Initiation and growth of ectopic neurites and meganeurites during postnatal cortical development in ganglioside storage disease

Developmental Brain Research, 51 (1990) 167-178 Elsevier

167

BRESD 51004

Initiation and growth of ectopic neurites and meganeurites during postnatal cortical development in ganglioside storage disease Steven U. Walkley 1, Henry J. Baker 2 and Mario C. Rattazzi 3 1Department of Neuroscience, Rose E Kennedy Center for Research in Mental Retardation and Human Development, Albert Einstein College of Medicine, Bronx, NY 10461 (U.S.A.), 2Department of Comparative Medicine, Bowman Gray School of Medicine, Wake Forest University, Winston-Salem, NC 27103 (U.S.A.) and 3Departments of Pediatrics and Research, North Shore University Hospital, Cornell University Medical College, Manhasset, NY 11030 (U.S.A.) (Accepted 25 July 1989) Key words: Ganglioside; Neurite growth; Dendrite; Pyramidal neuron; Cerebral cortex; Ganglioside storage disease

The incidence of cortical pyramidal neurons displaying meganeurites or enlarged axon hillocks with ectopic spines and neurites was evaluated developmentally using feline models of GM1 and GM2 gangliosidosis. Results of these studies demonstrated that the onset of ectopic neurite growth occurred after the elaboration of dendrites on cortical pyramidal neurons, and that the time of onset of this renewed dendritogenesis was similar in the two diseases, Initiation and growth of ectopic neurites also correlated in a general way with onset and progression of clinical deterioration in both diseases. In GM1 gangliosidosis there was a greater tendency toward formation of meganeurites, whereas in cats with GM2 gangliosidosis the growth of ectopic axon hillock neurites without meganeurites predominated. At end-stage disease in GM2 gangliosidosis, nearly 90% of pyramidal cells displayed some degree of axon hillock neurite growth as opposed to less than half this number for GM1 gangliosidosis cats at the same age. These data are consistent with the hypothesis that there are two separate driving forces behind these somadendritic abnormalities of pyramidal neurons in the gangliosidoses. Excessive intraneuronal accumulation of storage vacuoles accounts for the formation of meganeurites, whereas some type of intrinsic metabolic defect results in axon hillock neurite growth which in turn offers new surface area for synaptic input. Currently available data indicate that GM2 or GM3 ganglioside, or a closely related metabolic product other than GM1 ganglioside, may be primarily associated with the growth of ectopic dendritic processes on morphologically mature neurons in storage diseases.

INTRODUCTION T h e gangliosidoses are inborn errors of metabolism characterized by specific abnormalities in catabolic processes associated with ganglioside m e t a b o l i s m 13. Most c o m m o n l y this involves a defect in activity of a lysosomal hydrolase with subsequent, p r i m a r y storage of a particular ganglioside and secondary accumulation of a variety of o t h e r c o m p o u n d s . Golgi studies of ganglioside storage disorders, as well as o t h e r inherited and induced forms of neuronal storage disease in man and animals, have r e v e a l e d that specific types of neurons undergo distortion of somatic and axon hillock g e o m e t r y (meganeurite formation) and r e n e w e d dendrite growth 9"14"15'21"24"26-30. M e g a n e u r i t e s are distinct, storage vacuole-filled enlargements occurring specifically between the neuronal soma and axonal initial segment 25, and m a y or may not be a c c o m p a n i e d by ectopic spines or longer neuritic processes. Some neurons lack meganeurites but exhibit

ectopic spines or tufts of neurites on n o r m a l or slightly enlarged axon hillocks. New synapses have been docum e n t e d on m e g a n e u r i t e s ~5 and neurites 3°-32, and a variety of new connections have been suggested to contribute to this altered connectivity 22. T h e s e alterations in neuronal g e o m e t r y and connectivity have been hypothesized to underly the neuronal dysfunction which is characteristic of these diseases 15. M e g a n e u r i t e f o r m a t i o n and neurite growth have been found to be limited in distribution within the CNS of affected individuals, with cortical p y r a m i d a l neurons representing one type of cell chiefly involved in this process 22. A l t h o u g h m e g a n e u r i t e s and neurites have been found on cortical p y r a m i d a l neurons in most of the neuronal storage diseases studied to date ( G M 1 and G M 2 gangliosidoses 14"26"3°, sphingomyelin lipidosis type A 24, mucopolysaccharidosis type 128, and a - m a n n o s i d o s i s 27" 3~), it has been suggested that this process is most p r o n o u n c e d in the ganglioside storage disorders 15.

