Function coupling of otoferlin with GAD65 acts to modulate GABAergic activity.

Otoferlin, an integral membrane protein implicated in a late stage of exocytosis, has been reported to play a critical role in hearing although the underlying mechanisms remain elusive. However, its widespread tissue distribution infers a more ubiquitous role in synaptic vesicle trafficking. Glutamate, an excitatory neurotransmitter, is converted to its inhibitory counterpart, γ-aminobutyric acid (GABA), by L-glutamic acid decarboxylase (GAD), which exists in soluble (GAD67) and membrane-bound (GAD65) forms. For the first time, we have revealed a close association between otoferlin and GAD65 in both HEK293 and neuronal cells, including SH-SY5Y neuroblastoma and primary rat hippocampus cells, showing a direct interaction between GAD65 and otoferlin's C2 domains. In primary rat hippocampus cells, otoferlin and GAD65 co-localized in a punctate pattern within the cell body, as well as in the axon along the path of vesicular traffic. Significantly, GABA is virtually abolished in otoferlin-knockdown neuronal cells whereas otoferlin overexpression markedly increases endogenous GABA. GABA attenuation in otoferlin-knockdown primary cells is correlated with diminished L-type calcium current. This previously unknown and close correlation demonstrates that otoferlin, through GAD65, modulates GABAergic activity. The discovery of otoferlin-GAD65 functional coupling provides a new avenue for understanding the molecular mechanism by which otoferlin functions in neurological pathways.

Menin regulates spinal glutamate-GABA balance through GAD65 contributing to neuropathic pain.

BACKGROUND: Our previous work found that tumor suppressor menin potentiates spinal synaptic plasticity in the context of peripheral nerve injury-induced neuropathic hypersensitivity, but the underlying molecular mechanisms are not clear. We hereby assessed the role of menin in regulating the spinal balance between glutamate and GABA and its contribution to the pathological condition of nerve injury-induced hypersensitivity. METHODS: In spared nerve injury induced C57BL/6 mice, mechanical withdrawal threshold was measured with von Frey filaments after intrathecal administration of small interfering RNA (siRNA) of MEN1 or/and subcutaneous histone deacetylase (HDAC) inhibitors to control the level of glutamic acid decarboxylase 65 (GAD65). Immunoblotting and high-performance liquid chromatography were used to detect the level of protein expression and spinal glutamate and GABA, respectively. RESULTS: Genetic knockdown of spinal menin alleviated nerve injury evoked mechanical hypersensitivity, which was strongly associated with the alteration of the spinal level of GAD65 that resulted in an imbalance of glutamate/GABA ratio from the baseline ratio of 5.8 ± 0.9 (×10(-4)) to the peak value of 58.6 ± 11.8 (×10(-4)) at the day 14 after SNI (p < 0.001), which was reversed by MEN1 siRNA to 14.7 ± 2.1 (×10(-4)) at the day 14 after nerve injury (p < 0.01). In further, selective inhibitors of HDACs considerably reversed the ratio of spinal glutamate and GABA, and also alleviated the mechanical withdrawal threshold markedly. CONCLUSION: Our findings provide mechanistic insight into the contribution of the upregulated spinal menin to peripheral nerve injury induced neuropathic hypersensitivity by regulating glutamate-GABA balance through deactivating GAD65.

Glutamate acid decarboxylase 1 promotes metastasis of human oral cancer by ß-catenin translocation and MMP7 activation.

