Mounting evidence supports a key role played by estrogen or estrogen in synergy with an androgen, in the pathogenesis of prostate cancer (PCa). New experimental data suggest that this process could begin as early as prenatal life. During adulthood, estrogen carcinogenicity is believed to be mediated by the combined effects of hormone-induced, unscheduled cell proliferation and bioactivation of estrogens to genotoxic carcinogens. Increased bioavailability of estrogen through age-dependent increases in conversion from androgen could also be a contributing factor. Individual variations and race-/ethnic-based differences in circulating or locally formed estrogens or in tissue estrogen responsiveness may explain differential PCa risk among individuals or different populations. Estrogen receptor (ER)-alpha and ER-beta are the main mediators of estrogen action in the prostate. However, ER-beta is the first ER subtype expressed in the fetal prostate. During cancer development, ER-beta expression is first lost as tumors progress into high grade in the primary site. Yet, its reexpression occurs in all metastatic cases of PCa. A change in cytosine methylation in a regulatory CpG island located in the proximal promoter of ER-beta may constitute an "on/off" switch for reversible regulation of ER-beta expression. A variety of estrogenic/antiestrogenic/selective estrogen receptor modulator (SERM)-like compounds have been shown to use non-ERE pathways, such as tethering of ER-beta to NF-kappaB binding proteins, Sp2, or Ap1 for gene transactivation. These findings open new avenues for drug design that now focuses on developing a new generation of estrogen-based PCa therapies with maximal proapoptotic action but few or no side effects.