Epigenetics and precision medicine in cardiovascular patients: from basic concepts to the clinical arena

Abstract

Cardiovascular diseases (CVDs) remain the leading cause of mortality worldwide and also inflict major burdens on morbidity, quality of life, and societal costs. Considering that CVD preventive medications improve vascular outcomes in less than half of patients (often relative risk reductions range from 12% to 20% compared with placebo), precision medicine offers an attractive approach to refine the targeting of CVD medications to responsive individuals in a population and thus allocate resources more wisely and effectively. New tools furnished by advances in basic science and translational medicine could help achieve this goal. This approach could reach beyond the practitioners ‘eyeball’ assessment or venerable markers derived from the physical examination and standard laboratory evaluation. Advances in genetics have identified novel pathways and targets that operate in numerous diseases, paving the way for ‘precision medicine’. Yet the inherited genome determines only part of an individual’s risk profile. Indeed, standard genomic approaches do not take into account the world of regulation of gene expression by modifications of the ‘epi’genome. Epigenetic modifications defined as ‘heritable changes to the genome that do not involve changes in DNA sequence’ have emerged as a new layer of biological regulation in CVD and could advance individualized risk assessment as well as devising and deploying tailored therapies. This review, therefore, aims to acquaint the cardiovascular community with the rapidly advancing and evolving field of epigenetics and its implications in cardiovascular precision medicine.

Customized approaches for the management of cardiovascular disease

Clinicians have always intuitively individualized treatment to match the patient in front of them. But new tools permit far more precise tailoring of therapy beyond the practitioners’ ‘eyeball’ assessment or venerable markers derived from the physical examination and standard laboratory assessment. In early 2016, the then US President Barak Obama launched a Precision Medicine Initiative (https://obamawhitehouse.archives.gov/node/333101). The Mission Statement of this initiative read: ‘To enable a new era of medicine through research, technology, and policies that empower patients, researchers, and providers to work together toward development of individualized care’. This injunction, spurred by the availability of a growing array of novel technologies, incite practitioners to strive for an ever greater level of personalization.

The cardiovascular community takes just pride in the size of the clinical trials that inform medical practice and that dwarf those in many other specialties. Yet a tension prevails between large and ‘dumb’ trials and smaller ‘smarter’ studies. The inclusion criteria of many large-scale trials run countercurrent to the push to personalization. Current standard-of-care background therapy in the cardiovascular community has lowered event rates such that larger numbers or higher risk subjects are needed to see a signal. Thus, the selection of higher risk participants often drives the selection of entry criteria during trial design. These large studies tend to favour a ‘one-size-fits-all’ mentality, even though treated patients often do not match exactly the entry criteria for the randomized clinical trials that current guidelines enshrine.

The field of oncology boasts smaller studies, but much more commonly directed by biomarkers derived from advances in molecular biology and genetics than is the case for cardiovascular trials. The mining of contemporary cancer biology has yielded numerous targeted therapies, whose deployment often depends on companion diagnostic tests that reflect the fundamental mechanisms of the treatment. The successes of this approach have tamed many cancers, converting them into chronic diseases, and even effecting cures in some cases. This example should inspire the cardiovascular community to consider the concepts of development and deployment of therapies guided by biomarkers that narrow rather than broaden the target population. Perhaps it is time to add to the traditional ‘blockbuster drug’ model of broadly applicable therapies, a more targeted and tailored approach. Harnessing new tools furnished by advances in basic science might help in this regard. Advances in genetics have identified novel pathways and targets that operate in numerous diseases, paving the way for ‘precision medicine’. Mendelian randomization approaches can test the causal contribution of various mediators to disease. Yet the inherited genome determines only part of the risk profile and opportunities for personalized therapies that our patients present. Strictly genomic approaches do not take into account the world of regulation of gene expression by ‘epi’genetic changes—acquired modifications to the genome subject to influence by environment—a burgeoning field that adds a new dimension to our understanding of disease and potential new treatments at a growing pace (Figure 1). The application of epigenetics may advance individualized risk assessment and the development and deployment of tailored therapies. This review aims to acquaint the cardiovascular community with the rapidly advancing and evolving field of epigenetics that should contribute to our realizing the promise of ‘precision medicine’.

Figure 1.

Impact of genetics and epigenetics on cardiovascular phenotype. Genetic mutations acquired during the life course represent an irreversible process, whereas plastic epigenetic changes of DNA/histone complexes are reversible and amenable to pharmacological reprogramming.

Lessons from genetics

Considerable efforts have evaluated inborn genetic variation, which can certainly influence disease susceptibility.¹ Pharmacogenetic research has advanced markedly since the exploration of inherited differences in responses to drugs such as isoniazid and succinylcholine in the 1950s. Several clinical examples demonstrate what we have learned from the analysis of our genetic background and how adopting genetic information can apply to cardiovascular precision medicine.

