A Feed-Forward Dilator That Couples Myocardial Metabolism to Coronary Blood Flow
Objective— We tested the hypothesis that hydrogen peroxide (H2O2), the dismutated product of superoxide (O2·−), couples myocardial oxygen consumption to coronary blood flow. Accordingly, we measured O2·− and H2O2 production by isolated cardiac myocytes, determined the role of mitochondrial electron transport in the production of these species, and determined the vasoactive properties of the produced H2O2.
Methods and Results— The production of O2·− is coupled to oxidative metabolism because inhibition of complex I (rotenone) or III (antimycin) enhanced the production of O2·− during pacing by about 50% and 400%, respectively; whereas uncoupling oxidative phosphorylation by decreasing the protonmotive force with carbonylcyanide-p-trifluoromethoxyphenyl-hydrazone (FCCP) decreased pacing-induced O2·− production. The inhibitor of cytosolic NAD(P)H oxidase assembly, apocynin, did not affect O2·− production by pacing. Aliquots of buffer from paced myocytes produced vasodilation of isolated arterioles (peak response 67±8% percent of maximal dilation) that was significantly reduced by catalase (5±0.5%, P<0.05) or the antagonist of Kv channels, 4-aminopyridine (18±4%, P<0.05). In intact animals, tissue concentrations of H2O2 are proportionate to myocardial oxygen consumption and directly correlated to coronary blood flow. Intracoronary infusion of catalase reduced tissue levels of H2O2 by 30%, and reduced coronary flow by 26%. Intracoronary administration of 4-aminopyridine also shifted the relationship between myocardial oxygen consumption and coronary blood flow or coronary sinus po2.
Conclusions— Taken together, our results demonstrate that O2·− is produced in proportion to cardiac metabolism, which leads to the production of the vasoactive reactive oxygen species, H2O2. Our results further suggest that the production of H2O2 in proportion to metabolism couples coronary blood flow to myocardial oxygen consumption.
The coupling of blood flow to metabolism is the most important vasomotor adjustment for the regulation of oxygen delivery to metabolically active organ systems. This matching, termed metabolic dilation, or metabolic or active hyperemia, is critical to ensure adequate oxygen delivery for aerobic metabolism and adequate organ function.1 Although the factor or factors responsible for the coupling of flow to metabolism have been actively pursued for decades, no metabolite has been casually linked to the process of metabolic hyperemia or has withstood critical evaluation.1–3 Most investigations have pursued the idea that the metabolic regulation of flow is a negative feedback pathway, in which an imbalance between oxygen supply (delivered via flow) and oxygen demands, ie, demands exceed supply, results in the production of a metabolic dilator. The adenosine hypothesis was such a scheme, in which oxygen demands, in excess of supply would increase the production of adenosine through hydrolysis of ATP and subsequent dephosphorylation of ADP and AMP.1,4 However, the adenosine hypothesis has been largely refuted for normal metabolic dilation.2,3 Furthermore a problem with the negative feedback theories for metabolic regulation pertains to the problem that after blood flow is increased to match oxygen supply with demand, there is no error signal to sustain the dilation. We make this statement because in a negative feedback scheme the error signal for metabolic dilation is the metabolite produced when demand for oxygen exceeded supply; thus when oxygen supply has been rectified via dilation, the error signal is absent because the production of the metabolite would return to baseline. Accordingly, our hypothesis centered on a different scheme in which the production of a metabolic dilator would be a feed-forward system—one without an error signal—that would be directly linked to oxygen consumption. We hypothesized that H2O2 would link oxygen consumption and blood flow. Our hypothesis was based on observations showing that H2O2 is vasoactive,5,6 has a short half-life, because it is metabolized rapidly by catalase and it rapidly reacts with free thiol groups,7 and is freely permeable—all requirements of a metabolic dilator.8,9 Moreover the presence of catalase in the blood stream10,11 also would confine its vasoactive effects to the producing organ system, which would prevent any spill-over of the dilation to nonmetabolically active organ systems. And finally, exercise in skeletal muscle is associated with an increase in the production of reactive oxygen species, although the role of these species is not yet resolved.12
Materials and Methods
Measurement of Arteriolar Vasodilation and Isolation of Cardiac Myocytes
Wistar rats were used for studies of isolated arterioles and isolated cardiac myocytes. Rats were anesthetized with sodium pentobarbital (50 mg/kg, ip) and a mid-sternotomy was performed. The heart was excised and placed in 4°C buffered physiological salt solution (PSS). The bathing solution used for microvessel dissection had the following compositions (in mmol/L): NaCl 145.0, KCl 4.7, CaCl2 2.0, MgSO4 1.17, NaH2PO4 1.2, glucose 5.0, pyruvate 2.0, EDTA 0.02, 3-N-morpholino propanesulfonic acid buffer (MOPS) 3.0 and contained 1% bovine serum albumin (1 gm/100 mL). The solution was buffered to pH of 7.4 at 4°C. The PSS used to perfuse the vessels during the experiments was the same composition as mentioned above, but was buffered to pH of 7.4 at 37°C.
