Low Doses of Bisphenol A Promote Human Seminoma Cell Proliferation by Activating PKA and PKG via a Membrane G-Protein–Coupled Estrogen Receptor

1Endocrine-disrupting chemicals (EDCs) are hormone-like agents present in the environment that may alter the endocrine system of wildlife and humans. In particular, xenoestrogens have been hypothesized to be involved in developmental, reproductive, and malignant diseases by mimicking the natural hormone 17β-estradiol (E2) and interfering with endogenous endocrine regulation at specific periods, such as during fetal growth. Several organochloride pesticides—polychlorinated biphenyls (PCBs), phthalates, and bisphenol A (BPA)—used in the chemical industry have been considered as estrogenic EDCs. However, all of these EDCs have a very weak affinity for binding through the classical nuclear estrogen receptors (ERs), 1,000–2,000 times lower than that of E2. Studies of the nuclear transcriptional regulatory activities of nonphysiologic estrogens have mostly been unable to explain the actions of these chemicals in mediating endocrine disruption in animals and humans at the low picomolar or nano molar concentrations widespread in the environment. In the last few years, EDCs have been reported to act through hormone-independent mechanisms or through a nongenomic activation of membrane-initiated signaling pathways via membrane forms of ERs. Indeed, there is now convincing evidence that estrogens, in addition to the classical regulation of estrogen-responsive genes via nuclear ERs, are able to trigger rapid membrane activation of a variety of second-messenger–mediated signal transduction pathways, with possible implications for cell proliferation, apoptosis, or survival. However, the nature of these membrane ER(s), their relation to the classical ERs, and the precise signaling pathways that are activated remain to be elucidated. Moreover, fetal exposure to xenoestrogens is believed to be involved in male reproductive and developmental patho genesis. Diethylestylbestrol (DES), a potent synthetic estrogen used as an antiabortive drug in the 1970s, has a well-known deleterious effect in adults exposed in utero. DES can produce different developmental or carcinogenic effects in rodents when given during specific developmental windows, including fetal or perinatal periods, such as cryptorchidism or breast, prostate, or endometrial cancers. However, although indirect epidemiologic data show a constant increase of testicular cancer in young men and an increased relative risk via professional exposure to persistent organic pollutants, no experimental model has validated the possible carcinogenic role of exposure to xenoestrogens in developing a testicular germ cell cancer.

2BPA, initially produced like DES as a synthetic estrogen, has been rapidly and widely used as a cross-linking chemical in the manufacture of polycarbonate plastic and epoxy resins. Because of incomplete polymerization and degradation of the polymers by exposure to higher than usual temperatures, BPA leaches out from food and beverage containers, as well as from dental sealants. BPA is found in the serum, milk, saliva, and urine of humans at nanomolar concentrations. Remarkably, BPA has been measured in amniotic fluid at concentrations 5-fold higher than those measured in maternal plasma. Fetal and perinatal exposures to BPA in rodents have been shown to affect the brain, mammary gland, and reproductive tract, including hormone-dependent cancer. Although BPA induces an estrogenic effect through classical nuclear ERs at high concentrations and with a reduced affinity relative to E2, it also is able to trigger a non genomic effect in pancreatic islet, endothelial, and hypophysial cells and in breast cancer cells by initiating rapid responses at low concentrations.

3We recently reported that E2 coupled to bovine serum albumin (E2-BSA) stimulated the proliferation of human seminoma cells (JKT-1) in vitro through a G-protein–coupled nonclassical membrane ER (GPCR). In the present study, we investigated the hypothesis that BPA could stimulate seminoma cell proliferation through such a nongenomic action. We observed a promoting effect of BPA on seminoma cells through a rapid activation of cAMP-dependent protein kinase (PKA) and cGMP-dependent protein kinase (PKG) signaling pathways via a GPCR, illustrating that xenoestrogens, suspected to act as deleterious factors in breast and prostate cancers, could also act in this nongenomic pathway as possible promoting agents in testicular germ cell cancer.

