Peroxynitrite Disrupts Endothelial Caveolae Leading to eNOS Uncoupling and Diminished Flow-Mediated Dilation in Coronary Arterioles of Diabetic Patients

Diabetes mellitus (DM) is associated with impaired function of coronary resistance arteries. The consequence is a mismatch between myocardial blood supply and demand, leading to ischemic episodes in the diabetic heart (1). Reduced myocardial perfusion may develop even in the absence of occlusive coronary artery disease. For instance, Nitenberg et al. (2) have demonstrated that in DM patients with angiographically normal coronary arteries, cold pressor test–induced dilation was reduced, as estimated by coronary surface area during quantitative angiography. Intracoronary injection of acetylcholine (ACh), which normally dilates arteries, caused vasospasm preferentially in diabetic patients (3). In small coronary arteries dissected from the heart of diabetic patients, Miura et al. (4) and recently our group (5) demonstrated ACh-induced constriction. The underlying mechanism(s) responsible for coronary microvascular dysfunction remained poorly understood in human DM.

Flow-mediated dilation (FMD), the intrinsic regulatory mechanism of resistance arteries in response to increases in wall shear stress (WSS) (6), is one of the key determinants of myocardial perfusion (7). The endothelium-dependent FMD is responsible for normalizing WSS and provides a feed-forward mechanism for increases in myocardial blood flow and metabolite-induced vasodilation (6). The problem is that this regulatory mechanism often fails in the diseased heart. Early studies have shown that brachial artery FMD is reduced in patients with type 1 (8) and type 2 DM (9). Previously, we demonstrated that in a mouse model of type 2 DM (db/db mice), WSS-induced dilation is diminished in coronary arterioles, which is due to the reduced availability of nitric oxide (NO) (10,11). In DM, among other key pathological mechanisms responsible for reduced NO synthase (NOS) activation (12,13), the direct interaction of NO with superoxide anion (14) has received substantial attention. It is of particular importance that the NO-superoxide interaction not only reduces NO bioavailability but also generates the reactive peroxynitrite (ONOO⁻). ONOO⁻ has numerous detrimental effects in the cardiovascular system and plays a crucial role in the development of DM-associated vascular pathology (15). Emerging evidence indicates that ONOO⁻ preferentially targets vascular endothelium and specific subcellular compartments within (15). The footprint of enhanced ONOO⁻ production is 3-nitrotyrosine (3-NT) formation. 3-NT–modified proteins were detected even in the normal rat aorta endothelium, with a distinct localization to mitochondria (16). Some evidence indicates that in aging rat, 3-NT formation is enhanced and more widely distributed within subcellular compartments of the endothelium (17). Yet, it is unknown whether ONOO⁻ preferentially targets endothelial plasma membrane microdomains, such as caveolae, a specific type of cholesterol-rich membrane lipid raft, which plays a key role in WSS-induced, NO-dependent regulation of arterial diameter (18).

Accordingly, in this study, we set out to investigate alterations in coronary FMD in DM patients and to elucidate the impact of ONOO⁻, which we hypothesized preferentially targets endothelial caveolae. In this study, we demonstrate that NO-mediated coronary FMD is markedly reduced in patients with DM. We show that ONOO⁻ targets and disrupts endothelial membrane caveolae, which subsequently predisposes NOS for uncoupling in DM.

Protocols were approved by the institutional review board at our institution. Consecutive patients undergoing heart surgery were enrolled in this study. Patients were divided into two groups with or without documented DM. Patients were included irrespective of DM duration.

Video microscopy of isolated human coronary arterioles were performed as previously described (19). In brief, coronary arterioles (diameter ∼100 µm) were dissected from the right atrial appendages obtained from patients at the time of heart surgery. Arterioles were cannulated and pressurized (70 mmHg), and changes in diameter were measured with a video caliper (Colorado).

