Interactions between A_2A adenosine receptors, hydrogen peroxide, and K_ATP channels in coronary reactive hyperemia

Abstract

Myocardial metabolites such as adenosine mediate reactive hyperemia, in part, by activating ATP-dependent K⁺ (K_ATP) channels in coronary smooth muscle. In this study, we investigated the role of adenosine A_2A and A_2B receptors and their signaling mechanisms in reactive hyperemia. We hypothesized that coronary reactive hyperemia involves A_2A receptors, hydrogen peroxide (H₂O₂), and K_ATP channels. We used A_2A and A_2B knockout (KO) and A_2A/2B double KO (DKO) mouse hearts for Langendorff experiments. Flow debt for a 15-s occlusion was repaid 128 ± 8% in hearts from wild-type (WT) mice; this was reduced in hearts from A_2A KO and A_2A/_2B DKO mice (98 ± 9 and 105 ± 6%; P < 0.05), but not A_2B KO mice (123 ± 13%). Patch-clamp experiments demonstrated that adenosine activated glibenclamide-sensitive K_ATP current in smooth muscle cells from WT and A_2B KO mice (90 ± 23% of WT) but not A_2A KO or A_2A/A_2B DKO mice (30 ± 4 and 35 ± 8% of WT; P < 0.05). Additionally, H₂O₂ activated K_ATP current in smooth muscle cells (358 ± 99%; P < 0.05). Catalase, an enzyme that breaks down H₂O₂, attenuated adenosine-induced coronary vasodilation, reducing the percent increase in flow from 284 ± 53 to 89 ± 13% (P < 0.05). Catalase reduced the repayment of flow debt in hearts from WT mice (84 ± 9%; P < 0.05) but had no effect on the already diminished repayment in hearts from A_2A KO mice (98 ± 7%). Our findings suggest that adenosine A_2A receptors are coupled to smooth muscle K_ATP channels in reactive hyperemia via the production of H₂O₂ as a signaling intermediate.

the heart responds to acute ischemia by transiently increasing blood flow in a phenomenon called reactive hyperemia (8). This temporary reduction in coronary vascular resistance is mediated by chemical signals released into blood (38), including adenosine (35), which activates A₁, A_2A, A_2B, and A₃ receptors (29). Both A_2A and A_2B adenosine receptors (ARs) are expressed on endothelial and smooth muscle cells (2, 29, 33, 43). Pharmacological studies indicate that A_2A and A_2B receptors are the subtypes most likely involved in coronary reactive hyperemia (6, 10, 51); however, undesirable overlap in the pharmacological profiles of adenosine receptor antagonists obfuscates their relative contribution. This issue could be resolved with the gene-targeted approaches we have used previously to demonstrate pivotal roles for A_2A receptors in the regulation of coronary vascular function (28, 39, 41). Thus, to specifically determine the roles of A_2A and A_2B ARs in coronary reactive hyperemia, in addition to distinguishing their role from A₁ and A₃ ARs, we used Langendorff-perfused hearts from wild-type (WT), A_2A knockout (KO), A_2B KO, and A_2A/2B double knockout (DKO) mice, respectively.

Numerous mediators and end effectors of coronary reactive hyperemia have been proposed to function downstream of adenosine receptors, including H₂O₂ and ATP-dependent K⁺ (K_ATP) channels. The role of K_ATP channels in coronary reactive hyperemia is firmly established (3, 12), while the contribution of H₂O₂ and interactions with K_ATP channels are less clear. Some support for H₂O₂ in reactive hyperemia has been provided by studies of the reactive dilation of isolated coronary arterioles (18) and mesenteric reactive hyperemia (49). Additionally, it is reported that K_ATP channels mediate, to some degree, H₂O₂-induced dilation of arterioles from skeletal muscle and brain (24, 44). Thus we determined whether H₂O₂ couples adenosine receptor stimulation to K_ATP channel activity in coronary reactive hyperemia.

MATERIALS AND METHODS

Animals.

