GABAB receptor heterodimer-component localisation in human brain

Molecular Brain Research 77 (2000) 111–124 www.elsevier.com / locate / bres

Research report

GABA B receptor heterodimer-component localisation in human brain Andrew Billinton a , *, Antoinette O. Ige a , Alan Wise b , Julia H. White b , Graham H. Disney b , Fiona H. Marshall b , Henry J. Waldvogel c , Richard L.M. Faull c , Piers C. Emson a b

a Department of Neurobiology, Babraham Institute, Babraham, Cambridge CB2 4 AT, UK Receptor Systems, Molecular Pharmacology Unit, GlaxoWellcome, Medicines Research Centre, Gunnels Wood Road, Stevenage, Herts SG1 2 NY, UK c Department of Anatomy with Radiology, Faculty of Medicine and Health Sciences, University of Auckland, Private Bag 92019, Auckland, New Zealand

Accepted 15 February 2000

Abstract In recombinant cell lines, functional GABA B receptors are only formed by the heterodimerisation between two related G-protein coupled receptor proteins GABA B R1 (GBR1) and GABA B R2 (GBR2), whilst the individual GBR1 or GBR2 do not produce fully functional receptors. To determine whether the heterodimerisation occurs in vivo, novel polyclonal antibodies targeting the C termini of GBR1 and GBR2, were raised in different species, characterised, and used to determine the relative localisation of the reported heterodimer components in human brain tissue, using immunohistochemistry. The use of different species for the raising of the antisera allowed double immunofluorescent labelling of the receptors as an indication of GBR1 / GBR2 receptor co-localisation in human brain. The presence of both proteins is reported in cerebellum, hippocampus, cortex, thalamus and basal ganglia. Regions of the brainstem including pons and medulla, also express GBR1 and GBR2 protein. The double immunofluorescence demonstrated that GBR1 and GBR2 are co-localised in the human cerebellar cortex. Together these results suggest the widespread distribution of GABA B receptors in human brain, and that GABA B receptors GBR1 and GBR2 can exist in the same cell, and therefore may function as a heterodimer in the human brain.  2000 Elsevier Science B.V. All rights reserved. Themes: Neurotransmitters, modulators, transporters, and receptors Topics: GABA receptors Keywords: Antibody; Co-localisation; Confocal microscopy; GABA B receptors; Heterodimer; Immunohistochemistry

1. Introduction GABA B receptors which mediate slow inhibitory postsynaptic potentials (IPSPs), couple postsynaptically to inwardly rectifying potassium channels (K ir ), increasing 1 21 K flux, and presynaptically to calcium (Ca ) channels, 21 reducing inward Ca movement [3,39]. The first GABA B receptor GABA B R1 (GBR1), was recently cloned and shown to have similarity to metabotropic glutamate receptors [17]. However, recombinant expression of GBR1 failed to produce a functional GABA B receptor at the cell surface, which suggested that an accessory factor may be *Corresponding author. Tel.: 144-1223-496-000; fax: 144-1223-496022. E-mail address: [email protected] (A. Billinton)

required [9], such as a RAMP (receptor accessory membrane protein), as found in the case of the calcitonin receptor-like receptor (CRLR) [24]. Differing single transmembrane domain RAMPs conferred differential ligand specificity upon the CRLR receptor [24]. It is now known that to form functional GABA B receptors at the cell surface, GBR1 must heterodimerise with GABA B R2 (GBR2), a related G-protein coupled receptor discovered recently using expressed sequence tag (EST) homology based screening, and yeast two hybrid screening [14,18,20,30,42]. GBR2 has a similar molecular weight to GBR1 (¯120 kDa) and shares 35% sequence homology. Co-expression of GBR1 and GBR2 demonstrated cell membrane localisation of GABA B receptors, agonist affinity similar to native receptors and effective coupling to both adenylyl cyclase and K ir channels. Such a

0169-328X / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S0169-328X( 00 )00047-4

112

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

heterodimer represents a novel structure for a G protein coupled receptor, and introduces interesting possibilities for signal transduction. In this study, specific antisera raised to the C-termini of both GBR1 and GBR2 have been used to study the localisation of the components of the GABA B receptor heterodimer in human brain. Double fluorescence immunohistochemistry coupled with laser scanning confocal microscope imaging was used to investigate the co-localisation of GBR1 and GBR2 receptor proteins in neurologically normal human brain tissue in vitro. This work has, in part, been presented in abstract form [6].

2. Materials and methods

2.1. Antibody generation Antisera were raised to short synthetic peptides, specific for GBR1 (NH 2 –DGSRVHLLYK–COOH; residues 952– 961 of GBR1a) and GBR2 (NH 2 –VPPSFRVMVSGL– COOH; residues 930–941). Both sequences are located at the extreme end of the C-termini, respectively. The peptides were coupled to tuberculin protein derivative (TPD), and antisera raised by immunizing sheep (GBR1) or New Zealand white rabbits (GBR2) according to standard techniques [12]. All procedures were carried out under the appropriate UK Home Office licences.

2.2. Antibody purification GBR1 and GBR2 antisera used in the present study were purified from the crude antisera by immunoaffinity chromatography, as described previously [33], using the respective peptide coupled to sulfolink columns (Pierce).

