benzodiazepine receptors in the substantia nigra of the rat

0306-4522/92 $5.00 + 0.00 Pergamon Press Ltd 1BRO

Neuroscience Vol. 50, No. 2, pp. 355 370, 1992 Printed in Great Britain

THE REGIONAL, CELLULAR A N D SUBCELLULAR LOCALIZATION OF GABAA/BENZODIAZEPINE RECEPTORS IN THE SUBSTANTIA NIGRA OF THE RAT L. F. B. NICHOLSON,* R. L. M. FAULL,*q" H. J. WALDVOGEL* a n d M. DRAGUNOW + *Department of Anatomy and ~cDepartment of Pharmacology, School of Medicine, University of Auckland, Private Bag, Auckland, New Zealand Abstract--The regional, cellular and subcellular distribution of GABAA/benzodiazepine receptors was investigated by light and electron microscopy in the rat substantia nigra. The regional distribution and density of GABAA/benzodiazepine receptor subtypes (Type I and II) was studied using quantitative receptor autoradiography following in vitro labelling of cryostat sections with tritiated ligands. This was followed by a detailed study of the cellular and subcellular distribution and localization of GABAA/benzodiazepine receptors by light and electron microscopy using immunohistochemical techniques with a monoclonal antibody (bd-17) to the [~2.~subunits of the GABAA/benzodiazepine receptor complex. Finally, in situ hybridization histochemistry using 35S-labelled oligonucleotide probes was used to demonstrate the cellular distribution of mRNA for the ~ and ~2 GABAA receptor subunits in the substantia nigra. The results of the autoradiographic and immunohistochemical studies showed a close correspondence in the regional distribution of GABAA/benzodiazepine receptors in the substantia nigra. A moderate-tohigh density of receptors was present throughout the full extent of the substantia nigra pars reticulata with a very low density of receptors in the substantia nigra pars compacta. Quantitative autoradiographic studies showed that: (i) the pars reticulata contained mainly central Type I receptors; (ii) the highest density of receptors was present in the caudal pars reticulata (200 + 38 fmol/mg) with successively lower densities of receptors in the middle (176 + 31 fmol/mg) and rostral (150 + 26 fmol/mg) levels of the pars reticulata; and (iii) the density of receptors in the pars reticulata was reduced by 34% following 6-hydroxydopamine-induced degeneration of dopaminergic pars compacts neurons. At the cellular level, GABAA/benzodiazepine receptor immunoreactivity was localized in a punctate fashion on dendrites and neuronal cell bodies in the pars reticulata. At the subcellular level, GABAA/benzodiazepine receptor immunoreactivity was associated with the pre- and postsynaptic membranes of axodendritic synaptic complexes along the length of small-to-large sized smooth dendrites in the pars reticulata. Two types of immunoreactive axodendritic synaptic complexes were identified: most (about 80%) immunopositive synapses showed equal staining of the pre- and postsynaptic membranes and were associated with small (< 1.01tm) axon terminals containing few mitochondria and small, round-to-pleomorphic vesicles in synaptic contact with small, peripheral dendrites; less frequently (about 20%) immunopositive synapses showed a marked immunoreactive thickening of the postsynaptic membrane and were associated with large (>1.0/~m) axon terminals containing numerous mitochondria and mainly pleomorphic vesicles in synaptic contact with large mainstem dendrites. Finally, the in situ hybridization studies showed that cells in the pars reticulata, but not in the pars compacta, expressed mRNA for the :q (but not for the :~,) subunit of the GABA A receptor complex. These results provide a detailed morphological characterization of the regional, cellular and subcellular distribution of GABAA/benzodiazepine receptors in the rat substantia nigra. These findings provide an anatomical basis for GABA-mediated synaptic inhibition in the substantia nigra.

The m a j o r inhibitory n e u r o t r a n s m i t t e r in the brain is G A B A which exerts m a n y of its effects t h r o u g h a GABAA/benzodiazepine receptor complex. 1°'25'39 Over recent years, in vitro a u t o r a d i o g r a p h i c receptor binding studies using tritiated benzodiazepine receptor ligands have s h o w n the regional distribution of G A B A A/benzodiazepine receptors in the m a m m a l i a n brain (for review, see Ref. 35). Also, with the recent

tTo whom correspondence should be addressed. Abbreviations: CL218,872, 3-methyl-6-[3-trifluoromethylphenyl]-l,2,4-triazolol-[4,3-b]pyridazine; 6-OHDA, 6hydroxydopamine; PBS, phosphate-buffered saline; SN, substantia nigra; SNc, substantia nigra pars compacta; SNr, substantia nigra pars reticulata. 355

d e v e l o p m e n t of m o n o c l o n a l antibodies to the G A B A A / b e n z o d i a z e p i n e r e c e p t o r c o m p l e x , 1(i.3~,50 light- a n d electron-microscopic i m m u n o h i s t o c h e m i cal studies have been u n d e r t a k e n to d e m o n s t r a t e the cellular a n d subcellular localization o f G A B A A / benzodiazepine receptors in the central nervous system.4,35,36.~9.43,44,51

In the m a m m a l i a n brain the substantia nigra (SN) contains especially high c o n c e n t r a t i o n s of G A B A : 29 in particular, the substantia nigra pars reticulata (SNr) receives m a j o r G A B A e r g i c projections from the striatum 2s and the globus pallidus, 4~ and provides G A B A e r g i c projections to the t h a l a m u s 5'46 and superior colliculus. 5't4'49 Also, receptor autoradiographic and i m m u n o h i s t o c h e m i c a l studies have

356

L . F . B . NICHOLSON et al.

s h o w n t h a t the SN contains high c o n c e n t r a t i o n s of G A B A ^ / b e n z o d i a z e p i n e receptors (for review, see Ref. 35). In the present study we have u n d e r t a k e n a detailed light- a n d electron-microscopic investigation o f the regional, cellular a n d subcellular distribution o f G A B A A/benzodiazepine receptors in the SN o f the rat. In these studies we have used tritiated ligands a n d q u a n t i t a t i v e receptor a u t o r a d i o g r a p h y to show the regional distribution of receptors, a GABA:,subunit specific m o n o c l o n a l a n t i b o d y a n d i m m u n o histochemistry to show the cellular a n d subcellular localization of receptors, and subunit-specific oligonucleotide probes a n d in situ hybridization to identify the cells expressing m R N A for G A B A A / benzodiazepine receptors. EXPERIMENTAL PROCEDURES

