Physiological and Pathophysiological Insights of Nav1.4 and Nav1.5 Comparison

Mutations in Nav1.4 and Nav1.5 α-subunits have been associated with muscular and cardiac channelopathies, respectively. Despite intense research on the structure and function of these channels, a lot of information is still missing to delineate the various physiological and pathophysiological processes underlying their activity at the molecular level. Nav1.4 and Nav1.5 sequences are similar, suggesting structural and functional homologies between the two orthologous channels. This also suggests that any characteristics described for one channel subunit may shed light on the properties of the counterpart channel subunit. In this review article, after a brief clinical description of the muscular and cardiac channelopathies related to Nav1.4 and Nav1.5 mutations, respectively, we compare the knowledge accumulated in different aspects of the expression and function of Nav1.4 and Nav1.5 α-subunits: the regulation of the two encoding genes (SCN4A and SCN5A), the associated/regulatory proteins and at last, the functional effect of the same missense mutations detected in Nav1.4 and Nav1.5. First, it appears that more is known on Nav1.5 expression and accessory proteins. Because of the high homologies of Nav1.5 binding sites and equivalent Nav1.4 sites, Nav1.5-related results may guide future investigations on Nav1.4. Second, the analysis of the same missense mutations in Nav1.4 and Nav1.5 revealed intriguing similarities regarding their effects on membrane excitability and alteration in channel biophysics. We believe that such comparison may bring new cues to the physiopathology of cardiac and muscular diseases.

Voltage-gated sodium channels (Nav) constitute a family of 10 members in mammals, Nav1.1 to Nav1.9 and Nax, expressed in a large variety of tissues. In excitable cells such as striated myocytes, they initiate action potentials that, in heart as well as in skeletal muscles, trigger, and regulate the contraction. Because of their key role in this function, mutations impacting their activity have tremendous consequences. This review compares the knowledge accumulated in different aspects of the expression and function of Nav1.4 and Nav1.5 α-subunits, and focuses on "homologous" mutations i.e., in the same (aligned) amino acids of the skeletal muscle Nav1.4 and of the cardiac Nav1.5 leading to a large range of muscular and cardiac disorders also called channelopathies.
CLINICAL DESCRIPTION OF THE MAIN Nav1.4 AND Nav1.5 RELATED PATHOLOGIES Clinical Description of Nav1.4 Related Channelopathies Nav1.4, which is encoded by the SCN4A gene, is the poreforming subunit of the main sodium channel present in skeletal muscles. Nav1.4 related channelopathies that affect skeletal muscle excitability (Vicart et al., 2005;Jurkat-Rott et al., 2010;Nicole and Fontaine, 2015) are dominant diseases classified in two opposite groups as defined by the prevalent clinical symptoms: muscle stiffness and hypertonia (myotonia) episodes [non dystrophic myotonias (NDM)], and muscle weakness resulting in paralysis episodes (periodic paralyses; PP). It should be noted that similar clinical pattern are also associated with other channelopathies involving chloride channels (NDM) or calcium channels (PP). Table 1 summarizes the main classes of  (Trip et al., 2009;Raja Rayan and Hanna, 2010 McClatchey et al., 1992b;Ptácek et al., 1992;Hayward et al., 1996;Featherstone et al., 1998;Bouhours et al., 2004 SCM Stiffness at exertion (most often), permanently at rest (myotonia permanens), or acetazolamideresponsive myotonia Exertion Acetazolamide

Nav1.4-Related non Dystrophic Myotonias
Myotonia may occur at the beginning of effort and be alleviated (myotonia, with warm-up effect), or aggravated (paradoxical myotonia, also named paramyotonia) by continuing effort. Those myotonic or paramyotonic symptoms are associated with myotonic discharges analyzed with electromyographic investigations. NDM are opposed to dystrophic myotonias as observed in Steinert (Myotonic Dystrophy type 1, DM1) and PROMM (PROximal Myotonic Myopathy or Myotonic Dystrophy type 2, DM2) diseases. Among NDM, at least two entities differ clinically and electromyographically (Trip et al., 2009;Raja Rayan and Hanna, 2010).
-Paramyotonia congenita (PC) consists of cold-induced stiffness often associated with some weakness of face and extremities muscles, and paradoxical myotonia; it is associated with a progressive decrease of compound muscle action potential (CMAP) amplitude during repetitive short efforts test at EMG (pattern I according to Fournier, Fournier et al., 2004). -Sodium channel myotonias (SCM) regroup the remaining dominant sodium channel-related myotonias that are not significantly cold-sensitive or paradoxical, and do not exhibit any change of CMAP amplitude during repetitive short efforts test at EMG (pattern III according to Fournier); this SCM entity was initially termed "potassium-aggravated myotonia" as potassium load triggers myotonia in some cases. This group was further subdivided into three types: myotonia permanens, myotonia fluctuans, and acetazolamide-responsive myotonia. While this classification is not used in clinics, it has some relevance: myotonia permanens designates myotonia that is present permanently, even at rest; myotonia fluctuans designates myotonia that appears and disappears at some moment, with no systematic concomitance with exertion, a peculiar circumstance being exercise-induced delayed-onset myotonia, that occurs some time after exertion has stopped; acetazolamide-responsive myotonia is a treatment-related designation, that underlines the fact that some SCM are treatable by acetazolamide.

