Osteoporosis International

, Volume 19, Issue 11, pp 1517–1525

Sex hormones, their receptors and bone health

Authors

  • K. Venken
    • Bone Research Unit, Laboratory for Experimental Medicine and Endocrinology, Department of Experimental MedicineKatholieke Universiteit Leuven
  • F. Callewaert
    • Bone Research Unit, Laboratory for Experimental Medicine and Endocrinology, Department of Experimental MedicineKatholieke Universiteit Leuven
  • S. Boonen
    • Bone Research Unit, Laboratory for Experimental Medicine and Endocrinology, Department of Experimental MedicineKatholieke Universiteit Leuven
    • Bone Research Unit, Laboratory for Experimental Medicine and Endocrinology, Department of Experimental MedicineKatholieke Universiteit Leuven
Review

DOI: 10.1007/s00198-008-0609-z

Cite this article as:
Venken, K., Callewaert, F., Boonen, S. et al. Osteoporos Int (2008) 19: 1517. doi:10.1007/s00198-008-0609-z

Abstract

Sex steroids regulate skeletal maturation and preservation in both men and women, as already recognized in the 1940s by Albright and Reifenstein. The impact of gonadal insufficiency on skeletal integrity has been widely recognized in adult men and women ever since. In the context of their skeletal actions, androgens and estrogens are no longer considered as just male and female hormones, respectively. Androgens can be converted into estrogens within the gonads and peripheral tissues and both are present in men and women, albeit in different concentrations. In the late 1980s, sex steroid receptors were discovered in bone cells. However, the understanding of sex steroid receptor activation and translation into biological skeletal actions is still incomplete. Due to the complex metabolism, sex steroids may have not only endocrine but also paracrine and/or autocrine actions. Also, circulating sex steroid concentrations do not necessarily reflect their biological activity due to strong binding to sex hormone binding globulin (SHBG). Finally, sex steroid signaling may include genomic and non-genomic effects in bone and non-bone cells. This review will focus on our current understanding of gonadal steroid metabolism, receptor activation, and their most relevant cellular and biological actions on bone.

Keywords

Bone growthBone maintenanceOsteoporosisSex steroidsSex steroid receptorsSex steroid receptor signaling

Introduction

The lifetime risk for fragility fractures is considerably less in men (15%) than in women (40%); thus men are better protected against osteoporosis and osteoporotic fractures [1]. It has been generally accepted that sex steroids—both androgens and estrogens—play a key role in the establishment of this skeletal gender difference. According to the traditional concept of sexual dimorphic bone growth, androgens being considered as ‘male hormones’ stimulate, whereas estrogens being considered as ‘female hormones’ inhibit bone mineral acquisition during puberty. In addition, men lose less bone than women during aging because they do not experience an equivalent of menopause. More recently, it has become clear that androgens and estrogens cannot be considered as only ‘male’ or ‘female’ hormones, respectively. Androgens may act as prohormones of estrogens even within bone. Moreover, both androgens and estrogens and their respective receptors have specific actions on the different bone envelopes during bone growth and bone maintenance in both sexes. The detailed phenotypic description of several relevant knockout mouse models increased our insight in the role of sex steroid receptor activation for the growth and maintenance of different bone envelopes. Therefore, in the first two parts, the most important steps of sex steroid secretion, metabolism, transport and signaling will be reviewed. In a third part the relevance of androgen and estrogen signaling for biological skeletal action will be discussed.

