Multiple arrhythmic syndromes in a newborn, owing to a novel mutation in SCN5A

Abstract

Background

Mutations in the SCN5A gene have been linked to Brugada syndrome (BrS), conduction disease, Long QT syndrome (LQT3), atrial fibrillation (AF), and to pre- and neonatal ventricular arrhythmias.

Objective

The objective of this study is to characterize a novel mutation in Na_v1.5 found in a newborn with fetal chaotic atrial tachycardia, postpartum intraventricular conduction delay, and QT interval prolongation.

Methods

Genomic DNA was isolated and all exons and intron borders of 15 ion-channel genes were sequenced, revealing a novel missense mutation (Q270K) in SCN5A. Na_v1.5 wild type (WT) and Q270K were expressed in CHO-K1 with and without the Na_vβ1 subunit.

Results

Patch-clamp analysis showed ~40% reduction in peak sodium channel current (I_Na) density for Q270K compared with WT. Fast and slow decay of I_Na were significantly slower in Q270K. Steady-state activation and inactivation of Q270K channels were shifted to positive potentials, and window current was increased. The tetrodotoxin-sensitive late I_Na was increased almost 3-fold compared with WT channels. Ranolazine reduced late I_Na in WT and Q270K channels, while exerting minimal effects on peak I_Na.

Conclusion

The Q270K mutation in SCN5A reduces peak I_Na while augmenting late I_Na, and may thus underlie the development of atrial tachycardia, intraventricular conduction delay, and QT interval prolongation in an infant.

Introduction

The cardiac sodium current (I_Na) is responsible for the rapid depolarization of cardiomyocytes as well as for the propagation of the action potential throughout the heart. The current inactivates rapidly, although a small fraction of channels may reopen during the plateau phase of the cardiac action potential, resulting in a persistent or late inward current (I_NaL) (Maltsev et al. 1998). I_NaL is important in determining the duration of the cardiac action potential (Attwell et al. 1979).

I_Na is carried by the pore forming α-subunit, Na_v1.5, and the biophysical properties of Na_v1.5 are regulated by multiple proteins, such as the β-subunits (Na_vβ1–4) and anchoring proteins, all of which form a macromolecular complex. Some of these proteins also regulate targeting and trafficking of the macromolecular complex to the cell membrane (reviews by Meadows and Isom 2005; Abriel 2010). Several arrhythmia syndromes have been linked to mutations in genes encoding Na_v1.5 or in the ancillary subunits that contribute to I_Na. Brugada syndrome (BrS), conduction disease, and sick sinus syndrome have been linked to a reduction in I_Na (review by Zimmer and Surber 2008). Mutations resulting in a gain of function have been linked to Long QT Syndrome Type 3 (LQT3), as well as to prenatal and ventricular arrhythmias in newborns (Wang et al. 2008) and sudden infant death syndrome (SIDS) (Ackerman et al. 2001; Wedekind et al. 2001; Schwartz et al. 2002; Valdivia et al. 2002; Arnestad et al. 2007; Wang et al. 2007).

These syndromes may overlap in the same patient or among different members of a family carrying the same mutation (Bezzina et al. 1999; Veldkamp et al. 2000; Kyndt et al. 2001; Grant et al. 2002; Remme et al. 2008). Variations in phenotype expression are thought to be attributable to the severity of the disease-causing mutation as well as the possible co-existence of other genetic variations, including single nucleotide polymorphisms (SNP), which are not disease-causing, but which can affect arrhythmia susceptibility. For example, the SNP K897T in the K_V11.1 channel has been demonstrated to increase arrhythmia susceptibility in carriers of K_V11.1 mutations with low penetrance (Crotti et al. 2005; Nof et al. 2010).

Here, we describe a newborn with severe atrial tachycardia, conduction abnormalities, and QT interval prolongation. Genetic analysis revealed a substitution of a glutamine (Q) to a lysine (K) at position 270 (Q270K) in the 5th transmembrane segment (S5) of domain 1 (D1) of the Na_v1.5 sodium channel. The Q270K mutation resulted in a reduction in peak I_Na but a gain of function in late I_Na, compared with wild-type (WT), and thus is suspected to generate the substrate for both the slowed conduction and the LQT3 phenotypes observed in this patient.

