NaV1.5 knockout in iPSCs: a novel approach to study NaV1.5 variants in a human cardiomyocyte environment

Cardiomyocytes derived from patient-specific induced pluripotent stem cells (iPSC-CMs) successfully reproduce the mechanisms of several channelopathies. However, this approach involve cell reprogramming from somatic tissue biopsies or genomic editing in healthy iPSCs for every mutation found and to be investigated. We aim to knockout (KO) NaV1.5, the cardiac sodium channel, in a healthy human iPSC line, characterize the model and then, use it to express variants of NaV1.5. We develop a homozygous NaV1.5 KO iPSC line able to differentiate into cardiomyocytes with CRISPR/Cas9 tool. The NaV1.5 KO iPSC-CMs exhibited an organized contractile apparatus, spontaneous contractile activity, and electrophysiological recordings confirmed the major reduction in total Na+ currents. The action potentials (APs) exhibited a reduction in their amplitude and in their maximal rate of rise. Voltage optical mapping recordings revealed that the conduction velocity Ca2+ transient waves propagation velocities were slow. A wild-type (WT) NaV1.5 channel expressed by transient transfection in the KO iPSC-CMs restored Na+ channel expression and AP properties. The expression of NaV1.5/delQKP, a long QT type 3 (LQT3) variant, in the NaV1.5 KO iPSC-CMs showed that dysfunctional Na+ channels exhibited a persistent Na+ current and caused prolonged AP duration that led to arrhythmic events, characteristics of LQT3.

Generation of the NaV1.5 KO human iPSC line. The sgRNAs for CRISPR editing were designed using Predesigned Alt-R CRISPR-Cas9 guide RNA tool (Integrated DNA technologies website) and chosen according to the on-target and off-target scores. We have selected the following sgRNAs which have higher scores and the fewest and the non-cardiac off-targets: sgRNA-1 (5′-GAT TTC AAC CCT CCT AAC AT) and sgRNA-2 (5′-ACT TGT AGC TGA GAT CTG AG). The potential off-target sites are listed in the Supplementary Table S1 online.
The sgRNAs were cloned into the same modified px458 plasmid (Addgene plasmid 48,138) coding for the S. aureus Cas 9. Each sgRNA was under the control of its own U6 promoter. To generate CRISPR KO iPSCs, 7.5 µg of knockout plasmid (px458-sgRNA1-sgRNA2) was nucleofected into 800,000 iPSCs with 4D-Nucleofector (Lonza, Bâle, Switzerland) using P3 Primary cell kits (Lonza) and program CA-137. The cells were then seeded in 35-mm dishes in multiple dilutions in mTeSRplus supplemented with clone R (StemCell Technologies). After ~ 7 days in culture, each iPSC colony was manually picked using a micro-pipette under a Lynx EVO scope (Vision Engineering, CA, USA), and was transferred into two individual wells of two 96-well plates containing mTeSRplus medium. One plate used to extract the genomic DNA for PCR screening to detect positive clones, and www.nature.com/scientificreports/ the other was used to maintain and expand the positive clones. The PCR screening used the F1, 5′-GAC CAG AAA TGC ACT TGC TTC CTG TA and R1, 5′-GGT GTC TAT GAG AGT GGG CTT TGC T primers, which target regions surrounding the cutting sites of the sgRNAs. Positive clones, those with exon 6a removed, are characterized by a 682-bp PCR amplicon while negative clones are characterized by a 962-bp amplicon (Fig. 1B). One positive clone was chosen for the present study and was named Na V 1.5 KO.
cDNA analysis. RNA was extracted from Na V 1.5 KO iPSC-CMs and the control cell line on day 30 of maturation using Quick RNA mini-prep kits (Zymo Research, CA, USA). cDNA was synthesized using the Proto- Western blot analysis of Na V 1.5 KO and control iPSC-CM. Anti-Na V 1.5 (left panel) and the Anti-Pan Na V antibodies (right panel) were used to assess the expression of Na + channels. GAPDH staining was used as a loading control. Na V 1.5 KO (1) and Na V 1.5 KO (2) come from the same cell line but from two independent cardiac differentiation. This figure was made using MICROSOFT POWERPOINT  www.nature.com/scientificreports/ script II First Strand cDNA synthesis kit protocol (NEB, MA, USA). PCR analyses were performed using the F2, 5′-ACC TTC ACC GCC ATT TAC A and R2, 5′-CGT GGT CGG GGT TCT CGC CTGC) primers, which target regions surrounding the locus corresponding to exon 6a or 6b on Na V 1.5 mRNA. The PCR products were sent to a Sanger DNA sequencing facility (CHU de Québec, QC, Canada).
Western blotting. The proteins of the iPSC-CMs were extracted on day 30 of maturation. Briefly, the cells were washed with PBS and were lysed by scraping the cells into lysis buffer (10 mM Tris, 1% Nonidet P-40, 0.5% deoxycholic acid) supplemented with protease inhibitor cocktail (MilliporeSigma, ON, Canada). The cells were incubated in the lysis buffer for 1 h at 4 °C. Then, they were clarified by centrifugation at 18,000 g for 15 min at 4 °C. Protein concentrations were measured using the DC protein assay (BioRad, ON, Canada), with BSA as a reference. Protein extracts (20 µg) were denatured in 2X Laemmli buffer (MilliporeSigma) at 37 °C for 30 min. They were resolved on 4-12% gradient stain-free SDS-polyacrylamide gels (Bio-Rad) and were blotted on 0.45µm PVDF membranes (BioRad). The membranes were blocked and were incubated with rabbit anti-Na V 1.

Electrophysiology.
