INTRODUCTION — Estrogens act most importantly on the reproductive organs, but they also act on other organ systems such as cardiovascular, skeletal, immune, gastrointestinal, and neural sites [1-4]. Examples of the sites of responses to estrogen and the clinical consequences for the activity of estrogen in both females and males are summarized in the figure (figure 1). Their major actions are primarily genomic, mediated by nuclear estrogen receptors (ERs), but they also have non-genomic actions.
The molecular mechanisms of estrogen action will be reviewed here; the physiological actions of estrogens and estrogen analogues (selective estrogen receptor modulators [SERMs]) are discussed separately. (See "Normal menstrual cycle" and "Mechanisms of action of selective estrogen receptor modulators and down-regulators" and "Estrogen and cognitive function".)
ESTROGEN RECEPTORS — The genomic actions of estrogens are mediated via estrogen receptors (ERs), which are proteins that bind estrogens with high affinity and specificity. These receptors are members of a family of nuclear hormone receptors that include receptors that bind other steroids, thyroid hormone, and retinoids, and receptors such as peroxisome proliferator-activated receptor (PPAR), farnesoid X receptor (FXR), and liver X receptor (LXR) that mediate metabolic processes [5], as well as many "orphan" receptors for which no ligands have been identified. All these receptors function as ligand-modulated nuclear transcription factors [6-9].
Two ER molecules have been identified: the original ER-alpha, and later the ER-beta [10,11]. Their structures are similar to those of the other members of this family of receptors [12,13]. The key components are the C or DNA-binding domain, which binds with high affinity and specificity to DNA sequences (estrogen response elements [EREs]) to regulate transcription rates of target genes, and the E or ligand-binding domain, which binds estrogens and estrogen analogues. (See "Mechanisms of action of selective estrogen receptor modulators and down-regulators".)
The consensus ERE is a 13-base-pair inverted-repeat DNA sequence (GGTCAnnnTGACC), to which dimers of ER complexes bind with high affinity and specificity [14,15], with one receptor molecule in contact with each five-base-pair segment of the response element [16]. Dimerization of ER complexes is facilitated by receptor binding to a response element. Additional sequences located in the ligand-binding domain of the receptor also are involved in dimerization [11,14], as demonstrated by the formation of homodimers by truncated receptors containing only a ligand-binding domain [17].
The ERs contain two gene-regulatory regions, termed activation functions (AF): AF-1, located toward the amino-terminal end of the receptor, acts independent of ligand, whereas AF-2, located in the ligand-binding domain, is ligand dependent [10,11]. Steroid receptor co-regulators (SRCs) [18-22] interact with ERs and function together with massive multimeric complexes that facilitate access to chromatin, and dynamically coordinate the steps of transcription, including initiation, elongation, termination, and clearing or turnover of the transcriptional modulators (figure 2) [23-25]. Regulation of access to chromatin is facilitated or impeded in part by modifications of histones, the proteins that "package" DNA. Acetylation of histone tails is mediated by histone acetyl transferases (HATs), which increase access, and deacetylation by histone de-acetylases (HDACs), which increase chromatin compaction. SRC itself has HAT activity, and ER-SRC interacts with other HATs and with HDACs. Targeting of HDAC with inhibitors is an emerging therapy in breast cancer [26] and other cancers [27].
When an estrogen or estrogen analogue reaches the cell nucleus and binds to an ER, the conformation of the ligand-binding domain of the receptor is altered, allowing interaction with coactivator molecules if the ligand is an agonist, but preventing this interaction if the ligand is an antagonist, ultimately altering transcriptional rates of estrogen-responsive genes (figure 2) [5,17,19,28].
