It has been proposed that α-adrenoceptor vasoconstriction in coronary resistance vessels results not from α-adrenoceptors on coronary smooth muscle but from α-adrenoceptors on cardiac myocytes that stimulate endothelin (ET) release. The present experiments tested the hypothesis that the α-adrenoceptor-mediated coronary vasoconstriction that normally occurs during exercise is due to endothelin. In conscious dogs (n = 10), the endothelin ETA/ETB receptor antagonist tezosentan (1 mg/kg iv) increased coronary venous oxygen tension at rest but not during treadmill exercise. This result indicates that basal endothelin levels produce a coronary vasoconstriction at rest that is not observed during the coronary vasodilation during exercise. In contrast, the α-adrenoceptor antagonist phentolamine increased coronary venous oxygen tension during exercise but not at rest. The difference between the endothelin blockade and α-adrenoceptor blockade results indicates that α-adrenoceptor coronary vasoconstriction during exercise is not due to endothelin. However, in anesthetized dogs, bolus intracoronary injections of the α-adrenoceptor agonist phenylephrine produced reductions in coronary blood flow that were partially antagonized by endothelin receptor blockade with tezosentan. These results are best explained if α-adrenoceptor-induced endothelin release requires high pharmacological concentrations of catecholamines that are not reached during exercise.
- coronary blood flow
coronary blood flow is closely matched to myocardial oxygen consumption via a local metabolic mechanism and through feedforward β-adrenoceptor vasodilation during exercise (6, 8, 9, 21, 31). During exercise, α-adrenoceptor constriction competes with cardiac metabolic demand to limit coronary dilation (1, 8, 10, 12, 14, 24, 26).
Although α-adrenoceptor vasoconstriction has been demonstrated in small coronary arteries in vivo during β-adrenoceptor blockade (11, 15), only β-adrenoceptor vasodilation has been observed in isolated coronary arteries of comparable size (16, 25, 28). One proposed explanation for this paradox is that the α-adrenoceptors producing coronary vasoconstriction are not on vascular smooth muscle but on myocytes instead (5, 30). The α-adrenoceptor-activated myocytes then either release the potent coronary vasoconstrictor endothelin (ET) or some other factor that stimulates endothelin release from endothelial cells. In support of this hypothesis, isolated cardiac myocytes exposed to the α-adrenoceptor agonist phenylephrine release a substance that causes constriction of isolated coronary arterioles (30). The vasoconstriction could be blocked by pretreating either the isolated arterioles with an ETA receptor antagonist or the isolated myocytes with an α1-adrenoceptor antagonist. In anesthetized dogs, intracoronary phenylephrine or norepinephrine infusion during β-adrenoceptor blockade results in a constriction of coronary arterioles that can be blocked by an ETA receptor antagonist or by prazosin, an α1-adrenoceptor antagonist (5).
The present study tests the hypothesis that α-adrenoceptor-mediated vasoconstriction is due to endothelin under the physiological conditions of exercise. The present results indicate that some of the vasoconstriction produced by intracoronary injections of the α-adrenoceptor agonist phenylephrine is due to endothelin. However, endothelin does not contribute to α-adrenoceptor-mediated coronary vasoconstriction during exercise.
Acute closed-chest coronary perfusion preparation.
All animal experiments were conducted according to protocols approved by the Institutional Animal Care and Use Committee of the University of Washington. Male dogs (25–35 kg) were anesthetized with morphine (80 mg sc) followed by α-chloralose (100 mg/kg iv). Additional morphine (40 mg im) was given after 1 h, and chloralose (500 mg iv) was given approximately each hour. The animals were ventilated and given intravenous NaHCO3 to maintain blood gases within normal limits. Aortic pressure was measured via a fluid-filled catheter introduced via the right femoral artery. A stainless steel cannula was advanced via the right carotid artery and wedged into the circumflex coronary artery. The circumflex coronary artery was perfused with heparinized (750 U/kg iv) blood from the left femoral artery. A servo-controlled pump (22) maintained circumflex coronary artery pressure constant at 100 mmHg while circumflex blood flow was measured with an in-line ultrasonic transit time flow transducer (Transonics; Ithaca, NY). Intracoronary bolus drug injections were made just proximal to the cannula lumen. Cardiovascular variables were digitized and recorded continuously using Windaq software (Akron, OH). At the conclusion of experiments, the circumflex artery perfusion territory was stained with intracoronary crystal violet dye just before death with additional anesthesia followed by intravenous KCl. The stained myocardium was cut out and weighed. Coronary blood flow and myocardial oxygen consumption were normalized per gram of perfused myocardium.
