Abstract
Small-diameter vesicular glutamate transporter 3-lineage (Vglut3lineage) dorsal root ganglion (DRG) neurons play an important role in mechanosensation and thermal hypersensitivity; however, little is known about their intrinsic electrical properties. We therefore set out to investigate mechanisms of excitability within this population. Calcium microfluorimetry analysis of male and female mouse DRG neurons demonstrated that the cooling compound menthol selectively activates a subset of Vglut3lineage neurons. Whole-cell recordings showed that small-diameter Vglut3lineage DRG neurons fire menthol-evoked action potentials and exhibited robust, transient receptor potential melastatin 8 (TRPM8)-dependent discharges at room temperature. This heightened excitability was confirmed by current-clamp and action potential phase-plot analyses, which showed menthol-sensitive Vglut3lineage neurons to have more depolarized membrane potentials, lower firing thresholds, and higher evoked firing frequencies compared with menthol-insensitive Vglut3lineage neurons. A biophysical analysis revealed voltage-gated sodium channel (NaV) currents in menthol-sensitive Vglut3lineage neurons were resistant to entry into slow inactivation compared with menthol-insensitive neurons. Multiplex in situ hybridization showed similar distributions of tetrodotoxin (TTX)-sensitive NaVs transcripts between TRPM8-positive and -negative Vglut3lineage neurons; however, NaV1.8 transcripts, which encode TTX-resistant channels, were more prevalent in TRPM8-negative neurons. Conversely, pharmacological analyses identified distinct functional contributions of NaV subunits, with NaV1.1 driving firing in menthol-sensitive neurons, whereas other small-diameter Vglut3lineage neurons rely primarily on TTX-resistant NaV channels. Additionally, when NaV1.1 channels were blocked, the remaining NaV currents readily entered into slow inactivation in menthol-sensitive Vglut3lineage neurons. Thus, these data demonstrate that TTX-sensitive NaVs drive action potential firing in menthol-sensitive sensory neurons and contribute to their heightened excitability.
Significance Statement Somatosensensory neurons encode various sensory modalities including thermoreception, mechanoreception, nociception and itch. This report identifies a previously unknown requirement for tetrodotoxin-sensitive sodium channels in action potential firing in a discrete subpopulation of small-diameter sensory neurons that are activated by the cooling agent menthol. Together, our results provide a mechanistic understanding of factors that control intrinsic excitability in functionally distinct subsets of peripheral neurons. Furthermore, as menthol has been used for centuries as an analgesic and anti-pruritic, these findings support the viability of NaV1.1 as a therapeutic target for sensory disorders.
Introduction
Small-diameter dorsal root ganglion (DRG) neurons are sensory neurons that encode a diverse array of somatic sensations, including various forms of pain, thermosensation, itch and touch (Dubin and Patapoutian, 2010; McGlone and Reilly, 2010; Schepers and Ringkamp, 2010; Bautista et al., 2014; Liljencrantz and Olausson, 2014). This functional diversity is encompassed by small-diameter DRG neurons of the vesicular glutamate transporter 3 lineage (Vglut3lineage), which comprise approximately 15% of DRG neurons (Lou et al., 2013). For example, a subpopulation of Vglut3lineage neurons are unmyelinated, low threshold mechanoreceptors (C-LTMRs) that encode tactile stimuli (Seal et al., 2009). Furthermore, transient receptor potential melastatin (TRPM8)-expressing Vglut3lineage neurons are proposed to mediate oxaliplatin-induced cold hypersensitivity (Draxler et al., 2014). The diverse physiological processes in which these neurons have been implicated suggests they engage distinct transduction mechanisms to encode sensory information. Yet, the molecular determinants involved in transmitting electrical signals in discrete subpopulations of Vglut3lineage neurons remain poorly understood.
Following membrane depolarization, activation of voltage-gated sodium channels (NaVs) initiate action potentials. In sensory neurons, both action potential shape and discharge frequency transmit important information (Djouhri et al., 1998; Park and Dunlap, 1998; Liu et al., 2017), a concept that is clearly illustrated in small-diameter nociceptors. These neurons predominantly express tetrodotoxin (TTX)-sensitive NaV1.7 channels, as well as TTX-resistant NaV1.8 and NaV1.9 subunits. Many small-diameter, nociceptive DRG neurons exhibit a prominent TTX-resistant sodium current that produces a “shoulder” during action potential repolarization, therefore increasing action potential duration (Ritter and Mendell, 1992; Djouhri et al., 1998; Blair and Bean, 2002). The inactivation kinetics of TTX-resistant sodium currents during this shoulder likely allow for a greater contribution of high-voltage activated calcium channels, which may increase calcium entry and could be particularly relevant to neurotransmitter release at presynaptic terminals (Blair and Bean, 2002). Additionally, the kinetics of slow inactivation of TTX-resistant voltage-gated sodium channels in nociceptive neurons controls firing rate adaption in response to sustained depolarization (Blair and Bean, 2003; Choi et al., 2007). Thus, the molecular identity and biophysical properties of NaVs expressed in a given neuron impacts action potential firing and sensory coding.
Despite the contributions of small-diameter Vglut3lineage neurons to somatosensation, the NaVs that mediate action potential firing in these neurons remain unknown. Such information can provide important insights as to how developmentally related, yet functionally diverse, DRG neurons differentially engage NaVs to transmit sensory information. Accordingly, we asked whether subpopulations of small-diameter Vglut3lineage DRG neurons possess measurable differences in intrinsic excitability and, if so, whether such differences reflect the contributions of functionally distinct NaV subunits. Here, we show that small-diameter Vglut3lineage neurons activated by the cooling compound menthol exhibit heightened excitability compared with menthol-insensitive neurons, firing robust trains of TRPM8-dependent action potentials at room temperature. Furthermore, in vitro electrophysiological and pharmacological analyses revealed that unlike nociceptors, TTX-sensitive ion channels including NaV1.1, drive action potential firing and mediate excitability in these neurons.
