SCN5A variant that blocks fibroblast growth factor homologous factor regulation causes human arrhythmia

Significance

Cardiovascular disease remains the leading cause of mortality in the United States, and cardiac arrhythmia underlies the majority of these deaths. Here, we report a new mechanism for congenital human cardiac arrhythmia due to defects in the regulation of the primary cardiac Na_v channel, Na_v1.5 (SCN5A), by a family of signaling molecules termed fibroblast growth factor homologous factors (FHFs). Individuals harboring SCN5A variants that affect Na_v1.5/FHF interactions display atrial and ventricular phenotypes, syncope, and sudden cardiac death. The human variant results in aberrant Na_v1.5 inactivation, causing prolonged action potential duration and afterdepolarizations in murine myocytes, thereby providing a rationale for the human arrhythmia.

Abstract

Na_v channels are essential for metazoan membrane depolarization, and Na_v channel dysfunction is directly linked with epilepsy, ataxia, pain, arrhythmia, myotonia, and irritable bowel syndrome. Human Na_v channelopathies are primarily caused by variants that directly affect Na_v channel permeability or gating. However, a new class of human Na_v channelopathies has emerged based on channel variants that alter regulation by intracellular signaling or cytoskeletal proteins. Fibroblast growth factor homologous factors (FHFs) are a family of intracellular signaling proteins linked with Na_v channel regulation in neurons and myocytes. However, to date, there is surprisingly little evidence linking Na_v channel gene variants with FHFs and human disease. Here, we provide, to our knowledge, the first evidence that mutations in SCN5A (encodes primary cardiac Na_v channel Na_v1.5) that alter FHF binding result in human cardiovascular disease. We describe a five*generation kindred with a history of atrial and ventricular arrhythmias, cardiac arrest, and sudden cardiac death. Affected family members harbor a novel SCN5A variant resulting in p.H1849R. p.H1849R is localized in the central binding core on Na_v1.5 for FHFs. Consistent with these data, Na_v1.5 p.H1849R affected interaction with FHFs. Further, electrophysiological analysis identified Na_v1.5 p.H1849R as a gain-of-function for I_Na by altering steady-state inactivation and slowing the rate of Na_v1.5 inactivation. In line with these data and consistent with human cardiac phenotypes, myocytes expressing Na_v1.5 p.H1849R displayed prolonged action potential duration and arrhythmogenic afterdepolarizations. Together, these findings identify a previously unexplored mechanism for human Na_v channelopathy based on altered Na_v1.5 association with FHF proteins.

Encoded by 10 different genes, Na_v channel α-subunits regulate excitable membrane depolarization and are therefore central to metazoan physiology (1). Na_v channel function is critical for neuronal firing and communication (1, 2), cardiac excitation–contraction coupling (3), and skeletal and intestinal function (4, 5). The impact of Na_v channel function for human biology has been elegantly defined by nearly two decades of studies directly linking Na_v channel gene variants with human disease. To date, the field of human Na_v channelopathies has exploded to include wide spectrums of neurological [epilepsy (1), pain (6), ataxia (7)] and cardiovascular diseases (8) as well as myotonia congenital (9) and even irritable bowel syndrome (5). Though the majority of these diseases are based on gene variants in Na_v channel transmembrane segments that affect the channel pore or channel gating (10), a new paradigm for Na_v channelopathies has emerged based on variants that alter association of Na_v channels with essential regulatory proteins. To date, human Na_v channel gene variants linked with neurological and cardiovascular disease have not only provided new insight on the pathophysiology of excitable cell disease, but also identified and/or validated key in vivo Na_v channel regulatory pathways [syntrophin (11), ankyrin-G (12, 13), Na_v β1 (14), calmodulin (15), protein kinase A (16), and CaMKIIδ (17, 18)]. However, in many cases, whereas animal and cellular findings may strongly support the role of regulatory proteins for human Na_v channel function, human variants that may serve to validate the association have remained elusive, likely due to redundancy of signaling pathways or extreme severity of the disease.

Identified in the retina nearly two decades ago, fibroblast growth factor homologous factors (FHFs; FGF11–14) are a family of signaling proteins with key roles in ion channel regulation (19). Distinct from canonical FGFs that are secreted and bind to extracellular FGF receptors, FGF11–14 lack signal sequences and thus regulate intracellular targets. Currently, Na_v channels are the most characterized FHF target (20–22), and recent structural data mapped the FHF binding site to the Na_v channel C terminus (23, 24). Notably, FHFs display multiple roles in Na_v channel regulation, including expression, trafficking, and channel gating. However, each FHF appears to show unique regulatory roles for Na_v channel regulation that are cell type and Na_v channel isoform dependent. Though FHF signaling is complex, the roles of FHFs in vertebrate physiology are clearly illustrated by dysfunction of FHFs in human disease. To date, FHF loci have been linked with spinocerebellar ataxia 27, X-linked mental retardation, and cardiac arrhythmia (25–27). In animals, FHF deficiency results in severe neurological phenotypes associated with altered Na_v channel function (28). Despite the overwhelming biochemical, functional, and in vivo animal data linking Na_v channels and FHF proteins, and in contrast to many other Na_v channel regulatory pathways, there are surprisingly little data linking human Na_v channel variants with FHFs in any disease.