Nav1.4-Related Periodic Paralysis
Among PP, two distinct entities are recognized (Raja Rayan and Hanna, 2010): hypokalemic periodic paralysis (HypoPP) is characterized by a marked hypokalemia concomitant with paralysis episodes, and, on the opposite, hyperkalemic periodic paralysis (HyperPP) is associated with a tendency to high blood potassium levels during the paralysis episodes. From the electromyographic point of view, both are characterized by a marked decrease of CMAP amplitude after a 5 min-long effort (long effort test, also referred to as McManis test).
Overlap, borderline or mixed syndromes between PP and NDM or between their subtypes have been reported (McClatchey et al., 1992a;Sugiura et al., 2003;Webb and Cannon, 2008;Yoshinaga et al., 2012). The age at onset is usually in early to late childhood. Neonatal symptoms are not classically reported in the most frequent Nav1.4 channelopathies, but dominant de novo mutations are reported in moderate to severe neonatal clinical presentations such as severe neonatal episodic laryngospasm (SNEL) (Lion-Francois et al., 2010). In a general way, respiratory symptoms are not common in PP and NDM, however a small number of patients are exposed to laryngeal or diaphragmatic weakness or myotonia that may be symptomatic.
The minimal prevalence of skeletal muscle Nav1.4 channelopathies has been recently estimated to be 0.4:100,000 in England (Horga et al., 2013) and 1.4:100,000 in France. Mutations in Nav1.4 are mostly missense or rarely in-frame deletions or insertions, usually with a dominant effect. However exceptional recessive homozygosity (Arnold et al., 2015) and a possible recessive compound heterozygosity (Tsujino et al., 2003) have been reported in congenital myasthenic syndromes. A small number of canonical mutations account for a significant percentage of cases (Table 1), e.g., T1313M and R1448C/H for PC, T704M for HyperPP, V445M (Rosenfeld et al., 1997), V1293I (Koch et al., 1995), and G1306A/V/E for SCM, mutations of domains II and III S4 arginines (IIS4 and IIIS4) at position 669 (R>H), 672 (R>H/G/C/S), 1132 (R>Q) (Carle et al., 2006), 1135 (R>H) for HypoPP (Matthews et al., 2009). Mutations at IIS4 arginine 675 (R>Q/G/W) result in a special type of PP with both features of HyperPP and HypoPP . However, beside those frequent canonical mutations, more than 70 different missense mutations at more than 55 different positions in different domains of the protein have been reported in the literature as causative mutations for Nav1.4 channelopathies. The penetrance of Nav1.4 dominant mutations is variable for each mutation: it is high for HyperPP (T704M), PC (T1313M/A and R1448C/H) and SCM (V445M and V1293I) mutations, and lower, with cases of gender-related non-penetrance in pedigrees, for some other mutations such as HypoPP mutations at position 669 or 672 (Ke et al., 2013).
Clinical Description of Nav1.5 Related Channelopathies Nav1.5, which is encoded by the SCN5A gene, is the pore-forming subunit of the main cardiac sodium channel. Nav1.5 related channelopathies affecting cardiac excitability are dominant diseases that, similarly to Nav1.4 in the skeletal muscles, impact cardiac excitability through loss of function or gain of function effects on Nav1.5 activity. Table 2 summarizes the Nav1.5 related channelopathies that are discussed in this review, which only considers pathologies provoked by mutations in the same, i.e., aligned amino acids in Nav1.4 and Nav1.5 (cf. Part Comparison of Missense Mutations. Are there (dys-)Functional Homologies between Nav1.4 and Nav1.5?): the Brugada syndrome (BrS), the long QT syndrome (LQTS), and arrhythmic dilated cardiomyopathy. The latter includes a novel form of cardiac arrhythmia characterized by multifocal ectopic Purkinje-related premature contractions (MEPPCs), associated or not with atrial fibrillation and dilated cardiomyopathy. Consequently, Table 2 is not an exhaustive list of Nav1.5 related channelopathies.

