Hydrogen Peroxide Derived From Beating Heart Mediates Coronary Microvascular Dilation During Tachycardia
Objective— Coronary flow is closely correlated to the myocardial metabolic demand. We tested the hypothesis that hydrogen peroxide (H2O2) derived from beating hearts mediates metabolic coronary microvascular dilation.
Methods and Results— We used a bioassay method in which an isolated microvessel is placed on a beating heart to detect myocardium-derived vasoactive mediators. A rabbit coronary arterial microvessel (detector vessel [DV], n=25) was pressurized and placed on a canine beating heart. After intrinsic tone of DV had developed, we observed DV at rest (heart rate, 120 bpm) and during tachypacing (heart rate, 240 bpm) using an intravital microscope equipped with a floating objective. The tachypacing produced DV dilation by 8.2% (P<0.01 versus baseline), and the dilation was abolished by cell-impermeable catalase (a H2O2 scavenger, 500 U/mL). We performed myocardial biopsy at rest and tachypacing. The biopsy specimens were loaded with 2′,7′-dichlorodihydrofluorescein diacetate (10 μmol/L) to visualize H2O2, and observed with confocal microscopy. Dichlorofluorescein fluorescence was diffusely identified in the myocardium and the tachypacing increased the fluorescence intensity (P<0.01). Exogenous H2O2 caused vasodilation of arterial microvessels in vitro in a concentration-dependent manner that was abolished by catalase.
Conclusions— H2O2 derived from the beating heart mediates tachypacing-induced metabolic coronary vasodilation in vivo.
Coronary blood flow is linearly correlated to the myocardial oxygen consumption.1,2 The tight coupling between the cardiac metabolism and the flow conductance underlies the widely accepted assumption that the myocardium-derived vasoactive factor, so-called metabolic factor, rapidly regulates the vascular tone of coronary microvessels for matching the coronary flow to the cardiac metabolic state. However, the metabolic factor that is transmitted from the myocardium to the coronary microvessels has not been identified yet. Although there are many candidates for the mediators such as adenosine, prostanoids, autacoids, nitric oxide (NO), factors activating potassium channels, and so on, no substance solely explains the metabolic microvascular dilation.3,4
Although excessive reactive oxygen species (ROS) are produced in various pathological conditions such as ischemia/reperfusion and cardiac failure5,6 and are often hazardous for living organisms, increasing evidence has shown that ROS also play important roles as biological signals that mediate physiological phenomena at low concentrations.7,8 In the field of vascular biology, hydrogen peroxide (H2O2) has been known as one of the possible endothelium-derived hyperpolarizing factors.9 The cardiac myocyte is another major source of ROS.5,10 Superoxide (O2−·) and H2O2 are produced as byproducts of electron transfer reactions during normal aerobic metabolism. O2−· produced in mitochondria accounts for 1% to 2% of total consumed oxygen,11 and it is reduced to H2O2 by mitochondrial SOD. The steady-state levels of H2O2 inside the cardiac myocyte may allow a portion to diffuse out of the cell into the interstitial space and subsequently to the blood vessels because of its lipophilic property. It is possible that the diffused H2O2 is a signal of cardiac metabolism for the coronary microvessels. Furthermore, H2O2 potently dilates coronary arterial microvessels in endothelium-dependent and endothelium-independent fashion.12 Accordingly, we hypothesized that H2O2 derived from the beating heart mediates the dilation of the coronary microvessels during tachypacing.
We tested the hypothesis using a bioassay method developed in our laboratory,13–15 in which the isolated coronary vessel is placed on the beating heart, because the system enables us to separately control the beating heart and the coronary microvessels, and it is advantageous for detecting the myocardium-derived vasoactive signals.
Materials and Methods
The present study was approved by institutional ethical committee for animal experiments and conformed to the Guide for the Care and Use of Laboratory Animals, published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996).
The bioassay system consists of a pressurized isolated rabbit coronary microvessel for detection of vasoactive signals and a beating canine heart as previously described.13–15
For the preparation of the detector vessel (DV), male Japanese white rabbits (n=25, body weight 2.0±0.0 kg; Japan SLC, Hamamatsu) were anesthetized and anticoagulated with pentobarbital (50 mg/kg) and heparin (1000 U). After the rabbits were euthanized by bleeding from the carotid artery, the hearts were excised and immersed in chilled Krebs solution. A coronary arterial microvessel of the left ventricle (100 to 300 μm) was carefully isolated and used as DV. One end of the vessel was cannulated to a micropipette and the other end was ligated. DV was incubated in warm (38°C) Krebs solution until use. In 6 DVs, the endothelium was removed as previously described.16
Beagle dogs of either gender (n=25; body weight, 5.8±0.2 kg; NARC, Chiba, Japan) were used and prepared as described in our previous studies.13–15 Briefly, the dogs were anesthetized with α-chloralose (60 mg/kg, intravenous; Wako Pure Chemicals, Osaka, Japan) and artificially ventilated. The heart was exposed and paced at 120 bpm by left atrial pacing after the suppression of the sinoatrial node with formaldehyde injection.
