KDM8 epigenetically controls cardiac metabolism to prevent initiation of dilated cardiomyopathy

Figures

• 

  Fig. 1. Cardiomyocyte-specific loss of Kdm8 causes DCM leading to lethal heart failure. a, Hearts from 2-month-old and 6-month-old Kdm8fl/fl control and Kdm8fl/fl;Myh6-Cre mutant mice. Scale bar, 1 mm. b, Heart weight normalized to the tibia length of control, heterozygous (Kdm8fl/+;Myh6-Cre), and mutant mice. Data are mean ± s.e.m. Two-way ANOVA with multiple comparison and Sidak correction. n = 27 control, 5 heterozygous, and 25 mutant 2-month-old mice per group. n = 21 control, 7 heterozygous, and 8 mutant 4-month-old mice. n = 8 control, 6 heterozygous, and 8 mutant 6-month-old mice. c, Representative sections of three hearts from 2-month-old and 6-month-old control and mutant mice, stained with Masson’s trichrome. RA, right atrium; LA, left atrium; RV, right ventricle; LV, left ventricle; IVS, interventricular septum. Scale bar, 1 mm. d, Thickness of the wall of the LV, RV, and IVS, and diameter of the LV at end-diastole (EDD) measured from sections of 6-month-old hearts. Data are mean ± s.e.m. Two-tailed Student’s t-test. n = 4 controls and 6 mutants. e, Immunofluorescence of Col5a1 and α-actinin on sections of 2-month-old and 6-month-old hearts. Nuclei were counterstained with…

• 

  Fig. 2. Genes regulating mitochondrial oxidative metabolism are repressed before cardiac functional decline in Kdm8 mutant hearts. a, Heatmap clustering of genes differentially expressed in ventricles of 2-month-old Kdm8fl/fl;Myh6-Cre mutant vs Kdm8fl/fl control mice. Enriched GO terms and representative genes are shown. b, GSEA including all expressed genes. P values were obtained from GSEA. NES, normalized enrichment score. c, Western blot of electron transport chain complexes I (Ndufb8) and IV (Cytochrome C) on 2-month-old ventricles. d, Protein relative to histone H3 in c. Data are mean ± s.e.m. Two-tailed Student’s t-test. n = 4 control and 5 mutant hearts. e, Transmission electron micrographs of the left ventricle of 2-month-old controls and mutants. Arrowheads indicate mitochondria. Scale bar, 1 μm. f, Mitochondria per 10 μm2, and percentage area occupied by mitochondria per micrograph. Data are mean ± s.d. Two-tailed Student’s t-test. n = 5 mice per group. g, Heatmap of H3K36me2 (red) occupancy (normalized counts) in 2-month-old control and mutant hearts 1 kb upstream and downstream of the transcription start site (TSS). h, H3K36me2 enrichment as reads per kilobase per million…

• 

  Fig. 3. NAD+ synthesis metabolome is progressively altered at early stages of DCM. a, Reporter metabolite analysis using differentially expressed genes in ventricles of Kdm8fl/fl control and Kdm8fl/fl;Myh6-Cre mutant mice at 2 and 6 months of age. The left panel illustrates all 46 reporter metabolites detected in 2-month-old hearts. The right panel illustrates the top 55 out of 718 reporter metabolites. Upregulated metabolites are in red, and downregulated metabolites are in blue. Data were analyzed by piano using distinct-directional P values. n = 2 mice per group at 2 months, and 3 mice per group at 6 months. b, Metabolomic networks from liquid chromatography–mass spectrometry (LC–MS) analysis of ventricles of control and mutant hearts at 2 and 6 months of age. Significantly altered metabolites are circled in black. Upregulated metabolites are in red, and downregulated metabolites are in blue. Each time point was independently analyzed by two-tailed Student’s t-test. P values were adjusted for false discovery by the Benjamini–Hochberg method. n = 5 controls and 7 mutants at 2 months; n = 9 controls and 5 mutants at 6 months. c, Quantification of NAM, NAD+ and 1-methyl-NAM by LC–MS, and…

• 

  Fig. 4. NAD+ treatment blunts cardiac deterioration toward heart failure in Kdm8 mutant mice. a, Left ventricle ejection fraction, fractional shortening, and diameter at end-systole and end-diastole obtained from echocardiography on Kdm8fl/fl control and Kdm8fl/fl;Myh6-Cre mutant mice that were untreated or injected with NAD+ intraperitoneally for 2 months starting at 2 months of age. Data are mean ± s.e.m. Two-way ANOVA with Tukey’s multiple comparison correction. n = 3–9 mice per group (see Source data for details). *Statistically significant. ns = nonsignificant. b, Hearts of 4-month-old control and mutant mice that were untreated or treated with NAD+. Scale bar, 1 mm. Scale bars = 1 mm and 200 μm. c, Micrographs at the top correspond to representative sections of 4-month-old control and mutant hearts that were stained with Sirius Red/Fast Green. Scale bar, 1 mm. Micrographs at the bottom are close-ups of regions in the inset. Scale bar, 100 μm. The experiment was repeated independently on four untreated control, four NAD+-treated control, four untreated mutant, and three NAD+-treated mutant mice, with similar results. d, Diameter of the left ventricle measured from sections of…

