The E3 ubiquitin ligase MARCH2 protects against myocardial ischemia-reperfusion injury through inhibiting pyroptosis via negative regulation of PGAM5/MAVS/NLRP3 axis

Abstract

Inflammasome activation and pyroptotic cell death are known to contribute to the pathogenesis of cardiovascular diseases, such as myocardial ischemia-reperfusion (I/R) injury, although the underlying regulatory mechanisms remain poorly understood. Here we report that expression levels of the E3 ubiquitin ligase membrane-associated RING finger protein 2 (MARCH2) were elevated in ischemic human hearts or mouse hearts upon I/R injury. Genetic ablation of MARCH2 aggravated myocardial infarction and cardiac dysfunction upon myocardial I/R injury. Single-cell RNA-seq analysis suggested that loss of MARCH2 prompted activation of NLRP3 inflammasome in cardiomyocytes. Mechanistically, phosphoglycerate mutase 5 (PGAM5) was found to act as a novel regulator of MAVS-NLRP3 signaling by forming liquid-liquid phase separation condensates with MAVS and fostering the recruitment of NLRP3. MARCH2 directly interacts with PGAM5 to promote its K48-linked polyubiquitination and proteasomal degradation, resulting in reduced PGAM5-MAVS co-condensation, and consequently inhibition of NLRP3 inflammasome activation and cardiomyocyte pyroptosis. AAV-based re-introduction of MARCH2 significantly ameliorated I/R-induced mouse heart dysfunction. Altogether, our findings reveal a novel mechanism where MARCH2-mediated ubiquitination negatively regulates the PGAM5/MAVS/NLRP3 axis to protect against cardiomyocyte pyroptosis and myocardial I/R injury.

Fig. 1. MARCH2 protein level is increased in hearts of patients with ICM and mice during myocardial I/R.

a Heatmap showing six differentially expressed E3 ligases in mouse hearts following I/R injury. b Histogram showing the relative mRNA expression of the six ubiquitin-related genes in the indicated groups. c mRNA levels of MARCH2 following I/R injury (45 min/9 h) were analyzed by RT-qPCR (normalized to β-actin). d Western blotting analysis of MARCH2 expression in healthy human hearts and patients with ICM. e Quantitated MARCH2 levels in hearts of healthy human and patients with ICM (n = 6 for each group). f Western blotting analysis of MARCH2 protein levels following I/R injury (45 min/9 h). g Quantitated MARCH2 levels following I/R injury (45 min/9 h). h Representative immunohistochemistry images showing MARCH2 protein levels in human hearts. i Representative immunohistochemistry images showing MARCH2 protein levels in sham- or I/R-treated mouse hearts. Dotted white lines indicate the boundary of the infarct area and border area. j, k Representative western blotting image (j) and quantification analysis (k) of MARCH2 expression in NMCMs subjected to hypoxia (6 h)/reoxygenation (3 h) (n = 6 for each group). l Western blotting analysis showed that MARCH2 protein could not be detected following the deletion of MARCH2 gene. m, n CK-MB level (m) and LDH activity (n) of WT and MARCH2 KO mice with or without myocardial I/R (45 min/24 h) (n = 6 for each group). o Representative images of heart sections by TTC/Evans Blue staining depicting infracted area. p Ratios of area at risk (AAR) to left ventricular (LV) area. q Infarct area normalized to AAR. Data are shown as means ± SEM. Statistical significance was examined by two-way ANOVA with Bonferroni post-test. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 2. MARCH2 KO exacerbates myocardial dysfunction after myocardial I/R injury.

a Representative M-mode recordings of echocardiography of WT and MARCH2 KO mice subjected to I/R. b–f Quantitative analyses of echocardiographic measurements performed after I/R injury. n = 6 mice per group. Left ventricular end-systolic diameter (LVESD, b), left ventricular end-diastolic diameter (LVEDD, c), left ventricular end-systolic volume (LVESV, d), ejection fraction (EF, e), and fractional shortening (FS, f). g–l Evaluation of mechanical properties of single, acutely isolated cardiomyocytes from I/R-treated mice. Resting cell length (g), peak shortening (h), +dL/dt: maximal velocity of shortening (i), -dL/dt: maximal velocity of re-lengthening (j), TPS₉₀: time-to-90% peak shortening (k); TR₉₀: time-to-90% re-lengthening (l). n = 60 cells from 3 mice per group. m Representative PI-stained images of myocardial sections from WT and MARCH2^–/– mice subjected to I/R. Green: PI-positive nuclei; Red: cTnI-stained cardiomyocytes; blue, DAPI-stained nuclei; Scale bar = 20 μm. n Quantitative analysis of PI-positive cells. Experiments were repeated three times with similar results. Data are shown as the means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 3. PGAM5 is identified to interact with MARCH2.

