Hydrogen Sulfide Induces Keap1 S-sulfhydration and Suppresses Diabetes-Accelerated Atherosclerosis via Nrf2 Activation

Diabetes leads to a marked increase in atherosclerosis (1). There is considerable evidence demonstrating that oxidative stress and inflammation are involved in the pathogenesis of diabetes and its complications, including atherosclerosis (2). It has been suggested that hyperglycemia-induced superoxide overproduction may be a key event in the activation of pathways involved in the pathogenesis of diabetic complications (2). Approaches that limit oxidative stress may therefore translate to reduced inflammation and hence atherosclerosis.

It is well established that nuclear factor erythroid 2–related factor 2 (Nrf2) is one of the most important cellular defense mechanisms against oxidative stress. Nrf2 is broadly expressed in tissues but is only activated in response to a range of oxidative and electrophilic stimuli (3). Upon oxidative stress, Nrf2 escapes Kelch-like ECH-associated protein 1 (Keap1)–mediated repression, translocates to the nucleus, binds to antioxidant response element, and induces the expression of a battery of antioxidant proteins, one of the most important of which is heme oxygenase 1 (HO-1) (4). Nrf2 has emerged as an important target in diabetes and related complications (5,6), and low-dose dh404, which is an analog of the Nrf2 agonist bardoxolone methyl, lowers oxidative stress and protects against diabetes-associated atherosclerosis (7). These studies suggest that augmentation of antioxidant defenses via upregulation of the Nrf2 pathway may be a novel target for the prevention and treatment of diabetic complications.

Hydrogen sulfide (H₂S) plays an important role in physiology and pathophysiology in several biological systems. Emerging data suggest that H₂S improves diabetic endothelial dysfunction (8), nephropathy (9), retinopathy (10), and cardiomyopathy (11). However, there are no published data on the potential effect of H₂S on accelerated atherosclerosis in diabetes.

Some recent studies indicate that H₂S is cytoprotective during myocardial ischemia-reperfusion injury in the setting of diabetes by alleviating oxidative stress, and the ability of H₂S to upregulate cellular antioxidants in the heart in a Nrf2-dependent manner (12–14). H₂S may therefore play an important role in diabetes-accelerated atherosclerosis, and the effects of H₂S may be mediated via activation of Nrf2. In the current study, we have characterized whether and how H₂S targets on Nrf2 against the development of diabetes-accelerated atherosclerosis.

LDLr^−/− mice, on a C57BL/6 background, were purchased from the Model Animal Research Center of Nanjing University. Nrf2^−/− mice, on a C57BL/6 background, were a gift from Hongliang Li (Renmin Hospital of Wuhan University). Nrf2^−/− mice were crossed with LDLr^−/− mice to obtain LDLr^−/−Nrf2^−/−. At 8 weeks of age, male mice were rendered diabetic by administering 60 mg/kg/day streptozotocin (STZ) intraperitoneally daily for 5 days. After STZ administration, diabetic mice were administered the H₂S donor GYY4137 (133 μmol/kg/day, i.p.) or vehicle and kept on a high-fat diet (HFD) for 4 weeks. The dose of GYY4137 used was based on previous publications (15). Nondiabetic LDLr^−/− or LDLr^−/−Nrf2^−/− mice were kept on a standard chow diet for 4 weeks as control. Mice were housed (n = 1/cage; n = 6/group) in metabolic cages for 48 h prior to metabolic analysis to acclimate. Body weight, food intake, water intake, and urinary output were determined.

All animal experiments were approved by the Committee on Animal Care of Nanjing Medical University and were conducted according to the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals. All studies involving animals are reported in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines.

Plasma samples were obtained from mice fasted for 6 h. Glucose was measured directly from the tail tip with a glucometer; plasma lipid levels were measured enzymatically using commercial kits (Zhong Sheng Bei Kong, Peking, China) according to the manufacturer’s instructions.

Plasma H₂S concentration was measured as described previously (15).

