Mitochondrial CypD Acetylation Promotes Endothelial Dysfunction and Hypertension

Abstract

Background: Nearly half of adults have hypertension, a major risk factor for cardiovascular disease. Mitochondrial hyperacetylation is linked to hypertension, but the role of acetylation of specific proteins is not clear. We hypothesized that acetylation of mitochondrial CypD (cyclophilin D) at K166 contributes to endothelial dysfunction and hypertension.

Methods: To test this hypothesis, we studied CypD acetylation in patients with essential hypertension, defined a pathogenic role of CypD acetylation in deacetylation mimetic CypD-K166R mutant mice and endothelial-specific GCN5L1 (general control of amino acid synthesis 5 like 1)-deficient mice using an Ang II (angiotensin II) model of hypertension.

Results: Arterioles from hypertensive patients had 280% higher CypD acetylation coupled with reduced Sirt3 (sirtuin 3) and increased GCN5L1 levels. GCN5L1 regulates mitochondrial protein acetylation and promotes CypD acetylation, which is counteracted by mitochondrial deacetylase Sirt3. In human aortic endothelial cells, GCN5L1 depletion prevents superoxide overproduction. Deacetylation mimetic CypD-K166R mice were protected from vascular oxidative stress, endothelial dysfunction, and Ang II-induced hypertension. Ang II-induced hypertension increased mitochondrial GCN5L1 and reduced Sirt3 levels resulting in a 250% increase in GCN5L1/Sirt3 ratio promoting CypD acetylation. Treatment with mitochondria-targeted scavenger of cytotoxic isolevuglandins (mito2HOBA) normalized GCN5L1/Sirt3 ratio, reduced CypD acetylation, and attenuated hypertension. The role of mitochondrial acetyltransferase GCN5L1 in the endothelial function was tested in endothelial-specific GCN5L1 knockout mice. Depletion of endothelial GCN5L1 prevented Ang II-induced mitochondrial oxidative stress, reduced the maladaptive switch of vascular metabolism to glycolysis, prevented inactivation of endothelial nitric oxide, preserved endothelial-dependent relaxation, and attenuated hypertension.

Conclusions: These data support the pathogenic role of CypD acetylation in endothelial dysfunction and hypertension. We suggest that targeting cytotoxic mitochondrial isolevuglandins and GCN5L1 reduces CypD acetylation, which may be beneficial in cardiovascular disease.

Keywords: blood pressure; cardiovascular diseases; hypertension; mitochondria; superoxides.

Figure 1:

Western blot of CypD acetylation, expression of CypD, Sirt3 and GCN5L1 in arterioles isolated from human mediastinal fat in patients with essential hypertension compared with normotensive subjects. CypD acetylation (Acetyl-CypD) was measured by CypD immunoprecipitation and then analyzed with anti-Acetyl-Lysine antibodies. Data were normalized by GAPDH. CypD, Sirt3, and GCN5L1 levels were compared using Student’s t-test and Mann-Witney test was used for Acetyl-CypD. Results are mean ± STD. *P=7.4×10⁻⁸, ^#P=7.7×10⁻⁵ and **P=9.3×10⁻⁶ vs Normotensive.

Figure 2:

Depletion of CypD or GCN5L1 acetylase prevents simulation of mitochondrial ${{\text{O}}_2}^{\underset{\_}{•}}$ but Sirt3 depletion leads to overproduction of mitochondrial ${{\text{O}}_2}^{\underset{\_}{•}}$. HAEC were treated with Ang II (10nM), TNFα (1 ng/ml) plus IL-17A (10 ng/ml) (ATI) for 24 hours and mitochondrial ${{\text{O}}_2}^{\underset{\_}{•}}$ was measured by MitoSOX and HPLC following accumulation of superoxide specific product of MitoSOX, mito-2-hydroxyethidium. Insert shows typical HPLC diagram of MitoSOX samples with superoxide specific product mito-2-hydroxyethidium (2OH-E⁺) and non-specific oxidation product mito-ethidium (E⁺). Data were analyzed using 2-way ANOVA and Tukey post-hoc multiple comparisons. Results are mean ± STD (n=7). *P=5.5×10⁻⁸ vs NS siRNA, **P=8.0×10⁻⁷ vs NS siRNA+ATI, *** P=5.0×10⁻¹² vs NS siRNA+ATI, ^§P=7.0×10⁻⁴ vs NS siRNA, ^#P=1.1×10⁻⁴ vs NS siRNA+ATI.