An Institutional Animal Care and Use Committee at West Virginia University School of Medicine approved all experimental protocols. We followed guidelines set forth by the American Physiological Society and National Institutes of Health regarding the care and use of laboratory animals. A_2A and A_2B KO mice, both backcrossed 12 generations to the WT C57BL/6 background (Jackson Laboratory; Bar Harbor, ME), were bred to generate A_2A/A_2B double heterozygotes. Double heterozygotes were intercrossed, 1/16 of the offspring were A_2A/A_2B DKO, and A_2A/A_2B knockout breeding pairs were established. Mice were caged in a 12:12-h light-dark cycle with free access to standard chow and water.

Langendorff-perfusion.

Mice (10–14 wk) were anesthetized with pentobarbital sodium (50 mg/kg ip) and hearts were excised into heparinized (5 U/ml) ice-cold Krebs-Henseleit buffer containing (in mM) 119 NaCl, 11 glucose, 22 NaHCO₃, 4.7 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 2.5 CaCl₂, 2 pyruvate, and 0.5 EDTA. Hearts beat spontaneously when retrogradely perfused (80 mmHg) with 37°C buffer bubbled with 95% O₂-5% CO₂. The left atrium was removed and the left ventricle drained. A fluid-filled balloon was inserted into the left ventricle and connected to a transducer for pressure measurements. Left ventricular diastolic pressure was adjusted to 2–5 mmHg. Coronary flow was measured with a probe (Transonic Systems; Ithaca, NY) in the aortic perfusion line. Hearts were paced to 420 beats/min and function allowed to stabilize. Hearts with persistent arrhythmias or developed pressure less than 80 mmHg were excluded from the study. Coronary flow and cardiac function were measured using a Power Lab data acquisition system (AD Instruments; Colorado Springs, CO). Hearts were subjected to 15 s of total flow occlusion to elicit hyperemia. Catalase, glibenclamide, and N^ω-nitro-l-arginine methyl ester (l-NAME) (Sigma Chemical; St. Louis, MO) were delivered into the aortic perfusion line using a microinjection pump (Harvard Apparatus; Holliston, MA) as 1% of coronary flow to achieve a final concentration of 1,250 U/ml (46, 49).

Patch clamp.

Single smooth muscle cells were enzymatically isolated from mouse aortas. For cell isolation, a HEPES-buffered saline was used and contained (in mM) 135 NaCl, 5 KCl, 1 MgCl₂, 0.36 CaCl₂, 10 glucose, 10 HEPES, and 5 Tris; pH 7.4. The aorta was placed in this solution plus (in mg/ml) 2 collagenase, 1 elastase, 2 bovine serum albumin, and 1 soybean trypsin inhibitor for 15 min at 37°C. The tissue was passed through the tip of a fire-polished Pasteur pipette to liberate single cells. Smooth muscle cells were resuspended in enzyme-free buffer, stored on ice, and used within 8 h. Cells were placed in a recording chamber atop an inverted microscope and perfused with a HEPES-buffered solution containing 140 mM K⁺. Patch pipettes (3–5 MΩ) were filled with 140 mM K⁺ solution supplemented with 5 mM Mg-ATP and 0.1 mM Na-GTP, pH 7.1. Inward K_ATP currents were recorded in the conventional whole cell mode (with Nernst equilibrium potential for K⁺ set to 0 mV). The membrane potential was held at −80 mV and ramped or stepped to potentials between −100 and +100 mV. Currents were low-pass filtered at 1 kHz and digitized at 5 kHz. Series resistance and whole cell capacitance were compensated as completely as possible through the amplifier (PC-505, Warner Instruments; Hamden, CT). This amplifier was interfaced to a computer with pClamp 9 software for data acquisition and analysis (Molecular Devices; Sunnyvale, CA). All currents are reported as glibenclamide-sensitive K_ATP conductance (and calculated as nS/pF); effects of drugs are represented relative to control.