The Brugada Syndrome
The BrS is a primary electrical disorder that is characterized by a specific ECG pattern consisting of ST-segment elevation followed by a negative T-wave in the right precordial leads (Brugada and Brugada, 1992), indicating abnormal electrical activity in the upper part of the right ventricle (right ventricular outflow tract). This ECG pattern is associated with an increased risk of sudden cardiac death (SCD) resulting from polymorphic ventricular tachyarrhythmias or ventricular fibrillation. The incidence of BrS in the general population is currently estimated at 1:2000 (Antzelevitch et al., 2005). This syndrome is 8-10 times more prevalent in males than in females and typically manifests during adulthood, with a mean age of SCD of 41 ± 15 years (Antzelevitch et al., 2005). BrS was first described as a monogenic disease, with autosomal dominant transmission. Although more than 20 genes have been proposed as causally related to BrS, mutations in these genes explain less than 30% of the cases (Crotti et al., 2012;Nielsen et al., 2013;Antzelevitch and Yan, 2015;Veerman et al., 2015). Around 25% of BrS patients possess a mutation in SCN5A. So far, ≈300 mutations in SCN5A have been reported as related to BrS (http://www.ncbi.nlm.nih.gov/clinvar). These mutations lead to a loss of Nav1.5 function and reduce Na + current (I Na ). Besides BrS, loss-of-function mutations in SCN5A also cause isolated cardiac conduction disease and sinus node dysfunction (Remme et al., 2008). ECG signs of conduction defects are also a common feature of BrS. The other genes identified so far are coding for proteins that are involved in generating or regulating the sodium current (Antzelevitch and Yan, 2015), the L-type calcium current (Antzelevitch et al., 2007;Burashnikov et al., 2010;Béziau et al., 2014) or the transient outward potassium current (Delpón et al., 2008;Giudicessi et al., 2011).  Laurent et al., 2012;Mann et al., 2012;Nair et al., 2012;Beckermann et al., 2014 This list is not exhaustive, but corresponds to pathologies caused by Nav1.5 mutations that are homologous to mutations in Nav1.4 (cf. Tables 4, 5).
If BrS was first described as a monogenic autosomal dominant disease, there is accumulating evidence suggesting that it follows a more complex genetic model. Concerning SCN5A, segregation studies performed in large affected pedigrees demonstrate that mutations in this gene are characterized by a low penetrance (47%). In some instances, a single SCN5A mutation can lead to different cardiac arrhythmia phenotypes in the same family or even in a single patient (Kyndt et al., 2001;Probst et al., 2009). Moreover, in some pedigrees, the absence of the familial SCN5A mutation is observed in some affected family members, suggesting other origins for the disease (Probst et al., 2009). Recently, a genome-wide association study in a large cohort of BrS patients has provided the proof of concept that common genetic variants outside the SCN5A gene, e.g., SCN10A and HEY2 loci in the reported study, may have a large effect on the development of the disease (Bezzina et al., 2013). Altogether, these data suggest that the BrS most probably involves combined contribution of different gene variants of variable impact.

The Long QT Syndrome
Congenital LQTS is defined by several criteria including a prolongation of the QT interval corrected for heart rate, i.e., QTc, to values above 440 ms in males and 460 ms in females, due to prolonged ventricular action potentials. LQTS patients are predisposed to ventricular polymorphic tachyarrhythmias (torsades de pointes) that may lead to syncope, seizure or SCD (Amin et al., 2013). The most common form of LQTS (also called Romano-Ward syndrome) is an autosomic dominant disease. Its incidence in the population worldwide is about 1:2000 (Schwartz et al., 2009). To date, genetic defects in 15 different genes have been found in 70% of the LQTS patients (Amin et al., 2013;Giudicessi and Ackerman, 2013). Similar to BrS, the disease penetrance is most often incomplete and highly variable, ranging from 25 to 100% (Priori et al., 1999;Viadero et al., 2011). This suggests that additional genetic and non-genetic factors may modify the clinical manifestations of a given LQTS-causing mutation. In recent years, numerous studies have shown that genetic variants play an important modulatory role in establishing the disease severity (Amin et al., 2013). Among non-genetic factors, hypokalemia, or treatment with drugs inhibiting K V 11.1 (hERG) channels as side effect are well known to favor arrhythmic events. Sex is also a well-known modifier of QT interval duration in LQTS. Post-adolescence and pre-menopause women have a lower repolarization reserve than men and are therefore more prone to QT interval prolongation and cardiac events. This is partially explained by the effects of sex hormones on cardiac ion channel expression and function (Tanabe et al., 1999;Zicha et al., 2003;Bai et al., 2005;Gaborit et al., 2010). The most common types of LQTS are LQTS1 (30-35% of patients; Ackerman et al., 2011), LQTS2 (25-40%), and LQTS3 (5-10%), due to defects in KCNQ1 (K V 7.1 channel), KCNH2 (K V 11.1), and SCN5A (Nav1.5) genes, respectively. Approximately 80% of all LQTS causal mutations are found in these three genes. Clinically, LQTS3 is characterized by unusually increased duration of the ST segment with a late appearance of the T wave (Moss, 2002). It is often more lethal, although less frequent, than LQTS1 and LQTS2 (Priori et al., 2003). Bradycardia and pauses occurring at rest or more particularly during sleep are often at the origin of the arrhythmias, although fatal tachycardia-induced arrhythmias have also been reported for a third of the patients. Most of the SCN5A mutations that were reported to be related to LQTS3 (≈200; http://www.ncbi.nlm. nih.gov/clinvar) alter the fast inactivation process of the channel, leading to persistent inward sodium current causing prolonged membrane depolarizations (Wang et al., 1995;George, 2005).