Intracoronary endothelin dose-response experiments.
The purpose of these experiments was to confirm the effectiveness of tezosentan (Ro 61-0612) as an endothelin antagonist (4). Tezosentan (1 mg/kg) was dissolved in 20 ml isotonic saline and administered intravenously. Experiments were done in the acute closed-chest preparation. Intracoronary endothelin injections (300 μl) were made at doses of 0.1–10 μg in control experiments and 0.1–30 μg in tezosentan-treated animals. The response to each injection was defined as the lowest coronary blood flow observed before partial recovery began. Because of irreversible coronary vasoconstriction and ventricular fibrillation at the highest endothelin dose, each animal was used for only one endothelin dose-response experiment, either a control (n = 3) or a tezosentan (n = 3) experiment. At the conclusion of these experiments, circumflex arterial flow was nearly zero and the perfusion territory could not be dyed. In this case, the circumflex territory weight was estimated from total heart weight using data from previous studies (7).
Intracoronary phenylephrine dose-response experiments.
These experiments tested the hypothesis that coronary α-adrenoceptor vasoconstriction results from endothelin release. The acute closed-chest coronary perfusion preparation was used. Intracoronary bolus injections of the α-adrenoceptor agonist phenylephrine (100 μl) were used to elicit α-adrenoceptor vasoconstriction. To eliminate any potential β-adrenoceptor effects that could cause either metabolic or feedforward coronary vasodilation, dogs received propranolol (3 mg/kg iv) before phenylephrine injections. Animals also received atropine (0.25 mg/kg iv) to prevent baroreceptor reflex-induced decreases in heart rate secondary to the hypertensive effect of phenylephrine. Two phenylephrine dose-response experiments were done in each dog. In the experimental dogs (n = 4), the second dose-response experiment was preceded by the administration of tezosentan (1 mg/kg iv). In the time-control dogs (n = 4), the tezosentan vehicle (20 ml saline) was given instead. For each phenylephrine injection, baseline circumflex flow was averaged for 15 s just before injection. The response to phenylephrine was defined as the maximum decrease in circumflex flow from baseline, which occurred within 30 s of injection. Circumflex flow was allowed to return to baseline after each injection. Phenylephrine doses >3 μg/kg were not used because systemic vasoconstriction caused long-lasting increases in arterial pressure and thus increases in coronary blood flow due to augmented myocardial oxygen consumption.
The experiments tested the hypothesis that endothelin causes the α-adrenoceptor-mediated coronary vasoconstriction previously observed during exercise. Male dogs (n = 10) underwent sterile surgery as described previously (32). Catheters were placed in the aorta and coronary sinus, and a perivascular flow transducer (Transonics) was placed around the circumflex coronary artery. A splenectomy was also performed to limit hematocrit changes during exercise. After at least 10 days recovery, dogs were studied at rest standing in a sling and during treadmill exercise. While the dogs rested in the sling, arterial and coronary venous blood samples were drawn for measurement of blood gases, oxygen content (TotalO2X, Hospex; Chestnut Hill, MA), and lactate as previously described (8, 32). Coronary blood flow, heart rate, and arterial pressure were measured continuously. While remaining in the sling, dogs then received an intravenous infusion of either vehicle (20 ml saline) or tezosentan (1 mg/kg). Five minutes after the completion of the infusion, postdrug baseline data were collected. Hemodynamic data and blood samples were subsequently collected during three exercise levels (3 mph, 5% grade; 4 mph, 10% grade; 5 mph, 15% grade). Blood samples were drawn while hemodynamic variables were stable at each exercise level. The exercise periods lasted ∼2 min, and the animals were allowed to rest between exercise runs. All 10 dogs underwent both a control (vehicle) experiment and an endothelin blockade (tezosentan) experiment.
Tezosentan (Ro 61–0612) (1 mg/kg, iv) was obtained from Actelion (Allschwil, Switzerland). Propranolol (3 mg/kg iv) and atropine sulfate (0.25 mg/kg iv) were purchased from Sigma. Phenylephrine was obtained from American Regent Laboratories (Shirley, NY). All of the above drugs were dissolved or diluted in isotonic saline. ET-1 (Sigma) was dissolved in 50% methanol in saline at a concentration of 100 μg/ml, and serial dilutions were made in isotonic saline.
Phenylephrine and endothelin dose-response curves before and after tezosentan were analyzed using GraphPad Prism software (San Diego, CA). Control and tezosentan responses were fit separately to a four-parameter Hill equation using the constraints that the minimum values, the maximum values, and the Hill exponent were the same for both curves. The calculated ED50 was the only unconstrained variable. The F-test for nonlinear regressions (23) was then used to determine whether the calculated ED50 was significantly different after tezosentan.