HDLs have shown several potentially beneficial cardiovascular properties affecting reverse cholesterol transport, endothelial function, and inflammation. Clinical results of several therapies that elevate HDL levels have proved disappointing. In contrast, clinical responses to the cholesteryl ester transfer protein (CETP) inhibitor dalcetrapib appears to depend on a genetic variant in the adenylate cyclase type 9 (ADCY9) gene. In retrospective analysis, patients with the AA genotype at polymorphism rs1967309 benefited from a 39% reduction in the composite cardiovascular endpoint with dalcetrapib compared with placebo (n = 5749 patients).² The prevalence of the AA genotype is 17%, except for African Americans in whom it is higher. In contrast, patients with the GG genotype had a 27% increase in events with dalcetrapib vs. placebo, whereas heterozygotes had a neutral response. The lack of detectable genetic effect for rs1967309 in the placebo arm alone and the significant gene-by-treatment arm interaction supported that this was primarily a pharmacogenomic marker of response to therapy. There were strikingly concordant findings with dalcetrapib compared with placebo in terms of atherosclerotic changes on imaging in the dal-Plaque-2 study, inflammatory status per the high-sensitivity C-reactive protein level, and cholesterol efflux.³ The ongoing Dal-GenE randomized clinical trial of 5000 patients with the AA genotype at rs1967309 in the ADCY9 gene will test prospectively the genotype-dependent effects of dalcetrapib on cardiovascular clinical outcomes.⁴

Variants in the genes that encode the beta-1 adrenergic receptor (ADBR1) and, to a lesser extent the beta-2 adrenergic receptor (ADBR2), associate with response to beta-blockers in patients with heart failure and those with hypertension.^5–12 Many studies have focused on the ADBR1 Arg389Gly variant. Because the 389Arg allele associates with greater production of cyclic adenosine monophosphate compared with the 399Gly allele,^11,13 ‘hyper-responders’ carrying the 399Arg allele might benefit to a greater extent from beta-blockers. An important substudy emerged from the BEST trial, which investigated the effect of the beta-blocker bucindolol in patients with heart failure. This substudy of 1040 patients showed that bucindolol reduced mortality and hospitalization in homozygotes for the Arg389 allele compared with the Gly allele carriers.¹⁴ Some studies have reported more equivocal results, including a substudy of 600 patients of MERIT-HF.¹⁵ Given the heterogeneity in pharmacological properties of beta-blockers, these inconsistencies could reflect differences between these agents or the analysis of a smaller sample size in the MERIT-HF substudy.

Hepatocyte uptake of statins depends on the solute carrier organic anion transporter family member 1B1 (SLCO1B1) gene, which encodes the organic anion transporting polypeptide (OATP1B1) protein. Reduced entry of statins in hepatocytes may increase the risk of statin-induced myotoxicity. The rs4149056 polymorphism in the SLCO1B1 gene associates with a greatly increased risk of developing rhabdomyolysis in patients treated with simvastatin.¹⁶ In contrast, SLCO1B1 variants did not associate with statin-induced myotoxicity in patients treated with rosuvastatin.¹⁷

Genetic information has greatly contributed to our understanding of the multifactorial aetiology of type 2 diabetes mellitus (T2D) as well as the individual response to treatment with glucose-lowering agents.¹⁸ Individuals with loss of function variants in CYP2C9, a gene encoding for Cytochrome P450 2C9, display greater glycaemic response to sulfonylureas than those carrying wild-type alleles.¹⁸ Along the same line, heterozygous mutations in hepatocyte nuclear factor 1 homeobox A (HNF1A)—a transcription factor involved in the regulation of several liver-specific genes—strongly associate with extreme sensitivity to sulfonylureas in T2D patients. Furthermore, a genome-wide association (GWA) study in 1024 Scottish patients with T2D unveiled a variant rs11212617 near the ataxia telangiectasia mutated (ATM) gene that associated significantly with glycaemic response to metformin.¹⁸

Taken together, these data indicate that the unveiling of inherited genetic variants can contribute to understanding of clinical drug responsiveness and patient outcome. The application of ‘epi’genetics promises to deepen such genetic insights substantially.

From genomics to post-genomics: the ‘epigenetic revolution’

In 1956, the British developmental biologist, Conrad Waddington, demonstrated the inheritance of a characteristic acquired in a population in response to an environmental stimulus.¹⁹ He found that changes in the environmental temperature or chemical stimuli could induce different thorax and wing structures in embryo fruit flies carrying an identical genetic background.²⁰ The term ‘epigenetics’ originally embraced the process by which a fertilized zygote develops into a mature, complex organism, but underwent expansion as findings showed that cells having the same DNA can exhibit differential modulation of gene activity. In this regard, Waddington’s intuition helped to unveil a key biological mechanism whereby heritable traits can associate not only with changes in nucleotide sequence but also with chemical modifications of DNA or of the proteins with which DNA interacts. The discovery of DNA methylation in bacterial genomes followed Waddington’s observations. Experiments in 1975 showed the transmission of epigenetic changes to daughter cells where it actively regulates gene expression.²¹

Epigenetic modifications fall into three main categories: (i) chemical modifications of DNA (i.e. methylation); (ii) post-translational modifications of histone tails; and (iii) regulation of gene expression by non-coding RNAs [i.e. microRNAs, PIWI-interacting RNAs, endogenous short interfering RNAs, long non-coding RNAs (lncRNAs)] (Figure 2).²² Histones regulation can modify gene expression by altering the accessibility of chromatin to transcription factors. Chromatin comprises chromosomal DNA packaged around histone proteins that form nucleosomes. Multiple interactions tightly link histones to DNA, rendering nucleosomes very stable under physiological conditions.

