UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Macro-glial specialization in the brain Thompson, Sharleen Grace 1986-06-30

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


831-UBC_1986_A6_7 T46.pdf [ 10.59MB ]
JSON: 831-1.0096827.json
JSON-LD: 831-1.0096827-ld.json
RDF/XML (Pretty): 831-1.0096827-rdf.xml
RDF/JSON: 831-1.0096827-rdf.json
Turtle: 831-1.0096827-turtle.txt
N-Triples: 831-1.0096827-rdf-ntriples.txt
Original Record: 831-1.0096827-source.json
Full Text

Full Text

MACRO-GLIAL SPECIALIZATION IN THE BRAIN BY SHARLEEN GRACE THOMPSON B. SC. The University of British Columbia, 1976 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF PSYCHIATRY, DIVISION OF NEUROSCIENCES We accept thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA OCTOBER 198 6 (c)Sharleen Grace Thompson, 1986 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date THESIS ABSTRACT This thesis examines the evidence for glial cell specialization. It starts with an historical description of the development of ideas about glial cells, demonstrating how each technological advance allowed an increase in understanding of various morphologically different types of glial cells and how each technique provided more evidence for glial heterogeneity. The most spectacular recent development is the increasing evidence for biochemical heterogeneity in cells in vivo, in different cell lines and in primary cultures from various regions. These biochemical differences have been found both among cells that are morpholgically similar and between different cell types. The results of three experiments which provide direct or indirect evidence for glial cell heterogeneity are presented. The first experiment is an anatomical analysis of the cellular localization of hemosiderin in rat brain. The results show primary localization to oligodendrocytes but not all oligodendrocytes as there are distinct regional differences in both density and numbers of oligodendrocytes staining. In the second experiment, an alternate route of glutamate formation from proline or ornithine via 1-pyrroline dehydrogenase is demonstrated and shown to be present in only a small subset of glial cells and not in other cell types. In the third experiment the glial heterogeneity concept is used to provide an alternate interpretation of all data on biochemical effects of thiamine deficiency in rat brain. i i The conclusion summarizes the contribution of the experiments to the already strong evidence for glial heterogeneity and suggests ways that assumptions of glial heterogeneity rather than homogeneity could affect research the neurosciences. TABLE OF CONTENTS Title Page i Thesis Abstract ii Table of Contents v List of Tables v List of Figures vi Table of Abbreviations viDefinition of Glia Cell Types viii-xii Introduction 1 History of Development of Today's Ideas on Structure and Function 1 Function of Glia 5 Glia and Neurotransmitters 11 New Techniques Enabling Advances in Understanding Glia 2(a) Tissue Culture 1 (b) Freeze Fracture Techniques 2(c) Markers 25 i) Fibrous Protein 2ii) Glutamine Synthetase 29 iii) Carbonic Anhydrase 3 0 iv) Other Markers 31 Glia Heterogeneity-Morphological 34 Developmental Differences - a source of Heterogeneity 4Heterogeneity in Tissue Cultures 49 (A) Developmental Changes in Culture 49 (B) Culture Conditions, Development and Heterogeneity 53 (C) Cell Development and Differentation in Response to Injury 61 (D) Heterogeneity Between Different Glia Not Explained By Development or Cultural Conditions 63 Heterogeneity Between and Within Glia Cell Lines: Different Areas Labelling In Vivo 65 Differences in Glial Cells from Different Areas of the Brain 77 Summary of Evidence for Biochemical Differences in Glia 86 Experimental Rationale and Abstract 88 Experiment 1 91 Experiment 2 117 Experiment 3 12Conclusion 143 Acknowledgements 144 References 5 - iv -LIST OF TABLES Table I Minor astrocyte cell marker Pg. 3 2-33 Table II Oligodendrocyte and myelin markers Pg. 35-36 Table III Effects of culture conditions on cell characteristics Pg. 59-60 Table IV Comparative values of glutamate uptake Pg. 70 Table V Comparative values of glutamine uptake Pg. 72 Table VI Comparative values of high affinity GABA uptake Pg. 73 Table VII Iron staining in rat brain by area Pg. 115-116 Table VIII Enzyme levels in control, thiamine deficient and recovered rats Pg. 136 - v -LIST OF FIGURES Fig. 1 Microscopic pictures of iron staining in rat brain. Pg. 104, 106 Fig. 2 Photographs of whole sagittal sections of iron staining in rat brain. Pg. 108 Fig. 3 Schematic diagram of Fig. 2 showing cellular types by area Pg. 110 Fig. 4 Half photographs and half schematic drawings of coronal section of iron staining in rat brain. Pg. 112, 114 Fig. 5 Schematic representation of the conversion of proline and ornithine to glutamate and GABA. Pg. 124 Fig. 6 1-Pyrroline dehydrogenase staining of Bergmann glia in cerebellum Pg. 12 6 Fig. 7 1-Pyrroline dehydrogenase staining of astrocytes of dentate gyrus. Pg. 12 6 Fig. 8 Proline oxidase staining of Bergmann glia in cerebellum Pg. 12 6 Fig. 9 GABA-T staining in thiamine deficient rat Pg. 142 - vi -TABLE OF ABBREVIATIONS FULL WORD ABBREVIATION 1ST PG. USED Acetylcholine ACh 15 Acetylcholinesterase AChE 15 Adenosine Triphosphate ATP 79 Adenosine-5-Triphosphatase ATPase 10 Calcium Ca++ 12 Catechol-O-Methyl Transferase COMT 15 Central Nervous System CNS 6 Choline Acetyltransferase CAT 89 Cyclic Adenine Monophosphate CAMP 18 Diaminobenzaldehyde DAB 89 Dibutyryl Cyclic Adenine Monophosphate dBcAMP 17 Dopamine DA 11 Electroencephalogram EEG 89 Gamma-Aminobutyric Acid GABA 9 Gamma-Aminobutyric Acid Transaminase GABA-T 15 Glial Fibrillary Acidic Protein GFAP 25 Glutamic Acid Decarboxylase GAD 89 Glutamine Synthetase GS 16 Histamine Type I Receptor HI 18 Histamine Type II Receptor H2 18 Magnesium Mg++ 76 Maximum Velocity of Reaction Vmax 12 Michaelis Constant (Concentration of Substate at 1/2 Vmax) KM 70 Monoamine Oxidase MAO 15 Niacine Adenine Dinucleotide NAD 117 Noradrenaline NA 11 Ornithine Oxo-Acid Aminotransferase OrnT 117 Potassium K+ 8 Pyrithiamine PT 127 1-Pyrroline-5-Carboxylate P5C 89 Pyrroline-5-Carboxylate Dehydrogenase Pro 89 1-Pyrroline Dehydrogenase PDH 89 Serotonin 5HT 11 Sodium Na+ 11 Thiamine Deficient TD 127 Thiamine Triphosphate TTP 128 Tricarboxylic Acid Cycle TCA CYCLE 17 - vii -DEFINITIONS OF GLIAL CELL TYPES OLIGODENDROCYTES A class of glia cell first stained and seen by Golgi after he invented his silver staining technique. There is considerable morphological heterogeneity within this group. They are traditionally classified by where they are located and how they associate with other cells or by their nuclear and cytoplastic densities. ASTROCYTES The second major class of glial cells. They are larger than oligodendrocytes, have pale staining nuclei and electron light cytoplasm. Cajal, the first to describe them divided them into two subclasses: fibrous and protoplasmic, based on the presence or number of fibers within the cell body. Those described by Cajal are now considered as OOastrocytes. £-ASTROCYTE May be intermediate type between an OC-astrocyte and light oligodendrocyte. - viii -MICROGLIA A class of glia originally defined by the silver carbonate method of Del Rio Hortega. Their origin and classification is highly controversial. They are not discussed in this paper. DISTINCT SUB-CLASSES OF GLIA BERGMANN GLIA Also called Golgi Epithelial Cells; they are glial cells with cell bodies located in or just below the Purkinje cell layer of the cerebellum and having radiating fibers extending upward through to the outer surface of the molecular layer. They share many of the development and biochemical characteristics of glia but also have many differences. MULLER GLIA CELLS A glia cell in the retina of the eye. Though not the only glia cell in the eye, they have been extensively studied and have considerable overlap of characteristics with glia of the central nervous system. - ix -RADIAL GLIA CELLS PITUICYTES EPENDYMAL CELLS TANYCYTES A developmental stage of many glia cells where the cell body has long arms extending perpendicularly to some outer surface. These radiating fibers may assist in guiding other cells to their correct place during development. In some areas the radial form may persist until adulthood. The Bergmann glia may be an example of this. Pituicytes are glial cells in the neurohypophysis. They have many characteristic of central glia. These glia-like cells line the ventricular system within the brain and central canal of the spinal cord. They may have special function in blood brain barrier, and production of cerebrospinal fluid. They have many characteristics of glia cells and may evolve from radial glia. Specialized glia with radiating processes that line the ventricle, particularly the third ventricle. They have many glia characteristics. - x -ENTERIC GLIA Glia-like cells of the enteric nervous system that are more like the central glia than the peripheral Schwann cells SCHWANN CELLS Cells of the peripheral nervous system that wrap the peripheral nerves with layers of their external membrane to insulate nerves from each other. In special cases, such as those in the eye, they grow with the optic nerve into the brain and are located centrally. Occasionally, as in the Schwann cells of the olfactory nerve, they have properties similar to the central glia. RESEARCH CELL TYPES GLIA CELL LINES Permanent cell cultures maintained in laboratories and originally created by transformation by certain viruses or chemicals. They are believed to be models of brain tumors and thus to certain characteristics of various types of brain tumors, and are extensively used in research because the lines are stable and can be purchased. They have well defined characteristics which cannot be assumed to be like those of untransformed glia cells in vivo but some normal characteristics have been retained. Some lines, such as C-l or C-6, may have characteristics of gliomas, while others may resemble neural tumors or astrocytomas. PRIMARY CULTURES Cultures recently derived from fetal or neonatal brain and cultured for short periods of time. During this time they develop through several changes of morphology and biochemical characteristics that can be manipulated by culture conditions. They are thus useful in trying to understand the characteristics of glia. Since culture conditions are never identical to those in vivo, many in vivo characteristics never develop. - xi i -Introduction This thesis examines glial cell literature for evidence of glial heterogeneity and then presents my results on specific glial staining and the effects of experimental manipulation on subsets of glia. The results show considerable evidence for macroglial heterogeneity, both on a regional and a cellular basis. My research shows, in two unrelated procedures, that only certain subsets of glial cell are stained, further supporting the evidence for biochemical differences between glial cells. The current results suggest that generalization from one glial system to another is no longer valid. If glial cell heterogenity exists, why is the evidence so late in coming, and why is there tremendous resistance to the acceptance of this idea? Much of our understanding of glial cell function is based on work done very early in this century. The early assumptions were so well accepted that more recent results have largely been ignored by neuroscientists. Basic neuroscience texts still do not devote more than a small amount of space to glia, giving little more than a simple description of the basic types and perhaps an historical note on their function. History of Development A look at the historical work done on glia will serve to introduce the topic of glial cell structural and functional - 1 -heterogeneity. The early researchers faced a number of problems which led to assumptions that formed biases which now prevent the acceptance of some of the findings of heterogeneity. Virchow in 1846 was the first to mention the existence of neuroglia in the brain. He thought that, since neurons did not appear to occupy all the space in the brain, there must be something holding the neurons together; this he called the nerve glue, or neuroglia and the German word came to be adopted. He did not see that the "glue" was composed of cells because the early cell preservation techniques were crude and neuroglia were the first cells to swell and disintegrate, which made them difficult to see under light microscopes. Historically the main reason for concentration on neurons was the difficulty in studying glia. The fact that the spaces between the nerves were occupied by cells now called glia was first observed by Golgi (1879) after he invented the Golgi silver staining method for those glial cells now called oligodendrocytes. In 1913 Cajal invented the gold sublimate method which he found stained another type of non-neural cell, the astrocyte, thus allowing the differentiating of two types of cells, the oligodendrodcytes, stained best by the Golgi technique, and the astrocyte. In 1919 del Rio Hortega invented the silver carbonate method which stained microglia, the third major type of glial cell. Although there were advances in the understanding of the development of these cells, there were no major additions - 2 -to the understanding of cell types until cells were first separated, the electron microscope was invented and cell specific markers became available. I have not included further discussion of the microglia in this thesis because I have no research to present on microglia and they are surrounded by considerable controversy. Oligodendrocytes are now understood to be small oval cells that comprise about 20% of the brain mass (Varon, 1978). Oligodendrocytes come in several different morphologies and have been classified by two methods. The first method is by where they are located and how they associate with other cells. Oligodendrocytes occur in rows in white matter, called intrafasicular glial, where their processes are associated with groups of myelinated nerve fibers. In grey matter, they may appear as independent or as satellite cells in close association with neurons. Oligodendrocytes are found to have a wide range of nuclear and cytoplasmic densities (Caley and Maxwell, 1968). On the basis of these electron microscopic densities, Mori and Leblond (1970) classified them into light, medium and dark oligodendrocytes, which seem to be progessing developmental stages (Mori and Leblond,1970) from light to dark, with concurrent reductions in the size and increase in the density of the nucleus, reductions in cytoplasmic volume, increasing complexity of rough endoplasmic reticula and Golgi organelles, and reduction in the number of processes. Their development parallels myelination. The final developmental product, the mature oligodendrocytes, are characterized by a - 3 -small electron-dense nucleus, often positioned eccentrically, scant and dense cytoplasm with a highly developed Golgi apparatus, stacks of rough endoplasmic reticula cisternae and lamellar bodies frequently associated with intracellular membranes, and a small group of processes containing microtubules but no gliofilaments. Astrocytes are the second of the major classes of glia. They comprise 20 to 25% (Varon, 1978) of brain tissue, are larger than oligodendrocytes, have pale staining nuclei and electron light cytoplasm, have numberous processes with glial filaments and accumulate glycogen granules under anoxic conditions. Cajal originally divided them into fibrous and protoplasmic types, based on location, morphology and function. Now, with the electron microscope, mature fibrous astrocytes are noted to have extensive, well organized cytoplasmic filaments (Palay et al. 1962), and protoplasmic astrocytes do not. There are also a wide variety of other cells with astrocyte characteristics. These will be discussed more fully when glial heterogeneity is being examined in a following section. In the early years several theories were put forward as to the function of neuroglia. Golgi (1894) thought that they nourished neurons because he observed that they had end feet that were opposed to capillaries. In 1885 Weigart (1895) suggested their function was to give a structural support. In 1896 Marinesco suggested that they had a histolytic role in the clearing of dying neurons. His, in 1887, was the first to view glial cells as providing guidance for the growth of the - 4 -nerve fibers during embryological development. Lugaro (19 07) was the first to speculate on their function as a detoxification filter between blood and brain and also suggested that they served to remove and chemically split compounds secreted by nerve endings. Lugaro rejected Golgi's nutritional hypothesis because he did not believe that dendrites were like roots to plants, and he also rejected their role in providing biochemical support for neurons since he still thought that they were basically the packing material for the more "noble" neurons. Cajal (1913) thought that they served to insulate nerve fibers and fiber bundles. These concepts of glial function remained intact until Kory et al. (1958) isolated glial cells and the electron microscope stimulated ultrastructural research. Function of Glia We now understand that the functions of glia are complex, but the currently understood functions include some of those assigned to glia by many of the earlier researchers. In order to understand glia heterogeneity their basic functions must be understood. The concept of structural support, as proposed by Weigart (1895), is no longer seriously thought of as a function even though the historical fact is frequently mentioned in texts without much elaboration. In fact, except when extensive gliosis has formed scar tissue, glia are perhaps softer than neurons as they are more suspectible to ischemia and mechanical disruption. They do, however, perform several - 5 -structural functions. The oligodendroglia do wrap the nerve fibers in the brain and spinal cord with many layers of their cell, membranes, forming the central nervous system (CNS) myelin. Glia first appear just prior to the time of myelination. The most rapid rate of myelination is synchronous with the most rapid proliferation and differentiation of oligodendrocytes. In humans, myelination starts at about 4 months gestation and continues till about age 2; it starts at the neuronal cell body and grows distally. This serves to speed the saltatory conduction of electrical impulses along the nerves and provides some structural strengthening of these delicate fibers. In the white matter the oligodendrocytes do provide a warp and woof like matrix with the nerve fiber bundles. Their end feet and dendrites also provide sheaths over all outer surfaces of the CNS. They do have many types of connections between their plasma membranes and some types of connections may provide structural support. Mariesco's original proposal of histolytic activity for the removal of dying neurons may be correct in that glia may absorb the debris of the dying neurons, but most of this seems to be done by macrophages that invade the area of damage. This does not mean that glia are not involved in the activity at damaged sites. When repairs are needed they proliferate, increase in size and change structure so that they are more fibrous and thus form a tough scar tissue at the core of the damaged area. They also wall off the damaged area of the brain from the overlying leptomeningeal cells. In fact it was - 6 -this scar tissue that gave us the original concept of a structural support role for glia. Cajal's idea that glia serve to insulate and isolate nerve fibers and bundles is still highly supported today. Not only are the nerves myelinated for increased efficiency but the synaptic terminals are also separated from each other by glial cells. Thin astrocytic processes break up the neuropil into mosaics of small regions each containing a synaptic field. A similar parcelling occurs around clusters of synaptic terminals. Astrocytic processes often intervene between cell types or neuronal groups, isolating neuronal surfaces in such a way as to prevent flow of impulses in a haphazard manner (Lasansky, 1971). These areas have many membrane specializations and seem to have continuous dynamic alterations (Wolff and Guldner, 1978) as if actively involved in the isolation procedure. Lugaro's (1907) concept of a detoxification filter can be compared to the minor role that the astrocytic end feet may play in the blood brain barrier. They are no longer believed to provide a barrier around capillaries but, because most in coming chemicals must go through their non-occluding and non-continuous junctions, they may have the first opportunity at the selection of incoming chemicals. Golgi (1894) originally thought that the glia provided biochemical support and nutrition for the neurons. This idea has evolved to include several different concepts. Satellite oligodendrocytes (those cell bodies lying near long axons) may be involved in neuronal nutrition: Freide, (1966) thought of - 7 -them as auxiliary metabolic units for the axons of neurons. Astrocytic end feet may be involved in the transport of substances inward to neurons and glia at the centre of the brain mass. In many cases neurons and glia do have complimentary metabolisms. Cultured neurons only survive a few days without glia unless nerve growth factors or brain extracts are added. Medium that has surrounded glia will support neurite growth (Ebendal and Jacobson, 1975). This means some soluble factor(s) must be involved in maintaining the neurons but this may not be a nutritional substance. Contrary to the classical assumptions on glia function, the metabolic rate of glia cells is now known to be quite high. Hertz (1978) showed that the early work done on glial cell lines and glial scar tissue had given erroneously low metabolic rates for glia. Energy metabolism in some types is comparable to that of neurons (Hertz, 1982). Glia have the majority of oxidative enzymes, and also have reductive enzymes, although astrocytes are lower in oxidoreductase enzymes than are oligodendrocytes. The rate of protein synthesis suggests that the astrocytes are making a considerable proportion of the total brain protein (White and Hertz, 1981). Oligodendrocytes have a higher oxygen utilization than astrocytes and consume much of the oxygen in white matter (Pevsner, 1979). Cummins et al. (1979) showed that the uptake by glia of 2 radioactive metabolism markers can be increased by potassium (K+) or high levels of neurotransmitters, indicating a metabolic responsiveness to their environment. - 8 -Glia may not only interact with neurons but may have some of the ion conductance and receptor properties traditionally associated with neurons. Astrocytes have a resting membrane potential that is slightly higher than that of neurons, being 70-90 mV, and that varies with the external K+ levels according to the Nernst equation (Pevzner, 1979). Thus they may have some role in the production of extracellular current. Bowman and Kimelberg (1984) showed that astrocytes can be depolarized in primary culture, a property previously thought to be exclusive to neurons. They depolarize in the presence of 0-aminobutyric acid (GABA), L-glutamate, D- and L-aspartate and kainic acid in culture. In vivo the amino acids have been reported to depolarize all astrocytes lying in the vicinity of neurons. But substances that reversibly block K+ conductance abolished the depolarization of glial cells (Bowman and Kimelberg, 1984). Therefore the glial cells may not have receptors for these amino acids, but depolarized because of the efflux of K+ from the neurons. This subject is still controversial. Lugaro (1907) was the first to postulate that glia remove and catabolize substances released from nerves. Now we know of many substances that are removed from the synaptic cleft by glia and understand such actions to be of major importance in the functioning brain. Glia also seem to control extracellular K+ levels. They function as a fine tuning mechanism after the neurons do most of the uptake. It can be shown that glia could take up enough K+ to clear the excess that leaks out of neurons but whether - 9 -they actually do this is still in question. K+ released from neurons causes an increase in extracellular K+ which eventually causes the neurons to fire (Prince, 1978). The glial sheets act as dams restricting the diffusion of K+. Glial cells remove K+ and minimize the spread of K+ to other regions, thus acting as a buffer zone (Trachtenberg and Pollen, 1970). Glia are ideally suited for this role because they have a high resting membrane potential, are selectively permeable to K+, are electrically excitable, and have irregular bodies with many processes. This uptake process may be by active transport because glia have an adenosine-5 triphosphatase (ATPase) that is specifically activated by K+ (Franck et al., 1978, Grisar and Schoffeniels, 1978, Grossman, 1978, and Prince et al., 1978). This ATPase is as sensitive as neuronal ATPase to ouabain (Walz and Hertz, 1982). The concept of an active role for glia in K+ homostasis has three prerequisites. First, that the K+ released from neurons leads to a build-up of extracellular K+; this is unanimously accepted. Second, that the excess K+ is removed by surrounding cells, not through diffusion; this has now been demonstrated and probably requires energy as it has been shown that excess K+ leads to a transient increase in respiration in microdissected glial cells, bulk-prepared astroglia, or cultured astrocytes if the cultured cells are 3-4 weeks old (Hertz, 1982) . Third, that K+ uptake into astrocytes is more intense than into neurons and is further increased by an increase in the extracellular K+ level beyond the resting level; the large K+ content in astrocytes and high membrane - 10 -potentials unequivocally show that these cells are able to accumulate large amount of K+. Hertz and Chaban (1982) showed that astrocytes have uptake rates higher than neurons in cultures. This thus satisfies the third criteria of active glial transport. They also showed this uptake is inhibited by ouabain in both C-6 cells (a much studied glial cell line) and primary cultures of astrocytes, indicating that a sodium (Na+) -K+ exchange catalyzed by the Na+,K+-ATPase exists. Although uptake is reduced by oubain, it is not completely abolished. Therefore another mechanism must exist which oubain does not inhibit; this mechanism is probably dependant on carbonic anhydrase because acetazolamide, an inhibitor of carbonic anhydrase, inhibits some potassium uptake into cells (Hertz and Chaban, 1982). Glia and Neurotransmitters Glial cells also seem to be involved in many aspects of neurotransmitter function. As Lugaro (1907) suggested, glia may take up substances released by neurons. They may actively take up neurotransmitters by mechanisms that are not always identical to the uptake into neurons. They are capable of catabolizing some of these neurotransmitters and there is evidence of receptors on some types of glia. A. Active uptake of several neurotransmitters has been demonstrated in glial cell populations. Using primary cultures of astrocytes, Schousboe (1978) showed that glia can take up noradrenaline (NA), dopamine (DA), and serotonin - 11 -(5HT) . Hertz (1982) showed that this occurred in an energy-dependent manner requiring both Na+ and K+. Hansson et al. (1984a) did not confirm this data for DA. Several of the amino acid neurotransmitters also seem to be taken up by glia. Levi et al. (1982) and Wilkin et al. (1982) showed that the uptake of amino acids by slices of cerebellum was predominantly into astroglial cells rather than neurons. A high affinity uptake of glycine into glia has been repeatedly demonstrated (Hokfelt and Lungdahl, 1971, Matus and Dennison, 1971, Henn, 1976). GABA has repeatedly been shown to be taken up by glial cells (Henn, 1976, Currie and Kelly, 1981). Hansson et al. (1984b) and Larson et al. (1980) showed that this uptake was Na+ dependent. There is controversy over whether glial uptake of GABA is greater or less than into neurons. Balcar et al. (1982) found the uptake into glia less intense than into neurons, whereas Schousboe (1978) calculated, based on the maximum velocity (Vmax) in astrocytes in primary culture, that the rate could be 2 to 6 times higher than into neurons and could be increased by calcium (Ca++) or low K+. Schousboe (1981) also reviewed the work of many others and found that cultured astrocytes exhibit a Vmax comparable to that found in brain slices. L-Glutamate and D- or L-aspartate appear to share common transport systems. High affinity uptake of glutamate or D-aspartate into glia has been repeatedly demonstrated autoradiographically (Hokfeld and Ljungdahl, 1972, Schon and Kelly, 1974, Lasher, 1975, McLennan, 1976, Currie and Kelly, - 12 -1981). The glial uptake requires the presence of both Na+ and K+ and is both energy and temperature dependent (Schousboe, 1978) . Uptake of glutamate has also been demonstrated into many cultured cell lines (Hamberger, 1971, Henn et al., 1974, Weiler et al., 1979). It has also been demonstrated in astrocytes prepared by gradient centrifugation (Faivre-Baumann et al., 1974, Henn et al., 1974, Balcar et al., 1977, Pfeiffer et al., 1970), including astrocytomas (Snodgrass and Iversen, 1974), retinal Muller cells (Bruun and Ehinger, 1974, White and Neal, 1976) and astrocytes in primary culture (Schousboe et al., 1977b, Hertz et al., 1979, and Balcar and Houser, 1978). Schousboe (1978) showed that the glutamate Vmax for astrocytes in primary culture was much higher than that for glutamate uptake into sensory ganglia or glial cell lines. In some astrocyte cell lines it may be high enough to keep pace with release from neurons and also high enough so that the glutamate may be their only fuel source (Hertz, 1979). In some of the other glial cell lines this rate is probably not enough to provide the sole fuel source. We know that transmitter uptake in glia is not a general phenomenom because neutral amino acids other than GABA glycine materials, and closely related materials, have no high affinity uptake into glia. B. Although the uptake of neurotransmitters has been demonstrated, there is evidence that some of this uptake is not identical to that which occurs in neurons in either - 13 -character or quantity. The uptake of monoamines into primary cultures, for instance, is at lower rates than into neuronal cultures (Hertz, 1982) but the uptake of L-glutamate is higher (Schousboe, 1978a). However a large number of glial cell lines show high affinity uptakes similar in rate to those of neurons (Edwards et al., 1979). The mechanism can also be quite different. For example, Waniewski and Martin (1983) found that 4-acetamido-41-isothiocyano-2,2'-disulfonic acid stilbene, an inhibitor of anion exchange, was a potent and selective inhibitor of L-glutamic acid uptake by cultured glioma cell line and rat brain astrocytes but did not affect synaptosomal uptake. Therefore glutamate tranport systems differ between neurons and glia. Ramaharo-Brandro et al. (1982) report that neuronal and glial glutamate carriers exhibit differences in terms of both substrate specificity and in terms of dependency on mono- or di-valent cations. Only neuronal uptake is dependant on both Na+ and Ca++, and is therefore more susceptible to changes in external ionic concentrations. Furthermore, astrocyte uptake of glutamate was found to be non-competitively inhibited by D-aspartate whereas uptake by granule cells was competitively inhibited. Uptake of glutamate in astrocytes from prefrontal cortex was coupled to 1 Na+ ion in contrast to 2 for the granule cell. Glial cells exhibited no K+ induced release of glutamate in contrast to neurons (Drejer et al., 1982). A similar difference between neuronal and glial uptake can - 14 -be shown for GABA uptake. Kelly and Dick (1978) showed that -alanine is a specific blocker of GABA uptake in glia but not neurons and a cyclohexaneamine derivative is a blocker specific for GABA uptake into neurons. Thus the two GABA uptake systems are biochemically different. Glial cells can also release some neurotransmitters and this release can be demonstrated to be different in some cases from that of neurons. C. Glia cells can also possess catabolic enzymes. Although acetylcholine (Ach) uptake has not been demonstrated into glia or neurons, acetylcholinesterase (AChE) activity can be found in certain clonal lines of glial cells (C-6) (Vernadakis and Arnold, 