Materials & Methods
Key Resources
Animals
Animal use was conducted according to guidelines from the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals and was approved by the Institutional Animal Care and Use Committee of Columbia University Medical Center. Mice were maintained on a 12 h light/dark cycle, and food and water was provided ad libitum. Slc17a8iCre(stock #018147, Grimes et al., 2011) and Rosa26Ai14 mice (stock #007914, Madisen et al., 2010) were obtained from Jackson Laboratories and bred to produce Slc17a8iCre;Rosa26Ai14mice. Genotyping was performed through Transnetyx. Adult Slc17a8iCre;Rosa26Ai14mice (4-16 weeks old) of either sex were used for all experiments.
DRG culture preparation
DRGs were harvested from Slc17a8iCre;Rosa26Ai14 mice and transferred to Ca2+-free, Mg2+-free Hank’s solution (HBSS, Gibco) containing the following (in mM): 137.9 NaCl, 5.3 KCl, 0.34 Na2HPO4, 0.44 K2HPO4, 5.6 glucose, 4.2 NaHCO3, 0.01% phenol red. Processes were trimmed to reduce the amount of plated non-neuronal cells. Ganglia were treated with collagenase (1.5 mg/ml, Type P, Sigma-Aldrich) in HBSS for 20 min at 37°C followed by 0.05% Trypsin-EDTA (Gibco) for 3 min with gentle rotation. Trypsin was neutralized with culture media (MEM, with L-glutamine, Phenol Red, without sodium pyruvate) supplemented with 10% horse serum (heat-inactivated, Gibco), 10 U/ml penicillin, 10 μg/ml streptomycin, MEM vitamin solution (Gibco), and B-27 supplement. Serum-containing media was decanted and cells were triturated using a fire-polished Pasteur pipette in a serum-free MEM culture media containing the supplements listed above. Cells were plated on laminin-treated (0.05 mg/ml) glass coverslips, which had previously been washed in 2N NaOH for at least 4 h, rinsed with 70% ethanol and UV-sterilized. Cells were then incubated at 37°C in 5% CO2. Cells were used for electrophysiological experiments 16-24 h after plating.
HEK cell culture and transfection
Stably transfected HEK293 cell lines expressing human NaV1.1 (Kahlig et al., 2010), NaV1.6 (Dr. Lori Isom, University of Michigan), or HEK293 cells transiently transfected with a cDNA construct containing human NaV1.7 (Dr. Manu Ben-Johny, Columbia University) were used. The NaV1.7 plasmid was sequenced following transformation and extraction (Genewiz, see Mendeley dataset). HEK cells were grown in DMEM (GIBCO 11995) containing 10% FBS (Thermo Fisher Scientific A3840101), 1% Penicillin-Streptomycin (Thermo Fisher Scientific 15-140-122). Media for HEK cells stably expressing NaV1.1 or NaV1.6 also contained 400 μg/ml G418 (Fisher Scientific 10-131-035) to select for transfected cells. A calcium phosphate protocol was used to co-transfect NaV1.7 and green fluorescent protein into HEK 293 cells. Briefly, 2 M CaCl2, cDNAs, and sterile water were mixed together and added dropwise to a 2x solution of hepes buffered saline. The final solution was added dropwise to HEK293 cells that were plated the day before on glass cover slips coated with 0.05 mg/ml laminin. Cells were incubated at 37°C in 5% CO2 with the transfection solution for 3 h, followed by two washes with sterile phosphate buffered saline and addition of new cell culture media. Electrophysiological recordings were performed 24-72 hours post transfection.
Electrophysiology
Whole-cell voltage- and current-clamp recordings made from small-diameter (capacitance ≤ 25 pF), TdTomato-expressing (Vglut3lineage) dissociated DRG neurons and HEK cells were performed with patch pipettes pulled from standard borosilicate glass (1B150F-4, World Precision Instruments) with a P-97 puller (Sutter Instruments). For neuronal recordings, patch pipettes had resistances of 3-6 MΩ when filled with an internal solution containing the following (in mM): 120 K-methylsulfonate, 10 KCl, 10 NaCl, 5 EGTA, 0.5 CaCl2, 10 HEPES, 2.5 MgATP, pH 7.2 with KOH, osmolarity 300 mOsm. For HEK cell recordings, patch pipettes had resistances of 1.5-3 MΩ when filled with an internal solution containing the following (in mM): 140 CsF, 10 NaCl, 2 MgCl2, 0.1 CaCl2, 1.1 EGTA, 10 HEPES, pH 7.2 with CsOH, osmolarity ∼310 mOsm. Seals and whole-cell configuration were obtained in an external solution containing (in mM): 145 NaCl, 5 KCl, 10 HEPES, 10 glucose, 2 CaCl2, 2 MgCl2, pH 7.3 with NaOH, osmolarity ∼320 mOsm. Series resistance was compensated by 80%. In experiments with DRG neurons where currents from voltage-gated sodium channels were recorded, after the whole-cell configuration was established and neurons were tested for sensitivity to menthol, a modified external solution was applied (in mM): 105 NaCl, 40 TEA-Cl, 10 HEPES, 2 BaCl2, 13 glucose, 0.03 CdCl2, pH 7.3 with NaOH, osmolarity ∼320 mOsm. This modified solution was used in HEK cell recordings following seal acquisition in a standard external solution. All solutions used were allowed to warm to ambient temperature before each experiment to ensure all recordings were made at room temperature (20°C-23°C). After each experiment, the recording chamber was thoroughly cleaned with Milli-Q water. For experiments using NaV inhibitors, drugs were applied to neurons for 1 min in current-clamp mode in the absence of injected current; therefore, cells were recorded at their intrinsic Vm.
