Estrogen receptors and human disease: an update

Abstract

A myriad of physiological processes in mammals are influenced by estrogens and the estrogen receptors (ERs), ERα and ERβ. As we reviewed previously, given the widespread role for estrogen in normal human physiology, it is not surprising that estrogen is implicated in the development or progression of a number of diseases. In this review, we are giving a 5-year update of the literature regarding the influence of estrogens on a number of human cancers (breast, ovarian, colorectal, prostate, and endometrial), endometriosis, fibroids, and cardiovascular disease. A large number of sophisticated experimental studies have provided insights into human disease, but for this review, the literature citations were limited to articles published after our previous review (Deroo and Korach in J Clin Invest 116(3):561–570, 2006) and will focus in most cases on human data and clinical trials. We will describe the influence in which estrogen’s action, through one of or both of the ERs, mediates the aforementioned human disease states.

Estrogen receptors

Elucidating the functions of the estrogen receptors (ER) in normal and diseased states is important for the development of relevant therapeutic strategies. Two main forms of ER exist, ERα and ERβ, which are encoded by separate genes, Esr1 and Esr2, respectively. They are transcribed from different chromosomal locations and multiple splice variants exist for these receptors. Each receptor has distinct tissue expression patterns, post-translational modifications, and cellular localization in normal and disease states. The ERs are classical hormone nuclear receptors and members of the nuclear receptor super family having the functional structural domains A-F (Mangelsdorf et al. 1995; Hewitt et al. 2009; Burns et al. 2011). The A/B-domain contains the activation function 1 (AF-1), which is hormone-independent. The C-domain holds the DNA-binding domain, the D-domain or hinge region has nuclear localization sequences and interacts with AP-1, and the E/F-domain is the ligand-binding domain with the ligand-dependent activation function, AF-2. Over the last 5 years, great strides have been made to understand the post-translational modifications, polymorphisms, methylation status, and localization of ER. Most notably and much beyond the scope of this review, sequencing of chromatin immunoprecipitation enriched chromatin fragments (ChIP-seq) has addressed the relationship between ER chromatin interactions and responses to estradiol (Fullwood et al. 2009; Welboren et al. 2009a, b; Grober et al. 2011; Hewitt et al. 2012). The aforementioned studies reveal significant ER binding to chromatin without ligand, increased binding of ER with ligand and binding of ER quite distal to the promoter start sight or in intronic regions. These detailed studies are providing a vast amount of data that will be a part of the continued elucidation of the various mechanisms and function of the ERs, which are critical in the development of therapeutics. In this review, we will focus on the alterations of the ERs in human disease.

Breast cancer and ER

A plethora of clinical and experimental data is available regarding the ERs and breast cancer; therefore, this review will focus on the literature pertaining to human disease and clinical data. Most notably, breast cancer incidence has decreased in recent years, most likely as a result of the decline in the number of women receiving hormone replacement therapy due to negative health findings from the Women’s Health Initiative (Chlebowski et al. 2009). The localization of hormone nuclear receptors in normal breast tissue has not been well studied; consequently, Li et al. studied the localization of ERα, ERβ, progesterone receptor (PR)-A and PR-B, and androgen receptor (AR) in normal human mammary gland tissue. Samples from premenopausal women show that ERα, PR-A, PR-B, and AR localize mostly to the inner layer of epithelial cells lining acini and intralobular ducts and to myoepithelial cells in the external layer of interlobular ducts (Li et al. 2010). ERβ was found to be more widespread, having staining of stromal, epithelial, and myoepithelial cells in acini and ducts. Many studies have correlated the nuclear receptor status of the tumor to treatment regimens and survival outcomes.

Breast cancer survival and ERα

Estrogens play a central role in breast cancer development with ERα status being the most important predictor of breast cancer prognosis. ERα has long been determined to be a prognostic marker for breast cancer and increased survival is seen with ERα-positive status as these tumors respond to anti-estrogen therapy. The exception, however, is tumors which overexpress a 36-kDa variant of ERα, termed ERα-36. ERα-36 has no intrinsic transcription activity, but localizes predominantly to the plasma membrane and cytoplasm of cells and mediates the membrane-initiated non-genomic signaling pathway. The overexpression of ERα-36 is associated with poor disease-free survival, and patients are less likely to benefit from tamoxifen treatment (Shi et al. 2009; Lin et al. 2010; Vranic et al. 2011). Current studies associating breast cancer with ERα-positive status focus on post-translational modifications to ERα in hopes to find better therapies by understanding the ERα responsiveness. Multiple phosphorylation sites in ERα can be detected in breast tumor biopsy samples. The relationship of ERα phosphorylation (pS-104/106-ERα, p-S118-ERα, p-S167-ERα, p-S282-ERα, p-S294-ERα, p-T311-ERα, and p-S559-ERα) to clinical outcome after tamoxifen therapy suggests high phosphorylation status is associated with increased mortality (Skliris et al. 2010; Murphy et al. 2011). ERα-S305 phosphorylation positive breast cancers are resistant to adjuvant tamoxifen treatment, while ERα-S305 phosphorylation negative tumors have improved recurrence-free survival with tamoxifen treatment (Holm et al. 2009; Kok et al. 2011). Consequently, blocking ERα-S305 phosphorylation may represent a new therapy strategy (Barone et al. 2010). ERα status and decreased p44/42 mitogen-activated protein kinase (MAPK)/ERK1/2 activity are also factors that suggest a successful response to treatment (Generali et al. 2009). More specifically, ERα low phosphorylation at S118 and high phosphorylation of S167 are associated with improved disease-free survival (Yamashita et al. 2008; Motomura et al. 2010). A phosphorylated form of membrane ERα is proposed to characterize an invasive cancer state and suggests transcription-independent cellular responses to estrogen are involved in disease progression (Mintz et al. 2008). The interplay of ERα and MAPK to integrate signaling inputs occurs by estrogen-occupied ERα activating and interacting with ERK2 which colocalizes ERK2 and ERα at chromatin-binding sites (Madak-Erdogan et al. 2011). Phosphorylation modifications of ERα affect survival in ERα-positive breast cancer and could be used to distinguish patients who are more likely to benefit from endocrine therapy alone from those who may be resistant.

