Effect of exercise training on nitric oxide and superoxide/H₂O₂ signaling pathways in collateral-dependent porcine coronary arterioles

Abstract

Endothelial nitric oxide (NO) synthase (NOS) has been shown to contribute to enhanced vascular function after exercise training. Recent studies have revealed that relatively low concentrations of reactive oxygen species can contribute to endothelium-dependent vasodilation under physiological conditions. We tested the hypothesis that exercise training enhances endothelial function via endothelium-derived vasodilators, NO and superoxide/H₂O₂, in the underlying setting of chronic coronary artery occlusion. An ameroid constrictor was placed around the proximal left circumflex coronary artery to induce gradual occlusion in Yucatan miniature swine. At 8 wk postoperatively, pigs were randomly assigned to sedentary (pen-confined) or exercise-training (treadmill-run: 5 days/wk for 14 wk) regimens. Exercise training significantly enhanced concentration-dependent, bradykinin-mediated dilation in cannulated collateral-dependent arterioles (∼130 μm diameter) compared with sedentary pigs. NOS inhibition reversed training-enhanced dilation at low bradykinin concentrations in collateral-dependent arterioles, although increased dilation persisted at higher bradykinin concentrations. Total and phosphorylated (Ser¹¹⁷⁹) endothelial NOS protein levels were significantly increased in arterioles from collateral-dependent compared with the nonoccluded region, independent of exercise. The H₂O₂ scavenger polyethylene glycol-catalase abolished the training-enhanced bradykinin-mediated dilation in collateral-dependent arterioles; similar results were observed with the SOD inhibitor diethyldithiocarbamate. Fluorescence measures of bradykinin-stimulated H₂O₂ levels were significantly increased by exercise training, independent of occlusion. The NADPH inhibitor apocynin significantly attenuated bradykinin-mediated dilation in arterioles of exercise-trained, but not sedentary, pigs and was associated with significantly increased protein levels of the NADPH subunit p67phox. These data provide evidence that, in addition to NO, the superoxide/H₂O₂ signaling pathway significantly contributes to exercise training-enhanced endothelium-mediated dilation in collateral-dependent coronary arterioles.

an imbalance between vasodilator and vasoconstrictor influences of endothelium is a critical consequence of the pathogenic process of endothelial dysfunction, which is characterized by decreased vasodilation, a proinflammatory state, and smooth muscle proliferation (2, 4). An important feature of this altered vasoreactivity is impaired nitric oxide (NO) bioavailability and/or elevated oxidant stress (22, 30). Previous studies have shown that exercise training enhances endothelium-dependent relaxation and endothelial NO synthase (eNOS) mRNA expression and protein levels in coronary arteries and arterioles of control animals (29, 45). Exercise training has also been reported to enhance the contribution of NO and K⁺ channel activity to basal tone, as well as increase protein content of eNOS and phosphorylated eNOS (Ser¹¹⁷⁹) in small coronary arteries from pigs subjected to chronic coronary artery occlusion/stenosis (19). Exercise training-induced improvements in vascular function have been shown to contribute to enhanced myocardial perfusion and contractile function in vascular disease states (16, 17). However, despite remarkable evidence for therapeutic benefits of physical activity, the primary mechanisms by which regular exercise improves vascular function in the setting of coronary artery disease have not been fully elucidated.

While reactive oxygen species (ROS) have been implicated in the development of clinical pathophysiology of the cardiovascular system, additional studies have suggested that, at relatively low concentrations, ROS, such as superoxide and H₂O₂, can function as physiological signaling molecules (8, 9, 11, 27). Furthermore, recent studies have revealed that inhibition of NADPH oxidase reduced bradykinin-induced superoxide and H₂O₂ production (26) and attenuated endothelium-dependent dilation in human coronary arterioles (26) and rat skeletal muscle arterioles (46). Interestingly, a small number of recent reports indicate that the ROS H₂O₂ may function as a potential mediator of exercise training-induced adaptations in vascular reactivity (46). Thus we sought to assess the contribution of the NO and superoxide/H₂O₂ signaling pathways to exercise training-enhanced, endothelium-dependent dilation in collateral-dependent coronary arterioles and to begin to explore the underlying mechanisms that may contribute to adaptations in these signaling pathways.

MATERIALS AND METHODS

Experimental animals and surgical procedures.

All procedures were in accordance with “Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research and Training” and were approved by the Institutional Animal Care and Use Committee at Texas A & M University in accordance with the Association for the Accreditation of Laboratory Animal Care procedures. In addition, all protocols and methods conformed to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals [Department of Health and Human Services Publication No. (NIH) 85-23, Office of Science and Health Reports, Bethesda, MD]. Adult female Yucatan miniature swine (Sinclair Research Center, Auxvasse, MO) were surgically instrumented with an ameroid constrictor around the proximal left circumflex coronary (LCX) artery, as described in detail previously (21, 33). Animals were preanesthetized with glycopyrrolate (0.004 mg/kg im), midazolam (0.5 mg/kg im), and ketamine (20 mg/kg im), and surgical anesthesia was induced with 3% isoflurane. Animals were then intubated, and anesthesia was maintained with 2–3% isoflurane-balance O₂ during aseptic surgery. During the surgical procedure, animals received the following drugs as necessary: pancuronium (0.1 mg/kg) or vecuronium bromide (0.1 mg/kg) and lidocaine (1 mg/kg iv). Immediately following surgery, pigs received ketoprofen (Ketofen, 3.0 mg/kg iv). Prior to surgery and during surgical recovery, animals received buprenorphine hydrochloride (0.1 mg/kg iv) or butorphanol tartrate (0.5 mg/kg) every 3–6 h as needed for pain relief. The antibiotic Naxcel (cetiofur sodium, 4 mg/kg im) was administered 24 h before surgery, immediately prior to surgery, and for 2 days following surgery. For the most efficient use of our pigs, we utilize multiple tissue preparations (e.g., vascular, cardiac, skeletal muscle, blood, and cerebral) from each animal, so that numerous experiments can be conducted concurrently. Thus, while a large number of pigs were used for the studies described here, we make every effort to maximize the use of these animals.