Flow debt (equal to baseline flow rate × occlusion time) and repayment area [equal to integral of hyperemic area above baseline flow during the first minute of reactive hyperemia] were calculated as previously reported (6, 10, 51). Since absolute coronary flow rates change proportionally with heart mass and metabolic rate, the repayment area and debt area data are represented as milliliters per gram of wet heart weight (ml/g) and baseline and peak flow data are presented as milliliters per minute per gram of wet heart weight (ml·min⁻¹·g⁻¹). Each reactive hyperemia graph represents the average of all studies.

Statistical analysis.

Statistical analyses were made with t-test and one-way analysis of variance (ANOVA) as indicated. Results were considered significant when P < 0.05. Values are means ± SE from n number of animals.

RESULTS

A_2A receptors are important in coronary reactive hyperemia.

We used hearts from WT, A_2A and A_2B KO mice to test the involvement of adenosine receptor subtypes in coronary reactive hyperemia. Hearts from A_2A/2B DKO mice were used to test the combined role of A_2A and A_2B ARs known to induce coronary vasodilation. Baseline flow in WT hearts (n = 6) was 19.8 ± 0.9 ml·min⁻¹·g⁻¹, and 4.9 ± 0.2 ml/g of flow debt was incurred during a 15-s occlusion. Peak hyperemic flow was 42.7 ± 2.0 ml·min⁻¹·g⁻¹, and the repayment volume was 6.3 ± 0.4 ml/g. Thus flow debt was repaid 128 ± 8% in hearts from WT mice (Fig. 1). Repayment volume was reduced in A_2A KO mice (n = 8; Fig. 1). Specifically, while baseline flow (19.4 ± 1.6 ml·min⁻¹·g⁻¹), debt volume (4.9 ± 0.4 ml/g), and peak flow (42.9 ± 1.7 ml·min⁻¹·g⁻¹) were similar to WT, the repayment volume was only 4.5 ± 0.2 ml/g (P < 0.05 vs. WT). That is, flow debt was repaid only 98 ± 9% in hearts from A_2A KO mice (just 77 ± 7% of the response seen in WT).

Fig. 1.

Reactive hyperemia is reduced in hearts lacking the adenosine A_2A receptor. A: group data show baseline flow and the hyperemic response following a 15-s total occlusion of coronary flow. Responses were similar in hearts from wild-type (WT) (n = 6) and A_2B knockout (KO) (n = 5) mice, but were significantly reduced in hearts from A_2A KO (n = 6) and A_2A/A2B double knockout (DKO) (n = 8) mice. B: repayment of flow debt was reduced 22–29% in A_2A KO and A_2A/A2B DKO mice (*P < 0.05 vs. WT by one-way ANOVA; detailed in text) but not in A_2B KO (n = 5).

In contrast, hyperemic responses in hearts from A_2B KO mice (n = 5) were indistinguishable from those in WT (Fig. 1). There were no differences in baseline flow (20.3 ± 1.1 ml·min⁻¹·g⁻¹), debt volume (5.1 ± 0.3 ml/g), peak flow (44.2 ± 0.5 ml·min⁻¹·g⁻¹), or repayment volume (6.1 ± 0.1 ml/g). Flow debt in hearts from A_2B KO mice was repaid (123 ± 13%), representing 95 ± 9% of the response in hearts from WT mice. Responses in hearts from A_2A/2B DKO mice (n = 8) were similar to A_2A KO mice (Fig. 1). That is, repayment volume (4.9 ± 0.2 ml/g) and the repayment of flow debt (105 ± 6%) were reduced (P < 0.05 vs. WT). Resting flow (18.6 ± 0.6 ml·min⁻¹·g⁻¹), debt volume (4.7 ± 0.1 ml/g), and peak flow (.6 ± 1.2 ml·min⁻¹·g⁻¹) were not different from WT, but the repayment of flow debt in hearts from A_2A/2B DKO mice was only 82 ± 5% of the response seen in WT (Table 1).

Table 1.

