Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Aug 1.
Published in final edited form as: J Mol Cell Cardiol. 2020 Jun 10;145:30–42. doi: 10.1016/j.yjmcc.2020.06.001

Global analysis of histone modifications and long-range chromatin interactions revealed the differential cistrome changes and novel transcriptional players in human dilated cardiomyopathy

Chia-Feng Liu 1, Armen Abnousi 2, Peter Bazeley 2, Ying Ni 3, Michael Morley 5, Christine S Moravec 1, Ming Hu 2, W H Wilson Tang 1,4
PMCID: PMC7483840  NIHMSID: NIHMS1604774  PMID: 32533974

Abstract

Background:

Acetylation and methylation of histones alter the chromatin structure and accessibility that affect transcriptional regulators binding to enhancers and promoters. The binding of transcriptional regulators enables the interaction between enhancers and promoters, thus affecting gene expression. However, our knowledge of these epigenetic alternations in patients with heart failure remains limited.

Methods and Results:

From the comprehensive analysis of major histone modifications, 3-dimensional chromatin interactions, and transcriptome in left ventricular (LV) tissues from dilated cardiomyopathy (DCM) patients and non-heart failure (NF) donors, differential active enhancer and promoter regions were identified between NF and DCM. Moreover, the genome-wide average promoter signal is significantly lower in DCM than in NF. Super-enhancer (SE) analysis revealed that fewer SEs were found in DCM LVs than in NF ones, and three unique SE-associated genes between NF and DCM were identified. Moreover, SEs are enriched within the genomic region associated with long-range chromatin interactions. The differential enhancer-promoter interactions were observed in the known heart failure gene loci and are correlated with the gene expression levels. Motif analysis identified known cardiac factors and possible novel players for DCM.

Conclusions:

We have established the cistrome of four histone modifications and chromatin interactome for enhancers and promoters in NF and DCM tissues. Differential histone modifications and enhancer-promoter interactions were found in DCM, which were associated with gene expression levels of a subset of disease-associated genes in human heart failure.

Keywords: Epigenetics, histone modifications, chromatin interactions, heart failure

1. Introduction:

Heart failure (HF) is a leading cause of morbidity and mortality around the world [1]. When a heart is failing, it undergoes structural and functional remodeling, which results in various genomic and transcriptional changes and reprogramming in the heart [2]. Accumulating evidence has indicated that epigenetic regulations play an essential role in the development and pathogenesis of HF [3–5]. Specifically, post-translational modifications (PTMs) on histones play a critical role in the epigenetic regulation of gene expression. Acetylation and methylation are two major histone PTMs that can affect the chromatin structure and accessibility to transcriptional regulators. These changes allow or prevent transcriptional regulators, such as transcription factors (TF) and co -activating proteins, to bind to cis-regulating elements of a gene, e.g., promoters and enhancers. Gene expression levels can be dramatically increased through the cooperation between promoters and distal enhancers forming enhancer-promoter interactions, and such interactions are facilitated by the binding of TFs [6]. Enhancers can be located thousands of nucleotides away from their associated genes to regulate the transcription activity of those genes [6]. Additionally, multiple enhancers can form large clusters, namely super-enhancers (SE), with a high level of enhancer activity and can serve as a transcriptional hub allowing binding of master regulators, coactivators, and chromatin factor [7]. SEs are highly linked to genes essential for cell identity as well as disease development [7]. The 3D chromatin interactions (chromatin interactome) may mediate enhancer activity, and it may also influence the action of SEs in disease development.

Animal studies have shown that chromatin architecture is linked to HF, including long-range chromatin interactions [8, 9]. However, our understanding of the regulation of chromatin interaction and histone modifications in human HF remains unclear. Herein, we leveraged next-generation sequencing approaches to investigate transcriptome, epitome, and long-range chromatin interactions for enhancer and promoter in the genome of the human non-failing (NF) and dilated cardiomyopathy (DCM) left ventricular (LV) tissues.

2. Materials and methods

For full details, please see the expanded material and methods.

2.1. Patients and tissue samples

The study used left ventricular tissues from three male patients who underwent orthotopic cardiac transplantation for dilated cardiomyopathy (DCM) and three male patients who died of non-cardiac causes as non-failing (NF) heart control groups for chromatin immunoprecipitation sequencing (ChIP-seq). Two of the left ventricular tissues in each group from the same DCM and NF patients were used for long-range chromatin interactions by proximity ligation-assisted ChIP-seq (PLAC-seq). The study was approved by the Cleveland Clinic Institutional Review Board.

2.2. Chromatin immunoprecipitation-sequencing (ChIP-seq) and quantification-PCR (ChIP-qPCR)

The ChIP-seq procedures followed the protocol as previously described with some modifications [10]. Five nanograms of ChIP-ed DNA was used per ChIP reaction. TruSeq ChIP library preparation kit was used for library preparation (Illumina) and single-end reads were obtained using a HiSeq 4000 platform (Genomics Core Facility, University of Chicago). ChIP-seq results were validated by real-time PCR using a Quantstudio 3 Real-Time PCR system (Thermo Fisher Scientific) and SYBR Green. Primers used for ChIP-qPCR are listed in Table S2.

2.3. ChIP-seq data analysis

Sequence reads were aligned to the human reference genome (hg19) using Bowtie2 (Galaxy version 2.3.4.2) with the default setting [11, 12]. Additionally, eight histone modification ChIP-seq and two input fastq files from left ventricular tissues of two non-failing (NF) human donors were downloaded from Gene Expression Omnibus (GEO) database (SAMN00847538 and SAMN00847532) and were aligned as described above. Ngs.plot was used to generate the average peak profiles of histone modification across the genomic regions of interest for each sample [13]. MACS 2 was used for peak calling (nomodel; extension size: 200; FDR [q-value] cutoff: 0.05) [11, 14]. The correlation of ChIPseq was performed using plotCorrelation in Deeptool program (Galaxy Version 3.3.2.0.0) and using merged bam files of the individual from the same group ( n=4 in NF, n=3 in DCM). HOMER was used for super-enhancer detection and peak annotation [15]. Genomic Regions Enrichment of Annotation Tool (GREAT) [16] and Ingenuity Pathway Analysis (IPA, QIAGEN Inc., https://www.qiagenbioinformatics.com/) were used for gene ontology analysis.

2.4. Proximity Ligation-Assisted ChIP-seq (PLAC-seq)

PLAC-seq and library construction has followed the protocol as described in Fang et al. with some modifications [17]. The quantification of the purified libraries was confirmed using a dsDNA measurement (Qubit TM dsDNA HS assay, Q32832, ThermoFisher Scientific) with a Qubit fluorometer (Qubit 4 “Fluorometer, ThermoFisher Scientific). A Bioanalyzer (Agilent 2100 Bioanalyzer, Agilent) was used to determine the size distribution of the libraries. Paired-end reads were obtained using a HiSeq 4000

We applied the Model-based analysis of the PLAC platform (Genomics Core Facility, University of Chicago).

2.5. PLAC-seq data analysis

We applied the Model-based analysis of the PLAC-seq (MAPS) pipeline to analyze PLAC-seq data as previously described[18]. PLAC-seq and ChIP-seq data were visualized using WashU Epigenome Browser [19].