Tissue collection A total of 40 male albino Wistar rats weighing 250-300 g were used in this study on the distribution of GABAA/ benzodiazepine receptors in the SN: 25 for the autoradiographic demonstration of receptors, three for the immunohistochemioal localization of benzodiazepine receptors, and 12 for the localization of mRNA for GABA A receptor subunits using in situ hybridization. Autoradiographic localization qf GABA a/benzodiazepine receptors Rats were killed by decapitation and the brains removed and immediately frozen on dry-ice. The brains were sectioned at 16/zm on a cryostat in the coronal or sagittal plane and the sections thaw-mounted onto gelatine/chrome alum-coated slides. The slide-mounted sections were processed for the autoradiographic localization of the GABA A/benzodiazepine receptors as previously described. 8'51 Briefly, GABAA/benzodiazepine receptors were labelled by incubating the sections in 50 mM Tris-HC1 (pH 7.4) containing either 1 nM [3H]flunitrazepam (84 Ci/mmol, Amersham) or the benzodiazepine receptor antagonist [3H]Ro15-1788 (82.8Ci/mmol, New England Nuclear). Adjacent sections were incubated with [3H]flunitrazepam in the presence of CL218,872 (a ligand with high affinity for the Type I and low affinity for the Type II benzodiazepine receptors). The sections were washed (2 × 1 min in Tris-HCl buffer with a final dip in ice-cold distilled water) and dried under a stream of cold air. All procedures were carried out at 4°C. Nonspecific binding was determined by incubation of slides in the presence of 1/~M clonazepam. Once dry, the slides were brought to room temperature, taped into X-ray cassettes and apposed with 3H-sensitive Hyperfilm (Amersham) and exposed in the dark at 4°C for six to 12 weeks. The films were developed in D19, washed, fixed and dried. They were printed using standard photographic procedures to yield autoradiograms in which the autoradiographic labelled receptors appear as white dots on a dark background. After processing for autoradiography, selected sections were Nissl-stained to show the distribution of cell bodies. In order to investigate the proportion of receptors on dopaminergic neurons in the SN, seven rats were anaesthetized with sodium pentobarbitone (50 mg/kg, i.p.) and unilateral injections of 2/~1 of 6-hydroxydopamine (6OHDA; Sigma; 12#g of 6-OHDA in 2/zl of 0.9% NaCI + 0 . 2 % ascorbic acid) 6 were made into the right substantia nigra pars compacta (SNc) using stereotaxic procedures (coordinates: - 5.8 mm posterior, 1.7 mm lateral to the midline and 7 mm ventral to the pial surface). After

a survival period of 13 weeks the animals were kdled and processed for the autoradiographic localization of GABA A/benzodiazepine receptors as described above. After processing for autoradiography selected sections were also Nissl-stained with Cresyl Violet to determine the extent of cell loss in the pars compacta. Quantification ofautoradiograms. The optical densities of the autoradiographic labelling in the SN were determined directly from the autoradiograms using a spot densitometer described in our previous studies on benzodiazepine receptors in the human spinal cord. 7 Briefly, a microprojector was used to project an image of the autoradiogram (magnification × 12) onto a photo-sensitive diode such that the optical density of 100-gm diameter spots could be measured. The control labelling in the presence of 1/~M clonazepam was used to zero calibrate the photo-sensitive diode. Readings from the various regions of the SN were made by moving the autoradiogram in the microprojector. In each case, over 100 optical densitometric readings were taken from sections selected from six different levels of the SN. These optical density values were then converted to nCi tritium bound/mg protein using a standard curve derived from the optical density reading of the autoradiographic standards in which log nCi/mg was plotted against log optical density. These values were then converted to femtomoles of tritiated-ligand bound/mg tissue using the known specific activity of the labelled ligand. 4v All figures on the density of GABAA/benzodiazepine receptors labelled will therefore be given as femtomoles [3H]ligand per mg tissue which for convenience will be abbreviated to fmol/mg.

Immunohistochemical localization q/' GABA A/benzodiazepine receptors The rats were deeply anaesthetized with sodium pentobarbitone (50mg/kg, i.p.) and perfused intracardially with physiological saline followed by 500ml of fixative (4% paraformaldebyde, 0.1% glutaraldehyde in 0.t M phosphate buffer at pH 7.4). The brains were immediately removed, postfixed for 2-12h in the perfusate and the midbrain sectioned in the frontal plane on a Vibratome at 60-/zm sections. The sections were collected in phosphatebuffered saline (PBS) and processed for immunohistochemistry using a monoclonal antibody, bd-17, against the fl2,3 subunits of the GABAA/benzodiazepine receptor complex. The preparation and characterization of this antibody has been previously described) °'t2'3~3~39The localization of the antibody was visualized using standard immunohistochemical methods 45 as detailed below. Sections for light microscopy and electron microscopy were treated in the same way except that the sections for light microscopy had 0.1% Triton-X added to their washes while those for electron microscopy were first treated with 50% ethanol for 20 min to aid penetration of the antibody, All sections were then incubated for I h in 20% goat serum followed by incubation in the primary antibody (a hybridoma solution of bd-17) 39 for three days at room temperature (continuous agitation). The sections were then incubated in secondary antibody (biotinylated goat antimouse IgG, Sigma) at 1:400 overnight, then in ExtrAvidin" peroxidase conjugate (Sigma) at 1:I000 for 4 h. Following each of these steps the tissue was washed (3 x 10 rain) in PBS. The sections were then reacted for 15 min with 0.05% 3,3-diaminobenzidine tetrahydrochloride (Sigma) and 0.01% H202 (pH 7.4) in 0.1 M phosphate buffer, pH 7.4, to visualize the reaction product. Light-microscopic sections were washed in PBS, mounted on chrome alum slides, rinsed in distilled water and dehydrated through a graded alcohol series to xylene and coverslipped with DePeX (Serva). The reacted sections for electron microscopy were treated as follows: postfixed for 1 h in 1% osmium tetroxide in 0.1 M phosphate buffer, washed in buffer, dehydrated in a graded series of ethanols to 100% ethanol, washed in propylene oxide, impregnated with 812 resin (TAAB) overnight and

357

GABAA/benzodiazepine receptors in the substantia nigra then flat-embedded in fresh resin according to the method of Aides and Boone.t Blocks of tissue from the SNr were dissected out and glued onto blank blocks. Thin sections were cut on an ultramicrotome (LKB Ultratome IV), collected on single slot formvar coated grids, counterstained with Reynolds lead citrate and viewed in a Hitachi H7000 transmission electron microscope. Control sections for nonspecific labelling were also processed as above except that the primary antibody was substituted with 1% goat serum. As a further control for immunohistochemical processing artefacts, two rat brains were processed for electron microscopy as outlined above except that all immunohistochemical procedures were omitted. In situ hybridization histochernistry The distribution of GABA A receptor mRNAs for the ~l and ~2 subunits was investigated using in situ hybridization histochemistry and 35S-labelled oligodeoxyribonucleotide probes complementary to least conserved sequences in the cytoplasmic loop between the putative membrane-spanning domains M3 and M4. The oligonucleotide probes were kindly provided by Dr J. G. Richards (Hoffman La Roche, Basel); they were prepared on a DNA synthesizer by Med. prob A.S., Oslo. The a~-probe (54mer) was complementary to nucleotides 1144~1197 and the e2-probe (50mer) was complementary to nucleotides 1137 1186.32 Sense cq and ct2 35-base oligonucleotides, exactly complementary to the :q and ~2 antisense probes, were used as controls. The probes were 35S-labelled as described by Persohn et al. 32 The in situ hybridization method is detailed below. Rats were killed by decapitation; the brains removed and immediately frozen on dry-ice. Brains were then sectioned at 12 ~m in the coronal plane on a cryostat and mounted on slides previously coated with 2% 3-aminopropylthriethoxysilane solution in acetate. 34 Sections were stored at - 20 'C

Sections were processed for the localization of mRNA for the respective subunits of the GABAa receptor as described by Persohn et al? 2 Briefly, sections were brought to room temperature 1 h before hybridization then incubated in 50/~1 of hybridization buffer containing the 35S-labelled probe (3 x l0 s c.p.rn.). Sections were covered with Fujifilm" coverslips and incubated in a moist chamber at 3T'C for 24 h. Sections were then washed, dehydrated in ethanol and exposed (for up to 10 days) to sheet film (Hyperfilm, B-Max ", Amersham). The film was developed in Kodak PL12 then transfered to Kodak rapid fix. Following development, selected sections were counterstained with Cresyl Violet. RESULTS The principal aims of this study were to investigate the light-microscopic and electron-microscopic localization of GABAA/benzodiazepine receptors in the SN of the rat midbrain. First, the distribution of GABAA/benzodiazepine receptors was studied using autoradiography following in vitro labelling of cryostat sections with tritiated ligands. This was followed by a detailed study using light and electron microscopy to reveal the cellular and subcellular localization of the GABAA/benzodiazepine receptor using immunohistochemical techniques with a monoclonal antibody to the ~2.3 subunits of the receptor complex. Finally, in situ hybridization histocbemistry was used to identify the presence of the subunits of the GABAA/benzodiazepine receptor in the SN of the rat.

