Search

The striatum is the main input nucleus of the basal ganglia, mediating motor and cognitive functions. Striatal projection neurons are GABAergic medium spiny neurons (MSN), expressing either the dopamine receptor type 1 (D

Periglomerular (PG) cells in the rodent olfactory bulb are heterogeneous anatomically and neurochemically. Here we investigated whether major classes of PG cells use γ-aminobutyric acid (GABA) as a neurotransmitter. In addition to three known subtypes of PG cells expressing tyrosine hydroxylase (TH), calbindin D-28k (CB), and calretinin (CR), we identified a novel PG cell population containing the GABAA receptor α5 subunit. Consistent with previous studies in the rat, we found that TH-positive cells were also labeled with antibodies against GABA, whereas PG cells expressing CB or the α5 subunit were GABA-negative. Using GAD67-GFP knockin mice, we found that all PG cell subtypes expressed GAD67-GFP. Calretinin labeled the major fraction (44%) of green fluorescent protein (GFP)-positive cells, followed by TH (16%), CB (14%), and the α5 subunit (13%). There was no overlap between these neuronal populations, which accounted for ≈85% of GAD67-GFP-positive cells. We then demonstrated that PG cells labeled for TH, CB, or CR established dendrodendritic synapses expressing glutamic acid decarboxylase (GAD) or the vesicular inhibitory amino acid transporter, VGAT, irrespective of their immunoreactivity for GABA. In addition, CB-, CR-, and TH-positive dendrites were apposed to GABAA receptor clusters containing the α1 or α3 subunits, which are found in mitral and tufted cells, and the α2 subunit, which is expressed by PG cells. Together, these findings indicate that all major subtypes of PG cells are GABAergic. In addition, they show that PG cells provide GABAergic input to the dendrites of principal neurons and are interconnected with other GABAergic interneurons, which most likely are other PG cells. J. Comp. Neurol. 502:990–1002, 2007. © 2007 Wiley-Liss, Inc.

GABA(B) receptors mediate slow inhibitory GABAergic neurotransmission. They are encoded by two distinct subunits, GABA(B1) (GBR1) and GABA(B2) (GBR2), with two major isoforms of GBR1, GBR1a and GBR1b, arising from differential promoter usage. Heterodimerization of GBR1 and GBR2 is essential for GABA(B) receptor function, as shown in recombinant expression systems and in GBR1(-/-) mice. GABA(B) receptors are highly expressed during ontogeny, prior to synaptogenesis, but their developmental function remains elusive. Here we investigated the postnatal development of GABA(B) receptors in rodent brain, focusing on potential differences in the spatial and temporal expression pattern of GBR1 and GBR2. Immunohistochemistry with subunit-specific antibodies revealed a widespread staining for GBR1a and GBR2 in neonatal rodent brain. During the first 2 weeks, these two subunits exhibited largely overlapping regional distribution, but with profound distinctions in cellular and subcellular localization. The adult-like pattern was established during the third week, with a prominent up-regulation of GBR1b, extensively codistributed with GBR2. Several unexpected features were noted at early stages, notably, a selective GBR2 staining of axonal tracts, such as the corticothalamic projection, and a prominent GBR1 expression in astrocytes. The specificity of the antibody labeling was verified in GBR1- and GBR2-knockout mice. In addition, the analysis of these mutants revealed a partial preservation of GBR2 staining in GBR1(-/-) mice and vice versa. Altogether, the results suggest a functional role for GBR1 and GBR2 proteins in immature brain in addition to their contribution to dimeric GABA(B) receptor complexes.

Clustering of gamma aminobutyric acid (GABA)A receptors to postsynaptic sites requires the presence of both the γ2 subunit and gephyrin. Here, we analyzed by double-immunofluorescence staining the colocalization of gephyrin and major GABAA-receptor subtypes distinguished by the subunits α1, α2, α3, or γ2 in adult rat brain. By using confocal laser scanning microscopy, GABAA-receptor subunit staining revealed brightly stained clusters that were colocalized with gephyrin-positive clusters of similar size and distribution in several brain regions, including cerebellum, hippocampus, thalamus, and olfactory bulb. In addition, a diffuse staining was observed for GABAA-receptor subunits in the neuropil, presumably representing extrasynaptic receptors. Overall, only few gephyrin-positive clusters were not colocalized with GABAA-receptor subunit clusters. Electron microscopic analysis in cerebellar cortex confirmed the selective postsynaptic localization of gephyrin. High-resolution images (voxel size, 50 × 50 × 150 nm) were restored with an iterative image deconvolution procedure based on a measured point-spread function to analyze the colocalization between GABAA-receptor subunits and gephyrin in individual clusters. This analysis revealed a considerable heterogeneity in the micro-organization of these presumptive GABAergic postsynaptic sites. For instance, whereas gephyrin- and γ2 subunit-positive clusters largely overlapped in the cerebellar molecular layer, the colocalization was only partial in glomeruli of the granule cell layer, where small gephyrin clusters typically were “embedded” in larger GABAA-receptor clusters. These findings show that gephyrin is associated with a majority of GABAA-receptor subtypes in brain, and document the usefulness of image deconvolution for analyzing the structural organization of the postsynaptic apparatus by fluorescence microscopy. J. Comp. Neurol. 420:481–498, 2000. © 2000 Wiley-Liss, Inc.

