Molecular and Physiological Evidence for Functional-Aminobutyric Acid ( GABA )-C Receptors in Growth Hormone-secreting Cells *

The neurotransmitter -aminobutyric acid (GABA), released by hypothalamic neurons as well as by growth hormone(GH) and adrenocorticotropin-producing cells, is a regulator of pituitary endocrine functions. Different classes of GABA receptors may be involved. In this study, we report that GH cells, isolated by laser microdissection from rat pituitary slices, possess the GABA-C receptor subunit 2. We also demonstrate that in the GH adenoma cell line, GH3, GABA-C receptor subunits are not only expressed but also form functional channels. GABA-induced Cl currents were recorded using the whole cell patch clamp technique; these currents were insensitive to bicuculline (a GABA-A antagonist) but could be induced by the GABA-C agonist cis-4-aminocrotonic acid. In contrast to typical GABA-C mediated currents in neurons, they quickly desensitized. Ca i recordings were also performed on GH3 cells. The application of either GABA or cis-4-aminocrotonic acid led to Ca transients of similar amplitude, indicating that the activation of GABA-C receptors in GH3 cells may cause membrane depolarization, opening of voltage-gated Ca channels, and a subsequent Ca influx. Our results point at a role for GABA in pituitary GH cells and disclose an additional pathway to the one known via GABA-B receptors.

␥-Aminobutyric acid (GABA) 1 is widely distributed in the central nervous system (1). There, generally, it inhibits neuronal firing and contributes to stabilization of the membrane resting potential by acting on GABA-A, -B, and -C receptors. The term "GABA-C" receptor, which refers to bicuculline-and baclofen-insensitive ionotropic GABA receptors formed by subunits, is controversial, and GABA-C receptors may simply be a subset of GABA-A channels (2). GABA-C receptors are located in certain areas of the central nervous system and in the retinas of various species. They form Cl Ϫ channels, assumed to organize in either homo-or heteromers of the different subunits (3). Outside of the central nervous system and retina, the expression of GABA-C receptor subunits was re-ported based on RT-PCR analysis performed in rat peripheral tissues, namely in gonadal endocrine tissues, adrenal gland, placenta, and small intestine (4), and was also found by immunohistochemistry in human neuroendocrine midgut tumor cells (5). In these cells, GABA-C receptors were shown to be functional by studying Ca 2ϩ i levels and hormone release. In addition, functional GABA-C receptors were also observed in pituitary thyroid-stimulating hormone (TSH) cells using electrophysiological techniques (6). GABA, produced by growth hormone-(GH) secreting cells (7), acts as an autocrine regulator of GH levels via GABA-B receptors (8). In the present study, we report that GH cells also express GABA-C receptor subunits, which form functional receptors in a rat GH-producing cell line, GH3.

MATERIALS AND METHODS
Animals-Pituitary glands and brains were obtained from Sprague-Dawley rats bred at the Technische Universität Mü nchen. They were painlessly killed under ether anesthesia, according to institutional animal care guidelines. The tissues were removed and processed as described previously (7,8).
Culture Procedures of Rat GH3 Cells-The culture procedures applied for GH3 cells were described (8). Briefly, the cells were grown in F12-Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% fetal calf serum (PAA Laboratories, Linz, Austria). Since GH3 cells produce GABA, the patch clamp and Ca 2ϩ measurements required regular renewal of the medium. Positive recordings were obtained only up to day 6 of culture, when the cells did not yet form a confluent monolayer.
RNA Isolation and RT-PCR-RNA isolation and RT-PCRs from GH3 cells, laser-dissected GH and TSH cells, or rat tissues were performed as reported (8). Oligonucleotide primers (Table I) were chosen to encompass exon-intron boundaries to detect possible genomic DNA contamination; for amplifying 2 cDNA from the laser-dissected samples, nested primers were required. For all experiments, the nature of the amplified cDNAs was confirmed by direct sequencing using one of the oligonucleotide primers (AGOWA, Berlin, Germany).