Exercise training.

At 8 wk postoperatively, animals were randomly assigned to a sedentary (n = 48) or an exercise-training (n = 44) group. Sedentary pigs were confined to their pens, while exercise-trained animals underwent a progressive treadmill exercise-training program, 5 days/wk for 14 wk, as described previously (12, 15, 20). At termination, skeletal muscle citrate synthase activity and heart-to-body weight ratio were measured to evaluate effectiveness of the exercise-training regimen, as described previously (15, 21).

Preparation of coronary arterioles.

After the 14-wk exercise-training protocol or sedentary confinement, pigs were anesthetized using xylazine (Rompun, 2.25 mg/kg im), ketamine (35 mg/kg im), and pentothal sodium (30 mg/kg iv), and heparin was administered (1,000 U/kg iv). Animals were intubated and ventilated with room air, and a left lateral thoracotomy was performed in the fourth intercostal space. Hearts were removed, placed in Krebs bicarbonate buffer (0–4°C), and weighed. Visual inspection at the ameroid occluder during dissection of the LCX artery indicated 100% occlusion in all animals used in this study. With the aid of a dissection microscope, size-matched arterioles (∼130 μm) were isolated from subepicardial regions of the nonoccluded left anterior descending (LAD) artery and the collateral-dependent LCX artery in areas free from infarct.

Microvessel cannulation and experimental protocols.

Isolated arterioles were transferred to a Lucite vessel chamber containing physiological saline solution, cannulated, and pressurized for assessment of vascular reactivity, as described in detail previously (18). Arterioles underwent a 1-h equilibration period, during which the vessels established a stable level of basal tone. Arterioles were further preconstricted with endothelin-1 until a preconstriction level of ∼30–70% of maximal diameter was attained. For experiments in which pharmacological inhibitors were utilized, arterioles were preconstricted to the same level (∼30–70%) using the inhibitor plus endothelin-1, as previously described (18). Pharmacological inhibitors included the NOS inhibitor N^ω-nitro-l-arginine methyl ester (l-NAME, 300 μM), the H₂O₂ scavenger polyethylene glycol (PEG)-catalase (1,000 U/ml), an inhibitor of SOD [diethyldithiocarbamate (DETC), 1 mM], and an inhibitor of NADPH oxidase (apocynin, 100 μM). Additional vehicle control data were collected in the presence of the PEG compound at the same concentration (2.5 mg/ml) used in the PEG-catalase studies. Concentration-response curves were determined in response to cumulative concentrations of bradykinin or nitroprusside. All drugs were added directly to the tissue bath. Because our preliminary experiments suggested that arterioles subjected to repeated exposures to bradykinin exhibit tachyphylaxis, each arteriole underwent only a single concentration-response curve to bradykinin in the absence or presence of selective inhibitors.

Immunoblots.

Additional coronary arterioles (∼100–150 μm diameter, 10–15 mm total length) were dissected from the nonoccluded and collateral-dependent myocardial regions, quick-frozen in liquid N₂, and stored at −80°C for immunoblot analysis. Arterioles were homogenized in 40 μl of 2× lysis buffer [20 mM Tris·HCl, 50 mM NaCl, 0.1% Triton X-100, 1% each protease and phosphatase inhibitor cocktails (catalog nos. 539131 and 524625, respectively, Calbiochem), and 3 mM EGTA] by freeze-thaw cycles and vortexed approximately six to eight times. Protein concentration was determined by bicinchoninic acid protein assay kit (Pierce). Arteriole lysate (12 μg of total protein) was subjected to 12.5% SDS-PAGE, transferred to polyvinylidene difluoride membranes, and probed overnight with primary antibody, as described in detail previously (12). Primary antibody dilutions were as follows: eNOS (1:250), phosphorylated eNOS (Ser¹¹⁷⁹, 1:750), the NADPH oxidase subunits Nox1 (1.1 μg/ml), Nox2 (1:500), Nox4 (1:500), p47phox (1:1,250), and p67phox (1:250), and β-actin (1:5,000) at 4°C overnight. After they were washed, the membranes were incubated with the appropriate horseradish peroxidase-conjugated species-specific anti-IgG (diluted 1:50,000–1:100,000 depending on primary antibody) for 2 h at 25°C. Peroxidase activity was detected using SuperSignal West Dura substrate. Scanning densitometry was used to quantify signal density from luminograms. Normalization for potential loading differences was accomplished using the ratio of densitometry signals for proteins of interest to β-actin.

Fluorescent detection of H₂O₂.

