AIMSWe have compared the endothelin receptor subtype affinity (K(D)) and selectivity of four structural classes of antagonists (peptide, sulphonamide-based, carboxylic acid-based, myceric acid-based) in human cardiovascular tissues to determine whether these are predicted by values reported for human cloned receptors. Additionally, affinities (K(B)) for these antagonists, determined in ET-1-mediated vasoconstriction assays in human blood vessels, were used to identify discrepancies between K(B) and K(D) determined in the same tissues.MAIN METHODSCompetition binding experiments were carried out in sections of human left ventricle, coronary artery and homogenates of saphenous vein to determine K(D) values for structurally different ET(A)-selective (FR139317, BMS 182874, S97-139, sitaxentan, ambrisentan) and mixed (PD142893, Ro462005, bosentan, L-749329, SB209670) antagonists. Schild-derived values of antagonist affinity were obtained in vascular functional studies.KEY FINDINGSWhen compared with previously reported data in human cloned endothelin receptors, those antagonists reported to be ET(A)-selective exhibited even greater ET(A) selectivity in human ventricle (BMS 182874, sitaxentan, ambrisentan) that expressed both receptor subtypes. Those antagonists reported to have <100 fold selectivity in cloned receptors (PD142893, Ro-462005, bosentan, SB209670, L-749329) did not distinguish between receptor subtypes in human left ventricle. For antagonists where we determined affinity in vascular functional and binding assays (Ro462005, bosentan, BMS 182874, L-749329, SB209670) there was no correlation between the degree of discrepancy in K(B) and K(D) and structural class.SIGNIFICANCEFor an antagonist to retain ET(A)-selectivity in vivo it may be necessary to identify those compounds that have at least 1000 fold ET(A):ET(B) selectivity in in vitro assays.