Involvement of NADPH oxidase in A_2A adenosine receptor-mediated increase in coronary flow in isolated mouse hearts

Abstract

Adenosine increases coronary flow mainly through the activation of A_2A and A_2B adenosine receptors (ARs). However, the mechanisms for the regulation of coronary flow are not fully understood. We previously demonstrated that adenosine-induced increase in coronary flow is in part through NADPH oxidase (Nox) activation, which is independent of activation of either A₁ or A₃ARs. In this study, we hypothesize that adenosine-mediated increase in coronary flow through Nox activation depends on A_2A but not A_2BARs. Functional studies were conducted using isolated Langendorff-perfused mouse hearts. Hydrogen peroxide (H₂O₂) production was measured in isolated coronary arteries from WT, A_2AAR knockout (KO), and A_2BAR KO mice using dichlorofluorescein immunofluorescence. Adenosine-induced concentration-dependent increase in coronary flow was attenuated by the specific Nox2 inhibitor gp91 ds-tat or reactive oxygen species (ROS) scavenger EUK134 in both WT and A_2B but not A_2AAR KO isolated hearts. Similarly, the A_2AAR selective agonist CGS-21680-induced increase in coronary flow was significantly blunted by Nox2 inhibition in both WT and A_2BAR KO, while the A_2BAR selective agonist BAY 60-6583-induced increase in coronary flow was not affected by Nox2 inhibition in WT. In intact isolated coronary arteries, adenosine-induced (10 μM) increase in H₂O₂ formation in both WT and A_2BAR KO mice was attenuated by Nox2 inhibition, whereas adenosine failed to increase H₂O₂ production in A_2AAR KO mice. In conclusion, adenosine-induced increase in coronary flow is partially mediated by Nox2-derived H₂O₂, which critically depends upon the presence of A_2AAR.

Introduction

Coronary flow is tightly regulated to maintain a consistently high level of myocardial oxygen extraction over a wide range of myocardial demands [1–3]. This tight regulation is dependent on numerous vasoconstrictor and vasodilator influences, exerted by the autonomic nervous system, endothelium, and myocardium [3–6]. Adenosine, a well-known locally released metabolite, has been postulated as one of the important agents responsible for coronary vascular tone regulation in various conditions [3, 5, 7–9], among which adenosine has also been thought to regulate resting coronary flow [10, 11]. However, the role of adenosine in regulating resting coronary flow remains controversial [3]. This discrepancy may be due to different animal models, differences in species, and/or different agonists and antagonists used in these studies. It is well established that the coronary effects of adenosine are mediated through the activation of its four subtypes of receptors, namely A₁, A_2A, A_2B, and A₃ receptors (ARs) [9, 12, 13]. The role of these receptors has been studied in various species [12]. With an overall coronary vasodilation, A_2AAR predominantly and A_2BAR minimally contribute to dilating coronary vasculature [14–16], whereas A₁AR and A₃AR counteract the effects of A_2A/A_2BARs resulting in a diminished coronary vasodilation [15, 17, 18]. At post-receptor levels, several effector pathways of adenosine-mediated coronary flow have been reported, such as activation of nitric oxide (NO) pathway [19, 20], cyclic adenosine 5′-monophosphate (cAMP)-dependent pathway [21], activation of potassium channels [9, 21], and the involvement of H₂O₂ [19]. However, the downstream effectors linked to the activation of ARs in this process are not completely understood.

