Differential regulation of DNA methylation versus histone acetylation in cardiomyocytes during HHcy in vitro and in vivo: an epigenetic mechanism

Abstract

The mechanisms of homocysteine-mediated cardiac threats are poorly understood. Homocysteine, being the precursor to S-adenosyl methionine (a methyl donor) through methionine, is indirectly involved in methylation phenomena for DNA, RNA, and protein. We reported previously that cardiac-specific deletion of N-methyl-d-aspartate receptor-1 (NMDAR1) ameliorates homocysteine-posed cardiac threats, and in this study, we aim to explore the role of NMDAR1 in epigenetic mechanisms of heart failure, using cardiomyocytes during hyperhomocysteinemia (HHcy). High homocysteine levels activate NMDAR1, which consequently leads to abnormal DNA methylation vs. histone acetylation through modulation of DNA methyltransferase 1 (DNMT1), HDAC1, miRNAs, and MMP9 in cardiomyocytes. HL-1 cardiomyocytes cultured in Claycomb media were treated with 100 μM homocysteine in a dose-dependent manner. NMDAR1 antagonist (MK801) was added in the absence and presence of homocysteine at 10 μM in a dose-dependent manner. The expression of DNMT1, histone deacetylase 1 (HDAC1), NMDAR1, microRNA (miR)-133a, and miR-499 was assessed by real-time PCR as well as Western blotting. Methylation and acetylation levels were determined by checking 5′-methylcytosine DNA methylation and chromatin immunoprecipitation. Hyperhomocysteinemic mouse models (CBS+/−) were used to confirm the results in vivo. In HHcy, the expression of NMDAR1, DNMT1, and matrix metalloproteinase 9 increased with increase in H3K9 acetylation, while HDAC1, miR-133a, and miR-499 decreased in cardiomyocytes. Similar results were obtained in heart tissue of CBS+/− mouse. High homocysteine levels instigate cardiovascular remodeling through NMDAR1, miR-133a, miR-499, and DNMT1. A decrease in HDAC1 and an increase in H3K9 acetylation and DNA methylation are suggestive of chromatin remodeling in HHcy.

hyperhomocysteinemia (HHcy) has been associated with increased risk of venous thrombosis (10), myocardial infarction (29), peripheral vascular disease (7), and coronary artery disease (4). Clinical studies have demonstrated that patients with high levels of homocysteine (Hcy) in the blood are at higher risk of heart diseases. Vizzardi et al. (33) found raised Hcy levels in the plasma of 123 patients with dilated cardiomyopathy compared with 85 healthy control subjects. In a prospective study of plasma Hcy levels and risk of myocardial infarction, Stampfer et al. (29) found that 271 men who subsequently developed myocardial infarction had higher levels of Hcy in blood compared with healthy controls. In rats, HHcy for 10 wk induces cardiac remodeling and diastolic dysfunction (15). Though the detailed mechanisms are not well understood, HHcy has been shown to increase oxidative stress in the vasculature (34). Becker et al. (2) showed that HHcy inhibits nitric oxide-dependent regulation of cardiac O₂ consumption in vitro, through increased production of superoxide radical (O₂⁻) by activating NADPH oxidase. HHcy increases oxidative stress through the activation of N-methyl-d-aspartate receptor-1 (NMDAR1) (12), which leads to an influx of Ca²⁺ and generation of reactive oxygen species (ROS) and reactive nitrogen species (RNS). During HHcy, cardiac-specific deletion of NMDAR1 reduces oxidative stress, restores contractility in cardiomyocytes (25), and ameliorates matrix metalloproteinase-9 (MMP9)-mediated mitophagy (30). Antagonists to NMDAR1, like dizocilpine, have been reported to ameliorate sudden cardiac death by preventing MMP9 activation (20, 21).

In HHcy, there is activation of MMP9, which are involved in extracellular matrix remodeling (13, 26, 32). In the extracellular matrix of heart, the balance of collagen and elastin is maintained by MMP9/tissue inhibitor of matrix metalloproteinase ratio. MMPs remain latent in the healthy state; however, they get activated under stress conditions, and there is an imbalance of the elastin-collagen ratio, which leads to fibrosis and, eventually, heart failure (18, 31). Rosenberger et al. (28) have reported that during a Hcy-rich diet, an increase in MMP9 activity and decreased expression of connexins lead to reduced ventricular performance in mice.

