We examined the relative contribution of endothelial and vascular smooth muscle-derived prostaglandins and endothelium-derived relaxing factor in modulating both the large coronary artery and resistance vessel responses to thromboxane in vivo. Vascular responses to the thromboxane analogue U46619 were measured in four separate experimental protocols: 1) The vascular responses were measured in the presence and absence of intact endothelium to examine the role of endothelium-derived vasodilators. 2) Responses were measured in the presence of intact endothelium before and after inhibition of cyclooxygenase with indomethacin to examine the role of endothelial and vascular smooth muscle-derived prostaglandins. 3) Responses were measured after endothelial removal before and after indomethacin to examine the role of vascular smooth muscle-derived prostaglandins. 4) Responses were measured after indomethacin and before and after removal of endothelium to examine the role of endothelium-derived relaxing factor. In anesthetized dogs (n = 41) that underwent constant pressure perfusion of the left anterior descending coronary artery (LAD), LAD diameter was measured with sonomicrometer crystals, and coronary flow was measured with an electromagnetic flow probe. Intracoronary infusion of U46619 (0.01-1.0 microgram/min) produced a dose-dependent constriction of LAD. Constriction of the LAD was augmented after endothelial removal, after indomethacin treatment in both the presence and absence of endothelium, and after removal of the endothelium in the presence of indomethacin. Inhibition of prostaglandin synthesis had the greatest effect of augmenting constriction of LAD to thromboxane. Coronary flow was decreased by U46619 only in the presence of indomethacin.(ABSTRACT TRUNCATED AT 250 WORDS)