Journal of Neurophysiology

Tetrodotoxin-resistant voltage-dependent sodium channels in identified muscle afferent neurons

Renuka Ramachandra, Stephanie Y. McGrew, James C. Baxter, Esad Kiveric, Keith S. Elmslie

Abstract

Muscle afferents are critical regulators of motor function (Group I and II) and cardiovascular responses to exercise (Group III and IV). However, little is known regarding the expressed voltage-dependent ion channels. We identified muscle afferent neurons in dorsal root ganglia (DRGs), using retrograde labeling to examine voltage-dependent sodium (NaV) channels. In patch-clamp recordings, we found that the dominant NaV current in the majority of identified neurons was insensitive to tetrodotoxin (TTX-R), with NaV current in only a few (14%) neurons showing substantial (>50%) TTX sensitivity (TTX-S). The TTX-R current was sensitive to a NaV1.8 channel blocker, A803467. Immunocytochemistry demonstrated labeling of muscle afferent neurons by a NaV1.8 antibody, which further supported expression of these channels. A portion of the TTX-R NaV current appeared to be noninactivating during our 25-ms voltage steps, which suggested activity of NaV1.9 channels. The majority of the noninactivating current was insensitive to A803467 but sensitive to extracellular sodium. Immunocytochemistry showed labeling of muscle afferent neurons by a NaV1.9 channel antibody, which supports expression of these channels. Further examination of the muscle afferent neurons showed that functional TTX-S channels were expressed, but were largely inactivated at physiological membrane potentials. Immunocytochemistry showed expression of the TTX-S channels NaV1.6 and NaV1.7 but not NaV1.1. NaV1.8 and NaV1.9 appear to be the dominant functional sodium channels in small- to medium-diameter muscle afferent neurons. The expression of these channels is consistent with the identification of these neurons as Group III and IV, which mediate the exercise pressor reflex.

  • NaV1.8
  • NaV1.9
  • A803467
  • dorsal root ganglion neurons
  • exercise pressor reflex

muscle afferents are important regulators of motor and cardiovascular physiology. The Group I (Aα) and II (Aβ) afferents provide sensory feedback to control ongoing motor activity (Houk 1974). The Group III (Aδ) and IV (C) afferents mediate the exercise pressor reflex that is normally triggered by mechanical and metabotropic signals associated with skeletal muscle contraction (Kaufman and Hayes 2002). This reflex is an important neural mechanism regulating cardiovascular adjustments to exercise (Kaufman and Hayes 2002) and has important clinical implications because inappropriate activation by certain disease states such as peripheral vascular diseases (e.g., claudication) can adversely affect the cardiovascular system (Baccelli et al. 1999; Bakke et al. 2007; Smith et al. 2006). Thus an understanding of the receptors and ion channels that mediate excitability of these muscle afferents could facilitate development of selective treatments to suppress the exercise pressor reflex and normalize cardiovascular function in patients with peripheral vascular disease.

Voltage-dependent sodium (NaV) channels are the source of electrogenesis in excitable cells and contribute to excitability of muscle afferents. There are nine genes that code for NaV channels (Catterall et al. 2005; Rush et al. 2007). Of these, five are known to be expressed in adult peripheral sensory neurons, NaV1.1, 1.6, 1.7, 1.8, and 1.9 (Benn et al. 2001; Rush et al. 2007). These channels can be broadly classified on the basis of their sensitivity to tetrodotoxin (TTX) as TTX resistant (TTX-R) and TTX sensitive (TTX-S). Of the five NaV channels that are present in adult sensory neurons, three are TTX-S, while NaV1.8 and NaV1.9 are TTX-R. NaV1.1 has a widespread distribution in dorsal root ganglion (DRG) neurons, but its function in sensory neurons has not been fully elucidated (Rush et al. 2007). However, NaV1.1 is localized to the initial axon segment and nodes of Ranvier in spinal neurons (Duflocq et al. 2008). NaV1.6 and 1.7 are necessary for action potential initiation and transmission in both central and peripheral neurons (Rush et al. 2007). NaV1.8 is present in small to medium unmyelinated and thinly myelinated DRG neurons (Djouhri et al. 2003; Renganathan et al. 2001) and is well studied for its role in nociception and chronic pain (Dong et al. 2007; Jarvis et al. 2007; Zimmermann et al. 2007). NaV1.9 is primarily expressed in unmyelinated DRG neurons and has been shown to play a role in nociception (Fang et al. 2002).

Little is known regarding the NaV channels involved in electrogenesis of muscle afferents. Here we investigated the NaV channels expressed in muscle afferent neurons that were identified by retrograde transport of a dye injected into the triceps surae muscle. TTX-R channels appear to be the dominant NaV channel type in the majority of muscle afferent somas from which we recorded. TTX-S channels are also present but appear to be largely inactivated at physiological resting membrane potentials. Our evidence supports the expression of both NaV1.8 and 1.9 in these neurons. NaV1.8 is highly expressed in small and medium C and Aδ sensory neurons (Djouhri et al. 2003; Puhl and Ikeda 2008), which suggests that these muscle afferents were Group III (Aδ) and IV (C) that mediate the exercise pressor reflex (Kaufman and Hayes 2002; Kaufman et al. 1983).

MATERIALS AND METHODS

Labeling of muscle afferent neurons.

Adult male Sprague-Dawley rats were obtained from Hill Top Laboratories (Scottdale, PA). The rats were housed in a temperature-controlled room (24 ± 1°C) with a 12:12-h light dark cycle and fed a standard diet and tap water ad libitum. All procedures were reviewed and approved by the Institutional Animal Care and Use Committee and followed National Institutes of Health guidelines.

