Familial atrial fibrillation mutation M1875T-SCN5A increases early sodium current and dampens the effect of flecainide

Abstract Aims Atrial fibrillation (AF) is the most common cardiac arrhythmia. Pathogenic variants in genes encoding ion channels are associated with familial AF. The point mutation M1875T in the SCN5A gene, which encodes the α-subunit of the cardiac sodium channel Nav1.5, has been associated with increased atrial excitability and familial AF in patients. Methods and results We designed a new murine model carrying the Scn5a-M1875T mutation enabling us to study the effects of the Nav1.5 mutation in detail in vivo and in vitro using patch clamp and microelectrode recording of atrial cardiomyocytes, optical mapping, electrocardiogram, echocardiography, gravimetry, histology, and biochemistry. Atrial cardiomyocytes from newly generated adult Scn5a-M1875T+/− mice showed a selective increase in the early (peak) cardiac sodium current, larger action potential amplitude, and a faster peak upstroke velocity. Conduction slowing caused by the sodium channel blocker flecainide was less pronounced in Scn5a-M1875T+/− compared to wildtype atria. Overt hypertrophy or heart failure in Scn5a-M1875T+/− mice could be excluded. Conclusion The Scn5a-M1875T point mutation causes gain-of-function of the cardiac sodium channel. Our results suggest increased atrial peak sodium current as a potential trigger for increased atrial excitability.


Introduction
Atrial fibrillation (AF), the most common cardiac arrhythmia, is characterized by episodes of irregular and uncoordinated atrial electrical activity. It is associated with ischaemic stroke, cardiovascular death, and frequent hospitalizations. 1 A range of different common factors, such as heart failure, diabetes, and increased formation of fibrosis, can damage the atria. 2 These factors interact with a pre-existent, potentially inherited substrate to result in AF. 2 Inherited forms are characterized by early onset of the condition. Pathogenic variants in several genes associated with cardiomyopathies have been identified in familial AF, including variants in sarcomeric and cell-cell contact genes and others. 3 Within the group of ion channel genes, variants leading to dysfunction of the cardiac sodium channel (Na v 1.5) are associated with familial AF. 3,4 The pore-forming α-subunit of the Na v 1.5 is encoded by the SCN5A gene. Variants in SCN5A have been linked to several cardiac conditions, including AF. 4 A missense SCN5A point mutation, Met1875Thr (M1875T), located in the C-terminus of the channel protein, was linked to autosomal dominant familial AF that spanned three generations. 5 Atrial ectopy was evident in mutation carriers in adolescence, and persistent AF occurred as early as 27 years of age. Analyses in the human cell line HEK293 heterologous expression system 5 suggested an enhanced function of the mutated Na v 1.5 channel.
To investigate the impact of the M1875T mutation in the heart and atria, we generated and characterized a novel knock-in murine model (Scn5a-M1875T +/− ). Mice that were heterozygous for the M1875T mutation were viable and studied herein. We investigated hearts of these mice from the whole organ in vivo to the level of the single cardiac cell ex vivo.

Scn5a-M1875T murine model
Mice heterozygous for the knock-in mutation M1875T in the Scn5a gene (Scn5a-M1875T +/− ) were generated by T-C point-mutating exon 28 of the cardiac sodium channel SCN5A gene using pSCN5a_targ3 targeting vector and CRISPR/Cas9 system in murine embryonic stem (ES) cells ( Figure 1A). Mutation-harboring ES cells were characterized using Southern blot analysis and sequenced to exclude genomic rearrangements ( Figure 1B), 6 and injected into B6D2F1 mouse blastocysts. Sequencing analysis of the mutation-containing region using DNA from adult wildtype (WT) and heterozygous Scn5a-M1875T +/− mice on C57Bl/6J × 129sv hybrid genetic background are shown in Figure 1C. Detailed steps of the generation are further explained in the supplement.
Methionine at position 1875 of the human Na v 1.5 protein sequence corresponds to position 1877 of the murine sequence. The latter is therefore point-mutated in this model. However, to underpin the bedside-to-bench nature of this investigation, we use the human annotation and refer to it as 'Scn5a-M1875T' throughout this manuscript.
Mice were bred on an FVB or 129/sv genetic background and housed in individually ventilated cages with sex-matched littermates (2-5 mice/cage), under standard conditions: 12 h light/dark circle, 22°C and 55% humidity. Food and water were available ad libitum. The health status of mice used in the study was monitored daily and prior to experiments.
Functional experiments were conducted on hearts of male and female young adult mice of pure background (8-20 weeks), heterozygous for the knock-in mutation M1875T in the Scn5a gene (Scn5a-M1875T +/− ) and their WT littermates. Mice of both sexes and background were used evenly and analysed jointly, to make the data more widely applicable.

