Neuropathy-Associated NaV1.7 Variant I228M Impairs Integrity of Dorsal Root Ganglion Neuron Axons*

An underlying cause cannot be identified in a substantial proportion of cases of SFN, which are traditionally classified as idiopathic.¹^–³ Small-diameter peripheral axons and IENF are known to express voltage-gated sodium channel Na_V1.7.⁵ Faber et al. recently identified gain-of-function Na_V1.7 variants (single amino acid substitutions) in ~30% of patients with biopsy-confirmed idiopathic SFN.⁶ Patch-clamp studies of these Na_V1.7 variant channels demonstrated altered biophysical properties, resulting in gain of function at the channel level and in spontaneous firing and increased evoked firing frequency of dorsal root ganglion (DRG) neurons, which contribute to spontaneous and evoked pain. However, less is known about the molecular substrates for axonal injury and loss of IENF in idiopathic SFN.

IENF are known to express the Na-Ca exchanger-2 (NCX2) in addition to Na_V1.6, Na_V1.7, Na_V1.8, and Na_V1.9.⁵ Increased Na⁺ influx into axons expressing variant Na_V1.7 channels might be expected to trigger calcium influx into IENF and their small-diameter parent axons via reverse (Ca²⁺-importing) Na-Ca exchange, particularly in view of their short length constant and high surface-to-volume ratio.⁷^,⁸ Here, we developed a tissue culture model in which we could assess the effect of SFN-associated variant Na_V1.7 channels on DRG neurons in vitro. We demonstrate decreased length of DRG neurites expressing variant I228M channels. We also show that a use-dependent sodium channel blocker, and blocking of reverse (Ca²⁺-importing) operation of the Na-Ca exchanger, are protective. These observations suggest that gain-of-function variants of Na_V1.7 associated with SFN can contribute to impaired regeneration and/or degeneration of sensory axons through a cascade involving sodium channel activity and reverse Na-Ca exchange.

Results

SFN-Associated Na_V1.7 Variants Reduce the Length of DRG Neurites in Vitro

To examine the effect of SFN-associated Na_V1.7 channel variants on neurites of primary sensory neurons, we transfected DRG neurons in vitro with Na_V1.7 wild-type (WT) and variants I228M,⁹ M932L/V991L (ML/VL),⁶ and I720K⁶ (cotransfected with green fluorescent protein [GFP] to enable identification of transfected cells). As demonstrated in a representative 10 × 10 field-of-view montage image, cultures contained numerous GFP-positive neurons 3 days after transfection, with robust GFP signal in cell bodies as well as neurites (figure 1A). Examples of neurons transfected with WT, I228M, ML/VL, and I720K are shown at increased magnification in figure 1B.

Figure 1 Neurite length of neurons expressing Na_V1.7 wild-type (WT) and I228M, M932L/V991L (ML/VL), and I720K channels. (A) Large-field montage image consisting of a 10 × 10 field-of-view montage image of a dorsal root ganglion culture 3 days after transfection with Na_V1.7 WT + green fluorescent protein (GFP) constructs, with GFP signal as white. Dotted lines distinguish individual field-of-view captures. Scale bar: 1,000µM. (B) Increased magnification of individual neurons transfected with Na_V1.7 WT, I228M, ML/VL, and I720K constructs demonstrates reduced neurite length of I228M-transfected neuron compared to WT. Scale bar: 250µM. (C) Quantifications of the total neurite length/neuron calculated from large-field images and averaged for each condition. Pairwise comparisons between neurites from neurons expressing WT channels and channel variants I228M, ML/VL, and I720K are presented. Data are normalized to WT values and presented as mean ± standard error of the mean. *p < 0.05.

