Protective role of NRF2 in macrovascular complications of diabetes

Abstract Macrovascular complications develop in over a half of the diabetic individuals, resulting in high morbidity and mortality. This poses a severe threat to public health and a heavy burden to social economy. It is therefore important to develop effective approaches to prevent or slow down the pathogenesis and progression of macrovascular complications of diabetes (MCD). Oxidative stress is a major contributor to MCD. Nuclear factor (erythroid‐derived 2)‐like 2 (NRF2) governs cellular antioxidant defence system by activating the transcription of various antioxidant genes, combating diabetes‐induced oxidative stress. Accumulating experimental evidence has demonstrated that NRF2 activation protects against MCD. Structural inhibition of Kelch‐like ECH‐associated protein 1 (KEAP1) is a canonical way to activate NRF2. More recently, novel approaches, such as activation of the Nfe2l2 gene transcription, decreasing KEAP1 protein level by microRNA‐induced degradation of Keap1 mRNA, prevention of proteasomal degradation of NRF2 protein and modulation of other upstream regulators of NRF2, have emerged in prevention of MCD. This review provides a brief introduction of the pathophysiology of MCD and the role of oxidative stress in the pathogenesis of MCD. By reviewing previous work on the activation of NRF2 in MCD, we summarize strategies to activate NRF2, providing clues for future intervention of MCD. Controversies over NRF2 activation and future perspectives are also provided in this review.


| INTRODUC TI ON
Diabetes is predicted to be the seventh leading cause of death in the world in 2030. 1,2 Macrovascular complications of diabetes (MCD)-basically ischaemic heart disease, cerebrovascular disease and peripheral vascular disease-develop in over a half of the diabetic population, resulting in high morbidity and mortality. 3 This poses a severe threat to the world's public health. Despite the successful control of hyperglycaemia, hypertension and hyperlipidaemia, the diabetic patients are still at risk of developing MCD. 4 Therefore, it is crucial to identify more viable drug targets and develop more effective approaches, in order to prevent or slow down the pathogenesis and progression of MCD.
Oxidative stress is a key mechanism by which diabetes induces its complications. 5 Under diabetic condition, excessive reactive oxygen species (ROS) are produced, causing detrimental cellular events, such as formation of advanced glycation end products (AGEs) and overexpression of receptor for AGEs (RAGE), along with activation of polyol pathway, hexosamine pathway and protein kinase C (PKC). 5,6 These contribute to the pathogenesis and progression of MCD.
The transcription factor nuclear factor (erythroid-derived 2)like 2 (NRF2) plays a critical role in cellular defence against oxidative stress. 7 NRF2 turns on the transcription of various antioxidant genes, producing cellular antioxidants 7 that act as scavengers of free radicals and prevent the oxidative stress-driven pathogenesis of MCD. NRF2 is negatively regulated by Kelch-like ECH-associated protein 1 (KEAP1) in the cytoplasm. KEAP1 restricts NRF2 from nuclear translocation on one hand and facilitates proteasomal degradation of NRF2 on the other. 8 Therefore, small molecule-induced structural inhibition of KEAP1 protein-a canonical way to activate NRF2-has become a research hotspot in the past two decades, with the protective outcome verified in animal models of MCD.
In addition to structural inhibition of KEAP1 protein, activation of NRF2 in MCD has been achieved through other approaches. In this review, we summarize and discuss the various strategies arisen to activate NRF2 and their outcome in MCD, with the aim of providing insights into future management of MCD.

| PATHOPHYS I OLOGY OF M CD
The diabetes-driven atherosclerosis is the main cause of MCD.
The typical pathological features of atherosclerosis in arteries are characterized by generation of fibrosis, proliferation and structural derangement of smooth muscle cells, thickening of tunica media, accumulation of lipid, formation of plaque, calcification and thrombosis. 9 Under microscope, adhesion of blood cells to the endothelium and extension of interendothelial connection may be observed, indicating the diabetes-enhanced adhesion and permeability of the endothelium. 10 The arterial pathophysiological events induced by diabetes ( Figure 1) eventually lead to vascular dysfunction and ischaemic complications.

