A functional androgen receptor is not sufficient to allow estradiol to protect bone after gonadectomy in estradiol receptor–deficient mice

Although the role of estradiol in maintaining bone mass is well established, the relative contributions of the estradiol receptors ERα and ERβ and of the androgen receptor (AR) remain controversial. To determine the role of ERα-mediated, ERβ-mediated, and non–ER-mediated mechanisms in maintaining bone mass, gonadectomy and estradiol treatment were studied in ER-knockout mice. Estradiol treatment of ovariectomized ERαβ^–/– mice failed to prevent bone loss, precluding significant effects of estradiol on bone through non–ER-signaling pathways. In contrast, estradiol prevented ovariectomy-induced bone loss in ERβ^–/– mice, as in WT males and females, indicating that ERα is the major mediator of estradiol effects in bone. No response of bone to estradiol was detected in orchidectomized ERα^–/– mice, suggesting estradiol cannot protect bone mass via the AR in vivo. In contrast to female ERαβ^–/– and male ERα^–/– mice, female ERα^–/– mice were partially protected against ovariectomy-induced bone loss by estradiol, confirming that ERβ mediates estradiol effects in bone, but only in females and with a lower efficacy than ERα. We conclude that ERα is the main effector of estradiol’s protective function in bone in both male and female mice, and that, in its absence, AR is not sufficient to mediate this response.

The decrease in circulating sex hormone levels after menopause is the major cause of osteoporosis in women and is initially associated with a high level of bone turnover (1). Both the bone loss and the high bone turnover can be prevented by estradiol treatment (2). Similarly, estradiol appears to regulate bone volume in men, since osteoporosis is observed in men and male rodents deficient in aromatase and in one man with a truncated, nonfunctional estradiol receptor-α (ERα) (3–5). Tissue response to estradiol is classically considered to be mediated by two ER subtypes: ERα and ERβ. While both subtypes are expressed in bone (6–8), their relative functional importance remains unclear. Adding further complexity, it has also been suggested, based upon in vitro data, that estradiol may act in a nongenotropic manner via the androgen receptor (AR) in bone (9, 10). To address these questions, several groups, including ours, generated mutant mice in which ER subtypes were inactivated (11–13). Some of these mutants, however, express shortened ER transcripts that are able to bind estradiol and that have been shown recently to mediate at least the endothelial effects of estradiol (14–17). In contrast, the knockouts used here do not express any shortened transcripts with either ligand- or DNA-binding ability (18); this allowed us to determine the response of each ER-mediated or non–ER-mediated pathway to estradiol in vivo. Note that in order to distinguish between these models, we refer to the full knockouts as ERα^–/–, ERβ^–/–, and ERαβ^–/–, and to the incomplete knockouts as αERKO, βERKO, and αβERKO following notation used previously.

Using the mutant mice in which ERs are fully deleted, we previously showed that deletion of ERα, in both ERα and ERαβ male knockouts, resulted in high trabecular bone volume (BV/TV) and trabecular bone mineral density (Tb.BMD) in the presence of low bone turnover. Whereas ERβ deletion in males did not result in any modification of bone mass or turnover (13), both ER subtypes appeared to regulate bone turnover in female mice. High BV/TV and Tb.BMD were observed in ERα^–/– and ERβ^–/– females, while, in the absence of both receptors, BV/TV and Tb.BMD were low (13). The bone phenotype in these mice was complicated, however, by effects of ER deletion on the reproductive system; very high levels of circulating testosterone were observed in male and female ERα^–/– and male ERαβ^–/– mice, and very high circulating estradiol was detected in female ERα^–/– mice (13).

In order to avoid such endocrine feedback mechanisms, and to determine the signaling pathways by which estradiol can protect bone from gonadectomy-induced bone loss, male and female knockout mice were gonadectomized and treated with estradiol. In addition, the contribution of elevated sex hormone levels to the bone phenotype of male and female ERα^–/– mice was determined by treatment of gonadectomized ERα^–/– mice with testosterone, antiestrogens, or antiandrogens. Our results clearly establish that in male mice, ERα is the only receptor that mediates the protective effects of estradiol, while in female mice, ERβ also plays a minor role. No response to estradiol via the remaining AR could be detected, even at suprapharmacological estradiol concentrations.