H₂O₂, a molecule within the reactive oxygen species (ROS) family, has been proposed to serve as a pivotal vasodilator in coronary flow regulation in human, canine, porcine, and murine coronary vasculature [19, 22–26]. NADPH oxidases (Nox) are the major source of ROS in the vasculature that plays both physiological and pathophysiological roles in the control of vascular tone [27–29]. The Nox family consists of seven members, Nox1-5, Doux1, and Doux2, among which Nox2 (gp91 Phox) forms the major source of H₂O₂ in cells stimulated with growth factors or cytokines under normal circumstances [30, 31]. Indeed, in the coronary vasculature, Nox2 has been proposed to be a functionally relevant source of H₂O₂ that mediates agonist-induced vasodilation [32]. Several recent studies have shown the interaction between adenosine and ROS via the regulation of Nox activity. For instance, inhibition or deletion of Nox2 leads to the attenuation of adenosine vasoconstrictor responses in aortas [33] and renal arterioles [34] and adenosine vasodilator responses in cerebral arteries [35]. However, little is known about the involvement of Nox2-derived ROS in adenosine-mediated flow response in the coronary circulation. Although activation of A₁ and A₃AR has been shown to have vascular effects through ROS [33, 36], our recent studies have already excluded the involvement of A₁ and A₃ARs in ROS-mediated regulation of coronary flow [37]. Therefore, with the focus on the adenosine signaling, the first aim of the present study was to further investigate whether A_2A and/or A_2BAR are involved in Nox2-derived ROS-mediated coronary vasodilation. Given the previous findings that adenosine A_2AAR coupled to H₂O₂ contributes to coronary reactive hyperemia [38] and Nox2 acts as a functional source for H₂O₂ resulting in coronary vasodilation [32], the second aim of the present study was to examine whether Nox2-derived ROS production contributing to adenosine-mediated coronary flow is through H₂O₂ and which adenosine receptors are required in this process. We performed exogenous infusion of adenosine in Langendorff-perfused isolated hearts as well as immunohistochemistry in isolated coronary arteries from wild type (WT), A_2AAR knockout (KO), and A_2BAR KO mice for these studies.

Methods

Materials

EUK134 was purchased from Cayman Chemical (Ann Arbor, MI, USA), gp91 ds-tat was purchased from Anaspec (Fremont, CA, USA), and BAY 60-6583 was obtained as a gift from Bayer AG (Leverkusen, Germany). All other chemicals (adenosine, CGS-21680 and 2′,7′-dichloro-fluorescein diacetate (DCFH-DA)) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Stock solutions of adenosine, BAY 60-6583, CGS-21680, EUK134, and DCFH-DA were made in dimethyl sulfoxide (DMSO), whereas gp91 ds-tat was dissolved in distilled water.

Animals

All experimental protocols were approved by the Institutional Animal Care and Use Committee at West Virginia University School of Medicine. We followed guidelines set forth by National Institutes of Health regarding the care and use of laboratory animals. WT mice (C57BL/6 background) were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). A_2A and A_2BAR KO mice on an inbred C57BL/6 background were generously provided by Dr. Catherine Ledent (Universite Libre de Bruxelles, Belgium) and Dr. Stephen Tilley (University of North Carolina, Chapel Hill, NC, USA), respectively. Mice were kept in cages with 12:12-h light-dark cycles and maintained on a standard laboratory diet with access to water ad libitum. The absence of A_2A and A_2BARs at mRNA level in A_2AAR KO and A_2BAR KO mice has been confirmed by our previous studies using PCR in isolated aortas, mesenteric arteries, and coronary arteries [16, 20, 39–41]. With difficulties in collecting large amount of tissues for Western blot, one of our recent studies has confirmed the absence of A_2AAR protein level in A_2AAR KO mice in isolated coronary arteries using immunohistochemistry [40]. Our functional data also confirmed the lack of CGS-21680 (a selective A_2AAR agonist)-induced coronary flow response in isolated hearts of A_2AAR KO mice and the lack of BAY 60-6583 (a selective A_2B AR agonist)-induced coronary flow response in isolated hearts of A_2BAR KO mice [16, 20].

Langendorff-perfused mouse heart preparation

Mice (14–18 weeks) of either sex were anesthetized with pentobarbital sodium (50 mg kg⁻¹, i.p.). Mice were weighed before hearts were rapidly removed into heparinized (5 U ml⁻¹) 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. After removal of the surrounding tissue, the aorta was rapidly cannulated with a 20-gauge, blunt-ended needle; then, the heart was continuously perfused with 37 °C buffer bubbled with 95 % O₂/5 % CO₂ at a constant perfusion pressure of 80 mmHg [20, 40]. Subsequently, through an opening in the left atrium, a fluid-filled balloon made of plastic wrap was inserted into the left ventricle across the mitral valve. The balloon was connected to a pressure transducer for continuous measurement of left ventricular developed pressure (LVDP) and heart rate. The heart was then immersed in a water-jacketed perfusate bath maintained at 37 °C and beats spontaneously. Coronary flow was continuously measured with an ultrasonic flow probe (Transonic Systems, Ithaca, NY, USA) placed in the aortic perfusion line. A PowerLab Chart data acquisition system (AD Instruments, Colorado Springs, CO, USA) was used for data acquisition. Hearts were allowed to equilibrate for 30 to 45 min before starting experimental protocols. Hearts with persistent arrhythmias were excluded from the study. As a positive control, a 15-s flow occlusion resulted in a twofold increase in coronary flow over baseline [19, 38], indicating that buffer-perfused hearts have minimum hypoxia and coronary vessels have the capacity to further dilate.

