Recent data indicate that acute alcohol exposure can have a preconditioning-like protective effect on the heart. We investigated the effect of ethanol exposure shortly before regional ischemia in an infarct model. Both in the open-chest rabbit and in the isolated rabbit heart, exposure of the heart to ethanol significantly reduced infarct size, but only if the alcohol were washed out or sufficiently metabolized before the onset of ischemia. If ethanol were still present during ischemia, it could not only prevent its own protective effect, but also abolish protection induced by ischemic preconditioning or the mitochondrial K(ATP) channel activator diazoxide. In the in vitro model, we tested for possible mediators of ethanol-induced protection and made comparisons to the signaling cascade of ischemic preconditioning. Neither adenosine receptor blockade with 8-(p-sulfophenyl) theophylline, scavenging of free radicals with N-2-mercaptopropionyl glycine, nor closure of K(ATP) channels with glibenclamide affected ethanol's protective effect. However, either a PKC inhibitor or a protein tyrosine kinase inhibitor could completely block ethanol-induced infarct size reduction. Both the protective and anti-protective effects of ethanol had a threshold of about 5 mM. Thus, ethanol-induced protection is mediated by protein kinase C and at least one protein tyrosine kinase, but, in contrast to ischemic preconditioning, is not triggered by either adenosine receptors, free radicals, or K(ATP) channels. Ethanol can only exert its protective effect if it is removed before the onset of ischemia. If still present during ischemia, ethanol has the opposite effect, and inhibits preconditioning by an as yet unidentified mechanism.