Previous investigations demonstrated that melatonin exerts an oncostatic action on estrogen-responsive breast cancer, both in vitro and in vivo. Nevertheless, the pro-apoptotic effect of melatonin is still a matter of debate. An experimental study was undertaken to focus on melatonin-related apoptosis and to identify the apoptotic pathways involved. Whole cell-count, flow-cytometry analysis and proteins involved in apoptotic pathways [p53, p73, murine double minute 2 (MDM2), caspases-9,-7,-6, cleaved-poly ADP ribose polymerase (PARP), Bcl-2, Bax and apoptotic inducing factor (AIF)] were investigated in human MCF-7 breast cancer cells treated with physiological (1 nM) concentration of melatonin. Melatonin exerts a significant growth-inhibitory effect on MCF-7 cells, becoming evident after 72 hr and thereafter increasing linearly up to 144 hr. In this model, the growth-inhibition is transforming growth factor beta 1 (TGF beta 1)-dependent and it might be reversed by adding an anti-TGF beta 1 antibody. Melatonin induces a significant rise in apoptotic rate, at both 24 and 96 hr. The anti-TGF beta 1 antibody almost completely suppresses melatonin-related late apoptosis; however, early apoptosis is unaffected. Early programmed cell death is associated with a significant increase in the p53/MDM2 ratio and in AIF release, without modifications in caspase activity or cleaved-PARP levels. Activated caspases-9 and -7 and cleaved-PARP increased significantly at 96 hr, concomitantly with a down-regulation of the Bcl-2/Bax ratio. These data suggest that two distinct apoptotic processes are triggered by melatonin in MCF-7 cells: an early, TGF beta 1 and caspase-independent response, and a late apoptotic TGF beta 1-dependent process in which activated-caspase-7 is likely to be the terminal effector.