Intraparenchymal grafting of cerebellar cell suspensions to the deep cerebellar nuclei of pcd mutant mice, with particular emphasis on re-establishment of a Purkinje cell cortico-nuclear

Anat Embryo1 (1992) 185:409-420 Anatomy and Embryology �9 Springer-Verlag 1992

Intraparenchymal grafting of cerebellar cell suspensions to the deep cerebellar nuclei of pcd mutant mice, with particular emphasis on re-establishment of a Purkinje cell cortico-nuclear projection Lazaros C. Triarhou 1, Walter C. Low 2, and Bernardino Ghetti I

1 Laboratory of Cellular and Molecular Neuropathology, Department of Pathology, and Program in Medical Neurobiology, Indiana University School of Medicine, Indianapolis, IN 46202, USA 2 Departments of Neurosurgery and Physiology, and Program in Neuroscience, University of Minnesota Medical School, Minneapolis, MN 55455, USA

Accepted December 9, 1991

Summary. In transplanting embryonic cerebellar grafts to the cerebellar cortex of "Purkinje cell degeneration" (pcd) mutant mice to replace missing Purkinje cells (PC), donor PC leave the graft and migrate to the molecular layer of the host. However, PC axons do not always reach the deep cerebellar nuclei of the host, which would be a key element in restoring much of the necessary inhibitory cortico-nuclear projection associated with normal cerebellar function. Rather, grafted PC axons often innervate a region containing deep cerebellar nuclei neurons inside the transplant, while the perikaryon migrates to the host molecular layer. In the present study, aimed at re-establishing a PC innervation of the deep nuclei, we implanted E12 cerebellar cell suspensions intraparenchymally to the deep cerebellar mass of the hosts. The development of grafted PC was monitored with 28-kDa calcium-binding protein (CaBP) immunocytochemistry at various times after transplantation. At short survival times (5 days after grafting), grafts were confined to the site of the original injection. At longer survival times (7-32 days after grafting), grafted PC formed a migratory stream that reached the cerebellar cortex of the host. The most robust graft development was seen 1 month after grafting, the longest survival time allowed in this series of experiments. At that time, clusters of donor PC were found both in the deep nuclei parenchyma and aligned along cortical folia. The orientation of the dendritic trees of PC that had migrated to the cortex was toward the pia. A CaBP-immunoreactive fibre plexus innervated the host deep cerebellar nuclei. The stream of grafted PC extended from the deep cerebellar nuclei to the cerebellar cortex of the host, indi- cating that donor PC could establish their axonal contacts in the deep nuclei and then move to their final

Offprint requests to: L.C. Triarhou, Department of Pathology (Neuropathology), Indiana University School of Medicine, Medical Science Building A-142, 635 Barnhill Drive, Indianapolis, IN 46202-5120, USA

cortical locality, thus recapitulating a migratory path normally taken during cerebellar ontogeny. It appears therefore that both from the pathophysiological and on- togenetic standpoints, the deep cerebellar nuclei represent the appropriate site for PC implantation in cerebellocortical atrophy.

Key words: Cerebellar graft - Deep cerebellar nuclei - Neurological mutant mice "Purkinje cell degeneration" (pcd)

Introduction

The model most widely used to counteract heredodegenerative ataxia by means of neural grafting is the pcd mutant mouse (Sotelo and Alvarado-Mallart 1986, 1987a, b, 1991 ; Triarhou et al. 1987a, 1989, 1991 ; Gar- dette et al. 1988, 1990; Sotelo 1988; Chang et al. 1989; Ghetti et al. 1990; Sotelo et al. 1990). This mutant is characterized by a normal anatomical development of the cerebellum, followed by a virtually complete loss of Purkinje cells (PC), which takes place between postna- tal days 17 and 45 (Mullen et al. 1976; Landis and Mul- len 1978). Granule cells (GC) of the cerebellum (Ghetti et al. 1978; Triarhou et al. 1985), deep cerebellar nuclei neurons (Wassef et al. 1986; Triarhou et al. 1987b), and inferior olivary neurons (Ghetti et al. 1987; Shojaeian et al. 1988; Triarhou and Ghetti 1991) also degenerate to some extent; however, the loss of such neurons is only partial and is thought to represent a transsynaptic event (Ghetti et al. 1990). The site of action of the mutant gene appears to be intrinsic to PC (Mullen 1977).

Normally, PC constitute the only projection neuron of the cerebellar cortex (Eccles et al. 1967; Palay and Chan-Palay 1974; Ito 1984). All of the remaining cortical neurons are interneurons, functioning to modulate PC

410

Normal Cerr

PC

!

�9 DCN

Cerel~etl~r" O~[F'L

(9 I I

C~raf[ ~ CerebeZl~.r r

', /

. . . . . . . . . . . . . . . . . . . . . . . . . . . : = : : . - : : ' _ ' . . . . . . . . . . . . . . . . . . . . .

~c PC

f ~f . . . . .

(3 Fig. 1 a-d

~DCN

Cere~c|[~T" 0 ~ .

6C

'~f 4 1~ f cf

Q Cert~elta~

DC]~

. . . . . ~

~reffl~ i~ Deep Cerz~ell~Y" ]~eles

|

---:;i;2 \ .

/ r r

DCN . .

Cere~ellar

411

activity. PC are also modulated by afferent olivocerebel- lar climbing fibres. Mossy fibres indirectly affect PC activity through the mediation of G C and parallel fibres, which establish synapses on PC dendrites. Signals are transmitted f rom cortex to deep cerebellar nuclei through PC axons. In turn, axons of neurons of the deep nuclei transmit impulses outside the cerebellum (Fig. I a).