4JKT-1, a human testicular pure seminoma cell line developed from the testis of a 40-year-old man, expresses placental alkaline phosphatase (PLAP), a classical seminoma marker and more recently described markers. Specific embryonic stem cell markers for the JKT-1 cell line have been described previously. JKT-1 cells were maintained in Dulbecco’s modified Eagle medium (DMEM; Gibco BRL, Cergy Pontoise, France) supplemented with penicillin/streptomycin (1%), sodium pyruvate (2%), and 10% fetal bovine serum (FBS) in a humidified 5% CO2 atmosphere at 37°C. Cells were seeded in six-well plates (0.6 × 106 cells/well). After 48 hr, the cells were washed and estrogen-starved overnight in phenol red–free DMEM (Sigma, Lyon, France) supplemented with 1% charcoal-stripped FBS. We then added E2 (Sigma), freshly prepared E2-BSA (Sigma) devoid of free E2 removed by filtration, ICI 182780 (ICI; Falsodex; Astra-Zeneca, Birmingham, UK), BPA, DES, dichlorodiphenyltrichloroethane (DDT), DMSO (Sigma), or ethanol (as vehicle control) at different concentrations (picomolar to micromolar) according to previous studies and at environmentally relevant BPA concentrations and incubated them for 24 hr.

5We harvested cells using trypsin and counted them using Vi-CELL software (Beckman Coulter, Margency, France). Results are expressed as the percent of variation compared with control.

6We grew JKT-1 cells in 10-cm dishes at a density of 4.9 × 106 cells per dish. After 48 hr, the cells were washed with phosphate-buffered saline (PBS) and incubated overnight in phenol red–free DMEM/0.1% BSA; cells were then exposed to ligands for different time periods. We purchased forskolin from Sigma and PD 98059 (PD), H89, KT5823, NF449, and pertussis toxin (PTX) from VWR Calbiochem (San Diego, CA, USA). After washing cells with PBS, we lysed cell pellets in ice-cold lysis Brij buffer/NP-40 [50 mM Tris-HCl (pH 7.5), 1% NP-40, 1% Brij 96 (Fluka, St. Quentin Fallavier, France), 1 mM Na3VO4, 10 mM β-glycerophosphate, 10 mM NaF, 2 mM EDTA, and protease inhibitors (Complete; Roche Diagnostics, Meylan, France)]. Lysates were sonicated 7 sec on ice twice and then centrifuged for 15 min at 14,000 rpm. Equal amounts of whole protein extract were resolved on a 9% SDS-polyacrylamide gel. Proteins were transferred to a polyvinyl difluoride Immobilon-P membrane (Millipore, Saint Quentin en Yvelines, France) and probed with the antibodies against phosphorylated cAMP response-element–binding protein (phospho-CREB; Cell Signaling, Boston, MA, USA) and phosphorylated retinoblastoma protein (phospho-Rb; BD Pharmingen, Le Pont de Claix, France). After the blots were stripped, we verified equal loading of proteins by reprobing the same blots with anti-actin antibody (Cell Signaling).

7Results of cell count or densitometric analysis are expressed as percentages of variation compared with control. We used a nonparametric Mann–Whitney test for statistical analysis.

8BPA stimulated JKT-1 cell proliferation at very low doses. The dose–response curve had an inversed U-shape, with a weak stimulation at 10−6 M that was still present at 10−12 M, and a maximum effect around 10−7/10−9 M (Figure 1A). This stimulation was reproduced by E2-BSA, an impermeable E2 conjugate (Figure 1B), but not by E2, which triggered a significant decrease of cell proliferation at a physiologic intratesticular concentration of 10−9 M (Figure 1B), in agreement with our previous study.

