Formation of cholinergic synapses by intrahippocampal septal grafts as revealed by choline acetyltransferase immunocytochemistry

151

Brain Research, 369 (1986) 151-162 Elsevier BRE 11559

Formation of Cholinergic Synapses by Intrahippocampal Septal Grafts as Revealed by Choline Acetyltransferase Immunocytochemistry D.J. CLARKE l, F.H. GAGE 2.* and A. BJORKLUND 2

t University Department of Pharmacology, Oxford ( U.K.) and 2Department of Histology, University of Lund, Lund (Sweden) (Accepted July 30th, 1985)

Key words: cholinergic synapse - - choline acetyltransferase - - immunocytochemistry - transplantation - - hippocampus - - electron microscopy

The ultrastructural features of the contacts established by intrahippocampal grafts of foetal septal/diagonal band neurones in the dentate gyrus and the CA 1 region of the previously denervated host hippocampus have been analysed with electron microscopic immunocytochemistry using a monoclonal antibody to choline acetyltransferase (CHAT). The results show that the grafted ChAT-positive neurones are capable of forming extensive synaptic contacts with neuronal targets in areas of the dentate gyrus and CA1 which normally receive such innervation. While all types of contacts normally found in association with the granule and pyramidal cell layers were also present in the graft-reinnervated specimens, the quantitative relationship between somatic and dendritic synapses was abnormal. Thus, the ChAT-immunoreactive synapses on cell bodies, which amounted to only a few percent in the normal animal, constituted over 60% in the grafted animals. Conversely, synapses on dendrites which constituted over 90% in the normal dentate were reduced to less than 40% in the grafted animals. The postsynaptic targets of the graft-derived cholinergic synapses included dendrites and cell bodies of dentate granule cells and CA1 pyramidal cells. This supports previous electrophysiologicalstudies and indicates that the septal grafts may be able to modulate host hippocampal function via direct efferent connections onto the granule and pyramidal neurones in the host hippocampal formation. INTRODUCTION Previous studies 5-7 have shown that grafts of foetal septal/diagonal band (S-DB) cholinergic neurones can re-establish a new acetylcholinesterase (ACHE) positive input to the h i p p o c a m p a l formation in rats with lesions of the intrinsic s e p t o h i p p o c a m p a l cholinergic pathways. The S-DB tissue is grafted either as a solid piece, placed within a surgical cavity in the fimbria-fornix, or as dissociated cell suspensions injected directly into the h i p p o c a m p a l formation. In both procedures the grafted cholinergic neurones ex-. tend fibres into the previously d e n e r v a t e d hippocampal target, and within 2 - 4 months a new laminated A C h E - p o s i t i v e terminal pattern is established, which closely mimics that of the normal S-DB cholinergic innervation. There is evidence from behavioural studies 15.19.27, electrophysiological recordings 27,40, measurements of acetylcholine turnover 6 and region-

al glucose metabolism 22 that the 'new' septohippocampal projection f o r m e d from the grafts is functional. It remains unclear, however, to what extent the ability of septal grafts to influence the function of the host h i p p o c a m p u s d e p e n d s on the formation of specific synaptic contacts, or whether the grafted neurones m a y function in a m o r e non-specific, neurohumoral m a n n e r by the non-synaptic release of the transmitter from their ingrowing axonal terminals. With the recent introduction of specific and sensitive antisera to the acetylcholine synthesizing enzyme, choline acetyltransferase (CHAT), the possibilities for studying the ultrastructural features of cholinergic synapses in the brain by immunocytochemistry has been greatly i m p r o v e d 48. M o r e specifically, this approach has recently been used to characterize the cholinergic innervation of the rat dentate gyrus at the ultrastructural level 11. In the present study we have applied C h A T immunocytochemistry

* Present address: Dept. of Neurosciences, UCSD, La Jolla, California 92093, U.S.A. Correspondence: D.J. Clarke, Univ. Department of Pharmacology, Oxford OX1 3QT, England. 0006-8993/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)

152 to study the ultrastructural contacts of the graft-derived cholinergic innervation in rats with intrahippocampal S-DB transplants. MATERIALS AND METHODS A total of 19 young female Sprague-Dawley rats (ALAB, Stockholm, Sweden; 180-200 g at the time of surgery) were used.

