J Appl Physiol 97: 404-415, 2004;
doi:10.1152/japplphysiol.01345.2003
8750-7587/04 $5.00
INVITED REVIEW
HIGHLIGHTED TOPICS
Skeletal and Cardiac Muscle Blood Flow
Matching coronary blood flow to myocardial oxygen consumption
Johnathan D. Tune,2
Mark W. Gorman,1 and
Eric O. Feigl1
1Department of Physiology and Biophysics, University of Washington, Seattle, Washington 98195; and 2Department of Physiology, Louisiana State University Health Sciences Center, New Orleans, Louisiana 70112
 |
ABSTRACT
|
|---|
At rest the myocardium extracts 
75% of the oxygen delivered by coronary blood flow. Thus there is little extraction reserve when myocardial oxygen consumption is augmented severalfold during exercise. There are local metabolic feedback and sympathetic feedforward control mechanisms that match coronary blood flow to myocardial oxygen consumption. Despite intensive research the local feedback control mechanism remains unknown. Physiological local metabolic control is not due to adenosine, ATP-dependent K+ channels, nitric oxide, prostaglandins, or inhibition of endothelin. Adenosine and ATP-dependent K+ channels are involved in pathophysiological ischemic or hypoxic coronary dilation and myocardial protection during ischemia. Sympathetic 
-adrenoceptor-mediated feedforward arteriolar vasodilation contributes 25% of the increase in coronary blood flow during exercise. Sympathetic 
-adrenoceptor-mediated vasoconstriction in medium and large coronary arteries during exercise helps maintain blood flow to the vulnerable subendocardium when cardiac contractility, heart rate, and myocardial oxygen consumption are high. In conclusion, several potential mediators of local metabolic control of the coronary circulation have been evaluated without success. More research is needed.
adenosine; ATP-dependent K+ channels; nitric oxide; norepinephrine; feedback control; feedforward control
THE MYOCARDIUM HAS A VERY limited anaerobic capacity, and the heart is dependent on a continuous supply of oxygen from the coronary circulation. Myocardial oxygen consumption increases whenever there is tachycardia and may increase fivefold in the transition from rest to exercise. Powerful mechanisms increase coronary blood flow whenever myocardial oxygen consumption is increased. Without these mechanisms, decreased cardiac output, hypotension, arrhythmias, and death would result. The purpose of the present paper is to review the mechanisms that match oxygen delivery to myocardial oxygen consumption via coronary blood flow.
There are numerous factors that influence the contractile state of coronary vascular smooth muscle, such as the extracellular potassium and calcium ion concentrations (76). However, this brief review is restricted to factors that have been postulated to be involved in coronary vasodilation when myocardial oxygen consumption is augmented. The dominant influence on coronary blood flow is local metabolic feedback control. As will be seen, despite decades of research the local metabolic coronary controller remains unknown. In contrast, feedforward adrenergic control of coronary blood flow contributes significantly to exercise coronary vasodilation and to maintaining flow to the vulnerable subendocardium. The sine qua non of a negative feedback control mechanism is the presence of an error signal. An error signal is the difference between the momentary value of a sensed variable and a set point. The error signal then activates a mechanism that results in an adjustment that restores the regulated variable back toward the set point in a feedback manner.
Because resting left ventricular myocardial oxygen consumption is great (60 µl·min–1·g–1) with a high oxygen extraction (75%) and a low coronary venous oxygen tension (18 Torr), the error signal is assumed to be related to the balance between oxygen supply and oxygen consumption. The balance between oxygen delivery via coronary flow and myocardial oxygen consumption is given by the coronary venous oxygen tension. Oxygen tension is preferable to oxygen content (or extraction) or hemoglobin saturation because the local metabolic chemical mediator will most likely be sensitive to the thermodynamic potential of oxygen (oxygen tension) rather than blood oxygen content. Also, hemoglobin saturation changes as the hemoglobin dissociation curve shifts with carbon dioxide tension, pH, and temperature.
A useful method for analyzing coronary vascular responses is a graph of coronary venous oxygen tension vs. myocardial oxygen consumption as introduced by Heyndrickx et al. (95) and illustrated in Fig. 1. Line A in Fig. 1 shows the normal modest decline in coronary venous oxygen tension as myocardial oxygen consumption increases from rest to exercise. If this line were horizontal, then the match between oxygen supply and consumption would be perfect despite a severalfold increase in myocardial oxygen consumption. The widening divergence from horizontal of line A reflects an increasing error signal as myocardial oxygen consumption increases. Line B indicates the expected result when a mechanism involved in exercise coronary hyperemia is experimentally inhibited. The key point is the difference in slopes between lines A and B. The logic is that the greater the oxygen metabolism along the x-axis, the greater will be the fall in coronary venous oxygen tensions on the y-axis if local metabolic vasodilation is attenuated. In other words, if the coupling between coronary blood flow and myocardial oxygen consumption is experimentally inhibited, then the error signal will increase as oxygen metabolism increases. Line C in Fig. 1 indicates the expected effect if a tonic vasodilator effect is experimentally blocked. There is a decrease in coronary venous oxygen tension at rest and during exercise without a difference in slopes compared with the normal control line A. In this case, the error signal does not grow larger as oxygen consumption increases.
