Which of the reactive oxygen species (ROS) is/are responsible for the cytotoxicity of 95% O2 for rat distal fetal lung epithelial cells in vitro is unknown. The cytotoxicity of O2 has previously been shown (Am J Physiol 258:L165-L172,1990) to be reduced by pretreatment with liposomes containing superoxide dismutase (SOD) and catalase (CAT), or pretreatment with N-acetylcysteine (Arch Biochem Biophys 313:131,1994) to maintain intracellular non-protein thiols. We have recently found that 100 μM Trolox, a vitamin E analogue, is also completely protective (P<0.05). This protection can not be attributed to superoxide scavenging. Though pretreatment with 25 U/ml polyethylene glycol (PEG)-conjugated SOD and 50 U/ml CAT (U/ml) is also partially protective (P<0.05), treatment with PEG-SOD alone, or the SOD mimic TEMPO (0-100 μM), is not. Liposomal CAT, sufficient to increase cell enzyme activity by 50%, was partially protective (P<0.05) against O2-mediated cytotoxicity, while 100 μM Trolox or 10 μM U74389G were partially protective (P<0.05) against 50 μM H2O2-mediated injury, but not injuries generated with t-butyl or cumene hydroperoxides. Trolox could also act to limit peroxynitrite formation, but the peroxynitrite scavenger lipoic acid is not protective against 95% O2-mediated cytotoxicity. Hydroxyl radical production is significantly increased (P<0.05) with 95% O2, and Trolox may also limit formation of this ROS. Together, these findings are consistent with H2O2 formation being a critical component of 95% O2-mediated cytotoxicity which can be reduced by scavenging with either CAT or Trolox.