Exonic mutations in SCN9A (Na_V1.7) are found in a minority of patients with erythromelalgia

2.1 Patients’ inclusion/exclusion criteria

Over a period of five years, 350 patients were referred to the Section of Clinical Neurophysiology, Oslo University Hospital for neurophysiological evaluation of distal pain in their extremities (possible EM). The patients underwent further neurological and basic neurophysiological examinations, including QST, EMG, and neurography. The diagnosis of EM was based on the presence of episodic pain in the extremities, either occuring spontaneously or induced by warming, standing or exercise, warm and red feet during attacks and pain alleviated by cold. Patients were also included as having EM-like symptoms with either ongoing or episodic pain and with either red or warm feet. All patients were above 18 years and were asked about all previous and current diseases, the use of medication as well as characteristics of pain, and investigated clinically. The study was approved by the regional Ethical committee and all patients signed informed consents. Forty-eight patients fulfilled the clinical criteria for classical primary EM or EM-like symptoms. 37 out of 48 patients were from independent families and 11 patients originated from 5 small nuclear families, and were mother-daughter or brother-sister.

2.2 Quantitative sensory testing

Threshold temperatures for the sensation of warmth, cold, heat pain and cold pain were determined using a computerized Ther-motest (Somedic A/B, Sweden).

2.3 Polymerase chain reaction (PCR) and Sanger-sequencing

After having obtained ethics approval and informed consent, peripheral blood samples (10 ml) were taken from all the individuals and genomic DNA was isolated for genetic analysis. DNA from 96 healthy controls was from Cheshire County in UK.

PCR primers were designed using Oligo 5.0 software and amplified PCR fragments ranged from 300 to 900 bp in length covering the coding exons including the known splicing variants and at least 50 bp of exon-intron boundaries. Sanger sequencing data was collected using standard methods on an ABI3130xl sequencer, and sequence traces were analyzed after assembly using DNA Star v5.

2.4 Genomic capture and high throughput sequencing (next generation sequencing)

Sequencing of selected gene regions (Table 1) was performed using a custom SureSelect XT solution phase target enrichment kit (Agilent Technologies, Inc.), followed by sequencing via a HiSeq2000 (Illumina, Inc.). Approximately three micrograms of high MW DNA from each sample was prepared, captured and enriched according to the manufacturer’s protocols. Selected material was amplified to incorporate relevant Illumina adapter and index sequences and samples pooled and run as sets of 12 per HiSeq2000 channel in 100bp paired end mode and the resulting data demultiplexed by index into sample-specific read sets (FastQ files). All DNA capture, preparation and sequencing steps were performed by a 3rd party sequence service provider (Source: Bioscience, Nottingham, UK). Data analysis (mapping and variant calling) was primarily performed using CLCBio Genomics Workbench v4.6 (CLCBio, Aarhus, Denmark). Additional assessment of read set quality was performed using FastQC.

3.1 Genetics

In total, 29 genes were screened for mutation in this study including all the a subunits (SCN1A, 2A, 3A, 4A, 5A, 7A, 8A, 9A, 10A and 11A) and β subunits (SCN1B, 2B, 3B and 4B) of sodium channels, transient receptor potential channels (TRPA1, TRPV1, TRPM8), neurotrophic factors (NGF, NGFR, BDNF, GDNF, NTRK1) and other known pain-related genes (CACNG2, KCNS1, COMT, P2RX3, TAC1, TACR1 and WNK1) using next generation sequencing technique combined with classical Sanger sequencing (Table 1). Beside the mutations in Na_V 1.7, more than twenty rare genetic variants were identified in the genes coding for sodium channels (Na_V1.8, Na_V1.9, Na_V1.6, Na_V1.5, Na_V2.1, SCN1B, SCN3B), transient receptor potential channel (TRPA1, TRPV1), and for WNK1 and NGFR (Table 2A). No mutation was found in the genes coding for Na_V1.1, Na_V1.2, Na_V1.3, Na_V1.4, SCN2B, SCN4B, NGF, BDNF, GDNF, NTRK1, TRPM8, P2RX3, CACNG2, KCNS1, COMT, TAC1 and TACR1. All of the rare genetic variants identified in this study were further confirmed individually by classical Sanger sequencing and additionally, 96 healthy controls were screened routinely for all the variants.

