The G protein-coupled receptor GPR30 mediates c-fos up-regulation by 17beta-estradiol and phytoestrogens in breast cancer cells.

A growing body of evidence concerning estrogen effects cannot be explained by the classic model of hormone action, which involves the binding to estrogen receptors (ERs) alpha and ERbeta and the interaction of the steroid-receptor complex with specific DNA sequences associated with target genes. Using c-fos proto-oncogene expression as an early molecular sensor of estrogen action in ERalpha-positive MCF7 and ER-negative SKBR3 breast cancer cells, we have discovered that 17beta-estradiol (E2), and the two major phytoestrogens, genistein and quercetin, stimulate c-fos expression through ERalpha as well as through an ER-independent manner via the G protein-coupled receptor homologue GPR30. The c-fos response is repressed in GPR30-expressing SKBR3 cells transfected with an antisense oligonucleotide against GPR30 and reconstituted in GPR30-deficient MDA-MB 231 and BT-20 breast cancer cells transfected with a GPR30 expression vector. GPR30-dependent activation of ERK1/2 by E2 and phytoestrogens occurs via a Gbetagamma-associated pertussis toxin-sensitive pathway that requires both Src-related and EGF receptor tyrosine kinase activities. The ability of E2 and phytoestrogens to regulate the expression of growth-related genes such as c-fos even in the absence of ER has interesting implications for understanding breast cancer progression.

Estradiol (E2) 1 and natural estrogen-like compounds, such as genistein and quercetin, bind to and activate estrogen receptors (ER) ␣ and ␤, which in turn regulate the expression of target genes directly and/or indirectly via protein-protein interactions with other transcription factors (1)(2)(3)(4)(5)(6)(7). Although it is often, but not always, straightforward to link the physiological effects of estrogens to the genomic model of the ligand-receptor complex activity, considerable controversy still exists on the ability of E2 to elicit transcriptional responses independently of ER␣ and ER␤.
A variety of molecules located close to the plasma membrane, such as the G protein-coupled receptors (GPRs) (8 -11), Src, Ras, and Raf (12)(13)(14), have recently been recognized to mediate multiple extracellular stimuli, such as those induced by E2, in a cell context-specific manner (15)(16)(17)(18). For instance, it has been proposed that in breast cancer cells, E2 induces ERK phosphorylation via the orphan GPR, named GPR30 (9). Interestingly, such responses elicited by E2 do not require ER expression and/or localization at the membrane level, as shown on the contrary by other investigations in various cell types (8, 14, 19 -23). In this respect, the potential of E2 to trigger different pathways that integrate cell surface signaling with gene transcription has attracted an increased interest toward the identification of agonistic or antagonistic compounds.
The proto-oncogene c-fos plays a relevant role in the regulation of normal cell growth, differentiation, and cellular transformation processes (24 -27). Besides, c-fos represents a prototypical "immediate early" gene since its expression is rapidly induced by different extracellular stimuli including mitogens and hormones (28 -43). The nuclear protein encoded by c-fos interacts with c-jun family members to form the heterodimeric activating protein-1 transcription factor complex (24 -27). The fos-jun heterodimers binding to activating protein-1 sites located within mammalian gene promoters regulates gene expression in a specific manner depending on cellular and promoter context as well as interacting proteins (44 -49). Moreover, c-fos and other members of the fos family, such as fosB, fra-1, and fra-2, bind to sites identified in the regulatory region of target genes modulating the late response expression of critical factors for cell cycle re-entry, such as cyclin D1 (Ref. 50 and references therein).
The transcription of c-fos is controlled by multiple cis-elements present in the gene promoter: the cAMP-response element that binds to cAMP-response element-binding protein (51), the Sis-inducible enhancer that is recognized by the signal transducers and activators of transcription (STAT) group of transcription factors (52), the serum-response element that mediates c-fos induction by growth factors, and other extracellular stimuli leading to activation of MAPK pathways (53)(54)(55)(56)(57). The serum-response element initially binds a dimer of the serum-response factor, the binding of which recruits the ternary complex factors including Elk-1 and serum-response factor accessory protein 1 and 2 (54). Several studies have shown that ER␣ also activates c-fos expression in breast cancer cells and that the hormone-sensitive site is localized to a 240-bp region (Ϫ1300 to Ϫ1060) (30,32,36,38,39,43,45). Further analysis demonstrated an imperfect palindromic estrogen-re-sponsive element (ERE) within this sequence that is able to bind ER␣ in gel mobility shift assays but is not sufficient for transactivation (58). Thereafter, it was shown that ER␣ must interact with Sp1 at a GC-rich site downstream from the imperfect palindromic ERE (58). In a subsequent study (59), both growth factors and E2 seem to converge on the serum-response element. In the latter case, however, the activation by E2 involved non-genomic signaling of ER␣ through the MAPK signaling pathway and phosphorylation and binding of Elk-1 to the serum-response element.