Mean total neurite length/neuron was quantified from large-field images for WT-, I228M-, ML/VL-, and I720K-transfected neurons. There was an ~20% reduction (p < 0.05) in length of neurites of I228M-expressing neurons as compared to those transfected with WT channels (see figure 1C; WT: 1,483 neurons from n = 25 large-field images; I228M: 1,436 neurons from n = 25 large-field images). ML/VL- and I720K-transfected neurons displayed a trend toward reduced neuritic length of 7% and 6%, respectively, which did not reach statistical significance (see figure 1C; ML/VL: 2,333 neurons, n = 30; WT: 2,394 neurons, n = 29; I720K: 1,522 neurons, n = 24; WT: 1,285 cells, n = 24).

I228M Channels Do Not Induce DRG Neuron Death

To determine whether I228M channels induce cell death of DRG neurons in addition to reducing neurite length, we assessed neuron viability 3 days after transfection with WT or I228M, using ethidium-homodimer 1 as a marker for dead/dying cells.¹⁰ Within the population of transfected neurons (identified with GFP signal 1 day post-transfection), 99% of WT-channel–expressing neurons and 98% of I228M-expressing neurons remained viable in cultures at 3 days after transfection (figure 2), demonstrating that neurons expressing I228M channels in this in vitro model are not preferentially susceptible to cell death at a time (3 days in culture) when neurite length is significantly reduced.

Figure 2 Cell viability of neurons 3 days after transfection with Na_V1.7 wild-type (WT) or I228M channels. Top left: Neurons 3 days after transfection with I228M + green fluorescent protein (GFP) constructs. Top right: Same field after incubation with EthD-1, a marker for dead or dying cells. Lower left: Merged image shows that GFP-positive cells do not colabel with EthD-1. Lower right: Quantification of viable cells at 3 days after transfection (cells positive for GFP and negative for EthD-1 are considered viable). Data are presented as mean ± standard error of the mean, where n = number of large-field images, and number of cells analyzed is indicated in parentheses. Scale bar: 100µM.

Sodium Channel Blocker Carbamazepine Protects Neurites Expressing I228M and Has No Effect on WT-Containing Neurites

To determine whether the reduction in neurite length of I228M-transfected neurons could be attenuated by blockade of sodium channel activity, cultures were treated for 3 days with the use-dependent sodium channel blocker carbamazepine (CBZ) at a concentration (10 μM) previously shown to protect central nervous system (CNS) axons from anoxic injury.¹¹ After 3 days of CBZ treatment, neurites of I228M-transfected neurons exhibited increased length compared to untreated neurons in parallel cultures (figure 3A; left panel). Mean total length/neuron was increased by ~25% in CBZ-treated I228M-transfected neurons compared to untreated neurons (see figure 3A; top right), and the length of CBZ-treated neurites approached that of WT-expressing neurons. As another control, we assessed the effect of CBZ on WT-transfected neurons with no mutant channels; 10 µM CBZ did not alter neurite length of WT-transfected neurons (see figure 3A; bottom right), consistent with a protective effect of CBZ through blockade of hyperactive I228M channel activity.⁶

Figure 3 Effect of carbamazepine (CBZ) and KB-R7943 on I228M-induced reduced neurite length. (A) Left panel: Indvidual neurons transfected with I228M, untreated or treated with 10µM CBZ. CBZ-treated neurons display increased neurite length compared to untreated neurons. Right panel: Quantification of the total neurite length/neuron calculated for neurons expressing I228M (top) or wild-type (WT; bottom), untreated or treated with CBZ. CBZ significantly increases neurite length in I228M-transfected neurons but does not alter neurite length of WT-transfected neurons. (B) Left panel: Individual neurons transfected with I228M, untreated (top) or treated (bottom) with 0.5µM KB-R7943. KB-R7943–treated neurons exhibit increased neurite length compared to untreated neurons. Right panel: Quantification of the total neurite length/neuron calculated for neurons expressing I228M (top) or WT (bottom) channels, untreated or treated with KB-R7943. KB-R7943 significantly increases neurite length in I228M-transfected neurons but does not affect neurite length of WT-transfected neurons. Data are presented as mean ± standard error of the mean. Scale bar: 200µM. *p < 0.05.

