Next generation sequencing applications for cardiovascular disease

Abstract

The Human Genome Project (HGP), as the primary sequencing of the human genome, lasted more than one decade to be completed using the traditional Sanger’s method. At present, next-generation sequencing (NGS) technology could provide the genome sequence data in hours. NGS has also decreased the expense of sequencing; therefore, nowadays it is possible to carry out both whole-genome (WGS) and whole-exome sequencing (WES) for the variations detection in patients with rare genetic diseases as well as complex disorders such as common cardiovascular diseases (CVDs). Finding new variants may contribute to establishing a risk profile for the pathology process of diseases. Here, recent applications of NGS in cardiovascular medicine are discussed; both Mendelian disorders of the cardiovascular system and complex genetic CVDs including inherited cardiomyopathy, channelopathies, stroke, coronary artery disease (CAD) and are considered. We also state some future use of NGS in clinical practice for increasing our information about the CVDs genetics and the limitations of this new technology.

Key messages

• Traditional Sanger’s method was the mainstay for Human Genome Project (HGP); Sanger sequencing has high fidelity but is slow and costly as compared to next generation methods.

• Within cardiovascular medicine, NGS has been shown to be successful in identifying novel causative mutations and in the diagnosis of Mendelian diseases which are caused by a single variant in a single gene.

• NGS has provided the opportunity to perform parallel analysis of a great number of genes in an unbiased approach (i.e. without knowing the underlying biological mechanism) which probably contribute to advance our knowledge regarding the pathology of complex diseases such as CVD.

Keywords:

• Cardiomyopathies
• cardiovascular genetics
• channelopathies
• complex disease
• coronary artery disease
• next generation sequencing
• stroke

1. Introduction

Decoding the human genome is required to discover the role of genetic information for determining the development, structure and function of the human body, i.e. the genetic changes could affect the human health, behaviour, personality and etc. The Human Genome Project (HGP) [1] has offered the first returns on the promise of genomic medicine and it has been used as a reference tool to interrogate the variations found among individual genomes. Sanger sequencing has high fidelity but is slow and costly as compared to next generation methods. As a result, unsurprisingly, HGP cost $3 billion, and lasted more than one decade to sequence the first human genome [2,3]; while NGS technologies are approaching us to routinely analyze whole genome sequence (WGS) for just a $1000 per patient [4–7]. NGS tool may now be applied either targeted, i.e. disease specific gene panels and WES or non-targeted, i.e. WGS.

In addition to the availability of genomic information, HGP led to an acceleration in scientific progress and technological advances [8]. With advert of NGS, we could sequence the human genome to find the variations and to explore how their variations affect our life. NGS, thus, helps us in understanding effective ways to improve human health, and in paving the way for “personalized” medicine [9,10]. Most recently, the concept of “personalized” or “precision” medicine has received much attention [11], though some experts argue that “whole-genome testing hurts more than it helps” [9,12].

NGS has been shown to be successful in identifying novel causative mutations of single gene disorders and common diseases of cardiovascular system [13]. Genetic testing is an essential step for the disease management and genetic counselling for finding at risk asymptomatic family members. Cardiovascular Mendelian diseases include familial hypercholesterolemia, cardiomyopathies, primary arrhythmias, congenital heart disease (CHD), thoracic aortic aneurysms and dissections (TAAD). Moreover, NGS is being increasingly important in common cardiovascular diseases (CVDs) because, unlike genome wide association studies (GWASs) that provide only known single nucleotide polymorphism (SNP) data, it can provide more data including common and rare variants, indels (insertions and deletions) and copy-number variations (CNVs) [14].

Here, we summarize the current findings from the recent studies of NGS on Mendelian and in complex genetic CVDs, such as inherited cardiomyopathies, channelopathies as well as coronary artery disease (CAD) and stroke. We also discuss the challenges of NGS in clinical practice and its future perspectives in improving our understanding of the genetics of CVDs.

2. Sequencing technologies

First-, second- and third-generation sequencing technologies have been reviewed elsewhere [15–19]. In this review, we briefly describe traditional Sanger sequencing (first-generation) and compare it with high-throughput sequencing technologies.

2.1. Conventional Sanger’s method

The “chain-termination” or dideoxy technique, a major breakthrough in the progress of DNA sequencing technology, was introduced by Sanger in 1977 [20]. This method is the “gold standard” for clinical research sequencing with an accuracy of 99.99% and it is still used to validate NGS data. However, the disadvantages are low throughput and high costs per sample. For Sanger technique, automated DNA sequencing machines use fluorescent instead of radioactive labels (permitting the reaction to occur in single vessel) and capillary based electrophoresis instead of gel electrophoresis (permitting the separation to occur in single lane); however, they still are cloning and sequencing each DNA fragment in one at a time (low throughput).

2.2. Next-generation sequencing methods

NGS sequencing is a term used to describe a variety of massively parallel DNA sequencing technologies which allowing for simultaneous sequencing of a lot DNA and RNA molecules. As shown in Figure 1, despite differences in technical details, all the NGS instruments share common features. For example, they all need processes of library preparation, sequencing and imaging and data analysis. Details of NGS sequencing chemistry have been discussed elsewhere [15,21–24]. Commonly used DNA sequencing methods in NGS systems have been compared in Table 1.

Figure 1. Overview of the primary steps in new generation sequencing (NGS) workflow. All NGS systems show some common features. The genomic DNA is randomly fragmented and a library is prepared to enable massively parallel sequencing. The individual library fragments are either clonally amplified by emulsion PCR (Roche and Life Technologies) or by solid surface bridge amplification (Illumina), or not amplified (Third-generations PacBio and MinION). In third-generation DNA sequencing, a single DNA molecule is sequenced without the need for amplification. Flow cell sequencing of templates creates fluorescent, luminescent, or proton signals that in turn either generates detectable images or are detected by pH detectors. Then, these obtained signals will be processed into sequence reads that assembled and aligned subsequently by using special bioinformatics analyses.

Table 1. DNA Sequencing methods used by NGS systems.