Here we provide, to our knowledge, the first evidence that human Na_v channel gene variants that alter FHF binding result in potentially fatal human disease. We describe a five-generation kindred with a history of atrial and ventricular arrhythmias, cardiac arrest, and sudden cardiac death. Genetic testing revealed a SCN5A variant, resulting in p.H1849R, in affected family members. The identified SCN5A variant p.H1849R is novel and located at a site in the Na_v1.5 C-terminal domain identified to associate with FHFs. Notably, the human p.H1849R variant markedly altered interaction with FHFs, and functional analysis of the variant identified Na_v1.5 p.H1849R as a gain-of-function variant. Further, consistent with LQT3 phenotypes observed in the family, expression of this variant resulted in prolonged action potential duration and arrhythmogenic afterdepolarizations. Together, our findings define a previously unidentified mechanism for human Na_v channelopathies based on loss of Na_v1.5 association with FHF proteins and further confirm the critical link between these intracellular proteins and Na_v channels in excitable cells.

Results

Na_V1.5 Variant p.H1849R Associates with LQT, Atrial Fibrillation, Ventricular Tachycardia, and Sudden Cardiac Arrest.

We identified a previously uncharacterized SCN5A variant associated with atrial fibrillation (AF), long QT, and sudden cardiac arrest. The proband is a 27-y-old male who suffered sudden cardiac arrest at work while moving boxes. Upon resuscitation with CPR and automated external defibrillator (AED) shock, initial ECG recordings presented episodes of AF, evident by the lack of P-waves and a varying R–R interval, accompanied by prevalent premature ventricular contractions (PVCs; Fig. 1A, white arrows). Subsequent interrogations revealed a prolonged QT interval, with a corrected QT (QT_c) up to 496 ms (sinus rhythm), augmented by ST segment changes and episodes of AF. Procainamide challenge of the proband was negative for Brugada (BrS) ECG pattern, and did not demonstrate J-point elevation or QRS prolongation. Serial echocardiograms were unremarkable in terms of ventricular function and atrial and ventricular dimensions. A diagnosis of LQT3 was made after serial rest ECGs disclosed persistent QTc prolongation. Based on phenotypes and witnessed arrest, the proband was implanted with dual chamber implantable cardioverter defibrillator (ICD). The ICD has appropriately fired multiple times in response to sustained ventricular tachycardia/fibrillation (Fig. 1B). Notably, several family members have presented cardiac abnormalities coincident with a family history of sudden cardiac death (SCD; Fig. 1C). Due to the high suspicion for primary arrhythmia disease and documented family history of SCD, the proband underwent targeted genetic testing for identified arrhythmia genes (KCNQ1, KCNH2, SCN5A, ANK2, KCNE1, KCNE2, KCNJ2, CACNA1C, CAV3, SCN4B, AKAP9, and SNTA1). No known disease-associated variants were identified. However, secondary testing identified a heterozygous A > G nucleotide substitution in exon 28 of the SCN5A gene, producing a histidine-to-arginine change in Na_v1.5, the primary cardiac Na_v channel (Fig. 1 D and E; Na_v1.5 p.H1849R). The variant was not previously identified as disease causing, and has not been observed in the general population (minor allele frequency was 0.0 in ExAC and ClinVar). Follow-up ECGs consistently demonstrated periods of prolonged QTc, mild ST segment changes, AF, and PVCs.

Fig. 1.

Human Na_V1.5 H1849R variant associates with LQT and atrial fibrillation. (A) ECG records taken following resuscitation of a 27-y-old male experiencing sudden cardiac arrest. ECGs demonstrate atrial fibrillation and premature ventricular contractions (white arrows). (B) Malignant ventricular arrhythmia in proband requiring ICD discharge (note “S”). (C) Family pedigree of the proband (arrow) denoting members harboring the H1849R variant (+ symbol) and associated phenotypes. LQT, long QT, SCA, sudden cardiac arrest. (D) Chromatograms denoting the nucleotide change (A > G) resulting in a histidine-to-arginine change at amino acid position 1849 of Na_v1.5 in proband. (E) Membrane topology of Na_v1.5 protein denoting location the H1849R variant. (F) Na_v1.5 H1849 is highly conserved across species.

Cascade screening showed that the proband’s mother and sister harbored the Na_v1.5 p.H1849R variant (Fig. 1C). ECGs from the proband’s mother show borderline prolonged QTc (QTc > 460 ms) with slight ST changes following exertion. Results from an implantable loop recorder (ILR) document several episodes of nonsustained AF with atrial and ventricular pacing, and a rate-dependent bundle branch block. Succeeding interrogations of the ILR demonstrated continued periods of AF, summating in a visit to the emergency room for a 2-h period of recorded symptomatic AF. Patient’s sister reported a history of syncope, palpitations, and chest tightness following caffeine. Episodes of supraventricular and ventricular ectopy were recorded on a Holter monitor, and subsequently an ILR was placed for future interrogation. After positive genetic screening in the proband’s sister, further screening for the proband’s niece and nephew was indicated, and consent was given for testing. Genetic screening of the niece (5 y) was positive for the p.H1849R variant, and follow-up was initiated for electrophysiology study testing. A diagnosis of LQT3 was made after serial-rest ECGs disclosed persistent QTc prolongation (QTc 471 ms). Intermittent late-peaking T waves were also appreciated. Other significant family history included the proband’s maternal grandfather who had SCD in his sleep at the age of 28, and a maternal paternal great grandfather who had SCD in his sleep at the age of 52. No autopsies were performed on these individuals, and no DNA is available for evaluation.

