Role of Nrf2 Signaling in the Regulation of Vascular BK Channel β1 Subunit Expression and BK Channel Function in High-Fat Diet–Induced Diabetic Mice

Diabetes is closely associated with increased risks of cardiovascular diseases, which are the leading cause of death in the U.S. The large-conductance calcium-activated potassium (BK) channels are abundantly expressed in coronary arterial smooth muscle cells (SMCs), playing an important role in regulating coronary circulation and myocardial perfusion (1–4). Functional vascular BK channels are composed of the pore-forming α-subunits and the regulatory β1-subunits (BK-β1) in 4:4 stoichiometry. The BK-β1, encoded by the KCNMB1 gene, is a key determinant of BK channel electrophysiology. The presence of BK-β1 enhances BK channel α-subunit sensitivity to Ca²⁺, allowing channel activation in the physiological range of Ca²⁺ concentrations and membrane potentials (5–7). Activation of BK channels hyperpolarizes the membrane potentials of vascular SMC, leading to the reduction of intracellular Ca²⁺ concentration and vascular relaxation. However, in response to increased oxidative stress, the vascular BK channel function is impaired, with reduced BK-β1 protein expression being a common feature in the vascular SMCs of prediabetic, type 1 diabetic, and type 2 diabetic animals (8–13).

The nuclear factor-κB (NF-κB) is a family of transcription factors that includes RelA (p65), RelB, c-Rel, and p105/p50 (NF-κB1) or p100/p52 (NF-κB2) subunits (14). Under baseline conditions, p65 is bound to an inhibitory NF-κB subunit (IκB) that keeps it sequestered in an inactive state in the cytoplasm of cells. Phosphorylation of IκB by IκB kinase promotes IκB degradation through the ubiquitin-proteasome system (UPS), which in turn releases p65 and facilitates the nuclear translocation of the p65/p50 or p65/p52 dimeric complex (14). In addition, the NF-κB/p65 can be directly activated by reactive oxygen species (ROS) through phosphorylation of NF-κB/p65 (15). We have recently identified that the muscle ring finger protein 1 (MuRF1), a muscle-specific E3 ligase, is one of the target genes of NF-κB in vascular SMCs (16). An increase in the protein expressions of NF-κB/p50 and NF-κB/p65 was responsible for MuRF1-mediated BK-β1 protein degradation in type 1 diabetic mouse arteries (16). However, the upstream signaling leading to dysregulation of NF-κB expression in diabetic vessels is unclear.

The nuclear factor erythroid-2–related factor 2 (Nrf2) signaling plays a critical role in the maintenance of intracellular redox homeostasis by regulating multiple downstream antioxidant genes and phase II detoxification enzymes, which include NADPH dehydrogenase quinone 1, glutathione-disulfide reductase, glutathione translocase, thioredoxin, thioredoxin reductase 1, heme oxygenase-1 (HO-1), superoxide dismutase, catalase, and glutathione peroxidase (13,17–20). Activation of Nrf2 has been shown to be protective against hyperglycemia-induced, ROS-mediated apoptosis and cell damage in renal, cardiac, and vascular cells (21). C57BL/6 mice fed a high-fat diet (HFD) are an established diet-induced obese/diabetic mouse model (22–25) with metabolic features similar to those of human type 2 diabetes (9,25–27). In this study, we found that Nrf2 protein expression was significantly downregulated with an associated decrease of BK-β1 expression in vascular SMCs of HFD mice, resulting in coronary BK channelopathy and coronary vasculopathy. Pharmacological activation of Nrf2 not only suppressed the NF-κB/MuRF1-mediated protein degradation of BK-β1, but also augmented BK-β1 mRNA expression, protecting the BK channel function and BK channel–mediated coronary vasodilation in HFD mice. Hence, Nrf2 signaling represents a therapeutic target for cardiovascular complications in obesity and type 2 diabetes.

Male mice (C57BL/6J) at 4 weeks of age were purchased from The Jackson Laboratory. These animals were fed with a 60% fat diet (HFD) or 10% fat diet (low-fat diet [LFD]) for 6 months. In the dimethyl fumarate (DMF) treatment studies, after 6 months eating an HFD or LFD animals were randomly divided into four groups (DMF-treated LFD, placebo-treated LFD, DMF-treated HFD, and placebo-treated HFD) and were treated with DMF (25 mg/kg/daily) or placebo (same volume of vehicle) by gavage for 10 days (28,29). Tail blood pressures were measured using a CODA Non-Invasive Blood Pressure System (Kent Scientific Corporation, Torrington, CT), and blood glucose levels were determined using an Accu-CHEK Aviva Glucose Meter (Roche Diabetes Care Inc., Indianapolis, IN). All protocols were approved by the Institutional Animal Care and Use Committee of the Mayo Clinic (Rochester, MN).

