An Up-to-Date Overview of the Complexity of Genotype-Phenotype Relationships in Myotonic Channelopathies

Myotonic disorders are inherited neuromuscular diseases divided into dystrophic myotonias and non-dystrophic myotonias (NDM). The latter is a group of dominant or recessive diseases caused by mutations in genes encoding ion channels that participate in the generation and control of the skeletal muscle action potential. Their altered function causes hyperexcitability of the muscle membrane, thereby triggering myotonia, the main sign in NDM. Mutations in the genes encoding voltage-gated Cl− and Na+ channels (respectively, CLCN1 and SCN4A) produce a wide spectrum of phenotypes, which differ in age of onset, affected muscles, severity of myotonia, degree of hypertrophy, and muscle weakness, disease progression, among others. More than 200 CLCN1 and 65 SCN4A mutations have been identified and described, but just about half of them have been functionally characterized, an approach that is likely extremely helpful to contribute to improving the so-far rather poor clinical correlations present in NDM. The observed poor correlations may be due to: (1) the wide spectrum of symptoms and overlapping phenotypes present in both groups (Cl− and Na+ myotonic channelopathies) and (2) both genes present high genotypic variability. On the one hand, several mutations cause a unique and reproducible phenotype in most patients. On the other hand, some mutations can have different inheritance pattern and clinical phenotypes in different families. Conversely, different mutations can be translated into very similar phenotypes. For these reasons, the genotype-phenotype relationships in myotonic channelopathies are considered complex. Although the molecular bases for the clinical variability present in myotonic channelopathies remain obscure, several hypotheses have been put forward to explain the variability, which include: (a) differential allelic expression; (b) trans-acting genetic modifiers; (c) epigenetic, hormonal, or environmental factors; and (d) dominance with low penetrance. Improvements in clinical tests, the recognition of the different phenotypes that result from particular mutations and the understanding of how a mutation affects the structure and function of the ion channel, together with genetic screening, is expected to improve clinical correlation in NDMs.


MYOTONIC DISEASES
Electrical properties of the skeletal muscle fiber membrane (in a wide variety of different organisms) are characterized by a high resting membrane permeability for Cl − ions. These Cl − ions are transported by Cl − ion channels belonging to the CLC family, which includes Cl − ion channels and Cl − /H + exchangers that are found in all phyla from bacteria to mammals (6)(7)(8). The ClC-1 channel is specifically present in skeletal muscle where it accounts for the chloride conductance, G Cl , which amounts to ∼80% of the resting membrane conductance in resting muscles (9)(10)(11)(12)(13)(14)(15). Due to this high conductance and because the equilibrium potential of Cl − is close to the muscle resting membrane potential, ClC-1 conducts membrane currents that inhibit muscle excitability (6). The effective Cl − homeostasis is central in the generation and propagation of the action potential in muscle fibers.
Indeed, it is believed that the prominent role of G Cl in action potential repolarization and membrane potential stabilization is related to the large cell size and the peculiar t-tubular system of skeletal muscle. If repolarization were mediated exclusively by K + channels (as in most neurons), the extracellular K + concentration in the restricted t-tubular space would rise significantly, leading to depolarization and eventual inactivation of Na + channels (15). In fact, the "myotonic runs" of repeated action potential firing are partially caused by K + accumulation (15). Importantly, a still open question is whether ClC-1 channels are actually preferentially located in the t-tubular membrane or in the surface sarcolemma (16)(17)(18)(19).