Arrhythmic Dilated Cardiomyopathy
Dilated cardiomyopathy (DCM) is characterized by systolic dysfunction and, in most patients, left ventricular enlargement or dilatation. It has been associated with the mutations of more than 30 genes, including SCN5A (McNair et al., 2011;Hershberger et al., 2013). Sixteen SCN5A mutations are linked to familial or sporadic cases with DCM with various types of arrhythmias, for example, sinus node dysfunction, conduction delay, and atrial and/or ventricular tachy-arrhythmias (Amin, 2014). Among arrhythmic DCM, the MEPPC syndrome is a recently-described autosomal dominant form of cardiac arrhythmia (Laurent et al., 2012). It is characterized by frequent premature ventricular contractions (PVCs) originating from various ectopic foci along the fascicular-Purkinje system occasionally associated with dilated cardiomyopathy, non-sustained ventricular tachycardias (NSVTs), and sudden death. A similar phenotype was first reported in 2003 by Bezzina and collaborators in a newborn boy and his diseased sister, both genotyped with Nav1.5 W156X and R225W mutations (Bezzina et al., 2003). Both parents and an elder sibling, each one carrier of one or the other mutation, were asymptomatic. For the sister, arrhythmias being the cause of the DCM is unlikely because persistent arrhythmias were only present for a short period. Two other mutations in Nav1.5 (R222Q and R225P) have been linked to this MEPPC syndrome in several families (Laurent et al., 2012;Mann et al., 2012;Nair et al., 2012;Beckermann et al., 2014). In these families, dilated cardiomyopathy, when present, was suggested as a consequence of severe primary electrical dysfunctions.

Phenotypic and Genotypic Overlap between Cardiac and Skeletal Muscle Sodium Channelopathies?
A recently published study shows that patients carrying (or not) SCN4A causative mutations, present with mixed phenotype (BrS and myotonic features) (Bissay et al., 2015). Although SCN4A transcripts are present in human ventricles (Péréon et al., 2003), it is difficult to understand how the gain of function SCN4A mutations can be compared to the loss of function of SCN5A mutations classically associated with Brugada, as discussed in the study of Bissay and collaborators. Another study on a unique family described four patients carrying a SCN4A mutation and presenting with PC (Péréon et al., 2003), two of them having slightly prolonged QTc interval. Both PC and LQTS3 are associated with a gain of function of Nav1.4 and Nav1.5, respectively. In this case, it is tempting to hypothesize that the mutant Nav1.4 channels present in the heart are responsible for the QT prolongation. Identifying more families with such overlap phenotypes would help to confirm the potential mutual influence of both channels on the pathogenesis of cardiac and muscular diseases.

CHANNEL MOLECULAR BASES AND GENE EXPRESSION
Voltage-gated sodium channels consist of an α-subunit, constituting the pore, and accessory β-subunits controlling the expression and activity of the pore-forming subunit. Nav1.4, the most frequent Nav α-subunit expressed in the skeletal muscle is a glycosylated transmembrane protein of 1836 amino acids and has an apparent molecular weight of approximately 260 kDa (George et al., 1992a,b). Nav1.5, the most frequent Nav cardiac α-subunit is 2015-2016 amino acid long, depending on the splice variants, and has a similar apparent molecular weight (Gellens et al., 1992;Makielski et al., 2003;Balasuriya et al., 2012).