BACKGROUND: Glutamate decarboxylase 1 (GAD1), a rate-limiting enzyme in the production of γ-aminobutyric acid (GABA), is found in the GABAergic neurons of the central nervous system. Little is known about the relevance of GAD1 to oral squamous cell carcinoma (OSCC). We investigated the expression status of GAD1 and its functional mechanisms in OSCCs. METHODS: We evaluated GAD1 mRNA and protein expressions in OSCC-derived cells using real-time quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) and immunoblotting analyses. To assess the critical functions of GAD1, i.e., cellular proliferation, invasiveness, and migration, OSCC-derived cells were treated with the shRNA and specific GAD1 inhibitor, 3-mercaptopropionic acid (3-MPA). GAD1 expression in 80 patients with primary OSCCs was analyzed and compared to the clinicopathological behaviors of OSCC. RESULTS: qRT-PCR and immunoblotting analyses detected frequent up-regulation of GAD1 in OSCC-derived cells compared to human normal oral keratinocytes. Suppression of nuclear localization of ß-catenin and MMP7 secretion was observed in GAD1 knockdown and 3-MPA-treated cells. We also found low cellular invasiveness and migratory abilities in GAD1 knockdown and 3-MPA-treated cells. In the clinical samples, GAD1 expression in the primary OSCCs was significantly (P < 0.05) higher than in normal counterparts and was correlated significantly (P < 0.05) with regional lymph node metastasis. CONCLUSIONS: Our data showed that up-regulation of GAD1 was a characteristic event in OSCCs and that GAD1 was correlated with cellular invasiveness and migration by regulating ß-catenin translocation and MMP7 activation. GAD1 might play an important role in controlling tumoral invasiveness and metastasis in oral cancer.

Adenovector GAD65 gene delivery into the rat trigeminal ganglion produces orofacial analgesia.

BACKGROUND: Our goal is to use gene therapy to alleviate pain by targeting glial cells. In an animal model of facial pain we tested the effect of transfecting the glutamic acid decarboxylase (GAD) gene into satellite glial cells (SGCs) of the trigeminal ganglion by using a serotype 5 adenovector with high tropisms for glial cells. We postulated that GABA produced from the expression of GAD would reduce pain behavior by acting on GABA receptors on neurons within the ganglion. RESULTS: Injection of adenoviral vectors (AdGAD65) directly into the trigeminal ganglion leads to sustained expression of the GAD65 isoform over the 4 weeks observation period. Immunohistochemical analysis showed that adenovirus-mediated GAD65 expression and GABA synthesis were mainly in SGCs. GABAA and GABAB receptors were both seen in sensory neurons, yet only GABAA receptors decorated the neuronal surface. GABA receptors were not found on SGCs. Six days after injection of AdGAD65 into the trigeminal ganglion, there was a statistically significant decrease of pain behavior in the orofacial formalin test, a model of inflammatory pain. Rats injected with control virus (AdGFP or AdLacZ) had no reduction in their pain behavior. AdGAD65-dependent analgesia was blocked by bicuculline, a selective GABAA receptor antagonist, but not by CGP46381, a selective GABAB receptor antagonist. CONCLUSION: Transfection of glial cells in the trigeminal ganglion with the GAD gene blocks pain behavior by acting on GABAA receptors on neuronal perikarya.

Temporal migration of gonadotrophin-releasing hormone-1 neurones is modified in GAD67 knockout mice.

Gonadotrophin-releasing hormone (GnRH-1) neurones reside in the forebrain and regulate gonadal function via the hypothalamic-pituitary-gonadal axis. Disruption of this axis results in reproductive dysfunction. During embryonic development, GnRH-1 neurones migrate from the nasal pit through the nasal/forebrain junction (NFJ) into the developing brain. Prenatally gamma-aminobutyric acid (GABA) is excitatory and has been shown to play a role in nervous system development. Both in vivo and in vitro experiments suggest that GABA inhibits migration of GnRH-1 neurones. The present study examines the migration of GnRH-1 neurones in GAD67 knockout (KO) mice to further elucidate the role of GABA on GnRH-1 neuronal development. Three stages were examined, embryonic day (E)12.5, E14.5 and E17.5. GnRH-1 cell number and location were analysed by immunocytochemistry and in situ hybridisation histochemistry. The total number of GnRH-1 immunopositive cells was similar between wild-type (WT) and KO mice. However, significant differences were found in the overall distribution of GnRH-1 immunopositive cells in GAD67 KO compared to WT mice at all stages. Subsequent analysis by area revealed differences occurred at the NFJ with an increase in GnRH-1 cells in GAD67 KO at E14.5 and a decrease in GnRH-1 cells in GAD67 KO at E17.5. Comparable counts for cells expressing GnRH-1 transcript and protein were obtained. These data indicate that attenuated levels of GABA accelerate GnRH-1 cell migration in nasal areas as well as movement of GnRH-1 cells into the central nervous system at the NFJ.