1980). Glia also possess higher specific activities of the monoamine catabolizing enzymes, monoamine oxidase (MAO) (Hazama et al., 1976, Hansson and Sellstrom, 1983) and catchol-O-methyl transferase (COMT) than found in whole brain. The presence of COMT and MAO has been shown in several glial cell lines which suggests they have the ability to inactivate catecholamines (Silberstein et al., 1972). The degradative enzyme for GABA, GABA transaminase (GABA-T), has been demonstrated in glia. Bulk prepared glia and cultured astrocytes stain for GABA-T (Sellstrom et al., 1977, Tardy et al., 1978), as do astrocytes cultured from neonatal brain (Schousboe et al., 1972). GABA-T activity in glial cultures, however, is lower than in cerebral hemispheres (Hansson and Sellstrom, 1983) and lower than in neurons (Kelly - 15 -and Dick, 1978). The glutamate eventually generated by the breakdown of GABA and from other sources can also be catabolized in glia. In fact glutamine synthetase (GS), the degradative enzyme which plays a major role in the chemistry of glutamate, is found only in astrocytes. For a number of years glutamate has been thought to exist in two pools, one in the neurons from which glutamate is released when the neurons fire, and a second smaller pool in the glia, where glutamate is converted into glutamine with the help of GS. This glutamine is then released to be taken up by neurons and reconverted to glutamate. This schema remains highly controversial. The restriction of GS activity in the brain in vivo to astrocytes (Norenberg and Martinez-Hermandez, 1979) and the high activity of this enzyme in primary cultures of astrocytes (Schousboe et al., 1980) are consistent with the concept that any glutamate accumulated in astrocytes, is to a large extent converted to glutamine. Some researchers, however, have found GS activity low in glial cells relative to whole brain (Nicklas and Browning, 1978) but this is a function of the age of the glial cells. High GS activity matures late in development. In accordance with the late maturing of GS, the rate of glutamine synthesis is faster in 3 week cultures than 1 week old ones but it does not increase in response to di-butyl cyclic adenine mono-phosphate (dBcAMP) which is generally thought to cause maturation. Two other glutamate-metabolizing enzymes, glutamate dehydrogenase and glutamate oxaloacetate transferase, which - 16 -convert glutamate to °\ -ketoglutarate, are also present in astrocytic cultures at high activities (Schousboe et al., 1980a). This suggests that glutamate accumulated in astrocytes may be converted to tricarboxylic acid (TCA) cycle constituents and thus be a metabolic substrate. This would mean that the glutamate to glutamine to glutamate loop would not be completed. Support for such an alternative route was supplied by studies of the fate of radioactive glutamate in developing cultures of mouse astrocytes (Potter et al., 1982); the radioactivity of glutamine never exceeded that of its precursor glutamate indicating the other metabolic routes must exist. Possible roles for glia in helping to dispose of peptide neurotransmitters have not yet been widely investigated. However, Lentzen and Palendker (1983) showed, by using specific enzyme inhibitors and examining the products, that glial cells have the ability to degrade enkephalin, a peptide neurotransmitter, by both aminopeptidase and membrane bound enkephalinase A. D. There is considerable evidence that some glia can possess receptor or binding sites for various neurochemicals. For example, on selectively destroying the Muller glia cells in the retina, Memo et al. (1981) were able to show a selective loss of DA and 5HT binding sites, suggesting that Muller cells carry these receptors. Henn and Henn (1980) showed dopamine binding sites on astrocytes that are linked to adenylate cyclase and stimulate - 17 -cAMP formation, which is blocked by antipsychotic drugs. Astrocytes prepared from areas rich in dopamine show dopamine binding that can be blocked by antipsychotic drugs (Hansson et al., 1984), and antipsychotic effectiveness is correlated with their ability to displace dopamine (Hertz, 1981) in bulk prepared cells. The potency of the drugs in blocking the formation of cAMP in astrocytic cultures is also said to be well correlated to their effectiveness as antipsychotics; this is not true with neuronal preparations where the antipsychotic action seems more closely related to the dopamine binding sites which are not closely coupled with 3'-5'cyclic adenine monophosphate (cAMP). Hosli et al. (1984) showed that astrocytes cultures from rat brainstem and spinal cord had histamine type 1 (HI) and histamine type 2 (H2) receptors. The HI agonist thiazolethylamine produced mainly depolarizations while impromidine, a H2 agonist, caused hyperpolarizations. There are many other instances of glia cells interactions with drugs which indicate that glial cells may have receptors for the drugs and show receptor mediated responses similar to those seen in neurons. A recent article by Hertz and Richardson (1984) reviewed the data on this topic. C-6 cells, a glioma cell line, increase their levels of cAMP in response to NA or isoproterenol (Gilman and Nirenberg, 1971). A similar increase was noted in the human glioma line number 1181 (Clark and Perkins, 1971). Both these cell types must then possess all the known components of the cAMP regulating system (Perkins et al., 1971) and have receptors - 18 -for adrenergic drugs. However, adrenergic receptors in astrocytes and neurons may show different pharmacological profiles (Bender and Hertz, 1984). Chronic exposure to adrenergic drugs can cause a down regulation of adrenergic receptors on glia (Hertz and Richardson, 1984) as on neurons. For example, chronic exposure of astrocytic cell lines to isoproterenol leads to a decreased accumulation of cAMP and a decreased response to some drugs with ^-agonist properties (Hertz and Richardson, 1983) . Various antidepressant drugs, such as doxepin (Hertz and Richardson, 1983) and imipramine (Whitaker et al., 1983), are bound to or taken up by intact astrocytes but this might be because of the lipophilic nature of these drugs. Antidepressants might also interact with the Ot7- and ^-adrenergic receptor sites on astrocytes. These sites are known to exist and an interaction of antidepressants with .-adrenoreceptors is evident since such drugs inhibit isoproterenol-induced stimulation of cAMP production (Hertz and Richardson, 1983) . The binding of /3-adrenergic ligands to C-6 and astrocytoma cell lines is also inhibited by all groups of antidepressants but not by anxiolytic or antipsychotic drugs (Hertz et al., 1982b). Henn in 1980 demonstrated binding of the benzodiazepine, diazepam, to astrocytes. This selective binding can be better demonstrated with another benzodiazepine, R05-4864, because it dissociates less rapidly from astrocytes than from neuronal binding sites (Shoemaker et al., 1983). - 19 -Hertz and Mruerji (1980) showed a large amount of specific diazepam binding on primary astocytes in culture. Diazepam may be displaced by other benzodiazepines or by high concentrations of barbituates. Thus these drugs may be acting through the same receptors. Barbituates suppress potassium-induced stimulation of the oxygen uptake which occurs in brain slices and in astrocytes but not in neurons. This might be the in vitro manifestation of the barbituate-induced reduction in normal metabolic rate. Barbituates also inhibit GABA uptake into astrocytes and this may be one basis of their anti-convulsant action. It was predicted that non-barbituates which inhibit GABA uptake into glia might be effective anticonvulsants, and this was later found to be true for the drug THPO (Meldrum et al., 1982). Thus the effects of various drugs on astrocytes may be similar to those on neurons, have a different profile, or be selective for astrocytes. In some cases, the drug-glia interaction may be more clinically relevant than the drug-neuron interaction. New techniques enabling advances in understanding glia Our understanding of glia and how they exhibit heterogeneity only came about because of new techniques developed over the past decade or so. A) Tissue Cultures The understanding of glia has progressed in the last decade because of recent advances in techniques for separating pure, homogeneous samples. These are now routinely prepared, - 20 -using gradient centrifugation of tissue culture. The astrocytes prepared by gradient centrifugation from fresh tissue are normal astrocytes but may be contaminated by other cell types and debris and their functional integrity may be impaired. Astrocytes in culture are of two general types: established cell line cultures that are transformed cells which do not represent true glia types, and primary cultures, fresh tissue cultures that are treated by procedures which select for certain cell types. They are usually prepared from immature brain so their differentiation must occur during culture. Primary cultures are quite homogeneous, being less than 5% non-specific, and are believed to be functionally similar to in vivo glia of the type selected, most frequently astrocytes. The knowledge on heterogeneity that these techniques have added are particularly about regional heterogeneity or differences between various cell types. B) Freeze Fracturing Techniques There is now a freeze fracture technique which allows more detailed electron microscope views of cell surfaces than previously available. This new development has led to some evidence of glial heterogeneity. Freeze fracture is a technique in which cells are frozen and mechanically fractured; the fractured surface is replicated with platinium and carbon, which reveals the texture of the fracture lines upon examination by transmission electron microscopy. This examination yields several types of structural information. It can show the existence and - 21 -organization of the filaments in the cytoplasm, confirm and show differences in junction between cells, and reveal the existence of repeating patterns of bumps of unknown function on cell membranes. This technique can show heterogeneity within cellular parts, among cell types, and among cells of the same type from different areas. The astrocytic cell processes can be distinguished from those of other glial cell types on the basis of both 10 nm. cytoplasmic filaments and the characteristic membrane structure (Massa and Mugnaini, 1982). The membranes of oligodendrocytes and astrocytes have differing intra-membranous particles. Waxman and Black (1984) examined nodes of Ranvier in adult rat optic nerve and found most had astrocytic processes surrounding them. The cytoplasm of these astrocytes contains 10 nm filaments. The external faces are characterized by orthogonal arrays of pits with a centre to centre periodicity of 6 nm, which corresponds to particles on their protoplasmic or inner membrane faces. The density of particles is similar to that in periparenchymal astrocytic membranes and less than in pericapillary astrocytic and subpial astroglial membranes. Waxman and Black (1984) showed that the orthogonal arrays and gap junctions pattern can be used to identify these astrocytic processes. Anders and Brightman, (1979) showed that these orthogonal arrays of particles increase in number from embryonic day 20 on in rats. They also showed reactive astrocytes not only had an increased number of particles but that they were also - 22 -rearranged to a more highly ordered structure compared to that seen in normal astrocytes. The number of orthogonal arrays of particles on astrocytic membranes increase where they are in contact with non-neuronal tissue (Wujek and Reier, 1981, Anders and Brightman, 1979, Landis and Reese, 1981). But Landis and Reese (1981) did not find orthogonal assemblies in the C-6 glioma cell line. It is not clear whether this means that not all glia have them or that a glioma line, modified under the C-6 culture conditions, will not have them. Gotow (1984) found that filipin, a chemical that produces a characteristic disruption of membranes by acting on the cholesterol in the membranes, had less effect on orthogonal array-crowded astrocytic membranes contacting the basal lamina than on other membrane areas. This means either that these membrane areas contain less cholesterol or that the cholesterol is somehow protected from the filipin. Such areas also contain less alkaline phosphatase and Na+,K+-ATPase, which are both associated with the membrane transport normally found in perivascular processes. This suggests a regional specialization of the astrocytes in vivo. Gotow suggested that the orthogonal arrays may be structural in function, as they occur specifically where active transport is less, or that they are involved in forming a barrier to cholesterol and proteins. He found orthogonal arrays only in astrocytes and ependymal cells. In cultures they appear on all the surfaces and therefore may develop on surfaces exposed to large - 23 -extracellular spaces. The freeze fracture studies can also be used to define various types of junctions. Gap junctions occur between astrocytes and between oligodendrocytes and astrocytes but not between oligodendrocytes, but adjacent oligodendrocytes do form tight junctions (Massa and Mugnami, 1982). Saint Marie and Carlson (1983) used freeze fracture techniques to describe glia heterogeneity in the retina of the compound eye of the house fly. Cells in each layer of the retina had a characteristic pattern of three types of junctions (gap junctions, tight junctions, and septate junctions) and desmosomes which may be equivalent to the orthogonal arrays. Glial cells of each of the layers had characteristic patterns and densities of these features as well as of shapes and physical relationships to the other cells of the layer. The various types of contacts may have different functions: gap junctions - intracellular communication; tight junctions - occlusion of extracellular material; septate - firm but flexible adhesion or tissue impedance; desmosomes - intercellular adhesion. The differing patterns thus imply that the cells have differing functions. This type of work may well be extended in the future to glia of the CNS. C) Markers A variety of markers have been found that allow distinctions to be made between and within the various classifications of glia cells. These markers have provided a - 24 -wealth of information on glial heterogeneity. Some markers can be used to identify as glia, cell types that were not previously so classified, other markers can be used to identify subsets and to provide clues as to heterogeneity of glial function. There have been several good reviews of glial markers (Roots, 1981, Schachner, 1982) but they generally do not emphasize the glial heterogeneity revealed but rather the use of some markers for defining purity of cultures or similar purposes. 1) Fibrous proteins of astrocytes a) Glial fibrillary acidic protein Astrocytes are most reliably identified by the presence of glial filaments under the electron microscope. The Cajal gold method specifically stains these fibers (Cajal, 1913). They are composed of the most studied of glial specific markers, glial fibrillary acidic protein (GFAP). GFAP was originally isolated from multiple sclerosis plaques (Uyeda et al., 1972) and CNS gliotic areas (Bignami et al., 1972), and can be readily stained by immunohistochemistry. It is the principle constituent of the filaments that develop in astrocytes that may have some function in maintaining the shape of astrocytes. Duffy et al. (1982) looked at GFAP in human astrocytoma cells in culture and found a relationship between the shape of the astrocytes and the location of the GFAP. Spindle shaped cells had abundant GFAP in body and processes, whereas in round or polyhedral astrocytomas the GFAP was largely perinuclear. As processes developed, GFAP extended in dense parallel arrays. - 25 -There may also be a relationship between motility and GFAP. Stellate cells in culture, with extensive parallel arrays of GFAP fibers, were more rigid while spindle cells, without these parallel arrays were constantly extending and retracting processes. Salm et al. (1982) showed that rat pituicytes, which are very GFAP positive, do not have glial filaments, as do other astrocytes. This means that GFAP does not have to be organized into filaments to give positive GFAP staining in cells. Suess and Pliska (1981) found that the pituicytes remained strongly GFAP positive even after transplant of the pituitary to a region under the kidney capsule where there are no neural influences. GFAP is found mainly in mature astrocytes. It is denser in the fibrous subtype but is also in protoplasmic astrocytes. It is also found in several cell types related to astrocytes, including radial glial cells (Bignami and Dahl, 1974), the Bergmann glia, enteric glia and a small percentage of glia in the peripheral nervous system (Jessen et al., 1984). Kennedy (1982) noted that the question of whether all astrocytes contain GFAP has yet to be answered. A permanent increase in GFAP content rapidly follows injury and precedes astrogliosis (Bignami and Dahl, 1975). In cultured astrocytes GFAP staining can be increased by glia maturation factor (Lim et al., 1977), brain extracts or dBcAMP which causes maturation. In some cell types it may only be expressed transiently; in humans it is only expressed in ependymal cells between week 13 and full term, and is also - 26 -transiently in tanycytes (Roots, 1981). Heterogeneity can be found in subsets of cells. Dahl et al. (1982) noticed that antiGFAP stained only a subset of Schwann cells in rat sciatic nerve. These were mainly surrounding non-myelinated axons, and increased in number during Wallerian degeneration. There may also be heterogeneity of the GFAP within the CNS. GFAP has been shown to be composed of various chemically different proteins although the subtypes have the same molecular weight in peripheral and central glia. The problems with GFAP staining are mainly those of producing the antibody itself and the cross-reactivity that impurities cause. b) Vimentin There are two other proteins, related to but different from GFAP, which are found in some glia and are constituents of fiber systems within the cell. One of these is vimentin. Antibodies to vimentin and GFAP were used in a double labeling experiment to examine astrocytic filaments in development and wounding (Pixley and DeVellis, 1984). Filaments stained for vimentin only in newborn rats and for GFAP in 2 0 day and older rats, with a gradual switch in between. Stab wounds were made to cortical areas at a time when there were normally no vimentin-positive cells in the region. Vimentin only occurred at the edge of the wound. This led to the hypothesis that vimentin occurs when there is contact with wide open spaces and is lost when such contact with these spaces disappear. All vimentin-positive cells seem - 27 -to have at least one portion of the cell in contact with CSF, eg. ependymal cells, tanycytes, Bergmann glial fibers, Muller cells, and radial glial. The disappearance of vimentin positive cells correlates with the loss in extracellular volume. Cells in culture, where there is much extracellular fluid, develop vimentin regardless of origin. This hypothesis would explain the appearance of vimentin only in cells at the edge of the wound, close to the fluid tissue boundary. In the adult rat, vimentin was only observed in fibroblasts, cells of relatively large blood vessels, ependymal cells, and astrocytes. At embryonic rat day 11, the vimentin was observed only in radial fibers, ventricular cells, and blood vessels. In culture Schnitzer et al. (1981) found GFAP and vimentin occurring together in the cells which have fibroblast morphology. c) Desmin Dahl and Bignami (1982) showed that desmin forms a third fiber system in glia cells. It was uniformly distributed in astrocytes of brain and spinal cord and in Muller cells. A comparison with GFAP showed that both were similarly localized in brain and spinal cord but not in the fibers of Muller cells. 2) Glutamine sythetase GS is a major astrocyte marker which catalyses the reaction: glutamate + ammonia + ATP > Glutamine + ADP + Pi It may thus be involved in glutamate metabolism and ammonia - 28 -detoxification. Martinez-Hernandez et al. (1977) used immunohistochemical techniques to show that GS is located exclusively over glial cells in the brain and Muller cells of the retina. Sarthy and Lam (1978) confirmed its location in Muller cells. Electron microscope studies of rat brain have revealed a localization in both protoplasmic and fibrous astrocytes (Norenberg and Marinez-Hernandez, 1979, Kennedy, 1982). Norenberg (1983) also described GS staining in ependymal cells, Bergmann glia, perikaryon, vascular end feet in the glia limitans and astrocytic processes which ended just beneath the ependymal surface. This distribution supports a function for glia of providing a barrier against ammonia. Schousboe et al. (1977b) found high activity of this enzyme in primary astrocyte cultures. But Nicklas and Browning (1978) found that the C-6 astrocytoma cell line had very low GS activity. Its localization to glia implicates glia as the location of the small glutamate compartment (Norenberg and Marinex-Hernandez, 1979), as discussed previously. Since primitive glia do not contain GFAP, but do contain GS, the latter may be a better marker of astrocytes. Gliotic scar tissue, however, does not contain GS; this may lead to buildup of ammonia and glutamate which could explain the epileptogenic nature of such tissue (Norenberg, 1983). There are significant regional variations in the intensity of GS staining in astrocytes. The molecular layers of the cerebellum and hippocumpus are particularly heavily stained (Norenberg, 1979). - 29 -3) Carbonic anhydrase Carbonic anhydrase combines reversibly CO2 and H2O to form carbonic acid, hydrates aldehyde groups to alcohols, and acts as an esterase. It thus may be involved in the regulation of pH, secretory activities and movement of ions. It develops in the brain at the same time as glial proliferation. It used to be considered an astrocyte specific marker but is now known to be also on Muller cells (Sarthy and Lam, 1978) and oligodendrocytes as well (Ghandour et al., 1979, 1980, Rousell et al., 1979, Mandel et al., 1978). Primary cultures of astrocytes have carbonic anhydrase (Kimelberg et al., 1978b), but C-6 cells do not (DeVellis and Brooker, 1973). Kimelberg et al. (1982) found the histochemical method for carbonic anhydrase stained astrocytes in the monolayer of primary cultures from rat cerebral hemispheres, as well as cells believed be be oligodendrocytes above this layer. Carbonic anhydrase exists in several isoenzymes that have different amino acid sequences and are immunologically distinct. Antibodies to two of these isoenzymes (Ca-1 and CA-2) may stain distinct subsets of glia. It was thought that CA-1 might be astrocyte specific and CA-2 oligodendrocyte specific. However, in culture cDbAMP caused astrocytes to differentiate and stain for CA-2 as intensely as oligodendrocytes. - 30 -4) Other markers There are a number of markers that have been reported to stain astroglia predominantly. Many of these are isolated reports with very little confirmation. Table I summarizes many of these recent or poorly substantiated findings. There are a number of markers that used to be considered astrocyte specific and now are known to be on other cell types as well or whose specificity is now highly controversial. - 31 -MARKER TABLE I: MINOR ASTROCYTE CELL MARKERS TYPE OF CELL AUTHORS Non-neuronal Astrocytes cerebellum including Langley & Ghandour enolase * Bergmann glia & cytoplasmic processes (1982)  OL-2 -Glycoprotein Astrocytes; astrocytomas not glioblastomas Langley et al. (1982) Tamm-horsefall glycoprotein Ependymal cells & astrocyte processes Zalc et of Bergmann fibers or astrocytic feet in contact with blood vessels or meninges al. (1984) Sulfogalactosyl ceramide (SGC) ** 11 Zalc et al. (1984) Ml - antigen Distinguishes sub-cerebellar astrocytes Lagenaur some but not all GFAP+ cells • et al. (1980) C-l - antigen Only processes of Bergmann glia & Muller Sommer et al. (1981) cells, and ependymal cells, but not other astrocytes except in early postnatal astrocytes of white matter Purkinje cell layer and radially oriented structures of telencephalon, pons, pituitary and retina  IgG - RAN 2 # Astrocyte precursor cells, ependymal Bartlett et al. (1981) cells, Muller cells . leptomeningeal cells  * Catalyzes oxidation of phosphoglyceric acid to phosphopyruvate ** Cl" possibly involved in Cl transport # IgC made by antibody secreting hybridomas and defined by antibody TABLE I (continued): MINOR ASTROCYTE CELL MARKERS MARKER TYPE OF CELL AUTHORS THY 1 Only subtypes of astrocytes that are also glactocerebroside+ Schnitzer and Schachner (1981) THY 1 Two cell lines Kemshead et al. (1982) Antigen A2B5 Immature astrocytes; oligodendrocytes neurons; has considerable regional, developmental and species variation Schnitzer & Schachner (1982) S 100 Protein * (controversial) specific astrocyte marker OR in oligodendrocytes, endothelial cells and neurons also Gandour et al. (1981a) Hyden et al. (1980) Hansson et al. (1976) SSEA-1 Glycolipid ** antigen Subtypes of astrocytes. In early mouse cerebellum only in external granular layer and molecular layer; later only small areas in molecular area Lagenaur et al. (1982a) Lagenaur et al. (1982b) N11N1 Monoclonal # antibody Human fetal brain cultures and primate spinal cord; subtypes 80-90% GFAP+ in some cells only Dickson et al. (1983) 308 Monoclonal # antibody Differences found between GFAP+ and GFAP- astrocytes Dickson et al. (1983) * May be involved in binding Ca++ & movement of monovalent cations or GABA transport ** Originally found on surface of F9 tetracarcinoma cell # Originally raised to human neuroblastoma As can be seen from Table I, most of these minor markers for astrocytes are only specific for subsets of astrocytes. Thus, with the development of marker technology came considerable evidence for glial heterogeneity. More research on each of these markers may be the basis for future classification systems for astrocytes. Oligodendrocytes and CNS myelin also have their markers that are more or less specific. Table II summarizes data available on markers that could or have been considered oligodendrocyte markers. Future research will use these and other oligodendrocyte markers yet to be found to start to classify oligodendrocytes into subtypes based on heterogeneity seen between areas, species, and physical cell types. Since some markers are not only tools but are also direct indicators of glial structure and biochemistry, the wide variety of markers stainable in various subsets of glia is indicative of great functional heterogeity of glial cells. With advances in techniques for staining, using markers, freeze fracture, and electron microscopy, there came a new understanding of the morphological diversity of glia and how this diversity can sometimes be correlated to the biochemical diversity that can be found. - 34 -TABLE II: MARKERS FOR OLIGODENDROCYTES AND MYELIN MARKER LOCATION AUTHORS Galactocerebroside Myelin specific PNS + CNS oligodendrocytes epithelial cell of ventricles and choroid plexus Raff et al. (1978) Myelin basic protein * Early oligodendrocytes and myelin sheaths Sternberger et al. (1978) Hartman et al. (1979) Roussel & Nussbaum(1981) Myelin basic protein Cultures of galactocerebroside + cells Bhat et al. (1981) Anti-proteolipid antiserum Myelin sheaths and actively myelinating oligodendrocytes Agrawal & Hartman (1979) Wolfgram Protein WI & W2 Central myelin & oligodendrocytes Labourdette et al. (1979) 2',3'-Cyclic nucleotide 3'-phosphohydrolase ** Inner & outer most sheath myelin Species specific, not in fish; Its function may only be coincidentally related to myelin Not in old oligodendrocyte cultures Szuchet and Stefansson (1980) 01—>04 # Oligodendrocytes of early postnatal cerebellum, cerebrum, spinal cord, optic nerve & retina: 01 & 02, and 03 & 04 occur in different developmental times in different areas Sommer & Schachner (1981) Schachner (1982) * Antigen defined ** Catalyses hydrolysis of 2',3'-cyclic nucleosides to the 2•-nucleotides; major component Wolfgram protein # Four monoclonal antibodies from mouse myeloma immunized with white matter from corpus callosum TABLE II (continued): MARKERS FOR OLIGODENDROCYTES AND MYELIN MARKER LOCATION AUTHORS MAG Oligodendrocytes, Schwann cells and certain areas of periaxonal region of the central and peripheral myelin sheath Sternberger et al. (1979) Succinic dehydrogenase Oligodendrocytes and astrocytes Mossakowski(1962) Butyryl Cholinesterase * More active in oligodendrocytes than in other glia (1954) Cavangh & Thompson , Oehmichen (1980) Antimyelin antiserum Myelinated fibers, medium and light oligodendrocytes; not dark (mature) astrocytes, Golgi epithelium cells, Bergmann fibers and some subependymal cells Roussel & Nussbaum (1983) P-2 Myelin specific protein Rabbit CNS myelin: more in caudal areas Highest in spinal cord, lowest in frontal cortex; only in larger diameter axons Trapp et al. (1983) Glycerol-3-phosphate 01igodendrocytes DeVellis et al. (1978) TU-01 ** Cerebellar glia cells; only in microtubules; smooth endoplasmic reticulum outer mitochondrial membrane; ribosomal rosettes Hajos & Rostomian (1984) * Non-specific cholinesterase ** Anti-tubulin antibody Glial Heterogeneity - Morphological Given the development of new tools and recent improvements of old tools the heterogeneity, as we now know it, can be discussed. The most obvious first way to consider heterogeneity would be to discuss the visual differences between cells. A brief discription of glial cell types has already been given in the introduction but will be developed further here as there appears to be a variation in morphology within the classical glia types. Cajal originally divided astrocytes into fibrous and protoplasmic types, based on location, morphology and function. Several other distinct astroglial cell types have been discribed but it remains to be seen whether they are really distinctive cell types, transitional forms reflecting developmental stages, adaptions to local physical or chemical environments, or a reversible expression of a natural astrocytic plasticity. The morphological types which have been described have not been catagorized into usable subgroups. Most are still catagorized as mature astrocytes by the appearance of cytoplasmic filaments in the electron microscope, as originally described by Palay et al. (1962). The distinction between astrocytes and oligodendrocytes may sometimes be difficult. Both can be stained by the silver carbonate method so that the chemistry of their cytoplasms must be similar. Reyners et al. (1982) described a highly radiosensitive cell that has ultrastructural characteristics intermediate between the normal protoplasmic astrocyte and the - 37 -light oligodendrocyte. It is present in significant numbers and may be a multipotential reserve glial cell, capable of replacing oligodendrocytes or microglial cells. This beta-astrocyte, as it is called, can be distinguished from alpha-astrocytes, as normal astrocytes are called, by its irregularly shaped nucleus, denser ribosomal cover of the endoplasmic reticulum, coarser lysosomal morphology, lack of cellular processes, and total absence of gliofilaments. The ^-astrocytes are never found near the outer membranes of the blood vessels but, like microglia, are frequently found in perivascular areas, and are frequently satellites to nerve cells. Koenig and Barron (1963) also noted that there is a continuum of transitional forms between oligodendrocytes and reactive astrocytes. Oligodendrocyte morphology has long been sub-divided into sub-types. Rio Hortega (1928) was the first to describe differences in oligodendrocytes. He classified oligodendrocytes into perineuronal and interfasicular, the latter cells being frequently aligned in rows with soma close to one another. Both of these subclasses were heterogeneous in terms of size and shape of the cells bodies, in the number and features of their processes and in the number and size of the axons with which the cells were associated. He thus subclassified them into 4 prototypes. Most perineuronal and perivasular oligodendrocytes are type 1: these are small cells 15-20 fjim in diameter with many fine processes that show great variability in the manner in which they emanate from the cell but exhibit little