INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ENDOTHELIUM-DERIVED NITRIC OXIDE (NO) causes activation of soluble guanylyl cyclase in vascular smooth muscle. The resultant increase in cGMP and cGMP-dependent protein kinases causes vasodilation through modulation of calcium channels and by decreasing the calcium sensitivity of the vascular smooth muscle contractile proteins (24). The response to activation of guanylyl cyclase is terminated by several cyclic nucleotide phosphodiesterases that inactivate cGMP. Phosphodiesterase type 5 (PDE5), which selectively degrades cGMP but does not catabolize cAMP (18, 36), has recently received increased attention because of the availability of sildenafil, a selective inhibitor of PDE5, for treatment of erectile dysfunction (5). PDE5 is found in high concentration in the corpus cavernosum; sildenafil increases cGMP levels and causes smooth muscle relaxation (5). PDE5 is also found in the vascular smooth muscle of epicardial coronary arteries (36). In isolated coronary artery segments, sildenafil has been reported to cause increased intracellular levels of cGMP and relaxation (36). However, the effect of sildenafil on coronary resistance vessels has not been reported. Because the increased coronary blood flow rates during exercise or reactive hyperemia would be expected to cause an increase of shear-mediated endothelial NO production, inhibition of PDE5 might amplify the coronary vasodilation associated with these high-flow states. Consequently, the purpose of this study was to determine whether PDE5 inhibition with sildenafil causes an increase of the coronary resistance vessel dilation, which occurs in response to exercise or during the reactive hyperemia after a brief coronary artery occlusion.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Studies were performed on seven adult mongrel dogs (20-26 kg/wt) trained to run on a motor-driven treadmill. All experiments were performed in accordance with the "Guiding Principles in the Care and Use of Laboratory Animals" as approved by the Council of the American Physiological Society and with prior approval of the Animal Care Committee of the University of Minnesota.

Surgical preparation. The dogs were anesthetized with pentobarbital sodium (30-35 mg/kg), intubated, and ventilated with a mixture of oxygen (30%) and room air (70%). Respiratory rate and tidal volume were set to keep arterial blood gases within the physiological range. A left thoracotomy was performed in the fifth intercostal space, and the heart was suspended in a pericardial cradle. A polyvinyl chloride catheter measuring 3.0 mm internal diameter (ID) filled with heparinized saline was inserted into the internal thoracic artery and advanced into the ascending aorta. A similar catheter was introduced into the left ventricle through the apex and secured in place. A solid-state micromanometer (model P5, Konigsberg Instruments, Pasadena, CA) was also introduced into the left ventricle at the apex. A final catheter was introduced into the right atrial appendage, manipulated into the coronary sinus ostium, and advanced into the great cardiac vein until the catheter tip could be palpitated within 1 cm of the interventricular sulcus. This method allowed selective sampling of coronary venous blood draining the myocardium perfused by the left anterior descending coronary artery (LAD). Approximately 1.5 cm of the proximal LAD was dissected free, and a Doppler velocity probe (Craig Hartley, Houston, TX) was positioned around the artery. Immediately distal to the velocity probe, a hydraulic occluder was placed around the vessel. A silicone catheter (0.3 mm ID) bonded to a larger silicone catheter (1.6 mm ID) was introduced into the LAD immediately distal to the hydraulic occluder. The pericardium was then loosely closed, and the catheters and electrical leads were tunneled subcutaneously to exit at the base of the neck. The chest was closed in layers, and the pneumothorax was evacuated. Catheters were flushed daily with heparinized saline. The catheters, electrical leads, and occluder tubing were protected with a nylon vest. Postoperative analgesia was provided with the use of butorphanol 0.4 mg/kg sc every 4-6 h.

