Heart and Circulatory Physiology

Coronary blood flow regulation in exercising swine involves parallel rather than redundant vasodilator pathways

Daphne Merkus, David B. Haitsma, Tse-Yeung Fung, Yvette J. Assen, Pieter D. Verdouw, Dirk J. Duncker


In dogs, only combined blockade of vasodilator pathways [via adenosine receptors, nitric oxide synthase (NOS) and ATP-sensitive K+ (KATP) channels] results in impairment of metabolic vasodilation, which suggests a redundancy design of coronary flow regulation. Conversely, in swine and humans, blocking KATP channels, adenosine receptors, or NOS each impairs coronary blood flow (CBF) at rest and during exercise. Consequently, we hypothesized that these vasodilators act in parallel rather than in redundancy to regulate CBF in swine. Swine exercised on a treadmill (0–5 km/h), during control and after blockade of KATP channels (with glibenclamide), adenosine receptors [with 8-phenyltheophylline (8-PT)], and/or NOS [with Nω-nitro-l-arginine (l-NNA)]. l-NNA, 8-PT, and glibenclamide each reduced myocardial O2 delivery and coronary venous O2 tension. These effects of l-NNA, 8-PT, and glibenclamide were not modified by simultaneous blockade of the other vasodilators. Combined blockade of KATP channels and adenosine receptors with or without NOS inhibition was associated with increased H+ production and impaired myocardial function. However, despite an increase in O2 extraction to >90% during administration of l-NNA + 8-PT + glibenclamide, vasodilator reserve could still be recruited during exercise. Thus in awake swine, loss of KATP channels, adenosine, or NO is not compensated for by increased participation of the other two vasodilator mechanisms. These findings suggest a parallel rather than a redundancy design of CBF regulation in the porcine circulation.

  • coronary circulation
  • vasoconstriction
  • dilation
  • nitric oxide
  • adenosine
  • ATP-sensitive K+ channel

coronary blood flow (CBF) is tightly coupled to myocardial O2 consumption ( MV̇O2) to maintain a consistently high level of myocardial O2 extraction (MEo2; Refs. 18, 31). This tight coupling has been proposed to depend on periarteriolar O2 tension (Po2), signals released from cardiomyocytes (e.g., adenosine), and/or the endothelium [e.g., nitric oxide (NO)], but the contributions of each of these regulatory pathways and their interactions are still incompletely understood. Studies on the canine coronary circulation indicate that ATP-sensitive K+ (KATP) channels contribute to basal tone in coronary resistance vessels but are not mandatory for the increase in CBF that occurs during increased myocardial metabolic demands such as exercise (11, 42). Adenosine and NO are not mandatory for CBF regulation either at rest or during exercise but instead mediate the coronary vasodilation in response to exercise when KATP channels are blocked (13, 24). These findings suggest a regulatory design in the coronary circulation of the dog in which there is a redundancy in vasodilator pathways that can mediate coronary vasodilation when one system fails (5).

In contrast, in human and porcine coronary circulation, single blockade of adenosine (10, 14, 15), NO (8, 41), or KATP channels (7, 17) results in impaired coronary arterial inflow both at rest and during increased metabolic demands produced by exercise. An increased MEo2 is thereby necessitated, which results in a decreased coronary venous Po2 (PvO2; Refs. 7, 8, 10). These findings suggest that in swine and humans, these vasodilator mechanisms may act in a parallel rather than a redundant fashion as occurs in dogs. If this is true, then combined blockade of these vasodilator pathways should produce additive increases in coronary vasomotor tone (Fig. 1A) rather than a synergistic effect (Fig. 1B). To test this hypothesis, we investigated the integrated contribution of KATP channels, adenosine, and NO to the regulation of CBF in swine at rest and during exercise.

Fig. 1.

Concept of parallel vs. redundant regulation of tone of coronary resistance vessels by multiple vasodilator pathways. A: in case of a parallel design, combined administration of inhibitors of vasodilator pathways A and B should produce an additive increase in coronary vasomotor tone. Thus the decrement in coronary venous O2 tension (CVPO2) in response to inhibition of vasodilator pathway A is independent of the presence of inhibition of vasodilator pathway B. B: in case of a redundancy design, combined administration of inhibitors of vasodilator pathways A and B should produce an amplified (synergistic) increase in coronary vasomotor tone. Thus the decrement in coronary venous Po2 (PvO2) in response to inhibition of vasodilator pathway A is enhanced in the presence of inhibition of vasodilator pathway B. MV̇O2, myocardial O2 consumption.


