|Home | About | Journals | Submit | Contact Us | Français|
Bradykinin (BK) is a key mediator regulating coronary blood flow. It is degraded by angiotensin-converting enzyme (ACE), but what is unknown is whether enhanced tissue ACE activity interferes with BK-induced coronary vasodilation in obesity.
Coronary arterioles (~100 μm) were isolated from rats on a normal or high-fat diet (HFD) and from lean or obese patients undergoing heart surgery (n=74). We found that BK-induced dilation was diminished in the coronary arterioles of HFD rats, when compared with controls. When administered in vitro, the ACE inhibitor, captopril, restored the coronary dilation response to BK in HFD rats, but did not affect control responses. Abundant ACE expression was detected in coronary endothelium, which was associated with increased ACE activity in HFD arterioles, as measured by increased response to the ACE substrate, angiotensin I. Moreover, we found that in the coronary arterioles of obese patients, BK-induced dilation was augmented by in vitro captopril administration. Correspondingly, ACE activity was increased in the coronary arterioles of obese patients when compared with the non-obese. Logistic regression analysis revealed that obese patients taking ACE inhibitors prior to surgery exhibited an enhanced dilation response to BK.
We demonstrated augmented tissue ACE activity in the coronary arterioles of obese subjects, which leads to reduced coronary dilation response to BK. We provide a rationale for ACE inhibitor therapy in obese patients to improve dilatation of coronary microvessels.
Bradykinin (BK) is continuously produced in the heart by the tissue kinin-kallikrein system,1 and is considered to be a key endogenous regulator of coronary blood flow. In this context, Groves et al demonstrated that intracoronary infusion of the BK B2 receptor antagonist, HOE 140, reduced the diameter of epicardial coronary arteries and decreased coronary blood flow in patients without significant coronary occlusion.2 In patients with no signs of coronary artery disease, intracoronary infusion of BK increased coronary artery diameter and elevated coronary blood flow.3 It is known that the rate of BK production increases in response to ischemic insults,4 aiming to maintain coronary dilation and tissue perfusion.5 However, this mechanism often fails in the diseased heart because of the unresponsiveness of coronary resistance arteries to exogenous or endogenous BK.6 Obesity and metabolic syndrome have been shown to have a detrimental effect on the coronary microcirculation of patients undergoing percutaneous coronary intervention.7 Previously, we have shown that BK-induced coronary arteriolar dilation is reduced in obese-normotensive patients, when compared with lean-normotensive individuals.8 The underlying mechanisms responsible for impaired BK-dependent regulation of coronary vascular resistance in obesity are not fully understood.
In the circulation, BK is readily cleaved and inactivated by angiotensin-converting enzyme (ACE). The presence of tissue ACE has been described in cardiac myocytes and coronary vessels.9 In the myocardium, ACE plays a crucial role in various homeostatic pathways, including cell growth, extracellular matrix formation and apoptosis.10 In pathological conditions, such as atherosclerosis, hypertension and obesity, upregulation of tissue ACE is known to contribute to morphological changes in the heart by initiating cardiac and vascular hypertrophy.11 These adverse effects are mainly mediated by an increased ACE-dependent, localized production of angiotensin II (AngII).11 Less is known about the functional changes that may occur in coronary resistance arteries as a consequence of tissue ACE activation. The important ACE end-product, AngII, normally dilates coronary resistance arteries through activation of type 2 AngII receptors (AT2Rs).12,13 However, recent studies revealed that AT2R activation may also lead to constriction of resistance arteries in various disease states.14 In this context, Zhang et al have shown that AngII induces constriction of coronary arteries obtained from dogs fed a high-fat diet (HFD), an experimental model of obesity and metabolic dysfunction.15 These authors concluded that activation of tissue ACE could lead to enhanced production of AngII, thus promoting coronary vasoconstriction in obesity. This mechanism remains unconfirmed in human obesity.