Secretion, metabolism and transport of sex steroids

Androgens are C-19 steroids secreted from the testes in men and from the adrenals in both men and women. With respect to the biological activity of androgens, testosterone (T) is the most important androgen in men, of which approximately 95% is secreted by the testis. In addition, the adrenal cortex secretes large amounts of weakly active androgens such as dehydroepiandrosterone (DHEA), DHEA sulfate (DHEA-S), and androstenedione. T can act through its receptor, the androgen receptor (AR), either directly or indirectly after conversion to the more potent 5α-dihydrotestosterone (DHT) by the enzyme 5α-reductase in peripheral tissues [2]. In addition, T has the unique feature that it can be converted into 17β-estradiol (E2) by the P450 aromatase enzyme, even within bone, and subsequently can exert its effects through estrogen receptor α (ERα) or β (ERβ) [3]. Also the adrenal androgens are an important source of substrate for the extragonadal synthesis of potent sex steroids; i.e., these C-19 androgens can be metabolized to estrone (E1) by the aromatase enzyme or to T by steroid sulfatase, 17β-hydroxysteroid dehydrogenase (17β-HSD) and/or 3β-HSD (Fig. 1). Thus, depending on the relative expression of P450 aromatase, 5α-reductase, 17β-HSD, 3β-HSD and steroid sulfatase, T and C-19 androgens may either predominantly activate the AR or the ERs. As these enzymes are all expressed in bone tissue, local metabolism of androgens in bone may be of significant physiological importance. Thus, sex steroids are in a large part synthesized locally in peripheral tissues, providing individual target tissues with the means to adjust synthesis and metabolism of sex steroids to their local requirements. Only about 20% of E2 is directly secreted by the testis in men. The other 80% of male E2 is derived from androgen conversion in peripheral tissues, mainly adipose tissue [4].
https://static-content.springer.com/image/art%3A10.1007%2Fs00198-008-0609-z/MediaObjects/198_2008_609_Fig1_HTML.gif
Fig. 1

Overview of the metabolism and action of sex steroids in men. T, testosterone; DHT, dihydrotestosterone; E2, 17β-estradiol; 17β-HSD, 17β-hydroxysteroid dehydrogenase; 3β-HSD, 3β-hydroxysteroid dehydrogenase

In women, the steroidogenic pathways are essentially the same as in men. The ovaries express large amounts of the P450 aromatase enzyme, which converts androgens into estrogens within the ovaries. Therefore, also 95% of 17β-estradiol is directly secreted from the ovaries in premenopausal women. When ovarian estrogen production ceases after menopause, circulating estrogen concentrations are solely derived from the conversion of adrenal androgens into estrogens, mainly within fat. Estrogen concentrations in postmenopausal women are generally even lower than in elderly men. In contrast to women, total T concentrations decrease only marginally (about maximum 1% per year) during aging in men, whereas total E2 concentrations remain constant. In men, the majority of T is bound to sex hormone binding globuline (SHBG) (50–60%, with high affinity) and albumin (40–50%, with low affinity) whereas only 1–2% of T remains free [5]. The albumin-bound T and the free T, often referred to as the bioavailabe T (bioT), represent the fractions available for biological action. Unbound T diffuses passively through the cell membranes into the target cell, where it binds to the specific AR [6]. SHBG already increases markedly in middle aged men, resulting in important reductions of bioavailable T and E2 in elderly men even in the presence of only modest reduction of T secretion [7]. In contrast to men, only 20–40% of E2 in women is bound to SHBG, making its biological significance less important than in men.

Sex steroid receptors and sex steroid receptor signaling

Sex steroid receptors

The finding that AR, ERα and ERβ are expressed in human, rat and mouse osteoblasts, osteoclasts, osteocytes and growth plate chondrocytes supports the notion that sex steroids can act on the skeleton through direct stimulation of their receptor [8]. These receptors are composed of three independent but interacting functional domains; the NH2-terminal or A/B domain, the C or DNA-binding domain (DBD), and the D/E/F or ligand-binding domain (LBD). The N-terminal domain encodes a ligand-independent activation function (AF-1), a region of the receptor involved in protein-protein interactions and transcriptional activation of target gene expression. The COOH-terminal or ligand-binding domain (LBD) mediates ligand binding, receptor dimerization, nuclear translocation, and transcription of target gene expression.

Genomic sex steroid signaling

In the absence of hormones, sex steroid receptors are kept in an inactive state and are sequestered in a multiprotein complex by means of chaperone molecules such as heat shock proteins (Hsp) Hsp70 and Hsp90 within the cytoplasm. Hormone binding to the receptor induces a conformational change, resulting in dissociation of the chaperone proteins and translocation of the receptor to the nucleus and increased phosphorylation (Fig. 2). Then, the hormone-bound receptor dimerizes with another hormone-bound receptor. This dimer binds with high affinity to specific DNA sequences, i.e., estrogen or androgen responsive elements (ERE, ARE), respectively, located within the regulatory region of the target gene [9]. The DNA-bound receptor contacts the general transcription apparatus either directly or indirectly through co-regulatory proteins [10]. These co-regulatory proteins may include either factors enhancing transactivation (coactivators such sex steroid co-activator 1, 2 and 3) or factors reducing transactivation (corepressors). The process of gene modulation through this classical mechanism takes at least 30–45 minutes, and the time required to produce significant levels of protein is in the order of hours [11].
https://static-content.springer.com/image/art%3A10.1007%2Fs00198-008-0609-z/MediaObjects/198_2008_609_Fig2_HTML.gif
Fig. 2