Reverse Na-Ca Exchange Contributes to Reduced Neurite Length of 1228M-Transfected Neurons

Increased axonal Na⁺ influx has been shown to trigger reverse operation of the Na-Ca exchanger, producing injurious intra-axonal Ca²⁺ overload and axonal dysfunction in the CNS and peripheral nervous system.¹²^,¹³ To investigate whether reverse Na-Ca exchange contributes to reduced neurite length in I228M-transfected neurons, we treated cultures with KB-R7943 at 0.5 μM, a concentration that inhibits reverse but not forward operation of NCX.¹⁴ Following 3 days of treatment, I228M-transfected neurons exhibited substantially longer neurites compared to untreated I228M-transfected cells (see figure 3B; left panel). Quantification of mean total neurite length/neuron for I228M-transfected neurons demonstrated a ~25% increase with KB-R7943 treatment compared to untreated I228M-transfected neurons (see figure 3B; top right). We also assessed the effect of KB-R7943 on WT-transfected neurons with no mutant channels, as a second control. WT-transfected neurons displayed similar mean total neurite length/neuron for KB-R7943-treated and untreated cultures (see figure 3B; bottom right).

Discussion

Loss of IENF is a defining characteristic and valuable tool for diagnosis of SFN,¹^–³ but molecular mechanisms underlying the axonal loss are still not understood. IENF are the distal ends of unmyelinated C- and thinly myelinated Aδ-nerve fibers that exit dermal nerve bundles and branch into fine-caliber free nerve endings within the epidermis. The epidermis is characterized by constant turnover of keratinocytes, which migrate from basal layers to the epidermal surface, so that the IENF run through a highly dynamic terrain. Substantial evidence suggests continuous axonal remodeling, regeneration, and sprouting of IENF within the mature, nonpathological epidermis.¹⁵^,¹⁶ In accordance with this schema, reduced IENF density in SFN has been proposed to result from both impaired growth or regenerative capacity of axons and axonal degeneration.⁴ The present results, although based on observations of an in vitro model and not differentiating between impaired regenerative capacity and axonal degeneration, show that expression of a gain-of-function Na_V1.7 variant associated with idiopathic SFN results in decreased length of DRG neurites in vitro, suggesting that the variant channel can injure sensory axons in one or both ways.

Faber et al. recently demonstrated gain-of-function Na_V1.7 variants in ~30% of patients with idiopathic SFN.⁶ Na_V1.7 is preferentially and abundantly expressed in nociceptive DRG neurons and is, along with sodium channels Na_V1.6, Na_V1.8, and Na_V1.9, present in small peripheral axons and their IENF.⁵ Small diameter is known to increase sensitivity to small changes in sodium conductance or influx, because it imposes a high input resistance, short electrotonic and diffusional length constant, and high surface-to-volume ratio.⁷^,⁸ Fine-caliber IENF are thus likely to be particularly vulnerable to gain-of-function changes of SFN-associated variant Na_V1.7 channels, and to the increased firing frequency and spontaneous firing of neurons expressing these channels.⁶

Our results demonstrate reduced length of neurites from neurons transfected with I228M variant Na_V1.7 channels. We also observed a trend toward shorter neurites for ML/VL and I720K variant-transfected neurons, compared to those transfected with WT channels after 3 days in vitro, implying an effect of these Na_V1.7 variant channels on neurite length. The significant reduction in neurite length for I228M-transfected neurons, and the trend toward reduced neurite length of ML/VL and I720K transfected neurons, may reflect differences in functional properties of DRG neurons expressing these variant channels. Electrophysiological recordings from DRG neurons in vitro showed a 29% prevalence of spontaneous firing in neurons transfected with I228M channels,⁹ whereas ML/VL and I720K channels exhibited smaller proportions (15%, 17%, respectively) of spontaneous firing neurons.⁶ Interestingly, the I228M variant was found in patients with proximal (face, scalp) pain at onset, whereas the I720K and ML/VL variants were identified in patients with early distal (feet) pain.⁶^,⁹