It is well-known that excitation-contraction coupling, triggered by the nerve-impulse induced muscle action potential, is orchestrated by multiple factors (mainly ion channels) that lead to the release of Ca 2+ from the sarcoplasmic reticulum. However, changes in the electrical properties of an excitable muscle, that can occur by both acquired or inherited bases, are well-recognized causes of muscle malfunction in humans (20). Reduction in muscle excitability may be the major cause of muscle weakness leading to fatigue (21), while hyperexcitability may lead to a sustained bursts of discharges that cause involuntary after-contractions, a phenomenon known as myotonia, the classical and leading sign of several hereditary diseases of skeletal muscle (20,22). Therefore, myotonia is a clinical sign of skeletal muscle that results from an increased excitability of the muscle fiber membrane such that a single nerve stimulus triggers a burst of repetitive action potentials causing a delay in the temporal course of muscle relaxation (23). In patients, this condition can be detected as both electrical and clinical myotonia (24,25). Electrical myotonia is detected on electromyographic (EMG) tests as repetitive muscle fiber potential discharges, with waxing and waning frequency and amplitude with a firing rate between 20 and 80 Hz, while clinical myotonia is physically demonstrated by slowed muscle relaxation during repetitive hand grip, eye closure, or after tapping various muscles, such as the finger extensors (26).
Symptomatically, patients experience myotonia as stiffness at the beginning of motion, better exemplified during initial attempts at relaxation of hand grip (grip myotonia) and following percussion of the muscle located in the bases of the thumb (percussion myotonia) (27). Myotonia can ameliorate with repeated movements, a phenomenon known as warmup, but also, it can get worse with activity and become paradoxical (28,29).
Diseases associated with this sign are collectively termed myotonias and according to other clinical features (progressive muscle wasting and weakness, dystrophic changes), they are classified as: (1) dystrophic myotonias (DM) and (2) nondystrophic myotonias (NDM). Myotonic dystrophy type 1 (DM1) and type 2 (DM2), caused by the expansion of unstable microsatellites, belong to the first group, and are beyond the scope of this review. A group of five diseases, collectively called myotonic channelopathies, caused by mutations in voltage-gated Na + and Cl − channel genes, belong to the second group (30). This review is focused on the latter group of diseases.

NON-DYSTROPHIC MYOTONIAS
Dominantly or recessively inherited disorders caused by ion channel dysfunction include myotonia congenita (MC) (Thomsen's disease and Becker type myotonia), paramyotonia congenita (PC), hyperkalemic periodic paralysis (hyperPP) with myotonia and the sodium channel myotonias (SCM). These diseases show distinctive clinical features that allow separating them from the myotonic dystrophies. They are divided into two groups: the chloride and the sodium channelopathies. The first group, called myotonia congenita (MC), is subdivided in Thomsen's disease and the Becker generalized myotonia. The other group of three diseases listed above belong to the sodium channelopathies (24,25,(30)(31)(32).

Chloride Channelopathies
For precise control of muscle contraction by nerve activity, in normal conditions, a single nerve action potential triggers only a single muscle action potential. This is to a large extent guaranteed by the large G Cl that aids in repolarizing the action potential and, once repolarized, continues to stabilize the membrane potential, preventing thereby the insurgence of a train of action potentials (33). Reduction of the Cl − conductance in the skeletal muscle causes myotonia congenita (MC), the most common ion channel disease (34). The first description of this pathology dates back to the late nineteenth century, when Danish physician Asmus Julius Thomas Thomsen described it for himself and for some of his family members, with an autosomal dominant inheritance pattern (35). The dominant form of MC was called Thomsen's disease (DMC) after that (36). Almost a century after that first description, German professor Peter Emil Becker described a MC variant with autosomal recessive inheritance pattern, a variant that was called recessive generalized myotonia or Becker myotonia (RMC) (37). MC is electrophysiologically characterized by presenting increased excitability of the muscular fiber, which is due to repetitive action potentials of the muscle membranes; this is reflected in clinical myotonia, muscular stiffness (that is worse after rest) and hypertrophy (38,39). Myotonia in MC is clinically highly variable, ranging from myotonic discharges only detectable on the EMG test to disabling muscle stiffness at an early age (5). The leg muscles are commonly affected and handgrip myotonia is detected in about ¾ of patients (22). Almost every skeletal muscle in the body might show muscular stiffness, but it is ameliorated by exercise (warmup phenomenon). The clinical picture depends on whether the disease is inherited as autosomal dominant or autosomal recessive. RMC is more common (in most countries) and severe than DMC (5,24,25,(40)(41)(42). In DMC, age of onset is usually at birth or very early in infancy. The child might show unusually defined muscles in the extremities, delayed relaxation of the eyelids after forceful closure following sneezing or during crying, and hypertrophy is rare in childhood but common in adulthood (31,43). Severity varies from mild to moderate, there is no progression of the symptoms, and patients can experience a normal life (43). In RMC, age of onset is usually later than in DMC (25,43), but it has been also reported that age of onset could be earlier in RMC (44). Myotonia is generalized with moderate to pronounced hypertrophy, where many patients use to have a body-builder appearance due to hypertrophy triggered by the involuntary after-contractions (20,43). Myotonia is more severe in the recessive variant and usually, typical transient muscular weakness (lasting from seconds to as long as 30 min) is also observed (10,45), which may lead to recurrent falls (31,34). Muscle strength is normal in this variant but the disease slowly progresses in some patients (34,46).

