KCNH2 pharmacogenomics summary

Overview

The KCNH2 gene, or human ether-a-go-go related gene (hERG), codes for a potassium voltage-gated ion channel [1,2]. The current through the channel is termed the rapid component of the cardiac delayed rectifier (I_Kr). The KCNH2 gene is located on chromosome 7 and has 15 exons. Mutations and variants of KCNH2 are one cause of congenital long QT syndrome (LQTS), a rare syndrome that carries an increased risk of cardiac arrhythmias, including the polymorphic ventricular tachycardia termed torsades de pointes (TdP), which can be fatal [3,4]. There has also been an association between KCNH2 variants and sudden infant death syndrome [5]. Variants in many other genes, including KCNQ1, KCNE2, and SCN5A can cause congenital LQTS. However, the syndrome of drug-induced LQTS is most often caused by the block of the hERG channels encoded by the KCNH2 gene [2–4,6,7]. Other rarer mechanisms for drug-associated QT prolongation and TdP have been reported [6,8]. In addition, other conditions, such as heart block or severe electrolyte abnormalities, can also cause QT prolongation and TdP; collectively, the drug-induced and other forms are termed ‘acquired LQTS’ (aLQTS). For the remainder of this summary, the gene KCNH2 and the encoded protein, hERG, will be used interchangeably.

The hERG channel consists of six helical transmembrane domains, S1–S6 [4]. The structural determinants responsible for channel block have been identified by alanine scanning experiments and homology modeling [9–11]. These studies found that amino acids Y652 and F656 in the pore region of the channel on helix S6 are important for the binding of drugs and inhibitors [9–11].

There are more than 100 reported mutations in the KCNH2 gene related to congenital LQTS. Information on these mutations can be found on several online websites, including: the Online Mendelian Inheritance in Man database webpage for KCNH2 (http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=152427), connections for heart hERG polymorphisms (http://www.fsm.it/cardmoc/hergpoly.htm), connections for heart hERG mutations (http://www.fsm.it/cardmoc/hergmut.htm), and LQTS db hERG mutations (http://www.ssi.dk/graphics/html/lqtsdb/herg.htm). In addition, gene deletions and duplications have been observed in patients with congenital LQTS [12,13].

However, there are very few variants and amino acid changes that have been clearly associated with drug-induced hERG-related LQTS. A number of studies have strongly supported the idea that variation, not only in KCNH2 but also in other cardiac ion channels and associated genes, may predispose individuals to aLQTS [14,15]. In addition, several population studies have reported that KCNH2 haplotypes modulate variability in the QT interval [16–18]. A more detailed discussion of two KCNH2 variants related to drug response can be found later in this summary.

hERG/I_Kr inhibitors

Virtually all drugs that cause drug-induced QT prolongation are KCNH2/I_Kr blockers [4,7]. Eight drugs (astemizole, sertindole, terfenadine, cisapride, grepafloxacin, terodiline, lidoflazine, levomethadyl) have been removed from the market because of the risk of aLQTS and fatal TdP [1,19]; and a ninth, droperidol, has received highly restrictive labeling [1].

As a result of these events, testing for hERG blocking activity and subsequent evaluations for QT interval prolonging potential are routine in the pharmaceutical industry and such screening has resulted in halting the drug development of compounds that exhibit these potentially undesirable effects [4,20].

Inhibitors of hERG/I_Kr include amiodarone [14], astemizole [1,19,21–23] and its metabolite desmethylastemizole [22], cisapride [1,9,19,21,23,24], disopyramide [25], dofetilide [26,27], erythromycin [28,29], fluoxetine [30], grepafloxacin [1,19,21,23], haloperidol half maximal inhibitory concentration (IC₅₀) approximately 63 nmol/l [31], hydroxyzine [32], ibutilide [25]; levomethadyl [1,23], lidoflazine IC₅₀ less than 37 nmol/l [1,31], methadone [33]; mibefradil [23], moxifloxacin [28], perhexiline [29], pimozide IC₅₀ approximately 18 nmol/l [34], prenylamine IC₅₀ approximately 590 nmol/l [31], probucol [35], quinidine [25,29]; risperidone IC₅₀ 167 nmol/l [34], sertindole IC₅₀ approximately 3 nmol/l [34], IC₅₀ approximately 210 nmol/l [31], sotalol [25,29], telithromycin [28], terfenadine IC₅₀ less than 52 nmol/l [31], [1,9,21,23], terodiline [1,19,21], thioridazine IC₅₀ approximately 191 nmol/l [34], IC₅₀ approximately 224 nmol/l [36], ziprasidone IC₅₀ approximately 169 nmol/l [34].

Weak inhibitors of hERG/I_Kr (IC₅₀ >1 μmol/l) include arsenic trioxide I_Kr approximately 300 μmol/l [31], chlorpheniramine IC₅₀ approximately 13 μmol/l [31], cimetidine IC₅₀ greater than 10 μmol/l [31], doxepin IC₅₀ approximately 4 μmol/l [37], loratadine IC₅₀ approximately 4 μmol/l [31], lovastatin IC₅₀ approximately 7 μmol/l [31], olanzapine IC₅₀ approximately 6013 nmol/l [34], pentamidine Iherg approximately 1 mmol/l [31], procainamide IC₅₀ approximately 139 μmol/l [38], pyrilamine IC₅₀ approximately 6 μmol/l [31], quetiapine IC₅₀ approximately 5765 nmol/l [34], sparfloxacin (fluoroquinolone) IC₅₀ approximately 18 μmol/l [34,39].

Drugs that prolong QT interval by reducing cell surface KCNH2 expression include pentamidine [40], arsenic trioxide [41].

KCNH2 variants and their functional consequences

The discussion below focuses on two well-studied variants related to drug-induced LQTS: KCNH2: K897T (Lys897Thr); rs1805123; KCNH2:1670A > C and KCNH2: R1047L (Arg1047Leu); rs36210421; KCNH2:2120G > T.

Minor allele frequencies for both variants are reported in Table 1 at end of the summary.

Table 1.

Table of minor allele frequencies for variants

	Those with aLQTS [14]	Those without aLQTS [14]	TN-control [14]	NHGRI-control [14]	African–American with SIDS [42]	African–American control [42]	Caucasian with SIDS [42]	Caucasian control [42]	Danish population [43]
K897T (%)	14	16	12	13	5.9	4	17.8	24
R1047L (%)					0	0.5	4.2	1.5	4

aLQTS, acquired long QT syndrome; NHGRI, National Human Genome Research Institute; SIDS, sudden infant death syndrome; TN, Tennessee (population).

KCNH2: K897T (Lys897Thr); rs1805123; KCNH2:1670A > C

K897T (rs1805123) has been shown, in several studies, to be associated with longer [17,44], or shorter QT intervals [45–47]. K897T was also shown to create a phosphorylation site that inhibited channel activity, independent of drug binding [48]. But, in another small study, the impact of common KCNH2 polymorphisms, including K897T as well as P967L, R1047L (rs36210421) and Q1068R were found to have no significant differences in cisapride IC₅₀ values or Hill coefficients (compared with wild type) [42]. The K897 allele has been associated with higher incidence of atrial fibrillation, in a study conducted without drugs [49].

KCNH2: R1047L (Arg1047Leu); rs36210421; KCNH2:2120G > T

This variant has been implicated in sensitivity to the I_Kr blocker dofetilide [26]. However, in in vitro studies using variant protein, sensitivity to another blocker, cisapride, was similar to wild-type protein [42]. Another study showed that the R1047L mutation impaired K + current density [50]. Additional information is available at http://www.pharmgkb.org/search/annotatedGene/kcnh2/