Contribution of tetrodotoxin-sensitive, voltage-gated sodium channels (Na_V1) to action potential discharge from mouse esophageal tension mechanoreceptors

Abstract

Action potentials depend on voltage-gated sodium channels (Na_V1s), which have nine α subtypes. Na_V1 inhibition is a target for pathologies involving excitable cells such as pain. However, because Na_V1 subtypes are widely expressed, inhibitors may inhibit regulatory sensory systems. Here, we investigated specific Na_V1s and their inhibition in mouse esophageal mechanoreceptors—non-nociceptive vagal sensory afferents that are stimulated by low threshold mechanical distension, which regulate esophageal motility. Using single fiber electrophysiology, we found mechanoreceptor responses to esophageal distension were abolished by tetrodotoxin. Single-cell RT-PCR revealed that esophageal-labeled TRPV1-negative vagal neurons expressed multiple tetrodotoxin-sensitive Na_V1s: Na_V1.7 (almost all neurons) and Na_V1.1, Na_V1.2, and Na_V1.6 (in ∼50% of neurons). Inhibition of Na_V1.7, using PF-05089771, had a small inhibitory effect on mechanoreceptor responses to distension. Inhibition of Na_V1.1 and Na_V1.6, using ICA-121341, had a similar small inhibitory effect. The combination of PF-05089771 and ICA-121341 inhibited but did not eliminate mechanoreceptor responses. Inhibition of Na_V1.2, Na_V1.6, and Na_V1.7 using LSN-3049227 inhibited but did not eliminate mechanoreceptor responses. Thus, all four tetrodotoxin-sensitive Na_V1s contribute to action potential initiation from esophageal mechanoreceptors terminals. This is different to those Na_V1s necessary for vagal action potential conduction, as demonstrated using GCaMP6s imaging of esophageal vagal neurons during electrical stimulation. Tetrodotoxin-sensitive conduction was abolished in many esophageal neurons by PF-05089771 alone, indicating a critical role of Na_V1.7. In summary, multiple Na_V1 subtypes contribute to electrical signaling in esophageal mechanoreceptors. Thus, inhibition of individual Na_V1s would likely have minimal effect on afferent regulation of esophageal motility.

INTRODUCTION

Voltage-gated sodium channels (Na_V1s) are essential for action potential (AP) initiation from the peripheral nerve terminals of sensory afferent nerves (1). A stimulus acting on its receptor leads to the graded depolarization of the terminal membrane (generator potential) which in turn activates Na_V1s and initiates propagating APs. There are nine Na_V1s α subtypes, of these Na_V1.1, 1.2, 1.3, 1.6, 1.7, 1.8, and 1.9 are expressed in neurons (1, 2). Pharmacologically, Na_V1s are divided into those inhibited by tetrodotoxin (TTX-sensitive, Na_V1.1, 1.2, 1.3, 1.4, 1.6, and 1.7) and those not inhibited by TTX (TTX-resistant, Na_V1.5, 1.8, and 1.9) (1).

Na_V1 subtype-selective inhibitors are being developed for the treatment of various neurological conditions, in particular pain signaling and sensory disorders in the peripheral nervous system (2). However, because many Na_V1 subtypes are widely expressed throughout the peripheral nervous system, peripherally acting inhibitors may have unwanted effects on regulatory sensory systems. It is therefore crucial to obtain information on the role of various Na_V1 subtypes in various types of afferent nerves. Here, we focused on a major afferent nerve population in the gastrointestinal system—tension mechanoreceptors. Although tension mechanoreceptors innervate the whole gastrointestinal tube from the esophagus to rectum, we have focused on tension mechanoreceptors in the esophagus where they play the most prominent role in regulation of motility compared with more distal segments of the gastrointestinal (GI) tract. Esophageal tension mechanoreceptors are vagal afferents that are stimulated by low-threshold mechanical distension (3–8), which regulate esophageal motility including swallowing, secondary peristalsis, sphincter function, and vomiting (9–12). Dysfunctional activity in these afferents may contribute to ineffective esophageal motility and lower esophageal sphincter function associated with spastic disorders of the esophagus and noncardiac chest pain (12–14) and thus Na_V1s are potential therapeutic targets. On the other hand, targeting Na_V1s in other peripheral systems may interfere with normal esophageal tension mechanoreceptor function, thus impeding the function of the upper gut. We, therefore, determined the specific Na_V1s regulating AP initiation and conduction in tension mechanoreceptors of the esophagus.

