INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

CORONARY BLOOD FLOW DURING exercise is closely matched to myocardial metabolism. This linkage is thought to occur largely through the release of metabolic vasodilators by cardiac myocytes, forming a negative feedback control loop that keeps tissue oxygen tension relatively constant (15). It has recently been proposed that exercise-induced coronary vasodilation results, in part, from feedforward -adrenoceptor vasodilation mediated by catecholamines (11, 33). The hypothesis is that the increase in sympathetic activity during exercise stimulates both increased cardiac metabolism and direct coronary vasodilation.

In the dog heart, feedforward adrenoceptor vasodilation has been demonstrated during intracoronary norepinephrine infusion (33). However, intracoronary infusion may expose arterial vascular smooth muscle to higher norepinephrine concentrations than exist during more physiological interventions such as exercise. The direct vascular effects of intracoronary norepinephrine might therefore be exaggerated relative to the metabolic effects. Duncker et al. (11) demonstrated feedforward -adrenoceptor vasodilation in exercising pigs. Unlike humans and dogs, however, pigs do not exhibit -adrenergic coronary vasoconstriction during exercise (11). The goal of the present study was to determine the extent of feedforward sympathetic -adrenoceptor-mediated coronary vasodilation in exercising dogs.

The rationale used in the present study is that there are three main components that control coronary flow during exercise: 1) local metabolic vasodilation, 2) -adrenoceptor-mediated vasoconstriction, and 3) -adrenoceptor-mediated vasodilation. During -adrenoceptor blockade, coronary vasodilation consists of the local metabolic plus -adrenoceptor vasodilator components, whereas combined - and -adrenoceptor blockade produces a relatively "pure" local metabolic coronary vasodilation. Therefore, differences between the -adrenoceptor blockade and  + -adrenoceptor blockade conditions reflect the -adrenoceptor contribution. The feedforward sympathetic vasodilation hypothesis predicts that the myocardial oxygen supply-to-consumption ratio will be reduced when -adrenoceptors are blocked. Compared with -adrenoceptor blockade,  + -adrenoceptor blockade should result in lower coronary venous oxygen tensions at equivalent myocardial oxygen consumptions. In addition, arterial and coronary venous plasma adenosine concentrations were measured to evaluate the role of adenosine as a local metabolic negative-feedback vasodilator before and during adrenergic receptor blockade.

	METHODS AND MATERIALS

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

Surgical preparation. Experiments were performed on 10 adult male mongrel dogs (weight: 23-34 kg) that were taught to run on a motorized treadmill. Preanesthesia (0.05 mg/kg acepromazine and 0.06 mg/kg atropine, subcutaneous) was administered 30 min before intravenous induction of anesthesia with 5.75 mg/kg ketamine and 0.3 mg/kg diazepam. A surgical plane of anesthesia was maintained by mechanical ventilation with 0.5-3.0% isoflurane gas. Utilizing sterile technique, a splenectomy was performed through a midline abdominal incision to minimize changes in hematocrit during exercise. After this procedure, a left lateral thoracotomy was performed in the fifth intercostal space. With the use of a modified Seldinger technique, a polyurethane catheter was implanted into the descending thoracic aorta to measure aortic blood pressure and obtain arterial blood samples. A second polyurethane catheter was placed in the coronary sinus via a purse string suture in the right atrial appendage for coronary venous blood sampling. Catheter materials and dimensions have been described in detail by Tune et al. (49). The circumflex coronary artery was dissected free, and a flow transducer (see Pressure and flow measurement) was placed around the artery. No instruments were implanted in the myocardium, and no surgical stitches were placed in the ventricles to avoid injuring tissue that might release adenosine. A chest tube was placed to evacuate the pneumothorax, and the chest was closed in layers. The catheters and the flow transducer wire were tunneled subcutaneously and exteriorized between the scapulae. Both the abdominal and thoracic incisions were infiltrated with 2.5% bupivacaine, and 0.01 mg/kg Buprenex (Reckitt and Colman) was administered intramuscularly to minimize postoperative pain. Antibiotic (20 mg/kg cefazolin) was administered twice daily for 5 days. The animals received a multivitamin, baby aspirin (81 mg), and a dietary iron supplement (324 mg) daily. A nylon jacket (Alice King Chatham, Hawthorne, CA) was placed on the animals to protect the catheters and the flow transducer wire. Catheters were flushed daily with a high-viscosity solution of 50% glucose containing penicillin G (65,000 U/ml) and heparin (1,300 U/ml). The animals were allowed at least 10 days for recovery before experiments were conducted.