Materials and methods

Genetic analysis

Genomic DNA was prepared from peripheral blood lymphocytes of the patient and available family members. All known exons of SCN5A, SCN1B, SCN3B, KCNH2, KCNQ1, KCNJ2, KCNE1, KCNE2, KCNE3, KCND3, KCNIP2, KCNJ11, CACNA1C, CACNB2b, and CACNA2D1 were amplified with intronic primers and sequenced in both directions. No mutations were detected except in SCN5A. The mutation was absent in 200 unrelated control individuals with the same ethnic background. All clinical and genetic studies conformed to the principles outlined in the Declaration of Helsinski and written informed consent was obtained according to protocols approved by the local Institutional Review boards.

Cloning

Mutagenesis was performed on a plasmid containing cDNA for hNa_v1.5 that refers to isoform 2 (NM_000335) and is known as hH1c (i.e., it lacks Q1077). Of note, nomenclature of patient mutations refers to isoform 1. The point mutation Q270K (CAG → AAG) in hNa_v1.5 was introduced using mutated oligonucleotide extension (PfuTurbo Polymerase, Stratagene, La Jolla, California, USA) from a plasmid template (hNav1.5 in pcDNA3.1), digested with DpnI (Fermentas, St. Leon-Roth, Germany) and transformed into E. coli XL1 Blue cells. Na_vβ1 in pcDNA3.1 was obtained by subcloning from pIRES-hβ1-CD8: a kind gift of H. Abriel (University of Bern, Switzerland). All constructs were verified by DNA sequencing of the cDNA insert (Macrogen, Inc., Seoul, Korea).

Electrophysiology

Chinese hamster ovary (CHO-K1) cells were transfected with 2 μg hNa_v1.5, hNa_v1.5_Q270K, or hNa_v1.5 + hNa_v1.5_Q270K in a 1:1 molar ratio. For co-expression of hNa_v1.5 or hNa_v1.5_Q270K with hNa_vβ1, 1 μg Na_v1.5 (WT or Q270K) was mixed with 1 μg hNa_vβ1; eGFP (0.2 μg) was included to identify transfected cells. Transfections were performed using Lipofectamine and Plus reagent (GIBCO, Invitrogen, Carlsbad, Calif.) according to the manufacturer’s instruction. Cells were kept in Dulbecco’s modified Eagle’s medium (University of Copenhagen, Denmark) supplemented with 10% fetal calf serum (GIBCO) and 40 mg/L l-proline at 37 °C in 5% CO₂. Currents were recorded 2–3 days post-transfection using a MultiClamp 700B amplifier and MultiClamp Commander (Molecular Devices, Axon Instruments, Sunnyvale, Calif.). Cells were superfused with a solution containing (in mmol/L): 130 NaCl, 5 KCl, 1.8 CaCl₂, 1.0 MgCl₂, 2.8 Na acetate, 10 HEPES, 10 glucose, pH adjusted to 7.3 with NaOH. Patch pipettes were fabricated from borosilicate glass capillaries (Module Ohm, Herlev, Denmark) and had resistances between 1.5–2 MΩ when filled with pipette solution (mmol/L): 10 KCl, 105 CsF, 10 NaCl, 10 HEPES, 10 EGTA, and 5 TEA–Cl, pH adjusted to 7.2 with CsOH. Electronic compensation of series resistance to 70%–85% was applied to minimize voltage errors. All analog signals were acquired at 10–50 kHz, filtered at 6 kHz, digitized with a Digidata 1440 converter (Axon Instruments), and stored using pClamp10 software (Axon Instruments). All recordings were made at room temperature (20–22 °C). Data were analysed using pClamp10 software (Axon Instruments) and Prism 4.03 (GraphPad Software, Inc., La Jolla, Calif.). Mean ± SEM values are shown, statistical significance (*, p < 0.05) was evaluated as appropriate by one-way ANOVA followed by Student–Newman–Keuls post test, or by paired or unpaired Student’s t test.

Drugs

Tetrodotoxin (TTX, 10 μmol/L) was added from a 100 mmol/L stock of TTX-citrate (Tocris Bioscience, Bristol, UK) dissolved in water that was double-distilled using the Milli-Q system (Millipore Corp., Bedford, Massachusetts, USA). Ranolazine (Gilead, Palo Alto, Calif.) was made as a stock solution of 10 mmol/L DMSO, and used at a final concentration of 50 μmol/L.

