The role of nitric oxide (NO) in the etiology of ulcerative colitis is controversial with reports of the improvement and aggravation of colonic lesions by inducible NO synthase (iNOS) inhibitors. In the present study, we compared the effect of the selective iNOS inhibitor aminoguanidine and the nonselective NOS inhibitor N(G)-nitro-L-arginine methyl ester (L-NAME) on a dextran sulfate sodium (DSS)-induced model of colitis in rats. Experimental colitis was induced by a 3% DSS-solution added to drinking water for 7 days. Aminoguanidine (5 approximately 20 mg/kg) and L-NAME (10 mg/kg) were administered p.o. twice daily for the first 3 days, the last 3 days or all 6 days of DSS treatment. Body weight and severity of colitis (diarrhea, bloody feces) were observed over a period of 7 days. DSS treatment resulted in severe colonic lesions, accompanied by diarrhea, bloody feces, decrease of body weight and colon shortening. All of the parameters investigated improved significantly with aminoguanidine treatment at 20 mg/kg for 6 days or the last 3 days of DSS-treatment, but L-NAME did not significantly affect the colitis during these periods. When L-NAME or aminoguanidine was given in the first 3 days of DSS treatment, the colonic lesions were slightly aggravated by L-NAME but not affected by aminoguanidine. The expression of iNOS mRNA was observed from the 3(rd) day of DSS treatment. These results suggested that endogenous NO exerts a biphasic influence on DSS-induced colitis, depending on the NOS isoenzyme; a beneficial effect of NO derived from constitutive NOS and a detrimental effect of NO produced by iNOS in the development of colitis.