branching. Each process abuts a nerve - 38 -fiber. They are found in the cerebrum, cerebellum, and medulla. Most interfasicular oligodendrocytes are type 2: they have fewer and thicker processes than type 1, range from 20 to 40^ra in diameter, differ in the manner in which the processes come out from the cell, and are only in white matter. Type 3 oligodendrocytes are distinguished by their large size and few but large processes. They are most frequently found as mono- or bi-polar cells associated with large axons. They are less numerous than types 1 or 2. Type 4 are either mono- or bi-polar and have highly elongated bodies and are attached directly to axons. These sub-types show similar deposition of silver carbonate, being dense throughout the cell body. These subtypes therefore have similar biochemistry for this stain. Another classification scheme for oligodendrocytes is based on the wide range of nuclear and cytoplasmic densities (Caley and Maxwell, 1968, Mori and Leblond 1970) . They used these density differences to classify them into light, medium and dark oligodendrocytes. It is not known how much of an overlap there is between the two classification systems. It is highly probable that the increasing density is a function of developmental maturity. Many recent data suggest that a wide variety of cells not traditionally thought of as being glia share characteristics of the classical glia and usually resemble astrocytes. Ependymal cells contain vimentin and resemble astrocytes in morphology. Pituicytes of the neurohypophysis are GFAP-positive and - 39 -retain this characteristic even when no longer under neural influence. Muller cells of the retina have long been recognized as being glia, though they are not the only glia in the retina. They are GFAP-positive, and vimentin-positive, and contain both GS and carbonic anhydrase (Linser and Muscona, 1981). The two enzymes differ markedly in developmental profile (Linser and Muscona, 1981) . In early embryonic chicks, carbonic anhydrase is in all retina cells, but during development the specificity of Muller cells increases with the final disappearance from amacrine cells only on the 16th day. Bussow (1980) showed that Muller cells of the retina are not like glia of other areas as they seem to have a specialized function. They are throughout the entire thickness of the retina and their basal processes align with the nerve fiber to form septi that fasiculate the axons of the ganglionic cells. In the inner and outer plexiform layer of the retina of Macaque monkeys there are, based on morphology, at least two other glia cell types besides Muller cells (Boycott and Hopkins, 1981). Bussow (1980) found astrocytes that are not homogeneously distributed throughout the layers of the retina, but have processes only in the ganglion and nerve fiber layers that run parallel to the axons. He also saw in the optic nerve fibrous astrocytes with processes that run perpendicular to the axon bundles, these might be considered specialized glia for the ganglionic cell axons. Barber and Lindsay (1982) found that the Schwann cells of the olfactory and vomeronasal nerves are more closely related - 40 -to central astrocytes and glial cells of the myenteric plexus than to Schwann cells of other parts of the periphery because they react to antibodies to both GFAP and a 49k dalton glial filament protein from human brain. These glia, which grow from the periphery into the CNS along the nerve as it develops into brain, are traditionally called Schwann cells because of their peripheral origin. They are, however, also morphologically different from true peripheral Schwann cells in that they do not ensheath axons individually but extend tongues of cytoplasm which branch between the axons and separate them into bundles. They also have no basement membrane surrounding individual cells, and contain no collagen. Tanycytes are specialized glia-like cells with radially oriented processes that line parts of the ventricles, especially the third ventricle. They are GFAP-positive at an earlier stage than are glia and continue so throughout development (Basco et al., 1981, DeVietry et al., 1981). They remain GFAP+ into adulthood, in which regard they are like Bergmann glia of the cerebellum, some similar cells of the cortex, hippocampal glia of the dentate gyrus, and normal astrocytes. They, like other glia in contact with cerbrospinal fluid, are vimentin-positive. Bergmann glia, or Golgi epithelial cells as they are sometimes called, are specialized glia of the cerebellum with cell bodies around the Purkinje cells and radially oriented processes extending thoughout the molecular layer. They stain for most astrocytic markers. They evolve from fibers in the - 41 -molecular layer of the early cerebellum (Fulop et al., 1979) that transform into Bergmann glial cells after guiding granule cells to their final position (Rakic, 1971). Contestabile and Andersen (1978) studied Bergmann glial cells and found a different profile of enzyme activities than in regular astrocytes. They found high activities of nonspecific L-glutamate dehydrogenase, glucose-6-phosphate dehydrogenase and TPNH-tetrazolium reductase, but low activity of lactate dehydrogenase, and no succinate dehydrogenase. Therefore, the Bergmann glia are low in citric acid cycle enzymes and high in those of the pentose phosphate shunt. Bergmann glia are not the only glia in the cerebellum. Hatten et al. (1984) studied glial cells of different types in early postnatal mouse cerebellum. There were two types that were GFAP positive: one type had two to three neurons associated with it and resembled Bergmann glia, and the other was larger, had shorter arms in which it clustered a dozen or more neurons and resembled astrocytes of the granular layer. There were also galactocerebroside-positive glia which did not associate with cerebellar neurons during the time studied. Time lapse photography revealed extensive migration along the arms of the Bergmann-like astrocytes but not the stellate astrocytes. Radial glial cells have been described by a number of researchers. They transform into distinguishable types of glia, are related to astrocytes, Bergmann glia, Muller cells and tanycytes which all contain GFAP at some point in their development and have similar morphology. The immature radial - 42 -glia are classified as astrocytic because of the bundles of filaments and accumulation of glycogen in their cytoplasm (Rakic, 1972), but they may also develop into oligodendrocytes. Akers (1977) and Cajal (1929) both describe the developmental changes that occur in radial glia in the cortex. Transitional forms have been noted between radial glia and astrocytes (Schmechel and Rakic, 1979) and between radial glia and oligodendrocytes (Choi et al., 1983). Radial glia have been shown to transform into mature astrocytes (Rakic, 1972, Cajal, 1929, Schmechel and Rakic, 1979), and most astrocytes pass through a radial glial phase. They may also develop from ependymal and subependymal layers by way of glioblasts and astroblasts, as will be discussed in the next section. Using GFA to follow the developing brain in monkey, Levitt and Rakic (1980) showed the evolution of radial glial cells as they fanned out from reticular and subreticular zones to the pial surface where they had end feet that stayed until the transition to astrocytes. Radial glia are traditionally supposed to be the guidelines that neurons use to grow along during development. However, in the developing mouse cortex and hippocampus, Woodhams et al. (1981) noticed that cortical plate formation and time of the first appearance of GFAP+ radial glia did not correlate very well. A clear association is evident in the late stages but not in the early stages. This argues against a primary role in cortical plate formation. Choi and Lapham (1980) found two types of radial glial - 43 -cells in the developing cerebellum of the human fetus at 9 weeks, which is earlier than previously thought. The lower ones extended from the ventricular area to the vascular walls of the intermediate layer, the upper ones from below the Purkinje cell layer to the pia. The latter, at 20 weeks, closely resembled Bergmann glia. Seress (1980) examined the radial glial cells of postnatal rat brain. Until day 10 radial glial cells were seen in the walls of the third and forth ventricles, and had very variable morphology in different regions. After day 10, only tanycytes were seen in the wall of third venticle, showing the postnatal development of tanycytes. ¥ Developmental Differences - A Source Of Heterogeneity Glia cells can have very different developmental histories. Understanding the development is still not complete and has undergone many changes. In 1888 Golgi, using the Golgi staining technique, proposed that the precursors of all non-neuronal cells, the spongioblasts, arose from columnar epithelial cells in the walls of the neural tube. His (1889) first postulated the theory of 2 types of germinal cells in the neural tube, germinal cells which are neuron producing cells and spongioblasts which are glial producing. Lenhosseck (1891) was the first to prove that glial cells were epithelial rather than mesenchymal in origin and differentiate from the primative ependyma just as neurons do. Schaper (1897) argued that germinal cells produce a bipotential cell that migrated away from the layer to differentiate into glial or neuronal - 44 -cells. It was not until Cajal (1909-1911), in his classical studies of the spinal cord of the chick embryo, found transitional cells from neuroectoderm to mature neuroglia that the origin of the neuroectoderm was firmly established. Cajal noted three cell types: neurons appeared first, astrocytes and an unidentified third type appeared later. This third type was later elucidated by del Rio Hortega (1919) who used the silver carbonate method to identify them as microglia. The problem with this early work is that the staining was not really adequate for developmental work. The gold sublimate method of Cajal selectively stained astrocytes well but the silver stains were not completely selective for oligodendrocytes and neither of these two methods stained undifferentiated cell precursors. Penfield (1932) noted that some of the ependymal cells were derived from neuroepithelial cells which had processes extending to the external limiting membrane and were known as supportive spongioblasts. These cells lost their processes and became subpial astrocytes which then formed attachments to blood vessels. This was thus an alternate route than from bipotential cells for glial development. Until the development of the electron microscope, the ependymal cell was thought to be the precursor of the macroglia. The current theory is that glia develop from neuroectoderm after neuroblast formation declines. The glioblasts give rise to both astrocytes and oligodendrocytes. Using thymidine injected after birth in rats from day 1-21, and sacrificed at day 25, glial genesis in the somatosensory - 45 -cortex was observed (Ichikawa and Hirate, 1982). It occured from an inside to outside manner in the first two weeks. The glioblast development, however, is still in doubt. His's original theory of two precursor cells was challenged after 90 years by Fujita (1980), who showed that these two cell types are nothing but the same cell in different phases of the mitotic cycle. In the early stage of development the neural tube is composed only of matrix cells (stage I). Out of these develop post-mitotic cells that are the neuroblasts (stage II). When neuroblast production comes to an end, stage III begins which is the production of glioblasts and ependymal cells, which migrate and mature into oligodendrocytes or astrocytes and resting microglia. Fujuita et al. (1981) concluded that resting microglia evolved from neuroectoderm and can give rise to fibrous astrocytes upon injury. Sturrock (1976) observed four different types of immature glia in the corpus callosum of mice. They are the early glioblast, small glioblast, large glioblast, and young astrocyte. He proposed the following sequence of development: early glioblast >large glioblast > light oligodendrocyte >medium oligodendrocyte >dark oligodendrocyte early glioblast >small gliobast >young astrocyte >mature astrocyte Skoff (1980) disputed the concept that gliogenesis occurs only after neurogenesis. He cites the observation that radial glia exist to guide the neurons to their place, though this is now in question in the earlier stages (Woodhams et al., 1981). - 46 -He proposed that astrocytes and oligodendrocytes originate from astroblasts and oligodendroblasts, not glioblasts. Astrocytes can divide during development (Hajos et al., 1981) and, once formed, can divide but oligodendrocytes do not. Keenikova et al. (1979) stained both basic and acidic proteins during development and noticed successive changes in the ratio of basic to acidic, suggesting successive populations of glia types. Polyclonal antibodies to vimentin have shown the existance of subpopulations of astrocytes during devleopment (Dahl et al., 1981, Shaw et al., 1981, Yen and Field, 1981). Other developmental profiles seem to exist in some systems. Raff et al. (1984) describe three types of glial cells in the optic nerve that appear at different times. Type 1 astrocytes appear at embryonic day 16, oligodendrocytes at postnatal day 2 and type 2 astrocytes on postnatal day 10. In culture it was shown that the oligodendrocytes and type 2 astrocytes came from a different cell type than type 1. Others showed that the optic nerve astroblasts can be traced back to ventricular cells without going through glioblast stage. This is evidence for alternate gliogenesis in brain. The explanation for these grossly different theories of gliogenesis may be that different researchers observe different populations of glia which are heterogeneous in their development. These different developmental profiles may be a major source of confusion in interpreting the literature on glia heterogeneity. Not only are there different patterns of - 47 -gliogenesis but there are different patterns of development of biochemical abilities within these cells. Some biochemical phenomena develop early, such as vimentin, whereas some develop late. For example potassium is noted to have a stimulatory effect on the Na+,K+-ATPase only in very late ontogenic stages (Grisar, 1979) or in older astrocyte cultures (Moonen and Franck, 1977) Another example of differing biochemical maturation is the incorporation of radioactive glucose into aspartate, glutamate, and glutamine which develops late, as it is much less pronounced in brains from immature animals than in adult brain (Van den Berg, 1970). The development of metabolic compartmentation coincides well with the time period when conversion of glucose carbon into these amino acids intensifies (Patel and Balazs, 1974). The increased intensity of glutamate uptake into cultured astrocytes occurs at the same age (Schousboe et al, 1976, Hertz et al., 1979). Also GS levels rise in vivo at the same age and the same occurs in cultured astrocytes (Hertz et al., 1978). Therefore some glial differences may evolve along with the relative late development of metabolic compartmentalization. Observations like this means that caution must be used when interpreting the biochemical differences reported for glia as some hetergeneity may be due to the development stage of the cells used in the research. Glycogen storage changes within radial glial cells of developing rat brain is another example of developmental changes that occur. Such storage first appears on embryonic - 48 -day 14 in the choroid plexus and in the radial glial cells of midbrain and medullary raphe (Bruckner and Biesold, 1981). These cells retained the highest capacity throughout development but other radial glia also showed some glycogen storage as they developed. Glycogen storage then decreased to adult levels by postnatal day 21. This might indicate that glycogen is used as an energy source in perinatal metabolism. In other work, Colmant (1965) noticed increases in acid phosphatases, DPNH- and TPNH- tetrazolium reductases, succinate dehydrogenase, 5'nucleotidase, phosphamidase, and ^-naphthol esterase in oligodendrocytes during postnatal development. Lagenaur et al. (1980) used the antibody they designated Ml to distinguish subgroups of astrocytes in mouse cerebellum. Staining appeared in white matter at day 7 and lasted until adulthood but in Bergmann glia and in the granular layer on day 10 and lasted only a short time. In culture it is in some but not all GFAP+ cells. Neurons and glia are committed to cell lines prior to cessation of division but in most of these schemes the final differentiation occurs throughout the parenchyma, thus close to the cells and blood vessels they might eventually interact with, allowing for the local biochemical climate to induce variability. - 49 -Heterogeneity in tissue cultures A) Developmental changes in culture Primary cultures of glia provide some pertinant information on developmental questions and heterogeneity. Primary cultures are thought to mimic closely the in vivo situation and develop or redevelop many in vivo characteristics. Massa et al. (1983), for instance, showed that oligodendrocytes designated B3,f even redevelop membrane specializations such as tight junctions. Cultures allow experimental manipulations and developmental observations to be carried out. Several researchers have monitored primary glial cultures of various ages for changes with time in culture. These are believed to mimic developmental changes in vivo. Fedoroff et al. (1984a) examined newborn mouse astrocytes in cultures, originally plated at low density, longitudinally from three days to four weeks. The earliest astrocyte precursor cells or glioblasts are closely apposed epithelial cells that rarely have junctions. Their scanty cytoplasm contains many free ribosomes but few microfilaments. The cells of the next stage of astrocyte lineage, proastroblasts, are flat and separate from each other to a variable degree. They have intercellular junctions associated with microfilaments and contain singly dispersed intermediate filaments. The proastroblasts gradually differentiate into astroblasts which have a similar morphology except that they also contain bundles of intermediate fibers. When neuroblast production comes to an end, the third stage begins which is - 50 -the production of astroblasts and ependymal cells. These migrate and mature into olgodendrocytes or astrocytes and resting microglia. The mature fibrous astrocytes have well defined processes and distinct perikarya. This study showed that the route from radial glia to fibrous astrocytes is not the only route. It supports the general observation of the ventricular or subventricular origin of astrocytes. These observations also illustrates how lack of definition as to culture conditions or cellular stage could lead to observations of heterogeneity which would in reality be different stages. Marker changes also occur and compound the difficulties in research. Schousboe et al. (1980) showed that GFAP increased with time in astrocyte cultures to exceed the level in 4 week old whole forebrain. Labourdette and Marks (1975) showed that S100 is synthesized mainly after differentiation in the C-6 line. Changes in enzymes also occur during development. Schousboe et al. (1980) monitored various enzymes in primary cultures of astrocytes from the cortex of mice or rats. Na+,KH—ATPase reached its peak at 2-3 weeks in culture but the stimulatory effects of K+ did not occur until 4 weeks. This parallels the in vivo changes. Lactate dehydrogenase peaks at two weeks in culture to a level above that of adult brain, then drops to the adult level. The iso-enzyme pattern of lactate dehydrogenase changes from immature to mature from one to three weeks. GABA-T in astrocyte cultures drops in the first week but then increases back to levels comparable to - 51 -those in neonatal mouse brain. Carbonic anhydrease was low but found to increase toward in vivo levels in differentiated cultures. COMT and MAO increased with time in primary astrocyte cultures (Hansson and Sellstrom, 1983). Levi and Ciotti (1983) showed GABA but not D-aspartate uptake was restricted to mature stellate astrocytes in culture. Therefore GABA transport is a differentiated phenomenon or is due to a subset in culture that become prominent. Meller and Waelsch (1984) studied cultures of cells from embryonic brain for a year. Anti-GFAP and myelin basic protein were used to identify glial cell types. Four cell types were observed and monitored: flat epitheloid cells that were GFAP- and either myelin basic protein+ or -, astroglial cells which were 92% GFAP+, and oligodendrocytes which were myelin basic protein+. The astrocytes originate continuously where as the oligodendrocytes differentiate every 2 0 to 3 0 days. Vernadakis and Mangoura (1983) compared cultures from newborn and adult mice and found that those from newborn mice had both oligodendrocytes and astrocytes, as determined by markers, and these both increased in culture. On the other hand, in cultures from adult mice only the astrocytes increased and these for only a few days, whereas the oligodendrocytes decreased in number. This parallels the astrogliosis seen in aging brain. Other people have shown a variety of cell types in primary astrocyte cultures. Wilkin et al. (1983) examined primary - 52 -cultures made from cerebellar astrocytes which were GFAP positive and were of two distinct morphological types. One class was stellate in shape with radially distributed fine processes while the other was varied in shape, being either polygonal or elongated. They both could incorporate thymidine and therefore were capable of division. Both took up aspartate but only the stellate cells took up GABA. The stellate cells disappeared over the 12 days of culture but lasted longer in lower density cultures, possibly undergoing a change in shape following cell to cell interactions. Non-stellate cells that did show a weak GABA uptake ability lost this at later stages. c-AMP, which increases stellate morphology, did not increase GABA uptake, suggesting c-AMP is not a true agent of differentiation and that morphology is not a true indicator of biochemical function. The fact that non-stellate cells continued to divide but stellate did not may indicate that these are two different types of astrocytes, but other factors such as state of committal to differentiation at time of plating, or presence of particular types of neurons may be factors. B) Effect of culture conditions on cell development Some of the variability in cell type may be due to slight differences in culture mediums. Morrison and DeVellis (1983) attempted to study this by using a chemically defined medium. They found that a chemically defined medium produced purer and more controlled cultures that were 95% astrocytes (GFAP+) and only 1% + for fibronectin, a marker for meningeal or - 53 -endothelial cells. The cells were morphologically differentiated and positive for both S-100P and GS, indicating at least some biochemical differentiation. Not all astrocytes respond to differentiating factors in the same way. Raff et al. (1983) describe two types of astrocytes in culture of white matter, both being GFAP+. Type 1 are fibroblast-like, did not bind tetanus toxin or the monoclonal antibody A2B5, were stimulated to divide by bovine pituitary extract or epidermal growth factor and are also found in grey matter. Type 2 had a neuron-like morphology, bound tetanus toxin and A2B5, and were not stimulated by bovine pituitary extract or epidermal growth factor. Type 1 could be converted to neuron-like morphology in the presence of dBcAMP, pituitary extract or brain extracts, especially in serum-free medium, but did not gain the specific binding properties of type 2. In neonatal cultures most of the type 2 cells developed from GFAP- cells which were induced to express GFAP by culture conditions. Culture conditions can thus induce changes in morphology. If changes in culture conditions can induce changes in morphology, understanding the mechanism could give us an understanding of in vivo cell differences. Much work has been done on how culture conditions influence cultures. Trimmer et al. (1984) explored the culture conditions which influence the cellular composition of cerebral cortical cultures. A decrease of plating density, increased age of the animals and supplementation of the cortical cultures with meningeal fibroblasts all caused an increase in fibronectin - 54 -staining, and a decrease in GFAP, an increase in£2 adrenergic receptors and a decrease in Astroglial cells normally express both types of binding sites, with 60% fil and 40% jS2. Goldman and Chiu (1984) showed that astrocytes plated at high density reached higher densities quickly, had small perikarya and several long processes that were GFAP+ and contained less actin and more intermediate filaments, whereas those that were plated at low initial densities did not increase in cell number, were flat and polygonal, stained for GFAP and retained large amounts of cytoskeletal actin relative to intermediate filaments. These results were mirrored in the rates of synthesis of these cytoskeletal proteins. Both of these forms take time to develop from thin spindle-shaped cells with a few narrow processes. Lindsey et al. (1982), using astroglial cells from the corpus callosum, showed that astrocytes change shape as the cells approach confluency. Differences in culture conditions can exist within the same culture. Fedoroff et al. (1983) found a cell in normal cultures without dBcAMP which forms on top of the layer of precursor cells; this cell is smaller than the lower cells but resembles the dBcAMP stimulated large astrocyte. Both contain GFA and vimentin, with vimentin developing first. Such cells seem to develop spontaneously where there are special conditions at the interface between the cell confluent layer and the medium. Thus much of the heterogeneity that exists is really the result of the same cell responding to differing conditions. - 55 -Culture differences can be used to select for subsets of cells. A subset of oligodendrocytes were selected by their inability to plate on plastic culture plates, but only on polylysine coated plates. These were oligodendrocytes as they were 98% galactocerbroside +, are highly differentiated and remain so in culture. This is shown by high levels of CNPase activity, high incorporation of H2SO4 into sulfides, and a lipid metabolism that mimics that associated with myelinogenesis, i.e., the presence of myelin associated glycoproteins and myelin basic protein. If culture conditions can change cells, perhaps we are looking at bipotential or multipotential cells. They have been postulated to exist in vivo and have been demonstrated in cultures. Raff et al. (1984) describe a progenitor cell that differentiates into an oligodendrocyte if cultured in a serum free medium and an astrocyte if cultured with fetal calf serum. Galactocerebrosidase was used as a marker for oligodendrocytes and GFAP as a marker for astrocytes. The cell contains vimentin filaments which it retains if it v becomes an astrocyte and loses if it becomes an oligodendrocyte. The commitment is reversible for 1 to 2 days. Noble and Murray (1984) found the same or very similar cells in optic nerves of neonatal rats. These were stimulated to divide by the presence of purified type 1 astrocytes or soluble factors from such astrocytes, producing a large number of progenitor cells and oligodendrocytes. Noble and Murray - 56 -speculated that this subpopulation may be the source of the optic nerve oligodendrocytes and type 2 astrocytes but not type 1 astrocytes. They had a profile of antigens identical to the cells reported by Raff et al. (1983). Juurlink et al. (1981) found that immature epithelial-like cells that form type A colonies in culture come mainly from the subventricular zone and develop into type C cultures which are morphologically different. Since these respond to dBcAMP in the same ways as astrocytes do, the authors propose that type A cells are astrocyte progenitor cells, probably equivalent to the pale cells from the subventricular zone. As postnatal devleopment occurs, the number of colony-forming cells decrease. Thus there are also interactions between cell type and culture conditions. Another example of interaction was shown by Yu and Hertz (1982) who found that the proportion of MAO type A to type B decreased in mouse brain primary astrocyte cultures after treatment with dBcAMP. At 31 days, untreated cells express mainly type A but dBcAMP treatment causes 30% expression of type B. This increase in type B is similar to that seen with increasing age of the culture or in adult rats. Since most cell lines express one or the other of these enzymes, but not both, this finding may be an example of true induced differentiation that parallels that occurring in vivo. Pruss et al. (1982) found that astrocytes in culture respond to fibroblast growth factor and Schwann cell mitogen while oligodendrocytes will only do so in suspended cultures but not in primary cultures. This may be related to the - 57 -astrocytes ability to divide after injury. Since the content of serum change with age of the animal, in vivo serum changes may well be what controls glial differentiation, with different cells originating from the same progenitor cells at different times dictated by the changes in serum. Table III lists some of the effects in primary cultures of changes of various substances in the culture medium. - 58 -TABLE III: EFFECTS OF CULTURE CONDITIONS ON CELL CHARACTERISTICS MEDIUM CHANGES PROBABLE MECHANISM CELL TYPE OBSERVED CHANGES AUTHORS dcAMP Through Primarily Flat epitheloid-like cells change cAMP Astrocyte to larger stellate-type that Cultures resemble astrocytes Increase GFAP and Vimentin Actin increases Actin less organized Microtubules organized into bundles extending into processes Level of most enzymes increase Levels of GS decreased Change in pattern of protein synthesis Increased effect of potassium stimulation on Na+,K+-ATPase MAO and COMT increased Proportion of MAO Type A to Type B Fedoroff et al. (1984b); Hansson and Ronnbeck (1983) Ciesielski-Treska et al. (1984) Ciesielski-Treska et al. (1982b) Ciesielski-Treska et al. (1982a) Schousboe et (1980a) al. White Se Hertz (1981) Kimelberg et al. (1978a) Hansson & Sellstrom (1983) Yu & Hertz (1982) decreased as in older cell cultures or in vivo with age. Astrocytes to stain for CA-II as intensely as oligodendrocytes Kimelberg et al, (1982) Decreased GABA uptake with both Vmax and K affected Hansson et al. (1984b) TABLE III (continued): EFFECTS OF CULTURE CONDITIONS ON CELL CHARACTERISTICS MEDIUM PROBABLE CELL CHANGES MECHANISM TYPE OBSERVED CHANGES AUTHORS dcAMP Probably through AMP Horse Serum Fetal calf serum removal Hydrocortisone Prostaglandin PGE1 C-6 Primary astroglial C-6 [S 100] protein increased Increased aspartate aminotransferase Binding pattern of concavalin-A becomes confluent Astrocyte cultures from non-serum Tabuchi et al. (1981) Tardy et al (1981) Tabuchi et al, (1981) Fischer et al. containing medium begin to express GFAP (1982) Primary astrocyte cultures Retraction of cell soma; extension Hansson & Ronnback of cell processes; decreased 3H valine (1983) incorporation into protein; decreased total soluable protein; decreased protein secretion (all returned by soluble brain extracts) Increased glutamate dehydrogenase and GABA-T (as in aged mice) Schousboe et al, (1980a) COMT and MAO increased Increased Na+,K+,ATPase and GS Through cAMP Hansson & Sellstrom (1983) Schousboe et al. (1980) Increased GABA-T and aspartate amino- Tardy et al. transferase; same morphological changes (1981) as with GFAP Glia maturation Surface Glioblasts factor receptor Astroglial maturation Lim (1977) Ito et al. (1982) Cytosine Mitotic Cerebellar Astroglial maturation arabinoside inhibitor neuronal cultures Leu et al. (1983) C) Cell development and differentiation in response  to injury Another kind of cell differentiation occurs in response to injury. The type of response to some extent is variable, depending on the type of injury, the age of the animal and the location of the injury. In general astrocytes increase in number, in size and in number of processes in response to injury to become reactive astrocytes. All oxidoreductive enzymes become more active, as do most other enzymes (Oehmichen, 1980). The increase occurs earlier for those enzymes involved in