Results

Case presentation

Owing to polyhydramnios and fetal tachycardia, the female patient was prematurely born at week 34 by Cesarian section. Her birth weight was 2.04 kg. Fetal tachycardia was first diagnosed at 29 weeks gestation, and best characterized by fetal echocardiography as nonsustained but incessant chaotic atrial tachycardia or atrial fibrillation, with atrial rates as rapid as 500–600 beats per minute (bpm) and variable ventricular response (Fig. 1A). The irregularity of the atrial rhythm was not consistent with fetal atrial flutter. Runs were interrupted by brief periods of sinus rhythm at 140 bpm. The rhythm was controlled with maternal enteral sotalol (120 mg twice daily). Subsequent fetal echocardiograms demonstrated mostly sinus rhythm (Fig. 1B), with short bursts of regular atrial tachycardia at rates of up to 480 bpm with variable atrioventricular (AV) conduction. The patient’s mother was asymptomatic, and her first pregnancy was carried to full term without any problems.

Fig. 1.

(A) Two-dimensional fetal echocardiogram at 29 weeks gestation showing a frame from sector scan (above) and accompanying m-mode (below). Illustrated are mechanical atrial events at a very rapid and irregular rate (indicated by upward pointing arrows) and slower mechanical ventricular events (indicated by downward pointing arrows). (B) Fetal echocardiogram using the same format and obtained 3 days later after administration of sotalol to the mother, illustrating a normal atrial rate (indicated by upward pointing arrows) and a sinus pause. A, atrium; V, ventricle.

The patient’s initial rhythm within moments of delivery was atrial flutter, which degenerated into ventricular fibrillation (Figs. 2A and 2B). During atrial flutter, the ventricular rate was within the normal range (144 bpm) owing to 3:1 AV block. Thirty minutes post-natally, and following defibrillation and a single intravenous dose of amiodarone (5 mg/kg), the patient became asystolic. Successful resuscitation included transesophageal atrial pacing, owing to persistent bradycardia. There was intraventricular conduction delay with QRS duration as long as 119 ms during both atrial pacing and normal sinus rhythm, up to 7 days post-natally, and well beyond plausible effects of transplacental sotalol and the single small intravenous dose of amiodarone (Fig. 2C). This observation was in the presence of β-adrenergic agents, during normal electrolyte and acid-base status and after marked improvement in ventricular function by echocardiography. Echocardiography also revealed a bicuspid aortic valve. The patient’s initial course was further complicated by pulmonary hemorrhage and the need for oscillatory assisted ventilation, also by bifrontal subdural, interventricular, and periventricular hemorrhages, as well as hemodynamic instability, requiring multiple vasopressor agents (intra venous dopamine, epinephrine, and vasopressin).

Fig. 2.

(A) Twelve lead ECG of the proband soon after birth showing atrial flutter. (B) This deteriorated into ventricular fibrillation, requiring defibrillation. (C) Sinus rhythm on the 5th day of life, demonstrating intraventricular conduction delay. (D) During treatment with amiodarone on the 10th day of life, there was prolongation of the QTc interval (491 ms in V5 during period of stable sinus rhythm). (E) Also on the 10th day of life, following a short–long–short QRS sequence, torsade de pointes occurred. (F) Twelve lead ECG at 23 months of age while drug-free, showing normal QRS duration and QTc interval prolongation (466 ms). Pacing spikes account for pseudofusion.