	Pyrosequencing	Sequencing by synthesis	Sequencing by ligation	Ion semiconductor sequencing	SMRT sequencing	Nanopore Sequencing
Read length	700 bp	50–600 bp	35–60 bp	Up to 400 bp	10–15 kb maximum >40 kb	5.4 kb average (Up to 300 kb reported)
Reads per run	1 million	200–500 million Up to 2.5–3 billion	1.2–1.4 billion	Up to 80 million	500–1000 mega	4.4 million
Time per run	24 h	1–11 days	1–2 weeks	2 h	0.5–4 h	0.17–48 h
Accuracy^a	99.99%	98% 99.9% (Phred30)	99.9%	98%	87%	∼90% (up to 99% consensus)
Cost per one million bases	$10	$0.05–$0.15	$0.13	$1	$0.13–$0.60	$0.11–$0.50
Advantages	High throughput Large read length generation	Overcomes homopolymer issue due to terminated nucleotides Extremely efficient	Low cost per base	Time-efficient Cost-effective	Single molecule templates Lower cost per base Easy sample preparation Faster runtime Simplified primary data analysis	Very long reads Affordable equipment (MinION starter kit is only 1000 USD) Portable (Palm sized)
Disadvantages	Homopolymer errors Signal amplification step is needed High reagent cost Expensive as compared to its competitors	Need for signal amplification step increase in error rate of machine with increasing read length Equipment can be very expensive Requires high concentrations of DNA	Slower than other competitors Issues in sequencing palindromic sequences	Homopolymer errors Still a “wash-and-scan” system Requiring PCR amplification Overall read length is limited	Moderate throughput. Equipment can be very expensive	Lower throughput than other competitors Only 90% single read accuracy
Platform example	454 GS FLX (Roche)	HiSeq 2000 (Illumina)	SOLiDv4 (Applied Biosystems^b)	Ion torrent (Life Technologies)	PacBio (Pacific Biosciences)	MinION (Oxford Nanopore)

NGS: next-generation sequencing; SMRT: Single-molecule real-time.

^a Single read not consensus.

^b Was later acquired by Life Technologies, now part of Thermo Fisher Scientific.

3. NGS applications in cardiovascular disorders

Most CVDs have a clear genetic component in their aetiologies. Apart from very rare Mendelian Cardiovascular cases with strictly monogenic dominant inheritance, a large number of patients are considered to have polygenic/multifactorial forms in which two or more genetic defects located in the same or different genes as well as environmental factors cause the pathology. Also, symptoms of a monogenic disorder can be modified by common polymorphisms leading to change in prognosis [13]. Given this complexity, genetic molecular diagnostic tools with comprehensive approach, are needed.

NGS has provided the opportunity to perform parallel analysis of a great number of genes which probably contribute to advance our knowledge regarding the pathology of complex diseases such as CVD. It may also be a useful method in identifying rare variants in smaller families. Examples of Mendelian disorders in cardiovascular medicine include familial hypercholesterolemia, hypertrophic and familial dilated cardiomyopathies and channelopathies (i.e. Brugada and long QT syndrome) [25] while the most common CVDs in clinical practice are complex traits such as CAD and stroke that are resulted from complicated gene–gene and gene–environmental interactions [26]. Molecular genetic testing has not only been used as a research tool, but also has recently entered into clinical diagnostic settings due to its potential to provide more individualized and informative counselling for families [27].

3.1. Familial hypercholesterolemia

Familial hypercholesterolemia (FH) is a common hereditary lipoprotein disorder caused by defects in the hepatic uptake and degradation of low-density lipoprotein cholesterol (LDL-C) via the LDL-receptor pathway, resulting in severely elevated LDL-C levels from birth, enhanced atherosclerosis progression and premature atherosclerotic CVD [28].

Early diagnosis and therapy with lipid-lowering medications significantly reduces the risk for such events [29]; therefore, early identification of patients at risk is the mainstay of prevention for atherosclerotic CVD, primarily CAD [30]. FH, an autosomal dominant disease, is occurring in 1:500 people in most countries (the prevalence is thought to be between 1/500 and 1/200); homozygous FH is much rarer, occurring in 1 in a million births [31,32]. However, regarding the major cause of death in worldwide is CAV, many FH individuals may simply be neglected among the large number of any CAD individuals that caused by common risk factors [33]; thus, these two prevalence rates likely represent underestimates because many affected individuals are not diagnosed in most countries [34]. Autosomal dominant FH is commonly caused by mutations in the LDL-receptor gene (LDLR), and the genes encoding apolipoprotein B (APOB) and proprotein convertase subtilisin/kexin type 9 (PCSK9). According to various reports, the yield of genetic testing including these three genes markedly varies from 28% to 88% [35]. The FH Causative mutations may also exist in numerous secondary genes such as APOE, ARH and LIPA. In two independent studies using WES, the APOE gene, previously shown to be associated with LDL-C levels [36], has been found to be implicated in autosomal dominant FH [37,38]. FH with autosomal recessive inheritance of is caused by mutations in LDLRAP1 [39].

NGS advances are changing our perception of heterozygous FH about prevalence assessment and appropriation of polygenic effects [40]. Recently, Wang et al. [40] found that monogenic FH-causing mutations detected by targeted NGS were present in near half of the individuals with severe hypercholesterolemia, defined as LDL-C l > 5.0 mmol/L. When copy number variations and individuals with extreme polygenic scores were included, the percentage of individuals with monogenic mutations increased to 53.7 and 67.1%. In patients that are suspected of FH, the Consensus Statement of the European Atherosclerosis Society recommends performing genetic experiment if available, if any causative mutation is recognized, subsequent cascade testing should be done in the family [33]. A comprehensive detection approach including targeted NGS, a testing for copy number variations, and scores of polygenic trait, can diagnose most patients [40].