In addition, it is possible that the upregulated ACE in coronary microvessels interferes with the dilator effects of endogenously produced BK, a pathological mechanism that may also limit myocardial perfusion. In support of this scenario, Kuga et al demonstrated that the diameter of epicardial coronary arteries is increased after intracoronary infusion of the ACE inhibitor, enalaprilat, in patients without significant coronary stenosis.3 Previous studies have shown that systemic administration of an ACE inhibitor improves vasodilator responses in animal models of obesity.16,17 For instance, Russell et al reported that in ramipril-treated JCR:LA-cp obese rats, coronary blood flow in response to BK was significantly enhanced.17 It is important to note that in those studies only indirect evidence has been provided for the upregulated tissue ACE in coronary microvessels in obesity. Beneficial effects of systemic ACE inhibition could be related to the blood pressure (BP)-lowering effect and/or improved insulin resistance in obesity.16,17 Whether upregulated tissue ACE directly interferes with the dilator function of coronary resistance arteries in obesity remains unknown.
Thus, in the current study we set out to elucidate the direct vascular effects of ACE inhibition, with the aims of providing evidence for the upregulation of coronary microvascular ACE in obesity and clinical relevance to obese patients. To this end, BK-induced dilator responses were investigated in isolated coronary arterioles in a rat model of diet-induced obesity.
All protocols were approved by the Institutional Animal Care and Use Committee at Georgia Health Sciences University, Augusta, Georgia, USA. Male Wistar rats (n=24) were purchased from Charles River Laboratories (USA) and maintained in the University’s animal care facility with a 12-h light/dark cycle and free access to food and water. Rats were maintained on standard rat chow (n=12) or HFD (60% fat, 58Y1, TestDiet, PMI Nutrition International, n=12) for 12 weeks as described previously.18
Protocols involving human subjects were approved by the Institutional Review Board at Georgia Health Sciences University, Augusta, Georgia. Consecutive patients (n=74) who underwent cardiac surgery were enrolled in this study.
With the use of microsurgical instruments and an operating microscope, coronary arterioles (~1 mm in length) from the second branch of the septal artery (~1.5 mm in length) of the rat18 or from each patient’s right atrial appendage8 were isolated and transferred to organ chambers containing 2 glass micropipettes filled with physiological salt solution (PSS) comprising (in mmol/L): 110.0 NaCl, 5.0 KCl, 2.5 CaCl2, 1.0 MgSO4, 1.0 KH2PO4, 5.5 glucose, and 24.0 NaHCO3, equilibrated with a gas mixture of 21% O2–5% CO2-balanced nitrogen, at pH 7.4. The vessels were cannulated at both ends, and the micropipettes were connected with silicone tubing to a pressure servo control system (Living Systems Instrumentation, USA) to set the intraluminal pressure to 80 mmHg. Temperature was set at 37°C by a circulating bath temperature controller (Cole Parmer, USA). Changes in arteriolar diameter were continuously recorded with a microangiometer (Colorado Video, USA), connected to a microscope (Olympus BX2, Olympus, USA).18
Cumulative concentrations of BK (0.1–100 nmol/L), and AngII (1–100 nmol/L) were administered to the isolated (rodent or human) coronary arterioles, and changes in diameter were continuously measured. BK-induced responses were also obtained after in vitro incubation with captopril (10 μmol/L for 30 min) or NG-nitro-L-arginine methyl ester (L-NAME, 200 μmol/L for 30 min). AngII-induced responses were reassessed after in vitro incubation with either AngII receptor 1 (AT1R) antagonist, losartan (10 μmol/L for 30 min), and after co-incubation with losartan and the AT2R antagonist, PD 123,319 (10 μmol/L for 30min).
Immunohistochemistry was carried out as described previously.18 Briefly, heart samples from control (n=4) and HFD obese rats (n=4), as well as from non-obese (n=3) and obese (n=3) patients were embedded and frozen in OCT compound (Tissue Tek, Electron Microscopy Sciences, USA). Consecutive sections (8-μm thick) of rat hearts were immune-labeled with anti-ACE mouse primary antibody (1:1,000, ab77990, Abcam Inc, USA), anti-α-smooth muscle actin rabbit primary antibody (1:400, ab5694, Abcam Inc), as well as rabbit polyclonal anti-BK B1 receptor (BK1R, 1:50, sc-25484, Santa Cruz Biotechnology, USA) and mouse monoclonal anti-BK B2 receptor (BK2R, 1:50, 610451, BD Transduction Laboratories, USA). Moreover, consecutive sections (8-μm thick) of human heart samples were immune-labeled with anti-ACE mouse primary monoclonal antibody (1:400, ab77990, Abcam Inc). Primary antibodies were visualized using Cy3-labeled anti-mouse (1:400, A10520, Invitrogen, USA) and Cy5-labeled anti-rabbit (1:250, A10524, Invitrogen) secondary antibodies correspondingly. DAPI was used for nuclear staining. For nonspecific binding, the primary antibody was omitted. Images were acquired by using a Zeiss Axio Imager M2 microscope equipped with an Apotome (Carl Zeiss Microscopy, USA).
