We tested the hypothesis that selective adenosine A(3)-receptor stimulation reduces postischemic contractile dysfunction through activation of ATP-sensitive potassium (K(ATP)) channels. Isolated, buffer-perfused rat hearts (n = 8/group) were not drug pretreated (control) or were pretreated with adenosine (20 microM), 2-chloro-N(6)-(3-iodobenzyl)-adenosine-5'-N-methyluronamide (Cl-IB-MECA; A(3) agonist, 100 nM), Cl-IB-MECA + 8-(3-noradamantyl)-1,3-dipropylxanthine (KW-3902; A(1) antagonist, 5 microM), Cl-IB-MECA + glibenclamide (Glib; K(ATP)-channel blocker, 0. 3 microM), or Glib alone for 12 min before 30 min of global normothermic ischemia followed by 2 h of reperfusion. After 2 h of reperfusion, left ventricular developed pressure (LVDP, %baseline) in control hearts was depressed to 34 +/- 2%. In hearts pretreated with Cl-IB-MECA, there was a statistically significant increase in LVDP (50 +/- 6%), which was reversed with coadministration of Glib (37 +/- 1%). Control hearts also showed similar decreases in left ventricular peak positive rate of change in pressure (dP/dt). Therefore, the A(3) agonist significantly attenuated postischemic cardiodynamic injury compared with the control, which was reversed by Glib. Cumulative creatine kinase (CK in U/min) activity was most pronounced in the control group (10.4 +/- 0.6) and was significantly decreased by Cl-IB-MECA (7.5 +/- 0.4), which was reversed by coadministration of Glib (9.4 +/- 0.2). Coronary flow was increased during adenosine infusion (160% of baseline) but not during Cl-IB-MECA infusion. Effects of Cl-IB-MECA were not reversed by the specific A(1) antagonist KW-3902. We conclude that cardioprotection afforded by A(3)-receptor stimulation may be mediated in part by K(ATP) channels. Cl-IB-MECA may be an effective pretreatment agent that attenuates postischemic cardiodynamic dysfunction and CK release without the vasodilator liability of other adenosine agonists.