Distribution of Fos immunoreactivity in the rat brain after freezing or escape elicited by inhibition of glutamic acid decarboxylase or antagonism of GABA-A receptors in the inferior colliculus.

It has been shown that electrical stimulation of the central nucleus of the inferior colliculus (IC) at freezing or escape thresholds activates different neural circuits in the brain. Since electrical stimulation activates cell bodies and fibers of passage it is necessary to use chemical stimulation that activates only post-synaptic receptors. To examine this issue in more detail, we took advantage of the fact that GABAergic neurons exert tonic control over the neural substrates of aversion in the IC. Reduction of GABA transmission in this structure was performed with the use of semicarbazide - an inhibitor of the GABA synthesizing enzyme glutamic acid decarboxylase (GAD) - and the GABA-A receptor antagonist bicuculline. Depending on the dose employed local infusions of semicarbazide (6.0 microg/0.2 microl) or bicuculline (40 ng/0.2 microl) into this region caused freezing and escape, respectively. The results obtained showed that freezing behavior induced by semicarbazide was associated with an increase in Fos expression in the dorsomedial column of the PAG (dmPAG) only, while bicuculline-induced escape was related to widespread increase in Fos labeling, notably in the periaqueductal gray, hypothalamus nuclei, amygdaloid nuclei, the laterodorsal nucleus of thalamus (LD), the cuneiform nucleus (CnF) and the locus coeruleus (LC). Thus, the present data support the notion that freezing and escape behaviors induced by GABA blockade in the IC are neurally segregated: acquisition of aversive information of acoustic nature from the IC probably uses the dmPAG column as a relay station to higher brain centers whereas bicuculline-induced escape activates structures involved in both sensory processing and motor output of defensive behavior. These results support the existence of distinct neural circuits mediating the sensory and motor responses of the defense reaction. The extent of the brain activation during freezing appears to be limited to the anatomical connections of the dmPAG, whereas an overall activation of the limbic system predominates during escape behavior induced by IC stimulation.

Structural and functional analysis of cysteine residues in human glutamate decarboxylase 65 (GAD65) and GAD67.

Previously, we have shown that brain glutamate decarboxylase (GAD) is greatly inhibited by sulfhydryl reactive reagent suggesting cysteine residue(s) may play an important role in GAD function. In this report, we determined the role of cysteine residues in the recombinant human 65-kDa GAD isoform (hGAD65) and 67-kDa GAD isoform (hGAD67), using a combination of matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry and site-directed mutagenesis. Here, we report that cysteine 446 (C446) in hGAD65 is important for its activity and is present as free sulfhydryl group. This conclusion is based on the following observations: (i) mutation of C446 in hGAD65 to alanine reduced hGAD65 activity by more than 90%, (ii) MALDI-TOF analysis of the non-reduced, trypsin-digested GAD65 revealed that C446 is present as a free sulfhydryl group as indicated by a peak at m/z (mass/charge) 647.3446 (peptide 443-448) and, when GAD65 was treated with sulfhydryl reagent, N-ethylmaleimide (NEM), the peak is shifted to m/z 772.3702,a mass increase of 125.1 daltons (Da) as a result of modification of cysteine by NEM. Parallel studies have also been conducted with hGAD67. Cysteine 455 was found to be important for GAD67 activity.

Stiff-person syndrome.

