Polymorphism type
Sequence location
Predicted protein and potential functional effects
Occurrence in genome
Potential disease impact
Nonsense
Coding
Prematurely truncated, most likely loss of protein function
Very low
High
Missense, non-synonymous
Coding, non-conserved
Altered amino acid chain, mostly similar protein properties
Low
Low (to high)
Missense, non-synonymous
Coding, conserved
Altered amino acid chain, mostly different protein properties
Low
Medium to high
Rearrangements (insertion/deletion)
Coding
Altered amino acid chain, mostly different protein properties
Low
High
Sense, synonymous
Coding
Unchanged amino acid chain, rarely an effect on exon splicing
Medium
Low (to medium)
Promotor and regulatory sequences
Non-coding, promotor/UTR
Unchanged amino acid chain, but may affect gene expression
Low to medium
Low to high, depending on site
Intronic nucleotide exchange (<40 bp)
Non-coding, splice/lariat sites
Altered amino acid chain, failed recognition of exonic structure
Low
Low to high, depending on site
Intronic nucleotide exchange (>40 bp)
Non-coding, between introns
Unchanged amino acid chain, rarely abnormal splicing or mRNA instability, site for gene rearrangements
Medium
Very low
Intergenic nucleotide exchange
Non-coding, between genes
Unchanged amino acid chain, may effect gene expression, site for gross rearrangements
High
Very low
Genetic association studies (or: case-control studies) are an analysis of statistically significant relationships between SNP alleles and phenotypic differences. The power of a genetic association study is a direct function of the number and quality of the SNPs used to screen a population for phenotypic variability. SNPs and haplotypes can vary in their prevalence among different populations. Thus, a SNP associated with a particular phenotype or quantitative trait in one population may not have the same frequency or effect in another population, e.g., when the population is of different ethnicity, age or gender. Large datasets of chromosomal SNPs have been published since 2000 [1, 13–17], along with improved methods to screen immense numbers of SNP candidates. More than three million variants have been reported and are catalogued in public databases (e.g., http://www.ncbi.nlm.nih.gov/projects/SNP/). Newer techniques allow high-throughput genotyping to study simultaneously large numbers of SNP loci (currently: >4.0 M markers per sample/chip; e.g., HumanOmni5-Quad, Illumina Inc.) and are based on matrix-assisted laser desorption ionisation time-of-flight (MALDI-TOF; e.g., Sequenom MassARRAY), pyrosequencing, or hybridisation.
A huge and as yet unsolved problem is the identification of clinically relevant mutations in a plethora of functionally neutral single nucleotide polymorphisms. Common SNPs can be filtered out through comparison with genomes and exomes from healthy individuals (e.g., http://www.1000genomes.org/) [18] or dbSNP (see http://www.ncbi.nlm.nih.gov/projects/SNP/, despite an increasing contamination with clinically relevant mutations), but this approach is not possible for the many rare SNPs in the human genome. Indeed, comprehensive NGS-based re-analyses have recently found that 12 % of the previously reported mutations are not disease-causing itself [19]. In principle, large-scale whole-genome sequencing may reduce the number of novel variants from 3.4 million to a mere 150.000 per genome. Therefore, sequencing 100,000 individuals and comparing the results with their complete medical records (e.g., see http://www.personalgenomes.org/), would identify the vast majority of changes that do not give rise to disease. A comparable project is the NHLBI GO Exome Sequencing Project that focuses to discover novel genes and mechanisms contributing to heart, lung and blood disorders by pioneering the application of next-generation sequencing of the protein coding regions of the human genome across diverse, richly-phenotyped populations. These datasets and findings – obtained from potentially affected individuals – are also shared with the scientific community (Exome Variant Server; http://evs.gs.washington.edu/EVS/) and now showed that rare variants (e.g., less than 1 in 1,000 alleles) can be commonly identified in many cardiovascular genes. The presence of these databases, in contrary, enhances the need for certified and proven mutation databases for genes, such as such as the Human Gene Mutation Database (HGMD, see http://www.hgmd.org/), or potentially the Human Variome Project (http://www.humanvariomeproject.org/) or the Human Genome Organization (http://www.hugo-international.org/). Without theses, the clear clinical significance of many genetic variants and the role of the relevant genes in disease may remain uncertain for a long time. Similarly to SNPs and recently being more recognized [2], also many genomic imbalances were recurrently detected and were found in both, patients and healthy individuals (see central databases like Decipher, http://decipher.sanger.ac.uk/ or the Database of Genomic Structural Variations, http://www.ncbi.nlm.nih.gov/dbvar).