The I228M mutation substitutes a residue within the DIS4 segment of the channel, which is invariant in all human sodium channels except Na_V1.9, suggesting that it may play an important role in determining the functional properties of the Na_V1.7 channel. Although the I228M substitution is listed as a single nucleotide polymorphism in one database (Craig Ventor Human Genome) and was reported in <0.3% of control chromosomes in another series ¹⁷ and in 0.1% of chromosomes in the 1000 Genome Project (rs71428908), Estacion et al.⁹ did not find it in a control panel from 100 healthy ethnically matched controls (200 chromosomes). I228M displays substantially impaired slow inactivation, which increases the non-inactivated fraction of Na_V1.7 channels at potentials positive to −80mV, including resting potential.⁹ At the cellular level, expression of I228M depolarizes DRG neuron resting potential by ~5mV, induces spontaneous firing in nearly 30% of these cells, and doubles the firing rate in response to graded suprathreshold stimuli.⁹ Increased sodium influx into axons expressing mutant channels, and/or impaired calcium extrusion by NCX from them, would be predicted from each of these functional changes. The increased window current is predicted to produce sustained sodium influx, and spontaneous activity and increased evoked firing rates superimpose additional sodium influx associated with incremental action potentials.¹⁸ Moreover, because the Na-Ca exchanger is electrogenic, depolarization biases exchange against calcium efflux and toward calcium influx, as has been shown in axons within the anoxic optic nerve and peripheral axons subjected to anoxia, where persistent sodium influx has been shown to trigger reverse operation of the Na-Ca exchanger, causing an injurious intra-axonal Ca²⁺ overload and axonal dysfunction.¹²^,¹³ Our results suggest that NCX (present in peripheral axons and IENF as isoform NCX2⁵) contributes via reverse (Ca²⁺-importing) exchange to I228M-induced reduction in neurite length.

As with all in vitro models of neurodegenerative disorders, cultured neurons do not fully recapitulate the in vivo situation. In vitro models cannot fully mimic the time-dependent (years postnatal) or length-dependent (early effect on axons approximately 1m long for length-dependent, 10cm for neuropathies with early facial involvement) patterns of axonal damage occurring in patients with SFN. Indeed, the I228M variant has been reported in a small number of unaffected control chromosomes¹⁷ and in 2 unaffected children of a patient with I228M-associated SFN (ages younger than age of onset of neuropathy in affected individuals).⁹ Nevertheless, the altered channel biophysics and proexcitatory effects of SFN-associated Na_V1.7 variants⁶^,⁹ suggest that, at a minimum, they act as risk factors that predispose to development of SFN. Our results show a significant reduction in neurite length for DRG neurons transfected with the I228M Na_V1.7 variation, and a less pronounced trend toward reduction in neurite length in DRG neurons transfected with ML/VL and I720K. One possible explanation for this effect of variant channels on neuritic length after 3 days in culture might be that higher-than-normal overall levels of WT (endogenous) and variant (transfected) Na_V1.7 channels offset the brief time in culture and result in an accelerated pathogenic effect.

Irrespective of whether our in vitro assay reproduces in vivo time-dependence or fully reproduces the in vivo pattern of length-dependent axonal degeneration, we observed significant protective effects of the use-dependent sodium channel blocker CBZ and of reverse Na-Ca exchange (KB-R7943) on neurite length for DRG neurons transfected with I228M. These observations suggest that, similar to their injurious role in anoxic white matter¹³ and peripheral nerve¹² axons, activity of gain-of-function variant sodium channels and the Na-Ca exchanger in reverse mode may contribute to axonal injury in SFN. Assessment of IENF density, comparing progression of SFN with and without treatment with sodium channel blockers or blockers of Na-Ca exchange, might make it possible to determine whether this mechanism is at play in vivo, over years and in axons many centimeters in length, in human subjects with SFN.

