- Email: [email protected]
Formation of synaptic graft-host connections by noradrenergic locus coeruleus neurons transplanted into the adult rat hippocampus
EXPERIMENTAL
NEUROLOGY
110,258-267
(1990)
Formation of Synaptic Graft-Host Connections by Noradrenergic Locus Coeruleus Neurons Transplanted into the Adult Rat Hippocampus Yuzo MURATA,*
TANEMICHI
*Department
CHIBA,**~ PATRIK BRUNDIN,~
of Anatomy, Saga Medical School, Nabeshima, of Medical Cell Research, Section of Neurobiology,
Press,
Saga 840-01, Japan; University of Lund,
MATERIALS
Inc.
INTRODUCTION Grafts of noradrenergic neurons, obtained from the pontine locus coeruleus region of rat fetuses, have been 1 Present address: Department Chiba University, Inohana, Chiba 0014-4686190 $3.00 Copyright 0 1990 by Academic Press, All rights of reproduction in any form
of Anatomy, School 280, Japan.
of Medicine,
258 Inc. reserved.
AND OLLE LINDvALLt
and TDepartment Sweden
used to correct for noradrenergic deficits in the brain or spinal cord in rats with a preceding neurotoxic lesion of the intrinsic noradrenergic system (2,3,4,6,19,22,23). Previous studies have shown that the grafted locus coeruleus neurons can provide a new noradrenergic terminal network in a previously denervated target area (3, 4,19), and there is evidence from several experimental models that the grafts can be functional in the normalization of noradrenaline-dependent behavioral deficits, such as spinal reflexes and locomotor activity in the lesioned spinal cord (6, 23), avoidance learning in aged rats (9), open field locomotion in 6-hydroxydopamine (B-OHDA)-lesioned rats (22), and development of epileptic phenomena in 6-OHDA lesioned (2) or fimbriafornix lesioned rats (7). The prominent axonal outgrowth observed from grafted noradrenergic neurons suggeststhat the graftinduced functional effects may depend on the establishment of functional graft-host connections. Although there is electrophysiological evidence that grafted locus coeruleus neurons can form functional contacts with the host (4, 21), these efferent projections have so far not been investigated ultrastructurally. Moreover, it is not known to what extent the ability of intracerebrally implanted noradrenergic neurons to reinnervate the host depends on whether or not the intrinsic noradrenergic innervation of the grafted area is removed prior to transplantation. In the present study we have used antisera against tyrosine hydroxylase (TH) and noradrenaline (NA) on the light and electron microscopic levels to investigate the growth and synaptic connectivity of noradrenergic locus coeruleus neurons, transplanted as a cell suspension into the hippocampal formation, in the presence or absence of an intrinsic noradrenergic afferent input.
Transplants of cell suspension obtained from the locus coeruleus region of 13- to 14-day-old rat fetuses were implanted into the hippocampal formation of intact adult rats or rats from which the noradrenergic afferents to the hippocampus had been removed by bilateral 6-hydroxydopamine (6-OHDA) injections into the dorsal tegmental noradrenergic bundle. The growth noradrenergic axons into the host hippocampus from the implant was studied at 4-8 months after surgery by immunohistochemistry using antisera raised against tyrosine hydroxylase or noradrenaline. In the animals with an intact noradrenergic system the host noradrenergic afferents were removed by bilateral dorsal bundle lesions 2 weeks before sacrifice. Fine axon-like fibers (diameter about 0.3 pm) and thick dendrite-like fibers (diameter about 1.3 pm), labeled immunohistochemitally, were abundant and spread far from the graft. By electron microscopy, immunolabeled axon-like fibers formed mostly symmetrical synaptic contacts with nonlabeled spines and shafts of dendrites in the host. Labeled dendrite-like fibers of presumed graft origin penetrated deep into the host neuropil and received abundant afferents from nonlabeled axon terminals. The extent of graft-derived noradrenergic axons and the synapses established with the host hippocampal neurons were similar in the chronically denervated animals and in the animals where the intrinsic noradrenergic afferents had been left intact until 2 weeks before sacrifice. The results show that implanted embryonic noradrenergic neurons are able to innervate the hippocampus in both the presence and the absence of an intact intrinsic noradrenergic innervation and that the ingrowing axons form abundant synaptic connections with the host hippocampal neurons under both conditions, Dendritic processes from the grafted noradrenergic neurons that extend deep into the host tissue may receive a reciprocal synaptic host afferent input. o 1990 Academic
ANDERS BJijRRLuND,t
AND METHODS
Surgery. The study was performed on young adult female Sprague-Dawley rats, X30-200 g at the time of surgery. There were two groups. The first group (n = 8) received injections of 8 pg 6-hydroxydopamine HC1(6OHDA) dissolved in 4 ~1ascorbate-saline (0.2 mg ascorbic acid/ml) bilaterally into the tegmental noradrener-
FORMATION
OF
SYNAPTIC
GRAFT-HOST
CONNECTIONS
259
= 9) had an intact noradrenergic system at the time of transplantation and was treated with 6-OHDA as above 2 weeks before perfusion (i.e., at least 4 months after transplantation) to eliminate the host noradrenergic innervation of the hippocampal formation. This time inwas terval between 6-OHDA injection and perfusion chosen to obtain a complete degeneration of the host locus coeruleus terminals (14). The completeness of the 6-OHDA lesions was confirmed by immunohistochemistry, demonstrating only single remaining terminals in the neocortex of the rats included in the study. The re-
C
FIG. 1. TH-immunolabeled (a) and NA-immunolabeled (b) cell bodies and outgrowing fibers of a locus coeruleus graft implanted into the CA3 area of hippocampal formation of a rat which had received a bilateral 6-OHDA lesion of dorsal NA bundle 2 weeks before transplantation (8 months post-transplant survival). (a, b) From adjacent sections. The graft occurred attached to the alveus (a) and thick dendrite-like fibers (arrows) and thin axon-like fibers (arrowheads) were seen to extend into the host (left) from the transplant (T). a, alveus; o, oriens layer; p, pyramidal cell layer. Scale bar = 100 pm. FIG. 2. TH-immunolabeled cell bodies and outgrowing fibers in a locus coeruleus graft (4 months survival) implanted into the hippocampal formation of a rat which received a bilateral 6-OHDA lesion of the dorsal NA bundle 2 weeks before sacrifice. A thick dendrite-like fiber (arrow) can be seen to run more than 500 pm into the host hippocampus. Scale bar = 100 Wm.
gic bundle. The coordinates were (tooth-bar = 0) A = +0.5 mm from the interaural line, L = -+l.l mm, and V = 6.5 mm down from dura. The injections were given 2 weeks before transplantation. The second group (n
FIG. 3. Distribution of NA-immunolabeled fibers in the normal hippocampus (C) and in hippocampi (D, I) reinnervated by locus coeruleus implants, 8 months survival. The specimen D was noradrenergically denervated with 6-OHDA prior to transplantation. The specimen I was denervated with 6-OHDA 2 weeks prior to sacrifice. The graft-derived fibers in both cases had their densest distribution in the CA3 and the dentate hilus. The grafts, delineated with a dashed line, in both cases were located in the CA3 area. Scale bar = 1 mm.
260
MURATA
ET
AL.
