Stimulation of TM3 Leydig cell proliferation via GABAA receptors: A new role for testicular GABA

The neurotransmitter gamma-aminobutyric acid (GABA) and subtypes of GABA receptors were recently identified in adult testes. Since adult Leydig cells possess both the GABA biosynthetic enzyme glutamate decarboxylase (GAD), as well as GABAA and GABAB receptors, it is possible that GABA may act as auto-/paracrine molecule to regulate Leydig cell function. The present study was aimed to examine effects of GABA, which may include trophic action. This assumption is based on reports pinpointing GABA as regulator of proliferation and differentiation of developing neurons via GABAA receptors. Assuming such a role for the developing testis, we studied whether GABA synthesis and GABA receptors are already present in the postnatal testis, where fetal Leydig cells and, to a much greater extend, cells of the adult Leydig cell lineage proliferate. Immunohistochemistry, RT-PCR, Western blotting and a radioactive enzymatic GAD assay evidenced that fetal Leydig cells of five-six days old rats possess active GAD protein, and that both fetal Leydig cells and cells of the adult Leydig cell lineage possess GABAA receptor subunits. TM3 cells, a proliferating mouse Leydig cell line, which we showed to possess GABAA receptor subunits by RT-PCR, served to study effects of GABA on proliferation. Using a colorimetric proliferation assay and Western Blotting for proliferating cell nuclear antigen (PCNA) we demonstrated that GABA or the GABAA agonist isoguvacine significantly increased TM3 cell number and PCNA content in TM3 cells. These effects were blocked by the GABAA antagonist bicuculline, implying a role for GABAA receptors. In conclusion, GABA increases proliferation of TM3 Leydig cells via GABAA receptor activation and proliferating Leydig cells in the postnatal rodent testis bear a GABAergic system. Thus testicular GABA may play an as yet unrecognized role in the development of Leydig cells during the differentiation of the testicular interstitial compartment.


Background
Gamma-aminobutyric acid (GABA) is the most important inhibitory neurotransmitter in the vertebrate central nervous system. In addition to its well established function as neurotransmitter, locally synthesized GABA and GABA receptors are also present in endocrine organs, for example, in somatotrophs (GH-producing cells) of the anterior pituitary lobe [1][2][3] and in pancreatic islet cells [4][5][6]. In both endocrine tissues GABA is regulating the synthesis and the release of hormones. The release of glucagon and growth hormone was shown to be controlled by GABA in an auto-/paracrine manner.
In our previous work we recently identified another GABAergic system located in adult Leydig cells in rodent and human testis [7]. Since Leydig cells possess both isoforms of the GABA synthesizing enzyme glutamate decarboxylase, GAD65 and GAD67, vesicular GABA transporter (VGAT), as well as several GABA A and GABA B receptor subunits, GABA may act as an auto-/paracrine molecule regulating Leydig cell function. Some evidence for a role in release of testosterone came from pharmacological studies in rat Leydig cells, which respond to GABAergic stimulation with increased testosterone production [8,9]. What other roles GABA may have in endocrine Leydig cells and which GABA receptors are mediating these actions are not known.
In the central nervous system evidence for a non-synaptic, trophic role of GABA in neurogenesis during embryonic development is mounting [10][11][12][13][14][15]. Thus GABA stimulates progenitor cells to proliferate in different regions of the developing brain [16][17][18][19][20]. Since these neuronal progenitor cells are also capable of synthesizing GABA and possess GABA receptors, GABA executes this trophic function in an auto-/paracrine fashion [21]. Further non-synaptic actions of GABA in the developing brain that are evolving include regulation of migration and motility of embryonic neurons [22][23][24]. While in general, cellular responses to GABA are mediated through GABA A , GABA B and GABA C receptors and the intracellular signaling pathways associated with them [25], in respect to both cell proliferation and migration in the developing brain, contribution of GABA A receptors was reported [18,19,21,26,27]. Thus, although its precise regulation may depend on the region and cell type affected, GABA emerges as an important signal for cell proliferation and migration.
In view of this role of GABA in the brain, the question arises, whether GABA may influence cell proliferation processes in the testis, for example in Leydig cells, which bear GABA receptors [7]. In the testis of adult mammals, however, Leydig cells have only a marginal turnover rate and show low mitotic activity [28][29][30]. Due to the fact that Leydig cells in postnatal testis proliferate to a much greater extend than in adult testis [31][32][33][34][35], we sought to study postnatal testes of mice and rats at age of five-six days after birth. At this point of development two distinct types of Leydig cells are found, namely steroidogenic fetal Leydig cells with a typical rounded morphology clustered together in groups and spindle-shaped mesenchymal precursor cells of adult Leydig cells, which are located primarily in peritubular regions. The latter are not able to synthesize steroids, but are strongly proliferating and differentiate to Leydig progenitor cells during the second postnatal week in rodents [36][37][38]. Fetal Leydig cells may also increase in number during postnatal development, albeit to a smaller degree [31,39]. Thus, the endocrine compartment of postnatal testis bears developing and highly proliferating cells of the adult Leydig cell lineage. Therefore, in this study we addressed the questions whether a local GABAergic system is present in postnatal testis and may be involved in proliferation of Leydig cells.

