Objective Gain-of-function mutations in Nav1.9 have been identified in three families with rare heritable pain disorders, and in patients with painful small-fibre neuropathy. Identification and functional assessment of new Nav1.9 mutations will help to elucidate the phenotypic spectrum of Nav1.9 channelopathies.
Methods Patients from a large family with early-onset pain symptoms were evaluated by clinical examination and genomic screening for mutations in SCN9A and SCN11A. Electrophysiological recordings and multistate modelling analysis were implemented for functional analyses.
Results A novel Nav1.9 mutation, p.Arg222His, was identified in patients with early-onset pain in distal extremities including joints and gastrointestinal disturbances, but was absent from an asymptomatic blood relative. This mutation alters channel structure by substituting the highly conserved first arginine residue in transmembrane segment 4 (domain 1), the voltage sensor, with histidine. Voltage-clamp recordings demonstrate a hyperpolarising shift and acceleration of activation of the p.Arg222His mutant channel, which make it easier to open the channel. When expressed in dorsal root ganglion neurons, mutant p.Arg222His channels increase excitability via a depolarisation of resting potential and increased evoked firing.
Conclusions This study expands the spectrum of heritable pain disorders linked to gain-of-function mutations in Nav1.9, strengthening human validation of this channel as a potential therapeutic target for pain.
Statistics from Altmetric.com
Voltage-gated sodium channels (Navs) play essential roles in electrogenesis within the nervous system. Nav1.9 produces a tetrodotoxin-resistant (TTX-R) persistent sodium current characterised by hyperpolarised activation compared with other Navs, resulting in a large window current that has a depolarising effect around the resting potential of dorsal root ganglia (DRG) neurons.1 ,2 The ultra-slow kinetics and large window current suggest that Nav1.9 channels act as threshold channels,1 ,3 lowering the threshold for single action potentials and increasing repetitive firing.4
Nav1.9 channels are preferentially expressed in small-diameter (<30 μm diameter) somatosensory neurons of DRG and trigeminal ganglia5 and in intrinsic myenteric neurons.6 Within cutaneous afferents, Nav1.9 is expressed in functionally identified nociceptors.7 ,8 Although Nav1.9 is predominantly expressed in non-peptidergic cutaneous DRG neurons,9 ,10 it has recently been reported in peptidergic visceral DRG neurons innervating the colon.11 The distribution of Nav1.9 in cutaneous, visceral and myenteric neurons is consistent with its contribution to nociception and to regulation of gastrointestinal functions.
There is genetic support for a role of Nav1.9 in nociception and gastrointestinal function in humans. Mutations in Nav1.9 were described in two families with pain in distal extremities12 and in one family with cold-aggravated painful symptoms.13 Sporadic mutations of Nav1.9 have been identified in patients with painful small-fibre neuropathy.14 ,15 Paradoxically, one gain-of-function mutation in Nav1.9 has been reported in patients with congenital insensitivity to pain (CIP), albeit without a clear mechanistic basis for the reduced excitability of sensory neurons.16 ,17 An additional Nav1.9 mutation has been reported in a case of familial CIP associated with chronic diarrhoea;18 however, functional analysis of this mutation was not reported. The expression of Nav1.9 in nociceptors and myenteric neurons is consistent with the symptoms of distal extremity pain and gastrointestinal disturbances reported by patients carrying Nav1.9 mutations.12 ,14 ,16–18
Although a spectrum of pain disorders have been linked to Nav1.9 channelopathies, additional studies are needed to unmask the molecular and cellular bases of Nav1.9-related symptoms. We report here the identification and characterisation of a new Nav1.9 mutation, p.Arg222His, in a family with early-onset pain symptoms in distal extremities including joints, and gastrointestinal disturbances. This study expands the spectrum of heritable genetic disorders linked to gain-of-function mutations in Nav1.9.
Detailed methods are available in online supplementary files.