2.3. Western blotting Immunoblot analyses were performed to validate the specificity of the two antisera for GBR1 and GBR2, as described previously [34]. Membranes were prepared from HEK293T cells transfected with GBR1a and GBR2, GBR1b and GBR2 or GBR2 alone, and from fresh frozen human frontal cortex, hippocampus and cerebellum by homogenisation of samples in ten volumes of ice cold buffer (50 mM HEPES, 1 mM EDTA, 1 mM EGTA, 1 mM PMSF, pH 7.4) and centrifugation at 2000 rpm for 30 min at 48C to remove cell debris and nuclei. Supernatants were collected and centrifuged for a further 10 min at 18 300 rpm. Resulting pellets were resuspended in buffer and centrifuged again to a total of three washes. Finally, the pellets were resuspended in buffer and protein content determined using bicinchoninic acid, according to the method of Smith et al. [37]. Extracted proteins were denatured with sodium dodecyl suphate (SDS) and b-mercaptoethanol, and equal amounts

loaded (5 mg) and separated by SDS–PAGE gel electrophoresis, then transferred to nitrocellulose membrane. These were blocked with 5% milk / Tris buffered saline (TBST) prior to a 1-h incubation with GBR1 or GBR2 antiserum in 5% milk / TBST. Following extensive washes, membranes were incubated with peroxidase labelled secondary antibodies for 1 h and extensively washed before ECL detection (Amersham).

2.4. Immunohistochemistry Samples of perfusion-fixed neurologically normal human brain tissue blocked into different regions were obtained from the New Zealand Neurological Foundation Brain Bank (University of Auckland). The clinical details of these patients are included in Table 1. Freezing sledge microtome sections were cut at 50 mm and collected free-floating in 0.1 M phosphate-buffered saline (PBS). Endogenous peroxidase activity was inactivated and sections pre-blocked with 2% (v / v) normal rabbit (GBR1) or goat (GBR2) serum in 0.1 M PBS, 0.3% Triton X-100 at room temperature for 60 min. After washing, the sections were incubated overnight at 48C in primary antiserum (diluted to 1:4000 (GBR1) and 1:800 (GBR2) in 0.1 M PBS, 0.3% Triton X-100). Following washing, sections were incubated with biotinylated rabbit anti-sheep (GBR1) or goat anti-rabbit (GBR2) secondary antibodies (Vector Laboratories) diluted 1:200 in 0.1 M PBS, 0.3% Triton X-100, and processed with preformed avidin-biotinylated horseradish peroxidase complex (Vector) for 45 min. Bound antibodies were detected with 0.5 mg / ml 3,39-diaminobenzidine tetrachloride (DAB) and 0.03% H 2 O 2 (Vector). Stained sections were mounted from distilled water onto charged microscope slides (Superfrost Plus, BDH), allowed to dry overnight then coverslips mounted with DePeX mounting medium (BDH). As controls for antibody specificity, serial sections were incubated with 0.1 M PBS, 0.3% Triton X-100 in place of primary antibody, with pre-immune serum, or with antiserum pre-absorbed with the appropriate peptide against which the antiserum was raised (overnight at 48C, 100 mg / ml) prior to use. Table 1 Summary of clinical details of human tissue cases used in the present study a Case no.

Age (years)

Gender

Cause of death

PMI (h)

4168 4265 4310 4236 4334 2932 3198

84 71 72 76 60 50 52

F M M F M F M

MI IHD MI ICM MI Pancreatitis IHD

14 15 10 18 12 4 15

a

F, female; h, hours; ICM, ischaemic cardiac myopathy; IHD, ischaemic heart disease; M, male; MI, myocardial infarction; PMI, post-mortem interval.

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

Images of whole sections were produced using an Agfa Snapscan 1236 scanner with transparency adapter and scanning directly from slides. Microscope images were captured using an Olympus BX50 microscope with a video camera attachment, and Olympus software.

2.5. Double immunofluorescence Tissue sections were incubated in 1% (v / v) normal serum in 0.1 M PBS, 0.3% Triton X-100 for 60 min at room temperature. Incubation was carried out with primary antisera sequentially overnight at 48C with detection carried out the following day. GBR1 antibody was detected using fluorescein isothionate (FITC)-labelled donkey anti-sheep IgG (Jackson ImmunoResearch) at 1:50, and GBR2 antiserum detected using Texas Red labelled goat anti-rabbit IgG (Vector Laboratories) at 1:50. Sections

113

were mounted from distilled water onto charged microscope slides (Superfrost Plus, BDH), dried and coverslipped in VectorShield (Vector) aqueous mountant. Fluorescence was visualised with a laser scanning confocal microscope system (Biorad MRC1024). Sections were viewed using a 603 oil immersion objective, and subsequent image processing carried out on a PC using Confocal Assistant software.