2.6. RNA-seq and data analysis

RNA-seq analysis used the dataset from Myocardial Applied Genomics Network (MAGNet). Total RNAs from patients’ and donors’ left ventricles were extracted as previously described [10, 20]. Pair-end reads were mapped to the human hg 19 reference genome and then for transcript assembling and quantification using STAR package (v2.5.2b) with default parameters [21] . Reads uniquely mapped were considered for further analysis. DEseq2 (v3.8), a program using count based matrices to identify differentially expressed genes, was used to determine gene signatures [22]. To improve the reliability and accuracy of differential expression analysis, only genes with raw counts > 5 in all individually sequenced samples were examined. Student’s t-test and Benjamini-Hochberg false discovery rate tests were performed for each of the differentially expressed genes across the groups. The count values were normalized to Transcripts Per Kilobase Million (TPM). Changes >1.5 fold with an adjusted p<0.05 were considered differentially expressed between groups.

2.7. qRT-PCR assay

Total RNA of human LV tissues was prepared using RNeasy fibrous tissue mini kit followed manufacturer instruction (Qiagen, 74704). cDNA was synthesized using SuperScript™ IV First-Strand Synthesis System (Thermo Fisher Scientific, 18091050). cDNA samples were amplified with inventoried TaqMan® Gene Expression Assays (Table S3) and analyzed by a Quantstudio 3 Real-Time PCR system (Thermo Fisher Scientific). Reactions were analyzed in triplicate, and relative mRNA levels were calculated using the 2−ΔΔCT method and normalized to those of 18S ribosomal RNA.

2.8. Western blot

Protein extracts were prepared from human left ventricles as decribed[23]. The total lane density of transferred proteins stained with Ponceau S was used to control for loading/transfer differences. Primary antibodies used were as follows: SOX4 (Diagenode, CS129100), BCL6B(Abgent, AP20369c), SMYD1 (Abcam, ab32489), CELEC3B ( Abcam, . ab108999), SERPINE1 (Abcam, ab66705) and SLC27A3 (Thermo Fisher Scientific 12943–1-AP). Secondary antibodies coupled to Alexa Fluor 680 (Invitrogen Molecular Probes) or IRDye 800 (LI-COR Biosciences) were used, and the Odyssey CLS infrared imager system (LI-COR Biosciences) was used for visualization of Western blot signals. Odyssey version 1.2 imaging software was used to process all images.

3. Results:

3.1. Promoter and enhancer histone landscape changes in DCM left ventricles

To understand the role of epigenetic regulations in HF, we generated genome-wide chromatin-state maps for four major histone modifications using human left ventricle heart (LV) tissues of 6 males from patients suffering from HF due to dilated cardiomyopathy (DCM, n=3) and from non-heart failing donors (NF, n=3). We carried out chromatin immunoprecipitation followed by sequencing (ChIP-seq) using antibodies against H3K27ac for active enhancer mark, H3K4me1 for poised and active enhancer mark, H3K4me3 for active promoter mark and H3K27me3 for the repressive mark. After the initial quantification and alignment of the reads to the human genome, we first performed the global analysis for histone modifications (Figure 1A). While the average signals for H3K27ac, H3K4me1, and H3K27me3 are similar among the samples, we observed that the active promoter mark, H3K4me3, was significantly decreased more than 45% in LV samples from DCM compared to the ones from NF (Figure 1B) except one of the NF samples (ID: NF3), whose H3K4me3 level was similar to DCM ones. After re-examining the donor’s demographics and clinical characters, we found that the left ventricular ejection fraction (LVEF) of NF3 was only 0.37, which was moderately below normal, suggesting this donor might have developed HF or cardiomyopathy without symptoms (Table S1). Therefore, we excluded the data from this donor for downstream analysis.

Figure 1. Genome-wide ChIP-seq analysis revealed the differential display of H3K27ac and H3K4me3 signatures between NF and DCM.

Figure 1.

Open in a new tab

(A) Diagram of the experimental workflow. (B) Average profiles of histone modifications from five NF and three DCM samples. (C) Venn diagram showing the numbers and percentage of histone modification enriched regions overlapped between NF (n=4) and DCM (n=3). (D) Biological replicates (NF [n=4] and DCM [n=3]) were merged and Spearman correlation coefficients of ChIP-seq sequence bam data of H3K27ac, H3K4me3, H3K4me3, H3K27me3 and inputs were calculated. The number in the square denotes Spearman rho for each comparison. The dark blue dash line indicates the comparison of H3K4me3 between NF and DCM. NF: non-failing, DCM: dilated cardiomyopathy. GEO: data from Gene Expression Omnibus, TSS: transcription start site.

In addition, we obtained two NF male histone modification ChIP-seq data from the GEO database and compared the average peak profile with our data. The average peak profiles of these NF samples were consistent with our data (Figure 1 B, NF4 and NF5). To increase the power of data analysis, we performed a meta-analysis by combining our NF1 and NF2 data with the data generated from the GEO samples (NF 4 & 5) for all downstream analysis. We detected 46,166 and 10,950 differential H3K27ac and H3K4me3 enriched regions, respectively, between two groups. Additionally, 60% of the H3K27ac-enriched regions in NF overlapped with 73.6% of those in DCM, whereas 91% of the H3K4me3-enriched regions in NF were commonly found to overlap with 84.5% of the H3K4me3 peaks in DCM. (Figure 1C). Differential enriched regions for H3K4me1 and H3K27me3 were also identified (Figures 1C). Consistent with the average signal data, in which the average H3K4me3 signal was lower in DCM compared to that in NF (Figure 1B), the correlation analysis of aligned sequences revealed that the sequence profiles for H3K4me3 had a lower correlation between two groups compared to other histone modifications (Spearman Rho, H3K4me3: 0.81, H3K27ac: 0.96, H3K4me1: 0.98, H3K27me3: 0.97) (Figure 1D).

3.2. The Action of SMYD1 contributes to the global reduction of H3K4me3 signals

We hypothesize that the lower H3K4me3 signaling globally present in the DCM hearts is due to the alteration of activity or expression of the histone methyltransferases (HMT) and demethylases (HDMs) in DCM hearts. Therefore, we extracted the RNA level of all the HMTs and HDMs from the RNA-seq data and compared these between NF and DCM groups. Among all the HMTs and HDMs, we found that the RNA level of SMYD1, a cardiac/muscle-specific HMT for H3K4, was significantly downregulated at 1.6 fold in DCM compared to NF hearts. (Figure 2A). The lower SMYD1 RNA level was confirmed by qRT-PCR (Figure 2A). However, the SMYD1 protein level showed no significant difference between NF and DCM groups (Figure 2B), suggesting other mechanisms may be responsible for the decease of H3K4me3 enrichment. As SMYD1 methylates H3K4 at promoter regions and we observed the reduction of H3K4me3 signal at SMDY1 downstream genomic targets in DCM samples than that in the normal NF ones [24, 25] (Figure 2C), such as PPARGC1A (Peroxisome proliferator-activated receptor gamma coactivator 1α) and CASQ2 ( Calsequestrin 2) (Figure 2D & E), we hypothesize that the decrease of the H3K4me3 at promoter regions may occur due to a failed recruitment of SMYD1 to the promoter region of its downstream genes. Therefore, we performed a quantitative chromatin immunoprecipitation assay by_real-time PCR (ChIP-qPCR) for SMYD1 and measured the SMYD1 enrichment at promoter regions, upstream and downstream genomic regions of PPARGC1A and CASQ2. We detected a higher enrichment of SMYD1 at promoter regions of PPARGC1A and CASQ2 in NF comparing to that in DCM. In contrast, the SMYD1 enhancement remains low at detected intragenic and upstream zones of both gene loci ( Figure 2 F & G), which is consistent with the observation from the ChIP-seq data (Figure 2C–D & E).