Sodium Channelopathies
The upstroke of the action potential is mediated by opening of voltage-gated Nav1.4 sodium channels that generate an inward Na + current that renders the cells positive inside (depolarization) (31). Malfunction of these channels cause several human hereditary diseases of the skeletal muscle. Diseases belonging to this group are: paramyotonia congenita (PC), hyperkalemic periodic paralysis (HyperPP) with myotonia and the sodium channel myotonias (SCM). Similar to MC, these diseases are characterized by increased muscle membrane excitability that leads to repetitive action potentials; however, the underlying physiological defect is different compared with the chloride channelopathies (24,47). Symptoms in sodium channelopathies are episodic and vary from patient to patient and also from time to time on each affected individual (47). Clinical myotonia, affecting hands, face, upper and lower extremities is generalized, going from absent to very severe, depending on the disease, and usually gets worse with exercise; while electrical myotonia may be diffusely present in all muscle and in some cases, may worsen due to cold exposure, which may also induce weakness in some patients, and clinical myotonia. The length of the attack is very variable, lasting from minutes up to 7 days. Muscle hypertrophy is not as common as in the chloride channelopathies, but it is present in variable degrees in some patients. Age of onset of the disease is usually in the first decade and it might progress later in life in some patients (25,27,(48)(49)(50).
PC was first described by Eulenburg (51). Age of onset of first symptoms is at birth or during the first decade of live, affecting mainly the muscles from neck, face and upper limbs. As indicated by its name, PC shows paradoxical myotonia, which gets worse with repeated movements and/or cold exposure. PC show variable weakness, which becomes evident after myotonic stiffness attacks, the major symptom in PC (20,47). PC patients may experience HyperPP weakness-like phenotype after pronounce cooling or vigorous exercise (20).
In HyperPP, age of onset of the first symptoms (recurrent episodes of weakness as the main symptom) is during childhood. Symptoms are triggered by ingestion of K + -rich food, rest after vigorous exercise, and other environmental factors (47,(52)(53)(54). The episodes of weakness vary from minutes to several hours, with normal weakness between attacks. Patients affected with HyperPP frequently have myotonia (symptomatically in 12.5% or by EMG in 50% of patients), which can be from mild to moderate, particularly with the onset of a weakness attack. Some patients develop paramyotonia, highlighting the extensive clinical overlapping between these two conditions (20,47,53,55).
SCM is less well-defined with onset of first symptoms from childhood, as in acetazolamide-responsive myotonia, to adolescence, as in myotonia fluctuans (47). Myotonia can be present from mild (in myotonia fluctuans) to severe (in myotonia permanens affecting swallowing and breathing); no periodic paralysis has been observed in SCM patients, while there are some patients that have shown potassium-aggravated myotonia (K + -sensitive myotonia) (20).