Isolated Coronary Arterioles
Single arterioles were dissected from the left ventricle.13,14 A portion of the LV was removed and several arterioles of the appropriate size were located under a dissecting microscope. Each arteriole with surrounding ventricular muscle was excised, transferred to a temperature-controlled dissection dish (4°C) containing PSS, and dissected free of the muscle tissue. Side branches were tied off using 11-O suture. The vessel was transferred to a lucite chamber and cannulated at both ends using micropipettes that had matched resistances. The arteriole was tied to each pipette using 11-O suture. The preparation was then transferred to the stage of an inverted microscope. Leaks were assessed by measuring pressure at zero flow, which should equal pressure in the inflow reservoir pressure, when there were no leaks. Any preparations with leaks were excluded. Agents (supernatant, catalase) were administered in the bath.
Isolated Cardiac Myocytes
Cardiac myocytes were enzymatically isolated from rat hearts. After excision of the heart, the aorta was cannulated, and the preparation was suspended in a perfusion apparatus. The LV was initially perfused (retrograde from the aorta) at 37°C with oxygenated, calcium free-HEPES buffer (pH 7.45 [titrated with 5 mol/L NaOH]) in mmol/L: 10 HEPES, 30 taurine, 113 NaCl, 4.7 KCl, 0.6 KH2PO4, 0.6 Na2HPO4, 1.2 MgSO4, 0.032 Phenol Red, 12 NaHCO3, and 10 KHCO3 to rinse out residual blood and eliminate contraction. After cessation of contractile activity, the perfusion was switched to a buffer of the above constituents along with 0.25 mg/mL liberase blendzyme I (Roche), 0.14 mg/mL trypsin, and 12.5 μmol/L CaCl2. After perfusion of the heart for 10 to 12 minutes and identification of isolated myocytes in perfusate from the heart, hearts were detached from the perfusion apparatus, and placed in a “stop” solution containing the perfusion buffer with 10% BSA and 12.5 mmol/L Ca2+. The heart was minced into small pieces that were further titrated in stop buffer. After microscopic confirmation of the presence of myocytes the cells were filtered and placed in a 50 mL conical tube. CaCl2 was added in a series of 4 steps to arrive at a final concentration of 1.9 mmol/L. Cells were pelleted by centrifugation (1500 rpm) for 5 minutes, and the supernatant was discarded. Cells were resuspended in the stop buffer with calcium and then small aliquots were used for cell counts (hemocytometer) to enable dilution or concentration (via centrifugation) to a final concentration of 100 000 cells per ml. Viability of the myocytes was determined by trypan blue exclusion and a rod-like configuration. On average well over 70% of the cells exhibit rod-like configuration.
In Vitro Generation of O2·− and H2O2
Superoxide was measured in supernatant of stimulated myocytes by using electron paramagnetic spectroscopy (EPR) with the cyclic hydroxylamine, CP-H (1-hydroxy-3-carboxy-pyrrolidine),15 which penetrates into the cells and provides an index of intracellular superoxide production (Alexis Biochemicals). Suspensions of enzymatically isolated cardiac myocytes were treated with CP-H (2 mmol/L for 20 minutes) in 2 mL of the physiological salt solution (pH 7.4) under the following conditions: basal conditions (unstimulated), electrical stimulation at 400 minutes−1, and electrical stimulation in the presence of rotenone (2 μmol/L), antimycin (2 μmol/L), myxothiazol (100 nmol/L), FCCP (1 μmol/L, p-trifluoromethoxy carbonyl cyanide phenyl hydrazone), apocynin (0.3 mmol/L) or the superoxide dismutase (SOD) mimetic, Mn(III)tetrakis(4-benzoic acid)porphrin chloride (Mn-TBAP, 25 μmol/L). At 20 minutes (unstimulated or stimulated) cell-free supernatant (0.5 mL) was removed and immediately snap frozen in liquid nitrogen (LN2). Samples were stored in LN2 until EPR spectroscopy is performed. For EPR measurements, samples were rapidly thawed in a water bath at 37°C and aspirated into glass capillaries (ID 1 mm) and read at room temperature using a Bruker EMX spectrometer. The EPR spectrum settings were as follows: modulation amplitude 1.0 gauss, scan time 83 seconds, time constant 163 ms and microwave power 40 mW, field sweep 60 gauss, microwave frequency 9.78 GHz, receiver gain 5×103, center field 3485 gauss. Superoxide quantification from the EPR spectra was determined by double integration of the peaks, with reference to a standard curve generated from horseradish peroxidase generation of the anion from standard solutions of hydrogen peroxide, using p-acetamidophenol as the cosubstrate.