Pressure and flow measurement. A coextruded polyurethane catheter was used in the aorta so that a high-fidelity Mikro-tip catheter pressure transducer (3-Fr; SPR-524, Millar Instruments, Houston, TX) could be inserted at the time of the experiment to measure aortic blood pressure (14, 20). The pressure transducer was introduced into the aortic catheter through a hemostatic control valve (Tuohy-Borst adapter, Mallinckrodt Medical, St. Louis, MO) that allowed arterial blood samples to be withdrawn while maintaining a fluid-tight seal.

Blood sampling. Arterial and coronary venous blood samples were collected simultaneously into heparinized glass syringes that were immediately sealed and placed on ice. The samples were analyzed for hydrogen ion concentration, carbon dioxide tension, and oxygen tension with an pH/blood gas analyzer (model 1306, Instrumentation Laboratories, Waltham, MA). Oxygen content was determined using the fuel-cell method (Total O₂X, Hospex, Chestnut Hill, MA). In addition, a portion from both the arterial and coronary venous blood samples was transferred into NaF-coated vials to prevent glycolysis, and lactate concentration was determined with a YSI lactate analyzer (model 1500, Yellow Springs Instruments, Yellow Springs, OH). Myocardial oxygen consumption (µl O₂ · min¹ · g¹) was calculated by multiplying coronary blood flow per gram of perfused tissue by the arterial-coronary venous difference in oxygen content. Percent myocardial lactate extraction was calculated as the difference in arterial and coronary venous lactate concentration divided by the arterial lactate concentration.

Estimation of cardiac interstitial adenosine concentration. Cardiac interstitial adenosine concentration was estimated using a four-region (plasma, endothelial cell, interstitial space, parenchymal cell), axially distributed mathematical model (29, 30, 46). The model describes the effects of blood flow and adenosine transport and exchange between tissue regions, as well as cellular production and consumption, on the relationship between arterial, venous, and interstitial adenosine concentrations. This model has been previously used to estimate interstitial adenosine concentrations in vivo, and the constraints and assumptions have been described extensively (30, 46). The model accounts for myocardial blood flow heterogeneity and the change in heterogeneity that occurs with changes in blood flow. In addition, the model is constrained with previous estimates of capillary adenosine transport and metabolism adjusted for the level of coronary blood flow. Interstitial adenosine concentration was estimated using the measured values of coronary plasma flow and arterial plasma adenosine concentration by adjusting cellular adenosine production in the model to fit the measured coronary venous plasma adenosine concentration. Interstitial adenosine concentration was calculated at rest and during steady-state exercise conditions, with and without administration of adrenoceptor antagonists.

Experimental protocol. After an overnight fast, the dogs stood in slings while baseline data were collected. Coronary blood flow, aortic pressure, and heart rate were recorded while arterial and coronary venous blood samples were collected for measurement of blood gases, oxygen content, and lactate and adenosine concentrations. While remaining in the sling, the dogs received an intravenous infusion of either vehicle, phentolamine (l mg/kg), or phentolamine plus propranolol (1 mg/kg, each ). Five minutes after completing this infusion, postdrug baseline data were collected.

Drugs. -Adrenoceptor blockade was induced with 1 mg/kg of phentolamine. Combined - and -adrenoceptor blockade was produced with the same dose of phentolamine combined with 1 mg/kg of propranolol, administered simultaneously. Phentolamine and propranolol were dissolved in isotonic saline. The vehicle (control) experiments in this study were also part of concurrent studies using water-insoluble drugs. The vehicle contained, in ml, 0.5 1 N NaOH, 0.5 ethanol, and 0.5 propylene glycol in 28.5 ml of 5% glucose.

Data analysis. Unless otherwise stated, results are presented as means ± SE. Resting values before drug or vehicle administration are presented in Table 1 but were not included in the statistical analysis of drug and exercise effects. For each outcome variable, the control, -blockade, and  + -blockade groups were compared using a two-way ANOVA with 10 "blocks" (dogs) and 12 "treatments" (drug-exercise combinations). When treatment means differed from equality at the P < 0.05 level according to a two-way ANOVA, the Student-Newman-Keuls test was used to test all paired comparisons between means. To compare the slopes of the coronary sinus oxygen tension vs. myocardial oxygen consumption lines, the slopes were first determined individually for each dog and each drug. These slopes were then compared using two-way ANOVA and the Student-Newman-Keuls test. The average slopes were plotted, and the regression lines were centered at the mean oxygen consumption and mean venous oxygen tension of the data set being plotted. Statistical tests were performed using SigmaStat software (SPSS) for two-way ANOVA and multiple comparisons.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