The increase in oxidative stress in the form of ROS and RNS during HHcy inside the cells leads to the activation of some microRNAs (miRNAs) that regulate DNA methylation and acetylation (6). Chen et al. (6) have shown that increased oxidative stress leads to upregulation of MMP-2/MMP9 genes via upregulation of microRNA (miR)-29b. The role of miRNAs in Hcy-induced cardiac remodeling has been demonstrated by Mishra et al. (23). In diabetic cardiomyocytes, miR-133a regulates DNA methylation by regulating the levels of DNA methyltransferases (DNMTs) (5). Liang et al. (19) have reported that both activity and mRNA levels of DNMT1 increase in HHcy. In a study by Kinoshita et al. (17), it was demonstrated that plasma Hcy levels are associated with genome-wide DNA methylation. The authors used high-resolution DNA methylation array (450K) to show that DNA methylation levels were positively correlated with Hcy levels near CpG sites. Zhang et al. (37) have shown that high concentrations of Hcy (100–300 μM/l) induces hypermethylation, whereas low Hcy concentrations (10–30 μM/l) induce hypomethylation in the CpG islands of the DDAH2 gene (37). Evidence suggests that DNA methylation has emerged as an important phenomenon in human diseases such as atherosclerosis and cancer, which are represented by aberrant DNA methylation.

Although Hcy has been associated with aberrant DNA methylation, the role of epigenetic mechanisms in cardiomyocytes remains elusive. There are very few reports that demonstrate histone acetylation or other histone modifications during HHcy. Hence, the main aim of this study is to elucidate the epigenetic mechanisms comprising acetylation vs. methylation in cardiomyocytes under high-Hcy conditions.

MATERIALS AND METHODS

Animals and Experimental Design

Standard procedures were followed for animal care and health as per guidelines of the National Institute of Health and approved by the Institutional Animal Care and Use Committee, University of Louisville. The CBS +/− heterozygous animals were confirmed by genotyping and fed a methionine-rich diet to induce HHcy as done earlier in our laboratory (16). For control, we used wild-type (WT) C57BJ/L6 mice. For in vitro experiments, we used Dr. Claycomb's HL-1 cardiomyocyte cell line, which we have described earlier (5). The HL-1 cells, derived from AT-1 mouse atrium, retain the electrophysiological and biochemical properties of cardiomyocytes (8, 35). We treated HL-1 cells with Hcy and MK801 in a dose-dependent manner and selected six treatment groups: 1) control, 2) 5 μM Hcy, 3) 100 μM Hcy, 4) 100 μM Hcy + 10 μM MK801, 5) 10 μM MK801, and 6) 1 μM MK801.

In Vitro Studies in HL-1 Cardiomyocytes

mRNA, miRNA isolation, and quantitative real-time reverse-transcription PCR.

For total RNA extraction, TRIzol reagent (Life Technologies) was used according to the manufacturer's instructions. For cDNA synthesis, MultiScribe reverse transcriptase kit (Qiagen) was used with 2 μg total RNA using random primers. Quantitative real-time PCR was done with Sybr Green PCR Master mix (Qiagen) and Stratagene Mx3000P real-time PCR machine. The sequences of the primers used for real-time PCR of MMP9, DNMT1, histone deacetylase 1 (HDAC1), NMDAR1, and GAPDH are given in (Table 1). The fold expression was calculated as the cycle threshold difference between control and sample normalized with the housekeeping gene GAPDH. To quantify the expression of miR-133a and miR-499, real-time PCR assays from Qiagen were used, and snoRD-72 was used as an internal control.

Table 1.