Hydrogen peroxide was measured using an electrochemical detection system (Apollo 4000 from World Precision Instruments) from 0.5 mL aliquots of the supernatant derived from the suspensions of cardiac myocytes. Samples of supernatant were obtained from unstimulated myocytes, stimulation at 200 or 400 minutes−1, stimulation + catalase. During each treatment, stimulation was continued for 20 minutes in the presence of the drug, and samples of supernatant were obtained, frozen, and stored until measurements of H2O2 could be made using the Apollo 4000. Each electrode was calibrated using serial dilutions of H2O2, and the current recorded from the supernatant was then calculated as [H2O2].
The conditioned supernatant buffer from these various treatments was used not only for assessment of H2O2 and O2·− generation but also to establish vasoactive effects in isolated arterioles. In addition to these protocols, we also treated the supernatant with catalase (500 to 1000 U/mL) to determine whether the catabolism of H2O2 would affect vasodilation. We also determined whether 4-aminopyridine (4-AP, 0.3 mmol/L) would block the vasodilatory effects of the supernatant. We recently found that H2O2-induced coronary vasodilation of isolated arterioles, relaxation of coronary arterial rings, and of in vivo coronary blood flow was antagonized by 4-AP, indicating that H2O2 produces dilation via activation of Kv channels.16 To eliminate the possibility that the supernatant may be producing dilation via stimulation of endothelial NO production, we compared vasodilation to supernatant in control preparations to those treated with L-NAME (3×10−4 mol/L).
Measurement of MVO2 and In Vivo H2O2
Measurements of coronary and systemic hemodynamics were completed as described previously.17 Dogs (n=5) were sedated with morphine (3 mg/kg sc) and anesthetized with α-chloralose (100 mg/kg iv). The animals were intubated and ventilated with room air supplemented with oxygen. A catheter was placed into the thoracic aorta via the right femoral artery to measure aortic blood pressure. Another catheter was inserted into the right femoral vein for injection of supplemental anesthetic, heparin and sodium bicarbonate. The left femoral artery was catheterized to supply blood to an extracorporeal perfusion system used to perfuse the left anterior descending coronary artery (LAD). The heart was exposed via a left lateral thoractomy and placed in a pericardial cradle. A catheter was then situated in the interventricular vein via the right atrial appendage (and advanced retrograde through the coronary sinus and great cardiac vein) for coronary venous blood sampling. The LAD was isolated distal to its first major diagonal branch. After heparin administration (500 U/kg), the LAD was cannulated with a stainless steel cannula connected to the extracorporeal perfusion system with an in-line Transonics flow probe to measure coronary blood flow. Coronary perfusion pressure was maintained at 100 mm Hg throughout the experimental protocol by a servo-controlled roller pump. Arterial and coronary venous blood samples were collected when hemodynamic variables were stable and immediately sealed and placed on ice. The samples were analyzed in duplicate for pH, pco2, po2, hematocrit, and oxygen content with an Instrumentation Laboratories automatic blood gas analyzer (GEM Premier 3000) and CO-Oximeter (682) system. Myocardial oxygen consumption (MVO2) was calculated by multiplying coronary blood flow by the arterial-coronary venous difference in oxygen content.
An H2O2 electrode (100 μm diameter) connected to an Apollo 4000 (World Precision Instruments) was inserted into the myocardium in the LAD perfusion territory via a stab wound made by an 18 gauge needle. Each electrode was calibrated in vitro with known concentrations of H2O2 and the current recorded during the in vivo experiments was then calculated as [H2O2] and correlated to the levels of coronary flow and MVO2. After situation of the electrode and the allowance of several minutes to insure stability, the myocardium was paced at varying rates, or norepinephrine was infused i.v. to produce changes in heart rate and aortic pressure as maneuvers designed to increase myocardial oxygen consumption.