METHODS

Animals

Wild-type C57BL/6J mice were purchased from Jackson Laboratory (No. 000664). Knockin Pirt-Cre [Pirttm3.1(cre)Xzd, kind gift from Dr. Xinzhong Dong, Johns Hopkins University; 15] were crossed with the Ai96(RCL-GCaMP6s) mice (B6;129S6-Gt(ROSA)26Sor^{tm96(CAG-GCaMP6s)Hze}/J, No. 024106, Jackson Laboratory) to produce Pirt-GCaMP6s mice that express the cytosolic Ca²⁺ reporter CaMP6s in all vagal neurons (16). Genotype of the offspring was confirmed using polymerase chain reaction. Offspring were weaned at 21 postnatal days and up to five littermates were housed per cage under normal condition (20°C, a 12-h dark/light cycle). Mice were provided with standard rodent chow and water ad libitum. All experiments were performed with approval from the University of South Florida Institutional Animal Care and Use Committee (AAALAC No. 000434).

Electrophysiological Single-Fiber Recordings of AP Discharge from Tension Mechanoreceptors in the Mouse Esophagus

The ex vivo preparation for recording mouse esophageal vagal afferents was modified from our previously published guinea pig preparation (8, 17). Wild-type male mice were killed by CO₂ inhalation and exsanguination, and the esophagus and its extrinsic vagal innervation (including the jugular-nodose vagal ganglia) was dissected out. The esophageal mucosa was left intact. The preparation was placed in a two-compartment Sylgard-lined Perspex recording dish: the esophagus was mounted in the tissue compartment; and the rostral vagus, including its vagal ganglion, was drawn through and pinned in the recording compartment. Dow Corning high vacuum grease was used to seal the opening containing the vagus nerve. Both compartments were separately superfused (3 mL/min) with heated (35°C) Krebs bicarbonate solution (KBS, in mM: 118 NaCl, 5.4 KCl, 1 NaH₂PO₄, 1.2 MgSO₄, 1.9 CaCl₂, 25 NaHCO₃, and 11 dextrose, gassed with 95% O₂-5% CO₂) containing indomethacin (3 µM) and atropine (1 µM). Polyethylene tubing was inserted 3–5 mm into the cranial and caudal ends of the esophagus and secured to enable mechanical distension with isobaric intraluminal pressure generated by a gravity-driven fluid system (5, 10, or 30 mmHg). An aluminosilicate glass microelectrode, filled with 3 M NaCl solution, was inserted into the vagal ganglion, near the cell bodies of vagal afferents. To identify distension-sensitive afferents, intraluminal pressure of 30 mmHg for 20 s was applied. If necessary, the microelectrode was manipulated within the ganglion to increase the signal-to-noise ratio of APs discharged from an individual fiber. The recorded APs were amplified (Microelectrode AC amplifier 1800; A-M Systems, WA), filtered (0.3 kHz of low cut-off and 1 kHz of high cut-off), and monitored on an oscilloscope (TDS3045B; Tektronix, OR). The scaled output from the amplifier was captured and analyzed by a Macintosh computer using Spike2 software (Electronic Design Ltd., Cambridge, UK). AP discharge was quantified offline and recorded in 1-s bins. Spike2 generated waveform templates of the APs in each recording to determine the number of unique AP shapes. Only recordings displaying a single AP shape, representing the AP discharge from a single afferent fiber were included in the analyses. Only fibers that responded throughout the entire 20 s in a nonadapting manner under control conditions were included in the dataset.