Although FH is the most common and severe genetic form of hypercholesterolemia, mutations in proteins required for the elimination of sterols can also cause hypercholesterolemia [41]. A rare, autosomal recessive disorder, sitosterolemia, is caused by mutations in the ABCG5 and ABCG8 genes that phenotypically overlap with FH but it aetiologically is a different entity [42]. Cholesterol levels can be extremely elevated in some sitosterolemia patients but the diagnosis should be confirmed by increased plasma levels of plant sterols [43]. An interesting case supporting the importance of broad NGS analysis in identification of gene defect responsible for severe hypercholesterolemia has been reported by Rios et al. [41]. They performed WGS by using sequencing-by-ligation method to sequence the genome of an 11-month-old breast-fed girl with hypercholesterolemia with an atypical presentation. She presented with xanthomas and very high plasma cholesterol levels, no family history of FH, and her parents had normal plasma cholesterol levels. Based on this presentation, either an autosomal recessive disorder or a de novo mutation could be suggested. However, sitosterolemia was ruled out by documenting a normal plasma sitosterol-to-cholesterol ratio and known genetic causes of severe high plasma cholesterol levels were priorly ruled out by sequencing the responsible genes (LDLRAP, LDLR, PCSK9, APOE and APOB). Despite a normal plasma sitosterol-to-cholesterol ratio, WGS demonstrated two non-sense mutations in ABCG5 (Q16X and R446X) indicating the presence of sitosterolemia. This diagnosis was consequently confirmed by the finding of high plasma levels of plant sterols in a blood sample obtained after the infant had been weaned [41].

3.2. Hereditary cardiomyopathies

Based on structural and functional changes to the cardiac muscle, most of cardiomyopathies are grouped into specific morphological and functional phenotypes as follows: hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy (DCM), restrictive cardiomyopathy (RCM), arrhythmogenic right-ventricular cardiomyopathy (ARVC)/arrhythmogenic ventricular cardio-myopathy and unclassified cardiomyopathies containing left-ventricular non-compaction (LVNC); each phenotype is then sub-grouped into familial and non-familial forms [44]. This classification system assigns patients with identified de novo mutations to the familial group because they can subsequently transmit their disorder to their offspring. The majority of familial cardiomyopathies are monogenic diseases that are caused by a rare mutation in a specific gene [44].

The inherited cardiomyopathies are heterogeneous diseases and two of their major types are HCM and DCM [13,45]. HCM is inherited, in over half of cases, as an autosomal dominant genetic trait while about 20–35% of DCM cases show a family history with autosomal dominant inheritance pattern; some familial cases of DCM can be explained by an autosomal recessive or X-linked recessive trait [46–49]. Since the identification of a beta cardiac myosin heavy chain gene missense, MYH7, mutation as a molecular basis for familial HCM in 1990 [50], there has been a significant progress in exploring the genetic background of hereditary cardiomyopathies, proving it to be a highly heterogenic disease caused by mutations in at least 30 different genes [13]. HCM is mainly caused by numerous mutations in different genes encoding sarcomeric proteins whereas the pathways resulting in DCM show a great deal of variety, involving genes encoding sarcomere, Z-disk, nuclear lamina proteins, intermediate filaments and the dystrophin-associated glycoprotein complex [46,51,52]. WES, as a powerful tool, allowed the identification of several novel genes playing role in cardiomyopathy including mutations within MRPL3, MRPL44 and AARS2 in mitochondrial cardiomyopathies [53–55] and GATAD1 mutation in the autosomal recessive form of DCM [56].

Although limited, detecting the causative genes and mutations underlying cardiomyopathies has prognostic implications. For example, some gene variants of cardiomyopathies are linked with early onset disease manifestation, an overall poor prognosis or a high incidence of sudden cardiac death [13,57–59]. Additionally, about 5% of HCM patients with a detected genetic cause have more than one mutation in sarcomeric genes in whom a more severe phenotype and poor prognosis is observed [60,61] whereas patients with negative genetic tests more probably have a mild phenotype [62]. It has also been reported that a deletion in ACE gene encoding angiotensin-converting enzyme to be associated with a faster progression of hypertrophy and higher incidence of sudden cardiac death in HCM indicating that HCM phenotype may be modified by common polymorphisms [63,64].

Regarding hereditary cardiomyopathies, the usage of genetic information in clinical practice has been limited for two reasons: first, a number of different mutation frequencies have been reported; and, second, clinical manifestations are highly heterogeneous [65]. Hence, a specific genetic test in these patients is clinically of essential importance. A targeted “multiple causative genes” resequencing appears to be the right approach both for research and for diagnostic development. In recent years, the knowledge of the underlying genetic causes has substantially produced by NGS techniques. HCM, the most common inherited cardiac disease, is shown to be related to 20 genes and a total number of about 1400 distinct mutations [66], although environmental factors, such as age, sex and lifestyle can modulate clinical presentation of the disease [67–70]. Numerous causative mutations detected so far almost exclusively are missense variants resulting in structurally abnormal polypeptides that disrupt normal sarcomere function. Two leading causes of HCM, mutations in MYH7 and MYBPC3 (cardiac myosin-binding protein C), are collectively liable for almost 80% of cases [60]. In 2011, Meder et al. [71] were first who employed a targeted NGS approach in screening patients with hereditary cardiomyopathy. They applied a 47-gene panel to detect mutations in 10 patients with HCM and DCM and detected 27 new possibly damaging mutations [71]. The authors organized a microarray based on target enrichment followed by SOLiD NGS for an efficient and “time- and cost-efficient” genetic detection of cardiomyopathies [71]. Thereafter NGS method has entered the clinical diagnosis of cardiomyopathies [72]. In a large NGS study, 223 unrelated patients with HCM were studied by massive parallel resequencing on Illumina GAIIx in order to analyze coding, intronic and regulatory regions of 41 cardiovascular genes [73]. Excluding titin, 152 potentially causative variants in sarcomeric or associated genes (89 novels) were identified in 64% of patients. In cases in comparison with controls, an excess of rare non-synonymous single-nucleotide polymorphisms (nsSNPs) was observed in four sarcomeric genes (MYH7, MYBPC3, TNNI3 and TNNT2), with 34% of patients carrying candidate variants in desmosomal and ion channel proteins [73].

