Perspective: Nonreproductive Sites of Action of Reproductive Hormones

Estrogens and androgens, the female and male sex hormones responsible for gender dimorphism and reproduction, play also a very important role in tissues and organs that are not directly involved in procreation. In fact, withdrawal of the effects of estrogens at menopause from nonreproductive tissues like the skeleton, the cardiovascular system, and the brain, is a major risk factor for the development of osteoporosis, coronary artery disease, stroke, and perhaps neurodegenerative diseases like Alzheimers. On the other hand, continuous exposure of reproductive tissues to estrogen during the postreproductive part of life, is a risk factor for the development of breast, ovarian, and uterine cancer. During the last decade, there have been significant advances in our understanding of the mechanism of action of sex steroids on the skeleton, the cardiovascular system, and the central nervous system. The purpose of this perspective is to provide a brief outline of these advances with an emphasis on the emergence of several thematic similarities in the action of sex steroids on nonreproductive tissues. For more thorough discussion of this topic, the reader is referred to recent review articles (1–5). Estrogens and androgens influence the development of the skeleton and the linear growth of bones during puberty. In addition, they maintain homeostasis of the adult skeleton during its continuous regeneration by the process of bone remodeling. Linear skeletal growth is governed by the chondrocytes of the growth plate. Maintenance of adult bone mass is controlled by the basic multicellular units—temporary anatomic structures comprising osteoclasts in the front and osteoblasts in the rear. Bone remodeling consists of focal cycles of resorption followed by formation. The rate of remodeling or turnover depends on the number of cycles, but the effects on bone mass depend on the focal balance within each cycle. Estrogens and androgens decrease the number of remodeling cycles by attenuating the birth rate of osteoclasts and osteoblasts, in part by regulating the production of cytokines (6–8). The focal balance between resorption and formation depends on the lifespan of osteoclasts and osteoblasts, reflecting the timing of apoptosis. Apoptosis of osteoclasts and osteoblasts is modulated by sex steroids in opposite directions. Specifically, sex steroids promote osteoclast apoptosis but prevent the apoptosis of osteoblasts and osteocytes (9, 10). Consequently, loss of estrogens in females or androgens in males after maturity leads to an increased rate of bone remodeling and tilts the balance of resorption and formation in favor of the former; thus, loss of bone mass, deeper than normal osteoclast erosion cavities (leading to the removal of some cancellous elements entirely, with the remainder being more widely separated and less well connected), and increased propensity to fractures. Although two recent clinical trials have failed to show a protective effect of HRT in postmenopausal women with preexisting coronary heart disease (11, 12), an extensive body of evidence indicates that estrogens do have atheroprotective properties that result from a variety of actions. Such actions include favorable changes in lipoprotein metabolism and oxidation, modulation of endothelial permeability to low density lipoprotein, as well as regulation of the expression of cytokines and cell adhesion molecules, vascular smooth muscle cell and neointimal proliferation and migration, calcification and platelet adhesion, and aggregation. Evidence from animal and human studies has strongly suggested that estrogens, at physiological levels, also cause vasodilatation (3). This effect is shared by androgens. This finding is consistent with reports that normotensive men have higher plasma testosterone than hypertensive and that angina pectoris can be successfully treated with acute injections of testosterone (13, 14). In the case of either class of sex steroids, the effect on vasodilatation is mediated by activation of endothelial nitric oxide synthase (eNOS), as well as through increased expression of the eNOS and prostacyclin synthase genes, which regulate vascular tone (15, 16). Dihydrotestosterone, on the other hand, has been shown to increase human monocyte adhesion to endothelial cells, a proatherogenic effect that is mediated at least in part by increased endothelial cell-surface expression of vascular cell adhesion molecule 1 (17); this is perhaps one of several reasons for the higher incidence of coronary artery disease in men compared with premenopausal women. In the CNS, besides regulating gonadotropin and PRL secretion and sexual behavior, estrogens influence verbal fluency and memory, fine motor skills, coordination of movement as well as mood (4, 5, 18). In addition, they seem to decrease the risk of neurodegenerative diseases and attenuate injury by suppressing the effects of neurotoxic or ischemic stimuli or increasing the resilience of the brain to injury (19, 20). Several mechanisms may account for these effects. Estrogens selectively enhance the growth and differentiation of axons and dendrites (neurites) in the developing brain, an effect that may be recapitulated following injury later in life. Estrogens also exert antiapopotic effects on neuronal