glycolysis or the hexose monophosphate shunt than for those of the citric acid cycle, such as succinic dehydrogenase (Friede, 1966, Oehmichen, 1980). The enzyme increases are permanent as they persist even in old scars. Morphological changes in astrocytes in response to ischemic injury were examined (Petito and Babiak, 1982) and found to occur within 4 0 mins. after injury. These changes consisted of expansion and increased number of mitochondria, cytoplasm and rough ER, suggesting increased metabolic activity. The number of astrocytic nuclei also increased very early. Murabe et al. (1981) found that only astrocytes changed morphology in response to kainic acid-induced damage in the hippocampus. They first swelled, then filaments developed. Polynuclear astrocytes extended processes in areas vacated by the neurons. Astrocytes appeared to have phagocytic activity. - 61 -Primary cultures derived from kainic acid lesioned rat striatum lead to 2 morphologically distinguishable cell types (Van Alstyne et al., 1983). They were mainly (95%) composed of large flat cells with ill defined junctions and no cellular processes but 5% of the cells were small with processes. Upon treatment with dBcAMP, the large immature cells transform to the smaller type. These newly derived smaller cells exhibit cell-specific markers (galactocerbroside on 10% and GFAP on 80%), plus some fetal characteristics. Therefore the larger cells were glioblasts that were in the kainic acid damaged tissue. Freide (1966) showed that the oxidoreductive enzymes in oligodendrocytes are more active than in resting astrocytes but they increase to above the oligodendrocytic level in reactive astrocytes. Although oligodendrocytes do not proliferate, they do grow and their oxidoreductase enzymes do become more active in response to trauma (Ibrahim et al., 1974). Colmant (1965) noticed increases in acid phosphatases, DPNH- and TPNH-tetrazolium reductases, succinate dehydrogenase, 5 'nucleotidase, phosphamidase, and/5-naphthol esterase. If too much damage occurs, the oligodendrocytes will die. Even with knowing that there are several sources of variation between cells that can explain much observed heterogeneity, there is still heterogeneity that does not seen to be due to these variables. Cell types in vivo and in culture are found to have a number of biochemical differences - 62 -that remain unexplained. If there are biochemical differences in glia they must subserve some differences in function. D) Heterogeneity between different glia not explained  by development or culture conditions Different glial systems even have different cellular densities depending on their origin (Henn, 1980), and ultrastructural heterogeneity has been described (Mori and Lebond, 1970). Schachner et al. (1977) found that GFAP was located in variable places in astrocytes of the mouse cerebellum. The label was found in cells around the glomerular complexes in the granular layer, in radial fibers in the molecular layer, in the sheath surrounding Purkinje cells, and in astrocytic end feet impinging on meninges and blood vessels; in white matter cell bodies there was diffuse cytoplasmic label and elongated strings of label. Mize et al. (1981) observed tritiated GABA labelling in the superior colliculus of cats and noted that dark oligodendrocytes and astrocytes accumulated GABA moderately, while light oligodendrocytes and microglia did not. The dark oligodendrocytes wrap around presynaptic terminals and are therefore likely candidates for the removal of GABA. Hosli and Hosli (1978) observed this barrier function working in the cultured glial cells of dorsal root ganglia. In mixed cultures the glia, not the neurons, would take up GABA, but, if the neurons were isolated they took GABA up better than the glial cells. This was interpreted as meaning that the glial - 63 -cells normally act as a buffer zone, forming a barrier which prevents the uptake into neurons. Levi et al. (1983) correlated the morphology expressed by astroglial cells in post-natal cerebellar, interneuron-enriched primary cultures with the ability of these cells to accumulate putative amino acid neurotransmitters. The cultures were originally composed mostly of undifferentiated GFAP-containing cells, but, during the next 12 days, the number of stellate astrocytes increased to be 70-80% of the astrocytes present and they were larger. Levi et al. noted that aspartate accumulated, as shown by 3H-D-aspartate autoradiography, into the undifferentiated GFAP-containing cells but that 3H-GABA was accumulated in substantial amounts by the stellate astrocytes. Early astrocytes of other shapes stained only lightly for GABA, and, even within the stellate population the extent of GABA labeling was variable from one cell to another. Autoradiographic examinations and determinations of the IC 50s for GABA uptake inhibitors consistently indicated that the GABA transport systems present in stellate astrocytes did not have the features generally attributed to a glial transport system but instead matched that of the inhibitory interneurons present in the culture. They noticed in neuron-enriched cultures that astrocytes may lose their ability to take up GABA as cultures grow older, even though the stellate morphology is maintained. Freide (1966) noted that not all oligodendrocytes have the same enzyme activity. Satellite cells, for example, have a marked cytochrome oxidase activity which is not observed in - 64 -other oligodendrocytes. Szuchet and Yim (1984) found an oligodendrocyte line they designated B3,f, which was morphologically homogeneous in culture, but had anti-myelin associated glycoprotein galactocerebrosial staining that varied from weak to strong between cells. Some of the differences reported in various astrocyte cultures may be due to interspecies differences. These have been shown in a few systems. Low rates of potassium uptake were observed in young rat astrocytes (Kimelberg, 1979). The rate was higher in primary cultures of chick astrocytes, where it was almost totally inhibited by ouabain (Latzovits, 1978). Mouse brain astrocytes had a much higher rate (Hertz, 1978d). This last observation seems to be a true species variation since the experiments were done in the same laboratory. Heterogeneity between and within glial cell lines There is a very extensive body of research on differences between different glial cell lines, but the research is not without its problems. The glial cell line first produced was C-6 (Benda, 1968). Now there are many types of glial cell lines frequently studied. They show characteristics that are believed to some extent to resemble normal glia. These established cell lines have several advantages: they are readily available, relatively easy to maintain for long periods of time and, because their characteristics are relatively stable, they can be compared between laboratories. However, because they were - 65 -originally transformed by chemicals or viruses, their characteristics are not entirely like glia found in normal brain. They may actually be quite different. For example, the metabolic rate of glial cells was originally believed to be quite low, based on early work done on early glial cell lines and glial scar tissue that gave erronously low metabolic rates for glia (Hertz, 1978b). They may have characteristics of two or more types of glia cell or may even have neuronal characteristics. For example, glutamate is transported into a large number of glial cell lines which have a high affinity uptake similar to that seen in neurons (Edwards et al., 1979) but not in primary cultures of glia. Such findings mean that extrapolation can not be made from glial cell lines to normal glia without corroborative evidence. They remain, however, useful tools for preliminary research because of their ease of use. Glial cell lines are defined as glia because of markers or other charcteristics they share with glia. They can sometimes be easily distinguished from neuronal cell lines but these distinctions are not always clear. Specific antigens such as NS-1 (Schachner, 1974), GI and G2 (Stallcup and Cohn, 1976) are considered glial surface markers because they are on the surface of glial tumor lines but not neuronal tumor lines and thus are used in defining new glial lines. Shine et al. (1981) found more ^-glutamyl transpeptidase in glial cell lines than in neural ones. Wilson et al. (1981) worked extensively to define glial and neuronal cell lines. He used antisera against - 66 -pseudoneuronal and pseudoglial cell lines to define the relationship between the classic cell lines and between each other. For example, the N4 antigen was expressed by the pseudoneuronal cell lines and by 7/10 neuronal lines. Pseudoneuronal and neuronal lines were further related by the finding of similar Na+ and K+ channels. On the other hand, pseudoneuronal cell lines and pseudoglial cell lines were found to be related because both possess antigens called NG1 and NG2. Wilson et al. concluded that there must be developmental linkages between neuronal and glial cell lines. Osborn et al. (1981) found that glial lines differed from each other and from primary astrocyte cultures in expression of GFAP. Cultures of normal biopsied human glial material showed no GFAP+ after seven doublings but the glial line U251 MG showed 3% and U333CG/343 MG 98% GFAP+ cells. Osborn et al. found the difference between primary cultures and glial cell lines seemed permanent as it did not reverse in response to dBcAMP. This difference may be because of changes in the genetic marker for GFAP in transformed cells or because there was a subpopulation of cells containing different genetic material that thrived in culture. The basic biochemical level of functioning in most glioma cell lines is lower than in primary cultures of glia. The C-6 line (Kimelberg, 1974) and other glioma cell lines (Hertz, 1977) , for example, have been found to have a lower Na+,KH—ATPase activity and a lower than Nernstian slope for potassium uptake than primary cultures of glial cells from the cerebellum which show a classic Nernstian slope (Sugaya et - 67 -al., 1979). NN cells are found to be less responsive to K+ stimulation (Ciesielski-Treska, 1976) than primary cultures. Glioma cell lines also have a much lower lipid content than bulk-separated astroglia (Norton et al., 1975). Not all systems function at a lower level, however; for example C-6 cells have a higher glycolytic rate than primary astrocyte cultures as shown by the higher incorporation of glucose into lactate. Their ability to maintain a higher amount of ATP in the absence of oxygen may be related to their higher glycolytic rate (Passonneau et al., 1978). Carbonic anhydrase seems to be enriched in astroglia (Roussel et al., 1979, Kimelberg et al., 1978b), but it is not in C-6 cells (DeVellis and Brooker, 1973). Various cell lines are found to have different levels of MAO and proportions of types. Astrocytes possess higher MAO activities than brain, and activities in C-6 cells (Murphy et al., 1976) are even higher. This enzyme exists in two forms: C-6 and most other glial cell lines contain only form A (Haber and Hutchison, 1976) but astrocytes in primary culture contain both, especially after exposure to dBucAMP (Hertz, 1982). Evidence for compartmentation of glutamate metabolism is not as strong in glial cell lines as in primary cultures. C-6 did not seem to show evidence of compartmentalization in properties such as GS activity but bulk isolated glial cells did. High affinity uptake of glutamate has been demonstrated autoradiographically in glial cell lines (Faivre-Baumann et al., 1974, Henn et al., 1974, Balcar et al., 1977, Pfeiffer et - 68 -al., 1976), including astrocytoma (Snodgrass and Iversen, 1974). This glutamate uptake has been shown to be sensitive to Na+ stimulation in C-6 cells (Henn, 1975). The uptake system may be different from that in primary cultures. Calcuim is not required for glutamate uptake into primary astrocytes cultures (Schousboe et al., 1977b) or the NN glial cell line (Balcar et al., 1977) but is for uptake into C-6 glioma cells (Faive-Bauman et al., 1974). Table IV indicates that the Vmax values for glutamate uptake are generally higher for primary cultures of astrocytes than for uptake into astrocytes prepared by gradient centrifugation, or glial cell lines. - 69 -Table IV: Comparative Values of Glutamate Uptake (from Hertz, 1979) Cell Description Km(^M) Vmax* Reference Astrocytes in primary culture 22 0 0.8 A " II II II 50 5.9 B " " " 11 30-90 3-7.5 C " " 11 11 10-20 0.4-0.6 D C-6 glioma 15 0.4 E " " 66 F NN glioma cells 14 0.07 G " 11 " 12-19 0.02-0.03 H MGM-LM glioma cells 2 0 0.3 I 138 MG glioma cells 65 0.14 J Bulk prepared astrocytes 12 F " 11 » 12 K " " " 10 0.06 L Bulk prepared cerebellar astrocytes 15 0.2 M Retina 21 3.5 N * ( pi mol/min per g protein) 2 Reference codes A = Schousboe et al., 1977b, B = Hertz et al., 1978b, C = Hertz et al., 1979b, D = Balcar and Hauser, 1978, E = Faivre-Bauman et al., 1974, F = Henn et al., 1974, G = Balcar et al., 1977, H = Balcar et al., 1978, I = Stewart et al., 1976, J = Walum and Weiler, 1978, K = Henn, 1976, L = Weiler et al., 1979, M = LeCampell and Shank, 1978, N = White and Neal, 1976, 0 = Schousboe et al., 1979, P = Henn and Hamberger, 1971, Q = Lasher, 1975, R = Schrier and Thompson, 1974, S = Hutchison et al., 1974, T = Schousboe et al., 1977a. - 70 -Baetge et al. (1979) also reviewed the research on a wide variety of glial cell lines and found different uptake rates for glutamate. Two glial cell lines, B28 and BE11, had very high glutamate uptake rates and another two, B15 and Bill, had only moderately high uptake; some glutamate uptake, however, occurred in most of the other cell lines. Even though their uptake rates varied, the basic mechanism did not seem to vary. They had the same specificity and were coupled to Na+ in identical ways and the Km was the same. Schousboe (1978b) found differing glutamate Km's between cell lines and primary cultures (12 0 mM for NN cells and 18 mM in primary cultures) and the influence of Na+ on the uptake of glutamate differed from cell line to cell line. He also found that the uptake was Ca++-dependent in C-6 and some other glioma lines, but not in NN, primary cultures of astrocytes or bulk prepared glial cells. Hertz (1979) also examined glutamine uptake into various preparations. Again the kinetic constants varied from preparation to preparation but the glioma line was about the average of the normal lines as indicated in Table V. - 71 -Table V: Comparative values of glutamine uptake into different glial preparations Cell Type Km(^M) Vmax* Reference Bulk-prepared astrocytes 63 0 0.16 L Astrocytes in primary culture 3 3 00 5.0 0 " " " " 150 0.2 D D138 MG glioma cell line 490 2.9 J *(amol/min per 100 mg protein (for references see Table IV) GABA uptake also varies between normal glia and tranformed glial cell lines. Henn (1975) found C-6 cells to have a high affinity uptake for GABA which was Na+ sensitive. Schousboe (1981) compared the work of many others and found that cultured astrocytes exhibit a Vmax comparable to that found in brain slices and in neurons derived from the cerebellum but that C-6 cells had a much lower capacity though it was still high affinity uptake. Hertz (1979) reviewed much of the literature (Table VI) and similarly concluded the capacity of C-6 was lower than that of astrocytes in culture. - 72 -Table VI: Comparative Values of high affinity uptake of GABA into various glial preparations Cell Type Km(>JM) Vmax* References Bulk prepared astrocytes 0.27 P " " " 0.6 K Cultured cerebellar glia 0.29 0.0001-0.0002 Q C-6 glioma 32 0.002 R " " 0.22 0.0001 S " " 50 K Cultured cerebral astrocytes 40 0.035 T " " 11 45 0.040 B * |Jmol/min per g wet weight (for references see Table IV) C-6 also have a very low activity of GABA-T compared to cultured astrocytes and bulk prepared glia (Nicklas and Browning, 1977) and much lower activity than found in the brain of mice of similar age (Nicklas and Browning, 1978). Different cell lines can also show structural differences. Pilkington et al. (1982) showed that 3 cell lines derived from a spontaneous murine astrocytoma differed in the number and ratio of 10 nm filaments and 234 nm microtubules and that these differences were related to the degree of differentiation of the cell line. However, even a given cell line can vary morphologically within and between lots. Benda (1978) even noted that C-6 cells in a single plate display differences in morphology, patterns of colony formation, and patterns of biochemistry such as accumulation of S-lOOp. These characteristics can be manipulated by serums, plating - 73 -density and other factors. For example, in serum-less medium the ability to accumulate S-lOOp is lost (Pfeiffer et al., 1970). High levels of fetal calf serum bring out a selective increase in amino acid uptake and morphological changes (Logan, 1976). Glia maturation factor (bovine) has more effect on normal cells than on C-6 tumor cells and it has to be present within a critical time factor that matches the period of postnatal gliogenesis (Kato et al., 1981). There are situations where there appears to be spontaneous differentiation under exactly the same culture conditions of cell lines which differ biochemically and morphologicially. For example, three distinct types of astrocytic cell clones came out of established cultures of 8-day postnatal mouse cerebella (Alliot and Pressac, 1984). They were all GFAP+ but differed morphologically. Type 1 had small somata, several short processes, were pseudodiploid and were thought to resemble fibrous astrocytes. Type 2 bound monoclonal antibodies BSP-3, M2 and M3, and had small somata, with two proceses, one of which was long and thin. They were thought to resemble Golgi epithelial cells. Type 3 had large flat somata, no processes, were heterodiploid, and were thought to resemble filamentous astrocytes. These characteristics were all stable in culture over time and thus represent true differentiation. Cell lines can differ from each other in basic biochemistry. For example, C-6 has higher levels of S-lOOp than do neurons, and has an Na+,K+ pumping action similar to many neurons but higher than other glial cell lines. Another - 74 -example was found by Shine et al. (1981). They showed that y-glutamyl transpeptidase, which is thought to be involved in the transport of amino acids across membranes, in the activation of biopeptides and in the detoxification of various substances, has a tremendous variation between glial cell lines. It is highest in C-6 and lowest in human A1B1. Cell lines differ in various transmitter systems as well. AChE activity can be found in only certain clonal lines of glial cells such as C-6 (Vernadakis and Arnold, 1980). Certain cell clones exist that are particularly high in one or another putative amino acid transmitter. Cambier et al. (1983) created a glycine-enriched astrocytes clone, K55, derived from mouse cerebellar astrocyte cultures transformed by simian virus-40. They found that a high percentage of the astrocytic cell clones, derived from mouse cerebellar culture by simian virus-4 0 or by spontaneous transformation, contain high amounts of glycine (Cambier and Pessac, 1983), while the oligodendrocyte-like clones were high in alanine. They also noted that astrocytic cell clones used glutamine differently than did the other cell types (Cambier and Pessac, 1983). Schousboe (1978a) found that C-6 astrocytoma and primary cultures of astrocytes have a high capacity for taurine uptake while that in the NN line was lower. Drummond and Phillips (1977) found differences in amino acid levels in different cell lines which were not well correlated with the cell class. The amino acid levels were dependent on tissue culture conditions but, if these conditions were carefully controlled, some statistically - 75 -significant differences were still found. GABA levels were found to be particularly high in both C-6 and B92 glial lines. Glutamate levels in various cell lines varied between 50.8 to 158 nmol/mg protein and glutamine levels from 0.8 to 107 nmol/mg protein. Statistically significant differences were also observed for aspartate, proline, glycine, alanine, valine, cystathionine, isoleucine, and leucine. The uptake of amino acids by these different clones does not necessarily vary in the same way as the levels. Schier and Thompson (1974) examined uptake of putative neurotransmitters by three cultured glial cell lines. The cell lines exhibited similar rapid uptake of glutamate and Na+ dependent uptake of GABA, as well as pyridoxal-dependent GABA synthesis and excretion. Taurine uptake occurred in all three, with each showing a fast saturable component and a slow non-saturable component which varied in magnitude between the cell lines. There was one cell line which could maintain a high concentration gradient of taurine. Synthesis of taurine from cysteine was only found in one of these lines. Glial cell lines may also respond differently to drugs. Elkouby et al. (1982) found that two glial cell lines, NN astrocytoma and C-6 glioma, responded differently to the hormones hydrocortisone and thyroxine. The activity of Ca++,-Mg++ ATPase increased in the NN line but decreased in C-6 in response to these hormones. Another drug, dexametiasone, can be used to induce GS in only a subset of C-6 cells (Holbrook et al., 1981). Bigner et al. (1981) examined various characteristics of - 76 -fifteen lines of human cells traditionally thought of as being gliomas. They showed a wide variety of human leukocyte antigen phenotypes. All but two, which were from a black patient, had type B glucose-6-phosphate dehydrogenase isoenzymes. Only four could be transplanted into athymic mice, two of which grew and then regressed. Only two were GFAP+. Thus each line had a unique profile. Cell line research therefore shows a variety of types of heterogeneity. The significance is unknown but there are a number of sources of variance that may explain some of the differences. First, being transformed cells, they may be expressing some new genotype. Second, the transformed cell may be expressing different parts of the genotype than is normally expressed by the parent cell resulting in a mixing of characteristics. Third, these various glial types may be derived from different types of parent cells and retain the differences. Different subtypes of the progenitor cells may be related to variables we have already discussed or perhaps to the areas of the brain from which the cell came. Differences in glial cells from different areas of the brain There has been a wide variety of research that has shown regional heterogeneity in glial cells. Much of the data were generated by people who did not set out to show differences between regions or are minor observations in a paper on another topic. There are probably many more examples buried in the literature. Such differences have not been emphasized in indices to the literature because it was not until recently - 77 -that an interest in this subject developed . It has been known for a long time that the morphology of glial cells varies between different areas of the brain. Examples include glial cells that are specialized enough to have specific names, such as Bergmann glia, whose variance in morphology has been previously discribed. In addition, there are areas of brain where glial cells appear morphologically different but have not been given specific names. Astrocytes of the hippocampus, for example, have a characteristic shape that is different from that seen in other areas. Astrocytes are known to have several different types of GFAP of different molecular weights and different solubilities in water, with those of high molecular weights being the least water soluble (Eng, 1982). These forms are unevenly distributed in the brain even though they are all carried on the same gene (Gheuens et al., 1984). Since GFAP is known to influence the shape of astrocytes, this might be part of an explanation for some of the shape differences. GFAP varies not only in structure but in its schedule of appearance during development. Weir et al. (1984) measured GFAP in olfactory bulbs, forebrain and cerebellum of rats during development, using a double antibody radioimmunoassay. Each brain region showed a different pattern of development for GFAP. At birth, GFAP protein in the olfactory bulb was 85 times that in forebrain, and 485 times that in cerebellum. The increase in GFAP corresponded with maturation more than proliferation. The pattern of increase in GS activity was similar to that of GFAP in the forebrain and olfactory bulbs - 78 -but differed markedly in the cerebellum. In the cerebellum the maximum increase in GFAP occurred after the peak of astroglial proliferation and 1 week before maximum acquisition of GS and S-100 protein. The distribution of astroglial contacts on the surface of neurons varies greatly among brain areas as well as among different types of neurons (Guldner and Wolff, 1973, Peters and Palay, 1965, and Wolff, 1965). Neurons and synapses may even be wrapped differently by several layers of glial lamellae (Guldner and Wolff, 1973, Palay, 1966, Specek, 1968, and Szentagothai, 1970). Palay and Chan-Palay (1974) showed, for example, that Purkinje cells are largely covered by Bergmann glia in contrast to cerebellar interneurons which are not wrapped. Wolff and Guldner (1978) found that electrical stimulation produced swelling of astrocytic processes in the neocortex. Since this experimentally produced feature of cortical astrocytes exists normally in certain other astroglial cells it is suggested that variations of the structure and arrangement of astroglial processes between different brain regions may reflect neuronal activities. There is, however, some evidence (discussed below) that these characteristics are not just responses to neuronal influences but are stable characteristics of the glia in various regions. There have been numerous observations of differences in number of glial cells in various brain areas. Szeligo and Leglond (1977) not only found differences in numbers but also showed that handling or enriched environments caused increases - 79 -in the numbers of astrocytes and oligodendrocytes in only certain layers of the cortex and not in other areas, such as the corpus callosum. Oehmichen (1980) reported that astrocytes are observed in varying densities in the CNS and that the functional activity in the resting state is quantitatively different depending on location. For example, strong phosphorylase activity has been found in those areas that have a tendency to accumulate glycogen (Mossakowski and Penar, 1972, Oehmichen, 1980). Others have confirmed the variability from area to area of glycogen storage in radial glial cells of developing rat brain (Bruckner and Biesold, 1981). Not only can glycogen storage be seen in different concentration in various areas but the ratio of glia vs. neuronal incorporation of precursors into glycoconjugates varies from area to area. Higher incorporation levels were found in the supraoptic and arcuate nucleus and lowest in cerebellum. Other indices of glial metabolism can also vary. Glucose uptake varies widely from area to area. Thompson et al. (1980) showed that creatine kinase BB isoenzyme was localized only to astrocytes of the white matter of human cerebrum. This enzyme is normally associated with cells that have high adenosine triphosphate (ATP) regenerating capabilities such as cells involved in transport or contractile systems. Therefore astrocytes of the white matter would appear to have specialized functions. DeVellis et al. (1967) found that there is a regional - 80 -difference in the inducability of glucose phosphate dehydrogenase with the cerebellum and brainstem showing higher levels of induction then the cerebral hemispheres. This can not be explained by developmental time tables for glia. Kreutzberg and Hussain (1982) showed that Muller cells of the external retinal layers but not the internal layers have 51-nucleotidase on their membranes. This enzyme functions to hydrolyse monophosphates such as AMP. The reason for this difference is unknown but the enzyme has not been found on astrocytes of other areas. There is a large literature on differences in various transmitter-related indices between glia isolated from various regions of the brain. Differences between prefrontal cortex and visual cortex have been found in content of catcholamines (Bjorklund et al., 1978), and in membrane binding for naloxone, diazepam and a muscarinic ligand quinuclidinyl benzilate (Divac and Braestrup, 1978). Hansson et al. (1984a) showed that astroglial cultures from various regions of the brain showed increased cAMP after incubation with dopamine or apomorphine; the increase could be blocked by a dopamine antagonist. Such increase was most pronounced in a subpopulation of cells from the striatum and least in cells from the brain stem. Astrocytes prepared from areas rich in dopamine show dopamine binding that can be blocked by the dopamine antagonists chloropromazine, haloperidol, and other antipsychotic drugs, but astrocytes from non-dopamine containing parts of the brain do not have this ability (Hansson et al, 1984b). This means that there - 81 -must be specialized cells in dopamine-rich areas and that these characteristics of such specialized glia are stable in culture where they are not under neuronal influence. Hansson (1984) measured the activities of both MAO and COMT in primary astroglial cultures from newborn rat brain cultivated from six different regions and in brain homogenates from these same regions. The areas compared were the cerebral cortex, striatum, hippocampus, brain stem, and cerebellum. MAO activity was higher in the cultures from the striatum than in those from the other brain regions. Striatal homogenates showed the same trend which conflicts with the results of Hazama et al. (1976) who found no differences in the homogenates. COMT activity was the same in neonatal cultures and adult brain homogenates and also showed regional differences. The lowest activity was found in the brain stem, with higher levels in the cortex, striatum and cerebellum and the highest in the hippocampus. Henn and Henn (1980) found that glia from the caudate had a much higher number of haloperidol binding sites and more dopamine sensitive adenylate cyclase than those from other brain regions. Even so, the binding sites are located on only a fraction of the astroglial cells of the caudate. A very surprising finding was that by Denis-Donini et al. (1984) who showed that different glial populations affect the morphology of mouse mesencephalic dopaminergic neurons. Glial monolayers cultured from the striatal or the mesencephalic region of the embryonic brain were used to grow dopaminergic neurons from the mesencephalon. On mesencephalic glial cells - 82 -the majority of the dopamine neurons developed a great number of highly branched and varicose neurites, whereas on striatal glia they only exhibited one long, thin, linear neurite. The morphology of the underlying glia was not very different but they were not equally stained with GFAP, thus showing some heterogeneity in the level of expression of glial filament. Thus the classic assumption of glia only responding to their neuronal environment is actually found to be reversed. Goldlefter (1976) found that the periventricular glia of the hypothalamus were positive with gonadotrophin or to Gomori's stain and such staining increased on treatment with dopamine. Thus the glia of this area respond to neurotransmitter and to a hormone produced by surrounding cells. Schousboe (1978b) found that high affinity uptake of GABA occurred in peripheral ganglia, rat retina, glioma cell lines, spinal cord explant cultures, and primary cultures of glial cells from the cerebellum and cerebrum but not from other areas of the brain. It was only in glial cell cultures from the cerebellum and cerebrum that the level of uptake was comparable to that in brain slices. This astrocytic uptake was different from the neuronal system since it was selectively inhibited by |3-proline but not by two selective inhibitors of neuronal GABA uptake. Glial cells may also vary in their response to neurotransmitters. Krnjevic and Schwartz (1967) first showed that GABA applied iontophoretically caused depolarization of some, but not all, glial cells in the cortex. They could not, - 83 -however, rule out the possibility that the selectivity depended on proximity to GABA-depolarized neurons which released K+ that, in