The inhibitory amino acid gamma-aminobutyric acid (GABA) is widely distributed in the basal ganglia. It plays a critical role in the functioning of the striatum as it is the transmitter of projection neurons and sub-populations of interneurons, as well as afferents from the globus pallidus. Some of the factors controlling GABA transmission are the type(s) of GABA receptor expressed at the site of transmission, their subunit composition, and their location in relation to GABA release sites. To address these issues, we examined the sub-cellular localization of subunits of the GABA(A) receptor in the striatum of the rat. Sections of freeze-substituted, Lowicryl-embedded striatum were immunolabelled by the post-embedding immunogold technique with antibodies specific for subunits of the GABA(A) receptor. Immunolabelling for alpha1, beta2/3, and gamma2 GABA(A) receptor subunits was primarily located at symmetrical synapses on perikarya, dendrites, and spines. Quantitative analysis of the distribution of immunolabelling for the beta2/3 subunits revealed that the majority of membrane associated immunogold particles were at synapses and that, on average for the whole population, they were evenly distributed across the synapse. Double labelling for the beta2/3 subunits and for GABA itself revealed that receptor-positive synapses were formed by at least two populations of terminals. One population (59.3%) of terminals forming receptor-positive synapses was positive for GABA, whereas the other (40.7%) had low or undetectable levels of GABA. Furthermore, the post-synaptic neurons were characterised on neurochemical and morphological grounds as both medium spiny neurons and GABA interneurons. Triple immunolabelling revealed the co-localization of alpha1, beta2/3, and gamma2 subunits at some symmetrical axodendritic synapse. It is concluded that fast GABA(A)-mediated transmission occurs primarily at symmetrical synapses within the striatum, that the populations of boutons giving rise to receptor-positive synapses are heterogeneous, and that previously reported co-existence of different subunits of the GABA(A) receptor at the cellular level also occurs at the level of individual synapses.

The cellular distribution of utrophin, the autosomal homologue of dystrophin, was investigated in developing and adult rat and mouse brain by in situ hybridization and immunohistochemistry. Digoxigenin-labeled cRNA probes complementary to N-terminal, rod-domain, and C-terminal encoding sequences of utrophin were used to differentiate between full-length and short C-terminal isoforms. Largely overlapping distribution patterns were seen for the three probes in neurons of cerebral cortex, accessory olfactory bulb, and several sensory and motor brainstem nuclei as well as in blood vessels, pia mater, and choroid plexus. The C-terminal probe was detected in addition in the main olfactory bulb, striatum, thalamic reticular nucleus, and hypothalamus, suggesting a selective expression of G-utrophin in these neurons. Western blot analysis with isoform-specific antisera confirmed the expression of both full-length and G-utrophin in brain. Immunohistochemically, only full-length utrophin was detected in neurons, in close association with the plasma membrane. In addition, intense staining was seen in blood vessels, meninges, and choroid plexus, selectively localized in the basolateral membrane of immunopositive epithelial cells. The expression pattern of utrophin was already established at early postnatal stages and did not change thereafter. Double-labeling analysis revealed that utrophin and dystrophin are differentially expressed on the cellular and subcellular levels in juvenile and adult brain. Likewise, in mice lacking full-length dystrophin isoforms (mdx mice), no change in utrophin expression and distribution could be detected in brain, although utrophin was markedly up-regulated in muscle cells. These results suggest that utrophin and dystrophin are independently regulated and have distinct functional roles in CNS neurons.

The regional and cellular localisation of gamma-aminobutyric acid(A) (GABA(A)) receptors was investigated in the human basal ganglia using receptor autoradiography and immunohistochemical staining for five GABA(A) receptor subunits (alpha(1), alpha(2), alpha(3), beta(2, 3), and gamma(2)) and other neurochemical markers. The results demonstrated that GABA(A) receptors in the striatum showed considerable subunit heterogeneity in their regional distribution and cellular localisation. High densities of GABA(A) receptors in the striosome compartment contained the alpha(2), alpha(3), beta(2, 3), and gamma(2) subunits, and lower densities of receptors in the matrix compartment contained the alpha(1), alpha(2), alpha(3), beta(2,3), and gamma(2) subunits. Also, six different types of neurons were identified in the striatum on the basis of GABA(A) receptor subunit configuration, cellular and dendritic morphology, and chemical neuroanatomy. Three types of alpha(1) subunit immunoreactive neurons were identified: type 1, the most numerous (60%), were medium-sized aspiny neurons that were immunoreactive for parvalbumin and alpha(1), beta(2,3), and gamma(2) subunits; type 2 (38%) were medium-sized to large aspiny neurons immunoreactive for calretinin and alpha(1), alpha(3), beta(2,3), and gamma(2) subunits; and type 3 (2%) were large sparsely spiny neurons immunoreactive for alpha(1), alpha(3), beta(2,3), and gamma(2) subunits. Type 4 neurons were calbindin-positive and immunoreactive for alpha(2), alpha(3), beta(2,3), and gamma(2) subunits. The remaining neurons were immunoreactive for choline acetyltransferase (ChAT) and alpha(3) subunit (type 5) or were neuropeptide Y-positive with no GABA(A) receptor subunit immunoreactivity (type 6). The globus pallidus contained three types of neurons: types 1 and 2 were large neurons and were immunoreactive for alpha(1), alpha(3), beta(2,3), and gamma(2) subunits and for parvalbumin alone (type 1) or for both parvalbumin and calretinin (type 2); type 3 neurons were medium-sized and immunoreactive for calretinin and alpha(1), beta(2, 3), and gamma(2) subunits. These results show that the subunit composition of GABA(A) receptors displays considerable regional and cellular variation in the human striatum but are more homogeneous in the globus pallidus.