The study was approved by the institutional review board of Fundación para el Estudio de las Enfermedades Neurometabólicas. Informed consent was obtained prior to the initiation of the study. The index case (case IV.2 figure 1) is an adult female who presented with a history of lifelong intense pain episodes localised primarily in the distal extremities, especially in the joints of fingers, wrists and ankles. DNA was isolated from peripheral blood, and SCN9A and SCN11A exon screening was performed using methods described previously.19
The green fluorescent protein (GFP)-2A-tagged wild-type (WT) construct (pcDNA3-Nav1.9, WT) was previously described.14 The p.Arg222His mutation (referred to as R222H within figures) was introduced using QuikChange II XL site-directed mutagenesis (Stratagene) and confirmed by Sanger sequencing.
Superior cervical ganglion neuron isolation and transfection
Superior cervical ganglion (SCG) neurons which do not express endogenous TTX-R currents, but where Nav1.9 expresses at high levels, were used as an expression platform.15 Protocols for care and use of animals were approved by the Institutional Animal Care and Use Committee (IACUC) at the Veterans Administration Connecticut Healthcare System, West Haven. SCG neurons were isolated from neonatal (birth to 5 days old, gender not determined) Sprague-Dawley rats and transfected by electroporation as described previously.15
DRG neuron isolation and transfection
Voltage-clamp recordings at 22±1°C, from small SCG neurons (<25 µm) within 40–48 hours after transfection, were used to assess effects of the mutation on function of the Nav1.9 channel. Current-clamp recordings obtained at 22±1°C from small DRG neurons (<30 µm) 40–48 hours after transfection as previously described21 allowed us to study effects of the mutant channels on the firing properties of DRG neurons.
Multistate structural modelling
Electrophysiological data were analysed using Fitmaster (HEKA Elektronik) and Origin 8.5.1 (Microcal) and presented as mean±SE. Statistical significance was determined by Student's t-tests (current-clamp except firing frequency and spontaneous activity), Mann-Whitney test (firing frequency) or two-proportion z-test (comparison of proportion of cells producing spontaneous activity).
Case history and mutation detection
An adult female proband (subject IV. 2, figure 1A) presented with a history of lifelong, intense episodic pain in distal extremities, especially in the joints of the fingers, wrists and ankles. From the age of 1–15 years, she experienced frequent pain episodes (2–8 times per day), each lasting 5–40 min. She did not report rectal, ocular or submaxillary pain. There were no clear triggers. During these pain episodes she felt an urge to defecate, and frequently experienced diarrhoea. No changes in skin colour or sweating accompanied pain episodes. Her daughter (subject V. 4, figure 1A) presented with more frequent and more severe symptoms. Between the ages of 2 and 9 years she experienced >15 pain episodes per day. A second daughter (subject V. 5, figure 1A) reported less intense symptoms. Lamotrigine (100 mg, twice a day) resulted in 70–80% relief of pain in both daughters (V. 4 and V. 5).
Family history revealed a large family with an autosomal dominant pain disorder (figure 1A). All affected members had a normal neurological examination, including normal sensitivity to light touch, pinprick, vibration sense and joint position sense. Pain attacks were occasionally triggered by high ambient temperatures, but for most attacks no trigger could be identified. Pain severity in affected individuals diminished with age (after age of 16–18 years). All of the affected family members who provided DNA were symptomatic at the time of DNA collection. Some affected family members obtained relief by cooling their hands, while others experienced relief by applying pressure (wearing tight gloves). Ibuprofen and diclofenac did not provide relief.