Effects of gonadectomy. Significant trabecular bone loss was observed after gonadectomy in all male knockouts and in female ERα^–/– and ERβ^–/– mice (Figure 1, a and b) and was associated with elevated bone turnover (see Figures 2c, 3c, 5b, 6c). In ERαβ^–/– females, the already low BV/TV was not reduced further (Figure 1b), nor was there any change in osteoblast surface or osteoclast surface as percentages of trabecular bone surface (ObS/BS, OcS/BS) after ovariectomy (mean ObS/BS ± SEM: sham-operated, 19.9% ± 5.0%; ovariectomized, 23.9% ± 5.5%; mean OcS/BS ± SEM: sham-operated, 14.1% ± 2.7%; ovariectomized, 14.1% ± 1.5%).

Figure 1

Response of WT and ER knockout mice to gonadectomy. Mice were sham-operated (black bars) or gonadectomized (white bars) at 12 weeks of age, and bones were collected after 8 weeks. (a) Tb.BMD and BV/TV were significantly reduced in male mice of all genotypes after orchidectomy. (b) Tb.BMD and BV/TV were significantly reduced after ovariectomy in female mice of all genotypes except ERαβ^–/–. Values are mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 vs. sham-operated of the same genotype.

Figure 2

Male ERα^–/– mice do not respond to estradiol (E₂) treatment. (a) Representative von Kossa–stained sections of the proximal tibia from WT and ERα^–/– males sham-operated, orchidectomized (Orx), or orchidectomized and treated with slow-release pellets delivering 50 μg/kg/d estradiol (Orx+E₂). (b) BV/TV and Tb.BMD were increased after estradiol treatment in WT mice, but estradiol treatment did not alter BV/TV or Tb.BMD in ERα^–/– males. (c) ObS/BS, bone formation rate expressed as a percentage of trabecular bone surface (BFR/BS), and OcS/BS were all elevated after orchidectomy in WT and ERα^–/– males. This was prevented by estradiol treatment in WT but not in ERα^–/– males. The increased ObS/BS and OcS/BS were prevented by estradiol treatment in WT males, but not in ERα^–/– males. BFR/BS was further elevated in WT males treated with estradiol but remained unchanged in treated ERα^–/– males. Values are mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 vs. sham-operated of the same genotype; ⁺⁺P < 0.01, ⁺⁺⁺P < 0.001 vs. Orx of the same genotype.

Figure 3

Testosterone prevents orchidectomy-induced bone loss in WT and ERα^–/– males. Antiandrogen treatment partially reproduced the effects of orchidectomy. (a) Representative von Kossa–stained sections of the proximal tibia from WT (left) and ERα^–/– (right) males that were sham-operated (Sham), orchidectomized (Orx), orchidectomized and treated with 10 mg/kg/d subcutaneous propiotestosterone (Orx+T), or sham-operated and treated subcutaneously with the antiandrogen RU58642 (30 mg/kg/d) (Sham+aA). (b) BV/TV and Tb.BMD were protected from bone loss after orchidectomy by testosterone treatment. Antiandrogen treatment partially reproduced the bone loss observed after orchidectomy when administered to sham-operated males of both genotypes. (c) The increased ObS/BS, BFR/BS, and OcS/BS after orchidectomy were prevented by testosterone treatment in both WT and ERα^–/– males. Antiandrogen treatment increased these markers of bone turnover in sham-operated males of both genotypes. Values are mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 vs. sham-operated of the same genotype; ⁺P < 0.05, ⁺⁺P < 0.01, ⁺⁺⁺P < 0.001 vs. Orx of the same genotype.

Figure 5

Estradiol treatment affected bone volume in a similar manner in WT and ERβ^–/– females, but effects were less dramatic in ERα^–/– females. (a) WT, ERα^–/–, and ERβ^–/– females were ovariectomized and treated with 1, 10, and 100 μg/kg/d E₂; ERα^–/– females were also treated with 1,000 μg/kg/d. Estradiol treatment of ovariectomized (Ovx) WT and ERβ^–/– females prevented trabecular bone loss associated with ovariectomy. In ERα^–/– females, estradiol treatment at 100 μg/kg/d was the only dosage at which estradiol prevented the reduction in Tb.BMD and trabecular thickness (TbTh), and still this effect was only partial. (b) In ERα^–/– females, estradiol treatment reduced ObS/BS, OcS/BS, and osteoid volume (OV/BV) in a dose-dependent manner. Any effects of estradiol treatment on these parameters in WT and ERβ^–/– females were less pronounced.Values are mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 vs. sham-operated of the same genotype; ⁺P < 0.05, ⁺⁺P < 0.01, ⁺⁺⁺P < 0.001 vs. Ovx of the same genotype.

