Propranolol Attenuates Late Sodium Current in a Long QT Syndrome Type 3-Human Induced Pluripotent Stem Cell Model

Background Long QT syndrome type 3 (LQT3) is caused by gain-of-function mutations in the SCN5A gene, which encodes the α subunit of the cardiac voltage-gated sodium channel. LQT3 patients present bradycardia and lethal arrhythmias during rest or sleep. Further, the efficacy of β-blockers, the drug used for their treatment, is uncertain. Recently, a large multicenter LQT3 cohort study demonstrated that β-blocker therapy reduced the risk of life-threatening cardiac events in female patients; however, the detailed mechanism of action remains unclear. Objectives This study aimed to establish LQT3-human induced pluripotent stem cells (hiPSCs) and to investigate the effect of propranolol in this model. Method An hiPSCs cell line was established from peripheral blood mononuclear cells of a boy with LQT3 carrying the SCN5A-N1774D mutation. He had suffered from repetitive torsades de pointes (TdPs) with QT prolongation since birth (QTc 680 ms), which were effectively treated with propranolol, as it suppressed lethal arrhythmias. Furthermore, hiPSCs were differentiated into cardiomyocytes (CMs), on which electrophysiological functional assays were performed using the patch-clamp method. Results N1774D-hiPSC-CMs exhibited significantly prolonged action potential durations (APDs) in comparison to those of the control cells (N1774D: 440 ± 37 ms vs. control: 272 ± 22 ms; at 1 Hz pacing; p < 0.01). Furthermore, N1774D-hiPSC-CMs presented gain-of-function features: a hyperpolarized shift of steady-state activation and increased late sodium current compared to those of the control cells. 5 μM propranolol shortened APDs and inhibited late sodium current in N1774D-hiPSC-CMs, but did not significantly affect in the control cells. In addition, even in the presence of intrapipette guanosine diphosphate βs (GDPβs), an inhibitor of G proteins, propranolol reduced late sodium current in N1774D cells. Therefore, these results suggested a unique inhibitory effect of propranolol on late sodium current unrelated to β-adrenergic receptor block in N1774D-hiPSC-CMs. Conclusion We successfully recapitulated the clinical phenotype of LQT3 using patient-derived hiPSC-CMs and determined that the mechanism, by which propranolol inhibited the late sodium current, was independent of β-adrenergic receptor signaling pathway.

A number of studies reported genotype-phenotype correlations among those three major genotypes (Moss et al., 2000;Schwartz et al., 2001;Priori et al., 2004;Hobbs et al., 2006), and genetic testing for patients with LQT is highly recommended for identifying carriers in the families and determining the appropriate choice of gene-specific treatment (Ackerman et al., 2011;Skinner et al., 2019). Patients with LQT1 and LQT2 suffer from cardiac events that occur during exercise or emotional stress (Schwartz et al., 2001(Schwartz et al., , 2012. They are usually treated with β-blockers, as they are highly effective in reducing cardiac event and mortality rates (Schwartz et al., 2001;Priori et al., 2004). By contrast, patients with LQT3 experience cardiac events during rest or sleep (Schwartz et al., 2001(Schwartz et al., , 2012, and β-blocker therapy has resulted less effective or even harmful in preventing those cardiac events, according to small cohort studies and the clinical features of LQT3 (Moss et al., 2000;Schwartz et al., 2001;Priori et al., 2004). Recently, a large clinical cohort study, which consisted of 406 patients with LQT3, demonstrated that β-blocker therapy reduced the number of cardiac events in female patients >1 year old (Wilde et al., 2016), thus β-blocker therapy has regained importance as an optimal medicine for LQT3. However, the pharmacological mechanism by which β-blockers benefit LQT3 patients is still unclear.
In the present study, we investigated the cellular mechanism by which β-blockers affected late sodium currents in human LQT3 cardiomyocytes (CMs). To this end, we established a human induced pluripotent stem cell line (hiPSCs) from a young male carrying SCN5A-N1774D genotype and whose repetitive torsades de pointes (TdPs) were effectively treated with propranolol. In N1774D-hiPSC-derived CMs, prolonged action potential durations (APDs) and increased late sodium current at baseline were attenuated after propranolol administration. More importantly, we determined that late sodium current inhibition by propranolol was unrelated to β-adrenergic receptor signaling pathway in the LQT3 hiPSC-CM model.

