Methods
A round-shaped wound was created in the monolayers of rabbit gastric epithelial cells and the wound restitution was monitored by measuring the speed of cell migration. The degree of cellular apoptosis (Tunel, index%=positive cell numer/total cell number) was assessed using Becton-Dickinson FACScan flow cytometry. The monolayers were incubated with or without NO-donor sodium nitroprusside (SNP, 300mM). The potential role of cGMP as a second messenger of NO was investigated using 8-Br-cGMP. In order to assess whether the effects of NO are mediated by its reaction with superoxide (O₂^.-), yielding peroxynitrite (ONOO^-)₃, we studied the effects of O₂^.- sources diethylthiocarbamate (DDC, superoxide dismutase inhibitor) and pyrogallol (PG, an O"d2^.- generator) with or without SNP. We further studied the effects of O₂^.- scavengers allopurinol (AP, an inhibitor of xantine oxidase), and superoxide dismutase (SOD, catalyzes O₂^.- to H₂O₂) together with catalase (C, eliminates H₂O₂) with or without of SNP.

Results
Retardation of wound healing caused by SNP was enhanced by DDC and PG, and ameliorated by O₂^.- scavengers AP and SOD+C (Table 1). 8-Br-cGMP had no significant effect on wound restitution (cell migration (mm/24 h) 546±16 vs. 440±68, and apoptosis index % 8.2±3.4 vs. 9.4±2.4 in cGMP and controls, respectively).

The migration speed in the controls, SNP, SNP+DDC, SNP+PG, SNP+SOD and SNP+AP groups were 440±68 357±90* 232±99*§ 52±95*§ 406±90*§ and 388±88*(mm/24 h), consecutively, and the apoptosis index (%/24 h) 9±2 23±4* 23±2* 31±7*§ 15±5*§ and 15±3*, consecutively (*p< 0.001 as compared to controls, §p< 0.001 as compared to SNP, N=6 each).

Conclusions
SNP retards gastric restitution by inhibiting cell migration and inducing cellular apoptosis. These effects are not mediated by cGMP, but (ONOO^-)₃ or its derivatives may contribute to this process.