Histone Demethylase KDM5A Is Activated and Pathogenic in a Mouse Model of LMNA-Cardiomyopathy

Visual Abstract

Highlights

• •Histone lysine demethylase KDM5A is reactivated in lamin A–cardiomyopathy (LMNA-CMP).
• •Cardiomyocyte-specific deletion of Kdm5a improves cardiac function, prolongs survival, and attenuates fibrosis and cell death in a mouse model of LMNA-CMP.
• •Loss of Kdm5a restores dysregulated gene programs involved in oxidative phosphorylation, fatty acid metabolism, and myogenesis.
• •KDM5A represses the transcription factors TBX5 and ESRRG by reducing H3K4me3 at their promoters, leading to suppression of downstream target genes.
• •KDM5A thus represents a potential epigenetic therapeutic target for heart failure.

Summary

Heart failure is associated with widespread epigenetic remodeling in cardiomyocytes, but the chromatin regulators driving this process remain poorly defined. We identify the histone H3K4 demethylase KDM5A as a key epigenetic mediator of heart failure in a cardiomyocyte-specific lamin A–deficient mouse model. Disease-associated activation of KDM5A leads to redistribution of promoter histone H3 lysine 4 trimethylation and repression of genes regulating sarcomere structure and mitochondrial oxidative metabolism. Genetic deletion of Kdm5a restores histone H3 lysine 4 trimethylation enrichment, reactivates cardiac and metabolic transcriptional programs, improves cardiac function, and prolongs survival. These findings identify KDM5A-mediated chromatin remodeling as a previously unrecognized mechanism of heart failure and highlight KDM5A as a potential therapeutic target.

Heart failure (HF) is a major cause of mortality and morbidity worldwide.^1,2 Regardless of the underlying etiology, HF development and progression share several hallmark features, including mitochondrial dysfunction, metabolic dysregulation, and alterations in sarcomere composition.3, 4, 5, 6, 7 These pathological changes are often preceded and/or accompanied by extensive molecular remodeling, which affects cardiac myocyte structure and function and are regulated by epigenetic modifications.8, 9, 10, 11, 12 Establishment of chromatin/epigenetic changes is mediated by networks of histone-modifying enzymes that write, read, and erase histone marks in a locus and cell type–specific manner.^13,14 Despite advances, key challenges remain in identifying effective epigenetic targets and clarifying how distinct regulators drive HF pathogenesis. Given the heterogeneity of HF, defining core molecular mechanisms involved in its pathogenesis could serve as additional therapeutic targets for interventions.

The histone H3 lysine 4 trimethylation (H3K4me3), a ubiquitous epigenetic mark, is typically enriched at the promoters of active genes, and is a determinant of gene expression. It is dynamically regulated by histone methyltransferases and demethylases.13, 14, 15 The KDM5 family of histone demethylases are evolutionarily conserved members of the JUMONJI family and the primary regulator of the H3K4 methylation state. The KDM5 family of proteins catalyzes the removal of methyl groups from H3K4me3/2 and thereby acts as a repressor of gene expression, although evidence also supports roles in transcriptional activation through recruitment of co-regulators.^13,16, 17, 18

We recently showed increased lysine demethylase 5A (KDM5A) expression and activity in human HF samples and in mouse models of HF.^19,20 However, the functional and biological significance of activation of the KDM5A in cardiomyocytes (CMs) in HF remains unknown. Therefore, to determine the pathogenic role of KDM5A in HF, we used a well-characterized mouse model of HF generated on the deletion of the Lmna gene specifically in cardiac myocytes (Myh6-Cre:Lmna^F/F, aka lamin A/C-cardiomyopathy [LMNA-CMP] mice).^9,21 These mice exhibit premature death within 3 to 4 weeks after birth, severe cardiac dysfunction, and increased myocardial fibrosis and apoptosis.

Our results show that KDM5A protein levels are increased in cardiac myocytes in the LMNA-CMP and its target genes involved in cardiac structure and function, and oxidative phosphorylation (OXPHOS) are suppressed. Genetic deletion of the Kdm5a gene specifically in cardiac myocytes restores expression of the genes and transcription factors involved in cardiac structure and function, and OXPHOS in the LMNA-CMP mice. The findings were partially corroborated by increased genome-wide H3K4me3 methylation of the corresponding genes and transcription factor loci. Partial restoration of the molecular remodeling of the myocardium was associated with prolonged survival, improved cardiac function, and reduced myocardial fibrosis and cell death. Thus, our study uncovered a key role for KDM5A-mediated epigenetic regulation in mitochondrial homeostasis and cardiac function in HF.

Methods

Anesthesia and euthanasia

All echocardiography and tissue isolation procedures were conducted under general anesthesia, induced with 3% inhaled isoflurane and maintained at 0.5% to 1.0% throughout.^9,20, 21, 22 For CM and whole-heart isolations, mice were euthanized by cervical dislocation while under anesthesia.

Data Availability

RNA-seq and CUT&RUN data have been deposited in GSE309726 and additional processed data are presented in Supplemental Tables 1 to 3.

Mice used in the study

Wild-type (C57BL/6J), Myh6-Cre (C57BL/6J), and Kdm5a floxed mice (JB6.129S6[Cg]-Kdm5a^tm1Kael/J) were obtained from Jackson Laboratory. Lmna^F/F and Kdm5a^F/F mouse lines were backcrossed to C57BL/6J wild-type mice to homogenize the genetic background. Cardiac-specific Lmna knockout mice (Myh6-Cre:Lmna^F/F) were generated by crossing Lmna^F/F mice with Myh6-Cre transgenic mice. Myh6-Cre:Kdm5a^F/F mice were generated by breeding Myh6-Cre mice with Kdm5a^F/F mice and were subsequently crossed with Lmna^F/F mice to generate cardiac-specific double-knockout mice (Myh6-Cre:Lmna^F/F:Kdm5a^F/F). Genotyping was performed by polymerase chain reaction using DNA extracted from mouse tails, following incubation in buffer at 55 °C for 15 minutes and 95 °C for 3 minutes. Polymerase chain reaction was conducted with 2 μL of extract using a ready-mix blue extract. Oligonucleotide primers are listed in Supplemental Table 4.

Sex as a biological variable

Age-matched male and female mice were included in all experimental groups. Data were analyzed with sex incorporated as a biological variable to evaluate potential sex-dependent differences in the observed phenotypic outcomes.

Survival analysis

Kaplan-Meier survival curves were generated for each group, using approximately equal numbers of male and female mice.

Echocardiography

Cardiac function was assessed using a Vevo 1100 system (Fujifilm VisualSonics) under 1% isoflurane anesthesia, maintaining a heart rate of ∼500 beats/min. M-mode images were analyzed to measure interventricular septal thickness, posterior wall thickness, left ventricular end-diastolic diameter (LVEDD) and end-systolic diameter over 3 to 4 cardiac cycles.^9,20,22 To minimize bias, echocardiographic measurements were performed by investigators blinded to experimental group allocation.

CM isolation

CMs were isolated from P14-P21 mice following published protocols.^9,20,22,23 Briefly, hearts were mounted on a Langendorff system and perfused with collagenase (1.2 mg/mL, Worthington cat #LS004176) at 4 mL/min for 10 to 12 minutes. Ventricles were minced in a stop buffer (10% calf serum, 12.5 μM CaCl₂, 2 mM ATP), filtered (100 μm), and centrifuged at 300 rpm for 4 minutes. The CM pellet was resuspended in appropriate buffers for downstream applications.

Immunoblotting

Protein expression in whole heart and CM extracts was analyzed by immunoblotting (IB). Tissues and/or CM were homogenized in RIPA buffer containing benzonase (250 units/mL buffer) with protease/phosphatase inhibitors (Roche #4693116001 and #49068459001). Lysates were nutated at 4 °C for 30 minutes and sonicated (3 cycles of 30 seconds ON/OFF). Supernatants were collected after centrifugation (14,000 rpm, 15 minutes, 4 °C). Extracts (30-40 μg) were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to polyvinyl difluoride membranes at 4 °C overnight. Antibodies used in the article are listed in Supplemental Table 4.