The locations of the mutations in the various proteins are shown in Fig. 1. Among the α-subunits of voltage dependent sodium channels we found four mutations (P187L, I228M, I848T, N1245S) in Na_V1.7 (SCN9A) in 7 out of 48 patients with EM-like symptoms (Table 2B). The most common mutation associated with EM, I848T, is also common in our patients. Four patients who are from two independent families (mother and daughter) are heterozygous for the mutation. P187L is located in the extracellular loop between DIS1 and DIS2 and N1245S in the cytoplasmic loop between DIIIS2 and DIIIS3 (Fig. 1).

Fig. 1

Rare genetic variants in the genes identified in the patients. Locations of the mutations are marked with black dots.

Four rare genetic variants within the Na_V1.8 gene (E285K, M650K, L741V and V763I) were identified (Table 2A). E285K is located in the loop between DIS5 and D1S6, while the M650K is located in the cytoplasmic loop between domains I and II (Fig. 1).

L741V and V763I are situated in the transmembrane segments DIIS3 and DIIS4 which are extremely conserved (100%) between human, dog, mouse and rat (Figs. 1 and 2). The voltage sensors are formed by S1 to S4 in each domain which are crucial for sensing changes in membrane potential.

Fig. 2

Comparison of amino acid sequence between different species in the region where rare variants were found in Na_V1.8 and 1.9 (position highlighted in red).

Two rare genetic variants within the Na_V1.9 gene (N1169S, I1293V) were found (Table 1). N1169S is located in the small loop between DIIIS4 and DIIIS5 which is supposed to be important for the regulation of channel activity (Fig. 1). I1293V is in the transmembrane segment DIIIS6, a region for ion pore formation that is highly conserved between different species (human, dog, mouse and rat) and between all sodium channels (Fig. 2).

In total, four rare genetic variants, T220I, P648L, S1103Y and P2006A were found in Na_V1.5 with P648L and P2006A being de novo (Table 1, Fig. 1). Patients with mutations did not show any signs of cardiac defects.

Five rare genetic variants in the Na_V2.1 (V133I, Q236P, D803V, S1154Y and I1599V) were identified in this study and all these variants are neither present in public domain datasets nor in our 96 healthy controls (Table 2, Fig. 1).

Two rare variants, R125C in SCN1B and L10P in SCN3B, were found in two patients respectively (Table 2). The variant, R125C, may create a new position for the disulfide bridge and alter the secondary structure of the extracellular domain of SCN1B (Fig. 1). Mutation L10P is located at a highly conserved residue in the extracellular domain. This variant was also detected when 96 healthy controls were screened, but with an extremely low allele frequency (0.005, 1/192 chromosomes). No mutation within SCN2B and SCN4B was detected in this study.

3.2 Clinical characterization

We separated our patients according to results of the genetic analysis into those with mutations that have been proven to be linked to EM (marked in grey; n =4), mutations with a known gain of function in Na_V1.7 (marked in light grey; n =1), mutations in genes that have a close link to axonal excitability of nociceptors but for which functional characterizationis lacking (Na_V1.7, Na_V1.8, Na_V1.9, SCN1B, SCN3B; n =10), mutationsin genes that have aclose link to nociceptors (Na_V1.6, TRPA1, TRPV1; n =5), mutations in genes that have no close link to nociceptors (italics, Na_V1.5, Na_V2.1; n =7), and patients that show ednomutations in the analyzed genes (n =21) (Table 2B). The percentage of patients that reported incidence of similar symptoms among their relatives was higher in the I848T mutation patients (4/4) as compared to patients without evidence for mutations (9/21). The other groups of patients with mutations did not differ significantly from those without evidence for mutations in their family history (Table 2B).

While the age of onset was low (3–5 years) in the patients with the I848T mutations, no significant difference in age of pain debut was found in the other groups. Similarly, maximal intensity of pain (NRS-value) was highest in the I848T group, but did not differ significantly in the other patient groups. Pain descriptors did not contribute to differentiate the groups. In all groups, warmth and physical activity were the most prominent eliciting or worsening factors, without significant differences between the groups (see also Table 2).

3.3 Results of clinical examination

Two patients (7.4%) with mutations showed clinical signs of sen-sory neuropathy (reduced sensibility to light touch in distal parts of legs), but with normal findings on neurography. A total of seven patients (36.8%) without mutations displayed signsof sensory neuropathy (verified by sensory neurography in two of the patients, one more pronounced with sensory conduction velocities of 33 and 36 m/s, sural nerve bilaterally). Allodynia to light touch distally in the lower extremities was present in one of the patients in each group, whereas pin-prick hyperalgesia was demonstrated in seven patients with mutations and in four patients without mutations.

