Sodium channels, the electrogenisome and the electrogenistat: lessons and questions from the clinic

Abstract

In the six decades that have followed the work of Hodgkin and Huxley, multiple generations of neuroscientists and biophysicists have built upon their pivotal contributions. It is now clear that, in mammals, nine genes encode nine distinct voltage-gated sodium channels with different amino acid sequences and different physiological and pharmacological properties. The different sodium channel isoforms produce a multiplicity of distinct sodium currents with different time-dependent characteristics and voltage dependencies, which interact with each other and with the currents produced by other channels (including calcium and potassium channels) to shape neuronal firing patterns. Expression of these sodium channel isoforms is highly dynamic, both in the normal nervous system, and in the injured nervous system. Recent research has shed light on the roles of sodium channels in human disease, a development that may open up new therapeutic strategies. This article examines the pain-signalling system as an example of a neuronal network where multiple sodium channel isoforms play complementary roles in electrogenesis and a strong link with human disease has been established. Recent research suggests that it may be possible to target specific sodium channel isoforms that drive hyperexcitability in pain-signalling neurons, thereby providing new therapeutic strategies for chronic pain, and providing an illustration of the impact of the Hodgkin–Huxley legacy in the clinical domain.

Stephen Waxman received his MD and PhD degrees at the Albert Einstein College of Medicine. His mentors in physiology included Michael V. L. Bennett, Dominick Purpura and Patrick Wall. After training as a neurologist, he worked at Harvard, MIT and Stanford prior to moving to Yale, where he served from 1986 until 2009 as Chairman of Neurology. He currently is the Bridget Flaherty Professor of Neurology, Neurobiology and Pharmacology, and Director of the Center for Neuroscience & Regeneration Research at Yale. His research is aimed at finding new therapies for disorders of the nervous system. Having unravelled the contribution of ion channels to remissions in multiple sclerosis and having identified, for the first time, a sodium channel gene that controls pain sensitivity in humans, he continues to work on the pathobiology of ion channels. In 2009 he delivered The Physiological Society's Annual Review Prize Lecture, an honour that he shares with his heroes Andrew Huxley, John Eccles and Alan Hodgkin.

Introduction

Sixty years have passed since Hodgkin and Huxley demonstrated the role of sodium channels in action potential electrogenesis and predicted many of the properties of these channels (Hodgkin & Huxley, 1952). They made these pivotal discoveries in the squid giant axon, decades before the advent of patch-clamp recording methods, at a time when molecular biology was in its infancy and modern computers were not available. Hodgkin and Huxley could not see the channels and had no direct knowledge of their molecular structure. Nevertheless, they were able to predict, with prescient accuracy, many of the properties of these remarkable transmembrane proteins. The Hodgkin–Huxley formulation remains, to this day, one of the foundations of modern neuroscience. Importantly, it has not only provided a basis for an understanding of electrical signalling in the normal nervous system, but also has contributed to an understanding of neurological disorders that include certain forms of epilepsy (Reid et al. 2009; Meisler et al. 2010), migraine (de Vries et al. 2009; Pietrobon, 2010), neuromuscular disorders such as myotonia and periodic paralysis (Venance et al. 2006; Cannon, 2010; Jurkat-Rott et al. 2010; Matthews et al. 2010), and neuropathic pain (Dib-Hajj et al. 2010; Raouf et al. 2010).

In the decades that followed the work of Hodgkin and Huxley, and with an accelerating pace over recent years, several generations of neuroscientists and biophysicists have built upon their contribution. We now know that, in mammals, nine genes encode nine distinct voltage-gated sodium channels (or more strictly, nine distinct sodium channel isoforms), all sharing a common overall design motif, but with different amino acid sequences and different physiological and pharmacological properties (Catterall et al. 2005). The different channels have been termed, using consensus nomenclature, Na_V1.1–Na_V1.9. These different channels have different voltage dependencies and kinetics, and play multiple roles in electrogenesis, some activating to produce inward currents at or above action potential threshold, others activating in a subthreshold range. It has become clear that expression of sodium channels is not a fixed or static process but, on the contrary, is highly dynamic. And, still building upon the findings of Hodgkin and Huxley in the lowly squid, insights are being gained into the roles of sodium channels in human disease, a development that may open up new therapeutic strategies.

