The ability of various leukotrienes to alter the susceptibility of the rat gastric mucosa to injury by 20% ethanol and the possible mechanism of action were examined using an ex vivo gastric chamber preparation. Intraarterial infusions of leukotriene B4 or N-acetyl leukotriene E4 (0.01-1.0 microgram/kg per min for 10 min) had no significant effect on the extent of damage induced by topically applied 20% ethanol. However, infusion of leukotriene C4, D4, or E4 (1.0 micrograms/kg per min) significantly increased ethanol-induced damage, as measured macroscopically, histologically, and functionally. Lower doses of leukotriene C4, D4, or E4 were without significant effect in this model. The increase in damage induced by these three leukotrienes could be blocked by pretreatment with either of two structurally unrelated leukotriene D4 antagonists (L-649,923 or L-660,711). The augmentation of damage by leukotriene C4 was not affected by pretreatment with indomethacin or with a specific thromboxane A2-receptor antagonist (L-670,596). At the dose that increased ethanol-induced damage, none of the leukotrienes tested significantly altered gastric vascular permeability, as measured by Evan's blue leakage. However, using laser-Doppler flowmetry, leukotrienes C4 and D4 were found, when administered intraarterially at doses in the 0.05-1.0 micrograms/kg per min range, to produce dose-dependent reductions of gastric blood flow while N-acetyl leukotriene E4 was without effect and leukotriene B4 induced slight increases. The effects of leukotrienes C4 and D4 on gastric blood flow could be inhibited by the two leukotriene D4 antagonists but not by the thromboxane antagonist. These results demonstrate that although they do not produce damage by themselves, leukotrienes C4, D4, and E4 are capable of augmenting ethanol-induced injury to the gastric mucosa. Changes in vascular permeability do not appear to play a role in the mechanism of action of the leukotrienes, while their effects on gastric blood flow are likely to be important. Under certain condition it is therefore possible that local release of leukotrienes could, at least in part through reducing vascular perfusion, predispose the surrounding tissue necrosis.