In breast cancer, ER, PR, human epidermal growth factor receptor 2 (HER2) and Ki67 are biological makers for predicting prognosis and treatment decisions, but other nuclear receptors/transcription factors also play a role in breast cancer progression. The expression of ERα and PR in normal mammary tissue adjacent to the pathological tissue compared to women with benign breast disease shows that the overexpression of ERα and PR is associated with reduced breast cancer risk, but increased age at the diagnosis of hormone receptor-positive tumors is associated with higher disease-specific mortality (Lagiou et al. 2009; van de Water et al. 2012). In HER2-positive tumors, ERα and PR are associated with AR co-expression and lower proliferative activity, while AR-negative/ERα-negative tumors were associated with highest proliferative activity and histological grade (Lin Fde et al. 2012). Suggesting that co-expression of AR and ERα provides a protective effect. Tumors with high ERα expression and low Ki67 labeling are associated with tamoxifen and aromatase inhibitor treatment as a first-line endocrine therapy; however, low levels of ERα are a determinant of tamoxifen resistance in ERα-positive cancers (Endo et al. 2011; Kim et al. 2011).

Breast cancer survival and ERβ

Both ERα and ERβ are normally present in the mammary gland. ERα is a prognostic marker in breast cancer, but the role of ERβ is less clear. ERβ is regulated by transcriptional and post-transcriptional modifications; ERβ mRNAs are transcribed from three promoters and variants of ERβ1, ERβ2 and ERβ5 play a role in breast cancer (Smith et al. 2010). Immunohistochemical expression of ERα and ERβ in a breast cancer tumor tissue array demonstrated that 78 % of tumors are ERα-positive and 50 % of tumors are ERβ-positive. In the 512 tumors analyzed, the ERα-positive status correlates with survival and treatment responses, but no overall prognostic significance was demonstrated for ERβ (Borgquist et al. 2008). Although in a separate study, ERβ expression was found associated with ERα expression (Marotti et al. 2010). In contrast, a more detailed study of ERβ status examining nuclear and cytoplasmic expression of ERβ1, ERβ2, and ERβ5 suggests that the cellular localization of these isoforms differentially affects patient outcome (Chen et al. 2007; Mandusic et al. 2007; Shaaban et al. 2008; Yan et al. 2011). Examination of ERβ isoforms from laser microdissected human breast cancers show that the levels of the ERβ isoform expression are cell type and cellular compartment-dependent (Cummings et al. 2009). Cytoplasmic ERβ2 expression predicts worse overall survival, but positive staining for ERβ1 nuclear is associated with better survival (Lin et al. 2007; Sugiura et al. 2007; Honma et al. 2008; Shaaban et al. 2008). In ERα-positive/PR-positive tumors, patients with higher ERβ levels had longer survival; however, no association of ERβ levels and survival was found with ERα-negative/PR-negative tumors (Maehle et al. 2009). In contrast, Gruvberger-Saal et al. suggest from their study that in ERα-negative tumors, ERβ expression responds to tamoxifen and increases overall survival (Gruvberger-Saal et al. 2007). In contrast, co-expression of ERβ and HER2 associates with poorer prognosis in primary breast cancers (Qui et al. 2009). No correlation between ERβ levels and classical prognosticators for breast cancers are found; however, in node-positive cancers, ERβ-positive is related to the need for a more aggressive clinical course (Novelli et al. 2008). Similarly, in ERα-negative tumors, a correlation between ERβ protein expression and Ki67 is seen (Rosa et al. 2008). One study has demonstrated that nuclear ERβ-S105 phosphorylation is associated with better survival even in tamoxifen-resistant cases (Hamilton-Burke et al. 2010). Overall, the data suggest that ERβ is playing some role in breast cancer that is independent of ERα. A more detailed characterization of ERβ as well as that of other hormone receptors will be required to determine the role for ERβ as a predictor for survival.

Breast cancer and ER’s interaction with transcription factors

Endocrine resistance is a major hurdle to hormonal treatments for breast cancer, but recent studies have focused on finding ER-interacting proteins to potentially target as new therapeutic treatments. Tumors with higher IGF-1 and ERα take longer to develop tamoxifen resistance (Chong et al. 2011). In contrast, CUE domain-containing protein-2 (CUEDC2) expression and ERα expression are inversely correlated with high CUEDC2 expression being a predictor of poor responsiveness to tamoxifen treatment (Pan et al. 2011). Fork head proteins (FOX) belong to a group of pioneering transcription factors that interact with ER. FOXA1 and ERα predict favorable outcomes in breast cancer patients (Bernardo et al. 2010). FOXA1 influences genome-wide interactions between ERα and chromatin. In ERα-responsive breast cancer cells, response to tamoxifen depends on FOXA1 and ERα interaction, but in tamoxifen-resistant cells, ERα binding was ligand-independent but dependent on FOXA1 (Hurtado et al. 2011; Ross-Innes et al. 2012). FOXM1 has also been identified as a transcriptional target of ERα and may play a role in mitogenic function (Millour et al. 2010). Another genomic pioneering factor, pre-B cell leukemia homeobox 1 (PBX1) drives ERα signaling and underlies progression of breast cancer. PBX1-transcriptional program is associated with poor outcome in patients and a study in cells with PBX1 knockdown no longer proliferate after estrogen (Magnani et al. 2011). Future therapies may be aimed to target ERα-interacting partners.

