University of Colorado at Boulder
THE ROLE OF ASTROCYTES IN NEURAL REGENERATION OF THE CENTRAL NERVOUS SYSTEM
Troy S. Knapp
Introduction to Neuroscience 11:0012:00 M/W/F Dr. Tim Smock December 11, 1995
ABSTRACT Astrocytes display properties that are both inhibitory and beneficial to neural regeneration. Unfortunately, the exact nature of these inhibitory properties as of yet are unknown. Two hypotheses have been offered. One is based on the mechanically inhibitory properties of astrocyte scars. The other is based on the chemically inhibitory properties of a yet unknown substance. Fortunately, a great deal more is known about the beneficial properties of astrocytes. The goal of this paper is threefold: First, to review both the inhibitory and beneficial properties of astrocytes in neural regeneration. Second, to critique the strength of the more prevalent arguments found in the literature that deal with these inhibitory and beneficial properties. Third, to propose an experiment designed to draw out yet another beneficial property of astrocytes. REVIEW OF THE CURRENT LITERATURE Almost a century ago the German anatomist Rudolf Virchow recognized that brain cells could be divided into two distinct categories: (1) Neurons and (2) Neuroglia (Levitan et al, 73) Neuroglia is by far the most numerous of the cell types, out numbering neurons by a factor of approximately ten to one (Bignami et al, 1.) In fact, according to an anonymous corespondent in Nature about half of the volume of the vertebrate brain is composed of glial cells (Bignami et al). Until the 1920's neuroglia was believed to be a single functional unit that served only as a putty providing structural support to adjoining
neurons (Streit et al, 54.) At this time Pio del RioHortega developed a silver carbonate stain that made possible differentiation of the three types of glial cells, astrocytes, oligodendrocytes and microglia (Streit et al). Each type of glial cell has a specialized function relatively independent of the other. Here we are most interested in the actions of astrocytes and their contribution to neural regeneration in the CNS. Astrocytes perform a myriad of functions in the CNS, including, maintaining a stable neuronal microenvironment, uptake of amino acids, production of growth factors, and protection from oxygen toxicity. Astrocytes also seem to play an inhibitory role in the neural regeneration of the CNS. Whether this role is mechanical or chemical is still a matter of debate. The main function of astrocytes is to assure the stability of the neuronal microenvironment (Bignami et el, 31). In both gray matter and white matter we find that astrocytes are ideally located for carrying out this task. In gray matter astrocytes surround neurons and their processes, while in white matter astrocytes mainly surround the oligodendrocytic product, myelin. Further, astrocyte processes form a continuous lining on the surface of the brain and of blood vessels entering the brain from the leptomeninges. Astrocytes exhibit gap junctions with other astrocytes in both white and gray matter, forming a functional syncytium equilibrating changes in concentrations of ions and small solutes (Bignami et al, 36). These gap junctions facilitate astrocytes in maintaining the neuronal
microenvironment.
Potassium homeostasis is the clearest example
of astrocytes maintaining the neuronal microenvironment. The amount of potassium [K+] released during neuronal activity, such as an action potential, is relatively small, though it results in a significant increase in extracellular [K+]. These excess levels of [K+] are taken up by the astrocyte and dispersed through the syncytium (Bignami et al, 36). The effectiveness of this syncytium is evidenced by an increase in local extracellular [K+] being redistributed through the syncytium to distant areas where extracellular [K+] is lower (Brightman et al, 113). Research indicates that this syncytium may be responsible for the protection astrocytes seem to receive from some types of neuronal toxicity. Investigators found, using Ltranspyrolidine 2, 4dicarboxlic acid (transPCDA), that astrocytes were generally neuroprotective under excitotoxic conditions (Dugan et al, 3). The rational being that the functional syncytium easily dissipated the toxicity away from the region of highest concentration. Astrocytes also mediate toxicity by uptake of amino acid neurotransmitters. For example, astrocytes show a much grater uptake capacity for glutamate, one of the main excitatory neurotransmitters in the brain, then do neurons. Regional differences do exist in this respect however. This uptake capacity for glutamate is not surprising as this amino acid, like many other acidic amino acids, are neurotoxic (excitotoxins) (Bignami et al, 36).