In additional studies, bradykinin-mediated changes in H₂O₂ were detected in real time with the cell-permeable fluorescent indicator 5-(and 6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate (DCF). Pressurized coronary arterioles (∼100–150 μm) were loaded intraluminally by delivery of DCF (5 μM) into the arteriole through a cannulating micropipette. Flow was stopped, and the dye was incubated in the arteriole for 10 min. Flow was resumed for an additional 10 min to eliminate the fluorescent indicator from the arterioles and perfusion pipettes. The arteriole was allowed to equilibrate for an additional 10 min prior to start of experimentation. Albumin-free solutions were utilized to avoid potential interference with DCF fluorescence detection (47). After equilibration, a region of interest was selected, and changes in fluorescence intensity were observed (NIS Elements, Nikon) using an epifluorescence microscopy system. Arterioles were excited with a 175-W xenon arc lamp with a 475-nm interference filter (Lambda DG-4, Sutter Instruments). Fluorescence emission was captured at 515 nm every 30 s and reflected to an interline transfer, progressive-scan, cooled charge-coupled device video camera (CoolSNAP HQ, Photometrics) with a dichroic mirror. The microscope was equipped with an ×10 oil immersion objective, numerical aperture of 0.3. Basal fluorescence was obtained for 3 min; then arterioles were exposed to bradykinin (10⁻¹² or 10⁻¹⁰ M), and fluorescent images were acquired every 30 s for 5 min. An equal number of arterioles from each of the four vessel treatment groups received 10⁻¹² or 10⁻¹⁰ M bradykinin. Arterioles were exposed to exogenous H₂O₂, PEG-catalase, and nitroprusside in various combinations to verify specificity of DCF. Although DCF is oxidized by other peroxides in addition to H₂O₂, other investigators have demonstrated complete inhibition of agonist-stimulated fluorescence by addition of catalase (3, 24, 51).

Solutions and pharmacological agents.

l-NAME, apocynin, PEG-catalase, PEG, and DETC were purchased from Sigma. Endothelin-1 was purchased from Peninsula Laboratories. DCF was obtained from Invitrogen. Primary antibodies against the following proteins were utilized for these studies: eNOS (catalog no. 610297) and phosphorylated eNOS (Ser¹¹⁷⁹, catalog no. 612392) purchased from BD Biosciences; Nox1 (catalog no. abs5831) from Abcam; Nox4 (catalog no. NB110-58851) and β-actin (catalog no. NB400-501) from Novus; and Nox2 (catalog no. SC-20782), p47phox (catalog no. SC-14015), and p67phox (catalog no. SC-7662) from Santa Cruz Biotechnology. Physiological saline solution contained 145 mM NaCl, 4.7 mM KCl, 2.0 mM CaCl₂, 1.17 mM MgSO₄, 3.0 mM MOPS, 1.2 mM NaH₂PO₄, 5.0 mM glucose, 2.0 mM pyruvate, 0.02 mM EDTA, and 1% bovine serum albumin, pH 7.4.

Data presentation and analysis.

Citrate synthase, body weight, and heart-to-body weight ratio were compared between sedentary and exercise-trained pigs using Student's t-tests. Vasodilation responses to bradykinin are presented as the percent increase in internal diameter relative to the maximal possible dilation as follows: % = [(D_ss − D_pre)/(D_max − D_pre)]/100, where D_ss is the steady-state internal diameter in response to each bradykinin concentration, D_pre is the baseline diameter following preconstriction, and D_max is the maximal diameter of the arteriole determined by exposure to nitroprusside at the end of each experiment. Bradykinin- and nitroprusside-mediated dilation were evaluated by repeated-measures two-way ANOVA. Immunoblot and DCF fluorescence were analyzed using two-way ANOVA. Mean differences were ascertained using Bonferroni's multiple comparison tests when a main effect was significant by ANOVA. For all analyses, P ≤ 0.05 was considered significant. Values are means ± SE; n reflects the number of animals studied. When more than one arteriole from the nonoccluded or collateral-dependent myocardial region was used from one animal in identical protocols, the responses from these arterioles were averaged before data analyses were conducted.

RESULTS

Effectiveness of the exercise-training program.

Efficacy of the 14-wk progressive treadmill exercise-training regimen was demonstrated by significant increases in skeletal muscle oxidative enzyme capacity and an increased heart-to-body weight ratio in exercise-trained compared with sedentary animals. Citrate synthase activity increased significantly in muscle from exercise-trained (n = 44) compared with sedentary (n = 48) pigs as follows: deltoid muscle (44.3 ± 1.6 vs. 34.4 ± 1.1 μmol·min⁻¹·g⁻¹) and the medial (44.2 ± 1.6 vs. 35.1 ± 1.3 μmol·min⁻¹·g⁻¹), lateral (39.2 ± 1.0 vs. 29.9 ± 0.8 μmol·min⁻¹·g⁻¹), and long (31.4 ± 1.2 vs. 26.0 ± 0.8 μmol·min⁻¹·g⁻¹) heads of the triceps brachii muscle. Although body weight did not differ between sedentary and exercise-trained animals at the time of death (35.1 ± 0.7 and 33.7 ± 0.7 kg, respectively), heart-to-body weight ratio was significantly greater from exercise-trained (n = 44) than sedentary (n = 48) pigs (5.3 ± 0.1 vs. 4.6 ± 0.1 g/kg).

Characteristics of arterioles.

Maximal (passive) intraluminal diameters (D_max) of cannulated coronary arterioles, determined using nitroprusside (100 μM), were not significantly different between the four treatment groups (Table 1). The level of preconstriction was also similar across these groups (Table 1), and the concentration of endothelin-1 required to attain this level of preconstriction was not significantly different between arteriole groups (Table 1).

Table 1.

Dimensional characteristics of cannulated coronary arterioles

	n	Maximal Diameter, μm	Preconstriction, %	ET-1, nM
Nonoccluded
Sedentary	46	129 ± 4	60 ± 3	0.27 ± 0.02
Exercise-trained	40	134 ± 5	62 ± 3	0.26 ± 0.02
Collateral-dependent
Sedentary	44	135 ± 6	58 ± 3	0.30 ± 0.03
Exercise-trained	44	133 ± 4	63 ± 2	0.27 ± 0.03

Values are means ± SE; n, number of arterioles. %Preconstriction, percentage of maximal diameter to which arterioles were preconstricted; ET-1, concentration of endothelin-1 required to obtain %preconstriction. No significant differences exist.

