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

Abstract

Hydrogen sulfide (H2S) has been shown to have powerful antioxidative and anti-inflammatory properties that can regulate multiple cardiovascular functions. However, its precise role in diabetes-accelerated atherosclerosis remains unclear. We report here that H2S reduced aortic atherosclerotic plaque formation with reduction in superoxide (O2 (-)) generation and the adhesion molecules in streptozotocin (STZ)-induced LDLr(-/-) mice but not in LDLr(-/-)Nrf2(-/-) mice. In vitro, H2S inhibited foam cell formation, decreased O2 (-) generation, and increased nuclear factor erythroid 2-related factor 2 (Nrf2) nuclear translocation and consequently heme oxygenase 1 (HO-1) expression upregulation in high glucose (HG) plus oxidized LDL (ox-LDL)-treated primary peritoneal macrophages from wild-type but not Nrf2(-/-) mice. H2S also decreased O2 (-) and adhesion molecule levels and increased Nrf2 nuclear translocation and HO-1 expression, which were suppressed by Nrf2 knockdown in HG/ox-LDL-treated endothelial cells. H2S increased S-sulfhydration of Keap1, induced Nrf2 dissociation from Keap1, enhanced Nrf2 nuclear translocation, and inhibited O2 (-) generation, which were abrogated after Keap1 mutated at Cys151, but not Cys273, in endothelial cells. Collectively, H2S attenuates diabetes-accelerated atherosclerosis, which may be related to inhibition of oxidative stress via Keap1 sulfhydrylation at Cys151 to activate Nrf2 signaling. This may provide a novel therapeutic target to prevent atherosclerosis in the context of diabetes.

Figure 1

Effects of H₂S on atherosclerosis in HFD-fed diabetic LDLr^−/− mice. Diabetic LDLr^−/− mice were fed an HFD and received daily intraperitoneal injection of saline or H₂S donor GYY4137 (133 μmol/kg/day) for 4 weeks. A: Schema of experimental procedure. B: Lesion areas shown were quantified using ORO staining of the thoracoabdominal aorta. C and D: Representative ORO- and H-E–stained images and quantification of aortic sinus sections from LDLr^−/− (n = 6), STZ+HFD (n = 9), and STZ+HFD+GYY4137 (n = 8) mice. Scale bars, 200 μm. E and F: Frozen sections of aortic root were stained for antimacrophage (green) and DAPI (blue). Dotted lines indicate the boundary of lesion and aortic tunica intima. Quantitative data in the graph represent the positively stained area percentage of plaque (n = 6). Scale bars, 100 μm. Data shown are mean ± SEM. ***P < 0.001 vs. LDLr^−/− mice; ###P < 0.001 vs. STZ+HFD mice. L, lumen.

Figure 2

Effects of H₂S on ROS formation and Nrf2, VCAM-1, and ICAM-1 expression in aortas from HFD-fed diabetic LDLr^−/− mice. A: Representative DHE fluorescence image of aortic tissue from LDLr^−/−, STZ+HFD, and STZ+HFD+GYY4137 mice. Scale bars, 100 μm. B: Quantification of DHE fluorescence image of A. ***P < 0.001 vs. LDLr^−/− mice; ##P < 0.01 vs. STZ+HFD mice. n = 6. C: Representative immunostaining for Nrf2 (green) and DAPI (blue) of aorta. Scale bars, 50 μm. D: mRNA levels of HO-1 in the aortas of LDLr^−/− (n = 6), STZ+HFD (n = 6), and STZ+HFD+GYY4137 (n = 8) mice, as determined by quantitative RT-PCR analysis. mRNA levels of VCAM-1 (E) and ICAM-1 (F) in the aortas of LDLr^−/− (n = 6), STZ+HFD (n = 6), and STZ+HFD+GYY4137 (n = 6) mice. G: Representative VCAM-1 and ICAM-1 immunostaining of aortic arch section (with arrows). Scale bars, 100 μm. *P < 0.05, **P < 0.01, and ***P < 0.001 vs. LDLr^−/− mice; #P < 0.05 and ##P < 0.01 vs. STZ+HFD mice.

