Can prostate-specific antigen be used as a valid end point to determine the efficacy of chemotherapy for advanced prostate cancer?

It is current practice in many clinical trials evaluating new chemotherapy regimens for the treatment of advanced prostate cancer to use prostate-specific antigen (PSA) decline as a response criteria with the assumption that the level of PSA reflects the efficacy of chemotherapy. Advanced prostate cancer is heterogeneous; therefore, the validity of PSA decline as a measurable end point was studied in advanced human prostate-cancer cell lines: androgen-sensitive LNCaP and androgen-insensitive PC3 cells. Each cell line was grown for 4 days with escalating doses of Adriamycin or vinblastine. Cell counts, intracellular PSA concentrations, and secreted PSA levels were determined daily for 4 days. Untreated LNCaP cells had constant secretion of PSA per cell. In contrast, LNCaP cells treated with Adriamycin or vinblastine had an 80% reduction in cell numbers and a 3-fold increase in secreted PSA per cell by day 4. In contrast, PC3 cells had a different response to Adriamycin and vinblastine. Both drugs reduced cell numbers by 97% of control values and suppressed PSA production in the remaining viable cells by 4 days in culture. Thus, prostate-cancer cell production of PSA is variable with chemotherapy and the PSA level may not accurately reflect the actual tumor response to chemotherapy.