Human coronary arterial SMCs and the SmBM culture medium were purchased from Lonza Walkersville Inc. (Walkersville, MD). Cells between passages 5 and 8 were transduced with adenoviral green fluorescent protein (GFP) vectors carrying the Nrf2 gene (Ad-GFP-Nrf2) or Nrf2 short hairpin RNAs (shRNAs) (Ad-GFP-Nrf2 shRNA) at 50 multiplicity of infection (MOI) for 48 h as we previously described (11,16). Transduction of Ad-GFP or Ad-GFP-U6-Scramble-RNAi served as controls. All adenoviral vectors were obtained from Vector BioLabs (Malvern, PA).

Immunoblot analysis was performed in isolated arteries from mice and cultured coronary SMCs, as we reported previously (9,11,16). Rabbit anti-Nrf2 (1:200; catalog #sc-722; Santa Cruz Biotechnology), rabbit anti–HO-1 (1:200; catalog #sc-10789; Santa Cruz Biotechnology), and mouse anti-MuRF1 (1:200; catalog #sc-398608; Santa Cruz Biotechnology), mouse GAPDH-horseradish peroxidase (HRP) (1:5,000; catalog #HRP-6004; Proteintech), and mouse anti–β-actin-HRP (1:10,000; catalog #A3854; Sigma-Aldrich, St. Louis, MO) antibodies were used in this study. Rabbit anti–BK-β1 antibody (1:200) was custom made as previously reported (4). NF-κB family member antibody sample kit (1:200; catalog #4766S) was purchased from Cell Signaling Technology Inc. (Danvers, MA). Rabbit antiphospho-NF-κB p65 (Ser276) (1:200; catalog #PA5–37718) was obtained from ThermoFisher Scientific (Rockford, IL). The optical density of the bands was analyzed with Scion Image Software (Scion Corporation, Frederick, MD). Protein expression was expressed as relative abundance normalized to GAPDH or β-actin.

Single coronary SMC isolation and BK current recordings were performed as previously described (11,16). Whole-cell BK currents were defined as the iberiotoxin (IBTX; 100 nmol/L)-sensitive K⁺ current component, and the inside-out single BK channel currents were determined by their unitary current amplitude and Ca²⁺ sensitivity, as previously described (30,31). For whole-cell BK channel recordings, the pipette solution contained the following (in mmol/L): KCl 140, MgCl₂ 0.5, Na₂ATP 5.0, Na₂GTP 0.5, HEPES 10.0, EGTA 1.0, CaCl₂ 0.465 (∼200 nmol/L free Ca²⁺) at pH 7.38. The bath solution contained the following (in mmol/L): NaCl 145.0, KCl 5.6, MgCl₂ 1.0, CaCl₂ 1.0, HEPES 10.0, and glucose 5.0 at pH 7.40. For inside-out single BK channel recordings, the pipette solution contained the following (in mmol/L): KCl 140.0, CaCl₂ 1.0, MgCl₂ 1.0, EGTA 1.0, and HEPES 10.0 at pH 7.4. The bath solution contained the following (in mmol/L): KCl 140.0, MgCl₂ 1.0, HEPES 10.0, EGTA 1.0, and CaCl₂ 0.465 at pH 7.38. Experiments were performed at room temperature (22°C).

Isolation of total RNA, RT-PCR, and quantitative real-time PCR was performed as previously described (16). The reaction underwent a 40-cycle amplification with the following conditions: denaturalization for 15 s at 94°C; annealing for 30 s at 55°C; and extension for 30 s at 70°C. Copy numbers of the target gene were expressed as 2^-ΔCt (where ΔCt = Ct of target gene − Ct of internal control gene, and Ct represents the threshold cycle). The forward and reverse primer sequences are listed as follows:

• Human BK-β1 forward: 5′-CTTCTCCGCACCTCGGGGGA-3′; reverse: 5′-CGGTCAGCAGGAAGGTGGGC-3′;

• Human GAPDH forward: 5′-ACCACAGTCCATGCCATCACTGCC-3′; reverse: 5′-ACCAGGAAATGAGCTTGACAAAGT-3′;

• Mouse BK-β1 forward: 5′-TTCTGCACCTCAAGTCAACG-3′; reverse: 5′-TACTTCTGAGCCGCCAAGAT-3; and

• Mouse β-actin forward: 5′-TGTTACCAACTGGGACGACA-3′; reverse: 5′-CTGGGTCATCTTTTCACGGT-3′.