To evaluate atherosclerotic lesions, en face whole and histological sections were used for analysis. The entire aorta attached to the heart was dissected and stained with Oil Red O (ORO; Sigma-Aldrich, St. Louis, MO) (16). Lesions within the sinus were visualized after staining with ORO as well as hematoxylin-eosin (H-E) and quantitated as described previously (16).

Mouse peritoneal macrophages were isolated from male C57BL/6 or Nrf2^−/− mice as described previously (17). Peritoneal cells were collected by lavage and seeded in DMEM–low glucose medium (Gibco, Grand Island, NY) with 10% FBS (Gibco).

Human umbilical vein endothelial cells (HUVECs) were isolated from umbilical cords according to a previously described method (18). The endothelial cells were cultured in Endothelial Cell Medium (ScienCell, Carlsbad, CA).

EA.hy926 endothelial cells were purchased from American Type Culture Collection (Rockville, MD) and were cultured in DMEM–low glucose medium with 10% FBS.

Confluent cells (80–85%) were incubated with d-glucose (25 mmol/L; Sigma-Aldrich) plus oxidized LDL (ox-LDL; 50 mg/L; Yiyuan Biotechnology, Guangzhou, China), in the absence or presence of GYY4137 (50 or 100 μmol/L), sulfide-depleted GYY4137 (SDG; dissolved GYY4137 left unstoppered at room temperature for 5 days), or ZYJ1122 (a structural analog of GYY4137 lacking sulfur) for 24 h.

EA.hy926 cells were transfected with small interfering RNA (siRNA) oligonucleotide against Nrf2 (sense: 5′AAGAGUAUGAGCUGGAAAAACdTdT-3′, antisense: 5′GUUUUUCCAGCUCAUACAUUCdTdT-3′; Genepharma) or negative control siRNA. HO-1 expression was silenced by HO-1 siRNA mix that was purchased from Santa Cruz Biotechnology. The plasmid pcDNA3-flag-Keap1 purchased from Addgene (Cambridge, MA) was termed as Keap1-WT. Single mutation at Cys151, -273, or -288 to Ala (Haibio, Shanghai, China) was confirmed by DNA sequencing. EA.hy926 cells were transfected with expression vectors using the Lipofectamine 3000 reagent (Invitrogen).

Macrophages were fixed with 4% paraformaldehyde and stained using 0.5% ORO. Images of cells were acquired using a light microscope (Nikon, Tokyo, Japan).

Superoxide production in tissue sections of upper descending thoracic aorta and cells was detected by dihydroethidium (DHE) assay according to the manufacturer’s instructions. In brief, cells or tissue were incubated with DHE for 30 min. Fluorescence was measured with a Nikon TE2000 inverted microscope and quantified using Image-Pro Plus analysis software.

Sections or cells were fixed and permeabilized and then blocked and incubated with antibody against Nrf2, vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1), CD31, α-SMA, or macrophage (Abcam, Cambridge, MA). After additional washing, sections or cells were incubated with directly conjugated fluorescent secondary antibodies and DAPI (Invitrogen). Fluorescence was imaged using a Nikon TE2000 inverted microscope. Positive cells and total cells were quantified in five different sections from six different mice of each genotype, using Image-Pro Plus analysis software.

HO-1, ICAM-1, VCAM-1, thioredoxin (Trx), and glutamate cysteine ligase catalytic subunit (GCLC) mRNA expression was quantified by real-time PCR (RT-PCR) with forward and reverse primers (Supplementary Table 1).

Whole-cell, cytosolic, and nuclear proteins were extracted using RIPA buffer (Sigma-Aldrich) or a nuclear and cytoplasmic extraction kit (Thermo Fisher Scientific). Western blotting was performed as described previously (19). Primary antibodies included anti-Nrf2 (Santa Cruz Biotechnology, CA), anti-Nrf2 (phospho S40; Abcam), anti–HO-1 (Bioworld Technology Inc., Nanjing, China), anti–VCAM-1 (Abcam), anti–ICAM-1 (Abcam), anti–histone H3 (Bioworld Technology, Inc.), and anti-GAPDH (Abcam). Band intensities were analyzed using ImageJ 1.25 software.