Immunocytochemistry-GH3 cells were cultivated on glass coverslips (2 ϫ 10 4 cells/coverslip) for 1 day. They were then fixed and handled as described previously (11). For immunolocalization of GABA-C receptors, a polyclonal antibody produced in rabbit was used (diluted 1:50; courtesy Dr. Enz, Institut fü r Biochemie, Universitä t Erlangen-Nü rnberg, Germany). This antiserum was raised against 1 but was shown to recognize 2 as well (12). Immunoreactivity was visualized using a fluorescein isothiocyanate-labeled secondary goat antirabbit * This work was supported by a grant from Eli Lilly International Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Calcium Measurements-Ca 2ϩ measurements were performed on GH3 cells, up to day 6 of culture and before reaching confluency, as described (13). Briefly, the cells were loaded with Fura-2/AM (2,5 M, dissolved in Me 2 SO) for 30 min at 37°C in a standard external solution consisting of 140 mM NaCl, 2.7 mM KCl, 1.5 mM CaCl 2 , 1 mM MgCl 2 , 6 mM glucose, and 12 mM HEPES (pH 7.3). Fluorescence measurements were performed with the Zeiss Fast Fluorescence Photometry System (MPM-FFP, Zeiss, Oberkochen, Germany). The excitation wavelength was switched, at 400 Hz, between 340 and 380 nm using appropriate interference filters (bandwidth, 10 nm). The emitted light (505-530 nm) was monitored after averaging with a final time resolution of 80 ms. Ca 2ϩ levels are given in the figures as fluorescence ratios obtained from alternating excitation at 340 and 380 nm.
Drug Application-For both patch clamp and Ca 2ϩ measurements, a combination of global and local bath perfusion was installed that generated a continuous fluid stream containing the agent but confining it to a small volume. Fast pressurized perfusion systems equipped with magnetic valves were used for drug application as described previously (13,14). At each experiment, bath solution was applied first to test for mechanical interference by the mere approaching flow of solutions. GABA, cis-4-aminocrotonic acid (CACA), bicuculline, and baclofen (Tocris, Ellisville, MO) were utilized at a concentration of 100 M as reported previously (6).

Expression of GABA-C Receptors in Rat Pituitary GH Cells and in GH3
Cells-GABA-C receptors are present outside the central nervous system and retina, for example in the pituitary gland. There, TSH cells were shown to possess functional GABA-C channels (6). Therefore, we used TSH cells as a reference for our experiments. Rat pituitary sections were immunostained either for GH or for TSH and subsequently submitted to laser microdissection followed by RT-PCR experiments (Fig. 1). GH and TSH immunoreactive cells were harvested (Fig. 1A), RNA was extracted, and nested RT-PCRs for 2 were performed (Fig. 1B). Since additional tissue surrounds the cells of interest after microdissection, control samples in which GH/ TSH immunoreactive cells were destroyed by laser shots before excision were also analyzed. The 2 subunit was detected in GH cells and, as expected, in TSH cells, but not in the controls.
GH3 cells are derived from a rat GH pituitary adenoma (15) and are widely used as a model for the study of pituitary somatotrophs. We examined whether GH3 cells also possess GABA-C receptors. RT-PCR experiments identified 1 and 2 subunit mRNAs in GH3 cells (Fig. 2A). GABA-C receptor protein at GH3 cell membranes was shown by immunocytochem-istry (Fig. 2B). The antiserum, which recognizes both 1 and 2 , showed a staining, which was confined to the plasma membrane of the cells (Fig. 2B, left panel). Omitting the antiserum resulted in a weak homogeneous staining pattern (Fig. 2B, right panel). Further controls were performed by incubating the cells with rabbit normal serum (dilutions ranging from 1:10,000 to 1:40,000). These experiments led to a homogeneous nonspecific staining within the cells (data not shown) and argue for the specificity of the membrane-associated immunoreactivity obtained using the anti-GABA-C receptor antiserum.