Figure 2.

Environmental factors and epigenetics. Over time, an array of environmental factors significantly contributes to build our individual epigenetic background that includes DNA methylation changes, post-translational histone modifications and altered expression of non-coding RNAs.

DNA methylation

Methylation of DNA mainly occurs through attachment of methyl group to the C5 position in the cytosine-paired-with-guanine (CpG) dinucleotide sequences. CpG sequences localize in promoter regions rather than coding regions of genes.²³ CpG methylation suppresses gene transcription by directly impeding the binding of transcription factors to DNA or, indirectly, by recognition of methylated sites by chromatin modifying enzymes.²⁴ DNA methyltransferases (DNMTs: DNMT1, DNMT3a, and DNMT3b) mediate much DNA methylation. DNMT1 recognizes hypermethylated DNA thus maintaining methylation status during replication. In contrast, the methyl-writing enzymes DNMT3a and DNMT3b are responsible for de novo methylation.²⁵

Histone modifications

Post-translational modification of histones includes methylation, acetylation, ubiquitination, and phosphorylation. These modifications may cluster in different patterns to regulate the shift from a compact (heterochromatin) to an open (euchrotmatin) chromatin structure, or vice versa.²⁶ Families of chromatin remodellers acetylate and methylate histones: histone acetyltransferases (HATs) and histone deacetylases (HDAC) for acetylation and histone methyltransferases (HMTs) and demethylases (HDMs) for methylation. These histone-writing and -erasing enzymes interact selectively with DNA-methylated regions and thus enable gene repression or transcription.²⁷ Histone acetylation by HATs generally associates with enhanced gene expression, whereas HDACs exert opposite effects. SIRT1, the most studied mammalian Sirtuin, represents a clear example of how histone modifications impact cellular function by orchestrating key biological processes (i.e. metabolism, longevity) through the deacetylation of a number of enzymes and transcriptional switchers (i.e. PGC-1α, NF-κB, FOXO) as well as histones (i.e. H3K9 and H3K56).²⁸ The more complex process of histone methylation may result in different chromatin states according to the methylated residue and the number of added methyl groups. However, we can now associate specific methylation patterns to activation or silencing of specific genes. For example, mono-methylation of lysine 4 on histone 3 (H3K4me) licences a key pro-inflammatory complex, nuclear factor kappa-B (NF-κB).²⁹

Non-coding RNAs

Recent research has demonstrated a key role for the non-coding genome in genetic programming and gene regulation during development as well as in health and in CVD.³⁰ About 98% of the human genome does not encode proteins but can engage in transcription producing numerous non-coding RNAs (ncRNAs) that exert important regulatory and structural functions.³¹ Based on size, ncRNAs can be subdivided into 2 major groups: (i) small ncRNAs (sncRNAs, <200 nucleotides long) including microRNAs, PIWI-interacting RNAs, and endogenous short interfering RNAs and (ii) long non-coding RNAs (lncRNAs), which have a length between 0.2 kb and 2 kb. A growing body of evidence implicates ncRNAs in the pathogenesis of CVD and as biomarkers of cardiovascular damage.³¹ The high inter-individual diversity of chromatin architecture and the ncRNA landscape points to considering epigenetic modifications as tools to individual cardiovascular risk and customize treatments.

The epigenetic landscape in cardiovascular disease

The acquisition of epigenetic signals during an individual’s lifespan results mostly from environmental factors, including the intrauterine milieu, diet, atmospheric pollutants, smoking, urban noise, and, last but not least, the social, cultural, and economic circumstances encountered (Figure 2).²³ Most epigenetic modifications show stability and appear durable and thus able to affecting gene expression along the arc of ageing.^32,33 Modifications of the epigenetic landscape by environmental cues may perturb cardiovascular homeostasis and influence endothelial dysfunction, vascular ageing, or cardiomyocyte behavior.³⁴ Such modifications may also influence cardiovascular risk factors including dyslipidaemia, hypertension, obesity, and diabetes.