Understanding the interplay of ERβ and interacting partners is more in its infancy than ERα. Examination of interacting partners of ERβ was done in MCF-7 cells, and the interactome revealed 303 proteins that co-purify with ERβ nuclear extracts (Nassa et al. 2011). The identification of these interacting partners should pioneer a number of studies to more adequately understand the biological role of ERβ in cancer. Important early studies show the presence of ERβ attenuates the transcriptional activity of ERα (Matthews et al. 2006). Since that time, an interesting study with MCF-7 cells engineered to express ERα only, ERβ only, or ERα and ERβ examined ER-binding sites. The results reveal substantial binding site overlap when the ERs were expressed alone, with many fewer sites shared with both ERs present (Charn et al. 2010). Each ER restricts the binding of the other with ERα being more dominant to displace ERβ to new sites less enriched for estrogen response elements (ERE) (Charn et al. 2010). Additionally, the knockdown of ERβ in MCF-10 and MCF-7 cells results in increased growth in a ligand-independent manner while overexpression of ERβ allows MCF-7 cells to respond to tamoxifen and inhibit proliferation (Hodges-Gallagher et al. 2008; Treeck et al. 2010). In a mouse model, over-expression of ERβ in xenograft tumors from T47D cells reduced tumor growth and angiogenesis (Williams et al. 2008). These results regarding ERβ suggest this ER exerts anti-proliferative effects on breast cancer.

Breast cancer and miRNA association with ER

MicroRNAs (miRNAs) are small noncoding RNAs that regulate gene expression in different physiological and pathological process often negatively by hybridizing to complementary sites of target transcripts. Many miRNAs are implicated in several human cancers, including breast cancer, associated with the loss of tumor suppressor miRNAs or the overexpression of oncogenic miRNAs (O’Day and Lal 2010). Here, we discuss the association of ERs and miRNAs in breast cancer. miR-22 represses ERα expression by directly targeting its 3′ untranslated region (UTR), which induces mRNA degradation (Pandey and Picard 2009). ERβ1 is often down-regulated in breast cancers, and analysis of primary breast tumors demonstrates this may be due to a negative correlation between miR-92 and ERβ1 mRNA and protein (Al-Nakhle et al. 2010). In vitro experiments confirm this finding by showing that overexpression of miR-92 decreases ERβ1 via direct targeting of the ERβ 3′ UTR (Al-Nakhle et al. 2010).

Distinct expression of miRNAs is found between primary ERα-positive and ERα-negative breast cancer tumors and cell lines. Cells transfected with miRNA-221 and miRNA-222 have decreased ERα expression and increased resistance to tamoxifen treatment (Zhao et al. 2008). Acquired resistance to therapy occurs in a majority of breast cancer patients. Rao et al. show that prolonged fulvestrant (a selective estrogen receptor downregulator) therapy increases miR-221 and -222 expression, which results in the dysregulation of multiple oncogenic pathways associated with drug resistance (Rao et al. 2011). miR-221 and -222 are negatively modulated by ERα. A negative feedback loop may contribute to the proliferative advantage and migratory activity of breast cancer cells, which may promote the ERα-positive to ERα-negative tumor status transition (Di Leva et al. 2010). A microarray analysis for estrogen-regulated miRNAs in MCF-7 cells found that estrogen induces 21 miRNAs and represses seven of them, together potentially regulate 420 estradiol-regulated and 753 non-estradiol-regulated genes at the post-transcriptional level (Bhat-Nakshatri et al. 2009). A similar study in MCF-7 cells also found a subset of miRNAs regulated by estrogen and further demonstrates clinical relevance as pre-miR18a is upregulated in ERα-positive compared to ERα-negative breast cancers (Castellano et al. 2009). These studies suggest roles for miRNAs in the regulation of ER activity and consequently in the development and progression of breast cancers.

Breast cancer and ER single-nucleotide polymorphisms (SNPs)

Studies have examined single-nucleotide polymorphisms (SNPs) for ERα and ERβ to determine whether they confer risks for breast cancer development. Women with sporadic breast cancer more frequently (odds ratio (OR) = 1.99) carry the CC genotype of ERβ promoter SNP rs2987983 (Treeck et al. 2009). A large cohort examining four htSNPs that tag the six major haploytpes of ERβ demonstrated that the inherited variants in ERβ are not associated with an appreciable (OR > 1.2) change in breast cancer risk in Caucasian women (Cox et al. 2008). In contrast, in sporadic breast cancer, decreased risk is associated with the LL and the SL genotypes in c. 1092 + 3607(CA)(10–26) in ERβ (OR = 0.013) (Tsezou et al. 2008). The rs1801132 and rs2234693 ERα polymorphisms confer a slight decreased breast cancer risk for CC and CC/CT carriers (Li et al. 2010). C325G SNP in ERα is associated with increased cancer risk (OR = 2.28) in women 50 years and younger; however, overall susceptibility to breast cancer was not found (Siddig et al. 2008). SNP rs2046210 at 6q25.1 in the ERα promoter only shows a weak association with breast cancer in Chinese and European ancestry women (Zheng et al. 2009). In a study of more than 55,000 breast cancer patients to look at SNP rs3020314 in ERα intron 4 found no large risks for increased breast cancer susceptibility in European populations (Dunning et al. 2009). The expansive number of patients analyzed for ER SNPs suggests different SNPs may alter a woman’s risk for breast cancer.

