Phenotypic Severity of SCN5A-Related Bradycardia Is Independent of Dominant-Negative and Coupled Gating Effects

Abstract

BACKGROUND:

Pathogenic SCN5A variants are associated with inherited arrhythmias such as long-QT syndrome, Brugada syndrome, and sick sinus syndrome. While Na_v1.5, an α-subunit of the cardiac sodium channel encoded by SCN5A, has been considered to function as a monomer, recent studies reveal that a reduction of sodium current in wild-type Na_v1.5 can be caused by dimerization with loss-of-function mutated Na_v1.5 through dominant-negative effects. However, the clinical significance of the dominant-negative effect remains unclear.

METHOD:

We genetically screened a family who presented with sick sinus syndrome and sudden cardiac death. Whole-cell patch-clamp study using HEK293 (human embryonic kidney) cells coexpressing wild-type and variant SCN5A was performed. Channel dimerization was assessed by coimmunoprecipitation and proximity ligation assays. Also, the effects of difopein, a high-affinity inhibitor of Na_v1.5 interaction via 14-3-3 proteins, were evaluated.

RESULTS:

The proband carried compound heterozygous variants p.T1396P and p.G833R. The whole-cell mode patch-clamp techniques demonstrated that the p.T1396P showed a dominant-negative effect on the peak sodium currents (37% decrease in I_Na) and altered gating properties (5.6-mV shift in steady-state inactivation) when expressed with wild-type SCN5A. These effects were abolished by difopein. p.G833R showed no dominant-negative or coupled gating effect but still formed dimers. The proband developed earlier and more severe bradycardia than the mother, who only carries p.T1396P, suggesting that loss of coupled gating effect contributed to the severe phenotype.

CONCLUSIONS:

Our findings suggest that coupled gating may be physiologically important for normal Na_v1.5 function, and its loss can exacerbate disease severity.

What is Known?

• Loss-of-function variants of the Na_v1.5 channel can interfere with the function of wild-type channels via dominant-negative or coupled gating effects, but their clinical significance has remained unclear.

• Na_v1.5 channel dimerization and coupled gating are reported to be mediated through 14-3-3 protein interactions.

• Coupled gating has primarily been discussed as a mechanism explaining the dominant-negative effect caused by loss-of-function variants.

What the Study Adds

• A patient carrying compound heterozygous variants (p.G833R, which lacks coupled gating/dominant-negative effects, and p.T1396P, which has both) presented with earlier onset and more severe bradycardia compared with their mother, who carried p.T1396P alone.

• The p.G833R variant retained its ability for physical dimerization (confirmed by coimmunoprecipitation

  and proximity ligation assay) but exhibited no electrophysiological coupled gating effect. This suggests that the Na_v1.5 channel physical interaction is functionally independent of coupled gating.

• Coupled gating may be a necessary physiological mechanism for normal Na_v1.5 function in vivo, and its loss can exacerbate disease severity.

The voltage-gated sodium channels are expressed in various excitable cells and are essential for the generation and propagation of action potentials.¹ Among them, the human Na_v1.5, encoded by the SCN5A gene, plays a crucial role in cardiac excitability. Pathogenic variants in SCN5A are associated with a spectrum of inherited arrhythmic diseases, including long-QT syndrome, Brugada syndrome (BrS), and sick sinus syndrome.^2–4

Unlike potassium channels that form functional oligomers, the α-subunit of Na_v1.5 has long been considered to form an ion pore as a monomer. However, recent studies show that coexpression of wild-type (WT) and loss-of-function (LOF) types of the SCN5A variants can reduce peak sodium currents. This phenomenon, known as a dominant-negative (DN) effect, suggests that LOF variants can interfere with WT channel function. Adding to that, the coupled gating effect has also been suggested through the demonstration that the voltage dependence of the WT channel was affected by LOF variants. The DN and coupled gating effect are reported to be mediated through channel dimerization, facilitated by 14-3-3 protein interactions.^5–8 Despite these findings, the clinical relevance of the DN effect has yet to be uncovered. A previous study reported a higher penetrance in affected families carrying DN variants, but no clear association with phenotype severity was observed.⁸ Thus, we aim to find DN SCN5A variants in various genetic arrhythmias. In this study, we found a family who showed sick sinus syndrome and sudden cardiac death harboring 2 SCN5A variants: p.T1396P and p.G833R. The patch-clamp study revealed that p.T1396P, but not p.G833R, showed a DN and coupled gating effect. We further investigated the effects of these variants on activation and inactivation kinetics. In addition, we also assessed the association between disease phenotype severity and the presence of these effects.

