Reactive oxygen species assay-based risk assessment of drug-induced phototoxicity: Classification criteria and application to drug candidates

We have previously demonstrated that the phototoxic potential of chemicals could be partly predicted by the determination of reactive oxygen species (ROS) from photo-irradiated compounds. In this study, ROS assay strategy was applied to 39 marketed drugs and 210 drug candidates in order to establish provisional classification criteria for risk assessment of drug-induced phototoxicity. The photosensitizing properties of 39 model compounds consisting of phototoxic and non-phototoxic chemicals, as well as ca. 210 drug candidates including 11 chemical series were evaluated using ROS assay and the 3T3 neutral red uptake phototoxicity test (NRU PT). With respect to marketed drugs, most phototoxic drugs tended to cause type I and/or II photochemical reactions, resulting in generation of singlet oxygen and superoxide. There seemed to be a clear difference between phototoxic drugs and non-phototoxic compounds in their abilities to induce photochemical reactions. A plot analysis of ROS data on the marked drugs provided classification criteria to discriminate the photosensitizers from non-phototoxic substances. Of all drug candidates tested, 35.2% compounds were identified as phototoxic or likely phototoxic on the basis of the 3T3 NRU PT, and all ROS data for these phototoxic compounds were found to be over the threshold value. Furthermore, 46.3% of non-phototoxic drug candidates were found to be in the subthreshold region. These results verify the usefulness of the ROS assay for understanding the phototoxicity risk of pharmaceutical substances, and the ROS assay can be used for screening purposes in the drug discovery stage.