Patch clamp experiments with the iPSC-CMs were performed at room temperature using an Axopatch 200B amplifier and pClamp software v10 (Molecular Devices, CA, USA). Macroscopic Na + currents and APs were recorded using the whole-cell configuration of the patch clamp technique in voltage-and current-clamp modes, respectively. The pipettes were made from borosilicate glass capillaries (Sutter Instrument, CA, USA) and were fire polished. For the voltage-clamp experiments, the pipettes were coated with HIPEC (Dow-Corning, MI, USA) to minimize electrode capacitance. The pipettes were filled with a solution containing (in mmol/L): 35 NaCl, 105 CsF, 10 EGTA, and 10 HEPES. The pH was adjusted to 7.4 with CsOH. To record biophysical parameters, the bath solution was composed of (in mmol/L): 105 NMDG, 35 NaCl, 2 KCl, 1.5 CaCl 2 , 1 MgCl 2 , 10 D-glucose, 10 HEPES, 10 TEA-Cl, and 0.01 Nifedipine. For persistent Na + currents recordings, the bath solution was replaced by (in mmol/L): 140 NaCl, 2 KCl, 1.5 CaCl 2 , 1 MgCl 2 , 10 D-glucose, 10 HEPES, 10 TEA-Cl, 0.01 nifedipine, and 20 µmol/L TTX (LATOXAN, France). The pH of the bath solutions was adjusted to 7.4 with methanethiosulfonic (MTS) acid. Series resistance and cell capacitance were corrected. Na + currents were filtered at 5 kHz, digitized at 10 kHz, and stored on a microcomputer equipped with an AD converter (Digidata 1440A, Molecular Devices). P/4 leak subtraction was used prior to applying pulse stimulations.
For the current-clamp experiments, the patch pipettes (resistance 2-5 mΩ) were filled with a solution containing (in mmol/L): 10 NaCl, 122 KCl, 1 MgCl 2 , 1 EGTA, and 10 HEPES. The pH was adjusted to 7.3 with KOH. The bath solution (external current clamp) was composed of (in mmol/L): 154 NaCl, 5.6 KCl, 2 CaCl 2 , 1 MgCl 2 , 8 D-glucose, and 10 HEPES. The pH was adjusted to 7.3 with NaOH. The APs were recorded at 0.5, 1, 2 and 2.5 Hz stimulation frequencies. The holding potential during recording was maintained at − 80 mV. The duration of stimulation pulse was 3 ms with 0.5-1.5 nA injected current depending on the cell.

Optical mapping of iPSC-CMs in monolayers.
Optical mapping techniques were used to study the membrane potential and calcium waves on iPSC-CMs monolayers. The iPSC-CMs monolayers were reconstituted with the same number of cells over the experiments. iPSC-CMs were dissociated at days 12 to 15 of maturation with STEMdiff™ Cardiomyocyte Dissociation kit (StemCell Technologies). The monolayer was then reconstituted with 350 000 cells seeded on a 13 mm TC coverslip (Sarstedt) coated with hESC-qualified Matrigel®. The STEMdiff™ Cardiomyocyte Maintenance medium was replaced every two days for at least 15 days post-reconstruction. The iPSC-CMs (30-60 days of maturation) were feed with fresh medium one hour before www.nature.com/scientificreports/ optical recording. The cells were stained for 15 min with 10 µmol/L of di-4-ANEPPS (Thermo Fisher Scientific) for membrane potentials recording or for 30 min with 5 µmol/L of Rhod-2-AM (Thermo Fisher Scientific) for intracellular Ca 2+ transient imaging in the incubator (37 °C, 5% CO 2 ). Both dyes were diluted in Tyrode solution (in mmol/L: 154 NaCl, 5.6 KCl, 2 CaCl 2 , 1 MgCl 2 , 8 D-glucose, and 10 HEPES; pH 7.3). The cells were washed with the Tyrode solution and incubated for an additional 10 min or 30 min before optical or Ca 2+ imaging, respectively. Membrane voltage and intracellular Ca 2+ transients were mapped at high-speed (500 frames/sec) using CMOS N256 camera (MiCAM03, Brainvision, SciMedia Ltd, U.S.A) at a field of view of 15 mm. The imaging system includes a 530 nm green LED light source (LEX2-LZ4-G) with a stable intensity of 360 mW/cm 2 , led by an optical fiber and followed by a imaging cube containing a collimator, a dichroic mirror (560 nm), a bandpass 531/40 nm excitation filter (50 mm, BrightLine®, Semrock) and a long-pass 665 nm emission filter (50 mm, Andover Corporation), as well as a lens system with a maximum aperture of f/1.4 aligned in epifluorescence configuration for di-4-ANEPPS and Rhod-2, AM imaging. The recordings were performed at 37 °C using TCW1 worming plate controlled by TC02 temperature controller (Multichannel Systems, BW, Germany). Bipolar platinum/iridium electrodes (World Precision Instruments) positioned at the lower edge of the preparation were used to pace the monolayer with a stimulus generator (STG4002, Multichannel Systems) at cycle of 500 ms to 1500 ms using a bipolar square pulse of ± 8 V of a duration of 10 ms. 10 µmol/L of blebbistatin (Sigma-Aldrich) was added to the recording solution during membrane voltage recording to prevent cardiomyocytes contraction artefacts in the fluorescence signal. Raw optical signals were processed using Brainvision Workbench software to analyse the electrophysiological measurements, calculate the conduction velocities and generate the maps.

Statistical analysis.
