OBJECTIVE The vasoconstrictor effect of endothelins (ET) is mediated by endothelin A (ETA) and endothelin B (ETB) receptors. Furthermore, ETB receptor stimulation results in release of vasodilators. Hence, ETA receptor antagonists should attenuate ET-mediated vasoconstriction. The aim of the present study was to evaluate and compare the effects of BQ-123, an ETA receptor antagonist, and bosentan, an ETA and ETB receptor antagonist, on coronary vasomotor tone, left ventricular systolic function and ET-1 efflux in the presence or absence of myocardial ischaemia/reperfusion. METHODS Isolated rat hearts were perfused using a Langendorff preparation. Global ischaemia was induced on average by 68 +/- 2% (+/- standard error of the mean) reduction of a baseline perfusion flow-rate 10 min after introduction of ET antagonists. Thirty minutes of ischaemia was followed by 30 min reperfusion. ET-1 efflux in coronary perfusate was measured by radioimmunoassay. RESULTS In non-ischaemic hearts (n = 7), BQ-123 (10(-6) M) perfusion induced a progressive decrease in coronary flow-rate compared with control group. This flow reduction persisted after wash-out of BQ-123. In contrast, bosentan (10(-5) M, n = 7) induced no change in perfusion rate. In the absence of ET antagonists (n = 16), there was a 22 +/- 6% post-ischaemic increase in perfusion flow-rate. BQ-123 (n = 5) but not bosentan (n = 12) abolished this post-ischaemic increase in flow-rate. Neither BQ-123 nor bosentan induced significant variation in force of contraction. In ischaemic hearts, ischaemia per se induced a transient decrease in force of contraction. Bosentan significantly (P < 0.05) accentuated and BQ-123 tended to accentuate (P = 0.06) this decrease in force of contraction during ischaemia. Bosentan but not BQ-123 significantly impaired the recovery of systolic function during reperfusion (P < 0.05). Both BQ-123 and bosentan perfusion increased ET-1 efflux rate to 730 +/- 188% and 315 +/- 81% respectively. This effect was abolished during ischaemia for BQ-123, but not for bosentan. CONCLUSIONS In isolated perfused rat hearts, both BQ-123 and bosentan increased ET-1 efflux, but only BQ-123 exerted vasoconstrictor effects. These results thus generated the hypothesis that: (1) ET-1 release within the coronary vascular bed may be physiologically subject to negative feedback regulation mediated via ETA receptors; (2) ETA receptor antagonists increase ET-1 efflux, which may lead to net vasoconstriction via unopposed ETB stimulation. Furthermore, the negative inotropic effects observed during ischaemia suggest that ET is critical to the maintenance of systolic function during ischaemia.