The effects of hydrogen peroxide (H2O2) on the action of basally produced endothelium-derived relaxing factor (EDRF) were investigated by measuring cGMP accumulation in single and cocultures of calf pulmonary artery endothelial cells (CPAEs) and rabbit pulmonary artery smooth muscle cells (RPASMs) as a model for determining the contribution of EDRF dysfunction to altered vascular tone and reactivity frequently associated with oxidant-induced vascular injury. Higher cGMP levels in long-term cocultures (20.4±1.8 pmol/mg protein/15 min) than in single-cell cultures (CPAE, 9.6±0.9 pmollmg protein/15 min; RPASM, 3.7±0.2 pmol/mg protein/15 min), and CPAE-induced increases (fivefold) in intracellular RPASM cGMP content in short-term cocultures suggest basal release of EDRF. Basal generation and release of an l-arginine-derived endothelial labile factor accounted for the increases in cGMP, since the response was completely blocked by pretreatment of CPAEs withNGmonomethyl l-arginine. Pretreatment of long-term cocultures with H2O2for 30 minutes resulted in a dose-dependent (0.5–2 mM) decrease in cGMP formation (49–79%o). To determine the effects of H2O2on EDRF synthesis, transport, and RPASM responsiveness, CPAEs or RPASMs were selectively pretreated with H2O2before establishment of short-term cocultures. In cocultures of H2O2-pretreated CPAEs with untreated RPASMs, RPASM cGMP levels were reduced, suggesting a decrease in EDRF production rather than deterioration of EDRF during transport, because cGMP levels were unaffected by posttreatment with oxygen radical scavengers during coculture. Pretreatment of RPASMs with H2O2attenuated the untreated CPAE-induced, the putative EDRF S-nitroso-l-cysteine-induced, or the nitroprusside-induced increases in RPASM cGMP levels. This attenuation was prevented by pretreatment with either dimethylthiourea, deferoxamine, or dithiothreitol, suggesting a mechanism of H2O2action involving iron-catalyzed formation of intracellular hydroxyl radicals and their attack on cellular thiols. H2O2diminution of cGMP accumulation was not associated with lytic cell injury in the experimental time frame, because morphology and51Cr release from prelabeled RPASMs and CPAEs were unchanged.