Eighteen 2–3-mo-old crossbred Landrace × Yorkshire swine (10 males and 8 females) were entered into the study. The investigation conformed with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health [DHEW Publication No. (NIH) 85-23, Revised 1996, Office of Science and Health Reports, DRR/NIH, Bethesda, MD 20205] and was performed with the approval of the Animal Care Committee of Erasmus Medical Center. Adaptation of animals to laboratory conditions started 1 wk before surgery.

Surgical Procedures

Swine were sedated with ketamine (30 mg/kg im), anesthetized with thiopental (10 mg/kg iv), intubated, and ventilated with a 1:2 mixture of O2/N2O to which 0.2–1% (vol/vol) of isoflurane was added (7, 9). Anesthesia was maintained with midazolam (2 mg/kg + 1 mg·kg-1·h-1 iv) and fentanyl (10 μg·kg-1·h-1 iv). Under sterile conditions, the chest was opened, and fluid-filled polyvinylchloride (PVC) catheters were inserted into the aortic arch, pulmonary artery, and left atrium for blood pressure measurement, blood sampling, and drug administration. A Konigsberg pressure transducer was inserted into the left ventricle for recording of left ventricular (LV) pressure and change in LV pressure with time (LV dP/dt), and a PVC catheter was inserted for calibration of the Konigsberg transducer signal (7, 9). An ultrasonic transit-time perivascular flow probe (2.5 or 3.0 mm, Transonic Systems) was placed around the proximal left anterior descending coronary artery (LAD) to measure CBF. A small angiocatheter was inserted into the anterior interventricular vein to allow sampling of coronary venous blood. Electrical wires and catheters were tunneled subcutaneously to the back, the chest was closed, and the animals were allowed to recover for 1 wk (7, 9).

Experimental Protocols

Studies were performed 10–20 days after surgery with animals exercising on a motor-driven treadmill. Four exercise protocols were performed on different days and in random order.

KATP channels. With swine (n = 9) lying quietly on the treadmill, resting hemodynamic measurements including LV pressure, LV dP/dt, mean aortic pressure (MAP), left atrial pressure (LAP), and CBF were obtained, and blood samples were collected. Hemodynamic measurements were repeated and rectal temperature was measured while animals stood on the treadmill. Subsequently, a five-stage (1, 2, 3, 4, and 5 km/h) exercise protocol was started. Hemodynamic variables were continuously recorded, and blood samples were collected during the last 30 s of each 2–3-min stage. Blood samples were maintained in iced syringes until the conclusion of each exercise trial. Measurements of Po2 (mmHg), Pco2 (mmHg), and pH were then immediately performed with a blood gas analyzer (model 505, Radiometer, Acid-Base Laboratory), and O2 saturation and hemoglobin content (Hb, in g/100 ml) were measured with a hemoximeter (OSM2, Radiometer; Refs. 9, 10). After completing the exercise protocol, swine were allowed to rest on the treadmill for 90 min. After resting, the animals received glibenclamide [3 mg/kg iv (7)], and the exercise protocol was repeated.

Adenosine and KATP channels. At 90 min after the swine (n = 8) had completed a control exercise trial, the animals received the adenosine receptor antagonist 8-phenyltheophylline [8-PT, 5 mg/kg iv (10, 24)] and started a second exercise trial; 90 min later, the animals received 8-PT [2.5 mg/kg iv (24)] and glibenclamide (3 mg/kg iv) and began a third exercise trial.

NO and adenosine. At 90 min after the swine (n = 10) had completed a control exercise trial, the animals received the NO synthase (NOS) inhibitor Nω-nitro-l-arginine [l-NNA, 20 mg/kg iv (8)] and began a second exercise trial; 90 min later, animals received 8-PT (5 mg/kg iv) and the exercise was repeated.

KATP channels, NO, and adenosine. Swine (n = 6) underwent an exercise trial in the presence of 8-PT (5 mg/kg iv) and l-NNA (20 mg/kg iv). Animals received 8-PT (2.5 mg/kg iv) and glibenclamide (3 mg/kg iv) 90 min later, and exercise was repeated.