cells that are mediated by activation of mitogen-actived protein (MAP) kinases or protein kinase A and protein kinase C, down-regulation of the expression of neurotrophin receptors, or by altering free radical production or free radical action on cells. Estrogens may also regulate the cholinergic system by inducing acetylcholinesterase and its activity or the receptor for nerve growth factor (NGF), reduce the tone of dopaminergic neurons, and have antinociceptive and analgesic actions. Androgens protect neuronal cells from oxidative stress, and also alter the expression of the opioid receptor (21, 22). In addition, aromatizable as well as nonaromatizable androgens alter the morphology, survival, and axonal regeneration of lower motor neurons (23); and, androgen receptor dysfunction causes degeneration of spinal and bulbar motor neurons (Kennedy’s disease) (24). A striking difference between nonreproductive and reproductive sites of sex steroid action is the lower level of receptor expression in the former (10- to 50-fold). This phenomenon has been extensively documented for a long time, but never explained. Consistent with the effects of sex steroids on skeletal growth and remodeling, receptors for estrogens (ERα and ERβ) or androgens (AR) are present in chondrocytes, osteoblasts, and bone marrow stromal cells as well as in osteoclasts and their progenitors. Remarkably, however, the level of expression of ERα, ERβ, or AR in bone cells is at least 10-fold lower compared with reproductive tissues. In addition, the distribution of the receptors in bone cells does not vary by gender, as similar levels of ER and AR have been found in bone cells from males and females. ERα and ERβ, as well as AR, have been identified in vascular endothelial and smooth muscle cells (including cells from the coronary arteries, aorta and peripheral veins) as well as myocardial cells and macrophages. Nonetheless, as is the case with bone cells, the concentration of the estrogen receptors in cells of the cardiovascular system is very small—as little as 50-fold less compared with cells of reproductive tissues (25). Moreover, there does not seem to be different distribution of the ER in cells from males or females. The atheroprotective effects of estrogen are mediated to a large extent through the ERα, but ERα-independent mechanisms may be also involved (26). Receptors for ERα, ERβ, or AR have been demonstrated in the pituitary and hypothalamus as well as in nonhypothalamic brain regions such as the hippocampus, basal forebrain, brain stem, the causate-putamen and substantia nigra, the olfactory lobe, amygdala, cortex, midbrain, central brain, cerebellum and cholinergic, serotonergic, and catecholaminergic neurons. The level of estrogen receptor expression varies among different regions of the brain, but by and large, with the exception of the pituitary that exhibits moderate to high levels, ERα or ERβ expression in neuronal cells such as glial, is very low or undetectable. The levels of the expression of these receptors in brain may also vary at different stages of life, as indicated by the transient expression of ERα in the cerebral cortex during neonatal development—a period of dramatic neurogenesis and differentiation—but the receptors virtually disappear thereafter (20). Unlike reproductive tissues where development, growth, maintenance, and function depend on estrogens in females and androgens in males, several lines of evidence indicate that in nonreproductive tissues this specificity is greatly relaxed and that nonreproductive sites are responsive to either class of sex steroids irrespective of gender. Until recently it was thought that low doses of estrogens are responsible for the pubertal growth spurt in both boys and girls. Nonetheless, from the case of a man with estrogen resistance due to a mutation of the ERα gene (27) and two men with aromatase deficiency (28, 29), it has become clear that androgens, if nonaromatized to estrogen, can also stimulate growth directly at the level of the epiphyseal chondrocyte. Indeed, all three men achieved their genetic potential for height at a normal age of 16–17 yr, rather that at a later age, as could be expected with hypogonadism. On the other hand, high doses of estrogens inhibit growth by inhibiting cell proliferation at the hypertrophic zone or by inducing terminal differentiation of proliferating chondrocytes, apoptosis of hypertrophic chondrocytes, and vascular and osteoblast invasion into the growth plate. Because estrogen-resistant men in their mid to late twenties have tall stature and continued linear growth as a result of delayed skeletal maturation, androgens may be ineffective in inducing epiphyseal maturation and closure. Consequently, estrogens must be responsible for the closing of the epiphyses in both sexes (30). The decreased bone mass in the male with the mutant estrogen receptor and increased bone mass after treatment with estrogens in the two males with aromatase deficiency have also suggested that estrogens derived by peripheral aromatization of androgens are critical for the maintenance of bone mass in postpubertal life in men (31). However, in all three cases, the decreased bone mass and the increased levels of bone markers in young males with estrogen deficiency, in the face of androgen sufficiency, could well be the result of failure