maximal target size was not due to degradation of the RNA since gel electrophoresis showed total RNA to be intact.

60

Elliott J. Mufson et al.

Taken together, these findings support the concept that chemical modifications may be, at least in part, responsible for the inability to amplify long fragments by RT-PCR from RNA in fixed tissue. A few of these modifications may be reversed by heating extracted RNA species to 70  C or by digesting with proteinases [86]. However, prolonged incubation in formalin may make these chemical modifications irreversible. It is also possible that extended fixation may also partially degrade RNA. When performing expression profiling experiments, whenever possible, researchers should attempt to minimize the effects of fixation by selecting small target fragments to amplify via qPCR and/or microarray analysis or by using random primers to reverse transcribe the RNA.

7

Laser Capture Microdissection (LCM) Procedures The implementation of high-throughput microaspiration devices over the last few years has enabled rapid accession of single cells and homogeneous cellular populations for downstream genomic and proteomic analyses. Specifically, LCM is a strategy for acquiring pure cell populations from histochemically and/or immunocytochemically labeled cells from in vivo and in vitro sources [60, 88, 89]. LCM, a procedure developed originally at the National Institutes of Health [90, 91], and employed commercially by Arcturus (Thermofisher Scientific, South San Francisco, CA, USA), Leica Microsystems (Buffalo Grove, IL, USA) and Palm MicroBeam (Carl Zeiss, Jena, Germany), has become a widely used and reproducible technique. Two principal means of LCM include infrared laser capture and ultraviolet laser cutting (Figs. 1 and 2). When combined they provide a sensitive and specific means for the isolation of enriched cell populations. In addition, electrophysiology rigs can also be modified to aspirate cells from fixed tissue sections with minor modifications [36, 50]. Each method allows microdissected cells and their processes to be examined microscopically to confirm the identity and quality of isolated cell population(s) with minimal disruption of the surrounding neuropil [34, 93, 94]. RNA and DNA can be extracted from microdissected cells and utilized as input sources for downstream genomics applications such as microarray analysis (Fig. 3), RNA sequencing, and quantitative real-time PCR (qRT-PCR) [89, 94, 97]. LCM has also been increasingly utilized to collect cells for downstream proteomic analyses including two-dimensional gel electrophoresis, mass spectroscopy-based methods, and antibody-based protein chips [50, 51, 98–101]. Tissue used for LCM can be either fresh frozen or fixed in alcohol or aldehydes [50, 60]. If the samples are intended for downstream RNA analysis, RNase-free techniques should also be observed at every step of the experimental process to reduce naturally occurring RNases [64]. Please note that the

Fixation Protocols for Neurohistology

A

61

B Thermolabile polymer on bottom surface of cap

Infrared laser melts themolabile polymer

Cells of interest bound in polymercell composite

Properly melted polymer spots

Poor spot

Capture force Sheer force

Photo courtesy of Arcturus Biosciences

Substratum force

Fig. 1 Laser capture microdissection using a near-infra red (IR) laser source and a thermoplastic cap for collection. (A) An inert RNase-free plastic cap is placed above sample preparation. A near-IR laser is pulsed (rectangle with a dotted outline), activating the thermoplastic film on the cap, which adheres to cells of interest. When the cap is lifted, the cells of interest are removed from the tissue section in a precise manner. Physical forces involved are shown. (B) Micrographs of the melted polymer spots and a representative microdissected single cell. DNA, RNA, and proteins can be extracted from the microdissected cells for further downstream genomic and/or proteomic analyses. Images reprinted from Ref. [92] by permission of © Nature Publishing Group

required water purity (converting to resistivity of 18.2 MΩ cm, Fig. 3A) applies to 25  C, and the corresponding value at 20  C should be close to 24 MΩ cm. Gloves should be worn at all times, and all work areas and instrumentation should be wiped down with a solution that de-activates the RNases, such as RNase AWAY (Sigma-Aldrich, St. Louis, MO, USA). Regardless of the source of the tissue, human or animal, the specimen should be fixed as quickly as possible in order to prevent degradation of biomolecules, changes caused by postmortem enzyme activation (e.g., nucleases and proteases) [102]. In addition, cellular pathways become activated as a result of excision from its host, and this may have an impact on the biological profiles. To stabilize the biomolecules and inactivate enzyme activity, rapid fixation of tissues is required, either by freezing or placing them into chemical fixatives such as 4% paraformaldehyde or 10% neutral buffered formalin (NBF).

62

Elliott J. Mufson et al.

A

B

C

Fig. 2 Laser capture microdissection using an ultraviolet (UV) laser source and gravity collection. (A) Neuron identification for microaspiration. (B) A UV laser source precisely cuts around the desired neuron in the tissue section with negligible damage to the tissue. (C) The dissected material lands safely in the microfuge tube cap through gravity. The sample is now ready for extraction in the microfuge tube for downstream analyses. Bubbles on the right panel are higher power insets of the images on the left panel. Used with permission from © Barrow Neurological Institute (Phoenix, AZ, USA) 7.1

Cryofixation

When fixing tissue samples for LCM by freezing, it is recommended to embed the tissue in a matrix such as optimal cutting temperature compound or OCT (Sakura Finetek, Hatfield, PA, USA). Embedding the tissue prevents freezing artifacts from forming in the tissue during storage, and makes the tissue sectioning process easier. Once embedded in OCT, the sample is ready to be frozen. While it is common to use liquid nitrogen to freeze tissues for homogenate preparations, this method is not recommended when frozen tissues will be sectioned. The cryostat should also be avoided as a

Fixation Protocols for Neurohistology

PARAFFIN REMOVAL

XYLENE REMOVAL

C

B

A

63

REHYDRATION

*

Place stained tissue slide onto LCM apparatus for single cell (cell population) microdissection of labeled cells

D

Immerse slides in xylene (2x)

Grade through alcohols: 100% (2x) 95% (1x) 70% (1x) 50% (1x)

Rinse in RNAse-free, ultra-pure, deionized

Remove slide from coplin jar and apply histochemical or immunostaining solutions to visualize selected cell types

High-density cDNA microarray

water (3x) Custom-designed cDNA array RNA amplification or cDNA library construction

E

RNA sequencing (analysis of variance)

Fig. 3 Diagram showing an optimized staining protocol for formalin-fixed paraffin-embedded tissue sections for use in laser-capture microdissection (LCM). (A, B) Deparaffinization and staining. Subsequent (∗) dehydration of the stained section (in an increasing ethanol series), and immersion in xylene are not shown for brevity. (C) Microdissection. The gel shown next to the microfuge (Eppendorf) tube is a control demonstrating the specificity of the amplified RNA obtained from the dissected cell. (D) High-density cDNA microarray generated from neurofibrillary tangles-bearing (red) and normal (green) hippocampal (CA1) neurons of Alzheimer’s disease patients; yellow (an overlap of red and green) indicates expression in either cell type of a particular gene, each represented by a single spot. Custom-designed cDNA array (following hybridization with radiolabeled antisense RNA [aRNA] from single hippocampal CA1 neurons) represents expression profiles of different proteins. (E) Coefficient of variation plotted against expression level, obtained from mRNA isolated from single hippocampal CA1 neurons. Figures in D are reproduced from Refs. [34, 50, 95] by permission of © Springer and © Wiley, figure E is reproduced from Ref. [96])

means of freezing tissues, as the cryostat does not freeze the tissue rapidly enough to maintain biomolecular stability or to inactive enzyme activity. To optimally freeze tissues in OCT, the embedded tissue should be placed on top of crushed dry ice or in bath of cooled isopentane (using either dry ice or liquid nitrogen). Once the tissue has been frozen, it can be sectioned or stored at 70  C or below. Liquid nitrogen is not suitable (due to the gaseous N2 layer acting as thermal insulation, thus slowing down the freezing process). 7.2 Aldehyde Fixation

Formalin-fixation and subsequent paraffin embedding has long been a standard means of preserving tissue samples and is a popular specimen type for studies using LCM. Along with the preservation benefits of tissue fixation, paraffin embedding serves as a supporting

64

Elliott J. Mufson et al.

matrix making it easy to cut very thin and consistent sections. Various types of fixatives have been used to process tissue into paraffin blocks [60, 83, 85]. Although the most common fixative used in pathology labs is 10% NBF, most neuroscience laboratories use 4% paraformaldehyde [36, 76]. Formalin fixes the tissue by cross-linking macromolecules within, which in turn modifies and degrades biomolecules, primarily RNA, and as a result presents challenges to their extraction and downstream processing. Until recently commercial products were unavailable for the difficult extraction and processing of RNA from formalin-fixed tissues. Several products now exist that allow researchers to deal with formalin-fixed tissues and process them for various downstream analysis methods, such as gene expression (Paradise® Plus Reagent System, MDS Analytical Technologies) or protein analysis (Liquid Tissue®, Expression Pathology, Rockville, MD, USA). Optimally, tissue blocks to be formalin-fixed and paraffin-embedded should not be greater than 5 mm in thickness and should be fixed at room temperature for no more than 24 h. Once the tissue is fixed in formalin it should proceed immediately to processing into paraffin blocks. LCM has also been increasingly utilized to collect cells for downstream proteomic analysis and RNA seq, which requires careful handling and many cells for a successful assay [51, 101]. 7.3

Tissue Staining

Fixed paraffin-embedded or frozen tissue sections mounted onto a room temperature slide (stored at 70  C) are now ready (after paraffin removal by xylene, and rehydration) for staining and dehydration (Fig. 3). Staining of the tissue sections is important as it facilitates the visualization of cells during the microdissection process. However, optimal staining is inversely related to optimal biomolecule preservation. Standard histological staining protocols often entail prolonged protocols with extended aqueous steps, both of which increase opportunity for the degradation of biomolecules, primarily RNA. Both histochemical and immunohistochemical protocols can be compatible with many downstream analysis methods for LCM samples [36, 92, 94, 103–105], and there are several commercially available staining products optimized for use with LCM and downstream gene expression analysis (e.g., HistoGene® Frozen and ImmunoFluorescence Kits, MDS Analytical Technologies). Moreover, the choice of which histological stain to use for histochemical detection is important for gene expression studies in brain. Our group determined that of five histochemical stains (cresyl violet, thionin, hematoxylin and eosin, silver stain, and acridine orange) used in combination with an expression profiling paradigm that included regional and single cell analyses within the hippocampus of postmortem human brains and adult mice, three were viable and two yielded poor results [76]. Specifically, cresyl violet, thionin, hematoxylin and eosin performed similarly (and also did not differ from neurofilament

Fixation Protocols for Neurohistology

65

immunocytochemistry) in terms of gene expression profile detection from adjacent sections. In contrast, both silver stain and acridine orange, two reagents that intercalate into nucleic acids, did not enable microarray analysis [76]. Common to all LCM staining procedures are their short staining times, a requirement to maintain the biomolecule integrity, especially RNA. While not shown in Fig. 3, graded dehydration steps are typical in LCM staining protocols, as the ethanol removes all excess water and lipids from the stained tissue and helps to maintain the tissue morphology. Xylene is a typical final step, and serves to remove the ethanol from the tissue.

8

Summary In summary, the use of fixed tissues has enabled exciting hypothesis testing in postmortem human brain tissues as well as in relevant animal and cellular models of neurodevelopmental, neurodegenerative and neuropsychiatric disorders. Several histochemical stains, either alone or in combination with immunocytochemical preparations, can be employed in association with modern microaspiration techniques, such as LCM, as well as with downstream genomic and proteomic applications. Although there are several caveats and overall limitations to RNA and protein detection in fixed tissues, the pitfalls are outweighed by the research benefits when investigators employ these tissues and techniques properly and in the context of known work.

Acknowledgments National Institutes of Health grants AG014449, AG043375, and AG107617; the Alzheimer’s Association and Barrow Neurological Institute Barrow Beyond supported this work. We thank Megan Gautier for technical assistance. References 1. Gere C (2003) A brief history of brain archiving. J Hist Neurosci 12(4):396–410. https:// doi.org/10.1076/jhin.12.4.396.27916 2. Pole T (1813) The anatomical instructor. Smith & Davy, London 3. Reil JC (1809) Arch Physiol 9:136–146, 147–171, 172–195, 195–208. https://bio diversitylibrary.org/page/13752924 4. Schmahmann JD, Pandya DN (2007) Cerebral white matter - historical evolution of facts and notions concerning the organization of

the fiber pathways of the brain. J Hist Neurosci 16(3):237–267. https://doi.org/10. 1080/09647040500495896 5. Clarke E, O’Malley CD (1968) The human brain and spinal cord. A historical study illustrated by writings from antiquity to the twentieth century. University of California Press, Berkeley 6. Barbara JG (2007) Louis Ranvier (1835–1922): the contribution of microscopy to physiology and the renewal of French

66

Elliott J. Mufson et al.

general anatomy. J Hist Neurosci 16:413–431. https://doi.org/10.1080/ 09647040600685503 7. Fox CH, Johnson FB, Whiting J, Roller PP (1985) Formaldehyde fixation. J Histochem Cytochem 33(8):845–853. https://doi.org/ 10.1177/33.8.3894502 8. Blum F (1893) Der Formaldehyd als H€artungsmittel. Zeitschrift fu¨r Wissenschaftliche Mikroskopie und fu¨r Mikroskopische Technik 10:314–315. https://bio diversitylibrary.org/page/4637180 9. Mesulam MM (1982) Tracing neural connections with horseradish peroxidase. International Brain Research Organisation, New York. ISBN: 978-0471100294 10. Palay VC, Palay SL (1982) Cytochemical methods in neuroanatomy, Neurology and neurobiology, vol 1. Alan R. Liss, New York 11. Berger SL, Kimmel AR (eds) (1987) Guide to molecular cloning techniques, Methods in enzymology, vol 152. Academic (Elsevier), Orlando, FL. ISBN: 978-0121817756 12. Gerfen CR (1997) Basic neuroanatomical methods. In: Current protocols in neuroscience (core publication, unit 1.1). Wiley, New York, p 11. https://doi.org/10.1002/ 0471142301.ns0101s23 13. Volpicelli-Daley LA, Levey A (2003) Immunohistochemical localization of proteins in the nervous system. In: Current protocols in neuroscience (Suppl 25, unit 1.2). Wiley, New York, p 17. https://doi.org/10.1002/ 0471142301.ns0102s25 14. Zaborszky L, Wouterlood FG, Lanciego JL (2006) Neuroanatomical tract-tracing 3. Molecules, neurons, and systems. Springer, New York. https://doi.org/10.1007/0-38728942-9 15. Gibbs RB (1996) Expression of estrogen receptor-like immunoreactivity by different subgroups of basal forebrain cholinergic neurons in gonadectomized male and female rats. Brain Res 720(1–2):61–68. https://doi.org/ 10.1016/0006-8993(96)00106-0 16. Conner JM, Muir D, Varon S, Hagg T, Manthorpe M (1992) The localization of nerve growth factor-like immunoreactivity in the adult rat basal forebrain and hippocampal formation. J Comp Neurol 319(3):454–462. https://doi.org/10.1002/cne.903190310 17. Mufson EJ, Conner JM, Varon S, Kordower JH (1994) Nerve growth factor-like immunoreactive profiles in the primate basal forebrain and hippocampal formation. J Comp Neurol 341(4):507–519. https://doi.org/10.1002/ cne.903410407

18. Stefanini M, De Martino C, Zamboni L (1967) Fixation of ejaculated spermatozoa for electron microscopy. Nature 216 (5111):173–174. https://doi.org/10.1038/ 216173a0 19. Kiernan JA (2000) Formaldehyde, formalin, paraformaldehyde and glutaraldehyde: what they are and what they do. Microscopy Today 8(1):8–12. https://doi.org/10. 1017/S1551929500057060 20. Kiernan JA (2015) Histological and histochemical methods: theory and practice, 5th edn. Scion, Banbury, UK. ISBN: 978-1907904325 21. Pollock NJ, Wood JG (1988) Differential sensitivity of the microtubule-associated protein, tau, in Alzheimer’s disease tissue to formalin fixation. J Neurol Neurosurg Psychiatry 51 (11):1387–1390. https://doi.org/10.1177/ 36.9.2841371 22. Riederer BM, Porchet R, Marugg RA, Binder LI (1993) Solubility of cytoskeletal proteins in immunohistochemistry and the influence of fixation. J Histochem Cytochem 41 (4):609–616. https://doi.org/10.1177/41. 4.8450200 23. Vickers JC, Riederer BM, Marugg RA, Bue´eScherrer V, Delacourte A, Morrison JH (1994) Alterations in neurofilament protein immunoreactivity in human hippocampal neurons related to normal aging and Alzheimer’s disease. Neuroscience 62(1):1–13. https://doi.org/10.1016/0306-4522(94) 90310-7 24. Battifora H, Kopinski M (1986) The influence of protease digestion and duration of fixation on the immunostaining of keratins. A comparison of formalin and ethanol fixation. J Histochem Cytochem 34(8):1095–1100. https://doi.org/10.1177/34.8.2426335 25. Kitamoto T, Ogomori K, Tateishi J, Prusiner SB (1987) Formic acid pretreatment enhances immunostaining of cerebral and systemic amyloids. Lab Investig 57(2):230–236. 2441141 26. Podkletnova I, Alho H (1993) Ultrasoundamplified immunohistochemistry. J Histochem Cytochem 41(1):51–56. https://doi. org/10.1177/41.1.8417112 27. Shi SR, Imam SA, Young L, Imam SA, Taylor CR (1996) Use of pH 9.5 Tris-HCl buffer containing 5% urea for antigen retrieval immunohistochemistry. Biotech Histochem 71(4):190–196. https://doi.org/10.3109/ 10520299609117158 28. Shiurba RA, Spooner ET, Ishiguro K, Takahashi M, Yoshida R, Wheelock TR,

Fixation Protocols for Neurohistology Imahori K, Cataldo AM, Nixon RA (1998) Immunocytochemistry of formalin-fixed human brain tissues: microwave irradiation of free-floating sections. Brain Res Protocol 2(2):109–119. https://doi.org/10.1016/ S1385-299X(97)00029-9 29. Gown AM, de Wever N, Battifora H (1993) Microwave-based antigenic unmasking: a revolutionary new technique for routine immunohistochemistry. Appl Immunohistochem 1:256–266 30. Schneider DR, Felt BT, Goldman H (1982) On the use of microwave radiation energy for brain tissue fixation. J Neurochem 38 (3):749–752. https://doi.org/10.1111/j. 1471-4159.1982.tb08694.x 31. Cummings BJ, Mason AJ, Kim RC, Sheu PCY, Anderson AJ (2002) Optimization of techniques for the maximal detection and quantification of Alzheimer’s-related neuropathology with digital imaging. Neurobiol Aging 23(2):161–170. https://doi.org/10. 1016/S0197-4580(01)00316-5 32. D’Andrea MR, Reiser PA, Polkovitch DA, Gumula NA, Branchide B, Hertzog BM, Schmidheiser D, Belkowski S, Gastard MC, Andrade-Gordon P (2003) The use of formic acid to embellish amyloid plaque detection in Alzheimer’s disease tissues misguides key observations. Neurosci Lett 342 (1–2):114–118. https://doi.org/10.1016/ S0304-3940(03)00252-0 33. Long DJ II, Buggs C (2008) Microwave oven-based technique for immunofluorescent staining of paraffin-embedded tissues. J Mol Histol 39(1):1–4. https://doi.org/10.1007/ s10735-007-9093-6 34. Ginsberg SD, Che S, Counts SE, Mufson EJ (2006) Single cell gene expression profiling in Alzheimer’s disease. NeuroRx 3(3):302–318. https://doi.org/10.1016/j.nurx.2006.05. 007 35. Ginsberg SD, Alldred MJ, Che S (2012) Gene expression levels assessed by CA1 pyramidal neuron and regional hippocampal dissections in Alzheimer’s disease. Neurobiol Dis 45 (1):99–107. https://doi.org/10.1016/j. nbd.2011.07.013 36. Ginsberg SD (2014) Considerations in the use of microarrays for analysis of the CNS. In: Reference module in biomedical sciences 2014 (introduced by Richerson GB). Elsevier, Amsterdam, p 7. https://doi.org/10.1016/ B978-0-12-801238-3.04662-6 37. Ginsberg SD, Che S (2014) Methods and compositions for amplification and detection of microRNAs (miRNAs) and noncoding RNAs (ncRNAs) using the signature sequence

67

amplification method (SSAM). Recent Adv DNA Gene Seq 8(1):2–9. https://doi.org/ 10.2174/2352092208666141001154206 38. Rosene DL, Roy NJ, Davis BJ (1986) A cryoprotection method that facilitates cutting frozen sections of whole monkey brains for histological and histochemical processing without freezing artifact. J Histochem Cytochem 34(10):1301–1315. https://doi.org/ 10.1177/34.10.3745909 39. Rosene DL, Mesulam MM (1978) Fixation variables in horseradish peroxidase neurohistochemistry. I. The effect of fixation time and perfusion procedures upon enzyme activity. J Histochem Cytochem 26(1):28–39. https://doi.org/10.1177/26.1.413864 40. Meryman HT (1971) Cryoprotective agents. Cryobiology 8(2):173–183. https://doi.org/ 10.1016/0011-2240(71)90024-1 41. Meryman HT (1974) Freezing injury and its prevention in living cells. Annu Rev Biophys Bioeng 3:341–363. https://doi.org/10. 1146/annurev.bb.03.060174.002013 42. Karnovsky MJ (1965) A formaldehydeglutaraldehyde fixative of high osmolarity for use in electron microscopy. 5th Annual meeting of the American Society for Cell Biology, Philadelphia, 10–12 Nov 1965. J Cell Biol 27(2):137A–138A. https://jstor.org/sta ble/1604673 43. Hayat MA (2000) Principles and techniques of electron microscopy. Biological applications, 4th edn. Cambridge University Press, Cambridge, UK. ISBN: 978-0521632874 44. Scheff SW, Price DA, Schmitt FA, Mufson EJ (2006) Hippocampal synaptic loss in early Alzheimer’s disease and mild cognitive impairment. Neurobiol Aging 27 (10):1372–1384. https://doi.org/10.1016/ j.neurobiolaging.2005.09.012 45. Scheff SW, Price DA, Schmitt FA, DeKosky ST, Mufson EJ (2007) Synaptic alterations in CA1 in mild Alzheimer’s disease and mild cognitive impairment. Neurology 68 (18):1501–1508. https://doi.org/10.1212/ 01.wnl.0000260698.46517.8f 46. Scheff SW, Price DA, Schmitt FA, Roberts KN, Ikonomovic MD, Mufson EJ (2013) Synapse stability in the precuneus early in the progression of Alzheimer’s disease. J Alzheimers Dis 35(3):599–609. https://doi.org/ 10.3233/JAD-122353 47. Scheff SW, Price DA, Ansari MA, Roberts KN, Schmitt FA, Ikonomovic MD, Mufson EJ (2015) Synaptic change in the posterior cingulate gyrus in the progression of Alzheimer’s disease. J Alzheimers Dis 43

68

Elliott J. Mufson et al.

(3):1073–1090. https://doi.org/10.3233/ JAD-141518 48. Crisan D, Mattson JC (1993) Retrospective DNA analysis using fixed tissue specimens. DNA Cell Biol 12(5):455–464. https://doi. org/10.1089/dna.1993.12.455 49. Krafft AE, Duncan BW, Bijwaard KE, Taubenberger JK, Lichy JH (1997) Optimization of the isolation and amplification of RNA from formalin-fixed, paraffin-embedded tissue: the Armed Forces Institute of Pathology experience and literature review. Mol Diagn 2 (3):217–230. https://doi.org/10.1016/ S1084-8592(97)80032-X 50. Ginsberg SD, Hemby SE, Mufson EJ, Martin LJ (2006) Cell and tissue microdissection in combination with genomic and proteomic applications. In: Zaborszky L, Wouterlood FG, Lanciego JL (eds) Neuroanatomical tract tracing 3: Molecules, neurons, and systems. Springer, New York, pp 109–141. https://doi.org/10.1007/0-387-28942-9_4 51. Drummond ES, Nayak S, Ueberheide B, Wisniewski T (2015) Proteomic analysis of neurons microdissected from formalin-fixed, paraffin-embedded Alzheimer’s disease brain tissue. Sci Rep 5:15456. https://doi.org/10. 1038/srep15456 52. Brail LH, Jang A, Billia F, Iscove NN, Klamut HJ, Hill RP (1999) Gene expression in individual cells: analysis using global single cell reverse transcription polymerase chain reaction (GSC RT-PCR). Mutat Res 406 (2–4):45–54. https://doi.org/10.1016/ S1383-5726(98)00009-0 53. Crino PB, Trojanowski JQ, Dichter MA, Eberwine J (1996) Embryonic neuronal markers in tuberous sclerosis: single-cell molecular pathology. Proc Natl Acad Sci U S A 93 (24):14152–14157. https://doi.org/10. 1073/pnas.93.24.14152 54. Eberwine J, Spencer C, Miyashiro K, Mackler S, Finnell R (1992) Complementary DNA synthesis in situ: methods and applications. Methods Enzymol 216:80–100. https://doi.org/10.1016/0076-6879(92) 16011-8 55. Brooks-Kayal AR, Jin H, Price M, Dichter MA (1998) Developmental expression of GABA(A) receptor subunit mRNAs in individual hippocampal neurons in vitro and in vivo. J Neurochem 70(3):1017–1028. https://doi.org/10.1046/j.1471-4159. 1998.70031017.x 56. Kacharmina JE, Crino PB, Eberwine J (1999) Preparation of cDNA from single cells and subcellular regions. Methods Enzymol

303:3–18. https://doi.org/10.1016/ S0076-6879(99)03003-7 57. Van Deerlin VM, Ginsberg SD, Lee VMY, Trojanowski JQ (2002) The use of fixed human postmortem brain tissue to study mRNA expression in neurodegenerative diseases: applications of microdissection and mRNA amplification. In: Geschwind DH, Gregg JP (eds) Microarrays for the neurosciences. MIT Press, Cambridge, MA, pp 201–235. ISBN: 0262072297 58. Dorph-Petersen KA, Pierri JN, Perel JM, Sun Z, Sampson AR, Lewis DA (2005) The influence of chronic exposure to antipsychotic medications on brain size before and after tissue fixation: a comparison of haloperidol and olanzapine in macaque monkeys. Neuropsychopharmacology 30:1649–1661. https://doi.org/10.1038/sj.npp.1300710 59. Tecott LH, Barchas JD, Eberwine JH (1988) In situ transcription: specific synthesis of complementary DNA in fixed tissue sections. Science 240:1661–1664. https://doi.org/10. 1126/science.2454508 60. Goldsworthy SM, Stockton PS, Trempus CS, Foley JF, Maronpot RR (1999) Effects of fixation on RNA extraction and amplification from laser capture microdissected tissue. Mol Carcinog 25(2):86–91. https://doi.org/10. 1002/(SICI)1098-2744(199906)25:23.0.CO;2-4 61. Koopmans M, Monroe SS, Coffield LM, Zaki SR (1993) Optimization of extraction and PCR amplification of RNA extracts from paraffin-embedded tissue in different fixatives. J Virol Methods 43(2):189–204. https://doi. org/10.1016/0166-0934(93)90076-4 62. Foss RD, Guha-Thakurta N, Conran RM, Gutman P (1994) Effects of fixative and fixation time on the extraction and polymerase chain reaction amplification of RNA from paraffin-embedded tissue. Comparison of two housekeeping gene mRNA controls. Diagn Mol Pathol 3(3):148–155. https:// journals.lww.com/molecularpathology/toc/ 1994/09000 63. Shibata D, Kurosu M, Noguchi TT (1991) Fixed human tissues: a resource for the identification of individuals. J Forensic Sci 36 (4):1204–1212 64. Farrell RE Jr (2010) RNA methodologies, 4th edn. Academic, London. ISBN: 9780123747273 65. Weiss LM, Chen YY (1991) Effects of different fixatives on detection of nucleic acids from paraffin-embedded tissues by in situ hybridization using oligonucleotide probes. J

Fixation Protocols for Neurohistology Histochem Cytochem 39(9):1237–1242. https://doi.org/10.1177/39.9.1918942 66. Ben-Ezra J, Johnson DA, Rossi J, Cook N, Wu A (1991) Effect of fixation on the amplification of nucleic acids from paraffinembedded material by the polymerase chain reaction. J Histochem Cytochem 39 (3):351–354. https://doi.org/10.1177/39. 3.1704393 67. Greer CE, Peterson SL, Kiviat NB, Manos MM (1991) PCR amplification from paraffin-embedded tissues. Effects of fixative and fixation time. Am J Clin Pathol 95 (2):117–124. https://doi.org/10.1093/ ajcp/95.2.117 68. Barton AJ, Pearson RC, Najlerahim A, Harrison PJ (1993) Pre- and postmortem influences on brain RNA. J Neurochem 61 (1):1–11. https://doi.org/10.1111/j.14714159.1993.tb03532.x 69. Leonard S, Logel J, Luthman D, Casanova M, Kirch D, Freedman R (1993) Biological stability of mRNA isolated from human postmortem brain collections. Biol Psychiatry 33 (6):456–466. https://doi.org/10.1016/ 0006-3223(93)90174-C 70. Coombs NJ, Gough AC, Primrose JN (1999) Optimisation of DNA and RNA extraction from archival formalin-fixed tissue. Nucleic Acids Res 27(16):e12. https://doi.org/10. 1093/nar/27.16.e12-i 71. Roberts L, Bowers J, Sensinger K, Lisowski A, Getts R, Anderson MG (2009) Identification of methods for use of formalin-fixed, paraffinembedded tissue samples in RNA expression profiling. Genomics 94(5):341–348. https:// doi.org/10.1016/j.ygeno.2009.07.007 72. Ordway GA, Szebeni A, Duffourc MM, Dessus-Babus S, Szebeni K (2009) Gene expression analyses of neurons, astrocytes, and oligodendrocytes isolated by laser capture microdissection from human brain: detrimental effects of laboratory humidity. J Neurosci Res 87(11):2430–2438. https://doi.org/10. 1002/jnr.22078 73. Counts SE, Chen EY, Che S, Ikonomovic MD, Wuu J, Ginsberg SD, Dekosky ST, Mufson EJ (2006) Galanin fiber hypertrophy within the cholinergic nucleus basalis during the progression of Alzheimer’s disease. Dement Geriatr Cogn Disord 21 (4):205–214. https://doi.org/10.1159/ 000090906 74. Counts SE, He B, Che S, Ginsberg SD, Mufson EJ (2008) Galanin hyperinnervation upregulates choline acetyltransferase expression in cholinergic basal forebrain neurons in Alzheimer’s disease. Neurodegener Dis 5

69

(3–4):228–331. https://doi.org/10.1159/ 000113710 75. Counts SE, He B, Che S, Ginsberg SD, Mufson EJ (2009) Galanin fiber hyperinnervation preserves neuroprotective gene expression in cholinergic basal forebrain neurons in Alzheimer’s disease. J Alzheimers Dis 18 (4):885–896. https://doi.org/10.3233/ JAD-2009-1196 76. Ginsberg SD, Che S (2004) Combined histochemical staining, RNA amplification, regional, and single cell cDNA analysis within the hippocampus. Lab Investig 84:952–962. https://doi.org/10.1038/labinvest. 3700110 77. Ginsberg SD, Crino PB, Hemby SE, Weingarten JA, Lee VMY, Everwine JH, Trojanowski JQ (1999) Predominance of neuronal mRNAs in individual Alzheimer’s disease senile plaques. Ann Neurol 45(2):174–181. https://doi.org/10.1002/1531-8249( 199902)45:23.0.CO;2E 78. Ginsberg SD, Crino PB, Lee VMY, Eberwine JH, Trojanowski JQ (1997) Sequestration of RNA in Alzheimer’s disease neurofibrillary tangles and senile plaques. Ann Neurol 41 (2):200–209. https://doi.org/10.1002/ana. 410410211 79. Ginsberg SD, Galvin JE, Chiu T-S, Lee VMY, Masliah E, Trojanowski JQ (1998) RNA sequestration to pathological lesions of neurodegenerative disorders. Acta Neuropathol 96(5):487–494. https://doi.org/10.1007/ s004010050923 80. Tournier I, Bernuau D, Poliard A, Shoevaert D, Feldmann G (1987) Detection of albumin mRNAs in rat liver by in situ hybridization: usefulness of paraffin embedding and comparison of various fixation procedures. J Histochem Cytochem 35 (4):453–459. https://doi.org/10.1177/35. 4.3546490 81. Eastwood SL, Harrison PJ (1999) Detection and quantification of hippocampal synaptophysin messenger RNA in schizophrenia using autoclaved, formalin-fixed, paraffin wax-embedded sections. Neuroscience 93 (1):99–106. https://doi.org/10.1016/ S0306-4522(99)00096-2 82. Mizuno T, Nagamura H, Iwamoto KS, Ito T, Fukuhara T, Tokunaga M, Tokuoka S, Mabuchi K, Seyama T (1998) RNA from decades-old archival tissue blocks for retrospective studies. Diagn Mol Pathol 7 (4):202–208. https://doi.org/10.1097/ 00019606-199808000-00004

70

Elliott J. Mufson et al.

83. Stanta G, Bonin S (1998) RNA quantitative analysis from fixed and paraffin-embedded tissues: membrane hybridization and capillary electrophoresis. BioTechniques 24 (2):271–276. https://doi.org/10.2144/ 98242st04 84. Muciaccia B, Vico C, Aromatario M, Fazi F, Cecchi R (2015) Molecular analysis of different classes of RNA molecules from formalinfixed paraffin-embedded autoptic tissues: a pilot study. Int J Legal Med 129(1):11–21. https://doi.org/10.1007/s00414-0141066-1 85. Takagi H, Shibutani M, Kato N, Fujita H, Lee KY, Takigami S, Mitsumori K, Hirose M (2004) Microdissected region-specific gene expression analysis with methacarn-fixed, paraffin-embedded tissues by real-time RT-PCR. J Histochem Cytochem 52(7):903–913. https://doi.org/10.1369/jhc.3A6215.2004 86. Masuda N, Ohnishi T, Kawamoto S, Monden M, Okubo K (1999) Analysis of chemical modification of RNA from formalin-fixed samples and optimization of molecular biology applications for such samples. Nucleic Acids Res 27(22):4436–4443. https://doi.org/10.1093/nar/27.22.4436 87. Greer CE, Lund JK, Manos MM (1991) PCR amplification from paraffin-embedded tissues: recommendations on fixatives for long-term storage and prospective studies. PCR Methods Appl 1(1):46–50. https://doi.org/10. 1101/gr.1.1.46 88. Dolter KE, Braman JC (2001) Small-sample total RNA purification: laser capture microdissection and cultured cell applications. BioTechniques 30(6):1358–1361. https://doi. org/10.2144/01306pf01 89. Ehrig T, Abdulkadir SA, Dintzis SM, Milbrandt J, Watson MA (2001) Quantitative amplification of genomic DNA from histological tissue sections after staining with nuclear dyes and laser capture microdissection. J Mol Diagn 3(1):22–25. https://doi.org/10. 1016/S1525-1578(10)60645-9 90. Bonner RF, Emmert-Buck M, Cole K, Pohida T, Chuaqui R, Goldstein S, Liotta LA (1997) Laser capture microdissection: molecular analysis of tissue. Science 278 (5342):1481–1483. https://doi.org/10. 1126/science.278.5342.1481 91. Emmert-Buck MR, Bonner RF, Smith PD, Chuaqui RF, Zhuang Z, Goldstein SR, Weiss RA, Liotta LA (1996) Laser capture microdissection. Science 274(5289):998–1001. https://doi.org/10.1126/science.274.5289. 998

92. Espina V, Wulfkuhle JD, Calvert VS, VanMeter A, Zhou W, Coukos G, Geho DH, Petricoin EF III, Liotta LA (2006) Laser-capture microdissection. Nat Protoc 1 (2):586–603. https://doi.org/10.1038/ nprot.2006.85 93. Hemby SE, Ginsberg SD, Brunk B, Arnold SE, Trojanowski JQ, Eberwine JH (2002) Gene expression profile for schizophrenia: discrete neuron transcription patterns in the entorhinal cortex. Arch Gen Psychiatry 59 (7):631–640. https://doi.org/10.1001/arch psyc.59.7.631 94. Fend F, Emmert-Buck M, Chuaqui R, Cole K, Lee J, Liotta LA, Raffeld M (1999) Immuno-LCM: laser capture microdissection of immunostained frozen sections for mRNA analysis. Am J Pathol 154(1):61–66. https:// doi.org/10.1016/S0002-9440(10)65251-0 95. Ginsberg SD, Hemby SE, Lee VMY, Eberwine JH, Trojanowski JQ (2000) Expression profile of transcripts in Alzheimer’s disease tangle-bearing CA1 neurons. Ann Neurol 48 (1):77–87. https://doi.org/10.1002/15318249(200007)48:13.0. CO;2-A 96. Schafer MJ, Dolgalev I, Alldred MJ, Heguy A, Ginsberg SD (2015) Calorie restriction suppresses age-dependent hippocampal transcriptional signatures. PLoS One 10:e0133923. https://doi.org/10.1371/journal.pone. 0133923 97. Suarez-Quian CA, Goldstein SR, Pohida T, Smith PD, Peterson JI, Wellner E, Ghany M, Bonner RF (1999) Laser capture microdissection of single cells from complex tissues. BioTechniques 26(2):328–335. https://doi. org/10.2144/99262rr03 98. Simone NL, Paweletz CP, Charboneau L, Petricoin EF III, Liotta LA (2000) Laser capture microdissection: beyond functional genomics to proteomics. Mol Diagn 5 (4):301–307. https://doi.org/10.1054/ modi.2000.19808a 99. Craven RA, Banks RE (2002) Use of laser capture microdissection to selectively obtain distinct populations of cells for proteomic analysis. Methods Enzymol 356:33–49. https://doi.org/10.1016/S0076-6879(02) 56921-4 100. Moule´dous L, Hunt S, Harcourt R, Harry JL, Williams KL, Gutstein HB (2003) Proteomic analysis of immunostained, laser-capture microdissected brain samples. Electrophoresis 24(1–2):296–302. https://doi.org/10. 1002/elps.200390026 101. Freeman WM, Hemby SE (2004) Proteomics for protein expression profiling in

Fixation Protocols for Neurohistology neuroscience. Neurochem Res 29 (6):1065–1081. https://doi.org/10.1023/ B:NERE.0000023594.21352.17 102. Matsuo ES, Shin RW, Billingsley ML, Van deVoorde A, O’Connor M, Trojanowski JQ, Lee VM (1994) Biopsy-derived adult human brain tau is phosphorylated at many of the same sites as Alzheimer’s disease paired helical filament tau. Neuron 13(4):989–1002. https://doi.org/10.1016/0896-6273(94) 90264-X 103. Okuducu A, Janzen V, Hahne JC, Ko Y, Wernert N (2003) Influence of histochemical stains on quantitative gene expression analysis after laser-assisted microdissection. Int J Mol

71

Med 11(4):489–453. https://doi.org/10. 3892/ijmm.11.4.449 104. Bi WL, Keller-McGrandy C, Standaert D, Augood SJ (2002) Identification of nitric oxide synthase neurons for laser capture microdissection and mRNA quantification. BioTechniques 33(6):1274–1283. https:// doi.org/10.2144/02336rr01 105. Wang H, Owens JD, Shih JH, Li MC, Bonner RF, Mushinski JF (2006) Histological staining methods preparatory to laser capture microdissection significantly affect the integrity of the cellular RNA. BMC Genomics 7:97. https://doi.org/10.1186/14712164-7-97

Chapter 3 Three-Dimensional Atlases of Insect Brains Basil el Jundi and Stanley Heinze Abstract The morphological structure of the nervous system is ultimately the basis of its function. Analyses of the anatomical layout of brain areas, single neuron morphologies, and the synaptic connectivity of neurons are therefore essential for a comprehensive understanding of the computational processes implemented in neuronal networks. Insect brains have long served as models to examine neuronal circuits that process sensory information, provide the substrates for learning and memory, or generate motor patterns that drive well-studied behavior. The relatively small number of neurons these brains are composed of (up to one million) and their small overall size make them easily accessible for physiological and anatomical research. To aid the comparison of results within and across species, and thus make it possible to relate function to anatomical structure, printed brain atlases have been used as a common frame of reference for many decades. In recent years, digital, three-dimensional atlases were generated to provide geometrical as well as conceptual reference systems for the brains of several insect species. In this review we compare the different approaches for generating such three-dimensional atlases. We highlight the key problems that must be overcome during this process and the solutions that have been found to achieve this. The advantages and limitations of the different strategies are discussed, and the applications that have so far resulted from the implementation of these atlases are described. Key words Immunohistochemistry, Fixation, Confocal microscopy, Volumetric analysis, Quantitative neuroanatomy, Standardization, Neural networks, Central complex

1

Introduction The function of the nervous system is intimately linked to its structure, and the former cannot be understood without investigating the latter. Underlining the importance of anatomical work, the beginning of modern neuroscience was effectively marked by the insight that the nervous system consists of many millions of neurons instead of a continuous network of fibers—a revelation that was based on the ability to selectively label neurons [1]. The technique that initiated this development, the Golgi silver impregnation, was invented in the late 1800s and has been extensively used to define the structure of neurons, describe their anatomical variations, and infer the nature of their connections (e.g., [2]). The

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_3, © Springer Science+Business Media LLC 2020

73

74

Basil el Jundi and Stanley Heinze

many thousands of drawings by S. Ramo´n y Cajal still provide the foundation of modern neuroscience [1]. By compiling information about many individual neurons into brain atlases, researchers were able not only to classify cell types of different brain regions but also to draw functional conclusions about sensory pathways, infer connections between brain areas, and analyze the intrinsic organization of those regions. Printed atlases of the human brain and other vertebrate brains served as a reference system by providing a set of coordinates, which enabled researchers to unambiguously refer to specific structures (e.g., [3–5]). Additionally, the wide use of such atlases encouraged a common nomenclature system for each species and therefore facilitated efficient communication and exchange of data originating from different laboratories [6–8]. Although early work, for example, by Ramo´n y Cajal or Bertil Hanstro¨m, was also performed on invertebrates [9], insect brain research has long lacked a similarly systematic approach. Rather, individual researchers worked on many different species across the field of insect neuroscience, each with their own system for classifying anatomical information within their research topic [10]. The first attempt to comprehensively map the brain of an insect was the atlas of the house fly brain by Strausfeld [11]. Based on Golgi impregnations and classic neuropil stainings, all regions of the fly brain were displayed and named, and the neurons found in these areas were classified. Unmatched in aesthetic beauty and completeness in any species ever since, this seminal work has served as the prototype of an insect’s brain atlas and has guided countless anatomists over more than three decades. Nonetheless, much work in insect neuroscience has been performed on other species, particularly on the fruit fly, but also on bees, locusts, cockroaches, beetles, moths, butterflies, and many more [1, 9]. As no framework similar to the fly atlas exists in any of these species, several attempts have been made in recent years to overcome this problem and to generate standardized atlases of the brains of a variety of insects [12– 22]. Owing to technological progress both in imaging techniques as well as in data processing abilities, all new attempts to provide a common framework for anatomical data in insects have been carried out with the goal of generating three-dimensional (3D) atlases, in which morphological data on neurons can be collected. This chapter will review these attempts and give an overview about the specific problems encountered when generating standard brain atlases of insects, as well as the ways in which these problems have been solved in different species. The possibilities but also the limitations of such approaches will be highlighted, and ways to further optimize the use of this new generation of insect brain atlases will be suggested.

Insect Brain Atlases

2

75

Why Do We Need Three-Dimensional Atlases? When two-dimensional brain atlases are used to visualize data from three-dimensional brains, the transformation of the displayed information from brain slices to the actual configuration in the brain has to be performed by the researcher’s imagination. This leads to two problems. Firstly, although the generation of the two-dimensional projection views (e.g., with the camera lucida technique) can be very accurate, all information about the z-dimension is lost so that any back-transformation from two to three dimensions becomes highly subjective. Secondly, to obtain meaningful insights, it requires substantial training and anatomical background knowledge that is not readily available to many research groups that focus on functional studies of brains. However, the morphological layout of the brain is the physical substrate for all neural processes examined by physiological or genetic methods. Thus, a unifying brain atlas, which ties knowledge from the diverse disciplines of insect neuroscience to a shared anatomical framework, can act as a conceptual anchor in each species, thereby facilitating crossdisciplinary research. In this context, it is of major importance that visualizing brains and neurons in three dimensions makes morphological knowledge more accessible to researchers without extensive anatomical background. This enables linking structure with function in research fields, in which this is not typically done, so that new insights are likely to emerge. That linking structure and function can, indeed, lead to new insights, particularly in insect brains, has been shown in many examples, for example, in the moth antennal lobe [23, 24], the bee antennal lobe [25], or in the locust central complex [26]. The most obvious use for a 3D atlas of an insect brain is the possibility of combining morphological data on single neurons obtained from many individual brains into a common frame of reference [13, 20, 21]. Functionally related neurons can thus be visualized simultaneously and spatial relations between cells can be quantified, for example, via distance mapping to other neurons [15, 27]. In this way, network properties, including potential synaptic links of the involved neurons, can be identified or at least narrowed down [28]. Additionally, normalized lengths of neural connections, number and orientation of neural elements within networks, and relative diameters of branches can be measured directly. Based on this data, potential feedback connections, delay lines, local processing circuits, sites of convergence or divergence, and mapping of sensory information can be predicted. By including the structural constraint imposed by such anatomical data, more realistic modeling of computations performed by defined neural networks can be obtained.

76

Basil el Jundi and Stanley Heinze

Besides single neurons, immunohistochemical stainings against neurotransmitters [19] as well as data from in situ hybridizations (e.g., [29]) can also be registered into standardized brain atlases. In brain areas such as the medulla of the optic lobe or the central body of the central complex, different transmitters (particularly neuropeptides and biogenic amines) exhibit expression patterns that reveal a complex, stratified neuroarchitecture of these brain regions [30]. These strata serve as reliable orientation cues for mapping the localization of arborization trees of individual neurons and define anatomical as well as functional subdomains within neuropils (e.g., [31, 32]). By superimposing image stacks of many of these expression patterns within the same, normalized frame of reference, an increasingly refined neurochemical map of the brain of an insect species can be generated, not only for transmitters but potentially also for their receptors and components of intracellular signaling cascades. On the one hand, this will make it possible to define arborization domains of neurons in well-described areas of the brain with high precision, as well as enable us to assign functional subdivisions within the so-called unstructured neuropils that occupy vast areas of an insect’s central brain. On the other hand, this neurochemical map of the brain will facilitate identification of brain regions that contain high levels of many neuromodulators, and thus identify areas where synaptic plasticity might be particularly high. Combined with mapping single neurons to these regions, predictions about state-dependent modulations of these neurons’ physiology can be made that might serve as the basis for subsequent functional experiments. Furthermore, the size and shape of standardized neuropils of the atlas already provide valuable information in their own right. These data have not been directly accessible in any previous atlas based on two-dimensional images and can now serve as a reference for intra- and interspecific comparisons (e.g., [12, 13, 16–19, 33]). When standardized atlases are compared between species, it is essential that the same set of homologous neuropils is included in each species and similar criteria for their definition are used. For comparison within one species, it is important that the brains obtained to generate the atlas correspond to a consistent state of the animals used (e.g., same age, sex, experience). This defined set of neuropil shapes and the defined distribution of volumes can then be treated as a static baseline dataset for further comparisons. If naive, adult animals are used just after completion of initial developmental processes, the resulting standard can serve as a reference to quantify the structural correlates of experience-dependent plasticity, sensory deprivation, genetic manipulations, or different behavioral states (such as gregarious versus solitary states, or migratory versus nonmigratory life phases) (e.g., [12, 33]). Additionally, sex-specific differences in brain areas are accessible and quantifiable in the same way (e.g., [14, 16, 34]). Again, by combining such

Insect Brain Atlases

77

broad structural data with data of single neurons that arborize in regions exhibiting interesting differences, individual neurons might be identified as candidates for playing a role in the examined functional phenomena. In summary, there are a variety of reasons why 3D standardized brain atlases are an important addition to the repertoire of insect neuroscience. How useful they are as tools, however, depends on the quality of the atlas and the methods used to generate it. Both of these aspects heavily depend on the quality of the images used as basis for the 3D reconstructions that comprise the atlas. Which aspects of tissue preparation have proven to be particularly challenging, especially in the brains of large insects, will be highlighted in the following section.

3

Methodical Challenges Most 3D atlases of insect brains rely on image segmentation for visualizing distinct boundaries between different brain regions. To achieve this, the intrinsic contrast within nervous tissue of unstained samples has to be increased. In all atlases of insect brains over the last decade, this has been achieved via immunohistochemical fluorescence labeling of synaptic markers (Fig. 1). The treated brains are processed for confocal microscopy and optical sections are acquired. These image stacks are finally used to reconstruct triangulated surface models of brain regions. Despite the fact that insect brains are small compared to vertebrate brains, immunohistochemistry performed on unsectioned brains, for example, wholemount preparations, is a challenging task [35]. Several problems result from the fact that whole insect brains are much thicker than the sections that are classically used for immunohistochemical methods. First, antibodies have to be able to penetrate the tissue completely to avoid gradients in staining intensities from the edge to the center of the brains. This process requires that the tissue itself as well as the neural sheath around the brain is made permeable for the used antibodies. As this is heavily influenced by tissue fixation protocols, a balanced protocol that ensures tissue stability, antigen accessibility, epitope preservation, and tissue permeability has to be found for each species. Second, thick tissue is difficult to image with a confocal microscope at high resolution. Problems include the limited working distance of most objectives as well as absorption and scattering of excitation and emission light within the tissue. The latter problem can be partly solved by clearing the tissue, for example, with methyl salicylate [36]. As this process mostly requires dehydration of the used brains, shrinkage of tissue takes place and raises the question as to how faithfully the final images represent the situation in the living brain [13, 35, 37]. Finally, staining protocols involving whole-mount preparations tend to be very slow, allowing

78

Basil el Jundi and Stanley Heinze

Fig. 1 Neuropil staining (anti-synapsin) in different insect whole-mount brain preparations. All images are frontal views of single optical sections from confocal image stacks at the level of the central complex (CX). They illustrate the staining quality obtained by different staining protocols. (A) Brains of four insect species (see B-E) depicted to scale to highlight their relative sizes. (B) Brain of the red flour beetle (Tribolium castaneum). (C) Brain of the hawkmoth Manduca sexta. The antibody permeability was improved by using methanol during fixation. (D) Brain of the cockroach (Leucophaea maderae [now: Rhyparobia maderae]). The neural sheath was treated with collagenase to facilitate antibody permeability. (E) Brain of the sweat bee (Megalopta genalis). Sufficient antibody penetration was achieved by adding zinc ions to the fixative solution. OL optic lobes. Images B–D are based on data originally employed in other studies: B [18], C [16] and D [19]

for only a limited number of optimization runs for each species, if results are to be obtained within a reasonable amount of time. For that reason, each method used for a particular species always represents a trade-off between many variables and should never be regarded as the final, optimal protocol possible. Obtaining confocal microscopy images from large, thick samples is not trivial either, and the optimal imaging settings are again a trade-off between many factors. The goal for each standardized atlas is to contain information at the highest possible resolution in all three dimensions. The resolution limit is mostly defined by the numerical aperture of the used objectives, which is of particular importance in the axial dimension. As objectives have a limited working distance and field of view, the correct choice of objective

Insect Brain Atlases

79

requires considerable thought. Additionally, resolution is modulated by image contrast, which is diminished in thick samples such as whole-brain preparations. Hence, optimizing photon gain, while minimizing negative effects like photobleaching, is crucial for optimal resolution. The digital nature of images also requires that pixel size is carefully selected, as pixel size affects the photon gain during imaging and defines the size (in terms of memory) of the final image. Knowing the limits of available computers and software with respect to the largest file sizes that can be processed, eventually defines the voxel size at which a sample should be imaged. Once the images are acquired, the next challenge is image segmentation (Fig. 2). For this, a variety of software exists, both open source and commercial, to perform manual, semiautomatic, or automatic image segmentation (e.g., Amira, Imaris, FIJI, NeuroLucida, MatLab). A decision has to be made at this stage, as to which brain structures should be included in the atlas. It has to be considered that the new atlas to be generated should be compatible with already existing ones to enable meaningful comparisons. Also, the applications intended for each particular atlas, and the methods needed to achieve these, have to be considered when generating the atlas. This point is particularly important as the final result will always be a trade-off between many parameters, and no single atlas will be optimal for every single purpose. Finally, a digital 3D atlas of a brain is only useful if it is available for a wide range of users and is produced in a way that allows for easy incorporation of new data and additional features [6]. Therefore, the platform intended for publishing and distribution of an atlas should be suited for its anticipated use. The effort that has to be put into developing a user-friendly, highly visible, interactive, widely available, and visually appealing platform is considerable and should not be underestimated. The different ways these challenges have been tackled over the past decade, the limitations and strengths of the different solutions, and suggestions as to how future atlases of insect brains could benefit from these experiences will be reviewed in the following sections.

4

Different Approaches to Standardizing Insect Brains

4.1 Staining Protocols for Large Brains

As outlined above, the quality of a 3D insect brain atlas is firstly determined by the staining quality of the brain preparations. Only when sufficient contrast is generated, the subsequent image segmentation can be achieved satisfactorily and standardization algorithms can perform successfully. Although valid for all standardization protocols, staining contrast and intensity particularly influence the quality of those standard atlases generated by techniques that use confocal image stacks directly for brain

80

Basil el Jundi and Stanley Heinze

Fig. 2 The different methods to create insect brain atlases. All methods initially transform individual brains onto a template brain by using a rigid or affine transformation. Then the brains are aligned by a non-rigid transformation. The VIB (virtual insect brain) method [52] uses segmented image stacks (3D models) to create the standard brain. Individual 3D models are globally transformed by using the centers of all reconstructed neuropils as reference points. Afterward, the individual neuropils are locally transformed during the diffusion transformation. The final 3D standard model is generated by a threshold segmentation of the probability map. In addition, the VIB method applies all transformation parameters to the raw image stacks to generate a standardized grayscale image stack. The CMTK (Computational Morphometry Toolkit) method [54] generates the standard atlas based on raw images of individual brains. The standard brain is generated by an intensitybased algorithm using an affine transformation followed by a non-rigid transformation. In a modified version of the CMTK method, the ISA protocol, the non-rigid transformation is repeated several times (iterative averaging) to increase standardization quality. Afterward, all displacement and deformation parameters are applied to the corresponding segmented images. The FlyCircuit method [20] uses distance maps during the non-rigid transformation to generate the standard atlases. Similar to the VIB method, segmented image stacks are used for alignment. The BrainAligner technique [21] uses a reliable landmark algorithm to align the individual brains and is directly applied to raw image stacks

Insect Brain Atlases

81

alignment (see below; Fig. 2). All insect atlases to date—with the exception of Drosophila—are based on immunohistochemical labeling with an antibody against the presynaptic vesicle protein Synapsin [38]. In Drosophila, the antibody nc82, recognizing the protein Bruchpilot [12, 14, 21, 34], and an anti-DLG antibody (directed against Drosophila discs large protein; [20]) have been used to visualize the different compartments of the brain. All used antibodies have in common that they highlight brain regions with high synaptic density. In insects, such areas are free of cell bodies and separate from regions that mostly consist of fibers (tracts, bundles, and fascicles). Together, these “neuropils” comprise the bulk of the insect brain and contain many defined regions with easily recognizable shape, even across species (Figs. 1 and 3). The basic staining protocol consists of five principal steps: (1) fixation; (2) permeabilization; (3) antibody incubation; (4) dehydration/clearing; and (5) embedding in mounting medium. In Drosophila and other small species (e.g., the red flour beetle Tribolium castaneum), the permeabilization step can be omitted while still retaining adequate staining quality (Fig. 1B). Fixation. After dissection, the brains need to be immediately fixated to avoid loss of staining quality due to protein degradation. A standard cross-linking fixative consists of 4% formaldehyde solution, which is applied to the brains for 2–3 h at room temperature or overnight at 4
C. However, especially in larger insect brains, the antibody permeability of whole-mount preparations is substantially reduced after a formaldehyde fixation [39], resulting in a gradient of decreasing staining intensity toward the center of the brain [35]. Additionally, very strong fixation (e.g., overnight with 4% formaldehyde) leads to masking of antigens and reduced staining intensity compared to a shorter fixation with the same fixative [35]. To overcome these problems, in locust and monarch butterfly brains, increased permeability was obtained by adding zinc to the fixative, while reducing overfixation by decreasing the formaldehyde concentration from 4% to 1% [32, 35, 40]. Zn-ions coagulate cytoplasmic proteins and thus facilitate passage of antibodies through the tissue. Alternatively to formaldehyde, methanol has been used in some vertebrates as a non-cross-linking, coagulant protein fixative [41]. In insect tissues, such as the thoracic ganglia of the locust Schistocerca gregaria and the brain of the hawkmoth Manduca sexta, a combination of methanol and formaldehyde was applied as a fixative solution [16, 42–44]. In addition to its capacity as a fixative, methanol also eliminates cellular membrane lipids and thus contributes to improved antibody permeability of the tissue. Permeabilization. Increasing the antibody permeability is an essential step to achieve an adequate staining quality in large insect brains. A variety of strategies have been applied to accomplish this. In the brains of the cockroach Leucophaea maderae [19] and the locust Schistocerca gregaria [15] the ganglionic sheath was

82

Basil el Jundi and Stanley Heinze

Fig. 3 Standardized atlases of insect brains. (A) Frontal views of 3D surface models of all so far generated standardized atlases of insect brains, drawn to scale. Homologous brain regions (neuropils) are shown in

Insect Brain Atlases

83

enzymatically digested by using the enzyme collagenase (Fig. 1D). In contrast, in the brains of the honeybee Apis mellifera [13] and the moth Heliothis virescens [17], in the thoracic ganglia of the dragonfly Libellula luctuosa [45], and in another approach for the locust Schistocerca gregaria [35], high antibody permeability was achieved by a dehydration–rehydration protocol, that is, an increasing ethanol series directly followed by a decreasing ethanol series. In the moth and bee brain, the permeability was further increased by a short treatment with xylene in between the ethanol series. In the dragonfly thoracic ganglia and in moth brains, an additional permeabilization step with collagenase followed after the ethanol series. Ott [35] alternatively improved the antibody permeability in locust by applying a dimethylsulfoxide (DMSO)–methanol mixture for 45–60 min in conjunction with the formaldehyde–zinc fixation protocol. This approach was subsequently also adopted for monarch butterfly brains [32] as well as for brains of the solitary sweat bee Megalopta genalis (unpublished data by S.H.) (Fig. 1E). In the brain of Manduca sexta, no additional step was necessary as permeabilization was sufficiently achieved by the methanol–formaldehyde fixation, although a gradient in staining intensity toward central brain regions remained (Fig. 1C) [16]. Antibody incubation. The adjustment of antibody incubation times is crucial for achieving an even staining quality and depends heavily on brain size. To prevent tissue degradation during antibody incubation, all long incubation steps are carried out at 4
C. Smaller brains need relatively short incubation times that range ä Fig. 3 (continued) identical colors across species. Only standardized neuropils are shown (see panel B for neuropil labels). Due to their small size, the brains of the red flour beetle (Tribolium castaneum) and the fruit fly (Drosophila melanogaster) are additionally shown as enlargements. For Drosophila, the original standardized atlas is shown on the right [12, 52]. A more comprehensive standard atlas of Drosophila, generated recently, is shown further left and includes all brain neuropils [22]. The brain regions collectively referred to as “unstructured neuropils” (gray) were only separated into their individual compartments in Drosophila (left), while for the remaining species they were either not included at all or were reconstructed as a single large compartment. For Tribolium castaneum and the hawkmoth (Manduca sexta), the outer brain boundary is additionally displayed. (B) Brain neuropils included in all standard atlases (from top to bottom: mushroom body, central complex, antennal lobe, optic lobe) are shown side by side to illustrate interspecies differences in neuropil layout, arrangements of subcompartments, and level of detail used for standardization. All neuropils within each species are shown at their correct relative sizes. To aid comparison between species, the mushroom bodies have been matched in size. Abbreviations: CA calyx, PED pedunculus, VL/ML vertical/ medial lobe, CBU/CBL upper/lower division of the central body, PB protocerebral bridge, NO noduli, AL antennal lobe, LO lobula, LOP lobula plate, LOX lobula complex, iLO/oLO inner/outer lobula, ME medulla, LA lamina, AME accessory medulla. All images are based on data kindly provided by researchers who originally employed them in other studies: Leucophaea [now: Rhyparobia] maderae, H. Wei, ISA method [19]; Schistocerca gregaria, U. Homberg, ISA method [15]; Apis mellifera, J. Rybak and R. Menzel [13]; Heliothis virescens, P. Kvello and H. Mustaparta [17]; Tribolium castaneum, M. Kollmann and J. Schachtner [18]; Drosophila melanogaster, A. Jenett [22, 52]; Manduca sexta, B. el Jundi [16]

84

Basil el Jundi and Stanley Heinze

between 1 and 3 days (Drosophila melanogaster), and 1 and 2 days for primary and secondary antibodies (Tribolium castaneum) [18, 20, 21]. To ensure that the antibodies can reach central regions of the brain, incubation times are increased up to 6 days for the primary antibody and up to 4 days for the secondary antibody in larger brains [13, 15, 16, 32]. In addition, a preincubation/blocking step (e.g., with normal goat serum, NGS) can be added before incubating the samples with the primary antibody. This increases the staining contrast by reducing unspecific binding of antibodies to nontargeted epitopes within the sample. To further reduce background staining, NGS is often added to the antibody solutions as well. Additionally, detergents such as Triton X-100 or the powerful solvent DMSO are frequently added to the incubation solution, as they further improve tissue permeability. After each incubation step, the brains need to be extensively rinsed with buffer. To ensure that excess antibodies are washed out of central regions of the brain, the washing times are much longer in wholemount preparations of larger brains (6–8 times 20–30 min), compared to the corresponding washing times in smaller brains (6–8 times 5–15 min). Clearing of tissue. In whole-mount preparations, the disparity between the refractive indexes of brain tissue protein, lipids, and water causes light scattering, which reduces the image quality in microscopy. Therefore, the tissue of whole-mount preparations needs to be rendered transparent, if sufficiently bright signals should be detected from central brain regions. Most often, this is achieved by replacing the water in the preparation by a medium with a refractive index close to that of tissue proteins (optical clearing). One of the most efficient clearing agents is methyl salicylate [36]. It has been used particularly in conjunction with larger brains [13, 15, 16, 19, 32] but was also successfully applied to Tribolium castaneum brains [18]. Clearing with methyl salicylate requires preceding dehydration of the tissue. This can be conducted by using an ascending ethanol series. However, repeated treatments with ethanol, for example, when permeabilization has been achieved by the same means, can lead to tissue shrinkage and severe distortions of the brain tissue during the clearing process [35, 37]. This issue has been solved in dragonflies by applying an increasing 2,20 -thiodiethanol (TDE, in buffer) instead of the ethanol series to the samples [45]. In locusts, an increasing glycerol series has been used, followed by a treatment with pure ethanol, and subsequently clearing the brain with methyl salicylate [35]. Alternatively, Drosophila brains were also cleared with glycerol-based mounting media such as Vectashield (Vector Laboratories) [12, 21] or FocusClear (CelExplorer Labs) [20]. In general, it should be emphasized that this clearing step is a reversible procedure and is mostly required as an intermediate step enabling infiltration of the tissue with the mounting (nonpolar)

Insect Brain Atlases

85

medium after dehydration in ethanol (polar solvent). Finally, the transparency of the brain tissue is mostly determined by the mounting medium, largely owing to refractive index matching. Mounting. The final step in the generation of whole-mount preparations consists of mounting the brains between two coverslips. For this, a proper mounting medium needs to be chosen. As mentioned above, the clearing power of the mounting medium largely determines the transparency of the tissue. Often, mounting media that maximize tissue transparency have a refractive index closely matching that of glass. This is highly desirable, as it ensures a refraction-free light path from the sample to the objective via the glass coverslip, given that an immersion medium (typically cedar oil) of very similar refractive index is used. Additionally, an ideal mounting medium should allow for permanent storage of the preparation. While Drosophila brains are usually mounted using glycerol or a glycerol-based embedding medium such as Vectashield [12, 21] or MountClear (CelExplorer Labs) [20], Ott [35] used methyl salicylate (otherwise typically used as a clearing medium) to embed locust brains and Gonzalez-Bellido and Wardill [45] used TDE to mount dragonfly thoracic ganglia. When using these embedding mediums, the coverslips need to be sealed around the perimeter to enable short-term storage over several months. Most of the whole-mount preparations used to generate the standardized insect brain atlases covered in this chapter have been mounted using Permount (Fisher Scientific). This embedding medium uses xylene as solvent and dries into a solid resin over time. It is thus ideal to permanently store brain preparations for several years without the need for sealing the perimeter of the coverslips. However, such solid mounting media also have some disadvantages. The brain samples need sufficient time to dry (2–3 days) and equilibrate in the medium. Before an equilibrium is reached, the mismatch in the refractive indices of the clearing solution retained in the brain (e.g., methyl salicylate) and the mounting medium itself will adversely affect microscopic image quality when high-power lenses are used. Furthermore, air bubbles around the specimen can arise during drying of the mounting medium due to evaporation of the solvent (xylene), potentially reducing image quality as well. Additionally, it is noteworthy that anhydrous, nonpolar mounting media (e.g., Permount, methyl salicylate) favor the use of cyanine-based fluorescence dyes coupled to secondary antibodies. These dyes have been reported to be several times brighter in conjunction with nonpolar mounting media compared to newer Alexa Fluor dyes or DyLight dyes (Jackson ImmunoResearch, 2014). 4.2 Confocal Imaging

Obtaining images with the confocal microscope is not trivial and the optimal settings used for imaging represent a trade-off between many factors, particularly when dealing with large, thick samples. A standardized atlas should ideally be based on image data obtained at

86

Basil el Jundi and Stanley Heinze

the highest possible resolution, in all three dimensions. Resolution is diffraction-limited in a confocal microscope and depends on the wavelength of light (λ), and—most importantly—the numerical aperture (NA) of the objective used (lateral resolution limit: rlateral ¼ 0.37λ/NA). Unlike in wide-field microscopy where only emission wavelength affects resolution, both excitation and emission wavelengths affect it in confocal microscopy, with the excitation one becoming dominant at larger pinholes (over ~1 Airy unit). A value between them (approx. their geometric average) is thus frequently used to calculate resolution [46, 47]. In the z-dimension, the resolution (axial resolution limit) is considerably poorer than the lateral one and worsens with the square of NA (rax2 ial ¼ 1.28λ·n/NA ) [46], where n is the refractive index of the immersion medium (air in the case of a dry lens). The latter formula is an approximation valid only for dry objective lenses with NA under ~0.60. Additionally, both formulas only hold for a pinhole diameter equal to a small fraction (~1/4 or less) of 1 Airy unit, a condition that can only rarely be fulfilled in practice, due to signalto-noise constraints. More often, the confocal lateral resolution is closer to the widefield one, which is no more than ca 1.4 times worse [46, 48]. Confocal axial resolution (optical sectioning capability) also worsens with increasing pinhole size [46, 47]. Thus, the choice of a high numerical aperture objective is crucial to gain good axial resolution (Fig. 4A). It also defines the distance between the successive focal planes, that needs to be chosen before acquiring a confocal image stack. For the sampling along the z-axis to be adequate (i.e., complying with the NyquistShannon theorem discussed in Chapter 14 by Riley et al.), the distance between the successive focal planes should not exceed one half of the axial resolution limit (which determines the thickness of an “optical section”). Often, the distance is set to even smaller values so that the optical sections overlap. Another point to consider with respect to z-scaling is z-dimension shrinkage due to refractive index mismatch between the mounting and immersion media (or air in case of a dry lens). Similar to looking into a pool of water, the size of objects along the z-axis is shortened when a low refractive index immersion medium (e.g., water, n ¼ 1.33) or no medium at all (dry lens, n ¼ 1) is used to image a sample mounted in a high refractive index medium (~1.50). The shortening factor is approximately equal to the ratio of the two refractive indices [49]. Modern microscopes offer automatic z-dimension rescaling during image acquisition. Alternatively, the acquired images can instead be adjusted (stretched) afterward. This process is paramount for obtaining correct volume data and brain proportions. The same sampling rule as for the z-axis applies along the x- and y-axes (pixel size should be at least twice smaller than the lateral resolution limit). By adjusting the zoom settings of the confocal

Insect Brain Atlases

87

Fig. 4 Quality of images obtained with objectives of different resolving power and their effect on corresponding 3D reconstructions of neuropils and single neurons. (A) Illustration of confocal resolution for objectives of

88

Basil el Jundi and Stanley Heinze

microscope, the pixel size can be changed without affecting the axial resolution. Generally, undersampling can reduce image contrast and cause aliasing artifacts. However, as axial resolution represents the limiting factor for the overall quality of the final brain atlas, it is often not necessary to strictly adhere to this rule in the x and y dimensions. In fact, undersampling can help minimize photobleaching. Unfortunately, high numerical aperture objectives usually feature high magnification and small working distance, meaning that they have a small field of view and cannot reach the interior of thick samples. Hence, for imaging of large brains, only high-quality, low-magnification objectives (up to 25) are suitable. Additionally, the resolution limit is also modulated by contrast, which is considerably lower in thick samples due to scattering of excitation and emission light above and below the focal plane. The theoretical resolution limit thus often cannot be reached in whole-brain imaging conditions, and the highest practically achievable resolution depends on maximizing photon gain while detecting the emission light; one can increase the dwell time of laser beam at each pixel, the laser intensity, or the detector gain. However, high laser intensities and long dwell times increase photobleaching and generally damage the specimen (including regions outside the focal plane, which have not been imaged yet), while long dwell times dramatically increase the time required for data acquisition (confocal scanning). Too high values of detector gain will amplify the dark current of the detectors and reduce the signal-to-noise ratio. ä Fig. 4 (continued) different numerical apertures (NA). The z-dimensions represent the effective thickness of one optical section at a pinhole size of one Airy unit. The xy-dimensions represent the lateral resolution. The values denoted are the optimal voxel dimensions obtained from the Zeiss-LSM5 microscope software while taking into account excitation wavelength, and refractive index mismatch [49] between mounting medium and immersion medium (air in case of a dry lens). They do not necessarily match the theoretical resolution limits [46, 47]. (B) Single optical sections of an anti-synapsin labeled brain of the sweat bee Megalopta genalis (top, frontal view; bottom, ventral view), obtained with a 10 objective, resampled to the resolution used for reconstruction (3  3  3 μm). Note the lack of detail in the z-dimension. (C) 3D surface reconstructions based on the image stack shown in B. (D, E) Same as B/C, but images obtained with a 25 oil immersion objective from the same preparation. Note the additional detail and increased precision in images and reconstructions. (F) Maximal intensity projection of a confocal image stack obtained from a neurobiotin filled central-complex neuron of the monarch butterfly. Shown is the arborization tree in the lateral accessory lobe of a CL1 neuron. Images were obtained with a 25 oil immersion objective directly from a whole-mount preparation. (G) Skeletonized neuron reconstruction resulting from images shown in F. The distance along the neurites is illustrated in false colors (from blue to red). (H, I) Same as F/G, but images were obtained with a 40 oil immersion objective after rehydration and sectioning (120 μm thickness) of the brain. Note the markedly improved resolution in the z-dimension. (J) The improved quality of the images in H enabled creating a triangulated surface reconstruction of the neuron, upon generating the skeleton tree. Images F–J are based on data originally employed in another study [33]. Abbreviations: NA numerical aperture, CBU/CBL upper/lower unit of the central body, PB protocerebral bridge, NO noduli, AOTU anterior optic tubercle, LU/UU lower/ upper unit

Insect Brain Atlases

89

Finally, the analog detector signals are digitized, yielding a finite number of pixels, whose dimensions define the lateral resolution in the image, as described above (resolution  double the pixel size). Small pixel sizes increase the electronic size of an image, and more powerful computers are consequently required for data analysis. This will also limit the highest practically achievable resolution of atlases of large brains because current computers and software cannot handle file sizes above certain limits. 4.3 Image Processing: Stitching of Image Stacks

Small insect brains can usually be imaged in one image stack with a confocal microscope. However, due to the thickness and width of large insect brains, several image stacks are required to cover the whole brain, even when using a 10 objective. Thus, the resulting individual datasets have to be merged into one continuous image stack. To achieve this merging of numerous image stacks reliably, a combination of manual and semimanual tools of the software Amira (FEI, Visualization Sciences Group) has been frequently used (e.g., [16]). More recently, fully automated image stitching has become available in the ImageJ implementation FIJI [50, 51], yielding excellent results with much less potential for human error. If the resulting image stack exceeds the data capacity of the used 3D-reconstruction software or of the computer, the data stack has to be resampled to appropriately larger voxel sizes at the expense of image resolution.

4.4 Neuropil Reconstruction

Some standardization techniques (VIB and FlyCircuit, see Subheading 4.5) require 3D reconstructions of neuropils prior to brain registration. In this step, the obtained confocal image stacks serve as reference to reconstruct individual neuropils (image segmentation). For all insect atlases, the Amira software was used for this process. The extent of each neuropil has to be manually defined by explicitly assigning the voxels of the raw image stack with corresponding neuropil identities. An efficient strategy to achieve this without the need to manually label each individual voxel is to label the neuropil boundaries of only a limited number of optical sections in each spatial plane (x-y, x-z, y-z). The final 3D model can then be obtained by interpolating the regions in between the labeled optical sections with a radial basis function algorithm. However, this interpolation algorithm only generates an optimally smooth surface of the brain area if the image stack voxels are at least approximately cubic. Otherwise (e.g., relative voxel dimensions 1  1  5), the data stack should be appropriately resampled (e.g., to relative voxel dimensions 1  1  1). As neuropil boundaries are manually defined by the experimenter, the segmentation step requires detailed knowledge about the general anatomy of the insect brain, and reproducibility as well as quality of reconstruction depends heavily on the experimenter’s experience.

90

Basil el Jundi and Stanley Heinze

4.5 Standardization Protocols

Standard atlases are virtual brains calculated by aligning individual brains of the same species into an average brain. Several standardized atlases of insect brains have been generated over the last few years and, depending on the scientific question, each of them was created by one of four main standardization methods (Fig. 2). The atlas can either be generated using segmented image stacks of brains or by directly aligning raw image stacks. Although all four protocols differ substantially in registration algorithm and processing speed, the general standardization procedure is the same: Individual brains are first transformed based on six degrees of freedom (translation and rotation [rigid transformation]) or nine degrees of freedom (translation, rotation and anisotropic scaling [affine transformation]). Afterward, standard brains are generated by a non-rigid transformation (Fig. 2). Before the brains are aligned, an image stack of an individual brain (segmented or raw) is selected as reference brain. This template provides the initial reference for alignment of the remaining brains. The careful choice of an adequate reference brain is crucial, as all other brains are transformed to match that template during the rigid transformation phase. In the following paragraphs the characteristics of the different methods will be outlined, and their individual advantages and limitations are highlighted. The Virtual Insect Brain (VIB) protocol [52] has been originally used to create a standard brain for Drosophila melanogaster [12] and was later employed to create standard atlases of the brains of the desert locust [15], the hawkmoth [16], the red flour beetle [18], and the Madeira cockroach [19] (Fig. 3A). The standard atlases are calculated based on manually segmented image stacks (voxels were assigned with neuropil identities; Fig. 2). In a first step, global disparities in size, orientation, and position of the brain models are compensated (center transformation). Afterward, each individually segmented neuropil is rigidly transformed to the corresponding neuropil of the template brain by ignoring all other brain structures (local rigid transformation). Hereby, the overlap of the neuropils are maximized between template and individual brain and all segmented neuropils are merged to create a so-called “probability map” [12]. The standardized 3D atlas can be derived from this probability map by a threshold segmentation of this image stack. In addition, the transformation parameters are applied to the corresponding, raw images (Fig. 2). The VIB method automatically calculates mean volumes of all neuropils and the probability map reveals shape variabilities of brain areas. A method utilizing the Computational Morphometry Toolkit (CMTK) has been first applied to stacks of segmented images, similar to the VIB method [53]. Later, the method was directly adapted to raw image stacks to further improve the transformation accuracy and to avoid the unnecessary effort needed to create 3D brain models (Fig. 2) [54]. First, an affine transformation

Insect Brain Atlases

91

with nine degrees of freedom is applied to raw image stacks using an intensity-based algorithm that quantifies similarities between voxels [55, 56]. While applying the same intensity-based algorithm as during the affine registration, the brains are thereafter non-rigidly aligned by a so-called B-spline free-form deformation model [53, 57]. The basis of this model is a grid of uniformly spaced control points in the images that are shifted independently of neighboring control points during the non-rigid transformation. By an additional interpolation of the space between the control points homologous structures of different raw image stacks are aligned in a fully automated manner by deformation fields [53]. In several studies the non-rigid transformation step of the CMTK method has been applied iteratively to the original image stacks [13, 15, 17, 19, 33, 58]. Hereby, each iteration was performed with a newly created averaged image stack as template for the new non-rigid transformation (Iterative Shape Averaging, ISA [59]). The basic idea of this version of the CMTK method is to increase the transformation quality by using a continuously improving, averaged image stack as template, instead of an individual reference brain. The non-rigid transformation step can be applied iteratively until a satisfactory quality of the averaged image stack is reached [15]. This iterative process often results in an increasingly sharper and better-defined image stack of the brain [13]. In a further step, the deformation parameters that describe the transformations necessary to warp each raw image stack into the final result can be used to transform corresponding 3D reconstructions to a standardized 3D brain model. This is equally applicable to corresponding images of GFP expressing GAL4-lines (Drosophila) or immunohistochemical stainings present in the registered brains. The CMTK method enabled the generation of standard brain atlases of the honeybee [13], the fruit fly [14], the desert locust [15], the tobacco budworm (Heliothis virescens; [17]), and the Madeira cockroach [19] (Fig. 3A). In Drosophila, thousands of different GAL4-expression lines and a theoretically infinite number of single neurons generated by the “mosaic analysis with a repressible cell marker” (MARCM) flipout technique [60] create an ever-increasing catalog of anatomical data. Handling of these large amounts of data would be much simplified by a standardized brain atlas serving as common frame of reference to register all data into a single brain. To create an adequate 3D model for this purpose, the FlyCircuit standard atlas has been developed [20]. In the process of standardization, only two structures are standardized in a first step: the outer and inner brain surfaces. Similar to the VIB method, segmented image stacks are used for standardization and are aligned using a shape-based averaging method. First, reconstructed outer and inner brain surface models are transformed via a coarse, rigid transformation to a

92

Basil el Jundi and Stanley Heinze

“pseudo-average model” of those structures. Euclidian distance maps are generated for each pseudo-average model to measure the distances of voxels to the inner or outer brain surface. Finally, a cumulative distance map of all pseudo-average models defines the average surface of the 3D standard model. In a second step the mushroom bodies are standardized to compute the final FlyCircuit standard brain. They are standardized using an alignment method described by Wu et al. [61]. Individual mushroom body models are manually reconstructed from high-resolution images of the same set of brains. The average orientation of the principal axes of the lobes of the mushroom body models is determined and the mushroom bodies are non-rigidly aligned to these average axes. All other shown neuropils are segmented in a representative individual brain that exhibited the best fit with respect to the averaged brain surface and mushroom bodies, and are visualized together with the standardized structure. This allowed for the creation of two standardized Drosophila brain atlases, a male atlas and a female atlas [20]. Similar to the CMTK method, the BrainAligner algorithm is directly applied to raw image stacks (Fig. 2) [21]. This method enables a fast and fully automated alignment of brain images and corresponding expression patterns of GAL4 lines. The standardized atlas is again achieved in two steps: After an initial global affine alignment, image stacks are subjected to a local, nonlinear alignment. This nonlinear alignment is based on a “reliable landmarkmatching algorithm.” The algorithm compares and matches about 150 landmarks (number and location of landmarks are manually defined by the experimenter) between the template and individual image stacks. For this purpose, several parameters such as voxel intensity and correlation of local image patches are employed. Only landmarks that are located at similar places with respect to other landmarks (reliable landmarks) are considered for alignment via nonlinear geometrical warp. While the method allows for a quick alignment of image stacks to a standardized raw image stack, it has not been used to generate a standardized 3D model brain as in the other three methods. Although all methods share a similar standardization protocol, each has its own limitations and advantages. Comparison between the ISA version of the CMTK method and the VIB techniques by using the same set of locust and cockroach brains showed that a VIB standard brain is not an ideal platform for neuron registration but is optimized for quantitative volumetric comparison of neuropils [15, 19] (see Subheading 7). In contrast, the CMTK method is more suited to create a brain atlas that enables the registration of individual neurons into a common frame of reference [14, 33, 58, 62]. However, this technique, especially the ISA version, is relatively time-consuming and is thus not practical for aligning large numbers of brains. Chiang et al. [20] and Peng et al. [21] generated 3D standard atlases of Drosophila by techniques that require

Insect Brain Atlases

93

much less computational time and thus enable aligning of several hundred Drosophila brains. The BrainAligner and CMTK methods use raw image stacks for standardization. Due to the higher information content of raw images compared to segmented image stack, these algorithms yield much higher precision. Peng et al. [21] compared the BrainAligner method with the CMTK technique and argues that the BrainAligner method generates the most accurate standardized brain atlas, with the least computational demands and least running speed. This landmark-based algorithm, as applied to fly brains [21], therefore seems to be a promising and fast approach to generate future standard atlases.

5

Standardization of the Central Complex When dealing with brains of large insects, two factors limit the quality of the resulting brain atlas. First, the working distance of objectives used in confocal microscopy becomes smaller with increasing resolving power. Second, the computational power needed to average images increases dramatically at high resolution, that is, when using small voxel sizes. The problem of limited working distance of high-power objectives is not easily avoided. Currently, high-resolution imaging of whole-mount brains of large species is limited by a 25 long-working-distance objective (NA 0.8) providing lateral resolution of ca. 200–350 nm and an axial resolution of 1.3–2.3 μm, depending on wavelength (400–700 nm). The problem of limited computational power is less severe and will likely be solved in the future by more powerful hardware and software to handle large datasets. However, it currently sets the lower limit of voxel size in atlases covering whole brains of large insects to roughly 2  2  2 μm. Both problems have been tackled by generating standardized versions of sets of functionally connected neuropils rather than atlases of whole brains (Fig. 5). The neuropils chosen were the central complex (CX) and regions closely associated with it (additionally in the desert locust: lateral accessory lobes (LAL) [58] additionally in the monarch butterfly: LAL and the anterior optic tubercles) [33]. While together they span the center of the central brain, each of these brain areas consists of several subunits, whose neuroarchitecture is highly conserved across insects [9, 63]. They play a crucial role in encoding skylight compass cues in migratory and nonmigratory insects [64–69]. Data obtained in Drosophila have shown that the CX is needed for memorizing behaviorally relevant features of the visual environment and for computing a robust head-direction signal [70–74]. It is also critically involved in controlling motor programs that guide walking behavior [75– 80]. Combined, these data conceptually place the CX at the boundary of sensory processing and motor control and make it a likely

94

Basil el Jundi and Stanley Heinze

Fig. 5 Standardized atlases of neuropils involved in compass navigation. (A) Single optical section of a confocal image stack used for image segmentation in the monarch butterfly. Segmented neuropils are highlighted in color. The complete data stack was generated by merging several individual stacks. The yellow lines indicate the abrupt boundaries of the individually scanned confocal image stacks. (B) Same section as in A, but all grayscale values further away than 10 voxels (1  1  1 μm each) from any neuropil of interest were removed (masking). (C) Similar section as B, upon applying a standardization procedure (iterative shape averaging of 10 preparations). (D) Standardized atlas of the compass neuropils of the monarch butterfly (oblique view). (E) Standardized atlas of the central complex and lateral accessory lobes (LAL) of the desert locust (oblique view). (F) Different strategies for standardization of the upper division of the central body (CBU) in the monarch butterfly (left) and the locust (right). In the locust, all layers of the CBU (I–III) have been included in the atlas, while in the monarch butterfly, the individual layers (I–IV, top left) have not been included (lower left). (G) Tangential neurons of the lower division of the central body (CBL) registered into the standard atlas of the monarch butterfly, illustrating the emergence of a stratified subcompartmentation of the CBL, based solely on registered neurons. Abbreviations: ir immunoreactivity, AOTU anterior optic tubercle, UU/LU/NU upper/ lower/nodular unit, SP strap, BU bulb, LBU/MBU lateral/medial bulb, PB protocerebral bridge, NO noduli, ds/vs dorsal/ventral shell. All images are based on data originally employed in other studies on monarch butterfly [33] and desert locust [58]

Insect Brain Atlases

95

place in the insect brain to be responsible for selecting an appropriate behavior in response to the current sensory situation [81]. Because understanding this structure will likely provide new insights into the general mechanisms of how insect brains work, the CX has received considerable attention over the last decade. The fact that the neuroarchitecture of the CX is tightly linked to physiological characteristics of CX neurons [26] was yet another reason that a 3D atlas of these brain areas became desirable. The methods to achieve this differed between locusts and monarch butterflies. In locusts, the reconstructions were based on confocal image stacks obtained from thick vibratome sections (250 μm) of 20 brains. The brains were labeled with antibodies against synapsin and serotonin, using the staining protocol from whole-mount preparations, albeit with shorter incubation times and without collagenase digestion [58]. The contiguous sections were scanned with a 20 oil immersion objective (NA 0.7; Leica). After the resulting image stacks were manually registered and merged into one continuous stack at a final voxel size of 2  2  2 μm, they were used to reconstruct the CX-neuropils in unprecedented detail. In this way, the three horizontal layers of the upper division of the central body (CBU), as well as the compartments of the noduli and the subdivisions of the LAL (Fig. 5E, F), were included. As synapsin labeling was not sufficient to delineate the boundaries of all CBU layers, the latter were defined based on serotonin labeling. In the monarch butterfly, images were acquired directly from whole-mount preparations using a 25 oil immersion, long distance objective (NA 0.8; Zeiss). Therefore, no physical sectioning was needed to acquire image stacks covering the CX, the LAL, and the anterior optic tubercles at high resolution [33] (Fig. 5D). The ten used brains were labeled with antibodies against synapsin and treated identically to preparations used for low-resolution whole-brain reconstructions. In contrast to the locust, no subcompartments were reconstructed within the LAL, the noduli, and the CBU (Fig. 5D, F). The final voxel size used in the monarch butterfly was 1  1  1 μm. Although the use of sectioned brains solved the problem of limited working distance of the available objectives in the locust, an additional problem emerged. Because of anisometric shrinkage of neuropil structures of different tissue densities, the section faces were frequently distorted and, due to unavoidable tissue loss during the cutting process, neighboring sections could not be precisely aligned. Nevertheless, acceptably merged image stacks that covered the CX as well as the LAL were obtained in the majority of preparations. Their quality was tested by verifying that the neuropil volumes of these high-resolution reconstructions resembled those of the low-resolution locust brain atlas [58]. In both the monarch butterfly and the locust, the ISA implementation of the CMTK technique was used to align raw image stacks and to calculate a

96

Basil el Jundi and Stanley Heinze

standardized 3D model of the neuropils. As this protocol utilizes automatically detected image contrast to determine the transformation parameters, a problem occurs that is absent in whole-brain reconstructions. As the image stacks do not include the brain edges, their boundaries end abruptly within brightly stained tissue. Thus, the boundaries of the stacks provide a high-contrast “pseudo-landmark” within the final images, which is independent of neuropil location (Fig. 5A, yellow lines). To prevent the CMTK-algorithm to align the boundaries of the image stacks rather than the stained neuropil structures, each image stack was cropped at a fixed distance (10 voxels in all three dimensions) beyond the reconstructed neuropil surfaces. This procedure thus provided meaningful outer boundaries for all raw image stacks, which were subsequently used as input for the standardization algorithm (Fig. 5B). As registration of neurons into a common frame of reference was the goal for both species, it is possible—in retrospect—to evaluate advantages and disadvantages of each approach for practical use. Generally, including subcompartments of neuropils provides additional landmarks for a more precise neuron registration process (see next section), particularly within large neuropils like the LAL or the CBU. However, to use the full detail of the standardized locust CX, the included subcompartments of the atlas require that these structures are reconstructed for each registered neuron. As some of the subtle subcompartment boundaries are only visible with anti-serotonin and anti-synapsin double labeling, the required effort is substantial. Although this problem can be overcome when neuropil subcompartments are combined into single 3D structures, this eliminates the advantages of including subcompartments in the first place. In the monarch butterfly, the only boundaries used for the creation of the standardized neuropils were those that were also recognizable based on background staining alone, that is, without antibody labeling. Therefore, dye-injected neurons can be directly registered based on whole-mount preparations, without additional histological processing. Interestingly, after numerous neurons had been registered into this common frame of reference, layers and subcompartments became apparent solely by inspecting the branching patterns of the registered neurons [33] (Fig. 5G). Although not assessed quantitatively, defining the outer boundaries of neuropils thus appears to provide sufficient information for mapping their internal structures. Without noticeable loss of precision, this simpler method has therefore yielded a higher number of neuron registrations in the monarch butterfly (55 neurons; [33]) compared to the locust (8 neurons; [58, 82]), even though the locust standard CX has been available for a substantially longer time.

Insect Brain Atlases

6

97

Registration of Single Neurons The registration of neurons into a standardized atlas is one of the crucial steps required to analyze neuronal networks within a common reference system. The basic idea is to fit cells from individual brains into the standardized atlas by displacing and deforming the neurons as needed. Some studies registered neurons as they appear in the raw image stacks [14, 21] while others used 3D models for the same purpose (e.g., [13, 15, 17, 33, 58, 62]). Registration of 3D reconstructed neurons requires three main steps: image acquisition, 3D reconstruction of the neuron, and registration of the reconstruction into the atlas.

6.1 Generation of Image Data

Neurons in insect brains often exhibit large branching patterns combined with minute neurites, making the comprehensive imaging of individual cells challenging, as high resolution as well as large fields of view are required. To investigate potential connections between neurons at the cellular level, the imaging resolution should be as high as possible. In Drosophila, confocal images of neurons can be directly obtained from whole-mount preparations at high resolution (40 objective). Whole-mount preparations of larger brains can only be imaged using objectives with a large working distance, which typically implies low magnification (e.g., 25 objectives) (Fig. 4A, F). The elongated shape of the point-spread function in confocal imaging results in a blur in the z-dimension of the image stacks (Fig. 4F). To overcome this problem and to obtain high-resolution images of the neurons in question, Permount embedded, whole-mount brain preparations can be rehydrated by a decreasing ethanol series and subsequently embedded in albumin/gelatin to be cut into thick sections with a vibratome (130 μm) [31, 58, 83, 84]. After additional staining steps (for details, see Ref. [31]), these brain sections are mounted between two coverslips. Due to their reduced thickness compared to wholemount preparations, each section can be imaged by using objectives with shorter working distances. This results in image stacks with a much higher resolution (Fig. 4F, H). Depending on the size of the neuron, neurites and ramifications can be distributed over several sections. Therefore, each part of the neuron needs to be imaged separately and the complete neuron can only be visualized after merging all image stacks. As each brain section shows individual tissue shrinkage and distortions [37] the merging process requires manual alignment of the physically non-overlapping image stacks and inaccuracies are hard to avoid. On the other hand, in addition to enabling high-resolution imaging, this method enables combining the neuron labeling with staining against neurotransmitters or synaptic markers (often necessary for the following registration step described below) [31]. Taken together, depending on the required

98

Basil el Jundi and Stanley Heinze

final resolution and the size of the imaged brains, neurons can be reconstructed either based on images directly acquired from wholemount preparations, or reconstructed from high-resolution image stacks obtained from brain sections. 6.2 Neuron Reconstruction

Several methods have been implemented to reconstruct individual neurons in 3D in the insect brain [85–87]. In Drosophila, automated neuron tracing software has been created in conjunction with the FlyCircuit database, with the aim to reconstruct a large number of neurons [86]. First, the gray values of the image stack of the neuron are threshold-segmented and transformed into a binary matrix of either positive (contains neuron) or negative voxels (no neuron present). Based on the cell body as the initiation point, the neuron is reconstructed by calculating the shortest path from the cell body to each terminal of the neuron along the positive voxels of the binary matrix [20, 88]. This enables a relatively fast extraction of the midlines of neural fibers and an overall accurate visualization of the main branches of the cell, however, without any information about fiber diameters in the 3D model. Furthermore, this method is not suited to reconstruct detailed branching patterns and small arborizations, as they might be eliminated in the threshold-based preprocessing steps [88]. A similar basic idea for neuron reconstruction has been used in the neuron tracing tool implemented in the V3aaD (formally referred to as V3D) software [89, 90]. By marking the beginning and terminals of the neuron, the program automatically traces the fiber paths in between these points with high accuracy [87, 89]. Although easy to use for highquality image stacks showing arborization trees of low density, it has not been tested with intricate, dense branching patterns. The most widely used method to reconstruct neurons is the skeletonize tool, available as plug-in for the Amira software [13, 14, 17, 19, 33, 58, 62, 82, 83, 91–95] and introduced by Schmitt et al. [85]. This method performs a semiautomated reconstruction of neurons by successively fitting the diameter and midline of individual elements of a manually (or semimanually) traced neuron skeleton to the raw image stack by an intensity-based algorithm [85, 96]. This results in an accurate 3D model of the cell, with a precisely traced neural path and an adjusted diameter for each segment of the neuron. For the visualization, the original image stack of the neuron can be used as reference to generate a triangulated surface representation of the neuron, or the skeleton tree can be visualized directly (Fig. 4H–J). In contrast to the other methods, this tracing technique is relatively time-consuming and requires manual placing of many skeleton segments before fitting [85]. While the methods by Peng et al. [87] and Lee et al. [88] offer a tool to automatically reconstruct the main branches of cells in a fast way, Schmitt et al. [85] created a semiautomated neuron tracing tool that is best suited to comprehensively reconstruct individual neurons from high-resolution image stacks.

Insect Brain Atlases

6.3 Registration of Single Neurons

99

Registration of neurons into a standard atlas always requires two channels. One of them is the raw image stack of the neuron, or the 3D model of the cell itself (target channel), while the other one is an image stack consisting of either a neuropil marker immunostaining or the segmented neuropils (reference channel). In general, neurons are registered into the standardized brain based on the reference channel. The neuropils of the reference channel are positioned and deformed until they are accurately aligned with the corresponding brain areas of the standard atlas. The displacement and deformation parameters are then applied to the target channel to register the neurons into the standard atlas (Fig. 6). In Peng et al. [87] and Jefferis et al. [14], raw grayscale image stacks from whole-mount Drosophila brain preparations were directly used to register neuropil stainings into the standard atlas (reference channel). Afterward, the resulting displacement and deformation parameters from each individual registration were applied to the respective target channel to align imaged GAL4 lines into a common reference frame. For this registration, the authors used the same alignment algorithm as for the creation of the standard brain (see Subheading 4 and Fig. 2). In a second approach in Drosophila, Chiang et al. [20] also relied on raw image stacks as reference channel. They used a global affine registration with 12 (4  3) degrees of freedom (translation, rotation, scaling, shearing) to roughly transform the reference channel to the atlas and subsequently applied these transformation parameters to the raw image stack of the target channel. Afterward, the neurons were reconstructed from the transformed target channels with the tracing program described above [86] (see Subheading 6.2). Both methods are highly suited to quickly register a large number of cells into an atlas, but due to the omission of fine details of neural branches from the automated reconstruction, they do not seem to be sufficient to analyze potential connectivities of cells at fine spatial scales. However, the registration of a large number of neurons into the standard atlas via the method of Chiang et al. [20] revealed locally restricted brain areas with common innervation patterns that were defined as local processing units (LPUs) (see Subheading 7). Most frequently, the segmented image stacks of innervated neuropils were used to map neurons into standard atlases [13, 15, 17, 19, 33, 58, 62, 82, 91–95, 97]. This method requires the 3D reconstruction of the innervated brain areas, in addition to the neuron reconstruction. The segmented image stack of these brain areas then serves as reference channel (Fig. 6B). In an initial step, the neuropils of the reference channel are transformed to match the corresponding standardized brain areas by an affine transformation. Subsequently, these neuropils are deformed by an elastic (non-rigid) transformation in several iterations, resulting in a close match of neuropil shape between reference channel and the atlas (Fig. 6C). The elastic transformation is based on a B-spline

100

Basil el Jundi and Stanley Heinze

Fig. 6 Registration of single neurons into standardized brain atlases. (A) Maximal intensity projection (frontal view) of a CPU1 neuron of the monarch butterfly’s central complex. (B) Skeletonized reconstruction of the neuron shown in A, together with 3D surface reconstructions of the neuropils innervated by the cell (based on background fluorescence; oblique view). (C) Illustration of the two step protocol used for matching the raw neuropil reconstruction to corresponding neuropils in the standardized atlas. First, an affine transformation is performed that matches orientation and size of the neuropils. Second an iterative elastic transformation is carried out that matches local shape differences of the neuropils to the standard atlas. Both steps result in transformation parameters, which are subsequently used to transform the skeleton tree of the neuron. (D) Illustration of the neuron transformation achieved by applying the transformation parameters obtained in C. (E) Final result of the neuron transformation process. The registered neuron is displayed with all relevant neuropils of the standardized atlas. Abbreviations: CBU/CBL upper/lower division of the central body, PB protocerebral bridge, LAL lateral accessory lobe. All images are based on data originally employed in another study [33]

(free-form deformation) algorithm applied on uniformly distributed control points in the segmented image stack. The obtained transformation parameters are then applied to the target channel to accordingly deform the 3D model of the neuron

Insect Brain Atlases

101

(Fig. 6D). Finally, the elastically transformed neuron is integrated into the standardized atlas (Fig. 6E). Generally, this registration method requires a high amount of computational power and is time consuming when applied to the full resolution dataset. This can be solved by downsampling the data of the reference channel during the elastic registration step [19, 33]. However, the greater the voxel size of the reference channel, the lower the precision of the registered location of the neuron in the target channel. To overcome the computational power limitations in another way, el Jundi et al. [58] registered individual branching areas of the neurons stepby-step with the highest possible accuracy, at the expense of adequate registration of the neurites interconnecting the individual arborization trees. One of the most time-consuming steps of the registration process is the manual segmentation of images to assign neuropil boundaries in the reference channel. To reduce this time and to minimize the potential of human error during this process, Rybak et al. [62] created a statistical surface model (SSM) based on the standard atlas of the honeybee. The model provides an a priori information about the location of neuropil boundaries, thus enabling an automatic, model-based segmentation of defined neuropils. These are then used to obtain the displacement and deformation parameters necessary for registration of the target channel. Taken together, the precision of neuron registrations into 3D brain atlases is determined by two independent factors: First, the precision of the neuron reconstruction, and second, the accuracy of the registration process itself. The first factor depends on the quality of the target channel (e.g., neuron image resolution). Different reconstruction methods inevitably represent a trade-off between the amount of information extracted from available images and the amount of time needed to do that. In principle, it should be possible to combine different methods for neuron reconstruction with different mapping algorithms to achieve the best possible outcome for a specific scientific question. Ultimately, it is the scientific question the 3D brain atlas aims to solve, that dictates the necessary accuracy for the registration of neurons into a standard atlas.

7

Functional Applications Having explored the methodical aspects of how insect brain atlases have been generated in a diverse set of species, in the next section we will point out functional conclusions that were drawn from these atlases. Additionally, we will highlight the limitations of each approach, concluding that not every insect brain atlas is universal.

102

7.1

Basil el Jundi and Stanley Heinze

VIB Protocol

The first approach for generating standardized brain atlases that has been used in a variety of species is the Virtual Insect Brain (VIB) protocol. Standard atlases have been obtained in this way for the fruit fly, a hawkmoth (Manduca sexta), the red flour beetle, the desert locust, and the Madeira cockroach. The original aim of the VIB protocol was to generate a standard atlas that could be used as a reference for analyses of absolute neuropil volumes within a species and between species. This protocol is optimized for such comparisons, because neuropils are not scaled to match a reference brain during the standardization procedure of the VIB protocol [15, 19]. Hence, each VIB-standard brain provides a fixed population of neuropil volumes, which can be used as a static reference for comparisons to different developmental stages, more experienced individuals, or mutant strains. This has been exemplified for Drosophila through comparing different wild-type strains with one another, as well as comparing structural brain mutants to the Drosophila standard brain [12]. Interestingly, a trend toward less elaborate brains, particularly in higher processing centers such as the mushroom body, was observed in wild-type strains that had been kept in captivity for many generations. The comparison between well-characterized mutants (e.g., “reduced optic lobes”) and the standard brain revealed subtle differences in neuropil volumes that had not been previously described, indicating that comparisons using volumetric standard atlases lead to more comprehensive descriptions of structural brain defects in mutant fly strains. Comparison between male and female standard atlases has been performed in the hawkmoth Manduca sexta, reliably reproducing the previously known sexual dimorphism within antennal lobes [16]. Using the same method, no difference between the sexes was found for other brain regions of the same species, or for any part of the brain in the beetle Tribolium castaneum [18], suggesting that no such dimorphism exists in these insects on the level of neuropil volumes. Importantly, as the actual standard brain, that is, the standardized 3D model, is not required for analysis, the defined set of brains used as input for the VIB-protocol can also be directly used for volumetric analysis. Additionally, the comparison of VIB and ISA protocols has shown that the VIB protocol is not suited to map reconstructions of single neurons into the standard atlas [15, 19]. This is mostly due to the fact that neuropils are transformed independently of each other with this method, that is, the space in between the defined neuropils is disregarded. As this space is traversed by neurons that connect different neuropils, no meaningful path can be calculated for these connecting fibers in the VIB standard atlas. Taken together, as more accurate and more widely applicable methods for generating standardized 3D models of insect brains have become available and the standardization

Insect Brain Atlases

103

procedure per se is not required for any volumetric analysis, the VIB protocol has in principle become obsolete and is unlikely to be the method of choice in any future application. 7.2

CMTK Method

The method used for the majority of standard brain atlases so far was the CMTK method (mostly implemented in the ISA protocol). Intended as a platform for registration of single neurons into a common frame of reference, it had been developed for the honeybee [13], and was adopted for the moth Heliothis virescens [17], as well as for the standardized central complexes (CXs) of the locust and the monarch butterfly [33, 58]. This method was also used to generate secondary standard atlases for the desert locust and the Madeira cockroach, with the aim to compare the result with the already existing VIB standard atlas [15, 19]. The reference brain used for the registration of higher-order olfactory neurons in Drosophila was also generated using this algorithm, albeit without iterative shape averaging [14]. As the aim of these atlases was primarily to register neurons into a common frame of reference to detect potentially interacting neurons, knowing the precision of this registration process is crucial for drawing conclusions about possible connectivity. If certain neurons are in close proximity in the standard atlas, how likely is this also the case in individual brains? While the whole-brain atlases were all based on confocal image stacks obtained with a 10 objective (either dry, water immersion, or oil immersion), the standardized CXs were based on data imaged with a 20 or 25 objective. These a priori differences in image quality (Fig. 4A–E) as well as limited computational power resulted in trade-offs with respect to resolution of the final atlases. Voxel sizes of the atlases ranged from 5.9  5.9  3 μm in the locust (whole brain) to 1  1  1 μm in the monarch butterfly (“compass neuropils”), with the remaining atlases featuring intermediate values (Heliothis: 1.1  1.1  3.9 μm; honeybee: 3.8  3.8  3.8 μm; cockroach: 2.93  2.93  2 μm; locust CX: 2 2  2 μm). When registering neurons, the overall precision about the location of arborizations in the standard atlas is limited by the lowest effective resolution during any step of the registration process. Consequently, it is typically the z-dimension of the used image stacks that is the limiting factor, even though resolution of the images in the xy plane is markedly better. The z-dimensions of voxels should ideally be at least twice smaller than the optical resolution limit of the used objectives (Nyquist-Shannon sampling theorem; cf. Fig. 4A). In some standard brains (locust whole brain, cockroach; both imaged with a 10 oil immersion objective), the data are oversampled. This so-called “empty magnification” does not increase their precision beyond the optical resolution in the zdimension. As neurons are frequently scanned with high-power objective (e.g., with a 40 objective after rehydration and sectioning; [58]), high-resolution raw data are then mapped onto a much

104

Basil el Jundi and Stanley Heinze

less precise standard atlas, misleadingly suggesting a high precision of the overall dataset. Importantly, when evaluating the location of two registered neurons, no conclusion about the proximity of fibers can be drawn beyond the lower resolution limit imposed by the voxel size employed in neuropil registration. Neurons are often mapped onto the standard atlas by registering the segmented images of the neuropils innervated by the neurons of interest using a two-step protocol (affine and elastic registration). The use of segmented images rather than raw image data provides more flexibility in the way how images are obtained, so that the protocols to stain the reference channel of individual neurons do not have to match the procedures originally used to create the standard. For example, neuropils can be reconstructed based on background staining or autofluorescence. Whereas such raw data cannot be registered onto a standardized raw image stack that is based on anti-synapsin staining, the segmented images based on background fluorescence can be easily registered onto segmented images based on anti-synapsin labeling. However, with increasing distance from well-defined neuropil surfaces the registration becomes increasingly restricted to the affine transformation [33]. This means that local deformations present within these under-represented areas in individual brains are not corrected and fiber positions are only imprecisely mapped onto the standard, without an easy way to quantify this imprecision. Therefore, the highest precision (limited by the voxel size) only applies to registered neurons arborizing in close proximity to reconstructed surfaces of well-defined neuropils. Additionally, when neuropil boundaries are imposed on brain regions that are only ambiguously defined (e.g., some boundaries of the LAL or the subesophageal ganglion), the ambiguity of each individual reconstruction is largely ignored by imposing a sharp, consensus boundary. This can lead to severe stretching and misplacement of arborizations and to a mismatch between the apparent and real precision of the data in the vicinity of this boundary. Consequently, this means that methods like distance mapping between individually registered neurons cannot be used to prove synaptic contacts between these cells, as the uncertainty in the standardized data is greater than the precision required to identify individual synapses. These methods can only be used to exclude synaptic contacts between neurons that are sufficiently far apart from one another. Overall, these points underline the fact that to draw conclusions about potential connectivity of neurons embedded in the standard atlas one always has to be aware of the (im)precision and quality of the data originally used to generate the atlas. Despite these limitations, the different CMTK standard atlases (in particular the ISA implementation) have all been successfully used to register individual neurons into a common frame of reference (Fig. 7). In the honeybee, work has focused on neurons of the

Insect Brain Atlases

105

Fig. 7 Standardized brain atlases as a prerequisite for functional insights. (A) Reconstructions of polarization sensitive neurons of the monarch butterfly registered into the standardized atlas of the compass neuropils. Top: input and intermediate stage. Bottom: output stage. The combined display in the same frame of reference revealed input and output projections as well as detailed comparison of neural pathways involved in sky compass navigation to data known from the desert locust. (B) Schematic illustration of the sky compass network of the central complex in the desert locust. Names of neuron types are shown in italics. Neurons interconnecting the anterior optic tubercles (AOTU) were omitted for clarity (but included in A). Double arrows above the protocerebral bridge (PB) indicate mapping of polarized-light tuning directions of TB1 and CPU1 neurons. The asterisks (∗) mark the main output projections. Colors are identical to those used for neurons shown in A. (C) Registration of an input neuron (giant fan-shaped neuron [GFN]) to the upper division of the central body (CBU) and an output neuron (CPU1) into the standardized atlas of the locust’s central complex.

106

Basil el Jundi and Stanley Heinze

olfactory pathway and included local interneurons of the antennal lobe, antennal lobe projection neurons, and extrinsic neurons of the mushroom bodies [13, 62, 97]. These studies provide valuable visualizations of previously described neurons and illustrate the intricate anatomical organization of the dual olfactory pathway to the mushroom body calyx and the lateral horn. Similarly, olfactory and gustatory neurons have been registered into the standard atlas of the moth Heliothis virescens [91–93], while visual interneurons were mapped onto the standardized brain and CX of the locust [15, 58, 82]. However, the collected data in these atlases so far represent a relatively small set of neuronal morphologies that provide a proof-of-principle, but need to be complemented by more data to allow for novel, functional conclusions. Nevertheless, in two cases previously unknown pathways and new projection patterns of neurons have been revealed by registration of many cells into a standard atlas. Firstly, in the monarch butterfly, more than 55 neurons have been mapped onto a standardized version of the compass neuropils [33]. The registered neurons revealed two parallel pathways from different compartments of the anterior optic tubercle, via distinct regions of the bulbs, to separate layers of the CBL (Fig. 7A). The existence of these CBL strata and their distinct supply of input fibers clearly emerged from neuron registration and were thereafter confirmed by a detailed analysis of individual neurons. Similarly, it has been shown that different types of columnar output neurons from the CX project to separate, nonoverlapping regions of the LAL. Overall, the standard atlas of the monarch butterfly’s compass neuropils was successfully used to map major input and output pathways of the CX, refine the compartmentation of neuropils, and characterize the 3D layout of major tracts and fiber bundles supplying the CX. Secondly, in an elegant study in Drosophila, Jefferis and colleagues [14] mapped a large number of antennal lobe projection neurons onto a standardized reference brain. These neurons supply the mushroom body calyx and the ä Fig. 7 (continued) This mapping revealed overlap of presumed output fibers of the GFN cell with presumed input fibers of the CPU1 neuron in layer II of the CBU, potentially allowing for information exchange between these cell types. (D) Registration of fibers exhibiting PDF immunoreactivity in the optic lobe of the cockroach (Leucophaea [now Rhyparobia] maderae). The 3D information obtained via registration of transmitter systems can be used to quantitatively analyze projection patterns, identify candidate target regions of immunoreactive cells, and serve to reveal brain areas coexpressing other neuromodulators. (E, F) Identification and characterization of the heterolateral projection pattern of CPU1 neurons of the monarch butterfly central complex. (E) Dorsofrontal view of multiple CPU1 neurons mapped onto the standardized atlas of the compass neuropils. (F) Schematic view of the arborization pattern inferred from neuron registrations. Other abbreviations: UU/LU/ NU upper/lower/nodular unit, BU bulb, LBU/MBU lateral/medial bulb, CBL lower division of the central body, LAL lateral accessory lobes, dLAL/vLAL dorsal/ventral LAL, ME medulla, LA lamina, AME accessory medulla, aLobl anterior lobelet, PDF pigment dispersing factor. Images A-E are based on data originally employed in other studies: A, E [33, 64], B [65], C [58], and D [19]. Data for D were kindly provided by H. Wei, and a view near-identical to that shown in [19] is reproduced by permission of Wiley

Insect Brain Atlases

107

lateral horn with olfactory information. Through the registration process the authors uncovered that different types of olfactory information were likely mapped in a topographic way to different regions of the mushroom body calyx and to the lateral horn, an otherwise featureless neuropil. Additionally, comparing averaged lateral horns of male and female flies, regions exhibiting sex-specific local size differences were identified, suggesting that these neuropil compartments might be involved in pheromone processing. Apart from registration of single neurons, a comprehensive mapping of the expression profile of a neuroactive substance has been carried out in the Madeira cockroach, based on immunofluorescence labeling [19]. The substance chosen was the neuropeptide pigment-dispersing factor (PDF), which is thought to constitute the major output signal of the circadian clock network. Due to limits in computational power, the registration had to be performed on a coarsely downsampled dataset (voxel size 8  8  8 μm). Nevertheless, the result provides an impressive visualization of a complete neuropeptide system in an insect brain (Fig. 7D). With this unprecedented 3D information available, the authors were able to measure distances between all PDF expression domains in the cockroach brain. Based on these measurements they formulated new hypotheses with respect to encoding of circadian information, which would not have been possible without the used standard atlas. According to the authors, all distances between major PDF immunoreactive fiber plexi were equal or integer multiples of one another. This morphological feature (one could refer to it as “quantal spacing”) was postulated by the authors as being essential to maintain phase information of neuronal activity in these cells. This could be a prerequisite for the synchronizing effect PDF exerts on bilateral neuronal ensembles in the cockroach brain [98, 99]. In the monarch butterfly, volumetric comparisons have also been performed using the population of compass neuropils that were used to create the ISA-based standard atlas [33]. Although the final volumes of the standard atlas do not provide meaningful data, as they are the result of isomorphic scaling of all individual brains to match a preselected reference brain [13, 15], the raw data used as input to the ISA protocol nevertheless do represent the range of neuropil volumes that comprise the standard atlas. This defined distribution of reference volumes for each neuropil was therefore used to compare the volumes in the standard atlas (generated from naive butterflies) to those of experienced or inexperienced migratory forms of the monarch butterfly, inexperienced aged animals, and nonmigratory monarchs that had been experienced local foragers [33]. Interestingly, an age-related increase in absolute neuropil volumes was identified that was independent of sensory or behavioral experience of the animals. When only relative sizes of

108

Basil el Jundi and Stanley Heinze

neuropils were considered, enlargement of the protocerebral bridge was shown to correlate with migratory experience of the butterflies, consistent with the proposed role of this brain region in representing skylight compass cues [26, 64]. Overall, volumetric comparisons of neuropils between species have so far not been performed with enough scrutiny to enable major new insights into how brain organization on the level of neuropils correlates with the behavior of a particular species. Although a standardized brain atlas per se is not required to perform such analyses [100, 101], it nevertheless has the potential to significantly facilitate volumetric comparisons by providing a definite set of volumes that can serve as a reference distribution to all subsequent work. However, such comparisons are severely hampered by differences in the way these brain atlases have been generated. For example, the CX was separated into its compartments in some species (e.g., [15, 16, 18]), but not in others (e.g., [13, 17]) (Fig. 3B). In some species, the unstructured protocerebrum was not included in the atlas (e.g., [18, 19]; Fig. 3A), while the lamina of the optic lobes was only included in two of the seven standard atlases [16, 19] (Fig. 3B). At last, the volume of the antennal lobe was defined in most species as the combined volumes of the central antennal lobe neuropil and the glomeruli volumes. Contrary to that, in Heliothis virescens the antennal lobe volume was defined to exclusively contain the glomerular layer [17, 92]. These differences imply that relative neuropil sizes are difficult to compare between species. Ideally, identical structures reconstructed using the same criteria for defining neuropil boundaries should be used in all species. This would enable not only direct volumetric comparison, but also registration of data from one species onto data from another one, thus allowing identification of more subtle differences in shape and neuropil location. 7.3

FlyCircuit Atlas

Recently, one of two competing approaches to provide a comprehensive map of all neurons comprising the Drosophila brain has yielded a massive database, named FlyCircuit [20]. To enable registration of all data into a single frame of reference a new standard atlas was generated. More than 16,000 single neuron data-stacks, generated by MARCM flip-out technique [60], were mapped into this reference brain. This dataset then made it possible to identify all major tracts and fiber pathways of the fly brain and was used to reveal brain regions likely serving as functional units. These so-called local processing units (LPUs) were based on the projection fields of local interneurons and the presence of defined projections to and from other brain areas. In this way, the fly brain was divided into 41 LPUs and 6 smaller regions without local interneurons, named “hubs.” These new brain divisions matched previously defined neuropil boundaries in many areas, but also revealed differences and mismatches with previous neuropil concepts

Insect Brain Atlases

109

[10, 20, 22]. While further research is needed to prove whether LPUs do indeed represent true functional compartments, the LPU concept as such seems to be an elegant way to predict (or at least estimate) function from anatomical data. Although the protocols used for neuron registration into the FlyCircuit atlas offer the possibility for local affine as well as elastic registration, only global affine registration was used for the initial mapping of all 16,000 neuron morphologies [20]. This poses some limitations: As only very few structures have been used as registration landmarks, the neurons are only coarsely matched to the reference brain. Individual variations in shape and size of neuropils are thus not taken into account, thus limiting the possible conclusions about the overlap of arborization trees on spatial scales below the LPU-level. Additionally, automated neuron reconstruction and annotation tools mark neuropils as innervated by a reconstructed neuron. This results in relatively numerous false-positive results, when querying the database for neurons innervating a particular brain region. Additionally, the LPU-based neuropil definition leads to different boundaries of brain regions with identical names (compared to earlier definitions) and also assigns new names to regions with unchanged boundaries. Therefore, interpretation of neuronal arborization data and comparison to previous results from other species has to be performed with great care. Nevertheless, the FlyCircuit database so far represents the most comprehensive and most readily accessible collection of single neuron morphologies in any insect species and is thus of immense value to the scientific community not only in the field of Drosophila neurobiology. 7.4 BrainAligner Method

Most recently, the second approach aimed at comprehensive mapping of the Drosophila brain has resulted in a high-resolution (voxel size: 0.58  0.58  0.84 μm) standardized atlas based on averaging 295 individual brains [21]. This approach was fully automated and optimized for high-throughput. It has been used to register many thousands of GAL4 expression lines, each labeling a different set of Drosophila brain neurons, into a common frame of reference [21, 22]. The used algorithm relies on gray-value based landmarks manually placed throughout the image stack according to the local density of image information, that is, each brain region can be defined by as many (or as few) landmarks as its complexity demands. Such an approach has two key advantages: First, during the generation of the atlas, the registration processes are orders of magnitude faster than during any other image-based registration algorithm (e.g., the ISA protocol), as not every voxel has to by registered, but still provide a superior, nearly distortion-free output [21]. Second, when the same algorithm is used to register neurons or groups of labeled neurons into the atlas, high registration accuracy is not limited to the surfaces of well-defined neuropils but is distributed evenly throughout the preparation. This overcomes the

110

Basil el Jundi and Stanley Heinze

above-described limitations of protocols based on reconstructed neuropil surfaces. The quality and versatility of this approach has been demonstrated by registering Drosophila brains with overlapping GAL4 expression domains into the same frame of reference and comparing the output with biological co-expression of the same lines [21]. The registration results predicted the biological coexpression with high accuracy, thus validating the biological relevance of the registration results and opening up a path toward mapping of single neurons, immunostaining data, in situ hybridization patterns, and so on into the same standardized brain atlas. Given that the BrainAligner software is freely available and can be implemented on virtually any computer, the method is likely the most promising tool for quick and high-quality registration of insect brains. Nevertheless, it still has to be shown, whether this algorithm, which has been developed using high-contrast and highresolution confocal image stacks from Drosophila, is equally useful for aligning image stacks from larger insect brains, which usually exhibit lower staining quality and much lower resolution, particularly in the z-dimension. One weakness inherent to all standard atlases is the implicit assumption that neurons from individual brains within a given species are more or less invariant. The concept of mapping neurons from many individuals onto a common frame of reference only leads to meaningful data if a neuron from one brain can be replaced by a corresponding cell from another brain. However, this is clearly an over-simplification, as exposure to different environments, developmental conditions, or learning experiences will lead to different behavior among individual animals, which must result from more or less subtle changes of neuronal morphology. Accordingly, it has been shown in honeybees and ants that foraging experience and age differences result in changes in the density of connections at the input stage to the mushroom body calyx (e.g., [102, 103]). Differences in the summed experience of an insect during its life may thus lead to major differences in neuronal branching patterns and consequently to structural changes in neuropils. Therefore, one needs to firstly know the state of the animals used to generate the standard atlas. Secondly, it is equally important to keep track of the state of the individual from which the neuron to be registered is reconstructed. If, for example, a standard atlas is based on animals that have been inbred and deprived of sensory input for many generations, it may not be ideally suited for mapping neurons from wild-caught animals of the same species. Likewise, single neuron morphologies obtained from individuals with very different sensory experiences might suggest false connectivities when registered into a common frame of reference, as in reality, such connections might never occur. Ideally, in each species, a separate atlas would be generated for both sexes, for naive animals, as well as

Insect Brain Atlases

111

for each state of the animal that bears neurobiological relevance. This is clearly not achievable in many species. The anticipated application for each atlas thus has to be known before selecting the animals to generate it. If a standard brain is obtained from a population of mixed sex, age and experience, it might be a suitable average atlas to visualize all neurons of that species at low resolution, but it will clearly not be suited to compare differences in branching patterns of cells that correlate with any of those features. Potential applications of 3D atlases that have not been explored so far include the quantitative comparison of differences in neuropil shapes between related species, (e.g., diurnal versus nocturnal representatives of the same genus). This requires that homologous structures represented in each atlas are included for all considered species and that they have been reconstructed using the same criteria to define boundaries. Only then can the interspecies registration of brain regions, required for these analyses, work reliably. Such differences would highlight specific regions of the brain, in which adaptive changes have occurred between the examined species. These findings would hence provide ideal starting points for physiological studies. With more powerful computers and recently developed fast algorithms (e.g., the BrainAligner), the limitations that used to restrict the generation of standardized 3D brain atlases to only a few model species are no longer in place. Broad, comparative studies attempting to correlate individual species’ lifestyles with the characteristics of their brains have thus become feasible and are likely to shed more light on the principles underlying the evolution of brains and behavior in insects.

