Prenatal patency of the ductus arteriosus is maintained mainly by prostaglandin (PG) E(2). Accordingly, the vessel is endowed in its muscular component with a complete, cyclooxygenase (COX) and PGE synthase (PGES), system for the synthesis of the compound. COX1 is better expressed than COX2, particularly in the premature, but COX2 is more extensively coupled with microsomal PGES (mPGES). No evidence was obtained of either COX being coupled with cytosolic PGES (cPGES). Functionally, these data translate into a differential constrictor response of the ductus to dual, COX1/COX2, vs. COX2-specific inhibitors (indomethacin vs. L-745,337), with the latter being less effective specifically prior to term. This difference, however, subsides upon treatment with endotoxin and the attendant upregulation of COX2 and mPGES. Furthermore, when studied separately, COX1 and COX2 prove to be unevenly responsive to indomethacin, and an immediate and fast developing contraction of the vessel occurs only when COX2 is inhibited. Deletion of either COX gene results into upregulation of NO synthase, and a similar compensatory reaction is expected when enzymes are suppressed pharmacologically. We conclude that PGE(2) and NO can function synergistically in keeping the ductus patent. This arrangement provides a possible explanation for failures of indomethacin or ibuprofen treatment in the management of the prematurely born infant with persistent ductus. Coincidentally, it opens the way to new therapeutic possibilities being based on interference with the NO effector or a more selective disruption, possibly having mPGES as a target, of the PGE(2) synthetic cascade.