References

1 Bednarik J, Vlckova-Moravcova E, Bursova S, et al. 2009. Etiology of small-fiber neuropathy. J Peripher Nerv Syst 14: 177–183.
2 Devigili G, Tugnoli V, Penza P, et al. 2008. The diagnostic criteria for small fibre neuropathy: From symptoms to neuropathology. Brain 131(pt 7): 1912–1925.
3 Lacomis D. 2002. Small-fiber neuropathy. Muscle Nerve 26: 173–188.
4 Bursova S, Dubovy P, Vlckova-Moravcova E, et al. 2012. Expression of growth-associated protein 43 in the skin nerve fibers of patients with type 2 diabetes mellitus. J Neurol Sci 315: 60–63.
5 Persson AK, Black JA, Gasser A, et al. 2010. Sodium-calcium exchanger and multiple sodium channel isoforms in intra-epidermal nerve terminals. Mol Pain 6: 84.
6 Faber CG, Hoeijmakers JG, Ahn HS, et al. 2012. Gain of function Na_V1.7 mutations in idiopathic small fiber neuropathy. Ann Neurol 71: 26–39.
7 Donnelly DF. 2008. Spontaneous action potential generation due to persistent sodium channel currents in simulated carotid body afferent fibers. J Appl Physiol 104: 1394–1401.
8 Waxman SG, Black JA, Kocsis JD, Ritchie JM. 1989. Low density of sodium channels supports action potential conduction in axons of neonatal rat optic nerve. Proc Natl Acad Sci USA 86: 1406–1410.
9 Estacion M, Han C, Choi JS, et al. 2011. Intra- and interfamily phenotypic diversity in pain syndromes associated with a gain-of-function variant of Na_V1.7. Mol Pain 7: 92.
10 Gladman SJ, Huang W, Lim SN, et al. 2012. Improved outcome after peripheral nerve injury in mice with increased levels of endogenous omega-3 polyunsaturated fatty acids. J Neurosci 32: 563–571.
11 Fern R, Ransom BR, Stys PK, Waxman SG. 1993. Pharmacological protection of CNS white matter during anoxia: Actions of phenytoin, carbamazepine and diazepam. J Pharmacol Exp Ther 266: 1549–1555.
12 Lehning EJ, Doshi R, Isaksson N, et al. 1996. Mechanisms of injury-induced calcium entry into peripheral nerve myelinated axons: Role of reverse sodium-calcium exchange. J Neurochem 66: 493–500.
13 Stys PK, Waxman SG, Ransom BR. 1992. Ionic mechanisms of anoxic injury in mammalian CNS white matter: Role of Na+ channels and Na(+)-Ca2+ exchanger. J Neurosci 12: 430–439.
14 Watanabe Y, Koide Y, Kimura J. 2006. Topics on the Naþ/Ca2þ exchanger: Pharmacological characterization of Naþ/Ca2þ exchanger inhibitors. J Pharmacol Sci 102: 7–16.
15 Cheng C, Guo GF, Martinez JA, et al. 2010. Dynamic plasticity of axons within a cutaneous milieu. J Neurosci 30: 14735–14744.
16 Verze L, Paraninfo A, Ramieri G, et al. 1999. Immunocytochemical evidence of plasticity in the nervous structures of the rat lower lip. Cell Tissue Res 297: 203–211.
17 Singh NA, Pappas C, Dahle EJ, et al. 2009. A role of SCN9A in human epilepsies, as a cause of febrile seizures and as a potential modifier of Dravet syndrome. PLoS Genet 5: e1000649.
18 Chow CC, White JA. 1996. Spontaneous action potentials due to channel fluctuations. Biophys J 71: 3013–3021.

Neuropathy-Associated Na_V1.7 Variant I228M Impairs Integrity of Dorsal Root Ganglion Neuron Axons *

Materials and Methods