The incubated DV was gently placed on the beating left ventricle of the dog and pressurized to 80 cm H2O to produce intrinsic tone. DV and the heart surface were kept moist throughout the experiment by continuous suffusion of Krebs solution (38°C) at a rate of 1 to 1.5 mL/min.
For visualization of DV placed on the beating left ventricle, an intravital microscope system equipped with a floating objective and a charge-coupled device camera (HV-D28S; Hitachi, Tokyo) was used. This optical system was originally developed in our laboratory17 for the visualization of epimyocardial coronary microvessels in vivo. The spatial resolution of this optical system was 2 μm. Epi-illumination with a mercury lamp was applied to obtain the DV images. Obtained images were captured on a personal computer monitor, and the inner diameters were measured with Scion Image (version β4.0.2; Scion, Frederick, Md). The detailed description of the bioassay method is shown in online data supplement (please see http://atvb.ahajournals.org).
Experiments were performed ≈1 hour after DV was pressurized to 80 cm H2O on the beating heart, when all hemodynamic variables and blood gas parameters had become stable and the intrinsic tone of DV had developed.
Protocol 1 (vehicle group, n=11)
After the stable intrinsic tone of DV (with endothelium, n=5; without endothelium, n=6) was established at the resting condition (heart rate=120 bpm), the images of DV were collected. Thereafter, the heart rate was increased to 240 bpm and the images of DV were collected 10 minutes after the tachypacing. The efficacy of endothelial denudation was verified by the lack of vasodilation to 1 μmol/L acetylcholine (Sigma, St. Louis, Mo). At the end of the protocol, sodium nitroprusside (100 μmol/L; Wako) was suffused for 5 minutes to induce the maximal dilation and DV images were again collected.
Protocol 2 (n=14)
To investigate the involvement of H2O2, NO, KATP channels, and adenosine, we suffused cell-impermeable catalase (500 U/mL; Sigma; catalase group, n=6), Nω-nitro-l-arginine ([LNNA] NO synthase inhibitor, 100 μmol/L; Sigma) with glibenclamide ([GC] a KATP channel inhibitor, 5 μmol/L; Wako) (LNNA+GC group, n=4), and 8-phenyltheophylline ([8-PT] a selective adenosine receptor antagonist, 10 μmol/L; Sigma; 8-PT group, n=4) onto DV throughout the experiments. Ten minutes (catalase) or 30 minutes (LNNA+GC and 8-PT) after the pretreatment, the images of DV were collected before and after the tachypacing as in protocol 1. In protocol 2, all DVs were with endothelium.
In 5 anesthetized open-chest beagle dogs, epimyocardial biopsy of the left ventricle was performed at the heart rates of 120, 160, 200, and 240 bpm. The obtained specimens were loaded with 10 μmol/L 2′,7′-dichlorodihydrofluorescein diacetate (Sigma) to visualize H2O2 as 2′,7′-dichlorofluorescein (DCF) fluorescence.18 The fluorescent images were acquired using a laser scanning confocal microscope (C1; Nikon) and the fluorescence intensity was quantitated (AxioVision v220.127.116.11; Carl Zeiss, Jena, Germany). Detailed description of the method is in online data supplement.
We performed immunohistochemistry for sarcomeric α-actinin to stain the myocardium to examine the existence of DCF signal in the myocardium. The frozen myocardial sections (thickness, 5 μm) were incubated with anti-sarcomeric α-actinin antibody (1:250 mouse monoclonal IgG1; Sigma). The sections were incubated with Alexa 546-conjugated goat anti-mouse IgG1 secondary antibody (1:250; Molecular Probes, Eugene, Ore). Anti-sarcomeric α-actinin is specific for α-skeletal and α-cardiac muscle actinin, and does not bind to nonsarcomeric muscle elements such as the connective tissue and smooth muscle.19 The fluorescence image was acquired using a laser scanning confocal microscope (LSM5 PASCAL; Carl Zeiss). The detailed description of the method is shown in online data supplement.
