*Article* **DNMT3B System Dysregulation Contributes to the Hypomethylated State in Ischaemic Human Hearts**

**Estefanía Tarazón 1,2,\*,†, Lorena Pérez-Carrillo 1,†, Isaac Giménez-Escamilla 1, María García-Manzanares 1,3, Luis Martínez-Dolz 1,2,4, Manuel Portolés 1,2 and Esther Roselló-Lletí 1,2,\***


**Abstract:** A controversial understanding of the state of the DNA methylation machinery exists in ischaemic cardiomyopathy (ICM). Moreover, its relationship to other epigenetic alterations is incomplete. Therefore, we carried out an in-depth study of the DNA methylation process in human cardiac tissue. We showed a dysregulation of the DNA methylation machinery accordingly with the genome-wide hypomethylation that we observed: specifically, an overexpression of main genes involved in the elimination of methyl groups (*TET1*, *SMUG1),* and underexpression of molecules implicated in the maintenance of methylation (*MBD2*, *UHRF1)*. By contrast, we found *DNMT3B* upregulation, a key molecule in the addition of methyl residues in DNA, and an underexpression of miR-133a-3p, an inhibitor of *DNMT3B* transcription. However, we found many relevant alterations that would counteract the upregulation observed, such as the overexpression of *TRAF6*, responsible for Dnmt3b degradation. Furthermore, we showed that molecules regulating Dnmts activity were altered; specifically, SAM/SAH ratio reduction. All these results are in concordance with the Dnmts normal function that we show. Our analysis revealed genome-wide hypomethylation along with dysregulation in the mechanisms of addition, elimination and maintenance of methyl groups in the DNA of ICM. We describe relevant alterations in the DNMT3B system, which promote a normal Dnmt3b function despite its upregulation.

**Keywords:** ischaemic cardiomyopathy; DNMT3B; DNA methylation; non-coding RNA

#### **1. Introduction**

Ischaemic cardiomyopathy (ICM) is an important cause of mortality worldwide, associated with the development of heart failure (HF) [1]. Therefore, there is a great interest in the study of ICM to obtain a better understanding of the pathophysiology underlying the disease. One of the main research foci is epigenomics, which has allowed the identification of epigenetic alterations in ICM patients [2–4]. These epigenetic changes are associated with differential expression at the mRNA level of genes described in the development of ICM [5–7]. In addition, it has been observed that the epigenetic alterations produced during disease development could be valuable diagnostic and prognostic markers, and serve as therapeutic targets [8–11].

The most studied epigenetic modification is DNA methylation. It is a complex process, characterised by two main mechanisms: the addition and the elimination of methyl residues

**Citation:** Tarazón, E.; Pérez-Carrillo, L.; Giménez-Escamilla, I.; García-Manzanares, M.; Martínez-Dolz, L.; Portolés, M.; Roselló-Lletí, E. DNMT3B System Dysregulation Contributes to the Hypomethylated State in Ischaemic Human Hearts. *Biomedicines* **2022**, *10*, 866. https://doi.org/10.3390/ biomedicines10040866

Academic Editor: Tânia Martins-Marques

Received: 11 March 2022 Accepted: 5 April 2022 Published: 7 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

193

in DNA. The addition of methyl groups to the cytosine residues of DNA, which is associated with gene repression, is directly catalysed by three members of the DNA methyltransferase (DNMT) family. These molecules require S-adenosylmethionine (SAM), a methyl group donor, to carry out the reaction, which is then transformed into S-adenosyl-L-homocysteine (SAH). On the other hand, the removal of methyl groups from DNA is a complex process since several enzymatic reactions are required [12]. Moreover, in recent years, a large number of molecules implicated in DNA methylation have been described [13]. These key molecules can be classified based on their main function, as molecules involved in the addition, elimination and maintenance of the methyl residues on DNA or regulators of the methylation process. Other epigenetic mechanisms, such as post-translational modifications of histones and RNA-based mechanisms, mainly miRNAs, can also interact with the DNA methylation machinery. This allows complex gene regulation through the different components of the epigenome [14,15].

In the context of HF, an increase in the expression of the de novo *DNMTs*, *DNMT3A* and *DNMT3B*, has been observed [16], *DNMT3B* being the main DNA methyltransferase expressed in human and mouse hearts [17]. Controversially, greater hypomethylation has been observed in regions associated with genes that encode proteins in these patients [18]. Moreover, blood genome-wide DNA hypomethylation has been described to be a risk factor for ischaemic heart disease [19].

Therefore, due to the lack of knowledge and the controversies described related to the DNA methylation process in ICM, this study aims to elucidate the state of the DNA methylation machinery. To carry it out, we analyzed the expression of the main genes involved in DNA methylation in ischaemic hearts. Furthermore, to clarify the role of *DNMT3B* upregulation, we studied the state of the DNMT3B system and the genome-wide methylation state.

#### **2. Materials and Methods**

#### *2.1. Cardiac Tissue Samples*

Left ventricular (LV) tissue samples from 40 explanted human hearts were used in our experiments, some of them matching between different analyses. The specific sample size of each individual study is reported in Table 1. Patients were diagnosed with ICM based on the following inclusion criteria: prior documented episodes of acute myocardial infarction, an echocardiography showing normal contractility segments coexisting with other dyskinetic or akinetic segments and an electrocardiography showing signs of ischemia or myocardial necrosis. There were no signs of existence of a primary valvular disease. The ICM characteristics for inclusion in the study were obtained from clinical history, hemodynamic study, electrocardiograms and Doppler echocardiography data. All the data was collected by physicians who were blind to the subsequent analysis of the LV function. Patients were functionally classified following the New York Heart Association criteria and were receiving medical treatment in accordance with the guidelines of the European Society of Cardiology [20].

All controls (CNTs) had a normal LV function (LV ejection fraction >50%) and no history of any cardiac disease. Samples were obtained from non-diseased donor hearts that had been rejected for cardiac transplantation due to size or blood type incompatibility. All heart donors died either from a cerebrovascular accident or in a motor vehicle accident.

Tissue samples were collected from near the apex of the left ventricle, maintained in 0.9% NaCl, and preserved at 4 ◦C for a maximum of 4.4 ± 3 h after coronary circulation loss. The samples were stored at −80 ◦C until further use. Appropriate handling and rapid sample collection and storage led to high-quality samples (DNA ratios 260/280 ~1.8 and 260/230 ~2.0, RNA ratio 260/280 ~2.0 and RNA integrity number ≥9).

This study was approved by the Ethics Committee (Biomedical Investigation Ethics Committee of La Fe University Hospital of Valencia, Spain) and was conducted in accordance with the guidelines of the Declaration of Helsinki [21]. Signed informed consent was obtained from each patient or, in the case of CNT subjects, from their relatives.


**Table 1.** Clinical characteristics of ischaemic cardiomyopathy patients.

BMI: body mass index; ICM: ischaemic cardiomyopathy; LVEF: left ventricular ejection fraction; LVEDD: left ventricular end-diastolic diameter; LVESD: left ventricular end-systolic diameter; NYHA: New York Heart Association. Data are presented as the mean ± SD.\* The same patients were used to measure Dnmts nuclear activity.

#### *2.2. mRNA Extraction and Sequencing*

For this analysis, 23 samples were used (ICM, *n* = 13; and CNT, *n* = 10). RNA isolation and RNA-sequencing (RNA-seq) procedures and analyses have been extensively described previously by Roselló-Lletí et al. [22]. Briefly, RNA extractions were performed using a PureLink™ Kit (Ambion Life Technologies; Waltham, MA, USA) and cDNA libraries were obtained following Illumina's recommendations. Transcriptome libraries were sequenced on the SOLiD 5500 XL (Applied Biosystems; Waltham, MA, USA) platform. All data presented in this manuscript has been deposited in the NCBI's Gene Expression Omnibus (GEO) database and are accessible through the GEO series accession number GSE55296.

