v THE PRESENCE OF NAKED PURKINJE CELL DENDRITIC SPINE SPECIALIZATIONS AND OTHER PATHOLOGICAL CHANGES IN TEN DAY SALINE AND MET HYLAZOXYMETHANOL- INJECTED SWISS ALBINO MICE Thesis for the Degree of M. S. MICHIGAN STATE UNIVERSITY THOMAS H. HARTKOP 1976 Q h 1:. int: WW'HO‘AoVIM'm‘u-ux 'szo -~. :4”! m ‘ ‘5‘ ‘5' . _ P.‘ h ’I' ~. I 'r "’:‘- . ‘\ L4 2 3.31.1. -' :_ j: ‘ l f‘ r . ‘2 o ,": i‘nquI-vfifpj»: ‘r‘f'uar‘v‘b \xzu,..ss~ .1... u¢¢»JJ “I Q I O . ‘Iv'o s-r: .- - ‘11": 1Q ABSTRACT THE PRESENCE OF NAKED PURKINJE CELL DENDRITIC SPINE SPECIALIZATIONS AND OTHER PATHOLOGICAL CHANGES IN TEN DAY SALINE AND METHYLAZOXYMETHANOL-INJECTED SWISS ALBINO MICE BY Thomas Henry Hartkop Postnatal Swiss albino mice, Webster strain, treated at day zero with methylazoxymethanol acetate (MAM) or saline 0.05 ul/gm body weight, were sacrificed at 10 days of age. Attention was directed towards the differentiation of synaptic structures as well as to physical, histological and ultrastructural alterations. Naked Purkinje cell dendritic spines with postsynaptic spine Specializations were observed in both control and treated ten postnatal day mice. Many granule cells in their definitive locations exhibited parallel fiber-Purkinje cell contacts. A planimetric evaluation of the midsagittal surface area of the vermis revealed an average reduction of 60% in the MAM-treated mice. Dislocation of Purkinje cell somata, random orienta- tion of their apical poles and alterations within the Purkinje cell dendrites in 10 day MAM-treated mice were observed. The vertical processes of the Golgi epithelial Thomas Henry Hartkop (Bergmann) cells were reduced in length passed in an oblique direction towards the meninges, and contained fewer than normal warty excrescences. Necrotic debris was observed within the external differentiating cell layer in 10 day MAM-injected mice. No evidence of degeneration of pre- synaptic terminals was observed in either group. This study would appear to strengthen the hypothesis that Purkinje cell dendritic Spine specializations do not require permanent presynaptic contact by parallel fibers for development. THE PRESENCE OF NAKED PURKINJE CELL DENDRITIC SPINE SPECIALIZATIONS AND OTHER PATHOLOGICAL CHANGES IN TEN DAY SALINE AND METHYLAZOXYMETHANOL-INJECTED SWISS ALBINO MICE by Thomas H. Hartkop A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Anatomy 1976 Copyright by Thomas Henry Hartkop 1976 ii DED ICATI ON I dedicate this manuscript to my wife, Michele, my mother, Georgina R. Hartkop, and to the memory of my father, Dr. Henry H. Hartkop. iii ACKNOWLEDGMENTS I wish to express my sincere thanks to my major Professors, Dr. Thomas W. Jenkins and Dr. Margaret 2. Jones for their guidance and friendship during the past years. Appreciation is also due to Dr. Lawrence M. Ross and Dr. John E. Wilson for serving on my committee. I would like to also thank Donna Craft, Mary E. Gardner, Mrs. Nina Miller, Denise I. Stanke and Warren Taylor for their technical advice and assistance and my wife Michele A. Hartkop and Rebecca McMahon for typing and proofreading preliminary drafts of this manuscript. iv TABLE OF CONTENTS Page INTRODUCTION . . . . . . . . . . . . . . . . . . 1 REVIEW OF THE LITERATURE . . . . . . . . . . . . 7 Physical Examination. . . . . . . . . . . . 7 Macroscopic Examination . . . . . . . . . . 9 Microscopic Changes . . . . . . . . . . . . 11 Purkinje cells . . . . . . . . . . . . 11 Golgi epithelial (Bergmann) cells. . . 14 External differentiating plus internal granule cell reduction and regeneration 15 Synaptogenesis. . . . . . . . . . . . . . . 16 Structure of a mature axodendritic synapse. . . . . . . . . . . . . . . . l6 Ultrastructure of Purkinje cells, granule cells, and their processes . . 18 Synaptogenesis: development of parallel fiber-Purkinje cell contacts. 20 Growth cones . . . . . . . . . . . . . 21 Naked Purkinje cell dendritic spine specializations. . . . . . . . . . . . 22 Durability of naked Purkinje cell dendritic spine specializations. . . . 26 Necrotic debris. . . . . . . . . . . . 27 Whorls . . . . . . . . . . . . . . . . 28 Methylazoxymethanol Acetate (MAM) . . . . . 29 MATERIALS AND METHODS. . . . . . . . . . . . . RESULTS. . . . . . . . . . . . . . . . . . . . Physical Examination. . . . . . . . . . . Macroscopic Examination . . . . . . . . . Light Microscopic Examination . . . . . . Ultrastructural Examination . . . . . . . DISCUSSION . . . . . . . . . . . . . . . . . . Physical Examination. . . . . . . . . . . Macroscopic Examination . . . . . . . . . Light Microscopic Examination . . . . . . Purkinje cells 0 C O O O O O O C O 0 External differentiating cell reduction and regeneration in MAM-injected mice. Golgi epithelial (Bergmann) Astroglia. Necrotic debris in MAM-injected mice Ultrastructural Examination . . . . . . . Ultrastructural overview of the cerebellar cortex. . . . . . . . . . Presence of naked Purkinje cell dendritic spine specializations. . . Durability of naked Purkinje cell dendritic spine specializations. . . Growth cones . . . . . . . . . . . . Necrotic debris. . . . . . . . . . . whorls O O O O O O O O O O O O O O O 31 35 35 37 40 43 71 71 72 73 73 79 80 81 82 82 85 86 87 87 SUMMARY AND CONCLUSIONS. APPENDIX A - EXPERIMENTAL PROCEDURES APPENDIX B - PILOT STUDIES BIBLIOGRAPHY S. 89 91 103 109 LIST OF FIGURES Figure 1. 2. 10. 11. Postural positions of mice . . . . . . . . A comparison of areas and shapes of sagittal sections through the cerebellar vermis in saline and MAM-injected mice . . Morphologic comparison of_the vermis in ten day saline and MAM-injected Swiss albino mice, A. T32-2, B. T43-10. . . . . Comparison of cerebellar folia in ten day saline and MAM-injected mice, A. T32—5, B. T62-8 Bl, C. T63-2 Bl G5 D..T63-4Bl...............-.. The random array of Purkinje cell somata in the molecular and reduced internal granule cell layers in ten day MAM-injected mouse cerebellums contrasted to control, A. T60-1 Bl, B. T 63.2 B1. 0 o o o o o o o Disorientation of Purkinje cell apical poles in MAM-injected ten day mice, A. T63-2, Bo T6]-4 0 o o o o o o o o o o o o o o o 0 Comparison of Purkinje cells and dendrites in ten day saline and MAM-injected mice (Rapid Golgi technique), A. T62-7, Bo T65-8 o o o o o o o o o o o o o o o o o Disorientation of Purkinje cell dendrites in ten day MAM-injected mice, A. T65-8, Bo T65-8 o o o o o o o o o o o o o o o o o Golgi epithelial (Bergamnn) cells, A. T62-7, Bo 1.65-8, T65-8 o o o o o o o o o o o o 0 External differentiating cell layer and molecular layer in control ten day mice, 1.62-8 Bl G5. 0 o o o o o o o o o o o o o 0 Molecular layer in ten day control mice, 1.62-8 Bl G5. 0 o o o o o o o o o o o o o o vifi Page 36 39 46 48 50 52 56 58 60 62 List 0 Figure 12. 13. 14. 15. 16. 17. 18. 19. 