Despite the placement of temporary epicardial atrial pacing wires by limited sternotomy and attempts at overdrive atrial pacing, recurrent episodes of hemodynamically unstable atrial tachycardia required DC cardioversion while the patient was 1–7 days old. Intravenous amiodarone was loaded while the patient was 7–10 days old, for a total of 30 mg/kg, but resulted in marked QTc prolongation (Fig. 2D). Despite normal acid-base status and serum electrolytes, torsade de pointes (TdP) following long-short ventricular sequences (Fig. 2E) occurred on 4 occasions when the patient was 10-12 days old, requiring discontinuation of amiodarone. Quinidine-sulfate (10 mg/kg every 6 h) was administered enterally. Following the first dose of quinidine, she had no further atrial or ventricular arrhythmias. To limit future pause-dependent TdP, a permanent dual-chamber pacemaker was implanted on day 17, and she was discharged from the hospital when she was 37 days old. At 23 months of age, the patient had experienced no episodes of syncope, seizures, or other signs of paroxysmal tachycardias. Her parents had discontinued quinidine at approximately 6 months of age. Twelve lead electrocardiogram in the drug-free state showed QT interval prolongation with a QTc interval of 466 ms (Fig. 2F), but normalization of the intraventricular conduction delay. In summary, this infant displayed 3 clinical electrophysiologic abnormalities, which were thought to be independent of antiarrhythmic drug effects and metabolic disturbances: fetal irregular atrial tachycardia or atrial fibrillation, post-natal intraventricular conduction delay, and QT interval prolongation

Analysis of the patient’s DNA showed a heterozygous C to A transition at nucleotide position 808 in exon 7 in SCN5A (Fig. 3A). This transition predicted a substitution of a highly conserved glutamine (Q) to a lysine (K) at position 270 (Q270K) in S5 of domain I of Na_v1.5 (Fig. 3B).

Fig. 3.

Electropherogram of the patient’s DNA showing a heterozygous C to A transition at nucleotide position 808 in exon 7 in SCN5A (A), resulting in a substitution of glutamine (Q) to a lysine (K) at position 270 (Q270K) in S5 of the first domain (D1) of the Na_v1.5 sodium channel, marked by a dot in (B).

Electrophysiological characterization of Nav1.5_Q270K

WT Na_v1.5, Q270K or WT/Q270K (1:1 ratio to mimic the heterozygote state) were expressed in CHO-K1 cells. Both WT and Q270K channels exhibited a fast activating and inactivating current (Fig. 4A). However, Q270K had a reduced peak current density (Fig. 4B) and double exponential fits to the current decay showed that both τ_slow and τ_fast were significantly slower for Q270K (Fig. 4C). The relative weights of the pre-exponential factors for τ_fast/τ_total were calculated (Fig. 4D), showing that the relative contribution of the fast component was reduced for Q270K. For WT/Q270K an intermediate phenotype was found.

Fig. 4.

Whole-cel1 current recordings for Na_v1.5 wild type (WT) (n = 8), WT/Q270K (1:1 ratio, n = 6), and Q270K (n = 8). (A) Representative recordings. (B) Peak current density as a function of voltage. The asterisks indicate significant difference between WT and Q270K: *, p < 0.05; **, p < 0.01; and ***, p < 0.001. The difference between WT and WT/Q270K did not reach statistical significance. (C) Double exponential functions fit to the current decay and the resulting time-constants (τ_slow and τ_fast) plotted as a function of voltage. (D) The relative weight of the pre-exponential factors for τ_fast/τ_total as a function of voltage.

Steady-state activation curves were calculated by normalizing the currents shown in Fig. 5 to (V_m–E_K). The V_1/2 of steady-state activation was significantly shifted to more depolarized values for Q270K compared with WT, with V_1/2 = −34.4 ± 0.2 mV (n = 8) for WT, and V_1/2 = −28.6 ± 0.3 mV (n = 7) (p < 0.0001) for Q270K (Fig. 5A). Steady-state inactivation was addressed by a prepulse-pulse protocol (Fig. 5A). For WT, the V_1/2 of inactivation was −71.2 ± 0.3 mV (n = 9) and for Q270K significantly shifted to V_1/2 = −61.3 ± 0.8 mV (n = 6, p < 0.0001). The shift in V_1/2 for activation and inactivation resulted in larger window currents (magnified part of Fig. 5A). Within this window of voltages, a small percentage of channels are not inactivated and are thus available for activation.

Fig. 5.

Activation and steady state-inactivation relation for Na_v1.5 wild type (WT) (n = 9) and Q270K (n = 6). (A) Activation curves were calculated by normalizing the currents shown in Fig. 4 to (V_m–E_K). Steady-state inactivation was determined using the protocol shown in the inset. Peak currents at −20 mV were normalized and plotted against the conditioning potential, and Boltzmann equations were fit to the data points. The overlap regions of the 2 relationships are enlarged. (B) Time-dependent recovery from inactivation for WT and Q270K. Currents were activated by a 2-pulse protocol from a holding of either −120 mV, −100 mV, or −80 mV. The fraction of recovered current was plotted as a function of interpulse interval, and single exponential functions were fitted to the data.