The SCN4A gene which encodes Nav1.4 is composed of 24 exons, all containing coding sequence. No alternative splicing events have been reported in the literature. Nav1.5 is encoded by the SCN5A gene, composed of 28 exons, among which exons 2-28 contain the coding sequence. Exon 1 and part of exon 2 encode the 5 ′ untranslated region (UTR) while exon 28 contains the 3 ′ -UTR (Wang et al., 1996). Intron 2 of SCN4A and intron 3 of SCN5A are AT-AC type I introns. Intron 21 of SCN4A and intron 25 of SCN5A are AT-AC type II introns (Wu and Krainer, 1999). All other introns are canonical GT-AG introns. Unlike Nav1.4, mRNA variants of Nav1.5 are detected in the heart of mammals, resulting from alternative splicing. In human and murine hearts, 3 ′ -UTRs present two different splicing variants, generating short or long poly-adenine tails (Shang and Dudley, 2005). In addition, three rare variants were identified only in human, corresponding to alternative splicing of exon 28A by exons 28B-28D coding for truncated and nonfunctional forms of Nav1.5 . To date, only the mechanisms of this splice site are understood. They involve interactions with two splicing factors, the RBM25 and LUC7F3 proteins (Gao et al., 2011;Gao and Dudley, 2013). Four and three splice variants, which differ from the canonical non-coding sequence, were described for the 5 ′ UTR of human and mouse SCN5A mRNAs, respectively. These transcripts originate from the alternative splicing encompassing exons 1 (designated 1A, 1B, 1C, and 1D) and 2, and are preferentially expressed in the heart as compared with other tissues. Also, a neonatal isoform containing a neonatal exon 6A of 31 nucleotides has been reported. This form presents a difference of seven amino acids in the S3-S4 loop of domain I, in comparison with exon 6 of the adult form (Rook et al., 2012). Ventricular myocardial analysis displayed abnormal splicing of SCN5A exon 6, characterized by over-expression of this neonatal isoform, in one patient who present DCM with conduction system disease (Wahbi et al., 2013). These findings suggest a potential implication of mis-splicing of SCN5A in the cardiac defect observed in this patient.
Two distinct sodium currents and channels were historically described in skeletal muscle depending upon the developmental and innervation status of the myofiber. SkM1, the TTX-sensitive sodium channel expressed in innervated adult myofibers, corresponds to Nav1.4 and is the main skeletal muscle sodium channel (Trimmer et al., 1989;Kallen et al., 1990). SkM2, the TTX resistant sodium channel expressed in immature and denervated myofibers, corresponds to Nav1.5. In rodents, SCN4A expression increases just after birth concomitantly with the decrease of SCN5A gene expression (Stocksley et al., 2005). SCN4A expression is not sensitive to myofiber denervation by contrast to SCN5A gene expression, which was found to be upregulated in response to denervation (Awad et al., 2001).
The SCN4A promoter contains distinct positive-acting promoter E-box and negative-acting repressor E-box that cooperate to yield specific gene expression in differentiated skeletal myofibers (Kraner et al., 1998(Kraner et al., , 1999. It is suggested that the muscle specificity of SCN4A expression result from the binding of two basic helix-loop-helix transcription factors (bHLH) of the muscle-specific MyoD family, myogenin and MRF4 for initiation and maintenance, respectively, to the positive-acting promoter E-box located upstream the translation initiation site. NFI would be another major regulator of SCN4A gene expression acting in concert with bHLH factors, especially MRF4 (Hebert et al., 2007). The density of Nav1.4 is around 20 times higher at the neuromuscular junction (NMJ), in part as a result of local mRNA accumulation (Stocksley et al., 2005). Although the promoter element responsible for the transcriptional regulation of subsynaptic genes in response to neuronal factors at the NMJ is the N-box (TTCCGG) (Méjat et al., 2003), no N-box is present within the promoter of SCN4A, suggesting the involvement of other regulatory elements.
Similarly to alternative splicing, more is known concerning the regulation of the SCN5A promoter, compared with SCN4A. After the identification of a first promoter region for human SCN5A which includes multiple positive and negative cisacting elements extending into intron 1 (Yang et al., 2004), two other promoter regions for murine SCN5A (designated P2 and P3) containing two distinct cardiac-specific enhancer regions were identified and functionally characterized (Shang and Dudley, 2005). In human and rat, the segment immediately upstream of the major transcription start site contains three GC boxes that could serve as binding sites for the Sp1 transcription factor, which are homologous to the CACC boxes recognized in promoters of muscle restricted genes, and an Ebox binding site for bHLH factors (Yang et al., 2004). The human sequence also includes an additional C-rich motif which is recognized as a major regulator of expression in myocytes. Further, Yang and collaborators have characterized a binding site for GATA in intron 1, which is also known as a key regulator of gene expression in the heart. Surprisingly, variants in SCN10A (encoding Nav1.8 of which expression is extremely low in heart and undetectable in atrioventricular bundle) are associated with alterations of cardiac conduction parameters and BrS (van den Boogaard et al., 2014). Van den Boogaard and collaborators have shown that the SCN10A variants act more likely through an alteration SCN5A gene expression level. They have demonstrated that a cis-regulatory element located in SCN10A gene -which is immediately located next to SCN5A-was able to interact with both SCN5A and SCN10A promoters. Furthermore, they described, using healthy human heart samples, a direct correlation between the SCN5A (but not SCN10A) expression and the presence of the rs6801957 riskassociated SNP in the SCN10A intronic enhancer. Together, their data provided a genomic mechanism explaining how a common genetic variant at SCN10A locus influences cardiac physiology and predispose to Brs.