Generation and Characterization of LQT3-hiPSCs
We collected peripheral blood mononuclear cells from a patient carrying SCN5A-N1774D after obtaining the written informed consent, and employed an integration-free method using episomal vectors to generate hiPSCs (Okita et al., 2013). We used three different patient-derived iPSCs clones in this study. As control hiPSCs lines, we used two different lines (201B7 and 253G1) generated from a healthy donor (Takahashi et al., 2007;Nakagawa et al., 2008). This study was approved by the Kyoto University ethics review broad and conformed to the principles of the Declaration of Helsinki.
The pluripotency of established LQT3-hiPSCs were assessed using immunostaining and teratoma assay (Yamamoto et al., 2017;Wuriyanghai et al., 2018). Briefly, the hiPSC colonies were fixed with 4% paraformaldehyde for 20 min at 4 • C. The cells were permeabilized in 0.2% Triton X-100 (Nacalai Tesque, Kyoto, Japan) and blocked with 5% FBS. The samples were stained overnight at 4 • C with the following primary antibodies: mouse monoclonal anti-OCT3/4 (1:50; Santa Cruz Biotechnology, Delaware, CA, United States), mouse monoclonal anti-SSEA4 (1:200; Santa Cruz Biotechnology), and mouse monoclonal anti-TRA 1-60 (1:200; Santa Cruz Biotechnology). The secondary antibody was donkey anti-mouse Alexa fluor 488 (1:1000, Invitrogen, Carlsbad, CA, United States). The nuclei were stained with DAPI (1:2000, Wako Pure Chemical Industries, Osaka, Japan). The specimens were observed under a fluorescence microscope (Biozero BZ-9000; KEYENCE, Osaka, Japan). For teratoma assay, the hiPSCs were injected into severe combined immunodeficiency disease scid/scid mice under the testis capsule. Tumor samples were surgically dissected at 8 weeks, fixed in 10% formalin and stained with hematoxylin and eosin. All animal experiments were performed in accordance with the "Guide for the Care and Use of Laboratory Animals" (2011) of the National Institutes of Health and the Regulation on Animal Experimentation at Kyoto University, and approved by Ethics Committee of Kyoto University.

Cardiomyocyte Differentiation
Cardiomyocyte differentiation was induced with an embryoid body (EB) differentiation system as previously described (Yang et al., 2008;Wuriyanghai et al., 2018). The hiPSCs aggregated to form EBs and cultured in suspension for 20 days. On the 20th day of culture, EBs were dispersed into small clusters using collagenase B (Roche Diagnostics GmbH, Mannheim, Germany) and trypsin EDTA (Nacalai Tesque) and were plated onto 0.1% gelatin coated dishes. After day 20, hiPSC-CMs were maintained in DMEM/F12 with 2% fetal bovine serum, 2 mmol/L L-glutamine, 0.1 mmol/L non-essential amino acids, 0.1 mmol/L β-mercaptoethanol, 50 U/ml penicillin, and 50 µg/ml streptomycin. The medium was renewed every 2-3 days. Differentiated 6-8-week-old hiPSC-CMs were enzymatically dispersed into single cells using collagenase B and trypsin EDTA. The cells were plated on 0.1% gelatin coated glass cover slips. Patch-clamp experiments were performed 4-7 days after the procedures. Cardiac differentiation was performed more than five times with each hiPSCs clone and the data pooled from different lines or clones among the control and N1774D group were analyzed.
The current-voltage relationships were fit with the Boltzmann equation: where I is the peak sodium current during the test pulse potential V. The parameters estimated by the fitting are reversal potential (V rev ), maximum conductance (G max ), and slop factor (k). Steady-state availability was fit to the Boltzmann equation: I/I max = [1 + exp (V−V 1/2 )/k] −1 to determine the membrane potential for V 1/2 (half-maximal inactivation) and k (slop factor). I max is the maximum peak sodium current. Data for the time course of recovery from inactivation were fitted by a biexponential equation: , where A f and A s represent fractions of each components, respectively. Time course of entry into the slow inactivation state and development of closed-state inactivation were fitted with a single exponential equation: where the transfer rate of sodium channels from closed-state to inactivated closedstate without an intervening opening state was measured by a double pulse protocol. Time course for development of closedstate inactivation was fit with a single exponential equation, Late sodium currents were recorded for 800 ms at −10 mV before and after 20 µM tetrodotoxin (TTX; Nacalai Tesque) application. The pipette solution contained (in mM): 5 NaCl, 70 CsCl, 40 C S Asp, 10 EGTA, 10 HEPES, 5 MgATP, pH 7.3 (adjusted with CsOH).
To assess the effect of propranolol for APD and sodium current, recordings were performed before and after the administration of 5 µM propranolol hydrochloride (FUJIFILM Wako, Osaka, Japan).

