The human androgen receptor (hAR) is a member of the nuclear hormone receptor (NHR) superfamily. Because the activated hAR positively stimulates the growth of prostate cancer (PC), its inactivation remains the central goal of current PC therapies. Androgen antagonists (antihormones) bind to and inactivate the hAR. One underlying mechanism that we have identified is that the hAR-mediated transactivation is inhibited through recruitment of co-repressors that bind to hAR in the presence of androgen antagonists (Dotzlaw et al., 2002). Co-repressors are cofactors required for target gene repression originally identified for those transcription factors having a silencing function (Burke & Baniahmad, 2000). Interestingly, the recruitment of co-repressors to the hAR is dependent on the response element of the target genes, suggesting a structural influence of the DNA that might impose conformational changes on the hAR, and also implies that not all hAR target genes will be repressed to a similar extent by treatment with androgen antagonists. The known androgen antagonists act either as the complete antagonists hydroxyflutamide and casodex (Cas, bicalutamide) or as the partial antagonists cyproterone acetate (CPA) on hAR. Testing for binding of co-repressors in the presence of these different types of androgen antagonists revealed that the binding of co-repressors to hAR on a given response element is dependent on the type of antagonist used. For example, the SMRT co-repressor is recruited to the prostate specific antigen promoter but not to mouse mammary tumour virus (MMTV) promoter in the presence of CAS, whereas SMRT is recruited to the MMTV promoter in the presence of CPA (Dotzlaw et al., 2002; Shang et al., 2002; Kang et al., 2004). This indicates that in addition to the response element specificity, the type of ligand induces a specific AR-structure that dictates the binding specificity of co-repressors to hAR as well. Unlike to other members of the NHR, it is the amino(N-) terminus of hAR that harbours the major transactivation function. The previously known interaction domain of co-repressors with other NHR was identified to be the receptor carboxy(C-) terminus that includes the ligand binding domain. In contrast to other nuclear hormone receptors, we found that co-repressors such as SMRT, NCoR and Alien interact with the hAR N-terminus (Dotzlaw et al., 2002, 2003; Hodgson et al., 2005, and unpublished). Thus, both the co-repressors and the co-activators bind to the same AR-domain. At the functional level, the efficacy of inhibition of hAR by partial antagonists, such as CPA, was shown to be dependent on the level of co-repressors in the cell; the higher the level of co-repressors, the more potent the CPA-mediated inhibition of hAR (Dotzlaw et al., 2002). This strongly suggests that co-repressors mediate the inhibition of hAR by androgen antagonists. Insights into the interplay of co-repressors with co-activators have been gained by comparing pure agonists with the partial agonistic activity of CPA, which when employed in the absence of any agonist, is able to induce AR transactivation. However, the extent of AR-mediated transactivation is weaker than that induced by pure agonists. This suggests that in addition to co-repressors, co-activators bind AR in the presence of CPA as well. Surprisingly, we found that co-repressors also bind hAR in the presence of pure androgen agonists, as assayed by mammalian one hybrid and chromatin immuno-precipitation techniques. Furthermore, co-repressor binding in the presence of agonist did not lead to a significant repression of hAR (Dotzlaw et al., 2002 and unpublished). This suggests that the interplay of co-activators and co-repressors is also dependent on the type of ligand. But two central questions arose that yet remain unanswered. One is how the binding of co-repressors to agonist-bound hAR does not result in significant gene repression. The other question is how the interplay between co-repressors and co-activators act together to regulate the activity of hAR. Interestingly, using the partial antagonist CPA, a competition between co-activators and co-repressors binding was observed. Consistently with this, increasing the
Read full abstract