Resting flow, debt area, peak flow, and flow repayment area

	Resting Flow. ml·min−1·g−1	Debt Area, ml/g	Peak Flow, ml·min−1·g−1	Repayment Area, ml/g
WT	19.8 ± 0.9	4.9 ± 0.2	42.7 ± 2	6.3 ± 0.4
A2A KO	19.4 ± 1.6	4.9 ± 0.4	42.9 ± 1.7	4.5 ± 0.2*
A2B KO	20.3 ± 1.1	5.1 ± 0.3	44.2 ± 0.5	6.1 ± 0.1
A2A/2B KO	18.6 ± 0.6	4.7 ± 0.1	41.6 ± 1.2	4.9 ± 0.2*
WT + Cat	21.4 ± 2.1	5.3 ± 0.5	38.6 ± 2.4*	4.4 ± 0.4*
A2A KO + Cat	16.8 ± 0.9*	4.2 ± 0.2*	36.3 ± 1.4*	4.1 ± 0.4*
WT + l-NAME	11.7 ± 0.8*	2.9 ± 0.24*	31.9 ± 0.9*	2.7 ± 0.3*
WT + l-NAME + Cat	11.2 ± 1.3*	2.8 ± 0.30*	26.8 ± 1.6*	1.9 ± 1.2*

All values are means ± SE.

WT, wild type; KO, knockout; Cat, catalase; l-NAME, N^ω-nitro-l-arginine methyl ester.

^*P < 0.05 compared with WT (n = 5–7).

Adenosine activates K_ATP channels through A_2A receptors.

We tested the effect of K_ATP channel blockade on coronary reactive hyperemia response of WT isolated hearts. Glibenclamide (10 μM) significantly reduced WT (n = 5) baseline flow from 17.3 ± 0.5 to 12.8 ± 0.5 ml·min⁻¹·g⁻¹ and flow debt from 4.3 ± 0.1 to 3.2 ± 0.1 ml/g. Peak hyperemic flow and the repayment volume were reduced by glibenclamide from 35.9 ± 0.8 to 29.3 ± 1.1 ml·min⁻¹·g⁻¹ and 6.9 ± 0.4 to 2.7 ± 0.4 ml/g, respectively. Thus flow debt was repaid 161 ± 12% in hearts from WT mice while this was reduced to 87.5 ± 11.2% in the presence of glibenclamide (Fig. 2).

Fig. 2.

Reactive hyperemia is attenuated in the presence of glibenclamide. A: group data present baseline flow and the hyperemic response following a 15-s total occlusion of coronary flow. Responses were significantly reduced in hearts with K_ATP channel blockade by glibenclamide (10 μM, n = 5). B: repayment of flow debt was reduced by 54.4% in WT mice treated with glibenclamide (*P < 0.05 vs. WT by unpaired t-test) (n = 5).

We isolated smooth muscle cells from the aortas of WT, A_2A and A_2B KO, and A_2A/2B DKO mice and used whole cell patch-clamp to determine whether A_2A receptors couple to K_ATP channels. Recordings were made in symmetrical 140 mM K⁺ so that inward linear K_ATP current was more easily resolved. Effects of adenosine (10 μM) were compared with pinacidil (10 μM) as a positive control. Further, adenosine-activated currents were blocked with glibenclamide (10 μM) to confirm that K_ATP channels mediated them. Adenosine activated current in smooth muscle cells from WT mice (Fig. 3A), but not A_2A KO mice (Fig. 3B). Responses in smooth muscle cells from A_2B KO mice (Fig. 3C) were similar to WT, while responses in cells from A_2A/2B DKO mice (Fig. 3D) were like those from A_2A KO mice. Group data (cells from n = 4–5 mice in each group; Fig. 3E) show that adenosine-induced K_ATP current was diminished in smooth muscle cells from mice lacking the adenosine A_2A receptor (P < 0.05 vs. WT by one-way ANOVA). Importantly, however, functional K_ATP channel expression was not different as pinacidil-activated K_ATP current was similar in smooth muscle cells from all four strains of mice.

Fig. 3.