turn, depolarized the nearby glia. GABA-T, the degradative enzyme for GABA, showed no regional differences (Hansson, 1984) in primary cultures from the cerebral cortex, striatum, hippocampus, brain stem, and cerebellum of newborn rat brain. Glutamate indices have also been measured and found to vary regionally. Autoradiographical studies of glutamate or D-aspartate high affinity uptake (Currie and Kelly, 1981) showed extensive uptake into cerebellar glia, especially Bergmann glia, and that this decreased after transection of certain projections. This implied that these differences are a result of influences of neurons on glia, not the stable glia characteristics that other research was indicated. They also noted other differences in glutamate uptake from different regions. Hansson (1983) also used autoradiography to show regional differences in uptake. She found that glutamate, and to a lesser extent, aspartate, was taken up readily in cultures from the cerebral cortex, hippocampus, and striatum and, to a lesser extent, in cultures from the brainstem and cerebellum. This is evidence for stable glial differences in glutamate uptake. Valine, an amino acid which is incorporated mostly into protein, was used as an internal control and was found to be accumulated to the same extent in the various primary cultures. Schousboe (1978a) showed a range in values in glutamate uptake by astrocytes cultured from different brain regions. The Vmax ranged from 8 nmol./min/mg cell - 84 -protein in cells from whole cerebrum to 60 nmol./min/mg in cells cultured from cerebral cortex, with Km varying from 22 0^ M to 50 JAM. Schousboe and Divac (1979) further showed that the glutamate uptake in primary astrocyte cultures from neonatal mice after three weeks in culture was greater in cells originating from the prefrontal cortex and neostriatum than in those originating from the occipital cortex or cerebellum. These results generally correlate with the synaptosomal uptake of glutamate in these regions and indicate that this glial characteristic was stable for at least three weeks in culture without neuronal influences. Drejer et al. (1982) did a similar experiment and found the following Vmax values for astrocytes: prefrontal cortex -13.9, occipital cortex - 11.4, neostriatum - 27.3, and cerebellum - 5.8 nmol/min/mg cell protein. There were only minor differences in Km between regions except in the neostriatum where it was slightly higher. Differences in Vmax and not Km mean that there are differences in the number but not in the properties of the transport sites. Again the authors noted the apparent relationship between the regional ability of glia to accumulate glutamate and number of glutaminergic terminals. Glycine is an important inhibitory transmitter at the spinal level but not in the forebrain. It has been found that gradient-separated astrocytes from spinal cord, but not those from frontal cortex, show a high affinity uptake of glycine (Henn, 1980). Others have confirmed that the distribution of glial transport systems for glycine follows the same - 85 -distribution as glycine (Hokfelt and Lungdahl, 1971, Matus and Dennison, 1971) . The conclusions of these observations on GABA, glutamate, and glycine is that there are probably differences in the numbers of uptake sites in glial cells in various brain regions and that these are stable in culture. Moreover, the glial uptake seems to correlate to some extent with the regional density of the amino acid boutons. Schousboe et al. (1980b) suggested that this glial heterogeneity must be taken into account in the interpretation of neurochemical changes resulting from specific neuronal degrenerations. For example, the effects of gliosis after kainic acid lesions could seriously affect interpretation of biochemical changes. Summary of evidence for biochemical differentiation in glia The evidences for regional biochemical differences in glia or cultured glia is thus quite strong. Some differences may be because of direct effects of surrounding neurons but some are stable in culture after the effect of the neurons is no longer there. These stable differences may be integral parts of the genetic makeup of these glia or may be initiated at some critical developmental point by its environment. Questions of this nature have not yet been answered. There is evidence that reverse effects may be operative. Paterson et al. (1977) showed that glial cells release some factor that influences the amount of neurotransmitter synthesized by sympathetically derived neurons either by co-cultured or conditioned medium. They also found that C-6 and sympathetic - 86 -satellite cells both influence growth and development of cholinergic synapses and ACh synthesis. There is also some evidence of species variations. There are, for example, considerable differences in the rate of potassium uptake in astrocytes cultured from chick, rat or mouse brain. Thus future research must be extremely careful in transfering experiments from one species to another. If these biochemical differences between glia of different areas and species stand the test of time, then the difference must be explored further and considered in much of the on-going neurochemical research. In experimental conditions causing damage leading to gliosis, some of the biochemical changes will undoubtedly be found to be due to glial changes. Research on many diseases may have to consider glia as being possibly involved in the etiology. There are already research findings in some diseases that point to this. For example, Carter (1981) observed that GS activity was reduced in Huntington's disease in some areas where it could not be accounted for by cell loss. It has also been observed that thiamine deficient models of Wernicke-Korsakoff's syndrome produce damage first in glial cells of certain areas of the brain (Collins, 1968; Collins and Converse, 1970). Research aimed at identifying differences in glia has yielded much. But there is also in the vast literature on staining of brain cells many coincidental reports of staining of subsets of glia; such reports tend to be buried in the generalized literature because interest in glia has been so little compared to the interest in neurons. - 87 -EXPERIMENTAL RATIONALE AND ABSTRACT I have used two unrelated staining procedures that stain predominantly glial cells, but not all glial cells, only subsets of them. I have also looked at a model of Wernicke-Korsakoff's syndrome that demonstrates that the disease may damage the glial cells of only some areas and before neuronal damage occurs in these areas. In Experiment 1, hemosiderin, a form of iron, was examined in the brains of rats using a Prussian Blue followed by diaminobenzidine (DAB) procedure. The areas of the brain containing the various types of cellular and non-cellular staining were mapped. Iron was found to be predominantly located in or on oligodendrocytes, but not in all areas as there was a distinct regional pattern of staining. There was also some staining in neurons, ependymal cells and astrocytes of specific and restricted areas, and various levels of background staining. The background staining is probably terminal boutons on unstained cells or neuronal or glial processes. The results are compared to the known anatomy of several neurotransmitter systems. Significant overlap of the location of iron staining was noted with GABA, dopamine, endorphins and enkephalins. In Experiment 2, a modification of the method of Van Gelder (1965) for histochemical staining of GABA-T containing cells was used to stain cells containing some enzymes catalyzing a possible alternative route for glutamate production in brain: from proline or ornithine which is - 88 -oxidized to glutamate via l-pyrroline-5-carboxylate (P5C) by 1-pyrroline dehydrogenase (EC;PDH). PDH has been demonstrated in several bacteria and mammalian systems (Fig. 5) and, in our experiment, was found to be exclusively in glial cells such as the Bergmann glia of the cerebellum and astrocytes of the hippocampus. P5C can be formed from proline by the action of proline oxidase (pyrroline-5-carboxylate reductase, EC,PrO). This enzyme was also localized exclusively in glial cells but the staining was much less distinct. Both of these experiments provide additional evidence of glial cell specialization. Experiment 3 only postulates glial involvement in thiamine deficiency as the technique does not allow for cellular histochemistry. Pyrithiamine, a thiamine phosphokinase inhibitor, was fed to rats on a thiamine-deficient diet to create an animal model of Wernicke's encephalopathy. Symptoms of weight loss, ataxia, and loss of righting reflex were produced in rats in ten days. At this time some rats were sacrificed and the rest of the rats were returned to a normal diet, to be sacrificed only when their weight had returned to their pre-experimental level. Rats used for biochemical measurements were sacrificed by cervical fracture,the brains dissected into eight regions, and glutamic acid decarboxylase (GAD) and choline acetyltransferase (CAT) activity measured on the brain homogenates. Other rats were perfused for histological observation of GABA-T, by a modification of a - 89 -method of Van Gelder (1965). GAD activity was found to be significantly reduced in symptomatic rats in the thalamus > cerebellum > pons/medulla > and mid-brain. GABA-T staining was found to be similarly reduced, with greatest losses in the thalamus > inferior colliculus > pons > and medulla. CAT activity was not significantly altered in any brain areas. Upon return to a normal diet, recovery of GAD was significant only in the thalamus, while GABA-T staining recovered at least partially in all areas affected. These results are discussed in terms of glial specificity and effects these new assumptions might have on the interpretation of the results. - 90 -EXPERIMENT 1 There have been few studies examining the cellular distribution of iron in brain but iron may have an important role in the brain, and may be involved in several disease processes. The role of iron in the CNS is not yet understood but low dietary iron is known to have a number of effects on brain function, including effects on the electroencephalogram (EEG) (Tucker, 1982), and disturbances in circadian rhythm, thermoregulation, motor activity (Youdim et al., 1981) and decreased attentiveness (Youdim et al., 1980). The mechanism of production of symptoms in iron deficienty is thought to be at least partly through neurotransmitters although the reduced capacity of the blood to carry oxygen may have an indirect effect on the brain. Chronic, slightly elevated levels of iron have been shown to be toxic to both ACh and GABA neurons (Swainman, 1984). Iron deposits can also occur in certain diseases such as in Hallervorden-Spatz (Bronson, 1980), Huntington's disease or Parkinson's disease (Swainman, 1981). Non-heme iron exists in two forms in the brain: ferritin is iron held in storage by a protein forming globules in lysosomes in some parts of the brain, and hemosiderin, which is ferric hydroxide granules deposited more evenly in cell bodies and processes. Hemosiderin is probably the form active in the brain. Hemosiderin releases ferric iron on exposure to hydrogen chloride and potassium ferrocynanide can react with the ferric iron producing ferric ferrocyanide (Prussian Blue). This classic Perl's reaction can be intensified using a - 91 -procedure of Nguyen-Legros et al (1980). Diaminobenzidine is added to the Prussian Blue, allowing the Prussian Blue to act as a catalyst for the oxidation of DAB by hydrogen peroxide forming an intense brown deposit where the iron is. This is the procedure we used to stain for iron. We did a detailed map and analysis of both cellular and non-cellular iron which allowed correlation of iron with known neurotransmitter anatomy. METHOD A method similar to that of Nguyen-Legros et al. (1980) was used. Two solutions were made up just before experimental procedures were started. Solution A: 4% hydrogen chloride. Solution B: 4% ferrocyanide. Brains that had been perfused with phosphate buffered saline followed by 4% formaldehyde/4% gluteraldehyde, and stored for at least three days in the same fixatives, were used. Slices were cut on a cryostat at 50 and reacted for twenty minutes in a 50% mixture of solutions A and B. If the reaction proceeded correctly, the solution should be yellow not blue or green. The slices were then washed in 0.1 M phosphate buffer, pH 7.4, for 3 to 5 min. While the sections were washing, the DAB reagent was made up. 2 0 mg DAB was mixed into 100 ml tris buffer pH. 7.6, and 2 drops of hydrogen peroxide (30%) are added. The DAB (3.3'-diaminobenzidine tetrahydrochloride monohydrate 97%) was obtained from Aldrich. - 92 -The sections were placed in this reaction mixture for 10 min. in the dark, then taken out and washed, mounted, dehydrated and coverslipped. (The darkness of the stain depends on the amount of H2O2 and the time in this reaction mixture). RESULTS Several types of staining were seen. There were various levels of background staining without clear cellular morphology present (Fig. la & b), areas of high background with definite cellular staining (Fig. Ic & d), and areas with very low background staining but with definite cellular staining (Fig. Ie & f) as well as gradations in between. There were also areas of neuronal staining (Fig. lg) and of astrocytic staining (Fig. lh). There were gradations in the background staining of various areas such that it was a matter of judgement to decide which areas were to be called high, medium, low or no background staining. I feel that the background staining is probably a combination of staining of cellular processes and nerve endings. Fig. 2 shows photographs of sagittal sections of iron stained sections showing the density of stain. Fig. 3 gives maps corresponding to the photographs showing where there was individual cellular staining (circles), or high or medium levels of background staining (dots) or both. Fig. 4 presents coronal sections and the corresponding maps. Because of the judgemental nature of the mapping, the photographs may be useful in providing more detailed information as to the - 93 -density of the background staining than the maps and table can provide, but the photographs must be used cautiously in this regard since some dark areas may just reflect a high density of the cellular staining. All areas containing cellular staining are marked on the schematic maps and are judged to be nonambiguous. Most of the cellular staining is thought to be of oligodendrocytes but there are isolated individual cells that are probably neuronal (Fig. lg). There were also limited areas in the olfactory bulb and olfactory tract that had what appeared to be staining of fibrous astrocytes on a low background area (Fig. Ih) and other areas in the olfactory bulb that had what appeared to be a mixture of stained fibrous astrocytes and neurons or oligodendrocytes in a high background area. There were also regions in the area postrema and around the ventricles where the staining appeared to be predominantly in epithelial cells. Table VIII summarizes the areas showing various types of staining. DISCUSSION The most interesting observation that can be made from our results is that glial cell staining is not the same in all areas of the brain. This uneven distribution of stained glial cells tends to support further other observations of glial cell specializtion. This must indicate that glial cells are biochemically different in their iron metabolism and in iron-related functions, whatever they may be. The function of - 94 -iron is not understood in the brain, but our observations of regional heterogeneity in iron density and cellular location may be correlated with other information in an attempt to assess in which transmitter systems iron-rich glial cells may be involved. There are numerous theories proposed as to how iron is involved in the brain. Early iron localization studies using Turnbull blue (Spatz, 1922, Diezel, 1954) localized iron to the glial cells of the globus pallidus, and the substantia nigra, and, to a lesser extent, the red nucleus, striate body and Luys1 body, all structures of the extrapyramidal system. Spatz noted that iron deposits occurred in diseases involving the extrapyramidal system such as Parkinson's, Hallervorden Spatz's and Huntington's diseases. Based on these observations it was proposed that iron may be involved in dopamine metabolism because of the known importance of dopamine in the extrapyramidal system. Supporting evidence included observations that low iron caused a reduced hypothermic effect of D-amphetamine and increased apomorphine induced stereotypic behaviour (Youdim et al., 1981). Both effects are mediated by dopamine systems. It was postulated that iron may function as a cofactor for tyrosine and tryptophan hydroxylases (Youdim et al., 1984) or may be involved in dopamine receptor functions (Youdim et al., 1980). My findings are similar to those in Spatz's early work except there is no staining in the subthalamus (Luys body). My findings do show some correlation with dopamine distribution but there are areas high in dopamine that do not - 95 -have specific iron staining and areas of iron staining where there are no known dopamine tracts, projections or cell bodies. One of the major dopamine pathways is that from the zona compacta of the substantia nigra and cells just medial to it to the caudate, putamen, globus pallidus, olfactory tubercle, nucleus accumbens, and lateral amygdala nucleus and frontal cortex. In my findings the substantia nigra has stained glial cells, as do the caudate-putamen, globus pallidus, amygdala and olfactory tubercle and all these areas have high or medium background staining as well. But I find no iron staining in cells medial to the substantia nigra, and only medium background staining in the nucleus accumbens. Small branches of this dopamine system are supposed to ascend to the frontal cortex, anterior cortex, and septum. I find no staining in any part of the cortex although there is a uniform low background level. There is, however, medium background staining in some septal areas as well as neuronal staining in the lateral septum. There are other dopamine pathways such as the one from the arcuate nucleus of the hypothalamus to the median eminence. My findings show the arcuate nucleus has oligodendrocyte staining on a medium background. I did not stain sections containing the median eminence but Hill and Switzer (1984) found a high concentration of iron stained ependymal cells in that region. There are cells in the medial dorsal nucleus of the hypothalamus that are thought to be dopaminergic that project to the thalamus and zona incerta. I find that the medial - 96 -dorsal nucleus of the hypothalamus has stained oligodendrocytes with a high background, the thalamus has medium staining and some areas with positive oligodendrocytes, and the zona incerta has stained oligodendrocytes but no background staining. There are also dopamine interneurons in the hypothalamus, brain stem and olfactory bulb. These are all areas that contain some background staining with stained cells in the olfactory bulb. Thus all areas of dopamine cell bodies except the area medial to the substantia nigra also contain stained oligodendrocytes but in the dopamine terminal areas there is everything from positive staining of various cell types to no cell staining, and a range from low to high in background staining. It may be relevant that the dopaminergic areas which show the least iron staining are generally those of the A10 system in which dopamine and cholecystokinin are colocalized. My evidence is somewhat supportive of iron involvement in dopamine metabolism, particularly around non-peptidergic dopamine cell bodies, but the lack of a total match means that iron does not exist exclusively in association with dopamine. Other researchers have tried to correlate iron distribution with GABA neuroanatomy. Glutamate-binding protein is known to contain iron and is required for GABA regeneration (Michaelis et al., 1982). Francois et al. (1981) observed that the GABA striato- or pallido- nigral and cerebellar cortical pathways overlap significantly with iron - 97 -distribution. They also noted that areas high in GAD such as in the superior colliculus and the nucleus interpeduncularis, were also high in iron (Francois et al., 1981.). My observations confirm their findings of iron in all these areas except the cortex, where we find only low to medium background and no cellular staining, and the cerebellar cortex, where there is only medium background staining. Hill and Switzer (1984) did a study similar to mine using the same technique and found similar but not identical results and concluded that high iron concentrations in glia overlapped most significantly with areas high in GAD and GABA; these areas included the ventral pallidum, globus pallidus, substantia nigra pars reticulata, and cerebellar nuclei. They pointed out that injections of GABA into the globus pallidus led to reductions in iron in the ipsilateral ventral pallidum, globus pallidus and substantia nigra (Hill, 1984). They thought, however, that the distribution of iron indicated it was not exclusively related to GABA but might be involved in other neurotransmitter systems such as enkephalins. My results do not support the involvement of iron in GABA as strongly as do those of Hill and Switzer. GABA is thought to be the transmitter of the Purkinje cells of the cerebellum which project to the cerebellar nuclei and of the cerebellar basket cells, Golgi cells and the stellate cells, which are all wholly contained in the cerebellar grey matter. However, my results do not show any cellular staining in the cerebellar cortex and only a medium amount of background staining. Although I do find significant oligodendrocyte and possibly - 98 -neuronal staining in the cerebellar nuclei. Overall, this does not provide strong evidence for the involvement of iron in cerebellar GABA systems. The pattern of hippocampal staining is consistent with iron's involvement in GABA processes in that nucleus as the only area of staining is a narrow band of medium background staining around the middle of hippocampal layers where there are basket cells which are GABAergic. The high levels of stained cells and background in the globus pallidus and the pars reticulata of the substantia nigra are consistent with the well established GABAergic projections between these two structures. GABA is also the neurotransmitter of interneurons of the olfactory bulb which might be consistent with the observations of cellular staining for iron in that region. However, GABA is so ubiquitous in brain that if all GABA systems were associated with iron-rich glia or other structure, one would expect a far more even distribution of iron than found in this or previous studies. There is little correlation between areas of high [3H]-GABA uptake (Iversen and Schon, 1973), and high iron staining areas, except that the substantia nigra is high in both. Thus my results only give limited support to iron involvement in GABA metabolism in some areas of brain with these areas including those in which Hill showed reductions in iron after pallidal injection of GABA. Several researchers have suggested a connection between 5HT and iron. It was observed that iron-deficient synaptosomes take up less 5HT than normal synaptosomes and, when iron is returned to the diet, uptake increased (Kaladhar - 99 -and Rao, 1982). This phenomena extended to offspring of iron deficient mothers (Kaladher and Rao, 1983). These authors suggest an iron dependent serotonin binding protein or some other involvement of iron in vesicular storage of 5HT. Tamir et al. (1976) noted that the binding of serotonin by serotonin binding protein was enhanced by Fe2+. Most 5HT neurons are located in the raphe or reticular system and project to the neostriatum, cortex, thalamus, hippocampus, cerebellum, preoptic nucleus, septal nuclei or pons. Our results show positive cells or, at least, medium background staining in all the above areas except the cortex, but again there is no consistent staining pattern differentiating areas of projection and cell bodies. The ependymal cells lining the third ventricle are serotonergic and stain heavily for iron which might be interpreted as some support for the involvement of iron in 5HT systems. There is a striking overlap of iron distribution with some aspects of enkephalin neuroanatomy. There are both iron staining and enkephalin cells in the lateral septum, bed nucleus of the stria terminalis, striatum, hypothalamus, amygdala, substantia nigra, medial vestibular nucleus, nucleus of the spinal tract of the trigeminal, and the periaquaductal gray, although in the bed nucleus of the stria terminalis and the spinal tract of the trigeminal the staining is only a medium background staining. P-Endorphins and related substances also have a similar and extensive overlapping pattern with iron distribution. Our evidence indicates that, if iron is involved in any - 100 -specific neurotransmitter system, it is not involved in a simple way. It may be involved in two or more transmitter systems, or be involved in some other, as yet unhypothesized, processes. Our evidence does not eliminate any of the theories previously advanced but neither does it wholly support any one of them either. A recent study (Y. Noda unpublished) examined the effects of a 20 month normal, iron deficient, or iron abundant diet on three enzymes: CAT, GAD, and tyrosine hydroxylase (TH). The results showed that GAD and, to some extent, TH activity is inversely related to the amount of iron in the diet in all brain regions examined. CAT activities were unaffected. Noda thought that iron must be essential for both GABAergic and catcholaminergic systems but concluded that excessive iron might result in degeneration of the neurons. Iron deposits can be harmful and do occur in some diseases as mentioned earlier, and in the same structures that are normally high in iron. High levels of iron may be harmful because it can lead to the generation of oxygen free radicals, OH1 due to the iron mediated coupling of O2 and H2O2/ the so called Haber-Weiss reaction. The presence of iron in subsets of glial cells might suggest that: (1) the iron is necessary for some function of these particular glial cells; or (2) the iron is essential for certain types of neurons, and the glial cells surrounding them are either supplying iron to these neurons or scavenging it from the extracellular space around them. The most significant finding of this research is the - 101 -extensive localization of hemosiderin to glia and the fact that this localization is different in various brain regions. - 102 -Figure 1: Various Types of Staining for Iron in Rat Brain Fig. IA Midbrain areas showing several densities of background staining without any individually stained cells. Calibration bar = 1000 /um. Fig. IB Band of medium staining with no cells above the pyramidal cell layer of the hippocampus. Calibration bar = 300 jam. Fig. IC Area in the globus pallidus with moderately heavy background staining and clearly stained cells, probably oligodendrocytes. Calibration bar = 300 fKm. Fig. ID Strands of light and dark background staining with heavily stained oligodendrocytes among dark strands in the striatum. Calibration bar = 300 pirn. - 103 -- 104 -Fig. IE Several stained oligodendrocytes in a lightly stained area of the striatum. Calibration bar = 300^on. Fig. IF Interfasicular oligodendrocytes against light background staining of corpus callosum. Calibration bar = 300 /^m. Fig. 1G Lightly stained neurons on a light background in lateral septum. Calibration bar = 300 /im. Fig. IH Area in olfactory bulb with light background staining showing probable astrocytic staining. Calibration bar = 300 ijim. - 105 -- 106 -Fig. 2 Photographs of sagittal sections of whole rat brain. Fig. 2a 0.5mm of the midline. Fig. 2b, 1.2 mm of the midline. Fig. 2c, 2.9 mm off the midline. Calibration bar = 1 mm. - 107 -- 108 -Fig. 3 Schematic diagrams of figure 2. Circles indicate area of cellular staining, and dots indicate high background staining. Calibration bar = 1 mm. (See Table VII for abbreviations.) - 109 -- 110 -Fig. 4 Half photographs and half schematic drawing of coronal sections of rat brain. Fig. 4a, 3.2 mm anterior to bregma. Fig. 4b, 1.4 mm anterior to bregma. Fig. 4c, 0.6 mm anterior to bregma. Fig. 4d, 2.0 mm posterior to bregma. Calibration bar = 1mm. (See Table VII for abbreviation explanations) - Ill -- He! Fig. 4 (Continued) Half photographs and half schematic drawing of coronal sections of rat brain. Fig. 4 e, 4.0 mm posterior to bregma. Fig. 4f, 6.0 mm posterior to bregma. Fig. 4g, 8.8 mm posterior to bregma. Fig. 4h, 11.4 mm posterior to bregma. Circles indicate areas of cellular staining and dots of high background staining. Calibration bar = 1 mm. (See Table VII for abbreviation explanations) - 113 -- 114 TABLE VII: IRON STAINING IN VARIOUS AREAS OP THE BRAIN Table VII summarizes the type of staining in various structures. All structures not mentioned have no cells and low or no background staining. H, M, L - high, medium, low background staining, 0 - oligodendrocytes, N - neurons, A - astrocytes,E - epithelial cells. Brain Structure Symbol Background Cell Types Nucleus Accumbens Septi ACB M none Central Amygdala ACE M 0 Anterior Hypothalamic Area AHA M 0, N? Lateral Amygdaloid Nucleus AL M none Accessory Olfactory Bulb AOB M 0, N? Area Postrema AP H E? Arcuate Nucleus of Hypothalamus ARH M 0 Bed Nucleus of the Anterior BCA M 0 Commissure Bed Nucleus of Stria Terminalis BST M none Anterior Commissure CA none Strings Corpus Callosum CC none Strings Cerebellar Grey CG M none Inferior Colliculus CIF M 0 Caudate Putamen CPUH in Strings 0 Superior Colliculus CS M 0 Commissure of the Superior CSC none 0 Colliculus Lateral Cuneate Nucleus CUL M none Decussations of Medial Lemniscus DLM M 0 Dorsal Medial Nucleus of DMH H 0 the Hypothalamus Dorsal Raphe DR M none Endopeduncular Nucleus EP L 0 External Plexiform Layer EPL L 0, A of Olfactory Bulb Fornix FX M in strands non< Geniculate Body G M none Globus Pallidus GP H 0, N? Nucleus Gracilis GR L 0? Hypothalamus (all other areas) H M 0 Habenular Nucleus HN H none Hippocampus CA 3 HP M none Islands of Calleja IC H none Inferior Olfactory Bulb IGL M 0?, A Interpeduncular Nucleus IP H 0 Locus Coeruleus LC M 0 Lateral Hypothalmic Area LHA M none Lateral Lemniscus LL H 0 Dorsal Nucleus of the LLD M 0 Lateral Lemniscus Medial Lemniscus LM L 0 Lateral Septal Nucleus LS M N Lateral Nucleus of Thalamus LT M 0 Medial Forebrain Bundle MFB M 0 - 115 TABLE VII (continued) Brain Structure Symbol Background Cell Lateral Mammillary Nucleus ML H none Medial Mammillary Nucleus MM M 0 Medial Septal Nucleus MS M none Nucleus Accumbens NA M none Cochlear Nucleus NC M none Dentate Nucleus ND M 0 Fastigial Nucleus NF M 0,N? Interpositus Nucleus of NI M 0 Cerebellum Prepositus Nucleus NPH M 0 Posterior Nucleus of Thalamus NPT M 0 Red Nucleus NR L 0 Nucleus of Spinal Tract of NTST M none the Trigeminal Nerve Lateral Vestibular Nucleus NVL M 0 Medial Vestibular Nucleus NVM M 0 Superior Vestibular Nucleus Spinal Vestibular Nucleus Inferior Olivary Nucleus NVS M none NVSP M none OL M 0 Optic Tract OT none 0 Pons P H none Posterior Hypothalamus PH L none Pretectal Area PRT M 0 Lateral Preoptic Area POA H 0 Periventricular Grey PVG L 0 Paraventricular Hypothalamus PVH M none Reticular Formation RF none 0 Rhomboid Nucleus of Thalamus RH M N Reticular Nucleus of Thalamus RT M 0, N? Suprachiasmic Nucleus SC M N? Stria Medullaris Thalami SM H none Substantia Nigra SN H 0 Supraoptic Nucleus of the SO M N? Hypothalamus Solitary Nucleus SOL M none Superior Olivary Complex SOC M 0 Thalamus (all other areas) T M none Intermediate Olfactory Tract TO I none A? Nucleus Triangularis Septi TS H none Olfactory Tubercle TUO M 0 Ventral Nucleus of Thalamus VE M 0 Vertromedial Hypothalamus VMH M none Ventral Tegmental Nucleus VTN M none Zona Incerta AI none 0 Olfactory Nerve I H 0, N?, A? Vermian Lobule Cl M none Nucleus of the Third Nerve III L 0? Facial Nerve VII M 0 Hypoglossal Nucleus XII M none - 116 -EXPERIMENT 2 1-Pyrroline dehydrogenase (EC; PDH) has been shown in several bacterial and mammalian systems to be a key enzyme in the pathways from ornithine and proline to glutamate (Figure 5). Ornithine is converted to glutamic acid semialdehyde by ornithine ^-transaminase (Ornithine - oxo-acid aminotransferase, EC, OrnT) and the