Retrograde labeling with the lipophilic dye 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI) was used to identify muscle afferent neurons. Rats (wt 150–400 g) were anesthetized with an intraperitoneal injection of a mixture of ketamine, xylazine, and acepromazine. The calf area of both legs was shaved, and the left and the right triceps surae muscles were injected with 100 μl of 1.5% DiI in dimethyl sulfoxide (DMSO). The injections were made parallel to the longitudinal axis of the muscle for better dye distribution. Cutaneous afferents were labeled by the same method except that DiI was injected into the skin of the L4 and L5 dermatomes. The rats were killed 4–5 days after injection for neuronal isolation.

Neuronal isolation.

The triceps surae muscles receive sensory innervation from neurons in the lumbar DRGs L4 and L5 (Smith et al. 2006). These DRGs were obtained after the DiI-injected rat was killed with CO2 followed by decapitation. The DRG neurons were dissociated by incubation of the ganglia at 37°C for 60 min in an enzyme mixture containing (mg/ml) 1 trypsin, 0.7 collagenase, and 0.1 DNase in Earle's balanced salt solution that was gassed with 95% O2-5% CO2. The cells were washed in minimum essential medium (MEM) containing 10% fetal bovine serum (FBS) and plated onto polylysine-coated glass coverslips. The isolated neurons were maintained overnight at 37°C in a 5% CO2 incubator in MEM with 10% FBS and 1% penicillin-streptomycin.

Solutions.

DiI-positive DRG neurons were recorded in a solution containing (in mM) 45 NaCl, 100 N-methyl-d-glucosamine (NMG)-Cl, 4 MnCl2, 10 Na-HEPES, and 10 glucose, pH 7.4 and osmolarity 320 mosM. The reduced extracellular Na+ concentration (50 vs. 140 mM Na+) provided smaller NaV currents to greatly improve neuronal membrane voltage control. When testing the sodium sensitivity of the current we switched to a zero-sodium external solution containing (in mM) 145 NMG-Cl, 4 MnCl2, 10 NMG-HEPES, and 10 glucose, pH 7.4 and osmolarity 320 mosM. The pipette solution contained (in mM) 104 NMG-Cl, 14 creatine-PO4, 6 MgCl2, 10 NMG-HEPES, 5 Tris-ATP, 10 NMG2-EGTA, and 0.3 Tris2-GTP, pH 7.4 and osmolarity 300 mosM. Ion channel blockers were made up as stock solutions in either water or DMSO and diluted into the external solution to make the working concentration. TTX and CdCl2 were soluble in water, but A803467 (A803) was dissolved in DMSO prior to addition to the external solution. The maximal DMSO concentration was 0.02%. All solutions applied to a given neuron contained the same DMSO concentration so that the only change was presence or absence of the blocker. Solutions were applied with a gravity-fed perfusion system with six inputs and a single output. The solution exchange time was 2 s for this system. For the A803 dose-response study, each A803 concentration was applied for 6 min in a sequential manner from low to high concentration. This method was required because of the poor recovery from A803 block in our recordings. For the CdCl2 dose-response study, each CdCl2 concentration was applied for 2 min and the current was allowed to fully recover between each CdCl2 dose.

Measurement of ionic currents.

DiI-labeled DRG neurons (see Figs. 4 and 6) were identified with a Nikon Diaphot microscope with epifluorescence and were voltage-clamped with the whole cell configuration of the patch-clamp technique. Pipettes were pulled from Schott 8250 glass (King Precision Glass, Claremont, CA) on a Sutter P-97 puller (Sutter Instruments, Novato, CA). Currents were recorded with an Axopatch 200A amplifier (Molecular Devices, Sunnyvale, CA) and digitized with an ITC-18 data acquisition interface (Instrutech, Port Washington, NY). Experiments were controlled by a Power Macintosh computer (Apple Computer, Cupertino, CA) running S5 data acquisition software written by Dr. Stephen Ikeda (National Institute on Alcohol Abuse and Alcoholism, Bethesda, MD). Leak current was subtracted online with a −P/4 protocol. Recordings were carried out at room temperature, and the holding potential (HP) was −80 mV unless otherwise noted.

Data analysis.

Data were analyzed with IgorPro (WaveMetrics, Lake Oswego, OR) running on a Macintosh computer. Cell diameter was calculated with the cell capacitance as measured by the Axopatch circuitry and assuming that the neuron was a sphere (Harper and Lawson 1985) with a specific membrane capacitance of 1 μF/μm2 (Ho and O'Leary 2011). Dose-response data were fit by using Hill's equation to obtain the IC50. Group data are calculated as means ± SD throughout this report. Student's t-test was calculated to determine significant differences (P < 0.05).

Immunocytochemistry.

Neurons were fixed with 4% formaldehyde and permeabilized with 2% Tween 20 in phosphate-buffered saline (PBS); normal goat serum was used for blocking. Neurons were incubated overnight with primary antibody for NaV1.8 (mouse, Abcam, Cambridge, MA) or NaV1.1, NaV1.6, NaV1.7, or NaV1.9 (rabbit, Alamone Labs, Jerusalem, Israel) (1:500). Control neurons were incubated overnight in blocking solution lacking NaV primary antibodies. Neurons were washed and incubated with the appropriate secondary antibody, which was Alexa Fluor 488 IgG goat anti-rabbit or Alexa Fluor 635 IgG goat anti-mouse (Invitrogen, Carlsbad, CA). Neurons were visualized and images captured with a Nikon ECLIPSE 80i epifluorescence microscope. Fluorescence intensity and cell size were measured from individual neurons with ImageJ (rsbweb.nih.gov/ij/index.html). Cell size was calculated as cross-sectional area, and diameter (d) was calculated as d=2×(area/) (Harper and Lawson 1985). Fluorescence intensity values were displayed as histograms in IgorPro and fit by using a Gaussian distribution to obtain the mean and SD. The threshold for positive labeling of test neurons was the control mean intensity + 2 × control SD.