Tracings of coronary blood flow and aortic pressure from one dog under each experimental condition, during rest and at the highest level of exercise, are presented in Fig. 1. Figure 2 shows the predrug rest, postdrug rest, and exercise results for coronary blood flow, myocardial oxygen consumption, heart rate, mean aortic pressure, lactate extraction, and coronary venous carbon dioxide tension. During resting conditions, vehicle infusion had no effect on any of these variables. Compared with control, -adrenoceptor blockade decreased mean aortic pressure and increased heart rate, whereas combined  + -adrenoceptor blockade caused only a decrease in mean aortic pressure. During exercise, combined - and -adrenoceptor blockade significantly reduced coronary blood flow, myocardial oxygen consumption, heart rate, and mean aortic pressure compared with control. In relation to control, -adrenoceptor blockade during exercise increased myocardial oxygen consumption at the lower exercise levels but not at the highest level. Mean aortic pressure was significantly lower than control under all conditions after -adrenoceptor blockade and after  + -adrenoceptor blockade. Lactate extraction was lower during -adrenoceptor blockade, both at rest and during exercise. These data and others are summarized in Table 1.

View larger version (36K): [in this window] [in a new window]   Fig. 1.   Recordings of coronary blood flow and aortic pressure in 1 dog at rest and during level 3 exercise. Control, data obtained after vehicle infusion; -blockade, data collected after -adrenoceptor blockade with phentolamine (1 mg/kg);  + -blockade, after phentolamine plus propranolol (l mg/kg, each) administration; HR, heart rate.

View larger version (42K): [in this window] [in a new window]   Fig. 2.   Average (means ± SE) results of coronary blood flow (A), heart rate (B), myocardial O2 consumption (C), mean aortic pressure (D), coronary sinus CO2 tension (E), and lactate extraction (F) from 10 dogs at rest, after drug or vehicle infusion, and at levels 1-3 of exercise intensity. All 10 dogs received all treatments. * P < 0.05 vs. response during control conditions at the same exercise level.

Because adrenoceptor blockade affects myocardial oxygen consumption, responses to exercise must also be evaluated in relation to myocardial oxygen consumption. Figure 3 contains a plot of coronary sinus oxygen tension vs. myocardial oxygen consumption under the three different experimental conditions. During control conditions, exercise resulted in a moderate decrease in coronary sinus oxygen tension with increasing myocardial oxygen consumption [slope = 0.023 mmHg/(µl O₂ · min¹ · g¹)]. During  + -adrenoceptor blockade, this relationship shifts to lower oxygen tensions and has a significantly more negative slope [0.045 mmHg/(µl O₂ · min¹ · g¹), P < 0.05]. During -adrenoceptor blockade, the slope of the relationship is more positive than during control exercise [0.004 mmHg/(µl O₂ · min¹ · g¹), P < 0.05].

View larger version (18K): [in this window] [in a new window]   Fig. 3.   A: mean values ± SE of coronary venous oxygen tension during rest and 3 levels of exercise plotted against myocardial O2 consumption with individual regression lines for the 3 treatments. B: regression lines are repeated with significant differences in slopes indicated (2-way ANOVA and Student-Newman-Keuls test). Coronary venous O2 tension indicates the match between coronary O2 delivery and myocardial O2 consumption. The steep slope of combined - and -adrenoceptor blockade indicates a modest match by local metabolic factors in the absence of adrenergic mechanisms. The difference in slopes between the  + - and -adrenoceptor blockade treatments demonstrates feedforward -adrenoceptor-mediated coronary vasodilation. The difference in slopes between the control and -adrenoceptor blockade demonstrates the well-established feedforward -adrenoceptor-mediated coronary vasoconstriction.

Plasma adenosine concentrations. Arterial plasma, coronary sinus plasma, and estimated interstitial adenosine concentrations are presented in Fig. 4. During control conditions and during  + -adrenoceptor blockade, coronary sinus plasma adenosine did not increase significantly above resting levels at any exercise level. The estimated interstitial adenosine concentration in these two series did not reach the vasoactive threshold of ~117 nM (46). During -adrenoceptor blockade, however, coronary venous adenosine increased substantially in 7 of 10 dogs at the two highest exercise levels. Due to the large variability, this increase was statistically significant only at the second exercise level. Interstitial adenosine paralleled the venous adenosine and reached vasoactive concentrations at the two highest exercise levels.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Arterial (A), coronary venous (B), and estimated interstitial (C) concentrations of adenosine. Compared with resting levels, neither arterial, venous, nor interstitial adenosine concentration increased during exercise under control conditions or after  + -adrenoceptor blockade. During -adrenoceptor blockade, venous and estimated interstitial adenosine concentrations increased relative to control at the highest levels of exercise. Due to large variability, the increases were statistically significant only at the second level of exercise (two-way ANOVA and Student-Newman-Keuls test). Horizontal line at an interstitial concentration of 117 nM represents previously determined threshold for adenosine-induced vasodilation (46). * P < 0.05.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