9DDT, an organochloride pesticide with antiandrogenic activity, had no effect on JKT-1 cell proliferation (Figure 1B). DES, a potent synthetic estrogen, acted as a suppressor in a dose-dependent manner, as did E2 (Figure 2). When BPA was combined with either E2 or DES to verify synergic or antagonist effect, BPA neutralized the suppressive effect of E2 (Figure 2) and completely prevented the DES suppressive effect at the dose tested (10−9 M). However, E2-BSA plus BPA induced the same effect of each alone, indicating a lack of synergic or antagonistic effect and likely a similarity in the activated pathways (Figure 2).

10Activation of PKA was necessary for BPA to promote JKT-1 cell proliferation because H89, a specific PKA inhibitor, totally prevented an increase in cell proliferation (Figure 3A). Extracellular stimuli elicit changes in gene expression in target cells by activating intracellular protein kinase cascades that phosphorylate transcription factors within the nucleus. CREB is one of these factors that activates gene transcription after the phosphorylation of serine 133 induced by a variety of protein kinases, including PKA and extracellular-signal-regulated kinase 1/2 (ERK1/2). Using an anti–phospho-CREB antibody that recognizes phosphorylated serine 133, we observed a very rapid (5 min), BPA-induced activation of CREB in JKT-1 cells, with maximum activation at 30 min (Figure 3B). This activation was PKA dependent because H89 completely abolished CREB phosphorylation (data not shown).

11Rb is a nuclear factor that participates in the regulation of the cell cycle, interfering with cyclin action when nonphosphorylated. This suppressive effect is prevented through phosphorylation of Rb. In JKT-1 cells, BPA induced a rapid (4 hr) and intensive phosphorylation of Rb (Figure 4), leading to Rb inactivation during BPA stimulation.

12We recently reported that JKT-1 cells express ER-β but not ER-α. By immunofluorescence, subcellular fractionation, and Western blot, we found that the ER-β receptor had an apparent intra-cytoplasmic localization without any evident membrane location. To determine whether ER-β was involved in the estrogenic or xenoestrogenic activation of JKT-1 cell proliferation, we tested the effect of ICI, a pure ER antagonist. ICI completely counteracted the suppressive effects of E2 and DES on JKT-1 cell proliferation (Figure 5), supporting an ER-β–dependent mechanism. However, ICI did not prevent the promoting effect of BPA or of E2-BSA, supporting our hypothesis that the rapid effect induced by these two ligands was likely not dependent on a classical ER (Figure 5).

13GPCRs have been proposed to be involved in triggering membrane action of steroids, including estrogens. PKA activation is usually a Gαs protein-dependent mechanism, so we tested the effect of NF449, a Gαs inhibitor (Figure 6). NF449 blocked the promoting effect of BPA, illustrating the Gαs dependence of the PKA activation and the G-protein–coupled nature of the receptor involved in BPA stimulation. We have previously shown that E2-BSA, which triggers an effect quite similar to that of BPA, also induces a Gαi-dependent activation of the mitogen-activated protein kinase (MAPK)/ERK1/2 pathway. For this reason, we also studied the effect of PTX, an inhibitor of Gαi/Gαq protein, during BPA-induced JKT-1 proliferation. This toxin prevented the BPA-induced increase of cell proliferation (Figure 7).

14BPA promotion of JKT-1 cells did not seem to involve ERK1/2 activation because PD, an inhibitor of MAPK kinase, did not prevent BPA-enhanced proliferation (Figure 8). We therefore tested the PKG pathway, which is known to be Gαi/Gαq dependent and is involved in BPA activation of calcium influx in pancreatic islet α cells. KT5823, an inhibitor of PKG activation, prevented BPA-induced JKT-1 proliferation (Figure 6).

15These results strongly support the participation of a membrane GPCR involving both the Gαs and Gαi subunits. Figure 9 summarizes the signaling pathways activated during BPA induced JKT-1 proliferation.