During single-fiber recordings, the esophagus was distended sequentially to 5, 10, and 30 mmHg for 20 s, with 2 min separating each distension. The main protocol was as follows: 1st control distension test, rest for 15 min, 2nd control distension test, Na_V1 inhibitor A perfused for 30 min followed by distension test, Na_V1 inhibitor A and B perfused for a further 30 min followed by distension test, TTX (300 nM) perfused for 15 min followed by distension test. Drugs were perfused into the esophageal chamber alone, reaching the target Na_V1 from the abluminal side. Stocks of Na_V1 inhibitors PF-05089771, ICA-121341, and LSN-3049227 were made up in DMSO and diluted in KBS for final concentrations of 3 µM, 10 µM, and 10 µM, respectively. PF-05089771 is a selective inhibitor of Na_V1.7 (18). ICA-121341 is a selective inhibitor of mouse Na_V1.1, Na_V1.3, and Na_V1.6 (19, 20). LSN-3049227, aka compound-801, is a selective inhibitor of mouse Na_V1.2, Na_V1.6 and Na_V1.7 (20, 21). For each condition, the number of APs for each distension (20 s) was summated and presented in raw form. The number of APs for each distension was also normalized to the number of APs during the 1st control test. The normalized responses under different inhibitor conditions were compared using a two-way ANOVA with Tukey’s multiple comparisons. Only fibers that had reproducible responses to the 1st and 2nd control distensions were included in the analyses. To avoid washout issues of the Na_V1 inhibitors, only one fiber was studied per mouse (thus the n number refers to both number of fibers and number of mice). Although all fibers included in the analysis had nonadapting responses under control conditions, some fibers had responses to the 20-s esophageal distension in the presence of Na_V1 inhibitors that appeared to adapt. To determine the extent to which the Na_V1 inhibition impacted the dynamics of the response to 20-s distension, the AP discharge data were re-analyzed into two periods: the 1st 10 s and 2nd 10 s. The raw data was again normalized as aforementioned to the responses in the given set for the 1st control test. The normalized responses for the 1st and 2nd 10 s periods under different inhibitor conditions were compared using a two-way ANOVA with Tukey’s multiple comparisons.

The Δ% inhibition for each inhibitor was calculated by subtracting the normalized response (for the complete 20 s distension) following the incubation with the given inhibitor from the normalized response (for the complete 20 s distension) preceding the inhibitor incubation. As such the Δ% inhibition evoked by a particular inhibitor could be compared for when the inhibitor was given alone and when given in combination with other inhibitors. Differences in these Δ% inhibition values for a given inhibitor would suggest redundancy in the role of Na_V1s in AP discharge. For example, if the normalized response for fiber A during control 2 was 100, and the response following LSN-3049227 alone was 50, and the response following the LSN-3049227 and ICA-121341 combination was 5, and if the response for fiber B during control 2 was 100, and the response following ICA-121341 alone was 80, the calculated inhibition for ICA-121341 when used alone (fiber B) and when in combination with LSN-3049227 (fiber A) would be 20% and 45%, respectively. If the Δ% inhibition was greater than 25%, the fiber was categorized as being inhibited. We then calculated the % of fibers that were sensitive to a given inhibitor: 100 × (number of fibers inhibited by 25% or more/the total number of fibers tested).

Single-Neuron RT-PCR

Wild-type male mice (n = 4 animals) were anesthetized with ketamine (100 mg/kg ip) and xylazine (10 mg/kg ip). Nerves innervating the mouse esophagus were retrogradely labeled by injecting 4 × 10 µL of 10 mg/mL fluorescently labeled wheat germ agglutinin (WGA-488, Thermo Fisher) in between the muscle and mucosal layers of the esophagus. Mice were euthanized one week later by inhalation of CO₂ and exsanguinated. Both vagal ganglia were dissected out and dissociated enzymatically using 2 mg/mL collagenase type 1 and 2 mg/mL dispase II in 1× HBSS at 37°C for 50 min, as described previously (18). Following sequential mechanical trituration, neurons were washed and resuspended in 10 mL of L-15 medium containing 10% FBS. Then 20 µL of the cell suspension was seeded onto coverslips coated with poly-d-lysine and laminin and incubated at 37°C. As a positive control, we transferred 2 µL of the cell suspension into PCR tubes containing 1 µL RNaseOUT.

Coverslips with dissociated vagal sensory neurons were superfused with KBS. Esophageal-labeled neurons were identified by labeling with WGA-488, as determined by fluorescence (470 nm excitation and 525 nm emission) and an Olympus BX51WI microscope. Individual neurons were collected with a borosilicate glass pipette by applying a gentle negative pressure. Each single neuron was stored in PCR tube containing 1 µL RNaseOUT and stored at −20°C. The KBS surrounding the vagal neurons was sampled (∼2 µL) as a negative control for the RT-PCR.