Figure 2. SMYD1 contributes to the decrease of H3K4me3 enrichment at promoters.

Figure 2.

Open in a new tab

(A) Bar graphs showing the SMYD1 RNA levels in NF (n=162) and DCM (n=162) determined using RNA-seq (right) and quantification RT-PCR (left, n=6). Data are shown as mean± standard errors. Asterisks, p<0.0001 (Student T-test). The dot represents the relative RNA level normalized to 18s for each biological sample. The mean of biological replicates for each group with standard errors after normalized to 18s was shown. (B) Immunoblot analysis of human NF and DCM left ventricle lysates for SMYD1 and Ponceaus S staining (for actin) of the blot used for normalization. N=6. NF, non-failing heart; DCM, dilated cardiomyopathy, Pon S, Ponceau S. (C) Average profiles of H3K4me3 at SMYD1 know downstream genomic targets from five NF and three DCM samples (D-E) H3K4me3 peak profile at promoter regions of CASQ2 (D) and PPARGC1A (E). Angled arrows indicate the transcription start sites. Gray bars indicate the detect genomic regions in ChIP-qPCR. (F-G) ChIP-qPCR analysis of the SMYD1 enrichment at promoter, intragenic and upstream regions of CASQ2 (F) and PPARGC1A (G). Bar graphs show the percentages of enrichment with respect to 10% of input (non-immunoprecipitation DNA). Data are mean ± standard errors of biological replicates (n=3). Asterisks, p<0.05 (Student T-test).

3.3. ChIP-seq analysis discovered potential novel players in HF

We next carried out gene ontology (GO) analysis using Genomic Regions enrichment of Annotation (GREAT) to identify the biological events associated with the H3K27ac and H3K4me3-enriched regions across NF and DCM genomes [16]. We found that the common H3K27ac-enriched regions between NF and DCM were involved in regulating homeostasis and physiology of the hearts, such as mitochondrial transport, titin binding and actin filament bundle (Figure S1). While the NF-specific H3K27ac-enriched regions were primarily associated with the heart muscle-specific cellular components (myofibril, sarcomere and contractile fiber), DCM-specific H3K27ac enriched regions were associated with the GO terms similar to the ones found in the overlapped peaks (Figure S1). GREAT analysis for H3K4me3 enriched regions showed that the common H3K4me3 enriched regions in both NF and DCM were linked to the biological events that involved normal activities and components for hearts, e.g., ATP synthase, copper ion transport and succinate dehydrogenase activity. Interestingly, the DCM-specific H3K4me3 peaks were linked to the binding of the transcription factors for SMAD and E-box as well as to skeletal and cardiac muscle hypertrophy (Figure S2). Consistent with the GREAT analysis, de novo motif analysis for DCM-specific H3K4me3 enriched regions also identified the binding motif for E2F7, an E-box binding protein within these regions (Figure S3). While several interferon-regulatory factors (IRF) were identified as the top known DNA binding motifs within the H3K4me3 peaks for both groups, the top DNA motifs discovered within H3K27ac peaks between NF and DCM are distinctive (Figure 3A–D). The top DNA motifs found within H3K27ac-marked enhancers for NF are myocyte enhancer factor 2 (MEF2) transcription factors, such as MEF2A, B, and C, which are known for regulating cardiac gene expression during development and adulthood [26]. The top DNA-binding motifs found in enhancers from DCM are in the ETS-domain transcription factor family such as FLI1, ERG and ETVs (Figure 3B). ETS transcription factors have been shown to play an essential role during vasculature development, extracellular matrix organization, as well as induction of cardiac remodeling [27, 28]. Therefore, the enrichment of the binding sites for these ETS factors within enhancers of failing hearts suggested their roles in cardiac failure. We also discovered the DNA binding motifs for the transcription factors whose gene expression levels were up-regulated ≥1.5 fold in DCM compared to NF based on RNA-seq analysis (Figure 3E–G). These genes included the known transcription factors shown to have a role in the pathological process for HF, such as the STAT protein family and NFkB, and also contained TFs where the role in the failing heart remains unclear. These include SRY-related high-mobility-group box (SOX) transcription factor 4, SOX4, and the BTB-Zinc finger family protein, B-cell lymphoma 6 (BCL6/BCL6B) (Figure 3E–G). Because IRF and ETS families contain two large groups of transcription factors, we selected the genes, IRF1 and STAT4, which are most significantly upregulated in DCM in comparison to NF among each family group based on RNA-seq data for further validation. Quantification RT-PCR assay revealed that_the RNA levels of IRF1 and STAT4 were upregulated in DCM compared to NF hearts. However, it is not statistically significant due to the variation among samples (Figure 3H). The RNA level of SOX4 and BCL6B was significantly increased in DCM samples in comparison to NF ones (Figure 3H). The protein level for BCL6B and SOX4 was confirmed by Western blot. In the NF samples, the SOX4 protein was present. However, the SOX4 protein level was significantly increased in DCM samples (Figure 3I). BCL6B protein was barely detected in NF samples but was increased in DCM samples ( Figure 3J ).

Figure 3. Known motif analysis revealed the top DNA binding motifs within the H3K27ac- and H3K4me3- enriched regions identified from ChIP-seq.

Figure 3.

Open in a new tab

(A-B) Top motifs identified from H3K27ac-enriched regions in NF (A) and DCM (B) samples. (C-D) Top motifs identified from H3K4me3-enriched regions fin NF (C) and DCM (D) samples. The top 5 DNA motifs were shown. The percent sites indicate the percent of the enriched regions overlapping with the motif. (E) Selected DNA binding motifs identified using HOMER in the H3K27ac-enriched region of the DCM samples for the transcription factors whose gene expression levels were up-regulated ≥1.5 fold in DCM in comparison to NF. Percent sites indicate the percent of the H3K27ac-enriched regions overlapping the motif. (F) Selected DNA binding motifs identified using HOMER in the H3K4me3-enriched region of the DCM samples for the transcription factors whose gene expression levels were up-regulated ≥1.5 fold in DCM in comparison to NF. (G) Bar graphs showing the RNA levels in NF (n=162) and DCM (n=162) determined using RNA-seq for IRF1, STAT4, BCL6B, and SOX4. Data are shown as mean± SE. Asterisks, p<0.0001 (Student t-test). TPM: Transcripts Per Million reads. (H) Quantification RT-PCR analysis for IRF1, STAT4, BCL6B, and SOX4. The dot represents the relative RNA level normalized to 18s for each biological sample (n=6). The mean of biological replicates for each group with standard errors after normalized to 18s was shown. The asterisk indicates p < 0.05 (Mann Whitney test). (I-J) Immunoblot analysis of human NF and DCM left ventricle lysates for SOX4 (I) BCL6B (J) and Ponceaus S staining (for actin) of the blot used for normalization. SOX4 (3.889 ± 0.7532; p=0.0022) and BCL6B (6.639 ± 2.215; p=0.0376) protein level in DCM is significant increased in comparison with NF group. N=6. NF, non-failing heart, DCM, dilated cardiomyopathy, Pon S, Ponceau S. Asterisk indicates p < 0.05 (Mann Whitney test).