The degeneration of PC, as seen in pcd mutant mice and in humans with cerebellar a t rophy (Holmes 1907), deprives climbing fibres and cortical interneurons of a major synaptic target. However, the associated interrup- tion of the cortico-nuclear projection would be an im- por tant factor contributing to the pathophysiology of the ataxic syndrome (Fig. 1 b). PC utilize 7-aminobutyric acid (GABA) as their neurotransmitter (Obata 1969), and thus exert a powerful inhibition on deep nuclei neurons. In pcd mutant mice, the virtually complete loss of PC and their axons leads to a substantial reduction of the GABAergic input to the deep nuclei. A residual innervation by GABAergic boutons, estimated to be 15% of the normal value, persists in the pcd deep nuclei (Wassef et al. 1986); such axons could arise either f rom local interneurons or f rom inhibitory neurons projecting to the inferior olivary complex (Nelson et al. 1984; An- gaut and Sotelo 1987; Nelson and Mugnaini 1989) with collaterals in the deep cerebellar nuclei. However, the absence of 85% of the GABAergic boutons, associated with the loss of PC, would lead to a decreased inhibition of firing of deep nuclei neurons toward their post-cerebellar targets.

Due to the specificity of the cerebellar network ( "po in t - to -po in t" system, Sotelo and Alvarado-Mallar t 1986), the importance of precise re-connections following grafting of primordial cerebellar tissue to replace missing PC in pcd mutan t mice has been emphasized (Sotelo and Alvarado-Mal lar t 1986, 1987a, b, 1991; Gardet te et al. 1988; Sotelo et al. 1990). In brief, when a graft is placed into the host cerebellar cortex or between two adjacent cortical folia:

<

Fig. 1 a-d. Hypothetical scheme of Purkinje cell (PC) relationships with neurons of the deep cerebellar nuclei (DCN) in four different conditions. In normal cerebellum (a), PC axons supply an inhibitory input to DCN neurons, which in turn provide the cerebellar axonal output to post-cerebellar nuclei. In cortex, climbing and mossy fibre afferents (cfand mf, respectively) regulate PC activity, the latter with the mediation of granule cells (GC) and their axons, the parallel fibres (p J). In pcd mutant cerebellum (b), PC degenerate, thus depriving DCN neurons of the inhibitory c0rtico-nuclear input. For purposes of simplicity, only PC loss is depicted in the mutant, which is complete, while changes that result secondarily from PC loss and are only partial are not shown. When a cerebellar cell suspension is placed into the cerebellar cortex of the mutant (e), donor PC leave the graft and migrate into the cortex of the recipient; however, they often leave their axons inside the graft, showing a preference for their co-grafted DCN cells. The DCN of the host rarely receive axonal innervation from grafted PC. Following intraparenchymal graft placement into the DCN of the host (d), donor PC are close enough to the DCN of the host to be able to supply an inhibitory axonal innervation to the previously denervated DCN neurons

1. Donor PC leave the graft and migrate into the molecular layer of the host. 2. Donor PC receive synaptic input f rom host parallel and climbing fibres, which follows a normal developmental pattern. 3. Such afferent inputs to PC are functional as studied by electrophysiology.

The most serious limitation encountered so far in re- establishing the disrupted cerebellar circuitry is that grafted PC axons do not always reach host deep nuclei neurons. This problem can be attributed to the following reasons : 1. The distance which PC axons have to cover may be too long; it has been estimated that if donor PC are more than 600 gm away f rom the host deep nuclei, they do not innervate them (Sotelo 1988; Sotelo and Alvara- do-Mallart 1991). 2. The positioning of the axonal and dendritic poles of grafted PC during their migration is likely to be in- fluenced by two types of chemical cues: an affinity of PC axons for co-grafted deep nuclei cells, and an affinity of PC dendrites for the host molecular layer (Chang et al. 1989; Sotelo etal . 1990; Fig. 1c).

A PC innervation of the deep cerebellar nuclei of the host would be one key element in reinstating the missing inhibitory cortico-nuclear projection associated with normal cerebellar function. Ideally, for a complete restoration of function, grafted PC would also have to reach the molecular layer and to receive inputs f rom the surviving parallel and climbing fibres.

Having these considerations in mind, the aim of the present study was to re-establish a PC innervation of the deep cerebellar nuclei by implanting cerebellar cell suspensions directly into the deep nuclei parenchyma of the host (Fig. 1 d). A preliminary account of this work has been presented (Triarhou et al. 1989).

Materials and methods

Animals. Recipient animals were obtained from the colony of mutant mice maintained at Indiana University Medical Center, and originally established from pcd heterozygous mice purchased from Jackson Laboratory (Bar Harbor, Me.). Wild-type and mutant mice are maintained on a congenic C57BL/6J stock. Homozygous pod mutants (pcd/pcd) were obtained by crossing pairs of heterozy- gotes or homozygous females with heterozygous males. Animals were kept on a 12-h dark-light cycle and were provided with food and water ad libitum. Nine pcd homozygotes, 3-4 months old at the time of implantation, were used as graft recipients. Five of the mice received bilateral, and four unilateral cerebellar cell suspension grafts into the deep cerebellar mass. Mice were perfused at 5, 7, 15, 21, 29, and 32 days after grafting. Additional normal (+ /+ ) and non-grafted pcd mutant mice were used to study PC innervation and denervation of the deep nuclei. All experimenta- tion reported in this article was conducted in compliance with the guidelines of the National Institutes of Health concerning the care and use of laboratory animals for experimental procedures.

Embryonic cell suspension. Pairs of normal ( + / + ) C57BL/6J mice were placed in the same cage for a 36-h mating period. Pregnant dams were selected 11 days after separation from the males (i.e. on about gestational day 12) and anaesthetised with sodium pentobarbital (50 mg/kg intraperitoneally). A Caesarean section was per-

412

formed, and the embryos were removed from the uterine horns. Cerebellar primordia were dissected out from a total of 32 fetuses (crown-rump length 9-11 ram) with the help of fine iridectomy scissors, and collected in ice-cold Hank's balanced salt solution (HBSS). A cell suspension with several large aggregates was pre- pared under aseptic conditions essentially according to Bj6rklund et al. (1980) as modified by Chang et al. (1989). In brief, enzymatic digestion was carried out with 1 : 300 collagenase (Sigma Chemical Co., St. Louis, Mo.) and 1:300 dispase (Boehringer-Mannheim Biochemicals, Indianapolis, Ind.) in HBSS for about 10 min. Enzy- matic activity was quenched by the addition of HBSS with 15% fetal calf serum. The mixture was pelleted down by centrifugation at 180 g for 5 min. The pellet was resuspended in HBSS and spun down by centrifugation. Cell viability, determined in 5-gl aliquots using a haemacytometer and the trypan blue exclusion principle, was about 90% in the more dissociated portion of the suspension. Cell concentration was adjusted to 50,000 viable cells/gl for grafting.