Intraluminal flow was induced by changing perfusion pressure with equal degree but in opposite direction, in three steps (ΔP: 25–50–90 cmH₂O). Fig. 1A shows that ΔP elicited increases in flow (∼8 to ∼35 μL/min), which resulted in increases in WSS (from ∼5 to ∼15 dyne/cm²). To assess the role of NO, ONOO⁻, and NOS cofactor, tetrahydrobiopterin (BH₄) in affecting coronary FMD, coronary arterioles were incubated with NOS inhibitor, Nω-nitro-l-arginine methyl ester (l-NAME, 2 × 10⁻⁴ mol/L for 30 min), selective ONOO⁻ sequesters, iron(III)tetrakis(N-methyl-4'pyridyl)porphyrin-pentachloride (FeTMPyP, 10⁻⁴ mol/L) and uric acid (10⁻⁴ mol/L for 30 min), or stable BH₄ precursor, sepiapterin (10⁻⁵ mol/L for 30 min), respectively, and coronary FMD was reassessed. In other protocols, after incubation with FeTMPyP or sepiapterin, vessels were exposed to additional l-NAME, and coronary FMD was obtained again. In separate protocols, arterioles were exposed to exogenous ONOO⁻ (10⁻⁵ mol/L) and FMD was assessed. To investigate the role of caveolae, arterioles were preincubated with methyl-β-cyclodextrin (mβCD, 2 × 10⁻³ mol/L for 60 min), which is known to disrupt caveolae (20).

Superoxide production in coronary arteries was measured by using the dihydroethidium (DHE) method, as described previously (10,11). In brief, cryosections (8-μm thick) from snap-frozen human atrium were incubated with DHE (10 μmol/L for 30 min) with or without pretreatment with superoxide scavenger Tiron (5 mmol/L for 30 min). Slides were washed in PBS, and sections were covered with DAPI-Vectashield. Fluorescence intensity measurements (excitation BP545/25; emission BP605/70) of oxidized, nuclei-trapped DHE were performed in nuclei in the vessel wall and normalized to nuclei area.

Atrial appendages from non-DM and DM patients were fixed in 4% paraformaldehyde and paraffin embedded. Consecutive sections (8-μm thick) were blocked with normal donkey serum (1 h) and immune labeled with monoclonal anti–caveolin 1 (Cav-1) (1:100, overnight 4°C; Abcam) and polyclonal anti–3-NT (1:1,000, overnight 4°C; Sigma-Aldrich) antibodies. Immunofluorescent labeling was performed with corresponding Cy3 and Cy5 secondary antibodies (Jackson ImmunoResearch). DAPI was used for nuclear staining. For nonspecific binding, the primary antibody was omitted. Images were collected with fluorescent microscopy. The colocalization between Cav-1 and 3-NT was calculated with linear regression and Pearson correlation coefficient (r), as described by Zinchuk et al. (21), in DM and non-DM patients by an investigator unaware of sample identity.

In additional experiments, in situ proximity ligation assay (PLA; Duolink; Sigma-Aldrich) was performed to detect colocalization of Cav-1 and 3-NT according to manufacturer instructions. In brief, PLA reactions were performed using the same primary antibodies (anti–Cav-1 [1:100] and anti–3-NT [1:1,000]), followed by a pair of oligonucleotide-labeled secondary antibodies (orange PLA probes). In this assay, PLA probes create a positive signal only when the epitopes of the target proteins are in close proximity (<40 nm). The signal from each detected pair of PLA probes was then counted using fluorescence microscopy (excitation BP545/25; emission BP605/70). For negative controls, primary antibodies were omitted.

The right atrial appendages were fixed in 4% formaldehyde and 0.2% glutaraldehyde in 0.1 mol/L sodium cacodylate buffer (pH 7.4). Sections were cut with Leica EM UC6 ultramicrotome (Leica Microsystems, Inc., Bannockburn, IL). Sections were blocked in normal goat serum (1 h) and incubated with primary rabbit anti–3-NT polyclonal antibody (1:10, overnight 4°C; Sigma-Aldrich) or anti–Cav-1 polyclonal antibody (1:100, overnight 4°C; Cell Signaling) and with additional gold-labeled (10 nm) secondary antibody. Sections were observed in a JEM 1230 transmission electron microscope (JEOL USA, Inc.). Counting of anti–3-NT-immunogold and anti–Cav-1-immunogold particles within coronary endothelial cells was evaluated by an investigator unaware of sample identity.

Electron microscopy of atrial samples was performed as described above. Images selected for analysis contained at least one erythrocyte to ensure accuracy in identifying the endothelial cell layer. The number of caveolae per μm² of cell membrane was counted, as described previously (22), and compared in DM and non-DM patients by an investigator unaware of sample identity. In separate protocols, atrial samples were exposed to increasing concentrations of exogenous ONOO⁻ (1, 10, and 100 μmol/L), and the number of endothelial caveolae was calculated.