The glutamate that astrocytes uptake is used for ammonia detoxification, yet another example of astrocytes maintaining the neuronal microenvironment. The evidence supporting this is twofold. First, the brain enzyme responsible for the formation of glutamine from glutamate and ammonia is exclusively localized in astrocytes (Murphy et al, 383). Secondly, the swelling and vacuolation of astrocytic nuclei is the most prominent finding in patients dying of hepatic coma, believed to be caused by excess levels of blood ammonia (Murphy et al). As Bignami discussed the requirements of maintaining the neuronal microenvironment are so much more stringent in gray matter than in white matter that one would expect gray matter and white matter astrocytes to be different in this respect. Synaptic activity in the integrating zone of neurons found in gray matter are presumably more sensitive to small changes in the microenvironment then the conduction of action potentials found in white matter. Bignami therefore concluded that cerebral white matter is more comparable to peripheral nerve then to gray matter in respect to maintaining neuronal microenviroments. Astrocytes exhibit receptors for several types of growth factors as well as appear to produce both Nerve Growth Factor (NGF) as well as Basic Fibroblast Growth Factor (bFGF). bFGF is *
known to promote not only the survival of neuronal cells, but * More specifically bFGF has a wide range of tissue distribution and
the broadest spectrum of biological activities. Due to striking physiochemical properties several different factors are lumped under the term bFGF. They include: Astroglial growth factor , Heparinbinding growth factor class II and tumor angiogenesis factor.
also the proliferation and differentiation of nonneuronal cells like astrocytes (Enokido et al, 106). Since bFGF is found to have a positive growth action in both astrocytes and neurons the question is raised as to which cell is exhibiting an influence on the other. Bignami notes that two possibilities exist as to the source of these neurotrophic factors. They are either produced by the innervated target or by the glial cells responsible for these targets (p, 34). Some confusion exists in the literature as to this point. The uncertainty being what quantity of neurotrophic factors are produced in astrocytes verses neurons. It is well established, however, that neurons are the main site of NGF expression in normal CNS tissue. Though, tissue evidence suggests that astrocytes may by the source of NGF in damaged CNS tissue (Bignami, 34). The rational for this being that NGF mRNA levels have been found to be increased in primary astrocyte cultures stimulated by several cytokins and bFGF. Furukawa, however, asserts that healthy astroglial cells are known to synthesize NGF (p.42). In addition to the production of NGF and bFGF astrocytes have been found to contain receptor sites for epidermal growth factor (EGF) (Huff, 659). The investigators found that binding of EGF by the astrocytes was saturable, specific and not competed by NGF or bFGF. They did find, however, a 70% reduction in EGF binding when the astrocytes were pretreated with bFGF for an 18 hour period. This lead them to the conclusion that bFGF may
serve as an off switch for the EGF mitogenic signal in astrocytes. While the exact physiological significance of neurotrophic synthesis (whether in healthy or damaged CNS tissue) is unclear, it is clear that astrocytes are one of the sources providing neurotrophic factors to neurons. Regardless of the conditions under which astrocytes produce trophic factors, it has been securely documented that neurons require a glial environment to grow. Dugan and colleagues found this by developing "pure" neuronal cultures from mouse neocortex to study the effect of glial cells on the response of neurons to injury. They found that neuronal cultures grew best on a glial base, coming to the conclusion that cortical neurons require a glial conditioned medium in order to survive (p, 4545). Mark Noble came to the same conclusion in his study on the developmental biology of the optic nerve. Noble compared the growth of optic neurons on a variety of substrates including, optic astrocytes, schwann cells, skin fibroblasts, and cardiac myocytes. He found that dissociated neurons plated onto astrocytes grew as if they preferred the astrocytic surface to any other surface availalbe (p, 9). Noble saw extensive crossing over of neuronal processes, and an occasional instance of processes running in parallel for short distances (p, 9). Growth factors are not the only astrocyte product that are beneficial to neuronal survival. Astrocytes have been found to release pyruvate which has also been determined to have a positive effect on neuronal survival (Selak et al, 23).