Sodium channels play central roles in electrogenesis in almost all types of neurons within the mammalian nervous system. This article will focus on the pain-signalling system as an example of a neuronal network where multiple sodium channel isoforms play different and interdependent roles in electrogenesis and a strong link to human disease has been established. In discussing this emerging area, this article will illustrate how observations at the laboratory bench can inform clinical medicine, and how clinical observations can suggest hypotheses and propel questions about the basic biology of sodium channels.

The electrogenisome: especially complex in dorsal root ganglion neurons

The electrogenisome, or the set of molecules that confer electrical excitability, is complex in all cell types, but is especially rich within dorsal root ganglion (DRG) neurons, where a robust ensemble of voltage-gated sodium, calcium and potassium channels interact with each other and with other molecules, to produce multiple sub- and suprathreshold currents. In the aggregate, these currents shape the cells’ firing patterns.

Multiple sodium channel isoforms are present within these cells, particularly within nociceptive DRG neurons. As might be expected from the pivotal roles of sodium channels in electrogenesis, many of the drugs currently used to treat neuropathic pain (e.g. carbamazepine, mexiletine) are sodium channel blockers; other drugs used to treat pain, while initially designed to target other molecules, have sodium channel-blocking effects (Yang et al. 2009). Few, if any, of the existing sodium channel blockers exert their actions in a strictly focal manner, specifically on a single sodium channel isoform, but many exert differential effects on particular isoforms.

Importantly, nociceptors express several sodium channel isoforms that are not present at high levels within muscle, cardiac tissue, or the CNS. These ‘peripheral sodium channels’, Na_V1.7 (Toledo-Aral et al. 1997), Na_V1.8 (Akopian et al. 1996), and Na_V1.9 (Dib-Hajj et al. 1998b), play important roles in nociception (Rush et al. 2006) and one of them, Na_V1.7, is the first sodium channel to be identified as a contributor to pain in humans (Waxman & Dib-Hajj, 2005; Dib-Hajj et al. 2010). Each of these three peripheral sodium channels has unique functional attributes. Na_V1.7 activates in response to small slow depolarizations close to resting potential so as to produce its own depolarization (Fig. 1). Na_V1.7 thus has the capability of amplifying inputs such as generator potentials (Cummins et al. 1998; Herzog et al. 2003), and sets the gain on nociceptors (Waxman, 2006). The tetrodotoxin (TTX)-resistant sodium channel Na_V1.8, which was initially called SNS (sensory neuron specific) because of its discovery in peripheral sensory neurons where it is preferentially expressed (Akopian et al. 1996), is relatively resistant to inactivation by depolarization (Fig. 2A) and recovers rapidly from inactivation. Na_V1.8 thus produces repetitive firing in depolarized DRG neurons (Renganathan et al. 2001; Fig. 2B–D). Another tetrodotoxin-resistant sodium channel, Na_V1.9, is characterized by very slow activation and inactivation with a large overlap centred near resting potential (Cummins et al. 1999); this channel contributes a sodium conductance at rest that modulates the excitability of DRG neurons (Herzog et al. 2001; Baker et al. 2003).

Figure 1. Na_V1.7 amplifies small depolarizing stimuli.

Na_V1.7 exhibits slow closed-state inactivation, and thus can respond to small graded subthreshold potentials such as ramp stimuli, amplifying these depolarizations and bringing the cell closer to action potential threshold. From Cummins et al. (1998) with permission from the Society for Neuroscience.

Figure 2. Na_V1.8 supports repetitive firing in depolarized DRG neurons.