Experimental protocol. Studies were performed ~10 days after surgery for measurement of systemic hemodynamics and coronary blood flow at rest and during exercise before and after administration of sildenafil. When the dogs stood quietly on the treadmill, resting hemodynamics and coronary blood flow were recorded, and 2 ml of blood were withdrawn anaerobically from the aortic and coronary venous catheters and maintained on ice until blood gas analysis could be performed (within 30 min after sample collection). Subsequently, a three-stage treadmill exercise protocol was begun (stage 1: 4.8 km/h at 0% grade; stage 2: 6.4 km/h at 0% grade; and stage 3: 6.4 km/h at 10% grade). Each exercise stage lasted 3 min. Left ventricular (LV) and aortic pressures and coronary blood flow were measured continuously. Aortic and coronary venous blood samples were withdrawn during the last 30 s of each exercise stage when hemodynamics had reached a steady state. After the dogs were allowed a rest of a minimum of 20 min after completion of exercise so that heart rate, aortic, and LV pressures had returned to the preexercise resting control values, the reactive hyperemic responses to coronary occlusions 10 s in duration were recorded in duplicate; a 3-min interval was allowed between occlusions. Subsequently, the dogs were given an oral dose (2 mg/kg) of sildenafil. One hour after sildenafil administration, hemodynamic measurements and coronary blood flow were recorded during resting conditions and the three-stage exercise protocol was then repeated as described above. After the dogs rested a minimum of 20 min after completion of exercise, we repeated the reactive hyperemic responses to coronary occlusions were repeated as described above. To ensure that exercise had no effect on the reactive hyperemic response, the effect of sildenafil on reactive hyperemia after 10-s coronary artery occlusions was examined in three dogs on a separate day during which no exercise was performed. In three dogs, the effect of sildenafil was also examined on the reactive hyperemia after 5- and 20-s coronary artery occlusions. At the conclusion of the study, 3 ml of blood were withdrawn from the aortic catheter and immediately centrifuged at 3,000 rpm for 10 min at 4°C. The plasma was frozen at 70°C for determination of sildenafil concentration using high-performance liquid chromatography with mass spectrometric detection (35).

Endothelium-dependent vasodilation. In six dogs the effect of sildenafil on endothelium-dependent NO-mediated vasodilation was examined. Hemodynamic measurements and coronary blood flow were obtained with the dogs standing quietly in a sling. The increases in coronary blood flow produced by intracoronary infusion of acetylcholine (3.75 to 75 µg/min) was observed. After completion of these measurements, we administered sildenafil as an oral dose of 2 mg/kg. Thirty minutes after drug administration, coronary blood flow responses to intracoronary infusions of acetylcholine (3.75 to 75 µg/min) were repeated.

Hemodynamic measurements. Aortic pressure was measured with a fluid-filled pressure transducer positioned at midchest level. LV pressure was measured with the micromanometer calibrated with the fluid-filled LV catheter. The first time derivative of LV pressure (dP/dt) was obtained via electrical differentiation of the LV pressure signal. Coronary blood velocity was measured with a Doppler flowmeter system (Craig Hartley). Data were recorded on an eight-channel direct-writing oscillograph (Coulbourn Instruments, Lehigh Valley, PA).

Myocardial oxygen consumption. The blood specimens were maintained in iced syringes until the dogs completed each exercise trial. Measurements of PO₂, PCO₂, and pH were then immediately performed with a blood gas analyzer (model 113, Instrumentation Laboratory, Lexington, MA). Hemoglobin content was determined by the cyanmethemoglobin method. Hemoglobin oxygen saturation was calculated from the blood PO₂, pH, and temperature by use of the oxygen dissociation curve for canine blood. Blood oxygen content was computed as (hemoglobin × 1.34 × % O₂ saturation) + (0.0031 × PO₂). Oxygen consumption in the region of myocardium perfused by the LAD was calculated as the product of myocardial blood flow and the difference in oxygen content between aortic and coronary venous blood.