To study the relationships between pacing-induced changes between MVO2 and H2O2, and between H2O2 and coronary blood flow, we initially made measurements of the variables at baseline, then increased heart rate via electrical pacing to attain rates of 140, 160, and 200 minutes−1. Once a steady state was attained at each rate, measurements were acquired. To study the relationships between norepinephrine-induced changes between MVO2 and H2O2, and between H2O2 and coronary blood flow, we initially made baseline measurements, then infused norepinephrine i.v. in doses of 16, 32, and 64 μg/kg/min to achieve changes in systemic hemodynamics. Once a steady state was attained, measurements were acquired.
In 4 preparations, catalase (15 000 U/min, intracoronary) was infused under basal conditions over 3.5 minutes to reduce cardiac H2O2 levels independently of metabolism and during the last 30 sec tissue H2O2 and coronary flow were measured. This protocol was designed to better show cause and effect, ie, examine whether coronary flow followed the anticipated reduction in H2O2 induced by catalase.
In 5 preparations, 4-AP was administered intracoronary (calculated coronary plasma concentration of 0.3 mmol/L) to block Kv (4-AP-sensitive) ion channels. Measurements of coronary flow, arterial and coronary sinus oxygen tensions, and oxygen content were made to construct the relationship between coronary blood flow and myocardial oxygen consumption. This protocol was designed to block the ionic mechanism by which H2O2 produces coronary vasodilation (via activation of 4-AP sensitive ion channels) and establish whether this antagonism would corrupt the relationship between oxygen consumption and coronary blood flow, and that between oxygen consumption and coronary sinus po2. To ascertain that 4-AP was not producing any nonspecific effects, we performed 15-second coronary occlusions and determined whether the ability of the coronary circulation to dilation (measurement of peak-to-resting flow ratio) was affected by 4-AP.
We used a 1-way ANOVA followed by a Tukey post hoc test to determine differences among the different interventions for O2·− and H2O2 production. A 2-way ANOVA for repeated measures followed by Tukey post hoc test was used to determine differences in vasodilation resulting from the various interventions. An unpaired t test was used to compare relaxation to supernatant in the absence or presence of L-NAME. One-way ANOVA with repeated measures was used to determine changes in coronary flow, in vivo levels of H2O2, and myocardial oxygen consumption. A least squares regression analysis was used to establish the correlation coefficients between coronary blood flow and myocardial oxygen consumption (MVO2), and myocardial oxygen consumption and coronary sinus po2. A paired t test was used to determine the effects of catalase on coronary blood flow and H2O2 levels in the myocardium. Peak-to-resting flow ratios (basal to reactive hyperemic flow) were compared by an unpaired t test. A probability value of less than 0.05 was used to establish statistical significance.
Production of O2·− and H2O2 During Elevated Metabolism
Figure 1A through 1C shows the production of O2·− in isolated myocytes using electron paramagnetic spectroscopy (EPR) with the spin trap, CP-H (1-hydroxy-3-carboxy-pyrrolidine).15 Figure 1A shows EPR signals for O2·− in aliquots of supernatant from the myocyte suspensions after stimulation at 400 minutes−1 in the absence (left) or presence (right) of the superoxide dismutase (SOD) mimetic, MnTBAP.MnTBAP significantly attenuated the signal, which demonstrated the component related to O2·−. Figure 1B shows the production of O2·− under control conditions (nonstimulated myocytes) versus stimulation at 400 minutes−1. The production of O2·− was increased during augmented metabolism by pacing. Figure 1C summarizes superoxide generation from the integrated EPR signals under basal conditions (n=6) and during stimulation at 200 and 400 minutes−1 (n=−5 to 6).
EPR measurements of O2·− production from the isolated myocytes during control conditions (nonstimulated), during pacing at 400 minutes−1, or pacing in the presence of rotenone, antimycin, myxothiazol, apocynin, and FCCP are shown in Figure 2A (n=5 to 11). Production of O2·− by nonstimulated myocytes was not significantly affected by any of the inhibitors (not shown). In stimulated myocytes, antimycin or rotenone produced significant increases in superoxide production. In contrast, myxothiazol or FCCP decreased production of O2·− by the stimulated myocytes; whereas apocynin had no effects on production of the free radical. The concentration of H2O2 in the supernatant of the myocytes preparations (n=8 to 15) was increased (from nonstimulated) during stimulation at 200 and 400 minutes−1 (Figure 2B). In nonstimulated myocytes the concentration of H2O2 in the supernatant was 0.2±0.1 μmol/L, and this increased to 4.3±0.4 μmol/L and 9.5±0.7 μmol/L at 200 and 400 minutes−1, respectively. We also found that some of the interventions that affected O2·− production during pacing affected H2O2 production in a similar direction. During pacing at 400 minutes−1, apocynin did not change H2O2 concentrations in the supernatant (8.5±0.5 [apocynin+pacing] versus 9.5±0.7 μmol/L [pacing]), whereas myxothiazol significantly decreased production (P<0.05) during pacing (9.5±0.7 μmol/L [pacing] versus 4.8±0.6 μmol/L [myxothiazol+pacing]). In a few experiments, 500 U/mL of catalase was added to the samples, and the signal for H2O2 was absent, indicating that H2O2 was being measured.