Langendorff experimental protocols

After equilibrium, baseline coronary flow, heart rate (HR), and LVDP were measured. Adenosine concentration response curves (10⁻⁸–10⁻⁵ M) were acquired in perfused hearts from WT, A_2AAR KO, and A_2BAR KO mice. Each concentration of adenosine was infused for 5 min followed by a minimum of 5 min of perfusion for drug washout [16, 37]. In separate experiments, the specific Nox2 inhibitor gp91 ds-tat (1 μM) [42] or the superoxide dismutase and catalase-mimicking drug EUK134 (50 μM) was perfused for 20 min before acquiring adenosine concentration response curves (10⁻⁸–10⁻⁵ M) in perfused hearts from those mice [37].

In another two experimental groups, the selective A_2AAR agonist CGS-21680 concentration response curves (10⁻¹⁰–10⁻⁶ M) were acquired in perfused hearts from WT, A_2AAR KO, and A_2BAR KO mice, while the selective A_2BAR agonist BAY 60-658 concentration response curves (10⁻¹⁰–10⁻⁶ M) were acquired in perfused hearts only from WT mice. In separate experiments, the specific Nox2 inhibitor gp91 ds-tat (1 μM) was perfused for 20 min before acquiring CGS-21680 and BAY 60-658 concentration response curves (10⁻¹⁰–10⁻⁶ M), respectively [16, 37]. All compounds were infused at a rate of 1/100 ml min⁻¹ of the coronary flow through an injection port directly proximal to the aortic cannula using a microinjection pump (Harvard Apparatus, Holliston, MA, USA) [16, 20, 37].

Fluorescence detection of H₂O₂ in mouse coronary arteries

Left coronary arteries (with diameter of 50–120 μm) were isolated from WT, A_2AAR KO, and A_2BAR KO mice. Then, coronary arteries were incubated in DCFH-DA (10 μM) prepared in DMEM (ATCC, Manassas, VA, USA) for 30 min at 37 °C followed by 10-min wash [40]. Vessels were then pinned on a layer of silicone gel lying on a plastic culture dish and incubated with DMEM buffer maintained at 37 °C. Baseline control images were obtained by optical xyz sectioning with a vertical depth of 1 μm using a Zeiss water immersion objective (W N-AchroplanX40/0.75 numerical aperture) on a confocal microscope (LSM 510, Zeiss, Heidelberg, Germany). Once the baseline control images were acquired, adenosine (10 μM) was added into the culture dish and images were taken at 5 min (in some vessels at both 5 and 10 min) after adenosine incubation. We previously showed that a maximal fluorescence intensity of DCFH-DA was reached after 15 min of adenosine incubation [40], and there was no difference in fluorescence intensity between 5 and 10 min in the present study (data not shown). The data at 5 min after adenosine incubation was used for analysis. In separate sets of experiments, the arteries were treated with the Nox2 inhibitor gp91 ds-tat (1 μM) for 20 min and baseline control images were obtained before the addition of adenosine. After 5 min of adenosine incubation, additional images were taken. Hydrogen peroxide (200 μM) served as the positive control at the end of experiments [37, 40]. Only the bottom half of vessels (more proximal to the scanner) were scanned by xyz sectioning, in order to avoid the quenching as much as possible. ImageJ software was used for fluorescence image analysis. Stacks of regions of interest (ROIs) were selected on the basis of the outline of the vessel indicated by the fluorescence.

