In the present study, the ability of intact cardiac valvular endothelial cells to release vasodilatory prostanoids and endothelium-derived relaxing factor was investigated. Endothelium-denuded canine coronary arteries were used for bioassay and contractile force recording. Insertion of small segments of cardiac valve (20-30 mm2) with intact endothelium into endothelium-denuded coronary arterial rings did not markedly alter the sensitivity nor magnitude of the coronary artery contractile response to KCl. In contrast, the prostaglandin F2 alpha (PGF2 alpha)-induced contraction was significantly depressed (70% decrease in magnitude and 216% increase in ED50), compared with contraction in the absence of valvular endothelium (5.52 +/- 0.49 g and ED50 of 1.18 +/- 0.02 microM, respectively). These alterations in PGF2 alpha-induced contractions were reduced to 38% decrease in magnitude and +66% in ED50 in the presence of 5 microM indomethacin. Addition of acetylcholine (0.1-30 microM) into these endothelium-denuded coronary artery/valve preparations resulted in a dose-dependent relaxation, reaching a maximum of -59.9 +/- 1.6% (mean +/- SEM of seven vessels). Preincubation of valvular endothelium with 5 microM indomethacin also reduced these acetylcholine-induced valvular endothelium-dependent relaxations to 40.4 +/- 5.5% (mean +/- SEM of 13 vessels). Addition of hemoglobin (3 microM) further attenuated relaxation to -16.0 +/- 7.7% (mean +/- SEM of 14 vessels), while superoxide dismutase (20 units/ml) potentiated the relaxant response to -81.3 +/- 9.4% (mean +/- SEM of 11 vessels) in the presence of indomethacin. These findings suggest that there is a continuous basal release of vasodilatory prostanoids and endothelium-derived relaxing factor from the valvular endothelium, which can be further stimulated with acetylcholine and superoxide dismutase, and inhibited by indomethacin and hemoglobin.