Estrogen hormone physiology: reproductive findings from estrogen receptor mutant mice.

Estrogen is a well-known female steroid hormone synthesized from the ovary that controls the estrous or menstrual cycle in females, therefore estrogen is imperative for female reproduction. Estrogen is not only important in female reproduction but also in male reproduction and in numerous other systems including the neuroendocrine, skeletal and immune systems in males and females. Along with the influence of estrogen on many physiological processes, it is also implicated in many different diseases including obesity, metabolic disorder, cancer, osteoporosis, endometriosis and fibroids [ 1 , 2 ]. The predominant mechanism of estrogen action is through nuclear estrogen receptor (ER) expression in estrogen target organs [ 3 ]. The biological effects of estrogen are mediated through two distinct ER proteins, ERα and ERβ (ERs). These receptors are encoded from separate genes on different chromosomes and the expression profiles in the tissues are different. The predominant expression tissues for ERα include: liver, uterus, mammary gland, pituitary, hypothalamus, cervix, and vagina. ERβ expression is more limited and predominant expression tissues include: ovary, lung, and prostate [ 4 ]. Controlling the functions of ERα and ERβ is the basis for many therapeutic interventions to estrogen related diseases. Thus it is quite important to reveal the physiological role of ERα and ERβ in the tissues and the in vivo functionality of ER proteins.

The ERs are comprised of six structural domains: an amino-terminal domain (A/B-domain), a DNA binding domain (DBD; C-domain), a hinge region (D-domain), a ligand-binding domain (LBD; E-domain), and a carboxyl-terminal domain (F-domain) [ 3 ]. The C and E domains carry a high-degree of homology between ERα and ERβ; however the A/B, D and F domains are divergent [ 5 , 6 ]. The A/B-domain contains the transcription activation function 1 (AF-1) which is reported to be important for ligand-independent transactivation [ 7 ]. The LBD or E-domain of ERs contains the transcription activation function 2 (AF-2) that is important in ligand-dependent transcriptional regulation [ 7 ]. The helix 12 in the LBD is the core of AF-2 and the configuration of helix 12 is changed by ligands to either active (agonist bound) or inactive (antagonist bound) forms for transcription regulation [ 8 , 9 ]. Multiple ER-mediated transcription regulation mechanisms have been characterized. In the “classical” mechanism of estrogen action, hormone bound ERs directly bind to a specific DNA sequence called the estrogen responsive element (ERE) through the DBD or C-domain of ER [ 10 ]. Estrogen can modulate other transcription factor functions through tethered ER with other transcription factors such as c-Jun and Sp1. In the “tethered” mechanism, ER does not bind to a DNA element directly but forms a transcription activation complex on the AP-1 and Sp1 responsive elements [ 11 , 12 ]. It is also known that growth factors can activate estrogen-independent ERα mediated transcription and the AF-1 has an important role in the growth factor mediated ERα activation [ 13 ]. Furthermore, while ERα is a well-characterized nuclear transcription regulator, it is believed that the ERα protein is also involved in extranuclear non-genomic signal transduction [ 14 , 15 ].