20. f figures continued Naked Purkinje cell dendritic spine specializations in ten day MAM-injected mice, A. T63-2 Bl G2, B. T63-2 Bl G6 . . . Naked Purkinje cell dendritic spine specializations in ten day control Growth cones in control ten day mice, TGO-lo o o o o o o o o o o o o o o o o o 0 Residual MAM-induced cellular destruction, A. T63-1 B1 62, B. T63-2 B1 62, Co T63-2 B1 G2 0 o o o o o o o o o o o o o Membranous whorls in ten day saline- injected mice, A. T62-8 B1 GS, Bo T62-8 Bl G5 0 o o o o o o o o o o o o o Toe marking system . . . . . . . . . . . . HartkOp Counting Chamber . . . . . . . . . Hartkop Electron Microscopic Measurement Nomograph O O O O O O O O O O O O O O O O 0 Naked Purkinje cell dendritie Spine specialization in 100 day control mouse. . ix Page 64 64 66 68 70 94 100 102 108 Chart 1. LIST OF CHARTS Flow diagram for determining the presence of naked Purkinje cell dendritic spines and other pathological changes in 10 day saline and methylazoxymethanol-injected Swiss albino mice. . . . . . . . . . Average area of the vermis in saline and MAM-injected mice. . . . . . . . Physical examination rating scale. . Injection Record 0 O O O O O O O O O Perfusion Record . . . . . . . . . . Electron microscopic data sheet. . . page INTRODUCTI ON The cerebellar cortex is an excellent subject for experimentation because of its histological homogeneity, and the recent advances in knowledge regarding its normal development. Moreover, the synaptic architecture of the cerebellum has been delineated from both morphological and functional standpoints. Thus, cerebellar elements and their synaptic contacts can be studied with some measure of confidence (Hirano and Zimmerman, 1973). During the development of the cerebellar cortex, dif- ferentiated cells migrate from the external differentiating cell layer and assume their definitive locations as internal granule, stellate or basket cells. Ultrastructurally, granule cells that have completed their migration into the internal granule cell layer have left behind, during their excursion, an unmyelinated ascending axon that passes into the molecular layer to its point of bifurcation. From the point of bifurcation, parallel fibers are thereby developed in a plane parallel to the folial surface with numerous varicosities, containing synaptic vesicles, that make synap- tic contact with the Purkinje cell dendritic spine specializ- ations, basket, stellate, or Golgi II dendrites (Palay and Chan-Palay, 1974; Herndon, 1964; del Cerro and Snider, 1972; Larramendi, 1969). Larramendi (1969) described the develop- ment of parallel fiber-Purkinje cell dendritic Spine synapses within the upper molecular layer of 14 day old mice. He 1 2 stated that pre- and postsynaptic specializations form after the pre- and postsynaptic elements have fused. Sub- sequent to Larramendi's (1969) report, the deletion of pre- synaptic elements by a variety of pathological processes has been shown to lead to the differentiation of naked Purkinje cell dendritic spine specializations (Altman and Anderson, 1972, 1973; Herndon, Margolis, and Kilham, 1971; Herndon, 1968; Llinas, Hillman and Precht, 1972; Hirano‘ and Dembitzer, 1973, 1975; Hirano and Jones, 1972; Hirano, Dembitzer and Jones, 1972; Mouren-Mathieu and Colonnier, 1969, Sotelo, 1973, 1975). Cycasin (methylazoxymethanol glucoside) and MAM (methylazoxymethanol acetate), the agent used in this study, have been shown to destroy actively dividing cells within the external differentiating cell layer when injected into mice, rats, or hamsters, resulting in a decrease in parallel fibers and their synaptic contacts (Hirano, gt 31., 1972; Hirano and Jones, 1972; Jones, Yang and Michelsen, 1972a; Jones, Michelsen and Yang, 1973). Concomitantly with aberrant synaptogenesis, the administra- tion of MAM results in physical, histological and ultra- structural alterations by the tenth postnatal day. Such changes can be monitored by assessing the physical maturity, macrosc0pic changes within the cerebellar vermis, and histological alterations in the cerebellar cortex of 10 day treated and control mice. Questions concerning synapto- genesis such as the development and durability of the naked 3 Purkinje cell dendritic Spine specializations which result from MAM treatment, can be addressed by comparison of Similarities and differences of ultrastructural character- istics of MAM and saline groups. The physical and behavioral maturity of zero day cycasin-treated mice has been previously reported at dif- ferent postnatal ages (Jones, 33 31., 1973). In this experiment, sixty percent of 70 treated mice exhibited some aberrant physical, motor or postural changes, sometime during their initial 25 postnatal days. Recovery of func- tion occurred in less severely affected animals even though the cerebellum was severely damaged (Jones, 33 31., 1973). Studies in genetic mutants have Shown an appearance of cerebellar dysfunction before synaptogenesis is completed in the Weaver mouse (Rakic and Sidman, 1973c). These studies have supported the need for assessments of physical and behavioral characteristics like those carried out in this investigation. The definitive histologic changes seen in MAM- injected mice consist of destruction within the external differentiating cell layer, which causes a concomitant decrease in the foliation in midsagittal sections of the cerebellum (Chanda, 3E 31., 1973; Jones, 33 31., 1972a, 1973; Woodward, 22,§l°' 1975). A reduction in the size of the vermis has also been observed in animals subjected to a virus infection of feline panleukopenia (Herndon, 31 31., 4 1971, Llinas, 33 31., 1972), genetic defects (Caddy and Biscoe, 1975; Hirano and Dembitzer, 1975; Rakic and Sidman, 1973b, 1973c), and x-irradiation in rats (Altman and Anderson, 1971, 1972; Altman, 33 31., 1971; Anderson and Altman, 1971). A planimetric evaluation of the midsagittal sections of the vermis in rats was made by Woodward, 33 31. (1975). This method was used to compare the size reductions observed in midsagittal sections of the vermis in 10 day MAM injected mice and the size reductions previously reported in treated rats. Modifications of this method were used in this study to document the extent of external differentiating cell layer destruction after MAM treatment. 