3. Results

3.1. Western blotting Immunoblotting was carried out using affinity purified antisera. Lysates from transfected HEK293T cells demonstrate a lack of cross-reactivity between the GBR1 antiserum and GBR2 protein (lane 1), and lanes 2 and 3 show bands at 130 and 100 kDa corresponding to the splice variants GBR1a and GBR1b, respectively (see Fig. 1). Bands higher up in lanes 2 and 3 may represent GBR1a / GBR2 and GBR1b / GBR2 heterodimers. Both GBR1a and GBR1b bands are visible in lysates from human brain regions (lanes 4–6). Lanes 7–12 show an equivalent blot probed with the GBR2 antiserum, indicating detection of GBR2 and not the GBR1 splice variants. Again, the higher molecular weight band seen in lanes 10–12 may represent heterodimers.

3.2. Immunohistochemistry Antisera specificity controls performed showed that specific staining was absent when the antibody was preabsorbed with the peptide against which it was raised (Fig. 2). Omission of the primary antibody and pre-im-

Fig. 1. Western blot analysis of GABA B transfected cells and human brain membrane. The upper blot was probed with GBR1 antiserum, and the lower blot with GBR2 antiserum. Proteins on the lower blot were run in the same order as the upper blot. Lanes 1–3 and 7–9 contained membranes from HEK293T cells transfected with GBR2 alone, GBR1a1 GBR2 or GBR1b1GBR2. Lanes 4–6 and 10–12 contained membranes from human cortex, cerebellum, and hippocampus. The GBR1 antiserum (upper blot) shows differential bands for GBR1a (|130 kDa) and GBR1b (|100 kDa) and bands for both molecular weights in the tissue membranes. The GBR2 alone transfected lane (1) is clear. The GBE2 antiserum (lower blot) shows bands of approximately 120 kDa across all lanes, with GBR2 appearing as a doublet in transfected cell membranes.

Fig. 2. Specificity controls for GBR1 and GBR2 antisera. Specific signal was removed in cerebellar sections by incubation of the antisera with the relevant peptide overnight at 48C prior to use. Sections shown are: (a) GBR1 antiserum at 1:4000, (b) GBR1 antiserum at 1:4000 preabsorbed with 100 mM GBR1 peptide, (c) GBR2 antiserum at 1:800, and (d) GBR2 antiserum at 1:800 preabsorbed with 100 mM GBR2 peptide. Scale bars represent 2 mm.

114

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

mune serum substitution of the primary antibody achieved similar results (data not shown).

GBR2 staining (Fig. 3e), though the white matter was not itself immunoreactive (Fig. 3a, b).

3.3. Visual cortex

3.4. Caudate nucleus, putamen and globus pallidus

GABA B receptor staining was consistent throughout the grey matter of the cortex, with no differences observed between the gyrus and sulcus (Fig. 3a, b). Immunoreactivity for both receptors was present throughout the cortical layers, with a strongly defined border between layers IV and V, which was particularly evident for GBR2 staining (Fig. 3c). Layers I and II of the visual cortex, contained neuropil staining for both GBR1 and GBR2. Some GBR1 immunoreactive pyramidal cells were evident in layer III, along with stellate cells in layer IV (Fig. 3d). Stellate cells in layer IV were also strongly GBR2 immunoreactive (Fig. 3e). Striations of fibres running through the pyramidal cell layers into the white matter were particularly visible with

GBR1 and GBR2 immunoreactivity were predominantly neuropil in the caudate nucleus and putamen (Fig. 4a, b), though the occasional neurone was visualised with GBR1 antiserum (Fig. 4). The caudate nucleus showed the strongest staining for GBR1 and GBR2 in this region. The globus pallidus showed cellular staining for GBR1 (Fig. 4d), though GBR2 demonstrated a more diffuse neuropil staining.

3.5. Thalamus GBR1 and GBR2 immunoreactivity was evident throughout thalamic regions (Fig. 5a, b). GBR1 staining

Fig. 3. GBR1 and GBR2 immunoreactivity in the visual cortex. (a, b) Show GBR1 and GBR2 immunoreactivity, respectively, throughout the grey matter. (c) Shows the layers in the human visual cortex. (d) Depicts GBR1 immunoreactivity in layer III / IV pyramidal (arrows) and stellate (arrowheads) cells, and (e) shows GBR2 immunoreactivity in layer III / IV stellate cells. Scale bars: (a, b) 5 mm, (c) 1 mm, (d, e) 50 mm. I–VI represent layers of the cortical grey matter; wm: white matter.

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

115

Fig. 4. GBR1 and GBR2 immunoreactivity in caudate nucleus and putamen. (a, b) Show GBR1 and GBR2 immunoreactivity, respectively, in caudate nucleus. (Cd), putamen (Pu), and external globus pallidus (EGP). Cell bodies are visibly stained by GBR1 in caudate nucleus (c) (arrows), globus pallidus (d) (arrows) and more rarely in the putamen (e), where processes were also visible. ic, internal capsule. Scale bars: (a, b) 5 mm, (c, e) 25 mm, (d) 50 mm.

was intense over cell bodies, and particularly in the reticular, ventral and subthalamic nuclei (Fig. 5e, g, i). The neuropil was also darkly stained in the reticular nucleus (Fig. 5c). Cellular staining was much weaker for GBR2, with a general homogeneous neuropil stain predominating, though faintly stained neurones were visible in the reticular, ventral and subthalamic nuclei (Fig. 5f, h, j). The mediodorsal thalamic nucleus showed a darkly stained neuropil that was sparsely populated with GBR1 immunoreactive cells, with the occasional cell being faintly stained with the GBR2 antiserum (Fig. 5c, d). The internal capsule was devoid of immunoreactivity for either receptor (Fig. 5a, b).