3.3. Differential super-enhancers were found in DCM hearts

Super-enhancers (SE) are large clusters of enhancers that are typically bound by transcriptional regulators such as TFs and co-activators in high density. It was shown that the SE-driven transcriptional program is often involved in cell fate determination and in the development of diseases [7]. To understand if the SE-associated transcriptional program is involved in HF, we performed SE analysis using the signal obtained from H3K27ac ChIP-seq for both NF and DCM hearts. SE analysis revealed that DCM possessed fewer SEs than NF (DCM: 427, NF: 713) (Figure 4A). The essential cardiac-related genes were found linked to SEs in the heart. These genes include WASF2 (WAS protein family member 2), an actin-regulating/binding factor that is essential for cardiac development [29], NCOR2 (nuclear receptor corepressor 2), a transcriptional coregulator essential for normal cardiac developme nt [30], ENG (endoglin), a glycoprotein primarily associated with vascular endothelium and essential for normal heart development [31], ZMIZ1(zinc finger MIA-type containing1), a member of the protein inhibitor of the activated STAT (PIAS) family, required for proper vascular development, LMNA (Iamin A/C), TNNT3 (troponin T3), and MEF2A (myocyte enhancer factor 2A). We also uniquely identified one and two SE-associated genes whose gene expression level changed at least 1.5-fold in NF and DCM, respectively - SERPINE1 (serpin family E member 1) in NF, and SLC27A3 (solute carrier family 27 member 3) and CLEC3B (C-type lectin domain family 3 member B) in DCM. (Figure 4A & B). The RNA and protein levels of these three unique SE-associated factors were confirmed by qPCR and immunoblotting (Figure 4C &D). Although qPCR assay did not detect a significant increase of CLEC3B RNA level in DCM, its protein level is considerably enriched in DCM hearts (Figure 4D).

Figure 4. Super-enhancer analysis and comparison between NF and DCM.

Figure 4.

Open in a new tab

(A) Ranking of the enhancers detected in NF and DCM according to the normalized H3K27ac ChIP-seq signals. The number of super-enhancers (SE) detected in NF and DCM is shown. A few genes associated with SEs are indicated. (B) Bar graphs showing the RNA levels in NF (n=162) and DCM (n=162) determined using RNA-seq for one unique SE-associated gene identified from NF (SERPINE1) and two from DCM (SLC27A3 & CLEC3B). Data are shown as mean± SE. Asterisks, p<0.0001 (Student T-test). TPM: Transcripts Per Million reads. (C) Quantification RT-PCR analysis for SERPINE1, CLEC3B, and SLC27A3. The dot represents the relative RNA level normalized to 18s for each biological sample (n=6). The mean of biological replicates for each group with standard errors after normalized to 18s was shown. The asterisk indicates p < 0.05 (Mann Whitney test). (D) Immunoblot analysis of human NF and DCM left ventricle lysates for SERPINE1, CLEC3B, and SLC27A3 and Ponceaus S staining (for actin) of the blot used for normalization. SERPINE1 (0.4169 ± 0.0275; p=0.0022) protein is significantly decreased in DCM. CLEC3B (9.441 ± 3.074; p=0.0022) and SLC27A3 (1.533 ± 0.2633; p=0.0411) protein level in DCM is significantly increased in comparison with NF group. N=6. NF, non-failing heart, DCM, dilated cardiomyopathy, Pon S, Ponceau S. Asterisk indicates p < 0.05 (Mann Whitney test). (E) Venn diagram showing the number of differential genes associated with super-enhancers between NF and DCM. (F) Top transcription factor binding motifs identified in NF-specific SEs (SE found only in NF), in NF and DCM shared super-enhancers (SE found in both NF and DCM), and in DCM-specific super-enhancers (SE found only DCM). NF:non-failing, DCM: dilated cardiomyopathy.

A SE consists of a group of enhancers that often span across more than 10 kb of the genome, which may include multiple genes associated with that SE. The common enhancer annotation method will only annotate one SE to the closest gene , and this may overlook other genes nearby the same SE. Therefore, we also performed enhancer annotation using all H3K27ac peaks within the SE and assigned each H3K27ac enriched region to the closest gene. By doing so, we found that there are 2,181 associated with SE in the heart. There are 953 genes commonly associated with SE found in both NF and DCM tissues.

In addition, 738 genes were associated with NF-specific SEs, whereas 490 genes were specifically associated with SEs found only in DCM (Figure 4E). To explore what transcription factors may bind to these SEs, we performed motif analysis using the H3K27ac-enriched regions within SEs. As expected, the NF-specific SEs were enriched for DNA binding motifs corresponding to cardiac-specific master TFs (MEF2 and GATA) (Figure 4F). However, the DCM-specific SEs contains DNA binding motifs for zinc finger proteins (EGR1 and SCL/TAL1) that were shown to be involved in the process of HF (Figure 4F) [32]. Ingenuity pathway analysis (IPA) for 2,181 SE-associated genes revealed that the top GO-terms associated with the common and NF-specific SE genes were related to normal cardiovascular system development and function, e.g., the morphology of the heart and cardiovascular systems, and regulation of heart contraction (Figure 5A– C and Table S4–5). These genes included cardiac-specific master TFs (GATA4 and MEF2A), and cytoskeleton and filament associated genes (LMNA, ACTN4, and MYHs). The top GO-terms found in DCM-specific SE genes were related to connective tissue development and function, e.g., migration of fibroblast and fibrogenesis. The cardiac-related GO terms associated with the DCM-specific SE genes were associated with its cardiac remodeling and fibrosis (endothelial cells movement and vascular tissue development) (Figure 5C and Table S6) [33].

Figure 5. Gene ontology analysis of the super-enhancer-associated genes in NF and DCM using Ingenuity pathway analysis.

Figure 5.

Open in a new tab

Bar charts showing the top GO terms identified in the physiological system development and function for genes associated super-enhancers present only in NF (A), share between NF and DCM (B) and in DCM (C). Only the GO terms with p-value ≤ 0.01 were shown. The selected cardiac genes from denoted functional classes (1–4) are listed based on gene categories. The specific functional annotations and GO terms for the denoted functional classes are present in the bottom.

3.4. Differential long-range interactions of enhancers and promoters identified in DCM

The mammalian genome normally folds in three-dimensional (3D) architecture allowing cis-regulatory elements, such as enhancers and promoters, for accurate gene expression. Studies have demonstrated that aberrant chromatin interactions can lead to various diseases, including cardiovascular diseases [8, 9, 34, 35]. To understand the changes of chromatin interactions in HF, we performed proximity ligation-assisted ChIP-seq (PLAC-seq) for promoters and enhancers in the human NF and DCM hearts. Genomewide-PLAC-seq analysis using model-based analysis of PLAC-seq interactions (MAPS) pipeline revealed that 36.3% of H3K27ac-MAPS overlapped with 37.8% of H3K4me3-MAPS in NF, whereas 34.8% and 24.1% overlapped for H3K27ac- and H3K4me3- MAPS in DCM, respectively (Figure 6A). Comparison between NF and DCM showed that 40% of H3K27ac-MAPS overlapped with 67% of DCM’s, and 62% of H3K4me3-interactions in NF overlapped with 69% of that in DCM (Figure 6B). Moreover, DCMs have 60% fewer H3K27ac-interactions and 72% less for H3K4me3 than NF samples, suggesting fewer enhancer-promoter interactions are present in human HF myocardium. Furthermore, the interaction strength, judged by the false discovery rate (which is associated to the ratio for the observed number of read counts to expected read counts for each interaction) was significantly lower in DCM compared to the NF in both H3K27ac and H3K4me3 interactions, suggesting potential interruption of chromatin interaction in DCM hearts (Figure 6C). The genes associated with H3K27ac and H3K4me3 interactions displayed higher expression levels than genes without any interaction. On average, genes involved in the H3K4me3-interactions showed a higher expression level in DCM than that in NF, but no significant difference between NF and DCM for genes linked to H3K27ac-interactions (Figure 6D).