First-strand cDNA was synthesized with Invitrogen’s SuperScriptIII First-Strand Synthesis System for RT-PCR by following the manufacturer’s instructions. Primers and dNTP mix were added into each neuron-containing PCR tube. Samples were incubated at 75°C for 10 min and then placed on ice for at least 1 min. cDNA synthesis mix containing 10× RT buffer, MgCl₂, DTT, and SuperScript III RT was added into each sample. Samples were incubated at 50°C for 50 min, followed by 85°C for 5 min. cDNA was stored at −20°C until used for PCR amplification. cDNA was amplified with HotStarTaq DNA polymerase for 50 cycles of denaturation at 94°C for 30 s, followed by annealing at 60°C for 30 s and extension at 72°C for 1 min. Customized intron-spanning primers for mouse Na_V1s, P2X₂, and TRPV1 (Table 1) were used as described previously (19, 20). Products were visualized in 1.5% agarose gel with GelRed, with a 100-bp marker.

Table 1.

Mouse primers used for single-cell reverse-transcription polymerase chain reaction

Gene	Primer	Sequence 3′ to 5′	Product Size	NCBI Reference Sequence
P2X2	Forward	GGGGCAGTGTAGTCAGCATC	241 bp	NM_153400
	Reverse	TCAGAAGTCCCATCCTCCA
TRPV1	Forward	TCACCGTCAGCTCTGTTGTC	229 bp	NM 001001445
	Reverse	GGGTCTTTGAACTCGCTGTC
NaV1.1	Forward	AGACAGCATCAGGAGGAAGG	118 bp	NM_018733.2
	Reverse	GGAGAACAGGGAACCACGA
NaV1.2	Forward	TTTTCGGCTCATTCTTCACA	305 bp	NM_001099298.2
	Reverse	CATCTCTTGGCTCTGGTCGT
NaV1.3	Forward	AGACAGAGGGAGCACTTGGA	200 bp	NM_018732.3
	Reverse	CTATTGCGTCTTGGGGAAAA
NaV1.4	Forward	TCATCTTCCTGGGTTCCTTC	206 bp	NM_133199.2
	Reverse	ATCTGCCTCCTCTCCACCTT
NaV1.6	Forward	AGGCAGCAAAGACAAACTGG	157 bp	NM_001077499.2
	Reverse	GCAGCACTTGAACCTCTGG
NaV1.7	Forward	ATGCTCTTCTTTGCGGTTTC	381 bp	NM_001290674.1
	Reverse	CGGCTTCTTCCTGCTCTTTT

AP Conduction in Esophageal Vagal Afferents—GCaMP6s Studies

AP conduction in vagal afferents triggers transient increases in cytosolic Ca²⁺ in the neuronal soma within the vagal ganglion. CGaMP6s fluorescence increases with [Ca²⁺]_cytosolic, and is therefore an indirect indicator of electrical activity (21). Here, we used GCaMP6s imaging of esophageal-specific vagal neurons to determine the role of specific Na_V1s in the propagation of APs. The activity of individual vagal sensory neurons (i.e., cytosolic Ca²⁺ transients) was determined using GCaMP6s imaging of the vagal ganglia of Pirt-GCaMP6s mice as described previously (22). The nerves innervating the esophagus of male Pirt-GCaMP6s mice were retrogradely labeled with WGA-555 under ketamine-xylazine anesthesia as aforementioned. One week later, mice were killed by CO₂ inhalation and exsanguination, and the vagal nerve together with the jugular-nodose vagal ganglia was dissected out. The vagus nerve was then cleaned and desheathed using fine forceps, carefully avoiding any breaks or tears of the nerve fiber tracts. A two-compartment chamber was used to mount the nodose ganglia in one compartment and the vagus nerve in another. The two compartments were separately superfused with KBS (4 mL/min at 37°C). There was no fluid exchange between the two compartments, which were sealed using Vaseline. The vagal ganglia were pinned in the second compartment for two-photon imaging using the Olympus FVMPE-RS multiphoton system. The tissue was excited with 960 nm (InSight X3 IR laser) and the emitted light was split into two paths with specific filter cubes and distinct photomultiplier tubes (PMT) containing gallium arsenide phosphide (GaAsP): an FVG filter cube and 488 PMT for GCaMP6s, and an FGR filter cube and 555 for WGA. The ×25 water immersion objective was positioned directly above the nodose ganglia in the small chamber. A z-stack (100 µm) of the ganglia was visualized (∼1,000 GCaMP6s-positive neurons) using Olympus Fluoview software, which allowed for automated increases in laser percentage and PMT voltage depending on the depth of the z-plane into the ganglia. Live images were recorded every 12 s. We have previously shown that GCaMP6s responses to vagal electrical stimulation using a concentric electrode (PI-NE-100, Microprobes) with a A-V Systems 2100 stimulator at 20 V (0.2 ms square pulse) are quasilinear between 1 Hz and 10 Hz stimulation when given in 5-s trains (22). Here, we recorded GCaMP6s fluorescence at baseline and during vagal electrical stimulation (5 Hz for 20 s) under control conditions, and following 30-min treatment with Na_V1 blockers [either PF-05089771 (3 µM) or ICA-121341 (10 µM)], and following 15-min treatment with 300 nM TTX. The Na_V1 inhibitors were added to the vagus nerve compartment. Lastly, we determined the GCaMP6s responses to capsaicin (1 µM) added to the ganglionic compartment. An n of 3 animals were used for the PF-05089771 studies. An n of 3 animals were used for the ICA-121341 studies.

