An appropriate model to study the complex process of prostate tumorigenesis needs to reflect at least some aspects of the human disease presented in the clinic. The progression of prostate cancer includes the development of high-grade prostatic intraepithelial neoplasia (HGPIN, the precursor lesion for usual peripheral zone prostate cancer) (1 ), invasion, growth and potential dedifferentiation of organ confined tumor (2 ), extension outside of the prostate (either before or after local therapy), development of metastasis, and the progression from androgen-dependent (AD) to androgen-independent (AI) disease (3 ). Clinical observations have shown that androgen withdrawal has the greatest impact on the development and progression of prostate cancer. Therefore, androgen deprivation has been the gold standard in treating patients who have advanced prostate cancer (4 ,5 ). More than 80% of such patients show a favorable response, as evidenced by a decrease in serum prostate-specific antigen (PSA) levels and tumor regression, but tumor growth could resume despite continuous treatment (6 ). Although the progressive development of AI is poorly understood, the androgen receptor (AR) appears to be pivotal in the progression from AD to AI prostate cancer. Four possible mechanisms by which the AR may be actively involved in the development of AI include (1) mutations in the AR (5 ,7 –11 ); (2) AR amplification (12 ,13 ); (3) ligand-independent activation of AR through other signaling pathways (14 –22 ); and (4) altered activity of AR coregulators (either co-activators or co-inhibitors) (22 –24 ).