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request. This study was approved by the institutional review board of Shiga University of Medical Science (G2011-128). All patients provided written informed consent before genetic investigations in accordance with the Declaration of Helsinki. Detailed methods, including genetic analyses, electrophysiological recordings, biochemical assays, and statistical procedures, are provided in the Supplemental Material.

Results

Clinical Characteristics

Figure 1A shows the pedigree of our patients. The proband is a 32-year-old female who experienced a sudden loss of consciousness late at night. She presented with recurrent episodes of ventricular tachycardia (VT) of right ventricular origin and severe sinus bradycardia (Figure 1B). In the acute phase, her echocardiography revealed an aneurysmal change in the left ventricular apex, which was diagnosed as Takotsubo cardiomyopathy because of its characteristic morphology. However, she had no apparent suspicious external triggers for the event, such as emotional stress. Moreover, the VT morphology did not correspond to the echocardiographic findings as there were no structural abnormalities in the right ventricle. Although the echocardiographic abnormality was completely resolved over time, the patient’s sinus bradycardia persisted into the chronic phase, and an implantable cardioverter defibrillator was implanted due to the initial VT. Since the device implantation, she has remained free from VT recurrence. In addition, over more than 10 years of follow-up, she never developed any findings suggestive of arrhythmogenic right ventricular cardiomyopathy, such as right ventricular structural abnormalities or epsilon waves on ECG. Her mother also had a history of sinus bradycardia in her 40s after taking a class I antiarrhythmic agent due to paroxysmal atrial fibrillation. Shortly after administration, she lost consciousness due to sinus bradycardia, and temporary cardiac pacing was performed (ECGs were not available). Her sinus rhythm subsequently recovered without any permanent pacing device. At the age of 70 years, she manifests only right bundle branch block and first-degree atrioventricular block (Figure 1B). The proband’s elder sister was found to have bradycardia and left ventricular enlargement at school-based routine ECG screening in childhood. She underwent a pacemaker implantation at the age of 9. At 17 years old, she died suddenly, possibly due to tachycardia, though the electrogram data from the pacemaker was unavailable. The proband’s elder brother and father have no history of arrhythmic events, and their ECGs are normal.

Figure 1.

A family pedigree and the SCN5A variants. A, Pedigree of an affected family harboring 2 SCN5A variants. Filled symbols indicate individuals with bradyarrhythmia. Squares represent males; circles represent females. B, ECG recordings of individuals I-2 and II-3 (a: sinus bradycardia and b: ventricular tachycardia). C, Topology of Na_v1.5 indicating the locations of the 2 identified variants. ICD indicates implantable cardioverter defibrillator; and PM, pacemaker.

Identification of SN5A Variants

We identified 2 SCN5A variants, c.2497 G>A (p.G833R) and c.4186 A>C (p.T1396P), in the proband in a compound heterozygous fashion (Figure 1A and 1C). The minor allele frequency of p.G833R is 0.029% in East Asian, and p.T1396P has not been found in general populations according to the gnomAD, version 4.1.0 (https://gnomad.broadinstitute.org/gene/ENSG00000183873?dataset=gnomad_r4). Her mother carried only the heterozygous p.T1396P, and her father carried the heterozygous p.G833R. The genomic sample from the proband’s deceased sister was unavailable. According to the American College of Medical Genetics and Genomics guidelines,⁹ both variants were classified as variants of unknown significance. No variants were detected in desmosome-related genes.

Electrophysiological Characteristics of the Variant Channels

Patch-clamp analysis was performed in HEK293 cells transfected with various combinations of WT- and mutant-SCN5A plasmid (Figures 2 and 3; Table). The transfection conditions are shown in Figure 2A. The amounts of plasmids were 0.5 µg for each type (ie, WT and variants), and the total amount of plasmid was adjusted to 1.5 μg with the pEGFP-C3 plasmid. When only SCN5A p.T1396P plasmids (0.5 µg) were transfected, the cell did not generate I_Na, indicating that p.T1396P is an LOF variant (Figure 2B, filled purple triangle). In the cells cotransfected with WT and p.T1396P plasmids (0.5 µg each), the peak I_Na decreased by 37% (Figure 2B, pink triangle) compared with the cells transfected with the WT plasmid (0.5 µg). This result suggested that the p.T1396P variant showed a dominant-negative effect.