Animals
Testes and other tissues were obtained from adult, 3-6 months old (n = 12; Sprague-Dawley, Wistar) and five-six days old male rats (n = 14; Sprague-Dawley), as well as from adult (n = 4; BALB/c) male mice, which were bred at the Technische Universität München, Germany. Testes were also obtained from five-six days old male mice (n = 9; BALB/c), which were bred at the Instituto de Biología y Medicina Experimental, Buenos Aires, Argentina. According to the National Institute of Health Guide for the Care and Use of Laboratory Animals, they were painlessly killed under ether anesthesia by exsanguinations and organs were rapidly removed. Testes were either frozen until isolation of mRNA and preparation for GAD activity measurements, or fixed in Bouin's solution overnight at 4°C and then embedded in paraffin.
Cell proliferation studies TM3 cells (5 × 10 3 cells per well) were plated on 96-well plates (Nunc, Wiesbaden, Germany) and incubated for 24 h in the presence or absence of GABA, isoguvacine, baclofen, bicuculline and phaclofen. One experiment included 32 replicate wells per treatment. As previously described [45,46], cell proliferation was determined by using the CellTiter 96 AQ ueous One Solution cell proliferation assay (Promega, Mannheim, Germany). The specificity and sensitivity of this method was previously evaluated in our lab by comparison with a [ 3 H]thymidine incorporation assay [45].

RNA preparation and RT-PCR
Isolation of RNA from rodent testes, as well as RT and PCR for GAD65/67, VGAT and GABA receptor subunits were performed as described previously [47]. Conditions of PCR amplification consisted of 30 or 35 cycles (94°C for 30 s, 55°C for 30 s, 72°C for 60 s, followed by final extension for 5 min at 72°C). Oligonucleotide primers, as specified in Table 1, were synthesized according to published sequences. Verification of cDNAs was achieved by direct sequencing [47].

Immunocytochemistry
TM3 cells were cultivated on glass cover slips (2 × 10 4 cells per cover slip) for 1 day. They were then fixed and handled as previously described [49]. For immunolocalization an antiserum recognizing GAD65/67 and an antiserum recognizing VGAT was carried out overnight at 4°C (diluted 1:1000 in 0.02 M potassium phosphate buffered saline containing 2% goat non-immune serum, pH 7.4). Immunoreactivity was visualized using a secondary polyclonal goat anti-rabbit antiserum (diluted 1:200; Dianova, Hamburg, Germany) labeled with fluorescein isothiocyanate (FITC). For control purposes either the specific antiserum was omitted or incubations with rabbit non-immune serum (dilution 1:10.000) were carried out instead. Sections were examined with a Axiovert microscope (Zeiss, Oberkochen, Germany), equipped with a FITC filter set.