Statistical analysis

Langendorff baseline data for WT, A_2AAR KO, and A_2BAR KO groups were compared using one-way ANOVA followed by post hoc analysis using Bonferroni’s test. The effects of drug treatment on concentration response curves of adenosine, CGS-21680, and BAY 60-6583 were analyzed using two-way ANOVA followed by post hoc analysis using Bonferroni’s test. Since the absolute coronary flow changes proportionally with heart mass, the coronary flow was presented as ml min⁻¹ g⁻¹ wet heart weight [16, 37, 43]. With regard to imaging, the mean fluorescence intensity of each ROI (including those on both endothelial cells and smooth muscle cells) was calculated by subtraction of the background signal and changes in fluorescence intensity were presented as a ratio normalized to control [40]. Image data were analyzed using paired t test. All the data are presented as mean ± SEM; n represents the number of animals. Statistical significance was accepted when P < 0.05.

Results

Baseline function in isolated WT, A_2AAR KO, and A_2BAR KO mouse hearts

Table 1 summarizes the baseline functional parameters for heart rate, LVDP, and coronary flow in WT, A_2AAR KO, and A_2BAR KO mice after 30 min of equilibration of isolated hearts. Average heart weight to body weight ratio was significantly different between A_2AAR KO and A_2BAR KO mice, which is due to a heavier body weight in A_2BAR KO mice as compared to A_2AAR KO mice (27 ± 0.7 g in A_2BAR KO mice vs. 24 ± 0.5 g in A_2AAR KO mice), while the heart weight between the groups was comparable. However, no significant differences were observed in heart rate, LVDP, and baseline coronary flow from those mice (P > 0.05, by one-way ANOVA) (Table 1).

Table 1.

Baseline data for WT, A_2AAR KO, and A_2BAR KO mouse hearts

	WT (n = 21)	A2AAR KO (n = 15)	A2BAR KO (n = 17)
Age, week	16 ± 0.3	15 ± 0.3	16 ± 0.4
No. of mice	21	15	17
BW, g	25 ± 0.8	24 ± 0.5	27 ± 0.7
HW, mg	96.2 ± 2.6	99.1 ± 2.6	98.4 ± 2.4
HW/BW, %	0.39 ± 0.01	0.42 ± 0.01	0.37 ± 0.02*
CF, ml min−1 g−1	18 ± 0.6	18 ± 0.9	20 ± 0.4
HR, beats min−1	437 ± 10	391 ± 8	446 ± 10
LVDP, mmHg	90 ± 4	90 ± 5	80 ± 4

All parameters were collected after 30 min of equilibration in a Langendorff preparation. Values are means ± SEM

WT wild type, KO knockout mice, BW body weight, HW heart weight, CF coronary flow, HR heart rate, LVDP left ventricular developed pressure

*P < 0.05 compared to A_2AAR KO

Effects of Nox2 inhibition and A_2AAR/A_2BAR deletion on adenosine-induced increase in coronary flow

Adenosine produced a concentration-dependent increase in coronary flow in WT mice (E_max: 38.13 ± 0.78 ml min⁻¹ g⁻¹, data obtained from both Figs. 1a and 2a), which was significantly attenuated by Nox2 inhibition with gp91 ds-tat (Fig. 1a). In agreement with our previous studies [37], these findings indicate that adenosine-induced increase in coronary flow in WT mice is mediated in part through Nox2 activation.

Fig. 1.

Effects of NADPH oxidase (Nox2) inhibition on adenosine-induced increase in coronary flow. Shown are the effects of gp91 ds-tat (1 μM) on adenosine (Ado) concentration response curves (10⁻⁸ to 10⁻⁵ M) in isolated hearts of wild type (WT) (a, n = 6), A_2A receptor (AR) knockout (KO) (b, n = 6), and A_2BAR KO mice (c, n = 6). Values are mean ± SEM. *Significant difference vs. corresponding control points using two-way ANOVA followed by post hoc analysis using Bonferroni’s test (P < 0.05)

Fig. 2.