1 Random dislocation of Purkinje cell somata has been Shown in animals that have exhibited an external differen- tiating and internal granule cell loss due to genetic defects (Hirano and Dembitzer, 1975; Rakic and Sidman, 1973b; Sax, Hirano and Shofer, 1968), virus infection with feline panleukopenia (Herndon, 33 31., 1971; Llinas, 33_31., 1972), x-ray (Altman and Anderson, 1971, 1972; Altman, Anderson, and Strop, 1971; Anderson and Altman, 1971), and MAM-treatment (Chanda, Woodward and Griffin, 1973; Herndon, 33 31., 1971; Jones, 33 31., 1972a, 1973; Shimada and Langman, 1970; Woodward, 33 31., 1975). Various theories have been prOposed to explain the Purkinje cell apical pole orientation observed in 10 day saline and MAM-injected' mice. 5 With respect to number and orientation, changes in Purkinje cell primary dendrites have been described. Woodward, 33 31. (1975) suggested a plausible hypothesis for the formation of multiple primary dendrites. The presence of many massive Purkinje cell primary dendrites in MAM-injected mice will be discussed in relation to this hypothesis.' The altered number and random three dimen- sional orientation of secondary and tertiary branches and branchlets with spines on Purkinje cell somata have been previously reported in a variety of animals (Altman and Anderson, 1971, 1972b, l972d; Hamori, 1969; Mouren-Mathieu and Colonnier, 1969; Shofer, Pappas, and Purpura, 1964; Woodward, 33 31., 1975). The Purkinje cell dendrite and its dendritic processes will be described after MAM treatment. In a developmental and comparative light microscopic study, Swiss albino mice were injected with cycasin (0.5 mg/gm body weight), MAM (0.05 ul/gm body weight) or physio- logical saline (0.05 ul/gm body weight) within the first I 24 hours after birth. As early as Six hours after injec- tion with the MAM, cellular necrosis was evident within the external differentiating cell layer. Cycasin and MAM- injected mice both exhibited extensive necrosis in the external differentiating cell layer by two days post- injection. By the fifth post-injection day, both the external differentiating and internal granule cell layers 6 were markedly diminished, while the misaligned Purkinje cells were scattered in the molecular layer (Jones, 33_31., l972a, 1973). External differentiating cells that escape MAM-destruction may undergo prolific regeneration to replenish the external differentiating cell layer. The cells within this layer will eventually migrate inward to repopulate the internal granular cell layer (Chanda, 33 31., 1973; Shimada and Langman, 1970; Woodward, 33 31., 1975). The previously described destruction as well as the presence of persistent necrotic debris, and subsequent regeneration within the cerebellar cortex will be discussed and con- trasted to that observed in the 10 day MAM-injected mice. Bergmann fibers appear to be implicated in the migra- tion of differentiating granule cells by maintaining a side- by-Side aposition, thereby directing the granule cells toward their definitive location in the internal cell layer (Rakic, 1971; Sotelo and Chargeux, 1974). Anatomical variations in the Bergmann astroglia observed in this study will be compared and contrasted to research previously reported. Also, the presence or absence of growth cones, necrotic debris and whorls that inhabit areas of the external differentiating, molecular and internal granule cell layers at some stage in development of the cerebellar cortex, will be compared and contrasted to structures observed within control and MAM-injected mice. LITERATURE REVIEW In early studies on the carcinogenicity of cycad meal and cycasin (methylazoxymethanol glucoside), Hirano and Shibuya (1967) accidentally found that mice injected with cycasin (0.5 mg/gm body weight) at day zero exhibited physical signs due to a neurological deficit. They reported that the mortality rate in mice injected with cycasin_was 15% by postnatal days 5-15, and 33% by postnatal days 20-33. Eighty percent of the mice Showed ataxia and posterior paralysis by postnatal day 20, and these Signs were irreversible by postnatal day 33. PHYSICAL EXAMINATION In experiments to determine the effects of cycasin or MAM on neonatal Swiss albino mice, many physical changes were observed (Jones, 1971). These experiments showed that there was a delay in the appearance of fur or the opening of the eyes by a few days, even though the climbing abilities or righting abilities were not impaired (Jones, 1971; Jones, 33 31., 1972a; Jones, 33 31., 1973). Jones, 33_31., (1973) performed sequential neurological evaluations through 25 days of age on cycasin-injected mice. Sixty percent of the treated mice exhibited some aberrant physical, motor or poStural changes sometime during their initial 25 7 8 postnatal days, with only thirty percent remaining sympto- matic at 25 days of age. By ninety days, none of the less severely affected cycasin treated mice showed symptoms, indicating an apparent recovery from their previous neuro- logical deficits. Lane (1964) described Weaver mutant mice which exhibited clinical neurological Signs similar to those observed in cycasin and MAM-treated mice (Jones, 33 31., l972a). At about eight to ten days the first clinical neurological Signs in the Weaver were exhibited with the instability of gait. After several steps the animal would fall to one Side. At the second or third week the incoordi- nation abated Slightly with more stable locomotion. The adults appeared to compensate for their neurological defect by lowering their bodies to the surface they were upon, and widening their stance in order to maintain an upright position. Their limbs were poorly coordinated, especially the hind limbs. They also exhibited a fine rapid tremor of the extremities and trunk that was superimposed upon their Slow swaying movements. The neurological signs in the Weaver do not appear to be worsened with age (Lane, 1964). Staggerer mice, another neurological mutant, exhibits partial recovery from neurological deficits. These mice are physically smaller than normal mice during their first postnatal week, and exhibit unsteadiness when attempting to 9 walk. This unsteadiness is even more apparent during the second and third weeks, with the mice exhibiting a transient hyperextension of the limbs, and a reduction in activity and vigor which leads to ataxia associated with a slight tremor of the limbs. In the adult stage, Staggerer mice are physically smaller and less active, with poor self-grooming habits, but their tremor appears to abate (Sax, Hirano and Shofer, 1968). A tendency towards partial recovery from cerebellar deficits has also been observed in rats irradiated during the first postnatal days (Altman, 33 31., 1971; Anderson and Altman, 1972), in acute azide poisoning of monkeys (Mettler and Sax, 1972), and in hemicerebellectomized rats (Smith, Parks, and Lynch, 1974). A theory for the reduc- tion of cerebellar deficits suggests that the recovery is due to the great plasticity in the neocortex. This plas- ticity allows for the formation of new or altered synaptic contacts on the presumably remaining pathways, thereby compensating for partial cerebellar damage (Altman, 33 31., 1971; Smith, 35 31., 1974). MACROSCOPIC EXAMINATION The definitive histologic changes seen in the