Quantitative real-time polymerase chain reaction

Gene expression was quantified by quantitative polymerase chain reaction using gene-specific primers and SYBR Green master mix, normalized to GAPDH. Primer sequences are in Supplemental Table 4.

RNA sequencing

RNA sequencing (RNA-seq) was performed from P18-P20 isolated CMs. RNA was extracted with the RNeasy Mini Kit; samples with RNA Integrity Number >8.7 were used. Ribosomal RNA was removed, and libraries were generated for 150 base pairs paired-end sequencing (Illumina NextSeq500, Genewiz). Read alignment was completed using the Pluto Bioinformatics platform. Differentially expressed genes (q value <0.05) were obtained using limma-voom.

Cleavage under targets and release using nuclease assay

Cleavage under targets and release using nuclease (CUT&RUN) assay was performed by the Epigenomics Profiling Core at MD Anderson Cancer Center with some modifications.²⁴ Briefly, CMs isolated from P18-P20 mice were mildly crosslinked with 0.1% formaldehyde for 1 minute and quenched with 125 mM glycine. Nuclei isolated from ∼500,000 cells were immobilized on Concanavalin A–coated magnetic beads (Bangs Laboratories) and permeabilized using wash buffer containing 0.02% digitonin (Promega). Bead-bound nuclei were incubated with rabbit immunoglobulin G (Millipore) and H3K4me3 (EpiCypher) antibodies overnight at 4 °C. The next day, targeted chromatin digestion was achieved by pAG-MNase (EpiCypher) binding and incubation with CaCl₂ at 4 °C followed by purification of DNA using MinElute columns (Qiagen). Library preparation was carried out using the NEBNext Ultra II DNA Library Prep Kit (New England Biolabs) according to the manufacturer's instructions, and sequencing was performed on an Illumina NextSeq 500 platform to obtain 15 to 20 million reads per sample.

Sequencing reads were aligned to the mouse reference genome (GRCm39), and BAM files were used for peak calling with MACS2 using standard parameters (effective genome size = 1.87e + 09, band width = 300, model fold = [10, 30], and q value <0.05). Peak quality assessment and enrichment profiling were conducted using the computeMatrix and plotHeatmap functions in deepTools. Heatmaps were generated from log-transformed normalized values [log(CPM+1)]. Differential peak analysis between genotypes wild-type (WT) vs Myh6-Cre:Lmna^F/F and Myh6-Cre:Lmna^F/F:Kdm5a^F/F vs Myh6-Cre:Lmna^F/F was performed using DiffBind. Peaks were annotated and their genomic distribution was analyzed using Chip seeker, and visualization was performed with deepTools. Read alignment was completed using the Pluto Bioinformatics platform, and all downstream analyses, including peak calling and differential analysis, were conducted on a local Galaxy instance or in R.

Pathway analysis

Gene set enrichment analysis (GSEA) was performed on normalized counts using MSigDB gene sets, with 1,000 permutations and q < 0.05, or using the enrich R online tool.

Apoptosis

Hearts from P21 mice were fixed in 10% formalin overnight, dehydrated through ethanol gradients, and embedded in paraffin.^9,20,22,23,25, 26, 27 Sections (5 μm) were placed on Superfrost slides. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assays were conducted using the In-Situ Cell Death Detection Kit (Roche #11684795910) following deparaffinization and proteinase K digestion (20 μg/mL, 20 minutes, room temperature). Nuclei were counterstained with 1 μg/mL 4′,6-diamidino-2-phenylindole (DAPI). Images were captured with a Zeiss Axioplan microscope and quantified in ImageJ from ×40 high-magnification fields (6-9 sections/heart, ≥5 mice/group). Quantitative analysis was performed in a blinded manner.

Fibrosis

Sections were stained with picrosirius red for 1 hour, washed in ethanol, and mounted.^9,20,22,23,25, 26, 27 Collagen volume fraction (CVF) was analyzed in 10 fields/6 sections/5 mice/group using ImageJ. Quantitative analysis was performed in a blinded manner.

Statistics

For genomic studies, differential gene expression and peak enrichment analyses were performed using a false discovery rate using the Benjamini-Hochberg method at a threshold of 5% (q value <0.05). Pathway enrichment analyses were conducted with Ingenuity Pathway Analysis, applying a z-score cutoff of ±2 and a significance threshold of P < 0.05. For other studies, the normality of continuous variables was assessed using the Shapiro-Wilk test. Normally distributed variables were compared between 2 groups using the 2-sample t-test and across multiple groups using 1-way analysis of variance followed by Tukey's post hoc pairwise comparison. Non-normally distributed variables were compared using the Mann-Whitney test between 2 groups and by the Kruskal-Wallis test followed by Dunn's post hoc test for multiple comparisons across groups. Categorical variables were analyzed using the chi-square test. Survival analyses were performed using Kaplan-Meier survival curves. For molecular and histological analyses, 1 mouse per litter was randomly selected to preserve statistical independence. Survival and echocardiographic assessments included multiple animals from the same litter but selected randomly; data were analyzed assuming independence between observations. Survival was compared using the log-rank (Mantel-Cox) test. In Kaplan-Meier curves, stepwise declines indicate deaths and symbols denote censored animals alive at the time of analysis. Mice were euthanized at humane endpoints defined by institutional veterinary staff or died spontaneously. All statistical analyses were performed using R statistical software (version 4.5.3; R Core Team 2026) and/or GraphPad version 10.3.1(GraphPad Software).

Study approval

All animal experiments were conducted following the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at the University of Texas Health Science Center at Houston (AWC-24-0021). Supplemental Tables 1 to 3

Results

KDM5A is induced and its targets are suppressed in CMs in LMNA-CMP

The expression of the KDM5 family of proteins, including KDM5A, progressively declines in the heart in the postnatal period.²⁸ In contrast, in a mouse model of HF generated on deletion of the Lmna gene (Lmna^−/−), KDM5A is markedly induced in the postnatal heart, and the expression of its target genes is suppressed.²⁰ To investigate the role of KDM5A in the pathogenesis of HF, we used a CM-specific LMNA-CMP mouse model in which the Lmna gene is selectively deleted in CMs using the Myh6-Cre system (Myh6-Cre:Lmna^F/F).⁹ These mice develop cardiac dysfunction characterized by marked CM apoptosis and mortality by approximately 4 weeks of age.⁹ Heterozygous deletion of the Lmna gene results in a similar cardiac phenotype but evolves more gradually, with disease progression occurring over a 6- to 9-month period.^9,29,30 Given the shared phenotypic features but an accelerated course, the homozygous CM-specific Lmna knockout model was selected for the studies.

Analysis of whole-heart lysates from Myh6-Cre:Lmna^F/F mice by IB showed increased KDM5A expression level compared with WT controls (Figures 1A and 1B). Similarly, the KDM5A protein level was increased in CMs isolated from the Myh6-Cre:Lmna^F/F mice as compared with the WT myocytes (Figures 1C and 1D). Given that KDM5A is a histone demethylase predominantly localized to the nucleus, the expression level of KDM5A in the CM nuclear protein extracts was analyzed. IB analysis showed an increased level of KDM5A in Myh6-Cre:Lmna^F/F mice (Figures 1E and 1F) without appreciable change in total H3K4me3 (Supplemental Figures 1A and 1B). However, transcript levels of KDM5A putative target genes involved in OXPHOS, fatty acid oxidation (FAO), including the estrogen-related receptor (ERR) family of transcription factors, were reduced in the Myh6-Cre:Lmna^F/F CMs (Figure 1G). Taken together, these findings demonstrate that KDM5A levels, including in the myocyte nuclei, are increased and its downstream targets, including those involved in metabolism and myocyte structure, are suppressed in a mouse model of HF.

Figure 1.