Adenosine-induced K_ATP current is reduced in smooth muscle cells lacking the adenosine A_2A receptor. Representative traces show that adenosine (10 μM) activated glibenclamide-sensitive K_ATP current in smooth muscle cells from WT (A) and A_2B KO (C) mice, but not A_2A KO (B) or A_2A/A2B (D) DKO mice. For group data (E), K_ATP conductance (nS/pF) was normalized to control for cells from n = 4–5 mice in each group. *P < 0.05 vs. WT by one-way ANOVA.

H₂O₂ activates K_ATP current in smooth muscle.

We isolated smooth muscle cells from the aortas of WT mice and used whole cell patch clamp to determine whether H₂O₂ activates K_ATP current. Current was recorded in the symmetrical 140 mM K⁺ solutions described above. Figure 4A shows current sensitive to glibenclamide (10 μM) under control conditions and when activated by H₂O₂ (1 mM) and pinacidil (10 μM). Group data (cells from n = 5 mice) are shown in Fig. 4B; K_ATP conductance (nS/pF) was 0.018 ± 0.006 under control conditions and significantly increased to 0.070 ± 0.026 and 0.142 ± 0.033 with 1 mM H₂O₂ and with 10 μM pinacidil, respectively (both P < 0.05 vs. control by one-way ANOVA). To determine the role of K_ATP channels in H₂O₂-mediated coronary vasodilation, we examined the effect of glibenclamide (10 μM) on changes in flow induced by H₂O₂ (10 μM). In our isolated heart system, glibenclamide significantly decreased the baseline coronary flow (79 ± 6%) and consistent with the patch-clamp data, our data showed that glibenclamide significantly reduced the H₂O₂-induced increase in coronary flow from 164 ± 16 to 79 ± 6% (n = 4, Fig. 4C).

Fig. 4.

H₂O₂ activates K_ATP channels in smooth muscle cells. A: representative trace showing effect of H₂O₂ (1 mM) to increase K_ATP current. B: group data (cells from n = 5 mice) illustrate effect of H₂O₂ on K_ATP current relative to pinacidil (10 μM). C: group data (n = 4) demonstrate the effect of glibenclamide (10 μM) on H₂O₂-mediated increase in coronary flow of WT hearts. * and # indicate P < 0.05 vs. WT and H₂O₂ effect, respectively.

Adenosine-induced coronary dilation involves H₂O₂.

We used catalase to determine the role of endogenous H₂O₂ production in adenosine-induced coronary vasodilation. Adenosine-induced coronary vasodilation is catalase-sensitive (Fig. 5A), indicating a role for endogenous H₂O₂ production. Specifically, adenosine (1 μM) increased coronary flow (284 ± 53%) and catalase (1,250 U/ml) reduced the baseline flow to 86 ± 2%; however, catalase reduced the adenosine response to 89 ± 13% (P < 0.05 vs. control by paired t-test). Catalase reduced left ventricular developed pressure (Fig. 5C), but had no effect on heart rate (Fig. 5B). These data suggest that 1) H₂O₂ is involved in adenosine-mediated effects on cardiac contractility or 2) cardiac function decreased secondary to a reduction in coronary flow. Since we have previously shown that adenosine does not increase the coronary flow in A_2A/2B KO mice (39), this experiment was not performed on double knockout mice.

Fig. 5.

H₂O₂ dilates the coronary circulation and adenosine-induced vasodilation is catalase-sensitive. A–C show group data (n = 4) illustrating the catalase sensitivity of adenosine-induced coronary dilation (A), contractility (C), and heart rate (B) in WT mice. LVDP, left ventricular developed pressure. *P < 0.05 vs. adenosine alone.

H₂O₂ contributes to coronary reactive hyperemia.

We used hearts from WT and A_2A KO mice to determine whether H₂O₂ links A_2A receptor activation to coronary flow during reactive hyperemia. To do so, reactive hyperemia in hearts from both strains was compared with and without catalase (1,250 U/ml) to degrade H₂O₂ (Fig. 6). In hearts from WT mice, catalase significantly decreased the repayment of flow debt (from 128 ± 8 to 84 ± 9%; P < 0.05; n = 6). In hearts from A_2A KO mice, catalase had no effect on the already diminished repayment of flow debt (98 ± 9 vs. 98 ± 7%; n = 6, Table 1).