Effects of exercise training and occlusion on bradykinin-mediated dilation.

We examined the effect of exercise training and chronic coronary occlusion on bradykinin-mediated dilation in coronary arterioles isolated from nonoccluded and collateral-dependent myocardial regions. Our data revealed that exercise training did not alter bradykinin-mediated dilation in arterioles from the nonoccluded region (Fig. 1A). In contrast, bradykinin-mediated dilation was significantly enhanced in collateral-dependent arterioles from exercise-trained compared with sedentary animals (Fig. 1B). Bradykinin-mediated dilation was not statistically different in collateral-dependent compared with nonoccluded coronary arterioles of sedentary and exercise-trained pigs. Evaluation of IC₅₀ revealed that sensitivity to bradykinin was not altered by exercise training in the nonoccluded region (Fig. 1A, inset). In contrast, exercise training significantly increased the sensitivity of arterioles from the collateral-dependent region to bradykinin (Fig. 1B, inset).

Fig. 1.

Effect of chronic occlusion and exercise training on bradykinin-mediated vasodilation in porcine coronary arterioles. A: bradykinin-mediated dilation was not significantly altered by exercise training in arterioles of the nonoccluded region. B: exercise training significantly enhanced bradykinin-mediated dilation in the collateral-dependent region. IC₅₀ values (insets) revealed that sensitivity to bradykinin was not altered by exercise training in the nonoccluded region but was significantly increased in collateral-dependent arterioles of exercise-trained (EX) compared with sedentary (SED) pigs. Values are means ± SE of number of animals in parentheses. *P ≤ 0.05.

Role of NO in bradykinin-mediated vasodilation.

We evaluated the contribution of NO to exercise-training-enhanced, bradykinin-mediated vasodilation using the NOS inhibitor l-NAME. NOS inhibition significantly attenuated bradykinin-mediated dilation in nonoccluded and collateral-dependent arterioles of sedentary and exercise-trained animals. NOS inhibition similarly reduced bradykinin-mediated dilation in arterioles from the nonoccluded region of sedentary and exercise-trained animals (Fig. 2A). In collateral-dependent arterioles, l-NAME corrected the exercise-training-enhanced, endothelium-mediated dilation at lower bradykinin concentrations. However, l-NAME did not reverse the training-enhanced vasodilation at higher bradykinin concentrations (Fig. 2B), and significant vasodilation persisted that was insensitive to NO synthase (NOS) inhibition in collateral-dependent coronary arterioles from trained animals. In the presence of l-NAME, IC₅₀ values were similar in arterioles from the nonoccluded region of sedentary and exercise-trained pigs (Fig. 2A, inset). In contrast, enhanced sensitivity to bradykinin in arterioles from the collateral-dependent region of exercise-trained pigs persisted in the presence of l-NAME (Fig. 2B, inset).

Fig. 2.

Effect of the nitric oxide synthase ·(NOS) inhibitor nitro-l-arginine methyl ester (l-NAME, 300 μM) on bradykinin-mediated dilation of arterioles. NOS inhibition significantly attenuated dilation in nonoccluded (A) and collateral-dependent (B) arterioles of sedentary and exercise-trained pigs. In collateral-dependent arterioles, NOS inhibition corrected the exercise-training-enhanced, endothelium-mediated dilation at lower bradykinin concentrations; training-enhanced vasodilation persisted at higher bradykinin concentrations. IC₅₀ values revealed that, in the presence of l-NAME, sensitivity to bradykinin was similar in arterioles from the nonoccluded region of sedentary and exercise-trained pigs (A, inset). Enhanced sensitivity to bradykinin in arterioles from the collateral-dependent region of exercise-trained pigs persisted in the presence of l-NAME (B, inset). Values are means ± SE of number of animals in parentheses. *P ≤ 0.05.

Smooth muscle responsiveness to nitroprusside.

To examine the possibility that arteriolar smooth muscle responsiveness to NO was altered by occlusion or exercise training, we determined concentration-response curves of coronary arterioles to the endothelium-independent NO donor nitroprusside (10⁻¹⁰–10⁻⁴ M). Concentration-dependent dilation responses to nitroprusside were not significantly different in coronary arterioles isolated from the collateral-dependent and nonoccluded myocardial regions of sedentary or exercise-trained animals (data not shown).

Total and phosphorylated eNOS protein levels.

On the basis of our finding that NO contributed to exercise-training-enhanced dilation at lower bradykinin concentrations, we examined the effect of occlusion and exercise training on protein levels of total and phosphorylated (Ser¹¹⁷⁹) eNOS. Figure 3 provides representative immunoblots for total and phosphorylated eNOS, as well as β-actin as loading control, using arterioles from the nonoccluded and collateral-dependent regions of sedentary and exercise-trained pigs. Total (Fig. 3A) and phosphorylated (Fig. 3B) eNOS protein levels were significantly increased in coronary arterioles from the collateral-dependent region compared with the nonoccluded region. Exercise training did not significantly alter total or phosphorylated eNOS protein levels.

Fig. 3.

Effect of chronic occlusion and exercise training on total endothelial NOS (eNOS) and phosphorylated eNOS (p-eNOS, Ser¹¹⁷⁹) protein levels. Total (A) and phosphorylated (B) eNOS protein levels were significantly increased in coronary arterioles from the collateral-dependent region compared with the nonoccluded region, independent of exercise-training status. Protein was quantified by densitometry analysis, normalized to β-actin, and expressed relative to the density of nonoccluded arterioles of sedentary pigs. Values are means ± SE; n, number of animals. *P ≤ 0.05.

Role of H₂O₂ in bradykinin-mediated vasodilation.