| ROLE OF OXIDATIVE S TRE SS IN THE PATHOG ENE S IS OF M CD
Oxidative stress reflects an imbalance between the status of ROS and the antioxidant ability of a biological system. Diabetes induces generation of ROS through formation of AGEs, overexpression of RAGE and activation of polyol pathway, hexosamine pathway and PKC. 5 Upon diabetes, the excessive ROS exceed the scavenging capacity of the cellular antioxidant system, resulting in damage to proteins, lipids and DNAs. 11 Moreover, the diabetes-induced ROS provokes inflammation, which in turn exacerbates oxidative stress. This vicious circle formed by ROS and inflammation 12 contributes to fibrosis and calcification in the plaque, 13,14 during a later stage of atherosclerosis.
Endothelial dysfunction is a critical pathophysiological event prior to MCD, with oxidative stress and inflammation as major contributors ( Figure 1). 15 Once formed, typical pathological features of atherosclerosis-such as fibrosis, tunica media thickening, plaque and calcification-are impossible to be reversed. It is therefore crucial to improve diabetes-induced endothelial dysfunction, the effect of which may efficiently prevent or slow down atherogenesis. In this regard, targeting oxidative stress is a viable strategy.

| PROTEC TIVE EFFEC TS OF NRF2 ON MCD
As supplementation of antioxidants, such as vitamin E, vitamin C, coenzyme Q10, alpha-lipoic acid, L-carnitine and ruboxistaurin, has proven non-beneficial to diabetic complications, 16 attention has been transferred to activation of the inner cellular antioxidant capacity. NRF2-the governor of the cellular antioxidant defence system-activates the transcription of downstream antioxidant genes by binding antioxidant response element in the promoter regions of the antioxidant genes, such as haem oxygenase 1 (Hmox1), NAD(P) H dehydrogenase (quinone 1) (Nqo1), glutathione (Gsh), superoxide dismutase (Sod), gamma-glutamylcysteine synthetase and catalase (Cat). [17][18][19] K E Y W O R D S complications, diabetes, macrovascular, NRF2, oxidative stress NRF2 is negatively regulated by KEAP1. 7,8,20,21 Oxidative stress disrupts critical cysteine residues in KEAP1, resulting in the release of NRF2 from the KEAP1-NRF2 complex. 22 Hence, a short-term oxidative stress generated upon diabetes activates NRF2 as a compensatory protective mechanism through which the cells protect against hyperglycaemia-induced injuries. 23 However, accumulating evidence has demonstrated that cardiovascular NRF2 antioxidant signalling is impaired after a long-term exposure to hyperglycaemia. [24][25][26][27][28][29] This decompensatory effect exposes the cells to more severe injuries. Thus, activation of NRF2 is beneficial to the vasculature under both shortterm and long-term diabetic conditions prior to atherosclerosis, especially the latter.
To date, many studies have reported the protective effects of NRF2 activation on diabetic complications, such as diabetic cardiomyopathy, nephropathy and retinopathy. However, less is known for the impact of NRF2 activation on MCD, as summarized below.

| Sulforaphane (SFN)
SFN derives from broccoli sprouts 30 and modifies specific cysteine residues in KEAP1 protein, thereby changing its conformation.
This enables NRF2 to dissociate from KEAP1, promoting NRF2 nuclear translocation and preventing proteasomal degradation of NRF2. 31 The beneficial effects of SFN on MCD have been reported. 32 In a mouse model of type 2 diabetes, SFN activated NRF2 antioxidant signalling in the aorta and attenuated the diabetes-induced oxidative stress, inflammation, apoptosis, cell proliferation, thickening of the tunica media and accumulation of collagen in the aorta. 32 In an animal model of non-obese type 2 diabetes (Goto-Kakizaki rats), SFN reversed the diabetes-repressed expression of aortic NRF2, attenuated the production of aortic O 2 and AGEs and improved nitric oxide (NO)dependent and nitric oxide nitric oxide-independent vasorelaxation. 29 Moreover, SFN inhibits nuclear factor-kappa B (NF-κB) 33 -a key pro-inflammatory factor in diabetes-induced vascular inflammation.
Crosstalk may exist between NRF2 and NF-κB. HO-1, a potent antioxidant downstream of NRF2, has been shown to repress NF-κB activity by reducing cellular labile iron content. 34 NF-κB, in turn, suppresses NRF2-induced HO1 production. 35 The predominant activity of NF-κB under diabetes might explain the impaired NRF2 antioxidant activity in the vasculature under a long-term diabetic condition. Given that NF-κB is a transcription factor, it would be interesting to explore the effect of NF-κB on the expression of Keap1 in further studies.
The finding that SFN improves insulin resistance in patients with type 2 diabetes 36 suggests that SFN may prevent MCD at much earlier stages including diabetes and obesity.