Surgery All rats were subjected to a unilateral aspirative lesion of the fimbria-fornix and the overlying cingulate cortex (including the supracallosal striae) under methyl hexital (Brietal, Lilly) anaesthesia, as described previously 7,1s.43. This lesion, which is done under visual control in the operating microscope, reliably produces a virtually complete cholinergic denervation of the dorsal and caudal parts of the ipsilateral hippocampal formation, while about 15% of the cholinergic innervation of the ventral (temporal) part is spared 7,18. The lesioned rats were divided into 3 groups: (a) Lesioned-only controls: 5 rats received no transplants and served as a control for the efficiency of the lesion to remove the ChAT-immunoreactive fibres in the hippocampus. (b) S-DB suspension grafts: 8 rats received two injections of a cell suspension into the denervated dorsal hippocampus (ipsilateral to the lesion) in the same surgical session as the fimbria-fornix lesion. The suspension was prepared from the S-DB region of 14-15 day old rat foetuses of the same strain according to the technique of Bj6rklund et al. s. Each injection consisted of 3/~1 and was deposited stereotaxicaUy over 5 min with a Hamilton microsyringe at the following two sites: (i) A +4.5 mm (rostral to interaural line), L 3.5 mm, V 3.7 mm below dura; (ii) A +3.0 mm, L 3.7 mm, V 3.7 mm. Incisor bar was set at the level of the interaural line. (c) S-DB so.lid grafts: 6 rats received a graft of SDB tissue obtained from one 15-16-day-old rat foetus of the same strain in the same surgical session that the fimbria-fornix cavity was made, according to the technique of Stenevi et al. 43. The tissue piece was placed on the dorsal thalamic surface, which formed the bottom of the aspiration cavity, in contact with the cut rostral hippocampal surface. The cavity was

then filled with gel-foam and the wound closed.

lmmunocytochemistrv Four months following the transplantation surgery the animals were anaesthetized and then perfused through the heart first with approximately 50 ml 0.9% saline (w/v) and then with 250-3(10 ml of fixative containing 2% paraformaldehyde (Aldrich) and 0.1% glutaraldehyde (Taab) in 0.1 M phosphate buffer (PB, pH 7.4) over a period of 15-18 rain. A wide bore needle (18 gauge) and a fast flow rate of fixative through the animal ensured good retention of antigenicity and ultrastructural preservation. No postfixation was required using this method. A tissue block containing the hippocampal formation and the graft was dissected out. In many cases it was possible to locate the solid transplants from the surface of the intact brain. The hippocampus was then sectioned at 70 /*m in cold PB using a Vibratome (Oxford Instruments) and sections collected into small vials containing cold phosphate-buffered saline (PBS, pH 7.4). Sections were thoroughly washed in cold PBS prior to incubation for 2(/-24 h at 4 °C in the primary antibody, which was a rat-mouse hybrid monoclonal antibody to CHAT2% diluted 1:500 in PBS. The immunocytochemical procedure is essentially the same as previously described 9.11.21 and employs the peroxidase-antiperoxidase (PAP) method developed by Sternberger a4. The second antibody in this reaction is a rabbit anti-rat lgG (Miles) used at a working dilution of 1:40 (in PBS) in which sections were shaken constantly for 12-14 h at 4 °C. Sections were then transferred to a rat PAP (Cuello, 1:30 dilution in PBS; or Sternberger-Meyer 1:100 dilution in PBS). Both PAP compounds required only a 3-h incubation period at room temperature. The peroxidase end product was visualized by reaction with 3,3'-diaminobenzidine-HCl (Sigma: 0.05% in Tris-HCl buffer, pH 7.4) using 0.01% H202 as a second substrate. Sections were viewed in the light microscope', sections containing immunoreactive neurones were immersed in 1% OsO4 (in PB, pH 7.4) which both enhances the reaction product and acts as final fixation step for electron microscopy. The sections were then dehydrated and flat-embedded in epoxy resin (Durcupan, ACM Fluka) 42. Counter-staining of the material with uranyl acetate (1%) for electron microscopy was included at the

153

Fig. 1. A: low power light micrograph showing integration of a S-DB suspension graft with host hippocampus. The approximate boundary of the graft is shown by dashed lines and numerous intensely ChAT-positive neurones are visible within the graft (arrow heads). Host granule cell layers (gcl) and CA4 are also indicated to reference the graft position; lgb, host lateral geniculate body. B: higher magnification light micrograph of a group of ChAT-immunoreactive neurones (arrow beads) situated at the suspension graft/ host interface. C: high power light micrograph of an immunostained neurone in the graft tissue. Both proximal and distal dendrites (d) are ChAT-positive. Other immunoreactive neurones and processes are also seen. D - F : high power light micrographs of ChAT-fibre staining (arrow heads) in the granule cell layer (gcl) under various experimental conditions. D represents the density of immunoreactive punctate structures (arrow heads) within the granule cell layer ipsilateral to the S-DB graft, whereas E represents the dentate gyrus contralateral to that of D and therefore exhibits the pattern of immunostaining reminiscent of the normal animal, granule cell layer (gcl) and molecular layer (ml). F shows the almost complete absence of immunoreactive fibres in the granule cell layer ipsilateral to the fimbria-fornix lesion, in a lesioned-only animal; scales: A, 180/~m; B: 60/~m; C, 20/~m; D - F , 15/~m.