Reproducibility of exercise response. In accordance with previously published data from our laboratory (9, 10), no significant differences in the exercise response were found when animals underwent two or three consecutive exercise trials with administration of saline before exercise periods 2 and 3 (data not shown).

Data Analysis

Hemodynamic data were digitally recorded and analyzed off-line for heart rate (HR), mean aortic blood pressure, mean left atrial blood pressure, and maximum rate of rise of LV pressure (LV dP/dtmax) (9, 10). Because alterations in LV dP/dtmax can be influenced by changes in HR and hemodynamic loading conditions, LV dP/dtmax has been presented as a function of the rate-pressure product (HR × LV systolic pressure).

Blood O2 content (in μmol/ml) was computed as (Hb × 0.621 × O2 saturation) + (0.00131 × Po2). Myocardial O2 delivery (MDo2) was computed as the product of LAD coronary blood flow and arterial blood O2 content; MV̇O2 in the region of myocardium perfused by the LAD was calculated as the product of coronary blood flow and the difference in O2 content between arterial and coronary venous blood. MEo2 was computed as the ratio of the arteriovenous O2 content difference and the arterial O2 content. Finally, myocardial H+ production was computed as the product of CBF and the difference between the H+ content (10-pH) of arterial and coronary venous blood. To correct for any alterations in MV̇O2, the parameters MDO2, MEO2, coronary PvO2, and H+ production are all presented as a function of MV̇O2.

ANOVA for repeated measures or ANCOVA was used as appropriate. Post hoc testing was performed using Dunnett's test (exercise effect) or unpaired t-test (drug effect). Statistical significance was accepted when P < 0.05. Because no differences were found between male and female swine, data from both sexes were pooled. Data are means ± SE.


Exercise resulted in increases in HR (up to 80% of maximum HR) and left atrial pressure, whereas mean aortic blood pressure was minimally affected (Fig. 2). The exercise-induced increase in MV̇O2 was accommodated by a commensurate increase in MDo2 [principally mediated by an increase in CBF (Fig. 2) as arterial Hb levels increased by only 10% (not shown)] so that MEo2 was maintained constant at ∼80% (Fig. 3). l-NNA and glibenclamide increased mean arterial blood pressure both at rest and during exercise; this was associated with a (probably baroreceptor reflex-mediated) marked decrease in HR that was accompanied by a slight decrease in MV̇O2 and a more marked decrease in CBF (Fig. 2). Although 8-PT had no effect on mean arterial blood pressure, it resulted in small increases in HR at rest and during exercise. These changes were probably due to blockade of adenosine receptors located on the sympathetic nerve terminals that act to inhibit the release of catecholamines. The increases in HR were associated with slight elevations in MV̇O2 during control conditions (P = 0.14) and in the presence of l-NNA (P < 0.05).

Fig. 2.

Systemic hemodynamics, coronary blood flow, and MV̇O2 in exercising swine. Glib, glibenclamide; 8-PT, 8-phenyltheophylline; (l-NNA), Nω-nitro-l-arginine; HR, heart rate; MAP, mean aortic blood pressure; LAP, left atrial pressure; CBF, coronary blood flow. *P < 0.05 vs. control; †P < 0.05 vs. 8-PT; ‡P < 0.05 vs. l-NNA; ¶P < 0.05 vs. l-NNA + 8-PT; §P = 0.08 vs. control.

Fig. 3.

Myocardial O2 balance and left ventricular (LV) contractile function in exercising swine. MDo2, myocardial O2 delivery; MEo2, myocardial O2 extraction; LV dP/dtmax, maximum rate of rise of LV pressure. Rate-pressure product = HR × LV systolic pressure. *P < 0.05 vs. control; †P < 0.05 vs. 8-PT; ‡P < 0.05 vs. l-NNA; ¶P < 0.05 vs. l-NNA + 8-PT.