In patch clamp recordings, APs exhibited diverse shapes that can be attributed to different cardiomyocyte cell types, that are, ventricular, atrial, and nodal cells. For statistical analyses, we decided to pool ventricular and atrial-like APs and exclude nodal-like APs. The nodal-like APs were excluded based on their very distinctive shape, i.e., no plateau and low overshoot. All statistical analyses were performed using PRISM8 software (GraphPad, CA, USA). The normality of distribution was determined using D' Agostino-Pearson normality test. When the distribution is normal, the data are expressed as mean ± SEM (standard error of the mean). If not, the data are expressed as median ± quartiles (25 % and 75 %) with the min to max values. When 3 groups were compared, statistical significance was determined by one-way ANOVA with Dunnet's post hoc test, or a two-way ANOVA for two independent variables. When only two groups were compared, a two-tailed unpaired Student's t-test was used. Where data were not normally distributed, a nonparametric test (Mann-Whitney or Kruskal-Wallis test) was performed. All the statistical tests were performed using a 95 % confidence interval and the differences were considered significant beyond the risk threshold 0.05% (*P < 0.05, **P < 0.01, ***P < 0.001).

Results
CRISPR-Cas9 strategy for Na V 1.5 knockout. To selectively and permanently suppress Na V 1.5 channel expression in iPSCs, we used CRISPR-Cas9 gene editing technology. The strategy was to remove exon 6a in the SCN5a gene using two sgRNAs that precisely target the regions flanking this exon (Fig. 1A). We used PCR amplifications of the SCN5a genomic locus encompassing the sgRNA target sites to screen for positive iPSC clones. Figure 1B shows the positive PCR result of the Na V 1.5 KO cell line used in the present study. The Na V 1.5 KO cells had a shorter amplicon than the control cells (not exposed to Cas9), confirming that exon 6a in SCN5a had been deleted. A genetic screening by quantitative PCR was performed of the Na V 1.5 KO iPSCs to verify the chromosomes status ( Supplementary Fig. S1 online). 8 most mutated regions in iPSCs were analysed and compared to a genomic DNA control provided in the hPSC Genetic Analysis Kit. The NaV1.5 KO iPSCs has the same copy number that control except for the X chromosome where it was divided by two. The NaV1.5 KO cell line comes from a male, containing one copy of X chromosome, while the genomic DNA control comes from a female. So, the KO line does not appear to have genetic abnormalities.
To verify that Na V 1.5 protein expression was suppressed in the cells, the Na V 1.5 KO iPSCs were differentiated into cardiomyocytes. After 30 days of maturation, RNA and Western blot analyses were performed. The Na V 1.5 cDNA PCR amplification revealed a shorter amplicon, confirming that the Na V 1.5 mRNA was free of exon 6a and 6b as shown by the sequencing chromatogram (Fig. 1C,D). The deletion of the two exons 6 in the mRNA instead of only one as designed by our strategy can probably be explained by a mis-splicing event. The deletion of the exon 6a may have compromised splicing sites for the exon 6b since they are only 140 bp apart in the SCN5a gene. The Na V 1.5 cDNA sequencing chromatogram validated the shift in the reading frame of Na V 1.5 and the formation of stop codons in the region corresponding to exon 7 (only the first stop codon is shown in Fig. 1 D). The Western blot analysis were performed using two primary antibodies (Fig. 1E). The anti-Na V 1.5 antibody which recognize the intracellular loop of Na V 1.5 channel, between domains I and II and the anti-Pan Na V antibody which recognize the intracellular loop between domains III and IV, identical in all isoforms of Na V 1 in all vertebrates. The unedited blots are available at different exposures in Supplementary Figure S6 and S7. The immunostaining shows that Na V 1.5 protein expression was completely suppressed in the KO iPSC-CMs. Our CRISPR construction makes it possible to stop the translation of the protein at the end of domain I, that is why we cannot observe a truncated Na V 1.5 protein in the KO iPSC-CMs. The immunostaining with an anti-Pan Na V antibody that recognizes all channel subtypes showed that isoforms other than Na V 1.5 were expressed but at much lower levels (Fig. 1E). We used this new Na V 1.5 KO iPSC line in the present study.
Intact contractile structures in Na V 1.5 KO iPSC-CMs. The Na V 1.5 KO iPSCs were successfully differentiated into cardiomyocytes (Supplementary Videos S1 and S2 online). No difference in the differentiation course of the KO cells was observed compared to the control cells. The Na V 1.5 KO iPSC-CMs started to sponta- www.nature.com/scientificreports/ neously beat around day 8 to 10 of differentiation as did the control cells. After 30 days of maturation, the beating rate of the Na V 1.5 KO iPSC-CMs (0.46 ± 0.30 Hz) was also similar to that of the control cells (0.46 ± 0.19 Hz) (Fig. 5I). The morphologies of the cardiomyocytes differentiated from the Na V 1.5 KO iPSC line were then characterized to rule out the possibility that the loss of Na V 1.5 had affected sarcomeric organization. Immunofluorescence staining for myosin light chain 2v (MLC2v) and cardiac troponin T (cTnT) revealed that the contractile proteins in Na V 1.5 KO iPSC-CMs exhibited the same normal organization as in the control cells (Fig. 2).
Voltage and current-clamp analyses of Na V 1.5 KO iPSC-CMs. The Na V 1.5 channel is by far the main contributor to Na + currents that initiate and trigger the rise of APs in cardiomyocytes. With the KO of Na V 1.5, we expected that the iPSC-CMs would produce a very small Na + current. We thus performed voltage-clamp recordings in dissociated iPSC-CMs cells to measure the Na + and Ca 2+ currents in the Na V 1.5 KO and control cells. The Na V 1.5 KO iPSC-CMs did not produce any recordable Na + current, whereas the control cells generated a mean current amplitude of − 3200 pA, as shown by the representative current traces (Fig. 3A). When Figure 2. Immunofluorescence staining of cardiac markers in Na V 1.5 KO iPSC-CMs. NaV1.5 KO iPSC-CMs with organized contractile proteins, including myosin light chain 2v (MLC2v, green staining) and cardiac troponin T (cTnT, red staining). The bottom panels show merged images. Scale bar: 20 µm. Images were acquired using a Zeiss ImagesM2 LSM confocal microscope and were processed using ZEN software (Zeiss). This figure was made using IMAGEJ 1. www.nature.com/scientificreports/ the Na + concentration was increased from 35 to 140 mmol/L in the external recording solution, the Na V 1.5 KO iPSC-CMs produced a small Na + current of -670 pA, as shown by the representative current trace (Fig. 3A). At this Na + concentration, the amplitudes of the Na + currents in the control cells were too large (> − 20,000 pA) to be recorded. Given that the Na V 1.5 KO cell line generated beating cardiomyocytes, it should also produce Ca 2+ currents that are responsible for excitation-contraction coupling in cardiac muscles. The Na V 1.5 KO produce a Ca 2+ current which has the same biophysical properties that control iPSC-CMs ( Supplementary Fig. S2 online).  www.nature.com/scientificreports/ In the Na V 1.5 KO iPSC-CMs, the small residual Na + current was blocked by 50 nmol/L TTX ( Supplementary  Fig. S3 online). The Ca 2+ current activated several milliseconds after the Na + current was blocked with 5 µmol/L nifedipine, a Ca 2+ channel blocker.