None of the patients displayed increased skin temperature at the time of investigation. The skin temperature in the dorsum of the feet did not differ significantly between the two groups (in the group with mutation: median 28.6°C, (27.7–30.4 (quartiles)) (in the group without mutation: median 29.6°C (28.0–30.7) (p = 0.34, Mann–Whitney U-test).

4.1 Variants of targets with proven link to EM

The degree of neuronal hyperexcitability induced by Na_V1.7 mutations correlates with earlier onset and more severe symptoms [7,10,26]. In addition to familiar cases of EM also de-novo mutations of Na_V1.7 have been identified in patients without family history [27,28]. In our study a larger cohort of patients diagnosed with primary erythromelalgia has been genotyped providing an estimate of the prevalence of Na_V1.7 mutations among EM patients. Among the 48 patients there were 5 with a gain of function mutation of Na_V1.7 that has already been functionally characterized. Four patients carried the I848T mutation known tocause a massive shift of activation threshold and increased ramp current. The proportion of 4 in 48 is close to 1 in 15 [17] or 1 in 13 [33] reported earlier. In one patient the I228M variant of Na_V1.7 was identified that has been found in healthy controls and epilepsy patients [46], but recently was also linked to small fibre neuropathy and neuropathic pain [21,23,42]. The occurrence of small fibre neuropathy in EM has already been described [40]. Basically, small fibre neuropathy might imitate symptoms of EM or represent an independent pathophysiological endpoint of hyperexcitability leading to EM and, in certain mutations also to neuropathy. The gain of function mutations in Na_V1.7 induced by the I228M mutation suggests overlapping pathophysiologic mechanism between EM and small fibre neuropathy, i.e. increased excitability underlies pain and increased sodium influx per action potential might – after sodium-calcium exchange – cause axonal degeneration [42]. The clinical picture concerning severity of small fibre neuropathy vs. intensity of chronic pain varies between carriers of I228M variant with either dominance of pain, dominance of neuropathy or even absence of clinical symptoms [21]. Our patient carrying the I228M variant had typical painful, red and hot feet, but thepain relief upon cooling was mild and the pain was reported as constant rather than in attacks – this ambiguity of clinical symptoms might reflect the “in-between” type of hyperexcitability generated by this variant. Interestingly, a similar overlap of clinical symptoms and sodium channel characteristics of the mutated channel has also been described between EM and PEPD [19,20].

Two more variants of Na_V1.7, N1245S and P187L, were identified in our patients. N1245S has been reported before as rare variant (minor allele frequency: 0.0041; NCBI, SNP database) whereas the P187L mutation has not been described yet. In the absence of functional characterization, the clinical relevance of these Na_V1.7 variants remains unclear.

4.2 Variants of targets with probable link to axonal excitability

In contrast to earlier attempts [17] we found rare variants of Na_V1.8 and Na_V1.9 in our EM patients. Na_V1.8 is found in small-diameter C-fibre associated sensory neurons [8], but is also expressed in non-nociceptive primary afferent fibres [45] and in intracardiac neurons [25,48]. Na_V1.8 has been shown to have a specific role in the detection of noxious thermal, mechanical and inflammatory stimuli [1]. Recently, several reports indicate that SNP rs6795970 (A1073V) within the DII/III intracellular loop of, is associated with altered electrophysiological properties and abnormalities of cardiac conduction [30]. Most interestingly, among 104 patients with painful small fibre neuropathy nine were identified carrying mutations of Na_V1.8 with 3 of those mutations (Leu554Pro, Ala1304Thr, Cys1523Tyr) proven to result inagain offunction [24]. One of those patients also showed paroxysmal reddening of the soles; however, some pain relief was achieved by warming. Thus, none of these patients would have been diagnosed as EM.

We identified the rare variants L741V, M650K, and E285K of Na_V1.8inour patients. However, all of them have reported before as rarevariant (unknown minor allele frequency; NCBI, SNP database). Therefore, the variants found in our patient could be either unas-sociated polymorphisms or the disease penetration of the variants is low.

Na_V1.9 contributes to the persistent tetrodotoxin-resistant current in small diameter DRG neurons and to the persistent thermal hypersensitivity and spontaneous pain behaviour following the exposure to inflammatory mediators (inflammation-induced hyperalgesia). In our patients the rare variants N1169S and I1293V were identified, with N1169S, but not I1293V having been reported as SNP before (unknown minor allele frequency; NCBI, SNP database).