Fluoview-derived images were analyzed offline using ImageJ. Each nodose ganglia was imaged up to a depth of 100 µm, in ten 10-µm thick z-planes. We analyzed either odd- or even-numbered z-planes. This limited the possibility that a given mouse neuron (which are ∼20 µm in diameter) was counted twice. Each analyzed z-plane was used to mark the regions of interest (ROIs) of neurons that responded to electrical stimulation under control conditions. GCaMP6s fluorescence was recorded for each ROI before (baseline) and during electrical stimulation and during capsaicin treatment. The mean baseline response was subtracted from the maximum fluorescence during electrical stimulation to yield a delta-evoked GCaMP6s response during control conditions and during treatment with the specific Na_V1s inhibitor. Only WGA-positive neurons that responded to electrical stimulation with an increase in GCaMP6s fluorescence >1.5-fold over baseline were included. In these studies, 99 out of the 427 WGA-labeled neurons (23.1%) responded to electrical stimulation by this criterion. Preliminary studies showed that increased percentages of neurons could be activated using greater stimulating voltages, but this could lead to thermal damage of the nerve at its contact with the concentric electrode. WGA-labeled neurons were grouped according to their sensitivity to capsaicin. The delta GCaMP6s fluorescence during Na_V1 inhibitor treatment was also normalized to control conditions (Response_Normalized = 100 × Response_Inhibitor/Response_Control). GCaMP6s responses were compared using a one-way ANOVA. A neuron was considered completely blocked, partially blocked, or unaffected by the Na_V1 inhibitor if its Response_Normalized was <20%, <80%, or >80%, respectively.

Data and Statistical Analysis

For all experiments, data were compiled using Excel and GraphPad software. In all cases, P < 0.05 was considered significant.

RESULTS

AP Discharge in Esophageal Tension Mechanoreceptors is TTX-Sensitive

Esophageal tension mechanoreceptors encode innocuous esophageal distention in all species up to 60 mmHg (3–8). In our initial studies of mouse mechanoreceptors, we recorded similar numbers of APs in response to esophageal distention to 30 and 60 mmHg but responses to 60 mmHg were variable (data not shown). We therefore studied responses to distention to 5, 10, and 30 mmHg (20 s each, every 2 min). This protocol yielded highly reproducible responses to distention (Figs. 1 and 2, n = 6 fibers). Incubation for 30 min with the vehicle DMSO (0.15%) had no effect on the response distention (Figs. 1 and 2). Importantly, responses to esophageal distention to 5, 10, and 30 mmHg were completely abolished by 15-min incubation with 300 nM TTX (Figs. 1 and 2, P < 0.05). In separate experiments, responses to esophageal distension to 5, 10, and 30 mmHg were significantly inhibited by 100 nM TTX (Fig. 3, A–D, n = 6, P < 0.05), whereas 30 nM TTX only inhibited esophageal distension by 30 mmHg (Fig. 3, A, E–G, n = 4, P < 0.05). Thus, our data indicates that AP discharge from esophageal mechanoreceptors in response to distension is completely dependent on TTX-sensitive Na_V1s.