Figure 6

Testosterone treatment of ERα^–/– females prevents ovariectomy-induced bone loss and high bone turnover. Antiandrogen treatment, but not antiestrogen treatment, partially reproduces the effects of ovariectomy in ERα^–/– females. (a) Representative von Kossa–stained sections of proximal tibiae from sham-operated mice, ovariectomized (Ovx) mice, ovariectomized mice implanted with pellets delivering 2.5 mg/kg/d subcutaneous propiotestosterone (Ovx+T), and sham-operated females treated subcutaneously with the antiestrogen RU58668 (10 mg/kg/d) (Sham+aE) or the antiandrogen RU58642 (30 mg/kg/d) (Sham+aA). (b) Testosterone treatment (light gray bars) prevented the ovariectomy-induced reduction in Tb.BMD and BV/TV in ERα^–/– females. Antiandrogen treatment (dark gray bars), but not antiestrogen treatment (medium gray bars), reduced Tb.BMD and BV/TV to levels not significantly different from post-ovariectomy levels. (c) Testosterone treatment prevented the ovariectomy-induced increase in ObS/BS, BFR/BS, and OcS/BS in ERα^–/– females. Antiandrogen, but not antiestrogen, treatment increased ObS/BS and OcS/BS. Values are mean ± SEM. **P < 0.01, ***P < 0.001 vs. sham-operated of the same genotype; ⁺P < 0.05, ⁺⁺P < 0.01, ⁺⁺⁺P < 0.001 vs. Ovx of the same genotype.

ERα is required for bone response to estradiol in males. Treatment of WT orchidectomized male mice with high-dose estradiol (50 μg/kg/d, raising circulating levels to between 100 and 200 pg/ml) not only prevented orchidectomy-associated bone loss but also elicited an increase in bone mass compared with that of sham-operated animals (Figure 2, a–c). BV/TV was elevated approximately fourfold (Figure 2b), and Tb.BMD was almost doubled (Figure 2b); these effects were associated with a reduction in bone turnover, indicated by lowered ObS/BS and OcS/BS (Figure 2c), as well as reduced serum osteocalcin and urinary deoxypyridinoline cross-links (data not shown). Surprisingly, despite these reductions in bone turnover, bone formation rate was elevated in estradiol-treated male mice, possibly due to reduced removal of fluorochrome labels in the absence of normal osteoclastic bone resorption. In contrast, however, estradiol treatment did not prevent trabecular bone loss or the increased bone turnover associated with orchidectomy in ERα^–/– males (Figure 2, a and c). Furthermore, estradiol treatment of intact (non-orchidectomized) ERα^–/– males did not alter bone turnover, trabecular or cortical BMD, BV/TV, or cortical thickness (data not shown). These observations indicate that estradiol (and/or aromatized testosterone) is unable to play an osteoprotective role in males in the absence of ERα, despite the presence of functional ARs. Thus, the bone loss observed after orchidectomy in these mice most likely relates to loss of protective effects of testosterone that act on the skeleton directly via the AR (see below).

Bone loss after orchidectomy of ERα–/– mice is due to lack of testosterone. We therefore investigated the role of the high levels of circulating testosterone detected in ERα^–/– males (5 ± 1.9 vs. 1.2 ± 0.4 ng/ml) (13) by treating male orchidectomized WT and ERα^–/– mice with a high dose of testosterone (10 mg/kg/d, giving 22.8 ± 3.4 and 33.3 ± 3.0 ng/ml circulating levels, respectively) or with antiandrogens. Testosterone treatment maintained BV/TV and Tb.BMD in both WT and ERα^–/– males (Figure 3, a and b). Testosterone treatment also maintained the low bone turnover characteristic of ERα^–/– males and reduced WT bone turnover to levels equivalent to those in ERα^–/– males (Figure 3, a–c). Treatment of non-orchidectomized WT and ERα^–/– males with pure antiandrogens (RU58642 and RU50818) partially mimicked the effects of orchidectomy by reducing Tb.BMD and BV/TV and increasing bone turnover (effects of RU58642 are shown in Figure 3, a–c). Increased bone resorption was confirmed in WT and ERα^–/– males treated with antiandrogens by an increase in urinary deoxypyridinoline cross-links (data not shown). In ERα^–/– males, in which circulating testosterone is markedly elevated (see above) (13), antiandrogens were also effective in mimicking castration. This suggests that it is indeed the high level of circulating testosterone that is osteoprotective in ERα^–/– males and establishes the functionality of their ARs, thereby confirming that the protective effects of estradiol on the male skeleton require the presence of ERα.