Fig. 6.

Catalase (Cat)-sensitive portion of reactive hyperemia is absent in hearts lacking the adenosine A_2A receptor. A: group data show baseline flow and hyperemic responses in hearts from WT and A_2A KO mice in the absence or presence of catalase (1,250 U/ml). Hyperemic responses were reduced by catalase in hearts from WT (n = 7) but not A_2A KO (n = 6) mice. Catalase reduced the repayment of flow debt 30% in hearts from WT (B) mice (*P < 0.05 by unpaired t-test), but had no effect in hearts from A_2A KO mice (C).

H₂O₂ contributes to coronary reactive hyperemia independent of nitric oxide (NO).

Adenosine A_2A and A_2B receptors can also signal through NO (32). NO is a well-known vasodilator involved in regulating baseline coronary flow and reactive hyperemia. Thus, to separate the role of H₂O₂ from NO, we performed reactive hyperemia experiments on hearts from WT mice with l-NAME (50 μM, a NO synthase inhibitor) and l-NAME plus catalase (Fig. 7). l-NAME significantly decreased the repayment of flow debt (73 ± 8% of control; n = 5, P < 0.05). Importantly, however, addition of catalase to l-NAME further reduced the repayment of flow debt (57 ± 9%; P < 0.05). These data indicate that the effect of catalase is on H₂O₂ signaling, not NO signaling.

Fig. 7.

Catalase-sensitive portion of reactive hyperemia is independent of nitric oxide. A (n = 4–9) shows group data of baseline flow and hyperemic responses in hearts from WT mice in the presence and absence of N^ω-nitro-l-arginine methyl ester (l-NAME) (5 × 10⁻⁵ M, nitric oxide synthase inhibitor) or l-NAME + catalase (1,250 U/ml). B shows the further reduced hyperemic response in the presence of both l-NAME and catalase (68.5%) relative to l-NAME alone (53.1%) compared with WT. *P < 0.05 and #P < 0.05 vs. control and l-NAME effect, respectively.

DISCUSSION

Our study provides several new substantial lines of evidence regarding mechanisms of coronary reactive hyperemia. First, using gene-targeted mice, we offer a definitive comparison of the roles of A_2A and A_2B receptors in coronary reactive hyperemia. Our finding that A_2A receptor plays a dominant role in adenosine-mediated coronary reactive hyperemia substantially advances knowledge gained from studies using adenosine receptor antagonists (6, 10, 51). Second, with our catalase experiments, we provide the first evidence of a role for H₂O₂ in reactive hyperemia in the intact heart. These data add significantly to previous observations of roles for H₂O₂ in mesenteric reactive hyperemia (49) and reactive dilation in isolated coronary arterioles (18). Third, through a combination of experiments, we offer the original observation that H₂O₂ couples A_2A receptor stimulation to K_ATP channel activation in coronary reactive hyperemia.

Previously, two general strategies have been employed to determine the role of adenosine and its receptors in coronary reactive hyperemia: 1) enzymatic degradation of adenosine (e.g., infusing adenosine deaminase) or 2) pharmacological antagonism of adenosine receptors (e.g., administering aminophylline). With these tools, it had been demonstrated previously that adenosine plays a small, but significant role, in coronary reactive hyperemia in dogs, lambs, and pigs (4, 6, 10, 23). Because two of the four known adenosine receptor subtypes, A_2A and A_2B, are linked to coronary vasodilation, a logical question centers on which receptor subtype(s) is/are involved. Gene deleted knockout mice have been an important tool to dissect physiological signaling pathways, e.g., elucidating the role of receptor through its absence. Using PCR, we confirmed the absence of A_2A and A_2B receptors in knockout mouse tail, aorta and coronary arteries (data not shown). Therefore, the present results obtained using adenosine receptor knockout mice support the view that adenosine and the A_2A receptor mediate a portion of coronary reactive hyperemia.