• 

  Fig. 5. Tbx15 represses a mitochondrial gene network when derepressed in cardiomyocytes. a, Differentially expressed genes in 2-month-old Kdm8fl/fl;Myh6-Cre mutant vs Kdm8fl/fl control hearts. Gray dots represent log2[fold change] > 1 adjusted P < 0.05. n = 2 hearts per group. Data were analyzed using DESeq2. b, qPCR of Tbx15, Acta1, Nppa and Nppb on ventricles of 2-month-old control and mutant hearts. Data are mean ± s.e.m. Two-tailed Student’s t-test. n = 5 mice per group. c, Western blot of Tbx15 on 6-month-old ventricles. d, Protein relative to histone H3. Data are mean ± s.e.m. Two-tailed Student’s t-test. n = 5 mice per group. e, Immunofluorescence of Tbx15 in heart sections. Nuclei were counterstained with DAPI. Scale bar, 20 μm. f, H3K36me2 tracks (top) and mRNA fragments (bottom) on Tbx15 in control and Kdm8 mutant ventricles. g, Luciferase activity of the Nampt promoter in HEK293T normalized to pRL-TK, with increasing amounts of a plasmid encoding His-tagged TBX15 (TBX15-His). Data are mean ± s.d. One-way ANOVA with Tukey’s multiple comparison correction. n = 3 wells per group. h, Genes differentially expressed in iPSC-CMs transfected with TBX15-His. Gray dots represent…

• 

  Fig. 6. TBX15 upregulation tracks with stronger downregulation of mitochondrial metabolism regulatory genes in human hearts affected by DCM. a, GSEA comparing 149 DCM-affected and 113 control human hearts. Each circle represents a gene set, and blue circles represent downregulated gene sets. All gene sets with Q-value < 0.01 are shown. b, Normalized expression of KDM8 in control and DCM-affected hearts. Thick dotted lines denote the median, and thin dashed lines denote the quartiles. Data were analyzed using DESeq2. c, Unsupervised hierarchal clustering of genes in the NAD+ pathway. Up, significantly upregulated; Down, significantly downregulated; Ns, not significant; Ne, not expressed. d, Normalized expression of NAMPT and SARM1 in control and DCM-affected hearts. Thick dotted lines denote the median, and thin dashed lines denote the quartiles. Data were analyzed using DESeq2. e, Unsupervised hierarchal clustering of all genes that encode mitochondrial proteins (obtained from GO term Cellular Component). Three clusters are identified. MitoSevere clusters DCM samples with the lowest expression of mitochondrial genes. f, Unsupervised hierarchal clustering of genes dysregulated in…

• 

  Fig. 7. Schematic of the function of Kdm8 in cardiac metabolism control. In normal cardiomyocytes, Kdm8-mediated H3K36me2 demethylation in the Tbx15 gene body associates with repression. In contrast, H3K36me2 preferentially demethylates promoters of active genes encoding mitochondrial proteins like Nampt. In Kdm8-deficient cardiomyocytes, H3K36me2 is enriched in the body of upregulated genes like Tbx15 and in promoters of genes encoding mitochondrial proteins. Upon gene derepression, Tbx15 with associated corepressor complexes represses genes controlling mitochondrial function, leading to decreased NAD+ that precedes ATP depletion and cardiac deterioration toward DCM and heart failure.

• 

  Extended Data Fig. 1. Kdm8 demethylates H3K36me2 in the heart. a, Percentage of genotypes of male offspring from control Kdm8fl/fl females crossed with mutant Kdm8fl/fl;Myh6-Cre males. Data was analyzed by two-tailed chi-squared test. b, qPCR of Kdm8 normalized to Rpl13a in whole ventricles of 2-month-old control, heterozygous (Kdm8fl+;Myh6-Cre), and mutant mice. Error bars denote the mean + /- s.d. Data was analyzed by one-way ANOVA with Tuckey’s multiple comparison correction. n = 5 hearts per group. c, qPCR of Kdm8 in isolated cardiomyocytes (CMs) and in the non-cardiomyocyte cell fraction (non CMs) of 2-month-old control and mutant hearts. Error bars denote the mean + /- s.d. Data was analyzed by two-way ANOVA. n = 3 control and 5 mutant hearts. qPCR in b and c were repeated twice with similar results. d, Western blot of Kdm8 in whole ventricles at 6 months of age. e, Band intensity of Kdm8 protein normalized to histone H3 (H3) (shown in d) and analyzed by unpaired two-tailed Student’s t-test. n = 5 hearts per group. Error bars denote the mean + /- s.e.m. f, Histone demethylase (HDM) activity of Jumonji domain-containing proteins, shown as fluorescence units, on extracts of control and mutant…