Figure 3

Effects of H₂S on diabetes-accelerated atherosclerosis in LDLr^−/−Nrf2^−/− mice. Diabetic LDLr^−/−Nrf2^−/− mice were fed an HFD and received daily intraperitoneal injection of saline or GYY4137 (133 μmol/kg/day) for 4 weeks. A and B: Representative ORO- and H-E–stained images and quantification of aortic sinus sections from LDLr^−/−Nrf2^−/−, STZ+HFD, and STZ+HFD+GYY4137 mice (n = 5). Scale bars, 200 μm. C: Representative DHE fluorescence image of aorta. Scale bars, 100 μm. D: Quantification of DHE fluorescence image of C. *P < 0.05 and ***P < 0.001 vs. LDLr^−/− mice; ##P < 0.01 and ###P < 0.001 vs. LDLr^−/−Nrf2^−/− mice. n = 5–7. E: Representative VCAM-1 and ICAM-1 immunostaining of aortic arch section (with arrows). Scale bars, 100 μm. mRNA levels of VCAM-1 (F) and ICAM-1 (G) in the aorta of LDLr^−/−Nrf2^−/−, STZ+HFD, and STZ+HFD+GYY4137 mice (n = 5). H: mRNA levels of HO-1 in the aorta (n = 5). Data shown are mean ± SEM. *P < 0.05 and **P < 0.01 vs. control.

Figure 4

Effects of H₂S on HG+ox-LDL-treated primary peritoneal macrophages. Isolated peritoneal macrophages from C57BL/6 mice were treated with d-glucose (D-Glu) (25 mmol/L) plus ox-LDL (50 mg/L) in the presence or absence of GYY4137 (50 or 100 μmol/L) for 24 h. A: Macrophages incubated as above and stained with ORO. Scale bars, 20 μm. B: Representative DHE-stained images showing ROS generation in each condition. Scale bars, 50 μm. C: Quantification of DHE fluorescence image of B. ***P < 0.01 vs. untreated control; #P < 0.05 and ##P < 0.01 vs. treatment with HG+ox-LDL. n = 5. D: Immunohistochemistry was performed on macrophages stained with antibody directed against Nrf2 (green) and DAPI (blue). Scale bars, 20 μm. Western blotting analysis and quantification of cytoplasmic (E) and nuclear (F) Nrf2 protein. GAPDH and histone H3 were used for normalization for cytoplasmic and nuclear proteins, respectively (n = 4). G: Western blotting analysis and quantification of HO-1 protein expression (n = 4). Data shown are mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001 vs. untreated control; ##P < 0.01 vs. treatment with HG+ox-LDL.

Figure 5

Effects of H₂S on HG+ox-LDL–treated primary peritoneal macrophages from Nrf2^−/− mice. Isolated peritoneal macrophages from C57BL/6 (WT) and Nrf2^−/− mice were treated with d-glucose (D-Glu) (25 mmol/L) plus ox-LDL (50 mg/L) in the presence or absence of GYY4137 (100 μmol/L) for 24 h. A: Macrophages from Nrf2^−/− mice were stained with ORO. Scale bars, 20 μm. B: Representative DHE-stained images of macrophages from WT and Nrf2^−/− mice. Scale bars, 50 μm. C: Quantification of DHE fluorescence image of B. ***P < 0.001 vs. WT control; ###P < 0.001 vs. WT with HG+ox-LDL; &&&P < 0.001 vs. Nrf2^−/− control. n = 3. D: Western blotting analysis and quantification of HO-1 protein expression in macrophages from Nrf2^−/− mice (n = 4). Data shown are mean ± SEM. *P < 0.05 vs. untreated control.