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 1. Graphs of coronary venous oxygen tension vs. myocardial oxygen consumption are very useful in interpreting coronary vascular responses. Line A shows the normal modest decline in coronary venous oxygen tension as left ventricular myocardial oxygen consumption increases from rest to exercise. Line B shows the expected result if a coronary vasodilator mechanism activated during exercise is inhibited. The key point is the difference in slopes between lines A and B. Line C shows the expected result if a tonic vasodilator mechanism present at rest and during exercise is inhibited. In this case the error signal does not grow larger during exercise, resulting in a line parallel to line A but at lower oxygen tensions.
|
|
The relationships between myocardial oxygen consumption, coronary blood flow, and coronary venous oxygen tension during exercise for the right and left ventricles are shown in Fig. 2. Figure 2A illustrates that oxygen consumption is less in the right ventricle than in the left ventricle, and this is associated with a higher coronary venous oxygen tension (84, 90, 198, 208). Figure 2A also demonstrates that, as right ventricular oxygen consumption increases from rest to exercise, the initial effect is an increasing oxygen extraction (falling venous oxygen tension). The augmented oxygen extraction in the right ventricle occurs with little change in flow when venous tension is above 20 Torr, as shown in Fig. 2B. Conversely, in the left ventricle (where resting venous oxygen tension is 18 Torr), increases in myocardial oxygen consumption result in modest decreases in venous oxygen tension (Fig. 2A) and large increases in coronary flow (Fig. 2B).
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 2. Relationships between myocardial oxygen consumption, coronary blood flow, and coronary venous oxygen tension during exercise for canine right and left ventricles. A: oxygen consumption is lower in the right ventricle than in the left, and this is associated with higher coronary venous oxygen tensions. B: as right ventricular oxygen consumption increases during exercise, the initial response is increased oxygen extraction (decreased venous oxygen tension) with little increase in flow. However, there is an inflection point between oxygen tensions of 15 and 20 Torr at which flow increases sharply with decreasing venous oxygen tension. Resting left ventricular venous oxygen tension is 18 Torr and sits near this inflection point. These results suggest that the left and right ventricles may share common flow control mechanisms and that a powerful vasodilator mechanism is recruited when coronary venous oxygen tension falls below 15–20 Torr. Data from Refs. 84, 90, 198, 208.
|
|
The interesting intersection between the right and left ventricles in Fig. 2 suggests that they may share common flow control mechanisms and that a powerful vasodilator mechanism is recruited when coronary venous oxygen tension falls below 15–20 Torr.
The postulated coronary control mechanisms are individually considered below.
|
OXYGEN AND CARBON DIOXIDE
|
|---|
The simplest local metabolic vasodilators to consider are oxygen and carbon dioxide. Arterial hypoxemia and hypercarbia both result in coronary vasodilation (25, 76); however, the question is whether local tissue concentrations of these gases control coronary blood flow. Broten et al. (25) used coronary venous oxygen and carbon dioxide tensions as a surrogate for tissue values to correlate with coronary blood flow when arterial blood gases were varied by employing membrane oxygenators. A striking synergistic effect between coronary venous oxygen and carbon dioxide tensions and coronary blood flow was observed. That is, high carbon dioxide levels potentiated the effect of hypoxia and hypoxia potentiated the effect of hypercarbia in a greater than simply additive manner. However, the synergistic effect was able to predict only 40% of coronary vasodilation during cardiac pacing (25) and 23% of the conductance change during autoregulation (24). Those correlative experiments do not distinguish between the respiratory gases acting directly on vascular smooth muscle and there being an intermediate transmitter such as adenosine. The lack of 100% correlation between coronary flow and coronary venous respiratory gas tensions suggests that factors other than oxygen and carbon dioxide are involved in coronary control, or that coronary venous values are an incomplete index of tissue values.
|
ADENOSINE HYPOTHESIS
|
|---|
In 1963, Berne (20) and Gerlach et al. (81) independently proposed that adenosine is the vasodilatory metabolite that links changes in coronary blood flow to myocardial metabolism. The hypothesis predicts that increases in myocardial oxygen consumption decrease myocardial oxygen tension to stimulate the release of adenosine from cardiomyocytes. The resulting increase in cardiac interstitial adenosine concentration activates coronary vascular smooth muscle adenosine receptors, which increases coronary blood flow and augments oxygen delivery. This increase in oxygen delivery acts to restore myocardial oxygen tension to a normal operating level, thereby decreasing cardiac adenosine production in a negative feedback manner. Thus, in short, the cardiac interstitial concentration of adenosine controls coronary blood flow to match myocardial oxygen delivery with myocardial oxygen consumption, thereby maintaining myocardial oxygen tension within the normal physiological range (21).
Adenosine and local metabolic coronary blood flow control.