Ovarian cancer and ER

Ovarian cancer has one of the highest rates of mortality of the gynecological malignancies, and female oncogenesis is most likely attributed to the asymptomatic nature of the disease and inadequate treatment strategies. Ovarian cancer is often not detected until the disease is of a late stage and symptoms of abdominal pressure, pelvic pain, persistent indigestion, and/or loss of appetite are recognized (Zahedi et al. 2012). The standard of care includes surgery and chemotherapy, but this often results in toxicities and disease relapse; therefore, it is of importance to find methods to detect this disease earlier. Many human studies over the last few years have focused on both ERs in the understanding of survival outcomes, associations with SNPs and molecular responses by utilizing ovarian cancer cell models.

In reproductive age women, ERα is present in the ovarian stroma and thecal cells, ovarian surface epithelium and in corpus luteum; however, the localization of ERα in the ovaries of postmenopausal women was unknown (Brodowska et al. 2007). In contrast, ERβ is localized predominately to the granulosa cells of the ovary (Brandenberger et al. 1998). In order to make hormone treatments of ovarian cancer more effective, the normal localization of ERα in the ovary of postmenopausal women was examined. ERα is found in the ovarian surface epithelium, in epithelial inclusion cysts and stroma of women who had their last menstruation less than 5 years before the study, while women who had their last menstruation more than 5 years before the study had decreasing levels of ERα in the ovary.

Ovarian cancer and survival

The levels of ERα and ERβ and their relationship in expression levels to other hormone receptors have been widely studied recently. Epithelial cells in advanced serious ovarian cancer tissue have lower levels of ERβ and PR, but not ERα when compared to normal ovarian tissue (Issa et al. 2009; De Stefano et al. 2011). ERα expression is associated with progression-free survival and cause-specific survival as these patients can be selected for specific anti-hormonal therapy similar to hormone-responsive breast cancers (Burges et al. 2010). The Nurses’ Health Study reported that in epithelial ovarian tumors, PR expression is less likely to be invasive, of a lower grade and stage than PR-negative tumors, while ERα status in these tumors did not associate with any pathologic features of the tumor (Hecht et al. 2009). Experimental models designed to study the mechanism of action of ERα in ovarian cancer suggest roles for leptin as an induced signal transducer and activator of transcription 3 (STAT-3) binding to ERα. Down-regulation of E-cadherin by the ERα regulation of Snail and Slug, as well as reciprocal roles of estrogen, interleukin-6 (IL-6), and IL-8 are all being considered as cellular actions (Park et al. 2008; Salonen et al. 2009; Yang et al. 2009; Choi et al. 2011).

On the other hand, loss of ERβ expression correlates with shorter overall survival of ovarian cancer and may be a feature of malignant transformation as the expression of ERβ decreases with cancer progression (Chan et al. 2008; Halon et al. 2011). A reason for this transformation may be that ERβ expression is predominately nuclear in normal ovarian tissue, but often ovarian cancers exhibit cytoplasmic ERβ immunopositivity. ERβ levels are inversely correlated with metastasis-associated gene 1 (MTA1) expression and in an in vitro cell model, overexpressing MTA1 reduces ERβ levels (Dannenmann et al. 2008). Re-expression of ERβ in an ERβ-null ovarian clear cell adenocarcinoma cell line inhibits proliferation, migration, and motility (Zhu et al. 2011). The increased loss of ERβ with ovarian cancer progression suggests a protective role for ERβ in the ovary.

Ovarian cancer and SNPs and methylation

SNPs and methylation have been examined to determine whether they may be risk factors for ovarian cancer. A comprehensive study of epithelial ovarian cancer in American white women examined 13 different SNPs in ERα and found the strongest association of rs2295190 for the mucinous subtype of ovarian cancer (OR = 1.32) (Doherty et al. 2010). The ERβ rs1271572 polymorphism has an OR = 1.35 for the risk of development of invasive ovarian carcinoma in women <50 years and was strongest among women who had never used contraceptive steroids (Lurie et al. 2009, 2011). A study to characterize the role of haplotype diversity in ERβ with the risk of ovarian cancer found five haplotypes with a frequency of >5 % in white subjects, but no association was observed for the risk of ovarian cancer (Leigh Pearce et al. 2008). Promoter methylation analysis of ERβ from treatment insensitive ovarian cancer cell lines revealed that promoter methylation is cell type-specific and that the loss of ERβ isoform expression and inactivation correlates with aberrant promoter methylation (Suzuki et al. 2008; Yap et al. 2009). Overall, a slight increase in ovarian cancer risk may be attributed to SNPs or altered methylation status of ERα or ERβ.