ASSOCIATED/REGULATORY PROTEINS
Although expression of Nav1.5 or Nav1.4 α-subunits alone results in the generation of functional channels in heterologous expression systems, it is now quite clear that the regulation of gating and/or expression of the Nav subunits substantially relies on a variety of other accessory/regulatory proteins (Abriel, 2010;Rook et al., 2012). Interestingly, the alignment of Nav1.5 and Nav1.4 amino acid sequences could facilitate the identification of novel associated/regulatory proteins of the counterpart channel subunit. In addition, this direct sequence comparison has contributed, as for Navβ1 (Makita et al., 1996), and will certainly continue to contribute to localizing the structural determinants involved in the channel regulation. In this respect, Table 3 and Figure 1 recapitulate the Nav1.5 or Nav1.4 amino acid sequences previously identified to mediate interaction with associated/regulatory proteins, and indicates the corresponding sequences in the other Nav α-subunit. Whereas, a number of Nav1.5 interacting proteins, with their binding sites in the channel subunit, have been described in the literature (see references in Table 3), very little is known for Nav1.4 (Figure 1). Nevertheless, it is interesting to note that the amino acid sequence similarity obtained for some binding sites is high, suggesting the possibility that both channel subunits share the same associated/regulatory proteins. This is the case for example of calmodulin, which associates with the very well conserved (100% sequence similarity) IQ-motif on both Nav1.5 and Nav1.4 C-terminal domains (Tan et al., 2002;Young and Caldwell, 2005). Most of the proteins shown to interact with Nav1.5 on a site conserved in Nav1.4 are ubiquitously expressed (dynamitin, 14-3-3, CaMKII, MOG-1, calmodulin, FGF, PTPH1/PTPN3, SAP97), suggesting that an interaction with Nav1.4 may take place in the skeletal muscle cells (Marfatia et al., 2001;Blair et al., 2006). When not ubiquitously expressed, proteins known to interact with Nav1.5 are also expressed in skeletal muscle (α actinin2) that argue for a possible interaction with Nav1.4 (Foley and Young, 2014). Conversely, weaker sequence similarity may suggest different affinities, sites or absence of interaction/regulation. This is the case for example of Navβ1 for which the region within D1/S5-S6 that confers regulation of Nav1.4 in Xenopus oocytes (Makita et al., 1996) is not very well conserved in Nav1.5 (63.1% sequence similarity) which is also regulated by Navβ1, suggesting that the structural determinants of the interaction of Nav1.5 or Nav1.4 with Navβ1 are different. Finally, it is striking to note the complete absence of the PY-motif from the C-terminus of Nav1.4. This suggests that the regulation of Nav1.4 channel internalization and/or degradation is achieved through different mechanisms as compared to Nav1.5 for which cell surface expression is regulated through the ubiquitin-proteasome pathway (van Bemmelen et al., 2004;Rougier et al., 2005). These mechanisms remain to be identified.

FIGURE 1 | (A)
Schematic representation of the localization of the binding sites of associated/regulatory proteins identified for Nav1.4 (blue), Nav1.5 (pink), or both (purple). Nav1.4 interaction with ankyrin is only suggested by chimeric constructs, it remains to be studied with full length proteins (Lemaillet et al., 2003). DI to DIV, domains I to IV; ID, intracellular interdomains; N-ter and C-ter, N-and C-terminus ends. (B) Table presenting the amino-acid (aa) numbering and length of specific regions and domains in Nav1.4 and Nav1.5 proteins. It is noticeable that Nav1.4 IDI-II total amino-acid length is shorter than Nav1.5 IDI-II.
patch-clamp in heterologous expression systems, mutated Nav1.5 channels are showing different types of loss of function, such as a decrease in current density, a positive shift in the activation curve, a negative shift in the inactivation curve, or a loss of regulation by PKA (Tarradas et al., 2013;Zeng et al., 2013;Aiba et al., 2014). Mutations in SCN5A have also been found in patients presenting with LQT syndrome (named LQTS3 when SCN5A is mutated). As opposed to BrS, mutated channels in LQTS3 patients show a gain a function, mainly through an increase in a persistent Na + current (cf. Part Clinical Description of the Main Nav1.4 and Nav1.5 Related Pathologies). As a result, BrS mutations are associated to membrane hypo-excitability, whereas LQTS3 mutations are associated to prolonged action potential, referred here as membrane hyper-activity.