The development of multiple genetic models has led to an increase in understanding and knowledge of the physiological roles of estrogen receptors. These models include mice lacking functional ERα (αERKO), ERβ (βERKO) or both estrogen receptors (αβERKO). Several methods have been utilized to generate the ERKO mice. Currently, the cre-loxP system is the most widely used method to create whole tissue KO or tissue selective KO mice. The first αERKO mouse ( Esr1 tm1KSK ) was generated by an insertion method to disrupt ERα expression. Namely, a neo expression cassette was inserted into exon 2 to create the frame shift mutation, resulting in the disruption of functional ERα protein expression [ 16 ]. Unexpectedly, the Esr1 tm1KSK mouse expresses a trace level of N-terminal truncated mutant ERα (E1-ERα) which is generated by non-controlled alternative splicing [ 17 ]. We have reevaluated the ERα function using the Ex3αERKO ( Esr1 tm4.2KSK ) mice, in which exon 3 is excluded by the cre-loxP system to disrupt the functional ERα expression and the major phenotypes of both ERα null strains are nearly identical [ 18 , 19 ]. In our preliminary results, there are some differences between the two αERKO mouse lines, suggesting that the E1-ERα variant might have affected some physiological responses in the first αERKO mouse ( Esr1 tm1KSK ) line [ 18 , 19 ]. Estrogens stimulate the process of uterine epithelial proliferation and differentiation that is necessary for establishment and maintenance of pregnancy. Progesterone and many growth factors are also important for these processes. Establishment of a model with a null ERα (αERKO) has given us a model to determine and verify the physiological roles of ERα. Because ERα is the predominant factor for estrogen action in the adult mouse uterus, adult αERKO mice are infertile and have hypoplastic uteri [ 4 ]. A well-established model to examine uterine estrogenic response is to treat the ovariectomized (OVX) mouse with an agonist (i.e. estradiol or DES). In response to agonist treatment in a 3-day bioassay, uterine wet weight and the expression of estrogen responsive genes are induced by agonists in the wild type (WT) mouse. In contrast, αERKO uteri show neither uterine stimulation nor increased expression of estrogen-responsive genes [ 17 ]. αERKO uteri express residual progesterone receptor (PR) levels and respond to progesterone mediated gene stimulation but PR is not inducible and does not support embryo implantation [ 20 ]. αERKO mice have hemorrhagic cystic ovaries which are due to constitutively high serum luteinizing hormone (LH) levels causing the disruption of negative feedback regulation in the hypothalamic–pituitary–gonadal (HPG) axis. The treatment of an antagonist (antide) for gonadotropin releasing hormone reduces serum LH levels and prevents cyst formation in the αERKO ovaries [ 21 ]. The expression levels of several steroidogenic enzymes, including Cyp17a1 , Cyp19 , and HSD17 β 3 are elevated in the αERKO ovary. HSD17 β 3 is a testicular steroidogenic enzyme not observed in the WT ovary but is aberrantly expressed in the αERKO ovary [ 21 ]. αERKO females have elevated estradiol and androstendione levels in their serum and also have an elevated testosterone level [ 21 ]. Mammary gland development in the αERKO female is impaired and remains rudimentary after puberty when normal WT mammary gland development expands [ 18 ]. Adult male mice lacking ERα are infertile with lower sperm count and motility than WT mice and exhibit dilated seminiferous tubules [ 19 ]. The dilation is presumably due to loss of expression of efferent ductal testicular fluid reabsorption related proteins such as sodium–hydrogen exchanger 3 (NHE3/Slc9a3) and carbonic anhydrase 2 (Car2) [ 22 , 23 ]. ERα is not required by the male germ cells; demonstrated by transplantation experiments where male germ cells were put into germ cell-depleted WT testes and the recipients were able to sire offspring from the transplanted αERKO germ cells [ 24 , 25 ]. The serum hormone levels in male αERKO mice are also disrupted with elevated testosterone and LH levels [ 19 ]. Female βERKO mice are subfertile in continuous mating studies and have reduced litter sizes when compared to WT mice [ 26 ]. Mouse uteri have very low expression of ERβ and the uteri of βERKO mice are indistinguishable from WT with normal organization and development [ 4 , 27 ]. βERKO uteri show a comparable response to WT when challenged with estradiol in a 3-day bioassay [ 27 ]. ERβ is predominantly expressed in the granulosa cells of the ovarian follicles and βERKO mice have defects resulting in an inefficient and infrequent ovulatory response [ 27 , 28 ]. This deficiency is highlighted by the appearance of unruptured follicles following superovulation and a lower number of recovered oocytes after superovulation when compared to WT mice [ 26 ]. Granulosa cells of the βERKO mice did not differentiate fully after PMSG treatment and had reduced LH receptor expression, aromatase activity and estradiol synthesis [ 29 ]. These factors result in a deficient ovulatory response to gonadotropin treatment and failure of the cumulus–oocyte complex to expand fully [ 29 ]. In vitro studies of βERKO granulosa cells showed that ERβ is involved in induction of the cAMP pathway and this disruption may cause ovulation and fertility defects [ 29 ]. Male βERKO mice are fertile and overall exhibit a phenotype widely similar to WT littermates [ 27 , 28 ]. Mice lacking either ERα or ERβ have been instrumental in the study of estrogen action, however using these models does not account for the possible compensatory mechanisms provided by the opposite ER in each ERKO model. Therefore, mice lacking both ERs (αβERKO) were generated to study the potential compensatory and cooperative actions of ERα and ERβ. Mice lacking both ERα and ERβ display a phenotype similar to the αERKO mice. Females are infertile, have an altered endocrine environment, and exhibit deficient ovulation and male mice have dilated seminiferous tubules and deficient sperm [ 26 , 28 , 30 ]. The similar phenotype between the αERKO and αβERKO mice suggests that ERβ is not able to compensate for ERα functions in the single ERα knockout. However, the compensational function of ERα in the single ERβ knockout is still debatable. A very unique feature of the compound ER knockout adult female is the presence of seminiferous-like tubules in the ovary which develop post-pubertally and are not present in the ovary of either single ER knockout [ 28 , 30 ]. The cells found in these structures have features highly similar to Sertoli cells which are normally found only in the testes [ 28 , 30 ]. Along with physical commonality between the structures, the Sertoli-like cells found in the αβERKO aberrantly express genes known to be involved in Sertoli cell differentiation such as Sry-related transcription factor (Sox9) [ 30 , 31 ]. Sox9 is required for normal Sertoli cell differentiation in the development of the testis in both humans and rodents [ 32 , 33 ]. These unique features suggest that the double knockout granulosa cells undergo trans-differentiation and the structures change from ovarian features into more male-like structures [ 30 , 31 ]. This phenomenon has never been observed in αERKO mouse suggesting that granulosa cell ERβ might be involved in the prevention of granulosa cell trans-differentiation. While sex reversal of the ovarian structure has been shown in other strains/species, this model is noteworthy in that the differentiation occurs postnatally in αβERKO mice [ 30 ].