Effects of ROS scavenging on adenosine-induced increase in coronary flow. Shown are the effects of EUK134 (50 μM) on adenosine (Ado) concentration response curves (10⁻⁸ to 10⁻⁵ M) in isolated hearts of WT (a, n = 5), A_2AAR KO (b, n = 4), and A_2BAR KO mice (c, n = 6). Values are mean ± SEM. *Significant difference vs. corresponding control points using two-way ANOVA followed by post hoc analysis using Bonferroni’s test (P < 0.05)

Notably, inhibition of Nox2 by gp91 ds-tat in A_2AAR KO mice had no effect on adenosine-induced increase in coronary flow (Fig. 1b) but significantly attenuated the increased coronary flow in A_2BAR KO mice (Fig. 1c). All together, these findings indicate that adenosine-induced increase in coronary flow is partially mediated via activation of Nox2 and likely depends on A_2A but not A_2BARs.

Effects of ROS scavenging and A_2A/A_2BAR deletion on adenosine-induced increase in coronary flow

In accordance with previous studies from our laboratory [37], ROS scavenging with EUK134 markedly attenuated the adenosine-induced increase in coronary flow in WT mice (Fig. 2a). Similar to the effects of Nox2 inhibition, ROS scavenging with EUK134 had no effect on adenosine-induced increase in coronary flow in A_2AAR KO mice (Fig. 2b) but significantly attenuated the increased coronary flow in A_2BAR KO mice (Fig. 2c). All together, these findings indicate that ROS production is involved in adenosine-induced increase in coronary flow, which depends on A_2A but not A_2BAR activation.

Effects of Nox2 inhibition and A_2AAR/A_2BAR deletion on A_2AAR/A_2BAR agonist-mediated increase in coronary flow

To further investigate the role of A_2AAR in this process, the selective A_2AAR agonist CGS-21680 concentration response curves were acquired. CGS-21680 produced a concentration-dependent increase in coronary flow in WT, which was significantly blunted by Nox2 inhibition with gp91 ds-tat (Fig. 3a). A similar finding was also observed in A_2BAR KO mice in which CGS-21680-induced increase in coronary flow was significantly attenuated by the Nox2 inhibitor gp91 ds-tat (Fig. 3c). As expected, CGS-21680 failed to increase coronary flow in A_2AAR KO mice (Fig. 3b). To further exclude the involvement of A_2BAR, the selective A_2BAR agonist BAY 60-6583 concentration response curves were acquired in WT mice. BAY 60-6583 produced a concentration-dependent increase in coronary flow, which was not affected by the Nox2 inhibitor gp91 ds-tat (Fig. 4). Taken together, these findings indicate that adenosine-induced increase in coronary flow is partially mediated by Nox2-derived ROS, which requires A_2AARs.

Fig. 3.

Effects of Nox2 inhibition on adenosine A_2AAR agonist-induced increase in coronary flow. Shown are the effects of gp91 ds-tat (1 μM) on CGS-21680 (CGS, the A_2AAR selective agonist) concentration response curves (10⁻¹⁰ to 10⁻⁶ M) in isolated hearts of WT (a, n = 6), A_2AAR KO (b, n = 5), and A_2BAR KO mice (c, n = 5). Values are mean ± SEM. *Significant difference vs. corresponding control points using two-way ANOVA followed by post hoc analysis using Bonferroni’s test (P < 0.05)

Fig. 4.

Effects of Nox2 inhibition on adenosine A_2BAR agonist-induced increase in coronary flow. Shown are the effects of gp91 ds-tat (1 μM) on BAY 60-6583 (BAY, the A_2BAR selective agonist) concentration response curves (10⁻¹⁰ to 10⁻⁶ M) in isolated hearts of WT mice (n = 4). Values are mean ± SEM

Effects of Nox2 inhibition and A_2AAR/A_2BAR deletion on H₂O₂ production by adenosine in isolated coronary arteries