Figure 1.

Low-pressure distension evokes reproducible and nonadapting action potential (AP) discharge from esophageal tension mechanoreceptors which are completely inhibited by tetrodotoxin (TTX). Responses to distension to 5, 10, and 30 mmHg during controls and following 30-min incubation with DMSO (0.15%) and then 15-min incubation with TTX (300 nM). A: protocol. B: representative traces of AP discharge (in black), with individual waveforms of identified APs (in blue) demonstrating that responses were recorded from the same afferent.

Figure 2.

Esophageal mechanoreceptor responses to distension are tetrodotoxin (TTX)-sensitive. Responses of mechanoreceptors to distension to 5, 10, and 30 mmHg during controls and following 30-min incubation with DMSO (0.15%) and then 15-min incubation with TTX (300 nM) (n = 6). A: means ± SE the total number of action potentials (APs) during each test. B: number of APs normalized to the fiber’s response during the 1st control (red lines represent means ± SE). *Significant difference to control 2 (P < 0.05), #Significant difference to DMSO test (P < 0.05).

Figure 3.

Esophageal mechanoreceptor responses to distension are inhibited by low-dose TTX. Responses to distension to 5, 10, and 30 mmHg during controls and following 15-min incubation with 100 nM (B–D, n = 6) or 30 nM (E–G, n = 4) tetrodotoxin (TTX). A: protocol. B and E: representative traces. C and F: the total number of action potentials (APs) during each test. D and G: number of APs normalized to the fiber’s response during the 1st control. *Significant difference to control 2 (P < 0.05).

TTX-Sensitive Na_V1 Transcript Expression in Esophageal Mechanoreceptors

We next evaluated the expression of TTX-sensitive Na_V1s using single-cell RT-PCR in neurons projecting mechanoreceptors to the mouse esophagus. Based on previous studies, these were identified as vagal nodose neurons retrogradely labeled from the esophagus that lack the expression of TRPV1 (8, 23, 24). We found that mouse esophageal mechanoreceptors expressed multiple TTX-sensitive Na_V1s (Fig. 4A). Almost all the neurons tested (22 out of 23) expressed Na_V1.7 (Fig. 4B). Furthermore, almost all the neurons (21 out of 23) also expressed another TTX-sensitive Na_V1: ∼50% of the neurons expressed Na_V1.1, Na_V1.2, and Na_V1.6 (Fig. 4). Na_V1.1, Na_V1.2, and Na_V1.6 expression did not appear to be correlated. Finally, very few neurons expressed Na_V1.3 or Na_V1.4 (Fig. 4).

Figure 4.

Esophageal TRPV1-negative vagal neurons express multiple tetrodotoxin (TTX)-sensitive, voltage-gated sodium channel (Na_V1) transcripts. A: representative gels of individual dissociated vagal neurons (1–20) labeled from the esophagus using wheat germ agglutinin (WGA)-488. Transcripts for P2X₂ (positive control for nodose afferents), TRPV1, and TTX-sensitive Na_V1s were detected using targeted primers. B: the % of P2X₂-positive/TRPV1-negative neurons that express each TTX-sensitive Na_V1 (bottom). In addition, the percentage of Na_V1s in subsets of neurons that contain another Na_V1 (top).

AP Discharge in Esophageal Tension Mechanoreceptors is Dependent on Multiple Na_V1s

We hypothesized that Na_V1.1, Na_V1.2, Na_V1.6, and Na_V1.7 contribute to AP discharge from tension mechanoreceptors in response to esophageal distension. First, we determined the role of Na_V1.7, using a 30-min incubation of the selective Na_V1.7 inhibitor PF-05089771 (PF771, 3 µM) (25). PF771 alone had no effect on responses to 5 and 10 mmHg, but PF771 caused a minor yet significant decrease in the response to 30 mmHg (Fig. 5, P < 0.05, n = 8 fibers). These fibers were then subjected to a further 30-min incubation with PF771 (3 µM) in combination with ICA-121341 (ICA, 10 µM), which at this concentration selectively inhibits mouse Na_V1.1 and Na_V1.6 (19, 20). The combination of PF771 and ICA had a variable inhibitory effect on responses to esophageal distension, but this only reached significance for 30 mmHg (Fig. 5, P < 0.05, n = 8). In a separate set of recordings, we found that 30-min incubation with 10 µM ICA alone had no effect on responses to 5 and 10 mmHg, but caused a minor yet significant, decrease in the response to 30 mmHg (Fig. 6, P < 0.05, n = 12). These fibers were then subjected to a further 30-min incubation with ICA (10 µM) in combination with PF771 (3 µM). Consistent with our previous data, this combination of ICA and PF771 had a variable inhibitory effect on responses to esophageal distension that only reached significance for 30 mmHg (Fig. 6, P < 0.05, n = 12). These data indicate that multiple Na_V1s (Na_V1.1/Na_V1.6 and Na_V1.7) contribute to tension mechanoreceptors responses to esophageal distension. Furthermore, the residual responses in the presence of PF771 and ICA are likely mediated in part by Na_V1.2.