cere- bellums of MAM-injected mice consisted of destruction within the external differentiating cell layer (Jones, 33 31., 1972a). This destruction caused a concomitant decrease in the foliation as observed in midsagittal sections of the 10 cerebellum (Chanda, 33 31., 1973; Woodward, 33 31., 1975). Woodward, 33 31. (1975) planimetrically measured the sagittal sections of the vermis in rats at varying post- natal ages after injecting them with MAM (10 mg/Kg body weight) for four consecutive postnatal days. At day four, sagittal sections of the cerebellum were only 12% less in size than those of the control. At day seven, there was a 35% decrease in the MAM-injected rat sagittal vermis areas. At day 14, the reduction was diminished to 29%. Reduction in midsagittal sections of the vermis has also been reported in various animals due to: a virus infection with feline panleukopenia (Herndon, 33 31., 1971; Llinas, 33 31., 1972; Margolis and Kilham, 1970); genetic defects in Weaver mice (Hirano and Dembitzer, 1975; Rakic and Sidman, 1973b, 1973c); Staggerer mice (Hirano and Dembitzer, 1975); Lurcher mice (Caddy and Biscoe, 1975); x-irradiation in rats (Altman and Anderson, 1971, 1972; Anderson and Altman, 1972; Altman, 33 31., 1971); inhala- tion of ethanol (Bauer-Moffet and Altman, 1975); and hyper- thyroidism in rats (Lauder, Altman and Krebs, 1974). A 26% midsagittal vermis reduction has been observed within rats chronically subjected to high levels of ethanol (Bauer- Moffett and Altman, 1975). Of the genetic mutants, the heterozygous Weaver (+/wv) has shown a 5-10% midsagittal vermis reduction by the fourth postnatal day (Rakic and Sidman, 1973b), which increased to 20% at postnatal day 21 11 (Rakic and Sidman, 1973c). The homozygous Weaver (wv/wv) has concurrently Shown a 70% reduction in Size by postnatal day 21. The Lurcher mouse midsagittal vermis reduction has not been reported; but, published line drawings of the vermis show a size reduction similar to that observed within 'the homozygous Weaver. Sixteen day kittens, irradiated (200r) during the first two postnatal weeks, have exhibited up to a 58.5% midsagittal reduction in the size of the cerebellar ansiform lobule. The estimated maximal reduction of granule cells was 64% with 400r, while the same dosage destroyed very few Purkinje cells (Altman, Anderson and Wright, 1967). This concurs with the previously discussed studies showing that most of the reduction has been due to granule cell loss. Purkinj3 cells 1. Disorientation of Purkinje cell somata The previously described pathological processes that reduce midsagittal vermis areas, subsequently dislocate Purkinje cell somata which come to lie within the molecular and internal granule cell layers (Altman, 33_31., 1971; Anderson and Altman, 1972; Herndon, 33 31., 1971, Hirano and Dembitzer, 1975; Rakic and Sidman, 1973b, 19730; Sax, 33 31., 1968). Two theories have arisen that describe the dislocation of the Purkinje cell somata. In the first theory, Woodward, 33 31. (1975) proposed that MAM administered to newborn 12 mice causes reduction in the number of external differenti- ating cells, which Slows the production of internal granule cells. The decreased cell production delays the "folial expansion" that is essential in providing sufficient area for the development of a normal Purkinje cell layer. A second theory suggests that a glial reaction and rapid "astrocyte swelling", in response to external differenti- ating and internal granule cell degeneration and necrosis, cause the misalignment and disorientation of Purkinje cell somata (Jones and Gardner, 1976). 2, Purkinje cell apical pole orientation Altman and Anderson (1972) hypothesized that when the external differentiating cells and internal granule cells are destroyed by irradiation, the Purkinje cell changes seen are due to the absence of the interneurons and are not due to the direct damage produced by the radiation. With the destruction of the external differentiating cell layer the apical poles of the Purkinje cell somata which normally point towards the surface become randomly oriented. Altman and Anderson (1972) observed that the growth of the apical pole of the Purkinje cell "is an autonomous event, while the normal orientation of this growing structure depends on the presence and location of the external differentiating cell layer." With the autonomous growth, the random orientation of the Purkinje cell, primary dendrites, somata, and the 13 unusual shapes of the arborizing dendrites can be explained (Altman and Anderson, 1972; Woodward, 33 31., 1975). 3. Multiple and massive primary dendrites Woodward, 33 31. (1975) suggested a hypothesis for the formation of multiple primary dendrites. In the absence of parallel fibers due to the granule cell destruction, dendritic growth may spread from several early Purkinje somatic projections. With the late appearance of large numbers of parallel fibers in treated animals, one or more Purkinje cell processes may develop into a dendrite with! secondary and tertiary branches. If a particular dendrite does not reach a group of parallel fibers, multiple primary dendrites continue to grow and become progressively more difficult to reabsorb the larger they become. In the absence of a favorable environment, the immature growing process of the dendrites (filopodia) will attempt to reach one. After growing farther from the soma into a more suitable environment, the dendrite will then mature into secondary and tertiary branches and branchlets with spines (Woodward, 33 31., 1975). The same authors illustrated the random orientation of dendrites that in some cases become abnormally oriented perpendicular to (rather than within) the sagittal plane. After the initial cellular destruction by MAM or x-ray, the few remaining external differentiating cells begin their regeneration. 