Western immunoblot analysis was performed as described previously (5). In brief, coronary arterioles were homogenized in radioimmunoprecipitation assay buffer, and protein concentration was measured by Bradford assay. Equal amounts of proteins were loaded for gel electrophoresis. After blotting, a polyclonal anti–3-NT antibody (dilution 1:1,000) was used for the detection of 3-NT. Membranes were reprobed with anti–β-actin IgG (dilution 1:5,000) to normalize for loading variations. Corresponding horseradish peroxidase–labeled secondary antibody was used, and chemiluminescence was visualized autoradiographically.

Human coronary artery endothelial cells (HCAECs) (Lonza, Walkersville, CA) were grown in EBM-2 medium. Cells were maintained in normal (5.5 mmol/L) or high glucose medium (HG; 25 mmol/L for 24 h). HCAECs were then harvested for cell fractionation using ultracentrifugation. In brief, cells were washed in PBS and lysed in MSE extraction buffer (10 mmol/L Tris-HCl, 220 mmol/L d-mannitol, 70.1 mmol/L sucrose, 1 mmol/L EGTA, 0.025% BSA). Protein was measured (Bio-Rad), and equal amount of protein was centrifuged at 100,000g for 20 min. Supernatant was collected as cytosolic fraction, and pellet contained the membrane fraction. Equal amounts of total lysate and cytosolic and membrane fractions were loaded for Western blot analysis.

Protocols involving mice were approved by the Institutional Animal Care and Use Committee. Video microscopy and wire myography to assess coronary FMD (10,11) and relaxation of aorta were performed as described previously (23). In brief, the heart and thoracic aorta were excised from Cav-1 knockout (CavKo) (Cavtm1Mls/J; The Jackson Laboratory) and wild-type mice. Mouse coronary arteries (diameter ∼100 µm) were isolated and pressurized (70 mmHg). Arteries were preconstricted with thromboxane analog, U46619 (10⁻⁹ mol/L). Diameter changes of arterioles to increases in WSS and ACh (10⁻⁹ to 10⁻⁶ mol/L) were measured with video microscopy. Aorta sections (2 mm in length) were mounted on wire myograph to measure isometric force generation. Aorta were preconstricted with phenylephrine (10⁻⁶ mol/L), and relaxation was assessed in response to ACh (10⁻⁹ to 10⁻⁶ mol/L). Coronary arteries and aortic rings were incubated with sepiapterin (10⁻⁵ mol/L, 30 min) or sepiapterin plus l-NAME (2 × 10⁻⁴ mol/L, 30 min), and WSS- and ACh-induced responses were reassessed.

Statistical analyses were performed using IBM SPSS Statistics 19. Agonist-induced arteriole responses were expressed as changes in diameter as a percentage of the maximal dilation, defined as passive diameter of the vessel at 70 mmHg in calcium-free physiological salt solution. Aortic relaxations in mice were presented as percent changes in force after agonist administration. Statistical analysis was performed by repeated-measures ANOVA, followed by Tukey post hoc test. P < 0.05 was considered statistically significant. Data are expressed as mean ± SEM.

Patient demographics and clinical data are presented in Table 1. The age, sex, clinical parameters, underlying diseases, and medications were similar between the two populations, except the documented DM, glucose levels, and the use of insulin or oral antidiabetics. Patients with DM were more likely to undergo coronary artery bypass graft surgery than non-DM patients. Thus, the potential influence of antidiabetic medications and the type of the surgery on the measured end points cannot be entirely excluded in this study.

In isolated coronary arterioles, a spontaneous tone developed in response to 70 mmHg intraluminal pressure. There were no significant differences between the active (non-DM, 105 ± 7 µm; DM, 94 ± 7 µm) and passive (non-DM, 143 ± 8 µm; DM, 123 ± 7 µm) diameters in DM and non-DM patients. Incubation with sepiapterin, ONOO⁻ sequesters, FeTMPyP, uric acid, and/or l-NAME had no effect on the spontaneously developed tone (data not shown).

In coronary arterioles of non-DM patients, increases in intraluminal flow, via increasing WSS (Fig. 1A), elicited dilations, whereas coronary arterioles from DM patients exhibited a diminished FMD (Fig. 1B and C). The NOS inhibitor l-NAME reduced coronary FMD in non-DM, but not in DM, patients (Fig. 1D). Dilations to the NO donor sodium-nitroprusside were similar in DM and non-DM patients (Fig. 1E).