A, the voltage dependence of steady-state inactivation (estimated by measuring currents elicited by 20 ms test pulses to −10 mV after 500 ms inactivating prepulses ranging from −130 to −10 mV) for TTX-resistant currents (largely Na_V1.8 in this recording configuration) (filled squares) is depolarized compared to that of TTX-S currents (open circles) in DRG neurons. Modified from Cummins & Waxman (1997) with permission from the Society for Neuroscience. B–D, Na_V1.8 is a major contributor to the action potential upstroke in small DRG neurons. B, 1 s current injection in the presence of 250 nm tetrodotoxin (TTX) generates a train of TTX-resistant action potentials in DRG neurons (Fig. 2C). D, Na_V1.8^−/− DRG neurons are incapable of sustaining high frequency firing in response to identical stimuli. From Renganathan et al. (2001).

These peripheral sodium channel isoforms, as well as Na_V1.1 and Na_V1.6, interact in a complex manner to produce action potentials in DRG neurons. As shown in Fig. 3, under normal conditions, Na_V1.7 (together with Na_V1.6 and Na_V1.9 in some DRG neurons) operate in the range below the threshold for generation of all-or-none action potentials, and amplify small depolarizing inputs, bringing the neuron towards action potential threshold. Na_V1.8, activated at a relatively depolarized potential, contributes the majority of the inward current underlying the action potential in DRG neurons (Renganathan et al. 2001; Blair & Bean, 2002). As noted below, this mode of physiological coupling between Na_V1.7 and Na_V1.8 has important pathophysiological implications. Moreover, it appears likely that Na_V1.7 and Na_V1.8 interact to produce subthreshold membrane potential oscillations that can trigger ectopic repetitive firing in nociceptors (Yang et al. 2005; Kovalsky et al. 2009; Choi & Waxman, 2011).

Figure 3. Multiple sodium channel subtypes participate in electrogenesis in small DRG neurons.

Representation of subthreshold responses (dotted line) and overshooting action potentials solid from small DRG neurons. Under normal conditions, Na_V1.7 (and Na_V1.6 and/or Na_V1.9 in some cells) brings the neuron toward threshold (dashed line), and Na_V1.8 is largely responsible for the overshooting action potential with minor contributions of Na_V1.1, Na_V1.6 and Na_V1.7 to the action potential upstroke. A large depolarization is necessary for generation of the all-or-none action potential, due to the depolarized voltage dependence of activation of Na_V1.8. Modified from Rush et al. (2007).

Another sodium channel, Na_V1.3, is of special interest in terms of neuropathic pain because it is up-regulated within DRG neurons following peripheral nerve injury (Waxman et al. 1994; Black et al. 1999). Na_V1.3 displays three properties which can contribute to DRG neuron hyperexcitability: production of depolarizing responses to small ramp-like inputs, rapid recovery from inactivation (Cummins & Waxman, 1997), and production of a relatively large persistent current (Lampert et al. 2006).

Na_V1.7, Na_V1.8, Na_V1.9 and Na_V1.3 have become targets in the search for more effective pain pharmacotherapeutics, with the expectation that subtype-specific blockade of one or more of these channel isoforms might attenuate impulse activity within pain-signalling DRG neurons with little effect on CNS neurons or myocardium, thereby muting pain with minimal cardiac or CNS side effects. A possible caveat arises from occasional observations that have been interpreted as suggesting that these channel isoforms may also possibly contribute to activity of at least some types of CNS neurons (Whitaker et al. 2001; Blum et al. 2002; Holland et al. 2008; Singh et al. 2009; Estacion et al. 2010) or cardiac cells (Thimmapaya et al. 2005; Chambers et al. 2010). Expression at levels that contribute to function of the heart or brain, if confirmed in future studies, could pose a challenge for the clinical development of Na_V1.7-, Na_V1.8-, Na_V1.9-, or Na_V1.3-specific block as an approach to pain. However, as noted below, human subjects with channelopathy-associated insensitivity to pain lack functional Na_V1.7 channels, and do not appear to suffer from deficits other than absent pain sensitivity, suggesting that block of Na_V1.7 may be well-tolerated in humans. Moreover, methods may be developed that can selectively attenuate the excitability of nociceptors without affecting other types of neurons. One particularly novel approach selectively introduces a charged, membrane-impermeant sodium channel blocker such as the lidocaine derivative QX-314 through the TRPV1 channel via co-application of the sodium channel blocker and capsaicin (Binshtok et al. 2007).