Astrocytes also seem to play a role in protecting CNS neurons from oxygen toxicity, some of which is thought to be produced by microglia. These reactive oxygen species are believed to have the beneficial effect of damaging microbe membranes, proteins and DNA. Unfortunately, reactive oxygen species damage healthy cells in the same way (Streit et al, 61). These oxygen free radicals have been implicated as a potential cytotoxic mechanism responsible for nigral cell death in Parkinson's disease (JovoyAgid, 92). There are several enzymes in the CNS that act in concert to defend against this oxygen toxicity. Fist the superoxide radical is mutated to hydrogen peroxide by superoxide dismutase (SOD). The hydrogen peroxide is then decomposed to water or removed by glutathion peroxidase (GPX) which uses hydrogen peroxide to oxidize the reduced glutathion (JovoyAgid). Through immunocytochemical studies on human mesencephalon JovoyAgid found that GPX is found exclusively in astrocytic cells, showing that astrocytes are responsible for mediating oxygen toxicity. Knowing the functions of astroglial cells we now turn our attention to their role of neural regeneration in the CNS. A profound difference exists in the ability of neural regeneration in the CNS and the peripheral nervous system (PNS). This is demonstrated by the following, if an efferent dorsal root ganglion is cut the axon will regenerate normally reaching its peripheral target allowing functional recovery. However, if a CNS axon is severed the regenerative attempt will be abortive and non functional. A regenerating PNS axon will stop at the PNSCNS
junction. This apparently happens for one of two reasons. First, the lack of CNS regeneration is due to an active inhibition on the part of the CNS. Second, the CNS lacks regenerative factors normally found in the PNS. Astrocytes at the PNSCNS junction would share responsibility for the lack of neural regeneration in either of these two cases. One of the longest held hypothesis for the lack of neural regeneration in the CNS is astrocyte scar tissue resulting from CNS injury. The major mass of astrocytic scar tissue is formed from bundles of cytoplasmic intermediate filaments which are made of GFAP, an astrocytespecific protein (Bignami, 84). Upregulation of GFAP production is a main factor in the formation of glial scars. In this case astrocyte proliferation also occurs but is limited and confined to the area of injury. These astrocytic scars have been believed to be responsible for inhibition of neural regeneration (Bignami, 85). It has been suggested that increased GFAP expression by astrocytes is the most sensitive indicator of neural damage in the CNS. Farooque and colleagues found this to be true in rats. The investigators used immunohistochemistry to detect changes in the expression of GFAP in spinal tracts after using blocking weight techniques to induce spinal cord compression at the level of the eight and ninth thoracic vertebrae (Farooque et al, 41). The investigators found that within 24 hours post compression widespread astrocyte reaction occurred. Even mild compressions that did not produce any signs of dysfunction induced widespread astrocytic alterations. Further, the astrocyte response was more
marked in rats with more severe compression leading to more pronounced neurological deterioration (Farooque). Recent research has indicated that this astrocyte injury response is not entirely local. Studies by Janeczko point to the possibility that astrocytes migrate from peripheral areas to participate in the process. His evidence for this is that [3H]thymidinelabeled astrocytes at first scattered over a relatively wide area later became concentrated around CNS lesions (p, 236). According to Bignami several publications now suggest that normal glial cells (astrocytes and oligodendrocytes) have the ability to migrate in CNS tissue (p, 85). Astrocytes are found to exhibit scaring in both injuries that produce dysfunction and in injuries that do not. Yamada and colleagues have found the formation of astrocyte scar tissue does not create a major barrier in CNS neuronal regeneration in lower vertebrates. The researchers investigated axonal regeneration in the CNS using fine structural and histochemical aspects of the carp spinal cord, which was completely transected at the level of the dorsal fin. Fusion of the transected region and regeneration of axons was apparent at 26 days post lesion. By 115 days post lesion the rostral and caudal portion of the transected spinal cord were completely connected by the regeneration nervous tissue (p, 324). Horseradish peroxidase injected in the spinal cord at the portion caudal to the transection site was detected in the cytoplasm of large neurons located in the reticular formation of the midbrain (p, 325). This demonstrates that long axons regenerated through the ablation gap, indicating that
regenerating axons in carp spinal cord can pass through the glial scar bundle formed in the transected portion. Many of these regenerating axons were found to be in contact with astrocytes, indicating that glial cells do not play a major role as an obstacle for the prolongation of axons in the carp spinal cord (p, 325). Unfortunately, several observations are not readily explained by the scar hypothesis. First, extensive axonal growth may be observed in glial scars, as noted by Yamada and colleagues. Secondly, a study performed by Chi and Dahl on nerve grafting found that glial scars formed as the result of damage done during surgery at the interphase between brain and peripheral nerve implants do not prevent axons growing from the brain into the graft (p, 245). These inconsistencies are best accounted for by a second hypothesis. A chemical mechanism has been proposed by Liuzzi and Tedesch as the barrier to CNS neural regeneration. Knowing that regeneration of a peripheral axon stops at the junction to the CNS they proposed that astrocytes transmit some sort of 'stop' signal when contacted by a growing axon (Liuzzi, 4783). This signal would be similar to the physiological mechanisms that stop growth when axons reach their destination in development, except that in the first case the message is delivered by the astrocyte membrane and in the second case by the postsynaptic membrane (Bignami, 95). Unfortunately, this signal has not been found or identified. Oligodendrocytes however seem to exhibit a sort of stop signal on their surface that act as an axonal repellent