Vasoactivity of H2O2 Produced During Elevated Metabolism
To demonstrate that the levels of H2O2 being produced were capable of producing vasodilation, aliquots of fluid were removed from the myocyte suspension and administered to isolated coronary arterioles. Without stimulation, aliquots of supernatant from the myocytes did not produce significant vasodilation, but during electrical stimulation of 200 & 400 minutes−1, graded vasodilation was observed (Figure 3A, n=14 for each condition except catalase [n=11]). Administration of catalase to the myocyte chamber completely prevented the vasodilation during electrical stimulation (Figure 3A) suggesting that the vasodilatory activity was mediated by H2O2 because this activity was eliminated by degradation of this reactive oxygen species. Figure 3B illustrates the effects of the Kv channel antagonist, 4-aminopyridine (n=5 for each condition), on dilation to the supernatant. 4-AP substantially reduced the vasodilatory effects of supernatant from stimulated myocytes. Administration of L-NAME (3×10−4 mol/L), which blocked dilation to 10−5 mol/L acetylcholine, did not influence relaxation to the largest dose (500 μL) of supernatant from myocytes stimuated at 400 minutes−1 (difference of 3.9% between before and after L-NAME).
In Vivo Links Between H2O2, Myocardial Metabolism, and Coronary Blood Flow
The relationship between myocardial oxygen consumption (stimulated by pacing or norepinephrine) and myocardial levels of H2O2 shows a direct association (Figure 4A). A plot of coronary blood flow versus either H2O2 or MVO2 (Figure 4B) was also direct and significant (P<0.05), suggesting that the production of H2O2 is linked to myocardial oxygen metabolism. Calculation of the molar ratio of oxygen consumed versus H2O2 production (from the slope of the H2O2–MVO2 relationship) yielded a value of 23 nmol/L of H2O2 produced per μmol/L of O2 consumed. In the anesthetized preparations, coronary sinus pO2 values averaged 29±2 mm Hg under basal conditions and 25±2 mm Hg during the different levels of pacing or norepinephrine infusion. Arterial pressure and heart rates (under basal conditions) were 64±7 mm Hg and 112±7 minutes−1, respectively. Pressures did not change with pacing, but both arterial pressure (105±12 mm Hg) and heart rate (147±30 minutes−1) increased significantly during norepinephrine infusion (P<0.05).
Intracoronary catalase infusion did not affect systemic hemodynamics, but decreased both cardiac levels of H2O2 (from 75±19 to 52±16 μmol/L), and coronary flow (from 23±5 to 17±4 mL/min; both P<0.05, paired t test, Figure 5A). The percentage decreases in hydrogen peroxide levels and coronary flow induced by catalase were 30% and 26%, respectively. Infusion of 4-aminopyridine (intracoronary) produced dramatic effects on the relationship between coronary blood flow and myocardial oxygen consumption (MVO2; Figure 5B, left panel), and between coronary sinus po2 and MVO2 (Figure 5B, right panel). In both relationships, 4-AP significantly changed the relationship between the two variables by reducing coronary blood flow or decreasing coronary sinus po2 for a given level of oxygen consumption. 4-AP did not affect the magnitude of the reactive hyperemic response. The peak-to-resting flow ratio was 2.5±0.4 under control conditions and 2.8±0.4 during 4-AP. Also the magnitudes of the hyperemic flows (2.0 [control] versus 1.7 [4-AP] ml/min per g) were not different.