We next studied the impact of Na V 1.5 current losses on the properties of APs in iPSC-CMs. Na V 1.5 KO and control iPSC-CMs both triggered spontaneous APs in the gap-free mode of the current-clamp configuration (data not shown). We observed similar cell-subtypes repartition between control and Na V 1.5 KO iPSC-CMs: most ventricular-like cells and a minority of atrial-and nodal-like cells (data not shown). When stimulated at a stimulation frequency of 1 Hz, the Na V 1.5 KO cells triggered AP trains that respected the pace of stimulation seen in the control cells (Fig. 3B). However, the injected current to trigger an AP in the Na V 1.5 KO cells (0.97 ± 0.05 nA) was significantly increased compared to control cells (0.77 ± 0.04 nA; **p < 0.01 unpaired t-test). The following experiments were carried out on cardiomyocytes stimulated at 1 Hz in order to normalize the beating rate and enable further analysis. For each AP, several parameters were analysed, this include the overshoot, the dV/ dt max which corresponds to the opening of Na + channels (0 phase of AP), and the 20, 50 and 90% of repolarization which corresponds to the opening of Ca + and K + channels (1, 2 and 3 phases of AP).
As expected, several alterations in the APs from Na V 1.5 KO cells were identified. The APs from the Na V 1.5 KO iPSC-CMs exhibited a significant increase in duration (APD) when measured at 20, 50, and 90% of repolarization compared to the control cells (Fig. 3C,D). For example, the median of APD 90 was 442.5 ms with quartiles [333.9; 522.0] for the Na V 1.5 KO cells compared to 319.3 ms [200.3; 452.8] for the control cells, which was a significant increase of 122 ms. The AP overshoot and the upstroke velocity (dV/dt max ) were also affected in the Na V 1.5 KO iPSC-CMs, which influenced the amplitude and the time to peak of APs, as shown in the representative traces (Fig. 3E). The overshoot in the Na V 1.5 KO iPSC-CMs (37.0 ± 1.4 mV) decreased significantly by 14 mV compared to the control cells (51.4 ± 2.2 mV) (Fig. 3F). The main effect on APs was related to the dV/dt max , where a major decrease of 88 mV/ms was observed in the Na V 1.5 KO iPSC-CMs (15.2 ± 1.4 mV/ms) compared to the control cells (102.8 ± 7.7 mV/ms) (Fig. 3G).
From a pharmacological point of view, Na V 1.5 KO does not respond in the same way as the control iPSC-CMs. The APs from control iPSC-CMs were sensitive to TTX treatment ( Supplementary Fig. S4 online). Indeed, the addition of TTX decreased the dV/dt max of APs from control cells and this in a concentration-dependent way. For Na V 1.5 KO iPSC-CMs, the TTX had no blocking effect of the triggering of APs ( Supplementary Fig. S3  online). Moreover, the treatment with 1 µmol/L nifedipine blocks the APs from the Na V 1.5 KO but not APs from the control iPSC-CMs which just had a shortened duration ( Supplementary Fig. S3 and S5 online). These results confirm the Ca 2+ nature of the APs recorded from Na V 1.5 KO iPSC-CMs.
Voltage and calcium optical mapping of iPSC-CM monolayers. We next generated large (~ 1 cm diameter) iPSC-CM monolayers to study the biophysical impact of the loss of Na V 1.5 channels at level closer to heart tissue using the optical mapping (OM) technique. The conduction of electrical impulses through the iPSC-CM monolayers stained with the voltage sensitive dye di-4-ANEPPS was first analyzed. Figure 4A shows representative optical APs (OAPs) recorded in Na V 1.5 KO and control monolayers stimulated at 1 Hz. As with the patch-clamp recordings, both cell lines generated AP trains that followed the pace of the electrical stimulations. Activation maps were produced for the Na V 1.5 KO and control monolayers from these optical recordings (Fig. 4B). The maps notably showed that the conduction velocity (CV) was slower in the Na V 1.5 KO monolayers than in the control monolayers (Supplementary Video S3 online). When stimulated at 0.5, 1, and 1.5 Hz, the conduction velocities measured in the Na V 1.5 KO monolayers were 2.67 ± 0.16, 2.43 ± 0.13, and 2.12 ± 0.13 cm/s, respectively compared to 13.02 ± 1.7, 10.49 ± 1.3, and 8.77 ± 1.4 cm/s in the control monolayers, respectively (Fig. 4C). We next analyzed the parameters of the OAPs and found that the rise time, the duration, and the amplitude were affected in the Na V 1.5 KO monolayers at all three stimulation frequencies (Fig. 4D). The rise time of the OAPs was prolonged in the Na V 1.5 KO monolayers by an average of 5.7 ms. For example, the rise time was 16.5 ± 0.6 and 10.7 ± 1.0 ms for the Na V 1.5 KO and the control monolayers at 1 Hz, respectively. The duration of the OAPs was significantly shortened for the Na V 1.5 KO monolayers (Fig. 4E,F; Supplementary Video S3 online). At 1 Hz, the APD 50 and APD 80 were 356 ± 15 and 421 ± 14 ms for the Na V 1.5 KO monolayers, and 489 ± 24 and 585 ± 22 ms for the control monolayers. This represented 133 ms and 164 ms differences in the APD 50 and APD 80 , respectively. The amplitude of the OAPs also significantly decreased by 9 %, as calculated using the normalized fluorescence intensity, decreasing from 59,900 relative fluorescence units (RFU) in the control monolayers to 54,275 RFU in the Na V 1.5 KO monolayers at 1 Hz (Fig. 4G).