Over the last decades, the availability of relatively inexpensive techniques for genome-scale analysis of both DNA methylation and histone modifications led to an explosion of studies characterizing the impact of the epigenetic variations on CVD. A key contribution to our current understanding of cardiovascular epigenetics comes from gain- and loss-of-function experiments in animals. Genetic disruption of DNMTs (that establish or replicate DNA methylation) or MTHFR (related to methyl donor generation) in mice associate with DNA hypomethylation and subsequent increases in inflammatory mediators and formation of aortic fatty streaks.³⁵ Atherosclerosis-prone apoE^−/− mice develop specific changes in DNA methylation of transcribed gene sequences both in peripheral blood leucocytes and in the aorta before developing vascular lesions.³⁶ Along the same line, peripheral blood mononuclear cells (PBMCs) isolated form patients with atherosclerosis exhibit globally reduced DNA methylation.³⁷ Recent observations implicate somatic mutations in demethylases such as Tet2 that accrue with age in spurring the emergence of leucocyte clones that not only confer increased risk of haematological malignancies but also of CVDs.^38,39 Indeed, myocardial infarction and stroke account for much more of the increased mortality associated with these haematopoietic clones than do leukaemias.³⁹ Experimental work has established the causality of these methylase mutations in promoting plaque inflammation, probably due to epigenetic regulation of cytokine and chemokine genes.^39,40

Over the last few years, an array of studies has begun to link cardiovascular risk factors (i.e. ageing, hypertension, diabetes, and dyslipidaemia) to epigenetic modifications in human subjects. The discussion below provides early examples of such associations. Although their validation will require large prospective studies, the examples cited below provide intriguing glimpses of the potential of epigenetics to broaden our understanding of disease pathogenesis and new clinical tools as well.

Ageing

Alteration of epigenetic patterns in ageing—a phenomenon known as ‘epigenetic drift’—involves primarily a progressive decrease in global DNA methylation. The Normative Ageing Study showed a longitudinal decline in the average blood genomic DNA methylation of repetitive sequences, such as Alu and LINE-1, over 8 years of follow-up.⁴¹ Genome-wide studies in aged stem cells revealed a decrease in DNA methylation at the promoter of genes associated with self-renewal, whereas promoters of genes regulating differentiation were hypermethylated.⁴² Histone modifications such as acetylation of histone 3 at lysine 9 (H3K9Ac) as well as trimethylation of histone 3 at lysine9 (H3K9me3) and lysine 27 (H3K27me3) decrease with age and contribute to haematopoietic stem cells dysfunction and defective vascular repair.⁴³ The miRNA landscape in ageing tends towards increased miRNAs expression, leading to post-transcriptional suppression of many target genes and alterations in endothelial functions. Derailed expression of miR-29, miR-34a, miR-217, and miR-146 associates strongly with a decline of vascular and cardiac function in the elderly.^44,45 Age-dependent decreases in several lncRNAs, namely ANRIL, MALAT-1, MIAT, and Meg3, may lessen angiogenic potential.³¹

Hypertension

Peripheral blood leucocytes of patients with essential hypertension show a loss of global genomic methylation.⁴⁶ A recent genome-wide association and replication study of blood pressure phenotypes among 320 251 individuals of East Asian, European, and South Asian ancestry revealed that single-nucleotide polymorphisms influencing blood pressure associate strongly with methylation at multiple local CpG.⁴⁷ Another genome-wide methylation study on essential hypertension conducted in young African American males found increased methylation levels at two CpG sites in the SULF-1 gene involved in apoptosis and decreased methylation of the PRCP gene that regulates cleavage of angiotensin II and III.⁴⁸ Histone 3 methylation also controls the expression of genes related to hypertension. A polymorphism in DOT1L gene, encoding histone H3K79 methyltransferase, associated tightly with greater systolic and diastolic blood pressure response to hydrochlorothiazide in Caucasians.⁴⁹ Hypertensive patients have altered expression of miR-9 and miR-126 relative to healthy individuals in association with prognostic indices of hypertensive target-organ damage.⁵⁰ These miRs may serve as epigenetic biomarkers and therapeutic targets in patients with essential hypertension.

Type 2 diabetes

The involvement of epigenetics in T2D could help explain the long-lasting detrimental effects of hyperglycaemia despite optimal glycaemic control, the so-called ‘metabolic’ or ‘hyperglycaemic memory’. Several studies have shown that epigenetic signatures at the promoter of pro-oxidant and inflammatory genes contribute to persistent endothelial dysfunction, diabetic nephropathy, and retinopathy as well as atherosclerotic features even after restoration of normoglycaemia. ^27,51,52 Promoter methylation of the adaptor p66^Shc, involved in mitochondrial oxidative stress, is persistently reduced in PBMCs from T2D patients after glycaemic control and correlates with endothelial dysfunction and oxidative stress levels.⁵³ Saliva DNA from T2D patients with end-stage renal disease (ESRD) compared to that from patients with chronic kidney disease who did not progress to ESRD exhibits several differentially methylated genes.⁵⁴ We have recently reported that T2D patients carry a specific epigenetic pattern on histone 3, namely H3K4me.⁵⁴ This signature—driven by the methyltrasferase Set7—associates with NF-κB activation and subsequent overexpression of pro-atherosclerotic genes such as iNOS, COX-2, ICAM-1, MCP-1, and VCAM-1.⁵⁵ The histone 3 deacetylase SIRT1 declines in peripheral blood monocytes from patients with insulin resistance and coronary artery disease as well as in aged individuals, thus representing a key epigenetic route linking longevity, metabolism, and atherosclerosis.⁵⁶ Plasma miRNA profiling in a cohort of diabetic patients unveiled a profound down-regulation of miR-126.⁵⁷ Moreover, EPCs isolated from diabetics show reduced miR-126 expression, and exogenous miRNA-126 mimic restored EPCs angiogenic properties.^58,59