154 70% ethanol stage during the dehydration procedure. The slides were carefully examined in the light microscope and the transplants were located. In 9 rats in which the immunohistochemical staining was judged to be acceptable (3 lesioned-only, 4 with suspension grafts and 2 with solid grafts) the areas of the dorsal hippocampal formation were re-embedded for electron microscopy together with the corresponding region from the contralateral intact sides, which served as a within-animal control. Ultrathin sections (silver grey, 70 nm) were cut on a Reichert ultramicrotome and collected on formvar-coated, single-slot copper grids for examination in a Philips 201C electron microscope. To enhance contrast, sections were stained on the grids with lead citrate 36. RESULTS In all 6 grafted rats included in the study (4 animals in which the suspension technique was used and 2 animals which had received solid transplants) good surviving grafts were identified at the light microscopic level. The suspension grafts occurred as one or several tissue masses which were partly integrated into the host dorsal hippocampus, predominantly in the area of the dentate gyrus (Fig. 1A). The solid grafts had adhered to the rostral cut surface of the hippocampus and part of the tissue was also apparent within the choroidal fissure underlying the dorsal hippocampal formation. Numerous densely immunostained neurones were visible in both types of grafts and many ChAT-positive neurones were distributed along the border between the graft and host tissue (Fig. 1B). Proximal dendrites were also intensely immunoreactire and occasionally more distal portions could also be traced (Fig. 1C). Immunoreactive processes could be traced from the graft into neighbouring host hip-

pocampus. In the dentate gyrus fine, varicose fibres intermingled within the granule cell layer (Fig. 1D) and small, immunostained punctate structures were seen in close apposition to neuronal perikarya. In two animals, fibre-staining extended into CA l of the main body of the hippocampus where a similar pattern of immunostaining was observed within the pyramidal cell layer. When corresponding areas of the dorsal dentate gyrus on the contralateral sides of the grafted animals were examined, fine immunoreactive fibres were also observed, although in these instances, fibres appeared slightly above the granule cell layer, at the interface between the granule cell layer and the molecular layer (Fig. 1E), consistent with findings from unoperated control animals published previously 11. A completely different C h A T immunostaining was noted in the 3 lesioned-only animals. In the dentate gyrus and CA1 ipsilateral to the fimbria-fornix lesion, no apparent immunoreactive fibres could be found as shown in granule cell layer in Fig. IF. However, occasional immunostained perikarya were visible in the hilar region and in a few cases punctate structures, immunoreactive for CHAT, were observed close by. Material obtained from 3 suspension-grafted, 2 solid-grafted and 2 lesioned-only animals was then examined in the electron microscope. The intact and experimentally manipulated hippocampi were studied in parallel in all cases to ensure correct interpretation of the findings. From the 5 grafted animals, 131 immunoreactive boutons making synaptic contact in the granule cell layer of the dentate gyrus and the pyramidal cell layer of CA1 were identified and a classification on the basis of their postsynaptic targets and types of synaptic specialization is given in Table I. No obvious differences were observed between the two types of

Fig. 2. This figure displays the variety of postsynaptic targets which receive symmetrical membrane contacts from ChAT-immunoreactive boutons in the dentate gyrus of grafted animals. A: low power electron micrograph of a symmetrical synapse contacting the perikaryon of a granule cell, in a solid-grafted animal. The boxed area is shown at higher magnification in B; n, nucleus. B: higher magnification of the symmetrical axosomatic synapse (arrow) made by the intensely ChAT-immunoreactive bouton (asterisk) boxed in A. C: ChAT-positive bouton (asterisk) making a symmetrical synapse (arrow) with a neuronal perikaryon (p) in a suspension-grafted animal. D: immunoreactive houton (asterisk) forming an axodendritic synapse (arrow) onto a large dendritic shaft (d) in the granule cell layer of a suspension-grafted animal. E: dendritic shaft (d) in the granule cell layer of a solid-grafted animal receiving a symmetrical synapse (arrow) from an immunoreactive bouton (asterisk) and 2 asymmetric contacts (double arrow heads) from non-immunoreactive boutons (white stars); scales: A, 1/~m; B, C and E, 0.25/~m; D, 0.4/~m.

155

156

Fig. 3. Asymmetric synaptic specializations from immunoreactive terminals contacted all types of postsynaptic target in the dentate gyrus. A: an immunoreactive bouton (asterisk) forms asymmetric synapses (double arrow heads) with two spines (S I and Sz). One of these ($2) emerges from the main dendritic shaft (d) in a suspension-grafted animal. B: dendritic shaft (d) in the granule cell layer of a solid-grafted animal receiving a large asymmetric synapse (double arrow head) from a ChAT-immunostained bouton (asterisk). C: graft-derived ChAT-positive bouton (asterisk) with an asymmetrical contact (double arrow head) onto a spine head (s). This synapse probably represents an ~en passant' contact; the preterminal fibre (f) is also visible in this figure. D: a large cholinergic bouton (asterisk) forms an asymmetrical axosomatic contact (double arrow head) with a neuronal perikaryon (p); scales: A, B and D, 0.25 urn; C, 0.20ktm.

157 TABLE I

Number and percentage distribution of ChA T-immunoreactive boutons on different types of postsynaptic target The values from grafted animals (n = 5) represent pooled counts from dentate gyrus and CA1. The values from normal animals are derived from the dentate gyrus (data from Clarkell).

