Prostaglandins (PG) derived from COX-1 are essential for the maintenance of mucosal integrity but COX-2 isoform synthesizes PG at a site of inflammation. Recently, COX-2 mRNA expression was demonstrated at the ulcer edge during healing of chronic gastric ulcers but the role for expression of COX-2 and its products such as PGE(2) and cytokines including interleukin (IL-1beta) and tumor necrosis factor alpha (TNFalpha) in ulcer healing remains unknown. In this study, Wistar rats with gastric ulcers produced by serosal application of acetic acid (ulcer area 28 mm(2)) received daily treatment either with: (1) vehicle (saline); (2) NS-398 (10 mg/kg-d i.g.) and Vioxx (5 mg/kg-d i.g.), both, highly specific COX-2 inhibitors; (3) meloxicam (5 mg/kg-d i.g.), a preferential inhibitor of COX-2; (4) resveratrol (10 mg/kg-d i.g.), a specific COX-1 inhibitor; (5) indomethacin (5 mg/kg-d i.g); and (6) aspirin (ASA; 50 mg/kg-d i.g.), non-selective inhibitors of both COX-1 and COX-2. At day 3, 7, and 14 after ulcer induction, the animals were sacrificed and the area of gastric ulcers was determined by planimetry and histology, gastric blood flow (GBF) at ulcer base and margin was measured by H(2) clearance technique, and blood was withdrawn for measurement of plasma IL-1beta and TNFalpha levels. The mucosal biopsy samples were taken for the determination of PGE(2) generation by RIA and expression of COX-1, COX-2, IL-1beta, and TNFalpha mRNA by RT-PCR. In vehicle-treated rats, gastric ulcers healed progressively and at day 14 the healing was completed, accompanied by a significant rise in the GBF at ulcer margin. The IL-1beta, TNFalpha, and COX-1 mRNA were detected in intact and ulcerated gastric mucosa, whereas COX-2 mRNA were upregulated only in ulcerated mucosa with peak observed at day 3 after ulcer induction. The plasma IL-1beta level was significantly increased at day 3 and 7 but then declined at day 14 to that measured in vehicle-controls. Indomethacin and ASA, which suppressed PGE(2) generation both in the non-ulcerated and ulcerated gastric mucosa, significantly delayed the rate of ulcer healing and this was accompanied by the fall in GBF at ulcer margin and further elevation of plasma IL-1beta and TNFalpha levels, which was sustained up to the end of the study. Treatment with NS-398 and Vioxx, which caused only a moderate decrease in the PGE(2) generation in the non-ulcerated gastric mucosa, delayed ulcer healing and attenuated significantly the GBF at ulcer margin and PGE(2) generation in the ulcerated tissue, while raising the plasma IL-1beta and TNFalpha similarly to those observed in indomethacin- and ASA-treated rats. Resveratrol, which suppressed the PGE(2) generation in both non-ulcerated and ulcerated gastric mucosa, prolonged ulcer healing and this was accompanied by the fall in the GBF at the ulcer margin and a significant increase in plasma IL-1beta and TNFalpha levels. We conclude that (1) classic NSAID delay ulcer healing due to suppression of endogenous PG, impairment in GBF at ulcer area, and excessive cytokine expression and release, and (2) this deleterious effect of classic NSAID on the healing of pre-existing ulcers can be reproduced by selective COX-1 and COX-2 inhibitors, suggesting that both COX isoforms are important sources of PG that appear to contribute to ulcer healing.
Copyright 2001 Wiley-Liss, Inc.