Figure 6. Long-range chromatin interaction for promoters and enhancers using PLAC-seq revealed the alterations of epigenetic features in DCM.

Figure 6.

Open in a new tab

(A) Venn diagram showing the overlap between H3K27ac- and H3K4me3-long-range chromatin interactions identified by Model-based Analysis of PLAC-seq interactions (MAPS) pipeline in NF (right) and in DCM (left). (B) Venn diagram displaying the percentage and the number of NF- or DCM-specific and shared H3K27ac (left) and H3K4me3 (right) interactions. (C) Box plot of the comparison of the interaction strength (FDR, log2(-log transformation applied) of H3K27ac (left) and H3K4me3 interactions (right) between NF and DCM. The inner lines are median, and “+ “are means. All outliers are not presented. The lower FDR, the higher interaction strength of the interactions. Both interaction strength of the H3K27ac and H3K4me3 interactions in NF are higher than those in DCM. Asterisks, p<0.0001 (Mann-Whitney test). (D) Box plot of expression of different groups of genes. Genes with either H3K27ac or H3K4me3 interactions display higher expression than other genes in both NF and DCM. Asterisks, p<0.0001. (E) Bar graph showing the percentage of H3K27ac-MAPS or no-MAPS with SE versus without SE in both NF and DCM groups. (F) Bar graph showing the percentage of H3K4me3-MAPS or no-MAPS with SE versus without SE in both NF and DCM groups. The p-value and Odds ratio are shown (Fisher’s exact test). SE: Super-enhancers. FDR: false discovery rate.

Because 3D chromatin interaction modulates enhancer activity, we sought to determine if the enrichment of super-enhancers is associated with the long-range chromatin interactions in NF and DCM hearts. Bioinformatic analysis revealed that the ratio of H3K27ac-MAPS regions having SEs to that without SEs (14.33% in NF and 18.23 % in DCM) was significantly higher than the ratio of the no H3K27ac-MAPS regions having SEs to that without SEs (6.72% in NF and 6.54% in DCM) (p< 2.2e-16, Odds ratio =2.1321 in NF and 2.7886 in DCM) (Figure 6E). Similar results were observed in the comparison between H3K4em3_MAPS and no H3K4me3_MAPS for both NF and DCM groups (Figure 6F). Our data demonstrated that SEs are present in the genomic regions associated with long-range chromatin interactions. To determine if the three unique SE associated genes,SERPINE1, CLEC3B and SLC27A3 identified specifically in NF or DCM hearts display differential MAPS, we analyze the MAPS associated with the SE linked to these three genes. Several differential MAPS linked to the SEs associated with SPERINE1 and CLEC3B genes were identified but were not in the SLC27AC gene locus (Figure S4, gray vertical bar).

It has been well established that the fetal gene program is reactivated in HF [36]. One of the characteristic features in HF fetal gene reprogramming is the upregulation of natriuretic peptides, NPPA, and NPPB [37, 38]. We observed that differential interactions and level of the histone marks were also found in NPPA-NPPB gene loci (Figure 7A). The active enhancer and promoter signatures were both enriched in the DCM samples at NPPA-NPPB loci, suggesting the active transcription at these gene loci (Figure 7B, blue rectangle mark). Next we performed ChIP-qPCR to validate the enrichment H3K27ac and H3K4me3 ChIP signals at the promoter and enhancer regions of NPPB in the LVs from three independent individuals. The enrichment of both H3K27ac and H3K4me3 in DCM was significantly higher in DCM than that in NF samples. (Figure S5A). NPPA expression in DCM was increased more than 6.3 fold than that in NF, whereas NPPB expression in DCM was upregulated 2.0 fold than NF (Figure 7B left ). We confirmed their RNA levels using qRT-PCR in additional biological samples (Figure 7B right). Meanwhile, another feature of the fetal gene program is the switch of the expression from the adult isoform, α-myosin heavy chain (α-MHC, encoded by MYH6 gene) to the fetal isoform (β-MHC, encoded by MYH7 gene). In humans, β-MHC is predominantly expressed in the ventricle from the fetal stage to adulthood [36]. We observed that the H3K4me3 promoter mark was significantly lower in DCM at MYH6 promoter compared to that in NF Figure 7C, blue rectangle mark), consistent with the gene expression changes ([DCM/NF] ≥ 6.8 downregulated in RNA-seq). The RNA level of MYH7 was slightly increased in DCM (DCM/NF] ≥ 1.3 upregulated in RNA-seq). The RNA-seq data was confirmed by qRT-PCR (Figure 7D). A SE was present in NF but not in the DCM across MYH6-MYH7 gene loci and their neighborhood genes (Figure 7A, green bar). Although this SE was assigned to the SLC22A17 gene but not MYH6-MYH7, differential MAPS were identified. More H3K27ac-MAPS were identified in DCM than in NF, whereas there were more H3K4me3-MAPS in NF than in DCM (Figure 7C, gray vertical bar ). ChIP-qPCR assay validation showed that while there was no significant difference of H3K27ac enrichment between NF and DCM at MYH6 locus, the enrichment of H3K4me3 was significantly higher in NF than that in DCM, consistent with the ChIP-seq results (Figure S5B).

Figure 7. Epigenetic alteration at promoters and enhancers were correlated with gene expression at failure-associated gene loci (A, C, E).

Figure 7.

Open in a new tab

Comparison of the histone modifications and Model-based Analysis of PLAC-seq interactions (MAPS) profiles at NPPA-NPPB (A), MYH6-MYH7 (C), CTGF (E), POSTN (G) and their neighboring gene loci in NF and DCM. Gene body of interest is yellow highlighted with angled arrows at the transcription start sites. Genes of interest are shown in bold font. The horizontal light green bar denotes super-enhancers. Gray box indicates the differential H3K27ac- or H3K4me3-interactions between NF and DCM. The blue box indicates the alteration of promoter or enhancer histone modification signatures. For clarity, only the interactions associated with MYH6-MYH7, NPPA-NPPB, CTGF and POSTN gene loci were shown. (B, D, F) Showing gene expression level of MYH6 and MYH 7 (B) NPPA and NPPB (D) CTGF (F) and POSTN (H) using RNA-seq in NF (n=162) and DCM (n=162) (left) and qRT-PCR (right) (n=6). For qRT-PCR, the dot represents the relative RNA level normalized to 18s for each biological sample. The mean of biological replicates for each group with standard errors after normalized to 18s was shown. The asterisk indicates p < 0.05 (Mann Whitney test) for qRT-PCR and: p<0.0001 for RNA-seq. Data are shown as mean± standard error. TPM: Transcripts Per Million reads.