Postsynaptic t a r g e t s

Fimbria-fornix lesioned animals with septal grafts Symmetric Number

Perikarya Dendritic shafts Dendritic spines Small dendrites or spines Non-identified Total

Asymmetric Percent

Number

Normal Total

Percent

Number

Percent

Total number

Percent

78 27 5 -

70.9 24.5 4.6 0 0

10 4 7 -

47.6 19.0 33.4 0 0

88 31 12 -

67.2 23.7 9.1 0 0

4 51 17 5 2

5.1 64.6 21.5 6.3 2.6

110

100.0

21

100.0

131

100.0

79

100.0

grafts and t h e y are thus dealt with t o g e t h e r . T h e b o u -

T h e b o u t o n s thus r e s e m b l e C h A T - i m m u n o r e a c t i v e

tons e x a m i n e d w e r e relatively small ( 0 . 4 - 0 . 6 5 / ~ m di-

b o u t o n s d e s c r i b e d in d e n t a t e gyrus as well as in o t h e r

a m e t e r ) and w e r e of b o t h the t e r m i n a l (Figs. 2 A , C

areas in the n o r m a l rat brainll.48. All of the i m m u n o -

and 3A) and 'en passant' types (Figs. 3C and 4B).

stained b o u t o n s c o n t a i n e d fairly large, r o u n d vesicles

Fig. 4. Area CA1 of the hippocampal formation was also studied in two of the grafted animals. Many examples were found, having an almost identical distribution in their postsynaptic targets and types of synaptic specializations as in the dentate gyrus. A: a large CHATpositive bouton (asterisk) forming an 'en passant' symmetrical synapse (arrow) onto a neuronal perikaryon (p) in the pyramidal cell layer of CA1. The nucleus (n) is also indicated. B: asymmetric synapses (double arrow head) were also present in the pyrami~lal cell layer of CA 1. In this example the ChAT-immunoreactive fibre (f) forms an 'en passant' contact with a perikaryon (p). The asterisk indicates the piece of fibre forming the synaptic specialization, though no bouton-like structure is present. C: dendritic shafts (d) also received asymmetrical synapses (double arrow head) from immunostained boutons (asterisk) in CA1; scales: A, 0.2 am; B and C, 0.25 am.

158 and many contained at least one mitochondrion. Membrane specializations were of both the svmmetrical and asymmetrical types though no obvious differences in bouton morphology could be observed between those making symmetrical contacts (84%) and the minority (16%) which formed asymmetric synapses. At the electron microscopical level, those varicosities seen interspersed within the granule cell layer in the light microscope were found to be forming synaptic contact with neuronal perikarya and the vast majority of the identified synaptic boutons in this study contacted cell bodies (67%; Figs. 2A, B, C and 4A). Most of these contacts were made onto neurones identified as granule cells, while others had a distinctly different ultrastructural appearance 2s. This compares with only 5% of the boutons examined in a previous study in the dentate gyrus of control rats which formed axosomatic contacts 11. A smaller percentage of synapses were axodendritic in the grafted animals (24%) than in normal animals (65%) and a similar decrease in the number of synapses onto dendritic spines was also observed (cf. 21% in normal animals vs only 9% in grafted animals, Table I). The active site of the symmetrical contact ranged from 0.2-0.25 # m in length, a dimension very similar to that of control animals. However, the asymmetric

NORMAL

® \

.

I\

synapses had larger active sites, measuring 0.3-0.55 u m in length. Boutons forming asymmetric specializations, like those forming the symmetrical synapses, were found to contact neuronal perikarya, dendritic shafts and spines, though a greater proportion of the asymmetrical type were in contact with dendritic spines (Fig. 3A, B). Again a similarity is seen with the normal animal where many of the axospinous contacts, although classified as symmetric, had a greater density of postsynaptic thickening than a classically defined symmetrical synapse. In two grafted animals some immunoreactive varicosities were present in the CA1 and these were examined ultrastructurally. Essentially similar results to those of the granule cell layer were obtained, in that the majority of boutons contacted pyramidal neurone cell somata, as identified from their ultrastructural morphology and most synaptic specializations were of the symmetrical type (Fig. 4A). A few boutons did form asymmetrical contacts (Fig. 4B, C); however, their percentage of the total synaptic input in CA1 was similar to that seen in the total material given in Table I. In summary, the results indicate that cholinergic neurones in the grafts form synaptic contacts with

@ i

From ERC

SEPTAL GRAFT From ERC

From Sept a{ Graft

From Host Septum

) Fig. 5. Schematic representation of the proposed principal postsynaptic targets of the cholinergic innervation in the dentate gyrus in normal rats (A) and in denervated rats, reinnervated by S-DB grafts (B). Two types of target ceils are illustrated, a granule cell and a hilar interneurone. The synaptic input (non-cholinergic) from the entorhinal cortex (ERC), normally terminating on the distal dendrites of the granule cells, is also included.