In the skeletal muscle, a similar binary classification is observable among the 70 mutations identified so far, that are nearly exclusively missense or rarely in-frame deletions/insertions. NDM are linked to membrane hyperexcitability, often due to defective inactivation and hence a gain of function of Nav1.4 channel activity (Clarke et al., 2011). On the contrary, hypoPP is linked to membrane hypo-excitability, and is often due to the apparition of an aberrant current through the gating pore that can be a proton or a monovalent cation current (Sokolov et al., 2007). This so-called "omega" current (or gating pore current) causes paradoxical depolarization of myofibers in low K + , which inactivates Nav1.4 and renders myofibers non excitable. Seemingly paradoxical, hyperPP is associated with gain of function of Nav1.4 (as observed for myotonia) but loss of function on skeletal muscles (paralysis). As for myotonia, defective inactivation of Nav1.4 is often observed and favors membrane depolarization. The paradox is resolved if we consider that wild type Nav1.4 channels will be more inactivated due to a slightly more depolarized membrane, thus causing a loss of sarcolemmal excitability and myofiber paralysis (Cannon, 2015). The development of myotonia or hyperPP may depend on the degree of membrane excitability. This has been suggested for instance when in the same family, females carrying the M1370V mutation develop only a myotonia (PC) whereas males are presenting with both myotonia and hyperPP (Okuda et al., 2001).
Nav1.4 and Nav1.5 are similar. If we consider the aligned region between Nav1.5 and Nav1.4, which represents 95% of Nav1.4 sequence, 67% of the amino-acids are identical. Knowing that, one can wonder whether mutations have been identified at equivalent positions in both channels, and whether, in this case, the new amino-acid is the same, such as Q270K in both Nav1.4 and Nav1.5 or V445M in Nav1.4 and V411M in Nav1.5 (V445 is aligned with V411). It is possible to use an online  Mitrovic et al., 1995;Fleischhauer et al., 1998G1481E LQT3 Kapplinger et al., 2009 Domain IV S6 V1589M Overlap PC-SCM Heine et al., 1993;Mitrovic et al., 1994;Hayward et al., 1999V1763M LQT3 Chang et al., 2004Ma et al., 2013 Same amino acid substitutions occurring in both channels lead to consistent pathologies (in green) regarding membrane excitability.   Ptácek et al., 1992;Chahine et al., 1994;Mohammadi et al., 2003;Holzherr et al., 2014R1623Q LQT3 Kambouris et al., 1998Makita et al., 1998IVS4 R1448C PC Ptácek et al., 1992Chahine et al., 1994;Featherstone et al., 1998R1623Q LQT3 Kambouris et al., 1998Makita et al., 1998IVS4 R1448P PC Featherstone et al., 1998R1623Q LQT3 Kambouris et al., 1998Makita et al., 1998 IVS4 R1448S PC (mild) Bendahhou et al., 1999R1623Q LQT3 Kambouris et al., 1998Makita et al., 1998 Rojas et al., 1991;Cannon and Strittmatter, 1993;Hayward et al., 1999;Rojas et al., 1999M1766L LQT3 Valdivia et al., 2002Ye et al., 2003 Divergent amino-acid substitutions occurring in the two channels lead either to consistent (in green) or inconsistent (in red) pathologies regarding membrane excitability. compilation that has been proposed using a paralog annotation approach in order to retrieve homologous or nearly homologous variants in both genes (Ware et al., 2012;Walsh et al., 2014). If the same mutations of homologous residues exist, do they give rise to similar dysfunction on both channels? If yes, we can expect that both mutations give rise to the same change in membrane excitability. For instance, if a Nav1.5 mutation leads to hyper-activity of cardiac cells (LQTS3), the corresponding mutation in Nav1.4 may also give rise to a hyper-excitability phenotype of the skeletal muscle cells such as HyperPP, PC, or SCM. Tables 4, 5 present all the corresponding amino acids found to be mutated in patients with cardiac (Nav1.5) or neuromuscular (Nav1.4) pathologies. Table 4 and Figure 2A list the mutations for which the amino acid substitutions are the same, and Table 5 and Figure 2B those for which they are divergent. When looking at Table 4 and Figure 2A, it is striking to observe that all paralog mutations give rise to clearly consistent FIGURE 2 | Schematic representation of the equivalent Nav1.4/Nav1.5 amino acids with similar (A) or divergent (B) mutations in patients with skeletal (Nav1.4) or cardiac (Nav1.5) pathologies. Upward triangles indicate a gain of function on membrane excitability and downward triangles a loss of function on membrane excitability. Green and red triangles indicate consistent and inconsistent effects on Nav1.4 and Nav1.5 regarding membrane excitability, respectively. Upper amino acid number/letter, Nav1.4; Lower amino acid number/letter, Nav1.5. DI to DIV, domains I to IV; ID, intracellular interdomains; N-ter and C-ter, N-and C-terminus ends. functional effects, except one which is at first unclear, as detailed below. Indeed, all Nav1.4 mutations linked to membrane hyperexcitability (PC, SCM, and HyperPP) correspond to Nav1.5 mutations linked to membrane hyper-activity (LQTS3), except R675Q (R814Q in Nav1.5). The comparison between Nav1.4 R675Q and Nav1.5 R814Q is not obvious because the pathology induced by Nav1.4 R675Q mutation is difficult to classify as Normo/Hyper PP or Hypo PP . Indeed, patients experienced normal as well as decreased potassium levels concomitant to attacks. The rat ortholog of human Nav1.4 R675Q generates an omega current activated by depolarization when expressed in Xenopus oocytes (Sokolov et al., 2008) (cf. above). The omega current represents less than 1% of the peak pore current but it remains constant after slow inactivation of the pore current and requires high hyperpolarizations to deactivate. Therefore, it is suspected that this current, carried by Na + and K + ions, maintained during trains of action potentials and with a residual non-deactivated activity at resting potential could lead to sodium accumulation and a decrease in membrane excitability. It will be interesting to test whether the corresponding mutation in Nav1.5 is also responsible for an omega current. Moreover, the R675Q Nav1.4 mutation gives rise to a hyperpolarizing shift of the inactivation curve and a slower recovery from inactivation when expressed in HEK293 cells. Wu et al., 2014). Altogether, these observations suggested us to rank it as a hypo-excitability causing mutation, consistent with the BrS phenotype (loss of function) induced by the homologous Nav1.5 mutation R814Q. Table 4 and Figure 2A summarize the (dys-)functional homology between the equivalent mutant in Nav1.4 and Nav1.5. On the contrary, Table 5 and Figure 2B show that divergent amino acid substitution at the equivalent position leads to some inconsistencies (in red, 5/12). This suggests that the nature of the amino acid substitution is determinant for the direction of the functional net effect (loss or gain of function).