Recovery from inactivation of WT and Q270K was addressed by a 2-pulse protocol (Fig. 5B). At −120 mV and −100 mV, the recovery of Q270K and WT was similar, however at −80 mV, Q270K recovered significantly faster from inactivation than for WT, with τ = 5.8 ± 0.22 ms for WT (n = 8) and 4.8 ± 0.21 ms for Q270K (n = 6, p = 0.0077).

The interaction between Na_v1.5 and the Na_vβ1 has been reported to include the extracellular domain of Na_vβ1 and the extracellular loop between S5 and S6 in D1 and D4 of Na_v1.5 (Makita et al. 1996; McCormick et al. 1999). As the Q270K mutation resides in D1-S5 of Na_v1.5, suggesting that the Q270k mutation could interfere with interaction between Na_v1.5 and Na_vβ1, we repeated the functional studies in the presence of the Na_vβ1 subunit. Na_v1.5 WT, Q270K, or WT/Q270K were co-expressed with Na_vβ1 (1:1 ratio) in CHO-K1 cells. As Na_vβ1 increased current density in agreement with previous reports (Qu et al. 1995; Fahmi et al. 2001; Bezzina et al. 2003; Herfst et al. 2003), only half the amount of Na_v1.5 DNA was used compared with the experiments without Na_vβ1.

Similar to the results for experiments without Na_vβ1, peak current density was reduced for Q270K/Na_vβ1 compared with WT/Na_vβ1 (Fig. 6A), and WT/Q270K/Na_vβ1 had an intermediate phenotype (data not shown). The current decay (both τ_slow and τ_fast) was significantly slower for Q270K/Na_vβ1 over a range of voltages (Fig. 6B), and for Q270K/Na_vβ1 the relative weight of the fast component was lower (Fig. 6C).

Fig. 6.

Whole-cell current recordings for Na_v1.5 wild type (WT) (n = 11) and Q270K (n = 10) co-expressed with Na_vβ1 (1:1 ratio). (A) Representative currents and peak current densities shown as a function of voltage. (B) Double exponential functions were fit to current decay and the time-constants (τ_slow and τ_fast) are shown as a function of voltage. (C) Relative weight of the pre-exponential factors. (D) Activation curves were calculated by normalizing the currents shown in A to (V_m–E_K). Steady-state inactivation was measured and peak currents at −20 mV were normalized, plotted against the conditioning potential, and Boltzmann equations were fit to the data points. The overlap regions are enlarged. (E) Time-dependent recovery from inactivation for WT/ Na_vβ1 and Q270K/Na_vβ1. Currents were activated by a 2-pulse protocol from a holding of either −120 mV, −100 mV, or −80 mV. The fraction of recovered current was plotted as a function of interpulse interval and single exponential functions were fit to the data. *, p < 0.05; **, p < 0.01; and ***, p < 0.001.

Similar to the results in the absence of Na_vβ1, a significant shift to more positive potentials was observed in V_1/2 for activation and inactivation for Q270K/Na_vβ1 compared with WT/Na_vβ1. The V_1/2 of activation was −31.4 ± 0.4 mV for WT/Na_vβ1 (n = 10) and −26.7 ± 0.4 mV for Q270K/Na_vβ1 (n = 9, p < 0.0001). V_1/2 of inactivation was −77.7 ± 0.6 mV for WT/Na_vβ1 (n = 9) and −61.2 ± 0.5 mV for Q270K/Na_vβ1 (n = 8, p < 0.0001), (Fig. 6D), resulting in increased window currents. Time-dependent recovery from inactivation for WT/Na_vβ1 and Q270K/Na_vβ1 was investigated using holding potentials of −120 mV, −100 mV, and −80 mV. For all holding potentials, Q270K/Na_vβ1 recovered significantly faster than WT/Na_vβ1 (Fig. 6E).

