The authors state that they have no conflicts of interest.
The relative importance of AR and ER activation has been studied in pubertal male AR knockout and WT mice after orchidectomy and androgen replacement therapy, either with or without an aromatase inhibitor. AR activation dominates normal trabecular bone development and cortical bone modeling in male mice. Moreover, optimal periosteal bone expansion is only observed in the presence of both AR and ER activation.
Introduction: Androgen receptor (AR)–mediated androgen action has traditionally been considered a key determinant of male skeletal growth. Increasing evidence, however, suggests that estrogens are also essential for normal male bone growth. Therefore, the relative importance of AR-mediated and estrogen receptor (ER)–mediated androgen action after aromatization remains to be clarified.
Materials and Methods: Trabecular and cortical bone was studied in intact or orchidectomized pubertal AR knockout (ARKO) and male wildtype (WT) mice, with or without replacement therapy (3–8 weeks of age). Nonaromatizable (dihydrotestosterone [DHT]) and aromatizable (testosterone [T]) androgens and T plus an aromatase inhibitor (anastrazole) were administered to orchidectomized ARKO and WT mice. Trabecular and cortical bone modeling were evaluated by static and dynamic histomorphometry, respectively.
Results: AR inactivation or orchidectomy induced a similar degree of trabecular bone loss (−68% and −71%, respectively). Both DHT and T prevented orchidectomy-induced bone loss in WT mice but not in ARKO mice. Administration of an aromatase inhibitor did not affect T action on trabecular bone. AR inactivation and orchidectomy had similar negative effects on cortical thickness (−13% and −8%, respectively) and periosteal bone formation (−50% and −26%, respectively). In orchidectomized WT mice, both DHT and T were found to stimulate periosteal bone formation and, as a result, to increase cortical thickness. In contrast, the periosteum of ARKO mice remained unresponsive to either DHT or T. Interestingly, administration of an aromatase inhibitor partly reduced T action on periosteal bone formation in orchidectomized WT mice (−34% versus orchidectomized WT mice on T), but not in ARKO mice. This effect was associated with a significant decrease in serum IGF-I (−21% versus orchidectomized WT mice on T).
Conclusions: These findings suggest a major role for AR activation in normal development of trabecular bone and periosteal bone growth in male mice. Moreover, optimal stimulation of periosteal growth is only obtained in the presence of both AR and ER activation.
As a result of more pronounced periosteal bone formation, pubertal bone growth in men is associated with more radial bone expansion than in women.(1) This additional periosteal growth in males ultimately defines the cross-sectional area of the bone and thereby confers greater bone strength.(2) Androgen-mediated activation of the androgen receptor (AR) is considered a key determinant of this sex-specific pattern of periosteal growth.(3) In line with this concept, androgen-resistant rodents show reduced bone size because of a reduction in cortical thickness and periosteal perimeter.(4–6) A reduction in cortical bone area is also observed in the context of sex steroid deficiency, as induced by orchidectomy, in growing male rats and mice.(7) Moreover, trabecular bone is markedly affected by sex steroid deficiency or androgen resistance.(5,6)
Androgens can stimulate the skeleton not only through direct activation of the AR, but also indirectly, after aromatization into estrogens and subsequent activation of one or both estrogen receptors (ERα/β). In fact, increasing evidence suggests that estrogens may be critically involved in male skeletal growth. Men with natural mutations in the estrogen receptor α (ERα)(8) or the aromatase gene(9–13) show low BMD and failure to establish peak bone mass. Likewise, growing male rats treated with an aromatase inhibitor or male mice with inactivation of either the ERα or the aromatase enzyme have reduced bone size, as reflected by a decline in cortical area and periosteal perimeter.(14–16)
Transgenic mouse models offer the opportunity to study the relative importance of sex steroid receptors during skeletal growth and to determine the effects of orchidectomy and sex steroid replacement on trabecular and cortical bone compartments. In orchidectomized androgen-resistant mice, trabecular and cortical bone are not responsive to dihydrotestosterone (DHT) treatment.(6) Some(6) but not all(5) available data suggest only partial responsiveness to testosterone (T), supporting a role for aromatization of androgens into estrogens followed by ER activation. Intact male ERα knockout mice (ERαKO), on the other hand, show a cortical bone deficit but an increase in trabecular bone volume as a result of enhanced T levels acting through the AR.(17) In this animal model, orchidectomy-induced loss in trabecular and cortical BMD and cortical thickness is prevented by treatment with T,(17,18) consistent with a critical role for androgen-mediated AR activation. Therefore, it remains uncertain to what extent androgen action on male trabecular and cortical bone modeling depends on activation of the AR and/or aromatization of androgens into estrogens. To further address this question, we studied trabecular and cortical bone phenotype and the response to orchidectomy and replacement therapy in growing AR knockout (ARKO) mice and corresponding wildtype (WT) mice. To define the importance of AR activation on trabecular and cortical bone in growing male mice, we studied the effects of AR inactivation in comparison with orchidectomy. Additionally, to differentiate between direct (AR-mediated) and indirect (ER-mediated) effects, aromatizable and nonaromatizable androgens, as well as an aromatase inhibitor, were administered to orchidectomized ARKO and WT mice.