semialdehyde is in equilibrium with P5C which can also be formed from proline by the action of PrO. The P5C is oxidized by PDH to glutamate (Roberts, 1982). Glutamate is an important putative neurotransmitter in its own right and is also an immediate precursor of GABA. Most brain glutamate is formed from glucose through the tricarboxylic acid cycle but the route from ornithine or proline offers a possible alternative for a small glutamate pool. Proline is also a possible neurotransmitter which has been shown, when injected, to end up as GABA in glial cells (Van den Berg, 1970). One of the enzymes, PDH, has been purified from beef liver and is a mitochondrial enzyme that requires nicotinamide adenine dinucleotide (NAD) (Strecker, 1971). The other, PrO, has moderate activity in brain (Kawabata et al., 1980); although not fully characterized, it appears to be a membrane bound enzyme which also uses NAD (Boggess et al., 1978). Since both enzymes function in the presence of NAD, we thought they might be histochemically localized by variations of the technique developed for GABA-T by Van Gelder (1965). Van Gelder used the NADH produced during the metabolism of GABA to - 117 -reduce nitro blue tetrazolium to the dye formazan which stayed in the cells containing the GABA-T. Modifications of this technique with P5C or L-proline as a substrate were tried as a means of demonstrating the histochemical localization of PDH and PrO in brain. METHOD P5C was prepared from its precursor (supplied by Calbiochem of La Jolla, California) according to the manufacturer's directions: 1 gm of the precursor is dissolved in 33 ml of 6N HC1 and brought to 100 C for 45 min. The P5C was purified on a 150 x 3 0ml column of Dowex 50, 8% crosslinked, mesh 50-100 H+, using the procedures of Strecker (1960). The P5C was eluted and a portion of each fraction was analyzed for P5C by reaction with some o-aminobenzaldehyde and measurement of the absorbance at 440 nm. The samples showing presence of P5C were combined and lyophilized. Male Wistar rats weighing 250-350 gm obtained from Canadian Breeding Laboratories were perfused intracardially with 150 ml of ice cold 0.1M phosphate buffered saline (pH 7.4) followed by 2% gluteraldehyde/2% paraformaldehyde. Sections were cut on a vibratome and collected in 0.1M phosphate buffer. The free floating sections were stained for PDH by preincubating them for 2 0 min. at 37°C in the dark in 5 ml tris hydrogen-chloride 0.1M (pH 8.6), plus 0.5 ml NAD+ (10 mg/ml), plus 1.5 ml of solution containing 144 mg/ml NaCl, 2 0 mg/ml MgCl and 1 mg/ml KCN. After the preincubation, 10 mg of nitro blue tetrazolium were mixed in 0.25 ml dimethyl - 118 -sulfoxide which was added with 0.25 ml distilled water, followed by 0.2 ml of phenazine methosulfate (2 mg/ml) and 0.1-0.5 ml of 150 mg/ml of P5C. The incubation was continued at 37°C in the dark for 45 min. The reaction was stopped by transfer to a phosphate buffer. The sections were mounted on gelatin coated slides, dried at least overnight, dehydrated in xylene and coverslipped in Permamount. All solutions were in distilled water unless otherwise specified. The procedure for the L-proline oxidase staining was identical except that 0.1-0.5 ml of 250 mg/ml of commercially available L-proline was substituted from the P5C. Controls were done without L-proline or P5C. RESULTS All concentrations of P5C gave some background staining, but there was a much darker specific staining of certain types of cells. This was most evident in the cerebellum where a high proportion of Bergmann type astrocytes were darkly and distinctly stained. Although there was a high background level in the granular layer of the cerebellum, no cellular morphology was evident in that layer and the staining density never approached half of that in the Bergmann glial cells. Figure 6a shows the stained Bergmann glial cell bodies in the Purkinje cell layer and their fiber-like projections into the molecular layer. Figure 5b shows this staining is consistent throughout the cerebellar sections. The next most consistent and clear finding was in the pyramidal cell layer of the dentate gyrus of the hippocampus. Here the staining was light - 119 -but a distinct band of stained hippocampal astrocytes could be distinguished (Fig. 7). The only other clearly stained cells were occasional astrocyte-like cells of the corpus callosum and other prominent white tracts. PrO staining was much less distinct and was limited to the Bergmann glial cells (Fig. 8). Sections stained without either L-proline or P5C showed no cellular staining and only a faint pink background staining. DISCUSSION Our modifications of the Van Gelder technique for the histochemistry of GABA transaminase gave some indication of the probable localization of PDH and PrO. In both cases there was some non-specific background staining, but in neither case was it high enough to interfere with microscopic interpretation of specific cell staining. The technique could probably be used similarly for the histochemical localization of other NAD requiring enzymes. It may not, however, reveal all loci of such enzymes and is probably not suitable for quantitative analysis. Thus, for example, no cellular staining for PDH was seen in the cortex although appreciable, if relatively low, activity was found in that region on biochemical assay (Thompson et al., 1985). It is possible that the technique only gives clear staining of cells containing PDH at activities that approach the levels in the cerebellar Bergmann glial cells. Increasing the substrate concentration did not result in specific staining of more cell types. - 120 -The staining for proline was less distinct than that for PDH and may, in fact, be due to PDH since the product of PrO is P5C which could be acted on by PDH to produce glutamate and further NADH (Fig. 5). This possibility gains some support from the fact that the only definite cell staining for PrO was seen in the cerebellum although regional distribution data for PrO suggest the highest activities are in the midbrain and brain stem (Thompson et al., 1985). The most consistent histochemical finding is that only certain glial cell populations were stained. It can be argued that Bergmann glial cells are a special type of glial cell but the astrocytes of the hippocampus, although morphologically different to astrocytes in the rest of the brain, are not recognized as a specific subtype. This finding thus supports a growing body of evidence of glial cell specialization. It is tempting to speculate that the chemical specialization of glia is to provide materials important to the neurons in the vicinity. It is true that the densest staining (of Bergmann cells and of astrocytes in the hippocampal dentate gyrus) is in regions where important glutamate tracts as well as GABA interneurons are expected. On the other hand, there is considerable evidence that many corticofugal and cortical commissural tracts also use glutamate as a transmitter and that many cortical interneurons are GABAergic, but astrocytes staining for PDH or PrO were not seen in the cortical grey matter. Our research may be pertinent to the study of a number of familial conditions is which the peripheral metabolism of - 121 -proline or ornithine is known to be affected. In the hyperprolinemias there is a deficiency of PrO in type 1 (Haysaka et al., 1982) and of PDH in type II (Valle et al., 1974). A variety of neurological symptoms, including EEG abnormalities, convulsions, and mental deficiency, have been reported in such cases but the fact that many are asymptomatic suggests there is not a causal relationship (Molicca and Pavone, 1976). Nevertheless, it would be of interest to study brain region levels of these enzymes in post mortem tissue from such cases as compared with controls. The same might be true of cases of hyperornithinemia which have been reported to show atrophy of ocular tissue (Haysaka et al., 1982); since OrnT levels are normally much higher in the retina than in brain (Rao and Cotlier, 1984), the former may show the most drastic changes if this enzyme is deficient. - 122 -Fig. 5 Schematic representation of the conversions of proline and ornithine to glutamate and GABA. - 123 -H2C H2C CH-CH-COOH N H 2 H 2 N ~ C H 2 C H 2_ C H 2 CH ^QQQ|_J ORNITHINE H PROLIN / NAD PROLINE OXIDASE H,C-HC NADH -CH2 CH-COOH ORNITHINE ^-TRANSAMINASE ~ HC-CH2-CH2-CH-COOH NH-PYRROL! NE - 5- CARBOXYLATE GLUTAMIC ACID SEMIALDEHYDE / NA° I - PYRROLINE DEHYDROGENASE NADH HOOC-CH.-CH,-CH-COOH NH- G AD GLUTAMIC ACID HOOC-CH2-CH2-CH2NH2 GABA - 124 -Fig. 6 PDH staining in cerebellum (A) Bergmann glial cells showing fibers projecting up to the cerebellar molecular layer. Cell bodies are loosely arranged around the Purkinje (P) cell layer. Calibration bar = 100 Mm. (B) Same at lower magnification. Calibration bar = 300 jum. Fig. 7 PDH stained astrocytes in layer of dentate gyrus of hippocampus. Calibration bar = 300 Min. Fig. 8 PrO staining of Bergmann glial cells of cerebellum. Calibration bar = 100 pirn. - 125 -EXPERIMENT 3 Thiamine deficiency leading to Wernicke-Korsakoff's syndrome occurs among several populations of Western people, most commonly among alcoholics, but also in people on dialysis, people with intestinal absorption diseases (Sassaris et al., 1983), and the elderly (Iber et al., 1982). Thiamine deficiency (TD) can also lead to beriberi, a polyneuritis which can occur with congestive heart failure. Werniche's encephalopathy is a neurological disorder with symptoms of confusion, disturbances in ocular motility, pupillary alterations, nystagmus, and ataxia with tremors. Its symptoms are believed to be the direct result of a biochemical lesion which can largely be reversed by thiamine administration. Korsakoff's syndrome is characterized by impaired memory for recent events and polyneuritis. It occurs frequently with Wernicke's but does not reverse with thiamine therapy. Its thiamine resistant symptoms may be the result of structural damage because of repeated or long term thiamine deficiency. In humans the structural damage of Korsakoff's syndrome occurs as hemorrhagic lesions in the mammillary bodies, periventricular regions of the thalamus and hypothalamus, periaquaductal regions of the midbrain and floor of the fourth vertricle and in parts of the cerebellum. Wernicke's pathology is in similar structures if it is present. We used the pyrithiamine animal model in which rats are put on a thiamine deficient diet and given pyrithiamine (PT), - 127 -an antagonist of thiamine phosphokinase, the enzyme which converts thiamine to thiamine pyrophosphate. Using this model, symptoms of weight lost, ataxia, and loss of righting reflex occur in about 10 days and death in 14 days. PT produces lesions in the lateral vestibular nucleus, floor of the fourth ventricle, mammillary bodies, thalamus, inferior olive, and cerebellum; these are thus similar but not identical to the human patterns seen in Wernicke-Korsakoff's syndrome. Understanding the nature of the early biochemical lesions has been the goal of much research since 1936, when Peters proposed the biochemical lesion theory to explain the neurological effects of thiamine deficiency. Peters' original theory was that the biochemical lesion when found must explain two observations, the selective vulnerability of certain structures in the brain, and the reversibility upon treatment with thiamine. Explaining these observations remains important in current research. The enzymes for which thiamine triphosphate (TTP) is a co-enzyme, as well as several enzymes asociated with various neurotransmitters, have been examined by previous authors but the results do not explain fully the nature of the initial biochemical lesion. In this experiment we examined the synthetic enzymes for GABA and ACh and the degradative enzyme for GABA for their possible role in the initial biochemical lesion. ACh is one of the neurotransmitters previously studied in thiamine deficiency. A decrease in the TTP-dependent enzyme, - 128 -pyruvate dehydrogenase, which is essential for the production of acetyl-CoA and therefore of ACh, would theoretically lead to reduced synthesis of ACh and therefore reduced concentrations of ACh. Decreased synthesis of ACh has in fact been observed, but, although there were earlier reports (Hamel et al., 1980) of decreased ACh concentrations, most recent reports do not confirm this (Reynolds and Blass, 1975, Vorhees et al., 1977). The difference may lie in the speed at which the brain was fixed and the resultant extent to which the metabolically active pools of ACh are measured (Barclay et al., 1981). Since some of the pools are of little functional value, turnover is thought to be a better index of functional change (Cheney et al., 1977). Decreased turnover of ACh has been observed even in the presence of adequate levels of choline and CAT (Thornber et al., 1980). It is postulated that the decrease in pyruvate dehydrogenase in vivo in thiamine deficient animals is not enough to explain all the reduction in ACh synthesis. The levels of CAT (Bhatgat and Lockett, 1962, Heinrich et al., 1973, Reddy, 1982, Sacchi et al., 1978) and the activities of cholinesterase are reportedly not decreased (Takats et al., 1981). We examined the regional activities of CAT in controls, after the appearance of symptoms of thiamine deficiency, and after recovery to original weight. GABA has been found to be decreased in the whole brain, pons/medulla, midbrain, cortex and cerebellum prior to neurological symptoms in rats on pyrithiamine (Butterworth et al., 1979, Butterworth, 1982a). These findings have not been - 129 -confirmed by other researchers (Plaitakis et al., 1979). GABA high affinity uptake is not affected in any brain areas (Plaitakis, 1982). We examined both GAD and GABA-T in specific regions of the brain at the peak of the pyrithiamine induced symptoms and after return to original weight on a normal diet. If the lesion is fundamentally biochemical in nature, then there should be at least some recovery when thiamine is returned to the diet. There have been no studies done on either GABA or ACh enzymes to see if these change selectively and if they recover upon reintroduction of thiamine to the diet. If there is a biochemical lesion involving a particular enzyme, and the thiamine deficiency is stopped just before the onset of symptoms, there should be total recovery of enzymatic function; but if the critical time to stop is past, there may be residual damage due to prolonged biochemical disruption of the cell with some consequent cell death. Recovery of biochemical function might be explained in yet another way. If lesions are initiated in the glia, the recoverability of the early lesions may be because glial cells have the capacity to proliferate. The initial anatomical lesion appears to occur first in glial cells in the areas known to be most affected by thiamine deficiency such as the lateral vestibular nucleus (Collins, 1967). These early lesions consist of swelling of both glial cells and the myelin sheath (Robertson et al., 1968) and may involve astrocytes more than other cell types (Watanabe and Kanabe, 1978). Collins and Converse (1970) noted also that - 130 -the Bergmann glial fibers associated with degenerating neurons of the cerebellar molecular layer were the first to accumulate glycogen in thiamine deficiency. Since glia appear to be the first structures to change and since they do not change equally in all areas there may be fundamental differences in thiamine dependence of various glia. This experiment exemplifies a type of research where concepts of glial heterogeneity may be relevant to the interpretation of the data. METHOD Male Wister rats from Canadian Breeding Farms, weighing 300+12 gms, were given free access to water and commercially available thiamine deficient diet from Nutritional Biochemicals and were injected intraperitoneally with 0.5 mg/kg of pyrithiamine daily. The rats were housed individually in rooms with other rodents on a 12 hour on, 12 hour off light schedule. All rats were weighed daily and gross behavioural changes were noted. When rats exhibited signs of ataxia and loss of righting reflex, usually on day 10 or 11, they were either sacrificed for immediate use or were put on to a normal diet and given a few shots of thiamine hydrochloride (0.5 mg/kg intraperitoneally). These rats were kept until they had reattained their original weights, upon which time they had regained their righting reflex and had lost most of their ataxia. This was usually within two weeks. Rats for biochemical studies were sacrificed by cervical fracture. The brains were immediately removed and dissected - 131 -into eight regions: cerebellum, pons/medulla, neostriatum, midbrain, hypothalamus, thalamus, hippocampus, and cortex. Each tissue sample was homogenized in 0.3 ml or 10 volumes (whichever was less) of cold 0.25 M sucrose. Portions of the homogenate were used for determination of either CAT or GAD by methods described below. CAT was measured by a modification of the method of F. Fonnum (1969). 60-180 mg of tissue was activated by treatment with Triton X-100 and then incubated with acetyl-coenzyme A labelled with [C14]. The [14C] acetylcholine was absorbed onto an ion exchange column (IG50) and eluted with 3 ml of 4N acetic acid. Radioactivity in the eluant was counted. GAD activity was determined by a modification of the method of Lupien et al. (1968). L-(1-[14C]}-Glutamic acid is incubated with tissue homogenates in the presence of pyrrdoxal phosphase, and the [14C02] produced is trapped on hyamine hydroxide soaked paper, and the radioactivity counted. Separate rats, sacrificed by perfusion under deep barbituate anesthesia, were used for the GABA-T histochemistry which was done by a method a Van Gelder (1965) modified as follows. Rats anaethesized with sodium pentabarbital and perfused intracardially with 150 ml ice cold 0.1M phosphate buffer, pH 7.4, had their brains removed, kept in 0.1M phosphate buffer, sectioned at 50 /Um on an Oxford Vibratome and stained for GABA-T by preincubating free floating sections in the dark for 2 0 min. in a reaction mixture containing 5.0 ml tris HCl 0.1M, 0.2 ml of 250 mg/ml alpha-ketoglutarate, 1.5 ml of a solution containing 144 mg/ml NaCl, 2 0 mg/ml MgCl2, - 132 -and 1 mg/ml KCN, and 0.5ml 10 mg/ml NAD at pH 8.6. After the preincubation, 10 mg of nitro blue tetrazolium dissolved in 2.5 ml dimethyl sulfoxide and 2.5 ml water, 9.5 ml of lmg/5ml phenazine methosulfate and 0.2 ml of 250 mg/ml GABA are added to the pre-incubation medium. The sections are incubated for 45 min. at 37°C. The reaction is stopped with the transfer of these sections to 0.1M phosphate buffer. The sections are mounted on gelatin coated slides, air dried, dehydrated in xylene, and coverslipped with Permamount. RESULTS In symptomatic thiamine deficient (TD), pyrithiamine .treated rats GAD activities were found to be significantly decreased in four areas of the brain: the thalamus > cerebellum > pons/medulla > midbrain (see Table VIIIA). After body weight had returned to pre-experimental levels, there was significant recovery of GAD activity except in the thalamus. GABA-T staining was most dramatically reduced in the thalamus (see Fig. 9a, 9b), and next in the inferior colliculus. There was some loss in the pons and medulla, but no change in other areas of the brain including the cerebellum. After return to a normal diet, there is at least partial recovery of staining in all areas affected (see Fig. 9b, 9c). There was no significant change in CAT activity in any brain area (see Table VIIIB). - 133 -DISCUSSION Our finding of a specific loss of the two GABA related enzymes, GAD and GABA-T, in several brain areas is compatible with the findings of several other workers. It is compatible with the reduction of GABA in whole brain (Gaitonde, 1975, Gubler et al., 1974) in both PT and TD rats, and with the findings of reduced GABA concentrations in PT rats in the cerebellum (Butterworth et al., 1978, Butterworth, 1982a, Butterworth, 1982b), medulla/pons (Butterworth et al., 1978), and midbrain (Butterworth, 1982b). We did not, however, observe a decrease in GABA-T or GAD in cerebral cortex as Butterworth (1982b) did. Our findings are not compatible with those of Plaitakis et al. (1979) who found no change in cerebellum or pons/medulla in pyrithiamine treated rats. The thalamic changes in GABA we observed had not been reported elsewhere, but the periventricular region of the thalamus, where there is a high density of presumptive GABAergic neurons (Nagai et al., 1983), is an area which, like the cerebellum and midbrain, have notable histopathology in Korsakoff's syndrome. The fact that GAD remained reduced in the thalamus of rats put on a normal diet with thiamine supplementation suggests some structural damage to GABAergic systems in this area. It is interesting that the thalamus, which has the largest effects of thiamine deficiency on GAD and GABA-T, is also the region showing the largest drops in GAD and GABA on aging (McGeer and McGeer, 1982). The intracytoplasmic inclusions - 134 -found in the thalamus in thiamine-deficient mice have been said to be morphologically indistinguishable from those in aged mice (Aikawa et al., 1983). TABLE VIII: Enzyme Levels in Control, Thiamine-Deficient and Recovered Rats ( moles/hr-100 mg protein; Mean+S.D.; number rats in parentheses). Brain Area Controls (7) Thiamine Def.(7) Recovered (5) A. Glutamic Acid Decarboxylase Cerebellum 19. 39 + 1. 78 12. 14 + 1. 65# 17. 64 + 3. 25 Pons/Medulla 13. 24 + 1. 21 11. 38 + 0. 60** 12 . 92 + 1. 83 Neostriatum 15. 11 + 1. 06 12. 12 + 5. 60 14. 98 + 1. 24 Midbrain 19. 46 + 0. 95 13. 11 + 5. 21* 17. 74 + 3. 02 Hypothalamus 16. 28 + 1. 52 12. 13 + 3. 33 14. 26 + 2. 98 Thalamus 21. 34 + 1. 88 10. 02 + 2. 30# 14. 91 + 2. 99* Hippocampus 13 . 56 + 2. 15 13. 95 + 0. 97 13. 47 + 0. 85 Cortex 15. 56 + 2. 94 12 . 67 + 4. 29 13. 92 + 2. 84 Choline Acetyltransferase Cerebellum 1. 77 + 0. 26 1. 50 + 0. 22 1. 63 + 0. 24 Pons/Medulla 17. 90 + 2. 13 20. 32 + 2. 15 19. 74 + 1. 93 Neostriatum 33 . 06 + 5. 49 26. 93 + 2. 76 34. 32 + 3. 27 Midbrain 12. 08 + 1. 21 11. 54 + 1. 40 12. 09 + 1. 02 Hypothalamus 6. 19 + 0. 48 6. 93 + 0. 64 6. 37 + 0. 35 Thalamus 10. 54 + 1. 59 11. 08 + 1. 69 10. 91 + 0. 99 Hippocampus 11. 21 + 1. 34 10. 12 + 1. 35 11. 37 + 0. 85 Cortex 12. 76 + 3. 65 10. 90 + 2. 05 12 . 93 + 2. 64 #p<0.001, **p<0.001, *p<0.02 for comparison with controls. - 135 -The lack of change in CAT during thiamine deficiency is consistent with previous reports (Bhatgat and Lockett, 1962, Heinrich et al., 1973, Reddy, 1982, Sacchi et al., 1978, Thornber et al., 1980). This, combined with the fact that AChE is also unaffected in thiamine deficiency (Gibson et al., 1982, Takata et al., 1981) and that a decreased turnover of acetylcholine is observed even in animals showing normal levels of CAT (Sacchi et al., 1978, Thornber et al., 1980), is consistent with the belief that the amount of enzyme is not normally rate controlling and that factors such as decreased availability of acetyl coenzyme A (Vorhees et al., 1978) or an inhibitory effect of thiamine deficiency on acetylcholine release (Dunant and Eder, 1983, Eder et al., 1976) may play important roles. A hypothesis as to the mechanism of the losses in GAD and GABA-T must take into account the regional specificity observed. A specific loss of GAD is assumed to be because of destruction of GABAergic synaptosomes (Butterworth, 1982a) and perhaps of the GABA neurons themselves. If it is assumed that only neurons are involved, it is hard to see why GABA neurons are not destroyed equally in all areas. Why for instance is GAD not significantly reduced in the neostriatum or hippocampus where there are high concentrations of GABA neurons or interneurons? It has been suggested that the most affected areas are those with high turnover rates of thiamine and high oxidative metabolism which is dependent, for at least one step, on thiamine as a cofactor (Dreyfus, 1976). The - 136 -cerebellum is one such area (Ritchie et al., 1980, 1984). Decreased activity of pyruvate dehydrogenase, which is dependent on thiamine triphosphate as a coenzyme, would, for example, lead to decreased incorporation of glucose into amino acids and keto acids of both the TCA cycle and GABA shunt (Butterworth et al., 1978, Butterworth, 1982a). GAD activity, however, is not known to be affected by precursor availability. Another hypothesis involving only neurons is some interneuronal reaction. For example, the changes in cerebellar GAD might be secondary to changes in 5HT system which may innervate GABA neurons. Such a situation has been suggested (Chan-Palay et al., 1977) in the loss of serotonergic mossy fibers in contact with the cerebellar Purkinje cells which are GABAergic (Chan-Palay et al., 1977, Onodera et al., 1981, Plaitakis et al., 1978a, Plaitakis et al., 1978b, Plaitakis et al., 1979). Thus serotonergic changes might lead to GABA changes and GAD changes and these serotonergic neurons are known to be susceptible to thiamine deficiency (Plaitakis et al., 1978a). This hypothesis could presumably be tested by examining GAD levels in rats where lesions of the serotonergic neurons have been produced by other means such as 5,7-dihydroxytryttamine. The recovery of GAD in our results may indicate: biochemical reversal of changes which reduced the activity of GAD, regrowth of GABAergic synaptosomes containing GAD, or, if GAD is actually sensitive to precursor availability, the recovery of precursors. GABA-T loss and recovery could be - 137 -indicative of loss and recovery of synaptosomes which contain GABA-T to regulate GABA levels presynaptically. The regional specificity and partial reversibility of the changes in GABAergic systems analyzed under this scenario would indicate a possible role of thiamine in either GABA neurons or afferents to such neurons; there is no good explanation why all GABA neurons are not affected. If glial cell heterogeneity is taken into account and combined with the observations that the first changes that occur in the thiamine deficiency models are in glial cells, then the interpretation of our observations could be quite different. The recovery upon return of thiamine, a part of the definition of a biochemical lesion as defined by Peters, could depend upon proliferation of the remaining glial cells in the damaged areas, or of glial cells from surrounding areas to restore normal glial factors needed to support the GABAergic neurons. Since GABA-T is in glia as well as neurons, glial proliferation might help to explain the recoveries in GABA-T. The second part of Peters' definition of the biochemical nature of thiamine deficiency, the selective vulnerability of certain regions, may not be due to regional differences in neurons but be due to glial heterogeneity. It has been shown that only glial cells of certain areas show early thiamine deficiency changes. All the areas where GABA-T loss and recovery were noted are areas where glial cell damage occurs early; an anomaly is the cerebellum where there are early glial changes but no GABA-T losses. This area is also the - 138 -only one in which GAD losses and recovery do not seem to parallel GABA-T changes. This may be because any loss of neuronal GABA-T is concealed by GABA-T activity in the Bergmann glia which seem to contain unusually high concentrations of this enzyme which is found in both GABAergic neurons and glia (Nagai et al., 1983). In the cerebellum it is the glial cells of the molecular layer that are the first to change. Therefore the initial biochemical lesion may be in sub-types of glial cells leading to regional glial cell loss which in turn causes changes in neurons of the surrounding area. As reviewed in the main body of this thesis certain types of glia appear to have the ability to take up and metabolize glutamate and to form the glutamine required as a GABA precursor by GABAergic neurons. Changes in these symbiotic glia might lead to changes in the activity of GABA neurons. Butterworth (1982a) suggests that the types of glial cells may be important in determining the selective vulnerability of certain areas and notes that glial cell lines are more susceptible to thiamine deficiency than are neuronal lines. Thiamine pyrophosphatase activity was found to be very high in the plasma membrane of microglia, and oligodendrocytes and astrocytes also had significant staining in the Golgi apparatus (Murabe and Sano, 1981) so an association between thiamine and glia has been made. Other people have also suggested key roles for glia in thiamine deficiency. Butterworth (1982a) suggested that - 139 -selective changes in glial cell integrity may explain GABA changes in the lateral vestibular nucleus. He also postulated that the observed enhanced glutamate uptake in early thiamine deficiency may be explained by the proliferation of glial cells that occurs in damaged areas. In conclusion, if glial cell heterogeneity is assumed, the GABA enzyme changes we have observed may be the direct or indirect result of the early changes in a subtype of glial cells. This then serves to illustrate an example of the types or research where concepts of glial heterogeneity may be relevant to the interpretation of the results. - 140 -Figure 9. Sagittal sections of rat brains (at 2.5 mm from midline) stained for GABA-T. A, Control; B, Thiamine-deficient; C, Recovered, th, thalamus; p, pons; ic, inferior colliculus. - 141 -- 142 -CONCLUSION I have in this thesis reviewed the data on many morphologically defined types of glia or glial-like cells in the brain. These cell types have variable marker staining, vary biochemically, have different develpment profiles, and respond differently to different culture condition and to injury. Culture work shows even more variability. There are differences not only between cell lines but between primary cultures from different areas of the brain in cells that are morphologically similar. My experiments have added to this picture. Experiment 1 showed that glia can stain for iron with a distinct regional pattern of density and types of cell staining. This is just one more example of regional heterogeneity. Experiment 2 showed PDH, an enzyme only recently known to exist, can be stained for in a selected few glial cells. This would theoretically indicate that an alternate route of glutamate synthesis exists in these few selected glial cells. Experiment 3 illustrates how assumptions on the existence of glial heterogeneity may shed a different light on the interpretation of research data. There remains much research to be done on glial heterogeneity. I foresee that it is highly probable that a complimentary map of specific glia functions will be created with a complexity that may approach that now emerging for neurons. - 143 -ACKNOWLEDGEMENTS I would like to thank all the members of the U.B.C. Division of Neurological Research who were all very helpful, especially my advisor Dr. Edie McGeer whose warmth and generosity meant a tremendous amount to me. I would also like to thank the Huntington's Disease Society and the graduate student summer fund which supported me financially, and my husband who typed the document. The work on the iron experiment was supported by the Medical Research Council of Canada. Dr. Y. Noda, a scientist from the Chugai Research Laboratories, Tokyo, Japan, greatly assisted me with the iron research, and was the co-author of a paper submitted to J.Neurochem. that came out of this work. I would like to thank Dr. T.W. McBride for the use of his microscope. The pyrroline dehydrogenase experiments would not have been done without the original suggestion from Peter Wong, who also collaborated with a paper published in J.Neurochem. I was supported by the Garfield-Western Foundation and M.R.C. of Canada in this work. The thiamine experiment was supported by the M.R.C. of Canada and required the technical assistance of Mrs. Kim Singh. The data contained in the portion of my thesis formed the basis for an abstract for the International Association of Neurochemists meeting in Vancouver in 1983 which was published by the Journal of Neurochemistry Vol. 41 supplement, and a paper published in The Neurochemical Research. - 144 -REFERENCES Adams E. and Frank L. (1980) Metabolism of proline and hydroxyprolines. Ann. Rev. Biochem. 49, 1005-1061. Agrawal H.C. and Hartman B.K. (1979) Immunohistochemical localization of myelin proteolipid protein to oligodendrocytes in developing brain. Trans Amer. Soc. Neurochem. 