Reactive oxygen metabolites have been postulated to play an important role in both toxic and ischemic forms of acute renal tubular epithelial injury. In the present study, we examined the effect of enzymatically generated hydrogen peroxide on LLC-PK1 cells, a renal proximal tubule cell line. Exposure of LLC-PK1 cells to glucose and glucose oxidase (GO; which generates hydrogen peroxide) resulted in cytotoxicity (as measured by trypan blue exclusion) which was dose dependent and increased linearly over time to 81 +/- 5% at 180 minutes (8 +/- 1% at time 0; mean +/- SEM, N = 3 to 7). Catalase (which decomposes hydrogen peroxide) completely prevented the cytoxicity, confirming that the toxicity was due to hydrogen peroxide production. To assess whether the hydrogen peroxide toxicity was a direct effect or mediated by other toxic oxygen metabolites. several scavengers of reactive oxygen metabolites and iron chelators were used. Superoxide dismutase (a scavenger of superoxide) had no effect. Deferoxamine (DFO), an iron chelator, provided marked Protection (GO alone 45.9 +/- 4.4%; GO + DFO 13.0 +/- 2.0%; control 7.1 +/- 1.2%; N = 15 to 17, P < 0.001). Pretreatment with DFO (1 hr, then 2 washes to remove DFO before GO addition) also markedly inhibited the cytotoxicity, suggesting that DFO's effect was due to iron chelation. Two other metal chelators (dihydroxybenzoic acid and 1,10-phenanthroline) also significantly decreased the GO-induced cytotoxicity. However, three of four hydroxyl radical scavengers used (mannitol, dimethyl sulfoxide, sodium benzoate) did not significantly decrease cell death. Only dimethylthiourea provided protection. Similar results, including the protective effects of catalase and iron chelators, were obtained using adenine release as a measure of cytotoxicity. Taken together, these data indicate that in addition to direct cytotoxicity of hydrogen peroxide, iron plays a critical role in hydrogen peroxide-mediated cytotoxicity to LLC-PK1 cells.