(Bignami et al, 95). This was observed when dorsal root ganglion (DRG) axons are put in contact with sciatic nerve (a PNS nerve) and an optic nerve (a CNS nerve) in vitro. The DRG grows inside the sciatic nerve and avoids the latter optic nerve. This lead to the isolation of two inhibitory proteins fractions at 250 and 35 kD and to the demonstration that antibodies to these proteins promote axonal growth in the spinal cord (Schnell et al, 269). We have found that astrocytes perform a variety of functions in the CNS. Including maintaining a stable neuronal microenvironment, uptake of amino acids, production of growth factors, and protection from oxygen toxicity. Unfortunately, none of these functions elucidate hints as to why astrocytes appear to have an inhibitory effect on neural regeneration in the CNS. Both a mechanical and chemical hypothesis have been explored. The glial scar hypothesis has several unresolved contradictions. Therefore, the chemical hypothesis seems more solid. CRITIQUE OF THE LITERATURE Within the reviewed literature there are three areas lacking clarity. First, the effect of bFGF on astrocytes is unclear. Second, there is confusion in whether we find greater NGF production in healthy or damaged astrocytes. Third, though a physiological mechanism seems to prevent regenerating axons from crossing the PNSCNS barrier, none has been found or identified. As stated before bFGF is known to promote not only the survival of neuronal cells but also the proliferation and differentiation of nonneuronal cells like astrocytes (Enokido et al, 106). Both neurons and astrocytes express receptors for and
produce bFGF. Further, Enokido found that astroglia fibers increased in number with the addition of bFGF. Because we find both astroglia and neurons responding to and producing bFGF the question is raised as to which cell is having an action on the other. Or, does the possibility exist that they share a mutually beneficial relationship. It appears that the literature is silent on any mutually inclusive action bFGF may have on astrocytes and neurons. Secondly, uncertainty exists as to the production of NGF in astrocytes. Bignami has cited tissue evidence suggesting that astrocytes may be a source of NGF in damaged CNS tissue (Bignami et al, 34). His rational for this is the finding of NGF mRNA in primary astrocyte cultures. Furukawa states very clearly that healthy astrocytes are known to synthesize NGF in cultures. His evidence for this is that murine astrocytes synthesize and secrete molecules identical to murine submaxilary gland derived NGF with respect to molecular weight, isoelectric point, antigenicity and neurite promoting activity (Furukawa, 62). The question as to whether healthy or injured astrocytes produce NGF is of importance. If healthy astrocytes produce NGF then they could be considered to have a maintenance role in the CNS. If injured astrocytes produce NGF then they could be considered to have restorative role. The answer to this question goes to the very basis of defining the role astrocytes play in the CNS. Thirdly, Bignami asserts as a reasonable hypothesis that astrocytes possess on their surface a chemical that transmits a signal to growing axons that in effect tells them to stop their
regeneration. Bignami puts forth this hypothesis though he is unable to provide any information as to the nature of this chemical signal. This chemical hypothesis is found elsewhere in the literature though in no place is the structure or action of the chemical signal elucidated. The basis for Bignamis' hypothesis seems to be that oligodendrocyte appear to possess this chemically inhibitory property. Bignami states that in oligodendrocytes these inhibitory proteins, as well as antibodies for them have been identified (p, 95). Bignami is the only place in the literature that I have found a statement that the inhibitory proteins in oligodendrocytes have been identified. While the chemically inhibitory hypothesis appears reasonable it seems directly opposed to the other actions of astrocytes. We find astrocytes playing a variety of beneficial roles in relation to neurons. For example, the literature is very secure that astrocytes produce and maintain several types of neurotrophic factors, as well as protect neurons from oxygen and ammonia toxicity. In light of these beneficial aspects stating that astrocytes also possess a strong repellent to neuronal growth is difficult. This dilemma goes back to the previous dilemma raised by Bignami and Furukawa. Bignami felt that injured astrocytes were the source of astrocytic NGF production, while Furukawa believed that healthy astrocytes produced this NGF. This lead us to the question of whether astrocytes play a maintenance or restorative role in the CNS. The answer to this question in turn will provide hints as to which hypothesis against neural regeneration
(mechanical or chemical) is correct. If the role of astrocytes in the CNS is found to be restorative then the chemical inhibition hypothesis against neuronal regeneration seems out of place. The reasoning here being that a restorative role would be in opposition with a chemically inhibitory signal. If the role is indeed one of maintenance then the chemically inhibitory hypothesis of neuronal regeneration would appear valid. The thought here being that a chemically inhibitory signal on astrocytes provides a balance system against the neurotrophic factors, while the neurotrophic factors provide a check system against the chemically inhibitory signal. This arrangement would create a homeostatic state in which neither influence could exert its will unchecked. This would be in keeping with a maintenance role. Unfortunately, experimental verification of this check and balance theory of astrocyte chemical message and neurotrophic factor homeostasis would be difficult to carry out without knowing the make up or antibody factors of the chemically inhibitory signal. If NGF production was inhibited in a controlled environment one would expect the inhibitory chemical signal on the astrocyte to be free to express its action. This action would be to inhibit neuronal growth. However, it would be difficult to determine if any decrease in axonal sprouting or growth was due to the inhibitory chemical signal, or if it was due to a lack of NGF. If antibodies to the inhibitory chemical were known then an experiment to test this theory could most likely be carried out.