Non-target tissue cancers and ER

Colon cancer

Clinical and animal studies show that hormone replacement therapy reduces the risk of colon tumor formation. Research in the last 5 years has focused on the roles of ERα and ERβ in the development of colon cancer, but ERβ is denoted as the principal mediator in the colon (Giroux et al. 2011). ERβ and associated splice variants (ERβ1, ERβ2, and ERβ5) are highly expressed in normal colon tissue with less expression in early disease states and the lowest ERβ expression in colon adenocarcinomas (Castiglione et al. 2008). In familial adenomatous polyposis, decreased ERβ expression is related to the severity of disease and a mouse model of disease, Apc(Min/+), given an ERβ-selective agonist, diarylpropionitrile (DPN), results in a significant reduction in polyp multiplicity in both male and female experimental groups (Barone et al. 2010; Giroux et al. 2011). As prognostic markers in rodents and humans, ERβ knockout mice have increased the presence of mucin-depleted foci, an early event in colon cancer (Saleiro et al. 2010). Estradiol influences the physiology of non-cancerous colonocytes experimentally through ERβ during the initiation/promotion stage of disease development, which results in fewer pre-neoplastic lesions (Weige et al. 2009). In human colon cancer, ERβ along with other associated transcription factors are being studied to analyze ERβ’s role. Subsite-specific localization differences of ERβ expression in colonic epithelium have implications for the epidemiology of colon cancer (Papaxoinis et al. 2010). Low ERβ and high matrix metalloproteinases (MMP) 7 expressions are a risk factor in colon cancer. MMP7 is consistently expressed throughout cancer progression, but patients with higher ERβ nuclear expression have a higher 5-year survival rate (Fang et al. 2010). The co-expression of ERβ and proline, glutamate- and leucine-rich protein 1 (PELP1/MNAR) in epithelial cells of carcinomas is a favorable prognostic factor (Grivas et al. 2009). In insulin-resistant syndrome, a combined effect of insulin-like growth factor 1 (IGF1) and estrogens is correlated with colon cancers as a gradual increase in IGF-1 receptor expression and gradual decrease in ERβ expression is observed (Papaxoinis et al. 2007). These studies support a role for ERβ as a relevant biomarker of tumor progression and possible chemopreventive target in patients at risk for colonic neoplasia.

Even though ERα expression is extremely rare in colorectal tissue and its expression appears not to be associated with colon cancer, studies have addressed ERα (Grivas et al. 2009). Experimental studies of Apc(Min/+) mice lacking ERα showed an increased colon tumor burden and multiplicity compared to wild-type mice resulting in an elevated Wnt-β-catenin signaling pathway and resulting tumorigenesis (Cleveland et al. 2009). These studies suggest that estrogen’s protective action may be to suppress rather than stimulate the Wnt-β-catenin pathway in the colon. Decreased ERβ is observed in colonic polyps and colon cancer. ERα expression does not change in patients with colonic polyps; however, in colon cancer, ERα levels increase in men and not in women (Di Leo et al. 2008; Nussler et al. 2008). These data suggest sex-specific effects related to the ERα. In contrast, decreased levels of the ERα splice variants, ERα-36 and ERα-46, correlate with tumor stage and lymph node metastasis (Jiang et al. 2008). Examination of ERα promoter methylation status in sporadic colorectal cancer is found to be altered in normal background mucosa in association with vitamin B-12 status (Al-Ghnaniem et al. 2007; Hiraoka et al. 2010). Both ERα and ERβ are implicated in colon cancer, where ERβ is most likely a critical mediator and can be a target for mitigating cancer preventative effects.

Prostate cancer

Since our previous review, little research has been published regarding ERs and prostate cancer. Rodent studies show that estrogen signaling through ERα or ERβ is not needed for dioxin to inhibit prostatic epithelial budding, but a diet rich in sesame pericarp increases ERβ expression in the prostate and has a beneficial effect on the healthy status of the animal (Allgeier et al. 2009; Anagnostis and Papadopoulos 2009). In contrast, an ERβ antagonist selectively induced apoptosis in experimental castrate-resistant CD133+ basal cells and suggests a role for ERβ in prostatic stem cells and in prostate cancer. Estrogens are known to play a role in prostate carcinogenesis, but the expression of ERs throughout prostate carcinogenesis is unknown (Ellem and Risbridger 2007; McPherson et al. 2010; Risbridger et al. 2010). In general, low expression of ERα is detected in prostatic epithelial cells while ERβ is predominantly expressed in prostatic epithelial cells, but the ERα expression is highly variable in human tissues depending on disease status (Leav et al. 2001). Human studies have focused on SNPs of ERα and ERβ. ERα (rs1801132, rs2077647, rs746432, rs2273206, rs851982, and rs2228480) and ERβ (rs4986938, rs928554, rs8018687, and rs number not available for ESR2 5696 bp 3′ of STP A >G) SNPs from multiple cohorts with large study participation observed little evidence for any association of ERα or ERβ polymorphisms with prostate cancer risk (Chen et al. 2007; Chae et al. 2009). Clinical and basic science evidence linking estrogen and estrogen receptors to prostate cancer still remains an area of study.

Endometrial cancer

Endometrial cancer is the most common gynecological malignancy with risk factors being exposure to estrogens and having high body mass index (Collins et al. 2009). Recent human studies have focused on survival rates relative to hormone receptor status, association with other transcription factors and receptor SNPs. Examination of endometrial carcinomas by tissue microarray, fluorescence in situ hybridization (FISH), and immunohistochemistry demonstrates that the amplification of ERα is related to early-stage cancer, but in contrast, the absence of ERα correlates to death from disease (Lebeau et al. 2008; Jongen et al. 2009). Expression of ERα, PR-A, and PR-B is associated with lower-grade tumors, but patients with lower ratios of ERα/ERβ or PR-A/PR-B exhibit shorter disease-free survival (Jongen et al. 2009). A similar immunohistochemical analysis also demonstrates the loss of ERα, PR-A, and PR-B results in poorer survival, but the expression of ERβ does not correlate with survival (Shabani et al. 2007). Experimental studies show a similar correlation since mouse models of endometrial cancers show different severities of tumorigenicity when hormone receptors are lacking (reviewed in Yang et al. 2011a, b). ERα levels measured in metastatic endometrial carcinoma prior to hormonal treatment correlate with clinical response to tamoxifen and medroxyprogesterone acetate (Singh et al. 2007). Patient survival is associated primarily with ERα expression as these tumors are able to respond to anti-hormonal therapy.