The spatial changes in the kinetics and amplitude of Ca 2+ transients were next studied using the fluorescent probe Rhod-2-AM on cell monolayers. Figure 5A shows representative Ca 2+ signal recorded in Na V 1.5 KO and control monolayers when electrically stimulated at 1 Hz. Uniform propagation of Ca 2+ waves through the entire monolayer was observed in both cell lines, as shown by the representative Ca 2+ activation maps in Fig. 5B. As observed in OAPs recording, Na V 1.5 KO monolayers exhibited a significantly delayed and slower propagation of Ca 2+ transients and a compared to control monolayers (Supplementary Video S4 online). This was expected since the Ca 2+ transients are triggered during AP propagation through the Ca 2+ -induced Ca 2+ release (CICR) process. Thus, Ca 2+ waves propagation velocities (CaPV) were measured in Na V 1.5 KO (3.3 ± 0.6, 3.2 ± 0.4 and 2.7 ± 0.5 cm/s) and control (13.4 ± 1.5, 10.3 ± 0.93 and 7.4 ± 0.9 cm/s) at 0.5, 1 and 1.5 Hz pacing frequencies, respectively (Fig. 5C). Similarly, the Ca 2+ transient characteristics in Na V 1.5 KO monolayers were altered including its amplitude, durations (TD 50,80 ), half times to peak and decay speed (τ). A significant reduction of 10% in the Ca 2+ transients normalized amplitudes was observed between the two groups reflecting the maximum intracellular Ca 2+ concentrations, as measured at 64,335 RFU in the control and 58,337 RFU in the Na V 1.5 KO monolayers at stimulation frequency of 1 Hz (Fig. 5D). Furthermore, the durations of Ca 2+ transients were significantly shortened in Na V 1.5 KO iPSC-CMs at 0.  (Fig. 5E,F). More, the half times of Ca 2+ peak was shortened in Na V 1.5 KO iPSC-CMs at 0.5 Hz and 1 Hz, but not at 1.5 Hz (Fig. 5G). At 1 Hz, the half time of Ca 2+ peak was decreases by 36 ms with 252.3 ± 12.1 and 216 ± 5.8 ms for control and Na V 1.5 KO, respectively. As for the decay tau (τ), Na V 1.5 KO iPSC-CMs exhibit a faster decay of Ca 2+ transients (smaller τ) at 0.5 Hz and 1 Hz (Fig. 5H; Supplementary Video S4 online). As an example, the decay τ was 342.6 ± 24.5 in Na V 1.5 KO, and 402.5 ± 11.5 in control at 1 Hz. Finally, the spontaneous beating frequencies were measured before the addition of blebbistatin and electrical stimulation for each monolayer. The beating frequencies were found similar at ~ 0.46 Hz within the two groups (Fig. 5I).  www.nature.com/scientificreports/ www.nature.com/scientificreports/ Na V 1.5 channel re-expression in Na V 1.5 KO iPSC-CMs. We next used the Na V 1.5 KO cell line as an expression system to characterize the impact of a Na V 1.5 variant in a human cardiac environment. The known LQT3 variant delQKP 1507-1509 (delQKP) was used as a proof-of-concept in the present study. The Na V 1.5 KO iPSC-CMs were transiently transfected with either the cDNA Na V 1.5/delQKP construct or the Na V 1.5/WT construct. We conducted a voltage-clamp analysis and compared the biophysical properties of the mutated channel with those of the WT channel. The Na V 1.5/WT iPSC-CMs were also compared with the control iPSC-CMs. The Na V 1.5 KO iPSC-CMs transfected with Na V 1.5/WT (Na V 1.5/WT) or Na V 1.5/delQKP both produced a robust typical Na + current (Fig. 6A). The current densities in Na V 1.5/WT (− 192 ± 39 pA/pF) and Na V 1.5/delQKP (− 195 ± 43 pA) were similar as shown by the I/V curves (Fig. 6B). The Na V 1.5/WT tended to have slightly higher current densities compared to control iPSC-CMs (109 ± 10 pA/pF) but the difference was not significant (p = 0521) (Fig. 6B). The I/V curves were converted to conductance-voltage curves (G/V), which revealed no significative differences in steady-state activation between Na V 1.5/WT and Na V 1.5/delQKP (Fig. 6C). Compared to the control iPSC-CMs (− 35.9 ± 1.2 mV), the V 1/2 of activation of Na V 1.5/WT (− 41.7 ± 1.4 mV) was shifted significantly toward more negative voltages (Fig. 6C). An analysis of steady-state inactivation revealed a significant − 4.6 mV shift in the V 1/2 of Na V 1.5/delQKP (− 84.8 ± 0.4 mV) compared to Na V 1.5/WT (− 80.2 ± 1.1 mV) (Fig. 6D). No difference was seen in the V 1/2 of inactivation between Na V 1.5/WT and the control iPSC-CMs. The voltage dependence of inactivation was also significantly higher in Na V 1.5/delQKP (k v = − 5.8 ± 0.2) than in Na V 1.5/WT (k v = − 7.0 ± 0.2) as shown by the slope of the inactivation curve (Fig. 6D). We next used a double-pulse protocol to directly measure the kinetics of recovery from inactivation. The recovery kinetics, which were determined by fitting a single-exponential equation, were more than two times faster for Na V 1.5/delQKP (t rec = 3.9 ± 0.3 ms) than for Na V 1.5/WT (t rec = 9.9 ± 0.9 ms) (Fig. 6E). Na V 1.5/WT was slightly slower to recover than the control (t rec = 7.1 ± 0.5 ms). The decay of inactivation of the mutated channel was also affected. Na V 1.5/ delQKP exhibited three times faster inactivation than Na V 1.5/WT as calculated by fitting the current decays with an exponential function (Fig. 6F). For example, at − 40 mV, the time constants of decay were 1.1 ± 0.1 ms and 3.0 ± 0.3 ms for Na V 1.5/WT and Na V 1.5/delQKP, respectively. This faster inactivation rate can be clearly seen on the representative current traces in Fig. 6A. These results indicate that the inactivation process of the channels was compromised by the deletion of the three amino acids. This situation often generates the appearance of a persistent Na + current, which is what occurred with the variant channel. Na V 1.5/delQKP produced a persistent Na + current more than four times larger than Na V 1.5/WT, reaching 2.2 ± 0.3%, while