Mitochondrial oxidative stress contributes to diastolic dysfunction through impaired mitochondrial dynamics

Abstract

Diastolic dysfunction (DD) underlies heart failure with preserved ejection fraction (HFpEF), a clinical syndrome associated with aging that is becoming more prevalent. Despite extensive clinical studies, no effective treatment exists for HFpEF. Recent findings suggest that oxidative stress contributes to the pathophysiology of DD, but molecular mechanisms underpinning redox-sensitive cardiac remodeling in DD remain obscure. Using transgenic mice with mitochondria-targeted NOX4 overexpression (Nox4TG618) as a model, we demonstrate that NOX4-dependent mitochondrial oxidative stress induces DD in mice as measured by increased E/E', isovolumic relaxation time, Tau Glantz and reduced dP/dt_min while EF is preserved. In Nox4TG618 mice, fragmentation of cardiomyocyte mitochondria, increased DRP1 phosphorylation, decreased expression of MFN2, and a higher percentage of apoptotic cells in the myocardium are associated with lower ATP-driven and maximal mitochondrial oxygen consumption rates, a decrease in respiratory reserve, and a decrease in citrate synthase and Complex I activities. Transgenic mice have an increased concentration of TGFβ and osteopontin in LV lysates, as well as MCP-1 in plasma, which correlates with a higher percentage of LV myocardial periostin- and ACTA2-positive cells compared with wild-type mice. Accordingly, the levels of ECM as measured by Picrosirius Red staining as well as interstitial deposition of collagen I are elevated in the myocardium of Nox4TG618 mice. The LV tissue of Nox4TG618 mice also exhibited increased I_CaL current, calpain 2 expression, and altered/disrupted Z-disc structure. As it pertains to human pathology, similar changes were found in samples of LV from patients with DD. Finally, treatment with GKT137831, a specific NOX1 and NOX4 inhibitor, or overexpression of mCAT attenuated myocardial fibrosis and prevented DD in the Nox4TG618 mice. Together, our results indicate that mitochondrial oxidative stress contributes to DD by causing mitochondrial dysfunction, impaired mitochondrial dynamics, increased synthesis of pro-inflammatory and pro-fibrotic cytokines, activation of fibroblasts, and the accumulation of extracellular matrix, which leads to interstitial fibrosis and passive stiffness of the myocardium. Further, mitochondrial oxidative stress increases cardiomyocyte Ca²⁺ influx, which worsens CM relaxation and raises the LV filling pressure in conjunction with structural proteolytic damage.

Keywords: Cardiomyocyte; Diastolic dysfunction; Heart failure with preserved ejection fraction; Interstitial fibrosis; Mitochondrial dysfunction; NADPH oxidase 4; NOX4 inhibitor.

Graphical abstract

Fig. 1

Increased expression of NOX4 is associated with mitochondrial oxidative stress and dysfunction in LV myocardial samples from patients with a history of diastolic dysfunction. (A) Representative Western Blot analysis and quantification of NOX4 expression in human LV myocardial samples. (B) Representative immunofluorescence images and quantifications of NOX4. (C) MitoSOX and (D) ATP5G2/8-OHDG staining in human LV myocardial cross sections. (E) Citrate synthase activity was measured in protein lysates of LV myocardium. The data are presented as mean ± SEM, N = 3. Scale is 100 μm.

Fig. 2

Increased mitochondrial NOX4 expression leads to cardiac diastolic dysfunction in Nox4TG618 mice. Systolic and diastolic function parameters in wild-type and Nox4TG618 mice (A–G) were assessed by echocardiography. (A) Representative pulsed wave and tissue Doppler images. (B) Mitral valve (MV) E velocity (MV E), (C) MV septal annulus velocity E′ (E′), (D) ratio of MV E velocity to E′ velocity at septal annulus (E/E′), (E) isovolumic relaxation time (IVRT), (F) left atrial volume (LA vol), (G) ejection fraction (EF), (H) fractional shortening, (I) LV end-diastolic volume (LV EDV), and (J) mitral valve E to A ratio (N = 15). Pressure-volume loops analysis (K–O). (K) end-diastolic pressure (EDP), (L) end diastolic pressure volume relationship (EDPVR), (M) Tau Glantz (Τ Glantz), (N) rate of ventricular pressure decrease (dP/dT_min), and (O) ejection fraction (EF); (N = 7). The data presented as mean ± SEM.

Fig. 3

Increased mitochondrial NOX4 expression in mice increases mitochondrial oxidative stress, mtDNA damage, and leads to mitochondrial dysfunction in LV cardiomyocytes. Representative immunofluorescence images and quantification of LV cross sections from wild-type and Nox4TG618 mice stained for (A) NOX4 and counterstained with cardiomyocyte marker MYH, (B) MitoSOX Red, and (C) MitoPY1, (N = 9). (D) Mouse LV H₂O₂ levels were measured by Amplex Red assay (N = 7). (E) ATP5G2/8-OHdG colocalization, (N = 9). (F) Representative long PCR amplicon showing amplification of 10 kb mitochondrial DNA in mouse LV samples. The amplified mitochondrial DNA band was normalized to a short PCR amplicon of 0.22 kb and the band intensity was quantified by densitometry. (G) Data represent relative mitochondrial DNA copy number in LV tissues of wild-type and transgenic mice (N = 4). (H) Citrate synthase activity was measured in mouse LV lysates (N = 10). (I) Complex I activity measured in mouse LV lysates (N = 10). (J) Mitochondrial oxygen consumption rate (OCR) measured by Seahorse analysis (N = 5). (K) Mitochondrial bioenergetics parameters derived from OCR, (N = 5). The data are presented as mean ± SEM. Scale is 100 μm. . (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4