Whole animal knockout models have allowed tremendous insight into the physiological functions of the ERs. However, in the global knockout models it is not possible to look at the role of the ERs in specific tissues. With the advancement of the cre-loxP technology, the floxed ERα line can be crossed with other cre lines to produce mice lacking functional ER in specific cell types or tissues. Multiple ERα mouse models using this technique have been developed and studied [ 34 – 39 ]. In this review, we will discuss mice lacking ERα in uterine epithelial cells, referred to as the UtEpiαERKO mouse. Because the uterus is a primary target tissue for estrogen and is composed of different cell types, tissue or cell-type specific knockouts of this organ are instrumental in deciphering the role of ERα in uterine biology. In order to look at the individual roles of ERα in the stromal cells versus epithelium of the mouse uterus, a uterine epithelial-specific ERα knockout (UtEpiαERKO) mouse line was generated by crossing the floxed Esr1 mice ( Esr1 tm4.1KSK ) with Wnt7a -Cre mice [ 34 ]. The UtEpiαERKO female mice are infertile however they exhibit regular estrous cycles and ovaries possess all stages of follicular development including corpus lutea, indicating ovulation. Implantation was not observed in UtEpiαERKO uteri after either natural mating or embryo transfer suggesting ERα in the uterine epithelium is needed for embryo receptivity. Three days of estradiol treatment to the OVX UtEpiαERKO mice induced uterine wet weight and epithelium proliferation suggesting that ERα in the uterine stromal cells, but not the epithelium, regulates uterine epithelial cell proliferation. While estrogen induced uterine epithelium proliferation in UtEpiαERKO mice, apoptosis in the epithelial cells was significantly increased when compared to WT. Taken together these observations suggest that epithelial ERα is needed for proper endometrial epithelial cell functions.

It is impossible to examine the physiological role of ER functional domains using the ERKO mice since no protein is made. Therefore, mouse models have been made with genetically modified estrogen receptor domains to allow dissection of the physiological function of the ER domains. These include models with altered DNA binding domains (NERKI and EAAE) which disrupt binding to the estrogen responsive DNA element, as well as one model with a point mutation in the ligand binding domain (ENERKI) which disrupts estradiol mediated transcription activation by inhibiting estradiol binding [ 40 – 42 ]. Furthermore, mouse models have been made with the deletion of AF-1 or AF-2 (AF-1 0 and AF-2 0 ) and point mutations in the core AF-2 (AF2ERKI) [ 43 , 44 ]. In order to examine the physiological function of AF-1 and AF-2 of ERα, our laboratory established a knock-in mouse model (AF2ERKI, Esr1 tm3.1KSK ) that contains two point mutations in the AF-2 region, disrupting the AF-2 mediated transactivation [ 45 ]. Even though the AF2ER mutant ERα protein is expressed in AF2ERKI mice, the phenotypes of the AF2ERKI male and female are indistinguishable from αERKO mice, suggesting that the AF-2 is indispensable for ERα-mediated physiological responses. Growth factors, such as insulin like growth factor 1 (Igf1) and EGF are able to stimulate uterine cell proliferation without estrogen in OVX WT mice but not in the αERKO uteri [ 18 , 46 , 47 ]. This observation suggests that the growth factor dependent phosphorylation signals activate “ligand-independent” transcription through ERα in vivo [ 13 , 47 ]. However, Igf1 and EGF did not induce the uterine cell proliferation in the AF2ERKI [ 45 ], suggesting that the AF-1 activity of ERα is regulated by the “ligand-dependent” functional domain, AF-2. One unique characteristic of the AF2ERKI mouse model is that traditional antagonists such as fulvestrant/ICI182780 (ICI) and tamoxifen (Tam) act as agonists through the AF2ER mutant ERα [ 45 ]. Our in vitro experiments suggest that the AF-1 activity is necessary for the antagonist dependent AF2ER activation [ 45 ]. In a 3-day bioassay, ICI and Tam stimulated uterine growth and gene expression in a comparable way to WT treated with E2 [ 45 ]. In contrast, E2 induced prolactin ( Prl ) gene expression in the OVX WT mouse pituitary, however, ICI and Tam treatment were not able to stimulate Prl gene expression in the AF2ERKI pituitary [ 45 ]. These results indicate that the AF-1 functionality may be different between the uterine tissue and the pituitary. Furthermore, we were able to show that ER activation by Tam treatment to the male AF2ERKI was sufficient to partially rescue male fertility [ 48 ]. Thus the AF2ERKI mouse is a useful tool in determining the differential roles of AF-1 and AF-2 of ERα in the estrogen target tissues and hormone responsiveness.

Multiple mouse models with different mutations to the ERs have allowed great advances in deciphering the role of estrogen receptor in hormonal physiology as summarized in Table 1 (female) and Table 2 (male). Generation of ERKO mice demonstrated that loss of function through ER gene mutations is not lethal and phenotypes of these lines have given insight into the roles of the individual estrogen receptors. In particular, we now have advanced knowledge of the role and function of each estrogen receptor in female reproduction and other physiological processes which has led to therapeutic advances in estrogen receptor related diseases.