Background
Atrial fibrillation (AF) is the most prevalent sustained cardiac arrhythmia. It is responsible for considerable morbidity and mortality, and its population prevalence has reached epidemic proportions, affecting almost seven million patients in the European Union and the USA combined [
1‐
4].
In most cases AF is associated with cardiac risk factors such as hypertensive, ischemic, and/or structural heart disease [
1,
5]. However, 10-20% of patients suffering from AF are younger than 60 years of age and lack the traditional risk factors for AF. These patients are considered as having "lone" AF [
2]. The mechanisms underlying AF are not fully understood, but a heterogeneous model based on the interaction of multiple substrates and triggers is thought to underlie the pathophysiology of the disease. However, early-onset lone AF has been suggested to be a primary electrical disease caused by disturbances in ionic currents. Of note, a genetic cause of these types of electrical disturbances is becoming increasingly recognized [
6].
Identification of the genetic components of AF, and the importance of single nucleotide polymorphisms (SNPs) was recently shown in genome-wide association studies indicating that common variants also play a role in the development of AF [
7,
8]. This association between SNPs and AF was strongest in patients diagnosed at a younger age. There is evidence that variations in genes encoding ion channel subunits are associated with familial predisposition for AF. Several genetic reports have revealed mutations associated with AF in cardiac ion channels and accessory subunits [
6]. Most of these studies show that gain- or loss-of-function mutations in the genes encoding proteins contributing to cardiac depolarization, e.g.
SCN1-3B (involved in I
Na) [
6,
9], or cardiac repolarisation, e.g.
KCNQ1 (I
Ks),
KCNH2 (I
Kr),
KCNJ2 (I
K1) can lead to increased susceptibility to AF [
6]. These results support the two current conceptual models for AF. The first one being that cardiac action potential shortening functions as a substrate for re-entry wavelets in the atria [
10,
11] the second one proposing that a prolonged effective refractory period enhances the propensity for early after depolarization, and thereby increasing the susceptibility to AF [
12].
K
V7.1, the α-subunit of the I
Ks current, has repeatedly been associated with AF [
6]. Co-expression of its regulatory β-subunit KCNE1 changes the biophysical properties of the K
V7.1 channel dramatically [
13]. We hypothesized that early-onset lone AF is associated with mutations in
KCNE1.
Methods
An expanded Methods section is available in Additional file
1.
Study subjects
Consecutive patients with lone AF and onset of AF before 40 years (i.e. absence of clinical or echocardiographic findings of other cardiovascular diseases, hypertension, metabolic or pulmonary diseases) were included from eight hospitals in the Copenhagen region of Denmark [
9]. Healthy controls (216) were recruited from blood donors. The study conforms to the principles outlined in the Declaration of Helsinki and was approved by the Scientific Ethics Committee of Copenhagen and Frederiksberg (Protocol reference number KF 01313322). All included patients gave written informed consent.
Mutation screening
Genomic DNA was extracted from blood samples using the QIAamp DNA Blood Mini Kit (QIAGEN, Hilden, Germany). Oligonucleotide primers for exons and splice junctions were designed using the known sequence of human KCNE1 [Genbank:NG_009091.1]. All primers were designed with M13 tail sequences. DNA fragments amplified by Touchdown PCR were analyzed using a high-resolution melting curve analysis (Light Scanner, Idaho Technology, UT, USA). Fragments with melting curves differing from the curves of wild-type DNA were purified and directly sequenced using M13 primers and Big Dye chemistry (DNA analyzer 3730, Applied Biosystems, CA, USA). The identified variants were validated by the resequencing of a second PCR product.
A group of 216 ethnically matched healthy controls was screened employing high resolution melting curve analysis (Light Scanner, Idaho technology, Salt Lake City, USA), and bidirectional sequencing of genes previously associated with AF was performed (Additional file
1). A mutation was considered suspected being disease causing if criteria previously defined were met [
9].
Molecular biology
Site-directed mutagenesis introducing the mutations G25V (c.74 C > T) and G60D (c.179 G > A) into human KCNE1 cDNA [Genbank:NM_000219.3] and
in vitro transcription were performed using standard procedures. For a detailed description please refer to the Additional file
1.
Heterologous expression studies
We employed two-electrode voltage-clamp experiments using
Xenopus laevis oocytes expressing wild-type or mutant I
Ks and patch-clamp experiments using mammalian cells. A detailed description is available in Additional file
1.