#### *2.3. ncRNA Extraction and Sequencing*

For this analysis, 30 samples were used (ICM, *n* = 22; and CNT, *n* = 8). RNA extraction was carried out using the Quik-RNATM miniprep plus kit (Zymo Research; Irvine, CA, USA) and following the manufacturer's instructions. RNA quantification was performed using NanoDrop 1000 spectrophotometer and Qubit 3.0 fluorometer (Thermo Fisher Scientific; Horsham, UK). The purity and the integrity of the RNA samples were determined using a 0.8% agarose gel and the Agilent 2100 Bioanalyzer with RNA 6000 nano assay and small RNA assay kits (Agilent Technologies; Spain).

The cDNA libraries have been obtained following Illumina´s recommendations. Briefly, 3 and 5 adaptors were sequentially ligated to the RNA prior to reverse transcription and cDNA generation. The cDNA was enriched using PCR to create an indexed double-stranded cDNA library, and size selection was performed using a 6% polyacrylamide gel. The quality and quantity of the libraries were analyzed using a 4200 TapeStation (Agilent Technologies; Madrid, Spain) D1000 High-Sensitivity assay. The cDNA libraries were pooled and then sequenced using paired-end sequencing (100 × 2) in the Illumina HiSeq 2500 (Illumina; San Diego, CA, USA) sequencer.

#### *2.4. Validation for RT-qPCR*

Real-time quantitative polymerase chain reaction (RT-qPCR) was performed to validate the expression of *DNMT3B* in the LV samples (ICM, *n* = 14; and CNT, *n* = 8). One microgram of total RNA was reverse transcribed into cDNA using the M-MLV enzyme, as per the manufacturer's instructions (Invitrogen; Waltham, MA, USA). The resulting

cDNA was used as a template for RT-qPCR, which was performed in duplicate using the high-performance ViiATM 7 Real-Time PCR System thermal cycler, according to the manufacturer's instructions (Applied Biosystems; Waltham, MA, USA) and using TaqMan® (Thermo Fisher Scientific; Horsham, UK) probes: *DNMT3B* (Hs00171876\_m1) and *GAPDH* (Hs99999905\_m1). *GAPDH* was used as the reference gene. The relative expression of the DNMT3B gene was calculated according to the Livak method of 2−ΔΔCt [23].

#### *2.5. Enzyme-Linked Immunosorbent Assay*

Dnmt3b protein levels, Dnmts activity and the concentrations of S-adenosylmethionine (SAM) and S-adenosylhomocysteine (SAH) were determined via a specific sandwich enzyme-linked immunosorbent assay, as per the manufacturer´s instructions (DNMT3B Assay Kit ab113471 from Abcam (Cambridge, UK), DNMT Activity Quantification Kit (Colorimetric) ab113467 from Abcam, Human S-adenosylmethionine SAM ELISA kit MBS2605308 and Human S-adenosylhomocysteine SAH ELISA kit MBS2602645 from MyBioSource (San Diego, CA, USA)). The tests were quantified at 450 nm in a dual-wavelength microplate reader (Sunrise; Tecan, Tecan Ibérica Instrumentación S.L Barcelona, Spain) using Magellan version 2.5 software (Tecan).

Dnmt3b protein levels and Dnmts activity were measured in nuclear extracts from homogenate heart tissue (ICM, *n* = 8; CNT, *n* = 8). The isolation method of the nuclear fraction is detailed by Tarazón et al. [24]. Briefly, 100–150 mg of frozen left ventricle was used for the extraction, which was homogenized in an extraction buffer. The homogenates were subjected to a sucrose gradient and to successive ultracentrifugations until the isolation of the nuclear fraction. The concentrations of SAM and SAH were measured in cardiac tissue homogenates (ICM, *n* = 30; and CNT, *n* = 6). For the extraction, we used 25 mg of the frozen left ventricle. These were then homogenized with an extraction buffer in a FastPrep-24 homogenizer (MP Biomedicals; Irvine, CA, USA) with specifically designed Lysing Matrix D tubes. The homogenates were centrifuged, and the supernatant was aliquoted. The SAM and SAH tests have a limit of detection up to 5 ng/mL and 0.06 ng/mL, respectively. No significant cross-reactivity or interference between these metabolites and their analogues were observed.

#### *2.6. DNA Extraction and Infinium MethylationEPIC BeadChip*

For this analysis, 16 samples were used (ICM, *n* = 8; and CNT, *n* = 8). DNA isolation and DNA methylation analyses—using the Infinium MethylationEPIC BeadChip platform (Illumina; San Diego, CA, USA) to interrogate over 850,000 CpG sites—were performed and extensively described by Ortega et al. [7].

#### *2.7. Statistical Methods*

Data are expressed as the mean ± standard deviation (SD) of the mean for continuous variables and as percentage values for discrete variables. The Kolmogorov–Smirnov test was used to analyze the data distribution. Significant differences between the means of groups with a normal distribution were analyzed using the Student's *t*-test, while the non-parametric Mann–Whitney U test was used for data that was non-normally distributed. Finally, Pearson's correlation coefficient was calculated to determine the statistical relationship between variables. The molecules with *p* < 0.05 and fold > 1.20 were considered statistically significant. CpGs with a Δβ ≥ ±0.1 and *p* < 0.05 were considered as differentially methylated. All statistical analyses were performed using the SPSS software (version 20.0) for Windows (IBM SPSS Inc.; Endicott, NY, USA).

#### **3. Results**

#### *3.1. Clinical Characteristics of Patients*

This study included explanted heart tissue samples from CNT individuals and HF patients from ischaemic etiology. The six study populations with ICM were homogeneous with regard to their clinical characteristics (Table 1). ICM patients had a mean age of 56 ± 8 years, and 99% were male. They belonged to classes III–IV of the New York Heart Association functional classification and presented altered values in echocardiographic parameters. Furthermore, they were previously diagnosed with significant comorbidities, including hypertension and diabetes mellitus. CNT individuals had a mean age of 55 ± 17 years, and 65% were male. Comorbidities and other echocardiographic data were not available for the CNT group, in accordance with the Spanish Organic Law on Data Protection 15/1999.

#### *3.2. mRNA Expressio£n of Genes Involved in DNA Methylation*

Transcriptome-level differences between ICM and CNT samples were investigated using the mRNA-seq technology. We focused on the study of the genes involved in the DNA methylation machinery; these genes were classified based on their main function, relating to the addition, elimination, maintenance of methyl residues in DNA, or such as regulators of the DNA methylation process (Supplementary Table S1). We found alterations in several molecules of the methylation machinery (Figure 1a). Specifically, we observed the overexpression in *DNMT3B* gene (1.90 fold and *p* < 0.001), a molecule involved in the addition of de novo methyl groups in the DNA. Amongst the major genes analyzed associated with the elimination of methyl groups that were expressed in the heart, we observed that *TET1* (1.94 fold and *p* < 0.05) and *SMUG1* (1.36 fold and *p* < 0.05) were overexpressed in ICM patients. In addition, *MBD2* (−1.27 fold and *p* < 0.01) and *UHRF1* (−1.52 fold and *p* < 0.01) genes were underexpressed, both implicated in the maintenance of methyl groups. Finally, genes involved in the regulation of the methylation process were analyzed (Figure 1b). Specifically, we observed alterations in genes implicated in the regulation of the addition of methyl group, *AHCY* (−1.46 fold and *p* < 0.05) and *TRAF6* (1.31 fold and *p* < 0.05), and overexpression of genes related to the regulation of the elimination of methyl groups, *GADD45G* (2.33 fold and *p* < 0.001) and *OGT* (1.33 fold and *p* < 0.05).