l-NNA, 8-PT, and glibenclamide each caused coronary vasoconstriction at rest and during exercise. This limited MDo2 at a given level of MV̇O2 and necessitated an increase in MEo2, which thereby led to a decreased coronary PvO2 (Fig. 3). Combined administration of 8-PT + glibenclamide or l-NNA + 8-PT further impaired MDo2, which necessitated further increases in MEo2 and resulted in decreased coronary PvO2 (Fig. 3). The effects of glibenclamide during 8-PT administration were similar to those during control. This indicates that adenosine does not compensate for KATP channel blockade. Similarly, the effects of 8-PT were not modified by l-NNA, which indicates that adenosine does not compensate for NO inhibition. Finally, KATP channel blockade after l-NNA + 8-PT administration shifted the relation between MV̇O2 and MDo2 even further toward the line of identity with almost all (93 ± 1%) of the available O2 being extracted; the additional effects of KATP channel blockade were not statistically different from those during control or during 8-PT administration.

l-NNA, 8-PT, or the l-NNA + 8-PT combination did not affect the H+ production or the relation between rate-pressure product and LV dP/dtmax (Fig. 3). This indicates maintained aerobic metabolism and contractile function. Only glibenclamide produced an increase in H+ production during control, after 8-PT administration, and after l-NNA + 8-PT administration (Fig. 3). These glibenclamide-induced changes were accompanied by decreases in LV dP/dtmax particularly in the presence of 8-PT and l-NNA + 8-PT administration (Fig. 3).

The combination of l-NNA + 8-PT + glibenclamide significantly impaired the increase in CBF and hence MDo2 during exercise, which caused the relation between MV̇O2 and MDo2 to approach the line of identity (Fig. 4). Nonetheless, in the presence of l-NNA + 8-PT + glibenclamide, CBF still doubled in response to exercise. This indicates that despite near-maximal MEo2 values (93 ± 1%), metabolic vasodilator reserve could still be recruited to mediate the increase in CBF without aggravation of anaerobic metabolism and further impairment of contractile function.

Fig. 4.

Relation between MV̇O2 and MDO2 during control conditions and after blockade of nitric oxide synthesis, adenosine receptors, and KATP channels. Note that the relation approaches the line of identity after blockade (slope during control, 1.28 ± 0.03; slope after blockade, 1.08 ± 0.01; P < 0.05); *P < 0.05 vs. control.


The main findings in our study on awake swine are as follows: 1) the effects of l-NNA, 8-PT, and glibenclamide on MDo2 and MEo2 are independent of the patency of the other vasodilator systems studied, which suggests a design of CBF regulation that is parallel (additive) rather than redundant (backup) in nature; 2) combined blockade of KATP channels and adenosine receptors with or without NOS inhibition produces anaerobic metabolism accompanied by mild impairment of myocardial function; and 3) despite an increase in MEo2 to >90% at rest by combined l-NNA + 8-PT + glibenclamide, vasodilator reserve can still be recruited during exercise to mediate the residual exercise-induced increase in CBF.

Methodological Considerations

Coronary vascular tone and myocardial O2 balance. The normal heart is characterized by a high level (80%) of MEo2 under basal resting conditions (18, 31). Consequently, the ability of the coronary resistance vessels to dilate in response to increments in myocardial O2 demand is extremely important to maintain an adequate supply of O2. A sensitive way to study alterations in coronary vascular tone in relation to myocardial metabolism is to investigate the relationship between coronary PvO2 and MV̇O2 (18, 31). For example, an increase in coronary resistance vessel tone would limit CBF and hence restrict MDo2 and thereby force the myocardium to increase its MEo2 (to maintain MV̇O2), which results in a lower coronary PvO2. The coronary PvO2 thus represents an index of myocardial tissue oxygenation (i.e., the balance between MDo2 and MV̇O2) that is determined by the coronary resistance vessel tone.

Doses of drugs. The doses of 8-PT, glibenclamide, and l-NNA that we used in the present study were previously shown in our laboratory to effectively blunt the responses to exogenous agonists. Thus glibenclamide in a dose of 3 mg/kg iv attenuated the vasodilation produced by the KATP channel opener bimakalim (7), and l-NNA in a dose of 20 mg/kg iv blocked the NO-mediated vasodilation by ATP (8), whereas the vasodilation by nitroprusside remained unaltered in the presence of these antagonists. The dose of 8-PT (5 mg/kg) blocked the vasodilator response to intracoronary adenosine in anesthetized swine (6) as well as in awake dogs (3, 12). Furthermore, all three antagonists produced a significant increase in coronary resistance vessel tone, which indicates that the doses employed were effective.

