Estrogen Receptor β: Switching to a New Partner and Escaping from Estrogen

See allHide authors and affiliations

Science Signaling  12 Apr 2011:
Vol. 4, Issue 168, pp. pe19
DOI: 10.1126/scisignal.2001991


Estrogen receptor (ER) β, the “second” ER, plays a gatekeeper role by inhibiting cell proliferation, promoting apoptosis, and impeding the progression of prostate cancer. Ironically, its presumed ligand, 17β-estradiol, promotes cancer development in experimental models. The mechanisms underlying the interplay between estrogens and ERβ in prostate cancer remain largely unclear. Research on a previously unknown tethering partner of ERβ, Krüppel-like zinc finger transcription factor 5 (KLF5), and its downstream gene target (FOXO1) helps to unlock this puzzle. 17β-Estradiol is not required to maintain the tumor-suppressive function of ERβ in the prostate, a tissue with limited estrogen availability; moreover, the presence of 17β-estradiol abrogates ERβ- and KLF5-mediated signaling and promotes cellular proliferation. Future research into ERβ will likely involve this estrogen independency and the preference for binding nonclassical DNA elements through tethering. The development of ERβ-based therapies may lead to improved drug efficacy.

The nomenclature of a hormone nuclear receptor is traditionally defined by the type of ligand with which it interacts. ERα and ERβ were named on the basis of their ability to use 17β-estradiol (E2) as an endogenous ligand and their high sequence homology (1, 2). At the molecular level, the two ERs share similar DNA binding elements, co-regulators, co-mediators, and transcription machinery, and they even regulate a similar set of genes (3, 4). So far, our understanding of ERβ has been largely derived from data obtained on ERα. However, with the advancement of whole-genome screening techniques, we now know that some genes are predominantly controlled by ERβ (5). But why do we need two receptors responding to one ligand? This could be a mechanism to increase the functional diversity of estrogen signaling. For example, in tissues in which both ERα and ERβ are present, ERβ acts as a yin-yang partner to oppose the activities of ERα, possibly through heterodimerization with ERβ (6, 7). But what would happen when ERβ is present by itself, such as in the prostate epithelium?

Various clues have emerged in support of a tumor-suppressive function for ERβ in prostate carcinogenesis. First, expression of the gene encoding ERβ is gradually shut down in more aggressive, high–Gleason grade tumors through DNA methylation of the 5′ proximal promoter region (8). Second, ERβ-knockout mice exhibit epithelial hyperplasia in prostate epithelium (9). Furthermore, ectopic expression of ERβ in cells inhibits prostate cancer cell growth (10, 11), migration (12), and invasion (13), as well as suppressing epithelial-mesenchymal transition in prostate cells (14). Because the incidence of prostate cancer in humans increases with age, it is imperative to consider the role of estrogens in this process (15, 16). Owing to the gradual decline in testosterone production (17) and the increased in situ production of E2 (18), estrogen dominance increases with age. Moreover, in rodent models, estrogens are crucial risk factors for prostate carcinogenesis (1922). This leads us to two important questions: How do estrogen concentrations and ERβ abundance influence the risk of developing prostate cancer, and how does ERβ contribute to antiestrogen therapies for this cancer? This major gap in knowledge—how estrogens and ERβ interact with other factors to influence prostate carcinogenesis—remains to be filled.

An elegant study from the Yanagisawa group (23) reveals a potential mechanism that might fill the gap in data pertaining to how ERβ functions as a tumor suppressor in the prostate. They first discovered that the classical antiestrogen ICI 182,780 (ICI) suppresses human prostate cancer xenograft growth. ERβ functioned as a proapoptotic factor by maintaining FOXO1 expression through tethering on the FOXO1 promoter by Krüppel-like zinc finger transcription factor 5 (KLF5). Their data also showed that, in a cellular environment lacking E2 or with low E2 concentrations, ERβ still induced apoptosis by the KLF5 tethering mechanism. These findings contrast with the traditional view that E2 is needed in ERβ function. Even more unexpected is the finding that E2 not only has no effect on KLF5 tethering to the FOXO1 promoter but also promotes KLF5 protein degradation through recruitment of an E3 ubiquitin ligase, resulting in reduction of FOXO1 abundance. These observations are particularly germane in explaining the tumor-suppressive function of ERβ in the prostate—a tissue with low estrogen concentrations—because according to this report, the ERβ-KLF5-FOXO1 signaling pathway remains functional in promoting apoptosis of prostate cancer cells. These findings are clinically relevant because KLF5, ERβ, or FOXO1 alone, as well as KLF5 and ERβ together, were predictive of the prognosis for prostate cancer patients.