Over the past 40 years, numerous investigations have examined the role of adenosine in local metabolic coronary vasodilation during increases in myocardial oxygen consumption (13, 15, 21, 44, 51, 57, 62, 64, 65, 68, 82, 91, 104, 109, 128, 129, 133, 137, 198, 200, 204, 206). Results from these studies are inconsistent in that indexes of cardiac interstitial adenosine concentration were augmented during increased myocardial oxygen consumption in some experiments (13, 68, 82, 108, 132, 202) but not in others (115, 198, 200, 206). One possible explanation for these conflicting results is that the adenosine measurements reported in the early investigations did not reliably estimate the cardiac interstitial concentration of adenosine (16, 198). Blockade of endogenous adenosine receptors does not significantly decrease coronary blood flow at rest or during increases in myocardial oxygen consumption (15, 57, 104, 128, 198, 206). A possible explanation for the failure of adenosine receptor blockade to reduce coronary blood flow is that the cardiac interstitial adenosine concentration increases sufficiently to overcome the competitive receptor blockade, as would be predicted if adenosine were part of a high-gain negative feedback system (91, 94, 134). Therefore, it is apparent that to adequately test the adenosine hypothesis, it is essential to have a reliable estimate of cardiac interstitial adenosine concentration when adenosine receptors are blocked.
Recently, Tune et al. (198) quantitatively examined the contribution of adenosine to local metabolic coronary vasodilation by combining measurements of coronary venous adenosine concentration with adenosine receptor blockade during exercise in chronically instrumented dogs. Myocardial interstitial adenosine concentration was calculated from coronary blood flow, hematocrit, and the adenosine concentrations in arterial and coronary venous plasma using a mathematical model (112, 185). This model has been extensively tested and described (112, 185, 203). In testing this model, Stepp et al. (185) determined that there is a steep dose-response relationship between coronary blood flow and interstitial adenosine concentration and that the threshold for adenosine-mediated vasodilation is 117 nM. Tune et al. (198) found that coronary venous plasma adenosine concentration was little changed with exercise and the estimated interstitial adenosine concentration remained well below the threshold concentration necessary for coronary vasodilation. In addition, coronary venous and estimated interstitial adenosine concentrations did not increase to overcome the competitive adenosine receptor blockade of either 8-phenyltheophylline or 8-p-sulfophenyltheophylline. These findings indicate that adenosine is not required for local metabolic coronary vasodilation during physiological increases in myocardial oxygen consumption.
Tune et al. (198) also found that adenosine receptor blockade with 8-phenyltheophylline produced a significant decrease in coronary venous PO2 at a given myocardial oxygen consumption, i.e., parallel shift downward in the coronary venous PO2 vs. myocardial oxygen consumption relationship (Fig. 3A). This finding is consistent with earlier studies in humans (64), pigs (57), and dogs (200, 206), but not in a careful study by Bache et al. (15) in which no change was observed with 8-phenyltheophylline. The parallel shift downward suggests that adenosine might exert a tonic vasodilator influence on the coronary circulation at rest and during exercise. However, this finding is at odds with the cardiac adenosine measurements in that the estimated interstitial adenosine concentration remained well below the threshold value necessary for coronary vasodilation (198). This paradoxical finding could be due to a nonspecific effect of 8-phenyltheophylline because degradation of cardiac adenosine by intracoronary adenosine deaminase infusion does not affect coronary venous PO2 in unstressed dogs (111, 128, 139). Regardless of the parallel 8-phenyltheophylline effect, the fact that the slope of the coronary venous PO2 vs. myocardial oxygen consumption relationship was not steepened by adenosine receptor blockade indicates that adenosine is not required for local metabolic coronary vasodilation during exercise.
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 3. Graphs of coronary venous oxygen tension vs. myocardial oxygen consumption at rest and during exercise, under control conditions and after inhibition of several proposed exercise vasodilator mechanisms. A: blockade of adenosine receptors with 8-phenyltheophylline (8-PT). B: blockade of ATP-dependent K+ (KATP) channels with glibenclamide. C: inhibition of nitric oxide synthesis with N -nitro-L-arginine (LNA). D: simultaneous inhibition of adenosine, KATP channels, and nitric oxide. E: inhibition of prostaglandin synthesis with indomethacin. F: regression lines from the interventions in A–D plotted in 1 graph for comparison. In this panel the control line has been fit to the combined control data points from A–D. Data are from the references noted in each panel.
|
|
Adenosine and coronary blood flow control during ischemia.