Although the proband was first clinically diagnosed as having paroxysmal extreme pain disorder,24 sequencing all exons and flanking intronic sequences of the SCN9A gene revealed no mutations. Since the symptoms resembled familial episodic pain,12 the exons and intronic flanking sequences of the SCN11A gene were sequenced, demonstrating a mutation c.665G>A (NM_014139.2) in exon 5 of the SCN11A gene; this mutation causes the substitution of arginine 222 with histidine (p.Arg222His). This mutation was not found in exome sequencing databases (ESP and ExAc) and was predicted to be pathogenic by PolyPhen (disease causing; score: 0.999), MutationTaster (disease causing; p value: 1), SIFT (deleterious; score: 0) and Align GVGD (class C25; GV: 0.00—GD: 28.82). Sequencing of four additional symptomatic family members (IV. 1, IV. 4, V. 8 and V. 5; figure 1A) revealed the presence of the c.665G>A mutation, while an asymptomatic first cousin (V. 2; figure 1A) did not carry the mutation. Taken together, these results support the probable pathogenicity of the mutation.25
Multistate modelling provides a basis for understanding structural changes that underlie altered function of mutant channels. Arg222 is the positively charged residue in the first triad repeat R/KXX in transmembrane segment 4 of domain I (DI/S4), which acts as the voltage sensor for that domain (figure 1B), and is designated as R1. Sequence alignment shows an invariant arginine at this position in all human Navs. Substitution of arginine (positively charged at physiological pH) with histidine (predominantly neutral at physiological pH) in the voltage sensor of domain 1 (VSD1) is predicted to alter gating properties of the mutant channel.
As several negatively charged residues are present in the S2–S3 helices of Nav1.9 VSD1, we hypothesised that Arg222 might form ionic interaction with these residues in specific gating states, and that the mutation p.Arg222His might disrupt these ionic interactions, thus affecting channel gating.
To test this hypothesis, we employed multistate modelling, a tool that permits analysis of changes of transmembrane helices during channel gating and interactions between residues as the channel transits from closed to activated states.22 ,23 Nav1.9 VSD1 carries negatively charged residues that might interact with R1, including Glu163 and Glu173 of the S2 helix; Asp193 and Asp199 of the S3 helix. WT and mutant VSD1 of Nav1.9 were analysed in six gating states: (1) activated state; (2) early loss-of-activation state; (3) late loss-of-activation state; (4) resting/closed state; (5) early activation state; and (6) late activation state (figure 2).
To investigate possible interaction between Arg222 and other residues, we used Arg222 as a ‘ligand’ to probe surrounding residues as ‘receptors’.22 ,23 By plotting the ligand–receptor interaction, we constructed an interaction map between Arg222 and surrounding residues (figure 2). We observed strong ionic interactions of Arg222 with the negatively charged residues in four of the six gating states (except the full activated state and late activation state). In particular, Arg222 interacts with Glu163 in the early loss-of-activation state. In the late loss-of-activation state, Arg222 interacts with another negatively charged residue, Glu173. In the resting state, Arg222 interacts with two residues, Glu173 and Asp199. When the channel enters early activation state, Arg222 maintains its interaction with Asp199. However, after transit into late activation state and fully activated states, Arg222 no longer has any interaction with negatively charged residues (figure 2). Taken together, the structural modelling results predict that the mutation destabilises the closed state of the channels, thereby making it easier to open.
We used voltage-clamp recordings to test the prediction from multistate modelling that the p.Arg222His substitution in VSD1 alters gating properties of the Nav1.9 channel. Figure 3A, B shows representative Nav1.9 current traces recorded from SCG neurons expressing WT and p.Arg222His mutant channels, respectively. Although current density of p.Arg222His channels (54.2±9.8 pA/pF; n=12) was smaller than that of WT channels (74.5±11.7 pA/pF; n=15), the difference did not reach statistical significance. Figure 3C shows activation voltage dependence for WT and p.Arg222His channels. Compared with WT channels (−47.5±1.7 mV, n=15), the activation midpoint for p.Arg222His channels (−53.9±2.3 mV, n=12) was shifted by −6.4 