ERα is also required for bone responses to estradiol in females. In ERαβ^–/– females, not only did ovariectomy fail to decrease bone mass further (Figure 1), but estradiol, even at doses that increase bone mass in WT animals, failed to elicit any effect on trabecular bone (Figure 4) or cortical bone (not shown), indicating that the ability of the skeleton to respond to estradiol treatment had been entirely ablated with deletion of both ERs. Deletion of both receptors also prevented uterine response to estradiol (data not shown). Thus, the estradiol responses in bone and the uterus are clearly mediated by either or both ERs and cannot be mediated by non–ER-signaling pathways, such as AR. This allowed us to determine the physiological effects of estradiol on each receptor alone by treating the female single ER-knockout mice with estradiol.

Figure 4

Estradiol treatment of female ERαβ^–/– mice does not significantly alter trabecular bone mass or bone turnover. Intact ERαβ^–/– females were implanted with slow-release pellets delivering estradiol at 30 μg/kg/d. Tb.BMD and BV/TV were not increased by high-dose estradiol treatment. No markers of bone turnover were significantly altered. Shown here are ObS/BS, BFR/BS, and OcS/BS. Values are mean ± SEM.

ERβ facilitates the ERα-mediated responses of bone to estradiol in females. In contrast to the double knockouts, estradiol treatment elicited a skeletal response in WT and in both single-ER-knockout females. In WT females, as in male mice, trabecular BMD and BV/TV were significantly elevated compared with those of sham-operated mice in response to high-dose estradiol treatment, while low doses of estradiol simply prevented the bone loss associated with ovariectomy (Figure 5a). This increase in bone mass was associated with a reduction in bone turnover, measured both by histomorphometry (Figure 5b) and by serum osteocalcin and urinary deoxypyridinoline cross-links (data not shown, and Table 1). A similar response was observed in the absence of ERβ, although a tenfold higher dose of estradiol was required to prevent ovariectomy-induced bone loss in ERβ^–/– females (Figure 5a), indicating that the remaining receptor, ERα, mediates most of the response to estradiol observed in WT females. These data also suggest that there may be a cooperation between the two receptors in eliciting the skeletal response, ERβ facilitating the effects of ERα in bone. Indeed, even in the absence of ERα, protection against ovariectomy-induced changes in bone turnover was observed (Figure 5b), confirming that in the absence of ERα, ERβ can mediate some of the effects of estradiol in the bones of female mice. Interestingly, however, this does not translate into a measurable effect on bone volume (Figure 5b), probably because, as we suggested before (13), ERβ is not as efficient at reducing bone resorption (Figure 5 and Table 1).

Table 1

Effects of estradiol treatment on bone resorption in ovariectomized WT, ERα^–/–, and ERβ^–/– females, as measured by urinary deoxypyridinoline cross-links^A

ERα–/– females respond normally to testosterone treatment. Given that these results suggested that AR did not respond to estradiol in females, confirming what we observed also in males, we then tested for the presence of functional AR responses in these animals. Testosterone treatment in ovariectomized ERα^–/– mice maintained the high BV/TV and Tb.BMD characteristic of ERα^–/– females and, as in males, maintained a low bone turnover (Figure 6, a–c). Since ERα^–/– females demonstrate both high testosterone (4.87 ± 0.4 ng/ml) and high estradiol levels (96 ± 16 pg/ml), antiestrogen and antiandrogen treatment were then administered to determine whether either treatment could reproduce the effects of ovariectomy in ERα^–/– females. In non-ovariectomized female ERα^–/– mice, as expected, antiestrogen treatment did not significantly alter bone volume or bone turnover, confirming in these animals that even high levels of estradiol are not capable of mediating any bone response in the absence of ERα. In sharp contrast, antiandrogen treatment reduced BV/TV and induced an increase in bone turnover, confirming (a) that testosterone plays an important role, acting via AR rather than via ERs after aromatization, in determining the bone phenotype of ERα^–/– females (Figure 6 a–c), and (b) that the ARs expressed in ERα^–/– females are indeed functional.