In uterine carcinosarcoma, a rare and highly aggressive form of endometrial cancer, ERα and PR receptor expression are suppressed while ERβ expression is elevated and continues to increase with disease progression (Huang et al. 2009). In this cancer relative to normal endometrium, IGF-1 expression is reduced and HER2 increases, which supports a potential role for ERβ via crosstalk with HER2 in disease progression (Huang et al. 2009). Additionally, crosstalk may occur with G-coupled protein receptor 30 (GPR30) and ERβ as they are coordinately overexpressed and expression increases in advanced-stage disease (Huang et al. 2010). In endometrial cancers, crosstalk also occurs between steroid and prostaglandin pathways as cyclooxygenase 2 (COX-2) expression is higher in ERα-negative/low tumors and full length ERβ (ERβ1) and two ERβ variants (ERβ2 and ERβ5) are expressed regardless of tumor grade with most of the cells of these poorly differentiated cancers being ERβ-positive/ERα-negative (Collins et al. 2009). An inverse relationship between ERα and ERβ in endometrial cancers suggests ERβ antagonists may be useful therapeutics for ERβ-positive/ERα-negative/low cancers.

SNP analysis is commonly being used to analyze common genetic variations in patients with endometrial cancer. A positive association for the risk of cancer is seen in women who express ERα and CYP1B1 L432 V (Zhu et al. 2011). In a Swedish population, analysis of ERα gene promoter SNPs found five adjacent tagSNPs covering a 15-kb region at the 5′ end that associates with decreased risk for endometrial cancer; however, the variants did not associate with myometrial invasion or cancer survival rates (Einarsdottir et al. 2009). Polymorphic variations of ERα (rs2234670, rs2234693, and rs9340799) suggest decreased endometrial cancer risk most strongly with rs9340799 (OR = 0.75 for heterozygous and OR = 0.53 for homozygous) (Wedren et al. 2008). These results suggest that intronic variation in ERα, depending on the SNP, confers risk or protection to endometrial cancer.

Endometriosis and ER

Endometriosis affects 10–14 % of reproductive-aged women with symptoms such as dysmenorrhea, dyspareunia, and infertility. This number increases to 35–50 % in women with pelvic pain, infertility, or both (Galle 1989; Rawson 1991). Endometriosis is hypothesized to occur via retrograde menstruation, but because this occurs in greater than 90 % women, the lower incidence (10–14 %) of endometriosis suggests additional elements impact its etiology (Giudice and Kao 2004; Bulun et al. 2005). Endometriosis is hormonally responsive, and therapies for endometriosis aim to decrease ovarian estrogen production and or counteract estrogen effects with the use of gonadotropin-releasing hormone (GnRH) agonists, progestins (including oral contraceptives), and androgens (Giudice and Kao 2004). Endometriotic lesions are found throughout the peritoneal cavity attached to the ovaries, fallopian tubes, anterior and posterior cul-de-sac of the uterus, and surrounding ligaments (Giudice and Kao 2004). A definitive diagnosis of endometriosis is only made by positive histological evaluation of endometriotic lesions after laproscopic surgery and treatments only suppress disease symptoms; consequently, elucidating the cause of endometriosis is of critical importance. Current hypotheses regarding disease pathogenesis and pathophysiology suggest inherent defects in the immune system, the uterus, the endometrium, or even the peritoneal environment of women with endometriosis (Taylor et al. 1999; Sharpe-Timms 2001; Bischoff and Simpson 2004; Capellino et al. 2006; Bukulmez et al. 2008; Weiss et al. 2009; Kulak et al. 2011).

Peritoneal macrophages from women with endometriosis have increased inflammatory cytokines, which are associated with increased ERα expression (Montagna et al. 2008). Conversely, ERβ expression correlates with an increase of inflammatory cytokines in women with and without endometriosis (Montagna et al. 2008). Along with the altered ratio of ERβ to ERα mRNA, the inflammatory status in women with endometriosis also affects the aromatase levels in eutopic and ectopic endometrium (Bukulmez et al. 2008; Juhasz-Boss et al. 2011). The aforementioned studies suggest that paracrine factors are involved in disease regulation; therefore, a baboon model of endometriosis was used to address this hypothesis. In the baboon, aberrant distribution of ERα, ERβ, PR, and FKBP52 is seen in the eutopic endometrium and in the stromal cells, ERβ decreases in both glandular epithelium and stromal cells (Jackson et al. 2007). Further studies need to be done to determine the impact of paracrine signaling and if the signaling occurs from both eutopic and ectopic tissue or if one tissue type elicits stronger paracrine signals in the pathogenesis of endometriosis.

Ovarian endometriosis is found on the ovarian fossa, and studies have focused on alterations in ovarian function and infertility. Decreased fertility rates seen in women with endometriosis may relate to endometriosis found on the ovarian fossa. Granulosa cells from women with ovarian endometriosis have higher PR-A and ERα gene expression, lower-quality embryos, and decreased pregnancy rates than women with tubal infertility (Karita et al. 2011). A prospective study of ovarian endometriosis found increased levels of ERβ are associated with reductions in ectopic tissue expression of ERα, estrogen-related receptor α (ERRα) and ets-related gene (Erg) (Cavallini et al. 2011). They suggest from their findings that the up-regulation of ERβ is the principal ER involved in the progression of endometriosis. Another study focused on ovarian endometrioma samples find enzymes involved in estradiol formation and in progesterone inactivation are up-regulated, potentially resulting in increased local levels of mitogenic estradiol and decreased levels of protective progesterone (Smuc et al. 2009). Additionally, baboons with spontaneous endometriosis exhibit progesterone resistance, and the normal antagonism of progesterone on ERα is absent in animals with endometriosis (Wang et al. 2009). To address the complex nature of ovarian endometriosis, low-density array analysis was done on ovarian endometriosis samples. The results revealed that a number of differentially regulated genes are estrogen related with a large proportion being in the “secreted” and “extracellular region” categories (Vouk et al. 2011). These studies indicate that both ERs and alterations in hormonal milieu alter the paracrine environment and play a role in endometriosis pathogenesis.