The clinical use of SNPs is still far away from being established, at least in arrhythmia prediction. This might be related to some inherent limitations with SNP studies [20, 21]. The two major issues are statistical power and replication of genetic findings in another, independent population set of same origin to avoid population stratification. In association studies, the prevalence of genetic marker alleles in unrelated subjects with a certain phenotype and (unaffected) controls will be compared and aim to correlate differences in disease frequencies between groups (or in trait levels for continuously varying characters) with differences in allele frequencies at an SNP. Thus, the frequencies of the two variant forms (alleles) of an SNP are of primary interest for identification of genes affecting disease. The traditional ‘case-control’ approach assumes that any noted difference in allele frequencies is related to the outcome measured and that there are no unobserved confounding effects. Unfortunately, allele frequencies are known to vary widely within and between populations, irrespective of disease status. For an appropriate study, an adequate sample size of the groups and a relatively high frequency of the minor SNP allele (to facilitate detection of allele frequency differences between the investigated populations) are needed. Usually, haplotype tagging SNPs (tagSNPs) were selected on chip-based arrays to systematically analyze nearly every genes approach. Typical criteria for tagSNP selection are a pairwise-only tagging with r2 >0.8 and a minor allele frequency (MAF) >0.1. Studies with small sample sizes may commit type II errors, i.e., not declaring a statistically significant result when there may be a difference. These underpowered studies can be misleading because genes may be undetected, and reporting of the odds ratio and 95 %-confidence interval are recommended [22]. The term β is defined as the chance of making a type II error. Values for β are typically 10–20 %, meaning a power (1 − β) between 80 and 90 %. In contrast, a sample size that is much larger than required may declare small differences to be statistically significant and thus commit type I errors (i.e., declaring a statistically significant difference when it may not be present). The term α refers to the chance of making a type I error; usually, a level of 0.05 or less is chosen. Due to the increasing, but also inconsistent number of GWAS publications, proposed guidelines have been developed which should facilitate the quality of association studies [23, 24], including strategies to ascertain heritability and exact phenotyping of a trait, to perform population stratification of cases and controls (ethnicity, age and gender distribution), to select physiologically and genetically meaningful markers, to address the probability of association, and to replicate initial results in independent studies [25, 26]. A p-value <5 × 108 is considered as statistically significant for GWAS results. This quite stringent significance threshold, that is frequently used when studying samples of European ancestry, accounts for about 1,000,000 independent common variant tests in the human genome. To date, only a few of the several thousand published association studies strictly meet the criteria to ascertain a (‘true’) genetic association. For arrhythmogenic disorders, first studies exist [27–30], but the majority of data is still unreplicated by independent approaches. Differences in study outcome may be related to population stratification, study design, still inappropriate marker selection, and lack of statistical power [11]. Discovery of meaningful SNP markers [31], e.g., indicating an elevated risk of SCD, is still far from being established. Common weaknesses of many association studies include study design failing to adequately identify true positives while eliminating false positives, poorly defined phenotypes and sampling from heterogeneous patient populations, inappropriately matched controls, small sample sizes relative to the magnitude of the genetic effects, failure to account for multiple testing, population and sample stratification, failure to replicate marginal findings and overemphasizing interpretation of study results. In the past, the optimum study design for association studies has been discussed because, often, studies were prone to population stratification and biased or spurious results. Thus, replication of the findings from genetic association studies in other populations became a cornerstone for the data quality, and, so far, only a few studies merit these criteria. In this line, a shift from case-control and cohort studies towards family-based association designs has been noted. These study designs have fewer problems with population stratification, but have greater genotyping and sampling requirements, and data can be difficult or impossible to gather.
Analysis of SNPs in Cardiac Arrhythmogenesis: Towards a Dissection of Common ECG Traits
Phenotypic variation in arrhythmia development is well known from families with inherited, arrhythmogenic disorders that have demonstrated an important phenotypic spectrum of the same mutation in affected family members [32, 33]. Recent reports have highlighted the importance of a family history of sudden death as a risk for ventricular fibrillation (VF) in patients experiencing acute myocardial infarction (AMI), pointing to the possibility of a genetic predisposition. Familial aggregation demonstrated an increased risk of SCD among patients with a parental history of cardiac arrest [34, 35], but a clearly defined genetic basis is not known to date [36]. Sudden death was found to share the same profile of risk factors for coronary artery disease and, thus, was not specifically predictable in the general population. These observations are also recognized from in patients with more polygenic disorders, such as myocardial infarction, for which not every patient develops ventricular fibrillation during acute ischemia [37, 38]. In a case-control study in patients with a first ST-elevation myocardial infarction (STEMI) and similar infarct sizes and locations, it was recently shown that (cumulative) ST-segment elevation was significantly higher among cases and that familial sudden death occurred more frequently among cases than controls [37]. Two population-based studies of the late 1990s demonstrated an increased risk of SCD in first-degree relatives of SCD victims and provided some evidence that genetic components may be involved in SCD of unknown (probably atherosclerotic) origin [35, 39]. A family history of MI/SCD was associated with SCD (RR = 1.57), after adjustment for other common risk factors and person–years at risk among (first degree) relatives [39]. After differentiating between family history of MI or of SCD, the positive family history of early-onset SCD finally was associated with a 2.7-fold increase in risk of SCD. In victims of VF in the setting of their first, acute MI it has been also reported that SCD of degree relative is a strong risk factor for ventricular fibrillation (OR = 2.72) [37].