MATERIALS AND METHODS
ARKO mice were generated using the Cre/loxP technology, as previously described.(19) Their genetic background was C57Bl6/N. Genotyping was performed using PCR amplification.(19) Mice lived in conventional conditions: 12-h light/dark cycle, standard diet (1% calcium, 0.76% phosphate), and water ad libitum.
At the start of puberty (23 days of age), male WT and ARKO littermates were randomly divided in groups. Mice were either sham-operated (sham) or orchidectomized (orx) using sodium pentobarbital anesthesia. Orx mice were treated during an experimental period of 5 weeks with vehicle (V), DHT, T, or T plus aromatase inhibitor (anastrazole [Arimidex], 10 mg/kg/day). DHT (Fluka) and T (Serva) were administered using subcutaneous silastic implants (Silclear Tubing, Degania Silicone, Jordan Valley, Israel) in the cervical region. Vehicle animals received empty implants. Anastrazole is a potent, nonsteroidal, highly selective aromatase inhibitor with no intrinsic hormonal activity.(20) The substance was obtained from Astra Zeneca after formal approval of the protocol and was administered orally. Eight animals were included in each group. Body weight was measured weekly. All mice were injected intraperitoneally with the fluorochrome calcein at a 5-day interval and were killed 1 day after the second injection. Before the first calcein injection, mice were put in metabolic cages to collect urine for measurement of collagen cross-links. At death, serum was collected, stored at −20°C, and used for osteocalcin and IGF-I measurement. Femur and tibias were dissected and used to perform histomorphometric analysis. Efficacy of orx and DHT and T replacement in WT mice was verified by measurement of seminal vesicles wet weight immediately after death. Seminal vesicle weight expressed per gram of body weight was 6.2 ± 0.3 mg/g in the DHT-treated group and 8.3 ± 0.2 mg/g in the T-treated group compared with 5.7 ± 0.2 mg/g in sham-operated mice and 0.10 ± 0.04 in orx mice (p < 0.001). The greater androgenic potency of T may be explained by the constant and continuous release from the implants throughout the experimental period. The physiological significance of anastrazole was assessed in 8-week-old female mice that were given anastrazole (10 mg/kg/day, orally) for 2 weeks. The weight of the uterus per gram of body weight in Arimidex-treated females was 2.5 ± 0.1 versus 6.0 ± 0.2 mg/g in age-matched control females (p < 0.001).
The ethical committee of the Katholieke Universiteit Leuven approved all experimental procedures.
One femur and tibia were cleaned from surrounding tissue, immersed in Burckhardt's fixative (24 h, 4°C), kept in 100% ethanol, and embedded in methylmethacrylate.