Does an electrogenistat regulate neuronal excitability?

The presence of multiple subtypes of sodium channels within nociceptors reminds us that regulation of excitability of these cells must be a complex process. Under non-pathological conditions the firing properties of nociceptors, like those of most neurons, are usually maintained within a circumscribed range. This appears to be, at least in part, a result of homeostatic regulation of ion channel expression, post-translational modification, and/or interaction with binding partners or modulators.

While variations in the level of expression of any one of the sodium channel isoforms present within DRG neurons could in principle alter their level of excitability, computational models have demonstrated that multiple, distinct sets of membrane parameters can produce similar levels of activity. Thus changes in expression of channel ‘B’ can compensate for changes in expression of channel ‘A’ to maintain excitability within a particular range (Prinz et al. 2004). Experimental demonstrations of this type of homeostasis are provided by the observation that overexpression of the Shal potassium channel gene in invertebrate somatogastric ganglion neurons results in a large increase in I_A but little change in the neuron's firing properties, which are retained as a result of a compensatory increase in expression of a hyperpolarization-activated inward current (I_h). The compensatory change appears to occur at the transcriptional level, and may be a response to the presence of increased Shal protein within the cell (MacLean et al. 2003). Another example is provided by Purkinje neurons from Na_V1.6^−/− mice in which sodium current density is reduced in the long term, where an upregulation of calcium channels maintains excitability at close to its normal level (Swensen & Bean, 2005).

Changes in sodium channel expression have also been demonstrated in vivo in association with changes in the functional status of neurons. An example is provided by supraoptic magnocellular neurosecretory neurons within the hypothalamus, which alter their level of sodium channel expression as they switch from a quiescent state (in which they fire at <3 impulses per second) to an active state in which they produce high-frequency impulse bursts (≥10 per second) and release vasopressin in response to changes in brain osmolality (Tanaka et al. 1999). Figure 4 shows the up-regulated transcription of the genes encoding Na_V1.2 and Na_V1.6, and Fig. 5 shows the augmented amplitudes of currents produced by these channels that are observed when the magnocellular neurons switch to their high-frequency firing state. These observations provide a demonstration of molecular and functional remodelling of neurons, whereby these cells selectively activate specific ion channel genes and deploy functional channels to maintain the tuning of their electrogenisome.

Figure 4. Up-regulation of Na_V1.2 and Na_V1.6 in vivo in active supraoptic magnocellular neurons.

A, in situ hybridization with subtype-specific riboprobes for Na⁺ channel subunits Na_V1.2 and Na_V1.6 in the hypothalamic supraoptic nucleus (SON). Low levels of Na_V1.2 and Na_V1.6 mRNA are present in the control SON (left), and there is a distinct up-regulation of each of these transcripts after salt-loading (right), which increases osmolality and triggers high-frequency bursting in these cells. B, quantification of optical densities shows significant changes for Na_V1.2 and Na_V1.6 mRNA. (Bar = 100 μm.) Modified from Tanaka et al. (1999) with permission (©1999 National Academy of Sciences, USA).

Figure 5. Na⁺ currents are augmented in active SON neurons.