Genomic sex steroid signaling pathways. The ligand, androgen or estrogen, binds to its respective receptor, the androgen receptor (AR) or estrogen receptor (ER), respectively, which induces a conformational change. The receptor translocates to the nucleus and undergoes dimerization resulting in the formation of a transcriptionally competent complex that binds the response elements of target genes. ARE, androgen responsive element; ERE, estrogen responsive element; AP-1, activating protein-1; SP-1, specificity protein-1

Beside the ligand-dependent signaling mechanism of sex steroids, sex steroid receptor function can be modulated by signals other than sex steroids. This ligand-independent signaling mechanism depends on the ability of polypeptide growth factors, such as epidermal growth factor (EGF) and insulin-like growth factor-I (IGF-I), to activate sex steroid receptors and modulate the expression of target genes through their respective hormone response elements (ARE or ERE) [12]. The mechanisms by which sex steroid receptors and growth factor pathways converge have not been elucidated yet.

The above-mentioned mechanisms provide an explanation for the regulation of genes with a functional responsive element sequence within the promoter region. Reports of hormone-bound receptor activation of genes in which no responsive element sequence was present led to the discovery that sex steroid receptors can also modulate gene expression at alternative regulatory DNA sequences, such as activating protein-1 (AP-1) [1315] and specificity protein-1 (SP-1) [16] (Fig. 2). In this context, the sex steroid receptor is recruited to the specific promoter complex where it interacts with other DNA-bound transcription factors such as c-Jun, c-Fos, and other coactivator proteins.

Non-genomic sex steroid signaling

Genomic sex steroid signaling takes time. However, a variety of cell types are able to respond to sex steroids within seconds or minutes, which makes a classical genomic signaling unlikely. It suggests that sex steroids may elicit non-genomic effects, possibly through cell-surface receptors linked to intracellular signal transduction proteins [17, 18]. Binding sites for estrogens [19] and androgens [20] have been identified on plasma membranes. The membrane-initiated steroid signaling results in the activation of conventional second messenger signal transduction cascades, including activation of protein kinase A (PKA), protein kinase C (PKC), cellular tyrosine kinases, MAPKs, PI3K and Akt- and Src/Shc/ERK signaling [21, 22].

In the context of non-genomic sex steroid signaling, it was previously shown that androgens and estrogens can transmit antiapoptotic effects on osteoblasts in vitro with similar efficiency via either AR, ERα or ERβ [23]. These effects are mediated by Src/Shc/ERK signaling and seem sex non-specific [24]. Interestingly, the synthetic compound 4-estren-3α,17β-diol (estren) was described to increase bone mass in gonadectomized male and female mice through non-genomic signaling [25]. Moreover, these positive effects on bone occur without adverse effects on reproductive organs. However, later it was clearly shown that estren is able to increase seminal vesicles and uterine weight [26, 27] with even evidence for transcriptional (i.e., genomic) activity through the AR [2729]. In addition, in vivo data showed that no cross-reactivity exists between ARs and ERs and their ligands [30]. Although it has become clear that the mechanisms of sex steroids and their receptors are complex and that a wide diversity of cellular pathways may be involved, their significance with respect to bone metabolism has not been clarified yet. Therefore, classical genomic sex steroid signaling appears to be the most important mechanism of action of sex steroids on skeletal homeostasis.

Biological effects of sex steroids on bone growth

The production of sex steroids at the start of puberty is clearly linked with an increase of bone mineral acquisition during this period. Both adolescent boys and girls experience an increase in bone size resulting from an enlargement of the outer diameter of bone and a concomitant widening of the medulary diameter [31]. These changes in bone geometry occur in both sexes, but they occur to a substantially greater extent in males than in females. Bone growth involves growth in length and width by means of longitudinal bone formation and periosteal bone formation versus endosteal bone resorption, respectively. Overall, net endocortical bone resorption is less than periosteal bone formation resulting in radial bone expansion and cortical thickening. Investigation of the skeletal phenotype of relevant gonadoctomized and transgenic rodents has increased our understanding of the role of androgen and estrogen receptor signaling in bone formation and bone resorption at different bone envelopes.