Autoradiographic localization diazepine receptors

of

G A B A A/benzo-

The localization of GABAA/benzodiazepine receptors in the rat SN was investigated using [3H]flunitrazepam (a benzodiazepine receptor agonist with a high affinity for Type I and II receptors) and [3H]RoI5-1788 (a central benzodiazepine receptor antagonist with a high affinity for Type I and II receptors). The details on the anatomy and density of receptors in the two major subdivisions of the SN, SNr and SNc, are shown in coronal sections of the SN in Fig. 1, and in parasagittal sections in Fig. 2. The autoradiograms in both figures show a very similar overall distribution of GABAA/ benzodiazepine receptors in the SN. The autoradiograms in Fig. 1 show a moderate-to-high density of receptors throughout the full extent of the SNr; by contrast the SNc shows a very low density of receptors. Comparison of the density of receptors throughout the rostrocaudal extent of the SN, shown in Fig. la, c and e, clearly demonstrates that the intensity of receptor labelling is greatest in the caudal region of the SNr (Fig. le) when compared with the middle (Fig. Ic) and rostral (Fig. la) regions. Quantitative analysis of the autoradiograms from coronal sections of the SN incubated with [3H]flunitrazepam confirms this differential regional distribution in the density of receptors in the SNr. The highest densities of GABAA/benzodiazepine receptors were found in the caudal regions of the SNr. Here the density of receptors in the caudal SNr (200 + 38 fmol/mg; Fig. le) was 19% higher than that at mid-levels of the SNr (176 ___31 fmol/mg; Fig. lc), and over 30% greater than the density of receptors in the rostral SNr (150 + 26 fmol/mg; Fig. la). These quantitative studies reveal that the overall density of receptors throughout the SNr averaged 179 + 38 fmol/mg. This regional difference in the density of GABAA/benzodiazepine receptors in the SN is especially obvious in parasagittal sections of the brain (Fig. 2). The autoradiograms in Fig. 2a, b clearly show a higher density of receptors in the caudal region of the SNr (see insets, Fig. 2a, b). Quantitative comparative analyses of the density of ligand binding in parasagittal sections of the SNr incubated with [3H]RolS-1788 (Fig. 2a) and [3H]flunitrazepam (Fig. 2b) demonstrated that the density of receptors in the caudal SNr ([3H]RoI5-1788 = 259 +_ 43 fmol/mg; [3H]flunitrazepam = 267 + 51 fmol/mg) was substantially higher than the density of receptors in the rostral SNr ([3H]RoI5-1788 = 163 +_ 21 fmol/mg; [3H]flunitrazepam = 161 + 31 fmol/mg). In order to determine the proportion of Type I and Type II GABAA/benzodiazepine receptors in the SNr, the density of labelling in nigral sections incubated with [3H]flunitrazepam alone (a ligand with a high affinity for both Type I and II receptors; Fig. 2b) was compared with the density of labelling in adja-

358

I.. I-:. B. NICHOLSOYet al.

cent sections incubated with [3H]flunitrazepam in the presence of CL218,872 (a discriminating ligand with a preferential high affinity for Type I receptors; ~5~7 Fig. 2c). A visual comparison of the autoradiograms (Fig. 2b, c) suggests that the SNr contains mainly Type I GABAA/benzodiazepine receptors. These visual impressions are confirmed and extended by detailed quantitative comparative analyses of the autoradiograms. These studies show that the density of binding in parasagittal sections of the SNr incubated with [3H]flunitrazepam alone averages 217 _+ 68 fmol/mg (Fig. 2b), whereas adjacent sections incubated with [3H]flunitrazepam in the presence of CL218,872 (Fig. 2c) show a density of binding of 16 _+ 2.5 fmol/mg. In order to investigate the proportion of GABAA/benzodiazepine receptors which were localized on the cell bodies and dendrites of dopaminergic neurons in the SN, quantitative autoradiographic studies were undertaken on the SN in animals which had received unilateral ventral midbrain injections of 6-OHDA. On the lesioned side in these cases, Nisslstained sections of the midbrain showed a virtual total loss of neurons in the SNc (Fig. 3a) and the autoradiograms demonstrated an almost complete loss of receptors in the pars compacta (Fig. 3b). In the SNc on the normal side there was a moderate-tolow density of receptors (133 _+ 30 fmol/mg; Fig. 3b) whereas on the lesioned side the density of receptor labelling in the SNc was extremely low (11 7 fmol/mg; Fig. 3b). A loss of receptors was also evident in the SNr on the lesioned side. Quantitative studies showed that in comparison to the density of receptors in the SNr on the normal side (294 + 55 fmol/mg), the density of receptors in the SNr on the lesioned side (195 ___27 fmol/mg) was reduced by 34% and this reduction was due entirely to a loss of Type I GABAA/benzodiazepine receptors. Immunohistochemical localization o f G A B A A / benzo diazepine receptors

The distribution and localization of GABAA/ benzodiazepine receptors at the light- and electronmicroscopic levels were studied using standard immunohistochemical procedures and a monoclonal antibody (bd-17) specific for the /~2,~ subunits of the GABAA/benzodiazepine receptor complex. Details of the immunohistochemical localization of receptors in the SN are shown in Fig. 4 at the light-microscopic level and in Fig. 5 at the electronmicroscopic level.

Light microscopy. When sections which had been reacted with monoclonal antibody bd-17 were viewed at low-magnification in the light microscope (Fig. 4a), they showed a pattern of receptor immunoreactivity which was remarkably similar to the autoradiographic pattern of GABAA/benzodiazepine receptor distribution (Fig. la, c, e). The immunohistochemicat sections showed a moderate-to-high density of' receptor immunoreactivity throughout the SNr which was interrupted by the fascicles of myelinated fibres traversing the rostrocaudal length of especially the medial two thirds of the SNr, as shown in Fig. 4a. In contrast, the SNc showed a very low level of immunoreactivity. When viewed at higher magnification with the light microscope, the dense immunoreactivity in the SNr was localized on what appeared to be immunoreactive dendrites and the outer membrane of medium to large neuronal somata (small arrows Fig. 4c, e). At higher magnification, receptor immunoreactivity was distributed in a discrete punctate fashion along the outer membrane of the dendritic processes in the SNr (Fig. 4d). Sections in which the primary antibody was omitted from the immunohistochemical procedures showed no immunolabelling (Fig. 4b). Electron microscopy. Ultrastructural studies on the immunohistochemical localization of GABAA/ benzodiazepine receptors in the SN were undertaken on tissue blocks selected from the pars reticulata, the region which consistently showed the most intense immunoreactivity (Fig. 4a) and which is known to receive major GABAergic projections. 2~'4~ The electron-microscopic distribution of GABAA/ benzodiazepine receptors in the SNr is shown in a series of electromicrographs in Fig. 5 (Fig. 5b, d, f). In order to demonstrate the ultrastructural distribution of GABAA/benzodiazepine receptors in the SNr, results which were obtained from sections processed for immunohistochemistry in the presence of the primary antibody, bd-17 (Fig. 5b, d, f), were compared with the results from sections processed in the absence of the primary antibody (Fig. 5a, c, e). Comparison of these results consistently showed that receptor immunoreactivity in the SN was characterized by an increased density of staining which was only associated with the membranes of axons and dendrites (Fig. 5b, d, f). In virtually all instances, the immunoperoxidase reaction end-product was associa~ted with synaptic complexes where it was localized on both the pre- and postsynaptic membranes; in sections cut normal to the synaptic membranes there