BACKGROUND: Stiff-person syndrome (SPS) is a rare neurologic disorder with autoimmune features. It is characterized by progressive, severe muscle rigidity or stiffness most prominently affecting the spine and lower extremities. REVIEW SUMMARY: Superimposed muscle spasms result in simultaneous contraction of agonist and antagonist muscles which are detectable by electromyography (EMG) and relieved by administration of benzodiazepines. The exacerbation of SPS by emotional stressors often results in the referral of these patients for psychiatric assessment although this was more common before the discovery of an association with antibodies to glutamic acid decarboxylase (GAD antibodies). Formerly known as stiff-man syndrome, the female to male ratio is 2:1 and the principle paraneoplastic variant is associated with breast cancer. Although rare, this is a disease of middle age that severely curtails the functional capacity of those it strikes. It is frequently associated with diabetes and other autoimmune diseases. IVIg is recently demonstrated to be effective in the treatment of SPS; diazepam remains useful in managing the symptoms. CONCLUSIONS: This article summarizes the history of SPS, describes important clinical features, discusses management, touches upon areas of uncertainty, and postulates some avenues for research.

Enhancement of neuronal protection from oxidative stress by glutamic acid decarboxylase delivery with a defective herpes simplex virus vector.

We have developed defective herpes simplex virus 1 (HSV-1) vectors, based on amplicon plasmids with a replication-deficient mutant, as helper for the transfer of the glutamic acid decarboxylase (GAD67) or beta-galactosidase (beta-gal) gene as control directed by HCMV promoter into neuronal-like cells (PC12) and primary neurons. GAD67 protein was detected immunochemically, while GAD67 activity in virus-producing and nonproducing cell lines was detected enzymatically or by GABA release. Infection with GAD67-expressing amplicon vectors enhanced the resistance of PC12 cells to H(2)O(2). This protection was related to increased energy metabolism, as shown by MTT reduction and ATP level, and involved the GABA shunt, as shown by the reduction in ATP level seen in the presence of gamma-vinyl GABA (GVG), a specific GABA transaminase inhibitor. Level of glutathione (GSH), which requires ATP for its synthesis, was increased by the GAD67 transgene. The activity of glucose-6-phosphate dehydrogenase involved in the maintenance of the NADPH that can be used for the regeneration of the GSH pool, was increased by infection with amplicon vectors. Thus, replication-deficient HSV-1 and the GAD67 transgene have complementary neuroprotective effects and infection with GAD67-expressing amplicon vectors was able to protect nondifferentiated cortical neurons from glutamate toxicity mediated by oxidative stress. Such defective GAD67-expressing HSV-1, as neurotropic vector, should be helpful in neurodegenerative diseases implicating alterations of energy metabolism and oxidative stress in neuronal cells expressing GABA transaminase.

Differential regulation of GAD67, enkephalin and dynorphin mRNAs by chronic-intermittent L-dopa and A2A receptor blockade plus L-dopa in dopamine-denervated rats.

Adenosine A2A receptor antagonists have been proposed as an effective therapy in the treatment of Parkinson's disease. In the present study, we compared the modifications on striatal glutamate decarboxylase (GAD67), enkephalin, and dynorphin mRNA levels produced by a chronic-intermittent administration of L-3,4-dihydroxyphenyl-alanine (L-dopa) (6 mg/kg) with those produced by the adenosine A2A receptor antagonist SCH 58261 (5 mg/kg) plus L-dopa (3 mg/kg) in unilaterally 6-hydroxydopamine (6-OHDA)-lesioned rats. As previously reported, L-dopa (6 mg/kg) and SCH 58261 (5 mg/kg) plus L-dopa (3 mg/kg) produced the same degree of turning behavior after the first administration. However, while L-dopa (6 mg/kg) induced a sensitized turning behavior response during the course of the treatment, which indicated a dyskinetic potential, SCH 58261 (5 mg/kg) plus L-dopa (3 mg/kg) produced a stable turning behavior response, which was predictive of absence of dyskinetic side effects. Unilateral 6-OHDA lesion produced an elevation in striatal GAD67 and enkephalin mRNA levels and to a decrease in dynorphin mRNA levels. Chronic-intermittent L-dopa (6 mg/kg) treatment increased the striatal levels of GAD67, dynorphin, and enkephalin mRNA in the lesioned side as compared to the vehicle treatment. Chronic-intermittent SCH 58261 (5 mg/kg) plus L-dopa (3 mg/kg) as well as L-dopa (3 mg/kg) or SCH 58261 (5 mg/kg) alone did not produce any significant modification in GAD67, dynorphin, or enkephalin mRNA levels in the lesioned striatum as compared to the striatum of vehicle-treated rats. The results show that combined SCH 58261 plus L-dopa did not produce long-term changes in markers of striatal efferent neurons activity and suggest that the lack of modifications in GAD67 and dynorphin mRNA after SCH 58261 plus L-dopa might correlate with the lack of turning behavior sensitization which predicts drug dyskinetic potential.