16In this article we demonstrate for the first time that very low doses of BPA (picomolar or nanomolar) stimulate human seminoma cell proliferation by allowing a rapid, nongenomic, membrane-initiated activation of PKA and PKG signaling pathways associated with phosphorylation of the transcription factor CREB and the cell cycle regulator Rb. This promoting effect, similar to the one observed with E2-BSA but not with E2 alone, was triggered independently of classical ERs through a membrane receptor belonging to the GPCR family. The low concentrations of BPA able to produce such an effect give this observation environmental relevance and support the hypothesis of a possible contribution of xenoestrogenic fetal exposure to testicular germ cell carcinogenesis.

17Estrogens classically mediate their action after binding to nuclear receptors that act as transcription factors to modulate the activity of target genes by interacting with several DNA response elements. In addition to their ability to mediate gene transcription, estrogens also elicit rapid nontranscriptional effects by membrane-mediated signaling pathways leading to calcium influx, cAMP or nitric oxide production, phospholipase C activation, or inositol phosphate generation. The MAPK/ERK1/2 pathway can also be rapidly activated by estrogens in various cell types, such as endothelial, adipocyte, neuroblastoma, or breast cancer cell lines. Membrane activation of these rapid signaling cascades will then modulate gene transcription. We have recently reported that human seminoma cells express both classical ER-β and a membrane non classical estrogen GPCR. E2 has a high affinity for ER-β and triggers a suppressive effect in JKT-1 cells, whereas E2 coupled to BSA, which prevents membrane crossing, binds to an ncmER and promotes cell proliferation by activating rapid cell signaling, including PKA and MAPK pathways. BPA has a low affinity for ER-β, as described in several models, with a 1,000-fold weaker affinity than E2, and activates pancreatic islet, hypophysial, or endometrial cells through an ncmER. In JKT-1 cells, the differential affinity toward both receptors may explain the dose–response curve observed. At high micromolar concentrations, BPA may trigger a suppressive effect via ER-β as does E2, which neutralizes the nongenomic effect. At low concentration (10−9 M), this genomic effect is absent, allowing the non-genomic effect to be displayed because of the high affinity of BPA for the ncmER. When mixed together at this low concentration, BPA and E2 are mutually antagonistic, whereas DES, also a potent ligand for nuclear ER, at this low concentration only moderately counteracts the nongenomic BPA effect, possibly for conformational reason. Our model illustrates the paradoxical inversed U-shaped curve, explaining effects at very low doses, that has been described for BPA in several models (; vom; ), which could be produced by two different ERs and two different, genomic and non-genomic, mechanisms.

18We propose that the promoting effect occurs through nongenomic transduced activation of the PKA/CREB and the PKG pathways, as illustrated by the very rapid phosphorylation of CREB and the inhibition of both CREB phosphorylation and proliferation obtained with the PKA antagonist H89 and the PKG inhibitor NF449. Phosphorylated CREB will regulate cell-cycle–controlling genes as demonstrated by Rb phosphorylation. Estrogenic activation of the PKA/CREB pathway through an ncmER has already been described in several models. Concerning ERK activation, assessed the rapid changes in intracellular calcium levels induced by xenoestrogens in a pituitary tumor cell line and found that multiple membrane-initiated signaling pathways were activated. The differential patterns presented seemed to depend on the structure of the xenoestrogens and the conformation obtained with the membrane ER. In particular, BPA was one of the EDCs tested that did not induce ERK activation, as in our seminoma cell model. In another model of pancreatic islet α cells, showed that estrogens induce a rapid calcium influx through several pathways, including PKG. We therefore tested the PKG pathway, which involves for its activation a Gαq subunit, and found its contribution in BPA-induced JKT-1 proliferation as well, likely through a rapid intra cytoplasmic calcium increase. Then, we showed that the BPA proliferation-promoting effect on semi-noma cells in fact needed the two subunits Gαs and Gαi, which acted not in an opposite but in a complementary fashion, as already described in other models and supported in JKT-1 cells by the use of specific inhibitors. Indeed, activation of both the PKA and PKG pathways seemed to be necessary for the BPA-induced promoting effect, as we have previously shown for E2-BSA with the two PKA and ERK pathways. BPA, when mixed with E2-BSA, showed the same proliferative effect as each compound alone, without any synergistic or antagonistic effect. Despite the mild difference in the activated protein kinases, it is likely that this promoting effect induced by E2-BSA and BPA is mediated via the same GPCR membrane receptor, an ncmER.