In Vitro Studies
Coronary arterial microvessels (n=16, 198±8 μm, at maximal diameter) were isolated from rabbit hearts (n=9; body weight, 2.1±0.1 kg) and cannulated with dual glass micropipettes in a vessel chamber (CH/2/M; Living Systems Instrumentation, Burlington, Vt) containing Krebs solution (stop flow, 60 cm H2O of distending pressure, 38°C).
After the development of the spontaneous tone, the responses of coronary arterial microvessels to H2O2 (Wako) and sodium nitroprusside were examined, and the effect of extraluminal application of cell-impermeable catalase (500 U/mL) and the endothelial denudation were evaluated. Detailed description of the method is in the online data supplement.
Cell-impermeable catalase (with ≤0.01 mg/mL of thymol as impurity), PEG-catalase, acetylcholine, LNNA, sodium nitroprusside, and H2O2 were freshly dissolved in physiological salt solution to the desired concentrations. Glibenclamide was dissolved with dimethyl sulfoxide at first, and then the target concentration was obtained by dilution with Krebs solution. The final concentration of dimethyl sulfoxide was 0.005 vol%. 8-PT was dissolved in 80% pure methanol and 20% 0.2 mol/L NaOH. This solution was diluted to 10 μmol/L. A 10 mmol/L stock solution of 2′,7′-dichlorodihydrofluorescein diacetate was prepared in ethanol on every experimental day and freshly diluted to 10 μmol/L with phosphate-buffered saline for the experiments.
All variables are expressed as mean±SEM. In the bioassay studies, the vascular diameters were normalized to the maximal diameters defined as the vascular diameter in the presence of 100 μmol/L nitroprusside. When the spontaneous tone of DV did not develop (baseline diameter >90% of the passive diameter), the vessel was discarded. The diameter changes of DV and the differences in hemodynamic variables between baseline and tachypacing were compared by paired t test. Differences in the fluorescence intensity among each heart rate were statistically analyzed by 1-way ANOVA for repeated measures and Bonferroni corrections were applied. Differences in the maximal DV diameters, basal tone, and the percent change in the DV diameter from baseline among each group were compared by 1-way ANOVA. In the in vitro study, the vascular responses were expressed with the percentage of the maximal dilation caused by 100 μmol/L of nitroprusside, and the diameter changes were statistically analyzed with 2-way ANOVA for repeated measures. Post hoc analysis with Bonferroni corrections was applied to detect the concentrations of the significant difference. P<0.05 was considered statistically significant.
Hemodynamic and Blood Gas Data of Dogs
Although baseline mean aortic pressure was lower in vehicle group compared with other groups, it did not significantly change by tachypacing, and rate-pressure products (heart rate × systolic blood pressure) were doubled by tachypacing in each group (Table 1). The blood gases and pH were kept within physiological ranges during experiments.
Detector Vessel Responses to Cardiac Tachypacing
Vessel sizes and the basal tone of the detector vessels in the experimental groups are shown in Table 2. There were no differences in the vessel sizes of DV among 5 groups. There were no statistical difference in the developed basal tone between vehicle group (with endothelium) and any other groups. Endothelium denudation did not affect the basal tone.
Tachypacing significantly increased DV diameter in the vehicle group (Figure 1A). When the endothelium was removed, cardiac tachypacing again produced DV dilation (Figure 1B). In contrast, when cell-impermeable catalase was suffused onto DV, tachypacing-induced dilation was abolished (Figure 1C). In the presence of LNNA+GC or 8-PT, tachypacing resulted in DV dilation (Figure 1D, 1E). In the vehicle group, the percent change of DV by tachypacing was 8.2±0.8% with endothelium and 6.6±2.1% without endothelium, and there was no statistical difference between them (Figure 1F). When catalase was applied, the diameter change by tachypacing was −2.0±3.1% (P<0.05 versus vehicle groups). Neither LNNA+GC nor 8-PT inhibited the tachypacing-induced dilation (12.5±4.0% and 12.5±3.1%, respectively).
H2O2 Detection in the Myocardium
DCF fluorescence was increased by tachypacing (Figure 2A, 2B) and the fluorescence intensity increased in the heart rate-dependent manner (Figure 2C). The preincubation of the myocardium with PEG-catalase abolished the tachypacing-induced increase in the DCF fluorescence (Figure 2D), indicating the specificity of DCF signal for the H2O2 detection. In contrast, the cell-impermeable catalase, which was used in the bioassay experiments, did not decrease the myocardial DCF signal (Figure 2E).
