Androgen and bone mass in men*
Annie W.C. Kung
Department of Medicine, University of Hong Kong, Queen Mary Hospital, Hong Kong, China
J Androl 2003
Keywords: reactive oxygen species; testis; infertility; malondialdehyde
Androgens have multiple actions on the skeleton throughout life. Androgens promote skeletal growth and accumulation of minerals during puberty and adolescence and stimulate osteoblast but suppress osteoclast function, activity and lifespan through complex mechanisms. Also androgens increase periosteal bone apposition, resulting in larger bone size and thicker cortical bone in men. There is convincing evidence to show that aromatization to estrogens was an important pathway for mediating the action of testosterone on bone physiology. Estrogen is probably the dominant sex steroid regulating bone resorption in men, but both testosterone and estrogen are important in maintaining bone formation.
Androgens have multiple actions on the skeleton throughout life in both males and females. Androgens affect bone size, bone mass and bone remodeling and the mechanisms by which androgens exert these actions are complex.
2 Mechanisms of action on bone
Androgen receptors (AR) as well as estrogen receptors (ER) are present at physiological concentrations and binding affinity in both osteoblasts and osteoclasts [1, 2]. Binding of the sex hormones, both androgens and estrogens, to the receptor protein causes homo- or heterodimerization and conformational changes that allow coactivator proteins to interact with the receptor dimer. Alternatively, steroid hormone receptors suppress transcription by complexing with other transcriptional factors, thus preventing them from interacting with their target gene promoters [3-5].
3 Physiological action of androgens on bone
Androgens have direct effects on osteoblast function. Testosterone, dihydrotestosterone (DHT) and nonaromatizable androgens increase the proliferation of osteoblast-like cells in culture, and induce osteoblast differentiation [6, 7]. The mechanism of action is believed to be through a complex endocrine, paracrine and autocrine fashion. DHT increases the expression of TGF-b mRNA  and increases the mitogenic response to fibroblast growth factor and to IGF-II . Testosterone and DHT reduce cAMP response and prostaglandin E2 production in bone cultures exposed to stimulation with parathyroid hormone (PTH) [9, 10]. The effects of androgen on PTH action suggest that androgen could modulate bone turnover in response to PTH.
Testosterone also increases the lifespan of both osteoblasts  and osteoclasts  by affecting apoptosis . The effect of sex steroids on apoptosis is recently thought to be mediated through non-genotropic effects on bone, one of which is the activation of MAP kinases or PI3 kinases . It was observed that the apoptotic effects of sex hormones on osteoclasts and anti-apoptotic effects on osteoblasts and osteocytes were mediated through the nongenotropic function and resulted from activation of a Src/Shc/ERK signal transduction pathway. The AR or ER can transmit signals through the Src/Shc/ERK signaling pathway with similar efficacy irrespective of whether the ligand is an estrogen or an androgen.
Apart from the direct action from circulating androgens, there is evidence that the cellular effect of testosterone is mediated via its metabolites formed locally in bone as a result of local enzyme activities. Both aromatase and 5-a reductase are present in bone tissue [15-17]. Conversion of androgens to estrogens by the aromatase enzyme is important for many of the androgen actions on bone . At the tissue level, estrogen topically suppresses bone turnover and maintains a balance between bone formation and bone resorption . At the cellular level, estrogen increases osteoblast formation, differentiation, proliferation and function, but decreases osteoclast formation, activity and increases apoptosis, hence decreases osteoclast lifespan. The restraining effects on osteoclastogenesis result from actions of androgens and estrogens on cells of the bone marrow stroma/osteoblastic lineage and in particular, transcriptional down regulation of genes encoding cytokines and/or their receptors (e.g. IL-6, TNF, MCSF, osteopro-tegerin, IL-1RI/IL-1RII), which promote osteoclasto-genesis .
Sex steroids are responsible for skeletal growth and maturation. Skeletal size and volumetric BMD are similar in prepubertal girls and boys. Because of later onset of puberty and longer duration of growth spurts, boys acquire 10 % greater body weight and 25 % greater peak bone mass compared to girls. The greater bone mass in males is due to their greater bone size as testosterone promotes long bone growth, chondrocyte maturation, metaphyseal ossification, periosteal new bone formation and increase in calcium retention and incorporation into bone . The excess in periosteal bone apposition over endosteal bone resorption that occurs during the pubertal growth spurt increases both the size and the volumetric BMD in growing males.
Orchidectomy in young rats results in a reduction in cortical bone mass, which is due in part to a redution of periosteal bone formation. This is in contrast to oophorectomy, which results in an increase in periosteal apposition within 2 weeks~4 weeks [21, 22]. While estrogens appear to increase endosteal bone apposition, androgens have little of such action. These explain the larger cortical bone and thicker cortex seen in young-adult male than in women. By old age, there is an extensive loss of cortical bone due to endosteal resorption in both sexes, but because women undergo menopause, they lose more cortical bone than men. The endosteal resorption associated with aging is partially offset by periosteal apposition, which is 3 times greater in men than in women [23, 24]. A greater bone size in men confers greater mechanical advantage and the loss of bone strength during aging is less in men than women. Bone formed on the peoiosteal surface is biomechanically advantageous because it increases the cross-sectional movement of inertia and bending strength of the long bone .
4 Other actions of androgens
In addition to the direct effects on bone, androgens also affect the skeleton via actions on other tissues. For instance, androgens increase the muscle size and strength and increase mechanical stimuli on bone and these together contribute to reduce the propensity to falling. Androgens can also mediate their effect by increasing IGF-I, which has potent actions on bone . Some of the antiresorptive effect of testosterone is probably mediated by enhancing intestinal calcium absorption .
5 Pattern of bone loss in men
Men exhibit slow bone loss with aging, resulting in overall losses of about 20 %~25 % in both cortical and trabecular bone. Because men do not have the equivalent of menopause, they lack the early, accelerated phase that is induced in women by the precipitous fall in serum estrogens soon after menopause. The slow, age-related decline in bone loss is related to the slow decline in androgen and estrogen level in aging men. This leads to reduction in intestinal calcium absorption and renal calcium conservation and secondary hyperparathyroidism. A reduction in testosterone decreases the rate of periosteal bone apposition, although this is partially offset by the counter effect of estrogen deficiency.
In the unusual circumstances of acute hypogonadism, castrated men have a pattern of rapid bone loss similar to that of women after menopause . Similarly, in drug-induced androgen deficiency such as in patients with prostate cancer treated with LHRH agonist or anti-androgen agents, bone loss is evident within a few months of treatment . This rapid phase of bone loss is associated with an increase in remodeling and bone resorption, as seen in animal models of gonadal insufficiency. This is followed by a slower phase of gradual reduction in BMD leading to increased fracture risk.
6 Relative contributions of estrogen and testosterone in puberty
Recently, the traditional belief that bone mass in men is regulated by androgens have been called into question by rare experiments of nature. Smith et al.  reported a eunuchoid 28-year-old man with homozygous mutation of the estrogen receptor a gene associated with estrogen resistance. The patient had unfused epiphyses and severe osteoporosis despite normal levels of testosterone and DHT and elevated levels of estrogen. Carani et al.  and Bilezikian et al.  each reported a young adult man with aromatase gene mutation and failure to convert androgens to estrogens. Both men had elevated testosterone, DHT and androstenedione levels but undetectable estrogen levels, unfused epiphyses and low bone mass. Administration of estrogen to the two patients with aromatase deficiency resulted in marked improvement of bone mass but not in the patient with estrogen resistance. These cases serve to illustrate the importance of estrogen in establishing peak bone mass in men, and also highlight the possibility that subtle deficiencies in estrogen activity may contribute to low peak bone mass in some men.
In rats, the nonaromatizable androgens DHT decreased biochemical markers of bone turnover and urinary calcium excretion in young rats, although it is unclear whether these effects were due to the skeletal or extraskeletal actions of the androgen . In vitro, estrogen, testosterone and DHT stimulated osteoblast proliferation  and testosterone can prevent orchiectomyCinduced bone loss in ERa knockout mice .
7 Estrogens vs androgens in adulthood
The issue of the role of sex steroids in age-related bone loss has long been subjected to debate. While some studies reported a significant association between androgen concentrations and bone loss in older men [36-39], others failed to substantiate the relation between bone mass and androgens [40-43]. Recent epidemiology studies have demonstrated by multivariate analysis that estrogen, rather that testosterone, was the main predictor of BMD at all sites in older men, except for certain cortical bone sites in the appendicular skeleton [44-52]. Szulc et al.  showed that men with low levels of bio-17bE2 were associated with high levels of biochemical markers of bone turnover and low BMD. Khosla et al. also demonstrated in a group of aging men, the rate of bone loss from the forearm correlated with bioavailable estrogen rather than bioavailable testosterone .
The importance of estrogen in aging men is also demonstrated by experimental models of aged male rats. Orchidectomy and treatment with aromatase inhibitor in these animals produced a similar degree of bone loss . In orchidectomised aged male rats, there was a reduction in cancellous bone area at the tibia and vertebra. Both increase in osteoblast surface and osteoclast number were seen, suggesting that the bone loss in these animals were due to increased bone turnover and activation frequency, which in turn stimulate bone formation. This was accompanied by increase in the urinary excretion of calcium and N-telepeptide, a marker of bone resorption . Estradiol but not testosterone (total or free) was the only significant predictor of bone changes in these animals. Moreover, targeted deletion of the gene for either ERa or aromatase results in decreased BMD in male mice [56, 57]. Putting all the results together suggested that the major action of testosterone is mediated through aromatization to estrogen and binding to the ER.
Falahati-Nini  assessed the relative contributions of estrogen and testosterone on bone turnover by rendering elderly men hypogonadal with GnRH against lenprolide. Conversion of androgens to estrogen was blocked by administration of the aromatase inhibitor letrozole. The subjects then received replacement doses of testosterone and estrogen in turn. The result showed that estrogen prevented the increase in markers of bone resorption, whereas testosterone had only a minor effect. Bone formation markers were maintained by both testosterone and estrogen. The authors inferred that estrogen accounted for at least 70 % of the effect of sex steroids on bone resorption, whereas testosterone, for less than 30 %. Using a similar study design, Khosla et al.  observed that in elderly men treated with both GnRH agonist and aromatase inhibitor, testosterone replacement decreased osteoprotegerin (OPG) level by about 10 %, whereas estrogen replacement increased OPG by 18 %. OPG is a soluble decoy receptor secreted by the stromal-osteoblast lineage cells and serves to neutralize receptor activator of nuclear factor-kB ligand (RANKL) . RANKL is the important osteoblast-derived paracrine effector that stimulates all aspects of osteoclast function, i.e. the final mediator of osteoclasto-genesis. In view of the effect on OPG, Khosla et al. conclude that estrogen plays a more important role than androgen in inhibiting bone resorption in humans. In a similar study in younger individuals, Leder et al.  confirmed the increase in bone resorption following aromatase inhibition, even though an additional independent role of androgens on bone resorption was also observed. Also, Taxel et al.  demonstrated that elderly men being treated with aromatase inhibitor had a significant increase in bone resorption. Collectively, these results strongly suggest that estrogen is the dominant sex steroid regulating bone resorption, but both testosterone and estrogen are important in maintaining bone formation.
8 Bone mass in hypogonadal men
It is well known that abnormal gonadal function in adult men is associated with low bone mass and increased fracture risk . Loss of both cortical and trabecular bone is seen with a more intense loss in trabecular bone volume [63, 64]. Several studies suggest a correlation between bone mineral density and serum testosterone [65, 66], whereas others have not shown this association [37, 44, 67]. There are no prospective studies showing a positive relationship between testosterone and the rate of bone loss in men and no studies showing that men with low testosterone have rapid bone loss. Hypogonadism is a common finding in a significant percentage of men with low bone mass and a wide variety of causes of gonadal failure is associated with osteoporosis [68, 69]. Boonen et al.  reported that men with recent hip fractures had reduced testosterone levels compared to age-matched controls and in subjects with osteoporosis, about one-third had subnormal testosterone level .