Chemicals.

DiI, MEM, FBS, and penicillin-streptomycin were from Invitrogen. Tetrodotoxin citrate and A803467 were from Ascent Scientific (Princeton, NJ). All other chemicals were from Sigma (St. Louis, MO).

RESULTS

TTX sensitivity in muscle afferent neurons.

DiI-labeled muscle afferent neurons were selected for recording (see example in Fig. 4). As a first step toward identifying the NaV currents expressed in these neurons, we measured the effect of 300 nM TTX (Fig. 1C). This concentration will block all TTX-S channels without affecting TTX-R channels (Blair and Bean 2002, 2003; Catterall et al. 2005; Elliott and Elliott 1993; Ho and O'Leary 2011; Ogata and Tatebayashi 1993; Renganathan et al. 2001; Roy and Narahashi 1992). Using HP of −80 mV, we found only a few muscle afferent neurons (6/43) that showed strong block (>50%) by TTX. These neurons had cell diameters ranging from 35 to 47 μm, and we have termed these cells TTX-S. The NaV currents from these cells tended to activate and inactivate rapidly (Fig. 1A), with an average rise time (10–90% of maximum current) of 0.25 ± 0.22 ms and a fast inactivation time constant (τ) of 1.1 ± 1.3 ms (mean ± SD, n = 6). The other neurons (37/43) showed a much smaller NaV current block by TTX (<30%) at a HP of −80 mV, and we termed these cells TTX-R (Fig. 1C). The rise time of the NaV current in TTX-R neurons (Fig. 1B) was significantly larger than that of TTX-S neurons, with the 10–90% rise time = 0.63 ± 0.13 ms (P < 0.05 relative to TTX-S). The speed of inactivation was also significantly slower compared with TTX-S current, with a fast inactivation τ = 3.2 ± 1.6 ms (mean ± SD, n = 37, P < 0.05) at 10 mV.

Fig. 1.

Cell size distribution of tetrodotoxin-sensitive (TTX-S) and -resistant (TTX-R) neurons. A: example of voltage-dependent sodium (NaV) current from a TTX-S muscle afferent neuron. B: example of NaV current from a TTX-R muscle afferent neuron. C: % block of peak NaV current by 300 nM TTX plotted vs. cell diameter (calculated as described in materials and methods) for muscle afferent neurons. D: % block of peak NaV current by 300 nM TTX plotted vs. cell diameter (calculated as described in materials and methods) for cutaneous afferent neurons.

The diameter of the TTX-R muscle afferent neurons had a range of 27–50 μm, which spans that of the TTX-S neurons (35–47 μm; Fig. 1C). The majority of neurons (54%) had a diameter between 30 and 40 μm, while 19% had diameters in the 20–30 μm range and 27% had diameters in the 40–50 μm range. This distribution matches that found by Hu and McLachlan (2003), where the majority of muscle afferent neurons had diameters in the 30–40 μm range (∼38%), with fewer in the 20–30 (∼22%) and 40–50 (∼23%) μm ranges.

For comparison, we examined cutaneous afferent neurons (see materials and methods) to determine the NaV current TTX sensitivity (Fig. 1D). Of the 13 neurons recorded (HP −80 mV) only 1 showed strong TTX sensitivity, while TTX block of NaV current was <20% in the other 12 cutaneous afferent neurons (Fig. 1D). The diameter of TTX-R cutaneous afferent neurons ranged from 22 to 48 μm, with the majority of neurons having diameters <35 μm. The diameter of the single TTX-S neuron was 22 μm. Figure 1D shows that the neuronal diameters of the cutaneous afferents tended to be smaller than those of the muscle afferents (Fig. 1C) from which we recorded. The smaller diameter of cutaneous versus muscle afferent neurons is consistent with previous findings (Hu and McLachlan 2003).

TTX-S muscle afferent neurons.

The NaV current in TTX-S muscle afferent neurons (n = 6) was further examined by comparing TTX block at different times during the 25-ms step. The peak current was blocked by 96 ± 1.5% in response to 300 nM TTX, while the current at the end of the 25-ms step was not significantly affected (1.2 ± 1.4%) (Fig. 2, A and C). The complete block of the inactivating NaV current in these neurons demonstrates that 300 nM TTX blocks all TTX-S channels. The time course data show that TTX rapidly blocked peak current with recovery over a washout period of 10 min (Fig. 2B). We would have liked to investigate the source for the small TTX-R sustained current in these cells, but the low frequency of encountering such neurons prevented a more detailed study. However, the absence of inactivation over the 25-ms step suggests that NaV1.8 is not the source for this TTX-R current.

Fig. 2.

Differential block of peak vs. end current by TTX. A: TTX-S currents shown before (gray, Cntl) and during (black, TTX) application of 300 nM TTX and upon washout (gray, WO). Voltage protocol is shown at bottom. B: blocking time course of peak (black circles, Peak) and end (white squares, End) current by 300 nM TTX. Asterisks indicate the representative currents shown in A. C: mean % block (±SD) of peak and end current by 300 nM TTX (P < 0.05) in 6 TTX-S neurons.

A803467 blocked current in TTX-R muscle afferent neurons.