Data acquisition and analysis
Currents and voltages were acquired and analyzed using pClamp software (version 10, Molecular Devices). Recordings were obtained using an Axopatch 200b patch-clamp amplifier and a Digidata 1440A, and filtered at 5 kHz and digitized at 10 kHz. Analysis was performed using Clampfit 10 (Molecular Devices). All voltages were corrected for the measured liquid junction potential (−7 mV) between internal and external recording solutions. Phase plots were constructed from the first derivative of the somatic membrane potential (dV/dT) versus the instantaneous somatic membrane potential. Action potential threshold was calculated as the membrane potential at which the phase plot slope reached 10 mV ms-1 (Kress et al., 2008; Yu et al., 2008). Duration at the base was calculated by measuring the duration of the action potential starting at the resting membrane potential (Vm) and ending when the repolarization phase again passes the initial Vm. Following determination of menthol sensitivity in gap-free recording mode, some cells were not further analyzed using phase plot analysis due to low digital gain during recordings (n=3) or deteriorating cell health (n=2).
Pharmacology
TTX was from Abcam. PF 05089771 and ICA 121431 were from Tocris. PN3a was a generous gift from Dr. Irina Vetter (Institute for Molecular Biosciences, University of Queensland). All other chemicals were from Sigma-Aldrich.
Calcium imaging
Dissociated DRG neurons were loaded for 45 min with 10 mM Fura-2AM (Invitrogen), supplemented with 0.01% Pluronic F-127 (wt/vol, Invitrogen), in external solution. Images were acquired using MetaMorph software (version 7) and displayed as the ratio of 340 nm to 380 nm. Neurons were identified by eliciting calcium responses with a high potassium solution (140 mM) at the end of each experiment. Neurons were considered sensitive to an agonist if the average ratio during the 30 s following agonist application was ≥ 15% above baseline. Image analysis was performed using custom MATLAB scripts.
Multiplex in situ hybridization
DRG sections cut at 25-µm thickness were processed for RNA in situ detection using an RNAscope Fluorescent Detection Kit according to manufacturer’s instructions (Advanced Cell Diagnostics, Hayward, California, USA) with the following modifications: upon harvesting, DRG were fixed in 4% paraformaldehyde for 15 min and then incubated in 30% sucrose for 2 h at 4°C. DRG were embedded in Optimal Cutting Temperature Compound (Sakura) and stored at - 80°C until sectioned. The following RNAscope probes were used: Trpm8 (420451-C3, mouse), Scn1a (434181-C2, mouse), Scn8a (434191-C2, mouse), Scn9a (313341-C1, mouse), and Scn10a (426011-C2, mouse). In situ hybridization was followed by incubation at 4°C overnight with a rabbit anti-dsRed (1:3000, Clontech: 632475) primary antibody. Sections were then incubated at room temperature for 1 h with a goat anti-rabbit AlexaFluor 594-conjugated secondary antibody (Thermo Fisher Scientific, A-11012). Samples were mounted with Fluoromount-G (Fisher Scientific). Specimens were imaged in three dimensions (1-µm axial steps) on a Zeiss Exciter confocal microscope (LSM 5) equipped with a 40X, 1.3 NA objective lens. Images were analyzed using ImageJ software. Neurons considered positive for a given NaV subunit had signal equal to or greater than one standard deviation above background.
Experimental design and statistical analysis
Summary data are presented as mean ± standard deviation from n cells. For quantitative analysis of in situ hybridization data, at least three biological replicates per condition were used and the investigator was blinded to NaV subunit for analysis. Statistical differences between menthol-sensitive and menthol-insensitive populations were assessed using an unpaired Student’s t test (two-tailed) for normally distributed datasets. A Mann Whitney test was used for populations that did not conform to Gaussian distributions or had different variances. To estimate IC50 values for NaV antagonists, inhibitor versus response curves were fit with the following relation: Normalized current=100/(1+[Inhibitor]/IC50). Kinetic data were fit with single or double-exponential relations. The voltage-dependence of slow inactivation was fit with the Boltzmann equation: Fraction available=Minimum+[(Maximum-Minimum)/(1+exp(V50-Vm)/k)], where V50 denotes the membrane potential at which half the channels are inactivated and k denotes the Boltzmann constant/slope factor. Differences between fits were assessed with an Extra sum-of-squares F test. Statistical tests and fit parameters are listed in the Results and/or figure legends. Statistical significance in each case is denoted as follows: *P < 0.05, **P < 0.01; ***P < 0.001, and ****P < 0.0001. Statistical tests and curve fits were performed using Prism 7.0 (GraphPad Software).
Results
Menthol-sensitivity is restricted to Vglut3lineage DRG neurons
Vglut3lineage sensory neurons are a heterogeneous population. To identify functionally distinct subpopulations within this group, we tested the responsiveness of Vglut3lineage neurons to capsaicin, chloroquine and menthol (Figure 1, Figure 1-1), which activate nociceptors, pruritoceptors and cold receptors, respectively. We performed calcium microfluorimetry while applying various chemosensory stimuli to acutely cultured DRG neurons (<24 h) harvested from adult male and female Slc17a8iCre;Rosa26Ai14 mice. In these mice, neurons that express Vglut3 at any point during development are labeled with a TdTomato fluorescent reporter (Figure 1A). Neurons were identified by robust calcium responses to high-potassium depolarization (784 total DRG neurons, 331 Vglut3lineage, 453 non-Vglut3lineage, n=3 mice; Figure 1 B–D). Approximately 8% of Vglut3lineage neurons were activated by the TRPM8 agonist menthol, whereas no non-Vglut3lineage neurons responded to the compound (Figure 1E). Conversely, both populations contained neurons that were activated by the TRP vanilloid 1 (TRPV1) agonist capsaicin; however, comparatively fewer Vglut3lineage neurons were capsaicin-sensitive compared with non-Vglut3lineage neurons (∼9% vs. ∼46%, respectively). Few neurons of either group responded to both menthol and capsaicin (5/322 Vglut3lineage neurons), or to chloroquine (1/322 Vglut3lineage neurons and 4/460 non-Vglut3lineage neurons), a pruritogen that signals through MrgprA3 and TRP ankyrin 1 (TRPA1; Wilson et al., 2011). Interestingly, a comparison of un-normalized baseline fura-2 ratios showed that menthol-sensitive neurons had slightly elevated baseline calcium signals compared to menthol-insensitive neurons (F340/F380 = 0.52 ± 0.05 vs. 0.45 ± 0.03, n = 5 coverslips, P = 0.029, unpaired Student’s t test, two-tailed, Figure 1F). This analysis builds upon prior work (Draxler et al., 2014) by demonstrating that menthol sensitivity is restricted to the Vglut3lineage population.
