1. This study was designed to investigate the role of rat phosphodiesterase 3 (RPDE3) in regulation of liver metabolism in sepsis. We studied the effects of the phosphodiesterase 3 inhibitor (PDI), enoximone, alone and in combination with regulating factors of hepatic carbohydrate metabolism and bile secretion in the perfused liver of rats treated 4 h earlier with endotoxin. In addition, cyclic AMP and cyclic GMP levels were determined in the effluate and bile by radio immunoassay methods. 2. After endotoxin treatment, infusion of enoximone at three concentrations (1 microM, 10 microM) resulted in an increased glucose output from -1.4 +/- 0.9 to 7.8 +/- 2.5 mumol l-1 20 min-1. Bile acid-independent bile flow increased also, in a dose-dependent manner. 3. In untreated livers, cyclic AMP release increased in the effluate from 1000 +/- 73 fmol g-1 min-1 to 1710 +/- 143 fmol g-1 min-1 when enoximone (10 microM) was administered. In bile from untreated livers, the level of cyclic AMP was also significantly increased by enoximone. After endotoxin treatment, the enoximone (10 microM) effect on cyclic AMP levels in effluate and bile was greatly reduced. Levels of cyclic GMP in the effluate and bile appeared unchanged in the presence of enoximone. 4. During co-infusion of glucagon (1 nM) and enoximone (10 microM), cyclic nucleotide levels in the effluate and bile of livers after endotoxin treatment were determined. In the effluate, cyclic AMP release increased from 827 +/- 144 fmol g-1 min-1 to 17802 +/- 2821 fmol g-1 min-1 when glucagon was administered. The presence of enoximone enhanced cyclic AMP further to 41696 +/- 920 fmol g-1 min-1. The same changes in cyclic AMP release were found in bile. Levels of cyclic GMP in the effluate and bile were not significantly affected by the administration of glucagon and the PDI. 5. Glucose release was determined during glucagon, sympathetic nerves stimulation and phenylephrine administration in the presence and absence of enoximone. The addition of enoximone to glucagon increased glucose release by 8.2 +/- 2.8 mumol g-1 20 min-1, without alteration of lactate balance. The PDI enhanced the glycogenolytic effects of nerve stimulation and of phenylephrine, accompanied by a reduction in lactate production. 6. Enoximone significantly enhanced the bile acid independent bile flow after glucagon, nerves stimulation and after administration of phenylephrine. Bile acid secretion was unaffected by the PDI. The vasoconstrictor effect of nerve stimulation was reduced by the PDI. 7. We conclude that endotoxin treatment reduces the ability of the PDI, enoximone, to increase cyclic AMP release in the perfused liver. The significant increase in cyclic AMP release after stimulation with glucagon and enoximone favours the view that RPDE3 is involved in the degradation of cyclic AMP in the liver after exposure to endotoxin. Additionally, the inhibition of the RPDE3 results in glucose release, vasodilatation and choleresis in endotoxin pretreated livers.