TTX-R NaV1.8 channels are expressed in DRG neurons (Bulaj et al. 2006; Jarvis et al. 2007), and we hypothesized that a large fraction of the NaV current in our TTX-R muscle afferent neurons was generated by NaV1.8 channel activity. This hypothesis was supported by the slower activation and inactivation kinetics of the NaV current in TTX-R versus TTX-S neurons (Abdulla and Smith 2002; Blair and Bean 2002). We further tested this hypothesis by examining the effect of the NaV1.8 blocker A803 (Jarvis et al. 2007). A concentration of 300 nM A803 more effectively blocked peak NaV current (41 ± 15%) while inducing a smaller block of current at the end of the 25-ms step (22 ± 9%) in neurons (n = 22) exposed for at least 6 min (Fig. 3, A and C). The block of peak current by 300 nM TTX was only 6 ± 7% in the same 22 neurons, which shows that these muscle afferent neurons belonged to the TTX-R group (Fig. 3C). For the TTX-R cutaneous afferent neurons, NaV current block was 2 ± 4% by 300 nM TTX and 65 ± 12% by 300 nM A803 (n = 12). This A803 block is larger than that reported above for muscle afferent neurons but overlaps the A803 block of muscle afferent NaV current obtained in a separate set of experiments (see below).

Fig. 3.

TTX-R current is blocked by the NaV1.8 blocker A803467 (A803). A: TTX-R currents are shown before (gray, Cntl) and during (black, A803) application of 300 nM A803. Voltage protocol is shown at bottom. B: blocking time course of peak (black circles, Peak) and end (white squares, End) current by 300 nM TTX and A803. Note the absence of current recovery after washout of A803. Asterisks indicate the representative currents shown in A. C: mean % block (± SD, n = 22) of peak and end current by 300 nM TTX and 300 nM A803. D: dose-response curve for A803 is shown as % block of peak current vs. A803 concentration (30–1,000 nM). The curve was fitted with the Hill equation to yield the IC50.

The time course for 300 nM A803 block of NaV current in muscle afferent neurons was slow (Fig. 3B), with the average blocking τ = 4 ± 3 min (mean ± SD, n = 22). Unfortunately, we were not able to obtain recovery of NaV current during washout of A803 (Fig. 3B). The average recovery was only 13 ± 9% (n = 16) at 6 min after return to control external solution. Thus A803 was poorly reversible in our recordings.

The absence of recovery of NaV current from A803 forced us to use a cumulative method to assess the A803 dose-response relationship, in which concentrations were added in increasing order. The slow time course of block also complicated this measurement since we wanted to obtain data from multiple doses in the same neuron. We used an isochronic measurement technique in which the effect of each concentration was measured at the end of a 6-min application. These two methods have been previously combined to successfully assess the dose-response relationship of a slow ion channel blocker (Boland et al. 1994). TTX-R muscle afferents were exposed to 30, 100, 300, and 1,000 nM A803 (n = 12–22). Fractional block was plotted against A803 concentration and fitted with the Hill equation (Fig. 3D), which yielded an IC50 of 186 nM, a Hill coefficient of 0.7, and a 65% maximum block (Fig. 3D). Our IC50 of 186 nM is close to the previously published IC50 (140 nM) determined from rat DRG neurons (Jarvis et al. 2007), which supports the hypothesis that A803 is blocking NaV1.8 channels in our experiments. The presence of NaV1.8 channels in the muscle afferents was further investigated by using immunocytochemistry to gauge the expression of these channels (Fig. 4). Consistent with our recordings, the majority of identified muscle afferents (21/28, 75%) were stained with the NaV1.8 antibody. The distribution of cell diameters is shown for muscle afferent neurons that were positively (Fig. 4, top left) and negatively (Fig. 4, bottom left) stained by the NaV1.8 antibody. The unstained muscle afferents tended to have smaller diameters than the labeled muscle afferent neurons, with the majority of unlabeled neurons having diameters <30 μm. The NaV1.8-labeled muscle afferent neurons had diameters >20 μm with peak numbers in the 30–40 μm range, which fits with our electrophysiology data. Interestingly, even very large muscle afferent neurons (diameter > 40 μm) were stained with the NaV1.8 antibody, which also supports our electrophysiological findings.

Fig. 4.

Identified muscle afferent neurons are labeled by a NaV1.8 channel antibody. Top (NaV1.8): bright field (left), DiI (center), and NaV1.8 antibody labeling (right) images from same cells. Bottom (Cntl): bright field (right), DiI (center) and NaV1.8 negative control (right) images. Scale bar, 50 μm. Bar graphs show cell diameter distribution of the muscle afferent neurons stained with the NaV1.8 antibody (top) and the muscle afferent neurons that were not stained by the NaV1.8 antibody (bottom).

Recovery of TTX-S NaV channels at hyperpolarized voltages.

The absence of TTX-S currents was puzzling since previous studies have demonstrated strong block of muscle afferent transmission by TTX (Tsuchimochi et al. 2011). It has previously been demonstrated in DRG neurons that TTX-S NaV current can inactivate at voltages hyperpolarized to TTX-R channels (Leffler et al. 2002), and hyperpolarizing prepulses have been used to reveal TTX-S currents (Hong et al. 2004). We tested the idea that the TTX-S currents in our muscle afferent neurons were largely inactivated at our HP of −80 mV by hyperpolarizing the HP to −100 mV and retesting with 300 nM TTX. This hyperpolarization recovered a current that tended to activate more rapidly than the TTX-R current and was completely inactivated by the end of the 25-ms voltage step (Fig. 5A). Thus the amplitude of peak current was increased 40 ± 45% by hyperpolarizing the HP from −80 to −100 mV (n = 5), but current at the end of the step was not changed (12 ± 21%). From a HP of −80 mV, 300 nM TTX induced only a small block of peak current (4 ± 2%), but that block was significantly increased to 35 ± 15% (P < 0.05) upon changing the HP to −100 mV in the same five muscle afferent neurons (Fig. 5B). The diameters of these neurons ranged from 27 to 40 μm (mean 32 μm). TTX-S NaV channels are functionally expressed along with TTX-R channels in muscle afferent neurons but are largely inactivated at a HP of −80 mV.