In exercise experiments, control and tezosentan groups were compared using a two-way ANOVA with 10 “blocks” (dogs) and 10 “treatments” (drug-exercise level combinations). When treatment means differed at the P < 0.05 level, the Student-Newman-Keuls test was used to test all paired comparisons between means. In this study, it is only the effects of tezosentan at each exercise level that are of interest, and these are the only comparisons reported. To compare the slopes of the coronary sinus oxygen tension versus myocardial oxygen consumption plots, slopes were determined with the least-squares method for each dog after vehicle or tezosentan treatment. Resting values before vehicle or tezosentan administration were not included. The vehicle and tezosentan slopes were then compared using the paired t-test. The average slopes were plotted, and the linear regression lines were centered at the mean oxygen consumption and mean venous oxygen tension of the data set being graphed.
Figure 1 shows the results of the intracoronary endothelin dose-response experiments in the presence and absence of endothelin blockade with tezosentan. The ED50 was approximately sixfold higher after tezosentan (1.0 μg control vs. 5.9 μg after tezosentan, P < 0.001), demonstrating the effectiveness of tezosentan.
The results of the intracoronary phenylephrine dose-response experiments are shown in Fig. 2. Figure 2A demonstrates that tezosentan shifted the phenylephrine dose-response curve modestly rightward. The calculated control ED50 was 0.36 versus 0.90 μg/kg after endothelin blockade (P < 0.01). In time control experiments (Fig. 2B), there was no significant shift in the calculated ED50 (0.48 μg/kg in control vs. 0.27 μg/kg after tezosentan vehicle), demonstrating that the tezosentan effect was not due to the passage of time.
The results of the exercise experiments are presented in Table 1. As expected, exercise led to graded increases in coronary blood flow, myocardial oxygen consumption, and heart rate. The P values for changes due to exercise are not indicated in Table 1 to concentrate on the effects of endothelin blockade. The only hemodynamic or metabolic variables significantly affected by endothelin blockade were coronary venous oxygen tension and oxygen content under resting conditions. Both were increased whether compared with the separate vehicle experiments in the same dogs or to the pre-tezosentan values in the tezosentan experiments.
A sensitive way to elucidate vasodilator or vasoconstrictor influences on coronary blood flow is a plot of coronary sinus oxygen tension versus myocardial oxygen consumption. Figure 3A shows such a plot during exercise in both control and ETA/ETB receptor blockade experiments. Exercise under control conditions resulted in a moderate decrease in coronary sinus oxygen tension with increasing oxygen consumption, in agreement with previous studies in dogs (1, 8, 12, 31, 32). After endothelin receptor blockade, the slopes were not significantly different. Data from a previous study in this laboratory (Fig. 3B) demonstrate that α-adrenoceptor blockade with phentolamine significantly flattens the coronary venous oxygen tension line during exercise (8). A comparison of Fig. 3, A and B, demonstrates that endothelin receptor blockade does not resemble α-adrenoceptor blockade during exercise.
The primary goal of the present experiments was to test the hypothesis that α-adrenoceptor coronary vasoconstriction during exercise is mediated by endothelin. α-Adrenoceptor-mediated coronary vasoconstriction increases with exercise intensity in dogs (1, 8, 12, 14, 24, 31). If the endothelin hypothesis is true, then there should be more endothelin vasoconstriction during exercise than at rest. The slope of the coronary sinus oxygen tension versus myocardial oxygen consumption plot (Fig. 3A) should then be less steep after endothelin blockade, as is the case after α-adrenoceptor blockade (Fig. 3B). The results therefore do not support the endothelin hypothesis of α-adrenoceptor coronary vasoconstriction during exercise.
Nevertheless, the results of the intracoronary phenylephrine injections in the present study do support the hypothesis that α-adrenoceptor coronary vasoconstriction can be partly due to endothelin. Phenylephrine vasoconstriction was reduced after endothelin blockade with tezosentan (Fig. 2). How can the negative findings with exercise and the positive findings with intracoronary phenylephrine be reconciled? Because exercise results in norepinephrine rather than phenylephrine release, the different catecholamines could account for this discrepancy. However, DeFily et al. (5) elicited long-lasting arteriolar vasoconstriction with intracoronary norepinephrine infusion in the presence of β- and α2-adrenoceptor blockade. This response was similar to the phenylephrine response and was presumably due to endothelin release. A previous study (9) from this laboratory estimated that the myocardial interstitial norepinephrine concentration is ∼12 nM during strenuous exercise in dogs. This is roughly 100-fold lower than the plasma concentration of norepinephrine used to elicit endothelin-mediated α-adrenoceptor coronary vasoconstriction (5). There is no apparent connection between α-adrenoceptor coronary vasoconstriction and endothelin during exercise probably because exercise catecholamine concentrations are simply too low. Under pathophysiological conditions such as ischemia, myocardial interstitial norepinephrine concentrations reach ∼0.5 μM (17), which might be enough to activate endothelin release.