Statistics
Data are expressed as mean ± the standard error of measurement. Statistical comparisons were analyzed using appropriate student's t-test and paired t-tests. A p value of <0.05 was considered statistically significant.

Clinical Patient Profile
The proband was a 1-day-old infant who exhibited severe bradycardia at 27 weeks of gestation (Kato et al., 2014). His surface electrocardiogram showed an extremely prolonged QT interval (QTc 680 ms) with functional atrio-ventricular block and repeated TdPs immediately after his birth. He was suspected of long QT syndrome and thus propranolol was administered to suppress TdPs. But the treatment was not enough to suppress the electrical storm. Therefore, mexiletine was added and, as a result, TdPs disappeared.
Additionally, genetic testing identified a heterozygous SCN5A mutation, named p. N1774D, confirming the diagnosis of LQT3. Therefore, treatment of the patient with propranolol was discontinued. However, TdPs recurred after its interruption, and thus propranolol administration was resumed, and as a consequence, TdPs immediately disappeared.
Frontiers in Cell and Developmental Biology | www.frontiersin.org mutation (c.5319A > G) was confirmed in the patient-derived hiPSCs, but not identified in the control samples ( Figure 1C).

Prolonged APDs in N1774D-hiPSC-CMs
Action potentials of hiPSC-CMs were recorded using the current-clamp technique. Figure 2A shows typical ventriculartype APs. For this study, we measured APDs at 50% and 90% repolarization, APD 50 and APD 90 , respectively, maximum diastolic potential, and AP amplitude.

N1774D-hiPSC-CMs Exhibited Increased Late Sodium Current
The sodium channel currents in N1774D-hiPSC-CMs and control cells were recorded by using the voltage-clamp technique (Figure 3 and Table 2). Figure 3A illustrates representative whole-cell current traces of N1774D-hiPSC-CMs and control cells. As shown in Figure 3B, the peak sodium current densities of N1774D-hiPSC-CMs were significantly increased, compared to those of the control cells (N1774D: −333 ± 62 pA/pF, n = 15 vs. control: −175 ± 33 pA/pF, n = 15; p < 0.005; Table 2). Moreover, the steady-state activation of sodium channels, shifted to more negative potentials by 13 mV in N1774D-hiPSC-CMs, compared to that of the control cells ( Figure 3C and Table 2). No significant differences in the voltage dependence of inactivation and other kinetic properties were noted between control cells and N1774D-hiPSC-CMs.