A, family of traces from representative SON neurons freshly isolated from control (left) or salt-loaded (right) rats. The currents were elicited by 40 ms test pulses to potentials from −60 to 30 mV. Cells were held at −100 mV. B, ramp currents are elicited in MNCs by slow voltage ramps (600 ms voltage ramp from −100 to 40 mV). Left, the ramp current in a salt-loaded MNC before and after the addition of 250 nm TTX to the extracellular solution. TTX blocks the ramp current. Right, the TTX-sensitive ramp currents in representative control and salt-loaded MNCs. Leak currents recorded after application of 250 nm TTX were subtracted. C, the peak and ramp current densities (estimated by dividing the maximum currents by the cell capacitance) are larger in salt-loaded neurons (n = 29) than in control neurons (n = 34); note that the increase is proportionately greater for the ramp currents. Error bars indicate SEM; *P < 0.005. From Tanaka et al. (1999) with permission (©1999 National Academy of Sciences, USA).

These examples of homeostatic regulation of intrinsic neuronal excitability imply a need for an ‘electrogenistat’ within excitable cells, but its molecular workings are not yet understood. More is known about the effector pathway controlling ion channel expression than about the functional demand sensor. As with many intracellular control processes, changes in intracellular ion concentrations, particularly those for [Ca²⁺] and/or [Na⁺], may play a role. In rat muscle cells in vitro, electrical activity and increased cytosolic calcium decrease the level of sodium channel mRNA, with a consequent reduction in the number of functional sodium channels, providing a potential mechanism for feedback regulation of sodium channel density by electrical activity (Offord & Catterall, 1989). Calcium-mediated down-regulation of sodium channel expression, triggered by electrical activity, has also been reported in cultured cardiac myocytes (Chiamvimonvat et al. 1995). Dargent & Couraud (1990) have presented data indicating that, in developing brain neurons in vitro, Na⁺ influx triggers sodium channel internalization. Noting that this mode of channel internalization does not occur in astrocytes and is absent in adult brain, they have suggested that channel internalization in response to changes in intracellular [Na⁺] levels may provide a cell- and time-dependent mechanism that regulates the acquisition of electrical excitability during brain development (Dargent et al. 1994). The regulatory pathways may, in fact, be sensitive to temporal aspects of electrical activity such as the pattern of firing; it is known, for example, that cAMP-responsive element binding protein expression, and mitogen-activated protein kinase signalling, are sensitive to the pattern of electrical activity within neurons (Fields et al. 1997). Additional layers of complexity are added by evidence indicating that alternative splicing may alter channel properties (Chatelier et al. 2008; Choi et al. 2010) and that trophic factors and other molecules that modulate channels or their expression can rapidly modulate excitability. For example, brief exposure (as short as 1 min) to nerve growth factor (NGF), interferon-γ (IFNγ), epidermal growth factor (EGF) or basic fibroblast growth factor (bFGF) can evoke up-regulation of expression of Na_V1.7 within model systems such as PC12 cells (Toledo-Aral et al. 1995). Interaction with partner molecules (Liu et al. 2001; Okuse et al. 2002; Shah et al. 2004; Rush et al. 2005; Dib-Hajj et al. 2010) has also been shown to contribute to the regulation of expression of multiple sodium channel isoforms. Expression of these channel partners is dynamic and their levels may be regulated by electrical activity, introducing further complexity.

Lessons from pain syndromes due to sodium channel mutations

The Hodgkin–Huxley legacy has now reached the pain clinic: In 2004 inherited erythromelalgia, an autosomal dominant disorder in which patients experience excruciating, burning pain and erythema of the extremities in response to mild warmth, was identified as the first human pain disorder known to be produced by sodium channels. Also called the ‘man on fire syndrome’, inherited erythromelalgia is produced by gain-of-function mutations of the Na_V1.7 sodium channel which shift activation in a hyperpolarizing direction, slow deactivation, and enhance the channel's response to small depolarizations, thereby producing hyperexcitability of nociceptive DRG neurons (Dib-Hajj et al. 2005, 2010; Rush et al. 2006). Most cases of inherited erythromelalgia are unresponsive to pharmacotherapy although, as noted below, a few families house Na_V1.7 mutations that sensitize the channel to specific channel blockers (Choi et al. 2009; Fischer et al. 2009b). A second pain disorder with a distinct clinical phenotype of perirectal, periocular and perimandibular pain, paroxysmal extreme pain disorder, was subsequently shown to be produced by a different set of gain-of-function Na_V1.7 mutations which impair channel inactivation and enhance persistent current (Fertleman et al. 2006). Still other studies showed that loss-of-function mutations of Na_V1.7 produce channelopathy-induced insensitivity to pain, a disorder in which patients do not feel pain, even in response to events such as bone fractures, burns, dental abscesses or dental extractions that should cause pain (Cox et al. 2006; Goldberg et al. 2007).