Cardiac fibrosis is one of the critical pathological processes in cardiac remodeling leading to HF[33]. Two factors have been shown to play essential roles in cardiac remodeling, CTGF (connective tissue growth factor) and POSTN (Periostin). As expected, the expression of these two genes were significantly upregulated in DCM hearts ([DCM/NF] ≥ 1.8 for CTGF; [DCM/NF] ≥ 2.1 for POSTN) (Figure 7 F &H). The signals of H3K27ac and H3K4me3 at CTGF and POSTN gene loci were much more enriched in DCM than that in NF (Figure 7E & G, Figure S5C & D), which is consistent with the changes of gene expression. Interestingly, the long-range chromatin interactions were only discovered in POSTN locus in DCM but not in NF samples (Figure 7G), suggesting these long-range chromatin interactions were DCM-specific.

4. Discussion

We have established the genomic landscape of major histone modifications, and alterations of the long-range chromatin interactions for promoters and enhancers in human DCM and NF left ventricles, uncovering a distinct epigenetic difference in the cardiac genome and identified potential novel factors in DCM.

We observed that the global active promoter signature, H3K4me3, is significantly reduced in the DCM cardiac genome. This reduction is correlated to the downregulation of SMYD1 gene expression, an H3K4-specific lysine methyltransferase, in DCM. SMYD1 is essential for heart development[39], and it regulates cardiac metabolism by modulating transcription of mitochondrial biogenesis genes such as PPARGC1 in adult hearts [24]. Although we did not find alterations of SMYD1 protein levels between NF and DCM significantly, we demonstrated that the enrichment of SMYD1 to its known downstream genomic targets, e.g., the promoter region of PPARGC1 and CASQ2, was affected in DCM hearts (Figure 2), which was associated with the reduction of H3K4me3 signal at these genomic locations. Our data suggested that altered recruitment of SMYD1 to the promoter regions of its downstream targets may result in reduction of the H3K4me3 signal in DCM hearts. It remains unclear if the global H3K4me3 reduction observed in the current study is due to the alteration of SMYD1 actions in failing hearts. SMYD1 may have an important role in cardiac failure beyond regulating cardiac metabolism. Further investigations into the role of SMYD1 in epigenetic regulation in HF are warranted.

Additionally, the RNA level of SMYD1 was previously shown to be increased in the diseased hearts in humans[40], and the protein level was increased in pressure-overload cardiac hypertrophy and failure in mice [41]. However, previous SMYD1 expression in the human heart was detected by a semiquantitative real-time PCR with a few samples (n=3–in DCM in comparison to NF samples, which was a more robust and sensitive approach. Furthermore, we also observed the downregulation of the SMYD1 RNA level in patients with hypertrophic cardiomyopathy (Figure S6). Nevertheless, we did not find the difference in the protein level of SMYD1 between NF and DCM hearts (Figure 2B), suggesting the post-translational regulation of SMYD1 protein is present in diseased heart. It was known that rodent HF models may not reflect human diseases [42]. The discrepancy of SMYD1 expression between human and mice needs to be further determined.

Super-enhancers play a critical role in cell fate determination and the development of various diseases [7]. Our findings of the cardiac-specific master TFs linked to the NF-and common SE genes are consistent with the previous observations that SEs are often present near genes that possess cell-type-specific functions [7]. The top GO terms associated with pathological cardiac processes were identified in the DCM-specific SE genes, supporting the notion that SEs are associated with disease progression [7].

We reported three novel unique SE-associated genes, SERPINE1 for NF, as well as CLEC3B and SLC27A3 for DCM. SERPINE1, a serine protease inhibitor, was previously reported as a significantly downregulated gene in patients with idiopathic and ischemic cardiomyopathy [43, 44], which is consistent with our finding that the SE was only found in NF and not in DCM hearts. Its expression was 3.8-fold downregulated in DCM samples. The function of SLC27A3, a fatty acid transport protein, in the failing myocardium is largely unknown. CLEC3B, a plasminogen-binding protein, was reported to be a biomarker for low risk of HF and mortality. However, the role of CLEC3B and SLC27A3 in DCM has not been reported.

We revealed that SEs were preferentially present in the genomic regions with long-range chromatin interactions (Figure 6E&F), which is consistent of the concept that SEs are the function as a hub for massive binding for transcription regulators as they serve as the mediators to connect the 3D structure of chromatin [45, 46]. However, we only observed a few significantly differential long-range chromatin associated with SEs and the number of them our biological replicates (NF=2; DCM=2). Although we identified differential MAPS-interactions between NF and DCM at examined HF-associated gene loci (Figure 7), we did not find a significantly difference in the number of MAPS between NF and DCM within these gene loci except for POSTN. However, such data analysis requires larger biological samples to find the effect size on long-range differential chromatin for further association study [47]. Furthermore, whether or not these differential MAPS affect gene expression needs further investigations. Recent advances in genomic engineering using CRISPR/Cas9 approaches have opened up a new potential to study the role of chromatin interaction in regulating gene expression during development and diseases [48]. It will be interesting to study further the contribution of the alteration of chromatin interaction identified in the current study to targets gene expression, e.g., disease-associated interaction found in POSTN locus (Figure 7).

We have identified two novel TFs, SOX4, and BCL6/6B, whose DNA binding motifs were enriched in promoter regions, and their gene expression levels were significantly upregulated in DCM hearts. BCL6B was shown to play important roles in the immune response, repression of cancer, and to involve in ROS-dependent cellular events [49–51]. It was reported that BCL6B attenuates liver fibrosis and inflammation after liver injury [52]. SOX4 is expressed in developing and adult hearts and necessary for the proper development of the heart. It is a master regulator of epithelial-mesenchymal transition[53, 54]. SOX4 may involve in the TNF-alpha-associated inflammation process [55]. Since inflammation and fibrosis occurred in cardiac pathological processes, both BCL6B and SOX4 may be intricately involved and warrants further investigation as potential treatment targets for HF.

Several animal studies have shown that aberrant chromatin architecture, e.g., enhancer-promoter chromatin interactions and topologically associated domains, causes abnormal heart development as well as cardiology pathology [8, 9, 56]. However, in human HF, the long-range chromatin interaction in regulating gene expression is still not well understood. We found distinct epigenetic alterations and chromatin interactions in the HF gene loci, and these features correlated to the gene expression changes in DCM (Figure 7). However, variations were present among genes, which is consistent with previous findings [57]. Hearts consist of various cell types, and each type possesses cell-type-specific epigenetic features. Therefore, our data suggest the involvement of other mechanisms regulating these changes among genes in a cell type-specific manner.

We excluded one of the non-failing heart samples from the downstream epigenetic analysis due to the discovery of the reduction signal of H3K4me3, and LVEF was lower than usual. It has been postulated that epigenetic features, particularly histone modification signatures, can occur before the changes of gene expression [58]. In human ES cells, the H3K4me3 signal may” pre-condition” the chromatin structure prior to activating gene expression during development. It is unclear if similar processes occur during diseases. It was reported that 1 in 4 people might have developing HF without symptoms. Learning the epigenetic changes in the heart from people who may have developing cardiomyopathy or HF without symptom may be instrumental for exploring epigenetic targets for efficient upstream therapeutic strategies for HF in the future.

Additionally, several potential factors associated with epigenetic alterations were identified in the current study. Our study has provided valuable information on long-range chromatin interactions at promoters and enhancers in regulating gene expression and molecular features in normal and pathological human DCM. These data can also serve as a data source for any future study into epigenetic regulation on gene expression in HF. It will be necessary to further investigate the biological functions of these changes in HF. Additionally, identifying the epigenetic molecules causing these changes will open up the possibility of developing novel therapies for HF. The epigenetic signatures discovered from this study can serve as a database for developing diagnostic and targets for HF.