Propranolol Shortened APDs and Reduced Late Sodium Current in N1774D-hiPSC-CMs
We assessed the effect of propranolol on APDs and late sodium current using hiPSC-CMs. Propranolol shortened APDs in N1774D-hiPSC-CMs (APD 90 : 440 ± 37 ms at baseline vs. 332 ± 37 ms after propranolol treatment, respectively, n = 10; at   Data are expressed as means ± standard error of measurement. Parameters were calculated from fitting individual experiments illustrated in Figure 3. The data pooled from different lines or clones among the control and N1774D group were analyzed. † p < 0.01 vs. Control. A, fractional amplitude; τ, time constant; V 1/2 , midpoint potential; k, slop factor, and n, number of tested cells, respectively. 1 Hz pacing; p < 0.01; Figure 2 and Table 1), but propranolol did not significantly affect APDs in the control cells (APD 90 : 272 ± 22 ms at baseline vs. 267 ± 26 ms after propranolol treatment, respectively, n = 11; at 1 Hz pacing; p = 0.15; Figure 2 and Table 1). Figure 5A shows the representative traces of late sodium current at baseline and after administration of 5 µM propranolol to N1774D-hiPSC-CMs. Propranolol treatment decreased the ratio of late/peak sodium current by approximately 25%, from 0.53 ± 0.05 to 0.40 ± 0.06% (n = 7; p < 0.001; Figure 5B). These results suggested that propranolol attenuated late sodium current and thus shortened APDs in N1774D-hiPSC-CMs.

Propranolol Directly Inhibited the Late Sodium Current in N1774D-hiPSC-CMs
Finally, to examine whether propranolol affected the sodium channels through G protein cascade or not, we recorded late sodium currents by using intrapipette guanosine diphosphate βs (GDPβs), a G protein inhibitor (Figure 5C). The late sodium current was significantly reduced by treatment with propranolol and 1 mM intrapipette GDPβs (baseline, 0.43 ± 0.07% vs. propranolol 0.28 ± 0.07%, n = 8; p = 0.01; Figure 5D). However, the reduction rate of late sodium current by propranolol treatment was not statistically different after addition of intrapipette GDPβs (without GDPβs: 24.7 ± 4.8% vs. with GDPβs: 30.9 ± 6.9%; p = 0.48; Figure 5E). Thus, propranolol directly blocked sodium channels and did not affect them via G protein pathway through β-adrenergic receptors.
Additionally, the suppression of late sodium current by propranolol was larger than that of the peak current whether in absence or presence of intrapipette GDPβs (Figure 6). No significant difference in the suppression rate of the peak and the late sodium currents were observed whether in absence or presence of GDPβs (Figure 6).

Patient-Specific iPSC-CMs as a Cell
Model for Studying SCN5A-N1774D Associated LQT3 The efficacy of β-blockers has been controversial in patients with LQT3 (Moss et al., 2000;Schwartz et al., 2001;Priori et al., 2004). Recently, a large clinical study demonstrated the efficacy of β-blockers on patients with LQT3 (Wilde et al., 2016); Inset shows late sodium current between 500 and 600 ms. Tetrodotoxin-sensitive current was calculated by subtraction. (B) Mean late sodium current of control and N1774D channels. Late sodium current is presented as the percentage of late sodium current to peak sodium current. The data pooled from different lines or clones among the control and N1774D group were analyzed. The late sodium current was significantly increased. * p < 0.001, vs. control.  In the presence of intrapipette GDPβs, the rate of reduction was also larger in late sodium current compared to in peak sodium current. * p < 0.001, vs. peak sodium current.
however, the pharmacological mechanism of β-blockers in LQT3 remains unclear.
Recently, the iPSC technology enables us to analyze the disease-causing mechanism of inherited arrhythmia disorder using self-beating cardiomyocytes with the same genetic background as the patient, and several studies of hiPSC-based disease modeling of LQT3 have been reported (Davis et al., 2012;Fatima et al., 2013;Ma et al., 2013;Terrenoire et al., 2013;Spencer et al., 2014;Yoshinaga et al., 2019). To elucidate the mechanism, we established the LQT3-hiPSC-CM cell model from a patient carrying the SCN5A-N1774D mutation who was effectively treated with propranolol. In accordance with his clinical course, propranolol attenuated the increased of late sodium current and prolonged APDs in N1774D-hiPSC-CMs, compared to the control cells. In addition, by using GDPβs, an inhibitor of G proteins, we demonstrated that propranolol inhibited the late sodium current through a pathway other than β-adrenergic receptor signaling pathway. Thus, the efficacy and pharmacological mechanism of propranolol have never, to our knowledge, been hitherto elucidated in an LQT3-hiPSC-CM model.
In our previous report (Kato et al., 2014), we analyzed SCN5A-N1774D channels using a heterologous expression system in HEK 293 cells and determined that N1774D channels displayed increased peak sodium current densities (2.2 times), a hyperpolarized shift of steady-state activation curve by 7.9 mV, and increased late sodium current densities by 1.7 times, compared to those of wild-type channels. In our LQT3-hiPSC model, the steady-state activation curve was negatively shifted by 12.8 mV and the peak and late sodium current densities were 1.9 and 9.3 times greater, respectively, when compared to those of the control hiPSC-CMs (Figures 2, 3). Considering that our hiPSCs carried the heterozygous SCN5A-N1774D mutation, the electrophysiological properties of hiPSC-CMs showed similar trends but different degrees of change in sodium channel kinetics to those of the heterologous expression system, possibly it might result from the differences between the two cell types.

