Utilizing the isolated perfused rat liver, we examined several factors influencing efflux of glutathione [reduced glutathione (GSH) and glutathione disulfide (GSSG)] into perfusate and bile, including the effects of perfusate composition, oxygen delivery to the liver, and acivicin (AT-125), an inhibitor of gamma-glutamyl transferase activity. When livers were perfused with a recirculating Krebs-Ringer bicarbonate buffer only 7-26% of released glutathione was excreted into bile, mainly in its oxidized form (71-90% as GSSG). In contrast, when 20% bovine red blood cells or 20% fluorocarbon emulsion were utilized as perfusates, biliary glutathione accounted for a larger fraction of total hepatic efflux (16-41%), and only 39-65% was excreted as GSSG. To determine whether O2 delivery to the liver could explain some of these differences, biliary and sinusoidal efflux of glutathione were measured as O2 delivery was varied by 1) increasing the perfusion flow rate, 2) altering the concentration of fluorocarbon emulsion (5, 10, and 20%), and 3) changing the PO2 (95% O2-5% CO2 vs. 50% O2-5% CO2-45% N2). Under all experimental conditions, an increase in O2 delivery was accompanied by an increase in bile flow and in the concentration and rate of glutathione efflux into bile but no significant change in sinusoidal efflux of glutathione. Hepatic tissue GSH and GSSG levels were not affected by the various treatments. When gamma-glutamyl transferase activity was inhibited with AT-125, biliary glutathione increased to levels of approximately 50% of total hepatic efflux in fluorocarbon-perfused livers, and only 24-29% of the glutathione was excreted as GSSG.(ABSTRACT TRUNCATED AT 250 WORDS)