GAD activity measurements
Determination of GAD activity by measuring the production of radiolabeled carbon dioxide (CO 2 ) from 14 Cglutamate was performed as described previously [7,53].
In brief, TM3 cells, AtT20 cells and rat tissue samples were homogenized in 60 mM potassium phosphate buffer (pH = 7.1), containing 0.5% Triton X-100, 1 mM 2-aminoethyl-isothiouronium bromide and 1 mM phenylmethanesulphonyl fluoride (Sigma, Deisenhofen, Germany), centrifuged, and the supernatants were used in the assay. The assay was performed in a total reaction volume of 60 µl, containing 20 µl of sample and 0.1 mM EDTA, 0.5% Triton X-100, 0.1 mM DTT, 0.05 mM pyridoxal phosphate, 9 mM L-glutamate, 3.3 µCi/ml 14 C-glutamate (Biotrend, Köln, Germany, specific activity: 50-60 mCi/ mmol) and 60 mM potassium phosphate buffer. The reaction mix was incubated for 1 h at 37°C and then stopped by adding 100 µl of 10% trichloracetic acid per vial. The released CO 2 was absorbed on benzethonium hydroxidedrenched filter disks, and bound radioactivity was determined using a Tri-Carb 2100 liquid scintillation counter (Packard, Meriden, USA). The values obtained were normalized to protein content measured by DC protein assay (Bio-Rad GmbH, München, Germany) described above. Rat tissue samples that were heated to 95°C for 5 min served as negative controls.

Statistical analyses
Statistic analyses were performed using GraphPad Prism 3.02 (GraphPad Software, San Diego, USA). The results obtained in cell proliferation and GAD assay experiments were compared using one-way-analysis of variance ANOVA followed by Newmann-Keuls test. The results received in Western blot experiments were compared using one-way-analysis of variance ANOVA followed by Dunnett's test. Data shown are expressed as means+SEMs (standard error of the mean).

A GABAergic system is present in postnatal rat testis: Active GAD, VGAT and GABA A -α1 in postnatal rat Leydig cells
In postnatal rat testis immunohistochemical studies revealed that components of a local GABAergic system are present in interstitial cells (Figure 1). GAD65/67 and VGAT proteins were localized to the cytoplasm of interstitial cells. Because of their rounded morphology and clustered appearance these interstitial cells represent fetal Leydig cells. Protein of GABA A -α1 was immunolocalized not only to clustered fetal Leydig cells, but also to spindleshaped interstitial cells. When antisera against GABA B -R1 and GABA B -R2 were employed, specific immunoreactive signals were absent. All control panels probed with buffer alone or non-immune rabbit serum were negative.

Proliferation marker PCNA is localized in interstitial cells of postnatal rodent testis
To identify proliferating cells in adult and postnatal rat testes we probed testicular tissue sections of rats ( Figure 3) and mice (data not shown) with an antiserum against the proliferation marker PCNA. Specific immunoreactions were observed inside the seminiferous tubules on germ cells and in Sertoli cells, as well as in the cells of the interstitial compartment of all samples examined. In contrast, in adult rat testes we only occasionally found interstitial cells to be immunopositive for PCNA. In controls performed with buffer alone or buffer containing nonimmune mouse serum, respectively, specific immunoreactivity was absent.

The GABAergic system is also present in postnatal mouse testis: GAD67, VGAT and several GABA A receptor subunits in postnatal mouse testis
RT-PCR studies identified mRNAs of VGAT and GAD67, but not of GAD65 (data not shown). Furthermore, mRNAs of the GABA A receptor subunits α2, β1, β2, β3 and γ3 were readily detected ( Table 2). The mRNAs of the GABA A receptor subunits α1, α3, γ1, γ2 and of the GABA B receptor subunits R1 and R2 were not found in postnatal mouse testis in several RT-PCR experiments. Immunohistochemical experiments (using GAD and VGAT antisera) yielded results similar to the ones obtained in rat (data not shown). Figure 2 GAD is active in postnatal rat testis. GAD activity [cpm/µg protein] was measured in testicular tissue samples of 5-6 days old rats (D5-6, n = 10) and adult rats (Adult, n = 11). Tissue samples of adult rat testes (Co1, n = 6) and adult rat cerebella (Co2, n = 6) heated to 95°C for 5 min served as negative controls. GAD activity in adult testes is higher than GAD activity in postnatal testes. Columns with different superscripts are significantly (ANOVA/Newmann-Keuls, p < 0.001) different from each other and represent means+SEMs.