β-Blocker Therapy for Patients With LQT3
β-blocker therapy is effective for treatment of patients especially with LQT1 and LQT2, who experience cardiac events after adrenergic stimulation, such as exercise or emotional stress. In LQT1 and LQT2, β-blockers clinically reduce the rate of cardiac events and mortality (Moss et al., 2000;Schwartz et al., 2001;Priori et al., 2004). Therefore, based on the clinical evidence, β-blocker therapy is recommended as the first-line treatment for LQT1 and LQT2 (Priori et al., 2013).
In contrast, Schwartz et al. (2001) reported that the recurrence of syncope and death rates in patients with LQT3 were higher (50% and 17%, respectively, n = 18) than those with LQT1 and 2. Additionally, Priori et al. (2004) reported that 9 out of 28 (32%) LQT3 patients experienced a cardiac event while on β-blocker therapy at a rate four-fold higher than that of LQT1 patients. Even though the number of LQT3 patients enrolled in the study was small, β-blocker therapy in LQT3 was suggested to be ineffective in preventing fatal arrhythmias and even contraindicated or harmful due to the clinical feature of LQT3, that cardiac events mainly occur at rest, during sleep or bradycardia.
However, recently, the efficacy of β-blocker therapy was analyzed in a large multicenter LQT3 cohort study (Wilde et al., 2016). Wilde et al. (2016) investigated 406 LQT3 patients and demonstrated that β-blocker therapy reduced the risk of lifethreatening cardiac events in female patients >1 year of age. In the present study, our patient suffered from repetitive TdPs since birth and propranolol was definitely effective in suppressing those events. Despite accumulating clinical evidence that β-blocker therapy might be beneficial to LQT3 patients, the underlying mechanism by which β-blockers suppress malignant ventricular arrhythmias remains mostly unclear. In our LQT3-hiPSC model, prolonged APDs at baseline were attenuated after propranolol administration (Figure 5), which was consistent with the clinical efficacy of β-blocker therapy in LQT3.
Why Are β-Blockers Effective in Patients With LQT3?

Cardiac Sodium Channel and β-Adrenergic Stimulation
Cardiac sodium channels are known to be modulated by β-adrenergic stimulation via phosphorylation sites on protein kinases A and C, PKA and PKC, respectively (Grant, 2009). In cardiac sodium channels, PKA activation increased the peak sodium current density and shifted both steady-state activation and inactivation to the hyperpolarized direction (Matsuda et al., 1992;Ono et al., 1993). On the contrary, PKC activation reduced peak sodium current densities and caused a negative shift of the steady-state inactivation curve (Valdivia et al., 2009).
PKA activation did not affect the late sodium current in wild-type sodium channels. On the other hand, in LQT3-related mutant sodium channels, PKA activation had little or no effect on Y1795C and Y1795H channels, but it enhanced the late sodium current in KPQ, D1790G, R1623G, and V2016M channels (Chandra et al., 1999;Tateyama et al., 2003;Tsurugi et al., 2009;Chen et al., 2016).
In LQT3 mouse models, propranolol efficacy to prevent lethal arrhythmias has been contradictory. Fabritz et al. (2010) showed that chronic propranolol therapy did not suppress the carbachol-mediated ventricular arrhythmias, such as TdPs, in a heterozygous knock-in SCN5A-KPQ LQT3 mice. On the other hand, Calvillo et al. (2014) reported that propranolol pretreatment prevented carbachol-mediated malignant ventricular tachyarrhythmias in the SCN5A-KPQ LQT3 knock-in transgenic mice. Thus, the pharmacological mechanism of β-blockers in LQT3 mice and human cardiomyocytes derived from patients with LQT3 has remained uncertain until now. In the present study, we demonstrated that propranolol attenuated the prolonged APD, by reducing the late sodium current in the patient-derived LQT3-hiPSC-CMs, and provided new supportive evidence of its efficacy on LQT3 by using a human CM model.