Figure 6

Effects of H₂S on HG+ox-LDL–treated endothelial cells. HUVECs were treated with d-glucose (D-Glu) (25 mmol/L) plus ox-LDL (50 mg/L) in the presence or absence of GYY4137 (50 or 100 μmol/L) for 24 h. A: Representative DHE-stained images showing ROS generation. Scale bars, 100 μm. B: Quantification of DHE fluorescence image of A. **P < 0.01 vs. untreated control; #P < 0.05 and ##P < 0.01 vs. treatment with HG+ox-LDL. n = 5. C: Immunohistochemistry was performed on HUVECs stained with antibody directed against Nrf2 (green) and DAPI (blue). Scale bars, 20 μm. Western blotting analysis and quantification of cytoplasmic (D) and nuclear (E) Nrf2 protein. GAPDH and histone H3 were used for normalization for cytoplasmic (n = 5) and nuclear (n = 4) proteins, respectively. Western blotting analysis and quantification of VCAM-1 (F) (n = 4) and ICAM-1 (G) (n = 3) protein. H: mRNA levels of HO-1 (n = 5). I: Western blotting analysis and quantification of HO-1 protein expression (n = 4). Data shown are mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001 vs. untreated control; #P < 0.05, ##P < 0.01, and ###P < 0.001 vs. treatment with d-glucose plus ox-LDL.

Figure 7

Effects of H₂S on HG+ox-LDL–treated Nrf2 knockdown endothelial cells. EA.hy926 endothelial cells were transfected with control siRNA (Ctl siRNA) or Nrf2 siRNA for 24 h and then treated with d-glucose (25 mmol/L) and ox-LDL (50 mg/L) in the presence or absence of GYY4137 (100 μmol/L) for 24 h. A: Western blotting analysis and quantification of Nrf2 (n = 3). Data shown are mean ± SEM. ***P < 0.001 vs. Ctl siRNA control. B: Representative DHE staining images. Scale bars, 50 μm. C: Quantification of DHE fluorescence image of B. **P < 0.01 vs. Ctl siRNA control; #P < 0.05 vs. Ctl siRNA with HG+ox-LDL; &&&P < 0.001 vs. Nrf2 siRNA control. n = 5.

Figure 8

H₂S S-sulfhydrylated Keap1 at Cys151 to regulate Nrf2 transcription activity and reduce the generation of ROS in HG+ox-LDL–treated endothelial cells. A: EA.hy926 endothelial cells were treated with GYY4137 (100 μmol/L) followed by d-glucose (D-Glu) (25 mmol/L) plus ox-LDL (50 mg/L) stimulation for 24 h. Cell lysates were immunoprecipitated with an anti-Keap1 or an anti-IgG antibody (negative control) and blotted with an anti-Nrf2 antibody (top panel). An aliquot of total lysate was analyzed for Keap1, Nrf2, and GAPDH expression (bottom panel). B: EA.hy926 endothelial cells were treated with dithiothreitol (DTT) (1 mmol/L, negative control) or d-glucose (25 mmol/L) plus ox-LDL (50 mg/L) in the presence or absence of GYY4137 (100 μmol/L) for 2 h. S-sulfhydration on Keap1 was detected with the “tag-switch” method. C: After plasmid transfection of Keap1-WT or mutated Keap1 at Cys151, Cys273, and Cys288 for 24 h followed by GYY4137 (100 μmol/L) treated for another 2 h, S-sulfhydration on Keap1 was detected with the “tag-switch” method. D: Transfected cells were treated with d-glucose (25 mmol/L) plus ox-LDL (50 mg/L) in the presence or absence of GYY4137 (100 μmol/L) for another 24 h, cell lysates were immunoprecipitated with an anti-Keap1 antibody, and the immunoprecipitated proteins were subjected to immunoblot analysis with anti-Nrf2 antibodies (top panel). The total lysates were analyzed with anti-Keap1, anti-Nrf2, and anti-GAPDH antibodies (bottom panel). E and F: Transfected cells were treated with d-glucose (25 mmol/L) plus ox-LDL (50 mg/L) in the presence or absence of GYY4137 (100 μmol/L) for 24 h. Nuclear extracts prepared from cells were subjected to Western blotting analysis for detecting the nuclear localization of Nrf2 (n = 4). ROS accumulation was determined by the DHE assay. Scale bars, 50 μm. Data shown are mean ± SEM. **P < 0.01 vs. Keap1-WT–transfected cells treated with d-glucose plus ox-LDL. G: Quantification of DHE fluorescence image of F. **P < 0.01 vs. untreated Keap1-WT–transfected cells; ##P < 0.01 vs. Keap1-WT–transfected cells treated with d-glucose and ox-LDL; &&P < 0.01 vs. untreated C151A-transfected cells. n = 4.