Route of drug administration. The antagonists 8-PT, glibenclamide, and l-NNA were administered via the intravenous route to achieve stable blood levels of each antagonist in the coronary circulation. In contrast, intracoronary infusion at a constant infusion rate results in coronary artery blood concentrations that depend on the level of CBF. Thus when CBF increases during exercise, the blood concentrations of the antagonists decrease, which can result in a decreased efficacy of antagonism. Conversely, administration of inhibitors of vasodilator pathways via the intracoronary route is advantageous in that it is typically devoid of significant systemic hemodynamic effects (13, 24), whereas intravenous administration can result in marked hemodynamic changes that could potentially confound interpretation of the coronary PvO2 data.

For example, it could be argued that an increase in blood pressure increases MV̇O2 and thereby causes a decrease in coronary PvO2. Thus in dogs, the increase in MV̇O2 that occurs during exercise is associated with a decrease in coronary PvO2 (13, 18, 24, 31, 51). However, in swine, coronary PvO2 is remarkably stable over a wide range of MV̇O2 levels (710), whereas the antagonist-induced changes in MV̇O2 are small (likely because the decrease in HR balances the increase in aortic pressure). Moreover, by plotting coronary PvO2 as a function of MV̇O2, we would correct for any possible dependency of coronary PvO2 on the level of MV̇O2 (18, 31).

One could also argue that intravenous administration of antagonists limits myocardial O2 supply via modulation of the extravascular determinants of CBF and thereby decreases coronary PvO2. This is unlikely, however, because the increase in mean aortic pressure together with the decrease in HR would favor myocardial perfusion due to increased perfusion pressure and increased duration of diastolic perfusion time, respectively (18, 31). These hemodynamic changes would thus act to increase coronary PvO2, and therefore cannot explain the decreases in coronary PvO2. Importantly, during exercise, MEo2 and coronary PvO2 were maintained constant over a wide range of MV̇O2 levels despite marked changes in systemic hemodynamics.

Finally, it could be argued that the increase in aortic blood pressure influences autonomic nervous system activity (i.e., results in decreased sympathetic and increased parasympathetic activity), thereby influencing vasomotor tone in the coronary resistance vessels. However, we have previously shown that in resting swine, α- and β-adrenergic sympathetic control of coronary vasomotor tone is absent, which is caused by an inhibitory influence of the parasympathetic system (9). Consequently, under resting conditions where adrenergic control of coronary resistance vessel tone was already absent, a baroreceptor-mediated shift in autonomic balance toward further parasympathetic dominance would not have any effect on coronary PvO2. Furthermore, during exercise, MEo2 and coronary PvO2 were maintained remarkably constant over a wide range of MV̇O2 levels despite marked increments in sympathetic activity and withdrawal of parasympathetic activity (9). Therefore, any alteration in autonomic balance that was produced by the antagonists would not be expected to modify MEo2 and coronary PvO2.

Taken together, the effects of intravenous administration of l-NNA, 8-PT, and glibenclamide on MEo2 and coronary PvO2 do not appear to be related to alterations in systemic hemodynamics but rather represent direct effects on the coronary resistance vessels.

Myocardial O2 Balance During Exercise

In dogs, the increase in CBF (and hence MDo2) that occurs during exercise does not fully match the exercise-induced increase in MV̇O2, so that even during mild to moderate levels of exercise (<70% of maximum HR) an increase in MEo2 and hence a decrease in coronary PvO2 occurs (13, 18, 24, 31, 51). In contrast, in humans, minimal changes in MEo2 occur at mild to moderate levels of exercise although an increase in MEo2 and a decrease in coronary venous O2 content have been reported in humans during heavy exercise (>85% of maximum HR; Ref. 31). We previously showed that at mild to moderate levels of exercise, swine resemble humans more closely than dogs. However, in contrast to dogs and humans, swine also maintain a constant level of MEo2 and coronary PvO2 during heavy exercise, which indicates that the exercise-induced increase in MDo2 matches the increase in MV̇O2 (710). The decrease in coronary PvO2 that occurs in dogs during exercise is in part due to α-adrenergic vasoconstriction of coronary resistance vessels (18, 31), which is absent in swine (9) and has been proposed to represent a metabolic error signal that is necessary for the increase in CBF during exercise (18, 31). However, our studies on swine clearly indicate that a decrease in coronary PvO2 is not essential for the increase in CBF in the normal heart during heavy exercise.