14 The reorientation appears to be towards the regenerating cells (Altman and Anderson, 1972; Woodward, 33 31., 1975). After irradiation of the cerebellum, Altman and Anderson (1972) observed many thick blunt spines on massive primary dendrites forming synapses with climbing and mossy fibers and "pseudosynapses" with glia cells. They attrib- uted the lack of spiny branchlets due to the paucity of granule cells and their parallel fibers projecting into the molecular layer. They suggested that spine formation is dependent upon the presence of parallel fibers and is not an autonomous process. Golgi epithelial (Bergmann) astroglia As Bergmann astroglia develop, they send vertical ascending processes from the Bergmann fibers straight through the neuropil of the molecular layer toward the pial surface. During further differentiation, the vertical processes develop "excreSCences" 1 - 2 um long with drum- stick terminal enlargements in close proximity to passing parallel fibers (Palay and Chan-Palay, 1974). Each expan- sion comprises a sheath which encapsulates a Purkinje cell dendritic spine except for the area of its synaptic attach- ment with a parallel fiber (Rakic, 1971). Cerebellar abnormalities seen in Weaver mice may be associated with a reduced rate of granule cell migration (Rezai and Yoon, 1972). Rakic and Sidman (1973a, 1973b) proposed that the defective neuronal migration is secondary 15 to maldevelopment of Bergmann glia. Bignami and Dahl (1974) indicated that there is "an abnormality on the surface of the Bergmann fibers, preventing 'recognition' from migrating neurons, rather than the physical absence of these fibers serving as guidelines for neurons migrating through the molecular layer." External differentiating plus internal granule cell reduction and regeneration Methylazoxymethanol glucoside (cycasin) and methyl- azoxymethanol acetate (MAM) have been Shown to selectively destroy dividing microneurons when injected into mice, rats or hamsters (Calvet, Drian and Privat, 1974; Chanda, 33 31., 1973; Chanda, Woodward, and Griffin, 1975; Hirano, 33 31., 1972; Hirano and Jones, 1972; Jones, 33 31., 1973; Jones, Sweeley, and Yang, 1972b; Jones and Gardner, 1976; Shimada and Langman, 1970; Woodward, 33 31., 1975). In developmental and comparative light microsc0pic studies Jones, 33_31. (1972a, 1973) injected Swiss albino mice with cycasin (0.5 mg/gm body weight), MAM (0.05 ul/gm body weight) or physiological saline within the first 24 hours after birth. AS early as six hours, cellular necrosis was evident within the external differentiating cell layer in MAM-injected mice. Cycasin and MAM-injected mice both exhibited extensive necrosis in the external differentiating cell layer by three days postinjection. 'By the fifth post- injection day both the external differentiating and internal 16 granule cell layers were markedly diminished, while the misaligned Purkinje cells were scattered in the molecular and internal granule cell layers. The regeneration has also been described in MAM-treated hamsters (Shimada and Langman, 1970) and is similar to the cytological events described after x-irradiation (Altman, Anderson and Wright, 1969). Chanda, 33 31. (1973) con- firmed that regeneration had indeed taken place, by noting an increasing amount of DNA, RNA and protein during the recovery phase. They raised the question about the degree of function associated with regeneration. SYNAPTOGENES IS Synapse formation begins when transient contacts of growing axonal processes stop their migration and form permanent synaptic sites, composed of a synaptic cleft which joins the pre- and postsynaptic elements (Cotman and Banker, 1974). What is the structure of a mature axoden- dritic synapse? What are the ultrastructural features of parallel fiber-Purkinje cell dendritic spine synapses? How is the synaptic complex formed? What alterations of the synaptic complex are observed in pathological conditions? Structure 33 3 mature axodendritic synapse A mature axodendritic synapse is the site for trans- miSsion of impulses from an axon to a dendrite. The synapse consists of three structures: the synaptic cleft, a pre-, and post-synaptic element. The pre- and postsynaptic 17 elements are joined by a synaptic cleft approximately 160 A in width, with intercleft lines on each side of the cleft (60 A in width), resulting in a gap from pre- to postsynap- tic elements of approximately 225 to 300 A (Akert, 33_31., 1972). The cleft contains either fine, or smaller tufted densities, tipped with a knob that sometimes extends to the middle of the cleft but not entirely across it, or both of these structures (Gray, 1966). The presynaptic element is a dilated structure that contains several mitochondria and synaptic vesicles, 300 to 500 A in diameter (Gray and Guillery, 1966). These vesicles aggregate near the presynaptic thickening, which contains regularly spaced dense projections that appear to make up a lattice structure. The postsynaptic element consists of an electron dense thickening that may extend up to 200-500 A into the postsynaptic cytoplasm (Akert, 33 31., 1972; Gray and Guillery, 1966; Cotman and Banker, 1974). Gray (1959) described two types of synapses and distinguished them according to the width of the post- synaptic density. Type I is an asymmetrical synapse (like those observed at parallel fiber-Purkinje cell dendritic spine junctions) with a postsynaptic density approximately 200-500 A in width. The symmetrical Gray type II synapse is Similar to the type I, except that the postsynaptic density is 100-200 A in width. 18 Ultrastructure 33 Purkinje cells, granule cells, and their processes The mature Purkinje cell soma is 20-40 um in diameter. Its cytoplasm is rich in granular and agranular endOplasmic reticulum, hypolemmal cisternae (which are about 600 A beneath the plasmalemma and parallel to it), Golgi appara- tus, lysosomes, mitochondria, neurotubules, and neuro- filaments. The dendritic tree arises in one to four primary dendritic trunks from the apical pole of the cell and passes in a sagittal plane at right angles to the cerebellar folial surface. The purkinje dendritic tree contains primary, secondary (1-7 um in diameter at some points of bifurcation), and tertiary dendrites and branch- lets (0.5-2 um in diameter). Numerous spines project from the tertiary dendrites and branchlets. At its initial seg- ment, the Purkinje cell axon looks like an elongated part of the cell body because it contains not only neuro- filaments, but also endOplasmic reticulum, and numerous ribosomes that are usually absent in an axon. As the axon passes toward the medullary portion of the cerebellum, it sends off collaterals to the Lugaro cells, other Purkinje cells, Golgi II cells, and granule cells (Herndon, 1964; Palay and Chan-Palay, 1974). Granule cells that have completed their migration into the inner granule cell layer have perikarya about 7-10 um in diameter in mice and rats. A large nucleus is 19 surrounded by a thin rim of cytoplasm (0.1 to 0.3 um thick). Chromatin granules form clumps which are scattered through the nucleoplasm. The thin rim of cytoplasm contains: free ribosomes, a few mitochondria (0.1-0.3 by 2-4 um), membranes of endoplasmic reticulum (0.2-2.0 um in diameter), Golgi apparatus (on one side of the cell where there is an expansion of cytoplasm), and no Nissl substance. The unmyelinated ascending axon (0.1-0.3 um in diameter) of the granule cell passes upward to the molecular layer where it bifurcates to become the parallel fibers (0.1-0.2 um in diameter). At 1.7 to 2.5 um intervals along the parallel fibers there are varicosities (0.5-1.0 um in diameter) which contain a loose collection of round synaptic vesicles. Normally, these