Longitudinal sections of the undecalcified tibia were cut at 4 μm thickness using a rotation microtome (RM 2155 Autocut; Leica, Heidelberg, Germany) with a tungsten carbide blade (Leica, Nussloch, Germany). Sections were stained by a modified Goldner technique and subjected to static histomorphometry. Measurements were performed in the secondary spongiosa of at least three Goldner-stained sections, as previously described.(5,18) In each section, three consecutive fields were measured along the vertical axis of the central metaphysis, starting at a regular distance of the growth plate. Trabecular width and trabecular number were calculated according to the parallel plate model developed by Parfitt et al.(21)
Cross-sections of the undecalcified femur perpendicularly to the long axis were prepared at 200 μm thickness in the mid-diaphyseal region using the contact-point precision band saw (Exakt, Norderstedt, Germany). Sections were ground to a final thickness of 25 μm using a grinding system (Exakt). Sections were left unstained and subjected to dynamic histomorphometry. Three sections in the mid-diaphyseal region were measured by fluorescence microscopy, and the bone formation rate (BFR/B.Pm., μm2/μm/day) was assessed at both the endocortical and periosteal bone surfaces. The BFR was obtained by the product of mineral apposition rate (MAR, μm/day) and mineralizing perimeter per bone perimeter (Min.Pm./B.Pm., %). The mineralizing perimeter was calculated as follows: Min.Pm. = [dL + (sL/2)]/B.Pm., where dL represents the length of the double labels and sL is the length of single labels along the entire endocortical or periosteal bone surfaces. The MAR (μm/day) was calculated as the mean width of double labels, divided by interlabel time (5 days). The mineralizing perimeter is a measure for osteoblast number and MAR for the osteoblast activity. Also, the cross-sectional area (CSA), cortical area, cortical thickness, and endocortical and periosteal perimeters were measured on cortical cross-sections. All measurements were performed with a Kontron Image Analyzing computer (KS400 3.00; Kontron Bildanalyze, Munich, Germany) and a Zeiss microscope with a drawing attachment. Specific software was developed in collaboration with the manufacturer. Histomorphometric parameters are reported according to the recommended American Society for Bone and Mineral Research nomenclature.(21)
Trabecular and cortical volumetric BMD was assessed ex vivo by pQCT using the Stratec XCT Research M+ densitometer (Norland Medical Systems, Fort Atkinson, WI, USA). Slices of 0.2 mm thickness were scanned using a voxel size of 0.070 mm. One scan was taken 2 mm from the distal end of the femur, using contmode 1, peelmode 20, and a density threshold of 280 mg/cm3. The trabecular bone region was defined by setting an inner threshold corresponding to 30% of the total CSA. These metaphyseal scans were performed to measure trabecular volumetric density. A second scan was taken 7 mm from the distal end of the femur (an area containing only cortical bone) using separation mode 1 and a density threshold of 710 mg/cm3. These mid-diaphyseal scans were performed to determine cortical volumetric density, cortical thickness, and endocortical and periosteal perimeters.
Whole body DXA
Body composition was analyzed in vivo by DXA (PIXImus densitometer; Lunar Corp., Madison, WI, USA) using ultra-high resolution (0.18 × 0.18 pixels, resolution of 1.6 line pairs/mm) and software version 1.45. DXA was performed at the end of the experimental period.
Serum osteocalcin was measured by an in-house radioimmunoassay (RIA).(22) After acid-ethanol extraction, serum IGF-I concentrations were measured by an in-house RIA(23) in the presence of an excess of IGF-II (25 ng/tube). Urinary collagen cross-links (deoxypyridinoline [DPD]) were measured by high-performance liquid chromatography (HPLC) with fluorescence detection after acid hydrolysis.(24) The concentration of DPD was corrected for creatinine excretion, which was measured colorimetrically.
Statistical analysis of data was performed using NCSS software (Kaysville, UT, USA). One-way ANOVA, followed by Fisher's least significant difference multiple comparison test, and t-tests were performed to assess significance of difference between groups of the same genotype and between respective WT and ARKO groups. Data are represented as mean ± SE. p < 0.05 was accepted as significant.