**Figure 1.** mRNA relative expression levels of altered genes involved in the DNA methylation process in ischaemic hearts versus controls. (**a**) Methylation machinery of DNA: genes related to addition (*DNMT3B*), elimination (*TET1* and *SMUG1*) and maintenance (*MBD2* and *UHRF1*) of methyl groups to DNA. (**b**) Methylation machinery regulators (*AHCY*, *TRAF6*, *GADD45G* and *OGT*). The results were obtained by mRNA-sequencing SOLiD 5500XL platform. Data are presented as the mean ± SEM. a.u., arbitrary units. Ischaemic cardiomyopathy patients (*n* = 13; blue), controls subjects (*n* = 10; grey). \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.

#### *3.3. DNMT3B Regulation Analysis*

We validated the expression of *DNMT3B*, which plays a central role in the DNA methylation process and its overexpression contrasted with the general hypomethylation observed in these ICM patients. The results obtained by RT-qPCR corroborated the *DNMT3B* overexpression (2.22 fold and *p* < 0.01; Figure 2a) observed by RNA-seq (Figure 1a). We carried out a second RNA sequencing study, which allows the detection of ncRNAs and we observed alterations in several miRNAs previously described as regulators of *DNMT3B* expression. miR-133a-3p (Figure 2b), negative regulator of *DNMT3B* transcription, was reduced (−1.38 fold and *p* < 0.001). In addition, post-transcriptional regulators miRNAs of *DNMT3B* were altered (Figure 2c). Specifically, the miR-30d-5p (1.22 fold and *p* < 0.01) and miR-379-5p (1.41 fold and *p* < 0.05) were overexpressed, while the miR-29c-3p (−1.30 fold and *p* < 0.01) and miR-221-5p (−1.33 fold and *p* < 0.01) were underexpressed in ICM patients. Next, nuclear levels of the Dnmt3b protein were analyzed and we did not observe statistically significant changes between ICM and control individuals (Figure 2d).

**Figure 2.** DNMT3B system dysregulation. (**a**) *DNMT3B* mRNA levels in ischaemic hearts (*n* = 14) and controls (*n* = 8), using RT-qPCR. (**b**) miR-133a-3p expression levels in ischaemic hearts (*n* = 22) and controls (*n* = 8), using HiSeq 2500 platform. (**c**) miRNAs expression levels in ischaemic hearts (*n* = 22) and controls (n = 8), using HiSeq 2500 platform. (**d**) Dnmt3b nuclear protein levels in ischaemic hearts (*n* = 8) and controls (*n* = 8), using ELISA. (**e**) SAM/SAH ratio in ischaemic hearts (*n* = 30) and controls (*n* = 6), using ELISA. (**f**) lncRNA *H19* expression level in ischaemic hearts (*n* = 22) and controls (*n* = 8), using the Illumina HiSeq 2500 platform. (**g**) Dnmts nuclear activity in ischaemic hearts (*n* = 8) and controls (*n* = 8), using ELISA. Data are presented as the mean ± SEM. a.u., arbitrary units. Ischaemic cardiomyopathy patients (blue), controls subjects (grey). \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.

On the other hand, we analyzed different molecules that regulate the activity of Dnmts to determine whether their levels could favor the inhibition of Dnmts and, consequently, the global hypomethylation in ICM patients. Lower levels of SAM and higher levels of SAH are related to reduce methylation capacity, which is represented by the SAM/SAH ratio [25]. We measured SAM and SAH levels in human heart samples through ELISA analyses. The SAM/SAH ratio (Figure 2e) was reduced by 37% in ICM patients (*p* < 0.05). Specifically, we obtained a SAM/SAH ratio of 631.45 ± 163.26 in the CNT group and 461.55 ± 173.72 in ICM patients. Moreover, as previously shown, *AHCY* was underexpressed in ICM patients (Figure 1b). This gene encodes the S-adenosyl-L-homocysteine hydrolase (SAHH) protein, which is responsible for degrading SAH. Furthermore, through the sequencing of ncRNAs, we detected the levels of lncRNA *H19* (Figure 2f), which was overexpressed in ICM patients (1.68 fold and *p* < 0.01). *H19* is an SAHH inhibitor that favors the accumulation of SAH in the cells. In addition, the expression of *H19* correlated with the ejection fraction (r = 0.590 and *p* < 0.01). Next, we analysed Dnmts activity, observing absence of changes in ICM patients compared to control individuals (Figure 2g).

#### *3.4. State of Global DNA Methylation*

We analyzed global DNA methylation levels in ICM patients and CNT individuals. The β value and the *p* value were calculated for each detected CpG site, allowing the identification of 643 differentially methylated CpGs sites. We represented each CpG with a statistically significant value using a heat map (Figure 3a), which identified the ICM and CNT groups in two different methylation patterns. Specifically, the 82.9% of the CpGs sites were hypomethylated in ICM patients (Figure 3b). Furthermore, we analyzed the global methylation mean between both groups (ICM, mean β value = 0.481 ± 0.164 and CNT, mean β value = 0.569 ± 0.171; *p* < 0.0001 (Figure 3c)). These results showed genome-wide hypomethylation in ICM patients.

Next, we analyzed the position in the genome of the differentially methylated CpGs (Table 2). The 26.1% sites were distributed in the promoter proximal regions (TSS1500, TSS200 and 5´UTR), 44.3% in gene body and first exons, and 29.6% in 3´UTR and intergenic regions. In relation to CpG content and neighborhood context, the 54% of CpGs were present in open sea; the 27.2% in shores and shelves of CpG sites; and the 18.8% in CpGs sites. Although there is an evident greater number of a hypomethylated CpGs sites in ICM patients, we observed that the percentage of methylation is similar for hypermethylated and hypomethylated CpGs sites in each region, except for absence of differentially hypermethylated CpGs sites in 3´UTR regions and the marked reduction in the percentage of hypermethylation in the context of the TSS200 regions and islands.


**Table 2.** Comparison of statistically significant CpG site methylation (Δβ > 0.1 and *p* < 0.05) between ischaemic cardiomyopathy patients and control individuals.

#### **Table 2.** *Cont*.


**Figure 3.** Global CpGs hypomethylation in ischaemic cardiomyopathy patients. (**a**) Heat map representing global methylation in ischaemic hearts (*n* = 8) and controls (*n* = 8) for each statistically significant CpGs. The results are represented as individual β values (1 = hypermethylated (red) and 0 = hypomethylated (green)). (**b**) Bar graph illustrating the number of hypermethylated and hypomethylated CpGs in ischaemic hearts. (**c**) Box plot graph with global methylation mean. Data are presented as the mean ± SEM. \*\*\* *p* < 0.0001. The results were obtained by the 850 K Infinium MethylationEPIC BeadChip platform.

#### **4. Discussion**

Gene expression is regulated by pathological mechanisms involving DNA methylation and ncRNAs function. Both processes are closely related, but incomplete knowledge of these epigenetic alterations exists in ICM [26]. Moreover, controversial results have been published regarding DNA methylation in ICM. Specifically, increased expression of methyl group addition molecules contrasts with the DNA hypomethylation observed [16,18]. Thus, this study focused on the analysis of the state of the DNA methylation machinery in these patients. Our results showed an alteration of the DNA methylation machinery in ICM patients; specifically, we found a dysregulation in addition, elimination and maintenance of methyl groups accordingly with the genome-wide hypomethylation observed.