To assess the potential involvement of vasodilator ROS (H₂O₂) in adenosine-induced increase in coronary flow, DCFH-DA immunofluorescence was performed on intact isolated coronary arteries. As shown in Fig. 5, adenosine (10 μM) increased H₂O₂ generation in both WT and A_2BAR KO mice (Fig. 5a), as evidenced by increases in fluorescence intensity ratio in adenosine-treated tissues (1.36 ± 0.04 and 1.24 ± 0.02 times corresponding baselines, respectively, P < 0.05) as compared to corresponding baseline (Fig. 5b). H₂O₂ production was observed in both endothelial cells (ECs, arrow head) as well as smooth muscle cells (SMCs, arrow) (Fig. 1a). Of note, adenosine-increased H₂O₂ production was inhibited by the Nox2 inhibitor gp91 ds-tat in both WT and A_2BAR KO mice to baseline levels (Fig. 5a, b). More importantly, adenosine failed to increase H₂O₂ production from A_2AAR KO mice either in the presence or absence of Nox2 inhibition (Fig. 5a, b). Hydrogen peroxide (200 μM) added at the end of experiments further increased fluorescence (Fig. 5a), indicating that the dye did not reach the saturation point [40]. Taken together, these data further suggest that Nox2 is an important source for H₂O₂ production, which depends on the activation of A_2AARs by adenosine.

Fig. 5.

Effects of Nox2 inhibition on adenosine-induced H₂O₂ production in isolated coronary arteries. a Representative confocal fluorescence images showing changes in DCFH-DA (dichlorofluorescin) fluorescence intensity before (baseline) and after 5 min of adenosine (Ado) (10 μM) or 10 min of H₂O₂ (200 μM) stimulation in the presence and absence of gp91 ds-tat (1 μM) from WT, A_2AAR KO, and A_2BAR KO mice. Scale bar: 50 μm. b Adenosine-induced increase in H₂O₂ in isolated coronary arteries of WT (n = 8), A_2AAR KO (n = 5), and A_2BAR KO mice (n = 5). Values are mean ± SEM. *P < 0.05 vs. corresponding control (by paired t test)

Discussion

This study examined the hypothesis that A_2A and/or A_2BAR-mediated coronary vasodilation involves Nox2-derived ROS. With the focus on the adenosine signaling pathway, we measured changes in coronary flow in isolated hearts from WT, A_2AAR, and A_2B AR KO mice in response to Nox2 inhibition (gp91 ds-tat) or ROS scavenging (EUK134) during exogenous adenosine infusion. In addition, the downstream effector (H₂O₂) of ROS in this process was also assessed in isolated coronary arteries from WT, A_2AAR, and A_2B AR KO mice. The main findings of the present study are as follows: (1) either Nox2 inhibition with gp91 ds-tat or ROS scavenging with EUK134 attenuated the adenosine-induced increase in coronary flow in WT and A_2BAR KO mice but not in A_2AAR KO mice; (2) Nox2 inhibition blunted the A_2AAR agonist-increased coronary flow in WT and A_2BAR KO mice, whereas the A_2BAR agonist responses were unaltered by Nox2 inhibition in WT mice; (3) furthermore, adenosine-mediated increase in H₂O₂ formation in intact isolated coronary arteries of WT and A_2BAR KO mice was inhibited by Nox2 inhibition, whereas the H₂O₂ formation was not affected by adenosine either with or without Nox2 inhibition in A_2AAR KO mice. The implications of these findings are discussed below.

Adenosine-induced increase in coronary flow is mediated in part through Nox2-derived ROS

Adenosine is known to regulate coronary vascular tone in various species. We [20, 44] and others [10, 11, 45] previously demonstrated that adenosine also plays a physiological role in regulating coronary resting flow in human [11], swine [10], dogs [45], guinea pigs [46], and mice [20, 44], while some other groups failed to observe an endogenous role of adenosine [3, 47, 48], which leaves this scientific topic controversial. This discrepancy may be due to different animal models, differences in species, and/or different agonists and antagonists applied in these studies [19]. However, the application of exogenous adenosine to increase coronary flow for the investigation of adenosine receptor-mediated signaling pathways has been well established [37, 44]. Thus, in agreement with previous studies from our laboratory [37, 44, 49] and others [14], adenosine produced a large increase in coronary flow in isolated mouse hearts, supporting the overall effect of vasodilation by adenosine in the coronary circulation [14, 15, 37, 50, 51].