Fig. 1. A series of darkfield-type autoradiograms of the SN showing the distribution of GABAA/benzodiazepine receptors labelled with [3H]flunitrazepam from rostral (a), middle (c) and caudal (e) levels of the SN. For orientation purposes, the sections used to generate the autoradiograms in a, c and e have been counterstained for Nissl substance and fibres and are shown in b, d and f, respectively. The boundary between the SNr and the SNc is indicated by arrows. The autoradiograms show a moderate-to-high density of receptors throughout virtually the full extent of the SNr and a low density of receptors in the SNc. CC, crus cerebri; FR, fasiculus retroflexus; ML, medial lemniscus. Scale bar = I mm.

b

4

<

• ••?i :i~

360

L.F.B. N1CI-IOLSONet al.

was no detectable diffusion of reaction product into the synaptic cleft (Fig. 5b). Most of the immunopositive synapses showed equal immunoreactive staining of both the pre- and postsynaptic membranes (for example, boutons b~ in Fig. 5d). Less frequently, the immunoreactive staining on the synaptic membranes was unequal with the postsynaptic membrane showing a marked immunoreactive thickening on the cytoplasmic side of the membrane (see immunoreactive synapses identified as b2 in Fig. 5d, f). The extent and density of the immunopositive postsynaptic thickening is clearly evident when this type of immunoreactive synapse is compared with synapses in control sections; e.g., cf, the synapses in Fig. 5e, f. The immunopositive synapses were always associated with axodendritic contacts; no examples of immunoreactive axosomatic or axoaxonic synapses were found. The presynaptic axons associated with G A B A A/benzodiazepine receptor-immunopositive synapses were of two types. The majority (about 80%) were small (less than 1.0/~m in diameter) boutons (b~) containing few mitochondria and small, round-to-pleomorphic shaped vesicles (Fig. 5d). The second type of bouton (b2) was less numerous (about 20%); these boutons were characterized by their large size (greater than 1.0 # m in diameter) and contained numerous mitochondria and mainly pleomorphic vesicles (Fig. 5b, f). In the immunoreactive sections (Fig. 5b, d, f), the small-sized boutons (b~) were usually associated with immunopositive synapses showing equal staining of the pre- and postsynaptic membranes, whereas, the large-sized boutons (b2) were often associated with immunoreactive synapses showing a greater density of postsynaptic immunoreactive staining. These two bouton types (b 1 and b2) distinguished on the basis of bouton size were also recognized in control sections which were not processed for immunohistochemistry and those where the primary antibody was omitted from the procedure (Fig. 5a, c, e). In these control sections where the primary antibody was omitted, the synapses established by both types of bouton were symmetrical; i.e. they showed equal staining of the pre- and postsynaptic membranes. Indeed, in these control sections of the SNr, all synapses were of the symmetrical type. Type 1 boutons (bl) were small-sized (less than 1.0/~m in

diameter) containing few mitochondria and small, round-to-pleomorphic vesicles. They showed a widening of the synaptic cleft and the presence of equal pre- and postsynaptic thickenings (bj ; Fig. 5c); these boutons were the most common and accounted for 80% of the axon terminals seen. Type 2 boutons (b2) were large-sized (greater than 1.0/~m in diameter), contained numerous mitochondria and mainly pleomorphic vesicles (Fig. 5a, e), and showed preand postsynaptic thickenings of equal density (b2; Fig. 5a, e). In both the control and immunoreactive sections, the small-sized boutons (b~) established synaptic contact with small-diameter (less than 0.8/xm in diameter) dendrites (Fig. 5c, d). These synaptic boutons were invariably arranged in sequence, one next to the other, along the length (Fig. 5c, d) of smooth, spineless dendrites in the SNr. In contrast, the large-sized boutons (b2) in the control and immunoreactive sections established synaptic contact with large diameter (greater than 1.5 #m) dendrites (Fig, 5a, b). In situ hybridization studies The distribution of GABAA receptor mRNAs for the ~q and ~2 subunits was investigated using in situ hybridization and 35S-labelled antisense oligonucteotide probes. Transcripts of the ~ but not of the c~z subunit were detected in the SN. Figure 6 shows the distribution and localization of neurons expressing the ~l subunit in the SN: Fig. 6a is an autoradiogram showing the distribution of cells expressing ~ mRNA; and Fig. 6b is a Nissl stain of the section used to generate the autoradiogram in Fig. 6a and shows the boundary between the pars compacta and pars reticulata subdivisions of the SN. A careful comparison of Fig. 6a, b shows that virtually all of the neurons in the SNr express ~ subunit mRNA; in contrast, no cells in the pars compacta expressed ~ mRNA. DISCUSSION

This investigation demonstrates the regional, cellular and subeellular distribution and localization of GABAA/benzodiazepine receptors in the rat SN using autoradiographic and immunohistochemical

Fig. 2. A series of darkfield-type autoradiograms taken from serial parasagittal sections of the rat brain showing the distribution of GABA^/benzodiazepine receptors using: (a) [3H]Ro15-1788 (a central benzodiazepine receptor antagonist with a high affinity for Type I and II receptors); and [3H]flunitrazepam (FNZ) (a benzodiazepine receptor agonist with a high affinity for Type I and II receptors) in the absence (b) and presence (c) of CL218,872 (a ligand with a high affinity for Type I and a low atfinity for Type II receptors). The regions of the brain showing the SN are outlined by the boxes in a and b and these are illustrated at higher magnification in the insets. Comparison of the autoradiograms shows an identical distribution and similar intensity of labelling in the SNr in a and b, but an absence of labelling in the SNr in c, indicating that the SNr contains mainly central Type I GABAA/benzodiazepine receptors. In the rostral regions of the pars reticulata in a and b, note the lower density of receptor labelling and the presence of myelinated fibre bundles (see, e.g., double arrow in inset in b) as identified in adjacent myelin-stained sections. Scale bars = 1 mm.

362

I.. [: B. NICHOLSON el a/.

Fig. 3, The distribution of GABAA/benzodiazepine receptors using [~H]flunitrazepam (FNZ) in the SN of case DA7 which received a 6-OHDA lesion of the dopaminergic cells in the SNc on the right side of the midbrain 13 weeks prior to death. (a) Nissl stain of the section used in b; the asterisk in a indicates the marked neuron loss in the SNc. (b) Darkfield-type autoradiogram showing the distribution of GABAA/benzodiazepine receptors in the SN; the arrows indicate the boundary between the SNr and the SNc on both the left and right (lesioned) sides of the midbrain. In a, note the region of gliosis (large arrowhead) which marks the needle tract. Comparison of the pattern and density of labelling in the SN on the lesion and normal sides in b shows that on the side of the lesion there is a marked loss of receptors in the SNc and a moderate reduction in the density of receptors in the SNr. Scale bar = I ram,

Fig. 4. Light micrographs showing the immunohistochemical localization of GABAA/benzodiazepine receptors in the SN using the monoclonal antibody bd-17 specific for the fl2.3 subunits of the GABAA/benzodiazepine receptor complex. Panel a shows the receptors are densely localized in the SNr; the arrows indicate the boundary between the SNc and the SNr, and the region outlined by the box is shown at higher magnification in c. Panel c shows that receptors are found on both dendrites (large arrow) and neuronal cell bodies (small arrows). Panel d is a higher power photomicrograph of the region indicated by the large arrow in c and shows the punctate localization of dendritic receptors. Panel e shows the localization of receptor immunoreactivity on the surface of a neuronal cell body. Panel b is a control section where the primary antibody was omitted from the procedure. Scale bars = 1 mm (a, b): 50 FLm (c, e): 10pro (d).

