IT HAS been almost axiomatic that the establishment and maintenance of peak bone mass in women and men is, in large part, a function of their complement of sex steroids. The simplest formulation has placed androgens in a position of primacy for men and estrogen in a similarly important position for women. Men achieve greater bone mass than women because androgens are believed to be more important for the acquisition of bone size. With larger bones, especially in their cross-sectional diameter, men have greater areal bone density when skeleton maturity is reached. When true bone density is taken into account, that is, a volumetric quantity (g/cm3), bone density is not substantially different between the sexes. Nevertheless, the extent to which there is gender specificity in densitometric parameters has been explained primarily by the differential influences of androgens and estrogens on men and women, respectively. The discovery by Smith et al.1 of a man with a mutation of the estrogen receptor shattered these simplistic sex-steroid assignments. The young man described in this seminal report had a completely normal complement of androgens and androgen receptors but showed continuous longitudinal skeletal growth, even in young adulthood. He had severe osteoporosis. It was established subsequently that the genetic lesion was a point mutation in the α-estrogen receptor that had no affinity for endogenous or exogenous estrogens. This experiment of nature, a human gene knockout, if you will, necessitated a revision of the old truism that estrogens were relatively unimportant for male skeletal health. The old truth was further undermined when two men with aromatase deficiency were discovered.2-4 These men, discovered in their early adulthood, were born with point mutations in the aromatase gene responsible for the conversion of androgens to estrogens. Despite high levels of androgens, in one of the cases, and unmeasurable estrogen levels, in both cases, these men displayed a phenotype that was unmistakably similar to the man with the α-estrogen receptor defect. Longitudinal growth was continuous and they had severe osteoporosis. When estrogen was administered to these men with aromatase deficiency, longitudinal growth ceased and bone mass improved dramatically. In these two gene knockout experiments, in which the growing male skeleton had never been exposed to estrogens from birth either by estrogen receptor resistance or estrogen absence, the point was clearly made: estrogens are key for the control of longitudinal growth and for optimal mineralization of the male skeleton. Inactivating mutations of the α-estrogen receptor or the aromatase enzyme cannot address questions related to the importance of estrogens in maintenance of the male skeleton after peak bone mass has been achieved. A compelling body of other data confirmed that estrogens also are important in the maintenance of the male skeleton. Cross-sectional and longitudinal studies have shown that age-associated declines in male bone mineral density (BMD) are more directly related to declining estrogen levels than to declining androgen levels.5-7 These observations have been made both with respect to the slow decline in total sex steroid levels with aging as well as to more overt reductions in bioavailable estrogen and androgen levels that occur with aging.8-10 More dynamic investigations in which older men are rendered temporarily hypogonadal with gonidotropin-releasing hormone (GnRH) agonist therapy and in whom conversion from androgens to estrogens is blocked by aromatase inhibition have shown that the control of bone turnover in older men is under the predominant control of estrogens, with a minority of influence being attributed to androgens.11 The cumulative data argue for an important role for estrogen in the development and maintenance of bone mass in men. However, they do not discount an important effect of androgens in influencing bone size in the man, nor do these findings negate an important role for androgens in maintaining bone mass. In the counterpart syndrome, testicular feminization syndrome, in which the androgen receptor is defective, bone density in these genotypic males is reduced.12 In male rats, furthermore, it has been shown that after castration, either estrogens or nonaromatizable androgen can prevent bone loss.13, 14 Insights into the respective roles for estrogen and androgens in the male skeleton have been advanced further by observations in genetic models in which specific genes for the α-estrogen receptor and the β-estrogen receptor are knocked out singly or together as well as models in which the aromatase gene has been knocked out.15-17 The “experimental” flow is of historical interest, namely, from the initial characterizations of these gene knockouts in human subjects to mouse models. One must eventually return to the human gene knockout if there are inconsistencies between the animal and the human ones. The first discrepancy to be observed is that the mouse models are not as “clean” as the human models. The phenotypes of mice rendered deficient in the α- or β-receptor or deficient in aromatase activity do not mirror well the overt human phenotypes described previously. Besides less dramatic phenotypic changes, there are important differences between mouse and human models. For example, in the human, the lack of the α-estrogen receptor or the aromatase gene is associated with continuous longitudinal bone growth. In the