159 host hippocampus. Their postsynaptic targets are, however, proportionally different to that seen in the same area in normal rats and the contralateral sides of the brains of grafted animals in this study. Another anomaly also became apparent when the types of synaptic specializations were considered: the grafted animals had a small percentage of the input to all targets which was asymmetric, a type of cholinergic synapse rarely found in the normal dentate gyrus. DISCUSSION

Normal synaptic arrangement In normal animals both AChE histochemistry 2.45.46 and ChAT immunohistochemistry 11,20 indicate that the cholinergic afferents to the dentate gyrus terminate primarily in the supra- and infragranular zones, and to a lesser extent within the granule cell layer. From studies using lesions or anterograde tracing techniques it seems that the vast majority of these cholinergic afferents originate in the ipsilateral septai-diagonal band area 13,28,33,34. Complete transection of the septohippocampal axons running in the fimbria-fornix and the supracallosal striae, by an aspiration lesion of the type used in the present study, reduces ChAT activity in the dorsocaudal hippocampus and dentate gyrus by about 95% 6,18. Consistent with this, the present study showed the near total disappearance of ChAT-immunoreactive fibres in dorsocaudai hippocampus ipsilateral to the fimbria-fornix lesion. As summarized in the right-hand panel in Table I, ChAT immunocytochemistry of the normal dentate gyrus 11 has shown that the cholinergic fibres make predominantly symmetric synaptic contacts with dendrites and dendritic spines and to a lesser degree with cell bodies, within the supragranular zone. This is consistent with previous ultrastructurai studies using AChE-staining 4j or anterograde HRP tracing ~0, although these authors did not identify any symmetric synapses in their material. Chandler and Crutcher 10 reported a higher proportion of axosomatic contacts, but this discrepancy may be due to the fact that the major portion of the synapses characterized by these authors were located within the hilar zone. The postsynaptic targets of the normal cholinergic S-DB afferents clearly include the granule cells, but according to Chandler and Crutcherl0, hilar neurones, lo-

cated in the infragranular zone and pyramidal basket cells receive S-DB synapses as well.

Synaptic arrangement in the graft-reinnervated animals Light microscopically there was a marked recovery of ChAT positive fibres and terminals in the dentate gyrus and the CA1 region of the grafted animals, although the overall fibre density was lower than normal. Consistent with previous histochemical studies using AChE stainingS.7 the dense fibre input was around the granule cell layer in the dentate gyrus and around the pyramidal cell layer in CA1, i.e. regions which normally receive dense AChE-positive or ChAT-immunoreactive innervation. The ultrastructural findings show that the grafted cholinergic neurones are capable of forming abundant synapses with various types of neuronal targets in the previously denervated dentate gyrus. The synaptic arrangements were similar in solid-grafted and suspension-grafted animals. Together with similar results obtained with grafts of other types of neurones 3.17,35 the present observations provide evidence that extensive synaptogenesis from ingrowing axons is possible in the fully mature central nervous system, at least within a denervated region. The most striking finding in the present study was that all types of contacts normally occurring in association with the dentate granule cell layer were also present in the graft-reinnervated specimens, but that the quantitative relationships were abnormal. Thus, while synapses onto cell bodies normally constituted only about 5% in the normal dentate, they represented over 60% in the grafted animals, and the proportion of synapses onto dendrites (shafts or spines) was correspondingly reduced. Although the overall ChAT-positive synapse density was lower than normal it is our impression that the density of axosomatic synapses was higher than normal, but that they did not occur at the expense of the axodendritic contacts. This implies that the cell bodies in the area (part of which at least, were identical to granule cell perikarya) received a hyperinnervation from the S-DB grafts. The reason why this anomalous termination occurred is not clear. Since the granule cell somata normally are richly innervated, although not by cholinergic afferents, it seems possible that the formation of somatic contacts by the S-DB grafts could re-

16(I flect the filling of vacated terminal space created by the removal of non-cholinergic inputs (e.g. hypothalamohippocampal pathway, see ref. 14) by the timbria-fornix lesion. Interestingly, this perikaryal hyperinnervation pattern is similar to that found in the caudate-putamen in animals with intrastriatal grafts of foetal substantia nigra 17. In such animals the grafted dopamine neurones have been observed to form abnormal pericellular arrangements around the cholinergic giant interneurones, i.e. around cells which normally receive dopaminergic afferents out on their dendrites but very few axosomatic synapses. Since in the nigragrafted animal this hyperinnervation occurred selectively on the small subpopulation of giant neurones it was suggested that the perikaryal hyperinnervation may serve a functional purpose in that it could represent a powerful mechanism whereby the grafted dopaminergic neurones with a minimum of divergence could regain normal inhibitory control over striatal cholinergic transmission. According to this view, perikaryal hyperinnervation could provide a general mechanism whereby a subnormal density of regenerated afferents could restore relatively normal function. Another abnormality in the grafted animals was the relative abundance of asymmetric contacts. Other electron microscopical studies, using injection of anterograde tracers into the septal area 1°,38 have reported asymmetric types of synapses on cell bodies and dendrites in the hippocampus and dentate gyrus. Since asymmetric synapses are rarely seen at cholinergic terminals identified by ChAT immunocytochemistry 11~4s it seems possible that the asymmetric synapses that can be labelled anterogradely from the septum represent non-cholinergic afferents i,14,47. If so, it is conceivable that the asymmetric contacts formed by the grafts represent the taking over of asymmetric sites on target neurones vacated by the removal of non-cholinergic afferents 32. An alternative explanation may be that the asymmetric synapses on perikarya and dendritic shafts in the grafted animals reflect a relative immaturity of the graft-derived reinnervation. Ontogenetic studies on synaptogenesis in cortical regions may support this view 3° 32.