The quest to identify the causal links between metabolism and flow has spanned several decades. Although causal metabolites linking metabolism to flow have been suggested in the past, eg, adenosine,18 there has been no unequivocal proof or acceptance that a particular substance links metabolism to blood flow. Our results support the conclusion that one of the factors responsible for coronary metabolic dilation—the link between oxygen consumption and blood flow—is H2O2. Our conclusion is based on a number of observations. First, cardiac myocytes produced superoxide via forward electron flow in the myocardium, which would serve to link the production of superoxide with metabolic activity. Second, cardiac myocytes in vitro or in vivo appear to produce H2O2 proportionately with cardiac metabolism, and in vivo, there is a correlation between hydrogen peroxide and coronary flow. Third, vasodilatory properties of metabolites produced by metabolically active cardiac myocytes were abolished by catalase or by 4-aminopyridine suggesting the vasoactive metabolite was H2O2. Intracoronary infusion of catalase decreased both cardiac H2O2 levels and coronary flow in vivo, implying that the relationship between the two variables is causal. And finally, intracoronary administration of 4-aminopyridine reduced coronary flow or coronary venous oxygen tension for a given level of myocardial oxygen consumption implying that H2O2 was playing a pivotal role in vivo for coronary metabolic dilation. It is important to add that 4-aminopyridine did not affect the ability of the coronary circulation to dilate to an ischemic stimulus because dilation during reactive hyperemia was not compromised; suggesting that the effects of the Kv channel antagonist were not non-specific. Central to our conclusion and observations are considerations of our model, the measurements, and related reports in the literature.
Our results suggest a causal relationship between myocardial oxygen consumption and H2O2 production, and between H2O2 and either dilation of isolated arterioles or coronary blood flow. Although the in vitro data show the effects of hydrogen peroxide being formed by cardiac myocytes, the in vivo results are somewhat more complicated to interpret because the electrode is measuring a composite of hydrogen peroxide that can be produced from a variety of cells. Moreover, we cannot exclude the possibility that injury attributable to insertion of the electrode has influenced the measurements because they reflect a combination of interstitial levels, and levels being produced by punctured cells. Some of these concerns are lessened by the results showing that infusion of catalase decreases measured H2O2 levels, which suggest that the measurements are of a pool of H2O2 that is in ready equilibrium with the interstitium because it was susceptible to degradation by catalase. Despite the caveats, we can say that the in vivo measured levels of H2O2 in the myocardium were directly correlated to myocardial oxygen consumption, which reflects primarily oxygen metabolism of working cardiac myocytes, and that reductions in H2O2 in vitro or in vivo by catalase reduced dilation of isolated coronary resistance vessels or coronary blood flow, respectively. One caveat we are compelled to mention is that we did not measure oxygen consumption in vitro during electrical stimulation of the myocytes, but previously we have found that pacing induces contraction in the isolated myocytes and oxygen consumption increases about 30-fold and 40-fold when the cells are paced at 200 and 400 minutes−1, respectively.13 Thus we are confident in our assertion that pacing of the myocytes increased oxygen consumption, and the production of H2O2 is then linked to metabolism.
The suggestion that coronary metabolic dilation is mediated by H2O2 is further supported by the experiments using 4-aminopyridine, which blocks the vasodilatory actions of H2O2 in the coronary circulation.16 The extent of vasodilation produced in vitro (paced myocytes) was virtually abolished by 4-aminopyridine, which mimicked the inhibitory actions of catalase on dilation to the supernatant. The relationship between coronary blood flow and myocardial oxygen consumption was also significantly altered so that for any given level of oxygen consumption, coronary flow was less after 4-aminopyridine. We also would like to point out that the level of oxygen consumption we could attain during 4-aminopyridine was smaller than under control conditions. Our interpretation of this finding was that if metabolic dilation is prevented/blunted, ie, the increase in flow with metabolism, then by definition oxygen consumption (the product of flow and the arterial venous oxygen content difference) should be attenuated. The shift in the relationship between oxygen consumption and coronary sinus po2 also supports the conclusion that H2O2 produced metabolic dilation, because blockade of H2O2-induced vasodilation shifts this relationship so that for any level of oxygen consumption, the myocardium is relatively more hypoxic. The relationship between coronary venous po2 and myocardial oxygen consumption has been proposed as a way to study coronary metabolic dilation, because if a vasodilator metabolite linking metabolism and flow is blocked, the heart would become relatively more hypoxic.3,19 Our results are also consistent with this explanation. One of our observations also bears on these arguments. Specifically, during intracoronary catalase infusion, we found simultaneous reductions in both H2O2 and coronary blood flow; further suggesting that levels of H2O2 are directly linked to coronary flow. And finally, it is important to point out that the concentrations of H2O2 we found in vivo are capable of producing vasodilation. In arteries and arterioles, the H2O2 dose-response relationship for dilation has a threshold of 0.01 μmol/L and is maximal at 1000 μmol/L.16,20 Our measured values in vivo would be associated with significant dilation at all ranges, and near maximal dilation at the higher dose.
Another interesting aspect of our results is the lack of an effect of 4-AP on the magnitude of dilation after a reactive hyperemic response. It was important to show that 4-AP did not compromise coronary vasodilation to another stimulus. But the results imply that the basis for ischemic dilation (reactive hyperemia) is different than that resulting from metabolic dilation.