Erythromelalgia has begun to teach us lessons, not just about Na_V1.7, but also about its physiological partners. Episodic reddening of the skin due to abnormal neurovascular control accompanies pain in the affected limbs in inherited erythromelalgia. The dual manifestations of erythromelalgia mutations reflect their divergent functional effects on the two types of peripheral neurons, DRG neurons and sympathetic ganglion neurons, where Na_V1.7 is preferentially expressed. The L858H Na_V1.7 mutation, one of the earliest to be identified as a cause of inherited erythromelalgia, has been assessed by current-clamp in these two cell types and has been shown to produce hyperexcitability within DRG neurons and hypoexcitability within sympathetic ganglion neurons. This divergent effect of a single mutation is the result of the different cell backgrounds provided by DRG versus sympathetic ganglion neurons, specifically the selective expression of Na_V1.8 within DRG neurons, and its absence within sympathetic ganglion neurons (Rush et al. 2006). As described above, Na_V1.8 is relatively resistant to inactivation by depolarization. The opposing effects of the Na_V1.7 mutation in these two cell types arise from the 4–6 mV depolarization of resting membrane potential (presumably due to enhanced window current near resting potential as a result of increased overlap between activation and steady-state inactivation) that is produced by erythromelalgia mutations (Rush et al. 2006; Dib-Hajj et al. 2010). The way in which Na_V1.8 shapes the response to depolarization can be seen in Fig. 6, which shows that the depolarization produced in a sympathetic ganglion neuron by the L858H erythromelalgia Na_V1.7 mutation inactivates the sodium channels within these cells and reduces their excitability. Excitability can be rescued by co-expression of Na_V1.8 channels together with the mutant Na_V1.7 channels (Fig. 6A–C), an experimental manoeuvre which transforms sympathetic ganglion cells into DRG-like neurons (Rush et al. 2006).

Figure 6. Excitability of sympathetic ganglion neurons is reduced by erythromelalgia Na_V1.7 mutation L858H but can be rescued by coexpression of Na_V1.8.

A, suprathreshold responses recorded from representative superior cervical ganglion (SCG) neurons transfected with WT (continuous line), L858H (dotted line) and L858H plus Na_V1.8 (dashed line) channels. B, depolarized resting membrane potential (RMP) in cells transfected with L858H channels is maintained with coexpression on Na_V1.8. Dashed line indicates RMP in cells expressing WT Na_V1.7 alone. C, current threshold for action potential firing is reduced by coexpression of Na_V1.8 together with L858H, when compared to L858H alone (*P < 0.05). D, action potential overshoot in SCG neurons expressing L858H mutant channel is increased when Na_V1.8 is coexpressed with L858H (*P < 0.05). Thus, the expression of Na_V1.8, which has depolarized inactivation voltage dependence, allows SCG neurons to fire full overshooting action potentials, despite the depolarization induced by L858H Na_V1.7 erythromelalgia mutation. Modified from Rush et al. (2006) with permission (©2006 National Academy of Sciences, USA).

This experiment, motivated by observations in the clinic, demonstrates two important principles. The first is that the effects of ion channel mutations on neuronal function are not necessarily unidirectional or predictable on the basis of changes in channel function per se; on the contrary a single ion channel mutation can have divergent functional effects in different types of neurons. The second is that cell background and specifically the precise make-up of the electrogenisome can shape the functional effects of an ion channel mutation. In this case the presence or absence of Na_V1.8 determines the functional effects of an Na_V1.7 mutation at the cellular level.