A_2A and A_2B receptors both function in signaling in coronary smooth muscle and endothelium (31, 37, 40), but the relative role of these two receptor subtypes in coronary reactive hyperemia has remained, until now, obscure. Using an antagonist of A_2A receptors (SCH58261), it was suggested that 30–50% of flow repayment could be attributed to A_2A receptor activation (6, 51). Our results with hearts from A_2A KO and A_2A/_2B DKO mice indicate that the A_2A receptor is responsible for 22–29% of flow repayment, which is toward the lower end of the previous pharmacological estimates. Our results from A_2B KO and A_2A/2B DKO mice did not suggest a significant role for A_2B receptors in reactive hyperemia, similar to previous pharmacological studies (6, 51). Due to its low affinity for adenosine, A_2B receptors are thought to be activated under more severely ischemic conditions, such as during ischemia/reperfusion and preconditioning (13, 15).

We tested the involvement of H₂O₂ in reactive hyperemia and adenosine-induced coronary vasodilation. We demonstrate that catalase, which breaks down H₂O₂ into water and oxygen, significantly attenuates adenosine-induced coronary vasodilation. This suggests that H₂O₂ plays a critical role in adenosine-mediated signaling pathways. We also demonstrate that the repayment of flow debt was reduced by catalase, showing the contribution of endogenous H₂O₂ production to coronary reactive hyperemia. Furthermore, we suggest that H₂O₂ released during coronary reactive hyperemia is, at least partly, mediated through activation of A_2A adenosine receptors since catalase did not further decrease the flow repayment in A_2A KO hearts. This is not likely due to changes in H₂O₂ degradation in A_2A KO mice, as the concentration-response curve to exogenous H₂O₂ (100 nM to 80 μM) is unchanged compared with WT control or A_2B KO (data not shown). In the vasculature, NAD(P)H oxidases (Nox) are responsible for the majority of ROS production (21). Nox4 is shown to be the dominant isoform producing H₂O₂ and also expressed in vascular smooth muscle cells (45). Therefore, expression level of factors involved in vascular H₂O₂ production such as NOX4 in A_2A KO remain to be examined in future studies. Nevertheless, the idea that H₂O₂ is a mediator of coronary metabolic (36, 48) and endothelium-dependent (22, 27) vasodilation is gaining traction; however, no previous study had indicated a role for H₂O₂ in intact coronary reactive hyperemia. Two related observations had, however, been made in other vascular beds. Video microscopy of the cremasteric microcirculation indicated striking similarities in the pharmacological sensitivity of reactive hyperemia and exogenous H₂O₂ (46). A study using microspheres in the mouse mesenteric circulation indicated that ∼40% of hyperemic flow was catalase-sensitive, i.e., attributable to endogenous H₂O₂ production (49). Our data are in agreement and indicate that ∼35% of hyperemic flow in the mouse heart is catalase-sensitive.

We demonstrated previously that A_2A-mediated coronary vasodilation depends upon K_ATP channels, as it is blocked by glibenclamide (39). Here we show that 1) adenosine-induced coronary vasodilation depends upon H₂O₂, 2) H₂O₂ increases K_ATP channel current, and 3) increase in coronary flow by H₂O₂ is mediated through K_ATP channels. The latter finding extends the knowledge that K_ATP channels mediate, at least in part, H₂O₂-induced dilation of arterioles from skeletal muscle and brain (20, 24, 44). Further, it has been shown that H₂O₂ activates K_ATP channels in pancreatic beta cells and cardiac myocytes (16, 30). Nevertheless, the evidence of this mechanism in smooth muscle and the role of adenosine were lacking. We show for the first time that K_ATP current in smooth muscle cells is activated by H₂O₂. This adds to previous literature indicating that H₂O₂ activates K⁺ channels in coronary vascular smooth muscle cells including Ca²⁺-activated K⁺ channels (5, 42, 52) and voltage-dependent K⁺ channels (34). Additionally, glibenclamide and catalase both decreased the baseline coronary flow, with glibenclamide effect being about 7% more than catalase, suggesting that K_ATP channels may have additional pathways of activation in addition to adenosine-induced-H₂O₂ although H₂O₂ may be the major mediator in this activation.