Dyslipidaemia

Blood lipid profiles reflect both genetic and environmental factors.⁶⁰ Maternal and paternal dietary habits as well as the influence of intrauterine environment may influence the dietary habits and cholesterol levels of the offspring.⁶¹ The inheritance of specific epigenetic modifications on the promoter of genes that regulate glucose and lipid metabolism provides a potentially fascinating contributor to atherogenesis. Participants in the Dutch Hunger Winter Families Study who were exposed in utero to the 1944–45 famine, a condition associated with the development of obesity and insulin resistance in adulthood, displayed subtle blood methylation changes of insulin-like growth factor-2 (IGF-2) and leptin (Lep) genes compared with unexposed siblings.⁶² Heritable epigenomic changes may also contribute to the unexplained inter-individual postprandial lipaemia (PPL) variability, an independent risk factor for CVD. An epigenome-wide association study on 979 subjects challenged with a high-fat meal revealed that eight methylation sites encompassing five genes (LPP, CPT1A, APOA5, SREBF1, and ABCG1) associated significantly with PPL. Higher methylation at LPP, APOA5, SREBF1, and ABCG1 and lower methylation at CPT1A correlated with an increased TG–PPL response.⁶³ In another study from the same cohort, CPT1A methylation associated robustly with fasting very-low LDL-cholesterol and TG.⁶⁴ Finally, miR-33a/b may act as post-transcriptional regulators of lipid metabolism and insulin signalling, and their inhibition diminishes atherosclerosis by raising plasma HDL levels.⁶⁵

Myocardial infarction

Epigenetics may help to understand the considerable variability in mid- and long-term prognosis in patients who sustain myocardial infarction (MI). In the northern Sweden population health study, individuals with a history of MI showed differential DNA methylation at 211 CpG-sites representing genes related to cardiac function, CVD, cardiogenesis, and recovery after ischaemic injury.⁶⁶ Hence, epigenetic information may explain the alterations in cardiovascular gene expression trajectories and offer biomarkers for the follow-up of these patients. MI also associates with a profound deregulation of circulating miRNAs and lncRNAs. MiRNA-208b and miR-499 are highly increased in MI patients (>105-fold, P < 0.001), whereas they are nearly undetectable in healthy controls.⁶⁷ Elevation of plasma miRNAs observed during the 1st hour after the onset of symptoms indicates that these alterations might participate causally in the ischaemic event rather than representing a mere reaction to acute ischaemia. Patients presented <3 h after onset of pain showed positive miR-499 in 93% of patients and hs-cTnT in 88% of patients. Overall, miR-499 and hs-cTnT provided comparable diagnostic value with areas under the receiver operating characteristics curves of 0.97.⁶⁷ Consistently, a variety of studies showed that selected panels of circulating miRNAs may represent useful prognostic biomarkers in MI patients.^68,69 Circulating microRNAs may also permit differentiating Takotsubo cardiomyopathy from acute MI.⁷⁰ The expression of lncRNAs (ANRIL, KCNQ1OT1, MIAT, and MALAT-1) also rises significantly in PBMCs from patients with MI and independently predicts left ventricular dysfunction (ejection fraction ≤ 40%) at 4 months of follow-up.⁷¹

Stroke

Lower methylation of the TRAF3 gene—a member of the TNF receptor-associated factor (TRAF) protein family—correlated with increased platelet aggregation in stroke patients treated with clopidogrel (ρ = −0.29, P = 0.0075) as well as recurrence of ischaemic events.⁷² Another study from the same group showed that methylation of PPM1A gene, a member of the PP2C family of Ser/Thr protein phosphatase, associates with recurrent vascular events in aspirin-treated patients.⁷³ These studies provide proof of principle that—beyond genetic information—pharmacoepigenomics may enable precision approaches in cardiovascular medicine. Patients with acute ischaemic stroke show consistently altered circulating microRNAs, namely miR-125 b-2*, -27a*, -422a, -488, and -627, irrespective of age, disease severity, or confounding metabolic complications.⁷⁴ Profiling of lncRNAs in 266 whole-blood RNA samples from patients with ischaemic stroke and matched controls revealed a significant gender-sensitive dysregulation. Males with stroke compared with the controls showed 299 differentially expressed lncRNAs, whereas females had only 97 differentially expressed lncRNAs. Of note, some differentially expressed lncRNAs mapped close to genomic locations of previously identified stroke-risk genes, including lipoprotein(a)-like 2, ABO blood group, prostaglandin 12 synthase, and α-adducins.⁷⁵ The latter study furnishes hints for gender-specific epigenetic biomarkers in the setting of CVD.

