A G Protein–Coupled Receptor For Estrogen: The End Of The Search?

Estrogen is a steroid hormone involved in the regulation of a wide array of processes that include reproduction, sexual development, behavior, stress responses, bone integrity, neuroprotection, and cardiovascular health (1, 2). These effects take place in the central nervous system as well as in targeted tissues, such as the uterus and mammary gland (3).

In early studies, estrogen was considered a metabolic cofactor until the discovery of its first specific receptor, the estrogen receptor (ER), from extracts of rat uterus and vagina (4). The ER belongs to the steroid hormone nuclear receptor superfamily [reviewed in (2)]. Upon activation by their respective hormones, the members of this superfamily act as transcriptional factors modulating the expression of different genes. The ER was then named ERα when a second estrogen receptor, named ERβ, was isolated (5). ERα and ERβ mediate most of the estrogen-induced transcriptional effects; the actions of these receptors, however, are not redundant, but rather, complementary (2, 3). Another ER-related protein, named ERγ, has also been discovered (6).

The binding of estrogen triggers conformational changes in the ERs including their phosphorylation on serine and tyrosine residues, dissociation from the 90-kDa heat shock protein (hsp90), and subsequent ER dimerization. The ERα and ERβ receptors homodimerize, and can also heterodimerize to generate distinct functions. The receptor dimers interact with other transcriptional cofactors and with the estrogen response element (ERE), a specific regulatory DNA sequence present in the promoter of target genes (2, 3). The resulting suppression or enhancement of gene expression (leading to protein synthesis) allows the estrogen-mediated physiological responses to take place within hours after exposure to estrogen (2, 3). For these “genomic” effects, estrogen seems to directly modulate DNA using its ERs, and this process does not involve the second messenger signaling pathways.

Many estrogen-regulated effects, however, cannot be explained by this scheme of action. Some estrogen-targeted genes do not possess an apparent ERE (3), and many of the effects of estrogen occur rapidly within seconds to minutes (2, 3) and involve a variety of intracellular second messenger signaling pathways (7–9). Reports have suggested that these “rapid” and “nongenomic” events were mediated via ERα and ERβ (10, 11); however, growing evidence suggests the involvement of plasma membrane receptors, and especially G protein–coupled receptors (GPCRs).

As early as 1977, specific binding sites for estrogen were reported on the outer surface of isolated endometrial cells (12), but because of the difficulties in isolating and characterizing these sites, the existence of plasma membrane–associated ERs has remained controversial for over three decades. Also, the interests of researchers were focused more on estrogen-initiated genomic actions than on the nongenomic ones, which may have retarded progress in the elucidation of the membrane-related actions of estrogen.

In fact, a growing body of functional, biochemical and pharmacological evidence clearly suggest the presence of plasma membrane-associated estrogen receptors. One of these receptors was recently identified by different teams to be GPR30, an orphan GPCR (13, 14). Other membrane-associated estrogen receptors (mERs) have also been described.

A report that 17β-estradiol (E₂) activated the mitogen-activated protein kinase (MAPK) pathway in untransfected CHO-K1 and COS7 cells as well as in Rat2-fibroblasts (3), suggested that these cell lines may endogenously express an unidentified mER (3). It was also reported that E₂ activated the serine–threonine protein kinase B-Raf and the MAPK/ERK (extracellular-regulated kinase) pathway in cerebral cortical explants derived from ERα^−/− as well as from wild-type mice (8). Selective agonists to ERα and ERβ were unable to elicit ERK phosphorylation, and their antagonist ICI 182,780 was unable to block these actions, suggesting the presence of a new mER in mouse brain (8).

Another study showed that E₂ altered μ opioid and γ-amino butyric acid (GABA) mediated neurotransmission rapidly in hypothalamic neurons via a mechanism involving protein kinase C (15). These effects were mimicked by a membrane-impermeable complex of E₂ with albumin (E₂-BSA) and separately by a new selective modulator of E₂ termed STX (a diphenylacrylamide), which is unable to bind to ERα or ERβ, suggesting that the actions of E₂ were mediated by a novel, Gq-coupled mER (15). E₂ also binds to a specific receptor that activates protein kinase A (PKA), resulting in the μ-opioid receptor uncoupling from K⁺ channels (16). The estrogen-induced modulation of neuronal excitability has also been demonstrated for different GPCRs in β-endorphin, dopamine, and GABAergic neurons (17), as well as for acetylcholine-induced responses in GT1-7 cells involving a mER visualized on the cell surface through fluorescein-conjugated E₂-BSA (18).

A specific and unique mER, referred to as ER-X, was also characterized and shown to be developmentally regulated and distinct from ERα and ERβ (19). Its pharmacological profile was different from that of ERα and ERβ, with some characteristics opposite to those shown for these receptors. For example, the association of hsp90 is required for the inactive steady state of ERα, whereas, ER-X needs to be associated with hsp90 to mediate MAPK/ERK activation (3, 19). A heterodimeric estrogen-binding protein, named the putative ER (pER), demonstrated a high subnanomolar affinity for E₂ but was unable to bind other steroids, such as synthetic estrogens or antiestrogen. Depending on the cells examined, pER was localized at the plasma membrane and/or nuclear membrane, or in the cytoplasm and/or nucleus (20). Other new but unidentified mERs have been reported based on their ability to transduce functional responses, such as K⁺ channel activation, adenosine 3′,5′-monophosphate (cyclic AMP, cAMP) accumulation, and calcium mobilization (Table 1⇓) [reviewed in (2, 3)].