Another great example of the effect that NGS had on disclosing the genetic context of cardiomyopathies is the detection of a truncating mutation in the TTN gene as a main cause of DCM [74]. Since 1999, mutations in TTN, which encodes the giant muscle filament titin, has been found to cause familial DCM [75,76] but due to its large size (363 exons), the analysis of this gene by traditional Sanger sequencing was not easy so that scale of this causality remained underestimated for a long period of time. In a multicenter study that published in 2012, assessed mostly by NGS approach, Herman et al. [74] analyzed the full coding TTN sequence in patients with DCM, those with HCM, and in control group to evaluate the deleterious variants for co-segregation in the studied families. The truncating mutations of TTN, as a common reason of DCM, were determine in about 18% of sporadic idiopathic DCM and 25% of familial cases. The investigators concluded that incorporation of sequencing analysis of TTN into genetic screening should increase the sensitivity of genetic testing by around 50%, permitting earlier diagnosis and treatment to prevent progression of the disease in patients with DCM [74]. An independent study in 2014 confirmed this conclusion [77].

In one of the largest studies to date on the utility of broad gene panels in cardiomyopathy, a total of 766 DCM patients from the United States were genetically tested over 5 years in a laboratory of molecular diagnostics [78]. The patients were tested using gene panels of increasing size from 5 to 46 genes which their role in DCM had previously defined, beginning with a 5-gene Sanger panel to a NGS 46 gene Panel. By increasing gene panels’ size the clinical sensitivity for diagnosis of DCM increased by more than three times, from a range of 7.7–10% to a range of 27–37%, mostly because of the inclusion of the TTN gene. However, this rise in the sensitivity was at the expense of an increase in the percentage of patients receiving an inconclusive test result, from a range of 4.6–6.5% to a range of 51–61% [78]. In another large study conducted on >1400 participants with cardiomyopathies including HCM, DCM, ARVC, RCM and LVNC, a single exon resolution NGS-based copy number analysis was performed for up to 46 cardiomyopathy genes [79]. These authors, although identified clinically significant deletions and duplications in less than 1% of individuals, they concluded the added benefit of exon level deletion/duplication analysis was not costly significant in routine diagnostic testing [79]. The most recent studies enrolling smaller number of HCM patients confirmed the suitability of the target NGS methodology for clinical goals and the significant role of choosing proper patient in molecular genetic testing performance is more cost-effective [80,81].

ARVC, a rare familial cardiomyopathy that causes sudden death in young people and athletes rarely appears symptoms at youth and is a challenge to recognize it in early stages [82]. This heart-muscle disease in its advanced stages is difficult to be differentiated from DCM [83]. In a most recent study, targeted sequencing was performed in 14 ARVC cases using an “Illumina HighSeq 2000” with focus on 96 known cardiomyopathy and channelopathy genes for filtering. Approximately, 75% of cases with a certain diagnosis of ARVC based on 2010 task force criteria had desmosomal mutations while patients with a “possible” phenotype had variants in DCM related genes [84]. NGS-based panels are particularly attractive for ARVC diagnosis because these tools allow simultaneous molecular analysis of all the disease-related genes that is optimal for better explanation of the phenotype [85]. Recently, an NGS approach with a panel of multiple cardiomyopathy- and arrhythmia-associated genes has been successfully used to unravel the spectrum of genetic causes in patients with RCM [86].

3.3. Sudden cardiac death and channelopathies

Inherited arrhythmias are often called “channelopathies” because ion channel subunits or their regulating proteins are usually influenced by the underlying genetic defects. Genetic testing are gradually revolutionizing the disease management because of an exclusive role of NGS techniques for the detection of causative mutations for frequently fatal conditions, and screening of at risk relatives [87]. The mutation-specific testing for family members is recommend by experts whenever the genetic cause of disease has been established [82]. Clinical features as well as a unique genetic profile can characterize each of these cardiac channelopathies [88]. The determination of the cause of sudden cardiac death (SCD) is another application of comprehensive genetic testing. Sudden arrhythmic death syndrome (SADS) is defined as sudden death without an obvious cause, with inconclusive post-mortem results and negative toxicology tests. Genetic testing of KCNQ1, KCNH2, SCN5A and RYR2 genes in samples acquired after death, have shown that 35% of SADS are caused by ion channelopathies. The most common causes of SADS due to channelopathies are mutations of catecholaminergic polymorphic ventricular tachycardia (CPVT) (20%), long QT syndrome (LQTS) (15%) and Brugada syndrome (BrS) (<1%) [89].

There are several kinds of LQTS [90]. It is inherited in an autosomal dominant manner with reduced penetrance with an exception of a type referred to as “Jervell and Lange-Nielsen syndrome” which is associated with sensorineural deafness and is inherited in an autosomal recessive manner. Disease causing mutations can be identified in approximately 75% of cases. LQTS has been associated with hundreds of variations in more than 13 genes. Loss-of-function mutations in 2 genes, KCNQ1 and KCNH2, together accounts for almost 70% of cases while gain-of-function mutations in one gene, SCN5A, is responsible for another 5–10%; these three genes collectively consist up to 80% of clinically defined cases while the <5% remaining are due to mutations of 10 minor genes [82]. Conversely, in BrS, a disease with autosomal dominant inheritance, determined by ventricular fibrillation pre-disposition, conduction delays and nocturnal SCD [91,92], the yield of genetic testing is relatively low and disease causing mutations detection can be occurred in only 25 to 30% of cases [93]. Juang et al. [94] reported de novo mutations in patients with non-familial BrS and as stated by Omar et al. “many more genetic mutations are likely to be identified.” To date, BrS has been associated with hundreds of mutations in 17 genes; loss-of-function mutations in the SCN5A gene, that encodes the sodium ion channel alpha-subunit, account for the vast majority (75%) of the identified cases and cause 20–25% of BrS while the remaining genes accounts for a small number of individual cases encoding other ion channels (sodium, potassium and calcium) [95,96]. Moreover, BrS demonstrates incomplete penetrance, mixed phenotypes, and variable expressivity [97–99]. The manifestations of mixed phenotype, i.e. the combination of BrS with other cardiac arrhythmias, have been observed with LQTS [100], short QT syndrome (SQTS) [98,101], cardiac conduction disease [101] and other arrhythmias [93]. To overcome the genetic heterogeneity and consequently diagnostic confusion, Allegue et al. in 2010 [102] developed a novel simplified diagnostic approach for LQTS and BrS based on existing published mutations and Sequenom MassARRAY system. This system is very sensitive but NGS is perhaps even more sensitive than this and other sequencing methods [103]. NGS can be used to comprehensively analyze many arrhythmogenic mutations simultaneously and to identify pathogenic variants from genetic background noise in a fast and cost-efficient manner [104].

