Hepatocyte heterogeneity in response to extracellular adenosine

Metabolic and haemodynamic effects of adenosine were studied in antegrade and retrograde rat liver perfusions with influent nucleoside concentrations either below (i.e. 20 microM) or exceeding (i.e. 200-300 microM) the single-pass clearance capacity of the liver. Adenosine (20 microM) increased in antegrade perfusions the perfusion pressure and markedly stimulated prostaglandin D2, thromboxane B2 and glucose output, whereas in retrograde perfusions no pressure and eicosanoid response occurred and glucose output was stimulated only slightly. The perfusion-direction-dependent differences in the glucose and pressure response to adenosine (20 microM) were fully abolished in presence of ibuprofen (50 microM). When the adenosine concentration in influent was raised to 200-300 microM, i.e. to a concentration exceeding single-pass clearance of the nucleoside, the adenosine-induced prostaglandin D2 release was about 10-fold higher in retrograde perfusions than in antegrade perfusions. On the other hand, both adenosine (20-300 microM)-induced cyclic AMP (cAMP) and K+ release from the liver were not affected by the direction of perfusion, and maximal effects on cAMP release were observed at influent adenosine concentrations of 100 microM. The basal rate (adenosine absent) of prostaglandin D2 and thromboxane B2 release was about 10-fold higher in retrograde than in antegrade perfusion experiments, whereas the basal cAMP release from the liver was not affected by the direction of perfusion. Maximal adenosine-stimulated glucose output was significantly higher in antegrade than in retrograde perfusions at all adenosine concentrations tested (range 10-300 microM). Ibuprofen abolished this difference, indicating that eicosanoids liberated under the influence of adenosine contribute to the glycogenolytic response in antegrade, but not in retrograde, perfusion. Desensitization occurred following repetitive adenosine infusion; this was more pronounced for adenosine-induced prostaglandin release than for cAMP or K+ efflux. The data suggest the following. (i) Both cAMP and eicosanoids are involved in the stimulation of glycogenolysis by adenosine. (ii) Eicosanoids are probably liberated under the influence of extracellular adenosine from a portal pre-sinusoidal compartment and accordingly stimulate glycogenolysis only in antegrade perfusions. Thus signals derived from portal vein structures can modulate hepatocellular function. (iii) Contractile elements are probably located also inside the liver acinus. (iv) Eicosanoids released into the hepatic vein reflect less than 10% of hepatic eicosanoid formation, because of marked clearance by perivenous hepatocytes.