Electrophysiology of GABA-C Receptors in GH3 Cells-The whole cell patch clamp technique was applied to single GH3 cells to test the functionality of the putative GABA-C receptors. Application of 100 M GABA induced an inward whole cell current at a holding potential of Ϫ80 mV in about 70% of the cells (n ϭ 13; Fig. 3A). The GABA-induced current was rapidly and almost completely desensitizing during GABA application. GABA was also capable of eliciting the current in the presence of the GABA-A antagonist bicuculline (n ϭ 4; 100 M; Fig. 3C), even when bicuculline was applied about 1 min prior to GABA. There was no significant difference between the maximum amplitudes (paired t test) for 100 M GABA (0.8 Ϯ 0.3 pA/ picofarads; mean Ϯ S.D.) and for 100 M GABA ϩ 100 M bicuculline (0.9 Ϯ 0.8 pA/picofarads). The channel giving rise to the GABA-induced Cl Ϫ current was identified as GABA-C receptor because of its activation by the specific GABA-C receptor agonist CACA (100 M) in all GABA-sensitive cells tested (n ϭ 6; Fig. 3B). The specific peak current activated by GABA or CACA was in the range of 0.5-5.0 pA/picofarads and thereby  comparable with the GABA-C current observed in TSH cells (6). Activation of the current by GABA or CACA was repeatable several times in the same cell after washing with agonist-free bath solution for about 2 min (Fig. 3, A and B). The peak amplitude did not decrease in the course of the experiment in contrast to the rundown of the GABA-C receptor currents reported in TSH cells (6).
The Activation of GABA-C Receptors Provokes Intracellular Ca 2ϩ Transients-We performed fluorimetric measurements of cytosolic Ca 2ϩ concentrations in GH3 cells (Fig. 4). Both GABA and the GABA-C-specific agonist CACA (100 M of each) induced Ca 2ϩ transients of similar amplitude (n ϭ 9 cells). However, the use of baclofen (GABA-B receptors agonist) did not lead to Ca 2ϩ transients (data not shown). The Ca 2ϩ transients recorded in GH3 cells using either GABA or CACA are most probably due to Ca 2ϩ influx through voltage-gated Ca 2ϩ channels (16). Indeed, both KCl-and GABA-induced Ca 2ϩ transients are almost completely blocked by Gd 3ϩ (500 M), a blocker of voltage-dependent Ca 2ϩ channels (preliminary studies, data not shown). DISCUSSION In this study, the expression of GABA-C receptors is reported in endocrine cells, namely in the pituitary GH cells and in their corresponding cell line, GH3. These receptors are functional in GH3 cells: the current induced by GABA or by the typical GABA-C agonist, CACA, is characterized by its insensitivity to the GABA-A antagonist bicuculline and by its quick desensitization. The opening of GABA-C channels may induce a membrane depolarization as suggested by an increase in Ca 2ϩ i levels measured by fluorometry on GH3 cells.
In the pituitary gland, all types of GABA receptors are expressed (7,8,17), and GABA, produced either by the hypothalamus or by the pituitary itself (7,18,19), is known to be involved in the regulation of hormone levels (8,20,21). Interestingly GABA-C receptors were demonstrated previously in pituitary TSH cells (6). Therefore, these cells served as positive controls in our study. We detected the 2 subunit in GH and, as expected, in TSH cells. Boue-Grabot et al. (6) showed that TSH cells possess functional GABA-C channels, reported the absence of 1 in follicle-stimulating hormone, adrenocorticotropin, and prolactin cells, but did not investigate GH cells. By identifying 2 in TSH and GH cells, we confirm and extend their results. , and brain were reversetranscribed and used for PCR using specific primers for GABA-C 1 -and 2 -subunits. Brain samples served as positive controls; PCRs performed without template were negative (Co.). Sequencing of the PCR products confirmed their identity. B, the use of an antiserum recognizing both 1and 2 -subunits allowed localization of GABA-C receptors at GH3 cell membranes as seen by immunofluorescence microscopy. Immunoreactivity was detected in most cells; the arrow points to membrane-associated staining (left panel). In the control shown, the primary antiserum was omitted (right panel). The scale bar is equivalent to 10 m. We verified by RT-PCR and immunocytochemistry that GH3 cells also express GABA-C receptors. Then, performing patch clamp recordings on single GH3 cells, we could prove functional GABA-C receptors. The application of either GABA or CACA (100 M of each) induced a Cl Ϫ current insensitive to bicuculline, exhibiting an untypical fast desensitization but lacking rundown. Similar inactivating GABA-C currents were observed in rat TSH cells (6) but also in bipolar cells of the carp retina (22). Usually, GABA-C receptor activation is regarded to entail sustained Cl Ϫ currents (23)(24)(25) in contrast to the typically desensitizing GABA-A receptor (26 -28). Most likely, variations in the subunit composition (29), species-dependent protein sequence differences, or/and tissue-specific splicing variants (6) account for these different properties. In addition, we reported previously the presence of the GABA-A receptor subunit ␥ 2 in rat GH cells (7). The possibility that heterooligomerization might occur among and GABA-A subunits is under debate (30,31) and may also explain various electrogenic profiles.