mV (p<0.05). The activation slope factor of p.Arg222His channel (10.1±0.5 mV, n=12) was also significantly larger than for WT channels (8.2±0.4 mV, n=15, p<0.01). Figure 3D displays the steady-state fast inactivation curves of WT and p.Arg222His channels. Although fast inactivation midpoint for p.Arg222His was not significantly different from that of WT channels (WT: −46.7±1.5 mV, n=11; p.Arg222His: −45.6±2.1 mV, n=10), the fast inactivation slope factor for p.Arg222His channels (11.3±0.6 mV, n=10) was significantly larger than for WT channels (9.5±0.6 mV, n=11, p<0.05). The activation time-constants of p.Arg222His channels (n=12) were significantly smaller than those of WT channels (n=15, p<0.05) between −60 and 20 mV (figure 3E). However, inactivation time constants are not significantly different between WT (n=15) and p.Arg222His channels (n=12) in the voltage range from −25 to 20 mV (figure 3F). Deactivation time constants for WT (n=13) and p.Arg222His channels (n=10) are not significantly different in the voltage range from −110 to −60 mV (figure 3G). Steady-state slow inactivation midpoint for p.Arg222His channels (−84.1±3.2 mV, n=5) was not significantly different from that of WT channels (−82.9±2.4 mV, n=10; figure 3H). However, the slow inactivation slope factor of p.Arg222His channels (8.7±0.6 mV, n=5) was significantly larger than that of WT channels (6.8±0.3 mV, n=10, p<0.05). We also compared the response to slow ramp stimulation (which mimics the effects of subtle natural stimuli) between WT and p.Arg222His mutant channels. Although WT and p.Arg222His mutant channels produced large ramp currents, the normalised average peak ramp currents for p.Arg222His channels (53.3±3.1%, n=11) were smaller than that for WT channels (71.4±2.7%, n=13, p<0.001). Despite the hyperpolarising shift in V1/2 of activation of p.Arg222His channels, the voltage at which the peak of the ramp current occurs was not significantly shifted (WT: −43.2±1.8 mV, n=13; p.Arg222His: −44.1±1.8 mV, n=11). In the aggregate, these voltage-clamp results demonstrate enhanced and accelerated activation of mutant p.Arg222His channels, both proexcitatory Nav1.9 gating changes.
To assess the effects of p.Arg222His mutant channels on the firing properties of DRG neurons, we studied excitability of these neurons using current-clamp. The resting membrane potential of neurons that expressed p.Arg222His channels (−48.3±1.1 mV, n=27) was significantly depolarised (∼4 mV), compared with neurons expressing WT channels (−52.2±1.1 mV, n=23, p<0.05; figure 4C). Current threshold, assessed by injecting a series of currents in increments of 5 pA over 200 ms, was significantly smaller (174±17 pA, n=27, p<0.001) for neurons expressing p.Arg222His channels compared with neurons expressing WT (302±22 pA, n=23; figure 4D). However, there were no significant differences between cells expressing WT and p.Arg222His channels in input resistance (WT: 414±31 MΩ, n=23; p.Arg222His: 578±78 MΩ, n=27), action potential width at 0 mV (WT: 3.06±0.27 ms, n=23; p.Arg222His: 3.62±0.25 ms, n=27), voltage threshold (WT: −16.3±1.0 mV, n=23; p.Arg222His: −16.3±0.8 mV, n=27) or after-hyperpolarisation potential (WT: −68.3±1.7 mV, n=23; p.Arg222His: −66.0±1.7 mV, n=27). Although action potential amplitude of DRG neurons expressing p.Arg222His channels (98.2±0.9 mV, n=27) was significantly reduced compared with DRG neurons expressing WT channels (103.5±1.6 mV, n=23, p<0.01), the action potential overshoot was not significantly different (WT: 50.9±1.2 mV, n=23; p.Arg222His: 49.6±1.1 mV, n=27), consistent with the difference in the amplitude being due to the difference in resting potential.
We evaluated the effects of p.Arg222His mutant channels on repetitive firing of DRG neurons by injecting a series of 500 ms currents from 25 to 500 pA, in 25 pA increments. Figure 4A shows the responses of a representative neuron expressing WT or p.Arg222His mutant channels to 1X, 1.5X and 2X current threshold. A DRG neuron expressing WT channels generated a single spike in response to stimuli of 1X and 1.5X current threshold, and two spikes to injection of 2X current threshold. A representative DRG neuron which expressed p.Arg222His mutant channels produced two action potentials in response to stimuli at 1.5X and five action potentials at 2X current threshold. Data for the population of transfected DRG neurons demonstrate that p.Arg222His channels evoke significantly more action potentials compared with WT channels (figure 4B).