Study approval
All procedures were performed in compliance with the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes and conducted in accordance with rules and regulations for experiments with animals and approved by the UK Home Office (PPL number 30/2967) and by the institutional review board of University of Birmingham.

Electrocardiogram recordings in vivo
Non-invasive electrocardiograms (ECGs) were recorded in conscious young adult mice (8-19 weeks) using a tunnel system for gentle restraint (ecgTunnel, EMKA Technologies, Paris, France). 7 ECG recordings were analysed using ECGauto software (EMKA Technologies, Paris, France). ECGs were also recorded in isoflurane-sedated mice during echocardiography as below.

Echocardiography in vivo
Echocardiography was performed in sedated mice (0.5-2% isoflurane, supplemented with 100% O 2 ) using Vevo® 2100 system (VisualSonics, Amsterdam,  Netherlands) as reported previously. 8 Heart rate was maintained at 450 ± 70 bpm. Left atria (LA) were visualized in the parasternal long axis view in the plane of the aortic root. LA area and diameter were measured during pre-atrial contraction, using the P-wave of the limb ECG trace as a guide.

Histological analysis
Hearts were fixed in formalin and paraffin-embedded tissues were cut into slices of 4 µm. Sections were dewaxed, stained with hematoxylin and eosin for overviews and subsequently dehydrated, embedded and imaged on a NanoZoomer 2.0-HAT (Hamamatsu). Slides used for quantitative analysis were cooked in citrate buffer for antigen retrieval. Autofluorescence was quenched with a 0.25% Sudan black solution for 30 min and samples were blocked with 2% Bovine serum albumin (BSA)/ 2.2% Glycine for 1 h at room temperature (RT). Wheat germ agglutinin (WGA) with Alexa Fluor™ 488 Conjugate (1:200, W11261, Invitrogen) was applied for 2 h at RT.
Quantitative analysis of WGA-stained area and cardiac cell diameters as shown in Figure 2 was carried out using a published ImageJ plugin for atrial histological analysis (JavaCyte 9 ), with minor adjustments.
The LA was dissected free and cardiomyocytes were dissociated gently with fire-polished glass pipettes (2 to 1 mm diameter in sequence). Cells were re-suspended in 2 mL KB buffer and Ca 2+ was gradually reintroduced to the cell suspension incrementally over a period of 2 h to reach a final concentration of 1 mM. All experiments were performed within 8 h of isolation.

Whole-cell patch clamp electrophysiology of isolated atrial cardiomyocytes
Dissociated murine LA cardiomyocytes were plated on, and allowed to adhere to, laminin-coated coverslips (10 mm diameter) for at least 20 min. Coverslips were transferred to a recording chamber and were continually superfused at 3 mL.min −1 , with a low Na + external solution containing in mM; NaCl 10, KCl 4.5, C 5 H 14 CINO 130, CaCl 2 1, MgCl 2 1.2, HEPES 10 and glucose 10 (pH 7.4 with CsOH). To block L-type Ca 2+ currents, 2 mM NiCl 2 was added to the superfusate. Experiments were performed at 22 ± 0.5°C. Whole-cell patch clamp recordings were obtained in voltageclamp mode using borosilicate glass pipettes (tip resistances 1.5-3 MΩ).
For Na + current recordings, the pipette solution contained in mM: CsCl 115, NaCl 5, EGTA 10, HEPES 10, MgATP 5, Tetraethylammonium Chloride (TEACl) 20 and MgCl 2 0.5 (pH 7.2, KOH). Voltage-dependent Na + currents were evoked by 5 mV step depolarizations (100 ms) from a holding potential of −100 mV to test potentials ranging from −95 mV to +40 mV. Cells were excluded from analysis if there was no reversal of the sodium current by +40 mV. To investigate Na v 1.5 voltage-dependent inactivation kinetics, cells were subject to 500 ms pre-pulses ranging from −120 mV to −40 mV, followed by a 100 ms step to −30 mV. For Na v 1.5 time-dependent recovery kinetics, a standard two pulse protocol was used (−120 mV to −30 mV, 20 ms), with the time between the two pulses incrementally varying between 5 and 950 ms.
All recordings and analysis protocols were performed using an Axopatch 200B amplifier (Molecular Devices, USA) and digitized at 50 kHz using a CED micro1401 driven by Signal v6 software (Cambridge Electronic Design, Cambridge, UK). Series resistance was compensated, ranging between 60-100% for all cells. Experiments were terminated if series resistance abruptly changed or was above 10 MΩ.