The sodium channel a subunits determine the basic structure of the channel, while β-subunits modulate its properties. Genetic defects in the β subunits were reported to be associated with generalized epilepsy, with febrile seizures plus (SCN1B, [50]), cardiac conduction defect and Brugada syndrome 5 (SCN1B, [52]), Bru-gada syndrome 7 (SCN3B, [31]) and long QT syndrome 10 (SCN4B, [39] ). The R125C variant in the β1 subunit has been associated with epilepsy [41]. Albeit the β1 subunit can modify gating properties of Na_V1.7 and Na_V1.8 [38,49] its clinical role appears to be more obvious for epilepsy and cardiac arrhythmia [6]. In contrast, the β3 subunit is expressed mainly in small and medium size sensory neurons [6] and neuropathic pain models lead to its upregulation [12,44]. Despite these links to nociceptive fibres there are yet only reports of mutated β3 subunits leading to ventricular fibrillation via interaction with Na_V1.5 [32], but none that would be linked to chronic pain states.

4.3 Variants of targets with possible links to axonal hyperexcitability

Na_V1.6 is mainly expressed in myelinated afferents and has been shown to be up-regulated in neuropathic pain models in rodents [43]. Moreover, siRNA knock down of Na_V1.6 reduces mechanical hypersensitivity and bursting of primary afferents following inflammation of the dorsal root ganglion [55]. Mechanistically, resurgent currents leading to repetitive firing have been suggested to mediate amplification of afferent information including pain [34,47]. The rare variant I1583T of Na_V1.6 we found in our patient has not been reported as SNP before (NCBI, SNP database), however there is no information about potential functional consequences.

The transient receptor potential (TRP) channel family has been shown to play a dominant role with regard to thermo-and chemo-sensitivity [2]. Transient receptor potential A1 (TRPA1) is expressed by primary afferent sensory neurons of the pain pathway where it functions as a sensor for environmental and endogenous chemical irritants and contributes to inflammatory pain [2]. E179K (rs920829), a common polymorphism within the 4th ankyrin repeat of TRPA1, was shown to be associated with the somatosen-sory function in neuropathic pain patients [3], and a gain of function mutation N855S in the 4th transmembrane segment was confirmed to be the gene responsible for family episodic pain syndrome [35]. In our patients the variants N40S, R434C, and C608X were not reported as polymorphism (NCBI, SNP database), but there is also no functional data available to further judge their potential clinical relevance.

Transient receptor potential cation channel, subfamily V, member 1 (TRPV1) is a nonselective cation channel directly activated by harmful heat, extracellular protons, and vanilloid compounds. Elec-trophysiological results in TRPV1-nullmice (TRPV1–/–) indicate that TRPV1 is essential for selective modalities of pain sensation and for tissue injury-induced thermal hyperalgesia (temperature by generating inward membrane currents [5]). The variants found in our patients Q498X and F305S were not reported as polymorphism (NCBI, SNP database), but again no functional data are available.

4.4 Variants of targets with possible links to axonal hyperexcitability

Na_V1.5 is expressed exclusively in human heart and is responsible for the initial upstroke of action potential. Mutations in the SCN5A cause susceptibility to cardiac arrhythmias and sudden death in the long QT syndrome (LQT3, [51]). High incidence of abdominal pain in families with SCN5 A related cardiac channelopathy was observed and has led to the hypothesis that Na_V1.5 may contribute to the pathogenesis of functional bowel disorder [36]. Thus, the variants of Na_V1.5 found in our patients are most probably not related to their EM pain phenotype.

Sodium channel Na_V2.1 coded for by SCN6A is widely expressed and can bedetected in lung, heart, dorsal root ganglia and Schwann cells in the peripheral nervous system as well as in CNS. No mutation in Na_V2.1 was reported to be associated with human disease phenotype. Serine/threonine-protein kinase WNK1 plays an important role in regulating electrolyte homeostasis. In particular, it modulates sodium- and potassium-coupled chloride cotrans-porters. Mutations in WNK1 have been linked to higher sensitivity to thermal stimuli including cold pain [37], but not yet to human disease. Overall, the variants of Na_V2.1 and WNK1 therefore most probably represent polymorphisms without direct link to the pain phenotype of our patients.