Immunohistochemistry. The rats for immunohistochemistry were processed for the visualization of TH and NA according to McRae-Degueurce and Geffard (17). In brief, the animals were perfused with 30 ml of saline, followed by 500 ml of 5% glutaraldehyde with 1% sodium metabisulfite (SMB) in 0.1 A4 cacodylate buffer, pH 7.0-7.5. The brains were cut into small pieces and immersed in Tris-SMB buffer (0.05 A4 Tris and, 1% SMB pH 7.5) with 30% sucrose. The tissues were frozen in liquid nitrogen for several seconds and thawed. Sections (40 pm) were cut on a Vibratome (Oxford Instruments) and collected in Tris-SMB buffer. The sections were incubated in 1% sodium borohydride for 30 min at room temperature and washed several times in TrisSMB. They were then incubated at 4°C for at least 15 h with one of the two primary antisera raised in rabbits: anti-TH (dilution, 1:2000; Eugene Tech International, Inc., Allendale, SC) and anti-NA (dilution, 1:3000; SFRI Lavoratoire, Martignal, France). The antibodies were diluted in a solution of Tris-SMB containing 1% nonspecific serum. The immunocytochemistry procedure was then continued with the PAP method. From the second antibody incubation, Tris-NaCl (0.05 M Tris and 0.9% NaCl, pH 7.6) was used as the diluting buffer and rinsing buffer. Following the DAB reaction, the sections were rinsed briefly in 0.15 M phosphate buffer (pH 7.4), postfixed with 1% 0~0, in 0.1 M phosphate buffer for 1 h, dehydrated in ethanol, and embedded in Spurr’s resin between sheets of transparent plastic film (ACLAR). After light microscopic observation and photography, thin sections were made, mounted on Formvar-coated grids with a single hole, stained with uranyl acetate and lead citrate, and then analyzed by electron microscopy. Sections stained with the omission of the primary antiserum were used as controls for the specificity of the observed immunoreactivity. FIG. 4. NA-immunolabeled fibers in the CA3 area from the three animals shown in Fig. 3. C, normal control rat; D, grafted rat denervated prior to transplantation; I, grafted rat denervated 2 weeks before sacrifice. o, oriens layer; p, pyramidal cell layer; r, stratum radiaturn. Scale bar = 100 pm.
gion containing the locus coeruleus was dissected from the rhombencephalon of 13- to 14-day-old rat fetuses (crown-rump length, CRL, 10-13.5 mm). A suspension was made according to the procedure of Bjijrklund et al. (5) without the trypsin incubation step. Two 1.5- to 2-/*l aliquots of cell suspension were injected unilaterally into the hippocampal formation in eight 6-OHDA-pretreated rats and nine intact rats at the following stereotaxic coordinates (in mm): A = +4.3 (rostra1 to interaural line), L = 3.0, and V = 3.0 below dura. The rats were perfused for immunohistochemistry at 3-8 months after transplantation. Normal intact rats were proceeded in parallel.
RESULTS Light Microscopy In the TH-immunostained sections (Figs. la and 2) cell bodies were labeled intensely in the grafts, and thick dendrite-like fibers, thin axon-like fibers, and terminals occurred in large numbers in the host hippocampus surrounding the implant. The NA antiserum (Fig. lb) stained cell bodies and dendrite-like processes more lightly, but axon-like fibers and terminals were observed with the same intensity, abundance, and distribution as those visualized with the TH antiserum. The number of graft cell bodies visualized with the TH and NA antisera in adjacent sections (Fig. 1) was similar with both procedures, which suggests that the vast majority of the TH-positive neurons in the grafts were noradrenergic.
FORMATION
OF
SYNAPTIC
GRAFT-HOST
CONNECTIONS
261
1FIG. 5. NA-immunolabeled fibers in the dentate hilus (h) from the three animals shown in Fig. 3. In both the normal control (C) and the fibers are most prominent in the zone below the granule cell layer (g). Scale bar = 100 pm. w Ifted rats (D, and I) the immunolabeled
Fiber outgrowth occurred throughout the dorsal hippocampal formation (Fig. 3). It was particularly dense in the dentate hilus and the CA3 area where they formed a fine varicose terminal network in close association with the neuronal elements by the host (Figs. 4 and 5). The magnitude and distribution of graft-derived THand NA-positive fibers in the host hippocampus appeared to be similar in the rats denervated prior to transplantation and in those denervated 2 weeks before the perfusion (cf. pictures D and I in Fig. 3). In both groups, thick dendrite-like fibers (diameter about 1.3 pm) were seen to extend from the grafts into the host hippocampus for distance up to about 500 pm (Fig. 2). The distribution of NA- and TH-immunoreactive fibers in the grafted hippocampi was not clearly different from that seen in the normal controls. The overall fiber density was lower than normal in the grafted specimens. However, in the areas close to the grafts, which tended to be more densely innervated, the graft-derived terminal network was in some cases clearly denser than normal. The number of surviving immunolabeled neuronal somata varied between 20 and 300 per implantation in the rats denervated prior to the transplantation, with an average of 90.0 + 90.2 (mean + SD, n = B), and between 10 and 70 per implantation site, with an average of 38.8
+ 19.0 (mean -t SD, rz = 9) in the rats denervated 2 weeks before perfusion. The differences between the groups, however, were not statistically significant (Welch t test and Wilcoxon rank-sum test). Electron Microscopy The fine structural analysis was focused on the CA3 region and the hilus of the dentate gyrus, i.e., the areas that had the densest TH- and NA-positive fiber networks in the light microscope. Immunostained axonal boutons as well as coarse dendrite-like structures were observed with both antisera within the host. The THpositive structures were sometimes so heavily labeled that they obscured the internal structures of the stained profiles, whereas the NA-positive boutons typically were more lightly stained (e.g., Fig. 7). Consistent with the light microscopic picture, dendrite-like structures occurred in the areas close to the grafts, whereas the labeled boutons occurred also at greater distances from the grafts. Immunoreactive dendrite-like processes that penetrated deep into the host neuropil were seen to receive abundant synaptic contacts by immunonegative axon terminals (Fig. 6). The labeled boutons were frequently seen to form synaptic contacts with both spines (Figs. 7a and 9) and
262
FIG. 6. TH-immunolabeled dendrites that graft. The picture is from the stratum radiatum transplant survival). Scale bar = 1 pm.
MURATA
ET
AL.
form synapses (arrowheads) with unlabeled vesicle-filled terminals at the periphery of the CA3 area from a grafted animal denervated before transplantation (5 months
shafts (Figs. 7b and 8) of nonlabeled dendrites in both CA3 and the dentate hilus. In the cases where both the pre- and the postsynaptic elements were discernible, most synapses appeared to be of the symmetric type, and only rarely were clear asymmetrical synapses observed. Some of the boutons were large and the majority were small (about 0.3 pm in diameter), thus corresponding well to the fine-caliber varicose terminals that formed the fine varicose terminal networks at the light microscopic level. These graft-derived synaptic boutons had the same general morphological appearance as the noradrenergic synaptic terminals observed in the intact control hippocampi (Fig. 10). In the dentate hilus, labeled boutons were seen to make synapses with spines that, in addition, received synapses from a giant immunonegative mossy fiber terminal (Fig. 9). Since pyramidal cell dendrites constitute the predominant postsynaptic target of the mossy fibers, this arrangement indicates that the grafted noradrenergic neurons had innervated host pyramidal neurons. Detailed comparison between the hippocampi denervated prior to transplantation and those denervated 2 weeks before the perfusion revealed no clear-cut differences. Synaptic contacts appeared to be equally abundant in both situations and the types of synaptic con-
of the post-
tacts formed in the CA3 and dentate hilus were similar under both conditions.