Correspondence: S.U. Walkley, Department of Neuroscience, Rose E Kennedy Center, 1410 Pelham Parkway South, Bronx, NY 10461, U.S.A. 0165-3806/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

168 In the p r e s e n t study the i n c i d e n c e of e c t o p i c n e u r i t e

of age, and the GM2 gangliosidosis cats, 46-225 days (see Table I). In each case, the older animals represented the end-stage for each disease, with the GM2 gangliosidosis cats demonstrating a shorter survival time than the GM1 gangliosidosis cats. Selected neocortical gyri (coronal, sigmoidal, marginal, suprasylvian, ectosylvian and sylvian) were used to evaluate the incidence of pyramidal cells with specific types of morphological changes. Only cells clearly identifiable as pyramidal (based on cortical location, dendritic and somatic: characteristics, and axonal trajectory) were evaluated, and these included both superficial (supragranular) and deep pyramidal neurons. For the counting procedure, specific cortical gyri were identified and scanned at 100-250x magnification. Individual Golgi-impregnated cells throughout all cortical laminae were examined more closely and each adequately impregnated pyramidal neuron was classified according to the presence or absence of a meganeurite, and of ectopic spines or neurites on axon hillocks or on meganeuritcs (Table I). When necessary, higher magnification (400-1000x) was used to classify cells more accurately. The number of cells counted in each case varied from 100 to 300, generally depending on the quantity and quality of the Golgi impregnations. Exact cell counts are given for each animal in Table 1 and are presented in histogram form in Fig. 6A.B.

g r o w t h and m e g a n e u r i t e f o r m a t i o n was e x a m i n e d develo p m e n t a l l y in feline m o d e l s o f G M I a n d G M 2 gangliosidosis. T h e o v e r a l l o b j e c t i v e s of this i n v e s t i g a t i o n w e r e to establish

the

t i m e of o n s e t o f e c t o p i c m e m b r a n e

p r o d u c t i o n r e l a t i v e to n o r m a l p y r a m i d a l n e u r o n d e n d r i t i c m a t u r a t i o n , to e v a l u a t e the o n s e t and d e v e l o p m e n t of m e g a n e u r i t e f o r m a t i o n a n d n e u r i t e g r o w t h in r e l a t i o n s h i p to p r o g r e s s i o n o f clinical n e u r o l o g i c a l disease, and to assess w h e t h e r any significant d i f f e r e n c e s exist b e t w e e n t h e t w o diseases. MATERIALS AND METHODS Animals used in this study were derived from breeding colonies of cats known to be heterozygous for GM1 or GM2 gangliosidosis. The former disorder is characterized by deficiency of the lysosomal hydrolase, fl-galactosidase, and intraneuronal storage of GM1 ganglioside, and is similar to human juvenile GM1 gangliosidosis L ~0. The latter is characterized by a complete defect of lysosomal fl-hexosaminidase activity and storage of GM2 ganglioside, and is homologous with GM2 gangliosidosis type II (Sandhoff's disease) TM. Homozygous mutants used in this study were identified by a tail-tip or skin biopsy and enzyme assay shortly after birth and were maintained in the research colony until termination. All mutants were monitored carefully for onset and progression of neurological deficits and prior to euthanasia each animal was evaluated neurologically. Pentobarbital overdose was followed by perfusion with aldehydes or by whole brain immersion in 10% formalin for tissue fixation. Tissues were subsequently processed for Golgi rapid, Golgi-Kopsch, or Golgi-chloral hydrate procedures; specific details of these methods are given elsewhere 2~. Phenotypically normal littermates and other age-matched normal cats served as controls. Brains were cut mid-sagittally and sliced in the coronal plane for Golgi impregnation and identification of cortical gyri. The GM1 gangliosidosis cats used in this study were 24-325 days

RESULTS

Clinical progression o f disease K i t t e n s i d e n t i f i e d e n z y m a t i c a l l y shortly a f t e r birth as G M 1 o r G M 2 gangliosidosis m u t a n t s w e r e b e h a v i o r a l l y i n d i s t i n g u i s h a b l e f r o m l i t t e r m a t e s with n o r m a l o r interm e d i a t e e n z y m e levels. G r o w t h r a t e in the G M 2 gangliosidosis kittens a p p e a r e d

less t h a n that of n o r m a l

l i t t e r m a t e s by 4 w e e k s o f age, w h e r e a s t h e kittens with G M 1 s t o r a g e disease r e m a i n e d similar in size to their l i t t e r m a t e s until 6 m o n t h s o f age. T h e earliest signs of n e u r o l o g i c a l i m p a i r m e n t w e r e s e e n in the G M 2 ganglio-

TABLE I

Incidence of specific changes in cortical pyramidal neuron morphology in GMI and GM2 gangliosidosis mutants at various ages n, number of pyramidal cells counted; A - E , morphological changes as illustrated in Fig. 7. A, cells characterized by somata and axon hillocks of normal morphology; B, cells with enlarged axon hillocks but lacking spine or neurite growth; C, cells with meganeurites lacking spine or neurite growth; D, cells with meganeurites which also exhibit spines and/or neurites; E, cells with normal or enlarged axon hillocks which also display spines and/or neurites.