Herein, we show that early c-fos expression induced by E2 as well as the phytoestrogens genistein and quercetin is mediated at least in part by ER␣ in MCF7 breast cancer cells. However, the same agents are able to stimulate c-fos expression independently of ER expression in ER-negative breast cancer cells via GPR30.
Plasmids-The firefly luciferase reporter plasmids were c-fos and the deletion mutant c-fos⌬ERE (which lacks the ERE sequence) encoding Ϫ2.2-kb and Ϫ1172-bp 5Ј upstream fragments of human c-fos, respectively (gifts from K. Nose, Tokyo, Japan) (60) and Gal4-luc for the Gal4-Elk1 fusion protein. The latter two plasmids were described together with menin in our previous study (61). GPR30 and dominant negative ERK2 (DN/ERK2) expression vectors were kindly provided by R. Weigel (Philadelphia, PA) and M. Cobb (Dallas, TX). The Renilla luciferase expression vector pRL-TK (Promega, Milan, Italy) was used as a transfection standard.
Cell Culture-MCF7, MDA-MB-231, BT-20 human breast cancer cells were a gift from E. Surmacz (Philadelphia, PA), whereas SKBR3 human breast cancer cells were from Picard's laboratory (Genève, Switzerland). MCF7 and MDA-MB-231 cells were maintained in Dulbecco's modified Eagle's medium/F12 without phenol red supplemented with 10% fetal bovine serum (Invitrogen), whereas SKBR3 cells were maintained in RPMI 1640 without phenol red supplemented with 10% fetal bovine serum (Invitrogen). Cells were switched to medium without serum 24 h before transfections, 48 h before RT-PCR or immunoblot, 72 h before the evaluation of ERK1/ERK2 phosphorylation.
Transfections and Luciferase Assays-A total of 100,000 MCF7 cells were plated into 24-well dishes with 500 l of regular growth medium/ well the day before transfection. The medium was replaced with Dulbecco's modified Eagle's medium/F12 lacking phenol red and serum on the day of transfection, which was performed using FuGENE 6 reagent as recommended by the manufacturer (Roche Diagnostics) with a mixture containing 0.5 g of reporter plasmid, 0.1 g of effector plasmid where applicable, and 2 ng of pRLCMV. After 4 h, the medium was replaced again with serum-free Dulbecco's modified Eagle's medium lacking phenol red and containing 1 M of each treatment, and then cells were incubated for 5 h. Luciferase activity was measured with the dual luciferase kit (Promega) according to the manufacturer's recommendations. Firefly luciferase values were normalized to the internal transfection control provided by the Renilla luciferase activity. The normalized relative light unit values obtained from untreated cells were set as 100%, upon which the activity induced by treatments was calculated.
Western Blotting-Cells were grown in 10-cm dishes and exposed to ligands for 2 or 12 h before lysis in 500 l of 50 mM HEPES (pH 7.5), 150 mM NaCl, 1.5 mM MgCl 2 , 1 mM EGTA, 10% glycerol, 1% Triton X-100, 1% SDS, a mixture of protease inhibitors containing 1 mM aprotinin, 20 mM phenylmethylsulfonyl fluoride, and 0.2 M sodium orthovanadate. Protein concentration was determined by Bradford reagent according to the manufacturer's recommendations (Sigma). Equal amounts of whole protein extract were resolved on a 10% SDS-polyacrylamide gel, transferred to a nitrocellulose membrane (Amersham Biosciences), probed overnight at 4°C with the antibodies against c-fos, ␤-actin (Santa Cruz Biotechnology, Santa Cruz, CA), and pERK1/2, ERK2 (Cell Signaling Technology, Inc, Celbio, Milan, Italy) and then revealed using the ECL system (Amersham Biosciences).
Statistical Analysis-Statistical analysis was performed using analysis of variance followed by Newman-Keuls testing to determine differences in means. p Ͻ 0.05 was considered as statistically significant.