Thus, arrhythmia development may have a common and modifiable substrate in both, rare inherited (monogenic) and common (polygenic) forms of various arrhythmias and a positive family history can be noted in both. In addition, multiple factors – such as age, gender, and environmental condition – play an important role in the modulation of the phenotype. Structural and electrical remodeling during acute ischemia, altered hemodynamic loads, or changes in neurohormonal signaling are recognized key features that alter ion channel gene expression. Down-regulation of major repolarizing potassium currents, Ito, IKr, IKs, and IK1, has been described in several models of heart failure and resembles a condition of “acquired QT prolongation” and reduced, but reversible repolarisation reserve [40]. Cellular abnormalities through disturbances in the electrical cell-cell coupling and a local reduction of conduction velocity facilitate re-entrant ventricular arrhythmias. These cellular abnormalities can be found in the structurally diseased heart. The extent of genetically controlled variation is not clear to date, but it is of potential interest and under recent investigations. Of note, some studies already focused on associations of SNPs in ion channel genes and a relation with myocardial infarction. Since KATP channels are involved in membrane regulation during metabolic stress, studies focused to identify variants in the KCNJ11 gene associated with SCD after myocardial infarction [41]. These channels are composed of four pore-forming Kir6.2 (KCNJ11) subunits and four sulfonylurea receptor subunits (SUR2A); sarcolemmal KATP channels regulate membrane potential and action potential duration, whereas the mitochondrial KATP channels are involved in ischemic preconditioning. So far, two non-synonymous polymorphisms (R371H, P266T) in two highly conserved pore regions are known that showed altered modulation by intracellular ATP and protons and differences in channel density [42] and, thus, are potential candidates for genetically determined electrophysiologic differences under ischemic conditions. Interestingly, mutations in the KCNJ8 gene have been associated with idiopathic ventricular fibrillation [43–45]. Phase 2 re-entry is a key mechanism for ventricular fibrillation complicating acute myocardial infarction as well as arrhythmias associated with Brugada syndrome. In this line, a heterozygous SCN5A gene mutation (G400A, located in cis with the H558R polymorphism) was reported in a patient who developed an arrhythmic electrical storm during acute myocardial infarction and suggested a hidden genetic predisposition to the severity of arrhythmias that in the setting of acute myocardial ischemia [46]. Another study indicated significant changes (up to 63 % down-regulation) of sodium channel transcription in dependence of the SNP composition within the potential SCN5A promoter region when investigated by transient transfection of promoter-reporter constructs in CHO cells or in neonatal cardiomyocytes [47]. These results may further support a concept of interindividual variability in transcription of this cardiac ion channel gene and arrhythmogenesis.
Population–based studies for SCD are ongoing to highlight potential causes among patients with a positive parental history of cardiac arrest [34, 35], but a clearly defined genetic basis is not known to date [36]. In contrast to patients with cardiac dysfunction, in patients without intraventricular conduction defects or a normal cardiac function, QTc prolongation is a non-negligible risk factor for sudden cardiac death independent of age, history of myocardial infarction, heart rate, and drug use. This has been shown in the Rotterdam Study, a prospective population-based cohort study, in which 125 patients died of sudden cardiac death (mean follow-up 6.7 years) and a prolonged QTc interval had a threefold increased risk [36, 48, 49]. So far, first GWAS have been directed to evaluate the role of common genetic variation in modulation of SCD or VF risk [38, 50, 51]. Since the causes and confounding factors for SCD are diverse, it is not unexpected to note that in two GWAS studies reported [38, 50], did not confirm and replicate each other, i.e., the SNP at chromosome 2q24.2 (rs4665058) at the BAZ2B gene locus was not seen in the Dutch case-control set. Similarly, the CXADR gene signal was not detected in the study involving European ancestors; this may be due to several factors, including not only insufficient statistical power and random chance, but also differences in study design (population stratification) and phenotype definition. Future studies, entailing expression quantitative trait locus (eQTL) analysis in cardiac tissue as well as the genome-wide identification by ChiP-Seq of regulatory regions occupied by transcriptional enhancers and transcription factors [52] will shed additional light on the pathophysiology.
