In addition to suppressing breast cancer cell growth, retinoids potentiate growth inhibition in human breast cancer when tested in vitro and in vivo with tamoxifen and/or interferon. The purpose of this study was to ascertain the biologic effects of all-trans-retinoic acid (ATRA) administered alone and with tamoxifen +/- interferon and to identify the relationship between ATRA plasma concentrations and optimal biological dose (the lowest dose that produces a biological response). Three consecutive groups of 15 patients with locally advanced operable breast cancer were treated, in accordance with good clinical practice (GCP) requirements, with ATRA at 3 dose levels alone or with tamoxifen +/- alpha-interferon 2a at flat doses. After 3 weeks, the tumors were surgically removed. Biological parameters measured at the beginning (in biopsy tissue) and end (in surgical tissue) of the study were compared. The optimal biological dose for ATRA was 15 mg/m2/day. Treatments influenced tumor grade but not cell cycle kinetics (G0-G1 phase) or proliferation (Ki67 levels). ATRA induced progesterone receptors independent of dose level and co-administered drugs, but did not induce estrogen receptors when administered alone. Retinoic acid receptor (RAR)-alpha was not affected by treatment and RAR-alpha was moderately influenced whereas RAR-beta (concomitantly with transforming growth factor-beta) was induced in 33% of patients by ATRA alone. ATRA pharmacokinetics were dose- and time-dependent. Neither the ATRA + tamoxifen nor the ATRA + tamoxifen + interferon combinations potentiated the ATRA-induced biological changes. Future studies evaluating the role of RAR-beta as a biological marker of retinoid activity are warranted.