The androgen receptor as a direct target in male musculoskeletal maintenance

The major side effects of androgen deprivation are osteoporosis and sarcopenia that both may increase fracture risk. Androgens prevent bone and muscle loss in hypogonadal as well as orchidectomized rodents. However, the potential of androgens to restore hypogonadal osteoporosis has not yet been fully explored. The first aim of the introductory study was therefore to evaluate to what extent androgens (both testosterone (T), the main circulating male androgen, and a synthetic androgen MENT) were able to restore bone mass as well as lean mass in a preclinical animal model of androgen deprivation. Two months following orchidectomy the aged male rats were characterized by cortical and trabecular osteopenia, which was secondary to increased intramedullary bone resorption combined with decreased periosteal bone formation. Both T and MENT not only prevented further bone loss but also stimulated periosteal bone formation in this hypogonadal rat model. Moreover, both androgens also increased lean body mass, which could indirectly also stimulate bone formation. Our further aim was to explore to what extent the AR in bone cells of male mice was involved in the antiresorptive as well as in the osteoanabolic actions of androgens on respectively trabecular and cortical bone compartments. Global disruption of the androgen receptor in male mice was characterized by severe and early trabecular and cortical osteopenia again due to concomitant decreased periosteal bone formation as well as increased intramedullary bone resorption. Although the AR is expressed in different bone cells, the role of AR in these bone cells with respect to the suppressive effects on bone resorption as well as the stimulatory action on bone formation was not established. Using Cre-loxP technology, the AR was selectively and successfully deleted in osteocytes, which are the most abundant bone cells with the highest AR expression. For this purpose DMP1 (Dentin Matrix Acidic Phosphoprotein 1) Cre targeted expression was used which is expressed in terminally differentiated osteocytes. Bone phenotype of the ocy-ARKO mice -in analogy with other studies in late and early osteoblast differentiation stages- was only characterized by a late and moderate trabecular osteopenia without affecting cortical bone formation and/or excessive bone resorption. In order to further explore whether AR in osteoclasts could directly explain its suppressive effect, a selective ocl-ARKO was created with similar Cre-loxP technology using the cathepsin K Cre, which is expressed in mature osteoclasts. AR expression osteoclasts was however extremely low. The ocl-ARKO showed no bone phenotype and responded normally to androgen deprivation as well as replacement. Therefore, AR in mature osteoclasts did not appear to suppress bone resorption. Nevertheless, osteoclasts were increased in global ARKO mice. Moreover, bone marrow cell cultures of the global ARKO mice -in comparison with WT mice- also showed increased osteoclast generation as well as activity. To further explore the potential suppressive effect of AR in osteoblasts and in bone marrow cells on osteoclast generation and activity co-culture experiments were performed. Co-cultures of osteoblasts with bone marrow cells from respectively WT and ARKO showed increased osteoclastogenesis as well as osteoclast activity. These data indicated that bone marrow cells, and not osteoblasts, were essential for osteoclast generation and activity. Bone marrow however consists of hematopoietic cells, which can differentiate towards osteoclasts, and of stromal cells, which can differentiate into osteoblasts. AR mRNA expression was high in these stromal cells. Further gene expression analyses revealed no differences in RANKL/OPG ratio, which is the main determinant of bone resorption, in osteoblasts or stromal cells between ARKO and WT. Therefore the enhanced bone resorption observed in the global ARKO could not be due to higher RANKL expression and therefore more stimulation of osteoclastogenesis. Interestingly ARKO stromal cells exhibited higher expression of Runx2 mRNA compared to WT. Therefore, the AR might prevent early recruitment of bone marrow cells into the bone-remodeling cascade with minimal effect at later stages of osteoblast or osteoclast differentiation. Our in vitro observations might also explain the discordance of the modest bone phenotype characterized by late trabecular osteopenia in current selective bone ARKO models versus the severe early cortical as well as trabecular osteopenia in global ARKO. The final aim of this study was to investigate to what extent androgens increased the sensitivity to mechanical loading and hereby indirectly stimulated bone formation. As shown earlier, androgens indeed stimulated also muscle and hereby loading which could explain their stimulatory action on bone formation. In a mice tibia loading model the osteoanabolic response following mechanical loading was evaluated in mice who were androgen deprivated and mice who received androgen replacement (aromatizable testosterone or non-aromatizable dihydrotestosterone) immediately after orchidectomy. This osteoanabolic response was enhanced by androgen deprivation and decreased by androgen replacement hereby refuting the hypothesis that androgens increased mechanical sensitivity. On the contrary, mechanical sensitivity was reduced by androgens. Interestingly, such action was associated with changes in bone sclerostin expression, which is the major regulator of bone formation and which may open further therapeutic perspectives. Moreover, both androgens and loading, however independently also decreased bone resorption as well as RANKL-OPG expression. In conclusion, the osteoanabolic action of androgens could not be explained by AR in bone cells nor by enhancement of mechanosensitivity. The antiresorptive action was also not fully explained by its presence in bone cells but may be related to its action on bone marrow stromal cells, which orchestrate bone remodeling.status: publishe