Mouse coronary arteries (<100 μm in diameter) were dissected and vasoreactivity was measured as we have previously described (10,11). Only the vessels that showed no leak with 85% relaxation in response to 0.1 mmol/L nitroprusside and >35% constriction in response to 60 mmol/L KCl were used for experiments. BK channel–mediated coronary vasodilation was determined by the application of 100 nmol/L IBTX and subtracting the IBTX-insensitive component from total vasodilation.

Dehydrosoyasaponin-1 (DHS-1) was provided by Merck Research Laboratory (Boston, MA). Unless noted otherwise, all chemicals including DMF were purchased from Sigma-Aldrich. DMF was dissolved in DMSO and diluted with water into a 10 mmol/L stock solution.

Data were expressed as the mean ± SEM. A Student t test was used to compare data between two groups. A paired t test was used to compare data before and after treatment. One-way ANOVA, followed by the Tukey test analysis was used to compare multiple groups using SigmaStat 3.5 software (Systat Software Inc., San Jose, CA). A statistically significant difference was defined as P < 0.05.

After 6 months of HFD consumption, mice had significant increases in body weights, blood pressures, and random glucose and serum insulin levels, compared with those in control mice eating an LFD (Table 1).

We first compared the BK-β1 subunit–induced channel activation using a specific membrane-impermeable BK-β1 activator, DHS-1, in the excised membrane patches from freshly isolated coronary SMCs of LFD and HFD mice (Fig. 1A). The total BK single-channel open probability (nPo) was significantly decreased in HFD mice at baseline (0.09 ± 0.03, n = 23 cells) compared with that of LFD mice (0.33 ± 0.11, n = 23 cells, P < 0.001). Cytoplasmic application of 100 nmol/L DHS-1 robustly increased nPo to 0.68 ± 0.14 (n = 23 cells, P < 0.001 vs. baseline) in LFD mice. In contrast, the DHS-1 effects were modest though significant in HFD mice (0.17 ± 0.07, n = 23 cells, P < 0.05 vs. baseline, P < 0.001 vs. LFD with DHS-1).

Similar to streptozotocin-induced type 1 diabetic mice (16), HFD mice showed a significant reduction in BK-β1 protein expression by 57.1% (Fig. 1B), no change in BK-β1 mRNA expression (Fig. 1C), a 1.6-fold increase in MuRF1 protein expression, and a 51.8% reduction in Nrf2 protein expression in the arterial vessels of HFD mice compared with those of LFD mice (n = 6 mice for each group, P < 0.05) (Fig. 1B).

The BK channel–mediated coronary vasodilation was examined using NS-1619 (a BK channel activator) in LFD and HFD mice. As shown in Fig. 1D, NS-1619 dose-dependently dilated the coronary arteries of LFD mice with a half-maximal effective concentration of 0.09 ± 0.01 µmol/L and an efficacy of 71.72 ± 1.95% (n = 4 mice) but had reduced potency (half-maximal effective concentration 0.59 ± 0.06 µmol/L) and efficacy (19.16 ± 0.35%) in HFD mice (n = 4 mice, P < 0.001 vs. LFD for both). These results indicate that BK channel–mediated coronary vasodilation in HFD mice is impaired.

We have identified that MuRF1 is an E3 ligase responsible for BK-β1 protein degradation in vascular SMCs (16). To further study the role of Nrf2 in the regulation of MuRF1-dependent BK-β1 protein expression, we manipulated the Nrf2 gene expression in cultured human coronary SMCs by adenoviral delivery (50 MOI) of the Nrf2 gene and Nrf2 shRNA. We found that a 6.8-fold increase in Nrf2 expression by Ad-Nrf2 transduction produced a 2.1-fold enhancement in HO-1 expression, a 66.2% reduction in MuRF1 expression, and a 2.3-fold argumentation in BK-β1 protein expression (Fig. 2A). In contrast, a 58.9% knockdown of Nrf2 expression by Ad-Nrf2 shRNA resulted in a 41.2% decrease in HO-1 expression, a 3.0-fold upregulation of MuRF1 expression, and a 68.9% downregulation of BK-β1 protein expression (n = 9 culture plates for all, P < 0.05 vs. controls) (Fig. 2B). These results indicate that MuRF1 and BK-β1 protein expressions are tightly regulated by the Nrf2 gene in coronary SMCs.