We found a significantly increased DHE-detected, Tiron-inhibited superoxide production in coronary arteries of DM patients, when compared with non-DM (Fig. 2A and B). Moreover, the increased level of 3-NT in coronary arterioles indicated an augmented ONOO⁻ production in DM patients (Fig. 2C). Incubation of coronary arterioles with the selective ONOO⁻ scavengers FeTMPyP (Fig. 2D) or uric acid (Fig. 2E) significantly enhanced coronary FMD in DM, whereas exogenous application of ONOO⁻ reduced coronary dilation in non-DM patients (Fig. 2F).

The subcellular localization of 3-NT was investigated in coronary arterioles using immune electron microscopy and immunohistochemistry approaches. Whereas 3-NT immunogold labeling was less prominent within the endothelium of non-DM patients, in DM, the number of 3-NT immunogold particles was increased and more widely distributed, with a significantly increased presence in the luminal plasma membrane of endothelium (Fig. 3A and Table 2 for detailed quantification).

To test whether ONOO⁻ targets endothelial membrane caveolae, coimmunostaining of 3-NT and Cav-1 was performed. In coronary endothelial cells, a greater degree of colocalization between 3-NT and Cav-1 was found in DM patients, as indicated by Pearson correlations of fluorescence intensities (Fig. 3B and C). Colocalization of 3-NT and Cav-1 was confirmed by in situ PLA in further independent experiments (Fig. 3D and E).

We found no significant change in total Cav-1 expression between coronary arteries of DM and non-DM patients (Fig. 4A). When distribution of Cav-1 immunogold particles was investigated by electron microscopy, we found that membrane-localized Cav-1 was significantly reduced in DM endothelium (Fig. 4 and Table 2). To confirm these changes in the membrane Cav-1 content, we also used primary cultured HCAECs that were exposed to high glucose (25 mmol/L) concentrations. We found that exposure of HCAECs to high glucose significantly reduced membrane Cav-1 content, whereas it did not alter total and cytosolic level of Cav-1 (Fig. 4C and D).

Electron microscopy was used to evaluate the number of endothelial caveolae in DM and non-DM patients (Fig. 5A). We found a significantly reduced number of endothelial caveolae in DM patients, when compared with non-DM (Fig. 5B and C). Moreover, coronary arteries were exposed to exogenous ONOO⁻. ONOO⁻, in a dose-dependent manner reduced the number of endothelial caveolae (Fig. 5D and E). ONOO⁻, at the highest concentration, disrupted the integrity of the endothelial membrane, which was lacking any caveolae and showed signs of membrane blebbing.

We found that administration of sepiapterin, a stable precursor of BH₄ significantly enhanced coronary FMD in DM patients, in an NO-dependent manner, while it did not affect dilation in non-DM patients (Fig. 6A). Pharmacologic disruption of caveolae with mβCD entirely abolished FMD in coronary arterioles in non-DM patients, which was partially restored by additional administration of sepiapterin (Fig. 6B).

To provide additional evidence for the lack of caveolae contributing to endothelial NOS (eNOS) uncoupling, CavKO mice were used. A phenotypic characteristic of CavKO mice is the complete lack of endothelial caveolae (24), which we also observed in this study (data not shown). We found that increases in WSS elicited dilations in wild-type mice but caused constrictions in coronary arteries of CavKO mice, which was converted to dilation by sepiapterin, in an NO-dependent manner (Fig. 6C and D). There was no difference in ACh-induced coronary dilation in wild-type and CavKO mice either in the absence or presence of sepiapterin, and the responses were only partially inhibited by l-NAME (Fig. 6E). In the aorta of CavKO mice, administration of sepiapterin significantly enhanced relaxation to ACh, and the response was abolished by additional l-NAME, suggesting a greater contribution of NO to ACh response in the mouse aorta (Fig. 6F).

This study demonstrates that in coronary resistance, arteries of diabetic patients increased production of ONOO⁻ targets and disrupt endothelial caveolae, which in turn contributes to uncoupling of eNOS and results in a diminished FMD. This conclusion is supported by key findings. 1) In coronary arterioles of DM patients, ONOO⁻ sequester restored FMD, whereas in non-DM patients, exogenous ONOO⁻ reduced FMD. 2) 3-NT staining was localized at the endothelial plasma membrane in a close proximity to Cav-1. 3) Coronary arterioles of DM patients exhibited a reduced level of membrane-localized Cav-1 and also a reduced number of endothelial caveolae. 4) The BH₄ precursor sepiapterin enhanced coronary FMD after pharmacological or genetic (CavKO mice) disruption of caveolae. Finally, 5) sepiapterin restored coronary FMD in DM patients.