Functional implications Electrophysiological studies of the graft-to-host

connections in S-DB-grafted animals have been carried out in those two hippocampal regions, CAI and dentate gyrus, which were analysed in the present study. In a study using intracelhflar recordings in slices of the grafted hippocampus, Segal et al? ° could demonstrate synaptic responses in CA1 pyramidal neurones upon stimulation of the S-DB grafts. Two types of responses were recorded: first a slow, longlasting and voltage-dependent depolarization (persisting for 2-5 min) associated with an increase in spontaneous discharge of the excited neurones, and secondly a blockade of the afterhyperpolarization triggered by a burst of action potentials, possibly mediated through a blockade of a CaZ+-dependent potassium current. These responses, which mimic the ones seen after acetylcholine application in the normal animaP 2,39 were blocked by atropine and potentiated by physostigmine indicating that they were mediated by cholinergic muscarinic synapses established onto the CA1 pyramidal neurones. In another study, performed in S-DB-grafted animals in vivo with field potential electrodes, Low et al. 27 recorded evoked synaptic responses in the dentate gyms upon electrical stimulation of the S-DB graft and they provided evidence that graft stimulation could reproduce the kind of heterosynaptic facilitation of the dentate granule cells which is a characteristic feature of the normal cholinergic septohippocampal pathway 16,23.2429,37.These effects could be mediated via direct excitatory effects on the granule cells and/or via a disinhibitory mechanism4, 23, as illustrated in Fig. 5. Against the background of these electrophysiological experiments it seems likely that the cholinergic synapses demonstrated in the present study can restore the modulatory facilitation normally exerted by the cholinergic afferents at the level of both the dentate gyrus and the hippocampal CA1 region and it is likely that the granule cells and the pyramidal cells themselves are among the postsynaptic targets for the graft-derived cholinergic innervation. In the intact animal, one of the major functions of the cholinergic septal afferents seems to be a permissive or activating action which regulates the responses of the dentate granule cells and the hippocampal pyramidal cells to their major excitatory inputs, such as the massive cortical input via the entorhinal perforant path. As illustrated in Fig. 5A, the

161 position of the cholinergic termination on the proximal parts of the granule cell dendrites would be ideally suited to exert this m o d u l a t o r y effect on the entorhinal cortical input terminating m o r e distally on the dendrites. In line with previous electrophysiological results23, 2429, an alternative disinhibitory mechanism, m e d i a t e d by cholinergic synapses on hilar interneurones, has also been indicated in the figure. Cholinergic deafferentation, induced by the fimbria-fornix lesion, is thus likely to result in impaired transmission through the cortico-dentato-hippocampo-subicular circuitry, leaving the hippocampus in a non-responsive or dysfunctioning state. Re-establishment of cholinergic facilitation by the S-DB grafts may thus restore h i p p o c a m p a l responsiveness to, e.g., cortical inputs. Such a relatively non-specific m o d u l a t o r y effect of the graft-derived cholinergic synapses may explain both the general metabolic re-activating effect observed in septal grafted animals with measure-

REFERENCES 1 Baisden, R.H., Woodruff, M.L. and HooveL D.B., Cholinergic and non-cholinergic septo-hippocampal projections: a double-label horseradish peroxidase-acetylcholinesterase study in the rabbit, Brain Research, 290 (1984) 146-151. 2 Bakst, I. and Amaral, D.G., The distribution of acetylcholinesterase in the hippocampal formation of the monkey, J. Comp. NeuroL, 225 (1984) 344-371. 3 Beebe, B.K., M~allg~rd, A., Bj6rklund, A. and Stenevi, U., Ultrastructural evidence of synaptogenesis in the adult rat dentate gyrus from brainstem implants, Brain Research, 167 (1979) 391-395. 4 Ben-Ari, Y., Krnjevic, K., Reinhardt, W. and Ropert, N., Intracellular observations on the disinhibitory action of acetylcholine in the hippocampus, Neuroscience, 6 (1981) 2475-2484. 5 BjOrklund, A., Gage, F.H., Schmidt, R.H., Stenevi, U. and Dunnett, S.B., Intracerebral grafting of neuronal cell suspensions. VII. Recovery of choline acetyltransferase activity and acetylcholine synthesis in the denervated hippocampus reinnervated by septal suspension implants, Acta Physiol. Scand., Suppl., 522 (1983) 59-66. 6 Bjfrklund, A., Gage, F.H., Stenevi, U. and Dunnett, S.B., Intracerebral grafting of neuronal cell suspensions. VI. Survival and growth of intrahippocampal implants of septal cell suspensions, Acta Physiol. Scand., Suppl., 522 (1983) 49-58. 7 Bj6rklund, A. and Stenevi, U., Reformation of the severed septohippocampal cholinergic pathway in the adult rat by transplanted septal neurons, Cell Tissue Res., 185 (1977) 289-302. 8 Bj6rklund, A., Stenevi, U., Schmidt, R.H., Dunnett, S.B. and Gage, F.H., Intracerebral grafting of neuronal cell sus-

ments of regional glucose utilization 22, as well as the positive effects of i n t r a h i p p o c a m p a l S-DB grafts on maze-learning behaviour in rats with bilateral fimbria-fornix lesions t5.27. ACKNOWLEDGEMENTS