it was at 0.5 ± 0.2% of the transient current in Na V 1.5/WT (Fig. 6G). The nature of this non-inactivating current was confirmed by its block with 20 µmol/L TTX (Fig. 6H). The biophysical parameters of the voltage-clamp recordings are compiled in Supplementary Table S2 online. One of the major advantages of our new expression system using Na V 1.5/KO iPSC-CMs is the ability to study the impact of the mutation on APs directly. The results described above showed that the Na + current was restored but was affected when Na V 1.5/delQKP was transfected. We measured the impact of these biophysical effects on APs in current-clamp experiments. The first noticeable effect seen on APs after Na V 1.5/delQKP transfection was the increase in APD compared to those recorded in cells transfected with the WT channel. The representative AP traces in Fig. 7A clearly show this increase in duration for Na V 1.5/delQKP. The APD 50 and APD 90 , but not APD 20 , significantly increased by 101 and 150 ms, respectively, in Na V 1.5/delQKP following 1-Hz stimulations (Fig. 7B). Differences were also observed between control iPSC-CMs and Na V 1.5/WT iPSC-CMs. The APD 20 and APD 50 of Na V 1.5/WT were 81 and 92 ms smaller, respectively, than those of the control. The dV/dt max , which measures the kinetics of the rising phase of APs, was also affected and was significantly slower in Na V 1.5/delQKP (101.7 ± 6.2 mV/ms) than in Na V 1.5/WT (151.1 ± 10.9 mV/ms) (Fig. 7C). The dV/dt max of Na V 1.5/WT was significantly faster than that of the control (102.8 ± 7.7 mV/ms) (Fig. 7C). The AP overshoots were not affected, regardless of the iPSC-CMs studied, and maintained values of approximately 50 mV (Fig. 7D). The above results were determined when the iPSC-CMs were stimulated at 1 Hz. However, the current-clamp recordings were also conducted at stimulation frequencies of 0.5, 1.5, and 2 Hz. When stimulated at 1.5 Hz and above, the cells transfected with Na V 1.5/delQKP were more often unable to adapt to the rhythm of stimulation. In fact, at 1.5 Hz, 40% (8/20 cells) of the Na V 1.5/delQKP iPSC-CMs were unable to follow the rhythm compared to only 11% (2/18 cells) for Na V 1.5/WT. At 2 Hz, 90% (18/20 cells) of Na V 1.5/delQKP cells were unable to adapt to the rhythm of stimulation. Arrhythmic events such EADs (early afterdepolarizations) and DADs (delayed afterdepolarizations) were also observed on a regular basis in Na V 1.5/delQKP iPSC-CMs. Figure 7E shows examples of DADs recorded at 0.5 Hz in Na V 1.5/delQKP cells as well as EADs recorded in spontaneous APs.

Discussion
In this study, we used CRISPR-Cas9 genomic editing tool to successfully produce a novel iPS cell line in which the SCN5a gene coding for the cardiac Na + channel Na V 1.5 was knocked-out. This iPS cell line was used to produce cardiomyocytes (iPSC-CMs) that no longer express Na V 1.5 channels. We first characterized by electrophysiological recordings and optical mapping analysis the effect of the Na V 1.5 KO on the AP itself and their propagation in iPSC-CMs monolayer. Secondly, we re-introduced by transient transfection a WT Na V 1.5 channel or a Na V 1.5 LQT3 variant, delQKP, as a proof-of-concept to study Na V 1.5-linked variants in a human cardiomyocyte environment. Na V 1.5 KO iPSC-CMs exhibited spontaneous beats with a frequency comparable to the control cell line. Morphologically, we observed that Na V 1.5 KO iPSC-CMs express MLC2v and troponin T, two essential structural proteins involved in the formation of myosin and actin filaments, respectively. These proteins play an essential role in the organization of sarcomeres, and in the parallel alignment and movement of the filaments 11,12 . The suppression of Nav1.5 channels obviously did not affect the organized contractile apparatus to modify the contractility capabilities and rhythm of iPSC-CMs. www.nature.com/scientificreports/ , Na V 1.5/WT (n = 13) and Na V 1.5/delQKP (n = 9) iPSC-CMs. Na + current densities were measured by normalizing current amplitudes to membrane capacitance. Inset shows graph of I max . (C) Steady-state activation of Na + currents. Activation curves were generated using a standard Boltzmann distribution: G(V)/G max = 1/(1 + exp (− (V − V 1/2 )/k)). Inset shows graph of V 1/2 . (D) Steady-state inactivation of control (n = 14), Na V 1.5/WT (n = 13) and Na V 1.5/delQKP (n = 9). Inactivation currents were obtained using 20-ms test pulses to − 30 mV after a 500-ms pre-pulse to potentials ranging from − 120 to + 30 mV. The inactivation values were fitted to a standard Boltzmann equation: I(V)/I max = 1/(1 + exp ((V − V 1/2 )/k)) + C. Inset shows graph of V 1/2 and k slope (E) Recovery from inactivation values recorded from control (n = 14), Na V 1.5/WT (n = 12) and Na V 1.5/delQKP (n = 9) iPSC-CMs. The cells were depolarized to − 30 mV for 40 ms from a holding potential of − 100 mV to inactivate the Na + channels. Test pulses were then applied at − 30 mV for 20 ms to measure current amplitudes, with an interval ranging from 0.1 to 4000 ms. The resulting curves were fitted with a simple exponential equation: (A (exp (− t/τ) + C). Inset shows a graph of the recovery time constant (τ rec ). (F) The time constants of fast inactivation decay were plotted as a function of voltage for control (n = 14), Na V 1.5/ WT (n = 13) and Na V 1.5/delQKP (n = 9). The time constants were obtained using a simple exponential function. (G) Persistent Na + current/voltage relationships recorded in control (n = 9), Na V 1.5/WT (n = 11) and Na V 1.5/delQKP (n = 12) iPSC-CMs. The currents were obtained using 300 ms pulses