For instance, it is well recognized that sex steroids stimulate longitudinal bone growth at the start of puberty. More recently, it was established that this stimulatory action on enchondral bone formation depends on ERα activation in both sexes [8]. The latter effect may, at least partly, be attributed to indirect activation of the GH-IGF-I axis by estrogens [32]. An AR-dependent effect on enchondral bone formation remains rather uncertain as bone length in AR-disrupted mice and rats is unchanged [3336]. At the end of puberty, ERα activation is also responsible for epiphyseal growth plate closure in both sexes as illustrated by the absence of epiphyseal fusion in men with aromatase deficiency and ERα disruption [37, 38].

The greater male bone size primarily results from enhanced periosteal bone formation, affected by both androgens and estrogens [39]. In contrast with longitudinal bone formation, periosteal bone formation is affected by both androgen and estrogen actions: periosteal bone formation is significantly reduced following androgen deficiency in growing male rats, while it is increased in estrogen-deficient female rats leading to the traditional view of stimulatory androgens in males versus inhibitory estrogens in females on periosteal growth [40]. Studies in male mice and rats with either a natural mutation or genetic manipulation of the gene encoding the AR further support the role of direct AR-mediated androgen action on periosteal bone formation [3436]. Whether androgens also stimulate periosteal bone in females is less clear but likely as a similar osteoanabolic action of androgens in females has been demonstrated by some studies in female rats treated with an anti-androgen [41] and by observations of increased bone mass (even when corrected for bone size) in hirsute women [42]. Cortical bone mass in female mice is also increased by ERβ disruption, indicating that ERβ may inhibit periosteal bone expansion in females [43].

Since the observations that men suffering from either estrogen deficiency (due to a mutation in the aromatase gene) or resistance (secondary to a mutation in the ERα gene) have delayed skeletal development and impaired bone mineral gain [37, 38, 44], the crucial role of estrogens in male skeletal growth has received much attention. In addition, a significant increase in bone size of the distal radius has been observed following estrogen treatment in an adolescent aromatase deficient boy with (supra)normal T levels [44]. Recently, it has been shown that a combination of testosterone plus estradiol increased BMD to a greater extent than testosterone alone in an aromatase deficient man [45]. Moreover, this increased BMD was associated with an increased elevated cortical thickness of the radius and tibia, as measured by peripheral quantitative computed tomography (pQCT) [45]. Both reports challenge the traditional view of stimulatory androgens versus inhibitory estrogens on periosteal bone formation and indicate that periosteal bone formation in males may not be solely dependent on androgen action, but also, at least in part, on estrogen action. In this context, targeted deletion of the gene for either aromatase or ERα, but not ERβ, in male mice results in decreased cortical bone area and thickness [46]. Also, young growing male rats treated with an aromatase inhibitor show a reduced cortical bone growth [47]. However, these findings may be confounded by decreased serum IGF-I levels, again indicating that estrogen action may be rather indirect through modulation of the GH-IGF-I axis [8, 48].

During growth, not only periosteal but also medullary bone expansion is limited in females compared with males, i.e., endocortical contraction occurs in both female rodents and in women (relative to male) at the end of puberty, at least at some skeletal sites [49]. This suggests that estrogens are involved in bone mineral accumulation at the endocortical bone surface [50]. Therefore, estrogen receptor signaling has stimulatory and inhibitory effects at the enchondral, periosteal and endocortical bone surface whereas androgen receptor signaling stimulates bone formation at the periosteal surface.

Biological effects of sex steroids on bone maintenance

Sex steroids are not only essential for bone growth, but also for the maintenance of skeletal integrity as shown by skeletal changes following sex steroid deficiency in men and rodents. Sex steroid deficiency in both sexes always results in high bone turnover; i.e., osteoclast and osteoblast numbers and activity both increase, but the former exceeds the latter resulting in loss of bone mass and strength. Therefore, hypogonadal men, similar to postmenopausal women, develop osteoporosis. Androgens, in analogy with estrogen replacement in postmenopausal women, also prevent bone loss in hypogonadal men [51]. According to studies in transgenic mice, gonadal steroids are able to maintain trabecular bone mass by both AR and ERα activation in both sexes [8]. However, in physiological conditions, AR signaling appears to be the most important for maintenance of trabecular bone mass in males, whereas ERα signaling is the dominant pathway in females [36, 43].