GENETICS AND MUTATION IN MYOTONIC CHANNELOPATHIES Chloride Channelopathies
As mentioned above, MC is the most common hereditary ion channel disorder in humans, showing a prevalence between 1:23,000 and 1:50,000 for the recessive form (Becker Myotonia), while the dominant form (Thomsen's disease) is a bit less common in most countries (40,42,56). Both conditions are caused by mutations in the chloride voltage-gated channel 1 (CLCN1) gene (10,(57)(58)(59), which is located in chromosome 7q35 and encodes the voltage-gate chloride channel (ClC-1), belonging to the CLC family of chloride channels (60,61). Although the pathogenesis of MC is not fully understood, it is well-known that mutations in CLCN1 produce a reduction of the Cl − conductance that leads to membrane hyperexcitability, triggering repetitive action potentials (24,25,31). The channel conducts chloride ions over the entire physiological voltage ranges and is the major mediator of chloride conductance in skeletal muscle (13,14,31,62,63). Two subunits of the channel are required to come together to form the functional channel, and thus, work as double-barreled homodimers (64)(65)(66). The CLCN1 gene has 23 exons, with more than 200 different mutations described in this disease (4, 41, 43, 67, 68) (http:// www.hgmd.cf.ac.uk/ac/index.php). Mutations are found through the entire gene sequence, being present in the N-terminal, transmembrane, and C-terminal domains of ClC-1. Different types of mutations have been found in the CLCN1 gene, including nonsense, splice-site, missense, frameshift (insertion/deletions), and deletion/duplication mutations, with exon eight becoming a hot spot for DMC (20,41,67,(69)(70)(71). The recessive inheritance is conceptually explained by a loss-of-function effect caused by the mutations without significantly impacting on the formation or function of dimeric ClC-1 channels. On the other hand, the dominant inheritance is explained by a dominant-negative effect of mutated subunits on heteromeric mutant/WT channels. Most of the 200 different mutations identified and described behave as recessive, with the majority of the patients being compound heterozygous (carriers of two different recessive mutations). Only about 27 mutations have been associated with DMC, while about other 59 mutations have an unclear inheritance pattern, are sporadic or have been also shown to display a recessive inheritance pattern (http://www.hgmd.cf.ac.uk/ac/ index.php). Therefore, a clear distinction between dominant and recessive mutations is not always possible (5,39,41,43,(72)(73)(74). Thus, far, there is no other clinical phenotype associated with mutations in the CLCN1 gene.

Sodium Channelopathies
Na + channelopathies are not as common as Cl − channelopathies, showing a combined prevalence of about 1:100,000 (42). These disorders are caused by mutations in the sodium voltagegated channel alpha subunit 4 (SCN4A) gene, which is located in chromosome 17q23 and encodes the voltage-gated sodium channel (Nav1.4) of skeletal muscle (75)(76)(77). They are a heterogeneous group of autosomal dominant disorders with high penetrance (20,55). Mutations in SCN4A cause disruption of fast inactivation of the channel, which can be incomplete or slowed (78)(79)(80), leading to repetitive action potentials (myotonic runs) and consequent intracellular sodium accumulation that depolarizes muscle cells and can lead to inactivation of the Na + channels (25,31,32,47). Depending if depolarization is mild or not, myotonia or paralysis might appear, respectively (81). Nav1.4 is a channel formed by a single unit of Nav1.4 protein, which contains four repeated domains (DI-DIV), each one consisting of six transmembrane segments (S1-S6). The loops between S5-S6 segments from the four domains come together to form the ion-conducting pore, acting as a selective filter. Meanwhile, the S4 segment of each domain is in charge of sensing the voltage changes (31,32,47). The SCN4A gene has 24 exons, with about 83 different mutations described in the gene, but only about 65 of them have been associated with myotonia (40, 82-86) (http://www.hgmd.cf.ac.uk/ac/index.php). All SCN4A mutations correspond to missense mutations, with the single exception of a deletion/insertion mutation located in the splice site in intron 21 (87,88). All SCN4A myotonia mutations studied produce a gain-of-function