DISCUSSION The present results demonstrate that implanted fetal noradrenergic neurons from the brain stem locus coeruleus region can form synaptic contacts with host hippocampal neurons in the adult rat, both in intact animals and in animals where the intrinsic noradrenergic afferents to the hippocampus had been removed by 6-OHDA before transplantation. Previous studies, using the ALFA histofluorescence method, have shown that grafted neurons can reestablish fairly normal terminal innervation patterns in the adrenergically denervated hippocampus (3,4). In the present study we used immunohistochemical techniques with both TH and NA antisera. The results obtained with the two antisera were quite similar. The number of TH-immunostained and NA-immunostained neurons, as seen in alternating sections from the same specimens, were comparable, which suggests that the vast majority of the TH-immunostained neurons in the locus coeruleus grafts were noradrenergic. The principal difference between the results of the two antisera were that the TH antiserum stained
FORMATION
FIG. 7. NA-immunolabeled boutons unlabeled dendrites in the dentate hilus plant survival). Scale bar = 1 pm.
OF
SYNAPTIC
GRAFT-HOST
forming symmetric synaptic contacts of the host hippocampus. The animal
more powerfully than the NA antiserum, especially the neuronal soma and dendrites. The dendrite-like processes were thus observed more clearly with the TH antiserum. Also the axon-like fibers and boutons stained with the NA antiserum were generally less densely labeled than those stained with TH antiserum.
CONNECTIONS
(arrowheads) was denervated
263
with spines (SP, in a) and shafts (SH, in b) of before transplantation (8 months post-trans-
The synapses formed by the grafted locus coeruleus neurons were mostly of the symmetric type, contacting dendritic shafts or spines of immunonegative host neurons. Synapses onto cell bodies were not observed in the present material. The types of synapses formed appeared to be the same regardless of whether the hippo-
264
FIG. 8. with shafts noradrenergic
MURATA
TH-immunolabeled (a) and NA-immunolabeled (SH) of unlabeled dendrites in the stratum innervation was left intact until 2 weeks
ET
AL.
(b) boutons in a grafted rat forming synaptic contacts (arrowheads), respectively, radiatum (a) and the lacunosum moleculare layer (b) of the host CA3. The intrinsic before sacrifice (8 months post-transplant survival). Scale bar = 1 pm.
campus had been intact or adrenergically denervated at the time of transplantation. In the intact hippocampus, Frotscher and Leranth (12) and Milner and Bacon (18), using TH immunocytochemistry, have observed not only symmetric synapses on dendrites but also asymmetric contacts with spines as well as symmetric syn-
apses on cell bodies. The significance of these differences is unclear at this point. It seems possible that they could be explained either as an incomplete reinnervation of the denervated host neurons by the grafted neurons (i.e., a failure to reestablish certain types of contacts with cell bodies and spines) or the formation of
FORMATION
OF
SYNAPTIC
GRAFT-HOST
CONNECTIONS
265
F ‘IG. 9. A TH-immunolabeled bouton in a grafted rat forming a synapse (arrow) with a spine (SP) that also receives synapses (arrowhea ds) from a giant immunonegative mossy fiber terminal (MT) in the host dentate hilus. The symmetric or asymmetric nature of this synapse can not be cl early discerned in this case due to the dense presynaptic immunolabeling. The intrinsic NA innervation had been left until 2 weeks befc ore sacri fice (8 months post-transplant survival). Scale bar = 1 pm.
anomalous contacts, either onto abnormal targets or in abnormal positions with the correct target neurons. In the grafted animals which had an intact host noradrenergic input until 2 weeks before sacrifice it seems likely that the new synaptic input from the graft was formed as an addition to the intrisic innervation. In the present experiment the intrinsic noradrenergic innervation was removed before the perfusion to be able to positively identify the graft-derived TH-positive and NA-positive axons in the microscope. In a parallel study, now in progress, using the ALFA histofluorescence technique we have seen that locus coeruleus grafts implanted into intact, nondenervated targets form a varicose terminal network on top of the intrinsic one, thus creating an apparent hyperinnervation of the target area. In the hippocampus, this hyperinnervation has been estimated to amount to about 30-40% of the intrinsic innervation density. Also in cases where the transplantation has been made into an adrenergically denervated hippocampus, apparent hyperinnervation patterns are often formed in the areas close to the grafts. The ability of the grafted locus coeruleus neurons to innervate a target also in the presence of an intact host noradrenergic input is similar to what has previously been described for grafted serotonergic raphe neurons (1) and dopaminergic mesencephalic neurons (11,13). It is in contrast, however, to the behavior of grafts of more
highly specified neurons, such as retinal ganglion cells (16) and hippocampal granule cells and pyramidal neurons (24), which seem to be strictly dependent on the absence of the equivalent host innervation. This suggests that the growth of the monoaminergic neuron types may be less tightly regulated. As a consequence, however, grafted monoaminergic neurons would be more likely to exert functional effects also under conditions where the intrinsic homologous system is structurally intact (see, e.g., Ref. (13)). Electrophysiological studies in uivo (4) and in hippocampal slices in vitro (21) have provided evidence that the grafted locus coeruleus neurons form functional connections with the host hippocampus and that these connections, at least to some degree, are functionally similar to those of the normal noradrenergic afferent input. Although noradrenergic transmission in the brain may occur at both junctional and nonjunctional contacts (lo), it seems highly probable that at least part of the functional graft-host responses are mediated via the synaptic connections observed in the present study. The arrangement of the dendritic processes was an interesting feature of the grafted locus coeruleus neurons. In the TH-immunostained material, in particular, coarse tapering dendrites were seen to pass across the graft-host border deep into the host tissue. This arrangement, which has been observed also in grafts of
266
MURATA
ET
AL.
FIG. 10. Axo-dendritic associations of NA-immunolabeled terminals that form symmetric membrane specialization in the hilus of the dentate ! gyrus of a normal control animal. In a, a NA-immunolabeled terminal is seen to form a symmetric synapse (arrowhead) on a dendi :itic spine (I SP), which in addition receives a synapse from an unlabeled terminal (arrow). In b, the terminal is in contact (arrowhead) wit ;h a synapses from unlabeled terminals (arrow). Scale bar = 1 pm. dendrit ic shaft (SH), which also-receives
fetal dopamine neurons in the striatum (8,X), provides a possibility for reciprocal host-to-graft connections onto the transplanted noradrenergic neurons. Ultra-
structurally, these long projecting TH-positive dendrites were indeed seen to receive synapses of immunonegative boutons as they passed through the host neu-
FORMATION
OF
SYNAPTIC
ropil. Evidence for functional host afferent inputs to intrahippocampal grafted neurons has been obtained by Segal(20) from grafted raphe neurons in electrophysiological recordings from hippocampal slices in uitro. He observed excitatory postsynaptic responses in some graft neurons after stimulation in the host CA3 area. Such reciprocal graft-host connections may provide a morphological substrate for host afferent control of the activity of the ectopically placed grafts. Indeed, ongoing studies using the intracerebral microdialysis technique (Kalen et al., in preparation, and Bengzon et al., in preparation) indicate that the activity of grafted locus coeruleus neurons can be influenced from the host brain. ACKNOWLEDGMENTS We thank Gertrude Stridsberg for technical assistance. The study was supported by grants from the Swedish MRC (04X-3874 and 14X8666), the National Institutes of Health (NS-06701), and the Bank of Sweden Tricentary Fund.