Fig. 5.

Hyperpolarized holding potential (HP) reveals TTX-S NaV currents in TTX-R neurons. A: currents were recorded during voltage steps to +10 mV from HPs of −80 mV (gray, HP −80) and −100 mV (black, HP −100) and from HP of −100 mV in 300 nM TTX (black, HP −100 TTX). Note the similarity between NaV currents from HP −80 mV and HP −100 mV in TTX. Voltage protocol is shown at bottom. B: mean % block (±SD, n = 5) of peak and end current by 300 nM TTX at HP −80 mV and −100 mV. The only significant block by TTX was peak current from HP −100 mV (P < 0.05).

We were interested to determine which TTX-S NaV channels were expressed in muscle afferent neurons. Previous work has demonstrated that NaV1.1, 1.6, and 1.7 are the major TTX-S channels expressed in postnatal sensory neurons (Ho and O'Leary 2011; Lai et al. 2003; Rush et al. 2007). We found that muscle afferent neurons from adult rats were stained with antibodies for NaV1.6 and NaV1.7 (Fig. 6). For the DiI-positive muscle afferent neurons, 8 of 10 were positive for NaV1.6 staining and 5 of 7 were positive for NaV1.7 staining (separate experiments). The eight NaV1.6-positive neurons had diameters ranging from 23 to 35 μm, while the two NaV1.6-negative neurons had diameters of 34 and 56 μm. For NaV1.7, the diameters of the five positive neurons ranged between 26 and 36 μm, while the two negative neurons were 33 and 34 μm. In contrast to NaV1.6 and NaV1.7, none of the muscle afferent neurons (29–37 μm; 0/6) was positively stained by the NaV1.1 antibody (Fig. 7). Thus NaV1.6 and NaV1.7 appear to be the dominant TTX-S channels expressed in small to medium-sized muscle afferent neurons.

Fig. 6.

Identified muscle afferent neurons are labeled by NaV1.6 and NaV1.7 antibodies. Top row (NaV1.6) shows bright field (left), DiI (center), and NaV1.6 antibody labeling (right) images from the same cells, while second row (Cont) shows bright field (right), DiI (center), and NaV1.6 negative control (right) images. Third row (NaV1.7) shows bright field (left), DiI (center), and NaV1.7 antibody labeling (right) images, while bottom row (Cntl) shows bright field (right), DiI (center), and NaV1.7 negative control (right) images. Scale bars, 50 μm.

Fig. 7.

Identified muscle afferent neurons do not appear to express NaV1.1 channels. Top (NaV1.1): bright field (left), DiI (center), and NaV1.1 antibody labeling (right) images from the same cells. Bottom (Cntl): shows bright field (right), DiI (center), and NaV1.8 negative control (right) images. Scale bar, 50 μm.

A component of TTX-R current is A803 resistant.

The peak TTX-R current was blocked on average by ∼40% by 300 nM A803, while the current at the end of the voltage step was reduced by only ∼20% (Fig. 3C). A803 is relatively specific for NaV1.8 channels (Jarvis et al. 2007), which suggests that a small number of NaV1.8 channels remain active at the end of the 25-ms step. Since this end current is completely TTX insensitive, we hypothesized that at least part of the A803-insensitive current was generated by NaV1.9 channel activity. Consistent with this idea, the NaV1.9 currents are slowly inactivating (Coste et al. 2007) and the A803 resistant current in muscle afferent neurons is largely sustained over the 25-ms voltage step (Fig. 3A). As a first test of this idea, we examined the Na+ sensitivity of TTX-R current by switching the external Na+ concentration from 50 to 0 mM (in the presence of 300 nM TTX), which blocked peak current by nearly 100% but blocked end current by only ∼70% (Fig. 8, A and B). Thus NaV channels appear to generate only ∼70% of the sustained current. In separate set of experiments, we tested the sodium sensitivity of the current in the presence of 300 nM TTX and 300 nM A803. TTX blocked peak NaV current by 5 ± 4%, while A803 blocked current by 65 ± 5% in this set of experiments (n = 4). In the zero-Na+ external solution, the TTX-A803-resistant current was blocked by 87 ± 8% at peak and by 37 ± 11% (n = 4) at the end of the step (Fig. 8C). The apparent incomplete block of NaV1.8 channels by A803 (Fig. 3, C and D) complicates the determination of which NaV channel is the source for this Na+-sensitive current. However, it is likely that part of the sustained TTX-A803-resistant current is generated by NaV1.9 channel activity. Further support was obtained by staining muscle afferent neurons with a NaV1.9 antibody. We found positive NaV1.9 staining in all 10 muscle afferent neurons (diameters 22–45 μm) examined (Fig. 8, D and E).

Fig. 8.