Expression of KDM5A Protein Is Increased and Its Target Genes Are Suppressed in Myh6-Cre:Lmna^F/F Cardiomyocytes

(A) Immunoblot analysis of whole-heart lysates from wild-type and Myh6-Cre:Lmna^F/F mice showing KDM5A protein expression and normalization control GAPDH. (B) Quantification of KDM5A protein levels normalized to GAPDH, demonstrating increased expression in Myh6-Cre:Lmna^F/F hearts compared with wild-type (n = 5 wild-type and n = 4 Myh6-Cre:Lmna^F/F). (C) Immunoblot analysis of isolated cardiomyocytes from wild-type and Myh6-Cre:Lmna^F/F mice showing KDM5A expression. (D) Quantification of KDM5A protein levels from isolated cardiomyocytes normalized to Vinculin (VCL), confirming increased expression in Myh6-Cre:Lmna^F/F samples (n = 7 wild-type and Myh6-Cre:Lmna^F/F). (E) Immunoblot analysis of nuclear extracts from isolated cardiomyocytes of wild-type and Myh6-Cre:Lmna^F/F mice. (F) Quantification of nuclear KDM5A protein levels normalized to Histone H3, showing enhanced nuclear retention in Myh6-Cre:Lmna^F/F cardiomyocytes (n = 4 wild-type and Myh6-Cre:Lmna^F/F). (G) Transcript levels of representative KDM5A target genes in cardiomyocytes, showing suppression of genes in Myh6-Cre:Lmna^F/F samples as compared with wild-type (n = 4 wild-type and Myh6-Cre:Lmna^F/F). Normality of data distribution was assessed using the Shapiro-Wilk test. Normally distributed data are presented as mean ± SD and statistical significance was evaluated using Student's t-test [(B), (D), and (G)] and data that do not follow a normal distribution are presented as median with IQR (25th-75th percentile) and statistical significance was evaluated using the Mann-Whitney U test (F). The corresponding P values are shown. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. KDM5A = lysine demethylase 5A.

Kdm5a deletion improves survival and cardiac function in the LMNA-CMP mice

To elucidate the role of KDM5A induction in HF, we generated CM-specific Kdm5a single knockout mice (Myh6-Cre:Kdm5a^F/F). Both heterozygous and homozygous Kdm5a knockout offspring were born at the expected Mendelian ratios and exhibited no discernible phenotypes at birth. The efficiency of Kdm5a deletion was assessed by immunoblot analysis of CM lysates from WT, Myh6-Cre, Kdm5a^F/F, and Myh6-Cre:Kdm5a^F/F mice. KDM5A protein levels were significantly reduced in the knockout mouse hearts (Figures 2A and 2B).

Figure 2.

Conditional Deletion of Kdm5a in Cardiomyocytes Does Not Cause Overt Phenotypic Abnormalities

(A) Immunoblot analysis of isolated cardiomyocytes from P16 mice, which confirmed the absence of KDM5A protein following cardiomyocyte-specific deletion, normalized to VCL as a loading control (n = 3). (B) Quantitative assessment of KDM5A deletion. (C) Kaplan-Meier survival analysis showing that loss of Kdm5a in cardiomyocytes has no significant effect on overall survival compared with control groups. (D) Echocardiographic measurements obtained from 4 groups of mice assessing body weight, heart rate, left ventricular end-diastolic diameter (LVEDD), left ventricular end-systolic diameter (LVESD), and fractional shortening (FS) at 2 months of age. No significant differences were observed across genotypes. (E) Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining of myocardial sections from 2-month-old mice (n = 4 per group) showing apoptotic nuclei (green) and 4′,6-diamidino-2-phenylindole (DAPI) (blue). (F) Quantification of TUNEL⁺ nuclei expressed as percentage of total nuclei, showing no difference between the groups (n = 4). Normality of data distribution was assessed using the Shapiro-Wilk test. Normally distributed data are presented as mean ± SD and statistical significance was evaluated using one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test in (B) and one-way ANOVA in (D) LVEDD. Data that do not follow a normal distribution are presented as median with IQR (25th-75th percentile) in D (body weight, LVESD, FS, ejection fraction) and F. The corresponding significant P values are shown. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. LVEF = left ventricular ejection fraction; LVFS = left ventricular fractional shortening; other abbreviations as in Figure 1.

To determine the effect of Kdm5a deletion on survival, Kaplan-Meier survival analyses were performed, and cardiac function was assessed by transthoracic echocardiography. Myh6-Cre mice were included as controls to account for potential nonspecific effects of Cre recombinase and exhibited no detectable cardiac phenotype when evaluated at 2 months of age. CM-specific deletion of Kdm5a had no significant effect on survival up to 120 days of age, as shown in Figure 2C, and no effect on cardiac function at 2 months of age, as shown in Figure 2D and Table 1. Likewise, no increase in apoptosis was observed in any group, as assessed by TUNEL staining and quantitative analysis (Figures 2E and 2F). Our findings suggest that Kdm5a suppression is well tolerated in adult mice, consistent with its developmental downregulation in the postnatal period. To evaluate the impact of Kdm5a deletion in LMNA-CMP, we generated CM-specific double-knockout mice (Myh6-Cre:Lmna^F/F:Kdm5a^F/F). Because CM-specific deletion of Kdm5a alone did not cause a measurable abnormality in cardiac structure, function, survival, or apoptosis, the Kdm5a single knockout (Myh6-Cre: Kdm5a^F/F) group was not analyzed further in subsequent studies. These findings enable the determination of the effects of KDM5A activation on cardiac pathology in LMNA-CMP.

Table 1.

Echocardiographic Parameters in WT, Myh6-Cre, Kdm5a^F/F, and Myh6-Cre:Kdm5a^F/F Mice at 2 Months of Age

Groups	WT (n = 13)	Myh6-Cre (n = 15)	Kdm5aF/F (n = 12)	Myh6-Cre:Kdm5aF/F (n = 18)	P Value
Sex (M/F)	9/4	7/8	7/5	9/9	0.63
Age (days)	63.62 ± 2.06	63.40 ± 6.62	64 ± 5.34	66.56 ± 5.37	0.45
BW (g)	24.00 [22.00-28.00]	23.80 [19.10-25.60]	22.35 [20.63-27.50]	23.30 [21.15-27.63]	0.49
HR (bpm)	552 ± 40	537 ± 30	552 ± 45	548 ± 38	0.71
IVST (mm)	0.48 [0.47-0.52]	0.50 [0.48-0.55]	0.52 [0.48-0.54]	0.52 [0.51-0.54]	0.13
PWT (mm)	0.64 ± 0.06	0.59 ± 0.07	0.62 ± 0.10	0.65 ± 0.08	0.13
LVEDD (mm)	3.50 ± 0.27	3.33 ± 0.30	3.36 ± 0.40	3.43 ± 0.26	0.43
LVEDDi (mm/g)	0.15 [0.13-0.15]	0.15 [0.13-0.17]	0.15 [0.13-0.15]	0.15 [0.13-0.16]	0.66
LVESD (mm)	2.09 [2.00-2.27]	2.13 [1.95-2.40]	2.23 [1.85-2.39]	2.23 [1.96-2.41]	0.91
FS (%)	39.04 [36.53-40.54]	35.12 [32.45-36.62]	36.59 [33.36-39.71]	35.04 [30.85-40.10]	0.14
EF (%)	70.76 [67.35-72.37]	65.99 [62.12-67.94]	67.75 [62.81-71.07]	65.48 [59.50-71.79]	0.13
LVM (mg)	55.05 ± 11.71	47.87 ± 12.35	51.43 ± 14.30	54.86 ± 10.36	0.32
LVMi (mg/g)	1.82 ± 0.37	1.73 ± 0.33	1.78 ± 0.44	1.92 ± 0.32	0.46

Continuous data are presented as mean ± SD, and data that do not follow a normal distribution are presented as median with interquartile range [25th-75th percentile] and compared using one-way ANOVA or Kruskal-Wallis test as appropriate. Sex was compared using Fisher's exact test. P values for the primary group comparison are shown.

ANOVA = analysis of variance; bpm = beats/minute; BW = body weight; EF = ejection fraction; FS = fractional shortening; HR = heart rate; IVST = interventricular septal thickness; LVEDD = left ventricular end-diastolic diameter; LVEDDi = LVEDD indexed to body weight; LVESD = left ventricular end-systolic diameter; LVM = left ventricular mass; LVMi = left ventricular mass indexed to body weight; M/F = male/female; PWT = posterior wall thickness; WT = wild-type.

