Since androgen receptors (AR) stimulate growth and block regression of the prostate through direct binding to nuclear sites, androgen withdrawal therapies must effectively eliminate AR from nuclei to bring about regression of prostatic carcinoma. Using this criterion, we compared the effects of surgical castration with those of 12 different androgen-withdrawal regimens on the rat prostate. The most potent treatment in this regard, a combination of cyproterone acetate and low-dose diethylstilbestrol, caused an initial objective response rate of 98% in 51 patients with stage D prostate cancer. Thus, although assays for AR are useful for evaluating new drugs, the high initial response rates to treatment make them irrelevant for routine patient selection. Accordingly, the principal reason for studying AR is to determine the cellular and molecular processes leading to the eventual emergence of androgen-resistant stem cells. At the cellular level, tumor recurrence in the androgen-dependent Shionogi tumor system is characterized by a 30-fold enrichment of stem cells that possess cytoplasmic AR but that lack the ability to retain AR in the nucleus. While at the molecular level definitive results await the purification of AR and their genes, the occurrence of biologically active truncated and chimeric forms of AR, analogous to those that have been synthesized in vitro for other steroid receptors, would predictably yield an androgen-resistant phenotype.