The cells were harvested and lysed as previously described (20). Antibodies specific to Keap1 (Santa Cruz Biotechnology) or normal rabbit IgG were added to the supernatants followed by an incubation. Immune complexes were then precipitated with protein A agarose beads. Bound proteins were eluted by boiling with loading buffer and analyzed by Western blotting with anti-Nrf2 antibody.

Nrf2 and Keap1 S-sulfhydration was detected with the “tag-switch” method (21). The protein of Keap1 was pulled down with immunoprecipitation and treated with biotin-linked cyanoacetate. Samples were resuspended in Laemmli buffer, heated, and subjected to Western blotting analysis using anti-biotin antibody (Santa Cruz Biotechnology).

Data are expressed as mean ± SEM and were analyzed by one-way ANOVA followed by Newman-Keuls multiple comparison test as appropriate. All statistical analyses were performed using SPSS software, version 16.0. A value of P < 0.05 was considered statistically significant.

As expected, diabetic LDLr^−/− mice fed an HFD had lower body weight, higher plasma total cholesterol, triacylglycerol, urinary output, water intake, and food intake when compared with those fed standard chow, and these effects were unaffected by treatment with GYY4137 (Supplementary Table 2). Compared with LDLr^−/− mice, plasma H₂S concentration was reduced in HFD-fed diabetic LDLr^−/− mice, which could be significantly increased by administration of GYY4137 (Supplementary Table 2).

To determine the effect of H₂S on the formation of atherosclerotic lesions in STZ-diabetic LDLr^−/− mice, these animals were treated with GYY4137 or vehicle for 4 weeks, and an HFD was used to enhance atherogenesis (Fig. 1A). Initially, we measured the total aortic lesion area between the proximal ascending aorta and the bifurcation of the iliac artery by en face analysis of ORO-stained aortas. As expected, diabetic mice showed an increase in atherosclerotic plaques compared with the nondiabetic control. Treatment of diabetic LDLr^−/− mice with H₂S reduced lesion area (Fig. 1B and Supplementary Fig. 1). Similar results were confirmed by H-E and ORO staining in the aortic root (Fig. 1C and D). Immunofluorescence analysis of sections from the aortic root revealed that macrophage content was increased in diabetic LDLr^−/− mice, and this effect was attenuated by H₂S treatment (Fig. 1E and F). Collectively, these data demonstrate that exogenous H₂S decreases atherosclerotic lesions in diabetic LDLr^−/− mice.

Oxidative stress plays an important role in the pathogenesis of diabetes and its complications. To determine whether the protective role of H₂S in atherosclerotic lesions might relate to reduction of reactive oxygen species (ROS), we measured aortic superoxide formation by DHE assay. As expected, compared with nondiabetic LDLr^−/− mice, endothelial fluorescence was increased in diabetic LDLr^−/− mice. In contrast, endothelial superoxide production in diabetic LDLr^−/− mice was attenuated after GYY4137 administration (Fig. 2A and B). Several recent studies have identified Nrf2 as a critical transcription factor that regulates a battery of antioxidant genes in the face of oxidative stress (3). Recent work has also shown that H₂S regulates Nrf2 in myocardial tissue (12). To examine whether Nrf2 is activated in response to H₂S treatment, we investigated the intracellular localization of Nrf2. Immunofluorescence microscopy showed enhanced nuclear staining of Nrf2 in aortas of H₂S-treated in comparison with untreated diabetic LDLr^−/− mice (Fig. 2C). To further confirm the Nrf2 localization, we performed double staining for Nrf2 and CD31 (endothelial marker), α-SMA (smooth muscle cells marker), or macrophage marker and found that Nrf2 could be clearly shown colocalized with three markers in aorta. Also, GYY4137 treatment increased Nrf2 nuclear translocation in aortic endothelial cells, smooth muscle cells, and macrophages (Supplementary Fig. 2). In addition, the induction of expression of the Nrf2-related antioxidant defense enzyme HO-1 was substantially increased by H₂S (Fig. 2D); whereas other Nrf2 target genes, such as Trx and GCLC, were unchanged (Supplementary Fig. 3).