Successful deletion of Kdm5a and Lmna in CMs is illustrated in Figures 3A to 3C. Kdm5a deletion did not affect body weight in the Myh6-Cre:Lmna^F/F mice. Deletion of the Lmna gene, however, was associated with reduced body weight at 3 weeks of age, regardless of the Kdm5a deletion (Figure 3D, Table 2). We next assessed cardiac size and function by echocardiography at 3 weeks of age in the WT, Myh6-Cre:Lmna^F/F, and Myh6-Cre:Lmna^F/F:Kdm5a^F/F mice. As shown previously, the Myh6-Cre:Lmna^F/F mice had severe cardiac dilation and dysfunction.⁹ Deletion of the Kdm5a gene in the LMNA-CMP mice (Myh6-Cre:Lmna^F/F:Kdm5a^F/F) significantly improved left ventricular size and function, as evidenced by decreased left ventricular end-systolic diameters, LVEDD, and LVEDD indexes to body weight, and increased left ventricular fractional shortening and ejection fraction (Table 2, Figure 3D). Kaplan-Meier survival plot showed that deletion of the Kdm5a gene extended the median life span of Myh6-Cre:Lmna^F/F mice from 24 to 35 days and improved overall survival (log-rank, P < 0.001; Figure 3E). In agreement with these findings, as shown in Figure 3F, deletion of the Kdm5a gene attenuated the expression of the molecular markers of cardiac hypertrophy and HF in Myh6-Cre:Lmna^F/F:Kdm5a^F/F CMs as determined by quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR).

Figure 3.

Conditional Deletion of Kdm5a in Mice Attenuates Cardiac Dysfunction and Improves Survival of LMNA-Cardiomyopathy Mice

(A) Quantitative reverse-transcriptase polymerase chain reaction showing deletion of messenger RNA levels of Lmna and Kdm5a in the cardiac myocytes showing substantial reduction of Lmna in Myh6-Cre:Lmna^F/F and Myh6-Cre:Lmna^F/F:Kdm5a^F/F and Kdm5a in the latter group compared with wild-type controls. (B and C) Immunoblot and quantitative analysis showing reduction of lamin A (LMNA) in the Myh6-Cre:Lmna^F/F and Myh6-Cre:Lmna^F/F:Kdm5a^F/F and KDM5A in the Myh6-Cre:Lmna^F/F:Kdm5a^F/F group. The data confirmed deletion of both genes. (D) Body weight, heart rate, and selected echocardiographic indices were measured in wild-type, Myh6-Cre:Lmna^F/F, and Myh6-Cre:Lmna^F/F:Kdm5a^F/F mice (n = 14 per group) at P21. Parameters include body weight, LVESD, LVEDD, LVEDDi (LVEDD indexed to body weight), FS, and ejection fraction. (E) Kaplan-Meier survival curves demonstrating prolonged median and maximal survival in Myh6-Cre:Lmna^F/F:Kdm5a^F/F mice compared with Myh6-Cre:Lmna^F/F. (F) Transcript levels of selected markers of cardiac dysfunction measured by quantitative reverse-transcriptase polymerase chain reaction in isolated cardiomyocytes from P21 mice. Expression values were normalized to Gapdh n = 6 to 7. Normality of data distribution was assessed using the Shapiro-Wilk test. Normally distributed data are presented as mean ± SD and statistical significance was evaluated using one-way analysis of variance followed by Tukey's multiple comparison test in A,C,D and E (Nppb, Atp2a2, Myh6, and Myh7) and data that do not follow a normal distribution are presented as median with IQR (25th-75th percentile) and statistical significance was evaluated using the Kruskal-Wallis test with Dunn's post hoc test in (E) (Nppa, Acta1, Spp1, Cd44, Mmp9, and Sprf1). The corresponding P values are shown. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. Abbreviations as in Figures 1 and 2.

Table 2.

Echocardiographic Parameters in WT, Myh6-Cre:Lmna^F/F and Myh6-Cre:Lmna^F/F:Kdm5a^F/F at 3 Weeks of Age

Groups	WT (n = 14)	Myh6-Cre:LmnaF/F (n = 14)	Myh6-Cre:LmnaF/F :Kdm5aF/F (n = 14)	P Value
Sex (M/F)	8/6	7/7	10/4	0.62
Age (days)	20.00 ± 0.65	20.40 ± 0.91ac	21.36 ± 0.50ab	0.001
BW (g)	9.69 ± 2.20	7.86 ± 0.71a	9.1 ± 1.07	0.007
HR (bpm)	516 ± 20	510 ± 15	549 ± 47ab	0.003
IVST (mm)	0.39 ± 0.08	0.38 ± 0.08	0.40 ± 0.09	0.80
PWT (mm)	0.44 ± 0.03	0.41 ± 0.07	0.43 ± 0.08	0.47
LVEDD (mm)	2.70 ± 0.27	3.32 ± 0.30ac	2.86 ± 0.36b	<0.001
LVEDDi (mm/g)	0.29 ± 0.06	0.44 ± 0.06ac	0.32 ± 0.03b	<0.001
LVESD (mm)	1.51 ± 0.20	3.01 ± 0.35ac	2.10 ± 0.42ab	<0.001
FS (%)	43.96 ± 5.60	9.32 ± 4.60ac	28.63 ± 7.38ab	<0.001
EF (%)	81.92 ± 5.11	24.93 ± 10.40ac	60.04 ± 12.47ab	<0.001
LVM (mg)	20.40 ± 3.90	27.90 ± 7.80a	22.73 ± 5.88	0.007
LVMi (mg/g)	2.16 ± 0.41	3.53 ± 0.84ac	2.50 ± 0.61b	<0.001

Continuous data are presented as mean ± SD, and data that do not follow a normal distribution are presented as median with interquartile range (25th-75th percentile) and were compared using one-way ANOVA with Tukey's post-hoc test. Sex was compared using Fisher's exact test. P values for the primary group comparison are shown. Superscripts denote significant pairwise comparisons.

Abbreviations as in Table 1.

^aP < 0.05 compared to WT.

^bP < 0.05 compared to Myh6-Cre:Lmna^F/F.

^cP < 0.05 compared to Myh6-Cre:Lmna^F/F:Kdm5a^F/F.

Kdm5a deletion attenuates apoptosis in LMNA-CMP

The observed changes in cardiac function led us to analyze the apoptotic pathways, as cell death is an integral phenotype of LMNA-CMP^9,20,22 and HF in general. GSEA showed enrichment of genes involved in apoptosis in Myh6-Cre:Lmna^F/F, as compared with WT cardiac myocytes (Figure 4A). Deletion of the Kdm5a gene significantly suppressed pro-apoptotic gene expression signatures in Myh6-Cre:Lmna^F/F:Kdm5a^F/F mice, as demonstrated by GSEA plots derived from RNA-seq data (Figure 4A).

Figure 4.

Kdm5a Deletion Attenuates Apoptosis in Lamin A–Cardiomyopathy Hearts

(A) Gene set enrichment analysis of RNA-sequencing data showing significant induction of apoptotic gene signatures in Myh6-Cre:Lmna^F/F cardiomyocytes and reversal of this signature on Kdm5a deletion in Myh6-Cre:Lmna^F/F:Kdm5a^F/F cardiomyocytes. (B) Representative TUNEL staining of myocardial sections from P21 mice demonstrating increased apoptotic nuclei (green) in Myh6-Cre:Lmna^F/F hearts, counterstained with DAPI (blue). (C) Quantification of TUNEL⁺ nuclei expressed as percentage of total nuclei, showing a marked increase in Myh6-Cre:Lmna^F/F hearts compared with wild-type and a significant reduction on Kdm5a deletion. n = 6 to 8 per group. (D) Transcript levels of pro-apoptotic genes measured by quantitative reverse-transcriptase polymerase chain reaction from whole-heart RNA, normalized to Gapdh, showing upregulation in Myh6-Cre:Lmna^F/F and partial restoration in Myh6-Cre:Lmna^F/F:Kdm5a^F/F. n = 7 biological replicates per group were used. (E) Representative immunoblot analysis of cardiomyocyte extracts from P21 mice showing expression of Caspase-3 (CASP3), Bcl-2-associated X protein (BAX), Bcl2-associated agonist of cell death (BAD), and Receptor-interacting protein kinase-3 (RIPK3) across the 3 groups. (F) Quantification of immunoblot data (normalized to VCL) reveals increased expression of pro-apoptotic proteins in Myh6-Cre:Lmna^F/F hearts, which are attenuated following Kdm5a deletion. n = 7 to 8 biological replicates per group were used. Normality of data distribution was assessed using the Shapiro-Wilk test. Normally distributed data are presented as mean ± SD and statistical significance was evaluated using one-way analysis of variance followed by Tukey's multiple comparison test in B and C (Bid, Caspase 3, Gadd45b, Bcl2l2, Anxa1), and (F) (RIPK3). Data that do not follow a normal distribution are presented as median with IQR (25th-75th percentile) and statistical significance was evaluated using the Kruskal-Wallis test with Dunn's post hoc pairwise comparisons in C (Lgsl3, Gadd45a) and F (CASP3, BAD, and BAX). The corresponding P values are shown. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. Abbreviations as in Figures 1 and 2.

