The present study was undertaken to explore the role of nitric oxide (NO) in the pathogenesis of experimental non-steroidal anti-inflammatory drug (NSAID)-induced gastropathy. We assessed the role of NO inhibition and donation in indomethacin-induced gastric mucosal dysfunction. The stomach was perfused with vehicle (control) for 20 min, followed by indomethacin (10 mg ml-1 in 1 25 % sodium bicarbonate, pH 8 4) for 120 min. NG-nitro-L-arginine methyl ester (L-NAME, 5 and 10 mg kg-1, I.V. bolus), L-arginine, D-arginine (100 mg kg-1 I.V. bolus, 10 mg kg-1 h-1, 2 h infusion) and the NO donor glyceryl trinitrate (GTN) were given at the same time (20, 40 and 80 microg kg-1 min-1, 15 min infusion) as perfusion with indomethacin was started. Epithelial permeability was quantified by measuring blood-to-lumen clearance of 51Cr-labelled EDTA. Indomethacin caused a 20-fold increase in 51Cr-EDTA leakage compared with that of the control group. Treatment with L-NAME or L-arginine did not affect the indomethacin-induced alterations in mucosal permeability. Administration of GTN (20 microg kg-1 min-1) significantly reduced the indomethacin-induced mucosal dysfunction. By contrast, higher doses of GTN (80 microg kg-1 min-1) exacerbated epithelial dysfunction induced by indomethacin. Elevated levels of carbonyls and myeloperoxidase (MPO) observed after indomethacin administration were significantly reduced, to the control values, when GTN (20 microg kg-1 min-1) was administered along with indomethacin. These data suggest that NO from exogenous sources can exert a dual action on the integrity of the gastric mucosa challenged by indomethacin. Low doses of GTN can prevent mucosal dysfunction induced by indomethacin, while higher doses of GTN may exacerbate the increases in epithelial permeability.