3.6. Hippocampal formation In the hippocampal formation, GBR1 and GBR2 immunoreactivities were widely distributed (Fig. 6a, b).

Pyramidal cells bodies were distinctly stained for both GBR1 and GBR2 in CA2 and CA3, where they are more densely packed (Fig. 6c, d). CA1 pyramidal cells were less intensely stained for GBR1, and were barely visible when stained for GBR2. Stratum radiatum contained a few GBR1 stained interneurone cell bodies and weakly stained neuropil, stratum lacunosum moleculare more intense neuropil and sparse GBR1 stained interneurones. Stratum radiatum was less well defined for GBR2 immunoreactivity. The dentate molecular layer showed moderate neuropil staining for GBR1 and GBR2, and punctate staining over the granule cell layer for GBR1, which was weaker for GBR2 (Fig. 6e, f). The hilar region showed larger stained cell bodies for GBR1, which were not clearly visible for GBR2 (Fig. 6e, f). Pyramidal cell bodies were visible in the subiculum and entorhinal cortex for both receptors (Fig. 6g, h). The occasional large interneurone in stratum oriens was clearly stained for GBR1 (Fig. 6i), but GBR2 immunoreactivity was not apparent in the same cells.

116

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

Fig. 5. GBR1 and GBR2 immunoreactivity in thalamic regions. (a, b) Show GBR1 and GBR2 immunoreactivity, respectively, in mediodorsal (Md), reticular (Rt), subthalamic nuclei (STN) and ventral (V) regions of the human thalamus. (c, e, g, i) Represent GBR1 immunoreactive cell bodies and neuropil, and (d, f, h,j) represent GBR1 immunoreactive cell bodies and neuropil found in MD, Rt, STN and V thalamic regions, respectively. Scale bars: (a, b) 5 mm, (c–h) 25 mm.

3.7. Substantia nigra and red nucleus Substantia nigra pars compacta displayed highly visible melanin pigment-containing cells, which made it difficult to visualise specific staining, though the GBR1 antiserum did stain some of the neuropil (Fig. 7a, c, d). The pars

reticulata contained lighter non-melanin cells which could be distinguished as being GBR1 positive (Fig. 7e), though GBR2 staining was apparently absent from these cells. The red nucleus showed weak neuropil immunoreactivity and labelled cell bodies for GBR1 (Fig. 7f), which was much weaker for GBR2.

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

117

Fig. 6. GBR1 and GBR2 immunoreactivity in the human hippocampus. (a, b) Demonstrate the distribution of GBE1 and GBR2 immunoreactivity throughout the human hippocampus. Micrographs (c) and (d) show CA2 / 3 subfield pyramidal neurones (arrows), and (e) and (f) show the interface between the dentate molecular layer (mol), granule cell layer (gr) and dentate hilus (hil), for GBR1 and GBR2 immunoreactivity, respectively. (g, h) Show entorhinal cortex pyramidal neurones (arrows), and (i) shows a GBR1 stained stratum oriens interneurone and processes (arrows). DG, dentate gyrus; Ent, entorhinal cortex; Sub, subiculum; slm, stratum lacunosum–moleculare; so, stratum oriens; sp, stratum pyramidale; sr, stratum radiatum. Scale bars: (a, b) 5 mm, (c, d, g, h) 20 mm, (i) 25 m, (e, f) 50 mm.

3.8. Cerebellum GBR1 immunoreactivity was present in the granule cell, Purkinje cell and molecular layers of the cerebellar cortex (Fig. 8a, c). The granule cell layer showed patchy staining

for GBR1 immunoreactivity, whereas GBR2 staining was very weak in this layer (Fig. 8b, c, d). The molecular layer demonstrated a more uniform distribution of immunoreactivity, and included small, defined patches of intense staining scattered throughout the layer for both receptors.

118

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

Fig. 7. GBR1 and GBR2 immunoreactivity in human substantia nigra and red nucleus. (a, b) Show GBR1 and GBR2 immunoreactivity in human substantia nigra pars compacta (SNC), substantia nigra pars reticulate (SNR), and red nucleus (R). (c, d) Clearly show neuromelanin-containing cells, and some neuropil GBR1 immunoreactivity in (c). Faintly stained cells were also visible with GBR1 antiserum in SNR and R (e, f), which were not apparent with GBR2 antiserum. Scale bars: (a, b) 5 mm, (c–e) 25 mm.

Purkinje cells were individually stained for GBR1, and intensely for GBR2 with dendritic shafts and proximal processes visible (Fig. 8c, d). In the deep cerebellar nuclei, neurones were intensely stained for GBR1, as was the neuropil around them; though GBR2 immunoreactivity was predominantly neuropil (Fig. 8e, f).