varicosities articulate with Purkinje cell dendritic spines, basket cells, stellate, or Golgi cell dendrites (Gray, 1961; Herndon, 1964; Palay and Chan- Palay, 1974). Parallel fibers average 0.6 mm in length in the mouse and synapse with an average of 29 Purkinje dendritic Spines. Two hundred thousand to three hundred thousand parallel fibers course through each Purkinje cell arborization. The granule cell dendrites, located in the internal granule cell layer, are rather short, and synapse with the mossy as well as Golgi II axons and dendrites (Palay and Chan-Palay, 1974). The_junction between the parallel fiber and Purkinje cell dendritic spines is a Gray's type I asymmetrical 20 synapse. The synaptic cleft is widened to about 300 A, with occasionally occurring fine filaments crossing the synaptic cleft (Palay and Chan-Palay, 1974). The pre- synaptic parallel fiber varicosities that synapse with the Purkinje cell dendritic spines contain synaptic vesicles (240-440 A in diameter). These synaptic vesicles aggregate among tufted densities interwoven filaments (50 A in diameter), that extend from the axolemma. Mitochondria and microtubles are also found within the presynaptic vari- cosities. Thepostsynaptic elements are composed of spines attached to dendritic branchlets. They are 1.0-2.0 um long, witthulbous heads 0.4-0.6 um across on a stalk 0.2- 0.3 um but they occasionally contain cisternae and tubules among the fine filamentous matrix. The postsynaptic density (the postsynaptic element) is an electron dense structure that may extend up to 440 A into the post- synaptic cytoplasm (Palay and Chan-Palay, 1974). In adult animals, Bergmann astroglial processes that normally sur- round Purkinje cell dendrites, also encompass parallel fiber-Purkinje cell dendritic spine synapses. Synaptog3nesis: development 33 parallel fiber-Purkinje cell contacts Larramendi (1969) described parallel fiber-Purkinje cell dendritic spine synaptogenesis as taking place in stages in 14 day Swiss albino mice. In the first stage, 21 immature parallel fibers and ascending axons from the granule cells appear thicker in diameter than mature parallel fibers. They are densely packed without any inter- vening glia, contain numerous microtubules, and lack synaptic enlargements or swellings. As parallel fibers and Bergmann astroglia mature, glial processes associated with Purkinje cell dendrites expand to encompass nearby parallel fibers or axonal bundles. Numerous long, slender spines without membrane specializations, observed protruding from the Purkinje cell dendrites, grow in between the parallel fibers. After parallel fiber-Purkinje cell dendritic spines contact each other, membrane specializations form and a few synaptic vesicles adhere to the presynaptic site, barely enlarging the parallel fiber. Stage two, commences when the Purkinje cell dendritic spines appear to shorten and the well developed head slightly invaginates into the parallel fiber varicosity. The varicosity now contains many synaptic vesicles and a few mitochondria (Larramendi, 1969). Larramendi (1969) assumed that contact between these pre- and postsynaptic elements preceded the appearance of synaptic vesicles and pre- and postsynaptic membrane densities. Growth cones The early formation of "growth cones", suggested by del Cerro and Snider (1968) are recognized by a local accumulation of vesicles about 400-1100 A in diameter in an 22 area of the neuroblast near the cell membranes. A protru- sion of a short, thick process, containing numerous vesicles, precedes by elongation, with mitochondria and ribosomes moving into the proximal portion of the out- growing process. Within the molecular layer, both the Purkinje cell dendrites and parallel fibers end in "terminal growth cones"; "subterminal growth cones" precede spine formation (del Cerro and Snider, 1972; Johnson and Armstrong, 1970). Both terminal and sub—terminal growth cones are present in the Purkinje cell dendrites at 9-12 days and persist until adulthood (del Cerro and Snider, 1968). Naked Purkinje cell dendritic spine specializations Since Larramendi's report, naked Purkinje cell dendritic spine specializations have been shown to differen- tiate in the absence of a parallel fiber presynaptic element. This phenomenon is seen in various models when parallel fiber reduction is secondarily produced by cycasin or MAM (Hirano, 33 31., 1972; Hirano and Jones, 1972; Jones, 33 31., 1972; Jones, 33_31., 1973), viral infection with feline panleukopenia (Herndon, 33 31., 1971; Llinas, 33 31., 1972); gene mutations (Hirano and Dembitzer, 1973, 1975; Sotelo, 1973, 1975); x-ray (Altman and Anderson, 1972, 1973); or tissue culture (Seil and Herndon, 1970; Kim, 1975). Each pathological process has a different mechanism, but they all eventually cause the destruction 23 of the external differentiating cell layer and secondary reduction of internal granule cells, thereby producing an agranular cerebellum. In this setting, postsynaptic sites on Purkinje cell dendritic Spines appear to differentiate. Rats treated on postnatal days one through four with 333 (10 mg/kg body weight) exhibited extensive destruction of external differentiating cells during the first four postnatal days (Chanda, 33 31., 1973). At postnatal day seven the cerebellum in the treated animals exhibited a reduced foliation. By postnatal day 14 the external differentiating cell layer had regenerated and become thicker than in controls. By postnatal day 21 the external differentiating cell layer was almost totally absent due to the inward migration of cells, like in the controls (Woodward, 33 31., 1975). When cycasin (0.5 mg/gm body weight) and MAM (0.05 ul/gm body weight) were injected into zero day mice and examined 25 days postnatal, the molecular layer contained a significantly reduced parallel fiber population and a number of naked Purkinje cell dendritic Spine specializa- tions embedded in a matrix of astrocytic cytoplasm (Hirano, 33 31., 1972; Hirano and Jones, 1972; Jones, 33 31., 1973). Cats and ferrets injected at birth with feline pan- leukopenia virus exhibited naked Purkinje cell dendritic spine Specializations when sacrificed in adulthood. The virus primarily destroyed the external differentiating cell 24 layer and secondarily reduced this cell population. Deafferented Purkinje dendritic Spines were then enveloped by hypertrOphic glial cytoplasm of the Bergmann astrocytes (Herndon, 33 31., 1971). Hirano and Dembitzer (1973) examined the Purkinje cell spines of the mutant mouse, Weaver. In this mutant, parallel fibers may fail to form due to a reduced rate in the granule cell migration (Rezai and Yoon, 1972). Rakic and Sidman (1973a, 1973b) proposed that defective neuronal migration is secondary to maldevelopment of Bergmann