Regarding the mechanisms of elimination of methyl groups, we observed the overexpression of *TET1* and *SMUG1* genes, both molecules favoring DNA hypomethylation [27]. *TET1* plays a major role in the initiation of the active DNA demethylation process [28,29]. Furthermore, *OGT* has been related to the post-transcriptional regulation of *TET1* in animal models [30]. Loss of Ogt in early fetal cardiomyocytes leads to multiple heart developmental defects, including hypertrabeculation, biventricular dilation, atrial septal defects, ventricular septal defects and defects in coronary vessel development [31]. In this study, we observed overexpression of the *OGT* gene. Ogt adds residues of O-GlcNAc to Tet1, thereby stimulating its activity [32]. However, the impact of Ogt on the Tet family in human cells is slightly controversial and requires additional studies [33]. In addition, we observed overexpression of the *GADD45G* gene in ICM. This gene belongs to the *GADD45* gene family, which participates in the regulation of the active demethylation process, and its expression is stimulated by cellular stress [34]. Moreover, Lucas et al. [35] demonstrated that Gadd45g is up-regulated in murine cardiomyocytes subjected to simulated ischemia, and that it promotes the development and persistence of HF by inducing cardiomyocyte apoptosis. These results indicate a dysregulation in the process of elimination of methyl groups, which could favor the demethylation of DNA.

Furthermore, the *MBD2* and *UHRF1* genes were underexpressed in this study. These genes are responsible for reading the DNA methylation pattern and maintaining the methylated state associated with transcriptional repression [12]. Genetic alterations in *MBD2* and *UHRF1* have been linked to various cardiovascular pathologies such as atherosclerosis or arterial aneurism, which can trigger HF [36,37].

On the other hand, we observed overexpression of the *DNMT3B* gene in ICM patients. This gene encodes a methyltransferase responsible for adding methyl groups to DNA, thus promoting the methylated state. Previously, *DNMT3B* overexpression has been reported in the context of HF [16], which could be favored by hypoxic conditions [38]. *DNMT3B* is a molecule subject to significant regulation. At the transcriptional level, different miR-NAs regulators of its expression have been described in human and mouse cells [39–42]. Specifically, the miR-133a-3p binds to the *DNMT3B* promoter, reducing its transcription in cardiac cells [43], finding in our results the underexpression of miR-133a-3p in ICM patients. Furthermore, we observed alterations in several *DNMT3B* target miRNAs, which can induce dysregulation in the levels of mRNA, as well as at the protein level. Next, we analysed Dnmt3b nuclear protein levels. Our results did not show differences at the protein level between patients with ICM and control individuals. Yu et al. [44] have described in murine models of breast cancer that the ubiquitin E3 ligase Traf6 interacts with Dnmts, catalyzing their ubiquitination and inducing protein degradation. *TRAF6* mRNA levels were increased in ICM patients, being able to favor lysosomal degradation of Dnmt3b.

Moreover, the enzymatic activity of Dnmt3b is regulated by different molecules. Specifically, the SAH metabolite acts as an inhibitor of various methyltransferases, including Dnmt3b, thereby preventing the addition of methyl groups to DNA [45,46]. The accumulation of SAH in cells is caused by the SAM-dependent methylation, whose degradation product is SAH, and by the inactivation of SAHH, a molecule involved in the breakdown of SAH into adenosine and homocysteine. The results obtained in this study showed that the SAM/SAH ratio was reduced in ICM patients. SAH is a competitive inhibitor of Dnmts; therefore, the increase in SAH, together with the decrease in SAM, favors the binding of this molecule to one of its main targets. In addition, the reduction in SAM/SAH ratio has been linked to the development of vascular diseases and to DNA hypomethylation of the genome [47]. These results reinforce the existence of a generalized DNA hypomethylation in ICM patients. Furthermore, we observed the underexpression of *AHCY*, gene encoding SAHH and the overexpression of lncRNA *H19* in ICM patients. *H19* acts as an SAHH inhibitor, thus preventing SAH degradation [48]. However, the Dnmts activity was similar in ICM patients and control individuals. In recent years, the role of *H19* in the development and prognosis of HF has been extensively studied [49,50]. Several authors described altered *H19* expression in cardiovascular diseases and related it with hypertrophic growth inhibition in cardiomyocytes and alleviation of myocardial ischemia reperfusion injury [51–53]. We further observed that its expression related with ejection fraction, the central measure of left ventricular function, as previously has been described [52]. These results represent a promising possibility to identify new therapeutic strategies, such as *H19*-targeted delivery for clinical ICM therapy.

To date, alterations in the methylation pattern have been identified in ICM patients in specific regions of the genome [16]. Moreover, the analysis of genome-wide DNA methylation in human ICM was carried out, observing differential methylation in the CpG sites associated with gene promoters [54], as well as a greater hypomethylation in regions associated with genes that encode proteins in these patients [18]. We also analyzed the status of the CpG sites as a whole in ICM patients compared to control individuals, and our results showed a higher degree of hypomethylated CpGs in each of the genomic regions. In addition, the results obtained in the present study indicate an increase in the regulatory mechanisms that promote DNA demethylation. Consistently, our assessment of the global DNA methylation pattern revealed that explanted hearts from individuals with ICM presented genome-wide hypomethylation. A limitation of this study includes the use of cardiac tissue samples from patients with end-stage HF, who exhibit intrinsic variability that exists between individuals. The patients who participated in this study were also treated pharmacologically, and some therapies may have influenced the results. However, our study population was etiologically homogeneous and all individuals had been receiving medical treatment according to the guidelines of the European Society of Cardiology [20].

#### **5. Conclusions**

Our analysis revealed genome-wide hypomethylation concordant with dysregulation in the mechanisms of addition, elimination and maintenance of methyl groups in DNA observed in ischaemic hearts. We describe relevant alterations in the DNMT3B system, which promote a normal Dnmt3b function despite its upregulation.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biomedicines10040866/s1, Table S1: Genes involved in DNA methylation process in human heart.

**Author Contributions:** Conceptualization, M.P. and E.R.-L.; methodology, E.T. and L.P.-C.; validation, I.G.-E. and M.G.-M.; formal analysis, M.P.; investigation, E.T.; resources, L.M.-D.; writing—original draft preparation, E.T., L.P.-C., M.P. and E.R.-L.; writing—review and editing, I.G.-E., M.G.-M. and L.M.-D.; supervision, E.R.-L.; funding acquisition, E.T., L.M.-D. and E.R.-L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Institute of Health "Fondo de Investigaciones Sanitarias del Instituto de Salud Carlos III" [PI20/01469, PI20/00071, CP18/00145, CP21/00041, FI21/00034, FI21/00186], "Consorcio Centro de Investigación Biomédica en Red, M.P." [CIBERCV, under Grant CB16/11/00261], and co-funded by European Union (ERDF, "A way to make Europe") and European Social Fund (ESF, "The ESF invests in your future").

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee (Biomedical Investigation Ethics Committee of La Fe University Hospital of Valencia, Spain; protocol code 2016/0320, 15 November 2016).

**Informed Consent Statement:** All patients submitted a signed informed consent prior to their inclusion in the study, this consent included permission to extract and store their samples, as well as the publication of the results obtained from the research work.

**Data Availability Statement:** The mRNA-seq data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus and are accessible through GEO Series accession number GSE55296 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE55296, accessed on 11 March 2022).

**Acknowledgments:** The authors are grateful to Javier Moreno and Rosana Llobell for their technical support.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Review* **Cardiac Acetylation in Metabolic Diseases**

**Emilie Dubois-Deruy \*, Yara El Masri, Annie Turkieh, Philippe Amouyel, Florence Pinet \*,† and Jean-Sébastien Annicotte \*,†**

> Univ. Lille, Inserm, CHU Lille, Institut Pasteur de Lille, U1167-RID-AGE-Facteurs de Risque et Déterminants Moléculaires des Maladies Liées au Vieillissement, F-59000 Lille, France; yara.elmasri@pasteur-lille.fr (Y.E.M.); ani.turkieh@pasteur-lille.fr (A.T.); philippe.amouyel@pasteur-lille.fr (P.A.)