from − 100 to + 40 mV in + 5 mV increment. www.nature.com/scientificreports/ As expected, electrophysiological recording revealed a marked 97% reduction in Na + currents in the Na V 1.5 KO iPSC-CMs. A very small inward and fast-gated Na + current remains. By its high sensitivity to TTX, this inward current could probably be carried by Na V 1.7 channels which have already been identified by our group as slightly expressed in iPSC-CMs 13 . Such a reduction in Na + currents should have a major impact on the APs given their role in the triggering and the rising phase (phase 0) of the APs. Furthermore, the entry of Na + in the cardiomyocytes is normally followed by a cascade of ion channels activation such as K + and Ca 2+ channels that contribute to the plateau and the repolarisation phase of APs 14 . We can immediately notice that APs were still elicited in Na V 1.5 KO iPSC-CMs. Their resistance to TTX (1 µmol/L) but their complete inhibition by nifedipine (1 µmol/L) reveal the Ca 2+ nature of the APs in the KO cells. These Ca 2+ APs were characterized by a rising phase (phase 0) considerably slow due to the lack of Na + channel, as calculated by the dV/dt max from the electrophysiological recordings. Moreover, a stronger injected current was required to trigger Ca 2+ APs. Normally, during the phase 0, Na + channels contribute to more depolarization and the opening of Ca 2+ channels. In our situation where there are few Na + channels, a stronger depolarization is required to open Ca 2+ channels.
These alterations will certainly impact the propagation of the electrical signal through the myocardium. We take advantage of the properties of iPSC-CMs to form electrical coupled syncytia of monolayers and we have indeed found a drastically slowing of the conduction velocity in the KO. The role of Na V 1.5 in the cardiac conduction was already known as shown in a previous study with heterozygous (Scn5a +/− ) Na V 1.5 mice. These mice exhibited a 50% reduction in Na + current and had impaired atrioventricular conduction and delayed intra myocardial conduction 15 . Our optical mapping experiments confirmed the critical role of Na V 1.5 for controlling the speed of electrical wave propagation trough myocytes. The slowing of conduction velocity could be related to www.nature.com/scientificreports/ the lack of Na + channels in the intercalated discs, but, probably to connexins dysregulation, where their role must be studied in the context of Na V 1.5 KO. It is known that connexins interact physically with Na + channels and the lack of these channels in the intercalated discs may impede their role in establishing a normal conduction velocity. Another major parameter involved was the duration of the action potential. We found contradictory results according to the experiments carried out. In electrophysiological recordings, the duration of the AP was found to be prolonged while this was shortened in optical mapping analysis in the Na V 1.5 KO iPSC-CMs. In cardiomyocytes, Na V 1.5 influence the duration of the AP by the non-inactivating channels that re-open and generate a small inward current (persistent Na + current) during the plateau phase. Previous studies have shown that TTX, a Na + channel blocker, shortened the APD in dog and guinea pig ventricle heart muscle but not in atrial 1,16 . In addition, heterozygous (Scn5a +/− ) Na V 1.5 mice showed shorter APDs in the right ventricular epicardium but not in the left ventricular 17 . Studies with other Na + blockers such as flecainide, quinidine and lidocaine showed that the APD can also be increased depending on the stimulation frequency, tissue location and species [17][18][19] . To resume, the effect of reduction or the suppression of Na V 1.5 expression on APDs is not straightforward but influenced by the tissue and the Na + channel blocker used, among other factors. In this study, the results disparity between patch-clamp and optical mapping experiments could originate from the nature of the sample. In one case, the recordings were done in individual cells whereas in the other case they were performed on monolayers as behave as a functional syncytium. This highlights the complexity of the contributing factors that can influence the duration of the action potential. Further studies are warranted to understand more this difference.
To complete the Na V 1.5 KO iPSC-CMs characterization, we were interested to Ca 2+ homeostasis by measuring the Ca 2+ currents by the Patch-clamp technique and the propagation of Ca 2+ transients by optical mapping. The absence of Na V 1.5 does not modify the biophysical properties of L-type Ca 2+ channels but altered the propagation of Ca 2+ transients. Ca 2+ transients represent Ca 2+ release/reuptake cycles that coordinate contraction and relaxation of cardiomyocytes. Even though Na V 1.5 KO iPSC-CMs beat spontaneously and at rhythm similar to control, Ca 2+ propagation velocity has significantly decreased. This result is perfectly correlated with the reduction in conduction velocity and could be explained by alteration of the excitation-contraction coupling in the absence of Na V 1.5 in cardiomyocytes. The intracellular Ca 2+ release is triggered by the AP, which their conduction is delayed in the KO, which therefore also causes the observed delay in the wave of Ca 2+ release. Optical recordings on Na V 1.5 KO iPSC-CMs also showed a decrease in amplitude of Ca 2+ transients signal, probably indicating a reduction of Ca 2+ availability at each cycle. This could be explained by other Ca 2+ handling players such as Ryanodine channel or SERCA pump.