F I G U R E 1
Pathophysiology of macrovascular complications of diabetes (MCD). Hyperglycaemia causes formation of reactive oxygen species (ROS) and inflammation in the endothelium, leading to endothelial dysfunction as a critical first step towards MCD. Under diabetic condition, the permeability of the inflamed endothelium is increased, allowing recruitment of neutrophils and monocytes into the tunica intima, where the macrophages-differentiated from monocytes-engulf lipids and become foam cells that gradually form a plaque. Calcium can be deposited in the plaque. Thrombus forms at the location where the plaque breaks. Apoptosis is induced, and fibrosis is accumulated.
Smooth muscle cells proliferate, thereby thickening the tunica media. Smooth muscle cells can migrate into tunica intima through damaged internal elastic membrane, contributing to atherogenesis. Red characters, detrimental processes resulting in atherosclerosis

| Dh404
Dh404 is a derivative of bardoxolone methyl. The latter was tested in clinical trials for treatment of diabetic nephropathy (DN). 37 Dh404 activates NRF2 via modification of KEAP1, 38

| Dimethyl fumarate (DMF, BG-12)
DMF is a known NRF2 activator and has been used for clinical treatment of multiple sclerosis. 39  Moreover, this effect of DMF was abolished when NRF2 was knocked down in VSMCs. 40 This study may indicate a potential protective effect of DMF on MCD. The approval of DMF in clinical use has granted DMF a unique advantage in its potential application in MCD.

| Tert-butyl hydroquinone (tBHQ)
tBHQ activates NRF2 by targeting Cys-151 within KEAP1 protein. 41 tBHQ was reported to ameliorate diabetes-driven atherosclerosis in apolipoprotein E-deficient mice. 42 In this study, tBHQ was found to enhance NRF2 activity in macrophages and VSMCs within atherosclerotic lesions, promoting autophagic activity. This led to decrease in size, extension and lipid content of the atheroma plaques, as well as reduction of lesional macrophages, foam cell size and chemokine expression.

| Non-canonical ways to activate NRF2 in MCD
In recent years, strategies towards NRF2 activation in MCD have evolved from structural inhibition of KEAP1 to other regulatory mechanisms, shedding light upon novel targets and approaches for the intervention of MCD.

MicroRNA-200a (miR-200a)/Keap1 mRNA
We and others have reported that miR-200a targets Keap1 mRNA, leading to degradation of Keap1 mRNA and activation of NRF2. 24,43,44 Recently, our group found that the miR-200a/Keap1/ NRF2 axis played an essential role in protecting against diabetesinduced endothelial dysfunction. 45 In aortic ECs isolated from MiR-200a/KEAP1/NRF2 not only provides direct protection to the endothelium, but also benefits the endothelium through a remote route by its regulation of fibroblast growth factor 21 (FGF21).
FGF21 regulates glucose and lipid metabolism and has potential for the treatment of diabetes. 46 FGF21 was shown to prevent aortic pathologies in STZ-induced diabetic mice. 46 In OVE26 type 1 diabetic mice, Fgf21 mRNA and protein levels were found to be increased in the liver and plasma, leading to protection against the diabetes-induced aortic fibrosis and inflammation. 47 Further investigation showed that inhibition of HDAC3 with RGFP-966 enhanced hepatic miR-200a expression, the effect of which decreased Keap1 mRNA and protein levels, thereby promoting NRF2-mediated Fgf21 gene expression. FGF21, produced in the liver, was then released to the plasma and protected the diabetic aorta. 47