GRAFT-HOST
AZMITIA, E. C., M. J. PERLOW, M. J. BRENNAN, AND J. M. LAUDER. 1981. Fetal raphe and hippocampal transplants into adult and aged C57BL/6N mice: A preliminary immunocytochemical study. Brain Res. Bull. 7: 703-710.
2.
BARRY, D. I., I. KIKVADZE, P. BRUNDIN, T. G. BOLWIG, A. BJ~RKLUND, AND 0. LINDVALL. 1987. Graftednoradrenergic neurons suppress seizure development in kindling-induced epilepsy. Proc. Natl. Acad. Sci. USA 84: 8712-8715.
267
10.
DESCARRIES, L., AND A. BEAUDET. 1983. The use of radioautography for investigating transmitter-specific neurons. In Handbook of Chemical Neuroanatomy (A. Bjiirklund and T. Hiikfelt, Eds.), Vol. 1, pp. 286-364. Elsevier, Amsterdam.
11.
DOUCET, G., P. BRUNDIN, S. SETH, Y. MUFZATA, R. E. STRECKER, L. C. TRIARHOU, B. GHETTI, AND A. BJ~~RKLUND. 1989. Degeneration and graft-induced restoration of dopamine innervation in the weaver mouse neostriatum: A quantitative radioautographic study of [3H]dopamine uptake. Exp. Brain Res. 77: 552-568.
12.
FROTSCHER, M., AND C. LEF~ANTH. 1988. Catecholaminergic innervation of pyramidal and GABAergic nonpyramidal neurons in the rat hippocampus. Histochemistry 88: 313-319.
13.
GAGE, F. H., S. B. DUNNETT, U. STENEVI, AND A. BJ~RKLUND. 1983. Aged rats: Recovery of motor impairments by intrastriatal nigral grafts. Science 22 1: 966-969.
14.
LIDBRINK, P., AND G. JONSSON. 1974. Noradrenaline nerve terminals in the cerebral cortex: Effect on noradrenaline uptake and storage following axonal lesion with 6-hydroxydopamine. J. Neurochem. 22: 617-626.
15.
MAHALIK, T. J., T. E. FINGER, I. STROMBERG, AND L. OLSON. 1985. Substantia nigra transplants into denervated striatum of the rat: Ultrastructure of graft and host interconnections. J. Comp. Neurol. 240: 60-70.
16.
MCLOON, S. C., AND R. D. LUND. 1980. Specific projections retina transplanted to rat brain. Exp. Brain Res. 40: 273-282.
17.
MCRAE-DEGUEURCE, A., AND M. GEFFARD. 1986. One perfusion mixture for immunocytochemical detection of noradrenaline, dopamine, serotonin and acetylcholine in the same rat brain. Brain Res. 376: 217-219.
18.
MILNER, T. A., AND C. E. BACON. 1989. Ultrastructural localization of tyrosine hydroxylase-like immunoreactivity in the rat hippocampal formation. J. Comp. Neurol. 281: 479-495.
REFERENCES 1.
CONNECTIONS
of
3.
BJBRKLUND, A., H. NORNES, AND F. H. GAGE. 1986. Cell suspension grafts of noradrenergic locus coeruleus neurons in rat hippocampus and spinal cord: Reinnervation and transmitter turnover. Neuroscience 18: 685-698.
19.
NORNES, H., A. BJ~~RKLUND, AND U. STENEVI. 1983. Reinnervation of the denervated adult spinal cord of rats by intraspinal transplant of embryonic brain stem neurons. Cell Tissue Res. 230: 15-35.
4.
BJ~RKLUND, A., M. SEGAL, AND U. STENEVI. 1979. Functional reinnervation of rat hippocampus by locus coeruleus implants. Brain Res. 1'70:409-426. BJBRKLUND, A., U. STENEVI, R. H. SCHMIDT, S. B. DUNNED, AND F. H. GAGE. 1983. Intracerebral grafting of neuronal cell suspensions. Acta Physiol. &and. (Suppl.) 522: l-75. BUCHANAN, J. T., AND H. 0. NORNES. 1986. Transplants of embryonic brainstem containing the locus coeruleus into spinal cord enhance the hindlimb flexion reflex in adult rats. Brain Res.
20.
SEGAL, M. 1987. Interactions between grafted serotonin neurons and adult host rat hippocampus. Proc. N. Y. Acad. Sci. 495: 284295.
21.
SEGAL, M., E. AZMITIA, A. BJBRKLUND, V. GREENBERGER, AND G. RICHTER-LEVIN. 1988. Physiology of graft-host interactions in the rat hippocampus. In Progress in Brain Research (D. M. Gash and J. R. Sladek, Jr., Eds.), Vol. 78, pp. 95-101. Elsevier, Amsterdam.
22.
SEMENOVA, T. P., E. A. GROMOVA, N. I. GFUSCHENKO, I. V. NESTEROVA, A. V. KULIKOV, G. N. SMIRNOVA, T. M. TRETYAK, A. G. BRAGIN, AND 0. S. VINOGRADOVA. 1987. Behavioural, biochemical and histochemical effects of locus coeruleus transplantation in rats with neurotoxic lesions of the catecholaminergic system. Neuroscience 22: 993-1002.
23.
YAKOVLEFF, A., A. ROBY-BRAMI, B. GUEZARD, H. MANSOUR, BUSSEL, AND A. PRIVAT. 1989. Locomotion in rats transplanted with noradrenergic neurons. Bruin Res. Bull. 22: 115-121.
24.
ZHOU, C.-F., G. RAISMAN, AND R. J. MORRIS. 1985. Specific patterns of fibre outgrowth from transplants to host mice hippocampi, shown immunohistochemically by the use of allelic forms of Thy-l. Neuroscience 16: 819-833.
5.
6.
381:225-236. 7.
BUZSAKI, G., G. PONOMAREFF, F. BAYARDO, T. SHAW, AND F. G. GAGE. 1988. Suppression and induction of epileptic activity by neuronal grafts. Proc. Natl. Acad. Sci. USA 85: 9327-9330.
8.
CLARKE, D. J., P. BRUNDIN, R. E. STRECKER, 0. G. NILSSON, A. BJ~RKLUND, AND 0. LINDVALL. 1988. Human fetal dopamine neurons grafted in a rat model of Parkinson’s disease: Ultrastructural evidence for synapse formation using tyrosine hydroxylase immunocytochemistry. Exp. Brain Res. 73: 115-126. COLLIER, T. J., D. M. GASH, AND J. R. SLADEK, JR. 1988. Transplantation of norepinephrine neurons into aged rats improves performance of a learned task. Brain Res. 448: 77-87.
