Prostate cancer is a major health issue in westernized countries, being considered a prototypical age-related, androgen-dependent tumor. However, data on the association between circulating androgens and prostate cancer have been inconsistent and mostly not compatible with the androgen hypothesis. In addition, plasma androgen-to-estrogen ratio appears to decrease with age, suggesting that estrogens may also have a role. Results from our own and others' studies suggest that circulating steroids cannot be considered representative of their actual intraprostatic levels. This is a consequence of the expression and/or activity of steroid enzymes, including 17beta-hydroxysteroid dehydrogenase (17beta-HSD), 5alpha-reductase, 3alpha/3beta-HSD, and aromatase, which may eventually lead to a differential tissue accumulation of steroid derivatives having distinct biological activities. Interestingly, many of the genes encoding for steroid enzymes are highly polymorphic in nature, although only a few studies have investigated their relation with prostate cancer and the data presently available are inconclusive. Locally produced or metabolically transformed estrogens may differently affect proliferative activity of prostate cancer cells. In our studies, estrogen may either stimulate or decrease prostate cancer cell growth, also depending on the receptor status. In particular, an imbalance of ERalpha and ERbeta expression may be critical to determine the ultimate estrogen effects on prostate cancer cell growth. Furthermore, evidence is accumulating that estrogens regulate gene transcription through an array of estrogen-response elements (EREs) and non-EREs, either ligand-dependent or -independent. This is further complicated by the presence of receptor isoforms, distinct cofactor interaction, and potential heterodimerization. Based on this combined evidence, a hypothetical model of prostate cancer progression is presented.