In addition, Ad-Nrf2 transduction significantly increased BK-β1 mRNA levels in cultured human and mouse coronary SMCs by 2.6-fold and 1.6-fold, respectively (n = 9 culture plates for both, P < 0.05 vs. Ad-GFP) (Fig. 2C), suggesting that the upregulation of BK-β1 protein expression by Nrf2 was also attributed to the enhanced mRNA transcription in vascular SMCs.

Figure 3A illustrates the protein expressions of NF-κB subunits in the arteries of LFD and HFD mice. There was a 1.7-fold increase in NF-κB/p65 protein level (n = 6 mice, P < 0.05 vs. LFD) but no change in that of NF-κB/p50 (n = 6 mice, P = NS vs. LFD) in HFD mice. To determine the effects of Nrf2 on NF-κB protein expression, we transduced Ad-Nrf2 in human coronary SMCs. After a 48-h transduction, the protein levels of NF-κB/p50, NF-κB/p65, and phospho-NF-κB/p65(S267) were reduced by 64.5%, 61.2%, and 58.0%, respectively (n = 8 culture plates) compared with cells transduced with Ad-GFP (n = 7 culture plates, P < 0.05). Hence, NF-κB protein expressions are profoundly regulated by Nrf2 in arterial SMCs.

We examined the effects of DMF (a U.S. Food and Drug Administration–approved Nrf2 activator) on the protein expressions of Nrf2, HO-1, NF-κB/p65, MuRF1, and BK-β1 in cultured human coronary SMCs. A 12-h incubation with 10 μmol/L DMF produced a 1.9-fold and 1.6-fold augmentation, respectively, in Nrf2 and HO-1 protein expressions in human coronary SMCs, accompanied by a 43.9% and 63.8% reduction in NF-κB/p65 and MuRF1 expressions and a threefold increase in BK-β1 protein levels (n = 6 culture plates for each group, P < 0.05 vs. controls) (Fig. 4A). Moreover, the BK-β1 mRNA expression was also upregulated by 15.3% in cells treated with DMF compared with cells treated with vehicle (n = 9 culture plates, P < 0.05) (Fig. 4B). These results indicate that the upregulation of BK-β1 protein expression by DMF is a result of increased mRNA transcription and attenuated protein degradation of BK-β1 in vascular SMCs.

We further investigated the effects of DMF in HFD mice in vivo. We found that a 10-day course of treatment with DMF by oral administration significantly reduced the body weight (51.1 ± 1.4 g with placebo vs. 38.6 ± 3.4 g with DMF, n = 10 for both groups, P < 0.05), and blood glucose levels (284.9 ± 16.0 mg/dL with placebo vs. 217.8 ± 17.9 mg/dL with DMF, n = 10 for both groups, P < 0.05) in HFD mice compared with those with placebo treatment. Interestingly, DMF had no significant effect on the body weights and blood glucose levels in LFD mice: 28.9 ± 0.6 g with placebo vs. 28.7 ± 0.5 g with DMF (n = 9 for both groups, P = NS); and 168.1 ± 11.3 mg/dL with placebo vs. 150.7 ± 7.6 mg/dL with DMF (n = 9 for both groups, P = NS), respectively.

DMF treatment markedly enhanced the protein expressions of Nrf2 and BK-β1 in the arteries of HFD mice by 1.5-fold and 1.8-fold, respectively, whereas those of NF-κB/p50, NF-κB/p65, and MuRF1 were significantly downregulated by 56.8%, 75.5%, and 66.3% compared with HFD mice that received placebo treatment (n = 5 mice for each group, P < 0.05) (Fig. 5).

Figure 6A shows the representative tracings and current-voltage curves of IBTX-sensitive K⁺ currents recorded from freshly isolated coronary SMCs of HFD mice treated with DMF or placebo. BK current density was significantly increased in DMF-treated mice. It was 45.1 ± 8.4 pA/pF (n = 8 cells) with placebo and 73.1 ± 13.9 pA/pF with DMF (n = 8 cells, P < 0.05 vs. placebo) at 100 mV, which was similar to that in LFD mice without DMF treatment. Coronary vasodilation response to NS-1619 was impaired in placebo-treated HFD mice (n = 8), but it was preserved by DMF administration in HFD mice to a similar level to that in LFD control mice (n = 8) (Fig. 6B).