Isolated coronary arterioles were incubated with AngI (100 nmol/L, for 15 min) in the absence or presence of the ACE inhibitor, captopril (10 μmol/L), and diameter changes were measured. Tissue ACE activity was then expressed as the captopril-inhibited percent change in diameter in response to AngI.
Logistic regression was used to predict patient characteristics potentially associated with coronary arteriolar dilation to BK. Dilation >40% to BK (10 nmol/L) was considered an endpoint for this analysis. Patient characteristics considered significant (P<0.10) in the chi-square univariate analysis were included in multivariate models. PASW Statistics version 18 was used for this analysis. All other statistical analyses were performed using GraphPad Prism Software (USA) by 2-way repeated-measures ANOVA followed by Tukey’s post hoc test or Student’s t-test as appropriate. Data are expressed as mean ± SEM. Agonist-induced arteriolar response was expressed as the change in arteriolar diameter as a percentage of the maximal dilation defined as the passive diameter of the vessel at 80-mmHg intraluminal pressure in a Ca2+-free medium. P<0.05 was considered statistically significant.
Experimental obesity was induced in rats by 12 weeks of HFD. Their body weight increased and serum levels of insulin, cholesterol, and glucose were elevated, as reported previously.18 In the isolated coronary arterioles, spontaneous tone developed in response to an increase in intraluminal pressure of 80 mmHg without the use of any vasoactive agent. The active diameter of coronary arterioles obtained from control and HFD rats was 123±12 and 128±17 μm, respectively, whereas the passive diameter (in the presence of calcium-free solution) was 175±19 and 187±18 μm, respectively (no significant difference between groups).
We found that cumulative concentrations of BK elicited diminished dilation in the coronary arterioles of HFD rats, whereas BK substantially dilated vessels from the control animals (Figure 1A). Incubation in the presence of the NO synthase inhibitor L-NAME did not significantly affect BK-induced dilation in the coronary arterioles of control rats (Figure 1B), whereas L-NAME ameliorated constriction by BK in the coronary arterioles of HFD rats (Figure 1C). Also, we found similar expression patterns of the BK B1 and B2 receptors in the control and HFD rats, mainly localized to the endothelial layer of coronary arterioles (Figure 1D).
Incubation in the presence of captopril did not affect the basal diameter of isolated coronary arterioles (113±15 and 114±20 μm in control and HFD arterioles, respectively). We found that in the presence of captopril, the dilation response to BK was restored in the coronary arterioles of HFD rats, close to the control level (Figure 2B). Captopril did not affect BK-induced dilation in control vessels (Figure 2A). In association with these functional changes, we detected abundant immunostaining of ACE in coronary endothelial cells in both control and HFD rats (Figure 2C). As assessed by the response to the ACE precursor, AngI, we observed a trend toward increased vascular ACE activity in HFD arterioles, when compared with the responses of control vessels (Figure 2D).
Moreover, we found that administration of AngII elicited dose-dependent dilation of coronary arterioles with a similar magnitude in both control and HFD rats. AngII-induced dilation was not affected by the AT1R antagonist, losartan, whereas dilation was abolished in both groups by the simultaneous presence of losartan and the AT2R antagonist, PD 123,319, (Figures 3A,B).
Patients’ characteristics are summarized in Table 1. Patients from whom coronary arterioles were obtained for functional studies had a mean body mass index (BMI) of 35.6±1.8 (obese) or 25.9±0.5 (non-obese). Obese patients had significantly increased serum glucose levels, but there was no major difference in the other clinical parameters studied. Significantly more of the obese patients suffered from kidney disease, received β-blockers and underwent coronary artery bypass graft surgery (Table 1).
In the isolated human coronary arterioles from non-obese and obese patients, spontaneous tone developed in response to an intraluminal pressure of 80 mmHg, with no significant difference between groups (active diameter: 77±4 μm and 74±6 μm, respectively; passive diameter: 118±5 μm and 111±6 μm, respectively).