Functional ARs are expressed at high levels in ER knockout mice. Given that all the results reported above indicated that estradiol cannot act in bone via the AR, despite its verified functionality, we needed to verify the levels of expression of the AR in the bone of these genetically altered mice. Real-time RT-PCR on RNA extracted from WT and knockout femora demonstrated that AR mRNA was elevated twofold in ERα^–/– females, but present at normal levels in all other mutants. Expression levels of ERα, ERβ, and aromatase were not significantly altered in either male or female knockouts (Table 2). The functionality of these ARs in bone was demonstrated in the experiments reported above that show the ability of testosterone, but not estradiol, to elicit the proper bone responses after gonadectomy. Thus, even at high levels of AR expression, there were no protective effects of estradiol through the AR in female or male ERα^–/– mice.

Table 2

Deletion of ERα, in ERα^–/– females, results in increased expression of AR mRNA

The importance of estrogens for the development, maturation, and maintenance of the skeleton is now well established; however, the exact contribution of each receptor subtype remains elusive. Initial analysis of the bone phenotype of ERα^–/–, ERβ^–/–, and ERαβ^–/– mice revealed that while ERβ was totally dispensable in males, both receptor subtypes participate in the control of bone turnover in females (13). In the present report, we show that in male mice, estradiol regulates trabecular bone remodeling exclusively through ERα, while in females both ER subtypes influence bone turnover and trabecular structure. Thus, in both males and females, ERα is the main mediator of estradiol’s protective effects in bone, and the AR, although expressed at high levels and functional in response to androgens in the same animals, is not able to mediate the effects of estradiol on bone in vivo in the absence of ERα. We conclude that either the AR cannot mediate the bone responses to estradiol in vivo, or ERα is required to confer this property to the AR.

In male mice, trabecular bone loss occurred after orchidectomy in all ER knockouts, indicating that male gonadal hormones are responsible for the maintenance of BV/TV even in the absence of estrogen receptors. This suggests that it is indeed testosterone binding to the AR, and not estradiol derived from aromatization of testosterone and binding to ERs, that maintains bone mass in male mice. Kousteni et al. (9, 10), using in vitro studies, have recently suggested that estradiol, acting indifferently via the AR or the ERs, prevents osteoblast apoptosis via non-genotropic pathways. Whether or not this mechanism translates in vivo into the prevention of bone loss associated with gonadectomy in males and females remained to be tested. In apparent contradiction to this hypothesis, however, we show here that even a maximal dose of estradiol that is able to increase bone mass in intact mice is unable to prevent gonadectomy-induced bone loss when administered to male ERα^–/– mice. These in vivo data confirm that ERα is required for the osteoprotective effects of estradiol in male mice and exclude a role of the AR in mediating these responses. Thus, the reported antiapoptotic effects of estradiol mediated through non-genotropic pathways (9, 10) are either irrelevant to bone responses in vivo or insufficient to protect against bone loss when ERα is inactivated. It remains, however, possible that the AR participates in these responses but only when ERα is expressed.

Furthermore, the fact that, in the absence of ERα, high doses of estradiol were unable to prevent orchidectomy-induced bone loss in males further demonstrates that, like AR, and in contrast to the situation in females, ERβ is unresponsive to estradiol in males. This finding is consistent with our previous result showing the lack of a bone phenotype in male ERβ^–/– mice, and the identical phenotypes of male ERα^–/– and ERαβ^–/– mice (13).

In contrast to estradiol, however, and as expected, testosterone is capable of maintaining bone volume in males, independent of ERα. We show here that testosterone treatment prevented orchidectomy-induced bone loss in both WT and ERα^–/– males. This is consistent with effects of testosterone in vitro (23), in hypogonadic men (24, 25), and in αERKO mice (26), clearly showing that testosterone inhibition of bone turnover and prevention of bone loss are not exclusively mediated by aromatization of testosterone into estradiol. Thus, these results show that the high circulating testosterone level in male ERα^–/– and ERαβ^–/– mice is the main cause of the elevated BV/TV and low bone turnover observed in the gonad-intact knockouts (13). This was further confirmed by the fact that we can reproduce the effects of orchidectomy on bone by treating intact ERα^–/– males with antiandrogens. While trabecular bone loss and high bone turnover were observed after antiandrogen treatment, this response was milder in WT males than in ERα^–/– males. The weight of seminal vesicles was similarly decreased in antiandrogen-treated WT and ERα^–/– males (not shown), and circulating testosterone levels were not altered in either strain. Thus, in contrast to ERα^–/– males, in which trabecular bone mass is regulated strictly by androgens, both androgens and estrogens, acting through their respective specific receptors, regulate trabecular bone mass in WT males.

