SIRT6 Protects Smooth Muscle Cells From Senescence and Reduces Atherosclerosis

Abstract

Rationale: Vascular smooth muscle cell (VSMC) senescence promotes atherosclerosis and features of plaque instability, in part, through lipid-mediated oxidative DNA damage and telomere dysfunction. SIRT6 (Sirtuin 6) is a nuclear deacetylase involved in DNA damage response signaling, inflammation, and metabolism; however, its role in regulating VSMC senescence and atherosclerosis is unclear.

Objective: We examined SIRT6 expression in human VSMCs, the role, regulation, and downstream pathways activated by SIRT6, and how VSMC SIRT6 regulates atherogenesis.

Methods and results: SIRT6 protein, but not mRNA, expression was markedly reduced in VSMCs in human and mouse atherosclerotic plaques, and in human VSMCs derived from plaques or undergoing replicative or palmitate-induced senescence versus healthy aortic VSMCs. The ubiquitin ligase CHIP (C terminus of HSC70-interacting protein) promoted SIRT6 stability, but CHIP expression was reduced in human and mouse plaque VSMCs and by palmitate in a p38- and c-Jun N-terminal kinase-dependent manner. SIRT6 bound to telomeres, while SIRT6 inhibition using shRNA or a deacetylase-inactive mutant (SIRT6^H133Y) shortened human VSMC lifespan and induced senescence, associated with telomeric H3K9 (histone H3 lysine 9) hyperacetylation and 53BP1 (p53 binding protein 1) binding, indicative of telomere damage. In contrast, SIRT6 overexpression preserved telomere integrity, delayed cellular senescence, and reduced inflammatory cytokine expression and changes in VSMC metabolism associated with senescence. SIRT6, but not SIRT6^H133Y, promoted proliferation and lifespan of mouse VSMCs, and prevented senescence-associated metabolic changes. ApoE^-/- (apolipoprotein E) mice were generated that overexpress SIRT6 or SIRT6^H133Y in VSMCs only. SM22α-hSIRT6/ApoE^-/- mice had reduced atherosclerosis, markers of senescence and inflammation compared with littermate controls, while plaques of SM22α-hSIRT6^H133Y/ApoE^-/- mice showed increased features of plaque instability.

Conclusions: SIRT6 protein expression is reduced in human and mouse plaque VSMCs and is positively regulated by CHIP. SIRT6 regulates telomere maintenance and VSMC lifespan and inhibits atherogenesis, all dependent on its deacetylase activity. Our data show that endogenous SIRT6 deacetylase is an important and unrecognized inhibitor of VSMC senescence and atherosclerosis.

Keywords: atherosclerosis; inflammation; metabolism; muscle cells; telomere.

Figure 1.

SIRT6 (Sirtuin 6) protein expression is reduced in vascular smooth muscle cells (VSMCs) in human atherosclerosis and undergoing replicative and palmitate-induced senescence. A, Immunohistochemistry and (B) quantification for α-smooth muscle cell (SMC)-actin (blue) and SIRT6 (brown) of VSMCs from human aorta (n=9) and carotid artery plaques (n=6; unpaired t test). High-power images of outlined areas indicate SIRT6⁺/α-SMC-actin⁺ cells (arrows); scale bar=250 µm (high-power views) and 25 µm (low-power views). C, Quantitative polymerase chain reaction (QPCR) for SIRT6 mRNA expression of human plaque VSMCs (n=6), or aortic VSMCs, either actively dividing (n=6), undergoing replicative senescence (RS; n=6) or palmitate-induced stress-induced premature senescence (SIPS; n=3; 1-way ANOVA, Dunnett post hoc). D, Western blot for SIRT6 and p16^ink4a of human plaque VSMCs (middle; n=3) and aortic VSMCs, either actively dividing (left) or undergoing RS (right; n=4) with quantification (1-way ANOVA, Dunnett post hoc). β-actin was used as loading control. E, Western blotting for SIRT6 and p16^ink4a and quantification of human VSMCs (hVSMCs) undergoing SIPS (unpaired t test, n=3). F, Immunohistochemistry and (G) quantification for α-SMC-actin (blue) and SIRT6 (brown) of VSMCs in plaques of chow or high-fat diet (HFD)-fed ApoE^−/− (apolipoprotein E) mice (unpaired t test, n=6–7). High-power images of outlined areas indicate SIRT6⁺/α-SMC-actin⁺ cells (arrows); scale bar=150 μm (high-power views) and 10 μm (low-power views). H, QPCR for SIRT6 mRNA of mouse aortas of chow or HFD-fed ApoE^−/− mice (unpaired t test, n=3). Data are shown as mean±SEM with nominal P<0.05 or multiplicity adjusted P<0.05.

