To β or Not To β: Estrogen Receptors and Ovarian Function

Much of what we know about hormone action in the vertebrate endocrine system comes from studies of estrogen and its signaling pathways, and, in turn, a good deal of this information derives from studies of the ovary. The identification of an estrogenic activity produced in the ovary dates to the studies of Allen and Doisy in 1923 (1) and eventually led to the crystallization of estrone in 1929 (2). The ability to prepare labeled [H]estradiol allowed the identification of specific target tissues, solidified the receptor concept, and was crucial to the identification of the estrogen receptor, the first hormone receptor (3, 4). Even at these early stages, it was recognized not only that estrogens are made in the ovary, but that the ovary is itself a target for estrogen action (5, 6). With the identification of a second estrogen receptor, ER , in 1996 (7) and the finding that this form predominates over ER in the ovary (8, 9), new interest emerged in the intraovarian roles of estrogen and the estrogen receptors that mediate these effects. Definitive evidence regarding the importance and relative contributions of the two estrogen receptors in the ovary began to accumulate after the targeted disruption of estrogen receptors (10), (11, 12), or the combination (12, 13) in mice. The story of estrogen, its receptors, and the ovary continues to be written with the study by Couse et al. (14), which provides new insights into the ovarian molecular phenotypes of mice lacking each estrogen receptor, in this issue of Endocrinology. In many respects, the ER knockout mice ( ERKO) have the most severe ovarian phenotype, in which follicles fail to mature or ovulate and form hemorrhagic cysts, leading to infertility (10, 12). However, subsequent studies revealed that much of the ERKO ovarian phenotype could be explained by the lack of estrogen-mediated negative feedback on pituitary LH secretion, resulting in chronically elevated LH levels and enhanced ovarian steroidogenesis (15, 16). Indeed, chronic treatment with a GnRH antagonist to suppress LH reverses the cystic ovarian phenotype, and immature ERKO mice can be successfully ovulated with exogenous gonadotropins before the onset of overt LH hypersection (17), although in the ER knockout model generated by Dupont et al. (12) no ovulations in response to exogenous gonadotropins were observed. In contrast to the ERKO phenotype, gonadotropins and steroids are largely normal in ERKO mice (16), implicating ER as mediating most of the negative feedback effects of estrogens on pituitary gonadotropin secretion. The ERKO mouse ovaries appear grossly normal, with follicles at all stages of development but fewer corpora lutea. In agreement with this finding, these female mice exhibit reduced fertility, although there appears to be a range that extends from mild subfertility to complete infertility (11, 12). The ERKO mice also fail to respond to exogenous gonadotropins and exhibit mature follicles containing “trapped” oocytes, suggesting a deficiency in the response to the LH/ human chorionic gonadotropin (hCG) ovulatory stimulus. Not surprisingly, the combined ERKO mice are also infertile and have the attenuated folliculogenesis and anovulatory phenotype characteristic of the ERKO mice (12, 13). However, an unexpected finding was the apparent transdifferentiation of granulosa cells toward a male Sertoli cell phenotype in the ERKO mice, a phenotype that is also seen in mice with a targeted disruption of the cyp19 gene encoding aromatase, the critical enzyme in estrogen biosynthesis, generating a complete estrogen-deficient state (18–20). All of this work is suggestive of a primary role for ER in regulating follicle development and ovulation but does not provide for a complete mechanistic understanding of the roles or targets of ER in the ovary. The studies by Couse et al. (14) in this issue of Endocrinology begin to address these issues with a detailed analysis of the molecular phenotypes of the ERKO and ERKO mice in the preand periovulatory period. The overall picture that emerges is that the predominant ER -expressing cells of the ovary, the granulosa cells, exhibit an attenuated response to FSH (or pregnant mare serum gonadotropin, PMSG) in the ERKO ovary and, as a consequence, are not able to respond appropriately to the LH (hCG) stimulus to initiate cumulus expansion, follicle rupture, and ovulation. Supporting these conclusions regarding aberrant PMSG-induced granulosa cell differentiation are the findings that two classical markers of granulosa cell differentiation, the LH receptor and aromatase, show attenuated or delayed expression in the ERKO ovary. Although overall ovarian LH receptor mRNA levels are more or less normal, there is a selective failure of PMSG-induced LH receptor expression in the granulosa cells, and aromatase induction is delayed and does not occur until after the hCG stimulus, resulting in reduced estradiol secretion in the preovulatory period. Thus, the follicle is delayed in its maturation and is not poised to respond to the ovulatory stimulus. As a consequence, several key genes known to be critical for ovulation, including prostaglandin synthase 2 (21) and the progesterone receptor (22), fail to be appropriately induced by LH-hCG in the ERKO mouse ovaries. Morphological examination of the ERKO ovaries points to a failure of LH to induce expansion of the cumulus-oocyte complex in a subset of the preovulatory follicles, consistent with attenuated LH action and a periovulatory defect in the follicle. It is curious that only some follicles exhibit the reduced cumulusoocyte complex expansion, and it seems reasonable to specAbbreviations: ER, Estrogen receptor; ERKO, ER knockout; hCG, human chorionic gonadotropin; PMSG, pregnant mare serum gonadotropin.