to achieve peak bone mass because of defective skeletal growth during development as well as the fact that these individuals never underwent puberty—not to loss of bone mass, as is the case with the common forms of osteoporosis. In support of this contention, humans with complete androgen insensitivity, due to mutations in the androgen receptor gene on the X-chromosome, have decreased bone mass, in spite of the elevated estrogen levels (32); and women with androgen insensitivity syndrome had decreased bone mineral density even when fully complying with estrogen therapy (33). Moreover, androgens, including nonaromatizable ones, like dihydrotestosterone, have identical effects to those of estrogens on the biosynthetic activity and the birth as well as the death of bone cells in vitro and in vivo, at least in rodents (7, 34). Even more importantly, androgens can prevent bone loss in ovariectomized rodents (35–38). However, in a recent study of elderly men estrogen loss seemed to be more critical for a short-term increase in bone resorption markers than loss of androgens; whereas both estrogen and androgen loss were needed to elicit a short-term decrease in bone formation markers (39); these quick changes probably reflect the pro- and antiapoptotic effects of sex steroids on osteoclasts and osteoblasts. Heretofore, there has been no explanation for the puzzling efficacy of either class of sex steroids in females and males. Nonetheless, recent studies from our group have elucidated a potent antiapoptotic effect of estrogens or androgens on osteoblasts and osteocytes, which can be transmitted by the ERα, ERβ, or the AR with similar efficiency irrespective of whether the ligand is an estrogen or an androgen (10). We suspect that these findings are relevant to the efficacy of either class of sex steroids in females and males. We also speculate that our findings may account for the weak and puzzling skeletal phenotype of the deletion of the ERα or ERβ genes in mice (40). Indeed, these mice exhibit a decrease in the length and diameter of the femur in both females and males as well as low bone turnover, which are diametrically opposite to the increased length of long bones and the high turnover of patients with hypogonadism, or the humans with the ERα or the aromatase mutations. An equivocal bone phenotype has been also observed in aromatase-deficient (ArKO) mice (41). While possible that the minor phenotypic changes in these mutant mice may be due to loss of estrogenic effects on skeletal development or compensatory changes in the production of estrogens or androgens, an alternative explanation is that the difference between the receptor deficient vs. the hormone deficient mice might be due, in part, to the ability of the AR to transmit bone protective effects of estrogens. The effect of estrogens in vascular smooth muscle cells from males and females also appears to be gender independent. Indeed, estrogen administration improves vascular reactivity in men, produces a significant increase in the response to the endothelial vasodilator acetylcholine (Ach), prevents carotid injury in castrated male rats, and reduces plaque formation in transgenic male mice expressing low levels of apoE (16, 42–44). Furthermore, the man with the homozygous mutation of the ERα gene also has abnormal vascular reactivity (45). Like estrogens, administration of testosterone produces dose-dependent relaxation of the thoracic aorta of rats and rabbits and attenuates the contractile response of the aorta to phenylephrine (13, 16). Remarkably, the effects of testosterone occur in both males and females in a gender-independent fashion. Furthermore, the effects of testosterone persist in the mouse model of testicular feminization (tfm); and, are demonstrable in the presence of an androgen receptor antagonist or an aromatase inhibitor (13). This evidence raises the possibility that the vasodilating effects of androgen may be mediated by a sex-nonspecific mechanism, perhaps through the ER. In some regions of the brain, the male and female response to estrogens differs, but in others, estrogens can regulate similarly gene expression in males and females. Moreover, administration of estrogens and progestins improves the outcome after cerebral ischemia and traumatic brain injury in experimental models. This neuroprotective effect of estrogen administration extends to males as well (46). With the only possible exceptions of the regulation of TGFβ in bone and p75NTR and trkA (47, 48) in neuronal cells via putative EREs (Fig. 1A), the actions of estrogens on the skeleton and the CNS seem to result primarily either from protein-protein interaction between the estrogen receptor and other transcription factors such as NF-κB, AP-1, or C/EBP (49) (Fig. 1B), or via rapid nongenotropic mechanisms (Fig. 1C). The same seems true for the cardiovascular system. Some direct effects of estrogen on the blood vessels may be exerted through ERE-independent genomic actions, but the vast majority of the effects of estrogen on the vasculature, including relaxation of coronary arteries, vasodilatation and protection of blood vessel injury to hypoxic insults, occur via rapid nongenotropic actions. Paradigms of genotropic and nongenotropic actions of the estrogen receptor. Cartoon A depicts the protein-DNA interaction of the estrogen receptor in the