**\*** Correspondence: emilie.deruy@pasteur-lille.fr (E.D.-D.); florence.pinet@pasteur-lille.fr (F.P.);

jean-sebastien.annicotte@inserm.fr (J.-S.A.)

† These authors contributed equally to this work.

**Abstract:** Lysine acetylation is a highly conserved mechanism that affects several biological processes such as cell growth, metabolism, enzymatic activity, subcellular localization of proteins, gene transcription or chromatin structure. This post-translational modification, mainly regulated by lysine acetyltransferase (KAT) and lysine deacetylase (KDAC) enzymes, can occur on histone or non-histone proteins. Several studies have demonstrated that dysregulated acetylation is involved in cardiac dysfunction, associated with metabolic disorder or heart failure. Since the prevalence of obesity, type 2 diabetes or heart failure rises and represents a major cause of cardiovascular morbidity and mortality worldwide, cardiac acetylation may constitute a crucial pathway that could contribute to disease development. In this review, we summarize the mechanisms involved in the regulation of cardiac acetylation and its roles in physiological conditions. In addition, we highlight the effects of cardiac acetylation in physiopathology, with a focus on obesity, type 2 diabetes and heart failure. This review sheds light on the major role of acetylation in cardiovascular diseases and emphasizes KATs and KDACs as potential therapeutic targets for heart failure.

**Keywords:** acetylation; heart; obesity; diabetes; heart failure; enzymes

#### **1. Introduction**

Obesity is a worldwide epidemic associated with several public health challenges including type 2 diabetes, hypertension, obstructive sleep apnea, dyslipidemia and cardiovascular disease, and notably heart failure. The prevalence of these disorders (e.g., obesity, type 2 diabetes and heart failure) is still increasing and heart failure remains the most common cause of cardiovascular morbidity and mortality in the world [1]. Interestingly, although heart failure could provoke systolic or diastolic cardiac dysfunction, obesity and type 2 diabetes are more often characterized by cardiac hypertrophy and diastolic dysfunction [2]. Moreover, diastolic heart failure remains one of the more challenging diseases to treat and, thus, a better understanding of the physio(patho)logical mechanisms involved in metabolic heart disease is necessary to propose new pharmacological therapies. In this context, epigenetic or protein post-translational modifications, including phosphorylation, methylation, O-GlcNAcylation, ubiquitylation, and acetylation, could represent interesting targets since they modulate gene or protein expression during metabolic diseases [3].

Protein acetylation is a highly conserved mechanism that was first described 60 years ago, in which an acetyl group is covalently attached to the ε-amino group of lysine residues [4]. This modification induces important changes to the protein structure at its lysine residue, by altering its charge status and adding an extra structural moiety [5]. These post-translational modifications can regulate protein–protein interactions, stability and function. Acetylation was first described to regulate chromatin structure and histone

**Citation:** Dubois-Deruy, E.; El Masri, Y.; Turkieh, A.; Amouyel, P.; Pinet, F.; Annicotte, J.-S. Cardiac Acetylation in Metabolic Diseases. *Biomedicines* **2022**, *10*, 1834. https://doi.org/10.3390/ biomedicines10081834

Academic Editors: Tânia Martins-Marques, Gonçalo F. Coutinho and Attila Kiss

Received: 11 July 2022 Accepted: 28 July 2022 Published: 29 July 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

207

activity, associated to epigenetic-mediated gene regulation [4]. Since then, several proteomic analyses enabled to identify more than a thousand acetylation sites in both histone and non-histone proteins [6–8], and, interestingly, revealed that the subcellular distribution of acetylation is tissue dependent [9]. Focusing on cardiovascular and metabolic diseases, Table 1 indicates some of the most common proteins modified by cardiac acetylation.

**Table 1.** Most common proteins modified by cardiac acetylation in cardiovascular and metabolic diseases.


KAT: lysine acetyltransferase; KDAC: lysine deacetylase; LCAD: long chain acyl CoA dehydrogenase; SCAD: short chain acyl CoA dehydrogenase; β-HAD: L-3-hydroxy acyl-CoA dehydrogenase; OPA1: optic atrophic 1; PGC-1α: peroxisome proliferator-activated receptor-gamma coactivator; TBX5: T-Box transcription factor 5; VGLL4: vestigial-like 4; TEAD1: TEA Domain Transcription Factor 1; GATA4: GATA-binding factor 4; MEF: myocyte enhancer factor; SOD2: superoxide dismutase 2; Prx1: peroxiredoxin 1; Nrf2: nuclear factor erythroid-2-related factor 2; eNOS: endothelial nitric oxide synthase; β-MHC: beta-myosin heavy chain; TnI: troponin I; LKB1: liver kinase B1; GCN5L1: general control of amino acid synthesis 5 like-1; SIRT: sirtuins; HDAC: histone deacetylase; mPTP: mitochondrial permeability transition pore.

#### **2. Cardiac Acetylation**

*2.1. Regulation of Lysine Acetylation*

Lysine acetylation is a reversible mechanism that is regulated by the dynamic actions of lysine acetyltransferase (KATs) and lysine deacetylase (KDACs) enzymes [43]. There are three categories of KATs: the MYST family, the Gcn-5-related N-acetyltransferases (GNATs) and the E1A-associated protein of 300 kDa/CREB binding protein (p300/CBP) family [44]. KDACs are broadly categorized into four classes based on function and DNA sequence similarity. Class I (reduced potassium dependency 3 family), II (histone deacetylase 1 family) and IV (HDAC11) are considered as classical KDACs with zincdependent active sites, whereas class III enzymes are a family of silent information regulator 2-like nicotinamide adenine dinucleotide (NAD+)-dependent deacetylases/mono-ADPribosyl transferases known as sirtuins 1 to 7 [44]. Interestingly, the level of expression of KATs and KDACs follows a specific tissue distribution, suggesting specific functions in these different organs [45]. For example, in physiological conditions, KATs are weakly expressed in vascular tissue, except for the lysine acetyltransferase KAT2B, which is highly expressed [45]. Conversely, a database mining approach analyzing the expression profile of 164 enzymes in human and murine tissues highlighted heart as tissue in which KATs and KDACs are in the highest varieties [45]. In addition to tissue-specific expression, both KATs and KDACs have precise subcellular localizations related to their specific targets (Figure 1) and can regulate each other. For example, cardiomyocyte overexpression of SIRT6 decreases P300 levels, where SIRT6 promotes ubiquitination and degradation of P300 [46]. Moreover, cardiomyocyte overexpression of SIRT3 increases SIRT1 expression [27].

**Figure 1.** Subcellular localization of cardiac KATs and KDACs. This figure summarizes the subcellular localization of the most common lysine acetyltransferases (KATs, blue) and lysine deacetylases (KDAC, orange) and their cardiac targets. Ac: acetylated form; HDAC: histone deacetylase; SIRT: sirtuins; β-MHC: beta-myosin heavy chain; TnI: Troponin I; prx1: peroxiredoxin 1; LCAD: long chain acyl CoA dehydrogenase; SCAD: short chain acyl CoA dehydrogenase; β-HAD: L-3-hydroxy acyl-CoA dehydrogenase; SOD2: superoxide dismutase 2; CypD: Cyclophilin D; OPA1: optic atrophic 1; TBX5: T-Box transcription factor 5; VGLL4: vestigial-like 4; GATA4: GATA-binding factor 4; MEF: myocyte enhancer factor; Nrf2: nuclear factor erythroid-2-related factor 2; GCN5L1: general control of amino acid synthesis 5 like-1; CBP: CREB binding protein.