Sarcomeric α-actinin and H2O2 staining were performed for the same section (Figure 3A, 3B), and the superimposed image (Figure 3C) revealed that H2O2 was colocalized with sarcomeric α-actinin, indicating that H2O2 was located in the myocardium.
In Vitro Studies
The spontaneous vascular tone was developed in the isolated coronary arterial microvessels without any pharmacological preconstriction (66±2% of their maximal diameter). The microvessels dilated in response to the extraluminal application of H2O2 in a concentration-dependent manner and the dilation was abolished in the presence of cell-impermeable catalase (Figure 4A). Sodium nitroprusside caused dose-dependent dilation, and cell-impermeable catalase was without effect on the dilator responses (Figure 4B). We confirmed that the time-controlled microvascular responses to H2O2 in the absence of catalase were identical (data not shown). Endothelial denudation did not affect the vasodilation in response to H2O2 (Figure 4C).
The metabolic microvascular regulation is a complicated phenomenon because the flow increase by the metabolic factor increases shear stress in upstream vessels resulting in the enhanced release of the endothelium-derived relaxing factors, and the consequent endothelium-derived relaxing factor-induced dilation increases the distending pressure of the downstream microvessels, which could also affect microvascular tone.20 Furthermore, the enhanced metabolism of the heart is often accompanied by the neurohumoral activation such as the sympathetic nervous system. All of those changes modulate the microvascular diameter changes and make it difficult to characterize the metabolic factor from the myocardium during the increased metabolism in vivo. However, the metabolic coronary dilation can be observed only in the in vivo setting. Our bioassay method resolves this dilemma because we can separately control the beating heart and DV. That is, the distending pressure and flow state (stop flow) of DV were kept constant throughout the experiment and DV was separated from the bloodstream and nervous controls. Thus, the present bioassay system is advantageous for the detection of the beating heart-derived metabolic factor.
H2O2 as a Vasoactive Mediator for Metabolic Dilation
H2O2 might have derived from DV in response to another mediator released from the beating heart. If it was the case, the vascular endothelium or vascular smooth muscle must have been responsible for the H2O2 release. However, the former is unlikely because endothelial denudation did not have any effect on the detector vessel dilation. The latter is also unlikely, because we used cell-impermeable catalase in the bioassay. If H2O2 produced in the vascular smooth muscle itself caused dilation, cell-impermeable catalase should not have abolished the dilation. Accordingly, the beating heart, not DV, is responsible for the H2O2 release.
DCF fluorescence study showed that H2O2 in the beating heart is increased as the heart rate increases. Double staining of DCF fluorescence and sarcomeric α-actinin strongly suggests that the enhanced H2O2 release takes place in the myocardium, because the anti-sarcomeric α-actinin does not bind to the vascular smooth muscle.19 Although we did not quantitate the exact tissue concentration of H2O2, Saito et al10 have recently measured the H2O2 concentration in canine hearts using H2O2 electrode, reporting that it was ≈50 μmol/L at baseline (heart rate=140 bpm). In vitro studies in the present studies indicate the vasodilator effect of H2O2 on the coronary microvessels and the concentration-diameter curve for H2O2 demonstrates that 10 to 100 μmol/L of H2O2 produce coronary microvascular dilation. The magnitude of the microvascular dilation to the tachypacing was consistent with previous in vivo observations for the size class of microvessels we used (≈200 μm).21 All of these data support the concept that H2O2 released from beating myocardium is a physiological mediator of metabolic coronary microvascular dilation. However, the possibility that H2O2 produced in the vascular endothelium of the beating heart takes part in the DV dilation during tachypacing cannot be ruled out because earlier studies showed that the increased shear stress stimulates H2O2 production from the vascular endothelium.22
The dilation of DV without endothelium was comparable with that of DV with endothelium, demonstrating that the metabolic microvascular dilation is endothelium-independent. The contribution of the endothelium to the H2O2-induced dilation is controversial. Thengchaisri et al12 have recently demonstrated that H2O2 induces endothelium-dependent and endothelium-independent coronary arteriolar dilation in pigs, and suggested that the endothelial cyclooxygenase-1 partly mediates the H2O2 dilation. However, Rogers et al23 reported that denuding the endothelium of canine coronary arteries or arterioles did not affect the H2O2 dilation and that 4-aminopyridine sensitive KV channels had a role in the coronary arteriolar dilation to H2O2. The mechanism of H2O2-induced dilation may be species-dependent or vessel size-dependent.