T

.

t_

\

C~

364

L.F.B. N1CHOLSONet

methods at the light- and electron-microscopic levels. The autoradiographic results provide a quantitative assessment of the regional distribution of G A B A A/ benzodiazepine receptors and the immunohistochem-

al.

ical material has allowed qualitative studies to be undertaken at the light- and electron-microscopic levels. The results demonstrate that the highest concentrations of GABAA/benzodiazepine receptors are

Fig. 5a~:t

GABAA/benzodiazepine receptors in the substantia nigra

365

Fig:. 5e f Fig. 5. A series of electron micrographs from the SNr of the rat showing the morphology and axodendritic synaptic boutons from tissue which has been processed immunohistochemically in the absence (a, c, e) or presence (b, d, f) of the monoclonal antibody bd-17 to the fl2.3subunits of the GABA A/benzodiazepine receptor complex. Panels a and b show the appearance of large (b2) synaptic boutons (most are greater than 1.0 ~m in diameter), containing numerous mitochondria and mainly pleomorphic vesicles. These boutons are in synaptic contact with large (greater than 1.5 pm in diameter) dendrites (d), cut in transverse sections. Panels c and d show the appearance of small (less than 1.0/~m in diameter) synaptic boutons (b~) containing few mitochondria and small, round-to-pleomorphic vesicles in synaptic contact with small (less than 0.8/~m in diameter) dendrites (d) cut in longitudinal section. Panels e and f show the appearance of large (greater than 1.0 #m in diameter) synaptic boutons (b2) at high magnification. All identifiable axodendritic synaptic contacts (see arrows) in control sections showed equal pre- and postsynaptic membrane densities and can be classified as symmetrical in type (a, c, e). Sections processed in the presence of the primary antibody showed immunoreactivity localized almost exclusively on synaptic membranes where it was present on both the presynaptic and postsynaptic membranes (arrows, b, d, f). Most of the immunopositive synapses associated with the large boutons (b2) are characterized by the presence of an increased postsynaptic thickening (b, d, f). Scale bars = 0.5 l~m. found in the SNr, and that GABAA/benzodiazepine receptors are localized within morphologically distinct axodendritic synapses in the SN. In addition, in situ hybridization studies showed that cells in the pars reticulata expressed m R N A for the ~L subunit of" the G A B A A receptor complex. These findings are discussed together with the possible role of these receptors in basal ganglia function. Regional distribution ceptors

of GABAA/benzodiazepine re-

One of the major objectives of this study was to demonstrate the regional distribution and density of GABAA/benzodiazepine receptors in the SN using quantitative autoradiographic methods. These studies extend and complement other autoradiographic studies on the distribution of G A B A A /

benzodiazepine receptors in the rat brain which have been either general studies investigating the overall distribution of receptors in the brain 2~'~2"~5"55or studies primarily directed towards investigating receptor changes in the SN following lesions in the striaturn. 18'3~In the present study we have investigated the regional distribution, density and neuronal localization of GABAA/benzodiazepine receptor subtypes in the two major subdivisions of the SN. Our autoradiographic and immunohistochemical studies show that a moderate-to-high density of receptors was present throughout the full extent of the SNr with a very low density of receptors in the SNc (Figs 1, 2). Quantitative autoradiographic analysis showed that the highest density of receptors was present in the caudal pars reticulata with successively lower densities of receptors in the middle and rostral

366

L. t:. B. NI('HOLSONet al.

Fig. 6. (a) A darkfield-type autoradiogram showing the distribution of GABAAreceptor cq subunit mRNA in a coronal section of the SN in the rat as revealed by in situ hybridization using a 35S-labelled,antisensc 53mer oligonucleotide probe. Panel b is a Nissl stain of the section used for in situ hybridization in a~ The small arrows in a indicate cells showing cq hybridization signal in the SNr which correspond to the Nissl-stained SNr cells indicated by the small arrows in b. The large arrowheads in (a) and (b) indicate the boundary between the SNc and the SNr. In a, note the intense ~j signal in the SNr cells and the absence of ~-specific mRNA in the SNc. Scale bar = I mm.

levels of the pars reticulata. This heterogeneous distribution of receptors in the pars reticulata was especially obvious in parasagittal sections of the brain (Fig. 2a, b) and may reflect a regional variation in the density of termination of the GABAergic striatonigral and pallidonigral projections to the pars reticulata 28'41'42and/or a decreased density of receptors due to the presence of myelinated fibre fascicles which traverse the rostral regions of the SNr (Fig. 2b). Quantitative analysis of the receptor subtype distribution showed that the SNr contained mainly Type I GABAA/benzodiazepine receptors (93%) with minimal numbers of Type II receptors (7%). These findings are consistent with previous studies showing that the SN is enriched with Type I receptors. ~'~~6 It is interesting to note that the SN has a major involvement in the control of movement and that other subcortical structures such as the cerebellum and globus pallidus, which are also involved in the control of movement, are similarly enriched with Type I GABAA/benzodiazepine receptors. 8'9 These findings would tend to suggest that Type I receptors may be involved in the muscle relaxant and ataxic effects of benzodiazepines. However, the motor effects of benzodiazepines have been related to the Type II receptors while Type I receptors have been implicated in mediating the anti-anxiety effects of the benzodiazepines.26'56 Further studies are clearly needed to elucidate the functional significance and the role of Type I receptors in the SN. One of the most interesting findings in the present study is the demonstration that the density of Type

I GABAA/benzodiazepine receptors in the pars reticulata is reduced by 34% following 6-OHDAinduced degeneration of dopaminergic neurons in the SN (Fig. 3). These findings clearly demonstrate that approximately one third of the Type I GABAA/benzodiazepine receptors in the pars reticulata are localized on dopaminergic neurons. Since the majority of the cell bodies of the dopaminergic neurons in the SN are aggregated in the pars compacta lz and these neurons send long dendritic branches ventrally into the pars reticulata] 33~~-~it can be concluded that about one third of the Type l GABAA/benzodiazepine receptors in the pars reticulata are, in all likelihood, localized on the dendritic processes of dopaminergic neurons which extend to the pars reticulata. Further evidence for this conclusion is provided by the demonstration that GABAergic striatonigral fibres contact dopaminergic dendrites within the pars reticulata. 2 In conclusion, these findings suggest that approximately two-thirds of the GABAA/benzodiazepine receptors in the pars reticulata are involved with the modulation of nondopaminergic pars reticulata neurons; the remaining third are localized on dopaminergic dendrites and modulate the activity of pars compacta neurons which have a major projection to the striatum? ~6~7 Cellular and subcellular benzodiazepine receptors

localization

o f GABAA/

One of the major contributions of the present study is the demonstration of the cellular and subcellular distribution and localization of GABAA/

