This study aimed at characterizing the influence of endothelium on noradrenaline release from the canine pulmonary artery. Tritium overflow from intact or endothelium-free Vessels preloaded with 0.2 mu mol.1(-1 3)H-noradrenaline was evoked by electrical stimulation (1 Hz during 5 min) or potassium (25-100 mmol.1(-1)). The fractional release of tritium evoked by electrical stimulation was increased by removing the endothelium [from 1.7 (1.2; 2.4) to 2.7(2.3; 3.2)x 10(-5).pulse(-1), n = 10; P < 0.05]. Neither N-G-nitro-L-arginine methyl ester (L-NAME) (up to 300 mu mol.l(-1)) nor indomethacin (up to 30 mu mol.l(-1)), nor endothelin-1 (up to 30 nmol.l(-1)), nor suramin (up to 300 mu mol.l(-1)) changed tritium release evoked by electrical stimulation. In contrast, the selective A(1)-adenosine antagonist 1,3-dipropyl-8-cyclopenlylxanthine (DPCPX) (3.3-33 nmol.l(-1)) concentration-dependently increased, and the selective A(1)-adenosine agonist N-6-cyclopentyladenosine (CPA) (3.3-100 nmol.l(-1)) concentration-dependently decreased the evoked release of noradrenaline. Since the effects of DPCPX were observed in endothelium-intact tissues only, it may be concluded that adenosine secreted by the endothelium activates prejunctional release-inhibiting A(1)-receptors. Tetraethplammonium (TEA) (3.3-33 mmol.l(-1)) enhanced tritium overflow evoked by electrical stimulation more in endothelium-free than in endothelium-intact vessels, indicating that some K+-channel opener is involved in the inhibitory role of endothelium on noradrenaline release. Since it had been previously shown that A(1)-adenosine receptors are coupled to K+-channels, it is suggested that adenosine may inhibit noradrenaline release through the opening of K+-channels. In conclusion. the results show that in the canine pulmonary artery, adenosine is a good candidate for the endothelium-dependent inhibitory factor which is responsible for the reduction of noradrenaline release evoked by electrical stimulation.