BACKGROUND Management of prostate cancer that either is detectable by prostate specific antigen (PSA) measurements after curative intent or has spread outside of its capsule is a serious problem. Innovative, nontoxic approaches to the disease are required. One approach might be therapy with retinoids. Retinoid activities are mediated by two distinct families of transcription factors: the retinoic acid receptors (RARs) and retinoid X receptors (RXRs), which can induce transcriptional activation through specific DNA sites or by inhibiting the transcription factor AP-1 that usually mediates cellular proliferative signals. The RARs require heterodimerization with RXRs. RXRs can form either heterodimers or homodimers; and the latter can bind to DNA response elements that are distinct from those bound by the RAR/RXR heterodimers. METHODS A series of novel synthetic retinoids that selectively interact with RXR/RXR homodimers or RAR/RXR heterodimers, or that selectively inhibit AP-1 activity without activating transcription were evaluated for their ability to inhibit clonal growth of three human prostate cancer cell lines (PC-3, DU-145, and LNCaP). RESULTS Several notable findings were: 1) RXR-selective retinoids, such as SR11246, were able to inhibit the clonal growth of prostate cancer cells. In contrast, SR11246 had little effect on clonal growth of myeloid leukemic cells. 2) RAR-selective retinoids also inhibited clonal growth of prostate cancer cells. 3) The retinoid (SR11238) with potent anti-AP-1 activity had no effect on the clonal growth of prostate cancer cells. CONCLUSIONS This study shows that both RXR- and RAR-selective retinoids are worthy of further study and may be candidates for future clinical trials in prostate cancer.