Gonadal hormones also played an osteoprotective role in single-ER-knockout females, but not in ERαβ^–/– females, in which bone response to estradiol treatment was fully blocked. This result contrasts starkly with observations of ovariectomy-induced bone loss and osteoprotective actions of estradiol in αβERKO mice (27), in which shortened transcripts of ERα that lack 64 residues located mainly in the A/B region are still expressed (14, 15). This finding confirms that these shortened ERα’s respond to estradiol in a physiologically relevant manner and indicates that the activation function AF1 of ERα is not required to mediate at least some of the effects of estradiol on trabecular bone. We, however, show here that full deletion of both receptors in ERαβ^–/– mice prevents the response to estradiol, thus excluding any effect of estradiol on bone metabolism in vivo through the AR (9, 10) or through non–receptor-mediated pathways in females, as in males.

Since deletion of both ERs suppressed all effects of estradiol treatment in bone, estradiol responses of single ER knockouts can be interpreted as the physiological effects of the remaining receptor. In ERβ^–/–, where only ERα is expressed, any effect of estradiol treatment is mediated through ERα, and vice versa. In WT mice, both receptors may respond to estradiol, and any effect observed would be the sum of these responses. Activation of ERβ in vivo, however, seems to require a higher concentration of estradiol than does activation of ERα. This is consistent with in vitro studies showing that much higher levels of estradiol are required to activate estrogen-dependent transcription of an estrogen-responsive-element reporter construct via ERβ than via ERα (28). Our results, however, show that the presence of ERβ facilitates the response of ERα to estradiol on bone cells, although, alone, ERβ can maintain low bone turnover but not bone mass, probably because of its effects on bone resorption (13). Indeed, the response of trabecular bone to estradiol treatment was somewhat attenuated in ERβ^–/– mice, suggesting that ERβ does play a role in the WT female response to estradiol. In ERα^–/– females, where ERβ is the only receptor subtype present, a dose-dependent reduction in bone turnover was observed in response to estradiol treatment, indicating that estradiol can inhibit bone turnover via ERβ in vivo, but this did not translate into an increase in bone mass. The better protection of BV/TV in WT than in ERβ^–/– females suggests that ERα and ERβ functions are not competitive but are most likely synergistic when both are expressed. This apparently contradicts our previous results showing decreased bone resorption in ERβ^–/– females (13). However, since ovarian structure is altered, we cannot exclude alterations in other ovarian steroids, leading to decreased bone resorption in these females via the remaining ERα. The results presented in this study indicate, in addition, that ERβ is less efficient than ERα at decreasing bone resorption. Thus, in states of increased bone turnover such as that in ovariectomized ERα^–/– females, administered estradiol acting via ERβ can decrease both bone formation and bone resorption but does the latter less efficiently, leading to an imbalanced situation that favors bone resorption.

Since testosterone levels are also elevated in ERα^–/– females, the increased BV/TV and low bone turnover observed may also relate to a high level of circulating testosterone (although, notably, this level was not as high as that in ERα^–/– males). As in male mice, testosterone treatment of ERα^–/– females protected against both the bone loss and the high bone turnover associated with ovariectomy. This suggests that in females, as in males, testosterone is able to maintain a low level of bone turnover and high BV/TV.

Finally, the question of whether high testosterone or high estradiol levels determine the bone phenotype in intact ERα^–/– females was investigated by treatment with pure antiestrogen or pure antiandrogen to determine which was able to reproduce the effects of ovariectomy. While antiestrogen treatment did not significantly alter BV/TV or bone turnover, antiandrogen treatment partially reproduced the effects of ovariectomy, confirming that testosterone is the major hormone responsible for the high BV/TV in intact ERα^–/– females. However, since testosterone levels in ERα^–/– females are much lower than those in ERα^–/– males (13), and circulating levels of estradiol are very high, we cannot exclude that estradiol participates in the protection of bone via ERβ in these mice.