Age (days)

24 31 43 46 76 95 107 152 168 173 224 225 248 272 325

Disease type

GM1 GM1 GM1 GM2 GM1 GM2 GM2 GM2 GM2 GM1 GM1 GM2 GM1 GM1 GM1

n

100 100 200 100 300 150 200 200 200 175 200 200 300 200 200

Percent distribution of morphological changes A

B

C

D

E

100.0 99.0 84.0 90.0 81.0 77.0 42.5 3l .5 38.0 27.0 14.0 8.0 7.0 9.5 7.(7

0.0 0.0 6.0 0.0 9.0 0.0 0.0 6.0 2.5 22.0 17.0 3.0 6.0 5.5 2.0

0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.5 10.0 9.5 0.0 13.0 3.0 7.0

0.0 0.0 0.0 0.0 0.0 0.0 2.0 0.0 2.5 11.0 20.5 4.0 49.0 17.5 67.(I

0.0 1.0 10.0 10.0 10.0 23.0 55.5 62.5 56.5 30.0 39.0 85.0 25.0 64.5 17.0

169 sidosis kittens at 6 weeks, at which time careful observation revealed slight head and hindleg tremors. The GM1 gangliosidosis kittens showed similar signs at onset, but these did not occur until 12-14 weeks of age. By 13-15 weeks, the GM2 gangliosidosis animals demonstrated conspicuous neurological impairment consisting of mild ataxia and awkwardness in motor acts, as well as increased head, truncal, and limb tremors. Similar signs in the cats with GM1 gangliosidosis did not appear until

16-18 weeks. By this age in GM2 gangliosidosis neurological deficits were marked, with increased ataxia and some tendency towards spasticity, especially in the forelimbs. By 22-24 weeks cats with GM2 gangliosidosis were generally unable to stand and spent more time in lateral recumbency. The GM1 gangliosidosis cats at this age were again much less affected by their disease although they did demonstrate marked ataxia, tremor and impaired vision. Exaggerated acousticomotor re-

\ A C

D B O

E

L

i

0

50~m

Fig.1. Camera lucida illustration of Golgi-impregnated pyramidal (A,C-E) and non-pyramidal (B) neurons in neocortex of a 31-day-old GM1 gangliosidosis cat. Cells were characterized by lush dendritic spine growth and occasional somatic spines or 'protospines' (see cell D, arrow), all of which were essentially like those seen in normal cats at this age. One deep pyramidal neuron (E, arrow), also possessed larger spine-like growths at the axon hillock, which was slightly enlarged. This type of change was not seen in normal tissues. Axons are denoted by a in this and subsequent camera lucida drawings.

170

\

B

A

C

50~m Fig. 2. Camera lucida illustration of Golgi-impregnated pyramidal neurons in neocortex of a 46-day-old G ~

g~mgliosidosis cat. The ceils appear

normal except for a possibly inappropriate distribution of numerous spine-like processes at the initial portion of the axon in cell A (arrow), and prolific axon hillock neurite growth on the deep pyramidal cell C (arrow).

sponsiveness (increased startle response) was also evident at this age, whereas this clinical sign was absent in GM2 gangliosidosis cats. Muscle mass, appetite and general condition remained good through onset of spastic quadriplegia at 7-9 months of age in the GM1 gangliosidosis cats. With considerable supportive care the animals with GM2 gangliosidosis could be maintained until 7-8 months of age, whereas those with GM1 gangliosidosis survived 11-12 months.

Meganeurite formation and neurite growth Golgi impregnation data from all animals are summa-

rized in Table I and Figs. 6 and 7. Golgi studies on cerebral cortex of the youngest animals available for this study (GM1 gangliosidosis kittens at 24 and 31 days of age) indicated that axon hillock neurite growth was initiated only after the normal elaboration of dendritic processes on cortical pyramidal neurons (Fig. 1). Although GM2 gangliosidosis kittens of this same age were not available, similar findings in GM1 and GM2 ganglioside storage disease at 43 and 46 days, respectively, suggest a similar, early course for the latter (Table t, Fig. 2). In both diseases, the first pyramidal neurons to display axon hillock neurite growth tended to be located

171 in deeper laminae (i.e. layer V) and to be larger in size than non-neurite bearing cells (Figs. 1E and 2C). Meganeurites were not encountered on any pyramidal neurons at this early age, but aspiny, enlarged axon hillocks were found on larger pyramidal cells of layers III and V in the

GM1 gangliosidosis cats. The vast majority (84-90%) of pyramidal neurons in these younger mutants appeared to be of normal morphology. A few cells also displayed spine-like processes over their cell bodies and/or large numbers of spines at the distal axon hillock-initial

C

E

B

50urn

Fig. 3. Camera lucida illustration of Golgi-impregnated pyramidal (A-C, E) and non-pyramidal (D) neurons in neocortex of a 95-day-old GM2 gangliosidosis cat. Several of the pyramidal cells shown have abnormal axon hillock-initial segment regions. Cell A displays spine-like processes (arrow), cell B short neurites (arrow) and cell E, longer neuritic processes (arrow).