Although adenosine is not responsible for functional coronary hyperemia under normal physiological conditions, adenosine does contribute to coronary vasodilation when the myocardium is ischemic (121). Coronary vasodilation during exercise is significantly reduced by combined degradation of endogenous adenosine with adenosine deaminase and blockade of adenosine receptors with 8-phenyltheophylline in chronically instrumented dogs with a coronary stenosis (coronary perfusion pressure = 40 mmHg) (54, 121). These findings indicate that as long as the change in coronary blood flow is sufficient to prevent ischemia, adenosine release is not elevated and thus does not contribute to local metabolic coronary vasodilation. However, if changes in coronary blood flow do not sufficiently match changes in myocardial metabolism, such as in patients with a critical coronary stenosis, adenosine is released by the ischemic myocytes in an attempt to augment coronary blood flow and oxygen delivery. This increase in cardiac adenosine release not only increases coronary blood flow but has several other effects that decrease myocardial injury during ischemia, as recently reviewed by Headrick et al. (92). The cardioprotective effects of adenosine include 1) decreased myocardial contractility (negative inotropic effect) (17, 26, 43, 48–50); 2) decreased cardiac arrhythmias (negative chronotropic and dromotropic effects) (6, 45, 46, 73, 123); 3) decreased platelet aggregation (7, 63, 97, 98); 4) increased myocardial glucose metabolism and energetics (8, 69, 71, 77, 102, 119, 120); 5) antioxidant effect (decreased superoxide production) (37, 127, 174); 6) reduced infarct size (preconditioning) (10, 124, 132, 192, 199); and 7) attenuation of myocardial stunning (171, 179, 210).
In summary, recent experimental evidence indicates that adenosine is not the physiological local metabolic vasodilatory metabolite that links coronary blood flow to myocardial metabolism. However, adenosine plays a significant role in increasing coronary blood flow and protecting the ischemic myocardium from irreversible damage.
|
KATP CHANNELS
|
|---|
ATP-dependent K+ (KATP) channels were first identified in 1983 by Noma (158). These channels are present in a variety of tissues including vascular smooth muscle, arterial endothelial cells, cardiac muscle, skeletal muscle, brain, and pancreatic -cells (9, 35, 107, 144, 153, 158, 182). KATP channels are activated by alterations in the cellular metabolic state which, in turn, initiate membrane hyperpolarization and relaxation of vascular smooth muscle (153).
Interestingly, KATP channels are involved in adenosine-mediated coronary vasodilation because the sulfonylurea derivative glibenclamide (KATP channel antagonist) significantly impairs adenosine-induced vasodilation (2, 18, 34, 40, 60, 93, 113, 154, 160, 183). Early investigations have shown that KATP channels play a role in hypoxia-induced coronary vasodilation (40, 151) and coronary reactive hyperemia (11, 34, 60). However, because hypoxia and ischemia result in adenosine release from the myocardium (see ADENOSINE HYPOTHESIS above), the blunting of hypoxic or ischemic coronary vasodilation by glibenclamide most likely represents inhibition of adenosine-mediated vasodilation (see MULTIPLE CONTROLLERS below).
Under normal baseline conditions, blockade of KATP channels with glibenclamide results in a decrease in coronary blood flow (12–25%) with little or no change in myocardial oxygen consumption (39, 58, 60, 72, 100, 137, 147, 172, 177, 183), demonstrating that KATP channels contribute to basal coronary vascular tone. In contrast, KATP channels are not required for local metabolic vasodilation during increases in myocardial oxygen consumption (55, 58–60, 137, 172, 173). Richmond et al. found that glibenclamide did not attenuate the increase in coronary blood flow to paired cardiac pacing (172) or treadmill exercise (173) in dogs. They found that the relationship between coronary venous PO2 and myocardial oxygen consumption was shifted downward in a parallel manner by glibenclamide (Fig. 3B), indicating that KATP channels exert a tonic coronary vasodilator influence at rest and during exercise. These results are consistent with other studies in dogs (58–60) and pigs (55, 137). Also, these channels do not appear to be required for maintaining coronary blood flow constant when coronary perfusion pressure is altered, i.e., pressure-flow autoregulation (183), although glibenclamide treatment slows the response to a step change in coronary perfusion pressure (39). Taken together, these findings demonstrate that KATP channels have a tonic resting vasodilator effect in the coronary circulation that is also present during exercise. The evidence for greater coronary KATP channel opening during exercise is inconsistent, indicating that other factors are important for local metabolic control when myocardial oxygen consumption is augmented.
|
NITRIC OXIDE
|
|---|
Nitric oxide is formed from L-arginine by nitric oxide synthase and is released by endothelial cells (146). Nitric oxide release is augmented by agonists such as acetylcholine and bradykinin (5, 146, 167) and also by mechanical stimulation such as shear stress (27, 28, 110, 114, 116, 184), pulsatile flow (30), and axial strain (12).
Nitric oxide has been implicated as a factor in matching coronary blood flow with changes in myocardial metabolism. However, inhibition of nitric oxide synthesis with arginine analogs results in little (66, 70, 126, 130, 155) or no change (4, 22, 30, 41, 53, 101, 162, 165, 180, 181, 196, 202) in coronary blood flow at rest or when myocardial oxygen consumption is elevated. Studies also consistently show that nitric oxide is not required for exercise-induced coronary vasodilation (4, 22, 53, 101, 180, 196, 202). However, nitric oxide synthesis inhibition does decrease coronary venous oxygen tension at a given level of myocardial oxygen consumption (4, 22, 53, 101, 180, 196, 202), indicating that nitric oxide exerts a tonic vasodilator influence at rest and during increases in myocardial metabolism (Fig. 3C). The primary vasodilatory effect of nitric oxide is probably located in large coronary arteries because numerous studies have demonstrated a significant reduction in epicardial coronary diameter when nitric oxide synthesis is inhibited at rest (32, 33, 87, 105, 122, 155, 163) and during increases in myocardial oxygen consumption (19, 30, 66, 70, 170, 202). Studies by Bernstein et al. (22) and Traverse et al. (195) indicate that the exercise-induced epicardial coronary vasodilation is mediated by an increase in endothelial-mediated nitric oxide production at high levels of myocardial oxygen consumption. This increased nitric oxide production is most likely due to augmented shear stress that results from the high coronary flow rates (27, 28, 110, 114, 116, 184, 201). The vasodilation of epicardial coronary arteries by nitric oxide is physiologically beneficial in that it acts to prevent excessive shear stress on coronary endothelial cells and to preserve arteriolar vasodilator reserve (106).