Inherited erythromelalgia may also hold lessons that could be relevant to the development of personalized, genomically based pharmacotherapy for pain. Although most patients with inherited erythromelalgia are refractory to pharmacotherapy, Fischer et al. (2009a) have described an unusual kindred in which pain is relieved in multiple family members by carbamazepine. Patch-clamp studies of the V400M mutation in this family demonstrated changes in activation and steady-state inactivation. Notably, carbamazepine, at concentrations within the human therapeutic range, normalizes the voltage dependence of activation and inactivation in the mutant channels, although it does not affect these parameters in wild-type Na_V1.7. Choi et al. (2009) reported another patient, carrying a V872G mutation of Na_V1.7, with erythromelalgia that was sensitive to treatment with mexiletine, and showed that the V872G mutant channel displays stronger use-dependent fall-off of current following exposure to mexiletine, compared to wild-type channels. The unusual patterns of response of the mutant channels to sodium channel-blocking agents in these patients provide a molecular correlate to their clinical phenotype of pharmacoresponsive pain, and suggest that genomically based pharmacotherapy for pain may be a tractable goal.

Does the electrostatic set-point change in disease?

Patients with inherited erythromelalgia or paroxysmal extreme pain disorder harbour gain-of-function mutants of Na_V1.7 throughout their lives. Most of these patients experience life-long pain beginning in infancy or childhood, presumably because their DRG neurons fire at higher than normal frequencies throughout their lives. Conversely, patients with channelopathy-associated insensitivity to pain exhibit their pain-free phenotype beginning in infancy, and nociceptive neurons in these patients presumably fire at lower than normal frequencies throughout their lives. Whether there are changes in the set-points of the electrogenistats within DRG neurons or higher-order neurons along pain-signalling pathways in patients with these disorders is not yet known.

Two interesting patterns of response to sodium channel blockade that may bear on this question have been seen in patients with inherited erythromelalgia due to the V872G Na_V1.7 mutation. In the first of these cases (Choi et al. 2009), pain began at age 5, and was markedly reduced during a 1 month period of treatment with mexiletine at age 7. Treatment was then suspended because of parental concern about potential side effects. Despite cessation of treatment, however, there was sustained improvement, with substantially reduced pain 1 year following suspension of treatment. This case raises the question: was the set-point of the electrogenistat in this patient re-set to a lower level, closer to normal, as a result of a transient period of sodium channel blockade and attenuation of high-frequency firing early in life?

Another interesting phenomenon is illustrated by a second patient with erythromelalgia associated with the V872G mutation who began to experience severe burning pain at age 7. Mexiletine was initiated at age 33, and provided substantial improvement for 2 months. After 2 months, however, the effect subsided, and the patient relapsed to his previous state of severe burning pain despite continued administration of the sodium channel blocker. The literature contains descriptions of a number of similar cases, where there was an initial, but unsustained, response to sodium channel blockade (Harty et al. 2006). In cases such as this, the beneficial response usually lasts for several months and then abates. While transient relief in cases such as these could represent a placebo response, it may alternatively represent an initial pharmacological effect of sodium channel blockade which results in early pain relief, followed by a tachyphylactic loss of effectiveness. If this is the case, the question arises: are changes in neuronal excitability, possibly due to changes in sodium channel expression, triggered in some patients by treatment with sodium channel blockers?

The effects of sodium channel blockers on sodium channel expression have not been examined in DRG neurons. However, it is known that exposure to the sodium channel blocker TTX triggers up-regulated sodium channel expression in developing CNS neurons (Dargent & Couraud, 1990) and that transcription of sodium channel genes is upregulated by chronic treatment with phenytoin in CNS neurons in mice with epilepsy (Sashihara et al. 1994). It is possible that, as a consequence of sustained hyperactivity along pain-signalling pathway over years, the electrogenistat is re-set to a high set-point either within DRG neurons, or within postsynaptic neurons at higher levels within the pain-signalling system, e.g. within dorsal horn or thalamic neurons. If this is the case, sodium channel blockers might trigger a compensatory up-regulation of sodium channel expression, which could maintain neuronal firing rates at high levels, thus limiting the long-term effectiveness of subtype-specific sodium channel blockade. Development of transgenic mouse models of inherited erythromelalgia, or of induced pluripotent stem cell (iPS) models, might permit this putative mechanism to be assessed in vivo even prior to human studies.