The majority of menthol-sensitive DRG neurons have small somata and give rise to unmyelinated axons (Takashima et al., 2007; Dhaka et al., 2008). Thus, we targeted small-diameter, Vglut3lineage neurons with a membrane capacitance (Cm) of ≤ 25 pF for functional analysis. Neurons that did not meet these criteria were not analyzed further by electrophysiology. Using gap-free current-clamp recordings, we asked whether these neurons fire action potentials in response to menthol application (100 μM). Half of small-diameter Vglut3lineage neurons (31/62 neurons) fired trains of action potentials in response to menthol application (Figure 2A). A subset of neurons were subsequently exposed to 1 mM menthol, which activates TRPM8 ion channels but inhibits TRPA1 (Karashima et al., 2007; Xiao et al., 2008). All neurons examined showed a dose-dependent increase in menthol-evoked firing rates (Figure 2B), suggesting that menthol elicits firing through TRPM8 rather than TRPA1 in Vglut3lineage DRG neurons. We noted that menthol-sensitive neurons were among the smallest DRG neurons in vitro, whereas menthol-insensitive neurons were more varied in size (Figure 2C). Consistent with this observation, the distribution of Cm among these menthol-sensitive neurons was well fit by a single Gaussian distribution (R2 = 0.986; 8.1 ± 2.9 pF; Figure 2D). Conversely, menthol-insensitive neurons were better fit by a double Gaussian distribution, (R2 = 0.818), with the two populations having means of 8.7 ± 3.4 and 20.6 ± 2.0 pF (Figure 2D). These data suggest that menthol-sensitive Vglut3lineage neurons are a more homogenous subpopulation compared with menthol-insensitive Vglut3lineage neurons.
Menthol-sensitive Vglut3lineage neurons fire robustly at room temperature
We next asked whether intrinsic excitability properties differed between menthol-sensitive and insensitive Vglut3lineage neurons. During gap-free recordings, we noted that 87% (27/31) of menthol-sensitive Vglut3lineage neurons exhibited unusually robust action potential firing prior to menthol application. Two firing patterns were observed, sustained and phasic firing (Figure 2E-F). Of the menthol-sensitive neurons that exhibited non-evoked activity, 44% exhibited phasic firing, whereas 56% maintained sustained firing during gap-free recordings. Menthol-sensitive neurons with sustained action potential discharges showed higher average firing frequencies compared with burst firing frequencies of phasic neurons (Figure 2F). By contrast, few menthol-insensitive neurons exhibited non-evoked firing during gap-free recordings (4/31), and these produced only occasional action potentials. This ongoing activity in menthol-sensitive neurons is consistent with the elevated baseline flura-2 fluorescence observed during calcium microfluorimetry experiments (Figure 1F), as well as ex vivo data showing sustained firing upon cold or menthol-stimulation of TRPM8-expressing DRG neuron receptive fields (Jankowski et al., 2017). Together, these results suggest that menthol-sensitive Vglut3lineage population have heightened excitability under our in vitro recording conditions and that, within the menthol-sensitive Vglut3lineage population, firing properties vary.
The ability of menthol-sensitive neurons to fire robustly at room temperature, an activating stimulus for TRPM8 (McKemy et al., 2002; Andersson et al., 2004; Tajino et al., 2011; Fujita et al., 2013; Morenilla-Palao et al., 2014; Jankowski et al., 2017; Pertusa et al., 2018), led us to ask whether this firing was dependent upon TRPM8 ion channels. We applied the selective inhibitor, PBMC (25 nM, Knowlton et al., 2011) to menthol-sensitive neurons and analyzed its effect on action potential firing at room temperature. Whereas vehicle application did not have a significant effect on firing rates, PBMC drastically reduced action potential firing at room temperature within 2 min of drug application (n = 3 neurons per group, P = 0.0035, unpaired Student’s t test, Figure 2G-H). Thus, activation of TRPM8 ion channels mediates robust ongoing action potential firing in menthol-sensitive Vglut3lineage neurons.
Collectively, these data demonstrate that menthol-sensitive neurons are highly excitable subpopulation of small-diameter Vglut3lineage DRG neurons, capable of maintaining sustained action potential firing in vitro at room temperature.
Intrinsic excitability differs between Vglut3-lineage menthol-sensitive and menthol-insensitive neurons
To investigate the heightened intrinsic excitability in menthol-sensitive DRG neurons, we compared responses of menthol-sensitive and -insensitive Vglut3lineage neurons to 500 ms current injections using phase plot analysis (Figure 3A-B). The threshold for action potential firing in menthol-sensitive neurons was significantly hyperpolarized (−28.5 ± 6.6 mV, n = 28) compared with menthol-insensitive neurons (−22.2 ± 10.4 mV, n = 31, P = 0.0269, unpaired Student’s t test; Figure 3C). Menthol-sensitive neurons also fired more action potentials in response to a current injection of 50 pA (Figure 3D). Action potential duration at the base (see Methods) was significantly shorter in menthol-sensitive neurons compared with menthol-insensitive neurons (Figure 3E). Interestingly, in gap-free recordings prior to menthol application, menthol-sensitive neurons had significantly more depolarized membrane potentials (Vm) than menthol-insensitive neurons (−45.5 ± 4.4 mV vs. −51.2 ± 5.8 mV, n = 31 for each group, P < 0.0001, Mann Whitney test, Figure 3F). Thus, menthol-sensitive Vglut3lineage neurons maintain a Vm that more closely borders action potential threshold as compared with menthol-insensitive Vglut3lineage neurons. Conversely, action potential amplitude, and membrane voltage sag did not differ between the two populations (Figure 3G-H). Together, these data provide evidence that menthol-sensitive Vglut3lineage neurons have more excitable membrane properties than menthol-insensitive Vglut3lineage neurons.