In addition, cofactors play an essential role to modulate acetylation. Several shortchain acyl-CoA products, such as acetyl-coenzyme A (acetyl-CoA), are important metabolite intermediates generated during catabolism of various energy fuels that modulate cardiac acetylation. Indeed, acetyl-CoA acts as a substrate for KATs. Moreover, auto-acetylation

of mitochondrial proteins can occur efficiently at the pH of the matrix (pH ≥ 7.5) in actively respiring mitochondria [9]. On the other hand, levels of NAD+, an important cofactor for sirtuins, may impact protein acetylation [5]. Indeed, NAD<sup>+</sup> levels, synthesized from metabolic pathways, connect their enzymatic activity to metabolism and provide a metabolic link between the sirtuins activity, energy homeostasis and stress responses.

#### *2.2. Physiological Roles of Lysine Acetylation*

Lysine acetylation can modulate cell growth, metabolism, protein–protein interactions, protein stability, subcellular localization, gene transcription, chromatin structure or enzymatic activity [6,7]. This post-translational modification plays an essential role in cardiac physiology (Figure 2).

**Figure 2.** Physiological roles of cardiac acetylation. Ac: acetylated form; KATs: lysine acetyltransferases; KDACs: lysine deacetylase; ROS: reactive oxygen species.

#### 2.2.1. Heart Development and Cardiac Ageing

Acetylation can regulate embryonic heart development and cardiac progenitor fate [47,48]. TBX5, a T-box family transcription factor, is involved in heart and forelimb development and is acetylated by KAT2A and KAT2B at lysine 339, leading to stimulate its transcriptional activity [18]. Knockdown of *KAT2A* and *KAT2B* in zebrafish severely impairs heart development and induces pericardial oedema by inhibiting TBX5 [18]. Besides KATs, the lysine deacetylase enzymes HDAC4 and HDAC5 are also able to interact with TBX5 to exert a repressive role on cardiac genes transcription through deacetylation [19]. Furthermore, TBX1, another T-box family transcription factor expressed in cardiac progenitors, represses myocyte enhancer factor 2c *(Mef2c)* gene expression by reducing histone 3 acetylation on lysine 27 [48]. Moreover, acetylation of vestigial-like 4 (VGLL4) at K225 by P300 negatively regulates its binding to the transcription factor TEA Domain Transcription Factor 1 (TEAD1), leading to decreased neonatal cardiomyocytes proliferation and cardiomyocytes necrosis [20]. Inhibition of HDAC1 was also described to decrease the proliferation cardiomyocyte in zebrafish [49]. Indeed, mutation in *HDAC1* gene that induces protein instability is associated with decreased cardiomyocyte proliferation, suggesting an important role of HDAC1 during heart growth [49].

In addition, sirtuins are well-described regulators of aging and have been associated with longevity [50–52]. Indeed, mice deficient for *SIRT1* [53,54], *SIRT3* [55] or *SIRT6* [56] display a shorter life span with severe cardiac damage, such as hypertrophy and fibrosis. SIRT1 is well known to be cardioprotective by decreasing oxidative stress and inflammation, which promotes cardiomyocyte survival [54]. Moreover, SIRT3 is required for cardiomyocyte survival under several stress conditions such as serum starvation, genotoxic and oxidative stress [57]. Altogether, these studies demonstrate that KDACs and KATs exert crucial functions during heart development and cardiac ageing.

#### 2.2.2. Cardiac Contraction

Recently, reversible acetylation of sarcomeric proteins has been described as a mechanism regulating cardiac function. The comparison of lysine acetylation patterns from rats as well as from human skeletal muscle biopsies revealed that 80% of the proteins involved in muscle contraction were acetylated [9]. Moreover, proteomic analyses suggested that HDAC6 is localized in Z-disks and acts as a sarcomeric protein desacetylase [58]. Among them, acetylation can impact the β-myosin heavy chain (lysine 34, lysine 58, lysine 213, lysine 429, lysine 951 and lysine 1195) [31], titin [33], CapZβ1 (lysine 199) [34] or cardiac troponin I [35], directly affecting cardiac contraction.

2.2.3. Role of Lysine Acetylation in Cardiac Energy Metabolism and Mitochondrial Activity

Proteomic analyses suggested that the subcellular distribution of lysine-acetylated proteins is tissue dependent [9]. For example, the heart and muscles, both high energydemanding organs, are tissues with the largest fraction of mitochondrial protein acetylation [9]. Calorie restriction or changes in nutrition specifically affect the mitochondrial acetylome, but not the cytosol or nucleus [59], suggesting different roles for lysine acetylation in the mitochondria, nucleus and cytoplasm. Indeed, mitochondrial lysine acetylation was described to modulate cell metabolism by regulating fatty acid β-oxidation, the tricarboxylic acid cycle, the urea cycle, and oxidative phosphorylation [60].

In the early newborn period, an important change in myocardial energy substrate metabolism occurs with an increase in fatty acid β-oxidation [47], associated with an increase in long chain acyl CoA dehydrogenase (LCAD) and L-3-hydroxy acyl-CoA dehydrogenase (β-HAD) acetylation and activation [12]. At the cardiac level, mitochondrial general control of amino acid synthesis 5 like-1 (GCN5L1) triggers LCAD and β-HAD acetylation [10,11] whereas SIRT3 deacetylates both LCAD and β-HAD [13,14]. The implication of GCN5L1 in cardiac metabolism was confirmed by genetic deletion approaches. Indeed, knockdown of *GCN5L1* in H9c2 cardiomyoblasts decreases maximal mitochondrial respiration and activity of proteins involved in fatty acid oxidation [12].

However, there is no consensus about the effects of lysine acetylation on fatty acid β-oxidation. Indeed, some studies suggested an inhibitory effect [43] whereas others suggested a stimulatory effect [13,14]. As an example, it was reported that an impaired fatty acid β-oxidation was linked to a reduced acetylation and activity of LCAD in the liver [13] or heart [14] of SIRT3 knockout mice. On the other hand, cardiac mitochondrial protein hyperacetylation induced by *SIRT3* deletion is associated with an increased fatty acid β-oxidation rates [43]. This discrepancy could be due to organ, physio(patho)logical status (e.g., diabetes, obesity or newborn period) or acetylated proteins themselves, and are well described in the following review [43].

The implication of sirtuins in metabolic regulation is well detailed in a recent review [61]. Indeed, SIRT6 was described as a gatekeeper of glucose metabolism in cardiomyocytes. SIRT6 partial depletion decreased mitochondrial respiration whereas SIRT6 overexpression enhanced basal oxygen consumption [62]. Conversely, fatty acid uptake increases in *SIRT6* deficient cardiomyocytes and is decreased in SIRT6-overexpressed cardiomyocytes [63]. Of note, proteins involved in mitochondrial biogenesis can also be acetylated, such as optic atrophy 1 (OPA1) through SIRT3 which is able to activate OPA1 by deacetylation of the residues lysine 926 and lysine 931 [15]. Finally, a cross-talk between SIRT1, peroxisome proliferator-activated receptor gamma coactivator 1 alpha (PGC-1α) and AMP-activated protein kinase (AMPK) has been described to regulate cardiac metabolism [54]. Indeed, PGC-1α is a transcription factor playing a central role in the regulation of cellular energy metabolism, mitochondrial biogenesis and oxidative phosphorylation. SIRT1 could interact with PGC-1α and increase its expression levels [16]. Moreover, AMPK increases cellular NAD+ levels and subsequently increases SIRT1 activity, resulting in PGC-1α deacetylation and activation [64].

#### 2.2.4. Mitochondrial Oxidative Stress

It is well described that several mitochondrial proteins involved in oxidative stress are acetylated in the heart [60]. Oxidative stress is defined by a production of reactive oxygen species (ROS) higher than anti-oxidant capacities. Indeed, although different molecular processes may contribute to global oxidative stress, the majority of ROS originates from the mitochondrial compartment in the heart. Excessive ROS production occurs during mitochondrial dysfunction and induces irreversible damage to mitochondria, defining them as significant contributors to the development of cardiovascular disease [65,66]. In this context, we recently described that SIRT3 deacetylates and activates the mitochondrial superoxide dismutase 2 (SOD2), but there is a decreased interaction between SIRT3 and SOD2 induced SOD2 acetylation on lysine 68 and its subsequent inactivation, leading to mitochondrial oxidative stress and dysfunction in hypertrophied cardiomyocytes [27]. Moreover, SIRT3 inhibition increased oxidative stress in neonatal rat cardiomyocytes [27] or human aortic endothelial cells [26].