GAD is active in postnatal rat testis
PCNA in postnatal and adult rat testis

TM3 Leydig cells possess active GAD67, VGAT and GABA A
receptor subunits α1, α2 β1, β3 and γ1 To examine whether mouse-derived TM3 cells may serve as model for proliferating Leydig cells, we first determined whether TM3 cells are able to produce GABA and possess GABA receptors. Specific immunocytochemical staining against GAD65/67 and VGAT protein was observed in TM3 cells (Figure 4). Specific immunoreactivity was absent in controls performed with buffer alone or buffer containing non-immune rabbit serum, respectively. RT-PCR and Western blot experiments confirmed these results and revealed GAD67, but not GAD65, in TM3 cells ( Figure 5). GAD67 protein in TM3 cells was found to be enzymatically active ( Figure 6) with an assayed GAD activity of 3.25 ± 0.24 cpm/µg protein (n = 11). TM3 cells (n = 6) and rat testicular tissue (n = 6), both heated to 95°C for 5 min, served as negative controls. GAD activity of TM3 cells was significantly (ANOVA/Newmann-Keuls, p < 0.001) higher than GAD activity (0.46 ± 0.05 cpm/µg protein) of AtT20 cells (n = 9), another mouse cell line. GAD activities of AtT20 cells and of negative controls (inactivated by boiling) were not significantly different from each other (ANOVA/Newmann-Keuls, p > 0.05).
Furthermore, mRNAs of the GABA A receptor subunits α1, α2, β1, β3 and γ1 were readily detected in TM3 cells (Table   2). In contrast the mRNAs of the GABA A receptor subunits α3, β2, γ2 and γ3, as well as the mRNAs of the GABA B receptor subunits R1 and R2 were not found in TM3 cells in several RT-PCR experiments.

GABA and GABA A agonist isoguvacine increase cellular content of PCNA in TM3 cells
Levels of the proliferation marker PCNA was determined using Western blotting after stimulation with GABA or GABA receptor agonists/antagonists in TM3 cells ( Figure  7). TM3 cells were incubated 0, 5, 10, 15, 30 min with GABA, GABA+bicuculline, isoguvacine, isoguva-cine+bicuculline, baclofen and baclofen+phaclofen, respectively, and PCNA content was semiquantitatively determined by Western blotting. Signals were normalized and thus corrected for minor loading differences with the help of the results obtained for β-Actin (n = 5 experiments per treatment). Stimulation with GABA or GABA a agonist isoguvacine lasting for 15 min significantly (ANOVA/ Dunnett's, p < 0.001) increased PCNA content in TM3 cells compared to untreated samples (0 min stimulation). This effect was blocked by co-incubation with GABA a antagonist bicuculline. Stimulation with GABA B agonists or antagonists (baclofen, and baclofen+phaclofen) did not alter PCNA content in TM3 cells (data not shown).

GABA induced TM3 cell proliferation is mediated by GABA A receptor
In order to investigate whether GABA A receptor activation is not only able to increase PCNA content in TM3 cells, but indeed can stimulate TM3 cell proliferation, we performed proliferation assays (Figure 8). TM3 cells were incubated for 24 h with GABA (n = 22), isoguvacine (n = 21), baclofen (n = 25), GABA+bicuculline (n = 19) and isoguvacine+bicuculline (n = 13). GABA significantly (ANOVA/Newmann-Keuls, p < 0.05) increased TM3 cell proliferation up to 124.3 ± 4.4% compared to untreated controls (n = 36; 100%). Stimulation with isoguvacine also significantly (ANOVA/Newmann-Keuls, p < 0.05) increased cell proliferation up to 120.7 ± 4.1%, but baclofen treatment did not result in significant alteration in cell proliferation. Further evidence for involvement of GABA A receptors was provided by blocking of the proliferative effects of GABA and isoguvacine by GABA A antagonist bicuculline (ANOVA/Newmann-Keuls, p < 0.05).

