Histone lysine dimethyl-demethylase KDM3A controls pathological cardiac hypertrophy and fibrosis

Abstract

Left ventricular hypertrophy (LVH) is a major risk factor for cardiovascular morbidity and mortality. Pathological LVH engages transcriptional programs including reactivation of canonical fetal genes and those inducing fibrosis. Histone lysine demethylases (KDMs) are emerging regulators of transcriptional reprogramming in cancer, though their potential role in abnormal heart growth and fibrosis remains little understood. Here, we investigate gain and loss of function of an H3K9me2 specific demethylase, Kdm3a, and show it promotes LVH and fibrosis in response to pressure-overload. Cardiomyocyte KDM3A activates Timp1 transcription with pro-fibrotic activity. By contrast, a pan-KDM inhibitor, JIB-04, suppresses pressure overload-induced LVH and fibrosis. JIB-04 inhibits KDM3A and suppresses the transcription of fibrotic genes that overlap with genes downregulated in Kdm3a-KO mice versus WT controls. Our study provides genetic and biochemical evidence for a pro-hypertrophic function of KDM3A and proof-of principle for pharmacological targeting of KDMs as an effective strategy to counter LVH and pathological fibrosis.

Fig. 1

KDM3A promotes cardiomyocyte hypertrophy in vitro. a Venn diagram showing differentially expressed genes that are involved in histone methylation and are either up or downregulated in PDE5-Tg versus control littermates, PED5-Tg treated with sildenafil (Reversal) versus control littermates, and reversal versus PDE5-Tg mice after TAC. b KDM3A mRNA in human patients with hypertrophic cardiomyopathy (HCM) (n = 5) and normal controls. *p < 0.05 (t test). c Immunofluorescence micrographs of NRVMs transduced with adenoviruses expressing either LacZ or Kdm3a. Cells were stained with phalloidin. Scale bar, 100 μm. d Relative cell size from cells in c. n = 3 ± SEM, *p < 0.05 (t test). e Relative fold change of mRNA of gene associated with hypertrophic remodeling in NRVMs transduced with either Ad-LacZ or Ad-Kdm3a (n = 3 ± SEM). mRNAs were normalized against internal Gapdh. *p < 0.05 (t test). f–g NRVMs were transfected with control siRNA or Kdm3a specific siRNA, treated with or without PE. Cells were fixed and stained with phalloidin for measurement of cell size (g) or harvested for measurement of relative mRNA of Kdm3a and fetal gene markers (g). mRNA were normalized against internal GAPDH. n = 3 ± SEM, *p < 0.05 (ANOVA)

Fig. 2

KDM3A promotes TAC-induced hypertrophic remodeling. WT and Kdm3a-Tg (Tg) mice were subjected to Sham (S) and TAC (T) surgery. Echocardiograph was performed on mice and hearts were harvested after 6 weeks for histological and biochemical analysis. a H&E staining of histologic sections of WT and Kdm3a-Tg mouse hearts. Scale bar, 1 mm. b HW/BW, c relative myocyte cell size, d LW/BW, and e relative fibrotic area of WT and Kdm3a-Tg mouse hearts. f Left ventricular end diastolic diameter (LVEDD), left ventricular end systolic diameter (LVESD), percent of fractional shortening (FS%), and heart rate (HR, beat/min). g Relative mRNA of canonical fetal gene markers (Nppa, Nppb, and Myh7), Fhl1, and Col1a2. n = 6-10 ± SEM. ^{*, #}p < 0.05 (ANOVA). *WT TAC vs. WT Sham. ^#, Tg-TAC vs. WT-TAC

Fig. 3

Kdm3a-deficiency protects mice again TAC-induced hypertrophic remodeling. WT and Kdm3a KO (KO) mice were subjected to Sham and TAC surgery. Hearts were echoed and harvested 6 weeks post-surgery for histological and biochemical analysis. a H&E staining of histologic sections of WT and KO mouse hearts. Scale bar, 1 mm. b HW/BW, c relative myocyte size, and d percent of fibrotic area of WT and KO mouse hearts. e LVEDD, f LVESD, g percent FS and heart rate of WT and KO mouse hearts. h Relative mRNA of canonical fetal gene markers (Nppa, Nppb, and Myh7), Fhl1, and Col1a2. mRNA transcripts were measured by qRT-PCR, normalized against internal Gapdh, and expressed relative to Sham WT mice. n = 6-10 ± SEM. *^{, #}p < 0.05 (ANOVA). *, WT TAC vs. WT Sham. ^#, Tg-TAC vs. WT-TAC

Fig. 4

KDM3A promotes transcription of a pro-fibrotic gene program. a Volcano plot of differentially expressed genes in Kdm3a KO mouse hearts compared to those in WT mouse hearts at 6 weeks post-TAC surgeries. Timp1 is among the most downregulated genes in KO hearts. b Heatmaps of differentially expressed genes in KO vs WT and Kdm3a-Tg vs WT mouse hearts at 6 weeks post-TAC surgery, showing differentially expressed genes are inversely regulated in KO vs Tg hearts. c GO analysis of differentially downregulated genes in KO vs WT TAC mouse hearts. Relative mRNA of genes involved in fibrosis in Kdm3a^-/- (d) and Kdm3a-Tg (e) mouse hearts 6 weeks post-Sham and TAC surgery. mRNA transcripts were measured by qRT-PCR, normalized against internal Gapdh, and expressed relative to Sham WT mice. n = 5 ± SEM. *^{, #}p < 0.05 (ANOVA). *, WT TAC vs. WT Sham. ^#, Tg-TAC vs. WT-TAC