8

Implementing Data from Other Imaging Techniques into Standard Atlases Within each species a unified reference atlas can also be used for registration of image data obtained with methods other than confocal microscopy. Particularly interesting in this context are image stacks resulting from block-face electron microscopy and microcomputed tomography (μCT), as both of them contain information about the 3D layout of the brain. A number of challenges in relation to reliable image registration limit the possible conclusions from implementing these results into an atlas based on lightmicroscopy. For registration, landmarks have to be defined that are valid across the different imaging methods. Given the rather big differences in contrast and resolution among the imaging methods (relative to confocal imaging of synaptic markers), this is likely difficult to achieve. Upon successfully registering the image stacks, the precision in locating image features within the stacks is limited by the resolution of the atlas itself. The implication is that no reliable conclusion can be drawn from comparing two independently registered image stacks within a distance smaller than twice

112

Basil el Jundi and Stanley Heinze

the voxel size of the atlas. This is a minor problem for methods with similar or lower resolution than confocal microscopy (μCT, lightsheet microscopy [also referred to as selective plane illumination microscopy, SPIM]), but has to be taken into account when interpreting registered electron microscopic images. For example, although neurons reconstructed from two independent block-face electron microscopical image stacks can be registered into a standard atlas, no conclusion can be drawn about any potential interactions between these cells at distances below twice the voxel size of the atlas. Nevertheless, with imaging techniques that utilize undissected or even living brains (e.g., μCT), important information about the native size and orientation of neuropils is obtained. This is largely impossible with fixed samples commonly employed in confocal microscopy. Generally, the implementation of a variety of image sources (imaging modalities) into a single insect brain atlas can serve as a convenient method for accessing such data on a common platform. This holds not only for 3D data stacks, but also for images and image series obtained with classical techniques, such as Bodian’s staining, Golgi impregnation, or neurons visualized through diaminobenzidine (DAB) precipitation [7, 8]. A wealth of such data is available for many species, but these are often inaccessible to large parts of the scientific community. An easy access to the multimodal imagery of brain regions of interest would surely make an atlas more informative and comprehensive.

9

Databases All 3D insect brain atlases generated so far were originally designed as tools for collecting morphological information about a particular species across research laboratories and compiling it in a format suitable for future functional studies. Unfortunately, even about a decade after the first standardized 3D brain atlases were implemented, almost none of them have acclaimed a wide user basis, and most have been used exclusively by the laboratory that has generated the atlas in the first place. The only exceptions to this trend are the recently added databases for Drosophila. Nevertheless, as all other atlases have been proven to be adequately suited for their original purpose, the fact that they have not gained a wider acceptance among the scientific community cannot be entirely linked to problems in the atlas design. Rather, one problem may lie in the 3D image data formats, which are quickly becoming obsolete due to rapid progress in digital media. This generates incompatibility problems and prevents atlas data from being used as long-term reference [6]. Second, another major factor that prevents the wider acceptance of 3D atlases even for short-term use is the lack of an accessible, highly visible, and easy-to-use platform enabling fast

Insect Brain Atlases

113

data retrieval, as well as simple ways of contributing data by third parties. Such online databases have been attempted for most atlases (e.g., [13, 15, 16, 18]) but have remained incomplete, do not allow data upload, or are no longer accessible with modern internet browsers. The reasons for this development are numerous and have been extensively reviewed by Ito [6]. In the fruit fly Drosophila, recent 3D atlases are complemented by databases that begin to address the above described problems. Additionally, a substantial effort is made to provide a conceptual framework for the anatomical understanding of the fly brain. Consequently, a unified, comprehensive nomenclature system for all regions and compartments of the fly brain has been developed, thus serving as the central reference for the large scientific community working on Drosophila neurobiology [10]. A standardized, morphological 3D atlas (i.e., a geometrical reference) is much more likely to appeal to a broad user community within this conceptual reference system. Similar efforts have only recently been made in other insect species, aided by the fact that the new Drosophila nomenclature system also includes brain regions not found in flies and can thus serve as a more general reference for all insect brains. Due to the multitude of examined species and the relatively small number of researchers working on each of them, a collaborative effort to build up a comprehensive database of different insect brains, aimed at comparative research, seems to be the only way to ensure long-term use of the 3D insect brain atlases so far generated in other species than Drosophila. Such a platform has been recently developed and is implemented as the InsectBrainDatabase (IBdb), a freely available online tool to deposit, manage, and share anatomical and functional data from insect brains.1 This database is suited to serve as a central reference for interspecies comparison of brain areas, homologous neurons, expression patterns of transmitters, and so on. It highlights one of the main advantages of research covering a broad range of insects, namely, the richness in neural solutions, their differences, and the concepts shared across species, all reflecting evolutionary history or adaptations to different behavioral strategies and habitat choices.

10

Conclusions Over the last decade, much progress has been made in optimizing staining protocols, the processes of image segmentation, and methods for 3D reconstructions of insect brains. The resulting data have been used to generate a variety of standardized 3D brain atlases in a

1

www.insectbraindb.org

114

Basil el Jundi and Stanley Heinze

range of insect species of many different shapes and sizes. Upon comparing these atlases, no single optimal protocol can be proclaimed as universally applicable to all species. Rather, new attempts for standard atlases should consult previous work in closely related species and use these protocols as starting point for optimizing all methodical steps. To generate a standardized atlas, the choice of the most appropriate data-processing protocol has to be guided by the intended use of the atlas. The chosen method will likely be a tradeoff between available processing power, computational time, desired resolution, and achievable staining quality. In parallel to developing 3D brain atlases, an equally impressive progress has been made in 3D reconstructions based on individually stained neurons, or groups of labeled neurons (immunolabeling or genetic labeling). Several ways of how such data can be obtained and integrated into standardized atlases have been highlighted in this review. Importantly, an atlas is most useful if the displayed structures are easy to compare with the existing atlases of other species. Only if this compatibility is ensured, one can gain meaningful functional insights based on interspecies anatomical comparisons and thus optimally explore the richness of insect diversity. Such atlases will ultimately have the power to facilitate our understanding of how insect species have “rewired” the neural networks in their brains to adapt to different environments and to adopt different behavioral strategies.

Acknowledgments For kindly providing data originally employed in other studies on standardized atlases of various species, we are thankful to Hongying Wei, Arnim Jenett, Martin Kollmann, Joachim Schachtner, Pa˚l Kvello, Hanna Mustaparta, Ju¨rgen Rybak, and Randolf Menzel (for details, see figure legends). We are grateful to Maxim Telle for carrying out 3D reconstructions of the brains of Megalopta genalis. We also thank Eric Warrant, Julia Schuckel, William Wcislo, and the staff at the Smithsonian Tropical Research Institute in Panama City for essential support with the work on Megalopta genalis, funded by grants to S.H. by the Swedish Research Council (VR 621-2012-2213) and a Marie Curie Intra-European Fellowship (IEF 2012-327901). We also thank two anonymous referees for valuable suggestions that have improved this chapter, as well as Radek Pelc for excellent editing and improvements to the microscopy sections.

Insect Brain Atlases

115

Glossary2 3D-reconstruction The process of generating a virtual, threedimensional representation of a real object. Affine transformation A linear transformation characterized by 6 degrees of freedom in 2D (2 translation, 1 rotation, 2 scaling, 1 shearing) or 12 degrees of freedom in 3D (3 translation, 3 rotation, 3 scaling, 3 shearing). An affine registration has often been applied using only 9 degrees of freedom (3 translation, 3 rotation, 3 scaling). Amira A 3D-analysis software by Visualization Sciences Group (now part of FEI, Hillsboro, OR, USA). It enables image processing, image segmentation, and 3D visualization. Average brain The result of a process in which multiple individual brains are registered onto a ►template brain and merged into a single, averaged brain. An average brain consists either of averaged raw image stacks (grayscale values) or averaged segmented image stacks (neuropil reconstructions) and is often used as a standardized atlas (►standard brain). BrainAligner An automatic registration program capable of aligning 3D image stacks using a landmark-matching algorithm; written by the research group of Hanchuan Peng (Howard Hughes Medical Institute, Janelia Research Campus, Ashburn, VA, USA), and available online [21]. Computational Morphometry Toolkit (CMTK) An imageprocessing software comprising a set of tools for computational analysis of biomedical images (image processing, correction and registration); written by Torsten Rohlfing (Stanford University, CA, USA). Many standardized atlases are based on the registration tools of this software freely available online.3 Elastic/non-rigid transformation A nonlinear transformation enabling (in contrast to the ►rigid transformation) local stretching of the image in order to align (register) local image features with corresponding features in a reference image. FIJI A freely available image-processing package based on the opensource software ImageJ. It comprises a large variety of plug-ins specifically designed for bio-imaging applications, and is available online.4 FlyCircuit The name of a project in the research group of A.-S. Chiang (University of Taiwan, Republic of China). The project generated a large database comprising more than 16,000 individual The ►symbol denotes ►another entry in the glossary. http://www.nitrc.org/projects/cmtk 4 http://fiji.sc/Fiji 2 3

116

Basil el Jundi and Stanley Heinze

Drosophila brain neurons registered into a standard atlas. The FlyCircuit website provides a search engine for these neurons and allows the user to search for cells based on the brain areas they innervate. In addition, the website5 enables uploading neuron image stacks and to register these neurons online into the FlyCircuit atlas (via affine registration). It is also linked to the Virtual Fly Brain homepage,6 thus offering a comprehensive and detailed anatomical description of the Drosophila brain. Image registration or Image alignment The process of transforming an image (or image stack) onto a reference image (or image stack) by matching features of the image in question to the corresponding features of the reference image by applying transformation algorithms (rigid, affine, non-rigid). Image segmentation The process of subdividing an image into different compartments (e.g., into different brain areas). During this process each ►voxel of an image stack is assigned a unique identity, indicating that it belongs to one particular compartment (e.g., brain region) only. The result is a segmented image stack, in which only the ►voxel identity is preserved, as original intensity values are no longer relevant. Iterative Shape Averaging (ISA) protocol An image-registration protocol designed for insect brain standardization that uses the open-source ►CMTK image registration software package. It enables averaging of multiple image stacks after aligning them with respect to a reference image stack. For alignment, an initial rigid transformation is followed by a non-rigid transformation, iteratively applied to the original image stacks (an average image stack from the last iteration serves as a reference for the next one). Neuropil An anatomical term for a defined area within the insect brain. Neuropils are devoid of cell bodies and consist of axons, dendrites and glial cells. One can distinguish synaptic neuropil and fibrous neuropil. Fibrous neuropil comprises tracts, fiber bundles, and commissures, and contains only very sparse synaptic contacts, while synaptic neuropils are the principal areas in which synaptic contacts between neurons are formed. These regions are highlighted by synaptic markers (neuropil labeling). Numerical aperture (NA) A fixed parameter of the microscope objective, which determines its resolving power (resolution) that is proportional to 1/NA (in x-y-dimension), or approximately to n/NA2 (in z-dimension); n is refractive index of the environment immediately adjacent to the objective lens (immersion liquid or air). Point spread function (PSF) The three-dimensional diffraction pattern resulting from imaging an object that is much smaller than the

5 6

http://www.flycircuit.tw http://www.virtualflybrain.org

Insect Brain Atlases

117

resolution limit of the objective used. Fluorescent microspheres are frequently used to measure PSF in a fluorescence (confocal) microscope. The size of the point spread function determines the theoretical resolution limit of a given imaging setup in the x-y and zdimension (see ►Numerical aperture). Resolution is directly proportional to the wavelength of the emitted light (in a confocal microscope, it also depends on the wavelength of the excitation light). Reference brain The individual (real) brain that is used as a template for the registration process during standardization. As the final standard brain is heavily influenced by the size and anatomy of the reference brain, it is critically important to carefully consider parameters such as staining quality and deviations in brain size or neuropil locations from the population average when deciding which brain to select as the reference brain. Registration (of two images) A synonym for the process of their alignment. Rigid transformation A linear transformation that only involves translation and rotation. In two dimensions, a rigid transformation consists of three degrees of freedom (2 translations, 1 rotation). A rigid transformation in three dimensions consists of six degrees of freedom (3 translations, 3 rotations). Skeletonize A plug-in for the ►Amira software created by J. F. Evers (University of Heidelberg, Germany). It is a semiautomatic 3D neuron reconstruction tool available on request from Jan Felix Evers.7 Skeleton tree A three-dimensional representation of a neuron. Each skeleton tree comprises a hierarchical sequence of branches connected by branch points. Skeleton trees created with the ►skeletonize tool consist of linear strings of branch elements (“snaxels”), each defined by its midline and diameter. Standard brain A virtual brain that represents the typical, mean brain layout of a defined population of a species. It can either be realized in form of the individual brain that most closely resembles the population average, or it can be an actual average brain. Standard insect brains are available for the following species: Drosophila melanogaster (for available downloads, see ►FlyCircuit and ►VIB-protocol), Apis mellifera, Schistocerca gregaria, Manduca sexta (available upon request from Basil el Jundi, University of Wu¨rzburg, Germany), Tribolium castaneum, Heliothis virescens, Leucophaea maderae (now: Rhyparobia maderae); available in an online supplement to [19]). Honeybee and desert locust standard brains are available online at the Insect Brain Database.8

7 8

https://www.cos.uni-heidelberg.de/index.php/j.evers?l¼_e www.insectbraindb.org

118

Basil el Jundi and Stanley Heinze

Standardization, in this context The process of registering image stacks of a multitude of individual brains onto one another to generate an averaged brain atlas. This minimizes individual differences and ensures that the difference between each individual brain (of the ones used to create the standard) and the standard brain is smaller than the difference between any two individual brains. If the used brains are a representative sample from the overall population, this relation should hold up for all brains of that species. Stitching The process of combining two or more image stacks into a single contiguous image stack, based on identifying mutual image information present in the overlapping regions of the original image stacks. Template brain, in this context An image stack that serves as the initial reference for the alignment of other image stacks. To generate a ►standard brain, all image stacks of individual brains are transformed onto the image stack of the template brain during the registration process. Triangulated surface model A vector model of a three-dimensional surface defined by a finite number of triangles. Vaa3D A 3D-analysis software (previously called V3D) created in the laboratory of Hanchuan Peng (Howard Hughes Medical Institute, Janelia Research Campus, Ashburn, VA, USA). It enables processing of large image stacks (e.g., their stitching), visualization of 3D data, and neuron reconstruction, and is available online.9 Vector field A three-dimensional matrix of vectors. It assigns a vector to each voxel of an image stack, indicating the amplitude and direction of local transformations. Virtual Insect Brain (VIB) protocol A standardization method created in the research group of Martin Heisenberg (University of Wu¨rzburg, Germany). The VIB script is available online as a plugin for ImageJ (freeware) and the ►Amira software.10 Voxel A three-dimensional equivalent of pixel (volume element). While a pixel represents the elementary component of a two-dimensional raster image, voxel is the element of a regular orthogonal grid in a three-dimensional space. Its location is therefore defined by three coordinates (x, y, and z).

9

http://www.vaa3d.org http://www.neurofly.de

10

Insect Brain Atlases

119

References 1. Clarac F, Pearlstein E (2007) Invertebrate preparations and their contribution to neurobiology in the second half of the 20th century. Brain Res Rev 54:113–161. https://doi.org/ 10.1016/j.brainresrev.2006.12.007 2. Ramo´n y Caja´l S (1888) Morfologia y conexiones de los elementos de la retina de las aves. Revista Trimestral de Histologı´a Normal y Patolo´gica 1(1):11–16. http://zaguan. unizar.es/record/4583?ln¼en 3. Wullimann MF, Rupp B, Reichert H (1996) Neuroanatomy of the zebrafish brain: a topological atlas. Birkh€auser, Basel. https://doi. org/10.1007/978-3-0348-8979-7 4. Valverde F (1998) Golgi atlas of the postnatal mouse brain. Springer, Vienna. https://doi. org/10.1007/978-3-7091-6501-0 5. Woolsey TA, Hanaway J, Gado MH (2013) The brain atlas: a visual guide to the human central nervous system, 3rd edn. Wiley, Hoboken, NJ. ISBN: 978-1118438770 6. Ito K (2010) Technical and organizational considerations for the long-term maintenance and development of digital brain atlases and web-based databases. Front Syst Neurosci 4:26. https://doi.org/10.3389/fnsys.2010. 00026 7. Rybak J (2013) Exploring brain connectivity in insect model systems of learning and memory. In: Menzel R, Benjamin PR (eds) Invertebrate learning and memory, Handbooks of behavioral neuroscience, vol 22. Academic (Elsevier), London, pp 26–40. https://doi. org/10.1016/B978-0-12-415823-8.000046 8. Rybak J (2012) The digital honey bee brain atlas. In: Galizia CG et al (eds) Honeybee neurobiology and behavior. Springer, New York, pp 125–140. https://doi.org/10. 1007/978-94-007-2099-2_11 9. Strausfeld NJ (2012) Arthropod brains. Harvard University Press, Cambridge, MA. ISBN: 9780674046337 10. Ito K, Shinomiya K, Ito M, Armstrong JD, Boyan G, Hartenstein V, Harzsch S, Heisenberg M, Homberg U, Jenett A, Keshishian H, Restifo LL, Ro¨ssler W, Simpson JH, Strausfeld NJ, Strauss R, Vosshall LB, Insect Brain Name Working Group (2014) A systematic nomenclature for the insect brain. Neuron 81:755–765. https://doi.org/10. 1016/j.neuron.2013.12.017 11. Strausfeld NJ (1976) Atlas of an insect brain. Springer, Berlin. https://doi.org/10.1007/ 978-3-642-66179-2

12. Rein K, Zo¨ckler M, Mader MT, Gru¨bel C, Heisenberg M (2002) The Drosophila standard brain. Curr Biol 12:227–231. https:// doi.org/10.1016/S0960-9822(02)00656-5 13. Brandt R, Rohlfing T, Rybak J, Krofczik S, Maye A, Westerhoff M, Hege H-C, Menzel R (2005) Three-dimensional average-shape atlas of the honeybee brain and its applications. J Comp Neurol 492:1–19. https:// doi.org/10.1002/cne.20644 14. Jefferis GSXE, Potter CJ, Chan AM, Marin EC, Rohlfing T, Maurer CR, Luo L (2007) Comprehensive maps of Drosophila higher olfactory centers: spatially segregated fruit and pheromone representation. Cell 128:1187–1203. https://doi.org/10.1016/ j.cell.2007.01.040 15. Kurylas AE, Rohlfing T, Krofczik S, Jenett A, Homberg U (2008) Standardized atlas of the brain of the desert locust, Schistocerca gregaria. Cell Tissue Res 333:125–145. https://doi.org/10.1007/s00441-0080620-x 16. el Jundi B, Huetteroth W, Kurylas AE, Schachtner J (2009) Anisometric brain dimorphism revisited: implementation of a volumetric 3D standard brain in Manduca sexta. J Comp Neurol 517:210–225. https://doi.org/10.1002/cne.22150 17. Kvello P, Løfaldli BB, Rybak J, Menzel R, Mustaparta H (2009) Digital, threedimensional average shaped atlas of the Heliothis virescens brain with integrated gustatory and olfactory neurons. Front Syst Neurosci 3:14. https://doi.org/10.3389/neuro. 06.014.2009 18. Dreyer D, Vitt H, Dippel S, Goetz B, el Jundi B, Kollmann M, Huetteroth W, Schachtner J (2010) 3D standard brain of the red flour beetle Tribolium castaneum: a tool to study metamorphic development and adult plasticity. Front Syst Neurosci 4:3. https://doi.org/10.3389/neuro.06.003. 2010 19. Wei H, el Jundi B, Homberg U, Stengl M (2010) Implementation of pigmentdispersing factor-immunoreactive neurons in a standardized atlas of the brain of the cockroach Leucophaea maderae. J Comp Neurol 518:4113–4133. https://doi.org/10.1002/ cne.22471 20. Chiang A-S, Lin C-Y, Chuang C-C, Chang H-M, Hsieh C-H, Yeh C-W, Shih C-T, Wu J-J, Wang G-T, Chen Y-C, Wu C-C, Chen G-Y, Ching Y-T, Lee P-C, Lin H-H, Wu C-C, Hsu H-W, Huang Y-A, Chen J-Y,

120

Basil el Jundi and Stanley Heinze

Chiang H-J, Lu C-F, Ni R-F, Yeh C-Y, Hwang J-K (2011) Three-dimensional reconstruction of brain-wide wiring networks in Drosophila at single-cell resolution. Curr Biol 21:1–11. https://doi.org/10.1016/j.cub. 2010.11.056 21. Peng H, Chung P, Long F, Qu L, Jenett A, Seeds AM, Myers EW, Simpson JH (2011) BrainAligner: 3D registration atlases of Drosophila brains. Nat Methods 8:493–500. https://doi.org/10.1038/nmeth.1602 22. Jenett A, Rubin GM, Ngo T-TB, Shepherd D, Murphy C, Dionne H, Pfeiffer BD, Cavallaro A, Hall D, Jeter J, Iyer N, Fetter D, Hausenfluck JH, Peng H, Trautman ET, Svirskas RR, Myers EW, Iwinski ZR, Aso Y, DePasquale GM, Enos A, Hulamm P, Lam SCB, Li H-H, Laverty TR, Long F, Qu L, Murphy SD, Rokicki K, Safford T, Shaw K, Simpson JH, Sowell A, Tae S, Yu Y, Zugates CT (2012) A GAL4-driver line resource for Drosophila neurobiology. Cell Rep 2:991–1001. https://doi.org/10. 1016/j.celrep.2012.09.011 23. Namiki S, Kanzaki R (2011) Heterogeneity in dendritic morphology of moth antennal lobe projection neurons. J Comp Neurol 519:3367–3386. https://doi.org/10.1002/ cne.22754 24. Namiki S, Takaguchi M, Seki Y, Kazawa T, Fukushima R, Iwatsuki C, Kanzaki R (2013) Concentric zones for pheromone components in the mushroom body calyx of the moth brain. J Comp Neurol 521:1073–1092. https://doi.org/10.1002/ cne.23219 25. Krofczik S, Menzel R, Nawrot MP (2008) Rapid odor processing in the honeybee antennal lobe network. Front Comput Neurosci 2:9. https://doi.org/10.3389/neuro.10. 009.2008 26. Heinze S, Homberg U (2007) Maplike representation of celestial E-vector orientations in the brain of an insect. Science 315:995–997. https://doi.org/10.1126/science.1135531 27. Meseke M, Evers JF, Duch C (2009) Developmental changes in dendritic shape and synapse location tune single-neuron computations to changing behavioral functions. J Neurophysiol 102:41–58. https:// doi.org/10.1152/jn.90899.2008 28. Egger R, Dercksen VJ, de Kock CPJ, Oberlaender M (2014) Reverse engineering the 3D structure and sensory-evoked signal flow of rat vibrissal cortex. In: Cuntz H et al (eds) The computing dendrite: from structure to function. Springer, New York, pp 127–145. https://doi.org/10.1007/978-1-46148094-5_8

29. Miller JA, Ding SL, Sunkin SM, Smith KA, Ng L (2014) Transcriptional landscape of the prenatal human brain. Nature 508:199–207. https://doi.org/10.1038/nature13185 30. N€assel DR, Homberg U (2006) Neuropeptides in interneurons of the insect brain. Cell Tissue Res 326:1–24. https://doi.org/10. 1007/s00441-006-0210-8 31. Heinze S, Homberg U (2008) Neuroarchitecture of the central complex of the desert locust: intrinsic and columnar neurons. J Comp Neurol 511:454–478. https://doi. org/10.1002/cne.21842 32. Heinze S, Reppert SM (2012) Anatomical basis of sun compass navigation I: the general layout of the monarch butterfly brain. J Comp Neurol 520:1599–1628. https://doi.org/10. 1002/cne.23054 33. Heinze S, Florman J, Asokaraj S, el Jundi B, Reppert SM (2013) Anatomical basis of sun compass navigation II: the neuronal composition of the central complex of the monarch butterfly. J Comp Neurol 521:267–298. https://doi.org/10.1002/cne.23214 34. Cachero S, Ostrovsky AD, Yu JY, Dickson BJ, Jefferis GSXE (2010) Sexual dimorphism in the fly brain. Curr Biol 20:1589–1601. https://doi.org/10.1016/j.cub.2010.07. 045 35. Ott SR (2008) Confocal microscopy in large insect brains: zinc-formaldehyde fixation improves synapsin immunostaining and preservation of morphology in whole-mounts. J Neurosci Methods 172:220–230. https:// doi.org/10.1016/j.jneumeth.2008.04.031 € 36. Spalteholz W (1911) Uber das Durchsichtigmachen von menschlichen und tierischen Pr€aparaten. S. Hirzel, Leipzig. http://nbnresolving.de/urn:nbn:de:hbz:38m:1-534 37. Bucher D, Scholz M, Stetter M, Obermayer K, Pflu¨ger HJ (2000) Correction methods for three-dimensional reconstructions from confocal images: I. Tissue shrinking and axial scaling. J Neurosci Methods 100:135–143. https://doi.org/10.1016/ S0165-0270(00)00245-4 38. Klagges B, Heimbeck G, Godenschwege TA, Hofbauer A, Pflugfelder GO, Reifegerste R, Reisch D, Schaupp M, Buchner S, Buchner E (1996) Invertebrate synapsins: a single gene codes for several isoforms in Drosophila. J Neurosci 16:3154–3165 39. N€assel DR (1996) Advances in the immunocytochemical localization of neuroactive substances in the insect nervous system. J Neurosci Methods 69:3–23. https://doi. org/10.1016/S0165-0270(96)00016-7

Insect Brain Atlases 40. Ott SR, Rogers SM (2010) Gregarious desert locusts have substantially larger brains with altered proportions compared with the solitarious phase. Proc Biol Sci 277:3087–3096. https://doi.org/10.1098/rspb.2010.0694 41. Dent JA, Polson AG, Klymkowsky MW (1989) A whole-mount immunocytochemical analysis of the expression of the intermediate filament protein vimentin in Xenopus. Development 105:61–74. PMID:2806118. https://www.ncbi.nlm.nih.gov/pubmed/ 2806118 42. Ott SR, Elphick MR (2003) New techniques for whole-mount NADPH-diaphorase histochemistry demonstrated in insect ganglia. J Histochem Cytochem 51:523–532. https:// doi.org/10.1177/002215540305100414 43. Huetteroth W, Schachtner J (2005) Standard three-dimensional glomeruli of the Manduca sexta antennal lobe: a tool to study both developmental and adult neuronal plasticity. Cell Tissue Res 319:513–524. https://doi. org/10.1007/s00441-004-1016-1 44. Huetteroth W, el Jundi B, el Jundi S, Schachtner J (2010) 3D-reconstructions and virtual 4D-visualization to study metamorphic brain development in the sphinx moth Manduca sexta. Front Syst Neurosci 4:7. https://doi. org/10.3389/fnsys.2010.00007 45. Gonzalez-Bellido PT, Wardill TJ (2012) Labeling and confocal imaging of neurons in thick invertebrate tissue samples. Cold Spring Harb Protoc 2012:969–983. https://doi. org/10.1101/pdb.prot069625 46. Amos B, McConnell G, Wilson T (2012) Confocal microscopy. In: Egelman EH (ed) Comprehensive biohysics, Biophysical techniques for characterization of cells, vol 2. Elsevier, Amsterdam, pp 3–23. https:// doi.org/10.1016/B978-0-12-374920-8. 00203-4 47. Wilhelm S (2010) Confocal laser scanning microscopy. Carl Zeiss Microscopy GmbH, Jena. Brochure 60-1-0030/e. ISBN: 9783940885029. https://www.zeiss.com/ microscopy/int/downloads/brochuredownloads (go to “Light Microscopy”) 48. Pawley JB (2006) Fundamental limits in confocal microscopy. In: Pawley JB (ed) Handbook of biological confocal microscopy. Springer, New York, pp 20–42. https:// doi.org/10.1007/978-0-387-45524-2_2 49. Diaspro A, Federici F, Robello M (2002) Influence of refractive index mismatch in high-resolution three-dimensional confocal microscopy. Appl Opt 41(4):685–690. https://doi.org/10.1364/AO.41.000685

121

50. Preibisch S, Saalfeld S, Tomancak P (2009) Globally optimal stitching of tiled 3D microscopic image acquisitions. Bioinformatics 25:1463–1465. https://doi.org/10.1093/ bioinformatics/btp184 51. Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, Preibisch S, Rueden C, Saalfeld S, Schmid B, Tinevez J-Y, White DJ, Hartenstein V, Eliceiri K, Tomancak P, Cardona A (2012) Fiji: an open-source platform for biologicalimage analysis. Nat Methods 9:676–682. https://doi.org/10.1038/nmeth.2019 52. Jenett A, Schindelin JE, Heisenberg M (2006) The Virtual Insect Brain protocol: creating and comparing standardized neuroanatomy. BMC Bioinformatics 7:544. https:// doi.org/10.1186/1471-2105-7-544 53. Rohlfing T, Brandt R, Maurer CRJ, Menzel R (2001) Bee brains, B-splines and computational democracy: generating an average shape atlas. In: Procceeding of IEEE workshop MMBIA 2001 (Mathematical methods in biomedical image analysis), Kauai, HI, USA, 9–10 Dec 2001, pp 187–194. https:// doi.org/10.1109/MMBIA.2001.991733 54. Rohlfing T, Brandt R, Menzel R, Maurer CR (2004) Evaluation of atlas selection strategies for atlas-based image segmentation with application to confocal microscopy images of bee brains. NeuroImage 21:1428–1442. https://doi.org/10.1016/j.neuroimage. 2003.11.010 55. Studholme C, Hill D, Hawkes DJ (1997) Automated three-dimensional registration of magnetic resonance and positron emission tomography brain images by multiresolution optimization of voxel similarity measures. Med Phys 24:25–35. https://doi.org/10. 1118/1.598130 56. Studholme C, Hill D, Hawkes DJ (1999) An overlap invariant entropy measure of 3D medical image alignment. Pattern Recogn 32:71–86. https://doi.org/10.1016/ S0031-3203(98)00091-0 57. Rohlfing T, Maurer CR (2003) Nonrigid image registration in shared-memory multiprocessor environments with application to brains, breasts, and bees. IEEE Trans Inf Technol Biomed 7:16–25. https://doi.org/ 10.1109/TITB.2003.808506 58. el Jundi B, Heinze S, Lenschow C, Kurylas A, Rohlfing T, Homberg U (2010) The locust standard brain: a 3D standard of the central complex as a platform for neural network analysis. Front Syst Neurosci 3:21. https:// doi.org/10.3389/neuro.06.021.2009

122

Basil el Jundi and Stanley Heinze

59. Rohlfing T, Brandt R, Menzel R, Russakoff DB, Maurer CR Jr (2005) Quo vadis, atlasbased segmentation? In: Suri JS, Wilson DL, Laxminarayan S (eds) Handbook of biomedical image analysis. Springer, New York, pp 435–486. https://doi.org/10.1007/0-30648608-3_11 60. Lee T, Luo L (1999) Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron 22:451–461. https://doi.org/10.1016/ S0896-6273(00)80701-1 61. Wu C-C, Chen C-Y, Chang H-M, Chiang A-S, Chen Y-C (2009) Improved two-level model averaging techniques in Drosophila brain modeling. In: Wada T, Huang F, Lin S (eds) Advances in image and video technology. Springer, Berlin, pp 921–931. https:// doi.org/10.1007/978-3-540-92957-4_80 62. Rybak J, Kuß A, Lamecker H, Zachow S, Hege H-C, Lienhard M, Singer J, Neubert K, Menzel R (2010) The digital bee brain: integrating and managing neurons in a common 3D reference system. Front Syst Neurosci 4:30. https://doi.org/10.3389/ fnsys.2010.00030 63. Homberg U (2008) Evolution of the central complex in the arthropod brain with respect to the visual system. Arthropod Struct Dev 37:347–362. https://doi.org/10.1016/j. asd.2008.01.008 64. Heinze S, Reppert SM (2011) Sun compass integration of skylight cues in migratory monarch butterflies. Neuron 69:345–358. https://doi.org/10.1016/j.neuron.2010. 12.025 65. Homberg U, Heinze S, Pfeiffer K, Kinoshita M, el Jundi B (2011) Central neural coding of sky polarization in insects. Philos Trans R Soc B 366:680–687. https://doi. org/10.1098/rstb.2010.0199 66. el Jundi B, Pfeiffer K, Heinze S, Homberg U (2014) Integration of polarization and chromatic cues in the insect sky compass. J Comp Physiol A 200:575–589. https://doi.org/10. 1007/s00359-014-0890-6 67. Pfeiffer K, Homberg U (2014) Organization and functional roles of the central complex in the insect brain. Annu Rev Entomol 59:165–184. https://doi.org/10.1146/ annurev-ento-011613-162031 68. el Jundi B, Warrant EJ, Byrne MJ, Khaldy L, Baird E, Smolka J, Dacke M (2015) Neural coding underlying the cue preference for celestial orientation. Proc Natl Acad Sci U S A 112:11395–11400. https://doi.org/10. 1073/pnas.1501272112

69. Homberg U, el Jundi B (2014) Polarization vision in arthropods. In: Werner L, Chalupa M (eds) New visual neuroscience. MIT Press, Cambridge, MA, pp 1207–1217. ISBN: 9780262019163 70. Liu G, Seiler H, Wen A, Zars T, Ito K, Wolf R, Heisenberg M, Liu L (2006) Distinct memory traces for two visual features in the Drosophila brain. Nature 439:551–556. https:// doi.org/10.1038/nature04381 71. Pan Y, Zhou Y, Guo C, Gong H, Gong Z, Liu L (2009) Differential roles of the fan-shaped body and the ellipsoid body in Drosophila visual pattern memory. Learn Mem 16:289–295. https://doi.org/10.1101/lm. 1331809 72. Ofstad TA, Zuker CS, Reiser MB (2011) Visual place learning in Drosophila melanogaster. Nature 474:204–207. https://doi.org/ 10.1038/nature10131 73. Seelig JD, Jayaraman V (2013) Feature detection and orientation tuning in the Drosophila central complex. Nature 503:262–266. https://doi.org/10.1038/nature12601 74. Seelig JD, Jayaraman V (2015) Neural dynamics for landmark orientation and angular path integration. Nature 521:186–191. https://doi.org/10.1038/nature14446 75. Strauss R (2002) The central complex and the genetic dissection of locomotor behaviour. Curr Opin Neurobiol 12:633–638. https:// doi.org/10.1016/S0959-4388(02)00385-9 76. Neuser K, Triphan T, Mronz M, Poeck B, Strauss R (2008) Analysis of a spatial orientation memory in Drosophila. Nature 453:1244–1247. https://doi.org/10.1038/ nature07003 77. Harley CM, Ritzmann RE (2010) Electrolytic lesions within central complex neuropils of the cockroach brain affect negotiation of barriers. J Exp Biol 213:2851–2864. https://doi. org/10.1242/jeb.042499 78. Triphan T, Poeck B, Neuser K, Strauss R (2010) Visual targeting of motor actions in climbing Drosophila. Curr Biol 20:663–668. https://doi.org/10.1016/j.cub.2010.02. 055 79. Guo P, Ritzmann RE (2013) Neural activity in the central complex of the cockroach brain is linked to turning behaviors. J Exp Biol 216:992–1002. https://doi.org/10.1242/ jeb.080473 80. Martin J, Guo P, Mu L, Harley CM, Ritzmann RE (2015) Central complex control of movement in the freely-walking cockroach (Blaberus discoidalis). Curr Biol

Insect Brain Atlases 25:2795–2803. https://doi.org/10.1016/j. cub.2015.09.044 81. Heinze S (2015) Neuroethology: unweaving the senses of direction. Curr Biol 25: R1034–R1037. https://doi.org/10.1016/j. cub.2015.09.003 82. Rosner R, Homberg U (2013) Widespread sensitivity to looming stimuli and small moving objects in the central complex of an insect brain. J Neurosci 33:8122–8133. https://doi.org/10.1523/JNEUROSCI. 5390-12.2013 83. el Jundi B, Pfeiffer K, Homberg U (2011) A distinct layer of the medulla integrates sky compass signals in the brain of an insect. PLoS One 6:e27855. https://doi.org/10. 1371/journal.pone.0027855.g011 84. el Jundi B, Homberg U (2012) Receptive field properties and intensity-response functions of polarization-sensitive neurons of the optic tubercle in gregarious and solitarious locusts. J Neurophysiol 108:1695–1710. https://doi.org/10.1152/jn.01023.2011 85. Schmitt S, Evers JF, Duch C, Scholz M, Obermayer K (2004) New methods for the computer-assisted 3-D reconstruction of neurons from confocal image stacks. NeuroImage 23:1283–1298. https://doi.org/10.1016/j. neuroimage.2004.06.047 86. Lee PC, Ching YT, Chang HM, Chiang AS (2008) A semi-automatic method for neuron centerline extraction in confocal microscopic image stack. In: Proceedings of the 5th IEEE international symposium on biomedical imaging (ISBI 2008), Paris, 14–17 May 2008, pp 959–962. https://doi.org/10.1109/ISBI. 2008.4541157 87. Peng H, Ruan Z, Atasoy D, Sternson S (2010) Automatic reconstruction of 3D neuron structures using a graph-augmented deformable model. Bioinformatics 26: i38–i46. https://doi.org/10.1093/bioinfor matics/btq212 88. Lee PC, Chuang C-C, Chiang A-S, Ching YT (2012) High-throughput computer method for 3D neuronal structure reconstruction from the image stack of the Drosophila brain and its applications. PLoS Comput Biol 8: e1002658. https://doi.org/10.1371/jour nal.pcbi.1002658 89. Peng H, Ruan Z, Long F, Simpson JH, Myers EW (2010) V3D enables real-time 3D visualization and quantitative analysis of large-scale biological image data sets. Nat Biotechnol 28:348–353. https://doi.org/10.1038/nbt. 1612

123

90. Peng H, Bria A, Zhou Z, Iannello G, Long F (2014) Extensible visualization and analysis for multidimensional images using Vaa3D. Nat Protoc 9:193–208. https://doi.org/10. 1038/nprot.2014.011 91. Kvello P, Jørgensen K, Mustaparta H (2010) Central gustatory neurons integrate taste quality information from four appendages in the moth Heliothis virescens. J Neurophysiol 103:2965–2981. https://doi.org/10.1152/ jn.00985.2009 92. Løfaldli BB, Kvello P, Mustaparta H (2010) Integration of the antennal lobe glomeruli and three projection neurons in the standard brain atlas of the moth Heliothis virescens. Front Syst Neurosci 4:5. https://doi.org/ 10.3389/neuro.06.005.2010 93. Zhao X-C, Kvello P, Løfaldli BB, Lillevoll SC, Mustaparta H, Berg BG (2014) Representation of pheromones, interspecific signals, and plant odors in higher olfactory centers; mapping physiologically identified antennallobe projection neurons in the male heliothine moth. Front Syst Neurosci 8:186. https:// doi.org/10.3389/fnsys.2014.00186 94. Zhao X-C, Pfuhl G, Surlykke A, Tro J, Berg BG (2013) A multisensory centrifugal neuron in the olfactory pathway of heliothine moths. J Comp Neurol 521:152–168. https://doi. org/10.1002/cne.23166 95. Boerner J, Duch C (2010) Average shape standard atlas for the adult Drosophila ventral nerve cord. J Comp Neurol 518:2437–2455. https://doi.org/10.1002/cne.22346 96. Evers JF, Schmitt S, Sibila M, Duch C (2005) Progress in functional neuroanatomy: precise automatic geometric reconstruction of neuronal morphology from confocal image stacks. J Neurophysiol 93:2331–2342. https://doi. org/10.1152/jn.00761.2004 97. Kirschner S, Kleineidam CJ, Zube C, Rybak J, Gru¨newald B, Ro¨ssler W (2006) Dual olfactory pathway in the honeybee, Apis mellifera. J Comp Neurol 499:933–952. https://doi. org/10.1002/cne.21158 98. Reischig T, Petri B, Stengl M (2004) Pigment-dispersing hormone (PDH)immunoreactive neurons form a direct coupling pathway between the bilaterally symmetric circadian pacemakers of the cockroach Leucophaea maderae. Cell Tissue Res 318:553–564. https://doi.org/10.1007/ s00441-004-0927-1 99. Schneider N-L, Stengl M (2005) Pigmentdispersing factor and GABA synchronize cells of the isolated circadian clock of the cockroach Leucophaea maderae. J Neurosci

124

Basil el Jundi and Stanley Heinze

25:5138–5147. https://doi.org/10.1523/ JNEUROSCI.5138-A-04.2005 100. Sivinski J (1989) Mushroom body development in nymphalid butterflies: a correlate of learning? J Insect Behav 2:277–283. https:// doi.org/10.1007/BF01053299 101. Gronenberg W, Ash LE, Tibbetts EA (2008) Correlation between facial pattern recognition and brain composition in paper wasps. Brain Behav Evol 71:1–14. https://doi.org/ 10.1159/000108607

102. Stieb SM, Muenz TS, Wehner R, Ro¨ssler W (2010) Visual experience and age affect synaptic organization in the mushroom bodies of the desert ant Cataglyphis fortis. Dev Neurobiol 70:408–423. https://doi.org/10.1002/ dneu.20785 103. Groh C, Lu Z, Meinertzhagen IA, Ro¨ssler W (2012) Age-related plasticity in the synaptic ultrastructure of neurons in the mushroom body calyx of the adult honeybee Apis mellifera. J Comp Neurol 520:3509–3527. https://doi.org/10.1002/cne.23102

Chapter 4 Neuroanatomical Tracing Based on Selective Fluorochrome Expression in Transgenic Animals Floris G. Wouterlood Abstract The present chapter focuses on new developments in neuroanatomical tracing techniques. Two recent innovations are highlighted: the use of transgenic animals and neurochemical fingerprinting. In the brains of transgenic cre-mice, expression of a fluorescent protein can be induced in small populations of neurons via injection of a nonreplicating virus containing a plasmid coding for enhanced yellow fluorescent protein (eYFP). Upon transfection and insertion of the plasmid into the neuron’s DNA, the plasmid is only expressed after a “repair” by DNA-recombinase of the genome in the transfected neuron. Essential here is that the recombinase protein is expressed only in neurochemically specific neurons. A small group of neurochemically specific neurons can thus be indirectly “instructed” to start manufacturing a fluorescent label. The fluorescence accumulates in all processes of the transfected neurons, including the fibers projecting into other regions of the brain. This method is illustrated in parvalbumin-cre and choline acetyltransferase-cre mice and complemented with the second innovation discussed in the chapter: neurochemical fingerprinting, that is, screening of axon terminals by a combination of tracing and immunofluorescence–confocal laser scanning microscopy. This type of screening is performed by confocal laser scanning microscopy whose high resolution is required for colocalization studies at the axon terminal level. Key words Neuroanatomical tracing, Molecular biology, Optogenetics, Fluorescent protein, Biotinylated dextran amine, Immunofluorescence, Confocal laser scanning

Abbreviations 5-HT AAV BDA ChAT ChAT+/
CLSM CNS Cre eYFP eYFP
eYFP+

5-Hydroxytryptamine (serotonin) Adeno-associated virus Biotinylated dextran amine Choline acetyltransferase ChAT (immuno)positive/negative Confocal laser scanning microscopy Central nervous system Recombinase (“causes recombination”) protein Enhanced yellow fluorescent protein Non-eYFP expressing eYFP expressing

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_4, © Springer Science+Business Media LLC 2020

125

126

Floris G. Wouterlood

FP HDB NA NeuN PBS PHA-L PV PV+/
PV-cre VDB

1

Fluorescent protein Horizontal limb of the diagonal band (of Broca) Numerical aperture Neuronal nuclear antigen Phosphate buffered saline Leucoagglutinin from Phaseolus vulgaris Parvalbumin PV (immuno)positive/negative (Mouse), coexpressing parvalbumin and DNA recombinase Vertical limb of the diagonal band (of Broca)

Introduction One of the aims of neuroanatomy is to collect morphological information about neuronal networks. In this respect, three conditions must be considered before selecting suitable techniques. First, individual neurons are cells with processes spreading in three dimensions, that in many, many cases are much longer than the thickness of the histological sections in which the cell bodies are located. Second, neuronal networks usually consist of participating neurons with different physiological and neurochemical phenotypes. Third, neuronal networks involve scale: they can be extremely simply organized (e.g., in a monosynaptic reflex arc), they can consist of relatively few interconnected neurons such as in cerebellar cortex, or they may include a large number of participating neurons such as for instance in the barrel cortex module. All these features translate to suitable combinations of techniques and approaches required to study the neuronal networks. For instance, if one is interested in the structural interconnectivity of individual cells within a given neuronal network it is necessary to distinguish each cell from the other ones via a unique identifier. The ultimate goal is to “create” a network wherein every single neuron has its own label, thus bearing similarity to the “brainbow” mouse model developed by Weissman et al. [1]. This translates to multistaining techniques where each “color” acts as a unique identifier for a particular neuronal species. The above conditions also affect a number of fundamental choices, for example, the decision to study networks exclusively via a whole mount approach (e.g., [2–4]), thick slices [5], procedures progressing from thick brain slices via resectioning to histological sections [6], or conventionally via sections obtained on a microtome or a cryostat. When the focus is on synapses, an electron microscopy approach employing serial ultrathin sections or block abrasion is one of the options [7]. The present chapter is limited to the classical histological approach, that is, to 25–40 μm thick histological sections.

Selective Fluorochrome Expression Tracing

127

Anatomical network analysis in the central nervous system (CNS) at the cellular level requires labeling techniques to track connectivity in such a way that the individual neurons participating in the network become identifiable with a unique marker. When this condition is met the axons belonging to these neurons can be traced to the sites where they form physical contacts with the processes belonging to other neurons in the network. The situation is made more complex by two additional conditions. First, if one applies a general immunohistochemical labeling method the entire population of neurons expressing the marker will massively respond. Second, with a classical neuroanatomical tracing technique not all types of neuron are visualized equally well. The latter condition requires a short explanation: Classical tracing methods typically yield good results when they target projection neurons whose axons extend far away from the tracer injection spot. By contrast, they usually generate poor results with interneurons with short axons or with axon ramifications limited to a small tissue volume, in particular neurons occupying a brain volume which is equal to or smaller than the envelope of the tracer’s injection spot. Even small injections (diameter 100–200 μm) with an anterograde tracer make it next to impossible to distinguish locally distributing interneurons from projection neurons because their local axonal distributions mix up (Fig. 1, bottom frame). Finally, classical tracing methods inherently suffer from the drawback that the uptake of the tracer substance by neurons occurs indiscriminately. The injection site typically includes neurons expressing a plethora of neurochemical phenotypes. All these neurons become labeled with the tracer, regardless of their neurochemical makeup. Within the injection spot, neurons expressing neurotransmitter “X” may equally well pick up the same tracer in an injection spot as neighboring neurons expressing another neurotransmitter (“Y”), a peptide or calcium binding protein, and start transporting it (Fig. 2).

2

Neuroactive Substance-Specific Tracing: Retrograde Approaches The ability to trace neurons of only one neurochemical variety by which the neurotransmitter is identified, the so-called neurotransmitter-specific tracing, has been the dream for many neuroanatomists from times immemorial. Several techniques have been developed, and can be grouped into two categories: retrograde and anterograde techniques. Retrograde neurotransmitter-specific tracing has been proposed more than 30 years ago [8] but for technical reasons only rarely applied in its original form (e.g., [9]). This method is based on injecting a radioactive isotope-tagged neurotransmitter, followed by its uptake in axon terminals and retrograde transport. It requires a radioautographical detection

128

Floris G. Wouterlood

Fig. 1 Elements of a neuron network made visible with neuroanatomical tracing. (Top) Schematic network. Both nuclei, nc “A” and nc “B” contain projection neurons (Pr) and interneurons (In). Nucleus “A” projects to “B” and vice versa. (Middle) A retrograde tracer deposited in nucleus “B” only marks projection neurons in the partner nucleus. (Bottom) An anterograde tracer in nucleus “A” labels both projection neurons and interneurons in the injection nucleus “A,” and fibers in the partner nucleus. Based on the tracer contents alone, the types of neuron (Pr and In) cannot be distinguished from each other

with long exposure times and, of course, proper facilities and radiation containment measures to conduct the experiments. Fortunately, there are suitable alternative retrograde techniques. By far the most popular one relies on retrograde fluorescence tracing

Selective Fluorochrome Expression Tracing

129

Fig. 2 Indiscriminate uptake of a neuroanatomical tracer. (Top) A nucleus (nc “A”) may contain neurons of all neurochemical phenotypes (indicated by different colors), for example, glutamatergic (Glu), GABAergic (GABA), dopaminergic (DA), cholinergic (ACh), serotoninergic (5-HT), or those containing parvalbumin (PV). Combinations of neurotransmitters and neuropeptides are common. (Bottom) Neuroanatomical tracing (e.g., by biotinylated dextran amine, BDA) labels all these different cells, fibers and axon endings in the same “color”

followed by fluorescence immunohistochemistry. This approach (Fig. 3, bottom frame), introduced in 1980 [10], has gained widespread popularity [11–16]. As retrograde neuroanatomical tracing methods employ low-molecular weight, highly diffusible fluorescent compounds (e.g., FluoroGold, MW 532.6 from Santa Cruz

130

Floris G. Wouterlood

Fig. 3 Retrograde tracing method for establishing the neurochemical identity of labeled neurons. (Top) A simple model of neuronal network consisting of three different chemical phenotypes indicated by different grayscale level. (Bottom) A combination of retrograde fluorescent tracing (injection in nucleus “B”) and immunofluorescence (in nucleus “A”) makes it possible to establish the neurochemical identity of projection neurons (Pr) in “A,” with the aid of fluorescent labels [red and green] coupled to appropriate antibodies [orange])

Biotechnology, Dallas, TX, USA),1 they tend to produce relatively big injection sites, roughly 100–500 μm in diameter. Such a large injection site makes it impossible to study interneurons with short axons. Within the injection spot, these cells cannot be distinguished from long-projection neurons (Fig. 1, middle frame). Interneurons located in the nucleus hosting the projection neurons that send fibers to the injection site do not become labeled, simply because their axons do not reach the injection site. Interneurons thus escape attention, this being the drawback of retrograde fluorescent tracing combined with immunofluorescence.

1

https://www.scbt.com/scbt/product/fluorogold-trade-223769-64-0

Selective Fluorochrome Expression Tracing

3

131

Neuroactive Substance-Specific Tracing: Anterograde Approaches This category includes more options than the retrograde one. In short, the following approaches have been introduced.

3.1 Neurotoxin Treatment Combined with Anterograde Tracing

A way to conduct “selective” tracing is to treat a brain area with a neurotoxin in such a way that only neurons belonging to one neurochemically specific type survive. The connectivity of the survivors can then be reliably traced with a neuroanatomical tracer [17]. A point of critique here is that the cellular network in the injection site is radically damaged by the forced elimination of all chemically nonspecific cells.

3.2 Intra- or Pericellular Approach

Until fairly recently the only patent way to visualize interneurons was via an intra- or pericellular approach, typically in a neurophysiological setting. Filling individual neurons, postrecording, with a dye reveals the morphological details of the cells and their axonal collateral network. The neurophysiological recordings provide functional data, and their close correlation with morphological ones makes them particularly valuable. Spectacular results have been obtained in visual cortex, barrel cortex, hippocampus, substantia nigra and striatum [18–21]. Intracellular recording and dye-filling is a method applicable only to a limited number of neurons per experiment. The yield is low, typically only few cells in each brain slice or sacrificed test animal.

3.3 Neurochemical Fingerprinting of BDA-Labeled Fibers

We developed in our laboratory an optical screening technique, and have coined it “neurochemical fingerprinting.” This technique consists of anterograde neuroanatomical tracing enhanced by an immunofluorescence screening procedure. Its principle is illustrated in Fig. 4. A neuroanatomical tracer (BDA, biotinylated dextran amine [22] or PHA-L, a leucoagglutinin isolated from the kidney bean Phaseolus vulgaris [23]), is focally injected into a particular brain area, taken up by neurons and transported via the axons all the way to their terminal boutons. Note that the uptake of the tracer is indiscriminate with respect to the neurochemical identity of the cells in the injection spot. The essence of neurochemical fingerprinting is that brain sections are subjected to a multiple immunofluorescence procedure that starts with visualizing the transported tracer via a reaction with a fluorochromated streptavidin. The streptavidin procedure is combined with incubation of the sections with screening cocktails of primary antibodies against neuroactive substances suspected of being present in the labeled neurons. This approach has originally been introduced as an electron microscopy technique (e.g., [24, 25]), and with the improved resolution offered by advanced confocal laser scanning microscopy the procedure can now be successfully conducted by fluorescence microscopy [26].

132

Floris G. Wouterlood

Fig. 4 The concept of neurochemical fingerprinting. (Top) Identification of proteins in nucleus B (nc “B”) after anterograde tracer injection in nucleus A (nc “A”); see text for details. (Bottom) Successful fingerprinting: (A, B) Optical sections in YZ and XY planes of a raw image stack (series). Green, vesicular glutamate transporter 1 (fluorochrome excited by 488 nm). Red, BDA (fluorochrome excited by 543 nm). Orange (arrows), colocalization of the two fluorescence signals. (C) 3D computer reconstruction of the boxed area shown in B. Note the aggregate of green voxels located inside a varicosity on the fibers labeled in red (arrow). The contents of the varicosity suggests that this is a glutamatergic ending

After the incubation with the primary antibodies we continue with a matching cocktail of fluorochromated secondary antibodies. The sections are then inspected under a confocal laser scanning microscope equipped with one imaging channel for the tracer, and as many additional imaging channels as possible, ideally matching the number of additional reporter fluorochromes included in the screening. Each channel is configured for a specific fluorochrome: laser excitation wavelength, filter block/cube (dichroic mirror/ beamsplitter and barrier filter) and detector. The section is scanned in all these channels at very high resolution to visualize all fluorochromes in the terminal boutons of the fibers. Scanning should be conducted in “sequential” mode to radically exclude cross talk [27]. Z-scanning is necessary for extracting 3D information. As a rule of thumb an axon terminal must be at least 10 pixels across in the images to enable meaningful analysis [26]. The aim of the

Selective Fluorochrome Expression Tracing

133

whole task is to determine at the resolution level of individual axon terminals the extent of “colocalization” of the tracer with a particular reporter fluorochrome included in the fingerprinting. Merging of the corresponding frames of the Z-stacks from different channels provides an indication whether the labeled fiber contains a signal specific for one of the reporter fluorochromes [26, 28]. 3D reconstruction or computer analysis of the dataset to detect significant colocalization is often helpful to decide whether fluorescence associated with a reporter molecule is located inside a BDA labeled axon terminal or just outside. An example is shown in Fig. 4A–C. Neurochemical fingerprinting has the advantage that it makes it possible to determine the neurochemical identity of the labeled fibers. However, the procedure is tedious and time consuming since very high resolution is required (see Subheading 7.4 for details), Because neurochemical fingerprinting is basically a screening procedure it produces many negative results. One has to be lucky to have presented to the tracer-labeled fibers the right cocktail of specific primary antibodies. When lucky, the method may produce spectacular results (e.g., [28]). However, neurochemical fingerprinting is not restricted only to anterograde tracing; it is also applicable to neurons selectively expressing fluorescent proteins. In Subheading 9 of the present chapter the neurochemical fingerprinting procedure is briefly discussed and illustrated, to verify the specificity of eYFP expression in transgenic cre mice (details in Ref. [29]).

4

Virus-Based Tracing Twenty-five years ago Ugolini and coworkers introduced a novel technique for tracing fiber connectivity, based on the idea to use viruses [30]. The viral particles act as the probe. The beauty of virus tracing lies in the fact that the probe is self-amplifying and crosses synapses, hereby spreading the infection into synaptically related neurons, that is, neuronal networks. The process is time-dependent and in proportion to the number of synapses crossed. It is important to note that interneurons do not escape attention with this method as they get infected as well. The chief disadvantage of viruses is that they are in principle nasty pathogenic agents causing local encephalitis, and may even kill cells. They replicate inside infected neurons and when replication runs out of control they produce an avalanche of infected neurons that dooms the experiment. Caution must be applied to animals and experimenters alike because of the biohazardous nature of some viruses. As the CNS appears to contain mostly networking neurons forming contacts with each other at various levels of organization, native virus tracing may encounter difficulty in surmounting these barriers. Much effort has been put into getting these viruses in check for the purpose of good and clean tracing [31, 32].

134

Floris G. Wouterlood

In the original concept of virus application the virus particles themselves acted as the probe signalling the tracts and pathways along which specific brain areas are connected to each other. Antibodies generated against the virus coat proteins were necessary to detect the spread of the virus and to reveal the morphological details of the infected cells. Transport and multiplication of the virus particles is necessary in this concept to raise the probe concentration to a detectable level.

5

Fluorescent Protein as a Reporter Molecule and the Virus as a Vehicle Molecular biology and its optogenetics branch have delivered many new approaches suitable to study neuronal networks. Optogenetic techniques are nowadays commonly used to study cellular dynamics via optically delivered stimuli. Molecular biology has for this purpose produced genetically modified organisms whose cells metabolize (synthesize) a photoactivatable protein that, when illuminated, triggers cellular activity [33, 34]. As photoactivatable proteins do not always fluoresce the group of fluorescent proteins (FPs) is more interesting for neuroanatomy. Anyway, optogenetics can be credited with delivering as a by-product the central principle of a new generation of tracing techniques based on the very fact that neurons expressing a gene coding for a particular neuroactive substance are supplied with an extra gene coding for a fluorescent protein (FP). As the current method is the result of an evolution of ideas and molecular biological experiments, we will now briefly survey its historical timeline. First to appear were mice which only expressed the green fluorescent protein (GFP) in specific groups of CNS neurons (e.g., thy1 gene expressing cells [35]). Then came the neurotransmitter-specific GFP mice [36], FP-dopaminergic neurons [37], FP-GAD67 neurons [38], 5-HT-cre mice [39], DA-cre knockin mice [40], various types of FP neurons [41], including cholinergic [42] and parvalbumin [43] FP neurons, dopamine transporter (DAT)-GFP-mice [44], 5-hydroxytryptamine-GFP (5-HT-GFP) mice [45], and adenosine receptor (A2AR-Cre) mice [46]. Once these mice were available the fiber connectivity of the FP expressing neurons could be traced. An “illuminating” example of results obtained with such an approach is the mapping study in developing and adult Thy1-eYFP-H expressing transgenic mice [47]. Next came the development of mice possessing neurons expressing fluorescent proteins of various spectral variants [48– 50] (as surveyed in [51]), thus paving the way for mapping studies similar as in the Thy1-eYFP-H mice. One should be aware that in these transgenic animals all cells with the corresponding neurochemical identity express GFP or one of its spectral variants. The same happens in the multispectral brainbow mice [1]. The

Selective Fluorochrome Expression Tracing

135

consequence of this property is that all FP expressing cells light up green simultaneously upon fluorescence excitation (illumination). While for the study of the general population of any neurochemically specific neurons the FP-expressing mice represent an enormous step forward, the very property that all cells with a particular genetic (i.e., neurochemical) makeup turn green becomes a disadvantage when it comes to tracing. To study neuronal networks it is essential to selectively label individual neurons or small groups of them. Take for instance the case of parvalbumin (PV). This calcium-binding protein is interesting here as it represents a marker in a subpopulation of distinct GABAergic neurons [52], such as cortical basket and chandelier cells [53, 54]. In the cerebellum PV is expressed by Purkyneˇ (Purkinje) cells and interneurons [52]. In the septum-diagonal band area of the forebrain, PV neurons intermingle with cholinergic neurons; cells of both types project from the septum to the hippocampus which, in turn, contains its own (local) populations of PV interneurons. What if it were possible to manipulate a transgenic mouse in such a way that only a small, discrete portion of neurochemically specific neurons, for example, small groups of exclusively PV neurons or cholinergic neurons could be instructed to start manufacturing a FP species? Braz et al. [55] developed a gene construct to drive expression of wheat germ agglutinin in neurons containing cre (tyrosine recombinase) gene in their genome, using a lox site flanking a stop codon. Kuhlman and Huang [56] adapted this solution to enable expression of a fluorescent protein in chemically specific neurons. They experimented with cre-recombinase knock-in PV (transgenic) mice wherein neurons were transfected via a focal injection of an adenoassociated virus (AAV) harboring a loxP-STOP-loxP cassette in between the CMV (cytomegalovirus) promoter and the DNA sequence coding for fluorescent protein expression, for example, eYFP (the “cre-lox” approach illustrated in Fig. 5). After the AAV injection, PV neurons in the injection site promptly started to express, and accumulate eYFP [56]. AAVs are employed in molecular biology for gene delivery to neurons [34, 58]. This approach was later perfected by Sohal and coworkers [57] who used double-floxed inverted open-reading frames to achieve the same objective, that is, a complete cre dependence of the expression of the floxed gene. After the injection and transfection the plasmid is locally selectively expressed, specifically in neurons expressing the recombinase gene (Figs. 6 and 7). The cre breakthrough experiments were originally designed with an optogenetic purpose in mind [57, 59] and thus it took some time before the neuroanatomical community fully recognised the potential of the double-floxed AAV approach.

136

Floris G. Wouterlood

Fig. 5 Principle of virus-mediated selective fluorescence protein expression in PV-cre mice. (A) Under normal conditions, the two proteins, parvalbumin (PV) and recombinase (cre) are coexpressed. PVpr and IRES are promoters. (B) Upon transfection, an insert of a double-floxed, eYFP coding gene is introduced into the genome. The recombinase reinserts the eYFP gene in the correct transcription direction only in PV-containing cells as only those coexpress cre. (C) Upon recombination the eYFP gene is expressed and eYFP starts to accumulate in the cell. Modified after [57]

6

Principle of Selective Fluorochrome Expression Selective fluorochrome expression tracing requires two fundamental conditions to be fulfilled. The first one is the availability of animal strains containing a cre-recombinase sequence inserted into part of their genome, as close as possible to the gene coding for the desired neuroactive substance. In this chapter we illustrate the method mainly with parvalbumin-cre (PV-cre) transgenic mice, and to a lesser degree with choline acetyltransferase-cre (ChAT-cre) transgenic mice (see Subheading 7 for details). The second condition is the availability of an adeno-associated virus (AAV) that carries the double-floxed eYFP plasmid construct. Also here a commercial supplier is active, named Adgene.2 Both conditions together are known as cre-lox. The essence of the cre-lox approach is that only neurons that express the PV gene also express the recombinase gene. Only the recombinase protein is able to recombine the lox sites and reverse the eYFP sequence of the plasmid incorporated in the host cell’s genome. The result of the recombinase activity is that only PV positive, cre expressing neurons in the injection site start producing the fluorescent protein. None of the other cells express the recombinase gene (Fig. 6 shows what categories of expression are encountered in practice). eYFP transfected neurons expressing another neuroactive substance than PV (“nonPV expressing cells”) are thus in principle incapable of expressing

2

www.addgene.org

Selective Fluorochrome Expression Tracing

137

Fig. 6 Venn diagram explaining the cre-lox approach in parvalbumin-cre (PV-cre) mice carrying the cre (“cause recombination”) gene for recombinase. (A) The entire population of PV expressing cells exhibiting a wide variety in the amount of PV synthesized. (B) Ideally, all transfected PV neurons should start expressing eYFP, so that artifacts are not encountered. However, the expression of eYFP is not equally strong in all transfected neurons. (C, D) Cells designated as optically “silent” (in terms of eYFP fluorescence) in the injection area are for instance cholinergic, glutamatergic or peptidergic neurons, that is, neurons that possess but do not express their PV gene. Suspicious is category “c,” that is, the subpopulation of PV neurons located in the injection site and therefore presumably transfected. These cells nevertheless do not express eYFP (false negative). (E) Transfected, explicitly non-PV neurons may start expressing eYFP (false positive)

eYFP, that is, they remain “optically silent” (category “d” in Fig. 6) in spite of being transfected, as discussed in detail in Subheading 8. Kuhlman and Huang [56] effectively demonstrated their approach by injecting various variants of their cre-dependent AAV into the cerebral cortex of PV-cre mice and another strain of transgenic mice, the glutamatergic-cre (Emx1-cre knock-in) mice

138

Floris G. Wouterlood

Fig. 7 Principle of tracing based on selective expression of fluorescent protein. (Top) A simple model of neuronal network consisting of three different chemical phenotypes (different grayscale levels). (Middle) Focally injected virus with eYFP gene transfects all neurons. Only cells with identity 1 express recombinase, and only in these cells the polarity of the eYFP-coding DNA is reversed and the inserted eYFP code translated. (Bottom) In cells with identity 1, eYFP starts to be expressed, eventually filling the entire neuron, including its axon and terminals

developed by Guo et al. [60]. The induced fluorescence in successfully transfected neurons appeared to be so strong that even fine details on the processes of transfected neurons became visible, including the terminal boutons of basket neurons in PV-cre mice,

Selective Fluorochrome Expression Tracing

139

or dendritic spines on cortical pyramidal glutamatergic neurons [56] in the Emx1-cre mice. These morphological features are thus visualized in a manner similar to dark-field microscopy, as they shine very bright against dark background.

7

Tracing via Selective Fluorochrome Expression We have conducted several tracing experiments by following the cre-lox approach pioneered by Kuhlman and Huang [56], and the technology refined by Sohal et al. [57]. Cre knockin recombinant mice [49] can be obtained from The Jackson Laboratory.3 We worked with strain B6;129P2-Pvalbtm1(cre)Arbr/J (PV-cre mice), and in a parallel series of experiments with strain B6;129S6Chattm1(cre)Lowl/J (ChAT-cre mice). The PV-cre and ChAT-cre mice make an interesting testing set as it is known that in rodents part of the fiber input into the hippocampus originates from populations of GABAergic [61], PV-containing and cholinergic neurons in the medial septumdiagonal band area [62, 63] (as reviewed in [64]). The GABAergic and cholinergic cells in the septum/diagonal band are a typical example of a mixed neuronal population (Fig. 8A). Anterograde (nonselective) neuroanatomical tracing through injections of PHA-L or BDA into various spots in the septum/ diagonal band thus inevitably results in labeling of both GABAergic and cholinergic fibers projecting into the hippocampus. Detailed topographic investigation of chemically specific fiber endings in the hippocampus would be dependent on time-consuming neurochemical fingerprinting of the fibers labeled with a neuroanatomical tracer. Selective fluorochrome expression, however, eliminates the need for neurochemical fingerprinting altogether, and for that very reason should be considered a sophisticated neuroanatomical tracing tool, complementary to the traditional ones. As such, it enables visual “dissection” of a single neurochemical species.

7.1

3

Virus Injection

https://www.jax.org/

Injection anesthesia remains an option but owing to its flexibility gas anesthesia is nowadays much more common: initially 2% isoflurane, after stabilization 1.5% isoflurane. The virus is injected via a stereotaxic approach. An excellent description of the surgical procedure in mice is provided in Ref. [65]. In brief, the periosteum of the cranium is anesthetized with lidocaine and an opening in the skull drilled at the desired anteroposterior and mediolateral coordinates to lower the micropipette with the virus vector to its target area (coordinates according to the mouse brain atlas [65]). Using a standard mechanical injection equipment (steel injection cannula,

140

Floris G. Wouterlood

Fig. 8 Triple-immunofluorescence staining in PV-cre mouse #2, using antibodies against GFP, choline acetyltransferase (ChAT) and parvalbumin (PV). Low-power confocal imaging in the vertical limb of the diagonal band (VDB); sequential scanning mode, Z-projection (XY plane) images. The boxed areas are shown in Fig. 9 at higher resolution. (A) “Green” channel with eYFP expressing (eYFP+) neurons. (B) Cholinergic (ChAT+) neurons visible in the “red” channel. (C) PV expressing (PV+) neurons visible in the “infrared” channel. (D) False-colored merged image (eYFP green, ChAT blue and PV red). The eYFP+ neurons are also PV+ while all ChAT+ cells appear to be eYFP
and PV
(expressing neither eYFP nor PV). The intensity of immunostaining varies a lot among individual cells (a-d): (a) strong eYFP and strong PV signal; (b) strong eYFP and moderate PV signal; (c) strong eYFP and weak PV signal. (d) Strong eYFP but no PV signal (“leaking” artifact)

0.2 mm internal diameter) we injected 0.5 μl virus solution (serotypes AAV2 and AAV5; EF1α-DIO-EFYP-WPRE-pA [note the EF1α promoter], double floxed; a gift from prof. Karl Deisseroth, Department of Bioengineering, Stanford University) at a rate of 100 nl/min into various loci in the septum-diagonal area. Instead of a steel cannula a glass micropipette can be used [57, 66] and iontophoresis instead of mechanical injection [66]. After delivery of the virus the injection cannula was left in situ for 7 min and slowly

Selective Fluorochrome Expression Tracing

141

retracted stepwise. The animals were allowed to recover and returned to their housing unit. During the postinjection period transfected PV-expressing cells begin to make eYFP and “distribute” it through their processes. It takes several weeks before sufficient eYFP has been accumulated to render all the cellular processes detectable under fluorescence illumination. Harris et al. [66] recommend a minimum of 2 weeks. In our mice, wherein eYFP expression is enhanced by the presence of the EF1α promoter, we adhered to a 5-week postinjection survival period to be certain that even the smallest projection fibers in long projection systems are well labelled with eYFP [67] (e.g., the septohippocampal pathway and the projection from the basal forebrain into the olfactory bulb). Sohal et al. [57] reported a 15 days postinjection survival period for PV-cre mice whereas Dautan et al. [68] reported 2–4 weeks in ChAT-cre rats. Note that with classical neuroanatomical tracing the survival period is designed to allow adequate transport of the tracer. Depending on the length of the fiber projections, sufficient transport of, for example biotinylated dextran amine (BDA [69]) requires a survival period of 7–14 days postsurgery. The AAV experiments and classical neuroanatomical tracing are thus comparably time-consuming, as demonstrated in cre-rats [68] and in cre-mice [70]. To the best of author’s knowledge the transport mechanism governing the spreading of eYFP through the cellular processes and into the axons has not been elucidated yet. 7.2

Sacrifice

After the appropriate survival (eYFP accumulation) period, the mice in our experiments were deeply anesthetized with 2% isoflurane followed by 1.6 g/kg (of body weight) urethane. Their thorax was opened and via a hollow needle inserted in the left cardiac ventricle a small volume of phosphate-buffered saline (PBS), pH 7.6 was perfused to flush out erythrocytes, followed by approximately 250 ml of PBS, pH 7.6, in which 4% paraformaldehyde had been dissolved. This fixative can be supplemented with a trace amount of glutaraldehyde [68]. Brains dissected from the skull can be sectioned according to standard sectioning procedures after a short postsacrifice postfixation in fresh fixative. Postfixation times range between 4 [40] and 24 h (present study). For instance, we have sectioned cre-mouse brains on a vibrating microtome and on a freezing microtome. We usually collect sections in buffer for further free-floating processing, while others collect sections on microscope slides for further (on-slide) incubation. From this point on, different histological procedures can follow. Our favorable one includes 40–50 μm thick frontal sections in a free-floating incubation regime. At this stage we inspect sections cut at the anteroposterior level of the AAV injection spot, in order to evaluate in a fluorescence microscope under blue illumination the extent of virus infection and eYFP

142

Floris G. Wouterlood

expression. Sections can be stored in buffer in a refrigerator, with some sodium azide added to prevent growth of microorganisms, for at least half a year before serious signal deterioration (eYFP fluorescence) sets in. However, during a prolonged storage in a refrigerator the immunohistochemical properties of the tissue usually progressively deteriorate. A good alternative procedure is to cryoprotect the brains in 30% sucrose or in buffered 20% glycerin/ 2% dimethylsulfoxide [71], then to section the brain on a freezing microtome and store the sections in the cryoprotectant in a freezer at
20  C or colder until further processing. 7.3 Immunohistochemistry: eYFP Stabilization and Additional Markers

One of the attractive properties of selective fluorochrome expression is that a plethora of immunohistochemical processing techniques can be applied to the sections. Native eYFP expression can be so strong that it does not require immunohistochemical stabilization per se as storage in an aqueous environment is sufficient. However, we prefer immunofluorescence stabilization of eYFP via incubation with an anti-GFP antibody (recognizing eYFP as well), for the sake of creating collections of permanent slides. A very attractive option is to “counterstain” with a tinctorial fluorescent stain or to conduct a multiple immunofluorescence procedure to detect additional markers. Last but not least, immunohistochemistry can be conducted for control purposes. Controls are necessary at the start of any project involving transgenic cre mice or rats to determine the “grade” of the transgenic animal strain, and to answer for example a fundamental question as to whether the PV-cre animals do indeed express eYFP in PV neurons and whether all eYFP-expressing neurons coexpress PV. We conduct for this purpose a multi-immunofluorescence staining including (a) antibody against GFP, (b) antibody against the neurochemical marker associated with the particular cre mouse (PV in this case), and (c) additional antibody of choice. Because in our model the population of interest consists of a mixture of PV neurons and cholinergic neurons, the third antibody in our experiments was directed against choline acetyltransferase (ChAT), a reliable marker of cholinergic neurons. Here follows the triple immunofluorescence protocol, tailor-made for subsequent threechannel imaging with confocal laser scanning microscope. Results are shown in Figs. 8 and 9. Panel D in Fig. 8 is such a control; it shows that eYFP expression does not occur in ChAT-neurons. At the same time, panels E and F of Fig. 8 show that PV+, eYFP
neurons occur among PV+, eYFP+ neurons (false negative).

7.3.1 Triple Immunofluorescence (GFP-PV-ChAT) Staining Protocol

Incubation vehicle: Tris buffered saline (TBS), pH 7.6 (room temperature value), containing 0.2% Triton X-100 (TBS-TX); all rinsing and incubation was conducted with TBS-TX. N.B.: Triton X-100 is a trademark of Sigma-Aldrich (St. Louis, MO, USA).

Selective Fluorochrome Expression Tracing

143

Fig. 9 High-resolution confocal images of PV-cre mouse #2 neurons shown in Fig. 8; sequential scanning mode, Z-projection (XY plane) images. (A–C) represent individual imaging channels while (D–F) are merged images. (A) Neurons with a various degree of eYFP expression (eYFP+) visible in the “green” channel (a–e). (B) Cholinergic (ChAT+) neurons visible in the “red” channel. (C) Parvalbumin-containing (PV+) neurons (a–c, e, f) visible in the “infrared” channel: (c) very weak PV signal; (d) no signal at all (PV
). (D) False-colored merged image of the eYFP (green) and ChAT (magenta) channels; there is no colocalization. (E) False-colored merged image of the eYFP (green) and PV (magenta) channels. The colocalization is either strong (a, b), weak (c, e) or none (d, f). (F) “Classical” false-colored merged image with eYFP (green), ChAT (blue) and PV (red). (a, b) Strong colocalization of eYFP and PV (cells appear yellow). (c) Strong eYFP signal colocalizing with weak PV signal. (d) Strong eYFP, no PV signal (“leaking” artifact, possibly a mutation). (e) Weak eYFP, strong PV signal. (f) No eYFP (eYFP
) but detectable PV signal (false negative artifact). The conclusion is that most but not all eYFP+ neurons are also PV+ and vice versa (strong colocalization). All eYFP+ and/or PV+ cells (a–f) are ChAT
(not expressing ChAT) while at the same time all ChAT+ cells (unmarked) are eYFP
and PV
(i.e., no colocalization)

144

Floris G. Wouterlood

Brain sections were recovered from the
20  C storage freezer, thawed, rinsed and incubated. In between the incubation steps we always rinsed for 3 10 min. All rinsing and incubation was carried out under continuous gentle agitation of the sections on a rocking plateau, at room temperature. l

Block the sections for 1 h in a blocking solution (5% normal donkey serum [NDS] in TBS-TX).

l

Incubate overnight in a cocktail of three primary antibodies:

l

1:1000 anti-green fluorescent protein (GFP) (rabbit; MerckMillipore, cat. № AB3080) mixed with 1000 anti-parvalbumin (PV) (mouse; Swant, cat. № PV-28) and 1:250 anti-choline acetyltransferase (ChAT) (goat; Chemicon, cat. № AB144P), with 5% normal goat serum (NGS) added.

l

Incubate for 2 h in a cocktail of secondary antibodies (all from Invitrogen-Molecular Probes [Eugene, OR, USA]): 1:400 donkey anti-rabbit-Alexa Fluor™ 488 mixed with 1:400 donkey anti-mouse-Alexa Fluor™ 633 and 1:400 donkey anti-goatAlexa Fluor™ 543. Note 1. There is a very good reason to select the fluorochromeconjugated secondary antibody excitable by the 488 nm laser, Alexa Fluor 488 (absorption/emission maximum at 495/519 nm) to detect the anti-GFP primary antibody bound to it. The fluorochromated secondary antibody associated with the marker (eYFP) should preferably fluoresce in the same wavelength range as the marker itself so that the two spectrally coincide. This helps to avoid or at least reduce fluorescence cross talk phenomena [26, 27] mentioned below (Note 2). Spectrally, eYFP has an absorption maximum at 514 nm (with a shoulder at 488 nm) and an emission maximum at 527 nm [72].

l

Note 2. Selection of a 633 nm fluorochrome-conjugated secondary antibody for the detection of bound anti-PV was also governed by the spectral properties of eYFP. As we need to confirm that the PV and eYFP fluorescent immunosignals do indeed colocalize, any false positivity must be excluded. False positivity may occur here, for instance because of the so-called excitation cross talk [26, 27]. A good countermeasure is to select for PV a fluorochrome with an absorption peak as far away as possible from that of eYFP, and/or to use sequential scanning (see Subheading 7.4). After rinsing in TBS-TX, do a final rinse in Tris–HCl, pH 7.6 and mount on glass slides from Tris–HCl, pH 7.6, with 0.1% gelatin added (Oxoid).

Selective Fluorochrome Expression Tracing

145

l

Thoroughly dry the slides, dip for 10 s into laboratory grade toluene (Merck) and coverslip using Entellan (Merck) as mounting medium.

l

Alternative mounting: sections can be mounted from Tris–HCl, pH 7.6 on slides and immediately wet-coverslipped in an antifading mounting medium: 2.4 g Mowiol 4-44 (Clariant, Muttenz, Switzerland)4 and 6.0 g glycerin (Fluka) in 18 ml 100 mM Tris, pH 8.5.

The above model procedure stains for eYFP, PV and ChAT in a single section. Alternative immunohistochemical staining may of course include any neurochemical substance-specific antibody used jointly with the anti-GFP one. In our lab, we have successfully used the following combinations: GFP-ChAT, GFP-NeuN, GFP-PV, GFP-calretinin, and GFP-somatostatin. 7.4 Confocal Laser Scanning Microscopy

Owing to its better resolution (compared to a standard fluorescence microscope) and superior fluorescence signal separation capability the confocal laser scanning microscope (CLSM) is our preferred instrument to analyze immunofluorescence staining. At our institution we have access to a Leica TCS-SP2 AOBS model (Leica Microsystems, Heidelberg, Germany) equipped with appropriate filter cubes for visual fluorescence inspection and with Ar/Kr and HeNe lasers for scanning in several channels (details in refs. [26, 27]): – “Green” (488 nm excitation, 500–540 nm emission bandpass filtering) – “Red” 1 (543 nm excitation, 550–585 nm emission bandpass filtering) or – “Red” 2 (594 nm excitation, 605–630 nm emission bandpass filtering) – “Infrared” (633 nm excitation, 650 nm emission longpass filtering) All multichannel scanning was always performed using the scanner’s “sequential” mode, that is, with the detection channels activated sequentially to avoid cross talk phenomena. Low magnification scanning was performed with a 20 dry objective lens to delineate the location of eYFP expressing, transfected neurons relative to the non-eYFP expressing (e.g., nontransfected; see details in Subheading 8), PV and ChAT immunopositive neuronal subpopulations in the septal/diagonal band area. Note that high resolution (63 glycerin immersion objective lens, NA 1.3) is necessary to determine whether all eYFP-expressing neurons do indeed

4

A similar compound (Mowiol 4-88) is available from Sigma (cat. #81381).

146

Floris G. Wouterlood

coexpress PV or ChAT (in PV-cre or ChAT-cre mice, respectively), and vice versa. Details of the neurons and their processes were usually scanned while employing the 63 immersion lens and additional 8 electronic zoom. Scanning was mostly performed as Z-scanning, with frame dimensions 1024  1,024 pixels at 8-bit sampling and Z-increment 122 nm. Images to be used for illustration purposes were Z-scanned with the 20 dry lens at 2048  2048 pixels or with the 63 immersion lens at 1024  1024 pixels. Images of fibers and fiber terminals acquired using the 63 immersion objective were deconvoluted with Huygens Professional™ software (Scientific Volume Imaging, Hilversum, The Netherlands.5 For further computer processing and 3D reconstruction we used ImageJ [73] and Amira™ (FEI Visualization Sciences Group software).6 Figures for the present chapter were prepared with Adobe Photoshop™ software (version 6.0) and Xara™ software.

8

Results of Control Incubations Neurons and structures expressing (nonexpressing) eYFP are further referred to as eYFP+ (eYFP
), parvalbumin-immunopositive (negative) neurons as PV+ (PV
), and ChAT-immunopositive (negative) neurons as ChAT+ (ChAT
). The injection sites are represented by small groups of highly eYFP+ neurons at injection-targeted locations in the diagonal band. An example is shown in Fig. 8. As briefly mentioned in Subheadings 7.3 and 7.4, one of the first tasks in these experiments is to verify that in the PV-cre mice only parvalbumin-positive (PV+) neurons are expressing eYFP, and vice versa. Indeed, almost all eYFP+ neurons in the injection sites were also PV-positive. However, there is a great deal of variability in the amount of both eYFP and PV expressed in neurons (Fig. 6). The mix of intensely, moderately and weakly fluorescing neurons is well visible in the merged images (Figs. 8D and 9D–F). When one of the markers is expressed much more strongly than the coexpressed one the color balance in the merged images is dominated by the strongly “expressed” color. The merged images must therefore be evaluated with great care. Making a distinction between weak and no PV immunosignal poses a challenge, given the enormous sensitivity and contrast enhancement capabilities of confocal imaging instruments. An example is shown in Fig. 8 (cell “c”). The eYFP fluorescence of this cell is very strong indeed (panel A). The cell shows absolutely no ChAT-

5 6

https://svi.nl http://www.vsg3d.com/

Selective Fluorochrome Expression Tracing

147

immunosignal (panel B), but in the PV-associated channel a very weak immunosignal is present (panel C). The problem is that in the merged image (panel D), the PV signal is very strongly dominated by eYFP fluorescence. After the initial scanning with a low-power, dry 10 objective we switch to the 63 immersion objective and Z-scan the boxed area again, sequentially with the three lasers (each corresponding to one marker), deconvolute the images and prepare a set of merged images. At this point, a plethora of eYFP and PV fluorescence signal combinations emerges (Figs. 8 and 9): – Strong eYFP signal combined with strong PV signal (cells a and b), – Strong eYFP signal combined with weak PV signal (cell c), – Strong eYFP signal and no PV signal (cell d), – Weak eYFP signal combined with strong PV signal (cell e), and finally – No eYFP signal combined with strong PV signal (cell f). The fact that some cells (e.g., “f”) lack eYFP signal may have any of the following causes: – These are PV neurons that have not been transfected. – These are PV neurons that have been transfected but for some reason eYFP recombination has failed, or eYFP expression has been halted, or eYFP expression has not been sufficient to reach immunodetection level. One explanation may be that neurons must be transfected by a certain number of virus particles in order to start producing and accumulating sufficient eYFP. Note here that the vehicle virus (in contrast to “wild” viruses) does not multiply once inside a neuron. The ChAT immunopositive cells in the vertical diagonal band (VDB), scattered among the PV neurons, showed no trace of either eYFP (Fig. 9D) or PV (image not shown). eYFP+/PV
neurons (such as “d” in Figs. 8 and 9) were infrequently encountered in the septum-diagonal band. We also noted such cells in the olfactory tubercle adjacent to the VDB. In experiments with PV-cre mice of the same strain treated with virus injection in cerebral cortex, we noted spurious eYFP+/PV
pyramidal cells located at some distance from the injection spot. In PV cre-mice we thus suspect “leaking” (possibly a mutation), that is, recombinase activity yielding eYFP expression in transfected but otherwise non-PV neurons. Leaking can be a major source of error in this type of neuroanatomical tracing. These issues are further discussed in Subheading 11. The sites targeted by virus injections in the present study: the vertical (VDB) and horizontal (HDB) limbs of the diagonal band of Broca were full of eYFP expressing PV+ cell bodies (Fig. 8A, C, D).

148

Floris G. Wouterlood

Contingents of eYFP+ fibers could be easily traced via the precommissural fornix and fimbria fornicis into the hippocampus. These fibers displayed terminal distribution patterns in the hilus of the dentate gyrus and in all layers of hippocampal fields CA1 and CA3. Other eYFP+ fibers traveled from the diagonal band nuclei in anterior direction, close to and parallel with the brain surface, toward taenia tecta. Signs of retrograde transport of the virus were absent. Retrograde transport would have been revealed by the presence of eYFP expressing neurons elsewhere in the brain, notably in areas projecting to the diagonal band.

9

Neurochemical Fingerprinting in Axons of eYFP-Expressing PV Neurons As in the PV-cre mouse the fluorescent protein is expected to be expressed only in PV neurons, there would be (intuitively) no pressing need to screen fibers expressing eYFP also for the presence of PV. However, one of the conditions to be satisfied for the “standard” neurochemical fingerprinting (such as the one illustrated in Fig. 4B, C) to be successful is a strong expression of the “fingerprinted” marker in the fibers and axon terminals. Vesicular transporters, as the example in Fig. 4 shows, are typically strongly expressed in axon terminals and thus suitable for fingerprinting. The variability in PV expression encountered in the PV-cre mice, ranging from weak to strong expression, suggests that neurochemical fingerprinting with anti-PV antibodies may yield many false negatives. Figure 10 as an example of high-magnification neurochemical fingerprinting with anti-PV antibodies (red) and anti-GFP antibodies in the cerebral cortex of AAV-injected PV-cre mice illustrates that this artifact does indeed occur [29].

10

A Note on Track Labeling Track labeling is an artifact encountered in brain tissue along the injection cannula track, when cells with the expected molecular phenotype become virus-transfected and start expressing eYFP. Ideally, the AAV virus should be introduced only into the desired target site, that is, not along the injection track itself. Such cells are referred to as track cells. We noted this phenomenon especially in ChAT-cre mice where the caudate-putamen complex had to be dorsoventrally penetrated in order to deliver AAV into the cholinergic basal forebrain nuclei that are located “deep,” that is, at the ventral aspect of the brain. The striatum contains big ChAT+ interneurons described 30 years ago with ChATimmunohistochemistry alone [74, 75]. Dautan et al. [68] successfully transfected cholinergic interneurons in ChAT-cre rats, verified the transfection with ChAT immunofluorescence and noted that

Selective Fluorochrome Expression Tracing

149

Fig. 10 Neurochemical fingerprinting in PV-cre mice. Sections immunostained with anti-GFP/488 nm secondary (green) and anti-PV/594 nm secondary (red). (A) Visualization in XY, XZ, and YZ planes, merged raw confocal images. (B) Merged XY image, upon deconvolution. Some (arrows) but not all “green” eYFP-labeled fibers contain “red” PV immunofluorescence signal

the transfected ChAT+ neurons displayed the same morphological features as classically immunostained cholinergic interneurons (Fig. 5 in Ref. [68]). In all of our ChAT-cre mice, virus-injected in the nucleus basalis region, several sections of the caudate-putamen complex contained a dorsoventrally oriented column of eYFP+ neurons immediately dorsal to the injection site, thus locating the injection track. Control staining with antibodies against ChAT revealed that these eYFP + neurons were also ChAT+. Track labeling can affect the outcome in a false-positive way and thus should be suppressed as much as possible. The artifact may be significantly reduced by using smaller needles, a longer waiting period before retraction, and/or the use of glass needles with small tips.

11

Prospects Owing to its inherent neurochemical selectivity the fluorochrome expression presented here is indeed a very attractive tool for neuroanatomists. It may be even considered as a completely new generation of tracing method [29]. Classical neuroanatomical techniques are indiscriminate with respect to the neurochemical phenotype of the traced neurons (Fig. 2). With BDA tracing the chemical identification of the labeled fibers can be carried out only after the tracing procedure itself, via a screening referred to as “neurochemical fingerprinting” (Fig. 4), which is very time- and

150

Floris G. Wouterlood

resource-consuming. The enormous advantage that the cells in the transgenic animals forced via AAV transfection to express fluorescent protein (FP) belong to a particular neurochemical species can be fully exploited. The FP expressing neurons can be studied immediately, after only a brief initial neurochemical screening carried out to be certain that the transgenic animal strain used is indeed specific and that no neurons of another neurochemical phenotype inadvertently express recombinase and thus the FP. When neurons are organized in neurochemically heterogeneous nuclei of the type outlined in Fig. 2, and observed in the experiments presented here (Figs. 8 and 9), the selective fluorochrome expression is extremely helpful to sort out anatomical relationships between the different neurons and the relationships they maintain with the surrounding nuclei. Double selective fluorochrome expression may shed light on cellular interactions between neurochemically different neurons. A combination with BDA tracing adds an additional dimension. The compatibility of BDA tracing and selective fluorochrome expression has already been investigated in mice [70], as well as the compatibility of selective fluorochrome expression and retrograde tracing in rats [68]. One may try to inject a PV-cre mouse in one brain location, for instance the VDB, with AAV delivering a gene responsible for eYFP expression in septohippocampal PV+ neurons, while also applying a second injection of another AAV into the hippocampus to deliver a gene responsible for the expression of a “red” fluorescent protein (RFP) in intrahippocampal PV+ interneurons. Indeed, RFPs suitable for expression in mouse brain do exist (e.g., [76, 77]). The next step might be to induce via a third AAV injection an expression of an optogenetic protein, for example, channelrhodopsin, to enable modulation of neuronal activity or intracellular pathways by optogenetic procedures [78–80]. This would “cross-link” anatomy with electrophysiology, and perhaps even with behavioral research. Cellular networks may thus be subjected to truly physiological (functional) inquiries. A requirement for the selective fluorochrome expression to be successful is that the transgenetic mouse or rat strain is free from “aberrant” neurons, that is, cells not belonging to the desired neurochemical category, while expressing cre and therefore the transfected gene coding for FP. Such spurious nonspecific expression (“leaking”) may be due to a genetic accident (mutation) during transfection. Although the expression of floxed genes is strictly cre-dependent [81, 82] it is necessary in every single experiment to verify that cre expression is indeed regulated as intended [49]. This is crucial because of the myriad of new cre-lines that keep appearing in the realms of optogenetics and neuroscience. The suspicion in the present PV-cre mice of spurious eYFP positive, PV-negative neurons underscores the need to verify the expected

Selective Fluorochrome Expression Tracing

151

“specificity” of the cre-strains, and to develop reliable and stable, that is, error-free neuroactive marker cre-mice and cre-rat strains [83]. It is equally important to keep track of false negatives, that is, neurons of interest belonging to the neurochemically specific neuronal population but not being transfected. Finally, as we are dealing with standard immunohistochemistry with no special requirements for fixation, one can envisage countless combinations of selective fluorochrome expressions, with all the tools nowadays available to cell biologists, such as receptor and channel mapping, neurotransmitter and neuropeptide immunohistochemistry, calcium binding mapping, and vesicular transporter immunofluorescence.

Acknowledgments The author is grateful to Dr. Bernard Bloem (McGovern Institute for Brain Research, Massachusetts Institute of Technology, Cambridge, MA, USA) for contributing material with AAV2 virus injections in ChAT-cre mice, Dr. Antonio Luchicchi (Department of Integrative Neurophysiology, Center for Neurogenomics and Cognitive Research [CNCR], Neuroscience Campus Amsterdam, VU University Amsterdam, The Netherlands) for contributing material with AAV5 virus injections in PV-cre mice, Prof. Karl Deisseroth (Department of Bioengineering, Stanford University, Stanford, CA, USA) for donating the AAV2 virus serotype, and Ms. Yvonne Galis for her expert assistance with immunofluorescence. Dr. Radek Pelc (editor) is acknowledged for his suggestions that were basic to Fig. 6. References 1. Weissman TA, Sanes JR, Lichtman JW, Livet J (2011) Generating and imaging multicolor brainbow mice. Cold Spring Harb Protoc 2011(7):763–769. https://doi.org/10.1101/ pdb.top114 2. Chung K, Wallace J, Kim SY, Kalyanasundaram S, Andalman AS, Davidson TJ, Mirzabekov JJ, Zalocusky KA, Mattis J, Denisin AK, Pak S, Bernstein H, Ramakrishnan C, Grosenick L, Gradinaru V, Deisseroth K (2013) Structural and molecular interrogation of intact biological systems. Nature 497(7449):332–337. https://doi. org/10.1038/nature12107 3. Miyasaka N, Arganda-Carreras I, Wakisaka N, Masuda M, Su¨mbu¨l U, Seung HS, Yoshihara Y (2014) Olfactory projectome in the zebrafish forebrain revealed by genetic single-neuron

labelling. Nat Commun 5:3639. https://doi. org/10.1038/ncomms4639 4. Tomer R, Ye L, Hsueh B, Deisseroth K (2014) Advanced CLARITY for rapid and highresolution imaging of intact tissues. Nat Protoc 9:1682–1697. https://doi.org/10.1038/ nprot.2014.123 5. Kononenko N, Witter MP (2012) Presubiculum layer III conveys retrosplenial input to the medial entorhinal cortex. Hippocampus 22:881–895. https://doi.org/10.1002/hipo. 20949 6. Wouterlood FG, Goede PH, Arts MPM, Groenewegen HJ (1992) Simultaneous characterization of efferent and afferent connectivity, neuroactive substances and morphology of neurons. J Histochem Cytochem 40:457–465. https://doi.org/10.1177/40.4. 1552182

152

Floris G. Wouterlood

7. Maco B, Cantoni M, Holtmaat A, Kreshuk A, Hamprecht FA, Knott GW (2014) Semiautomated correlative 3D electron microscopy of in vivo-imaged axons and dendrites. Nat Protoc 9(6):1354–1366. https://doi.org/10. 1038/nprot.2014.101 8. Streit P (1980) Selective retrograde labeling indicating the transmitter of neuronal pathways. J Comp Neurol 191:429–463. https:// doi.org/10.1002/cne.901910308 9. Kiss J, Csa´ki A, Bokor H, Kocsis K, Kocsis B (2002) Possible glutamatergic/aspartatergic projections to the supramammillary nucleus and their origins in the rat studied by selective [3H]D-aspartate labelling and immunocytochemistry. Neuroscience 111:671–691. https://doi.org/10.1016/S0306-4522(02) 00037-4 10. Van der Kooy D, Steinbusch HWM (1980) Simultaneous fluorescent retrograde axonal tracing and immunofluorescent characterization of neurons. J Neurosci Res 5:479–484. https://doi.org/10.1002/jnr.490050603 11. Bigl V, Woolf NJ, Butcher LL (1982) Cholinergic projections from the basal forebrain to frontal, parietal, temporal, occipital, and cingulate cortices: a combined fluorescent tracer and acetylcholinesterase analysis. Brain Res Bull 8:727–749. https://doi.org/10.1016/03619230(82)90101-0 12. Skirboll L, Ho¨kfelt T, Norell G, Phillipson O, Kuypers HG, Bentivoglio M, CatsmanBerrevoets CE, Visser TJ, Steinbusch H, Verhofstad A, Cuello AC, Goldstein M, Brownstein M (1984) A method for specific transmitter identification of retrogradely labeled neurons: immunofluorescence combined with fluorescence tracing. Brain Res Rev 8:99–127. https://doi.org/10.1016/ 0165-0173(84)90001-8 13. Datiche F, Cattarelli M (1996) Catecholamine innervation of the piriform cortex: a tracing and immunohistochemical study in the rat. Brain Res 710:69–78. https://doi.org/10. 1016/0006-8993(95)01279-6 14. Zappone CA, Sloviter RS (2001) Commissurally projecting inhibitory interneurons of the rat hippocampal dentate gyrus: a colocalization study of neuronal markers and the retrograde tracer Fluoro-gold. J Comp Neurol 441:324–344. https://doi.org/10.1002/cne. 1415 15. Chandler DJ, Waterhouse BD (2012) Evidence for broad versus segregated projections from cholinergic and noradrenergic nuclei to functionally and anatomically discrete subregions of prefrontal cortex. Front Behav Neurosci 6:1–9. https://doi.org/10.3389/fnbeh.2012.00020

16. Chandler DJ, Lamperski CS, Waterhouse BD (2013) Identification and distribution of projections from monoaminergic and cholinergic nuclei to functionally differentiated subregions of prefrontal cortex. Brain Res 1522:38–58. https://doi.org/10.1016/j.brainres.2013.04. 057 17. Halberstadt AL, Balaban CD (2008) Selective anterograde tracing of nonserotonergic projections from dorsal raphe nucleus to the basal forebrain and extended amygdala. J Chem Neuroanat 35:317–325. https://doi.org/10. 1016/j.jchemneu.2008.02.006 18. Chang HT, Kitai ST (1985) Projection neurons of the nucleus accumbens: an intracellular labeling study. Brain Res 347:112–116. https://doi.org/10.1016/0006-8993(85) 90894-7 19. Sik A, Tamamaki N, Freund TF (1993) Complete axon arborization of a single CA3 pyramidal cell in the rat hippocampus, and its relationship with postsynaptic parvalbumincontaining interneurons. Eur J Neurosci 5:1719–1728. https://doi.org/10.1111/j. 1460-9568.1993.tb00239.x 20. Prensa L, Parent A (2001) The nigrostriatal pathway in the rat: a single-axon study of the relationship between dorsal and ventral tier nigral neurons and the striosome/matrix striatal compartments. J Neurosci 21:7247–7260. https://doi.org/10.1523/JNEUROSCI.2118-07247.2001 21. Klausberger T, Magill PJ, Ma´rton LF, Roberts JD, Cobden PM, Buzsa´ki G, Somogyi P (2003) Brain-state- and cell-type-specific firing of hippocampal interneurons in vivo. Nature 421(6925):844–848. https://doi.org/10. 1038/nature01374 22. Veenman CL, Reiner A, Honig MG (1992) Biotinylated dextran amine as an anterograde tracer for single- and double-label studies. J Neurosci Methods 41:239–254. https://doi. org/10.1016/0165-0270(92)90089-V 23. Gerfen CR, Sawchenko PE (1984) An anterograde neuroanatomical tracing method that shows the detailed morphology of neurons, their axons and terminals: immunohistochemical localization of an axonally transported plant lectin, Phaseolus vulgaris leucoagglutinin (PHA-L). Brain Res 290:219–238. https:// doi.org/10.1016/0006-8993(84)90940-5 24. Freund TF, Antal M (1988) GABA-containing neurons in the septum control inhibitory interneurons in the hippocampus. Nature 336 (6195):170–173. https://doi.org/10.1038/ 336170a0 25. Charara A, Smith Y, Parent A (1996) Glutamatergic inputs from the pedunculopontine

Selective Fluorochrome Expression Tracing nucleus to midbrain dopaminergic neurons in primates: Phaseolus vulgaris-leucoagglutinin anterograde labeling combined with postembedding glutamate and GABA immunohistochemistry. J Comp Neurol 364:254–266. https://pubmed.ncbi.nlm.nih.gov/8788248 26. Wouterlood FG, Belie¨n JAM (2014) Translation, touch and overlap in multi-fluorescence confocal laser scanning microscopy to quantitate synaptic connectivity. In: Bakota L, Brandt R (eds) Laser scanning microscopy and quantitative image analysis of neuronal tissue, Neuromethods, vol 87. Springer, New York, pp 1–36. https://doi.org/10.1007/978-1-4939-03818_1 27. Wouterlood FG (2006) Combined fluorescence methods to determine synapses in the light microscope: multilabel confocal laser scanning microscopy. In: Zabo´rszky L, Wouterlood FG, Lanciego JL (eds) Neuroanatomical tracttracing 3: molecules, neurons, and systems. Springer, New York, pp 394–435. https://doi. org/10.1007/0-387-28942-9_13 28. Boulland J-L, Jenstad M, Boekel A, Wouterlood FG, Edwards RH, Storm-Mathisen J, Chaudhry FA (2009) Vesicular glutamate and GABA transporters sort to distinct sets of vesicles at a symmetric synapse. Cereb Cortex 19:241–248. https://doi.org/10.1093/ cercor/bhn077 29. Wouterlood FG, Bloem B, Mansvelder HD, Luchicci A, Deisseroth K (2014) The fourth generation of neuroanatomical tracing techniques: exploiting the offspring of genetic engineering. J Neurosci Methods 235:331–348. https://doi.org/10.1016/j.jneumeth.2014. 07.021 30. Ugolini G, Kuypers HG, Strick PL (1989) Transneuronal transfer of herpes virus from peripheral nerves to cortex and brainstem. Science 243:89–91. https://doi.org/10.1126/ science.2536188 31. Ugolini G (2008) Use of rabies virus as a transneuronal tracer of neuronal connections: implications for the understanding of rabies pathogenesis. In: Dodet B, Fooks AR, Mu¨ller T, Tordo N (eds) Towards the elimination of rabies in Euroasia (joint OIE/WHO/ EU conference, Paris, 27–30 May 2007), ISBN:978-3805585507. Developments in Biologicals (Basel), 131:493–506. [a] https:// www.ncbi.nlm.nih.gov/pubmed/18634512, [b] https://www.karger.com/Book/Home/ 235783, [c] https://www.oie.int/fileadmin/ Home/eng/Conferences_Events/sites/ rabies2007/files/Rabies_presentations.htm 32. Ugolini G (2010) Advances in viral transneuronal tracing. J Neurosci Methods 194:2–20.

153

https://doi.org/10.1016/j.jneumeth.2009. 12.001 33. Barth AL (2007) Visualizing circuits and systems using transgenic reporters of neural activity. Curr Opin Neurobiol 17:567–571. https://doi.org/10.1016/j.conb.2007.10. 003 34. Luo L, Callaway EM, Svoboda K (2008) Genetic dissection of neural circuits. Neuron 57:634–660. https://doi.org/10.1016/j.neu ron.2008.01.002 35. Feng G, Mellor RH, Bernstein M, KellerPeck C, Nguyen QT, Wallace M, Nerbonne JM, Lichtman JW, Sanes JR (2000) Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP. Neuron 28:41–51. https://doi.org/10.1016/S08966273(00)00084-2 36. Aller MI, Jones A, Merlo D, Paterlini M, Meyer AH, Amtmann U, Brickley S, Jolin HE, McKenzie AN, Monyer H, Farrant M, Wisden W (2003) Cerebellar granule cell Cre recombinase expression. Genesis 36:97–103. https:// doi.org/10.1002/gene.10204 37. Zhao S, Maxwell S, Jimenez-Beristain A, Vives J, Kuehner E, Zhao J, O’Brien C, de Felipe C, Semina E, Li M (2004) Generation of embryonic stem cells and transgenic mice expressing green fluorescence protein in midbrain dopaminergic neurons. Eur J Neurosci 19:1133–1140. https://doi.org/10.1111/j. 1460-9568.2004.03206.x 38. Tamamaki N, Yanagawa Y, Tomioka R, Miyazaki J, Obata K, Kaneko T (2003) Green fluorescent protein expression and colocalization with calretinin, parvalbumin, and somatostatin in the GAD67-GFP knock-in mouse. J Comp Neurol 467:60–79. https://doi.org/ 10.1002/cne.10905 39. Zhuang X, Masson J, Gingrich JA, Rayport S, Hen R (2005) Targeted gene expression in dopamine and serotonin neurons of the mouse brain. J Neurosci Methods 43:27–32. https://doi.org/10.1016/j.jneumeth.2004. 09.020 40. B€ackman CM, Malik N, Zhang YJ, Shan L, Grinberg A, Hoffer BJ, Westphal H, Tomac AC (2006) Characterization of a mouse strain expressing Cre recombinase from the 30 untranslated region of the dopamine transporter locus. Genesis 44:383–390. https:// doi.org/10.1002/dvg.20228 41. Gong S, Doughty M, Harbaugh CR, Cummins A, Hatten ME, Heintz N, Gerfen CR (2007) Targeting Cre recombinase to specific neuron populations with bacterial artificial chromosome constructs. J Neurosci

154

Floris G. Wouterlood

27:9817–9823. https://doi.org/10.1523/ jneurosci.2707-07.2007 42. von Engelhardt J, Eliava M, Meyer AH, Rozov A, Monyer H (2007) Functional characterization of intrinsic cholinergic interneurons in the cortex. J Neurosci 27:5633–5642. https://doi.org/10.1523/jneurosci.4647-06. 2007 43. Tanahira C, Higo S, Watanabe K, Tomioka R, Ebihara S, Kaneko T, Tamamaki N (2009) Parvalbumin neurons in the forebrain as revealed by parvalbumin-Cre transgenic mice. Neurosci Res 63:213–223. https://doi.org/10.1016/j. neures.2008.12.007 44. Zhou W, Lee YM, Guy V, Freed CR (2009) Embryonic stem cells with GFP knocked into the dopamine transporter yield purified dopamine neurons in vitro and from knock-in mice. Stem Cells 27:2952–2961. https://doi.org/ 10.1002/stem.216 45. Vucurovic K, Gallopin T, Ferezou I, Rancillac A, Chameau P, van Hooft JA, Geoffroy H, Monyer H, Rossier J, Vitalis T (2010) Serotonin 3A receptor subtype as an early and protracted marker of cortical interneuron subpopulations. Cereb Cortex 20:2333–2347. https://doi.org/10.1093/ cercor/bhp310 46. Zhang JP, Xu Q, Yuan XS, Cherasse Y, Schiffmann SN, de Kerchove d’Exaerde A, Qu WM, Urade Y, Lazarus M, Huang ZL, Li RX (2013) Projections of nucleus accumbens adenosine A2A receptor neurons in the mouse brain and their implications in mediating sleep-wake regulation. Front Neuroanat 7:43–57. https:// doi.org/10.3389/fnana.2013.00043 ˜ o C, 47. Porrero C, Rubio-Garrido P, Avendan Clasca´ F (2010) Mapping of fluorescent protein-expressing neurons and axon pathways in adult and developing Thy1-eYFP-H transgenic mice. Brain Res 1345:59–72. https:// doi.org/10.1016/j.brainres.2010.05.061 48. Livet J, Weissman TA, Kang H, Draft RW, Lu J, Bennis RA, Sanes JR, Lichtman JW (2007) Transgenic strategies for combinatorial expression of fluorescent proteins in the nervous system. Nature 450:56–62. https://doi.org/10. 1038/nature06293 49. Madisen L, Zwingman TA, Sunkin SM, Oh SW, Zariwala HA, Gu H, Ng LL, Palmiter RD, Hawrylycz MJ, Jones AR, Lein ES, Zeng H (2010) A robust and high-throughput Cre reporting and characterization system for the whole mouse brain. Nat Neurosci 13:133–140. https://doi.org/10.1038/nn. 2467 50. Taniguchi H, He M, Wu P, Kim S, Paik R, Sugino K, Kvitsiani D, Fu Y, Lu J, Lin Y,

Miyoshi G, Shima Y, Fishell G, Nelson SB, Huang ZJ (2011) A resource of Cre driver lines for genetic targeting of GABAergic neurons in cerebral cortex. Neuron 71:995–1013. https://doi.org/10.1016/j.neuron.2011.07. 026 51. Asrican B, Augustine GJ, Berglund K, Chen S, Chow N, Deisseroth K, Feng G, Gloss B, Hira R, Hoffmann C, Kasai H, Katarya M, Kim J, Kudolo J, Lee LM, Lo SQ, Mancuso J, Matsuzaki M, Nakajima R, Qiu L, Tan G, Tang X, Ting JT, Tsuda S, Wen L, Zhang X, Zhao S (2013) Next-generation transgenic mice for optogenetic analysis of neural circuits. Front Neural Circuits 7:160–184. https://doi. org/10.3389/fncir.2013.00160 52. Celio MR (1986) Parvalbumin in most gamma-aminobutyric-acid containing interneurons in the rat cerebral cortex. Science 213:995–997. https://doi.org/10.1126/sci ence.3945815 53. Hendry SHC, Jones EG, Emson PC, Lawson DEM, Heizmann CW, Streit P (1989) Two classes of cortical GABA neurons defined by different calcium binding protein immunoreactivities. Exp Brain Res 76:467–472. https:// doi.org/10.1007/BF00247904 54. Fonseca M, Soriano E, Ferrer I, Martinez A, ˜ o´n T (1993) Chandelier cell axons identiTun fied by parvalbumin-immunoreactivity in the normal human temporal cortex and in Alzheimer’s disease. Neuroscience 55:1107–1116. https://doi.org/10.1016/0306-4522(93) 90324-9 55. Braz JM, Rico B, Basbaum AI (2002) Transneuronal tracing of diverse CNS circuits by Cre-mediated induction of wheat germ agglutinin in transgenic mice. Proc Natl Acad Sci U S A 99:15148–15153. https://doi.org/10. 1073/pnas.222546999 56. Kuhlman SJ, Huang ZJ (2008) Highresolution labeling and functional manipulation of specific neuron types in mouse brain by Cre-activated viral gene expression. PLoS One 3:e2005. https://doi.org/10.1371/jour nal.pone.0002005 57. Sohal VS, Zhang F, Yizhar O, Deisseroth K (2009) Parvalbumin neurons and gamma rhythms enhance cortical circuit performance. Nature 459:698–702. https://doi.org/10. 1038/nature07991 58. White MD, Milne RV, Nolan MF (2011) A molecular toolbox for rapid generation of viral vectors to up- or down-regulate neuronal gene expression in vivo. Front Mol Neurosci 4:8. https://doi.org/10.3389/fnmol.2011. 00008

Selective Fluorochrome Expression Tracing 59. Deisseroth K, Feng G, Majewska AK, Miesenbo¨ck G, Ting A, Schnitzer MJ (2006) Next-generation optical technologies for illuminating genetically targeted brain circuits. J Neurosci 26:10380–10386. https://doi. org/10.1523/JNEUROSCI.3863-06.2006 60. Guo H, Hong S, Jin XL, Chen RS, Avasthi PP, Tu YT, Ivanco TL, Li Y (2000) Specificity and efficiency of Cre-mediated recombination in Emx1-Cre knock-in mice. Biochem Biophys Res Commun 273:661–665. https://doi. org/10.1006/bbrc.2000.2870 61. Ko¨hler C, Chan-Palay V, Wu JY (1984) Septal neurons containing glutamic acid decarboxylase immunoreactivity project to the hippocampal region in the rat brain. Anat Embryol 169:41–44. https://doi.org/10.1007/ BF00300585 62. Amaral DG, Kurz J (1985) An analysis of the origins of the cholinergic and noncholinergic septal projections to the hippocampal formation of the rat. J Comp Neurol 240:37–59. https://doi.org/10.1002/cne.902400104 63. Kiss J, Patel AJ, Baimbridge KG, Freund TF (1990) Topographical localization of neurons containing parvalbumin and choline acetyltransferase in the medial septum-diagonal band region of the rat. Neuroscience 36:61–72. https://doi.org/10.1016/03064522(90)90351-4 64. Teles-Grilo Ruivo LM, Mellor JR (2013) Cholinergic modulation of hippocampal network function. Front Synaptic Neurosci 5:2. https://doi.org/10.3389/fnsyn.2013.00002 65. Paxinos G, Franklin KBJ (2001) The mouse brain in stereotaxic coordinates. Academic, New York 66. Harris JA, Oh SW, Zeng H (2012) Adenoassociated viral vectors for anterograde axonal tracing with fluorescent proteins in nontransgenic and cre driver mice. Curr Protoc Neurosci (Suppl) 59:Unit 1.20. https://doi.org/ 10.1002/0471142301.ns0120s59 67. Bloem B, Schoppink L, Rotaru DC, Faiz A, Hendriks P, Mansvelder HD, van de Berg W, Wouterlood FG (2014) Topographic mapping between basal forebrain cholinergic neurons and the medial prefrontal cortex in mice. J Neurosci 34:16234–16246. https://doi.org/ 10.1523/jneurosci.3011-14.2014 68. Dautan D, Huerta-Ocampo I, Witten IB, Deisseroth K, Bolam JP, Gerdjikov T, MenaSegovia J (2014) A major external source of cholinergic innervation of the striatum and nucleus accumbens originates in the brainstem. J Neurosci 34:4509–4518. https://doi.org/ 10.1523/jneurosciI.5071-13.2014

155

69. Reiner A, Honig MG (2006) Dextran amines: versatile tools for anterograde and retrograde studies of nervous system connectivity. In: Zaborszky L, Wouterlood FG, Lanciego JL (eds) Neuroanatomical tract-tracing 3: molecules, neurons, and systems. Springer, New York, pp 304–335. https://doi.org/10. 1007/0-387-28942-9_10 70. Wang Q, Henry AM, Harris JA, Oh SW, Joines KM, Nyhus J, Hirokawa KE, Dee N, Mortrud M, Parry S, Ouellette B, Caldejon S, Bernard A, Jones AR, Zeng H, Hohmann JG (2014) Systematic comparison of adenoassociated virus and biotinylated dextran amine reveals equivalent sensitivity between tracers and novel projection targets in the mouse brain. J Comp Neurol 522:1989–2012. https://doi.org/10.1002/ cne.23567 71. Rosene DL, Roy NJ, Davis BJ (1986) A cryoprotection method that facilitates cutting frozen sections of whole monkey brains for histological and histochemical processing without freezing artifact. J Histochem Cytochem 34:1301–1315. https://doi.org/10.1177/ 34.10.3745909 72. Wachter RM, Elsliger M-A, Kallio K, Hanson GT, Remington SJ (1998) Structural basis of spectral shifts in the yellow-emission variants of green fluorescent protein. Struct Fold Des 6:1267–1277. https://doi.org/10.1016/ S0969-2126(98)00127-0 73. Rasband WS (1997–2004) ImageJ. National Institutes of Health, Bethesda (MD, USA). http://rsb.info.nih.gov/ij/ 74. Armstrong DM, Saper CB, Levey AI, Wainer BH, Terry RD (1983) Distribution of cholinergic neurons in rat brain: demonstrated by the immunocytochemical localization of choline acetyltransferase. J Comp Neurol 216:53–68. https://doi.org/10.1002/cne.902160106 75. Phelps PE, Houser CR, Vaughn JE (1985) Immunocytochemical localization of choline acetyltransferase within the rat neostriatum: a correlated light and electron microscopic study of cholinergic neurons and synapses. J Comp Neurol 238:286–307. https://doi.org/10. 1002/cne.902380305 76. Runyan CA, Sur M (2013) Response selectivity is correlated to dendritic structure in parvalbumin-expressing inhibitory neurons in visual cortex. J Neurosci 33:11724–11733. https://doi.org/10.1523/jneurosci.2196-12. 2013 77. Adamah-Biassi EB, Zhang Y, Jung H, Vissapragada S, Miller RJ, Dubocovich M (2014) Distribution of MT1 melatonin receptor promoter-driven RFP expression in the

156

Floris G. Wouterlood

brains of BAC C3H/HeN transgenic mice. J Histochem Cytochem 62:70–84. https://doi. org/10.1369/0022155413507453 78. Zhang F, Gradinaru V, Adamantidis AR, Durand R, Airan RD, De Lecea L, Deisseroth K (2010) Optogenetic interrogation of neural circuits: technology for probing mammalian brain structures. Nat Protoc 5:439–456. https://doi.org/10.1038/nprot.2009.226 79. Fenno L, Yizhar O, Deisseroth K (2011) The development and application of optogenetics. Annu Rev Neurosci 34:389–412. https://doi. org/10.1146/annurev-neuro-061010113817 80. Miesenbo¨ck G (2011) Optogenetic control of cells and circuits. Annu Rev Cell Dev Biol 27:731–758. https://doi.org/10.1146/ annurev-cellbio-100109-104051

81. Sauer B (1998) Inducible gene targeting in mice using the cre/lox system. Methods 14:381–392. https://doi.org/10.1006/ meth.1998.0593 82. Sauer B (2002) Cre/lox: one more step in the taming of the genome. Endocrine 19:221–228. https://doi.org/10.1385/ ENDO:19:3:221 83. Witten IB, Steinberg EE, Lee SY, Davidson TJ, Zalocusky KA, Brodsky M, Yizhar O, Cho SL, Gong S, Ramakrishnan C, Stuber GD, Tye KM, Janak PH, Deisseroth K (2011) Recombinase-driver rat lines: tools, techniques, and optogenetic application to dopamine-mediated reinforcement. Neuron 72:721–733. https://doi.org/10.1016/j.neu ron.2011.10.028

Chapter 5 Optical Imaging Probes for Amyloid Diseases in Brain Pratyush Kumar Mishra, Myeong-Gyun Kang, and Hyun-Woo Rhee Abstract As amyloid fibril accumulation currently represents the key pathological evidence for a number of neurodegenerative disorders, a range of optical imaging probes have been developed in the past to visualize these infectious proteins. To some extent, they make it possible to characterize neuronal diseases in a postmortem state, that is, on sections of patients’ brains. However, many of the probes are of limited use for in vivo brain imaging, due to the lack of blood–brain barrier permeability and absorption/scattering of weak fluorescence signals deep in the neural tissue. Recently developed probes with significantly improved imaging properties partly amend such problems. These include near-infrared, two-photon, and phosphorescent imaging probes, as well as those selective for certain amyloid types, and for detecting enzymatic activity linked to the amyloid disease progression. Obviously, this has important medical implications, as a disease can be more reliably diagnosed and its progress detected at an early stage. Key words Alzheimer’s disease, Amyloid, Optical imaging probes, Near-infrared fluorescence, Phosphorescence, Two-photon microscopy

Abbreviations AD APP Aβ BBB CR HOMO LUMO PET PiB ThS/ThT

Alzheimer’s disease Amyloid precursor protein Amyloid beta Blood–brain barrier Congo red Highest occupied molecular orbital Lowest unoccupied molecular orbital Positron emission tomography Pittsburgh compound B Thioflavin S/T

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_5, © Springer Science+Business Media LLC 2020

157

158

1

Pratyush Kumar Mishra et al.