In the present study, we investigated the key molecules that determine gamma-aminobutyric acid (GABA)ergic signal transduction in the parabrachial/Kolliker-Fuse complex (PB/KF) by means of immunocytochemistry and in situ hybridization. Our data demonstrate a dense plexus of GABA-immunoreactive (-ir) varicosities throughout the nuclei of the PB and the KF. The number of neurons expressing GAD65 or GAD67 mRNA was fairly low in the PB, whereas caudally in the KF an accumulation of GAD-expressing neurons was observed. The GABA transporter-3 (GAT-3) was detected in all parts of the PB/KF, whereas immunolabeling for GAT1 was not observed. All nuclei of the PB and the KF exhibited immunoreactivity for the gamma2-, alpha2-, and alpha3-subunits of the GABA(A) receptor. Gamma2-ir was strong and similar in all PB/KF nuclei. In contrast, alpha2-labeling was particularly intense in the superior lateral PB, and alpha3-labeling was most prominent in the external lateral and external medial PB, compared with the remaining nuclei. With respect to the subcellular localization, we found gamma2-ir in cell bodies and higher order dendrites, whereas alpha2- and alpha3-ir was predominantly found in cell bodies. Immunolabeling for the beta2/3- and the alpha1-subunit was seen in cell bodies and presumed dendritic profiles. The staining intensity was strongest in the dorsal lateral PB. Most importantly, the external lateral PB and the waist area were totally devoid of beta2/3- and alpha1-ir. Our data suggest that neural processing in the PB/KF is under a strong GABAergic inhibition that is apparently mediated by different types of GABA(A) receptors in functionally different pathways through the PB/KF.

The distribution of gamma-aminobutyric acid(A) (GABA(A)) receptors was investigated in the basal ganglia in the baboon brain by using receptor autoradiography and the immunohistochemical localisation of the alpha1 and beta2,3 subunits of the GABA(A) receptor by light and electron microscopy. In the caudate-putamen, the alpha1 subunit was distributed in high densities in the matrix compartment, and the beta2,3 subunits were more homogeneously distributed; the globus pallidus showed lower levels of the alpha1 and beta2,3 subunits. Four types of alpha1 subunit immunoreactive neurons were identified in the baboon striatum: the most numerous (75%) were type 1 medium-sized aspiny neurons; type 2 (2%) were large aspiny neurons with an indented nuclear membrane located in the ventral striatum; type 3 neurons were the least numerous (1%) and were comprised of large neurons in the ventromedial regions of the striatum; and type 4 (22%) neurons were medium to large aspiny neurons located in striosomes. At the ultrastructural level, alpha1 and beta2,3 subunit immunoreactivity was localised in the neuropil of the striatum in both symmetrical and asymmetrical synaptic contacts. In the globus pallidus, alpha1 and beta2,3 subunits were localised on large neurons and were found in three types of synaptic terminals: type 1 terminals were small and established symmetrical synapses; type 2 terminals were large; and type 3 terminals formed small synaptic terminals with subjunctional dense bodies. These results show that the subunit composition of GABA(A) receptors varies between the striosome and the matrix compartments in the striatum and that there is receptor subunit homogeneity in the globus pallidus.

The tubulin-binding protein gephyrin is essential for the formation of postsynaptic glycine-receptor clusters in cultured spinal neurons. In addition, there is increasing evidence that gephyrin can also be present at nonglycinergic synapses. Here we analyzed immunocytochemically the subcellular localization of gephyrin in the main olfactory bulb of the rat and compared its distribution with that of gamma-aminobutyric acid (GABA) and of two major GABA(A)-receptor subunits. Gephyrin was selectively localized to the postsynaptic side of symmetric synaptic junctions, where the presynaptic terminals contained GABA. Moreover, gephyrin colocalized extensively with the alpha1 and gamma2 subunits of the GABA(A) receptor. In contrast, gephyrin was not detected at presumed glutamatergic synapses. These results indicate that gephyrin is not uniquely associated with glycine receptors, but can also be found at distinct GABAergic synapses. Thus, they raise the possibility that gephyrin is involved in anchoring certain GABA(A)-receptor subtypes in the postsynaptic membrane.