In the second part of this study, we have re-introduced by DNA transfection the Na V 1.5 WT channel or the Na V 1.5 delQKP LQT3 variant in the Na V 1.5 KO iPSC-CMs and compared their biophysical effects. The transfection of the Na V 1.5/WT channel in KO iPSC-CMs successfully restored Na + channel expression and the fast upstroke velocity of APs. In fact, the upstroke velocity was even faster in Na V 1.5/WT than in control. We hypothesize that this difference could be explain be the higher level of Nav1.5 channels due to its overexpression produced by the plasmid transfection.
SCN5a mutations leading to LQT3 are generally associated with a gain of channel function, characterized by an increase in the persistent Na + current leading to delayed repolarization and sustained APD. DelQKP is the deletion of three amino acids (QKP) at positions 1507-1509 on Na V 1.5 20 . This mutation is located in the DIII-DIV linker region of Na V 1.5, which is essential in the inactivation process of the channel 21 . This variant has already been characterized by using non-cardiac cells and heterozygote knock-in delQKP Scn5a Δ/+ mice 20,22 . In the present study, our approach was to use iPSC-CMs in order to provide a human and a native cardiac cell environment to evaluate and validate the effects of the Na V 1.5 variant. Our results in iPSC-CM confirmed that the QKP deletion cause a destabilization of the inactivated state of the channel, that is: a hyperpolarized shift of the inactivation, a faster inactivation decay time, a faster reactivation and finally an increased persistent Na + current. We obtained similar biophysical alterations of the channel that those reported previously in non-cardiac cells (tsa201 cells) and in cardiomyocytes from delQKP Scn5a Δ/+ mice 20,22 .
The strength of our model lies in the possibility of recording cardiac APs and measuring the effect of Na V 1.5 variants. We show for the first time that iPSC-CMs transfected with Na V 1.5/delQKP exhibit the essential LQT3 features. Their APs presented a delayed and a prolonged repolarization phase, leading to arrhythmic events such as EADs and DADs. The prolongation of the AP is most likely caused by the increase in the persistent Na + current seen in voltage-clamp recordings The continuous entry of Na + into the cell keeps the Ca 2+ channels activated and increases the release of Ca 2+ from the sarcoplasmic reticulum 23 . As consequence, the duration of the phase 2 of the AP increases. Moreover, the increase in APD may be caused also by the increased rate of Na V 1.5 recovery after inactivation as demonstrated for the I1768V variant 24 . The arrhythmic events caused by the lengthening of the AP were previously reported in the delQKP Scn5a Δ/+ mouse model. Therefore, our new-approach using Na V 1.5 KO iPSC-CMs is capable of reproducing what seen in mouse model but in cardiomyocytes-like of human origins.
An increase in the duration of APs at the cellular level results in a prolongation of the QT interval on the electrocardiogram of patients with LQT syndrome. Moreover, patients are at increased risk for the development of polymorphic ventricular tachycardia, specifically torsade de pointes (TdP) 25 . Thanks to our model, we were able to show that the delQKP variant induces a prolongation of APD at the origin of LQT, and more interesting, causes EADs, responsible for TdP 26 .
Interestingly, during electrophysiological recordings, we observed differences between the control and Na V 1.5 KO iPSC-CMs transfected with Na V 1.5/WT channel. Indeed, the Na + current exhibited a hyperpolarized threshold of activation, faster kinetic of inactivation and slower recovery from inactivation compared to control cells (iPSC-CMs not knocked-out for Na V 1.5). iPSC-CM transfected with Na V 1.5/WT also generated APs with a faster depolarization phase and a shortened repolarization phase (APD 20 and APD 50 only). As mentioned earlier, these differences may be caused by channel overexpression on iPSC-CMs. But interestingly, it could also be explained by the differences between the adult and the fetal isoform of Na V 1.5 27 . Although the efficiency of iPSC www.nature.com/scientificreports/ differentiation to cardiomyocytes has increased considerably with novel differentiation protocols, the level of iPSC-CM maturity remains an issue. This low level of maturity is characterized by the expression of numerous fetal isoforms to the detriment of adult isoforms, including those of Na V 1.5 10,28 . Thereby, when we compared the transfected Na V 1.5/WT with the control iPSC-CMs, in fact we mostly compared the effects between the adult and the fetal isoforms. However, when we compared the transfected Na V 1.5/WT with the transfected Na V 1.5/ delQKP, both channels were in the adult isoform background. This is an additional advantage of the new Na V 1.5 KO iPSC-CM, i.e., the ability to study Na V 1.5 variants in the desired background, either adult or fetal isoforms.

Conclusion
In conclusion, thanks to our Na V 1.5 KO iPSC model, we were able first to characterize the impact of the KO of Na V 1.5 on APs as well as on the propagation of voltage and Ca 2+ waves on iPSC-CMs. Secondly, as a proofof-concept, we were able to directly study the impacts on the APs of a known mutation linked to LQT3. The workflow of our novel approach is straightforward. The iPSC-CM from the Na V 1.5 KO cell line can be frozen and thawed on demand to study any Na V 1.5 variants by a simple transfection, as easily than with an expression system such as HEK293 cells with however several advantages as described above. Na V 1.5 variants and drug effects can now be straightforwardly and easily studied in a cardiac cellular environment. We hope that in the future, our new tool will help in the diagnosis and treatment of patients suffering of cardiovascular diseases.