An A803-insensitive component of TTX-R current is blocked in zero Na+. A: currents in the presence of 300 nM TTX before (gray, Cntl) and during (black, Zero Na+) application of zero-Na+ external solution. Voltage protocol is shown at bottom. B: mean % block (±SD, n = 6) of peak and end TTX-R current by zero-Na+ external solution. These currents were recorded in the presence of 300 nM TTX. C: NaV current was recorded in presence of 300 nM A803 and 300 nM TTX. Mean % block (±SD, n = 4) of peak and end current by zero-Na+ external solution (P < 0.05) is shown. D: identified muscle afferent neurons labeled by a NaV1.9 channel antibody: bright field (left), DiI (center), and NaV1.9 antibody labeling (right) images from the same cells. E: images from the same control cells are shown for bright field (left), DiI (center), and NaV1.9 negative control. Scale bar, 50 μm.

The Na+-insensitive current is largely sensitive to Cd2+.

Several properties of the current observed in zero-Na+ external solution suggested that the source could be voltage-dependent calcium (CaV) channel activity, including slower current activation than that of NaV current, and the current was noninactivating over the 25-ms step (Fig. 9A). It seems likely that a small current could arise from Mn2+ permeation, since we have previously noted measurable current through CaV2.2 channels carried by 5 mM Mg2+ or 5 mM La3+ (V. Yarotskyy and K. S. Elmslie, unpublished observations). Since there are multiple CaV channels expressed in DRG neurons (Lu et al. 2010), we chose the broad-spectrum CaV channel blocker Cd2+ to test CaV channel involvement. In zero Na+, 300 μM Cd2+ blocked 80% of the current measured at the end of the 25-ms step (Fig. 9, A and B). The dose-response relationship was well fit by the Hill equation to yield a Hill coefficient of 1.1 and an IC50 of 16 μM, which is consistent with CaV channel block (Carbone et al. 1990; Lansman et al. 1986). NaV channels are blocked by Cd2+, but the IC50 is typically >100 μM (Coste et al. 2007; Ikeda et al. 1986; Sheets and Hanck 1992). When examined in 50 mM Na+ external solution, 300 μM Cd2+ blocked only a small fraction (35%) of the end current, with an IC50 of 9 μM and a Hill coefficient of 0.8 (Fig. 9, C and D), which are both similar to those obtained in zero-Na+ external solution. Thus the same contaminating CaV current appears to be present in external Na+. One surprise was that an additional component of Cd2+ block was not observed when NaV currents were recorded (50 mM Na+ vs. 0 Na+ external), since it was reported previously that NaV1.9 current was blocked by Cd2+ with an IC50 of 233 μM (Coste et al. 2007).

Fig. 9.

Current in zero-Na+ external solution is blocked by Cd2+. A: currents recorded in zero-Na+ external solution before (gray, Cntl) and during (black, Cd2+) application of 300 μM Cd2+ and upon recovery (gray, WO). Voltage protocol is shown at bottom. B: Cd2+ dose-response curve shows mean % block (±SD, n = 3–7) for concentrations ranging from 3 to 300 μM. Curve was fitted with the Hill equation to yield the IC50. C: current in 50 mM Na+ external solution before (gray, Cntl) and during (black, Cd2+) application of 300 μM Cd2+ and upon recovery (gray, WO). Voltage protocol is shown at bottom. D: Cd2+ dose-response curve shows mean % block (±SD, n = 6) for concentrations ranging from 30 to 300 μM. Curve was fitted with the Hill equation to yield the IC50.

DISCUSSION

We investigated NaV currents in identified muscle afferents to determine the dominant channel types expressed. Surprisingly, we found that TTX-S channels dominate in only ∼14% of our recorded muscle afferent neurons, while TTX-R channels supplied the dominant NaV current in the remaining 86% of neurons studied. Functional TTX-S channels were expressed in neurons expressing the TTX-R channels, but the TTX-S channels were largely inactivated at our standard HP of −80 mV. The slower rise time and inactivation of the TTX-R current were consistent with the current being generated by NaV1.8 channels (Blair and Bean 2002), which was supported by the sensitivity of this current to the NaV1.8-specific blocker A803 (Jarvis et al. 2007). In addition, the majority of muscle afferent neurons were positively stained by a NaV1.8 antibody. A portion of the TTX-R current was not sensitive to block by A803 but was blocked by removal of external Na+. This current was sustained throughout the 25-ms voltage step, which is consistent with NaV1.9 as the source for this current. Immunocytochemistry experiments confirmed that an antibody directed against NaV1.9 positively stained muscle afferent neurons. This work demonstrates that NaV1.8 and 1.9 are expressed in identified muscle afferents and suggests that these NaV channels play an important role in the excitability of the muscle afferents. NaV1.8 and NaV1.9 are highly expressed in C and Aδ nociceptors (Djouhri et al. 2003; Fang et al. 2002). These afferents from muscle are termed Group IV (C) and Group III (Aδ). On the basis of the neuronal size and expression of these NaV channels, it appears that the majority of our recorded muscle afferents were Group III and IV, which together mediate the exercise pressor reflex (Kaufman and Hayes 2002; Kaufman et al. 1983).

TTX-sensitive currents in identified muscle afferents.

The NaV current in only 14% of the recorded neurons was strongly blocked (>50%) by TTX when the HP was −80 mV, while the TTX-S current in the vast majority of muscle afferent neurons was <30% of the total NaV current. However, we were able to recover TTX-S current by hyperpolarizing the HP to −100 mV, which increased the TTX-induced block to 35% from 4% at HP = −80 mV. Thus, even though TTX-S NaV channels are expressed in muscle afferent neurons, they are mostly inactivated at HP −80 mV. The reported resting membrane potential in NaV1.8-expressing DRG neurons is between −60 and −90 mV, with most neurons between −70 and −80 mV (Blair and Bean 2002). Thus the TTX-S channels in these neurons would be strongly inactivated at physiological resting membrane potentials.