Conventional immunohistochemistry provides little evidence for the synaptic localization of ionotropic neurotransmitter receptors, suggesting that their epitopes are not readily accessible in situ. Here, we have adapted antigen retrieval procedures based on microwave irradiation to enhance the immunohistochemical staining of gamma-aminobutyric acid type A (GABA[A]) and N-methyl-D-aspartate (NMDA) receptor subunits in rat brain tissue. Microwave irradiation of fixed tissue produced a marked reduction of nonspecific staining, allowing an improved detection of GABA(A) receptor subunits. However, staining of NMDA receptor subunits remained suboptimal. In contrast, microwave irradiation of cryostat sections prepared from fresh tissue resulted in a major enhancement of both NMDA and GABAA receptor subunit staining. The diffuse, partially intracellular signals were largely replaced by numerous, intensely immunoreactive puncta outlining neuronal somata and dendrites, highly suggestive ofsynaptic receptors. In hippocampus CA1-CA3 fields, the NR2Aand NR2B subunit positive puncta exhibited an extensive colocalization in the stratum oriens and radiatum, whereas pyramidal cell bodies, which receive no excitatory synapses, were unstained. In addition, the NR2A subunit, but not the NR2B subunit, was selectively detected on pyramidal cell dendrites in the stratum lucidum of CA3, suggesting a selective targeting to sites of mossy fiber input. For the GABAA receptor subunits, the most striking change induced by this protocol was the selective staining of the axon initial segment of cortical and hippocampal pyramidal cells. The alpha2 subunit immunoreactivity was particularly prominent in these synapses. In control experiments, the staining of cytoskeletal proteins (neurofilaments, glial fibrillary acid protein) was not influenced by prior microwave irradiation. The enhancement of cell-surface-associated staining is therefore strongly suggestive of an 'unmasking' of subunit epitopes by the microwave treatment. These results reveal a remarkable specificity in the synaptic targeting of NMDA and GABAA receptor subunits in hippocampal and neocortical neurons, suggesting that individual neurons can express multiple receptor subtypes in functionally distinct synapses.

Spinally projecting neurons of the ventromedial medulla (VMM) compose an important efferent pathway for the modulation of nociception. These neurons receive a substantial gamma-aminobutyric acid (GABA)-ergic input, but the GABA receptor that mediates this input is unknown. This study examined the distribution of GABA(A) receptor alpha1 and alpha3 subunits in serotonergic and nonserotonergic neurons of the VMM that project to the dorsal horn in the rat. A pledget of Gelfoam soaked in Fluoro-Gold was placed at the thoracolumbar junction of the spinal cord to label spinally projecting neurons. Alternate sections of the medulla were then incubated with a mixture of antisera to either serotonin and the alpha1 subunit, or to serotonin and the alpha3 subunit of the GABA(A) receptor. Nearly 30% of spinally projecting neurons in the VMM were immunoreactive for the alpha1 subunit. A similar percentage of spinally projecting neurons in the VMM were immunoreactive for the alpha3 subunit, although diffuse cellular labeling combined with intense staining of processes in the neuropil precluded a rigorous semi-quantitative estimation of this population. No alpha1-subunit-immunoreactive neurons colocalized serotonin. In contrast, serotonergic neurons were immunoreactive for the alpha3 subunit. However, these double-labeled neurons were a modest percentage of the serotonergic population. A small percentage of spinally projecting serotonergic neurons was immunoreactive for the alpha3 subunit. These results suggest that significant numbers of spinally projecting serotonergic and nonserotonergic neurons of the VMM possess GABA(A) receptors that differ in their respective subunit compositions and that both classes of neurons may mediate the antinociception produced by the microinjection of GABA(A) receptor antagonists in the VMM.

Gamma aminobutyric acid (GABA)A-receptors are expressed in fetal mammalian brain before the onset of synaptic inhibition, suggesting their involvement in brain development. In this study, we have analyzed the maturation of the GABAA-receptor in the marmoset monkey forebrain to determine whether distinct receptor subtypes are expressed at particular stages of pre- and postnatal ontogeny. The distribution of the subunits alpha 1, alpha 2, and beta 2,3 was investigated immunohistochemically between embryonic day 100 (6 weeks before birth) and adulthood. Prenatally, the alpha 2- and beta 2,3-subunit-immunoreactivity (-IR) was prominent throughout the forebrain, whereas the alpha 1-subunit-IR appeared in selected regions shortly before birth. The alpha 2-subunit-IR disappeared gradually to become restricted to a few regions in adult forebrain. By contrast, the alpha 1-subunit-IR increased dramatically after birth and replaced the alpha 2-subunit in the basal forebrain, pallidum, thalamus, and most of the cerebral cortex. Staining for the beta 2,3-subunits was ubiquitous at every age examined, indicating their association with either the alpha 1- or the alpha 2-subunit in distinct receptor subtypes. In neocortex, the alpha 1 -subunit-IR was first located selectively to layers IV and VI of primary somatosensory and visual areas. Postnatally, it increased throughout the cortex, with the adult pattern being established only during the second year. The switch in expression of the alpha 1- and alpha 2- subunits indicates that the subunit composition of major GABAA-receptor subtypes changes during ontogeny. This change coincides with synaptogenesis, suggesting that the emergence of alpha 1- GABAA-receptors parallels the formation of inhibitory circuits. A similar pattern has been reported in rat, indicating that the developmental regulation of GABAA-receptors is conserved across species, possibly including man. However, the marmoset brain is more mature than the rat brain at the onset of alpha 1-subunit expression, suggesting that alpha 1-GABAA-receptors are largely dispensable in utero, but may be required for information processing after birth.