Introduction Alzheimer’s disease (AD) is the most common cause of dementia, significantly impairing the quality of life. AD has become a major cause of deaths in USA, and estimated 11–16 million people are expected to be affected by 2050 [1]. It is characterized by aggregated proteins on neuronal cells, referred to as amyloid fibril accumulation or amyloid plaques, originally mistaken for granules of starch (amylum in Latin). Formation of the amyloid (or amyloidlike) plaques in the brain is also a hallmark of more than 20 other neurodegenerative diseases, including Parkinson’s, Huntington’s, and Creutzfeldt
Jakob diseases [2, 3]. Amyloid fibril accumulations are caused by different neuronal protein or peptide aggregations in different neurodegenerative disease models. For example, in AD, the amyloid-beta (Aβ) peptide, a 36–43 amino acid peptide derived from the amyloid precursor protein (APP), forms protein aggregates referred to as senile plaques [4]. In Parkinson’s disease, alpha-synuclein, a presynaptic vesicle protein, aggregates to form insoluble fibrils known as Lewy bodies [5]. In Creutzfeldt-Jakob disease, prion protein (PrP), itself a normal cellular membrane protein, turns into an infectious form (PrPSc) indigestible by cellular proteases; it has more β-sheets essential in the formation of amyloid fibrils and plaques [6]. It is a well-known fact that the fibrillar proteins implicated in the neurodegenerative diseases have abundant β-sheet secondary structures and form continuous β-sheet fibers and accumulations (Fig. 1), and eventually visible plaques. Although it is still ambiguous whether the amyloid plaques are the cause or a mere effect of the neurodegenerative disorders [7], amyloid fibril accumulation of certain neuroproteins has become an obvious pathological hallmark of such medical conditions. In a test tube assay, higher-order Aβ oligomers were found to function as templates recruiting additional monomeric Aβ and toxic Aβ oligomers; this so-called “amyloid cascade hypothesis” describes the progression of AD in patients’ brains [7]. Such cascade kinetics is also encountered in other neurodegenerative disease models, such as Parkinson’s disease. Given that the amyloid plaques are currently the key pathological evidence of the neurodegenerative disorders, a range of imaging probes have been developed to visualize these infectious proteins. Such probes proved useful in correctly characterizing the neuronal diseases in the postmortem, histological sections of patients’ brains, and some of them also in diagnosing the neurodegenerative diseases at an early stage directly in vivo in experimental animals. They are also crucial in the quest to develop, by fluorescence-based high-throughput screening, specific inhibitors of Aβ peptide formation, thus paving the way to an appropriate

Optical Imaging Probes for Amyloids

159

Fig. 1 A schematic model of amyloid-β (Aβ1–42) peptide aggregation mechanism, consistent with Raman spectra recorded during that process. Reproduced from Ref. [48] by permission of © Wiley

treatment. Furthermore, with recent advances in photodynamic therapy [8], imaging probes can also be employed as therapeutic agents [9]. The last few years have thus seen the development of significantly improved imaging probes, as surveyed in the present chapter. To put their functionality in context, traditional imaging probes are briefly described first. This review is, of course, far from exhaustive, and we refer readers to other papers of this kind [10– 14].

2 2.1

Traditional Imaging Probes Congo Red

Congo Red (CR) is a diazo dye first synthesized in 1883 by Paul Bottiger, and was originally intended as a cellulose (textile) staining dye [15]. It has two sulfonate groups rendering an anionic charge. As a diazo dye, it is not very strongly fluorescent (excitation/ emission ~500/~600 nm) but can be used as a strong red stain in bright-field (i.e., nonfluorescence) imaging. In biochemistry and medicine, CR is the standardized probe to label amyloid plaques, usually visualized owing to birefringence [16], that is, in polarization microscopy [3, 17], and it is still used in post mortem analyses of AD patients’ brains. CR’s binding to amyloid-β peptide aggregates is very strong and specific, which is due to a combination of electrostatic interactions between CR’s anionic sulfonate groups and cationic amino acid residues in the antiparallel protein β sheets, and hydrophobic interactions between CR’s aromatic ring and the hydrophobic surface of amyloid fibrils. However, CR is a weak fluorophore, thus its sensitivity is very low compared to other fluorescent dyes. Based on CR’s shape, several fluorescent dyes (e.g., X-34) have been developed [18]. As shown in Fig. 2, these dyes showed bright blue fluorescence in transgenic Caenorhabditis elegans [19]. In addition, CR is not brain–blood barrier (BBB)permeable, this being a crucial factor for detecting the amyloid proteins in living mammalian brain.

160

Pratyush Kumar Mishra et al.

A

B

50 µm

Congo Red

C

D

X-34

50 µm

Fig. 2 Structure of Congo Red (CR) and X-34, its stronger-fluorescence derivative. (A, B) Isolated CR-stained amyloid-like protein aggregates formed by proteolytic cleavage of glutathione S-transferase (GST)-huntingtin fusion protein (GST-HD51) implicated in Huntington’s disease (HD). (A) Bright-field image. (B) Polarizationmicroscopy image (240  240 μm) indicating ordered molecular structure. (C, D) Transgenic Caenorhabditis elegans model of Alzheimer’s disease. (C) DIC Nomarski image. (D) Epi-fluorescence image (X-34 staining in vivo). Arrows show X-34 reactive deposits, numerous in constitutive amyloid-β expression animal (CL2006) but not in temperature-induced expression animal (CL4176). Reproduced from Refs. [3, 18, 19] by permission of © Cell press, © SAGE publication and © Elsevier, respectively 2.2 Thioflavin T and Pittsburgh Compound B

Thioflavin T (ThT) is a small benzothiazole dye frequently used in the analysis of aggregated amyloid proteins [20]. ThT’s binding affinity is slightly weaker than CR’s. ThT emits blue fluorescent light (λmax ¼ 474 nm) which increases several-fold upon binding to the amyloid protein. As ThT binding to the amyloid protein is very specific, its structure has inspired a rational design of other probes, such as Pittsburgh compound B (PiB), a radioactive analogue of ThT [21]. PiB is a BBB-permeable positron emission tomography (PET) probe actively used in clinical trials, to image the amyloid plaques directly in vivo. Thioflavin S (ThS) as an analogue of ThT is derived from it; methyl group of ThS is conjugated to another benzothiazole ring with sulfonic acid (Fig. 3). It binds to amyloid fibrils and increases their fluorescence emission, however, it does not generate

Optical Imaging Probes for Amyloids

Thioflavin T

Pittsburgh compound B

161

Thioflavin S

Fig. 3 Chemical structure of Thioflavin T, Pittsburgh compound B (PiB) and Thioflavin S. The radioactive carbon (11C) in PiB enables positron emission tomography (PET) imaging

wavelength shift in excitation or emission spectra. This characteristic of Thioflavin S results in high background signal. Like PiB, many PET dyes are currently being developed to image the amyloid proteins. As PET technology is very sensitive, it can be used even when the dye is localized deep in the body, including the brain. However, the PET probes have obvious limitations, such as very short lifetime (~20 min), and the inability to discriminate between the bound and unbound state, so that additional steps are required to wash the unbound probe from the brain tissue. In addition, trained and experienced staff are required in order to obtain meaningful imaging data using the PET probes. To overcome these limitations, other probes have been developed as outlined below. Their novel properties include far-red/ near-IR fluorescence enabling multiphoton excitation, long lifetime, higher selectivity in sensing relevant hallmark proteins of different neurodegenerative disorders, or the ability to detect enzymatic activity associated with AD progression. These features improve the selectivity and sensitivity of in vivo imaging of small toxic plaques in living brain tissue.

3

Long-Wavelength and Near-Infrared Fluorescent Imaging Probes For in vivo optical imaging, probes should ideally emit light at long wavelengths (600–700 nm), that is, close to the near-IR (NIR) region, offering the advantage of reduced scattering of light which can thus penetrate considerably deeper through the living brain tissue (up to several centimeters in the NIR region [22]) and generates weaker autofluorescence. NIR excitation is required in multiphoton microscopy (see next section). Sometimes, these probes are referred to (somewhat confusingly) as “NIR probes” as their spectra extend to the NIR region (700–3000 nm) although “far-red” or “long-wavelength” probes would usually be more appropriate terms when working with visible (typically red) light. Although often intended primarily for whole-animal (i.e., mesoscopic) imaging [23] these probes can, of course, be employed in microscopic imaging as well [24].

162

Pratyush Kumar Mishra et al.

Furthermore, when considering brain imaging in living organisms, probes should be able to pass the BBB as they are introduced into the blood stream by intravenous injection. They should be sufficiently small (up to ~600 Da), relatively lipophilic and should not bear ionic groups [24], as required by the so-called H-bonding rules [25]. Some sources specify ~400 Da as the size limit [25]. 3.1

2C40

To achieve long-wavelength fluorescence, fluorophores should have a very small energy gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) [24]. Commercially available dyes such as cyanine (Cy5, Cy7) are well known fluorophores of this kind. Inspired by these organic dyes’ conjugated structure, Chang’s group developed probes capable of detecting amyloids in the brain [25]. Over 320 styryl compounds were obtained by a diversity-oriented synthesis on solid phase, and exposed them to amyloid proteins, either in or out of the amyloid fibril state. Based on this screening, several useful compounds have been identified; they exhibited very bright fluorescence and a spectral shift upon binding to the fibril-state amyloid proteins. One of these dyes, 2C40, features yellow fluorescence (~565 nm) which shifts closer to the NIR region upon binding to insulin amyloid fibrils (~590 nm), and increases 20and 27-fold in the presence of the amyloid-β peptide (Aβ42 and Aβ40, respectively). Using this dye, amyloid plaques could be selectively imaged in mouse brain slices by fluorescence microscopy (Fig. 4). From colocalization staining experiments with thioflavin S, 2C40 was found to preferentially label the core of the amyloid deposits. However, because of the probe’s strong positive charge, it cannot be used in vivo as it lacks BBB permeability.

3.2

NIAD-4

In 2005, Swager’s group reported a highly BBB-permeable far-red fluorescent probe, NIAD-4, suitable for imaging Aβ aggregates [24]. It is derived from strong Aβ fibril binders such as CR, ThT, and the BBB-permeable probe, PiB. To ensure BBB permeability and binding affinity to the Aβ fibrils, a dicyanomethylene group with a slightly anionic BBB-permeable functional group selectively coordinating with the slightly cationic surface of the Aβ fibrils is attached. A highly polarizable thiophene linker is employed to narrow down the HOMO-LUMO gap and achieve NIR fluorescence (λmax  600 nm). Indeed, NIAD-4 shows high affinity to the Aβ aggregates (Ki ¼ 10 nM) and its NIR fluorescence and high BBB-permeability make it a suitable probe for in vivo imaging of Aβ plaques in the brains of transgenic AD mice (Fig. 5).

3.3

DANIR

In 2014, Saji’s group reported yet another new NIR fluorescent imaging probe, DANIR, suitable for the Aβ plaques [23]. Interestingly, its chemical structure is completely different from the previously known Aβ plaque dyes. DANIR employs a very simple

Optical Imaging Probes for Amyloids

163

2C40

A

B

Fig. 4 Chemical structure of 2C40, and staining of a brain slice from Alzheimer’s disease mouse. (A) Labeling of amyloid deposits (arrows) with 2C40. (B) Confocal microscopy images (in pseudocolors) of amyloid deposits labeled with 2C40 (red) and Thioflavin S (ThS, blue). Magnification/scale bar not specified in the original source. Reproduced from Ref. [25] by permission of © Wiley

chemical π-conjugated donor (N,N0 -dimethylamino benzene) to acceptor (dicyanomethylene group) architecture, yet exhibits a fairly strong NIR fluorescence at 665 nm and good binding affinity (Kd ¼ 27 nM) for the Aβ1–42 aggregates. Para-positioning of the N, N0 -dimethylamino group on the benzene ring results in a good electron-pushing group but also acts as a well-known Aβ binder. Furthermore, a small molecular weight (249.31 Da) and optimal lipophilicity (log P ¼ 3.37; P is partition coefficient) lend this probe good BBB permeability and fast clearance rates, essential for in vivo imaging of Aβ aggregates in living brain, as demonstrated in transgenic AD mice (Fig. 6). 3.4 Curcumin-Based Probes

It is a well-known phenomenon that curcumin (yellow) reacts with boric acid to form rosocyanine (red). A probe suitable for nearinfrared imaging, CRANAD-2 (abbreviation derived from author’s name) was prepared from curcumin by incorporating a difluoroboronate and N,N-dimethyl group substitution at para position of the phenyl ring. Its λmax  760 nm in methanol, and displays a large Stokes shift in phosphate-buffered saline (λmax(ex)  640 vs. λmax  805). Upon binding to Aβ (1–40) aggregates (em) (Kd ¼ 38.69 nM), a 70-fold fluorescence enhancement was encountered [26] (Fig. 7).

164

Pratyush Kumar Mishra et al.

NIAD–4

B

A

1 mm

30 µm

Fig. 5 Chemical structure of NIAD-4, and staining of amyloid deposits (plaques) in an aged transgenic (APP) Alzheimer’s disease mouse overexpressing amyloid precursor protein. (A) Coronal section of brain labeled with a 10 μM solution of NIAD-4 in DMSO/propylene glycol. (B) In vivo fluorescence detection of both cerebrovascular amyloid angiopathy (a blood vessel is shown) and senile plaques by multiphoton microscopy, immediately following an intravenous injection of 2 mg/kg of NIAD-4. Reproduced from Ref. [24] by permission of © Wiley

Recently, a radioactively (18F)-labeled variant of CRANAD2 has been reported as “probe 2.” It is suitable for both optical and PET (radio) imaging. As it lacks the N,N-dimethyl groups at one end it exhibits smaller Stokes shift; λmax  650. This new probe makes it possible to visualize sectioned brain samples (Fig. 7). Its disadvantage is reduced brain permeability, possibly because it is rapidly metabolized. Its pharmacokinetics thus requires further investigation [27].

4

Two-Photon and Phosphorescent Imaging Probes Two-photon fluorescence microscopy represents yet another attractive tool for in vivo imaging of Aβ plaques. While it offers resolution down to the optical diffraction limit of ~200 nm it does not enable whole-animal imaging (cf. Subheading 3 above). It also employs NIR illumination (two low-energy photons instead of one high-energy photon as in conventional, single-photon confocal microscopy), and imaging is possible to a considerably greater

Optical Imaging Probes for Amyloids

Thioflavin S

DANIR

A

B

C

D

Cy5 filter

165

GFP filter

Fig. 6 Chemical structure of DANIR and ex vivo fluorescence images of brain slices. (A) Double-transgenic (APPswe/PSEN1) Alzheimer’s disease mouse overexpressing human amyloid precursor protein and presenilin 1. (C) Wild-type control mouse, after injection of DANIR. (B, D) Amyloid-β (Aβ) plaques visualized by colocalization staining of the same sections with thioflavin S. Filter sets as indicated in the image plate, magnification/scale bar not specified in the original source. Reproduced from Ref. [23] by permission of © American Chemical Society

depth than with single-photon excitation operating in a visible light range [28], up to 1 mm in living brain tissue [29]. The highly localized nature of the two-photon excitation process results in less photobleaching and fairly low photodamage, thus allowing for longer real-time observations, with low autofluorescence background. Probes with sufficiently high two-photon action crosssection are required [30]. 4.1

SAD1

Recently, Kim’s group reported a successful development of a two-photon NIR fluorescent probe, SAD1, for Aβ plaques, that has been used for brain imaging directly in vivo, at a depth of 400 μm [29]. SAD1 has a conjugated structure of benzothiazole

166

Pratyush Kumar Mishra et al.

CRANAD-2

Probe 2

A

B

C

D

Fig. 7 Chemical structures of CRANAD-2 and Probe 2, and fluorescence images of brain slices. (A, B) Doubletransgenic mouse (Tg APP/PS-1) with abundant Aβ aggregates stained with the probes. (C, D) Age-matched wild-type mouse (no or very few aggregates), staining with Aβ specific antibody. Scale bars, 100 μm. Reproduced from Ref. [27] (Creative Commons licence)

ring and N-methyl aminonaphthalene. The benzothiazole ring, also present in PiB, lends the probe increased binding affinity for Aβ1–42 aggregates (Kd ¼ 17 nM). The N-methyl-aminonaphthalene group makes the probe suitable for two-photon excitation (two-photon action cross section 170 Goeppert–Mayer at 750 nm excitation). Furthermore, this dye’s small molecular weight and moderate lipophilicity translates to good BBB permeability, thus enabling in vivo Aβ plaque imaging in brain tissue of transgenic AD mice (Fig. 8). 4.2

[Ru(bpy)2dppz]2+

In 2011, Martı´’s group reported a new kind of amyloid probe based on ruthenium complex dyes, [Ru(bpy)2dppz]2+ where

Optical Imaging Probes for Amyloids

167

OH

N S N H

SAD1

40 μm 50 μm 200 μm Fig. 8 Chemical structure of SAD1 and 3D reconstructed two-photon excitation microscopy images of frontal cortex of transgenic (5XFAD) Alzheimer’s disease mouse overexpressing human amyloid precursor protein and presenilin-1, after intraperitoneal injection of 10 mg/kg SAD1. Color-inverted image reproduced from Ref. [29] by permission of © Royal Society of Chemistry

“bpy” ¼ 2,20 -bipyridine and “dppz” ¼ dipyrido[3,2-a:20 ,30 -c]-phenazine) [31]. Ruthenium metal complexes have been studied in a wide variety of applications, including DNA detection [32], cell viability tests [33] and cell imaging, but have rarely been utilized for peptide sensing. Although [Ru(bpy)2dppz]2+ shares almost no structural similarity with traditional amyloid fibril binders such as CR or ThT, it specifically binds to the fibrillar states of Aβ1–42 and its photoluminescence increases up to 50-fold upon addition of Aβ1–42 aggregates. In aqueous solution, free [Ru (bpy)2dppz]2+ does not emit any photoluminescence due to the population of a low-lying dark state. However, when it interacts with amyloid fibrils, microenvironmental changes lead to the population of a luminescent state. In further computational studies, binding modes between this dye and amyloid fibrils were investigated [34]. The hydrophobic cleft between Val18 and Phe20 of Aβ1–42 was proposed as a plausible binding site for [Ru (bpy)2dppz]2+, which could lead to photoluminescence during the “switch-on” state. The reason why ruthenium dyes are prominent in this field is because [Ru(bpy)2dppz]2+ has exceptional properties, such as long luminescence lifetime, which cannot be achieved by organic dyes. In organic dyes, the lifetime of fluorescence emission resulting from a singlet excitation state is usually between picoseconds (10
12 s)

168

Pratyush Kumar Mishra et al.

A

B

C

N N

N

N

N

N

Ru2+

N

N

[Ru(bpy)2dppz]2+

D

E N N

N

N

N

N

Ru2+

N

N

[Ru(phen)2dppz]2+ Fig. 9 Dyes based on ruthenium complexes. (A, D) Chemical structure of [Ru(bpy)2dppz]2+ and [Ru (phen)2dppz]2+ (phosphorescent dyes). (B, C) Their emission spectra in the presence of Aβ monomers before incubation (blue curves), and Aβ fibrils after 500 min incubation in a fluorescent (rhodamine B-containing) medium (red curves). (B) Steady-state. (C) Time-resolved (350–700 ns gating after light excitation). (E) Fluorescence detection in human neuroglioma H4/α-syn-GFP cells of alpha synuclein (αS) aggregation implicated in Parkinson’s disease and induced here by MG132, a proteasome inhibitor (2 μM, 16 h). Magnification/scale bar not specified in the original source. Reproduced from Refs. [31, 37] by permission of © American Chemical Society

and nanoseconds (10
9 s). However, ruthenium dyes, as well as many other metal-complexed luminescent dyes, have longer lifetimes, ranging from microseconds (10
6 s) to milliseconds (10
3 s), following triplet excitation. We call these luminescent events phosphorescence [35]. Unlike nonphosphorescent dyes such as ThT, [Ru (bpy)2dppz]2+ can detect Aβ fibrillization even in the presence of another organic dye such as rhodamine B, by employing a timegating approach. In time-resolved emission spectra (obtained with the aid of a 350–700 ns gating window), luminescence intensity could be selectively detected from the ruthenium dyes, even after the background organic dye’s fluorescence has almost completely decayed [31] (Fig. 9). Although this phenomenon is of little use to monitor Aβ fibrillization per se (rhodamine B is not needed for that) it would be of

Optical Imaging Probes for Amyloids

169

great advantage for screening Aβ fibril growth inhibitors. These often have conjugated planar structures with binding affinity for hydrophobic β-sheet fibril surfaces, and this coplanar structure brings about some level of fluorescence emission. For example, potential drugs such as curcumin and quercetin exhibit significant fluorescence interference during Aβ fibrillization assay when used jointly with ThT, making such assays unreliable [36]. The long lifetime of [Ru(bpy)2dppz]2+ makes it possible to use the abovementioned time-gating approach to “dissect” in time the phosphorescence intensity of fibrillized Aβ peptides even in the presence of strong fluorescence background from the tested drugs. Likewise, monitoring of the Aβ formation directly in situ (i.e., in a cellular environment) is likely to require simultaneous staining with organic dyes, to monitor cellular events linked to Aβ formation. Such correlative approach may possibly shed more light on the amyloid cascade hypothesis mentioned above [7]. Ruthenium dyes could also be used in fluorescence-lifetime imaging microscopy (FLIM) capable to discriminate background cellular, endogenous autofluorescence on the basis of phosphorescence lifetime. 4.3 [Ru(phen)2dppz]2+

5

In 2012, Martı´’s group reported another ruthenium dye, [Ru (phen)2dppz]2+ (“phen” ¼ 1,10-phenanthroline, “ddpz” ¼ as above) offering favorable light-switching properties and even stronger photoluminescence intensity than [Ru(bpy)2dppz]2+. [Ru(phen)2dppz]2+ was successfully employed in real-time, in vitro monitoring of alpha-synuclein (αS) fibrillization implicated in Parkinson’s disease [37]. Furthermore, the [Ru(phen)2dppz]2+ complexes were also employed to visualize chemically induced αS aggregation directly in situ, in human neuroglioma cells overexpressing green-fluorescent-protein-fused alpha-synuclein (α-synGFP) and treated with MG-132, a proteasome inhibitor capable of inducing aggregation of misfolding-prone proteins (Fig. 9). It should be noted that of the two currently available ruthenium dyes described above, only [Ru(phen)2dppz]2+ emits red fluorescence (640 nm) suitable for deep-tissue imaging. Both of the dyes have a strong positive charge, which is not suitable for in vivo imaging requiring more neutral molecules, as explained above.

Selective Imaging Probes for Different Types of Amyloid Fibrils As mentioned in Introduction, many kinds of neurodegenerative disorders are accompanied by the presence of amyloid fibrils, or formation of other neuronal protein aggregates. As traditional amyloid staining probes exhibit unspecific binding affinity toward

170

Pratyush Kumar Mishra et al.

any amyloid β-sheet structure [38], it is rather difficult to discriminate between the various types of amyloid plaques (deposits), and to correctly identify the type of neurodegenerative disease if using only the traditional imaging probes. For example, the current amyloid-staining probe PiB has proved very useful as a diagnostic tool for the amyloid plaques but its PET signal is of no use in discriminating between the amyloid-associated disorders it accompanies, and the same applies to Congo Red (Fig. 2). From this viewpoint, fluorescence imaging probes have strong advantages over the PET probes as the former can provide more specific information, depending on the kind of amyloid fibrils they have affinity for. 5.1

ANCA

In 2012, Yang’s group was successful in discriminating between the amyloid-β plaques and prion (PrPSc) deposits in mouse brain, using three newly developed optical probes, namely, amino-naphthalenyl 2-cyanoacrylate (ANCA)-based probes [39]. ANCA probes are fluorophores with molecular rotors generating green fluorescence emission through a photon-induced dipole mechanism. An electron-rich amino group attached to the naphthalene moiety pushes the electron density to the electron-withdrawing cyano ester group through the π-conjugated system of the naphthalene ring. In its unbound state, the ANCA probes exhibit weak fluorescence because their photon-excited energy is easily dissipated through molecular motion, such as rotations around single bonds between donor (amino-) and acceptor (cyano ester) groups. However, when the probes are bound to a protein, these rotations are restricted and energy of the excited state is released through fluorescence emission. Overall, the ANCA probes’ fluorescence considerably increases upon binding to amyloid fibrils or plaques and, interestingly, its emission wavelength maximum is different upon binding to Aβ plaques and prion plaques (λmax ¼ 535 and 554 nm, respectively for the ANCA-1 probe). These findings are explicable by different properties of the binding pockets of Aβ and prion fibrils, each representing a microenvironment with a somewhat different polarity (prions being more polar than Aβ). Polarity of each protein’s binding pockets can be expressed as a dielectric constant value. The dielectric constant of Aβ and prion plaques is roughly similar to that of diethyl ether (ε0 ¼ 4.27) and tetrahydrofuran (ε0 ¼ 7.52), respectively, implying the binding pocket of Aβ has a more polar environment than that of prion plaques. Using this probe, the authors were able to spectrally discriminate between Aβ deposits in the hippocampus of a transgenic AD mouse and PrPSc deposits in the corpus callosum of a prion-infected mouse (Fig. 10).

Optical Imaging Probes for Amyloids

ANCA-1

171

Congo Red

D

A

F

C 100 µm

100 µm

E

B

Fig. 10 Chemical structure of ANCA-1 probe employed here to stain unfixed cryo-sections (true-color imaging). (A) Aβ deposits in the hippocampus of a transgenic Alzheimer’s disease mouse. (B) Deposits of an infectious form (PrPSc) of prion protein in the corpus callosum of an experimentally infected mouse. (C) Ex vivo fluorescence spectra of stained deposits showing distinct emission maxima for Aβ and PrPSc. (D, E) Deposits of the same origin as in A and B, upon Congo red staining. (F) Fluorescence emission spectra showing no separation between Aβ and PrPSc. NFI, normalized fluorescence intensity. Reproduced from Ref. [39] by permission of © American Chemical Society 5.2

BODIPY-Zn-DPA

Tau is an abundant microtubule-associated protein found in neurons. Protein-protein interactions between tau and microtubules are controlled by the phosphorylation status of serine/threonine residues of the tau protein. However, under pathological conditions, tau is unusually hyperphosphorylated (at more than 30 sites) and does not bind to microtubules; instead, it accumulates in brain tissue in insoluble aggregates referred to as neurofibrillary tangles (NFTs). Jointly with the senile plaques (which are due to amyloid-β peptide aggregation), NFTs are the representative pathological markers of many neurodegenerative disorders, including Alzheimer’s disease [40] It is thus obvious that selective imaging of NFTs to differentiate them from other fibril proteins is likely to contribute to a better understanding, and possibly diagnosing of neurodegenerative diseases in which NFTs are implicated. Several imaging probes for NFTs have been developed so far but none features high enough selectivity. In 2009, Hamachi’s group developed selective fluorescent probes for neurofibrillary tangles (NFTs), taking an advantage of the fact that NFTs are formed from hyperphosphorylated proteins [41], These probes feature a very specific binding affinity for phosphorylated regions, lent to them by a binuclear Zn(II)-dipicolylamine (DPA) complex. The Zn-DPA complex itself is not fluorescent but specifically binds

172

Pratyush Kumar Mishra et al.

phosphoester groups in aqueous solution (Kd ~ 10–80 μM) [42]. Two Zn-DPA complexes are attached to one green fluorescent boron-dipyrromethene (BODIPY) fluorophore in order to increase the binding affinity for proximal bis-phosphorylated sites, using cooperative forces. Several bis-phosphorylated tau protein sites were tested in vitro, including tau(204–217), tau(227–238) and tau(394–403) peptides. Its bis-phosphorylated sites with three amino acid residues between them generally exhibited strong binding affinity for the Zn-DPA probe (Kd between 4 and 40 μM) and yielded a ~1.5-fold increase in 500–600 nm fluorescence. Interestingly, this probe could not bind to other bis-phosphorylated tau peptides which had fewer (1) or more (5) amino acid residues between the phosphorylated sites, suggesting that the two Zn-DPA complexes of the probe are “correctly” separated so that they can simultaneously coordinate with the two bis-phosphorylated sites (positions i, i + 4). Of course, this probe has no binding affinity for monophosphorylated and nonphosphorylated peptides, resulting in the ultrahigh selectivity of the Zn-DPA probe to discriminate between hyperphosphorylated tau proteins and senile plaques of nonhyperphosphorylated Aβ plaques. Indeed, in vitro studies revealed that this imaging probe very selectively stained phosphorylated tau oligomers but not Aβ plaques while thioflavin (ThT) strongly stained both of them (Fig. 11). Selective staining of NFTs was also confirmed directly in situ, in AD patient’s brain slices; the regions stained with the Zn-DPA probe perfectly overlapped with those stained with antiphosphorylated Tau monoclonal antibody [41]. Overall, selective fluorescent amyloid/prion/NFT imaging probes such as ANCA and Zn-DPA may prove very important as diagnostic tools, owing to their potential to distinguish between medical conditions with very similar symptoms and pathological characteristics. However, their molecular weight is over 400 Da (cf. Subheading 3) and the Zn-DPA complex additionally carries a high cationic charge. This prevents them from efficiently crossing the blood–brain barrier (i.e., being employed directly in vivo), which poses a challenge for even more suitable probes to be developed in the future.

6

Selective Imaging Probes for Enzymatic Activity

6.1 Probes for β-Secretase

Aβ peptides are the result of a sequential cleavage of amyloid precursor protein (APP), a transmembrane protein on the cell surface, by two proteases, β- and γ-secretase. The former one, also referred to as β-site APP cleaving enzyme (BACE) is a membraneanchored aspartic protease that cleaves APP in the extracellular domain (N-terminal region) [43]. After that, γ-secretase subsequently cleaves the C-terminus of the processed APP, releasing

Optical Imaging Probes for Amyloids

173

BODIPY-Zn-DPA A

B

C

D

10 µm

E

F

G

H

10 µm

Fig. 11 Chemical structure of BODIPY-Zn-DPA and fluorescence microscopy analysis of in vitro aggregates of phosphorylated-Tau (“pTau”), nonphosphorylated Tau (“Tau”), and Aβ1–42 stained with (A–D) thioflavin T and (E–H) BODIPY-Zn-DPA. In B and F, staining of p-Tau was conducted in the presence of sodium pyrophosphate (PPi, 100 μM), a strong binding competitor of the Zn-DPA complex. Reproduced from Ref. [41] by permission of © American Chemical Society

the toxic Aβ peptides. The first cleavage by BACE is critical for the generation of Aβ peptides because APP processed by α-secretase could not generate Aβ peptides even if γ-secretase was present. Therefore, imaging probes for the BACE activity in living neurons or brain tissue are important to monitor AD progress. Furthermore, these probes have the potential to screen specific inhibitors based on intensity changes of the probe’s fluorescence, following inhibition of BACE activity. These inhibitors are prospective drugs capable of preventing Aβ peptide generation [44]. Generally, the imaging probes for BACE activity are composed of a BACE-cleavable specific peptide sequence (EVNLDAEFK) and two attached fluorophores featuring high Fo¨rster Resonance Energy Transfer (FRET) efficiency. In 2007, Luo’s group successfully developed a genetically encodable BACE substrate using

174

Pratyush Kumar Mishra et al.

monomeric versions of cyan fluorescent protein (mCerulean) and yellow fluorescent protein (mCitrine). This probe exhibited good FRET efficiency when the two fluorescent proteins were closely linked together. However, when the linkage was cleaved by BACE, the FRET efficiency (of energy transfer from mCerulean to mCitrine) dropped [45]. A spatiotemporal dynamics of BACE activity in living cells/tissues can thus be followed by monitoring the probe’s FRET efficiency. In 2012, Franz’s group reported a synthetic peptide-based FRET probe, β-MAP, for measuring BACE activity in living cells [46]. A BACE-cleavable substrate peptide consisting of 13 amino acids (EVNLDAHFWADRK) was synthesized, fluorescent 7dimethylamino-coumarin-4-acetic acid (DMACA) attached to the N-terminus of the peptide, and a 4-(dimethylaminoazo)benzene4-carboxylic acid (DABCYL) group appended to the lysine amino group of the peptide. In this system, DABCYL quenches the fluorescence of the proximally linked fluorophore, DMACA, through FRET so that intact β-MAP exhibits negligible fluorescence. However, when BACE cleaves it, the DMACA domain is released from DABCYL and becomes fluorescent. A polyethylene glycol (PEG) linker is attached to DABCYL, lending the probe more flexibility and solubility in aqueous solution. A dihydrocholesterol group is attached at the C-terminus so that the probe is targeted to cellular membranes, including lipid rafts, the site of Aβ oligomerization [47]. The probe can even reach intracellular vesicles (Fig. 12) whose acidic environment activates BACE and enhances DMACA fluorescence. Gene silencing (knockdown) experiments proved that the cleavage is indeed driven by the BACE activity. High (although not 100%) specificity of staining with β-MAP has been documented in living HeLa cells. When treated with Axon 1125, a BACE inhibitor, a significant decrease in β-MAP’s fluorescence was encountered with increasing inhibitor concentration. To the best of author’s knowledge, the β-MAP probe has not been applied to neurons yet. 6.2 Probe for Amyloid-Binding Alcohol Dehydrogenase

Amyloid peptide-binding alcohol dehydrogenase (ABAD) is an important metabolic enzyme involved in oxidative reactions for β-oxidation of fatty acids, isoleucine catabolism, and steroid metabolism. Due to its multifunctionality, it bears many other names, including 17β-hydroxysteroid dehydrogenase 10 (17β-HSD10 or simply HSD10), endoplasmic reticulum-associated amyloid-binding protein (ERAB) [48], human brain short chain L-3-hydroxyacyl-CoA dehydrogenase (SCHAD) [49], and human type II hydroxyacyl-CoA dehydrogenase (HADH II) [50]. ABAD has been shown to exhibit strong protein-protein interactions with the amyloid-β peptide. The ABAD-Aβ peptide complex was investigated by, surface-plasmon resonance, immunoprecipitation and X-ray crystallography. In vitro, Aβ

Optical Imaging Probes for Amyloids

DMACA

BACE substrate

175

Membrane anchor

DABCYL PEG

O O

O N H

E

V

N

L

D

A

H

F

W

A

D

R

H K N HN

N

O O

R 5

3

O

D O

O

NH2

O

β-MAP N N N

Fig. 12 Chemical structure of β-secretase membrane-anchored probe (β-MAP) and time-lapse images of a HeLa cell stained with βMAP (200 nM). DMACA fluorescence from cleaved β-MAP (λexc/em ¼ 390/470 nm) is localized here primarily to intracellular vesicles as their acidic environment both enhances it and activates the β-site amyloid precursor protein (APP) cleaving enzyme (BACE). The rate of fluorescence increase over time reflects the BACE activity. Reproduced from Ref. [46] by permission of © Wiley

peptide strongly inhibits the oxidative activity of purified ABAD protein by distorting its binding site for a cofactor, nicotinamide adenine dinucleotide (NAD), thus impairing ABAD’s normal function. At cellular level, coexpression of ABAD and mutant amyloid precursor protein has been shown to stimulate the generation of reactive oxygen species, eventually leading to mitochondrial dysfunction and cell death [51, 52]. It is obvious that a measurement of ABAD activity (e.g., by optical means) can indicate intracellular Aβ peptide generation. In 2007, Sames’s group developed the first imaging probe of this kind, cyclohexenyl amino naphthalene alcohol [53], later abbreviated as CHANA [54]. Interestingly, CHANA (λmax ex/em ¼ 313/444 nm, quantum yield ~44% in buffer; and almost nonfluorescent in chloroform) is effective as a probe not because it would bind to ABAD; it is oxidatively converted by ABAD to a highly fluorescent, greenish product, cyclohexenyl amino naphthalene ketone, CHANK (Fig. 13). It is brightly fluorescent in chloroform mimicking cell

176

Pratyush Kumar Mishra et al.

membranes (λmax ex/em ¼ 385/510 nm, quantum yield ~70%) and almost nonfluorescent in water. This very property (combined with the spectral shift between CHANA in buffer and CHANK in chloroform) efficiently eliminates background fluorescence in cellular imaging. In total, five fluorescent aminonaphthalene probes were synthesized, all similar to estradiol in shape, with an electronconjugated en-ol group attached to the 2-position of naphthalene. This conjugated structure results in the generation of a fluorogenic response during oxidation of alcohol (en-ol) to ketone (en-one) by purified ABAD. As ketones usually function as electronwithdrawing groups, naphthalene’s electron-donating (amino) and withdrawing (ketone) groups comprise an efficient “pushpull” system capable of yielding strong fluorescence [53]. As CHANA has one chirality center in its structure, two enantiomers designated as (+)-CHANA and (
)-CHANA are present in it as a racemic mixture [41, 42] employed in [41]. In 2010, the same group has reported that (
)-CHANA, synthesized and isolated using asymmetric (chiral) catalysts, exhibited properties favorable for ABAD’s activity measurement. (
)-CHANA preferentially reacted with ABAD (kcat/Km ¼ 48 μM
1 min
1) while (+)CHANA was consumed by ABAD very slowly (kcat/ Km ¼ 2.5 μM
1 min
1) [54]. The (
)-CHANA enantiomer was successfully employed to monitor metabolic activity in living ABAD-transfected HEK293T cells. After only 0.5 h incubation with (
)-CHANA, the bright greenish fluorescence could be observed in ABAD-transfected cells, while nontransfected (“null”) cells exhibited negligible fluorescence even after 3 h of incubation. The opposite has been shown to be the case for the (+)-CHANA enantiomer, implying that while (
)-CHANA is specifically recognized by ABAD, (+)-CHANA is consumed by other endogenous cellular oxidases or dehydrogenases. The specificity of (
)-CHANA staining for ABAD was further confirmed in ABAD-transfected human neuroblastoma SK-N-SH cells, by treating them not only with Aβ42 but additionally also with AG18051, an irreversible inhibitor of ABAD [54]. Colocalization of (
)-CHANA and MitoTracker Deep Red, a mitochondrial dye has also been demonstrated at/near the mitochondrial membranes of human neuroblastoma SK-N-SH cells (Fig. 13), suggesting that ABAD imported by mitochondria is still functional [54]. This observation is consistent with results from other studies of the ABAD-Aβ complex in mitochondria [51, 52], and is particularly important for understanding its role in mitochondrial dysfunction, as related to Alzheimer’s disease. Overall, probes capable of monitoring BACE and ABAD enzyme activity in real time directly in living cells, have a potential to provide deeper insight into the mechanisms of neurodegenerative disorders. They can also serve as a platform for efficient high-

Optical Imaging Probes for Amyloids

177

CHANK, fluorescent product

CHANA Probe

A

20 µm

C

B

Fig. 13 Chemical structure of CHANA probe designated here as a racemic mixture of (+)-CHANA and (
)CHANA, and the principle of amyloid peptide-binding alcohol dehydrogenase (ABAD)‘s activity detection by fluorescence (adapted from Ref. [46]). CHANA, cyclohexenyl amino naphthalene alcohol. (A) Colocalization image (right) of (
)-CHANA (green, center) and MitoTracker Deep Red (red, left) in human neuroblastoma SKN-SH cells. (B) Close-up of colocalized fluorescence staining in the mitochondrial membrane. (C) Fluorescence intensity of MitoTracker and CHANK in a single mitochondrion, along the transect line shown in B. Reproduced from Ref. [54] by permission of © American Chemical Society

throughput screening to identify/develop specific inhibitors of these metabolic enzymes’ activities strongly linked to the progression of such medical conditions.

7

Superresolution Imaging Probes While the probes listed in the previous sections can, of course, be used in structured-illumination microscopy (SIM, resolution

178

Pratyush Kumar Mishra et al.

PD-NA-TEG

A

B

Fig. 14 Chemical structure of PD-NA-TEG and images of transgenic mouse (Tg6799) brain slice (9-month-old male). (A) Conventional fluorescence microscopy. (B) Super resolution microscopy. Reproduced from Ref. [55] by permission of © Royal Society of Chemistry

ca. 100 nm), they are not suitable for localization-based superresolution imaging (PALM/STORM) capable of revealing much smaller details. The recently synthesized PD-NA-TEG probe overcomes this limitation, and a resolution of 28 nm was achieved. Application of the probe to a brain slice of transgenic (Tg6799) mouse, revealed that the Aβ deposits are not randomly aggregated, but originate from a single point and grow radially [55] (Fig. 14). This nucleation may be due to protein β-sheet transition. Revealing this type of morphological detail may in the future help in finding a way of arresting the growth of Aβ aggregates.

8

Summary The present chapter covers neurodegenerative disorders associated with amyloid deposition in brain tissue from the viewpoint of recently developed probes for optical imaging. As described, they offer multiple advantages over PET probes although only few of them have been clinically tested yet. Novel multifunctional probes

Optical Imaging Probes for Amyloids

179

compatible with NIR imaging, two-photon excitation, blood–brain barrier permeability, fast clearance rates, and discrimination between specific fibril proteins are needed to bring us closer to understanding the fundamental principles governing the onset of neurodegenerative disorders.

Acknowledgments This work was supported by the National Research Foundation of Korea (NRF-2018K2A9A2A08000087, NRF-2019R1A2C 3008463) and the Organelle Network Research Center (NRF-2017R1A5A1015366). HWR also acknowledges support from the KBRI basic research program through Korea Brain Research Institute funded by Ministry of Science and ICT (17-BR-01, 19-BR-04-01). References 1. Sloane PD, Zimmerman S, Suchindran C, Reed P, Wang L, Boustani M, Sudha S (2002) The public health impact of Alzheimer’s disease, 2000-2050: potential implication of treatment advances. Annu Rev Public Health 23:213–231. https://doi.org/10.1146/ annurev.publhealth.23.100901.140525 2. Eisenberg D, Jucker M (2012) The amyloid state of proteins in human diseases. Cell 148:1188–1203. https://doi.org/10.1016/j. cell.2012.02.022 3. Scherzinger E, Lurz R, Turmaine M, Mangiarini L, Hollenbach B, Hasenbank R, Bates GP, Davies SW, Lehrach H, Wanker EE (1997) Huntingtin-encoded polyglutamine expansions form amyloid-like protein aggregates in vitro and in vivo. Cell 90:549–558. https://doi.org/10.1016/S0092-8674(00) 80514-0 4. Chow VW, Mattson MP, Wong PC, Gleichmann M (2009) An overview of APP processing enzymes and products. NeuroMolecular Med 12:1–12. https://doi.org/10.1007/ s12017-009-8104-z 5. Recchia A, Debetto P, Negro A, Guidolin D, Skaper SD, Giusti P (2004) Alpha-synuclein and Parkinson’s disease. FASEB J 18:617–626. https://doi.org/10.1096/fj.030338rev 6. Glatzel M, Abela E, Maissen M, Aguzzi A (2003) Extraneural pathologic prion protein in sporadic Creutzfeldt-Jakob disease. New Engl J Med 349:1812–1820. https://doi. org/10.1056/NEJMoa030351

7. Hardy J (2006) Alzheimer’s disease: the amyloid cascade hypothesis: an update and reappraisal. J Alzheimers Dis 9(Suppl.3):151–153 8. Allison RR, Moghissi K (2013) Photodynamic therapy (PDT): PDT mechanisms. Clin Endosc 46:24–29. https://doi.org/10.5946/ce. 2013.46.1.24 9. Allison RR, Downie GH, Cuenca R, Hu XH, Childs CJH, Sibata CH (2004) Photosensitizers in clinical PDT. Photodiagn Photodyn Ther 1:27–42. https://doi.org/10.1016/ S1572-1000(04)00007-9 10. Ono M, Saji H (2015) Recent advances in molecular imaging probes for β-amyloid plaques. Med Chem Commun 6(3):391–402. https://doi.org/10.1039/c4md00365a 11. Fu H, Cui M (2018) Fluorescent imaging of amyloid-β deposits in brain: an overview of probe development and a highlight of the applications for in vivo imaging. Curr Med Chem 25(23):2736–2759. https://doi.org/ 10.2174/0929867325666180214110024 12. Verwilst P, Kim HS, Kim S, Kang C, Kim JS (2018) Shedding light on tau protein aggregation: the progress in developing highly selective fluorophores. Chem Soc Rev 47 (7):2249–2265. https://doi.org/10.1039/ c7cs00706j 13. Xu MM, Ren WM, Tang XC, Hu YH, Zhang HY (2018) Advances in development of fluorescent probes for detecting amyloid-beta aggregates. Acta Pharmacol Sin 37 (6):719–730. https://doi.org/10.1038/aps. 2015.155

180

Pratyush Kumar Mishra et al.

14. Peng C, Wang X, Li Y, Li H-W, Wong MS (2019) Versatile fluorescent probes for nearinfrared imaging of amyloid-b species in Alzheimer’s disease mouse model. J Mater Chem B 7(12):1986–1995. https://doi.org/10. 1039/c9tb00161a 15. Frid P, Anisimov SV, Popovic N (2007) Congo red and protein aggregation in neurodegenerative diseases. Brain Res Rev 53:135–160. https://doi.org/10.1016/j.brainresrev.2006. 08.001 16. Marcus A, Sadimin E, Richardson M, Goodell L, Fyfe B (2012) Fluorescence microscopy is superior to polarized microscopy for detecting amyloid deposits in Congo red-stained trephine bone marrow biopsy specimens. Am J Clin Pathol 138:590–593. https://doi.org/10.1309/ AJCP6HZI5DDQTCRM 17. Howie AJ, Brewer DB (2009) Optical properties of amyloid stained by Congo red: history and mechanisms. Micron 40:285–301. https://doi.org/10.1016/j.micron.2008.10. 002 18. Styren SD, Hamilton RL, Styren GC, Klunk WE (2000) X-34, a fluorescent derivative of Congo red: a novel histochemical stain for Alzheimer’s disease pathology. J Histochem Cytochem 48:1223–1232. https://doi.org/10. 1177/002215540004800906 19. Drake J, Link CD, Butterfield DA (2003) Oxidative stress precedes fibrillar deposition of Alzheimer’s disease amyloid β-peptide (1-42) in a transgenic Caenorhabditis elegans model. Neurobiol Aging 24:415–420. https://doi. org/10.1016/S0197-4580(02)00225-7 20. Khurana R, Coleman C, Ionescu-Zanetti C, Carter SA, Krishna V, Grover RK, Roy R, Singh S (2005) Mechanism of thioflavin T binding to amyloid fibrils. J Struct Biol 151:229–238. https://doi.org/10.1016/j. jsb.2005.06.006 21. Klunk WE, Engler H, Nordberg A, Wang Y, Blomqvist G, Holt DP, Bergstro¨m M, Savitcheva I, Huang GF, Estrada S, Ause´n B, Debnath ML, Barletta J, Price JC, Sandell J, Lopresti BJ, Wall A, Koivisto P, Antoni G, Mathis CA, La˚ngstro¨m B (2004) Imaging brain amyloid in Alzheimer’s disease with Pittsburgh compound-B. Ann Neurol 55:306–319. https://doi.org/10.1002/ana.20009 22. Weissleder R, Pittet MJ (2008) Imaging in the era of molecular oncology. Nature 452:580–589. https://doi.org/10.1038/ nature06917 23. Cui M, Ono M, Watanabe H, Kimura H, Liu B, Saji H (2014) Smart near-infrared fluorescence probes with donor-acceptor structure

for in vivo detection of β-amyloid deposits. J Am Chem Soc 136:3388–3394. https://doi. org/10.1021/ja4052922 24. Nesterov EE, Skoch J, Hyman BT, Klunk WE, Bacskai BJ, Swager TM (2005) In vivo optical imaging of amyloid aggregates in brain: design of fluorescent markers. Angew Chem Int Ed 44:5452–5456. https://doi.org/10.1002/ anie.200500845 25. Li Q, Lee J-S, Ha C, Park CB, Yang G, Gan WB, Chang Y-T (2004) Solid-phase synthesis of styryl dyes and their application as amyloid sensors. Angew Chem Int Ed 43:6331–6335. https://doi.org/10.1002/anie.200461600 26. Ran C, Xu X, Raymond SB et al (2009) Design, synthesis, and testing of difluoroboronderivatized curcumins as near-infrared probes for in vivo detection of amyloid-β deposits. J Am Chem Soc 131(42):15257–15261. https://doi.org/10.1021/ja9047043 27. Kim H, Im YH, Ahn J et al (2019) Synthesis and in vivo characterization of 18F-labeled difluoroboron-curcumin derivative for β-amyloid plaque imaging. Sci Rep 9(1):6747. https://doi.org/10.1038/s41598-01943257-9 28. Chung C, Srikun D, Lim CS, Chang CJ, Cho BR (2011) A two-photon fluorescent probe for ratiometric imaging of hydrogen peroxide in live tissue. Chem Commun (Camb) 47:9618–9620. https://doi.org/10.1039/ c1cc13583j 29. Heo CH, Kim KH, Kim HJ, Baik SH, Song H, Kim YS, Lee J, Mook-Jung I, Kim HM (2013) A two-photon fluorescent probe for amyloid-β plaques in living mice. Chem Commun (Camb) 49:1303–1305. https://doi.org/10. 1039/c2cc38570h 30. Helmchen F, Denk W (2005) Deep tissue two-photon microscopy. Nat Methods 2:932–940. https://doi.org/10.1038/ nmeth818 31. Cook NP, Torres V, Jain D, Martı´ AA (2011) Sensing amyloid-β aggregation using luminescent dipyridophenazine ruthenium (II) complexes. J Am Chem Soc 133:11121–11123. https://doi.org/10. 1021/ja204656r 32. Murphy CJ, Barton JK (1993) Ruthenium complexes as luminescent reporters of DNA. Methods Enzymol 226:576–594. https://doi. org/10.1016/0076-6879(93)26027-7 33. Jime´nez-Herna´ndez ME, Orellana G, Montero F, Portole´s MT (2000) A ruthenium probe for cell viability measurement using flow cytometry, confocal microscopy and timeresolved luminescence. Photochem Photobiol

Optical Imaging Probes for Amyloids 72:28–34. https://doi.org/10.1562/00318655(2000)0722.0.CO;2 34. Cook NP, Ozbil M, Katsampes C, Prabhakar R, Martı´ AA (2013) Unraveling the photoluminescence response of light-switching ruthenium(II) complexes bound to amyloid-β. J Am Chem Soc 135:10810–10816. https:// doi.org/10.1021/ja404850u 35. Lewis GN, Kasha M (1944) Phosphorescence and the triplet state. J Am Chem Soc 66:2100–2116. https://doi.org/10.1021/ ja01240a030 36. Hudson SA, Ecroyd H, Kee TW, Carver JA (2009) The thioflavin T fluorescence assay for amyloid fibril detection can be biased by the presence of exogenous compounds. FEBS J 276:5960–5972. https://doi.org/10.1111/j. 1742-4658.2009.07307.x 37. Cook NP, Kilpatrick K, Segatori L, Martı´ AA (2012) Detection of α-synuclein amyloidogenic aggregates in vitro and in cells using light-switching dipyridophenazine ruthenium (II) complexes. J Am Chem Soc 134:20776–20782. https://doi.org/10. 1021/ja3100287 38. Khurana R, Uversky VN, Nielsen L, Fink AL (2001) Is Congo red an amyloid-specific dye? J Biol Chem 276:22715–22721. https://doi. org/10.1074/jbc.M011499200 39. Cao K, Farahi M, Dakanali M, Chang WM, Sigurdson CJ, Theodorakis EA, Yang J (2012) Aminonaphthalene 2-cyanoacrylate (ANCA) probes fluorescently discriminate between amyloid-β and prion plaques in brain. J Am Chem Soc 134:17338–17341. https:// doi.org/10.1021/ja3063698 40. Binder LI, Guillozet-Bongaarts AL, GarciaSierra F, Berry RW (2005) Tau, tangles, and Alzheimer’s disease. Biochim Biophys Acta Mol basis Dis 1739:216–223. https://doi. org/10.1016/j.bbadis.2004.08.014 41. Ojida A, Sakamoto T, Inoue M, Fujishima S, Lippens G, Hamachi I (2009) Fluorescent BODIPY-based Zn(II) complex as a molecular probe for selective detection of neurofibrillary tangles in the brains of Alzheimer’s disease patients. J Am Chem Soc 131:6543–6548. https://doi.org/10.1021/ja9008369 42. Ojida A, Mito-Oka Y, Sada K, Hamachi I (2004) Molecular recognition and fluorescence sensing of monophosphorylated peptides in aqueous solution by bis(zinc(II)-dipicolylamine)-based artificial receptors. J Am Chem Soc 126:2454–2463. https://doi.org/10. 1021/ja038277x 43. Vassar R (1999) Beta-secretase cleavage of alzheimer’s amyloid precursor protein by the

181

transmembrane aspartic protease BACE. Science 286:735–741. https://doi.org/10. 1126/science.286.5440.735 44. Ghosh A, Gemma S, Tang J (2008) β-Secretase as a therapeutic target for Alzheimer’s disease. Neurotherapeutics 5:399–408. https://doi. org/10.1016/j.nurt.2008.05.007 45. Lu J, Zhang Z, Yang J, Chu J, Li P, Zeng S, Luo Q (2007) Visualization of beta-secretase cleavage in living cells using a genetically encoded surface-displayed FRET probe. Biochem Biophys Res Commun 362:25–30. https://doi.org/10.1016/j.bbrc.2007.07. 145 46. Folk DS, Torosian JC, Hwang S, McCafferty DG, Franz KJ (2012) Monitoring β-secretase activity in living cells with a membraneanchored FRET probe. Angew Chem Int Ed 51:10795–10799. https://doi.org/10.1002/ anie.201206673 47. Kim S-I, Yi J-S, Ko Y-G (2006) Amyloid beta oligomerization is induced by brain lipid rafts. J Cell Biochem 99:878–889. https://doi.org/ 10.1002/jcb.20978 48. Du Yan S, Shi Y, Zhu a, Fu J, Zhu H, Zhu Y, Gibson L, Stern E, Collison K, Al-Mohanna F, Ogawa S, Roher a, Clarke SG, Stern DM (1999) Role of ERAB/L-3-Hydroxyacyl-coenzyme A dehydrogenase type II activity in Aβ-induced cytotoxicity. J Biol Chem 274:2145–2156. https://doi.org/10.1074/ jbc.274.4.2145 49. He X-Y, Merz G, Mehta P, Schulz H, Yang S-Y (1999) Human brain short chain L-3-hydroxyacyl coenzyme A dehydrogenase is a singledomain multifunctional enzyme: Characterization of a novel 17-hydroxysteroid dehydrogenase. J Biol Chem 274:15014–15019. https:// doi.org/10.1074/jbc.274.21.15014 50. Oppermann UCT, Salim S, Tjernberg LO, Terenius L, Jo¨rnvall H (1999) Binding of amyloid β-peptide to mitochondrial hydroxyacylCoA dehydrogenase (ERAB): Regulation of an SDR enzyme activity with implications for apoptosis in Alzheimer’s disease. FEBS Lett 451:238–242. https://doi.org/10.1016/ S0014-5793(99)00586-4 51. Lustbader JW, Cirilli M, Lin C, Xu HW, Takuma K, Wang N, Caspersen C, Chen X, Pollak S, Chaney M, Trinchese F, Liu S, Gunn-Moore F, Lue L-F, Walker DG, Kuppusamy P, Zewier ZL, Arancio O, Stern D, Yan SS, Wu H (2004) ABAD directly links Aβ to mitochondrial toxicity in Alzheimer’s disease. Science 304:448–452. https:// doi.org/10.1126/science.1091230 52. Yan SD, Stern DM (2005) Mitochondrial dysfunction and Alzheimer’s disease: role of

182

Pratyush Kumar Mishra et al.

amyloid-beta peptide alcohol dehydrogenase (ABAD). Int J Exp Pathol 86:161–171. https://doi.org/10.1111/j.0959-9673.2005. 00427.x 53. Froemming MK, Sames D (2007) Harnessing functional plasticity of enzymes: a fluorogenic probe for imaging 17β-HSD10 dehydrogenase, an enzyme involved in Alzheimer’s and Parkinson’s diseases. J Am Chem Soc 129:14518–14522. https://doi.org/10. 1021/ja072601x

54. Muirhead KEA, Froemming M, Li X, Musilek K, Conway SJ, Sames D, GunnMoore FJ (2010) (
)-CHANA, a fluorogenic probe for detecting amyloid binding alcohol dehydrogenase HSD10 activity in living cells. ACS Chem Biol 5:1105–1114. https://doi. org/10.1021/cb100199m 55. Wang Y-L, Fan C, Xin B et al (2018) AIE-based super-resolution imaging probes for β-amyloid plaques in mouse brains. Mater Chem Front 2(8):1554–1562. https://doi. org/10.1039/C8QM00209F

Chapter 6 Chemical Clearing of GFP-Expressing Neural Tissues Klaus Becker, Saiedeh Saghafi, Christian Hahn, Nina J€ahrling, and Hans-Ulrich Dodt Abstract In the course of the recent rapid development of optical-sectioning and tomographic microscopy techniques, such as confocal microscopy, light-sheet microscopy, or optical coherence tomography (OCT), chemical tissue clearing experienced a renaissance. As most organic solvents commonly employed in tissue dehydration and clearing quench the green fluorescent protein (GFP) and related genetic markers, microscopy of chemically cleared GFP-expressing specimens remains a challenging task. In this chapter we review a variety of tissue-clearing methods and describe in detail a protocol that well preserves GFP fluorescence while maintaining tissue transparency. Key words Light-sheet microscopy, Mouse brain, Hippocampus, Spinal cord, Tissue clearing, Dibenzyl ether, Tetrahydrofuran

Abbreviations BABB BHT DBE PBS THF

1

Benzyl alcohol/benzyl benzoate Butyl hydroxytoluene Dibenzyl ether Phosphate-buffered saline Tetrahydrofuran

Introduction

1.1 Evolution of Chemical Tissue Clearing

The history of chemical clearing (or, to be more precise, optical clearing by chemical means) of biological preparations dates back to the beginning of the twentieth century when a German anatomist Werner Spalteholz published a pioneering study about this topic [1]. The clearing and brightening effect on biological tissues of some organic solvents and ethereal oils, such as Canada balsam, clove oil, carbon disulfide, and some other substances, was described even earlier, for example, by Lundvall [2, 3]. However,

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_6, © Springer Science+Business Media LLC 2020

183

184

Klaus Becker et al.

Spalteholz was the first who investigated this effect systematically and understood the fundamental physical principle that matching the average refractive index of a sample to that of an incubation liquid soaking it markedly reduces light scattering generated by intra-tissue variations of the refractive index; consult Chapter 7 by Colarusso [4] for details. As a result, the specimens become translucent or even transparent, providing light absorption by any chromophores that may be present in them is only moderate [1, 5, 6]. Guided by this simple rule, Spalteholz tested various organic liquids, most of them ethereal oils, for their potential to optically clear organs isolated from humans and animals. In 1911 he published his results in German in a slim book entitled “On making human and animal specimens transparent” [1]. By screening various mixtures of organic solvents, largely composed of methyl salicylate, benzyl benzoate or isosafrole, Spalteholz found that a proportion of about 5:3 vol. parts of methyl salicylate/benzyl benzoate (or 3:1 vol. parts of methyl salicylate and isosafrole) is suitable for clearing of many anatomical preparations. This particular mixture became a standard for a commercial production of the so-called “Spalteholz preparations” by the Anatomical Institute of the Museum for Hygiene in Dresden, for which Spalteholz was occasionally working as a consultant. Although Spalteholz’s clearing method yielded high-grade anatomical demonstration samples sold to universities, anatomical institutions and schools worldwide, its scientific impact remained quite limited outside the anatomical education domain [7]. Distribution of these preparations had stopped in 1971 and Spalteholz’s tissue-clearing technique slid into oblivion. In the following decades chemical tissue clearing was only rarely revisited, mainly in the field of embryology and developmental biology. Owing to the rapid development of novel optical-sectioning and tomographic techniques allowing large fields of view and high depths of focus, such as optical projection tomography (OPT) [8], optical transmission tomography, optical emission tomography [9], and light-sheet microscopy [10–12] developed later on, chemical tissue clearing found new applications in deep-tissue imaging of large samples such as entire mouse brains or embryos [9, 10, 13–16]. Most clearing protocols are based on matching the refractive index of the immersion medium to that of the organic components of the tissue—typically the proteins. These solvents can be (a) lipophilic (non-polar) compounds easily penetrating the cell membranes or (b) hydrophilic (polar) substances composed of sugars, polyalcohols, or iodine compounds. In some cases the membrane permeability is enhanced by addition of detergents such as Triton X-100, sodium dodecyl sulfate (SDS), or saponin. While the first type (a) of clearing recipes requires a careful prior tissue dehydration, the second type (b) is directly applicable to fixed or unfixed tissues. Table 1 lists the chemical clearing techniques since 1897.

Chemical Clearing of GFP-Expressing Neural Tissues

185

Table 1 A historical overview of chemical tissue-clearing techniques

Refs. Author(s)

Dehydration medium

Clearing medium

Comments

Ethanol, benzene

Mixtures of benzene (mixed with Severely toxic vapors, very unpleasant smell of peppermint oil) and carbon hydrogen sulfide disulfide

[1]

Spalteholz Ethanol, 1911 benzene

Mixtures of methyl salicylate with Severe quenching of GFP benzyl benzoate or isosafrole fluorescence by methyl salicylate

[17]

Drahn 1922

Solutions of naphthalene in tetraline

[18]

Dent et al. Ethanol 1989

[19]

Chiang et al. 2002

None

Distributed under the trade Mixture of unknown name FocusClear. High composition, probably cost (~$180/5 mL), does containing DMSO, diatrizoate not work in large acid, EDTA, glucamine, specimens such as whole NADP, sodium diatrizoate and embryos or mouse brains polyoxyalkalene derivatives

[20]

Staudt et al. 2007

None

2,20 -Thiodiethanol

Applicable to cultured cells, almost no clearing effect in tissues

[21]

Efimova 2-Butoxy and ethanol Anokhin 2009

75% Aqueous solution of sodium/meglumine diatrizoate

Specimens have to be dehydrated first; rehydration takes place in the clearing medium

[22]

Tsai et al. 2009

None

60% Sucrose solution, membrane Transparency increase quite moderate, usable only in permeabilization with 2% (v/v) thin tissue slices Triton X-100

[23]

Hama et al. 2011

None

“ScaleA2” 4M urea, 10% (w/v) Incubation time of several glycerol and 0.1% (w/v) Triton weeks, strong tissue X-100 swelling yielding high fragility

[13]

Ertu¨rk et al. 2012

Tetrahydrofuran BABB (composition see above) (THF), waterand peroxidefree

[24]

Becker et al. 2012

Tetrahydrofuran Dibenzyl ether (DBE), waterand peroxide-free (THF), waterand peroxidefree

[2, 3] Lundvall 1904, 1905

Ethanol

Toxic vapors, intense smell of antimoth powder

“Murray’s clear” 1:2 vol. parts of Standard clearing medium for benzyl alcohol and benzyl whole-mount preparations. benzoate (BABB) Limited GFP compatibility

(continued)

186

Klaus Becker et al.

Table 1 (continued)

Refs. Author(s)

Dehydration medium

[25]

None

Chung et al. 2013

Clearing medium

Comments

Owing to high permeability Electrophoretic removal of cell of cleared samples and membranes after embedding in excellent preservation of polyacrylamide hydrogel for proteins molecular stabilization of morphological phenotyping and wholestructures. Final clearing is mount immunostaining of achieved via incubation in cleared samples is possible FocusClear or glycerol

1.2 Solvent-Based (Lipophilic) Tissue Clearing with Benzyl Alcohol/Benzyl Benzoate (BABB)

To map the distribution of intermediate filament proteins during embryonic development, A. Murray developed a clearing medium consisting of a mixture of 1 vol part benzyl alcohol (BA) and 2 vol parts benzyl benzoate (BB), commonly abbreviated as BABB [18, 26]. Since BABB was found to be compatible with 40 ,6-diamidino-2-phenylindole (DAPI) and other fluorescent markers, it became a standard tissue-clearing medium for deep-tissue imaging of whole-mount preparations in embryological research. Compared to methyl salicylate which also found a widespread use as a tissue-clearing medium [27], BABB generally provides better and faster clearing results even though its refractive index (~1.56) is very close to that of methyl salicylate (~1.54). BABB’s much lower polarity (water solubility 0.0154 g/L vs. ~0.7 g/L for methyl salicylate) enabling better penetration through cell membranes, lipidic (non-polar) in nature is the most likely explanation. Unfortunately, BABB significantly quenches the fluorescence of GFP over time. Therefore, to achieve sufficiently high fluorescence intensities, incubation times should be kept as short as possible [10]. Furthermore, high GFP-expression rates are required. Recently, it was demonstrated that substitution of BABB with dibenzyl ether (DBE) yields superior GFP fluorescence and better tissue transparency [24]. A DBE-based, GFP-friendly dehydration and clearing protocol for large samples such as mouse brains or entire embryos is described in Subheading 1.4.

1.3 Water-Based (Hydrophilic) Tissue-Clearing Reagents

In recent years various attempts have been made to develop waterbased clearing media inherently not requiring tissue dehydration and applicable to both fixed and unfixed specimens. CellExplorer Labs (Taiwan) distribute a water-miscible tissueclearing medium with a brand name FocusClear containing dimethyl sulfoxide (DMSO), diatrizoate acid, EDTA, glucamine, NADP, sodium diatrizoate and polyoxyalkalenes in unspecified proportions [19] FocusClear has a refractive index of 1.45 and was successfully applied to clear, for example, insect brains [28]

Chemical Clearing of GFP-Expressing Neural Tissues

187

and murine intestine [29]. Its price of about $180/5 mL vial makes it significantly more expensive than custom-made clearing media, thereby limiting its use to small samples such as tissue slices. FocusClear preserves fluorescence of GFP and related markers [30, 31] but poorly penetrates lipid-rich myelinated tissue. Staudt el al. [20] introduced 2,20 -thiodiethanol (TDE) as another hydrophilic mounting medium for cultured cells. Its refractive index of ~1.52 can be adjusted by dilution with water (n ¼ 1.33) in any proportion. Due to its very low penetration rate it cannot be used in large objects such as whole-mount preparations. While TDE can partly quench GFP fluorescence the fluorescence of some other markers such as Cy3 can even be amplified [20]. In a similar approach, Efimova and Anokhin [21] used a customized ~75% aqueous mixture of sodium diatrizoate and meglumine diatrizoate to clear entire mouse brains. Due to the high content of iodine (a heavy element) diatrizoic acid exhibits high X-ray opacity. Its salts are thus used as contrast agents for intestinal medical radiography and are commercially available, for example, under the trade names Hypaque 76% (Amersham Health, USA), and Gastrografin or Urografin (Bayer, Germany). These products contain meglumine diatrizoate and sodium diatrizoate in varying proportions, and all have a high refractive index of about 1.45. However, due to high polarity, their penetration through the cell membranes is very slow in nondehydrated tissues. Efimova and Anokhin [21] circumvented this problem by a prior dehydration of mouse brains in an ascending concentration row of 2-butoxy ethanol. While dehydration may accelerate the diffusion of diatrizoate solutions into the tissue by permeabilizing cell membranes, potential advantages of a water-based tissue-clearing medium such as the elimination of dehydration-induced shrinking artifacts and some gain of time do not seem to outweigh the drawbacks. Tsai et al. [22] cleared murine brain slices in 60% aqueous solutions of sucrose, after permeabilizing the cell membranes using the detergent Triton X-100. While the clearing effect may be sufficient in some applications, the increase of tissue transparency that can be achieved is quite modest, so that only thin samples such as brain slices can be processed. Although not tested yet we do not expect the described procedure to have any major effect on the fluorescence of GFP and other genetic markers. Recently, Hama et al. [23] presented “Scale,” an aqueous solution consisting of urea, glycerol and a small addition of the detergent Triton X-100, as a clearing medium. Hama et al. suggest different variations of this solution, named ScaleA2, ScaleU2, and ScaleB4. ScaleA2 contains 4M urea and 10% (w/v) glycerol and 0.1% (w/v) Triton X-100; ScaleU2 contains 4M urea, 30% (w/v) glycerol, and 0.1% (w/v) Triton X-100; and ScaleB4 contains 8M urea and 0.1% (w/v) Triton X-100. The pH of ScaleB4 was

188

Klaus Becker et al.

adjusted to 8.7 [23]. Presently, ScaleA2 is distributed by Olympus under the brand name “Scaleview-A2”, jointly with specialized long working distance (WD) immersion objectives (25/1.0 at WD 4 mm; 25/0.90 at WD 8 mm). These are corrected for the refractive index of ScaleA2 (~1.38). In contrast to most other clearing liquids causing tissue shrinkage and hardening, specimens cleared with Scale exhibit pronounced swelling, up to double the original volume. According to Hama et al. [23], their clearing procedure does not show any appreciable quenching effect on the fluorescence of EGFP and other fluorescent proteins (mAG1, YFP, DsRed, and mCherry) while it is inevitable at least to a certain degree when nonpolar solvents such as BABB or DBE are used. Samples cleared with ScaleA2 generally exhibit a soft and fragile texture so that they have to be handled very carefully; this problem can be solved by switching to one of the other two variants (ScaleU2 or ScaleB4) [23]. At 656 nm, the refractive index of ScaleA2 is 1.38, which is much lower compared to other tissue-clearing media designed to match the refractive index of proteins (~1.53). This implies that the clearing effect of Scaleview does not rely on refractive-index matching but presumably on structural modifications of proteins such as collagen [5] and the pronounced tissueswelling effect due to the absorption of water. As with all other aqueous solutions, clearing with Scale requires long incubation times, for example, about 2 weeks for E13.5 (13.5-day-old) mouse embryo. Generally, for ScaleA2 and ScaleU2 the authors report incubation times in the order of days/weeks and weeks/ months, respectively [23]. While Scale may sufficiently clear E13.5 mouse embryos and very young mouse brains below P14 (postnatal day 14), it does not clear highly myelinated regions. In our experiments, poor or even no clearing was obtained in large or fatty specimens, such as nervous tissue with all variations of Scale. This is not surprising as myelin layers consist of a fatty substance and can thus only be penetrated by lipophilic solvents. We failed to clear muscle samples of about 5–10 mm edge length by incubation in ScaleA2, ScaleU2 or ScaleB4 or for about 10 weeks [24]. After this time span the specimen clearly began to show signs of structural decomposition. At first, water-based tissue-clearing media appear promising as they do not require prior specimen dehydration, thus preventing possible shrinking artefacts. However, as long incubation times are needed there is generally no gain in the clearing rate compared to the procedures involving dehydration. To our knowledge, there is currently no alternative to nonpolar organic solvents when clearing heavily myelinated or fatty samples (e.g., entire mouse brains or human tumor biopsies of ~10 mm edge length). However, this situation may change owing to a novel clearing approach presented in 2013 by Chung et al. [25]. In their method referred to as “CLARITY” the specimen morphology is well

Chemical Clearing of GFP-Expressing Neural Tissues

189

preserved in a first step by perfusion with a mixture of acrylamide, bisacrylamide, and formaldehyde that can be thermally polymerized to an elastic hydrogel. By incubation at 37  C the monomers are further cross-linked to proteins, nucleic acids and some other small, low molecular weight cellular compounds, thus forming a hybrid, mechanically stable molecular mesh. Since lipids are not integrated into the hydrogel mesh, they can be extracted from the sample via a modified form of SDS-PAGE electrophoresis in a second step. The polymerized molecular mesh preserves the biomolecules and fine structural features such as membrane-localized proteins, synapses and spines, thus preventing their displacement during this extraction procedure. After lipid extraction the specimens are incubated for some time in an aqueous clearing medium such as FocusClear (CellExplorer Labs, Taiwan), or 80% glycerol to preserve their final transparency [25]. Chung et al. [25] have also achieved an excellent conformational preservation of functional proteins and a high permeability for aqueous solutions, making the cleared samples accessible to molecular phenotyping as well as whole-mount immunostaining, which is not possible with other clearing techniques described before. 1.4 Chemical Clearing of GFP-Expressing Tissues with Tetrahydrofuran and Dibenzyl Ether

Generation of GFP-expressing transgenic animals enables histological studies of neuronal networks with unprecedented accuracy [32]. Therefore, GFP became one of the most common genetic reporters in molecular biology, medicine and cell biology. However, GFP is of limited compatibility with common dehydration and tissue-clearing techniques. For example, ethanol and BABB as frequently used tissue-dehydration and clearing media considerably quench the fluorescence of GFP and related markers. This is a severe drawback which can turn clearing, staining and embedding of GFP-expressing material into a rather difficult task [33]. To overcome this problem we systematically screened potential dehydration and clearing media (Table 2) and demonstrated in GFP-labeled mouse brains and spinal cords that dehydration with tetrahydrofuran (THF) yields stronger fluorescence with a markedly reduced background [13, 24]. We also found that clearing with dibenzyl ether (DBE, Fig. 1B) better preserves fluorescence and achieves higher tissue transparency [24]. Combined, dehydration with THF and clearing with DBE is well suited to GFP-expressing specimens and is applicable to mouse spinal cords, entire mouse embryos, entire Drosophila and other specimens; mouse hippocampi are shown in Fig. 2. Since THF and DBE are chemically inert ethers, it may be speculated that it is this chemical quality that is responsible for their good GFP compatibility.

190

Klaus Becker et al.

Tissue shrinking is an unavoidable side-effect of any dehydration procedure and is slightly higher in THF than in ethanol. However, this shrinkage is the same (about 20%) in each direction (i.e., isotropic) for murine spinal cord [24] and the shrinking artifacts we observed were no greater than in ethanol. The contribution of the clearing medium (BABB or DBE) to the total tissue shrinkage is negligible [13, 24]. Clearing with DBE proceeds faster than with BABB. This is likely due to the lower viscosity of DBE translating to faster tissue penetration. Compared to the dynamic viscosities of benzyl alcohol and benzyl benzoate (6.4 cP and 10 cP, respectively) the viscosity of DBE (5.3 cP) is rather low. An additional benefit of DBE over BABB is its lower toxicity. Indeed, it is used as a stabilizer in perfumes and as a food-flavoring agent. In rats no toxic effects of a 196 mg/kg/day dose of DBE were found [35]. Furthermore, DBE (~$40/L) is considerably cheaper than BABB (BA ~$300/L, BB ~$100/L).

Table 2 Dehydration and clearing chemicals tested Dehydration Dehydration medium Methanol

H3C

OH

H3C

CH3

GFP fluorescence after clearing with BABBa Severely reduced Not detectable

Acetone O OH H3C

O

O

2-butoxyethanol

CH3

Severely reduced Dimethyl formamide

N H3C

CH3

O

Severely reduced S

H3C

Severely reduced

CH3

Dimethyl sulfoxide (DMSO)

H2O

O

Dioxane

Severely reduced

O

O

Tetrahydrofuran

Well preserved

Chemical Clearing of GFP-Expressing Neural Tissues

191

Clearing nb

Clearing medium O O

Clearing efficiency and GFP fluorescence after ethanol dehydration

1.54 Poor clearing, fluorescence strongly reduced CH3

Methyl salicylate

OH

OH

1.54 Poor clearing, fluorescence reduced

Benzyl alcohol

1.57 Poor clearing, fluorescence reduced O

Benzyl benzoate

O CH3

Trans-anethole

O H3C CH3 O

CH3

Isoeugenol

1.56 Good clearing and preservation of fluorescence; brownish discolorations 1.58 Good clearing and preservation of fluorescence; strong brownish discolorations

HO

O

CH3

1.57 Poor clearing, fluorescence strongly reduced

Isosafrole

O Br

Br

1,5-bromopentane

Br

1.56 Good clearing and preservation of fluorescence; significant toxicity and high volatility

Bromobenzene

O

1.51 No clearing at all

Dibenzyl ether

1.56 Good clearing and preservation of fluorescence

Most of the substances found to be GFP-incompatible possess hydroxyl or carbonyl groups. Adapted from Ref. [24] a Benzyl alcohol/benzyl benzoate b Refractive index (rounded to 0.01)

2

Preparation of Chemical Agents THF as well as DBE can become contaminated with peroxides over time due to contact with oxygen [24]. After long-term storage, DBE can additionally become contaminated with benzaldehyde [36] (Fig. 3). Since both of these substances quench GFP fluorescence even at low concentrations, THF and DBE should always be purified before use by column chromatography and stored in brown glass bottles containing a molecular sieve of 3 A˚ pore as a protection from water, and access of oxygen prevented by covering the liquid with argon as an inert gas.

192

Klaus Becker et al.

Fig. 1 Chemical structures of tetrahydrofuran and dibenzyl ether. Interestingly, both compounds are ethers lacking any functional groups 2.1 Tetrahydrofuran and Dichloromethane

Peroxide removal from THF was carried out by column absorption chromatography (Fig. 4A) with basic activated aluminum oxide, ca. 250 g/L) [24, 37]. However, chromatography also removes the stabilizer BHT normally present in commercially available THF. For safety reasons, generation of dangerous amounts of peroxides due to exposure to sunlight or oxygen must be prevented by replacing BHT after chromatography, for example, by adding 250 mg/L into the receiver flask (Fig. 4A-3) protected from light by an aluminum foil. Insufficiently stabilized THF can explode with fatal consequences! Peroxide concentrations were estimated using Quantofix test stripes. Chemicals l Tetrahydrofuran (THF, Sigma-Aldrich 186562). l

Aluminum oxide (basic-activated Brockmann I grade, SigmaAldrich 199443).

l

Butylhydroxytoluene (BHT, Sigma-Aldrich W218405).

l

Calcium chloride (Sigma-Aldrich C1016).

l

Dichloromethane (Carl Roth, Germany, 7334).

l

Quantofix Peroxide 25 test stripes (Sigma-Aldrich Z249254).

l

Argon gas.

Equipment l Brown-glass storage bottle. l

Dropping funnel with pressure compensation (Fig. 4A-1).

l

Chromatography column (Fig. 4A-2).

l

Two-necked round bottom flask (Fig. 4A-3) wrapped in aluminum foil.

l

Drying tube filled with calcium chloride (Fig. 4A-4).

l

Rubber and glass joints.

l

Silicon tubes (2).

Chemical Clearing of GFP-Expressing Neural Tissues

193

Fig. 2 Ultramicroscopy-based reconstructions of isolated hippocampi from thy-1 EGFP-M mouse. Dibenzylether (DBE) clearing yields stronger fluorescence and better visibility of details. (A1–A3) Left hippocampus dehydrated with tetrahydrofuran (THF) and cleared with benzyl alcohol/benzyl benzoate (BABB). (B1–B3) Right hippocampus (from the same mouse) dehydrated with THF and cleared with DBE. CA1 cornu ammonis region one, gl granular cell layer, py pyramidal cells. All images were acquired with an Olympus objective XLFluor 4/0.28 and an imaging setup described earlier [10, 13, 15] and in Chapter 11 by Saghafi [34]. Adapted from Ref. [24] 2.2

Dibenzyl Ether

Compared to THF, DBE’s viscosity and boiling point are higher, so the removal of its peroxides was carried out using a different setup (Fig. 4B). The filter funnel was filled with ~250 g of activated aluminum oxide per liter, and suction was applied to the receiver flask [24]. Peroxide concentrations were estimated using Quantofix test stripes. The presence of unwanted aldehydes (such as benzaldehyde, Fig. 3) and ketones was checked by applying Brady’s test

194

Klaus Becker et al.

Fig. 3 Chemical decomposition of dibenzyl ether (DBE) by oxygen. Oxygen can transform dibenzyl ether into an instable peroxide, which then further reacts, yielding benzaldehyde and benzyl alcohol

Fig. 4 Peroxide removal apparatus. (A) Peroxide removal from tetrahydrofuran (THF) and dichloromethane. (1) Dropping funnel with pressure compensation. (2) Chromatography column filled with basic-activated aluminum oxide. (3) Two-necked round-bottom flask filled with butylhydroxytoluene (BHT) as a stabilizer of THF. (4) Drying tube filled with calcium chloride. (B) Peroxide removal from dibenzyl ether and benzyl alcohol/ benzyl benzoate. (1) Filter unit with a filter plate (16–40 μm pore size). (2) Vacuum-tight filtering flask. Reprinted from Ref. [24]

[38]. Chemically, the amino group of 2,4-dinitrophenylhydrazine (Brady’s reagent) reacts with them by addition-elimination of carbonyl groups, yielding insoluble hydrazine manifesting itself as an orange precipitate.

Chemical Clearing of GFP-Expressing Neural Tissues

195

Chemicals l Aluminum oxide (basic-activated Brockmann I grade, SigmaAldrich 199443). l l

Dibenzyl ether (DBE, Sigma-Aldrich 108014). Molecular sieve (3 A˚ mesh, Sigma-Aldrich 208582).

l

Quantofix Peroxide 25 test stripes (Sigma-Aldrich Z249254).

l

Brady’s Reagent (2,4-dinitrophenylhydrazine, Sigma-Aldrich D199303; 4 g dissolved in 8 mL of concentrated sulfuric acid, 90 mL of methanol and 10 mL of water).

Equipment l Brown-glass storage bottle. l

2.3 Formaldehyde/ PBS

Bu¨chner funnel with filter plate of 16–40 μm pore width (Fig. 4B-1).

l

Vacuum-tight filtering flask (Fig. 4B-2).

l

Silicon tubes and joints.

l

Vacuum pump LABOPORT (KNF, USA).

For 1 liter of fixative 40 g of paraformaldehyde powder (a solid polymer of formaldehyde) was added and dissolved in phosphatebuffered saline (PBS) at 60  C while stirring; a slightly basic pH (NaOH) is required. Since GFP fluorescence is stabilized under basic conditions, pH was adjusted to 7.8. Formaldehyde solution (4%) was kept at 4  C and only stored for a few days. Please consult Chapter 2 by Mufson [39] for properties of paraformaldehyde. Chemicals l Deionized water.

3

l

Paraformaldehyde (Sigma-Aldrich 6148).

l

PBS 10 mM (Dulbecco, Biochrom AG, Germany L182).

l

Sodium hydroxide (NaOH).

Perfusion of Mice and Dissection of Neural Tissues 14 days old thy-1 EGFP-M (c57/bl6) mice [40] were killed by asphyxiation in CO2, transcardially perfused post mortem with at least 20 mL of PBS (pH 7.8) containing 10 Units/mL of heparin until blood was removed, and perfusion-fixed by 100 mL of 4% formaldehyde in PBS (pH 7.8), as follows: The mouse (with its head pointing away from the experimentalist) was pinned ventral-

196

Klaus Becker et al.

side-up onto the dissection plate and its ventral part moistened with ethanol. Its abdominal wall and chest were opened, a perfusion cannula carefully inserted into the left heart ventricle and fixed with a microvascular clamp. Finally, the right heart auricle was opened with small scissors. The flexible tube of the cannula was clamped in a peristaltic pump. The mouse should only be perfused slowly (1.5–2.0 mL/min) with PBS and formaldehyde (both on ice). No major blood vessels, lung or liver should be hurt during the dissection. After perfusion the brain was removed from the skull and postfixed in 4% formaldehyde/PBS (pH 7.8 at 4  C) overnight. Next day, the specimens were rinsed 3 in PBS (at room temperature, 30 min each) and immediately dehydrated. Animal care and euthanasia were carried out in accordance with the local ethic guidelines and animal protection laws. Hippocampi were dissected from the brain by splitting it into hemispheres, detached by dilation of the flanking ventricles using Dumont #5 forceps, and post-fixed overnight in 4% formaldehyde/ PBS (pH 7.8), rinsed 3 in PBS (15 min each), and immediately dehydrated. Spinal cord was dissected by removing the viscera and cutting the vertebral column, for example, above the lumbar vertebrae by using appropriate scissors. The vertebral column was carefully cut out and the spinal cord taken out by cutting the spinal nerves using Dumont #5 forceps and spring scissors. Spinal cords were processed in the same way as hippocampi. Chemicals/Equipment l Heparin (Sigma-Aldrich, Germany).

4

l

Carbon dioxide supply and an asphyxiation chamber.

l

Butterfly cannula (Nipro 21G, with a flexible tube).

l

Microvascular clamp (S&T, Switzerland, P-3, 70007).

l

Peristaltic pump (Ismatec ISM796B).

l

Spring scissors (Fine Science Tools, 15006-09, 10 mm bladed angled side).

l

Standard dissection equipment (small scissors for opening the mouse, larger scissors for decapitation, dissection plate with pins, forceps for handling, appropriate spatula for removing the brain from the skull).

Tissue Dehydration and Clearing Dehydration and clearing of the dissected neural tissues was carried out at room temperature under an extraction hood in 15 mL glass

Chemical Clearing of GFP-Expressing Neural Tissues

4.1

4.2

Dehydration

Clearing

197

vials filled with ca. 10 mL of dehydration/clearing solution and fitted with a tight lid. A rotary mixer was employed to facilitate penetration of the solution into the samples. The fixed brains were dehydrated as a whole or split into hemispheres. To minimize shrinkage of brain, hippocampi, and spinal cord, water was removed by incubating in an ascending concentration row of water-diluted or pure peroxide-free THF, as outlined below. For spinal cord, we additionally recommend a 1-h incubation step in dichloromethane after dehydration, to remove myelin and facilitate subsequent clearing. l

Brain: 50, 70, 80, 96 vol.%, finally 3 pure agent (100%), 12 h/ step.

l

Hippocampi: 50, 80, 96 vol.%, finally 3 100%, 1 h/step, last step overnight.

l

Spinal cord: 50, 80, 96 vol.%, finally 3 100%, 1 h/step, last step overnight.

Clearing of the dehydrated specimens was carried out in peroxidefree DBE until becoming transparent, as outlined below: l

Brain: 1–2 days, three exchanges of clearing medium.

l

Hippocampi: ca. 1 day, two exchanges of clearing medium.

l

Spinal cord: ca. 1 day, one exchange of clearing medium.

References € 1. Spalteholz W (1911) Uber das Durchsichtigmachen von menschlichen und tierischen Pr€aparaten. S. Hirzel, Leipzig. http://s2w. hbz-nrw.de/zbmed/content/titleinfo/ 555354 € 2. Lundvall H (1904) Uber demonstration embryonaler Knorpelskelette. Anatomischer Anzeiger (Jena) 25:219–222. BHL: 11889861. https://biodiversitylibrary.org/ page/11889861 3. Lundvall H (1905) Weiteres u¨ber demonstration embryonaler Skelette. Anatomischer Anzeiger (Jena) 27:520–523. BHL: 11893322. https://biodiversitylibrary.org/ page/11893322 4. Colaruso P (2020) The properties of light governing biological microscopy. In: Pelc R, Walz W, Doucette JR (eds) Neurohistology and imaging techniques, Neuromethods, vol 153. Springer (Humana Press), New York, pp

201–223. https://doi.org/10.1007/978-10716-0428-1_7 5. Rylander G, Stumpp OF, Milner TE, Kemp NJ, Mendenhall JM, Diller KR, Welch AJ (2006) Dehydration mechanism of optical clearing in tissue. J Biomed Opt 11 (4):041117. https://doi.org/10.1117/1. 2343208 6. Tuchin VV (2005) Optical clearing of tissues and blood. SPIE Press, Bellingham (WA, USA). ISBN: 9780819460066 7. Keller PJ, Dodt H-U (2011) Light sheet microscopy of living or cleared specimens. Curr Opin Neurobiol 22(1):1–6. https://doi. org/10.1016/j.conb.2011.08.003 8. Sharpe J, Ahlgreen U, Perry P, Hill B, Ross A, Hecksher-Sorensen J, Baldock R, Davidson D (2002) Optical projection tomography as a tool for 3D microscopy and gene expression

198

Klaus Becker et al.

studies. Science 296(5567):541–545. https:// doi.org/10.1126/science.1068206 9. Oldham M, Sakhalkar H, Oliver T, Allan JG, Dewhirst M (2008) Optical clearing of unsectioned specimens for three-dimensional imaging via optical transmission and emission tomography. J Biomed Opt 13(2):021113. https://doi.org/10.1117/1.2907968 10. Dodt H-U, Leischner U, Schierloh A, J€ahrling N, Mauch CP, Deininger K, Deussing JM, Eder M, Zieglg€ansberger W, Becker K (2007) Ultramicroscopy: three-dimensional visualization of neuronal networks in the whole mouse brain. Nat Methods 4 (4):331–336. https://doi.org/10.1038/ nmeth1036 11. Huisken J, Swoger J, Del Bene F, Wittbrodt J, Stelzer EH (2004) Optical sectioning deep inside live embryos by selective plane illumination microscopy. Science 305 (5686):1007–1009. https://doi.org/10. 1126/science.1100035 12. Voie AH, Burns DH, Spelman FA (1993) Orthogonal-plane fluorescence optical sectioning: three-dimensional imaging of macroscopic biological specimens. J Microsc 170 (Pt3):229–236. https://doi.org/10.1111/j. 1365-2818.1993.tb03346.x 13. Ertu¨rk A, Becker K, J€ahrling N, Mauch CP, Hojer CD, Egen JG, Hellal F, Bradke F, Sheng M, Dodt H-U (2012) Threedimensional imaging of solvent-cleared organs using 3DISCO. Nat Protoc 7(11):1983–1995. https://doi.org/10.1038/nprot.2012.119 14. J€ahrling N, Becker K, Scho¨nbauer C, Schnorrer F, Dodt H-U (2010) Threedimensional reconstruction and segmentation of intact Drosophila by ultramicroscopy. Front Syst Neurosci 4(1):1–6. https://doi.org/10. 3389/neuro.06.001.2010 15. Becker K, J€ahrling N, Kramer ER, Schnorrer F, Dodt HU (2008) Ultramicroscopy: 3D reconstruction of large microscopical specimens. J Biophotonics 1(1):36–42. https://doi.org/ 10.1002/jbio.200710011 16. Sakhalkar HS, Dewhirst M, Oliver T, Cao Y, Oldham M (2007) Functional imaging in bulk tissue specimens using optical emission tomography: fluorescence preservation during optical clearing. Phys Med Biol 52(8):2035–2054. https://doi.org/10.1088/0031-9155/52/ 8/001 17. Drahn F (1922) Ein neues Durchtr€ankungsmittel fu¨r histologische und anatomische Objekte. Berl Tierarztl Wochenschr 38:97–100 18. Dent JA, Polson AG, Klymkowsky MW (1989) A whole-mount immunocytochemical analysis

of the expression of the intermediate filament protein vimentin in Xenopus. Development 105(1):61–74. PMID: 2806118. https:// www.ncbi.nlm.nih.gov/pubmed/2806118 19. Chiang AS (2002) Aqueous tissue clearing solution. Patent US6472216B1. https:// patents.google.com/patent/US6472216B1 20. Staudt T, Lang MC, Medda R, Engelhardt J, Hell S (2007) 2,2’-Thiodiethanol: a new water soluble mounting medium for high resolution optical microscopy. Microsc Res Tech 70 (1):1–9. https://doi.org/10.1002/jemt. 20396 21. Efimova OI, Anokhin KV (2009) Enhancement of optical transmission capacity of isolated structures in the brain of mature mice. Bull Exp Biol Med 147(1):3–6. https:// doi.org/10.1007/s10517-009-0463-9 22. Tsai PS, Kaufhold JP, Blinder P, Friedman B, Drew PJ, Karten HJ, Lyden PD, Kleinfeld D (2009) Correlations of neuronal and microvascular densities in murine cortex revealed by direct counting and colocalization of nuclei and vessels. J Neurosci 29(46):14553–14570. https://doi.org/10.1523/JNEUROSCI. 3287-09.2009 23. Hama H, Kurokawa H, Kawano H, Ando R, Shimogori T, Noda H, Fukami K, SakaueSawano A, Miyawaki A (2011) Scale: a chemical approach for fluorescence imaging and reconstruction of transparent mouse brain. Nat Neurosci 14(11):1481–1488. https:// doi.org/10.1038/nn.2928 24. Becker K, J€ahrling N, Saghafi S, Weiler R, Dodt H-U (2012) Chemical clearing and dehydration of GFP expressing mouse brains. PLoS One 7(3):e33916. https://doi.org/10.1371/ journal.pone.0033916 25. Chung K, Wallace J, Kim SY, Kalyanasundaram S, Andalman AS, Davidson TJ, Mirzabekov JJ, Zalocusky KA, Mattis J, Denisin AK, Pak S, Bernstein H, Ramakrishnan C, Grosenick L, Deisseroth K (2013) Structural and molecular interrogation of intact biological systems. Nature 497 (7449):332–337. https://doi.org/10.1038/ nature12107 26. Klymkowsky MW, Hanken J (1991) Wholemount staining of Xenopus and other vertebrates. In: Kay BK, Peng HB (eds) Xenopus laevis: practical uses in cell and molecular biology, Meth Cell Biol, vol 36. Academic Press, New York, pp 419–441. https://doi.org/10. 1016/S0091-679X(08)60290-3 27. Skinner RA (1986) The value of methyl salicylate as a clearing agent. J Histotechnol 9 (1):27–28. https://doi.org/10.1179/his. 1986.9.1.27

Chemical Clearing of GFP-Expressing Neural Tissues 28. Chiang AS, Liu YC, Chiu Sl HSH, Huang CY, Hsieh CH (2001) Three-dimensional mapping of brain neuropils in the cockroach Diploptera punctata. J Comp Neurol 440(1):1–11. https://doi.org/10.1002/cne.1365 29. Fu YY, Lin CW, Enikolopov G, Sibley E, Chiang AS, Tang SC (2009) Microtome-free 3-dimensional confocal imaging method for visualization of mouse intenstine with subcellular-level resolution. Gastroenterologia 137(2):453–465. https://doi.org/10.1053/j. gastro.2009.05.008 30. Liu YC, Chiang AS (2003) High-resolution confocal imaging and three-dimensional rendering. Methods 30(1):86–93. https://doi. org/10.1016/S1046-2023(03)00010-0 31. Hara M, Noiseux N, Bindokas V (2005) High resolution optical imaging of infarction in intact organs. BioTechniques 39(3):373–376. PMID: 16206909. https://www.ncbi.nlm.nih. gov/pubmed/16206909 32. Tsien RY (1998) The green fluorescent protein. Annu Rev Biochem 67:509–544. https://doi.org/10.1146/annurev.biochem. 67.1.509 33. Ward W (2005) Biochemical and physical properties of green fluorescent protein. In: Chalfie M, Kain SR (eds) Green fluorescent protein, 2nd edn. Wiley Interscience, New York, pp 39–65 34. Saghafi S, Becker K, J€ahrling N, Hahn C, Dodt H-U (2020) Ultramicroscopy of nerve fibers and neurons: fine-tuning the light sheets. In: Pelc R, Walz W, Doucette JR (eds) Neurohistology and imaging techniques,

199

Neuromethods, vol 153. Springer (Humana Press), New York, pp 325–339. https://doi. org/10.1007/978-1-0716-0428-1_11 35. Burdock GA, Ford RA (1992) Safety evaluation of dibenzyl ether. Food Chem Toxicol 30 (7):559–566. https://doi.org/10.1016/ 0278-6915(92)90189-R 36. Eichel FG, Othmer DF (1949) Bezaldehyde by autoxidation of dibenzyl ether. Ind Eng Chem 41(11):2623–2626. https://doi.org/10. 1021/ie50479a054 37. Dasler W, Bauer CD (1946) Removal of peroxides from organic solvents. Ind Eng Chem Anal Ed 18(1):52–54. (incorporated into Anal Chem). https://doi.org/10.1021/ i560149a017 38. Brady OL, Elsmie GV (1926) The use of 2,4-dinitrophenylhydrazine as a reagent for aldehydes and ketones. Analyst 51:77–78. https://doi.org/10.1039/AN9265100077 39. Mufson EJ, Perez SE, Kelley CM, Alldred MJ, Ginsberg SD (2020) Fixation protocols for neurohistology: neurons to genes. In: Pelc R, Walz W, Doucette JR (eds) Neurohistology and imaging techniques, Neuromethods, vol 153. Springer (Humana Press), New York, pp 49–71. https://doi.org/10.1007/978-10716-0428-1_2 40. Feng G, Mellor RH, Bernstein M, KellerPeck C, Nguyen QT, Wallace M, Nerbonne JM, Lichtman JW, Sanes JR (2000) Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP. Neuron 28 (1):41–51. https://doi.org/10.1016/S08966273(00)00084-2

Chapter 7 The Properties of Light Governing Biological Microscopy Pina Colarusso Abstract Just as the telescope allowed us to view the heavens, the microscope has opened vistas within the microcosm of the natural world. Here we review the properties of light and its interaction with matter that govern image formation in optical microscopy. We also offer a glimpse into the history of the understanding of light and demonstrate how different models are applied to explain its behavior. We lay the foundation by defining key physical characteristics of light, and then move onto describe phenomena associated with light that are important for biological applications of optical (light) microscopy. Reflection and refraction are explained using a simplified framework known as geometrical optics. The more general wave representation is then used to describe interference and diffraction, the phenomena that arise when light interacts with itself either as it propagates freely or after it meets a macroscopic object such as an aperture or edge. Polarization is also defined using the wave model of light. We also cover the relevance of scattering absorption, and dispersion in optical microscopy. We conclude with a short account of the relationship between the wavelength of light and the perception of color by the human eye. Key words Absorption, Diffraction, Dispersion, Interference, Polarization, Refraction, Scattering

1

Introduction Optical microscopy harnesses the interactions of visible electromagnetic radiation (400 and 700 nm) and matter to reveal objects or features that are invisible to the unaided eye. We know this narrow sliver of the electromagnetic spectrum as “light,” and we will use both “optical” and “light” microscopy interchangeably here. Optical microscopes form magnified images from one or more lenses. The lenses used in standard light microscopes are based on the optical phenomenon of refraction, the bending of light at interfaces, to provide enlarged images of the specimen. Refraction also occurs in the specimen itself, and other light–matter interactions such as reflection, absorption, scattering, and interference contribute to the final appearance of the image. For example, the absorption of light creates vivid contrast in an optical microscopy image of a histological specimen stained with hematoxylin and eosin, while the interference of light leads to the topographic appearance of cells

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_7, © Springer Science+Business Media LLC 2020

201

202

Pina Colarusso

viewed with differential interference contrast (DIC) microscopy. The image formed by an optical microscope thus can be understood as the physical evidence of the many light–matter interactions underlying this technique. The aim of this chapter is to review the fundamental properties of light as well as the key light–matter interactions that occur when visualizing samples using optical microscopes. It is directed to the biologist without a specialized background in optics. Of course, it is possible to use an optical microscope with little knowledge or appreciation of the properties of light. Yet insights grounded in the properties of light and light–matter interactions have sparked a number of powerful approaches to visualize the microscopic world including now-standard techniques such as phase contrast [1] or more recent innovations such as multiphoton [2] or super-resolution [3] microscopy. A solid grounding in the basic properties of light is an advantage when troubleshooting problems with optical microscopes, making it a quick and systematic process that facilitates research. In addition, a working knowledge of the physics underlying image formation creates common ground between biologists and physicists, creating freer exchange of knowledge and expertise that is critical for further advances in optical microscopy. The nature of light remains elusive. What is light? How can we define it? As early as the tenth century, Hasan Ibn al-Haytham (Alhazen) proposed that light was composed of particles that travel in straight lines or rays [4]. Newton also postulated that light was composed of particles, and this view remained dominant until the nineteenth century. The pendulum swung toward a wave interpretation by the early 1800s with the experiments carried out independently by Young and Fresnel. By the 1860s Maxwell developed a comprehensive theory that united electricity and magnetism; the theory is laid out in an elegant set of equations known today as “Maxwell’s equations” [5]. These equations showed unequivocally that light is a wave and provided a coherent framework that could account for empirical results that had been gathered over centuries. The acceptance of Maxwell’s equations into the scientific canon was solidified by further experimental evidence, and the wave model of electromagnetic radiation remained ascendant until the twentieth century. Maxwell’s equations make several important predictions about the nature of light. A central tenet is that light is composed of oscillating electromagnetic waves travelling at a constant speed (in an environment of constant/homogeneous refractive index). We know this speed as a fundamental physical constant, the “c, the speed of light in vacuum.” Maxwell’s equations also predict the existence of electromagnetic waves outside the visible region; an early experimental justification of their validity was provided by Hertz in 1888, when he discovered radio waves and showed that they could propagate in free space at the speed of light [6]. Another

Properties of Light

203

important result was that electromagnetic waves can exert a force on, and induce motion of, electrically charged particles. Charged particles in turn can also generate electromagnetic waves when they move with acceleration or deceleration (e.g., when they are forced by magnetic field to move on a curved path in a synchrotron). Taken as a whole, Maxwell’s equations and the supporting experimental observations of the nineteenth century form the foundation of what is known as classical electromagnetism [7]. Note that the terms “light” and “electromagnetic radiation” and “electromagnetic waves” are used interchangeably in this chapter. In the early twentieth century, it became clear that Maxwell’s equations could not account for some observations. Puzzling phenomena such as the Balmer series, the emission of light at discrete wavelengths from hydrogen, blackbody radiation, and the photoelectric effect could not be explained using classical electromagnetic theory. The gaps in the classical model sparked a revolution in the understanding of light by the renowned physicists such as Niels Bohr, Max Planck, Albert Einstein, Erwin Schro¨dinger, and Louis de Broglie. Quantum mechanics was able to reconcile diverse puzzling phenomena by treating electromagnetic radiation as quantized; the basic unit of quantization became known as a photon. Inspired by the idea that light is quantized, de Broglie proposed that matter particles also could behave like waves. He predicted that a moving particle can exhibit wave-like behavior; however, the wave-like properties of matter are observable only when the particle is exceedingly small or cold. Electron microscopy is a practical application of the wave nature of electrons. The remarkable work by de Broglie and his contemporaries led to a theory that blurs the distinction between matter and energy, and predicts that both electromagnetic radiation (including visible light) and particles are both waves and particles. Quantum mechanics, in short, reconciled a light–particle dichotomy that had been a source of controversy for hundreds of years. Nonetheless quantum mechanics places an important restriction on the dual wave-particle nature: one can design an experiment to observe light or matter behaving entirely as a particle or as a wave, but it is not possible to observe both manifestations at the same time. This central tenet in quantum mechanics is known as the complementarity principle. Currently, the most rigorous theoretical framework for describing the nature of light and its interactions with matter is provided by quantum electrodynamics (QED). The curious reader is directed to the popular account of QED by Richard Feynman [8]. This chapter follows standard practice and uses the wave and the particle properties of light to explain the different phenomena underlying optical microscopy. First, a simplified version of the wave representation known as geometrical optics is used to describe

204

Pina Colarusso

reflection and refraction, which are the basis of the interactions of light with optical devices such as mirrors and lenses. A more general form of the wave representation is used to describe interference and diffraction, which arise when light interacts with itself either as it propagates freely in a homogeneous medium or after it impinges on a macroscopic object such as an aperture or edge. Polarization is also defined using the wave nature of light. Similarly, the wave representation will provide the context for detailing the most common types of scattering of light by objects. The particle nature of light will be invoked when describing light absorption (due to the excitation of electrons in molecules) and its relaxation mechanisms at the level of the specimen.

2

Basic Concepts Electromagnetic radiation is composed of electric and magnetic waves that oscillate perpendicular to each other; they are known as transverse waves because the direction of vibration is perpendicular to their direction of propagation (Fig. 1A). These perpendicular waves are known as the electric and magnetic fields. Electromagnetic radiation in a perfect vacuum travels at the speed of light c, a fundamental physical constant that is defined to be 299,792,458 m/s [9]. Monochromatic light consists of a single frequency f and is related to its wavelength λ by the following equation: c f ¼ ð1Þ λ Electromagnetic waves are classified according to their frequency f, and span the range from radio waves (lowest frequency f, longest wavelength λ) to gamma rays (highest frequency f, shortest wavelength λ). A schematic overview of the electromagnetic spectrum is shown in Fig. 1B. Visible light comprises a narrow slice of the electromagnetic spectrum that can be sensed by the human eye, ranging approximately from 400 nm (violet) to 700 nm (red). The wavelength decreases as the frequency increases, moving from red to violet. As mentioned previously, the electromagnetic wave contains a magnetic field oscillating perpendicular to the electric component. However, the magnetic forces involved in optical microscopy are much smaller than the electrical forces. Most effects observed in optical microscopy can be explained by considering the electric field alone, and thus the magnetic field will not be considered any further [10]. Classical electromagnetism cannot account for some properties of light; in fact, light–matter interactions such as absorption or fluorescence require describing light as being composed of

Properties of Light

B

α

E0

z

205

E

-E0

A

B

Airy disc diameter

Radio

Micro

IR

Visible

UV

X-rays

γ-rays

103

10-2

10-5

10-6

10-8

10-10

10-12

Wavelength (metres)

A B

Resolution

C D

Fig. 1 Wave nature of light. (A) An idealized electromagnetic wave. The electric field E and magnetic field B oscillate perpendicular to each other and propagate in a direction (z) that is perpendicular to the planes of vibration. For clarity, only one of the two components is considered in other diagrams (Figs. 1B, C, 2D, and 3). (B) An overview of the electromagnetic spectrum. Note that as the frequency increases, the energy increases, and the wavelength decreases. (C) A pictorial representation of two idealized monochromatic waves (A and B) described by Eq. 11. E0 is the amplitude of the wave, and α ¼ 0 for wave A. (D) The two-dimensional Airy diffraction pattern that arises from the diffraction through a circular aperture, or by microscopic imaging of a single bright spot on dark background (e.g., a fluorescently labeled biomolecule). Resolution marked in the image is according to the Rayleigh criterion

photons. The energy of a photon is directly proportional to its frequency and inversely proportional to its wavelength: c ð2Þ E ¼ hf ¼ h λ where h ¼ 6.626015  1034 J s is another fundamental physical constant known as Planck’s constant. Visible light, like all electromagnetic radiation, slows down when it propagates in a medium other than a perfect vacuum. The index of refraction of a medium is denoted by n and is defined as the ratio of the speed of light in a vacuum to the speed of light in the material c ð3Þ n¼ v

206

Pina Colarusso

where n and ν are defined for a specific wavelength. The refractive index n of a material is often reported at a wavelength of 589.3 nm, the central wavelength D between the D1 and D2 doublet lines in the atomic emission spectrum of sodium. The refractive index of a substance at this wavelength is designated nD. Immersion oils used on microscope objectives are often labeled with the refractive index measured at D. Recent advances in optical technology have shown that it is possible to fabricate materials with negative refractive indices [11, 12]. Known as “meta-materials,” the development is significant because the refractive index of a material greatly influences the resolving power (or resolution) of an optical microscope. If materials with negative refractive indices live up to their promise, they will revolutionize optical microscopy.

3

Reflection and Refraction Thus far light has been treated as travelling in a straight line in a vacuum or uniform medium. In standard biological optical microscopy, light does not propagate in a homogeneous medium, but instead interacts with the optical components such as lenses, filters, and mirrors as well as the biological specimen itself. That is, light travels in straight lines unless it encounters a boundary between two or more media. The precise nature of the boundary defines the subsequent light–matter interactions such as reflection, refraction, absorption, interference, and scattering. When light encounters an object such as a mirror or lens that has dimensions much greater than its wavelength, its path can be modeled using geometrical optics. In this model, the propagation of light is simplified by representing it as a ray pointing in the direction of travel. The distance traveled by light in a material also is often compared to the distance that would be traveled by light in a vacuum. In any given medium, the length of the light ray’s path, d, is given by d ¼ vt

ð4Þ

where v is the velocity and t is the time traveled. We have seen in Eq. 3 that v ¼ c/n, so it follows that c d¼ t ð5Þ n nd ¼ ct

ð6Þ

The quantity nd has units of length and is known as the optical path. It represents the distance light would have traveled in a vacuum in the time that it takes to travel a distance d in the medium (light travels faster and thus further in a vacuum than in standard materials). The optical path is a concept that recurs in many areas of

Properties of Light

207

optical microscopy and is applied to explain image formation using techniques such as phase-contrast (described in Chapter 10 by Pelc et al.) and differential interference contrast microscopy. Geometrical optics is often used to describe the reflection and/or refraction that occurs when light encounters an interface much larger than its wavelength. First let us consider reflection which occurs when light encounters a highly polished surface such as a mirror. Reflection can be depicted as two rays, representing the incident and reflected ray, respectively. A line drawn perpendicular to the reflecting interface from the point of incidence is known as the normal. The plane of incidence is defined as the plane containing the incident and reflected rays as well as the normal. The Law of Reflection states that the angle of incidence ϕi equals the angle of reflection ϕr (both are measured with respect to the normal): ϕi ¼ ϕr

ð7Þ

Whenever light travels from one medium to another, it will undergo both reflection and refraction if the refractive index differs between the two media. To illustrate, consider a slab of glass surrounded by air, where nair ¼ 1 and nglass ¼ 1.5. From the everyday experience of looking through a window during the daytime, we would expect most of the light to travel through the glass. However, a small fraction of the light is reflected at both air-glass interfaces. In fact, at normal (perpendicular) incidence, glass reflects about 4% of the incident light at each interface. As light travels through a slab of glass, 4% of the incident light will be reflected from the first surface, while 4% of the light (i.e., 0.04  0.96) will be lost from the second interface as the light exits the glass. This type of a back-of-the-envelope calculation is often used to calculate the efficiency of optical instruments such as the light microscope. Analytically, reflection is described by Fresnel’s equations [13]. Because of reflection, the more optical elements that are used in a microscope, the less efficient the transmission through the microscope. Despite the light loss, note that additional optical elements are often included in microscopes to provide benefits that outweigh the losses in transmitted light. These include features such as wavelength selection and improved image quality. Furthermore, the losses from microscope optics are much less than from plain glass because these specialized devices are treated with antireflection coatings (discussed in Subheading 4) to maximize the efficiency of the required transmission or reflection. Light can also undergo refraction at a boundary (Fig. 2A). Indeed, the dependence of the speed of light on the medium means that light will bend or refract when it encounters an interface between two materials with different refractive indices. As with

208

Pina Colarusso

n2

1

2

θ2 1 θ1

A n1 > n2

n1

2

1+2

1+2

n1 > n2 nA

D

nB

Core (n1) Cladding (n2)

B

C

Total internal reflection dA = thickness

dB = thickness

Fig. 2 Refraction, reflection and interference. (A) Snell’s Law of Refraction. Dotted line represents total internal reflection. (B) Total internal reflection governing propagation of light in an optical fiber, thus lending it lightguiding capability. (C) Multilayer reflector diagram denoting refractive indices (nA, nB). Perpendicularly to the plates, the strongest reflection (constructive interference) occurs for λ ¼ 4nAdA ¼ 4nBdB. (D) Constructive and destructive interference of light waves 1 and 2. Drawings A-D were provided by Radek Pelc

reflection, the extent of refraction is measured as the angle between the direction of its propagation and the normal to the point of incidence. Light will bend toward the normal if the refractive index increases at the interface, while it will bend away from the normal if the refractive index decreases. This optical property is encapsulated by Snell’s law, in which the angle of incidence θ1 is related to the angle of refraction θ2 by n1 sin θ1 ¼ n2 sin θ2 ,

ð8Þ

where n1 and n2 are the refractive indices in the first and second medium, respectively. The interested reader is directed toward introductory books on optics for the derivations of the law of reflection and Snell’s law from Fermat’s principle [10, 13]. Snell’s law also leads to a significant property that forms the basis of several optical devices: Once the incident ray reaches a critical angle, the ray is no longer refracted but instead is reflected at the boundary. This process is known as total internal reflection (Fig. 2A, B) and can occur when the index of refraction for the first medium n1 exceeds the index of refraction of the second medium, n2. In this case, the angle of refraction is larger than the angle of  incidence and thus will reach 90 before the angle of incidence reaches this value. The minimum angle of incidence leading to

Properties of Light

209

total internal reflection is known as the critical angle θc. Snell’s law can be rearranged so that n1 sin θc ¼ n2 sin 90∘ ¼ n2   n θc ¼ arcsin 2 n1

ð9Þ ð10Þ

The reflected ray forms a thin evanescent wave at the boundary between the two media that dies off exponentially as the distance from the boundary increases. This phenomenon has been exploited for the transmission of light through fiber optics and is the basis of total internal reflection fluorescence microscopy (TIRF) [14, 15]. Total internal reflection is also essential for light-guiding capability of optical fibers and cables (Fig. 2B). Light can bend gradually in an environment with a continuously changing refractive index, which can reduce the effect of dispersion. Examples include gradient-refractive-index (GRIN) lenses (discussed in Chapter 1 by Wouterlood and Langer) and some optical fibers.