Interestingly, consistent with their longer duration action potentials, 44% of menthol-insensitive neurons had a pronounced “shoulder” during the repolarization phase of the action potential (Figure 3A-B). This shoulder was completely absent in menthol-sensitive neurons. The presence of a shoulder is attributed to sodium currents mediated by TTX-resistant NaV1.8 and NaV1.9 channels (Blair and Bean, 2002). Thus, a differential contribution of NaV subunits to action potential firing in these two populations of Vglut3lineage neurons could underlie the observed differences in excitability.
Differences in NaV current slow inactivation kinetics in small-diameter Vglut3lineage neurons
NaV slow inactivation has been linked to adaptation of action potential firing in small-diameter DRG neurons, whereby sequestration of tetrodotoxin-resistant NaVs subunits NaV1.8 and NaV1.9 in the slow inactivated state restricts the duration of action potential discharges in response to sustained stimulation (Blair and Bean, 2003). As we found that menthol-sensitive Vglut3lineage neurons are capable of maintaining prolonged action potential discharges for several minutes in vitro (Figure 2) and fire robustly in response to current injection (Figure 3), we hypothesized that NaV slow inactivated states are unstable in menthol-sensitive Vglut3lineage neurons. To test this model, we first measured NaV entry into slow inactivation by delivering a conditioning pulse from −100 mV to 0 mV for 50–1600 ms between 3-ms test steps to −20 mV (Figure 4A). Consistent with our hypothesis, entry of NaV currents into the slow inactive state was almost fourfold slower in menthol-sensitive neurons compared with menthol-insensitive neurons (τ = 1485 ms, n = 6 vs. τ = 376.5 ms, n = 5; P < 0.0001, Extra sum-of-squares F test.) Notably, after a 1600 ms conditioning pulse, ∼65% of the initial NaV current in menthol-sensitive neurons was still present, whereas only ∼21% of the current remained in menthol-insensitive neurons. Thus, NaV currents in menthol-sensitive neurons are resistant to slow inactivation.
We next analyzed recovery from slow inactivation by delivering a 3-ms test pulse to −20 mV, followed by a 1-s conditioning step from −100 mV to 0 mV, and a second 3-ms test pulse given at recovery intervals of increasing duration. Recovery time constants from both populations were well-fit with double-exponential functions. Consistent with the slow inactivated state being unstable in menthol-sensitive Vglut3lineage neurons, NaV currents in these cells recovered faster from slow inactivation than those in menthol-insensitive neurons (Figure 4B); however, note that the 1-s conditioning pulse did not drive all channels into the slow inactivated state. Sodium currents in menthol-sensitive Vglut3lineage neurons recovered from slow inactivation with an average weighted time constant of 244.3 ms, whereas menthol-insensitive Vglut3lineage neurons recovered slower, with an average weighted time constant of 311.2 ms (P < 0.0001, Extra sum-of-squares F test). Indeed, after 50 ms, ∼80% of the NaV current had recovered in menthol-sensitive neurons, compared with ∼50% in menthol insensitive neurons. The steady-state voltage-dependence of slow inactivation was comparable in menthol-sensitive and -insensitive Vglut3lineage neurons (Figure 4C). These data suggest that in menthol-sensitive neurons, slow inactivated states are less stable across membrane voltages. Finally, recovery from fast inactivation was not distinguishable between menthol-sensitive and -insensitive Vglut3lineage neurons (Figure 4D).
Collectively, these data indicate that the slow inactivated state of NaVs expressed in menthol-sensitive Vglut3lineage neurons is unstable, which could explain the capacity of these neurons to sustain action potential firing for prolonged periods of time. Moreover, the kinetics of slow inactivation we obtained for NaV currents in this population of small-diameter neurons suggest they do not rely upon TTX-resistant NaVs, which readily enter into the slow inactivated state (Blair and Bean, 2003; Choi et al., 2007).
NaV expression profiles in small-diameter Vglut3lineage DRG neurons
Given the functional differences in membrane excitability and NaV currents between menthol-sensitive and -insensitive Vglut3lineage neurons, we next asked whether these two populations have distinct expression profiles of NaVα subunits, nine of which are encoded in the mammalian genome (NaV1.1-NaV1.9; Catterall, 2012). To do so, we performed single-molecule multiplex in situ hybridization experiments (Figure 5). Menthol-sensitive Vglut3lineage neurons were identified based on TRPM8 mRNA expression. Similarly sized, small-diameter Vglut3lineage neurons lacking TRPM8 expression were considered menthol-insensitive neurons.
We focused our analysis on NaV1.1, NaV1.6, NaV1.7 and NaV1.8 subunits, which are commonly found in adult murine DRG neurons (Figure 5A-D; Black et al., 1996; Ho and O’Leary, 2011). Quantification of NaV mRNA staining from 848 DRG neurons (n = 3 animals) revealed broad and comparable expression of TTX-sensitive NaV1.1, NaV1.6, and NaV1.7 subunits between TRPM8+ and TRPM8- small-diameter Vglut3lineage neurons (Figure 5E-F). Transcripts for the TTX-resistant isoform NaV1.8, although widely expressed, was lower in TRPM8+ compared with TRPM8- Vglut3lineage neurons, [59% (47/80) vs. 96% (97/101), respectively]. These data suggest that differential expression of NaVs at the mRNA level cannot account for the differences in excitability observed between menthol-sensitive and -insensitive Vglut3lineage DRG neurons.
TTX-sensitive NaVs mediate action potential firing in menthol-sensitive Vglut3lineage neurons
Considering the overlap in NaV mRNA expression between putative menthol-sensitive and -insensitive neurons, we next used a pharmacological approach to assess the complement of functional NaV isoforms in these two populations. We first asked whether evoked firing from menthol-sensitive Vglut3lineage neurons is blocked by TTX (Figure 6A). A 1-min application of TTX (0.3 or 1 µM) abolished action potential firing in menthol-sensitive Vglut3lineage neurons (Figure 6B). On the other hand, 1 µM TTX abolished action potential firing in only 3/11 menthol-insensitive Vglut3lineage neurons. The inhibitory effect of TTX on action potential firing was significantly greater in menthol-sensitive neurons compared with menthol-insensitive neurons (P = 0.0053, unpaired Student’s t test; Figure 6B). Thus, these results demonstrate that menthol-sensitive and -insensitive Vglut3lineage neurons have functionally distinct complements of NaV subunits, with TTX-sensitive channels driving action potential firing in the former and TTX-resistant channels playing a major role in spike firing in the latter.