GABAA-receptors display an extensive structural heterogeneity based on the differential assembly of a family of at least 15 subunits (alpha 1-6, beta 1-3, gamma 1-3, delta, rho 1-2) into distinct heteromeric receptor complexes. The subunit composition of receptor subtypes is expected to determine their physiological properties and pharmacological profiles, thereby contributing to flexibility in signal transduction and allosteric modulation. In heterologous expression systems, functional receptors require a combination of alpha-, beta-, and gamma-subunit variants, the gamma 2-subunit being essential to convey a classical benzodiazepine site to the receptor. The subunit composition and stoichiometry of native GABAA-receptor subtypes remain unknown. The aim of this study was to identify immunohistochemically the main subunit combinations expressed in the adult rat brain and to allocate them to identified neurons. The regional and cellular distribution of seven major subunits (alpha 1, alpha 2, alpha 3, alpha 5, beta 2,3, gamma 2, delta) was visualized by immunoperoxidase staining with subunit-specific antibodies (the beta 2- and beta 3-subunits were covisualized with the monoclonal antibody bd-17). Putative receptor subtypes were identified on the basis of colocalization of subunits within individual neurons, as analyzed by confocal laser microscopy in double- and triple-immunofluorescence staining experiments. The results reveal an extraordinary heterogeneity in the distribution of GABAA-receptor subunits, as evidenced by abrupt changes in immunoreactivity along well-defined cytoarchitectonic boundaries and by pronounced differences in the cellular distribution of subunits among various types of neurons. Thus, functionally and morphologically diverse neurons were characterized by a distinct GABAA-receptor subunit repertoire. The multiple staining experiments identified 12 subunit combinations in defined neurons. The most prevalent combination was the triplet alpha 1/beta 2,3/gamma 2, detected in numerous cell types throughout the brain. An additional subunit (alpha 2, alpha 3, or delta) sometimes was associated with this triplet, pointing to the existence of receptors containing four subunits. The triplets alpha 2/beta 2,3/gamma 2, alpha 3/beta 2,3/gamma 2, and alpha 5/beta 2,3/gamma 2 were also identified in discrete cell populations. The prevalence of these seven combinations suggest that they represent major GABAA-receptor subtypes. Five combinations also apparently lacked the beta 2,3-subunits, including one devoid of gamma 2-subunit (alpha 1/alpha 2/gamma 2, alpha 2/gamma 2, alpha 3/gamma 2, alpha 2/alpha 3/gamma 2, alpha 2/alpha 5/delta).(ABSTRACT TRUNCATED AT 400 WORDS)

The rat spinal cord receives noradrenergic (NA) projections from the locus coeruleus (LC) and the A5 and A7 groups. In contradiction to previous statements about the distribution of descending NA axons, we have recently proposed that in the rat LC neurons project primarily to the dorsal horn and intermediate zone, whereas A5 and A7 neurons project to somatic motoneurons and the intermediolateral cell column. The aim of the present study was to determine the funicular course and terminal distribution of descending NA axons from the LC and from the A5 and A7 groups. The organization of the coeruleospinal projection was analyzed by using the anterograde tracer Phaseolus vulgaris leucoagglutinin in combination with dopaminehyphen;β-hydroxylase immunohistochemistry. The trajectory of A5 and A7 axons was studied in spinal cord sections of rats following ablation of the coeruleospinal projection with the neurotoxin DSP-4. To assess the relative contribution of the LC and the A5 and A7 groups to the NA innervation of the spinal cord, unilateral injections of the retrograde tracer True Blue were made at cervical, thoracic, and lumbar levels, and retrogradely labeled NA neurons were identified by dopamine-β-hydroxylase immunofluorescence. The results of the anterograde tracing experiments confirm our previous findings that LC neurons project most heavily to the dorsal horn and intermediate zone. Analysis of horizontal sections revealed that LC axons descend the length of the spinal cord within layers I and II. In contrast to the intragriseal course of LC fibers, A5 and A7 axons travel in the ventral and dorsolateral funiculi and terminate in the ventral horn and the intermediolateral cell column. Retrograde transport studies indicate that the contribution of the A5 and A7 groups to the NA projection to the spinal cord is greater than that of the LC. We conclude that descending axons of the LC and A5 and A7 groups differ in their course and distribution within the spinal cord. The documentation of a definite topographic order in the bulbospinal NA projections suggests that the LC and the A5 and A7 groups have different functional capacities. The LC is in a position to influence the processing of sensory inputs, in particular nociceptive inputs, whereas A5 and A7 neurons are likely to influence motoneurons.