Other anti-oxidant enzymes were described to be acetylated in the heart, such as peroxiredoxin 1 [28] in the mitochondria or nuclear factor erythroid 2-related factor 2 [29] in the nucleus.

#### **3. Implication of Cardiac Acetylation in Metabolic Heart Disease**

Post-translational acetylation triggers modification of the activity of several proteins that occurs in obesity, diabetes and early stage of heart failure as detailed below [67,68]. Table 2 summarizes the cardiac modulation of KATs and KDACs expression levels reported in metabolic heart disease.


**Table 2.** Modulation of KATs and KDACs expression in heart failure and metabolic diseases.

KATs: lysine acetyltransferases; KDACs: lysine deacetylases; SIRT: sirtuins; HDAC: histone deacetylase; GCN5L1: general control of amino acid synthesis 5 like-1.

#### *3.1. Cardiac Hypertrophy*

Cardiac hypertrophy is a consequence of genetic, mechanic or neurohormonal changes contributing to heart failure progression. Interestingly, total KAT activity was found to be increased in the hearts of mice subjected to phenylephrine (PE), an hypertrophic stimulus [23]. In this context, the increased expression of P300 and KAT2B acetylases may explain hypertrophied cardiomyocytes under PE [23,25,46,69]. In addition, PE-induced hypertrophy also decreases SIRT6 expression, leading to an increase in histone H3 acetylation on lysine 9 (H3K9) [25,46,69]. Interestingly, PE-induced hypertrophy is decreased by treatment with polyunsaturated fatty acid [69] or anacardic acid [23] in association with a decreased H3K9 acetylation. In parallel, overexpression of P300 induces an increased size of cardiomyocytes and a modification of their myofibrillar organization [69], whereas inhibition of P300 by

siRNA attenuates hypertrophic response in neonatal rat cardiomyocytes [46]. On the other hand, overexpression of SIRT6 [46] or SIRT3 [27] decreased cardiomyocytes hypertrophy. As an example, NAD inhibits oxidative stress and hypertrophy induced by PE in cardiomyocytes or by angiotensin II in mice [36]. Indeed, NAD requires SIRT3 to deacetylate and activate liver kinase B1 (LKB1) and its target AMPK [36]. Furthermore, a decreased SIRT2 protein expression was reported in hypertrophic hearts from mice whereas cardiac-specific SIRT2 overexpression protected the hearts against angiotensin II–induced hypertrophy [37]. As shown for SIRT3, SIRT2 deacetylates LKB1 at lysine 48, these promote the phosphorylation of LKB1 and the subsequent activation of AMPK signaling [37]. Moreover, SIRT7 levels increase in myocardial tissues after pressure overload induced by transverse aortic constriction in mice [21]. Cardiomyocyte-specific deletion of *SIRT7* exacerbates hypertrophy and fibrosis induced by transverse aortic constriction, suggesting an antihypertrophic role of SIRT7 [21]. Indeed, SIRT7 deacetylates the hypertrophy response transcription factors such as GATA-binding factor 4 (GATA4) in cardiomyocytes [21]. Conversely, P300 acetylates GATA4 and activates its DNA binding activity, inducing cardiac hypertrophy and heart failure [22].

Finally, HDAC3 expression and activity increase during hypertrophy [80] whereas a decreased binding of HDAC3 to myofibrils is observed in PE-induced hypertrophy in neonatal rat cardiomyocytes that increases the acetylation of CapZβ1 at lysine 199 and alters myofibril growth during cardiac hypertrophy [34].

#### *3.2. Cardiac Fibrosis*

Cardiac fibrosis, defined as an excessive deposition of extracellular matrix in the cardiac muscle, is an important contributor to several heart diseases. Interestingly, in isoproterenol-injected mice, KAT2B activity was found to be increased only in cardiac fibroblasts, whereas no effects were observed in cardiomyocytes isolated from these mice [39]. In this model, KAT2B induces acetylation of SMAD2 and, subsequently, activates SMAD2 and the transforming growth factor β signaling pathway [39]. SMAD3 could also be acetylated during cardiac fibrosis [41,42]. Activation of SIRT1 by resveratrol [41], nicotinamide mononucleotide [42] or geniposide [70] could decrease SMAD3 acetylation and fibrosis. Moreover, inhibition of P300 appears to exert anti-fibrotic functions [82].

Decreased α-tubulin acetylation and increased HDAC6 were also described in cardiac fibroblasts or the heart from isoproterenol-induced fibrosis. Interestingly, treatment of cardiac fibroblasts with tubastatin A, an HDAC6 inhibitor, restored α-tubulin acetylation and decreased fibrosis [83].

#### *3.3. Heart Failure*

Cardiac hypertrophy and fibrosis are both involved in heart failure development by activation of several intracellular signaling pathways, leading to left ventricular remodeling with systolic dysfunction. Despite the best modern therapeutic management, left ventricular remodeling remains independently associated with heart failure and cardiovascular death at long-term follow-up after myocardial infarction [84]. Moreover, alterations in cardiac energy metabolism, both in terms of changes in energy substrate preference and decreased mitochondrial oxidative metabolism and ATP production, are key contributors to heart failure development [5].

An increase in P300 expression is observed during ischemia/reperfusion injury [85] and myocardial infarction [69]. Interestingly, it was described that activation of P300 acetylase, leading to an increase in acetylated H3K9 induced by myocardial infarction, is reversed by treatment with polyunsaturated fatty acid [69]. Indeed, eicosapentaenoic acid and docosahexaenoic acid directly blocks the histone acetyltransferase activity of P300 and subsequently reduces both hypertrophy and fibrosis induced by myocardial infarction in rats [69]. Other P300 inhibitors were also described to be cardioprotective, such as *Ecklonia stolonifera* Okamura extract [86], curcumin [87] and metformin [88]. Moreover, KAT2B levels are also increased by ischemia/reperfusion injury [73].

The expression and activity of sirtuins are also highly impacted during heart failure. As an example, a decrease in SIRT3 and SIRT6 expressions were observed in human failing hearts that induces a global increase in protein acetylation [63,77]. Proteomic analysis also identified an increased acetylation of mitochondrial proteins induced by transverse aortic constriction [89].

Inhibition of SIRT3 was also described to induce mitochondrial oxidative stress and hypertrophy, notably by increasing the inactive form of SOD2 (acetylated on lysine 68) in hypertrophy [27] or following ischemia/reperfusion [90]. Moreover, deletion of SIRT2 exacerbates cardiac hypertrophy and fibrosis and decreases cardiac ejection fraction and fractional shortening in angiotensin II-infused mice by inhibition of AMPK activation, whereas cardiac-specific SIRT2 overexpression reversed this phenotype [37]. SIRT1 expression and activity are decreased in the heart following ischemia/reperfusion [30]. SIRT1 inhibition was also reported to increase endoplasmic reticulum stress-induced cardiac injury by decreasing eukaryotic initiation factor 2 alpha (eIF2α) deacetylation on lysine 143 in cardiomyocytes and in adult-inducible SIRT1 knockout mice [91].

Acetylation of contractile proteins is also modified during heart failure. For example, acetylation of β-MHC on lysine 951 was decreased in both ischemic and non-ischemic failing hearts [31].

