LBD Dimerization of the Androgen Receptor but Not N/C Interaction Is Crucial for Normal Male Development in Mice

The androgen receptor (AR) is a nuclear receptor with a main role in the development and maintenance of the male phenotype. To execute its role as transcription factor, the AR forms homodimers. Three dimerization modes have been described for the AR: one via the DNA binding domain, a second via the ligand binding domain (LBD) and a third via interactions between the LBD and the aminoterminus of the AR (N/C). Based exclusively on in vitro data, all three dimerization modes seem to contribute to full AR activity, albeit to a different extent. The in vivo role of the dimerization modes, however, remains unknown. To study the physiological relevance, we generated two mouse models using a CRISPR/Cas9 approach, in which either the N/C interaction (AR(NoC)) or LBD dimerization (AR(Lmon)) was disrupted. Surprisingly, the male AR(NoC) mice have a normal phenotype, indicating that the N/C interaction is not crucial for male development. In contrast, AR(Lmon) males have an external female phenotype with cryptorchid testes and high levels of circulating testosterone (T), androstenedione and luteinizing hormone (LH) (6-, 13- and 45-fold higher, respectively). They have no prostate, seminal vesicles or epididymis, illustrating the importance of LBD dimerization during male development. Phenotyping the AR(Lmon) model furthermore provided evidence of a crucial role for the AR in bone homeostasis as well as steroidogenesis. The AR(Lmon) males display a severe bone phenotype, similar to that of complete AR knockout (ARKO) mice. The bone phenotype of ARKO was postulated to be mainly due to lower estrogen levels. However, in contrast to ARKO mice, AR(Lmon) mice have high circulating levels of T, which can still function as prohormone for estradiol and support bone function via the ERα. Immunohistological analysis of AR(Lmon) testes showed hyperplasia of the Leydig cells and residual spermatogenesis. Analysis of the steroidogenic pathway revealed that while the expression of most genes is increased, the expression of Hsd17b3, encoding the enzyme responsible for conversion of androstenedione into T, is low in AR(Lmon) testis. Reporter assays confirmed that the promotor of this gene is indeed upregulated by the AR itself. In conclusion, our work uncovers the physiological role of the N/C interaction and LBD dimerization of the AR. It furthermore demonstrates a direct role for AR in male bone development independent of T aromatization into estrogens. Finally, we show that the AR controls the final step in the synthesis of its own ligand. In contrast to the in vitro data, N/C interaction is not crucial for male development in vivo. The AR(Lmon) model illustrates that LBD dimerization could be an excellent new therapeutic target for inhibiting AR activity for example in advanced prostate cancer that has developed resistance to the current AR-targeting therapies.