The primer sequences of genes for RT-PCR

Genes	Forward Primers	Reverse Primers
MMP9	5′-TACAGGGCCCCTTCCTTACT-3′	5′-CCACATTTGACGTCCAGAGA-3′
GAPDH	5′-ACCACAGTCCATGCCATCAC-3′	5′-TCCACCACCCTGTTGCTGTA-3′
DNMT1	5′-AGGGAAAAGGGAAGGGCAAG-3′	5′-CCAGAAAACACATCCAGGGTCC-3′
HDAC1	5′-GACTGGACCCTCTGTCATTA-3′	5′-TTTCTCCTAAAGTGCAGCTC-3′
NMDAR1	5-GAATGATGGGCGAGCTACTCA-3	5-ACGCTCATTGTTGATGGTCAGT-3
MMP9 (ChIP) −518 Transcription start site	5′-AGATGAAGCAGGGAGAGGAAGC-3′	5′-CCTCCAGAGGTCAGCCAA-3′

MMP9, matrix metalloproteinase 9; DNMT1, DNA methyltransferase 1; HDAC1, histone deacetylase 1; NMDAR1, N-methyl-D aspartate receptor-1; ChIP, chromatin immunoprecipitation.

Western blot analysis.

For protein extraction cells were harvested in radioimmunoprecipitation assay buffer (Boston Bioproducts) with added protease inhibitors and phenylmethanesulfonyl fluoride (Calbiochem, La Jolla, CA), and protein estimation was done by the Bradford method. The protein lysate (30 μg) was denatured in SDS sample buffer (2% SDS, 10 mM dithiothreitol, 60 mM Tris·HCl pH 6.8, bromophenol blue 0.1%) and loaded on 12% PAGE. The separated proteins were transferred to polyvinylidene difluoride membrane by an electro-transfer apparatus (Bio-Rad) and blocked with 5% nonfat dry milk for 1 h at room temperature. The membrane was probed with primary and secondary antibodies and developed with ECL Western blotting detection system (GE Healthcare, Piscataway, NJ), and the image was recorded in the gel documentation system ChemiDoc XRS system (Bio-Rad, Richmond, CA). The membranes were stripped and probed with anti-GAPDH antibody (Millipore, Billerica, MA) as control. The data were analyzed with Image Lab densitometry software (Bio-Rad) and normalized to GAPDH.

Gelatin zymography.

The protein lysate (30 μg) was incubated with SDS sample buffer (without reducing agents) at 37°C for 30 min and loaded on 10% SDS-PAGE gel prepared with 2% gelatin. After protein separation, the gel was washed three times in 2.5% Triton X-100 for 20 min each to remove SDS and incubated in activation buffer (5 mmol/l Tris·HCl pH 7.4, 0.005% vol/vol Brij-35, and 1 mmol/l CaCl₂) for 24 h at 37°C. The gel was stained in Coomassie and destained to observe clear bands as a result of proteolytic activity of MMPs for gelatin on a stained background along with controls. The imaging was done with gel documentation system (Bio-Rad, Hercules, CA), and data were analyzed with image lab software (Bio-Rad, Hercules, CA).

2′,7′-Dichlorofluorescein diacetate treatment and flow cytometry.

Oxidative stress in the cells was determined by carboxy 2′,7′-dichlorofluorescein diacetate (DCFDA) (Molecular Probes) treatment. Cells were incubated at 37°C in DCFDA at 5 μM in PBS for 30 min then washed with PBS. The DCFDA-treated cells were observed with flow cytometry (Accuri CFlow Plus system) with gating applied. Prior to flow cytometry, live cells were counted by Bio-Rad TC10 automated cell counter using Trypan blue solution. We measured the oxidative stress in both normal and Hcy-treated cells by calculating the number of cells stained with DCFDA from the forward scattered graph. DCFDA-treated cells along with 4′,6-diamidino-2-phenylindole were also observed under laser scanning confocal microscope (Olympus FluoView1000).

Chromatin immunoprecipitation assay.

For chromatin immunoprecipitation (ChIP) assay we used the ab500 kit from Abcam and followed the manufacturer's protocol. In brief, the cells were collected, fixed with paraformaldehyde, and lysed with the buffers A and B provided by the manufacturer. The cells were centrifuged and resuspended in buffer C, and the resulting DNA was sheared with a sonicator (sonic dismembrator model 100, Fisher Scientific) to produce DNA fragments of size 200–1,000 bp. The sheared DNA was incubated with H3K9Ac antibody (provided in the kit) overnight and next day mixed with beads for immunoprecipitation. After immunoprecipitation the DNA was purified and checked with PCR.

In Vivo Studies in CBS+/− Mouse Heart

Immunohistochemistry of CBS and WT mouse heart tissue.

