1. The antimigraine drug, sumatriptan, contracts the human coronary artery and, in some patients, elicits chest symptoms (e.g. pressure and pain), particularly after subcutaneous administration. We studied the effects of the thromboxane A2 (TxA2) analogue, U46619 and endothelin-1 on contractile responses to sumatriptan in the human isolated coronary artery as well as the role of endogenously produced TxA2 and endothelin-1 in contractions evoked by sumatriptan. 2. In the presence of U46619 (1 and 3 nM), mean concentration-response curves to sumatriptan in the human coronary artery were shifted vertically due to the initial contraction by U46619, but when this initial contraction was subtracted from the response to sumatriptan, no significant augmentation was observed. However, analysis of the degree of augmentation in individual arterial segments revealed that the augmentation was variable and related inversely to the Emax of sumatriptan in the absence of U46619 (r = 0.78 and 0.81 for 1 and 3 nM, respectively; P < 0.05). 3. Treatment with the TxA2 receptor antagonist, SQ30741 (100 nM), or incubation of vessel segments with aspirin (10 microM), significantly reduced responses to sumatriptan; in aspirin-treated vessel segments, SQ30741 failed to decrease further the contractions to sumatriptan. The decrease in Emax of sumatriptan by both SQ30741 and aspirin correlated significantly with the Emax of sumatriptan without SQ30741 (r = 0.74; P < 0.01) or aspirin (r = 0.94; P < 0.01). In aspirin-treated vessel segments, responses to sumatriptan were significantly augmented in the presence of U46619 (3 nM; P < 0.05). 4. The specificity of SQ30741 was demonstrated by its ability to antagonize coronary artery contractions to U46619 (pA2: 7.54 +/- 0.30), but not endothelin-1. Similarly, incubation with aspirin (10 microM) did not affect contractile responses to endothelin-1, but significantly reduced TxA2 production in coronary artery segments as judged by a decrease in thromboxane B2 (TxB2) from 4.77 +/- 0.98 to 1.38 +/- 0.36 ng g-1 2 h-1. 5. Endothelin-1 (1 nM) did not significantly augment contractions to sumatriptan; there was also no relationship between the degree of augmentation and the control Emax of sumatriptan in the absence of endothelin-1. Furthermore, unlike SQ30741 or aspirin, a high concentration (100 nM) of the non-selective ETA/ETB receptor antagonist, SB 209670, failed to affect contractile responses to sumatriptan. However, SB 209670 potently antagonized coronary artery contractions induced by endothelin-1 with a pA2 of 8.84 +/- 0.32. 6. Compared to control vascular segments, endothelial denudation did not reduce TxA2 production (with endothelium = 2.56 +/- 1.38 vs. without endothelium = 12.32 +/- 4.94 ng TxB2 g-1 2 h-1), suggesting that the production of TxA2 is not confined to the endothelium. The sumatriptan-induced contractions were also unaffected by endothelial denudation. 7. The results of the present study suggest that endogenously produced TxA2 enhances contractions to sumatriptan in the human isolated coronary artery. Such a mechanism may play a role in causing chest symptoms after sumatriptan by potentiating coronary vascular contraction by sumatriptan in vivo.