Effects of AR inactivation versus orchidectomy on trabecular bone
AR inactivation resulted in a pronounced trabecular bone phenotype. Sham ARKO mice showed a significant reduction in trabecular BMD (−61%, p < 0.001; Fig. 1A) and trabecular bone volume (B.Ar./T.Ar., %) (−68%, p < 0.05; Fig. 1B).The latter was the result of a significant decrease in trabecular number (−63%, p < 0.05; Fig. 1C) but not in trabecular width (Fig. 1D). Orchidectomy (orx) or androgen deficiency in WT mice induced a trabecular bone phenotype similar to the phenotype observed with AR inactivation; trabecular BMD (−69%, p < 0.001), trabecular bone volume (−71%, p < 0.05), and number (−58%, p < 0.05) were all significantly decreased after orx (Figs. 1A–1C). The reduction in trabecular bone volume was associated with an increase in urinary DPD in sham ARKO mice (63 ± 8 nM/mM creatinine) and orx WT mice (52 ± 3 nM/mM) compared with sham WT mice (28 ± 3 nM/mM; p < 0.001). Longitudinal growth, as assessed by femoral length, was not changed by AR inactivation (15.2 ± 0.2 mm in sham ARKO mice) or orx (15.0 ± 0.1 in orx WT mice) compared with sham WT mice (15.2 ± 0.2 mm).
Effects of AR inactivation versus orx on cortical bone
Periosteal growth of the midfemoral shaft was assessed by dynamic histomorphometric analysis. Not only AR inactivation but also orx significantly decreased cross-sectional area (−13% and −9%, respectively, p < 0.05; Fig. 2A), cortical area (−18% and −12%, respectively, p < 0.05; Fig. 2B), cortical thickness (−13% and −8%, respectively, p < 0.05; Fig. 2C), and periosteal perimeter (−6% and −5%, respectively, p < 0.05; Fig. 2D). This resulted from a significant decrease in periosteal bone formation rate (Ps.BFR/B.Pm.) in ARKO mice and orx WT mice (−50% and −26%, respectively, p < 0.05; Fig. 3A), which was associated with a reduced periosteal osteoblast number in both sham ARKO mice and orx WT mice compared with sham WT mice, as indicated by a significantly reduced periosteal mineralizing perimeter (Ps.Min.Pm./B.Pm.) (−50% and −39%, respectively, p < 0.01; Fig. 3B). The periosteal MAR, a measure for osteoblast activity, was not affected by either AR inactivation or orx (data not shown).
Sham ARKO mice and orx WT mice also gained significantly less body weight during puberty (from 3 to 8 weeks of age) compared with sham WT mice (p < 0.001; Fig. 4). In addition, lean body mass, but not fat mass, was significantly decreased in sham ARKO and orx WT mice at the end of puberty (−12% and −15%, respectively, p < 0.001; Table 1).
Table Table 1.. Body Composition in Sham-Operated WT and ARKO Mice and Orchidectomized WT Mice
Effects of DHT, T, and aromatase inhibitor on trabecular and cortical bone modeling
Both DHT and T prevented the orx-induced loss of trabecular BMD in WT mice (p < 0.001 versus orx +V WT; Fig. 5). Administration of the aromatase inhibitor (AI) did not affect T action on trabecular BMD in WT mice. High serum osteocalcin and urinary DPD levels, reflecting the orx-induced high bone turnover state, were suppressed after treatment with DHT, T, and T + AI in WT mice (Table 2). In contrast to the effects observed in WT mice, the low trabecular BMD in ARKO mice was not affected by nonaromatizable (DHT) and aromatizable (T) androgens or T + AI. The high rate of bone turnover in ARKO was present in each condition (Table 2).