Individual Vasodilator Pathways and Myocardial O2 Supply

Individual blockade of NOS with l-NNA or adenosine receptors with 8-PT results in a decrease of MDo2 that necessitates an increase in MEo2 and hence leads to a decrease in coronary PvO2 in swine at rest and during exercise (8, 9). However, neither vasodilator pathway is compulsory for the exercise-induced increase in MDo2 (8, 9); this agrees well with studies on humans (4, 14, 15, 32, 36, 40). In contrast, blocking NO synthesis (2, 24, 45) or adenosine receptors (3) does not affect CBF and MEo2 in dogs either at rest or during exercise, although this is not an unequivocal finding (49, 51).

Blocking KATP channels with glibenclamide results in a decreased MDo2, an increased MEo2, and a decreased coronary PvO2 in dogs (11, 42) and swine (7) both under resting conditions and during exercise. However, KATP channel blockade does not affect exercise-induced coronary vasodilation in either species (7, 11, 42). Thus although KATP channels clearly contribute to the regulation of tone in coronary resistance vessels, they are not mandatory for exercise-induced vasodilation.

The lack of effect of blocking a single vasodilator pathway on the exercise-induced coronary vasodilation could be interpreted to suggest that the pathway does not contribute to coronary vasodilation. Alternatively, blockade of one vasodilator mechanism could be compensated for by increased contribution of another pathway.

Integrated Control of Myocardial O2 Supply by Multiple Vasodilator Pathways

Adenosine and NO. A compensatory role for adenosine in the regulation of CBF when NOS is blocked is suggested by studies in isolated guinea pig hearts during reactive hyperemia (30), and in open-chest dogs during metabolic vasodilation produced by cardiac pacing (33). The decrease in CBF produced by NG-nitro-l-arginine methyl ester was exaggerated following pretreatment with 8-PT and resulted in regional myocardial contractile dysfunction and blunted lactate extraction (33). In contrast, in awake dogs, Ishibashi et al. (24) failed to observe any effect of 8-PT on MDo2 and coronary PvO2 in awake dogs pretreated with l-NNA either at rest or during exercise, which indicates that NO and adenosine either alone or in combination are not mandatory for CBF regulation at rest and during exercise. Also in awake dogs, coronary venous adenosine concentrations were unchanged following NOS inhibition (49). These findings suggest that in awake dogs that are free from anesthesia and acute surgical trauma, loss of NO is not compensated for by an adenosine-mediated vasodilator influence either at rest or during exercise.

The present study shows that both NO and adenosine exert vasodilator influences in the coronary resistance vessels of awake swine. However, loss of NO- or adenosine-mediated vasodilator influence was not compensated for by an increased vasodilator influence via the other pathway, as the effects of l-NNA and 8-PT were additive. Thus the effect of blockade of adenosine receptors on MDo2 and MEo2 appears to be independent of NOS activity. In isolated porcine coronary arterioles, exogenously administered adenosine produced vasodilation that was in part mediated via NO (22, 23). Although our data suggest that endogenous adenosine acts in parallel and not via NOS in the porcine heart in vivo, we cannot exclude the possibility that, after NOS inhibition, adenosine levels increased sufficiently to maintain the vasodilatory influence via direct actions on the vascular smooth muscle and that this occurred possibly through KATP (23) or voltage-dependent K+ (Kv) channels (22). The decreases in MDo2 and coronary PvO2 produced by combined adenosine receptor blockade and NOS inhibition were not associated with evidence of anaerobic metabolism or loss of LV contractile function. This suggests that the increase in MEo2 could fully compensate for the impediment in coronary arterial inflow.