4

Interference and Diffraction Thus far, we have simplified the propagation of light by using geometrical optics. Although rays are an adequate representation of the interactions of light through optical devices such as mirrors and lenses, they cannot be applied to other phenomena such as interference and diffraction. Instead these phenomena are treated using the wave model of electromagnetic radiation. As a starting point, let us review the basic properties of an ideal propagating electromagnetic wave, which is considered to be infinite, monochromatic, and vibrating in a single plane. Such a wave can be described by a periodically oscillating function, as shown below:   ! ! 2π E ðz, t Þ ¼ ξ E 0 cos z  2πf t  α , ð11Þ λ where E0 is the amplitude, z is the position in the direction of propagation, λ is the wavelength, t is time, f is the frequency (2πf ! is often denoted ω), α is the phase, and ξ is the unit electric field vector (whose direction is known as the polarization, discussed in Subheading 5). Figure 1C is a pictorial representation of Eq. 11 at one time point for two different values of α (0 for waveform A). The phase α is seen as the offset from a reference point at z ¼ 0 expressed as a fraction of the wave’s cycle (2π total). Another important property is the intensity of a wave, which is proportional to the square of the absolute value of the amplitude, |E0|2 or simply E02. We perceive the intensity as brightness, and it is averaged out over time (i.e., it is smaller than E02) as the oscillations occur at frequencies that are too

210

Pina Colarusso

high to be detected by the human eye. The phase α also cannot be directly perceived by the human eye. When two or more waves overlap, a new wave is formed that is the algebraic sum of the individual waves. The superposition of waves leads to interference, a phenomenon that cannot be described using the ray model of light. Interference is related to the phase α associated with the propagation of a wave (Eq. 11 and Fig. 1C). As shown in Fig. 2D, when two waves of identical frequency have phases that are identical or offset by integer multiples of 2π, they are referred to as being “in phase” and the resulting wave has the same frequency but double the amplitude. This type of wave superposition is known as constructive interference. If the waves have phases that differ by odd multiples of π, the resulting wave is a flat line; this case corresponds to destructive interference. Of course, these two cases are the simplest possible ones, and the waveforms 1 and 2 in Fig. 2D are slightly offset from being exactly in phase and out of phase. More generally, light is polychromatic and has finite duration. In this case, the interference patterns that result are more complex, but general principles used in their analysis are the same. The interference of light is the foundation for various techniques and devices used in optical microscopy. Usually, interference is produced by recombining light that has traveled through different optical paths. Recall that the optical path is defined as nd (Eq. 6). Imagine that a monochromatic beam of light is split into two, after which each component beam travels through a different optical path. The two component beams will then experience a phase shift relative to each other. If the two beams are then recombined, constructive or destructive interference will occur if the light waves are out of phase by multiples of 2π or odd multiples of π, respectively. Consider a thin optical coating placed on a reflective surface such as glass. The light will reflect from the top of the coating (primary beam) as well as from the bottom of the coating at the glass interface (secondary beam). The coating thickness can be engineered to introduce phase shift of π between the primary and secondary beams, causing destructive interference. This condition leads to minimized reflections from the surface of the material, and indeed is the basis for many types of antireflection (multi)coatings used on optical devices such as eyeglasses. Multireflector plates (Fig. 2C) found in some biological tissues function similarly, and can give rise to iridescence (dependence of detected color on the viewing angle). Interference can be used to visualize thin and transparent biological specimens that are difficult to visualize using standard brightfield microscopy. When imaged with an optical microscope, cells and tissues introduce phase shifts in the incident light waves because they are composed of organelles, extracellular matrix, and

Properties of Light

211

other components with varying refractive indices and thicknesses. As mentioned earlier, these phase differences are not apparent to the human eye, or to standard detectors such as cameras. For example, an epithelial monolayer immersed in a standard buffer is almost invisible under standard bright-field illumination. However, if the light traveling through different optical paths is forced to interfere constructively or destructively, the monolayer becomes visible because the resulting image contains regions of bright and dark. In this way, phase differences are converted to intensity variations that are discernible to the human eye or to a standard detector. Interference lies at the heart of the phase contrast microscope (described in detail in Chapter 10 by Pelc et al.), and in interference reflection and fluorescence interference contrast microscopy. Interference is also the basis of differential interference contrast microscopy and the advances in super-resolution microscopy [16]. Diffraction is another phenomenon related to interference that can be described using the wave model of light. Diffraction occurs when light encounters an obstacle or inhomogeneity as it travels. Light will travel in a straight line in a homogeneous medium; if it encounters an aperture or an edge, however, it will bend and interfere with itself. You can try a simple experiment by taking two opaque pencils or pens and holding them up to a light fixture (it helps if the pencils or pens are black). Hold the pencils so that they are parallel to each other and a few millimeters apart; close one eye, and focus on the gap as you move the pencils away from your face. Dark bands or fringes should be visible in the gap between the pencils. This simple experiment demonstrates that light does not travel in straight lines, but instead scatters off the edges of the pencils and interferes with itself. It is possible to model the various diffraction patterns that arise with different edges and apertures using wave mechanics. Optical microscopy is often called a “diffraction-limited” technique because this phenomenon limits the spatial resolution attainable as long as all other variables are optimal (e.g., no bubbles in the immersion oil or dried buffer on the objective). Images are degraded by diffraction in all standard optical microscopes. When Young reported his ground-breaking experiments on the interference of light in 1803, he immediately recognized the importance of diffraction in optical microscopy. He wrote “a central dark spot, and a light spot surrounded by a darker circle, may respectively be produced in the images of a semitransparent and opaque corpuscle; and impress us with an idea of a complication of structure which does not exist” [17]. A so-called Abbe diffraction apparatus was employed to demonstrate such effects in microscopy, as described in Chapter 8 by Pelc. The principal contribution to diffraction in optical microscopy arises from the propagation of light through the condenser and the objective (for applications such as epifluorescence, only the

212

Pina Colarusso

objective is considered). The diffraction pattern resulting from light through a circular aperture such as a microscope objective is approximated by an Airy function; in two dimensions, it is known as the Airy diffraction pattern. It consists of a central bright (“Airy”) disk surrounded by alternating bright and dark rings that decrease in intensity as the distance from the centre of the disk increases. The radius of the Airy disc defines resolution according to the Rayleigh criterion (Fig. 1D); see Chapter 9 by Wilson et al. for details. An optical microscopy image can be considered as the sum of the individual points within the image; it follows that each point is associated with a three-dimensional diffraction pattern. Generally, the individual diffraction patterns are not obvious unless small distinct particles such as receptors, beads, or quantum dots are being imaged. In this case, diffraction rings will be visible around the particle, and will be especially prominent if the specimen is brought in and out of focus. The original description of the effect of diffraction on the ultimate spatial resolution obtainable by optical microscopy is credited to Ernst Abbe whose original paper [18] is also available in English [19]. Specialized optical techniques such as confocal [20], deconvolution [21], and super-resolution [3] microscopy are designed to minimize the effects of diffraction.

5

Polarization We have considered light as a ray and then as a wave. We now will consider the orientation of the electric field vector as it propagates ! through space. As shown in Fig. 1A, the electric field E can be ! represented as a vector. By examining Eq. 11, ξ denotes the unit vector (direction) of the electric field at any instantaneous point in time as the light wave travels along z. The orientation of the electric field vector is known as the polarization. If the electric field vibrates in one plane as in Fig. 1A, light is referred to being linearly (or plane) polarized, as depicted in Fig. 3. ! Although this polarization state exists, in the most general case ξ will rotate around the direction of propagation. When the tip of the electric vector traces out a circle instead of an ellipse, the light is known as circularly polarized. Further classification is provided by the direction of rotation; when observing the wave head-on (Fig. 3 bottom), light is labeled as right circularly polarized if the electric component is rotating clockwise and left circularly polarized if it is rotating counterclockwise. The most general polarization state is known as elliptical polarization because the tip of the electric field vector traces an ellipse; note that elliptically polarized light can also be designated right or left elliptically polarized analogously to circularly polarized light. Finally, a group of electromagnetic waves all exhibiting the identical orientation of the electric components that are oscillating in phase is known as polarized light

Properties of Light

213

Fig. 3 Polarization. Unpolarized, linearly polarized, and circularly polarized light shown schematically. Linear polarizer converts unpolarized light to linearly polarized one. Quarter-wave (λ/4) plate converts linearly polarized light to circularly polarized one. At the bottom, views down the light ray are shown (adapted from https://en.wikipedia.org/wiki/Waveplate)

(elliptically, circularly, or linearly). Unpolarized light consists of a mixture of electric field vectors pointing in random directions. In nature, most light is partially polarized and contains a mixture of both polarized and unpolarized light. The polarization state is used to produce contrast in many different types of optical techniques such as differential interference contrast and polarization microscopy. However, the human eye and most detectors are not sensitive to polarization, thus indirect methods have been developed that reveal the polarization state of light through the discrimination of brightness (intensity) or wavelength. This fundamental property of light has also been exploited in the design of fluorescent probes that are also sensitive to polarization. In addition, the polarization of light is important to consider when working with optical microscopes that include gratings or detectors that are sensitive to polarization, as the efficiency of these components can vary with the polarization of light. Typically, optical microscopy techniques based on polarization make use of optical elements known as polarizing filters, or simply polarizers. Filters that produce linear, circular, or elliptically polarized light have been applied in optical microscopy. When working with optical microscopes, it is often useful to know how to determine the transmission axis of a linear polarizer. A linear polarizer will transmit light only along one axis while blocking the other

214

Pina Colarusso

components (producing linearly polarized light). If light is directed through two identical linear polarizers at normal (perpendicular) incidence, and one filter is rotated while the other is held fixed, the transmitted light intensity will vary periodically from minimum to maximum brightness. The points of maximum transmission correspond to both polarization axes being aligned parallel to each other, so that light of the same linear polarization is passing through both filters. The angle at which there is minimum transmission of light corresponds to an orthogonal orientation of transmission axes of both polarizers. The polarizers in this position are referred to as “crossed” [13]. Indeed, it is possible to determine the quality of the filters by taking the ratio of the intensity of the light transmitted when the polarizers are crossed to the intensity of the incident light. This ratio is known as “maximum extinction”; inexpensive Polaroid sheets for example will have extinction ratios of about 1000 while the high-quality filters used in polarization or differential interference microscopy, for example, have extinction ratios of between 10,000 and 100,000 [22]. Although we have been reviewing the individual properties of light in isolation, it is important to note that light–matter interactions can affect one or more properties of light at the same time. Take reflection, for example. Light undergoing reflection from a dielectric (i.e., electrically nonconductive) surface such as glass or water can undergo a change in polarization or in phase shift. The extent of the changes in polarization and phase that occur upon reflection depend on the refractive indices of the media and the angle of incidence (described by the Fresnel equations, which can be found in any standard textbook on optics) [13, 23]. As a practical example, polarized sunglasses work because they remove the reflected light (glare) for surfaces such as water or ice. This glare is linearly polarized in the horizontal direction, so the sunglasses are simply oriented to accept vertical polarizations and reject the horizontally polarized glare. If you own a pair of polarizing sunglasses, rotate your head sideways and you should see the glare start to emerge. Unlike humans, crustaceans and insects are capable of detecting polarized light.

6

Scattering and Absorption Thus far, we have considered reflection, refraction, and transmission using geometrical optics, and have treated interference, diffraction, and polarization using the wave model of light. Two other phenomena are important contributors to the images in optical microscopes, namely, scattering and absorption. Scattering and absorption are related in that they lead to extinction, the attenuation of light in its direction of propagation. However, they arise

Properties of Light

215

from distinct light–matter interactions. While scattering leads to a deviation of light from its original trajectory, absorption leads to a reduction of the energy in the propagating beam. Scattering can occur over all frequencies for any set of particles, while absorption can occur only at specific frequencies corresponding to differences in energy states that are characteristic for chemical constituents of the substance. Scattering also occurs on a much faster time scale than absorption [24]. 6.1

Scattering

Scattering occurs when the propagating light wave encounters an inhomogeneity with dimensions on the order of its wavelength or smaller (when the size of the object is larger than the wavelength of light, behavior of light can be described by reflection and refraction). The interaction with the inhomogeneity, also knows as a scattering locus, causes the light to deviate from its original trajectory. Typically, the light is scattered in many different directions though most of the light continues along its original trajectory. In optical microscopy, most samples are complex mixtures of materials with various refractive indices and geometries; different organelles in cells, for example, provide multiple scattering loci. A hallmark of scattering is the resulting polarization of light. If the incoming light is linearly polarized, the scattered light maintains the same polarization. Scattered light also can be polarized even if the incoming light is not. Unpolarized light remains unpolarized in the original direction of travel, but the light emitted at right angles to the axis of propagation is linearly polarized. Intermediate angles are partially polarized. Detailed descriptions of the resulting polarization of scattered light can be found in several references [13, 25]. Scattering is known as elastic or inelastic. Elastic scattering yields photons at the same frequency as the incident photons, while inelastic scattering yields photons that are higher or lower in frequency. Inelastic scattering will not be considered further, though it is at the heart of novel microscopy techniques such as coherent anti-Stokes Raman scattering (CARS) microscopy [26]. Two important types of elastic scattering are Rayleigh (the proper attribution should name Tyndall as well) and Mie scattering. Rayleigh scattering occurs when the particles are much smaller than the wavelength of the incident light (by about a factor of ten or greater). The degree of scattering depends on several variables, but the key result is that Rayleigh scattering is proportional to 1/λ4, the inverse of the fourth power of the wavelength of the impinging light. The strong dependence on wavelength means that violet and blue light is scattered to a much greater extent than red light (because the wavelength of light increases going from violet to red). Rayleigh scattering of sunlight by oxygen and nitrogen molecules in the upper atmosphere is the reason the sky is blue on

216

Pina Colarusso

a clear day.1 In biological systems, Rayleigh scattering also occurs at the cell membrane as well as at certain regions of the eye. With Mie scattering, the size of the particles is comparable to the wavelength of the impinging light. Mie scattering shows a weak dependence on wavelength compared to the strong dependence shown in Rayleigh scattering. An example of Mie scattering can be seen overhead on a cloudy day; clouds are composed of large particles that scatter all wavelengths equally, and thus appear white. Scattering contributes to the appearance of cells and tissues imaged with an optical microscope. Some techniques such as darkfield microscopy are based solely on the scattering of light by the specimen. In some areas of biological imaging, light scattering is considered as a valuable intrinsic signal that yields useful information about the composition of the sample, and has been exploited in areas such as tissue diagnostics, as for example in the determination of precancerous changes in biopsy specimens [27]. While scattered light carries the information about specimen structure and is thus essential for image formation, in optical microscopy, excessive scattering is often viewed as an undesirable effect that leads to blur, thus degrading the overall quality of the image, and limiting the depth to which specimens can be imaged [28]. Indeed, various specimen preparation techniques have been devised to minimize scattering [29]. These are collectively referred to as “optical clearing,” described in Chapter 6 by Becker et al. From everyday experience, we know that ordered solids reflect and refract light with minimal scattering. Consider a clear block of ice or glass, for example. Light can be reflected or refracted as it impinges on this solid, but one would expect minimal scattering. It turns out that scattering is appreciable only when there are local fluctuations in the spatial arrangement of the particles. In an ordered solid such as ice, light is also scattered, but it destructively interferes in all directions except in those predicted by the laws of reflection or refraction (Fig. 2C). A small impurity in the ice or glass such as a piece of dust or a bubble, however, would be observable as scattering of light from its original trajectory. Sometimes confusion can arise between the terms scattering and diffraction. Diffraction can be viewed as a macroscopic manifestation of scattering. However, subtle differences in the usage of the two terms exist. Diffraction usually refers to scattering from a uniform barrier, or a set of regularly spaced obstacles, that typically leads to small changes in the trajectory of the propagating light. Scattering can lead to large deviations of the light trajectory, and additional parameters such a polarization and the optical properties 1

The reader may wonder why the sky is not violet rather than blue, as shorter wavelengths are more strongly scattered (violet has a shorter wavelength than blue). The blue sky arises because there is a modest peak in the solar emission in the blue above the earth’s atmosphere. Furthermore, our eyes are more sensitive to blue than violet [25]

Properties of Light

217

of the inhomogeneity are considered. As a concrete example, we refer to the diffraction pattern associated with a microscope objective (the Airy disc, Fig. 1D) compared to scattering of light by a mitochondrion. In periodic objects such as diatom frustule (see Chapter 8 by Pelc), interference of scattered light yields distinct diffraction patterns, similar in nature to X-ray diffractogram of DNA. 6.2

Absorption

Quantum mechanics specifies that atoms and molecules occupy discrete energy states, and that they can undergo transitions between different energy states by absorbing or releasing photons. Historically, Bohr showed that hydrogen can absorb and release photons of certain frequencies (wavelengths) but not others (recall that a photon’s energy is equal to hf (¼hc/λ) as shown in Eq. 2). The photons absorbed and released have energies that correspond to the differences in energy levels of the hydrogen atom that can be detected as sharp lines in the absorption or emission spectra of hydrogen [30]. Photons with energies in the visible region typically involve transitions between electronic energy levels of atoms and molecules. Electronic transitions in molecules are more complicated than those in atoms because these transitions are accompanied by changes in the vibrational and rotational energy levels as well [31]. These transitions between energy levels can only occur when two conditions are met: (1) the incoming photon must have an energy (frequency) close to the difference in energy between the levels (the probability of absorption increases the better the energy match); (2) the transition occurring between the two energy states for both absorption and emission must satisfy various selection rules derived from quantum mechanics. The absorption of photons leads to the attenuation of the incoming electromagnetic radiation. Often these transitions are mapped out by measuring the attenuation of incident light as a function of wavelength with a spectrometer. The resulting plots are known as absorption spectra. For atoms, the absorption occurs over a very narrow frequency range compared to the wide frequency ranges seen for molecules arising from vibrational and rotational motion. For this reason, we often refer to “lines” in atomic spectra while we refer to “bands” in molecular spectra. In biological microscopy, the absorbing species are generally molecular. They can be extrinsic labels such as histological dyes (Chapter 1 by Wouterlood and Langer) or fluorescent tags (Chapter 4 by Wouterlood and Chapter 5 by Mishra et al.), and intrinsic components, such as melanin or hemoglobin. The absorbance A is directly proportional to the concentration of the constituents absorbing in the sample, according to Beer’s law. It is expressed as follows:

218

Pina Colarusso

A ¼ εb c

ð12Þ

where ε is the molar extinction coefficient at a specific wavelength of light, b is the physical path length (sample thickness) and c is the concentration of the molecular species. The absorbance for a mixture is additive, comprising the sum of the individual absorbances of the molecular species within the sample. Note that Eq. 12 is linear only for relatively moderate concentrations of an analyte roughly corresponding to values of A between 0.2 and 0.9. Once an atom or molecule absorbs a photon, it will eventually release the energy and return to the original energy state. For atoms, the visible photon absorbed will be reemitted as a photon of the same energy. For molecules, however, the photon release can occur through several relaxation mechanisms, such as fluorescence, phosphorescence, chemiluminescence, or heat. Fluorescence is the most important relaxation mechanism in optical microscopy. The absorbance is directly related to the refractive index n of a material. Earlier we briefly mentioned that the refractive index n is dependent on wavelength. From Eq. 3 n is the ratio of the speed of light in a vacuum to the speed of light in a material at a specific wavelength. It follows then that the velocity of light through a material depends on the wavelength. The refractive index of a substance tends to be higher for wavelengths closer to an absorption band. It follows then that light with a wavelength close to the absorption band travels slower than light with a wavelength further from the absorption band. In glass, for example, red light propagates faster than blue light because glass has an absorption band in the ultraviolet region, which borders on the violet-blue edge of the visible spectrum [32]. The dependence of refractive index on wavelength means that light passing through a refractive element such as a prism is separated into its constituent wavelengths. This arises because the individual wavelengths of light experience different refractive indices and are refracted to different degrees, as specified by Snell’s law (see Eq. 8). Violet light, for example, experiences a higher refractive index than red light, meaning that it will be refracted to a greater extent by a glass prism. This phenomenon as known as chromatic dispersion. Chromatic dispersion is an important consideration in optical microscopy. To illustrate, it is important to choose a microscope objective with minimal chromatic dispersion (also known as chromatic aberration) in order to achieve proper image registration across all wavelengths. Otherwise, the images recorded at different wavelengths will be misaligned and lead to erroneous interpretation of data, such as when examining the relationships between different molecular constituents. Furthermore, chromatic dispersion leads to broadening of multiwavelength light pulses as they propagate through optical elements such as prisms or fiber optics. This is relevant to techniques such as multiphoton microscopy.

Properties of Light

6.3 A Note about Color

219

Most sensors used in optical microscopy allow for discrimination among different wavelength regions of the visible spectrum through prisms, gratings, filters, or specialized coatings or masks on the detector elements themselves. This wavelength discrimination of course lies at the heart of techniques such as fluorescence microscopy, in which different cell types, organelles, proteins or other cellular constituents can be identified by using tags that emit over different wavelength ranges. However, detectors do not sense “color” as a normal human would, but instead detect the intensity of light within well-defined spectral intervals. Although grayscale images are common in microscopy, it is often useful to work with color images. We now turn our attention to a brief description of the connection between the spectral properties of an object and the human perception of color. Visible light, as mentioned previously, ranges between approximately 400 and 700 nm, which is the range of electromagnetic radiation that can be detected by the normal human eye. In dim lighting conditions, the rod cells in the retina predominate in visual perception. The rods contain a sole pigment, rhodopsin, and thus human vision under these conditions is monochromatic. In higher light, the cones contribute to visual perception. The cone cells in the eye contain pigments that are sensitive to long, middle, and short wavelengths, with absorption peaks at approximately 420, 530, and 560 nm [33, 34]. Normal human color vision is thus known as trichromatic. We sense color because our brain is wired to interpret the stimulation of the cone receptors in the retina [34–36]. However, the exact mechanism by which the cone response to light is translated to the sensation of color is not fully understood. By comparison, sea mammals are monochromatic, other mammals except primates are dichromatic, and fish and most reptiles are tetrachromatic. Some shrimps can detect 12 colors [37]. However, discrimination between wavelengths is worse than in (trichromatic) humans, probably due to less advanced post-processing of color input in the brain. In the physical environment, the cones in the retina are stimulated by the different visible wavelengths that surround us, from the colors of a rainbow to the output of a laser pointer. The brain then interprets the physical stimulation of the cones as “blue” or “yellow” depending on the relative stimulation of the S, M, and L cones. Most color sensations, however, are not associated with pure spectral colors in the natural world. Objects appear colored because they selectively absorb some wavelengths and reflect or transmit others. The cones in the retina are stimulated to varying degrees, and in turn, the human brain perceives this sensation as chromatic color. A computer projector produces yellow light by a combination of red and green light, but our eyes cannot distinguish this from pure yellow light. Objects that reflect or transmit all

220

Pina Colarusso

wavelengths equally are perceived as black, grey or white, depending on their relative intensities. Black, grey, and white are known as achromatic colors. Intriguingly, the perception of color does not arise simply from the relative stimulation of the S, M, and L cone pigments, but also depends on the context in which the wavelengths are observed. In short, the brain does not always interpret equivalent spectral information from the same object as having the same color [34]. Coupled with the variation and defects in color perception in the human population, it should be recognized that any technique that relies on the visual observation to gauge the appearance of color has the potential for misinterpretation [38]. Nevertheless, the use of color as a tool to discriminate among different components is ubiquitous in optical microscopy, particularly in histology, which is based on the use of color to differentiate among the varied components in a specimen. The sample preparation methods are usually standardized so that the majority of the population will interpret the color information in a similar fashion. Furthermore, most histological specimens are examined under very similar lighting, typically white-light that approximates natural illumination by sunlight. We have described spectral colors as well as colors that arise from absorption, reflection and transmission. Effects such as interference and scattering can also lend color to objects. This type of coloration is known as structural color [39]. Colors arising from interference can be used to provide analytical information in polarization microscopy [40]. Note that the same processes giving color to microscopic specimens also lead to the brilliant blue of a summer sky (Rayleigh scattering), rainbows (dispersion), and the wonderful iridescence of a peacock feather or a butterfly wing (interference).

7

Conclusion In this chapter, we have explored some fundamental properties of light and its interactions with matter, with a focus on light microscopy. Many of the concepts introduced here are considered more fully in other chapters. Snell’s law is the basis for magnification by an objective and eyepiece, and diffraction influences the ultimate spatial resolution of optical microscopy. Other phenomena such as interference and polarization are shown to be important keystones in both bright-field and differential interference microscopy. Scattering and absorption are important processes that generate image contrast, especially when observing tissue sections in histological preparations. We have painted the properties of light in broad strokes yet we hope that it provides a solid foundation for an understanding of the detailed descriptions of optical microscope techniques found in the other chapters of this volume. On a closing

Properties of Light

221

note, interested readers are encouraged to delve into one of the many excellent textbooks on optics and vision to learn more about this fascinating and fundamental area.

Glossary Absorption Attenuation of light intensity upon passing through a medium or an object. Diffraction Macroscopic manifestation of ►scattering, due to ►interference of scattered light with itself. A diffraction pattern (due to light scattering in the specimen) may be viewed at the back focal plane of an objective lens. It is the Fourier transform of specimen structure. Dispersion The change in ►refractive index as a function of wavelength. Dispersion shows a greater change for wavelengths that are close to absorption bands for a material. GRIN lens A gradient-of-refractive-index lens featuring a gradient of ►refractive index, which improves its imaging capability. It is employed, for example, in deep brain (endoscopic) imaging. Interference Superposition of two or more light waves. Direct and diffracted light interfere in the microscope to form image. Optical path difference Also referred to as “optical thickness,” it is the product of physical thickness (of the object) and ►refractive index difference between the object and the medium surrounding it. Optical path length The distance of the light traveled in a material, multiplied by its ►refractive index. It represents the distance light would have traveled in a vacuum. Point of incidence The point at which an electromagnetic wave impinges on an interface from one material to another. Often used within the ray model of light. Polarization A property that describes the orientation of the electric field vector in an electromagnetic wave. In the general case, polarization is elliptical, but simple orientations such as linear (plane) and circular polarization also exist. Ray model A simplified model used to describe the propagation of light. The wave motion is simplified by depicting it as a ray in the direction of propagation. Refraction The bending of light as it propagates from one medium to another, differing from each other by ►refractive index. Refractive index The ratio of the speed of light in a vacuum to the speed of light in a material.

222

Pina Colarusso

Scattering Deviation of illuminating (direct) light from its path by localized inhomogeneities such as cell organelles. Sometimes used as a synonym of ►diffraction. Wave model A more accurate representation of light, in which the electric and magnetic fields are viewed as vibrating at right angles to each other as well as at right angles to their axis of propagation.

References 1. Zernike F (1955) How I discovered phase contrast. Science 121(3141):345–349. https:// doi.org/10.1126/science.121.3141.345 2. Denk W, Strickler JH, Webb WW (1990) Two-photon laser scanning fluorescence microscopy. Science 248(4951):73–76. https://doi.org/10.1126/science.2321027 3. Formsiero EF, Rizzoli SO (eds) (2014) Superresolution microscopy techniques in the neurosciences, Neuromethods, vol 86. Springer, New York. https://doi.org/10.1007/978-162703-983-3 4. Crombie AC (1990) Expectation, modeling and assent in the history of optics: Part I. Alhazen and the medieval tradition. Stud Hist Phil Sci 21(4):605–632. https://doi. org/10.1016/0039-3681(90)90035-7 5. Sengupta DL, Sarkar TK (2003) Maxwell, Hertz, the Maxwellians, and the early history of electromagnetic waves. IEEE Anten Propag Mag 45(2):13–19. https://doi.org/10.1109/ MAP.2003.1203114 6. Adawi I (1989) Centennial of Hertz radiowaves. Am J Phys 57(2):125–127. https:// doi.org/10.1119/1.16106 7. Staley R (2008) Worldviews and physicists’ experience of disciplinary change: on the uses of ‘classical’ physics. Stud Hist Philos Sci A 39 (3):298–311. https://doi.org/10.1016/j. shpsa.2008.06.002 8. Feynman RP (1985) QED: the strange theory of light and matter. In: Alix G (ed) Mautner memorial lectures. Princeton University Press, Princeton (NJ, USA). ISBN: 9780691083889 9. International Bureau of Weights and Measures (2006) Le Syste`me International d’Unite´s (SI) ¼ The International System of Units, 8th edn. Bureau International des Poids et Mesures, Se`vres. ISBN: 9282222136 10. Jenkins FA, White HE (1976) Fundamentals of optics, 4th edn. McGraw-Hill, New York. ISBN: 978-0070323308

11. Pendry JB (2004) Negative refraction. Contemp Phys 45(3):191–202. https://doi.org/ 10.1080/00107510410001667434 12. Pendry JB, Smith DR (2006) The quest for the superlens. Sci Am 295(1):60–67. https:// www.ncbi.nlm.nih.gov/pubmed/16830681 13. Hecht E (2002) Optics, 4th edn. AddisonWesley, Reading (MA, USA). ISBN: 9780805385663 14. Burghardt TP, Hipp AD, Ajtai K (2009) Around-the-objective total internal reflection fluorescence microscopy. Appl Opt 48 (32):6120–6131. https://doi.org/10.1364/ AO.48.006120 15. Axelrod D (2001) Selective imaging of surface fluorescence with very high aperture microscope objectives. J Biomed Opt 6(1):6–13. https://doi.org/10.1117/1.1335689 16. Shtengel G, Galbraith JA, Galbraith CG, Lippincott-Schwartz J, Gillette JM, Manley S, Sougrat R, Waterman CM, Kanchanawong P, Davidson MW, Fetter RD, Hess HF (2009) Interferometric fluorescent super-resolution microscopy resolves 3D cellular ultrastructure. Proc Natl Acad Sci U S A 106(9):3125–3130. https://doi.org/10.1073/pnas.0813131106 17. Young T (1804) The Bakerian lecture: experiments and calculations relative to physical optics. Philos Trans 94:1–16. https://doi. org/10.1098/rstl.1804.0001 18. Abbe E (1873) Beitr€age zur Theorie des Mikroskops und der mikroskopischen Wahrnehmung. (Schultze’s) Arch Mikroskop Anatom (Bonn) 9:413–468. https://doi.org/10. 1007/BF02956173. BHL: 13290331. https://biodiversitylibrary.org/page/ 13290331 19. Abbe E (1875) A contribution to the theory of the microscope, and the nature of microscopic vision. Proc Brist Natural Soc 1(Pt2):200–261. https://biodiversitylibrary.org/page/ 7506100

Properties of Light 20. Pawley JB (2006) Handbook of biological confocal microscopy, 3rd edn. Springer, New York. https://doi.org/10.1007/978-0-387-455242 21. Swedlow JR (2007) Quantitative fluorescence microscopy and image deconvolution. Methods Cell Biol 81:447–465. https://doi.org/ 10.1016/S0091-679X(06)81021-6 22. Fish KN (2009) Total internal reflection fluorescence (TIRF) microscopy. Curr Protoc Cytom 50(1): Unit 12.18. https://doi.org/ 10.1002/0471142956.cy1218s50 23. Saleh BEA, Teich MC (2007) Fundamentals of photonics, 2nd edn. Wiley, Hoboken (NJ, USA). ISBN: 978-0471358329 24. Prasad PN (2003) Introduction to biophotonics. Wiley, Chichester (UK) and Hoboken (NJ, USA). ISBN: 0471287709 25. Zare RN (1995) Laser experiments for beginners. University Science Books, Sausalito (CA, USA). ISBN: 978-0935702361 26. Evans CL, Xie XS (2008) Coherent anti-Stokes Raman scattering microscopy: chemical imaging for biology and medicine. Annu Rev Anal Chem 1:883–909. https://doi.org/10.1146/ annurev.anchem.1.031207.112754 27. Boustany NN, Boppart SA, Backman V (2010) Microscopic imaging and spectroscopy with scattered light. Annu Rev Biomed Eng 12:285–314. https://doi.org/10.1146/ annurev-bioeng-061008-124811 28. Ntziachristos V (2010) Going deeper than microscopy: the optical imaging frontier in biology. Nat Methods 7(8):603–614. https:// doi.org/10.1038/nmeth.1483 29. Dickie R, Bachoo RM, Rupnick MA, Dallabrida SM, Deloid GM, Lai J, Depinho RA, Rogers RA (2006) Three-dimensional visualization of microvessel architecture of wholemount tissue by confocal microscopy. Microvasc Res 72(1–2):20–26. https://doi.org/10. 1016/j.mvr.2006.05.003 30. Bohr N (1922) The structure of the atom. In: Nobel lectures (physics), 1922–1941, pp 7–47. Elsevier, Amsterdam (1965). ISBN:

223

9781483222486. https://www.nobelprize. org/prizes/physics/1922/bohr/lecture/ 31. Harris DC, Bertolucci MD (1989) Symmetry and spectroscopy: an introduction to vibrational and electronic spectroscopy. Dover Publications, New York. ISBN: 978-0486661445 32. Fox M (2010) Optical properties of solids. Oxford master series in physics (2nd ed). Oxford University Press, New York. ISBN: 978-0199573370 33. Goldstein EB (2009) Sensation and perception, 8th edn. Barnes & Noble, Belmont (CA, USA). ISBN: 9780495601500 34. Conway BR, Chatterjee S, Field GD, Horwitz GD, Johnson EN, Koida K, Mancuso K (2010) Advances in color science: from retina to behavior. J Neurosci 30(45):14955–14963. https://doi.org/10.1523/JNEUROSCI. 4348-10.2010 35. Conway BR (2009) Color vision, cones, and color-coding in the cortex. Neuroscientist 15 (3):274–290. https://doi.org/10.1177/ 1073858408331369 36. Neitz J, Neitz M (2008) Colour vision: the wonder of hue. Curr Biol 18(16): R700–R702. https://doi.org/10.1016/j.cub. 2008.06.062 37. Thoen HH, How MJ, Chiou T-H, Marshall J (2014) A different form of color vision in mantis shrimp. Science 343(6169):411–413. https://doi.org/10.1126/science.1245824 38. Rubin LR, Lackey WL, Kennedy FA, Stephenson RB (2009) Using color and grayscale images to teach histology to color-deficient medical students. Anat Sci Educ 2(2):84–88. https://doi.org/10.1002/ase.72 39. Kinoshita S, Yoshioka S (2005) Structural colors in nature: the role of regularity and irregularity in the structure. ChemPhysChem 6 (8):1442–1459. https://doi.org/10.1002/ cphc.200500007 40. Murphy DB, Davidson MW (2012) Fundamentals of light microscopy and electronic imaging, 2nd edn. Wiley, New York. https:// doi.org/10.1002/9781118382905

Chapter 8 Beyond Brightfield: “Forgotten” Microscopic Modalities Radek Pelc Abstract “Forgotten” microscopic modalities, devices, and accessories derived from or complementary to brightfield microscopy are briefly surveyed. These include off-axis illumination, schlieren contrast, Abbe diffraction apparatus, Rheinberg illumination, darkfield and phase-contrast microscopy combined, incidentillumination microscopy, camera lucida, and comparison microscopy. Examples of their use are shown. While most of them are no longer or only rarely available or used, they are still important for proper understanding of image formation, contrast generation, and data interpretation in microscopy. In some cases, they are superior to their more modern counterparts. Key words Comparison microscope, Condenser, Conjugate planes, Darkfield, Incident illumination, Ko¨hler illumination, Objective, Rheinberg illumination, Schlieren contrast

1

Introduction The late professor David John Hugh Cockayne (1942–2010), former president of the International Federation of Societies for Microscopy, claimed1 that it is vital to read the very original papers in the field. Ernst Abbe (1840–1905) is, of course, a classical example, as his main paper, although translated to English [1] shortly after the German original [2] was published in 1873, is very rarely quoted (forty times less often than the original according to Google Scholar [earliest citation dated 1996]). A possible explanation may be that over 98% of the English-speaking population can also read German with confidence . . . Whatever is the case, Colin J. R. Sheppard has recently done justice to Abbe in this sense [3]. It is worth noting that a number of previously popular microscopic modalities are either extinct or repeatedly “rediscovered,” as exemplified in Subheading 2. August Ko¨hler suggested a new way of illuminating microscopic specimens [4, 5] some time after the introduction by Ernst

1

16th International Microscopy Congress (IMC16), Sapporo, Japan (3-8 Sept 2006)

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_8, © Springer Science+Business Media LLC 2020

225

226

Radek Pelc

Abbe of a better condenser [6, 7]. Both of them worked jointly with Carl Zeiss in the newly founded company, Carl Zeiss AG in Jena, Germany. Ko¨hler’s main aim was to facilitate photomicrography so that homogeneous distribution of light intensity in the viewing field is achieved. Previously, the so-called “critical” (Nelson) illumination was more popular, and a ribbon filament lamp was the preferred light source [8]. The chief advantage of the Ko¨hler illumination method is that the light source does not have to be homogeneous for the viewing field to be evenly bright, so that an ordinary light bulb can be used; the tungsten wire coil is not seriously degrading the image (a diffuser is sufficient to mitigate that). This can be achieved by positioning the light source in the condenser front focal plane. However, as this would result in exposing the specimen to excessive heat, it is more practical to place the light source further away, in an optically conjugate plane (Fig. 1); the collector lens (Fig. 2A, B) serves this purpose. The light source in the microscope is typically fixed, and optimal condenser position must be adjusted for each objective lens. This adjustment when changing objectives was easier in the so-called pancratic (“all-mighty”) condenser that used to be part of Zeiss microscopes (e.g., NfpK or Amplival); the condenser position did not need to be vertically changed, and an internal zooming system was used to achieve Ko¨hler illumination.

2

Off-Axis Illumination and Schlieren (Modulation) Contrast An important implication of the Ko¨hler illumination principle is that the condenser can be partly obstructed at its front focal plane (aperture diaphragm level), without introducing luminance inhomogeneity into the image. In a microscope properly adjusted for Ko¨hler illumination, and its condenser diaphragm fully open (ideally matching the numerical aperture [NA] of the objective), an axially symmetrical light beam illuminates the specimen. Nonabsorbing objects such as unstained living cells are rendered in minimal contrast, and very thin ones such as filopodia or lamellipodia are hardly visible. Decentering the condenser diaphragm, or asymmetrically obstructing it results in contrast enhancement. An example is shown in Fig. 2C, D obtained in a slightly different way, in that an accessory lens (rather than the diaphragm) of the condenser is offset. The contrast enhancement is greater in optical thicker objects. Condensers in microscopes made till ~1960s by Carl Zeiss (e.g., NfpK from 1960s) and Meopta (e.g., C36Bi) were equipped with a laterally shifting diaphragm (Fig. 2E) to achieve this effect [9, 10]. In author’s experience, however, a shifting straight-edge diaphragm (Fig. 2F) is more efficient, and available

“Forgotten” Microscopic Modalities

Camera

A

227

Primary image plane

Fluorescence lamp

Exit pupil of the eyepiece Primary image plane

Halogen lamp

Objective back focal plane Specimen plane Condenser front focal plane (aperture diaphragm)

Condenser focusing

CONDENSER CENTERING

Field diaphragm

Lamp filament

B

C

D CONDENSER FOCUSING AND CENTERING INCORRECT

CONDENSER CENTERING INCORRECT

CONDENSER PROPERLY ADJUSTED

Fig. 1 Optically conjugate planes of an upright microscope. (A) Imaging and aperture planes are marked in blue and red, respectively. Adapted from Ref. [39] by permission of © Elsevier. Note the original designation of the “field diaphragm” was “field-of-view diaphragm” (Sehfeldblende in German). The framed text is reproduced from Ref. [5]. (B–D) Field diaphragm as viewed through the eyepiece while centering the condenser (images by Dr. Lisa Cameron); see Appendix 2 for details

in the RCH condenser2 made by Lambda Praha (originally Meopta, Czechoslovakia). Examples of its use may be found elsewhere [11, 12]. Image contrast can be further improved by employing another asymmetric diaphragm in the objective back focal plane (optically conjugate with the condenser diaphragm), the so-called schlieren diaphragm or modulator [13]. As surveyed elsewhere [11] this 2

Relief contrast after Hostounsky´. Dr. Zdeneˇk Hostounsky´ (1925-2013) was a protozoologist and insect pathologist at Czechoslovak Academy of Sciences in Prague, and a founding member of The Stentor Institute.

228

Radek Pelc ID Obj SPECIMEN

B C

STAGE

Con

EYEPIECE

AD

D

FD

OBJECTIVE

CL

A

CONDENSER

COLLECTOR LENS

E

CONDENSER

10 µm

F

CONDENSER SHIFTING IRIS DIAPHRAGM SHIFTING EDGE DIAPHRAGM

Fig. 2 Adjustment of a microscope for Ko¨hler and off-axis illumination. (A) Original diagram adapted from Refs. [4], [5]; copyright expired 70 years after author’s death (August Ko¨hler, 1866–1948). Ko¨hler’s portrait is as published by Zeiss Microscopy (Jena, Germany). AD Aperture diaphragm (in condenser), CL Collector lens, Con Condenser, FD Field diaphragm, ID Iris diaphragm (inside objective), Obj Objective lens. (B) Field (imaging) and aperture planes. Adapted from Ref. [40]; the origin of this drawing (recently shown in Ref. [41]) may be traced to Carl Zeiss materials at least 75 years old [10]. (C, D) Off-axis (oblique) illumination obtained by offsetting a condenser accessory lens. Reproduced from Ref. [42]. C Condenser aperture diaphragm viewed through a centering telescope. D Unstained protozoon (Peranema trichophorum) and its separately contrast-optimized flagellum, otherwise hardly visible in axially symmetrical illumination. (E, F) Off-axis (oblique) illumination setups utilizing a shifting diaphragm. E When needed, the shifting-iris diaphragm is swung into a working position under the condenser. Reproduced from Ref. [9]. F Dedicated off-axis illumination condenser with a shifting-edge diaphragm. Reproduced from Ref. [12]

“Forgotten” Microscopic Modalities

229

simple modality was “rediscovered” approx. every 25 years since it was first published [14], most likely because the papers describing it usually lacked micrographs. Eventually, it matured into Hoffman modulation contrast [15], a direct competitor of a noticeably more costly differential interference contrast (DIC Nomarski) invented more than 20 years earlier [16]. Single-sideband edge-enhancement microscopy represents yet another variant [17, 18]. At this point, it should be emphasized that in images of greater optical thickness such as cell clusters, bigger cells or tissue replicas the setup employing no modulator (i.e., off-axis illumination) is more suitable (data not shown). In that case the objective aperture fulfills the role of a modulator to some extent [19] when the condenser diaphragm is open a bit more than required by the numerical aperture of the objective. This represents a slight departure from the optimal Ko¨hler illumination (i.e., condenser diaphragm setting exactly matching the numerical aperture of the objective) as stray light may start contributing to image formation. However, this is usually more than compensated for by improved image contrast. It should be noted that the improvement is not very significant as the objective aperture is circular rather than straight (cf. the text above on the condenser shifting diaphragms (Fig. 2E, F)).

3

Abbe Diffraction Apparatus The abovementioned modulation (schlieren) contrast relies on selectively filtering certain diffraction maxima (orders) at the objective back focal plane, with the aim to improve image contrast. In brightfield microscopy, there is no filtering. Darkfield microscopy, on the other hand, represents the other extreme, in that all direct (undiffracted) light, also referred to as the 0th order diffraction component, is blocked. Examples are shown in Subheadings 4 and 5. A classical (the simplest) example of the importance of diffractive phenomena in microscopic image formation is shown in Figs. 3 and 4. Preventing the 1st and higher diffractive orders (maxima) from contributing to image formation results in complete disappearance from images of the structures (periodically spaced dots) represented by them (Fig. 3B). Likewise, merely increasing the illumination wavelength λ (e.g., by switching from blue to red light) renders previously visible structures invisible (Fig. 3C, D) as resolution is inversely proportional to wavelength (resolution limit ¼ λ/2NA [1, 2]). The first of these phenomena can be demonstrated with the aid of an objective with a built-in iris diaphragm. For example, the Nikon 100/0.501.30 objective offers the option to gradually

230

Radek Pelc

A

B

C

D

2 µm 1st 1st 1st

1st

1st

1st 0th

0th 1st

1st

1st

1st

1st

1st

NA = 1.40 (full) Green light

NA = 0.40 Green light

OPENING OF THE OBJECTIVE APERTURE

NA = 0.80 Blue light

NA = 0.80 Red light

BLOCKED 1st DIFFRACTION MAXIMA

Fig. 3 Diffractive nature of microscopic image formation. Diatom (Pleurosigma angulatum) viewed with a highresolution, oil-immersion objective (60/0.40  1.40) fitted with an iris diaphragm making it possible to adjust its numerical aperture anywhere between maximum (1.40 in A) and minimum (0.40 in B); an intermediate value (0.80) is shown in C and D. (A, C) Inclusion of the 1st order diffraction maxima makes the periodic structure visible. (B, D) Exclusion of the 1st order diffraction maxima renders the periodic structure invisible. (C, D) Dependence of resolution on wavelength: blue light resolves the structure, red light does not. Adapted from Ref. [43]

reduce its numerical aperture from 1.30 down to 0.50 (resolution worsens 1.30/0.50 ¼ 2.6-fold). In order to convince the scientific community that the diffraction theory of image formation [1, 2] is indeed valid [20] the so-called Abbe Diffraction Apparatus was designed, and later also made commercially available by Carl Zeiss [8, 21]. Nowadays, it may be occasionally found on eBay. Its components are shown in Figs. 4 and 5, including the effects on images of passing or blocking specific diffraction maxima (Fig. 4). False structures (lines nonexistent in the specimen) appear in the image if only the 1st diffraction maxima are blocked because the spatial frequency (line density) seemingly increases. One may say that the 0th and 2nd diffraction order rays “do not know” the 1st diffraction order rays are missing, and interfere with each other as usual, but this time forming a nonrealistic image. Similar demonstrations have been described [22] and illustrated [23] in greater detail elsewhere, most extensively by Kurt Michel [24]. They would be surely of great benefit in microscopy training courses. Earlier, the equipment was referred to as the “Abbe Demonstration Microscope” [8] or the “Pulfrich-Abbe

“Forgotten” Microscopic Modalities

NO IMAGE

REALISTIC IMAGE

231

FALSE IMAGE

IMAGE PLANE

OBJECTIVE back focal plane diaphragm

S1

DIFFRACTION FUNNEL

0th only

-1st, 0th, +1st

-2nd, 0th, +2nd

DIFFRACTION MAXIMA passed

CONDENSER front focal plane diaphragm

Fig. 4 Abbe diffraction apparatus in action. This simple device used to be commercially available from Carl Zeiss at least since 1937. The objective back focal plane diaphragms are inserted into a slider (S1) in the “diffraction funnel” fitted above the objective of an upright microscope. Depending on which diffraction maxima are allowed to pass the image looks different (0th order maximum yields no image at all, passing only the second order maxima (i.e., blocking the first order one) results in twice higher density of the lines in the image (i.e., every other line in the image is an artifact). A hypothetical object consisting of vertical parallel lines is considered here, similar to those engraved in the “Diffraction Plate” that is part of the original equipment (Fig. 5). The diffraction funnel is reprinted from Carl Zeiss catalogue, “Der Diffraktions-Apparat nach Abbe” (Druckschrift “Mikro 11-432-1” dated 1940)

Demonstration Microscope” [25]. Please note that the phase plates shown in Fig. 5 were not included in the early models as phasecontrast microscopy was only invented in early 1930s.

4

Rheinberg Illumination Until the invention of phase-contrast microscopy [26] darkfield and schlieren microscopy were dominating the realm of imaging unstained (nonabsorbing) objects. A modality combining darkfield and brightfield microscopy became known as Rheinberg illumination [27]. It can render objects under investigation in a color of choice, against a background of another color. This is achieved by using concentric color filters (Fig. 6) inserted into the condenser filter holder. If the central disc is black and sufficiently large the brightfield component is not present, and a darkfield image in a color of choice is obtained (Fig. 6C) [28]. A dedicated condenser called “Mikropolychromar” (Fig. 6A) used to be manufactured by Carl Zeiss since 1933 [8], and found its

232

Radek Pelc

LAMPHOUSE SLIT

SLIDERS S1 S2

Eyepiece pinhole (‘diopter’)

Clip-M = Clip-on magnifier D-Fun = Diffraction funnel Substage diffuser

COLOR FILTERS

Clip-M CONDENSER

Ground glass insert

D-Fun

Eyepiece x16

DIFFRACTION PLATE (testing specimen)

Achromatic objective x6.3/0.20

Fig. 5 Abbe diffraction apparatus at rest. The “diffraction plate” hosts microscopic gratings as testing objects. The red box highlights (left) five aperture stops (masks), and (right) two phase (“retardation”) plates and corresponding substage condenser diaphragms. Each of the masks and phase plates fits in the S1 slider inserted into the diffraction funnel (D-Fun) mounted above an objective. The S2 slider is fitted with a variable diaphragm, and enables experiments shown in Fig. 3 even if the objective is not fitted with the iris diaphragm. To directly examine the (intermediate) image of the specimen the eyepiece may be replaced with a groundglass insert. The diffraction pattern generated by the specimen is inspected either with a clip-on magnifier (Clip-M) above the eyepiece (with an effect of either a centering [“phase”] telescope or Bertrand lens) or with an eyepiece pinhole (“diopter”). Details may be found in Ref. [21]. Image (unmarked) reprinted by permission of © The Trustees of the National Museums of Scotland (Edinburgh, UK) where this piece of equipment (made ca. 1970) is held

use, for example, in studying intracellular motility [29]. EastmanKodak Co. was supplying Wratten-Rheinberg filter sets for this microscope [8, 30]. Thread cells of hagfish slime gland [31, 32] could be conveniently visualized under Rheinberg illumination using custom-made color filters (Fig. 6D). It should be noted though that the Mikropolychromar is in fact a simple condenser, not a dedicated darkfield one, and as such only performs well at smaller magnifications. An excellent recent review about Rheinberg illumination is available [33]. A dedicated darkfield condenser capable of mixing brightfield and darkfield images was also made and referred to as the “QuickChange-Over” condenser (Fig. 7A) [30], implying the ease of switching between brightfield and darkfield. The author has not encountered any images arising from its use. An interesting

MIKROPOLYCHROMAR

“Forgotten” Microscopic Modalities

233

B

A

C BRIGHT FIELD DARK FIELD

500 µm

x5

BF

20 µm

50 µm SEM

D

x2 20 µm

A

B

C

D

Fig. 6 Rheinberg illumination. Color discs shown at the bottom (A-D) were used to obtain the images. (A) The Mikropolychromar condenser (aplanatic) capable of mixing brightfield and darkfield images. As it is not a dedicated darkfield condenser it is suitable for low-power objectives only (cf. Fig. 7a). Reproduced from Carl Zeiss catalogue, “Mikropolychromar” (Druckschrift “Mikro 493/II” from 1938). (B) Lens-cleaning paper. Original magnification 50, image width ca. 2 mm. Courtesy of © Stephen W. Downing (University of Minnesota Medical School-Duluth Campus [Duluth, MN, USA]), originally presented in a 1980 photomicrography competition (https://www.nikonsmallworld.com/people/steve-downing). (C) Proboscis of a house cricket (Acheta domesticus) in pure darkfield (the brightfield component is blocked). Courtesy of © Stefano Barone (Diatom Lab, Italy; www.diatomshop.com, www.testslides.com). Image originally presented in a 2014 photomicrography competition (https://www.nikonsmallworld.com/people/stefano-barone), and a similar one elsewhere [28]. (D) Two thread cells of a hagfish slime gland, with the threads unwinding. Corresponding brightfield (BF) and SEM images are shown. Courtesy of © Stephen W. Downing (University of Minnesota Medical School-Duluth Campus [Duluth, MN, USA])

234

Radek Pelc DF

DF BF

Control arm BF

Lamelae

D

ID

ID ED c

c

p

W. Watson & Sons (London)

TRAVISS STOP (expanding diaphragm, ED)

p

A

B

Fig. 7 Darkfield microscopy aided by Traviss stop. (A) Dedicated darkfield condenser for mixing darkfield and brightfield illumination. It used to be manufactured by Leitz, and marketed as the “Quick-Change-Over” condenser. BF Brightfield. C Central rays for darkfield illumination. D Diffusely reflecting surface generating brightfield illumination. DF Darkfield. ED Expanding diaphragm (Traviss stop). ID Iris (aperture) diaphragm. p Peripheral rays for brightfield illumination. (B) The Traviss stop made the use of a darkfield condenser more straightforward, even though originally not part of the type shown here. Image A is adapted from Ref. [10]. Image B is based on Ref. [34] (copyright expired 70 years after author’s death; Edmund J. Spitta, 1853–1921 [20]), as adapted in Ref. [44] (reprinted by permission of © Macmillan Magazines/Springer)

accessory of darkfield condensers is the so-called Traviss expanding stop, essentially an iris diaphragm working in reverse (Fig. 7B). Jointly with the standard iris diaphragm, light annulus of any diameter and thickness can be produced. This is helpful in correctly adjusting darkfield illumination [34]. The Traviss stop used to be manufactured by W. Watson & Sons (London) [8].

5

Heine Condenser and Incident Illumination The examples shown in the previous section (Rheinberg illumination) illustrate the capabilities of darkfield microscopy. A more advanced option was available in the form of the Heine condenser (Fig. 8A, B). This was a variant of the cardioid condenser (the shape of its mirror is derived from the cardioid curve), that is, a dedicated darkfield condenser performing well not only at small magnifications. Additionally, it enabled easy switching not only to brightfield but also to phase-contrast imaging [35, 36]. The cardioid condenser alone (inset in Fig. 8B), that is, without the phase-contrast modality add-on, makes it possible to visualize, for example, single unstained microtubules (Fig. 8C), and to follow their dynamic instability (Fig. 8D). Darkfield imaging provides better contrast in images of single microtubules than other suitable label-free modalities such as interferometric scattering [37] or interference reflection [38] microscopy.

“Forgotten” Microscopic Modalities

235

HEINE CONDENSER

Z = phase ring

cardioid condenser

L’ = illuminating light

BRIGHT BRIGHT PHASE FIELD CONTRAST FIELD

DARK FIELD

single microtubule

DARK FIELD

5 µm

Distance (µm) of microtubule tip from a reference point 6 4 2

0

A C

10

Time (min)

20

B D

Fig. 8 Heine condenser. (A, B) The Heine condenser is based on the cardioid (i.e., dedicated darkfield) condenser, and makes it possible to gradually change from one modality to another (brightfield, darkfield and phase contrast). The main image is reproduced from Leitz catalogue № 51.3-5a/Engl.–X/60/FY/L (early 1950s). The photograph of the condenser (complete with a screwable lens for use of immersion oil) was taken by Peter Ho¨bel (Erlangen, Germany; http://www.mikroskopie-ph.de/index-Heine.html). The ray diagram (bottom right) is reproduced from Ref. [10]. (C, D) Time-lapse imaging of a single unstained microtubule by darkfield microscopy, using the cardioid condenser (inset in B). Microtubule growth and shortening (dynamic instability) can be followed; microtubule tip is the plus end. Adapted from Ref. [45] by permission of © Nature Publishing Group (Springer)

More recently, a revival of the combined illumination scheme has been presented in an incident illumination setup (Fig. 9) better suited to inspect, for example, tissue surfaces. As it also provides darkfield and phase-contrast images, complementary image information can be conveniently obtained (Fig. 9B, C). It is inspired by the “Ultropak” device (Leitz) which did not offer the phasecontrast modality.

236

Radek Pelc

Diaphragm

Phase contrast Bright field

BRIGHT FIELD

B

Phase contrast Phase plate

~100 µm

C

Specimen Phase plate

A

ULTROPAK Phase contrast Dark field

DARK FIELD

Fig. 9 Incident illumination. (A) Vertical illuminators inspired by the “Ultropak” (Leitz) which was not fitted with a phase plate. Various modalities (color-coded in the image) and their combinations may be obtained, depending on the annular diaphragm(s) used (small, large or both). The drawings are adapted from Refs. [46, 47] by permission of © Cambridge University Press. The photograph is reproduced from Leitz catalogue № 513-36a/Engl (1965). (B, C) Filamentous algae in incident illumination; note complementary rendering in brightfield, darkfield, and phase-contrast. Reprinted from Ref. [46] by permission of © Cambridge University Press

6

Camera Lucida and Comparison Microscopy The drawing attachment (tube) often referred to as camera lucida, literally “light chamber,” used to be commonplace in microscopy laboratories even when photomicrography was already widely used. When using it, the microscopist is simultaneously viewing the cell under the microscope, and his/her drawing of that cell (Fig. 10A). Semi-transparent mirror (beam splitter) project the drawing into the eyepieces. Naturally, both images should be of comparable brightness, and built-in rotatable polarizing filters, for example, in camera lucida made by Carl Zeiss (“Zeichenapparat”) facilitate that. Nowadays, software is available to skeletonize cell images, that is, to convert their grayscale representations (typically 8-bit, 0 to 255) to line-drawing type images (one-bit models). Nevertheless, it is still often more convenient to draw the images manually with camera lucida (Fig. 10B). The drawings of neurons by Ramo´n y Cajal (1852–1934) of course represent a classical example.3 A need often arises to view two similar scenes (specimens) simultaneously. For this purpose, a so-called comparison microscope may be conveniently employed. It in fact consists of two separate microscopes connected with a special (dual) eyepiece

3

The Beautiful Brain (Abrams Books 2017, ISBN: 9781419722271)

“Forgotten” Microscopic Modalities

237

A Beam splitter

OLM

B Drawing projected to eyepieces

Fig. 10 Camera lucida or drawing attachment. (A) Example of fitting to Wild M20-EB microscope. Adapted from Ref. [48] (© Ian Walker). (B) Camera lucida drawings of freshly isolated chick retina photoreceptor cells (cones) aligned by the outer limiting membrane (OLM). The color spots are oil droplets acting as color filters aiding color vision. Adapted from Ref. [49] (©Lo´pez-Lo´pez et al.)

A A

Microscope #1

EYEPIECE

Microscope #2

B

WILD-TYPE

50 µm

UV cones Rods

MUTANT

Fig. 11 Comparison microscopy. (A) The comparison eyepiece fits two microscopes working in concert. The Zeiss drawing is reproduced from Ref. [10]. The optical diagram was drawn by Tama´s Szo˝cs (https://en. wikipedia.org/wiki/Comparison_microscope). (B) An example of use. As the author of the present chapter is not in possession of the comparison eyepiece separately obtained images are shown: Retinae of zebrafish larvae, wild-type versus lots-of-rods mutant (lor p25bbt). Reproduced from Ref. [50] by permission of © Natl. Acad. Sci. USA

tube capable of merging the two images into one (Fig. 11A). An example of comparison microscopy is shown in Fig. 11B in which the retina of wild-type and mutant zebrafish larvae is presented. As digital images can be easily acquired and displayed, the comparison eyepiece dating to at least 75 years ago [10] is

238

Radek Pelc

nowadays hardly encountered in biological laboratories. However, it is still useful in situations when only one digital camera is available, or when routine comparative observations are made. Indeed, dedicated comparison microscopes, inseparable from each other, have been commercially available at least since 1935. Examples of situations where comparative microscopy is useful include pathological, biological, forensic, and industrial laboratories specializing, for example, in the following [8]: 1. Examining tissue in health and disease 2. Identifying powdered adulterated drugs 3. Biological systematics (inspecting unknown vs. type specimen) 4. Identifying crystals, hair, or textile fibers from a crime scene 5. Comparing optical performance of two microscopes

Acknowledgments The author is grateful to Dr. Petro Khoroshyy (Czech Academy of Sciences, Prague) and Dr. Floris G. Wouterlood (Amsterdam University Medical Centers) for helpful comments, and acknowledges support via Ministry of Education projects: Chiral Microscopy (LTC17012) and ChemBioDrug.4

Appendix 1 (Hands-on Demonstration) Diffraction and Resolution Inspect a diatom specimen5 using an objective fitted with a built-in iris diaphragm (e.g., Nikon 100/0.50  1.30) at different settings (i.e., different effective numerical aperture, NA). In this way, it is possible to artificially reduce the objective’s resolving power to the extent (NA ¼ 0.40 in our example) that the finest details in the diatom image completely disappear (Fig. 3B). Alternatively, use an ordinary objective (having no iris diaphragm) jointly with a diffraction funnel fitted under the ocular head of an upright microscope. Insert S2 slider (with a built-in iris diaphragm) into the funnel (Figs. 4 and 5). Inspect the objective back focal plane using a centering (“phase”) telescope or Bertrand lens6 to monitor the diffraction maxima of different orders while closing and opening the diaphragm in (or above) the objective lens. Typically, 0th and 1st 4

CZ.02.1.01/0.0/0.0/16_019/0000729 Available, for example, from Diatom Lab (www.diatomshop.com, www.testslides.com) 6 The Bertrand lens is an extra focusable lens that, when inserted into the optical path, works in conjunction with the eyepiece to form a small telescope to give a magnified view of the objective back focal plane. 5

“Forgotten” Microscopic Modalities

239

order maxima will be visible (the former representing direct or undiffracted light). Note that the diffraction patterns may be clearly observed only with the condenser aperture diaphragm closed as much as possible (down to a “pinhole”). Illuminate the diatom specimen with light of different wavelengths. Blue light (shorter wavelength, ca. 450 nm) is more likely to resolve fine details than red light, ca. 700 nm) as the resolving power is inversely proportional to NA. For this effect to be sufficiently prominent, the objective (or S2 slider) iris controlling the effective NA needs to be set to an appropriate value (0.80 in the example shown in Fig. 3C, D), as checked by viewing the diffraction patterns in the objective back focal plane. Be aware that different diatom species feature structures of different periodicities.

Appendix 2 (Exercise) ¨ hler Illumination Setting Ko (minor adaptation of a text originally written by Dr. Lisa Cameron, Light Microscopy Core Facility, Duke University, Durham, NC, USA) Following is a step-by-step protocol for Ko¨hler illumination with transmitted light. An interactive version is also available at the online Microscopy U(niversity)7 and elsewhere8. For information on focusing the light source, please see the “Focusing the light source” section further below. First, open all diaphragms. Raise the condenser to its highest point (on an upright microscope). Put a well-stained specimen on the stage, and inspect it with a low power (10) objective. Focus the objective lens to obtain a sharp image of the specimen by using the coarse and fine focus controls. This first step sets the correct relationship between the specimen and objective lens. On a binocular microscope, each user may need to adjust the eyepieces for their own eyes for optimal focus. At least one eyepiece will have an adjustment collar. Use one eye to look down the microscope and focus on some detail in the specimen while keeping the eye which uses the eyepiece with adjustable collar closed. Then switch and use the other eye. Turn the adjustable eyepiece collar to focus the same detail in the image as sharp as before. This procedure is referred to as diopter adjustment and is recommended every time a microscope is used, for the sake of microscopist’s visual comfort.

7 8

https://www.microscopyu.com/tutorials/kohler https://micro.magnet.fsu.edu/optics/timeline/people/kohler.html

240

Radek Pelc

Setting Ko¨hler illumination is possible without it if using one eye only. Focus the condenser by first closing the illuminated field diaphragm and then adjust the height of the condenser with the condenser focus knob until a sharp image of the field diaphragm is seen superimposed on the image of the specimen (Fig. 1C). Make sure that the condenser diaphragm is wide open. This adjustment sets the correct relationship between the condenser lens and the specimen. If the microscope has an Abbe condenser, this image will likely have a fringe of color around the field aperture. Center the condenser lens. To do this, make the image of the field diaphragm concentric with the field of view (Fig. 1D) using the condenser centering screws. This adjustment makes the optical axis of the condenser lens coincide with that of the microscope as defined by the field diaphragm and the objective lens. Adjust the area of the field that is illuminated. Open the field diaphragm until its image is just outside the field of view; readjust the condenser centering if necessary (as you open it). This ensures that illumination falls only on the area of specimen within the field of view, and that the diameter of the primary image is only a little larger than the field-limiting diaphragm as seen by the eyepiece. This prevents light from falling on the internal walls of the microscope to be scattered to produce hot spots and haze, reducing contrast in the final image. Adjust the aperture diaphragm (illuminating aperture) in the condenser. To do this, remove the eyepiece, or turn the Bertrand lens into position if available—look down the microscope tube from ca. 100 mm above the tube, and observe the back focal plane of the objective, the disc of light at the base of the tube. More conveniently, use a centering (“phase”) telescope in place of the eyepiece, in the same way as during adjustment of the annular diaphragm for phase-contrast imaging. Close the aperture diaphragm until the image of the iris is approximately 70–80% of the viewing field (the aperture of the objective). Replace the eyepiece (or remove the Bertrand lens). The working (effective) aperture of the condenser is now slightly smaller than the aperture of the objective lens. Do not close the diaphragm too far; this will cause a serious deterioration in the quality of the image. Adjust the brightness of illumination using the control on the lamp power supply, or by inserting neutral density filter(s). These are usually found along the base of the microscope between the lamp and the field diaphragm. The microscope optical adjustments or diaphragms should not be used to control brightness. This will adversely affect the quality of the image. For instance, if the condenser diaphragm is closed too much, the image will appear too contrasty, as refractile structures will be highlighted too much due to diffraction effects; and with it wide open, there will be glare due to stray light (internal reflections). The resolution is poor in both.

“Forgotten” Microscopic Modalities

241

In a microscope with absolutely no internal reflections the setting is optimal when the effective numerical aperture of the condenser (adjustable by its diaphragm) matches the NA of objective in use. As such microscopes do not exist in reality the abovementioned setting of ca. 70–80% is recommended. Image contrast is slightly improved yet the diffraction artifacts thus introduced are minimal. For a higher power objective: Rotate the nosepiece to the 40 dry objective. Owing to parfocality of objective design, the 40 objective should be almost in focus after aligning the microscope for 10; it was not the case in very old microscopes. As before, focus and center the image of the field diaphragm using the condenser focus knob and the condenser centering knobs. The aperture of the field diaphragm will need to be readjusted. Remove an eyepiece (ideally replace it with the centering [“phase”] telescope), or use the Bertrand lens to observe the back focal plane of the objective. Notice that the area illuminated for the low power objective is much smaller than the diameter of the back aperture of the 40 objective. Adjust the condenser diaphragm so that the effective NA of the condenser is about the same as the objective NA. For a high-power oil immersion objective: Rotate the nosepiece so that a high-power oil immersion objective is near-vertical. Just before it is clicked into place, stop and add oil to the coverslip, as close as possible to the optical axis (light beam coming from condenser prealigned at smaller magnifications, see above). Be sure the oil droplet does not have any bubbles. Use immersion oil provided by the microscope manufacturer, as there are some slight differences. Ideally, the refractive index of the oil, coverslip, and objective lenses should be the same. Click the oil immersion objective into place. The space between the front lens of the objective and the coverslip should now be filled with oil. Remove an eyepiece (ideally replace it with the centering [“phase”] telescope) or use the Bertrand lens to view the back aperture of the objective. Open the condenser diaphragm to almost fill the objective aperture. Replace the eyepiece (or remove the Bertrand lens) and observe the specimen. Adjust the field diaphragm until the edge just matches the field of view. Strictly speaking, the condenser should again be readjusted as above (focus and centering). However, switching from 40 to 100 objective will rarely misalign the condenser beyond tolerable limit.

242

Radek Pelc

Focusing the light source: Remove the diffuser from the lamp housing or along the base of the microscope stand, if possible, in order to see bulb and filament. Lamp illumination should fill most of the front aperture of the condenser. Put a sheet of lens paper on the specimen stage to help visualize the area of illumination. Focus light on the lens paper by moving the lamp-focusing knob. Then remove the eyepiece (ideally replace it with the centering [“phase”] telescope) or insert the Bertrand lens to view the back focal plane of the objective. Be sure the lamp filament is centered and focused in the plane of the condenser diaphragm. Adjust the collector lens on the lamp housing. N.B.: Many student microscopes and more recently released modern research ones do not have illumination bulb adjustments, but are designed to deliver even illumination. References 1. Abbe E (1874/1875) A contribution to the theory of the microscope, and the nature of microscopic vision. Proceedings of the Bristol Naturalists’ Society (new series) 1 (Pt 2):200–261 (translated and commented on by H. E. Fripp). https://bio diversitylibrary.org/page/7506100 2. Abbe E (1873) Beitr€age zur Theorie des Mikroskops und der mikroskopischen Wahrnehmung. (Schultze’s) Archiv fu ¨ r Mikroskopische Anatomie (Bonn) 9:413–468. https://doi. org/10.1007/BF02956173. BHL:13290331. https://biodiversitylibrary.org/page/13290 331 3. Sheppard CJR (2017) Resolution and superresolution. Microsc Res Tech 80(6):590–598. https://doi.org/10.1002/jemt.22834 4. Ko¨hler A (1893) Ein neues Beleuchtungsverfahren fu¨r mikrophotographische Zwecke. Zeitschrift fu¨r Wissenschaftliche Mikroskopie und fu¨r Mikroskopische Technik 10 (4):433–440. https://biodiversitylibrary.org/ page/3013561 5. Ko¨hler A (1894) New method of illumination for photomicrographical purposes. J Roy Microsc Soc (volume not numbered, April issue):261–262. https://biodiversitylibrary. org/page/49603250 6. Abbe E (1873) Ueber einen neuen Beleuchtungsapparat am Mikroskop. (Schultze’s) Archiv fu¨r Mikroskopische Anatomie (Bonn) 9:469-480. https://doi.org/10.1007/ BF02956177. BHL:13290387. https://bio diversitylibrary.org/page/13290387 7. Abbe E (1875) A new illuminating apparatus for the microscope. Monthly Microscopical J

13:77–82. https://biodiversitylibrary.org/ page/14276730 8. Needham GH (1958) The practical use of the microscope. Charles C Thomas, Springfield, IL. https://lccn.loc.gov/57010440 9. Metzner P (1928) Das Mikroskop, 2nd edn. Franz Deuticke, Leipzig/Vienna 10. Wolf J (1944) Mikroskopicka´ technika. Vesmı´r, Prague 11. Hostounsky´ Z, Pelc R (2006) An efficient way of high-contrast, quasi-3D cellular imaging: off-axis illumination. J Biochem Biophys Meth 68(1):23–30. https://doi.org/10. 1016/j.jbbm.2006.03.016 12. Hostounsky´ Z, Pelc R (2007) Phase-contrast versus off-axis illumination: is a more complex microscope always more powerful? Adv Physiol Educ 31(2):232–235. https://doi.org/10. 1152/advan.00028.2006 13. Axelrod D (1981) Zero-cost modification of bright field microscopes for imaging phase gradients on cells: Schlieren optics. Cell Biophys 3 (2):167–173. https://doi.org/10.1007/ BF02788132 14. To¨pler A (1866) Ueber die Methode der Schlierenbeobachtung als mikroskopisches Hu¨lfsmittel, nebst Bemerkungen zur Theorie der schiefen Beleuchtung. (Poggendorf’s) Ann Phys Chem 127(4):556–580. [Ann Phys 203 (4):556–580]. https://doi.org/10.1002/ andp.18662030405 15. Hoffman R, Gross L (1975) The modulation contrast microscope. Nature 254:586–588. https://doi.org/10.1038/254586a0

“Forgotten” Microscopic Modalities 16. Padawer J (1968) The Nomarski interferencecontrast microscope. An experimental basis for image interpretation. J Microsc (Oxf) 88 (Pt3):305–349. https://doi.org/10.1111/j. 1365-2818.1968.tb00616.x 17. Bretschneider F, Teunis PFM (1994) Reducedcarrier single-sideband microscopy: a powerful method for the observation of transparent microscopical objects. J Microsc (Oxf) 175 (Pt2):121–134. https://doi.org/10.1111/j. 1365-2818.1994.tb03475.x 18. Inoue´ S, Spring K (1997) Video microscopy: the fundamentals. Plenum Press (Springer), New York. ISBN: 978-0306455315 19. Murphy DB, Davidson MW (2012) Fundamentals of light microscopy and electronic imaging, 2nd edn. Wiley-Blackwell, New York. https://doi.org/10.1002/ 9781118382905 20. Bradbury S (1996) The reception of Abbe’s theory in England. Proc Roy Microsc Soc 31 (4):293–299. http://www.microscopist.co. uk/wp-content/uploads/2017/04/Thereception-of-Abbes-Theory-in-England.pdf 21. Nuttall B (2008) Image formation in the microscope: the Abbe diffraction apparatus. Bull Sci Instr Soc 96:27–29 22. Fripp HE (1876/1877) On professor Abbe’s experimental demonstration of diffraction spectra in the microscope. Proc Bristol Naturalists’ Soc (new series) 2(Pt 1):24–38. BHL:7545726. https://biodiversitylibrary. org/page/7545726 23. Stephenson JW (1877) Observations on professor Abbe’s experiments illustrating his theory of microscopic vision. Monthly Microscopical J 17:82–88. BHL:14157879. https://biodiversitylibrary.org/page/ 14157879 24. Michel K (1964) Die Grundzu¨ge der Theorie des Mikroskops in elementarer Darstellung (Physik und Technik), vol 1. Wissenschaftliche Verlagsgesellschaft, Stuttgart 25. Wetzel RA, Needham GH (1928) Some photographs of optical images after Abbe. Washington Meeting of the American Physical Society (Washington, D.C., 20–21 Apr 1928). Phys Rev 31(6):1108. https://doi.org/10. 1103/PhysRev.31.1108 26. Zernike F (1955) How I discovered phase contrast. Science 121(3141):345–349. https:// doi.org/10.1126/science.121.3141.345 27. Rheinberg J (1896) On an addition to the methods of microscopical research, by a new way of optically producing colour-contrast between an object and its background, or between definite parts of the object itself. J

243

Roy Microsc Soc (volume not numbered, August issue):373–388. https://bio diversitylibrary.org/page/49597776 28. Gautier C (2012) Proboscis (“tongue”) of a cricket (Gryllus campestris) under Rheinberg illumination. The Scientist (Image of the Day, 30 Oct 2012) https://www.the-scientist.com/ image-of-the-day/image-of-the-day-40285 and Nikon Small World https://www. nikonsmallworld.com/people/christiangautier 29. Bu¨nning E (1935) Einige Beobachtungen u¨ber die Ko¨rnchenstro¨mung in den Staubfadenhaaren von Tradescantia, nebst Bemerkungen u¨ber die Verwendbarkeit des Mikropolychromars bei zellphysiologischen Studien. Zeitschrift fu¨r wissenschaftliche Mikroskopie und fu¨r mikroskopische Technik 52(2):166–172 30. Shillaber CP (1959) Photomicrography in theory and practice, 5th edn. John Wiley & Sons, New York (Chapman & Hall, London) 31. Downing SW, Spitzer RH, Salo WL, Downing JS, Saidel LJ, Koch EA (1981) Threads in the hagfish slime gland thread cells: Organization, biochemical features, and length. Science 212 (4492):326–328. https://doi.org/10.1126/ science.212.4492.326 32. Downing SW, Spitzer RH, Koch EA, Salo WL (1984) The hagfish slime gland thread cell. I. A unique cellular system for the study of intermediate filaments and intermediate filamentmicrotubule interactions. J Cell Biol 98 (2):653–669. https://doi.org/10.1083/jcb. 98.2.653 33. Havics AA (2014) Contrast methods in microscopy: Rheinberg illumination. The Microscope 62(4):157–169 34. Spitta EJ (1907) Microscopy. The construction, theory and use of the microscope. E. P. Dutton & Co, New York € 35. Gottschewski GHM (1953) Uber die Anwendungsmo¨glichkeiten der neuen LeitzPhasenkontrasteinrichtung nach Heine. Zeitschrift fu¨r Wissenschaftliche Mikroskopie und fu¨r Mikroskopische Technik 61(4):185–200 36. James P (2003) The Heine condenser (Part 1 and 2). Micscape, issue 95 http://www. microscopy-uk.org.uk/mag/sep03ind.html and 97 http://www.microscopy-uk.org.uk/ mag/nov03ind.html 37. Andrecka J, Arroyo JO, Lewis K, Cross RA, Kukura P (2016) Label-free imaging of microtubules with sub-nm precision using interferometric scattering microscopy. Biophys J 110 (1):214–217. https://doi.org/10.1016/j.bpj. 2015.10.055

244

Radek Pelc

38. Mahamdeh M, Simmert S, Luchniak A, Sch€affer E, Howard J (2018) Label-free highspeed wide-field imaging of single microtubules using interference reflection microscopy. J Microsc 272(1):60–66. https://doi.org/10. 1111/jmi.12744 39. Rottenfusser R (2013) Proper alignment of the microscope. Meth Cell Biol 114:43–67. https://doi.org/10.1016/B978-0-12407761-4.00003-8 40. Bradbury S, Evennett PJ, Haselmann H, Piller H (1989) RMS dictionary of light microscopy (RMS Handbook №15). Oxford University Press and Royal Microscopical Society, Oxford. ISBN: 978-0198564218 41. Sanderson J (2019) Understanding light microscopy. Wiley (RMS series), Hoboken, NJ. ISBN: 978-0470973752 42. Wessenberg H, Reed MK (1971) The use of oblique illumination in microscopic observations of living protozoa. Trans Am Microsc Soc 90(4):449–457. https://jstor.org/stable/ 3225459 43. Michel K (1943) Grundzu¨ge der Mikrophotographie. In: Zeiss Nachrichten, Sonderheft (special issue), vol 4, 2nd edn. Kommissionsverlag Gustav Fischer, Jena, pp 1–192 44. Amos WB (2000) Lessons from the history of microscopy. Nat Cell Biol 2(8):E151–E152. https://doi.org/10.1038/35019639 45. Horio T, Hotani H (1986) Visualization of the dynamic instability of individual microtubules

by dark-field microscopy. Nature 321 (6070):605–607. https://doi.org/10.1038/ 321605a0 46. Piper T, Piper J (2012) Variable phase darkfield contrast – a variant illumination technique for improved visualizations of transparent specimens. Microsc Microanal 18(2):343–352. https://doi.org/10.1017/ S1431927612000153 47. Piper T, Piper J (2013) Variable phase brightfield contrast – an alternative illumination technique for improved imaging in transparent specimens. Microsc Microanal 19(1):11–21. https://doi.org/10.1017/ S1431927612013323 48. Walker I (2005) Wild M20-EB microscope. Micscape, issue 122 http://www.microscopyuk.org.uk/mag/dec05ind.html 49. Lo´pez-Lo´pez R, Lo´pez-Gallardo M, Pe´rez´ lvarez MJ, Prada C (2008) Isolation of chick A retina cones and study of their diversity based on oil droplet colour and nucleus position. Cell Tiss Res 332(1):13–24. https://doi.org/10. 1007/s00441-007-0572-6 50. Alvarez-Delphin K, Morris AC, Snelson CD, Gamse JT, Gupta T, Marlow FL, Mullins MC, Burgess HA, Granato M, Facool JM (2009) Tbx2b is required for ultraviolet photoreceptor cell specification during zebrafish retinal development. Proc Natnl Acad Sci U S A 106 (6):2023–2028. https://doi.org/10.1073/ pnas.0809439106

Chapter 9 Stereomicroscopy in Neuroanatomy Erin E. Wilson, William Chambers, Radek Pelc, Paul Nothnagle, and Michael W. Davidson Abstract The main benefits of the stereomicroscope are that it is designed in a modular motif, allowing for a wide range of accessories such as stands, eyepieces, objectives, and illuminating bases for a wide variety of contrast enhancement techniques. Often utilized to study the surfaces of specimens, the stereomicroscope frequently uses incident (reflected) illumination, permitting the observation of specimens that would normally be too thick or opaque. Translucent and transparent objects can be successfully imaged with a number of transmitted illumination methods depending on the observer’s needs, and fluorescence stereomicroscopes are being increasingly used for three-dimensional observation. An excellent working distance, ranging from 3 to 5 cm to even 20 cm in certain models, and wide field of view that these models feature, are critical factors in the observation of a far-ranging variety of biological specimens. Examples of nervous and other tissues are presented in this review, for example, retina in living fish, paraneurons in shrimp, innervation of murine heart, and YFP-expressing regenerating nerves in the cornea of transgenic mice. Key words Barrel distortion, Common main objective, Convergence, Diascopic illumination, Episcopic illumination, f-number, Greenough, Keystone effect, Perspective distortion

Abbreviations CMO f-number FN NA YFP

1

Common main objective A parameter inversely proportional to NA of the objective Field number (of the eyepiece) Numerical aperture Yellow fluorescent protein

Introduction The first microscope utilizing stereoscopic vision was designed and built by Che´rubin d’Orle´ans in 1671, but his invention was actually a pseudostereoscopic system that achieved image erection only through the application of supplemental lenses

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_9, © Springer Science+Business Media LLC 2020

245

246

Erin E. Wilson et al.

[1]. Stereomicroscopy languished at this stage of development for more than 150 years, after which Sir Charles Wheatstone authored a treatise on binocular vision [2], providing a catalyst for further development. Over the next 80 years, Francis Herbert Wenham and John Ware Stephenson used achromatic prisms to build what are considered the first truly functional stereomicroscopes [3]. The stage was set for American instrument designer Horatio S. Greenough, who at the dawn of the industrial age in the 1890s, unveiled an innovative design featuring inverting (erecting) prisms, that was to become the forefather of modern stereomicroscopes. About half a century later, the first modern stereomicroscope was introduced in the United States by the American Optical (Spencer) company of Southbridge (MA, USA). Dubbed the Cycloptic®, this breakthrough design featured a die-cast aluminum housing, a consistent working distance (at four inches, one of the longest), and an internal magnification changer, which enabled the observer to increase the objective magnification from 0.7
to 2.5
in five steps. As is the case with any innovative new product, American Optical soon faced competition. In 1959, Bausch and Lomb introduced its own stereomicroscope, which had a significant new advance that would ultimately revolutionize stereomicroscopy: continuously variable, or zoom magnification. Named the StereoZoom, the new instrument was the first stereomicroscope without erecting prisms. Similar in size and shape to the Cycloptic, the StereoZoom also had a comparable magnification range (0.7
to 3
) with similar working distances. By the 1960s, several Japanese companies, including Olympus, Nikon, and Unitron, had rolled out their own zooming stereomicroscopes as they sought to increase their presence in the USA. As a group, the American, Japanese and European microscope manufacturers continued advancing the development of “bigger and better” stereomicroscopes offering a plethora of new features. These advances were accelerated by the invention of high-speed computers, which enabled optical designers to tackle the complex problem of creating an effective variable magnification zoom lens system. Modern stereomicroscope designs are characterized by high numerical aperture objectives that yield high resolution images featuring a minimum amount of flare and geometrical distortion. The observation tubes will accommodate high-eyepoint eyepieces having a field of view up to 26 millimeters, with a diopter adjustment that allows the image and reticle (graticule) to be merged into focus simultaneously. Additionally, many models have high zoom ratios (up to 15
and beyond) that provide a wide magnification range (between 2
and 540
), and reduce the necessity to change objectives. For comfort, ergonomic features incorporated into the microscope designs help reduce fatigue during long hours of

Stereomicroscopy in Neuroanatomy

247

operation, and new accessories enable modern stereomicroscopes to image specimens that were impractical just a few years earlier. Working in concert, the human eyes and brain produce what is referred to as stereoscopic vision, which provides spatial, threedimensional images of objects, all because of the brain’s interpretation of the two slightly different images received from each of the retinas. On average, human eyes are about 65 mm apart, and each eye perceives an object from a somewhat different perspective that differs by a few degrees. When transmitted to the brain, the images are fused together, but manage to retain a high degree of depth perception, a truly remarkable occurrence. The stereomicroscope takes advantage of this ability to perceive depth by transmitting twin images that are inclined by a small angle (convergence, usually between 10 and 12 ) to yield a legitimate stereoscopic effect [4].

2

Stereomicroscope Designs Stereomicroscopes can be roughly split into two basic families, each of which possesses positive and negative characteristics. The older stereomicroscopic system, named after the inventor Greenough (see Sect. 1), utilizes twin body tubes that are inclined to produce the stereo effect. Specimens are thus imaged using two separate compound microscope optical trains, each with an eyepiece, objective, and intermediate lens elements. A newer system, termed the common main objective (CMO), utilizes a single large objective that is shared between two individual optical channels consisting of a pair of eyepiece tubes and lens systems [4, 5]. Either type of microscope can be equipped with step-type individual lenses to change magnification, or a continuously variable zoom-type magnification system. Here we address the pluses and minuses of both the Greenough and common main objective stereomicroscope designs (Fig. 1).

2.1

Greenough

The Greenough stereomicroscope design consists of two identical and symmetrical optical systems, each containing a separate eyepiece and objective arranged in accurate alignment within a single housing (Fig. 1). A principal advantage of this design is the high numerical apertures (translating to better resolution) that can be obtained. Indeed, the objectives are very similar in configuration to those utilized in classical compound microscopes. In general, the lower portions of the body tubes, containing the slender objectives, are tapered and converge at the best focus of the object plane. The upper end of the body tubes projects a pair of images into the observer’s eyes, normally with a pair of standard eyepieces. The size, focus, rotation, and centering of the two images must be held constant within very tight tolerances, in order for the eyes to view

248

Erin E. Wilson et al.

x y z

Greenough Design

Common Main Objective (CMO) Design

Fig. 1 Comparison of Greenough and common main objective (CMO) microscope designs and illumination pathways through the optical train. Images of the inverting prisms may be misleading due to highly subjective perception of depth (prism orientation). Reprinted from Ref. [6] by permission of © Florida State University. Bottom Inverting (erecting) prisms of Reichert Mak stereomicroscope (Greenough design) reprinted from a catalogue (4. “Mak” K I–II D, 11/62) published by a no-longer-existing company (C. Reichert Optische Werke AG, Vienna)

essentially the same scene. The lone departure is the slightly different viewing angle at which each image is projected onto the retina. A pair of erecting prisms or mirror system is utilized to de-rotate and invert the (already inverted) magnified image received from the objectives and present it to the observer as it would appear (in terms of orientation) without a microscope [4]. The body tubes are built to provide a straight line-of-sight in some designs (Fig. 1 left), while others enlist the aid of additional prisms to allow inclination of the tubes and a more natural (“ergonomic”) viewing position for the microscopist. Because the imageforming light rays pass through the complex lens system on center, the quality of the image is symmetrical about its center, as is the case with most compound microscopes. Additionally, correction for optical aberrations in Greenough-type microscopes is less difficult

Stereomicroscopy in Neuroanatomy

249

Fig. 2 Perspective distortion (dome or globe appearance) of the specimen results from a combination of keystone and barrel distortion. Note the distortion of the left- and right-eye images is slightly exaggerated here, and the perspective distortion is inherent to CMO design stereomicroscopes only, and almost absent in high-end models. The coin images are reprinted from Ref. [6] by permission of © Florida State University

than with CMO designs, because the lenses are smaller, axially symmetrical, and do not rely heavily on light rays passing through the objective periphery. Due to the oblique separation of each body tube from a common axis (by half convergence), a distortion artifact arises in the Greenough microscope [7]. Referred to as the “keystone effect,” it causes the area on the left side of the right eye to appear slightly smaller than that on the right-hand side of the same image, and of course the reverse is true for the left eye’s image (Fig. 2). Keystone distortion arises from the fact that the intermediate images produced by each body tube are inclined with respect to the specimen plane, and tilted relative to each other, so that only the central regions are in simultaneous focus at identical magnifications. The result is that peripheral portions of the viewing field are focused either slightly above or below the actual specimen plane and have very small differences in magnification, although the human eye usually compensates for this effect and it is often not noticeable to the microscopist. During prolonged observations periods, however, fatigue and eyestrain can be accelerated by the keystone effect. The tiny change in magnification and focus across the viewing field in Greenough stereomicroscopes might be noticed in a photograph or video image produced through one side of the instrument, especially if the object is primarily flat and rectilinear. In photomicrography (or digital imaging), focus discontinuities brought on by the inclination angle are easily compensated by

250

Erin E. Wilson et al.

tilting either the specimen or one of the beam paths so that the optical axis is perpendicular to the specimen. When undertaking measurements with a reticle, the linear eyepiece grid should be oriented in the “x” direction (Fig. 1). Another solution is to tip the specimen five or six degrees so that it is perpendicular to the optical axis of the channel actually employed in photomicrography. 2.2 Common Main Objective (CMO)

Stereomicroscope designs with a CMO are based on the refracting action of a single, large diameter objective lens, through which both the left and right channels view the specimen [7]. Each channel operates as an independent optical train parallel to the other (this is the reason they are also known as parallel microscopes; see Fig. 1), and there is collimated light between the individual channels and the objective (the image is projected to infinity, i.e., the optics is infinity-corrected). Such an arrangement makes it easier to assure that convergence of the left and right optical axes coincide with the focal point in the specimen plane. Because this parallel axis arrangement is usually extended to include the eyepieces, the left and right images are viewed by the microscopist’s eyes with little or no convergence. An important advantage of the CMO system is that the optical axis of the objective is perpendicular to the specimen plane, and there is no inherent tilt of the image at the eyepiece focal plane. Despite the fact that in most situations there are the usual 10 to 12 of convergence at the specimen, the brain is not used to interpreting three-dimensional images without perceived convergence (at eyepiece level). This leads to a unique anomaly that is inherent in CMO stereomicroscopes: the center portions of the specimen appear to be slightly elevated, so that a flat specimen now appears to have a convex shape. For example, a coin would have the appearance of being thicker in the center. Artifacts of this nature are referred to as a perspective distortion but should not cause concern unless the microscope is utilized to judge flatness or height (Fig. 2). Specimens with complex or rounded shapes, while displaying a certain amount of perspective distortion, often do not appear to be distorted when viewed through the stereomicroscope. Perspective distortion is sometimes referred to as doming or the globular effect, and results from a combination of keystone and barrel (the opposite of pincushion) distortion. As an example presented in Fig. 2 is a slightly exaggerated illustration of how a US Lincoln penny, a disc-shaped flat coin, would appear in a stereomicroscope with severe perspective distortion. The original penny is shown at the top of the illustration to have a flat surface. Just beneath are the images projected simultaneously by the microscope to both the left and right eyes, which demonstrate an asymmetrical barrel distortion directed toward the central axis of the microscope. The final result is perception of a dome- or globe-shaped object when the images from both eyepieces are projected onto the retinas

Stereomicroscopy in Neuroanatomy

251

and fused together in the brain. Most high-end research grade CMO stereomicroscopes produced by the major manufacturers have virtually eliminated this artifact, but it still occurs in some less expensive microscopes. A frequent artifact encountered with CMO stereomicroscopes is that small amounts of the off-axis aberrations, including astigmatism, coma, and lateral chromatic aberration, appear in the center of each image [4]. Circumstances of this sort occur because each optical channel is receiving a significant part of light rays from an off-center region of the large objective. The effect is generally not noticed when both eyes are used to view the specimen, but a photomicrograph may have asymmetric geometry across the field. For the most part, the chromatic aberrations are difficult and expensive to correct, especially considering the large size and volumes of glass used in manufacture of the objectives. Some CMO stereomicroscope designs have made this a nonissue by providing the facility to offset the large objective, positioning its center on the axis of either the left or right side optical channel. Other microscope designs even provide a means for replacing the large objective with a conventional infinity-corrected objective that can be utilized to view and photograph specimens at high magnifications (and numerical apertures). The infinity optical system is considered to be the greatest design feature as well as the main practical advantage of a CMO stereomicroscope [4]. A collimated light pathway, with two parallel axes for the channels, exists between the objective and removable head/observation tube assembly (labeled “infinity space” in Fig. 3A). The optical system allows the effortless introduction of accessories, including beamsplitters, coaxial episcopic illuminators, photo or digital video intermediate tubes, drawing tubes, eyelevel risers, and image transfer tubes into the space between the microscope body and the observation head. It is also possible to place these accessories in the space between the objective(s) and the zoom body, although this is rarely done in practice. Because the optical system produces a parallel bundle of light rays between the body and microscope head, the added accessories do not introduce significant aberrations or shift the position of images observed in the microscope. Such versatility is not available in stereomicroscopes designed around the Greenough principles (with the two optical channels not being parallel), and the same would be the case even if their optics was infinity-corrected, which (to the best of authors’ knowledge) is not the case anyway. 2.3 Greenough and CMO Compared

Because there are no universally accepted criteria for comparing performance between the stereomicroscope systems, it is difficult to determine which of the two designs (CMO or Greenough) is superior. CMO stereomicroscopes, in general, have a greater light-gathering power than the Greenough-design and are often

252

Erin E. Wilson et al.

Infinity space

Measuring reticle

A

B

Fig. 3 Anatomy of a stereomicroscope. (A) Cutaway view of the internal components and optical train of a typical common main objective (CMO) stereomicroscope. Assorted stereomicroscopy attachment lenses spanning a wide range of magnifications are shown at the bottom. (B) Zoom stereomicroscope configuration with continuously variable magnification range. Both images are reprinted from Ref. [6] by permission of © Florida State University

more highly corrected for optical aberration. Some observations and photomicrography might best be conducted utilizing a CMO microscope, while other situations may call for features exclusive to the Greenough design. Greenough microscopes are typically employed for “workhorse” applications, such as dissecting biological specimens, soldering miniature electronic components, and similar routine tasks. These microscopes are relatively small, inexpensive, very rugged, simple to use, and easy to maintain. In contrast, CMO microscopes are generally utilized for more complex applications requiring advanced optical and illumination accessories. The wide spectrum of accessories available for these microscopes lends to their strength in the research arena. Another consideration is the economics of microscope purchase, especially on a large scale. A CMO stereomicroscope can cost several times more than a Greenough microscope, which is a chief consideration for customers who may require dozens or even hundreds of microscopes.

Stereomicroscopy in Neuroanatomy

3

253

Objectives and Eyepieces The total magnification of a stereomicroscope is the product of the objective and eyepiece magnifications, plus that contributed by any intermediate or external auxiliary magnifying lens systems.

3.1

Objectives

In the simplest microscopes, the objectives (or single objective in a CMO design) are permanently mounted in the lower body housing, and magnification can only be altered by introducing eyepieces of varying power. Slightly more complex microscopes have interchangeable objectives that allow total magnification factors to be adjusted by using a higher or lower power objective or by substitution eyepieces of differing magnification. Objectives in these models are mounted by screw threads or clamps, which enable relatively quick changeover to a new magnification. Mid-level stereomicroscopes are equipped with either a sliding objective housing, or a rotating turret accommodating several matched sets of objectives to produce varying magnification factors. The operator simply twists the turret to position a new paired set of objectives beneath the channel tubes [4]. Microscopes utilizing this design were once very popular, but they are rarely manufactured today. Stereomicroscopes offering the highest levels of quality come equipped with a zoom lens system or a rotating drum containing Galilean telescopes that are utilized to increase and decrease overall magnification [4, 5, 7]. The rotating drum system functions as an integral intermediate tube containing several paired sets of lenses that can be installed into the optical pathway by rotating the drum. In most models, positive de´tentes are employed to act as “click stops” to secure the lens mounts into correct alignment, and are marked to notify the operator of the new magnification factor. The drum usually has a pair of empty lens mounts that are devoid of auxiliary lenses and can be positioned into the optical path to allow use of the objective and eyepiece combination without additional magnification. Zoom lens systems (Fig. 3B) provide a continuously variable magnification range [8]. Their key advantage is the elimination of the blank-out that occurs with possible visual loss of spatial relationships between specimen features when magnification is changed in discrete, stepped settings. Some of the older literature refers to zoom systems as pancratic (“almighty”) systems after the Greek words pan for “all” and kratos for “power.” Zoom ratios vary between 4:1 and 15:1, depending upon the microscope age, manufacturer, and model. In general, a zoom lens system is comprised of a minimum of three lens groups, enlisting two or more elements (lenses) for each group. One element is fixed within the channel tube, while the other two are smoothly translated up and down by

254

Erin E. Wilson et al.

precision cams. Following the zoom system, inverting prisms and a tube lens are utilized to erect and relay the image before projecting it into the eyepieces. Several of the newer stereomicroscope models employ a positive click-stop that alerts the microscopist at selected magnification positions in the zoom range. A distinction of this sort is essential for calibration of the magnification level at a given power step, a feature often found useful when making measurements. In the early days of stereomicroscopy, zoom lens systems were equipped with a magnification range of approximately 7
to 30
. The magnification factors slowly increased as optical performance improved in this class of microscopes, and more recent student microscopes now feature zoom ranges between 2
and 70
. Mid-level stereomicroscopes have zoom magnification factors with an upper magnification limit between 250
and 400
, while high-end research microscopes come with zoom systems that can reach in excess of 500
in magnification. The wide magnification range is complemented by a depth of field (Tables 1 and 2) and working distances that are significantly larger than are found in compound microscopes having equivalent magnifications. The working distance on modern stereomicroscopes varies between 20 and 140 mm, depending upon the objective magnification and zoom ratio [4, 5, 7]. With the addition of specialized auxiliary attachment lenses, working distances of 300 mm or more can be achieved. Viewing field diameters are also much wider than those attainable with compound microscopes. Auxiliary attachment lenses can be fitted to the objective barrel on specially designed stereomicroscopes (Fig. 3A). For the most part, the attachment lenses are threaded to rotate into a matching thread set on the front of the objective barrel. Other (and much Table 1 Typical stereomicroscope (Nikon SMZ1500) objective specifications

a

Objective

Color codea

Numerical apertureb

Working distance (mm)

ED Plan 0.5


Red

0.045

155

ED Plan 0.75


Yellow

0.68

117

ED Plan 1


White

0.09

84

ED Plan 1.5


Green

0.14

50.5

ED Plan 2


Blue

0.18

40

HR Plan Apo 0.5


n/a

0.066

136

HR Plan Apo 1


n/a

0.131

54

HR Plan Apo 1.6


n/a

0.21

24

Color coding varies from company to company At the highest zoom setting (
11.25)

b

Stereomicroscopy in Neuroanatomy

255

Table 2 Depth of field in a modern stereomicroscope Depth of field (micrometers) Objective

Zoom factor

Numerical aperture

f-number

10


15


20


30


HR Plan Apo 1
(Nikon)

0.75 1 2 4 6 8 10 11.25

0.023 0.029 0.052 0.085 0.104 0.118 0.128 0.131

21.7 17.2 9.6 5.9 4.8 4.2 3.9 3.8

1348 820 239 80 48 35 28 26

1072 655 193 66 41 30 24 22

934 573 170 59 37 27 22 21

796 491 147 52 33 25 21 19

Note the depth of field depends on both the zoom factor (i.e., numerical aperture/f-number of the objective lens) and the eyepiece magnification (10
to 30
)

A

B

OBJECTIVE

FUNDUS LENS

VESSELS

50 µm

Fig. 4 The use of an ordinary fundus lens (in lieu of a dedicated attachment lens) to study fish eye retina under a stereomicroscope. (A, B) Optical setup showing the fundus lens held immediately above the eye of anaesthetized fish by a micromanipulator arm. Stereomicroscope objective is shown in A. (C) Vasculature (outlined in white) is employed as an in vivo landmark to navigate within the eye of a transgenic zebrafish expressing GFP (green) and mCherry (red) in UV- and blue-sensitive photoreceptor cells (cones), respectively. All three images are excerpts from Ref. [9], based on a Creative Commons license

rarer) versions attach to the barrel with a clamping device. Such lenses enable the microscopist to either increase or decrease the magnification of the primary objective. When image quality is not the most pertinent factor, attachment lenses are useful because optical corrections cannot be as accurately performed due to the fact that the lens is not mounted in the identical position each time it is attached. In addition, attachment lenses modify the objective working distance (the distance between the specimen and the objective front lens element). A lens that increases (decreases) the microscope magnification will also simultaneously render a shorter (longer) working distance. In lieu of a nondedicated auxiliary lens, an ordinary fundus lens may be used, for example, to visualize retina in living fish (Fig. 4).

256

3.2

Erin E. Wilson et al.

Eyepieces

Modern stereomicroscopes are equipped with standardized widefield high-eyepoint eyepieces that are available in magnifications ranging from 5
to 30
in approximately 5
increments. Most of these eyepieces can be utilized with or without eyeglasses, and protective rubber cups are available to prevent contact between the microscopist’s eyeglasses and the lens of the eyepiece. Eyepieces generally are equipped with a diopter adjustment to allow simultaneous focusing of the specimen and measuring reticles with both eyes, and binocular microscope heads now have movable tubes that enable the operator to vary the inter-pupillary distance between eyepieces over a range of 55–75 mm. The inter-pupillary adjustment is often accomplished by rotating the prism bodies with respect to their optical axes. Because the objectives are fixed in their relationship to the prisms, the adjustment does not alter the stereoscopic effect. Microscopists who wear eyeglasses to correct for shortsightedness and differences in vision between eyes do not have to remove them for microscopy. Eyeglasses worn only for close-up work should be removed during observation because the microscope produces the image at some distance. The field of view (FOV) is determined by the objective magnification and the size of the fixed field diaphragm in the eyepiece [10]. When the magnification is increased in a stereomicroscope, the size of the field of view is decreased if the eyepiece diaphragm diameter is held constant. Increasing the size of the eyepiece diaphragm opening (this must be done during manufacture) will increase the field of view at fixed magnification. In most stereomicroscope and compound microscope eyepieces, the physical diameter of the field diaphragm (located either in front or behind the eyepiece field lens) is measured in mm and referred to as the field number (often abbreviated as FN), and should not be confused with the f-number described in Subheading 4. The actual physical size of the field diaphragm and apparent optical field diameter can vary in eyepiece designs having a field lens below the diaphragm. Measuring and photomicrography reticles are placed in the plane of the eyepiece field diaphragm, so as to appear in the same optically conjugate plane as the specimen. The viewing field diameter is quantitatively determined by dividing the field number (FN) of the eyepiece, usually inscribed on its housing, by the magnification power of the objective. Included in the calculation should also be the zoom setting and any additional accessories inserted into the optical path that may have a magnification factor. However, the eyepiece magnification is not included, which is a relatively common mistake made by novices in microscopy. When a wider field of view is desired, the microscopist should choose eyepieces with a higher field number. In the lower magnification ranges, stereomicroscopes have substantially larger fields of view than classical laboratory compound microscopes. The typical viewing field with a 10
eyepiece and a

Stereomicroscopy in Neuroanatomy

257

low power objective (0.5
) is around 65–80 mm in diameter, almost double the size observed (about 40 mm) with a compound microscope at comparable magnification. Strong illumination is required, which is often challenging.

4

Resolution and Depth of Field The overall resolving power is limited by the optical parameters of the microscope as well as the density of photoreceptive cells and nerve endings in the retina, much like the limiting grain size in photographic film or the pixel density in a charge-coupled device (CCD) digital camera. Resolution of the stereomicroscope itself is determined by the wavelength of illumination and the numerical aperture (NA) of the objective. The smallest distance discernible by the microscope itself between two specimen points is typically given by the Rayleigh (1a) or Abbe/Sparrow (1b) criterion [10], Resolution ðd1 Þ ¼ 0:61  λ=ðn  sin θÞ

ð1aÞ

Resolution ðd2 Þ ¼ 0:50  λ=ðn  sin θÞ

ð1bÞ

where d1 is the smallest practically resolvable distance between two points, d2 is the theoretical resolution limit, λ is the illumination wavelength (a mixture centered around 550 nm when white light is employed), unless fluorescence is used (in which case λ is the emission wavelength), n is the refractive index of the medium immediately adjacent to the objective (which is always air in stereomicroscopy, n ¼ 1), and θ is the objective one-half angular aperture at focus. The term “n · sin θ” is referred to as numerical aperture (NA). As an example, a modern Nikon SMZ1500 stereomicroscope equipped with a 1.6
apochromatic objective with a numerical aperture of 0.21, will have a resolution of approximately 1.6 μm when the specimen is illuminated with light having a wavelength of 550 nm. Objective lenses manufactured for CMO stereomicroscopes typically vary in magnification from 0.5
to 2
, with three or four intermediate values. Parameters of typical stereomicroscope objectives at varying magnification are presented in Table 1. In the past, several manufacturers have assigned color codes to their stereomicroscope objective magnification values. The sole determining factor of the resolving power of stereomicroscope objectives is the objective numerical aperture, and it is not influenced by optical parameters of the eyepiece. Overall resolution will not be affected when exchanging 10
eyepieces for 20
or higher magnification eyepieces, although specimen detail that is not visible at the lower magnification will often be revealed when the eyepiece magnification is increased. This is due to limited resolving power of camera (which may result in undersampling)

258

Erin E. Wilson et al.

or human eye (ca. 1 angular minute). The highest power eyepieces (30
or higher) may thus yield the so-called empty magnification. In order to gauge and compare the performance of one microscope to another, the resolution value is often expressed in terms of black and white line pairs per millimeter (lp/mm). In the case of the Nikon 1.6
/0.21 objective discussed above, the resolution approaches 630 lp/mm (~1.6 μm) under optimum conditions, when using white light. Ranging in power from 0.3
to 2
, auxiliary attachment lenses can alter the working distance and resolving power of a stereomicroscope optical system. Roughly speaking, the resolving power influence is proportional to the magnification factor of the attachment lens. The field of view diameter is inversely proportional to the magnification factor, while the depth of field decreases with the numerical aperture and eyepiece magnification (Table 2). Working distance also decreases with magnification, but the function is not linear and difficult to compute. In addition, use of the auxiliary lenses will not have significant impact on image brightness (in most cases). Objective lenses designed for general photography are rated with a system that is based on f-numbers (abbreviated “f”) rather than numerical aperture (NA), see Table 2. These two values actually express the same quantity: the light gathering ability of a photography lens or a microscope objective. The f-number can also be calculated by dividing the focal length of the lens system by the aperture diameter. f  number ðf Þ ¼ 1=ð2  NAÞ

ð2aÞ

NA ¼ 1=ð2  f Þ

ð2bÞ

In a stereomicroscope objective itself, the aperture diameter is fixed. This is similar to the situation with conventional compound microscope objectives, with the exception of older, high-NA ones; some of them have a built-in iris diaphragm making it possible to reduce the aperture diameter (and thus NA), with the aim to make the objective more suitable for dark-field imaging. As the stereomicroscope magnification is increased or decreased by changing the zoom factor, the focal length (and thus also the f-number) is also altered accordingly. Two objectives having the same magnification can have different focal lengths due to variations in tube lens and zoom-channel aperture specifications. In late model microscopes objective focal lengths have been reduced in order to increase the total system numerical aperture (i.e., improve resolution). As an example, the Nikon SMZ-U stereomicroscope 1
objective has a focal length of 100 mm, while the later model SMZ1500 microscope employs a focal length of 80 mm for an objective having similar magnification and optical corrections. The difference between the two

Stereomicroscopy in Neuroanatomy

259

microscope designs is the size of the zoom system aperture. When interchanging objectives having the same magnification but different focal lengths, an additional factor must be introduced into total magnification calculations to correct for the focal length differences. In all forms of optical microscopy, the concept of depth of field is perhaps most important in stereomicroscopy, and is strongly influenced by the total magnification of the instrument, including the contribution from both the objective and auxiliary attachment lenses [4]. At a magnification of 50
, using a 1
objective (numerical aperture 0.10), 10
eyepieces, and a zoom factor of 5, the depth of field exhibited by a typical stereomicroscope is approximately 55 μm. If a 2
attachment lens is added, the new magnification will be 100
, but the depth of field drops to about 14 μm, a substantial decrease from the value (55 μm) without the auxiliary lens. Unless resolution is important it is wiser to change the eyepiece magnification from 10
to 20
to achieve the added magnification so as to partly retain the larger depth of field value. Increasing the objective numerical aperture through enhanced optical correction (e.g., from achromat to apochromat) will also produce a modest decrease in field depth. Depth of field values for a Nikon apochromatic 1
objective are presented in Table 2, as a function of zoom factor and eyepiece magnification. Some stereomicroscopes combine high resolution with high depth of field by using different optics in each of the two channels of the CMO objective [11]. Enhancing the depth of field (and contrast) can be achieved by reducing the size of the double iris diaphragm positioned between the objective and the eyepieces (Fig. 3A). A wheel or lever is used to open and close this diaphragm in the microscope body housing. There are actually two diaphragms, one for each of the channels, in the CMO stereomicroscope design. Depth of field and numerical aperture variations, as a function of diaphragm opening size, is presented in Table 3 for the Nikon plan apochromatic 1
objective at the highest zoom magnification factor (11.25). Closing the iris diaphragms will of course also reduce overall light intensity, with implications for exposure times in photomicrography. At some point, depending upon the optical configuration of the microscope, the image begins to degrade and specimen details exhibit diffraction phenomena while minute structural details (featuring high spatial frequencies) disappear as higherorder diffraction maxima are blocked and cannot contribute to image formation. The best setting is a balance between maximum specimen detail and maximum contrast as seen in the eyepieces, on film or in digital images.

260

Erin E. Wilson et al.

Table 3 Depth of field and numerical aperture (NA) in stereomicroscope objective (same as in Table 2), as a function of closing down the double-iris diaphragm between the eyepieces and the objective (NA of 0.131 represents a fully opened diaphragm at the highest zoom setting [
11.25]) Depth of field (micrometers) Numerical aperture

10


15


20


30


0.131

26

22

21

19

0.095

44

39

37

35

0.063

89

83

79

76

Eyepiece magnification is specified (10
to 30
)

Table 4 Main illumination schemes in stereomicroscopy, with light sources listed Illumination direction and source Illumination type (Modality)

Transmitted (diascopic)

Incident (episcopic or reflected)

Bright field

Typically built-in Fig. 5A, C

Built-in or removablea Figs. 7B, 8, and 9C

Oblique (relief) contrastb

Typically built-inc Fig. 5B, D

Built-in or removabled Fig. 7B

Dark field

Built-in or removablee Figs. 5E and 6

Typically removable Fig. 7B, C

Fluorescence

n/af

Figs. 4C and 9A, B, D

a

Vertical and coaxial illuminators are especially suitable to routinely inspect specimens in exactly the same way Utilizes off-axis illumination c Marketed under various names such as “oblique coherent contrast” (Nikon) or “Rottermann contrast” (Leica). The origin of the name “Rottermann” remains elusive even after repeated questioning of Leica representatives d Oblique episcopic (i.e., inclined incident) illumination provided, for example, by Nikon G-ICIL LED (oblique and coaxial episcopic illuminator in a single setup) e An example of the removable option is, for example, the LED ringlight placed under the specimen on a transparent support (e.g., Petri dish) f A theoretically possible but impractical option b

5

Illumination Schemes Selection of a method for specimen illumination (as summarized in Table 4) and the determination of its effectiveness in revealing features of interest is one of the more critical aspects of observation and applies to all forms of optical microscopy. Stereomicroscopes are often utilized to examine specimens under both incident (episcopic) and transmitted (diascopic) illumination schemes, employing a variety of appropriately positioned light sources.

Stereomicroscopy in Neuroanatomy

261

There are situations in which incident and transmitted light sources are combined to take advantage of particular specimen characteristics in a manner that most effectively reveals the features of interest. Many of the specimens examined with stereomicroscopes are three-dimensional, and require a significant degree of creativity on the part of the microscopist to most effectively illuminate the specific details of interest. Despite their obvious differences, a number of similarities exist between the challenges faced in illumination for stereomicroscopy and those encountered in close-up or macrophotography using conventional camera and lens combinations. The lower magnifications utilized with stereomicroscopy overlap the reproduction ratios possible using traditional camera lenses coupled to extension devices, or specialized macro lenses, and many specimens can be effectively imaged with either type of equipment. In addition, many of the illumination techniques that have been proven useful in macrophotography can be applied to the stereomicroscope, and vice versa. Techniques used in stereomicroscopy commonly vary considerably from those developed for “standard” compound microscopes employed in conventional optical microscopy, especially with regard to many of the illumination strategies. Because of the far-ranging variety of specimens the stereomicroscope is designed to accommodate, there exists no single optimum illumination strategy that is the correct choice. Each specimen under examination can be illuminated by a variety of different schemes, and employing a seemingly infinite number of variations or combinations. For a given specimen or object, although there may be several possible illumination schemes capable of producing acceptable results, a single approach may be discovered that, after extensive tweaking, produces exceptional results. Careful consideration of the characteristics of the specimen must be made in selecting an illumination strategy to suit the needs of visual observation, photomicrography, or digital imaging. The most important characteristic is usually the degree of opacity of the specimen, and this will determine the basic type of illuminator that should be employed—incident (episcopic), transmitted (diascopic), or in some cases, a combination of both. For the most part, opaque specimens are typically illuminated from above (with incident light), using orientations ranging from on-axis (parallel to the microscope optics) to highly oblique (up to a 90 incident angle from the optical axis), as required to reveal the features or characteristics of interest. Once it has been determined which general category of illuminator best fits the specimen’s opacity, a number of other factors should be considered to further refine particular variations on the basic illumination scenario that will likely produce the desired results.

262

Erin E. Wilson et al.

In most cases, opaque specimens benefit from incident (reflected) illumination, while translucent and transparent objects usually produce the best results with some variation of transmitted illumination (bright-field, polarized, oblique, or dark-field). However, this is not always the case as translucent objects may benefit from having at least a portion of their illumination directed from a source placed above them. Aside from opacity, a number of secondary factors should be considered in planning a lighting strategy, including the basic physical characteristics of the specimen, the type of information that is required from the examination, digital or photographic imaging requirements, and how the information will be utilized. In choosing and configuring illumination in a manner that will reveal the desired information, it is important to consider the topography and composition of the specimen. Nematodes, organoids, and crustacean eyes (to name but a few examples) all behave differently with regard to their appearance under different lighting conditions. Specific environmental requirements of some specimens may affect their suitability for illumination with various source types. For example, excised living tissue and living aquatic organisms typically require immersion in buffer/water during observation. Some specimens (e.g., cornea) may produce artifacts by reflecting images of the light source (or sources) into the microscope objective (Fig. 9C). The most difficult specimens may even require a special lighting technique just to be rendered visible. Sensitivity of the specimen material to heat or ultraviolet light is another important factor that must be considered as both are significant emission components of some illumination sources. Light sensitivity may require limiting the amount of time that the specimen is illuminated. If so the choices of possible illumination techniques become far more restricted. A similar problem occurs if a specimen is being studied for observation or recording of a transient, or short-lived, phenomenon or property, in which case the intensity of the illumination may become the principal factor in choosing a lighting strategy. Illumination strategy is often strongly influenced by the purpose of the microscope examination, or the specific kind of information that is required from the specimen being studied. A variety of schemes may be necessary to reveal minute details, larger features, or gross characteristics. Depending upon the information that must be obtained from a given specimen, the illumination technology employed can be extremely simple, or much more elaborate, and may require a combination of techniques. For example, if the only important property of a specimen is its color, then the lighting employed can be very simple physically, and is only required to provide accurate color rendition. If both color determination and topography are important, then more attention must be paid to lighting geometry so that all features of interest are revealed.