Data analysis. Heart rate, LV and aortic pressures, and coronary velocity were measured from the strip-chart recordings. In five of the dogs, the coronary Doppler velocity was calibrated using an injection of 15-µm diameter microspheres labeled with ¹⁴¹Ce, ⁵¹Cr, or ⁹⁵Nb (NEM, Boston, MA), while the coronary Doppler signal was simultaneously recorded. At the conclusion of the study, duplicate myocardial specimens were obtained from the region of LV perfused by the anterior descending coronary artery for radioactive counting and determination of blood flow per gram of myocardium as previously described (15). A calibration factor calculated as blood flow per gram of myocardium determined with microspheres per coronary Doppler shift was then used to convert the coronary velocity to blood flow per gram of myocardium throughout the study. In the remaining two dogs, coronary blood flow (Q) was computed from the Doppler shift using the equation Q = 2.5 × f × d², where f is the Doppler shift (kHz) and d is the diameter of the coronary artery (3). In these dogs, the mass of myocardium perfused by the LAD was obtained by multiplying LV weight × 0.43, which has been reported to represent the average fraction of LV perfused by the LAD (31). Blood flow per gram of myocardium was then determined as LAD coronary artery flow per LAD perfused myocardium. Total LAD blood flow during reactive hyperemia was determined by electrical integration of the Doppler velocity tracing. Reactive hyperemia flows were calculated as follows: blood flow debt (ml) = control blood flow rate (ml/s) × duration of occlusion (s); excess reactive hyperemia flow (ml) = total flow during reactive hyperemia (ml)  [control blood flow rate (ml/s) × duration of reactive hyperemia (s)]; blood flow debt repaid during reactive hyperemia = excess reactive hyperemia flow (ml)/blood flow debt (ml). The duration of reactive hyperemia was taken as the time from release of the coronary occlusion to the point at which flow returned to within 5% of control.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Systemic hemodynamics. The hemodynamic responses to graded exercise during control conditions and after administration of sildenafil are shown in Table 1. During control conditions exercise caused significant increases in heart rate, LV systolic pressure, and LV dP/dt_max. Mean aortic pressure was significantly increased during exercise stage 3 compared with resting conditions. After administration of sildenafil, resting heart rate, LV systolic pressure, LV dP/dt_max and the rate-pressure product were unchanged. Two-way ANOVA testing for the effect of sildenafil considering rest and all three exercise stages simultaneously demonstrated that mean aortic pressure was significantly less after sildenafil treatment; however, this difference was not significant when rest and each exercise stage were tested individually (Table 1). After sildenafil exercise caused significant increases in heart rate, LV systolic pressure, LV dP/dt_max, and the rate-pressure product that were not different from those during control conditions.

Coronary hemodynamics. Coronary responses to graded treadmill exercise during control conditions and with sildenafil are shown in Table 1 and Fig. 1. Exercise caused significant increases of myocardial blood flow and oxygen consumption and a significant decrease in coronary venous oxygen tension. During control conditions exercise caused a significant increase of hemoglobin from 13.1 ± 0.7 g/dl at rest to 14.6 ± 0.5 g/dl during the heaviest level of exercise (P < 0.05). Sildenafil caused a significant decrease of hemoglobin at rest and during exercise (P < 0.05 by two-way ANOVA), but did not interfere with the normal increase of hemoglobin from rest to exercise (hemoglobin after sildenafil increased from 12.5 ± 0.8 g/dl at rest to 13.9 ± 0.6 g/dl during the heaviest level of exercise; P < 0.05). Myocardial blood flow tended to be higher after sildenafil, and this achieved statistical significance when rest and all three exercise stages were considered simultaneously by using two-way ANOVA (mean increase = 0.26 ± 0.05 ml · min¹ · g¹; P < 0.05). However, the slightly higher myocardial blood flow rates after sildenafil were almost exactly matched by the slightly lower hemoglobin, so that the relationship between coronary oxygen delivery (myocardial blood flow × arterial oxygen content) and myocardial oxygen consumption was unchanged by sildenafil (Fig. 2). Coronary venous oxygen tension was not significantly altered by sildenafil (Fig. 3).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Relationship between myocardial blood flow and myocardial oxygen consumption at rest and during graded treadmill exercise. Data were obtained during control conditions and 60 min after oral administration of sildenafil, 2 mg/kg. Values are means ± SE; n = 7 dogs.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Relationship between oxygen delivery (arterial oxygen content × myocardial blood flow) plotted against myocardial oxygen consumption at rest and during graded treadmill exercise. Data were obtained during control conditions and 60 min after oral administration of sildenafil, 2 mg/kg. Values are means ± SE; n = 7 dogs.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Relationship between coronary venous (CV) oxygen tension and myocardial oxygen consumption at rest and during graded treadmill exercise. Data were obtained during control conditions and 60 min after oral administration of sildenafil, 2 mg/kg. Values are means ± SE; n = 7 dogs.