Figure 3:

Effect of mito2HOBA on CypD acetylation, mitochondrial CypD, GGCN5L1 and Sirt3 levels. Wildtype Sham or Ang II-infused mice were supplied with mito2HOBA in drinking water (0.1 g/L). (A) Typical Western blots of isolevuglandins-protein adducts (isoLGs-K ab), Sirt3, GCN5L1 and Acetyl-CypD in aortic mitochondria normalized by mitochondrial Complex IV. (B-D) Densitometry of Sirt3, GCN5L1 and GCN5L1/Sirt3 ratio. Data were analyzed using 2-way ANOVA and Turkey post-hoc multiple comparisons. Results are mean ± STD. (B) *P=3.9×10⁻⁴ vs Sham, **P=0.018 vs Ang II. (C) *P=2.9×10⁻⁶ vs Sham, *P=2.5×10⁻⁵ vs Ang II. (D) *P=1.0×10⁻¹⁰ vs Sham, ** P=1.5×10⁻⁹ vs Ang II. (E) Systolic blood pressure. Data were analyzed using 2-way ANOVA with repeated measurements. Results are mean ± STD. *P=0.0001 vs Sham, **P=0.001 vs Ang II (n=8).

Figure 4:

Deacetylation mimic CypD-K166R reduces mPTP opening (A) and improves mitochondrial respiration (B). (A) Addition of CaCl₂ to mitochondria above Ca²⁺ retention capacity led to mPTP opening and mitochondria swelling. *P=0.0013 vs WT+Suc, **P=0.0046 vs WT+Suc, ^#P=6.7×10⁻⁵ vs WT+Suc+PC. (B) Respiratory control ratio (State 3/State 4 ratio) was measured in isolated kidney mitochondria with succinate (Suc) or succinate + palmitoyl-carnitine (Suc+PC). *P=0.040 vs WT+Suc, **P=0.043 vs WT+Suc, ^#P=1.2×10⁻⁴ vs WT+Suc+PC. Data were analyzed using 2-way ANOVA with a Tukey post-hoc multiple comparisons. Results are mean ± STD.

Figure 5:

Systolic blood pressure (A), endothelial dependent relaxation (B) and mitochondrial superoxide (C, D) in wildtype and CypD-K166R mice. (A) Blood pressure was measured by telemetry. *P=1.0×10⁻¹⁰ vs WT, **P=3.6×10⁻⁵ vs WT+Ang II. (B) Relaxation was measured by isometric tension. Aortic relaxation data were analyzed using 2-way ANOVA with repeated measurements. **P=0.002 vs WT+Ang II. (C) Mitochondrial superoxide was measured by mitoSOX/HPLC. *P=6.2×10⁻⁷ vs WT, **P=6.1×10⁻⁵ vs WT+Ang II. Superoxide data were analyzed using 2-way ANOVA and Tukey post-hoc multiple comparisons. (D) Aortas were incubated ex vivo in DMEM for 24-hours prior to superoxide analysis. *P=4.9×10⁻⁹ vs WT, **P=2.6×10⁻⁷ vs WT+Ang II+IL17A+TNFα. Results are mean ± STD (n=8).

Figure 6:

Systolic blood pressure, aortic relaxation, mitochondrial ${{\text{O}}_2}^{\underset{\_}{•}}$ and endothelial nitric oxide in aorta from Sham or Ang II-infused WT and Ec^GCN5L1KO mice (Ang II 0.7 mg/kg/day, 14 days). (A) Blood pressure was measured by telemetry. P=0.001 vs WT, **P=0.0001 vs WT+Ang II. (B) Endothelial-dependent relaxation was measured by myography. P=5.2×10⁻⁹ vs WT, **P=4.8×10⁻⁸ vs WT+Ang II. Blood pressure and myography data were analyzed using 2-way ANOVA with repeated measurements. (C) Mitochondrial ${{\text{O}}_2}^{\underset{\_}{•}}$ was measured by MitoSOX and HPLC in isolated aortas. P=4.6×10⁻⁷ vs WT, **P=0.00031 vs WT+Ang II. (D) Endothelial NO was analyzed by NO spin trap FeDETC₂ and EPR (insert). P=1.0×10⁻¹⁰ vs WT, **P=4.4×10⁻⁸ vs WT+Ang II. Superoxide and NO data were analyzed using 2-way ANOVA and Tukey post-hoc multiple comparisons. Results are mean ± STD.

Figure 7:

Analysis of glycolysis by lactate production in aortas from Sham and Ang II-infused wildtype, CypD-K166R and Ec^GCN5L1KO mice. Mice were infused with Ang II (0.7 mg/kg/day) for 14 days. Isolated intact and denuded (DN) aortas were placed in DMEM organoid culture for analysis of lactate. Data were analyzed using 2-way ANOVA and Tukey post-hoc multiple comparisons. (A) *P=1.0×10⁻¹⁵ vs WT, **P=1.6×10⁻⁸ vs WT+Ang II, ***P=2.5×10⁻⁸ vs WT+Ang II. (B) ^#P=0.01 vs WT Intact, *P=1.0×10⁻¹⁵ vs WT, **P=1.0×10⁻¹⁵ vs WT+Ang II Intact, ^§P=0.045 vs Ec^GN5L1KO Intact, ***P=1.0×10⁻¹⁵ vs WT+Ang II Intact, ^&P=0.047 vs Ec^GN5L1KO+Ang II Intact. Results are mean ± STD.