3.9. Pons and medulla Strong cellular staining for GBR1 was seen throughout the grey matter of the pons, particularly in the pontine nuclei (Fig. 9a, g), whereas GBR2 produced predominantly neuropil staining (Fig. 9b, h). The locus coeruleus was intensely stained for both receptors, though some of this staining was due to endogenous melanin (Fig. 9c, d). Cells of the Raphe nucleus were strongly stained for GBR1, but only very weakly stained for GBR2 (Fig. 9e, f). Fig. 10a,b shows a section through the upper part of the medulla oblongata. The inferior olive and accessory nuclei are clearly stained and at the light microscopic level, cell bodies are more visible for GBR1 in the strongly stained neuropil than for GBR2 (Fig. 10e, f). Also, intense staining

for both GBR1 (cellular and neuropil) and GBR2 (weaker cellular and neuropil) was present in the dorsal cochlear nucleus, inferior and medial vestibular nuclei and hypoglossal nucleus (Fig. 10a–d). GBR2 staining was particularly intense in the medullary reticular formation (Fig. 10b). The arcuate nucleus presented weak cellular staining for GBR2 in the moderately immunoreactive neuropil (Fig. 10h). This immunoreactivity was more pronounced for GBR1, with cell bodies being clearly visible (Fig. 10g).

3.10. Double immunofluorescence Images taken at the microscopic level using a 603 objective of a human cerebellar section double labelled for GBR1 (green) and GBR2 (red) is shown in Fig. 11a and b, respectively. Weaker fluorescence can be seen for both GBR1 and GBR2 in the molecular and granule cell layers, and a Purkinje cell (arrow) with associated process (arrowhead) is clearly labelled by both antibodies. When the images are merged using the Confocal Assistant software, a yellow colour results from co-localisation of GBR1 and GBR2, particularly over the Purkinje cell, but also over

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

119

Fig. 8. GBR1 and GBR2 immunoreactivity in the cerebellum. (a, b) Show GBR1 and GBR2 immunoreactivity, respectively, in cerebellar cortex and dentate nucleus (Dt). Higher magnification images (c) and (d) revealed the staining of Purkinje cells (arrows) as well as molecular (Mol) and granule cell (Gr) layers for GBR1 and GBR2 immunoreactivity, respectively. (e, f) Show strong cellular immunoreactivity for GBR1 in the dentate nucleus and predominantly neuropil staining for GBR2. Scale bars: (a, b) 5 mm, (c–f) 50 mm.

molecular and granule cell layers (Fig. 11c). Intense spots of immunofluorescence are presumed to be autofluorescent lipofuscin granules, which accumulate along the granule cell–molecular layer border.

4. Discussion This study demonstrates the specific immunodetection of GABA B receptors GBR1 and GBR2 in regions of perfused fixed, neurologically normal human brain. Western blotting demonstrated the specificity of the antisera raised for human GABA B receptors and showed a lack of cross reactivity of the antisera between GBR1 and GBR2. Immunohistochemistry controls demonstrated specificity of

the antisera for the peptide sequences to which the antisera were raised. The existence of mRNA for GBR1 and GBR2 in the human brain regions examined in this study has been reported [27]. In situ hybridisation studies in rat brain have shown GBR1, but not GBR2 mRNA to be abundant in the striatum [14,18,20,30], leading to the suggestion that GBR2 is not expressed in this region. However, the present study finds GBR1 and GBR2 immunostaining, predominantly in the neuropil in the equivalent human brain region. This can easily be reconciled by the fact that the mRNA may be present in cells extrinsic to the caudate, and the protein transported to axon terminals and dendrites within the caudate itself. This view is consistent with findings from studies on rat striatal GABA B receptors,

120

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

Fig. 9. Immunohistochemical localisation of GBR1 and GBR2 in human pons. Neuromelanin-containing cells are clearly visible in the locus coeruleus (LC) with both GBR1 (c) and GBR2 (d) antisera. GBR1 darkly stained cell bodies in the raphe (Ra) (e), but these were barely visible when stained with GBE2 antiserum (arrows) (f). The pontine nuclei possessed smaller, more rounded GBR1 immunoreactive cell bodies (arrows) (g), which were again barely visible when stained with GBR2 antiserum (arrows). Scale bars: (a, b) 5 mm, (c–h) 50 mm.

which suggest that the majority of these receptors are located on afferent terminals. Lesion studies in rodents have demonstrated that few GABA B receptors are intrinsic to the caudate, as lesions of cortical and nigral inputs reduce GABA B receptor binding in the rat striatum, whereas direct striatal lesions had no significant effect [19,20]. Electrophysiological investigations have indicated a lack of functional postsynaptic GABA B receptors on striatal output neurones, which further supports this suggestion [35,36]. However, several recent studies suggest the presence of GABA B receptors on subpopulations of striatal interneurones [13,43].

Strong thalamic expression of GBR1 is consistent with rat in situ hybridisation and immunohistochemical data [7,23,25]. The strongly GBR1 stained cells and neuropil of the reticular thalamic nuclei was in contrast to that shown in the rat [25], and may reflect differences between the antisera raised. Thalamic regions receive afferents from the cortex, which may account for the level of neuropil staining in this region. The role of the thalamus and GABA B receptors in absence seizures is widely recognised [10,38] and the presence of GABA B receptors in human thalamus would be consistent with their possible involvement in human absence seizures.