case of the transcriptional regulation of an ERE containing gene, like Lactoferrin. Cartoon B depicts the protein-protein interaction responsible for the transcriptional regulation of a gene that does not contain an ERE, like IL-6. By contrast to these genotropic actions. Cartoon C depicts a nongenotropic action involving the interaction of a membrane-associated form of the ER (that could be the classical ER or a shorter splice variant) with the Src/Shc/ERK signaling pathway. Although in this third model the ligand-activated receptor protein is depicted as a monomer, it may well be that in fact the membrane-associated receptor, as it is the case with the one localized in the nucleus, interacts with components of the signal transduction pathway as a dimer. Studies of estrogen and androgen action on neuronal, endothelial, and bone cells have all independently elucidated a remarkably similar mechanism that is mediated via the activation of MAP kinase signaling pathways. Indeed, estrogens elicit a rapid and sustained activation of ERK1 and ERK2 in explants of the cerebral cortex, which result from increased B-raf and MEK kinase activity. Similarly, estrogens activate eNOS in human endothelial cells through ERα, by activation of the PI3-kinase-Akt pathway. Remarkably, it was demonstrated that this effect results from direct interaction of the receptor protein with the p85 regulatory subunit of the PI3-kinase (50). The relaxation of the rat thoracic aorta by testosterone is also mediated via NO release (16). The antiapoptotic effect of estrogens and androgens on osteoblast and osteocytes is mediated via a Src/Shc/ERK signaling cascade and a region of the classical receptor that is distinct from the one responsible for the genotropic actions of the ligand-activated protein (10). Indeed, using ERα as a paradigm, we determined that this effect requires only the ligand binding domain and is eliminated by nuclear targeting of the receptor protein. Interestingly, expression of constitutively active Src or stimulation of the endogenous Src/JNK pathway enhances transcriptional activation of the estrogen-ER complex and strongly stimulates the otherwise weak activation by the unliganded ER, through a Raf/MEK/ERK and a MEKK/JNKK/JNK signaling cascades (51, 52). Thus, Src may be a pivotal protein for the function of the ER both in its genotropic and nongenotropic modes of action. Estrogens phosphorylate Src within seconds in mammary tumor cells, and ER or AR coimmunoprecipitate with Src in prostate cancer cells (53, 54). This evidence taken together with the finding that the antiapoptotic signal of the ligand binding domain of the ER is preserved when targeting this protein to the membrane, but it is lost when targeting it to the nucleus (10), strongly that these actions are mediated via the or a of the classical receptor that is with the plasma (Fig. 1C). In support of this have shown that both ERα and ERβ can be in the cell and that a form of estrogen endothelial cells from apoptosis and their function via rapid activation of the MAP Several of the MAP kinase signaling including and are in that are in the protein and signal transduction are found in a variety of cell including neuronal, endothelial, and cells ERα has been shown recently to coimmunoprecipitate with Furthermore, a of ERα has been in with both and eNOS in human endothelial cells it is very that nongenotropic activation of signaling by sex steroids is mediated via the classical receptors, or perhaps a that is localized in the and in within hormones exert effects on nonreproductive tissues like the skeleton, the and CNS Although their mechanism of action in these tissues has to be several striking to from their actions on reproductive tissues. nonreproductive tissues low levels of receptors and exhibit a of of the levels of the estrogen or androgen receptor in cells from males vs. females. in nonreproductive there appears to be relaxed specificity in the of females and males to estrogens or androgens. there are nongenotropic effects of estrogens and androgens in nonreproductive tissues via putative membrane-associated within receptors that may transmit the of estrogens or androgens in a sex-nonspecific effects may account at least in part for the relaxed gender specificity of the of nonreproductive tissues to estrogens or androgens. The nongenotropic effects of sex steroids on cells from nonreproductive tissues like bone can be demonstrated in vitro with in spite of the low levels of receptor expression in such cells. This is in contrast to genotropic effects that if at are to elicit even with of the steroids in nonreproductive tissues. This that the receptor expression in nonreproductive tissues may be for nongenotropic actions of sex steroids but for genotropic actions. it is possible that the low receptor expression in cells from nonreproductive tissues is of these tissues from the sex hormones during and the their ability to to the effects of sex In we that understanding of the and similarities of the mechanism of action of sex steroids in nonreproductive vs. reproductive together with evidence that one can nongenotropic from genotropic of the sex steroid receptors with (10), may provide of the effects of sex steroids on nonreproductive tissues during postreproductive life, or effects that are no needed are on the reproductive tissues.