Chromatin immunoprecipitation combined with next generation sequencing, called ChIP-seq, allows comprehensive analysis of all the sites of interaction between ER-alpha and DNA in a cell or tissue. These types of studies have indicated that most ER-alpha binding does not occur at gene promoters, but often in distant regions that act as transcriptional enhancers, and can be more than 100,000 base pairs from transcribed genes [29,30]. It was through this approach that FoxA1's role as a pioneer factor, facilitating ER-alpha binding in breast cancer cells, was discovered [31-33]. Although estrogen greatly increases ER-alpha recruitment to chromatin and is needed to regulate transcription of target genes, ER-alpha ChIP-seq analysis of mouse uterine tissue demonstrated that some ER-alpha is pre-bound to DNA in the absence of estrogen, with the number of sites per gene increased with hormone treatment [34].
Alternative mechanisms of action — In addition to the classic mechanism of ER function, ERs can sometimes regulate expression of genes that lack an estrogen-response element. As an example, in cultured cells, ERs can modulate the transcriptional activity of heterodimers of the transcription factors fos and jun, resulting in activation of reporter genes containing activator protein 1 (AP-1) elements [35]. In addition, complexes of ERs and specificity protein 1 (SP1) molecules can activate some reporter genes, such as those for c-FOS and transforming growth factor alpha, which lack a full estrogen-response element, by binding respectively to half sites of these response elements and guanine cytosine (GC)-rich sequences in the regulatory regions of these genes [13,36-38]. How much importance such gene regulatory mechanisms contribute to physiological responsiveness will require further studies with additional models. A mouse with ER-alpha mutations that prevent its interaction with DNA lack estrogen responses, including uterine and mammary gland development and growth [39,40]. In addition, these mice lack estrogen-induced gene regulation, indicating that ER-alpha DNA-binding independent mechanisms of response are not sufficient to support normal physiologic functions.
ERs can also function independently of estrogen. Specifically, both epidermal growth factor (EGF) and insulin-like growth factor-I (IGF-I), acting via their extracellular membrane bound receptors, can couple cell signaling mechanisms to stimulate transcription of ER target genes in the absence of estrogen [13,41-43]. Thus, cross-talk and signal amplification can be coupled between membrane-bound growth factor receptor pathways and nuclear receptors. Such coupling mechanisms have been a factor in breast cancer diagnosis and therapy [44]. Although data are limited, ER-alpha is believed to be an agonist and growth promoting, while ER-beta is felt to be inhibitory towards breast cancer advancement and possibly others [45,46].
NON-GENOMIC ACTIONS OF ESTROGENS — Estrogens bind with high affinity to other cell components, including plasma membranes [47,48]. Some effects of estrogen, such as rapid induction of mitogen-activated protein (MAP) kinase and extracellular signal-regulated kinase (ERK) pathways, appear to involve direct action of estrogen receptors (ERs) at the plasma membrane rather than genomic modulation [49-56].
As these rapid effects occur without ER-gene interaction, they are called "non-genomic," although the signals initiated by these mechanisms ultimately result in regulation of genes. These responses are observed in diverse tissues, including the cardiovascular system, central nervous system, and in breast cancer cells. Several models have been developed that describe mechanisms for these non-genomic effects [57-59].
In some cases, the same ER molecule (ER-alpha or ER-beta) that interacts with genes in the nucleus becomes associated with the plasma membrane by interacting with membrane proteins such as caveolin, growth factor receptor-binding proteins such as Shc, or with guanosine (G) nucleotide-binding proteins. Alternatively, a G protein-coupled receptor, GPR30 or GPER, may be activated at the membrane by estrogen [60-64]. The physiological significance of these rapid effects and how they are integrated with the nuclear responses to estrogen are important questions that are currently being investigated.
PHYSIOLOGIC ROLES OF ESTROGENS
Clinical examples of estrogen insensitivity — The ability to produce gene-targeted mice and rats in which the estrogen receptor (ER) genes or the gene for aromatase, the enzyme that converts androgens to estrogens, have been knocked out or altered to selectively disrupt signaling in specific cells or disable one aspect of estrogen signaling [39,65-86], and the identification of a limited number of patients with mutations in these genes, has allowed study of the physiological processes that require estrogen in vivo [87-90]. The mechanisms of ER-mediated regulation of target genes have been extensively characterized using isolated cells or cell components.
