Cardiomyocyte-specific deletion of the mitochondrial transporter Abcb10 causes cardiac dysfunction via lysosomal-mediated ferroptosis

Abstract

Heart function is highly dependent on mitochondria, which not only produce energy but also regulate many cellular functions. Therefore, mitochondria are important therapeutic targets in heart failure. Abcb10 is a member of the ABC transporter superfamily located in the inner mitochondrial membrane and plays an important role in haemoglobin synthesis, biliverdin transport, antioxidant stress, and stabilization of the iron transporter mitoferrin-1. However, the mechanisms underlying the impairment of mitochondrial transporters in the heart remain poorly understood. Here, we generated mice with cardiomyocyte-specific loss of Abcb10. The Abcb10 knockouts exhibited progressive worsening of cardiac fibrosis, increased cardiovascular risk markers and mitochondrial structural abnormalities, suggesting that the pathology of heart failure is related to mitochondrial dysfunction. As the mitochondrial dysfunction was observed early but mildly, other factors were considered. We then observed increased Hif1α expression, decreased NAD synthase expression, and reduced NAD+ levels, leading to lysosomal dysfunction. Analysis of ABCB10 knockdown HeLa cells revealed accumulation of Fe2+ and lipid peroxides in lysosomes, leading to ferroptosis. Lipid peroxidation was suppressed by treatment with iron chelators, suggesting that lysosomal iron accumulation is involved in ferroptosis. We also observed that Abcb10 knockout cardiomyocytes exhibited increased ROS production, iron accumulation, and lysosomal hypertrophy. Our findings suggest that Abcb10 is required for the maintenance of cardiac function and reveal a novel pathophysiology of chronic heart failure related to lysosomal function and ferroptosis.

Keywords: heart failure; lysosomes; mitochondria.

Figure 1. Abcb10 deletion in mouse cardiomyocytes causes cardiac dysfunction and shortened lifespan

(A) Western blot and real-time qPCR analysis of Abcb10/Abcb10 expression in WT and Abcb10cKO hearts at 6 months. Gapdh was used as internal control (WT, n=4; Abcb10 cKO, n=6). (B) Relative mRNA expression of cardiac failure markers Anf and β-mhc in WT and Abcb10cKO mouse hearts at 6, 10, and 12 months of age (WT, n=4–6; Abcb10 cKO, n=5–6). (C) Representative images of Masson's trichrome staining of heart sections from WT and Abcb10 cKOs at 10 months. Fibrosis in hearts was quantified by measuring the blue staining per tissue area (6 sections per sample). Scale bars for whole heart sections = 500 µm; scale bars for sections below = 100 µm. (WT, Abcb10 cKO: n=4). (D) Heart weight-to-body weight ratios of 9- and 11-month-old mice (WT, n=5; Abcb10 cKO, n=4). (E) Left ventricular diameter to total diameter ratio of ventricle mid-region heart cross sections (WT, Abcb10 cKO: n=3). (F) Survival curve for male WT and Abcb10 cKO mice (WT male: n=17, female: n=31; Abcb10 cKO male: n=18, female: n=31). In A–E, error bars are presented as mean ±SD. Statistical significance was assessed using the Student’s t-test; *P<0.05, **P<0.01, ***P<0.001.

Figure 2. Functional characteristics of mitochondria in Abcb10 cKO hearts

(A) Quantitative real-time qPCR expression analysis of Fgf21 and Gdf15 (biomarkers for mitochondrial disorders) in the hearts of 6- to 12-month-old WT and Abcb10 cKO mice (WTs, n=4–6; Abcb10 cKOs, n=5–6). (B) Quantitative real-time qPCR expression of integrated stress gene (Atf3, Atf4, Chop and Trib3) in heart from 10-month-old Abcb10 cKO and WT mice (WT: n=4, Abcb10cKO: n=6). (C) Western analysis of OXPHOS proteins in the hearts of 10-month-old WT and Abcb10 cKO mice. CoxI and CoxIII (mitochondrial DNA encoded, Complex IV), Ndufa9 and Ndufb8 (complex I), Sdha (complex II), Uqcrc1 (complex III) Atp5a (complexV). GAPDH was used as an internal control (WT, n=4; Abcb10 cKO, n=6). (D) Mitochondrial oxygen consumption rates (OCRs) in mitochondrial fractions from WT and Abcb10 cKO hearts from 8 months old. Results represent mean ± SD (WT, Abcb10 cKO, n=3). In A–D, error bars are presented as mean ±SD. Statistical significance was assessed using the Student’s t-test; *P<0.05, **P<0.01, ***P<0.001.