Molecular mechanism of the antiresorptive action of sex steroids

The underlying mechanism of the antiresorptive action of sex steroids during gonadal insufficiency has been studied in great detail in bone marrow (co-)cultures. In these cultures, sex steroids downregulate osteoclast precursors (osteoclastogenesis) in bone marrow. This action indirectly relates to the regulation of the production of multiple cytokines by different cell types in the bone marrow microenvironment [52]. In this context, the antiresorptive action of estrogen appears to be mediated by the downregulation of cytokines which are also involved in osteoclast formation, such as interleukin-1 (IL-1) and tumor necrosis factor α (TNFα) produced by monocytes, and interleukin (IL-6) and granulocyte macrophage-colony stimulating factor (GM-CSF) produced by stromal cells and osteoblasts [53] (Fig. 3). Moreover, transforming growth factor β (TGFβ) mediates the stimulatory effect of E2 on the induction of osteoclast apoptosis in vitro and in vivo, and it directly acts on the osteoclast to decrease its activity [54]. E2 can also suppress osteoclastogenesis through enhanced production of OPG and decreased RANKL by osteoblasts [55]. Interestingly, a direct ERα-mediated anti-apoptotic effect on osteoclasts has been described in female but not male osteoclast-specific ERα knockout mice [56]. Androgens also decrease osteoclast formation and resorption through increased production of OPG by osteoblasts [55, 57]. In addition, androgens—at least in vitro—may also directly reduce osteoclastogenesis independently of osteoblasts [58, 59]. In vitro data also indicate that sex steroids stimulate proliferation and differentiation of osteoblasts and inhibit osteoblasts apoptosis [52, 60].
https://static-content.springer.com/image/art%3A10.1007%2Fs00198-008-0609-z/MediaObjects/198_2008_609_Fig3_HTML.gif
Fig. 3

Schematic representation of the anti-resorptive action of estrogen on bone. Estrogen downregulates the production of cytokines involved in bone resorption, such as IL-1, TNFα, IL-6, GM-CSF, M-CSF. Down regulation of these cytokines subsequently decreases the number and activity of osteoclasts. Estrogen also decreases TGFβ, resulting in a decreased osteoclast number and activity and increased osteoclast apoptosis. Estrogen increases OPG, the decoy receptor of RANKL in binding to its receptor RANK on osteoclasts. Beside the indirect action of estrogen on osteoclast number and activity, estrogens may also directly decrease osteoclast apotosis, at least in females

Sex steroids and age-related bone loss

In both sexes, age-related cortical bone loss is caused by lower periosteal bone formation versus higher endocortical bone resorption. In elderly men and women, ongoing periosteal bone formation is unable to offset endosteal bone resorption, resulting in cortical thinning [61, 62]. In aged rodents, it appears that periosteal bone is less responsive to androgens than estrogens, as androgen deficiency and replacement hardly affects periosteal bone in ageing rodents [63, 64], while estrogens increase cortical thickness through endocortical contraction/apposition [30]. To what extent periosteal bone formation and endocortical bone resorption in elderly men are also related to age-related changes in gonadal steroid concentrations, remains unclear.

Age-related trabecular bone loss in men is primarily caused by thinning of trabeculae, whereas trabecular bone loss in postmenopausal women is typically characterized by a disruption of trabeculae. The resulting loss of connectivity has a much more profound impact on bone strength compared with trabecular thinning leading to similar decreases in trabecular bone volume [65]. The acceleration of trabecular bone loss in women after menopause (as well as after ovariectomy) is caused by the high bone turnover due to estrogen deficiency. Likewise, the same type of trabecular bone loss is observed in hypogonadal men.