There is one clinical case of ER resistance in a male patient who harbored a mutation in the ESR1 gene resulting in a premature stop codon and no expression of ER protein. Clinical observations of this 28-year-old patient were similar to the experimental ER-knockout animal models [87-89]. The male patient was infertile, with low normal sperm counts and low sperm viability, and gonadotropin dysregulation with elevated serum estrogen levels. He also had open epiphyses, osteopenia, obesity, insulin resistance, and premature coronary artery disease, highlighting the importance of estrogen in male physiology. The patient was homozygous for this mutation, which was not spontaneous but resulted from a consanguineous marriage of parents who were second cousins [91]. Genetic analysis of the family members showed Mendelian distribution with two heterozygous sisters who were fertile and non-phenotypic, similar to the animal models.
More recent reports describe additional cases of both male and female individuals with mutations in ESR1. In one case, an 18-year-old female with estrogen insensitivity presented with bilateral intermittent pelvic pain, primary amenorrhea, absence of breast development, high serum estradiol (3500 pg/mL [12,850 pmol/L]), mildly elevated gonadotropin levels, and enlarged multicystic ovaries [90]. She had a point mutation in her ESR1 gene resulting in alteration of a glutamic acid residue in the ligand-binding domain that affects the ability of the ER-alpha protein to interact with coactivator proteins [92]. Unlike the original male patient who did not express ER protein, this patient expresses a mutated ESR1 protein that is functionally inactive, not due to altered hormone binding, but to compromised coactivator interactions. (See "Causes of primary amenorrhea", section on 'Estrogen resistance'.)
Estrogen insensitivity was also described in a consanguineous family with three affected siblings (one male and two females with similar clinical presentations to those described above) who carried a homozygous ESR1 mutation within the ligand-binding region that impairs the ability to bind estrogen and decreases estrogen responsive gene transcriptional activity [93]. The parents (who were first cousins), and one additional child, were heterozygous for the mutation and were unaffected. Mutations in the ESR2 gene have also been reported in both male and female patients [94,95]. These clinical cases of estrogen resistance and insensitivity indicate that mutations in the ESR1 or ESR2 gene are not lethal as was previously proposed [96].
As might be expected, the most important defects in these patients and in mouse models are in the reproductive system. Both male and female mice lacking the ER-alpha gene or ER-alpha and -beta genes are infertile, whereas in animals lacking only the ER-beta gene the males are fertile and the females are subfertile [97,98]. Both male and female ER-alpha knockout rats are infertile as well [99]. Mice lacking the membrane guanosine (G) nucleotide-binding protein-coupled receptor GPR30 are fertile, indicating that this receptor is not necessary for reproductive function [100]. In mice lacking aromatase, the males are infertile and the females are anovulatory [101,102].
Gene targeted models
Female reproduction — Deletion of ER-alpha or aromatase in female mice (or deletion of ER-alpha in rats) results in complete infertility. Deletion of ER-beta in mice results in subfertility, as a result of failure of several components necessary for successful reproduction. Normally, gonadotropin secretion by the pituitary is regulated by a negative feedback loop in which rising serum estrogen concentrations down-regulate transcription of the gene for the beta-subunit of LH, and withdrawal of estrogen by ovariectomy results in an increase in LH secretion. Thus, mice with deletion of ER-alpha or aromatase, rats with ER-alpha deletion, and patients lacking aromatase or ER-alpha have high serum LH concentrations. In contrast, serum LH concentrations are normal in ER-beta-deficient mice, indicating that the negative feedback is mediated by ER-alpha. Serum follicle-stimulating hormone (FSH) levels in the ER-alpha-deleted female mice and rats are normal, and FSH increases in wild type (WT) and ER-alpha-deleted female mice following ovariectomy [103]. Female mice with mutations that disrupt ER-alpha DNA-binding activity, signaling via activation function (AF)-1 or AF-2, or membrane signaling are also infertile, indicating the importance of all components of ER-alpha [39,80-86].
Ovary — Both ER-alpha and -beta are present in normal mouse ovaries, but their distribution varies. ER-beta is found predominantly in the granulosa cells of the follicles, whereas ER-alpha is found principally in the thecal and interstitial regions [67].