Millat et al. [105] applied NGS for major LQTS genes sequencing on a population of 30 previously studied suspected individuals to survey the practicability, specificity and sensitivity of NGS. They showed that, as compared to before used mutation finding methods, the NGS is a very efficient, rapid and low-cost high-throughput mutation recognition technique and can be used in clinical laboratories. In 2017, Chae et al. [106] established for the first time a targeted multi-gene NGS custom panel of 13 LQTS-associated genes by using the Ion PGM platform and validated the system for routine use in clinical genetic diagnosis of LQTS and other genetic diseases. NGS can also be used in clinically diagnosed BrS cases with no identified genetic mutations for both diagnostic and prognostic purposes as well as for screening for mutations in relatives. For example, NGS methodologies have been recently applied to identify novel candidate genes associated with BrS [107] and to identify novel potential genetic associations between BrS and other cardiac disorders [108,109].There are other examples emphasizing the need for NGS technologies to detect common and/or rare genetic mutations that may have various phenotypic manifestations. In a cohort of 35 patients from nine different Spanish centres causes of unexplained cardiac arrest, by using NGS, were found to be BrS in seven, CPVT in five, short QT syndrome in one of these patients among other cardiac channelopathies [110]. The main gene responsible for up to 50% of CPVT cases is the ryanodine receptor (RYR2) [111]. The autosomal recessive inheritance pattern of CPVT is linked with the CASQ2 gene, which its mutations are the cause of 3–5% of cases. The CPVT is associated with sudden infant death syndrome and this happening might be the first appearance of it. Therefore, CPVT genetic evaluation is important for index cases at birth and for all their family members as early as possible [82]. Given the large size of the RYR2 gene and the vital importance of time-efficient genetic testing, the application of NGS-based diagnostics in the case of CPVT appears particularly attractive. In a most recent study, Bosch et al. [111] identified a rare novel variant (p.11I > S_RyR2) as a cause of familiar CPVT using NGS technology which was later confirmed by conventional Sanger sequencing.

3.4. Congenital heart disease

Congenital heart diseases (CHDs) are defined as structural or functional abnormalities occurring during intrauterine life that are present at birth [112]. Until recently patients with CHD would not survive long enough to have children and as a result, most patients had a negative family history [112]. The dramatic improve in paediatric cardiology and surgical interventions in the past years have led to a significant increased survival of patients with CHD so that multi-generation families with multiple affected members are started to being observed [113,114]. To conduct more effective genetic counselling for CHD families, it is essential to improve our knowledge in regard to the genetics of CHD. Genetic testing in CHD patients is not only important in assessing risk for reproduction and other family members [113], but also in predicting implication of another organ and/or future complications. For instance, patients with the NKX2.5 mutation who underwent a successful operation for cardiac defect can still be at risk of heart block [115].

As the most common type of birth defect, CHD has a complex aetiology with both environmental and genetic causes. This heterogeneous disorder is related with long-identified chromosomal abnormalities such as Down syndrome and chromosome 22q11 deletion, as well as with mutation in over 50 human disease genes [116]. Nevertheless, a large proportion of CHD-associated mutations are clustered in a few developmental genes including, NKX2–5, GATA4 and NOTCH1 [117]. Notably, no causal chromosomal abnormality or mutation is detected in the majority of patients with CHD [118].

The process of discovering the genetic basis of CHD is particularly challenging due to the scarcity of large pedigrees, the genetic heterogeneity, variable expressivity and variable penetrance [119]. However, new technologies such as NGS have already contributed in identifying causative mutations among CHDs and may greatly advance our knowledge of the genetic factors contributing to CHD. By doing exome sequencing on DNA samples from two cousins of a large family with pleiotropic CHD, a novel MYH6 mutation was identified as a genetic cause of CHD which was the first use of NGS in this regard [120].

In another WES study in 13 parent-offspring trios and 112 unrelated individuals with non-syndromic atrioventricular septal defect (AVSD), the investigators identified the new CHD-associated gene NR2F2 with several disease-causing variants, which turned out to account for other forms of CHD, including tetralogy of Fallot (TOF), ventricular septal defect (VSD), aortic valve stenosis (AS) and coarctation of the aorta (CoA) [121]. In 2013, Zaidi et al. [122] published a notable study wherein the authors combined a case–control study with WES of trios. In 362 trios of healthy parents and affected offspring, they identified a significant increase in de novo coding variants in potentially causal genes involved in histone methylation expressed in the developing heart. The study was important in contributing significantly towards the overall understanding of CHD and genes particularly involved in developmental pathways. However, they found no cause for CHD in the majority of families included in this study and estimated that de novo point and small insertion/deletion mutations can contribute to approximately 10% of severe CHD cases [122]. A subsequent study on sporadic forms of CHD also identified a significant increase in de novo coding variants across all genes including genes highly expressed in the developing heart [123]. In a more recent study, the investigators applied exome sequencing to both syndromic and non-syndromic sporadic forms of CHD [124].

Some other researches have successfully used NGS in CHD by concentrating on members families with assumed Mendelian states of CHD [120,125–130]. In 2014, Blue et al. [131] reported the results of screening a targeted panel of 57 candidate genes in a cohort of 16 families with a strong history of CHD versus over 6000 control subjects. The authors found variants of unknown significance in 25% of the studied families in addition to a pathogenic variant in five families (31%).