Thus, in the study of the effects of receptor knockouts in bone, the physiology of the whole animal needs to be considered. This is especially true in the case of the ER knockouts, in which a reproductive phenotype is observed (18) and in which testosterone and estradiol levels are elevated by ERα deletion (13). Taking these changes into consideration, we have determined that ERα is the major receptor regulating bone response to estradiol in both male and female mice. ERβ is able to play a minor protective role in female mice but does not participate in the response to estradiol in males. Furthermore, and most importantly, given the recent hypothesis that estradiol can act via the AR to protect bone, our results clearly demonstrate that the AR, although expressed and functional in ERα^–/– males and females, cannot substitute for ERα to mediate estradiol protection against gonadectomy-induced bone loss in male or female mice.

The authors thank Pierre Chambon, Sonia Dupont, and Andree Krust for generating the knockout mice. We also thank Jennifer Juel for technical assistance in histology and morphometry. This work was supported in part by NIH grant DE-04724 to R. Baron, and in part by grants from Aventis Pharma and ProSkelia Pharmaceuticals to P. Clément-Lacroix and R. Baron.

Conflict of interest: The authors have declared that no conflict of interest exists.

Nonstandard abbreviations used: estradiol receptor (ER); androgen receptor (AR); trabecular bone volume (BV/TV); trabecular bone mineral density (Tb.BMD); peripheral quantitative computerized tomography (pQCT); osteoblast surface as percentage of trabecular bone surface (ObS/BS); osteoclast surface as percentage of trabecular bone surface (OcS/BS); bone formation rate as percentage of trabecular bone surface (BFR/BS).