If the inhibitory chemical signal could be inhibited then NGF would be expected to express its effect unchecked. This could be easily checked by looking for any increases in axonal growth or sprouting in a controlled environment. The answers to a great many questions are resting on the discovery of the structure of the proposed inhibitory chemical signal on astrocytes. The most exciting discovery in recent literature is that of astrocytes having the ability to migrate in the CNS from peripheral areas to lesion areas as discovered by Janeczko and verified by Bignami (p, 236: p, 85). In light of the many beneficial aspects that astrocytes have been shown to exhibit in neuronal maintenance, and possibly in neural regeneration, this discovery is even more exciting. Naturally, the question is raised, if astrocytes migrate to injured areas and exhibit their beneficial properties can they be placed at the site of injury and remain viable so as to exhibit their beneficial properties? PROPOSED EXPERIMENT In hopes of gaining further understanding of the potentially beneficial aspects of transplanted astrocytes I propose the following experiment. In fifteen male Wistar rats of the same age (approximately 90 days) create a 1mm lesion in the white matter of the right cerebral hemisphere underlying the cerebral cortex. At 30 minutes post lesion inject .25mL of pure cultured *
* The rational for creating a lesion in the white matter is based on
Bignami's hypothesis that cerebral white matter is less sensitive then gray matter to small changes in the microenvironment. Because white matter is more robust to changes in the microenvironment accidental contamination due to the surgery will play less of a role as a confounding variable.
astrocytes (as derived by Dugan and colleagues (p, 4546)) that have been labeled with [3H]thymidine. At 40 days post lesion kill the animals and fix the brains in Bouins's fixative and section at 5 µm intervals in the coronal plane. Label 5 of the coronal sections according to the method developed by Janeczko and Bignami
Knapp 1
(Janeczko, p. 237). Examine a 100 x 100 µm area of lesion in #
each of the 5 coronal sections. Tally both the number of labeled and unlabeled astrocytes in these sections as well as the number of astrocytes in the corresponding contralateral unlesioned area. Further, estimate the number of labeled astrocytes placed into the lesion area. From these numbers perform a Students two tailed ttest to determine the following: (1) If there is a significant difference between the number a labeled astrocytes added to the lesion area verses the number of labeled astrocytes found in the lesion area after the animal was killed. A significant difference here would suggest one of two things. First, that the astrocytes are incapable of growing into the lesion area. Second, that there is some maximum number of astrocytes that the area is able to support and that number had been exceeded. Finding no significant difference would suggest that the labeled astrocytes are capable of growing into the existing neural structure. (2) If there is a significant difference between the total number of astrocytes on the lesion side verses the unlesioned side. A significant difference may indicate that the labeled astrocyte culture was able to grow into the existing structure. No significant difference between the two sides may indicate that there is some maximum level of astrocytes that a # The method is as follows: stain the coronal sections
immunocytochemically by the peroxidaseantiperoxidase method according to Van Noorden. Then prepare autoradiographs from the immunocytochemically stained sections by the dipping technique using illford K2 emulsion, expose for 21 days, develop and stain with Harris' hematoxylin and eosin.