Figure 4E shows representative spontaneous firing recorded from a neuron expressing p.Arg222His mutant channels. The proportion of spontaneously firing DRG neurons expressing p.Arg222His channels was 51% (28 out of 55 cells), significantly larger for neurons expressing WT channels (18%; 5 out of 28 cells; p<0.01, z-test). These results demonstrate that mutant p.Arg222His channels make DRG neurons hyperexcitable.
While a large number of mutations in Nav1.7 have been linked to pain disorders in humans, fewer mutations in Nav1.9 have been described thus far.26 Interestingly, gain-of-function mutations in Nav1.9 have been linked to painful12–15 and painless channelopathies.16–18 We report here the identification and characterisation of a novel familial Nav1.9 mutation, p.Arg222His, causing a painful channelopathy. The substitution p.Arg222His within the DI/S4 segment of the channel significantly hyperpolarises channel activation. The expression of the p.Arg222His mutant channel in DRG neurons depolarised of resting potential by 4 mV, reduced threshold and increased frequency for evoked firing in these neurons, and increased the fraction of DRG neurons that fire spontaneously.
Nav1.9 is present all along the length of primary afferents from peripheral terminals in the skin, along the fibre shafts within the sciatic nerve and the central terminals within the dorsal horn, acting as a threshold channel which regulates firing of these afferents.26 Pain symptoms associated with the p.Arg222His mutation in Nav1.9 are consistent with hyperexcitability of DRG neurons that express mutant channels. Gastrointestinal symptoms reported by these patients might be explained by the effect of the mutation on visceral afferents that are known to express Nav1.9 channels.6 ,27 A specific role for Nav1.9 in regulating gastrointestinal function is supported by demonstrating that this channel is important for responses to mechanical stimulation and mechanical hypersensitivity of visceral afferents innervating the colon when challenged with inflammatory mediators identified in inflammatory bowel disorder or Crohn's disease.11 Symptoms in patients described here improved with age after adolescence; however, little is known about age-dependent changes in the clinical symptoms that accompany channelopathies, and the molecular/cellular correlates of this improvement are not known.
The p.Arg222His mutation enhances activation by significantly hyperpolarising channel activation by 6.4 mV, increasing the current of the mutant channel in the voltage domain defined by the overlap of activation and steady-state inactivation of the channel (window current), and displays accelerated activation. At the cellular level, expression of the p.Arg222His in DRG neurons depolarises resting potential of DRG neurons by 4 mV. We previously showed that depolarisation of resting potential of DRG neurons by ∼5 mV increases excitability.14 In agreement with our published findings, expression of p.Arg222His in DRG neurons reduced threshold and increased frequency for evoked firing, and increased the fraction of neurons that fire spontaneously, providing a biophysical basis for the pain reported in these patients.
We used multistate modelling to explore the structural changes within the mutant Nav1.9 channel that underlie its enhanced function. While our homology model is extrapolated from potassium channels, it provides insights about probable intramolecular structural changes underlying altered function of the mutant p.Arg222His channel because of the similarities of basic ion channel activation. Our results predict state-dependent ionic interactions of Arg222 with negatively charged residues (figure 2). Although it is difficult to predict the local pH of p.Arg222His in the folded channel and whether it changes during gating, it is reasonable to assume that, under physiological conditions, histidine residues are predominantly neutral (with a relatively small proportion of channels carrying positive charges at this position). As strong ionic interactions between Arg222 and negatively charged residues only occur in gating states other than late activation and fully activated states, Arg222 is likely to stabilise channel closed conformations, increasing the energy barrier to open the channel. Thus, p.Arg222His substitution makes the channel easier to open, thereby shifting channel activation in the hyperpolarising direction. This prediction is consistent with our observations showing that p.Arg222His hyperpolarises voltage-dependent activation of Nav1.9 channel by 6.4 mV. Another substitution at Arg222 (p.Arg222Cys) has been reported as a rare variant in a multiethnic population (1/120 806 alleles, ExAC database). Based on multistate modelling, a substitution of Arg222 by Cys is also predicted to destabilise the closed state of the channel; the degree of destabilisation (as reflected by hyperpolarised activation) and whether a pain phenotype might have been present in carriers of the p.Arg222Cys variant are not known.