Atrial optical mapping
Optical mapping of the LA was conducted as previously described. 13 Isolated whole hearts were loaded on to a vertical Langendorff apparatus and perfused with a standard KH solution. Hearts were perfused at 4 mL.min −1 (equilibrated with 95%O 2 /5%CO 2 and heated to 36-37°C, pH 7.4). Hearts were loaded with 25 µL of voltage sensitive dye Di-4-ANEPPS at a concentration of 5 mg/mL, diluted in 1 mL of KH solution and delivered via bolus port injection over 3-5 min. The LA was then isolated and pinned in a superfusion chamber containing 37°C KH solution for transfer to the optical mapping setup, anterior surface facing up.
In the optical mapping system, atria were superfused with KH solution (95%O 2 /5%CO 2 , 36-37°C) containing contraction uncoupler Blebbistatin (35 µM). For imaging, atria were illuminated by two dual LEDs at 530 nm. A 630 nm long-pass filter was used to separate emitted fluorescence, imaged using an ORCA flash 4.0 CMOS camera (Hamamatsu, Japan). Images were acquired at a framerate 0.987 kHz and pixel size of 71 µm/pixel. 2 Atria were paced using bipolar platinum electrodes delivering 2 ms pulses at twice diastolic threshold (minimum voltage required to elicit APs).
One-minute baseline recording was taken following a 10 min equilibration period to ensure contraction uncoupling and temperature restabilization. During imaging, atria were initially paced at 330 ms pacing cycle length (PCL). A 'ramp' pacing protocol was then initiated, in which the atria were paced at 120 ms PCL for 100 stimuli and then PCL was reduced from 120 ms to 80 ms in 10 ms intervals every 20 stimuli. After taking baseline recordings, LED illumination was switched off and the superfusion solution replaced with an identical solution containing flecainide, a clinically-used sodium channel blocker 14 at a concentration of 1 µM and then 5 µM (or control solution without flecainide for time control experiments).
Subsequent recordings were then made as described above after 20 min superfusion with 1 µM flecainide solution and then further 15 min with 5 µM flecainide solution. Atria were paced at 330 ms PCL continuously in dark conditions between recordings.
From these recordings, APD and conduction velocity (CV) were mapped across the LA using ElectroMap software. 13 Atria were removed from analysis at a given PCL if loss of 1:1 capture ratio with pacing stimuli (i.e. missed beats) was observed.

Statistics
For all murine experiments presented herein, experimenters were blinded to the genotype of the littermate pairs during data collection and analysis. Student's t-tests were used for singular comparisons for normally distributed data. A hierarchical nested t-test ( Figure 3) and Mann-Whitney test (see Supplementary material online, Table S3) were used where appropriate.

Viable heterozygous Scn5a-M1875T +/− mice show normal cardiac size, structure, and basic function
The point mutation previously identified in patients with early familial AF was successfully introduced to exon 28 of the mouse Scn5a gene via homologous recombination of a targeting vector. The vector contained the T-C point mutation (CRISPR/cas9-mediated) resulting in methionine-threonine exchange in the Na v 1.5 protein (Figure 1; see Supplementary material online, Figure S1A). Offspring from both WT × heterozygote (=Scn5a-M1875T +/-) and heterozygote × heterozygote pairings were viable. No homozygous Scn5a-M1875T −/− offspring were born (see Supplementary material online, Figure S1B), suggesting embryonic lethality, as previously reported for other Scn5a mutations. 15 Accordingly, the ratio of WT and heterozygous animals shifted from 1:2 (expected) to approximately 1:3 when heterozygous animals were crossed (see Supplementary material online, Figure S1B). The ratio of male:female sex in offspring approximated 1:1 as expected (see Supplementary material online, Figure S1C). Age-matched young adult WT and Scn5a-M1875T +/− mice displayed similar heart rate, PR-, QRS-and QT-interval in ECGs recorded awake (see Supplementary material online, Table S1) and during sedation (see Supplementary material online, Table S2). Echocardiography and histological examination excluded overt differences in structure (see Supplementary material online, Table S3; Figure 2A). Neither atrial extracellular matrix content nor atrial cardiomyocyte cell diameter were affected by the mutation (Figure 2B and C; see Supplementary material online, Figure S2A). Accordingly, pro atrial natriuretic peptide protein expression (proANP) was detected in right atria as expected but was not elevated in ventricles (see Supplementary material online, Figure S2B).