| Modulation of c-Jun N-terminal kinase (JNK)
JNK has been shown to regulate vascular NRF2 expression and function upon diabetes. However, the role of JNK in control of On the contrary, JNK was found by other groups to negatively regulate NRF2. C66-a novel curcumin analogue-has shown NRF2activating efficacy in both the aortas 51 and kidneys 24 of STZ-induced diabetic mice. C66 was primarily identified to be an inhibitor of JNK. 51 Inhibition of JNK by either C66 or SP600125 activated NRF2 expression and function and attenuated the diabetes-induced aortic pathological injuries to a similar extent, 51 suggesting that JNK negatively regulates NRF2 in the aorta. Moreover, a recent study showed that C66 predominantly targeted JNK2, as both Jnk2 gene deletion and C66 treatment could similarly activate NRF2 and alleviate diabetes-induced aortic oxidative stress, inflammation and fibrosis. 43 Supporting these findings, our group found that SP600125 prevented DN through activation of NRF2, with enhanced expression of Hmox1 and Nqo1. 52 SP600125 hampered the HG-induced phosphorylation of JNK and c-Jun, repressing Keap1 gene expression. 52 This inhibitory impact of SP600125 on KEAP1 was viable because SP600125 induced remarkable nuclear translocation of NRF2. 52 Supporting the inhibitory effect of SP600125 on Keap1 gene expression, 36 c-Jun-binding sites were found between −3000 bp and −1 bp within the promoter region of the mouse Keap1 gene. 52 Luciferase reporter assay should be helpful to further confirm the regulatory effect of JNK on Keap1 gene transcription.
As reviewed above, controversies exist in defining the role of JNK in NRF2 activity in the vasculature. These controversies may be owing to the differences between cell types (BACEs, HUVECs, whole aortas and mouse mesangial cells) and (or) disease models (stimulation with AGEs vs HG/hyperglycaemia). Further studies are needed to elucidate the regulatory effect of JNK on NRF2 in MCD.

| Transcriptional activation of Nfe2l2
Recently, NRF2 activation has been achieved in diabetes-induced endothelial dysfunction at the transcription level. Sodium butyrate (NaB), a polyunsaturated fatty acid found in food, was primarily reported by our group to activate renal Nfe2l2 gene expression in an STZ-induced type 1 diabetic mice. 53 We further found that NaB improved the diabetes-induced aortic endothelial dysfunction in the wild-type, but not Nfe2l2 gene knockout, mice. 28 Mechanistically, in HG-treated aortic endothelial cells, NaB elevated Nfe2l2 mRNA and protein levels without facilitating NRF2 nuclear translocation, an effect distinct from that of SFN. Further, NaB inhibited the activity of histone deacetylase (HDAC) and increased the occupancy of the transcription factor aryl hydrocarbon receptor and the co-activator P300 at the Nfe2l2 gene promoter. Moreover, the P300 inhibitor C646 completely abolished these efficacies of NaB. 28 These results suggest that NaB prevented diabetes-induced aortic endothelial dysfunction through HDAC/P300-mediated transcriptional activation of Nfe2l2.

| Preservation of NRF2 protein
Preservation of NRF2 protein from degradation is another efficacious way to maintain cellular NRF2 protein level, facilitating NRF2 activation. MG132-a proteasome inhibitor-preserved NRF2 protein and attenuated aortic pathological injuries in OVE26 diabetic mice. 54 MG132, administered 3 months starting from the age of 3 months old, was shown in this study to lower the diabetes-provoked aortic levels of tumour necrosis factor-alpha (TNF-α), plasminogen activator inhibitor-1, 3-nitrotyrosine and 4-hydroxynonenal. Moreover, MG132 ameliorated the diabetes-induced thickening and structural derangement of the aortic wall. 54 It is noted that the pathological aortic changes at the initiation of the treatment (3-month-old OVE diabetic mice) were not evident. MG132 should have yielded greater beneficial effects if administered at earlier stages of diabetes, since once formed, it is impossible to reverse the diabetes-induced typical pathological features. 55

Zinc (Zn)
Zn is an essential trace element that has antioxidant activity. Zn deficiency in endothelial cells enhanced inflammatory response and oxidative stress. 56,57 On the contrary, Zn supplementation benefitted the aortas of 3-month-old OVE26 diabetic mice, 58 including the attenuation of aortic fibrosis, oxidative stress, inflammation, apoptosis and proliferation. 58 Moreover, Zn elevated both mRNA and protein levels of Nfe2l2, 58 suggesting that Zn might regulate Nfe2l2 gene expression at the transcription or post-transcription levels.
The same group performed a further study to investigate the effect of Zn supplementation on DN. In OVE26 type 1 diabetic mice, Zn induced renal AKT and glycogen synthase kinase-3 beta (GSK-3β) phosphorylation with a decrease in proto-oncogene tyrosine-protein kinase Fyn (Fyn), a nuclear exporter of NRF2. 59

