The aim of this study was to determine whether endogenous adenosine has antiarrhythmic effects on ischemia-induced ventricular tachyarrhythmias. We therefore modulated the effect of endogenous adenosine in isolated rat hearts using four different approaches. First, interstitial adenosine was elevated by metabolic inhibition with either EHNA (erythro-9-(2-hydroxy-3-nonly)adenine) or acadesine [5-amino-1-beta-D-imidazole-4-carboxamide). Second, cardiac effects of A1 adenosine receptors were allosterically enhanced with PD81,723 (2-amino-4,5-dimethyl-3-thienyl)[3-(trifluoromethyl)phenyl]-methanone . Third, endogenous adenosine release was suppressed with NBMPR (S-(4-nitrobenzyl)-6-thioinosine), and fourth, adenosine receptor subtypes were blocked with antagonists of different selectivity. Regional ischemia, induced by coronary artery ligation, caused ventricular fibrillation of a reproducible kind in about 20% of untreated hearts with a low calcium concentration in the perfusion medium (0.80 mmol/l CaCl2) and in about 75% with high calcium (1.85 mmol/l) within an observation period of 30 min. At high calcium, EHNA (1 and 10 micromol/l) and acadesine (500 micromol/l) suppressed the occurrence of ventricular fibrillation from 68% (controls) to 47%, 33% and 38%, respectively. Conversely, PD81,723 (10 micromol/l) did not influence the occurrence of ventricular fibrillation. At low calcium, NBMPR (0.1 and 1 micromol/l) resulted in a concentration-dependent rise of ventricular fibrillation from 13% (controls) to 40% and 57%, respectively. The adenosine receptor antagonists theophylline (100 micromol/l), XAC (Xanthine Amine Congener; 1 micromol/l) and 8-PT (8-phenyltheophylline; 1 micromol/l) caused a rise in the occurrence of ventricular fibrillation from 25%, 15% and 18% (controls) to 57%, 39% and 44%, respectively, and the selective A2a receptors antagonist CSC (8-(3-chlorostyryl)caffeine; 5 micromol/l) from 20% to 56%. Conversely, the selective A1 receptor blocker DPCPX (8-cyclopentyl-1,3-dipropyl-xanthine; 1 micromol/l) was ineffective. NBMPR or EHNA concentration-dependent suppressed or increased ischemia-induced adenosine overflow, respectively, in a concentration-dependent manner, whereas the adenosine receptor antagonists did not influence adenosine overflow. We conclude that endogenous adenosine is an antiarrhythmic mediator accumulating in acute ischemic myocardium to a level which effectively decreases the occurrence of ventricular fibrillation by an A2 adenosine receptor activation in the isolated rat heart.