Experiments were performed in isolated non-working guinea pig hearts perfused according to the Langendorff technique (95% O2, 5% CO2), to evaluate the relative contribution of the coronary endothelium to the formation of cardiac adenosine during hypoxia, hypercapnia, and acetylcholine infusion. For this purpose the adenine-nucleotides of the coronary endothelium were prelabeled by perfusion of isolated hearts with 3H-adenosine (10(-8) M) for 35 min. Changes in the relative specific radioactivity (RSA) of adenosine released into the coronary effluent perfusate were used to assess changes in the relative contribution of the coronary endothelium and cardiomyocytes to total cardiac adenosine release. Hypoxic perfusion (15% O2) doubled coronary flow and increased total adenosine release by about two orders of magnitude and in addition, substantially increased the release of 3H-adenosine. The RSA of adenosine, however, was consistently depressed. During hypercapnic acidosis (9% CO2) the increase in coronary flow was associated with only a small and transient rise in cardiac adenosine release, and did not influence the formation of 3H-adenosine. In the unpaced heart, acetylcholine (10(-7) and 2 X 10(-6) M) dose-dependently increased coronary flow and the release of both adenosine and 3H-adenosine. Within the first minute, the RSA of adenosine was increased, but thereafter was decreased relative to control. In the paced heart, the effects of acetylcholine (2 X 10(-6) M) were greatly attenuated. Increasing coronary flow by bradykinin and isosorbide dinitrate or decreasing heart rate by (-)N6-phenylisopropyl-adenosine did not significantly affect effluent perfusate concentration of adenosine or its RSA.(ABSTRACT TRUNCATED AT 250 WORDS)