The question was addressed whether methylenediphenyl-4,4'-diisocyanate (MDI), a bifunctional electrophile, can induce DNA double-strand breaks (DSB) by repair of interstrand DNA crosslinks or whether DSB are the result of cell death. Cultured human lung epithelial cells (A549) were treated with MDI, methylene-4,4'-dianiline (MDA; a potential hydrolysis product of MDI), the nitrogen mustard melphalan, and the detergent Triton X-100. All chemicals were dissolved in ethylene glycol dimethyl ether which was added to a cell monolayer covered with phosphate-buffered saline. After 2 h, the treatment solution was exchanged against medium, and 8, 24, and 72 h after treatment initiation, the induction of DNA double-strand breaks was assessed by pulsed-field gel electrophoresis. At the same time, the viability was determined with the MTT test (intracellular reduction of the tetrazolium dye MTT). At the 8-h time point, 1 and 10 microM melphalan induced DSB without concomitant effect on cell viability. With all other chemicals, the dose-response curves for DNA fragmentation and viability were mirror images. Approximate 50% lethal concentrations were 200, 3000, and 100 microM for MDI, MDA, and Triton X-100, respectively. For these chemicals, the observed DSB were the consequence of extragenomic damage in the course of cell death rather than of an interaction with DNA. The mechanistic difference of melphalan was supported by analysis of nuclear morphology. Apoptotic bodies were observed only after melphalan treatment, whereas MDI and Triton X-100 produced only irregular clumping of chromatin (72-h time point). DNA fragment length analysis showed a time-independent pattern, with sizes between 1 and 4 Mbp for melphalan, while MDI, and Triton X-100 induced smaller DNA fragments in a time-dependent manner. It is concluded that DSB observed in cells treated with MDI are unlikely the result of DNA crosslink formation.