Interneurons of the cerebellum granule cell layer (GCL) form distinct populations. Golgi cells extend dendrites in the molecular layer (ML) and innervate granule cells. In contrast, Lugaro cells have dendrites confined to the GCL but innervate interneurons in the ML, and globular cells have both their dendrites and axons in the ML. The latter cells were described recently and remain poorly characterized. Although several neurochemical markers have been associated selectively with GCL interneurons, it is unclear how they relate to their morphological classification and neurochemical phenotype (glycinergic and/or γ-aminobutyric acid [GABA]ergic). Here, we performed a detailed characterization of GCL interneurons in mice expressing enhanced green fluorescent protein (GFP) in glycinergic and GABAergic neurons, respectively. By using immunofluorescence for metabotropic glutamate receptor 2 (mGluR2) and neurogranin as markers, we demonstrate the existence of five non-overlapping subsets of Golgi cells: about 65% are glycinergic/GABAergic and co-express both markers. Two small subsets (5–10%) also contain both neurotransmitters but express only mGluR2; they are distinguished by cell body size and location in the GCL. The fourth subset (15%) is GABAergic only and expresses neurogranin. The fifth subset (5%) is glycinergic only and lacks both markers. Thus, the heterogeneity of Golgi cells suggests that they belong to specific functional circuits and are differentially regulated by mGluRs and Ca2+-calmodulin-dependent signaling pathways. In contrast to Golgi cells, Lugaro and globular cells are glycinergic/GABAergic and lack mGluR2 and neurogranin. They each represent at least 15% of GCL interneurons and extensively innervate stellate and basket cells, but not Purkinje cells, emphasizing their contribution to inhibitory control of ML interneurons. J. Comp. Neurol. 500:71–83, 2007. © 2006 Wiley-Liss, Inc.

Targeted deletion of the alpha1 subunit gene results in a profound loss of gamma-aminobutyric acid type A (GABA(A)) receptors in adult mouse brain but has only moderate behavioral consequences. Mutant mice exhibit several adaptations in GABA(A) receptor subunit expression, as measured by Western blotting. By using immunohistochemistry, we investigated here whether these adaptations serve to replace the missing alpha1 subunit or represent compensatory changes in neurons that normally express these subunits. We focused on cerebellum and thalamus and distinguished postsynaptic GABA(A) receptor clusters by their colocalization with gephyrin. In the molecular layer of the cerebellum, alpha1 subunit clusters colocalized with gephyrin disappeared from Purkinje cell dendrites of mutant mice, whereas alpha3 subunit/gephyrin clusters, presumably located on dendrites of Golgi interneurons, increased sevenfold, suggesting profound network reorganization in the absence of the alpha1 subunit. In thalamus, a prominent increase in alpha3 and alpha4 subunit immunoreactivity was evident, but without change in regional distribution. In the ventrobasal complex, which contains primarily postsynaptic alpha1- and extrasynaptic alpha4-GABA(A) receptors, the loss of alpha1 subunit was accompanied by disruption of gamma2 subunit and gephyrin clustering, in spite of the increased alpha4 subunit expression. However, in the reticular nucleus, which lacks alpha1-GABA(A) receptors in wild-type mice, postsynaptic alpha3/gamma2/gephyrin clusters were unaffected. These results demonstrate that adaptive responses in the brain of alpha1(0/0) mice involve reorganization of GABAergic circuits and not merely replacement of the missing alpha1 subunit by another receptor subtype. In addition, clustering of gephyrin at synaptic sites in cerebellum and thalamus appears to be dependent on expression of a GABA(A) receptor subtype localized postsynaptically.

Gephyrin is a postsynaptic scaffolding protein involved in clustering of glycine- and GABA(A) receptors at inhibitory synapses. The role of gephyrin in GABAergic synapses, the nature of its interactions with GABA(A) receptors, and the mechanisms of targeting to GABAergic synapses are largely unknown. To gain further insights into these questions, the formation of GABA(A) receptor and gephyrin clusters and their distribution relative to presynaptic terminals were investigated in immature cultures of embryonic hippocampal neurons using triple immunofluorescence staining. GABA(A) receptor clusters, labeled for the alpha2 subunit, formed independently of gephyrin clusters, and were distributed on neurites at constant densities, either extrasynaptically or, to a lesser extent, postsynaptically, apposed to synapsin-I-positive axon terminals. In contrast, gephyrin clusters were always associated with GABA(A) receptors and were preferentially localized postsynaptically. Their density increased linearly with the extent of innervation, which developed rapidly during the first week in vitro. These results suggested that GABA(A) receptor clustering is mediated by cell-autonomous mechanisms independent of synapse formation. Their association with gephyrin is dynamically regulated and may contribute to stabilization at postsynaptic sites. Labeling for vesicular glutamate transporters revealed that most synapses in these immature cultures were presumably glutamatergic, implying that postsynaptic GABA(A) receptor and gephyrin clusters initially were located in "mismatched" synapses. However, clusters appropriately localized in GABAergic synapses were distinctly larger and more intensely stained. Altogether, these results demonstrate that the targeting of GABA(A) receptor and gephyrin clusters to GABAergic synapses occurs secondarily and is regulated by presynaptic factors that are not essential for clustering.