Identification of the TTX-S NaV channels expressed in muscle afferent neurons was done by immunocytochemistry. Previous work has demonstrated expression of TTX-S NaV1.1, 1.6, and 1.7 channels in adult sensory neurons (Ho and O'Leary 2011; Lai et al. 2003; Rush et al. 2007). Using antibodies directed against these channels, we found strong labeling by the NaV1.6 and NaV1.7 antibodies but no labeling of muscle afferent neurons by the NaV1.1 antibody. We would have liked to support this finding by using pharmacological agents to identify functional channels. Unfortunately, the specific blockers are not available to differentiate among TTX-S channels. However, recent work shows that the use of multiple μ-conotoxins may allow the identification of functional TTX-S channels (Wilson et al. 2011).

Two recent reports demonstrate strong suppression of the exercise pressor reflex by TTX applied either directly to the dorsal root (Tsuchimochi et al. 2011) or injected into the DRGs (Wang et al. 2011), which demonstrates that the TTX-S channels are active at least in the axons of the Group III and IV afferents. While we cannot yet explain the differences between our study and those examining axonal responsiveness, it is possible that axonal TTX-S NaV channels could be associated with different auxiliary subunits that could depolarize the inactivation vs. voltage relationship relative to those channels expressed in the soma. Previous work has demonstrated that coexpression of NaV β-subunits can shift the NaV channel inactivation-voltage relationship (Johnson and Bennett 2006; Johnson et al. 2004; Vijayaragavan et al. 2004). For example, coexpression of the TTX-S channel NaV1.7 with the NaV β1-subunit shifts voltage-dependent inactivation ∼20 mV depolarized to increase channel availability (Klugbauer et al. 1995).

Noninactivating TTX-R NaV current.

A portion of the sodium current did not inactivate during our 25-ms voltage steps and was insensitive to TTX. This current was also largely insensitive to A803. This combination of traits was consistent with NaV1.9 as the source for the noninactivating current (Coste et al. 2007; Cummins et al. 1999, 2000). Unfortunately, NaV1.9 channel-specific blockers do not exist, which prevented the pharmacological assessment of these channels. As an alternative, we assessed the Na+ sensitivity of this current and found a 70% reduction of end current in our zero-Na+ external solution, as expected for NaV1.9. The remaining sustained current was Na+ insensitive and likely resulted from the activity of CaV channels since this current was strongly blocked by Cd2+. We also used immunocytochemistry to assess expression of NaV1.9 in muscle afferent neurons and found that all of the neurons examined were immunoreactive for NaV1.9. Together these results support the functional expression of NaV1.9 in NaV1.8-expressing muscle afferent neurons.

Previous work has demonstrated NaV1.9 expression in Aδ (Group III) and C (Group IV) afferent neurons (Fang et al. 2002). The voltage-dependent and kinetic properties of NaV1.9 channels suggest that their activity contributes to resting membrane potential and action potential initiation, but the activity of these channels alone does not generate action potentials (Cummins et al. 2007; Herzog et al. 2001). Thus enhancement of NaV1.9 channel activity can increase neuronal excitability (Baker 2005), and knockdown of these channels reduces inflammation-induced hyperalgesia (Amaya et al. 2006; Lolignier et al. 2011).

NaV1.8 channels generate the main TTX-R current.

The majority of identified muscle afferents from which we recorded (27- to 50-μm diameter) expressed a TTX-R current that activated and inactivated significantly more slowly than the TTX-S currents. These properties are consistent with the identification of the source for this current as NaV1.8 channels (Blair and Bean 2002), which was further supported by the sensitivity of the current to the NaV1.8 blocker A803 (Jarvis et al. 2007). We also used immunocytochemistry to show that the majority of identified muscle afferent neurons were labeled with a NaV1.8 antibody. One surprise was that the muscle afferent neurons not labeled by the NaV1.8 antibody had diameters less than ∼30 μm, with 50% of neurons (3/6) with diameters between 25 and 30 μm not labeled. This contrasts with our electrophysiology data, since all six muscle afferent neurons with diameters in the 25–30 μm range expressed TTX-R current that was blocked by A803. One possible explanation for this discrepancy is that we did not record enough neurons in this size range to observe TTX-S neurons, but this seems unlikely if 50% of the neurons fail to express NaV1.8. Another possibility is that the smaller neurons are not permeabilized by our procedure as well as the larger neurons, which would limit antibody access to the intracellular epitope. Our procedure must permeabilize the neuron to allow antibody access, but we lose DiI labeling if there is too much permeabilization. More work is required to investigate the NaV channels expressed in these smaller muscle afferent neurons.

Both our electrophysiology and immunocytochemistry experiments found strong expression of NaV1.8 channels in medium (30–40 μm)- and large (>40 μm)-diameter muscle afferent neurons. Previous work has shown preferential expression of NaV1.8 channels in small-diameter (<35 μm) C and Aδ neurons, but these were largely cutaneous afferents (Djouhri et al. 2003) and muscle afferents have been shown to have larger diameters than cutaneous afferents (Hu and McLachlan 2003). Our electrophysiological recordings support the larger diameter for muscle afferent neurons. The majority of the cutaneous afferents from which we recorded had diameters <35 μm, while the majority of muscle afferent neurons had diameters >30 μm. Considering this and the fact that medium-sized muscle afferent neurons (30–40 μm) express both NaV1.8 and NaV1.9, it seems that the size range of Group III (Aδ) and Group IV (C) muscle afferent neurons is larger than that for cutaneous afferents.

The expression of NaV1.8 by large-diameter (>40 μm) muscle afferent fibers was surprising but was a consistent result in both our electrophysiology and immunocytochemistry data. However, there is increasing evidence that NaV1.8 channels can be expressed in a wide range of sensory neurons, including Aβ afferents. In a study of human DRG, postmortem samples (male 57–72 yr old) showed NaV1.8 antibody staining in up to 80% of large-diameter neurons (60–80 μm) and staining in afferents (Aβ fibers) innervating cutaneous Meissner's corpuscles (Coward et al. 2000). In addition, animal studies have demonstrated expression of NaV1.8 in large-diameter DRG neurons in mouse (48%, diameter >40 μm; Renganathan et al. 2000) and rat (39%, area > 1,200 μm2; Novakovic et al. 1998). This was further supported in a mouse study demonstrating NaV1.8 expression in DRG neurons expressing Neurofilament 200 (a marker of large-diameter neurons) and in sensory afferents innervating Meissner's corpuscles (Aβ fibers) (Shields et al. 2012). Thus at least some Group II (Aβ), and perhaps Group I, muscle afferent neurons appear to express NaV1.8.

A803 blocked on average only ∼50% of the peak current at the highest concentration used (1 μM). We were reluctant to use higher concentrations since previous work determined the IC50 to be 140 μM for rat NaV1.8 and showed that the block became nonselective at higher concentrations (Jarvis et al. 2007). Jarvis et al. (2007) demonstrated that block by A803 was incomplete at inactivating voltages hyperpolarized to −70 mV, which suggests that the inactivation state of NaV1.8 channels is a critical factor. Thus our −80 mV HP likely limited the maximal block we could achieve using A803. The impact of channel inactivation could also explain the large variance we found in A803 block of NaV current (e.g., Fig. 3D), since small changes in channel inactivation state could have a large impact on block. It is also likely that A803-insensitive NaV currents such as NaV1.9 and TTX-S currents further limited the maximal block (most A803 block was measured in the absence of TTX). While these factors impacted our ability to determine the full magnitude of NaV1.8 current in these neurons, it is clear that NaV1.8 channels generate a substantial fraction (50% or more) of the total NaV current in these identified muscle afferents.

NaV1.8 channels play a dominant role in action potential generation in the soma of small-diameter neurons (Blair and Bean 2002; Renganathan et al. 2001; Rush et al. 2007). Expression of NaV1.8 is strong in soma of these neurons (Djouhri et al. 2003), but expression is normally low in the axons (Gold et al. 2003; Pinto et al. 2008). This pattern of expression appears to be mimicked by muscle afferents. We demonstrated strong NaV1.8 expression in the soma of small to medium identified muscle afferents (likely Group III and IV), while TTX-S channels appear to dominate action potential propagation in Group III and IV axons (Tsuchimochi et al. 2011; Wang et al. 2011), since axonal application of TTX blocked the exercise pressor reflex. However, NaV1.8, along with NaV1.7 and NaV1.9, channels have been shown to be expressed in most cutaneous free nerve endings (presumably nociceptors), which suggests that these channels play a role in sensory transduction (Persson et al. 2010). Thus, even though NaV1.8 channels do not participate in axonal action potential propagation in nondisease states, they could play an important role in action potential initiation.

Role for NaV1.8 channels in the exercise pressor reflex.

NaV1.8 has been shown to play a role in increasing excitability of nociceptors in chronic nerve injury models of neuropathic pain (Gold et al. 2003; Roza et al. 2003; Thakor et al. 2009). In addition, A803 and another NaV1.8 blocker, μO-conotoxin MrVIB, produce analgesia in animal models of neuropathic and inflammatory pain when systemically applied, which supports a role for NaV1.8 in contributing to chronic pain (Bulaj et al. 2006; Ekberg et al. 2006; Jarvis et al. 2007). Further support comes from analgesic effects produced by knockdown of NaV1.8 with antisense RNA in animal models of neuropathic pain (Joshi et al. 2006; Ruangsri et al. 2011). Together these results suggest that chronic disease states could alter the distribution of NaV1.8 in Group III and IV afferents to increase excitability of the exercise pressor reflex. However, one model of peripheral vascular disease, 72-h muscle ischemia, failed to reduce the effect of TTX to block the exercise pressor reflex (Tsuchimochi et al. 2011). While this model does increase cardiovascular responses to muscle contraction (Tsuchimochi et al. 2010), it is possible that 72 h is insufficient to produce changes in NaV channel distribution in Group III and IV neurons. Chronic nerve injury models that do produce such changes are typically measured 7–10 days after injury (Gold et al. 2003; Jarvis et al. 2007). It would be interesting to see whether longer durations of muscle ischemia could affect changes in axonal NaV channel distributions. One possible mechanism is the release from the damaged tissues of inflammatory mediators such as tissue necrosis factor-α (TNF-α) and chemokine CCL2, which have both been shown to increase NaV1.8 channel activity in uninjured sensory neurons (Belkouch et al. 2011; He et al. 2010).

GRANTS

This work was funded by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant AR-059397 (K. S. Elmslie).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: R.R. and K.S.E. conception and design of research; R.R., S.Y.M., J.C.B., and E.K. performed experiments; R.R., S.Y.M., J.C.B., E.K., and K.S.E. analyzed data; R.R., S.Y.M., J.C.B., E.K., and K.S.E. interpreted results of experiments; R.R., S.Y.M., J.C.B., and K.S.E. prepared figures; R.R. and K.S.E. drafted manuscript; R.R., S.Y.M., J.C.B., E.K., and K.S.E. edited and revised manuscript; R.R., S.Y.M., J.C.B., E.K., and K.S.E. approved final version of manuscript.

REFERENCES

View Abstract