We next aimed to dissect the specific contributions of individual TTX-sensitive NaV subunits to action potential firing in menthol-sensitive Vglut3lineage neurons. A metabolite of TTX, 4,9-anhydro-TTX (AH-TTX), has been reported to selectively block NaV1.6 channels (Rosker et al., 2007); however, its effect on NaV1.1 channels was not examined. Accordingly, we analyzed inhibition by AH-TTX of sodium currents HEK cells stably transfected with human NaV1.1 channels (Kahlig et al., 2010). The average peak amplitude for NaV1.1 currents recorded from this cell line was −2499 ± 1499 pA (n = 15). The dose-response curve obtained showed an apparent IC50 of 120.7 nM (n = 3-5 observations per concentration, Figure 6C). Indeed, there was notable block of NaV1.1 currents by 200 nM AH-TTX (Figure 6D), which is within the range of concentrations typically used to block NaV1.6 (100–300 nM, Rosker et al., 2007; Hargus et al., 2013; Barker et al., 2017). Thus, we were unable to use this reagent to examine a specific role for NaV1.6 channels in action potential firing in menthol-sensitive Vglut3lineage neurons.
A role for NaV1.1 in menthol-sensitive Vglut3lineage neurons
The TTX-sensitive channels NaV1.1 and NaV1.7 are both expressed in adult DRG neurons and have been implicated in various forms of pain processing (Cummins et al., 2004; Nassar et al., 2004; Osteen et al., 2016). Whether or not they function small-diameter Vglut3lineage DRG neurons has yet to be determined. We therefore investigated the contribution of these subunits to action potential firing in menthol-sensitive neurons.
We first tested ICA 121431, an inhibitor of the NaV1.1 and NaV1.3 channels (Figure 7A-B). NaV1.3 is expressed at only low levels in uninjured adult rat, mouse and human DRGs (Waxman et al., 1994; Felts et al., 1997; He et al., 2010; Usoskin et al., 2015; Chang et al., 2018). Thus, we used ICA 121431 as a blocker of NaV1.1 channels in adult mouse DRG preparations (McCormack et al., 2013). Application of 500 nM ICA 121431 drastically reduced action potential firing in menthol-sensitive Vglut3lineage neurons (baseline: 37.1 ± 7.8 Hz, post-ICA 121431: 4.6 ± 4.9 Hz; n = 11, Figure 7B). Action potential firing was also reduced in a subset of menthol-insensitive Vglut3lineage neurons (5.5 ± 4.4 to 2.0 ± 1.9 Hz; n = 8). Like TTX, however, the effect of ICA 121431 was significantly greater in menthol-sensitive compared with menthol-insensitive neurons (P = 0.0363, Figure 7B). The specificity of ICA 121431’s block of NaV1.1 currents was confirmed by testing inhibition of recombinant NaV1.1, NaV1.6 and NaV1.7 mediated currents (IC50 = 25.2 nM, 2.7 µM, and 2.8 µM respectively, n = 4-6 observations per concentration, Figure 7C-D). The average peak currents recorded for NaV1.6 and NaV1.7 channels were −1545 ± 1134 pA and −1147 ± 1391 pA (n = 6 and n = 9, respectively).
We next investigated the contribution of NaV1.7 channels to action potential firing in menthol-sensitive neurons using PF 05089771 (25 nM; Alexandrou et al., 2016; Theile et al., 2016), a selective blocker of this channel. PF 05089771 had little effect on mean firing rates in menthol-sensitive Vglut3lineage neurons (Figure 8A-B). Moreover, when cells were analyzed as a percentage of control firing, the effects of PF 05089771 did not differ between menthol-sensitive and menthol-insensitive neurons. A dose-response curve performed in HEK cells transiently transfected with recombinant NaV1.7 showed an apparent IC50 of PF 05089771 for inhibition of NaV1.7 of 10.7 nM (n = 4-5 observations per concentration, Figure 8C), consistent with published values (Alexandrou et al., 2016; Theile et al., 2016). Thus, at the concentration used in this study, PF 05089771 blocked roughly 70% of the NaV1.7 mediated current (Figure 8D). We also tested the spider venom toxin Pn3a (300 nM), a structurally unrelated NaV1.7 antagonist whose mechanism is distinct from that of PF 05089771 (Deuis et al., 2017). Consistent with results obtained using PF 05089771, Pn3a had no effect on action potential firing in menthol-sensitive neurons (control: 38.7 ± 5.0 Hz, after Pn3a perfusion: 34.0 ± 4.0 Hz, n = 3, Figure 8B). Together, these results demonstrate that action potential firing in menthol-sensitive Vglut3lineage neurons depends upon TTX-sensitive NaVs including NaV1.1.
NaV1.1 channels are critical determinants of entry into slow inactivation in menthol-sensitive Vglut3lineage neurons
As NaV currents in menthol-sensitive Vglut3lineage neurons are resistant slow inactivation, we next asked if this biophysical feature depended upon the activity of NaV1.1 channels. To accomplish this, we analyzed rates of entry into, and recovery from, slow inactivation in menthol-sensitive Vglut3lineage neurons in the presence of 500 nM ICA 121431. Analysis of whole-cell currents found that the ICA-sensitive component was 38.3% ± 20.2% of the total NaV current in these neurons (Figure 9A-B). In line with NaV1.1 channels being critical to the excitability of menthol-sensitive DRG neurons, we found the rate of entry into slow inactivation drastically increased in the presence of ICA 121431. The previously observed rate of 1485 ms fell to 327.7 ms when NaV1.1 channels were blocked (Figure 9C; n = 10). Conversely, the average weighted time constant of recovery from slow inactivation more than doubled [without ICA 121431: 311.2 ms, versus with ICA 121431: 686.4 ms (τ1 = 725.1 ms, τ2 = 59.8 ms), n = 6, Figure 9D]. Linear regression analysis showed a significant correlation between the magnitude of the ICA-sensitive current and the rate of entry into slow inactivation (r2 = 0.43, P = 0.04, n = 10, Figure 9E). On the other hand, the amplitude of the ICA-sensitive current did not correlate with the rate of recovery from slow inactivation (r2 = 0.37, P = 0.20, n = 6, Figure 9F). These data demonstrate a new role for NaV1.1 in setting the rate of NaV current entry into slow inactivation in sensory neurons. Collectively, our results support a role for NaV1.1 channels as key mediators of excitability in menthol-sensitive Vglut3lineage neurons.
Discussion
Small-diameter Vglut3lineage DRG neurons are a heterogeneous population that encode distinct somatic senses. This study reveals two important findings about the functional heterogeneity in such neurons. First, menthol-sensitive Vglut3lineage DRG neurons possess a unique excitability profile, which allows them to maintain prolonged spike discharges. Second, TTX-sensitive NaVs mediate action potential firing in these sensory neurons, with a notable contribution of NaV1.1. We propose that cation influx through TRPM8 ion channels produces an excitatory drive that activates NaV1.1 ion channels at room temperature. Once activated, these channels cycle through open and fast-inactivated states, with the majority of channels bypassing long-lived slow inactivated states. This is likely attributable to unique features of NaV1.1-containing macromolecular complexes in menthol-sensitive neurons, including association with auxiliary proteins or posttranslational modifications (Aman and Raman, 2007), with the end result being continuous action potential firing (Figure 10). Thus, menthol-sensitive Vglut3lineage DRG neurons represent a highly excitable population of small-diameter sensory neurons in which action potential firing depends upon TTX-sensitive NaV complexes.
Prior work has focused on the role of potassium channels as excitability breaks in in TRPM8-expressing sensory neurons. A molecular profiling study identified the TASK-3 leak potassium channel as highly enriched in TRPM8+ DRG neurons and suggested that inhibition of this channel decreases cold activation thresholds (Morenilla-Palao et al., 2014). There was only a modest effect, however, of TASK-3 genetic deletion on intrinsic excitability. H-current (Ih), a conductance mediated by hyperpolarization-activated cyclic nucleotide-gated (HCN) channels, is reportedly pronounced in cold-activated sensory neurons (Viana et al., 2002; Orio et al., 2009). Furthermore, genetic deletion of Hcn1 converts firing patterns in cold-sensing optic nerve fibers from regular to burst spiking (Orio et al., 2012). Consistent with previous studies, we observed a prominent sag ratio, a current-clamp readout of Ih, in menthol-sensitive Vglut3lineage neurons; however, our study revealed no difference in sag ratio between menthol-sensitive and -insensitive Vglut3lineage neurons. Thus, TASK-3 and HCN channels, though important for cold detection, are unlikely to mediate the differences in intrinsic excitability between these two populations. Future studies are needed to determine whether other potassium conductances, such as those mediated Kv1 (Madrid et al., 2009; Gonzalez et al., 2017a; Gonzalez et al., 2017b), contribute to differences in intrinsic excitability between menthol-sensitive and - insensitive DRG neurons.
In addition to potassium channels, previous studies investigated TTX-resistant NaVs in menthol-sensitive DRG neurons. NaV1.8 is found in ∼90% of small-diameter DRG neurons (Shields et al., 2012) and has been implicated in menthol-sensitized cold responses (Zimmermann et al., 2007). However, NaV1.8 null mice show normal physiological and behavioral responses to cold (Luiz et al., 2019). A recent study identified a subpopulation of DRG neurons that are both Vglut3lineage and NaV1.8lineage (Patil et al., 2018). These neurons possess properties to similar the menthol-sensitive neurons analyzed in the present work, including fast action potential durations, insensitivity to capsaicin, and small somata. Moreover, NaV1.9, the other TTX-resistant NaV isoform, was reported to be expressed in nociceptors that respond to cooling, as well as contribute to pain perception in response to noxious cold (Lolignier et al., 2015). In that study, however, NaV1.9 mRNA co-expressed with only ∼20% of TRPM8+ DRG neurons. The proportion of adult menthol-sensitive neurons that express functional NaV1.8 or NaV1.9 protein is unknown; nonetheless, our pharmacological studies indicate that TTX-resistant NaVs do not drive action potential firing in menthol-sensitive neurons under our experimental conditions.
Instead, we provide evidence that functionally distinct NaVs contribute to the different excitability profiles of menthol-sensitive and -insensitive Vglut3lineage DRG neurons. Although multiplex in situ hybridization data showed widespread expression of several NaV transcripts, our pharmacological analysis revealed a more restricted functional contribution, with NaV1.1 comprising over one third of the total NaV current and mediating most action potential firing in menthol-sensitive Vglut3lineage DRG neurons. Conversely, TTX-resistant channels dominated in menthol-insensitive neurons. In DRG, NaV1.1 is reported to be predominantly localized to medium-diameter neurons that mediate mechanical pain (Osteen et al., 2016). Interestingly, that study showed that ∼40% of trigeminal neurons that express functional NaV1.1 channels also exhibit menthol-evoked calcium transients. Our study extends these findings by demonstrating that action potential firing in menthol-sensitive Vglut3lineage DRG neurons is dependent upon TTX-sensitive NaVs, with the NaV1.1/NaV1.3 antagonist ICA 121431 dramatically reducing firing rates. Furthermore, while TTX-sensitive NaV1.7 channels are important to the function of small-diameter nociceptors and pain signaling (Cox et al., 2006; Minett et al., 2012; Yang et al., 2018), these channels are likely inactivated at the resting membrane potential of menthol-sensitive neurons. Indeed, the V1/2 of inactivation of NaV1.7 is roughly −75 mV (Alexandrou et al., 2016). Conversely, the V1/2 inactivation of NaV1.1 is approximately −17 mV (Aman et al., 2009), a membrane potential that is much more depolarized than the resting potential of menthol-sensitive Vglut3lineage DRG neurons in our study. Thus, we have identified a new role for TTX-sensitive NaV1.1 channels in action potential firing in small-diameter DRG neurons.
