Methotrexate, an anti-rheumatic agent, has recently been reported to show an anti-inflammatory action via ecto-5'-nucleotidase- and adenosine-dependent mechanisms. Because ecto-5'-nucleotidase contributes to the production of adenosine and adenosine has a potent cardioprotective effect against ischemia/reperfusion injury, we investigated whether methotrexate or MX-68 [N-1-((2,4-diamino-6-pteridinyl) methyl)-3,4-dihydro-2H-1,4-benzothiazine-7- carbonyl]-N-2- aminoadipic acid] could reduce infarct size via adenosine-dependent mechanisms. In beagle dogs, the left anterior descending coronary artery was perfused through a bypass tube, which was occluded for 90 minutes followed by 6 hours of reperfusion. The size of infarcts was assessed by TTC staining. MX-68 reduced infarct size compared with that in untreated dogs (13.7 +/- 1.9 versus 38.6 +/- 5.3%, P < 0.01). This effect was completely blunted by either the adenosine receptor antagonist 8-sulfophenyltheophylline (8-SPT) (45.0 +/- 4.6% and 46.8 +/- 5.8% in the 8-SPT and MX-68 + 8-SPT groups, respectively) or by the ecto-5'-nucleotidase inhibitoralpha,beta-methylenadenosine 5'-diphosphate (AMP-CP) (44.0 +/- 4.5% and 46.7 +/- 5.8% in the AMP-CP and MX-68 + AMP-CP groups, respectively). Methotrexate also reduced infarct size to a level comparable with that in the MX-68 group, and its effect was also blunted by 8-SPT. There were no significant differences of collateral blood flow or risk area between the groups. We conclude that methotrexate and its derivative (MX-68) both limit infarct size via adenosine-dependent mechanisms.