Na_v1.5 p.H1849R Alters Association with FHFs.

p.H1849 is located in the cytoplasmic C terminus of Na_v1.5 (Na_v1.5 CT) and conserved across species (Fig. 1 E and F). This specific residue was previously identified as forming part of the binding pocket for FHFs (23). A rendering of the interaction surfaces for FGF13 and the Na_v1.5 CT demonstrates a critical association between FGF13 Arg57 with its interaction pocket anchored by residue H1849 on the Na_v1.5 CT (Fig. 2 A and B). We tested the impact of p.H1849R on Na_v1.5/FHF interactions. Coimmunoprecipitation experiments using detergent-soluble lysates from mouse heart confirmed interaction of Na_v1.5 with FGF13 (Fig. 2 C and D). Further, pull-down experiments from detergent-soluble lysates from mouse heart demonstrated direct association of GST–Na_v1.5 C terminus with FGF13 (Fig. 2E). The p.H1849R variant altered the interaction with FHFs. First, in the context of a GST–Na_v1.5 CT fusion protein, the variant reduced binding for FGF13 from detergent-soluble lysates from mouse heart (Fig. 2E). We observed no binding of GST alone for FGF13. Second, in vitro binding assays demonstrated similar results with all FHFs: radiolabeled FGF12, FGF13, and FGF14 all showed reduced interaction with GST–Na_v1.5 CT harboring the p.H1849R variant or GST alone (Fig. 2F). We quantified the impact of the Na_v1.5 p.H1849R variant on FHF binding using isothermal titration calorimetry. Purified WT Na_v1.5 C terminus and FGF13 associated with high affinity (K_d = 17 nM; Table S1; Fig. 2G). In contrast, purified Na_v1.5 C terminus p.H1849R showed ∼100-fold decrease in affinity for purified FGF13 (K_d = 1.9 µM; Table S1; Fig. 2 H and I). In summary, Na_v1.5 p.H1849R is associated with familial arrhythmias and significantly alters the interaction of Na_v1.5 with FHF family proteins.

Fig. 2.

Human Na_v1.5 p.H1849R variant blocks association with FHF proteins. (A) Schematic of Na_v1.5 denoting FHF binding domain in in red. (B) Interaction surfaces of FGF13 and Na_v1.5, denoting the critical Arg57 on FGF13 (blue) and its interaction pocket anchored by His1849 (yellow). (C) Na_V1.5 Ig coimmunoprecipitated FGF13 from detergent-soluble lysates of murine heart. (D) Conversely, FGF13 Ig coimmunoprecipitated Na_V1.5 from detergent-soluble lysates of murine heart. (E) Immobilized GST-Na_v1.5 C terminus (CT) associated with FGF13 from detergent-soluble lysates of murine hearts. In contrast, p.H1849R CT did not demonstrate significant binding. No interaction was observed between FGF13 and GST alone. (F) ³⁵S-labeled FGF12, 13, and 14 directly associate with GST-Na_V1.5 CT but not GST-Na_v1.5 p.H1849R. (G and H) Isothermal titration calorimetry data demonstrated a 10²-fold decrease in binding affinity for FGF13 to the H1849R C-terminal domain (CTD) relative to WT CTD (K_d = 1.88 ± 0.11 μM and 16.6 ± 0.7 nM, respectively; n = 3, P < 0.05). (I) Qualitative rendering of binding affinity of Na_v1.5 CT H1849R relative to WT Na_v1.5 CT.

Table S1.

Summary of ITC data

Titrant	Cell	Kid, nM	ΔH, Kcal/mol	ΔS, cal⋅mol−1⋅deg−1	N value
FGF13	Nav1.5 amino acids 1773–1940	16.6 ± 0.7 (3)	−2.009 ± 0.2 (3)	28.7 ± 0.58 (3)	0.84 ± 0.08 (3)
FGF13	Nav1.5 amino acids 1773–1940 HR	1877 ± 108 (3)	−5.43 ± 0.55 (3)	7.7 ± 1.87 (3)	0.82 ± 0.07 (3)

Numbers in parentheses denote the number of experiments (n = 3).

Creation of in Vivo Model to Investigate Human SCN5A Variants.

To analyze the consequences of altered interaction between the Na_v1.5 p.H1849R variant and FHFs on select I_Na properties, we created a new mouse model for inducible cardiomyocyte Na_v1.5 silencing. Scn5a knock-in mice were created with loxP sites flanking Scn5a exons 10 and 13 (Fig. S1) to silence Scn5a expression in the presence of Cre recombinase. Following birth, cardiomyocytes isolated from neonatal Scn5a^f/f mice were cultured and analyzed for Na_v1.5-dependent current ± Adv-Cre transduction. Though we observed robust I_Na in mock-transduced myocytes, we detected only negligible I_Na in cultures transduced with Adv-Cre (Fig. S1). Consistent with these data, we observed diminished Na_v1.5 by immunoblot and immunostaining of Adv-Cre–infected myocytes (Fig. S1). Despite I_Na loss, myocytes remained healthy and viable. Notably, introduction of exogenous WT Na_v1.5 in these cultures resulted in rescue of I_Na to approximately physiological levels (Scn5a^f/f myocyte I_Na,peak: 115.2 ± 8.4 pA/pF; Scn5a^f/f + Adv-Cre + Na_v1.5 I_Na,_peak: 142.2 ± 18.2 pA/pF). Thus, this new system is sufficient to analyze I_Na properties of putative Na_v1.5 variants.

Fig. S1.

Creation and validation of Scn5a^f/f model. (A) Schematic of the Scn5a mouse allele. Exons are represented by light gray boxes, exon numbers are indicated. Red triangles denote loxP sites flanking exons 10–13 in targeting construct. Addition of Cre recombinase results in excision of exons 10–13 and loss of Na_v1.5 protein expression in myocytes. (B and C) I_Na recorded from neonatal myocytes from Scn5a^f/f animals cultured 72 h ± adenovirus-expressing Cre recombinase (Adv-Cre). (D) I_Na current/voltage relationship for myocyte cultures ± Adv-Cre. (E) Immunoblot analysis of isolated from Scn5a^f/f mice incubated with Adv-Cre show reduced Na_v1.5 compared with mock-treated myocytes. (F) Immunostaining of Na_v1.5 in untreated and treated myocytes (α-actinin used as positive marker of myocytes).