To assess the extent of apoptosis, we performed a TUNEL assay, which showed an increased number of TUNEL-positive nuclei in the myocardium of Myh6-Cre:Lmna^F/F mice compared with WT (Figures 4B and 4C). Conversely, the number of TUNEL-positive nuclei was markedly reduced in Myh6-Cre:Lmna^F/F:Kdm5a^F/F mice, as shown in Figures 4B and 4C. qRT-PCR analysis of the transcripts from whole heart was performed that showed upregulation of cell death gene in LMNA-CMP and partial rescue on Kdm5a deletion (Figure 4D). To determine if the effects are mediated from reduced cell death in CMs and whether deletion of the Kdm5a gene affected cell death protein in the LMNA-deficient CM, we detected the expression levels of selected cell death proteins, namely Caspase-3, Bcl2-associated agonist of cell death, Bcl-2-associated X protein, and Receptor-interacting protein kinase-3 in isolated CMs from the WT, Myh6-Cre:Lmna^F/F, and Myh6-Cre:Lmna^F/F:Kdm5a^F/F mice and control groups. Expression levels of Caspase-3, Bcl2-associated agonist of cell death, Bcl-2-associated X protein, and Receptor-interacting protein kinase-3 were elevated in CMs isolated from the Myh6-Cre:Lmna^F/F compared with those from the WT mice (Figures 4E and 4F) and partially rescued on Kdm5a deletion (Myh6-Cre:Lmna^F/F:Kdm5a^F/F) mice. Finally, we determined the expression levels of selected cell death proteins in CMs at an early disease stage (P14) before the onset of overt cardiac dysfunction. The expression levels of selected cell death markers were unchanged at P14 as shown in Supplemental Figures 2A and 2B. The finding suggests that the cell death is likely a downstream effect of Kdm5a target dysregulation in LMNA-CMP and by inference, attenuation of cell death likely reflects a downstream consequence of restoring KDM5A-regulated gene programs.

Kdm5a deletion rescues myocardial fibrosis in Myh6-Cre:Lmna^F/F mice

To determine the effect of Kdm5a deletion on cardiac phenotypes, cardiac fibrosis was assessed by staining thin myocardial sections with picrosirius red, and the CVF was calculated, which is used as a quantitative marker of fibrosis. The CVF was significantly increased in Myh6-Cre:Lmna^F/F hearts, comprising 9.8% ± 3.8% of the myocardium (Figures 5A and 5B), compared with approximately 1% observed in WT hearts. This increase was notably attenuated in Myh6-Cre:Lmna^F/F:Kdm5a^F/F mice, as indicated by a significant reduction in CVF to 1.5% ± 0.85% (Figure 5B). qRT-PCR analyses of transcripts associated with fibrosis revealed increased expression levels of Tgfb1, Tgfb2, and Col1a1 among others in the Myh6-Cre:Lmna^F/F hearts, which were significantly attenuated on Kdm5a deletion in Myh6-Cre:Lmna^F/F:Kdm5a^F/F hearts (Figure 5C).

Figure 5.

Kdm5a Deletion Rescues Myocardial Fibrosis in Myh6-Cre:Lmna^F/F Mice

(A and B) Representative picrosirius red–stained myocardial sections from 3-week-old wild-type, Myh6-Cre:Lmna^F/F, and Myh6-Cre:Lmna^F/F:Kdm5a^F/F mice, with corresponding quantitative analysis of collagen volume fraction (CVF). CVF was markedly elevated in Myh6-Cre:Lmna^F/F hearts (9.8% ± 3.8%) compared with wild-type (∼1%), and this increase was significantly attenuated in Myh6-Cre:Lmna^F/F:Kdm5a^F/F hearts (1.5% ± 0.85%). Collagen volume fraction was quantified using ImageJ. For each genotype, n = 5 mouse hearts were used. (C) Transcript levels of fibrosis-associated genes in wild-type, Myh6-Cre:Lmna^F/F, and Myh6-Cre:Lmna^F/F:Kdm5a^F/F hearts, measured by quantitative reverse-transcriptase polymerase chain reaction. n = 8 biological replicates per group were used. Normality of data distribution was assessed using the Shapiro-Wilk test. Normally distributed data are presented as mean ± SD and statistical significance was evaluated using one-way analysis of variance followed by Tukey's post hoc pairwise comparison in B and C (Tgfb1, Tgfb2, Col1a1, Col3a1, Col6a1, Ctgf, and Vim). Data that do not follow a normal distribution are presented as median with IQR (25th-75th percentile) and statistical significance was evaluated using the Kruskal-Wallis test with Dunn's post hoc pairwise comparisons in C (Tgfb3, Col5a2, and Postn). The corresponding P values are shown. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

Deletion of Kdm5a attenuates dysregulation of gene expression in LMNA-CMP

To determine the molecular mechanisms underlying the rescue of HF phenotypes on deletion of the Kdm5a gene, we performed RNA-seq analysis on isolated cardiac myocytes from WT, Myh6-Cre:Lmna^F/F, and Myh6-Cre:Lmna^F/F:Kdm5a^F/F CMs. Principal component analysis of the RNA-seq data revealed distinct genotype-dependent segregation of the samples (Figure 6A). Analysis of the RNA-seq data between the Myh6-Cre:Lmna^F/F and WT cardiac myocytes led to the identification of 6,609 differentially expressed genes, composed of 3,258 upregulated and 3,351 downregulated genes (Figure 6B). Deletion of Kdm5a attenuated dysregulation of gene expression in Myh6-Cre:Lmna^F/F:Kdm5a^F/F cardiac myocytes, as 1,905 genes were differentially expressed compared with Myh6-Cre:Lmna^F/F (Figure 6C). Of these, 1,458 differentially expressed genes were considered rescued as their expression level was normalized upon Kdm5a deletion, that is, was similar to that in the WT myocytes (Figure 6D). GSEA of known KDM5A target genes indicated their suppression in the Myh6-Cre:Lmna^F/F and their restoration following the deletion of the Kdm5a gene (Figure 6E), which supports the role of KDM5A-mediated epigenetic remodeling in the pathogenesis of HF.

Figure 6.