Our results show that current decay was significantly slower for the Q270K mutation, and that window currents were increased. This prompted us to address the question as to whether Q270K has a larger persistent or late current than WT. WT/Na_vβ1 and Q270K/Na_vβ1 currents were activated by a 1 s step to −45 mV in the absence or presence of 10 μmol/L TTX for both Q270K and WT channels (Figs. 7A and 7B). Currents recorded in presence of TTX were digitally subtracted from controls, and the ratio between TTX sensitive peak and late currents were calculated, revealing a 3-fold increase in Q270K late current compared with WT (Fig. 7C).

Fig. 7.

Effect of 10 μmol/L tetrodotoxin (TTX) on Na_v1.5 wild type (WT) (n = 5) and Q270K (n = 5) co-expressed with Na_vβ1 (1:1 ratio). Currents were activated by a 1 s step to −45 mV from a holding potential of −120 mV at a frequency of 0.25 Hz. (A) Representative WT recordings in absence or presence of TTX. For comparison, the peak and late current are shown at different scales. (B) Representative Q270K recordings. (C) Summarized data of the effect of TTX on current density. Currents in presence of TTX were digitally subtracted from currents recorded in the absence of TTX to determine the TTX-sensitive current. Late current was measured as mean current after 950–990 ms and the ratio between the TTX-sensitive peak and late current was calculated for WT and Q270K. *, p < 0.05 and **, p < 0.01.

As the Q270K mutation affected several gating processes of the Na_v1.5 channel, we measured I_Na using action potential voltage clamp techniques (Magyar et al. 2004). It is well established that I_to is absent in the neonatal ventricle in virtually all mammalian species (Jeck and Boyden 1992; Plotnikov et al. 2004; Shimoni et al. 1997) and it is likely that newborn humans also lack ventricular I_to. In the next next series of experiments, a previously recorded neonatal canine ventricular action potential waveform was used to activate WT/Na_vβ1or Q270K/Na_vβ1 current (Fig. 8A). Holding potentials of either −60 mV or −120 mV were used. At −60 mV, Na_v1.5 was fully inactivated (Fig. 8A) and the recorded current represents endogenous current. The recordings obtained at −6 0 mV were digitally subtracted from the recordings obtained at −120 mV (Fig. 8B). During the initial part of the action potential, the fast activating and inactivating I_Na transient dominated; however, during the plateau phase a persistent small I_Na current was present. Similar to the findings using TTX, the ratio of late current relative to peak current was 3-fold higher for Q270K compared with WT (Fig. 8C).

Fig. 8.

Wild type (WT) (n = 7) and Q270K (n = 6) co-expressed with Na_vβ1 (1:1 ratio). An action potential previously recorded from an isolated neonate canine ventricular cardiomyocyte paced at 1 Hz was used as command (A). (B) The command action potential was modified so that either a holding potential of −120 mV or −60 mV was imposed. (C) The recordings obtained using −60 mV as holding was digitally subtracted from the recordings obtained at −120 mV. For comparison, the peak and late current are shown at different scales. (D) Summary data of peak and late WT/Na_vβ1 and Q270K/Na_vβ1 current, as well as the ratio between late and peak current. Late current was measured as mean current 50 to 55 ms after the peak, corresponding to a potential of 22–23 mV. *, p < 0.05 and **, p < 0.01.

Ranolazine (50 μmol/L) has been shown to preferentially block late Na_v1.5 current, both for WT and for LQT3 mutants (Fredj et al. 2006). We next tested the effect of 50 μmol/L ranolazine on WT/Na_vβ1 and Q270K/Na_vβ1. Currents were activated by a −20 mV depolarization and recorded in the absence and presence of ranolazine (Figs. 9A-9B). Ranolazine had a minor but significant effect on both WT and Q270K peak currents (Fig. 9C). Similar to previous reports (Fredj et al. 2006), we found a relatively larger block of the late current for both WT and Q270K. The currents recorded in the presence of ranolazine were digitally subtracted from controls and the ranolazine-sensitive peak current, and late currents were calculated as well as the ratio between ranolazine sensitive peak and late current (Fig. 9C).

Fig. 9.