Na_v1.5 p.H1849R Alters Na_v1.5 Steady-State Inactivation in Primary Myocytes.

FHF proteins regulate Na_v channel steady-state inactivation (21, 29), and knockdown of FHFs in cardiomyocytes results in a hyperpolarizing shift in steady-state interaction (24, 27). We therefore used the Scn5a^f/f myocyte system to test the human Na_v1.5 H1849R variant for alterations in I_Na inactivation properties. As noted above, I_Na levels in Adv-Cre–infected Scn5a^f/f myocytes (GFP-positive) are rescued by expression of exogenous Na_v1.5 (Fig. 3A). Of note, expression of Na_v1.5 p.H1849R at equivalent levels resulted in an increased I_Na density relative to WT Na_v1.5 (Fig. 3A; increased ∼34%; P < 0.05). Further, we observed a significant ∼6.7-mV depolarizing shift in the V_1/2 of voltage-dependent inactivation of the p.H1849R variant compared with the WT channel (−101.2 vs. −107.9 mV; P < 0.05; Fig. 3B). Because FHF knockdown in cardiomyocytes leads to a hyperpolarizing shift in steady-state inactivation, these results suggested that the p.H1849R variant maintains an altered interaction with FHFs despite the reduced affinity noted in binding experiments with the isolated Na_v1.5 CT and FHFs (Fig. 2). We observed no difference in voltage dependence of activation of WT Na_v1.5 compared with Na_v1.5 p.H1849R (Fig. 3B), consistent with the lack of effect of FHFs on Na_v1.5 interaction in cardiomyocytes (24). In summary, consistent with past work linking FHFs with regulation of Na_v channel steady-state inactivation (21, 24, 27, 29), Na_v1.5 p.H1849R displays aberrant steady-state inactivation in primary myocytes.

Fig. 3.

Nav1.5 p.H1849R results in gain of function. (A) I_Na elicited from Na_v1.5 p.H1849R-expressing myocytes (blue, n = 8) displayed significantly larger I_Na density relative to cells expressing WT Na_V1.5 (black, n = 7; P < 0.05). I–V relationship for cells silenced for endogenous Na_V1.5, but untransfected, shown in red (n = 5). (B) Boltzmann fits of the voltage-dependent inactivation (dashed lines) of the p.H1849R channels (blue circles) demonstrated a depolarizing shift in the V_1/2 relative to WT (black circles; −101.2 vs. −107.9 mV; P < 0.05). Boltzmann fits of the voltage-dependent activation profile (solid blue and black lines) showed similar V_1/2 values (−61.9 vs. −62.4 mV for WT and p.H1849R, respectively; P = N.S.). R² values for all fits >0.99. (C) I_Na elicited from Na_v1.5 p.H1849R and FGF12 expressed in HEK293 cells (blue, n = 14) displayed significantly larger I_Na density relative to currents from cells expressing WT Na_v1.5 (black, n = 11; P < 0.01 for test voltages as indicated). (D) Boltzmann fits of the steady-state voltage-dependent inactivation of the p.H1849R channels expressed with FGF12 (blue triangles, n = 13) demonstrated a depolarizing shift in the V_1/2 relative to WT expressed with FGF12 (black circles, n = 12; −67.7 vs. −76.2 mV; *P < 0.01). (E) Boltzmann fits of the voltage-dependent activation profile show similar V_1/2 values [−56.0 vs. −55.7 mV for WT (n = 11) and p.H1849R (n = 14), respectively; P = N.S.]. Vm in D and mV in E. (F) Scaled exemplar traces at a test potential of −40 mV and one-exponential fits of the decay phases for the p.H1849R channels expressed with FGF12 (blue solid line and dashed line, respectively) and for the WT Na_v1.5 channels expressed with (black solid line and dashed line, respectively). (G) Decay time, τ, of the fits as in H (n = 28 for p.H1849R and n = 21 for WT; *P < 0.01).

Na_v1.5 p.H1849R Displays Arrhythmogenic I_Na Properties.

Patients harboring SCN5A variants with LQT3 display signature QT interval prolongation associated with altered Na_v1.5 regulation. To date, two molecular mechanisms underlying aberrant I_Na regulation in LQT3 are described. LQT3 is primarily caused by SCN5A variants that alter the fast inactivation gate of I_Na, resulting in channels that may reopen at later stages of the action potential (30). This depolarizing “late” I_Na current thus extends ventricular depolarization, often resulting in arrhythmogenic afterdepolarizations. Alternatively, human LQT3 may also be linked with SCN5A variants that slow Na_v1.5 channel inactivation (individual channels with longer open state duration) (31, 32), again extending ventricular depolarization and creating a substrate for arrhythmogenic afterdepolarizations.