The ovaries of ER-alpha-deficient mice and rats develop hemorrhagic cysts and lack mature follicles and corpora lutea, indicating the absence of ovulation, but they secrete increased amounts of estrogen and testosterone [67,78,99], like the ovaries of transgenic mice that overexpress LH [104], indicating that the changes in the ovaries of the ER-alpha-deficient mice are due to LH hypersecretion. Similarly, mice with mutations that disrupt ER-alpha DNA-binding activity, signaling via AF-2, or membrane signaling also exhibit elevated LH and hemorrhagic cysts [39,80-86]. In support of this hypothesis, treatment of ER-alpha-deficient mice with a gonadotropin-releasing hormone antagonist (GnRH) decreases LH secretion and also prevents the ovaries from developing hemorrhagic cysts [105], similar to the effect from treatment of the ER mutant female patient.
The ovaries of mice with ER-beta deficiency are morphologically normal. Administration of gonadotropins to these animals also leads to formation of fewer oocytes than are produced by similarly treated WT mice, indicating a role for ER-beta in ovulation as well. When ER-beta-deficient follicles were cultured in vitro, they grew normally until they reached the antral stage, at which time their growth was significantly diminished. In addition, their ovulation rate was significantly lower than that of WT follicles, indicating that in the case of ER-beta deficiency, the defect is intrinsic to the follicle [106].
Uterus — The initiation and maintenance of pregnancy are dependent upon ovarian hormones. The pre-ovulatory peak of estrogen secretion is important for proliferation of the uterine epithelium in preparation for implantation, while rising progesterone secretion after ovulation is required for implantation of the embryo and the formation of decidual tissue in the stroma.
Hypoplastic, immature uteri are seen in ER-alpha-deficient rats and mice, and in mice with aromatase deficiency [71,101] or mutations that disrupt either ER-alpha DNA-binding activity or signaling via AF-2 [39,78,81,85,99]. In contrast, mice with epithelial cell-selective deletion of ER-alpha, disruption of ER-alpha AF-1 signaling, or membrane signaling develop normal-sized uteri [80,107]. The ER-alpha-deficient uterus does not respond to estrogen treatment in terms of weight increase, epithelial proliferation, or induction of estrogen-responsive genes, and therefore cannot support implantation of donor embryos [67,78,99].
The aromatase-deficient mouse uterus is estrogen responsive, however, maximal epithelial growth response requires administration of exogenous estrogen in the peri- or postpubertal window. This indicates that during pubertal development estrogen is needed to optimize the later cellular responsiveness of the tissue [108]. Mice with mutations that disrupt ER-alpha DNA-binding activity lack uterine responses to estrogen [39,109]. Mice with selective deletion of ER-alpha in uterine epithelial cells initially exhibit a uterine growth response to estrogen, but are unable to maintain the growth, and show a blunted increase in uterine weight [80]. Estrogen induces a full uterine growth response in mice with disrupted membrane signaling, indicating that non-genomic signaling does not impact uterine response [86].
Mice with disruption in signaling via AF-2 lack uterine responses to estrogen, but exhibit growth and regulation of transcripts when the selective estrogen receptor modulator (SERM) tamoxifen or the pure antagonist ICI 182 780 (fulvestrant) are used because of their ability to activate the AF-1 function of this mutant receptor [81]. Mice with deletion of ER-alpha AF-1 function have a blunted response to estrogen [110]. All of these responses to estrogen are normal in ER-beta-deficient mice, and they can carry pregnancies to term [98]. Thus, ovarian dysfunction is not the only component of infertility in ER-alpha-deficient mice and rats.
In contrast, inefficient ovarian function in ER-beta-deficient mice alone can explain the subfertility in these animals. Overall, it appears that ER-alpha is the principal receptor subtype required for uterine reproductive function.