Because NOS inhibition did not fully correct the exercise-training-enhanced bradykinin-mediated dilation, we determined the contribution of the candidate endothelium-derived hyperpolarizing factor H₂O₂ to bradykinin-induced dilation using the cell-permeable H₂O₂ scavenger PEG-catalase. PEG-catalase significantly attenuated bradykinin-mediated dilation in all arteriole treatment groups (Fig. 4). Scavenging of H₂O₂ similarly reduced bradykinin-mediated dilation in arterioles from the nonoccluded region of sedentary and exercise-trained pigs (Fig. 4A). Importantly, PEG-catalase abolished the difference in bradykinin-mediated dilation between arterioles of the collateral-dependent myocardial region in exercise-trained and sedentary animals (Fig. 4B). In the presence of PEG-catalase, IC₅₀ values were similar in arterioles from the nonoccluded region of sedentary and exercise-trained pigs (Fig. 4A, inset). Importantly, treatment with PEG-catalase reversed the enhanced sensitivity to bradykinin observed under control conditions in arterioles from the collateral-dependent region of exercise-trained pigs (Fig. 4B, inset). Vehicle-control experiments revealed that bradykinin-mediated dilation was not significantly altered by inclusion of an equivalent concentration of the PEG compound in the tissue bath (n = 5).

Fig. 4.

Effect of the H₂O₂ scavenger polyethylene glycol-catalase (PEG-CAT) on bradykinin-mediated dilation of arterioles. PEG-CAT significantly attenuated dilation in nonoccluded (A) and collateral-dependent (B) arterioles of sedentary and exercise-trained pigs. Exercise-training-enhanced dilation in collateral-dependent arterioles was reverse in the presence of PEG-CAT. In the presence of PEG-CAT, IC₅₀ values were similar in arterioles from the nonoccluded region of sedentary and exercise-trained pigs (A, inset). Treatment with PEG-CAT reversed the enhanced sensitivity to bradykinin that was observed under control conditions in arterioles from the collateral-dependent region of exercise-trained pigs (B, inset). Values are means ± SE of number of animals in parentheses. *P ≤ 0.05.

Effects of SOD inhibition on bradykinin-mediated vasodilation.

We also examined the effect of the SOD inhibitor DETC on bradykinin-mediated dilation in nonoccluded and collateral-dependent arterioles from sedentary and exercise-trained pigs. DETC specifically inhibits the copper-containing SOD isoforms Cu,Zn-SOD and extracellular SOD (6), reducing the conversion of superoxide to H₂O₂. Treatment with DETC significantly diminished bradykinin-mediated dilation in all arteriole treatment groups (Fig. 5). SOD inhibition similarly attenuated bradykinin-mediated dilation in arterioles from the nonoccluded region of sedentary and exercise-trained pigs (Fig. 5A). Notably, DETC treatment had a more pronounced effect in collateral-dependent arterioles of exercise-trained than sedentary pigs (Fig. 5B), similar to that observed in the presence of PEG-catalase in Fig. 5B. Despite the more marked attenuation of bradykinin-mediated dilation with DETC after exercise training, a significant interaction effect between collateral-dependent arterioles of sedentary and exercise-trained pigs persisted, although post hoc analysis revealed that responses differed only at one concentration (3 nM) of bradykinin (Fig. 5B). In the presence of DETC, IC₅₀ values were similar in arterioles from the nonoccluded region of sedentary and exercise-trained pigs (Fig. 5A, inset). Similar to that observed with PEG-catalase, treatment with DETC reversed the enhanced sensitivity to bradykinin that was observed under control conditions in arterioles from the collateral-dependent region of exercise-trained pigs (Fig. 5B, inset).

Fig. 5.

Effect of the SOD inhibitor diethyldithiocarbamate (DETC) on bradykinin-mediated vasodilation in coronary arterioles. DETC significantly attenuated dilation in nonoccluded (A) and collateral-dependent (B) arterioles of sedentary and exercise-trained pigs. Exercise-training-enhanced dilation in collateral-dependent arterioles was nearly reversed in the presence of DETC. In the presence of DETC, IC₅₀ values were similar in arterioles from the nonoccluded region of sedentary and exercise-trained pigs (A, inset). Treatment with DETC reversed the enhanced sensitivity to bradykinin that was observed under control conditions in arterioles from the collateral-dependent region of exercise-trained pigs (B, inset). Values are means ± SE of number of animals in parentheses. *P ≤ 0.05.

Bradykinin-induced H₂O₂ production.

We utilized the fluorescent indicator DCF as a measure of bradykinin-stimulated H₂O₂ in cannulated arterioles (Fig. 6). Our preliminary studies with DCF demonstrated that treatment with PEG-catalase (1,000 U/ml) abolished bradykinin-stimulated DCF fluorescence and exogenous application of H₂O₂ (3 mM) markedly stimulated DCF fluorescence in arterioles, demonstrating specificity of the fluorescent dye for H₂O₂ (Fig. 6A), as confirmed previously by other investigators (3, 24, 51). Arterioles designated sedentary nonoccluded in Fig. 6 were obtained from the LAD-supplied region of occluded (n = 2), sham-operated (n = 1), or control, nonoccluded (n = 1) pigs. Mean responses from arterioles of the nonoccluded region (LAD) of occluded pigs were not different from the mean responses of the same region (LAD) of nonoccluded pigs. Arterioles from other treatment groups were collected from designated regions of occluded pigs. The peak cytosolic DCF signal in response to bradykinin stimulation was calculated by subtraction of basal fluorescence from the peak fluorescence response to bradykinin. Bradykinin stimulation significantly increased DCF fluorescence in coronary arterioles of exercise-trained animals to a greater extent than that observed in sedentary pigs, with no effect of occlusion (Fig. 6B).