Discussion
The present study shows that crucial components of a GABAergic system are present in the endocrine compartment of postnatal rodent testis and that GABA stimulates proliferation of TM3 Leydig cells via GABA A receptors. These results suggest that GABA may regulate cell proliferation of fetal Leydig cells and/or mesenchymal precursors of the adult Leydig cell lineage in an auto-/paracrine manner. We therefore suggest that GABA may contribute to the morphogenesis of the testis.
The RT-PCR results (+) in postnatal mouse testis (D5) and in TM3 cells were confirmed by sequencing; (-) indicates that the expected mRNA was not detected in several RT-PCR experiments.
Previously we and others demonstrated first details of a local GABAergic system in the endocrine compartment of adult rodent and human testis [7,54]. Adult Leydig cells possess enzymatically active GAD, VGAT and several GABA A and GABA B receptor subunits. Both isoforms GAD65 and GAD67 were present in rats and mice [7]. The functional significance of a testicular GABAergic system is not well known, but auto-/paracrine modulation of testosterone production in Leydig cells is a possibility suggested by studies describing stimulating effects of GABA on testosterone production in rats [8,9]. Since hormonal influences clearly govern steroid production of the adult testis, the modulatory effect of GABA on testosterone may however not be the main effect of GABA.
We rather speculated that GABA may exert trophic effects in the testis. This assumption was based on the trophic action of GABA via GABA A receptors in the developing brain. We focused in this study therefore on the postnatal testis, which bears proliferating cells including fetal Leydig cells and cells of the adult Leydig cell lineage.
It is widely accepted that two distinct Leydig cell populations are present during the first postnatal week in mouse and rat testis [33,34,[36][37][38][39], namely steroidogenic fetal Leydig cells and mesenchymal precursors of adult Leydig cells. The first mentioned form conspicuous clusters in the interstitium [31,34,55]. Although fetal Leydig cells Immmunocytochemical evidence for the presence of GAD and VGAT in TM3 cells Figure 4 Immmunocytochemical evidence for the presence of GAD and VGAT in TM3 cells. A cytoplasmatic staining pattern for GAD (glutamate decarboxylase) and VGAT (vesicular GABA transporter) was observed in TM3 cells (A, B). Controls included using non-immune serum (C) and omission of primary antiserum. Bars: 15 µm.