• 

  Extended Data Fig. 2. Kdm8 mutants develop dilated cardiomyopathy with no signs of concentric hypertrophy. a, Echocardiogram analysis of left ventricle (LV) anterior wall thickness at end-diastole, LV posterior wall thickness at end-systole, LV posterior wall thickness at end-diastole, and LV fractional shortening in control (Kdm8fl/fl), heterozygous (Kdm8fl/+;Myh6-Cre), and mutant (Kdm8fl/fl;Myh6-Cre) mice at 2, 4 and 6 months of age. Error bars denote the mean + /- s.e.m. Data was analyzed by 2-way ANOVA with Tuckey’s multiple comparison correction. * denotes comparisons between control and mutants, # between control and heterozygous hearts. n = 6 - 11 mice per group. b, Wheat germ agglutinin staining (WGA) on cross sections of the hearts of 2 and 6-month-old control and mutant mice. Scale bar, 100 μm. c, Cell surface area quantified from sections stained with WGA and analyzed by two-tailed Student’s t-test. Error bars denote the mean + /- s.e.m. n = 3 hearts per group. d, Lung sections of control and Kdm8 mutants at 4 and 6 months of age stained with H&E. Arrows point to areas of accumulated alveolar fluid. Scale bar = 100 μm. e, Alveolar area per 0.1 mm2 in lung sections Error bars denote the mean…

• 

  Extended Data Fig. 3. Loss of Kdm8 in cardiac progenitor cells causes dilated cardiomyopathy. a, qPCR of Kdm8 normalized to Rpl13a on 2-month-old control (Kdm8fl/fl) and mutant (Kdm8fl/fl;Myh6-Cre or Kdm8fl/fl;Nkx2-5-Cre) hearts. Error bars denote the mean + /- s.e.m. Data was analyzed by two-way ANOVA. n = 5 hearts per group. The experiment was repeated twice with similar results. b, Hearts of control and mutant mice at 2, 6, 20, and 30 months of age. Scale bar, 1 mm. c, Sections of 2, 6 and 30-month-old control and mutant hearts stained with H&E. Scale bar, 1 mm. d, Echocardiogram analysis of left ventricle (LV) ejection fraction, fractional shortening, diameter at end-systole and end-diastole, and isovolumetric relaxation time normalized to the time between heartbeats (IVRT/RR) in control and mutant mice. Error bars denote the mean + /- s.e.m. Data was analyzed by 2-way ANOVA with Sidak’s multiple comparison correction. n = 10 – 11 mice at 2, 4 and 6 months, n = 5 controls and 7 mutants at 20 months, and n = 3 controls and 4 mutants at end point (EP) (see Supplementary Table 2 for P values). * P < 0.05, ** P < 0.01, *** P < 0.001. e, Survival curve of control, and mutant mice analyzed by Log-rank…

• 

  Extended Data Fig. 4. Mitochondrial function regulators are downregulated at DCM onset, and mitochondrial DNA is decreased at end-stage DCM in Kdm8 mutant hearts. a, Gene ontology terms (cellular component) enriched amongst genes upregulated (yellow) and downregulated (purple) in ventricles of 2-month-old Kdm8 mutant hearts. Data was analyzed using DAVID v6.8 with Benjamini correction. b, qPCR of transcripts of components of electron transport (ETC) complexes and the tricarboxylic acids (TCA) cycle on ventricles of 2-month-old control (Kdm8fl/fl) and mutant (Kdm8fl/fl;Myh6-Cre) hearts, analyzed by two-tailed Student’s t-test with Hold-Sidak multiple comparison correction. n = 5 mice per group. Error bars denote the mean + /- s.e.m. c, qPCR on cardiomyocytes (CMs) and noncardiomyocytes (non CMs) isolated from control and mutant mice at 2 months of age. Error bars denote the mean + /- s.e.m. Data was analyzed by two-way ANOVA. n for Nampt, Ndufa3, and Cox6a2 = 5 control and 6 mutant hearts. n for Ogdh1 = 4 control and 5 mutant hearts. d, qPCR of mitochondrial DNA (mtDNA) normalized to the β-globin gene as representative of nuclear DNA, on the left (LV) and right ventricle (RV) of Kdm8fl/fl control and…

• 

  Extended Data Fig. 5. H3K36me2 is enriched at gene bodies of upregulated genes, and promoters of downregulated metabolic genes in Kdm8 mutant hearts. a. Proportion of genes with increased or decreased H3K36me2 that were downregulated or upregulated in 2-month-old Kdm8fl/fl;Myh6-Cre mutant hearts. b. Gene ontologies (GO) (biological process) enriched amongst genes whose promoters were enriched for H3K36me2 in 2-month-old Kdm8 mutant hearts. Data was analyzed using DAVID v6.8 with Benjamini correction. c. H3K36me2 CUT&Tag and RNAseq tracks of Acta1 and d. Ogdhl loci in 2-month-old Kdm8fl/fl control and Kdm8fl/fl;Myh6-Cre mutant hearts. The blue line indicates the promoter region upstream the transcription start site of Nampt. Units for CUT&Tag are reads per kilobase per million mapped reads (RPKM) and reads per million mapped reads (RPM) for RNAseq. e. GO terms (biological process) enriched amongst genes downregulated in 2-month-old Kdm8 mutant hearts. Data was analyzed using DAVID v6.8 with Benjamini correction.