Heart tissue from CBS+/− and WT mice was collected and embedded in disposable plastic tissue molds (Polysciences, Warrington, PA) containing tissue-freezing media (Triangle Biomedical Sciences, Durham, NC). The tissue was stored at −80°C until further use. Cryosectioning was done with a Cryocut (Leica CM 1850) to get tissue sections of 5 μm, which were placed on polylysine-coated slides for staining. Tissue fixation was done with paraformaldehyde and permeabilized with Triton X-100 (0.3%). Primary antibodies were added for DNMT1, HDAC1, and NMDAR1 at 1:250 dilution (Abcam) and incubated overnight at 4°C. After washing the slides with Tris-buffered saline (TBS) and TBS + Tween 20 and applying fluorescently tagged secondary antibodies, we mounted the slides and visualized them with a laser scanning confocal microscope (Olympus FluoView1000).

DNMT activity assay.

Nuclei were isolated from the heart tissue with the EpiQuik nuclear extraction kit (Epigentek, Farmingdale, NY), and DNMT activity was measured with the EPiQuik DNMT activity/inhibitor assay ultra kit (Epigentek) according to the manufacturer's instructions. The result was calculated by the formula:

$$\text{DNMT}\hspace{0.17em}\text{activity}\hspace{0.17em}(\text{OD}\cdot {\text{h}}^{\text{−1}}\cdot {\text{mg}}^{-1})=\frac{\text{sample OD}-\text{blank}\hspace{0.17em}\text{OD}}{\text{nuclear}\hspace{0.17em}\text{protein}\hspace{0.17em}\text{amount}\hspace{0.17em}(\mu \text{g})\times \text{h}}\times \mathrm{1,000}$$

where OD is optical density.

HDAC activity assay.

The HDAC activity assay was performed similarly to the DNMT activity assay with the EpiQuick HDAC Activity/Inhibition Assay Kit according to the manufacturer's instructions. The activity of HDAC enzyme was calculated was using the following formula

$$\text{HDAC activity}(\text{OD}\cdot {\text{h}}^{\text{−1}}\cdot {\text{ml}}^{-1})=\frac{\text{OD}(\text{control}-\text{blank})-\text{OD}\hspace{0.17em}(\text{sample}-\text{blank})}{\text{reaction}\hspace{0.17em}\text{time}\hspace{0.17em}(\text{0}\text{.5–1}\hspace{0.17em}\text{h})}\times \text{sample}\hspace{0.17em}\text{dilution}$$

5′-Methylcytosine levels in genomic DNA.

For quantification of 5-methylcytosine levels in the gnomic DNA, we used the 5-mc DNA ELISA kit (Zymo Research, Irvine, CA) as per manufacturer's instructions. In brief, the denatured genomic DNA was coated in the wells provided in the kit along with controls. This was followed by the addition of primary (monoclonal antibody specific to 5-methylcytosine) and secondary antibody. The result was calculated through a standard curve generated by the positive and the negative controls in the kit and expressed as 5-methylcytosine percentage for DNA samples. The results were taken as the average of six independent experiments.

Statistical Analysis

Student's t-test was used to compare all experimental results. A value of P < 0.05 was considered significant, and n = 4 for all experiments. The results were expressed as means ± SE. We also applied the Mann-Whitney U-test, which is a nonparametric test for comparing two groups, and P < 0.05 was considered significant. SPSS Ver. 16 was used to calculate the Mann-Whitney U-test.

RESULTS

In Vitro Studies on HL-1

Expression of NMDAR1 after Hcy and MK801 treatment.

We treated the HL-1 cells with Hcy in a dose-dependent manner. We found that, in both Western and real-time PCR, the expression of NMDAR1 increased with Hcy treatment (Fig. 1, A and B). But on treatment with MK801, the expression decreased. In the MK801 per se group, the expression further decreased compared with the Hcy + MK801 group. Hcy is an agonist to the NMDAR1, while MK801 is antagonist. In our study Hcy increased the expression of NMDAR1, while MK801 decreased its expression.

Fig. 1.