9.
B.
NEUROLOGY
110,258-267
(1990)
Formation of Synaptic Graft-Host Connections by Noradrenergic Locus Coeruleus Neurons Transplanted into the Adult Rat Hippocampus Yuzo MURATA,*
TANEMICHI
*Department
CHIBA,**~ PATRIK BRUNDIN,~
of Anatomy, Saga Medical School, Nabeshima, of Medical Cell Research, Section of Neurobiology,
Press,
Saga 840-01, Japan; University of Lund,
MATERIALS
Inc.
INTRODUCTION Grafts of noradrenergic neurons, obtained from the pontine locus coeruleus region of rat fetuses, have been 1 Present address: Department Chiba University, Inohana, Chiba 0014-4686190 $3.00 Copyright 0 1990 by Academic Press, All rights of reproduction in any form
of Anatomy, School 280, Japan.
of Medicine,
258 Inc. reserved.
AND OLLE LINDvALLt
and TDepartment Sweden
used to correct for noradrenergic deficits in the brain or spinal cord in rats with a preceding neurotoxic lesion of the intrinsic noradrenergic system (2,3,4,6,19,22,23). Previous studies have shown that the grafted locus coeruleus neurons can provide a new noradrenergic terminal network in a previously denervated target area (3, 4,19), and there is evidence from several experimental models that the grafts can be functional in the normalization of noradrenaline-dependent behavioral deficits, such as spinal reflexes and locomotor activity in the lesioned spinal cord (6, 23), avoidance learning in aged rats (9), open field locomotion in 6-hydroxydopamine (B-OHDA)-lesioned rats (22), and development of epileptic phenomena in 6-OHDA lesioned (2) or fimbriafornix lesioned rats (7). The prominent axonal outgrowth observed from grafted noradrenergic neurons suggeststhat the graftinduced functional effects may depend on the establishment of functional graft-host connections. Although there is electrophysiological evidence that grafted locus coeruleus neurons can form functional contacts with the host (4, 21), these efferent projections have so far not been investigated ultrastructurally. Moreover, it is not known to what extent the ability of intracerebrally implanted noradrenergic neurons to reinnervate the host depends on whether or not the intrinsic noradrenergic innervation of the grafted area is removed prior to transplantation. In the present study we have used antisera against tyrosine hydroxylase (TH) and noradrenaline (NA) on the light and electron microscopic levels to investigate the growth and synaptic connectivity of noradrenergic locus coeruleus neurons, transplanted as a cell suspension into the hippocampal formation, in the presence or absence of an intrinsic noradrenergic afferent input.
Transplants of cell suspension obtained from the locus coeruleus region of 13- to 14-day-old rat fetuses were implanted into the hippocampal formation of intact adult rats or rats from which the noradrenergic afferents to the hippocampus had been removed by bilateral 6-hydroxydopamine (6-OHDA) injections into the dorsal tegmental noradrenergic bundle. The growth noradrenergic axons into the host hippocampus from the implant was studied at 4-8 months after surgery by immunohistochemistry using antisera raised against tyrosine hydroxylase or noradrenaline. In the animals with an intact noradrenergic system the host noradrenergic afferents were removed by bilateral dorsal bundle lesions 2 weeks before sacrifice. Fine axon-like fibers (diameter about 0.3 pm) and thick dendrite-like fibers (diameter about 1.3 pm), labeled immunohistochemitally, were abundant and spread far from the graft. By electron microscopy, immunolabeled axon-like fibers formed mostly symmetrical synaptic contacts with nonlabeled spines and shafts of dendrites in the host. Labeled dendrite-like fibers of presumed graft origin penetrated deep into the host neuropil and received abundant afferents from nonlabeled axon terminals. The extent of graft-derived noradrenergic axons and the synapses established with the host hippocampal neurons were similar in the chronically denervated animals and in the animals where the intrinsic noradrenergic afferents had been left intact until 2 weeks before sacrifice. The results show that implanted embryonic noradrenergic neurons are able to innervate the hippocampus in both the presence and the absence of an intact intrinsic noradrenergic innervation and that the ingrowing axons form abundant synaptic connections with the host hippocampal neurons under both conditions, Dendritic processes from the grafted noradrenergic neurons that extend deep into the host tissue may receive a reciprocal synaptic host afferent input. o 1990 Academic
ANDERS BJijRRLuND,t
AND METHODS
Surgery. The study was performed on young adult female Sprague-Dawley rats, X30-200 g at the time of surgery. There were two groups. The first group (n = 8) received injections of 8 pg 6-hydroxydopamine HC1(6OHDA) dissolved in 4 ~1ascorbate-saline (0.2 mg ascorbic acid/ml) bilaterally into the tegmental noradrener-
FORMATION
OF
SYNAPTIC
GRAFT-HOST
CONNECTIONS
259
= 9) had an intact noradrenergic system at the time of transplantation and was treated with 6-OHDA as above 2 weeks before perfusion (i.e., at least 4 months after transplantation) to eliminate the host noradrenergic innervation of the hippocampal formation. This time inwas terval between 6-OHDA injection and perfusion chosen to obtain a complete degeneration of the host locus coeruleus terminals (14). The completeness of the 6-OHDA lesions was confirmed by immunohistochemistry, demonstrating only single remaining terminals in the neocortex of the rats included in the study. The re-
C
FIG. 1. TH-immunolabeled (a) and NA-immunolabeled (b) cell bodies and outgrowing fibers of a locus coeruleus graft implanted into the CA3 area of hippocampal formation of a rat which had received a bilateral 6-OHDA lesion of dorsal NA bundle 2 weeks before transplantation (8 months post-transplant survival). (a, b) From adjacent sections. The graft occurred attached to the alveus (a) and thick dendrite-like fibers (arrows) and thin axon-like fibers (arrowheads) were seen to extend into the host (left) from the transplant (T). a, alveus; o, oriens layer; p, pyramidal cell layer. Scale bar = 100 pm. FIG. 2. TH-immunolabeled cell bodies and outgrowing fibers in a locus coeruleus graft (4 months survival) implanted into the hippocampal formation of a rat which received a bilateral 6-OHDA lesion of the dorsal NA bundle 2 weeks before sacrifice. A thick dendrite-like fiber (arrow) can be seen to run more than 500 pm into the host hippocampus. Scale bar = 100 Wm.
gic bundle. The coordinates were (tooth-bar = 0) A = +0.5 mm from the interaural line, L = -+l.l mm, and V = 6.5 mm down from dura. The injections were given 2 weeks before transplantation. The second group (n
FIG. 3. Distribution of NA-immunolabeled fibers in the normal hippocampus (C) and in hippocampi (D, I) reinnervated by locus coeruleus implants, 8 months survival. The specimen D was noradrenergically denervated with 6-OHDA prior to transplantation. The specimen I was denervated with 6-OHDA 2 weeks prior to sacrifice. The graft-derived fibers in both cases had their densest distribution in the CA3 and the dentate hilus. The grafts, delineated with a dashed line, in both cases were located in the CA3 area. Scale bar = 1 mm.
260
MURATA
ET
AL.