To identify potential pathogenic mechanisms of the human Na_v1.5 H1849R variant in the complete absence of any potentially competing ionic currents, we analyzed I_Na signatures of WT Na_v1.5 and Na_v1.5 p.H1849R coexpressed with FGF12 in HEK293 cells (SI Materials and Methods). FGF12 was chosen as the FHF for this assay as it is the most abundant FHF in human heart (27). Consistent with findings in myocytes, we observed increased I_Na,peak for Na_v1.5 p.H1849R compared with WT Na_v1.5 (Fig. 3C; P < 0.05). Further, we observed a depolarizing shift in steady-state inactivation for Na_v1.5 p.H1849R compared with WT Na_v1.5 (Fig. 3D; ∼9-mV shift for Na_v1.5 p.H1849R relative to WT; −67.7 ± 1.9 vs. −76.8 ± 1.9 mV, respectively; P < 0.05). We did not observe differences in activation between Na_v1.5 and Na_v1.5 p.H1849R [Fig. 3E; WT: −57.1 ± 2.5 mV, p.1849R: −56.0 ± 1.5 mV; P = not significant (N.S.)]. Most notably, we observed a striking slowing of inactivation of Na_v1.5 p.H1849R compared with WT Na_v1.5 (Fig. 3 F and G; ∼fivefold increase in decay time for Na_v1.5 p.H1849R compared with WT Nav1.5; P < 0.05). In summary, the Na_v1.5 p.H1849R variant has a number of gain-of-function phenotypes contributing to its arrhythmogenic properties, most notably a slowed rate of inactivation.

Na_v1.5 H1849R Alters Myocyte Membrane Excitability.

Action potentials (APs) were measured in myocytes expressing WT Na_v1.5 or Na_v1.5 p.H1849R to define the relationship between the human variant and myocyte membrane excitability. In line with alterations in I_Na observed in both myocytes and heterologous cells, we observed a significant increase in AP duration (APD) at 90% repolarization (APD₉₀) for myocytes expressing Na_v1.5 p.H1849R compared with WT Na_v1.5-expressing myocytes (Fig. 4 A and B; P < 0.05). We observed no difference in resting membrane potential or peak transmembrane potential between the two groups (P = N.S.). Notably, consistent with the significant differences in APD₉₀, we observed spontaneous depolarizations (Fig. 4 C and D) in Na_v1.5 p.H1849R myocytes but not in myocytes expressing WT Na_v1.5 (WT Na_v1.5: 0% myocytes; Na_v1.5 p.H1849R: ∼43% myocytes; P < 0.05). In summary, consistent with altered I_Na and the human LQT3 phenotype, Na_v1.5 p.H1849R results in APD prolongation and arrhythmogenic spontaneous membrane afterdepolarizations.

Fig. 4.

Na_v1.5 p.H1849R prolongs AP duration and elicits spontaneous activity in murine myocytes. (A and B) Compared with myocytes expressing WT Na_v1.5, Na_v1.5 p.1849R-expressing myocytes display increased APD [APD₉₀; WT Na_v1.5: 33.1 ± 6.0 ms (n = 7 myocytes); Na_v1.5 p.H1849R: 70.4 ± 12.7 ms (n = 7 myocytes)]. (C and D) Unlike myocytes expressing WT Na_v1.5 (n = 7), Na_v1.5 H1849R-expressing myocytes displayed spontaneous activity during pacing (myocytes paced at 0.5 Hz in C) or spontaneous depolarizations following pacing [spontaneous activity observed following pause following 10 stimulations (red points) at 1.0 Hz in D].

SI Materials and Methods

Molecular Biology.

FGF14-1a (NM_004115), FGF14-1b (NM_175929), FGF12 (NM_021032), and FGF13 (NM_004114) were cloned from human heart cDNA library (Clontech) using unique primers and subcloned into pcDNA3.1+. Na_v1.5 pDsRed pIRES and Na_v1.5 pcDNA3.1+ were generous gifts from Jose Jalife, University of Michigan, Ann Arbor, MI. Na_v1.5 CT (amino acids 1771–2016) was cloned using unique primers and standard techniques and subcloned into pGEX-6P1 (Clontech). All sequences were verified before experimentation. The H1849R mutation was introduced into Na_v1.5 pDsRed pIRES, Na_v1.5 pcDNA3.1+, and Na_v1.5 CT pGEX-6P1 using the Agilent QuikChange II XL SDM kit and unique primers. All sequences were verified before experimentation.

Coimmunoprecipitation Experiments.

A total of 100 μg of detergent-soluble heart lysates were incubated with protein A beads overnight at 4 °C with gentle rotation. Beads were washed with PBS plus 150 mM NaCl, eluted, and electrophoresed. Immunoblotting using anti-Na_V1.5 Ig, anti-FGF12 Ig, anti-FGF13 Ig, and anti-FGF14 Ig was performed using standard molecular protocols. Blots were imaged using the BioRad gel documentation system.

In Vitro Binding Assays.