effect of Nav1.4, resulting in defects of channel inactivation or enhancement of activation, which explain the dominant inheritance pattern of the diseases (20,82). Mutations have been located through the entire gene sequence, but depending on the disease, they tend to group differentially. For instance, mutations associated with HyperPP with myotonia are generally located in the inner regions of the transmembrane segments or in the intracellular interlinking loops in repeat domains DII and DIV of the channel, eliciting a persisting inward sodium current, which impairs repolarization and increases membrane excitability (82,89). In PC, mutations have been found throughout the gene, with exon 24 appearing to be a hot spot (90), but their impairment of fast inactivation is less notorious than the one associated with HyperPP (81). Regarding SCM, although mutations have been found throughout the gene, there are more likely located in the N terminus of the channel, in repeat domain D1, particularly in the inactivation gate (82). Interestingly, in recent papers, the authors describe several SCN4A mutations, previously reported in unrelated myotoniapositive families, new or de novo mutations, that contribute to apnea during the physiological stress of seizures, severe respiratory failure or associated with paradoxical vocal fold motion (PVFM) (91)(92)(93). Neonatal laryngospasm and unusual distribution of myotonia and other NDM signs have also been reported in several NDM patients, who have been shown to carry different SCN4A mutations, such as G1306E, I693T, A799S, N1297K, and T1313M (although not all patients that carry these mutations show childhood or neonatal respiratory problems) (78,(94)(95)(96)(97)(98)(99). This expands the spectrum of phenotypes associated with mutations in the SCN4A gene. In addition to myotonic diseases associated with SCN4A mutations, other diseases, that are beyond the scope of this review, also present mutations in this gene, such as: hyperkalemic periodic paralysis without myotonia (HyperPP), hypokalemic periodic paralysis (HyppoPP), normokalemic periodic paralysis (NormoPP), and congenital myasthenic syndrome (CMS) (20,82). Recently, it has also been suggested that a subset of cases with sudden infant death syndrome (SIDS) might be due to SCN4A mutations, as one report (and the only one thus far) has found novel or very rare functionally disruptive SCN4A genetic variants associated with SIDS, although the authors indicate that new studies in other populations are required to confirm their finding (100).

COMPLEX GENOTYPE-PHENOTYPE RELATIONSHIPS IN MYOTONIC CHANNELOPATHIES
In order to provide accurate prognostic information to the patients and families affected with a hereditary disease, it is essential to have appropriated genotype-phenotype relationships. In the case of the non-dystrophic myotonias, this has been extremely difficult mainly due to two factors: (1) the wide spectrum of symptoms and overlapping phenotypes present in both groups (Cl − and Na + myotonic channelopathies) (24,25,47), and (2) both genes, CLCN1 and SCN4A, present high genotypic variability, with more than 200 or 65 different mutations already described, respectively (http://www.hgmd. cf.ac.uk/ac/index.php) (23,43). Although many mutations cause a unique and reproducible phenotype in all patients, some SCN4A and CLCN1 mutations cause similar phenotypes. Most worryingly, for several mutations, very different clinical phenotypes have been found in different carriers of the same mutation (23,25,47,101), severely compromising the genotypephenotype correlation in NDM. Another important limitation for the improvement of genotype-phenotype correlations has been the lack of a sufficient number of individuals carrying each mutation (23), in particular in the case of novel mutations [such as the very recent study that described seven novel CLCN1 mutations (68)], which makes the situation even more complex, not without mentioning all those cases that show myotonia or a myotonia-like phenotype but in which the mutation has not been found. It is worth mentioning though, that correlation of mutations with the clinical phenotype gives insights into the pathophysiology of human channelopathies, and although some correlation exists between specific mutations and the associated clinical manifestations, this is vague (47).
