AIM: To investigate the regulatory effect of electroacupuncture (EA) at Zusanli (ST36) on tumor necrosis factor-alpha (TNF-alpha) in rats with ulcerative colitis (UC), and further elucidate the therapeutic mechanism of EA on UC. METHODS: Thirty-two male Sprague-Dawley (SD) rats were randomly divided into four groups (n = 8): normal control group, UC control group, UC+ST36 group and UC+non-acupoint group. A solution containing ethanol and 2,4,6-trinitrobenzenesulfonic acid (TNBS) was instilled into the distal colon in the rat (at a dose of 100 mg/kg) to set up UC rat model. Rats in wakefulness state of UC+ST36 group were stimulated at ST36 by EA once a day, while those of UC+non-acupoint group were done at 0.5 cm beside ST36. After 10 d treatment, all rats were sacrificed simultaneously. Colon musocal inflammation and damage were assessed by measuring colon mass, morphologic damage score, colonic myeloperoxidase enzyme (MPO) activity, serum TNF-alpha and colonic TNF-alpha mRNA level. Morphologic damage score was examined under stereomicroscope. Colonic MPO activity was measured by spectrophotometer method. Serum TNF-alpha concentration was determined by radioimmunoassay (RIA). Colonic TNF-alpha mRNA expression level was analyzed by semiquantitative reverse transcription polymerase chain reaction (RT-PCR). RESULTS: Ratio of colonic mass/body mass (m(C)/m(B)) and activity of colonic MPO (mukat/g tissue) markedly increased (8.5 +/- 2.6 vs 2.5 +/- 0.4; 145 +/- 25 vs 24 +/- 8, P < 0.01 vs normal control group). Compared with normal control rats, serum TNF-alpha and colonic TNF-alpha mRNA level in UC control group were increased 2.5 fold (2 278 +/- 170 vs 894 +/- 248, P < 0.01) and 4.3 fold (0.98 +/- 0.11 vs 0.23 +/- 0.11, P < 0.01) respectively. After EA at ST36, m(C)/m(B) and MPO activity were reduced significantly (5.3 +/- 2.0 vs 8.5 +/- 2.6; 104 +/- 36 vs 145 +/- 25, P < 0.01, 0.05) compared with those of UC control group. Serum TNF-alpha and colonic TNF-alpha mRNA level were inhibited by EA stimulation at ST36 (P < 0.01). The inhibitory rate was 16% and 44% respectively. Morphologic damage score was also increased markedly in rat with UC (P < 0.01), whereas it was decreased by EA at ST36 (P < 0.05). There was no significant difference between UC control group and UC+EA at non-acupoint (P > 0.05). Furthermore, these parameters were highly correlated with each other (P < 0.01). CONCLUSION: Serum TNF-alpha concentration and colonic TNF-alpha mRNA expression level are increased significantly in UC rats in correlation with the severity of disease. It indicates that TNF-alpha is closely involved in the immune abnormalities and inflammatory responses in UC. EA at ST36 has therapeutic effect on UC by downregulating serum TNF-alpha and colonic TNF-alpha mRNA expression. High levels of TNF-alpha and its corresponding mRNA expression seem to be implicated in the pathogenesis of UC.