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

121

Fig. 10. Localisation of GBR1 and GBR2 immunoreactivity in human medulla (a, b). The GBR1 antiserum stained the neuropil and cell bodies in the inferior vestibular nucleus (IVe), inferior olive (IO) and arcuate nucleus (Arc) (c, e, g). cell bodies were visible above the neuropil staining for GBR2 immunoreactivity in these regions (arrows) (d, f, h). Scale bars: (a, b) 5 mm, (c–h) 50 mm.

The overall distribution of GABA B receptors in the human hippocampal formation is in agreement with the mRNA distribution demonstrated by in situ hybridisation [4]. In the present study, CA1 pyramidal cells were less intensely stained than CA2 / CA3 pyramidal cells, whereas the converse is true for the rat hippocampus (for GBR1, at least) [25]. Electrophysiological studies have implied that CA1 GABA B receptors are more likely to be found on processes distal from the pyramidal cell bodies [21]. GABA B receptors in the hippocampus are thought to be implicated in the regulation of long term potentiation

(LTP) [11], and the cognitive enhancing abilities of GABA B receptor antagonists have been demonstrated in several species including primates [28]. This effect would presumably involve an interaction with GBR1 as GBR2 has been shown to be insensitive to currently available GABA B receptor antagonists [14,18,42]. Functional GABA B receptors have been reported in the substantia nigra, with GABA B agonists producing hyperpolarisation of pars compacta neurones [35]; this has been substantiated autoradiographically and immunohistochemically in rat brain [8,25]. In fact, GABA B and

122

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

dopamine D 2 receptors are thought to activate the same K 1 conductance in pars compacta neurones [22], therefore their presence would be expected on these neurones. Cerebellar staining for GBR1 and GBR2 is consistent with ligand receptor autoradiography in human cerebellum [1,5], suggesting the primary location of GABA B receptors to be on dendrites in the molecular layer. Purkinje cells were more strongly stained for GBR2 than GBR1, which may reflect differences in trafficking rates out to the dendrites in these cells. Localisation of GABA B receptors in the human dentate nucleus appears to be in conflict with a reported lack of GABA B receptors in this region as demonstrated by receptor autoradiography [2]. This discrepancy is most likely due to the use of the less sensitive [ 3 H]GABA binding assay, rather than as here, specific antibodies, which would not have been available at the time. The presence of GBR1 and GBR2 in the pons and medulla are consistent with reports of GABA B receptor existence in brainstem nuclei. Baclofen has been reported to attenuate the firing and increase K 1 conductance of locus coeruleus neurones in the rat [31,32]. Also, 5-HT release from the rat raphe´ nuclei is reportedly reduced in the presence of GABA B agonists [40]. These provide indications of the existence of pre- and postsynaptic GABA B receptors in nuclei of the brainstem. The inferior olive climbing fibres project to and excite the Purkinje cells of the cerebellum, and are reported to possess GABA B receptors on their terminals [15,16,41]. The cell bodies evident in this study may indicate local inhibition of intrinsic olivary neurones, with neuropil staining representing GABA B receptors relating to cortical afferent terminals. Double fluorescence immunohistochemistry suggests colocalisation of GABA B receptors GBR1 and GBR2 in cerebellar cortex, supporting the notion that GBR1 and GBR2 may form functional heterodimers in this region [26]. None of the regions examined in the present study showed a complete mismatch between GBR1 and GBR2 immunoreactivity, simply differences in cellular and neuropilar localisation. This may simply reflect the characteristics of the two different antisera, or may relate to different trafficking rates for the two proteins away from the cell body into processes. C terminal variants of GBR1 and GBR2 may exist that are not recognised by the antisera used in the present study. Also, the existence of additional undiscovered protein partners for GBR1 or GBR2 subtypes, remains a possibility. Fig. 11. Confocal imaging of GBR1 / GBR2 double fluorescence immunohistochemistry. GBR1 (a, FITC, green) and GBR2 (b, Texas red) antisera detected using fluorescent secondary antibodies. Panel c is the product of merging (a) and (b), showing the co-localisation of GBR1 and GBR2 in a Purkinje cell (arrow) and Purkinje cell dendrite (arrowhead). Weaker fluorescence is also seen in the molecular layer (mol) and much weaker in the granule cell layer (Gr). Intense spots of fluorescence along the granule cell layer–molecular layer border may represent granules of the autofluorescent pigment lipofuscin. Scale bar: 50 mm.

Acknowledgements We gratefully acknowledge the support of GlaxoWellcome, the Health Research Council of New Zealand and the New Zealand Neurological Foundation. AB is sup-

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

ported by the BBSRC, and AOI is the recipient of a MRC / GlaxoWellcome CASE award.