The poor correlations are most evident in Cl − channelopathies, where the same mutation can be inherited as dominant or recessive with different clinical manifestations, for example, F167L (39,67,68,102,103), A313T (67, 104), or W433R (44) (see Table 1). By recording and analyzing the ion currents of heterologously expressed mutant channels in different in vitro expression systems (Xenopus oocytes or HEK cells), much progress has been made in understanding how specific mutations affect the function of a particular ion channel, in providing insights onto the mechanism for the inheritance pattern, but also in the role that the voltage-gated ion channels play in excitable tissues (20,25,105). These analyses have contributed to improve to some extent, the clinical correlations through a better understanding of the channel dysfunction and its associated clinical picture. Nevertheless, of the more than 200 or 65 different Cl − and Na + mutations, only about 80 and 30 different mutations, respectively, have been functionally characterized (http://www.hgmd.cf.ac.uk/ac/ index.php). However, these functional analyses have contributed to understand in a better way the recessive or dominant behavior of different CLCN1 mutation than the overlapping phenotypes and clinical variability shown by some specific mutation (39,67,102,103). For instance, in general, it has been reported that several recessively inherited CLCN1 mutations show biophysical defects like reduced open probability, reduced single-channel conductance, or biochemical instability, in a manner that does impinge in a significant manner on the formation or function of heteromeric mutants/WT channels (13,106,107). A simple example are early stop codon mutations which do not result in the expression of ClC-1 subunits (108). In this regard, it has to be remembered that to provoke myotonia, G Cl has to be lowered below roughly 30% (109)(110)(111)(112). On the other hand, several dominantly inherited CLCN1 mutations have been shown to exert in vitro a dominant negative effect in co-expression with WT (39,58,64,107,113,114). In many cases the dominant negative effect is mediated by a shift of the open probability of the "common gate" to more positive voltages (107,113,114). These properties provide a rational to explain the dominant inheritance. However, this does not explain the clinical variability observed in both RMC and DMC. Interestingly, there is a group of 12 CLCN1 mutations (see Table 2) in which the functional in vitro analyses have not been able to demonstrate differences with the WT channel, suggesting the presence of additional factors in skeletal muscle fibers, not present in oocytes/HEK cells, that are involved in the disease. Such factors could be related, for example, to subcellular targeting, a still open question for ClC-1 (18,19). For these mutations, a powerful experimental approach would be the generation and analysis of knock-in mice carrying the myotonia-related CLCN1 mutations, since myotonic mice are an established model for chloride channel myotonia (13,115). Such knock-in mice could provide important information that cannot be obtained in heterologous expression systems. Additionally, patient derived induced pluripotent stem cells (iPSCs), combined with suitable myogenic differentiation (116), or the use of CRISPR/Cas approaches in order to correct (genome editing) specific mutations (followed by evaluating the off-target activities of CRISPR/Cas systems), might provide future avenues of studying the impact or correction of ion-channel mutations. However, in particular for CLCN1, these approaches are complicated by the fact that CLCN1 expression (in vivo) requires fully differentiated and innervated muscle fibers (117). Importantly, many of these approaches also apply for the study of sodium channel mutations (see below).
Clinical correlations in sodium channel myotonic disorders are not as complex as the chloride channel myotonic disorders. As with MC, functional analyses in Na + myotonic channelopathies have contributed to the understanding of the pathophysiological mechanism and improvement of clinical correlation, which are more accurate in this case. These analyses have been able to provide that Na + channel mutations have four major effects on Nav1.4 function: (1) enhanced activation; (2) impaired slow inactivation; (3) impaired fast inactivation; and (4) accelerated recovery from fast inactivation. But, as with the CLCN1 mutations, while these effects explain the dominant behavior of the mutations (dominant gain-of-function effect), they contribute little to the explanation of the clinical variability seen in this group of NDM. Nevertheless, it is well-accepted that: (1) a large fraction of persistent current and an incomplete slow inactivation of the channel may cause a strong long-lasting depolarization, providing the bases for weakness in HyperPP; (2) a slowing of fast inactivation and an incomplete closure of the channel may explain the paradoxical myotonia characteristic of PC; and (3) an increased persistent fraction and/or slowing of fast inactivation might explain the slight depolarization that causes myotonia in affected patients with SCM [reviewed in (47)]. But this does not explain the clinical variation observed in different patients, or even the overlapping symptoms seen with MC. In the case of SCN4A mutations, to the best of our knowledge, there have not been reported mutations that, functionally, behave as the WT channel. Interestingly, there have been few patients carrying a SCN4A mutation who have been shown to