Although it is often, but not always, straightforward to link the transcriptional response of estrogens with the ligand-dependent genomic model of ER activity, the observation that the rapid signaling events mediated by GPR30 can also lead to the activation of transcriptional machinery has provided further insight into the complexity of estrogen function regardless of ER expression. In this vein, Kanda and Watanabe demonstrated that E2 through GPR30 upregulates nerve growth factor production by inducing c-
fos expression via cAMP in macrophages (
Kanda and Watanabe, 2003a). The same authors demonstrated that E2 induces cyclin D2 and Bcl-2 expression via protein kinase A-mediated CREB phosphorylation in keratinocytes (
Kanda and Watanabe, 2003b;
Kanda and Watanabe, 2004). In a similar vein, Hsieh
et al. recently showed that E2 attenuates hepatic injury after trauma-hemorrhage by upregulating Bcl-2 expression through a GPR30 and PKA-dependent pathway (
Hsieh et al., 2007).
The expression of the oncogene c-
fos, used as an early molecular sensor of estrogen action, provided further evidence of GPR30-dependent transcriptional activation by E2 in ER-positive MCF7 and more importantly, ER-negative SKBr3 breast cancer cells (
Maggiolini et al., 2004). This study proved that GPR30 signaling requires EGFR and occurs through rapid ERK1/2 phosphorylation in triggering the genomic response to estrogen notably in tumor cells devoid of ERs (
Maggiolini et al., 2004). Further extending these results, E2, the phytoestrogen genistein and the 4-hydroxylated metabolite of the SERM tamoxifen (OHT) induced the expression of c-
fos through the GPR30/EGFR/ERK signaling pathway and above all, also induced proliferation of thyroid tumor cells lacking ER (ARO cells) or cells expressing a non-transcriptionally active variant of ERα (FRO and WRO cells) (
Vivacqua et al., 2006a). Taking into account that thyroid cancer is three times more frequent in women than in men from the onset of puberty until menopause, when this ratio declines progressively (
Henderson et al., 1982), the gender-dependent difference observed worldwide (
Waterhouse et al., 1982) and the increased risk in women taking estrogens for gynecological disorders (
Persson et al., 1996), the GPR30 pathway may represent a new window to circumvent the classical ER-mediated biological thyroid cell response.
The agonist activity of E2 and OHT elicited through the GPR30/EGFR/ERK signaling pathway was also shown in endometrial cancer cells harboring WT ERα (Ishikawa) or its splice variant (Hec1A) (
Vivacqua et al., 2006b). In these cell contexts, OHT still retained the antagonist property on ERα activation by E2, yet mediated induction of c-
fos and cell proliferation in a GPR30-dependent fashion similar to E2 (
Vivacqua et al., 2006). These findings provided further insight into the molecular mechanisms potentially involved in the increased incidence of endometrial cancer in women treated with tamoxifen for breast tumors (
van Leeuwen et al., 1994). Interestingly, in patients with endometrial carcinoma, GPR30 overexpression positively correlated with EGFR levels, occurred more frequently in high-grade, biologically aggressive histological subtypes and was associated with poorer survival rates (
Smith et al., 2007).
The recent availability of the GPR30-specific agonist G-1 (
Bologa et al., 2006) represented a key experimental tool towards a precise demonstration of the estrogen-induced and GPR30-mediated transcriptional activation events involving cross-talk with ERα (
Albanito et al., 2007). Taking advantage of the lack of any detectable activity of G-1 on the classical ER and using as model systems ovarian cancer cells expressing both ERα and GPR30, it was observed that G-1, like E2, up-regulated diverse estrogen-responsive genes including c-
fos, pS2 and cyclins A, D1 and E; however, it failed to increase the ERα-target gene PR, which only responded to E2 treatment (
Albanito et al., 2007). These data were further corroborated using ER-negative and GPR30-positive SKBr3 cells, where G-1 like E2 stimulated c-
fos expression, but had no effect on PR expression (
Albanito et al., 2007). Together, these results suggested that estrogen-activated PR expression occurs specifically through ERα, while GPR30, possibly together with ERα (see below), mediates the transcriptional activation of the other genes (). In addition, E2 and G-1 used in combination in ovarian cancer cells did not show any further increase in the transcriptional activation of
c-fos compared to either compound alone, suggesting that a common pathway mediates the genomic response. Of note, knocking down GPR30 or ERα revealed a cross-talk between these estrogen receptors in the stimulation of c-
fos by G-1 and E2. The above-mentioned regulation of c-
fos was predictive of the ovarian cell growth observed silencing ERα or GPR30 expression, which interestingly prevented the proliferative effects induced by either ligand (
Albanito et al., 2007). In SKBr3 cells, which express GPR30 but lack ER, the knock-down of GPR30 was sufficient to block the growth stimulation by G-1 and E2. Overall, the findings indicate that cooperation between ERα and GPR30 may take place when both receptors are co-expressed, as also suggested by Sukocheva et al. (
Sukocheva et al., 2006).