Figure 5
RT-PCR and Western blot results: TM3 cells express GAD67 and VGAT. RT-PCR experiments (A) revealed that TM3 cells possess mRNA for GAD67, but lack GAD65 (data not shown), as well as VGAT. Expected product sizes were 399 bp for GAD67 and 302 bp for VGAT. Western blot analysis (B) revealed the presence of GAD67 in TM3 cells. Protein probes from mouse brain served as positive control and depict immunoreactivity for both GAD isoforms.
represent a differentiated cell population, there is evidence for a moderate increase in the number of these cells during the first two weeks of postnatal development [31,39]. In contrast, non-steroidogenic mesenchymal precursor cells of the adult Leydig cell lineage are located primarily in peritubular regions and proliferate strongly. They differentiate during the second postnatal week into progenitor cells and from the end of the third week on into newly formed, immature and then into fully functional mature adult Leydig cells [33,34,[36][37][38].
Our immunohistochemical findings indeed evidenced dramatic proliferative events in the postnatal testis. Interstitial and peritubular cells in the testis of five days old rats expressed the proliferation marker PCNA, which was used in testicular tissues before [56][57][58]. Among these cells are likely connective tissue cells and endothelial cells [59][60][61], but also fetal Leydig cells and mesenchymal precursors of adult Leydig cells, as judged by their typical location and morphology. Since the latter are undifferentiated in nature [34,[36][37][38], we could not use specific markers to distinguish them from other cell types.
Our study links Leydig cells proliferation and local testicular GABA synthesis. This is based first on the fact that we identified GABA synthesis and GABA A receptors in the postnatal testis of rodents, and second on the proliferative action of GABA and GABA A agonists in TM3 Leydig cells.
We identified only fetal Leydig cells, characterized by their rounded morphology and clustered appearance in the testicular interstitium, to possess GAD67 and VGAT. In contrast to adult testis, GAD65 was not detected. GAD67 was, however, found to be enzymatically active in rat testicular tissue of the same developmental stage. These two results together allow the conclusion that only fetal Leydig cells possess the pivotal molecules to synthesize and store GABA.
The present investigation provides insights into the possible targets of testicular GABA. As evidenced by RT-PCR studies, several GABA A receptor subunits are expressed in the postnatal testis. GABA B receptors were not found in postnatal testis, a result in contrast to our previous study in the adult rodent testis [7]. Immunolocalization of GABA A receptor subunits was hampered, due to availability of suitable antisera, but localization of GABA A -α1 revealed presence on rat fetal Leydig cells, but also on spindle-shaped interstitial cells. At least some of the last mentioned cells are very likely to represent mesenchymal precursor cells of the adult Leydig cell lineage. Thus according to the immunolocalization of GABA A -α1, both fetal Leydig cells and precursors of adult Leydig cells, are possible targets for GABA in the postnatal testis.
The number of fetal Leydig cells increases moderately in rodents during the first two weeks of postnatal development [31,39] and it is possible that GABA may mediate this effect. Another possibility is that GABA may modulate cell proliferation of mesenchymal precursor cells of adult Leydig cells or other GABA A receptor bearing testicular cell types. Based on our results in the present study, it is possible that GABA might even be a start signal leading to proliferation and differentiation of mesenchymal precursors of adult Leydig cells. Interestingly, this signal is as yet unknown [34,37,38]. Thyroid hormone may be involved, but participation of LH and androgens in the initiation of adult Leydig cell development was ruled out [37,39,[62][63][64][65][66].
Clearly, in-vivo evidence for such a crucial role of testicular GABA is as yet missing, but unequivocal evidence for a proliferative action of GABA via GABA A receptors was provided by cell culture experiments using TM3 Leydig cells, which possess GABA A receptor subunits. Involvement of GABA A receptors was suggested by the use of the pharmacologically well defined GABA A agonist isoguvacine and by the use of the GABA A antagonist GAD is active in TM3 cells Figure 6 GAD is active in TM3 cells. GAD activity [cpm/µg protein] was measured in TM3 cells (n = 11 batches of cells) and in AtT20 cells (n = 9 batches of cells), another endocrine mouse cell line used for control purposes. TM3 cells (Co1, n = 6 batches of cells) and testicular samples of adult rats (Co2, n = 6), both heated to 95°C for 5 min, served as additional negative controls. Testicular GAD activity in TM3 cells is higher than GAD activity in AtT20 cells or in controls. Columns with different superscripts are significantly (ANOVA/ Newmann-Keuls, p < 0.001) different from each other and represent means+SEMs.
bicuculline. Interestingly, this signaling pathway is in analogy to studies in the developing brain, where GABA also induces cell proliferation of neuronal progenitors and other neuronal cell types via activation of GABA A receptors [10][11][12][13].
In summary, a GABAergic system exists already in postnatal rodent testis and differs from the one in adult testis, since one of the two GAD isoforms as well as GABA B receptor subunits are missing. Nevertheless, it appears functional and our results suggest that GABA has similar roles in the developing brain and in the developing testis, namely to act as a trophic factor affecting the morphogenesis of crucial cells in these two organs.

Authors' contributions
CG carried out most of the experiments, participated in the study design, performed statistical analyses and drafted the manuscript. RFGD carried out part of RT-PCR, Western blotting and immunocytochemistry. AT and AK PCNA content in TM3 cells is increased by GABA and GABA A agonist isoguvacine Figure 7 PCNA content in TM3 cells is increased by GABA and GABA A agonist isoguvacine. Using Western blot analyses we examined PCNA content of TM3 cells after 0, 5, 10, 15, 30 min stimulation with GABA, GABA+bicuculline, isoguvacine and isoguva-cine+bicuculline, respectively. The PCNA content of TM3 cells after 15 min stimulation with GABA (A) is significantly higher compared to untreated TM3 cells or compared to TM3 cells stimulated for 15 min with GABA+bicuculline. After 15 min stimulation with isoguvacine (B) the PCNA content of TM3 cells is also significantly higher compared to untreated TM3 cells or compared to TM3 cells stimulated for 15 min with isoguvacine+bicuculline. Data represent means+SEMs of n = 5 independent experiments and were normalized to β-Actin levels. Columns with different superscripts are significantly (p < 0.001, ANOVA/ Dunnett's test) different from each other. Figure 7C and 7D depict representative Western blot experiments, respectively.
provided technical assistance. AM conceived of the study, and participated in its design, coordination and writing of the manuscript. All authors read and approved the final manuscript.