Immunohistochemistry. The rats for immunohistochemistry were processed for the visualization of TH and NA according to McRae-Degueurce and Geffard (17). In brief, the animals were perfused with 30 ml of saline, followed by 500 ml of 5% glutaraldehyde with 1% sodium metabisulfite (SMB) in 0.1 A4 cacodylate buffer, pH 7.0-7.5. The brains were cut into small pieces and immersed in Tris-SMB buffer (0.05 A4 Tris and, 1% SMB pH 7.5) with 30% sucrose. The tissues were frozen in liquid nitrogen for several seconds and thawed. Sections (40 pm) were cut on a Vibratome (Oxford Instruments) and collected in Tris-SMB buffer. The sections were incubated in 1% sodium borohydride for 30 min at room temperature and washed several times in TrisSMB. They were then incubated at 4°C for at least 15 h with one of the two primary antisera raised in rabbits: anti-TH (dilution, 1:2000; Eugene Tech International, Inc., Allendale, SC) and anti-NA (dilution, 1:3000; SFRI Lavoratoire, Martignal, France). The antibodies were diluted in a solution of Tris-SMB containing 1% nonspecific serum. The immunocytochemistry procedure was then continued with the PAP method. From the second antibody incubation, Tris-NaCl (0.05 M Tris and 0.9% NaCl, pH 7.6) was used as the diluting buffer and rinsing buffer. Following the DAB reaction, the sections were rinsed briefly in 0.15 M phosphate buffer (pH 7.4), postfixed with 1% 0~0, in 0.1 M phosphate buffer for 1 h, dehydrated in ethanol, and embedded in Spurr’s resin between sheets of transparent plastic film (ACLAR). After light microscopic observation and photography, thin sections were made, mounted on Formvar-coated grids with a single hole, stained with uranyl acetate and lead citrate, and then analyzed by electron microscopy. Sections stained with the omission of the primary antiserum were used as controls for the specificity of the observed immunoreactivity. FIG. 4. NA-immunolabeled fibers in the CA3 area from the three animals shown in Fig. 3. C, normal control rat; D, grafted rat denervated prior to transplantation; I, grafted rat denervated 2 weeks before sacrifice. o, oriens layer; p, pyramidal cell layer; r, stratum radiaturn. Scale bar = 100 pm.
gion containing the locus coeruleus was dissected from the rhombencephalon of 13- to 14-day-old rat fetuses (crown-rump length, CRL, 10-13.5 mm). A suspension was made according to the procedure of Bjijrklund et al. (5) without the trypsin incubation step. Two 1.5- to 2-/*l aliquots of cell suspension were injected unilaterally into the hippocampal formation in eight 6-OHDA-pretreated rats and nine intact rats at the following stereotaxic coordinates (in mm): A = +4.3 (rostra1 to interaural line), L = 3.0, and V = 3.0 below dura. The rats were perfused for immunohistochemistry at 3-8 months after transplantation. Normal intact rats were proceeded in parallel.
RESULTS Light Microscopy In the TH-immunostained sections (Figs. la and 2) cell bodies were labeled intensely in the grafts, and thick dendrite-like fibers, thin axon-like fibers, and terminals occurred in large numbers in the host hippocampus surrounding the implant. The NA antiserum (Fig. lb) stained cell bodies and dendrite-like processes more lightly, but axon-like fibers and terminals were observed with the same intensity, abundance, and distribution as those visualized with the TH antiserum. The number of graft cell bodies visualized with the TH and NA antisera in adjacent sections (Fig. 1) was similar with both procedures, which suggests that the vast majority of the TH-positive neurons in the grafts were noradrenergic.
FORMATION
OF
SYNAPTIC
GRAFT-HOST
CONNECTIONS
261
1FIG. 5. NA-immunolabeled fibers in the dentate hilus (h) from the three animals shown in Fig. 3. In both the normal control (C) and the fibers are most prominent in the zone below the granule cell layer (g). Scale bar = 100 pm. w Ifted rats (D, and I) the immunolabeled
Fiber outgrowth occurred throughout the dorsal hippocampal formation (Fig. 3). It was particularly dense in the dentate hilus and the CA3 area where they formed a fine varicose terminal network in close association with the neuronal elements by the host (Figs. 4 and 5). The magnitude and distribution of graft-derived THand NA-positive fibers in the host hippocampus appeared to be similar in the rats denervated prior to transplantation and in those denervated 2 weeks before the perfusion (cf. pictures D and I in Fig. 3). In both groups, thick dendrite-like fibers (diameter about 1.3 pm) were seen to extend from the grafts into the host hippocampus for distance up to about 500 pm (Fig. 2). The distribution of NA- and TH-immunoreactive fibers in the grafted hippocampi was not clearly different from that seen in the normal controls. The overall fiber density was lower than normal in the grafted specimens. However, in the areas close to the grafts, which tended to be more densely innervated, the graft-derived terminal network was in some cases clearly denser than normal. The number of surviving immunolabeled neuronal somata varied between 20 and 300 per implantation in the rats denervated prior to the transplantation, with an average of 90.0 + 90.2 (mean + SD, n = B), and between 10 and 70 per implantation site, with an average of 38.8
+ 19.0 (mean -t SD, rz = 9) in the rats denervated 2 weeks before perfusion. The differences between the groups, however, were not statistically significant (Welch t test and Wilcoxon rank-sum test). Electron Microscopy The fine structural analysis was focused on the CA3 region and the hilus of the dentate gyrus, i.e., the areas that had the densest TH- and NA-positive fiber networks in the light microscope. Immunostained axonal boutons as well as coarse dendrite-like structures were observed with both antisera within the host. The THpositive structures were sometimes so heavily labeled that they obscured the internal structures of the stained profiles, whereas the NA-positive boutons typically were more lightly stained (e.g., Fig. 7). Consistent with the light microscopic picture, dendrite-like structures occurred in the areas close to the grafts, whereas the labeled boutons occurred also at greater distances from the grafts. Immunoreactive dendrite-like processes that penetrated deep into the host neuropil were seen to receive abundant synaptic contacts by immunonegative axon terminals (Fig. 6). The labeled boutons were frequently seen to form synaptic contacts with both spines (Figs. 7a and 9) and
262
FIG. 6. TH-immunolabeled dendrites that graft. The picture is from the stratum radiatum transplant survival). Scale bar = 1 pm.
MURATA
ET
AL.
form synapses (arrowheads) with unlabeled vesicle-filled terminals at the periphery of the CA3 area from a grafted animal denervated before transplantation (5 months
shafts (Figs. 7b and 8) of nonlabeled dendrites in both CA3 and the dentate hilus. In the cases where both the pre- and the postsynaptic elements were discernible, most synapses appeared to be of the symmetric type, and only rarely were clear asymmetrical synapses observed. Some of the boutons were large and the majority were small (about 0.3 pm in diameter), thus corresponding well to the fine-caliber varicose terminals that formed the fine varicose terminal networks at the light microscopic level. These graft-derived synaptic boutons had the same general morphological appearance as the noradrenergic synaptic terminals observed in the intact control hippocampi (Fig. 10). In the dentate hilus, labeled boutons were seen to make synapses with spines that, in addition, received synapses from a giant immunonegative mossy fiber terminal (Fig. 9). Since pyramidal cell dendrites constitute the predominant postsynaptic target of the mossy fibers, this arrangement indicates that the grafted noradrenergic neurons had innervated host pyramidal neurons. Detailed comparison between the hippocampi denervated prior to transplantation and those denervated 2 weeks before the perfusion revealed no clear-cut differences. Synaptic contacts appeared to be equally abundant in both situations and the types of synaptic con-
of the post-
tacts formed in the CA3 and dentate hilus were similar under both conditions.