E2, Genistein, and Quercetin Transactivate c-fos Promoter
Constructs in MCF7 Cells-In our previous study (4), we demonstrated that an ERE reporter gene transfected in MCF7 cells responds to genistein and quercetin, like E2, through an ER␣mediated mechanism. Thus, in the present study, we first examined whether transiently transfected full-length human c-fos promoter (Ϫ2.2 kb) could respond to these phytoestrogens as reported previously for E2 (67). Fig. 1A demonstrates that genistein and quercetin mirror E2 activity. Activation by all three compounds is sensitive to the MAPK inhibitor PD 98059 and is unaffected by the PI3K inhibitor wortmannin (or LY 294002, not shown). In our previous investigation (61), we showed that the nuclear protein encoded by the tumor suppressor gene MEN1, named menin, inhibits ERK-dependent phosphorylation and activation of the ternary complex factor member Elk-1, which regulates c-fos gene expression. Therefore, we evaluated the action of menin on the transactivation property of estrogenic compounds using the above mentioned c-fos promoter assay and a corresponding reporter gene assay for monitoring the activation of Gal4-Elk-1. As expected, luciferase activity of both constructs was reversed in the presence of menin (Fig. 1, A and B). The transactivation of Elk-1 by E2 and phytoestrogens confirmed its involvement in c-fos expression, whereas the inhibitory effects of PD 98059 indicated that a MAPK-dependent mechanism is required for such a response (Fig. 1B). Interestingly, the transcriptional potential of these compounds was maintained using a mutant of the c-fos promoter lacking the ERE region (Ϫ1172 bp), suggesting that other pathways, not requiring an ER␣-ERE interaction, contribute to c-fos transactivation.
E2, Genistein, and Quercetin Rapidly Induce c-fos mRNA Expression in ER␣-positive MCF7 Cells-It has been largely reported (28 -33,35-43,45, 69 -76) that c-fos is promptly stimulated by E2 and various agents. To evaluate whether phytoestrogens mimic E2 in this regard and in the activation of the two well known target genes pS2 and cathepsin D (77), we performed a semiquantitative RT-PCR comparing the mRNA levels after standardization on the housekeeping gene encoding the ribosomal protein 36B4. A treatment for only 1 h with E2 and phytoestrogens up-regulated c-fos expression in ER␣-positive MCF7 cells (Fig. 2A). The same results were obtained for pS2 and cathepsin D with a 12-h exposure, whereas c-fos returned to levels resembling those of untreated cells (Fig. 2B).
E2, Genistein, and Quercetin Up-regulate c-fos Protein Levels in MCF7 Cells-The aforementioned results prompted us to evaluate whether c-fos protein levels follow the mRNA increase observed after a short (2-h) treatment of ER␣-positive MCF7 cells. A similar exposure with E2, genistein, and quercetin also resulted in a significant c-fos protein induction (Fig. 3A) that was dependent on protein synthesis, as ascertained by pretreating cells with the protein inhibitor cycloheximide (Fig.  3B). To verify the involvement of ER␣ in this process, we used the two ER antagonists ICI 182780 and OHT, which both partially blocked c-fos stimulation (Fig. 3, C and D). c-fos induction was also inhibited by the MAPK inhibitor PD 98059 and by transfection of an expression vector encoding DN/ERK2 (Fig. 3, E and F). Thus, both ER␣ and MAPK pathways may contribute to c-fos protein expression, although these two path-

GPR30-mediated Signals in Breast Cancer Cells 27012
ways cannot be considered as mutually exclusive (see "Discussion"). The inhibitors wortmannin (Fig. 3G) and LY 294002 (data not shown) did not alter the c-fos stimulation observed with the compounds used, suggesting that PI3K is not involved in this process. After a 12-h exposure with E2, genistein, and quercetin, the levels of c-fos were similar to those of untreated cells (Fig. 3H) paralleling the mRNA levels (Fig. 2B).