Based on the genes identified, they also identified a possible syndromic cause in three of these five families. In two other recent studies, the investigations applied exome sequencing to familial forms of CHD [132,133]. However, in practice, applying this approach can be quite challenging particularly so in small, dominant families in which there are considerable challenges in variant interpretation and in establishing its pathogenicity. For example, in a multigenerational family with bicuspid aortic valve and other forms of CHD, Martin et al. failed to identify any strong coding variants, despite the use of three commonly used strategies for variant selection and use of linkage analysis along with exome sequencing [128].

3.5. Thoracic aortic aneurysms and dissections

Thoracic aortic aneurysm, as a silent but dangerous disease, in many cases appears with acute aortic rupture or dissection, leading to a high mortality. It is reasonable to do active screening for detecting people at risk at the early stages because prophylactic medical and surgical measures may reduce the risk of thoracic aortic aneurysm and dissections (TAAD). Although more than 20% of cases of TAAD have a familial form with a usually autosomal dominant pattern of inheritance, an underlying genetic defect is found in about 20% of index cases with a limited penetrance. Familial TAAD can be non-syndromic or be part of a well-characterized genetic syndrome as a case for the Marfan syndrome, caused by FBN1 mutations and Loeys–Dietz syndrome, causing by either TGFBR1 mutations or TGFBR2 [134]. Four genes, including TGFBR2 and TGFBR1, MYH11 and ACTA2, have been identified to cause TAAD; ACTA2 gene mutations are the most common reason of familial TAAD, which accounts for up to14% of identified cases [134]. Sakai et al. [135] used two different technologies, resequencing array technology (ResAT) and NGS, to analyze eight genes associated with a syndromic aortic aneurysm and/or dissection (AAD) in 70 non-syndromic patients (35 thoracic, 30 abdominal and 5 both thoracic and abdominal ADD). Using the Illumina GAIIx platform, they identified eighteen variants with NGS as well as ResAT technologies and stated that ResAT cannot indicate short insertions/deletions (indel), and it is impractical to update custom arrays frequently. The authors also deduced that, though NGS needs advanced informatics methods, it can detect almost all mutations types [135]. Furthermore, in another study by using exome sequencing of two far affected relatives, a frame shift mutation in the SMAD3 gene was identified as the reason of vascular disease in TAAD alone families, along families with TAAD, the inheritance pattern of intracranial and other arterial aneurysms was autosomal dominant [136]. There are some newly published studies that applied NGS for molecular diagnosis of both syndromic and non-syndromic TAAD with diagnostic yield varying between 4 and 35% [137–140]. As stated by Poninska et al. [140], this low diagnostic yield indicates that in most of the patients with TAAD, the disease is caused by different factors such as genetic or environmental or a combination of these factors.

According to AHA guidelines [141], in patients with a family history of TAAD, sequencing of the ACTA2 gene is the reason for detecting if mutations in this gene account for the inherited predisposition (class IIa), and sequencing of other genes (TGFBR1, TGFBR2 and MYH11) may be considered (IIb). The same approach may be advisable in sporadic disease among young patients with no additional risk factors [142]. There is evidence that rare copy number variants (CNVs) contribute to the pathogenesis of TAAD by disrupting genes regulating vascular smooth muscle cell adhesion and contractility in 13% of sporadic and 23% of familial TAAD cases [143]. Similar to other complex traits CVDs, also the NGS technologies have been applied for the recognition of rare genetic variants in large abdominal aortic aneurysm sample groups [144]. Using SOLiD platform, the investigators found 681 coding variants; however, the majority of the discovered candidate novel variants were false positives [144].

3.6. Coronary artery disease and stroke

Coronary artery disease (CAD), as the leading cause and stroke are major causes of mortality and morbidity in both the industrialized and developing world; these two collectively account for two of every three deaths [145]. The Framingham study reported that a family history of heart disease is associated with CAD risk independently of other risk factors, showing a 2.4-fold higher risk in men and a 2.2-fold higher risk in women [146]. Myocardial infarction is more likely to cluster in families than stroke [147]; with heritability estimates as high as 60% for CAD [148]. Although Mendelian disorders, caused by rare mutations, are well-recognized and well-characterized for conditions such as familial dyslipidemia which are associated with CAD, the vast majority of CAD and stroke cases represent a multi-factorial complex disease. The phenotypes of complex or multifactorial diseases, unlike Mendelian disorders, do not have a single genetic cause. Rather, they are likely caused by many contributing factors including multiple genes as well as lifestyle and environmental factor, each with relatively small effects on gene expression and disease. As a result, the genetics in non-Mendelian CVDs, such as CAD and stroke, is still a difficult but highly promising conundrum. In 2007, genome-wide association study (GWAS) techniques provided the first evidence that common genetic variation on chromosome 9p21 influence the risk for CAD [149,150]. During the next 5 years, 36 genetic variants reported to be associated with higher CAD risk and each of these associations confirmed in populations independent of the discovery population [151]. In 2013, these loci for CAD risk underwent a meta-analysis in a total sample size of over 190,000, the CARDIoGRAMplusC4D Consortium [152]; Based on this sample size, the chances of even one of these 36 genetic risk variants being false is extremely improbable [153]. However, these risk variants altogether only account for about 10% of the expected heritability while it is estimated that about 50% of susceptibility for CAD is genetic [154]. The GWAS only provides population-attributable risks information and cannot be transferred to a CAD individual. In contrast to GWAS, the NGS results can be used to the affected individual directly. Additionally, in most cases, NGS is unbiased approach, in that it permits both rare (with a minor allele frequency (MAF) of <5%) and common variants (with MAF >5%) to be identified, makes it appropriate for studying heterogeneous, complex diseases such as CAD and stroke. This is because NGS is able to complement the common susceptibility variants established through the GWAS [155]. For example, in a study published in 2011, using the Illumina GA platform, sequence data from an approximately 240 kb region on chromosome 9p21 was analyzed in 47 unrelated individuals of European ancestry from the HapMap CEU population [156]. By comparing the findings of this NGS study with pilot data from the 1000 Genomes Project [157], the authors demonstrated for the first time that in spite of several gaps in sequence coverage which existed after the alignment to a reference genome, the targeted sequencing with NGS provides high sensitivity for rare variants [156].