DISCUSSION The present results demonstrate that implanted fetal noradrenergic neurons from the brain stem locus coeruleus region can form synaptic contacts with host hippocampal neurons in the adult rat, both in intact animals and in animals where the intrinsic noradrenergic afferents to the hippocampus had been removed by 6-OHDA before transplantation. Previous studies, using the ALFA histofluorescence method, have shown that grafted neurons can reestablish fairly normal terminal innervation patterns in the adrenergically denervated hippocampus (3,4). In the present study we used immunohistochemical techniques with both TH and NA antisera. The results obtained with the two antisera were quite similar. The number of TH-immunostained and NA-immunostained neurons, as seen in alternating sections from the same specimens, were comparable, which suggests that the vast majority of the TH-immunostained neurons in the locus coeruleus grafts were noradrenergic. The principal difference between the results of the two antisera were that the TH antiserum stained
FORMATION
FIG. 7. NA-immunolabeled boutons unlabeled dendrites in the dentate hilus plant survival). Scale bar = 1 pm.
OF
SYNAPTIC
GRAFT-HOST
forming symmetric synaptic contacts of the host hippocampus. The animal
more powerfully than the NA antiserum, especially the neuronal soma and dendrites. The dendrite-like processes were thus observed more clearly with the TH antiserum. Also the axon-like fibers and boutons stained with the NA antiserum were generally less densely labeled than those stained with TH antiserum.
CONNECTIONS
(arrowheads) was denervated
263
with spines (SP, in a) and shafts (SH, in b) of before transplantation (8 months post-trans-
The synapses formed by the grafted locus coeruleus neurons were mostly of the symmetric type, contacting dendritic shafts or spines of immunonegative host neurons. Synapses onto cell bodies were not observed in the present material. The types of synapses formed appeared to be the same regardless of whether the hippo-
264
FIG. 8. with shafts noradrenergic
MURATA
TH-immunolabeled (a) and NA-immunolabeled (SH) of unlabeled dendrites in the stratum innervation was left intact until 2 weeks
ET
AL.
(b) boutons in a grafted rat forming synaptic contacts (arrowheads), respectively, radiatum (a) and the lacunosum moleculare layer (b) of the host CA3. The intrinsic before sacrifice (8 months post-transplant survival). Scale bar = 1 pm.
campus had been intact or adrenergically denervated at the time of transplantation. In the intact hippocampus, Frotscher and Leranth (12) and Milner and Bacon (18), using TH immunocytochemistry, have observed not only symmetric synapses on dendrites but also asymmetric contacts with spines as well as symmetric syn-
apses on cell bodies. The significance of these differences is unclear at this point. It seems possible that they could be explained either as an incomplete reinnervation of the denervated host neurons by the grafted neurons (i.e., a failure to reestablish certain types of contacts with cell bodies and spines) or the formation of
FORMATION
OF
SYNAPTIC
GRAFT-HOST
CONNECTIONS
265
F ‘IG. 9. A TH-immunolabeled bouton in a grafted rat forming a synapse (arrow) with a spine (SP) that also receives synapses (arrowhea ds) from a giant immunonegative mossy fiber terminal (MT) in the host dentate hilus. The symmetric or asymmetric nature of this synapse can not be cl early discerned in this case due to the dense presynaptic immunolabeling. The intrinsic NA innervation had been left until 2 weeks befc ore sacri fice (8 months post-transplant survival). Scale bar = 1 pm.
anomalous contacts, either onto abnormal targets or in abnormal positions with the correct target neurons. In the grafted animals which had an intact host noradrenergic input until 2 weeks before sacrifice it seems likely that the new synaptic input from the graft was formed as an addition to the intrisic innervation. In the present experiment the intrinsic noradrenergic innervation was removed before the perfusion to be able to positively identify the graft-derived TH-positive and NA-positive axons in the microscope. In a parallel study, now in progress, using the ALFA histofluorescence technique we have seen that locus coeruleus grafts implanted into intact, nondenervated targets form a varicose terminal network on top of the intrinsic one, thus creating an apparent hyperinnervation of the target area. In the hippocampus, this hyperinnervation has been estimated to amount to about 30-40% of the intrinsic innervation density. Also in cases where the transplantation has been made into an adrenergically denervated hippocampus, apparent hyperinnervation patterns are often formed in the areas close to the grafts. The ability of the grafted locus coeruleus neurons to innervate a target also in the presence of an intact host noradrenergic input is similar to what has previously been described for grafted serotonergic raphe neurons (1) and dopaminergic mesencephalic neurons (11,13). It is in contrast, however, to the behavior of grafts of more
highly specified neurons, such as retinal ganglion cells (16) and hippocampal granule cells and pyramidal neurons (24), which seem to be strictly dependent on the absence of the equivalent host innervation. This suggests that the growth of the monoaminergic neuron types may be less tightly regulated. As a consequence, however, grafted monoaminergic neurons would be more likely to exert functional effects also under conditions where the intrinsic homologous system is structurally intact (see, e.g., Ref. (13)). Electrophysiological studies in uivo (4) and in hippocampal slices in vitro (21) have provided evidence that the grafted locus coeruleus neurons form functional connections with the host hippocampus and that these connections, at least to some degree, are functionally similar to those of the normal noradrenergic afferent input. Although noradrenergic transmission in the brain may occur at both junctional and nonjunctional contacts (lo), it seems highly probable that at least part of the functional graft-host responses are mediated via the synaptic connections observed in the present study. The arrangement of the dendritic processes was an interesting feature of the grafted locus coeruleus neurons. In the TH-immunostained material, in particular, coarse tapering dendrites were seen to pass across the graft-host border deep into the host tissue. This arrangement, which has been observed also in grafts of
266
MURATA
ET
AL.
FIG. 10. Axo-dendritic associations of NA-immunolabeled terminals that form symmetric membrane specialization in the hilus of the dentate ! gyrus of a normal control animal. In a, a NA-immunolabeled terminal is seen to form a symmetric synapse (arrowhead) on a dendi :itic spine (I SP), which in addition receives a synapse from an unlabeled terminal (arrow). In b, the terminal is in contact (arrowhead) wit ;h a synapses from unlabeled terminals (arrow). Scale bar = 1 pm. dendrit ic shaft (SH), which also-receives
fetal dopamine neurons in the striatum (8,X), provides a possibility for reciprocal host-to-graft connections onto the transplanted noradrenergic neurons. Ultra-
structurally, these long projecting TH-positive dendrites were indeed seen to receive synapses of immunonegative boutons as they passed through the host neu-
FORMATION
OF
SYNAPTIC
ropil. Evidence for functional host afferent inputs to intrahippocampal grafted neurons has been obtained by Segal(20) from grafted raphe neurons in electrophysiological recordings from hippocampal slices in uitro. He observed excitatory postsynaptic responses in some graft neurons after stimulation in the host CA3 area. Such reciprocal graft-host connections may provide a morphological substrate for host afferent control of the activity of the ectopically placed grafts. Indeed, ongoing studies using the intracerebral microdialysis technique (Kalen et al., in preparation, and Bengzon et al., in preparation) indicate that the activity of grafted locus coeruleus neurons can be influenced from the host brain. ACKNOWLEDGMENTS We thank Gertrude Stridsberg for technical assistance. The study was supported by grants from the Swedish MRC (04X-3874 and 14X8666), the National Institutes of Health (NS-06701), and the Bank of Sweden Tricentary Fund.