172

D

A

C

Fig. 4. Camera lucida illustration of Golgi-impregnated pyramidal (A,B,D) and non-pyramidal (C) neurons in neocortex of a 168-day-old GM2 gangliosidosis cat. The axon hillock-initial segment regions of pyramidal cells are characterized by short spines (cell A, arrow), longer neuritic processes (cell D, arrow), or meganeurites with occassional spines (cell B, arrow). Intrinsic cells (C) most commonly appeared normal.

Fig. 5. Photomicrographs of Golgi-impregnated pyramidal neurons in neocortex of GM 1 gangliosidosis cats at different ages illustrating different degrees of morphological change. A: layer III pyramidal cell in a 173-day-old animal. Cell has an enlarged axon hillock (arrow) w ~ appears free of ectopic spines and neurites. B: layer II pyramidal cell in a 248-day-old animal. Cell has an aspiny meganeurite (arrow). C: deep layer III pyramidal cell in a 248-day-old animal. Cell has a large meganeurite (large arrow) which has numerous spines and sh0rt neurites on its surface (small arrows). D,E: layer V pyramidal cell (and an adjacent intrinsic cell in D, curved arrow), in a 43-day-old animal. The pyramidal cell has an enlarged axon hillock (large arrow) which is covered with short neuritic projections (small arrows). Bars in A,C, E = 25 ~m; bar in B = 15/~m; bar in D = 65/~m.

...j

174 segment region of the axon (Fig. 2A). The latter change generally appeared in excess of initial segment spines observed in normal kittens at the same age. By 76 days in GM1 gangliosidosis there was increased evidence of intrasomatic storage as axon hillock enlargements were more common, although the amount of axon hillock neurite growth, at least in this particular animal available for study, was not greater than that in the 43-day-old animal. Meganeurites again were not observed at this age. By 95 days in the next oldest GM2 gangliosidosis cat, 23% of cortical pyramidal cells exhibited ectopic spine or neurite growth on normal or enlarged axon hillocks (Fig. 3). Aspiny, enlarged axon hillocks, as were evident in the younger GM1 gangliosidosis cat, were not seen, nor were meganeurites. In

A

loo]

9o;

8o r..

FA g

6o

~

5o

"5

4o

°' t

~o~ i 20.

t 24-

51

4.3

ii¸ 76 Days

B

17,3 of

224

248

272

525

Age

No,~ot Uo,pho,o~y (A)

'oo 1 !

$p;ny Megoneur;te, (o) ~on Hillock Neurlte. (£)

rc~4

c •6o~

DISCUSSION

L

~o~ i

'°i oL

GM2 gangliosidosis mutants at 107 and 152 days, the incidence of axon hillock neurite growth increased dramatically to involve over half of the impregnated pyramidal neurons. Aspiny, enlarged axon hillocks, but not meganeurites, were also seen at the latter age. The GM2 gangliosidosis animal at 168 days of age demonstrated a similar incidence of neurite growth (56%), but also occasional meganeurites, and the majority of these possessed spines or neurites (Fig. 4). In the GM1 gangliosidosis mutant of a comparable age (173 days), meganeurites were found on 21% of the pyramidal neurons, but only half of these possessed spines or neurites (the others being aspiny), and axon hillock neurite growth was found on only 30% of Golgiimpregnated pyramidal neurons. These differences in the relative incidence of meganeurites and neurite growth in the two diseases were even more apparent at 225 days of age. At this time, 89% of cortical pyramidal neuron axon hillocks in the GM2 gangliosidosis mutant showed some degree of ectopic spine or, more often, neurite growth. Axon hillocks were commonly enlarged, but were not separated from the cell body by a neck and thus were not classified as meganeurites. Less than 10% of the observed pyramidal cells demonstrated normal morphology. In the GM1 gangliosidosis cat at about the same age (224 days), slightly less than 60% of pyramidal cells displayed spine or neuritebearing axon hillocks or meganeurites. Another 25% of the pyramidal cells displayed swollen axon hillocks or meganeurites, but these lacked spines and neurites. In the older GM1 gangliosidosis animals (248,272, and 325 days), the incidence of pyramidal neurons with axon hillocks or meganeurites with spines and neurites increased steadily (74%, 82%, and 84%, respectively). In contrast, aspiny meganeurites and enlarged axon hillocks decreased in incidence from a peak at 225 days (26%), and by terminal disease were found on less than 10% of pyramidal cells. The frequency of normal appearing pyramidal neurons remained about the same (7-9%) for all of these older GM1 gangliosidosis mutants, and was similar to that of the GM2 gangliosidosis cat at end-stage disease (8%).