animal models, defects in these same genes are associated with reduced long bone growth. Moreover, in the animal models, it is not clear whether estrogen deficiency is associated with high or low bone turnover, whereas in the human models, high bone turnover is observed consistently. Nevertheless, the animal gene knockouts have been instructive, helping us to understand further the role of estrogens in male skeletal development and maintenance. In an article recently published by Vidal et al.,18 characterization of male mice deficient in the α-estrogen receptor (ERKO) or the β-estrogen receptor (BERKO) or both (DERKO) has helped to establish further a role for specific estrogen receptors in the acquisition of skeletal mass in the male. The importance of the α-estrogen receptor in a number of skeletal characteristics of the male was clearly shown. Foreshortened long bone growth was seen only in the ERKO and DERKO mice but not in the BERKO animals, clearly attributing this estrogenic function to the α-receptor, not the β-receptor. In addition, significant reductions in mice lacking the α-receptor, not the β-receptor, were observed in total body mineral content, femoral bone mineral content, and spine bone mineral content. The abnormalities persisted in the ERKO and DERKO mice when animal weight and skeletal length were accounted for. On the other hand, cancellous bone density as analyzed by peripheral quantitative tomography (pQCT) and by bone histomorphometry was unchanged in all animals. Cortical bone mineral content by pQCT was diminished in ERKO and DERKO animals as compared with wild-type (WT) animals. The reduction in cortical bone mineral content was caused by reduced cross-sectional area primarily because of smaller periosteal and endosteal circumferences. Similar to cancellous bone density, cortical volumetric density was unchanged. Mechanical strength was diminished in the ERKO animals with a tendency for a similar decline in the DERKO animals. Observations with these knockout mice clearly established differences in the α- and β-estrogen receptors with all abnormalities observed being clearly a function of the α-receptor, not the β-receptor. However, in a number of important respects, the knockout animals did not confirm expectations that cancellous or cortical bone density would be diminished in the α-receptor-deficient knockout animal. These points are important but not clearly understood. It could be that the estrogen receptors are just not important in these aspects of the male mouse skeleton. Alternatively, it is possible that compensatory mechanisms were induced to adjust for the receptor deficiency. Estrogen and testosterone levels are elevated in these animals. Elevated levels could have accounted for the lack of some of these expected changes. It is even possible that the elevated estrogens in these knockout animals was somehow activating a truncated version of the α-receptor that has been shown to be present in α-receptor knockout animals.19 Consistent with this point, Gentile et al. have shown that in DERKO animals, estrogens can rescue the loss in bone density after ovariectomy when very high levels of estrogens are used.20 In this issue of the Journal,21 Lindberg et al. have continued their study of the relative importance of the α- and β-estrogen receptor on the male mouse skeleton by taking a different tact. Recognizing that the compensatory adjustments in sex steroids caused by estrogen receptor deficiency may have masked observations that might otherwise have been made; they subjected WT and the same knockout animals, ERKO, BERKO, and DERKO, to orchidectomy at 7 months of age. Their experimental premise is that elimination of all sex steroids by orchidectomy would permit assignment of a specific role for estrogens when 17β-estradiol was used subsequently to prevent the abnormalities induced by the abrupt loss of androgens. However, it should be noted that this new model does not solve problems presented by the original model per se. The abrupt removal of androgens and estrogens at 7 months of age by orchidectomy will provide information about the relative importance of estrogen and its receptors in maintaining the skeleton of these animals rendered androgen deficient but does not address issues related to the respective roles of androgens and estrogens in establishing the male skeleton. Those issues will continue to be problematic in these knockout animals for the reasons given previously. After orchidectomy, total body areal BMD fell significantly in all groups. However, only in the animals with an intact α-estrogen receptor (WT, BERKO) could estrogens maintain bone density at baseline, preorchidectomy levels. It would appear valid to conclude that the lack of a β-estrogen receptor did not impair the effect of estrogen on the α-receptor and that the α-receptor alone was sufficient to prevent the loss in total body BMD due to orchidectomy. Similarly, animals with an intact α-estrogen receptor (WT, BERKO) were able to respond to estrogen such that femoral areal BMD and femoral bone mineral content were maintained. When femoral areal BMD and femoral bone mineral content were examined directly after orchidectomy and estrogen replacement, again only the mice with intact α-estrogen receptors (WT and BERKO) showed levels that were higher than postorchidectomy levels. It is unfortunate that baseline, preorchidectomy levels were not shown in this experiment. If one can compare such baseline