Figure 2.

SIRT6 (Sirtuin 6) protein stability is positively regulated by ubiquitin ligase CHIP (C terminus of HSC70-interacting protein). A, Western blot for SIRT6 and CHIP in human vascular smooth muscle cells (hVSMCs) treated with 0, 200, or 300 µmol/L palmitate (palm) for 24 or 48 h, and quantification (2-way ANOVA, Tukey post hoc, n=3). β-actin was used as loading control. B, Quantitative polymerase chain reaction (QPCR) for SIRT6 and CHIP mRNA expression of hVSMCs treated with 300 µmol/L palmitate or control BSA for 24 or 48 h (2-way ANOVA, Bonferroni post hoc, n=3–5). C, Immunoprecipitation for SIRT6 with hVSMCs treated with palmitate (p) or BSA (c) for 48 h, followed by Western blotting for CHIP and SIRT6. IgG1κ used as isotype control. D, Western blot for SIRT6 in hVSMCs treated with palmitate or BSA for 48 h, with/without proteasome inhibitor MG132 (1 µmol/L) or cycloheximide (CHX, 1 mg/mL), and quantification (2-way ANOVA, Tukey post hoc, n=3). E, Immunoprecipitation for SIRT6 with hVSMCs treated with palmitate (p) or BSA (c) for 48 h, with/without MG132 (1 µmol/L), followed by Western blotting for ubiquitin and SIRT6. F, Western blot with quantification for CHIP and SIRT6 expression in hVSMCs after transient silencing of CHIP (siCHIP) vs control siRNA (siCtrl) followed by 48 h palmitate or BSA treatment (1-way ANOVA, Tukey post hoc, n=3). G, SIRT6 half-life determined by Western blotting after siCHIP vs siCtrl at 0, 24, 48, and 72 h (n=3). H, Western blot for CHIP, Myc-tag, and SIRT6 in hVSMCs infected with a lentivirus expressing Myc-tagged CHIP vs empty vector (EV), treated with palmitate (or BSA) for 48 h, and quantification (2-way ANOVA, Tukey post hoc, n=3). I, QPCR for CHIP mRNA expression in hVSMCs treated with palmitate (or BSA) for 24 h with/without p38-inhibitor (p38 I, 1 µmol/L) or for 48 h with/without JNK (c-Jun N-terminal kinase)-inhibitor (40 µmol/L; 1-way ANOVA, Dunnett post hoc, n=3–4). J and K, Western blot analysis for phospho-p38, p38 (J) phospho-c-Jun, total c-Jun (K), CHIP and SIRT6 expression in hVSMCs treated with palmitate (or BSA) for 24 and 48 h, with/without (J) p38 I or (K) JNK I, and quantification (1-way ANOVA, Dunnett post hoc, n=3). Histone 3 was used as loading control. Data are shown as mean±SEM with multiplicity adjusted P<0.05.

Figure 3.