46

m

95

107 Days

of

152 Age

168

225

Fig. 6. Histograms demonstrating the relative proportions of pyramidal cells with the morphological changes documented in Table I and illustrated in Fig. 7 for cats with GM1 (A) and GM2 (B) gangliosidosis at different ages.

This investigation illustrates the remarkable degree of ectopic membrane production which occurs at the axon hillock region of cortical pyramidal neurons in the gangliosidoses. By terminal disease in both GM1 and GM2 gangliosidosis, more than 90% of these neurons have axon hillocks characterized by spiny meganeurites and/or secondary neurites. Furthermore, the time of initiation of these changes in pyramidal neuron morphol-

175

A

B

C

D

E

Fig. 7. Schematic illustration depicting the likely paths (a-e) taken by pyramidal neurons which develop the particular morphological changes characteristic of ganglioside storage disease. A, cells of normal morphology; B, cells with enlarged axon hillocks lacking ectopic spines and/or neurites; C, cells with meganeurites lacking spines and/or neurites; D, cells with meganeurites which display ectopic spines and/or neurites; E, cells with slightly enlarged axon hillocks with prolific growth of neurites and spines. ogy essentially coincides with onset of neurological deficits. This distorted neuronal geometry and altered synaptic connectivity have been suggested to contribute to neuronal dysfunction 15, although a compensatory function for the latter has also been suggested 21. However, the recently described predominance of neuroaxonal dystrophy in GABAergic neurons in ganglioside storage disease may be more readily correlated with the genesis of brain dysfunction in these diseases 23. Although the studies reported here do not directly answer this controversy, they do fully document the early onset of meganeurite and neurite growth and the lack of obvious degenerative changes in the soma-dendritic domain of pyramidal neurons until late in the disease course. Indeed, pyramidal cell dendrites revealed lush spine coverage, as well as length and orientation comparable to that observed in normal animals. Spine-like processes also were observed on somata, axon hillocks, and axonal initial segments of the younger animals used in this study. Although such spine distribution is not abnormal at this age, the occurrence of distally-placed initial segment spines was unusual in that they appeared to exceed in number those seen in control animals. Their relationship

to the other changes described here is unclear, and in older animals they were observed less commonly. The dominant change in pyramidal neuron morphology in ganglioside storage disease occurs at the axon hillock and a schematic summary of the development of these changes is given in Fig. 7. In GM2 gangliosidosis at end-stage disease, the vast majority of Golgi-impregnated pyramidal neurons displayed ectopic, dendritic spine-covered membrane at the axon hillock area, and this new membrane occurred almost exclusively as neurite growth rather than meganeurite formation. In work reported elsewhere 3°, these ectopic neurites have been shown to be recipients of new, primarily asymmetrical, synapses which appear identical to those reported earlier for ectopic neurites in GM1 gangliosidosis 32 and a-mannosidosis 3~. The morphology of these synapses makes them unlikely to be GABAergic, and instead suggests an excitatory function (see ref. 22 for discussion). Cats with GM1 gangliosidosis at end-stage disease displayed a similar incidence of axon hillock abnormalities as the GM2 gangliosidosis mutant at end-stage, but the majority of these cells displayed spiny meganeurites rather than neurites. Thus it appears that although there

176 is significant production of ectopic membrane at the axon hillock in both diseases, the form which this membrane takes in each disease is considerably different. Indeed, the predominance of meganeurite formation with storage of GM1 ganglioside is somewhat surprising, given that the total amount of ganglioside in feline GM1 and GM2 gangliosidosis brains at end-stage has been reported to be similar 7. To account for this difference, it could be surmised that other non-ganglioside or non-lipid constituents must be accumulating preferentially with GM1 ganglioside in storage vacuoles. Alternatively, the GM2 ganglioside may be more readily shunted elsewhere into other regions of pyramidal cells (for example, into the plasmalemma of cell somata, dendrites, or ectopic neurites), or is accumulating predominately in non-pyramidal cells. The changes at the axon hillock in these two diseases can be made to appear equivalent if pyramidal neurons with aspiny axon hillock enlargements and aspiny meganeurites are added to that number of cells with neurites and spiny meganeurites. But as has been argued elsewhere 2~, enlarged axon hillocks and meganeurites, when lacking spines, may not be recipients of new, asymmetrical synapses and thus may only represent cell volume increase to accommodate storage. Golgi and ultrastructural studies of ceroid lipofuscinosis support this view, as meganeurites are aspiny and also free of new synapse formation 34. If this distinction is correct for ganglioside storage disease, it suggests the importance of the metabolic defect in GM2 gangliosidosis in terms of the induction specifically of neurite growth and concomitant formation of new synapses. The suggestion is that there are two, essentially independent, driving forces occurring simultaneously in these storage diseases. Firstly, increased membrane production allows for axon hillock expansion and meganeurite formation to accommodate increased intraneuronal volume. Secondly, membrane is generated in the form of ectopic neurites which are accompanied by spines and synapses. Whether the key element in this second driving force is neurite growth, per se, or surface area for new synapse formation is not clear at present, but the existence of spines and synapses on some meganeurites suggests that the latter (i.e. new synapse formation) may be the critical event 21. Numercrus studies indicate an effect of GM1 ganglioside on induction of neurite growth on neurons in culture 12'16. Recently, however, a variety of ganglioside species were compared for their neuritogenic capacity in Neuro-2a cells in GM1 ganglioside was found not to be significantly more able to elicit this change in cell morphology than other gangliosides such as GM2 and GM3 s. A more meaningful way to gain insight into a possible relationship between gangliosides and dendrite