Stereomicroscopy in Neuroanatomy

263

Another important factor that must be taken into consideration in choosing specimen illumination is the photographic or digital imaging requirements. If a traditional film camera is employed to record images, the color temperature (and possibly other spectral characteristics) of the light source must be appropriate for the film used, in order for the specimen to be accurately represented. The intensity of the lighting must also be adequate to ensure exposures that are of reasonable duration for the camera/ film combination employed, which is an especially important factor in a clinical laboratory settings. Digital image capture systems require many of the same considerations as film systems, although white balance adjustment on the imaging device (digital camera) allows considerable latitude in matching camera response to the color characteristics of various light sources. Should video recording be conducted, illumination intensity may be an even greater issue. Finding the optimal match of the microscope and specimen with an illumination system often depends to a large degree upon the skill and training of the operators that will use the system, and the setting or type of environment in which it will be employed. In a clinical laboratory where only one type of tissue is routinely examined, the simpler lighting systems that are preset to a fixed configuration may be preferred, in contrast to a research environment typically requiring a more flexible lighting system, and more skilled staff. Ergonomic characteristics of the combined microscope and illumination system are an important consideration in the lighting systems since the design must be conducive to repetitive operations. Comfort and ease of use are undoubtedly important in any serious application of microscopy, although the work environment that perhaps best validates this concern is the clinical laboratory where a fatiguing or difficult-to-use lighting configuration can reduce the accuracy of critical specimen analyses, even when conducted by a skilled microscopist. 5.1 Transmitted Illumination

Transmitted bright-field illumination (Fig. 5A, C) is one of the most commonly utilized observation modes in all forms of optical microscopy, including stereomicroscopy, and is ideal for fixed, stained specimens or other types of samples having high natural absorption of visible light [16]. Collectively, partly (or fully) absorbing specimens, typically imaged under bright-field illumination, are termed amplitude objects because the amplitude (square root of intensity) of the illuminating light is reduced when light passes through the specimen. In transmitted illumination, the phase relationship between light rays passing unaffected through and around the partly absorbing object such as stained cell (illuminating or direct light, also termed the surround light wave) may be interpreted as being 180 out of phase with light diffracted by the

264

Erin E. Wilson et al.

Bright field illumination Off-axis (relief) illumination

A

B

BRIGHT FIELD ILLUMINATION

OFF-AXIS ILLUMINATION

~1 mm BRIGHT FIELD ILLUMINATION

C

Magnification unknown

OFF-AXIS ILLUMINATION

D

DARK-FIELD ILLUMINATION

E

UNDISCLOSED MODALITY

F

100 µm

Fig. 5 Off-axis (relief or oblique) transmitted illumination in stereomicroscopy is typically achieved by a sliding diaphragm in the illumination base, and marketed, for example, as “oblique coherent contrast” (Nikon) or “Rottermann contrast” (Leica). (A, B) Goat hair under horizontal (A) and vertical (B) illumination (relative to the fibers). (C–E) Hookworm (Ancylostoma caninum) under a stereomicroscope. The dark-field modality (E) represents an extreme form of the off-axis variant (D). (F) Another species of hookworm (Uncinaria spp.). Images A–E (including the optical setup) are adapted from Ref. [6] by permission of © Florida State University, image F is reprinted from Ref. [12] by permission of © Springer

cell (not taking into account phase shift by the same unstained cell). This results in destructive interference between the direct and diffracted light at the image plane, to produce a visible image of the cell appearing dark (due to absorption) with high contrast against bright background. In contrast, transparent specimens that do not absorb light, but produce only a phase change of light passing through are termed phase objects. Such specimens are often almost invisible (suffer from poor contrast) because the human eye (or camera) is insensitive to phase shifts of light waves. These are due to differences in

Stereomicroscopy in Neuroanatomy

265

C

B

Eyestalk

A

Fig. 6 Dark-field transmitted illumination in stereomicroscopy. (A) Transmitted bright-field and dark-field illuminator (“condenser”). (B) Transmitted dark-field illuminator (“condenser”). (C) Unstained ghost shrimp (Macrobrachium jelskii) eye in dark-field of unknown configuration. . Chromatophores (sometimes categorized as “paraneurons”) in the eyestalk are clearly discernible owing to their red and yellow pigmentation (erythrophores and xanthophores, respectively). Images A and B are reprinted from Ref. [6] by permission of © Florida State University), image C from Ref. [15] by permission of © Vito´ria Tobias Santos, Federal University of Rio de Janeiro (Macae´ campus)

optical thickness (physical thickness multiplied by refractive index difference between the specimen and its surrounding medium). A typical example is unstained living tissue which appears almost transparent under bright-field illumination (Fig. 5C). Optically thicker phase specimens can generate acceptable contrast (Fig. 5A) because parts of them are not exactly in focus. This phenomenon is well known to electron microscopists who routinely defocus their images a bit to gain contrast, at the expense of resolution. In stereomicroscopy, phase objects are to be examined under oblique (off-axis) or dark-field illumination described in Sects. 5.3 and 5.4, respectively. Other optical-contrasting modalities (phase, schlieren, modulation, or differential interference contrast) are not available in stereomicroscopy (cf. Subheading 5.2). However, a single-objective stereoscopic phase-contrast setup is described in Refs. [7, 8] in Chapter 10. An example of transmitted-light darkfield illumination is shown in Fig. 6. 5.2 Incident Illumination

One of the more important considerations in developing an incident lighting strategy for stereomicroscopy is the working distance of the microscope objective, which can seriously restrict the

266

Erin E. Wilson et al.

flexibility in positioning incident illuminators. Measured between the objective lens and the specimen, this distance covers a range of several centimeters (for lower aperture and magnification objectives) to only a few millimeters for the highest numerical aperture objectives. In the studio setting of general photography, the photographer has a significant amount of latitude in placing lights in nearly any arrangement that is necessary to fulfill the desired lighting effect. In contrast, the size of the “studio” under a stereomicroscope objective may measure just a few centimeters or millimeters, and impose severe limitations on the choices available in the lighting scheme [4]. A small working space not only restricts the type of illuminator that may be used, but also the range of angles from which light can reach the specimen field. Limited area between the objective front lens and specimen may force placement of illuminators farther off-axis than desired, and often prevents the elimination of shadows on rough-surfaced specimens. Microscopists have a vast range of illumination sources available and fiber optic illuminators are probably the most versatile and popular. Many different light source designs, fiber types, configurations, and accessory attachments are available (Figs. 7A, B and 8). A fiber optic light system can be configured to meet the stringent requirements of almost any application. Typically powered by high-intensity tungsten-halogen lamps, fiber optic illuminators are relatively bright sources, and by utilization of appropriate filters, can be color-balanced for video or still image recording. Configured as cold light sources (through the addition of infrared filters), fiber optic systems are much more suitable for investigations of heat-sensitive specimens than are basic incandescent illuminators. VERTICAL ILLUMINATORS (Fig. 8A) provide incident illumination (reflected brightfield) by the addition of a half-reflecting surface, which is placed beneath or above the microscope objective at a 45-degree angle to the optical axis [17]. The reflector directs light, from an illuminator placed at right angle to the optical axis, downward toward the specimen, while allowing light reflected from (and/or backscattered by) the specimen to pass back through the microscope optical system. In the stereomicroscope, half-reflecting mirrors are commonly employed to perform the beam-splitting function. Illuminators made for Greenough microscopes must be designed to accommodate each eye path (at specific angles to each other), and incorporate angled optical elements to satisfy this requirement. For a single light path (Fig. 8A), such as that utilized in photomacrography, the reflector can be simply a thick piece of glass. Vertical illuminators may incorporate either condensing lenses or diffusers between the light source and the half-reflecting mirror. In a condenser system, the rays from the light source are focused in

Stereomicroscopy in Neuroanatomy

A

267

B 1 2

3

C

0.5 mm

1 mm

0.5 mm

Fig. 7 Incident illumination in stereomicroscopy. (A) Circular multi-LED illuminator (just two LEDs at 0 and 180 are shown for clarity). (B) Circular illuminator (similar to A in design) combined with an external lamp (positions marked 1 to 3). (1, 2) Bright-field to off-axis (oblique) illumination. (2, 3) Off-axis (oblique) to darkfield illumination. (C) Dorsocranial surface of murine heart histochemically stained for acetylcholine esterase (intrinsic cardiac nerve plexus), illuminated from several directions (as in B2 and B3). Black dashed line denotes the sinoatrial node (SAN). .. Right-sided preganglionated nerves accessing the ► right ganglion cluster. .. DRA and RV subplexal epicardial nerves (I, II, and III) extending toward the SAN region. ►► Left ganglion cluster. (CS) Coronary sinus. (CV) Caudal cava vein. (DRA) Dorsal right atrial nerve plexus. (LA) Left atrium. (LAu) Left auricle. (LPV) Left pulmonary vein. (LV) Left ventricle. (MPV) Middle pulmonary vein. (RA) Right atrium. (RAu) Right auricle. (RCV) Right cranial vein. (RPV) Right pulmonary vein. (RV) Right ventricle. Images A and B are reprinted from Ref. [6] by permission of © Florida State University, image C from Ref. [13] by permission of © Elsevier

a fashion similar to incident-light Ko¨hler illumination. The illuminating rays converge, after being reflected from the beam-splitting mirror, at the exit pupil (rear aperture) of the objective lens. This type of system maximizes the numerical aperture of the

268

Erin E. Wilson et al.

Fig. 8 Incident illumination setups in stereomicroscopy. (A) Vertical illuminators, with a mirror below or above the objective lens (Greenough and CMO design, respectively). A condenser may additionally be included between the light source and the mirror. (B) Coaxial illuminator fully utilizing the numerical aperture of the stereomicroscope (illuminating light passing through the objective and the zoom lenses). Both images are reprinted from Ref. [6] by permission of © Florida State University

illumination path, producing images with relatively high contrast, superior resolution, and good rendition of minute surface detail. COAXIAL ILLUMINATORS (Fig. 8B) are similar in concept to the vertical illuminators, and produce comparable results in specimen lighting characteristics. A major difference, however, is that the illumination path for coaxial illumination lies within the optical system of the microscope, rather than between the microscope and specimen. It is described as through-the-lens (or incident bright-field) illumination. In this technique, the stereomicroscope objectives and zoom lenses act as its own condenser, in a manner similar to the function of classical metallurgical microscopes. The main advantage of this technology is that the illumination system numerical aperture is altered in concert with that of the objective. As magnification is increased in the zoom body of the stereomicroscope, the numerical aperture also increases, for both the imageforming and illumination pathways. Such a manifestation offsets the loss in image intensity with increased magnification that is characteristic of other lighting techniques, such as simple vertical illumination. Consequently, the field of view through the eyepieces is equally bright throughout the magnification range of the zoom optical system. 5.3 Off-axis (Oblique) Illumination

As mentioned in Sect. 5.1, specimens that are almost transparent and colorless may be almost invisible when viewed in the stereomicroscope using traditional transmitted (diascopic) bright-field illumination techniques. However, if the illumination is directed so

Stereomicroscopy in Neuroanatomy

269

that it originates from a single azimuth and strikes the specimen from an indirect angle, otherwise known as off-axis or oblique (transmitted or incident) illumination (Figs. 5B, D and 8B), details in the specimen may be revealed with much greater contrast and visual clarity than when the light is allowed to pass directly through specimen features along the optical axis of the microscope [10, 16, 18]. Refractive index gradients in the specimen bring about redistribution of light intensity in the image in such a way that one half of each object (e.g., a cell cluster) appears brighter than the other. The objective back aperture additionally acts as the so-called schlieren diaphragm (a simple modulator) that makes this effect stronger, especially in optically thinner objects such as single cells. The result is a relief-like specimen pattern having regions displaying shadows and highlights, much like that observed with the Hoffman modulation contrast (itself a refined form of schlieren microscopy [19]) or differential interference contrast (DIC Nomarski) technique in compound microscopes [16]. A variety of illumination scenarios have been employed to provide off-axis (oblique) directional lighting for observation of specimens with the stereomicroscope. Simple diascopic bases are often equipped with a tilting mirror that can be adjusted to provide a certain degree of oblique illumination, but the light is not easily controlled and does not provide a uniform field of view. More complex microscope stands (or bases) have additional control possibilities, including tilting mirrors that are not restricted to a single axis, and sliding mirror assemblies that can be inserted and removed from the light path. Some models use one or more sliding baffles to restrict the illumination geometry and ensure that only off-axis light strikes the specimen. Acceptable to excellent results can be achieved with any of these techniques on a wide variety of specimens, but most require a significant degree of manipulation involving the illumination pathway, which is often difficult to quantitatively analyze or reproduce. Involving more art than science, the results achieved with off-axis illumination in stereomicroscopy are highly dependent on the skill, experience, and patience of the microscopist. A large number of lighting schemes have been developed, which further complicate the issues surrounding this technique for enhancing specimen contrast. A variety of names have been adopted such as “oblique coherent contrast” (Nikon, Fig. 5) or “Rottermann contrast” (Leica). The apparent three-dimensional effect of off-axis illumination techniques does not represent the actual specimen geometry or typography, and should not be employed to conduct measurements of specimen dimensions. The true value of the off-axis illumination image is in revealing transitions in refractive index or other optical path differences within the specimen that enable the morphology and internal structural arrangement to be more clearly understood.

270

Erin E. Wilson et al.

5.4 Dark-field Illumination

For rendering unstained and transparent specimens clearly visible, dark-field illumination (Figs. 5E, 6, 7B, C) is one of the more simple and popular techniques. Prime candidates for dark-field observation often have refractive indices very close in value to that of their surroundings and are difficult to image with conventional methodology. In such situations, other optical-contrasting modalities (listed in Sect. 5.1) may additionally be used in standard (i.e., nonstereo) microscopes [16, 18–20]. Cytoplasm of living cells has a refractive index (n) typically ranging from 1.35 to 1.4 [21], resulting in a rather small refractive index difference from the surrounding aqueous medium (n ¼ 1.33). If necessary, the refractive index difference (and thus the optical path difference) may be modulated (typically reduced) by adding, for example, bovine serum albumin or dextran to the bathing solution (typically buffer or water) [22]. Illumination of specimens by dark-field requires blocking out of the central light rays along the optical axis of the microscope, which ordinarily pass through the specimen [18, 23]. Blocking these light rays allows only those oblique rays originating at large angles to strike the specimen positioned on the microscope stage. In a compound microscope equipped with a simple (Abbe-type) condenser system an annular diaphragm (same as required for highpower phase-contrast objectives) can be used with low-power objectives to obtain a good-quality dark-field image. In some dedicated dark-field condensers, the top lens is spherically concave. If no specimen is present on the stage, and the numerical aperture of the condenser is greater than that of the objective, the oblique rays cross and miss entering the objective front lens because of their obliquity. The field of view will appear dark. When a transparent specimen is placed on a glass microscope stage and observed under dark-field illumination, the oblique light rays cross the specimen and are scattered (diffracted), reflected, and/or refracted by the specimen, and some of them enter the objective. Certain parts of the specimen then appear bright on an otherwise black background. In terms of the Fourier optics, darkfield illumination removes the 0th order from the diffraction pattern formed at the back focal plane of the objective. The result is an image formed exclusively from higher-order diffraction maxima (due to scattering by the specimen), and is also responsible for the main limitation of dark-field observation. Because the image is composed entirely from scattered light, it is rich in glare and often distorted to varying degrees, so it cannot be considered a faithful geometrical reproduction of the specimen. Indeed, this very limitation stimulated the invention of phase-contrast microscopy. The best candidates for dark-field illumination in stereomicroscopy include minute living aquatic organisms or bigger protozoa. Nerves in a living murine heart (Fig. 7C), or chromatophores (sometimes categorized as “paraneurons”) in shrimp (Fig. 6C) or fish can be comfortably visualized as well. Care should be taken in

Stereomicroscopy in Neuroanatomy

271

preparing specimens for dark-field microscopy because features that lie above and below the plane of focus, primarily fingerprints, dust, fibers, and cleaning residue, can also scatter light and contribute to image degradation. In review, dark-field microscopy is still an excellent tool for both biological and medical investigations. The technique can be effectively utilized to view a wide spectrum of biomedical specimens and can often reveal details that are not visible with other illumination methodology. 5.5 Fluorescence Illumination

A

Dichroic mirror

Until the recent past, fluorescence illumination (Fig. 9) was an option available only on research-level compound microscopes equipped with specialized high-numerical aperture objectives.

Mercury lamp

Dichroic mirror

Camera port

B

Specimen

Excitation Emission (fluorescence)

C

D

0.5 mm

0.5 mm

Fig. 9 Fluorescence in stereomicroscopy. (A, B) Epifluorescence illumination (by mercury arc lamp) and lightpaths. Dichroic mirror reflects excitation light and transmits emission light. (C, D) Stereomicroscopy investigation of nerve regeneration in the cornea of thy1-YFP transgenic mouse expressing yellow fluorescent protein. (C) Nonfluorescence image of cornea immediately after a lamellar dissection surgery (red dotted line). . No-dissection sites hosting major nerve trunks. (D) In vivo maximum intensity projection image of the fluorescently stained nerves. Nerves originating from .. central and .. peripheral stromal nerve trunks. .. Nerve branching. .. Corneal apex. Images A and B are adapted from Ref. [6] by permission of © Florida State University, images C and D reprinted from Ref. [14] by permission of © ARVO

272

Erin E. Wilson et al.

The need for this technique in stereomicroscopy has escalated with the introduction of genetically encoded (tagged) fluorescent proteins such as green fluorescent protein (GFP) [24]. For example, different photoreceptor cell types (expressing either GFP and mCherry) in retina [9] may be easily distinguished from each other (Fig. 4), or neurons expressing yellow fluorescent protein (YFP) in mice [14] may be used to monitor neuronal regeneration (Fig. 9D). Mercury arc-discharge and metal halide lamps are often utilized as an excitation source in fluorescence microscopy and the intense ultraviolet radiation they generate can damage the retina of the observer’s eye [16]. To avoid this situation, many microscope manufacturers include a protection device on the microscope body that filters ultraviolet light bathing the specimen on the microscope stage. Other safety precautions include ultraviolet barrier filters in the observation path and stray-light protection surrounding the lamphouse. In addition, dummy filter carriers are often inserted into unoccupied filter positions in sliders and rotating frames designed for color (fluorescence) filters. Fluorescence stereomicroscopes usually come equipped with a light stop (shutter) positioned somewhere between the lamphouse and the vertical illuminator to block damaging ultraviolet radiation from the lamp when specimens are not being observed or imaged. Such a stop should always be inserted into the light path when observations are not being made. While the illumination lightpath (as shown in Fig. 9B) may be the same as in coaxial illuminators (Fig. 8B) fluorescence setups are available that are similar to those vertical illuminators (Fig. 8A) in which the excitation light bypasses the objective. The advantage of such setups is no absorption of the excitation light in the objective. Reflexes, or hot spots, can occur in the lower portion of the viewing field when observing specimens using higher magnification apochromatic objectives (1.6
to 2
) in fluorescence stereomicroscopy. An artifact of this sort is usually only visible at the lower zoom ratios and often disappears when the zoom factor is increased. In most cases, reflex problems do not occur with lower magnification objectives (0.5
and 1
), regardless of the optical correction factor, and are usually absent from high magnification objectives of lower correction (achromats and plan achromats).

6

Summary Considering the wide range of accessories currently available for stereomicroscope systems, this class of microscopes is extremely useful in a multitude of applications. Stands and illuminating bases are available from all of the manufacturers, and can be adapted to virtually any working situation. There are a wide range of

Stereomicroscopy in Neuroanatomy

273

objectives and eyepieces, enhanced with attachment lenses and coaxial illuminators that are fitted to the microscope as an intermediate tube. Working distances can range from 3 to 5 cm to as much as 20 cm in some models, allowing for a considerable amount of working room between the objective and specimen. Magnification is often thought of as the most important criterion for judging the performance of an optical microscope. This is far from true, because the correct magnification is the one sufficient for the task at hand and should not be unnecessarily exceeded. Many classical investigations into the basis of cellular structure and function are best conducted with classical transmitted- and incident-illumination compound microscopes. Submicrometer resolution (translating to magnifications exceeding a factor of
100) is required for these studies, which usually do not rely heavily on large depths of field for successful observation. On the other hand, a wide variety of specimens must be examined at smaller magnifications but require a larger depth of field with a high degree of contrast. Stereomicroscopes have characteristics that are valuable in situations where three-dimensional observation and perception of depth and contrast is critical to the interpretation of specimen structure. These instruments are also essential when micromanipulation of the specimen is required in a large and comfortable working space. The wide field of view and variable magnification displayed by stereomicroscopes are, of course, also useful for careful manipulation of delicate living organisms.

Acknowledgments The first four coauthors are grateful to the late professor Michael Wesley Davidson (1950–2015). The present chapter is largely derived from the MicroscopyU(niversity) website [6] that was his brainchild. RP acknowledges support via Ministry of Education projects: Chiral Microscopy (LTC17012), CzechBioImaging (LM2015062), and ChemBioDrug.1

1

CZ.02.1.01/0.0/0.0/16_019/0000729

274

Erin E. Wilson et al.

References 1. Bradbury S (1967) The evolution of the microscope. Pergamon Press, New York. ISBN: 9781483164328 2. Wheatstone C (1838) Contributions to the physiology of vision – part the first. On some remarkable, and hitherto unobserved, phenomena of binocular vision. Phil Trans Roy Soc Lond 128(1838):371–394. https://www. jstor.org/stable/108203 3. Turner GLE (1989) The great age of the microscope. The collection of the Royal Microscopical Society through 150 years. Adam Hilger, Bristol; CRC Press, New York. ISBN: 9780852740200 4. Seidenberg RL (1981) Stereomicroscopy: a review. Am Lab 13(April):114–125 5. Nothnagle PE, Rosenberger HE, Seidenberg RL (1973) Stereomicroscopy: some considerations. Canadian Res Dev 6(2):30–35 6. Stereomicroscopy (as part of “MicroscopyU: The source for microscopy education”). Molecular Expressions™ (National High Magnetic Field Laboratory, Florida State University, Tallahassee, FL, USA). https://www. microscopyu.com/techniques/ stereomicroscopy 7. Schlueter GE, Gumpertz WE (1976) The stereomicroscope: instrumentation and techniques. Am Lab 8(April):61–71 8. Mann A (ed) (1993) Selected papers on zoom lenses. SPIE Milestone Series (Thompson BJ, ed), volume MS 85. SPIE, Bellingham, WA. ISBN: 9780819413888 9. Duval MG, Chung H, Lehmann OJ, Allison WT (2013) Longitudinal fluorescent observation of retinal degeneration and regeneration in zebrafish using fundus lens imaging. Mol Vis 19:1082–1095. https://www.ncbi.nlm.nih. gov/pubmed/23734077 10. Inoue´ S, Spring KR (1997) Video microscopy: the fundamentals. Plenum Press, New York. ISBN: 978-0306455315 11. Schnitzler H, Zimmer K-P (2008) Advances in stereomicroscopy. Proc SPIE 7100:71000P. https://doi.org/10.1117/12.797409 12. Castinel A, Duignan PJ, Pomroy WE, Lyons ET, Nadler SA, Dailey MD, Wilkinson IS, Chilvers BL (2006) First report and characterization of adult Uncinaria spp. in New Zealand sea lion (Phocarctos hookeri) pups from the Auckland Islands, New Zealand. Parasitol Res 98 (4):304–309. https://doi.org/10.1007/ s00436-005-0069-8 13. Pauza DH, Saburkina I, Rysevaite K, Inokaitis H, Jokubauskas M, Jalife J, Pauziene

N (2013) Neuroanatomy of the murine cardiac conduction system. A combined stereomicroscopic and fluorescence immunohistochemical study. Auton Neurosci 172(1-2):32–47. https://doi.org/10.1016/j.autneu.2013.01. 006 14. Namavari A, Chaudhary S, Sarkar J, Yco L, Patel K, Han KY, Yue BY, Chang J-H, Jain S (2011) In vivo serial imaging of regenerating corneal nerves after surgical transection in transgenic thy1-YFP mice. Invest Ophthalmol Vis Sci 52(11):8025–8032. https://doi.org/ 10.1167/iovs.11-8332 15. Santos VT (2013) Macrobrachium (ghost) shrimp eye. Nikon Small World. https://www. nikonsmallworld.com/people/vitoria-tobiassantos 16. Murphy DB, Davidson MW (2012) Fundamentals of light microscopy and electronic imaging. Wiley-Blackwell, Hoboken, NJ. https://doi.org/10.1002/9781118382905 17. Abramowitz M (1990) Reflected light microscopy: an overview. Olympus America, Lehigh Valley, PA. Melville, NY 18. Abramowitz M (1987) Contrast methods in microscopy: transmitted light. Olympus America, Lehigh Valley, PA. Melville, NY 19. Axelrod D (1981) Zero-cost modification of bright field microscopes for imaging phase gradient on cells – schlieren optics. Cell Biophys 3 (2):167–173. https://doi.org/10.1007/ BF02788132 20. Bradbury S, Evennett PJ (1996) Contrast techniques in light microscopy. RMS Handbooks, № 34. BIOS Scientific Publishers, Oxford. https://doi.org/10.1046/j.1365-2818.1997. 00702.x 21. Tuchin V (2015) Tissue optics: light scattering methods and instruments for medical diagnostics (3rd edn). SPIE, Bellingham, WA. https:// doi.org/10.1117/3.1003040 22. Barer R, Joseph S (1954) Refractometry of living cells, Part I. Basic principles. Quart J Microsci Sci 95(4):399–423. J Cell Sci s3–95:399–423, https://jcs.biologists.org/ content/s3-95/32/399 23. Schlueter GE (1974) A high intensity transmitted light base for darkfield stereomicroscopy. American Laboratory 6(April):40–45 24. Shaner NC, Patterson GH, Davidson MW (2007) Advances in fluorescent protein technology. J Cell Sci 120(24):4247–4260. https://doi.org/10.1242/jcs.005801

Chapter 10 Conventional, Apodized, and Relief Phase-Contrast Microscopy Radek Pelc, Zdeneˇk Hostounsky´, Tatsuro Otaki, and Kaoru Katoh

Abstract Non-absorbing colorless specimens (phase objects) can be visualized either by chemical staining (histology) or by the so-called optical contrasting (staining). The latter is achieved by converting optical phase shifts within the specimen, invisible to human eye, to intensity differences in the microscopic image. Out of several available modes of microscopic phase visualization, conventional, apodized, and relief phase contrast are described in detail. A comparison with relief contrast alone, that is, the off-axis (schlieren) illumination mode is presented as well. Images of various phase specimens of biological origin are shown, demonstrating the strong and weak points of each mode. Physiological aspects of image comprehension, facilitated by shading and other visual cues to depth structure are briefly discussed. A phase-contrast microscope equipped with an objective hosting no phase annulus is presented; the latter is located in a pupil projection (optically relayed) plane, in an external attachment unit. This setup enables phase-contrast and, for example, conventional epi-fluorescence and total internal reflection fluorescence (TIRF) images to be acquired with a single objective lens. Such modality is demonstrated in growth cones of neuroblastoma–glioma hybrid mouse–rat cells and touch receptor neurons of Caenorhabditis elegans. Examples of phase-contrast imaging in electron and X-ray (synchrotron radiation) microscopy are also presented. Key words Apodization, Depth perception, Modulation contrast, Neuroblastoma–glioma cells, Quasi-3D images, Schlieren imaging, Shape-from-shading, Microtomography, Touch receptor neurons, Visual cues

1

Introduction to Optical Contrasting

1.1 Histological Versus Optical Staining

Neurohistology and histology in general are usually associated with stained specimens, that is, those featuring regions differing in color and/or absorption (Fig. 1D). Trans-illuminated stained glass (Fig. 1A) represents a macroscopic analogy. Staining introduces the so-called amplitude contrast into the images and such objects

Electronic supplementary material: The online version of this chapter (https://doi.org/10.1007/978-1-07160428-1_10) contains supplementary material, which is available to authorized users. Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_10, © Springer Science+Business Media LLC 2020

275

276

Radek Pelc et al.

PHASE CONTRAST

100 μm

B BRIGHT FIELD

A

D C

Fig. 1 Chemical (A) versus optical (B) staining (contrasting) of non-absorbing, trans-illuminated objects. (A) Stained glass (St Ludmila) in Mostov Castle chapel. (B, C) A diatom, reproduced from Ref. [1] by permission of © AAAS. Species, scale and objective details not stated, conventional phase contrast (positive type, as judged from the published [untrimmed] micrograph). (B) The oldest known phase-contrast micrograph (Zernike 1932), probably obtained with a setup shown in Suppl. S3e (right). (C) Bright-field image acquired with a narrow iris diaphragm. (D) Histological section from human colon stained with hematoxylin (objective,
10/0.25)

are referred to as “amplitude objects”; the amplitude of the illuminating light wave is modulated by the details in the specimen (Fig. 1D). To visualize those, an ordinary microscope is sufficient (bright-field illumination). On the other hand, there are situations when specimens cannot or should not be stained. However, unless stained, they often feature no or few absorbing structures, that is, there is little contrast in them. As such, they cannot be studied under the ordinary microscope unless the contrast is introduced into their images by optical means. Such optical contrast enhancement includes a number of modes [2, 3]: (1) Dark field; (2) Hoffman modulation contrast and other so-called schlieren (relief) illumination modes; (3) differential interference contrast (DIC) after Smith/Nomarski); (4) interference contrast (see Subheading 2.3), (5) polarization microscopy, and (6) phase contrast [1, 4–12] which is by far the most widespread modality and at the same time relatively simple. Its

Phase-Contrast Microscopy

277

DIE REIFETEILUNGEN (MEIOSE) BEI DER SPERMATOGENESE DER SCHNARRHEUSCHRECKE PSOPHUS STRIDULUS L.

MIKROLABORATORIUM CARL ZEISS JENA DR. KURT MICHEL

INSTITUT FÜR DEN

10 μm

PHASENKONTRASTVERFAHREN NACH ZERNIKE 1943

WISSENSCHAFTLICHEN FILM

AVAILABLE AS A VIDEO (Suppl. S1)

Fig. 2 Optical contrasting of chromosomes by phase-contrast imaging. A video still image (at 9 min 40 s out of ca. 11 min total) of second meiosis (telophase) during spermatogenesis in the creaking locust (Psophus stridulus L.), as recoded in unstained cells on a 16 mm film (one frame every 7.5 s) [13]. Objective,
40/0.65 (conventional phase-contrast [positive]). A portion is available as a video (every fourth frame only, 1 sec corresponds to 12 min 30 sec at 25 frames/sec) licensed by © IWF Go¨ttingen (now part of German National Library of Science & Technology, Hannover)

popularity received a powerful boost when unstained chromosomes were filmed during cell division [13]; a video may be inspected online (Suppl. S1 associated with Fig. 2). Unstained polytene (giant) chromosomes from salivary glands of Chironomus larvae could also be visualised (Suppl. S4a). Phase-contrast microscopy is best suited for optically very thin objects such as pseudopodia, lamellipodia or neuronal growth cones, and in this sense is complementary, for example, to DIC Nomarski microscopy. Many examples may be inspected online.1 In a broad sense, the optical contrasting (also referred to as optical contrast enhancement) may be referred to as optical “staining” because its effect is the same as if the specimens were actually stained histochemically. A narrower definition of “optical staining” only refers to the use of color filters in the optical path of the microscope, which however does not necessarily enhance contrast in the images. What is required is to convert the light waves’ phases, modulated by the specimen structures but undetectable by eye or camera (Fig. 1C), to amplitude differences that can be perceived and detected (Fig. 1B). Non-absorbing specimens are called “phase specimens,” absorbing ones “amplitude specimens.” The latter include those that were stained (histo)chemically (Fig. 1A, D). 1

http://microscopyu.com/galleries/dicphasecontrast/index.html

278

Radek Pelc et al.

A

OBJECTIVE

B

SLIDING DIAPHRAGMS (m, n)

C

RELIEF CONTRAST (off-axis illumination)

CONDENSER IRIS DIAPHRAGM

RELIEF DIAPHRAGM

D

BRIGHT FIELD (axial illumination)

100 μm

Fig. 3 Relief contrast (schlieren) imaging. A combination of A and B (an edge diaphragm in both condenser and objective) represents a simplified version of Hoffman modulation contrast. (A) The original version of reliefcontrast imaging device (a simple microscope) capable of “optically staining” objects with poor natural contrast. Two shifting relief (edge) diaphragms (m, n) are placed at the objective back focal plane; m is semi-absorbing. (B) Relief-contrast condenser; the relief diaphragm can be either engaged (") or disengaged (#). The objective is not modified at all [14]. (C, D) Potato starch granules obtained by fractionated sedimentation of starch powder in water, to reduce granule size spread. Images were acquired with the device shown in B. Arrows ("#) indicate the position of the relief diaphragm. Objective,
10/0.25. Other examples are shown elsewhere [14–17]. Image A is reprinted from Ref. [18] (copyright expired 70 years after author’s death [August To¨pler, 1836-1912]), image B from Ref. [14] by permission of © The American Physiological Society 1.2 Relief Contrast as an “Ancestor” of Phase Contrast

Relief-contrast imaging (Fig. 3) has been used in microscopy for over 150 years [18–20], certainly earlier than the diffraction theory of microscopic image formation has been formulated [21, 22], illustrated [23, 24] and correctly interpreted [25, 26]; cf. Ref. [20] in Chapter 8 by Pelc. Relief-contrast imaging typically utilizes one or two edge diaphragms inserted asymmetrically into the light beam. An advanced

Phase-Contrast Microscopy

279

form is known as Hoffman modulation contrast employing an asymmetric illumination slit diaphragm in the condenser and a graded-transmittance filter called “modulator” in the objective [27]. A more simple form is usually referred to as relief, off-axis, oblique, or asymmetric illumination (one edge diaphragm in the condenser, Fig. 3B) [14–17]. Various intermediate versions are listed elsewhere [3, 15]. Either of these setups belongs to the category of schlieren imaging whose origin can be traced to Robert Hooke’s experiments in seventeenth-century Oxford (1672, as documented by Joachim Rienitz). Relief contrast yields quasi-3D images featuring a relief impression, owing to shading patterns also referred to as streaks or schlieren (Fig. 3C and elsewhere in the present chapter). This effect, discussed in greater detail in Subheading 3.2, is a consequence of an asymmetry in the diffraction pattern at the objective back focal (transform) plane. It is achieved by the abovementioned asymmetric diaphragm(s) (Fig. 3A, B). The low sensitivity of dark field and relief/schlieren contrast in imaging very thin phase specimens (Figs. 4A, 12C, and 16C) was the very motivation for modifying it [28, 29], eventually yielding the phase-contrast microscopy (1953 Nobel Prize to Frits Zernike, a Dutch scientist) [1, 12]. The contrast is indeed much higher in thin objects (Fig. 4E). Contrary to a popular belief, the phase-contrast imaging had initially nothing to do with microscopy, let alone with cell imaging. Zernike originally portrayed the phase contrast as an improvement of the so-called knife-edge (schlieren) method (invented by Le´on Foucault in mid-nineteenth century) to trace minute irregularities on astronomical mirror surfaces [9, 10]. 1.3 Imitation Phase Contrast

It has been well known that images with a contrast comparable to that in the relief or even phase-contrast ones can be obtained in the ordinary bright-field illumination mode, simply by going slightly under or above focus (Fig. 4B, D). Such images are similar to phase-contrast ones (Fig. 4E). In electron microscopy, this approach is often referred to as “defocus phase contrast,” and routinely used. Contrast can be further enhanced by subtracting the underand over-focused images from each other. This effect is sometimes applied instead of using the phase-contrast microscope [30]. Details can be found in Subheading 4.1 and Fig. 4. The drawbacks of such calculated images are: (1) lower contrast than in phase contrast; (2) images cannot be viewed directly in the microscope; (3) image detail is somewhat compromised. Apart from defocus, image contrast also depends on phase shift and spatial frequency.

280

Radek Pelc et al.

RELIEF CONTRAST A

B

C

10 µm under focus PHASE CONTRAST H

PHASE CONTRAST

BRIGHT FIELD D

precise focus

E

10 µm above focus

NEGATIVE PHASE CONTRAST

IMITATION PHASE CONTRAST F=B-D

G=D-B

I

100 µm POSITIVE

POSITIVE

NEGATIVE

POSITIVE

Fig. 4 Thin, non-absorbing (phase) objects visualized in various imaging modes. (A–G) Microcrystals in pickled pepper brine evaporate. Note the poor contrast in the relief-contrast image (A) acquired with the device shown in Fig. 3B. Images of equal or superior contrast can be obtained under ordinary bright-field illumination, even without closing the condenser iris diaphragm too much, merely by slight under- or over-focusing (B, D). In B, the objective lens is 10 μm closer to the specimen than at precise focus (C). A subtraction of the out-of-focus images (B, D) from each other yields even stronger contrast (imitation phase contrast; F, G) which, however, is weaker than in genuine phase-contrast imaging (E). Image quality is compromised in B, D, F and G. Objectives,
10/0.25 (no phase plate, A–D) and
10/0.25 (conventional phase contrast [negative], E). Condenser iris diaphragm setting (effective NA) 0.45 (A) or 0.20 (B–D). (H, I) Amyloid-β aggregates prepared in vitro (specimen provided by MUDr. Anna Kubesˇova´, Psychiatric Centre Prague). Objective,
20/0.40 DL (conventional phase contrast [positive]). (H) Standard specimen. (I) Specimen prepared from frozen stock

2

Conventional Phase-Contrast Microscopy

2.1 Principle of Contrast Generation

Typically, a phase object has a refractive index that differs from that of the medium. As a result, it shifts the phase of an incident light wave because the speed of light through the object and the medium is different, in inverse proportion to the refractive index (n). For a given wavelength (λ), the phase shift (ε) induced by the phase object equals 2π(n1  n2)t/λ, where n1 and n2 are the refractive indices of the medium and phase object, respectively, and t is the object’s (physical) thickness. The wave-front becomes locally retarded or advanced (Fig. 5A), that is, it changes from Ψ ¼ sin (ωt) to Ψ ¼ sin (ωt  ε). The (object-to-medium) phase shift and its sign thus “encode” an information about the object’s structure (relative to the medium

A a

b = sin( t - )

PHASE OBJECT

= sin( t)

B

Phase shift, = ca -150°

OBJECTIVE BACK FOCAL (TRANSFORM) PLANE IMAGE PLANE

OBJECT PLANE

b

a

a

a c

c

c

(NO PHASE PLATE)

d

BRIGHT OBJECT ON BRIGHT BACKGROUND

BRIGHT FIELD a = background (surround) light in the image d = a + c = interfered light (phase object’s image)

PHASE OBJECT

(phase shift, = ca -60°)

a b

c

c

e

PHASE PLATE

(phase shift and attenuation of a only)

c

e

f BRIGHT OBJECT ON DARK BACKGROUND

NEGATIVE PHASE CONTRAST -i /2 e = 0.25 a e = background (surround) light in the image = direct light a upon passing through the phase plate (-90º

phase shift and amplitude / intensity reduction to 50% / 25%)

f = c + e = interfered light (phase object’s image)

100 μm

BOTH IMAGING MODES a = illuminating light = direct light through every point in the specimen (unperturbed wave) -i /3 (phase shift, = - /3 = -60°) b = all light passing through the specimen, b = a e c = light diffracted by the phase object (perturbed wave) = b - a

Fig. 5 Effects of a phase-retarding object on illuminating light and microscopic imaging. (A) Illuminating light ! ! wave (Ψ) passing through the phase object (grey box) inducing a phase shift (ε) of ca. 150 ; b ¼ a ei5π/6. (B) Vector diagrams approximately “modeling” contrast in images of radula isolated from Spanish slug (Arion lusitanicus), consistent with a phase shift (ε) of ca. 60 . The radius of each circle represents the amplitude ! of the illuminating light (vector a ). Transmittance of the phase annulus is ca. 25%. Objectives,
10/0.25 (no phase plate) and
10/0.25 (conventional phase contrast [negative]). The phase-contrast image is reproduced from Ref. [16] by permission of © SPIE. Cf. positive phase-contrast image of the same object (Fig. 10)

282

Radek Pelc et al.

surrounding it), in terms of its optical path difference (OPD) map; OPD ¼ (n1  n2)t. If the medium is air (n1 ¼ 1), OPD equals physical thickness (t) minus optical thickness (n2t). Positive (negative) phase shift and OPD mean the object is phase-advancing (phase-retarding). For example, a phase shift of 90 translates to an OPD of λ/4. As depicted in Fig. 5B, the illuminating light ! that was phaseshifted due to traversing the phase object (vector b ) can be com! ! pounded of direct light (vector a ) and diffracted light (vector c ), ! ! ! that is, b ¼ a + c . The direction and amplitude of direct light (the so-called geometric rays) is unaffected. On the other hand, the direction and amplitude of diffracted light varies according to the lateral dimensions of the object (Fig. 6) and the phase shift between the object and the surrounding medium (Figs. 5, 9, and 10), respectively. In the image plane, the direct and diffracted light interfere (Figs. 5B and 6). At bright-field conditions, the amplitude of the ! interfered light (vector d ) is always the same as that of the back! ground (vector a ), that is, the phase object is hardly visible, at least at sharp focus (cf. Fig. 4). One way of making the object visible is to (1) attenuate or completely block the direct light (diffracted light carrying the information about object’s structure should not be attenuated). Attenuation [33] takes place, for example, in schlieren imaging (relief contrast, Hoffman modulation contrast) [14–18, 27], blockage in darkfield illumination (see Chapter 8 by Pelc, Chapter 9 by Wilson et al., and Chapter 11 by Saghafi et al.). Another option is to (2) alter the relative phase (i.e., the phase difference) between the direct and diffracted light; this is the case in phase contrast. In either of these two possibilities (1, 2) the interfered light (direct ! and diffracted light combined, vector f ) representing the phase object in the image is no longer of the same intensity as that of the ! background (vector a ), and, compared to a bright-field image, the object becomes more clearly visible. However, the phase shifting alone (Fig. 17A) would generate an image of only moderate contrast, comparable to that obtained by attenuating the direct light. As a result, it is a combination of the phase shifting and attenuation of the direct light, that increases the contrast significantly, as described in the following section. 2.2 Phase and Amplitude Modulation

In order to alter the relative phase between the direct and diffracted light the two must be physically separated. This condition is fulfilled most efficiently at the objective back focal plane (Fig. 6). The diffracted light always fills much of the objective back focal plane, regardless of condenser diaphragm’s shape or size. On the other hand, if the condenser diaphragm is small or thin, for example, of annular or slit shape, the direct light passes the objective back focal plane only through a small area called “conjugate area” as it is

Phase-Contrast Microscopy ANNULAR DIAPHRAGM

APODIZED PHASE PLATE

283

CONVENTIONAL PHASE PLATE

Phase A … annulus … A (ring) B … Apodization … annuli C C see cross-sections below

CONDENSER LENS

OBJECTIVE LENS

TUBE LENS

Large PHASE OBJECT Small

ILLUMINATING LIGHT

OBJECT PLANE

TRANSFORM (objective back focal) PLANE

IMAGE PLANE Direct (surround) light Diffracted (object-specific) light

Periodic objects st

Diffracted light (+1 order) th

Direct light (0 order) st

Diffracted light (-1 order)

C B

C

A

A

B C

C

APODIZED PHASE CONTRAST

‘Feet’

CONVENTIONAL PHASE CONTRAST

Fig. 6 Principle of conventional and apodized phase-contrast microscopy: A selective treatment of direct and diffracted light. Direct light is modulated in the “A” area (phase shift and attenuation) optically conjugate with the annular diaphragm. The “A” area is referred to as “conjugate area,” the rest of the phase plate (also called diffraction plate) as “complementary area.” Diffracted light from small structures is not appreciably modulated as it passes mainly through the “C” areas in both conventional and apodized phase contrast. Diffracted light from large structures is attenuated by the apodization annuli (“B” areas, apodized phase contrast only) as originally proposed elsewhere. If the transmittance of the phase annulus is set to zero, the phase-contrast imaging becomes dark-field imaging. Reproduced from Ref. [16] by permission of © SPIE (layout slightly modified to save space); the apodized phase plate design has originally been proposed earlier [31, 32]

284

Radek Pelc et al.

conjugate with the diaphragm. A special filter of such shape is then employed to shift the phase of direct light, typically by 90 (i.e., λ/4). If the diaphragm is of annular shape it is called phase annulus (ring) (Fig. 6). The contrast between the object and the background in the image is further enhanced by attenuating the direct ! ! light (vector e smaller than vector a in Fig. 5B). Unlike in high-power objectives [8, 11], the objective back focal plane is not often located between lenses in low-power objectives, and the phase-shifting (dielectric) and light-attenuating (metallic) annuli thus do not have to be placed on any of the lenses compounding the objective. Instead, they are coated onto a transparent circular holder called phase plate (Fig. 6) typically inserted into the objective lens casing. The vector diagrams (Fig. 5B) make it possible to “calculate” contrast in the phase-contrast image, without having to resort to any mathematical apparatus at all. Otherwise, Bessel functions would be necessary for larger than very small phase shifts. Likewise, it is possible to deduce that in very thin objects (phase shift ε < ca. 15 , i.e., OPD < ca. λ/24), the highest contrast is obtained if the phase of the direct light is shifted by ca. 90 and most of its intensity attenuated at the objective back focal plane, that is, by the phase annulus. This deduction can also be made arithmetically. As mentioned in the previous section, the light wave of amplitude equal to unity, after passing through a phase-retarding or phase-advancing object inducing a phase shift ε (negative or positive value, respectively) may be described by Eq. 1a or 1b: Ψ ¼ sin ðωt  εÞ

ð1aÞ

Ψ ¼ sin ωt cos ε  cos ωt sin ε

ð1bÞ

At very small phase shifts (ε < ca. 15 ), sin ε  ε and Eq. 1b can be written as Eq. 2a, 2b or 2c: Ψ  sin ωt  ε cos ωt

ð2aÞ

Ψ  sin ωt  ε sin ðωt þ π=2Þ

ð2bÞ

Ψ  sin ωt þ ε sin ðωt  π=2Þ

ð2cÞ

The first term, sin ωt corresponds to the background (illuminating) light while the second one, sin (ωt  π/2) to the diffracted light. Obviously, their phases differ by π/2 ¼ 90 . The two components interfere with each other at the image plane. In order to create a high-contrast image, there are two possibilities: 1. To make them interfere destructively (“annihilate”). In that case, a dark image of a phase-retarding object is formed, against bright background (positive phase contrast). To achieve that, the two components must be brought to antiphase, by shifting

Phase-Contrast Microscopy

285

the direct light’s phase by ca. +90 (or that of diffracted light by ca. 90 ). 2. To make them interfere constructively, so that bright image of the phase object is formed, against a dark background (negative phase contrast). To achieve that, the two components must be brought to in-phase, by shifting the direct light by ca. 90 (or diffracted light by ca. +90 ). This situation is depicted in Fig. 5B. The above applies only to sufficiently thin phase-retarding objects. In phase-advancing objects, but also in thicker phaseretarding objects, the image rendering is opposite in terms of grayscale (positive-negative). These situations are explained in greater detail in Subheading 3.1. In both steps (1) and (2), the direct light should also be attenuated so that it is comparable in intensity to the diffracted light. However, at 75% attenuation (or 25% transmittance, a common value in commercially available microscopes), the optimal phase shift is only ca. 75 rather than 90 . This can be easily derived with the vector diagrams (Fig. 5B). This condition is sometimes referred to as Richter’s condition discussed in detail elsewhere [7, 8] or, in case of positive phase contrast and phase-retarding objects, “conditions of darkest contrast” [6]. Nevertheless, the contrast improvement thus achievable is not very marked, and the 90 phase shift in the annulus is the most commonly used value. The illuminating diaphragm and corresponding (conjugate) area at the objective back focal plane do not have to be of annular shape (Fig. 7E). Separation of direct and diffracted light is possible even with other shapes of the diaphragm (Fig. 7A–D) [11, 24, 26, 28, 29, 35, 39], as investigated most extensively by Zernike [11] and Beyer [35]. For example, a microscope made by C. Baker of Holborn Ltd (London) employed a diaphragm in the shape of a cross. However, the annular shape of the phase strip proved to be the most practical, for example, because it makes the alignment of the microscope very straightforward and the halo (shade-off) artifacts are spread in all angular directions in the image. The initially used shape was usually a straight strip (Fig. 7B), a logical extension of diffraction experiments that were carried out with amplitude gratings as test specimens. Their main aim was to verify the diffraction theory of microscopic image formation [21– 23, 25]; cf. Ref. [20] in Chapter 8 by Pelc. For that purpose, the so-called Abbe diffraction apparatus was devised [2, 23, 24, 26], complete with interchangeable slit masks inserted close to the objective back focal plane (Fig. 7F–H in this chapter, and Figs. 4 and 5 in Chapter 8). The effects on images of altering the relative phase of diffracted and direct light, and of attenuating the direct light, was also tested with such and similar devices (Fig. 7I). Although this is the very principle of phase-contrast microscopy it

DISC

STRIP

PROPELLER

GRID

ANNULUS

OBJECTIVE back focal plane

CONDENSER front focal plane

A

B

Burch and Stock Köhler and Loos (1942) (1941, 1947)

C

D

E

Beyer (1953)

Zernike (1942)

Zernike (1942)

PRE-ZERNIKE DIFFRACTION EXPERIMENTS

I MAGE

-1

+1

OBJECTIVE back focal plane

0

F

G

H

Objective masks (F-H) supplied with the 1876 DEMONSTRATION MICROSCOPE designed by Ernst Abbe at Carl Zeiss, Jena (Stephenson 1877)

CONDENSER front focal plane

I Bratuscheck (1892) Rheinberg (1904, 1905)

Fig. 7 Condenser diaphragms and optically conjugate areas at the objective back focal plane. (A–E) Examples of phase plates (1st row) and corresponding optically conjugate condenser diaphragms (2nd row) employed in phase-contrast microscopy [11, 28, 29, 35, 39]. The grey areas attenuate the direct (undiffracted) light and shift its phase, typically by 90 (corresponding to λ/4). (E) is the most common design [11] although not the only useful one (cf. Figs. 6 and 17). (F–H) Examples of aperture stops (masks) [21, 23, 24, 26] designed to verify Abbe’s diffraction theory of microscopic image formation [21–23, 25, 26], as part of a simple “Abbe diffraction apparatus.” Amplitude grating serves as a test specimen. (F) admits 0th order maximum (direct light) only; grid structure is not resolved. (G) additionally admits 1st order maxima (diffracted light); grid structure becomes correctly visible. (H) selectively admits 0th and 2nd order maxima; the grating appears twice denser (blocking of 1st order maxima brings about false structure). The condenser diaphragm in F–H is the same as in I (central slit). (I) An objective mask for testing the effect of altering the relative phase of direct (0th order maximum) and diffracted (1st order maxima) light, by employing Abbe’s glass wedge compensator [40, 41], in lieu of a phase plate introduced by Zernike and Lyot much later. The dotted pattern denotes either an experimental phase-difference modification (shift) between the direct and diffracted light [40, 41], or attenuation and phase shifting of direct light [33]

Phase-Contrast Microscopy

287

escaped attention at that time, probably because the test specimens were mostly [33] or only [40, 41] amplitude rather than phase gratings. Phase-contrast microscopy was thus nearly discovered in 1892 [33], as illustrated in Fig. 7I. A brief note about these very early experiments may be found in Suppl. S5 and [2, 6]. Zernike’s own work was originally patented/presented in 1932/1933 (Suppl. S5). As admitted by Zernike himself [1], Bernard Lyot, a French astronomer had also independently discovered the phase contrast. As mentioned above, the phase plate can be accessed through the back objective pupil in low-power objectives. A standard objective can thus be relatively easily converted to a phase-contrast one, by inserting a “homemade” phase plate. Recipes to make a phase plate can be found elsewhere [7–9, 11, 35, 39], the one provided by Burch [10] probably being the oldest one. 2.3 Halo (Shade-off) Artifacts

“Halo” is an artifact that surrounds the object’s (e.g., cell’s) image, much like the golden halo surrounds a saint’s head (Fig. 1A). “Shade-off” is an artifact inside the object’s image (internal halo). Both artifacts have the same cause, as described in Subheading 3.3 and Fig. 6, and manifest themselves as distortions in luminance distribution in the image. The halo and shade-off artifacts are inseparable from each other, so that “halo (shade-off)” is consistently used throughout the present chapter, rather than the less accurate (but common) “halo.” The specimen structure is often almost completely blanketed by these artifacts (Figs. 12A and 16A). In extreme cases, next to nothing can be deduced from the phasecontrast image (Fig. 12A). Intermediate situations are shown in Figs. 12B, 13A, and 16B. Many examples may be inspected online (see link in Fig. 15 legend). The intensity of diffracted light, and consequently the magnitude of the halo (shade-off) artifacts rise with the phase shift between the object and its surrounding medium. The artifacts can thus be often reduced or almost eliminated by replacing the medium surrounding the phase object, for example, air for oil. An example is shown in Fig. 12B, C (microcrystals mounted in air and oil, respectively. Unfortunately, the medium-replacement method is rarely useful when the phase objects are living cells, except when their refractive index needs to be determined by refractive index matching; the cell becomes almost invisible at that point [4]. In other cases, the halo (shade-off) artifacts have to be removed by optical means. The simplest, often sufficient way of halo (shade-off) artifact reduction in thicker objects is to use the relief contrast alone (Figs. 12A and 16A). Other options include variable (Fig. 11 and Suppl. S3), apodized (Figs. 15 and 16) and relief phase contrast (Fig. 13). Yet another possibility is the use of interference-contrast microscopy (not to be confused with differential interference

288

Radek Pelc et al.

contrast or DIC)2 in which no halo (shade-off) artifacts are present [4, 5]. The luminance level in each image point directly corresponds to an optical thickness value. A wider usage of the interference-contrast microscopy is prevented by a noticeably higher cost and complexity, compared to phase contrast. A combination of phase contrast and interference contrast used to be available as “Interphako” [7, 8] (Suppl. S3b). 2.4 Adjustment of a Phase-Contrast Microscope

Adjustment of a phase-contrast microscope is rather straightforward. The only thing to do is to make sure, with otherwise correctly adjusted condenser, that the annular diaphragm in the condenser and the phase annulus in the objective are well aligned. An interactive tutorial is available online.3 Briefly, the procedure is as follows (Fig. 8): 1. Upon setting up Ko¨hler illumination, shift the specimen on the stage so that only very few cells or other objects remain in the viewing field. 2. Select the correct annular diaphragm, that is, according the diameter of the phase annulus in the objective lens. The diaphragms are typically marked Ph1, Ph2, and so on. 3. Remove one eyepiece from its socket and replace it with a centering telescope (also called an “auxiliary microscope”) that is typically part of the phase-contrast set. Example views through the centering telescope are shown in Fig. 8A. In some microscopes, a Bertrand lens can be swung in or inserted (an “aperture-viewing unit”), in which case the telescope is not needed. 4. While turning the adjustment screws of the condenser annular diaphragm (not those of the condenser itself), make sure its bright image overlaps with the phase annulus in the objective, as shown in Fig. 8A [bottom-left cutout]. Note that for practical reasons, the dimensions of the annular diaphragm are such that its projection at the objective back focal plane is slightly narrower than the phase annulus. 5. Replace the centering “telescope” with the eyepiece. A common pitfall is associated with ignoring the step 1 above. If too many objects (especially strongly diffracting ones) remain in the viewing field during the adjustment, the procedure may turn quite difficult, if not outright impossible. Examples are shown in Fig. 8B–D. The highly periodic structures in the specimen diffract light that forms prominent patterns, thus splitting the single image of the phase annulus into 1 + 2 (Fig. 8B), 1 + 6 (Fig. 8C), or 1 + multiple (Fig. 8D) annuli. On the other hand, leaving no

2 3

http://microscopyu.com/galleries/dicphasecontrast/index.html https://www.microscopyu.com/tutorials/phase-contrast-microscope-alignment

Phase-Contrast Microscopy

MISALIGNED DIAPHRAGM

A

B

-1

0

289

+1

NO DIFFRACTING OBJECTS b

a

a

b’

ALIGNED DIAPHRAGM NO DIFFRACTING OBJECTS

50 μm

a+b NO SPATIAL FREQUENCY

C

A SINGLE SPATIAL FREQUENCY

D

x2

50 μm THREE SPATIAL FREQUENCIES

100 μm MULTIPLE SPATIAL FREQUENCIES

Fig. 8 Alignment of a phase-contrast microscope and effects of strongly diffracting objects. Combined views down the eyepiece (images) and the centering telescope (diffractograms at the objective back focal plane, insets). (A) No or few diffracting objects in the viewing field. Only the phase annulus (a) and the annular diaphragm (b) can be seen; stray light (b0 ) is a weak mirror image of b. Alignment (centering) of the annular diaphragm is straightforward. (B–D) Alignment is difficult if strongly diffracting (scattering) objects, particularly periodic ones, occupy most of the viewing field. Depending on the spatial frequencies and orientation of the corresponding periodic structures (ribs), the direct light (0th order, central bright annulus) is split into 1 + 2 (B), 1 + 6 (C) or multiple (D) annuli. In (B), the direct light (0th order) and 1st/+1st diffraction orders are marked. In (C), the periodicity of ca. 4.2 μm is less clearly discernible (thinner ribs in the cutout detail) as the corresponding diffracted light is right at the edge of the objective aperture and cannot fully contribute to image formation. Objects are body scales of silverfish (Lepisma saccharina), mounted in air. Objectives,
40/0.65 dry (B, C) and
20/0.40 (D), in both cases conventional phase contrast (negative)

objects at all in the viewing field may not be good either, especially in modern microscopes in which internal reflections are mostly eliminated; the phase annulus is often almost invisible in such cases.

290

3

Radek Pelc et al.

Specialized Phase-Contrast Modes

3.1 Positive, Negative and Variable Phase Contrast

As shown in Fig. 4 for the imitation phase contrast, either positive (“dark”) or negative (“bright”) contrast can be obtained. In real (non-imitation) phase contrast (positive/negative), the phase of the direct light is advanced/retarded by the phase plate, typically by 90 . Assuming the phase objects are optically thin (see below) and phase-retarding as, for example, cells in a physiological solution or water, positive/negative phase contrast renders them dark/ bright on bright/dark background (Fig. 9). These two situations are reminiscent of bright-field images of classical histological (stained) preparations (Fig. 1D) and fluorescence images, respectively. This might explain the current prevalence of positive phasecontrast objectives on the market (histology being older than fluorescence microscopy). Another possible explanation lies in the fact that unlike, for example, owls or other nocturnal animals accustomed to a dark sky, humans as diurnal creatures active during the day are used to a bright sky, this being an equivalent of the bright background in positive phase contrast microscopy. The dark/bright rendering demonstrated in Figs. 9 and 10, and described in detail in Glossary (entry ►Dark/bright contrast) takes place if the phase shift induced by the object is below a certain limit. As it can be easily proved with vector diagrams (Figs. 9 and 10), this limit is ca. 55 for positive phase contrast (assuming 25% transmittance and 90 shift in the phase annulus), and determines the so-called range of unreversed contrast. Close to the reversal point, the object is hardly visible. Beyond that limit, the rendering is opposite, that is, bright objects turn dark and vice versa. The range of non-reversed contrast is greater in negative phase contrast. At a phase shift of 180 (corresponding to λ/2) in the phase annulus, positive and negative phase contrast yield identical images. However, as it can be deduced from Fig. 5B, this option is impractical because of lower image contrast. In any case, caution is due when interpreting images of objects inducing phase shifts comparable to or greater than 360 , as phase wraps around (cf. Subheading 3.5). “Anoptral contrast” (a variant of negative phase contrast) used to be manufactured by Reichert at least since 1954 and supplied with, for example, Neozet, Biozet, and Zetopan microscopes. It employed phase plates made of soot featuring smaller reflections than metallic coatings common in other microscopes [7, 8, 38]. The implication of the above is that obtaining a low-contrast image in positive or negative phase contrast does not necessarily mean that the phase-contrast microscopy as such is unsuitable for that particular specimen. In order to make it possible to set any sign (positive or negative) and value of the phase shift, and any value of transmittance in the phase annulus, the so-called variable phase-

Fig. 9 Optically thin specimen in positive/negative phase contrast. Images are consistent with a phase shift (ε) of ca. 30 . Object’s rendering is complementary, with good contrast in both cases. (Top) Human buccal epithelial cells. Objectives,
10/0.25 DL and BM (conventional phase contrast [positive and negative, respectively]). (Middle) Vector diagrams approximately “modeling” contrast in the images (see Fig. 5B for details). The phase shift and transmittance of the phase annulus are +90 /90 (positive/negative phase contrast, at λ  530 nm) and ca. 25%, respectively. Luminance (approx. the light intensity recorded by a ! ! camera in each image!pixel) between and in the “phase objects” (cells) is proportional to | e |2 and | f |2, ! respectively; | e | and | f | are light vector amplitudes. (Bottom) An equivalent representation of the!vector ! ! diagrams, showing more clearly the relative phases of diffracted ( c ), background ( e ) and interfered ( f ) light

292

Radek Pelc et al.

Fig. 10 Optically semi-thin specimen in positive/negative phase contrast. Images are consistent with a phase shift (ε) of ca. 60 . The positive phase-contrast image is of poor quality as ε is near the contrast-reversal point. (Top) Radula of Spanish slug (Arion lusitanicus). Objectives,
10/0.25 DL and
10/0.25 (conventional phase contrast [positive and negative, respectively]). A color version of the boxed area is shown in Fig. 5B. A thicker portion of the same specimen is shown in Fig. 12A. A portion of the negative phase-contrast image is reproduced (upon conversion to B&W) from Ref. [16] by permission of © SPIE. Other examples may be inspected online (https://www.microscopyu.com/tutorials/positive-and-negative-phase-contrast). (Bottom) Vector diagrams approximately “modeling” contrast in the images (see Figs. 5B and 9 for details); “phase objects” are the periodically spaced denticles of the radula

Phase-Contrast Microscopy

293

contrast systems (Suppl. S3) have been developed [7, 8, 43], such as “Polanret” by American Optical [6] (Suppl. S3d) or its various modifications [44], including Cooke’s “Variable-amplitude phase contrast” (Suppl. S3e) and Nikon’s “Interference-phase contrast ” (Suppl. S3a). The latter system (which however does not have any built-in interferometer) utilizes a set of four optical elements jointly acting as a phase plate enabling to vary the ratio between the intensity of direct and diffracted light and their relative phase [43] (Fig. 11). These elements are located in a plane conjugate to the objective back focal plane (approximately its exit pupil). The “Interphako” system by Carl Zeiss makes it possible to combine phase-contrast with interference contrast or to switch to interference contrast alone which is free of the halo (shade-off) artifacts [7, 8] (Suppl. S3b).

VARIABLE PHASE PLATE (P1 + Q + PA + P2) V P2

P1 X

P1 = Polarizer (rotatable) Q = Quarter-wave plate PA = Polarizing aperture (fixed) P2 = Analyzer (rotatable)

P1

Q PA

EYEPIECE

P2

PA

OPTICAL RELAY (PUPIL PROJECTION) UNIT

BFP OBJECTIVE

V = Polarization direction in PA X = Slow or fast axis in Q VX angle = 45° (fixed)

SPECIMEN CONDENSER

Fig. 11 Variable phase contrast as an external (optical-relay) unit. The device is capable of functioning with standard (non–phase-contrast) objectives. P1, Q, PA, and P2 jointly act as a variable phase plate whose phase shift and transmittance can be altered to some extent by rotating P1 and P2. All four elements are placed in a physically accessible, optically relayed plane that is conjugate to the objective back focal plane (BFP). PA is sometimes referred to as “zonal polarizer.” Reproduced from Ref. [8] by permission of © Polish Scientific Publishers (PWN); layout slightly modified for clarity, originally described elsewhere [43]. The device may be inspected online (https://www.microscopyu.com/museum/interference-phase-microscope-accessory) and in Suppl. S3a

294

Radek Pelc et al.

All of the abovementioned devices represent external units that can be inserted between the ocular head and the nosepiece. Zeiss Jenaval4 represents a more simple version of such a system. Although none of them is commercially available any longer the concept of external phase plates (a pupil-projection, or optical-relay setup) has been recently reintroduced by Nikon and Leica (see Subheading 3.4). Any of the abovementioned variable systems is also capable of reducing the halo (shade-off) artifacts, as the ratio of direct and diffracted light can be modified. However, image contrast is reduced at greater transmittance values of the phase annulus, as it can be easily demonstrated with vector diagrams. For example in Fig. 9 (left), if transmittance increases from 25% (DL or ADL pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi lens) ! to 45% (DLL lens), vector e becomes 0:45=0:25 times ¼ !

!

!

ca:1:34 times bigger, thus affecting vector f ¼ | c + e | and reducing ! ! the background-to-object luminance ratio (| e |2/| f |2) from ca. 12 to ca. 9 (contrast reduced). If transmittance increases to 100% (as in the phase plate shown in Fig. 17A) the backgroundto-object luminance ratio becomes only ca. 3.4 (rather poor contrast). Contrary to that, apodization reduces the halo (shade-off) artifacts at the expense of only negligible contrast deterioration in the halo (shade-off) artifact-free regions. To some extent, the phenomena described in this paragraph may be inspected online (see link in Fig. 15 legend). Yet another system also referred to as “variable” is represented by a double phase annulus setup “Phv” (Fig. 17B, Suppl. S3c) described in Subheading 3.3. It is employed, for example, in variable relief (VAREL) contrast described in the following section. 3.2 Relief Phase Contrast

4

As shown in Figs. 12A and 16A, relief contrast is often sufficient in removing or reducing the halo (shade-off) artifacts. However, as explained earlier, it yields poor contrast in thin specimens, and phase contrast is preferable in such cases (Fig. 12C, 16C). A combination of relief contrast and phase contrast, referred to here as “relief phase contrast” (Fig. 13), features advantages of both modes. It can be easily setup by partially masking the annular diaphragm in a conventional phase-contrast condenser, for example, with a piece of black cardboard or thin metal sheet. An example of such an adaptation is shown in Fig. 13 (bottom right). Ideally, the phase annulus should also be adapted accordingly. However, as predictable from the theory of image formation in phase-contrast microscopy (Fig. 6), this would improve the image quality only marginally. Indeed, a phase-contrast objective may be used as a standard one if the annular diaphragm is disengaged (standard bright-field illumination).

http://mikroskop-online.de/Mikroskop%20Druckschriften.htm

CONVENTIONAL PHASE CONTRAST

RELIEF CONTRAST

A

50 μm IN AIR

B

100 μm IN OIL

C

O B J E C T T O M E D I U M P H A S E S H I F T

100 μm

CONDENSER DIAPHRAGM

CONVENTIONAL PHASE PLATE

CONDENSER DIAPHRAGM

PHASE PLATE UNNECESSARY

Fig. 12 Phase contrast and relief contrast as complementary imaging modes. Note the effect on image contrast and artifacts of varying the phase shift between the object and the surrounding medium (proportional to optical path difference, OPD). Each image pair represents an identical viewing field. Objectives,
10/0.25, with and without phase plate (conventional phase contrast [negative] and relief contrast, respectively). N.B.: In the relief-contrast mode, near-identical images are obtained with and without phase plate. (A) A highly indented portion of radula isolated from Spanish slug (Arion lusitanicus). The halo (shade-off) artifacts completely degrade the phase-contrast image. The relief-contrast image is reproduced by permission of © The American Physiological Society [14]. A thinner portion of the same specimen is shown in Figs. 5B and 10. (B) Microcrystals in pickled pepper brine evaporate mounted in air (large phase shift). Profound halo (shade-off) artifacts blanket the entire phase-contrast image. (C) The same specimen as in B (an immediately adjacent viewing field), mounted in paraffin oil (small phase shift); the phase-contrast image is artifact-free (cf. Subheading 4.2)

296

Radek Pelc et al. CONVENTIONAL PHASE CONTRAST

RELIEF PHASE CONTRAST

A

100 μm

B

100 μm

CONDENSER DIAPHRAGM

CONVENTIONAL PHASE PLATE

CONDENSER CONVENTIONAL DIAPHRAGM PHASE PLATE

Fig. 13 Relief phase contrast. In a thick specimen (A), the off-axis (relief) illumination introduces shading, a visual cue to depth structure, thus facilitating image comprehension. The image of a thin specimen (B) is largely unaffected by the asymmetric diaphragm modification. A similar, commercially available imaging setup is referred to as “variable relief contrast” or VAREL [34]. (A) Leaf epidermal replica from acid sorrel (Rumex acetosa), in a transparent acrylate resin. Objective,
10/0.25 (conventional phase contrast [negative]). Angular size of the illuminating diaphragm segment ca. 90 . Originally presented as a conference paper [17]. (B) Human osteoblast-like MG-63 cells (unstained) provided jointly by HloDr Nora Sˇesˇa´k-Kra´lı´kova´ and MUDr Lucie Bacˇa´kova´ (both of them at Institute of Physiology, Czech Academy of Sciences, Prague). Objective,
10/0.30 DLL (conventional phase contrast [positive]). Angular size of the illuminating diaphragm segment ca. 120 (rather than 90 shown in the image)

In a thick specimen (Fig. 13A), relief phase contrast reduces the halo (shade-off) artifacts (a relief contrast mode prevails). In a thinner specimen (Fig. 13B), it yields an image that is near-identical with a conventional phase-contrast one (a phase-contrast mode prevails). In this way, relief phase contrast broadens the range of specimen thicknesses that can be examined with a single imaging mode. Some of these data have been presented elsewhere [17].

Phase-Contrast Microscopy

297

Various other designs of this kind have been presented elsewhere [34, 37, 45–49] with images shown in [37, 45, 48, 49]. Candle soot employed in a variant of reduced-carrier single-sideband microscopy [45] both attenuates and shifts the phase of light [7, 8, 38]. Relief phase contrast has also been combined with apodization; no images are available [46, 47]. The only commercially viable form is VAREL (variable relief) contrast made by Zeiss [34]. An objective with two phase annuli is employed (Fig. 17B), the smaller/bigger one for conventional/relief-phase contrast; a 360 / 120 annular diaphragm is used (cf. Fig. 13). Except for the relief illumination aspect, the VAREL system capable to easily switch between bright-field, phase-contrast, and dark-field illumination is achieved in a similar way as in the Heine condenser5 by Leitz (Suppl. S4c). The relief illumination modes introduce into the images shading, one of the so-called visual cues to depth structure. These cues also include contours (silhouette) and texture, and facilitate image comprehension [50], as illustrated in Fig. 14. As assessed by functional magnetic resonance imaging (fMRI), horizontal stimuli (Fig. 14B) bring about a stronger activity in the visual cortex of human brain than vertical ones (Fig. 14A). This finding may have its origin in the evolution of visual perception mechanisms in human brain, in that the prevailing illumination direction (to which the brain is “adjusted” at rest) is vertical, from a single star, the Sun [51]. No such differences in the fMRI signal map could be observed with control stimuli featuring no shading (Fig. 14C, D). These phenomena are important in stereomicroscopy that also utilizes a binocular cue called parallax, and are discussed in greater detail elsewhere [3, 14–16]. It should be borne in mind that while relief contrast is suitable as an observation device, shapes of certain structures can be easily misinterpreted (boxed details in Fig. 16A). Consequently, it should not be used as a measuring device. In other words, relief phase contrast observation is useful in the initial screening of mediumthick specimens but should be complemented by a careful examination under a (non-relief) phase contrast. 3.3 Apodized Phase Contrast and Related Modes

5

Apodization is commonly used in optics to attenuate the diffraction effects at edges. In microscopy, it is employed in a refined form [52] of Hoffman modulation contrast [27], by using a fine transmittance grading of the modulation filter. In phase-contrast microscopy, apodization represents a relatively recent improvement [16, 17, 31, 32, 42, 47, 53, 54] and is described here in greater detail. It is capable of reducing the halo (shade-off) artifacts and to extend the usable optical path difference range (broadly speaking,

https://archive.org/details/Leitz

298

Radek Pelc et al.

SHADING STIMULI

CONTROL DISPLAYS

(c)

(a)

vertical stimulation (b)

(d)

horizontal stimulation (b) minus (a) fMRI signal

(d) minus (c) fMRI signal

Horizontal > Vertical

Vertical > Horizontal

fMRI signal origin (yellow box)

Red line denotes the plane of section (left)

Fig. 14 Effects of shading on the visual cortex activity in human brain. (Top) Shading stimuli (a, b) as visual cues to depth structure, and control displays without shading (c, d); stimulus orientation either vertical (a, c) or horizontal (b, d). (Bottom) Differential fMRI signal map obtained by subtracting the activation map for the vertical shading display (a) from that for the horizontal one (b); yellow indicates the strongest differential signal (dashed box). A comparison with control displays featuring no shading (c, d) is also shown; no differential signal (d minus c) could be detected in the brain. Reprinted from Ref. [50] by permission of © Elsevier (layout slightly modified for clarity)

optical or specimen thickness) of the specimens. The terminology is not entirely consistent as some authors refer to “apodization” as “modulation” instead [47]. The principles of halo (shade-off) artifact generation and its reduction by apodization are illustrated in Fig. 6. The greater the object size (laterally), the smaller its spatial frequency and the angle of light diffraction [55]. As a result, objects with large lateral

Phase-Contrast Microscopy

299

dimensions (e.g., whole cells) diffract light to rather small angles. The diffracted light thus traverses the phase plate very close to the phase annulus, and a substantial proportion passes through the phase annulus itself. The greater this proportion, the stronger the halo (shade-off) artifacts; ideally, only direct (undiffracted) light should pass through the phase annulus. Contrast in the affected regions is thus rather poor as the relative phase of direct and many diffracted photons is the same as in bright field. In very thin specimens (e.g., cell extensions spreading on a microslide), the artifacts are insignificant, regardless of their lateral dimensions. They only become prominent in those large objects that are also thick, or, to be more precise, in specimens featuring large enough phase shifts (proportional to optical path differences). This condition is fulfilled, for example, in cell clusters and illustrated in Figs. 12A, B and 16A. The reason is that, the magnitude of diffracted light increases with the object-to-medium phase shift, as mentioned above and illustrated with vector diagrams in Figs. 5B, 9 and 10. The halo (shade-off) artifacts can thus be reduced if the so-called apodization annuli (neutral density filters, i.e., filters inducing no phase shift) are placed immediately next to the phase annulus itself (Fig. 6). Ideally they should have a graded transmittance [32, 53]. The effects of apodization are demonstrated in Figs. 15 and 16. The fine “ridges” (or “spines”) shown in Fig. 16B are only visualized in apodized phase contrast. Note that the specimen is a leaf replica, thinnest at cell centers. Apparently, the phase shifts in the specimen are too small for relief contrast to be sensitive but too large for the conventional phase-contrast image to be free of the halo (shade-off) artifacts. Apodized phase contrast is capable of visualizing such structures well while relief contrast only reveals the overall cell shape. If the object is very thin (image almost free of the halo [shadeoff] artifacts) the contrast actually slightly deteriorates due to apodization (Fig. 16C). The reason is that, the apodization annuli attenuate some of the diffracted light. As a result, the direct light becomes relatively less attenuated (compared to the diffracted light) and contrast is reduced. Nevertheless, the contrast deterioration due to apodization is very small indeed, and this minor drawback is strongly outweighed by the profound halo (shade-off) artifact reduction in thicker objects, which often manifests itself as an apparent contrast increase in the image as a whole (Fig. 15). In Fig. 15, this is so despite the fact that the transmittance of the phase annulus is greater in the ADL than in the DM lens (25% vs. 14%, apodized vs. conventional phase-contrast image). As explained in Subheading 3.1, greater transmittance of the phase annulus result in lower contrast and, to some extent, also in halo (shade-off) artifacts reduction. A rigorous comparison of the DM and ADL images is thus impossible because ideally, either DM

300

Radek Pelc et al. CONVENTIONAL PHASE CONTRAST

APODIZED PHASE CONTRAST

20 μm

CONDENSER DIAPHRAGM

CONVENTIONAL PHASE PLATE

CONDENSER DIAPHRAGM

APODIZED PHASE PLATE

Fig. 15 Effect of apodization and phase annulus transmittance on image quality in phase contrast. Neuroblastoma–glioma hybrid (mouse–rat) glial cells (NG108-15). Halo (shade-off) artifacts are marked with arrows. Halo (shade-off) artifact reduction is mainly due to apodization, partly due to greater transmittance of the phase annulus (DM 14%, ADL 25%). Contrast was separately optimized in the main images and cutout details. Objectives (2002 catalogue models),
40/0.60 DM dry (conventional phase contrast [positive]) and
40/0.60 ADL dry (apodized phase contrast [positive]). Images (trimmed) are reproduced from Ref. [42] by permission of © The Biophysical Society of Japan (layout slightly modified for clarity). Other examples may be inspected online (https://www.microscopyu.com/techniques/phase-contrast/apodized-phase-contrast)

and ADM, or DL and ADL lenses should have been used in Fig. 15 (such combinations were commercially unavailable when the images were acquired). Examples of other designs aimed at halo (shade-off) artifact reduction, often trial-and-error attempts [6, 35, 36, 38] are presented in Fig. 17. Phase contrast employing a 100% transmittance phase plate (design A) used to be known as “B-minus contrast” [6]. It reduces the artifacts but also significantly deteriorates image contrast of details of any size, that is, including small ones. Design B (“strenger Phasenkontrast”) shown in Fig. 17 and Suppl. S3c is based on Beyer’s “variable phase contrast” system utilizing two annular diaphragms and two phase annuli (“Phv” objectives in older Zeiss microscopes). The halo (shade-off) artifacts are reduced not only at smaller annulus width (see above) but also at smaller annulus diameter, as described in detail elsewhere [7, 8, 35]. To some extent, the halo (shade-off) artifact removal can be achieved simply by reducing the width of the annular aperture

Phase-Contrast Microscopy

CONVENTIONAL PHASE CONTRAST

APODIZED PHASE CONTRAST

301

RELIEF CONTRAST

A

100 x 100 μm

B

200 x 200 μm

C

200 x 200 μm

CONDENSER CONVENTIONAL DIAPHRAGM PHASE PLATE

CONDENSER DIAPHRAGM

APODIZED PHASE PLATE

CONDENSER DIAPHRAGM

PHASE PLATE UNNECESSARY

Fig. 16 Effect of apodization in phase contrast, and a comparison with relief contrast. Note the dramatical image quality improvement by apodization in a thicker object (A) and only a minor contrast deterioration in a thin object (C). (A) Coach grass (Agropyron repens), leaf epidermal replica. (B) Catchweed (Galium aparine), leaf epidermal replica. (C) Human buccal epithelial cells. Objectives,
10/0.25 with no phase plate (relief contrast, A and B),
10/0.30 DLL (relief contrast, C),
10/0.25 DL (conventional phase contrast [positive], A–C) and
10/0.25 ADL (apodized phase contrast [positive], A–C). Images A and B (trimmed) are reproduced from Ref. [16] by permission of © SPIE. N.B.: In the relief-contrast mode, near-identical images are obtained with and without phase plate (C)

302

Radek Pelc et al.

OBJECTIVE

CONDENSER

A

B

C

D

Bennett et al. (1951)

Beyer (1953)

Peshkov (1955)

Soran and Diaconeasa (1957)

Phase shift (A)

Phase shift and attenuation (B, C, D)

Color filter (D)

Fig. 17 Alternative (non-apodizing) designs capable of partially reducing the halo (shade-off) artifacts (cf. apodized phase plate, Fig. 6). (A) “B-minus contrast” that used to be made by American Optical (Spencer) Co (Suppl. S4b); it is only suitable for thicker specimens yielding a high intensity of diffracted light [6]. (B) “Variable phase contrast” (Zeiss) employing “Phv” objectives nowadays used in VAREL contrast [34], utilized a very narrow phase annulus [35] (Suppl. S3c). (C) A significant portion of diffracted light is attenuated but also phase-shifted which is not ideal [36] although a more refined form works well [37]. (D) Conventional phase contrast combined with color (Rheinberg-type) dark field [38]

alone [37], with a device similar to Heine condenser (Suppl. S4c). However, as the width of the phase annulus is not correspondingly reduced, it is not an optimal solution (cf. Fig. 17C). Design C shown in Fig. 17 features an element of spatialfrequency filtering (cf. apodized phase plate in Fig. 6) but a significant portion of the diffracted light is subject not only to attenuation but also to the same phase shift as direct light, which eliminates the chief advantage of phase-contrast imaging. Only edges and very small structures (objects with high spatial frequencies) remain visualized in high contrast [36]. The efficiency of the design D shown in Fig. 17 ([negative] phase contrast combined with color dark field [Rheinberg-type] illumination) is demonstrated by its authors in epidermal cells of onion bulb scales [38]. In the opinion of the authors of the present chapter, results of comparable or better quality can be obtained in this specimen simply by using relief phase contrast or relief contrast alone (images not shown). Related to this design is a note about Heine condenser in Subheading 3.2 describing VAREL contrast. Yet another strategy to remove the halo (shade-off) artifacts is based on employing either a small circular off-axis light source (diaphragm) rotating about the microscope’s optical axis [32], or

Phase-Contrast Microscopy

303

an array of stationary small circular diaphragms randomly distributed across the entire condenser aperture [56]. In either case, phase discs conjugate with each of the circular aperture (s) are part of the phase plate (cf. Fig. 7A). Some recent variable phase-contrast systems are also mentioned in [56]. 3.4 Pupil-Projection Apodized Phase Contrast with Fluorescence

If a near-simultaneous phase-contrast and epifluorescence data acquisition is necessary, it is inconvenient (oil immersion objectives), and sometimes impossible (motile cells) to switch the objective lenses. In such cases, it is desirable to use an external phase plate located outside the objective lens casing, in a pupil-projection (optically relayed) plane [54]. Older systems (intended for variable phase-contrast rather than fluorescence microscopy) are described in Subheading 3.1 and in Suppl. S3. The combined phase-contrast and fluorescence image (Fig. 19f4) facilitates an interpretation of the fluorescence signal, especially in regions where it is weak and its location in the cell/ organism is unrelated to any clearly discernible structural landmarks, as it is often the case (Fig. 19f2, f3). The phase-contrast mode thus adds a “night-vision” quality to the often mostly dark fluorescence images. Apodization extends the specimen thickness that can be reasonably examined, owing to the halo (shade-off) artifact reduction (“fog-lights” effect). As such, the combination of phase-contrast and fluorescence microscopy is not new. However, to the best of authors’ knowledge, phase contrast has never been combined with total internal reflection fluorescence (TIRF) microscopy requiring a dedicated objective lens. Here we present such data acquired with a single objective lens. The optical layout employed in the present study is shown in Fig. 18 and is compatible with any Nikon Eclipse Ti series microscope (apodized phase annuli available for high-power objectives). Leica offers a similar system referred to as “Integrated (or Intermediate) Phase Contrast” (IPH), with the phase plates (conventional phase annuli available for objectives of any magnification) located in a special slider (“integrated interpupillary interface”). A number of points should be borne in mind when using or building a pupil-projection (optical-relay) setup, i.e., that accommodating an external phase plate: (1) The pupil-projection lens should be of sufficiently good quality to prevent deterioration in image quality. (2) If a mirror is inserted in the optical path (as, e.g., in the optical setup shown in Fig. 18) a special attention should be paid to its flatness. A nanometer-range precision is required to prevent interference effects degrading the images. (3) If polarizing filters are used (as, e.g., in the systems shown in Fig. 11 and Suppl. S3a, d, e) light intensity attenuation may become a problem as each polarizer reduces the intensity to 1/2. Moreover, as a quarter-wave

304

Radek Pelc et al.

PUPIL-PROJECTION (OPTICAL-RELAY) SETUP

STANDARD SETUP

EVANESCENT (TIRF) WAVE IN THE CELL

ANNULAR APERTURE

CELL

CONDENSER

COVERSLIP

SPECIMEN EXTERNAL APODIZED PHASE PLATE

CCD CAMERA

OBJECTIVE LENS APODIZED PHASE PLATE

ILLUMINATION THROUGH A HIGH-N.A. OBJECTIVE

TOTAL INTERNAL REFLECTION

CCD CAMERA

Fig. 18 A diagrammatic layout of an external apodized phase plate setup and TIRF recording. (Left) The external (pupil-projection) apodized phase plate is located in an optically relayed plane conjugate to the objective back focal plane (approx. the exit pupil in low-power lenses); cf. Suppl. S3. Apodized phase-contrast and fluorescence images (Fig. 19) can thus be acquired without switching the objective lens. A standard setup with the apodized phase plate inside the objective lens casing is shown for comparison. (Right) The principle of total internal reflection fluorescence (TIRF) recording. Illuminating light strikes the coverslip-specimen interface at a sharp (smaller-than-critical) angle, so that total internal reflection back to the glass occurs. Only an evanescent wave penetrates into the specimen, typically only ca. 200 nm deep. Near-membrane fluorescence only is thus excited in coverslip-adherent cells

plate is employed, a monochromatic light only may be used. Multiwavelength quarter wave plates are available but they are not perfect. 3.4.1 Neuronal Growth Cones of Neuroblastoma– Glioma Hybrid Mouse–Rat Cells (NG108-15)

For EGFP (excitation/emission 488/507 nm), GFP-HQ filter set (Nikon) was used: excitation filter (Ex 455–485 nm), a dichroic mirror (DM 495 nm), and a barrier filter (BA 500–545 nm). For mCherry (excitation/emission 587/610 nm), Texas Red filter set (Nikon) was used: excitation filter (Ex 540–580 nm), a dichroic mirror (DM 595 nm), and a barrier filter (BA 600–660 nm). In the TIRF mode, the specimen was illuminated through the periphery of a high-NA objective lens diaphragm (Fig. 18). This guarantees that the light strikes the specimen at a smaller-thancritical angle, and total reflection occurs. Only an evanescent wave penetrates into the specimen. As it decays exponentially with distance, its intensity is appreciable only within ca. 200 nm from the glass–cell interface. Near-membrane fluorescence only is thus excited in coverslip-adherent cells (Fig. 19C, E). Apodized phase-contrast images (Fig. 19A) were recorded in stained cells at 538–558 nm (peak value 548 nm, filter half-width 10 nm). This guaranteed that there was no effect of staining on the apodized phase-contrast images. In Fig. 19A–E, the recording

Phase-Contrast Microscopy

NG108-15 cell

FASCIN - EGFP

ACTIN - mCherry

* APODIZED PHASE CONTRAST

10 μm CONVENTIONAL FLUORESCENCE

Caenorhabditis elegans

F1

305

TIRF

TIRF

CONVENTIONAL FLUORESCENCE

TOUCH RECEPTOR NEURONS (mec- 4 :: GFP)

F2

*

F3

*

F4

x2

* APODIZED PHASE CONTRAST

CONVENTIONAL FLUORESCENCE

20 μm F1+F3 OVERLAY

Fig. 19 Near-simultaneous recording with a single objective lens of apodized phase-contrast (positive), conventional epifluorescence and TIRF images. (A–E) Neuronal growth cones (bright box in A) of a neuroblastoma–glioma hybrid mouse–rat cell (NG108-15); EGFP and mCherry staining for fascin and actin. Objective,
100/1.49. (F) Touch receptor neurons expressing recombinant GFP in Caenorhabditis elegans. Objective,
60/1.20. (F1, F2) Corresponding images from a z-stack (no deconvolution). (F3) Z-max intensity projection from the entire z-stack of deconvolved GFP fluorescence images, highlighting the neurons throughout the worm body. (F4) Combination (an overlay of F1 and a color version of F3). The star symbol (∗) denotes details in which contrast was separately optimized (A, F2 and F3). The co-authors of micrographs are Satoe Ebihara (A–E) and Motomichi Doi (F1–F4), both at Biomedical Research Institute AIST, Tsukuba, Japan. Micrographs A–C and F4 are reproduced by permission of © Nikon Corporation, Tokyo (Eclipse Ti-E prototype data 2007), setup as in Fig. 18 (external apodized phase plate; Ph-4 in A-E and Ph-3 in F1–F4)

sequence was A–B–C–D–E. The same cell type was also observed without staining (Fig. 15), using a standard phase-contrast setup (i.e., with the phase plate inside the objective). 3.4.2 Caenorhabditis elegans (Strain SK4005, Genotype zdIs5)

This is a transgenic strain expressing GFP in touch receptor neurons [57]; the promoter of mec-4 gene [58] is cloned into a GFP vector. GFP is thus expressed in cells in which the promoter of the mec-4 gene is active, that is, in touch receptor neurons, rendering them visible (fluorescent). However, the GFP expressed in this particular strain is slightly modified, making it more fluorescent than wild-type GFP but less so than EGFP.

306

Radek Pelc et al.

mec-4 (mechanosensory abnormality) gene itself encodes an amiloride-sensitive Na+ channel protein, degenerin (768 amino acids) required to sense gentle mechanical stimuli (e.g., touch) along the body wall. MEC-4 protein is paralogous to, and coexpressed with, MEC-10, in all six touch receptor neurons in C. elegans [58], as described in WormBase (wormbase.org). The filters (GFP-HQ set for GFP, and 548 nm filter for apodized phase contrast) were the same as used with the NG108-15 cells. Two z-series stacks were recorded (2
221 images, step size 200 nm), apodized phase contrast and GFP fluorescence. The z-position was controlled with 50 nm precision (“Perfect Focus System,” Nikon). The recording sequence was F1 ... F2 ... z-shift ... F1 ... F2 ... z-shift, and so on. The images shown in Fig. 19f1 and f2 represent corresponding frames from the z-stack (vertical position #153 in both images). The GFP fluorescence images recorded in C. elegans were processed in two steps: 1. deconvolution (“NIS-Elements” software, Nikon) of all images in the z-stack; 2. maximal intensity projection (“ImageJ” software, version 1.40g, National Institutes of Health, Bethesda, MD, USA) applied to the entire z-stack of deconvolved images. Each pixel in the resultant image (Fig. 19f3) contains the highest GFP fluorescence intensity detected along the z-axis. In this way, the touch receptor neurons were conveniently visualized and localized. The apodized phase-contrast image (Fig. 19f1), recorded in the already stained preparation at 538–558 nm (peak value 548 nm, filter half-width 10 nm), was overlaid with the maximum intensity projection fluorescence image (Fig. 19f3), yielding Fig. 19f4. The depth of field in high-magnification phase-contrast images is inherently small. This is particularly noticeable owing to the halo (shade-off) artifact reduction by apodization, thus enabling more details deep in the object to be assigned to a particular plane along the z-axis. To some extent, “optical sectioning” is thus possible. 3.5 Phase Contrast in Electron and X-Ray (Synchrotron) Microscopy

Much like in light microscopy, it is usually preferable in electron microscopy (EM) to observe a specimen in its natural state, that is, without staining. As image contrast is low in such specimens, for example, in cryosections or unstained protein particles, optical means of contrast enhancement are required. The possibility of phase-contrast imaging in electron microscopy has been known, at a theoretical level, for a long time. However, technological obstacles, mainly charging and thermal drift of the phase plate kept preventing the phase contrast from becoming a widespread mode of observation in EM. Instead, it was necessary to resort to the so-called defocus phase contrast, meaning of which is the same as “imitation phase contrast” in the present chapter (Fig. 4F, G). Recently, many of the problems related to the phase plate that can be made of vacuum-evaporated amorphous carbon film, have been

Phase-Contrast Microscopy

307

resolved, as described in brief surveys [59, 60]. Other-than-carbonfilm phase plates are also available, namely, electrostatic phase plates [60], an equivalent of some of the variable phase contrast systems in light microscopy (Suppl. S3a, d, e). Unlike in optical phase contrast, there is nothing like a narrow phase annulus in the EM version of phase-contrast setup. Instead, a phase plate with a tiny central hole for unscattered electrons is used (Fig. 20B). Such design is complementary to that shown in Fig. 7A, and is referred to as Zernike phase contrast (ZPC) by some authors [62, 63]. As unscattered electrons are not attenuated it represents an EM equivalent of the B-minus phase plate in light microscopy (see Subheading 3.3). A phase shift of 90 (retardation) is applied to scattered (diffracted) rather than unscattered electrons, yielding an equivalent of positive phase contrast in light microscopy. Its applications include, for example, single particle analysis of protein molecules (Fig. 20B) or visualization of bacteriophages (Fig. 20C, videos as Suppl. S2a, S2b). Owing to a greater contrast, higher effective resolution is achieved, and a considerably smaller number of particles need to be sampled to make a 3D reconstruction [62]. Note that the phase-contrast image shown in Fig. 20C was post-processed (contrast inverted), yielding bright objects on dark background, as in negative phase contrast [63]. For somewhat larger (thicker) objects, over ca. 50 nm, a halfcircle phase plate (Fig. 20A) is more beneficial as the contrast transfer function is greater at smaller spatial frequencies (i.e., for larger objects) than in ZPC. Some authors call it Coherent Foucault Imaging, thus highlighting the Foucault knife edge test, also used in light microscopy [18] (Fig. 3A) and to detect imperfections of astronomical mirrors [9, 10]. Recently, another term has been coined, Hilbert differential contrast (HDC), emphasizing the Hilbert transform inherent in it [59]. Much like the relief phase contrast in light microscopy (Fig. 13), HDC introduces a visual cue (shading) into the images (Fig. 20A). For this reason, the contrast is sometimes referred to as “topographic contrast” [61], absent in the ZPC images (Fig. 20B, C). Apart from the abovementioned phase plates, there is now a more modern type in which the phase shift is induced by the electron beam. Attempts have been made to adapt the phasecontrast theory to thicker objects [64–66], to include phaseamplitude ones inducing phase shifts greater than 360 (Nagayama K., Okazaki, Japan—personal communication). Except for EM tomography [63], a 3D reconstruction of a specimen is often impossible in EM as it requires a large number of serial sections to be prepared. In some cases, this problem can be solved by X-ray imaging. Compared to electrons (in EM), X-rays penetrate the specimen to a much greater depth and enable nondestructive examination. Highly collimated, intense X-ray sources (synchrotrons) are suitable in this respect. Owing to

PHASE-CONTRAST TEM

A

CONVENTIONAL TEM

ELECTRON BEAM SPECIMEN OBJECTIVE LENS

BACK FOCAL PLANE

PHASE PLATE

PROJECTION LENS S.E. U.E.

200 nm IMAGE PLANE

B

ELECTRON BEAM SPECIMEN OBJECTIVE LENS

BACK FOCAL PLANE

PHASE PLATE

PROJECTION LENS S.E. U.E.

20 nm IMAGE PLANE

C IMAGING SETUP AS IN B PLUS CRYO-EM TOMOGRAPHY

+ DNA

+ DNA

100 nm 3D TOMOGRAM AS A VIDEO

3D TOMOGRAM AS A VIDEO

(Suppl. S2 A)

(Suppl. S2 B)

Fig. 20 Phase contrast in transmission electron microscopy (TEM). Conventional TEM images (slightly defocused) are shown for comparison. SE, scattered electrons. UE, unscattered electrons. (A) Hilbert differential contrast (π-type asymmetric phase plate, phase shift 180 ); unstained freeze-substituted proximal tubular rat kidney epithelial cells (ultrathin section ca. 100 nm). (B) Zernike phase contrast (π/2-type symmetrical carbon phase plate ca. 32 nm thick, phase shift of SE  90 at 300 kV [positive phase contrast], central hole ca. 0.5 μm in diameter, not to scale); unstained ice-embedded GroEL chaperonin protein particles. (C) Zernike phase contrast (carbon phase plate ca. 24 nm thick, phase shift of SE  90 at 200 kV [positive phase contrast]) combined with cryo-electron tomography (70 tilt angle); unstained ice-embedded T4 bacteriophages with their capsids either empty (“+”, bright outline) or filled with DNA (bright contents), contrast inverted for clarity, arrows mark phage tails, images are 5 nm slices through reconstructed 3D tomograms. Optical diagrams (adapted) in A and B are reproduced from Ref. [59] by permission of © The Biophysical Society of Japan. Micrographs/videos (trimmed) are reproduced from Refs. [61] and [62, 63] by permission of © Kluwer/Springer (A) and © Elsevier (B, C), respectively

Phase-Contrast Microscopy

309

microtomographic reconstruction algorithms (SR-μCT, synchrotron radiation microscopic computed tomography), small organisms can be examined directly in vivo, at resolution in the μm range [67]. As absorption differences within the specimen are often small (and close to zero in phase objects), phase-contrast imaging is conveniently employed [68–70]. An example of a tomographic reconstruction from X-ray (synchrotron radiation) phase-contrast images of nervous tissue is shown in Fig. 21 (the trigeminal nerve of a rat). A phaseinterferometric setup was used [69]. In a way, it is analogous to the “Interphako” method in light microscopy [7, 8] (Suppl. S3b). High-resolution mass density mapping is possible, potentially enabling to detect, nondestructively, brain microtumors at an early stage.

4

Experimental Demonstrations

4.1 Imitation Phase Contrast

1. Prepare buccal epithelial cells as described in Subheading 5. Alternatively, use microcrystals prepared as described in the next demonstration. 2. At standard bright-field illumination conditions (condenser diaphragm set to match as closely as possible the numerical aperture of the objective lens), focus the image of the cells. At precise focus, the cells almost vanish (image contrast much worse than in Fig. 16C [right]). 3. Slightly under-focus the image (specimen a few μm closer to the objective than at perfect focus). Take an image (Image X) 4. Slightly over-focus the image (specimen a few μm farther from the objective than at perfect focus). Take an image (Image Y). 5. Subtract the two acquired images. To prevent intensity values becoming negative or saturating at level 255 (eight-bit grayscale), the subtraction formula should be one of the following: Positive imitation phase‐contrast image ¼ ½ðImage X  Image Y Þ þ 127 Negative imitation phase‐contrast image ¼ ½ðImage Y  Image XÞ þ 127 6. Compare the result with a genuine phase-contrast image (Fig. 16C), and with the examples shown in Fig. 4F, G. Various image-processing software can be used to perform such simple operations (e.g., ImageJ [freeware], analySIS, Adobe Photoshop). A dedicated software, QPI SDK (Quantitative Phase Imaging software development kit) has been designed specifically

310

Radek Pelc et al.

A

STORAGE RING

B

NaI detector interferometer

sample

BOOSTER RING BENDING MAGNET

CCD

WIGGLER

SR liquid cell

beam stop monochromator phase shifter

ELECTRON GUN LINAC UNDULATOR

MICROTOMOGRAPHY BEAMLINE

BEAMLINES

mass density

C 0.913 g/cm3

1.048 g/cm3

NERVE

NERVE

BRAIN

BRAIN

1 mm BRIGHT FIELD

PHASE - INTERFERENCE CONTRAST

Fig. 21 Phase-contrast interference microtomography of nervous tissue, as performed with synchrotron radiation. Microtumors may potentially be detected. (A) An interferometric setup. The interference pattern is modified by elastic forward diffraction of the specimen and recorded as a phase-shift projection. The liquid cell enables refractive index matching of the sample, the NaI detector monitors the interferometer setting and the “phase shifter” acts as a phase plate. By blocking the lower (reference) beam, phase contrast can be disengaged to generate an absorption image. SR, synchrotron radiation source (BW2 beamline, positron storage ring DORIS, Hamburg Synchrotron Laboratory DESY). (B) A diagrammatic layout of a synchrotron radiation (X-ray) source. LINAC, linear accelerator. Positrons may be used instead of electrons. (C) Paraffinembedded intracranial part of the trigeminal nerve of a rat, ~20 days after exposure on postnatal day 1 to Nethyl-N-nitrosourea (EtNU), a carcinogen. (Left) A classical histological section (cell nuclei stained with azan). (Right) X-ray phase-contrast microtomogram of unstained unsectioned specimen, obtained at 12 keV. Arrows denote areas of the highest mass density. Vertical separation of the two images 40 μm, voxel size 5.4 μm. Reproduced from Ref. [69] by permission of © The Biophysical Society (A, C), and from Ref. [70] by permission of © Blackwell (Wiley) (B). Layout is slightly modified for clarity (B, C)

for this purpose, that is, to obtain phase-contrast images without having to use a phase-contrast microscope [30]. It is claimed to be capable of generating other image types as well (Hoffman modulation contrast, differential interference contrast, and dark field), and

Phase-Contrast Microscopy

311

used to be available on a commercial basis from Iatia Vision Sciences. 4.2 Halo (Shade-off) Artifact Reduction by Medium Replacement

1. Place a droplet of salt solution on a microslide and let it evaporate. Microcrystals can be seen under the microscope. Brine from pickled vegetable (celery, pepper, cucumber, etc.) is another easy option (Figs. 4A–E and 12B, C). 2. Place a small drop of paraffin oil (refractive index ca. 1.43) at the edge of the evaporation residue area. 3. Cover ca. half of the evaporation residue with a coverslip while allowing the paraffin oil to fill the space between the microslide and the coverslip, that is, “flooding” the microcrystals in ca. half of the preparation. 4. Images of crystals that are not covered with paraffin oil are obscured by profound halo (shade-off) artifacts (Fig. 12B); the phase shift, proportional to the refractive index difference between the crystals and the air, is too large. 5. Images of crystals that are surrounded by paraffin oil are almost entirely halo (shade-off) artifact-free (Fig. 12C). This type of demonstration can also be used to highlight the complementary nature of phase-contrast and relief-contrast imaging [14, 16]. The procedure can be further refined by using other liquids of known refractive indices. Cargille Laboratories (Cedar Grove, NJ 07009-1289, USA) supply special optical fluids of welldefined refractive indices as well as specially prepared comminuted minerals and optical glasses, some of them already mounted on microslides. In living cells, this type of demonstration may be carried out using 21% bovine serum albumin6.

5

Specimens Invertebrate neurons of Caenorhabditis elegans (strain SK4005, genotype zdIs5) were studied directly in vivo. The strain (see Subheading 3.4.2 for details) was obtained from Caenorhabditis Genetics Center at University of Minnesota. C. elegans was then mildly anesthetized by 10 mM levamisole in the external medium. Neuroblastoma–glioma hybrid mouse–rat NG108-15 cells were obtained from American Tissue Culture Collection (ATCC) and stained with enhanced green fluorescent protein (EGFP) for fascin and with mCherry for actin. Human osteoblast-like MG-63 cells were obtained from European Collection of Cell Cultures (ECACC, Health Protection Agency, Salisbury, UK) and cultured at 37  C for two days at

6

Barer R (1954) Naturwissenchaften 41(9):206–208. https://doi.org/10.1007/BF00623016

312

Radek Pelc et al.

Department of Biomaterials, Institute of Physiology, Czech Academy of Sciences, Prague. Buccal epithelial cells were gently scraped from the inner cheek surface of one of the authors (RP) with a stainless steel spatula, dispersed with a surgical needle in physiological solution on a microglass sealed with nail varnish to a coverlip to prevent evaporation. Body scales of silverfish, commonly known as fishmoth or (urban) silverfish (Lepisma saccharina) were prepared as follows: A dead, dried-out adult silverfish was held above a water droplet (ca 5 mm in diameter) placed on a microslide. Silvery scales were removed from the body surface with a firm hair (eyelash) extracted from wild boar skin and fixed to a skewer as a holder, so that most of the scales were landing on the water surface. As soon as the water evaporated, the scales clang to the glass surface. This guaranteed that the scales were completely flattened and in focus even under a
40 objective. The group of scales was covered with a coverslip which was then sealed around its edges to the microslide with Canada balsam, forming a permanent specimen. Note that the silvery scales are only found in adult animals, after their third molting. Other objects (starch granules, microcrystals, slug radula, and plant leaf epidermal replicas) were prepared as described in the original papers [14, 16], Fig. 3 and Subheading 4.2, with the exception of amyloid-β aggregates (undisclosed).

6

Imaging Setups Light-microscopy setups only are shown below (EM and X-ray microscopy [Figs. 20 and 21] is not considered). Specifications not listed here may be found in figure legends. In phase-contrast setups, the illumination diaphragm in the condenser was always of annular type. Filters (wavelengths) are specified for phase-contrast images only as this parameter is relatively unimportant in relief contrast and bright field. Microscopes made by Lambda were fitted with Praktica VLC3 camera and a photographic eyepiece FUx4, and images acquired on Fuji Superia 100 color film. Microscopes made by Nikon were fitted with Nikon DS-Fi1 CCD camera, and images acquired at 2560
1920 resolution, except in (1) Fig. 4H, I where Nikon H-III film camera was used, and (2) Fig. 19 where cooled CCD MicroMAX 512 EBFT camera (Princeton Instruments [Roper Scientific], Trenton, NJ, USA) operating at 40  C and 512
512 resolution was used. Abbreviations: “NA” ¼ numerical aperture, “WD” ¼ working distance, “NCB” ¼ neutral color balance (bluish) filter.

Phase-Contrast Microscopy

6.1 Relief Contrast (Lambda) and Bright Field

313

Setup as in Fig. 3B; does not apply to Fig. 1C (reprinted image). Lambda Praha (formerly Meopta, Czechoslovakia) DN45-BH51 microscope (upright) with a dedicated relief-contrast (off-axis illumination), Abbe-type condenser (RCH-0115 or RCH-0128 model, NA 1.20). Objectives achromatic, no phase plate. Relief-contrast mode: relief diaphragm engaged and condenser iris diaphragm set to a value up to twice higher than objective NA, to increase the relative contribution to image formation of diffracted light (objective aperture acting as a schlieren diaphragm). Bright-field mode: relief diaphragm disengaged, condenser iris diaphragm set to a value matching the objective NA. l

6.2 Relief Contrast (Nikon)

Nikon Eclipse Ti-E microscope (inverted) with LWD condenser (NA 0.52). Objective planfluor originally designed for conventional phase contrast (positive). Relief diaphragm made of black cardboard and placed at the same vertical position as the (removed) annular diaphragm normally used in phase-contrast mode. l

6.3 Conventional Phase Contrast (Negative)

6.4 Relief Phase Contrast (Negative)

Figs. 1D, 3C, D, 4A–D, 5B [top], 12, and 16A, B

Fig. 16C

Lambda Praha (formerly Meopta, Czechoslovakia) DN45-BH51 microscope (upright) with an Abbe-type phase-contrast condenser NA 1.25 (normally used in DN45-BH47Ph model). Objectives planachromatic, transmittance of the phase annulus 25  2% as determined by exposure measurement with Nikon U-III photomicrography unit (phase-contrast versus standard planachromatic objectives
10,
20,
40, and
100). l

Figs. 4E, 8, 9, and 12: No filter

l

Figs. 5B [bottom] and 10: Ordinary (non-interference) green filter (λ  550 nm)

l

Fig. 13A: Tilting interference filter visually adjusted to pass green light

Microscope, condenser and objective same as in conventional phase contrast (negative). A relief diaphragm made of black cardboard and placed close to the annular diaphragm in the condenser (vertical separation ca. 5 mm), thus forming an angular illumination aperture (segment). Tilting interference filter visually adjusted to pass green light. l

Fig. 13A

314

Radek Pelc et al.

6.5 Relief Phase Contrast (Positive)

Nikon Eclipse Ti-E microscope (inverted) with LWD condenser (NA 0.52). Objective planfluor DLL (transmittance of the phase annulus ca. 45%). A relief diaphragm made of black cardboard and placed over the annular diaphragm (i.e., at the same vertical position), thus forming an angular illumination aperture (segment) in the condenser. NCB filter. l

6.6 Conventional and Apodized Phase Contrast (Positive)

6.7 Pupil-Projection Apodized Phase Contrast (Positive) with Fluorescence and TIRF

Fig. 13B

Except in Figs. 1B, 2 and 4H, I, Nikon Eclipse-series microscopes (inverted) with LWD/ELWD condenser (WD 30/75 mm, NA 0.52/0.30) were used as specified below. Transmittance of the phase annulus ca. 25% (DL and ADL lens) or ca. 14% (DM lens). Transmittance of the apodization annuli ca. 50% (ADL lens). Green interference filter (GIF, λ  530 nm) was used except in Figs. 1B and 2 (filter unknown [reprinted images]), and Figs. 9 and 16C (NCB filter). l

Fig. 1B: Microscope make and model unknown (reprinted image)

l

Fig. 2: Zeiss, model unknown (reprinted image)

l

Fig. 4H, I: Nikon E200 + turret condenser (dry, NA 0.90), planachromatic objective

l

Fig. 9: Ti-E + LWD condenser, planachromatic objectives

l

Fig. 10: TS100-F + ELWD condenser, achromatic objective

l

Fig. 15: TE2000-E + LWD condenser, planfluor ELWD objectives

l

Fig. 16A: TE2000-S + LWD condenser, achromatic objectives

l

Fig. 16B: TS100-F + ELWD condenser, achromatic objectives

l

Fig. 16C: Ti-E + LWD condenser, planachromatic objectives

Setup as in Fig. 18; Nikon Eclipse Ti-E (inverted) microscope with CLWD condenser (WD 7.2 mm, NA 0.78). External (optically relayed) apodized phase plate outside the objective lens casing. Transmittance values of the phase/apodization annuli not stated (prototype setup). Other details are listed in Subheading 3.4. l

Figure 19A–E: Apochromatic objective (oil immersion)

l

Figure 19f1–f4: Planapochromatic VC objective (water immersion)

Phase-Contrast Microscopy

7

315

Summary Conventional phase contrast is best suited for very thin specimens such as single neuronal extensions (growth cones); images of thicker objects suffer from halo (shade-off) artifacts. Apodized phase contrast, a relatively recent improvement in phase-contrast imaging, substantially reduces the halo (shade-off) artifacts. Its operational range, in terms of acceptable object-tomedium phase-shift values is greater than in conventional phase contrast but smaller than in relief (phase) contrast. Relief phase contrast can be easily obtained by a zero-cost adaptation of the annular condenser diaphragm, in conventional or apodized phase-contrast microscope. In many an application, this combination makes the best of both worlds in that it spans a wide range of object-to-medium phase shifts. Relief contrast alone (i.e., without the phase-contrast modality) is suitable for thick phase objects only. Its advanced form is referred to as Hoffman modulation contrast and either modality falls into the schlieren imaging category. Both relief contrast alone and relief phase contrast are capable of introducing into the images an important visual cue to depth structure (shading), thus facilitating image comprehension. If a simultaneous phase-contrast and fluorescence recording is necessary, a pupil-projection (optical-relay) setup is useful, in that the phase plate is external, that is, located outside the objective lens casing, in an attachment unit. This permits phase-contrast observation to be carried out with any (non-phase-contrast) objective lens of choice, including the highest-NA lenses suitable for total internal reflection fluorescence (TIRF) recording.

8

Online Resources Apart from web links listed in the main text and Online Supplements (S3, S4 and S5), the following are recommended to locate documents describing atypical (older) phase-contrast microscope models. They help to properly understand modern equipment.

7

l

The Zeiss Archives7

l

Carl Zeiss Jena microscope catalogues (1940–1989)8

l

Carl Zeiss microscope accessories9

https://www.zeiss.com/corporate/int/about-zeiss/history/archives http://mikroskop-online.de/Mikroskop%20Druckschriften.htm 9 http://mikroskop-online.de/Mikroskopiezusatzeinrichtungen.htm 8

316

Radek Pelc et al.

Acknowledgments The authors are grateful to Satoe Ebihara and Motomichi Doi (Biomedical Research Institute AIST, Tsukuba, Japan) for preparing specimens of NG108 cells and Caenorhabditis elegans, to W. Brad Amos (MRC Laboratory of Molecular Biology, Cambridge, UK) and Kuniyaki Nagayama (Okazaki Institute for Integrative Bioscience, Japan) for helpful comments. Petr Pithart (Lambda, Prague); Yoshiro Oikawa, Hiroyasu Tanaka, Masaaki Tamura, Ichiro Sase, Takaaki Okamoto, and Yoshimitsu Tuboi (Nikon, Japan); and Ivan Rozkosˇny´ (Nikon, Prague) kindly lent their microscopes. RP was supported by Ministry of Education projects “Chiral Microscopy” (LTC17012) and “ChemBioDrug”,10 KK by System Development Program for Advanced Measuring Technology (SENTAN) of Japan Science & Technology Agency (JST), and ZH by The Stentor Trust.

Glossary11 Amplitude objects Objects that attenuate light intensity (e.g., opaque granules or stained cells or tissue sections). To visualize them, phase contrast or other ways of ►optical contrasting (staining) is unnecessary. However, such objects typically also induce phase shifts (phase-amplitude objects); cf. ►phase objects. Anoptral contrast A variant of negative phase contrast (Reichert) that employed phase plates made of soot featuring smaller reflections than metallic coatings common in other microscopes. Apodized phase contrast Phase contrast enhanced by apodization. Neutral density (i.e., non–phase-shifting) filters of annular shape are placed immediately next to the phase annulus. Either two or multiple (graded-transmittance) apodization annuli may be used; only the former option is available on a commercial basis. Apodizing filters are sometimes referred to as modulation filters, the latter being a more general term. A-type/B-type phase plate Obsolete terms (coined by American Optical [Spencer] Co) meaning that a light-attenuating (metallic) layer is coated onto the ►conjugate/complementary area of the ►phase plate. B-minus phase plate An obsolete term for a ►phase plate yielding ►positive phase contrast (also referred to as ►dark contrast), in

10

CZ.02.1.01/0.0/0.0/16-019/0000729 Terminology in phase-contrast microscopy has not been entirely consistent over time. An additional complexity arises when various imaging modes are combined (e.g., relief phase contrast), and a specific terminology is used in electron microscopy.

11

Phase-Contrast Microscopy

317

which neither the ►conjugate nor complementary area is lightattenuating. Cf. ► Zernike phase contrast. Conjugate/complementary area Parts of a ►phase plate. The conjugate area (passing mainly direct or undiffracted light) is optically conjugate with the annular opening in the condenser diaphragm. The complementary area (passing diffracted light only) is the rest of the ►phase plate. Conventional phase contrast Non-apodized, non-relief phase contrast in light microscopy. In electron microscopy, phase-contrast observation is relatively new, thus the term “conventional” is irrelevant. Dark/bright contrast (or “dark/bright phase”) Older terms for ►positive/negative phase contrast, originally coined by American Optical (Spencer) Co and nowadays used by Nikon. The dark/ bright contrast terminology is consistent with sufficiently thin phase-retarding objects such as smaller cells in water. For example, dark-low-low (DLL), dark-low (DL), or dark-medium (DM) denotes a positive phase-contrast objective with a phase annulus of very low, low, or medium attenuation (transmittance ca. 45%, 25%, or 14%, respectively), rendering thin enough phaseretarding objects as less or more dark on bright background. The DL and DM objectives are also available in ►apodized versions (ADL and ADM). An ADH (apodized–dark–high) objective features transmittance of only ca. 6% and the highest phasevisualization sensitivity. Bright-medium (BM) denotes a negative phase-contrast objective with a medium-attenuation phase annulus (transmittance ca. 14%). Objectives currently made by Zeiss are mostly of positive type (marked “Ph1,” “Ph2,” etc.), and only rarely negative (“Ph2-”). A lens with two phase annuli (positive and negative) is also available (“Ph1 Ph2-”). In Olympus’s long-barrel (LB) series objectives (no longer commercially available), the following designations were used: PL/PLL (positive low/low-low) and NM/NH (negative medium/high). Objectives that used to be made by Leitz (“Pv” series) were either of “n” or “h” (positive), or “-h” (negative) type, with transmittance approximately equivalent to that of DL or DM, or BM lenses, respectively. Cf. ►anoptral contrast. Defocus phase contrast A term sometimes used in electron microscopy. Its meaning is the same as what is called “imitation phase contrast” in the present chapter, as achieved by slightly defocusing a bright-field image. Hilbert differential contrast A term sometimes used to denote a phase-contrast mode in electron microscopy, employing a π-type (λ/2 or 180 phase shift), asymmetric phase plate, suitable for relatively thicker objects (e.g., ultrathin sections or whole ice-embedded bacterial cells). Earlier, it was sometimes referred to as “Difference contrast TEM” (DTEM). Images are rendered in a

318

Radek Pelc et al.

quasi-3D appearance, similar to the ►relief phase contrast ones in light microscopy. Hoffman modulation contrast One of the so-called ►schlieren imaging modes. It employs an asymmetric slit diaphragm in the condenser and a graded-transmittance filter (“modulator”) in the objective. “Nikon Advanced Modulation Contrast” (NAMC) and Olympus’s “Relief Contrast” are its near-identical, double-slit variations; cf. ►relief contrast. Leica’s “Integrated (or Intermediate) Modulation Contrast” (IMC) allows for the modulator to be easily removed from the optical path as it is placed outside the objective (an ►optical-relay setup). A more simple variant, “Emboss Contrast” has been introduced by Nikon. Optical contrasting (staining) Visualization of refractive index differences within a specimen. The relevant imaging modalities include dark field (including ►Rheinberg illumination), ►Hoffman modulation contrast (and any other ►schlieren imaging mode, e.g., ►relief contrast), phase contrast, interference contrast, differential interference contrast (DIC) after Smith/Nomarski, and polarization microscopy. A narrower definition of “optical staining” only refers to the use of color filters in the optical path of the microscope, which however does not necessarily enhance the contrast in the images. Optical path difference (OPD, proportional to object-to-medium phase shift) Physical thickness (d ) of the object (e.g., a cell) multiplied by a refractive index difference (n1  n2) between the medium surrounding it (e.g., a physiological solution) and the object itself. For example, OPD of λ/6 translates to an object-to-medium phase shift of 60 . Optical thickness (or optical path length) is defined as n2d in non-absorbing objects. Often used (incorrectly) instead of OPD. Optical-relay setup See ►pupil-projection (optical-relay) setup. Phase objects Objects that shift only the phase of light, without attenuating its amplitude (e.g., unstained cells or tissue sections). Under a properly adjusted bright-field illumination, such objects are almost invisible. Chemical staining (as in classical histology or fluorescence microscopy) may be used to visualize them. An appropriate optical setup enabling ►optical contrasting (staining) represents a less invasive alternative. Phase plate (also referred to as “diffraction plate”) A glass plate at the objective back focal (transform) plane, with a thin coating of dielectric (i.e., phase-shifting) material either at the ►conjugate or complementary area. As the transform plane is often located between tightly packed lenses compounding the objective the coating is deposited directly on one of the lenses in that case. Cf. ►Atype/B-type phase plate.