Several recent studies showed the interaction between adenosine and ROS via the activation of Nox, which has been reported to be the major source of ROS in the vasculature and involved in vascular tone regulation [27, 28]. Among Nox isoforms, Nox2 has been extensively studied in various vascular beds. Nox2 inhibition or deletion attenuated the adenosine-mediated vasoconstriction in aortas [33] and renal arterioles [34]. In contrast, Nox2 inhibition or ROS scavenging blunted the adenosine-induced vasodilation in cerebral arteries [35] and coronary circulation [37], both of which suggest that adenosine is capable of generating ROS in these vasculatures. Consistent with our previous findings in the coronary circulation but in contrast to those in other vascular beds [33, 34, 37], Nox2 inhibition or ROS scavenging significantly attenuated the adenosine-mediated increase in coronary flow, suggesting that adenosine-mediated coronary vasodilation is in part attributed to Nox2-derived ROS. This heterogeneity might be due to vasoconstrictor ROS vs. vasodilator ROS [29, 31] generated from different vascular beds (coronary vs. renal artery) [34, 37]. Furthermore, in the coronary vasculature, the varying effects of ROS may depend on which source ROS are generated from (NADPH vs. uncoupled NO synthase) [31, 52, 53] or the pathophysiological conditions [32, 52, 53], e.g., diabetes, and the right ventricular hypertrophy [54–56]. In accordance with our previous studies [37], the effect of ROS scavenging with EUK134 was comparable with the effect of Nox2 inhibition with gp91 ds-tat in WT isolated hearts (P = 0.24), suggesting that the majority of ROS generated by adenosine is likely from Nox2, which leads to increases in coronary flow. Although EUK134 is capable of scavenging both vasoconstrictor and vasodilator ROS [57], the observation that EUK134 significantly attenuated the adenosine-mediated increase in coronary flow in one of our previous studies [37] as well as in the present study suggest that vasodilator ROS are likely generated by adenosine.

In the healthy coronary vasculature, several recent studies addressed a crucial role of NADPH oxidase in coronary vasodilation in response to various agonists. Indeed, downregulation of Nox activity by cytosolic p47phox knockdown or inhibition of Nox2 attenuated the coronary vasodilation to vascular endothelial growth factor (VEGF) [52] and bradykinin [32], respectively. In addition, Nox2 overexpression enhanced the coronary vasodilation in response to both VEGF and acetylcholine [53], further confirming that the vasodilator ROS is derived from Nox in this process. More importantly, Larsen and colleagues proposed that Nox2 is a functionally relevant source of H₂O₂ that mediates agonist-induced coronary vasodilation [32]. In accordance with this concept, we found in the present study as well as one of our previous studies [40] that adenosine-increased H₂O₂ formation in intact isolated coronary arteries was attenuated by not only Nox2 inhibition but also catalase [40]. The production of H₂O₂ induced by adenosine might stem from the intermediate ROS product namely, superoxide anion [27, 28, 31], as we previously reported that adenosine is also capable of producing superoxide [37]. Thus, it is likely that H₂O₂ acts as a downstream effector of superoxide contributing to adenosine-induced increase in coronary flow. Regarding the source of H₂O₂, we believe that H₂O₂ produced from EC and SMC of coronary arteries may play a major role in adenosine-mediated coronary flow. This idea is based on our previous [40] as well as present findings that the increased H₂O₂ was observed in both EC (arrow head) and SMC (arrow) (Fig. 5a). In addition, functional A_2AAR is reported to be expressed on both EC and SMC [9], which is in agreement with our previous data showing positive A_2AAR staining in both cell types of isolated mouse coronary arteries [40]. Therefore, it is reasonable to speculate that adenosine may act through A_2AAR in both cell types producing H₂O₂, and activation of A_2AAR-H₂O₂ pathway may result in different downstream vasodilator mechanisms, as activation of A_2AAR in different cell types of coronary artery stimulates either cAMP-PKA or cGMP-PKG signaling [9, 12, 58, 59]. On the other hand, the cross talk of these vasodilator mechanisms exists. For instance, there is a possibility that endothelium-derived H₂O₂ may elicit further H₂O₂ production in SMC through activation of Nox [60], independent of A_2AAR activation on SMC to activate PKG pathway, as H₂O₂ has been shown to be able to regulate both PKA and PKG pathways [61]. Taken together, these findings support Nox2 as a relevant source for H₂O₂ in the coronary vasculature [32] and suggest that the involvement of Nox2-derived ROS in adenosine-mediated coronary vasodilation is through H₂O₂ production.