growth may be to examine the characteristics of ectopic dendrite growth in neuronal storage diseases, including those with and those without primary or secondary disruption of metabolic pathways involving gangliosides. An underlying theme of this work is that the ectopic growth of dendritic membrane as seen in the neuronal storage disorders may simply be a recapitulation of the same process of dendritic initiation which occurred and then ceased to occur during late fetal development. If this is the case, it would be important to identify those metabolic events common to the neuronal storage diseases and to early brain development. As has been reviewed elsewhere 22, all neuronal storage diseases displaying ectopic neurite growth also exhibit evidence for changes in ganglioside patterns, and thus Purpura's original hypothesis suggesting a link between ganglioside metabolic defects and the process of renewed dendritogenesis remains cogent. In sphingomyelin lipidosis type A 2'33, mucopolysaccharidosis type 16, and in inherited and swainsonine-induced a-mannosidosis TM, abnormal patterns of gangliosides (specifically, elevations of GM2 and GM3) have been documented. Furthermore, in a-mannosidosis stacks of membranous leaflets consistent with the storage of gangliosides have been observed within the pathologic cytosomes of neurite-bearing pyramidal cells, whereas non-neurite bearing cells tend to have typical non-membranous, clear or floccule-filled inclusions 31. The data reported in this paper suggest that the driving force behind neurite growth is greater in GM2 than in GM1 gangliosidosis. Interestingly, feline GM1 gangliosidosis has been shown to be characterized by relative increases in GM2 and GM3 gangliosides 4, and synaptosomal preparations in GM1 gangliosidosis cats have been reported to have a high content of GM1 and GM2 gangliosides 35. Thus a common metabolic abnormality in all neuronal storage diseases characterized by ectopic neurite growth appears to be an increased relative amount of GM2, and possibly GM3, ganglioside. Indeed, GM1 ganglioside may not be elevated in all storage diseases with ectopic dendritogenesis which have been studied to date, and in the case of GM2 gangliosidosis it has in some cases been reported as decreased in relative and/or absolute quantities 711"~9. Other reports indicate small increases in GM1 ganglioside and large increases in GM3 ganglioside in GM2 gangliosidosis ~7. The same gangliosides known to be elevated in storage diseases with neurite growth (that is, GM2 and GM3) reach their peak levels concomitant with the period of initiation and early growth of dendrites during late fetal development, and are later found only in trace quantities in normal adult CNS 2°. This provides additional suggestive evidence linking the presence of GM2 and GM3 gangliosides, or closely associated metabolic products,

177 with the process of dendritic sprouting. A d m i t t e d l y , the m e c h a n i s m by which G M 2 and/or G M 3 ganglioside (or any ganglioside or r e l a t e d product) might participate in the initiation of dendrites is unknown. A l t h o u g h gangliosides have been suggested to act as acceptor molecules for n e u r o t r o p h i c substances 16, m o r e recent studies indicate that they m a y function as m o d u l a t o r s of receptors, r a t h e r than as receptors themselves. For example, G M 3 ganglioside m a y play a role as a m o d u l a t o r for e p i d e r m a l growth factor in fibroblasts 3. Thus, rather than acting directly as a neurite growth p r o m o t o r , certain gangliosides m a y act to facilitate the function of specific r e c e p t o r s for growth factors already present in brain. The elevation of these same gangliosides in neuronal storage diseases, p r e s u m a b l y not just in lysosomal storage vacuoles but i n c o r p o r a t e d into their n o r m a l p l a s m a l e m m a l d o m a i n s as well (e.g. see ref. 35), may (in the presence of r e c e p t o r s and growth factors) be linked in some way to r e n e w e d dendritogenesis. A l t e r n a t i v e l y , as m e n t i o n e d above, the driving force m a y be for new synapse f o r m a t i o n , possibly excitatory synaptic contacts 22, with ectopic neurite growth supplying the necessary surface a r e a for this input. T h e suggestion inherent in available data from the