cDNAs for Na_v1.5-CT and Na_v1.5-CT H1849R were PCR generated, cloned into pGEX-6P1 (Amersham Biosciences), and the sequences confirmed before experimentation. BL21(DE3)pLysS cells (Agilent) that were transformed with the pGEX-6P1 constructs were grown overnight at 37 °C in LB supplemented with ampicillin (50 μg/mL). The overnight cultures were subcultured for large-scale expression. Cells were grown to an optical density of 0.6–0.8 and induced with 1 mM isopropyl, 1 thio-α-d-galatcopyranoside (IPTG) for 4 h at 37 °C. Cells were centrifuged for 5 min at 6,000 × g, resuspended in lysis buffer [1 mM DTT, 1 mM EDTA, 40 μg/mL 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride, 10 μg/mL leupeptin, 40 μg/mL benzamidine, 10 μg/mL pepstatin] and frozen at −20 °C. Cells were lysed by thawing and homogenized by sonication. Cell debris was removed by centrifugation at 4,000 × g for 30 min at 4 °C. Supernatants were added to 1 mL equilibrated glutathione–sepharose 4B (Amersham Biosciences) and incubated overnight at 4 °C. The glutathione–sepharose solutions were washed with PBS containing 1% Triton X-100 (three times), PBS containing 500 mM NaCl (three times), and stored in PBS containing 1 mM NaN₃. Protein purification and molecular weights were verified by SDS/PAGE followed by Coomassie staining. For in vitro translation/binding assays, FGF14-1a, FGF14-1b, FGF12, and FGF13 constructs were in vitro translated using rabbit reticulocyte lysate, [³⁵S]-methionine (20 μCi Redivue l-[³⁵S]-methionine; GE Healthcare), T7 polymerase, and 1 μg plasmid DNA (TNT Coupled Rabbit Reticulocyte Lysate System; Promega). For binding experiments, in vitro translated products were incubated with immobilized GST, immobilized GST-Na_V1.5-CT, or immobilized GST-Na_V1.5-CT H1849R overnight at 4 °C in binding buffer (PBS + 750 mM NaCl, 0.1% Triton X-100). Reactions were washed three times in binding buffer, eluted, and separated by SDS/PAGE. Gels were stained with Coomassie to show presence of GST proteins and to verify comparable loading before drying the gel in a gel dryer (Bio-Rad Laboratories). Radiolabeled proteins were detected by standard autoradiography.

Pull-Downs.

For pull-down experiments, 100 μg whole-heart lysate were incubated with immobilized GST, immobilized GST-Na_V1.5-CT, or immobilized GST-Na_V1.5-CT H1849R in binding buffer (750 mM NaCl) overnight at 4 °C. Beads were washed with binding buffer, eluted, and proteins separated by electrophoresis. Proteins were transferred to nitrocellulose and incubated in primary antibody (anti-FGF12 Ig, anti-FGF13 Ig, and anti-FGF14 Ig) overnight at 4 °C. Blots were washed three times and incubated in secondary antibody for 2 h at 4 °C. Blots were developed using Clarity ECL kit (BioRad) and imaged using the BioRad gel documentation system.

Isothermal Titration Calorimetry.

Human Na_V1.5 C terminus (amino acids 1773–1940), with or without the H1849R mutation, and human FGF13U were cloned (separately) into pET28 (Novagen) and transformed into BL21 Escherichia coli. Protein expression was induced by 1 mM IPTG, and the protein was purified by metal affinity chromatography, followed by size-exclusion chromatography on a Superdex 75 10/300 (GE Healthcare Life Sciences). Protein concentration was determined by A₂₈₀ and the calculated extinction coefficient. ITC experiments were performed with the VP-ITC (MicroCal) at 20 °C. The cell solution containing the Na_V1.5 C terminus (wild type or H1849R mutant) was titrated with injections of solution containing FGF13U. ITC experiments were repeated with different preparations and slightly varied concentrations at least three times to confirm thermodynamic parameters and stoichiometry values. The binding isotherms were analyzed with a single site binding model using the MicroCal Origin version 7.0 software package (OriginLab Corporation), yielding binding enthalpy (ΔH), stoichiometry (n), entropy (ΔS), and association constant (K_a). Results are presented as mean ± SE.

Animals.

Mice used in these studies were neonatal (postnatal day 1 or 2) C57BL/6 mice and transgenic animals carrying a modified Scn5a gene. These modified animals contain loxP sites in the Scn5a gene, flanking exons 11 and 13.

Cardiomyocyte Experiments.

To generate the primary cardiomyocyte cultures, hearts were dissected from P1 or P2 mice and placed in 2 mL of Ham’s F-10. Twenty-four hours later, WT and Scn5a^f/f cardiomyocytes were infected with Cre recombinant adenovirus (expresses Cre recombinase, a type I topoisomerase from P1 bacteriophage that catalyzes site-specific recombination of DNA between loxP sites). At 24 h posttransduction, Scn5a^f/f cardiomyocytes were transfected either WT or H1849R Na_V1.5 using Lipofectamine 2000 (Invitrogen).

Immunoblots and Antibodies.

Wild-type murine mouse hearts were flash-frozen on liquid nitrogen and homogenized in homogenization buffer [0.025 M Tris, 0.15 M NaCl, 0.001 M EDTA, 1% Nonidet P-40, 5% (vol/vol) glycerol, protease inhibitor mixture (Sigma); pH 7.4] using a handheld homogenizer. Lysates were centrifuged at 100,000 × g for 30 min at 4 °C to remove debris and the protein quantitated using BCA analysis (ThermoScientific). Heart lysates, following quantitation by BCA assay (Pierce), were electrophoresed using 4–15% BioRad gels and transferred to nitrocellulose membranes. Membranes were blocked for >1 h at room temperature in 5% (wt/vol) nonfat dry milk and incubated in primary antibody overnight at 4 °C. Primary antibodies included anti-Na_V1.5 Ig (Covance), anti-FGF12 Ig (Abcam), anti-FGF14 (Millipore), anti-GAPDH Ig (Fitzgerald Industries), and anti-FGF13 Ig (Antibodies Inc.). Secondary antibodies used were donkey anti-mouse HRP and donkey anti-rabbit HRP (Jackson Laboratories). Densitometric analysis was performed using Photoshop software, and all data were normalized to GAPDH.

Immunostaining and Confocal Microscopy.