Table Table 2.. Bone Turnover Markers in WT and ARKO Mice
Both DHT and T significantly increased the periosteal mineralizing perimeter (Min.Pm) in orx WT mice (+101% and +141% versus orx + V WT, respectively, p < 0.001; Fig. 6A), but not the periosteal MAR (p = 0.56; Fig. 6B). As a result, the periosteal BFR was significantly enhanced by both DHT and T in orx WT mice (+100% and +163% versus orx + V WT, respectively, p < 0.001; Fig. 6C), along with a significant increase in cortical area (+18% and +38% versus orx + V WT, respectively, p < 0.001) and cortical thickness (+17% and +27% versus orx + V WT, respectively, p < 0.001; Table 3). Also, the periosteal perimeter (+10% versus orx + V WT, respectively, p < 0.001) and CSA (+22% versus orx + V WT, respectively, p < 0.001) were significantly increased by T (Table 3). Administration of T + AI partially reduced the effects of T on the periosteal Min.Pm. (−25% versus orx + T WT, p < 0.001) and periosteal BFR in orx WT mice (−34% versus orx + T WT, p < 0.001; Figs. 6A and 6C), resulting in a reduced cortical area (−17% versus orx + T WT, p < 0.001), thinner cortex (−7% vs. orx + T WT, p < 0.001), and smaller periosteal perimeter (−9% versus orx + T WT, p < 0.001; Table 3). This effect of T + AI on cortical bone of orx WT mice was associated with a significant decline in serum IGF-I levels (−21% versus orx + T WT, p = 0.01; Table 4). In contrast to the effects observed in WT mice, the periosteal bone surface of ARKO mice remained unresponsive to any treatment (Figs. 6A–6C; Table 3), and no significant changes in serum IGF-I could be detected (Table 4).
Table Table 3.. Parameters of Cortical Bone Width in WT and ARKO Mice
Table Table 4.. Serum IGF-I in WT and ARKO Mice
In this study in growing male mice, androgen unresponsiveness, in the context of AR inactivation and sex steroid deficiency induced by orx, induced a similar degree of trabecular bone loss. Both conditions resulted in a significant reduction of trabecular bone volume with fewer and widespread but normal-sized trabeculae. Aromatizable and nonaromatizable androgens prevented orx-induced trabecular bone loss to a similar degree in WT mice but not in ARKO mice. Moreover, inhibition of aromatase activity did not alter T action on trabecular bone. Taken together, these findings support the concept that AR activation is essential for normal development of trabecular bone in male mice.
It is generally accepted that both AR and ERα activation, independently and to a similar degree, affect trabecular bone mass in male mice. This so-called “dual mode of action” of androgens on trabecular bone has been derived from studies in male ERαKO and ARKO mice.(6,25) In ERαKO mice, the increase in trabecular bone volume is normalized to WT levels after administration of an antiandrogen, indicating that their enhanced trabecular bone mass is the result of the higher T levels acting directly through the AR.(17) In ERαKO mice, orx reduces trabecular bone mass to the same level as in WT controls, and this orx-induced trabecular bone loss is fully prevented by T,(17,18) again showing the important role of AR-mediated androgen action for normal trabecular bone mass in these animals. In orx WT mice, on the other hand, 17β-estradiol (E2) also increases trabecular bone volume and even seems to have an osteoanabolic effect.(18,25) However, the doses used in these experiments are pharmacologic and not representative of the significantly lower physiological levels in male mice (∼5 pg/ml). Further evidence for a possible role for aromatization of androgens into estrogens followed by ER activation in male mice comes from data obtained by Kawano et al.,(6) who showed that T partly prevents orx-induced bone loss in ARKO mice. However, this study evaluated areal femoral BMD only and did not allow to distinguish between trabecular and cortical bone; in addition, no aromatase inhibitor was used, and the experiment was not designed to determine the role of aromatization in T action on trabecular bone. Finally, in androgen-resistant testicular feminized male (Tfm) mice, administration of pharmacological doses of T did not increase serum E2, trabecular BMD, or trabecular bone volume.(5) This finding is consistent with the very limited expression of the aromatase enzyme in rodents; the major sites of expression are the gonads and the brain and thus peripheral aromatization is very poor.(26) With respect to trabecular bone, one might therefore hypothesize that the intrinsic response to androgens and estrogens is regulated differently in rodents and humans. Therefore, although both AR and ERα activation may be able to increase trabecular bone mass, our study shows that androgen acting directly through the AR is the major determinant of normal trabecular bone development, at least in physiological conditions in male mice. Aromatization and subsequent ER activation, on the other hand, seem to play a less important role in trabecular bone development (Fig. 7). Alternatively, one may also hypothesize that androgens affect trabecular bone indirectly. In this respect, it has been shown that androgens are able to upregulate the expression of the IGF-I receptor and, as a consequence, are able to enhance IGF-I action.(27,28) AR inactivation would therefore be associated with a decreased responsiveness to IGF-I. The concept that local actions of IGF-I may be important for normal trabecular bone development is supported by the finding that loss of IGF-I receptor signaling or IGF-I overexpression specifically in osteoblasts is associated with marked alterations in trabecular bone volume of young growing mice.(29,30)
In this study, we found androgen resistance and sex steroid deficiency to have similar negative effects on cortical area and thickness, as a result of a reduced bone formation at the periosteum. In WT controls, but not in ARKO mice, T and DHT stimulated periosteal bone formation with an increase in cortical area. In ARKO mice, periosteal bone remained unresponsive to either aromatizable or nonaromatizable androgens, again supporting the concept that AR activation also plays a major role in cortical bone modeling in male mice (Fig. 7).