Phase-Contrast Microscopy

319

Positive/negative phase contrast Imaging modes rendering thin enough phase-retarding (phase-advancing) objects as dark/bright (bright/dark), relative to the background. The phase annulus advances/retards the direct light, typically by 90 (λ/4). At a phase shift of 180 (λ/2) in the phase annulus, the two terms lose their meaning as positive and negative phase contrast yield identical images. Cf. ►anoptral contrast and ►dark/bright contrast. Pupil-projection (optical-relay) setup A system in which the objective back focal plane (approx. the exit pupil in low-power objectives) is optically relayed (duplicated) to a more accessible location. A ►phase plate may then be placed in an external, detachable unit rather than in the objective lens itself. This is of advantage when switching, for example, between total internal reflection fluorescence (TIRF) and phase-contrast modes as the need to exchange the objective lens is eliminated. Such systems are offered by Leica (“Integrated [or Intermediate] Phase Contrast,” IPH) and Nikon, and used to be available on Zeiss microscopes. Relief contrast One of the so-called ►schlieren imaging modes rendering ►phase objects in a relief-like (quasi-3D) appearance. It employs an asymmetric cut-off diaphragm in the condenser (offaxis illumination). A refined form of relief contrast is referred to as ►Hoffman modulation contrast. Relief phase contrast A combination of ►relief contrast and phase contrast, available from Zeiss as “variable relief” (VAREL) contrast. Not to be confused with ►variable phase contrast. Rheinberg illumination A refined form of dark-field illumination, in that the background is not dark but colored. Objects illuminated by another (contrasting) color are well visible. Schlieren imaging A collective term for ►optical contrasting (staining) modes employing one or two asymmetrical (relief, cutoff) diaphragms in the optical path (e.g., ►Hoffman modulation contrast and ►relief contrast). “Schlieren” (from German, meaning “streaks”) denotes shading patterns in the images which thus attain a relief-like (quasi-3D) appearance. Variable phase contrast A system enabling a gradual adjustment of the phase shift and transmittance, with a set of optical elements that emulate the ►phase plate. The most popular device of this kind is called “Polanret” by American Optical (Spencer) Co. This and other similar devices are no longer commercially available. An equivalent in electron microscopy is represented by a system employing an electrostatic (tunable) phase plate. Cf. ►relief phase contrast. Zernike phase contrast A term sometimes used to denote a phasecontrast mode in electron microscopy, employing a π/2-type (λ/4 or 90 phase shift), axially symmetrical phase plate with a tiny central opening for unscattered electrons (an equivalent of direct

320

Radek Pelc et al.

or undiffracted light in light microscopy). It is particularly suitable for very thin objects (e.g., protein particles). An equivalent of ►B-minus phase plate in light microscopy. References 1. Zernike F (1955) How I discovered phase contrast. Science 121(3141):345–349. https:// doi.org/10.1126/science.121.3141.345. Transcript of Nobel Lecture (https://www. nobelprize.org/prizes/physics/1953) 2. Needham GH (1958) The practical use of the microscope. Charles C. Thomas, Springfield, IL. https://lccn.loc.gov/57010440 3. Inoue´ S, Spring KR (1997) Video microscopy: the fundamentals. Plenum Press (Springer), New York. ISBN: 9780306455315 4. Ross KFA (1967) Phase contrast and interference microscopy for cell biologists. Edward Arnold, London. ISBN: 0713121335 5. Osterberg H (1955) Phase and interference microscopy. In: Oster G, Pollister AW (eds) Physical techniques in biological research, vol. I (Optical techniques). Academic Press, New York, pp 377–437. https://lccn.loc.gov/ 54011056 6. Bennett AH, Jupnik H, Osterberg H, Richards OW (1951) Phase microscopy: principles and applications. Wiley, New York. ISBN: 9780343275006 (2018 reedition by Franklin Classics). https://doi.org/10.5962/bhl.title. 5881 7. Pluta M (1975) Non-standard methods of phase contrast microscopy. In: Barer R, Coslett VE (eds) Advances in optical and electron microscopy, vol 6. Academic, London, pp 49–133. ISBN: 0120299062 8. Pluta M (1989) Phase contrast microscopy. In: Advanced light microscopy, Specialized methods, vol 2. PWN (Polish Scientific Publishers), Warsaw, pp 1–90. pp. 385–388. ISBN: 9780444989185 9. Zernike F (1934) Diffraction theory of the knife-edge test and its improved form, the phase-contrast method. Month Notic Roy Astron Soc 94(5):377–384. https://doi.org/ 10.1093/mnras/94.5.377 10. Burch CR (1934) On the phase-contrast test of F. Zernike. Month Notic Roy Astron Soc 94 (5):384–399. https://doi.org/10.1093/ mnras/94.5.384 11. Zernike F (1942) Phase contrast, a new method for the microscopic observation of transparent objects (Parts I and II). Physica (Amsterdam) 9(7):686–698. https://doi.org/ 10.1016/S0031-8914(42)80035-X, and 9

(10):974–986. https://doi.org/10.1016/ S0031-8914(42)80079-8 12. Tolansky S (1967) Frits Zernike 1888-1966. Biogr Mem Fellows Roy Soc 13:392–402. https://doi.org/10.1098/rsbm.1967.0021 13. Michel K (1943/1944/1958) Die Reifeteilungen (Meiose) bei der Spermatogenese der Schnarrheuschrecke (Psophus stridulus L.). Hochschulfilm C443/1944, Reichsanstalt fu¨r Film und Bild in Wissenschaft und Unterricht (RWU), Berlin 1944 (Wissenschaftlicher Film C443/1944, Institut fu¨r den Wissenschaftlichen Film, Go¨ttingen 1958). https://doi. org/10.7295/W9CIL39225. Video still images published in Zeiss Nachrichten 4 (9):236–251 (1943). http://www.archive. zeiss.de/objekt_start.fau?prj=zeiss&dm= Druckschriften&ref=86486 14. Hostounsky´ Z, Pelc R (2007) Phase-contrast versus off-axis illumination: is a more complex microscope always more powerful? Adv Physiol Educ 31(2):232–235. https://doi.org/10. 1152/advan.00028.2006 15. Hostounsky´ Z, Pelc R (2006) An efficient way of high-contrast, quasi-3D cellular imaging: off-axis illumination. J Biochem Biophys Meth 68(1):23–30. https://doi.org/10. 1016/j.jbbm.2006.03.016 16. Pelc R, Hostounsky´ Z, Otaki T (2008) Correlation between off-axis illumination and apodized phase-contrast: two complementary microscopic phase-imaging modes. J Biomed Opt 13(5):054067. https://doi.org/10. 1117/1.2966716 17. Pelc R, Hostounsky´ Z, Otaki T (2008) Reliefand apodized-phase-contrast imaging of biological specimens. 9th Asia-Pacific Microscopy Conference (APMC9), Jeju Island, Korea, 2-7 Nov 2008. Kor J Microsc 38(4Suppl):1073–1074 (on CD only) 18. To¨pler A (1866) Ueber die Methode der Schlierenbeobachtung als mikroskopisches Hu¨lfsmittel, nebst Bemerkungen zur Theorie der schiefen Beleuchtung. (Poggendorff’s) Annalen der Physik und Chemie (Leipzig) 127(4):556–580. https://doi.org/10.1002/ andp.18662030405 [Annalen der Physik (Berlin) 203(4):556–580] 19. Abbe E (1873) Ueber einen neuen Beleuchtungsapparat am Mikroskop. (Schultze’s) Archiv fu¨r Mikroskopische Anatomie (Bonn)

Phase-Contrast Microscopy 9:469–480. https://doi.org/10.1007/ BF02956177. BHL:13290387. https://bio diversitylibrary.org/page/13290387. English version: Ref. [20] 20. Abbe E (1875) A new illuminating apparatus for the microscope. Month Microsc J 13:77–82. https://biodiversitylibrary.org/ page/14276730 21. Abbe E (1873) Beitr€age zur Theorie des Mikroskops und der mikroskopischen Wahrnehmung. (Schultze’s) Archiv fu¨r Mikroskopische Anatomie (Bonn) 9:413–468. https://doi. org/10.1007/BF02956173 BHL:13290331. https://biodiversitylibrary.org/page/ 13290331. English version: Ref. 22 22. Abbe E (1874/1875) A contribution to the theory of the microscope, and the nature of microscopic vision. Proceedings of the Bristol Naturalists’ Society (new series) 1 (Pt2):200–261 (translation and comments by H. E. Fripp). https://biodiversitylibrary.org/ page/7506100 23. Stephenson JW (1877) Observations on professor Abbe’s experiments illustrating his theory of microscopic vision. Month Microsc J 17:82–88. BHL:14157879. https://bio diversitylibrary.org/page/14157879 24. Nuttall B (2008) Image formation in the microscope: the Abbe diffraction apparatus. Bull Sci Instrum Soc 96:27–29 25. Rheinberg J (1905) On the collected papers of Abbe, and microscope theory in Germany. J Quekett Microsc Club. (series 2) 9 (56):153–166. https://biodiversitylibrary. org/page/2200471 26. Michel K (1964) Die Grundzu¨ge der Theorie des Mikroskops in elementarer Darstellung (Physik und Technik, vol. 1, 2nd ed). Wissenschaftliche Verlagsgesellschaft, Stuttgart [3rd edition (1981) as ISBN: 3804706193] 27. Hoffman R (1977) The modulation contrast microscope: principles and performance. J Microsc (Oxf) 110(Pt3):205–222. https:// doi.org/10.1111/j.1365-2818.1977. tb00033.x 28. Ko¨hler A, Loos W (1941) Das Phasenkontrastverfahren und seine Anwendungen in der Mikroskopie. Naturwissenschaften 29(4):49–61. https://doi.org/10.1007/BF01476460. English version: Ref. 29 29. Ko¨hler A, Loos W (1947) The phase-contrast method and its application in microscopy. Textile Res J 17(2):82–95. https://doi.org/10. 1177/004051754701700203 30. Curl CL, Harris T, Harris PJ, Allman BE, Bellair CJ, Steward AG, Delbridge LMD (2004) Quantitative phase microscopy: a new tool for

321

measurement of cell culture growth and confluency in situ. Pflu¨gers Arch (Eur J Physiol) 448(4):462–468. https://doi.org/10.1007/ s00424-004-1248-7 31. Otaki T (2000 and 2001) Artifact halo reduction in phase contrast microscopy using apodization. Opt Rev 7(2):119–122 and 8 (4):284–286. https://doi.org/10.1007/ s10043-000-0119-5, https://doi.org/10. 1007/s10043-001-0284-1 32. Ellis GW (1988) Scanned aperture light microscopy. In: 46th Annual Meeting of the Electron Microscopy Society of America, Milwaukee (WI, USA), 7–12 Aug 1988. Proceedings (Bailey GW, ed), pp 48–49 (San Francisco Press, San Francisco). Brief version published in: Cell Motil Cytoskel 10(1-2):342. https:// doi.org/10.1002/cm.970100138 33. Bratuscheck K (1892) Die Lichtst€arke-Aenderungen nach verschiedenen Schwingungsrichtungen in Linsensystemen von grossem Oeffnungswinkel mit Beziehung zur mikroskopischen Abbildung. Zeitschrift fu¨r Wissenschaftliche Mikroskopie und fu¨r Mikroskopische Technik 9(2):145–160. https://biodiversitylibrary.org/page/ 4625575 34. Hohman B, Keller E (2001) VAREL: a new contrasting method for microscopy. Biomedical Products (renamed to Bioscience Technology), 1st January issue 35. Beyer H (1953). Untersuchungen u¨ber den Einfluß der Gestalt der Aperturblende auf die mikroskopische Abbildung beim Phasenkontrastverfahren. In: Jenaer Jahrbuch 1953, pp 162–209. VEB Carl Zeiss Jena and VEB Gustav Fisher, Jena 36. Peshkov ML (1955) A new type of objectives for anoptral microscopy, utilizing an opening, and a brief analysis of the principle of their operation. Biol Bull Rev (Moscow) 40 (3):372–378 (Uspekhi Sovremennoi Biologii, article in Russian) 37. Piper J (2010) Condenser aperture reduction phase contrast: a new technique for improved imaging in light microscopy. J Adv Microsc Res 5(1):1–10. https://doi.org/10.1166/jamr. 2010.1018 38. Soran V, Diaconeasa B (1957) The ultraanoptral microscope: an improvement of Wilska’s method. Naturwissenschaften 44 (17):465. https://doi.org/10.1007/ BF00640884 39. Burch CR, Stock JPP (1942) Phase-contrast microscopy. J Sci Instrum 19(5):71–75. https://doi.org/10.1088/0950-7671/19/ 5/302

322

Radek Pelc et al.

40. Rheinberg J (1904 and 1905) On the influence on images of gratings of phase differences amongst their spectra. J Roy Microsc Soc (volumes not numbered) Pt4:388–390 and Pt2:152–155. https://biodiversitylibrary.org/ page/2178964, https://biodiversitylibrary. org/page/2174836 41. Conrady A (1905) An experimental proof of phase reversal in diffraction spectra. J Roy Microsc Soc Pt2:150–152. https://bio diversitylibrary.org/page/2174834 42. Katoh K, Otaki T, Suzuki M (2004) Apodized phase contrast microscopy. Seibutsu Butsuri (“Biophysics”) 44(6):260–264. https://doi. org/10.2142/biophys.44.260 43. Nakamura K (1969) Optical microscope (in Japanese). In: Biophysical Society of Japan (ed) Saibou seibutsu butsuri kenkyu-ho (“Research methods in cell biophysics”). Yoshioka Shoten (Maruzen Press), Kyoto, pp 1–130 44. Pluta M (1989) Simplified polanret system for microscopy. Appl Optics 28(8):1453–1466. https://doi.org/10.1364/AO.28.001453 45. Bretschneider F, Teunis PFM (1994) Reducedcarrier single-sideband microscopy: a powerful method for the observation of transparent microscopic objects. J Microsc (Oxf) 175 (Pt2):121–134. https://doi.org/10.1111/j. 1365-2818.1994.tb03475.x 46. Stankewitz H-W (1998) Process and device for contrasting objects to be microscopically examined. USA patent US5708526. https:// patents.google.com/patent/US5708526 47. Takaoka H (1999 and 2000) Optical microscope which has optical modulation elements. USA patents US5969853 (https://patents. google.com/patent/US5969853) and US6130776 (https://patents.google.com/pat ent/US6130776) 48. Brief J (2005) Minimal dezentriert Phasenkontrast—Eine sehr einfache Methode zur Kontrastverbesserung bei hoher Vergro¨ßerung. Mikrokosmos 94(4):203–205. www.zobodat. at/pdf/Mikrokosmos_94_4_0001.pdf 49. Piper J (2007) Relief phase contrast: a new technique for phase-contrast light microscopy. Microsc Analysis 21(4):9–12. https://micros copy-analysis.com/magazine (issue 4 ¼ July) 50. Humphrey GK, Goodale MA, Bowen CV, Gati JS, Vilis T, Rutt BK, Menon RS (1997) Differences in perceived shape from shading correlate with activity in early visual areas. Curr Biol 7 (2):144–147. https://doi.org/10.1016/ S0960-9822(06)00058-3

51. Sun J, Perona P (1998) Where is the sun? Nat Neurosci 1(3):183–184. https://doi.org/10. 1038/630 52. Otaki T (2001) Phase object observation device. Japan patent application JP2001194592A (https://patents.google. com/patent/JP2001194592A) 53. Otaki T (2001) Phase contrast observation device. USA patent US6317261B1 (https:// patents.google.com/patent/US6317261B1) 54. Katoh K, Otaki T (2006) Imaging living samples with apodized phase contrast microscopy, a novel phase contrast microscopy. In 16th Intnl Microsc Congr (IMC16), Sapporo, Japan, 3–8 Sept 2006. Proceedings, p 64 55. Xu R (2002) Particle characterization: light scattering methods. Kluwer, New York. ISBN: 1402003579 56. Maurer C, Jesacher A, Bernet S, Ritsch-Marte M (2008) Phase contrast microscopy with full numerical aperture illumination. Opt Express 16(24):19821–19829. https://doi.org/10. 1364/OE.16.019821 57. Clark SG, Chiu C (2003) C. elegans ZAG-1, a Zn-finger-homeodomain protein, regulates axonal development and neuronal differentiation. Development 130(16):3781–3794. https://doi.org/10.1242/dev.00571 58. Lai C-C, Hong K, Kinnell M, Chalfie M, Driscoll M (1996) Sequence and transmembrane topology of MEC-4, an ion channel subunit required for mechanotransduction in Caenorhabditis elegans. J Cell Biol 133 (5):1071–1081. https://doi.org/10.1083/ jcb.133.5.1071 59. Danev R, Nagayama K (2006) Applicability of thin film phase plates in biological electron microscopy. Biophysics 2:35–43. https://doi. org/10.2142/biophysics.2.35 60. Nagayama K (2008) Development of phase plates for electron microscopes and their biological application. Eur Biophys J 37 (4):345–358. https://doi.org/10.1007/ s00249-008-0264-5 61. Danev R, Okawara H, Usuda N, Kametani K, Nagayama K (2002) A novel phase-contrast transmission electron microscopy producing high-contrast topographic images of weak objects. J Biol Phys 28(4):627–635. https:// doi.org/10.1023/A:1021234621466 62. Danev R, Nagayama K (2008) Single particle analysis based on Zernike phase contrast transmission electron microscopy. J Struct Biol 161 (2):211–218. https://doi.org/10.1016/j.jsb. 2007.10.015 63. Danev R, Kanamaru S, Marko M, Nagayama K (2010) Zernike phase contrast cryo-electron

Phase-Contrast Microscopy tomography. J Struct Biol 171(2):174–181. https://doi.org/10.1016/j.jsb.2010.03.013 64. Hopkins HH (1953) A note on the theory of phase-contrast images. Proc Phys Soc B 66 (4):331–333. https://doi.org/10.1088/ 0370-1301/66/4/411 65. Beleggia M (2008) A formula for the image intensity of phase objects in Zernike mode. Ultramicroscopy 108(9):953–958. https:// doi.org/10.1016/j.ultramic.2008.03.003 66. Danev R, Glaeser MG, Nagayama K (2009) Practical factors affecting the performance of a thin-film phase plate for transmission electron microscopy. Ultramicroscopy 109 (4):312–325. https://doi.org/10.1016/j. ultramic.2008.12.006 67. Westneat MW, Socha JJ, Lee W-K (2008) Advances in biological structure, function, and physiology using synchrotron X-ray imaging. Annu Rev Physiol 70:119–142. https:// doi.org/10.1146/annurev.physiol.70. 113006.100434

323

68. Fitzgerald R (2000) Phase-sensitive X-ray imaging. Physics Today 53(7):23–26. 53 (11):82. https://doi.org/10.1063/1. 1292471, https://doi.org/10.1063/1. 2405518 69. Beckmann F, Heise K, Ko¨lsch B, Bonse U, Rajewsky MF, Bartscher M, Biermann T (1999) Three-dimensional imaging of nerve tissue by X-ray phase-contrast microtomography. Biophys J 76(1):98–102. https://doi. org/10.1016/S0006-3495(99)77181-X 70. Betz O, Wegst U, Weide D, Heethoff M, Helfen L, Lee W-K, Cloetens P (2007) Imaging applications of synchrotron X-ray phasecontrast microtomography in biological morphology and biomaterials science. I. General aspects of the technique and its advantages in the analysis of millimetre-sized arthropod structure. J Microsc (Oxf) 227(Pt1):51–71. https://doi.org/10.1111/j.1365-2818.2007. 01785.x

Note Added in Proof12 71. Kubbinga H (2012–2016) Collected papers of Frits Zernike (1888–1966). vol I–IV. Grøningen University Press, Grøningen. ISBN: 9789081442831. https://openlibrary.org/ books/OL25241270M 72. Brice AT (1953) On the Zernike phase-contrast method of microscopy. Amer Biol Teacher 15(5):124–128. https://doi.org/10.2307/ 4438501 73. Jacquez JA, Biesele JJ (1954) A study of Michel’s film on meiosis in Psophus stridulus

12

L. Exp Cell Res 6(1):17–29. https://doi.org/ 10.1016/0014-4827(54)90144-5 74. Melly MA, Thomison JB, Rogers DE (1960) Fate of staphylococci within human leukocytes. J Exp Med 112(6):1121–1130. https://doi. org/10.1084/jem.112.6.1121 75. Rogers DE (1950s) Neutrophil chasing bacteria (video). https://embryology.med.unsw. edu.au/embryology/index.php/Movie_-_ Neutrophil_chasing_bacteria (comments by Thomas S. Stossel and Mark A. Hill)

A compilation of all Zernike’s papers is now available [71], including translations to English. A historical account of early time-lapse phase-contrast microscopy [13] may be found in Refs. [72, 73]. Phase-contrast recording of phagocytosis [74] may now be viewed as a video [75].

Chapter 11 Ultramicroscopy of Nerve Fibers and Neurons: Fine-Tuning the Light Sheets Saiedeh Saghafi, Klaus Becker, Nina J€ahrling, Christian Hahn, and Hans-Ulrich Dodt Abstract Light sheet generation approaches employed in the ultramicroscopy imaging technique are presented, and three different designs are discussed. These include slit-based as well as more advanced slit-free setups, including those based on aspheric optics, that is, inherently aberration-free. The performance of the light sheets thus generated is briefly described theoretically, and experimentally compared. Images of a chemically cleared fruit fly (Drosophila melanogaster), nerve fibers in an entire mouse embryo, and individual neurons in intact mouse hippocampi are presented. Key words Aspheric optics, Gaussian beam, Flattened Gaussian beam, Light-sheet microscopy, Ultramicroscopy

1

Introduction Virtual (i.e., optical/tomographical) sectioning of samples represents a noninvasive alternative to physical sectioning. Methods such as optical projection tomography (OPT), optical coherence tomography (OCT) and magnetic resonance imaging (MRI) are examples of such an approach [1]. While their penetration depth is greater, resolution is poorer than in microscopy [2]. Confocal and two-photon microscopy are well-established, high-resolution, point-scanning options capable to optically section uncleared (i.e., typically strongly scattering) specimens up to ~100 μm and ~1 mm deep, respectively [3]. Penetration depth of 3 mm has been achieved with confocal microscopy (i.e., singlephoton excitation) in a 10-days old (quite transparent) mouse embryo, with a recently developed setup referred to as “mesolens” [4]. Rather than point-by-point, ultramicroscopy (frequently referred to as light-sheet microscopy) performs the optical

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_11, © Springer Science+Business Media LLC 2020

325

326

Saiedeh Saghafi et al.

sectioning with a light sheet, and is thus considerably faster than confocal or two-photon microscopy, at the expense of spatial resolution. Providing the specimens are sufficiently transparent (i.e., featuring long enough transport mean free path [1]) even centimeter-sized objects can be analyzed. All of these approaches enable volumetric measurements and 3D visualization of specimens while relying on a computer reconstruction of a stack of images. Siedentopf and Zsigmondy are credited with pioneering the field of light-sheet microscopy back in 1903 [5, 6]. Richard Adolf Zsigmondy was a physical chemist active in the field of colloid chemistry, trying to visualize small colloidal particles (Ultramicronen) while Henry Siedentopf was an optical physicist at Carl Zeiss company. They illuminated colloidal solutions from the side (perpendicularly to the optical axis) as powerful enough dark-field condensers were unavailable at that time. From a construction (hardware) point of view, ultramicroscopy is essentially a variant of dark-field or fluorescence imaging, depending whether or not a fluorophore is used. A bulky carbon arc burner was employed as a light source. Using lenses, aperture stops (slits) and microscope objectives they generated a simple light sheet to illuminate a welldefined volume of minimal thickness. This type of an ultramicroscope was referred to as “slit ultramicroscope.” Although it was never widely used except in colloidal chemistry, Zsigmondy was awarded in 1925 a Nobel Prize in Chemistry for this work. Subsequent microscope development focused on making more efficient dark-field condensers such as the cardioid condenser [7, 8]. The Cassegrain condenser (akin to the Cassegrain telescope in design) or the so-called (luminous) spot-ring condenser had numerical aperture as high as 1.50 [8]. The cardioid condenser (named after the cardioid curve), also invented by Siedentopf, could be supplemented by an azimuthal (Szegvari’s) diaphragm suitable for highlighting elongate particles oriented in a particular direction [8, 9]. The history of dark-field microscopy dates back to 1830s [8, 10]. Jentzsch’s “ultrakondensor” (originally made by Ernst Leitz, Wetzlar, and later also by C. Baker, London) dating back to 1910 was laid upon the stage of an upright microscope. It was built around a flow-cell so that the liquid or gas sample could be illuminated more-or-less perpendicularly to the optical axis [10, 11], much like in the “slit ultramicroscope.” The term “ultracondenser” was applied to any condenser employed in ultramicroscopy, and the setup was referred to as “reflecting ultramicroscope.” Indeed, the condensers had to be double-reflecting (“bireflecting”), mirrorbased to achieve sufficiently high numerical aperture, that is, illumination near-perpendicular to the optical axis [8]. This concept has now been “reinvented” by Leica (the successor of Ernst Leitz) in the form of a mirror device (a DLS model) attached to the objective of an upright microscope, to obtain a “digital light sheet.”

Ultramicroscopy of Nerve Fibers and Neurons

327

In 1964, McLachlan described the use of light-sheet illumination in microscopy to visualize the surface of minerals and insect eyes [12]. A similar approach was employed in light-scanning photomicrography to visualize the surface of insects, with two- or three-sided illumination [7]. This setup, marketed as “Dynaphot” by Irvine Optical Corporation (Burbank, CA, USA) as a low-magnification imaging tool complementary to the scanning electron microscope. Other examples include macroscopic imaging of screws, coins and rifle bullets [7]. In 1993, Decre et al. reported the use of laser light sheets to investigate specific phenomena in fluid dynamics [13]. All of these setups employed light-sheet microscopy to visualize the surface of the objects. Voie pioneered the combination of light-sheet microscopy with fluorescence excitation to visualize a biological object [14]. A guinea pig cochlea made fluorescent by immersion into rhodamine was used as a study object. In order to reduce scattering and achieve better image quality, he made the cochlea transparent using a special clearing solution developed a century ago by Werner Spalteholz [15]. In 2002, Fuchs et al. reported the usage of a lightsheet microscope for microbial studies in oceanography [16]. Later on, research conducted by Stelzer’s group drew a significant attention to the field of light-sheet microscopy (since then often referred to as “selective plane illumination microscopy [SPIM]) by demonstrating that this technique is an excellent method for visualizing near-transparent transgenic fish embryos expressing green fluorescent protein (GFP) in muscles. They obtained images with semiisotropic resolution (i.e., approximately the same in all three directions) by combining light-sheet microscopy with acquisition of image stacks along different directions [17]. Thereafter, light-sheet microscopy (or ultramicroscopy) became popular among the biological fraternity. As most multicellular biological samples are not very transparent, it is essential to (optically) clear them to reduce light scattering primarily responsible for that. With rare exceptions, this is necessary if millimeter-sized (or greater) objects are to be imaged, and as clearing is typically achieved by chemical means, the process is referred to as “chemical clearing” [18]. The cleared (semitransparent) specimen is placed in a chamber filled with the clearing medium, illuminated and scanned by a very thin light sheet generated by a dedicated optical unit. The image quality in ultramicroscopy can be further enhanced by improving the optical parameters of the system, chiefly determined by its optical components. The present chapter describes three different light-sheet (ultramicroscopy) setups, and demonstrates their performance in imaging neural tissue. To pay tribute to the original inventors of the technique [5, 6], we adhere to its original name, that is, ultramicroscopy.

328

2

Saiedeh Saghafi et al.

Ultramicroscopy Setups To achieve good enough optical sectioning capability (i.e., resolution) of the ultramicroscope it is essential to use an ultrathin light sheet. The following sections describe how this can be achieved, in three different setups.

2.1

Slit-Based Setup

In a slit-based ultramicroscope, the illuminating beam is produced by a sapphire laser (500 mW, 488 nm) from Coherent (Dieburg, Germany). It is initially expanded by two lenses, the so-called Keplerian beam expander unit, and then guided toward a beamshaping unit housing a rectangular slit aperture and a cylindrical lens, as shown in Fig. 1. By truncating the expanded Gaussian beam, the aperture forms a semiuniform beam. The cylindrical lens then focuses it in one direction (X) only, that is, the beam remains practically unaltered in the other (perpendicular) direction (Y). This results in converting an axially symmetrical laser beam into a thin sheet of light (Fig. 1).

Z

Fluorescence light

Objective Cylindrical lens

X

Brain

Slit aperture

Fig. 1 Slit-based ultramicroscopy setup. The specimen (e.g., mouse brain) is submerged in a clearing medium and illuminated by two identical light sheets, each generated by a slit aperture and a cylindrical lens. Beam expanders (not shown) are placed before the slit apertures

Ultramicroscopy of Nerve Fibers and Neurons

SLIT-BASED SLIT-BASED 8mm × 10mm

2mm × 10mm

SLIT- FREE

SLIT- FREE

“UM-1”

“UM-2”

Z

A

329

Y X

B X

C Y STANDARD OPTICS

ASPHERIC OPTICS

Fig. 2 Theoretical (A, B) and measured (C) intensity distribution of light sheets produced by slit-based and alternative (slit-free) ultramicroscopy setups (normalized intensities in B, C are color-coded as in Fig. 3). The slit aperture (dimensions are shown) is parallel to the Y axis. Scale bars: X ¼ 150 μm, Y ¼ 500 μm (images expanded in X direction for clarity). Adapted from Refs. [20, 26] by permission of © Wiley Interscience

In such a system the optical properties of the light sheet are mostly determined by the shape and dimensions of the slit aperture [19]. A wide rectangular slit aperture produces a light sheet that is tightly compressed at the focus, but diverges rapidly away from it, that is, yielding a short Rayleigh range, and poor uniformity along the light sheet [20]. In contrast, a narrow aperture increases the Rayleigh range but also increases the thickness of the light sheet and of course reduces the overall light intensity (Fig. 2). The Rayleigh range is defined as the distance along the propagation axis (z) from the beam’s waist located at the focusing plane to the point where the beam’s cross-sectional area is approximately doubled. The uniformity of laser intensity along the focusing line (running perpendicularly to the beam propagation axis, i.e., in the Y direction), and the length of this line are of course important parameters in ultramicroscopy. The greater the uniformity of the light sheets the better its optical sectioning capability. Mathematically, the truncated Gaussian beam can be described with the aid of Heavyside step function and a finite sum of complex Gaussian functions [20–25], that is, as a generalized Huygens– Fresnel diffraction integral [20]. The intensity profiles of a symmetrical Gaussian beam (i.e., of the same width in Y and Y direction) upon passing through a rectangular slit aperture (either 8
10 mm or 2
10 mm), at near-field as well as far-field conditions, are shown in Fig. 3. The truncated Gaussian beam is turned into a light sheet by passing through a cylindrical lens. The light sheet is then scanned through the specimen, and the images (optical sections) thus obtained are combined into a 3D image which can be further

330

Saiedeh Saghafi et al.

Truncated beam 8mm x 10mm

Truncated beam 2mm x 10mm

Nomalized intensity

Gaussian beam (no truncation)

Fig. 3 Simulated intensity (I ¼ |E(x, y)|2) profiles of a normalized Gaussian beam and truncated Gaussian beams produced by two slit apertures (8 mm
10 mm and 2 mm
10 mm), at near-field and far-field (image plane). Only the far-field case is of practical relevance to ultramicroscopy; the side lobes are due to beam truncation (slits parallel to the Y axis, cf. Fig. 1) and adversely affect image quality. Adapted from Ref. [20] by permission of © Wiley Interscience

processed, much like in confocal microscopy. A slit-based ultramicroscopy setup is shown in Fig. 1. The optical characteristics of the light sheet, such as the uniformity of laser intensity along the line of focus and the length of this line are of critical importance in ultramicroscopy, and depend on the width of the slit aperture. A combination of a wide slit aperture (e.g., 8
10 mm) and a cylindrical lens of appropriate focal length (e.g., f ¼ 80 mm, lens from (Qioptiq, Munich, Germany) yields a very small beam spot size at the focal point. However, it quickly diverges, that is, has a short Rayleigh range. Furthermore, we are confronted with the fact that light sheet intensity is less uniform along the propagation axis (z). For small specimens, this type of light sheet can be used but for large (cm-sized) specimens such as entire mouse brain, it fails to provide images of sufficient quality. By reducing the width of the slit aperture, the Rayleigh range is extended, but the thickness of the light sheet inevitably increases [20]. A major loss of light intensity is another drawback of the slitbased setup which is nevertheless sufficient in many cases. Various designs employing lenses, wedges and prisms have been suggested to provide a suitable laser light distribution, without causing a major intensity loss [20]. Two of them are presented below. 2.2 Slit-Free Setup “UM-1”: Standard Optics

By using one spherical lens and three cylindrical lenses placed and oriented as shown in Fig. 4, the properties of the light sheet are markedly improved [20]. The light sheet is generated without truncating the beam, thus preserving light intensity. In addition, the beam uniformity along the line of focus is improved as no side

Ultramicroscopy of Nerve Fibers and Neurons

331

X

Z

Y

Fig. 4 Slit-free setup (UM-1) capable of generating the light sheet. One spherical (1) and three planoconvex cylindrical (2–4) lenses are employed. Orientation of the cylindrical lenses (0 or 90 ) is shown more clearly in the cutouts. Adapted from Ref. [20] by permission of © Wiley Interscience

lobes are generated at the focal point of the last lens; these would otherwise appear at the aperture edges. We used a specific type of Galilean beam expander that includes a planoconcave spherical lens (1) and a planoconvex cylindrical lens. (2) As the second lens is cylindrical, the beam expanded by the first lens converges in one direction (Y) while remaining unchanged in the perpendicular direction (X), that is, it adopts a semielliptical profile. The orientation of the second cylindrical lens (3) is perpendicular to the first cylindrical lens and controls the beam shape in the X direction while not affecting its convergence in the Y direction. The third planoconvex cylindrical lens (4) is oriented perpendicularly to the second cylindrical lens (3). The former converts the incident light with elliptical profile into a thin sheet with improved (longer) line of focus and Rayleigh range [20]. By adjusting the position of the first two cylindrical lenses, the length of the line of focus can be easily varied. Figures 5 and 6 (showing an entire mouse embryo and the eyes of Drosophila, respectively) demonstrate a marked improvement in resolution and overall image quality that can be achieved with the UM-1 setup, compared to the slit-based one. The threedimensional reconstructions (yielding the virtual optical sections shown in image plates) were performed on stacks of images, each of them acquired at a different position of the light sheet. In the next section, we introduce another optical system for light sheet generation that employs complex aspheric optical elements that reshape the initial beam into an ultrathin light sheet. This design not only preserves light intensity but also significantly improves resolution and other optical characteristics.

332

Saiedeh Saghafi et al.

SLIT-BASED SETUP

SAGITTAL PLANE

PARASAGITTAL PLANE

*

SLIT- FREE SETUP “UM-1”

1 mm

*

Fig. 5 Examples of image reconstructions obtained with the slit-based and slitfree ultramicroscopy setups. Virtual (optical) sections along two planes through a mouse embryo, with nerve fibers immunostained by anti-neurofilament 160 antibody conjugated with Alexa Fluor 488. Boxes mark the areas with the greatest improvement of image quality. (∗) Vibrissae innervation. Objective: Olympus XLFLUOR 4
/0.28 2.3 Slit-Free Setup “UM-2”: Aspheric Optics

One of the strongest motivations to use aspheric optics in any optical design is its ability to form a distortion-free image with minimum aberrations. Like the UM-1 setup described above, UM-2 is also slit-free, thus preserving light intensity. In this novel setup, we employed aspheric optical elements to generate a highly uniform, ultrathin light sheet while at the same time achieving the longest Rayleigh range, compared to both the slit-based and UM-1 setups. A detailed analytical description of its properties can be found elsewhere [26]. Importantly for ultramicroscopy, aspheric optics can efficiently reshape the laser beam. For example, two aspheric lenses facing each other can convert a beam of a Gaussian intensity profile into a semiuniform one. The Powell lens is another beam-shaping optical element. Being a specific form of an aspheric prism, it is defined by its fan angle and the refractive index of the lens, and is capable of reshaping a Gaussian beam into a bone-shaped one [26]. By placing

Ultramicroscopy of Nerve Fibers and Neurons

SLIT-BASED SETUP (standard optics)

333

SLIT- FREE SETUP “UM-1” (standard optics)

A

B X

Z

100 µm

Y

SLIT- FREE SETUP “UM-2” (aspheric optics)

C

D om

lo 100 µm

m

Fig. 6 Virtual (optical) sections through the facetted eye of fruit fly (Drosophila melanogaster), as reconstructed from a stack of images by Amira software. Note the markedly better rendering of ommatidia with the “UM-2” setup. (A, B) Reconstructions from 432 real optical sections. Adapted from Ref. [19] by permission of Wiley Interscience. (C) Interommatidian bristles. (D) Ommatidia (om), medulla (m) and lobula (lo). The near-horizontal stripes are due to optical sectioning by the light sheet in one particular direction (X axis perpendicular to the stripes, cf. Fig. 1). Objective (A–D): Olympus UPlanFL N 10
/0.30. Images C and D are adapted from Ref. [26] by permission of © Wiley Interscience

the Powell lens between two aspheric condenser lenses facing each other, an elliptical beam with a semiuniform intensity is formed from a Gaussian beam (Fig. 7). To utilize this elliptical beam for ultramicroscopy, it must be transformed into a thin light sheet with sufficient uniformity along the line of focus. Using two achromatic cylindrical lenses, we can generate a very thin sheet of light with optimized Rayleigh range and minimal aberrations. Figure 7 describes the measured intensity profile in the XY plane at focus (0 mm), and 4 mm down the optical axis (Z). The UM-2 setup includes a beam splitter directing the laser beam of a Gaussian intensity profile into two arms (50/50) guided toward two identical light-sheet generator units [26]. Each of them is placed at either side of a quartz chamber containing the chemically cleared specimen, and has five optical elements shown in Fig. 7. The first one (1) is an aspheric condenser lens focusing the

334

Saiedeh Saghafi et al. - 4 mm

cyl. lens 90°

X Y

cyl. lens 90°

0 mm

+ 4 mm

Fig. 7 Slit-free setup (UM-2) based on aspheric optics and capable of generating an ultrathin light sheet. (1, 3) Aspheric planoconvex condenser lenses (focal length f ). (2) Conic aspheric lens (e.g., the Powel lens). (4, 5) Achromatic planoconvex cylindrical lenses (focal length F and F0 , respectively). The beam-shaping effects of two planoconvex lenses alone (1+3) and the Powell lens alone (2) are shown in the bottom panels; note the near-homogeneous and bone-shaped intensity distribution, respectively. A similar bone-shaped distribution is obtained by a combination of lenses (1) and (2). Beam profiles in the XY plane are shown. The illuminating beam intensity is of a Gaussian profile (left), 4 and 0 mm denote a distance down the optical axis (Z) from the rear focal point of the entire setup (right). Adapted from Ref. [26] by permission of © Wiley Interscience. Cylindrical lenses (4 and 5): see a note in Fig. 4 legend

beam on the surface of the second element (2) (a conical aspheric lens, e.g., a Powell lens) having a fan angle smaller than 10 . The latter reshapes the Gaussian laser beam into a bone-shaped beam with a semiuniform intensity distribution at its edges. The third element (3) is another aspheric condenser lens placed at a distance of 2f from the Powell lens, where f is the focal length of the lens (identical to the first one). Combined, these three lenses convert the Gaussian beam into an elliptical beam with a semiuniform intensity distribution. The fourth element (4) is an achromatic cylindrical lens placed at a distance of F+F0 (where F and F0 is the focal length of the third and fourth lens, respectively). Finally, the fifth element (5) (achromatic cylindrical lens of focal length F0 ) pffiffiffi 0 placed at a distance of 2F from the fourth lens. Jointly, the fourth and fifth lenses generate a very thin light sheet, at the focal distance of the fifth lens. Both theoretical and experimental comparison of the light sheet in the slit-based, UM-1 and new (UM-2) setup shows a more homogeneous luminance distribution [26] and a longer Rayleigh range in the latest system inherently featuring minimum optical aberrations and maximum uniformity (Fig. 2). Providing

Ultramicroscopy of Nerve Fibers and Neurons

A

335

B

X

Z 200 µm

Y 100 µm

lo

C

100 µm

D

50 µm

Fig. 8 Pyramidal cells in an intact, chemically cleared mouse hippocampus expressing green fluorescent protein (GFP) under the control of thy-1 promoter, as visualized by a slit-free ultramicroscopy setup (“UM-2”) employing aspheric optics to generate the light sheet, and four different objectives (Olympus): (A) XLFLUOR 2
/0.14. (B) XLFLUOR 4
/0.28. (C) UPlanFL N 10
/0.30. (D) LUCPlanFL 20
/0.45. Image C is adapted from Ref. [26] by permission of © Wiley Interscience

the specimen is sufficiently transparent (optically cleared) the ultrathin light sheet, inherently causing minimum scatter from out-offocus planes, makes it possible to reliably inspect structures deep within the tissue. Figure 6 features both surface and inner structures of the head capsule of a fruit fly (Drosophila melanogaster); ommatidial lens facets, mechanosensory inter-ommatidial bristles, as well as inner structures such as medulla and lobula are clearly discernible. Image quality markedly improves from the slit-based to UM-1 to UM2 setups. The capability of the novel aspheric-optics setup (UM-2) is further demonstrated by visualizing pyramidal cells in the hippocampus (Fig. 8).

336

3

Saiedeh Saghafi et al.

Summary Three ultramicroscopy setups presented here yield 3D imagery of both outer and inner structures of chemically cleared biological samples. The novel aspheric-optics setup (UM-2), inherently free of optical aberrations, is clearly superior in terms of resolution and image rendering in general, which in authors’ opinion outweighs its higher cost and complexity. Its ultrathin and highly uniform light sheet is of particular importance if not-too-transparent specimens are to be examined, that is, when chemical clearing reaches its limitation, as it is often the case.

Appendix The complex light intensity in the X direction (i.e., perpendicularly to the slit) of an axially symmetrical Gaussian beam propagating in vacuum along the z-axis is described by Eq. 1 [20]:     w0 x2 x2 exp ikz  ik exp  2 þ iϕðz Þ E ðx, z Þ ¼ E 0 wðz Þ 2Rðz Þ w ðz Þ ð1Þ where E0 is the amplitude of the illuminating light w0 is the radius of beam waist at the focus (z ¼ 0) w(z) is the radius of the beam pffiffiffiffiffiffi ffi at distance z from the focus i is the complex unit, 1 k is the wave number (2π/λ, where λ is the wavelength of the illuminating light) R(z) is the radius of curvature of the beam at distance z φ (z) is the longitudinal phase delay at distance z "  2 # z Rðz Þ ¼ z 1 þ ð2Þ zR vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi u"  2 # u z w ðz Þ ¼ w 0 t 1 þ zR  ϕðz Þ ¼ arctan

z zR

ð3Þ

 ð4Þ

where zR is the Rayleigh range (Fig. 9), zR ¼

πw 20 : λ

ð5Þ

Ultramicroscopy of Nerve Fibers and Neurons

337

ZR

w0

2w0

Z

X

Fig. 9 Detail of a light beam at focus. Rayleigh range (zR) describes the beam expansion along the direction of propagation (Z), and is defined as the distance from the waist (diameter 2w0) to the point where the cross section of the beam is approximately doubled. For a Gaussian beam, this occurs at a beam diameter of pffiffiffi 2w 0 . Adapted from Ref. [20] by permission of © Wiley Interscience

At the beam waist (z ¼ 0), w(z) ¼ w0, R(z) ¼ R0, φ(z) ¼ φ0 and Eq. 1 becomes simplified as follows:     x2 x2 þ iϕ0 ð6Þ E ðx Þ ¼ E 0 exp  2 exp ik 2R0 w0 E(x) is the complex amplitude of light while its (measurable) intensity is   2x 2 ð7Þ I ðx Þ ¼ jEðxÞj2 ¼ E 20 exp  2 w0 This is a well-known Gaussian distribution. As the beam is axially symmetrical the same type of dependence is obtained for the Y direction, that is, the intensity can be generalized as follows:   2 x 2 þ y 2 2 2 ð8Þ I ðx, y Þ ¼ j E ðx, yÞj ¼ E 0 exp w20 However, in a slit-based system the Y direction (parallel to the slit) is of little relevance as the beam is truncated by the slit mainly in one direction only (X in our notation), and later also “flattened” (in the same direction) by the cylindrical lens. The effect on the Gaussian beam of the slit (width ¼ 2a in the X direction) is described by the Heavyside step function Ap(x) which equals 1 for a  x  a, and 0 elsewhere. A generalized Huygens–Fresnel diffraction integral then describes the truncated (modified Gaussian) beam [20–25]: Za E ðx Þ ¼ E 0 a



  ik  2 2 2 Ax 0  2xx 0 þ Dx dx 0 A p ðx 0 ÞE ðx 0 Þ exp  2B ð9Þ

where A–B and D are elements of the optical transfer matrix. Equation 9 is not dependent on C, the third element of the transfer

338

Saiedeh Saghafi et al.

matrix. In the Y direction, the beam is practically untruncated (assuming a relatively narrow slit), so the Heavyside step function is not applied. References 1. Ntziachristos V (2010) Going deeper than microscopy: the optical imaging frontier in biology. Nat Meth 7(6):603–614. https:// doi.org/10.1038/nmeth.1483 2. Santi PA, Johnson SB, Hillenbrand M, GrandPre PZ, Glass TJ, Leger JR (2009) Thin-sheet laser imaging microscopy for optical sectioning of thick tissues. Biotechniques 46(4):287–294. https://doi.org/10.2144/000113087 3. Helmchen F, Denk W (2005) Deep tissue two-photon microscopy. Nat Meth 2 (12):932–940. https://doi.org/10.1038/ nmeth818 4. Saini A (2012) New lens offers scientist a brighter outlook. Science 335 (6076):1562–1563. https://doi.org/10. 1126/science.335.6076.1562 5. Siedentopf H, Zsigmondy R (1903) Ueber Sichtbarmachung ultramikroskopischer Teilchen, mit besonderer Anwendung auf Goldrubingl€aser. (Drude’s) Annalen der Physik (4th ser) 10(1):1–39. https://doi.org/10. 1002/andp.19023150102 6. Zsigmondy R (1909/1914) Colloids and the ultramicroscope: a manual of colloid chemistry and ultramicroscopy John Wiley, New York. ISBN: 978-0548767832 (1st ed in two printings, translated by Alexander J). https:// openlibrary.org/works/OL1473918W/Col loids_and_the_ultramicroscope 7. Keller PJ, Dodt H-U (2012) Light sheet microscopy of living or cleared specimens. Curr Opin Neurobiol 22(1):1–6. https://doi. org/10.1016/j.conb.2011.08.003 8. Needhan GH (1958) The practical use of the microscope. Charles C. Thomas, Springfield, IL. https://lccn.loc.gov/57010440 9. Siedentopf H (1908) Die Sichtbarmachung von Kanten im mikroskopischen Bilde. Zeitschrift fu¨r wissenschaftliche Mikroskopie und mikroskopische Technik 25(4):424–431. https://doi.org/10.1007/BF01503978 (note by Steubing 1910). https://bio diversitylibrary.org/page/3756612 € 10. Jentzsch F (1910) Uber Dunkelfeldbeleuchtung; Der Ultrakondensor. Physikalische Zeitschrift 11(21/22):993-1001;1001–1002 (Proceedings of “82. Versammlung deutscher ¨ rtzte”. Ko¨nigsberg, Naturforscher und A 18–24 Sept 1910)

11. Chamot E´M (1915/1921) Elementary chemical microscopy (1st/2nd ed). John Wiley, New York, p. 71/122. Reprinted (2013) by Forgotten Books. ISBN: 9781330679081. https://www.forgottenbooks.com/en/ books/ElementaryChemicalMicroscopy_ 10091269 12. McLachlan D (1964) Extreme focal depth in microscopy. Appl Opt 3(9):1009–1013. https://doi.org/10.1364/AO.3.001009 13. Decre´ M, Buchlin JM (1994) An extra-thin light sheet technique used to investigate meniscus shapes by laser-induced fluorescence. Exp Fluids 16(5):339–341. https://doi.org/10. 1007/BF00195434 14. Voie AH, Burns DH, Spelman FA (1993) Orthogonal-plane fluorescence optical sectioning: three-dimensional imaging of macroscopic biological specimens. J Microsc 170 (3):229–236. https://doi.org/10.1111/j. 1365-2818.1993.tb03346.x € 15. Spalteholz W (1914) Uber das Durchsichtigmachen von menschlichen und tierischen Pr€aparaten. S. Hierzel, Leipzig. http://digital. zbmed.de/zbmed/content/titleinfo/555354 16. Fuchs E, Jaffe J, Long R, Azam F (2002) Thin laser light sheet microscope for microbial oceanography. Opt Exp 10(2):145–154. https://doi.org/10.1364/OE.10.000145 17. Huisken J, Swoger J, Del Bene F, Wittbrodt J, Stelzer EH (2004) Optical sectioning deep inside live embryos by selective plane illumination microscopy. Science 13 (5686):1007–1009. https://doi.org/10. 1126/science.1100035 18. Becker K, J€ahrling N, Saghafi S, Weiler R, Dodt H-U (2012) Chemical clearing and dehydration of GFP expressing mouse brains. PLoS One 7(3):1–6. https://doi.org/10.1371/jour nal.pone.0033916 19. Dodt H-U, Leischner U, Schierloh A, J€ahrling N, Mauch C-P, Deininger K, Deussing J-M, Eder M, Zieglg€ansberger W, Becker K (2007) Ultramicroscopy: three-dimensional visualization of neuronal networks in the whole mouse brain. Nat Methods 4 (4):331–336. https://doi.org/10.1038/ nmeth1036 20. Saghafi S, Becker K, J€ahrling N, Richter M, Kramer ER, Dodt H-U (2010) Image

Ultramicroscopy of Nerve Fibers and Neurons enhancement in ultramicroscopy by improved laser light sheets. J Biophotonics 3 (10–11):686–695. https://doi.org/10.1002/ jbio.201000047 21. Siegman AE (1986) Lasers. Oxford University Press, Mill Valley, CA. ISBN: 0198557132 22. Wen JJ, Breazeale MA (1988) A diffraction beam field expressed as the superposition of Gaussian beams. J Acoust Soc Am 83 (5):1752–1756. https://doi.org/10.1121/1. 396508 23. Zhao D, Zhang W, Wang S, Liu H, Zhu Q, Wei X (2003) Propagation of one dimensional off-axial Hermite-cosine-Gaussian beams

339

through an aperture paraxial ABCD optical system. Optik 114(49–51). https://doi.org/ 10.1364/JOSAA.26.002480 24. Ding D, Zhang Y (2004) Notes on the Gaussian beam expansion. J Acoust Soc Am 116 (3):1401–1405. https://doi.org/10.1121/1. 1781619 25. Longhurst RS (1968) Geometrical and physical optics, 2nd edn. Longmans, London 26. Saghafi S, Becker K, Hahn C, Dodt H-U (2014) 3D-ultramicroscopy utilizing aspheric optics. J Biophotonics 7(1–2):117–125. https://doi.org/10.1002/jbio.20130004

Chapter 12 Imaging and Electrophysiology of Individual Neurites Functionally Isolated in Microchannels Heinz D. Wanzenboeck, Petra Scholze, and Johann K. Mika Abstract The present chapter describes fabrication and utilization of neurite-isolation microelectrode arrays (NI-MEAs). The neurite-isolation device makes it possible to functionally isolate individual neurites from neuronal cell bodies as the neurites are forced to grow into narrow microchannels. Being microfluidic setups, the NI devices can be conveniently employed to carry out laser axotomy, study retrograde/ anterograde transport in single axons, investigate the mutual effects of soluble factors (e.g., amyloid-β peptides) secreted by different neuronal populations (each “housed” in a different macrochannel), or to create the so-called axonal diodes imposing unidirectional axonal connectivity, to name but a few options. A separately made MEA is assembled with the NI device so that each of its microchannels is connected to one microelectrode. Unlike in patch-clamp, electrophysiology performed with the aid of NI-MEA is “massively parallel” as many neurites can be recorded simultaneously, and the data acquisition is noninvasive as the electrodes are not forced into a physical contact with the neurites. The capability of such setup is demonstrated here by recording the effects of different concentrations of nerve growth factor (NGF) on the electrical activity of individual neurites of superior cervical ganglions in primary culture, during their outgrowth into the microchannels. Simultaneous microscopic (e.g., immunofluorescence) observation of multiple neurites and their somata is an inherent extra option permitting to correlate their electrical activity with other physiological parameters, recorded by optical means. Key words Neurite isolation device, Microfluidics, Multielectrode array, Superior cervical ganglion

Abbreviations AID Ig iPSC MEA NGF NI-MEA PDMS PBS SCG

Axon isolation device Immunoglobulin Induced pluripotent stem cells Micro/Multielectrode array Nerve growth factor Neurite-isolation MEA Polydimethylsiloxane Phosphate buffered saline Superior cervical ganglion

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_12, © Springer Science+Business Media LLC 2020

341

342

1

Heinz D. Wanzenboeck et al.

Introduction The versatility of microscopic imaging techniques—especially in combination with fluorescence immunolabeling—for the identification of cellular components of neuronal tissue is beyond doubt and has been demonstrated in many cases. However, understanding the physiology of nervous tissue is related to the electrical activity of neurons. In neuroscience the electrophysiological methods focus on the electrical measurement of action potentials. Initially recordings of large-scale electric signals such as electroencephalography provided fundamental insights into the brain function. With the introduction of the patch-clamp and voltage-clamp methods electrophysiology was extended to single-cell level and provided deeper insights into the signal transduction mechanisms. However, manual approaching of a specific neurite (i.e., axon or dendrite) with a diameter of only a few micrometers with several individually manipulated patch pipettes is a challenging task. As an alternative, systems based on microchip technology have been introduced that feature a high number of multiple electrodes for simultaneous recording of neuronal network activity. Most systems have electrodes with a diameter of several tens of micrometers which is about the size of neuron’s soma, but recently introduced microfabricated electrodes can be as small as 30 nm [1]. The present chapter focuses on the electrophysiology of neurites using microelectrode arrays (MEAs) with microelectrodes of 1 μm (diameter 50–1000 μm of patchpipette)

10–200 μm

0.1–20 μm

Typical electrode spacing

>10 μm

Variableb

>20 μm

>1 μm

Electrode placement

Manual contact to neuron

Growth of neurons

Growth of neurons

Growth of neurons

Microscopy setup

No restrictions

Restrictions applyc

Restrictions applyc

Incident illumination onlyd

a

Passive microelectrodes are conductive structures that transmit the signal to external electronics for amplification. Active microelectrodes are microelectrodes on a semiconductor chip that directly processes the recorded signal by amplification or filtering. Interdigitated electrodes consist of two parallel comb-like planar electrodes set several micrometers apart, in an interlaced manner b Typical electrode spacing ranges from the upper sub-micrometer range to several hundred micrometers. As the spacing influences the penetration depth of the electrical field into the medium the dimensions are adapted to each specific application c Transillumination (with inverted or upright microscope) may be used except when substrate and/or electrodes are not transparent (in which case incident illumination and upright microscope have to be used) d As the whole setup is not transparent, only upright microscope may be used

While three-dimensional electrode configurations enabling locally resolved measurement in three-dimensional tissue have also been developed the present study focuses on planar designs only. Depending on the size of the electrical contact pads the number of recording microelectrodes in commercial MEA systems is 60–256 microelectrodes, commonly arranged in regular arrays of 8  8 to 16  16 microelectrodes (Fig. 6). A special design combining microelectrodes with a microchannel setup is the so-called patch clamp on a chip approach. Instead of using a pipette (as with conventional patch clamping) a vacuum suction is applied through a microfabricated aperture of the patchclamp chip. In this aperture a microelectrode is placed to record the potential. As the vacuum suction through the microfabricated aperture sucks the cell membrane on to the microelectrode a high-resistance “gigaseal” is established between the microelectrode and the cell membrane [63] and membrane voltages can be measured on a microchip with very high precision [64–68].

Functional Isolation of Neurites in Microchannels

359

Fig. 6 Photograph (49  49 mm) and a schematic layout of a perforated microelectrode array 60pMEA100/ 30iR-Ti from Multi Channel Systems. The base material is a polyimide foil on a ceramic carrier

Fig. 7 SEM image of neurons growing on a complementary metal oxide semiconductor (CMOS)-based microelectrode array (MEA). Reprinted from Ref. [69] by permission of © IEEE

Another approach for high-resolution recording of action potentials is the use of “active” microelectrodes that have the amplifier (and further signal processing electronics) already implemented in the MEA substrate fabricated by a complementary metal oxide semiconductor (CMOS) technology. Instead of metal electrodes, silicon-based field-effect devices can be used, with neurons interacting capacitively via the transistor gate [70] (Fig. 7). Active CMOS-based multielectrode arrays with a sensitive area of 2 mm  1.75 mm featuring 11,011 platinum electrodes with a diameter of 7 μm and electrode center-to-center distances of 18 μm have been developed by Frey et al. [56]. A fully fledged microelectronic circuit allows for signal amplification (up to 80 dB), filtering (high pass: 0.3–100 Hz, low pass: 3.5–14 kHz) and analog-to-

360

Heinz D. Wanzenboeck et al.

digital conversion (8 bit) directly on the chip and has made it possible to electrically visualize action potential propagation in growing cortical neurons. The application of CMOS based MEAs has been restricted to dedicated research labs until 2014 when Multi Channel Systems has presented an “active” amplified MEA featuring 4225 recordings. Yet the high cost of a CMOS-MEA and the unresolved issue of imaging neuronal cultures on the nontransparent CMOS-MEAs limit its wider application.

5

Microfluidic Neurite Isolation Devices The possibility to functionally isolate neurites from their somata opens up new horizons in neurophysiology in part owing to the rise of microfluidic devices in the last decades of the twentieth century. The growth of neurites can now be studied in greater detail. Depending on the scientific question different microfluidic devices emerged to functionally separate and/or structure growing neurites in cell culture. Most devices share the principle that artificial physical barriers separate the neurites using microchannels—the so-called physical patterning. The geometry of these channels prevents the cells from entering the channels whereas the neurites will grow through them to the other part of the culture chamber. One of the first neurite isolation devices of this kind was reported about one decade ago [71]. This device was made by microfabrication and surface micropatterning techniques and consisted of a polydimethylsiloxane (PDMS) microfluidic part placed on a glass substrate. The basic concept (separating neurites from somata using artificial channels) is still in strong use to form asymmetric microchannels yielding “axonal diodes” (Fig. 8) imposing unidirectional axonal connectivity with 97% selectivity [72]. Alternatively, compartmentalized neuron-arraying microfluidic circuits are employed to yield neuronal networks with cellular inputs reduced by up to two orders of magnitude, thus facilitating data interpretation [73]. A similar design optimizing the number of cells in front of the microchannels was developed by Park et al. Their microchip has a round seeding reservoir in the middle and six reservoirs connected with it via microchannels. Effects of up to six neurotrophic factors can thus be investigated [74]. Investigation of retrograde/anterograde transport in single axons, and mutual effects of soluble factors (e.g., amyloid-β peptides) secreted by different neuronal populations (each “housed” in a different macrochannel) has also been reported [75]. Using a continuous medium flow neurons can be trapped in microwells, while their isolated neurites grow in the flow direction. Thus obtained “one-

Functional Isolation of Neurites in Microchannels

361

Fig. 8 Polarization of axonal growth by “axonal diodes” (asymmetrical, funnel-shaped microchannels) imposing unidirectional axonal connectivity. (A, B) Two macrochannels (blue) linked by five microchannels (orange). (C, D) Axons of cortical neurons can only grow through the microchannels in one direction; the seeding macrochannel is on the left. (E) Quantification of C vs. D. (F, G) Cortical neurons (red) projecting their axons from the left into a macrochannel seeded with striatal neurons (dark) serving as a target (0 and 7 days in culture). Scale bars, 50 μm. Reprinted from Ref. [72] by permission of © Royal Society of Chemistry

way-structured” neuronal network makes it possible to analyze cultured neuronal networks that are regulated both morphologically and functionally [76]. The neurite isolation device can also be conveniently employed to carry out laser axotomy (Fig. 9) [77]. The patterned microchannel networks can also be used to investigate the mechanical properties of functionally isolated

362

Heinz D. Wanzenboeck et al.

Fig. 9 Applications of microfluidics to study neurons of Caenorhabditis elegans. (A) Worms are immobilized and positioned for high-resolution axotomy with an ultrafast laser. (B) Worms are subjected to high-throughput sorting, based on optically recorded phenotypes. Reprinted from Ref. [77] by permission of © Nature Publishing Group

neurites such as stretching or deformation [78], or to stimulate neuronal growth in specific directions by applying mechanical stress [79]. Closed in vitro systems are also advantageous to examine the impact of concentration gradients between the cell seeding reservoir and the target one behind the microchannels. Bhattacharjee et al. presented a gradient-generating device to study the response of mammalian neurons to axon guidance factors [80]. Another device for a simultaneous generation of multiple gradients with gradually changed slopes was developed by Xiao et al. to investigate axonal growth rate and direction in response to substrate-bound laminin gradients with different slopes [81]. Apart from the physical patterning techniques to isolate neurites chemical patterning (surface immobilization) is also an important strategy when surface binding properties or a stand-alone system without physical limitation is required. The geometry of grown neurons can be easily controlled by the use of biomolecules, printed beforehand on the substrate using a microcontact printing technique [61], or by the use of a physiological protein pattern, with polystyrene acting as a repellent background for the cells [82]. New methods also include nanostructured chemical templates for local deposition of gold nanoparticles to pattern and guide neurons [83].

Functional Isolation of Neurites in Microchannels

363

In summary, either of the two methods to isolate neurites (i.e., physical and chemical patterning) has its advantages and disadvantages, depending on the relevant scientific question. Physical patterning yields distinctly separated neurites while misguiding of neurites cannot be fully excluded in chemical patterning. In patch-clamp or microscopy where close physical access to the neurites is necessary, chemical patterning will be the best option to functionally isolate them from cell bodies. Hereinafter we present an approach combining a neuriteisolation device with a microelectrode array, permitting to investigate the effects of added substances (such as growth factors) on signal propagation through a neurite, simultaneously in many neurons.

6

Neurite-Isolation Microelectrode Arrays

6.1 Design Considerations

As mentioned in Subheading 4 there are numerous possibilities to measure physiological activity of neurites, depending on the specific scientific problem to be tracked. Nevertheless, there are some essential considerations in every setup: (1) how to convert the outcome of the experiment to measurable values (optical, electrical, magnetic), (2) how to verify the results of new techniques with the established ones, and (3) how to adapt a device to yield additional parameters. Our task was to provide a platform for investigating the effects of pharmaceuticals on the electrical activity of growing neurites functionally isolated in microchannels, and to correlate this activity with their originating neurites by microscopic observations. For this purpose we have combined a neurite isolation device (NI-device) and a microelectrode array (MEA). We decided to fabricate both devices by ourselves so that we can quickly adapt them if needed. We first experimented with different setups of the NI-devices and MEAs. Figure 10 shows two models we mainly used, with four reservoirs and two macrochannels. The two setups differ in the number of microchannels and the shape, size and alignment of the electrodes. The so-called 35/25-NI-MEA consists of 35 microchannels between the two macrochannels, with one electrode in each microchannel and 25 macroelectrodes in the source macrochannel (where the neurons are seeded. The other one (30-2-NI-MEA) contains 30 microchannels with two electrodes in each microchannel. The design of both devices is presented in Fig. 11. Both designs were used for imaging and recording of spontaneous activity as well as nicotine-induced electrical activity of mouse sympathetic neurons. Designed as experimental (re)generating platform for neurites we demonstrate the sensitivity of the 35/25-NI-MEA using two different concentrations of nerve

364

Heinz D. Wanzenboeck et al.

Fig. 10 Design of two different NI-MEAs. The neurite-isolation (NI) device (blue) consists of four reservoirs, two of them pairwise connected by macrochannels to form a source (were the cells are seeded), and a sink compartment. 35/25-NI-MEA: The two macrochannels are connected by 35 microchannels acting as physical barriers for neuronal somata, thus allowing only neurites to enter. The underlying microelectrode array (MEA, red) is fitted with 25 circular macroelectrodes (in source macrochannel) and 35 square-shaped microelectrodes (each in a different microchannel) to record channel-specific neuronal activity. 30-2-NI-MEA: The two macrochannels (compartments) are connected by 30 microchannels, each harboring two microelectrodes

growth factor (NGF), which is a neurotrophin implicated in axonal growth and guidance. We carried out the recording over several weeks to demonstrate the effect of NGF during the growth of neurites: increased electrical activity and an increasing number of neurites in the microchannels. After multiple electrical measurements during several weeks, the cultures were chemically fixed and processed for immunocytochemistry using antibodies. To differentiate between cell bodies and neurites antibodies specifically directed against axonal (SMI31) and somatodendritic cytoskeletal proteins (MAP2) were used. 6.2 Fabrication, Recording and Imaging 6.2.1 Neurite Isolation Device Fabrication

The NI-device is a transparent microfluidic component fabricated from polydimethylsiloxane (PDMS, Fig. 11A). Further layout description refers to the 35/25-NI-MEA configuration inspired by a setup by Taylor et al. [84]. Four cylindrical wells (diameter: 4 mm; height: 4 mm) serve as reservoirs for the culture medium. Two of their receptacles are connected with each other by macrochannels (width: 4 mm; height: 120 μm) to form a source and sink compartment. The source and sink compartments are connected by

Functional Isolation of Neurites in Microchannels

365

Fig. 11 Concept, design and assembly of the 35/25-Neurite-Isolation-Multielectrode Array (NI-MEA). A-C show the fabricated device (left column), layout of its central part (middle column) and a microscopic image of the microchannels and/or microelectrodes (right column). (A) Neurite-Isolation (NI) device. (B) Multielectrode Array (MEA). (C) NI and MEA combined to 35/25-NI-MEA, with a microelectrode in each microchannel

35 microchannels (length: 450 μm). The microchannels act as physical barriers for neuronal somata, thus allowing only the growing neurites to enter. The fabrication of the NI-device includes a replica-mould process with a silicon master as a negative mould. The structures, each of a different height, are transferred with two different heights onto the silicon master using photoresists made by MicroChem (Westborough, MA, USA): SU-8 2005 for the microchannels, and SU-8 3050 for the macrochannels and the reservoirs. A passivation layer of methyltrichlorosilane (Sigma 679208) was coated onto the master to avoid sticking of PDMS to it, and to enable multiple casts.

366

Heinz D. Wanzenboeck et al.

The PDMS structure was fabricated using Sylgard 184 (Dow Corning, Midland, MI, USA). The base and the curing agent were mixed (10:1 v/v) and the viscous PDMS mixture was transferred onto the master until a height of 4 mm was reached. PDMS was cured for 10 min on a hotplate at 150
C after t demolding from the master; the reservoirs were mechanically punched out to facilitate that. Microchannels of finished 35/25-NI-MEAs had a width of 10.29  0.13 μm and a height of 4.32  0.09 μm (means  SEM, n ¼ 30). The small standard deviations clearly indicate that the fabrication process was precise and highly reproducible. 6.2.2 Microelectrode Array (MEA) Fabrication

The MEA is fabricated on a 49  49 mm borosilicate glass substrate. The 35/25-NI-MEA is fitted with 25 macroelectrodes and 35 microelectrodes on this substrate (Fig. 11B). The circular macroelectrodes (diameter: 100 μm) are played in the source macrochannel. The square-shaped microelectrodes (25  25 μm laterally) are positioned in the microchannels at a distance of 132.5 μm from the source macrochannel. The substrates were cleaned with acetone, isopropanol and nitrogen, and the electrodes’ boundaries were defined lithographically using AZ 5214 E photoresist (MicroChemicals, Ulm, Germany). The metal electrodes were fabricated by a lift-off process using a sputter system (Von Ardenne LS 320 S) resulting in a titanium layer of 20 nm thickness and a top gold layer of 140 nm thickness. To cover the interconnections between these electrodes and the outer electrodes (for the connection to the recording system) a 400 nm insulation layer of silicon nitride was thereafter applied on the substrate using plasma enhanced vapor deposition (PECVD, Oxford Plasmalab 80 plus). Upon exposing all electrodes by dry etching (RIE, Plasmalab System 100) using another photolithographic mask the finished MEAs were inspected by a light microscope.

6.2.3 NI-MEA Alignment

Precise alignment of the microelectrodes with the microchannels can be complicated due to shrinkage of PDMS during the fabrication of the NI-device. The cross-linking of PDMS during the curing process results in shrinkage of the total volume [85]. The exact shrinkage ratio was determined on a test series of NI-MEAs. The distance l between the middle and the outermost microchannel differed from the desired distance L (defined by the photomask, 2676.4 μm) by 2.40.2% (Eq. 1). dev ½% ¼

Ll  100 L

ð1Þ

The shrinkage was roughly uniform over the entire distance of the device. Based on this, we made a new photomask with a distance between each microchannel increased by 2.4%, thus compensating for the shrinkage.

Functional Isolation of Neurites in Microchannels

367

Using a Karl Suss MJB 3 mask aligner the microelectrodes were accurately aligned to the microchannels (Fig. 11C). To form closed macro- and microchannels, that is, to prepare the NI-MEA a tight seal between the NI-device and the MEA was obtained by oxygen plasma treatment using a plasma asher (TePla 100-E, Technics Plasma GmbH, Kirchheim, Germany). The connection to the MEA1060-Inv-BC amplifier (Multi Channel Systems, Reutlingen, Germany) is achieved by 60 square outer electrodes (2.2  2.2 mm laterally, spacing 0.2 mm). 6.2.4 Cell Culture Preparation

Superior cervical ganglion (SCG) cells were prepared as previously described in Subheading 2.7.3. NI-MEAs were kept in an incubator with 5% CO2 and 95% air at 36.5
C. Every 3 days, 50 μl of the culture medium (ca. 80% of its volume) was replaced in all reservoirs.

6.2.5 Electrophysiological Recordings

Electrophysiological recordings carried out at 37
C in a bathing solution consisting of 120 mM NaCl, 3.0 mM KCl, 2.0 mM CaCl2, 2.0 mM MgCl2, 20 mM glucose, 0.1 mg/ml bovine serum albumin (BSA) and 10 mM HEPES adjusted to pH 7.30 with NaOH, with the NI-MEAs connected to a MEA1060-Inv-BC amplifier (Multi Channel Systems, Reutlingen, Germany). Both spontaneous and induced (evoked by adding 3 μM (-)-nicotine, Sigma N3876) neuronal activity was recorded. Supression of electrical activity was achieved by including 1 μM tetrodotoxin (TTX, Latoxan L8503) in the bathing solution. Two minutes of activity were recorded nonstop on days 4, 7, 10, and 14 after plating. We used the MC_RACK v4.3.5 software (Multi Channel Systems, Germany) to record raw data subsequently processed MATLAB (MathWorks, Ismaning, Germany) while employing (1) template detection combined with threshold detection and (2) spike clustering using principal component analysis.

6.2.6 Imaging of Isolated Neurites

After the final recording the cultured SCG neurons were fixed and stained as described in Subheading 3.2. After washing with PBS/Triton, the reservoirs were filled with H2O and covered with a coverslip. Fluorescent images were acquired by a Leica TCS SP5 confocal microscope and the Leica Application Suite software.

6.3

As NI-MEAs are designed for high optical transparency cells can be monitored by phase-contrast and fluorescence microscopy. After seeding, SCGs cells become adherent within 2 h; satellite glial cells proliferate sufficiently large, and neurons (rendered as bright in positive phase contrast) start extending their neurites into microchannels within 2 days. After 2–3 weeks in culture, neuronal bodies

Results

6.3.1 Microscopic Observations

368

Heinz D. Wanzenboeck et al.

Fig. 12 Cultured neurons in neurite-isolation (NI) devices, without electrodes. The cell bodies, too large to enter the microchannels, are confined to the source macrochannel while neurites outgrow into the sink macrochannel. (A) Brightfield image of SCG neurons used in the present study. (B) Phase-contrast (positive type) image of rat E18 cortical neurons; the microchannels appear wider than they really are, due to the “halo” artifact inherent to phase-contrast imaging. Image B reprinted from Ref. [88] by permission of © Nature Publishing Group

are surrounded by confluent glial cells (Fig. 12). Cultures of this age were processed as described in Subheading 3.2.2. Figure 13 shows the middle part of a 35/25-NI-MEA. Staining with the axon-specific antibody SMI31 revealed a rich network of neurites in the source compartment and also the growth of axons through microchannels into the sink compartment. MAP2 labeled both cell bodies and dendrites of SCG neurons. Whereas cell bodies were confined to the source compartment, some antiMAP2-stained neurites extended into the sink compartment, occasionally associated with anti-SMI31-labeled axons (Fig. 13). In nonpolarized neurons the neurites are MAP2-positive at first. During differentiation MAP2 is confined to the initial segment of axons which indicates a high probability that neurites that are long enough to reach the sink compartment represent dendrites [86]. Thus, the culture conditions in NI-MEAs also seem to stimulate the dendritic growth of SCG neurons [87]. The ingrowth of neurites into microchannels was correlated with electrical recordings in four different 35/25-NI-MEAs, to confirm that the electrical measurements were indeed of neurite origin. In 93.6% of cases the neurites were observed in a microchannel at the end of the measurements after fixing and staining, and their activity was recorded beforehand with the relevant microelectrode.

Functional Isolation of Neurites in Microchannels

369

Fig. 13 Part of the 35/25-NI-MEA device showing the source macrochannel with cell bodies of SCG neurons and round macroelectrodes (bottom), the microchannels with microelectrodes (middle), and the outgrowing neurites in the sink macrochannel (top). The two images only differ in antibody staining (left, rabbit anti-MAP2; right, mouse anti-SMI31) 6.3.2 Electrical Activity of Cultured SCG Neurons

After the simultaneous electrical recordings of 120 s duration from the 35 microchannels on days 4, 7, 10, and 14 the spikes were identified and analyzed offline using a custom-made MATLAB program employing a combined threshold-spike shape detection algorithm. Figure 14 shows a typical activity map of a 35/25-NIMEA after inclusion of 3 μM nicotine in the bathing solution of the source compartment. Spike activity in response to nicotinic acetylcholine receptor (nAChR) activation was readily detected by microelectrodes located in the microchannels. The number of channels showing spike activity was dependent on the age of cultures as well as on the concentration of NGF added to the culture medium. Owing to favorable signal-to-noise ratio, action potentials with an amplitude as small as several hundred microvolts only could be detected (Fig. 14, electrodes 1–35 [microchannels]). As the microchannels are so narrow the neurites are in good electrical contact with the microelectrodes. The effective microelectrode area is defined by the overlap of the microelectrode (25  25 μm) and the microchannel (width 10.5 μm).

370

Heinz D. Wanzenboeck et al.

Fig. 14 Activity map showing detected action potentials (red spots) on each electrode of a 35/25-NI-MEA, over time. Electrodes #36 to #60 correspond to the 25 round macroelectrodes in the source (seeding) macrochannel. Due to greater cell-to-electrode distances and poor signal-to-noise ratio the recorded activity is much smaller than in the 35 microchannels where the 35 microelectrodes (#1 to #35) are in intimate contact with the neurites

Electrical activity recorded by macroelectrodes placed in macrochannels was, on the other hand, infrequent and difficult to detect due to small signal amplitudes (Fig. 14, electrodes 36–60 [single macrochannel]). Given the rich network of neurites detected by immunocytochemistry in macrochannels of the source compartment it is unlikely that neuronal processes would bypass the macroelectrodes. Due to the lower signal-to-noise ratio, caused by the much larger surface of macroelectrodes, the detection of sufficient signal amplitudes from macroelectrodes was rather difficult. 6.3.3 Activity of Neurites at Different NGF Concentrations

Neurotrophic factors facilitate the guidance of axons to the target cell [89]. Hence, we applied different NGF concentration to the source seeding and sink compartments to study such effects. NGF is not only essential for sympathetic neurons’ survival but also promotes the growth of neurites. NGF is also present in the tube formed by Schwann cells as a neuronal guidance cue [90]. The design of these experiments is shown in Fig. 8. Hence, cells in the source and sink compartment were kept in culture medium enriched with NGF at either 10 or 20 ng/ml.

Functional Isolation of Neurites in Microchannels

371

Fig. 15 Effect of two different concentrations of nerve growth factor (NGF) on the electrical activity of individual neurites growing through microchannels of the neurite-isolation microelectrode array (35/25-NI-MEA) comprising 35 microchannels. (A) Number of microchannels with recorded activity, regardless of discharge rate (data from three experiments). (B) Normalized activity (discharge rate) in 35 individual microchannels; the area under the curve represents the overall activity of all neurites for each NGF concentration

Recordings of the electrical activity induced by 3 μM nicotine showed that the number of active microchannels (i.e., those with detectable spike activity) increased with the age of cultures. Figure 15 shows this total number of microchannels where electrical activity was recorded but differences in the frequency of discharges of individual microchannels were not yet considered. Whereas the number of active microchannels represents a reliable indicator of neurite ingrowth into the microchannels, this parameter per se does not provide any information on the overall neuronal activity in a single microchannel. We thus rescaled the data by setting the discharge frequency in the most active channel of each NI-MEA to unity (Eq. 2), and channels were sorted by their activity (regardless of their physical position). Relative spike rate ðcurrent channelÞ ¼ spike rate ðcurrent channelÞ=spike rate ðthe most active channelÞ ð2Þ The channel-related distribution of normalized and averaged neuronal activity was determined as a histogram shown in Fig. 15. Despite identical culturing conditions we experienced a large variability among the three SCG cultures. A somewhat higher activity was detected at higher (20 ng/ml) NGF concentration is indicated in Fig. 15A. Interestingly, one neurite exposed to a low concentration (10 ng/ml) of NGF exhibited with a particularly high discharge rate. As a result, the relevant histogram shown in Fig. 15B is relatively steep.

372

Heinz D. Wanzenboeck et al.

In contrast NI-MEAs kept in a higher (20 ng/ml) NGF level appeared to have more microchannels with a higher discharge rate. At 20 ng/ml concentration of NGF the neuronal activity reached saturation already on day 7 (data not shown). By day 14, the shape of the histograms and thus the “distribution” of active channels within the NI-MEAs became quite similar for both NGF concentrations. The developed platform allows a more comprehensive experimental design to investigate the role of NGF in artificial bands of Bu¨ngner (endoneurial tubes) in neuronal regeneration.

7

Summary and Outlook The combined neurite-isolation and microelectrode-array approach presented here makes it possible to carry out correlative microscopical (e.g., immunofluorescence) and electrophysiological recording of many individual neurites simultaneously and is thus “massively parallel.” High-throughput testing of substances identified as promising in the pharmaceutical industry, or those potentially neurotoxic, would be a logical extension of this approach. To study the growth of neurites, primary neuronal cultures are ideal as they make it possible to investigate the emergence of a functionally inter-connected neuronal network starting from single neurons. Prospective drugs should thus be ideally screened on these cultures, thus paving the way, for example, to the treatment of neurodegenerative diseases or spinal cord injuries. In authors’ opinion, the following technological challenges are expected to be addressed in the near future: l

Placing more than one microelectrode inside the same neurite microchannel, in order to determine the propagation speed of an electrophysiological response to a given stimulus, within a single neurite.

l

Modification of the neurite-isolation device so that individual microchannels are vertically separated, to enable application of substances of interest directly to individual neurites (i.e., avoiding the macrochannels).

l

Thinner glass substrates of the microelectrode array, to enable compatibility with high-power microscope lenses for monitoring neurites in the microchannels.

Functional Isolation of Neurites in Microchannels

373

References 1. Fattahi P, Yang G, Kim G, Abidian MR (2014) A review of organic and inorganic biomaterials for neural interfaces. Adv Mater 26 (12):1846–1885. https://doi.org/10.1002/ adma.201304496 2. Giordano G, Costa L (2011) Primary neurons in culture and neuronal cell lines for in vitro neurotoxicological studies. In: Costa LG, Giordano G, Guizzetti M (eds) In vitro neurotoxicology. (Meth Mol Biol, vol 758. Humana Press, Totowa, NJ, pp 13–27. https://doi.org/ 10.1007/978-1-61779-170-3_2 3. Biedler JL, Helson L, Spengler BA (1973) Morphology and growth, tumorigenicity, and cytogenetics of human neuroblastoma cells in continuous culture. Cancer Res 33 (11):2643–2652. https://www.ncbi.nlm.nih. gov/pubmed/4748425 4. Biedler JL, Roffler-Tarlov S, Schachner M, Freedman LS (1978) Multiple neurotransmitter synthesis by human neuroblastoma cell lines and clones. Cancer Res 38(11):3751–3757. https://www.ncbi.nlm.nih.gov/pubmed/ 29704 5. Greene LA, Tischler AS (1976) Establishment of a noradrenergic clonal line of rat adrenal pheochromocytoma cells which respond to nerve growth factor. Proc Natl Acad Sci U S A 73(7):2424–2428. https://doi.org/10. 1073/pnas.73.7.2424 6. Hilgenberg LG, Smith MA (2007) Preparation of dissociated mouse cortical neuron cultures. J Vis Exp 10:562. https://doi.org/10.3791/ 562 7. Kramer D, Minichiello L (2010) Cell culture of primary cerebellar granule cells. Meth Mol Biol 633:233–239. https://doi.org/10.1007/ 978-1-59745-019-5_17 8. Mao L, Wang J (2003) Primary striatal neuronal culture. In: Wang J (ed) Drugs of abuse. (Meth Mol Med, vol 79. Humana Press, Totowa, NJ, pp 379–386. https://doi.org/ 10.1385/1-59259-358-5:379 9. Zeitelhofer M, Vessey JP, Xie Y, Tubing F, Thomas S, Kiebler M, Dahm R (2007) Highefficiency transfection of mammalian neurons via nucleofection. Nat Protoc 2 (7):1692–1704. https://doi.org/10.1038/ nprot.2007.226 10. De Biasi M (2002) Nicotinic receptor mutant mice in the study of autonomic function. Curr Drug Targets CNS Neurol Disord 1 (4):331–336. https://doi.org/10.2174/ 1568007023339148

11. Fischer H, Orr-Urtreger A, Role LW, Huck S (2005) Selective deletion of the α5 subunit differentially affects somatic-dendritic versus axonally targeted nicotinic ACh receptors in mouse. J Physiol 563(Pt 1):119–137. https:// doi.org/10.1113/jphysiol.2004.075788 12. Putz G, Kristufek D, Orr-Urtreger A, Changeux JP, Huck S, Scholze P (2008) Nicotinic acetylcholine receptor-subunit mRNAs in the mouse superior cervical ganglion are regulated by development but not by deletion of distinct subunit genes. J Neurosci Res 86(5):972–981. https://doi.org/10.1002/jnr.21559 13. David R, Ciuraszkiewicz A, Simeone X, Orr-Urtreger A, Papke RL, McIntosh JM, Huck S, Scholze P (2010) Biochemical and functional properties of distinct nicotinic acetylcholine receptors in the superior cervical ganglion of mice with targeted deletions of nAChR subunit genes. Eur J Neurosci 31 (6):978–993. https://doi.org/10.1111/j. 1460-9568.2010.07133.x 14. Scholze P, Ciuraszkiewicz A, Groessl F, Orr-Urtreger A, McIntosh JM, Huck S (2011) α4β2 nicotinic acetylcholine receptors in the early postnatal mouse superior cervical ganglion. Dev Neurobiol 71(5):390–399. https://doi.org/10.1002/dneu.20870 15. Ciuraszkiewicz A, Schreibmayer W, Platzer D, Orr-Urtreger A, Scholze P, Huck S (2013) Single-channel properties of α3β4, α3β4α5 and α3β4β2 nicotinic acetylcholine receptors in mice lacking specific nicotinic acetylcholine receptor subunits. J Physiol 591 (Pt13):3271–3288. https://doi.org/10. 1113/jphysiol.2012.246595 16. Blumenfeld H (2003) From molecules to networks: cortical/subcortical interactions in the pathophysiology of idiopathic generalized epilepsy. Epilepsia 44:7–15. https://doi.org/10. 1046/j.1528-1157.44.s.2.2.x 17. Luna LG (1968) Manual of histologic staining methods of the Armed Forces Institute of Pathology. McGraw-Hill (Blakiston Div.), New York. https://archive.org/details/ ManualHistologicalStaining 18. Coons AH, Creech HJ, Jones RN (1941) Immunological properties of an antibody containing a fluorescent Group. Exp Biol Med (Maywood) 47(2):200–202. https://doi.org/ 10.3181/00379727-47-13084p 19. Volgyi B, Bloomfield SA (2002) Axonal neurofilament-H immunolabeling in the rabbit retina. J Comp Neurol 453(3):269–279. https://doi.org/10.1002/cne.10392

374

Heinz D. Wanzenboeck et al.

20. Seifert U, Hartig W, Grosche J, Bruckner G, Riedel A, Brauer K (1998) Axonal expression sites of tyrosine hydroxylase, calretinin- and calbindin-immunoreactivity in striato-pallidal and septal nuclei of the rat brain: a doubleimmunolabelling study. Brain Res 795 (1–2):227–246. https://doi.org/10.1016/ S0006-8993(98)00298-4 21. Haykal-Coates N, O’Callaghan JP, Reinhard JF Jr, Jensen KF (1991) Pargyline and γ-butyrolactone enhance tyrosine hydroxylase immunostaining of nigrostriatal axons. Brain Res 556(2):353–357. https://doi.org/10.1016/ 0006-8993(91)90330-X 22. Hechler D, Nitsch R, Hendrix S (2006) Greenfluorescent-protein-expressing mice as models for the study of axonal growth and regeneration in vitro. Brain Res Rev 52(1):160–169. https://doi.org/10.1016/j.brainresrev.2006. 01.005 23. Zhou XY, Ji M, Liu JY, Tan MQ (2007) GFP gene transfection of dendritic cells mediated by recombinated adeno-associated virus. J Centr South Univ (Med Sci) 32(5):742–746. (in Chinese). https://www.ncbi.nlm.nih.gov/ pubmed/18007063 24. van den Pol AN, Spencer DD (2000) Differential neurite growth on astrocyte substrates: interspecies facilitation in green fluorescent protein-transfected rat and human neurons. Neuroscience 95(2):603–616. https://doi. org/10.1016/s0306-4522(99)00430-3 25. Cassereau J, Nicolas G, Lonchampt P, Pinier M, Barthelaix A, Eyer J, Letournel F (2013) Axonal regeneration is compromised in NFH-LacZ transgenic mice but not in NFH-GFP mice. Neuroscience 228:101–108. https://doi.org/10.1016/j.neuroscience. 2012.10.011 26. Kameda H, Furuta T, Matsuda W, Ohira K, Nakamura K, Hioki H, Kaneko T (2008) Targeting green fluorescent protein to dendritic membrane in central neurons. Neurosci Res 61(1):79–91. https://doi.org/10.1016/j.neu res.2008.01.014 27. Sesack SR, Hawrylak VA, Matus C, Guido MA, Levey AI (1998) Dopamine axon varicosities in the prelimbic division of the rat prefrontal cortex exhibit sparse immunoreactivity for the dopamine transporter. J Neurosci 18 (7):2697–2708. https://www.ncbi.nlm.nih. gov/pubmed/9502827 28. Hartline FF (1979) Biological applications for voltage sensitive dyes. Science 203 (4384):992–994. https://doi.org/10.1126/ science.424731 29. Ross WN, Reichardt LF (1979) Speciesspecific effects on the optical signals of

voltage-sensitive dyes. J Membr Biol 48 (4):343–356. https://doi.org/10.1007/ BF01869445 30. Baker BJ, Kosmidis EK, Vucinic D, Falk CX, Cohen LB, Djurisic M, Zecevic D (2005) Imaging brain activity with voltage- and calcium-sensitive dyes. Cell Mol Neurobiol 25 (2):245–282. https://doi.org/10.1007/ s10571-005-3059-6 31. Zochowski M, Wachowiak M, Falk CX, Cohen LB, Lam YW, Antic S, Zecevic D (2000) Imaging membrane potential with voltage-sensitive dyes. Biol Bull 198(1):1–21. https://doi.org/ 10.2307/1542798 32. Foust A, Popovic M, Zecevic D, McCormick DA (2010) Action potentials initiate in the axon initial segment and propagate through axon collaterals reliably in cerebellar Purkinje neurons. J Neurosci 30(20):6891–6902. https://doi.org/10.1523/jneurosci.0552-10. 2010 33. Popovic M, Gao X, Zecevic D (2012) Voltagesensitive dye recording from axons, dendrites and dendritic spines of individual neurons in brain slices. J Vis Exp 69:e4261. https://doi. org/10.3791/4261 34. Grinvald A, Ross WN, Farber I (1981) Simultaneous optical measurements of electrical activity from multiple sites on processes of cultured neurons. Proc Natl Acad Sci U S A 78(5):3245–3249. https://doi.org/10.1073/ pnas.78.5.3245 35. Kinoshita M, Ueda R, Kojima S, Sato K, Watanabe M, Urano A, Ito E (2002) Multiple-site optical recording for characterization of functional synaptic organization of the optic tectum of rainbow trout. Eur J Neurosci 16(5):868–876. https://doi.org/10.1046/j. 1460-9568.2002.02160.x 36. Tsytsarev V, Liao LD, Kong KV, Liu YH, Erzurumlu RS, Olivo M, Thakor NV (2014) Recent progress in voltage-sensitive dye imaging for neuroscience. J Nanosci Nanotechnol 14(7):4733–4744. https://doi.org/10.1166/ jnn.2014.9531 37. Mao C, Kisaalita WS (2004) Determination of resting membrane potential of individual neuroblastoma cells (IMR-32) using a potentiometric dye (TMRM) and confocal microscopy. J Fluoresc 14(6):739–743. https://doi.org/ 10.1023/B:JOFL.0000047224.41328.f8 38. Maric D, Maric I, Smith SV, Serafini R, Hu Q, Barker JL (1998) Potentiometric study of resting potential, contributing K+ channels and the onset of Na+ channel excitability in embryonic rat cortical cells. Eur J Neurosci 10 (8):2532–2546. https://doi.org/10.1046/j. 1460-9568.1998.00284.x

Functional Isolation of Neurites in Microchannels 39. Bullen A, Saggau P (1999) High-speed, random-access fluorescence microscopy: II. Fast quantitative measurements with voltagesensitive dyes. Biophys J 76(4):2272–2287. https://doi.org/10.1016/s0006-3495(99) 77383-2 40. Mohajerani MH, McVea DA, Fingas M, Murphy TH (2010) Mirrored bilateral slow-wave cortical activity within local circuits revealed by fast bihemispheric voltage-sensitive dye imaging in anesthetized and awake mice. J Neurosci 30(10):3745–3751. https://doi.org/10. 1523/jneurosci.6437-09.2010 41. Kibler AB, Jamieson BG, Durand DM (2012) A high aspect ratio microelectrode array for mapping neural activity in vitro. J Neurosci Methods 204(2):296–305. https://doi.org/ 10.1016/j.jneumeth.2011.11.027 42. Johansson F, Wallman L, Danielsen N, Schouenborg J, Kanje M (2009) Porous silicon as a potential electrode material in a nerve repair setting: Tissue reactions. Acta Biomater 5(6):2230–2237. https://doi.org/10.1016/j. actbio.2009.02.010 43. Renshaw S (ed) (2007) Immunohistochemistry. Scion Publishing, Bloxham 44. Herzog W, Weber K (1978) Fractionation of brain microtubule-associated proteins. Isolation of two different proteins which stimulate tubulin polymerization in vitro. Eur J Biochem 92(1):1–8. https://doi.org/10.1111/j.14321033.1978.tb12716.x 45. Be´barova´ M (2012) Advances in patch clamp technique: towards higher quality and quantity. Gen Physiol Biophys 31(2):131–140. https:// doi.org/10.4149/gpb_2012_016 46. Hodgkin AL, Huxley AF (1952) A quantitative description of membrane current and its application to conduction and excitation in nerve. J Physiol 117(4):500–544. https://www.ncbi. nlm.nih.gov/pubmed/12991237 47. Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ (1981) Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflugers Arch 391(2):85–100. https://doi. org/10.1007/BF00656997 48. Neher E, Sakmann B, Steinbach JH (1978) The extracellular patch clamp: a method for resolving currents through individual open channels in biological membranes. Pflugers Arch 375(2):219–228. https://doi.org/10. 1007/BF00584247 49. Neher E, Sakmann B (1992) The patch clamp technique. Sci Am 266(3):44–51. https://doi. org/10.1038/scientificamerican0392-44

375

50. Caldwelland JH, Milton RL (1995) The loose patch voltage clamp technique. In: Boulton A, Baker G, Walz W (eds) Patch-clamp applications and protocols, Neuromethods, vol 26. Humana Press, Totowa, NJ, pp 173–192. https://doi.org/10.1385/0-89603-3112:173 51. Messerli MA, Collis LP, Smith PJS (2009) Ion trapping with fast-response ion-selective microelectrodes enhances detection of extracellular ion channel gradients. Biophys J 96 (4):1597–1605. https://doi.org/10.1016/j. bpj.2008.11.025 52. Spira ME, Hai A (2013) Multi-electrode array technologies for neuroscience and cardiology. Nat Nanotechnol 8(2):83–94. https://doi. org/10.1038/nnano.2012.265 53. Thomas CA Jr, Springer PA, Loeb GE, Berwald-Netter Y, Okun LM (1972) A miniature microelectrode array to monitor the bioelectric activity of cultured cells. Exp Cell Res 74(1):61–66. https://doi.org/10.1016/ 0014-4827(72)90481-8 54. Liu M-G, Chen X-F, He T, Li Z, Chen J (2012) Use of multi-electrode array recordings in studies of network synaptic plasticity in both time and space. Neurosci Bull 28(4):409–422. https://doi.org/10.1007/s12264-012-12515 55. Hutzler M, Lambacher A, Eversmann B, Jenkner M, Thewes R, Fromherz P (2006) High-resolution multitransistor array recording of electrical field potentials in cultured brain slices. J Neurophysiol 96 (3):1638–1645. https://doi.org/10.1152/jn. 00347.2006 56. Frey U, Egert U, Heer F, Hafizovic S, Hierlemann A (2009) Microelectronic system for high-resolution mapping of extracellular electric fields applied to brain slices. Biosens Bioelectron 24(7):2191–2198. https://doi.org/ 10.1016/j.bios.2008.11.028 57. Berdondini L, Imfeld K, Maccione A, Tedesco M, Neukom S, Koudelka-Hep M, Martinoia S (2009) Active pixel sensor array for high spatio-temporal resolution electrophysiological recordings from single cell to large scale neuronal networks. Lab Chip 9 (18):2644–2651. https://doi.org/10.1039/ b907394a 58. Huys R, Braeken D, Jans D, Stassen A, Collaert N, Wouters J, Loo J, Severi S, Vleugels F, Callewaert G, Verstreken K, Bartic C, Eberle W (2012) Single-cell recording and stimulation with a 16k micro-nail electrode array integrated on a 0.18 μm CMOS chip. Lab Chip 12(7):1274–1280. https:// doi.org/10.1039/c2lc21037a

376

Heinz D. Wanzenboeck et al.

59. Lewicki MS (1998) A review of methods for spike sorting: the detection and classification of neural action potentials. Network 9(4): R53–R78. https://doi.org/10.1088/0954898X_9_4_001 60. Brown EN, Kass RE, Mitra PP (2004) Multiple neural spike train data analysis: state-of-the-art and future challenges. Nat Neurosci 7 (5):456–461. https://doi.org/10.1038/ nn1228 61. Offenh€ausser A, Bo¨cker-Meffert S, Decker T, Helpenstein R, Gasteier P, Groll J, Mo¨ller M, Reska A, Sch€afer S, Schulte P, Vogt-Eisele A (2007) Microcontact printing of proteins for neuronal cell guidance. Soft Matter 3 (3):290–298. https://doi.org/10.1039/ B607615G 62. Pushparaj VL, Shaijumon MM, Kumar A, Murugesan S, Ci L, Vajtai R, Linhardt RJ, Nalamasu O, Ajayan PM (2007) Flexible energy storage devices based on nanocomposite paper. Proc Natl Acad Sci U S A 104 (34):13574–13577. https://doi.org/10. 1073/pnas.0706508104 63. Klemic KG, Klemic JF, Reed MA, Sigworth FJ (2002) Micromolded PDMS planar electrode allows patch clamp electrical recordings from cells. Biosens Bioelectron 17(6–7):597–604. https://doi.org/10.1016/S0956-5663(02) 00015-5 64. Andersson H, van den Berg A (2004) Microtechnologies and nanotechnologies for singlecell analysis. Curr Opin Biotechnol 15 (1):44–49. https://doi.org/10.1016/j. copbio.2004.01.004 65. Bruhn BR, Liu H, Schuhladen S, Hunt AJ, Mordovanakis A, Mayer M (2014) Dual-pore glass chips for cell-attached single-channel recordings. Lab Chip 14(14):2410–2417. https://doi.org/10.1039/c4lc00370e 66. Fertig N, Blick RH, Behrends JC (2002) Whole cell patch clamp recording performed on a planar glass chip. Biophys J 82 (6):3056–3062. https://doi.org/10.1016/ s0006-3495(02)75646-4 67. Py C, Denhoff MW, Sabourin N, Weber J, Shiu M, Zhao P (2014) Priming and testing silicon patch-clamp neurochips. N Biotechnol 31(5):430–435. https://doi.org/10.1016/j. nbt.2014.04.003 68. Py C, Martina M, Diaz-Quijada GA, Luk CC, Martinez D, Denhoff MW, Charrier A, Comas T, Monette R, Krantis A, Syed NI, Mealing GA (2011) From understanding cellular function to novel drug discovery: the role of planar patch-clamp array chip technology. Front Pharmacol 2:51. https://doi.org/10. 3389/fphar.2011.00051

69. Hierlemann A, Frey U, Hafizovic S, Heer F (2011) Growing cells atop microelectronic chips: Interfacing electrogenic cells in vitro with CMOS-based microelectrode arrays. Proc IEEE 99(2):252–284. https://doi.org/ 10.1109/JPROC.2010.2066532 70. Fromherz P, Offenhausser A, Vetter T, Weis J (1991) A neuron-silicon junction: a Retzius cell of the leech on an insulated-gate fieldeffect transistor. Science 252 (5010):1290–1293. https://doi.org/10. 1126/science.1925540 71. Taylor AM, Rhee SW, Tu CH, Cribbs DH, Cotman CW, Jeon NL (2003) Microfluidic multicompartment device for neuroscience research. Langmuir 19(5):1551–1556. https://doi.org/10.1021/la026417v 72. Peyrin JM, Deleglise B, Saias L, Vignes M, Gougis P, Magnifico S, Betuing S, Pietri M, Caboche J, Vanhoutte P, Viovy JL, Brugg B (2011) Axon diodes for the reconstruction of oriented neuronal networks in microfluidic chambers. Lab Chip 11(21):3663–3673. https://doi.org/10.1039/c1lc20014c 73. Dinh ND, Chiang YY, Hardelauf H, Baumann J, Jackson E, Waide S, Sisnaiske J, Frimat JP, Thriel CV, Janasek D, Peyrin JM, West J (2013) Microfluidic construction of minimalistic neuronal co-cultures. Lab Chip 13(7):1402–1412. https://doi.org/10.1039/ c3lc41224e 74. Park J, Kim S, Park SI, Choe Y, Li J, Han A (2014) A microchip for quantitative analysis of CNS axon growth under localized biomolecular treatments. J Neurosci Methods 221 (0):166-174. doi:https://doi.org/10. 1016/j.jneumeth.2013.09.018 75. Poon WW, Blurton-Jones M, Tu CH, Feinberg LM, Chabrier MA, Harris JW, Jeon NL, Cotman CW (2011) β-Amyloid impairs axonal BDNF retrograde trafficking. Neurobiol Aging 32(5):821–833. https://doi.org/10. 1016/j.neurobiolaging.2009.05.012 76. Takayama Y, Kotake N, Haga T, Suzuki T, Mabuchi K (2012) Formation of one-waystructured cultured neuronal networks in microfluidic devices combining with micropatterning techniques. J Biosci Bioeng 114 (1):92–95. https://doi.org/10.1016/j.jbiosc. 2012.02.011 77. Hulme SE, Shevkoplyas SS, Samuel A (2008) Microfluidics: streamlining discovery in worm biology. Nat Meth 5(7):589–590. https://doi. org/10.1038/nmeth0708-589 78. Fior R, Maggiolino S, Lazzarino M, Sbaizero O (2011) A new transparent Bio-MEMS for uni-axial single cell stretching. Microsystem Technologies 17(10–11):1581–1587.

Functional Isolation of Neurites in Microchannels https://doi.org/10.1007/s00542-011-13258 79. Nguyen TD, Hogue IB, Cung K, Purohit PK, McAlpine MC (2013) Tension-induced neurite growth in microfluidic channels. Lab Chip 13(18):3735–3740. https://doi.org/10. 1039/c3lc50681a 80. Bhattacharjee N, Li N, Keenan TM, Folch A (2010) A neuron-benign microfluidic gradient generator for studying the response of mammalian neurons towards axon guidance factors. Integr Biol 2(11-12):669–679. https://doi. org/10.1039/c0ib00038h 81. Xiao RR, Zeng WJ, Li YT, Zou W, Wang L, Pei XF, Xie M, Huang WH (2013) Simultaneous generation of gradients with gradually changed slope in a microfluidic device for quantifying axon response. Anal Chem 85 (16):7842–7850. https://doi.org/10.1021/ ac4022055 82. Vogt AK, Lauer L, Knoll W, Offenh€ausser A (2003) Micropatterned substrates for the growth of functional neuronal networks of defined geometry. Biotechnol Prog 19 (5):1562–1568. https://doi.org/10.1021/ bp034016f 83. Gilles S, Winter S, Michael KE, Meffert SH, Li P, Greben K, Simon U, Offenh€ausser A, Mayer D (2012) Control of cell adhesion and neurite outgrowth by patterned gold nanoparticles with tunable attractive or repulsive surface properties. Small 8(21):3357–3367. https://doi.org/10.1002/smll.201200465 84. Taylor AM, Blurton-Jones M, Rhee SW, Cribbs DH, Cotman CW, Jeon NL (2005) A

377

microfluidic culture platform for CNS axonal injury, regeneration and transport. Nat Meth 2 (8):599–605. https://doi.org/10.1038/ Nmeth777 85. Lee SW, Lee SS (2008) Shrinkage ratio of PDMS and its alignment method for the wafer level process. Microsystem Technologies 14(2):205–208. https://doi.org/10.1007/ s00542-007-0417-y 86. de Lima AD, Merten MD, Voigt T (1997) Neuritic differentiation and synaptogenesis in serum-free neuronal cultures of the rat cerebral cortex. J Comp Neurol 382(2):230–246. https://doi.org/10.1002/(SICI)1096-9861( 19970602)382:23.0. CO;2-4 87. Lein P, Johnson M, Guo X, Rueger D, Higgins D (1995) Osteogenic protein-1 induces dendritic growth in rat sympathetic neurons. Neuron 15(3):597–605. https://doi.org/10. 1016/0896-6273(95)90148-5 88. Park JW, Vahidi B, Taylor AM, Rhee SW, Jeon NL (2006) Microfluidic culture platform for neuroscience research. Nat Protocols 1 (4):2128–2136. https://doi.org/10.1038/ nprot.2006.316 89. Bosse F (2012) Extrinsic cellular and molecular mediators of peripheral axonal regeneration. Cell Tissue Res 349(1):5–14. https://doi. org/10.1007/s00441-012-1389-5 90. Sofroniew MV, Howe CL, Mobley WC (2001) Nerve growth factor signaling, neuroprotection, and neural repair. Annu Rev Neurosci 24:1217–1281. https://doi.org/10.1146/ annurev.neuro.24.1.1217

Chapter 13 Consumer Versus Dedicated Digital Cameras in Photomicrography Jo¨rg Piper and Radek Pelc Abstract A number of consumer digital cameras (compact, bridge, single lens reflex [SLR], and system ones) are of sufficiently high quality to qualify as suitable for photomicrography and represent an affordable alternative to dedicated, high-end cameras typically equipped with very sensitive sensors. When the image sensor resolution is at least 6 or 8 megapixel digital images offer rendering of details that is comparable to conventional micrographs taken on a standard 36
24 mm film. In most situations, micrographs taken by high-end (SLR) or other cameras feature no obvious differences in quality, so that even compact or bridge cameras may be used in most cases. Otherwise, for example, in low-light conditions or when very large print formats are required, SLR camera may be needed owing to its low noise, superb resolution and high ISO speed range. Dedicated moderate-cost cameras equipped with CMOS sensors represent an optimal solution for high-resolution video clips and in situation when life-view images have to be presented on high-resolution screens. On the other hand, color images are better rendered by high-end system cameras and ordinary (consumer) cameras. Layout of photosensitive cells in the retina across taxonomical groups is presented as an analogy of image sensor designs. Key words Photomicrography, Compact camera, Bridge camera, Mirror-reflex camera, System camera, Sensor, Resolution

Abbreviations AEB APS CCD CMOS DRI DSLR EOS fps EOS/LER EV HDR/HDRR lp/mm

Automatic exposure bracketing Advanced photographic system Charge-coupled device Complementary metal oxide semiconductor Dynamic range increase Digital single-lens reflex (mirror reflex) camera Canon camera series (electro-optical system) Frames per second Canon EOS to Leica-R (lens adapter) Exposure value High dynamic range rendering Line pairs per millimeter

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_13, © Springer Science+Business Media LLC 2020

379

Jo¨rg Piper and Radek Pelc

380

MFT MP NA NMOS SLR

1

Micro Four Thirds Megapixel Numerical aperture N-type metal-oxide-semiconductor Single-lens reflex (mirror reflex) camera

Introduction For more than 150 years, images have been recorded from the microscope [1]. Fundamental technical basics of photomicrography in general have been compiled elsewhere [2, 3]. Needless to say, digital cameras are nowadays so common that the analog ones are close to extinction. Historically, the adaptation of Nikon Coolpix 950/990 series cameras to photomicrography can be regarded as a pioneering attempt offering for the first time the option to “go digital” in an easy and cost-efficient manner [4]. The obvious advantage of digital cameras is the option to immediately check image on the camera display, much like in Polaroid photographs heralded in the predigital era. A vast range of dedicated photomicrography cameras is available, with ultrahigh sensitivity and dynamic range, and extraordinarily low noise level (“high-end” segment models). This holds especially for cameras equipped with cooled sensors. On the other hand, such specialized cameras can only be used in photomicrography, and are rather expensive compared to “normal” (consumer) digital cameras, partly because they are produced in much lower numbers. The latter represent an attractive alternative. Modern consumer cameras are fitted with high-resolution sensors, and high ISO speed values can be achieved owing to efficient noise-reduction algorithms. High-end consumer digital cameras yield high-fidelity images, at least in routine situations (standard specimens and/or common illumination techniques). Nevertheless, not every type of a digital camera is suitable for photomicrography, and the present chapter aims to highlight that. Digital sensors are traditionally referred to with a “type” designation, for example, 1/2.300 (ca. 11 mm). However, this is not any of the sensor dimensions (6.2
4.6 mm), and the sensor diagonal (7.7 mm) is only approximately two-thirds of this figure (Fig. 1A). While 1 in. is defined as 25.4 mm under normal circumstances, for historical reasons it represents only ca. 16 mm when describing digital camera sensors [5]. These are typically smaller than the so-called full frame defined as the light-sensitive part of a standard photographic film (36
24 mm). In a 100 analog video camera tube such as vidicon (the ‘ancestor’ of digital image sensors), 25.4 and 16 mm was its outer and usable inner diameter, respectively.

Photomicrography

381

A

B

D

C

R

G

B

L

M

S

5’ arc

Fig. 1 Digital image sensors versus human retina. (A) Sensor sizes commonly employed in consumer digital cameras. The sensor diagonal is approximately two-thirds of the value stated in the figure (in inches). “Full Frame” denotes the sensitive area of a standard film (36
24 mm). (APS) Advanced photographic system. (MFT) Micro four thirds. (B, C) Anatomy of a color image sensor. Each pixel (square) is covered by a color filter (Red, Green or Blue). (B) Bayer model. A regular 2
2 pixel motif is repeated throughout the sensor. (C) X-trans model (Fuji). A highly aperiodic 6
6 pixel motif is employed to reduce Moire´ and other artifacts. (D) Photoreceptive cells (cones) in human perifovea under an ophthalmoscope. These are primarily but not exclusively sensitive to long (L), medium (M), or short (S) wavelengths. Angular size of the white ring segments drawn over L and M cells represents the frequency of each cell’s color insensitivity. The S cones were not evaluated. Image D is reprinted from Ref. [18] by permission of © AAAS

The resolution of an analog (film) camera is described by the maximum number of line pairs per millimeter (lp/mm) that can still be resolved in the film emulsion. The mean value is 50 lp/mm, and it is modulated by the lens quality (about 40 lp/mm for standard zoom lenses and 60 lp/mm for high-end lenses). In consumer digital cameras, 40/50/60 lp/mm translates to ca. 5.5/8.5/12 megapixel (MP). A digital camera equipped with a 8 MP sensor thus approximately corresponds to an average analog film camera.

382

Jo¨rg Piper and Radek Pelc

The smaller the physical size of the digital sensor, the greater the imaging artifacts, especially blooming and noise. The maximum usable ISO speed equivalent of a digital camera thus depends on sensor size, and in low-light conditions such as fluorescence or polarization microscopy, large sensors are preferred. While conventional photographic film consists of three or four layers (each sensitive to one of the basic colors) sensors employed in digital cameras typically consist of a single layer of photodiodes arranged in a rectangular array of pixels. To enable color imaging, each pixel is coated with a filter for one of the three primary colors, red, green and blue (RGB). Typically, these filters are arranged in repeating 2
2 or 6
6 patters (Bayer model and a more recently developed X-Trans sensor from Fuji, Fig. 1B, C). Light entering them is thus filtered in green, red or blue [6]. As light entering at least half of all pixels (50% and ~55% in the Bayer and Fuji models, respectively) is filtered in green, the effective number of available pixels (and thus image resolution) is much lower for specimens that are predominantly red or blue (see Subheading 4.4 for details). A living counterpart of the electronic image sensor is of course the retina and its array of photosensitive cells (Figs. 1D and 2). Their layout varies greatly across taxonomical groups (Fig. 2), and primitive organisms can have as little as 30 photoreceptive cells (Fig. 2A). The most commonly used sensor types are charge-coupled device (CCD), complementary metal oxide semiconductor (CMOS) and N-type metal-oxide-semiconductor (NMOS). While CCD sensors are characterized by low noise levels CMOS sensors exhibit high frame rates and are thus particularly suitable for video recordings or life views. NMOS sensors are fitted with micro lenses and filters consisting of other materials than in common CMOS sensors. It still remains to be seen whether this new “-MOS” sensor will really yield enhanced image quality.

2

Types of Consumer Digital Cameras Digital consumer cameras can be divided into four categories: (1) compact cameras, (2) bridge cameras, (3) single lens reflex cameras (SLRs), and (4) system cameras. The main characteristics of these construction types are described in detail below, and compiled in Table 1.

2.1 Compact Cameras

These cameras are equipped with the smallest sensors (1/2.300 ¼ 6.2
4.6 mm in most cases) and fitted with the smallest objectives so that they typically weigh very little (ca. 100–200 g). The noise level of these cameras is accordingly higher than in the other types, and the range of ISO speed values that can be reasonably used is limited. Objectives are mostly designed as zoom systems, and cannot be easily replaced (i.e., without completely

Photomicrography

SEA SQUIRT

A

FRUIT FLY

C1

2°

~10 µm

A

B1

PEARLEYE FISH

~5 µm

10 µm MORMYRID FISH

L1 = Lobe 1

ZEBRAFISH

D

Rod-like

B2

383

C2

UV cones Rods

50 µm

MARMOSET

E

L2 = Lobe 2

~5 µm

10 µm

10 µm Cones

S cone

Fig. 2 “Living camera” sensors: Photoreceptive cells across taxonomical groups. (A) Primitive eye (ocellus) with only 30 photoreceptive cells in the larva of a sea squirt, Ciona intestinalis (Ascidiacea, Tunicata). Reprinted from Ref. [19] by permission of © Wiley. (B) Ommatidium of a fruit fly (Drosophila). (B1) Seven photoreceptive cells (R1–R7). (B2) A layer above them. 1 , 2 , and 3 are pigment cells, and “c” are cone cells synthesizing the ommatidium lens (not shown in this section). Reprinted from Refs. [20, 21] by permission of © Elsevier. (C) Grouped retina with photoreceptive cells clustered into retinal cups. (C1) Accessory retina of deep-sea pearleye fish (Scopelarchus michaelsarsi), with rod-like photoreceptors only. (C2) Light-adapted retina of mormyrid fish (Gnathonemus sp.). The tapetal cup shown here is surrounded by 6 others, and includes 22 cones. The tiny bright dots between the cones are rods. Reprinted from Ref. [22] by permission of © Elsevier. (D) Zebrafish larva retina (5 days postfertilization) immunolabeled for UV opsin in cones (green) and an unknown epitope in rods (red). Reprinted from Ref. [23] by permission of © Natl Acad Sci USA. (E) Shortwavelength-sensitive (S) cones in the perifovea of marmoset, upon in situ hybridization with cDNA encoding human S-cone opsin (cells appearing dark). The unstained cells are other types of cones. Reprinted from Ref. [24] by permission of © The Optical Society of America

dismantling the camera). Typically, the optical zoom factor is ca. 3
to 5
but in the so-called mega zoom compact cameras, this factor can increase up to 10, 20, 30, or even 50
. In some of these cameras, an adapter tube can be mounted at the base of the objective so that converter lenses can be placed in front of the zoom lens. Just a few cameras have an objective fitted with a thread so that filters can be mounted. Displays are typically fixed at the back of the camera body, or they can be turned or tilted. Compact cameras are equipped with compur (leaf) shutters operating without any noticeable vibration.

Jo¨rg Piper and Radek Pelc

384

Table 1 Main features of compact, bridge, single lens reflex (SLR), and system cameras Compact cameras

Bridge cameras

SLR cameras

System cameras

Sensor type

CMOS

CMOS, CCD

CMOS, CCD, NMOS

CCD, CMOS

Sensor size

+

++

++++

+++

Noise level

++++

+++

+

++

Light sensitivity

+

++

++++

+++

ISO range

+

++

+++ / ++++

++ / +++

Resolution

+

++

+++

++

Shutter vibration

Low or absent

Low or absent

High

High

Removable lens

No

No

Yes

Yes

Flash shoe

No

Some models

Yes

Some models

Live view

Yes

Yes

Some models

Some models

Remote control

Some models

Some models

Most models

Most models

Video clips

Most models

Most models

Some models

Most models

Weight

+

++/+++

+++/++++

+/++

Figure

3 and 5B

6C

7

8

2.2

Bridge Cameras

As their name suggest, cameras of this type are intended to bridge the gap between compact cameras and those whose objectives can be replaced. The objectives (zoom lenses) of bridge cameras cannot be removed but the sensor size is typically greater (e.g., 2/300 ¼ 8.8
6.6 mm or 100 ¼ 13.2
8.8 mm) than in the compact ones. Thus, the noise is reduced, light sensitivity enhanced and higher ISO values become available. Lower noise levels are particularly useful in low-light situations (polarization, interference, phase-contrast, and fluorescence microscopy). Moreover, the variability of settings is higher than in the compact cameras. For example, images can be saved not only as JPG but also as RAW or TIF files. Some bridge cameras are fitted with a flash shoe so that an external flash can be used instead of a built-in (integrated) one (see Subheading 4.3 for details). This is of advantage when studying motile objects or fast processes. Bigger sensor size translates to a greater diameter of the objective lens and camera weight, typically 300–700 g. Most bridge cameras are also fitted with compur (leaf) shutters comparable to those of compact cameras so that any vibration due to shutter action is irrelevant (negligible). The other characteristics of compact cameras described in the previous section are also relevant for bridge cameras (Table 1).