Kdm5a Deletion in Lamin A–Cardiomyopathy Restores OXPHOS, FAO, and Myogenesis Gene Expression

(A) Principal component analysis of RNA-sequencing data from wild-type (black), Myh6-Cre:Lmna^F/F (red), and Myh6-Cre:Lmna^F/F:Kdm5a^F/F (green) cardiomyocytes showing distinct clustering of transcriptional profiles. Each dot represents an individual biological replicate. (B) Volcano plot of differentially expressed genes in Myh6-Cre:Lmna^F/F vs wild-type hearts, highlighting significantly upregulated (red) and downregulated (blue) genes. (C) Volcano plot of differentially expressed genes in Myh6-Cre:Lmna^F/F:Kdm5a^F/F vs Myh6-Cre:Lmna^F/F CM, showing transcriptional rescue on Kdm5a deletion. (D) The Venn diagram depicting the overlap of differentially expressed genes showing rescue after Kdm5a loss and heatmap of rescue gene showing widespread transcriptional alterations in Myh6-Cre:Lmna^F/F compared with wild-type, with partial restoration of expression patterns following Kdm5a deletion. (E) Gene set enrichment analysis plots for KDM5A target genes. Left: Myh6-Cre:Lmna^F/F vs wild-type, showing significant enrichment of KDM5A targets. Right: Myh6-Cre:Lmna^F/F:Kdm5a^F/F vs Myh6-Cre:Lmna^F/F, showing reversal of KDM5A-driven transcriptional repression. (F) Hallmark pathway enrichment analysis highlighting rescue of oxidative phosphorylation (OXPHOS), fatty acid oxidation (FAO), and myogenesis pathways on Kdm5a deletion. (G) Downregulated hallmark pathways in Myh6-Cre:Lmna^F/F:Kdm5a^F/F relative to Myh6-Cre:Lmna^F/F, including tumor necrosis factor (TNF)α signaling, hypoxia, epithelial-to-mesenchymal transition (EMT), and p53 pathways. (H) Ingenuity Pathway Analysis predicted upstream regulators that show inhibition in Myh6-Cre:Lmna^F/F:Kdm5a^F/F compared with Myh6-Cre:Lmna^F/F, including KDM5A, TP53, and HIF1A. (I) Ingenuity Pathway Analysis predicted activated upstream regulators in Myh6-Cre:Lmna^F/F:Kdm5a^F/F compared with Myh6-Cre:Lmna^F/F, including estrogen-related receptor alpha (ESRRA), and Peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC1A). Abbreviation as in Figure 1.

Pathway analysis of genes whose transcripts were rescued on deletion of the Kdm5a gene in LMNA-CMP mice showed a marked upregulation of genes involved in OXPHOS and myogenesis (Figure 6F), whereas those involved in inflammation and fibrosis were suppressed (Figure 6G). KDM5A was the top upstream regulator of the genes whose transcripts were rescued on deletion of the Kdm5a gene. In addition, TP53 and members of the SMAD family were among the upstream regulators whose targets were suppressed on deletion of the Kdm5a gene (Figure 6H); in contrast, targets of Peroxisome proliferator-activated receptor gamma coactivator 1-alpha (z-score = 5.1) and the ERR (z-score = 3.2) were induced (Figure 6I). Collectively, these data demonstrate that KDM5A deletion abrogates the pathogenic transcriptional program in Myh6-Cre:Lmna^F/F myocytes by restoring expression of genes involved in OXPHOS and myogenesis and inducing expression of the key transcription factors, such as the ERR family.

Genome-wide redistribution of H3K4me3 marks in LMNA-CMP CMs and their restoration on Kdm5a deletion

KDM5A preferentially targets cis-regulatory genomic regions that are enriched with H3K4me3 marks for their removal. Therefore, to elucidate the epigenetic changes associated with the deletion of the Kdm5a gene, we determined genome-wide H3K4me3 occupancy in cardiac myocytes isolated from WT, Myh6-Cre:Lmna^F/F, and Myh6-Cre:Lmna^F/F:Kdm5a^F/F mice by CUT&RUN assays. Approximately 15 million uniquely mapped reads per sample were obtained and analyzed, which led to the identification of 21,324 ± 2,550 H3K4me3 peaks in WT, 20,042 ± 1,917 peaks in Myh6-Cre:Lmna^F/F, and 23,775 ± 2,150 peaks in Myh6-Cre:Lmna^F/F:Kdm5a^F/F samples (Figure 7A). Genomic annotation revealed that differential H3K4me3 peaks were predominantly localized to promoters and transcription start site (TSS)-proximal regions, with additional enrichment at intronic and exonic regulatory elements (Figure 7B). Overall genomic distribution and peak length of H3K4me3 were comparable across all 3 genotypes (Figure 7C). Consistent with canonical H3K4me3 localization, average count per million signal profiles demonstrated strong enrichment at TSSs, and heatmap analysis showed robust H3K4me3 signal accumulation within ±3 kb of TSSs in all groups (Figure 7D, Supplemental Figure 3). Principal component analysis clearly separated WT, Myh6-Cre:Lmna^F/F, and Myh6-Cre:Lmna^F/F:Kdm5a^F/F samples, indicating genotype-dependent global differences in H3K4me3 landscapes (Figure 7E).

Figure 7.

Genome-Wide H3K4me3 Profiling Reveals Altered Promoter Enrichment in Lamin A–Cardiomyopathy and Rescue On Kdm5a Deletion

(A) Total number of H3K4me3 peaks identified by cleavage under targets and release using nuclease in wild-type (black), Myh6-Cre:Lmna^F/F (red), and Myh6-Cre:Lmna^F/F:Kdm5a^F/F (green) cardiomyocytes. Each dot represents an individual biological replicate; bars indicate mean ± SD. (B) Genomic annotation of H3K4me3 peaks across genotypes. (C) Mean H3K4me3 peak length per sample, showing similar distribution across genotypes. (D) Aggregate H3K4me3 signal profiles centered on genome-wide transcription start sites (TSSs ±5 kb), demonstrating promoter-proximal H3K4me3 enrichment. (E) Principal component analysis of genome-wide H3K4me3 profiles separates wild-type, Myh6-Cre:Lmna^F/F, and Myh6-Cre:Lmna^F/F:Kdm5a^F/F samples, indicating distinct chromatin states and partial normalization following Kdm5a loss. (F) Volcano plot showing differentially enriched H3K4me3 peaks in Myh6-Cre:Lmna^F/F relative to wild-type. Red, enriched; blue, depleted; gray, not significant. (G) Volcano plot of H3K4me3 changes in Myh6-Cre:Lmna^F/F:Kdm5a^F/F relative to Myh6-Cre:Lmna^F/F, showing restoration of promoter-associated H3K4me3 peaks. (H) Average profiles and heatmaps of H3K4me3 signal at peaks enriched in Myh6-Cre:Lmna^F/F:Kdm5a^F/F relative to Myh6-Cre:Lmna^F/F, demonstrating recovery of promoter H3K4me3 occupancy toward wild-type levels. (I) Gene ontology and pathway enrichment analysis of genes associated with rescued H3K4me3 peaks, revealing restoration of muscle contraction, cardiac conduction, metabolic, and nuclear receptor signaling programs. (J) Average profiles and heatmaps of H3K4me3 signal at peaks depleted in Myh6-Cre:Lmna^F/F:Kdm5a^F/F relative to Myh6-Cre:Lmna^F/F, indicating attenuation of aberrantly elevated promoter marks. (K) Pathway enrichment analysis of genes associated with depleted peaks, highlighting downregulation of RHO GTPase signaling, extracellular matrix organization, and vascular endothelial growth factor (VEGF)-associated pathways. H3K4me3 = histone H3 lysine 4 trimethylation.

Differential peak analysis between Myh6-Cre:Lmna^F/F and WT CMs identified 444 enriched and 862 depleted H3K4me3 peaks in the Myh6-Cre:Lmna^F/F group (Figure 7F). Notably, deletion of Kdm5a in the Myh6-Cre:Lmna^F/F background resulted in redistribution of H3K4me3 at 1,425 genomic loci, comprising 501 enriched and 924 depleted peaks in Myh6-Cre:Lmna^F/F:Kdm5a^F/F CMs (Figure 7G).

Heatmap and meta profile analyses revealed a subset of H3K4me3 peaks that were depleted in Myh6-Cre:Lmna^F/F CMs but restored on Kdm5a deletion (Figure 7H), suggesting these loci represent direct or primary targets of KDM5A-mediated demethylation. Pathway analysis showed that these rescued H3K4me3 peaks are enriched at genes essential for cardiac function (Figure 7I). Conversely, H3K4me3 peaks that were increased in Myh6-Cre:Lmna^F/F CMs and subsequently reduced in Myh6-Cre:Lmna^F/F:Kdm5a^F/F CMs (Figure 7J) likely reflected secondary effects associated with phenotypic rescue or indirect regulation through transcription factor–dependent mechanisms. Pathway analysis showed these peaks are localized at inflammatory and extracellular matrix–related genes implicated in fibrosis and cell death (Figure 7K, Supplemental Tables 2A and 2B).