Reactive hyperemia. During control conditions, a 10-s coronary occlusion was followed by a reactive hyperemia 40 ± 9.7 s in duration, during which a peak flow rate of 114 ± 9.7 ml/min was achieved, resulting in 424 ± 105% debt repayment. Sildenafil did not significantly alter debt repayment (361 ± 110%), the peak blood flow rate during reactive hyperemia (111 ± 9.6 ml/min), or the duration of reactive hyperemia (38 ± 9.8 s). In three dogs, in which reactive hyperemia after a 5-s coronary occlusion was examined on days when no exercise was performed, sildenafil also had no effect on the reactive hyperemia (control debt repayment = 322 ± 16%; debt repayment after sildenafil = 316 ± 24%). In three dogs, in which reactive hyperemia after a 10-s coronary occlusion was examined on days when no exercise was performed, sildenafil also had no effect on the reactive hyperemia (control debt repayment = 325 ± 19%; debt repayment after sildenafil = 307 ± 44%). In the three dogs in which 20-s coronary occlusions were performed, sildenafil did not alter the reactive hyperemia (control debt repayment = 305 ± 22%; debt repayment after sildenafil = 296 ± 18%).

Endothelium-dependent vasodilation. Intracoronary infusions of acetylcholine in doses of 3.75 to 75 µg/min had no effect on blood pressure (105 ± 6.0 mmHg) or heart rate (117 ± 5.5 beats/min). During control conditions coronary blood flow increased from 48.6 ± 4.4 ml/min at baseline to 130 ± 10.6 ml/min during infusion of acetylcholine at a dose of 75 µg/min (Fig. 4). Sildenafil did not significantly change heart rate or mean arterial pressure. However, the increase of coronary blood flow produced by intracoronary acetylcholine after sildenafil was significantly augmented (Fig. 4). Thus during control conditions an increase in coronary blood flow of 80 ml/min required a dose of acetylcholine of 61.2 µg/min, whereas after sildenafil an 80 ml/min increase in coronary blood flow was produced by a dose of acetylcholine of 11.3 µg/min.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Dose-response curve demonstrating the increase of left anterior descending coronary artery blood flow produced by graded intracoronary infusions of acetylcholine during control conditions and 60 min after oral administration of sildenafil, 2 mg/kg. Values are means ± SE; n = 7 dogs. *P < 0.05 comparing control vs. sildenafil.

Sildenafil plasma levels. Plasma sildenafil levels ranged from 383 to 789 nM (mean = 547 ± 77 nM). Because sildenafil is 84% bound to canine plasma protein (35), this represents a mean plasma-free sildenafil concentration of 87.5 ± 12.4 nM.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, PDE5 inhibition with sildenafil augmented the endothelium-dependent coronary resistance vessel dilation produced by acetylcholine. After sildenafil, coronary blood flow was slightly higher at rest and during exercise, but this was associated with a slightly lower hemoglobin. Thus sildenafil produced no change in the relationship between myocardial oxygen availability and myocardial oxygen consumption, or in coronary venous oxygen tension. Furthermore, sildenafil did not augment the reactive hyperemia that followed a brief coronary artery occlusion. These findings indicate that inhibition of PDE5 augments coronary resistance vessel dilation produced by an endothelium-dependent NO-mediated agonist but does not augment the resistance vessel dilation that occurs during the increased endothelial shear associated with the high coronary blood flow rates of exercise or reactive hyperemia. The implications of these findings will be discussed below.