Fig. 6.

Fluorescence measures of H₂O₂ in cannulated porcine coronary arterioles by 5-(and 6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate (DCF). A: representative fluorescence images of arterioles at baseline, in response to bradykinin in the absence and presence of PEG-CAT (1,000 U/ml), in response to exogenously applied H₂O₂ (3 mM) in the absence and presence of PEG-CAT, and in response to the NO donor nitroprusside (100 μM) in the presence of PEG-CAT. B: DCF fluorescence was significantly increased in nonoccluded and collateral-dependent coronary arterioles of exercise-trained compared with sedentary pigs. Values are means ± SE; n, number of animals. *P ≤ 0.05.

Effect of apocynin on bradykinin-mediated vasodilation.

In the vasculature, NADPH oxidase is considered a primary source of superoxide anion, the precursor to H₂O₂ production (13, 14, 39). We examined the effect of the NADPH oxidase inhibitor apocynin on bradykinin-mediated dilation. Pretreatment with apocynin did not significantly alter bradykinin-induced dilation in arterioles from the nonoccluded (Fig. 7A) or collateral-dependent (Fig. 7B) region of sedentary pigs. In contrast to these findings, apocynin significantly attenuated bradykinin-mediated dilation in arterioles from nonoccluded (Fig. 7C) and collateral-dependent (Fig. 7D) myocardial regions of exercise-trained animals. Treatment with apocynin did not significantly alter the sensitivity to bradykinin in arterioles from the nonoccluded or collateral-dependent regions of sedentary pigs (Fig. 7, A and B, insets). On the other hand, the concentration necessary to elicit 50% dilation in response to bradykinin was significantly increased in arterioles from the nonoccluded and collateral-dependent regions of exercise-trained pigs (Fig. 7, C and D, insets).

Fig. 7.

Effect of the NADPH oxidase inhibitor apocynin on bradykinin-mediated dilation of arterioles. Apocynin did not significantly alter bradykinin-mediated dilation in nonoccluded (A) or collateral-dependent (B) arterioles of sedentary pigs. Apocynin significantly attenuated dilation in nonoccluded (C) and collateral-dependent (D) arterioles of exercise-trained pigs. In the presence of apocynin, IC₅₀ values revealed that sensitivity to bradykinin was not altered in arterioles from the nonoccluded and collateral-dependent regions of sedentary pigs (A and B, insets). Apocynin significantly reduced sensitivity to bradykinin in arterioles from the nonoccluded and collateral-dependent regions of exercise-trained pigs (C and D, insets). Values are means ± SE of number of animals in parentheses. *P ≤ 0.05.

NADPH oxidase subunit protein levels.

We examined protein levels of the NADPH oxidase subunits p47phox, p67phox, and Nox4. In preliminary studies (Fig. 8A), we characterized the protein levels of Nox1, Nox2, and Nox4, specific homologs that are reported to be most abundant in vascular cells (26, 28). Our findings revealed that while we were able to detect Nox1 and Nox2 in pig tissues (e.g., brain, lung, and/or myocardium), the protein levels of these two homologs in pig arterioles were of very low abundance. On the other hand, Nox4 protein was abundant in pig arterioles and, thus, was further studied across arteriole treatment groups. Figure 8, B–D, shows representative immunoblots of Nox4, p47phox, and p67phox, as well as β-actin as loading control, using arterioles from the nonoccluded and collateral-dependent regions of sedentary and exercise-trained pigs. Immunoblot analyses for Nox4 (Fig. 8B) and p47phox (Fig. 8C) revealed no significant effect of occlusion or exercise training on arteriolar protein levels. In contrast, exercise training produced a small, but statistically significant, increase in p67phox protein levels (Fig. 8D), independent of occlusion.

Fig. 8.

Effect of chronic occlusion and exercise training on protein levels of NADPH oxidase subunits. Preliminary evaluation of the Nox family of isoforms revealed that Nox1 and Nox2 were poorly expressed in porcine coronary arterioles (PA), whereas Nox4 was abundant (A). Subsequent evaluation across groups revealed that neither Nox4 (B) nor p47phox (C) was significantly altered by occlusion or exercise training. In contrast, p67phox protein (D) was significantly increased by exercise training, independent of occlusion. Protein was quantified as described in Fig. 3 legend. PL, pig lung; PH, pig heart; PB, pig brain. Values are means ± SE; n, number of animals. *P ≤ 0.05.

DISCUSSION

In this study, we document the novel finding that exercise training increases the contribution of the superoxide/H₂O₂ signaling pathway to endothelium-dependent vasodilation in collateral-dependent coronary arterioles. Specifically, our data indicate that H₂O₂ significantly contributes to bradykinin-mediated dilation in arterioles from nonoccluded and collateral-dependent regions of sedentary and exercise-trained pigs. Importantly, we provide two data sets that suggest that the elimination of H₂O₂, via PEG-catalase or DETC, nearly abolished the exercise-training-enhanced dilation in collateral-dependent arterioles. Furthermore, H₂O₂ levels were significantly increased in arterioles from exercise-trained pigs, independent of occlusion. Our additional unique data reveal that the NADPH oxidase inhibitor apocynin significantly attenuated bradykinin-mediated dilation in arterioles from nonoccluded and collateral-dependent myocardial regions of exercise-trained, but not sedentary, pigs. Consistent with these findings, the NADPH oxidase subunit p67phox was increased by exercise training, independent of occlusion.