Adenosine and KATP channels. A compensatory role for adenosine in the regulation of CBF when KATP channels are blocked was first demonstrated in dogs (13, 43). After KATP channel blockade with intracoronary glibenclamide (50 μg·kg-1·min-1), which produced a 20% decrease in CBF and a 50% decrease in regional wall thickening (11, 13), adenosine receptor blockade resulted in a significant additional decrease in CBF and regional wall thickening in particular during exercise (13). In contrast, Richmond et al. (42) found that administration of an intravenous dose of 1 mg/kg glibenclamide, which had no effect on lactate extraction, did not affect coronary venous adenosine concentrations (P = 0.11) either at rest or during exercise. In that study, the dose of glibenclamide may have been too low, as the authors stated that a higher dose resulted in CBF oscillations. The latter was shown by Samaha et al. (43) to be due to cyclic release of adenosine that resulted from ATP breakdown, i.e., the presence of myocardial ischemia. In the present study, glibenclamide in an intravenous dose of 3 mg/kg (resulting in a free arterial plasma concentration of 1–2 μmol; Ref. 7) produced an increase in H+ production and a small decrease in LV contractile function, which may reflect incipient myocardial ischemia.

The effects of simultaneous blockade of KATP channels and adenosine receptors on MDo2 and MEo2 were approximately equal to the sum of the effects of the individual antagonists. This indicates that loss of KATP channel activity was not compensated for by an increased adenosine-mediated vasodilator influence. These findings suggest that adenosine and KATP channels act independently, i.e., in parallel, in swine. Adenosine may mediate its vasodilator effect on porcine coronary resistance vessels via both KATP (23) and Kv channels (22). Therefore, it is possible that following KATP channel blockade, adenosine levels increased sufficiently to maintain the vasodilatory influence via Kv channels. Conversely, it is possible that, after adenosine receptor blockade, KATP channel activity was maintained via alternate mediators such as prostacyclin (16, 26, 44).

In the presence of adenosine receptor blockade, additional KATP channel blockade resulted in increased H+ production and a more marked decrease in global LV function, which is consistent with the presence of myocardial ischemia. This suggests that the further impairment of MDo2 by combined blockade of adenosine receptors and KATP channels compared to KATP channel blockade alone was associated with development of more significant myocardial ischemia. These results are in agreement with studies on dogs in which adenosine receptor blockade amplified the KATP channel blockade-induced impairment of regional myocardial wall motion (13).

Adenosine, KATP channels, and NO. In awake dogs, inhibition of NOS alone or in combination with adenosine receptor blockade does not affect the relation between MV̇O2 and coronary PvO2 (24). However, combined blockade of adenosine receptors, NOS, and KATP channels markedly (50%) reduced CBF at rest and virtually abolished the exercise-induced coronary vasodilation (24). Those findings suggest that metabolic dilation of canine coronary resistance vessels is regulated via a complex of vasodilator systems that act in concert to match CBF and hence MDo2 to MV̇O2, so that when one system fails, backup systems take over to ensure adequate blood supply to the myocardium. The observation that only KATP channel blockade decreased CBF suggests that the principal coronary vasodilator pathway in the dog is the KATP channel with adenosine and NO primarily acting as backup systems (13, 24).

In contrast, in the present study on swine, the effect of blockade of KATP channels was not enhanced by adenosine receptor blockade and NOS inhibition. Metabolic regulation of CBF in swine appears therefore to be mediated through parallel pathways. Thus blocking these pathways simultaneously results in an additive rather than a synergistic (as is the case in dogs) effect on MDo2. After blockade of adenosine receptors and NOS, the glibenclamide-induced decrease in coronary PvO2 in resting swine tended to be somewhat less compared with the effects under control conditions and during adenosine receptor blockade (P = 0.12). This might reflect the vasodilation by NO and adenosine through KATP channels (23). The data must be interpreted with some caution, as MEo2 exceeded 90% and thereby reached the upper limit of MEo2 in the porcine heart (29).