Graft surgery. Recipient mutant mice were anaesthetised by intra- peritoneal injection of 1 ml/kg of a mixture containing 100 mg/ml ketamine HC1, 2.2 mg/ml acepromazine and 0.48 mg/ml atropine, and placed in a Kopf stereotaxic apparatus. Craniotomies (unilateral or bilateral) were performed at the region of the occipital bone. Cell suspensions were grafted by using a 10-gl Hamilton syringe. Stereotaxic coordinates had been determined in advance with cresyl violet injection into the cerebellum of a test animal; they were 2.2 mm posterior and 1.7 mm lateral (to the left or right side) fi'om lambda, and 2.3 mm dorsoventral from dural surface (incisor bar was set at - 5). Two microlitres of cell suspension (containing a total of 100,000 viable cells) were slowly injected into the parenchyma of the cerebellar hemisphere under manual control over a period of 5 min; the pipette was left in place for an additional 10 min before withdrawal. The overlying skin was sutured and animals were allowed to recover.

an oven at 70~ for 48 h. Mesas were trimmed on the plastic blocks under a dissecting microscope, and ultrathin sections were obtained and stained with lead citrate. Ultrastructural observations were carried out in a Philips 300 electron microscope.

Results

Host mutant cerebellum

The cerebe l lum ofpcd m u t a n t mice unde rgoes a p rogres - sive a t r o p h y ; the cerebel lar cor tex is decreased in thickness. PC are v i r tua l ly absen t f rom the cerebel la r cor tex o f pcd m u t a n t s af ter 45 days o f age. The loss o f PC results in a dene rva t i on o f the deep nuclei f r om afferent PC axon terminals . This can be seen in i m m u n o c y t o - chemica l p r e p a r a t i o n s label led wi th a n t i - C a B P ant i - bodies (Fig. 2). In n o r m a l mice, a dense ne tw ork o f im- m u n o p o s i t i v e C a B P te rmina l s is seen in the deep cerebellar nuclei (Fig. 2a) . In 23 -day-o ld pcd mutan t s , when an e s t ima ted 2 5 - 5 0 % o f PC have degene ra t ed (Mul len et al. 1976), one sees a pa r t i a l dep le t ion o f C a B P - i m m u - noreac t ive PC axon te rmina ls f rom the deep nuclei (Fig. 2b) . In pcd m u t a n t s b e y o n d 45 days o f age, the deep cerebel la r nuclei are devo id o f C a B P - i m m u n o r e a c - five nerve te rminals , owing to the v i r tua l ly comple t e absence o f PC (Fig. 2c). The n u m b e r o f deep nuclei neurons in pcd m u t a n t s is n o r m a l at 23 days o f age and decl ines by 21% at 300 days o f age (Tr i a rhou et al. 1987b).

Immunocytochemistry. The animals were heparinised (500 USP un- its i.p.), anaesthetised with pentobarbital sodium (50 mg/kg i.p.) and perfused transcardially with 10 ml of ice-cold, oxygen-en- riched, calcium-free Tyrode's buffer, followed by 100 ml of a mixture containing ice-cold 4% (w/v) formaldehyde, 0.05% (v/v) glutaraldehyde and 0.2% picric acid in 0.1 M sodium phosphate buffer (pH 7.4). Brains were removed and immersion-fixed in the same fixative without glutaraldehyde for 2 h at 4 ~ C. They were subse- quently rinsed in 20% sucrose dissolved in 0.1 M phosphate buffer (pH 7.4) for 1 day or longer. Sections, 50 gm in thickness, were cut on a freezing microtome and collected free-floating in Tris- buffered saline (pH 7.6).

Immunolabelling of sections with 28-kDa CaBP antibody was carried out en bloc. Three 10-rain rinses with Tris-buffered saline were carried out between the various incubations. The primary antibody (mouse monoclonal) was applied at a 1:5,000 dilution at 4 ~ C for 40 h. The production and characterisation of this antibody can be found in Cello et al. (1990). Following rinses, the secondary antibody (rabbit anti-mouse IgG) was applied at a 1 : 40 dilution at room temperature for 1 h. Mouse peroxidase anti-peroxidase (PAP) complex was used at a 1 : 200 dilution at room temperature for 1 h. After that step, sections were rinsed and reacted with 0.05% (w/v) 3,3'-diaminobenzidine-HC1 with 0.01% H202 for 10 min, and finally rinsed and mounted on acid-cleaned, gelatin- coated glass slides.

For electron microscopy, selected sections were rinsed in 0.1 M phosphate buffer pH 7.2, twice for 5 min and then impregnated en bloc with 2% osmium tetroxide in 0.1 M phosphate buffer for 1 h, rinsed in normal saline twice for 5 min, and contrasted with a saturated aqueous solution of uranyl acetate for 2 h. They were then dehydrated in graded ethanols (50%, 70%, 80%, 95%, 100% for 10 min each) and acetone (three times for 10 min), left in a 50% acetone: 50% Poly/Bed 812 mixture for 2 h, 100% Poly/Bed 812 for 2 h, flat embedded in Poly/Bed 812 and polymerized in

Graft survival

The d e v e l o p m e n t o f g ra f ted PC was m o n i t o r e d wi th im- m u n o c y t o c h e m i s t r y for CaBP, a PC-specif ic m a r k e r in n o r m a l cerebel lum. In 11 o f the 14 graf ts , C a B P - i m m u n - opos i t ive cells ( f rom now on referred to as g ra f ted PC) cou ld be ident i f ied (Fig. 3 and Table 1).