A solid line of evidence indicates that increased production of ONOO⁻ plays an important role in the development of cardiovascular complications in DM (15). ONOO⁻ is a powerful oxidizing agent that causes rapid depletion of sulfhydryl groups, causes DNA damage and protein nitration of aromatic amino acid residues leading to 3-NT formation (15). Cellular mechanisms to prevent the deleterious effect of ONOO⁻ are not clearly defined (25). Previous studies have shown a significantly enhanced formation of 3-NT in the coronary endothelium in sepsis (26), in the coronary artery wall in human transplant coronary artery disease (27), and in coronary vessels in an animal model of DM (28). In this study, we found an elevated superoxide anion and increased 3-NT formation in coronary microvessels of DM patients. Given that the two patient populations exhibited similar comorbidities, it is likely that among various confounding factors, DM and most likely high blood glucose concentration are the major contributors for enhanced ONOO⁻ formation in patients undergoing heart surgery. Importantly, in this study, we found that ONOO⁻, in a reversible manner, diminishes NO-dependent FMD in coronary arterioles of DM patients.

Next, we aimed to elucidate ONOO⁻-targeted subcellular mechanisms that are responsible for diminished NO availability in coronary arterioles of DM patients. Under physiological conditions, ONOO⁻ production, as detected by 3-NT, is restricted to the mitochondria of the vascular endothelium (16). WSS-dependent activation of eNOS, on the other hand, is primarily coupled to regulatory mechanisms at the endothelial plasma membrane, although some evidence indicates involvement of mitochondria in this process (29). Among others, platelet endothelial cell adhesion molecule-1 (30,31) and endothelial caveolae (18) located at the plasma membrane have been shown to mediate WSS-dependent vasodilation in rodents. Given that ONOO⁻ causes lipid peroxidation (15), we wondered whether ONOO⁻ preferentially targets and interferes with vasoregulatory mechanisms intrinsic to the endothelial plasma membrane. In this study, we found that 3-NT immunogold particles were more abundant in the plasma membrane of the coronary endothelium in DM patients. We also demonstrated that 3-NT staining was in close proximity (<40 nm) to Cav-1, the main scaffolding protein of caveolae. There are four tyrosine groups within Cav-1 that can be potentially nitrated. Whether 3-nitration of Cav-1 has any functional consequence is entirely unknown. To investigate possible effects of ONOO⁻ and Cav-1 interaction in DM, we measured Cav-1 expression in coronary arteries and found no significant changes between DM and non-DM patients. Interestingly, immune electron microscopy has revealed that the membrane-localized Cav-1 is significantly reduced in the coronary endothelium of DM patients. The reduced level of membrane Cav-1 was also confirmed in cultured human coronary endothelial cells that were exposed to high glucose concentrations for 24 h. Based on these results, we raised the possibility that alterations in the membrane-localized Cav-1 content may influence the formation and/or stability of endothelial caveolae in DM. Indeed, in this study, we found that DM patients exhibited a significantly reduced number of coronary endothelial caveolae. We also demonstrated that short-term administration of exogenous ONOO⁻ elicited a dose-dependent, significant caveolae loss. In line with this observation, an earlier study has shown that even a short-term myocardial ischemia, due to a temporary occlusion of the coronary artery in anesthetized dogs, elicited a significant loss of caveolae in coronary vessels (32). A later study by Peterson et al. (33) showed that bovine aortic endothelial cells acutely exposed to a superoxide-generating napthoquinolinedione resulted in loss of endothelial cell caveolae. Collectively, we propose that a reduced membrane-localized Cav-1 may lead to destabilization and/or disruption of endothelial caveolae in coronary arterioles of DM patients. We suggest that a direct interaction between ONOO⁻ and Cav-1 is responsible for this pathology. Whether ONOO⁻-induced 3-nitration of Cav-1 contributes to this process or only represents a “footprint” of Cav-1 targeting by ONOO⁻ has yet to be elucidated.