1. Turner, RT, Riggs, BL, Spelsberg, TC. Skeletal effects of estrogen. Endocr. Rev. 1994. 15:275-300.
   View this article via: PubMed CrossRef Google Scholar
2. Lindsay, R, et al. Long-term prevention of postmenopausal osteoporosis by oestrogen. Evidence for an increased bone mass after delayed onset of oestrogen treatment. Lancet. 1976. 1:1038-1041.
   View this article via: PubMed Google Scholar
3. Oz, OK, et al. Bone has a sexually dimorphic response to aromatase deficiency. J. Bone Miner. Res. 2000. 15:507-514.
   View this article via: PubMed CrossRef Google Scholar
4. Miyaura, C, et al. Sex- and age-related response to aromatase deficiency in bone. Biochem. Biophys. Res. Commun. 2001. 280:1062-1068.
   View this article via: PubMed CrossRef Google Scholar
5. Rochira, V, Balestrieri, A, Faustini-Fustini, M, Carani, C. Role of estrogen on bone in the human male: insights from the natural models of congenital estrogen deficiency. Mol. Cell. Endocrinol. 2001. 178:215-220.
   View this article via: PubMed CrossRef Google Scholar
6. Komm, BS, et al. Estrogen binding, receptor mRNA, and biologic response in osteoblast-like osteosarcoma cells. Science. 1988. 241:81-84.
   View this article via: PubMed CrossRef Google Scholar
7. Onoe, Y, Miyaura, C, Ohta, H, Nozawa, S, Suda, T. Expression of estrogen receptor beta in rat bone. Endocrinology. 1997. 138:4509-4512.
   View this article via: PubMed CrossRef Google Scholar
8. Vidal, O, Kindblom, LG, Ohlsson, C. Expression and localization of estrogen receptor-beta in murine and human bone. J. Bone Miner. Res. 1999. 14:923-929.
   View this article via: PubMed CrossRef Google Scholar
9. Kousteni, S, et al. Nongenotropic, sex-nonspecific signaling through the estrogen or androgen receptors: dissociation from transcriptional activity. Cell. 2001. 104:719-730.
   View this article via: PubMed Google Scholar
10. Kousteni, S, et al. Reversal of bone loss in mice by nongenotropic signaling of sex steroids. Science. 2002. 298:843-846.
    View this article via: PubMed CrossRef Google Scholar
11. Windahl, SH, et al. Female estrogen receptor beta–/– mice are partially protected against age-related trabecular bone loss. J. Bone Miner. Res. 2001. 16:1388-1398.
    View this article via: PubMed CrossRef Google Scholar
12. Vidal, O, et al. Estrogen receptor specificity in the regulation of skeletal growth and maturation in male mice. Proc. Natl. Acad. Sci. U. S. A. 2000. 97:5474-5479.
    View this article via: PubMed CrossRef Google Scholar
13. Sims, NA, et al. Deletion of estrogen receptors reveals a regulatory role for estrogen receptors-beta in bone remodeling in females but not in males. Bone. 2002. 30:18-25.
    View this article via: PubMed CrossRef Google Scholar
14. Couse, JF, et al. Analysis of transcription and estrogen insensitivity in the female mouse after targeted disruption of the estrogen receptor gene. Mol. Endocrinol. 1995. 9:1441-1454.
    View this article via: PubMed CrossRef Google Scholar
15. Denger, S, et al. ERalpha gene expression in human primary osteoblasts: evidence for the expression of two receptor proteins. Mol. Endocrinol. 2001. 15:2064-2077.
    View this article via: PubMed CrossRef Google Scholar
16. Krege, JH, et al. Generation and reproductive phenotypes of mice lacking estrogen receptor beta. Proc. Natl. Acad. Sci. U. S. A. 1998. 95:15677-15682.
    View this article via: PubMed CrossRef Google Scholar
17. Pendaries, C, et al. The AF-1 activation-function of ERalpha may be dispensable to mediate the effect of estradiol on endothelial NO production in mice. Proc. Natl. Acad. Sci. U. S. A. 2002. 99:2205-2210.
    View this article via: PubMed CrossRef Google Scholar
18. Dupont, S, et al. Effect of single and compound knockouts of estrogen receptors alpha (ERalpha) and beta (ERbeta) on mouse reproductive phenotypes. Development. 2000. 127:4277-4291.
    View this article via: PubMed Google Scholar
19. Van de Velde, P, et al. RU 58,668: further in vitro and in vivo pharmacological data related to its antitumoral activity. J. Steroid Biochem. Mol. Biol. 1996. 59:449-457.
    View this article via: PubMed CrossRef Google Scholar
20. Battmann, T, et al. Pharmacological profile of RU 58642, a potent systemic antiandrogen for the treatment of androgen-dependent disorders. J. Steroid Biochem. Mol. Biol. 1998. 64:103-111.
    View this article via: PubMed CrossRef Google Scholar
21. Teutsch, G, et al. Non-steroidal antiandrogens: synthesis and biological profile of high-affinity ligands for the androgen receptor. J. Steroid Biochem. Mol. Biol. 1994. 48:111-119.
    View this article via: PubMed CrossRef Google Scholar
22. Sims, NA, et al. Bone homeostasis in growth hormone receptor–null mice is restored by IGF-I but independent of Stat5. J. Clin. Invest. 2000. 106:1095-1103.
    View this article via: JCI PubMed CrossRef Google Scholar
23. Hofbauer, LC, Hicok, KC, Khosla, S. Effects of gonadal and adrenal androgens in a novel androgen-responsive human osteoblastic cell line. J. Cell. Biochem. 1998. 71:96-108.
    View this article via: PubMed CrossRef Google Scholar
24. Katznelson, L, et al. Increase in bone density and lean body mass during testosterone administration in men with acquired hypogonadism. J. Clin. Endocrinol. Metab. 1996. 81:4358-4365.
    View this article via: PubMed CrossRef Google Scholar
25. Anderson, FH, Francis, RM, Peaston, RT, Wastell, HJ. Androgen supplementation in eugonadal men with osteoporosis: effects of six months’ treatment on markers of bone formation and resorption. J. Bone Miner. Res. 1997. 12:472-478.
    View this article via: PubMed CrossRef Google Scholar
26. Vandenput, L, et al. Testosterone prevents orchidectomy-induced bone loss in estrogen receptor-alpha knockout mice. Biochem. Biophys. Res. Commun. 2001. 285:70-76.
    View this article via: PubMed CrossRef Google Scholar
27. Gentile, MA, Zhang, H, Harada, S, Rodan, GA, Kimmel, DB. Bone response to estrogen replacement in OVX double estrogen receptor-(α and β) knockout mice. J. Bone Miner. Res. 2001. 16(Suppl. 1):S146. (Abstr.)
28. Hall, JM, McDonnell, DP. The estrogen receptor beta-isoform (ERbeta) of the human estrogen receptor modulates ERalpha transcriptional activity and is a key regulator of the cellular response to estrogens and antiestrogens. Endocrinology. 1999. 140:5566-5578.
    View this article via: PubMed CrossRef Google Scholar