9 Androgen replacement in hypogonadal adult men
Testosterone therapy increased bone mineral density in hypogonadal men and the response in the trabecular bone (e.g. spine) appears to be more prominent than in the cortical bone (e.g. radius) [72, 73]. Treatment with testosterone for 18 months increased spine BMD by about 6 % in a group of adult men with hypogonadism, although the increase in radial BMD was insignificant [74,75]. The increased bone remodeling rate associated with hypogonadism declines with testosterone replacement and the biochemical markers of bone resorption also decrease with treatment. In other secondary causes of osteoporosis associated with hypogonadism, such as glucocorticoid excess, renal insufficiency, post-transplantation and alcoholism, testosterone replacement should be beneficial for bone health, but there is little published data in this area. Reid et al.  reported that testosterone therapy improved spine BMD in a small group of men receiving glucocorticoid. Whether testosterone supplement is equally effective in other conditions associated with low testosterone levels remain unclear.
Apart from the increase in bone mass, other additional positive effects of androgen replacement therapy include increase in muscle strength and lean body mass [77, 78], both of which help to promote bone health and reduce fracture risk. The most effective routes and doses of androgen administration for the prevention and treatment of bone loss in hypogonadal men remain uncertain. The transdermal route apparently is as effective as intramuscular administration in improving bone mass . An adequate dosage should be accompanied by a fall in the biochemical markers of bone resorption.
In those patients whose bone mass failed to improve significantly after adequate androgen replacement, antiresorptic therapy with the bisphosphonates has been shown to be effective . Whether low dose intermittent administration of PTH can produce similar anabolic action at trabecular bone as seen in eugonadal men with idiopathic osteoporosis is unclear .
More recent studies tried to dissect the non-genotropic actions of sex steroids from the classical genotypic-action. Administration of synthetic ligands which can dissociate AR or ER signaling through kinases from the classical transcriptional activity of these receptors in the nucleus to either male or female gonadectomised mice causes more significant increases in BMD than androgens and estrogens without affecting the reproductive organs . It is possible that agents of this sort may represent a novel class of gender neutral pharmacotherapeutics that may improve sex steroid replacement therapy in both sex.
10 Androgen therapy in eugonadal men
It has been hypothesized that androgens, with its anabolic action, may benefit osteoporotic men even in those with normal testosterone level. There are limited data in this aspect. A few small trials of androgen administration in older men suggest that there may be beneficial effects on muscle strength and improvement in body composition, but the effects on bone mass or biochemical indices of remodeling are inconclusive [81, 82]. Snyder et al.  studied a group of elderly men over 65 years and observed that increasing serum testosterone concentration with testosterone patch to the mid-normal range for young men did not increase lumbar spine BMD. However, in those men with subnormal pretreatment serum testosterone levels, testrosterone therapy did show a positive effect on BMD and reduced the markers of bone resorption. Thus, until more definitive data are available, testosterone should not be utilized for prevention or reversion of bone loss in aging men unless there is convincing evidence for androgen deficiency.
In summary, androgens have complex actions on the skeleton. However, the relative importance of androgen during different phases of skeleton development is still far from clear. There is fairly convincing evidence that aromatization to estrogen is an important mechanism for mediating the action of androgens on bone physiology. Little is known of how the genotropic and nongenotropic actions are coordinated. Finally, the usefulness of androgens or androgen analogues, including the non-aromatizable androgen in skeletal therapeutics is unknown. The usefulness of assessing estrogen level in metabolic bone disease associated with hypogonadism in male and for monitoring of testosterone therapy is unclear.
DS, Eriksen EF, Keeting PE, Wilson EM, Lubahn DB, French FS, et al.
Identification of androgen receptors in normal human osteoblast-like cells.
Proc Natl Acad Sci U S A 1989; 86:854-7.
to: Professor Annie Kung,
Department of Medi-cine, University of Hong Kong, Rm 420, Block K, Queen
Mary Hospital, 102 Pokfulam Road, Hong Kong, China.