Mammary tissue — At birth, the rodent mammary gland consists of a rudimentary ductal tree, which, as female rodents mature, elongates in response to estrogen and branches in response to progesterone to fill the stroma. In ER-alpha-deficient mice and rats, the ductal rudiment fails to elongate [78,99,111]. Similarly, no elongation is observed in mice with disrupted ER-alpha DNA-binding and AF-2 function [39,81]. In ER-beta-deficient mice, the gland develops normally, and these mice can nurse their young, indicating normal lactation function as well. These observations indicate that ER-alpha is required for normal mammary gland maturation and development.
Available data suggest that there is differential tissue growth regulation by different functional domains of the estrogen receptor, AF-1 in the uterus and AF-2 in the mammary gland.
Mating behavior — Mammalian fertility requires neuroendocrine control of behavior. In this regard, the lack of ER-alpha also leads to severe disruption in female mating behavior, whereas the lack of ER-beta does not [112,113], but does appear to result in an age-dependent increased anxiety [114].
Male reproduction — Male mice and rats with ER-alpha deficiency are infertile, demonstrating that estrogen has an essential physiological role in male reproduction. The testes of these mice and rats have dilated seminiferous tubules, due to lack of fluid resorption in the efferent ducts [66,79,99,115], and contain no sperm [116]. Male mice with disrupted ER-alpha DNA-binding or AF-2 activities exhibit the same deficiencies [39,82,117]. Treating AF-2-deficient ER-alpha male mice with tamoxifen can prevent complete infertility [82]. When germ cells from ER-alpha-deficient mice are transplanted into the testes of normal mice depleted of germ cells, the recipients can produce normal offspring [118,119]. This observation indicates that ER-alpha is not needed for sperm function but is required in the somatic cells of the male reproductive tract to allow maturation of sperm. Similarly, aromatase-deficient mice are infertile due to loss of spermatids and epididymal sperm, but have normal fluid resorption in their efferent ducts [75,120].
Although the lack of sperm alone results in male infertility, the mating behavior of mice with ER-alpha, ER-alpha and -beta, and aromatase deficiency is abnormal, while that of ER-beta-deficient mice is normal [113], indicating the importance of ER-alpha in normal male mating and aggressive behaviors. One exception to these observations is male-specific mounting behavior, which is not compromised in either individual knockout but is lost in the combined double knockout [121].
Combined deficiency of estrogen receptor alpha and beta — The phenotypes of mice with combined deficiency of ER-alpha and -beta are similar to those of ER-alpha-deficient mice [77,122], emphasizing the importance of ER-alpha in both male and female reproduction. One exception is the appearance of the ovaries [77,122]. In mice with the combined deficiency, the ovaries contain some normal follicles, each containing an oocyte, and the granulosa cells and surrounding theca are normal. However, many follicles are abnormal, and when viewed in a two-dimensional section appear similar to seminiferous tubules. These follicles most often lack an oocyte and granulosa cells, but rather have Sertoli-like cells, located along the basement membrane. These latter changes emerge progressively with age, suggesting that these tubular structures represent "ghosts" of follicles in which the oocyte died and the remaining granulosa cells were "trans-differentiated" into Sertoli-like cells. These unusual follicles are an indication that both ER-alpha and -beta are necessary to maintain normal follicular structures and granulosa cell differentiation.
Nonreproductive tissues — The increased risk of osteoporosis and cardiovascular disease in postmenopausal females indicates a role for estrogen in bone and cardiovascular tissue (figure 1) (see "Pathogenesis of osteoporosis", section on 'Estrogen' and "Menopausal hormone therapy and cardiovascular risk"). These tissues have been studied in ER- and aromatase-deficient mice.
●Bone – The femurs of mice with ER-alpha and combined -alpha/-beta deficiency, but not ER-beta deficiency, are short [123] compared with those of normal mice, and mice with these deficiencies have lower serum IGF-I concentrations. In addition, femoral diameter is smaller in affected female mice and femoral density is lower in affected male mice [67].
In ER-beta-deficient female mice, bone mineral content and density are increased, but they are normal in male mice [124], indicating that ER-beta may exert a negative effect on bone density in female mice. The major clinical problems in patients lacking aromatase or ER-alpha are osteoporosis or osteopenia and eunuchoid stature, due to lack of epiphyseal fusion, indicating the importance of estrogen in bone function in humans [72].