87 (abstract). Aikawa H., Suzuki K. and Iwasaki Y. (1983) Ultrastructural observations on the thalamic neuronal inclusions in young mice. Acta Neuropath. (Berl) 59, 316-318. Akers R.M. (1977) Radial fibers and astrocyte development in the rat cerebral cortex. Anat. Rec. 187, 520-521. Alliot F. and Pessac B. (1984) Astrocytic cell clones derived from established cultures of 8-day postnatal mouse cerebella. Brain Res. 306, 283-291. Anders J.J. and Brightman M.W. (1979) Assemblies of particles in the membrane of developing, mature and reactive astrocytes. J. Neurocytol. 8, 777-795. Baetge E.E., Bulloch K. and Stallcup W.B. (1979) A comparison of glutamate transport in cloned cell lines from the central nervous system. Brain Res. 167, 210-214. Balcar V.J., Borg J. and Mandel P. (1977) High affinity uptake of L-glutamate and L-aspartate by glial cells. J. Neurochem. 28, 87-93. Balcar V.J., Mark J., Borg J. and Mandel P. (1978) Abstr. Sym. GABA Biochemistry and CNS Functions. Eds. P. Mandel, F.V. DeFeudis and J. Mark. p. 20. Balcar V.J. and Hauser K.L. (1978) Transport of [3H] L-glutamate and [3H] glutamiune by dissociated glial and neuronal cells in primary culture. Proc. Eur. Soc. Neurochem. 1, 498. Barber P.C. and Lindsay R.M. (1982) Schwann cells of the olfactory nerves contain glial fibrillary acidic protein and resemble astrocytes. Neurosci. 7, 3077-3090. Barclay L.L., Gibson G.E. and Blass J.P. (1981) Impairment of behavior and acetylcholine metabolism in thiamine deficiency, J. Pharmarcol. Exp. Ther. 217, 537-543. Bartlett P.F., Noble M.D., Pruss R.M., Raff M.C, Rattray S. and Williams CA. (1981) Rat Neural Antigen-2: A cell surface antigen on astrocytes, ependymal cells, Muller cells, and Lepto-meninges defined by a monoclonal antibody. Brain Res. 204, 339-351. - 145 -Basco E., Woodhams P.L., Hajos F. and Balazs R. (1981) Immunocytochemical demonstration of glial fibrillary acidic protein in mouse tanycytes. Anat. Embryol. 162/ 217-222. Benda P., Lightbody J., Sato G. , Levine L. and Sweet W. (1968) Differentiated rat glial cell strain in tissue culture. Science 161, 370-371. Benda P. (1978) Rodent glial cell lines. In: Dynamic Properties of Glial Cells, Eds. E. Schoffeniels, G. Franck, L. Hertz, and D.B. Tower, Pergamon Press, New York, pp.67-81. Bender A.S. and Hertz L. (1984) Flunmrazepam binding to intact and homogenized astrocytes and neurons in primary culture. J. Neurochem. 43, 1319-1327. Bhat S., Barbarese E. and Pfeiffer S.E. (1981) Requirement for nonoligodendrocyte cell signals for enhanced myelinogenic gene expression in long-term cultures of purified rat oligodendrocytes. Proc. Nat. Acad. Sci. (USA) 78, 1283-1287. Bhatgat P.J.G. and Lockett M.F. (1962) The synthesis of acetylcholine by acetone dried powders from the brains of normal and of thiamine deficient rats, J. Phar. Pharmacol. 14, 37-40. Bignami A., Eng. L.F., Dahl D. and Uyeda CT. (1972) Localization of the glial fibrillary acidic protein in astrocytes by immunoflourescence. Brain Res. 43, 429-435. Bignami A. and Dahl D. (1974) Astrocytic-specific protein and neuroglial differentiation. An immunoflourescences study with antibodies to glial fibrillary acidic protein. J. Comp. Neurol. 153, 27-38. Bignami A. and Dahl D. (1975) Astrocyte-specific protein and radial glial in the cerebral cortex of newborn rat. Nature 252, 55-56. Bigner D.D., Bigner S.H., Ponten J., Westermark B., Mahaley M.S., Rouslahti E., Herschman H., Eng L.F. and Wikstrand C.J. (1981) Heterogeneity of genotypic and phenotypic characteristics of fifteen permanent cell lines derived from human gliomas. J. Neuropath. Exp. Neurol. 40, p.201-229. Bjorklund A., Divac I. and Lindvall O. (1978) Regional distribution of catchecholamines in monkey cerebral cortex. Evidence for a dopaminergic innervation of the primary prefrontal cortex. Neurosci. Lett. 7, 115-119. Boggess S.F., Koeppe D.E. and Stewart G.D. (1978) Oxidation of proline by plant mitochondria. Plant Physiol. 62, 22-25. - 146 -Bowman C.L. and Kimelberg H.K. (1984) Excitatory amino acids directly depolarize rat brain astrocytes in primary culture. Nature (Lond) 311, 656-659. Boycott B.B. and Hopkins J.M. (1981) Microglia in the retina of monkey and other mammals; its distinction from other types of glia and horizontal cells. Neurosci. 6, 679-688. Bronson R.T. and Schoene W.C. (1980) Spontaneous pallido-nigral accumulation of iron pigments and shperoid-1ike structures in Macaque monkeys. J. Neuropathol. Exp. Neurol. 34, 181-196. Bruchner G. and Biesold D. (1981) Histochemistry of glycogen deposition in perinatal rat brain: importance of radial glial cells. J. Neurocytol. 10, 749-757. Bruun A., and Ehinger B. (1974) Untake of certain possible neurotransmitters into retinal neurons of some mammals. Exp. Eye Res. 9, 435-447. Bussow H. (1980) The astrocytes in the retina and the optic nerve head of mammals: A special glia for the ganglionic cell axons. Cell Tissue Res. 206, 367-378. Butterworth R.F., Hamel E. and Barbeau A. (1978) Role of aspartic and glutamic acid in the ataxia produced by thiamine deficiency. Soc. Neurosci. Absts. 4. Butterworth R.F., Hamel E. , Lauderville F. and Barbeau A. (1979) Amino acid changes in thiamine deficient encephalopathy: some implications for the pathogenesis of Fredreich's ataxia. Can. J. Neurol. Sci. 6, 217-222. Butterworth R.F. (1982a) Regional amino acid neurotransmitter distribution in thiamine deficiency. In H.Z. Sable and C.J. Gubler (Eds.) Thiamine: 20 years of progress. N.Y. Acad. Sci., New York, pp.464-465. Butterworth R.F. (1982b) Review - Neurotransmitter function in thiamine deficiency encephalopathy. Neurochem. Internatl. 4, 1-16. Cajal R.S. (1909-1911) Histologie du systeme Nerveux de l'Homme et des Vertetres, 2 vol. Maloine, Paris. Cajal, S.R. (1913) Sobre un nuevo proceder de impregnacion de la neuroglia y sus resultados en los centros nerviosos del hombre y animales. Trab. Lab. Invest. Biol. Univ. Madrid 11, 219-237. Cajal R.S. (1929) Development of the cerebral cortex. In: L. Guth. (transl. 1960), Studies on vertebrate Neurogenesis, Charles C. Thomas, Springfield, Illinois, pp.325-335. - 147 -Caley D.W. and Maxwell D.S. (1968) An electron microscopic study of the neuroglia during postnatal development of the rat cerebrum. J. Comp. Neurol. 133, 45-70. Cambier D., Alliot F. and Pessac B. (1983) A glycine-enriched astrocytic cell clone derived from mouse cerebella transformed by Simian virus-40. J. Neurochem. 40/ 1262-1264. Cambier D. and Pessac B. (1983) Endogenous amino acids in permanent cell clones derived from mouse cerebellum. Abs #P-1. Abstracts of the Neuroscience Soc. Meeting/ 1983. Carter C.J. (1981) Loss of glutamine synthetase activity in the brain in Huntington's disease. Lancet 8223/ 782-783. Cavanagh J.B. and Thompson R.H.S. (1954) Demyelination. Brit. Med. Bull. 10, 47-52. Chan-Palay V., Plaitakis A./ Nicklas W. and Berl S. (1977) Autoradiographic demonstration of loss of labelled indoleamine axons of the cerebellum in chronic diet-induced thiamine deficiency. Brain Res. 138, 380-384. Cheney D.L./ Moroni F., Malthe-Sorenson D. and Costa G. (1977) Endogenous modulations of acetylcholine turnover rate. In D.J. Jenden (Ed), Cholinergic mechanisms and psychopharmacology, Plenum Press, New York/ pp.551-564. Choi B.H. and Lapham L.W. (1980) Evolution of Bergmann glia in developing human fetal cerebellum: a Golgi, electron microscopic and immunoflourescent study. Brain Res. 190/ 369-383. Choi B.H./ Kim R.C. and Lapman L.W. (1983) Do radial glia give rise to both astroglial and oligodendrocytes? Develop. Brain Res. 8, 119-130. Ciesielski-Treska J./ Roth-Schechter B., Beya D., Mandel P., Nissen C. and Hertz L. (1976) Rate of oxygen uptake by the NN cell line of hamster astroglia: Lack of effect by excess potassium. J. Neurochem. 26/ 197-199. Ciesielski-Treska J., Baber M. and Aunis D. (1982a) Microtubular organization of flat epitheloid and stellate process-bearing astrocytes in culture. Neurochem. Res. 7, 275-286. Ciesielski-Treska J., Guerold B. and Aunis D. (1982b) Immunofluorescence study on the organization of actin in astroglial cells in primary cultures. Neurosci. 7, 509-522. Ciesielski-Treska J., Ulrich G./ Mensch C. and Aunis D. (1984) Electrophoretic pattern and distribution of cytoskeletal proteins in flat-epitheloid and stellate process-bearing astrocytes in primary culture. Neurochem. Int. 6, 533-543. - 148 -Clark R. and Perkins (1971) Regulation of cAMP concentration in cultured human astrocytoma cells by catecholamines and histamine. Proc. Nat. Acad. Sci. (USA) 68, 2757-2760. Collins G.H. (1967) Glial cell changes in brainstem in thiamine deficient rat. Am. J. Pathol. 50, 791-814. Collins G.H. and Converse W.K. (1970) Cerebellar degeneration in thiamine deficient rats. Am. J. Pathol. 53, 219-233. Colmant G.W. (1965) Zerebrale Hypoxie. Thieme Verlag, Stuttgart. Contestabile A. and Andersen H. (1978) Methodological aspects of the histochemical localization and activity of some cerebellar dehydrogenase. Histochem. 56, 117-132. Cummins C.J./ Glover R.A. and Sellinger O.Z. (1979) Neuronal clues regulate uptake in cultured astrocytes. Brain Res. 170, 190- 193. Currie D.N. and Kelly J.S. (1981) Glial versus neuronal uptake of glutamate. J. Exp. Biol. 95, 181-193. Dahl D., Rueger D.C. and Bignami A. (1981) Vimentin, the 57,000 molecular weight protein of fibroblast filaments is the major cytoskeletal component in immature glia. Europ. J. Cell Biol. 24, 191- 196. Dahl D. and Bignami A. (1982) Immunohistological localization of desmin, the muscle type 100 A filament protein, in rat astrocytes and Muller glia. J. Histochem. Cytochem. 30, 207-213. Dahl D., Chi N.H., Miles L.E., Nguyen B.T. and Bignami A. (1982) Glial Fibrillary acidic protein in Schwann cells. J. Histochem. Cytochem. 30, 912-918. Denis-Donini S., Glowinsky J. and Prochiantz A. (1984) Glial heterogeneity may define the three-dimensional shape of mouse mesencephalic dopaminergic neurons. Nature (Lond) 307, 641-643. DeVelliss J., Schjeide O.A. and Clemente CD. (1967) Protein sythesis and enzymic patterns in the developing brain following head x-irradiation of new born rats. J. Neurochem. 14, 499-511. DeVellis J. and Brooker G. (1973) Induction of enzymes by glucocorticoids and catecholamines in rat glial cell line. In: Tissue Culture of the nervous system, S. Gato (Ed.) Plenum Press, New York, pp.231-246. DeVellis J., Mcginnis G., Breen G., Leveille P., Bennet K. and McCarthy K. (1978) Hormonal effects on differentiation in neural cultures. In: Cell, Tissue and Organ Culture in Neurobiology (Eds. S. Fredoroff and L. Hertz). New York, Acadmic Press, pp.485-512. - 149 -De Vitry F., Picart R./ Jacque C. and Tixier-Vidal A. (1981) Glial fibrillary acidic protein: a cellular marker of tanycytes in mouse hypothalamus. Dev. Neurosci. 4, 457-460. Dickson J.G./ Flanigan T.P., Kemshead J.T., Doherty P. and Walsh F.S. (1983) Identification of cell-surface antigens present exclusively on a sub-population of astrocytes in human foetal brain cultures. J. Neuroimmunol. 5, 111-123. Diezel P.B. (1954) Iron in the brain: a chemical and histochemical examination. In: Biochemistry of the Developing Nervous System (Whelsoh N., ed.) pp.145-152. Divac I. and Braestrup C. (1978) Distribution of naloxone/ diazepam and QNB binding in the monkey cerebral cortex. Neurosci. Abstr. 4/ 443. Dreyfus P.E. (1976) Thiamine deficient encephalopathy: thoughts on its pathogenesis. In C.J. Gubler, M. Fujiwara, and P.M. Dreyfus (Eds.) Thiamine, John Wiley & Sons/ New York/ pp.229-239. Drejer J./ Larsson O.M. and Schousboe A. (1982) Characterization of L-gutamate uptake into and release from astrocytes and neurons cultured from different brain regions. Exp. Brain Res. 47/ 259-269. Drummond R.J. and Phillips A.T. (1977) Intracellular amino acid content of neuronal, glial/ and non-neuronal cell culture: the relationship to glutamic acid compartmentalization. J. Neurochem. 29/ 101-108. Duffy P.E., Huang Y. and Rapport M.M. (1982) The relationship of glial fibrillary acidic protein to the shape, motility and differentiation of human astrocytoma cells. Exp. Cell Res. 139/ 145-157. Dunant Y. and Eder L. (1983) Functional relationships between thiamine and acetylcholine. J. Neurochem. 41, suppl. s31. Ebendal T. and Jacobson CO. (1975) Human glial cells stimulating outgrowth of axons in cultured chick embryo ganglia. Zoon. 3/ 169-172. Eder L./ Hirt L. and Dunant Y. (1976) Possible involvement of thiamine in acetylcholine release. Nature 264/ 186-188. Edwards E., Bulloch K. and Stallcup W.B. (1979) A comparison of glutamate transport in cloned cell lines from the central nervous system. Brain Res. 167, 210-214. Elkouby, A., Ledig M. and Mandel P. (1982) Effect of hydrocortizone and thyroxine of ATPase activities of neuronal and glial cell lines in culture. Neurochem. Res. 7, 387-397. - 150 -Eng L.F. (1982) The glial fibrillary acidic protein: the major protein constituent of glial filaments. Scand. J. Immunol. 15 (suppl.9), 41-51. Faivere-Bauman A., Rossier J. and Benda P. (1974) Glutamate accumulation by a clone of glial cells. Brain Res. 76, 371-375. Fedoroff A., White R., Neal J., Subrahmanyan L. and Kalnins V.I. (1983) Astrocyte cell lineage II. Mouse fibrous astrocytes and reactive astrocytes in cultures have vimentin- and GFA-containing intermediate filaments. Devol. Brain Res. 7, 303-315. Fedoroff S., Neal J., Opas M. and Kalnins V.I. (1984a) Astrocyte cell lineage. III. The morphology of differentiating mouse astrocytes in colony culture. J. Neurocytol. 13, 1-20. Fedoroff S., Mcauley W.A.J., Houle J.D. and Devon R.M. (1984b) Astrocyte cell lineage. V. Similarity of astrocytes that form in the presence of dBcAMP in culture to reactive astrocytes in vivo. J. Neurosci. Res. 12, 15-27. Fischer G., Leutz A. and Schachner M. (1982) Cultivation of immature astrocytes of mouse cerebellum in a serum free, hormonally defined medium. Appearance of the mature astrocyte phenotype after the addition of serum. Neurosci. Lett. 29, 297-302. Fonnum F. (1969) Radiochemical microassays for determination of choline acetyltransferase and acetylcholine esterase. Biochem. J. 15, 465-472. Franck G., Grisar T., Monen G. and Schoffeniels E. (1978) Potassium transport in mammalian astroglia. In: Dynamic Properties of Glial Cells. Eds., E. Schoffeniels, G. Franck, L. Hertz, and D.B. Tower. Pergamon Press, New York, pp.315-325. Francois C, Nguyen-Legros J. and Percheron G. (1981) Topographical and cytological localization of iron in the rat and monkey brains. Brain Res. 215, 317-322. Friede R.L. (1966) Topographic brain chemistry. Academic press, New York. Fujita S. (1980) Cytogenesis and pathology of neuroglia and microglia. Path. Res. Pract. 168, 271-278. Fujita S., Tsuchinashi Y. and Kitamura T. (1981) Origin, morphology and function of the microglia. In Glial and Neuronal Cell Biology, Alan R. Liss Inc., New York, pp.141-169. Fulop Z., Lakos I., Basco E. and Hajos F. (1979) Identification of early glial elements as the precursors of Bergmann glia: a Golgi analysis of the developing cerebellar cortex. Acta Morphologica Acad. Sci. Hung. 27, 273-280. - 151 -Gaitonde M.K. (1975) Conversion of [ul4C] threonine into 14C-labelled amino acids in brain of thaimine deficient rats. Biochem. J. 150, 28-95. Ghandour M.S., Langley O.K./ Vincendon G. and Gombos G. (1979) Double Labelling immunohistochemical technique provides evidence of the specificity of glial cell markers. J. Histochem. Cytochem. 27/ 1634-1637. Ghandour M.S./ Langley O.K./ Vincendon G./ Gombos F./ Filippi D./ Limozin N./ Dalmasso C. and Laurent G. (1980) Immunochemical and immunohistochemical study of carbonic anhydrase/ II in adult rat cerebellum: a marker for oliogodendrocytes. Neurosci. 5/ 559-572. Gheuens J., deSchutter E., Noppe M. and Lawenthal A. (1984) Identification of several forms of the glial fibrillary acidic protein, or -albumin by aspecific monoclonal antibody. J. Neurochem. 43, 964-970. Gibson G., Barclay L. and Blass J. (1982) The role of the cholinergic system in thiamine deficiency. In H.Z. Sable and C.J. Gubler (Eds.) Thiamine 20 years of progress, New York Academy of Sciences, pp.382-403. Gilman A. and Nirenberg M. (1971) Effect of catecholamines on the cAMP concentrations of clonal satellite cells of neurons. Proc. Nat. Acad. Sci. (USA) 68, 2165-2168. Goldlefter L. (1976) Studies on the structure and function of gomori-positive glial cells in the rat hypothalamus. Acta Anat. 95, 545-557. Goldman J.E. and Chiu F-C. (1984) Growth kinetics, cell shape and the cytoskeleton of primary astrocyte cultures. N. Neurochem. 42, 175-184. Golgi C. (1879) Di unanuova reasione apparentemente nera della cellule nervose cerebrali ottentuta col biochloruro di mercuric Archs. Sci. Med. 3, 1-7. Golgi C. (1894) Untersuchungen uber den feineren Ban des centralen und peripheren Nervensystem. Jena: Fisher. Gotow T. (1984) Cytochemical characteristics of astrocytic plasma membranes specialized with numerous orthagolal arrays. J. Neurocytol. 13, 431-448. Grisar T. (1978) L'astroglie dans les mechanismes de regulation des ions K+ au sein des espaces extracellulaires du systeme nerveux central. Ph.D Thesis, University of Leige, Belgium. - 152 -Grisar T.F.G. and Schoffeniels E. (1978) K+ activation mechanisms of the (Na+/K+) -ATPase of bulk isolated glia and neurons. In: Dynamic Properties of Glial Cells/ Eds./ E. Schoffeniels, G. Franck, L. Hertz, and D.B. Tower. Pergamon Press, New York, pp.359-369. Grossman R.G. (1978) Glial-neural interactions studied with intracellular injections of ions into cortical glia. In: Dynamic Properties of Glial Cells, Eds., E. Schoffeniels, G. Franck, L. Hertz, and D.B. Tower. Pergamon Press, New York, pp.105-113. Gubler C.J., Adams B.L., Hammond B., Yuan E.C., Guo S.M. and Bennion M. (1974) Effect of thiamine deprivation and thiamine antagonists on the level of -aminobutryic acid and on 2-oxoglutarate metabolism in rat brain. J. Neurochem. 22, 831-836. Guldner F.H. and Wolff R.J. (1973) Multilamellar astroglial wrapping of neuronal elements in the hypothalamus of rat. Experientia 29, 1355-1356. Haber B. and Hutchison H.T. (1976) Uptake of neurotransmitters and precursors by clonal cell lines of neuronal origin. In: Transport Phenomena in the Nervous System, Physiological and Pathological Aspects, Advances of Experimental Medicine and Biology, Vol. 69, Eds., G. Levi, L. Battistin and A. Lajtha. Plenum Press, New York, pp.179-198. Hajos F. and Rostomian M.A. (1984) Localization of a-tubulin immunoreactivity to cerebellar Bergman glia with the TU01 antibody. Histochem. 81, 297-299. Hamel E., Butterworth R.F., Jolicoeur F.B. and Barbeau A. (1980) Differences in (i) behavior (ii) brain amino acids distribution and (iii) reversibility by thiamine between thiamine deficient and pyrithiamine treated rats. Soc. Neurosci. Abstracts 6. Hamberger A. (1971) Amino acid uptake in neuronal and glial fractions from rabbit cerebellar cortex. Brain Res. 31, 169-178. Hansson E. (1983) Accumulation of putative amino acid neurotransmitters, monoamines and D-Ala-met-enkephalinamide in primary astroglial cultures from various brain areas, visualization by autoradiography. Brain Res. 289, 189-196. Hansson E. and Ronnbak L. (1983) Protein metabolism in an astroglial primary culture. Neurochem. Res. 8, 375-388. Hansson E. and Sellstrom A. (1983) MAO, COMT, and GABA-T activities in primary astroglial cultures. J. Neurochem. 40, 220-225. Hansson E. (1984) Enzyme activities of monoamine oxidase, cathechol-o-methyltransferase and -aminobutyric acid transaminase in primary astroglial cultures and adult rat brain from different brain regions. Neurochem. Res. 9, 45-57. - 153 -Hansson E., Ronnbak L. and Sellstrom A (1984a) Is there a dopaminergic glial cell. Neurochem. Res. 9, 679-689. Hansson E., Isacsson H. and Sellstrom A. (1984b) Characteristics of dopamine and GABA transport in primary cultures of astroglial cells. Acta Physiol. Scand. 121, 33-341. Hansson H.A., Persson L., Ronnback L. and Haglid K.G. (1976) Immunoelectron microscope study of the distribution of the S-100 protein in brain glial cells. Cytobios 15, 45-48. Hatton M.E., Liem R.K.H. and Mason CA. (1984) Two forms of cerebellar glial cells interact differently with neurons in vitro. J. Cell Biol. 98, 193-204. Hartman B.K., Agrawal H.C, Kalmbach S. and Shearer W.T. (1979) A comparative study of the immunohistochemical localization of basic protein to myelin and oligodendrocytes in rat and chicken brain. J. Comp. Neurol. 188, 273-290. Hayasaka S., Matsuzawa T., Shiono T., Mizuno K. and Ishiguro I. (1982) Enzymes metabolizing ornithine-proline pathway in bovine eye. Exp. Eye Res. 34, 635-638. Hazama H., Ito M., Hirano M. and Uchimura H. (1976) Monoamine Oxidase activities in neuronal and glial fractions from regional areas of the brain. J. Neurochem. 26, 417-419. Heinrich CP., Stadler H. and Weiser H. (1973) The effects of thiamine deficiency on the acetylcoenzyme A and acetylcholine levels in the rat brain. J. Neurochem. 21, 1273-1281. Henn F.A. and Hamberger A. (1971) Glial cell function: uptake of transmitter substances. Proc. Natn. Acad. Sci. (USA) 68, 2686-2690. Henn F.A., Goldstein M.N. and Hamberger A. (1974) Uptake of the neurotransmitter candidate glutamate by glia. Nature (Lond) 249, 663-664. Henn F.A. (1975) Glial transport of amino acid neurotransmitter candidates. In: Metabolic Compartmentation and Neurotransmission. S. Clark, D.D. Clarke and D. Schneider, Eds. Plenum Press, New York, p.91-97. Henn F.A. (1976) Neurotransmission and glial cells: a functional relationship? J. Neurosci. Res. 2, 271-282. Henn F.A. (1980) Separation of neuronal and glial cells and subcellular constituents. In: Advances in Cellular Neurobiology, S. Fredoroff and L. Hertz, Eds. Academic Press, New York, pp.373-403. - 154 -Henn F.A. and Henn S.W. (1980) The psyche-pharmacology of astroglial cells. Prog. Neurobiol. 15/ 1-17. Hertz L. (1977) Biolchemistry of glial cells. In: Cell, Tissue and Organ cultures in Neurobiology. Eds., S. Fedoroff and L. Hertz. Academic Press, New York, pp.39-71. Hertz L. (1978a) Kinetics of adenosine uptake into astrocytes. J. Neurochem. 31, 55-62. Hertz L. (1978b) Energy metabolism of glial cells. In: Dynamic Properties of Glial Cells. Eds., E. Schoffeniels, G. Franck, L. Hertz and D.B. Tower. Pergamon Press, New York, pp.121-131. Hertz L., Bock E. and Schousboe A. (1978a) GFA content/ glutamate uptake and activity of glutamate metabolizing enzymes in differentiating mouse astrocytes in primary cultures. Devel. Neurosci. 1, 226-238. Hertz L., Schousboe A., Boechler N., Mukerji S. and Fedoroff S. (1978b) Kinetic characteristics of the glutamate uptake into normal astrocytes in cultures. Neurochem.. Res. 3, 1-14. Hertz L. (1979) Inhibition by barbituates of an intense net uptake of potassium into astrocytes. Neuropharmacol. 18/ 629-633. Hertz L. , Baldonin F. and Schousboe A. (1979a) Serotonin receptors on astrocytes in primary culture, effects of methysergide and flouxetine. Can. J. Physiol. Pharmacol. 57, 223-226. Hertz L. lyu A., Svenneby G., Kuamme E., Fosmack H. and Schousboe A. (1979b) Absence of preferential glutamine uptake into neurons. An indication of a net transfer of TCA constituents from nerve endings to astrocytes? Neurosci Lett. 16, 103-109. Hertz L. and Mukerji S. (1980) Diazepam receptors on mouse astrocytes in primary culture: displacement by pharmacologically active concentrations of benzodiazepines or barbituates. Can. J. Physiol. Pharmacol. 58, 217-220. Hertz L. (1981) Functional interactions between astrocytes and neurons. In: Glial and neuronal cell biology, Ed., Alan R. Liss, Inc., New York, pp.45-58. Hertz L. (1982) Astrocytes. In: Handbook of neurochemistry, Vol.1 (Lajtha A., Ed.), 2nd Edn. Plenum Press, New York, pp.319-355. Hertz L. and Chaban G. (1982) Indications for an active role of astrocytes in potassium homeostatis at the cellular level: Potassium uptake and metabolic effects of potassium. In: Neuroscience approached through cell culture, pp.158-174. - 155 -Hertz L. and Richardson J.S. (1983) Acute and chronic effects of antidepressant drugs on B-adrenergic function in astrocytes in primary culture: An indication of glial involvement in affective disorders? Neurosci. Res. 9, 173-182. Hertz L. and Richardson J.S. (1984) Is neuropharmacology merely the pharmacology of neurons—or are astrocytes important too? Trends in Pharmacol. Sci. 5, 272-276. Hill J.M. and Switzer III, R.C. (1984) The regional distribution and cellular localization of iron in the rat brain. Neurosci. 11, 595-603. Hill J.M. (1984) Iron concentration reduced in ventral pallidium, Globus pallidus, and substantia nigra by GABA Transaminase inhibitor, gamma-vinyl GABA. Brain Res. His W. (1887) Die Enterwicklung der ersten Nervenbahnen beim menschlichen embryo. Uebersichliche darstellung. Arch Anat. Physiol. (Arch. Abt.) 368-378. His W. (1889) Die Neuroblastasten und deren entstehung im embryonalen mark. Abhandl. Kgl. Sachs. Ges. Wiss. Math. Phys. Cl. 15, 313-372. Hokfelt T. and Ljungdahl A. (1971) Light and electron microscopic autoradiography on spinal cord slices after incubation with labelled glycine. Brain Res. 32, 189-194. Hokfelt T. and Ljungdahl A. (1972) Application of cytochemical techniques to the study of suspected transmitters substances in the nervous system. In: Advances in Biochemical psychophamacology. Eds./ E. Costa and P. Greengard. Raven Press, New York, Vol.6, pp.1-36. Holbrook N.J./ Grasso R.J. and Hackney J.F. (1981) Glucocorticoid receptor properties and glucocorticoid regulation of glutamine synthetase activity in sensitive C-6 and Resistant C6H glial cells. J. Neurosci. Res. 6, 75-88. Hosli E. and Hosli L. (1978) Action and uptake of neurotransmsitters in CNS tissue culture. Rev. Physiol. Biochem. Pharmacology 81, 135-188. Hosli L./ Hosli E./ Schneider U. and Wiget W. (1984) Evidence for the existence of histamine HI and H2-receptors on astrocytes of cultured rat central nervous system. Neurosci. Lett. 48/ 287-291. Hutchison H.T., Werrbach K., Vance C. and Haber B. (1974) Uptake of neurotransmitters by clonal lines of astrocytes and neuroblastoma in cultures. I. transport of -aminobutyric acid. Brain Res. 66/ 265-274. - 156 -Hyden H., Lange P.W. and Larsson S. (1980) SlOO-glia regulation of GABA transport across the nerve cell membrane. J. Neurological Sci. 45, 303-316. Iber F.L./ Blass J.P., Brin M. and Leevy CM. (1982) Thiamine in the elderly. Relation to alcoholism and to neurological degenerative disease. Am. J. Clin. Nutr. 36, 1067-1082. Ibrahim M.Z.M., Khreis Y. and Koshayan D.S. (1974) The histochemical identification of microglia. J. Neurol. Sci. 22, 211-233. Ichikawa M. and Hirata Y. (1982) Morphology and distribution of postnatally generated glial cells in the somatosensory cortex of rat: an autoradiographic and electron microscope study. Dev. Brain Res. 4, 369-377. Ito J., Kato T./ Yamakawa Y./ Kato H./ Sakazaki Y., Lim R. and Tanaka R. (1982) Interaction of glia maturation factor with the glia cell membrane. Brain Res. 243, 309-314. Iverson L.L. and Schon F.E. (1973) The use of autoradiographical techniques for the identification and mapping of transmitter specific neurons in CNS. In: New Concepts in Neurotransmitter Mechanisms, Ed., A.J. Mandell. Plenum Press, New York, pp.153-193. Jurrlinks B.W., Fedoroff S., Hall L. and Nathaniel E.J.H. (1981) Astrocyte cell lineage. 1. Astrocyte progenitor cells in mouse neopallium. J. Comp. Neurol. 200, 375-391. Jesson K.R., Thorpe R. and Mirsky R. (1984) Molecular identify, distribution and heterogeneity of glial fibrillary acidic protein: an immunoblotting and immunohistochemical study of Schwann cells, satellite cells, enteric glia and astrocytes. J. Neurocytol. 13/ 187-200. Kaladhar M. and Narasinga Rao B.S. (1982) Effects of iron deficiency on serotonin uptake in vitro by rat brain synaptic vesicles. J. Neurochem. 38, 1576-1581. Kaladhar M. and Narasinga Rao B.S. (1983) Effects of maternal iron deficiency in rat on serotonin uptake in vitro by brain synaptic vesicles in the offspring. J. Neurochem. 40, 1768-1770. Kato T., Yamakawa Y., Lim R., Turriff D.E. and Tanaka R. (1981) Biological effects of Bovine glia maturation factor on glial cells in culture. Neurochem. Res. 6, 401-412. Kawabata Y.T., Katunuma N. and Sanada Y. (1980) Characteristics of proline oxidases in rat tissues. J. Biochem. 88, 281-283. Keenikova V.A., Pevzner L.Z. and Mularek O. (1979) Postnatal changes in the content of basic and acidic proteins in rat brain astroglia and oligodendroglia. Neurosci. 4, 1187-1193. - 157 -Kelly J.S. and Dick F. (1978) GABA in glial cells of the peripheral and central nervous system. In: Dynamic Properties of Glial Cells. Eds./ E. Schoffeniels, G. Franck, L. Hertz and D.B. Tower. Pergamon Press, New York, pp.183-192. Kemshead J.T., Ritter M.A., Cotmore S.F. and Greaves M.F. (1982) Human Thy-1 expression on the cell surface of neuronal and glial cells. Brain Res. 236, 451-461. Kennedy P.G.E. (1982) Neural cell markers and their applications to neurology. J. Neuroimmunol. 2, 35-53. Kimelberg H.K. (1974) Active potassium transport and (Na+ + K+) ATPase activity in culture glioma and neuroblastoma cells. J. Neurochem. 22, 971-976. Kimelberg H.K., Narumi S. and Bourke R.S. (1978a) Enzymatic and morphological properties of primary rat brain astrocyte cultures, and enzyme development in vivo. Brain Res. 153, 55-77. Kimelberg H.K., Biddlecome S., Narumi S. and Bourke R.S. (1978b) ATPase and carbonic anhydrase activities of bulk-isolated neuron, glia and synaptosome fractions from rat brain. Brain Res. 141, 305- 323. Kimelberg H.K., Bowman C, Biddlecome S., Bourke R.S. (1979) Cation transport and membrane potential properties of primary astoglial cultures from neonatal rat brains. Brain Res. 177, 533-550. Kimelberg H.K., Steig P.E. and Mazurkiewicz J.E. (1982) Immunocytochemical and biochemical analysis of carbonic anhydrase in primary astrocyte cultures from rat brain. J. Neurochem. 39, 734-742. Koenig H. and Barron K.D. (1963) Reactive gliosis-a histochemical study. J. Neuropath. Exp. Neurol. 22, 336-345. Korey S.R., Orchen M. and Brotz M. (1958) Studies of white matter. I. Chemical constitution and respiration of neuroglial and myelin-enriched fractions of white matter. J. Neuropathol. Exp. Neurol. 17, 430-438. Kreutzberg G.W. and Hussain S.T. (1982) Cytochemical heterogeneity of the glial plasma membrane: 51-nucleotidase in retinal Muller cells. J. Neurocytol. 11, 53-64. Krnjevic K. and Schwartz S. (1967) Some properties of unresponsive cells in the cerebral cortex. Exp. Brain Res. 3, 306- 319. Labourdette G. and Marks A. (1975) Synthesis of S-100 protein in monolayer cultures of rat-glial cells. Eur. J. Biochem. 58, 73-79. - 158 -Labourdette G., Roussel G., Ghandour M.S. and Nussbaum J.L. (1979) Cultures from rat brain hemispheres enriched in oligodendrocyte-like cells. Brain Res. 179, 199-203. Lagenaur C, Sommer I. and Schachner M. (1980) Subclass of astroglia in mouse cerebellum recognized by monolonal antibody. Dev. Biol. 79, 367-378. Lagenaur C., Masters C. and Schachner M. (1982a) Changes in expression of glial antigens Ml and Cl after cerebellar injury. J. Neurosci. 2, 470-476. Lagenaur C, Schachner M., Solter D. and Knowles B. (1982b) Monoclonal antibody against SSEA-1 is specific for a subpopulation of astrocytes in mouse cerebellum. Neurosci. Lett. 31, 181-184. Landis D.M. and Reese T.S. (1981) Membrane structure in mammalian astrocytes: a review of freeze-fracture studies on adult/ developing, reactive/ and cultured cells. J. Exp. Biol. 95, 35-48. Langley O.K. and Ghandour M.S. (1981) An immunocytochemical investigation of non-neuronal enolase in cerebellum: a new astrocyte marker. Histochem. J. 13, 137-148. Langley O.K., Ghandour M.S., Vincendon G., Gombos G. and Warecka K. (1982) Immunoelectron microscopy of 2-glycoprotein: an astrocyte-specific antigen. J. Neuroimmunol. 2, 131-143. Larsson O.M., Hertz L. and Schousboe A. (1980) GABA uptake in astrocytes in primary cultures: coupling with two sodium ions. Neurosci. Res. 5, 469-477. Lasansky A. (1971) Nervous function at the cellular level: glia. Ann. Rev. Physiol. 33, 241-256. Lasher R.S. (1975) Uptake of GABA by neuronal and non-neuronal cells in dispersed cell cultures of postnatal rat cerebellum. J. Neurobiol. 