Although glycine is a major inhibitory transmitter in the mammalian CNS, the role of glycinergic neurons in defined neuronal circuits remains ill defined. This is due in part to difficulties in identifying these cells in living slice preparations for electrophysiological recordings and visualizing their axonal projections. To facilitate the morphological and functional analysis of glycinergic neurons, we generated bacterial artificial chromosome (BAC) transgenic mice, which specifically express enhanced green fluorescent protein (EGFP) under the control of the promotor of the glycine transporter (GlyT) 2 gene, which is a reliable marker for glycinergic neurons. Neurons expressing GlyT2-EGFP were intensely fluorescent, and their dendrites and axons could be visualized in great detail. Numerous positive neurons were detected in the spinal cord, brainstem, and cerebellum. The hypothalamus, intralaminar nuclei of the thalamus, and basal forebrain also received a dense GlyT2-EGFP innervation, whereas in the olfactory bulb, striatum, neocortex, hippocampus, and amygdala positive fibers were much less abundant. No GlyT2-EGFP-positive cell bodies were seen in the forebrain. On the subcellular level, GlyT2-EGFP fluorescence was colocalized extensively with glycine immunoreactivity in somata and dendrites and with both glycine and GlyT2 immunoreactivity in axon terminals, as shown by triple staining at all levels of the neuraxis, confirming the selective expression of the transgene in glycinergic neurons. In slice preparations of the spinal cord, no difference between the functional properties of EGFP-positive and negative neurons could be detected, confirming the utility of visually identifying glycinergic neurons to investigate their functional role in electrophysiological studies.

The opener muscle of the dactyl of the walking leg of crayfish is innervated by one excitatory axon releasing glutamate and one inhibitory axon releasing GABA. Functional GABA(A) receptors are present postsynaptically on the muscle and presynaptically on terminals and release boutons of the excitatory axon, whereas presynaptic GABA(A) autoreceptors have not been reported on terminals or release boutons of the inhibitory axon. Using antibodies raised against mammalian GABA(A) receptor subunits alpha3 and beta(2/3), we obtained highly specific staining of the presynaptic membrane of the inhibitory bouton and of the postsynaptic membrane of the muscle. Using pre- and postembedding techniques, staining was localized to only presynaptic and postsynaptic membranes of synaptic active zones. We also found extrasynaptic receptor subunit immunoreactivity near (up to 100 nm) to the active zones. Staining with antibodies for the alpha3 and beta(2/3) subunits showed colocalization of particles of the two subunits. We suggest that presynaptic inhibitory boutons of the crayfish possess GABA(A)-like autoreceptors composed of at least the alpha3 and beta(2/3) subunits.

The cellular and subcellular distribution of four GABA(A) receptor subtypes, identified by the presence of the alpha1, alpha2, alpha3, or alpha5 subunit, was investigated immunocytochemically in dissociated cultures of hippocampal neurons. We addressed the questions whether (1) cell-type specific expression, (2) axonal/somatodendritic targeting, and (3) synaptic/extrasynaptic clustering of GABA(A) receptor subtypes was retained in vitro. For comparison, the in vivo distribution pattern was assessed in sections from adult rat brain. The differential expression of GABA(A) receptor subunits allowed to identify five morphologically distinct cell types in culture: the alpha1 subunit was observed in glutamic acid decarboxylase-positive interneurons, the alpha2 and alpha5 subunits marked pyramidal-like cells, and the alpha3 subunit labeled three additional cell types, including presumptive hilar cells. All subunits were found in the somatodendritic compartment. In addition, appropriate axonal targeting was evidenced by the intense alpha2, and sometimes alpha3 subunit labeling of axon-initial segments (AIS) of pyramidal cells and hilar cells, respectively. Accordingly, both receptor subtypes were targeted to AIS in vivo, as well. Synaptic receptors were identified by colocalization with gephyrin, a postsynaptic clustering protein, and apposition to presynaptic terminals labeled with synapsin I. In vitro and in vivo, alpha1- and alpha2-receptor subtypes formed numerous synaptic clusters, alpha3-GABA(A) receptors were located either synaptically or extrasynaptically depending on the cell type, whereas alpha5-GABA(A) receptors were extrasynaptic. We conclude that receptor targeting to broad subcellular locations does not require specific GABAergic innervation patterns, which are disturbed in vitro, but depends on protein-protein interactions in the postsynaptic cell that are both subunit- and neuron-specific.