Recently, different groups identified and characterized an orphan GPCR, GPR30, as a new estrogen-receptor involved in the rapid actions elicited by estrogen (13, 14). In late 1990s, we and others independently cloned GPR30 (21–23). Its deduced aminoacid sequence showed a heptahelical-serpentine transmembrane structure—one of the characteristics shared by all GPCR superfamily members. Based on its structural similarity with receptors that bind chemokines and angiotensin II, the endogenous GPR30 ligand was thought to be a peptide; however, when tested, neither chemotactic peptides nor angiotensin II or angiotensin VI bound to GPR30 (3). GPR30 is widely distributed in different tissues including heart, breast, lung, CNS, vascular endothelium, and leukocytes (2, 23), and in ER-positive tumor cell lines (24).

Because GPR30 is expressed in tissues and in tumors that respond to estrogen, and because G-protein inhibition alters the second messenger pathways elicited by estrogen in these tissues and tumors, different groups focused on the participation of GPR30 in rapid-onset estrogen-mediated actions. Thus, estrogen was shown to induce adenylyl cyclase (AC) and MAPK activation in MCF-7 breast cancer cells that express GPR30, but not in MDA-MB 231 cells that express ERβ but not GPR30 (7, 25). These effects, however, were restored in MDA-MB 231 cells transfected with GPR30 cDNA (7, 25), and involved G_βγ-subunits and downstream activation of a Src-related tyrosine kinase. ER antagonists including ICI 182,780 were able to induce ERK activation, suggesting that this estrogen-induced action was GPR30-mediated and was independent of ERα and ERβ (7, 25).

GPR30 also mediates increased c-fos expression elicited by E₂ or by the phytoestrogens genistein and quercitin, in breast cancer cells (26). Furthermore, increased c-fos expression was repressed in GPR30-expressing SKBR3 cells transfected with an antisense oligonucleotide against GPR30, and was reconstituted in GPR30-deficient MDA-MB 231 and BT-20 breast cancer cells transfected with GPR30 cDNA (26). Increased c-fos expression was sensitive to G_βγ− and pertussis toxin (PTX), and involved the tyrosine kinase activities of Src and the epidermal growth factor (EGF) receptor (26). Specifically, the ability of the agonists to increase c-fos expression in SKBR3 cells was completely abrogated when cells were pretreated with either PTX or the Src family tyrosine kinase inhibitor PP2 (26).

Recently, GPR30 was definitively identified as a mER by two different groups (13, 14). Thomas et al. culminated their investigations of this receptor by identifying it as a high-affinity, saturable, displaceable, single binding site for E₂ in membranes from breast cancer SKBR3 cells (13). They had previously established that these cells do not express ERα or ERβ but do endogenously express GPR30. These authors also showed that, unlike untransfected cells, Human Embryonic Kidney 293 (HEK 293) cells transfected with GPR30 cDNA possess a high-affinity binding site for estrogen, which, when activated, led to the activation of a stimulatory G protein and subsequent activation of AC (26). Taken together their results showed that GPR30 was a plasma membrane-associated GPCR, mediating estrogen-elicited rapid responses.

Revankar et al. showed that GPR30 activation by estrogen led to intracellular calcium mobilization and phosphatidylinositol- 3,4,5-trisphosphate accumulation in the nucleus (14). The most intriguing part of this study, however, was the finding that GPR30 was exclusively localized in the endoplasmic reticulum—results confirmed by differentially tagged GPR30 constructs and fluorescent estrogen derivatives. This differs from what was observed by Thomas et al. (13) and by others, who have observed GPR30 localized at the plasma membrane. This finding by Revankar et al. (14) also is in contrast with the plasma membrane-association and functionality usually described for GPCRs. It is a fact that processing is a critical step for GPCRs to exit from endoplasmic reticulum and to traffic to the plasma membrane. Some of these receptors, such as the GABA_B receptor, must to be associated with other GPCRs or chaperone proteins to reach the plasma membrane. Nevertheless, and taking into account all the processing steps, the finding by Revankar et al. (14), if confirmed, would make GPR30 unique, or at least confirm GPR30 as the first GPCRs that functions in a non-nuclear intracellular compartment.

We are faced with a growing body of compelling evidence for the presence of ERs that are completely different from the well-known ERs that have nuclear functions. These newly identified receptor(s) mediate estrogen-elicited rapid responses, and control different physiological, and probably, pathophysiological actions of estrogen. Accumulating evidence clearly identifies GPR30 as one of these receptors, responsible for rapid estrogen-elicited responses in different cell lines, whether it is localized at the plasma membrane or in intracellular compartments. Is GPR30, functioning as “the membrane receptor for estrogen,” responsible for all the rapid nongenomic responses of estrogen or are there different mERs representing a new subfamily of plasma membrane estrogen receptors? Further characterization of GPR30 and other putative mERs, their (sub)cellular and tissue distribution, as well as their mode of action, will contribute to better understanding of the complexity of estrogen receptor-mediated signal transduction in relation to the wide range of physiological roles played by estrogen. It may also help to clearly distinguish estrogen-elicited functions (genomic vs nongenomic), and may open new therapeutic windows for the drug discovery process to target the pathological as well as beneficial effects linked to estrogen and its receptors.

• © American Society for Pharmacology and Experimental Theraputics 2005