GABAA/benzodiazepine receptors in the substantia nigra benzodiazepine receptors in the SN by light and electron microscopy using immunohistochemical techniques and a monoclonal antibody (bd-17) to the /~2.3 subunits of the GABA A/benzodiazepine receptor complex. The results showed that at the cellular level, GABAA/benzodiazepine receptor immunoreactivity was localized in a punctate fashion on dendrites and neuronal cell bodies in the pars reticulata (Fig. 4). At the subcellular level, electron-microscopic studies showed that GABAA/benzodiazepine receptor immunoreactivity was associated with the preand postsynaptic membranes of axodendritic synaptic complexes along the length of small-to-large-sized smooth dendrites in the pars reticulata (Fig. 5b, d). The presence of immunoreactive material on the postsynaptic membrane is consistent with the postsynaptic localization of GABAA/benzodiazepine receptors. The significance of the presence of immunoreactivity on the presynaptic membrane has been discussed by previous investigators using monoclonal antibodies to localize G A B A A/benzodiazepine receptors in the central nervous system. 3-~'365~ It has been suggested that presynaptic immunoreactivity may be due to the diffusion or migration of reaction product from the post- to the presynaptic membranes or that it may indicate the presence of presynaptic benzodiazepine/GABA autoreceptors (see Ref. 51 for a detailed discussion). Two types of immunoreactive axodendritic synaptic complexes were identified in the immunoreactive sections. Most (about 80%) immunoposilive synapses showed equal staining of the pre- and postsynaptic membranes and were associated with small ( < 1.0 ~m) axon terminals containing few mitochondria with small, round-to-pleomorphic vesicles in synaptic contact with small, peripheral dendrites (Fig. 5d). Less frequently (about 20%) immunopositire synapses showed a marked immunoreactive deposit on the postsynaptic membrane and were associated with large ( > 1.0/tm) axon terminals containing numerous mitochondria and mainly pleomorphic vesicles in synaptic contact with large mainstem dendrites (Fig. 5b). Our studies of the morphological characteristics of axodendritic synaptic boutons in control sections processed in the absence of the primary antibody showed that, as in the immunoreactive sections, two types of axodendritic synaptic complexes could be identified. The morphology of the two types of synaptic boutons in the immunoreactive and control sections were identical except in the case of the large synaptic boutons (Fig. 5b, f). In the control sections these large boutons established symmetrical synaptic contact; in the immunoreactive sections the synaptic complexes associated with these large boutons were characterized by a marked immunoreactive deposit on the postsynaptic membrane of what is identifiable in control sections as a symmetrical synapse (cf. boutons in Fig. 5e, f). These findings show that there is a considerable difference in the antigenic properties of

367

the postsynaptic membrane associated with these two bouton types. The postsynaptic membrane of the synaptic complex associated with the large boutons shows enhanced antigenic properties for the bd-17 antibody compared with the postsynaptic membrane of the small boutons, These differences are sufficient to result in a considerable accumulation of immunoreactive material on the postsynaptic membrane of the symmetrical synaptic contact associated with these large boutons. These immunoreactive differences between the two bouton types may reflect fundamental differences in the GABAA/ benzodiazepine receptors associated with these two types of boutons. Our results clearly establish that GABA A/ benzodiazepine-immunoreactive synapses are associated with two morphologically distinct types of boutons in the SNr. In this respect it is interesting to compare the results of our studies with the recent detailed studies of Smith and Bolam 4~-4~on the pattern of termination of identified afferent fibres in the SN. In these excellent studies investigating the precise pattern of termination 4~ and convergence 4~ of pallidal and striatal projections on nigral neurons, Smith and Bolam showed that GABAergic pallidonigral fibres form large synaptic boutons (0.8-3.0 l~m in diameter, containing pleomorphic vesicles and large numbers of mitochondria) and establish symmetrical synaptic contact principally with large-diameter proximal dendrites and also on cell bodies, 4~42 whereas striatonigral terminals were generally small (0.3 ~1.8/~m in diameter, containing small round or pleomorphic vesicles with few mitochondria) and formed symmetric synapses predominantly with smalk distal dendrites. 42 These findings correlate very closely with our results and strongly suggest that the large- and smallsized boutons identified in our study, which are associated with GABA-immunoreactive synapses, represent the termination of pallidonigral and striatonigral fibres, respectively. This correlation is further strengthened by the results of a number of investigators that demonstrate that both of these projections utilize mainly GABA as their neurotransmitter (for references see Ref, 42). These findings therefore provide further collaborative evidence in support of our observations that the immunoreactivity associated with these synaptic complexes does indeed represent the distribution of GABA A receptors. As detailed above, the difference in the immunoreactivity that we have observed between the two immunoreactive bouton types in the SN may therefore reflect a morphological and functional difference in lhe receptors associated with the two bouton types and thus of the GABAergic input provided by the striatonigral and pallidonigral inputs to nigral neurons. Attention should however be drawn to problems associated with the interpretation of the localization of immunoreactivity at the ultrastructural level. For example, the light micrographs (Fig. 4c, e) show the

368

L. [:. B. NICHOLSON et al.

presence of immunoreactivity on cell somata, whereas at the electron microscopic level, no evidence of immunopositive axosomatic contacts was identified. These observations, together with the findings in previous studies showing the presence of GABAergic axosomatic contacts in the SNr, 4~'42 suggest that our electron microscopic immunohistochemical studies may not reveal the whole population of synapses containing GABAergic receptors. Alternatively, the GABAergic axosomatic synapses may not contain antigenic sites for the bd-17 monoclonal antibody, i.e. these synapses may not contain the fl2.3 subunits. Subunit composition o f G A B A A/benzodiazepine receptors in the substantia nigra

Molecular biological cloning studies on the constituent proteins of the GABAA/benzodiazepine receptors in the mammalian central nervous system have identified that the GABAA receptor complex is comprised of four different classes of subunits designated c~, fl, y and 6 (see Ref. 24 for references). It has now been established that there is a considerable heterogeneity of these various subunit proteins of the G A B A A receptor. Recent studies have shown the existence of at least six ct-subunits (cq 6), three flsubunits (/~1--3), tWO y-subunits (Yt,2) and one 3-subunit (for reviews, see Refs 23 and 30), and that there are regional differences in the specific subunit composition of GABAA receptors in various regions of the brain.23,s3,54

In our in situ hybridization studies using oligonucleotide probes for the ~ and e2 subunits, and immunohistochemical studies using a monoctonal antibody (bd-17) specific for the /~2,3 subunits, we have investigated the subunit composition of the G A B A A/benzodiazepine receptor complex in the SN of the rat. The results of these studies show that the cells in the SNr express m R N A for the ~j (Fig. 6) but not the ~2 subunit, and are immunoreactive for the fl2,3 subunits (Figs 4, 5) of the GABAA/ benzodiazepine receptor. These results are in agreement with previous in situ hybridization studies showing that cells in the pars reticulata express m R N A for the ~j and /~, subunits. 2°'23'37'4° Furthermore, cells in the pars reticulata have been shown to express m R N A for the Y2 subunitfl°'37 These collective findings suggest an ~1/3272 subunit configuration for G A B A A/benzodiazepine receptors in the pars reticulata of the SN of the rat. This particular receptor subtype has been studied in cultured cells and its pharmacological and electrophysiological properties are similar to endogenous receptors. 48 Since our studies have shown that over 90% of the G A B A A/benzodiazepine receptors in the pars reticulata are of the Type I variety (high affinity for triazolopyriadazines and /~ carbolineslS'17), then it would appear that the GABAA/benzodiazepine Type I binding site in the SNr in the rat may correspond with the ~ [322,2 receptor. This suggestion is consistent