Recently, a quantitative influence of ion channel gene variation on the myocellular repolarization has been described in twins [29] and in the general population [27, 28, 30]. Of note, the heritability reflecting the degree of variance in ECG indices between individuals is for the QTc interval in the range of 25–50 % and for the PR interval between 34 and 40 %, at least depending on the population set studied, see [53]. Genomic studies are currently on the way to narrow candidate these gene regions and to identify these variants (SNPs or haplotype constellations) in coding and non-coding sequences. An example has been shown very recently [54]; in the study by Amin and co-workers sequence variance of the 3′-UTR at the LQT-1 (KCNQ1) locus was investigated by microRNA binding sites that may influence KCNQ1 expression of mutant or wild-type allele. As a novel concept, three single nucleotide polymorphisms (rs2519184, rs8234, and rs10798) were associated in an allele-specific manner with QTc and symptom occurrence and, intriguingly, with concordant, but altered gene expression upon luciferase reporter assays. These data raised the idea that clinical disease expression may be a function of the ratio between normal and mutant allele expression and other factors [54].
It has long been surmised that drug–induced torsade de pointes is an acquired condition that may occur in the context of a mutation encoding for a cardiac ion channel gene responsible for repolarization. This was enhanced due to the recognition of LQTS gene mutation carriers with normal or nearly normal ECGs (incomplete or non-penetrance) [55, 56]. First reports on patients with drug-induced QT interval prolongation and LQTS ion channel gene mutations and were reported nearly a decade ago [57–59] and are listed in Table 21.2. In at least 15–20 % of patients LQTS genes mutations can be found [60], even some reports are indicative for a higher ratio [73, 74]. Altogether, these factors further diminish ‘repolarization reserve’ [75] to a critical extent and allow the generation of afterdepolarizations and triggered activity preceding torsade de pointes. The observation that in the majority of patients with idiosyncratic drug reactions and TdP development a LQTS gene mutation cannot be found, is possibly related to undetected mutations in these, predisposing variants or other target genes responsible for cardiac repolarization [11].
Table 21–2
Ion channel gene variants identified in patients with drug-induced QT prolongation or TdP occurrence
Gene | Current | Amino acid alteration | Drug/setting | Functional assay | Minor allele | Reference |
|---|---|---|---|---|---|---|
KCNE2 | IKr | T8E | Trimethoprim (TMX)/sulfa-methoxazole (SMX); quinidine; amiodarone | E8-MiRP1 weakly reduced IKr current peak density; (CHO cells); SMX and TMX had almost no effect on wild-type channels, but SMX was reported to inhibit more than 50 % of A8-MiRP1 at −40 mV. Mutant channels were 4× more sensitive to SMX than wild-type | 1.6 % | |
IKr | Q9E | Clarithromycin, low K+ | Rare in Caucasians, but not in Afro-Americans | |||
IKr | M54T | Procainamide | T54-MiRP1 significantly reduced IKr current peak density (CHO cells). No influence on drug-related channel inhibition was seen | Rare | [61] | |
IKr | M57T | Oxatomide | T57-MiRP1 significantly reduced IKr current peak density (CHO cells). No influence on drug-related channel inhibition was seen | Rare | [61] | |
IKr | A116V | Quinidine | V116-MiRP1 significantly reduced IKr current peak density (CHO cells). No influence on drug-related channel inhibition was seen | Rare | [61] | |
KCNH2 | IKr | M124T | Probucol | Co-expression of wild-type HERG and T124-HERG resulted in markedly smaller amplitudes of IKr (Xenopus oocytes). Probucol decreased the amplitude of the HERG tail current, decelerated the rate of channel activation, accelerated the rate of channel deactivation, and shifted the reversal potential to a more positive value | Rare | [64] |
IKr | R328C | Not reported | Rare | [65] | ||
IKr | P347S | Cisapride/clarithromycin | Rare | |||
IKr | R486H | Quinidine | Rareb | [57] | ||
IKr | A561P | Clobutinol | P561-HERG led to an intracellular trafficking defect; when co-expressed with wild-type HERG, voltage-dependence was shifted towards more negative potentials (3–3.5-mV). Clobutinol further blocked heteromeric channels | Rareb | ||
IKr
 Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
Get Clinical Tree app for offline access
|