SIRT6 (Sirtuin 6) delays senescence, binds to telomeres and regulates telomeric H3K9 (histone H3 lysine 9) deacetylation and DNA damage repair (DDR) signaling of human vascular smooth muscle cells (hVSMCs). A, Western blot for SIRT6, H3K9ac (acetylated H3K9), H3K27ac (acetylated H3K27), and p16 in hVSMCs expressing shRNA against SIRT6 (sh#1 and sh#2) or overexpressing SIRT6 (S6) or its catalytic inactive mutant (SIRT6^H133Y, S6H) with quantification (1-way ANOVA, Dunnett post hoc, n=3–5). Total histone 3 and β-actin were used as loading control. B, Representative cumulative population doublings (CPD) of SIRT6-depleted (top) or SIRT6/SIRT6^H133Y (bottom) hVSMCs. C, % 5-ethynyl-2’-deoxyuridine (EdU) incorporation and (D) % SAβG (senescence-associated beta galactosidase)-positive hVSMCs in late-passage experimental cell lines (p10-12; 1-way ANOVA, Dunnett post hoc, n=4). E, Telomere chromatin immunoprecipitation (Telo-ChIP) analysis for SIRT6 in hVSMCs overexpressing SIRT6 or expressing SIRT6^H133Y (1-way ANOVA, Tukey post hoc, n=4) at early passage (p4-6). F and G, Telo-ChIP analysis for H3K9ac (F), and 53BP1 (p53 binding protein 1; G) in SIRT6-depleted or SIRT6/SIRT6^H133Y hVSMCs at early passage (p4-6; 1-way ANOVA, Dunnett post hoc, n=3–4). Data are shown as mean±SEM with multiplicity adjusted P<0.05. SAβG indicates senescence-associated β galactosidase.

Figure 4.

SIRT6 (Sirtuin 6) prevents senescence-associated changes in cell metabolism. A–D, Seahorse extracellular flux profiles of human vascular smooth muscle cells (hVSMCs) expressing shRNA against SIRT6 (sh#1 and sh#2) or overexpressing SIRT6 (S6) or SIRT6^H133Y (S6H) in the presenescent (p4-6; A and B) and senescent (p10-12) stage (C and D). Seahorse tracings are presented as % of baseline, after normalization to protein (left). A and C, Fatty acid oxidation was measured by oxygen consumption rate (OCR) upon addition of palmitate (2 mmol/L) and etomoxir (240 µmol/L). B and D, Glycolysis was measured by extracellular acidification rate (ECAR) upon glucose (10 mmol/L), oligomycin (2 µg/mL), and 2-deoxyGlucose (2DG, 50 mmol/L). Maximal respiration and glycolytic capacity are shown in right panels (1-way ANOVA, Dunnett post hoc, n=3–5). Data are shown as mean±SEM with multiplicity adjusted P<0.05.

Figure 5.

SIRT6 (Sirtuin 6) partially suppresses inflammation in senescent human vascular smooth muscle cells (hVSMCs). A and C, Western blot for phospho-NF-κB (nuclear factor-κB) p65 and total NF-κB p65 of hVSMCs expressing shRNA against SIRT6 (sh#1 and sh#2), overexpressing SIRT6 (S6) or expressing SIRT6^H133Y (S6H) at presenescent (p4-6; A) or senescent (p10-12; C) stage (1-way ANOVA, Dunnett post hoc, n=3–4). B and D, Quantitative polymerase chain reaction (QPCR) for expression of proinflammatory cytokines of experimental cell lines at presenescent (n=3; B) or senescent stage (n=4; D; 1-way ANOVA, Dunnett post hoc). Data are shown as mean±SEM with multiplicity adjusted P<0.05. IL indicates interleukin; and MCP-1, monocyte chemoattractant protein 1.

Figure 6.