Figure 3. Structural characteristics of mitochondria in Abcb10 cKO hearts

(A) Representative electron microscopy images of the arrangement and area of mitochondria in heart tissue from 10-month-old WT and Abcb10 cKO mice. Quantification was performed by determining mitochondria area/total area per sheet (5–9 sheets per group); scale bars, 1 µm. (B) Representative electron microscopy images of heart tissue from 10-month-old WT and Abcb10 cKO mice and quantification of abnormal mitochondria (30 sheets per group). (C) Western analysis of mitochondrial fission and fusion proteins in hearts of Abcb10 WT and cKO mice from 10-month-old. Gapdh was used as an internal control (WT: n=4, Abcb10 cKO: n=6). (D) Western analysis of mitochondrial autophagy proteins Pink1 in hearts of WT and Abcb10 cKO mice from 10-month-old. (WT: n=4, Abcb10 cKO: n=6). In A-D, error bars are presented as mean ± SD. Statistical significance was assessed using the Student’s t-test; *P<0.05, **P<0.01.

Figure 4. Reduced NAD biosynthesis and NAD⁺ levels in Abcb10 cKO hearts

(A) LC-MS/MS metabolic analysis of nicotinamide adenine dinucleotide (NAD⁺), nicotinamide adenine dinucleotide, reduced form (NADH), nicotinamide adenine dinucleotide monophosphate (NADP), nicotinamide adenine dinucleotide monophosphate, reduced form (NADPH), nicotinamide mononucleotide (NaMN), and nicotinate in hearts from10-month-old WT and Abcb10 cKO mice (WT, n=4; Abcb10 cKO, n=6). (B) Quantitation of Nampt, Nmnat1, and Nmnat3 mRNA levels in hearts from 6- to 10-month-old WT and Abcb10 cKO mice, measured by real time qPCR (WT, n=4; Abcb10 cKO, n=4–6). (C) Western blotting of NAD-synthesizing enzymes and Hif-1α in hearts from 10-month-old WT and Abcb10 cKO mice. Gapdh was used as an internal control (WT, n=4; Abcb10 cKO, n=6). (D) ROS production by MitoSOX fluorescent probe from WT and Abcb10 knockout cardiomyocyte. Twenty cardiomyocytes were measured per mouse heart. Scale bar, 50 µm (WT, n=3; Abcb10 cKO, n =3). (E) Relative protein levels of oxidative stress protein 4-HNE and 3-nitrotyrosine (WT, n=4; Abcb10 cKO, n=6). In A–E, error bars are presented as mean ±SD. Statistical significance was assessed using the Student’s t-test; *P<0.05, **P<0.01, ***P<0.001.

Figure 5. Cardiac-specific Abcb10 deficiency causes impaired lysosomal function and accumulation of p62 in hearts

(A) Immunostaining of Lamp 2 (red) in heart sections with DAPI-stained nuclei (blue). Western blot showing increased Lamp 2 in 10-month-old WT and Abcb10 cKO hearts. Gapdh was used as an internal control. Error bars are presented as means ± SD (WT, n=4; Abcb10 cKO: n=6). (B) Immunostaining of Lamp2 in cardiomyocyte from each mouse heart. Relative Lamp2 intensity was measured in twenty cardiomyocyte from each mouse heart; scale bar, 50 µm. Error bars are presented as means ± SD (WT: n=3, Abcb10 cKO: n=3). (C) Detection of lipofuscin granules around the nuclei by autofluorescence in heart sections. Tissues were excited at a wavelength of 540 or 470 nm, and emission spectra were collected using a confocal microscope at wavelengths of 580–630 nm or 510–560 nm. Quantification of the ratio of autofluorescence appearance per DAPI staining area is presented in the right panel; scale bar, 20 µm. Error bars are presented as means ± SEM (WT and Abcb10 cKO, n =4; 12 sheets per group). (D) Immunostaining of Lamp 2 (red) and Galectin 3 (green) in 10-month-old WT and Abcb10 cKO hearts; scale bar, 20 µm. Quantification on the right shows the increased co-localization of Lamp 2 and galectin 3 in Abcb10 cKO hearts. Error bars are presented as means ± SD (WT and Abcb10 cKO, n=4; 10 sheets per group). (E) Western blot showing increased Galectin3 protein level in lysosomal fraction of 11-month-old WT and Abcb10 cKO heart. Vdac, located in the mitochondrial outer membrane, is used as loading control (WT: n=4, Abcb10 cKO: n=5). (F) Western blotting of Capthepsin B, Cathepsin D and Gapdh in 10-month-old WT, Abcb10 cKO hearts (WT: n=4, Abcb10 cKO: n=6). (G) The accumulation of autophagic marker protein p62 in 10-month-old Abcb10 cKO hearts. Gapdh was used as an internal control. Error bars are presented as means ± SD (WT, n=4; Abcb10 cKO, n=6).