GRAFT-HOST
AZMITIA, E. C., M. J. PERLOW, M. J. BRENNAN, AND J. M. LAUDER. 1981. Fetal raphe and hippocampal transplants into adult and aged C57BL/6N mice: A preliminary immunocytochemical study. Brain Res. Bull. 7: 703-710.
2.
BARRY, D. I., I. KIKVADZE, P. BRUNDIN, T. G. BOLWIG, A. BJ~RKLUND, AND 0. LINDVALL. 1987. Graftednoradrenergic neurons suppress seizure development in kindling-induced epilepsy. Proc. Natl. Acad. Sci. USA 84: 8712-8715.
267
10.
DESCARRIES, L., AND A. BEAUDET. 1983. The use of radioautography for investigating transmitter-specific neurons. In Handbook of Chemical Neuroanatomy (A. Bjiirklund and T. Hiikfelt, Eds.), Vol. 1, pp. 286-364. Elsevier, Amsterdam.
11.
DOUCET, G., P. BRUNDIN, S. SETH, Y. MUFZATA, R. E. STRECKER, L. C. TRIARHOU, B. GHETTI, AND A. BJ~~RKLUND. 1989. Degeneration and graft-induced restoration of dopamine innervation in the weaver mouse neostriatum: A quantitative radioautographic study of [3H]dopamine uptake. Exp. Brain Res. 77: 552-568.
12.
FROTSCHER, M., AND C. LEF~ANTH. 1988. Catecholaminergic innervation of pyramidal and GABAergic nonpyramidal neurons in the rat hippocampus. Histochemistry 88: 313-319.
13.
GAGE, F. H., S. B. DUNNETT, U. STENEVI, AND A. BJ~RKLUND. 1983. Aged rats: Recovery of motor impairments by intrastriatal nigral grafts. Science 22 1: 966-969.
14.
LIDBRINK, P., AND G. JONSSON. 1974. Noradrenaline nerve terminals in the cerebral cortex: Effect on noradrenaline uptake and storage following axonal lesion with 6-hydroxydopamine. J. Neurochem. 22: 617-626.
15.
MAHALIK, T. J., T. E. FINGER, I. STROMBERG, AND L. OLSON. 1985. Substantia nigra transplants into denervated striatum of the rat: Ultrastructure of graft and host interconnections. J. Comp. Neurol. 240: 60-70.
16.
MCLOON, S. C., AND R. D. LUND. 1980. Specific projections retina transplanted to rat brain. Exp. Brain Res. 40: 273-282.
17.
MCRAE-DEGUEURCE, A., AND M. GEFFARD. 1986. One perfusion mixture for immunocytochemical detection of noradrenaline, dopamine, serotonin and acetylcholine in the same rat brain. Brain Res. 376: 217-219.
18.
MILNER, T. A., AND C. E. BACON. 1989. Ultrastructural localization of tyrosine hydroxylase-like immunoreactivity in the rat hippocampal formation. J. Comp. Neurol. 281: 479-495.
REFERENCES 1.
CONNECTIONS
of
3.
BJBRKLUND, A., H. NORNES, AND F. H. GAGE. 1986. Cell suspension grafts of noradrenergic locus coeruleus neurons in rat hippocampus and spinal cord: Reinnervation and transmitter turnover. Neuroscience 18: 685-698.
19.
NORNES, H., A. BJ~~RKLUND, AND U. STENEVI. 1983. Reinnervation of the denervated adult spinal cord of rats by intraspinal transplant of embryonic brain stem neurons. Cell Tissue Res. 230: 15-35.
4.
BJ~RKLUND, A., M. SEGAL, AND U. STENEVI. 1979. Functional reinnervation of rat hippocampus by locus coeruleus implants. Brain Res. 1'70:409-426. BJBRKLUND, A., U. STENEVI, R. H. SCHMIDT, S. B. DUNNED, AND F. H. GAGE. 1983. Intracerebral grafting of neuronal cell suspensions. Acta Physiol. &and. (Suppl.) 522: l-75. BUCHANAN, J. T., AND H. 0. NORNES. 1986. Transplants of embryonic brainstem containing the locus coeruleus into spinal cord enhance the hindlimb flexion reflex in adult rats. Brain Res.
20.
SEGAL, M. 1987. Interactions between grafted serotonin neurons and adult host rat hippocampus. Proc. N. Y. Acad. Sci. 495: 284295.
21.
SEGAL, M., E. AZMITIA, A. BJBRKLUND, V. GREENBERGER, AND G. RICHTER-LEVIN. 1988. Physiology of graft-host interactions in the rat hippocampus. In Progress in Brain Research (D. M. Gash and J. R. Sladek, Jr., Eds.), Vol. 78, pp. 95-101. Elsevier, Amsterdam.
22.
SEMENOVA, T. P., E. A. GROMOVA, N. I. GFUSCHENKO, I. V. NESTEROVA, A. V. KULIKOV, G. N. SMIRNOVA, T. M. TRETYAK, A. G. BRAGIN, AND 0. S. VINOGRADOVA. 1987. Behavioural, biochemical and histochemical effects of locus coeruleus transplantation in rats with neurotoxic lesions of the catecholaminergic system. Neuroscience 22: 993-1002.
23.
YAKOVLEFF, A., A. ROBY-BRAMI, B. GUEZARD, H. MANSOUR, BUSSEL, AND A. PRIVAT. 1989. Locomotion in rats transplanted with noradrenergic neurons. Bruin Res. Bull. 22: 115-121.
24.
ZHOU, C.-F., G. RAISMAN, AND R. J. MORRIS. 1985. Specific patterns of fibre outgrowth from transplants to host mice hippocampi, shown immunohistochemically by the use of allelic forms of Thy-l. Neuroscience 16: 819-833.
5.
6.
381:225-236. 7.
BUZSAKI, G., G. PONOMAREFF, F. BAYARDO, T. SHAW, AND F. G. GAGE. 1988. Suppression and induction of epileptic activity by neuronal grafts. Proc. Natl. Acad. Sci. USA 85: 9327-9330.
8.
CLARKE, D. J., P. BRUNDIN, R. E. STRECKER, 0. G. NILSSON, A. BJ~RKLUND, AND 0. LINDVALL. 1988. Human fetal dopamine neurons grafted in a rat model of Parkinson’s disease: Ultrastructural evidence for synapse formation using tyrosine hydroxylase immunocytochemistry. Exp. Brain Res. 73: 115-126. COLLIER, T. J., D. M. GASH, AND J. R. SLADEK, JR. 1988. Transplantation of norepinephrine neurons into aged rats improves performance of a learned task. Brain Res. 448: 77-87.
9.
B.