Myocardial H+ production and LV contractile function were maintained during combined blockade of adenosine receptors and NOS but deteriorated when KATP channel blockade was added. These findings are consistent with the data of Ishibashi et al. (24), who showed that combined blockade of adenosine receptors, KATP channels, and NOS inhibition resulted in a marked loss of systolic wall thickening in dogs. In contrast to Ishibashi et al. (24) and the present study, Tune et al. (50) reported an unchanged arteriovenous H+ difference when all three vasodilator systems were blocked. The main differences between our study and that of Tune et al. (50) are the doses of 8-PT [3 mg/kg (50) vs. 5 mg/kg in our study] and glibenclamide [1 mg/kg (50) vs. 3 mg/kg in our study]. In addition, we used consecutive incremental levels of exercise that resulted in a total duration of the exercise protocol of ∼15 min, whereas Tune et al. (50) waited between each exercise level for hemodynamic variables to return to control resting values; this likely resulted in a longer duration of the total exercise protocol. Because both glibenclamide and 8-PT are competitive inhibitors of KATP channels and adenosine receptors, respectively, the longer duration of the protocol in combination with the lower dosage of the inhibitors and the increased recruitment of these systems during exercise may have resulted in incomplete inhibition of these vasodilator systems particularly during higher levels of exercise and may thus have resulted in an underestimation of their contribution to metabolic regulation of CBF.

Possible Mediators of Residual Vasodilation

Despite a near-maximal MEo2 value (>90%) and evidence for anaerobic metabolism and impaired LV function at rest, CBF and MDo2 could still increase in response to exercise in the presence of l-NNA, 8-PT, and glibenclamide. Apparently, vasodilator mechanisms that cannot be recruited at rest can be recruited during exercise. Such candidate mechanisms include β-adrenoceptors, prostacyclin, H+ ions, and endothelium-derived hyperpolarizing factor (EDHF). Although we previously found that β-adrenergic vasodilation plays a role in exercise-induced vasodilation in swine (9), most β-adrenergic vasodilation is mediated through NO (38), adenosine (48), and KATP channels (35) and therefore should have been susceptible to our inhibitors. It is therefore unlikely that β-adrenergic vasodilation contributed to the increase in CBF during exercise after all three vasodilator systems were blocked. Prostacyclin is produced in the coronary circulation and contributes to basal coronary tone as well as metabolic dilation in humans (4). Also, prostacyclin has been shown to contribute to the regulation of basal vascular tone in swine (1) and to flow-induced dilation in rats (28). Consequently, an increase in prostacyclin production may have contributed to the increase in CBF during exercise. H+ ions result in vasodilation through both KATP (25) and Ca2+-dependent K+ (KCa) channels (21) as well as voltage-dependent Ca2+ channels (27). Although part of the effect of H+ ions would therefore be blocked by the administration of glibenclamide, vasodilation as a result of increasing production of H+ ions may still be present through the other pathways. Because, especially, glibenclamide increased the arteriovenous H+ difference, H+ ions could have mediated the exercise-induced vasodilation that persisted in the presence of blockade of adenosine receptors, KATP channels, and NOS. Alternatively, EDHF may play a role in exercise-induced vasodilation. EDHF has been shown to contribute to vasodilation to pulsatile stretch (39) as well as agonists such as substance P, acetylcholine, and bradykinin (16, 19, 20, 34, 37, 46). However, information about physiological stimuli of EDHF production is lacking; i.e., it is unknown whether EDHF production changes with changing frequency of the pulsatile pressure in vivo. Preliminary data from our laboratory indicate that blockade of EDHF production with sulfaphenazole, an inhibitor of cytochrome P-450, does not affect coronary blood flow at rest or during exercise. Moreover, the contribution of EDHF is per definition measured after inhibition of NO and thus little is known about the interaction between those substances and whether they act in an additive or redundant fashion.

In conclusion, the present study suggests that regulation of CBF and myocardial O2 supply by adenosine, NO, and KATP channels encompasses a parallel rather than a redundancy design in swine. Because substantial vasodilator reserve was still present when all three vasodilator pathways were blocked, our results also indicate that pathways other than adenosine, NO, and KATP channels must be involved in the regulation of CBF in swine. Our results differ markedly from observations on dogs, where blockade of these three vasodilator pathways fully abolished the exercise-induced increase in CBF. In view of our findings that indicate important species differences in coronary vasomotor tone regulation, extrapolation of results obtained in animal studies to the human heart should be done with caution.


The authors express gratitude to R. Stubenitsky, R. Hoogendoorn, and B. Houweling for technical assistance.

D. J. Duncker is the recipient of an Established Investigator Stipend (2000T038) and D. Merkus is the recipient of a Postdoctoral Stipend (2000T042) of The Netherlands Heart Foundation.


  • 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.


View Abstract