a Comparison of cell type distributions among different groups showing changes in MP and cardiomyocyte clusters after I/R procedure. b, c KEGG enrichment analysis (b) and GO (c) analysis of scRNA-seq data from WT-I/R and MARCH2^–/–-I/R hearts. d, e Interaction between NLRP3 and ASC in mitochondria (mito-NLRP3 and mito-ASC) of cardiac tissues of MARCH2 KO and WT mice subjected to I/R was examined by IP-western blotting assay. IP with NLRP3 antibody (d) and IP with ASC antibody (e). f Representative Western blotting analyses of MARCH2, NLRP3, ASC, caspase-1 (procaspase1; cleaved caspase-1), GSDMD (full-length; N-terminal) in cardiac tissues of MARCH2 KO and WT mice subjected to I/R (45 min/9 h). g IL-18 release was measured by enzyme-linked immunosorbent assay (ELISA) in cardiac tissues of MARCH2 KO and WT mice subjected to I/R (45 min/9 h). h Schematic diagram showing MS analysis workflow for identifying targets of MARCH2. i IP analysis with anti-HA antibody and immunoblotting with antibodies of anti-Flag and anti-HA, respectively, in NMCMs transfected with PGAM5-HA or control vector along with MARCH2-Flag. j Endogenous IP of MARCH2 and PGAM5 in NMCMs. k Direct interaction between GST-PGAM5 and His-MARCH2 demonstrated by GST pull-down assays. Both input and pull-down samples were immunoblotted with anti-GST and anti-His antibodies. GST-PGAM5 and His-MARCH2 proteins were expressed in vitro. l Representative immunofluorescence images of MARCH2 and PGAM5 in NMCMs. Data are shown as the means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 4. MARCH2 E3 ligase promotes degradation of PGAM5 through K48-linked polyUb.

a Western blotting analysis of PGAM5-HA and MARCH2-Flag. HL-1 cardiomyocytes were transfected with vector expressing PGAM5-HA or MARCH2-Flag and treated with MCM + H/R. b Western blotting analysis of MARCH2 and PGAM5 proteins in HL-1 cardiomyocytes transfected with MARCH2 siRNA, siRNA-resistant MARCH2 or negative control and treated with MCM + H/R. c Analysis of PGAM5-HA protein levels in MCM + H/R-treated HL-1 cardiomyocytes transfected with either MARCH2 WT or its catalytically inactive mutant (C64/67 S, abbreviated CS). d, e Western blotting analysis of MARCH2 and PGAM5 protein levels and quantification analysis of PGAM5 level. HL-1 cardiomyocytes were transfected with MARCH2 (d) or siRNA-MARCH2 (e) and then treated with MCM + H/R and cycloheximide (CHX; 30 μm). f, g Corresponding quantitation graphs of relative PGAM5 degradation in MCM + H/R-treated HL-1 cardiomyocytes transfected with MARCH2 (f) or siRNA-MARCH2 (g) and CHX. The experiments were repeated three times. Error bars represent standard deviation. h Effects of the indicated polyUb on MARCH2-mediated PGAM5 ubiquitination. HL-1 cardiomyocytes were transfected with the indicated ubiquitin under MCM + H/R treatment. IP analysis with anti-HA antibody and immunoblotting with antibodies of anti-His and anti-HA. i Western blotting of the mapping analysis showing the binding domains of MARCH2 to PGAM5. Various truncated forms of MARCH2 (△1–55, △56–116, △117–214 and △215–246) were expressed and purified. j–l Effects of the indicated PGAM5 KR mutants (K88R, K141R, or combined K88-141R) on MARCH2-mediated PGAM5 ubiquitination (j), protein level (k), and degradation (l) under MCM + H/R treatment. PGAM5^–/– HL-1 cardiomyocytes were transfected with the indicated constructs, and PGAM5 protein levels were analyzed. Statistical differences were analyzed by unpaired Student’s t-test for comparison between two groups. *P < 0.05, **P < 0.01.