Heart failure

Genome-wide maps of DNA methylation and histone-3 lysine-36 trimethylation (H3K36me3) in cardiomyopathic and normal human hearts showed a variety of epigenomic patterns in important DNA elements of the cardiac genome in patients with advanced cardiomyopathy.⁷⁶

Interrogation of the cardiac methylome in patients with idiopathic dilated cardiomyopathy detected methylation differences in pathways related to heart disease, but also in genes with yet unknown function in heart failure, namely lymphocyte antigen 75 (LY75) and adenosine receptor A2A (ADORA2A).⁷⁷ Advanced heart failure patients have low circulating levels of miRNA-103, miRNA-142-3p, miRNA-30 b, miRNA-342-3p, and miRNA-652-3p and high levels of miRNA-499 and miRNA-508-5p.^78,79 Moreover, in such patients, low circulating levels of miRNA-423-5p portend poor long-term outcomes.⁸⁰

Towards individual epigenetic maps

Contrary to the common belief that an individual’s genome is no different than another, scientists have known for years that not all epigenomes are born equally. When it comes to epigenetic modifications—diversity is the norm rather than the exception. Constructed by way of chemical information specific determinants are codified for health and changed in disease. With this knowledge, precision medicine is evolving because of the significant advances enabling scientists and physicians fundamental and medical insights.⁸¹ The completion of the Human Genome Project furnished the cornerstone of the genetic blueprint and the mainstay of genomics research.^82,83 Within a decade of announcing the draft sequence, scientists had transitioned from understanding the structure of DNA to major accomplishments in genome biology. Next-generation sequencing technologies rapidly emerged bringing with it unprecedented opportunities to address the unmet need to understand genes in context with biology and disease.⁸⁴ Critical advances in bioinformatics accompanied those in genomics permitting processing and analysis of ever expanding data sets.⁸⁵ Despite these strides, the epigenomic world remains largely uncharted about precision medicine in cardiology. Although there remain substantial barriers to overcome, efforts have begun to accelerate the development cardiovascular epigenomics.⁸⁶

Although technological advances in epigenomics have begun to enable clinical applications, a key challenge remains how to cost-effectively perform whole-genome bisulfite sequencing (WGBS) for epigenome-wide association studies (EWAS, Table 1, Figure 3).

Table 1.

Abbreviations list and description of main epigenetic technologies and studies

Acronym	Definition	Description
ATAC-Seq	Assay for transposase-accessible chromatin with high-throughput sequencing	ATAC-seq is a high-throughput sequencing technology that furnishes a genome-wide map of chromatin accessibility. Specifically, it allows simultaneous, genome-wide information on the positions of: (i) open chromatin, (ii) transcription factor binding, (iii) nucleosomes in regulatory regions, and (iv) information on chromatin state annotation.
WGBS	Whole-genome bisulfite sequencing	WGBS is a next-generation sequencing technology used to determine the DNA methylation status of single cytosines. For WGBS, DNA samples have to be treated with sodium bisulfite, a chemical compound that converts unmethylated cytosines into uracil but leaves 5-methylcytosine residues unaffected. This technique measures single-cytosine methylation levels genome-wide and directly estimates the ratio of methylated molecules.
RNA-seq	RNA sequencing	RNA-seq is used to explore the cellular transcriptome. It allows the assessment of alternative gene spliced transcripts, post-transcriptional modifications, gene fusion, mutations/single nucleotide polymorphisms and dynamic changes in gene expression, or differences in gene expression in different groups or treatments. In addition to mRNA transcripts, RNA-seq can look at different populations of RNA including total RNA, small RNA, such as miRNA, tRNA, and ribosomal profiling.
ChIP-on-chip	ChIP-on-chip	ChIP-on-chip is a technology that combines chromatin immunoprecipitation (ChIP) with DNA microarray (chip). ChIP-on-chip is used to investigate interactions between proteins and DNA in vivo.
COMET	COMET	The COMET package is computational tool that allows to visualize EWAS results in a genomic region of interest. COMET provides a plot of the EWAS association signal and visualisation of the methylation correlation between CpG sites (co-methylation).
EWAS	Epigenome-wide association studies	EWAS are designed to investigate the association between a genome-wide set of quantifiable epigenetic marks (i.e. DNA methylation) and a particular identifiable phenotype/trait in large human cohorts. These sudies may help to elucidate whether an epigenetic perturbation is associated causally or consequentially with a given phenotype.
IHEC	International Human Epigenome Consortium	IHEC is a scientific organization, founded in 2010, which helps to co-ordinate global efforts in the field of epigenomics. IHEC member organizations are engaged in efforts to generate at least 1000 reference (baseline) human epigenomes from different types of normal and disease-related human cell types.
GTEx	Genotype-Tissue Expression project	This program is providing valuable insights into the mechanisms of gene regulation by studying human gene expression and regulation in multiple tissues from healthy individuals as well as in a variety of human diseases.