Migratory patterns of grafted PC

A t 5 days surv iva l t ime af ter t r ansp l an t a t i on , graf ts were conf ined to a wel l -c i rcumscr ibed region, c o r r e s p o n d i n g to the or ig ina l in ject ion site. The cluster o f g ra f ted PC was loca ted in p a r t in the p a r e n c h y m a o f the hos t deep cerebel la r nuclei and pa r t l y in the ad jacen t subcor t i ca l whi te m a t t e r (Fig. 3a). G r a f t s r eached d imens ions o f 100 lam AP, 600 g m DV, and 100-200 g m M L . G r a f t e d P C were sp ind le - shaped and were no t seen away f rom the graft .

A t 7 and 15 days survival , d o n o r PC f o r m e d a migra - to ry s t ream, c o m m e n c i n g f rom the c luster inside the deep cerebe l la r mass and ex tend ing to the cerebel la r cortex o f the hos t (Fig. 3 b, c). The cluster o f in jected cells reached on the average d imens ions o f 125 g m AP, 350 g m DV, and 150 g m M L . The m i g r a t o r y s t r eam was a b o u t 50-75 g m wide in the A P p lane and therefore ap- pea red th inner than the p o r t i o n o f the c luster tha t was inside the deep nuclei ; lengthwise, the m i g r a t o r y s t r eam

413

Fig. 2a-e. Innervation of the deep cerebellar nuclei by Purkinje axon terminals, as revealed by immunocytochemistry for calcium- binding protein, a In normal cerebellum, one sees plentiful immunoreactive terminals surrounding unlabelled neurons of the deep nucleus, b In the cerebellum of 23-day-old pcd mutants, where 25-50% of Purkinje cells have degenerated, one sees a substantial reduction in the number of calcium-binding protein immunoreac-

tive nerve terminals innervating the deep nuclei, e In the cerebellum of pcd mutants past 45 days of age, when over 99% of Purkinje cells have degenerated, one observes a virtually complete absence of calcium-binding protein immunoreactive nerve terminals in the deep cerebellar nuclei, denoting in essence a cortico-nuclear dener- ration. Only background staining is seen. Fields represent parasagittal sections from nucleus lateralis. Bar 50 gm

Table 1. Listing of recipient animals. Graft survival was monitored by immunopositivity of grafted Purkinje cells (PC) for calcium- binding protein (CaBP)

Animal Graft Graft Graft survival CaBP no. type side (days) PC

5681-13 Bilateral Right 5 + Left 5 +

6472-5 Unilateral Right 5 5642-10 Bilateral Right 7

Left 7 + 5687-9 Bilateral Right 15 +

Left 15 + 6472-1 Unilateral Right 15 6472-4 Unilateral Right 21 + 6472-6 Unilateral Right 29 + 5252 Bilateral Right 32 +

Left 32 + 5461 Bilateral Right 32 +

Left 32 +

o f grafted PC reached dimensions o f 600-700 gm in the DV plane.

The mos t robus t graft development was seen 1 m o n t h after graft ing (Figs. 3 d and 4). At tha t time, the cytologi- cal development o f grafted PC had proceeded, such that

the dendritic arbor iza t ion was extensive and covered mos t o f the thickness o f the host molecular layer. D o n o r PC were found in clusters extending f rom the deep nuclei pa renchyma to the cortical folia o f the host. D o n o r PC clustered over the entire thickness o f the colonized molecular layer; they formed zones two to five cells thick. Some of the t ranplanted PC remained in the vicinity o f the injection track, while mos t o f them had migra ted and occupied folia cont iguous to the track. The migra to- ry s t ream emerged f rom the in t raparenchymal ly injected cluster o f cells, and moved off between the subcortical white mat ter and the cerebellar surface. Graf ted CaBP- immunoreac t ive PC spread over a length o f up to 3,000 gm o f linear cortex. Graf ted PC spread to bo th sides o f the injection site to a radius o f 500-600 gm.

Dendritic development and orientation of grafted PC

The dendrit ic trees o f mature graf ted PC, located in the cortex, were a r ranged in a plane at an angle o f 90 ~ to the or ientat ion o f host parallel fibres, i.e. f lattened in the transverse plane with a m o n o p l a n a r disposition. The m a x i m u m dendrit ic extension was seen in sagittal sections, whereas it was minimal in coronal sections. On the other hand, the dendrit ic trees o f the major i ty o f

414

Fig. 3a-d. Intraparenchymal graft placement into the deep cerebellar nuclei of recipient pcd mutant mice. Immunocytochemical label- ling with calcium-binding protein antibodies reveals the distribution of donor Purkinje cells, since virtually all of the host Purkinje cells have been lost. a At 5 days after grafting, donor Purkinje cells are confined to a cluster (arrow) inside the host parenchyma. b At 7 days after grafting, a migratory stream of calcium-binding protein-immunoreactive cells (double arrows) is formed, starting at the cluster of the original graft site (single arrow), with a direction

toward areas of the cerebellar cortex, e At 15 days after grafting, a proportion of transplanted Purkinje cells (double arrows) has left the original site of injection (single arrow) and expanded to cover cortical layers, d At 30 days after grafting, donor Purkinje cells cover a substantial part of the host, along the shape of the cerebellar cortex; grafts placed bilaterally (arrows). Sagittal sections with unilateral grafts in a-e, coronal section with bilateral grafts in d. Bar 300 gm (a, b); 100 gm (e); 1,000 gm (d)

grafted PC located in the deep cerebellar nuclei were multiplanar.

Graf ted PC reached diameters of 4-6 gm at 5-7 days after grafting and 15-25 gm at 15-30 days. Donor PC had an elongated shape at 5 7 days' survival time. At later stages of development (15-30 days after grafting), donor PC had a spherical shape. The degree of dendritic differentiation progressed with time after grafting, such that immature processes seen at 7 days after grafting had developed into full dendritic trees by 15 days after

grafting and later (Fig. 5). These phases of dendritogene- sis correspond to the ones observed by Ram6n y Cajal (1926) during normal embryonic cerebellar development, and described in intracortical cerebellar grafts by Sotelo et al. (1990) and by Sotelo and Alvarado-Mal lar t (1991): they are, namely, the phase o f " fusiform cells" (5-9 days after grafting), the phase of "stellate cells with disor- iented dendrons" (10 11 days after grafting), and the phase of the "or ienta t ion and flattening of dendri tes" (11 14 days after grafting).

