Studies in animals have shown that serotonin constricts coronary arteries if the endothelium is damaged, but in vitro studies have revealed a vasodilating effect on isolated coronary segments with an intact endothelium. To investigate the effect of serotonin in humans, we studied coronary-artery cross-sectional area and blood flow before and after the infusion of serotonin in seven patients with angiographically normal coronary arteries and in seven with coronary artery disease. We measured the cross-sectional area of the coronary artery by quantitative angiography and coronary blood flow with an intracoronary Doppler catheter. Measurements were obtained at base line and during intracoronary infusions of serotonin (0.1, 1, and 10 μg per kilogram of body weight per minute, for two minutes). We repeated the measurements after an infusion of ketanserin, an antagonist of serotonin receptors that is thought to block the effect of serotonin on receptors in the arterial wall but not in the endothelium. In patients with normal coronary arteries, the highest dose of serotonin increased cross-sectional area by 52 percent (P<0.001) and blood flow by 58 percent (P<0.01). The effect was significantly potentiated by administration of ketanserin. In patients with coronary-artery atherosclerosis, serotonin reduced cross-sectional area by 64 percent (P<0.001) and blood flow by 59 percent (P<0.001). Ketanserin prevented this effect. Serotonin has a vasodilating effect on normal human coronary arteries; when the endothelium is damaged, as in coronary artery disease, serotonin has a direct, unopposed vasoconstricting effect. When considered with other evidence, these data suggest that platelet-derived factors such as serotonin may have a role in certain acute coronary ischemic syndromes. (N Engl J Med 1991; 324:641–8.)

We have speculated previously that the abrupt conversion from chronic stable to unstable angina and the continuum to acute myocardial infarction may result from myocardial ischemia caused by progressive platelet aggregation and dynamic vasoconstriction themselves caused by local increases in thromboxane and serotonin at sites of coronary artery stenosis and endothelial injury. Platelet aggregation and dynamic coronary artery vasoconstriction probably result from the local accumulation of thromboxane and serotonin and also relative decreases in the local concentrations of endothelially derived vasodilators and inhibitors of platelet aggregation, such as endothelium-derived relaxing factor (EDRF) and prostacyclin. With severe reductions in coronary blood flow caused by these mechanisms, platelet aggregates may increase, and an occlusive thrombus composed of platelets and white and red blood cells in a fibrin mesh may develop. When coronary arteries are occluded or narrowed for a sufficient period of time by these mechanisms, myocardial necrosis, electrical instability, or sudden death may occur. We believe that unstable angina and acute myocardial infarction are a continuum in relation to the process of coronary artery thrombosis and vasoconstriction. When the period of platelet aggregation or dynamic vasoconstriction at sites of endothelial injury and coronary artery stenosis is brief, unstable angina or non-Q wave infarction may occur. However, when the coronary artery obstruction by these mechanisms is prolonged for several hours, Q wave myocardial infarction results. Chronic endothelial injury and coronary artery stenosis are probably associated with the accumulation of platelets, white and red blood cells, and a fibrin mesh at the site of stenosis and endothelial injury.

Cyclic variations in coronary blood flow (CFVs) in dogs with experimental coronary artery stenosis and endothelial injury appear to result primarily from the aggregation of platelets at the site of stenosis followed by dislodgement and distal embolization. Using this canine model, we tested the hypotheses: (a) that thrombin is an important mediator of CFVs in dogs with coronary stenoses and endothelial injury; (b) that inhibition of thrombin with heparin, or MCI-9038, a selective thrombin inhibitor, abolishes CFVs in this model; and (c) that abolition of CFVs by thrombin inhibition is time dependent. CFVs, produced in open-chest dogs by placing a flow-reducing plastic constrictor around the left anterior coronary artery, were monitored for either 30 min (group I) or 3 h (group II) before treatment with either heparin or 4-methyl-1-(N2-[(3-methyl-1,2,3,4-tetrahydro-8-quinolinyl (MCI-9038). In group I, cyclic flow variations were abolished by heparin in 12 of 18 dogs and by MCI-9038 in 5 of 7 dogs. In group II, cyclic flow variations were not abolished by heparin in any of seven dogs and were abolished by MCI-9038 in only one of seven dogs. Thus, (a) thrombin appears to be an important mediator of cyclic flow variations in dogs with coronary artery stenosis and endothelial injury and (b) inhibition of thrombin abolishes CFVs after short but not prolonged periods of CFVs.

The goal of the present study was to demonstrate that intracoronary platelet deposition may trigger intense vasoconstriction of large epicardial coronary arteries in vivo and that this is largely mediated by thromboxane A2 and serotonin released by activated platelets. Cyclic flow variations (progressive declines in blood flow followed by sudden restorations of flow) due to recurrent intracoronary platelet activation and thrombus formation were induced by damaging the endothelium and placing a cylindrical constrictor on the left anterior descending coronary artery (LAD) in open-chest, anesthetized dogs. Coronary diameters were measured in vivo by means of ultrasonic crystals sutured on the LAD immediately distal to the constrictor (LAD1) and 1 cm below (LAD2) and on the circumflex coronary artery (Cx). Coronary artery diastolic diameters were measured continuously before and during cyclic flow variations and after they were abolished by administration of LY53857, a serotonin-receptor antagonist (group 1, n = 7), or SQ29548, a thromboxane-receptor antagonist (group 2, n = 7). During cyclic flow variations, at the nadir of coronary flow, LAD1 (a site of maximal platelet accumulation) cross-sectional area decreased by 52 +/- 10% and 38 +/- 6% in group 1 and 2 animals, respectively (p less than 0.001 compared with values recorded during a brief LAD occlusion obtained by a suture snare), whereas LAD2 (a site of minimal or no platelet accumulation) cross-sectional area did not differ from that recorded during the brief LAD occlusion. SQ29548 abolished cyclic flow variations in seven of seven dogs and LY53857 in six of seven, but they affected the increased coronary vasoconstriction differently: LAD1 cross-sectional area increased by 32 +/- 6% of the control value in SQ29548-treated animals, whereas it returned to baseline dimension values in the LY53857-treated group as these interventions also abolished the cyclic flow variations. We conclude that a marked coronary vasoconstriction may be triggered by local platelet deposition and that thromboxane A2 and serotonin are mediators of this vasoconstriction.