|
PROSTAGLANDINS
|
|---|
Prostaglandins are arachidonic acid metabolites that have been shown to be released into the coronary circulation during episodes of hypoxia (1), anoxia (23), and coronary artery occlusion (3). However, an important role of prostaglandins in physiological local metabolic coronary vasodilation has not been demonstrated (38, 65, 89, 96). Dai and Bache (38) found that inhibition of prostaglandins with indomethacin did not significantly affect resting coronary blood flow or attenuate exercise-induced coronary vasodilation. In addition, the relationship between coronary venous oxygen tension and myocardial oxygen consumption was unaltered by indomethacin (Fig. 3E), demonstrating that prostaglandins do not mediate functional coronary exercise hyperemia. These findings agree with the study of Edlund et al. (65), who found that ibuprofen did not significantly attenuate exercise hyperemia in normal, healthy humans. Prostaglandins may play a role in coronary blood flow control in patients with ischemic heart disease (52, 80). However, this is not a consistent finding (61, 161).
|
ENDOTHELIUM-DERIVED HYPERPOLARIZING FACTOR
|
|---|
Endothelium-derived hyperpolarizing factor (EDHF) is released by coronary endothelial cells and mediates vasodilation by hyperpolarizing vascular smooth muscle (79, 156). The presence of EDHF is based on the observation that the endothelium-dependent vasodilator acetylcholine hyperpolarizes and relaxes isolated coronary vessels in the presence of combined nitric oxide and cyclooxygenase blockade (78, 79, 108, 166, 168, 189). It appears that a component of EDHF is cytochrome P-450 metabolites of arachidonic acid, i.e., epoxyeicosatrienoic acids and hydroxyeicosatetraenoic acids, both of which modulate coronary vascular tone (29, 78, 88, 141, 156, 157, 159, 175, 205). Although EDHF has been shown to mediate coronary vascular relaxation, its role in local metabolic coronary vasodilation during increases in myocardial metabolism has not been investigated.
|
ENDOTHELIN
|
|---|
Endothelin-1 is an endothelium-derived peptide that has potent, long-acting vasoconstrictor properties (207). The role of endothelin in the coronary circulation has recently been reviewed by Lavallee and Thorin (118). Endothelin-mediated coronary vasoconstriction occurs via activation of ETA receptors (83, 125, 140, 164, 191). However, endothelin can also exert a vasodilator influence on coronary arteries through activation of ETB receptors, but the net effect of endothelin is vasoconstriction. The sensitivity for endothelin-mediated coronary vasoconstriction increases with decreasing vessel diameter (103, 117). Recent investigations that have examined the effects of endothelin receptor blockade consistently show that endothelin contributes to coronary vascular tone under resting conditions but that endothelin has little or no effect on coronary blood flow control as myocardial oxygen consumption is increased during exercise (136, 138, 187). This finding is not surprising in light of the fact that arterial and coronary venous endothelin concentrations are unchanged with exercise (138, 187). Thus inhibition of endothelium-mediated vasoconstriction contributes little to exercise hyperemia.
Interestingly, studies from the Chilian laboratory indicate that 1-adrenoceptor stimulation of cardiac myocytes promotes endothelin-mediated coronary vasoconstriction both in vitro (193) and in vivo (42). However, at present this mechanism has only been demonstrated with intracoronary infusions of pharmacological concentrations of 1-adrenoceptor agonists. Whether this mechanism of 1-adrenoceptor-mediated endothelin release is active under physiological conditions such as exercise has not been determined.
|
MULTIPLE CONTROLLERS
|
|---|
There are multiple mechanisms of coronary blood flow control, and one hypothesis was that when KATP channels or nitric oxide synthesis are inhibited, adenosine levels increase in compensation. The reason for postulating adenosine compensation is based on the observation that the addition of an adenosine receptor antagonist after prior KATP channel blockade (59, 60, 101) or nitric oxide synthesis inhibition (130, 190) decreased coronary blood flow and/or depressed the slope of the relationship between coronary venous oxygen tension and myocardial oxygen consumption. However, recent studies from the Feigl laboratory have demonstrated that inhibition of either KATP channels or nitric oxide synthesis does not significantly increase coronary venous or estimated cardiac interstitial adenosine concentration at rest and during exercise (172, 173, 196). In addition, the estimated interstitial adenosine concentration remained well below the threshold necessary for coronary vasodilation with or without KATP channel blockade (172, 173) or nitric oxide synthesis inhibition (196). These findings indicate that adenosine does not increase in compensation when KATP channels or nitric oxide synthesis is inhibited during exercise.