We thank Alicja Flasch, Lesley Hayes and Agneta Persson for expert technical assistance. We are also most grateful to Dr. Bruce H. Wainer, Chicago, for supplying the choline acetyltransferase antibody and to Prof. David Smith for helpful discussions and support throughout this work. The study was supported by grants from the Swedish M R C (04X-3874) and the National Institutes of Aging ( A G 03766). D.J.C. is a Horace Le M a r q u a n d and D u d l e y Bigg research fellow of the Royal Society. Financial support from an ETP Twinning G r a n t is also gratefully acknowledged.

pensions. I. Introduction and general methods of preparation, Acta Physiol. Scan&, Suppl., 522 (1983) 1-8. 9 Bolam, J.P., Wainer, B.H. and Smith, A.D., Characterization of cholinergic neurons in the rat neostriatum. A combination of choline acetyltransferase immunocytochemistry, Golgi-impregnation and electron microscopy, Neuroscience, 12 (1984) 711-718. 10 Chandler, J.P. and Crutcher, K.A., The septohippocampal projection in the rat: an electron microscopic horseradish peroxidase study, Neuroscience, 10 (1983) 685-696. 11 Clarke, D.J., Cholinergic innervation of the rat dentate gyrus: a light and electron microscopical study using a monoclonal antibody to choline acetyltransferase, Brain Research, in press. 12 Cole, A.E. and Nicoll, R.A., Acetylcholine mediates a slow synaptic potential in hippocampal pyramidal cells, Science, 221 (1983) 1299-1301. 13 Crutcher, K.A., Madison, R. and Davis, J.N.. A study of the rat septohippocampal pathway using anterograde transport of horseradish peroxidase, Neuroscience, 10 (1981) 1961-1973. 14 Dent, J.A., Calvin, N.J., Stanfield, B.B. and Cowan, W.M., The mode of termination of the hypothalamic projection to the dentate gyrus - - an EM autoradiographic study, Brain Research, 258 (1983) 1-11. 15 Dunnett, S.B., Low, W.C., Iversen, S.D., Stenevi, U. and Bj6rklund, A., Septal transplants restore maze learning in rats with fornix-fimbria lesions, Brain Research, 251 (1982) 335-348. 16 Fantie, B.D. and Goddard, G.V., Septal modulation of the population spike in the fascia dentata produced by perforant path stimulation in the rat. Brain Research, 252 (1982) 227-237. 17 Freund, T.F., Bolam, J.P., Bj6rklund, A., Stenevi, U., Dunnett, S.B. and Smith, A.D., Synaptic connections ofni-

162 gral transplants rcinnervating the host neostriatum: a THimmunocytochemical study, J. Neurosci., 5 (1985) 603616. 18 Gage, F.H., Bj6rklund, A. and Stenevi, U., Reinnervation of the partially deafferented hippocampus by compensatory collateral sprouting from spared cholinergic and noradrenergic afferents, Brain Research. 268 (1983) 27-37. 19 Gage, F.H., Bj6rklund, A., Stenevi, U., Dunnett, S.B. and Kelly, P.A.T., Intrahippocampal septal grafts ameliorate learning impairments in aged rats, Science, 225 (1984) 533-536. 20 Houser, C.R., Crawford, G.D., Barber, R.P., Salvaterra, P.M. and Vaughn, J.E., Organization and morophological characteristics of cholinergic neurons: an immunocytochemical study with a monoclonal antibody to choline acetyltransferase, Brain Research, 266 (1983) 211-215. 21 lngham, C.A., Bolam, J.P., Wainer, B.H. and Smith, A.D., A correlated light and electron microscopic study of cholinergic neurons that project from the basal forebrain to frontal cortex in the rat, J. Comp. Neurol., 239 (1985) 176-192. 22 Kelly, P.A.T., Gage, F.H., Ingvar, M., Lindvall, O., Stenevi, U. and Bj6rklund, A., Functional reactivation of the deafferented hippocampus by embryonic septal grafts as assessed by measurements of local glucose utilization, Exp. Brain Res.. 58 (1985) 570-579. 23 Krnjcvic, K., Reiffenstein, R.J. and Ropert, N., Disinhibitory action of acetylcholine in the rat's hippocampus: extracellular observations, Neuroscience, 6 (1981) 2465-2474. 24 Krnjevic, K. and Ropert, N., Electrophysiological and pharmacological characteristics of facilitation of hippocampat population spikes by stimulation of the medial septum, Neuroscience, 7 (1982) 2165-2183. 25 Laatsch, R.H. and Cowan, W.M., Electron microscopic studies of the dentate gyrus of the rat. I. Normal structure, J. Comp. Neurol., 128 (1966) 359-396. 26 Levey, A.I., Armstrong, D.M., Atweh, S.F., "Ferry, R.D. and Wainer, B.H., Monoclonal antibodies to choline acetyltransferase: production, specificity and immunohistochemistry, J. Neurosci., 3 (1983) 1-9. 27 Low, W.C., Lewis, P.R., Bunch, S.T., Dunnett, S.B., Thomas, S.R., Iversen, S.D., BjOrklund, A. and Stenevi, U., Functional recovery following transplantation of embryonic septal nuclei into adult rats with septohippocampal lesions: the recovery of function, Nature (London), 300 (1982) 260-262. 28 Lynch, G., Rose, G. and Gall, C., Anatomical and functional aspects of the septo-hippocampal projections. In K. Elliot and J. Whelan (Eds.), Functions of the Septo-Hippocampal System, Elsevier, Amsterdam. 1978, pp. 5-24. 29 McNaughton, N. and Miller, J.J., Medial septal projections to the dentate gyrus of the rat: electrophysiological analysis of distribution and plasticity, Exp. Brain Res., 56 (1984) 243-256. 30 Mates, S.L. and Lund, J.S., Spine formation and maturation of type 1 synapses on spiny stellate neurons in primate visual cortex, J. Comp. Neurol., 221 (1983) 91-97. 31 Mates, S.L. and Lund, J.S., Developmental changes in the relationship between type 2 synapses and spiny neurons in the monkey visual cortex, J. Comp. Neurol., 221 (1983) 98-105. 32 Matthews, D.A., Cotman, C. and Lynch, G., An electron microscopic study of lesion-induced synaptogenesis in the dentate gyrus of the adult rat. II. Reappearance of morphologically normal synaptic contacts, Brain Research, 115