Figure 6. Abcb10 deficiency results in decreased GSH/GSSG ratios and induces ferroptosis

(A) The mRNA levels of ferroptosis-related genes (Ptgs2, Chac1, and Ho-1) in heart tissue from 10-month-old WT and Abcb10 cKO mice were determined by real time-PCR (WT, n=4; Abcb10 cKO, n =6). (B) Western blot analysis of the levels of GPX4 and TFRC in 10-month-old WT and Abcb10 cKO hearts. Gapdh was used as an internal control (WT, n=4; Abcb10 cKO, n=6). (C) LC-MS/MS metabolomic analysis of GSH and GSSG in WT and Abcb10 cKO hearts. GSH/GSSG ratios were decreased in 10-month-old Abcb10 cKO hearts (WT, n=4; Abcb10 cKO, n=5). (D) Representative image of FerroOrange (cytoplasmic iron level) in cardiomyocyte from WT and Abcb10 cKO heart. Thirty cardiomyocytes from each mouse heart were measured; scale bar, 50 µm (WT and Abcb10 cKO, n=3). (E) Representative image of Liperfluo (lipidperoxide) in cardiomyocyte from WT and Abcb10 cKO heart. Forty cardiomyocytes from each mouse heart were measured; scale bar, 50 µm (WT and Abcb10 cKO, n=2). (F) Predicted pathogenesis mechanism of dilated cardiomyopathy in Abcb10 cKO mice. In A–E, error bars are presented as mean ± SD. Statistical significance was assessed using the Student’s t-test; *P<0.05, **P<0.01, ***P<0.001.

Figure 7. ABCB10 siRNA treatment causes intracellular iron accumulation and triggers lysosomal lipid peroxidation

(A) The mRNA levels of ferroptosis-related genes in HeLa cells treated with ABCB10 siRNA for 72 h (n=3). (B) Western blot analysis of ferroptosis-related proteins in HeLa cells treated with ABCB10 siRNA for 72 h. GAPDH was used as an internal control (n=3). (C) Detection of intracellular GSH and GSSH concentrations in HeLa cells treated with ABCB10 siRNA for 72 h. The GSH/GSSG ratio was decreased in ABCB10 siRNA-treated cells (n=3). (D) Intracellular Fe²⁺ was detected by FerroOrange using fluorescence microscopy. The relative fluorescence intensity of FerroOrange was increased in HeLa cells treated with ABCB10 siRNA (n=11); scale bar: 20 µm. (E) Detection of mitochondrial Fe²⁺ using Mito-FerroGreen in HeLa cells treated for 72 h with ABCB10 siRNA. The Mito-FerroGreen fluorescence signals were of similar intensity in WT and ABCB10 siRNA-treated cells (n=11); scale bar: 20 µm. (F) Representative confocal images of WT cells and HeLa cells treated with ABCB10 siRNA for 72 h and stained with LipiRADICAL Green (detection reagent for lipid radicals). LipiRADICAL Green relative fluorescence intensity was increased in ABCB10 siRNA-treated cells (n=28); scale bar: 20 µm. (G) LipiRADICAL Green and LysoTracker Red were used to co-stain HeLa cells treated with ABCB10 siRNA for 72 h. Representative colocalization images of lipid peroxidation in lysosomes; scale bar: 20 µm. (H) Intracellular lipid radical staining by LipiRADICAL Green in HeLa cells. WT: untreated, WT+DFO: HeLa and addition of 100 μM DFO for 3 h, ABCB10 KD: ABCB10 knockdown, ABCB10+DFO: ABCB10 knockdown and addition of 100 μM DFO for 3 h. The right panel shows the intensity of LipiRADICAL Green per cell; scale bar: 20 μm (n=13–19). Error bars are presented as mean ± SD. One-way ANOVA with Tukey’s multiple comparisons test, *P<0.05, **P<0.01, ***P<0.001. (I) Schematic diagram of a potential ferroptosis pathway in Abcb10 cKO hearts. Red arrows indicate increases or decreases in Abcb10 cKO, compared with levels in WT. In A–F, error bars are presented as mean ± SD. Statistical significance was assessed using the Student’s t-test; *P<0.05, **P<0.01, ***P<0.001.