Propranolol Has an Inhibitory Effect on Sodium Channels but Not Through β-Adrenergic Receptor Blockade in an LQT3-hiPSC Model
Propranolol was previously reported to have a unique blocking effect on sodium channels, which was not identified in other β-blockers (Wang et al., 2010). In a heterologous expression system using HEK293 cells, only propranolol, not metoprolol or nadolol, blocked the cardiac sodium channels similar to local anesthetics, and not by acting as β-blockers (Wang et al., 2010). They also demonstrated that the block depends on a critical D4/S6 residue, F1760, involved in local anesthetic effects. In addition, Bankston and Kass (2010) demonstrated that propranolol decreased late sodium current in F1473C and KPQ mutant channels. In this study, we showed no significant difference in the suppression rate of late sodium current by propranolol in presence, or not, of intrapipette GDPβs, an inhibitor of G proteins ( Figure 5E). This was found to be consistent with the previous reports that propranolol has a sodium channel blocking effect through interfering with the cascade other than G protein cascades (Bankston and Kass, 2010;Wang et al., 2010).
Interestingly, Bankston and Kass (2010) also reported that propranolol within 40 nM to 40 µM range preferential inhibited late but not peak sodium currents. In this study, administration of 5 µM propranolol reduced late sodium current, but it slightly decreased peak sodium current (Figure 6 and Table 3).
Regarding the plasma concentration of propranolol, it was reported that there was wide inter-individual variation from 0.04 to 4 µM in peak plasma concentrations, and an increase in the dose up to plasma levels of approximately 0.4 µM was suggested when sufficient therapeutic effects were not achieved (Woosley et al., 1979). Unfortunately, the plasma concentration of propranolol in the patient in this study was not available, and we recorded APs and late sodium current with only 5 µM propranolol using hiPSC-CM model. In addition, we did not examine the effect of propranolol on the kinetics of sodium channel current, which was demonstrated in a heterologous expression system (Wang et al., 2010). Therefore, further examination is needed to reveal the detailed relationship between propranolol dose and the effects on electrophysiological properties in hiPSC-CMs.

Study Limitations
In the present study, we did not employ the isogenic control line which will prove the phenotype is directly imputable to the mutation. In addition, we did not examine the efficacy of other Data are expressed as means ± standard error of measurement. p value for vs. with GDPβs in reduction rate. The data pooled from different clones in N1774D-hiPSC-CMs were analyzed. β-blockers. Another limitation is that hiPSC-CM is an in vitro cellular model; therefore, it is impossible to assess the role of sympathetic nervous system. Clinically, left cardiac sympathetic denervation was reported to inhibit cardiac events in LQT3 patients , which indicated that β-blockers would have an antiarrhythmic effect by suppressing sympathetic nervous system in LQT3.

Conclusion
We successfully recapitulated the LQT3 disease phenotype in the SCN5A-N1774D-hiPSC-CMs. Additionally, the propranolol efficacy and mode of action were demonstrated in this cell model. Specifically, propranolol blocked sodium channels by means of a unique mechanism unrelated to β-adrenergic signaling pathway, by preferentially inhibiting late rather than peak sodium current.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding author.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Ethics Committee of Kyoto University. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin. The animal study was reviewed and approved by Ethics Committee of Kyoto University.