Fig. 5

KDM3A activates the transcription of Timp1 in cardiomyocytes. Cardiomyocytes (CM) and cardiac fibroblasts (cFb) were isolated from WT and Kdm3a-Tg mouse hearts at week 6 post-Sham and TAC surgery. a–e Transcripts of fetal genes and fibrosis-related ECM genes were measured by qRT-PCR, normalized against internal Gapdh, and expressed relative to WT Sham CM. n = 3 ± SEM, *^{, #}p < 0.05 by student t test. *, WT TAC vs WT Sham. ^#, Tg-TAC vs WT-TAC. f Cardiomyocytes from Kdm3a-Tg mouse hearts at week 6 post-Sham and TAC surgery were isolated and used for ChIP assay with antibodies against KDM3A (left panel) or H3K9me2 (right panel). The relative occupancies of KDM3A and H3K9me2 at Timp1 promoter were normalized against Input and expressed relative to Sham control. n = 3 ± SEM, *p < 0.05 (t test). g Timp1-Luciferase reporter was transfected in combination with plasmids as indicate into 293 T cells. Cells were harvested 48 h after transfection. Luciferase activities were normalized against co-transfected β-galactosidase, expressed relative to vector-transfected cells. n = 3 ± SEM. (h-k) Kdm3a-Tg mice were transduced with adenovirus AAV9 expressing control or Timp1 shRNA and subjected to Sham or TAC surgery. Hearts were echoed and harvested 5 weeks post-surgery. h WB of TIMP1 and GAPDH, i HW/BW, j %FS, and k percent of fibrotic area of Sham (S), TAC (T) Kdm3a-Tg mice transduced with control or Timp1 shRNA. n = 7 ± SEM, *p < 0.05 (ANOVA)

Fig. 6

JIB-04 blocks TAC-induced hypertrophic remodeling and fibrosis. a Experimental protocol of JIB-04 and control vehicle (Veh) treatment of WT and Kdm3a-Tg mice. JIB-04/Veh were given to mice at either day 1 or 2 (exp 1) or 21 (exp 2) post-TAC. b %FS, c HW/BW ratio, d relative cell size, e percentage of fibrotic area, and f relative fold change of fetal genes and Timp1 of WT sham, WT TAC and Kdm3a-Tg TAC mouse heart at the end of experiment 1. n = 5-6 ± SEM, *, ^#, ^$, p < 0.05 (ANOVA). *, WT TAC vs WT Sham. ^#, JIB-04 WT TAC vs Veh WT-TAC. ^$, JIB-04 Tg TAC vs Veh TG TAC. g Western blot of TIMP1 in vehicle or JIB-04 treated WT mouse hearts at post-sham/TAC week 6 in experiment 1. h %FS over treatment period, i HW/BW, j relative cell size, and k % fibrotic area at end of experiment 2 of Sham and TAC Kdm3a-Tg mouse hearts treated with vehicle or JIB-04. n = 5 ± SEM, *p < 0.05 (ANOVA)

Fig. 7

JIB-04 protects mice against I/R injury. WT mice were subjected to I/R surgery and given vehicle (Veh) or JIB-04 as shown in Fig. 6a. Echocardiographs were performed weekly for 4 weeks to measure the fractional shortening (a). b Percent fibrotic area and c HW/BW ratio of mice treated with Veh or JIB-04 at 4 weeks post-I/R surgery. d Immunofluorescences of H3K9me2, H3K9me3, H3K4me3, and H3K27me3 of hearts treated with Veh or JIB-04 at week 4 post-I/R surgery. H3K9me2/3 in both cardiomyocyte (arrow head) and non-cardiomyocytes (arrow) were affected by JIB-04 treatment. Scale bar, 20 μm. e Quantification of percent of cells in d stained with methylated histones as indicated. Staining of methylated histones was normalized against Dapi stained cells. *p < 0.05, n = 5 ± SEM (ANOVA)

Fig. 8

JIB-04 targets H3K9me2 demethylase and fibrotic gene program. a Western blots of indicated proteins in lysates from WT mouse hearts treated with vehicle (Veh) or JIB-04 harvested at day 7 and day 42 post-TAC surgeries. b, c Western blots of indicated proteins from lysates of NRVMs (b) and human iPSC-CM (c) treated with various concentration of JIB-04. d Venn diagram showing overlapping of downregulated gene sets between KO vs WT and JIB-04 vs Veh treated mice. e GO analysis of 114 overlapping genes in d, showing significant enrichment of gene set in ECM and growth factor/integrin signaling. f qRT-PCR of representative genes in d. n = 3 ± SEM, *, ^#, ^$, p < 0.05 (ANOVA). *, WT TAC vs WT Sham. ^#, JIB-04 WT TAC vs Veh WT TAC. ^$, JIB-04 Tg TAC vs Veh TG TAC. g Schematics showing potential mechanism of action for KDM3A and inhibitor JIB-04