Data analysis
Data analysis was performed with Igor Pro (Wavemetrics, Lake Oswego, OR, USA) and GraphPad Prism (GraphPad Software Inc., San Diego, CA, USA). I/V-curves were constructed by measuring the current at the end of a voltage-step to potentials ranging from -100 to +60 mV for TEVC recordings or +40 mV for patch-clamp experiments, respectively. The data was plotted against the corresponding membrane potentials. Similarly, peak tail-currents, measured at -120 mV for TEVC or -40 mV for patch-clamping experiments following the depolarizing step, were plotted against the membrane potential of the depolarizing step to construct the activation curves. A Boltzmann function (I/I
max
= I
min
+ (I
max
-I
min
)/(1 + exp((V50-V)/k))) was fitted to the activation curves to obtain the potential of half-maximal activation (V50) and the slope factor (k). Activation of KV7.1/KCNE1 channels results in sigmoidal activation current traces. In order to compare the activation kinetics we determined the time needed to reach the half-maximal current level (t1/2). An estimate of the values of the time constants of channel deactivation (τ) was obtained by fitting a mono-exponential function to tail-current traces measured at -140 to -40 mV following a voltage-step to +40 mV for TEVC or 0 mV for patch-clamp experiments.
The frequency dependence of conduction was investigated using three different voltage protocols (60, 120, and 180 bpm). The amount of charge conducted by the IKs channel complex was calculated by integrating the area under the curve in the first 130 ms after the capacitive spike (10 ms) of the pulse at the 7th second. At this time point the current amplitude had reached a steady-state for all pacing frequencies tested. The amount of charge carried was normalized to the charge carried at 60 bpm.
Data are represented as mean ± SEM, unless otherwise indicated. Unpaired t-tests or ANOVA followed by Tukey's method of multiple comparisons were used as appropriate to compare the wild-type and mutated IKs channel complex. P-values below 0.05 were considered statistically significant.
A detailed description of all methods used is available in Additional file
1.
Discussion
Although AF is the most common cardiac arrhythmia, the fundamental molecular pathways in many cases remain undefined. To our knowledge, the present study is the first to report that mutations in the β-subunit KCNE1 are associated with AF.
Both mutations (G25V, G60D) were absent in 216 matched controls, in publicly available databases and in 2276 exoms from the Popgen population NHLBI Exome Sequencing Project
http://evs.gs.washington.edu/EVS/ supporting that these variants are not rare polymorphisms but disease causing mutations [
20]. The patient carrying mutation G25V had no family history of arrhythmia. The patient carrying mutation G60D had an interesting phenotype of AF and borderline QT
c interval that co-segregated from the mother to the son.
We used the Xenopus laevis oocyte system as it allows for better control of subunit ratios compared to transfection of mammalian cell lines such as CHO-K1. In X. laevis oocytes, KCNE1-G25V and KCNE1-G60D showed a gain-of-function for IKs both with respect to steady-state current levels, kinetics, and heart rate-dependent modulation of IKs (for G60D). When investigating the mutation in an experimental condition mimicking the heterozygous state of the patient, changes were not significant and hence the functional studies did not fully explain the phenotype, which could indicate that either the mutation is not causing the phenotype, or that the interaction is more complex. Yet, we showed gain-of-function effects for both identified mutations strengthening the notion that mutations in KCNE1 are associated with AF.
Furthermore, it has been shown that monoallelic expression is much more widespread than previously thought affecting 20% of human genes [
21]. In a study addressing 190 genes on chromosome 21,
KCNE1 was found to have 10 CG methylation sites rendering
KCNE1 a profoundly epigenetically regulated gene [
22] thereby making also the homozygote experimental condition highly relevant.
Though differences were less pronounced in mammalian cells, we still observed a gain-of-function of activation kinetics for G60D. Mutations leading to even mild gain-of-function of the I
Ks current have been described earlier in the context of lone AF, yet all mutations identified so far reside in the α-subunit of the channel complex [
18,
23]. Two studies have investigated a possible association between AF and the SNP G38S in the β-subunit KCNE1 in Chinese AF cohorts. Lai et al. reported a significant association of 38 G, which was not found in the later study [
24,
25]. Of note, others observed decreased I
Ks amplitudes and reduced surface expression with 38 G pointing to a loss-of-function of this variant [
26].