carry a second mutation in CLCN1 (see Table 3). These patients have shown an exacerbated or atypical Na + channel disease phenotype, suggesting that both mutation may act synergistically to influence the clinical and neurophysiological phenotype observed in those patients (130). For many years, it was common practice to screen just one among CLCN1 or SCN4A genes, based on the presumptive clinical phenotype. It is thus likely that a significant group of the first (or even some of recent cases) NMDs cases (with atypical/unclear phenotype, unclear inheritance pattern or dual inheritance-dominant or recessive) with molecular diagnosis, carry a second mutation in any of this ion channel genes, and, maybe even in other loci. Based on this and suggested by many authors, the recommendation is that in those NDM cases with atypical phenotypes or inheritance pattern, both CLCN1 and SCN4A genes are screened. However, thanks to new technologies and price reduction, it might now be possible to carry out deep sequencing, such as whole genome sequencing or genetargeted sequencing (by next generation sequencing-NGS), in these patients. These new approaches not only would allow to properly screen these genes and genotype NDM patients, but also, could contribute to identify gene modifiers that might be involved in modulating their phenotypes. In fact, very recently it was published the first report using whole genome sequencing for screening the CLCN1 gene in RMC patients (131). These  new approaches could be very useful to screen those patients with atypical phenotypes and those in which the effect of the identified mutation does not differ from the wild-type channel (see below). The molecular bases for the clinical variability present in myotonic channelopathies remain obscure. Yet, a deeper understanding is needed to improve genotype-phenotype correlations. Several hypotheses have been put forward to explain the clinical variability, which include: (a) differential allelic expression; (b) trans-acting genetic modifiers; (c) epigenetic, hormonal or environmental factors; and (d) dominance with low penetrance (20,47,(136)(137)(138). These putatively acting (alone or in combination) mechanisms might contribute to explain why the same mutation causes different degrees of channel dysfunction in different patients (74), and therefore elicit a modulation in the NDM phenotype. Importantly, to the best of our knowledge, none of these possibilities has been established experimentally in NMD patients or are under current investigation.
The above considerations suggest that it is not only important to study as many individuals or mutations as possible, but also to try to obtain functional data for all mutations. In addition, deeper insight is expected from bioinformatics and structural approaches combined with information on nearby mutations, aided by the recently obtained 3D structure of the ClC-1 protein (139,140). The results of these studies, in addition to explaining in a better way the different symptoms associated with particular mutations, would contribute to improving clinical correlations. Improvements in clinical tests, the recognition of the different phenotypes that result from particular mutations and the understanding of how a mutation affects the structure and function of the ion channel, together with genetic screening, is expected to improve clinical correlation in NDMs, which could ultimately be translated into better clinical management and to a better quality of life of affected patients and their families.
Finally, a solid knowledge of functional defects caused by specific mutations might help in the development of selective drugs targeted at a correction of such effects. For example, gating modifier drugs could be developed to invert the shift of the voltage-dependence of some dominant CLCN1 mutations. Several small molecule compounds are indeed known to interfere with the open probability of ClC-1 channels, suggesting the principal feasibility of such an approach (141,142). In particular, derivatives of clofibric acid have been shown to shift the voltagedependence of ClC-1 and the related ClC-0 to more positive voltages (141)(142)(143). The development of novel molecules that specifically bind to and stabilize the open state of the channel could be aided by the recent determination of the structure of ClC-1 (139,140). Another possible line of intervention regards patients with mutations that lead to protein folding defects. Such defects might be treated with small molecules that act as "correctors." Such an approach is already in clinical use in other diseases, such as cystic fibrosis (144). Since correctors are assumed to stabilize the folded state of the channel, it might be tempting to start the search of correctors using ClC-1 inhibitors, like 9-anthrazene carboxylic acid (9AC) (143). However, all known ClC-1 inhibitors are of rather low affinity and quite unspecific (8). Therefore, such correctors are likely only to be discovered using high-throughput screening, as done for the most common CFTR mutation causing cystic fibrosis (118). Indeed, generic correctors, such as 4-phenylbutyrate, miglustat, and sildenafil have proven unsuccessful for cystic fibrosis (118).
For preparing this review, we (both authors) have used our personal historical literature database of publications regarding CLCN1 and SCN4A related myotonia, as well as PubMed searches using the keywords "myotonia AND (CLCN1/CLC-1, SCN4A/Nav1.4, apnea, laryngospasm)."

AUTHOR CONTRIBUTIONS
All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.