n e u r o n a l storage disorders is that G M 2 and/or G M 3 gangliosides, rather than G M 1 , can be best c o r r e l a t e d with the initiation of dendritic neurites. This m a y only reflect relative potency, or alternatively, the subcellular location of specific gangliosides m a y be most critical in this process. Nevertheless, these and r e l a t e d findings are consistent with the possibility that the initiation of new dendrites on p y r a m i d a l neurons is influenced by the presence of particular gangliosides in their p l a s m a l e m m a . Certainly, it cannot be excluded that the presence of new dendrite growth itself leads to e l e v a t e d levels of these m i n o r gangliosides in storage disease, as might be suggested to explain their brief elevation during normal brain d e v e l o p m e n t . But it seems m o r e likely that the changes in ganglioside patterns in neuronal storage disorders are a direct consequence of the p r i m a r y metabolic defect, and that their correlation with renewed dendritogenesis suggests a pivotal role in the initiation of new d e n d r i t e growth and/or new synapse formation.

REFERENCES

pathol. Exp. Neurol., 22 (1973) 1-18. 11 Goldman, J.E., Yamanaka, T., Rapin, I., Adachi, M., Suzuki, K. and Suzuki, K., The AB-variant of GM2 gangliosidosis, Acta NeuropathoL, 52 (1980) 189-202. 12 Ledeen, R.W., Biology of gangliosides: neuritogenic and neuronotrophic properties, J. Neurosci. Res., 12 (1984) 147-159. 13 O'Brien J.S., The gangliosidoses. In J.B. Stanbury, J.B. Wyngaarden, D.S. Frederickson, J.L. Goldstein and M.S. Brown (Eds.), The Metabolic Basis of Inherited Disease, 5th edn., McGraw-Hill, New York, 1983, pp. 945-969, 14 Purpura, D.P. and Baker, H.J., Meganeurite and other aberrant processes of neurons in feline GM1 gangliosidosis, Brain Res., 143 (1978) 13-26. 15 Purpura, D.P. and Suzuki, K., Distortion of neuronal geometry and formation of aberrant synapses in neuronal storage disease, Brain Res., 116 (1976) 1-21. 16 Roisen F.J., Bartfeld, H., Nagele, R. and Yorke, G., Ganglioside stimulation of axonal sprouting in vitro, Science, 214 (1981) 577-578. 17 Rosengren, B., Mansson, J.-E. and Svennerholm, L., Composition of gangliosides and neutral glycosphingolipids of brain in classical Tay-Sachs and Sandhoff disease: more lyso-GM2 in Sandhoff disease?, J. Neurochem., 49 (1987) 834-840. 18 Siegel, D.A. and Walkley, S.U., Changes in ganglioside composition in inherited and swainsonine-induced ct-mannosidosis, J. Neuropathol. Exp. Neurol., 48 (1989) 350. 19 Suzuki, K., Suzuki, K., Rapin, I., Suzuki, Y. and Ishii, N., Juvenile GM2-gangliosidosis, Neurology, 20 (1970) 190-204. 20 Tettamanti, G., Brain gangliosides in development, Adv. Exp. Med. Biol., 13 (1971) 75-89. 21 Walkley, S.U., Further studies on ectopic dendrite growth and other geometrical distortions of neurons in feline GM1 gangliosidosis, Neuroscience, 21 (1987) 313-331. 22 Waikley, S.U., Pathobiology of neuronal storage disease. In J.R. Smythies and R.J. Bradley (Eds.), International Review of Neurobiology, Vol. 29, Academic, New York, 1988, pp. 191244.