Photomicrography

385

2.3 Single Lens Reflex Cameras (SLRs)

In a digital SLR (mirror reflex) camera, objective can be easily removed from the camera, thus enabling its easy fitting to a microscope (or telescope, if needed). Camera settings are highly variable and thus compatible with professional tasks. Flash shoe is a standard. The sensor size is much greater than in compact and most bridge cameras, thus yielding very low noise levels and very high light sensitivity, resolution and image fidelity. Advanced photographic system formats (e.g., APS-C, 22.3
14.9 mm) are implemented in many models while the high-end ones are fitted with fullformat sensors (36
24 mm). Focal length, diameter and weight of objectives rise with the sensor size. The camera body weight, typically 500–1000 g is mainly determined by its material. SLR cameras are fitted with slit shutters also referred to as focal-plane shutters. These enable shorter exposure times but bring about more vibrations than the compur (leaf) shutters. Additional vibrations are due to the mirror movement.

2.4

These are named so as they can be disassembled into a system of components. This includes a removable objective. However, as there is no mirror they are smaller and lighter, and thus often referred to as compact system cameras, mirror-less interchangeable-lens cameras or hybrid cameras. The very first one was introduced in 2009 by Olympus as a digital version of the Olympus PEN, the first ever compact film camera made in 1959 and utilizing a half-frame format (18
12 mm). Nowadays, most camera makers offer digital cameras of this category. Instead of having a mirror, a system camera is fitted with two coactive sensors (one of them displaying the image, and the other one capturing it). In most products, slit shutters are used, like in SLRs. The majority of them (e.g., Olympus E and Panasonic G series) comply with the Micro Four Thirds (MFT) standard (17.3
13 mm sensor) but other options are also available (e.g., Sony NEX cameras equipped with APS-C sensors).

3

System Cameras

Fitting a Camera to a Microscope Experienced users often come up with specific solutions for camera–microscope adapters when a lathe is available [7]. Even if the microscope image viewed though the eyepiece is well focused the micrograph may not be sharp, and vice versa. Such problems can be reduced or eliminated by varying the position of the camera or the eyepiece, or of course by manually focusing the camera objective (if present). When the camera sensor is too small only a very small part of the intermediate image enters the eyepiece and contributes to the micrograph. A converter lens placed between the camera objective

386

Jo¨rg Piper and Radek Pelc

and the microscope eyepiece may be useful. Otherwise, a camera fitted with a bigger sensor has to be used. The other extreme is a situation when the sensor is not completely filled by the image, that is, its marginal regions are not utilized. The micrograph thus appears as a (small) circle. In this case the resolving power of the microscope objective may not be fully utilized (depending on the numerical aperture and the final magnification). Additionally, a circular image typically requires cropping to a square or a rectangle. Photomicrographs may be affected by vignetting, especially when a camera is used with its objective on and/or when the sensor is placed beyond the appropriate distance required by the microscope optics. In particular, compact and bridge cameras fitted with the so-called mega zoom lenses may suffer from vignetting. In some cases, this effect can be reduced or eliminated by varying the focal length of the camera objective and/or the position of the camera body. Otherwise, a different camera model should be used. When a sensor is illuminated by oblique incident light, many more artifacts are generated than in a film, especially asymmetric vignetting, blooming, and loss of sharpness [6]. Digital cameras should thus be adjusted for photomicrography with maximum optical and mechanical precision. Compact and bridge cameras can only be connected to the microscope with their (zoom) lens on, which has to be centered directly over the ocular. In most cases, the portion of the image to be acquired can be modified with the aid of the integrated zoom. The position of the camera eyepoint (the front focal plane of the camera objective), the focal length of the camera lens, the lens-tosensor distance, and the minimum achievable distance between the lens and the ocular are not standardized. Many compact and bridge cameras are thus not suitable for photomicrography. The main characteristics of the various types of digital cameras are described below while specifying their advantages and disadvantages pertaining to photomicrography. 3.1 Compact Cameras

These cameras represent a convenient solution when the microscope is of a simple construction, that is, not equipped with a trinocular head. In this case, the eyepiece should be fitted (via a thread) to an adapter tube originally designed to hold converter (wide-field or telephoto) lenses, and serving here as an “interface” between the eyepiece and the camera. This tube is screwed onto the camera providing the latter is fitted with a suitable thread. Low camera weight and compur shutters introducing minimal vibrations represent a bonus here if the optical axis is not vertical (Fig. 3C); blurriness caused by vibration can be avoided. Several cameras of the Canon Powershot series, (e.g., Canon Powershot A95) can be connected with a microscope by use of the Canon adapter tube (Figs. 3A, B and 4F). This tube has to be fitted

Photomicrography

387

Fig. 3 Adjustment of a compact camera for photomicrography (Canon Powershot A95). (A) Adapter tube fitted to the camera. (B, C) Complete assembly fitted to a trinocular head of an upright (B) and stereo (C) microscope. (1) Photo tube of a trinocular head (Leitz/Leica, FSA series). (2) Connecting tube accommodating the photoocular. (3) Photo-ocular 10
(Promicron). (4) Annular adapter 28 $ 55 mm (Promicron). (5) Adapter tube (Canon B-52) for converter lenses

Fig. 4 Components required to couple a camera to a microscope. (A) Photo-ocular for spectacle wearers, Periplan GF 10
/18 (Leitz/Leica). (B) Photo-ocular 10
(Promicron). (C) Annular adapter 28 $ 55 mm (Promicron). (D) The same adapter (5) fitted to a photo-ocular (3) shown in A, and a stepping ring (6). (E) The same adapter fitted to a photo-ocular (4) shown in B (no stepping ring). (F) Adapter tube (Canon LA-DC52D) primarily designed to accommodate converter lenses for telephoto or wide-field photography (converter lens adapter). (1) 28 mm thread. (2) 55 mm thread. (3, 4) Photo-oculars (A, B). (5) Annular adapter shown in C. (6) Stepping ring 55 $ 58 mm (Hama)

to the base of camera’s zoom lens and its other end connected to an annular (ring) adapter (Fig. 4C) which has to be swiveled on it via an outer thread (55 mm). The annular adapter (Promicron, Germany) is also fitted with an inner thread (28 mm) to accommodate a photo-ocular providing the latter is fitted with the same thread (Fig. 4D, E), for example, Leitz/Leica Periplan GF photo-ocular

388

Jo¨rg Piper and Radek Pelc

Fig. 5 Alternative ways of adjusting compact cameras for photomicrography. (A) Adapter for Canon Powershot S80. (1) Adapter tube for converter lenses. (2) Stepping ring 37 $ 52 mm (Hama). (3) Annular adapter 52 $ 28 mm (Promicron). (4) Photo-ocular Periplan GF 10
/18 (Leitz). (B) A simplistic approach: Casio Exilim EX-Z 110 mounted with an external universal “digiscoping” adapter (Meade) connected to a phototube (monocular setup)

10
/18 for spectacle wearers (Fig. 4A) or Promicron photo-ocular 10
(Fig. 4B). The complete photomicrography setup weighs only ca. 340 g (excluding batteries) so that it is compatible with any type of an ocular head (monocular, binocular, trinocular). Variations of this configuration are of course possible (or necessary), and an appropriate stepping ring (e.g., 55 $ 58 mm, Fig. 4D) and connecting tube may additionally be required. Another adapter tube is needed for the Canon Powershot S80 series fitted with a different thread (37 mm) for converter lenses. Therefore, a stepping ring (37 $ 52 mm) is required to accommodate the annular adapter (Promicron, this time with 52 mm outer thread) carrying the photo-ocular (Fig. 5A). When a compact camera cannot be fitted with an adapter tube, simple adapters have to be used, such as the Meade universal “digiscoping” adapter suitable for small compact cameras, as shown in Fig. 5B. This is the cheapest and simplest way to connect a compact digital camera to a microscope, providing the camera entry pupil is adequate. Without batteries, the camera (Casio) weighs only 136 g but as the Meade adapter itself is considerably heavier (350 g) monocular or trinocular (vertical) tubes are preferred. Even the simple cameras integrated in mobile phones can be fitted to a microscope in a similar way. Thus, for instance, an iPhone can be adjusted for photomicrography by use of a special adapter (http://www.skylightscope.com).

Photomicrography

389

Fig. 6 Adjustment of a bridge camera for photomicrography (Olympus Camedia C-7070). (A) Annular adapter 45 $ 52 mm (Promicron). (B) Vario photo-ocular (Leitz/Leica) fitted to the annular adapter (4). (C) Complete assembly. (1) Photo tube of a trinocular head (Leitz/Leica, FSA series). (2) Vario photo-ocular. (3) Ocular zoom ring. (4) Annular adapter; (4a) 45 mm thread. (5) Adapter tube for converter lenses (Olympus). (6) Power supply cable 3.2

Bridge Cameras

Being heavier than the compact ones, bridge cameras should preferably be used with a vertical phototube, that is, with a monocular or trinocular setup. Olympus Camedia C-7070, a typical bridge camera, can be fitted to a microscope in the same way as the above-described Canon Powershot A95, providing suitable adapters are used. A Promicron ring adapter fitted with two different threads (52 and 28 mm), and the photo-oculars described above (Fig. 4A, B) may be used. Alternatively, this camera can also be used with a Leitz/ Leica vario photo-ocular fitted with a 45 mm thread. A different annular adapter (45 $ 52 mm) is thus needed (Fig. 6A). Such adapters can be custom-made, for example, by the Promicron Company. Figure 6B shows the coupling of the vario ocular with the annular adapter. A complete assembly with the camera (including the Olympus adapter tube) is shown in Fig. 6C. A dedicated solution (kit) for fitting digital cameras to a microscope is offered, e.g., by Olympus. For example, Olympus Camedia C-5050 may be coupled to Olympus microscope (e.g., BX-51) via the following three elements: C-3040-ADU adapter (to be fitted to the camera), U-CMAD3 (a C-mount adapter) and U-TV1
2 (a video port with a C-mount, to be fitted to the phototube of trinocular head). Details may be found in Olympus catalogues.

390

Jo¨rg Piper and Radek Pelc

3.3 Single Lens Reflex Cameras (SLRs)

The camera should ideally be used with its objective removed but it is not necessary if high-quality objectives are available. Three options are available: 1. (/) Both the camera objective and the microscope eyepiece are removed. The camera body is fitted to the microscope in a well-defined position so that the intermediate image is directly projected onto the camera sensor. This is acceptable when the intermediate image is fully corrected so that a compensating ocular is unnecessary. 2. (/+) The camera objective is removed while the eyepiece is retained. In this case, the eyepiece acts as a substitute for the camera objective. This option is preferred when the intermediate image is affected by aberrations, especially in older, finitetube-length (i.e., not infinity-corrected) microscopes so that a compensating ocular is needed to obtain high-quality images. It is referred to either as “photo-ocular” or “projection ocular” or “projective.” Although often treated as synonyms their design is somewhat different, and the “projection ocular” was originally designed for microprojection [8]. 3. (+/+) Both the camera objective and the eyepiece are retained. In this assembly, all settings and functions of the camera are preserved, even though only with camera objectives suitable for photomicrography, that is, those that can be mechanically handled, preferably with a constant focal length (f) of medium value such as 35, 50, or 60 mm (at 35 mm equivalent), or zoom lenses with small zoom factors (3
or 5
, for instance). Selection of the best option (1–3) is dictated by the microscope quality, user preferences and the imaging task/modality. Canon EOS series cameras are typical SLRs. All models fitted with an APS-chip can be used in photomicrography without removing their objective so that all available camera settings can be utilized. When a camera equipped with an APS sensor is combined with f ¼ 35 mm (i.e., moderately wide-angle) lens, full field of view of a conventional wide-field ocular (e.g., 10
/18) is captured in the image. When a lens with a focal length of 50 or 60 mm is used, the image occupies ca. 60–70% of the viewing field. In most cases, this is an advantage as the remaining (peripheral) part of the viewing field is often not very sharp anyway. The best results are achieved when Canon camera bodies are combined with Leica-R lenses (ideally Summicron-R 2:2/35 mm, Summilux-R 1:1.4/50 mm or Macro-Elmarit-R 1:2.8/60 mm). These fixed-focal-length lenses (“Leica-R” or LER) can be fitted to any Canon camera body by a suitable bayonet adapter (e.g., Novoflex EOS/LER), and their filter thread can accommodate the above-mentioned annular (ring) adapters (Promicron). The camera body (optionally fitted with an ergonomic [angle] viewfinder) can

Photomicrography

391

B

AirShutter

A

C

Fig. 7 Adjustment of single lens reflex (SLR) cameras for photomicrography. (A) SLR camera (Canon EOS 350D). (1) Phototube of a trinocular head (Leitz/ Leica, FSA series). (2) Connecting tube accommodating the photo-ocular. (3) Photo-ocular for spectacle wearers, Periplan GF 10
/18 with a 28 mm thread (Leitz/Leica). (4) Annular adapter 28 $ 55 mm (Promicron) bolted to a stepping ring 55 $ 58 mm (Hama). (5) Leica Summicron-R 2.0 lens ( f ¼ 35 mm). (6) Bayonet adapter EOS/LER (Novoflex). (7) Remote control cable (Canon RS-60E3). (8) Angle viewfinder with integrated 2.5
loupe (Seagull), combined with an ocular extender (Canon EP-EX15). (B) SLR camera fitted to the front port of an inverted microscope (Nikon Diaphot TMD). (C) Modern SLR cameras enable wireless transfer of digital images to a remote screen or mobile phone, for example, via CASE/AirShutter or CamRanger. This is convenient, for example, if micrographs are to be acquired in a sterile cell-culture box where cabling is undesirable. Image C is adapted from camfere.com

thus be coupled to the Periplan or Promicron photo-ocular or Leitz/Leica vario photo-ocular. An example of such setup is shown in Fig. 7A. Other examples include SLR camera fitted to an inverted microscope (Fig. 7B). Modern SLRs can even transfer images remotedly, to a portable display or a smartphone (Fig. 7C). This is convenient, for example, if the microscope and camera have to be placed in a hypoxic chamber for prolonged periods of time (to study effects of hypoxia on living cells) and cannot be accessed. Unless using flash (see Subheading 4.3 for details), exposure should be shorter than 1/500 s or longer than 1/8 s, especially at higher magnifications. Another way to reduce vibrations is to engage mirror lock-up; in this mode, the mirror flips up well before the shutter opens. Most SLR cameras can be operated by a remote

392

Jo¨rg Piper and Radek Pelc

Fig. 8 System cameras and exposure metering setups. (A) Digital camera (Nikon 1, MFT sensor format) fitted to a microscope via a C-mount adapter. Neither a camera objective nor the photo-ocular is present, that is, the intermediate image is projected directly onto the camera sensor. (B) Analog camera (Nikon FDX-35) complete with an exposure-metering unit (Nikon H-III). (C) Historical analog camera (Leica M) coupled to Leitz Ortholux microscope (modified from Ref. [25]). (1) Exposure meter, Leitz Gossen Microsix-L. (D) Automatic exposuremetering device, Leitz Combiphot. (2) Shutter cable (for the automatic exposuremetering device). (3) Central (axially symmetrical) shutter. (4) Bayonet (Leitz)

or wireless shutter switch (synonym: wireless remote controller) so that the camera itself is not touched by the user when triggering the shutter. Nevertheless, the shutter itself causes additional vibrations which cannot be eliminated by the mirror lock-up. These can be minimized by inserting into the photomicrography assembly a central (axially symmetrical) shutter that used to be a commonplace in older setups (Fig. 8D). When a conventional ocular for spectacle wearers is used it should be fitted with a soft grommet (a rubber ring cushioning the ocular against the photo tube it is inserted in) to reduce vibrations due to shutter action. Some SLR camera models also enable a live-view mode, including on a high-resolution computer monitor (if a connection from camera A/V or DVI/HDMI socket is available). Models fitted with full-frame (36
24 mm) sensors can be used in exactly the same way as analog SLR cameras operating with standard films.

Photomicrography

3.4

System Cameras

3.5 Dedicated Moderate-Cost Photomicrography Cameras

393

These cameras can be adjusted for photomicrography in the same way as the mirror reflex (SLR) ones as their objective can also be detached. Although there are no mirrors, vibrations of the slit or focal-plane shutters typically fitted in these cameras remain an issue. In this sense, the lower weight of these cameras represents an advantage. The sensors are generally smaller than in high-end SLRs. Some system cameras such as Nikon 1 developed in 2011 (Fig. 8A) enable computer-controlled image capture so that vibrations due to the shutter action are eliminated; the camera shutter is permanently inactive in this case. Analog cameras complete with an exposure-metering unit were also made although the unit could usually be separated from the camera (Fig. 8B, C). A number of manufacturers offer dedicated photomicrography cameras at low or moderate cost. Like consumer cameras, they are equipped with CCD or CMOS sensors and must be connected to a computer via USB or Firewire cables; digital live images can be seen directly on a computer monitor. A dedicated image capture software has to be installed, and the computer itself must of course be capable of displaying images and ideally also video sequences. While photomicrographs are not significantly better than those obtained with high-quality standard consumer cameras uncompressed videos can be captured in full resolution (same as that of still images). When inspected on a high-end monitor (e.g., 4 MB ¼ 2560
1600 pixel) the perceived resolution is likely to be much higher than in full HD format (1920
1080 pixel). Figure 9 shows an example of such camera (DFK 72AUC02 from The Imaging Source, Bremen, Germany) fitted with a 5 MP (2592
1944 pixel) CMOS color sensor (ca. 1/2.000 in size). It is connected to a PC via USB 2.0 cable.

Fig. 9 A dedicated moderate-cost photomicrography camera setup. (1) Photoocular, Periplan 10
(Leitz/Leica). (2) Additional ocular, Plo¨ssl 0.5
/12.5 mm acting as a converter lens (Teleskop Service Ransbach GmbH, Ransbach, Germany). (3) Camera; DFK-series, 72AUC02 model fitted with CMOS sensor, 2592
1944 pixel (The Imaging Source). (4) USB cable

394

Jo¨rg Piper and Radek Pelc

3.6 Camera Types by Image Comparison

4

A stained histological section of human eye, showing neuronal tissue of the retina is used here to compare bright-field images obtained with a compact, bridge, single lens reflex (SLR) and dedicated moderate-cost photomicrography camera (Fig. 10). In this particular case, images of satisfactory quality can be obtained with all camera types. Color rendering is reasonable as the image background is almost free from any yellowish taint. The only exception is the dedicated camera offering somewhat poorer color fidelity (Fig. 10D). However, the black-and-white versions of the same images appear to be near-identical in all four cases. However, the above conclusions are unlikely to hold for low-light conditions such as fluorescence or polarization microscopy of specimens with low birefringence; cameras equipped with large image sensors are expected to yield much better results in such cases.

Step-by-Step Photomicrography

4.1

Focusing

Resolution of the camera display should be as high as possible if used for focusing. Ergonomy-wise, a swivel display is preferred when the camera is fitted to the phototube of a trinocular head. When the digital camera is equipped with HDMI-, AV-, or USB-based video outputs, life view can also be obtained on external monitors and images taken by use of capture software. When the camera is used jointly with its own lens the autofocus should be turned off. Usually, the lens should be focused to infinity. In special cases, other settings may be preferred, to improve image quality (e.g., suppression of vignetting). In compact and bridge cameras, images can be focused using the integrated LCD display. In SLR and system cameras, their LCD display can only be used to focus images when the so-called life view mode is engaged. Otherwise, a viewfinder must be used (an angle variant is often convenient) and focusing is further facilitated by a 2.5
loupe integrated in it.

4.2

Exposure

A camera employed in photomicrography should always be operated remotedly to avoid vibrations due to touching the shutter release button. A self-timer (delayed exposure) or a shutter release cable are the simplest solutions. Especially when single lens reflex or system cameras are used, the mirror/shutter itself can cause vibrations. Their effect can be eliminated by using very long or very short exposure times (see Subheading 3.3 for details), or by using flash (see Subheading 4.3 for details). Typically, micrographs are correctly exposed when automatic metering modes of the digital camera are used (full-aperture multispot or center-weighted metering, partial-aperture metering at center, or single-spot metering). If necessary, the automatic exposure setting can be overriden, in most cameras up to 2 EV

Photomicrography

395

50 µm Vitreous body

Internal limiting membrane Optic fibers

Interneurons Internal plexiform layer

ORIGINAL IMAGES

CONVERTED TO B&W

Fig. 10 Comparison of camera performance in bright field. Leica Dialux 20 microscope, planachromatic objective
40/0.65, Leitz/Leica Periplan photo-ocular 10
, halogen light source 50 W. (Top) Anatomy of human retina employed as a testing specimen. (Bottom) Color images were acquired, and then converted to black & white. (A) Compact camera, Casio Exilim EX-Z110 (CCD sensor 6 MP, 1/2.500 ¼ 5.1
3.8 mm). (B) Bridge camera, Olympus Camedia C-7070 (CCD sensor 7.1 MP, 1/1.800 ¼ 7.1
5.3 mm). (C) Single lens reflex (SLR) camera, Canon EOS 350D body (CMOS sensor 8 MP, APS-C type, 1/0.600 ¼ 23
15 mm) fitted with a Leica lens (Summicron 1:2.0, f ¼ 35 mm). (D) Dedicated moderate-cost photomicrography camera, The Imaging Source DFK 72AUC02 (CMOS sensor 5 MP, 1/2.000 ¼ 6.4
4.8 mm)

396

Jo¨rg Piper and Radek Pelc

(exposure values). Automatic exposure bracketing (AEB) may also be useful in this respect; an image sequence is taken automatically at different EV setting. Image brightness can of course be adjusted by varying the intensity of the light source. The ISO speed equivalent should be kept as low as possible to keep noise at minimum. In most cases, ISO speed set to 100 will yield optimal results. 4.3

Use of Flash

Flashlight is of course essential to take micrographs of motile cells. In system and SLR cameras, it also helps reducing vibration artifacts due to shutter/mirror action. To use flash in micrography, the microscope stand should ideally be fitted with a bayonet so that various types of light sources (housings) can be accommodated, such as a mirror housing fitted with a semitransparent mirror (beamsplitter). Serving as an interface between the microscope and the light source, this type of housing can accommodate both flash unit and a standard light source (e.g., halogen bulb), each fitted via a separate bayonet. In this way, specimens can be inspected using light bulb illumination and photographed with flashlight. The “microflash device” made by Ernst Leitz (Wetzlar) is an example of such assembly (Fig. 11). In its reflex housing, an

A

B

C

N 1.4

1.FN50

N4

N 16

4.FN0.4

Fig. 11 “Microflash” device (Leitz/Leica). (A/B) Lamp housing (accommodating a standard light source, 11) detached/attached. (1) Bayonet for connecting the device to the microscope stand of an upright microscope. (2) Bayonet for connecting a lamp housing accommodating a standard light source (11) to the microflash device. (3) Flash clamp. (4) Handle for inserting/removing (5) an additional mirror (reflecting 100%). (6) Semitransparent mirror (reflecting 10%). (7) Collector lens. (8) Capillary flash tube. (9) Cylindrical reflector. (10) Flash cable. (11) Standard light source (bulb). (12) Collector lens. (C) Set of gray filters (Leitz) to regulate flashlight intensity; these are inserted into the microscope condenser filter holder. The drawing (B) is adapted from Ref. [26]

Photomicrography

397

additional, 100%-reflecting plane mirror fitted near the (fixed) semi-transparent mirror can be inserted into the light path. In that case, 100% of standard light source intensity is used for inspection/focusing while the flashlight is blocked. This is convenient at low-light conditions such as darkfield, interference-contrast, or polarization microscopy. On the other hand, when the mirror is removed from the light path, only 10% of the standard light source intensity is utilized for inspection/focusing (usually sufficient in bright field and phase contrast) while nearly 100% of the flashlight can reach the specimen. The flash light intensity can be regulated with a set of several gray filters; these can be combined at will to finely adjust exposure. When an automatic flash is employed the through-the-lens (TTL) exposure metering is used (flash is cut off once enough light has been “accumulated” by the image sensor), which often yields satisfactory results (a TTL flash cable is required). In some cases, the ISO value has to be shifted to achieve good results (e.g., ISO of the flash switched to 400 or 800 when the camera itself is set to ISO 100). When a Canon Speedlite flash (with off-camera shoe TTL cable) is used, the ISO shift is unnecessary. In some cameras, flash can be synchronized even at ultrahigh shutter speeds (e.g., 1/4000 s) so that images are sharp even at the highest magnifications most prone to vibrations. Obviously, high-speed synchronization is recommended to counter the effects of vibrations. 4.4 Monochrome Techniques: Color vs B&W Images

In most cases, the automatic white balance (AWB) mode yields good results although in specific situations, the setting may need to be modified (white balance for artificial light, day light, flash light etc). In specimens stained only in red or blue, image quality can be improved simply by illuminating them with green light, typically by using a green filter as it absorbs red/blue. In addition to increasing image contrast during visual inspection (red/blue structures are rendered black), 50% rather than only 25% of all pixels in the color sensor (Bayer model chip, Fig. 1B) contribute to the image. Otherwise, the signal of the remaining 75% of all pixels is assigned by an automatic interpolation, which reduces contour sharpness. Color fringes lining linear structures and Moire´ effects are additional disadvantages of this design. To counter that, a highly aperiodic pixel motif (X-Trans model) with quasi-randomly assigned colors has been introduced by Fuji (Fig. 1C). Monochromatic light also eliminates chromatic aberration of the microscope itself. Human eye is most sensitive to green light, so that at least subjectively, green images show maximum detail and tonal values.

398

Jo¨rg Piper and Radek Pelc

Especially in monochromatic images (including fluorescence ones) the dominant color can be out of the used color gamut (a complete subset of colors), this being a so-called out-of-gamut phenomenon. The visual information can also be significantly enhanced in these cases when images are taken in (or converted to) black & white [9]. 4.5 Video, Slow Motion, and TimeLapse Imaging

Modern digital cameras, especially compact and bridge ones usually also enable video capture in high-definition (HD, 1280
720 pixels) or full HD (1920
1080 pixels) standard. The video formats vary, but in most models, videos are formatted as AVI, MOV, MPEG4 (MP4) or QuickTime. When they are captured in full HD quality, still frames they are composed of can be extracted as 2 megapixel images. Some modern digital cameras enable video capture and single-shot photography to be carried out quasisimultaneously so that images can be taken at maximum resolution during video capture. In automatic run, videos are well exposed in most cases, and autofocus should be switched off to enable manual focusing. AC adaptors integrated in some cameras (e.g., Olympus Camedia C-7070) are advantageous as they eliminate the need for exchanging batteries, for example, during a lengthy time-lapse imaging. Some types of digital cameras enable high-speed video capture so that fast-moving objects can be tracked. This has been taken to an extreme in Casio EX-F1 bridge camera developed in 2008; video clips with frame rates as high as 300, 600 or 1200 fps (frames per second) can be recorded. When inspected by software operating at the usual frame rate of 30 fps, time stretching of 10
, 20
, and 40
is achieved. Even very fast processes could be quantitatively analyzed at the macro- and microscale [10].

4.6 Image Acquisition and Processing

Fundamental image acquisition parameters should of course be adjusted for optimal image quality, that is, minimum compression (ideally none) and image size. In theory, the image size should be at least 6–8 megapixel in order to make full use of the objective’s resolving power even in images acquired at the lowest magnification-to-NA (numerical aperture) ratio. The aim is to comply with the Nyquist-Shannon criterion (see Fig. 2 in Chapter 14 by Riley et al.). In practice, however, up to 12 megapixel is often advisable for proper rendering of the smallest details. Some of the other parameters may be adjusted as well if needed (brightness and contrast, white balance, color saturation) but caution is due here to enable rigorous image comparison in the future. The sharpness setting should be used with extra care as it may affect rendering of fine structures. Upon image acquisition, contrast would be typically optimized (again) by adjusting image histogram; other-than-linear stretching

Photomicrography

399

(such as “gamma” correction) is to be used with great care to prevent over-optimistic conclusions arising from image comparison. In some cases fine details can be further enhanced when images are converted to black & white (see Subheading 4.4). This is especially so when the RGB color channels can be individually adjusted (e.g., in ImageJ freeware, or the RGB-to-Gray plug-in in Adobe Photoshop). Apart from these standard techniques, two other image-processing options deserve attention: deep-focus image stacking and high dynamic range rendering (HDR/HDRR) or dynamic range increase (DRI). For deep-focus image stacking, a stack of images acquired at different focal planes is needed. The single images are then electronically superimposed with a dedicated software combining together only those image regions that are in focus. Image thus created is mostly free from details that are not sharp, regardless of specimen thickness and objective’s depth of field. Various freeware (e.g., Picolay, CombineZ) and shareware (e.g., Helicon Focus) can be used for this purpose [11, 12]. Both HDR and DRI (the latter referred to as “exposure blending”) are suitable for improving image quality, especially in underand overexposed areas. A set of images of the same viewing field is required at different exposures so that each region of interest is well exposed at least in one image. The resulting image stack has to be processed with HDR or DRI software so that only well exposed image areas contribute to the reconstructed image. Various freeware are available: DRI Tool, Image Stacker, Easy HDR, and Picturenaut. “Photomatix Pro” and “FDR Tools Advanced” are suitable for HDR rendering, and both of them are available as shareware. More details about these techniques and their practical use may be found elsewhere [12, 13]. Finally, “color integrity” and “calibration” are additional factors that have to be taken into account. In all photomicrographs, colors are influenced by a number of factors: the light source, the camera algorithm for image processing, the color gamut used, and the particular type of the monitor used for presentation and/or observation of photomicrographs and/or live-view images. It is thus recommended to use dedicated hardware and software for color calibration. A sophisticated system based on a standardized multicolor calibration slide has recently been reported [14]. In fluorescence microscopy, fading (photo bleaching) is an additional problem leading to a gradual reduction in the emitted light intensity. Additionally, calibration algorithms have been developed so that the negative effects of bleaching can be mitigated or avoided [15]. Of course, digital postprocessing of micrographs may also lead to a ‘pseudo-improvement’ or accentuation of artifacts so that the final image may strongly deviate from the specimen’s original

400

Jo¨rg Piper and Radek Pelc

Reticle

Photocell

Projective

Lamp

Fi lm

Magnification changer (Optovar)

Fig. 12 A touch of history: Photomicroscope I (Zeiss), 1950s. Before reaching the eye, light passes through two semitransparent mirrors (beamsplitters) as light intensity is split between the photocell (to determine exposure time), film, and eyepiece. Adapted from Ref. [2]. Details may be found elsewhere [27]

appearance. This prompted the development of ethical guidelines for digital image processing [16, 17]. A film camera “embedded” in the microscope stand is shown in Fig. 12 as a historical curiosity.

Acknowledgments The authors are grateful to prof. Jan Valenta (Faculty of Mathematics and Physics, Charles University, Prague) for helpful comments. RP acknowledges support via Ministry of Education projects: Chiral Microscopy (LTC17012) and ChemBioDrug.(1) References 1. Overney NL, Overney GT (2011) The history of photomicrography. Normand and Gregor Overney, California. http://www.microscopyuk.org.uk/mag/artmar10/go-no-historyphotomicro.html

1

CZ.02.1.01/0.0/0.0/16_019/0000729

2. Lawson D (1972) Photomicrography. Academic Press, London and New York. ISBN: 0124397507

Photomicrography 3. Inoue´ S, Spring K (1997) Video microscopy: the fundamentals. Plenum Press (Springer), New York. ISBN: 9780306455315 4. Evennett P (2000) The new photomicrography. Proc Roy Microsc Soc 35:253–256. www.quekett.org/wp-content/uploads/ 2015/09/Evennett_digitalcamera.pdf 5. Bockaert V (2003) Sensor and pixel sizes. In: 123 of Digital Imaging (chapter 1). http:// www.123di.com/123di_contents.php 6. Altmann R (2003) The sensor. In: Digital photography and image processing (in German). Midas, Zu¨rich, pp 20–24. ISBN: 3907020642 7. Kinch RJ (online) Making digital camera microscope adapters. http://www.truetex. com/micad.htm 8. Needham GHN (1958) The practical use of the microscope. Charles C. Thomas, Springfield, IL. https://lccn.loc.gov/57010440 9. Piper J (2014) Preparing monochromatic images for publication: theoretical considerations and practical implications. Microscopy Today 22(1):18–24. https://doi.org/10. 1017/S1551929513001132 10. Nachtigall W (2015) High-velocity movements (...). Part 1: High speed registrations of fruit explosions in Impatiens (in German). Mikroskopie 2(2):73–78. https://doi.org/10. 5414/MKX0065 11. Piper J (2008) Use of software to enhance depth of field and improve focus in photomicrography. Microsc Anal 113(May):15–19. https://microscopy-analysis.com/magazine 12. Piper J (2010) Software-based stacking techniques to enhance depth of field and dynamic range in digital photomicrography. In: Hewitson TD, Darby IA (eds) Histology protocols, Methods in molecular biology, vol 611. Springer (Humana Press), New York, pp 193–210. https://doi.org/10.1007/978-160327-345-9_16 13. Piper J (2009) Image processing for the optimization of dynamic range in photomicrography. Microsc Anal 117(Jan):5–9. https:// microscopy-analysis.com/magazine 14. Foster B, Sedgewick J (2014) Color integrity: is what you see what you saw? Microsc Today 22(1):12–16. https://doi.org/10.1017/ S1551929513001272 15. Zwier JM, Van Rooij GJ, Hofstraat JW, Brakenhoff GJ (2004) Image calibration in fluorescence microscopy. J Microsc (Oxford) 216 (1):15–24. https://doi.org/10.1111/j.00222720.2004.01390.x 16. Hayden JE (2000) Digital manipulation in scientific images: some ethical considerations. J Biocommun 27(1):11–19. https://www.ncbi. nlm.nih.gov/pubmed/10916744

401

17. Cromey DW (2010) Avoiding twisted pixels: ethical guidelines for the appropriate use and manipulation of scientific digital images. Sci Eng Ethics 16(4):639–667. https://doi.org/ 10.1007/s11948-010-9201-y 18. Sabesan R, Schmidt BP, Tuten WS, Roorda A (2016) The elementary representation of spatial and color vision in the human retina. Sci Adv 2:e1600797. https://doi.org/10.1126/ sciadv.1600797 19. Horie T, Orii H, Nakagawa M (2005) Structure of ocellus photoreceptors in the ascidian Ciona intestinalis larva as revealed by an antiarrestin antibody. Develop Neurobiol 65 (3):241–250. https://doi.org/10.1002/neu. 20197 20. Hardie RC, Juusola M (2015) Phototransduction in Drosophila. Curr Opin Neurobiol 34:37–45. https://doi.org/10.1016/j.conb. 2015.01.008 21. Larson DE, Liberman Z, Caga RL (2008) Cellular behavior in the developing Drosophila pupal retina. Mechan Develop 125 (3–4):223–232. https://doi.org/10.1016/j. mod.2007.11.007 22. Francke M, Kreysing M, Mack A, Engelmann J, Karl A, Makarov F, Guck J, Kolle M, Wolburg H, Pusch R, von der Emde G, Schuster S, Wagner H-J, Reichenbach A (2014) Grouped retinae and tapetal cups in some Teleostian fish: Occurrence, structure, and function. Progr Retin Eye Res 38:43–69. https://doi.org/10.1016/j.preteyeres.2013. 10.001 23. Alvarez-Delphin K, Morris AC, Snelson CD, Gamse JT, Gupta T, Marlow FL, Mullins MC, Burgess HA, Granato M, Facool JM (2009) Tbx2b is required for ultraviolet photoreceptor cell specification during zebrafish retinal development. Proc Natl Acad Sci U S A 106 (6):2023–2028. https://doi.org/10.1073/ pnas.0809439106 24. Martin PR, Grunert U, Chan TL (2000) Spatial order in short-wavelength-sensitive cone photoreceptors: a comparative study of the primate retina. J Opt Soc Am A 17(3):557–567. https://doi.org/10.1364/JOSAA.17.000557 25. Ernst Leitz GmbH (1968) Microsix-L exposure meter. Factory print 540-21b (German) or 54-22 (English), Wetzlar (Germany). http://microscope.database.free.fr/540.html 26. Ernst Leitz GmbH (1966): Microflash device. Factory print 540-27 (German) or 54-25 (English), Wetzlar (Germany). http://micro scope.database.free.fr/540.html 27. Walker D (2007) A tour around the Zeiss Photomicroscope III. Micscape, issue 145. http://www.microscopy-uk.org.uk/mag/ nov07ind.html

Chapter 14 Digital Micrographs in Pathology Roger S. Riley, Jorge Almenara, and Christine E. Fuller Abstract Digital image capture, storage, searching, retrieval, processing, and manipulation are surveyed. These are detailed by describing how to select suitable digital camera, perform digital photomicrography, and process digital images. Medical applications include education, telepathology aided by automated slide scanners, computer-assisted image analysis of immunohistochemical specimens, and quality assurance and control in histopathology. Image formats and commonly encountered limitations of digital photography and photomicrography are described. Documentation includes listings of suppliers of camera-to-microscope couplers, digital still cameras, slide scanners, and image-processing software. Signal-to-noise (S/N) ratio. Abbreviations of image formats are shown in Tab.2. Key words Digital camera, Image processing, Medical illustration, Photomicrography, Automated slide scanner, Telepathology, Virtual microscopy

Abbreviations ADC CAIA CCD CMOS CMYK DICOM FOV FT-CCD HDRI HSI IT-CCD PACS RGB

Analog–digital converter Computer-aided image analysis Charge-coupled device (image sensor type) Complementary metal oxide semiconductor (an image sensor type) Cyan-magenta-yellow-black (a subtractive color mixing model) Digital imaging and communications in medicine (an image format) Field of view Frame-transfer CCD High dynamic range imaging Hue saturation intensity (a color mixing model for image processing) Interline-transfer CCD Picture archiving and communication system Red-green-blue (an additive color mixing model)

Electronic supplementary material: The online version of this chapter (https://doi.org/10.1007/978-1-07160428-1_14) contains supplementary material, which is available to authorized users. Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1_14, © Springer Science+Business Media LLC 2020

403

404

Roger S. Riley et al.

(D)SLR (Digital) single-lens reflex (camera) S/N Signal-to-noise (ratio) Abbreviations of image formats are shown in Tab.2.

1

Introduction The development of digital imaging in the latter part of the twentieth century resulted in a rapid obsolescence of film photography. This has had a remarkable impact on medical research, education, and patient care, since high-quality digital images can be rapidly and conveniently acquired without the high direct cost and environmental impact associated with photographic films. These digital images can be immediately transferred into print, digital reports, publications, or transferred to other individuals, archived and later retrieved, annotated with descriptive text or audio files, or used in other ways. In fact, the technology for the capture, storage, and retrieval of digital images has become technically advanced to the point that an entire glass microscope slide can be digitized and the images restored on a computer screen to accurately simulate the original experience of visualizing the slide. This technique, referred to as “virtual microscopy” is leading to major changes in biomedical research, education, and patient care [1, 2]. The techniques and applications of digital photography will be explored in this chapter, with an emphasis on applications in the neurosciences. Many reviews of digital photography and imaging have been recently published [3–7].

2

Basic Principles of Digital Imaging The camera is a device to capture electromagnetic energy at wavelengths that corresponds to human vision. The digital camera (“digicam”) consists of a series of optical lenses to capture and focus the electromagnetic energy. Solid-state computer chips, called image sensors, convert the light energy into a digital format. A variety of magnetic and non-magnetic devices store the captured information for display, enhancement, printing, and other applications [3, 8–10]. Portable, self-contained digital cameras have removable storage devices to save the captured images as well as additional computer circuitry to display the images on a selfcontained and/or external video monitor. For permanent storage or additional manipulation, the images can be printed or transferred to a laptop or desktop computer system. Some studio and scientific digital cameras lack internal storage devices and are directly interfaced to an external computer system for camera manipulation and image storage. Scientific and medical photography employs two basic types of digital cameras for different purposes. Digital still cameras capture single images, while digital video cameras acquire multiple

Digital Micrographs

405

contiguous frames. Both digital still and digital video cameras can be used for gross photography, macrophotography, or attached to a microscope for photomicrography. Special modifications of the digital photomicrography camera have been developed for low light microscopy, fluorescence microscopy, and other unique forms of microscopy. Specialized digital cameras termed scanners are used to convert printed material or silver halide film into a digital format. Digital camera technology has rapidly progressed since its commercial introduction in the mid-1990s. Although the early digital cameras produced suboptimal image quality, of the first digital cameras, image quality is not an issue with modern multi-megapixel digital cameras and associated equipment, such as high-resolution printers. In this regard, present digital photography equipment can create prints 1600  2400 or larger that rival the quality of conventional medium format film. In addition, some digital cameras permit storage of the “raw” unprocessed data captured by the camera, so that exposure, contrast, white balance, and other image parameters can be easily modified. 2.1 The Image Sensor

The discoveries of two groups of scientists lead to a silicon semiconductor chip that detects and captures light. Drs. Sangster and Teer of the Philips Research Labs invented a device (BucketBrigade Device, or BBD) in 1969 that transfers electronic charges from one transistor to another. In 1970, two Bell Laboratory Scientists, Willard S. Boyle and George E. Smith, expanded this principle into a proposed data storage apparatus, the Charge Coupled Device (CCD) [11]. In October 2009, Drs. Boyle and Smith were awarded the Nobel Prize in Physics for their efforts [12]. Although the CCD had limited usefulness in data storage, astronomers and other scientists soon recognized an unanticipated feature of the CCD: its great sensitivity to light. CCDs were incorporated into television cameras and flatbed scanners in the mid-1970s, into scientific, industrial, and military cameras in the 1980s, and into mass-produced, consumer-oriented digital cameras in the mid-1990s. The enormous commercial success of the digital camera has motivated rapid improvements in technology, leading to progressively improved image quality. Presently, digital cameras utilize the CCD, as well as a later derivative, the complementary metal–oxide–semiconductor (CMOS) sensor.

2.2 CCD and CMOS Image Sensors

Image sensors are placed at the focal plane of the lens system of a digital camera. The CCD consists of separate image sensing and storage sections. The image sensor is comprised of a thin piece of semiconductor material (i.e., silicon wafer) coated with an array of microscopic (4–30 μm) light-sensitive metal–oxide–semiconductor capacitors (MOS capacitors) termed photodiodes or photosites [8, 13, 14]. Known as photoelectric conversion, this process converts the light information into digital format. During the exposure phase, photons of light falling on a CCD generate electric charges

406

Roger S. Riley et al.

Microlens

Silicon substrate

Light

Potential well Photodiode

Fig. 1 Anatomy (cross section) of a digital camera chip (left), and a digital image capture and storage flowchart (right). Reproduced from Ref. [177] by permission of © Wiley

in a cumulative manner (photoelectric effect), such that the voltage generated is proportional to the number of photons (Fig. 1). The analog voltage signals from each photodiode are temporarily stored until the exposure is complete, and during the readout phase they are consecutively transferred to a readout register, output amplifier, and finally to an analog-to-digital converter (ADC) chip. The ADC chip converts voltage variations (i.e., brightness) into discrete binary numbers [13] (Fig. 1). CCDs can be subclassified into full-frame, frame-transfer, interline, and linear types, depending on the relative arrangement of the image sensor and temporary storage areas [8, 9]. In a full-frame CCD, the entire CCD array is exposed to light, and the acquired electrical charge is sequentially shifted across the array to a serial register. A mechanical shutter is used to block further light exposure during the readout phase. These sensors are relatively slow but commonly used in high-end digital cameras because of their high capture density. The frame-transfer CCDs (FT-CCDs) are composed of a dedicated image area and a separate masked storage area. The data from the acquired image shifts to the storage area in a few milliseconds and can be used for another exposure. Although a mechanical shutter is not required, some reduction in image quality termed “image smearing” may occur from light exposure during readout. In spite of this disadvantage, these sensors have high frame rates

Digital Micrographs

407

and exposure efficiency, and are commonly used in high-quality video cameras. The interline-transfer CCD (IT-CCD) is a variation of the frame transfer CCD with interlaced, alternating rows of photosensitive and masked storage arrays. Interline transfer CCDs are commonly used for digital microscopy due to their speed and light sensitivity [8]. Linear CCDs consist of a single row of pixels that is scanned across the plane of an image by a stepper motor. Devices with this technology are slow and best used for stationary objects but can produce high-resolution images and are primarily used in scanners and digital camera scanner backs. CMOS sensors are modifications of a type of semiconductor widely used in the computer industry for CPUs, memory modules, and other computer components. The CMOS sensor, unlike the CCD, directly incorporates circuitry for charge-to-voltage conversion, signal amplification, noise-correction, digitization, and/or other features (very large scale integration, VLSI). Since they can be produced in existing fabricating facilities that produce other computer semiconductor components, CMOS image sensors also have lower production costs than CCDs. In theory, VLSI leads to smaller, more reliable, less expensive cameras with low power consumption, faster frame rates, few artifacts, and the ability to include “higher-level” camera functions such as image stabilization and wireless control. Unfortunately, the presence of additional circuitry on the CMOS image sensor also leads to a smaller proportion of the photosite that is sensitive to light and therefore to lower light sensitivity. The “fill factor” for a CMOS image sensor is on the order of 30%, while CCDs typically have fill factors >90%. Increasing the size of the CMOS pixel partially compensates for the lower fill factor. However, modern CMOS-based high-resolution digital cameras produce images equivalent to those of high-resolution CCD cameras. Silicon photodiodes capture only the brightness of focused light and are insensitive to color. Therefore, the native image from a CCD represents the distribution of electrons on the chip in the form of a gray scale that varies from pure white to pure black. Presently, red, green, and blue optical filters are used to transform the gray scale information into a color image. The number, type, and arrangement of photosites and color filters are critical features that establish the applications of the CCD. The area array and the linear array are the two most common types of arrangement of the photosites and color filters. The area array is most widely used in digital cameras for both still images and video. These digital cameras have image sensors with photodiodes arranged in a rectangular grid, and light from the entire scene reaches the sensor at the same time. The cost and image quality of these cameras depends on the number of image sensors, and the arrangement of the color filters. “One-chip, one-shot”

408

Roger S. Riley et al.

digital cameras are the most common and least expensive. They have a single image sensor with minute red, green, or blue filters fixed over the individual photosites in a predefined pattern [14]. Each photosite captures the brightness of its assigned color during an exposure, and software interpolation is subsequently used to calculate the two colors that the sensor did not record from the color information of surrounding pixels. Inaccurate computer interpolation can make these cameras prone to an artifact termed color aliasing, but progressive advances in software technology have largely resolved this issue in the more expensive digital cameras. “One-chip, three-shot” digital cameras have a liquidcrystal tunable filter or four-position rotating color wheel placed over the image sensor to obtain color information. Generally, a neutral filter position is used for composition and focusing. These cameras can have excellent resolution and color fidelity at a relatively low price, but can image only stationary objects in color. “Three-chip” or “three-CCD” digital cameras have separate image sensors optimized for red, green, or blue light and use a beam-splitting filter to simultaneously generate three copies of an incoming scene [14]. These cameras can image moving subjects and have the highest image resolution and most accurate color fidelity, but can be bulky and expensive. Linear arrays (“sticks”) of photosensors are used in scanners, high-end studio cameras for still photography, and in some specialized digital cameras for photomicrography. Many scanning cameras use a single array of photosensors to sequentially scan the scene only one row at a time with red, green, and blue filters. The more expensive scanning cameras have three rows of sensors, each covered by the same color scheme. Scanning digital cameras are capable of very high resolution and color fidelity, given that 10,000 or more pixels may be present in each array. Unfortunately, they are relatively slow and useful for only constantly illuminated and motionless subjects. A popular variant of the scanning digital camera, the “moving one chip camera” uses a one-chip, one-shot image sensor, which moves in subpixel range in both X and Y directions to increase the resolution. Although the final image depends upon computer interpolation, the resolution can still be very high. Several high-quality photomicrography digital cameras, including the Nikon DXM-1200 (Nikon Corporation, Tokyo, Japan) and the Carl Zeiss Axiocam (Carl Zeiss, Gottingen, Germany) use moving one-chip mechanisms. 2.3 Advantages and Disadvantages of Image Sensors

The technical advantages and disadvantages of the digital image sensor over conventional photographic film has been a matter of contention since the conception of the digital camera. Essentially, the major advantages of the image sensor include light sensitivity, dynamic range, and linearity, while the major disadvantages include image resolution and the tendency to generate digital artifacts. The

Digital Micrographs

409

light sensitivity, or quantum efficiency, of an image sensor is the fraction of incident photons that are captured and converted into an electronic signal [14]. The dynamic range is usually specified as the ratio of the maximum number of electrons accumulated before the signal is saturated (i.e., full well capacity, FWC) to the readout noise [14]. The quantum efficiency of most high-quality CCDs is approximately 40–70%, although some special CCDs for scientific applications have quantum efficiencies as high as 90%. This compares to a quantum efficiency of approximately 20% for a conventional video camera, 3% for the human eye, and less than 3% for a film camera. Digital camera manufacturers utilize the International Organization for Standardization (IOS or more commonly ISO) filmequivalent number to rate the light sensitivity of image sensors in consumer and professional digital cameras. ISO numbers for image sensors used in existing commercial digital cameras vary from 50 to 25,600, while film stock ISO varies from 6 to 3200. The more sensitive an image sensor is to light, the higher is the ISO number. All digital cameras have controls to automatically or manually adjust the light sensitivity, so that the sensitivity of the image sensor can be increased in low light conditions. Unfortunately, increased sensitivity is usually achieved by amplifying the signal from the image sensor, which increases the amount of noise and decreases the quality of the final image. Since the acquisition of high-quality images in dim light is a much desired feature of photographers, digital camera manufacturers have devoted much attention to improving the usable ISO by modifications in the electronic and computer processing of the image sensor output. The light sensitivity of the silicon photodiode extends across the visible wavelengths and into the near infrared, making infrared photography possible. Pixel binning is often used to increase the sensitivity of a CCD in low incident light situations, and improves the signal to noise (S/N) ratio [8, 14]. Binning is accomplished by combining the charge from several adjacent pixels into a “superpixel.” The charge from the selected pixels is accumulated until the exposure is complete; readout then occurs. Binning results in decreased resolution for increased sensitivity in situations of low incident light where a usable image could not otherwise be obtained. For example, 3  3 binning results in a ninefold improvement in the signal intensity and a ninefold reduction in noise over single pixel capture [14]. Dedicated photomicrography cameras, especially those designed for low-light situations, such as fluorescence microscopy, usually have user selectable binning options. The dynamic range (i.e., intensity resolution) of an image sensor is the range of light level to which it is sensitive, or the ratio of the most intense signal to the smallest resolvable signal that can be generated by a digital camera [8, 14]. Under ideal

410

Roger S. Riley et al.

conditions, the dynamic range of current high-quality CCDs is theoretically 106, or an order of magnitude of 20,000:1. However, the dynamic range of most consumer digital cameras is much less, on the order of 450:1 to 500:1. For comparison, the potential dynamic range of the human eye is about 109, and its dynamic range is about 30,000:1 under normal conditions. The dynamic range of photographic film is approximately 103, or 7.5 magnitudes. The dynamic range of an image sensor is usually expressed in practical terms as a gray scale resolution (i.e., 8-bit, 12-bit, 16-bit, 24-bit, etc.). The dynamic range is one of the most critical features of an image sensor for scientific applications, such as for quantitative image analysis, where the accurate measurement of very small differences in light levels is essential. Fortunately, the dynamic range and sensitivity of most image sensors, unlike photographic film, is linear across the visible wavelength of light and range of light intensities. Pixel dimension refers to the size of each pixel comprising the CCD array. For optimal image quality in digital photomicrography, each pixel comprising a CCD array should be between 6.5 and 7.5 μm in dimension. However, CCDs with pixel dimensions of 9–10 μm provide adequate image quality with optical magnifications above 20, plus their field of view is larger than that of a CCD array with smaller pixels. For example, Hand calculated that a 1.4 megapixel imager comprised of 6.5 μm pixels samples roughly half the field of view of an array with 9.5 μm pixels [15]. The major disadvantages of several image sensors are their smaller physical size relative to photographic film, and their tendency to generate noise and other electronic artifacts. Image sensor sizes are presently limited by their cost, manufacturing capability, and the availability of electronic and computer technology to read and process the millions of electronic pulses generated during an exposure. However, at the time of this writing, digital cameras with 35 mm-sized image sensors are available with more than 24 million pixels (6048  4032 pixels), while the most advanced, commercially available medium format-sized digital camera has an 80.64  80.64 mm image sensor with 85 million pixels (85 megapixel, 85 Mpixel, 9216  9216 pixels). Unfortunately, the resolution of the CCD cannot be directly compared to that of photographic film, since CCDs are composed of discrete pixels, while photographic film consists of a light-sensitive dye emulsion. The resolution of the photographic film is expressed in lines per mm or per inch. Fine grain photographic film has a resolution in excess of 4000 lines per inch, while the human eye can discriminate about 1600 scan lines [16]. Practically, images produced by 35 mm or medium-format cameras with resolutions about 25 Mpixel are equivalent to fine grain photographic film. In addition, to size, the image quality of both photographic film and digital image sensors is affected by other factors. For example, grain affects the quality of

Digital Micrographs

411

film photography, while electronic image sensors generate noise and other artifacts that can impair image quality. Table 1 summarizes the most important instrumentation-related problems in digital photography. Electronic “background” noise generated by an image sensor includes thermal noise, photon noise, and fixed pattern noise. Thermal noise (dark current) is the most important, and is thermally dependent upon the electronic noise generated by an image sensor, even in the absence of light [8, 14]. Dark current noise appears as “hot pixels” or white dots in images photographed at room temperature with a long exposure time. This is the primary limiting factor for the image sensor dynamic in low light conditions. Dark current effects double in intensity for each 10  F increase in ambient temperature and can be reduced by cooling the image sensor. “Cooled” digital cameras are essential for photography in dim light conditions, such as astronomy and certain forms of photomicrography, such as darkfield and fluorescence microscopy. Liquid nitrogen is used by some astronomers to cool their CCDs for long exposures, but Peltier-cooled digital cameras are more practical for biological sciences. Peltier coolers are electrically driven heat pumps, thermal cyclers, or thermoelectric coolers that consist of paired semiconductors sandwiched between two ceramic plates. The passage of current through the semiconductor results in heat energy from one side of the Peltier cooler moving to the other, creating a temperature differential of up to 120 . These devices are based on a principle discovered in 1834 by French physicist Jean Charles Athanase Peltier, who found that passing current along a circuit containing dissimilar materials results in a refrigerating effect. Exposure times of up to 30 s are usually possible if the CCD is cooled to 0  C. Digital cameras equipped with image intensifiers (“intensified digital cameras”) are also available for low-light imaging. Noise that develops during amplification and digitization of the analog signal include readout amplifier noise, reset noise, I/f noise, and quantization noise. The amount of electronic noise is expressed by the “signal-to-noise” (S/N) ratio, which varies from about 200:1 (46 dB) in a typical video camera to >100,000:1 (100 dB) in a very high-resolution digital camera. Other than dark current, noise generated by amplification of the stored charge on each photodiode into analog voltage (i.e., “readout noise”) is the second major source of electronic noise [14]. Readout noise has been substantially reduced by recent advances in CCD design and manufacture, but is still a consideration in high-resolution scientific photography. Readout noise increases in proportion to readout speed and is lowest in slow-scan digital cameras. Bad pixels are photodiodes that return incorrect signals. Image sensor fabrication is not perfect, and the presence of a few flawed pixels among the millions of photodetectors comprising a modern

412

Roger S. Riley et al.

Table 1 Limitations in digital photography (reprinted from author’s own paper [177]) Problem

Cause and characteristics

Possible solutions

Bad pixels

Photodetectors that produce Software interpolation of images or inaccurate data. May appear as subtraction of a “dark frame.” consistent bright or dark spots in Some newer digital cameras an image. The most common form include software to “map out” of bad pixel is the “stuck pixel” or bad pixels “hot pixel,” which appears as a fixed colored spots in images recorded at long exposure times

Blooming (“Light spillover”) (“Streaking”)

Modern image sensors incorporate Results when the charge in a anti-blooming gates and overflow photosite exceeds its storage wells to eliminate excess capacity (“oversaturation”) and electrons. Sensitivity and dynamic overflows to an adjacent photosite. range may be reduced by antiAppears as bright vertical streaks, blooming measures white halos, or spots in an image with extreme exposure values

Color aliasing artifacts (“Christmas tree lights”)

Software correction may be possible Inaccurate averaging between adjacent pixels. Most apparent in enlarged images and diagonal lines

Fringing

Software correction using a flat field Wavelength-dependent, wave-like with a wavelength corresponding patterns in a digital image due to closely to the image reflections of incident light in the CCD or associated filters. Most obvious when the incident light source contains a strong component at a single wavelength. A particular problem in astronomy

Non-linearity

Exceeding the linear range (“oversaturating”) a photodiode so that a further increase in light intensity does not cause a corresponding increase in signal intensity

Decrease light intensity or combine multiple short exposure into a single image

Pixelation

Blocks of color most apparent as jagged diagonal lines in an image. Caused by overcompression of an image with a “lossy” technique, overenlarging an image, or capturing a detailed image with a low-resolution CCD

Using a camera with a higher resolution CCD. Reduce the level of compression or capture the image in TIFF or CCD RAW format

Pixel sensitivity

Small variations in light sensitivity of Calibration of the image sensor with an image of an evenly illuminated individual photodiodes due to light source (flat field) manufacturing imperfections. Significant in astronomy or (continued)

Digital Micrographs

413

Table 1 (continued) Problem

Cause and characteristics

Possible solutions

scientific applications where quantitative data is acquired Readout signal noise and bias

Electronic noise generated during amplification of the photodiode signal into analog voltage. Bias is an offset, false signal generated during signal amplification

Design of CCDs and amplification circuitry. Bias is removed with data from bias strips and bias frames

Stepping

Lines of different densities and colors, particularly in shadowed regions of scanned images. May result from light source variations or electronic noise

Upgrade scanner

Thermal noise

Random, thermally dependent noise Some digital cameras utilize electronic dark noise subtraction produced by image sensors in the to minimize the problem. Image absence of light. Principal sensor cooling by forced air, limitation of dynamic range thermoelectric, or cryogenic in obsolete or very sensitive digital techniques is required to cameras. Appears as random white minimize thermal noise in digital dots in images, most visible on cameras for many scientific images taken at long exposures applications. Image sensor design can reduce problem

Vignetting (see Table 4 for details)

Dimming of objects at the edge of an “Zoom-in” with the digital camera (with a reduction in the field of image due to mechanical view); find an appropriate interference. A particular problem camera/microscope adapter with digital photomicroscopy utilizing fixed lens digital cameras

image sensor is not unusual. Unfortunately, since CCD data is read in a “bucket brigade” style, one bad pixel can compromise an entire row or column. Bad pixels are the most obvious during low-light, high ISO exposures. Digital image processing software can sometimes recognize a bad pixel and replace the data by interpolation from the surrounding photosensors. Image sensors are graded for quality based on the number of bad pixels and other features. Blooming is the buildup of a pixel charge beyond the full well capacity, resulting in corruption of the surrounding pixels by the excess charge [14]. Anti-blooming mechanisms are an inherent feature of the CCD. Imperfections in CCD fabrication also result in small, wavelength-dependent, random variations in the sensitivity of the individual photodiodes comprising an [entire] array. The variation in pixel sensitivity is inconsequential for most digital photography applications but can be calibrated for special circumstances where

414

Roger S. Riley et al.

precise, flat field imaging is required. This calibration is performed by imaging a uniformly illuminated source while measuring the variation in individual pixel readout. “Fringes” are wavelengthdependent, wave-like patterns in a digital image caused by multiple reflections of the incident light in the CCD or associated filters. Fringing is obvious when the incident light contains a strong, single-wavelength component, and can be corrected with a flat field illuminated under the same wavelength conditions. 2.4 Fundamental Characteristics of Digital Images

The brightness and color information recorded by a digital camera or scanner is used by the computer or printer to recreate an image of the original scene. The data recorded by each photocell comprises a minute portion of the final image that appears as a tiny square termed a picture element or “pixel.” The process of creating a grid of pixels is termed “bit mapping,” because each pixel has a numbered address, and is stored in an area of memory called a bit-map. Therefore, digital (raster) images are commonly referred to as “bit-maps.” The number and size of the pixels (spatial resolution), the amount of brightness information (color depth, pixel depth, brightness resolution), and the manner in which the image is displayed or printed largely determines how closely a bit-mapped image resembles the original scene. The term “resolution” is widely used to indicate the number of photocells comprising an image sensor and the number of pixels in a raster image, but “resolution” properly refers to the ability of a lens to resolve close lines on a test chart. Precisely, the optical resolution of a digital camera or scanner refers to the actual number of photosites, while the “interpolated resolution” indicates the actual number of photosites plus those generated through interpolation. Most consumer digital cameras have 16–24 megapixels, and the maximum resolution of professional digital SLR cameras is currently 61 megapixels. In comparison, the resolution of 35 mm color slide film is estimated at 10–20 megapixels, while the human eye has a resolution of approximately 120 megapixels. The “resolution” of a bit-map image is commonly expressed as the total number of pixels (i.e., 1,920,000 pixels or 1.92 megapixels), or by its dimensions in pixels (i.e., 1600  1200 pixels), where the first number is the number of pixels across the screen (columns) and the second number is the number of pixels down the screen (rows). The resolution of printers is usually stated in dots per inch (dpi), which vary from 300 dpi to 2400 dpi. A similar term, pixels per inch (ppi), is also applied to the resolution of computer monitors. Presently, most monitors display images at the relatively low resolution of 72 ppi. Most image sensors have a rectangular array of pixels, but the relative dimensions of the array can vary between image sensors. The aspect ratio of an image is the ratio of its height to its depth. Although a 1.33:1 aspect ratio is common for image sensors and

Digital Micrographs

415

display monitors at common resolutions (640  480 pixels, 800  600 pixels, 1024  768 pixels), 35 mm film and 400  600 photo paper have 1.5:1 aspect ratios, and 8  10 in. photo paper has an aspect ratio of 1.29:1. Practically, this indicates that digitally acquired images may have to be adjusted (cropped) for different purposes. Some digital photomicrography cameras feature the ability of a user-selectable aspect ratio. The cones of the human eye are sensitive to red, green, and blue colors, and all other colors are recognized as additive mixtures of those three primary colors. Transmissive electronic display devices, such as computers and television, use the RGB color model, which emits mixtures of red, green, and blue light. The RGB color model is additive, such that equal amounts of all colors produce white. Theoretically, any color in the visible spectrum can be produced by merging the proper amounts of red, green, or blue, but electronic devices are unfortunately capable of displaying only a limited portion, or gamut, of the visible spectrum [17]. The interrelationship between colors in a color model is visualized by mapping a three dimensional “color space” that contains all possible colors. Color spaces were first defined by the Commission International de l’E´clairage (CIE) in 1931 to mathematically describe all colors visible to the human eye [18]. Several subsets of the RGB color model called “working color spaces” were devised for specific purposes, including image capture and editing [3, 18]. The sRGB IEC61966-2.1 color space most closely simulates the gamut of most electronic display devices and is widely used in so-called “prosumer” (professional consumer) digital cameras and images prepared for display on the Internet. The Adobe RGB 1998 color space has a wider gamut meant to encompass the colors attainable on a color printer, and is often used during professional image capture and color manipulation to retain a more accurate representation of the original scene. The RGB color model is device independent, but the appearance of a specific RGB value may vary between different devices or even in the same device over time. Due to these variations, accurate color display from a given display device requires periodic calibration against a standard color chart. The result of the calibration is a “color profile” for that specific device at the time of calibration. Unlike electronic displays, which utilize the additive RGB color model, color printers are subtractive devices based on the CMYK (cyan-magenta-yellow-black) color model. Computer algorithms termed “color management modules” perform calculations to convert the colors of an image from one color space to another [18]. The interconversion of RGB displays images to CMYK for print publication and can lead to changes in hue and saturation. This is because the CMYK color space has a slightly smaller gamut of colors than the RGB color space, and RGB hue values may be slightly different than those in CMYK [19]. This problem is most

416

Roger S. Riley et al.

notable for fluorescence images, which have a “weak and sickly” appearance when the additive color scheme of the RGB images is converted into the subtractive CMYK color Scheme [19]. Durcan and Hinchcliffe have described appropriate means to avoid this problem when multicolor fluorescent images are needed [19]. The CIE L∗a∗b∗ color system was reported to reproduce the color of H&E-stained and Papanicolaou-stained tissue better than the RGB or HIS models [20]. Digital cameras record brightness information about the red, green, and blue colors of a scene in three channels. Another major determinant of image quality is the amount of color information obtained. The color depth refers to the number of computer bits used to record information about each color, while tonal balance refers to the brightness dynamics of an image [21]. Eight bits per color (8 bits red, 8 bits green, and 8 bits blue; 24-bit color) is the minimum color depth for adequate color reproduction and permits the display of 256 shades/color, or 16.7 million total color values (2563). Each channel (i.e., red, green, and blue) of an 8-bit color digital image consists of multiple gray spots that can vary in 255 levels from black (level 0) to white (level 255). However, recently developed consumer digital cameras capture images with 10 bits/color (1024 levels/RGB channel, 30-bit color), while professional digital cameras and scanners commonly achieve either 12 bits/color (4096/RGB channel, 36-bit color) or 14 bits/color (16,384 levels/RGB channel, 42-bit color). Unfortunately, the extra color information is usually downsized to 8-bits/color for display, since the human eye is theoretically incapable of differentiating more than 16.7 million colors. Several mathematical algorithms (image formats) are available to store image sensor data in digital form [19] (Table 2). Since each format has advantages and disadvantages, most digital cameras offer several user-selectable formats (i.e., JPEG, JPEG 2000, TIFF, CCD RAW, etc.). The Joint Photographic Experts Group (JPEG) format is universally applied in digital photography, including the storage of 24-bit color photographic images on the web. JPEG uses a “lossy” compression scheme. Although lossy compression selectively removes information from the file, the user can select the degree of compression. Low to intermediate compression of an original image produces dramatic reductions in file size with very little degradation of image quality. A more recent iteration of JPEG, termed JPEG2000 permits improved compression performance and introduces fewer artifacts. However not all image software will recognize the files [22, 23]. The Tagged Image File Format (TIFF) and Portable Network Graphics (PNG) formats were originally developed for image transfer in the graphic arts, and are widely used today in desktop publishing. There are a variety of TIFF formats, some of which use “lossless” image compression, which does not compress the image

Digital Micrographs

417

Table 2 Features of common image storage formats (adapted from author’s own paper [177]) Format

Primary application

Advantages

Disadvantages

BMP Windows Bit map

Raster graphics format for Windows

Very simple and not File size very big, as under patent no compression is protection, i.e., applied compatible with a vast majority of graphical software

EPS Encapsulated PostScript

Best file format for placing color images in page layout documents. Files include a low resolution preview image for screen display and image data written in Postscript

Rapid screen rendering Large files may be slow to print and high resolution printed output on PostScript compatible printers. Selectable preview display (none, 1-bit/8-bit TIFF, 1-bit/8-bit JPEG) and encoding (ASCII, binary, JPEG). PostScript color management

EXIF Exchangeable Image File

Format utilized by digital cameras to store image and camera-specific information in image file headers. Developed by JEITA (Japan Electronics and Information Technology Industries Association) Technical Standardization Committee

Flexible format, can record exposure information (shutter speed, aperture), the time and date the image was taken, and other information. Information is stored in file header and transferred with file. The new EXIF 2.2 standard includes information for printers to enable them to perform accurate image adjustment

Specific application required to read EXIF data

FlashPix

Panoramic and “zoomable” images

Rapid display of zoomable images

Image display requires FlashPix format plug-in and Microsoft OLE (continued)

418

Roger S. Riley et al.

Table 2 (continued) Format

Primary application

Advantages

Disadvantages

GIF Graphics Interchange Format

Common online storage format for images with transparency effect (graphic type images, animated sequences) and small images with limited color data

Support for alpha mask channel and transparency

Limited to 8-bit (256 color) images.

JPEG Joint Photographic Experts Group JPEG2000

“Lossy” Efficient image Universal image compression compression format storage format in method removes for digital images, digital cameras and “extra” image data level of compression is graphic software and can cause user-selectable. An 8industry. A image degradation to 10-fold file modified JPEG and artifacts, compression is format especially if an typically used without (JPEG2000) was image is visible degradation of recently developed recompressed. image quality by the Digital Detailed images Imaging Group do not compress (DIG) effectively

PDF Portable Document Format

Popular, flexible format for electronic publishing and prepress

PhotoCD

Proprietary process of Images can be opened to Only specialized facilities with a multiple sizes. A Kodak Corporation photo CD transfer printed index sheet to place digitized station can with thumbnails of files of photographs produce the images is included onto a CD-ROM. PhotoCD Images are adjusted CD-ROMs. for color and Images cannot be compressed to opened by every 4.5 Mb graphics program

Specialized software Independent of fonts is required to write and computer PDF files operating system, relatively small file sizes, high output resolution. Files can be opened and viewed with a freeware application (Adobe Acrobat Reader). Various levels of protection are available

(continued)

Digital Micrographs

419

Table 2 (continued) Format

Primary application

Advantages

Disadvantages

PICT

Standard Macintosh image file format. Utilizes lossless Run Length Encoding

Efficient file compression without image quality degradation

Pixel size limitations, limited application in non-Macintosh computers. Best for images with large, flat areas of color

PSD Photoshop Document format

Native format of Photoshop

Image layer can be easily Very large file size added or removed

PNG Portable Network Graphics

Not widely utilized Efficient image Recently developed at present. Some compression, support format for online graphic for alpha mask images storage and applications and channels, display older web transparency, and browsers do not 32-bit images support PNG format

RAW (CCD-RAW)

File format utilized by Files are stored at high RAW formats are specific for each resolution (10- or professional digital digital camera 12-bits), contain cameras to store manufacturer. more information original, Software than JPEG or TIFF uncompressed file (“acquire files, and are relatively information with module”) must be small. Artifacts no in-camera available to open produced by processing (“digital files. Image in-camera processing negative”) postprocessing is or image compression required are avoided

TIFF Tagged Image File Format

Relatively large Submission of images Flexible, industryimage file size. standard image format to labs and service Byte order differs read by most bureaus. Most with computer operating systems. digital cameras operating systems Support for alpha utilize TIFF as the channels and paths. format for storing “Lossless” uncompressed compression (LZW (“lossless”) image compression) may be used

or cause any image degradation. The file sizes of uncompressed TIFF images are relatively large in comparison to those produced by the JPEG scheme, but the LZW TIFF compression algorithm can considerably reduce the size of some images. A few digital cameras permit the storage of image data in the raw, unprocessed

420

Roger S. Riley et al.

form (CCD RAW, CRW) for later software modification in a desktop computer. The unprocessed format produces relatively small files, decreases the time required to process and store images, and provides the option of later image software modifications. In addition, the original image can be reprocessed at a future time to incorporate the latest advances in software technology. Generally, the original image should be stored in an uncompressed “lossless” format, while images for display, Internet broadcasting, or transfer and should be converted to “lossy” format. Furthermore, images in “lossy” formats should not undergo further manipulation, given that information is lost with every file open/save operation. As of yet, biomedical researchers and pathologists have not widely embraced other professional image storage and retrieval formats, including the Digital Imaging and Communications in Medicine format (DICOM). DICOM, developed through the collaboration of the American College of Radiology (ACR) with medical industry companies and the National Electrical Manufacturers Association (NEMA), is the current standard for radiographic images [24, 25]. Since most healthcare institutions utilize existing Picture Archiving and Communication Systems (PACS) for DICOM, there is vast interest in integrating images from anatomic pathology into this system [26, 27]. 2.5 Digital Still Cameras

A scientist interested in digital photography has the option of either adapting a commercially existing digital still or video camera for their specific requirements, or acquiring a digital camera especially manufactured for a specific application. Commercial prosumer, or point-and-shoot, digital still cameras have non-interchangeable, integral lenses, while advanced amateur and professional single lens reflex (SLR) incorporate removable, interchangeable lenses. The major advantages of the commercial prosumer camera models are their versatility and relatively low cost. These cameras can be adapted to existing microscope systems for photomicrography, or easily removed from the microscope for macrophotography or other purposes [28, 29]. The image resolution of integral lens digital cameras is generally 16–24 Mpixel, and the cost of the cameras is generally less than a $1000. “Prosumer” or “Professional-grade” single-lens reflex (DSLR) digital still cameras can be retrofitted for digital photography. These provide much higher image quality results than the integral-lens prosumer digital cameras because of the CCDs larger physical side and the more sophisticated image processing circuitry. These cameras are mounted to a microscope with a “T-mount,” and can be easily removed from the microscope for high-quality macrophotography, copy work, or other applications. Furthermore, these cameras utilize existing lenses, flash units, and other SLR camera accessories that may already be available in the laboratory. The newest SLR digital cameras offer image resolutions of 18–61

Digital Micrographs

421

megapixels and provide a live “through-the-lens” image for composition and focusing through the optical viewfinder and the LCD screen. A low power (2 or 4) eyepiece magnifier or auxiliary focus finder is still recommended with these cameras to reduce focusing errors at low magnification. The multiplier effect arises because many SLR-type digital cameras utilize preexisting 35 mm camera bodies with a CCD mounted at the film plane. The optics of these cameras are optimized for 35 mm film, and most existing CCDs are smaller than a 35 mm film frame, so the image projected on the CCD is physically larger than the size of the CCD causing only the center portion of the image to be captured. Practically, this multiplies the effective focal length of any mounted lens so that a conventional 50 mm lens mounted on a digital camera with a multiplication factor of 1.5 behaves like a 75 mm lens, while a 100 mm lens behaves closer to a 150 mm lens. Modification of existing camera stands and other equipment used for film photography may be necessary to accommodate this difference. The magnification factor does not apply to several recently developed SLR digital cameras that have a “full frame” CCD equal in size to a 35 mm frame (24  36 mm). Dedicated photomicrography cameras provide the ultimate resolution, flexibility, and ease of use for critical photomicrography [9, 13, 30]. The features of several selected and dedicated still photomicrography cameras are shown in Table S1. The best of these cameras presently offers image resolutions of 12–20 megapixels, 42 bit RGB color depth, a dynamic range of 1:2500 or higher, and special hardware or software features to eliminate color fringing, moire´, blooming, and other problems. Most of these cameras incorporate a “C-mount” interface to a microscope, and some include Peltier cooling for fluorescence and darkfield microscopy. They usually interface to a personal computer through a USB or PCI-interface card and include advanced computer functions for image acquisition, adjustment, annotation, measurement, archiving, and reporting. Modern “virtual microscopy” imaging technology has also resolved the limited value of single still photomicrography images. Automated robotic slide scanners (i.e., whole slide imagers, digital slide systems, virtual microscopes, wide field imagers) permit selected portions or entire glass slides, to be scanned and digitized at high magnification for subsequent examination with special viewing software. The software provides a “virtual” microscope experience by permitting panning and zooming, as well as capture of selected fields, printing, text or voice annotation. “Virtual slides” are stored on servers or DVD-ROMs, and can be transferred to Internet-based servers for streaming access on the Internet. There is enormous potential of virtual microscopy for education, clinical service, and research. However, this challenges the present

422

Roger S. Riley et al.

capabilities of image data storage, viewing, and transfer because the data from each slide may exceed 30 Gb or more. For clinical applications, this technology also raises new legal and ethical questions about data protection, sharing, and availability. Currently, several companies supply robotic slide scanners, image viewing, storage, and manipulation software, Internet server capacity, and other hardware and software products for virtual microscopy [31, 32] (Table S2). In the absence of a robotic slide scanner, multiple sequential single images of portions of a microscope slide can be obtained with a still photomicrography camera and “stitched” together using commercial software for processing it into a virtual slide. Of particular interest to neuroscience researchers using confocal microscopy are companies (MicroBrightField, Inc., MBF Bioscience, VT, USA)1 dedicated to supplying image processing software for the construction of 3D images from serial sections, confocal stereology, and the quantification of fluorescently labeled cells. 2.6 Digital Video Cameras

1

Digital video cameras offer the advantage of high-resolution, “realtime” WYSIWYG (“What You See Is What You Get”) output for situations where a good video image display is necessary, such as during a video conference. Still images may be captured from the video stream with a “frame grabber board.” Video cameras with three-chip technology offer good to excellent color resolution and adequate to good low light sensitivity. However, adapting these cameras to an existing microscope system is generally more difficult than applying digital still cameras. Compatibility of the computer software and image processing board is often a critical issue because of the requirement of a computer interface. The optical interface with the microscope can also be complex, and these cameras are prone to image display problems, such as RF interference and color balancing problems. Some digital video cameras will accept video or SLR camera lenses, but their bulk and necessity for a computer interface are limiting factors for applications other than studio photography and photomicrography. Some of the major factors to consider in selecting video cameras are the camera resolution in TV lines, the type of output (i.e., RGB, S-video, composite), and the signal-tonoise (S/N) ratio. The practical problems in performing video microscopy are the topic of several reviews [33]. Some manufacturers have modified existing video technology to provide high-resolution photomicroscopic digital cameras [34]. These cameras are directly connected to a computer, either through an image transfer board or through a conventional computer interface (DVI, USB, Firewire). The most sophisticated of

http://www.mbfbioscience.com

Digital Micrographs

423

the digital imaging cameras are the video hybrid imagers [15, 35]. These cameras provide high-resolution 3-CCD images with complete WYSIWYG functionality, TWAIN(2) compliance, and “real time” image viewing for image composition, focusing, or small-group conferencing. Most of these cameras are physically small and easy to mount to a microscope trinocular head. Fluorescence photomicrography is possible with hybrid video cameras because of their excellent light sensitivity, relatively cool operation, and short-duration integration times. Peltier cooling is available in some hybrid cameras adapted for imaging of low light level fluorescence images. Some recent models also offer “real time” color correction, image enhancement, and resolution up to 1920  1080 at 60 frames/s. 2.7 Image Storage, Searching and Retrieval

A major advantage of digital still and many digital video cameras is the ability to accumulate and temporarily store large numbers of images in the camera for later transfer to a computer. The images are usually stored on removable, reusable store media that can be inserted into a computer or printer. The number of images that can be stored depends on the capacity of the storage device, the number of pixels comprising the final image, the type of storage format chosen by the user, and the degree of compression applied. However, present high-resolution (10–20 megapixel) images stored at acceptable compression ratios each require a 5–20 Mb of storage space. When the memory on a storage device is exhausted, it can be removed and replaced with a blank one so that acquiring a storage capacity is limited only by the size of the budget. Although professional digital cameras directly linked to a computer do not have the same limitations imposed by removable storage media, the camera is restricted to one location. Compact flash and secure digital (SD) are currently the most popular types of portable image memory media.

2.8 Image Processing and Manipulation

The images acquired by a digital camera must be transferred, or downloaded, into a computer for long-term storage and further manipulation. During the downloading process, images are transferred from the camera storage device into the computer storage device. This can be accomplished by connecting the digital camera to a computer and directly downloading the images from a storage device, or by removing the storage device from the camera and using a computer-connected adapter to transfer the images. Since most digital cameras include internal serial, parallel, USB (Universal Serial Bus), or Institute of Electrical and Electronics Engineers (IEEE) 1394 (Firewire) ports, downloading can occur simply by

2

A way of regulating communication between software and digital cameras or scanners. The word TWAIN (not an acronym in fact) is from Kipling’s The Ballad of East and West: “... and never the twain shall meet...” It is meant to reflect the difficulty of connecting scanners and personal computers at that time (early 1990s).

424

Roger S. Riley et al.

linking the camera port to the corresponding computer port by a cable, and then using a computer program to view, select, and transfer the images. The speed of image transfer varies from 0.12 Megabits per second (Mbps) for serial devices to 25 Mbps for Firewire ports. Some digital cameras also incorporate wireless (infrared or cellular) transfer technology that eliminates the need for a physical connection between the camera and computer. Card readers or adapters are popular for the frequent transfer of images into the computer, because they are usually faster and eliminate the need for constantly connecting and disconnecting the digital camera to the computer. The PC card slot in a laptop computer provides a convenient and rapid way of transferring images into these computers. Adapters are available for most popular types of flash memory that permit the insertion of the memory card into the PC card slot. A variety of card readers are available for desktop computers without PC card slots. These utilize the serial, parallel, USB, or Firewire ports and have reading slots for one to several types of flash memory cards. A special adapter is also available for Smartmedia cards that utilize the 3.500 floppy disk slot of a computer. An almost endless number of software manipulations can be performed on captured digital images to remove artifacts and optimize them for special purposes, such as printing. Image manipulation and printing is discussed below. 2.9

Image Printing

A variety of high-quality and relatively inexpensive color and blackand-white laser printers are [presently] available for the home and small business office, although liquid inkjet printers are the most common. High-quality (“photo-quality”) inkjet printers are available for less than $400, and produce a full-color letter-sized photorealistic print in several minutes for less than $1.00. These printers spray small “microdroplets” of liquid ink onto a paper surface by a mechanical or piezoelectric process as the print head travels across and down a sheet of paper. The printer can print a dot of color, print several dots on top of each other, or leave the dot blank (white) at each of the several million points on the paper. The liquid dye is stored in small, user-replaceable cartridges. In addition to the quality of dye and the number of spray nozzles, print quality also depends upon the type of paper used. While liquid dye soaks into ordinary printer paper, high-resolution printing requires specially treated glossy paper. Although most inkjet prints fade with time, recently introduced special dyes and papers can last 30 years or more. Although “desktop” inkjet printers cannot match the resolution, color accuracy, or longevity of a color printing press or photographic print, the performance–price ratio continues to increase with rapid technical innovations in the field. Inkjet printers have a multitude of educational abilities and can be used to proof documents for submission to a service bureau. The

Digital Micrographs

425

major disadvantages of personal color printers, compared to color monitors and photographic film, are their reproduction of a smaller range of color, or their smaller color gamuts [36]. Fortunately, there are several ways to optimize the printed output. Calibration of the color monitor to the printer is an absolute requirement so that the behavior of the printer can be accurately predicted from the appearance of an open file on the screen. If adjustments are needed to produce more accurate print output, the parameters of the printing program can be changed to send different instructions to the printer, or the parameters of a copy of the original file can be altered. Electronic darkroom software programs, such as Photoshop or PaintShop Pro are usually used to print digital images, but there are several supplemental software “plug-ins,” which greatly expand the capability of these programs. Other types of color printers include dye sublimation (dye-sub), solid ink-jet, thermal-wax, thermo autochrome, snapshot, Fujix Pictrography, Iris inkjet printers, and Fiery color servers. Generally, they are much more expensive than inkjet printers, but are capable of extremely high resolution. High-quality photographic prints for publication, insurance documentation, or other applications can be economically obtained from online service bureaus with digital photographic printers [37]. Film recorders are devices that render digital images to color or black and white film, either as a film negative or a transparency at resolutions of 2000 lines to 8000 lines of resolution [36]. In a film recorder, the output file is displayed on a small cathode ray tube (CRT) that is used to expose a frame of film in a line-by-line fashion. Film recorders are available in pathology departments, university multimedia support departments, and commercial slide imaging service bureaus. Most film recorders produce 35 mm slide film; larger output film recorders are also available that can image 4  5-in. film to produce poster-sized prints, or 8  10-in. overhead transparencies for presentation. To assure accurate color reproduction, the gamma of the monitor used to develop the slide must match the gamma of the film recorder’s CRT. The output file must also have the proper aspect ratio and be the proper size. If the slides include text, font compatibility between the computer of origin and the film recorder is crucial.

3

Practical Considerations in Scientific Imaging

3.1 Selecting a Digital Camera

Digital photomicrography of the peripheral blood smear requires a digital camera, a high-quality microscope, a mount for attaching the camera and microscope, and computer hardware and software for processing and storing the digital images. Fortunately a multitude of digital camera, microscope, microcomputer, and software options are available for the scientist interested in digital

426

Roger S. Riley et al.

photography. The selection of the appropriate components of a digital camera system are dependent upon a number of factors, including existing microscopic resources, the anticipated applications of the digital camera, specimen types, the expected application of the digital images, the availability of funds, and the knowledge and prior experience of the individuals involved. Both digital still cameras and digital video cameras are suitable for photomicrography, although the features of each type of camera are different. A professional SLR-type digital camera with a removable lens or a prosumer digital camera with good close-up capability is recommended for brightfield photography if the camera will be removed from the microscope and utilized for gross or general photography. In contrast, digital video and dedicated photomicrography cameras are permanently fixed to the microscope but permit a “live” high-resolution image to be displayed on a high-resolution monitor(s) for simultaneous viewing. These cameras also permit imaging at low light levels, morphometric measurements, and other beneficial features. Digital camera models with high light sensitivity, short integration times, and Peltier cooling are required for the acquisition of fluorescence or darkfield images, when imaging times of greater than 1 s are common. This is especially true for fluorescence in situ hybridization (FISH) image capture, where tiny intranuclear fluorescent dots are being photographed. The necessary spatial resolution for a photomicrographic digital camera is determined by the field of view and objective numerical aperture. At a given magnification, increasing the number of pixels beyond the minimum requirement (oversampling) increases the color accuracy but not the spatial resolution [38]. The Nyquist– Shannon criterion states that two pixels per resolution limit are required in order to fully utilize the resolving power of the microscope. In bright-field microscopy, it equals λ/2NA (where NA is the numerical aperture of microscope objective) (Fig. 2). Other technical factors of importance include the camera noise and sensitivity, dynamic range, and temporal resolution [8]. In addition to the technical features of a selected digital camera system, the potential purchaser should consider practical features including the cost, user convenience, reliability, the availability of training and technical support, and the cost and availability of repair [8]. Computer software is a separate consideration. Nearly all dedicated photomicrography systems include software for processing the acquired images, but third-party software programs may offer more advanced and/or superior user friendly features. Another consideration for some users is the capability of software to program and run customized scripts for automated image acquisition and analysis [8]. Product literature, trade shows, and manufacturer demonstrations can provide helpful information, but obtaining personal information from other users and firsthand experience is advised prior to the investigation of an advanced digital camera system.

Planapochromat 0.45

5

Planfluor

4 0.75

Achromat

0 x4

x10

x20

x40

1.25

1.40

1.30

0.65

0.95

0.40

1

0.50

0.10

0.25

0.30

0.13

3 2

427

Objective type

0.75

6

0.20

REQUIRED CAMERA RESOLUTION (Megapixel)

Digital Micrographs

x100

OBJECTIVE MAGNIFICATION

Fig. 2 Requirements for digital image resolution in brightfield images if resolving power of the microscope is to be fully utilized (λ ¼ 580 nm, aspect ratio of image sensor 5/4, projection lens 0.63). Typical combinations of objective magnification and numerical aperture (NA) are shown. NA is the number in/above the columns. Based on data published elsewhere [38], and assuming the Nyquist–Shannon criterion to be satisfied, that is, pixel size not exceeding ½ of resolution limit [178] 3.2 Performing Digital Photomicrography

The major requirements for optimal digital photomicrography are identical to the requirements for silver halide photography, namely, thin (2–3 μm), evenly sectioned, well-stained slides and properly aligned, evenly illuminated microscopes with clean, high-quality lenses. The basic principles of microscopy, photomicrography, and microscopy optimization have been extensively reviewed by Vetter and others [39–54]. The manufacturer’s manual should also be consulted for particular microscope and digital camera models. Environmental vibrations pose a serious, and often unrecognized, problem for photomicrography. To obtain high-quality images, a photomicrography system should be placed in a quiet room on a sturdy table with four legs. If vibrations remain a problem, the end of each table leg can be embedded in a small container of sand, or special vibration-dampening equipment and techniques can be employed [55–57]. For occasional non-critical photomicrography, the front lens of the fixed-lens digital camera may be set for infinity focus, manually placed at the focal point of the microscope eyepiece, and held in a steady position with the shutter tripped [29, 58]. The LCD monitor of the camera is used for fine composition and evaluation of the final image. However, for suitable images, an adapter (relay lens) is required to hold the camera and mount it to the eyepiece or preferably, optical port of the microscope. Unfortunately, poor image quality, chromatic aberration, vignetting, and other problems can result from optical incompatibilities between the adapter, camera, and microscope [59].

428

Roger S. Riley et al.

The field of view (FOV, angle of view) is a critical consideration in the selection of a relay adapter. The FOV is the angular extension of an image captured by a camera, which is determined by the focal length of the camera lens and the size of the image sensor. The microscope has a circular FOV, while the FOV of the digital camera is square or rectangular, so the image sensor of the camera must be matched to the relay adapter with the goal of capturing the largest area possible from the camera’s circular FOV without vignetting. This requires balancing the magnification and physical diameter of the relay lens, the optimal placement of the camera lens, the camera aperture, and the lens magnification (“zoom factor”). The LE-Adapter (LensPlus, Redding, CA, USA) is the simplest and least expensive type of adapter. The LE-Adapter consists of a threaded coupling ring and a lens adapter ring. The coupling ring is threaded into the filter threads of the digital camera lens, while the eyepiece is inserted into the lens adapter and secured with setting screws. Lastly, the digital camera–LE-adapter assembly is inserted into the sleeve of the microscope eyepiece. A special relay lens (Nikon MDC Lens, Nikon Co, Tokyo, Japan) for the Nikon Coolpix 900 series digital cameras are also a popular low-priced alternative to dedicated photomicrography cameras. The MDC Lens consists of a 10 eyepiece with male 28 mm threads compatible with the female mounting thread of Nikon Coolpix cameras, while the relay (projection) lens is inserted into the sleeve of the microscope eyepiece. Several commercial firms specialize in the development and manufacture of camera mounts, couplers, and relay lens for many consumer-grade digital cameras (Table 3). Adapters can also be constructed for a specific camera by an optical facility if care is taken to provide a stable support for the camera; the user avoids damaging the optics of the camera and microscope, and completely eliminates extraneous light. A computer interface is not required to capture images since these cameras have integral data storage capability. Once the camera is properly mounted and the microscope optimized for photomicrography, digital photomicrography with most consumer integral lens digital cameras is performed with the camera in the aperture priority mode and the focus adjusted to the infinity setting. The aperture is set to the lowest f-number, and the zoom setting is adjusted to assure the desired image field of view while avoiding vignetting. An AC adapter is preferred over internal batteries to reduce weight and provide a convenient, continuous power supply. The metering, image adjustment, contrast, sharpening, and other functions of the camera are adjusted to optimize image quality [60]. Digital photomicrography with interchangeable lens, SLR-type digital cameras are similar to that with filmbased SLR except for the availability of digital software functions, the ability to immediately view captured images on the LCD monitor, and the special problems imposed by digital photomicrography

Digital Micrographs

429

Table 3 Suppliers of camera mounts, couplers, and relay lenses (∗) Supplier

Product(s) a

Edmund Optics Barrington, NJ, USA

Universal digital camera microscope adapter coupler for popular Canon, Fujifilm, Nikon, and Olympus integral lens digital cameras

GreatScopesb High Point, NC, USA

Filter-treaded mounts for many integral lens cameras; inexpensive digital single lens reflex (DSLR) relay adapters; “professional” multielement relay lenses for DSLRs and camcorders

Martin Microscope Companyc Easley, SC, USA

MDSLR proprietary widefield projection lens for select DSLRs, MM-SLR adapter for select DSLRs, MM multielement relay lens for select integral lens cameras and camcorders

Microscope Worldd Carlsbad, CA, USA

Meiji camera adapters for integral lens digital cameras and DSLRs

Optics Planete Northbrook, IL, USA

Wide variety of camera mounts and adapters

PAX-itf Villa Park, IL, USA

C-mount adapter for LOMO brand microscopes used to be on the portfolio

Qioptiq Imaging Solutionsg Rochester, NY, USA

Optem camera couplers for a wide variety of integral lens digital cameras, DSLRs, and dedicated digital photomicroscopy cameras

ScopeTronixh Berlin, Germany

“MaxView™ Plus” digital camera attachment system for integral lens digital cameras

Spectra Servicesi Ontario, NY, USA

Camera couplers for a wide variety of digital still and video cameras

SpotImagingj Direct image projection (DIP) microscope adapters (a division of Diagnostic Instruments) Sterling Heights, MI, USA Zarf Enterprisesk Spokane, WA, USA

Proprietary lens adapters for a wide variety of digital still and video cameras

(∗) as of July 2019 a http://www.edmundoptics.com/onlinecatalog/DisplayProduct.cfm?productid¼2416 b http://www.greatscopes.com c http://www.martinmicroscope.com d https://www.microscopeworld.com/t-digital_video_adapters.aspx e http://www.opticsplanet.com f https://www.paxit.com/microscope-cameras/microscopes-and-imaging-hardware/ g http://www.qioptiq.com/optem-microscopy-couplers.html h http://www.scopetronics.com i https://spectraservices.com/category/microscope-camera-couplers.html j http://www.spotimaging.com/microscope-adapters-2/ k http://www.zarfenterprises.com

(see below). Most dedicated photomicrography camera systems include specialized software for image optimization and capture. Some knowledge of photography is required to obtain an image with proper exposure and accurate color rendition

430

Roger S. Riley et al.

[60]. For example, photomicrography using the “auto” settings of attached, handheld digital cameras results in a dark image since the white balance is not optimized for the tungsten or halogen illumination systems of microscopes and the camera averages the scene for a gray scale image. To correct this problem, the camera’s white balance must be manually set on a blank, fully illuminated microscope field. The level of illumination should not be changed after setting the white balance, because this will change the color temperature of the light, and alter the white balance. For convenience, many microscope illumination systems have a “photography” illumination setting to assure consistency between images. The autoexposure settings of both still and dedicated photomicrography cameras work well for most images. However, the exposure compensation of +1 to +2 is required to achieve a white background in specimens with an “open” background, such as peripheral blood smears, bone marrow aspirates, and cytopathology specimens. Most digital cameras also provide an option to “spot meter” individual cells or portions of the image to produce correct exposure values. This option should be used whenever possible to achieve the best exposure. Fortunately, digital cameras provide an immediate preview of the captured image so that incorrect white balance, exposure settings, or focus can be corrected. Exposure, color balance, and some focus problems can be corrected later in an image analysis program, but it is much easier to assure proper exposure at the time of imaging. If available, the option to capture image data in native “RAW” format is recommended, since this format provides more flexibility for further image optimization. Photomicrographic images of suboptimal quality can result from inherent limitations of the equipment used or improper photography technique. Problems common with any photomicrography system include uneven illumination, out-of-focus images at low magnification, loss of important image detail, poor color rendition, and inconsistent exposures [61, 62]. Problems unique to digital photomicrography include vignetting, a small FOV, and a dark image background. However, a major advantage of the digital camera is the ability to immediately view captured images, permitting the rapid recognition and correction of imaging problems. There is increasing interest in standardized techniques of image optimization and quality assessment, particularly for the wholeslide imaging of human tissue performed for diagnostic purposes [62–64]. Practical solutions for common imaging problems in digital photomicrography are further discussed below and summarized in Table 4. Vignetting is the appearance of dark, rounded edges on a digital image. Vignetting results when a rectangular digital image is obtained through a circular lens, and the lens diameter of the digital camera is greater than the diameter of the microscope

Digital Micrographs

431

Table 4 Common problems in digital photomicroscopy (adapted from author’s own paper [177]) Problem

Cause

Solution

Dark images

Fast camera shutter speed

Increase illumination; use camera manual settings to increase exposure time

Dark image background

Improper white balance

Manually set white balance prior to obtaining images; use exposure compensation; use spot metering

Dark spots

Dust and debris on microscope, camera lens, or CCD

Carefully clean all lenses in the optical path, use a microscope dust cover when system not in use. Clean CCD (professional digital cameras only)

Loss of image detail

Improper setting of aperture diaphragm (lack of contrast with excessively large aperture diaphragm, diffraction with excessively small aperture diaphragm); vibrations and “camera shake;” poorly focused image; low digital camera resolution

Adjust aperture diaphragm to achieve best balance of detail and contrast; use “digital zoom” or focusing aids; use remote shutter release; place microscope/camera system on a sturdy surface; acquire a digital camera with higher resolution

Out of focus areas (particularly at image periphery)

Nonplanar microscope objectives, lens aberrations in camera or microscope

Try different objectives and tube (relay) lens, realign optical path of microscope

Reflections, bright spots, or concentric rings

Internal reflections from multiple lens surfaces and/or non-coated lens surfaces

Set camera to “aperture priority” mode and set aperture to largest lens opening (smallest f-number); try different objectives and tube (relay) lens; realign optical path of microscope

Small field of view

Relatively small physical dimensions of CCD

“Zoom-out” to maximize field of view, while avoiding vignetting

Uneven illumination

Microscope is misaligned for Ko¨hler Realign microscope optics to ensure Ko¨hler illumination illumination, or microscope design does not conform to Ko¨hler illumination. Common causes of misalignment include an off-center substage condenser or lamp, improper use of the diaphragms, and out-of-focus condenser

Vignetting

Relatively small diameter of relay lens; narrow camera aperture

Set camera to “aperture priority” mode and set aperture to largest lens opening (smallest f-number); “zoom-in”; obtain relay lens with wider diameter

432

Roger S. Riley et al.

eyepiece opening. This is frequently a problem with digital photomicrography rather than film photomicrography because of the relatively small size of the CCD in most consumer digital cameras. To avoid vignetting, it is recommended to use a relatively small camera aperture and a properly seated relay lens with sufficient magnification to fill the diagonal of the film frame. Digital cameras with zoom capability can also be “zoomed in” to fill the frame, although only the central portion of the lens area will be recorded in the final image. Out of focus images at low magnification results because low power objectives (2, 4) have a very small depth of focus (0.03–1.4 μm) at the image plane. In contrast, focusing is not as critical for higher power objectives with a larger depth of focus (20—4.75 μm, 40—12.0 μm) [15, 35]. The depth of focus describes the vertical tolerance in placing the camera for the image to remain sharp. It should not be confused with the depth of field which decreases with objective’s resolving power, and equals the distance between the nearest and most distant objects in the specimen, that still appear sharp. Prosumer still and video cameras require manual focusing using the microscope focusing controls. Since the integral LCD monitor of these cameras is small and suboptimal for accurate image focusing, most have a “video-out” feature to permit the live image to be streamed to a larger monitor or laptop computer, and some permit the image to be magnified for manual focusing. If a larger monitor or laptop computer is not available, a 2 or 4 magnifying glass or a special LCD magnifying viewer such as the Xtend-a-View Pro™ (Photosolve, Saratoga, CA, USA) can be used to magnify the image on the LCD monitor for more accurate focusing. Well-focused images are difficult to obtain using microscopes with low power objectives (i.e., 2, 5, 10). If an image with a large field of view and low magnification is needed, several images taken at medium power (i.e., 20, 40) can be easily stitched together using special software (see discussion below). Automatic focusing with appropriately equipped microscopes and special computer software is the optimal method to achieve sharp image focus. Using automatic focusing, some software programs will also capture multiple images at sequential focus planes and automatically merge the images into an “extended focus” image. Dedicated photomicrography systems without automatic focusing capability usually include a variety of aids to assist with optimal focus, such as focus indicators and zoom magnification of selectable portions of the image. Digital slide scanners offer another option for the capture of images of glass microscope slides at low magnification. These scanners are designed for imaging mounted slides or unmounted film, but some manufacturers offer special holders for microscope slides. Tissue sections can also be mounted on 200  200 glass slides and

Digital Micrographs

433

scanned in the same manner as a 35 mm slide. An area of up to 24  36 mm can be imaged, or the “preview” function of the scanner software can be used to select a specific region of the tissue section to scan. Images prepared by this technique provide excellent sharpness, even illumination, and superior color reproduction compared to conventional images taken with a camera or microscope [65–69]. 3.3 Macroscopic Photography

The immediate availability of digital images of macroscopic specimens is extremely advantageous for the neurosciences laboratory, since they can be printed, annotated with descriptive text, or used in other ways to facilitate evaluation of the corresponding microscopic sections, quality assurance, or conferencing. For example, hard copies of complex macroscopic specimens, especially those with resection margins, are commonly used in the pathology gross room to “map out” the areas sampled for microscopic examination. The optimal photography of gross specimens requires extensive experience, high-quality imaging equipment, an excellent illumination system, and detailed knowledge of camera operation and the principles of photography (Fig. 3). Since most prosumer digital cameras with integral lens have excellent “close-up” or “macro” capability, these cameras can be utilized for macroscopic pathology without further modification. If this is not suitable, a variety of lens attachments are often available for special specimens. The excellent macrophotography lenses, lighting equipment, and other equipment manufactured by Canon Inc. (Tokyo, Japan), Nikon Corporation (Tokyo, Japan), and other manufacturers can be used for macroscopic photography with SLR-type digital cameras, although the “multiplier effect” or “magnification factor” must be considered. Proper lighting is critical for sharp, evenly illuminated macroscopic images with excellent definition, contrast, and DOF. Fiberoptic illumination using a Xenon flash source is the ultimate light

A

B

Fig. 3 Macroscopic images of human brain before (A) and after (B) sectioning. The boxed areas in A were separately contrast-optimized

434

Roger S. Riley et al.

source for macrophotography. If possible, neuroscientists desiring optimal macroscopic imaging should consult a company specializing in scientific macrophotography, such as Visionary Digital (Palmyra, VA, USA). Advanced, digitally based workstations for gross biological specimens have been described by several companies [70]. The flatbed scanner can also be used to capture gross images of biological specimens. As described by Mai et al., the specimen is placed on the surface of the scanner bed covered by a protective transparent sheet of plastic or projector transparency film [71]. With proper lighting, high-resolution, shadowless images for surfaces with a depth up to 30 mm can be obtained. Matthews and Denney described a modified “wet scanning” technique to improve the quality of images with wet highlights [72]. They mounted a Perspex box with a clear 3 mm base on a scanner, partially filled the box with distilled water, immersed the specimen in the water, and then scanned the specimen using conventional PC hardware and software [72]. 3.4 Software Manipulation of Digital Images

Software manipulations of digital images are performed to (1) adjust the tonal balance, hue, saturation, or other image properties to more closely resemble the original image, (2) improve image resolution or recover small and/or dim objects, (3) remove image flaws and artifacts such as noise, dust, and scratches, (4) correct uneven background brightness, (5) extract specific information to perform further analysis, such as quantitative analysis or measurement of various parameters, (6) change the size (resample), format, resolution, or orientation of an image for a publication, e-mail, print, 35 mm slide, web page, or other document, (7) crop an image to remove extraneous information or emphasize the key area, (8) add text, arrows, scale bars, and/or voice annotation to an image to clarify or add additional information, (9) merge multiple single images together to create a panoramic image, 3D stereo image, montage, or animated display, and (10) organize, sort, rename, rate, assign keywords or comments, print, and store [21, 73–77]. Many of these enhancements can be achieved with consumer photo-editing (“electronic darkroom”) computer programs such as Photoshop and Lightroom (Adobe Systems Inc., San Jose, CA, USA), Aperture (Apple, Inc., Cupertino, CA, USA), and LightZone (Light Crafts, Inc., Palo Alto, CA, USA) aided by numerous free and commercial plug-ins, scripts, and actions for image sharpening, color correction, noise removal, and other enhancements. Some software programs, termed image content managers, function as “digital light boxes” or “image databases” and permit images to be displayed in a variety of ways, organized, sorted, printed, and stored in collections on hard drives, CD-ROM media, tape, or other storage devices. Most of these programs

Digital Micrographs

435

permit complex searches of image collections for specific files. In addition, there are many commercial software programs specifically written for the organization, editing, and/or analysis of gross, photomicrographic, and other biomedical images. Selected commercial scientific imaging software programs are listed in Table S3. In addition to these commercial applications, there is ImageJ, a complete, open-source, Java-based image manipulation program written by Wayne Rasband, available in the public domain from the National Institutes of Health for a variety of computer operating systems [78, 79]. ImageJ is frequently updated and is accompanied by more than 400 plug-ins to implement specific image processing features. Software enhancement is most often performed to restore the tonal balance of an image lost during the capture process. The brightness and contrast controls of Adobe Photoshop or other image processing programs are commonly used for this purpose, although some information may be lost by this technique. Fortunately, both the levels control and the curves control avoid information loss and permit better management of the enhancement process. The “levels control” in Adobe Photoshop provides a visual representation of brightness values in the form of a histogram, with brightness values (0–255 in eight-bit images) represented on the horizontal axis, and the number of pixels at each brightness

A

Raw

Raw

A

B

0

Raw

x2

255

B

Fig. 4 Examples of digital image manipulations to increase contrast. (A) Normalization (linear stretching of histogram). (B) Equalization (nonlinear deformation of histogram) to balance the proportion of dark and bright tones. The left part of the histogram is now less “dense” (see inset) so that the overall histogram “density” is uniform across the 0–255 range

436

Roger S. Riley et al.

value on the vertical axis (Fig. 4). Sliders below the histogram permit the bright, midrange, and dark values of an image to be adjusted separately. Adjusting the midrange values to change the brightness of an image without changing the shadow or highlight extremes or color composition is the only enhancement required for many images. This is analogous to changing the “gamma” on a computer monitor. If necessary, the shadow or highlight values can also be changed to reveal image detail lost in the black or white areas of the image. This can be accomplished by moving sliders or by using special “eyedroppers” to select the darkest region of an image for assignment as “black” and the lightest region for assignment as “white.” “Histogram equalization” is a histogram manipulation that can be performed on images that have underutilized gray scale values (i.e., large valleys). These images often show more detail when some of the gray level “peaks” are reassigned (Fig. 4) and the valleys compressed. The “curves control” in Adobe Photoshop is more complex than the brightness/contrast controls but permits more precise tonal correction, since any point along the 255-level tonal scale of each channel can be changed [21]. The curves control of an original, unaltered image consists of a two-dimensional line graph with a slanted line running through it at an angle of 45 . The horizontal and vertical axes of the graph represent tonal values in versus tonal values out. Clicking on the graph creates a point with the line flowing through. Moving the point changes of the brightness of the image is similar to moving the midrange point of the levels control. A new point can be placed on the graph with another click. With practice, curves can be created which thoroughly optimize the brightness and contrast of an image. Furthermore, curves can be saved, reloaded, and applied to other images. Image detail can usually be enhanced with this technique without altering the color composition. The hue of an image can be changed by controls that effectively rotate all of the colors in the image around the color wheel. Saturation, or color purity, can be changed by a separate control in image manipulation programs. The hue and saturation values of images from high-resolution digital cameras rarely require adjustment. Most scenes consist of an almost infinite gradation of color and tone. Conversion of a scene into a digital image with a finite number of ordered pixels results in an inherent loss of detail perceived as blurring or softening. Motion of the camera or the object being photographed can further add blurring. Fortunately, special software algorithms can usually restore much of the lost detail. The “unsharp masking” algorithms included in Photoshop and other image manipulation programs work by finding pairs of adjacent pixels, which differ in brightness by a specified amount (e.g., edges). They also and increase the contrast of these pixels and any others that fall within a certain radius. The brightness difference (threshold) can be specified, as well as the percentage by which the contrast of the edge pixels is increased (amount) and the number of

Digital Micrographs

437

pixels around each edge which are sharpened (radius). The optimal setting for these parameters depends on the type of image and the resolution of the final output device (on-screen display, ink-jet printer, laser printer, etc.). Deconvolution (“deblurring”) is a more complex process that restores lost image detail by extracting information from blurred portions of an image using nearestneighbor algorithms. Many dedicated photomicrography software programs incorporate deconvolution routines. Deconvolution is especially advantageous for low-light situations, such as fluorescence microscopy [80–82]. Small image defects such as dust and scratches can be removed with special subroutines (“filters”) in Photoshop or other programs, or by manually replacing the defective areas with pixels from the surrounding region. Two companies provide scanning and imaging software with sophisticated routines to remove dust, scratches, or other unwanted small objects. Software programs with this technology include the Silver Fast’s Smart Removal of Defects or SRD (LaserSoft Imaging, Kiel, Germany) and Digital ICE Technology (Eastman Kodak Company, Rochester, NY). Non-uniform microscope illumination can usually be avoided by using a highquality microscope carefully adjusted to achieve Ko¨hler illumination. However, even with these steps, the final image can sometimes show an uneven background that was not detected during image acquisition. The optimal means to resolve this problem, and restore a uniform background, is to acquire images from an illuminated area of the slide lacking the specimen. A new image must be acquired for each change in magnification or illumination. These background images have the same irregular illumination as the specimen images, and can be used by image processing programs, such as ImageJ, to correct the background. If background images are not available, a free, Java-based application program for the Windows operating system, known as the Background Subtraction Toolkit, is available that generates the background from user chosen control points in the image, and produces a uniform replacement background from this information [83, 84]. Photographic images are commonly stored in the Photoshop native (PSD), JPEG, BMP, or TIFF formats, but there are more than 145 available image formats. Although all image manipulation programs and content managers allow for opening a variety of formats and saving the images in different formats, some programs specialize in the conversion of images from one format to another. If the relevant information is only present in a portion of an image, this area can be selected with “cropping” tools and resaved as a separate image. Alternatively, “masking” tools permit desired objects, such as a cell or a person, to be selected, removed, and/or transferred to another image. Masking is especially useful for removing bloody or unattractive backgrounds from digital images of biological specimens. Information, such as symbols,

438

Roger S. Riley et al.

arrows, text, or even voice annotations can be added to digital images with most image manipulation, word processing, and website creation/management programs. The capability to manipulate or even to dramatically alter digital images with digital darkroom software is a major ethical and legal concern because these images can play an essential role in patient care and scientific research [85–90]. However, the problem, as Suvarna and Ansary emphasize, is not the software bur rather the “intention of the individual to deliberately falsify an image [91].” The problem of fraudulent images is likely to multiply as digital images become even more widely accepted for medical and scientific applications. A possible solution for publishers is to randomly audit individuals submitting papers or to require authors to submit both the original and modified versions of digital images. In view of the possibility of inadvertent or fraudulent misrepresentation of digital images, Pritt et al. and others have proposed the following general guidelines when modified digital images of pathologic specimens are utilized for any application [92, 93].

3.5 Panoramic, Continuous, and Object Photography

l

“Global adjustments” to the entire image in color, brightness, and/or contrast are acceptable if these changes will more accurately duplicate the original image.

l

The manipulation of focal areas of an image, such as changing the background or removing selected cells or structures, is usually not acceptable.

l

Multiple images should not be merged to create a single continuous image. If multiple images are placed into a single figure, they should be delineated by lines, boxes, inserts, and so on.

l

Annotations should be placed in such a manner to preserve the major features of the image and not alter its interpretation.

l

A description of any digital manipulation in a manuscript should be described in the “methods.”

l

The original, unaltered images should be archived and available to interested parties upon request.

One of the most promising applications of digital photography is for medical education [94]. Panoramas are “images with unusually wide fields of view that extend far beyond a single camera snapshot” that convey a better sense of realism than images of normal width [95]. Circular and spherical panoramas are popular and widely used to present immersive views or “virtual tours” of landscape photographs, real estate, and architecture. However, flat or planar panoramas can also be prepared from a mosaic of single images of a microscope slide (Fig. 5), high-altitude photographs, maps, or similar data. Panoramic pictures are prepared by merging or “stitching” multiple detailed overlapping images into a larger image. When panoramas are saved in Quicktime VR movie or similar format,

Digital Micrographs

439

Fig. 5 A “panorama” image stitched from three individual micrographs obtained as a mosaic. In some cases, stitching software requires the individual images to overlap

the viewer can observe the entire image, “zoom-in” for a highresolution view of a particular area, or move horizontally or vertically. The term “virtual microscopy” has been applied to panoramic images of microscope slides, since the viewing experience is analogous to using a microscope in real time. Panoramic images are automatically obtained by the newer virtual microscopy systems, but scientists without access to these systems can use Photoshop or other special “stitching software” to prepare similar images. These programs require multiple images of a scene that overlap by 30% to 50%. Some programs will only prepare mosaics from images taken on the same horizontal plane (single-row panorama), while other programs will stitch images from mosaics, multiple rows (multi-row panorama), or even from a sphere (spherical panorama). Many stitching programs will not only seamlessly stitch digital images but also remove artifacts in the original photographs and those created by the stitching process. In continuous photography, a sequential series of pictures are combined into a movie or animated gif to show an action sequence. Some digital cameras permit time lapse photography by taking a series of pictures at specified intervals. Object photography is performed by taking multiple images of a stationary object at specified angles around its periphery to produce three-dimensional (3D) images. With the proper software, a view at any angle can be displayed, or the images can be combined into an animated gif or movie, so that the object appears to rotate in space. Small objects are usually photographed on a special turntable termed an object rig. The object is mounted in the center of the turntable, which is rotated a precise number of degrees, as photographs are taken from a digital camera at a fixed position. Other resources have described the creation of an image library of 3D models of normal and diseased human organs [96]. The most recent development in digital imaging is the acquisition and software manipulation of multiple images of a subject to obtain improved resolution, dynamic range, depth of field, or

440

Roger S. Riley et al.

contrast. Collectively, these techniques are known as high dynamic range imaging (HDRI) [94]. In the biological sciences, the technique of “focus stacking” is increasingly applied to compensate for the limited depth of field of the microscope [97, 98]. This versatile technique uses a bracketed “stack” of partially focused images, each obtained at a slightly different focus, by brightfield, fluorescence, confocal laser scanning, or other forms of microscopy [97]. The focused portions of each image are then combined to produce a single, sharply focused image using special imaging analysis software with extended depth of focusing algorithms. Some imaging processing software programs permit live 3D manipulation of the image stack, including the creation of representative “slices” at different depths of focus. Recent examples of the utilization of this technology in the neurosciences include the construction of representations of complex bundles of neural processes, spatial distribution patterns of newborn neurons, the automated detection and analysis of individual neurons, and the determination of the three-dimensional (3D) morphology of neurons in the cortex [99–102]. 3.6 Preparing Digital Images for Electronic Publishing

Electronic publishing is “distributing information via computer instead of paper” [103]. Following the computer revolution of the past two decades, high-quality electronic publishing hardware and software is available to nearly every profession, and much of the general public in the developed countries. With electronic publishing software and hardware, it is not only possible to distribute digital images and written documents on the Internet but also through e-mail messages, locally via intranets and the CD-ROM. The major advantage of electronic publishing is speed and flexibility. Documents can be shared in seconds rather than hours or days, and document creation does not depend upon dictating machines, secretaries, or printers. Presently, electronic publishing is standardized around two major formats, HyperText Markup Language (HTML) and Adobe Portable Document Format (PDF). HTML is most widely utilized format for Internet-based publications, but it can also be used for other forms of document sharing. HTML is extremely flexible, since documents can vary greatly in complexity, and include formatted text, hypertext links, images, and animations. PDF documents are mainly used to share documents via CD-ROM and e-mail. Since the PDF format is based on the Adobe Postscript language used by some printers, a document prepared in any software program can be compactly stored in its original form as a PDF file, and then reproduced on screen or in printed format by anyone with a free viewing program termed the Acrobat Reader. PDF documents can also include interactive and multimedia components such as hypertext links, text searching, and “fill-in-the-blank” forms. Images for publication in electronic documents or websites should meet the existing submission guidelines of the publisher.