Figure 2.

Electrophysiological analysis of SCN5A variants. A, Transfection conditions. In all groups, the total plasmid amount was adjusted to 1.5 µg. The meaning of each shading pattern in the bar graph is shown in the legend inset. B through D, Current-voltage (I-V) relationships for cells expressing various plasmid combinations. The test pulse protocol is shown in the inset. E, Time constant of current decay for various test potentials. F and G, Steady-state inactivation and activation curves. The test pulse protocol for inactivation is shown in the inset. GR indicates G833R; TP, T1396P. *P<0.05 and **P<0.01 by 1-way ANOVA followed by the Tukey post hoc test vs wild type (WT).

Figure 3.

Electrophysiological properties with Difo (difopein). A through C, Time constant of current decay in cells coexpressing Difo. D through F, Steady-state inactivation curve in cells coexpressing Difo. G, Bar graphs representing peak I_Na density at −40 mV with or without coexpressing Difo. GR indicates G833R; n.s., not significant; and TP, T1396P. *P<0.05 and **P<0.01 by 1-way ANOVA followed by the Tukey post hoc test vs wild type (WT).

Table.

Electrophysiological Properties

In contrast, I_Na in the cells transfected with p.G833R (0.5 µg) was comparable to that of WT (Figure 2C, brown inverted triangle). When p.G833R (0.5 µg) was cotransfected with WT (0.5 µg), the peak I_Na was significantly larger than in cells transfected with WT (0.5 µg) alone (Figure 2C, green square) and comparable in the cell transfected with WT (1.0 µg; Figure S1A; Table). Interestingly, when p.G8333R (0.5 µg) was cotransfected with p.T1396P (0.5 µg), the peak I_Na did not decrease (Figure 2D, blue diamond), indicating that p.T1396P did not exhibit DN effect on the p.G833R channels. To further investigate the DN effect of p.T1396P, we also tested a condition in which p.T1396P was expressed at a 2-fold higher level relative to p.G833R. For this purpose, we prepared 2 conditions: cells transfected with 0.5 µg each of SCN5A-G833R and T1396P (1:1), and cells transfected with 0.5-µg G833R and 1.0-µg T1396P (1:2). The total amount of plasmid was adjusted to 2.0 μg with pEGFP-C3 plasmid for this experiment. When comparing the peak I_Na density at –40 mV, we observed no significant difference between these 2 conditions (Figure S1B).

We also assessed the effects of p.T1396P on various characteristics of the WT channel. When p.T1396P plasmids (0.5 µg) were cotransfected with WT plasmids (0.5 µg), the fast component of current decay was prolonged, and the voltage dependence of steady-state inactivation was shifted rightward by 5.6 mV (Figure 2E and 2F; Table), while the voltage dependence of activation was not affected (Figure 2G). Despite the p.T1396P channel being nonfunctional, it affected WT channels, indicating that these 2 channels can function by interacting with each other (coupled gating).

Difopein Prevented the Interactions Between WT and Variant Na_v1.5

Previous studies reported that difopein inhibits 14-3-3 proteins that play a role in the dimerization of Na_v1.5 channels.^6,10 To investigate this further, we coexpressed this peptide with various combinations of WT and variant SCN5As. Coexpression of difopein with WT and p.T1396P channels reversed the prolonged current decay and rightward shift of the steady-state inactivation curve. In contrast, difopein did not affect these parameters in cells transfected with p.G833R (Figure 3A through 3F). In addition, the steady-state inactivation for cells expressing only WT was also shifted leftward by coexpression of difopein (Figure S1C).

Furthermore, the peak I_Na in the cells transfected with WT and p.T1396P plasmids increased by 1.5fold (Figure 3G) with difopein coexpression. In contrast, difopein did not affect I_Na in the cells expressing WT and p.G833R or transfected with p.G833R and p.T1396P. In cases where a dominant-negative effect was not observed, coexpression of difopein did not increase current densities, suggesting that the increase of the peak current density is via cancelation of the DN effect by T1396P.