415

The diameter of host deep nuclei neurons that were innervated by CaBP immunoreactive bouton-like swell- ings was on the average 10 gm, as measured in immunocytochemical preparations counterstained with toluidine blue (Fig. 6c, d). The varicosities of CaBP-immunoreactive PC axon terminals had a diameter of 2-3 gin. Elec- tron microscopic observations 1 month after transplantation showed CaBP-immunoreactive axon terminals in synaptic contact with unlabelled neurons of the host deep cerebellar nuclei (Fig. 7).

Fig. 4. A section of pcd cerebellum with a graft at I month's survival time. Immunocytochemistry with calcium-binding protein antibody. The graft occupies a little less than one-half of the host cerebellum at this particular parasagittal level. Bar 300 gm

The developed dendritic trees of PC that had migrated to the cerebellar cortex of the host were oriented upward, i.e. toward the pia, and their axons were oriented toward the host subcortical white matter (Fig. 5 f). Such an orientation mimicked the normal anatomical location of PC in the cerebellum.

Somatic investment of deep cerebellar neurons by PC axons

A PC axonal plexus innervated the host deep cerebellar nuclei already at 5 days after grafting and persisted at 32 days after grafting (Fig. 6). Although the same stereotaxic coordinates were used for all implantations, there were small variations among the precise placement of each graft. One reason for this is that when a large blood vessel is seen directly under the needle in the pial surface, then the injection site is slightly altered in order to avoid piercing the vessel and causing bleeding. Nevertheless, considering the technical problem associated with the small size of the deep cerebellar mass in the mouse, all of the grafts were found either directly inside the deep nuclei or in their immediate vicinity in the white matter. The relative amount of innervation supplied to each nucleus depended on the precise location of the graft. In general, the plexus of PC-derived CaBP-immunoreactive fibres covered the lateral aspect of the nucleus medialis, most of the nucleus interpositus, and the medial portion of the dentate nucleus.

Discussion

Cerebellar transplants have been used in pcd mutant mice to address several different issues; implantation site varied in each study in accordance with the question asked. The contribution of the present paper is discussed below, in the context of previously published work that has formed the basis for the present study.

Previous studies used solid EI4-E15 cerebellar primordia from normal donor embryos, grafted into the cerebellomedullary cistern of adult pcd mutants, to test the possible effects of the humoral environs related to the genetically-determined degeneration of PC and its chemical consequences on grafted wild-type ( + / + ) PC (Triarhou et al. 1987a). The successful growth of genetically-normal PC in the pcd cerebrospinal fluid, as well as the organization of donor tissue into a typical trila- minar structure, indicated that the adult pcd cerebellar microenvironment does not contain factors that could prevent the survival and histotypic differentiation of normal PC.

In another set of previous studies, solid cerebellar grafts (Sotelo and Alvarado-Mallart 1987b) or cerebellar cell suspensions (Sotelo and Alvarado-Mallart 1986, 1991 ; Gardette et al. 1988, 1990; Chang et al. 1989; So- telo et al. 1990) were grafted into the pcd cerebellar cortex (in most cases between two adjacent cortical folia) to replace missing PC. The integration of donor PC into the synaptic circuitry of the host (mainly by receiving afferents from host parallel and climbing fibres) led to the notion that neural transplantation could be used to counteract heredodegenerative ataxia. Particular emphasis has been placed on the fact that such an amelioration would be associated with the reconstruction of a neural system characterized by "point-to-point" synaptic con- nectivity (Sotelo and Alvarado-Mallart 1986).

The present study introduces a new site of graft placement, intraparenchymally into the deep cerebellar mass of the host. The main findings from utilizing this ap- proach can be summarized as follows. Firstly, a new PC axonal innervation is provided to the denervated deep nuclei of the host. Secondly, a substantial proportion of intraparenchymally grafted PC are found to oc- cupy cortical localities, most probably through migrating there. The size of the cortical migratory stream of grafted PC progressively increases with time after transplantation. The source of the PC axonal plexus in the host deep cerebellar nuclei may either be PC that have stayed in the deep nuclei or PC that have migrated to

416

Fig. 5a-f. Dendritic development of grafted Purkinje cells at 5 (a), 7 (b), 15 (c), 21 days (d), and 1 month after transplantation (e, f). Notice the correct orientation of Pttrkinje cell dendrites toward the pial surface in f; ML, molecular layer. All sections are parasagittal. Fields in a - e were taken from the intraparenchymal portion of the graft; the field in f is from the cerebellar cortex of the host, where grafted PC are now located following their migration. Bars 30 gm

417

Fig. 6a-d. Innervation of the host deep cerebellar nuclei (arrow- heads) by calcium-binding protein immunoreactive axons emanat- ing from grafted Purkinje cells at 15 days (a) and 1 month (b--d) after transplantation. Arrows in a and b show the graft border. Single arrow in e shows the intra-nuclear cluster of grafted Purkinje cells, while double arrows point to those that migrated to the cortex.

Sections in c and d were lightly counterstained with toluidine blue to reveal the cell bodies of host deep nuclei neurons. Sections in a and b are parasagittal, in e and d coronal. Abbreviations in c: M, nucleus medialis; I, nucleus interpositus; L, nucleus lateralis. The field in a is from nucleus lateralis, and in b and d from nucleus interpositus. Bar 50 gm (a, b, d), 200 gm (e)

the host cerebellar cortex or both. The idea that grafted PC which have migra ted to the cortex are those that establish synaptic contacts in the host deep nuclei needs to be demons t ra ted directly, e.g. by retrogradely label- ling cortically located PC bodies by injection o f tracer in the deep nuclei. However , one m a y speculate and pu t for th the hypothesis tha t grafted PC establish contac t

with cells o f the deep cerebellar nuclei and then migrate to the host cerebellar cortex.