In this report, we show that pituitary GH cells express GABA-C receptors and that they are functional in a corresponding tumor cell line, GH3. What could be the physiological role of GABA-C receptors in GH-secreting cells? A gut neuroendocrine tumor cell line STC-1 was also shown to possess functional GABA-C receptors (5,6). In STC-1 cells, GABA increases Ca 2ϩ i levels by acting on GABA-C receptors. It is also known that the activation of GABA-B receptors influences Ca 2ϩ i concentrations in neurons (32). Since GH3 cells bear functional GABA-B receptors as well (8), one can suppose that GABA may control Ca 2ϩ i levels via both metabotropic and ionotropic mechanisms. To test this hypothesis, we performed fluorimetric measurements of cytosolic Ca 2ϩ concentrations in GH3 cells. We showed that both GABA and CACA (100 M of each) induced Ca 2ϩ transients of similar amplitude. However, preliminary experiments using baclofen (a GABA-B receptor agonist) did not result in Ca 2ϩ transients. These observations argue for a control of Ca 2ϩ i levels by GABA via ionotropic routes. A possible mechanistic explanation for the Ca 2ϩ transients can be deduced from results obtained in developing neurons. There, GABA acts as a trophic substance. Via GABA-A channels, GABA can depolarize the cell membrane when the Cl Ϫ reversal potential is positive to the resting membrane potential (33)(34)(35). Thus, it can provoke Ca 2ϩ transients via the activation of voltage-operated Ca 2ϩ channels (5,36). The Ca 2ϩ transients recorded in GH3 cells, using either GABA or CACA, are most probably due to Ca 2ϩ influx through voltage-gated Ca 2ϩ channels (16). Indeed, preliminary results showed that both KCland GABA-induced Ca 2ϩ transients are almost completely blocked by Gd 3ϩ (500 M), a blocker of voltage-dependent Ca 2ϩ channels. We propose that GABA action on GABA-C receptors leads to membrane depolarization and Ca 2ϩ influx in GH3 cells, suggesting an excitatory function of GABA in endocrine cells.
In summary, the presence of multiple GABA receptor subtypes in GH cells, with different properties and GABA sensitivities, suggests that GABA may have several functions in these cells (e.g. regulation of chloride conductance, membrane potential, control of Ca 2ϩ i concentrations, modulation of hormone secretion). Our present results are in accordance with recent reports indicating new roles of neurotransmitters, including GABA, in various nonneuronal tissues. This is best illustrated in the pancreas, where the neurotransmitter gluta-mate is secreted by ␣-cells and triggers the Ca 2ϩ -dependent exocytosis of GABA from ␤-cells. Once secreted, GABA in turn binds to GABA-A receptors on ␣-cells, where it acts as a paracrine inhibitor for glucagon secretion (37). Thus, signaling molecules originally thought to be restricted to the central nervous system appear to be produced and active in important endocrine systems, namely the anterior pituitary and the pancreas.