Enhanced neuronal protection from oxidative stress by coculture with glutamic acid decarboxylase-expressing astrocytes.

Astrocytes expressing glutamic acid decarboxylase GAD67 directed by the glial fibrillary acidic protein promoter were shown to provide enhanced protection of PC12 cells from H(2)O(2) treatment and serum deprivation in the presence of glutamate. In addition, they protected non-differentiated, but not differentiated, embryonic rat cortical neurons from glutamate toxicity. Glutamic acid decarboxylase (GAD)-expressing astrocytes showed increased glutathione synthesis and release compared to control astrocytes. These changes were due to GAD transgene expression, as transient expression of a GAD antisense plasmid resulted in partial suppression of the increase in glutathione release. In addition to the previously demonstrated increases in NADH and ATP levels and lactate release, GAD-expressing astrocytes show increased antioxidant activity, explaining their ability to protect neurons from various injuries.

Cellular and molecular roles of beta cell autoantigens, macrophages and T cells in the pathogenesis of autoimmune diabetes.

Type I diabetes, also known as insulin-dependent diabetes mellitus (IDDM) results from the destruction of insulin-producing pancreatic beta cells by a progressive beta cell-specific autoimmune process. The pathogenesis of autoimmune IDDM has been extensively studied for the past two decades using animal models such as the non-obese diabetic (NOD) mouse and the BioBreeding (BB) rat. However, the initial events that trigger the immune responses leading to the selective destruction of the beta cells are poorly understood. It is thought that beta cell autoantigens are involved in the triggering of beta cell-specific autoimmunity. Among a dozen putative beta cell autoantigens, glutamic acid decarboxylase (GAD) has been proposed as perhaps the strongest candidate in both humans and the NOD mouse. In the NOD mouse, GAD, as compared with other beta cell autoantigens, provokes the earliest T cell proliferative response. The suppression of GAD expression in the beta cells results in the prevention of autoimmune diabetes in NOD mice. In addition, the major populations of cells infiltrating the islets during the early stage of insulitis in BB rats and NOD mice are macrophages and dendritic cells. The inactivation of macrophages in NOD mice results in the prevention of T cell mediated autoimmune diabetes. Macrophages are primary contributors to the creation of the immune environment conducive to the development and activation of beta cell-specific Th1-type CD4+ T cells and CD8+ cytotoxic T cells that cause autoimmune diabetes in NOD mice. CD4+ and CD8+ T cells are both believed to be important for the destruction of beta cells. These cells, as final effectors, can kill the insulin-producing beta cells by the induction of apoptosis. In addition, CD8+ cytotoxic T cells release granzyme and cytolysin (perforin), which are also toxic to beta cells. In this way, macrophages, CD4+ T cells and CD8+ T cells act synergistically to kill the beta cells in conjunction with beta cell autoantigens and MHC class I and class II antigens, resulting in the onset of autoimmune type I diabetes.

Effect of injection of antisense oligodeoxynucleotides of GAD isozymes into rat ventromedial hypothalamus on food intake and locomotor activity.

