1. Early observations of endothelium-dependent relaxation and endothelium-derived relaxing factor from 1980 to 1987. In 1980, Furchgott and Zawadzki demonstrated that the vascular relaxation induced by ACh was dependent on the presence of the endothelium and provided evidence that this effect was mediated by a labile humoral factor, later known as EDRF. Endothelium-dependent relaxation, which was subsequently demonstrated in many vascular preparations, including some veins, arteries, and microvessels, occurs in response to a variety of substances, such as ACh, adenine nucleotides, thrombin, substance P, the calcium ionophore A23187, and bradykinin. Other stimuli, such as hypoxia, increase in flow, and electrical stimulation, also cause endothelium-dependent relaxation of vascular tissue in vitro. Some agents, however, such as the nitrovasodilators, atrial natriuretic factor, bovine retractor penis inhibitory factor,/3-adrenergic agonists, and prostacyclin, induce vascular relaxation by endothelium-independent mechanisms (for reviews, see Furchgott, 1984; Griffith et a!., 1984; Busse et al., 1985; Moncada et al., 1986b). The humoral nature of EDRF was first demonstrated using a variety of pharmacological preparations in which the biologically active substance was transferred from a donor to a detector bioassay. One such system consisted of a “sandwich” arrangement of two rabbit aortic strips in which the EDRF donor (a strip with intact endothehum) was placed, intima! surface to intimal surface, next to the detector(a strip without endothelium; Furchgott, 1984). Another approach involved perfusion ofthe lumen of an intact rabbit isolated aorta, the effluent of which was used to superfuse endothelium-denuded vascular rings (Griffith et a!., 1984; Rubanyi et al., 1985). Stimulation of the donor aorta with ACh caused relaxation of the detector tissues. Finally, vascular endothelial cells, cultured on microcarriers and packed in the barrel of a syringe or a modified chromatography column, were perfused, and the effluent was used to superfuse a ring of canine coronary artery or a series of rabbit aortic strips denuded of endothelium(Cocks et a!., 1985; Gryglewski et a!., 1986a).

It was established, using techniques such as these, that EDRF was a very short-lived substance with a half-life of only seconds in oxygenated physiological salt solutions (Griffith et a!., 1984; Cocks et a!., 1985). Release of EDRF was observed under basal conditions as well as after stimulation with ACh (Griffith et a!., 1984; Rubanyi et a!., 1985; Martin et a!., 1985). The effects of EDRF were shown to be inhibited by Hb, methylene blue (Martin et al., 1985), and other agents such as dithiothreitol and hydroquinone(Griffith et a!., 1984) and to be mediated by stimulation of the soluble guanylate cyclase with the consequent elevation of intracellular cyclic GMP levels (Rapoport and Murad, 1983). Bioassay studies in which the source of EDRF, either fresh vascular tissue with endothelium(Rubanyi et a!., 1985) or vascular endothelial cells in culture(Cocks et al., 1985; Gryglewski et al., 1986a), was separated from the detector(endothelium-denuded vascular rings or strips) allowed the study of the effects of physical or chemical manipulation on the generation, stability, or actions of EDRF. It was found using such techniques that superoxide anions (O2) contribute to the instability of EDRF, because the effects of EDRF were prolonged by the addition of SOD(Gryglewski et a!., 198Gb; Rubanyi and Vanhoutte, 1986) and inhibited by Fe2 (Gryglewski et a!., 198Gb) and hyperoxia(Rubanyi and Vanhoutte, 1986). Furthermore, a number of compounds described as inhibitors of EDRF were shown to …