Effects of cGMP on coronary vasomotion. NO and sildenafil act on a common cGMP signaling pathway; one has the ability to increase cGMP synthesis and the other to elevate smooth muscle cGMP levels by selectively blocking cGMP degradation. As a result, the effect of PDE5 inhibition on coronary vessels might be expected to mimic the effect of NO donors. Therefore, it is of interest to review studies examining the effects of NO on the coronary circulation. Previous investigators have demonstrated that stimulation of endogenous coronary NO production (as with acetylcholine) (16, 20), intra-arterial infusion of authentic NO (8), or administration of NO donors (11) resulted in significant increases of coronary blood flow. Despite this evidence that NO can cause vasodilation of coronary resistance vessels, NO synthase inhibition with N^G-nitro-L-arginine (L-NNA), N^G-monomethyl-L-arginine (L-NMMA), or N^G-nitro-L-arginine methyl ester (L-NAME) did not decrease coronary blood flow in anesthetized (27) or awake (3, 4, 37) dogs, indicating that NO is not critical for maintaining coronary flow during basal conditions. This is in agreement with previous reports that coronary NO production is undetectable in vivo during resting conditions and that the cGMP content of unstimulated endothelial cells studied in vitro is very low. However, coronary NO production has been shown to increase during exercise, likely due to increased endothelial shear (4) and sympathetic nervous system activation of endothelial NO synthase (eNOS) by stimulation of endothelial ₂- and ₂-adrenoceptors (15). Nevertheless, blockade of NO production with NO synthase inhibitors did not impair the increase in coronary flow during treadmill exercise in normal dogs (3, 4). Furthermore, the relation between myocardial oxygen consumption and coronary flow was not altered by L-NNA, indicating that inhibition of NO production did not interfere with metabolic regulation of coronary vasomotor tone. Similarly, in the present study, blockade of cGMP degradation with sildenafil did not cause coronary resistance vessel dilation out of proportion to myocardial oxygen requirements. The apparent discrepancy between the demonstrated ability of NO to cause resistance vessel dilation in the heart and the failure of either decreasing coronary vascular cGMP levels by blocking NO synthesis or increasing cGMP levels by inhibiting PDE5 to alter coronary flow relative to myocardial oxygen demands may be explained by reciprocal vasomotor adjustments at several levels in the coronary microcirculation.