glia. Bignami and Dahl (1974) indicated that there is "an abnormality on the surface of the Bergmann fibers." This prevents the 'recognition' of migrating neurons, that use the fibers as guidelines during their migration through the molecular layer. Naked Purkinje cell dendritic spine specializations were clearly demonstrated in this genetic mutant (Hirano and Dembitzer, 1973). Within another mutant, the Staggerer, the Purkinje cells are abnormal. Tertiary dendritic branches, branch- lets and Spines are missing. The normal target for parallel fibers, which occasionally form presynaptic grids (a normal presynaptic element with synaptic vesicles facing a Bergmann astroglial process) is thus eliminated. Sec- ondary degeneration of granule cells occurs (Hirano and Dembitzer, 1975; Sotelo, 1973; Sotelo and Changeux, 1974). Some Purkinje cell dendritic spines on tertiary branches 25 and branchlets do appear at 23 postnatal days but were mostly absent by three months. Naked Purkinje cell dendritic spine Specializations were seen only in 3 month mice (Hirano and Dembitzer, 1975). Prolonged géirradiation, like panleuk0penia, cycasin, and MAM, destroys cerebellar granule cells in rats before migration (Altman, 33 31., 1967; Altman and Anderson, 1971, 1972; Anderson and Altman, 1972; Hopewell, 1974; Woodward, Hoffer and Altman, 1974). Purkinje cells still form their characteristic dendritic branching pattern and even form naked Purkinje cell dendritic Spine Specializations (Altman and Anderson, 1972, 1973). Cultivation of rat and mouse cerebellums has resulted in the growth and development of various types of cerebellar neurons, including Purkinje and granule cells (Kim, 1975; Seil and Herndon, 1970). Kim (1975), Seil and Herndon (1970) have observed the concurrent development of naked Purkinje cell dendritic Spine specializations. When rat cerebellums are cultivated in the presence of MAM, naked Purkinje cell dendritic Spine specializations are produced in the almost total absence of any other cells (Calvet, 33 31., 1974). These experiments are Significant because they Show that naked Purkinje cell dendritic spine specializations form autonomously and in the absence of parallel fibers. The capacity to initiate production of these specializations may therefore be intrinsic to the postsynaptic cell. 26 The appearance of naked Purkinje cell dendritic Spine postsynaptic densities implies either degeneration of pre- synaptic elements after the formation of normal synaptic contacts, or the formation of these dendritic spine special- izations by some other mechanism. Kim (1975) relates that if the first hypothesis is to be substantiated, degenerating presynaptic terminals should be demonstrated. If another mechanism is responsible for the development of these structures, degenerating presynaptic terminals should not be demonstrated. One of three possible mechanisms might explain the formation of Purkinje cell dendritic spine specializations in the absence of a presynaptic element. The first mechanism is the endowment of Purkinje cells with the ability to form dendritic Spine specializations indepen- dent of presynaptic input (33 3333 formation, Altman and Anderson, 1972). A second mechanism would involve the formation of dendritic Spine Specializations after the stimulation of synaptic contacts on a Purkinje cell by afferents from climbing fibers or contacts other than granule cell parallel fibers (Hamori, 1973; Sotelo, 33 31., 1975). The third mechanism would involve the induction of Purkinje cell dendritic spine Specializations by stimula- tion of a small number of synaptic contacts between parallel fibers and Purkinje cell dendritic spines (Hirano and Dembitzer, 1974). 27 Durability 33 naked Purkinje cell dendritic §pine specializations Naked Purkinje cell dendritic Spine specializations have been reported to be durable structures remaining intact for long periods of time (up to a year) (Hamori, 1968, 1973; Herndon, 1968; Herndon, 33 31., 1971; Hirano and Dembitzer, 1975; Sotelo, 1973, 1975). A relatively recent theory proposes that the naked Purkinje cell dendritic Spine Specializations on tertiary branches and branchlets are maintained by climbing fiber contacts upon the primary and secondary Purkinje cell dendrites (Hamori, 1973). As a result, Sotelo (1975) in contrast to Hamori, suggests that naked Purkinje cell dendritic Spine specializations are autonomous in their appearance and stability, which supports the theory of 33 3333 formation and independent maintenance of naked spine Specializations. Necrotic debris Ultrastructurally, cellular necrosis within the external differentiating cell layer has been observed as early as six hours following MAM-administration with extensive necrosis present three days postinjection. By the fifth postinjection day the external differentiating and internal granule cell layers have been Shown to be markedly diminished with necrosis not evident after five days post- natal in MAM or cycasin-injected mice (Jones, 33_31., 1972a, 1973). Jones and Brownson (1969) observed that necrotic 28 debris in the normal differentiating hippocampus was sur- rounded by a pale watery cytoplasm similar to the appearance of astrocytes. This phenomenon was also observed after thiOphen necrosis by Herndon (1968). Cell degeneration and necrosis has been observed within Golgi cells from normal 10 day mice (Larramendi, 1969) and within normal fetal rat central nervous system. These may be normal processes for the removal of a vestigial organ or cells (Maruyama and D'Agostino, 1967). A host of methods previously described has been used for the destruction of cells of the cerebellar external differentiating or internal granular cell layers. Necrosis has been a prominent feature at some point in the process followed by the removal of cellular debris. This probably is accomplished by phagocytic cells which migrate from the blood vessels. However, astrocytes assume a phagocytic role, too (Herndon, 1968; Jones 33 31., 1973). Whorls Membranous whorls composed of smooth concentric membranes were occasionally observed within astrocytes (Palay and Chan-Palay, 1974; Sotelo and Palay, 1971; Hirano, 33 31., 1972). They were encountered in both foot processes and bulbous terminals of Bergmann astroglia as well as within Bergmann astroglial fibers. The concentric lamellated whorls composed of collapsed cisternae appeared to be derived from endoplasmic reticulum and mitochondria 29 (Palay and Chan-Palay, 1974; Sotelo and Palay, 1971). They resembled concentric laminar membranes observed in axon terminals. Smooth membranes have been shown to have the capability of producing intracellular concentric whorls and have been shown to occur also in a variety of non-neural tissues (Sotelo and Palay, 1971). They