Supplementary Material

1

Highlights.

  • We establish the genomic cistrome and chromatin interactome in NF and DCM.

  • Global H3K4me3 enrichment is reduced in DCM heart.

  • Super-enhancers are enriched in genomic regions associated with long-range chromatin interaction

Acknowledgments

We thank Drs. Lin Li and Hsin-Yi Weng for statistical advice and Lifebanc of Northeastern Ohio for assistance in obtaining unmatched organ donor hearts for research.

Sources of Funding

This research work is supported by the Collins Family Fund, the Wortzman Family Fund and the Cleveland Clinic Research Program Committees Award. Dr. Tang is supported by grants from the National Institutes of Health (NIH) and the Office of Dietary Supplements (R01DK106000, R01HL126827). Dr. Hu is partially supported by the NIH grant U54DK107977. The content is solely the responsibility of the authors and does not necessarily represent the official views of NIH.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Data Availability

All raw ChIP-seq and PLAC-seq data can be found in the GEO database (GSE 135956)

Disclosures

All authors stated that they do not have any relevant relationships to disclose.

References:

  • [1].Savarese G, Lund LH, Global Public Health Burden of Heart Failure, Card Fail Rev 3(1) (2017) 7–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [2].Heusch G, Libby P, Gersh B, Yellon D, Bohm M, Lopaschuk G, et al. , Cardiovascular remodelling in coronary artery disease and heart failure, Lancet 383(9932) (2014) 1933–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [3].Alexanian M, Padmanabhan A, McKinsey TA, Haldar SM, Epigenetic therapies in heart failure, J Mol Cell Cardiol 130 (2019) 197–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [4].Papait R, Greco C, Kunderfranco P, Latronico MV, Condorelli G, Epigenetics: a new mechanism of regulation of heart failure?, Basic Res Cardiol 108(4) (2013) 361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [5].Liu CF, Tang WHW, Epigenetics in Cardiac Hypertrophy and Heart Failure, JACC Basic Transl Sci 4(8) (2019) 976–993. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Yu M, Ren B, The Three-Dimensional Organization of Mammalian Genomes, Annu Rev Cell Dev Biol 33 (2017) 265–289. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [7].Hnisz D, Abraham BJ, Lee TI, Lau A, Saint-Andre V, Sigova AA, et al. , Super-enhancers in the control of cell identity and disease, Cell 155(4) (2013) 934–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [8].Lee DP, Tan WLW, Anene-Nzelu CG, Lee CJM, Li PY, Luu TDA, et al. , Robust CTCF-Based Chromatin Architecture Underpins Epigenetic Changes in the Heart Failure Stress-Gene Response, Circulation 139(16) (2019) 1937–1956. [DOI] [PubMed] [Google Scholar]
  • [9].Rosa-Garrido M, Chapski DJ, Schmitt AD, Kimball TH, Karbassi E, Monte E, et al. , High-Resolution Mapping of Chromatin Conformation in Cardiac Myocytes Reveals Structural Remodeling of the Epigenome in Heart Failure, Circulation 136(17) (2017) 1613–1625. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Liu CF, Angelozzi M, Haseeb A, Lefebvre V, SOX9 is dispensable for the initiation of epigenetic remodeling and the activation of marker genes at the onset of chondrogenesis, Development 145(14) (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [11].Afgan E, Baker D, Batut B, van den Beek M, Bouvier D, Cech M, et al. , The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2018 update, Nucleic Acids Res 46(W1) (2018) W537–W544. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [12].Langmead B, Salzberg SL, Fast gapped-read alignment with Bowtie 2, Nat Methods 9(4) (2012) 357–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [13].Shen L, Shao N, Liu X, Nestler E, ngs.plot: Quick mining and visualization of next-generation sequencing data by integrating genomic databases, BMC Genomics 15 (2014) 284. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [14].Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, et al. , Model-based analysis of ChIP-Seq (MACS), Genome Biol 9(9) (2008) R137. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [15].Heinz S, Benner C, Spann N, Bertolino E, Lin YC, Laslo P, et al. , Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities, Mol Cell 38(4) (2010) 576–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].McLean CY, Bristor D, Hiller M, Clarke SL, Schaar BT, Lowe CB, et al. , GREAT improves functional interpretation of cis-regulatory regions, Nat Biotechnol 28(5) (2010) 495–501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].Fang R, Yu M, Li G, Chee S, Liu T, Schmitt AD, et al. , Mapping of long-range chromatin interactions by proximity ligation-assisted ChIP-seq, Cell Res 26(12) (2016) 1345–1348. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [18].Juric I, Yu M, Abnousi A, Raviram R, Fang R, Zhao Y, et al. , MAPS: Model-based analysis of long-range chromatin interactions from PLAC-seq and HiChIP experiments, PLoS Comput Biol 15(4) (2019) e1006982. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [19].Zhou X, Maricque B, Xie M, Li D, Sundaram V, Martin EA, et al. , The Human Epigenome Browser at Washington University, Nat Methods 8(12) (2011) 989–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Liu Y, Morley M, Brandimarto J, Hannenhalli S, Hu Y, Ashley EA, et al. , RNASeq identifies novel myocardial gene expression signatures of heart failure, Genomics 105(2) (2015) 83–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. , STAR: ultrafast universal RNA-seq aligner, Bioinformatics 29(1) (2013) 15–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Love MI, Huber W, Anders S, Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2, Genome Biol 15(12) (2014) 550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Li W, Kennedy D, Shao Z, Wang X, Kamdar AK, Weber M, et al. , Paraoxonase 2 prevents the development of heart failure, Free Radic Biol Med 121 (2018) 117–126. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [24].Warren JS, Tracy CM, Miller MR, Makaju A, Szulik MW, Oka SI, et al. , Histone methyltransferase Smyd1 regulates mitochondrial energetics in the heart, Proc Natl Acad Sci U S A 115(33) (2018) E7871–E7880. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Stewart MD, Lopez S, Nagandla H, Soibam B, Benham A, Nguyen J, et al. , Mouse myofibers lacking the SMYD1 methyltransferase are susceptible to atrophy, internalization of nuclei and myofibrillar disarray, Dis Model Mech 9(3) (2016) 347–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [26].Potthoff MJ, Olson EN, MEF2: a central regulator of diverse developmental programs, Development 134(23) (2007) 4131–40. [DOI] [PubMed] [Google Scholar]
  • [27].Hao G, Han Z, Meng Z, Wei J, Gao D, Zhang H, et al. , Ets-1 upregulation mediates angiotensin II-related cardiac fibrosis, Int J Clin Exp Pathol 8(9) (2015) 10216–27. [PMC free article] [PubMed] [Google Scholar]
  • [28].Siwik DA, Colucci WS, Regulation of matrix metalloproteinases by cytokines and reactive oxygen/nitrogen species in the myocardium, Heart Fail Rev 9(1) (2004) 43–51. [DOI] [PubMed] [Google Scholar]
