Indirect evidence suggests that oxygen radicals may contribute to ischemic preconditioning. We directly investigated whether exposure to oxygen radicals per se, in the absence of ischemia, could reproduce the beneficial effects of ischemic preconditioning on infarct size and on postischemic contractile dysfunction. In one branch of the study, isolated rabbit hearts underwent 30 minutes of total global ischemia and 45 minutes of reperfusion (n=6, control group). A second group, before ischemia/reperfusion, was exposed for 5 minutes to a low flux of oxygen radicals generated by purine/xanthine oxidase (P/XO), followed by a 15-minute washout (n=6). Oxygen radical pretreatment significantly improved postischemic recovery of contractile function. We then investigated in another branch of the study whether this preconditioning effect would also reduce infarct size and whether it was mediated by protein kinase C activation. Control hearts were subjected to coronary artery occlusion for 30 minutes, followed by 2.5 hours of reperfusion (n=6). A second group, before coronary occlusion, was exposed to oxygen radicals and washout as described (n=8). A third group was subjected to oxygen radical infusion, but an inhibitor of protein kinase C (polymyxin B, 50 micromol/L) was administered throughout subsequent ischemia (n=7). A fourth group was exposed to oxygen radicals in the presence of scavengers (superoxide dismutase, 250 U/mL; catalase 500, U/mL; n=8). Pretreatment with oxygen radicals markedly reduced infarct size, from 65+/-19% of risk region in controls to 12+/-4% (P<.05). Protein kinase C inhibition significantly attenuated this effect (infarct size, 37+/-9% of risk region; P<.05 versus P/XO; P=NS versus controls). Oxygen radical-induced preconditioning was prevented by scavengers (infarct size, 55+/-14% of risk region; P<.05 versus P/XO; P=NS versus P/XO+polymyxin B). Our data show that in the absence of ischemia, exposure to low concentrations of oxygen radicals can reproduce the beneficial effects of ischemic preconditioning on infarct size and postischemic recovery of left ventricular function. Thus, oxygen radicals might be potential contributors to ischemic preconditioning.