Previous studies have reported that H₂O₂ signaling contributes to endothelium-dependent bradykinin-mediated dilation in coronary arterioles of heart disease patients (26). However, our studies are the first to demonstrate that exercise training produces adaptations in H₂O₂ signaling that correct impaired endothelium-dependent dilation in coronary arterioles from ischemic hearts. In support of our data, studies in rat skeletal muscle arterioles revealed that exercise training increased NO and superoxide anion production in response to increased intraluminal flow and that enhanced superoxide, or its dismutation to H₂O₂, contributed to the exercise-training-induced improvements in flow-mediated dilation (46). Previous data reported by our group have also revealed that exercise training corrects impaired adenosine-mediated dilation in collateral-dependent arterioles, which also appeared to be mediated by NO and H₂O₂ (48). Our complementary evaluation of bradykinin-stimulated arteriolar H₂O₂ production revealed that exercise training increased H₂O₂ levels in arterioles from the nonoccluded and collateral-dependent region. Taken together with our vascular reactivity studies, our data indicate that enhanced H₂O₂ levels contribute significantly to adaptations in the coronary microcirculation following exercise training. Furthermore, adaptations in H₂O₂ signaling with exercise training appear to compensate for loss of endothelium-dependent vasodilation in the collateral-dependent region.

It is curious that while bradykinin-stimulated H₂O₂ levels were elevated in nonoccluded and collateral-dependent arterioles of exercise-trained pigs, dilation to bradykinin was augmented only in collateral-dependent arterioles after exercise training. Furthermore, no difference in the effect of PEG-catalase was noted in nonoccluded arterioles of sedentary and exercise-trained pigs. Although these findings are difficult to reconcile, we postulate that potential cross talk between various signaling pathways of vasodilation may alter the contribution of individual pathways, limiting overall dilation. Other investigators have cited reports of the interaction and interdependence of multiple pathways of endothelium-derived vasodilators (53). Furthermore, we reported previously that multiple mediators appear to be enhanced by exercise training in the underlying setting of ischemic heart disease and proposed that this may reflect redundant pathways that cooperate to preserve arterial relaxation/dilation (1). Alternatively, combined chronic occlusion and exercise training may stimulate increased expression and/or activity of signaling pathway components downstream of H₂O₂ that are not altered by exercise training alone, enhancing vasodilation in collateral-dependent arterioles of exercise-trained compared with sedentary pigs.

It is also interesting that, in the presence of DETC, dilation at high bradykinin concentrations was more pronounced in nonoccluded and collateral-dependent arterioles of exercise-trained than sedentary pigs. This effect did not attain statistical significance in arterioles from the nonoccluded region and attained statistical significance at only one concentration of bradykinin (3 nM) in arterioles from the collateral-dependent region. Because DETC specifically inhibits the copper-containing SOD isoforms Cu,Zn-SOD and extracellular SOD (6), activity of Mn-SOD may have contributed to the persistent bradykinin-mediated dilation at high concentrations in arterioles from exercise-trained pigs. In the presence of DETC, another vasodilatory mediator or pathway (H₂O₂-independent) also may have partially compensated to promote dilation at high bradykinin concentrations; however, one would speculate that the same mediator would have contributed in the presence of PEG-catalase also.

Our findings also support a role for an increased contribution of NOS to exercise-training-enhanced dilation at low bradykinin concentrations, despite the finding that enhanced dilation persisted in the presence of NOS inhibition at higher bradykinin concentrations. Our additional data demonstrate that eNOS and phosphorylated eNOS protein levels were significantly enhanced in coronary arterioles from the collateral-dependent region compared with the nonoccluded region, independent of exercise status. Thus the exercise-training-enhanced dilation at low bradykinin concentrations in the collateral-dependent region likely was not attributable to changes in eNOS or phosphorylated eNOS protein levels. These data suggest that another regulatory mechanism of eNOS enzyme activity likely played a role in the enhanced eNOS contribution at low bradykinin concentrations. Our eNOS protein data in coronary arterioles are consistent with our previous report that eNOS and phosphorylated eNOS protein levels were enhanced in epicardial artery segments by chronic occlusion, but not by exercise training (52). Furthermore, these findings are in contrast to our previous report that protein content of eNOS and phosphorylated eNOS (Ser¹¹⁷⁹) was elevated in small arteries from exercise-trained animals, with the greatest effect in collateral-dependent vasculature (19, 29). Interestingly, it has been reported previously that exercise training produces variable and heterogenous adaptations in total eNOS protein content depending on vessel size (29).

In the present study, apocynin significantly attenuated bradykinin-mediated vasodilation in arterioles from nonoccluded and collateral-dependent regions in exercise-trained animals. These findings suggest that NADPH oxidase provides a significant source of superoxide that either directly, or indirectly through its dismutation to H₂O₂, contributes to bradykinin-mediated dilation in coronary arterioles from exercise-trained pigs. These data are in contrast to previously reported rat skeletal muscle arteriole data, in which apocynin reduced flow-induced vasodilation in soleus muscle arterioles of young and old sedentary rats but not in those of exercise-trained animals (46). However, these investigators did demonstrate that exercise training augmented the relative increase in superoxide production by intraluminal flow, suggesting an exercise-training-induced increase in superoxide levels by a source other than NADPH oxidase (46).

