Although the urinary excretion of NAG and gamma-GTP has been employed as an index of renal injury, the mechanism of release of these enzymes remains unclear. In the case of renal ischemic injury, we recently reported that u-NAG and u-gamma-GTP were not correlated well with the degree of renal damage especially in the late phase of ischemic injury [1]. In order to elucidate the relationship between release of these enzymes and renal ischemic injury, we devised an in vitro model of anoxic injuries to renal tubular cells and measured the time-course release of NAG (a lysosomal enzyme), gamma-GTP and LAP (brush border enzymes) at 37 degrees C under five different aerobic and anaerobic conditions (95% O2, 20% O2, 10% O2, 3% O2 and 0% O2). The ATP and lactate levels in the system were also measured to estimate the metabolic state of the tubular cells. In the 3% O2, 10% O2 and 20% O2 groups, NAG, gamma-GTP and LAP were released into the media at a similar rate to that in the 95% O2 group, and anaerobic damage to the renal tubular cells was not observed. In the 0% O2 group, the NAG in the medium increased from 60 to 180 min at a significantly higher rate than that of the 95% O2 group, and a linear relationship was observed between NAG concentration and incubation time (r = 0.73), although NAG did not increase significantly at 30 min.(ABSTRACT TRUNCATED AT 250 WORDS)