Effect of 50 μmol/L ranolazine on wild type (WT) (n = 7) and Q270K (n = 4) co-expressed with Na_vβ1 (1:1 ratio). Currents were activated by −20 mV step from a holding potential of −120 mV at a frequency of 0.25 Hz. (A) Representative WT recordings in the absence or presence of ranolazine. For comparison, the peak and late current are shown at different scales. (B) Representative Q270K recordings. (C) Summarized data of the effect of ranolazine on peak current density and percentage block of peak and late currents. The ranolazine sensitive currents were determined by digitally subtracting currents in presence of ranolazine from control. The late currents were determined as mean currents 45 to 50 ms after the peak and the ratio between the ranolazine sensitive peak and late current was calculated for WT and Q270K. *, p < 0.05; **, p < 0.01; and ***, p < 0.001.

Discussion

In this study, we identified a Q270K substitution in Na_v1.5 in a neonate with severe cardiac arrhythmia, including fetal chaotic atrial tachycardia or fibrillation, intraventricular conduction delay, and prolonged QT intervals. Although the patient’s early history was complicated by pulmonary hemorrhage, intracranial hemorrhage, hemodynamic instability, and treatment with vasopressor agents, the chaotic atrial tachycardia existed in utero prior to maternal antiarrhythmic drug application, and intraventricular conduction delay would not be an expected sequela of those noncardiac co-morbidities in an otherwise normal electrolyte and acid–base state. At age 23 months, the QTc interval was 466 ms after antiarrhythmic drugs had been discontinued.

The 40% reduction of Q270K peak currents may explain the conduction delay. Discriminating intraventricular conduction delay from a true Brugada pattern may be less clear in infants compared with adults, as previously published electrocardiographic examples of BrS in children have not always conformed to strict diagnostic criteria (Priori et al. 2000; Beaufort-Krol et al. 2005). Despite equal genetic transmission of BrS, the clinical phenotype is 8–10 times more prevalent in males than in females, and BrS typically affects individuals in the 4th decade of life (Antzelevitch et al. 2005). Thus, the observation that the patient (a female newborn) exhibited a phenotype consistent with previously reported cases of childhood BrS is interesting. Experimental studies suggest that the presence of a prominent transient outward current (I_to) predisposes the myocardium to the development of the BrS by permitting the expression of a prominent phase 1 (Calloe et al. 2009). However, it is well established that I_to is virtually absent in the neonatal ventricle in all mammalian species studied (Jeck and Boyden 1992; Plotnikov et al. 2004; Trépanier-Boulay et al. 2004), and it is likely that newborn humans also lack I_to. This could account for the absence of a classic coved or saddleback ST segment pattern in this infant’s right precordial leads. This does not explain why quinidine, which is known to be antiarrhythmic in the setting of BrS, likely due to inhibition of I_to, was effective in suppressing the arrhythmias in this patient.

The Q270K mutation is located proximal to the extracellular loop of Na_v1.5 that has been shown to be involved in interaction with Na_vβ1 (Makita et al. 1996; McCormick et al. 1999), however, the effects of Q270K in absence or presence of Na_vβ1 were qualitatively similar, suggesting that the mutation does not interfere with Na_v1.5 and Na_vβ1 interaction.

Besides the reduction of Q270K peak current, inactivation was slower, and the V_1/2 of steady-state inactivation shifted resulted in larger window currents for the mutant. The slower decay of INa and increased window current could suggest that the Q270K mutation is interfering with fast inactivation of Na_v1.5. Fast inactivation of Na_v1.5 involves binding of a cluster of hydrophobic amino acids (the inactivation particle) positioned in the D3–D4 linker (West et al. 1992; Kellen-berger et al. 1997) to domains of S6 in D4 as well as the intracellular loops between D3 S4/S5 (Smith and Goldin 1997), D4 S4/S5 (McPhee et al. 1998), and D4 S6 (McPhee et al. 1995), and thereby closing the pore. Furthermore, inactivation is coupled to activation as movement of the voltage sensing S4 region of D4 generates the signal for inactivation (Chanda and Bezanilla 2002). Previously reported LQT3 mutations cluster in these regions (review by Zimmer and Surber 2008). However, mutations resulting in increased persistent current have been found outside these regions, such as the transmembrane regions (Wang et al. 2007), as well as the extracellular domain of D1 at S216L (Wang et al. 2007). The Q270K mutation resides in transmembrane segment S5 of D1, and how the Q270K destabilizes the inactivated state cannot be easily explained in terms of the current knowledge of sodium channel structure–function relationships, suggesting that other regions of the sodium channel are important for fast inactivation.

