Consistent with previous reports, sphingosine at a high concentration (5 microM) was cardiotoxic as evidenced by increased infarct size in response to ischemia/reperfusion in an ex vivo rat heart. Sphingosine 1-phosphate (S1P) at 5 microM was cardioprotective. However, at a physiologic concentration (0.4 microM) sphingosine as well as S1P was effective in protecting the heart from ischemia/reperfusion injury both when perfused prior to 40 min of ischemia (preconditioning) or when added to reperfusion media following ischemia (postconditioning). Protection by sphingosine and S1P was evidenced with both pre- and post-conditioning by a >75% recovery of left ventricular developed pressure during reperfusion and a decrease in infarct size from 45% of the risk area to less than 8%. When VPC23019, an S1P(1and3)G-protein coupled receptor antagonist, was added to the preconditioning or postconditioning medium along with S1P, it completely blocked S1P-induced protection. However, VPC 23019 did not affect the ability of 0.4 microM sphingosine to either precondition or postcondition hearts. Studies of preconditioning revealed that inhibition of protein kinase C with GF109203X blocked preconditioning by S1P. However, GF109203X did not affect preconditioning by 0.4 microM sphingosine. Likewise, cotreatment with the PI3 kinase inhibitor wortmanin blocked preconditioning by S1P but not by sphingosine. By contrast, inhibition of protein kinase G with KT5823 had no effect on S1P preconditioning but completely eliminated preconditioning by sphingosine. Also, the protein kinase A inhibitory peptide 14-22 amide blocked preconditioning by sphingosine but not S1P. These data reveal for the first time that sphingosine is not toxic at physiologic concentrations but rather is a potent cardioprotectant that utilizes a completely different mechanism than S1P; one that is independent of G-protein coupled receptors and utilizes cyclic nucleotide-dependent pathways.