Endometriosis and SNPs and methylation

A number of research groups have evaluated ERα or ERβ SNPs and the association with endometriosis. Examination of the RsaI polymorphism of the ERβ gene found a nine times higher frequency for the AG polymorphism genotype in the patients with endometriosis compared to controls (Silva et al. 2011). The ERβ gene +1730 G/A polymorphism is associated with the risk of infertility and/or endometriosis-associated infertility and a predisposing factor to endometriosis (Bianco et al. 2009; Zulli et al. 2010). In a Korean population, the ERβ gene +1730 G/A polymorphism was not associated with endometriosis (Lee et al. 2007). SNPs associated with ERα (rs2234693–T/C SNP, dinucleotide (TA)(n) repeat), and ERβ (dinucleotide (CA)(n) repeat) revealed ERα longer (TA)(n) repeats correlated with susceptibility to stage I–II endometriosis, and ERβ shorter (CA)(n) repeats were linked with endometriosis without infertility (Lamp et al. 2011). Two independent studies have confirmed that the ERα gene-397 (T/C) PvuII polymorphism predisposes women to approximately 2.6-fold increase in endometriosis (Hsieh et al. 2007; Govindan et al. 2009). Evaluation of intron 1 and exon 1 ERα gene polymorphisms was not associated with increased risk for endometriosis (Sato et al. 2008). Together, these data suggest that ER SNPs, specifically SNPs for ERβ, are associated with increased susceptibility for endometriosis.

ERβ promoter methylation in the CpG islands from −197 to +359 shows significantly higher methylation in primary endometrial cells versus endometriotic cells, suggesting that methylation status is responsible for differential expression of ERβ in endometriosis and endometrium (Xue et al. 2007). The same group demonstrates increased ERβ levels concomitant with lower ERα levels in endometriosis and experimentally demonstrates that knockdown of ERβ increases ERα while overexpression of ERβ decreases ERα levels (Trukhacheva et al. 2009). Thus, ERβ may serve as a therapeutic target for the treatment for endometriosis.

Fibroids and ER

Uterine leiomyomas or fibroids are the most common uterine tumors in women and occur during childbearing years. Fortunately, fibroids rarely progress to leiomyosarcomas (Strissel et al. 2007). They are more common in African–American women than Caucasian women (Peddada et al. 2008). Fibroids are often asymptomatic, but symptoms can include bleeding between menstrual cycles, menorrhagia, pelvic cramping with menstruation, and/or sensations of fullness or pressure in the lower abdomen. Experimental studies are being done to determine what stimulates fibroid growth. Ishikawa et al. (2010) grafted human uterine leiomyomata tissue beneath the renal capsule of immunodeficient mice and found that the fibroid tissue increases in size with estrogen plus progesterone and estrogen alone, but decreases in size with anti-progestin and progesterone withdrawal. The cause of fibroids is unknown, but their growth coincides with hormones and menstruation. Fibroids within the same woman often have different growth rates despite being in the same hormonal milieu demonstrating the difficulty in the study of fibroids (Peddada et al. 2008).

Recent studies on ERα or ERβ and fibroids have focused on receptor expression, SNPs, methylation, and phosphorylation. Fibroids express higher levels of ERα, ERβ, and PR than control myometrium, but postmenopausal fibroids have 2.5-fold higher ERβ expression with no difference in ERα or PR (Strissel et al. 2007; Bakas et al. 2008; Chakravarty et al. 2008). ERα is present in smooth muscle cells with a variation in subcellular localization, while ERβ is widely distributed in smooth muscle, endothelial, and connective tissue cells with nuclear location (Valladares et al. 2006). The link between uterine fibroids and increased ER expression compared to normal myometrium led to assessing polymorphisms and susceptibility to fibroids. ERα SNPs (rs9322331 and rs17847075) analyzed in Brazilian women do not differ in women with fibroids compared to controls (Villanova et al. 2006). In contrast, a study of Asian Indian women examining a T/C SNP in intron 1 and exon 2 boundary of ERα indicated a significant association of the C allele with endometriosis (OR = 2.66) and fibroids (OR = 2.08) (Govindan et al. 2009). Much more research is needed to understand the role ERs are playing in fibroids.

Phosphorylation and methylation status impacts ERα activity. The expression of ERα-phospho-S118 by p44/42 MAPK from fibroids taken during the proliferative phase is increased, and these fibroids exhibit higher PCNA expression compared to patient-matched myometria and secretory-phase fibroids (Hermon et al. 2008). Asada et al. investigated the DNA methylation status of ERα promoter region (–1188 to +229 bp) and indicate hypomethylation of the CpG sites in leiomyomas coincides with increased ERα mRNA levels (Asada et al. 2008). Further studies focusing on the hormonal responsivity, SNPs, and methylation status in fibroids are needed to draw full conclusions, but ERα appears to be a main factor in fibroids.