data from their earlier article,18 always a treacherous proposition, it appears that estrogen was not able to protect completely against some loss in femoral areal BMD and femoral bone mineral content. However, without direct data shown from the experiment at hand, it is difficult to be sure about such conclusions. This experiment would have been strengthened further if a nonaromatizable androgen were used in a separate experimental arm. If the idea is that estrogen, acting through the α-receptor, is interchangeable with androgen in functions such as those reported in Fig. 2 of Lindberg et al. (maintenance of postorchidectomy femoral areal density and femoral bone mineral content), it would have been desirable to have made the comparison directly. The inference, from the published literature, suggests but does not prove that androgens also would have been equally effective in restoring such deficiencies.1, 2, 22 The concern here is that because estrogen might not have maintained these parameters at baseline preorchidectomy levels, androgens conceivably might have been more effective. Alternatively, perhaps both an estrogenic and androgenic effect might have been required to maintain completely femoral areal density and femoral bone mineral content at preorchidectomy levels. Therefore, a more rigorous experiment design is needed before concluding that there is “redundancy” in the system, with α-estrogen action and androgens serving as interchangeable agonists to protect orchidectomy-associated bone loss. The need for this kind of experimental approach assumes even greater significance in light of recent provocative data from Kousteni et al.23 who showed that with respect to certain nongenomic aspects of sex steroid action such as apoptosis, androgens, estrogens, and their receptor loci that mediate these effects appear to be interchangeable. A possible loss of sex-steroid identity cannot easily account for the data in the model of Lindberg et al. because the restorative estrogen effect in these male mice was specifically a function of the α-estrogen receptor. Lindberg et al. nicely distinguish between cancellous and cortical elements by using pQCT and histomorphometry. Although it is stated that cancellous bone density was not diminished in ERKO animals, nevertheless, treatment with 17β-estradiol was able to prevent orchiectomy-associated declines only in the animals with the α-receptor (WT and BERKO). These findings were confirmed by histomorphometric analysis. Similarly, cortical bone parameters that were diminished in BERKO and DERKO animals were maintained after orchidectomy by estrogen only in the mice with the α-estrogen receptor. Again, we are not told what the preorchidectomy baseline values were so that whether estrogen was able to maintain completely or partially these parameters in reference to preorchidectomy values is not clear. The results of these elegant experiments clearly provide further confirmation of the importance of estrogens in developing and maintaining the male skeleton. They focus on the α-estrogen receptor as the most important in this respect. It is premature to conclude that estrogens and androgens are interchangeable sex steroids in this experimental paradigm because the direct head-to-head experiment was not conducted. Furthermore, one would have to show this interchangeability over a host of geometric, densitometric, and metabolic parameters, not just a selected few. Finally, what are we to conclude about the β-estrogen receptor? The data would seem to argue that this estrogen receptor has little if any role to play in male skeletal metabolism. However, in female animals age-related reductions in cancellous bone seem be diminished in animals lacking the β-estrogen receptor. These observations suggest that the β-estrogen receptor may play a permissive role in age-related bone loss or that by virtue of its absence, sensitivity to the protective effects of the α-receptor is enhanced.24, 25 More studies are needed on a putative role for the β-estrogen receptor in male skeletal metabolism. Nevertheless, the young man with the α-estrogen receptor defect, but presumably with a competent β-receptor, is osteoporotic.1 Thus, at least in the human, the β-estrogen receptor does not have an important role to play in the acquisition of peak bone mass. It is clear that the estrogen knockout experiments are providing a new paradigm with which to ask key questions of the role of estrogens in the skeletal development and skeletal maintenance of the male. The key questions to address with regard to sex steroids and the male skeleton can be summarized briefly. What specific roles do androgens and estrogens play in the development and maintenance of skeletal mass in the male? What quantitative and qualitative properties are functions of androgens or estrogens? Of functions governed by the estrogens, what specificity can be attributed to α-estrogen receptors versus β-estrogen receptors? Finally, are there any aspects in the male skeleton that can be served interchangeably by androgens and estrogens? The knockout paradigms nicely described in this article as well as those in the human along with careful physiological studies in human male subjects will help to shed even greater light on the fascinating interplay among the sex steroids and male skeletal metabolism.
No takes yet. Share an insight, caveat, or question.
John P. Bilezikian (2002) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: