A Novel SCN5A Variant Causes Temperature-Sensitive Loss Of Function in a Family with Symptomatic Brugada Syndrome, Cardiac Conduction Disease, and Sick Sinus Syndrome

Introduction: Brugada syndrome (BrS) is an inherited arrhythmia syndrome associated with an increased risk of sudden cardiac death. SCN5A is the most important disease-modifying gene for BrS, but many SCN5A variants have not been functionally characterized. Furthermore, the temperature dependency of SCN5A is only rarely explored in in vitro analyses. Methods: The clinical phenotype of the affected family was assessed by medical history, ECGs and ajmaline challenge. Whole-cell patch clamp recordings were performed on HEK 293T cells expressing Nav1.5-G1712S, a novel SCN5A variant found in the symptomatic family. Results: Three male family members had experienced sudden cardiac death, sudden cardiac arrest, and rhythmogenic syncopes. Beside a positive ajmaline challenge with demarcation of a Brugada type 1 ECG, 1 patient also showed evidence of symptomatic cardiac conduction disease and sick sinus syndrome (SSS). In patch clamp analyses, Nav1.5-G1712S generated reduced peak currents as compared to the wild type. At body temperature, Nav1.5-G1712S additionally exhibited an enhanced slow inactivation and an impaired recovery from inactivation. Conclusion: We conclude that G1712S is a pathogenic SCN5A loss-of function mutation at physiological temperature associated with an overlapping presentation of BrS, SSS, and cardiac conduction disease.


Introduction
Brugada syndrome (BrS) is an inherited arrhythmia syndrome associated with an increased risk of malignant cardiac arrhythmias. The variable phenotype of BrS makes an accurate individual risk stratification prior to the implementation of invasive preventive treatment options like an implantable cardioverter-defibrillator (ICD) or a pacemaker challenging. [1] Overlapping phenotypes with other inherited arrhythmia syndromes such as sick sinus syndrome (SSS), cardiac conduction disease and long QT syndrome complicate the management of these patients. [1,2] Genetic testing is being increasingly performed in symptomatic patients and their family members. A thorough understanding of the pathogenicity of any given variant may lead to a more efficient risk stratification. [1] SCN5A encodes for the -subunit of the cardiac sodium channel Nav1.5 and is probably the most relevant disease-modifying gene for BrS. [3] Variants of SCN5A are identified in about 20% of patients, but only some of them are actually considered causative. [4,5]. Typically, these variants associated with BrS display loss-of-function properties as determined by in vitro patch clamp. [1,6] These experiments are usually conducted at room temperature, which may be a critical limitation as some biophysical properties of sodium channels are temperature-sensitive. [7][8][9] We identified the SCN5A variant G1712S in a family with symptomatic BrS. This variant was already identified in a single patient with BrS [6], but it has not been characterized by means of in vitro electrophysiology and no clinical data have been published. In this study we present the phenotype of the affected family as well as biophysical properties of the mutant Nav1.5-G1712S.

Materials and Methods
Clinical analysis A family with BrS and familial sudden cardiac death was identified. All family members were encouraged to undergo complete clinical and genetic work-up including medical and family history, physical examination, repetitive baseline ECGs, stress test and ajmaline challenge. The study was approved by the ethics committee of Hannover Medical School (no. 1673-2013).

Mutagenesis
Human (h) Nav1.5 cDNA was PCR amplified and subcloned into the pTracer-SV40 vector (Invitrogen, Carlsbad, CA, USA). Nav1.5-G1712S and Nav1.5-G1712C mutants were generated by site directed mutagenesis of hNav1.5 cDNA in pTracer-SV40 using the quikchange lightning site-directed mutagenesis kit (Agilent, Waldbronn, Germany) according to the instructions of the manufacturer. Mutants were sequenced to verify intended amino acid exchanges, and to exclude further channel mutations.
Patch clamp experiments Na + currents from transfected HEK 293T cells were recorded using the whole-cell configuration of the patch clamp technique with a HEKA EPC9 patch-clamp amplifier (HEKA Elektronik, Lambrecht, Germany). Data were sampled at 20 kHz and filtered at 5 kHz. Data were recorded using the Pulse software (v 8.80, HEKA Electronik) and analyzed with the Fitmaster software (HEKA Electronik, Lambrecht, Germany). Curve fitting and statistical analyses were performed using Origin 8.5 software (Microcal Software, Northampton, MA). The bath solution contained (mM): 70 NaCl, 70 Choline Chloride, 3 KCl, 1 CaCl2,1 MgCl2, 10 HEPES and 15 glucose. The pH was adjusted to 7.4 with tetraethylammonium hydroxide (TEA-OH). The pipette solution contained (mM): 140 CsF, 10 NaCl, 10 HEPES, 1 EGTA, pH 7.4 (CsOH). Pipettes were pulled from borosilicate glass (GB 100 TF-8P, Science Products, Berlin, Germany) and their resistance ranged from 1.0 to 2.5 MOhm. Serial resistance (5.2 ± 2.4 MOhm) was compensated by 60-80% to minimize voltage errors. Capacitance artifacts were compensated using the amplifier circuitry. Patch clamp experiments were carried out at 22-24°C (room temperature), or at 35-37°C. For the experiments at physiological temperature, the external solution was preheated to 37°C and during the experiments the temperature was maintained using a heating plate filled with temperature-controlled solution.
Statistical analysis Clinical and electrophysiological data were presented as mean ± S.D. and mean ± S.E., respectively. Chi-square test and Fisher's exact test were used for comparison of categorical variables. Statistical significance was calculated with one-way ANOVA with post hoc Tukey-Test. A p < 0.05 was considered statistically significant. Calculations were performed using Origin 8.5 (Microcal Software, Northhampton, MA). Data are presented as mean ± standard error of mean (S.E.M.).

Family history
The index patient (III-4, Fig. 1A and B) survived sudden cardiac arrest at the age of 11 months. The initial rhythm detected was ventricular fibrillation. The boy suffered severe brain damage after 1h of cardiopulmonary resuscitation. The parents were asymptomatic, but the mother (II-2) had lost two pregnancies, and her father (I-1) had died of sudden cardiac arrest at the age of 38. The initial diagnostic workup of the boy (III-4) including echocardiography, repeat ECGs and Holter ECG did not result in any diagnosis. Genetic testing revealed a heterozygous missense point mutation on SCN5A, leading to the exchange of a glycine (G) to serine (S) at position 1712 (Gly1712Ser; 5134 G>A). The same SCN5A variant was detected in the mother. After considering the family history and according to her personal wish, the mother received an ICD for primary prophylaxis. Two years later she gave birth to male twins. One was a stillbirth, and the other (III-6) was born healthy but tested positive for SCN5A-G1712S. At the age of 3 years, the surviving twin suffered from a suspected rhythmogenic syncope during a febrile episode and received an ICD/pacemaker for secondary prophylaxis. After the ICD implantation both patients (II-2 and III-6) have not experienced any recurrences of spontaneous arrhythmia nor any device-associated complications. No relevant clinical information was available from II-1 and his children.

Clinical reassessment
At presentation in our outpatient clinic, we reviewed all clinical reports and ECGs available. While there were no pathological findings in the mother, several resting ECGs of her sons III-4 and III-6 were abnormal. In repeat ECGs III-6 showed a bradycardic sinus rhythm and prolonged PQ and QRS intervals for his age, with nonspecific depolarization and repolarization abnormalities (Fig. 1C). III-4 showed intermittent incomplete right bundle branch block with ST-elevations in V1 and V2 (Fig. 1D). Review of the emergency report of III-6 at the time of the syncope revealed the initial detected rhythm as bradycardia with an escape rhythm of 40/min. Intrahospital telemonitoring showed episodes of sinus bradycardia as well as a 2nd degree AV-block (Fig. 1E). None of the reviewed ECGs showed a relevant prolongation or shortening of the QT interval. Due to the suspected familial channelopathy, we performed an ajmaline challange in the mother. The test was positive with a Brugada type 1 ECG after administration of 35 mg of ajmaline ( Fig. 2A), allowing the diagnosis of familial BrS. III-6 also produced a positive ajmaline (30 mg) test performed at the age of 11, e.g. ajmaline induced coved-type ECG elevations in the right precordial leads. At the same time, he developed sinus arrest with stimulation of the ICD/pacemaker (Fig. 2B). Taken together, we found the clinical manifestation of ventricular fibrillation in III-4, SSS as well as cardiac conduction disease in III-6, sudden cardiac death in I-1 and an asymptomatic phenotype in the only female variant carrier II-2.

Discussion
We identified the missense mutation G1712S in SCN5A in a family with symptomatic BrS and found that the mutation causes a loss-of-function as Nav1.5-G1712S generated smaller inward currents as compared to Nav1.5-WT. Surprisingly, an enhanced slow inactivation and an impeded recovery from fast and slow inactivation of Nav1.5-G1712S became evident only when experiments were conducted at physiological temperatures. Thus, the mere analysis at room temperature may have promoted misinterpretation of Nav1.5-G1712S as a polymorphism with a more modest loss-of-function phenotype. In fact, similar biophysical properties have been proposed as pathogenic for other Brugada-causing Nav1.5 mutations. [2,9] Together with the clinical information and previously published in silico analysis, Nav1.5-G1712S can be declared as a pathogenic mutation with high probability to cause BrS as well as possibly SSS and cardiac conduction disease. [5] G1712S was previously reported to be associated with a Brugada phenotype in a single patient, and an in silico analysis had proposed a likely pathogenicity. However, that report lacked both clinical data and an electrophysiological characterization. [6] In fact, around 10% of published Brugadaassociated SCN5A-variants are suspected to be non-pathogenic. [4] The observation that some patients with a BrS phenotype completely lack Nav1.5-mutations has raised questions about the relevance of Nav1.5 mutations as causative for BrS [11]. This notion is supported by studies revealing rather complex, and in part Nav1.5-independent genetic backgrounds of BrS [12]. Nevertheless, our findings support the notion that any novel mutation detected warrants an extensive analysis, including electrophysiological experiments in order to define its possible pathogenicity. [5]. Performing these electrophysiological experiments in transiently transfected HEK 293T cells however, has several limitations. It may conceal properties of Nav1.5-G1712S expressed in native cardiomyocytes, known to express important interacting proteins which could be relevant for BrS. Our analyses did not contain -subunits which modify the function of Nav1.5 [13]. The question if G1712S has an impact on the interaction with -subunits, warrants future exploration. While a considerable number of Brugada-associated mutations have been identified within the poreloops and S5-S6 segments of Nav1.5, variants in these regions were rarely found in healthy controls. [6] By demonstrating the loss-of-function phenotypes of both Nav1.5-G1712S and G1712C located within the pore-loop of domain IV, our data support this correlation between topology and pathogenicity. An important study from Li and co-workers recently mapped known mutations of Nav1.5 associated with cardiac disorders based on a structure revealed by Cryo-EM [14]. In fact, Nav1.5-G1712S was included in this study and was reported to be located in a segment which stabilizes the selectivity filter. A close-by mutation, Nav1.5-D1714G was found to display similar effects on current density as well as slow inactivation in functional studies. [9] In regard to the temperature-sensitivity found in our study, Amin et al. reported similar observations on Nav1.5-D1714G. Mutations in other regions have also shown temperature-dependent properties. [8] Previous studies have associated corresponding sites on other domains with altered slow inactivation properties. [15] Although the process of slow inactivation remains incompletely understood, a conformational change of the pore-region is thought to be an underlying mechanism. [16] Voltagedependent activation and steady-state fast inactivation were not affected in Nav1.5-G1712S. These channel properties have been associated with distinct regions on SCN5A. [17] We show that the G1712S exchange did influence recovery from fast inactivation. While this property has not been linked to the pore-region, the mechanism is poorly understood and seems to be influenced by various locations. Although a correlation between genotype and phenotype in BrS does exist, the evidence allowing genetics-based risk stratification and decision making is insufficient. [18] With 3 out of 3 male family members being symptomatic with sudden cardiac arrest or syncope, our data support the notion that loss-of-function mutations on SCN5A located within the pore-region are associated with a severe phenotype. [18,19] The occurrence of sinus arrest during ajmaline challenge further demonstrates the severity of the phenotype. [20] Ajmaline unmasking sinus node dysfunction has been described before. [21] Since overlap syndromes between BrS and SSS exist and may be caused by common SCN5A mutations, G1712S may cause BrS as well as SSS. This association has been described for only a few mutations, but none within the pore-loop of domain IV so far. SSS tends to manifest earlier in life than the Brugada phenotype and shows male predominance. Some patients with SSS and BrS were even implanted a pacemaker instead of an ICD. [22][23][24] Of note, the patient III-6 not only showed features of SSS, but also of cardiac conduction disease with prolonged PQ and QRS intervals at rest and documented intermittent 2nd degree AV block, further supporting the relevance of bradycardia in this case. Considering this and the ECG recordings of III-6, it is likely that the rhythmogenic syncope was caused by bradycardia rather than tachycardia. Consequently, no further events occurred after implantation of the ICD/pacemaker. Van den Berg et al. have previously reported several cases of bradycardia-associated death in patients with BrS. [21] Together with our observation, these data suggest that clinicians should look out for clinically relevant bradycardia in patients with BrS. This is especially important when considering implantation of an ICD without antibradycardia function such as the subcutaneous ICD, which is increasingly considered in young patients with channelopathies. [1] Conclusions Taken together, we characterized a new pathogenic SCN5A mutation, G1712S. It was associated with symptomatic BrS, cardiac conduction disease and SSS requiring pacemaker therapy. The mutant caused a loss-of-function phenotype, but some effects were only obvious at physiological temperatures. Together with previous reports, our data suggest that reliable functional studies should be carried out at physiological temperature in addition to room temperature. Moreover, our findings stress the relevance of overlap syndromes and bradycardia in patients with BrS. Representative resting ECGs of III-6 (C, at the age of 7y) and III-4 (D, at the age of 3y). E. Intrahospital ECG monitoring of III-6 after syncope. Fig. 2. A, B. Ajmaline challenges in II-2 (A) and III-6 (B). Modified ECG with V1-6 precordial (V1-3 3rd ICS, V4-6 4th ICS) before (l) and after (r) administration of 30 and 35 mg ajmaline respectively. Fig. 3. A. Cartoon of the -subunit Nav1.5 displaying the location of the residue G1712 in the pore-forming loop. B, D. Representative current traces displaying voltage-dependent activation of Nav1.5-WT (black, n= 19 at both temperatures), Nav1.5-G1712S (red, n= 20 respectively 22) and Nav1.5-G1712C (blue) at room (B, 22-24°) or near body (D, 35-37°C) temperature. Cells were held at -120 mV and 100 ms long depolarizing pulses were applied as depicted in the inset. C, E Peak current amplitudes of Nav1.5-WT, Nav1.5-G1712S and Nav1.5-G1712C mediated currents displayed as dot plots. Data are shown as mean ± S.E.M. F, G. Normalized current-voltage plots of Nav1.5-WT and Nav1.5-G1712D recorded at room temperature (F) or at body temperature (G). H, I. Conductancevoltage relationships converted from the peak IV-plot shown figure in F and G. Data were fitted with the Boltzmann function. Fig. 4. A, B. Steady-state fast inactivation plots for Nav1.5-WT (n= 18, respectively 19) and Nav1.5-G1712S (n= 18 at both temperatures) recorded at room (A, 22-24°) or near body (B, 35-37°C) temperature. Cells were held at -150 mV and fast inactivation was induced by 100 ms long depolarizing pulses followed by the test-pulse at 0 mV (see inset). Mean data were normalized and fitted with the Boltzmann function. C, D. Plots demonstrating recovery from fast inactivation of Nav1.5-WT (n= 22, respectively 25) and Nav1.5-G1712S (n= 18, respectively 17) recorded at room (C, 22-24°) or near body (D, 35-37°C) temperature. Cells were held at -150 mV and two 20 ms long consecutive pulses to -10 mV with a variable interval ranging from 0.1 to 1683 ms were applied (see inset). Normalized data were fitted with a single exponential to obtain the time constant ( ). E, F. Plots displaying voltage-dependent slow inactivation for Nav1.5-WT (n= 19, respectively 9) and Nav1.5-G1712S (n= 13, respectively 12) recorded at room (E, 22-24°) or near body (F, 35-37°C) temperature. Cells were held at -150 mV and slow inactivation was induced by 10 s long depolarizing pulses followed by an inter-pulse at -150 mV allowing recovery from fast inactivation and the test-pulse at 0 mV (see inset). Mean data were normalized, and a B-spline was drawn between data points to guide the eye. G, H. Plots demonstrating recovery from inactivation of Nav1.5-WT (n= 10, respectively 11) and Nav1.5-G1712S (n= 11, respectively 12) recorded at room (G, 22-24°) or near body (H, 35-37°C) temperature. Cells were held at -150 mV and inactivation was induced by a 10 s long pre-pulse at 0 mV followed by recovery at -150 mV for a variable duration before the test-pulse to 0 mV was applied. Normalized data were fitted with a double exponential to obtain the time constants and 2.