In the ventromedial hypothalamus (VMH), gamma-aminobutyric acid (GABA) plays a role in regulating feeding and running behaviors. The GABA synthetic enzyme, glutamic acid decarboxylase (GAD), consists of two isozymes, GAD65 and GAD67. In the present study, the phosphorothioated antisense oligodeoxynucleotides (ODNs) of each GAD isozyme were injected bilaterally into the VMH of male rats, and food intake, body weight and locomotor activity were monitored. ODNs were incorporated in the water-absorbent polymer (WAP, 0.2 nmol/microliter) so that ODNs were retained at the injection site. Each antisense ODN of GAD65 or GAD67 tended to reduce food intake on day 1 (day of injection=day 0) though not significantly. An injection combining both antisense ODNs significantly decreased food intake only on day 1, but body weight remained significantly lower than the control for 5 days. This suppression of body weight gain could be attributed to a significant increase in locomotor activity between days 3 and 5. Individual treatment with either ODNs did not change locomotor activity. The increase in daily locomotor activity in the group receiving the combined antisense ODNs occurred mainly during the light phase. Neither vehicle (WAP) nor control ODN affected food intake, body weight and locomotor activity. Histological studies indicated that antisense ODN distributed within 800 micron from the edge of the area where WAP was located 24 h after the injection gradually disappeared within days, but still remained within 300 micron m distance even 7 days after the injection. Antisense ODN was effectively incorporated by all the cell types examined, i.e., neurons, astrocytes and microglias. Further, HPLC analysis revealed that antisense ODNs of GAD isozymes, either alone or combined, decreased the content of GABA by 50% in VMH 24 h after the injection. These results indicate that suppression of GABA synthesis by either of the GAD isozymes is synergistically involved in suppressing food intake and enhancing locomotor activity in rat VMH.

A role for glutamate decarboxylase during tomato ripening: the characterisation of a cDNA encoding a putative glutamate decarboxylase with a calmodulin-binding site.

A tomato fruit cDNA library was differentially screened to identify mRNAs present at higher levels in fruit of the tomato ripening mutant rin (ripening inhibitor). Complete sequencing of a unique clone ERT D1 revealed an open reading frame with homology to several glutamate decarboxylases. The deduced polypeptide sequence has 80% overall amino acid sequence similarity to a Petunia hybrida glutamate decarboxylase (petGAD) which carries a calmodulin-binding site at its carboxyl terminus and ERT D1 appears to have a similar domain. ERT D1 mRNA levels peaked at the first visible sign of fruit colour change during normal tomato ripening and then declined, whereas in fruit of the ripening impaired mutant, rin, accumulation of this mRNA continued until at least 14 days after the onset of ripening. This mRNA was present at much lower levels in other tissues, such as leaves, roots and stem, and was not increased by wounding. Possible roles for GAD, and its product gamma-aminobutyric acid (GABA) in fruit, are discussed.

GABA production in rat islets of Langerhans.

Homogenates of pancreatic islets catalyzed breakdown of L-glutamate to GABA with a rate of 0.24 +/- 0.04 nmol.min-1 x mg-1 protein at 37 degrees C. The formation of GABA was stimulated by addition of pyridoxal phosphate in the range 0.05-1 microM (0.97 +/- 0.02 nmol.min-1 x mg protein-1 at a saturating cofactor concentration), which indicates that the process was catalyzed by glutamic acid decarboxylase. The half-maximal effect was obtained with 0.1 microM PLP. Kinetic analyses of the results showed that the Vmax and Km for the reaction were 1.12 nmol.min-1 x mg protein-1 and 0.66 mM, respectively. The pH optimum was 7.0. Subcellular fractionation revealed that 51% of GAD activity was present in the cytosol, 17% in microsomes, 9% in secretory granules, 5% in mitochondria, and 11% in cell debris. Comparison of the kinetic properties of the cytosolic and microsomal forms of the enzyme showed that their Km for glutamate was the same, but that the cytosolic GAD had a lower Km for PLP. GABA synthesis in the nominal absence of PLP was enhanced by malate (twofold increase at 5 mM) and citrate (threefold increase at 5 mM), but was unaffected by ATP and chloride. However, if the islet homogenate was prepared and incubated in the presence of PLP, neither malate nor citrate influenced enzyme activity. Aspartate and AOA were powerful inhibitors of glutamate breakdown. Freshly isolated islets contained approximately 4 mM GABA, whereas the concentration was < 0.1 mM in whole pancreas.(ABSTRACT TRUNCATED AT 250 WORDS)