There is clinical and experimental evidence that monoamine neurons respond to lesions with a wide range of compensatory adaptations aimed at preserving their functional integrity. Neurotoxin-induced lesions are followed by increased synthesis and release of transmitter from residual monoamine fibers and by axonal sprouting. However, the fate of lesioned neurons after long survival periods remains largely unknown. Whether regenerative sprouting may contribute significantly to recovery of function following lesions which induce cell loss has been questioned. We have previously analyzed the response of locus coeruleus (LC) neurons to systemic administration of the noradrenergic (NE) neurotoxin N-(2-chloroethyl)-N-ethyl-2-bromobenzylamine (DSP-4) to adult rats. This drug causes ablation of nearly all LC axon terminals within 2 weeks after administration, followed by a profound loss of LC cell bodies 6 months later. The present study was conducted to determine the fate of surviving LC neurons and to characterize their potential for regenerative sprouting during a 16 month period after DSP-4 treatment. The time-course and extent of LC neuron degeneration were analyzed quantitatively in Nissl-stained sections, and the regenerative response of residual neurons was characterized by dopamine-β-hydroxylase immunohistochemistry. The results document that LC neurons degenerate gradually after DSP-4 treatment, cell loss reaching on average 57% after 1 year. LC neurons which survive the lesion exhibit a vigorous regenerative response, even in those animals in which cell loss exceeds 60–70%. This regenerative process leads progressively to restoration of the NE innervation pattern in the forebrain, with some regions becoming markedly hyperinnervated. In stark contrast to the forebrain, very little reinnervation takes place in the brainstem, cerebellum and spinal cord. These findings suggest that regenerative sprouting of residual neurons is an important compensatory mechanism by which the LC may regain much of its functional integrity in the presence of extensive cell loss. Furthermore, regeneration of LC axons after DSP-4 treatment is region-specific, suggesting that the pattern of reinnervation is controlled by target areas. Elucidation of the factors underlying recovery of LC neurons after DSP-4 treatment may provide insights into the compensatory mechanisms of central neurons after injury and in disease states. © 1992 Wiley-Liss, Inc.

Dendrites of multipolar relay neurons in the lateral geniculate nucleus of the marmoset (Callithrix jacchus), at various ages from birth to adulthood, were studied in rapid Golgi preparations. The dendrites were analyzed by means of three-dimensional computer reconstructions and decomposed into intermediate and terminal segments, both of which were further classified according to their centrifugal order. Measurements were made of the number of segments per dendrite, the total length of dendrites, and the mean length of intermediate and terminal segments. In adult marmosets, there are four stem dendrites on average per neuron, and each dendrite divides into a mean of 14 segments. Between birth and 6 weeks of age, the mean dendritic length doubles, mainly because of changes in terminal segments. There is a significant decrease in dendritic length into adulthood. The total number of stem dendrites does not change after birth, but during the first postnatal week dendrites lose distal segments, after which there is a significant increase in the number of segments of orders 3 to 7. The mean length of intermediate segments does not change with age, nor with order, whereas the length of terminal segments increases from 50 to 120 microns from birth to 6 weeks of age, and then decreases to the adult value of 80 microns. In conclusion, during the period of most rapid visual development, important morphological changes occur in geniculate relay-cell dendrites, involving essentially terminal segments. These observations correlate well with changes of geniculate volume and neuronal density.

The distribution and morphology of serotoninergic neurons in the marmoset (New-World monkey) brainstem were studied by immunocytochemistry and computer-assisted threevdimensional reconstruction. The cytoarchitectonic localization of serotoninergic neurons was ascertained in series of adjacent immunostained and Nissl-stained sections, and the extent and shape of the serotoninergic nuclei were visualized by computer reconstruction. The overall distribution of the immunoreactive neurons is comparable to that already described for several species of primates. The serotoninergic nuclei are spatially well segregated into an anterior and a posterior group. The anterior group, in the mesencephalon and the rostral pons, contains the largest population of serotoninergic neurons. These neurons are not confined to the raphe nuclei near the midline, but rather expand laterally in the reticular formation. This expanded distribution of the neurons in the anterior group results in a partial fusion of the nuclei. In some nuclei, particularly the median raphe, subdivisions can be clearly delineated on the basis of the distinct morphology of the neurons and of their clustering. The neurons of the posterior group, in the caudal pons and the medulla, are almost all contained within the limits of the raphe nuclei. The serotoninergic neurons located in the reticular formation form a lateral column, which is clearly separated from the serotoninergic neurons found near the midline. Immunoreactive axons are distributed throughout the brainstem, but they innervate certain motor and sensory nuclei more densely. It was consistently found in newborn animals that the overall immunoreactive axonal network was richer than in juveniles or adults, suggesting that there may be a major modification in the function of the serotoninergic system around birth.