with the results of previous tn situ hybridization studies 54 and pharmacological studies 19"33 showing that the cq/~xyz (where fix is any/~ subunit) receptor displays a pharmacology resembling that of the Type I site. Finally, as detailed and discussed above, it is especially interesting that a proportion (up to 30%) of the Type I GABAa/benzodiazepine receptors in the SNr are localized on dopaminergic neurons (presumably on the dendrites of dopaminergic neurons with cell bodies localized in the pars compacta), and that cells in the pars compacta do not express m R N A for either the cq (Fig. 6) or ~2 subunits. Indeed, previous studies have shown that pars compacta neurons express m R N A for the ~3 subunit.37 These findings suggest that the Type 1 GABAA/ benzodiazepine receptors localized on dopaminergic neurons may in fact have a different subunit composition to the Type I receptors localized on nondopaminergic pars reticulata neurons. Our results are consistent with recent molecular biological studies demonstrating a heterogeneity in the subunit composition of GABAA/benzodiazepine receptors in the mammalian brain (for review see Ref. 19). It is tempting to speculate that, from an ultrastructural point of view, this suggested heterogeneity in the subtype configuration of G A B A a receptors on dopaminergic and non-dopaminergic cells in the SN may be reflected in detectable morphological differences in GABAergic synapses on these two neuronal subtypes. This suggestion is particularly interesting in the light of our electron-microscopic immunohistochemical studies showing the presence of two morphologically distinct types of GABAA/benzodiazepine immunopositive axodendritic synaptic complexes in the pars reticulata of the rat. Our results immediately pose the question: are these two types of axodendritic synaptic complexes identified at the electronmicroscopic level localized on dopaminergic and non-dopaminergic neurons in the pars reticulata, respectively? Furthermore, if so, do these morphological differences represent GABAergic synapses containing receptors of different subunit composition? Unfortunately our study cannot resolve these intriguing questions. However, as discussed above, comparison of our results on the ultrastructural morphology of these two types of immunopositive boutons with Smith and B o l a m ' s 41"42findings suggests that the large boutons may be pallidonigral terminals and the small boutons striatonigral terminations. Our suggestions that these two boutons may be localized separately on dopaminergic and nondopaminergic neurons is not wholly supported by Smith and Bolam's 42 findings that some pallidonigral and striatonigral terminals converge onto the same nigrotectal pars reticulata neuron. Nevertheless, the possibility that ultrastructural differences in synaptic morphology may reflect subunit receptor differences

GABAA/benzodiazepine receptors in the substantia nigra raises the exciting possibility that subcellular synaptic m o r p h o l o g y identifiable with the electron microscope may reflect molecular biological characteristics o f n e u r o t r a n s m i t t e r receptors in the brain.

369

Acknowledgements--This study was supported by grants

from the Health Research Council of New Zealand, the New Zealand Neurological Foundation and the New Zealand Lottery Board.

REFERENCES

1. Aides L. D. and Boone T. B. (1984) A combined flat-embedding, HRP histochemical method for correlative light and electron microscopic study of single neurons. J. Neurosci. Res. 11, 27 34. 2. Bolam J. P. and Smith Y. (1990) The GABA and substance P input to dopaminergic neurons in the substantia nigra of the rat. Brain Res. 529, 57-78. 3. Carter D. A. and Fibiger H. C. (1977) Ascending projections of presumed dopamine-containing neurons in the ventral tegmentum of the rat as demonstrated by horseradish peroxidase. Neuroscience 2, 569-576. 4. De Bias A. L., Vitorica J. and Friedrich P. (1988) Localization of the GABA h receptor in the rat brain with a monoclonal antibody to the 57,000 M r peptide of the GABA A receptor/benzodiazepine receptor/C1- channel complex. J. Neurosci. 8, 602~14. 5. Di Chiara G., Porceddu M. L., Morelli M., Mulas M. L. and Gessa G. L. (1979) Evidence for a GABAergic projection from the substantia nigra to the ventromedial thalamus and to the superior colliculus of the rat. Brain Res. 176, 273 284. 6. Dragunow M., Leah J. D. and Faull R. L. M. (1991) Prolonged and selective induction of Fos-related antigen(s) in striatal neurons after 6-hydroxydopamine lesions of the rat substantia nigra pars compacta. Molec. Brain Res. 10, 355-358. 7. Faull R. L. M. and Villiger J. W. (1986) Benzodiazepine receptors in the human spinal cord: a detailed anatomical and pharmacological study. Neuroscience 17, 791 802. 8. Faull R. L. M. and Villiger J. W. (1988) Multiple benzodiazepine receptors in the human basal ganglia: a detailed pharmacological and anatomical study. Neuroscience 24, 433-451. 9, Faull R. L. M., Villiger J. W. and Holford N. H. G. (1987) Benzodiazepine receptors in the human cerebellar cortex: a quantitative autoradiographic and pharmacological study demonstrating the predominance of type I receptors. Brain Res. 411, 379-385. 10. H~iring P., Stfihli C., Schoch P., Tak~cs B., Staehelin T. and M6hler H. (1985) Monoclonal antibodies reveal structural homogeneity of 7-aminobutyric acid/benzodiazepine receptors in different brain areas. Proc. natn. Acad. Sci. U.S.A. 82, 4837-4841. 11. Hedreen J. C. and Chalmers J. P. (1972) Neuronal degeneration in rat brain induced by 6-hydroxydopamine; a histological and biochemical study. Brain Res. 47, 1 36. 12. Houser C. R., Olsen R. W., Richards J. G. and M6hler H. (1988) Immunohistochemical localization of benzodiazepine/GABAA receptors in the human hippocampal formation. J. Neurosci. 8, 1370-1383. 13. Juraska J. M., Wilson C. J. and Groves P. M. (1977) The substantia nigra of the rat: a Golgi study. J. eomp. Neurol. 172, 585~00. 14. Kilpatrick 1. C., Collingridge G. L. and Starr M. S. (1982) Evidence for the participation of nigrotectal l'aminobutyrate-containing neurons in striatal and nigral-derived circling in the rat. Neuroscience 7, 207 222. 15. Klepner C. A., Lippa A., Benson D, I., Sano M. C. and Beer B. (1979) Resolution of two biochemically and pharmacologically distinct benzodiazepine receptors. Pharmac. Biochem. Behav. 11, 457-462. 16. Kuypers H. G. J. M., Kievit J. and Groen-Klevant A. C. (1974) Retrograde axonal transport of horseradish peroxidase in rat's forebrain. Brain Res. 67, 211-218. 17. Lo M. M., Strittmatter S. M. and Snyder S. H. (1982) Physical separation and characterization of two types of benzodiazepine receptors. Proc. natn. Acad. Sei. U.S.A. 79, 680~584. 18. Lo M. M. S., Niehoff D. L., Kuhar M. J. and Snyder S. H. (1983) Differential localization of type I and type II benzodiazepine binding sites in substantia nigra. Nature 306, 57~i0. 19. Lfiddens H. and Wisden W. (1991) Function and pharmacology of multiple GABA n receptor subunits. Trends"pharmac. Sci. 12, 49 51. 20. Malherbe P., Sigel E., Baur R., Persohn E., Richards J. G. and M6hler H. (1990) Functional characteristics and sites of gene expression of the e 1,/3 l,?2-isoform of the rat GABA n receptor. J. Neurosci. 10, 2330-2337. 21. Mennini T. and Gobbi M. (1990) Regional distribution of low-affinity GABA receptors coupled to benzodiazepine receptor subtypes in rat brain: an autoradiographic evaluation. Eur. J. Pharmac. 189, 143-148. 22. Miyoshi R., Kito S. and Mizuno K. (1985) Autoradiographic studies on benzodiazepine receptor subtypes in the rat brain. Jap. J. Pharmac. 38, 281-285. 23. M6hler H., Malherbe P., Draguhn A. and Richards J. G. (1990) GABAn-receptors: structural requirements and sites of gene expression in mammalian brain. Neurochem. Res. 15, 199 207. 24. M6hler H., Malherbe J. G., Richards J. G., Persohn E., Benke D., Barth M., Rhymer T. and Sigel E. (1991) GABAA-receptor gene expression and regulation. In Neurotransmitter Regulation of Gene Transcription, Fidia Research Foundation Series Vol. 7 (ed. Costa E.), pp. 111-124. Thieme Publishers, New York. 25. M6hler H., Richards J. G. and Wu J.-Y. (1981) Autoradiographic localization of benzodiazepine receptors in immunocytochemically identified 7-aminobutyrergic synapses. Proc. nam. Acad. Sci. U.S.A. 78, 1935 1938. 26. Montaldo S., Serra M., Concas A., Corda M. G., Mele S. and Biggio G. (1984) Evidence for the presence of benzodiazepine receptor subclasses in different areas of the human brain. Neurosci. Lett. 52, 263 268. 27. Nauta H. J. W., Pritz M. B. and Lasek R. J. (1974) Afferents to the rat caudoputamen studied with horseradish peroxidase. An evaluation of a retrograde neuroanatomical research method. Brain Res. 67, 219-238. 28. Nitsch C. and Riesenberg R. (1988) Immunocytochemical demonstration of GABAergic synaptic connections in rat substantia nigra after different lesions of the striatonigral projection. Brain Res. 461, 127 142. 29. Okada Y. (1976) Role of GABA in the substantia nigra. In GABA in Nervous System Function (eds Roberts E., Chase T. N. and Tower D. B.), pp. 235-243. Raven Press, New York. 30. Olsen R. W. and Tobin A. J. (1990) Molecular biology of GABAA-receptors. Fedn Am. Soc. exp. Biol. J. 4, 1469-1480.