Basal forebrain neurons and memory: a biochemical, histological, and behavioral study of differential vulnerability to ibotenate and quisqualate.

The differential vulnerability of basal forebrain cells to ibotenate (IBO) or quisqualate (QUIS) was investigated in rats. IBO was also coinjected with cystine (CYS) or zinc (Zn). Cortical choline acetyltransferase (ChAT) and glutamate decarboxylase (GAD) activity, neurotensin receptors, and high-affinity choline uptake sites were quantified in conjunction with radioimmunoassays for neurotensin, substance P, and somatostatin; immunocytochemistry for neurotensin-, somatostatin-, Leu-enkephalin-, and ChAT-positive cells; and in situ hybridization histochemistry of somatostatin, substance P, and enkephalin mRNAs. Compared with the performance of controls, continuous alternation performance in a T maze of IBO+Zn or IBO+CYS rats was better than that of IBO rats, whereas the performance of QUIS rats was unimpaired. Of those neurotransmitter systems examined, only ChAT-immunoreactive cells were vulnerable to IBO or QUIS. However, cholinergic cell loss did not correlate with impaired performance.

Gamma amino butyric acid (GABA) controls anterior pituitary hormone secretion.

Intraventricular injection (3V) of GABA can lead to a dose-related increase in plasma LH in ovariectomized (OVX) and OVX, estrogen-primed animals. Since the effects were blocked by the GABA receptor antagonist, bicucculine (B), they appear to be specific. These alterations in plasma LH were accompanied by alterations in hypothalamic LHRH, dopamine (DA) and norepinephrine (NE) content which suggests roles for all three compounds in the genesis of the increase in plasma LH. Since the DA receptor blocker pimozide (P) failed to block the elevation in LH induced by GABA it appears that the effect of GABA on LH release is mediated by NE. Others have found that the GABA agonist, muscimol, can lower plasma LH under certain conditions. Consequently, it appears likely that there may be opposite actions of GABA on LHRH release depending on the site of action within the hypothalamus. Intravenous (iv) injection of B in OVX rats produced an initial fall in plasma LH followed by a prolonged rise which again suggests several sites of action of GABA on LH release. GABA had no effect on FSH release consistent with separate control of this hormone. 3V injection of various doses of GABA produced a dose-related lowering of plasma TSH in OVX rats which appeared to be mediated by the dopaminergic system since P abolished the TSH-lowering of GABA. Following iv injection of B in normal male rats, there was a dramatic decline in TSH suggesting that under these conditions GABA would have a stimulatory action. Similar results were seen in OVX rats. The results indicate an important stimulatory action of GABA on TSH release in both males and OVX females. Perhaps the discrepancy between the results with B and GABA can be explained again by multiple sites of action of GABA of opposite sign. 3V injection of GABA induced a dose-related stimulation of growth hormone (GH) secretion; however, more recent evidence from other laboratories suggests that under certain conditions GABA has an inhibitory role in GH secretion. Again, we speculate that GABA had dual sites of action of opposite sign to affect GH. In contrast to the effects of GABA on these pituitary hormones, it appears to have a direct inhibitory effect on prolactin (Prl) secretion via the lactotrophs. Both stimulatory and inhibitory actions of GABA have been found following its injection into the brain. Studies with iv B also support dual actions on Prl release.