Comparison of Cell-Surface Expression of WT and Each Variant Na_v1.5

To assess the cell-surface expression of each mutated Nav1.5, we performed a cell-surface biotinylation assay (Figure 4A and 4B). The quantitative analysis of cell-surface expression exhibited no significant difference among WT and 2 variant channels, suggesting that both p.G833R and p.T1396P variants do not affect translation and membrane trafficking of Na_v1.5.

Figure 4.

Biotinylation and coimmunoprecipitation (co-IP) assays. A, Representative immunoblot images of biotinylated surface protein fractions and total protein fractions from cells transfected with wild-type or mutant SCN5A plasmids. B, Beeswarm plots showing individual data points from the biotinylation assay (n=4). Box plots are overlaid to indicate the median, interquartile range (box), and data distribution (whiskers). C, Representative immunoblot images from the co-IP assay. Shown are the co-IP fractions and corresponding total protein fractions from cells transfected with various plasmid combinations. D, Beeswarm plots with overlaid box plots representing individual data points from the co-IP assay (n>7). No Tfx indicates nontransfected cells. FLAG-GR/TP indicates FLAG-tagged G833R/T1396P; FLAG-WT/MT, FLAG-tagged WT/mutant; HA-GR/TP, HA-tagged G833R/T1396P; HA-WT/MT, HA-tagged WT/mutant; and n.s., not significant. *P<0.05 and **P<0.01 by 1-way ANOVA followed by the Tukey post hoc test vs wild type (WT).

WT and Variant Na_v1.5 Assembled Into Oligomers, Irrespective of Their Electrophysiological Characteristics

Next, to evaluate the oligomerization patterns of WT and variant channels, we performed coimmunoprecipitation assays using HA-tagged and FLAG-tagged WT and variant channel plasmids. Figure 4C shows that the FLAG-tagged Na_v1.5 was coprecipitated by the HA-tagged Na_v1.5, regardless of the electrophysiological effects. This result indicates that Na_v1.5 forms oligomers. This oligomerization was unaffected by the presence of difopein or by tag-swapping conditions (Figure 5).

Figure 5.

Coimmunoprecipitation (co-IP) assay with or without difopein. A, Representative immunoblot images from the co-IP assay. The co-IP fractions and corresponding total protein fractions from cells transfected with various plasmid combinations with or without difopein are shown. B, Beeswarm plots with overlaid box plots representing individual data points from the co-IP assay (n>4). Box plots filled with blue indicate cells cotransfected with difopein, while those filled with green indicate cells without difopein. FLAG-GR/TP indicates FLAG-tagged G833R/T1396P; FLAG-WT/MT, FLAG-tagged wild type/mutant; HA-GR, HA-tagged G833R; HA-WT/MT, HA-tagged wild type/mutant; and n.s., not significant.

In Situ Proximity Ligation Assay Exhibited Close Localization of WT and Mutated Na_v1.5

To further assess the assembly of Na_v1.5 in more physiological conditions, we performed an in situ proximity ligation assay (Figure 6) using HA- and FLAG-tagged constructs. Through the microscopic evaluation, we observed proximity ligation assay signals in the cells transfected with all pairwise combinations of SCN5A plasmids (Figure 6A), irrespective of the electrophysiological characteristics, suggesting the proximal expression of each tagged channel in living cells. This result supports the coimmunoprecipitation results, showing that Na_v1.5 channels form oligomers. As observed in the coimmunoprecipitation study, neither the presence of difopein nor the tag-swapping conditions affected the proximity ligation assay signal expression (Figure 7).

Figure 6.

Proximity ligation assay (PLA). A, Microscopic images showing HEK293 (human embryonic kidney) cells expressing various plasmid combinations. 4′,6-Diamidino-2-phenylindole (DAPI) was used for nuclei (blue); Duolink proximity ligation signals are shown in red. B, Beeswarm and overlaid box plots representing PLA signal areas normalized by cell numbers (n>9). FLAG-GR/TP indicates FLAG-tagged G833R/T1396P; HA-GR/TP, HA-tagged G833R/T1396P; HA-WT/MT, HA-tagged wild type/mutant; and n.s., not significant.

Figure 7.