The mutation G25V is located in the extracellular N-terminus of the channel protein. A recent study as implicated the extracellular juxtamembranous region of KCNE1 in gating [
27], however, little is known about the role of the proximal KCNE1 N-terminus. The mutation G60D resides in the transmembrane segment (TMS) of KCNE1 in close proximity of residues 57-59 ("the gating triplet") that are critical for the modulation of K
V7.1 channels [
28]. Several glycine residues located in the TMS including G60 seem instrumental for forming a curvature that locates threonine 58 [
29], which is critical for slow activation kinetics conferred by KCNE1 [
30]. Hence, the mutation G60D in this close proximity could be speculated to compromise the function of this critical amino acid and thereby speed up activation [
30‐
33].
Gain-of-function mutations of the I
Ks channel are expected to increase the repolarising potassium currents which could abbreviate the cardiac action potential duration as well as the effective refractory period in cardiomyocytes. The mutation could thereby create a profibrillatory substrate within the atrium [
10,
11]. I
Ks also plays a major role in ventricular repolarization. Gain-of-function of I
Ks would be expected to result in both action potential and QT
c shortening which we did not observe in the proband. It has been suggested that common variations in other genes may protect the patient from a shortening of QT
c [
31]. All mutation carriers in this study had not experienced ventricular arrhythmias, yet display a borderline long QT
c interval. Pai and Rawles suggested a link between AF and prolongation of the mean QT interval [
32]. The first study associating gain-of-function mutations in K
V7.1 with AF analyzed a large four-generation family. Nine of 16 family members diagnosed with autosomal dominant hereditary AF showed prolonged QT
c ranging from 450 to 530 ms [
33]. Vice versa, Ackermann and colleagues reported presence of early-onset AF in a cohort of congenital LQT patients underlining the evidence for a link between these two syndromes [
34]. Very recently, also a Brugada syndrome associated gain-of-function mutation in K
V4.3, the α-subunit underlying the transient outward current I
to, has been linked with QT
c prolongation in the affected family member [
35].
Despite the fact that proband 2 and his mother, both carrying the mutation G60D, had borderline long QT
c intervals, there was no further suspicion of LQTS in the two patients. Clinical evaluation revealed no family history of syncope or near-syncope and the proband had a one year implantable loop recorder without any ventricular arrhythmias detected. Of note, both our index probands had an ECG pattern of IRBBB, which has recently been shown to be associated with early-onset lone AF [
15] supporting the AF phenotype in the patients.
The exact mechanisms linking AF and QT prolongation and the different effect of I
Ks mutations remain to be elucidated. One possible explanation may be different composition of the I
Ks channel complex in atria and ventricles as suggested earlier to explain a mixed phenotype of AF and QT prolongation in a patient with a
KCNQ1 mutation [
18]. Also, chamber-specific interaction partners yet to be identified may modulate the effects of mutations in atria and ventricles. Furthermore, differences in triggers such as the well-documented beta-adrenergic stimulation [
36] or the more recently described modulation by natriuretic peptide precursor A may be involved [
23,
37].
Limitations
We limited our analysis to the KCNE1 encoding regions, and the possibility of mutations occurring in regions of the gene other than coding regions cannot be excluded. Though we performed genetic testing of all genes associated with AF earlier, we cannot exclude mutations in yet unknown genes. Furthermore, genetic testing of family members was limited as some were unavailable. Also, the number of probands was small, however, it should be noted that the cohort of young lone AF patients was well-defined. The functional analyses used conventional heterologous expression systems in which the environments differ from that in the native cardiomyocytes.
Acknowledgements
We are grateful to the patients and the cardiology departments at Frederiksberg, Hvidovre, Gentofte, Glostrup, Amager, Herlev and Bispebjerg hospitals for help with supply of patient data. We thank Pia Hagman and Amer Mujezinovic for technical assistance.
Funding sources
The Danish National Research Foundation, The John and Birthe Meyer Foundation, The Research Foundation of the Heart Centre Rigshospitalet, The Arvid Nilsson Foundation, and Director Ib Henriksens Foundation.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MSO performed mutation screening, participated in the study design and wrote the manuscript. BHB participated in the study design, performed heterologous expression studies for G60D and the statistical analysis, and drafted the manuscript. JBN acquired clinical data and drafted the manuscript. ABS and JPD performed heterologous expression studies for G25V. JJ participated in the mutation screening. HKJ, SH and JHS acquired clinical data and helped to draft the manuscript. NS performed molecular biology and sequence alignment, conceived the study, participated in its design and coordination and wrote the manuscript. All authors read and approved the final manuscript.