Effect of homocysteine (Hcy) on N-methyl-d aspartate receptor 1 (NMDAR1) expression. We found increase in the expression of NMDAR1 after Hcy treatment in HL-1 cells. The expression was confirmed by Western blot and real-time PCR. The data were normalized to GAPDH. A: Western blot showing a significant (P < 0.05, n = 4) increase in the expression of NMDAR1 with Hcy treatment compared with control (CT) (30 μg protein-equal loading). There was a significant decrease in the NMDAR1 expression after MK801 (NMDAR1 antagonist) treatment. B: real-time expression of NMDAR1 (2 μg RNA used to make cDNA) after Hcy treatment showed upregulation, while MK801 treatment downregulated NMDAR1. *P < 0.05 compared with CT; #P < 0.05 compared with 100 μM Hcy.

Oxidative stress in HHcy.

In the HHcy condition, there is generation of ROS, which leads to oxidative stress inside the cells. To measure oxidative stress in HL-1, we treated the cells with DCFDA (Molecular Probes). Cells with more oxidative stress take a higher amount of DCFDA. We found that cells treated with Hcy take more DCFDA, which shows that HHcy increases oxidative stress (Fig. 2A). When cells were treated with MK801, oxidative stress decreased. H₂O₂ treatment was taken as a positive control. Oxidative stress was also measured with the help of flow cytometry (Fig. 2B). Cells treated with DCFDA were subjected to flow cytometry in Accuri CFLOW plus flow cytometer to determine the number of cells stained with DCFDA. Live cell count was done with the Bio-Rad TC10 automated cell counter. Cells treated with H₂O₂ were taken as a positive control. We performed gating in the forward scattered plot to select a fixed area for all the treatments. We found that with Hcy treatment the number of counts increased in the flow data compared with control. When treated with MK801 + Hcy, the number of cells decreased in the flow, which showed that MK801 is decreasing oxidative stress. We also determined the expression of the superoxide dismutase (SOD) gene, which is a marker for oxidative stress (Fig. 2C). The expression of SOD increased with Hcy treatment and decreased on treatment with MK801.

Fig. 2.

Oxidative stress in hyperhomocysteinemia (HHcy). We found an increase in oxidative stress in HHcy. The oxidative stress was assessed by treatment with 2′,7′-dichlorofluorescein diacetate (DCFDA), a dye that binds to the superoxide radical generated inside the cells. The oxidative stress was further confirmed by flow cytometry by quantifying the cells after DCFDA treatment. A: 1st column shows the cells stained with 4′,6-diamidino-2-phenylindole (DAPI), which stains the nucleus blue. The 2nd column shows cells stained with DCFDA in green. The cells with Hcy (100 μM) treatment take more green stain (arrows). The treatment with MK801 decreased oxidative stress. H₂O₂ was used as a positive control. B: the 1st graph shows the forward scattered plot with gating applied, and the 2nd plot shows the count number in terms of peak area. The live cell count (∼2 × 10⁶) was first found by Bio-Rad TC10 automated cell counter using Trypan blue dye. Then oxidative stress was found to be more in Hcy (100 μM)-treated cells as the number of counts was more. Treatment with MK801 reduced oxidative stress. C: Western blot of the superoxide dismutase (SOD) protein showed increase in the expression, with Hcy treatment, which indicates increased oxidative stress. *P < 0.05 compared with control; #P < 0.05 compared with 100 μM Hcy.

Expression of miR-133a and miR-499.

In a previous study (5), we reported that in diabetic cardiomyocytes miR-133a is involved in regulating DNA methyltransferases. Hence we checked the expression of miR-133a in HHcy and found downregulation through real-time PCR (Fig. 3A). We also found decreased expression of miR-499 in HHcy (Fig. 3A). miR-499 has been extensively studied in heart failure. When we used MK801, the expression of both miRNAs upregulated and became comparable with control. We used mimics and inhibitors to check whether miR-133a and -499 regulate DNMT1. With miR-133a mimics we found a decrease in the expression of DNMT1, and with inhibitors it was increased (Fig. 3B). However, we could not find much change in the expression of DNMT1 with the use of miR-499 mimics and inhibitors.

Fig. 3.