Perhaps a better approach would be to study smooth muscle cells isolated from mouse coronary arteries or arterioles. This is because differences between large- and small-caliber vessels exist (e.g., effects on intravascular pressure and relative responses to metabolic, neurohumoral, myogenic, and flow-mediated mechanisms) (7, 11). Moreover, it has been demonstrated that differences in K⁺ channel expression may explain some segmental differences in vascular reactivity (1, 25). Importantly, however, there is no a priori reason to assume that, in this context, K_ATP channels in smooth muscle cells from the aorta and coronary arterioles would respond differently to adenosine and H₂O₂ signaling. Specifically, both conduit and resistance arteriolar smooth muscle cells express K_ATP channels (17, 19), adenosine A₂ receptors (9, 47), and relax or dilate to H₂O₂ (26, 27). In fact, we demonstrate here that K_ATP electrophysiology from aortic smooth muscle cells (Figs. 3 and 4) matches very well with functional responses from coronary resistance vessels (Fig. 4C). Thus, until demonstrated otherwise, it is justifiable to conclude that aortic smooth muscle cells serve as an excellent electrophysiological model for coronary arteriolar dilation, at least with regard to adenosine signaling, reactive oxygen species, and K_ATP channel regulation.

With regard to the limitations of our study, since we have used isolated heart system, the contribution of flow/shear stress- or pressure-dependent coronary mechanisms in the development of reactive hyperemia could not be clearly elucidated or isolated from the effect of local metabolic factors. Additionally, Krebs buffer used in crystalloid-perfused mouse heart, unlike blood, does not have O₂-carrying capacity and hence some may suggest that the heart would be ischemic (by definition) or hypoxic compared with circulating blood. Regardless, constantly oxygenated ex vivo isolated heart model is a well-established method that has been used to answer mechanistic questions. Data gathered from isolated heart reflect the overall end effect of shear/pressure stress- and metabolites-induced effects via eliminating the complexity of the role of neuronal and hormonal effects during coronary reactive hyperemia, thus allowing a better understanding of the coronary flow regulation.

Our study provides novel data to further understand mechanisms of coronary reactive hyperemia and adenosine-induced vasodilation. Using adenosine receptor KO mice, we show that A_2A receptors, but not A_2B receptors, are critical to coronary reactive hyperemia. This finding supports and extends previous studies using pharmacological tools. Additionally, we demonstrate a role for H₂O₂ in reactive hyperemia in the intact heart. This finding supports observations made in the microcirculation of skeletal muscle and brain and represents the first such report for the coronary circulation. Finally, our data suggest that H₂O₂ is a signaling molecule coupling the stimulation of A_2A receptors to the opening of K_ATP channels. This finding is entirely novel and will require additional research to unravel molecular mechanisms linking A_2A receptor activation to H₂O₂ production and finally to K_ATP channel opening. At present, a list of likely mechanisms might include A_2A-induced activation of NADPH oxidases (14), bioconversion of superoxide to H₂O₂, and redox regulation of the K_ATP channel or proteins that regulate it (50).

GRANTS

National Heart, Lung, and Blood Institute Grants HL-027339, HL-094447, HL-071802, and T32-HL-090610 supported this project.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: M.S.S. conception and design of research; M.S.S., X.Z., S.A., and G.M.D. performed experiments; M.S.S., X.Z., S.A., and G.M.D. analyzed data; M.S.S., X.Z., S.A., B.T., G.M.D., and S.J.M. interpreted results of experiments; M.S.S., X.Z., and G.M.D. prepared figures; M.S.S. and G.M.D. drafted manuscript; M.S.S., X.Z., B.T., G.M.D., and S.J.M. edited and revised manuscript; M.S.S., X.Z., S.A., S.L.T., C.L., B.T., G.M.D., and S.J.M. approved final version of manuscript.