NaV1.1 channels promote excitability and high frequency firing in several neuronal populations. In mouse models of irritable bowel syndrome and chronic visceral hypersensitivity, NaV1.1 is functionally upregulated, leading to hyperexcitability of mechanosensory fibers innervating the colon (Osteen et al., 2016; Salvatierra et al., 2018). Moreover, mutations in NaV1.1 are most frequently associated with inherited forms of epilepsy, including Dravet syndrome (Catterall et al., 2010). In this disorder, loss of NaV1.1 in hippocampal interneurons leads to reduced sodium current and attenuated action potential firing (Yu et al., 2006), resulting in disinhibition of hippocampal circuits that causes seizures (Oakley et al., 2013). To our knowledge, our results provide the first functional evidence for NaV1.1-dependent action potential firing in small-diameter somatosensory neurons.
Our data also indicate that the biophysical properties of NaV1.1-containing channel complexes could explain the heightened excitability of menthol-sensitive DRG neurons. NaV currents in these neurons entered into slow inactivation much more slowly than what has been reported for other DRG populations (Blair and Bean, 2003), with a time constant of ∼1.5 s (Figure 4A). This contrasts with capsaicin-sensitive nociceptors and IB4+ DRG neurons, where slow inactivation of TTX-resistant NaVs is reported to produce action potential adaptation in response to sustained depolarization (Blair and Bean, 2003; Choi et al., 2007). Importantly, application of ICA 121431 drastically enhanced the rate of entry into slow inactivation in menthol-sensitive DRG neurons (Figure 9C). Thus, the resistance of NaV1.1 currents to slow inactivation could be a mechanism by which menthol-sensitive neurons sustain action potential firing for extended periods of time.
Previous studies have reported that NaV1.1 channels are subject to use-dependent inactivation at high firing frequencies (Spampanato et al., 2001). It is therefore possible that in menthol-sensitive neurons, NaV1.1 α subunits associate with auxiliary proteins that destabilize inactivated states, such as the β4 subunit (Aman et al., 2009). Moreover, due to the lack of selective pharmacological tools, we were unable to test the contribution of NaV1.6 channels to action potential firing in small-diameter Vglut3lineage DRG neurons. It has been hypothesized, however, that synergistic activity of NaV1.1 and NaV1.6 is important for overcoming the high action potential threshold set by voltage-gated potassium channels of the KV1 family in pyramidal cells and GABAergic interneurons (Lorincz and Nusser, 2008). KV1 channels are also expressed in TRPM8+ trigeminal neurons, where they are proposed to determine thermal excitability (Madrid et al., 2009). We cannot rule out the possibility that action potential firing patterns in menthol-sensitive Vglut3lineage DRG neurons are tuned by the concerted actions of NaV1.1 and NaV1.6 channels that counterbalance an opposing Kv1 conductance, thus regulating the responsiveness of these neurons (Figure 10).
The finding that menthol-sensitive neurons are a subset of Vglut3lineage neurons raises the possibility that Vglut3 protein plays a role in synaptic transmission from TRPM8-expressing DRG neurons to second order neurons in the spinal cord. In contrast to this model, Vglut3- /- mice are reported to have normal responses to cold stimuli, indicating that Vglut3 protein is not required for TRPM8-dependent behaviors in mice (Draxler et al., 2014). Furthermore, in sensory neurons innervating the dura and cerebral blood vessels, TRPM8 and Vglut3 protein expression do not overlap in adult mice (Ren et al., 2018). Thus, we speculate that menthol-sensitive neurons express the Slc17A8 locus during development rather than in mature DRG.
Collectively, our data indicate that, unlike many other small-diameter DRG populations, action potential firing in in menthol-sensitive DRG neurons is dependent upon TTX-sensitive NaVs including NaV1.1. Genetic approaches using NaV1.1 null mutations are needed to define the exact contributions of this subunit to the function of menthol-sensitive neurons, as well as sensory-driven behaviors (Cheah et al., 2012). It also remains to be determined if NaV1.1 channels are viable therapeutic targets for pathologies that produce cold hypersensitivity. Additionally, menthol has been used for centuries as a topical analgesic and anti-pruritic. Indeed, it has been shown that TRPM8-expressing DRG neurons are required for inhibition of itch by cooling and furthermore, that topical application of menthol inhibits chloroquine-evoked itch behaviors (Palkar et al., 2018). Thus, targeting TTX-sensitive NaV1.1 channels in menthol-sensitive DRG neurons might prove to be a new direction for the treatment of various sensory disorders.
Author contributions
TNG and EAL performed calcium imaging and data analysis. TNG designed and preformed all electrophysiological experiments and data analysis. TAD assisted TNG with in situ hybridization and TAD performed quantitative analysis. TNG made the figures and wrote the first draft of the manuscript. EAL assisted with writing the manuscript. TNG and EAL edited the manuscript and all authors approved the manuscript. EAL and TNG acquired funding and EAL supervised the project.
Acknowledgements
This research was supported by NIAMS R01AR051219 (to E.A.L.). T.N.G. holds a Postdoctoral Enrichment Program Award from the Burroughs Wellcome Fund and was supported by NHLBI T32HL120826. Core facilities were supported by the Columbia University EpiCURE Center (NIAMS P30AR069632) and the Thompson Family Foundation Initiative in CIPN and Sensory Neuroscience. This project was initiated during the MBL Neurobiology Course with support from NINDS R25NS063307. Dr. Blair Jenkins, Mr. Javier Marquina-Solis and Dr. Adrian Thompson participated in preliminary studies at MBL. Thanks to Dr. Manu Ben-Johny and Dr. Lori Isom for sharing reagents, Dr. Irina Vetter for peptide toxins, Ms. Venesa Cuadrado for technical assistance, Ms. Rachel Clary for assistance with custom MATLAB routines, and Dr. Jon Sack and members of the Lumpkin laboratory for helpful discussions.
Footnotes
Conflicts of Interest: None to declare