A total of 388 H3K4me3 peaks were reciprocally regulated in Myh6-Cre:Lmna^F/F:Kdm5a^F/F compared with Myh6-Cre:Lmna^F/F (Supplemental Tables 2A and 2B). Incorporating the transcript data with the H3K4me3 peaks led to the identification of 189 genes whose transcript levels, as well as H3K4me3 peaks, were rescued. This subset included key cardiac transcription factors T-box 5 (TBX5) and members of the ERR family of proteins. The list of loci that showed concordant changes in both transcript levels and promoter H3K4me3 signal in Myh6-Cre:Lmna^F/F and were normalized on Kdm5a deletion is provided in Supplemental Figure 4 and Supplemental Table 3. Collectively, these findings suggest that KDM5A mediates targeted demethylation of H3K4me3 at promoter regions of genes involved in cardiac function and OXPHOS.

Suppression of cardiac and OXPHOS transcription factors TBX5 and ERR in LMNA-CMP

RNA-seq analysis revealed extensive downregulation of cardiac structural genes and FAO, and OXPHOS in LMNA-CMP CMs, changes that were restored on Kdm5a deletion. Among the genes that exhibited concordant promoter H3K4me3 alterations (Supplemental Tables 2 and 3), the transcription factors TBX5 and estrogen-related receptor gamma (ESRRG) emerged as key regulators, suggesting that they further mediate and reinforce KDM5A-driven gene suppression. Supporting this, H3K4me3 enrichment was markedly reduced at the promoters and cis-regulatory elements of Tbx5 (Figure 8A) in LMNA-CMP CMs and partially restored following Kdm5a deletion (Figure 8A). Importantly, qRT-PCR analysis was performed using CMs isolated from early-stage hearts at P14, before the onset of overt cardiac dysfunction. This analysis showed that Tbx5 transcript levels were already markedly reduced in LMNA-CMP CMs and were restored on Kdm5a deletion (Figure 8B). Likewise, downstream TBX5 target genes, including Irx4, Hopx, Myl2, Myl3, Tnni3, Tnnt2, Myo5c, Tcap, and several potassium channel genes were significantly suppressed in LMNA-CMP myocytes and re-expressed in Kdm5a-deficient hearts, as shown by qRT-PCR (Figure 8C). H3K4me3 and RNA levels for another transcription factor, namely Esrrg, showed reduction in LMNA-CMP and rescue on Kdm5a deletion (Figures 8D and 8E). ESRRG-regulated targets including Esrra and genes within the OXPHOS and FAO pathways, together with genes involved in mitochondrial translation and ribosome biogenesis, were consistently downregulated in LMNA-CMP CM and rescued in Kdm5a-deficient CMs (Figures 8F and 8G, Supplemental Figure 5). These findings are consistent with our previous data in induced pluripotent stem cell (iPSC)-derived CMs, where inhibition of KDM5 led to increased expression of Tbx5, Esrra, and Esrrg.²⁸ Together, these data indicate that KDM5A is associated with suppression of cardiac structural and mitochondrial metabolic programs, in part through promoter-proximal H3K4me3 remodeling and in part through transcriptional networks involving TBX5 and ESRRG.

Figure 8.

Deletion of Kdm5a Restores Expression of Cardiac and OXPHOS Transcription Factors and Their Genes in Lamin A–Cardiomyopathy

(A) Representative Integrative Genomics Viewer genome browser tracks showing H3K4me3 enrichment at cardiac transcription factor Tbx5 cis-regulatory elements and promoter region. Tracks shown on mm39 genome assembly. Signals from wild-type (black), Myh6-Cre:Lmna^F/F (red), and Myh6-Cre:Lmna^F/F:Kdm5a^F/F (green) hearts are displayed. (B) Quantitative reverse-transcriptase polymerase chain reaction analysis of Tbx5 transcript levels across genotypes (n = 7 per group), demonstrating suppression in Myh6-Cre:Lmna^F/F and rescue on Kdm5a deletion. (C) Quantitative reverse-transcriptase polymerase chain reaction analysis of downstream targets of T-box 5 (n = 7 per group), demonstrating restoration of expression in Myh6-Cre:Lmna^F/F:Kdm5a^F/F. (D) Representative Integrative Genomics Viewer genome browser tracks showing H3K4me3 enrichment at transcription factor Esrrg cis-regulatory and promoter region. Signals from wild-type (black), Myh6-Cre:Lmna^F/F (red), and Myh6-Cre:Lmna^F/F:Kdm5a^F/F (green) hearts are displayed. (E) Quantitative reverse-transcriptase polymerase chain reaction analysis of Esrrg transcript levels across genotypes (n = 7 per group), demonstrating suppression in Myh6-Cre:Lmna^F/F and rescue on Kdm5a deletion. (F and G) Quantitative reverse-transcriptase polymerase chain reaction analysis of ESRRG target genes encoding OXPHOS complex (Cox7a1, Cox7a2, Ndufa6, Ndufa3, Sdhb, Cox10, Ndufs2, Uqcr2, and Uqcr10) and fatty acid oxidation regulators (Cpt1b, Acadvl, Acadm, and Echdc3) in Myh6-Cre:Lmna^F/F, all of which were significantly restored on Kdm5a deletion (n = 6-7 per group). Gene set enrichment analysis demonstrating overall suppression of fatty acid metabolism and OXPHOS in Myh6-Cre:Lmna^F/F compared with wild-type and partially restored in Myh6-Cre:Lmna^F/F:Kdm5a^F/F. Normality of data distribution was assessed using the Shapiro-Wilk test. Normally distributed data are presented as mean ± SD and statistical significance was evaluated using one-way analysis of variance followed by Tukey's post hoc pairwise comparisons in B (Tbx5), C (HopX, Irx4, Myl3, Myl2, Myo5C, Tnni3, Kcnd2, Kcnq1, Kcnip2, and Kcnh2), E, and F (Esrra, Cox7a1,Cox7a2, Ndufa6, Sdhb, Ndufs2,Uqcr10, Cpt1b, Acadvl, and Acadm). Data that do not follow a normal distribution are presented as median with IQR (25th-75th percentile) and statistical significance was evaluated using the Kruskal-Wallis test with Dunn's post hoc pairwise comparisons in C (Tnnt2, Tcap, and Kcnj2) and F (Cox10, Uqcr2, Ndufa3, and Echdc3). The corresponding P values are shown. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. Abbreviations as in Figures 6 and 7.

Discussion

HF is a complex and multisystem disorder leading to cardiac dysfunction and representing the major cause of morbidity and mortality.^1,2 Irrespective of the initiating cause, HF progression is driven by convergent molecular derangements encompassing impaired mitochondrial oxidative capacity, disrupted sarcomere integrity, and widespread metabolic reprogramming.^3,7,10,31, 32, 33 In this work, we demonstrate KDM5A is upregulated in CMs in a mouse model of HF (Myh6-Cre:Lmna^F/F), coinciding with repression of genes required for cardiac contractility and OXPHOS. These changes likely contribute to impaired bioenergetics, cell death, and progressive cardiac dysfunction. We further show that genetic deletion of the histone demethylase KDM5A is sufficient to attenuate cardiac and OXPHOS gene dysregulation and lead to prolonged survival, improved cardiac function, attenuated myocardial cell death, and reduced fibrosis in a mouse model of HF.

Moreover, we have previously demonstrated that KDM5A is re-expressed and active in human HF samples, in Lmna^−/− mouse hearts, and in a mouse model of desmoplakin CMP.^19,20,34 Others have independently reported dysregulation of KDM5 family members in human HF samples.³⁵ Furthermore, KDM5A and B also regulate cardiac fibroblast function in angiotensin II–induced myocardial fibrosis.^36,37 In addition, we have reported the beneficial effects of chemical inhibition of the KDM5 family in iPSC-CMs, which support the potential translational utility of the findings.²⁸ Together, our data, in agreement with the existing data, support the generalizability of reactivation of KDM5A in HF and its translational utility as a potential therapeutic target.