●Vascular effects – Blood vessels of ER-deficient mice have been evaluated by studying the protective effect of estrogen against carotid injury [125-127]. The protective effect was lost in ER-alpha-deficient mice, but not in ER-beta-deficient mice [127], indicating that ER-alpha mediates the estrogen protection on atherosclerosis [128]. Mice deficient in ER membrane signaling also lacked estrogen effects on re-endothelialization, indicating that the membrane signaling mechanism of the ER mediates this response [86]. The male patient described above who lacked ER-alpha had endothelial dysfunction and evidence of coronary artery disease at age 31 years, despite normal serum cholesterol concentrations [88]. (See 'Physiologic roles of estrogens' above.)
●Brain – The effects of estrogen on the brain and cognitive function are reviewed separately. (See "Estrogen and cognitive function".)
●Obesity and insulin resistance – Obesity and insulin resistance greatly impact human health. There is evidence from human as well as animal models for a protective role of estrogen. Diabetes is less prevalent in females than in males until after menopause, and hormone therapy appears to lower the risk of type 2 diabetes mellitus [129,130] (see "Menopausal hormone therapy: Benefits and risks", section on 'Type 2 diabetes mellitus'). Additionally, studies in rodent models of insulin resistance or experimentally induced diabetes show sexually dimorphic patterns, with females often exhibiting reduced susceptibility, which is lost after ovariectomy. Estrogen treatment protects males and ovariectomized females [130]. In addition, mouse knockout models of ER-alpha or Cyp19, the enzyme that synthesizes E2, exhibit obesity and insulin resistance.
Mice lacking ER-alpha have increased adiposity and insulin resistance. Sites of action for estrogen in controlling such conditions involve brain regions, such as SF-1 and pro-opiomelanocortin (POMC) neurons [131,132], where estrogen regulates food intake, thermogenesis, and energy expenditure to provide the proper weight balance. Additionally, estrogen acts in peripheral tissues to suppress adipokine and cytokine levels that control inflammation responses that influence glucose and insulin actions in skeletal muscle, reflecting the known condition of insulin resistance seen in the metabolic syndrome [133]. Mice lacking ER-alpha are partially protected from experimentally-induced diabetes [134], suggesting that not all the protective effects are mediated by ER-alpha. This may be explained from experimental studies of the metabolic syndrome showing the ER-alpha AF-1 function can prevent obesity, but not insulin resistance suggesting that insulin sensitivity requires AF-2 function. Males with mutations in Cyp19 or ER-alpha also exhibit insulin resistance [87,130].
SUMMARY
●Estrogen receptors – Estrogens act on the reproductive organs, cardiovascular organs, and bone. Their major actions are genomic, mediated by nuclear estrogen receptors (ERs). Two ER protein molecules have been identified: the original ER-alpha and the later described ER-beta. (See 'Estrogen receptors' above.)
●Non-genomic actions – As these rapid effects occur without ER-gene interaction, they are called "non-genomic," although the signals initiated by these mechanisms ultimately result in regulation of genes. (See 'Non-genomic actions of estrogens' above.)
●Knockout models – The ability to produce gene-targeted mice in which ER genes or the aromatase gene (the enzyme that converts androgens to estrogens) have either been knocked out, deleted selectively in specific cell types, or mutated to disrupt particular ER functions, has allowed expanded study of the physiological processes that require estrogen in vivo. Identification of patients with mutations in these genes has shed further light on the clinical physiological actions of estrogens. (See 'Physiologic roles of estrogens' above.)
●Reproduction – As might be expected, the most important defects in these patients (and in knockout models) are in the reproductive system. Both male and female mice lacking the ER-alpha gene or ER-alpha and -beta genes are infertile, whereas in animals lacking only the ER-beta gene the males are fertile and the females are subfertile. In mice lacking aromatase, the males are infertile and the females are anovulatory. (See 'Female reproduction' above.)
Do you want to add Medilib to your home screen?