E2, Genistein, and Quercetin Stimulate c-fos Protein Expression in ER-negative SKBR3 Cells-Are E2 and phytoestrogens able to trigger c-fos stimulation independently of ER in breast cancer cells, and if so, what is the signal transduction pathway involved in such a response? The partial inhibitory effects exerted by ICI 182780 and OHT in ER␣-positive MCF7 cells (Fig. 3, C and D) prompted us to further investigate possible mechanisms regulating c-fos expression in ER-negative SKBR3 cells. Surprisingly, E2 as well as genistein and quercetin induced c-fos protein expression in SKBR3 cells as in MCF7 cells (Fig. 4, A and B). The ER antagonists ICI 182780 and OHT, which by themselves had no effect on c-fos stimulation (data not shown), did not alter the response to these treatments (Fig.  4, C and D), whereas the MAPK inhibitor PD 98059 and DN/ ERK2 (see above) completely reversed the response (Fig. 4, E and F). Next, as observed in MCF7 cells, the PI3K inhibitors wortmannin (Fig. 4G) and LY 294002 (data not shown) did not modify c-fos induction. Considering that a large variety of extracellular ligands signal through GPRs and result in ERK1/2 activation (9, 78, 79), we examined whether GPR mediates c-fos induction by E2, genistein, and quercetin in ER-negative SKBR3 cells. These cells express higher levels of this orphan receptor as compared with other breast cancer cell lines as observed in the present study (data not shown) and a previous study (80). After ligand-receptor interaction, the G␣-subunit protein dissociates from the heterotrimeric G␣␤␥ complex, leading to MAPK activation through a mechanism sensitive to pertussis toxin and tyrosine kinase inhibitors. Interestingly, pretreatment with either pertussis toxin or the Src family tyrosine kinase inhibitor PP2 completely abrogated the ability of the tested compounds to stimulate c-fos expression in SKBR3 cells (Fig. 4, H and I). A previous study (9) reported that GPR30 can promote the tyrosine phosphorylation of the EGF receptor (EGFR), which in turn leads to ERK1/2 activation. It is worth noting that the EGFR kinase inhibitor, tyrphostin AG-1478, blocked c-fos induction by E2 and phytoestrogens (Fig. 4J), confirming an involvement of the EGFR in this pathway. To analyze the ligand specificity of the rapid c-fos response observed in SKBR3 cells, we performed treatments with other steroids, including dexamethasone, progesterone, and R1881. These all failed to induce c-fos expression (Fig. 4K), as did 17␣-estradiol (data not shown). After treatment of cells with E2, genistein, and quercetin for 12 h, the rapid increase of c-fos was no longer noticeable (Fig. 4L).
To provide evidence that c-fos stimulation depends on MAPK in both ER␣-positive MCF7 cells and ER-negative SKBR3 cells, we evaluated ERK1/2 phosphorylation after a very short treatment (5 min) with E2, genistein, and quercetin. Interestingly, all compounds were able to activate ERK1/2 in both breast cancer cell lines without changes in total ERK2 (Fig. 5, A and  B), suggesting that c-fos stimulation can occur via MAPK independently of ER expression.
Furthermore, to investigate the involvement of GPR30 in the up-regulation of c-fos by E2, genistein, and quercetin, we transfected a specific GPR30 antisense oligonucleotide, which abrogated the response to agents used, whereas a control oligonucleotide had no effect (Fig. 6). This result clearly suggested that GPR30 mediates c-fos induction by E2 and phytoestrogens in SKBR3 cells.

E2, Genistein, and Quercetin Up-regulate c-fos in MDA-MB 231 and BT-20 Cells Transfected with GPR30 -Complemen-
tary evidence for a role of GPR30 was provided using MDA-MB 231 and BT-20 breast cancer cells, which do not express GPR30 (9,80). It is worth noting that all compounds acquired the ability to induce c-fos after transfecting cells with an expression plasmid encoding GPR30 but not in the presence of the empty vector (Fig. 7). DISCUSSION The genomic response to the binding of ER␣ with its cognate ligand E2 is a relatively long term effect, which involves an appropriate fit of the hormone in the receptor binding pocket to produce a complex of sufficient stability and lifetime (81). As a consequence, multiple protein-protein and protein-nucleic acid interactions trigger the enzymatic reactions required for the remodeling of the chromatin architecture and changes in the rate of gene transcription (81). Nonetheless, a variety of cell types respond to estrogens in a very short time, making the classical ER-mediated mechanism of action unlikely (82). Thus, the ability of estrogens to interact with a putative membraneassociated ER has been associated with many rapid effects (83), although it is uncertain whether these activities are mediated by proteins similar or distinct from those involved in the well known transcriptional events occurring after E2 stimulation. Furthermore, the potential for signaling cross-talk between the classical ER-mediated mechanism of action and alternative signal transduction pathways triggered by E2 still remains to be explored.