Atrial Scn5a-M1875T +/− cardiomyocytes have an augmented peak sodium current density
To determine the impact of the M1875T point mutation on peak I Na amplitude and Na v 1.5 channel gating properties, left atrial cardiomyocytes from WT and Scn5a-M1875T +/− mice (8-13 weeks) were isolated and whole-cell patch clamp recordings were performed.
The M1875T variant increased I Na over test potentials ranging from −100 to +40 mV ( Figure 3A and B). At a peak test potential of −30 mV, mean left atrial cardiomyocyte I Na density was higher in Scn5a-M1875T +/− (−28.0 ± 1.1 pA/pF, n = 107 cells) than in WT littermates (−22.8 ± 0.7 pA/pF, n = 109 cells, P < 0.0001; Figure 3C). The elevation in I Na was also apparent when recordings were grouped by heart  Figure S3A).
Whole-cell I Na voltage-dependent inactivation and time-dependent recovery kinetics were not altered in Scn5a-M1875T +/− cardiomyocytes compared to WT (see Supplementary material online, Figure S3B and C). Na v 1.5 expression in hearts of Scn5a-M1875T +/− mice at the mRNA and the protein level revealed no difference at either the whole cell or isolated membrane fraction level (see Supplementary material online, Figure S4).

APs from Scn5a-M1875T +/− atria have a larger amplitude and faster peak upstroke velocity
APs were measured in whole left atrial tissue isolated from WT and Scn5a-M1875T +/− mice (9-13 weeks) using sharp microelectrodes.
The RMP was not different between genotypes (100 ms PCL: WT −72.4 ± 0.6; Scn5a-M1875T +/− −73.1 ± 0.6 mV). Atrial activation times were also similar (100 ms PCL: WT 4.9 ± 0.2; Scn5a-M1875T+/− 4.9 ± 0.2 ms) (see Supplementary material online, Table S4). Only at the long PCL of 1000 ms, the APD at 50 and 90% repolarization was shorter in Scn5a-M1875T +/− left atria than in WT ( Figure 4D-F), while the APD at 30% and 70% repolarization was not significantly different (see Supplementary material online, Table S4). Similarly, optical mapping data showed no APD differences at PCLs tested. We performed analyses of APD restitution (see Supplementary material online, Figure S5), which suggest that an effect of the M1875T mutation may be a blunting of APD restitution. mutation on atrial conduction. Left atria were superfused with the open channel sodium channel blocker flecainide (1 µM, clinically used concentration) to determine the response of Scn5a-M1875T +/− and WT left atria. While CV was unchanged at baseline, flecainide slowed conduction less in Scn5a-M1875T +/− left atria (CV difference at 100 ms PCL −6 ± 1 cm/s, n = 12, P = 0.0357) than in WTs (CV difference at 100 ms PCL −10 ± 1 cm/s, n = 12, P = 0.0357; Figure 5A and B; see Supplementary material online, Figure S6). We also investigated flecainide-induced changes in left atrial refractoriness as it is known that flecainide induces post-repolarization refractoriness using rapid atrial pacing. Representative optical AP recordings show 1:1 capture in the Scn5a-M1875T +/− left atria with flecainide, while several stimuli in the WT left atria did not elicit APs ( Figure 5C). All time-controlled left atria (no flecainide, same experimental duration) were successfully paced with 1:1 capture down to 80 ms PCL. Loss of 1:1 capture began at longer PCLs in WT (7/12 atria lost 1:1 capture at PCLs ≤ 120 ms, Figure 5D

Main findings
Our study describes the effects of the familial AF mutation Scn5a-M1875T +/− in a newly generated murine model. Key findings are that the M1875T Scn5a mutation leads to an increased atrial AP upstroke velocity and amplitude, and a selective increase in the early cardiac sodium current (I Na ). Atrial cardiomyocyte capacitance and size, as well as cardiac size and function, are preserved. The effect of the sodium channel blocker flecainide is dampened in Scn5a-M1875T +/− atria. Our measurements in this new murine model confirm that a selective increase of I Na may lead to familial AF as observed in the family affected, LA following exposure to 5 µM flecainide. Time control data (green) shows minimum PCL at which 1:1 stimulus capture was maintained in atria that were not exposed to flecainide but had been under experiment conditions for the same time period (35 min from baseline recording). n = 12 per group at 330-100 ms PCL. Atria were excluded from further analysis at shorter PCLs if 1:1 capture was lost, only data in steady state was used. n = 5 for time control experiments. *P < 0.05, **P < 0.01, WT vs. Scn5a-M1875T +/− ; ANOVA statistics. Data are represented as box-whiskers with box limits denoting the interquartile range (IQR) and the whiskers projecting to min and max values (B) or as the mean ± SEM (D). and suggest that commonly used concentrations of sodium channel blockers may be less effective in familial forms of AF with a selective increase in I Na than in other types.