Knapp 2
given area can support, based on the assumption that the unlesioned side is close to this maximum density. The optimal results would be to find after analysis of the data that no significant difference was found in the first case while a significant difference was found in the second case. This would indicate a probability that astrocytes are capable of being placed into a damaged neuronal environment and surviving.
Knapp 3
Work Cited Bignami, Amico, and Dahl, Doris. Medical Intelligence Unit: Glial Cells in the Central Nervous System and their Reaction to Injury. Austin: R.G. Landes Company, 1994. Brightman, Milton and TaoCheng, JungHwa. "Mutually Imposed Structural Changes in Plasma Membranes of Astroglia and Brain Endothelium." Neurology and Neurobiology: Differentiation and Function of Glial Cells. Vol 55. New York: Alan R. Liss, Inc., 1990. Chi, N, Dahl, D. "Autologous Peripheral Nerve Grafting into Murine Brain as a model for Studies of Regeneration in the Central Nervous System." Experimental Neurology. 79 (1983): 245264. Dugan, L, Bruno, V, Amagasu, S, Giffard, R. "Glia Modulate the response of murine cortical neurons to excitotoxicity: Glia exacerbate AMPA neurotoxicity." Journal of Neuroscience 15(6) (1995): 45454555. Enokido, Y, and Hatanaka, H. "Neurotrophic Factors Rescue Neuronal Cell Death caused by Oxygen Toxicity in Culture." Neurotrophic Factors. Taniguchi Symposia on Brain Sciences, No. 15. Bocca Raton: CRC Press, 1992. Farooque, M, Badonic, T, Olsson, Y, and Holtz, A. "Astrocytic Reaction after Graded Spinal Cord Compression in rats: Immunohistochemical Studies on Glial Fibrillary Acidic Protein and Vimentin." Journal of Neurotrauma 12(1), 1995: 4145. Furukawa, Yoshiko. "Regulation of Nerve Growth Factor Synthesis by Caechol Derivatives." Neurotrophic Factors. Tanigachi Symposia on Brain Sciences, No 15. Bocca Raton: CRC Press, 1992. Huff, K. "Astrocyte Binding of Epidermal Growth Factor." Neural Development and Regeneration: Cellular and Molecular Aspects. NATO ASI Series H: Cell Biology, Vol 22. 1987.
Knapp 4
JavoyAgid, F. "Factors Associated to Dopaminergic Cell Death in Parkinson's Disease." Trophic Regulation of the Basal Ganglia: Focus on Dopamine Neurons. WennerGren International Series, Vol 62. Britan: Butler and Tanner, Ltd. 1994. Janeczko, K. "Spatiotemporal Patterns of the Astroglial Proliferation in rat Brain Injured at the Postmitotic Stage of Postnatal Development: a Combined Immunocytochemical and Autoradiographic study. Brain Research. 485, 1989: 236243. Kaczmarek, Leonard, and Levitan, Irwin. The Neuron: Cell and Molecular Biology. New York: Oxford University Press, 1991. KincaidColton, Carol, and Streit, Wolfgang. "The Brain's Immune System." Scientific American. Nov 1995. Liuzzi, F, and Tedeschi, B. "Axoglial Interactions at the Dorsal Root Transitional Zone Regulate Neurofiliment Protein Synthesis in Axotomized Sensory Neurons." Journal of Neuroscience. 12 (1992): 47834792. Murphy, S, and Pearce, B. "Functional Receptors for Neurotransmitters on Astroglial cells." Journal of Neuroscience 22 (1987): 381394. Noble, Mark. "Developmental Biology of the Optic Nerve." Neural Development and Regeneration: Cellular and Molecular Aspects. NATO ASI Series H: Cell Biology, Vol 22. 1987. Selak, I, Skaper, S, Varon, S. "Pyruvate Participation in the low Molecular Weight Trophic Activity for CNS Neurons in Glia Conditioned Media." Journal of Neuroscience 5 (1985): 2328. Schnell, L. Schwab, M. "Axonal Regeneration in the Rat Spinal Cord Produced by an Antibody Against Myelinassociated Neurite Growth Inhibitors. Nature. 343 (1990): 269272. Van Noorden, S. "Tissue preparation and immunostaining techniques for light microscopy." Immunocytochemistry: Modern Methods and Applications. Wright Bristol (1986): 2653.
Knapp 5 Yamada, H, Miyake, T, and Kitamura, T. "Regeneration of Axons in the Carp Spinal Cord." Zoological Science. 12(3) (1995): 325 332.