6, 597-608. Latzkovits L. (1978) Neuronal-glial interactions in potassium transport. In: Dynamic Properties of Glial Cells. Eds., E. Schoffeniels, G. Franck, L. Hertz and D.B. Tower. Pergamon Press, New York, pp.327-LeCampbell M. and Shank R.P. (1978) Glutamate and GABA uptake by cerebellar granular and glial cell enriched fractions. Brain Res. 153, 618-622. Lenhossek M. von (1891) Zur Kenntnis der Neuroglia des menschlichen Ruckenmarkes. Verh. Anat. Ges. 5, 193-221. - 159 -Lentzen H. and Palenker J. (1983) Localization of the thiorphan-sensitive endopeptidase, termed enkaphalinase A, on glial cells. FEBS Lett. 153/ 93-97. Levi G., Gordon R.D., Gallo V., Wilkin G.P. and Balazs R. (1982) Putative acidicamino acid transmitters in the cerebellum I. Depolarization-induced release. Brain Res. 239, 425-445. Levi G. and Ciotti M.T. (1983) Glutamate and GABA localization and evoked release in cerebellar cells differentiating in culture. In: Glutamine, glutamate and GABA in the Central Nervous System. Eds. L. Hertz, E. Kvanne, E.G. McGeer and A. Schousboe. Alan R. Liss Inc., New York, p.493-508. Levi, G., Wilkin G.P., Ciotti M.T. and Johnstone S. (1983) Enrichment of differentiated, stellate astrocytes in cerebellar interneuron cultures as studied by GFAP immunoflouresence and autoradiographic uptake patterns with H3-D-aspartate and H3-GABA. Brain Res. 312, 227-241. Levitt P. and Rakic P. (1980) Immunoperoxidase localization of glial fibrillary acidic protein in radial glial cells and astrocytes of the developing Rhesus monkey brain. J. Comp. Neurol. 193, 815-840. Lim R., Turriff D., Troy S. and Kato T. (1977) Differentiation of glioblasts under the influence of glia maturation factor. In: Cell, Tissue and Organ Culture in Neurobiology. Eds., Federoff S. and Hertz L. Academic Press, New York, pp.223-236. Lindsay R.M., Barker P.C, Sherwood M.R.C, Zimmer J. and Raisman G. (1982) Astrocyte cultures from adult rat brain. Derivation, characterization and neurotrophic properties of pure astroglial cells from corpus callosum. Brain Res. 243, 329-343. Linser P. and Moscona A.A. (1981) Carbonic Anhydrase C in the neural retina: transition from generalized to glia-specific localization during embryonic development. Proc. Natl. Acad. Sci. (USA) 78, 7190-7194. Logan N. (1976) Amino acid transport by 2 glial cell lines and by proliferating glia. Exp. Neurol. 53, 431-4 Lundgren D.W. and Ogur M. (1973) Inhibition of yeast 1-pyrroline-5-carboxylate dehydrase by common amino acids and the regulation of proline catabolism. Biochem. Biophys. Acta. 292, 246-257. Lugaro E. (1907) Sulle funzioni della nevroglia. Riv. Patol. Nerv. Ment. 12, 225-233. Lupien P.J., Hinse CM. and Berlinguet L. (1968) Determination of glutamic acid decarboxylase activity in rat brain. Anal. Biochem. 24, 1-8. - 160 -Mandel P., Rouseel G., Delaunoy J.P. and Nussbaum J.L. (1978) Wolfgram proteins/ oligodendroglial cell and myelin markers/ carbonic anhydrase/ a glial marker. In: Dynampic Properties of Clial Cells, E. Schoffeniels, L. Hertz and D.B. Tower (Eds.). Pergamon Press, Oxford, pp.267-274. Marinesco M.G. (1896) Lesions des centres nerveux produitespar la toxine de Bacillus botulinus. CR. Soc. Biol. 48, 989-991. Martinex-Hernandex A., Bell K.P., Norenberg M.D. (1977) Glutamine synthetase: glial localization in brain. Science 195, 1356-1358. Massa P.T. and Mugnaini E. (1982) Cell junctions and intramembrane particles of astrocytes and oligodendrocytes: a freeze-fracture study. Neurosci. 7, 523-538. Massa P.T., Szuchet S. and Mugnaini E. (1983) Tight junctions and intramembrane particles of cultured oligodendrocytes. Biophys. J. 41, 69. Matus A.I. and Dennison M.E. (1971) Autoradiographic localization of tritated glycine at 'flat' vesicle synapses in spinal cord. Brain Res. 32, 195-197. McGeer E.G. and McGeer P.L. (1982) Neurotransmitters in normal aging. In: D. Piatt (Ed.) Geriatrics 1, Springer-Verlag, Berlin, Heidelberg, pp.263-282. McLennan H. (1976) The autoradiographical localization of L-[3H]glutamate in rat brain tissue. Brain Res. 115, 139-144. Meldrum B.S., Croucher M.J. and Krogsgaard-Larsen P. (1982) In: Problems in GABA res. from Brain to Bacteria. Eds., Okada Y. and Roberts E. Excerpta Medica, Amsterdam, pp.182-191. Meller K. and Waelsch M. (1984) Cyclic morphological changes of glial cells in long-term culture. J. Neurocytol. 13/ 29-47. Memo M., Riccardi F., Tcabuchi M. and Spano P.F. (1981) Localization of GABA and Dopamine Receptor sites in retinal glial cells using DL-a-aminoacidic acid. In: GABA and Benzodiazapine Receptors, E. Costa, CD. Chiara, G. Gessa, Eds. Raven Press, New York, pp.41-51. Michaelis E.K., Belieu R.M., Grubbs R.D., Michaelis M.L. and Chang H.H. (1982) Differential effects of metal ligands on synaptic membrane glutamate binding uptake systems. Neurochem. Res. 7, 423-436. Miller R.H and Raff M.C (1984) Fibrous and Protoplasmic astrocytes are biochemically and developmentally distinct. J. Neurosci. 4, 585-592. - 161 -Mize R.R./ Spencer R.F. and Sterling P. (1981) Neurons and Glia in cat superior colliculus accumulate (3H) gamma-aminobutyric acid. J. Comp. Neurol. 202, 385-396. Molicca F. and Pavone L. (1976) Hyperprolinemia: a disease that does not need treatment? Acta. Pediat. Scan. 65, 206-208. Moonen G. and Franck G. (1977) Potassium effect on Na+, K+-ATPase activity of cultured new born rat astoblasts during differentiation. Neurosci. Lett. 4, 263-267. Mori S. and Leblond CP. (1970) Electron microscopic identification of three classes of oligodendrocytes and a preliminary study of their proliferative activity in the corpus callosum of young rats. J. Comp. Neurol. 139, 1-30. Morrison R.S. and DeVellis J. (1983) Differentiation of purified astrocytes in a chemically defined medium. Develop. Brain Res. 9, 337-345. Mossakowski M.J. (1962) The activity of succinic dehydrogenase in glial tumors. J. Neuropath Exp. Neurol. 21, 137-146. Mossakowski M.J. and Penar B. (1972) Some aspects of the histochemistry of the reactive glia. Neuropath. Pol. 10, 317-323. Murable Y. and Sano Y. (1981) Morphological studies on neuroglia, I, Electron microscope identification of silver-impregnated glial cells. Cell Tissue Res. 216, 557-568. Murphy D.L., Donelly C.H. and Richelson E. (1976) Substrate and inhibitor related characteristics of monoamine oxidase in C6 rat glial cells. J. Neurochem. 26, 1231-1235. Nagai T., McGeer P.L. and McGeer E.G. (1983) Distribution of GABA-T-intensive neurons in the rat forebrain and midbrain. J. Comp. Neurol. 218, 220-238. Nguyen-Legros J., Bizot J., Bolesse M. and Pulicani J.P. (1980) Noir de Diaminobenzidine ; Une nouvelle methode histochemique de revelation du fer exogene. Histochemistry 66, 239-244. Nicklas W.J. and Browning E.T. (1978) Amino Acid metabolism in glial cells homeostatic regulation of intra- and extra-cellular mileeu by C-6 glioma cells. J. Neurochem. 30, 955-963. Noble M. and Murray K. (1984) Purified astrocytes promote the in vitro division of a bipotential glial progenitor cell. EMBO J. 3, 2243-2247. Norenberg D. and Martinez-Hernandex A. (1979) Fine structure localization of glutamine synthetase in astrocytes of rat brain. Brain Res. 161, 303-310. - 162 -Norenberg D. (1983) Immunohistochemistry of glutamine synthetase. In: Glutamine, Glutamate and GABA in the Central Nervous System, Ed./ Alan R. Liss Inc., New York, pp.95-111. Norton W.T., Abe T., Poduslo S.E. and DeVries G.H. (1975) The lipid composition of isolated brain cells and axons. J. Neurosci. Res. 1, 57-55. Oehmichen N. (1980) Enzyme-histochemical differentiation of neuroglia and microglia: A contribution to the cytogenesis of microglia and globoid cells. Path. Res. Pract. 168/ 344-373. Onodera K., Ogura Y. and Kisara K. (1981) Characteristics of muricide induced by thiamine deficiency and its suppression by antidepressants or intreventricular serotonin. Phys. Behav. 27, 847-859. Osborn M., Ludwig-Festl M., Weber K. , Bignami A./ Dahl D. and Bayreuther K. (1981) Expression of glial and vimentin type intermediate filaments in cultures derived from human glial material. Differentiation 19/ 161-167. Palay S.L., McGee-Russell S.M./ Spencer G.Jr. and Grillo M.A. (1962) Fixation of neural tissue for electron microscopy by perfusion with solutions of osmium tetroxide. J. Cell Biol. 12/ 385-410. Palay S.L. (1966) The role of neuroglia in the organization of the central nervous system. In: Nerve as Tissue. K. Rodahl and Issektz/ Jr./ Eds. Hoeber-Harper/ New York/ pp.3-10. Palay S.L. and Chan-Palay V. (1974) Cerebellar Cortex. Springer-Verlag, Berlin, Heidelberg, New York, p.348. Pellegrino L.J., Pellegrino A.S. and Cushman A.J. (1979) A Stereotaxic Atlas of the rat brain, Second edition. Plenum Press, New York. Passonneau J., Schwartz J. and Lust W. (1978) Some aspects of intermediate metabolism in glioma cells in culture. In: Dynamic Properties of Glia Cells. Eds., E. Schoffeniels, G. Franck, L. Hertz and D.B. Tower. Pergamon Press, Oxford, pp.133-142. Patel A.J. and Balazs R. (1974) Factors affecting the development of metabolic compartmentation in the brain. In: Metabolic Compartmentation and Neurotransmission. Eds., S. Berl, D.D. Clarke and D. Schneider. Plenum Press, New York, pp.363-383. Patterson P.H./ Chun L.L.Y. and Reichardt L.F. (1977) Cellular Neurobiology, Alan R. Liss, New York, pp.94-103. Penfield W. (1932) Neuroglia: normal and pathological. In: Penfield W., Ed., Cytology and Cellular Pathology of the Nervous System, Vol.11, pp.423-479. - 163 -There is no page 16*+. Perkins J., Maclntyre E., Riley W. and Clark R. (1971) Adenyl cyclase/ phosphodiesterase and cAMP dependant protein kinase of malignant glial cells in culture. Life Sci. 10, 1069-1080. Peters A. and Palay S.L. (1965) An electromicroscopic study of the distribution and patterns of astroglial processes in the central nervous system. J. Anat. (Lond) 99, 419. Petito CK. and Babiak T. (1982) Early proliferative changes in astrocytes in postischemic noninfarcted rat brain. Ann. Neurol. 11, 510-518. Pevzner L.Z. (1979) Functional biochemistry of the neuroglia, (translated by Brian Tiplay). Consultants Bureau, New York. Pfeiffer S., Herschman H., Lightbody J. and Sato G. (1977) Synthesis by a clonal line of rat glial cells of a protein unique to the nervous system. J. Cell Physiol. 75, 329-339. Pilgrim C. and Reisert I. (1981) The glial cell as a site of glycoconjugate synthesis in the brain. Prog. Clin. Biol. Res. 19/ 59-64. Pilkington G.J./ Lantos P.L., Darling J.L. and Thomas D.G.T. (1982) Three cell lines from a spontaneous murine astrecytoma show variation in astrocytic differentiation. Neurosci. Lett. 34, 315-320. Pixley S.K. and DeVellis J. (1984) Transition between immature radial glia and mature astrocytes studied with monoclonal antibody to vimentin. Develop. Brain Res. 15, 201-209. Plaitakis A., Nicklas W.J. and Berl S. (1978a) Thiamine deficiency: Selective impairment of the cerebellar serotonergic system. Neurol. 28, 691-698. Plaitakis A., Van Woert M.H., Hwang E.C and Berl S. (1978b) The effect of acute thiamine deficiency on brain tryptophan, serotonin and 5-hydroxyindole actic acid. J. Neurochem. 31, 1087-1089. Plaitakis A., Nicklas W.J. and Berl S. (1979) Alterations in uptake and metabolism of aspartate and glutamate in brain of thiamine deficient animals. Brain Res. 171, 489-502. Plaitakis A., Hwang E.C./ Van Woert M.H., Sziligyl P.I.A. and Berl S. (1982) Effect of thiamine deficiency on brain neurotransmitter systems. In: Thiamine 20 yearss of progress, H.Z. Sable and C.J. Gubler (Eds.) Acad. Sci., New York, pp.367-380. Potter R.L., Yu A.C, Schousboe A. and Hertz L. (1982) Metabolic fate of [U-14C]-labeled glutamate in primary cultures of mouse astrocytes as a function of development. Dev. Neurosci. 5, 278-284. - 165 -Prince D.A., Pedley T.A. and Ransom B.R. (1978) Fluctuations in ion concentration during excitation and seizures. In: Dynamic Properties of Glial Cells. Eds./ E. Schoffeniels/ G. Franck/ L. Hertz and D.B. Tower. Peramon Press, New York/ pp.281-303. Pruss R.M./ Bartlett P.F./ Gavrilovic J./ Lisak R.P. and Rattray S. (1982) Mitogens for glial cells; a comparison of the cultured astrocytes/ oligodendrocytes and Schwann cells. Dev. Brain Res. 2/ 19-35. Raff M.C./ Mirsky R., Fields K.L., Lisak R.P., Dorfman S.H., Silverberg D.H., Gregson N.A./ Leibowitz S. and Kennedy M.C. (1978) Galactocerebroside is a specific cell surface antigenic marker for oligodendrocytes in culture. Nature (Lond) 274/ 813-815. Raff M.C, Abney E.R., Cohen J., Lindsay R. and Noble M. (1983) Two types of astrocytes in cultures of developing rat white matter: differences in morphology, surface gangiosides and growth characteristics. J. Neurosci. 3, 1289-1300. Raff M.C, Williams B.P. and Miller R.H. (1984) The in vitro differentiation of a bipotential glial progenitor cell. The EMBO J. 3, 1857-1864. Rakic P. (1971) Neuro-glia relationships during granule cell migration in developing cerebellar cortex. A Golgi and electron microscopic study in Macacus rhesus. J. Comp. Neurol. 141, 283-312. Rakic P. (1972) Mode of cell migration to the superficial layers of fetal monkey neocortex. Comp. Neurol. 145, 61-81. Ramaharobandro N., Borg J., Mandel P. and Mark J. (1982) Glutamine and glutamate transport in cultured neuronal and glial cells. Brain Res. 244, 113-121. Rao G.N. and Cotlier E. (1984) Ornithine delta-aminotransferase activity in retina and other tissues. Neurochem. Res. 9, 555-562. Reddy T.S. (1982) Effect of thiamine deficiency on 2', 3', cyclic nucleotide 3'-phosphohydrolase activity in rat CNS. Nutr. Reports Internl. 25, 285-287. Reyners H., Gianfelici de Reyners E. and Maisin J.R. (1982) The beta astrocyte: a newly recognized radiosensitive glial cell type in the cerebral cortex. J. Neurocytol. 11/ 967-983. Reynolds S.F. and Blass J.P. (1975) Normal levels of acetyl coenzyme A and of acetylcholine in the brain of thiamine deficient rats. J. Neurochem. 24, 185-189. - 166 -Rindi G./ Patrini C, Coraincioli V. and Reggiani C. (1980) Thiamine content and turnover of some rat nervous regions using label thiamine as a tracer. Brain Res. 181/ 369-380. Rindi G., Comincioli V., Reggiani C. and Patrini C. (1984) Nervous tissue thiamine metabolism in vivo. II. Thiamine and its phosesters dynamics in different brain regions and sciatic nerve of the rat. Brain Res. 293, 329-342. Rio Hortega P. del (1919) El tercer alemento de los centros nerviosos. Bol. Soc. Espan. Biol., 69-120. Rio Hortega del P. (1928) Terecera aportacion al conocimiento morphologicoe interpretacion funcional de la oligodendroglia. Real Soc. Espanola de Historia Natural 14, 4-122. Robertson D.M., Wasan S.M. and Skinner D.B. (1968) Ultrastructural features of early brain stem lesions of thiamine deficient rats. Am. J. Pathol. 52, 1081-1087. Roots B.I. (1981) Comparative study on glial markers. J. Exp. Biol. 95, 167-180. Roussel G./ Delaunoy J.P., Nussbaum J.L. and Mandel P. (1979) Demonstration of a specific localization of carbonic anhydrase C in the glial cells of rat CNS by an immunohistochemical method. Brain Res. 160, 47-55. Roussel G. and Nussbaum J.L. (1981) Comparative localization of Wolfgram Wl and myelin basic proteins in the rat brain during ontogenesis. Histochem J. 13, 1029-1047. Roussel G. and Nussbaum J. (1983) Immunohistochemical study with an anti meylin serum: a marker for all glial cells except "dark" oligodendrocytes. J. Neurocytol. 5, 209-226. Sacchi O., Ladinsky H./ Prigioni I., Consolo S., Peri G. and Perri V. (1978) Acetylcholine turnover in the thiamine depleted superior cervical ganglion of the rat. Brain Res. 151/ 609-614. Saint Marie R.L. and Carlson S.D. (1983) Glial membrane specialization and the compartmentalization of the lamina ganglionaris of the housefly compound eye. J. Neurocytol. 12/ 243-275. Salm A.K./ Hatton G.I. and Nilaver G. (1982) Immunoreactive glial fibrillary acidic protein in pituicytes of the rat neurohypophysis. Brain Res. 236/ 471-476. Sarthy P.V. and Lam D.M.K. (1978) Biochemical studies of isolated glial (Muller) cells form the turtle retina. J. Cell Biol. 78, 675-684. - 167 -Sarthy P.V. (1983) Release of [3H] -aminobutyric acid from glial (Muller) cells of the rat retina: effects of K+, veratridine, and ethylenediamine. J. Neurosci. 3, 2494-2503. Sassaris M., Meka R., Miletello G./ Nance C. and Hunter F.M. (1983) Neuropsychiatric syndromes after gastric partition. Am. J. Gastroenterol. 78, 321-325. Schachner M. (1974) NS-1 (nervous system antigen-1), a glial-cell-specific antigenic component of the surface membrane. Proc. Nat. Acad. Sci. 71, 1795-1799. Schachner M., Hedley-White E.T., Hsu D.W., Schoonmaker G. and Bignami A. (1977) Ultrastructural localization of glial fibrillary acidic protein in mouse cerebellum by immunoperoxidase lebelling. J. Cell Biol. 75, 67-73. Schachner M. (1982) Cell type-specific surface antigens in the mammalian nervous system. J. Neurochem. 39, 1-8. Schaper A. (1897) Die fruhesten diffirenzierrungsvorgange im centralnervensystem. Arch. Entw-Mech. Organ 5, 81-132. Schier B.K. and Thompsom E. (1974) On the role of glial cells in the mammalian nervous system. J. Biol. Chem. 249, 1769-1780. Schmechel D.E. and Rakic P. (1979a) A Golgi study of radial glial cells in developing monkey telencephalon: morphogenesis and transformation into astrocytes. Anat. Embryol. 156, 115-152. Schnitzer J., Pranke W.W. and Schachner M. (1981) Immunocytochemical demonstration of vimentin in astrocytes and ependymal cells of developing and adult mouse nervous system. J. Cell Biol. 90, 435-447. Schnitzer J. and Schachner M. (1981) Expression of thy-1, H-2 and NS-4 cell surface antigens and tetanustoxin receptors in early postnatal and adult mouse cerebellum. J. Neuroimmunol. 1, 429-456. Schnitzer J. and Schachner M. (1982) Cell type-specificity of neural cell surface antigen recognized by monoclonal antibody A2B5. Cell Tissue Res. Schon F. and Kelly J.S. (1974) Autoradiographic localization of [3H]GABA and [3H]glutamate over satellite glial cells. Brain Res. 66, 275-288. Schousboe A., Lisy V. and Hertz L. (1976) Postnatal alterations in effects of potassium on uptake and release of gluamate and GABA in rat brain cortex slices. J. Neurochem. 26, 1023-1027. - 168 -Schousboe A., Hertz L. and Svenneby G. (1977a) Uptake and metabolism of GABA in astrocytes cultured from dissociated mouse brain hemispheres. Neurochem. Res. 2, 217-229. Schousboe A., Svenneby G. and Hertz L. (1977b) Uptake and metabolism of glutamate in astrocytes cultured from dissociated mouse brain hemispheres. J. Neurochem. 29, 999-1005. Schousboe A. (1978a) Glutamate, GABA and taurine in cultured normal glial cells. In: Dynamic Properties of Glial Cells. Eds., E. Schoffeniels, G. Franck, L. Hertz and D.B. Tower. Pergamon Press, New York, pp.173-182. Schousboe A. (1978b) Effects of GABA-analogues on the high affinity uptake of GABA in primary cultures. In: Advances in Experimental Medicine and Biology, Vol.123, GABA Biochemistry and CNS Function. Eds., P. Mandel and F.V. deFeudis. Plenum Press, New York, pp.219-237. Schousboe A. and Divac I. (1979) Differences in glutamate uptake in astrocytes cultured from different brain regions. Brain Res. 179, 407-409. Schousboe A., Hertz L., Svenneby G. and Kvamme E. (1979) Phosphate activated glutaminase activity and glutamine uptake in astrocytes in primary culture. . J. Neurochem 32, 943-950. Schousboe A., Nissen C, Bock E., Sapirstein V.S., Juurlink B.H.J, and Hertz L. (1980a) Biochemical development of rodent astrocytes in primary cultures. In: Tissue Culture in Neurobiology, E. Giacobini, A. Vernadakis. A. Shahar, Eds. Raven Press, pp.397-409. Schousboe A., Drejar J. and Divac I. (1980b) Regional heterogeneity in astroglial cells, implications of neuronal glial interactions. Trends Neurosci. 14, 13-14. Schousboe A. (1981) Transport and metabolism of glutamate and GABA on neurons and glial cells. Intnat. Rev. Neurobiol. 22, 1-45. Schousboe A., Larsson O.M., Drejar J., Krogsgardre-Larsen P. and Hertz L. (1983) Uptake and release processes for glutamine, glutamate and GABA in cultured neurons and astrocytes. In: Glutamine, Glutamate and GABA in the Central Nervous System. Eds., L. Hertz, E. Kvamme, E.G. McGeer and A. Schousboe. Alan R. Liss Inc., New York, pp.297-315. Schrier B.K. and Thompson E.J. (1974) On the role of glial cells in the mammalian nervous system. J. Biol. Chem. 249, 1769-1780. Sellstrom A., Sjoberg L.K.B. and Hamberger A. (1975) Neuronal and glial systems for -aminobutric acid metabolism. J. Neurochem. 25, 393-398. - 169 -Seress L. (1980) Development and structure of the radial glia in the postnatal rat brain. Anat. Embryol. 160/ 213-226. Shaw G., Osborn M. and Weber K. (1981) An immunofluorescence microscopical study of the neurofilament triplet proteins, vimentin and glial fibrillary acidic protein within the adult rat brain. Europ. J. Cell Biol. 26, 68-82. Shine H.D., Hertz L., deVellis J. and Haber B. (1981) A fluorometric assay for [g]-Glutamyl transpeptidase: demonstration enzymatic activity in cultured cells of neural origin. Neurochem. Res. 6, 453-463. Shoemaker H., Boles R.G., Horst W.D. and Yamamura H.I. (1983) Specific high affinity binding sites for [3H]Ko5 - 4864 in rat brain and kidney. J. Pharmacol. Exp. Ther. 225, 61-69. Silberstein S., Shein H. and Bern K. (1972) Catechol-O-methyl transferase and monoamine oxidase activity in culture rodent astrocytoma cells. Brain Res. 41, 245-248. Skoff R.P. (1980) Neuroglia: A re-evaluation of their origin and development. Path. Res. Pract. 168, 279-300. Snodgrass S. and Iversen L. (1974) Amino acid uptake into human brain tumors. Brain Res. 76, 1769-1780. Sommer I. and Schachner M. (1981) to oligodendrocyte cell surfaces: CNS. Dev. Biol. 83, 311-327. Monoclonal antibodies (ol-04) An immunological study in the Sommer I., Lagenaur C. and Schachner M. (1981) Recognition of Bergmann glial and ependymal cells in the mouse nervous system by monoclonal antibody. J. Cell Biol. 90, 448-458. Spacek J. (1968) A contribution to the electron microscopy of neuroglia: II. Relation between glial and nerve cells. Cs. Fysiol. 17, 275-276. Spatz H. (1922) Eisennachweis im gehirin, besonders in zentren des extraphyramidalen systems. Z. Ges. Neurol. Psychiat. 77, 261-390. Stallcup W.B. and Cohn M. (1976) Correlation of surface antigens and cell type in cloned cell lines from rat central nervous system. Exp. Cell Res. 98, 285-297. Sternberger N.H., Itoyama Y., Kies M.W. and Webster H.deF. (1978) Immunocytochemical method to identify basic protein in rnyelin-forming oligodendrocytes of newborn rat central nervous system. J. Neurocytol. 7, 251-263. - 170 -Stewart R.M./ Martuza R.L., Baldessarini R.J. and Kornblith P.L. (1976) Glutamate accumulation of human gliomas and menigiomas in tissue culture. Brain Res. 12, 549-559. Strecker H.J. (1960) The interconversion of glutamic acid and proline II - the preparation and properties of l-pyrroline-5-carboxylic acid. J. Biol. Chem. 235, 2045-2050. Strecker H.J. (1971) l-pyrroline-5-carboxylate dehydrogenase. In: Methods in Embryology. Eds., H. Tabor and C.W. Tabor. Academic Press, New York/ 17B., pp.262-265. Sturrock R.R. (1976) Light microscopic identification of immature glial cells in semithin sections of the developing mouse corpus callosum. J. Anat. 122, 521-537. Suess U. and Pliska V. (1981) Identification of the pituicytes as astroglial cells by indirect immunofluorescence-staining for the glial fibrillary acidic protein. Brain Res. 221/ 27-33. Sugaya E., Sekiya Y., Kobori T. and Noda Y. (1979) Glial membrane potential and extracellular potassium concentration in cultured glial cells. Exp. Neurol. 66, 403-408. Swainman K.F. (1984) Grant proposal effect of iron on brain tissue and its relationship to Parkinsonism. Szeligo V. and Leblond CP. (1977) Response to the three main types of glial cells of cortex and the corpus callosum in rats handled during suckling or exposed to enriched, control or improverished environment following weaning. J. Comp. Neurol. 172, 247-264. Szentagothis J. (1970) Glomerular synapses, complex synaptic arrangements and their operational significance. In: The Neurosciences, 2nd Study Program. F.O. Schmitt et al., Eds. New York, Rockefeller Univ. Press, pp. 427-443. Szuchet S. and Stefansson K. (1980) In vitro behavior of isolated oligodendrocytes. Adv. Cell Neurobiol. 1, 313-346. Szuchet S. and Yim S.H. (1984) Characterization of a subset of oligodendroctyes separated on the basis of selective adherence properties. J. Neurosci. Res. 11/ 131-144. Tabuchi K., Furuta T., Norikane H., Tsuboi M., Moriya Y. and Nishimoto A. (1981) Evaluation of the drug-induced morphological differentiation of rat glioma cells (C-6) from the aspects of S-100 protein level and Con A binding pattern. J. Neurological Sci. 51, 119-130. - 171 -Takats A., Komoly S., Fazekas A., Tarczy M., Papp M. and Auguszt A. (1981) Comparative histochemical and immunofluorescent observations in thiamine deficient encephalopathies. Acta Neuropathol. (Berl)/ supp.VII, 50-51. Tamir H., Klien A. and Rapport M.M. (1976) Serotonin binding protein: enhancement of binding by Fe2+ and inhibition by drugs. J. Neurochem. 26, 871-878. Tardy M., Bardackdjiam J. and Gonnard P. (1978) GABA metabolism in cultured glial cells. In: Advances in Experimental Medicine and Biology 23, 177-187. Tardy M., Fages B., Rolland B., Bardardjian J. and Gonnard P. (1931) Effect of prostiglandins and dibutyrtl cyclic AMP on the morphology of cells in primary astroglial cultures and on metabolic enzymes of GABA and glutamate metabolism. Experentia 37, 19-21. Thompson R.J., Kynoch P.A.M. and Sarjant J. (1980) Imraunohistochemical localization of creatine kinase BB isoenzyme to astrocytes in human brain. Brain Res. 201, 423-426. Thompson S.G. and McGeer E.G. (1985) GABA-transaminase and glutamic acid decarboxylase change in the brain of rats treated with pyrithiamine. Neurochem. Res. 10/ 1653-1660. Thompson S., Wong P., Leung S. and McGeer E. (1985) Regional distribution in rat brain of l-pyroline-5-carboxyline dehydrogenase and its localization to specific glial cells. J. Neurochem. 45, 1791-1796. Thornber E.J., Dunlop R.H. and Gawthorne J.M. (1980) Thiamine deficiency in the lamb: changes in thiamine phosphate esters in the brain. J. Neurochem. 35, 713-717. Trachtenberg M. and Pollen D.A. (1970) Neuroglia: biophysical properties and physiologic function. Science 167, 1248-1250. Trapp B.D., Itoyama Y., Macintosh T.D. and Quarles R.H. (1983) P2 protein in oligodendrocytes and myelin of the rabbit central nervous system. J. Neurochem. 40, 47-54. Trimmer P.A./ Evans T., Smith M.M., Harden T.K. and McCarthy K.D. (1984) Combination of immunocytochemistry and radioligand receptor assay to identify B-adrenergic receptor subtypes on astroglial in vitro. J. Neurosci. 4, 1598-1606. Tucker D.M., Sandstead H.H., Swenson R.A., Sawler B.G. and Penland J.G. (1982) Longitudinal study of brain function and depletion of iron stores in individual subjects. Physiol. Behav. 29/ 737-740. - 172 -Uyeda C.T., Eng L.F. and Bignami A. (1972) Immunological study of the glial fibrillary acidic protein. Brain Res. 37, 81-89. Valle D.J./ Phang J. and Goodman S. (1974) Type 2 hyperprolinemia absence of delta! l-pyrroline-5-carboxylic acid dehydrogenase activity. Science 85, 1053-1054. Van Alstyne D., Smyrnis E.M. and Paty D.W. (1983) Differentiation of glioblasts from adult brain. Neurosci. Letts. 40, 327-332. Van dan Berg CJ. (1970) Compartmentation of glutamate metabolism in the developing brain. Experiments with labelled glucose, acetate, phenylalanine, tyrosine, and proline. J. Neurochem. 17, 973-983. Van Gelder N.M. (1965) The histochemical demonstration of -aminobutryric acid metabolism by reduction of a tretrazolium salt. J. Neurochem. 12, 231-237. Varon S. (1978) Macromolecular glial cell markers. In: Dynamic Properties of Glial Cells. Eds., E. Schoffeniels/ G. Franck/ L. Hertz and D.B. Tower. Pergamon Press, New York, pp.93-103. Varon S.S. and Somjen G.G. (1979) Neuron-glial interactions. Neurosci. Res. Prog. Bull. 17, 8-239. Vernadakis A. and Arnold E.B. (1980) Age related changes in neuronal and glial enzyme activities. Adv. Cellular Neurobiol. 1/ 229-283. Vernadakis A. and Mangoura D. (1983) Glial enzymes: primary glial cultures from newborn and adult mouse brain. Abs. P-6/ J. Neurochem. 41/ supp. Virchow R. (1946) Ueber das granulierte asnchen der wandungen der gerhirnventrikel. Allg Z. Psychiat. 3/ 423-450. Vorhees CV., Schmidt D.E./ Barrett R.J. and Schenker S. (1977) Effects of thiamine deficiency on acetylcholine levels and utilization in vivo in rat brain. J. Nutr. 107, 1902-1908. Vorhees C.V., Schmidt D.E. and Barrett R.J. (1978) Effects of pyrithiamin and oxythiamin on acetylcholine levels and utilization in rat brain. Brain Res. Bull. 3, 493-496. Walum E. and Weiler C. (1978) Kinetics of glutamate, glutamine, and leucine transport in cultured neuroblastoma and glioma cells. Proc. Eur. Soc. Neurochem. 1, 499. Walz W. and Hertz L. (1982) Ouabain-sensitive and ouabain-resistant net uptake of potassium into astrocytes and neurons in primary cultures. J. Neurochem. 39, 70-77. - 173 -Waniewski R.A. and Martin D.L. (1983) Selective inhibition of glial versus neuronal uptake of L-glutamic acid by SITS. Brain Res. 268, 390-392. Watanabe I. and Kanabe S. (1978) Early edematous lesions of pyrithiamine induced acute thiamine deficient encephalopathy in mouse. J. Neuropathol. 37, 401-413. Waxman S.G. and Black J.A. (1984) Freeze-fracture ultrastructure of the perinodal astrocyte and associated glial junctions. Brain Res. 308, 77-87. Weigart F. (1895) Beitrage zur kennins der normalen menschlichen neuroglia. Weisbrod, Frankfurt/ AM Main. Weiler C.T./ Nystrom B. and Hamberger A. (1979) Characteristics of glutamine vs glutamate transport in isolated glia and synaptosomes. J. Neurochem. 32/ 559-565. Weir M.D., Patel A.J., Hunt A. and Thomas D.G.T. (1984) Developmental changes in the amount of glial fibrillary acidic protein in three regions of the rat brain. Dev. Brain Res. 15/ 147-154. Whitaker P., Vint C. and Morin R. (1983) [3H]imipramine label sites on brain astocglial cells not related to serotonis uptake. J. Neurochem. 41, 1319-1323. White F.P. and Hertz L. (1981) Protein synthesis by astrocytes in primary cultures. Neurochem. Res. 6, 353-364. White R.D. and Neal M.J. (1976) The uptake of L-glutamate by the retina. Brain Res. Ill, 79-93. Wilkin G.P., Garthwaite J. and Balazs R. (1982) Putative acidic amino acid transmitters in the cerebellum II. Electromicroscopic localization of transport sites. Brain Res. 244, 69-80. Wilkin G.P./ Levi G., Johnstone S.R. and Riddle P.N. (1983) Cerebellar astroglial cells in primary culture: expression of different morphological appearances and different ability to take up [3H]D-asparytate and [3H]GABA. Dev. Brain Res. 10, 265-277. Wilson S., Baetge E. and Stallcup W.B. (1981) Antisera specific for cell lines with mixed neuronal and glial properties. Dev. Biol. 83, 146-153. Wolff J. (1965) Elektronen mikroskopische untersuehungen uber die struktur und gestalt von astrozyten fortsatzen z. Zell Forsa 66, 811-822. - 174 -Wolff J.R. and Guldner F.H. (1978) Perisynaptic astroglial reactions to neuronal activity. In: Dynamic Properties of Glial Cells. Eds., E. Schoffeniels, G. Franck, L. Hertz and D.B. Tower. Pergamon Press, New York, pp.115-118. Woodhams P.L., Basco E., Hajos F., Csillag and Balazs R. (1981) Radial glia in the developing mouse cerebral cortex and hippocampus. Anat. Embryol. 163, 331-343. Wurjek J.R. and Reier P.J. (1984) Astrocytic membrane morphology: differences between mammalian and amphibian astrocytes after axotomy. J. Comp. Neurol. 222, 607-619. Yen S.H. and Fields K.L. (1981) Antibodies to neurofilament, glial filament and fibroblast intermediate filament proteins bind to different cell types of the nervous system. J. Cell Biol. 88, 115-126. Yoneda Y. and Roberts E. (1982) A new synaptic biosynthetic pathway of proline from ornithine and its negative feedback inhibition by proline. Brain Res. 239, 479-488. Youdim M.B.H., Green A.R., Bloomfield M.R., Mitchell B.D., Heal D.J. and Grahame-Smith D.B. (1980) The effects of iron deficiency on brain biogenic monoamine biochemistry and function in rats. Neuropharmacology 19, 259-267. Youdim M.3.H., Yehuda S. and Ben-uriah Y. (1981) Iron deficiency-induced circadian rhythem reversal of dopaminergic mediated behaviors and thermoregulation in rats. Eur. J. Pharmacol. 74, 295-301. Youdim M.B.H., Ashkenazi R., Ben-shachar D. and Yehuda S. (1984) Modulation of dopamine receptor in the striatum by iron: behavioral and biochemical correlates. In: Advances in Neurology, Vol.40 (Hassler R.G. and Christ C.G., Eds.). Raven Press, New York, pp.159-170. Yu P.H. and Hertz L. (1982) Differential expression of Type A and Type B monoamine oxidase of mouse astrocytes in primary cultures. J. Neurochem. 39, 1492-1495. Zalc B., Collet A., Monge M., Oilier-Hartmann P., Jacque C, Hartmann L. and Baumann N.A. (1984) Tamm-Horsfall protein, a kidney marker is expressed on brain sulfogalactosylceramide-positive astroglial structures. Brain Res. 291, 182-187. - 175 -


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items