At last, we focused on five equivalent mutations that have been studied extensively in patch-clamp in both Nav1.4 and Nav1.5. Table 6 shows changes in each biophysical parameter for these mutations. When looking at the direction of the functional effects (gain or loss of function), we observe two major points. First, a strikingly similar functional effect of the same mutations in both channels. Second, all the gain of function mutations leading to hyper-activity/excitability provoke an increase in the persistent current, when measured, suggesting that this mechanism plays a major role in the pathogenesis of Nav channelopathies.
Recently, an omega current has been observed in Nav1.5 mutant channels identified in patients presenting with arrhythmic DCM or MEPPC (Gosselin-Badaroudine et al., 2012b;Moreau et al., 2015b). This omega current, due to mutations of arginine in the S4 of domain I, is similar to the one observed in Nav1.4 (Sokolov et al., 2007;Struyk et al., 2008;Francis et al., 2011;Gosselin-Badaroudine et al., 2012a;Groome et al., 2014). This further strengthens the functional similarity between Nav1.4 and Nav1.5 in pathophysiological situations. A common feature of MEPPC, is an increase in window current provoked by the Nav1.5 R225W, R222Q, and R225P mutations, increasing cardiac excitability of the fascicular-Purkinje system (Laurent et al., 2012;Mann et al., 2012). Another common feature of two of these mutations: R222Q and R225W is the presence of an omega current. This Nav1.5 omega current may be responsible for the peculiar cardiac phenotype (Moreau et al., 2015a), similar to the omega current of Nav1.4 being responsible for the hypoPP phenotype, through sodium accumulation and a decrease in membrane excitability (Sokolov et al., 2008). Indeed, most of the SCN5A mutations linked to DCM are located in the voltage sensor domain (VSD) as pointed by McNair et al. (2011). However, in some cases DCM may be secondary to arrhythmias and window current increase. For instance, preventing arrhythmias by quinidine improved the ventricular function (ejection fraction) in patients with the Nav1.5 R222Q mutation, via a decrease in the window sodium current (Laurent et al., 2012). The use of specific inhibitor of the alpha pore and the omega (or gating pore) current would allow to test for the respective role of the altered gating (activation, inactivation) and the omega current on the development of the pathology. Noteworthy, the various localization of the Nav1.4 mutations giving rise to omega current (in domains I, II, and III) strongly suggests that similar mutations in Nav1.5 will be identified in domain II and III in addition to the ones already identified in domain I (Moreau et al., 2015b).
To conclude, given the sequence similarity between Nav1.4 and Nav1.5, any characteristics described for one channel subunit may shed light on the properties of the counterpart channel subunit, such as the presence of specific protein partners, or the effects of a specific amino acid substitution. One can argue that the effect of a mutation on Nav1.4 is difficult to compare with Nav1.5 since the different molecular and cellular environment may drastically modify the effect of the mutation. Nevertheless, we noticed that the same mutation lead to comparable effect regarding membrane hypo or hyperexcitability (Table 4 and Figure 2A). This suggests that the cellular environment is usually not able to invert the effect of a mutation from gain to loss of function phenotypes and reciprocally. Such comparison between Nav1.4 and Nav1.5 will probably draw more and more interest, to address the challenge of interpreting and understanding pathogenicity of rare SCN4A or SCN5A variants revealed by next-generation sequencing studies (Arnold et al., 2015;Bergareche et al., 2015;Coll et al., 2015).