PDE isoenzymes in vascular smooth muscle. Previous studies using the PDE5 inhibitors zaprinast (23) or E4021 (29) demonstrated an increase in cGMP levels in isolated coronary artery segments and dose-dependent dilation of the large epicardial coronary arteries in awake pigs (1, 29). Similarly, Wallis and co-workers (36) reported that sildenafil resulted in a significant increase of cGMP concentration in isolated canine coronary artery segments in vitro. In the present study, PDE5 inhibition with sildenafil produced only a slight, although significant, augmentation of coronary resistance vessel dilation in response to exercise. Failure of sildenafil to produce a greater increase of myocardial blood flow may have been the result of alternate pathways for degradation of cGMP. In addition to PDE5, four other PDE enzymes have been identified in vascular smooth muscle (26, 36). Vascular smooth muscle cGMP-hydrolyzing activity is mainly due to PDE1, a calmodulin-dependent PDE, and PDE5, which is a calcium-calmodulin-independent cGMP-specific PDE (2). Of the total cGMP hydrolyzing activity, PDE1 constituted 73% in porcine and ~80% in bovine coronary artery (2), suggesting that PDE1 would play an important role in regulating cGMP levels. Vinpocetine, a selective inhibitor of PDE1, has been shown to cause concentration-dependent relaxation of isolated arterial vessel segments (13). Increased levels of cAMP in vascular smooth muscle can also cause vasodilation. cAMP hydrolyzing activity is mainly due to PDE4, a cAMP-specific PDE, and PDE3, a cGMP-inhibited PDE (more abundant than PDE4). Selective PDE3 inhibitors have been shown to produce concentration-dependent relaxation of isolated canine coronary artery segments (14). Increased cGMP activity has the potential to enhance the effects of cAMP by inhibiting PDE3 activity. However, Wallis et al. (36) reported that the increase in cGMP in vascular smooth muscle caused by sildenafil did not produce a change in cAMP. PDE4, the cAMP-specific PDE, is insensitive to cGMP. In contrast to PDE3, PDE4 inhibitors used alone caused only weak relaxation of blood vessels (21), suggesting that PDE4 is less important than PDE3 in regulating vasomotor tone under physiological conditions. PDE2 can hydrolyze both cAMP and cGMP but is present only in very small amounts in vascular smooth muscle. Sildenafil has a high affinity for PDE5 and a low affinity for the other PDE isoenzymes in vascular smooth muscle (36). Thus the IC₅₀ for inhibition of human PDE1, PDE2, PDE3, PDE4, and PDE5 was (in nM) 280, 6,800, 16,200, 7,200, and 3.5, respectively, indicating that sildenafil is highly selective for PDE5 (36). The mean plasma-free sildenafil concentration of 87.5 ± 12.4 nM in the present study would have provided a high degree of blockade of PDE5 with relatively little inhibition of the other PDE isoenzymes. The modest effect of sildenafil on myocardial blood flow in the present study may have occurred because sildenafil inhibits principally PDE5, with much less effect on PDE1, which provides an alternate pathway for the degradation of cGMP.

Effect of sildenafil on reactive hyperemia. In the present study, sildenafil had no effect on the reactive hyperemia that followed a brief coronary occlusion. In contrast, in isolated guinea pig hearts (22), open-chest dogs (39), and awake dogs (3, 7), NO synthase inhibition with L-NNA (3), L-NMMA (39), or L-NAME (22) decreased total reactive hyperemia flow principally by attenuating the late phase of the hyperemia response. Despite this previous evidence that NO contributes to coronary reactive hyperemia, in the present study, sildenafil did not result in an increase of total reactive hyperemia flow, the peak flow rate, or the duration of reactive hyperemia. Failure of sildenafil to augment reactive hyperemia suggests that an alternate pathway, such as PDE 1, can effectively degrade cGMP after blockade of PDE 5 (36). Alternatively, it is possible that metabolic vasomotor adjustments at the arteriolar level are able to counter any increase in cGMP-mediated vasodilation of the coronary arteries.

Effect of NO and sildenafil on myocardial oxygen consumption. Several investigators have suggested that NO can contribute to regulation of myocardial oxygen consumption. Hintze and co-workers (4, 32) reported that blockade of NO production in vivo and in vitro (32) led to significant increases in myocardial and skeletal muscle oxygen consumption. Conversely, stimulating NO-endogenous production with bradykinin or ramiprilat, or administering an NO donor, resulted in decreases in oxygen consumption in primate myocardial tissue (12). However, studies from other laboratories have reported that blockade of NO synthesis resulted in a decrease (33) or no change (10, 28) in myocardial oxygen consumption. In the present study, sildenafil had no significant effect on myocardial oxygen consumption, or on the increase in oxygen consumption that occurred in response to exercise. Sildenafil had no effect on LV dP/dt_max, in agreement with a previous report demonstrating that sildenafil had no effect on contractility in isolated canine right ventricular trabecular muscle (36). These results suggest that this dose of sildenafil had negligible effects on the PDE isoenzymes that catabolize cAMP in myocardial myocytes. Sildenafil did cause a modest decrease of aortic pressure with no change in heart rate. Similarly, in normal men intravenous sildenafil (40-200 mg) produced transient decreases in systolic and diastolic blood pressures of 10 and 7 mmHg, respectively, with no change in heart rate (17, 40).