(1976) 23-41. 33 Mellgren, S.I. and Srebro, B., Changes in acetylcholinestcrase and distribution of degenerating fibres in thc hippocampal region after septal lesions in the rat, Brain Research, 52 (1973) 19-36. 34 Mosko, S., Lynch, G. aml Cotman, C.W., The distribution of septal projections to the hippocampus of the rat, J. Comp. Neurol., 152 (t973) 163-174. 35 Raisman, G. and Ebner, F.F., Mossy fibre projections into and out of hippocampal transplants, Neuroscience, 9 (t983) 783-801. 36 Reynolds, E.S., The use of lead citrate at high pH as an electron opaque stain in electron microscopy, J. Cell Biol.. 17 (1963) 208-212. 37 Robinson, G.B. and Racine, R.J., Heterosynaptic interactions between septal and entorhinal inputs to the dentate gyrus: long-term potentiation effects, Brain Research, 249 (1982) 162-166. 38 Rose, A.M., Hattori, T. and Fibiger, H.C., Analysis of the septohippocampal pathway by light and electron microscope autoradiography, Brain Research, 108 (1976) 170-174. 39 Segal, M., Multiple actions of acetylcholine at a muscarinic receptor studied in the rat hippocampal slice, Brain Research, 246 (1982) 77-87. 40 Segal, M., Bj6rklund, A. and Gage, F.H., Intracellular analysis of cholinergic synapses between grafted septal nucleus and host hippocampal pyramidal neurons. In A. Bj6rklund and U. Stenevi (Eds.), Neural Grafting in the Mammalian CNS, Elsevier, Amsterdam, 1985, pp. 389-399. 41 Shute, C.C.D. and Lewis, P.R., Electron microscopy of cholinergic terminals and acetylcholinesterase-containing neurons in the hippocampal formation of the rat, Z. ZellJorsch. Mikrosk., 69 (1966) 334-343. 42 Somogyi, P., Hodgson, A.J. and Smith, A.D., An approach to tracing neuron networks in the cerebral cortex and basal ganglia. Combination of Golgi-staining, retrograde tracing of horseradish peroxidase and anterograde degeneration of synaptic boutons in the same material, Neuroscience, 4 (1979) 1805-1852. 43 Stenevi, U., Bj6rklund, A. and Svendgaard, N.-Aa., Transplantation of central and peripheral monoamine neurons to the adult rat brain: techniques and conditions for survival, Brain Research, 114 (t976) 1-20. 44 Sternberger, L.A., Hardy, P.H., Cuculis, J.J. and Meyer, H.G., The unlabeled antibody enzyme method of immunocytochemistry: preparation and properties of antigen-antibody complex (horseradish peroxidase-antihorseradish peroxidase) and its use in identification of spirochetes, J. Histochem. Cytochem., t8 (1970) 315-333. 45 Storm-Mathisen, J., Quantitative histochemistry of acetylcholinesterase in rat hippocampal region correlated to histochemical staining, J. Neurochem., 17 (1980) 739-750. 46 Storm-Mathisen, J. and Blackstad, T.W:, Cholinesterase in the hippocampal region. Distribution and relation to architectonics and afferent systems, Acta Anat., 56 (1964) 216-253. 47 Vincent, S.R. and McGeer, E.G., A substance P projection to the hippocampus, Brain Research, 215 (1981) 349-351. 48 Wainer, B.H., Bolam, J.P., Freund, T.F., Henderson, Z., Totterdell, S. and Smith, A.D., Cholinergic synapses in the rat brain: a correlated light and electron microscopic study employing a monoclonal antibody against choline acetyltransferase, Brain Research. 308 (1984) 69-76.