Baicalin
Baicalin is the major component found in Scutellaria baicalensis root, exhibiting anti-inflammatory activity. 66 Baicalin was reported to restore the activity of hyperglycaemia-impaired aortic NRF2 signalling in STZ-induced diabetic mice. 66

| Other mechanisms to activate vascular NRF2 antioxidant signalling
Several other mechanisms have been studied.

| Insulin-like growth factor 1 (IGF-1)
In a mouse model of vascular ageing, liver-specific knockdown ofIGF-1 decreased vascular oxidative stress resistance by impairing NRF2 signalling. 67 In the aortas of IGF-1-deficient mice, the expression of Nfe2l2 and downstream glutamate-cysteine ligase catalytic subunit, Nqo1 and Hmox1 was decreased. 67 When challenged with HG, the expression of these NRF2 downstream genes was activated in the wild-type, but not IGF-1-deficient, aortas. 67 This study indicates that IGF-1 positively regulates NRF2. IGF-1 activation or overexpression in animal or cell models of MCD could be helpful to confirm IGF-1 as an activator of NRF2.

| HO1
HO1 is a known downstream antioxidant of NRF2 and has shown beneficial effects on endothelial cells and animal models of vascular disease. 68

| NRF2-AC TIVATING S TR ATEG IE S BORROWED FROM D IAB E TE S AND C ARDIOVA SCUL AR COMPLIC ATIONS OTHER THAN M CD
NRF2 activation in diabetes and cardiovascular complications other than MCD, such as DN, cardiomyopathy and retinopathy, has implications on MCD.

SIRT1 is an HDAC and benefits diabetes and complications. 71
Antioxidant efficacy contributes to SIRT1's protective effects. 71,72 In a mouse model of STZ-induced diabetes, SIRT1 and NRF2 anti- P53 is a transcription factor that activates the expression of miR-34a which targets Sirt1 mRNA. 75 By inhibiting SIRT1 in the presence of either the P53 inhibitor pifithrin-α or the miR-34a inhibitor, our other work found that inhibition of P53/miR-34a improved diabetes-induced endothelial dysfunction through activation of SIRT1. 75 Taken together, these studies have found a SIRT1/P53/miR-34a circuit that acts upstream of NRF2 antioxidant signalling and controls endothelial function under diabetes.

| Mouse double minute 2 (MDM2)/P53
MDM2 is a suppressor of P53. 77 As P53 is primarily known as a guardian of the genome coordinating cellular responses to genotoxic stress, 78 MDM2 is considered as an oncogene in tumours. 77 However, under diabetic conditions, MDM2/P53 shows a different profile from that in carcinogenesis. We and others have found that P53 is elevated and activated upon diabetes, contributing to endothelial dysfunction, 75,79 nephropathy 26,80 and cardiomyopathy. 81 Moreover, renal Mdm2 expression is inhibited in STZ-induced diabetic mice. 26 Interestingly, inhibition of MDM2 by nutlin3a activated P53 and generated DN-like pathologies in the non-diabetic healthy mice. 26 Notably, inhibition of P53 by pifithrin-α activated renal NRF2 signalling and attenuated renal injuries in the wild-type, but not Nfe2l2 gene knockout, diabetic mice. 26 In HG-treated mouse mesangial cells, P53 gene silencing completely abolished nutlin3a's inhibitory effect on NRF2 signalling. Together, these findings demonstrate that MDM2 controls NRF2 antioxidant signalling via inhibition of P53 in DN, providing MDM2 as a novel target.

Kim et al investigated whether NRF2 is involved in the protective
effect of GLP-1 on oxidative stress-induced beta-cell apoptosis. 85 In this study, NRF2 antioxidant signalling was activated by exendin-4 (EX4)-an agonist of GLP-1-in palmitic acid-or hydrogen peroxidestimulated beta cells. 85 NRF2 was required for the protective effects of EX4 on ROS production and insulin secretion, as these effects were abolished when Nfe2l2 was knocked down. 85 Mechanistic study revealed that EX4 phosphorylated PKC-δ. By using PKC-δ siRNA in the presence of EX4, PKC-δ was found to mediate the EX4induced activation of NRF2. Hence, PKC-δ might be a positive regulator upstream of NRF2 in beta cells under oxidative stress.