First, to assess the effect of obesity on dilator function we compared BK-induced dilation in coronary arterioles isolated from obese and non-obese patients, and found no significant difference in the magnitude of BK-induced coronary dilations between the 2 groups (Figure 4A). Also, incubation with the NO synthase inhibitor, L-NAME, had no significant effect on BK-induced dilation of arterioles from obese or non-obese patients (Figures 5A,B).
In protocols similar to those performed in rats, we found that captopril administration in vitro significantly enhanced the dilation response to BK of arterioles from obese, but interestingly, not from non-obese patients (Figures 4B,C). An abundant ACE immunostaining was detected in the endothelial layer of coronary arterioles from both groups (Figures 4D,4E). We also found significantly increased ACE activity in the coronary arterioles of obese patients when compared with arteries from non-obese individuals (Figure 4F).
Moreover, AngII elicited the same magnitude of dilation of coronary arterioles from the obese and non-obese patients (Figures 5C,D). Incubation with the AT1R antagonist, losartan, enhanced AngII-induced dilation of coronary arterioles from obese patients only at the highest AngII concentration (10–7mol/L) (Figure 5D); losartan did not affect AngII-induced dilation of arterioles from non-obese patients. Simultaneous incubation and presence of losartan and the AT2R antagonist, PD 123,319 completely abolished AngII-induced dilation in both groups.
Because the BK-induced vasodilator response can be affected by multiple pathological factors, we performed a series of logistic regression analyses of various cofactors in order to identify those that were predictive of dilatation of coronary arterioles in the obese patient population. Patient age, sex, and use of statins, angiotensin-receptor blockers, aspirin, antidiabetic agents and calcium-channel blockers were considered in the univariate analysis. Statins and ACE inhibitors were the only variables in the univariate logistic regression that were suggestive (P<0.1) of ≥40% dilation in response to BK (Table 2A). Based on this, statins and ACE inhibitors were included in the multivariate logistic regression model to evaluate ACE inhibitors as independent predictors of improved vessel dilation response to BK. Multivariate logistic regression, after controlling for the effect of statins, demonstrated that obese patients with prior use of ACE inhibitors exhibited a significantly higher dilation response to BK than those not on ACE inhibitors. The use of statin therapy was highly predictive of improved vessel dilation response to BK (Table 2B).
Because ACE inhibitors can potentially alter vasomotor function indirectly through their BP-lowering effect, we also performed additional linear regression analyses to test for association between BP and BK-induced vasodilation. We did not detect any significant association between systolic or diastolic BP (R2: 0.011, P=0.417, R2: 0.001, P=0.807, respectively) and the magnitude of BK-induced dilation.
This study demonstrated enhanced tissue ACE activation in coronary resistance arteries in obesity, a pathological alteration that interferes with BK-mediated regulation of arteriolar diameter. This conclusion is supported by the finding that (1) HFD obese rats showed diminished BK-induced coronary arteriolar dilation, which was restored by in vitro administration of the ACE inhibitor, captopril; (2) similarly, in coronary arterioles from obese patients, in vitro administration of captopril enhanced the dilatory response to BK, which was associated with increased ACE activity of coronary vessels when compared with non-obese subjects; and (3) in our retrospective analysis, obese patients who were taking ACE inhibitors prior to cardiac surgery exhibited enhanced BK-induced dilation in coronary resistance arteries.
Previous studies have shown that systemic administration of ACE inhibitors improves vasodilator function in animals with experimental obesity and insulin resistance. For instance, in obese Zucker rats, oral administration of the ACE inhibitor, imidapril (10 mg/kg), was associated with increased liver and skeletal muscle blood flow.16 Acetylcholine (ACh)-induced dilation in epineural arterioles was improved in obese Zucker rats by treatment with enalapril (20 mg/kg).19 In a study by Duarte et al, endothelium-dependent relaxation of isolated aorta was studied in obese Zucker rats treated orally with either captopril or enalapril (50 mg/kg and 10 mg/kg, respectively). Interestingly, they found that only the sulfhydryl group containing captopril augmented the impaired endothelium-dependent aortic relaxation in obese animals.20 The beneficial effect of ramipril (1 mg/kg) in improving endothelial dysfunction of the aorta was also demonstrated in obese JCR:LA-cp rats.17 In that study, BK-stimulated coronary blood flow was also measured in isolated rat hearts. The authors found enhanced coronary flow in response to BK in ramipril-treated JCR:LA-cp rats, indicating the beneficial effect of ACE inhibitors on myocardial perfusion.17 These studies reveal that part of the beneficial vascular effects of systemic ACE inhibition is indirect and can be related to BP-lowering or improvement of insulin resistance. Indeed, it has been postulated that ACE inhibitors may increase insulin sensitivity in patients with type 2 diabetes.21 It remains unclear whether or not ACE inhibitors have a direct vascular effect, mediated by local inhibition of tissue ACE in the coronary circulation.