In another study, the researchers used WES by NGS to identify rare variants in the gene encoding adiponectin which is shown to be associated with large effects on families from the Insulin Resistance Atherosclerosis Family Study [158]. They proposed that this approach would be beneficial in detecting novel genes affecting complex traits in a wide range of family studies [158].

Data on stroke predisposition is relatively sparse mainly because of the genetic complexity of stroke. Several rare Mendelian stroke disorders and syndromes are known such as CADASIL CARASIL and MELAS [159]. CADASIL, the most common genetic ischemic stroke, is a dominantly inherited small vessel disease caused by more than190 known mutations of the NOTCH3 gene [160]. However, the majority of ischemic strokes are complex disease resulting from multiple genetic and environmental factors. While the first goal of exome sequencing was to identify rare causative mutations responsible for the Mendelian genetic disease, now WES is successfully using for pinpointing rare variations associated with ischemic stroke [161,162].

In one of the biggest exome studies to date, among 496 ischemic stroke patients from a wider exome sample of 4204 unrelated individuals, seven variants in the paraoxonase-1 gene were found to be associated with ischemic stroke wherein two were exclusively found in patients of European ancestry and one detected only in patients of African Ancestry [163]. Moreover, the availability of international projects such as the Genomics England 100,000 genomes project [164] with the aim to generate and store WES data, would lead to pinpointing multiple variants that have a small but significant effect on the pathophysiology of stroke.

3.7. Other cardiovascular diseases

The NGS technologies have also been used in other CVD-related conditions with complex traits, including thrombophilia, hypertension and dyslipidemia. In a family with thrombophilia record, NGS has been used to sequence genomic DNA by applying the Illumina platform for multigenic risk detection that cause inherited thrombophilia and achieving suitable pharmacological treatment [165]. In this study, 200 variants were diagnosed and evaluate with HapMap different populations. Moreover, as noted by Costa et al., association of multiple genes with numerous cardiovascular diseases, hypertension as a main risk factor in 69% of stroke and 49% of CAD, results in 50% of all cardiovascular mortality and its heritability suggested to be from 30% to 50% [166]. Finally, Sadananda et al. showed that NGS approach can reliably and accurately detect pathogenic variants in a large number of patients with low levels of HDL cholesterol − one of the most frequent lipid abnormalities and a key risk factor for CVD [167].

4. Non-coding RNA analysis in cardiovascular diseases

Profiling of gene expression evaluating disease versus healthy, is a worth approach for detection of possible biomarkers for prediction/diagnosis of disease intensity and treatment targets identification. Recently, gene expression survey by microarray has been developed in cardiovascular researches [168,169].

RNA-sequencing (RNA Seq.) is a very effective strategy for determining expression profile [170] by using different steps for entire RNA purification and following next generation sequencing [171]. It has a wide application spectrum, (e.g. gene regulation survey, the discovery of developmental steps and pharmacogenomics) [172], and can survey all aspects of RNA, (e.g. RNA–RNA interaction, RNA–protein binding and RNA structure) [173]. RNA Seq. has removed some difficulties of before used methods and let high accuracy concurrent research of transcriptome various layers, i.e. both coding and non-coding regions [174,175]. It can be performed by the same platforms for DNA Seq. also the same steps just some more, i.e. the summary of RNA Seq. is shown in Figure 2. However, according to the project goal, RNA sample preparing is different [176] and regarding various types of RNAs, there are some kinds of RNA Seq. such as total RNA, mRNA, targeted RNA, small RNA, single cell RNA, ribosome profiling and RNA exome capture seq.

Figure 2. Summary of the RNA-Sequencing (RNA-Seq.) steps workflow. The first step, for purifying and isolating cellular RNAs, cell are disturbed by chaotropic agents and detergents. After homogenization, total RNA can be isolated from cell debris and RNA subsets molecules are separated by specific protocols. After achieving suitable RNA, the RNA should be changed to double-stranded complementary DNA (cDNA).Next, the adaptors addition and DNA amplifying occur. The cDNAs are sequenced by sequencing platforms, i.e. Life’s 454 and SOIiD. Finally, short reads sequences are analyzed.

The Encyclopaedia of DNA Elements (ENCODE) project declared that only 3% of the genome is protein coding and about 85% is non-coding [177]. Furthermore, the GWAS finding represents although disease causing variants happen more in protein coding sequence but majority of variants (>80%) are in non-coding sequence, that's why induced recent complex disease aetiology researches to non-coding in addition to coding region [178]. There are some studies have shown non-coding RNAs (nc RNA) importance in human heart failure [179,180]. Micro RNAs (miRNAs) are the best nc RNA in heart characterization, which regulate mRNA expression [179]. MiRNA expression studies of the failing and control the human heart, detected mRNA expression alteration and proved miRNA profile can diagnosis samples with high accuracy [181]. The first document of miRNA signature in CAD was provided by Menno Hoekstra that proved some miRNAs, i.e. miR-198 can be an effective tool for determining of CAD risk in patients [182]. In a study in 2013 [183], it was elucidated that miR-22 increased expression in transgenic mice cardiomyocytes, result in heart failure and cardiac dilation [183]. Furthermore, Chini [184] reported some miRNAs (i.e. increased levels in heart failure: miR-423-5p, miR-320a, miR-22, miR-92b, miR-29b, miR-122 and miR-142-3p, reduced levels in heart failure: miR-107, miR-125b, miR-126, miR-139, miR-142-5p and miR-497) for heart failure diagnosis in his publication [184] that all of them except miR-320a, miR-22 and miR-92b were concluded before [185]. The miR-208 is another example that silencing of it, in heart failure rat model, induced survival while miR-208 overexpression, induce cardiomyocytes hypertrophy [186–188].