19Increasing evidence from different tissues and cell types has suggested that there are multiple mechanisms through which estradiol can stimulate rapid intracellular signaling. However, one of the main questions remains the nature of the coupled receptors. Different studies of nonclassical estrogen signaling in a variety of target cells, such as endothelial, neuronal, and pituitary cells, have strongly suggested that nuclear classical ER or ER-like proteins are candidates for the membrane ERs. In our model, however, this membrane receptor is unlikely to be a classical ER because ICI failed to inhibit cell proliferation; JKT-1 cell membranes do not express ERβ as we previously reported; and E2 and DES, a potent synthetic estrogen that binds to ER, do trigger not a promoting but a suppressive effect. showed that estrogen-induced ERK activation may occur in human breast cancer cells that do not express either ER-α or ER-β. GPR30, an orphan GPCR, has been proposed as a nonclassical estrogen receptor able to stimulate cancer cells. Thus, GPR30 is a candidate for our ncmER in JKT-1 seminoma cells, activated by E2-BSA or BPA and able to transduce the PKA and ERK or PKG signaling pathways. Moreover, GPR30 has been identified recently in mouse spermatogonia and has also been involved in estrogenic germ cell proliferation control. Seminoma cells are considered to be issued from transformed gonocytes or undifferentiated spermatogonia. Thus, we propose that GPR30 could represent the ncmER in JKT-1 seminoma cells, able to activate PKA, ERK, or PKG pathways. Studies of its expression, precise localization, and involvement in triggering an estrogenic promoting effect in human seminoma cells and human fetal or adult germ cells are now under way in our laboratory.

20It is now possible, as we recently proposed, to more comprehensively describe actions of estrogens at low concentrations on human seminoma cells, involving two different and opposite effects: first, a predominant long-lasting, suppressive effect, antagonized by ICI, thus likely mediated by ER-β, which involves a classical nuclear genomic pathway controlling cell cycle gene expression; and second, a rapid, nongenomic promoting effect triggered by PKA and ERK activation, which involves not a classical ER (not antagonized by ICI or tamoxifen), but rather a PTX-responsive non-classical membrane estrogen GPCR. The resulting impact on germ cells may depend on the relative expression of both receptors (ER-β and GPCR), the endogenous concentration of E2, and the respective binding affinity of the estrogenic compounds. In the presence of both receptors, as in JKT-1 seminoma cells, the suppressive effect of E2 may remain predominant because of a rapid dephosphorylation of ERK1/2 by an ER-β-dependent expression of protein phosphatase 2A, as described in neonatal rat cerebellar neurons. In contrast, in JKT-1 cells, BPA does not need the MAPK pathway, but rather the PKG pathway, which would have a more potent effect on the nongenomic pathway because BPA has a higher affinity for this receptor. Furthermore, human gonocytes, which do not express the active ER-β1 isoform until the prenatal period, may be exclusively sensitive to the membrane-mediated promoting effect if they do express the ncmER as do mouse spermatogonia. Progressive expression of ER-β1 during the perinatal period will prevent proliferation and enhance differentiation into spermatogonia. However, excessive fetal exposition to xenoestrogens with high affinity for the non classical estrogen GPCR, as shown here for BPA, may promote abnormal proliferation of gonocytes through the nongenomic pathway and thus contribute to malignant germ cell transformation, leading, as proposed by, to carcinoma in situ and then to testicular germ cell cancer, the most frequent cancer of the young men with an increasing incidence.

21This GPCR-mediated nongenomic action therefore represents a new basis for evaluating xenoestrogens such as BPA that could interfere with the developmental programming of fetal germ cell proliferation and/or differentiation when it crosses the placenta. Testing of EDCs will require a cell model that expresses this receptor alone and/or together with nuclear ERs.