Three diterpenoid quinones (royleanone- SAR 3, horminone- SAR 26, and acetyl horminone- SAR 43) isolated from the roots of Salvia officinalis L. were tested for their cytotoxic and DNA-damaging activity in human colon carcinoma cells Caco-2 and human hepatoma cells HepG2 cultured in vitro. Cytotoxicity was measured by the trypan blue exclusion technique and induction of apoptosis was evaluated by flow immunofluorocytometry after 30-300 min. exposure of HepG2 and Caco-2 cells to diterpenoid quinones and following 24 hr post-incubation in the culture medium. Induction of DNA breaks was measured after 60 min. exposure of cells to different concentrations of the compounds studied by the alkaline elution of DNA and by the Comet assay. Though all the quinones tested decreased the viability of the cells studied proportionally to the concentration and to the time of treatment (cytotoxicity= 30-60%), the increased level of apoptotic nuclei comparable to the level of apoptotic nuclei induced by a topoisomerase I inhibitor was proved only in HepG2 cells treated with 1x10(-4) mol/l SAR 26 or SAR 43. Either no or marginal increase of the level of apoptotic nuclei was observed in SAR 3-treated HepG2 cells and in SAR 3-, SAR 26- or SAR 43-treated Caco-2 cells. All compounds tested induced creation of DNA strand breaks in both cell types at concentrations >1x10(-7)-1x10(-6) mol/l. The occurrence of DNA strand breaks at different pH values as well as the kinetics of DNA breaks rejoining were evaluated only in colonic cells Caco-2. The Comet assay processed in parallel at pH 13.0 and pH 12.1 showed that strand breaks detected in SARs-treated colonic Caco-2 cells originated from alkali-labile sites, as induced DNA lesions were converted to DNA strand breaks only under strong alkaline conditions. The kinetics of DNA rejoining revealed that SARs-induced DNA breaks were repaired very slowly.