suggest that the overproduction of smooth membranes is characteristic of developing, growing or regenerating axons. In contrast, Spirals without any type of cistern separating the lamellar membranes have also been observed within saline and MAM- injected mice. Chan-Palay (1973) and Larramendi (del Cerro and Snider, 1969) believe that this is a fixation artifact, while del Cerro and Snider (1969) showed Similar structures in Purkinje cell dendrites and suggested the production was due to a toxic drug effect. METHYLAZOXYMETHANOL ACETATE (MAM) The initial studies of methylazoxymethanol glucoside (cycasin) were due to the interest in the toxicity appear- ing in sheep and cattle grazing on the cycad during long, hot, dry spells in Guam (Jones, 33 31., 1973). The result- ing experiments showed that a fatal neurological disorder occurred in cattle following ingestion of cycad leaves. Neurological diseases commonly observed in natives and their use of cycad flour as a food supplement, raised the question of associated central nervous system toxicity (Jones, 33 31., 1973; Whiting, 1963). Subsequent studies 30 failed to reveal an histopathologic basis for the neuro- logical disorders in cattle or to establish a definite connection with human neurological disease. Cycasin and its derivative, methylazoxymethanol acetate (MAM), were found to be hepatotoxic (Ganote and Rosenthal, 1968; Zedeck, 33 31., 1970), teratogenic (Spatz, Dougherty and Smith, 1967) carcinogenic (Hirano, 1969), and neurotoxic (Hirono and Shibuya, 1967) when administered to a host of animals. Jones and other investigators have shown that the neuro- toxic effects are confined to dividing cells in the central nervous system (Jones and Gardner, 1976). The glucoside or the acetate of methylazoxymethanol acts by methylating the 7 position of guanine in the brain nucleotides, causing disruption of DNA and RNA function '(Chanda, 33 31., 1975; Jones, 33 31., 1973; Matsumoto and Higa, 1966; Nagata and Matsumoto, 1969; Shank and Magee, 1967). MAM appears to have little neurotoxic effects on mature differentiated cells within the central nervous system since there are no cerebellar behavioral changes nor loss in cerebellar DNA if MAM-injections are given after 21 days of age in rats (Chanda, 33 31., 1975). Dividing cells are most susceptible to damage by MAM (Chanda, 33 31., 1975; Jones, 33 31., 1973). MATERIALS AND METHODS Newborn Swiss albino mice, Webster strain (obtained from Spartan Research Animals Inc., Haslett, Michigan) were selected randomly from Six litters (which had been reduced to 10 pups per litter) examined, weighed, marked for identification and injected with methylazoxymethanol acetate (MAM) (0.05 ul/gm body weight) or saline (0.05 ul/gm body weight) (Appendix A). Both (saline and MAM- injected) groups were housed in plastic shoebox cages containing pinechips. Controlled room temperature, humidity and a standard light-dark cycle were maintained. All mice in each group were sacrificed at 10 days of age after they were again weighed. Specific factors were assessed and recorded after physical examination (Appendix A). Eight mice from each group (the number of MAM-treated mice remaining alive at 10 days and the saline or MAM- injected group) were anesthetized with Nembutal (sodium pentobarbital, 0.05 mg/gm body weight). After thoracotomy and exposure of the heart, heparin (300 units/animal) and sodium nitrite (0.01 ml/gm body weight of a 1% solution) were injected into the left ventricle. The mice were then perfused (Appendix A) with 1% glutaraldehyde and 0.5% paraformaldehyde in a 0.12 M standard phosphate buffer containing 0.02 mM CaCl to establish a pH of 7.65 at 2 520 mOsM (using a technique modified from Palay and Chan- Palay, 1974). 31 32 Rapid Golgi impregnation: After perfusion for 20 - 30 minutes, three cerebellums from each group (saline or MAM-injected) were removed and immersed in a 2.33% potassium dichromate and 0.19% osmium tetroxide solution for ten days followed by one day in 0.75% silver nitrate at room tempera- ture (Rapid Golgi technique from Palay and Chan-Palay, 1974). They were then embedded in routine fashion in paraffin, sectioned at 75 - 125 um on a Sorval TC-2 tissue sectioner and mounted on Slides. Photomicrographs of the Golgi sec- tions and all other light microscopy were taken on a Zeiss PhotomicrOSCOpe II. A morphometric evaluation of sagittal sections of the vermis was made with the use of a Bausch and Lomb MicrOprojector and a Hartkop Counting Chamber (Appendix A), (modified from a Method according to Weibel, Kistler and Scherle, 1966). The remaining five mice from each group (saline or MAM-injected) were perfused for 20 - 30 minutes. The skulls were opened and the heads placed in fresh fixative at 4°C overnight. The following morning one-half of the cerebellum was postfixed in 10% buffered formalin, embedded in paraffin, sectioned, and mounted slides stained in routine fashion with Harris hematoxylin-eosin (Luna, 1968; Sheehan and Hrapchak, 1973) or Luxol fast blue-cresyl violet (Luna, 1968). The vermis from the second half of each cerebellum was postfixed for two hours in 2% osmium tetroxide in a 0.12 M standard phosphate buffer with a 3.5% dextrose, and 33 then rinsed in a 0.1 M sodium acetate solution. They were then stained 33f3133_with 0.5% uranyl acetate for 30 minutes at 4°C, dehydrated through graded alcohols, and embedded in Epon-araldite (modified from Palay and Chan- Palay, 1974). One micron thick sections were stained with 1% toluidine blue for study by light microscopy. Thin sections (600 - 800 A) were stained with lead citrate l - 2 minutes and uranyl acetate for 30 - 90 minutes according to the methods of Reynolds (1963), Venable and Coggeshall (1965). Ultrastructurally, the Purkinje cell dendrites, granule cell parallel fibers and their synapses were qualitatively assessed for the presence or absence of certain structures with the use of a Philips 201 electron microscope. The structures assessed were: naked Purkinje cell dendritic spines with a postsynaptic thickening, Purkinje cell dendritic spines, normal synaptic cleft of 200 - 300 A within a normal parallel fiber-Purkinje cell dendritic spine synapse, naked presynaptic varicosities, presynaptic terminals (varicosities on parallel fibers), growth cones, microtubules and postjunctional dense bodies (Appendix A, Electron Microscopic Data Sheet). The measure- ment of structures in the micrographs was accomplished with the use of a Hartkop Electron Microsc0pic Measurement Nomograph (Appendix A). 34 sooomozofiz zomeosqm onmmHzmfiEs omonnoem ouoneomm\\\\v JWWW\WoH:e E: H omzeon ems. ooamtzm mm>qu wmn.o . . empowers _ . vcoo zHee40 em oemommmm loomed: soon eo\H9-mo.oo oozasemxsxoosessemz sees-masomozn d mohz oszq4 meSm AsIBm2 024 HZHQCW M