Digital Micrographs

441

These guidelines are usually different from those for printed documents due to restrictions in the amount of information web browsers can process and the amount of information electronic monitors can display. To maintain reasonable bandwidth, web image file sizes should be as small as possible while still maintaining the integrity of the original image. A high-quality, noise-free image is a critical requirement. Generally, the image is then downsized, sharpened, converted to the sRGB colorspace, converted to an eight bit, 72 dpi image, and saved as a JPEG file. Since monitor screen sizes and Internet bandwidth are continually increasing, an image size of at least 800  600, or preferably 1024  768, is desirable for electronically displayed images. Although the narrow sRGB color gamut still approximates most of the default behavior of electronic monitors much better than the wider Adobe 1998 color gamut used for printing, images displayed in the Adobe color gamut will look dull. Photomicrographs should include internal scale markers, and the figure caption should include the magnification used to obtain the image. Image processing with an aim other than to restore the quality of the original image is generally not permitted for published scientific images. Adobe Photoshop is widely used for the preparation of images for electronic publication, but there are specialized programs for this purpose, such as the Photoshop plugin Web Presenter Pro (FM Software, Los Angeles).3 The international availability of original texts and images via the Internet and CD-ROM has made copyright issues a priority that must be considered by creators of documents for electronic publishing. In the USA, writers, inventors, artists, playwrights, musicians, and other creators of original works are protected via copyright protection by the U.S. constitution (Article I, Section 8), in order to “promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.” Duplication of all or any portion of a work without permission from the creator is a violation of copyright law and subject to fines and prosecution. This excludes materials which are no longer protected because of time or applications, which meet the doctrine of fair use, defined as “. . . the fair use of a copyrighted work, including such use by reproduction for purposes such as criticism, comment, news reporting, teaching, scholarship, or research.” Material that is not copyrighted is considered to be in the “public domain” and can be freely duplicated provided credit is given to the original author. In spite of the existence of the copyright law, many photographers apply additional protection to their works by using visible or invisible digital watermarks. Visible watermarks are translucent overlays

3

http://www.fredmiranda.com/software

442

Roger S. Riley et al.

applied to an image with the author’s name and other information. Invisible watermarks are imbedded information in the file itself that can be displayed by the appropriate software. 3.7 Preparing Digital Images for Print Publication

Digital imaging, computer graphics, and technological innovations in output technology have revolutionized the world of printing [104, 105]. Today, printing encompasses a number of output options, including home and small office printing, film recording, service bureau printing, and commercial printing [36]. In spite of the recent emphasis on producing graphics for electronic publication, scientists may need to output digital images to physical media for publication in scientific books, journals, posters, patient care reports, brochures, newspaper articles, or printed handouts used for a lecture or conference. The major goals for printed images are to fulfill the submission guidelines of the print publication while assuring that the final printed image is an accurate representation of the original image [106, 107]. The submission guidelines of the print publication should be carefully reviewed. Although most journals adhere to the general submission guidelines of the International Committee of Medical Journal Editors, they have their own specific requirements [108]. At a minimum, these should specify the size and/or aspect ratio of the final image, the resolution required, for the acceptable file format(s), and whether the image should be submitted in color or gray scale. Some publications accept color images but charge an extra publication fee. Print publications usually require raster graphics in the TIFF format and vector graphics in EPS. If color is acceptable, the color space should be specified. Some publications accept images in the RGB format and perform their own CMYK conversion, while others required a user-converted CMYK image. Once the submission requirements are confirmed, (application of) the computer processing of the digital images can be performed. This usually involves global processing to improve the brightness, contrast, and/or color balance, merging multiple images into a single image, and the addition of text, arrows, and other typographic material [107]. It is unacceptable to add, alter, or remove a portion of an image or a particular image feature, merge or splice multiple images to falsely imply that they represent a single field, or to alter an image in any other way to cause a different interpretation by the viewer. Resizing an image to meet submission requirements is necessary, but cropping should be performed with extreme caution. A journal may require that both the original and enhanced image be included at the time of publication submission, or may require submission if an inappropriate alteration is suspected by the editor or reviewers. The science of “digital forensics” has recently arisen, because of concerns about fraud in science and the press. Some scientific journals routinely apply special forensic software programs to all submitted images to detect fraudulent manipulation that may otherwise not be noticed.

Digital Micrographs

4

443

Medical Applications of Digital Images Medicine is an extremely visual science. Representative, highquality images of pathologic specimens are essential for research, teaching, and selected areas of patient care. The recent availability of digital images overcomes many of the past limitations of film photography and makes gross and photomicrographic images immediately available [90, 109–111].

4.1 Medical Education

Digital imaging is an essential tool for medical educators [112]. With digital images, students and health professionals are no longer restricted to using cumbersome books, printed atlases, and microscope/Kodachrome slide sets, and can access instructional material at any time via the Internet or CD-ROM (teleeducation) [113, 114]. In addition, digital presentations have largely replaced 35 mm slide-based [ones] in medical education, since they are relatively inexpensive and easy to prepare and update, and can be immediately provided as supplemental material for a lecture. Starting in the mid-1980s, single digital static gross and photomicrographic images have been utilized in pathology education. Several institutions are sharing their collections of digitized gross and photomicrographic images as large, high-quality Internetbased digital atlases for medical education. Notable examples include the following: 1. Internet Pathology Library for Pathology Education,4 University of Utah. 2. American Society of Hematology Image Bank [115].5 3. Urbana Atlas of Pathology,6 University of Illinois College of Medicine at Urbana-Champaign. 4. Synthesis of Digital Resources for Medical Education: Pathology,7 Histology8 etc. [116]. 5. Hypertext Atlas of Organ Pathology,9 Masaryk University, Czech Republic [117]. Although these websites primarily feature static images, there is great interest in the use of panoramic, object-based, and threedimensional (3D) gross and panoramic microscopic photography in medical education. In this regard, the current state-of-the-art system utilizes virtual microscopy, whole-slide imaging, and Java-

4

http://library.med.utah.edu/WebPath/webpath.html http://ashimagebank.hematologylibrary.org 6 https://www.med.illinois.edu/m2/pathology/PathAtlasf/titlepage.html 7 http://www.syndrme.org/index.php/pathology 8 http://www.syndrme.org/index.php/histology 9 https://atlases.muni.cz 5

444

Roger S. Riley et al.

based Web-based streaming of Flashpix images from a server with appropriate annotation and supplemental text using a “pan and zoom” client viewer [118]. Excellent examples of this technology include the Virtual Slidebox,10 an extensive educational site developed by Dee and collaborators at the University of Iowa, and the European Virtual Microscopy Network,11 sponsored by the Universities of Helsinki and Tampere. In addition, an interactive, multiresolution atlas of the brain (Brainmaps)12 was developed by Mikula and collaborators [119]. Other studies revealed that in 2007, approximately 33% of the academic pathology departments in the USA have implemented some form of virtual microscopy for medical education and that that only 30% were continuing to use glass microscope slides [118]. Several studies have proven the efficacy of virtual microscopy for teaching medical students and residents in comparison to conventional teaching with glass microscope slides [120–126]. For educational purposes, an unlimited number of virtual slides can be viewed through the Internet using browser technology giving users access to entire libraries rather than small areas of predefined regions of interest in microphotographs. This tool enhances the educational field in histology and pathology since the students are able to increase their ability to identify areas of diagnostic or histological patterns in the context of the whole sample. Virtual microscopy with the help of the web-based setting allows viewers (students and trainees) to share the same virtual slide without even being in the same room. There is economic impact that can be achieve since there is not need to equip laboratories with microscopes but with rather less expensive computers, thus lowering the training costs. Compatible software allows for digital slides to be stored in servers and the images retrieved by users anywhere. This type of software generally uses a standard browser and permits to the end user to look at the slide using on-screen functions. This function allows the students to digitally annotate the virtual slides and save the still images as JPG or TIFF files for later review. Another advantage of virtual slides is they ensure that each student sees the same slide as the rest of his classmates avoiding the discrepancies during testing coming from non identical glass slides.13 The always evolving field of virtual microscopy has expanded to new dimensions on how to show and how to interact with the images. A new concept was presented by the collaborative work between the Institute for Molecular Medicine Finland (FIMM) and Multitouch Ltd. that allows to rotate or zoom in and out the

10

https://www.mbfbioscience.com/iowavirtualslidebox https://doi.org/10.1007/s00428-009-0749-3 12 http://brainmaps.org 13 https://doi.org/10.1111/j.1600-0463.2011.02869.x 11

Digital Micrographs

445

scanned images using touch screen capabilities (see video online14). Digital images, particularly whole slide images, are increasingly used in continuing medical education, professional examinations, and proficiency testing. For example, many organizations, including the American Association of Neuropathologists (AANP), have replaced glass microscope and 35 mm slides with digital whole slide images for educational courses. This approach is not only convenient for the presenter and participants but reduces costs and provides access to valuable case material where limited tissue is available [6]. Although The American Board of Pathology has used static digital images in certification examinations for several years, they are increasingly adopting whole slide images. In pathology, digital images are now used by the American Society for Clinical Pathology in the Resident In-Service Examination (RISE) and other proficiency assessments. Interestingly, the College of American Pathologists (CAP) offers the Online Digital Slide Program in Dermatopathology and the Autopsy Pathology where the slides are reviewed and continuing medical education (CME) credits can be claimed. 4.2

Patient Care

4.2.1 Telepathology and Histologic Diagnosis Using Automated Whole-Slide Digitization

14

The issues encountered in using pathology images for patient care are more complex than those in pathology education, and only rudimentary progress has been made in this area. At the present, most interest concentrates on telepathology and value-added, image-enhanced specimen reporting. Telepathology is the sharing of microscopic images via a telecommunication device for remote primary diagnosis, expert consultation and consensus diagnosis, case conferencing, quality assurance, and education [127, 128]. The simplest form of telepathology is sending selected digital images of an interesting or perplexing case via an e-mail attachment to an expert scientist, a group of colleagues, or residents for educational purposes (static telepathology). Dynamic telepathology is more complex, and involves interactive imaging, where the consultant at a remote location controls the functions of a microscope at the transmitting site [129, 130]. First demonstrated in 1980, dynamic telepathology was of little interest to most scientists, since it required a microwave transmission link and special hardware and software [131]. However, in the 1990s, the technical innovations in digital imaging, computer software, and telecommunications greatly improved the practical feasibility of dynamic telepathology, and increasing numbers of pathology departments in the USA and other countries have reported applications in the remote diagnosis of frozen sections, surgical pathology, and cytopathology specimens [132–135]. More recently, the

http://www.youtube.com/watch?v¼ihaM3DvyUHE&feature¼related

446

Roger S. Riley et al.

advent of whole slide digitization and the virtual microscope slide is revolutionizing the practice of diagnostic pathology, with the virtual slides available for diagnostic evaluation or local/remote consultation via a network server or transportable media [127, 136]. This trend is in many ways analogous to diagnostic radiology, where digital imagers replaced silver halide film, and the resulting images are rapidly available and can be viewed and interpreted using PACS. Digital imaging is more difficult to incorporate into routine pathology workflow due to the continuing requirement for stained glass microscope slides and time delays in diagnostic interpretation attributable to the process of slide digitization. However, in systems developed at the Massachusetts General Hospital, robotic wholeslide imaging devices under the control of the laboratory information system (LIS) have been used to image bar-coded slides from the histology laboratory in a continuous or mini-batch flow. The resulting images, together with images of the specimen acquired in the gross room, are placed on a server and are available for review by the pathologist at a local or remote imaging workstation [63, 137]. If special stains are required, the resulting microscope slides are automatically imaged for further evaluation by the pathologist. Presently, the high-resolution images obtained from a 1  2 tissue specimen are equivalent to approximately 5000 Mpixel and require 0.5–1 GB of storage in a compressed JPEG or JPEG2000 state [6]. Further advances are rapidly expected in this field, particularly in the development of the optimal virtual slide workstation [27, 130, 138]. Unfortunately, unresolved issues of medical licensing and credentialing, patient confidentiality, reimbursement, insurance, malpractice liability and other medicolegal problems are currently delaying the more widespread use of telepathology [86, 91, 139– 142]. Pathologists performing diagnostic telepathology must be aware of the policies, regulations, and statutes of their institution and state, and those of the institution and state providing the images. Value-added pathology is the enhancement of the conventional written pathology report with supplemental data made possible through the use of computer technology [143, 144]. In addition to digital images, value-added enhancements include prognostic correlations, trend graphs, links to previous reports, expert reviews and opinions, clinical pathway information, and links to literature reports and Internet sites. Although anatomic scientists have not traditionally concerned themselves with information technology, the increasing competiveness of the market, pressure from technologically sophisticated colleagues in other fields, and other factors have led them to embrace image-enhanced reports, telepathology,

Digital Micrographs

447

and other modern technological innovations to improve patient care [111]. Although the integration of a few selected images into a report is of dubious value for patient care, the availability of large-scale, automated, bar-coded, multi-resolution imaging of entire slides into a networked database could dramatically alter the microscope-centered world of the surgical scientist and greatly improve their efficiency. Essentially freed from the microscope, scientists can simultaneously view multiple slides, compare different parts of the same slide concurrently, perform image analysis, annotate the slides with graphics or text, adjust the color, brightness, or other parameters of the slides to their preference. [111]. Current and previous patient data are immediately available for clinical consultations, patient care, educational conferences, and research applications throughout the institution without the need for the repeated archival and retrieval of glass slides. Peer review and professional consultations can be efficiently obtained, and expert consultations can be rapidly performed without the need for expensive recutting and transportation of glass microscope slides [111]. Digital cameras have been utilized from the hospital autopsy suite, forensic pathology laboratory, and surgical pathology gross room to document pathologic lesions and supplement verbal descriptions [70, 145–148]. In an academic autopsy, Belanger and coworkers found the digital camera to be practical, reliable, and cost-effective. With digital photography, the number of images taken increased nearly twofold per case, and the technology was readily accepted by both pathology residents and the technical staff [147]. In another study, the incorporation of color digital images into autopsy reports was perceived by clinicians to increase the value of the report and improve their understanding of the results [149]. The application of telepathology for the diagnosis of congenital cardiac malformations has been described [150]. 4.2.2 Computer-Assisted Image Analysis

The advent of immunohistochemistry (IHC) and related techniques in the early 1980s has profoundly changed the practice of anatomic pathology. These techniques permit the identification of a large and rapidly growing number of biological molecules of diagnostic and prognostic significance. The unaided visual interpretation of IHC-stained microscope slides stained was qualitative (i.e., negative or positive) for many years, when semiquantitative manual grading of staining intensity gradually became accepted for antigens such as HER2/neu [151–154]. Recent advances in image acquisition hardware and software technology have proven that decentralized computer-aided image analysis (CAIA) of IHC-stained slides were equivalent or superior to manual scoring, if standardized techniques of image acquisition were followed [153, 155, 156]. Although several biomarker companies have

448

Roger S. Riley et al.

achieved clearance by the US Food and Drug Administration (FDA) for the CAIA of breast biomarkers, this technique is still more time-consuming than manual CAIA interpretation. Substantial work remains to develop standardized image acquisition and analysis techniques, and to solve problems such as nonspecific staining and the interpretation of weak antigen expression [153]. A major advantage in the use of whole-slide images in the interpretation of special stains is that multiple images can be aligned and overlaid, permitting the interpretation of several markers in a selected region of the tissue, cell cluster, or even a single cell [153]. A computerized program for the automated determination of the Ki-67 labeling index in meningiomas was developed as a plug-in for the public domain software program ImageJ [157]. The mean labeling indices were not statistically different for the manual and automated methods, but the mean time effort for counting decreased from 374 s/image for the manual counts to 11 s/image for the automated index [157]. In neuropathology, automated computer-based systems have been developed for the grading of astrocytomas, the determination of Ki-67 labeling indices in meningiomas, and the classification of plaques and tangles [6, 157–160]. 4.2.3 Quality Assurance and Quality Control

Quality assurance in pathology is largely documented by peer review of glass microscope slides. In order to improve efficacy and decrease the extra workload associated with maintenance of the slides, several groups have evaluated digital images for intradepartmental or interfacility quality assurance [136]. In an early study, Cruz and coworkers captured up to 12 static digital images per case at an appropriate magnification, which another scientist archived for review. Image quality was inadequate in only 2.8% the cases, but the amount of time required by the referring scientist (4.5 min/ case) was the major disadvantage of the study [161]. In another early study, scientists evaluated semiautomated imaging of patient histopathology slides with bar coded labels as a technique to reduce the error rate in pathology [162]. The corrected report rate was reduced from 0.2% for conventional reports to 0.04% with the inclusion of images, presumably because the diagnostic errors were more obvious on proofreading and sign-out. Graham and collaborators utilized virtual slide telepathology to provide sameday quality assurance for academic subspecialty pathologists providing on-site surgical services at a small community hospital affiliated with a large academic teaching hospital [163]. The service permitted timely correction of diagnostic errors and improved the job satisfaction of the on-site pathologists. A potentially important application of digital photography for quality assurance and quality control in the laboratory is the documentation and teaching of medical procedures and techniques.

Digital Micrographs

449

With the ease and relatively low cost of digital photography, photographs or video clips can easily supplement written procedure manuals and quality assurance documents. This should reduce training time, increase the efficacy of the education process, and ultimately reduce medical errors. 4.3 Research and Scholarly Activity

Digital photography has greatly enhanced the ability of researchers to acquire many types of research data, as well as to prepare papers for presentation or publication. The advantages of the digital camera in research are similar to other applications and include rapid and cost-effective digital acquisition and the ability to immediately preview the results. In addition, the related techniques of quantitative image analysis provide powerful new, previously unavailable techniques of image analysis. For example, Bornfleth and collaborators used a one-chip true-color CCD camera with a triple-bandpass filter for comparative genomic hybridization imaging on metaphase chromosomes through the simultaneous registration of the three dyes Texas Red, FITC, and DAPI [164]. In a related study, van Der Laak et al. developed a standardized model for stain recognition and analysis using RGB data captured from immuno-doublestained tissue sections by transmitted light microscopy [165]. Through this technique, one can obtain RGB image data and optical density values from the tissue sections. The RGB data is transformed to a hue-saturation-intensity (HSI) color model, which is applied to the optical density values to obtain data regarding the density of each stain (hue-saturation-density transform). Digital imaging is used routinely in fluorescence in situ hybridization (FISH) testing performed on monolayer smear/cytospin cell preparations and thin sections of formalin-fixed, paraffinembedded tissue, the latter requiring a fluorescent microscope outfitted with motorized Z-stacker to allow for image capture in multiple sequential focal planes. Black-and-white, high-resolution Cohu CCD cameras are extremely helpful in fluorescence digital imaging, allowing for detection of dim signals under appropriate filter sets, with the resulting images reconstituted using imaging software-generated pseudocolors [166, 167]. A number of dedicated FISH imaging systems incorporate software allowing for optimal capture of the tiny signals in concert with suppression of background autofluorescence. Digital photography and automated image analysis has been used for the identification of fungal species growing on agar plates, and for the automated identification of tubercle bacilli in sputum [168, 169]. Ernsting used image analysis of digital pictures of the conjunctiva to accurately predict hemoglobin concentration, while Vrolijk and collaborators were able to detect in situ hybridization probe spots in interphase nuclei using centromere-specific probes with a non-cooled video-rated CCD camera [170, 171]. Numerous other examples of the innovative use of the digital camera can be found in scientific literature.

450

Roger S. Riley et al.

Digital photography also facilitates the creation of scientific publications and presentations [106]. Almost all journals nowadays accept digital images in file format, and inexpensive high-resolution prints can be prepared from those that do not accept files. Chen found that the average cost of a 5  7 in. black-and-white publication-quality print could be reduced from $8.50 to $1.00 with a printer, and a 30-panel, 4  8 foot (1.2  2.4 m) standardsized poster, and could be [manufactured] for approximately $100 [172]. Frank et al. evaluated the megapixel digital camera for producing publication-quality illustrations, while the impact of file size on quality for the production of publication quality black and white images of histopathologic specimens was evaluated by Barker et al. [173, 174]. The authors recommended a minimum file size of 2.8 Mb for a publication-quality 5  7 in. print at 100 and a 1.7 Mb file for a similar image at 460.

5

The Future of Digital Photography in Pathology The digital revolution will continue to bring continued technical innovations and changes to the profession of pathology, as well as to other health sciences. Since neuroscience is a highly visual scientific field, there are many advantages of the digicam over the conventional film camera in the research laboratory, classroom, and conference room, and more recently in the pathology gross room and frozen section suite. In the near future, the digital camera will be an essential part of pathology gross and microscopic workstation, and further developments in the engineering of laboratory management systems will thoroughly integrate digital images into specimen reporting and quality assurance. There will be continued advances in the technology of image enhancement, and scientists will routinely use various forms of quantitative image analysis [175]. In addition, globalization of pathology will lead to large network-based digital histology atlases and the rapid availability of expert assistance. Digital storage technology will continue to develop, with large, centrally located, professionally maintained file servers [176]. Advanced digital pattern recognition systems will ultimately be available to supplement the diagnostic abilities of the scientist.

References 1. Montalto MC (2008) Pathology RE-imagined. The history of digital radiology and the future of anatomic pathology. Arch Pathol Lab Med 132(5):764–765. https:// www.ncbi.nlm.nih.gov/pubmed/18466020 2. Soenksen D (2009) Digital pathology at the crossroads of major health care trends:

corporate innovation as an engine for change. Arch Pathol Lab Med 133(4):555–559. https://www.ncbi.nlm.nih.gov/pubmed/ 19391651 3. Russ JC (2006) The image processing handbook, 5th edn. CRC Press, Boca Raton. https://doi.org/10.1201/9780203881095

Digital Micrographs 4. Wu Q, Merchant F, Castleman K (2008) Microscope image processing, 1st edn. Academic Press, Burlington. ISBN: 9780123725783 5. Ford JC, Pinder KE, Ovalle WK, Li CH (2008) Pathology education in a multisite urban/rural distributed curriculum. Hum Pathol 39(6):811–816. https://doi.org/10. 1016/j.humpath.2008.02.009 6. Guzman M, Judkins AR (2009) Digital pathology: a tool for 21st century neuropathology. Brain Pathol 19(2):305–316. https://doi.org/10.1111/j.1750-3639. 2009.00264.x 7. Matsumoto B (2010) Practical digital photomicrography: photography through the microscope for the life sciences, 1st edn. Rocky Nook, Santa Barbara. ISBN: 9781933952079 8. Murphy DB, Davidson MW (2012) Fundamentals of digital imaging (Chapter 17). In: Fundamentals of light microscopy and electronic imaging, 2nd edn. Wiley-Blackwell, Hoboken, pp 259–282. https://doi.org/10. 1002/9781118382905.ch17 9. Spring KR (2007) Cameras for digital microscopy. Methods Cell Biol 81:171–186. https://doi.org/10.1016/S0091-679X(06) 81010-1 10. Spring K, Davidson MW (online) Fundamentals of digital imaging. Florida State University/Microscopy University. https://www. microscopyu.com/digital-imaging/fun damentals-of-digital-imaging 11. Boyle WS, Smith GE (1976) The inception of charge-coupled devices. IEEE Trans Elect Dev 23(7):661–663. https://doi.org/10. 1109/T-ED.1976.18470 12. Brumfiel G (2009) Nobel prize in physics awarded to light pioneers. Nature 461 (7265):707. https://doi.org/10.1038/ 461707a 13. Spring KR (2000) Scientific imaging with digital cameras. BioTechniques 29(1):70–76. https://doi.org/10.2144/00291bi01 14. Spring KR, Fellers TJ, Davidson MW (online) Introduction to charge-coupled devices (CCDs). Florida State University/Microscopy University. https://www.microscopyu. com/digital-imaging/introduction-tocharge-coupled-devices-ccds 15. Hand WG (online) Practical guide to digital imaging for microscopy. http://www. bioresearchonline.com/product.mvc/Practi cal-Guide-to-Digital-Imaging-for-Micros0002

451

16. Deziel DJ (2006) Documentation. In: ScottConner CEH (ed) The SAGES manual (Fundamentals of laparoscopy, thoracoscopy, and GI endoscopy), 2nd edn. Springer, New York, pp 78–83. https://doi.org/10.1007/0-38730485-1_9 17. Grey T (2006) Color confidence: the digital photographer’s guide to color, 2nd edn. Wiley, Indianapolis. ISBN: 978-0471786160 18. McHugh S (online) Color management: Understanding color spaces. https://www. cambridgeincolour.com/tutorials/colorspaces.htm 19. Durcan TM, Hinchcliffe EH (2007) Digital image files in light microscopy. Methods Cell Biol 81:315–333. https://doi.org/10.1016/ S0091-679X(06)81015-0 20. Bueno G, Gonzalez R, Deniz O, Gonzalez J, Garcia-Rojo M (2008) Colour model analysis for microscopic image processing. Diagn Pathol 3(Suppl 1):S18. https://doi.org/10. 1186/1746-1596-3-S1-S18 21. Evening M (2001) Adobe Photoshop 6.0 for photographers, 1st edn. Focal Press, Oxford. ISBN: 978-0240516332 22. Della Mea V, Bortolotti N, Beltrami CA (2008) A survey on non specialized off-theshelf JPEG2000 viewers for digital microscopy use. Diagn Pathol 3(Suppl 1):S20. https://doi.org/10.1186/1746-1596-3-S1S20 23. Tuominen VJ, Isola J (2007) The application of JPEG2000 in virtual microscopy. J Digit Imaging 22(3):250–258. https://doi.org/ 10.1007/s10278-007-9090-z 24. Graham RN, Perriss RW, Scarsbrook AF (2005) DICOM demystified: a review of digital file formats and their use in radiological practice. Clin Radiol 60(11):1133–1140. https://doi.org/10.1016/j.crad.2005.07. 003 25. Kahn CE Jr, Carrino JA, Flynn MJ, Peck DJ, Horii SC (2007) DICOM and radiology: past, present, and future. J Am Coll Radiol 4 (9):652–657. https://doi.org/10.1016/j. jacr.2007.06.004 26. Le Bozec C, Henin D, Fabiani B, Schrader T, Garcia-Rojo M, Beckwith B (2007) Refining DICOM for pathology--progress from the IHE and DICOM pathology working groups. Stud Health Technol Inform 129 (Pt 1):434–438. https://www.ncbi.nlm.nih. gov/pubmed/17911754 27. Vannier MW (2009) Medical imaging workstations: what is missing and what is coming? Arch Pathol Lab Med 133(4):542–546.

452

Roger S. Riley et al.

https://www.ncbi.nlm.nih.gov/pubmed/ 19391647 28. Lee E, Kim HK, Kim I (2000) Anatomic pathology image capture using a consumertype digital camera. Am J Surg Pathol 24 (7):1034–1035. https://doi.org/10.1097/ 00000478-200007000-00018 29. Tse CC (1999) Anatomic pathology image capture using a consumer-type digital camera. Am J Surg Pathol 23(12):1555–1558. https://doi.org/10.1097/00000478199912000-00014 30. Peres MR (1998) An overview of some professional digital cameras and their use on the light microscope. J Biol Photogr 66 (1):13–20. https://www.ncbi.nlm.nih.gov/ pubmed/9700933 31. Rojo MG, Garcia GB, Mateos CP, Garcia JG, Vicente MC (2006) Critical comparison of 31 commercially available digital slide systems in pathology. Int J Surg Pathol 14 (4):285–305. https://doi.org/10.1177/ 1066896906292274 32. Felten CL, Strauss JS, Okada DH, Marchevsky AM (1999) Virtual microscopy: high resolution digital photomicrography as a tool for light microscopy simulation. Hum Pathol 30(4):477–483. https://doi.org/10.1016/ S0046-8177(99)90126-0 33. Hinchcliffe EH, Sluder G (2007) Working with classic video. Methods Cell Biol 81:141–157. https://doi.org/10.1016/ S0091-679X(06)81008-3 34. Bergeron BP (1999) Considering a digital video camera? Must-have, might-want, and what’s that features. Postgrad Med 106 (5):33–34. https://doi.org/10.3810/pgm. 1999.10.15.759 35. Microscopy camera imaging fields (ophthalmic, dentistry, neurosurgery etc). Optronics, Goleta (CA, USA). https://www.optronics. com/microscope-camera-video-imagingapplications.html 36. Bouton GD, Kubicek G, Bouton BM (1998) Inside Photoshop 5. New Riders Publishing, Indianapolis. ISBN: 978-1562058845 37. Ratner D (2000) Real photographic prints from digital images. Dermatol Surg 26 (8):799–800. https://doi.org/10.1046/j. 1524-4725.2000.00114.x 38. Flynn BO, Davidson MW (online) CCD resolution for optical microscopy. Florida State University/Microscopy University. https:// www.microscopyu.com/tutorials/ccd-resolu tion-for-optical-microscopy

39. Vetter JP (1974) A systematic approach to colour photomicrography. 2. Cameras and photographic techniques. Med Biol Illus 24 (3):140–152. https://www.ncbi.nlm.nih. gov/pubmed/4468367 40. Gupta PK, Hinsch J (1983) Essentials of light microscopy and photomicrography. Pathol Annu 18(Pt1):331–358. https://www.ncbi. nlm.nih.gov/pubmed/6877877 41. Vetter JP (1984) Five common problems in color photomicrography. Pathologist 38 (3):163–170. https://www.ncbi.nlm.nih. gov/pubmed/10265487 42. Vetter JP (1987) Photomicrography: a translation into the vernacular. Part II-the specimen stage, the specimen, and the imageforming system. J Biol Photogr 55 (4):135–142. https://www.ncbi.nlm.nih. gov/pubmed/3436919 43. Vetter JP (1987) Photomicrography: a translation into the vernacular. Part I-the illuminating system. J Biol Photogr 55(3):79–85. https://www.ncbi.nlm.nih.gov/pubmed/ 3680140 44. Vetter JP (1988) Photomicrography: a translation into the vernacular. Part IV: producing high quality photomicrographs. J Biol Photogr 56(3):89–108. https://www.ncbi. nlm.nih.gov/pubmed/3182576 45. Vetter JP (1988) Photomicrography: a translation into the vernacular. Part III-the photographic system. J Biol Photogr 56(2):53–64. https://www.ncbi.nlm.nih.gov/pubmed/ 3391977 46. Vetter JP (1992) Biomedical photography, 1st edn. Focal Press, Woburn. ISBN: 9780240800844 47. Scott ML (1996) Exercises and calibrations in biomedical photomicrography. J Audiov Media Med 19(2):69–76. https://doi.org/ 10.3109/17453059609023008 48. Foster B (1997) Optimizing light microscopy for biological and clinical labs, 1st edn. Kendall/Hunt Publishing Company, Dubuque. ISBN: 978-0787235383 49. Goldstein DJ (1999) Understanding the light microscope: a computer-aided introduction, 1st edn. Academic Press, London and San Diego. ISBN: 978-0122886607 50. Wingate RJ (1999) Microscopy and photomicrography techniques. In: Sharpe PT, Mason I (eds) Molecular embryology, Methods Mol Biol, vol 97. Humana Press (Springer), Totowa (NJ, USA), pp 711–733. https:// doi.org/10.1385/1-59259-270-8:711

Digital Micrographs 51. Rost FWD, Oldfield RJ (2000) Photography with a microscope, 1st edn. Cambridge University Press, Cambridge. ISBN: 9780521770965 52. Bennin B (2001) Photomicrography for the multitude. Dermatol Online J 7(1):21. https://www.ncbi.nlm.nih.gov/pubmed/ 11354054 53. Salmon ED, Canman JC (2002) Proper alignment and adjustment of the light microscope. Curr Protoc Immunol 48(1):Unit 21. https://doi.org/10.1002/0471142735. im2101s48 54. Keller HE (2007) Proper alignment of the microscope. Methods Cell Biol 81:43–53. https://doi.org/10.1016/S0091-679X(06) 81003-4 55. Hill D (2001) Personal communication 56. Kennedy D (1992) A low-cost vibration damping platform for photomicrography. J Biol Photogr 60(2):71. https://www.ncbi. nlm.nih.gov/pubmed/1577678 57. Zieler HW (1992) Photomicrography versus vibration. J Biol Photogr 60(1):19. https:// www.ncbi.nlm.nih.gov/pubmed/1551821 58. Maude RJ, Koh GC, Silamut K (2008) Taking photographs with a microscope. Am J Trop Med Hyg 79(3):471–472. https://www.ncbi. nlm.nih.gov/pubmed/18784246 59. Hinsch J (2007) Mating cameras to microscopes. Methods Cell Biol 81:55–61. https://doi.org/10.1016/S0091-679X(06) 81004-6 60. Nordberg JJ, Sluder G (2013) Practical aspects of adjusting digital cameras. Methods Cell Biol 114:151–162. https://doi.org/10. 1016/B978-0-12-407761-4.00007-5 61. Abramowitz M, Davidson MW. Troubleshooting errors and faults in photomicrography. http://micro.magnet.fsu. edu/primer/photomicrography/errorindex. html 62. Kayser K, Gortler J, Metze K et al (2008) How to measure image quality in tissuebased diagnosis (diagnostic surgical pathology). Diagn Pathol 3(Suppl 1):S11. https:// doi.org/10.1186/1746-1596-3-S1-S11 63. Gilbertson JR, Ho J, Anthony L, Jukic DM, Yagi Y, Parwani AV (2006) Primary histologic diagnosis using automated whole slide imaging: a validation study. BMC Clin Pathol 6:4. https://doi.org/10.1186/1472-6890-6-4 64. Yagi Y, Gilbertson JR (2008) The importance of optical optimization in whole slide imaging (WSI) and digital pathology imaging. Diagn Pathol 3(Suppl 1):S1. https://doi.org/10. 1186/1746-1596-3-S1-S1

453

65. Burns BF (1997) Creating low-power photomicrographs using a 35 mm digital slide scanner. Am J Surg Pathol 21(7):865–866. https://doi.org/10.1097/00000478199707000-00017 66. Azumi N (1998) Creating low-power photomicrographs using a 35-mm digital slide scanner. Am J Surg Pathol 22(7):908. https://doi. org/10.1097/00000478-19980700000019 67. Gebert A, Werner K, Posselt W (1998) Use of a digital film scanner to enhance low-power bright field photomicrography. Anat Embryol (Berl) 198(5):435–438. https://doi.org/10. 1007/s004290050194 68. Ventura L, Leocata P, Colimberti P (1999) Digital scanning of histologic sections. Am J Surg Pathol 23(11):1435. https://doi.org/ 10.1097/00000478-199911000-00020 69. Ventura L, Chiominto A, Colimberti P, Sarra M, Dal Mas A, Leocata P (2000) Creating low-power photomicrographs by digital scanning of histological sections. Pathologica 92(1):9–12. https://www.ncbi.nlm.nih.gov/ pubmed/10842994 70. Leong AS, Visinoni F, Visinoni C, Milios J (2000) An advanced digital image-capture computer system for gross specimens: a substitute for gross description. Pathology 32 (2):131–135. https://doi.org/10.1080/ 003130200104385 71. Mai KT, Stinson WA, Swift J, Burns BF, Perkins DG (2001) Creating digital images of pathology specimens by using a flatbed scanner. Histopathology 39(3):323–325. https://doi.org/10.1046/j.1365-2559. 2001.1181c.x 72. Matthews TJ, Denney PA (2001) Digital imaging of surgical specimens using a wet scanning technique. J Clin Pathol 54 (4):326–327. https://doi.org/10.1136/jcp. 54.4.326 73. Kilbourne S (1999) A primer on digital imaging--post production for still photography: Part I. 1991. J Biol Photogr 67(1–4):49–54 [originally 59(2):43-48] https://www.ncbi. nlm.nih.gov/pubmed/10914282 74. Kilbourne S, Dodd R (1991) A primer on digital imaging--post production for still photography: part 2. J Biol Photogr 59 (3):125–132. https://www.ncbi.nlm.nih. gov/pubmed/1917825 75. Kilbourne S, Dodd R (1992) A primer on digital imaging--post production for still photography: part 3. J Biol Photogr 60(1):1–10. https://www.ncbi.nlm.nih.gov/pubmed/ 1551818

454

Roger S. Riley et al.

76. Ang T (2000) Silver pixels: an introduction to the digital darkroom, 1st edn. Amphoto, New York. ISBN: 978-0817458898 77. Davies A, Fennessy P (2001) Digital imaging for photographers, 4th edn. Focal Press, Oxford and Wolburn. ISBN: 9780240515908 78. Collins TJ (2007) ImageJ for microscopy. BioTechniques 43(Suppl 1):25–30. https:// doi.org/10.2144/000112517 79. Papadopulos F, Spinelli M, Valente S (2007) Common tasks in microscopic and ultrastructural image analysis using ImageJ. Ultrastruct Pathol 31(6):401–407. https://doi.org/10. 1080/01913120701719189 80. Scalettar BA, Swedlow JR, Sedat JW, Agard DA (1996) Dispersion, aberration and deconvolution in multi-wavelength fluorescence images. J Microsc 182(Pt1):50–60. https:// doi.org/10.1046/j.1365-2818.1996. 122402.x 81. Swedlow JR (2013) Quantitative fluorescence microscopy and image deconvolution. Methods Cell Biol 114:407–26. https://doi.org/ 10.1016/B978-0-12-407761-4.00017-8 82. Wallace W, Schaefer LH, Swedlow JR (2001) A workingperson’s guide to deconvolution in light microscopy. BioTechniques 31 (5):1076–1097. https://doi.org/10.2144/ 01315bi01 83. Parry-Hill MJ, Davidson MW (online) Background subtraction toolkit. Florida State University/Microscopy University. https:// micro.magnet.fsu.edu/primer/ digitalimaging/backgrounddownload.html 84. Spring KR, Russ JC, Parry-Hill MJ et al (online) Background subtraction. Florida State University/Microscopy University. https://micro.magnet.fsu.edu/primer/java/ digitalimaging/processing/bac kgroundsubtraction/index.html 85. Barry CJ, Yogesan K, Constable IJ, Eikelboom RH (1999) A case for electronic manipulation of medical images? J Audiov Media Med 22(1):15–20. https://doi.org/10. 1080/014051199102313 86. Dierks C (2000) Legal aspects of telepathology. Anal Cell Pathol 21(3-4):563904. https://doi.org/10.1155/2000/563904 87. Hayden JE (2000) Digital manipulation in scientific images: some ethical considerations. J Biocommun 27(1):11–19. https://www. ncbi.nlm.nih.gov/pubmed/10916744 88. Stanberry B (2000) Telemedicine: barriers and opportunities in the 21st century. J Intern Med 247(6):615–628. https://doi.org/10. 1046/j.1365-2796.2000.00699.x

89. Tsuchihashi Y, Okada Y, Ogushi Y, Mazaki T, Tsutsumi Y, Sawai T (2000) The current status of medicolegal issues surrounding telepathology and telecytology in Japan. J Telemed Telecare 6(Suppl 1):S143–S145. https://doi.org/10.1258/ 1357633001934465 90. Beals TF (2001) Digital imaging in anatomic pathology. Lab Med 32(6):327–331. https:// doi.org/10.1309/U23N-6N0T-795T-X65B 91. Suvarna SK, Ansary MA (2001) Histopathology and the ’third great lie’. When is an image not a scientifically authentic image? Histopathology 39(5):441–446. https://doi.org/10. 1046/j.1365-2559.2001.01312.x 92. Pritt BS, Gibson PC, Cooper K (2003) Digital imaging guidelines for pathology: a proposal for general and academic use. Adv Anat Pathol 10(2):96–100. https://doi.org/10. 1097/00125480-200303000-00004 93. Pritt B, Gibson P, Cooper K (2006) Digital manipulation of pathologic images. Int J Surg Pathol 14(1):107. https://doi.org/10. 1177/106689690601400121 94. Gulbins J, Gulbins R (2009) Stitching— increasing image coverage. In: Photographic multishot techniques: high dynamic range, super-resolution, extended depth of field, stitching, 1st edn. Rocky Nook, Santa Barbara, pp 97–134. ISBN: 978-1933952383 95. Benosman R, Kang SB (eds) (2001) Panoramic vision, 1st edn. Springer, New York. https://doi.org/10.1007/978-1-47573482-9 96. Crudele M, Clapworthy GJ, Dong F, Krokos MA, Vasilonikolidakis N, Salcito G (1999) Accessing a www reference library of 3D models of pathological organs to support medical education. Stud Health Technol Inform 68:532–537. https://www.ncbi.nlm.nih. gov/pubmed/10724945 97. Russ JC (2006) 3-D image visualization (Chapter 13). In: The image processing handbook, 5th edn. CRC Press, Boca Raton, pp 667–728. https://doi.org/10.1201/ 9780203881095 98. Gulbins J, Gulbins R (2009) Focus stacking— maximizing depth of field. In: Photographic multishot techniques: high dynamic range, super-resolution, extended depth of field, stitching, 1st edn. Rocky Nook, Santa Barbara, pp 53–96. ISBN: 978-1933952383 99. Lu J, Fiala JC, Lichtman JW (2009) Semiautomated reconstruction of neural processes from large numbers of fluorescence images. PLoS One 4(5):e5655. https://doi.org/10. 1371/journal.pone.0005655

Digital Micrographs 100. Kopel H, Meshulam M, Mizrahi A (2009) Three-dimensional distribution patterns of newborn neurons in the adult olfactory bulb. J Neurosci Methods 182(2):189–194. https://doi.org/10.1016/j.jneumeth.2009. 06.009 101. Kamali M, Day LJ, Brooks DH, Zhou X, O’Malley DM (2009) Automated identification of neurons in 3D confocal datasets from zebrafish brainstem. J Microsc 233 (1):114–131. https://doi.org/10.1111/j. 1365-2818.2008.03102.x 102. Oberlaender M, Bruno RM, Sakmann B, Broser PJ (2007) Transmitted light brightfield mosaic microscopy for three-dimensional tracing of single neuron morphology. J Biomed Opt 12(6):064029. https://doi. org/10.1117/1.2815693 103. Faulkner A (1999) Electronic publishing guide. In: The essential resource for electronic publishing, 1st edn. Adobe Press, San Jose. ISBN: 978-1568304694 104. Airey T (2001) Creative digital printmaking: A photographer’s guide to professional desktop printing, 1st edn. Watson-Guptill Publications, New York, ISBN: 978-0817437268 105. Johnson H (2004) Mastering digital printing, 2nd edn. Course Technology PTR, Boston. ISBN: 978-1592004317 106. Murphy DB, Davidson MW (2012) Digital imaging processing (Chapter 18). In: Fundamentals of light microscopy and electronic imaging, 2nd edn. Wiley-Blackwell, Hoboken, pp 415-449. https://doi.org/10. 1002/9781118382905.ch18 107. Vargas C (2002) Preparing scientific images for publication and display. In: Gallin JI (ed) Principles and practice of clinical research, 1st edn. Academic Press (Elsevier), San Diego and London, pp 427–437. ISBN: 978-0122740657 108. Uniform requirements for manuscripts submitted to biomedical journals: Writing and editing for biomedical publication. American College of Physicians, 2007. http://www. icmje.org/recommendations 109. Furness PN (1997) The use of digital images in pathology. J Pathol 183(3):253–263. https://www.ncbi.nlm.nih.gov/pubmed/ 9422979 110. Saltz JH (2000) Digital pathology-the big picture. Hum Pathol 31(7):779–780. https://www.ncbi.nlm.nih.gov/pubmed/ 10923911 111. Sholehvar D (2001) The age of digital imaging. Advance for Administrators of the Laboratory 10:51–56

455

112. Weinberg DS (1997) Digital imaging as a teaching tool for pathologists. Clin Lab Med 17(2):229–244. https://doi.org/10.1016/ S0272-2712(18)30220-8 113. Banjanovic B, Masic I (1999) Telemedicine and telematics in medical education. Med Arh 53(3):21–23. https://www.ncbi.nlm. nih.gov/pubmed/10870618 114. Lakatos J, Bodor T, Zidarics Z, Nagy J (2000) Data processing of digital recordings of microscopic examination of urinary sediment. Clin Chim Acta 297(1–2):225–237. https:// doi.org/10.1016/S0009-8981(00)00249-7 115. Afrin LB (1999) Web access to the American Society of Hematology slide bank. Blood 93 (7):2425–2426. https://www.ncbi.nlm.nih. gov/pubmed/10215352 116. Evans JA, Wagner U, Santos CM, Hennighausen L (2000) The interactive web-based histology atlas system. Oncogene 19 (8):989–991. https://doi.org/10.1038/sj. onc.1203279 117. Feit J, Matyska L, Ulman V et al (2008) Virtual microscope interface to high resolution histological images. Diagn Pathol 3(Suppl 1): S10. https://doi.org/10.1186/1746-15963-S1-S10 118. Dee FR (2009) Virtual microscopy in pathology education. Hum Pathol 40 (8):1112–1121. https://doi.org/10.1016/j. humpath.2009.04.010 119. Mikula S, Trotts I, Stone JM, Jones EG (2007) Internet-enabled high-resolution brain mapping and virtual microscopy. NeuroImage 35(1):9–15. https://doi.org/10. 1016/j.neuroimage.2006.11.053 120. Krippendorf BB, Lough J (2005) Complete and rapid switch from light microscopy to virtual microscopy for teaching medical histology. Anat Rec B New Anat 285(1):19–25. https://doi.org/10.1002/ar.b.20066 121. Scoville SA, Buskirk TD (2007) Traditional and virtual microscopy compared experimentally in a classroom setting. Clin Anat 20 (5):565–570. https://doi.org/10.1002/ca. 20440 122. Braun MW, Kearns KD (2008) Improved learning efficiency and increased student collaboration through use of virtual microscopy in the teaching of human pathology. Anat Sci Educ 1(6):240–246. https://doi.org/10. 1002/ase.53 123. Koch LH, Lampros JN, Delong LK, Chen SC, Woosley JT, Hood AF (2009) Randomized comparison of virtual microscopy and traditional glass microscopy in diagnostic accuracy among dermatology and pathology

456

Roger S. Riley et al.

residents. Hum Pathol 40(5):662–667. https://doi.org/10.1016/j.humpath.2008. 10.009 124. Farah CS, Maybury T (2009) Implementing digital technology to enhance student learning of pathology. Eur J Dent Educ 13 (3):172–178. https://doi.org/10.1111/j. 1600-0579.2009.00570.x 125. Coleman R (2009) Can histology and pathology be taught without microscopes? The advantages and disadvantages of virtual histology. Acta Histochem 111(1):1–4. https://doi.org/10.1016/j.acthis.2008.09. 003 126. Bruch LA, De Young BR, Kreiter CD, Haugen TH, Leaven TC, Dee FR (2009) Competency assessment of residents in surgical pathology using virtual microscopy. Hum Pathol 40(8):1122–1128. https://doi.org/ 10.1016/j.humpath.2009.04.009 127. Weinstein RS, Graham AR, Richter LC (2009) Overview of telepathology, virtual microscopy, and whole slide imaging: prospects for the future. Hum Pathol 40 (8):1057–1069. https://doi.org/10.1016/j. humpath.2009.04.006 128. O’Malley DP (2008) Practical applications of telepathology using morphology-based anatomic pathology. Arch Pathol Lab Med 132 (5):743–744. https://www.ncbi.nlm.nih. gov/pubmed/18466015 129. Mencarelli R, Marcolongo A, Gasparetto A (2008) Organizational model for a telepathology system. Diagn Pathol 3(Suppl 1): S7. https://doi.org/10.1186/1746-1596-3S1-S7 130. Krupinski EA (2009) Virtual slide telepathology workstation of the future: lessons learned from teleradiology. Hum Pathol 40 (8):1100–1111. https://doi.org/10.1016/j. humpath.2009.04.011 131. Wells CA, Sowter C (2000) Telepathology: a diagnostic tool for the millennium? J Pathol 191(1):1–7. https://pubmed.ncbi.nlm.nih. gov/10767711 132. Banach L, Stepien A, Schneider J, Wichrzycka-Lancaster E (2008) Dynamic active telepathology over National Health Laboratory service network, South Africa: feasibility study using Nikon Coolscope. Diagn Pathol 3(Suppl 1):S3. https://doi. org/10.1186/1746-1596-3-S1-S3 133. Della Mea V, Cataldi P, Pertoldi B, Beltrami CA (2000) Combining dynamic and static robotic telepathology: a report on 184 consecutive cases of frozen sections, histology

and cytology. Anal Cell Pathol 20(1):33–39. https://doi.org/10.1155/2000/734943 134. Weisz-Carrington P, Blount M, Kipreos B et al (1999) Telepathology between Richmond and Beckley veterans affairs hospitals: report on the first 1000 cases. Telemed J 5 (4):367–373. https://doi.org/10.1089/ 107830299311934 135. Evans AJ, Chetty R, Clarke BA et al (2009) Primary frozen section diagnosis by robotic microscopy and virtual slide telepathology: the university health network experience. Hum Pathol 40(8):1070–1081. https://doi. org/10.1016/j.humpath.2009.04.012 136. Rocha R, Vassallo J, Soares F, Miller K, Gobbi H (2009) Digital slides: present status of a tool for consultation, teaching, and quality control in pathology. Pathol Res Pract 205 (11):735–741. https://doi.org/10.1016/j. prp.2009.05.004 137. Gilbertson J, Yagi Y (2008) Histology, imaging and new diagnostic work-flows in pathology. Diagn Pathol 3(Suppl 1):S14. https:// doi.org/10.1186/1746-1596-3-S1-S14 138. Wienert S, Beil M, Saeger K, Hufnagl P, Schrader T (2009) Integration and acceleration of virtual microscopy as the key to successful implementation into the routine diagnostic process. Diagn Pathol 4:3. https://doi.org/10.1186/1746-1596-4-3 139. Ranson D (2007) Telemedicine and the law. J Law Med 15(3):356–359. https://www.ncbi. nlm.nih.gov/pubmed/18251419 140. Hill K (2006) Consent, confidentiality and record keeping for the recording and usage of medical images. J Vis Commun Med 29 (2):76–79. https://doi.org/10.1080/ 01405110600863365 141. Scheinfeld N (2004) Photographic images, digital imaging, dermatology, and the law. Arch Dermatol 140(4):473–476. https:// doi.org/10.1001/archderm.140.4.473 142. Leung ST, Kaplan KJ (2009) Medicolegal aspects of telepathology. Hum Pathol 40 (8):1137–1142. https://doi.org/10.1016/j. humpath.2009.01.009 143. Hahn AW, Leon MA, Klein-Leon S et al (1997) Delivery of laboratory data with world wide web technology. Biomed Sci Instrum 33:252–256. https://www.ncbi. nlm.nih.gov/pubmed/9731367 144. Matsen JM (1988) The regionalization of laboratory services at the University of Utah Medical Center. Associated regional and university pathologists Inc (ARUP). Arch Pathol Lab Med 112(9):957–959. https://doi.org/ 10.1038/461707a

Digital Micrographs 145. Oliver WR (1998) Image processing in forensic pathology. Clin Lab Med 18(1):151–180. https://doi.org/10.1016/S0272-2712(18) 30185-9 146. Cruz D, Seixas M (1999) A surgical pathology system for gross specimen examination. Proc AMIA Symp 32(12):236–240. https:// www.ncbi.nlm.nih.gov/pubmed/10566356 147. Belanger AJ, Lopes AE, Sinard JH (2000) Implementation of a practical digital imaging system for routine gross photography in an autopsy environment. Arch Pathol Lab Med 124(1):160–165. https://www.ncbi.nlm.nih. gov/pubmed/10629152 148. Blitzer HL, Jacobia J (2002) Forensic digital imaging and photography, 1st edn. Academic Press, London and San Diego. ISBN: 9780121064112 149. Pritt B, Gibson P, Cooper K, Hardin N (2004) What is a picture worth? Digital imaging applications in autopsy reports. Arch Pathol Lab Med 128(11):1247–1250. https://doi.org/10.1038/461707a 150. Tennstedt C, Sunkel-Wehrstedt K, Vogel M, Hufnagl P (2000) Diagnosis of congenital heart malformations - possibilities for the employment of telepathology. Anal Cell Pathol 21(3–4):768578 (pp 229-235). https://doi.org/10.1155/2000/768578 151. Feldman MD (2008) Beyond morphology: whole slide imaging, computer-aided detection, and other techniques. Arch Pathol Lab Med 132(5):758–763. https://doi.org/10. 1155/2000/768578 152. Skaland I, Ovestad I, Janssen EA et al (2008) Digital image analysis improves the quality of subjective HER-2 expression scoring in breast cancer. Appl Immunohistochem Mol Morphol 16(2):185–190. https://www.ncbi.nlm. nih.gov/pubmed/18227722 153. Friedberg RC, Pantanowitz L (2009) Practice evolution: decentralized computer-assisted immunohistochemical image analysis. Arch Pathol Lab Med 133(4):597–600. https:// www.ncbi.nlm.nih.gov/pubmed/19391658 154. Bernardo V, Lourenco SQ, Cruz R et al (2009) Reproducibility of immunostaining quantification and description of a new digital image processing procedure for quantitative evaluation of immunohistochemistry in pathology. Microsc Microanal 15 (4):353–365. https://doi.org/10.1017/ S1431927609090710 155. Fine JL, Grzybicki DM, Silowash R et al (2008) Evaluation of whole slide image immunohistochemistry interpretation in challenging prostate needle biopsies. Hum Pathol

457

39(4):564–572. https://doi.org/10.1016/j. humpath.2007.08.007 156. Turbin DA, Leung S, Cheang MC et al (2008) Automated quantitative analysis of estrogen receptor expression in breast carcinoma does not differ from expert pathologist scoring: a tissue microarray study of 3,484 cases. Breast Cancer Res Treat 110 (3):417–426. https://doi.org/10.1007/ s10549-007-9736-z 157. Kim YJ, Romeike BF, Uszkoreit J, Feiden W (2006) Automated nuclear segmentation in the determination of the Ki-67 labeling index in meningiomas. Clin Neuropathol 25 (2):67–73. https://www.ncbi.nlm.nih.gov/ pubmed/16550739 158. Glotsos D, Tohka J, Ravazoula P, Cavouras D, Nikiforidis G (2005) Automated diagnosis of brain tumours astrocytomas using probabilistic neural network clustering and support vector machines. Int J Neural Syst 15(1–2):1–11. https://doi.org/10. 1142/S0129065705000013 159. Chubb C, Inagaki Y, Sheu P et al (2006) BioVision: an application for the automated image analysis of histological sections. Neurobiol Aging 27(10):1462–1476. https:// doi.org/10.1016/j.neurobiolaging.2005. 08.023 160. Glotsos D, Kalatzis I, Spyridonos P et al (2008) Improving accuracy in astrocytomas grading by integrating a robust least squares mapping driven support vector machine classifier into a two level grade classification scheme. Comput Methods Prog Biomed 90 (3):251–261. https://doi.org/10.1016/j. cmpb.2008.01.006 161. Cruz D, Valenti C, Dias A, Seixas M, Schmitt F (2001) Digital image documentation for quality assessment. Arch Pathol Lab Med 125(11):1430–1435. https://www.ncbi.nlm. nih.gov/pubmed/11697997 162. McClanin S (2001) Error reduction in surgical pathology through simplified communication of the pathologist with the clinician. 5th national conference on Advancing Pathology Informatics, Imaging, and the Internet (APIII 2000), Pittsburgh, PA. Arch Pathol Lab Med 125(8):1014 (abstract). https:// www.archivesofpathology.org/toc/arpa/ 125/8 163. Graham AR, Bhattacharyya AK, Scott KM et al (2009) Virtual slide telepathology for an academic teaching hospital surgical pathology quality assurance program. Hum Pathol 40(8):1129–1136. https://doi.org/10. 1016/j.humpath.2009.04.008

458

Roger S. Riley et al.

164. Bornfleth H, Aldinger K, Hausmann M, Jauch A, Cremer C (1996) Comparative genomic hybridization imaging by the one-chip true-color CCD camera kappa CF 15 MC. Cytometry 24(1):1–13. https://doi. org/10.1155/2000/768578 165. van Der Laak JA, Pahlplatz MM, Hanselaar AG, de Wilde PC (2000) Hue-saturationdensity (HSD) model for stain recognition in digital images from transmitted light microscopy. Cytometry 39(4):275–284. https://www.ncbi.nlm.nih.gov/pubmed/ 10738280 166. Fuller CE, Schmidt RE, Roth KA et al (2003) Clinical utility of fluorescence in situ hybridization (FISH) in morphologically ambiguous gliomas with hybrid oligodendroglial/astrocytic features. J Neuropathol Exp Neurol 62 (11):1118–1128. https://doi.org/10.1093/ jnen/62.11.1118 167. Perry A, Fuller CE, Banerjee R, Brat DJ, Scheithauer BW (2003) Ancillary fish analysis for 1p and 19q status: preliminary observations in 287 gliomas and oligodendroglioma mimics. Front Biosci 8:A1–A9. https://doi. org/10.2741/896 168. Veropoulos K, Learmonth G, Campbell C, Knight B, Simpson J (1999) Automated identification of tubercle bacilli in sputum. A preliminary investigation. Anal Quant Cytol Histol 21(4):277–282. https://www.ncbi. nlm.nih.gov/pubmed/10560504 169. Do¨rge T, Carstensen JM, Frisvad JC (2000) Direct identification of pure Penicillium species using image analysis. J Microbiol Methods 41(2):121–133. https://doi.org/10. 1016/S0167-7012(00)00142-1 170. Ernsting K, Suner S, Jay G (2001) Use of digital imaging of conjunctiva to predict hemoglobin concentration. 2001 SAEM Annual Meeting, Atlanta (GA, USA), 6-9 May 2001. Acad Emerg Med 8(5):528–529. https://doi.org/10.1111/j.1553-2712. 2001.tb00169.x

171. Vrolijk J, Sloos WC, Verwoerd NP, Tanke HJ (1994) Applicability of a noncooled videorated CCD camera for detection of fluorescence in situ hybridization signals. Cytometry 15(1):2–11. https://doi.org/10.1002/cyto. 990150103 172. Chen MY, Ott DJ, Rohde RP, Henson E, Gelfand DW, Boehme JM (1997) Costeffective poster and print production with digital camera and computer technology. Am J Roentgenol 169(4):955–957. https://doi. org/10.2214/ajr.169.4.9308444 173. Frank MS, Dreyer KJ, Mehta A (1999) The megapixel digital camera: value for creating publication-quality illustrations. Am J Roentgenol 173(4):883–887. https://doi.org/10. 2214/ajr.173.4.10511141 174. Barker NJ, Zahurak M, Olson JL, Nadasdy T, Racusen LC, Hruban RH (1998) Digital imaging of black and white photomicrographs: impact of file size. Am J Surg Pathol 22(11):1411–1416. https://doi.org/10. 1097/00000478-199811000-00012 175. O’Brien MJ, Sotnikov AV (1996) Digital imaging in anatomic pathology. Am J Clin Pathol 106(4 Suppl 1):S25–S32. https:// www.ncbi.nlm.nih.gov/pubmed/8853053 176. McColl RI, Johnson A (2001) The comparative effectiveness of conventional and digital image libraries. J Audiov Media Med 24 (1):8–15. https://doi.org/10.1080/ 0140511020033306 177. Riley RS, Ben-Ezra JM, Massey D, Slyter RL, Romagnoli G (2004) Digital photography: a primer for pathologists. J Clin Lab Anal 18 (2):91–128. https://doi.org/10.1002/jcla. 20009 178. Lanni F, Baxter GJ (1992) Sampling theorem for square-pixel data. Proc SPIE 1660:140–147. https://doi.org/10.1117/ 12.59551

INDEX Underlined numbers denote figures/legends. Bold numbers denote main (A-Z listed) glossary entries. Numbers shown in italics denote tables.

A Abbe diffraction apparatus ........................ 211, 229, 230, 231, 232, 285, 286 Abbe diffraction limit................................................. 7, 38 Abbe, Ernst Karl............................... 7, 38, 212, 225, 286 Absorbance .................................................. 217, 218, 350 Absorption (optical).............................77, 144, 184, 201, 204, 206, 214, 215, 217–220, 221, 263, 264, 272, 275, 307, 309 Accessory retina............................................................. 383 Acetylcholine esterase ................................................... 267 Acheta domesticus (house cricket) ................................ 233 Actin...................................................................... 305, 311 Adapter tube........................................383, 386, 387–389 Adaptive optics ............................................................8, 26 Adeno-associated virus (AAV)............. 29, 135–137, 138, 140, 141, 148, 150 Advanced photographic system (APS)............... 381, 385, 390, 395 Airy disc ....................................................... 205, 212, 217 Airy unit.....................................................................86, 88 al-Haytham (Alhazen), Hasan Ibn............................... 202 Alzheimer’s disease (AD)................................56, 63, 158, 160, 163–165, 167, 171, 176 American Optical (Spencer) Co. ........................ 246, 293, 302, 317 Amira .................................................................79, 89, 98, 115, 117, 118, 146, 333 Amplitude object.........................................263, 275, 316 Amyloid beta (Aβ) ............................... 158–178, 280, 312, 360 binding alcohol dehydrogenase..................... 174–177 cascade hypothesis.......................................... 158, 169 fibril accumulation .................................................. 158 plaques ........................................... 158–160, 162, 170 precursor protein (APP) .......................158, 164–167, 172–173, 175 Analog to digital converter (ADC) .............................. 406 Ancylostoma caninum ................................................... 264 Anesthesia ............................................................... 53, 139 Annular (ring) adapter .................................387–389, 391

Anoptral contrast ........................................ 290, 317, 319 Anti-reflection coating ......................................... 207, 210 Antibody anti-MAP2 ..............................................352–353, 369 anti-SMI31 ............................................ 353, 368, 369 Apis mellifera (honeybee)....................................... 83, 91, 101, 103, 104, 110, 117 Apodization ......................................................... 283, 294, 297–299, 300, 301, 303, 306, 314, 316 Arborization tree .............................. 76, 88, 98, 101, 109 Area array (CCD).......................................................... 407 Arion lusitanicus ......................................... 281, 292, 295 Artifact(s) blooming (see Blooming) (color) aliasing.......................................... 88, 408, 412 (due to) diffraction ........................................ 231, 241 false negative/positive(ity) .................. 109, 137, 140, 142, 143, 144, 148, 149, 151 freezing (freeze-thaw).........................................54, 62 fringes(ing) ............................................. 240, 412, 414 halo (shade-off) (see Halo (shade-off) artifact) (in) image acquisition ..................................... 88, 381, 382, 386, 397, 407, 408, 410, 411, 412, 413, 416, 418, 419, 434 (in) image processing ............399, 418, 419, 424, 439 “leaking” (false positive)...................... 137, 140, 143, 147, 150 Moire´ ..................................................... 381, 397, 421 shrinking ......................................................... 187, 190 (in) stereomicroscopy ................... 249–251, 262, 272 track labeling .................................................. 148, 149 vibration............................... 383–386, 391–394, 396, 397, 427, 431 vignetting (see Vignetting) Aspect ratio................................. 414–415, 425, 427, 442 Aspheric optics .............................................329, 331–336 Astigmatism ................................................................... 251 Astrocyte fibrous ............................................................... 11, 116 protoplasmic .............................................................. 11 subpial end feet ......................................................... 11

Radek Pelc et al. (eds.), Neurohistology and Imaging Techniques, Neuromethods, vol. 153, https://doi.org/10.1007/978-1-0716-0428-1, © Springer Science+Business Media LLC 2020

459

NEUROHISTOLOGY

460 Index

AND IMAGING

TECHNIQUES

Astroglia.......................................................................3, 11 Autofluorescence................................................... 51, 104, 161, 165, 169, 449 Automatic exposure bracketing (AEB) .................................... 396 white balance (AWB) .............................................. 397 Auxiliary focus finder .................................................... 421 Axon degenerating ..................................................... 6, 9, 21 guidance................................................. 362, 364, 370 hillock ........................................................................ 15 initial axon segment .................................................. 15 isolation device (AID)........................... 342, 343, 356 mixed axon terminal ...........................................29, 34 nodes of Ranvier ....................................................... 33 unmyelinated ............................................3, 10, 15, 21 Axonal diode ........................................................ 360, 361 Axotomy ............................................................... 361, 362

B Background Subtraction Toolkit.................................. 437 Bands of Bu¨ngner ......................................................... 372 Barrel distortion ................................................... 249, 250 Bayer model................................................. 381, 382, 397 Bayonet (adapter/fitting) ............................390–392, 396 Beam expander ..................................................... 328, 331 Beamsplitter......................................................... 132, 236, 237, 251, 396, 400 Benzylalcohol benzylbenzoate (BABB) ..............185, 186, 188–191, 193 Benzyl benzoate .................................................. 184–186, 190, 191, 193, 194 ß-secretase.....................................................172–174, 175 Biogenic amine................................................................ 76 Biotinylated dextran amine (BDA) ........... 129, 131–133, 139, 141, 149, 150 Birefringence ........................................................ 159, 394 Birthday studies............................................................... 33 Bit mapping ................................................................... 414 Blooming ............................................................. 382, 386, 412, 413, 421 B-minus contrast ........................................ 300, 301, 302, 307, 316, 320 Bodian’s staining ........................................................... 112 Bouin’s................................................................ 52, 53, 58 Bouton en passant.............................................................15, 24 terminaux .................................................................. 24 Boyle, Willard Sterling .................................................. 405 Brady’s (reagent/test) ......................................... 193–195 Brain average ......................................90, 92, 115, 117, 118 reference ....................................................90, 91, 102, 103, 106–109, 117

standard .................................................74, 80, 90–92, 99, 102, 103, 111, 115, 117, 118 template .............................................80, 90, 115, 118 virtual ................................................................ 90, 117 BrainAligner ......................................................80, 92, 93, 109–111, 115 Brain macrophage ........................................................... 11 Bridge camera..............................................382, 384, 385, 386, 389, 394, 395, 398 Bright contrast ............................................290, 317, 319 Brightness resolution .................................................... 414 Bromobenzene .............................................................. 191 1,5-Bromopentane........................................................ 191 Bruchpilot (protein)........................................................ 81 B-spline free-form deformation ..................................... 91 Buccal epithelial cells ..........................291, 301, 309, 312 Bucket-Brigade Device ........................................ 405, 413 Bulb..................................................................30, 94, 106, 141, 226, 302, 396 Butylhydroxytoluene (BHT) ............................... 192, 194

C 2C40 ..................................................................... 162, 163 Caenorhabditis elegans .......................................... 31, 159, 160, 305, 306, 311, 316 Cajal’s silver stain .......................................................... 348 Calbindin ....................................................................... 350 Calyx ......................................................83, 106, 107, 110 Camera lucida......................................... 28, 75, 236, 237 Cardiac nerve plexus ..................................................... 267 cDNA...................................................................... 63, 383 Centering telescope ............................228, 232, 288, 289 Central body (of CX).............................................. 76, 83, 88, 94, 95, 100, 105, 106 Central complex (CX) .............................. 75, 76, 78, 83, 88, 93–96, 100, 103, 105, 106, 108 Chaperonin protein....................................................... 308 Charge-coupled device (CCD) .......................... 257, 304, 310, 312, 382, 384, 393, 394, 405–414, 416, 419, 420, 421, 423, 431, 432, 449 Chemical (tissue) clearing.................................... 183–197 Chemical patterning............................................. 362, 363 Chick retina ................................................................... 237 Chironomus.................................................................... 277 Choline acetyltransferase (ChAT) ...............................136, 140–149, 151 Chromatic aberration..........................218, 251, 397, 427 Chromatolysis................................................................ 8, 9 Chromatophores ......................................... 265, 270, 310 Chromosomes ...................................................13, 32, 33, 276, 277, 449 Ciona intestinalis .......................................................... 383 Circadian........................................................................ 107 CLARITY ............................................................... 25, 188

NEUROHISTOLOGY Climbing fibers................................................................ 16 “C-mount” interface..................................................... 421 Coaxial illuminators ...................251, 260, 268, 272, 273 Cochlea ................................................................... 34, 327 Collagenase................................................ 78, 83, 95, 347 Colloid ........................................................................... 326 Colon ............................................................................ 276 Color absorption................................................................ 220 depth ...................................................... 414, 416, 421 false (pseudo) .................................................. 88, 140, 143, 163, 449 management module .............................................. 415 perception....................................................... 219, 220 structural.................................................................. 220 vision ...................................................... 219, 237, 303 Color model CMYK.................................................... 415, 416, 442 RGB ...................................................... 382, 399, 415, 416, 421, 422, 441, 442, 449 Color space Adobe RGB 1998 ................................................... 415 sRGB IEC61966-2.1.............................................. 415 Color system CIE L*a*b*............................................................. 416 Coma ............................................................................. 251 Commissure............................................................ 24, 116 Common main objective (CMO) ...............................243, 247–253, 257, 259, 268 Compact camera ............................................................ 382–388, 394, 395, 398 flash .......................................................................... 423 Comparison microscopes............................ 236, 237, 238 Compass navigation ............................................... 94, 105 Complementary area.......................... 283, 316, 317, 318 Complementary Metal Oxide Semiconductor (CMOS) .......................................... 356, 359, 360, 382, 384, 393, 395, 405–408 Compur (leaf) shutter .......................................... 383–386 Computational Morphometry Toolkit (CMTK) ......... 80, 90–93, 95, 96, 103–108, 115 Computer-assisted image analysis ....................... 447–448 Computer interface card reader ............................................................... 424 FireWire port.................................................. 423, 424 parallel port ............................................................. 424 PC card slot ............................................................. 424 serial port ................................................................. 424 universal serial bus (USB)............. 393, 394, 421–424 Condenser Abbe....................................................... 226, 286, 313 aplanatic ................................................................... 233 cardioid .................................................. 234, 235, 326 Cassegrain................................................................ 326

AND IMAGING

TECHNIQUES Index 461

Heine ............................................. 234–236, 297, 302 pancratic................................................................... 226 Cone (cells) ................................................ 219, 237, 255, 304–305, 381, 383 Confocal (imaging/microscopy) cross talk ................................................ 132, 144, 145 history (see Naora and Minsky) image stack (see Image stack) imaging channel ............................................. 132, 143 multichannel scanning ............................................ 145 point spread function....................................... 97, 116 resolution (axial/lateral) ........................35, 86–89, 93 sequential scanning ........................................ 140, 143 Z-scanning............................................. 132, 146, 147 Congo red (CR).................................................. 159, 160, 162, 167, 170, 171 Conjugate area ....................................... 282, 283, 285, 286, 317 plane....................................................... 226, 227, 256 Connexins........................................................................ 34 Convergence................................. 75, 247, 249, 250, 331 Cornea ............................................................55, 262, 271 Cornu ammonis ............................................................. 193 Correlative light-electron microscopy (CLEM)........... 19, 35–37 Cortex agranular .................................................................... 16 dysgranular ................................................................ 16 granular.........................................................16, 38, 39 Cortical dopamine transporter ..................................... 350 Cortical neurons.......................................... 360, 361, 368 Counterstaining ................................. 16, 17, 19, 38, 142 Creaking locust (rattle grasshopper)............................ 277 Cresyl (echt) violet ................................................ 5, 6, 18, 20, 21, 64, 348 Creutzfeldt-Jakob disease ............................................. 158 Critical angle ............................................... 208, 209, 304 Cropping (of images).................................. 386, 437, 442 Crustaceans........................................................... 214, 262 Cryo-electron (EM) tomography ....................... 307, 308 Cryoprotectant ..........................................................54–55 Cryoprotection..........................................................54–55 Cryosectioning ...............................................37, 171, 306 Curcumin.............................................................. 163, 169 Curves control...................................................... 435, 436 Cyanine ...........................................................85, 162, 163 Cycloptic® (by American Optical) ............................... 246 Cylindrical lens .................................................... 328–331, 333, 334, 337 Cytoarchitectonics........................................................... 16

D Dale’s postulate ............................................................... 35 Dark contrast................................................................. 316 Dark current ........................................................... 88, 411

NEUROHISTOLOGY

462 Index

AND IMAGING

TECHNIQUES

Darkfield ................................................... 23, 24, 28, 139, 216, 229, 231, 232, 233–236, 258, 260, 262, 264, 265, 267, 270–271, 276, 279, 282, 283, 297, 302, 310, 318, 319, 326, 397, 411, 421, 426 de Broglie, Louis........................................................... 203 Deconvolution ....................................149, 212, 306, 437 Dedicated photomicrography camera................ 380, 393, 394, 395, 399, 409, 421, 426, 428, 429, 430, 432 Defocus...................................... 265, 279, 306, 308, 317 Deformation field............................................................ 91 Degenerin ...................................................................... 306 Dementia ................................................................ 15, 158 Demonstration microscope Abbe............................................................7, 230, 286 vs. phase plates................................................231, 286 Pulfrich-Abbe ................................................. 230–231 Dendritic spine .......................................... 15, 20, 36, 139 2-Deoxy-D-[1-14C]glucose (2DG) ............................... 32 Depth of field .......................................................254, 255, 257–260, 273, 306, 399, 432, 439, 440 Depth perception ................................................. 247, 248 Depth structure ...................................296, 297, 298, 315 Desert locust, see Schistocerca gregaria Diaminobenzidine (DAB) ..................................... 37, 112 Diaphragm (aperture, mask) annular (ring) ................................................ 283, 284, 285, 288, 289, 294, 297, 300, 313, 387, 390 aperture................................. 226, 227, 228, 234, 309 field ................................................227, 228, 254, 256 field (-of-view)................................................ 227, 256 iris................................. 228, 229, 230, 234, 258–260 relief ....................................................... 278, 313, 314 shifting ......................................................26, 228, 229 slit.......................................................... 232, 279, 282, 285, 286, 317, 328–330 Szegvari’s ................................................................. 326 Diatom......................................................... 217, 230, 233 Dibenzyl ether (DBE) .................................185–195, 197 Dichroic mirror ........................................... 132, 271, 304 Differential interference contrast (DIC).....................202, 207, 211, 213, 229, 265, 269, 276, 277, 287, 288, 310, 318 Diffracted light .............................................................221, 264, 282, 283, 284–287, 289, 293, 294, 298, 299, 302, 313, 317 Diffraction funnel .............................................................. 231, 232 limit ...............................................................7, 38, 164 -limited ............................................................. 86, 211 maxima/pattern ............................................ 225, 229, 230, 231, 259, 270, 286 “Digiscoping” adapter .................................................. 388 Digital atlas (for medical education)

Brainmaps ................................................................ 444 Hypertext Atlas of Organ Pathology ..................... 443 Urbana Atlas of Pathology ..................................... 443 Virtual Slidebox....................................................... 444 Digital ICE Technology ............................................... 437 Digital light sheet (DLS) .............................................. 326 DiI..............................................................................25, 37 3D reconstruction.............................................25, 26, 27, 29, 30, 77, 87, 89, 91, 97, 99, 113, 114, 115, 133, 307 Dimethylsulfoxid (DMSO) .....................................54–55, 83, 84, 142, 164, 185, 186, 190 DiO .................................................................................. 25 Diopter ........................................................ 232, 246, 256 Disc-large (DLG) protein............................................... 81 Disease models .............................................................. 158 Dispersion........................................... 209, 218, 220, 221 DNA .....................................................31, 57, 60, 61, 63, 135, 138, 167, 217, 308, 383 Doming effect ............................................................... 250 d’Orle´ans, Che´rubin ..................................................... 245 Double-floxed (AAV)..........................135, 136, 140, 150 Dragonfly...................................................................83–85 Drawing attachment ............................................236, 237 Drosophila (fruit fly) ................................... 31, 74, 81, 82, 83–85, 90–93, 97–99, 102, 108, 110, 113, 116, 117, 189, 331, 333, 335, 383 Dynamic range .................................................... 380, 399, 408–410, 412, 413, 421, 426, 439, 440 Dynamic range increase (DRI)..................................... 399 Dynaphot....................................................................... 327 Dynein ............................................................................. 22

E EGFP ..........................................188, 193, 196, 305, 311 Electric field (vector) ..........................209, 212, 213, 221 Electrodes amplified .................................................................. 360 on chip ............................................................ 356, 358 interdigitated ............................................................358 patch-clamp .................................................... 354, 357 Electroencephalography (EEG) ................. 342, 354, 355 Electromagnetic radiation (waves) .................. 202–205, 209, 217, 219 spectrum ................................................ 201, 204, 205 Electron microscopy ............................................. 2, 9, 19, 33, 35–37, 55, 111, 126, 131, 203, 279, 306, 317, 319 Electrotonic coupling ..................................................... 34 Embryo ............................................................25, 33, 184, 185, 186, 188, 189, 325, 327, 331, 332, 345 Endoneurial tube .......................................................... 372

NEUROHISTOLOGY Ependyma ..................................................................11–12 Epicardial nerves ........................................................... 267 Epitope preservation ....................................................... 77 Erythrophores ............................................................... 265 Ethereal oil ........................................................... 183, 184 EtNU (N-ethyl-N-nitrosourea) ................................... 310 Evanescent wave................................................... 209, 304 Evolution ............................................................. 111, 113, 134, 183–186, 297 Exposure blending ........................................................ 399 Exposure compensation........................................430, 431 Extinction ratio ............................................................. 214 Eyepoint ..............................................246, 256, 386, 396 eYFP absorption maximum .............................................. 144 antibody ......................................................... 142, 144, 145, 148, 149 counterstaining........................................................ 142 emission maximum ................................................. 144 expression ...................................................... 133–138, 141–143, 147, 150 immunohistochemistry .................................. 142–145 spectral properties ................................................... 144 stabilization .................................................... 142–145

F False connectivity .......................................................... 110 False negatives ..................................................... 137, 142, 143, 148, 151 Far-field (conditions) ........................................... 329, 330 Fascin .................................................................... 305, 311 Fiber optics .................................................. 209, 218, 266 Field number (FN)........................................................ 256 Field of view (FOV) ................................................ 78, 88, 246, 256, 258, 268–270, 273, 390, 410, 413, 426, 428, 430–432 FIF technique ............................................................34, 35 FIJI................................................................... 79, 89, 115 Fill factor........................................................................ 407 Film recorder ................................................................. 425 Filopodia........................................................................ 226 Fixation .......................................... 2, 5, 6, 31, 34, 49–65, 77, 78, 81, 83, 151, 351, 352 Fixed pattern noise........................................................ 411 Flagellum ....................................................................... 228 Flash shoe ..............................................................384, 385 Flatbed scanner .................................................... 405, 434 Fluorescence in situ hybridization (FISH)......... 426, 449 Fluorochromated secondary antibodies ....................... 24, 132, 144 Fluorochrome selection ....................................... 125–151 Fluorogold.............................................................. 24, 129

AND IMAGING

TECHNIQUES Index 463

FlyCircuit ...........................................................80, 89, 91, 92, 98, 108–109, 115, 116, 117 f-number ......................................255, 256, 258, 428, 431 FocusClear ..................................... 84, 185, 186, 187, 189 Focus stacking ............................................................... 440 Formaldehyde......................................... 5, 34, 51–56, 58, 59, 61, 64, 81, 83, 189, 195, 196, 351, 352 Frame-transfer CCD (FT-CCD) .................................. 406 Free-floating incubation ............................................... 141 Fresnel’s equations ............................................... 207, 214 Fringes ......................................................... 211, 397, 414 Fruit fly, see Drosophila Full frame (sensor size) ............................... 380, 381, 392 Full-frame CCD ................................................... 406, 421 Functional MRI (fMRI) ................................32, 297, 298

G Galilean beam expander ........................................................ 331 telescope .................................................................. 253 Ganglion (-a) cell .............................................................................. 10 cluster....................................................................... 267 thoracic ......................................................... 81, 83, 85 Gap junctions ..................................................... 11, 29, 34 Gas anesthesia................................................................ 139 Gaussian beam (profile) ................................................ 329, 330, 332–334 functions .................................................................. 329 Gene expression profiling .................................. 57–58, 65 Genes .................................................... 13, 31, 33, 49–65, 134, 135, 136–138, 150, 174, 305, 345, 346 Geometrical optics .....................203, 206, 207, 209, 214 GFP fluorescence reconstitution across synapses (GRASP) .............................................................. 31 Glia............................................................... 347, 367, 368 astrocyte........................................................ 10, 11, 12 ependyma............................................................. 11–12 glial fibrillary acidic protein ................................11, 12 microglia .............................................................. 10–13 oligodendrocyte ..................................................10, 12 radial .......................................................................... 12 scar ............................................................................. 11 Globular effect .............................................................. 250 Glutaraldehyde .................................................... 5, 51, 53, 55–57, 141, 351 Gnathonemus sp. ........................................................... 383 Golgi, Camillo.................................................6, 8, 12, 13, 17–20, 23, 27, 28, 30, 35, 36, 37 Golgi (silver) staining....................................... 12, 18, 20, 23, 28, 30, 35, 73, 74, 112, 348

NEUROHISTOLOGY

464 Index

AND IMAGING

TECHNIQUES

Granular cortex ........................................................ 16, 38 Granular (cell) layer .................................. 17, 36, 38, 193 Granule cell (layer)...................................... 16, 17, 30, 38 Green fluorescent protein (GFP) ............................25–27, 29–31, 33, 37, 91, 134, 140, 142, 144, 145, 148, 149, 165, 168, 169, 183–197, 255, 272, 304, 305, 306, 327, 335, 350 Greenough (design).......... 247, 248, 249–252, 266, 268 Greenough, Horatio S. ........................................ 246, 247 GroEL............................................................................ 308 Grouped retina .............................................................. 383 Growth cones ......................................277, 304, 305, 314

H Hagfish.................................................................. 232, 233 Halo (shade-off) artifact ............................ 285, 287–288, 293, 294, 295, 296–300, 302, 303, 306, 311, 315, 368 Hanstro¨m, Bertil ............................................................. 74 Hawkmoth, see Manduca sexta Heavyside step function................................................ 329 Heliothis virescens (tobacco budworm) .................. 82, 83, 91, 103, 106, 108, 117 High dynamic range imaging (HDRI) ..................................................... 440 rendering (HDRR) ................................................. 399 Highest occupied molecular orbital (HOMO) ........... 162 Hilbert differential contrast (HDC) ..........307, 308, 317 Hippocampal (CA1) neuron ...................... 20, 27, 29, 63 Hippocampus (-i)......................................... 5, 12, 16, 30, 39, 64, 131, 135, 139, 148, 150, 170, 171, 189, 193, 196, 197, 335, 345, 347 Histochemistry ...............................................7, 24, 51–53 Histogram equalization ................................................ 436 Histological procedure ....................................22, 52, 141 Hoffman modulation contrast ........................... 229, 269, 276, 278, 279, 282, 297, 310, 315, 318, 319 Honeybee, see Apis mellifera Hooke, Robert .............................................................. 279 Hookworm .................................................................... 264 Horseradish peroxidase (HRP) ................................22, 24 Hue ......................................................415, 434, 436, 449 Human brain .............................................. 53, 54, 56, 57, 64, 65, 74, 174, 297, 298, 433 Huntington’s disease ........................................... 158, 160 Huygens-Fresnel diffraction integral ........................... 329 Hydrogel................................................................186, 189 Hyperphosphorylated proteins............................ 171, 172 HyperText Markup Language (HTML)...................... 440

I Illumination “critical” (Nelson)................................................... 226

diascopic ................................................ 260, 268, 269 episcopic ................................................. 251, 260, 261 fiber optic................................................................. 266 incident ......................................... 234–236, 260–262, 265–269, 273, 358 Ko¨hler ................................................... 226, 228, 229, 267, 288, 431, 437 off-axis (oblique, relief) ................................ 226–229, 260, 264, 265–269, 278, 279, 296, 313, 319 Rheinberg .............................231–234, 302, 318, 319 through-the-lens ..................................................... 268 transmitted ............................................260, 262–265, 268, 269, 273 Image processing ............................................. 26, 27, 30, 89, 115, 309, 399, 400, 403, 413, 420, 422–424, 435, 437 segmentation ............................................... 77, 79, 89, 90, 94, 113, 115, 116 sensor .................................................... 381, 382, 394, 395, 397, 404–414 stack ..................... 26, 27, 28, 30, 76, 77, 78, 79, 80, 86, 88, 89–93, 94, 95–100, 103, 104, 109, 110–112, 115, 116, 118, 132, 326, 327, 331, 333, 399, 440 ImageJ............................................ 27, 89, 115, 118, 146, 306, 309, 399, 435, 437, 448 Imaris ............................................................................... 79 Immediate early gene expression ................................... 31 Immersion fixation ................................................. 51, 53, 54, 351 oil ........................................................................88, 95, 103, 206, 211, 230, 315 Immunocytochemistry...................................... 51–54, 65, 351, 364, 370 Immunofluorescence ......................................51, 64, 107, 130, 131, 140–145, 148, 151, 346, 372 Immunoglobulin (Ig) ................................................... 349 Immunohistochemistry (IHC)......... 23, 24, 51, 77, 129, 142–145, 148, 151, 351, 447 Immunoprecipitation.................................................... 174 Immunoreactivity ................................................... 94, 106 Immunostaining (-labeling) .............................12, 17, 18, 37, 63, 99, 110, 114, 140, 149, 186, 189, 332, 342, 348, 350–352, 354 Incubation liquid .......................................................... 184 Induced pluripotent stem cell (iPSC) .......................... 344 Infinity space ........................................................ 251, 252 Injection iontophoresis ........................................................... 140 mechanical ...................................................... 139, 140 micropipette ................................................... 139, 140 retraction ................................................................. 149 spot ....................................... 127, 130, 131, 141, 147

NEUROHISTOLOGY stereotaxic ......................................................... 22, 139 InsectBrainDatabase (IBdb) ......................................... 113 In situ hybridization .........................................51, 59, 76, 110, 383, 426, 449 Interference (optical) ................................. 201, 202, 204, 206, 207–220, 221, 229, 234, 264, 265, 269, 276, 287, 288, 293, 303, 310, 313, 314, 318, 384, 397 Interference-phase contrast ........................ 293, 310, 384 Interferometer ...................................................... 293, 310 Interline-transfer CCD (IT-CCD)............................... 407 Intermediate tube (photo/video) .............. 251, 253, 273 Internal plexiform layer ................................................ 395 Interneurons............................................. 16, 17, 30, 106, 108, 127, 128, 130, 131, 133, 135, 148–150, 395 Interphako (Zeiss)....................................... 288, 293, 309 Interpolated resolution ................................................. 414 Intracellular dye filling .................................................. 131 Isoeugenol ..................................................................... 191 Isoleucine catabolism .................................................... 174 Isosafrole....................................................... 184, 185, 191 Iterative Shape Averaging (ISA).......................80, 83, 91, 92, 94, 95, 102–104, 107, 109, 116

J Jentzsch’s “ultrakondensor” ........................................ 326

K Kaede ............................................................................... 33 Keystone distortion (effect).......................................... 249 Kidney................................................................... 131, 308 Kinesin ............................................................................. 22 Knock-in mice ............................................................... 137 Ko¨hler, August ............................................ 225, 226, 228

L Lamellar dissection....................................................... 271 Lamellipodia ......................................................... 226, 277 Lamina ................................. 4, 11, 30, 83, 106, 108, 346 Laser capture microdissection (LCM) .....................59–64 LE-Adapter .................................................................... 428 Leaf replica .................................296, 299, 301, 312, 365 Lens accessory ...................................................88, 226, 228 aspheric ........................................................... 332–334 attachment ................... 252, 254, 255, 258, 259, 273 Bertrand.......................................................... 232, 288 collector ................................................. 226, 228, 396 converter........................................ 383, 385, 387–389 fundus ...................................................................... 255 GRIN ....................................................... 25, 209, 221 macro ....................................................................... 261 mesolens .................................................................. 325

AND IMAGING

TECHNIQUES Index 465

ommatidium ............................................................ 383 planoconvex.................................................... 331, 334 telephoto......................................................... 386, 387 Lepisma saccharina............................................... 289, 311 Leucoagglutinin (PHA-L) .............................24, 131, 139 Leucophaea (Rhyparobia) maderae ......................... 78, 81, 82, 83, 106, 117 Levamisole ..................................................................... 311 Levels control ....................................................... 435, 436 Lewy bodies................................................................... 158 Libellula luctuosa............................................................. 83 Light sensitivity ...........................................262, 384, 385, 407–409, 412, 422, 423, 426 Light-sheet microscopy .......................... 8, 184, 325–336 Light-switching ............................................................. 169 Lightroom ..................................................................... 434 Linear array (CCD).............................................. 407, 408 Lobe accessory ..................................... 88, 93, 94, 100, 106 antennal .................................... 75, 83, 102, 106, 108 medial ........................................................................ 83 optic ....................................76, 78, 83, 102, 106, 108 vertical........................................................................ 83 Lobelet........................................................................... 106 Lobula....................................................83, 250, 333, 335 Local processing unit (LPU) .........................99, 108, 109 Lots-of-rods mutant ..................................................... 237 Lowest unoccupied molecular orbital (LUMO)......... 162 Luminescence lifetime .................................................. 167 Luxol fast blue (LFB) ...............................................19–21 Lyot, Bernard ....................................................... 286, 287

M Macrobrachium jelskii ................................................... 265 Macrophage..................................................................... 11 Madeira cockroach, see Leucophaea maderae Magnetic resonance imaging (MRI)...................... 26, 32, 297, 298, 325, 354 Manduca sexta (hawkmoth) .............................78, 81, 82, 83, 90, 102, 117 Marchi stain .................................................................8, 21 MARCM flip-out technique.................................. 91, 108 Marmoset....................................................................... 383 Masking .................................................... 53, 81, 94, 294, 351, 354, 436, 437 MatLab ...........................................................79, 367, 369 Maximum intensity projection ............................ 271, 306 mCherry ..................................................... 188, 255, 272, 304, 305, 311 Medulla ........................................... 76, 83, 106, 333, 335 Megalopta genalis (sweat bee) ............. iv, 78, 83, 88, 114 Meglumine diatrizoate......................................... 185, 187 Meiosis ........................................................................... 277 Mesolens ........................................................................ 325

NEUROHISTOLOGY

466 Index

AND IMAGING

TECHNIQUES

Metal-oxide-semiconductor (MOS) .......... 359, 382, 405 Meta-materials............................................................... 206 Methanol .................................................... 59, 78, 81, 83, 163, 190, 195, 351 Methyl salicylate ......................................... 23, 25, 77, 84, 85, 184, 185, 186, 191 Michel, Kurt .................................................................. 230 MicroBrightField .......................................................... 422 Micro-computed tomography (μCT) .......................... 111 Micro Four Thirds (MFT).......................... 381, 385, 392 Microelectrode array (MEA) ..................... 342, 345, 350, 355–360, 363–372 Microfabrication............................................................ 360 Microflash ...................................................................... 396 Microfluidics............................................... 342, 345, 350, 351, 356, 360–364 Microsphere................................................................... 117 Microtomography ................................................ 309, 310 Microtubule....................................................... 13–15, 22, 171, 234, 235, 252 Microtubule-associated protein (MAP)......................171, 352, 364, 368 Mikropolychromar ...................................... 231, 232, 233 Minsky, Marvin Lee ........................................................ 26 Modulation contrast .................................. 226–229, 269, 276, 278, 279, 282, 297, 310, 315, 318, 319 Modulator ..................................227, 229, 269, 279, 318 Moire´ (effect/artifact) ......................................... 381, 397 Molecular layer ...................................................17, 36, 39 Molecular motion ......................................................... 170 Monarch butterfly .............................................81, 83, 88, 93–96, 100, 103, 105, 106, 107 Monochromatic light..........................204, 210, 304, 397 Mormyrid fish......................................................... 34, 383 Morphometric ........................................................ 51, 426 MOS, see Metal-oxide-semiconductor Mounting.................................................... 81, 84–86, 88, 145, 187, 428 Moving one chip camera .............................................. 408 Multielectrode array (MEA)............................... 342, 345, 350, 355–360, 363–372 Multiphoton microscopy ............................ 161, 164, 218 Multiple sclerosis............................................................. 11 Murine heart ........................................................ 267, 270 Murray’s clear ................................................................ 185 Mushroom body .............................................83, 92, 102, 106, 107, 110 Myelin ................................................. 3–8, 10, 15, 19–21, 25, 36, 187, 188, 197

N Nanoparticle ........................................................... 23, 362 Naora, Hiroto ................................................................. 26 NA, see Numerical aperture

Near-field (conditions) ........................................ 329, 330 Near-infrared fluorescence................................... 161–163 Nematode ...................................................................... 262 Neural network ...................................................... 75, 114 Neurite............................................ 88, 97, 101, 341–372 Neurite isolation device ............................................. 342, 360–368, 372 MEA...................................... 342, 363, 364, 365–372 Neural progenitor cell, 344 Neuroanatomical tracing anterograde................................22, 24, 131–133, 139 neurotransmitter-specific ............................... 127, 134 retrograde fluorescence........................................... 128 selective fluorochrome expression................. 125–151 track labeling .................................................. 148–151 viruses ............................................................. 133–134 Neurobasal medium..............................................347, 348 Neurobiotin..................................................................... 88 Neuroblastoma........................................... 176, 177, 300, 304–305, 311, 343 Neuroblastoma-glioma .......................300, 304–305, 311 Neurochemical fingerprinting ................................................ 131–133, 139, 148, 149 map ............................................................................ 76 Neurodegenerative disease .............................54, 65, 158, 161, 169–171, 176, 178, 350, 372 Neurofibrillary tangles (NFT) .......................63, 171, 172 Neurofilament ................................ 15, 64, 325, 350, 353 Neurogenetics ................................................................. 49 Neuroglioma cells ......................................................... 169 NeuroLucida ................................................................... 79 Neuromodulator ..............................................34, 76, 106 Neurotrophin ......................................................... 52, 364 Nerve growth factor................................................... 52, 345, 347, 363–364, 369–372 plexus ....................................................................... 267 regeneration............................................................. 271 Neuron archetypal................................................................... 13 axon..................................................... 3, 8, 10, 13–16, 21–23, 24, 28, 29, 32, 34–35, 36 basket ........................................................16, 135, 138 chandelier ......................................................... 16, 135 cilia .......................................................................12, 15 CL1 ............................................................................ 88 endoplasmic reticulum....................6, 13, 32, 36, 348 dendrite............................................. 8, 13–17, 20, 24, 29, 32, 36, 174, 348 doctrine.........................................................2, 3, 8, 39 dendritic spine ..................................... 15, 20, 36, 139 giant fan shaped ...................................................... 105 Golgi apparatus .....................................................6, 13

NEUROHISTOLOGY hippocampal, CA1 ...................................... 20, 27, 29, 30, 37, 63, 148 input......................................................................... 105 microaspiration............................................. 60, 62, 65 motor end plate......................................................... 16 Nissl substance ....................... 13, 14, 16, 17, 36, 348 nucleolus...................................................6, 13, 14, 32 nucleus ...................................................13–16, 24, 29, 32, 34, 36, 129, 130 output ............................................ 105–107, 109, 110 polarization-sensitive .............................................. 105 Purkyneˇ (Purkinje) ...................................... 15, 16, 17, 19, 30, 35, 36, 135 touch receptor ................................................ 305, 306 Neuronal cell culture embryonic....................................................... 344, 345 immortalized .................................................. 343–345 neuron-like .............................................................. 345 primary................................................... 343, 345, 346 stem-cell derived ............................................ 343, 344 Neuronal circuit ......................................................................... 75 guidance cue............................................................ 370 network....................................................97, 126, 130, 133–135, 138, 189, 342, 343, 354, 357, 360, 361, 372 protein aggregates................................. 158, 160, 169 Neuropeptide ........................................76, 107, 129, 151 Neuropil compass......................................93, 94, 103, 105–108 fibrous ...................................................................... 116 homologous ........................................ 76, 82, 91, 113 synaptic .................................................................... 116 “unstructured” ...........................................76, 83, 108 Neurotransmitter ....................................... 11, 33–35, 76, 97, 127, 129, 134, 151, 344 NG108-15 cell ....................................300, 304–306, 311 NIAD-4 ................................................................ 162, 164 Nicotinic acetylcholine receptor (nAChR) .................. 369 Nissl bodies....................................................................... 6, 8 stain .................................................. 5, 6, 8, 14, 16–19 substance........................................13, 14, 16, 36, 348 Nodular unit........................................................... 94, 106 Noduli.......................................................... 83, 88, 94, 95 Nomarski ....................................160, 229, 269, 276, 277 Nuclei......................................................5, 16, 30, 32, 39, 128, 148, 150, 310, 348, 449 Normal goat serum (NGS).................................... 84, 144 N-type MOS (NMOS) .........................................382, 384 Numerical aperture (NA) adjustment by iris ........................................... 230, 234 of condenser ............................................................ 282 depth of field ..................................255, 258, 259, 260

AND IMAGING

TECHNIQUES Index 467

effective vs. nominal ......................................... 88, 280 f-number ..........................................................255, 258 magnification .................................................. 386, 398 resolution..........................78, 86, 116, 117, 230, 426 TIRF ..................................... 303–305, 314, 315, 319 zoom factor .................................. 255, 258, 259, 272, 383, 390, 428 Nyquist-Shannon (criterion/theorem)................ 86, 103, 398, 426, 427

O Oblique coherent contrast.......................... 260, 264, 269 Oblique (relief) contrast ............................................... 260 Ocellus ........................................................................... 383 Ocular extender............................................................. 391 Ommatidium (-a).......................................................... 383 One-chip, one-shot (camera) .............................. 407, 408 One-chip, three-shot (camera)..................................... 408 Ophthalmoscope ........................................................... 381 Opsin ............................................................................ 383 Optic fiber .......................................................................... 395 lobe .....................................76, 78, 83, 102, 106, 108 tubercle ................................. 88, 93, 94, 95, 105, 106 Optical clearing.................................................. 77, 81, 84, 85, 183–197, 216 coating ................................................... 207, 210, 219 coherence tomography (OCT) .................62, 63, 325 contrasting .......... 265, 270, 275–280, 316, 318, 319 emission tomography (OET) ................................. 184 imaging probes ............................................... 157–182 path difference......................................221, 269, 270, 277, 282, 295, 297, 299, 318 path length ............................................ 218, 221, 318 projection tomography (OPT)............. 184, 284, 325 relay....................................................... 293, 294, 303, 304, 315, 318, 319 section ................................................... 26, 30, 77, 78, 86, 88, 89, 94, 132, 184, 306, 328, 329, 331, 332, 333 thickness ............................... 221, 229, 265, 282, 288 tissue clearing (OTC) ..............................8, 23, 25, 26 transmission tomography ....................................... 184 Optically conjugate plane ........................... 226, 227, 256 Optogenetics .................................. 31, 33, 134, 135, 150 Organoids ...................................................................... 262 Osteoblast-like cell (MG63)................................ 296, 311 Outer limiting membrane (OLM) ............................... 237

P PaintShop ...................................................................... 425 Pancratic system ............................................................ 253

NEUROHISTOLOGY

468 Index

AND IMAGING

TECHNIQUES

Panoramic images ................................................ 434, 439 Paraformaldehyde ...................................... 51–56, 58, 59, 61, 64, 141, 195, 351, 352 Paraneuron ........................................................... 265, 270 Parkinson’s disease ...................................... 158, 168, 169 Particle nature (of light) ...................................... 203, 204 Pathological evidence.................................................... 158 Pearl-eye fish.................................................................. 383 Pedunculus ...................................................................... 83 Peltier cooling ............................................. 421, 423, 426 Peltier, Jean Charles Athanase ...................................... 411 Pepper brine ......................................................... 280, 295 Peppermint oil............................................................... 185 Peranema trichophorum ................................................ 228 Perception of depth ............................................. 248, 273 Perfusion....................................................... 5, 51, 53, 58, 189, 195–196, 351 Pericellular recording .................................................... 131 Perifovea ............................................................... 381, 383 Permeabilization................................81, 83, 84, 185, 351 Permount™...............................................................85, 97 Perspective distortion .......................................... 249, 250 Phase contrast apodized ................................................ 275–315, 316 bright .....................................................276, 278–282, 285, 288–290, 305, 307–310, 312, 317 conventional .......................................... 275–315, 317 dark ................................................................ 276, 279, 281–285, 290, 297, 302, 303, 310, 316, 317 defocus.......................................... 279, 306, 308, 317 with Heine condenser ........................... 294, 301, 302 imitation ............................... 279, 280, 290, 306, 310 interference........................................... 276, 277, 288, 293, 303, 310, 313, 314 negative................................................. 280, 281, 284, 285, 289, 290, 292, 295, 296, 302, 307, 310, 313, 319 positive..........................................276, 277, 280–282, 284, 285, 290, 292, 296, 300, 305, 307, 308, 310, 313, 314, 319 relief ....................................................... 275–315, 319 variable .................................................. 287, 290, 293, 294, 296, 297, 301, 319 Zernike.................................................. 276, 277, 279, 285, 286, 287, 308, 319 Phase object................................................ 264, 265, 280, 281, 282, 283, 285, 287, 290, 291, 292, 309, 315, 316, 318, 319 Phase plate (any type) ................................ 231, 232, 236, 280, 281, 283, 284, 286, 287, 290, 293, 294, 295, 296, 299, 300–302, 303, 304, 305, 306, 307, 308, 310, 313–317, 318, 319, 320 apodized ............................................... 283, 297, 299, 300, 301, 302, 304, 305, 314

A-type .............................................................316, 318 B-minus ............................... 300, 302, 307, 316, 320 B-type .............................................................316, 318 conventional ......................... 283, 295, 296, 300, 301 electrostatic..................................................... 307, 319 external ................................................. 294, 303, 304, 305, 314, 315 Phase shift................................................... 210, 214, 264, 279, 280, 281, 282, 283, 284, 285, 286, 287, 290, 291–293, 295, 299, 302, 307, 308, 310, 311, 315–317, 318, 319 Pheromone .................................................................... 107 Phosphorescence ......................................... 168, 169, 218 Phosphorylation ............................................................ 171 Pphotoactivatable protein............................................. 134 Photocell............................................................... 400, 414 Photoconversion ............................................................. 37 Photodiode ................................................. 382, 405, 406, 407, 409, 411, 412, 413 Photodynamic therapy .................................................. 159 Photoelectric conversion................................................................ 405 effect ........................................................................ 203 Photography continuous (time-lapse) ................................. 438, 439 macroscopic ............................................................. 433 object .............................................................. 438, 439 panoramic .............................................. 438, 439, 440 Photoluminescence .............................................. 167, 169 Photon noise ................................................................. 411 Photoreceptive cell ..............................257, 381, 382, 383 Photoreceptor cell ....................................... 237, 255, 272 Photoshop .......................................... 146, 309, 399, 419, 425, 434–437, 439, 441 Photo tube................................................... 387, 389, 392 Pigment cell ............................................................................ 383 dispersing factor (PDF) ......................... 106, 107, 418 Pincushion distortion ................................................... 250 Pittsburgh compound B (PiB) ........................... 160, 161, 162, 166, 170 Pixel binning..................................................................... 409 depth ........................................................................ 414 dimension ....................................................... 410, 414 Planck, Max ................................................................... 203 Plasmid .................................................... 29, 33, 135, 136 Pleurosigma angulatum................................................ 230 Point spread function (PSF) ......................... 97, 116, 117 Polanret ................................................................ 293, 319 Polarization microscopy ............................ 159, 213, 220, 276, 318, 382, 394, 397 Polarization of light circular ............................................................ 212, 213

NEUROHISTOLOGY elliptical........................................................... 212, 213 linear ............................................................... 212–215 Polarizer (polarizing filter) ................................. 213, 214, 236, 293, 303 Polaroid ................................................................ 214, 380 Polydimethylsiloxane (PDMS) ....................360, 364–366 Polyimide ....................................................................... 359 Polytene (giant) chromosomes .................................... 277 Positron emission tomography (PET) ................. 32, 160, 161, 164, 170, 178 Postfixation time .................................................... 53, 141 Potassium dichromate............................... 18, 19, 58, 348 Powell lens................................................... 332, 333, 334 Preganglionated nerve .................................................. 267 Pre-synaptic vesicle........................................................ 158 Printer dye sublimation ....................................................... 425 Fiery color server..................................................... 425 film recorder ............................................................ 425 Fujix Pictrography................................................... 425 inkjet ............................................................... 424, 425 Iris inkjet.................................................................. 425 solid ink-jet .............................................................. 425 thermal-wax ............................................................. 425 Prion protein ........................................................ 158, 171 Probability map .........................................................80, 90 Proboscis........................................................................ 233 Promoter cytomegalovirus (CMV) ......................................... 135 EFF1α ............................................................. 140, 141 Protease ................................................... 53, 61, 158, 172 Protocerebral bridge .........................................83, 88, 94, 100, 105, 108 Protocerebrum .............................................................. 108 Protozoon...................................................................... 228 Proximal tubule............................................................. 308 Pseudocolors ........................................................ 163, 449 Psophus stridulus ............................................................ 277 Pupil-projection .................................................. 293, 294, 303, 304, 314, 319 Purkyneˇ (Purkinje) cell ........................................... 15, 16, 17, 19, 30, 35, 36, 135, 353 Pyramidal cells ................................ 15, 30, 147, 193, 335

Q Quantal spacing............................................................. 107 Quarter-wave plate..............................213, 293, 303–304 Quasi-3D ..................................................... 279, 318, 319 Quenching (of fluorescence) ................................185, 188

R Radioactive amino acids............................................22, 23 Radioautography .................................22, 23, 24, 33, 127

AND IMAGING

TECHNIQUES Index 469

Radula ..................................................281, 292, 295, 312 Ramifications .......................................................... 97, 127 Ramo´n y Cajal, Santiago.................. 6, 8, 14, 23, 74, 236 Ranvier, nodes of............................................................. 33 Rasband, Wayne ............................................................ 435 Ray model .............................................................. 210, 221 representation................................................. 206, 207 Rayleigh range.....................................329–334, 336, 337 Read-out noise .............................................409, 411, 413 Real-time PCR (RT-PCR)........................................58–60 Recombinase activity............................................ 136, 147 Red flour beetle, see Tribolium castaneum Refraction ............................................. 85, 201, 204–208, 214–216, 221 Refractive index, negative ............................................. 206 Registration .......................................... 25, 89–92, 96–99, 100, 101, 103, 104, 105, 106–111, 115–116, 117, 118 Relay adapter .........................................................428, 429 Relief contrast............................................. 278, 279, 280, 282, 287, 294, 295, 296, 297, 299, 301, 302, 311–313, 315, 318, 319 Resolution axial/lateral........................................ 7, 86, 88, 89, 93 confocal........................................................ 35, 86, 87, 88, 89, 110, 143 diffraction-limited ........................................ 7, 86, 211 diffraction maxima ................................ 229, 230, 231 in digital image/sensor .................404, 408, 410, 427 isotropic ................................................................... 327 NA vs magnification ....................................... 386, 398 widefield..................................................................... 86 Resolving power, see Resolution Reticular doctrine..................................................................2, 39 formation ............................................................ 16, 39 Retina........................................................ 13, 31, 34, 219, 237, 247, 248, 250, 255, 257, 272, 356, 381, 382, 383, 394, 395 Retinal cups ................................................................... 383 Rhodamine ........................................................... 168, 327 Rhyparobia maderae, see Leucophaea maderae Ribonuclease.................................................................... 57 Rienitz, Joachim............................................................ 279 RNA .............................................................. 9, 13, 31, 33, 57–60, 61, 63, 64–65 Rod (cell) ....................................................................... 383 Rotterman contrast ..................................... 260, 264, 269 Ruthenium complex (dye).......................... 166, 168, 169

S Sacrifice ........................................... 22, 23, 131, 141–142 Salivary gland................................................................. 277

NEUROHISTOLOGY

470 Index

AND IMAGING

TECHNIQUES

Saturation ............................................................ 372, 398, 415, 434, 436, 449 Scale ScaleA2 ................................................... 185, 187, 188 ScaleB4 ........................................................... 187, 188 Scale U2.......................................................... 187, 188 Scaleview-A2.................................................................. 188 Scanning digital camera ................................................ 408 Scanning electron microscopy (SEM) ATUM-SEM.............................................................. 30 FIB-SEM ................................................................... 30 Scattering (of light) elastic ....................................................................... 215 inelastic .................................................................... 215 Mie .................................................................. 215, 216 vs. optical clearing ................................................... 216 Rayleigh ................................................. 215, 216, 220 Schistocerca gregaria (desert locust)....................... 81, 82, 83, 90, 91, 93, 94, 102, 103, 105, 117 Schlieren ..................................................... 226–229, 231, 265, 269, 276, 278, 279, 282, 313, 315, 318, 319 Schwann cells.......................................................... 10, 370 Schwann, Theodor ......................................................8, 10 Scopelarchus michaelsarsi ............................................... 383 Sea squirt ....................................................................... 383 Secure digital (SD)........................................................ 423 Seeding macrochannel ......................................... 361, 370 Sehfeldblende .................................................................. 227 Selective fluorochrome expression ................31, 125–151 Selective plane illumination microscopy (SPIM) ...................................................... 112, 327 Sensory pathway.............................................................. 74 Shading ....................................................... 279, 296, 297, 298, 307, 315, 319 Sheath ganglionic .................................................................. 81 neural ...................................................................77, 78 Sherrington, Charles Scott ............................................... 8 Shrimp ......................................................... 219, 265, 270 Siedentopf, Richard Adolf ............................................ 326 Signal to noise (S/N) ratio................................... 88, 369, 370, 409, 411, 422 Sihler’s stain..................................................................... 23 Silver chromate .......................................... 18, 19, 28, 36, 37 nitrate..........................................................18, 19, 348 stain ..................................................... 8, 9, 12, 14, 18, 20, 21, 23, 30, 35, 64, 65, 348 Silverfish................................................................ 289, 312 Single cell gene array....................................................... 54 Single-sideband edge enhancement ............................. 229 Sino-atrial node ............................................................. 267 Skeletonization .................................................88, 98, 100 Skeleton tree........................................... 88, 98, 100, 117

Slide scanner ................................................ 421, 422, 432 Slime gland ........................................................... 232, 233 SLR camera..................................................384, 385, 390, 391, 392–395, 396, 414, 420–422, 426, 428, 429, 433 Smart Removal of Defects (SRD) ................................ 437 Smith, George Elwood ................................................. 405 Snaxel ............................................................................. 117 Snell’s Law (of refraction) ..................208, 209, 218, 220 Sodium azide ................................................................. 142 Spalteholz preparations................................................. 184 Spalteholz, Werner ..........................23, 25, 183, 184, 185 Spanish slug ................................................. 281, 292, 295 Sparrow limit ................................................................. 257 Spatial frequency ........................................ 230, 259, 279, 289, 298, 302, 307 Spectra atomic ............................................................. 206, 217 lines vs. bands .......................................................... 217 molecular ................................................................. 217 Spike activity......................................................... 369, 371 Spot meter .............................................................430, 431 Standardization (of brain images) .......................... 79, 80, 83, 89, 90–94, 96, 102, 116, 117, 118 Starch (granules) ......................................... 158, 278, 312 Stepping ring ............................................... 387, 388, 391 Stereology...................................................................... 422 Steroid metabolism ....................................................... 174 Still cameras ..................................................404, 420–422 Stitching................................................ 89, 118, 438, 439 Strap ................................................................................. 94 Streptavidin........................................................... 131, 349 Striatal neuron ............................................................... 361 Stromal nerve ................................................................ 271 Superior cervical ganglion (SCG) ...................... 342, 345, 346, 347, 351–353, 356, 367–371 Super-resolution fluorescence microscopy .........................35, 131, 141, 145, 162, 164, 173, 178, 219, 272, 290, 303 imaging probes .......................................164, 177–178 microscopy............................................. 202, 211, 212 Surface-plasmon resonance .......................................... 174 Sweat bee, see Megalopta genalis Sylvius (de le Boe¨), Franc¸ois............................................. 4 Synapse ..............................................................3, 8, 9, 13, 14, 15, 29, 31, 32, 34, 104, 126, 133, 189 Synapsin ........................................ 78, 81, 88, 95, 96, 104 Synaptic link ............................................................................. 75 terminal..................................... 13, 15, 21, 22, 35, 36 Synchrotron................................203, 306, 307, 309, 310 Syncytium ..................................................................11, 39 System camera .............................................382, 384, 385, 392, 393, 394, 396

NEUROHISTOLOGY T Tapetal cup .................................................................... 383 T4 bacteriophage .......................................................... 308 Telepathology dynamic.................................................................... 445 static ......................................................................... 445 Telephoto ............................................................. 386, 387 Temporal resolution............................................. 350, 426 Tetrahydrofuran (THF) .......................................170, 185, 189, 190, 192–194 Tetraline......................................................................... 185 Tetrodotoxin ................................................................. 367 Therapeutic agents ........................................................ 159 Thermal noise........................................................411, 413 2,2’-Thiodiethanol (TDE) ..............................84, 85, 187 Thioflavin S/T (ThS/ThT).......................................... 160 Thread cell ............................................................ 232, 233 Three-chip (three CCD)...................................... 408, 422 Thy-1 EGFP-M ......................................................... 193, 195 promoter......................................................... 305, 335 (Tissue) clearing .................................................. 8, 23, 25, 77, 81, 84, 85, 183–189 T-mount ........................................................................ 420 Tobacco budworm, see Heliothis virescens Tomographic microscopy ............................................. 184 Tonal balance............................................... 416, 434, 435 To¨pler, August............................................................... 278 Total internal reflection (TIRF) ........................... 33, 208, 209, 303, 304, 305, 314, 319 Tract tracing ..............................................................51–53 Trans-anethole............................................................... 191 Transfection ......................... 29, 135, 136, 148, 150, 350 Transform plane .......................................... 279, 281, 318 Transformation affine (anisotropic) ............................. 80, 90, 99, 100, 104, 115, 116 center ......................................................................... 90 diffusion..................................................................... 80 elastic ....................................................... 99, 100, 115 local rigid ................................................................... 90 non-rigid.............................. 80, 90, 91, 99, 115, 116 rigid....................................................90, 91, 116, 117 Transgenic (mouse) ............................................ 125–151, 159, 160, 162, 164–167, 170, 171, 178 Transmission axis ........................................................................... 213 of light ............................................................ 209, 214 Transport anterograde.................................................21, 22, 360 bulk flow .................................................................... 22 centrifugal.................................................................. 21 centripetal .................................................................. 22

AND IMAGING

TECHNIQUES Index 471

fast anterograde ......................................................... 22 retrograde .................................................22, 127, 148 slow anterograde ....................................................... 22 speed .......................................................................... 22 Traviss (expanding) stop ............................................... 234 Triangulated surface model ...................................77, 118 Tribolium castaneum (red flour beetle) ................. 78, 81, 82, 83, 84, 90, 102, 117 Trigeminal nerve .................................................. 309, 310 Triple immunofluorescence ........................ 140, 142, 143 Triplet excitation ........................................................... 168 Triton X-100 ................................................ 84, 142, 184, 185, 187, 352 TWAIN (compliance) ................................................... 423 Two-photon microscopy................................................... 7, 26, 161, 164, 325, 326 probe............................................................... 164–169 Tyrosine hydroxylase..................................................... 350

U Ultramicronen ............................................................... 326 Ultramicroscope (-y) slit-based ......................................................... 328–330 slit-free ............................................................ 330–332 Ultropak ............................................................... 235, 236 Uncinaria ...................................................................... 264 Unsharp masking .......................................................... 436

V Vaa3D (or V3D) ....................................................98, 118 Value-added pathology ................................................. 446 VAREL (contrast) ...................................... 294, 296, 297, 302, 319 Vasculature..................................................................... 255 Vectashield .................................................................84, 85 Vector diagram .........................................281, 284, 285, 290, 291, 292, 294, 299 of electric field (see Electric field) field ..........................................................................118 GFP.......................................................................... 305 graphics.................................................................... 442 virus.......................................................................... 139 Vein ................................................................................ 267 Vertical illuminator .............................236, 266, 268, 272 Vesalius, Andreas ............................................................... 4 Vesicular transporter ......................................35, 148, 151 Vessel....................................................3, 9, 164, 196, 255 VIB protocol ......................................................... 90, 102, 103, 117, 118 Vibratome..................................................................95, 97

NEUROHISTOLOGY

472 Index

AND IMAGING

TECHNIQUES

Video.................................................. 249, 251, 263, 266, 276, 277, 307, 308, 380, 382, 384, 389, 393, 395, 398, 404 Video camera....................................................... 380, 404, 405, 407, 409, 411, 420, 422–423, 426, 429, 432 Video hybrid imager ..................................................... 423 Viewfinder ...........................................390, 391, 394, 421 Vignetting....................................................386, 394, 413, 427, 428, 430, 431, 432 Virchow, Rudolf Ludwig Carl ........................................ 10 Virtual microscopy............................ 421, 439, 443, 444, 446 slides...................................... 421, 422, 444, 446, 448 Virtual Fly Brain...................................... 80, 90, 102, 116 Virtual Insect Brain (VIB) ....................... 80, 89–92, 102, 103, 117, 118 Vision, color, see Color vision Visual cortex ............................................ 4, 131, 297, 298 Visual cues ..................................296, 297, 298, 307, 315 Vitreous body ................................................................ 394 Voltage-dependent ion channel (VDIC) ....................... 33 Voltage-sensitive dye ......................................32, 350, 354 Volumetric analysis............................................... 102, 103 Voxel ..................................................... 79, 88, 89–93, 94, 95, 96, 98, 103, 104, 107, 109, 112, 116, 118, 132, 310

phase shift ......................................280, 281, 291, 292 Weigert, Carl ...............................................................8, 19 Weigert stain................................................................8, 19 White balance ............................................. 263, 397, 398, 405, 430, 431 Whole-mount preparations ....................... 23, 77, 81, 84, 85, 88, 95–97, 185, 186, 187 Whole slide imagers ................................... 421, 430, 441, 445, 446, 448 Willis, Thomas................................................................... 4 Wireless shutter switch.................................................. 392 Wratten-Rheinberg filter............................................... 232 WYSIWYG ............................................................ 422, 423

W

Z

Waldeyer-Hartz (von), Heinrich Wilhelm Gottfried ...... 8 Watermarks ........................................................... 441, 442 Wave amplitude ........................................................ 205, 210 nature of light................................................. 204, 205 optics........................................................................ 203 particle duality ......................................................... 203 phase ........................................................................ 210

Zamboni’s........................................................................ 52 Zebrafish ...................................................... 237, 255, 383 Zernike, Frits .............................................. 276, 277, 279, 285, 286, 287, 307, 308, 319 Zinc ..................................................................... 78, 81, 83 Zsigmondy, Henry ........................................................ 326 Z-stack(er) ............................. 27, 28, 133, 305, 306, 449

X Xanthophores ................................................................ 265 X-ray...................................................................... 306–309 X-Trans model...................................................... 381, 397 Xylene ........................................................... 6, 18, 21, 63, 64, 65, 83, 85

Y Yellow fluorescent protein (YFP) ........................... 31, 33, 174, 188, 271, 272 Young, Thomas .................................................... 202, 211