ROS are produced by a variety of intracellular mechanisms, including mitochondrial sources, xanthine oxidase, NADPH oxidase, cytochrome P-450, cyclooxygenases, lipoxygenases, and uncoupled eNOS (2). NADPH oxidase is generally regarded as a predominant source of superoxide in the cardiovascular system (37, 40). In our study, exercise training significantly elevated the arteriolar protein level of the NADPH oxidase subunit p67phox, independent of occlusion. Previous studies have documented that exercise training did not alter p67phox protein levels in coronary arterioles from control animals (43), while p67phox was reduced and p47phox protein levels were not altered by exercise training in aortic endothelial cells isolated from the same animal model (42). However, exercise training has been reported to increase protein content of p47phox and p67phox, but not gp91phox, in rat neutrophils (32). Differences in animal models and cell and tissue types may account for the discrepancy in findings across the literature regarding the effects of exercise training on NADPH oxidase subunit expression levels. Furthermore, NADPH oxidase is composed of several subunits, so coordination of these subunits may contribute to altered enzyme activity, independent of protein content (42). In our study, exercise training resulted in increased protein levels of p67phox and an increase in the contribution of NADPH oxidase to bradykinin-mediated dilation, both independent of occlusion. Taken together, these data suggest that enhanced p67phox protein levels after exercise training may have contributed to increased NADPH oxidase activity.

In present study, we also found that Nox4 protein levels appeared to be more abundant than Nox1 and Nox2 in porcine coronary arterioles. Consistent with these data, previous studies also have reported that Nox4 is expressed at markedly higher levels than Nox2 in vascular endothelial cells (50), suggesting that Nox4 may play a more pronounced role in vascular endothelial cells. Nox4 is thought to act as an oxygen sensor and contribute to increased ROS generation and cell proliferation (23, 31). In our pig model, exercise training bouts likely cause periods of ischemia/hypoxia in the collateral-dependent region (41); thus we postulate that Nox4-induced superoxide generation may be altered in the collateral-dependent region, despite no change in protein levels. Additionally, it has been suggested that Nox4 may generate more H₂O₂ than superoxide and, thus, contribute directly to H₂O₂-mediated vasodilation (26, 34).

Study limitations.

Although DCF is often used experimentally to detect intracellular H₂O₂, literature suggests that it also reacts with other ROS and that confirmation of H₂O₂ detection by DCF is generally accomplished by treatment with PEG-catalase (7). On the basis of our findings that PEG-catalase eliminated the bradykinin-stimulated DCF signal in our cannulated microvessels, we postulated that the DCF signal was representative of H₂O₂. On the other hand, use of PEG-catalase may influence the equilibrium flux across the handling processes of other ROS through the creation of a H₂O₂ sink. For example, scavenging of H₂O₂ may increase the forward flux of superoxide through SOD, altering the concentration of other ROS that may have been detected by DCF. Similarly, use of PEG-catalase in our studies of bradykinin-mediated dilation may have also altered other ROS upstream of H₂O₂. Although use of the SOD inhibitor DETC provided additional evidence suggesting that the exercise-training-enhanced bradykinin-mediated dilation was attributable to H₂O₂, the specificity of DETC is equivocal (25).

Furthermore, inclusion of experiments with combined l-NAME and PEG-catalase treatment may have provided additional insight into a potential source of ROS in these studies, as well as the interaction of these two mediators. Drouin et al. (8) demonstrated that N-nitro-l-arginine (NOS inhibitor) + catalase had no additive inhibitory effects on acetylcholine-mediated dilation compared with catalase alone in mouse cerebral arterioles. They ultimately concluded that uncoupled eNOS was the source for superoxide/H₂O₂ generation in their model (8). On the basis of our data, we anticipate that pretreatment of arterioles with l-NAME + PEG-catalase would have an additive effect in arterioles from exercise-trained pigs. These findings would suggest that the PEG-catalase-sensitive component is, at least in part, independent of NOS. We base our prediction on our finding that bradykinin-mediated dilation had a significant component that was apocynin-sensitive (Fig. 7), suggesting that an NADPH oxidase-derived ROS contributed to bradykinin-induced dilation in arterioles from exercise-trained pigs. Because arterioles from sedentary pigs did not demonstrate sensitivity to apocynin, there is potential that pretreatment with l-NAME + PEG-catalase would not have an additive effect. Future studies will more thoroughly explore potential sources of ROS, as well as downstream signaling pathway components and their potential interaction, that may contribute to exercise-training-enhanced, endothelium-dependent vasodilation.

Clinical implications and conclusions.

Despite extensive evidence for therapeutic benefits of physical activity, the cellular and molecular mechanisms by which regular exercise improves vascular function in the setting of coronary artery disease are only beginning to be explored. Understanding adaptive responses of the coronary microcirculation to disease and interventional therapies is important in coronary artery disease, because these small arteries and arterioles determine coronary vascular resistance and, therefore, coronary blood flow (33). Historically, the production of ROS has been implicated in the development of cardiovascular pathophysiology (5, 38). Interestingly, accumulating evidence indicates that relatively low concentrations of ROS can function as important signaling molecules under physiological conditions and may be involved in normal regulation of vascular structure and function (10, 35, 36, 44, 49). Our current studies provide new evidence that exercise training appears to enhance the contribution of ROS to improvements in endothelium-mediated dilation in collateral-dependent coronary arterioles distal to chronic coronary artery occlusion/stenosis. These exciting findings suggest the novel concept that the superoxide/H₂O₂ signaling pathway, as well as the NO signaling pathway, may play critical roles in the adaptive responses of the coronary microcirculation in the diseased heart in response to exercise training.

GRANTS

These studies were supported by National Institutes of Health Grants R01-HL-064931 (C. L. Heaps and J. L. Parker) and R01-GM-046441 (Principal Investigator, Paul A. Lindahl).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

W.X. and C.L.H. are responsible for conception and design of the research; W.X. performed the experiments; W.X. and C.L.H. analyzed the data; W.X., J.L.P., and C.L.H. interpreted the results of the experiments; W.X. and C.L.H. prepared the figures; W.X. drafted the manuscript; W.X., J.L.P., and C.L.H. edited and revised the manuscript; W.X., J.L.P., and C.L.H. approved the final version of the manuscript.