We previously showed that pertussis toxin-sensitive G protein (GPTX) in the microvascular wall plays a role in the coronary microvascular dilation in response to the increased cardiac metabolism in vivo.21 It is possible that H2O2 activates a signaling pathway which involves GPTX, although how H2O2 links to GPTX is not clear at this point. There is a report that H2O2 acutely upregulates sphingosine-1-phosphate receptors, which activate GPTX.24 Further studies are needed to determine the signaling mechanisms of H2O2 derived from the myocardium.
Catalase we used in the bioassay and in vitro studies was cell-impermeable catalase. We validated that it does not decrease the intracellular H2O2, whereas PEG catalase does by performing DCF staining (Figure 2D, 2E). Therefore, it is plausible that the cell-impermeable catalase quenched H2O2 in the space between the beating heart and DV. These results are in concert with the concept that H2O2 is a transferable vasoactive mediator released from the beating heart to the coronary microvessels.
In the present experiments, none of the inhibitors of NO, KATP channels, or adenosine abolished the tachypacing-induced DV dilation. Our results are consistent with earlier studies. It has been shown that NO inhibitors do not impair the coronary flow increase during exercises25,26 or tachypacing.27 It is conceivable that NO does not play a major role in the metabolic flow increase, whereas it modulates redistribution of the coronary resistance from large microvessels to small ones during the increased cardiac metabolism.27 Our previous study showed that KATP channels are likely to play an important role in determining coronary microvascular basal tone but not in dilating coronary microvessels during tachypacing.28 Other researchers have also provided evidence that KATP channel blocker does not result in the impairment of flow response to increased cardiac metabolism.3 Furthermore, adenosine is unlikely to mediate the metabolic coronary vasodilation under the normal perfusion, whereas it plays a role during impaired coronary flow state.3
We suffused warm Krebs solution throughout the experiments at the rate of 1 to 1.5 mL/min to keep the bioassay system wet and warm. The space between the floating objective and the heart surface was always filled with the solution, which could dilute the vasoactive signals. Therefore, it is possible that the transferable vasomotor signals evaluated in our bioassay method may have been underestimated to some extent.
Because of the technical limitation, the size of DV was ≈150 μm in inner diameter with spontaneous tone. It is greater in size for the coronary microvessels that is most sensitive to metabolic factors.4 The vascular wall of the smaller vessels was too thin to be detected on the beating heart. However, the microvessels ≈150 μm bear significant resistance;29 therefore, we believe that the present results well demonstrate the physiological significance of myocardium-derived metabolic factor on coronary resistance regulation.
In the present study, we have demonstrated that H2O2, one of the ROS, is involved in the physiological regulation of coronary microvascular tone during enhanced cardiac metabolism. Redox homeostasis is delicately balanced by a network of ROS production systems and many scavenging systems. It has been recognized that ROS play pathological roles in many diseases. In the failing myocardium, oxidative stress is increased and appears to play a role in the pathogenesis.30,31 The imbalanced redox homeostasis in the myocardium may affect the production and diffusion of H2O2 in the myocardium, leading to the impairment of the coupling between the myocardial metabolic state and coronary flow regulation.
We have demonstrated that the metabolically stimulated heart releases H2O2 that mediates the coronary microvascular dilation. H2O2 may be one of metabolic factors that links the myocardial metabolic state to coronary flow regulation. The concept that ROS plays an important role as a physiological signal for coronary flow regulation may provide a new insight for the understanding of the coronary regulation and the pathophysiology of various diseased conditions.
The authors thank N. Yamaki for her excellent technical assistance.
Sources of Funding
This study was supported by grants-in-aid for scientific research from the Ministry of Education, Culture, Sports, Science and Technology, Tokyo, Japan (No. 16209027, 16659192), the Japanese Ministry of Health, Labor, and Welfare, Tokyo, Japan, the Japan Foundation of Cardiovascular Research, Tokyo, Japan, and Saito Gratitude Foundation (No. GK151112).
During the preparation of this manuscript, the study elucidating the role of H2O2 in metabolic coronary dilation was published.10
Original received August 23, 2006; final version accepted February 12, 2007.
Feigl EO. Coronary physiology. Physiol Rev. 1983; 63: 1–205.
Ide T, Tsutsui H, Kinugawa S, Utsumi H, Kang D, Hattori N, Uchida K, Arimura K, Egashira K, Takeshita A. Mitochondrial electron transport complex I is a potential source of oxygen free radicals in the failing myocardium. Circ Res. 1999; 85: 357–363.