Neonatal cardiomyocytes were washed with PBS (pH 7.4) and fixed in warm 2% paraformaldehyde (37 °C). Cells were blocked/permeabilized in PBS containing 0.075% Triton X-100 and 2 mg/mL BSA and incubated in primary antibody [anti-Na_V1.5 and anti–α-actinin (Sigma)] overnight at 4 °C. Following washes (PBS plus 0.1% Triton X-100), cells were incubated in secondary antibodies (Alexa Fluor conjugated; Invitrogen) for 2 h at 4 °C and mounted using VectaShield (Vector Labs). Cells were imaged on a LSM 780 confocal microscope (Carl Zeiss). Myocytes were imaged using identical confocal settings between genotypes. At least five myocytes were examined for each staining protocol.

HEK293 Cell Electrophysiology.

HEK293T cells were transfected with FGF12 and either Na_v1.5 p.H1849R or WT Na_v1.5 at 60% confluence using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. The total amount of DNA for all transfections was kept constant. Transfected cells were identified by GFP fluorescence. Na⁺ currents were recorded using the whole-cell patch-clamp technique at room temperature (20–22 °C) 48–72 h after transfection using an EPC 10 USB patch amplifier (HEKA Elektronik). The signal was filtered at 2.9 Hz and digitized at 20 Hz. Electrode resistance ranged from 3 to 4 MΩ. Capacitance and series resistance were adjusted (70–85%) to obtain minimal contribution of the capacitive transients. The bath solution contained (in mM): 124 NaCl, 2 CaCl₂, MgCl₂ 1, KCL 1, 20 tetraethylammonium chloride (TEA-Cl), 5 Hepes, 10 glucose. NaOH was added to achieve pH 7.3 (300–310 mOsm). The intracellular solution contained (in mM): 125 CsF, 10 NaCl, 10 Hepes, 15 TEA-Cl, 1.1 EGTA, and 0.5 Na-GTP, pH adjusted to 7.3 with CsOH (290–300 mOsm). Standard two-pulse protocols were used to generate the steady-state inactivation curves: from a holding potential of −120 mV, cells were stepped to 500-ms preconditioning potentials varying between −130 and −10 mV (prepulse), followed by a 20-ms test pulse to −20 mV. Currents (I) were normalized to I_max and fit to a Boltzmann function of the form I/I_max = 1/{1 + exp[(V_m − V_1/2)/k]} in which V_1/2 is the voltage at which half of Na_V1.5 channels are inactivated, k is the slope factor and V_m is the membrane potential. Data analysis was performed using Clampfit 10.2 software (Axon Instruments) and Origin 8. Results are presented as means ± SE; the statistical significance of differences between groups was assessed using a two-tailed Student’s t test and was set at P < 0.05.

Myocyte Electrophysiology.

I_Na currents were recorded from freshly isolated ventricular myocytes using whole-cell patch-clamp configuration using an Axopatch 200B amplifier and Digidata 1440A digitizer. Data acquisition and analysis was performed using pCLAMP software (version 10.3; Molecular Devices). Sodium currents (I_Na) were recorded at room temperature (20–22 °C) with pipette resistances <2.5 MΩ when filled with pipette filling solution containing (in mM): 5 NaCl, 135 CsF, 10 EGTA, 5 MgATP, 5 Hepes, pH 7.2. The extracellular bathing solution contained (in mM): 10 NaCl, 1 MgCl₂, 1.0 CaCl₂, 0.1 CdCl₂, 1 glucose, 127.5 CsCl, and 20 Hepes; pH was maintained at 7.4 with CsOH at room temperature. Appropriate whole-cell capacitance and series resistance compensation (≥60%) was applied along with leak subtraction. To assess the I_Na density, cells were held at −140 mV and stepped to various test potentials from −100 to +10 mV in 5-mV increments, with 200-ms duration pulses and 2,800-ms interpulse intervals. Voltage dependence of inactivation was assessed by holding the cells for 200 ms at −140 to −40 mV followed by a 50-ms test pulse to −40 mV to elicit I_Na. Interpulse interval was 2,700 ms. Inactivation profiles from all experiments were fit with a Boltzmann function to determine V_1/2. Statistical significance for I_Na activation/inactivation was assessed by a Student’s t test.

Action Potentials.

Borosilicate glass electrodes were pulled using a Brown–Flaming puller, yielding a tip resistance of 3–5 mΩ when filled with pipette solution containing (in mM): 120 K-aspartic acid, 8 KCL, 7 NaCl, 1 mM MgCl₂, 5 Mg-ATP, and 10 Hepes. Cells were superfused with an external solution containing (in mM): 5.4 KCl, 135 NaCl, 1 MgCl₂, 10 Hepes, 0.33 NaH₂PO4 × 2H₂O, 10 glucose, and 1 CaCl₂. Action potentials were recorded using an Axon 200B PatchClamp amplifier at room temperature. Briefly, cells were stimulated by with 2 nA/2 ms current pulses. Pulse trains of APs were elicited at 1 or 0.5 Hz. AP duration values were determined as the repolarization percent from the peak to baseline using Clampfit 10.4 software.

Statistics.

Data are presented as mean ± SEM. SigmaPlot 12.0 was used for statistical analysis. The Wilcoxon–Mann–Whitney U test was used to determine P values for single comparisons. One-way ANOVA was used for multiple comparisons with the Bonferroni test for post hoc testing. If the data distribution failed normality tests with the Shapiro–Wilk test, a Kruskal–Wallis one-way ANOVA on ranks was applied with a Dunn multiple-comparisons test for significant P values. Contingency data were analyzed using χ² test. The null hypothesis was rejected for P < 0.05.