#### *3.4. Obesity*

Obesity is a major cause of disability and is often associated with cardiac hypertrophy, fibrosis, type 2 diabetes, obstructive sleep apnea and alteration of cardiac metabolism. It has notably described an increase in genes involved in fatty acid oxidation [11]. Moreover, several studies have suggested an increase in acetylation raised from the non-enzymatic reaction of high levels of acetyl-CoA generated during a high-fat diet (HFD) and obesity [5]. Furthermore, hyperacetylation of mitochondrial proteins and metabolic inflexibility was reported in response to HFD or obesity. SIRT3 expression is decreased in left ventricles of obese patients and is associated with an increased level of brain natriuretic peptide (BNP), a marker of cardiac dysfunction and of protein acetylation [17]. It was also reported that exposure of cardiomyoblasts H9c2 to palmitate led to an increase in both SIRT3 and GCN5L1 RNA levels [11]. At the metabolic level, HFD induced an increase in SCAD and LCAD acetylation and activity [11]. Moreover, the acetylation level of α-tubulin on lysine 40 is increased in the hearts of HFD mice and a pharmacological activation of α-tubulin acetylation decreases glucose transport [92].

In the other hand, cardiac specific inhibition of SIRT6 in mice exposed to HFD-induced cardiac hypertrophy and lipid accumulation [93]. In this context, SIRT6 activated the expression of endonuclease G and SOD2, that could decrease oxidative stress and hypertrophy [93].

#### *3.5. Type 2 Diabetes*

As obesity, type 2 diabetes is also associated with several cardiac dysfunctions such as cardiac hypertrophy and fibrosis, and acetylation may be involved in these mechanisms. Indeed, SIRT1 expression and activity are decreased in the heart of diabetic rats, induced by HFD and streptozotocin injection and, conversely, an up-regulation of SIRT1 by adenovirus attenuates cardiac dysfunction and oxidative stress [30]. Moreover, in a model of sucrose-fed rats, cardiac dysfunction is associated with a decreased SIRT3 protein expression (despite an increase at RNA level) and an increased GCN5L1 (protein and RNA) and mitochondrial protein acetylation [11,17]. Decreased SIRT3 levels are also associated with an increased acetylation of SOD2, and an increased oxidative stress and apoptosis in heart of diabetic mice [79]. Interestingly, exposure of cardiomyoblasts H9c2 to high glucose concentration decreased SIRT3 [78] and increased GCN5L1, oxidative stress and autophagy mediated by cytoplasmic Forkhead box O1 (FOXO1) acetylation [11,76]. Conversely, expression of GCN5L1 is decreased in hearts of diabetic ZSF1 rats, a model characterized by obese animals with hyperglycemia, hyperinsulinemia and cardiac dysfunction [10]. This decreased GCN5L1 expression and activity, with no modulation of SIRT3, induced

a decrease of short- and long-chain acyl CoA dehydrogenases and a reduced respiratory capacity [10]. One explanation could be the transition from prediabetes, in which GCN5L1 becomes elevated to promote mitochondrial fatty acids oxidation, to an overt diabetic state, in which GCN5L1 expression is downregulated leading to an overall decrease in mitochondrial fuel oxidation [10].

On the other hand, OVE26 mice, a mouse model of type 1 diabetes, develop cardiac dysfunction with increased left ventricle diameters and fibrosis and decreased ejection fraction associated with an increase in HDAC3 activity, oxidative stress and inflammation [81]. Interestingly, the selective HDAC3 inhibitor RGFP966 reversed the cardiac phenotype in these mice [81]. In another model of type 1 diabetes, injection of streptozotocin in rats induced cardiac dysfunction, cardiac hypertrophy, fibrosis and inflammation [72] as well as oxidative stress [28]. This oxidative stress is due to an increased HDAC6 activity leading to a decrease in the acetylated form of peroxiredoxin 1, suggesting that acetylation of peroxiredoxin 1 is involved in this activation [28]. In this context, tubastatin A, a highly selective inhibitor of HADC6, may represent an interesting pharmacological target since treatment of diabetic rats with tubastatin A increased acetylation of peroxiredoxin 1 and decreased oxidative stress and cardiac dysfunction [28]. Conversely, SIRT1 expression and activity appear to be decreased in the left ventricles of streptozotocin-induced diabetic rats [72] and mice [38]. In these models, reduced SIRT1 expression is associated with acetylation of endothelial nitric oxide synthase [72] or Akt [38]. In addition, SIRT6 expression is also decreased in the hearts of diabetic mice [63], suggesting that sirtuin family members may play a crucial role in type 1 diabetes-associated cardiac dysfunctions.

#### *3.6. Pharmacological Modulation of Cardiac Acetylation*

Due to the major implication of KATs and KDACs in cardiac dysfunction induced by obesity, type 2 diabetes or heart failure, targeting these enzymes could be beneficial for patients.

First, inhibition of P300 appears to be a promising approach to inhibit cardiac hypertrophy and fibrosis induced by several stimuli. As an example, oral administration of eicosapentaenoic acid (1 g/kg) or docosahexaenoic acid (1 g/kg) one week after myocardial infarction preserved fractional shortening and decreased cardiac hypertrophy and perivascular fibrosis [68]. *Ecklonia stolonifera* Okamura extract, an algae traditionally used in Japanese foods, inhibits PE-induced hypertrophy in neonatal rat cardiomyocytes and restores fractional shortening and reduced cardiac hypertrophy and fibrosis by oral administration one week after myocardial infarction in rats [73]. Oral administration of curcumin (50 mg/kg/day) acts also as a P300 inhibitor and decreases posterior wall thickness, cardiac hypertrophy and perivascular fibrosis induced by hypertension [74]. Finally, metformin, that directly inhibits P300-mediated acetylation of H3K9, blocks PE-induced cardiomyocytes hypertrophy [75].

Based on the cardioprotective effects of SIRT3, it seems essential to develop new therapeutic strategy to increase its expression or activity. Recent reviews have well described these molecules [94,95]. For example, exogenous NAD is able to inhibit hypertrophy induced by PE in cardiomyocytes or by angiotensin II in mice [36]. Resveratrol, an activator of both SIRT1 and SIRT3, is also described to be cardioprotective [41].

#### **4. Conclusions**

Despite available therapies, cardiovascular diseases, and particularly ischemic diseases, still remain the first cause of mortality and morbidity in the world. Obesity and type 2 diabetes are some of the major risk factors of cardiovascular diseases. Acetylation is an essential mechanism involved in several processes contributing to cardiac diseases such as cell metabolism, gene transcription or enzymatic activity. In this context, targeting enzymes responsible of acetylation/deacetylation could be beneficial and is a very promising research area, as demonstrated by the development of inhibitor targeting P300 [82], modulators of SIRT6 [96] or agonists of SIRT3 [94,95] to treat cardiac dysfunctions.

**Author Contributions:** Writing—original draft preparation, E.D.-D., Y.E.M., A.T., F.P. and J.-S.A., funding acquisition, P.A., F.P. and J.-S.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** E.D.D. was supported by grants from region Hauts de France, CPER "Longévité" of Institut Pasteur de Lille and "Fondation Lefoulon Delalande". F.P. and E.D.D. received a grant by the French Government, managed by the National Research Agency (ANR) under the program "Investissements d'avenir" with the reference ANR-16-RHUS-0003. Y.E.M., E.D.D. and F.P. are members of FHU-CARNAVAL. This work was supported by Programme Hubert Curien AMADEUS (France-Autriche), ANR grant BETAPLASTICITY (ANR-17-CE14-0034 to J-S.A.), INSERM, CNRS, Institut Pasteur de Lille (CPER CTRL Melodie), Université de Lille, Fondation pour la Recherche Médicale (EQU202103012732), Conseil Régional Hauts de France and Métropole Européenne de Lille and Société Francophone du Diabète.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data sharing not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