We would like to mention an apparent discrepancy between measured H2O2 concentrations in the isolated myocytes concentrations versus those in vivo. We believe this occurred because of the dilutional effects of having 200 000 myocytes in 2 mL of buffer, of which the myocytes were only a small fraction of the total volume in the chamber.
We would be remiss to not mention a recent article that suggested reactive oxygen species are not produced during myocardial metabolism,21 which would suggest that they are not vasoactive mediators of coronary metabolic dilation. Our results differ and support the concept that these species are important mediators of metabolism. A likely reason why our conclusions differ from the report of Traverse et al21 relates to their method of measurements of free radicals. Traverse et al collected coronary sinus blood then added the spin trap in an attempt to measure O2·−. We would argue that because of the short half-life of O2·− attributable to spontaneous chemical dismutation, enzymatic dismutation, and its reaction with a variety of proteins, lipids, metals, and thiols, the several seconds required to collect the venous sample and add the spin trap would likely result in a negative experiment, because the superoxide would be extinguished. Also the activities of superoxide dismutases have been calculated to be of sufficient magnitude to prevent any efflux of native O2·− from the mitochondria.22 We believe our results of superoxide being produced in a parallel manner to metabolic activity in isolated cardiac myocytes, and the production of H2O2 both in vitro and in vivo proportionately to metabolism support the idea that that the production of these reactive oxygen species is linked to myocardial oxygen metabolism, and that they are vasoactive metabolites.
The production of H2O2 as a function of metabolism can be thought of as a feed-forward control system in that the organ is directly producing the dilator as a function of metabolism, and not when there is an imbalance between oxygen supply and demand. Several observations suggest that the source of H2O2 is from the generation of O2·− by mitochondrial electron transport. The results using the inhibitors of complex I (rotenone) and complex III (antimycin and myothizol) in the electron transport chain suggest that the production O2·− (and thus H2O2) is produced by forward electron flow driven by oxidative metabolism and the sources of O2·− are from complexes I and III. If the electron flow were reverse (from complex II to I), then rotenone would be expected to decrease the production of O2·−; however, this was not the observation. Both rotenone and antimycin are known to increase O2·− generation during forward electron flow,23–25 but myxothiazol reportedly has the opposite effect,24,25 although admittedly this point is controversial.26–28 Thus our results suggest that increases in the rate of forward electron flow in the mitochondria, which would occur with heightened myocardial oxygen metabolism, lead to the production of O2·−, which then is converted to the coronary vasodilator, H2O2. This conclusion is further strengthened by the results of studies using FCCP and apocyinin. Apocynin prevents cytosolic assembly of cytosolic NADPH oxidases, and reportedly does not affect mitochondrial O2·− production,29 which supports our conclusions that the mitochondria is the source of superoxide during pacing. FCCP is an ionophore that decreases the mitochondrial protonmotive force, which then uncouples mitochondrial electron transport and reduces O2·− production.23 Our observations also show that the increases in O2·− during pacing of isolated myocytes was decreased by FCCP—further supporting the concept that increased electron transport is the basis for increased superoxide production and, thus, hydrogen peroxide production during heightened myocardial metabolism.
Another issue that bears on our findings relates to the influences of nitric oxide on mitochondrial electron transfer and on the mitochondrial production of O2·− and H2O2. NO interferes with mitochondrial electron transfer at complex III, which increases the production of O2·− by this complex.30 Although the in vitro myocyte preparations are removed from the effect of endothelial derived NO, the intact heart is certainly subject to this effect. Thus in the intact heart, mitochondrial generation of O2·− and H2O2 may also be affected by the effects of NO on complex III. Although we cannot resolve the degree to which this effect contributes to the mitochondrial production of O2·− and H2O2 in vivo, we can state that the production of the reactive oxygen species in vivo or in vitro was directly linked to oxygen consumption.
Directly linking metabolic coronary vasodilation with a product of oxidative metabolism is a vastly different paradigm than the conventionally hypothesized negative feedback systems for metabolic flow control. We calculated that 23 nmol/L of H2O2 would be produced for every μmol/L of O2 consumed, thus linking H2O2 to myocardial oxygen consumption which is essential for a metabolic dilator. Moreover, the use of H2O2 for metabolic dilation appears to be an energetically conserved system, because the production of O2·− and H2O2 do not require additional energy for their production, and in effect, they are waste products of mitochondrial electron transfer. Within mitochondria, and between the inner and outer mitochondrial membranes, sufficient amounts of superoxide dismutase are present to metabolize O2·− into H2O222; thus the majority, if not all, O2·− produced during the course of oxygen metabolism will efflux from the mitochondria as H2O2. Although we did not measure H2O2 with all the interventions that affected mitochondrial electron transport, because of the high activities of mitochondrial superoxide dismutases, we would expect that the production of H2O2 would be proportional to O2·−. Indeed as our results suggest, myxothiazol and apocynin had similar effects, reduction and no effect, respectively, on the production of O2·− and H2O2 during pacing.