Cardiovascular disease and ER

Cardiovascular disease in women following menopause increases to the same levels found in men, suggesting a cardioprotective role for estrogens. However, hormone replacement therapy was found not to be cardioprotective in postmenopausal women (Rossouw et al. 2002, 2007; Farquhar et al. 2005; Manson et al. 2007). Much of the research published in the last 5 years regarding ERs and cardiovascular disease has been in animal models. We will first summarize the human studies and then focus on key animal data. Kjaergaard et al. examined whether the ERα IVS1-397T/C polymorphism affects high-density lipoprotein cholesterol response to hormone replacement therapy and the risk of cardiovascular disease, cancer of reproductive organs, and hip fracture (Kjaergaard et al. 2007). From 9,244 Danish individuals followed for 23–25 years, they concluded that the ERα IVS1-397T/C polymorphism does not contribute to disease. Both ERα and ERβ are present in the vasculature and in the human heart. ERα and ERβ mRNA expression from left ventricular specimens from patients with coronary heart disease and dilated cardiomyopathy are lower when compared to healthy heart donors (Leibetseder et al. 2010). Analysis of time dependency of receptor expression showed an 8-h period rhythm for ERβ in patients with dilated cardiomyopathy that was not observed in coronary heart disease patients. The increased ERα and ERβ mRNA expression in left ventricular specimens from patients with coronary heart disease and dilated cardiomyopathy may reflect a compensatory mechanism by the ERs to counteract the decline in ventricular function seen with these diseases (Leibetseder et al. 2010).

Animal studies for cardiovascular disease and ERs mainly focused on cardioprotection or ER’s crosstalk with other transcription factors. Estrogen is known to confer cardioprotection, but this protective role has been challenged due to numerous clinical trials demonstrating that hormone replacement therapy has negative cardiovascular consequences (as reviewed in our previous article Deroo and Korach 2006). Nevertheless, understanding the mechanistic role of ER in cardiovascular disease is important. Selectively, knocking out ERα reduces estrogen’s protection by increasing atherosclerotic plaques and serum cholesterol levels in ApoE mice (reviewed in Mendelsohn 2000; Rossouw et al. 2007). Loss of ERα specifically in coronary endothelial cells abolishes the protective action of estrogen to limit infarct size and coronary endothelial function (Favre et al. 2010). The loss of endothelial ERα suggests a critical role for ERα in the estrogen-induced prevention of endothelial dysfunction after ischemia/reperfusion. Three studies demonstrate both ERs contribute to the pro-angiogenesis effects of estrogen replacement therapy and can promote the mobilization and homing of bone marrow stem cells into the myocardium to preserve cardiac function after myocardial infraction (Hamada et al. 2006; Chen et al. 2009; Bolego et al. 2010). The mode of estrogen’s influences on vascular endothelial growth factor and its signaling machinery (VEGF receptors, Akt and eNOS) to affect the development of coronary microvasculature in the heart is via ERα, as wild type and ERβ knockout mice do not have a marked decrease in coronary capillary density like the ERα knockout mice (Jesmin et al. 2010). In volume-overloaded hearts of ovariectomized rats, estrogen treatment improves the tissue inhibitor of metalloproteinases (TMIP-2)/MMP-2 and TIMP-1/MMP-9 protein balance, restores ERα (not ERβ) expression, and prevents MMP-9 activation, perivascular collagen accumulation and development of heart failure (Voloshenyuk and Gardner 2010). In ovariectomized rats, estrogen replacement therapy increases ERα and ERβ expression relative to controls (Chen et al. 2009). The ERβ knockout mice have elevated blood pressure, and 8β-VE2, an ERβ selective ligand that does not promote uterine growth, lowers blood pressure in these mice superior to estradiol or the ERα selective agonist 16α-LE2 (Jazbutyte et al. 2008). ERβ’s involvement in myocardial infraction was demonstrated by treatment with DPN, an ERβ selective ligand, which increased myocardial functional recovery (Vornehm et al. 2009). The aforementioned animal studies would support a protective role for the ERs and agonist treatment in cardiovascular disease. Based on experimental studies, estrogen’s protection on neural ischemia and peripheral vascular disease is mediated through ERα while myocardial protection works through ERβ. Due to the complexity of human disease, future studies will be needed in animal models to recapitulate the numerous aspects affecting cardiovascular disease.

Concluding remarks

The loss or overexpression of the ERs contributes to disease development and progression. The focus of research over the last 5 years has focused mostly on ERs and cancer, endometriosis, and fibroids. Less human research has been done with cardiovascular disease, stroke, and Alzheimer’s disease due to the negative findings of the Women’s Health Initiative (Rossouw et al. 2002). Nonetheless, in the clinical setting, endocrine therapy with ER agonists and antagonists is often used for treatment. ER antagonists are critical first-line treatments for hormone-positive cancers and hormonal control of the menstrual cycle often decreases the symptoms of endometriosis. Available ER antagonists often have off target effects; therefore, much work has been done and needs to be done to define tissue-specific selective estrogen receptor modulators (SERMs). More importantly, understanding the basic mechanisms of how SERM’s work in relation to tissue selective actions and the responses of estrogen and ER will be a benefit toward developing more effective therapeutics. Moreover, the further knowledge of the disease hormone receptor expression levels, localization, and post-translational modifications will allow for targeted therapies by SERMs. For instance, detailed hormone receptor status, interacting partners, and post-translational modification of breast cancer reveal survival outcomes and these cancer subtypes may someday be targeted specifically to increase disease survival. Increasing our understanding of the mechanisms of action of ER in normal and disease states will hopefully lead to further treatments and increased disease survival or treatment for symptoms.

Abbreviations

ER

Estrogen receptor

PR

Progesterone receptor

AR

Androgen receptor

DPN

Diarylpropionitrile

MAPK

Mitogen-activated protein kinase

HER2

Human epidermal growth factor receptor 2

UTR

Untranslated region

SNP

Single-nucleotide polymorphism

miRNA

MicroRNA

FOX

Forkhead binding protein