Oxidative stress induces the expression of adhesion molecules such as VCAM-1 and ICAM-1, which promote the recruitment to, and accumulation of, inflammatory cells within the developing atherosclerotic lesion. Therefore, the levels of VCAM-1 and ICAM-1 in aorta were determined by RT-PCR and immunofluorescence. After 4 weeks on HFD, both VCAM-1 and ICAM-1 increased in diabetic LDLr^−/− mice, and this effect was abrogated by treatment with H₂S (Fig. 2E–G and Supplementary Fig. 4). Together, these results indicate that exogenous H₂S attenuates diabetes-accelerated atherosclerosis, most likely by maintaining redox balance via the Nrf2 pathway.

To further explore the pathophysiological significance of H₂S-induced Nrf2 activation in vivo, we mated LDLr^−/− mice with Nrf2^−/− mice to generate LDLr^−/−Nrf2^−/− mice. After injection of STZ and 4 weeks of HFD, with or without concomitant GYY4137 treatment, metabolic characteristics were assessed (Supplementary Table 3). Histological assessment of atherosclerotic lesions at the aortic sinus by ORO and H-E staining showed a marked increase of plaques in the aortic root from LDLr^−/−Nrf2^−/− diabetic mice fed HFD, and the aortic plaque area was now not reduced by H₂S treatment (Fig. 3A and B). Moreover, the expressions of superoxide, VCAM-1, and ICAM-1 were not reduced after treatment of GYY4137 in diabetic LDLr^−/−Nrf2^−/− mice (Fig. 3C–G and Supplementary Fig. 5). Complementary analyses of Nrf2 target gene levels in aorta revealed that the expression of HO-1 could not be augmented by H₂S in the presence of Nrf2 deficiency (Fig. 3H). These results demonstrate that Nrf2 is necessary for the inhibitory effect of H₂S to be exerted on diabetes-accelerated atherosclerosis in vivo.

Accumulation of cholesterol and cholesteryl esters in macrophages and subsequent foam cell formation is a critical early event in atherogenesis. To further investigate the molecular mechanisms underlying the effects of H₂S, we established a macrophage model in hyperglycemic and hyperlipidemic conditions in vitro, which replicates some of the characteristics of macrophages in the diabetes-accelerated atherosclerotic mouse model. Mouse peritoneal macrophages from C57BL/6 were incubated with high glucose (HG) plus ox-LDL (HG+ox-LDL), with or without GYY4137 for 24 h, after which foam cell formation and ROS production were measured by ORO staining and DHE assay, respectively. As expected, foam cell formation was induced in macrophages exposed to ox-LDL, and this effect was exaggerated by coincubation with HG (data not shown). Pretreatment with GYY4137 (50 or 100 μmol/L), but not SDG or ZYJ1122, abrogated this effect (Fig. 4A and Supplementary Fig. 6). In addition, superoxide generation was enhanced in HG+ox-LDL–stimulated macrophages (Fig. 4B and C), and this too was attenuated by pretreatment with H₂S.

Next, to test whether Nrf2 is involved in the effects of H₂S on macrophage function, we investigated its intracellular localization. Immunofluorescence microscopy showed enhanced nuclear staining of Nrf2 in cells treated with H₂S in comparison with vehicle-treated cells, in the presence of HG+ox-LDL (Fig. 4D). Western blotting analysis of cytoplasmic and nuclear protein extracts also indicated increased nuclear accumulation of Nrf2 protein in cells treated with H₂S (Fig. 4E and F), suggesting that Nrf2 is activated in response to H₂S exposure. Similarly, in the presence of HG+ox-LDL, H₂S-pretreated macrophages exhibited increased production of HO-1 (Fig. 4G).