370

L . F . B . N1CHOLSON el al.

31. Pan H. S., Penney J. B. Jr and Young A. B. (1984) Characterization of benzodiazepine receptor changes in substantia nigra, globus pallidus and entopeduncular nucleus after striatal lesions. J. Pharmac. exp. Ther. 230, 768 775. 32. Persohn E., Malherbe P. and Richards J. G. (1991) In situ hybridization histochemistry reveals a diversity of GABA,~ receptor subunit mRNAs in neurons of the rat spinal cord and dorsal root ganglia. Neuroscience 42, 497-507. 33. Pritchett D. B., L/iddens H. and Seeburg P. H. (1989) Type I and Type 11 GABAA-benzodiazepine receptors produced in transfected cells. Science 245, 1389-1392. 34. Rentrop M., Knapp B., Winter H. and Schweizer J. (1986) Aminoalkylsilane-treated glass slides as support 10t~m situ hybridization of keratin cDNAs to frozen tissue sections under varying fixation and pretreatment conditions. Histochem. J. 18, 271 276. 35. Richards J. G., M6hler H. and Haefely W. (1986) Mapping benzodiazepine receptors in the CNS by radiohistochemistry and immunohistochemistry. In Neurohistochemistrv: Modern Methods and Applications (eds Panula P., Paivarinta H and Soinila S.), pp. 629 677. Alan R. Liss, New York. 36. Richards J. G., Schoch P., H/iring P., Takacs B. and M6hler H. (1987) Resolving GABAA/benzodiazepine receptors: cellular and subcellular localization in the C.N.S. with monoclonal antibodies. J. Neurosci. 7, 1866 1886. 37. Richards G., Schoch P. and Jenck F. (1991) Benzodiazepine receptors and their ligands. In 5-HTtA Agonists, 5-HT~ Antagonists and Benzodiazepine: Their Comparative Behavioural Pharmacology (eds Rodgers R. J. and Cooper S. J.l. pp. 1-30. John Wiley, New York. 38. Rinvik E. and Grofova 1. (1970) Observations on the fine structure of the substantia nigra in the cat. Expl Brain Res. II, 229 248. 39. Schoch P., Richards J. G., H/iring P., Takacs B., St/ihli C., Staehelin T., Haefely W. and M6hler H. (1985) Co-localization of GABA A receptors and benzodiazepine receptors in the brain shown by monoclonal antibodies. Nature 314, 168-171. 40. Sequier J. M., Richards J. G., Malherbe P., Price G. W., Matthews S. and M6hler H. (1988) Mapping of brain areas containing RNA homologous to cDNAs encoding the 2 and/~ subunits of the rat GABAA 7-aminobutyrate receptor. Proc. nam. Acad. Sci. U.S.A. 85, 7815 7819. 41. Smith Y. and Bolam J. P. (1989) Neurons of the substantia nigra reticulata receive a dense GABA-containing input from the globus pallidus in the rat. Brain Res. 493, 160-167. 42. Smith Y. and Bolam J. P. (1991) Convergence of synaptic inputs from the striatum and the globus paltidus onto identified nigrocollicular cells in the rat: a double anterograde labelling study. Neuroscience 44, 45 73. 43. Soltesz I., Roberts J. D. B., Takagi H., Richards J. G., M6hler H. and Somogyi P. (1990) Synaptic and nonsynaptic localization of benzodiazepine/GABAA receptor/C 1 channel complex using monoclonal antibodies in the dorsal lateral geniculate nucleus of the cat. Eur. J. Neurosci. 2, 414--429. 44. Somogyi P., Takagi H.. Richards J. G. and M6hler H. (1989) Subcellular localization of benzodiazepine/GABAa receptors in the cerebellum of rat, cat and monkey using monoclonal antibodies. J. Neurosci. 9, 2197 2209. 45. Sternberger L. A. (1974) Immunoeytochemisto,, pp. 129 171. Prentice-Hall, New Jersey. 46. Straughan D. W., MacLeod N. K., James T. A. and Kilpatrick I. C. (1980) GABA and the nigrothalamic pathway, Brain Res. Bull. 5, 7--11. 47. Unnerstall J. R., Niehoff D. L., Kuhar M. J. and Patacios J. M. (1982) Quantitative receptor autoradiography using [3H]Ultrofilm: application to multiple benzodiazepine receptors. J. Neurosci. Meth. 6, 5973. 48. Verdoorn T. A., Draguhn A., Ymer S., Seeburg P. H. and Sakmann B. (1990) Functional properties of recombinant rat GABA a receptors depend upon subunit composition. Neuron 4, 919-928. 49. Vincent S. R., Hattori T. and McGeer E. G. (1978) The nigrotectal projection: a biochemical and uttrastructural characterization. Brain Res. 151, 159-164. 50. Vitorica J., Park D., Chin G. and De Blas A. L. (1988) Monoclonal antibodies and conventional antisera to the G A B A a receptor/benzodiazepine receptor/C1 channel complex. Neuroscience 8, 615~522. 51. Waldvoget H. J., Faull R. L. M., Jansen K. L. R., Dragunow M., Richards J. G., M6hler H. and Streit P. (1990) GABA, GABA receptors and benzodiazepine receptors in the human spinal cord: an autoradiographic and immunohistochemical study at the light and electron microscopic levels. Neuroscience 39, 361-385. 52. Wassef M., Berod A. and Sotelo C. (1981) Dopaminergic dendrites in the pars reticulata of the rat substantia nigra and their striatal input. Combined immunocytochemical localization of tyrosine hydroxylase and anterograde degeneration. Neuroscience 6, 2125 2139. 53. Wisden W., Morris B. J., Darlison M. G., Hunt S. P. and Barnard E. A. (1988) Distinct GABA A receptor ~ subunit mRNAs show differential patterns of expression in bovine brain. Neuron 1, 937-947. 54. Wisden W., Morris B. J., Darlison M. G., Hunt S. P. and Barnard E. A. (1989) Localization of GABA A receptor subunit mRNAs in relation to receptor subtypes. Molec. Brain Res. 5, 305-310. 55. Young W. S. and Kuhar M. J. (1980) Radiohistochemical localization of benzodiazepine receptors in rat brain. J. Pharmac. exp. Ther. 212, 337 346. 56. Young W. S., Niehoff D. L., Kuhar M. J., Beer B. and Lippa A. S. (1981) Multiple benzodiazepine receptor localization by light microscopic radiohistochemistry. J. Pharmac. exp. Ther. 216, 425-430. (Accepted 29 April 1992)