Mitochondrial dynamics are impaired in the left ventricle of Nox4TG618 versus wild-type mice. (A) Representative TEM images and quantification of mitochondrial size in mouse LV cross sections (N = 3). Representative Western blot analysis and quantification of proteins involved in mitochondrial fission (B), fusion (C) and mitophagy (D); (N = 4). (E) Representative immunofluorescent images and quantification of mouse LV sections stained with LAMP1/ATP5G2, (N = 9). (F) Representative comet assay images of cardiomyocytes isolated from LV samples and quantification of comet tails in Nox4TG618 versus wild-type mice (N = 3). (G) Flow cytometric analysis of LV cardiomyocyte apoptosis, % of cells positive for Annexin V (N = 5). Data presented as mean ± SEM. Scale is 100 μm.

Fig. 5

Increased mitochondrial NOX4 expression in Nox4TG618 mice is associated with upregulated L-type Ca²⁺current (I_CaL), increased expression of calpain 2 and fractured sarcomeric Z-discs in cardiomyocytes. (A) Representative recordings of the I_CaL in wild-type and Nox4TG618 mice elicited at 10 mV, (B) voltage-current curves (Vm), (C) conductance (G), (D) normalized conductance (G/Gmax), and (E) membrane capacitance. For I_CaL recordings, cardiomyocytes (n = 27 wild-type and 30 Nox4TG618) from 4 mice of each genotype were used. (F) A normalized weight for the heart based on the length of the tibia; (N = 4). (G) Representative Western blot analysis and quantification of expression of calpain1/2 (CAPN1/2) and calpastatin (CAST) (N = 6) in LV lysates. (H) Representative TEM images of murine LV cross sections with focus on Z-disc (N = 3). The data are presented as mean ± SEM.

Fig. 6

Myofibroblast activation and ECM synthesis is increased in the LV of Nox4TG618 mice. A) LV fibrosis was measured by staining LV cross sections with picrosirius red (N = 9). (B) Immunofluorescent analysis and quantification of collagen I deposition (N = 9). Myofibroblast activation in LV cross sections was measured by immunofluorescent staining for periostin (C), and ACTA2 (N = 9) (D). LV cross sections were counterstained with cardiomyocyte marker MYH7. Representative images of LV cross sections stained with WGA and quantification of cardiomyocyte cross-sectional area (CSA) (N = 9) (E). Concentrations of (F) TGFβ (N = 13) and (G) osteopontin (OPN, N = 7) in LV myocardium lysates, and (H) MCP-1 (N = 15) in mouse plasma. The data are presented as mean ± SEM. Scale is 100 μm. . (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 7

Mitochondrial overexpression of human catalase inhibits mitochondrial oxidative stress, reduces activation of myofibroblasts, ECM synthesis, and preserves diastolic function in Nox4TG618 mice. Representative images and quantification of (A) MitoSox Red and (B) ATP5/8-OHDG. (C–I) Echocardiographic assessment of systolic and diastolic function. (C) Ratio of MV E velocity to E′ velocity at septal annulus (E/E′), (D) left atrial volume (LA vol), (E) isovolumic relaxation time (IVRT), (F) ejection fraction (EF), (G) fractional shortening (FS), (H) LV end-diastolic volume (LV EDV), and (I) ratio of mitral valve E velocity to A velocity (E/A) (N = 7). Representative immunofluorescence images and quantification of (J) picrosirius red staining, (K) Collagen I, (L) periostin and (M) ACTA2 in mouse LV cross sections, counterstained with cardiomyocyte marker MYH (N = 7). The data are presented as mean ± SEM. Scale is 100 μm. . (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 8

Treatment with GKT137831 inhibits mitochondrial oxidative stress, decreases fibrosis and prevents development of diastolic dysfunction in Nox4TG618 mice. (A–G) Echocardiographic assessment of systolic and diastolic function. (A) Ratio of MV E velocity to E′ velocity at septal annulus (E/E′), (B) left atrial volume (LA vol), (C) isovolumic relaxation time (IVRT), (D) ejection fraction (EF), (E) fractional shortening (FS), (F) LV end-diastolic volume (LV EDV), and (G) ratio of mitral valve E velocity to A velocity (E/A) (N = 7). Representative images and quantification of (H) MitoSox Red staining, (I) ATP5G2/8-OHDG colocalization, (J) picrosirius red staining, (K) collagen I, (L) periostin and (M) ACTA2 in mouse LV sections, counterstained with cardiomyocyte marker MYH (N = 7). The data are presented as mean ± SEM. Scale is 100 μm. . (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 9

LV samples of diastolic dysfunction patients show increased fibrosis and myofibroblast activation/Representative immunofluorescent images and quantification of (A) picrosirius red staining, (B) collagen I, (C) periostin, and (D) ACTA2 in human LV cross sections and counterstained for cardiomyocyte marker MYH. Western blot analysis and quantification of expression of proteins involved in mitochondrial fission (E), fusion (F), and of calpain 1/2 (CAPN1/2) and calpastatin (CAST) (G) in human LV myocardial lysates. The data are presented as mean ± SEM, N = 3. Scale is 100 μm. . (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)