Figure 5.

Esophageal mechanoreceptor responses to distension are partially inhibited by PF771. Responses to distension to 5, 10, and 30 mmHg during controls, and following 30-min incubation with PF771 (PF, 3 µM), then following 30-min incubation with 3 µM PF and 10 µM ICA (n = 8). A: protocol. B: representative traces. C: means ± SE the total number of action potentials (APs) during each test. D: number of APs normalized to the fiber’s response during the 1st control (red lines represent means ± SE). *Significant difference to control 2 (P < 0.05), #Significant difference to responses following treatment with PF771 alone (P < 0.05).

Figure 6.

Esophageal mechanoreceptor responses to distension are partially inhibited by ICA. Responses to distension to 5, 10, and 30 mmHg during controls and following 30-min incubation with ICA (10 µM), then following 30-min incubation with 10 µM ICA and 3 µM PF771 (PF, n = 12). A: protocol. B: representative traces. C: means ± SE the total number of action potentials (APs) during each test. D: number of APs normalized to the fiber’s response during the 1st control (red lines denote means ± SE). *Significant difference to control 2 (P < 0.05), #Significant difference to responses following treatment with ICA alone (P < 0.05).

Next, we tested the inhibitory effect of LSN-3049227 (LSN, aka compound 801) on mechanoreceptor responses to esophageal distension. At 10 µM, LSN is expected to block mouse Na_V1.2, Na_V1.6, and Na_V1.7 (20, 21). Following 30-min incubation, LSN alone had no effect on responses to 5 and 10 mmHg, but caused a significant decrease in the response to 30 mmHg (Fig. 7, P < 0.05, n = 8). The residual responses in the presence of LSN are likely mediated in part by Na_V1.1. We then treated these fibers with a further 30-min incubation with LSN (10 µM) in combination with ICA (10 µM), which would be expected to block all the TTX-sensitive Na_V1 expressed in esophageal mechanoreceptors (Na_V1.1, Na_V1.2, Na_V1.6, and Na_V1.7). As expected, this combination robustly inhibited responses at all distension pressures (Fig. 7, P < 0.05, n = 8).

Figure 7.

Esophageal mechanoreceptor responses to distension are partially inhibited by LSN. Responses to distension to 5, 10, and 30 mmHg during controls and following 30-min incubation with LSN (10 µM), then following 30-min incubation with 10 µM LSN and 10 µM ICA (n = 8). A: protocol. B: representative traces. C: means ± SE the total number of APs during each test. D: number of action potentials (APs) normalized to the fiber’s response during the 1st control (red lines denote means ± SE). *Significant difference to control 2 (P < 0.05), #Significant difference to responses following treatment with LSN alone (P < 0.05).

Although only mechanoreceptors that responded in a nonadapting manner to distension under control conditions were included in the study, the responses of some fibers to distension in the presence of Na_V1 inhibitors appeared to adapt (see Fig. 5B and Fig. 7B). We re-analyzed the data, splitting the response to 20-s distension into two separate periods: the 1st 10 s and the 2nd 10 s of the distension (Fig. 8A). We plotted the means ± SE of the normalized responses for the 1st and 2nd 10 s periods for the fibers treated with PF771 (3 µM) and PF771 (3 µM) in combination with ICA (10 µM) (Fig. 8B), with ICA (10 µM) and ICA (10 µM) in combination with PF771 (3 µM) (Fig. 8C), and with LSN (10 µM) and LSN (10 µM) in combination with ICA (10 µM) (Fig. 8D). There appeared to be an overall trend for the response during the 2nd 10-s period to be more inhibited by PF771, ICA, and LSN, than the 1st 10-s period, especially at the highest distension pressure tested (30 mmHg). However, only the response during the 2nd 10-s period to 30 mmHg distension in the presence of LSN alone was significantly lower than the 1st 10-s period (P < 0.05) (Fig. 8D).