Additional experiments were performed to confirm the involvement of Nrf2 in the protective effect of H₂S, using mouse peritoneal macrophages isolated from Nrf2^−/− mice. Inhibition of foam cell formation and superoxide generation induced by HG+ox-LDL was attenuated by H₂S treatment in Nrf2 knockout (KO) cells (Fig. 5A–C). Consistent with these results, elevation of HO-1 expression by H₂S treatment was also abolished in the Nrf2 KO group (Fig. 5D). Furthermore, HO-1 knockdown by siRNA or inhibition by ZnPP (HO-1 inhibitor) also abrogated H₂S-mediated suppression of O₂⁻ generation and foam cell formation (Supplementary Fig. 7). These results demonstrate that the Nrf2/HO-1 pathway is responsible for the inhibitory effects of H₂S on HG+ox-LDL–induced foam cell formation and oxidative stress in macrophages.

It has been reported that endothelial dysfunction caused by lipotoxicity or hyperglycemia is mediated through several mechanisms, including increased oxidative stress and proinflammatory responses. Therefore, we measured the effect of H₂S on oxidative stress in endothelial cells by DHE assay. Stimulation of HUVECs with HG+ox-LDL for 24 h caused an increase in the production of superoxide, and this increase was alleviated by pretreatment with GYY4137 (50 or 100 μmol/L) (Fig. 6A and B) but not with SDG or ZYJ1122 (Supplementary Fig. 8).

To confirm whether the cytoprotective effect of H₂S against oxidative stress was also associated with Nrf2, we carried out immunofluorescence and Western blotting for Nrf2. GYY4137 had no effect on Nrf2 phosphorylation, but can increase the Nrf2 protein expression in the nuclear in HG+ox-LDL–stimulated endothelial cells (Fig. 6C–E and Supplementary Fig. 9A), implying that H₂S may promote phosphorylation-independent Nrf2 nuclear translocation. Consistent with our in vivo study, H₂S reduced the expression of VCAM-1 and ICAM-1 (Fig. 6F and G); in addition, the Nrf2 target gene HO-1 was also increased by H₂S pretreatment (Fig. 6H and I). To further clarify whether H₂S-induced downregulation of oxidative stress was dependent on activation of the Nrf2 pathway, EA.hy926 cells were transfected with Nrf2 siRNA for 24 h before H₂S and HG+ox-LDL treatment. Western blotting revealed that individual transfection with Nrf2 siRNA successfully reduced Nrf2 protein expression at 24 h posttransfection, as compared with negative control siRNA-transfected (Ctl siRNA) cells (Fig. 7A). Nrf2 knockdown abrogated H₂S-mediated suppression of ROS production induced by HG+ox-LDL in endothelial cells (Fig. 7B and C). Furthermore, inhibition of HO-1 expression or activity by siRNA or ZnPP abolished the protective effects of H₂S (Supplementary Fig. 10). Together, these results indicate that the antioxidative and anti-inflammatory effects of H₂S in the presence of HG+ox-LDL are partially mediated by the Nrf2/HO-1 pathway in endothelial cells.

