To test whether DNA injury contributes to TNF-induced cytotoxicity, we attempted to enhance DNA injury by inhibiting its repair and then assessing effects on cytotoxicity. DNA repair, assayed as unscheduled DNA synthesis, was first detected in TNF-sensitive targets by 2-3 h of incubation with TNF. Targets resistant to TNF cytotoxicity did not demonstrate significant repair replication. Repair preceded the detection of TNF-induced DNA injury, which was subsequently demonstrated by a double-stranded DNA fragmentation assay, sedimentation of DNA in neutral and alkaline sucrose gradients, and gel electrophoresis of extracted DNA. This suggested that early during exposure to TNF, DNA repair proceeds more rapidly than strand breakage. To inhibit repair, nontoxic concentrations of aphidicolin (inhibitor of DNA polymerase-alpha) and dideoxythymidine (inhibitor of DNA polymerase-beta and gamma) were used. Aphidicolin inhibited repair and consistently sensitized to TNF cytotoxicity, decreasing the ID50 for TNF at least 10- to 50-fold. In contrast, dideoxythymidine had no effect on repair or cytotoxicity. Deoxycytidine, which competitively inhibits binding of aphidicolin to DNA polymerase, blocked the sensitization in a concentration-dependent fashion. In targets sensitized with aphidicolin, TNF-induced strand breakage was accelerated, being detected by 4 h of culture in the sucrose gradient assay. Sensitization to TNF was not due to a heightened activation of poly (ADP-ribose) polymerase. These results indicate that TNF-induced strand breakage participates in TNF-induced cytotoxicity and that the level of DNA repair plays a role in determining relative sensitivity of targets.