A probable explanation for the discordant results is that the intracoronary glibenclamide infusions used by the Bache laboratory (59, 60, 101) result in myocardial ischemia, as demonstrated by decreased myocardial function (101). It is well recognized that ischemic myocardium releases adenosine, thus the adenosine effect is probably due to ischemia rather than a physiological compensatory effect. Similar intracoronary glibenclamide infusions have resulted in oscillations in coronary blood flow (150, 177). Although it has not been studied, a plausible mechanism for the coronary blood flow oscillations is that glibenclamide produces ischemia that increases adenosine levels and thus coronary flow, which relieves the ischemia and washes out the adenosine to return to ischemia in an oscillatory manner.
Combined blockade of adenosine receptors, KATP channels, and nitric oxide synthesis produces a marked parallel downward shift in the coronary venous oxygen tension vs. myocardial oxygen consumption graph (Fig. 3D) (197). This triple blockade resulted in coronary venous oxygen tensions below 7 Torr during exercise with an increase in coronary venous adenosine concentration. However, the adenosine levels did not increase sufficiently to overcome the adenosine receptor blockade. The interpretation is that triple blockade causes a decrease in resting coronary blood flow and that the added stress of exercise results in incipient subendocardial ischemia.
Another question is whether KATP channels and nitric oxide act on coronary blood flow in a linear additive or nonlinear synergistic manner (137). The data from pigs (137) and dogs shown in Fig. 3F suggest the effects are approximately additive.
|
SYMPATHETIC CONTROL OF CORONARY BLOOD FLOW
|
|---|
Sympathetic nervous system activation during exercise increases heart rate, cardiac contractility, and left ventricular afterload, all of which result in metabolic vasodilation. The local metabolic vasodilation obscures the direct effects of sympathetic stimulation on coronary blood vessels. These direct effects include both vasoconstriction through coronary -adrenoceptors and vasodilation via -adrenoceptors.
|
-ADRENOCEPTOR-MEDIATED FEEDFORWARD CORONARY VASODILATION
|
|---|
Direct sympathetic vasodilation of the coronary vessels is an attractive hypothesis for matching coronary flow and oxygen consumption during exercise, because in this case the same stimulus responsible for increasing oxygen consumption will simultaneously increase oxygen delivery. Thus -adrenoceptor vasodilation is a "feedforward" mechanism that does not require an error signal such as decreased cellular oxygen tension (142, 143). The potential for coronary -adrenoceptor vasodilation can be seen most clearly in isolated blood vessels. Small coronary vessels dilate in response to epinephrine and norepinephrine, even without prior -adrenoceptor blockade (169, 186, 209). Thus the net effect of sympathetic stimulation on coronary resistance vessels appears to be vasodilation. However, demonstration of -adrenoceptor coronary vasodilation during exercise is difficult because of the simultaneous presence of a large metabolic vasodilation. The key to separating these effects is to produce metabolic vasodilation free from direct vascular catecholamine effects. In conscious animals this has been achieved by exercise during simultaneous - and -adrenoceptor blockade (56, 84). These responses are then compared with exercise during -adrenoceptor blockade, a situation that includes both metabolic vasodilation and -adrenoceptor vasodilation. Thus the difference between these two conditions will reflect the influence of -adrenoceptor vasodilation. When coronary venous oxygen tension is plotted vs. myocardial oxygen consumption in these experiments (Fig. 4), it can be seen that venous oxygen tension is considerably lower at any given myocardial oxygen consumption when both - and -adrenoceptors are blocked. The conclusion is that coronary arteriolar -receptors contribute substantially to exercise vasodilation.
The experimental approach above has one drawback. The presence of -blockade increases neuronal catecholamine release, which may exaggerate feedforward -vasodilation. This problem has been addressed by estimating the vascular effects of the cardiac catecholamine concentrations measured during normal exercise without adrenergic blockade (85). Epinephrine is released by the adrenal glands and reaches the coronary circulation via the arterial plasma, so its effects can be approximated by exogenous epinephrine infusions in resting animals that match the arterial plasma concentrations during exercise. Plasma epinephrine concentrations reached during strenuous exercise increase myocardial oxygen consumption only slightly in resting dogs and do not change coronary sinus oxygen tension (85). Circulating epinephrine therefore contributes very little to feedforward sympathetic vasodilation.