Expression of microRNAs (miR)-133a and -499. A: we found 4- to 5-fold decreases in the expression of miR-133a and -499 with real-time PCR. snoRD-72 was used as an internal control. In our earlier reports (3) we showed that miR-133a is involved in regulating the DNA methylation in diabetic cardiomyocytes. Here we showed the involvement of miR-133a in HHcy condition. B: use of mimics and inhibitors for miR-133a and miR-499 to check whether these 2 microRNAs (miRNAs) regulate the level of DNA methyltransferase 1 (DNMT1). The level of DNMT1 was regulated by miR-133a mimics and inhibitors. *P < 0.05 compared with control; #P < 0.05 compared with 100 μM Hcy.

DNA methylation and histone modification in HHcy.

We checked the protein expression as well as the transcript of DNMT1 with Western and real-time PCR, respectively. We found an increase in the expression of DNMT1 RNA and protein (Fig. 4, A and B). For histone modification we checked the expression of HDAC1 with real-time PCR and Western, which showed a decrease in the expression of RNA and protein (Fig. 4, C and D). We also checked whether there is a change in histone 3 acetylation at lysine 9 with antibody specific for acetylated histone H3K9Ac (provided in the kit). We found that under HHcy there was an increase in the amount of H3K9 acetylation, which decreased upon treatment with MK801 (Fig. 4E).

Fig. 4.

Expression of methylation and acetylation genes. We looked into the expression of methylation (DNMT1) and acetylation genes [histone deacetylase 1 (HDAC1), H3K9Ac] through Western and real-time PCR. Western blot (A) and real time (B) showed an increase in the expression of DNMT1 in HHcy. There was decrease in the expression of HDAC1 through Western blot (C) and real-time PCR (D) in HHcy. E: the histone acetylation using H3K9Ac antibody showed an increase in acetylation at histone 3 at lysine 9. *P < 0.05 compared with control; #P < 0.05 compared with 100 μM Hcy.

Expression of MMP9.

Increase in oxidative stress leads to cardiovascular remodeling, which is due the induction of MMPs. We checked the expression of the MMP9 gene with semiquantitative PCR (Fig. 5A), as well as the protein with Western (Fig. 5B). We found an increase in the expression of the genes as well as protein after Hcy treatment. The expression decreased in the Hcy + MK801 group. These results were consistent with the zymography gel assays for the detection of the MMP9 protein activity (Fig. 5B). The zymography gels showed an increase in MMP9 activity on treatment with 100 μM Hcy. We further confirmed the histone acetylation at MMP9 gene with ChIP assay and found an increase in the amount of acetylation at the MMP9 gene (Fig. 5C).

Fig. 5.

The expression of remodeling gene matrix metalloproteinase 9 (MMP9). To determine the expression of the remodeling gene MMP9, we used semiquantitative PCR (A) and Western blotting (B). The MMP9 activity was determined by gelatin zymography (B). The data confirmed an increase in the activity of MMP9 in HHcy. C: chromatin immunoprecipitation (ChIP) and PCR using MMP9 specific primers. Hcy (100 μM) treatment increased the acetylation of H3K9 at the MMP9 promoter regions so the MMP9 inputs (template DNA) are increased in the immunoprecipitation mediated by H3K9Ac antibody, which showed increased amplification through PCR. *P < 0.05 compared with control; #P < 0.05 compared with 100 μM Hcy.

In Vivo Studies on CBS+/− Heart Tissue

Expression of NMDAR1, DNMT1, and HDAC1.

We checked the expression of NMDAR1, DNMT1, and HDAC1 through immunohistochemistry (IHC) staining of heart tissue and found an increase in the expression of NMDAR1 in HHcy in CBS+/− mouse (Fig. 6A). Additionally, there was an increase in the expression of DNMT1, which suggests increased methylation patterns in HHcy (Fig. 6A). We also found a decrease in HDAC1 expression, which suggests histone deacetylation (Fig. 6A). We also confirmed the expression of DNMT1 and HDAC1 by measuring the corresponding enzyme activity in the heart tissue. The results of the enzyme activities were consistent with the results obtained in the IHC studies (Fig. 6B).

Fig. 6.