Although the Nav1.9 p.Arg222His and p.Arg225Cys mutations are associated with familial episodic pain and expression of either mutant channel in DRG neurons renders these neurons hyperexcitable, the changes at the channel level leading to DRG neuron hyperexcitability appear to be different for these two mutations. p.Arg225Cys was reported to increase the current density of sodium channels, but the increase did not reach statistical significance and no other gating property was reported to be altered.12 As discussed above, p.Arg222His confers gain-of-function attributes on the Nav1.9 channel. The difference in the reported functional effects of the p.Arg222His mutation (this study) and the p.Arg225Cys mutation12 at the channel level is not well understood, but may reflect the use of different expression systems. Zhang et al12 used WT mouse DRG neurons to record the p.Arg225Cys current by voltage-clamp, while we used SCG neurons which do not produce endogenous TTX-R currents, permitting us to investigate the effect of the p.Arg222His mutation on Nav1.9 properties without contamination by endogenous currents. It is possible that endogenous TTX-R currents in WT mouse DRG neurons could have masked the effects of the p.Arg225Cys mutation on Nav1.9 in the Zhang et al12 study. Alternatively, a non-canonical effect of the p.Arg225Cys mutation, for example an effect on channel targeting or an indirect effect on other channels, may have contributed to hyperexcitability of DRG neurons that express this mutant channel. Additional experiments will be needed to distinguish between these alternative explanations.
Twelve Nav1.9 mutations have been linked to painful channelopathies, six of which have been functionally characterised12–15 (and this study). Although all patients with these six mutations showed painful symptoms, patients with p.Arg222His, p.Arg225Cys, p.Ala808Gly, or p.Val1184Ala mutations displayed symptoms in early childhood and the pain diminished with age; in contrast, patients with p.Ile381Thr, p.Gly699Arg or p.Leu1158Pro mutations started to experience pain in adulthood, and pain did not disappear with age at the last follow-up. All of these mutations are reported to significantly increase excitability of DRG neurons. However, the mutations have different effects on biophysical properties of the channel. p.Ile381Thr and p.Gly699Arg demonstrate enhanced activation and impaired fast inactivation, whereas p.Arg222His, p.Leu1158Pro and p.Val1184Ala hyperpolarise activation without a significant effect on fast inactivation. In contrast, p.Arg225Cys and p.Ala808Gly have no effects on channel gating properties. While there is a notable pattern of decreasing pain with age with the p.Arg222His, p.Arg225Cys, p.Ala808Gly and p.Val1184Ala mutations, there is no simple correlation between the channel phenotype and clinical phenotype.
In summary, we report the identification and characterisation of a novel Nav1.9 mutation from a multigenerational family. Pain in distal extremities and gastrointestinal disturbances that started at an early age subsided after adolescence. The gain-of-function changes in p.Arg222His mutant channels cause DRG neurons to become hyperexcitable, thus producing chronic pain. This mutation expands the repertoire of Nav1.9 channelopathies and provides a mechanistic explanation for the pain and gastrointestinal disturbances in human patients carrying this mutation.
While our study was under review, another study has reported Nav1.9 mutations pArg222His and pArg222Ser in two Japanese families. Okuda H et al PLoS One 11(5):e0154827.
The authors thank Palak Shah, Fadia Dib-Hajj, Peng Zhao and Lawrence Macala for technical assistance.
Contributors CH, YY, RHteM and JMP designed research, collected and analysed data, and wrote manuscript. JPHD, SGW and SDD-H designed and supervised project, analysed data, and wrote and edited manuscript.
Funding This work was supported in part by grants from the Rehabilitation Research Service and Medical Research Service, Department of Veterans Affairs (SDD-H, SGW). The Center for Neuroscience and Regeneration Research is a Collaboration of the Paralyzed Veterans of America with Yale University.
Competing interests None declared.
Patient consent Obtained.
Ethics approval The study was approved by the IRB of the Fundación para el Estudio de las Enfermedades Neurometabólicas, Buenos Aires, Argentina.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.