| PKR-like ER kinase (PERK) and thrombospondin 1
Endoplasmic reticulum (ER) stress and oxidative stress are important mediators of beta-cell failure in diabetes. 82

| DJ-1 (Parkinson disease protein 7, PARK7)
DJ-1, also known as PARK7, stabilizes NRF2 by preventing association of NRF2 with KEAP1. 88 In support of the positive regulatory effect of DJ-1 on NRF2, a recent study found that the protein levels of DJ-1 and NRF2/HO1 were increased in the kidneys of STZ-induced Sprague Dawley diabetic (4, 8 and 12 weeks of diabetes) rats. 89 The role of DJ-1 in the regulation of NRF2 antioxidant signalling in MCD warrants further investigation.
In summary, several targets identified in diabetes and its vascular complications other than MCD might shed light on future studies for MCD, as summarized in Table 2.

| NRF2 AC TIVATOR S IN CLINI C AL TRIAL S FOR VA SCUL AR COMPLI C ATI ON S OF D IAB E TE S
Although no NRF2 activator has been applied in clinical trials for MCD, the clinical trials of NRF2 activators used in diabetic complications other than MCD may provide clues for the future trials of MCD.

| Bardoxolone methyl in treatment of DN
The most well-known NRF2 activator tested in clinical trials for the treatment of diabetic complications is bardoxolone methyl. The initial application of bardoxolone methyl in treatment of type 2 diabetic patients with 3b-4 chronic kidney disease (CKD) yielded promising beneficial outcome. 90 The 20 patients with moderate-to-severe DN received bardoxolone methyl at 25 mg/d for 28 days, followed by a dose of 75 mg/d for additional 28 days. 90 By the end of the study, bardoxolone methyl dose-and time-dependently elevated estimated glomerular filtration rate and significantly reduced serum creatinine and blood urea nitrogen, accompanied by an increase in creatinine clearance. 90 Additionally, bardoxolone methyl blunted the biomarkers of vascular injury and inflammation. 90 It is noted that no severe side effect was observed in all the patients. 90 The success in the initial clinical trial of bardoxolone methyl brought light to the management of DN. Bardoxolone methyl was applied in a much larger phase 3 clinical trial in type 2 diabetic patients with stage 4 CKD thereafter. 91 However, the trial was terminated because of severe heart complications. 92 It is speculated that bardoxolone methyl may pharmacologically induce acute sodium and fluid retention, and therefore increase blood pressure and result in heart failure in patients with advanced CKD, despite its improvement of kidney injury in some of the patients. 92,93 Thus, attention should be paid to the off-target effects of a therapeutic in addition to its dose. Moreover, criticism has been raised regarding the use of the NRF2-activating drug at an inappropriate stage of the disease. 55,94 Application of the antioxidant approach should be more beneficial are impossible to be reversed, 55,94 whereas prevention of DN by NRF2 activators starting from the initiation of diabetes has proven successful in numerous experimental studies. 24,53,70,[94][95][96] The benefits earned and lessons learned from the use of bardoxolone methyl in the clinical trials of DN have a positive impact on the potential application of NRF2-activating approaches in future clinical trials of MCD, in which special attention should be paid to off-target effects and occasion for intervention.

| Curcumin in treatment of diabetic microangiopathy
Curcumin is a natural compound that derives from turmeric. 97 24 Therefore, C66 has a potential for clinical use in the management of MCD, based on its advantage in bioavailability and the solid preliminary work.

| CON CLUS I ON S AND PER S PEC TIVE S
In recent years, the strategies for NRF2 activation have evolved from the canonical structural inhibition of KEAP1 protein to others such as microRNA-induced inhibition of KEPA1 production, inhibition of proteasomal degradation of NRF2, inhibition of AKT/ GSK-3β/Fyn-mediated nuclear export of NRF2 and modulation of HDAC/P300-controlled Nfe2l2 gene transcription (Figure 2, red characters). In these novel strategies for NRF2 activation, epigenetic mechanisms such as microRNAs 45,48 and histone modifi- Taken together, NRF2 activation has a good potential in future clinical intervention of MCD. However, to date, the molecular mechanisms of NRF2 activation have not been fully understood. Further mechanistic studies are warranted. In addition, more novel and effective approaches should be developed as these will benefit the population with diabetes and MCD.

ACK N OWLED G EM ENTS
This work was supported in part by National Natural Science

CO N FLI C T S O F I NTE R E S T
None.