Clues f rom developmental studies suppor t such a pos- sibility. Dur ing normal cerebellar on togeny, generat ion o f deep nuclei neurons antedates the p roduc t ion o f PC (Al tman 1982; A l tman and Bayer 1985 a). After the ces- sation o f mitot ic divisions, deep nuclei cell bodies de-

418

Fig. 7. Synaptic contact (arrow) formed by a calcium- binding protein-immunoreactive bouton on the unlabelled soma of a neuron (iV) in the recipient deep cerebellar nuclei. The section from which this micrograph was taken was coronal and was taken from the nucleus interpositus. Bar i gm

scend from the nuclear transitory zone to the depth of the cerebellum, whereas PC follow a course of migration from the ventricular neuroepithelium of the cerebellar anlage to the cerebellar cortex (Ram6n y Cajal 1929; Miale and Sidman 1961; Altman and Bayer 1985a, b). It has been suggested that the portion of the PC which is to become the axon may maintain synaptic contact with cells of the deep cerebellar nuclei from the earliest migratory stages of cerebellar histogenesis onwards (Miale and Sidman 1961) and that the crossing of trajec- tories may allow PC to establish contact with deep nuclei neurons en route of their perikarya to the surface (Alt- man 1982).

In support of this line of reasoning, a more recent immunocytochemical study on the prenatal development of mouse cerebellum, marking PC with anti-spot 35 antibody, has shown an early contact between immature PC and deep nuclei cells. After E16, presumed axons of PC reach the region of the deep nuclei and PC move through it at E17 (which would correspond to about 5 days after grafting). Spot 35-immunoreactive fibres enter the region of the deep cerebellar nuclei, apparently arising from both migratory PC and from PC aready settled in the cortex (Yuasa et al. 1991).

There are parallels between PC migratory events during development in situ and following grafting. Follow- ing transplantation to the cerebellar cortex, donor PC often leave their axon inside the graft, while the perikaryon migrates to the host molecular layer (Chang et al. 1989); grafted PC axons fail to reach the host deep nuclei if grafted PC are more than 600 gm away from them (Sotelo 1988; Sotelo and Alvarado-Mallart 5991); and the orientation of the dendritic tree of grafted PC is inverted in reference to the host cerebellar cortex. How- ever, the dendritic tree is oriented toward the direction of the migration, since in the graft situation PC travel from the pia toward the white matter, which is the re-

verse pathway from that taken normally in the in situ development.

In studies with cell suspension grafts to the cerebellar cortex, PC invade the host molecular layer through a radial or oblique migration (Sotelo et al. 1990). Migrato- ry PC are provided with a leading process and its termi- nal growth cone; when descending processes of radially migrating neurons reach the host molecular layer-granule cell layer interface, their growth cones arrest their inward penetration. It has been postulated that one pos- sine mechanism preventing further PC fibre outgrowth could be the granule cell layer, which may act as an inhibitory barrier (Sotelo et al. 1990). However, even in the case that some PC axons might extend past the granule cell layer, they would then enter the subcortical white matter. There, growing axons might be blocked from further growth, in line with the findings of Caroni and Schwab (1988) and Schnell and Schwab (1990), who showed that oligodendroglial cells express cell surface components that are potent inhibitors of neurite outgrowth along fibre tracts. In particular, oligodendrocytes and CNS myelin express membrane-bound surface pro- teins of defined molecular mass 35,000-kDa (NI-35) and 250,000-kDa (NI-250) with highly nonpermissive substrate properties. By intracerebrally applying monoclonal antibodies IN-1 and IN-2, raised against neurite growth inhibitors NI-35 and NI-250, it has been possible to induce regenerating fibres in the rat spinal cord to elongate over a distance 7- to l l-fold longer than in control animals. A similar mechanism may be operating in the cerebellum since the axons of cortically grafted PC have to traverse substantial tracts of oligodendroglia. Other mechanisms preventing graft-derived PC fibre outgrowth to the deep nuclei may involve the temporal expression of extracellular matrix (ECM) molecules, such as neural cell adhesion molecule (NCAM), N-cad- herin, and the integrins tenascin and thrombospondin,

419

which are known to be involved in different phases during the course of cerebellar development (cf. Reichardt and Tomaselli 1991).

At the end of their migration, PC exhibit a bipolar shape and begin to build their dendritic trees. In mature grafts, most of the donor PC have inverted dendrites (Sotelo and Alvarado-Mal lar t 1986). The direction of the migration appears to coincide with the dendritic pole and be opposite to the axonal pole of the cell. It has been suggested that chemical or electrical cues may selec- tively attract grafted PC to their ultimate site in the molecular layer (Sotelo and Alvarado-Mallar t 1987b), according to the idea of positive neurotropism of Ram6n y Cajal (1910).

When placed into the deep cerebellar mass, donor PC are able to innervate the host deep nuclei and many PC dendritic trees have a correct orientation, because the direction of their migratory path coincides with the normal one, i.e. f rom deep sites toward the pia. In this case, transplanted PC recapitulate a migratory path normally taken during embryonic development of the cerebellum.

One issue that merits discussion is the inclusion of donor deep nuclei cells in the grafted tissue, regardless of site of placement in the host. Donor PC may show an affinity for co-grafted deep nuclei cells. However, in the case of intraparenchymal grafts, donor PC would have the opportuni ty to be in physical proximity to host deep nuclei cells as well. The existence of vacated postsynaptic sites on neurons of the host deep cerebellar nuclei renders that target also available to donor PC. Of course, at this point it is difficult to estimate the relative proport ions of donor and host deep nuclei neurons being innervated by the axons of grafted PC.

The behavioral effects of cerebellar transplants in cerebellar ataxia of PC type remain an open question. Part- ly due to the lack of behavioural tests as specific for cerebellar PC loss as e.g. the rotational model of nigrostriatal dopamine deficiency (Ungerstedt and Arbuthnot t 1970), it has not been studied so far. It remains to be established whether intra-cortical or intra-nuclear grafts or both can provide a beneficial effect on locomotor coordination and to what degree.