Proximity ligation assay (PLA) with Difo (difopein). A, Microscopic images showing HEK293 (human embryonic kidney) cells expressing various plasmid combinations with or without Difo. 4′,6-Diamidino-2-phenylindole (DAPI) was used for nuclei (blue); Duolink proximity ligation signals are shown in red. B, Beeswarm and overlaid box plots representing PLA signal areas normalized by cell numbers (n>5). Box plots filled with blue indicate cells cotransfected with Difo, while those filled with green indicate cells without Difo. FLAG-GR/TP indicates FLAG-tagged G833R/T1396P; HA-GR/TP, HA-tagged G833R/T1396P; and HA-WT/MT, HA-tagged wild type/mutant.

Discussion

In this study, we identified compound heterozygous SCN5A variants in a patient with severe sick sinus syndrome. Functional analysis demonstrated that the p.T1396P variant exhibited a severe LOF effect, whereas the p.G833R variant retained almost normal function comparable to that of the WT channel. Electrophysiological assays further revealed that the p.T1396P channel exerted a DN effect on the coexpressed WT channel. In addition, it delayed current decay and altered the voltage dependence of steady-state inactivation, suggesting a coupled gating effect. Notably, all these effects were abolished by coexpression of difopein, consistent with previous reports.⁶ Interestingly, the p.G833R variant was unaffected by the DN and coupled gating effects. Coexpression of difopein with the variant did not increase current densities.

However, in contrast to this lack of coupling in the p.G833R channel, the coimmunoprecipitation assay confirmed dimerization in all tested channel combinations. In addition, in situ proximity ligation assays confirmed the close spatial localization of Na_v1.5 channels in all tested combinations. These findings indicate that while Na_v1.5 channels physically interact in close proximity, coupled gating is functionally independent of physical dimerization.

Distinct Roles of Physical Interaction and Functional Coupling in SCN5A

The DN effect and electrophysiological coupling of LOF SCN5A variants have been previously reported in in vitro and in vivo studies.^5–7 The hypothesis of Na_v1.5 dimerization and coupled gating was originally proposed based on electrophysiological findings and coimmunoprecipitation experiments demonstrating physical interaction of 2 Na_v1.5 components.^5,6 However, recent evidence suggests that physical dimerization and electrophysiological coupling are distinct. The coupled gating is mediated by the 14-3-3 peptide and can be disrupted by difopein, whereas the dimerization remains unaffected.¹⁰ Our findings align with this, as we observed physical interactions between Na_v1.5 channels in various combinations, including an electrophysiologically uncoupled variant.

Recently, Sumino et al¹¹ reported that sodium channels form oligomers by moving their voltage sensor domains away from the pore domain and allowing them to interact with those of neighboring channels, thereby coordinating their gating activities, as observed by high-speed atomic force microscopy. This observation implies that sodium channels are interconnected through a broad structural component of the voltage sensor domain, which may not be easily dissociated. However, their coordinated electrical activity might be more finely tuned and stabilized by the 14-3-3 peptide.

In line with this hypothesis, our variant G833R, which resides in the S4-S5 linker connecting the voltage sensor domain and the pore domain, may impair the dynamic movement of the voltage sensor domain and, thereby, interfere with the finely regulated coupled electrical activity mediated through interaction with adjacent channels, while the robust physical association itself remains preserved.

Clinical Background and Unresolved Relationship Between LOF and DN Effects

Clinically, the degree of LOF of Na_v1.5 has been associated with disease severity for BrS.¹² However, among these LOF variants, the correlation between the DN effect and disease severity has yet to be established. Although a previous study reported high penetrance in BrS families carrying DN variants, a clear association with phenotypic severity has not yet been observed.⁸ Furthermore, in our analysis of the association between LOF-type SCN5A mutations and lethal arrhythmic events in patients with BrS,¹³ the event rate did not differ between patients with complete LOF missense variants and those with nonmissense LOF variants, such as truncation or nonsense variants that likely lack DN effects due to haploinsufficiency. Considering that many missense LOF variants have been reported to exhibit DN effects,⁸ these findings suggest that the presence of a DN effect does not necessarily correlate with worse clinical outcomes.¹³

Insights From a Compound Heterozygous Family: Potential Physiological Role of Channel Coupling

In contrast to previous studies that mainly compared disease severity between carriers with and without DN variants, this study focused on a single family carrying 2 distinct variants in a compound heterozygous fashion. Experimental models that attempt to characterize the phenotype of a single variant, both in the presence and absence of DN or coupling effects, are not feasible in practice and remain purely theoretical. However, the family in this study represents a rare example in which such a comparison is naturally possible. Although prior reports have focused on BrS, the phenotype here was sinus node dysfunction (sick sinus syndrome), representing a different but still informative context.