References [1] R.L. Albin, S. Gilman, Autoradiographic localization of inhibitory and excitatory amino acid neurotransmitter receptors in human normal and olivopontocerebellar atrophy cerebellar cortex, Brain Res. 522 (1990) 37–45. [2] R.L. Albin, R.H. Price, S.Y. Sakurai, J.B. Penney, A.B. Young, Excitatory and inhibitory amino acid binding sites in human dentate nucleus, Brain Res. 560 (1991) 350–353. [3] R. Andrade, R.C. Malenka, R.A. Nicoll, A G protein couples serotonin and GABA B receptors to the same channels in hippocampus, Science 234 (1986) 1261–1265. [4] A. Billinton, V.H. Baird, B. Bettler, N.G. Bowery, Alterations in GABA B receptor mRNA expression in hippocampal sclerosis associated with human temporal lobe epilepsy, Br. J. Pharmacol. 122 (1997) 255P. [5] A. Billinton, N. Upton, N.G. Bowery, GABA B receptor isoforms GBR1a and GBR1b, appear to be associated with pre- and postsynaptic elements respectively in rat and human cerebellum, Br. J. Pharmacol. 126 (1999) 1387–1392. [6] A. Billinton, A.O. Ige, A. Wise, J.H. White, G.H. Disney, F.H. Marshall, H.J. Waldvogel, R.L.M. Faull, P.C. Emson, Immunohistochemical localisation of GABA B receptors GBR1 and GBR2 in human brain, Soc. Neurosci. Abstr. 25 (1999) 1976. [7] S. Bischoff, S. Leonhard, N. Reymann, V. Schuler, R. Shigemoto, K. Kaupmann, B. Bettler, Spatial distribution of GABA B R1 receptor mRNA and binding sites in the rat brain, J. Comp. Neurol. 412 (1999) 1–16. [8] N.G. Bowery, A.L. Hudson, G.W. Price, GABAA and GABA B receptor site distribution in the rat central nervous system, Neuroscience 20 (1987) 365–383. [9] A. Couve, A.K. Filippov, C.N. Connolly, B. Bettler, D.A. Brown, S.J. Moss, Intracellular retention of recombinant GABA B receptors, J. Biol. Chem. 273 (1998) 26361–26367. [10] V. Crunelli, N. Leresche, A role for GABA B receptors in excitation and inhibition of thalamocortical cells, Trends Neurosci. 14 (1991) 16–21. [11] C.H. Davies, S.J. Starkey, M.F. Pozza, G.L. Collingridge, GABA autoreceptors regulate the induction of LTP, Nature 349 (1991) 609–611. [12] E. Harlow, P. Lane, Antibodies, A Laboratory Manual, Cold Spring Harbour Press, Cold Spring Harbour, 1988. [13] Y. Ikarashi, M. Yuzurihara, A. Takahashi, H. Ishimaru, T. Shiobara, Y. Maruyama, Modulation of acetylcholine release via GABAA and GABA B receptors in rat striatum, Brain Res. 816 (1999) 238–240. [14] K.A. Jones, B. Borowsky, J.A. Tamm, D.A. Craig, M.M. Durkin, M. Dai, W.J. Yao, M. Johnson, C. Gunwaldsen, L.Y. Huang, C. Tang, Q. Shen, J.A. Salon, K. Morse, T. Laz, K.E. Smith, D. Nagarathnam, S.A. Noble, T.A. Branchek, C. Gerald, GABA B receptors function as a heteromeric assembly of the subunits GABA B R1 and GABA B R2, Nature 396 (1998) 674–679. [15] K. Kannari, M. Tomiyama, M. Matsunaga, Reduction of GABAA and GABA B receptor densities in the cerebellar cortex from 3acetylpyridine-induced ataxic rats, Brain Res. 656 (1994) 413–416. [16] K. Kato, H. Fukuda, Reduction of GABA B receptor binding induced by climbing fibre degeneration in the rat cerebellum, Life Sci. 37 (1985) 279–288. [17] K. Kaupmann, K. Huggel, J. Heid, P.J. Flor, S. Bischoff, S.J. Mickel, G. McMaster, C. Angst, H. Bittiger, W. Froestl, B. Bettler, Expression cloning of GABA B receptors uncovers similarity to metabotropic glutamate receptors, Nature 386 (1997) 239–246.