1 Baker, H.J., Lindsey, J.R., McKhann, G.M. and Farrell, D.F., Neuronal GM1 gangliosidosis in a Siamese cat with fl-galactosidase deficiency, Science, 174 (1971) 838-839. 2 Brady, R.O., Sphingomyelin lipidoses: Niemann-Pick disease. In J.B. Stanbury, J.B. Wyngaarden, D.S. Fredrickson, J.L. Goldstein and M.S. Brown (Eds.), The Metabolic Basis of Inherited Disease, 5th edn., McGraw-Hill, New York, 1983, pp. 831-841. 3 Bremer, E.G. and Hakomori, S., Gangliosides as receptor modulators, Adv. Exptl. Med. Biol., 174 (1984) 381-394. 4 Byrne, M.C. and Ledeen, R.W., Regional variation of brain gangliosides in feline GM1 gangliosidosis, Exp. Neurol., 81 (1983) 210-225. 5 Byrne, M.C., Ledeen, R.W., Roisen, EJ., Yorke, G. and Sclafani, J.R., Ganglioside-induced neuritogenesis: verification that gangliosides are the active agents, and comparison of molecular species, J. Neurochem., 41 (1983) 1214-1222. 6 Constantopoulos, G. and Debakan, A.S., Neurochemistry of the mucopolysaccharidoses: brain lipids and lysosomal enzymes in patients with four types of mucopolysaccharidosis and in normal controls, J. Neurochem., 30 (1978) 965-973. 7 Cork, L., Munnell, J.E, Lorenz, M.D., Murphy, J.V., Baker, H.J. and Rattazzi, M.C., GM2 ganglioside lysosomal storage disease in cats with fl-hexosaminidase deficiency, Science, 196 (1977) 1014-1017. 8 Cork, L., Munnell, J.E, Baker, H,J. and Rattazzi, M.C., A comparative pathologic study of feline and human GM2 gangliosidosis. In H.M. Zimmerman (Ed.), Progress in Neuropathology, Vol. 4, Raven, New York, 1979, pp. 161-177. 9 Cummings, J.F., Wood, P.A., Walkley, S.U., De Lahunta, A. and DeForest, M.E., GM2 gangliosidosis in a Japanese spaniel, Acta Neuropathol., 67 (1985) 247-253. 10 Farrell, D.F., Baker, H.J., Herndon, R.M., Lindsay, J,R. and McKhann, G.M., Feline GM1 gangliosidosis: biochemical and ultrastructural comparisons with the disease in man, J. Neuro-

Acknowledgements. We wish to thank Mrs. Marie Buschke for her excellent technical assistance with the Golgi method and Dr. D. Siegel for discussion. This work was supported by NIH grants NS 18804 (S.U.W.), NS10967 (H.J.B.), and NS21404 (M.C.R.).

178 23 Walkley, S.U., Vulnerability of GABAergic neurons to neuroaxonal dystrophy in neuronal storage disorders, J. Neuropathol. Exp. Neurol., 48 (1989) 350. 24 Walkley, S.U. and Baker, H.J., Sphingomyelin lipidosis in a cat. II. Golgi studies, Acta NeuropathoL, 65 (1984) 138-144. 25 Walkley, S.U. and Pierok, A.L., Ferric ion-ferrocyanide staining in gangloside storage disease establishes that meganeurites are of axon hillock origin and distinct from axonal spheroids, Brain Res., 382 (1986) 379-386. 26 Walkley, S.U., Baker, H.J. and Purpura, D.P., Morphological changes in feline GM1 gangliosidosis. In EC. Rose and P.O. Behan (Eds.), Animal Models of Neurological Disease, Pitmans, London, 1980, pp. 419-429. 27 Walkley, S.U., Blakemore, W.E and Purpura, D.P., Alterations in neuron morphology in feline mannosidosis: a Golgi study, Acta NeuropathoL, 53 (1981) 75-79. 28 Walkley, S.U., Haskins, M.E. and Shull, R.M., Alterations in neuron morphology in mucopolysaccharidosis type I. A Golgi study, Acta Neuropathol., 75 (1988) 611-620. 29 Walkley, S.U., Prieur, D.J. and Ahern-Rindell, A.J., Prolific growth of ectopic axon hillock-associated neurites occurs on cortical pyramidal neurons in a newly discovered ovine model of

neuronal storage disease, Neuroscience, 22 (1987) $270 (abstr. l. 30 Walkley, S.U., Rattazzi, M.C. and Baker, H.J., Distribution of ectopic neurite growth and other geometrical distortions of neurons in feline GM2 gangliosidosis, Brain Res., in press. 31 Walkley, S.U., Siegel, D.A. and Wurzelmann, S., Ectopic dendritogenesis and associated synapse formation in swainsonine-induced neuronal storage disease, J. Neurosci., 8 (1988) 445-457. 32 Walkley, S.U., Wurzelmann, S. and Purpura, D.P., Uttrastructure of neurites and meganeurites of cortical pyramidal neurons in feline gangliosidosis as revealed by the combined Golgi-EM technique, Brain Res., 211 (1981) 393-398. 33 Wenger, D.A., Sattler, M., Kudoh, T., Snyder, S.P. and Kingston, R.S., Niemann-Pick disease: a genetic model in Siamese cats, Science, 208 (1980) 1471-1473. 34 Williams, R.S., Gott, I.T., Ferrante, R.L and Caviness, V.S., The cellular pathology of neuronal ceroid lipofuscinosis, Arch. NeuroL, 34 (1977) 278-305. 35 Wood, P.A., McBride, M.R., Baker, H.J. and Christian, S.T., Fluorescence polarization analysis, lipid composition, and Na, K-ATPase kinetics of synaptosomal membranes in feline GMI and GM2 gangliosidosis, J. Neurochem., 44 (1985) 947-956.