Figure 3.

Epigenetic-based technologies with potential application in cardiovascular patients. PBMCs, peripheral blood mononuclear cells; WGBS, whole-genome bisulfite sequencing; MSREs, methylation sensitive restriction enzymes; HPLC, high-performance liquid chromatography; ChIP, chromatin immunoprecipitation; miRNAs, microRNAs; RNA-seq, RNA sequencing.

The high cost of constructing reference methylomes encountered by the International Human Epigenome Consortium (IHEC) has represented a challenge.⁸⁷ The standard sequencing coverage misses many differential methylation sites. A technical advance in the quest to recover methylation sites from low-coverage WGBS uses an algorithm that captures almost twice the amount of informative bisulfite sequencing data.⁸⁸ Saturation analysis of WGBS that constructs reference methylomes assesses multiple features involving blocks of comethylation (COMET) and improves detection from existing data, through the development of computational tools such as COMETgazer and COMETvintage. Indeed, the improved identification and recovery of methylation sites combined with advances in computational analyses advance the goal of translating EWAS into clinical practice. Despite the development of WGBS and new informatic tools considerable challenges remain. The validation of informative differential methylation sites remains incomplete due to limitations in assay performance and standardization. To help close this gap, a multicentre study organized by the BLUEPRINT Project assessed the accuracy and robustness of methylation assays tailored for cost-effective, high-throughput biomarker development designed for large population studies.⁸⁹ This extraordinarily comprehensive resource identified strong concordance in assays based on amplicon-based bisulfite sequencing as well as bisulfite pyrosequencing. The reliable identification of epigenetic determinants for clinical applications could sharpen precision medicine.⁹⁰

The technological accomplishments made in recent years to construct reference methylomes requires dismantling sequences before genome reconstruction or ‘shaping’ as it pertains to chemical epigenetic information. Personal epi(genomics) is emerging from major projects that provide significant resources like those already described and the Genotype-Tissue Expression (GTEx) project.⁹¹ Mapping our epigenomic differences with technologies that catalogue sequence similarity is an important move forward for personal therapeutic design. Intense interest surrounds the clinical development of pharmacological compounds that chemically modify key epigenetic determinants. Mapping a new route borrowed from the field of oncology, histone deacetylase inhibitors show therapeutic potential in heart disease.⁹² However, our view of the regulatory determinants that control transcription are largely derived from cell experiments and small animal studies that draw mechanistic parallels but highlight important differences. Indeed, while fundamental discoveries can chart a course to understanding pathology, human data are still necessary to answer the remaining critical questions about cardiac disease. The existing gap for epigenomics is person-centred cardiovascular information⁹³ and the unmet need to distinguish gender-specific epigenetic differences.⁹⁴ In an emerging era of epi(genomic) medicine, the potential to inform on evidence-based responses can lead to significant advances in personalized cardiovascular care. For example, epigenetic variability exists between genders and this is evident when assessing response to pharmacological HDAC inhibitors.⁹⁵

Epigenetic therapies in the cardiovascular patient

The recent advances in epigenetic tools and the movement towards big data to dissect patient heterogeneity also raises the real need to assess distinct tissue types and specific cell lineages. This consideration has become more evident in the cardiac regeneration field. The highly specialized transcriptional programme during cardiomyocyte maturation remains poorly understood because of cellular heterogeneity. Single-cell RNA sequencing (RNA-seq) and unbiased clustering were recently used to characterize cardiomyocytes, fibroblasts, and endothelial cells from different regions of the heart throughout development.⁹⁶

The limited and modest gain in cardiac function observed in randomized clinical trials of cell therapy for myocardial ischaemia indicates the need for better fundamental understanding of the cells being transferred and the environment which receives them. Disease-mediated epigenetic alteration of functional genes in transferred cells might contribute to the disappointing functional outcomes in clinical trials of myocardial cell therapy. Epigenetic silencing of gene expression through chromatin remodelling depends largely on chromatin condensation resulting from DNA methylation of CpG residues by DNMTs and post-translational modifications in the histone tails including specific alterations in acetylation and methylation of specific lysine residues. Generally, DNA methylation of gene regulatory regions along with H3K9 deacetylation by HDACs alters chromatin structure and limits gene transcription.⁹⁷ Such repressive epigenetic modifications in autologous adult stem cells might compromise both their differentiation and functional activities and contribute to the general lack of substantive therapeutic benefit of cellular therapies. Indeed epigenetic silencing of genes via aberrant DNA methylation can occur in diabetes, hypertension, obesity, and heart failure.⁹⁸ Since many patients with heart disease also have such co-morbidities, epigenetic silencing of functional genes in their autologous stem cells may impair their ability to effect myocardial repair. Epigenetic reprogramming to modulate gene repressive epigenetic marks by the use of small molecules might mitigate such limitations to the functions of adult stem cells in ischaemic myocardial repair (Figure 4).