To further determine whether the protective effects of DMF on BK channel activities were mediated through BK-β1 function, we examined coronary BK channel activities in freshly isolated coronary SMCs in response to DHS-1 in HFD mice after a 10-day course of treatment with DMF by gavage. Figure 6C illustrates the single BK currents elicited at 60 mV from coronary SMCs of HFD mice treated with placebo or DMF. DMF treatment markedly increased the channel nPo at baseline compared with placebo treatment (0.09 ± 0.02 with placebo, n = 14 cells vs. 0.57 ± 0.15 with DMF, n = 14 cells, P < 0.001). Application of DHS-1 further enhanced the channel openings (nPo = 1.02 ± 0.21, n = 13, P < 0.001 vs. baseline of DMF-treated cells). In comparison, DHS-1 had no effects in placebo-treated cells (nPo = 0.10 ± 0.03, n = 13 cells, P = NS vs. baseline; P < 0.05 vs. DMF-treated cells). Importantly, BK channel activities in HFD mice after DMF treatment were comparable to those in LFD mice before and after exposure to DHS-1. The nPo was 0.45 ± 0.18 at baseline in LFD cells (n = 14 cells, P = NS vs. the baseline of DMF-treated HFD cells) and 1.03 ± 0.18 after exposure to DHS-1 (n = 14 cells, P = NS vs. DMF-treated HFD cells with DHS-1). Hence, our results demonstrate that treatment with DMF restores coronary BK channel function and BK channel–mediated coronary vasoreactivity in HFD mice in vivo.

In this study, we found that a 6 months of consumption of an HFD increased mouse body weight and blood pressure, and serum glucose and insulin levels, which is consistent with previous reports (22–25). Since the average HbA_1c in patients with diabetes who have poor glycemic control in the U.S. is ∼9% (32), corresponding to a blood glucose level of ∼212 mg/dL, we believe that the HFD mice with a random average blood glucose level of 230 mg/dL is an excellent animal model for studying type 2 diabetes. We found that BK-β1 protein expression was significantly downregulated in HFD mouse arterial SMCs, resulting in impaired BK channel function. BK-β1 homeostasis is dependent on the balance between BK-β1 protein turnover and synthesis. It has been reported that decreased BK-β1 protein expression was associated with the nuclear factor of activated T-cell c3 (NFATc3)–dependent inhibition of mRNA transcription in isolated mesentery arteries of HFD mice cultured with 20 mmol/L glucose for 48 h (13). In this study, we found that such a reduction of BK-β1 protein expression in coronary arteries of HFD mice was mainly due to accelerated proteolysis through the UPS because the BK-β1 mRNA levels were not altered in coronary arteries of HFD mice, which is similar to our findings in mice with streptozotocin-induced type 1 diabetes (9,11,16). The difference in findings might be due to the difference in experimental conditions and vessel bed specificity.

Nrf2 signaling is a master regulator of cellular redox status and detoxification responses (33,34). The function of Nrf2 is principally regulated by the Kelch-like ECH-associated protein 1 (Keap1), an adaptor protein for Cullin 3–dependent ubiquitination and degradation (35,36). Under normal conditions, Nrf2 is inhibited by binding to Keap1. Increase in oxidative stress or electrophile stimuli modifies specific cysteine residues in Keap1 and releases Nrf2 from binding with Keap1. The unbound Nrf2 translocates into the nucleus and heterodimerizes with small Maf proteins to form a transactivation machinery that binds to the promoter region of antioxidant response elements and electrophile response elements, which contain a Nrf2-binding motif [TGA(G/C)xxxGC] and/or a Maf-binding motif [TGCTGA(G/C)] (37, 38). Clinical studies (39,40) have revealed that the reduction of Nrf2 expression contributes to oxidative stress in patients with chronic type 2 diabetes. One of our important findings in this study is that the regulation of vascular BK-β1 protein expression in HFD mice by Nrf2 was mediated through the NF-κB/MuRF1-dependent proteolysis. The protein expression of Nrf2 and BK-β1 was markedly decreased in HFD mice, whereas that of MuRF1 was increased. These reciprocal relationships were confirmed by manipulations of Nrf2 expression in cultured coronary arterial SMCs: adenoviral expression of Nrf2 gene profoundly upregulated the protein expression of BK-β1 and significantly downregulated that of MuRF1, whereas the knockdown of Nrf2 expression by Nrf2 shRNA had the opposite effects. However, the molecular mechanisms underlying Nrf2-regulating BK-β1 protein expression could be complicated. We found that Nrf2 overexpression also upregulated the BK-β1 mRNA expression in cultured coronary SMCs. It is not surprising since the KCNMB1 gene contains several consensus sequences of the Nrf2- and Maf-binding motifs in its promoter. Whether Nrf2 directly increases the KCNMB1 mRNA transcription or indirectly regulates other transcriptional factors that are responsible for BK-β1 transcription is currently unknown but warrants investigation. Nevertheless, an increase in BK-β1 mRNA expression and decrease in BK-β1 protein degradation by Nrf2 signaling would greatly modulate BK channel function in coronary SMCs.