Generally, Nrf2 is retained in an unactivated state binding with Keap1 in the cytoplasm, which serves as an adaptor for the degradation of Nrf2. Nrf2 can be activated by physiological stimuli that disrupt Keap1-Nrf2 interactions leading to nuclear translocation of Nrf2 (22). To further explore the mechanisms of Nrf2 activation, we immunoprecipitated the cell lysate using an anti-Keap1 antibody and blotted for Nrf2. The results showed that GYY4137 decreased the Nrf2/Keap1 interaction in HG+ox-LDL–treated endothelial cells (Fig. 8A). S-sulfhydration, the addition of one sulfhydryl to the thiol side of the cysteine residue and formation of a persulfide group (R-S-S-H), has been identified as a novel posttranslational modification by H₂S in eukaryotic cells. However, the covalent modification in sulfhydration is reversed by reducing agents, such as dithiothreitol (23). We tested the S-sulfhydration of Nrf2 and found that H₂S donor GYY4137 or NaHS had no effect on Nrf2 S-sulfhydration (Supplementary Fig. 9B). We next investigated whether H₂S directly modified Keapl. After preincubation with GYY4137, EA.hy926 cells were treated with HG+ox-LDL and subjected to the “tag-switch” assay. There was stronger S-sulfhydration of Keap1 after GYY4137 incubation (Fig. 8B). To identify the S-sulfhydrated cysteine residue, Keap1 mutated at Cys151, Cys273, or Cys288 to alanine (C151A, C273A, or C288A) or wild type (WT) was transfected into endothelial cells. H₂S still enhanced S-sulfhydration on Keap1 after WT or mutated Keap1 at Cys288 but not at Cys151 and Cys273 overexpression (Fig. 8C). H₂S increased Nrf2 dissociation from Keap1 in HG+ox-LDL–treated endothelial cells after Keap1-WT and Keap1-C273A but not Keap1-C151A overexpression (Fig. 8D). Moreover, after Keap1 mutation at Cys151, H₂S failed to induce Nrf2 nuclear translocation or decrease the generation of superoxide (Fig. 8E–G). These findings indicate that S-sulfhydration of Cys151 in Keap1 is critical for Nrf2 activation in HG+ox-LDL–treated endothelial cells.

A complex interaction between inflammation, lipid deposition, monocytic infiltration, and endothelial dysfunction is responsible for the initiation and progression of diabetes-accelerated atherosclerosis (1). Experimental evidence for an antiatherosclerotic effect of H₂S has been obtained in numerous studies in hyperlipidemic animal and cell models (15,24–26), but the antiatherosclerotic effect in the context of diabetes has not been previously investigated. Recent data published by our group demonstrated that treatment with H₂S decreased aortic atherosclerotic plaque formation and partially restored aortic endothelium-dependent relaxation in ApoE^−/− mice fed an HFD (15). Exogenous H₂S improved endothelium-dependent relaxation in isolated vascular rings incubated with HG and attenuated hyperglycemia-induced DNA injury and improved cellular viability in bEnd.3 microvascular endothelial cells (8). These data suggested that exogenous H₂S might serve as a treatment option for diabetes-associated atherosclerosis. Indeed, in the current study, we found that H₂S supplementation reduces lesion area and macrophage infiltration in diabetic LDLr^−/− mice. In agreement with these findings, we also observed that H₂S treatment attenuated HG+ox-LDL–induced foam cell formation. Our study provides the first evidence that H₂S may prevent the development of diabetes-accelerated atherosclerosis, which is not related to any effects on circulating blood glucose or cholesterol.

Several pathological mechanisms have been proposed for diabetic vascular complications, including diabetes-accelerated atherosclerosis, such as increased polyol pathway flux, increased advanced glycation end product formation, and activation of protein kinase C, all of these in association with hyperglycemia-induced ROS accumulation (27). Endothelial cells and macrophages are both sources of ROS. Indeed, in our study, we demonstrated that STZ-treated LDLr^−/− mice fed an HFD showed an increase in atherosclerotic plaques compared with nondiabetic LDLr^−/− mice, accompanied by increased superoxide production in aorta, and this was further confirmed in HG+ox-LDL–treated macrophages and endothelial cells. The increase in ROS promotes the recruitment and accumulation of inflammatory cells to the developing atherosclerotic lesion. H₂S has also been shown to have powerful antioxidant properties. Exogenous H₂S attenuates the hyperglycemia-induced enhancement of ROS formation in endothelial cells and human U937 monocytes (8,28). In line with these findings, we observed that H₂S decreased superoxide generation in macrophages and endothelial cells cultured with HG+ox-LDL. Furthermore, we showed that superoxide production in the aortas of diabetic LDLr^−/− mice was reduced after H₂S administration. In this study, we show for the first time that inhibition of HG+ox-LDL–generated ROS with H₂S prevents the diabetes-induced increase in plaque area. Additionally, we found that H₂S attenuates the increase in aortic VCAM-1 and ICAM-1 expression.