Our results demonstrated that current decay was significantly slowed by the Q270K mutation, and that the window current was increased, which prompted us to investigate whether late current was also affected. An increased late current was confirmed by experiments with TTX. Since impedance is high during the action potential plateau (Spitzer et al. 2006), the opening of even a few channels can greatly influence action potential duration. The Q270K mutation significantly increased the fraction of late current, with I_NaL = 0.16 ± 0.03% (n = 6) for WT and 0.54 ± 0.01% (n = 6, p = 0.032) for Q270K. These numbers are in agreement with previous results showing residual current to be 0.15% of the peak current for WT Na_v1.5 channels (Dumaine et al. 1996). Also, in LQT3 mutant channels such as ΔKPQ, which cause a large increase in late currents, the fraction of late currents is still less than 5% (Dumaine et al. 1996).

Several reports have linked SCN5A mutations with increased I_NaL, and (or) increased window currents, to ventricular arrhythmias and SIDS (Ackerman et al. 2001; Lupoglazoff et al. 2001; Wedekind et al. 2001; Valdivia et al. 2002; Chang et al. 2004; Arnestad et al. 2007; Wang et al. 2008; Huang et al. 2009; Ruan et al. 2010). In a Norwegian study, 201 patients diagnosed with SIDS were screened for mutations in genes associated with LQTS. Mutations and rare variants were found in 26 cases: 4 patients harboured mutations in SCNA5, causing an increased I_NaL, and 4 had rare variants in SCNA5, also causing increased I_NaL (Arnestad et al. 2007), suggesting that increased I_NaL is an important genetic risk factor for SIDS. Although the proband initially had a normal baseline QTc of 421 ms, amiodarone treatment increased the QTc interval to 491 ms (Fig. 2D) and triggered several episodes of TdP (Fig. 2E). The torsadogenic effect of intravenous amiodarone during periods of normal electrolyte and acid–base status is very unusual in an infant, but could have been explained by the additional influence of intracranial hemorrhage. However, amiodarone-induced TdP has been associated with an increase in I_NaL in both experimental and clinical studies (Lehtonen et al. 2007; Remme and Bezzina 2007; Wegener et al. 2008; Wu et al. 2008) and the effect of amiodarone to induce TdP in our proband could be associated with the increased I_NaL induced by the Q270K mutation.

Gain of function, including positive shifts in steady-state inactivation of Na_v1.5, have also been linked to lone atrial fibrillation (Makiyama et al. 2008; Li et al. 2009) or atrial fibrillation in association with LQT3 (Benito et al. 2008). The increased I_NaL may cause early after-depolarizations and triggered activity (Benito et al. 2008; Makiyama et al. 2008), and thus may contribute to the development of atrial flutter in the patient.

The increase in persistent current (Fig. 7) prompted us to test the effect of ranolazine on the Q270K mutation. Ranolazine has previously been demonstrated to be a potent blocker of I_NaL induced by LQT3 mutations (Fredj et al. 2006; Bank-ston et al. 2007) and has been found to be effective in antagonizing the proarrhythmic effects of I_NaL in pharmacological models of LQT3 (Song et al. 2004; Wu et al. 2004). Ranolazine has also been shown to be effective in abbreviating QTc in patients with congenital LQT3 (Moss et al. 2008). We found that ranolazine preferentially blocked I_NaL for both WT and Q270K, however this preference was greater for the Q270K compared with WT, suggesting that ranolazine might be of benefit in this setting because of its actions to reduce the LQT3 phenotype without aggravating the conduction slowing produced by this mutation.

Conclusion

The Q270K mutation in Na_v1.5 reduces peak current density while augmenting late I_Na, and thus may underlie the development of neonatal life-threatening arrhythmias characterized by an overlap syndrome consisting of chaotic atrial tachycardia, intraventricular conduction delay, and QT interval prolongation.

Limitations

Although we used a constant molar ratio of multiple genes to ensure proper expression of both gene products, it is difficult to determine whether an individual cell actually expressed the 2 subunits in the same proportion. We assume that a cell incorporates both subunits based on our observations that when eGFP was co-transfected in a 1/10 ratio with the genes of interest, all cells expressing eGFP also expressed I_Na.