From squid to clinic: questions and future directions in sodium channel research

Building upon the pivotal discoveries of Hodgkin and Huxley in the squid giant axon, the last six decades have seen exciting progress in our understanding of mammalian and even human sodium channels. There has been a productive convergence of information derived by study of normal channels at the laboratory bench, and of pathological aspects of electrogenesis as studied in disease models.

As noted above, one important set of questions focuses on the details of the intracellular signalling pathways that regulate ion channel expression. Relatively little is known about the afferent arm of the feedback pathway, or about the sensor. While the available evidence suggests that some components of the signalling cascade may be sensitive to temporal aspects of electrical activity, we do not yet understand the precise mechanism and nature of this time dependency. Nor is it clear whether activity-driven changes in the set-point of the electrogenistat are permanent. There is a need for us to learn whether various components of the network are altered in disease. If sodium channel blockade, e.g. with pharmacological blockers, triggers changes in channel expression, we need to understand whether the changes are different for different cell types, or for different types of neurons, different channel isoforms, or different pathological insults.

Another important set of questions concerns the persistence of changes in ion channel transcription that accompany damage to the nervous system. Following injury to peripheral axons, up-regulation of Na_V1.3 expression within axotomized DRG neurons appears to be due, at least in part, to loss of access to a peripheral pool of nerve growth factor (NGF). The changes can, in part, be experimentally reversed by permitting the nerve to regenerate and reinnervate its original target (Dib-Hajj et al. 1998a) or by supplying NGF to the injured nerve tips (Leffler et al. 2002). These observations raise the question of whether, following re-setting of the electrogenistat after nerve injury in humans, the set-point can be returned therapeutically to its original level.

Still another set of questions focus on whether, following peripheral nerve injury, the electrogenistat may be reset at multiple levels along the neuraxis. Nerve injury triggers up-regulated expression of Na_V1.3 within higher-order dorsal horn neurons and thalamic neurons that participate in pain signalling, and these molecular changes appear to contribute to increases in the intrinsic excitability of these cells (Hains et al. 2004; Zhao et al. 2006). There is electrophysiological evidence for increased excitability of thalamic neurons in animals with experimental diabetic neuropathy (Fischer et al. 2009b). Studies in human subjects with limb amputation (which transects peripheral nerves) have also revealed increased burst firing of thalamic neurons, possibly due to changes in intrinsic electrogenic mechanisms (Lenz et al. 1998) and evidence derived from magnetic resonance spectroscopy suggests that there is thalamic dysfunction in patients with diabetic neuropathy (Selvarajah et al. 2008; Sorensen et al. 2008). We still do not understand the degree to which central hyperexcitability contributes to pain in these examples. If central generators contribute substantially to pain and operate in a manner independent of the afferent barrage from peripheral neurons in these settings, block of peripheral sodium channels (Na_V1.7, Na_V1.8, and/or Na_V1.9) might have less efficacy than expected. This problem might be overcome if molecular methods, such as targeted gene knock-down or other gene manipulation methods (Yeomans et al. 2005; Fink, 2011 in press), could be used to re-set the electrogenistat within central as well as peripheral neurons.

As new therapeutic candidates are identified and move toward clinical application, it will be increasingly important for clinical and translational scientists to maintain a dialogue with ion channel biologists. Clinical observation may provide important clues or answers and, conversely, basic research will likely propel the development of new and more effective medications. This work will continue to build upon the work of Hodgkin and Huxley, whose discoveries remain a bastion of basic neuroscience, and the Hodgkin–Huxley legacy will almost certainly have an increasingly important impact in the clinical domain.