Various organs, including heart, kidneys, liver or brain, respond to brief exposures to ischemia with an increased resistance to severe ischemia/reperfusion and this phenomenon is called "preconditioning". No study so far has been undertaken to check whether such short, repeated gastric ischemic episodes protect gastric mucosa against severe damage caused by subsequent prolonged ischemia/reperfusion and, if so, what could be the mechanism of this phenomenon. The ischemic preconditioning was induced by short episodes of gastric ischemia (occlusion of celiac artery from one to five times, for 5 min each) applied 30 min before prolonged (30 min) ischemia followed by 3 h of reperfusion or 30 min before topical application of strong mucosal irritants, such as 100% ethanol, 25% NaCl or 80 mM taurocholate. Exposure to regular 30-min ischemia, followed by 3-h reperfusion, produced numerous severe gastric lesions and significant fall in the gastric blood flow and prostaglandin E(2) generation. Short (5-min) ischemic episodes (1-5 times) by itself failed to cause any gastric lesions, but significantly attenuated those produced by ischemia/reperfusion. This protection was accompanied by a reversal of the fall in the gastric blood flow and prostaglandin E(2) generation and resembled that induced by classic gastric mild irritants. These protective and hyperemic effects of standard preconditioning were significantly attenuated by pretreatment with cyclooxygenase-2 and cyclooxygenase-1 inhibitors, such as indomethacin, Vioxx, resveratrol and nitric oxide (NO)-synthase inhibitor, N(G)-nitro-L-arginine (L-NNA). The protective and hyperemic effects of standard preconditioning were restored by addition of 16,16 dm prostaglandin E(2) or L-arginine, a substrate for NO synthase, respectively. Gastroprotective and hyperemic actions of standard ischemic preconditioning were abolished by pretreatment with capsaicin-inactivating sensory nerves, but restored by the administration of exogenous CGRP to capsaicin-treated animals. Gene and protein expression of cyclooxygenase-1, but not cyclooxygenase-2, were detected in intact gastric mucosa and in that exposed to ischemia/reperfusion with or without ischemic preconditioning, whereas cyclooxygenase-2 was overexpressed only in preconditioned mucosa. We conclude that: (1) gastric ischemic preconditioning represents one of the most powerful protective interventions against the mucosal damage induced by severe ischemia/reperfusion as well as by topical mucosal irritants in the stomach; (2) gastric ischemic preconditioning resembles the protective effect of "mild irritants" against the damage by necrotizing substances in the stomach acting via "adaptive cytoprotection" and involves several mediators, such as prostaglandin derived from cyclooxygenase-1 and cyclooxygenase-2, NO originating from NO synthase and sensory nerves that appear to play a key mechanism of gastric ischemic preconditioning.