Discussion

Cardiac ion channel dysfunction is tightly linked with congenital human arrhythmias. Though mutations in ion channel pore and gating regions represent the vast majority of pathogenic variants, a second class of arrhythmia variants alter the regulation of ion channels by cytosolic signaling or scaffolding proteins. In fact, these findings have revealed new paradigms for rare forms of human arrhythmia based on defects in ion channel-associated proteins. By integrating human clinical and genetic data with biochemical, cell biological, and electrophysiology findings, we identify a new mechanism for human cardiac Na_v channelopathy based on loss of binding to the FHF family of ion channel regulatory proteins. Individuals harboring the Na_v1.5 p.H1849R variant display LQT3 and AF. Functionally, Na_v1.5 p.H1849R displays an altered interaction with FHFs, resulting in increased I_Na peak, increased Na_v1.5 channel availability and, most critically, slowed inactivation. Together, these parameters provide a rationale for the human cardiac phenotypes. In summary, these new findings directly validate the FHF family of signaling proteins as critical for human cardiac excitability. Further, these data illustrate the likely multifunctional roles of FHF proteins for the regulation of membrane excitability across diverse cell types.

The FHFs have gained significant attention due to their role in regulating Na_v and Ca_v channels (24, 29, 33–36), as well as their links with human excitable cell disease. In neurons, FHF dysfunction has been linked with aberrant I_Na phenotypes and human FGF14 variants cause spinocerebellar ataxia 27 (25). Further, the FGF13 locus has been linked with nonspecific forms of X-linked mental retardation (26). Finally, relevant for this study, a missense variant in FGF12 linked with aberrant Na_v1.5 function was recently associated with Brugada syndrome (27). However, to date, no human Na_v channel variants that alter association with FHF proteins have been linked with human disease. Of note, though FGF12–14 all directly associate with Na_v channels, each display unique Na_v channel regulatory properties that may differentially tune Na_v channel expression, trafficking, current density, availability, and/or persistent current (I_Na,L), depending on the cell type and Na_v channel isoform (29, 33, 34). Given these data, the differences in clinical phenotypes are not surprising, but raise important caveats regarding data interpretation.

The molecular mechanisms underlying human LQT3 may be multifactorial but are generally classified into two categories. The primary mechanism underlying LQT3 is altered Na_v1.5 fast inactivation (30). As noted above, in this mechanism, mutant Na_v1.5 channels display an increased probability for reopening, resulting in inward depolarizing I_Na during the plateau phase of action potential (I_Na,late), lengthening of repolarization, and susceptibility to arrhythmogenic afterdepolarizations. However, though less common, slowed inactivation, resulting in Na_v1.5 channels with increased time in the open state, may also underlie LQT3 (31, 32). In our study, Na_v1.5 p.H1849R resulted in increased peak I_Na and increased I_Na availability (37, 38), consistent with known roles of FHFs in channel gating and/or internalization (21, 29). Most notably, the human Na_v1.5 p.H1849R variant displayed a striking signature of slow inactivation compared with WT Na_v1.5. Consistent with these data, compared with WT Na_v1.5, myocytes harboring the Na_v1.5 p.H1849R displayed increased APD and spontaneous afterdepolarizations even in the absence of adrenergic stimulation. Thus, our findings illustrate that Na_v1.5 p.H1849R is an I_Na gain-of-function mutation through the slowing of the rate of inactivation and increasing Na_v1.5 availability during the AP plateau phase, thus prolonging APD and producing proarrhythmic afterdepolarizations. These findings are consistent with prior reports linking I_Na slow inactivation and increased availability with both LQT3 (31, 32) and AF (38) as well as LQT associated with α1 syntrophin gene variants (37).

In summary, our data provide, to our knowledge, the first evidence for human disease based on Na_v channel gene variants that block FHF regulation. Our combined data support altered FHF binding as the mechanism for arrhythmia for Na_v1.5 p.H1849R. Further, recent unbiased structural data identify Na_v1.5 H1849 as a central residue for Na_v1.5/FHF association. Although the endogenous concentrations of FHFs and Na_v channel α-subunits in cardiomyocytes are not known, it is reasonable to assume that an ∼100-fold decrease in affinity between the core domain of the FHF and the mutant Na_v1.5 C-terminal domain (CTD), the key interaction sites identified on each molecule (23), would affect FHF-dependent channel function. This hypothesis is consistent with the previous observation that an affinity-reducing mutation within the FHF core, at the Na_v channel interaction site, similarly confers altered Na_v channel function in neurons (36). However, we note that this variant may also alter Na_v1.5 folding and biophysical function independent of, or in conjunction with, FHF-mediated effects. Again, it will be critical in future in vivo experiments to dissect these specific points. In closing, our findings underscore the complexity and heterogeneity in the presentation of congenital arrhythmia, particularly as it relates to Na_v1.5 and its growing list of regulatory proteins.

Materials and Methods

Approval for use of human subjects was obtained from the Institutional Review Board of Ohio State University, and subjects provided informed consent. Genomic DNA was extracted from peripheral blood lymphocytes of the proband, proband’s sister, mother, niece, and grandmother. A full LQTS genetic panel screened for mutations in 12 known disease-causing genes: KCNQ1, KCNH2, SCN5A, ANK2, KCNE1, KCNE2, KCNJ2, CACNA1C, CAV3, SCN4B, AKAP9, and SNTA1. Sequencing identified a heterozygous A > G nucleotide transition in exon 28 in the SCN5A gene, producing a histidine-to-arginine amino acid change at the 1849 residue in the Na_V1.5 protein.

For animal experiments, animals were handled according to approved protocols and animal welfare regulations of the Institutional Animal Care and Use Committee of The Ohio State University. Detailed descriptions of all materials and methods are provided in SI Materials and Methods.