Overall, androgens are indeed considered a key determinant of male cortical bone growth.(7) This assumption is based on observations in orx male rodents showing a decrease in periosteal circumference.(31) Additionally, observations of a female-like bone size in rodents and humans with inactive ARs have provided further support for a key role of AR activation in male radial bone expansion.(4–6,32) Interestingly, androgen resistance or deficiency also result in reduced body weight and lean body mass (a surrogate marker for muscle mass). Because muscle mass is an important determinant of mechanical loading,(33) which, in turn, stimulates bone formation, one might speculate that some of the effects of androgens and AR activation are indirect through stimulation of muscle mass and subsequent increased mechanical loading.
However, aromatization of androgens into estrogens and subsequent ER activation may be important as well in the process of male cortical bone modeling. Our finding of a blunting of the effect of T on periosteal bone growth by anastrazole, an AI, supports this possibility. Moreover, this effect was only observed in WT mice and was paralleled by a significant decline in serum IGF-I levels. In previous studies in male rats, administration of an AI was also associated with lower serum IGF-I levels and reduced CSA.(14,34) A similar decrease in serum IGF-I levels and cortical thickness has been observed in male ERαKO mice.(15) Likewise, administration of anastrozole to late pubertal boys decreased serum E2 as well as serum IGF-I(35) and, more recently, treatment with E2 was found to enhance periosteal bone expansion in an aromatase-deficient adolescent boy.(11) All these studies lend support to the hypothesis that estrogen-related changes in serum IGF-I may act to mediate estrogen action on the periosteal bone surface. It has been well established that growth hormone (GH) and IGF-I are major determinants of postnatal growth.(36,37) Mice with a disrupted GH receptor have dramatically decreased serum IGF-I levels and severely retarded skeletal growth rates.(38,39) In GHRKO mice, we previously showed that the periosteal bone surface is extremely sensitive to changes in circulating IGF-I levels, with periosteal bone growth being fully rescued after upregulation of serum IGF-I in GHRKO mice treated with E2, but not when exposed to T or DHT.(40) These findings suggest that the effect of aromatization of androgens into estrogens and subsequent ER activation is indirectly mediated through alterations of the GH/IGF-I axis (Fig. 7). An intriguing observation in this study was the lack of effect of the aromatase inhibitor on cortical bone modeling in ARKO mice. This observation remains unexplained. One might hypothesize that the severe cortical phenotype in ARKO mice is not further affected by inhibition of aromatase activity. Also, no changes in serum IGF-I were observed in ARKO mice that could have affected cortical bone growth.
In conclusion, this study in ARKO mice supports the concept that AR activation is a major determinant of trabecular bone development in physiological conditions, whereas aromatization and subsequent ER activation seem less important. Stimulation of bone formation at the periosteal bone surface and increases in bone size in male mice predominantly depend on activation of the AR as well. However, to obtain a maximal stimulation of periosteal bone growth in male mice, both AR and ER activation are required.
This study was supported by Katolieke Universiteit Leuven Grant OT/01/39 and Fund for Scientific Research-Flanders, Belgium Grants G.0417.03, G.0458.05, and G.0171.03. DV and SB are Senior Clinical Investigators of the Fund for Scientific Research-Flanders, Belgium. SB is a holder of the Leuven University Chair in Metabolic Bone Diseases, supported by Roche and GSK.