Prostaglandin H synthase 2 (PGHS-2), a highly inducible isoenzyme, is responsible for overproduction of the prostaglandins (PGs) in inflammatory sites. We established that among fish oil polyunsaturated fatty acids (PUFAs), eicosapentaenoic acid (EPA), but not docosahexaenoic acid (DHA), greatly decreased interleukin-1beta (IL-1beta)-induced PGHS-2 expression in human pulmonary microvascular endothelial cells (HPMECs). Lipoxygenase products 12 (S)-hydroperoxyeicosapentaenoic acid ((S)-HpEPE), 15 (S)-HpEPE and leukotriene (LT) D5 reproduced similar inhibitory effect, suggesting that they may be the intermediate metabolites responsible for PGHS-2 down-regulation by EPA. Accordingly, the EPA effect is prevented by nordihydroguaiaretic acid (NDGA) and by REV 5901, nonspecific and specific 5-lipoxygenase inhibitors, respectively. Besides, inhibition of cyclooxygenase activity by ibuprofen, indomethacin or aspirin was not able to prevent this effect. Moreover, cyclooxygenase metabolites of EPA (PGs D3, E3 and I3) markedly potentiate IL-1beta-induced PGHS-2 expression, probably by increasing intracellular cAMP levels. Peroxisome proliferator-activated receptors (PPARs) are known to be activated by fatty acids (FAs) such as EPA. We found here that HPMECs express only weak amounts of PPARalpha and PPARgamma whose activation by synthetic agonists, Wy-14,643 and ciglitazone, does not cause any inhibition of IL-1beta-induced PGHS-2 expression. This finding ruled out the involvement of PPARs in the EPA inhibitory effect. In addition, we established that EPA, which failed to inhibit nuclear factor-kappaB (NF-kappaB) activation, suppressed p38 mitogen-activated protein kinase (MAPK) phosphorylation in stimulated HPMECs. Our data demonstrate that EPA, unlike DHA, down-regulates PGHS-2 expression in HPMECs probably through its 5-lipoxygenase-dependent metabolites and advocates a beneficial role for this FA in limiting inflammatory response.