OBJECTIVE To investigate whether the neurohormone melatonin can prevent apoptosis caused by deprivation of oxygen, glucose, and serum (experimental ischemia) in cultured human retinal pigment (RPE) cells. METHODS Cultures of human RPE cells established from a variety of donors were grown to passage four and then subjected to experimental ischemia, with or without various substances, for up to 72 hours. Cells were examined for morphologic changes and breakdown of DNA, assessed by TdT-dTUP terminal nick-end labeling (TUNEL) and agarose gel electrophoresis. Changes in transcription and translation of various proto-oncogenes (bcl-2, TIAR, ICH-1S/1) were assessed by analysis of mRNA and protein levels, respectively. The effect of various substances on the iron-ascorbate-induced formation of reactive oxygen species (ROS) in chick retinal dissociates was also investigated. RESULTS Cultured human RPE cells on coverslips that were incubated in serum-free medium, glucose, and oxygen remained viable for up to 40 hours. Thereafter, there was a steady decrease in cell numbers and an increase in the number of cells labeled by the TUNEL method. By 72 hours 65% of cells remained attached to the coverslips, of which approximately 65% were TUNEL positive. Furthermore, most of the experimental ischemia-treated cells exhibited a shrunken appearance typical of apoptosis. Fragmentation of the DNA from cells in which ischemia was induced for 72 hours was also confirmed by agarose gel electrophoresis. Inclusion of 100 microM melatonin significantly decreased the amount of apoptotic cell nuclei after ischemia, but the effect was mild compared with that of fetal calf serum, which almost completely counteracted cell death. The action of melatonin was not prevented by 0.01 mM to 1 mM luzindole, a melatonin receptor antagonist. In addition, 100 microM ascorbate did not counteract ischemia-induced apoptosis. Treatment of RPE cells with 100 microM flupirtine gluconate for 72 hours caused an upregulation of the proto-oncogene protein Bcl-2 and a decrease in TIAR and ICH-1L proteins compared with that in control cells. Melatonin at 100 microM had no such effect. The levels of the mRNA transcripts for ICH-1L relative to those for ICH-1S were significantly decreased in cultures treated with 100 microM flupirtine or 100 microM melatonin when compared with levels in control cells. However, the effect of flupirtine was greater than that of melatonin. Ten micromolar ascorbate and 5 microM iron stimulated the formation of ROS in chick retinal cell dissociates. Ascorbate, melatonin, and flupirtine (all at 100 microM) blunted this response in the order flupirtine > melatonin >> ascorbate. Luzindole had no effect, alone or in the presence of melatonin. CONCLUSIONS The presented data show that melatonin counteracted ischemia-induced apoptosis in human RPE cells by a process that seemed to be independent of melatonin receptors. Moreover, melatonin and flupirtine counteracted iron-ascorbate-induced ROS formation and decreased the ratio of mRNA for ICH-1L and ICH-1S. However, melatonin was less potent than flupirtine in its action in each case, which suggests that either the two compounds act on different signaling pathways or that they act on the same pathway with differing potency. The failure to detect an effect of melatonin on the levels of Bcl-2, ICH-1L, and TIAR proteins when compared with the effect of flupirtine was probably caused by the sensitivity of the procedures. It is suggested that substances that can prevent ROS formation can potentially nullify apoptotic cell death, but this is difficult to detect experimentally when the substance has only a mild effect, such as in the case of ascorbate.