Fig. 5. PGAM5 mediates the regulation of MARCH2 on NLRP3 inflammasome assembly following myocardial I/R.

a Pearson correlation coefficients between MARCH2 and PGAM5 protein levels in ICM. b, c Representative western blotting (b) and quantitated analysis (c) of PGAM5 expression in NMCMs infected with vector or MARCH2 under MCM + H/R treatment. d Ubiquitination assays determining the ubiquitination of endogenous PGAM5 in the hearts of MARCH2 KO and WT mice subjected to I/R injury. e Immunohistochemistry for MARCH2, PGAM5, and GSDMD in hearts of MARCH2 KO and WT mice following I/R injury. f, g Representative western blotting image (f) and quantitated analysis (g) of PGAM5 in cardiac tissues of MARCH2 KO and WT mice subjected to I/R (45 min/9 h). h, i Interaction between NLRP3 and ASC was examined by IP-western blotting assay in cardiomyocytes with MARCH2 or PGAM5 overexpression following MCM + H/R challenge. IP with NLRP3 antibody (h), IP with ASC antibody (i). j Interaction between PGAM5 and MAVS was examined by IP-western blotting assay in cardiomyocytes with or without MCM + H/R treatment. k Representative immunofluorescence images of MAVS and PGAM5 in HL-1 cells. l Colocalization analysis of PGAM5–MAVS. Pearson’s R value (no threshold) was calculated by ImageJ Fiji software. n = 6 images from 3 biological replicates. Data are shown as the means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 6. PGAM5–MAVS co-condensates form under MCM + H/R challenge and PGAM5 promotes MAVS-dependent NLRP3 activation.

a Time-lapse images of HL-1 cells expressing PGAM5-mcherry. PGAM5 condensate fission and fusion is presented in the boxes. b FRAP analysis of PGAM5-mcherry condensates in HL-1 cells. c Quantification of FRAP in the bleached region of PGAM5-mcherry condensates, show as means ± SD (n = 6). d Time-lapse images of HL-1 cells expressing MAVS-EGFP. MAVS condensate fission and fusion is presented in the boxes. e FRAP analysis of MAVS-EGFP condensates in HL-1 cells. f Quantification of FRAP in the bleached region of PGAM5-mcherry condensates, show as means ± SD (n = 6). g FRAP analysis of MAVS-EGFP condensates in HL-1 cardiomyocytes without (upper) or with (lower) the presence of PGAM5. h Quantification of FRAP in the bleached region of MAVS-EGFP condensates with or without PGAM5, show as means ± SD (n = 6). i Representative fluorescent images of MAVS-EGFP droplets in the presence or absence of PGAM5. j Quantification of fluorescence intensity of MAVS-EGFP liquid droplets in i. k Interaction between NLRP3 and MAVS was examined by IP-western blotting assay in cardiomyocytes in the presence or absence of si-PGAM5 under MCM + H/R treatment. l Immunofluorescence of NLRP3 and MAVS in cardiomyocytes with or without PGAM5 knockdown under MCM + H/R treatment. m Colocalization analysis of NLRP3–MAVS. Pearson’s R value (no threshold) was calculated by ImageJ Fiji software. n = 6 images from 3 biological replicates. n–p Representative western blotting (n) and quantitated analyses of GSDMD (o) and caspase-1 (p) in cardiomyocytes infected with PGAM5 or si-MAVS under MCM + H/R treatment. q IL-18 release measured by ELISA in cardiomyocytes of the indicated groups underwent MCM + H/R treatment. r–t Representative western blotting (r) and quantitated analyses of GSDMD (s) and caspase-1 (t) in cardiomyocytes infected with MARCH2, PGAM5, or MAVS under MCM + H/R treatment. u IL-18 release measured by ELISA in cardiomyocytes of the indicated groups underwent MCM + H/R treatment. Data are shown as the means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 7. MARCH2 overexpression reduces I/R injury and preserves cardiac function via inhibition of PGAM5–MAVS-NLRP3 inflammasome pathway.