The expression of methylation and acetylation genes in vivo. The results obtained in the HL-1 cardiomyocyte cells were confirmed in vivo using CBS+/− heterozygous mice with HHcy. A: immunohistochemistry of heart tissues showed increase in the expression of NMDAR1 (green color 1st 2 rows) in CBS+/− mice compared with wild type (WT). Additionally, there was an increase in the expression of DNMT1 and decreased expression of HDAC1. The results correlated with the in vitro findings. B: DNMT and HDAC activity was assessed by ELISA (Epigentek kit). The results showed an increase in the DNMT activity, which was determined as optical density (OD)·h⁻¹·mg⁻¹, and a decrease in the HDAC activity (OD·h⁻¹·ml⁻¹). *P < 0.05 compared with WT; #P < 0.05 compared with WT.

5-Methylcytosine DNA methylation levels.

We assessed the levels of 5-methylcytosine in the genomic DNA of CBS+/− mice by ELISA using antibody specific to 5-methylcytosine. We found an increase in the level of 5-methylcytosine in HHcy CBS+/− mice compared with WT mice (Fig. 7). The data represent the average of six independent experiments. These results are consistent with our earlier results in brain tissue (16).

Fig. 7.

Assessment of 5-methylcytosine (mc) levels. The total DNA methylation level in HHcy CBS+/− was quantified by ELISA using the monoclonal antibody for 5-mc. We plotted a standard curve by using positive and negative controls provided in the 5-mc DNA ELISA kit (Zymo Research), and the absorbance of the sample was compared with the standard curve for methylation percentage. We found a 1.7% increase in the 5-mc levels in the genomic DNA isolated from heart, which suggests an increase in the methylation patterns in HHcy. The data represent the average of 6 independent experiments.

DISCUSSION

Our study aims to elucidate DNA methylation vs. histone acetylation epigenetic mechanisms in cardiomyocytes under high-Hcy conditions. Our hypothesis is that under high-Hcy conditions there is high expression of the NMDAR1 receptor, which leads to an increase in oxidative stress inside cardiomyocytes. Increased oxidative stress leads to the downregulation of miR-133a and -499, which regulate the expression of DNA methylation and histone modification enzymes. Finally, this leads to the altered expression of genes such as MMP9, which results in cardiac remodeling.

Hcy has been reported as agonist of the NMDAR1 receptor, which is expressed in cardiomyocytes. In this study we found increased expression of NMDAR1 in HL-1 cardiomyocytes and oxidative stress with high levels of Hcy (100 μM). There is prior evidence (1) that explains that agonists increase the expression of NMDAR1 through a feedback positive loop. We suggest that the increase in the expression of NMDAR1 is due to downstream events after the agonists bind to the receptor. We observed that there was an increase in oxidative stress after NMDAR1 activation by Hcy. The increase in oxidative stress increases NMDAR1 expression by a positive feedback loop. Another study by Betzen et al. (3) reported that exogenous ROS upregulate the expression of NMDAR1, which is due to a positive feedback loop. We also showed previously that cardiac-specific deletion of NMDAR1 restores contractility in cardiomyocytes in HHcy (25). The oxidative stress posed by Hcy is overcome by a compensatory mechanism to increase antioxidant enzymes such as superoxide dismutase. Moat et al. (24) have reported that an increase in plasma Hcy levels leads to an increase in activity of erythrocyte SOD. In patients with homocysteinuria, there is an increase in the SOD levels with increase in plasma Hcy (36). Mendes et al. (22) have reported that in male Wistar rats, HHcy induces cardiac dysfunction by increasing oxidative stress, which is followed by increased levels of the SOD enzyme. In agreement with these studies we found an increase in the levels of SOD enzyme in HL-1 cardiomyocytes with high Hcy (100 μM).

High levels of Hcy have been associated with DNA methylation, but the mechanisms are still being explored. Since Hcy is a precursor to S-adenosyl methionine, which is a universal methyl donor, high levels of Hcy lead to high levels of S-adenosyl methionine. Since S-adenosyl methionine is one of the substrates for DNMTs, high levels increase the activity of DNMTs, which is reflected in the global methylation experiment (Fig. 7). Kinoshita et al. (17) have shown that plasma Hcy levels are positively associated with DNA methylation (71.7%) in the CpG islands and negatively correlated at 28.3% CpG sites. Hence, there is a hypermethylation at CpG sites in the HHcy condition. Jia et al. (14) have shown Hcy induced hypermethylation of DDH2 promoter in endothelial cells along with an increase in the expression of DNMT1. Similarly, in aortic aneurysm, an increase in the expression of DNMT1 has been reported in HHcy (27). We also found an increase in the levels of DNMT1 enzyme activity as well as expression in HHcy.