The arterial plasma concentrations of norepinephrine are not very helpful, because the sympathetic nerves release norepinephrine directly into the interstitial space. It is the interstitial norepinephrine concentration to which the coronary arterioles are exposed. The interstitial norepinephrine concentration can be estimated when the arterial and coronary venous plasma concentrations, coronary plasma flow, and the coronary capillary permeability to norepinephrine are known (36). The interstitial concentration was estimated at 12 nM in strenuously exercising dogs (85). The effect of this concentration on arteriolar diameter was estimated from an in vitro norepinephrine dose-response curve (169). Assuming a cubic dependence of flow on arteriolar diameter (131), the exercise norepinephrine concentrations would increase coronary flow by 67%. Because this level of exercise increased coronary blood flow by 260%, feedforward -adrenoceptor vasodilation by norepinephrine can account for roughly 25% of coronary exercise hyperemia (85).
|
-ADRENOCEPTOR-MEDIATED FEEDFORWARD CORONARY VASOCONSTRICTION
|
|---|
Coronary vessels also possess -adrenoceptors capable of causing vasoconstriction during sympathetic activation. -Adrenoceptor vasoconstriction can be demonstrated in two ways:
1) Coronary sympathetic nerve stimulation during -adrenoceptor blockade lowers coronary blood flow and coronary venous oxygen tension, and this effect can be blocked by -adrenoceptor antagonists (67, 74, 75, 135, 152).
2) -Adrenoceptor blockade during sympathetic stimulation or exercise results in higher coronary blood flow and coronary venous oxygen tension at a given level of myocardial oxygen consumption, compared with control conditions (14, 47, 86, 95, 99, 145, 149, 178). Interestingly, some of the increased coronary blood flow and coronary venous oxygen tension after -adrenoceptor blockade during exercise is due to nitric oxide formation (188). The stimulus for increased nitric oxide formation in this case may be increased to-and-fro flow oscillations in the coronary arteries after -adrenoceptor blockade (148).
The simultaneous presence of -adrenoceptor vasoconstriction and -adrenoceptor vasodilation in the coronary circulation raises the question as to which one predominates during sympathetic activation. The answer depends on vessel size. Large coronary arteries in vitro contract in response to norepinephrine, whereas small coronary arteries relax under these conditions (169, 209). In vivo, -adrenoceptor vasoconstriction is limited to vessels >100 µm in diameter (31). Because arterioles are the primary site of vascular resistance, the overall direct effect of sympathetic activity on coronary vessels is a decrease in vascular resistance. It should be noted that even large coronary arteries dilate in response to exercise. -Adrenoceptor vasoconstriction limits the extent of this dilation rather than causing an actual reduction in diameter. Even so, -adrenoceptor vasoconstriction in larger vessels may not be in direct competition with -adrenoceptor vasodilation. During strenuous exercise, the vasodilated subendocardium is vulnerable to underperfusion due to high myocardial compressive forces and short diastoles. -Adrenoceptor vasoconstriction increases subendocardial flow under these conditions (99). The postulated mechanism is that -adrenoceptor vasoconstriction increases vascular stiffness and so decreases capacitance in medium-sized intramyocardial vessels. The result is a reduction in wasteful to-and-fro flow oscillations during the cardiac cycle (148).
In summary, sympathetic activation results in feedforward -adrenoceptor-mediated coronary vasodilation in arterioles that helps match oxygen delivery to myocardial oxygen consumption during exercise. There is also a simultaneous -adrenoceptor-mediated vasoconstrictor effect in medium and large coronary arteries that helps maintain subendocardial blood flow during exercise when there is tachycardia, augmented contractility, and high myocardial oxygen consumption.
|
CONCLUSION
|
|---|
The physiological local metabolic controller of coronary blood flow remains unknown. The use of a graph of coronary venous oxygen tension vs. myocardial oxygen consumption permits a critical analysis of the factors that control coronary blood flow when myocardial oxygen consumption is augmented, as during exercise. With this approach, the following conclusions may be drawn.
Numerous studies over the past four decades consistently demonstrate that adenosine acts as a pathophysiological coronary vasodilator when the myocardium becomes hypoxic or ischemic. Recent investigations reveal that adenosine acts as a cardioprotective agent during ischemia, producing preconditioning and ameliorating cardiac stunning. However, adenosine does not cause the physiological coronary vasodilation observed during exercise.
KATP channels contribute a vasodilator influence to resting coronary blood flow but are probably not responsible for local metabolic control during exercise.
Nitric oxide dilates large coronary arteries owing to the shear forces on endothelial cells but does not act as the local metabolic vasodilator during exercise.
The actions of adenosine, KATP channels, and nitric oxide are approximately additive, without one compensating for the other when it is blocked. In particular, in the absence of ischemia adenosine levels do not increase in compensation when KATP channels and nitric oxide synthesis are blocked.
Indomethacin treatment to block prostaglandin synthesis has little effect on coronary blood flow control.
Endothelin contributes a modest resting vasoconstrictor influence on the coronary circulation, but inhibition of endothelin is not the vasodilator mechanism in exercise hyperemia.
-Adrenoceptor-mediated feedforward dilation of coronary microvessels contributes 25% to the increased coronary flow observed during exercise.
A paradoxical -adrenoceptor-mediated vasoconstrictor effect in large- and medium-sized coronary arteries during exercise helps prevent subendocardial underperfusion when myocardial oxygen consumption, cardiac contractility, and heart rate are high.