In several neuroanatomical systems, graft-induced functional improvement has been effected by transplanting presynaptic neuronal populat ions (missing in the host) heterotopically into the area of their postsynaptic cells. As examples one can mention substantia nigra grafts to the caudate (Bj6rklund and Stenevi 1979), as well as septum (Bj6rklund and Stenevi 1977; Low et al. 1980), raphb (Segal and Azmitia 1986) and coeruleus grafts (Bj6rklund et al. 1976; Barry et al. 1987) to the hippocampus. In all of those studies, the corresponding intrinsic cells that are being replaced are generally located at a substantial distance f rom their targets (several millimeters or even 1 cm) and if transplanted homotopi- cally, they will not elongate their axons along the normal pathway to reach their target area f rom a distance (Bj6rklund et al. 1983). Moreover, when transplanted into the target, they do not migrate to their normal posi- tion. The situation is different in the cerebellum, since

the distance between presynaptic and target neurons is relatively shorter (compared e.g. to the nigrostriatal pathway). Another feature characteristic of PC is their substantial migratory potential, which in temporal terms coincides with the optimal harvesting age for grafting (E12). Both of these aspects emphasize the uniqueness of the cerebellar model in studying particular cellular mechanisms of neural transplantation.

Acknowledgements. We thank Dr. M.R. Celio, Fribourg University, Switzerland, for the gift of CaBP antibody; C.J. Alyea, R. Funk- houser and L. Graham for technical help; J. Demma for photo- graphic assistance; and Dr. C.D. Nordschow for support. A preliminary report on this work was presented at the 19th Annual Meeting of the Society for Neuroscience, Phoenix, Ariz., 29 Oc- tober-3 November, 1989. Supported in part by grants R01- NS14426 (B.G.) and R29-NS29283 (L.C.T.) from U.S. Public Health Service.

References

Altman J (1982) Morphological development of the rat cerebellum and some of its mechanisms. Exp Brain Res [Suppl] 6 : 8-49

Altman J, Bayer SA (1985a) Embryonic development of the rat cerebellum. II. Translocation and regional distribution of the deep neurons. J Comp Neurol 231:27-41

Altman J, Bayer SA (1985b) Embryonic development of the rat cerebellum. III. Regional differences in the time of origin, migration, and settling of Purkinje cells. J Comp Neurol 231:42- 65

Angaut P, Sotelo C (1987) The dentato-olivary projection in the rat as a presumptive GABAergic link in the olivo-cerebello- olivary loop. Neurosci Lett 83:227-231

Barry DI, Kikvadze I, Brundin P, Bolwig TG, Bj6rklund A, Lind- vall O (1987) Grafted noradrenergic neurons suppress seizure development in kindling-induced epilepsy. Proc Natl Acad Sci USA 84:8712-8715

Bj6rklund A, Stenevi U (1977) Reformation of the severed septohippocampal cholinergic pathway in the adult rat by transplanted septal neurons. Cell Tissue Res 185:289-302

Bj6rklund A, Stenevi U (1979) Reconstruction of the nigrostriatal dopamine pathway by intracerebral nigral transplants. Brain Res 177:555-560

Bj6rklund A, Stenevi U, Svendgaard N-A (1976) Growth of transplanted monoaminergic neurones into the adult hippocampus along the perforant path. Nature 262: 787-790

Bj6rklund A, Schmidt RH, Stenevi U (1980) Functional reinnerva- tion of the neostriatum in the adult rat by use of intraparenchymal grafting of dissociated cell suspensions from the substantia nigra. Cell Tissue Res 212:39-45

Bj6rklund A, Stenevi U, Schmidt RH, Dunnett SB, Gage FH (1983) Survival and growth of nigral cell suspensions implanted in different brain sites. Acta Physiol Scand [Suppl] 522:9-18

Caroni P, Schwab ME (1988) Antibody against myelin-associated inhibitor of neurite growth neutralizes nonpermissive substrate properties of CNS white matter. Neuron 1:85 96

Cello MR, Baler W, Schfirer L, Gregersen H J, de Viragh PA, Norman AW (1990) Monoclonal antibodies directed against the calcium binding protein Calbindin D-28k. Cell Calcium 11 : 599 602

Chang AC, Triarhou LC, Alyea CJ, Low WC, Ghetti B (1989) Developmental expression of polypeptide PEP-19 in cerebellar cell suspensions transplanted into the cerebellum ofpcd mutant mice. Exp Brain Res 76:639-645

Eccles JC, Ito M, Szentfigothai J (1967) The cerebellum as a neuronal machine. Springer, Berlin Heidelberg, 335 pp

Gardette R, Alvarado-Mallart RM, Crepel F, Sotelo C (1988) Elec- trophysiological demonstration of a synaptic integration of

420

transplanted Purkinje cells into the cerebellum of the adult Pur- kinje cell degeneration mutant mouse. Neuroscience 24: 777- 789

Gardette R, Crepel F, Alvarado-Mallart RM, Sotelo C (1990) Fate of grafted embryonic Purkinje cells in the cerebellum of the adult "Purkinje cell degeneration" mutant mouse. II. Develop- ment of synaptic responses: An in vitro study. J Comp Neurol 295:188-196

Ghetti B, Alyea CJ, Muller J (1978) Studies on the Purkinje cell degeneration (pcd) mutant: primary pathology and transneur- onal changes. J Neuropathol Exp Neurol 37:617

Ghetti B, Norton J, Triarhou LC (1987) Nerve cell atrophy and loss in the inferior olivary complex of ~ Purkinje cell degeneration" mutant mice. J Comp Neurol 260 : 409-422

Ghetti B, Triarhou LC, Alyea CJ, Low WC, Chang AC (1990) Timing of neuronal replacement in cerebellar degenerative ataxia of Purkinje cell type. Prog Brain Res 82:197-202

Holmes G (1907) A form of familial degeneration of the cerebellum. Brain 30:467-489

Ito M (1984) The cerebellum and neural control. Plenum Press, New York, 580 pp

Landis SC, Mullen RJ (1978) The development and degeneration of Purkinje cells in pcd mutant mice. J Comp Neurol 177:125- 144