Characterization of mice and mouse vascular smooth muscle cells (VSMCs) expressing human SIRT6 (Sirtuin 6) or SIRT6^H133Y. A, Western blot for hIRT6 and V5 in aortas of wild-type (WT) mice or mice expressing hSIRT6 (S6) or SIRT6^H133Y (S6H). B, Quantitative polymerase chain reaction (QPCR) analysis of SMC-rich tissues (aorta, heart, lung, and gut) for hSIRT6 in experimental mice (1-way ANOVA, nominal P<0.05 shown, n=4). C, Western blotting for human and mouse SIRT6, V5 and H3K9ac (acetylated histone H3 lysine 9) in aortic VSMCs isolated from experimental mice. D, QPCR for mouse endogenous SIRT6, SIRT1, SIRT2, SIRT3, and SIRT5 (1-way ANOVA, nominal P<0.05 shown, n=4). mVSMC culture lifespan assessed by cumulative population doublings (CPD; E) and cell proliferation by 5-ethynyl-2’-deoxyuridine (EdU) incorporation (F; 1-way ANOVA, Dunnett post hoc, n=4). G and I, Seahorse profiles for fatty acid oxidation (FAO; G) and glycolysis (I) and maximal respiration and glycolytic capacity (1-way ANOVA, Dunnett post hoc, n=3). H and J, QPCR for selected FAO (H) and glycolysis (J) genes. Data are shown as mean±SEM with multiplicity adjusted P<0.05. ACOX1 indicates acyl-CoA-oxidase-1; CPT1, carnitine-palmitoyltransferase-1; GLUT1, glucose transporter 1; and PDK1, pyruvate dehydrogenase kinase 1.

Figure 7.

Vascular smooth muscle cell (VSMC) SIRT6 (Sirtuin 6) protects against development of atherosclerosis. A, Oil Red O (ORO) staining of descending aorta and quantification of ORO-positive area in littermate ApoE^−/− (apolipoprotein E) (wild type [WT]), SM22α-SIRT6/ApoE^−/− (S6), and SM22α-hSIRT6^H133Y/ApoE^−/− (S6H) mice fed a high-fat diet (HFD) for 16 weeks (1-way ANOVA, Dunnett post hoc, n=7–8). Scale bar=3 mm. B, Plaque area at the maximum area of aortic root plaques of experimental mice (1-way ANOVA, Dunnett post hoc, n=13–16). C, Representative images of aortic root plaques stained with Masson trichrome. Asterisks and black arrowheads indicate necrotic cores and fibrous caps, respectively. Scale bar=150 µm. Analysis of cap/plaque (D), cap/core (E), and core/plaque (F) ratios (1-way ANOVA, Dunnett post hoc [D] and Kruskal-Wallis, Dunn post hoc [E–F], n=13–16). Analysis of % α-SMC-actin (G), % Mac-3 (H), and % collagen (I)–positive areas in aortic plaques (Kruskal-Wallis, Dunn post hoc [G and I] and 1-way ANOVA, Dunnett post hoc [H], n=13–16). J, Quantification of SAβG (senescence-associated β-galactosidase)-positive area in plaques of brachiocephalic artery (Kruskal-Wallis, Dunn post hoc, n=6–9). K, mRNA expression of p16 and senescence-associated secretory phenotype cytokines in aortas of experimental mice (1-way ANOVA, Dunnett post hoc, n=5). Data are shown as mean±SEM with multiplicity adjusted P<0.05. IL indicates interleukin; and SAβG, senescence-associated β galactosidase.

Figure 8.

Schematic overview of SIRT6 (Sirtuin 6) regulation in vascular smooth muscle cells (VSMCs) in atherosclerosis. In normal healthy aortic VSMCs, the ubiquitin ligase CHIP (C terminus of HSC70-interacting protein) binds to SIRT6 and promotes its stability. SIRT6 deacetylates histone 3 at telomeric DNA, preserves telomere integrity and protects against telomere damage, prolonging VSMC lifespan. In atherosclerosis, CHIP expression is reduced upon exposure to free fatty acids such as palmitate in a p38 and JNK (c-Jun N-terminal kinase)-dependent manner. As a result, SIRT6 is no longer protected from proteasomal degradation. Loss of SIRT6 binding (or its activity) leads to hyperacetylation of telomere histones and 53BP1 (p53 binding protein 1; NHEJ)-mediated telomere damage, which impairs telomere integrity and promotes senescence. Senescent VSMCs are characterized by a metabolic shift from fatty acid oxidation to glycolysis and increased expression of proinflammatory markers. JNK indicates c-Jun N-terminal kinase; and NHEJ, nonhomologous end joining.