The molecular mechanisms underlying the regulation of NF-κB expression by Nrf2 in diabetic vessels are currently unclear. It has been reported that the heme metabolites of HO-1, such as CO, bilirubin, and Fe²⁺, are the core of Nrf2-mediated NF-κB inhibition (41–43). On the other hand, NF-κB/p65 deprives CREBPs from Nrf2, resulting in the suppression of Nrf2 transcriptional activity (44). In addition, the phosphorylation of serine residues in NF-κB/p65 could directly promote NF-κB/p65 nuclear translocation (15). The interplay between the Nrf2 and NF-κB pathways was confirmed by increased protein expression and transcriptional activity of NF-κB/p65 in the hearts of Nrf2 KO animals (45). In this study, we found that the protein expressions of NF-κB/p65 and were upregulated and those of Nrf2 and HO-1 were downregulated in the arteries of HFD mice. Such changes in protein profiles are favorable for MuRF1-dependent BK-β1 protein degradation in HFD mice. We believe that ROS plays an important role in the regulation of Nrf2 and NF-κB activity. Decrease of Nrf2 activity promotes ROS generation and facilitates NF-κB/p65 protein phosphorylation, which in turn attenuates the effects of Nrf2 on BK-β1 transcription and the NF-κB/MuRF1-dependent BK-β1 protein degradation. Adenoviral expression of Nrf2 gene enhanced HO-1 protein expression and reduced NF-κB/p50 and NF-κB/p65 and phospho-NF-κB/p65(S267) protein expressions in human coronary SMCs. The scheme illustrating the role of Nrf2 signaling in the regulation of vascular BK-β1 expression is shown in Fig. 7.

It has been reported that DMF enhances Nrf2 transcriptional activity through covalently binding to the reactive cysteine residues of Keap1, thereby preventing the proteolytic degradation of Nrf2 in the UPS (46). Recently, DMF (Tecfidera) has been successfully used in the treatment of human relapsing multiple sclerosis with no major adverse events (47,48) but has not been applied to the treatment of diabetic vascular complications. We show that the treatment of human coronary SMCs and HFD mouse coronary arteries with this Food and Drug Administration–approved drug suppressed the NF-κB/MuRF1-induced BK-β1 protein degradation and enhanced BK-β1 mRNA transcription, in turn protecting coronary BK channel function and BK channel–mediated coronary vasodilation in HFD mice. Moreover, a 10-day course of treatment with DMF by oral administration resulted in significantly decreased body weights and serum glucose levels in HFD mice, whereas DMF had no such effects in LFD mice. These novel results provide the first evidence that Nrf2 activation is an exciting strategy for the treatment of BK channelopathy and vasculopathy in obese/diabetic animals.

Acknowledgments. The authors thank Merck Research Laboratory for providing DHS-1 for this work.

Funding. This study was supported by the National Institutes of Health and National Heart, Lung, and Blood Institute (grants HL-080118 and HL-74180); the American Diabetes Association (grants ADA-JFA-07-39, ADA 1-12-BS-119, and ADA 1-16-IBS-195); the Natural Science Foundation of Jiangsu Province (grant BK20140249 to Y.L.); the National Natural Science Foundation (grant 81470489 to Q.C.) of the People’s Republic of China; a Prospective Research Award from the Department of Cardiovascular Diseases; and a Pilot Grant from the Center for Biomedical Discovery, Mayo Clinic, Rochester, MN.

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

Author Contributions. T.L. conducted experiments and analyzed data and wrote and edited the manuscript. X.S., Y.L., Q.C., and X.-L.W. conducted experiments and analyzed data. H.-C.L. wrote and edited the manuscript. T.L. and H.-C.L. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 76th Scientific Sessions of the American Diabetes Association, New Orleans, LA, 10–14 June 2016.