a Representative M-mode recordings of echocardiography of WT mice injected with AAV9-cTnT-NC or AAV9-cTnT-MARCH2 (2 × 10¹¹ V.g/mouse) for 3 weeks and then subjected to I/R (45 min/24 h). b, c Quantitative analysis of echocardiographic measurements performed in AAV9-cTnT-NC or AAV9-cTnT-MARCH2 mice subjected to I/R (45 min/24 h) injury (n = 6 mice per group). Left ventricular (LV) ejection fraction (EF, b) and fractional shortening (FS, c). d Representative images of heart sections by TTC/Evans Blue staining depicting infracted area. e Ratios of area at risk (AAR) to left ventricular (LV) area. f Infarct area normalized to AAR (n = 6 mice per group). g, h LDH activity and CK-MB level in AAV9-cTnT-NC or AAV9-cTnT-MARCH2 mice with or without myocardial I/R (45 min/24 h) (n = 6 mice per group). i Ubiquitination assays determining the ubiquitination of endogenous PGAM5 in the hearts of AAV9-cTnT-NC or AAV9-cTnT-MARCH2 mice subjected to I/R injury. j Representative western blotting analysis of PGAM5, caspase-1 (procaspase1; cleaved caspase-1), GSDMD (full-length; N-terminal), and IL-18 in hearts of AAV9-cTnT-NC or AAV9-cTnT-MARCH2 mice subjected to I/R (45 min/9 h). k Quantitated analysis of PGAM5 in hearts of AAV9-cTnT-NC or AAV9-cTnT-MARCH2 mice subjected to I/R (45 min/9 h). l, m Interaction between NLRP3 and ASC in mitochondria of cardiac tissues of AAV9-cTnT-NC or AAV9-cTnT-MARCH2 mice subjected to I/R was examined by IP-western blotting assay. n IL-18 release was measured by ELISA in cardiac tissues of AAV9-cTnT-NC or AAV9-cTnT-MARCH2 mice subjected to I/R. Data are shown as the means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 8. PGAM5/MAVS signaling pathway is essential for the regulation of MARCH2 in I/R injury.

a Prior to I/R injury, WT mice were administrated with AAV9-cTnT-vector, AAV9-cTnT-MARCH2, AAV9-cTnT-PGAM5 or AAV9-cTnT-MAVS (2.0 × 10¹¹ V.g/mouse) for 3 weeks by tail vein injection. TTC/Evans Blue staining is used to depict infracted area. b Ratios of area at risk (AAR) to left ventricular (LV) area. c Infarct area normalized to AAR. No statistical significance for Vector-I/R vs MARCH2-PGAM5-I/R, no statistical significance for Vector-I/R vs MARCH2-MAVS-I/R. d, e Echocardiographic assessment of ejection fraction (EF, d) and fractional shortening (FS, e) in the indicated groups. No statistical significance for Vector-I/R vs MARCH2-PGAM5-I/R, no statistical significance for Vector-I/R vs MARCH2-MAVS-I/R. f, g LDH activity (f) and CK-MB level (g) in mouse with and without myocardial I/R. No statistical significance for Vector-I/R vs MARCH2-PGAM5-I/R; no statistical significance for Vector-I/R vs MARCH2-MAVS-I/R. h Representative western blotting results of MARCH2, PGAM5, mito-MAVS, caspase-1 (procaspase1; cleaved caspase-1), GSDMD (full-length; N-terminal), and IL-18 in hearts of the indicated group mice subjected to I/R (45 min/9 h). i IL-18 release was measured by ELISA in cardiac tissues of the indicated group mice subjected to I/R (45 min/9 h). No statistical significance for Vector-I/R vs MARCH2-PGAM5-I/R, no statistical significance for Vector-I/R vs MARCH2-MAVS-I/R. j A working model for MARCH2 function in I/R injury. MARCH2 interacts with and promotes the degradation of PGAM5 by facilitating its K48-linked polyUb, thus inhibiting the activity of MAVS/NLRP3 inflammasome pathway and cardiomyocyte pyroptosis. Data are shown as the means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.