Our findings suggest that key cardiac transcription factors, namely TBX5 and mitochondrial regulators ESRRG/estrogen-related receptor alpha (ESRRA), are targets of KDM5A among others. The roles of these factors in HF and especially in cardiac gene expression and mitochondrial function are well established.38, 39, 40 Consistent with this, our previous data in iPSC-CMs showed that inhibition of KDM5 increased H3K4me3 deposition at the ESRRA promoter, accompanied by elevated RNA and protein levels.²⁸ Furthermore, knockdown of ESRRA in iPSC-CM with concomitant KDM5 inhibition led to normalization of ESRRA targets in the iPSC-CMs. These data provide further evidence that KDM5A directly targets ERR for repression.²⁸ In addition, our current study uncovers the regulation of TBX5, a key cardiac transcription factor, by KDM5A, which is a novel finding.

Our findings uncover a previously unrecognized role of KDM5A in the pathogenesis of LMNA-CMP. Genome-wide profiling revealed a global redistribution of H3K4me3 peaks at promoter in LMNA-deficient CMs, despite the loss of lamin-associated domains, which are typically associated with transcriptional repression.^19,41,42 This suggests that KDM5A activation establishes a unique repressive state in LMNA-deficient CMs. Interestingly, attenuation of LMNA-CMP severity also has been reported following Bromodomain and Extra-Terminal (BET) bromodomain inhibition with JQ1, which similarly targets transcriptional programs as an epigenetic reader.^9,43 Although mechanistically distinct, the BET proteins exacerbate pathological programs by inducing transcriptional elongation, whereas the KDM5A seems to suppress cardiac and OXPHOS gene programs; however, both BET inhibition and KDM5A deletion converge on restoring cardiac function, suggesting that modulation of epigenetic readers and erasers represents a shared vulnerability in LMNA-driven CMP and HF. Whether combined targeting of these pathways would yield additive benefit remains an important question for future investigation.

The KDM5 family of histone demethylases regulates transcription by removing methyl groups from H3K4me3/2/1 and thereby suppressing expression. However, KDM5A can also activate or modulate gene expression by interacting with co-factors and cell-type–specific transcription factors.^18,44, 45, 46, 47 In this study, although several genes exhibited direct promoter-associated H3K4me3 changes in a KDM5A-dependent manner, a distinct subset showed restored expression on Kdm5a deletion without detectable changes in H3K4me3 deposits in their promoters. The former likely represent direct KDM5A targets, whereas the latter reflect downstream effects mediated through transcription factors such as TBX5, ESRRA, and ESRRG. We also observed that LMNA-CMP hearts, like most forms of HF, displayed upregulation of inflammatory and epithelial-to-mesenchymal transition–fibrosis-related gene programs. These changes may reflect indirect consequences of CM injury, cell death, and activation of secondary inflammation, as discussed previously. In particular, fibrosis in LMNA-CMP may arise from the induction of profibrotic signaling from diseased CMs, including increased expression of transforming growth factorβ1, pro-inflammatory cytokines. The beneficial effects of the deletion of Kdm5a on myocardial fibrosis may reflect partial restoration of the expression of genes involved in CM function, FAO, and OXPHOS.

Finally the previously described changes may be attributable to either indirect effects or as a result of H3K4me3-independent functions of KDM5A as part of a multiprotein complexes, including those containing retinoblastoma protein (RB1), Sin3B/NuRD complexes, and transcription factors including E2Fs and ZBTB20.^44,46,47 Such scaffolding interactions may also explain why Kdm5a deletion normalizes transcription even in the absence of measurable promoter H3K4me3 changes. In addition, KDM5A may contribute to disease progression by directly engaging pro-inflammatory transcription factors, repressing anti-inflammatory regulators, or indirectly activating stress-responsive pathways downstream of mitochondrial dysfunction and cell death. Collectively, these findings highlight the broad regulatory influence of KDM5A across multiple transcriptional networks in the failing heart and establish it as a driver of HF pathogenesis.

The LMNA-CMP model provides a severe and well-defined genetic model of CMP to interrogate the epigenetic mechanisms driving HF progression. Although complete loss of LMNA represents a severe phenotype, this model demonstrates the key pathological features seen in heterozygous LMNA deficiency and disease-causing LMNA mutations, including cardiac dysfunction, apoptosis, fibrosis, and mitochondrial abnormalities.^9,48,49 Because disease progression is rapid and severe, this model allows mechanistic insights to be obtained acutely. In this context, a partial improvement in cardiac function and survival are meaningful, given the severe nature of the disease and the multiple pathways involved in HF. Importantly, interventions that show benefit in this setting are likely to have equal or greater impact in milder, clinically relevant LMNA-CMP models.

The cause of death in the LMNA-CMP and the salubrious effects of the deletion of the Kdm5a gene in CMs in the LMNA-CMP mice could not be ascertained with certainty. We posit that increased mortality in the LMNA-CMP mice is likely due to cardiac systolic dysfunction resulting from increased CM cell death, inflammation, and fibrosis, and reduced expression of genes coding for sarcomere proteins and FAO. Likewise, the beneficial effect of deletion of the Kdm5a gene on survival likely reflects the attenuation or reversal of these molecular and cellular events in the Myh6-Cre:Lmna^F/F:Kdm5a^F/F mice.

Although the precise mechanism underlying KDM5A induction in CMs remains unknown, it is noteworthy that KDM5A has been identified as an interacting partner of RB1.50, 51, 52 Notably, RB1 expression is regulated by its interaction with LMNA.^53,54 Therefore, in LMNA-CMP and in HF, the loss of RB1 may stabilize KDM5A and facilitate its relocalization to gene promoters, thereby providing a plausible mechanism for its aberrant activation. Mice with the deletion of Kdm5a alone were not included in the mechanistic studies, because they did not show a discernible phenotype. Nevertheless, the possibility that some of the transcriptional or chromatin effects reflect baseline consequences of Kdm5a loss rather than disease-specific reversal could not be excluded.

Together, these findings implicate KDM5A as a key epigenetic regulator orchestrating pathological transcriptional programs in HF, encompassing genes governing mitochondrial metabolism, contractile function, and OXPHOS. Notably, systemic or CM-specific deletion of Kdm5a was well tolerated in adult mice,⁴⁵ underscoring the therapeutic viability of its inhibition. Thus, by revealing KDM5A as a driver of disease progression, this study highlights a rational avenue for intervention, supported by the growing repertoire of selective KDM5 inhibitors and RNA interference–based strategies capable of mitigating pathological remodeling in LMNA-CMP and broader HF contexts.

Funding Support and Author Disclosures

Dr Gurha is supported by National Institutes of Health (NIH), National Heart, Lung, and Blood Institute (NHLBI) award R01HL174481. Dr Marian is supported by NIH, NHLBI awards R01HL151737 and R01HL132401 and National Institute on Aging award R01AG082751. Dr Rouhi is supported by American Heart Association Career Development Award 23CDA1053101. This work was also supported by the John Dunn Foundation (Dr Gurha), Houston, Texas 77098, USA, and the Ewing Halsell Foundation (Dr Marian), San Antonio, Texas 78205, USA. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.

Perspectives.

COMPETENCY IN MEDICAL KNOWLEDGE: The findings of the present study inform clinicians and researchers that activation of the histone demethylase KDM5A contributes to the pathogenesis of LMNA-CMP by suppressing CM gene programs required for myocardial structure, metabolism, and function.

TRANSLATIONAL OUTLOOK: KDM5A is activated in LMNA-CMP and contributes to cardiac dysfunction, fibrosis, cell death, and premature mortality. Genetic deletion of Kdm5a in CMs in a mouse model of LMNA-CMP improves cardiac function and restores the expression of genes involved in oxidative phosphorylation, fatty acid metabolism, myogenesis, and cardiac transcriptional regulation. These findings identify KDM5A-dependent epigenetic dysregulation as a driver of HF and support further investigation of KDM5A inhibition as a potential therapeutic strategy for LMNA-CMP and possibly other forms of HF characterized by maladaptive epigenetic remodeling.