How disruption of endothelial caveolae by ONOO⁻ contributes to an impaired NO synthesis is not well understood. A study by Peterson et al. (34) demonstrated that Cav-1, eNOS, and GTPCH-I, the rate-limiting enzyme for BH₄ biosynthesis, are all localized in the endothelial caveolea. This and other studies indicated a crucial role for caveolae in the compartmentalization of enzymes that are required for NO synthesis (35). It is also known that the interaction between Cav-1 and eNOS is inhibitory (36–38). Loss of caveolae, through exposing endothelial cells to oxidized LDL, cyclosporine, or mβCD has been shown to displace eNOS from the plasma membrane, which is, however, associated with a diminished NO production (39–41). A study by Zhang et al. (42) showed that in endothelial cells, the plasma membrane–targeted eNOS was more efficient to produce NO when compared with the Golgi-targeted eNOS. Moreover, a recent elegant study demonstrated that a noninhibitory mutant of the Cav-1 scaffolding domain, without interfering with other functions of endogenous Cav-1, such as forming caveolae, significantly enhanced eNOS-derived NO synthesis (43). These studies indicated that under normal conditions, the interaction between Cav-1 and eNOS directs eNOS to the caveolae, and that this compartmentalization seems important for eNOS activation and NO production. Given that loss of caveolae structure/function may ultimately lead to an impaired NO synthesis. In order to furnish functional evidence for this scenario, in this study, human coronary arterioles were acutely treated with mβCD to disrupt vascular caveolae. We found that mβCD completely abolished NO-dependent coronary FMD in non-DM patients. Correspondingly, we found that increases in flow induced coronary arterial constrictions in CavKO mice. Our data are the first, to our knowledge, to provide evidence for the crucial role of caveolae in mediating WSS-induced dilation in human coronary arterioles.

The impaired flow-induced coronary dilation after caveolae disruption can be mediated by several mechanisms. Due to the observed close proximity of eNOS to GTPCH-I (34), we raised the possibility that the availability for eNOS cofactor BH₄ is limited after caveolae disruption, which ultimately leads to eNOS uncoupling in DM patients (44,45). Interestingly, in the CavKO mice, the ratio of BH₄ to BH₂ is reduced in the myocardium, which leads to eNOS uncoupling and may contribute to the cardiopulmonary phenotype of these mice (46). In this study, we demonstrated that in mβCD-treated vessels, the stable BH₄ precursor sepiapterin partially restored dilation to flow in human coronary arterioles. Moreover, in mice lacking caveolae, we found that after sepiapterin administration, flow-induced coronary dilation was restored to the control levels, whereas sepiapterin enhanced aortic relaxation. Importantly, the diminished FMD in coronary arterioles of DM patients was also augmented by sepiapterin administration. Collectively, these results are the first to demonstrate that impairment in flow-induced, NO-dependent coronary dilation in DM patients is due to eNOS uncoupling, which is likely mediated by ONOO⁻-dependent disruption of endothelial caveolae. It should be noted that the influence of ONOO⁻ on the function of eNOS might not be limited to the disruption of coronary endothelial caveolae. It has been shown that ONOO⁻ directly interacts and reduces the level of BH₄, as it has greater affinity for BH₄ than that of ascorbic acid and glutathione (47). Chen et al. (48) demonstrated that exposure of human eNOS to ONOO⁻ resulted in a dose-dependent loss of activity with a marked destabilization of the eNOS dimer. They found that both free and eNOS-bound BH₄ were oxidized by ONOO⁻; however, full oxidation of eNOS protein–bound BH₄ required significantly higher ONOO⁻ concentrations (48). Based on the presented data, we also support the direct interaction between ONOO⁻ and BH₄, which explains why ONOO⁻ scavenger and sepiapterin fully restore coronary dilation in DM patients, in spite of having a reduced number of caveolae. Based on the results of this study, we propose a novel “upstream” mechanism by which ONOO⁻ targets and disrupts coronary endothelial caveolae and that a disrupted endothelial caveolae predisposes BH₄ to ONOO⁻-dependent oxidation in DM (Fig. 6G). Restoring endothelial caveolae via ONOO⁻ sequestration and/or treatment with a stable BH₄ precursor may facilitate NO production in the diseased coronary artery, a therapeutic strategy, which may prove beneficial to improve cardiovascular morbidity and mortality in DM.

Acknowledgments. The authors thank Robert Smith, Penny Roon, and Perry Libby (Electron Microscopy and Histology Core Laboratory, Georgia Regents University) for their assistance in the electron microscopy study.

Funding. This study was supported by Grant R01-HL-104126 from the National Heart, Lung, and Blood Institute (to Z.B.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. J.C. and Z.B. participated in the design of the study, performed experiments and the statistical analysis, and wrote the manuscript. H.D., I.C., A.S., and A.F. performed experiments. V.S.P., V.K., and E.B.d.C. participated in the design of the study and wrote the manuscript. M.J.R. participated in the design of the study, performed experiments, and wrote the manuscript. Z.B. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.