In the heart, AngII is primarily synthesized in situ via the conversion of Ang I, a mechanism that appears to be mediated by tissue ACE rather than by the circulating enzyme.22 Although AngII has many adverse, mainly long-term, effects in the heart,11 it preferentially dilates coronary resistance arteries, primarily via activation of AT2R.12,13 The contribution of AngII and AT2R activation to the regulation of coronary arteriolar diameter in obesity is unclear. An earlier study by Zhang et al demonstrated that in dogs fed a HFD, AngII elicited constriction, but dilated the coronary arterioles of control animals.15 The present study shows that AngII-induced dilation in coronary microvessels with no significant difference in the overall magnitude of the dilation between control and HFD rats, as well as between non-obese and obese patients. In order to assess the potential contribution of AT1R and AT2R in this response, AngII-induced dilation was measured in the presence of both the AT1R blocker, losartan, and the AT2R blocker, PD 123,319. We found no significant effect of losartan in AngII-mediated dilation in HFD rats. In the coronary arterioles of obese patients, administration of losartan elicited a trend toward an enhanced AngII-induced dilation, whereas AngII-induced responses were entirely abolished by additional application of AT2R antagonist in both the HFD rats and obese patients. Collectively, these data suggest only a minor contribution of AT1R to the AngII-induced response and indicate preserved AT2R-dependent dilator signaling in obesity.
Thus, it seems that in obesity there are mechanisms other than increased AngII production that are primarily responsible for the impaired BK-mediated regulation of coronary microcirculation. Indeed, it has been postulated that the effects of ACE inhibitors are mainly attributable to an increase in the tissue level of BK in the microvasculature.23 In low, nanomolar concentrations, BK is converted by ACE into an inactive metabolite BK-(1–7), which is further converted into BK-(1–5).24 BK-(1–5) has no vasoactive effect, although it may inhibit thrombin-induced platelet aggregation.25 Kuga et al demonstrated that in epicardial coronary arteries BK-induced increases in diameter were further enhanced by intracoronary infusion of enalaprilat in patients without significant coronary stenosis.3 Given that, in our study we raised the hypothesis that obesity leads to increased activity of microvascular ACE, which mainly manifests as increased breakdown of the vasodilator BK. To furnish evidence for this scenario, coronary arterioles were dissected from the heart and the BK-induced vasomotor responses were investigated in isolated microvessels ex vivo. In this study design, the function of tissue ACE and its effect on BK-induced responses can be assessed independently of the myocardium and systemic circulation. We found that exogenous BK elicited diminished dilation of isolated coronary arterioles from HFD rats. There were no major changes in the expression of BK receptors (BK1R or BK2R) in lean or obese animals. It is known that BK, via activation of its receptors, may activate the synthesis of several vasoactive molecules in coronary arteries, including nitric oxide (NO), prostaglandins and endothelium-derived hyperpolarizing factor.26 Under certain pathological conditions, the lack of NO production may convert vasodilation to vasoconstriction.27,28 In this context, we have reported earlier that synthesis of endothelial NO is reduced in the coronary microvessels of obese rats.18 In the present study, we found that the NO synthase inhibitor, L-NAME, did not affect BK-induced dilation in isolated human coronary arterioles from either lean or obese patients. These findings are in accordance with previous observations obtained in isolated human coronary arterioles29 and indicate only minor, if any, involvement of NO in BK-induced coronary arteriolar dilation. Similar results were obtained in the control rat coronary arterioles, demonstrating the lack of contribution of NO to BK-induced dilation. Interestingly, in the coronary arterioles of HFD rats, BK-induced constriction was ameliorated in the presence of L-NAME, which suggests that in obesity NO synthase may become a source of a constrictor factor. Indeed, when NO synthase is uncoupled, it generates superoxide anion,30 which may facilitate the production of constrictor prostanoids in vascular endothelial cells.31 Thus, we concluded that BK receptors and changes in their initiated downstream signaling are unlikely to contribute to the diminished BK-induced coronary responses in HFD rats.