In our laboratory, we showed earlier that miRNAs are involved in cardiac remodeling in HHcy (23) and miR-133a regulates DNA methyl transferases in diabetic cardiomyocytes (5). The study states that a decrease in the level of miR-133a leads to an increase in the expression of DNMT1. Hence we aimed to explore the role of miR-133a in HHcy. In the present study we found a decrease in the levels of miR-133a and -499, which have been reported to be associated with cardiac dysfunction (9). To check whether this leads to increased DNMT1 expression we used mimics and inhibitors. When we used 133a mimics we found a profound decrease in DNMT1 expression, while with 133a inhibitors it increased. However, there was not much difference in the expression of DNMT1 with the use of 499 mimics and inhibitors (Fig. 3B). MiRNAs have been reported to target DNA methyltransferase enzymes such as DNMT3b. Chen et al. (6) have reported that miR-29b regulates DNMT3b in high oxidative-low density lipoprotein conditions.

Apart from DNA methylation, Hcy has also been associated with histone acetylation to alter the expression of cardiac remodeling genes. Narayanan et al. (27) have reported an increase in acetylation of histone 3 at lysine 9 in HHcy during aortic aneurysm and an increase in expression of the MMP9 gene. In a rat model of diet-induced HHcy, global protein and histone methylation levels are affected in a tissue-specific manner (11). The authors show that Hcy elevation elicits an H3R8me2a decrease in the brain. In our study on cardiomyocytes, we found an increase in H3K9 acetylation and a decrease in the expression of HDAC1. The transcriptional or epigenetic mechanism that might be responsible for a decrease in HDAC1 expression is an increase in methylation status of this gene. We have shown that there is an increase in the overall genomic DNA methylation in HHcy, which may lead to the methylation of HDAC1. Zhao et al. (38) have shown that there are 54 CpGs sites within the promoter CpG island (571 bp) of HDAC1 (GenBank accession no. NC_000070), which may get methylated and lead to decrease in expression. With ChIP assay, we found more acetylation of the MMP9 gene, which showed an increase in activation in HHcy conditions.

To confirm the results in vivo we performed the same experiments in the CBS+/− mouse, which has high levels of HHcy after being fed a methionine diet. We found increased levels of the DNMT1 and decreased levels of HDAC1 protein in the heart. We confirmed the results by performing DNMT and HDAC enzyme activities through ELISA. When HHcy was induced, the expression of NMDAR1 increased in mouse heart tissue. To determine DNA methylation levels we checked 5-methylcytosine levels through ELISA and found increased levels of 5-methylcytosine in CBS +/− heart tissue, which is suggestive of increased DNA methylation.

Based on our results and previous study we propose a mechanism of Hcy-induced cardiovascular risk, presented in Fig. 8. Hcy triggers epigenetic changes through the activation of NMDAR1 in cardiomyocytes followed by increased oxidative stress, which leads to decreased levels of in miR-133a and -499. We postulate that these changes together with an increase in the levels of DNMT1 and H3K9 acetylation, which are suggestive of chromatin modification and DNA methylation under HHcy, account for increased risk of cardiac dysfunction. This study provides a molecular insight into the epigenetic mechanisms posed by increased Hcy levels on cardiomyocytes, which can lead to cardiac dysfunction.

Fig. 8.

Proposed hypothesis. Represented is the proposed hypothesis for Hcy-induced mechanisms in cardiac remodeling. Hcy increased oxidative stress inside the cells by agonizing the NMDAR1 receptor leading to an increase in the DNMT1 expression and histone acetylation through the involvement of miR-133a. This further led to increase in the activity of MMP9 enzymes that induce cardiac remodeling.

GRANTS

Part of this study was supported by National Heart, Lung, and Blood Institute Grants HL-74185 and HL-108621.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: P.C., A.K., P.K.K., and A.F. performed experiments; P.C. and S.C.T. analyzed data; P.C. and S.C.T. interpreted results of experiments; P.C. prepared figures; P.C. drafted manuscript; P.C. and S.C.T. edited and revised manuscript; S.G. and S.C.T. conception and design of research; S.C.T. approved final version of manuscript.