Low WC, Lewis PR, Bunch ST, Dunnett SB, Thomas SR, Iversen SD, Bj6rklund A, Stenevi U (1980) Functional recovery following neural transplantation of embryonic septal nuclei in adult rats with septohippocampal lesions. Nature 300: 260-262

Miale IL, Sidman RL (1961) An autoradiographic analysis of histogenesis in the mouse cerebellum. Exp Neurol 4: 277-296

Mullen RJ (1977) Site ofpcd gene action and Purkinje cell mosaic- ism in cerebella of chimaeric mice. Nature 270: 245-247

Mullen RJ, Eicher EM, Sidman RL (1976) Purkinje cell degeneration, a new neurological mutation in the mouse. Proc Natl Acad Sci USA 73:208-212

Nelson BJ, Mugnaini E (1989) Origins of GABAergic inputs to the inferior olive. Exp Brain Res [Suppl] 17:86-87

Nelson B, Barmack NH, Mugnaini E (1984) A GABAergic cerebello-olivary projection in the rat. Soc Neurosci Abstr 10:539

Obata K (1969) GABA in Purkinje cells and motoneurons. Exper- ientia 25 : 1285

Palay SL, Chan-Palay V (1974) Cerebellar cortex: Cytology and organization. Springer, Berlin, 348 pp

Ram6n y Cajal S (1910) Algunas observaciones favorables a la hip6tesis neurotr6pica. Trab Lab Invest Biol Univ Madrid 8 : 63-135

Ram6n y Cajal S (1926) Sur les fibres moussues et quelques points douteux de la texture de l'~corce c~r6belleuse. Trab Lab Invest Biol Univ Madrid 24:215-251

Ram6n y Cajal S (1929) Studies on vertebrate neurogenesis. Guth L (transl). Thomas, Springfield, Ill., 1960, pp 251-321

Reichardt LF, Tomaselli KJ (1991) Extracellular matrix molecules and their receptors : Functions in neural development. Ann Rev Neurosci 14:531-570

Schnell L, Schwab ME (1990) Axonal regeneration in the rat spinal cord produced by an antibody against myelin-associated neurite growth inhibitors. Nature 343 : 269-272

Segal M, Azmitia EC (1986) Fetal raphe neurons grafted into the hippocampus develop normal adult physiological properties. Brain Res 364:162-166

Shojaeian H, Delhaye-Bouchaud N, Mariani J (1988) Stability o f inferior olivary neurons in rodents. I. Moderate cell loss in adult Purkinje cell degeneration mutant mouse. Dev Brain Res 38:211-218

Sotelo C (1988) Transplantation de neurones embryonnaires dans le cervelet de souris: Restauration de l'int6grit6 c6r6belleuse chez des souris avec ataxie h6r6do-d6g6n6rative. M6d Sci 8:507 514

Sotelo C, Alvarado-Mallart RM (1986) Growth and differentiation of cerebellar suspensions transplanted into the adult cerebellum of mice with heredodegenerative ataxia. Proc Natl Acad Sci USA 83:1135 1139

Sotelo C, Alvarado-Mallart RM (1987 a) Embryonic and adult neurons interact to allow Purkinje cell replacement in mutant cerebellum. Nature 327:421-423

Sotelo C, Alvarado-Mallart RM (1987b) Reconstruction of the defective cerebellar circuitry in adult Purkinje cell degeneration mutant mice by Purkinje cell replacement through transplantation of solid embryonic implants. Neuroscience 20:1-22

Sotelo C, Alvarado-Mallart RM (1991) The reconstruction ofcere- bellar circuits. Trends Neurosci 14:350-355

Sotelo C, Alvarado-Mallart RM, Gardette R, Crepel F (1990) Fate of grafted embryonic Purkinje cells in the cerebellum of the adult "Purkinje cell degeneration" mutant mouse. I. Develop- ment of reciprocal graft-host interactions. J Comp Neurol 295:165 187

Triarhou LC, Ghetti B (1991) Stabilisation of neurone number in the inferior olivary complex of aged ' Purkinje cell degeneration' mutant mice. Acta Neuropathol (Berl) 81:597-602

Triarhou LC, Norton J, Alyea C, Ghetti B (1985) A quantitative study of the granule ceils in the Purkinje cell degeneration (ped) mutant. Ann Neurol 18:146

Triarhou LC, Low WC, Ghetti B (1987a) Transplantation of cerebellar anlagen to hosts with genetic cerebellocortical atrophy. Anat Embryol 176:145-154

Triarhou LC, Norton J, Ghetti B (1987b) Anterograde transsynaptic degeneration in the deep cerebellar nuclei of Purkinje cell degeneration (pea') mutant mice. Exp Brain Res 66 : 577-588

Triarhou LC, Low WC, Ghetti B (1989) Intraparenchymal grafting of cerebellar cell suspensions to the deep cerebellar nuclei of pod mutant mice: Rationale and histochemical organization. Soc Neurosci Abstr 15:10

Triarhou LC, Low WC, Ghetti B (1991) Serotonin fiber innervation of cerebellar cell suspensions intraparenchymalty grafted to the cerebellum of ped mutant mice. Neurochem Res 17:475-482

Ungerstedt U, Arbuthnott GW (1970) Quantitative recording of rotational behavior in rats after 6-hydroxydopamine lesions of the nigrostriatal dopamine system. Brain Res 24:485-493

Wassef M, Simons J, Tappaz ML, Sotelo C (1986) Non-Purkinje ceil GABAergic innervation of the deep cerebellar nuclei: A quantitative immunocytochemical study in C57BL and in Pur- kinje cell degeneration mutant mice. Brain Res 399:12~135

Yuasa S, Kawamura K, Ono K, Yamakuni T, Takahashi Y (1991) Development and migration of Purkinje cells in the mouse cerebellar primordium. Anat Embryol 184:195-212

Intraparenchymal grafting of cerebellar cell suspensions to the deep cerebellar nuclei of pcd mutant mice, with particular emphasis on re-establishment of a Purkinje cell cortico-nuclear

Documents