The present experiments are the first to combine measurements of plasma adenosine and catecholamine concentration (17) with adrenergic blockade during exercise. The major findings are 1) -adrenoceptor-mediated feedforward coronary vasodilation is an important component of coronary vasodilation that matches coronary blood flow to myocardial oxygen consumption during exercise. 2) Interstitial adenosine concentration does not increase to compensate for the loss of -adrenoceptor-mediated coronary vasodilation during exercise. 3) Exercise during systemic -adrenoceptor blockade may cause myocardial ischemia as demonstrated by high coronary venous plasma adenosine concentrations in 7 of the 10 experimental animals.

Feedforward control may also be called yoked (parallel) or open-loop control. Feedforward control differs from feedback control in that an error signal is not involved. Activation of feedback control requires an error signal that is the difference between an actual reading and some set or operating point. Local metabolic control of coronary blood flow is an example of closed-loop, negative-feedback control that tends to keep myocardial oxygen tension constant (8). A feedforward system is activated by a direct signal that may be yoked to control two variables in parallel. The -adrenoceptor-mediated vasoconstriction and -adrenoceptor-mediated vasodilation observed in the present study are examples of feedforward control. The sympathetic discharge to the heart during exercise results in a parallel or yoked discharge to the pacemaker, ventricular myocardium, and coronary vessels. During exercise, the adrenergic activation of the pacemaker and myocardium results in a large increase in myocardial oxygen consumption, whereas the parallel activation of -adrenoceptor coronary vasodilation increases oxygen delivery to the heart without reference to a myocardial oxygen tension error signal. There are two theoretical strategies for obtaining rapid, accurate matching of coronary blood flow to myocardial oxygen consumption: 1) a high-gain, closed-loop feedback system or 2) adding open-loop feedforward control to a closed-loop feedback system with modest gain. High-gain feedback systems tend to be unstable, with overshoots and oscillations, or have very slow response times to avoid the instability. However, a combined feedforward and feedback control system is capable of rapid and accurate control (34). The present results demonstrate combined adrenergic feedforward and local metabolic feedback control of coronary blood flow during exercise.

Logic of the experimental design. The increase in coronary blood flow during exercise is thought to result largely from the release of local metabolic vasodilators that are secondary to the increase in myocardial oxygen consumption (15). Superimposed on the local metabolic vasodilation are the direct vascular effects of cardiac autonomic nerves and circulating catecholamines. It has been demonstrated many times that -adrenoceptor blockade decreases coronary vascular resistance during exercise (1, 18, 23, 26, 39, 43). This was originally interpreted to mean that the direct vascular effect of catecholamines during exercise is vasoconstriction, which is normally masked by the much larger local metabolic vasodilation. Such an analysis ignores the possibility that the vasodilation ascribed to local metabolic factors also includes a direct -adrenoceptor vasodilator influence of catecholamines on coronary resistance vessels.

Balance between oxygen supply and consumption. All of the responses described above must be evaluated in relation to myocardial oxygen consumption, the primary determinant of coronary blood flow. A sensitive way to elucidate vasodilator or vasoconstrictor influences on coronary blood flow is a plot of coronary sinus oxygen tension vs. myocardial oxygen consumption. Coronary sinus oxygen tension is an index of tissue oxygenation and reflects the balance between coronary oxygen delivery and myocardial oxygen consumption. On such a plot, a vasodilator influence will either shift the line upward or make the slope more positive. A vasoconstrictor influence will shift the line downward or make the slope more negative (33). Influences on baseline flow shift the lines in a parallel manner, whereas an intervention that affects exercise vasodilation will change the slope. An added virtue of such plots is that they facilitate comparison of experimental conditions in which heart rate and arterial pressure are different. Such differences affect coronary blood flow by changing myocardial oxygen consumption, which is accounted for in these plots. Compared with control exercise, -adrenoceptor blockade shifts the slope of the coronary venous oxygen tension vs. myocardial oxygen consumption plot toward higher oxygen tensions (Fig. 3), demonstrating the existence of -adrenoceptor vasoconstriction. Compared with -adrenoceptor blockade,  + -adrenoceptor blockade makes the slope of this plot significantly more negative. This demonstrates the presence of -adrenoceptor vasodilation during exercise. Quantitation of these effects in terms of blood flow is a complex issue explored in the accompanying study (17).