Involvement of Nox2-derived ROS in adenosine-induced increase in coronary flow requires the presence of A_2AARs

We [9, 16] and others [14, 15] have demonstrated that adenosine-mediated coronary vasodilation is predominantly attributed to A_2AAR as compared to vasodilator A_2BAR. In contrast, adenosine A₁ and A₃ARs counteract the adenosine-mediated coronary vasodilation [15, 17, 18]. Despite the findings that all four adenosine receptor subtypes are expressed and functional in the coronary vasculature [13], their roles in regulating coronary vascular tone through generation of ROS remain unclear. Given the fact that Nox has been suggested to be the major source of ROS in coronary vasculature [32, 52], it is not clear whether this leads to the generation of ROS by adenosine. We recently found that activation of A₃AR enhances Nox2-derived ROS contributing to vascular contraction in mouse aorta [33], suggesting a role of A₃AR in relation to ROS production. In contrast, in the coronary circulation, we have already excluded the involvement of A₁ and A₃ARs in increased flow that is coupled to the activation of Nox2 [37].

In the present study, we further investigated the role of adenosine A_2A and A_2BARs in this process. We found that adenosine A_2A but not A_2BARs are required for Nox2-mediated coronary vasodilation, as evidenced by the attenuation of adenosine-increased coronary flow by Nox2 inhibition (gp91 ds-tat) or ROS scavenging (EUK134) in both WT and A_2BAR KO mice but not in A_2AAR KO mice. Furthermore, the selective A_2AAR stimulation-induced increase in coronary flow was significantly attenuated by Nox2 inhibition in both WT and A_2BAR KO mice, whereas Nox2 inhibition failed to blunt the increased coronary flow upon the selective A_2BAR stimulation in WT mice. Taken together, these findings suggest that A_2AAR could be a potential mediator of adenosine-induced ROS generation contributing to increases in coronary flow. Moreover, in the isolated coronary arteries, adenosine-increased H₂O₂ formation was inhibited by Nox2 blockade in both WT and A_2BAR KO mice but not in A_2AAR KO mice, further suggesting that the involvement of Nox2-derived ROS in adenosine-induced increase in coronary flow requires the presence of A_2AAR. In contrast to coronary circulation, both A_2A and A_2BARs have been observed to generate ROS accounting for the vasodilation in cerebral circulation [35]. This discrepancy may be due to the different distribution, expression, and the downstream effectors of A_2A and A_2BARs between these vascular beds [12, 13].

A_2AAR-mediated coronary vasodilation through Nox2-derived ROS observed in the present study may bolster the functional influences of A_2AAR by integrating the vascular actions of the generated ROS to the previously known A_2AAR-mediated releases of other vasodilators (such as NO [20] and prostacyclin [62]). Furthermore, pertinent to this effect, it may also be possible that adenosine-generated ROS is involved in mediating reactive or functional hyperemia, despite the role of adenosine in functional hyperemia remaining controversial [19]. Indeed, the involvement of adenosine in either reactive or functional hyperemia has been shown to be mediated by endogenously released vasodilators, including NO, prostacyclin, and endothelium-derived hyperpolarizing factor [38, 40, 63, 64]. The influence and interaction of these vasoactive factors in generation of ROS by A_2AAR in the coronary circulation remain unclear. Although we previously reported that NO and H₂O₂ function in a parallel manner contributing to the reactive hyperemia [40], future studies focusing on this aspect will provide greater insights into the interaction of these vasoactive factors.

Conclusions and implications

The present study shows that adenosine induces generation of ROS (likely H₂O₂) through the activation of Nox2 resulting in coronary vasodilation. This whole process requires the presence of A_2AAR. These findings further support NADPH oxidase as a functionally relevant source of ROS for regulating coronary vascular tone. As both adenosine and ROS are generated and released under pathophysiological conditions [9, 65] and generated ROS mainly exerts a detrimental influence on the coronary vasculature [54–56], future studies addressing the role of A_2AAR in ROS-mediated coronary vascular tone regulation in pathophysiological conditions will bring us greater insights into mechanisms underlying ischemic heart disease.