Even if catecholamine concentrations in exercise are too low to stimulate endothelin release via an α-adrenoceptor mechanism, spontaneous endothelin release may cause coronary vasoconstriction. The ET receptor blockade results indicate that there is endothelin-mediated coronary vasoconstriction at rest (Table 1 and Fig. 3A) as previously observed in dogs (29) and pigs (19, 20). Measurements of plasma endothelin concentration have failed to detect endothelin release across the coronary circulation at rest or during exercise (20, 29). The resting coronary vasoconstriction uncovered by tezosentan or other endothelin antagonists with no detectable arteriovenous concentration gradient may reflect endothelin release from the abluminal surface of endothelial cells.
On the basis of a steeper coronary venous oxygen tension versus myocardial oxygen consumption line after endothelin blockade than during untreated control conditions, it has been suggested that endothelin production decreases during exercise and that this decrease contributes to normal exercise coronary vasodilation (19, 20). The steepening of the relationship was not confirmed in the present study (Fig. 3) and was not observed by Takamura et al. (29). However, even with a constant endothelin production rate, the small vasoconstrictor effect at rest would become concealed by the large coronary flow during exercise. Endothelial shear stress-induced nitric oxide vasodilation during exercise may also directly counteract the effect of a constant endothelin concentration (18, 27). Thus the observation that endothelin blockade produces smaller effects during exercise than at rest does not necessarily mean that the coronary endothelin production rate decreases during exercise.
Thus far only α1-adrenoceptors, as opposed to α2-adrenoceptors, have been implicated as stimulating endothelin release. The effects of endothelin blockade in Fig. 3A have been compared with those of the combined α1- and α2-antagonist phentolamine (Fig. 3B). Phentolamine blockade of α2- adrenoceptors may accentuate the change in slope of the coronary sinus oxygen tension versus myocardial oxygen consumption relationship for two reasons. First, α2-adrenoceptor blockade increases norepinephrine release (13) and may therefore increase feedforward β-adrenoceptor vasodilation (6, 8, 9, 21). Second, a significant amount of coronary α-adrenoceptor vasoconstriction may be mediated by α2-adrenoceptors in dogs (3, 11, 15). Thus it may be unrealistic to expect endothelin blockade to reproduce all of the slope change versus control seen in Fig. 3B. Nevertheless, selective α1-adrenoceptor blockade with prazosin during exercise produces results similar to phentolamine (1, 2). Thus it is unlikely that the negative result with endothelin blockade is due to a comparison with phentolamine rather than prazosin.
Although an endothelin receptor antagonist reduced α-adrenoceptor coronary vasoconstriction due to phenylephrine in the present study, the magnitude of the reduction was relatively small compared with earlier studies in which arteriolar diameter was measured (5, 30). The measurement of coronary blood flow in the present study as opposed to arteriolar diameter may account for some of this difference. Another factor may be the length of exposure to α-adrenoceptor agonists. The peak phenylephrine concentrations produced in the present study via bolus intracoronary injection are probably similar to those of Tiefenbacher et al. (30) and DeFily et al. (5) but were transient. DeFily et al. (5) infused intracoronary phenylephrine for 15 min, whereas Tiefenbacher et al. (30) exposed isolated myocytes to phenylephrine for 20 min. endothelin production may therefore increase with longer exposure to α-adrenoceptor agonists.
In conclusion, spontaneous endothelin release in dogs causes coronary vasoconstriction at rest but not during exercise. α-Adrenoceptor-mediated coronary vasoconstriction during exercise is not due to endothelin release. However, endothelin is partially responsible for α-adrenoceptor coronary vasoconstriction after intracoronary phenylephrine injections. These results are best explained if α-adrenoceptor-induced endothelin release requires high pharmacological concentrations of catecholamines that are not reached during exercise.
This study was supported by National Heart, Lung, and Blood Institute Grants HL-49822 and HL-07828.
Dr. Jean-Paul Clozel of Actelion Pharmaceuticals generously provided tezosentan.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 by the American Physiological Society