Scn5a-M1875T +/− sodium channels in the murine atrium
Our data show a gain-of-function variant, namely an increased early sodium current as evidenced by an augmented APA and upstroke velocity, and larger I Na in hearts with a Scn5a-M1875T +/− variant.
In contrast to findings in HEK293 cells, 5 there was no depolarizing shift of Na v 1.5 channel inactivation in murine left atrial cardiomyocytes observed. Instead, we show a similarly augmented I Na without alterations in channel gating properties. Generation of the murine mutant model allowed us to study mutated sodium channels in cardiac tissue within the presence of the greater protein complex including α-and β-subunits and other membrane proteins, a complete cardiomyocyte contractile apparatus and all other cardiac cell types.
SCN5A genetic variants reported show various underlying mechanisms mainly linked to channel dysfunction, defective channel trafficking, or protein complex formation. 16 Cardiac sodium channels have been observed to form dimers and different regions of the channel protein have been implicated in dimerization. 17 Structural analysis following the hypothesis of Na v 1.5 dimerization via C-terminal interaction reveals that the surface of residue Met1875 of one Na v 1.5 cytosolic C-terminus will interact with Ala1924 of a second Na v 1.5. 17 This prediction suggests that the location site containing the mutation in our model could provide a structural basis for altered Na v 1.5-Na v 1.5 channel interaction further to be investigated.
It is unlikely that the M1875T mutation increases the late sodium current (I Na, l ), as APs we measured were not prolonged by the mutation, in line with the initial report of the mutation in human embryonic kidney (HEK) cells. 5 A lack of AP prolongation clearly differs to findings in the gain-of-function mutation ΔKPQ-Scn5a +/− mutant murine model which shows an increase in I Na,l 11 with prolonged atrial and ventricular APD, especially at longer PCLs. This suggests that the M1875T mutation acts differently to, and is distinct from, SCN5A gain-of-function mutations leading to prolonged repolarization and long QT syndrome, in concordance with the broad spectrum of phenotypic outcomes resulting from mutations in the same ion channel gene. 18 It could be of interest to model the effects of the increased I Na including effects on restitution and on human atrial AP models, e.g. higher amplitude and upstroke velocity at fast rates, while a slight APD shortening was only seen at slower rates.

Effect of Scn5a-M1875T +/− on conduction and refractoriness
While defects in the cardiac sodium channel can also cause conduction disturbances, 16,18 this variant shows a dampened response to flecainide both on conduction and post-repolarization refractoriness. Flecainide is a clinically used sodium channel blocker that inhibits cardiac I Na via blocking the pores of open Na v 1.5 channels. 10,14 The differential response to flecainide is likely due to the increased early sodium influx through the mutated Na v 1.5 channels, enabling preservation of conduction and activation properties in atrial cardiomyocytes when sodium channels are inhibited, leading to an enhanced activation reserve.

Limitations
This study primarily investigated which alteration in sodium channels is the basis for the clinically observed AF in the family carrying the mutation. Spontaneous AF was not observed during the single time point ECG. Detailed investigation of the arrhythmia trigger and pro-arrhythmic effects in the model, e.g. using telemetry and electrophysiological studies, are warranted in the future. It would also be of interest to systematically investigate the effect of the mutation on ventricular I Na .
Murine models have limitations due to differences between mice and human, but due to consistency in sodium currents have been useful in characterizing sodium channel mutations. 8,11,15 Identified mutation-induced changes at the cellular and organ level in this model appear sufficient to explain AF in the original family. Studies in human cardiomyocytes and in atrial engineered heart tissue would be desirable to assess the effect of the M1875T mutation in human models in the future.

Conclusion
The Scn5a-M1875T +/− variant causes a selective increase in the early cardiac sodium current, leading to an increased activation reserve and reduced refractoriness, whilst structure and contractile function are preserved. These findings can explain the familial occurrence of atrial ectopy and AF in the absence of reported severe heart disease. More widely they suggest a mechanism by which altered cardiomyocyte sodium current can predispose to AF. Our data also show that the M1875T gain-of-function mutation decreases the effectiveness of sodium channel blockers such as flecainide, which may have implications for treatment.