  • [29].Yamazaki D, Suetsugu S, Miki H, Kataoka Y, Nishikawa S, Fujiwara T, et al. , WAVE2 is required for directed cell migration and cardiovascular development, Nature 424(6947) (2003) 452–6. [DOI] [PubMed] [Google Scholar]
  • [30].Jepsen K, Solum D, Zhou T, McEvilly RJ, Kim HJ, Glass CK, et al. , SMRT-mediated repression of an H3K27 demethylase in progression from neural stem cell to neuron, Nature 450(7168) (2007) 415–9. [DOI] [PubMed] [Google Scholar]
  • [31].Bourdeau A, Dumont DJ, Letarte M, A murine model of hereditary hemorrhagic telangiectasia, J Clin Invest 104(10) (1999) 1343–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [32].Neyses L, Pelzer T, The biological cascade leading to cardiac hypertrophy, Eur Heart J 16 Suppl N (1995) 8–11. [DOI] [PubMed] [Google Scholar]
  • [33].Travers JG, Kamal FA, Robbins J, Yutzey KE, Blaxall BC, Cardiac Fibrosis: The Fibroblast Awakens, Circ Res 118(6) (2016) 1021–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [34].Hall DD, Spitler KM, Grueter CE, Disruption of cardiac Med1 inhibits RNA polymerase II promoter occupancy and promotes chromatin remodeling, Am J Physiol Heart Circ Physiol 316(2) (2019) H314–H325. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [35].Solovei I, Wang AS, Thanisch K, Schmidt CS, Krebs S, Zwerger M, et al. , LBR and lamin A/C sequentially tether peripheral heterochromatin and inversely regulate differentiation, Cell 152(3) (2013) 584–98. [DOI] [PubMed] [Google Scholar]
  • [36].Dirkx E, da Costa Martins PA, De Windt LJ, Regulation of fetal gene expression in heart failure, Biochim Biophys Acta 1832(12) (2013) 2414–24. [DOI] [PubMed] [Google Scholar]
  • [37].Nakao K, Minobe W, Roden R, Bristow MR, Leinwand LA, Myosin heavy chain gene expression in human heart failure, J Clin Invest 100(9) (1997) 2362–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [38].Reiser PJ, Portman MA, Ning XH, Schomisch Moravec C, Human cardiac myosin heavy chain isoforms in fetal and failing adult atria and ventricles, Am J Physiol Heart Circ Physiol 280(4) (2001) H1814–20. [DOI] [PubMed] [Google Scholar]
  • [39].Phan D, Rasmussen TL, Nakagawa O, McAnally J, Gottlieb PD, Tucker PW, et al. , BOP, a regulator of right ventricular heart development, is a direct transcriptional target of MEF2C in the developing heart, Development 132(11) (2005) 2669–78. [DOI] [PubMed] [Google Scholar]
  • [40].Borlak J, Thum T, Hallmarks of ion channel gene expression in end-stage heart failure, FASEB J 17(12) (2003) 1592–608. [DOI] [PubMed] [Google Scholar]
  • [41].Franklin S, Kimball T, Rasmussen TL, Rosa-Garrido M, Chen H, Tran T, et al. , The chromatin-binding protein Smyd1 restricts adult mammalian heart growth, Am J Physiol Heart Circ Physiol 311(5) (2016) H1234–H1247. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [42].Riehle C, Bauersachs J, Small animal models of heart failure, Cardiovasc Res 115(13) (2019) 1838–1849. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [43].Huang H, Luo B, Wang B, Wu Q, Liang Y, He Y, Identification of Potential Gene Interactions in Heart Failure Caused by Idiopathic Dilated Cardiomyopathy, Med Sci Monit 24 (2018) 7697–7709. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [44].Yu A, Zhang J, Liu H, Liu B, Meng L, Identification of nondiabetic heart failure-associated genes by bioinformatics approaches in patients with dilated ischemic cardiomyopathy, Exp Ther Med 11(6) (2016) 2602–2608. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [45].Schmitt AD, Hu M, Jung I, Xu Z, Qiu Y, Tan CL, et al. , A Compendium of Chromatin Contact Maps Reveals Spatially Active Regions in the Human Genome, Cell Rep 17(8) (2016) 2042–2059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [46].Wang X, Cairns MJ, Yan J, Super-enhancers in transcriptional regulation and genome organization, Nucleic Acids Res 47(22) (2019) 11481–11496. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [47].Gorkin DU, Qiu Y, Hu M, Fletez-Brant K, Liu T, Schmitt AD, et al. , Common DNA sequence variation influences 3-dimensional conformation of the human genome, Genome Biol 20(1) (2019) 255. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [48].Guo X, Dean A, CRISPR/Cas9 offers a new tool for studying the role of chromatin architecture in disease pathogenesis, Genome Biol 19(1) (2018) 185. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [49].Morimoto H, Kanastu-Shinohara M, Ogonuki N, Kamimura S, Ogura A, Yabe-Nishimura C, et al. , ROS amplification drives mouse spermatogonial stem cell self-renewal, Life Sci Alliance 2(2) (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [50].Wang J, Dong L, Xu L, Chu ES, Chen Y, Shen J, et al. , B cell CLL/lymphoma 6 member B inhibits hepatocellular carcinoma metastases in vitro and in mice, Cancer Lett 355(2) (2014) 192–200. [DOI] [PubMed] [Google Scholar]
  • [51].Xu L, Li X, Chu ES, Zhao G, Go MY, Tao Q, et al. , Epigenetic inactivation of BCL6B, a novel functional tumour suppressor for gastric cancer, is associated with poor survival, Gut 61(7) (2012) 977–85. [DOI] [PubMed] [Google Scholar]
  • [52].Wang W, Huang P, Wu P, Kong R, Xu J, Zhang L, et al. , BCL6B expression in hepatocellular carcinoma and its efficacy in the inhibition of liver damage and fibrogenesis, Oncotarget 6(24) (2015) 20252–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [53].Schilham MW, Oosterwegel MA, Moerer P, Ya J, de Boer PA, van de Wetering M, et al. , Defects in cardiac outflow tract formation and pro-B-lymphocyte expansion in mice lacking Sox-4, Nature 380(6576) (1996) 711–4. [DOI] [PubMed] [Google Scholar]
  • [54].Tiwari N, Tiwari VK, Waldmeier L, Balwierz PJ, Arnold P, Pachkov M, et al. , Sox4 is a master regulator of epithelial-mesenchymal transition by controlling Ezh2 expression and epigenetic reprogramming, Cancer Cell 23(6) (2013) 768–83. [DOI] [PubMed] [Google Scholar]
  • [55].Bhattaram P, Muschler G, Wixler V, Lefebvre V, Inflammatory Cytokines Stabilize SOXC Transcription Factors to Mediate the Transformation of Fibroblast-Like Synoviocytes in Arthritic Disease, Arthritis Rheumatol 70(3) (2018) 371–382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [56].Gomez-Velazquez M, Badia-Careaga C, Lechuga-Vieco AV, Nieto-Arellano R, Tena JJ, Rollan I, et al. , CTCF counter-regulates cardiomyocyte development and maturation programs in the embryonic heart, PLoS Genet 13(8) (2017) e1006985. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [57].Papait R, Cattaneo P, Kunderfranco P, Greco C, Carullo P, Guffanti A, et al. , Genome-wide analysis of histone marks identifying an epigenetic signature of promoters and enhancers underlying cardiac hypertrophy, Proc Natl Acad Sci U S A 110(50) (2013) 20164–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [58].Perino M, Veenstra GJ, Chromatin Control of Developmental Dynamics and Plasticity, Dev Cell 38(6) (2016) 610–20. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1604774-supplement-1.pdf (2.4MB, pdf)

RESOURCES