The combination of the p.G833R variant, which does not have coupled gating nor DN effects, and the LOF-type p.T1396P variant did not result in a mild phenotype. In terms of current density, p.G833R would be expected to generate a higher density of I_Na in the proband’s cardiac tissue, unaffected by the DN effects of p.T1396P. However, in reality, the proband harboring both variants exhibited earlier onset and more severe bradycardia compared with the mother, who carried only p.T1396P. Her clinical course was also complicated by VT. Although genetic testing was not available for her elder sister (II-2), who had received a pacemaker, it is plausible that her death was likewise due to ventricular tachyarrhythmia. These severe clinical courses can be explained by 2 possibilities.

One possibility is the shift of inactivation by p.T1396P. The coexpression of WT and p.T1396P demonstrated a rightward shift of the steady-state inactivation curve via the coupled gating effect. This shift can induce an increase in window current, possibly counterbalancing the reduction of the peak I_Na density by the DN effect. In contrast, in the cells expressing p.G833R and p.T1396P, this shift was not observed, suggesting that the proband could not benefit from this compensatory mechanism, resulting in a severe and early onset phenotype.

Another possibility is that coupled gating may occur in normal channels and could play a significant physiological role in vivo. Indeed, in our coexpression assay with difopein, not only the cells expressing WT and p.T1396P but also the cells expressing WT channel alone exhibited a leftward shift in steady-state inactivation, suggesting that WT Nav1.5 functions with coupling and its voltage dependence is naturally rightward-shifted compared with that of the monomer state under normal physiological conditions. In heterozygous carriers of LOF variants and WT allele, the reduced I_Na caused by the LOF variant may be compensated by fully functional WT channels. In contrast, in compound heterozygotes with both an LOF variant and a variant lacking coupled gating function, such as p.G833R, the variant lacking coupled gating function may fail to adequately substitute for the LOF channels, thereby exacerbating the disease phenotype. If coupled gating is not a negative factor that causes DN effects, but rather a necessary mechanism for normal channels, then the severe phenotype observed in the proband in this study becomes more understandable.

At present, the physiological role of coupled gating and channel dimerization remains unclear. We hypothesize that high-density clustering of Nav1.5 channels at the perinexus,¹⁴ near cell-cell junctions, and their coordinated activation may be partly mediated by these mechanisms. The loss-of-coupling variant G833R, therefore, may cause a marked delay in action potential propagation across cardiac tissues. As proposed in BrS, localized conduction delay can trigger malignant ventricular tachyarrhythmias, which may have contributed to the development of VT in the proband and possibly her elder sister II-2. Further investigation under physiological conditions is warranted to elucidate the true impact of the coupled gating effect. In any case, further investigation under physiological conditions is required to elucidate the true impact of the coupled gating effect.

Research Limitations

The electrophysiological and immunological studies were examined only in cultured HEK293 cells; thus, the actual functional characteristics can be different because various channel modulators, such as β-subunit, cytoskeleton, and membrane proteins, exist in human cardiomyocytes. In addition, the genotype of the proband’s sister was not confirmed.

Article Information

Acknowledgments

The authors would like to thank Drs Mikiko Ohno and Akio Shimizu for their constructive advice. During the preparation of the work, the authors used Grammarly software and ChatGPT for English proofreading.

Sources of Funding

This study was supported by the Japan Society for the Promotion of Science, KAKENHI (Grants-in-Aid for Scientific Research) grants 18K15887 and 23K07528, a research grant from the Mochida Memorial Fund, the Suzuken Fund, and the Yamauchi-Susumu Research Fund for Dr Kato.

Disclosures

None.

Supplemental Material

Supplemental Methods

Figure S1

References 15

Supplementary Material

hae-19-e014270-s001.pdf ^{(256.5KB, pdf)}