Their study challenges some preconceptions regarding the function of ERβ. It is generally believed that ERβ relies on estrogens for executing its tumor-suppressive and pro-apoptotic functions. More studies are needed to determine whether ERβ actually requires a ligand to maintain its basal function or whether E2 is an endogenous ligand that triggers negative feedback on ERβ downstream signaling. Until now, ERβ has been defined as a second estrogen receptor, and most of the data for ERβ derive from insightful findings on ERα in research on women’s health. With the availability of this new information, a reevaluation of various dogmas about ERβ action, especially in an estrogen-limited environment, is warranted.

One implication is that E2 may not be required for ERβ functions. Studies from various laboratories have shown that ERβ resides in the nucleus (24, 25) and can recruit co-regulators and activate gene transcription even in the absence of E2 (26). Furthermore, non-E2 ligands such as hormone metabolites (5α-androstane-3β, 17β-diol) (12, 27) may exist at higher concentrations than E2 and might be the preferred ligand for ERβ. Alternative ligands could be responsible for the tissue-specific action of ERβ. Therefore, data from studies of the endogenous action of ERβ should be interpreted with extra caution with respect to whether this action is E2-dependent or E2-independent.

Another implication is that tethering may be a primary mode for ERβ action and that the classical estrogen receptor–responsive element (ERE) may not be the preferred DNA binding site for ERβ. In support of this reasoning, data from a global profile of ERβ occupancy in a breast cancer cell line indicate that ERE is not the major DNA binding element for ERβ (28, 29). Other alternative ERβ binding sites such as AP1 (28), Sp1 (30), and NF-κB (30) have been documented. Perhaps antiestrogens, phytoestrogens, or endogenous ligands act as agonists to direct ERβ to tether to nonclassical ER binding sites so as to regulate the cell cycle, apoptosis, cell growth, and bone formation (Fig. 1). However, in an estrogen-dominated environment, ERβ may switch over to ERE-based gene transcription. This would be consistent with the finding that ERβ isoforms can selectively promote ERβ transactivation on an ERE-based pS2 promoter only in the presence of E2 (31). Also, this promoter-switching mechanism (Fig. 1) may provide insight into why binding of phytoestrogens, such as genistein and apigenin, to ERβ produces outcomes different from those of estrogens (32).

Fig. 1

Ligand-dependent switching of ERβ action. Traditionally, in an estrogen-rich environment, ERβ directly binds to the ERE-based promoters, recruits co-activators (CoA) and co-mediators (CoM), and activates a set of genes. Without estrogens, ERβ interacts with co-repressors (CoR), co-mediators, or both on classical ERE-based promoters and causes gene silencing. In an estrogen-limited environment (left panel), such as the prostate, ERβ is recruited to target gene promoters containing Ap1-, NF-κB–, or GC-box elements through a tethering mechanism and transcriptionally activates different sets of target genes. Ligands such as androgen metabolites, antiestrogens, or phytoestrogens may or may not be required in this mode of action. Furthermore, if E2 is present, it may impede the tethering-mediated action of ERβ by inducing ubiquitin-mediated degradation of its tethering partner (KLF5, bottom right) and diminishing transcriptional activity at gene promoters. The thickness of the bent arrow depicts the degree of promoter transactivation.


A final implication concerns the design of ERβ-based therapeutics. If E2 is not a preferred ligand for ERβ, drugs designed with a focus on the optimal conformation of E2-bound ERβ and the classical ERE pathway (33) may need to be critically reevaluated. The energy-minimized E2-bound model may favor more rapid ERβ turnover, whereas antiestrogens such as ICI may protect ERβ and its tethering partner from degradation, thus enhancing the stability of these proteins. This new approach could provide an important future direction for in silico drug design, particularly if the ERβ-based therapies are used for men’s health or for postmenopausal women.

ERβ research, after being under the shadow of ERα for more than a decade, deserves closer scrutiny in future investigations.


View Abstract

Stay Connected to Science Signaling

Navigate This Article