We have demonstrated for the first time to our knowledge that E2 and the two major phytoestrogens genistein and quercetin are able to induce rapid c-fos up-regulation in breast cancer cells by a GPR30-mediated mechanism. In addition, the results obtained in MCF7 cells indicate that ER␣ contributes to this response, showing its potential to regulate the expression of both early and late genes. Collectively, these findings sug- gest that endogenous as well as environmental estrogens control relevant genomic events through diverse signaling pathways in both ER-positive and ER-negative breast cancer cells (Fig. 8).
Of note, using the SKBR3 cells as a model system, we have determined that the G protein-coupled orphan receptor GPR30, via rapid ERK1/ERK2 activation, mediates an early c-fos upregulation by E2 and even genistein and quercetin without the involvement of ER␣. Moreover, the GPR30-deficient MDA-MB 231 and BT-20 breast cancer cells can be converted to cells that up-regulate c-fos in response to E2 and phytoestrogens upon heterologous expression of GPR30.
Consistent with the above observations, agents able to block G protein signaling such as pertussis toxin or the Src familyspecific tyrosine kinase inhibitor PP2, as well the MAPK inhibitor PD 98059 or DN/ERK2, prevent c-fos induction by treatments. Considering that EGFR tyrosine kinase activity was required for GPR30-dependent ERK1/ERK2 activation by E2 (9), we also ascertained its role in the up-regulated expression of c-fos in SKBR3 cells. Indeed, the EGFR kinase inhibitor tyrphostin AG-1478 abrogated the response to E2 and phytoestrogens. Given the above described potential of estrogen-like compounds on c-fos modulation, we wanted to evaluate an eventual response to other classes of hormones. Steroid receptor agonists tested, such as dexamethasone, progesterone, R1881, or even 17␣-estradiol, did not promote an enhancement of c-fos, clearly suggesting that the GPR30-mediated cascade is specifically activated by E2 and phytoestrogens. This result obtained in SKBR3 cells was in agreement with a previous study (30), which demonstrated the lack of c-fos sensitivity to treatment with the same agents in MCF7 cells.
Could the pure antiestrogen ICI 182780 or the mixed agonist/antagonist OHT modify the c-fos expression per se or act as antagonists of the c-fos response observed in SKBR3 cells? Neither compound was able to elicit significant changes in the presence or absence of E2 or phytoestrogens in this model system, whereas they partially reversed the induction of c-fos in MCF7 cells. These findings indicate that E2, genistein, and quercetin at least follow two diverse mechanisms of action in MCF7 cells, one involving ER␣, whereas the other is MAPKmediated, as suggested by the partial inhibition obtained with antiestrogens or using PD 98059 and the dominant negative ERK. In the absence of ER␣, as in the case of SKBR3 cells, the MAPK pathway emerges as crucial for the specific stimulation of c-fos by estrogens and phytoestrogens. On the contrary, wortmannin and LY 294002 at a concentration allowing cell survival did not alter the efficacy of treatments. As a consequence, the PI3K pathway might not be involved in c-fos protein expression in our experimental conditions, which were rather different as compared with those followed by Duan et al. (84), which described an involvement of both MAPK and PI3K cascades in E2-induced c-fos stimulation evaluated at the mRNA level.
c-fos is the prototypic member of a family of transcription factors that regulate many relevant physiological processes per se (50) and/or after heterodimerization with jun family members to form the AP1 transcription factor, which binds to specific sites within the regulatory region of target genes, the expression of which might be critical for cell cycle progression (86). Thus, the up-regulation of c-fos by extracellular stimuli could represent an early molecular sensor associated with relevant biological responses, including those involved in cell proliferation.
The presence of ER␣ in breast tumor biopsies has been recognized as a positive prognostic marker that correlates with higher survival rates and lower risk of relapse (68). However, about half of ER␣ ϩ tumors fail to respond to antiestrogens such as tamoxifen from the beginning, and most of the remaining ones eventually become resistant to the same treatment. Cancer progression from estrogen-stimulated and tamoxifensensitive growth to ER␣-independent and tamoxifen-resistant proliferation is well documented but remains essentially unexplained (85). In this context, signaling pathways triggered by endogenous and environmental estrogens able to promote the expression of genes involved in proliferative events independently of ER␣ should be taken into account for establishing a more comprehensive approach to antiestrogen therapy for breast cancer.