123

[18] K. Kaupmann, B. Malitschek, V. Schuler, J. Heid, W. Froestl, P. Beck, J. Mosbacher, S. Bischoff, A. Kulik, R. Shigemoto, A. Karschin, B. Bettler, GABA B -receptor subtypes assemble into functional heteromeric complexes, Nature 396 (1998) 683–687. [19] G.J. Kilpatrick, M.S. Muhyaddin, P.J. Roberts, G.N. Woodruff, GABA B binding sites on rat striatal synaptic membranes, Br. J. Pharmacol. 78 (1983) 6P. ¨ ¨ [20] R. Kuner, G. Kohr, S. Grunewald, G. Eisenhardt, A. Bach, H.C. Kornau, Role of heteromer formation in GABA B receptor function, Science 283 (1999) 74–77. [21] J.C. Lacaille, S. Williams, D.D. Samulack, Inhibition of hippocampal pyramidal cells by GABA B interneurons, Pharmacol.Commun. 2 (1992) 88–92. [22] M.G. Lacey, N.B. Mercuri, R.A. North, On the potassium conductance increase activated by GABA B and dopamine D 2 receptors in rat substantia nigra neurones, J. Physiol. 401 (1988) 437–453. [23] X.Y. Lu, M.B. Ghasemzadeh, P.W. Kalivas, Regional distribution and cellular localization of g-aminobutyric acid subtype 1 receptor mRNA in the rat brain, J. Comp. Neurol. 407 (1999) 166–182. [24] L.M. McLatchie, N.J. Fraser, M.J. Main, A. Wise, J. Brown, N. Thompson, R. Solari, M.G. Lee, S.M. Foord, RAMPs regulate the transport and ligand specificity of the calcitonin-receptor-like receptor, Nature 393 (1998) 333–339. [25] M. Margeta-Mitrovic, I. Mitrovic, R.C. Riley, L.Y. Jan, A.I. Basbaum, Immunohistochemical localization of GABA B receptors in the rat central nervous system, J. Comp. Neurol. 405 (1999) 299– 321. [26] F.H. Marshall, K.A. Jones, K. Kaupmann, B. Bettler, GABA B receptors — the first 7TM heterodimers, Trends Pharmacol. Sci. 20 (1999) 396–399. [27] S.C. Martin, S.J. Russek, D.H. Farb, Molecular identification of the human GABA B R2: cell surface expression and coupling to adenylyl cyclase in the absence of GABA B R1, Mol. Cell. Neurosci. 13 (1999) 180–191. [28] C. Mondadori, J. Jaekel, G. Preiswerk, CGP36742: the first orally active GABA B blocker improves the cognitive performance of mice, rats, and rhesus monkeys, Behav. Neural Biol. 60 (1993) 62–68. [29] R. Moratalla, N.G. Bowery, Chronic lesion of corticostriatal fibres reduces GABA B but not GABAA binding in rat caudate putamen: an autoradiographic study, Neurochem. Res. 16 (1991) 309–315. [30] G.Y. Ng, J. Clark, N. Coulombe, N. Ethier, T.E. Hebert, R. Sullivan, S. Kargman, A. Chateauneuf, N. Tsukamoto, T. McDonald, P. Whiting, E. Mezey, M.P. Johnson, Q. Liu, L.F. Kolakowski, J.F. Evans, T.I. Bonner, G.P. O’Neill, Identification of a GABA B receptor subunit, gb2, required for functional GABA B receptor activity, J. Biol. Chem. 274 (1999) 7607–7610. [31] H.R. Olpe, M.W. Steinmann, R.G. Hall, F. Brugger, M.F. Pozza, GABAA and GABA B receptors in locus coeruleus: effects of blockers, Eur. J. Pharmacol. 149 (1988) 183–185. [32] S.S. Osmanovic, S.A. Shefner, Baclofen increases the potassium conductance of rat locus coeruleus neurons recorded in brain slices, Brain Res. 438 (1988) 124–136. [33] T. Phillips, A. Makoff, S. Brown, S. Rees, P. Emson, Localization of mGluR4 protein in the rat cerebral cortex and hippocampus, NeuroReport 8 (1997) 3349–3354. [34] T. Phillips, A. Makoff, E. Murrison, M. Mimmack, H. Waldvogel, R. Faull, S. Rees, P. Emson, Immunohistochemical localisation of mGluR7 protein in the rodent and human cerebellar cortex using subtype specific antibodies, Mol. Brain Res. 57 (1998) 132–141. [35] G.R. Seabrook, W. Howson, M.G. Lacey, Electrophysiological characterization of potent agonists and antagonists at pre- and postsynaptic GABA B receptors on neurones in rat brain slices, Br. J. Pharmacol. 101 (1990) 949–957. [36] G.R. Seabrook, W. Howson, M.G. Lacey, Subpopulations of GABAmediated synaptic potentials in slices of rat dorsal striatum are differentially modulated by presynaptic GABA B receptors, Brain Res. 562 (1991) 332–334.

124

A. Billinton et al. / Molecular Brain Research 77 (2000) 111 – 124

[37] P.K. Smith, R.I. Krohn, G.T. Hermanson, A.K. Malia, F.H. Gartner, M.D. Provenzano, E.K. Fujimoto, N.M. Goeke, B.J. Olson, D.C. Klent, Measurement of protein using bicinchoninic acid, Anal. Biochem. 150 (1985) 76–85. [38] O.C. Snead, Basic mechanisms of generalized absence seizures, Ann. Neurol. 37 (1995) 146–157. [39] T. Takahashi, Y. Kajikawa, T. Tsujimoto, G-Protein-coupled modulation of presynaptic calcium currents and transmitter release by a GABA B receptor, J. Neurosci. 18 (1998) 3138–3146. [40] R. Tao, Z. Ma, S.B. Auerbach, Differential regulation of 5-hydroxytryptamine release by GABAA and GABA B receptors in midbrain raphe nuclei and forebrain of rats, Br. J. Pharmacol. 119 (1996) 1375–1384.

[41] R. Vigot, J.M. Billard, C. Battini, Reduction of GABA inhibition in Purkinje and cerebellar nuclei neurons in climbing fibre deafferented cerebella of rat, Neurosci. Res. 17 (1993) 249–255. [42] J.H. White, A. Wise, M.J. Main, N.J. Fraser, G.H. Disney, A.A. Barnes, P. Emson, S.M. Foord, F.H. Marshall, Heterodimerization is required for the formation of a functional GABA B receptor, Nature 396 (1998) 679–682. [43] K.K.L. Yung, T.K.Y. Ng, C.K.C. Wong, Subpopulations of neurons in the rat neostriatum display GABA B R1 receptor immunoreactivity, Brain Res. 830 (1999) 345–352.