Taken together, our results suggest that the production of H2O2, which stems from the dismutation of O2·− that is formed during mitochondrial electron transport, is seminal in the coupling between oxygen metabolism to blood flow in the heart. Our results also demonstrate that the levels of H2O2 produced by cardiac myocytes, either in vitro or in vivo, are linked to oxygen metabolism, and are produced in sufficient amounts to be vasoactive. Also, enzymatic catabolism of H2O2 with catalase, or blockade of the dilator actions of H2O2, abrogate coronary metabolic dilation. Based on these results, we conclude that H2O2 serves as a feed-forward link between metabolism and blood flow in the heart.
Sources of Funding
The authors acknowledge the following sources of grant support from the National Institutes of Health (HL32788, HL65203, HL73755 [W.M.C.]; HL67804 [J.D.T.], COBRE RR18766 [C.Z., G.D.]), the American Heart Association (0455435B [C.Z.]), and Atorvastatin Research Award (2004-37 [C.Z.]).
S.S. and C.Z. contributed equally to this work.
Original received July 21, 2006; final version accepted September 22, 2006.
Feigl EO. Coronary Physiology. PhysiolRev. 1983; 63: 1–205.
Tune JD, Richmond KN, Gorman MW, Olsson RA, Feigl EO. Adenosine is not responsible for local metabolic control of coronary blood flow in dogs during exercise. Am Journal of Physiology - Heart & Circulatory Physiology. 2000; 278: H74–H84.
Tune JD, Gorman MW, Feigl EO. Matching coronary blood flow to myocardial oxygen consumption. J Appl Physiol. 2004; 97: 404–415.
Rubio R, Wiedmeier VT, Berne RM. Relationship between coronary flow and adenosine production and release. JMolCell Cardiol. 1974; 6: 561–566.
Rauchova H, Vokurkova M, Koudelova J. Developmental changes of erythrocyte catalase activity in rats exposed to acute hypoxia. Physiol Res. 2005.
Alexandra K. Adams TMB. The Role of Antioxidants in Exercise and Disease Prevention. The Physician and Sports Medicine. 2002; 30: 37–44.
Merkus D, Duncker DJ, Chilian WM. Metabolic regulation of coronary vascular tone: role of endothelin-1. Am J Physiol Heart Circ Physiol. 2002; 283: H1915–H1921.
Tiefenbacher CP, DeFily DV, Chilian WM. Requisite role of cardiac myocytes in coronary alpha1-adrenergic constriction. Circulation. 1998; 98: 9–12.
Rogers PA, Dick GM, Knudson JD, Focardi M, Bratz IN, Swafford J, Albert N, 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.
Knudson JD, Dincer UD, Dick GM, Shibata H, Akahane R, Saito M, Tune JD. Leptin Resistance Extends to the Coronary Vasculature in Prediabetic Dogs and Provides a Protective Adaptation against Endothelial Dysfunction. Am J Physiol Heart Circ Physiol. 2005; 289: H1038–1046.
Berne RM. The role of adenosine in the regulation of coronary blood flow. Circ Res. 1980; 47: 807–813.
Tune JD, Richmond KN, Gorman MW, Feigl EO. Control of coronary blood flow during exercise. Exper Biol Med. 2002; 227: 238–250.
Traverse JH, Nesmelov YE, Crampton M, Lindstrom P, Thomas DD, Bache RJ. Measurement of myocardial free radical production during exercise using EPR spectroscopy. Am J Physiol Heart Circ Physiol. 2006; 290: H2453–H2458.
Okado-Matsumoto A, Fridovich I. Subcellular distribution of superoxide dismutases (SOD) in rat liver: Cu,Zn-SOD in mitochondria. J Biol Chem. 2001; 276: 38388–38393.
Muller FL, Liu Y, Van Remmen H. Complex III releases superoxide to both sides of the inner mitochondrial membrane. J Biol Chem. 2004; 279: 49064–49073.
Lambert AJ, Brand MD. Inhibitors of the quinone-binding site allow rapid superoxide production from mitochondrial NADH:ubiquinone oxidoreductase (complex I). J Biol Chem. 2004; 279: 39414–39420.