These reports express the role of miRNAs as biomarkers in CADs is significant. NGS efficiently lead to identify novel miRNAs and define molecular mechanisms of miRNA involvement in cardiovascular disease [184]. In 2005, for the first time, small RNAs (about 2 million) sequencing of Arabidopsis thaliana was done [i.e. miRNA were recognized among the small interfering RNAs (siRNA)] and it was the entry to miRNA applications in molecular diagnosis [189]. In miRNAs as biomarkers concern, using NGS have advantages to traditional technologies, i.e. in per run, so much GB of data is generated, novel miRNAs are discovered [184], miRNA even with low expression can be detected, other ncRNA (i.e. tel-sRNAs (telomere specific small RNAs), PASRs (promoter-associated small RNAs) and crasiRNAs (centromere repeat-associated small interacting RNAs) can be received by NGS [190]. Ounzain et al. [179] have found some associated novel long non-coding RNAs (e.g. novlnc6 that [185] modulates Nkx2-5 and is associated with cardiac evolvement enhancer) in cardiovascular disease. In this study, novel lncRNAs relationship with heart miRNA such as mir139a, -499 and 30c [179] were realized. In a pilot study of DCM and RCM patients [191], heart transcriptome changes were evaluated among patients and healthy individuals by RNA-Seq, about 15,800 genes were detected (Ensembl v70) that express in heart and GENCODE Consortium annotated 500 of them as long non-coding RNA (lnc RNA). 140 lncRNA were deregulated in heart failure study group versus healthy group, among these deregulated lncRNA, 76% were down-regulated and the rest was up-regulated. Additionally, stated specific transcriptome changes of “oxidative phosphorylation”, “actin cytoskeleton regulation” and “focal adhesion” in cardiomyocytes. Also, genes expression quantification which belong to the “cytokine-cytokine receptor interaction”, NF signalling pathway” and “phagosome” were introduced as a RCM and DCM signature [191].

The recognition of novel miRNAs with specific targets can be the beginning of therapeutics development by miRNA targeting. Therapeutic strategies base on miRNA can be composed with other therapeutics (e.g. stem cells). Actually, MiRNA can regulate stem cells differentiation into cardiovascular cells. Differentiated cells of a patient can be converted to pluripotent stem cells (iPSCs) by reprogramming. There is evidence that miRNAs have a role in iPSCs induction [192]. Regarding cardiovascular diseases are complex, so it is not enough just concentrate on the DNA level and other levels (i.e. epigenetics, transcriptomics and proteomics) should be investigated [193].

5. Conclusion and future perspectives

The most genetic disease definitive detection is provided by recognition of a causal mutation. Sanger’s sequencing may efficiently detect the culprit mutation in disorders caused by a limited number of gene mutations; however, this conventional technique is not efficient for diagnosis of many single-gene disorders because the disease is highly genetically heterogenous or the responsible gene(s) has not been explored [41]. NGS by re-sequencing the whole exome (or genome) of patients has revealed that can be efficient approaches in Mendelian diseases studies [194]. In addition, applying these technologies to complex diseases, including CAD and other CVDs may result in exploring the genetic background of these conditions. Although GWAS have made great improvement in understanding the genetic causes of complex CVDs, at present, heritable risks for CVDs are modestly explained by genetic variants. This has raised the question of whether rare mutations (i.e. variants with allelic frequency <1%), which are not represented on commercial SNP arrays, at least partly explain the observed missing heritability [195]. The ability for detection of structural and rare variants, and the challenges in relation of variation different types with phenotype, can be reached by applying advanced technologies such as NGS. In 2009, the National Heart, Lung and Blood Institute (NHLBI), Exome Sequencing Project (ESP) was founded to discover rare, protein-coding variants associated with heart-, lung- and blood-related diseases. For determining rare variants related to CVDs, exome sequencing was used to diagnosis well phenotypic of cardiovascular cohorts [196]. Results from ESP and several similar recent WES projects suggest that data generation is not a significant technical problem for sequencing-based studies of rare-variant associations; nevertheless, the data-management, variant-calling step, QC and study of these data are challenging yet [196]. There is a significant consensus among the human geneticists which the way to manifest the biological processes that cause complex diseases is through large-scale consortia and data sharing [196]. As previously mentioned, WES and WGS have identified numerous rare variants important in both Mendelian and complex CVD conditions. Although, WGS provides better coverage than WES (the exome is almost 1% of the whole genome), due to the costs and low depth of sequence reads, it is still not feasible to use WGS in a large number of individuals and WES is possibly an intermediate available approach. WGS will definitely be the best option for genetic testing especially in cardiovascular disorders. Since non-coding variants play a notable role in increasing the risk of CVDs and the cost of sequencing continues to decrease, in the future WGS will eventually be more cost-effective than WES. WGS is useful in revealing the aetiology of unknown genetic diseases as well as in diagnosis of patients with atypical presentation of known diseases. Furthermore, new ways of detecting disease-associated biochemical pathways may be opened up by using functional genomics methods, such as RNA-seq and ChIP-seq. Unravelling the genetic causes of CVDs in addition to identification of environmental risk factors of these disorders provide new approaches for precision and efficient therapies and NGS technology shed light on the road of personalized medicine.

In conclusion, NGS technologies have made WGS a feasible way for gathering global genomic information. Currently, several NGS platforms are used in cardiovascular genetics that share general processing steps but vary in specific technical details determining their advantages or disadvantages. It has been recently revealed that NGS to have great potential for detecting novel causative mutations in various Mendelian diseases. It is also predicted that the NGS will be growingly essential in the study of multifactorial traits such as CVDs in which risk prediction thorough the identification and characterization of casual genes remains an important challenge for advance in prevention and treatment. However, the main limitation of WGS and WES application for diagnosis purpose is interpretation of the genetic results their use for routine genetic diagnosis could also be limited by technical issues.

Acknowledgements

This review was carried out in Genetic Research Laboratory, Genetic Research Centre, Rajaie Cardiovascular Medical and Research Centre, Iran University of Medical Sciences, Tehran, Iran.

Disclosure statement

The authors have no conflict of interest to declare.