Comparison with previous studies. The presence of -adrenoceptor vasoconstriction in the coronary circulation has been demonstrated in previous studies of exercising dogs (1, 18, 23, 26, 39, 43). The results in Fig. 3 confirm these findings. The purpose of the current study was to isolate -adrenoceptor vasodilation, for which -adrenoceptor blockade experiments were an important part of the experimental design.

-Adrenoceptor subtype. Feedforward -adrenoceptor vasodilation would be much simpler to demonstrate if the -adrenoceptors on coronary smooth muscle were of a different subtype than the cardiac myocyte -adrenoceptor. Whereas at least some of the coronary vascular receptors are ₂ (19, 33, 38, 48), there is abundant evidence for coronary vascular ₁-adrenoceptors as well (33, 40, 41, 47, 48). Given the presence of both receptor subtypes, it is clear that full pharmacological blockade of -adrenoceptor vasodilation requires a nonselective antagonist, such as the propranolol used in the present study. Selective blockade of ₂-adrenoceptors reduces coronary blood flow in dogs at a fixed exercise intensity (10, 32), but, without measurements of myocardial oxygen consumption and coronary sinus oxygen tension, it is not clear whether this is caused solely by loss of ₂-adrenoceptor vasodilation.

Location of - and -adrenoceptors. One of the earliest studies of isolated coronary arteries noted that norepinephrine contracts large vessels but dilates small ones (52). This may be explained by the distribution of - and -adrenoceptors along the coronary arterial tree. A subsequent in vivo study found -adrenoceptor vasoconstriction only in coronary arteries larger than 100 µm in diameter (9). Conversely, -adrenoceptor density varies inversely with coronary vessel diameter (36, 37). This reciprocal gradient in adrenoceptors favors sympathetic vasodilation in resistance vessels and vasoconstriction in distribution vessels. -Adrenoceptor vasoconstriction and -adrenoceptor vasodilation are, therefore, not truly antagonistic but appear to be spatially distributed so as to increase coronary blood flow and improve transmural flow distribution (26).

Study limitations. A limitation of the present study is that -adrenoceptor blockade with phentolamine increases cardiac and systemic norepinephrine release [concentrations are presented in the accompanying study (17)]. These higher norepinephrine concentrations may exaggerate feedforward -adrenoceptor vasodilation. A similar problem occurs if results during -blockade alone are compared with control exercise results (4, 24). In this case, elevated catecholamine release (45) would exaggerate the unopposed -adrenoceptor vasoconstriction component. Thus adrenergic blockade studies can show the presence of both - and -adrenoceptor-mediated effects but have inherent limitations when estimating their magnitude. Therefore, the studies in the accompanying study (17) were undertaken to estimate the -feedforward vasodilator contribution at normal exercise catecholamine concentrations.

Role of adenosine as a metabolic negative feedback vasodilator. In exercising dogs, the decrease in coronary sinus oxygen tension represents an error signal and implies that local metabolic feedback vasodilation is occurring. Adenosine has long been proposed as a metabolic vasodilator in the coronary circulation (6, 7, 15). During hypoxia or ischemia, there is compelling evidence in favor of such a role for adenosine (3, 15, 31, 42, 45). During normal exercise, the role of adenosine is far more controversial. An increase in coronary venous plasma adenosine during exercise has been reported (13) but was not confirmed in a later study (49). Adenosine receptor antagonists do not reduce coronary blood flow during exercise (2, 49). It has been suggested that adenosine levels may increase sufficiently to overcome the competitive blockade (21), but this was not found when adenosine concentration was measured in exercising dogs (49).

Adenosine during -adrenoceptor blockade. The high coronary venous and estimated myocardial interstitial adenosine levels observed in 7 of the 10 dogs during exercise, with -adrenoceptor blockade, indicates regional myocardial ischemia. If these values were representative of the whole myocardium, coronary blood flow would be at the maximum pharmacological value of the adenosine dose-response curve of ~ 4 ml · min¹ · g¹ (46) rather than the ~ 2 ml · min¹ · g¹ observed during exercise in the present study. The fall in myocardial lactate extraction during exercise with -adrenoceptor blockade (Fig. 2) also suggests regional, but not global, ischemia because overall lactate extraction remained positive. Regional adenosine release with phentolamine is not responsible for the increase in coronary venous oxygen tension compared with control exercise. Adenosine attenuates the fall in local oxygen tension in the region where it is released, but this would not increase oxygen tension above control and should have only a small effect on the global tension measured in the coronary sinus.