Droge W. Free radicals in the physiological control of cell function. Physiol Rev. 2002; 82: 47–95.
Thannickal VJ, Fanburg BL. Reactive oxygen species in cell signaling. Am J Physiol Lung Cell Mol Physiol. 2000; 279: L1005–L1028.
Saitoh S, Zhang C, Tune JD, Potter B, Kiyooka T, Rogers PA, Knudson JD, Dick GM, Swafford A, Chilian WM. Hydrogen peroxide: a feed-forward dilator that couples myocardial metabolism to coronary blood flow. Arterioscler Thromb Vasc Biol. 2006; 26: 2614–2621.
Thengchaisri N, Kuo L. Hydrogen peroxide induces endothelium-dependent and -independent coronary arteriolar dilation: role of cyclooxygenase and potassium channels. Am J Physiol Heart Circ Physiol. 2003; 285: H2255–H2263.
Sato K, Komaru T, Shioiri H, Takeda S, Takahashi K, Kanatsuka H, Nakayama M, Shirato K. Hypercholesterolemia impairs transduction of Vasodilator signals derived from ischemic myocardium. Arterioscler Thromb Vasc Biol. 2004; 24: 1–6.
Takahashi K, Komaru T, Takeda S, Sato K, Kanatsuka H, Shirato K. Nitric oxide inhibition unmasks ischemic myocardium-derived vasoconstrictor signals activating endothelin type A receptor of coronary microvessels. Am J Physiol Heart Circ Physiol. 2005; 289: H85–H91.
Ashikawa K, Kanatsuka H, Suzuki T, Takishima T. Phasic blood flow velocity pattern in epimyocardial microvessels in the beating canine left ventricle. Circ Res. 1986; 59: 704–711.
Ohba M, Shibanuma M, Kuroki T, Nose K. Production of hydrogen peroxide by transforming growth factor-beta1 and its involvement in induction of egr-1 in mouse osteoblastic cells. J Cell Biol. 1994; 126: 1079–1088.
Tanikawa T, Kanatsuka H, Koshida R, Tanaka M, Sugimura A, Kumagai T, Miura M, Komaru T, Shirato K. Role of pertussis toxin-sensitive G protein in metabolic vasodilation of coronary microcirculation. Am J Physiol Heart Circ Physiol. 2000; 279: H1819–H1829.
Rogers PA, Dick GM, Knudson JD, Focardi M, Bratz IN, Swafford AN Jr, Saitoh S, Tune JD, Chilian WM. H2O2-induced redox sensitive coronary vasodilation is mediated by 4-aminopyridine-sensitive K+ channels. Am J Physiol Heart Circ Physiol. 2006; 291: H2473–H2482.
Igarashi J, Miyoshi M, Hashimoto T, Kubota Y, Kosaka H. Hydrogen peroxide induces S1P1 receptors and sensitizes vascular endothelial cells to sphingosine 1-phosphate, a platelet-derived lipid mediator. Am J Physiol Cell Physiol. 2006; Aug 30:[Epub ahead of print].
Altman JD, Kinn J, Duncker DJ, Bache RJ. Effect of inhibition of nitric oxide formation on coronary blood flow during exercise in the dog. Cardiovasc Res. 1994; 28: 119–124.
Ishibashi Y, Duncker DJ, Zhang J, Bache RJ. ATP-sensitive K+ channels, adenosine, and nitric oxide-mediated mechanisms account for coronary vasodilation during exercise. Circ Res. 1998; 82: 346–359.
Jones CJ, Kuo L, Davis MJ, DeFily DV, Chilian WM. Role of nitric oxide in the coronary microvascular responses to adenosine and increased metabolic demand. Circulation. 1995; 91: 1807–1813.
Takeda S, Komaru T, Takahashi K, Sato K, Kanatsuka H, Kokusho Y, Shirato K, Shimokawa H. Beating myocardium counteracts myogenic tone of coronary microvessels: involvement of ATP-sensitive potassium channels. Am J Physiol Heart Circ Physiol. 2006; 291: H3050–3057.
Nakamura R, Egashira K, Machida Y, Hayashidani S, Takeya M, Utsumi H, Tsutsui H, Takeshita A. Probucol attenuates left ventricular dysfunction and remodeling in tachycardia-induced heart failure: roles of oxidative stress and inflammation. Circulation. 2002; 106: 362–367.