Encouraging progress has been made in understanding the many factors that control coronary blood flow, but much remains to be done.
|
GRANTS
|
|---|
The authors were supported by National Heart, Lung, and Blood Institute Grants HL-49822 and HL-07403 (E. O. Feigl and M. W. Gorman) and HL-67804 (J. D. Tune) and by an American Diabetes Association Career Development Award (J. D. Tune).
|
FOOTNOTES
|
|---|
Address for reprint requests and other correspondence: E. O. Feigl, Dept. of Physiology and Biophysics, Univ. of Washington School of Medicine, Box 357290, Seattle, WA 98195-7290 (E-mail: efeigl{at}u.washington.edu).
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.
|
REFERENCES
|
|---|
- Afonso S, Bandow GT, and Rowe GG. Indomethacin and the prostaglandin hypothesis of coronary blood flow regulation. J Physiol 241: 299–308, 1974.[Abstract/Free Full Text]
- Akatsuka Y, Egashira K, Katsuda Y, Narishige T, Ueno H, Shimokawa H, and Takeshita A. ATP sensitive potassium channels are involved in adenosine A2 receptor mediated coronary vasodilatation in the dog. Cardiovasc Res 28: 906–911, 1994.[Abstract/Free Full Text]
- Alexander RW, Kent KM, Pisano JJ, Keiser HR, and Cooper T. Regulation of postocclusive hyperemia by endogenously synthesized prostaglandins in the dog heart. J Clin Invest 55: 1174–1181, 1975.[ISI][Medline]
- Altman JD, Kinn J, Duncker DJ, and Bache RJ. Effect of inhibition of nitric oxide formation on coronary blood flow during exercise in the dog. Cardiovasc Res 28: 119–124, 1994.[Abstract/Free Full Text]
- Amezcua JL, Palmer RMJ, de Souza BM, and Moncada S. Nitric oxide synthesized from L-arginine regulates vascular tone in the coronary circulation of the rabbit. Br J Pharmacol 97: 1119–1124, 1989.[ISI][Medline]
- Amoah-Apraku B, Xu J, Lu JY, Pelleg A, Bruns RF, and Belardinelli L. Selective potentiation by an A1 adenosine receptor enhancer of the negative dromotropic action of adenosine in the guinea pig heart. J Pharmacol Exp Ther 266: 611–617, 1993.[Abstract/Free Full Text]
- Anfossi G, Russo I, Massucco P, Mattiello L, Cavalot F, Balbo A, and Trovati M. Adenosine increases human platelet levels of cGMP through nitric oxide: possible role in its antiaggregating effect. Thromb Res 105: 71–78, 2002.[CrossRef][ISI][Medline]
- Angello DA, Berne RM, and Coddington NM. Adenosine and insulin mediate glucose uptake in normoxic rat hearts by different mechanisms. Am J Physiol Heart Circ Physiol 265: H880–H885, 1993.[Abstract/Free Full Text]
- Ashcroft SJ and Ashcroft FM. Properties and functions of ATP-sensitive K-channels. Cell Signal 2: 197–214, 1990.[CrossRef][ISI][Medline]
- Auchampach JA and Gross GJ. Adenosine A1 receptors, KATP channels, and ischemic preconditioning in dogs. Am J Physiol Heart Circ Physiol 264: H1327–H1336, 1993.[Abstract/Free Full Text]
- Aversano T, Ouyang P, and Silverman H. Blockade of the ATP-sensitive potassium channel modulates reactive hyperemia in the canine coronary circulation. Circ Res 69: 618–622, 1991.[Abstract/Free Full Text]
- Awolesi MA, Sessa WC, and Sumpio BE. Cyclic strain upregulates nitric oxide synthase in cultured bovine aortic endothelial cells. J Clin Invest 96: 1449–1454, 1995.[ISI][Medline]
- Bacchus AN, Ely SW, Knabb RM, Rubio R, and Berne RM. Adenosine and coronary blood flow in conscious dogs during normal physiological stimuli. Am J Physiol Heart Circ Physiol 243: H628–H633, 1982.[Abstract/Free Full Text]
- Bache RJ, Dai XZ, Herzog CA, and Schwartz JS. Effects of nonselective and selective 1-adrenergic blockade on coronary blood flow during exercise. Circ Res 61: II36–II41, 1987.[Medline]
- Bache RJ, Dai XZ, Schwartz JS, and Homans DC. Role of adenosine in coronary vasodilation during exercise. Circ Res 62: 846–853, 1988.[Abstract/Free Full Text]
- Bassingthwaighte JB. Interstitial adenosine: the measurement, the interpretation. J Mol Cell Cardiol 24: 337–346, 1992.[CrossRef][ISI][Medline]
- Belardinelli L, Linden J, and Berne RM. The cardiac effects of adenosine. Prog Cardiovasc Dis 32: 73–97, 1989.[CrossRef][ISI][Medline]
- Belloni FL and Hintze TH. Glibenclamide attenuates adenosine-induced bradycardia and coronary vasodilation. Am J Physiol Heart Circ Physiol 261: H720–H727, 1991.[Abstract/Free Full Text]
TOP
ABSTRACT
OXYGEN AND CARBON DIOXIDE
