Although carbon monoxide (CO) has traditionally been viewed as a toxic gas, increasing evidence suggests that it plays an important homeostatic and cytoprotective role. Its therapeutic use, however, is limited by the side effects associated with CO inhalation. Recently, transition metal carbonyls have been shown to be a safe and effective means of transporting and releasing CO groups in vivo. The goal of the present study was to test whether a water-soluble CO-releasing molecule, tricarbonylchloro(glycinato) ruthenium (II) (CORM-3), reduces infarct size in vivo when given in a clinically relevant manner, i.e., at the time of reperfusion. Mice were subjected to a 30-min coronary artery occlusion followed by 24 h of reperfusion and were given either CORM-3 (3.54 mg/kg as a 60-min intravenous infusion starting 5 min before reperfusion) or equivalent doses of inactive CORM-3, which does not release CO. CORM-3 had no effect on arterial blood pressure or heart rate. The region at risk did not differ in control and treated mice (44.5 +/- 3.5% vs. 36.5 +/- 1.6% of the left ventricle, respectively). However, infarct size was significantly smaller in treated mice [25.8 +/- 4.9% of the region at risk (n = 13) vs. 47.7 +/- 3.8% (n = 14), P < 0.05]. CORM-3 did not increase carboxyhemoglobin levels in the blood. These results suggest that a novel class of drugs, CO-releasing molecules, can be useful to limit myocardial ischemia-reperfusion injury in vivo.