Recent studies indicate that H₂S may upregulate endogenous antioxidants through an Nrf2-dependent signaling pathway (12) to combat oxidative stress. To date, the role of Nrf2 in atherosclerosis remains controversial. Myeloid Nrf2 deficiency aggravates both early and late stages of atherosclerosis in LDLr^−/− mice (29,30). Ellagic acid improves oxidant-induced endothelial dysfunction and atherosclerosis partly via Nrf2 activation (31). In contrast to these reported protective actions, Nrf2 has also been ascribed as having potentially proatherogenic functions, in that ApoE^−/−Nrf2^−/− double KO mice exhibited reduced plaque (32). In diabetes-associated atherosclerosis, a novel analog of the Nrf2 agonist bardoxolone methyl has been found to reduce atherosclerotic lesions as well as oxidative stress and the proinflammatory mediators ICAM-1 and VCAM-1 in STZ-induced diabetic ApoE^−/− mice (7). Our data support previous findings regarding the protective actions of Nrf2 and suggest that H₂S can attenuate endothelial dysfunction, foam cell formation, and atherosclerosis in the context of diabetes, at least partially via the Nrf2/HO-1 pathway.

A widely accepted model for Nrf2 nuclear accumulation describes that a modification of the Keap1 cysteines leads directly to the dissociation of the Keap1-Nrf2 complex (33). Recently, one study suggested that Keap1 can be S-sulfhydrated at Cys151, which stimulates the dissociation of Nrf2 to enable its translocation to the nucleus (34). We found that Keap1 could be S-sulfhydrated at Cys151 and Cys273 simultaneously, but only the S-sulfhydration of Cys151 was involved in activation of Nrf2, which decreased the ROS generation to improve endothelial function. Kim et al. (35) found that thiol modification of Keap1 Cys288 is responsible for diallyl trisulfide–induced activation of Nrf2 signaling. However, in our study, Cys288 of Keap1 could not be S-sulfhydrated after treatment with GYY4137. This discrepancy may be attributed to different regulatory mechanisms in different cell types, the use of different H₂S treatment regiments giving rise to different kinetics of H₂S release. Nevertheless, this study demonstrates a significant role of Keap1 Cys151 S-sulfhydration in the protective effects of H₂S against diabetes-accelerated atherosclerosis.

In summary, our study provided definitive evidence that H₂S can lessen diabetes-accelerated atherosclerosis in LDLr^−/− mice and improve hyperglycemia/ox-LDL–induced injury in macrophages and endothelial cells. This protective effect of H₂S can partly be attributed to activation of Nrf2 via Keap1 S-sulfhydration at Cys151. Our findings suggest that activation of Nrf2 may be a potential novel therapeutic strategy against diabetes-associated vascular disease and that exogenous H₂S administration in the form of an H₂S donor (GYY4137) may be of therapeutic benefit in the setting of diabetes-associated atherosclerosis. Finally, our study provides new insight into the mechanisms responsible for the antiatherosclerotic effects of H₂S in the context of diabetes.

Acknowledgments. The authors thank Hongliang Li (Renmin Hospital of Wuhan University) for providing the Nrf2^−/− mice.

Funding. This work was supported by grants from the National Natural Science Foundation of China (81200197, 81170083, and 81330004) and the National Basic Research Program of China 973 (2011CB503903 and 2012CB517803).

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

Author Contributions. L.X. and Y.G. researched data, contributed to discussion, and edited the manuscript. M.W., S.Z., W.W., Y.M., Y.H., and Y.W. researched data. G.M. contributed to discussion and proofread the manuscript. G.L. designed the study and reviewed data. P.K.M., X.W., H.W., Z.Z., and Y.Y. reviewed the manuscript. A.F. contributed to discussion and rewrote the manuscript. Z.H. reviewed data and edited the manuscript. Y.J. designed the study, reviewed data, and edited the manuscript. All authors approved the final version of the manuscript. Y.J. 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.