Figure 4.

Epigenetics in cardiovascular precision medicine. Unveiling the individual epigenetic landscape provides an important snapshot of the epigenetic machinery that can be eventually employed to customize diagnostic and therapeutic approaches in primary and secondary prevention of cardiovascular disease. Available technologies for the study of the epigenome may furnish detailed epigenetic maps based on DNA–histone interactions and non-coding RNAs landscape. Individual epigenetic maps could represent a novel tool in the clinical practice to stratify cardiovascular risk beyond traditional or genome-based risk calculators. Epigenetic information also helps in deciphering inter- and intra-personal variation in individual drug response. Finally, adverse epigenetic patterns are amenable to pharmacological reprogramming of chromatin modifying drugs or non-coding RNAs.

Chemical modifiers of DNA demethylation (5-azacytidine) and histone acetylation (e.g. Trichostatin A, valproic acid) can induce multipotency by enhanced reprogramming of somatic cells for somatic cell nuclear transfer (SCNT) or induced pluripotent stem cell (iPSc) derivation.^99–101 Both drugs can change the fate of a given cell by chromatin remodelling, thus restoring the transcription of silenced genes including those that might reactivate pluripotency in somatic cells.^101–103 Specifically in the cardiovascular system, treatment with 5-azacytidine increases cardiomyocyte differentiation of mesenchymal stem cells (MSCs) yielding improved cardiac function after transplantation of 5-azacytidine–treated MSCs compared with control MSCs.^104–107 Similarly, inhibition of both DNA methylation and histone deacetylation by combined treatment of 5-azacytidine and valproic acid can enhance both the plasticity and function of endothelial progenitor cells and transplantation of treated cells boosting myocardial repair.¹⁰⁸ Multiple studies have reported similar strategies of epigenetic modification of various kinds of stem/progenitor cells using inhibitors of HDACs.^109–113 Specific inhibition of Class I HDACs by mocetinostat, augmented cardiac c-kit progenitor cell function including increased cardiomyogenic gene transcription and improved functional repair of ischaemic myocardium upon transplantation.¹¹⁴ Similarly, loss of HDAC4 function using interfering RNA improved functions of cardiac stem cells.¹¹⁵ Furthermore, inhibition of HDAC also enhanced differentiation of induced pluripotent stem cells into cardiomyocytes.¹¹⁶

Thus, ample evidence in the literature suggests that in vitro epigenetic reprogramming of a variety of stem progenitor cells can improve their functions including their ability to augment post-infarction myocardial function. Although not yet tried in the setting of clinical cell therapy, this approach of using small molecule modifier of DNA methylation and HDAC inhibition might enhance the clinical benefits of cell therapies for repair of the myocardium as well as other ischaemic tissues.

Conclusions

Environmental factors potently influence epigenetic variations and altered gene expression over the life course. Cigarette smoking, pollution, and high-fat diets all contribute to altered chromatin architecture, DNA methylation, as well as circulating and tissue levels of non-coding RNAs. Over the last decade, the launching of precision medicine initiatives aims to tailor health care on the basis of a person’s genes, lifestyle, and environment.¹¹⁷ Such personalized approaches usually employ panomics analyses to unmask the cause of individual patient’s disease and to develop specific therapeutic strategies. In this regard, patient-level ‘epigenomic’ information may contribute to sharpen risk assessments beyond traditional calculators and direct the use of chromatin-modifying drugs to restore the expression of salutary genes in endothelial cells, cardiomyocytes, and bone marrow-derived angiogenic cells (Figure 4).

Epigenetic marks in cardiovascular precision medicine should not be considered as self-standing layers of biological regulation but should be always interpreted in light of the individual genetic background. A growing body of evidence suggests the existence of a dynamic cross-talk between genetics and epigenetics. DNA methylation changes connect directly with genetically regulated gene expression variation, and, in turn, genetic variation significantly impacts on epigenetic routes and chromatin architecture.^118,119 Hence, future personalized approaches should arise from a full integration of genetic and epigenetic maps in an attempt to build a multidimensional view of our genome.

The causality of presumed epigenetic events and cardiovascular phenotype requires careful consideration. Indeed, epigenetic markers, like any other molecular marker, are vulnerable to confounding and reverse causation. In this respect, framework of Mendelian randomization—a process that interrogates the causal relationships between exposure, epigenetic marks, and outcome—can serve to establish meaningful hierarchies, thus helping to discriminate between epigenetic phenomena and epiphenomena.^120,121 The results of large epigenomic studies over the upcoming years will help to decipher the complex link between genetics, epigenetics, and CVD and define and validate the added value of epigenetic information into personalized cardiovascular therapies.