When isolated coronary arterioles were incubated with the ACE inhibitor, captopril, the reduced dilation response to BK was restored in HFD rats, close to a similar level observed in the control vessels. In humans, although there was no significant difference in the magnitude of BK-induced coronary dilation between the obese and non-obese patients, we found that captopril significantly enhanced the dilation response to BK preferentially in obese individuals. In association with these functional changes, we detected increased microvascular ACE activity in arterioles obtained from obese subjects. Collectively, these observations are the first evidence that ACE expressed in the wall of coronary resistance arteries interferes with BK-mediated dilation in obesity.
In order to further investigate the effect of obesity-dependent activation of coronary microvascular ACE and to reveal potential therapeutic implications, we retrospectively analyzed the effects of ACE inhibitor therapy on the BK-induced coronary response in obese patients. We found that the use of ACE inhibitors prior to cardiac surgery predicted an enhanced vasodilation response to BK. Statin therapy was also highly predictive of improved vessel dilation. In the multivariable logistic regression analysis, through controlling for statin therapy, we identified ACE inhibitor therapy as an independent predictor of improved vasodilation in obese patients. However, the possible influence of other modifying factors (such as the BP-lowering effect of ACE inhibitors) cannot be entirely excluded, which may limit the interpretation of these data. It should be also noted that in our isolated vessel studies the effect of the ACE inhibitor, captopril, on the BK-induced coronary responses was investigated irrespective of prior ACE inhibitor therapy. Given that, although our results suggest a direct vascular effect of captopril in obese patients, a conclusion in regard to the indirect effect on vasomotor behavior has limitations.
Clinical evidence indicates that ACE inhibitors provide end-organ protection independent of their BP-lowering effects. Two large clinical trials, HOPE32 and EUROPA,33 assessed the effectiveness of ACE inhibitor therapy and found that treatment with ramipril or perindopril was associated with a reduction in the risk of cardiovascular death, myocardial infarction and stroke or cardiac arrest in patients with high risk for cardiovascular diseases. Of note, when combined with ACE inhibitors, AT1R blockers seem to have minimal effect on the progression of cardiovascular disease as compared with ACE inhibitor therapy alone.34 These results indicate that ACE inhibitor therapy is particularly effective in reducing the incidence of cardiovascular events in high-risk patients. Whether or not obese patients particularly benefit from ACE inhibitor therapy in preventing cardiovascular complications is not entirely understood. It is known that obesity is frequently associated with left ventricular hypertrophy and congestive heart failure, pathological alterations that can be effectively targeted and treated with ACE inhibitors.35 These observations provide only indirect clinical evidence of the upregulation of the cardiac renin-angiotensin-system in obesity. There are recent studies indicating that obese patients exhibit enhanced activation of both systemic and tissue ACE. For example, Cooper et al showed that obese individuals have higher plasma ACE activity than non-obese individuals.36 In obese subjects with an average BMI of 35, serum ACE activity significantly decreased with only a modest weight loss, which was induced by caloric restriction over 5 weeks.37 In the study by Barton et al, a tissue-specific increase in ACE activity was found in the kidneys of obese patients.38 Moreover, Bitigen et al investigated the relationship between ACE genotype and left ventricular contractile dysfunction in obese patients and found that ACE DD genotype frequency, which is associated with elevated ACE activity, was increased in the obese subgroup with diastolic dysfunction.39 Similarly, Niemiec et al found that the ACE DD genotype increased the risk of coronary artery disease in obesity.40
In conclusion, we provide functional evidence for upregulated ACE activity in the coronary microvascular wall in obesity. We suggest that upregulation of ACE is primarily responsible for the diminished dilatory response to BK, which can be prevented by ACE inhibition to enhance vasodilation of coronary arterioles. Further clinical investigations may provide the rationale for ACE inhibitor therapy in obese patients, which may correlate with improved dilator function of coronary resistance arteries, thus preventing cardiovascular complications in obesity.
This study is supported by R01 HL104126 grant from the National Heart, Lung and Blood Institute. Z.B. acknowledges the support from the British Heart Foundation, Centre of Research Excellence, Oxford (RE/08/004). A.F. acknowledges receiving a Young Investigator Award from the Japanese Society for Microcirculation.