Nevertheless, GPR30 can be sufficient to signal alone in absence of ER as in SKBr3 breast cancer cells. In these latter cells, a recent study (
Albanito et al., 2008a) showed that the widespread environmental contaminant and endocrine-disruptor, atrazine, activates GPR30-dependent signaling, although in ovarian cancer cells, both GPR30 and ERα were required to induce c-
fos expression and cell proliferation in line with the aforementioned results obtained using E2 and G-1. Previous studies have shown that atrazine may exhibit an estrogen-like action increasing aromatase expression and activity without any direct agonism or antagonism of the classical ERs (
Fan et al., 2007a;
Fan et al., 2007b;
Heneweer et al., 2004;
Roberge et al., 2004;
Sanderson et al., 2001;
Tennant et al., 1994). In the ovarian cancer cells used by Albanito et al. (
Albanito et al., 2008a), atrazine neither showed an ability to interact with ERα nor stimulated aromatase activity. Interestingly, atrazine acted through both GPR30 and ERα via the EGFR/MAPK signaling pathway to trigger transcriptional activation and cell proliferation. The authors concluded that a complex interplay between GPR30 and ERα contributes to atrazine activity, which nevertheless is still elicited in presence of GPR30 alone as demonstrated in ERα-negative SkBr3 breast cancer cells. From these data it is reasonable to argue that the evaluation of estrogenic activity of phyto- and xenoestrogens should be extended to their potential ability to activate GPR30 signaling alongside the well-known agonist effects exerted through the classical ER-mediated genomic response.
In order to examine the effects of extranuclear estrogen-mediated pathways on global gene expression, Madak-Erdogan
et al. recently evaluated the action of E2 and estrogen-dendrimer conjugates (EDCs), which are unable to cross the nuclear membrane (
Harrington et al., 2006), in a genome-wide cDNA microarray analysis of MCF-7 breast cancer cells (
Madak-Erdogan et al., 2008). Approximately 25% of all E2-regulated genes responded to a 4 h treatment with EDC, independently of the recruitment of ERα to ERE. Furthermore, antiestrogens or ERα knockdown, as well as MAPK and c-Src kinase inhibitors, abolished the up-regulation of EDC-sensitive genes. Interestingly, the authors suggested that EDC signaling triggers transcriptional activation through rapid kinase activity via ERα without its recruitment to EREs. On the contrary, the authors failed to reveal any potential role for GPR30, possibly due to the experimental design and model system used.
A physiological role for GPR30-mediated transcriptional responses through cross-talk with ERα has recently been found in mouse spermatogonia GC-1 cells, which served to investigate the estrogen-mediated regulation of testicular function (
Sirianni et al., 2008). On the basis of an altered testicular phenotype in ERα knockout mice (
Das et al., 2000;
Eddy et al., 1996), which is less severe compared to aromatase knockout mice, Sirianni
et al. investigated the potential involvement of an estrogen-binding receptor different than the well-known ERs, such as GPR30, in estrogen signaling (
Sirianni et al., 2008). The authors demonstrated that E2 and G-1 activate the EGFR/ERK pathway leading to the stimulation of c-
fos and cyclin D1 expression as well as GC-1 cell growth. Interestingly, the proliferative effects induced by E2 and G-1 were abrogated using either the ERα antagonist ICI 182780 or silencing GPR30 expression. The results obtained are consistent with data recently reported by Bouskine
et al., demonstrating that E2, through the activity of a Gi protein, could induce rapid activation of ERK1/2 and PKA signaling pathways, which in turn are involved in the proliferation of human germ cell tumors (
Bouskine et al., 2008).
To this point, the aforementioned studies have broadened the molecular mechanisms through which estrogen signaling may involve the EGFR signaling pathway in modulating gene expression. In this regard, the genotropic activity of estrogen has been largely recognized to be mediated by the ERs, although many E2-responsive genes are key molecules participating in EGFR signaling (
Levin, 2003). Interestingly, the cell membrane-associated form of ER has been reported to couple with and activate diverse G proteins, thereby triggering biological responses via EGFR transactivation (
Levin, 2003;
Razandi et al., 2003). Moreover, E2 activation of GPR30 activates the EGFR signaling cascade (
Filardo et al., 2000) similar to other GPCR ligands (
Bhola and Grandis, 2008;
Rozengurt, 2007). Recently, Albanito et al. provided evidence in ER-negative breast cancer cells that EGF, by up-regulating GPR30 expression, engages E2 to potentiate the biological response to EGFR signaling (
Albanito et al., 2008b). This positive feedback loop between EGFR and GPR30 activity represents a further demonstration that signaling pathways may be able to auto-regulate the amplitude of their own activation to modulate signaling robustness in the face of variable inputs (
Freeman, 2000). The results obtained by Albanito et al. also demonstrated that the classical ERs may not be required for GPR30/EGFR cross-talk; however, GPR30 and ERα can cooperate in mediating the action of E2 as observed in ER-positive, GPR30-positive ovarian cancer cells (
Albanito et al., 2007). Finally, the clinical observation that GPR30 overexpression is associated with lower survival rates in endometrial cancer patients (
Smith et al., 2007) and higher risk of developing metastatic disease in patients with breast tumor (
Filardo et al., 2006) suggests an important involvement in carcinogenesis. Therefore, the expression levels of GPR30, which are regulated by EGF/EGFR signaling may characterize the estrogen sensitivity of these tumors in addition to predisposing tumors to an altered responsiveness to endocrine therapy.
Very recently, a set of genes has been identified that may contribute to the proliferative activities of GPR30 (
Pandey et al., 2009). Of particular interest was the prominent induction of connective tissue growth factor (CTGF) by both E2 and OHT through GPR30 signaling, which enhanced breast cancer cell migration. As CTGF was also induced by OHT in fibroblasts from breast tumor biopsies, this points to the potential involvement of GPR30 pathways in the aggressive behavior of breast tumors, particularly in response to endogenous estrogens or OHT used in (chemo)therapeutic interventions.