Overall, the gonadal androgen T or the adrenal androgen DHEAS only show weak or no correlations with BMD. Serum E2, in particular bioavailable E2, is positively associated with bone density, rates of bone loss and fracture occurrence in elderly men [7, 66]. In addition, E2 is negatively correlated with markers of bone resorption [7]. Interestingly, bone loss seems associated with a polymorphism in the aromatase gene in elderly men, even independently of circulating E2 concentrations [67]. Therefore, men with higher local estrogen synthesis may be better protected against bone loss. Aromatization of androgens into estrogens and subsequent activation of ER may have an important role in male skeletal homeostasis during aging. Whether or not there exists a threshold concentration for serum E2 below which bone loss increases in both men and women remains a matter of debate [68]. In addition, intracrine or paracrine actions of estrogens may be more relevant for skeletal homeostasis in the elderly than previously anticipated. In accordance, aromatase inhibitors cause bone loss and increased fracture risk in postmenopausal women illustrating the importance of estrogen synthesis in these women with already low circulating estrogen concentrations [69].

It is unknown whether the frequently observed subnormal concentrations of bioavailable T in elderly men contribute to bone loss. Accordingly, intervention trials have not yet been able to prove a benefit of T replacement in elderly men with partial androgen deficiency [70, 71]. Therefore, androgen replacement with the specific purpose of promoting bone health and preventing osteoporosis cannot be recommended in this patient group until extensive clinical data on long-term efficacy and safety become available [72].

Selective estrogen and androgen modulators

Estrogen replacement for osteoporosis in women, although clearly effective, remains restricted as a consequence of unwanted effects on uterus, breast and the cardiovascular system [73, 74]. Partial estrogen receptor agonists, having agonistic estrogen-like effects on bone and antagonistic effects on reproductive tissues, have replaced estrogen in the treatment of postmenopausal bone loss. Raloxifene, a second generation SERM, is approved as a drug for both prevention and treatment of postmenopausal osteoporosis. Clinical studies confirmed that raloxifene effectively reduced fracture risk, without adverse effects on reproductive tissues [75, 76]. The recognition of SERMs as agents able to elicit estrogenic effects in a tissue-specific manner remains a major challenge to the pharmaceutical industry to continue to develop new ER ligands that retain the beneficial effects in tissues such as bone and brain, but that lack the mitogenic and perhaps carcinogenic actions in the breast and uterus. SERMs are thought to exert different biological actions in different tissues compared to full agonists, such as E2, via distinct conformational changes in the ER on binding to ligand. These different conformations, in turn, result in tissue specificity, because those mechanisms involved in ER-dependent stimulation of transcription are thought to vary according to tissue and promoter context [77]. It has thus far not been possible to define a single unifying mechanism to explain why SERMs preferentially activate transcription in certain tissues but not others. The successful clinical application of SERMs stimulated a great interest in the discovery and development of selective AR modulators (SARMs). One uniform characteristic of these compounds is that they are not substrates for aromatase and 5α-reductase [78]. Unlike the agonistic-antagonistic properties of SERMs in target tissues, these AR ligands are known to act as full agonists in anabolic organs (bone and muscle) but as partial agonists in androgenic tissues (prostate and seminal vesicles). Preclinical data supporting the bone selectivity of these agents are still rather limited [78]. The goal of SARMs is to reproduce the beneficial effects of androgens in men and women without undesirable effects such as prostate stimulation or acne. The discovery of SARMs not only provides a potentially significant therapeutic advance for androgen replacement therapy [79], but also provides model compounds to further study the molecular mechanism of AR action.

Conclusion

It has now been well established that androgens and estrogens and their respective receptors have specific actions on the different bone envelopes during bone growth and bone maintenance in both sexes. Androgens through AR activation and estrogens through ER activation are both required for optimal stimulation of bone formation as demonstrated by many reports in mice and men. Beside their bone-anabolic actions, androgens and estrogens both exert anti-resorptive effects. Although an important anti-resorptive role has been attributed to ERα in female osteoclasts, the main target cell for AR-mediated anti-resorptive action in males remains to be determined. The search for compounds that selectively stimulate bone and muscle but not other androgenic/estrogenic tissues looks promising and confers a major advantage not only to the improvement of treatment modalities for osteoporosis and sarcopenia but also for the elucidation of the molecular mechanism of sex steroid action in bone cells. As the underlying mechanisms of sex steroids on bone appear extremely complex, future studies must focus on the understanding of different sex steroid receptor signaling pathways and their interaction as a potential drug target for anabolic and/or anti-resorptive therapy.

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2008