Figure 8.

Comparison of mechanoreceptor responses to the 1st and 2nd 10 s of esophageal distension. A: representative example of action potential (AP) discharge to 20 s distension, with the 1st and 2nd periods identified. B–D: means ± SE normalized AP discharge in response to the 1st (x-axis) and 2nd (y-axis) 10 s of esophageal distension for fibers treated with PF771 (3 µM) and PF771 (3 µM) in combination with ICA (10 µM) (B, n = 8), with ICA (10 µM) and ICA (10 µM) in combination with PF771 (3 µM) (C, n = 12), and with LSN (10 µM) and LSN (10 µM) in combination with ICA (10 µM) (D, n = 8). Responses to distension to 5, 10, and 30 mmHg are denoted by circles, squares, and triangles, respectively. The black line denotes a slope of 1. *Significant difference between responses to the 1st and 2nd 10 s of distension (P < 0.05).

In addition to clear stimulus intensity-dependent sensitivity to Na_V1 inhibition, there was substantial variation in the inhibitory effect of PF771, ICA, and LSN on mechanoreceptor responses to esophageal distension. This could be due in part to heterogeneous expression of Na_V1 transcripts (Fig. 4) or redundant recruitment of different Na_V1. We calculated (for the responses to the complete 20 s distension) the Δ% inhibition specifically evoked by PF771, ICA, and LSN when incubated alone, or when in combination with the other Na_V1 inhibitors (see methods for calculation) (Fig. 9, A–C). If a fiber’s normalized responses to distension were reduced by 25% or more by a given inhibitor, it was considered to be inhibited. We therefore calculated the % of mechanoreceptors whose responses were inhibited by >25% by each inhibitor treatment (Fig. 9, D–F). Using these analyses, we found that only LSN-induced inhibition was significantly increased at higher distension pressures (i.e., 30 mmHg) (Fig. 9, C and F), whereas there was no significant increase in the % inhibition by PF771 or ICA at 30 mmHg (Fig. 9, A, B, D, and E). Although the % inhibition caused by PF771 appeared to increase in some fibers when in combination with ICA, this did not reach significance for the entire population (Fig. 9, A and D, P > 0.05). In addition, the % inhibition caused by ICA was little changed when PF771 was added to the buffer (Fig. 9, B and E, P > 0.05). Interestingly, the % inhibition caused by ICA was significantly increased when in combination with LSN (Fig. 9, B and E, P < 0.05). These data suggest that there is little redundancy between PF771-inhibited channels and ICA-inhibited channels, but that there is redundancy between ICA-inhibited channels and LSN-mediated channels (although not at 30 mmHg, where LSN alone is sufficient to almost abolish the response).

Figure 9.

Evidence of redundancy in the role of specific tetrodotoxin (TTX)-sensitive, voltage-gated sodium channels (Na_V1s) in mechanoreceptor responses to distension. Δ% inhibition for PF771 (PF, A), ICA (B), and LSN (C) alone or in combination with other Na_V1 inhibitors (see methods for calculation) (red lines denote means ± SE). Horizontal dotted line denotes the designated threshold to indicate inhibition. *Significant difference across distension pressures (P < 0.05), #Significant difference in the inhibitory effect of the blocker when used in combination with other inhibitors compared with the blocker alone (P < 0.05). Percentage of fibers inhibited by PF771 (PF, D), ICA (E), and LSN (F) alone or in combination with other Na_V1 inhibitors, as defined by the 25% threshold value shown in A–C.

AP Conduction in Esophageal Vagal Afferents—GCaMP6s Studies

In our studies of distension-evoked activation of the tension receptors in the esophagus, the Na_V1 blockers likely acted on the nerves within the tissue where access to the nerve endings is relatively unimpeded. In that case, the drugs would have little access to the vagal axons per se because of the thick relatively impervious sheath covering the vagus nerves (17, 26). To identify the Na_V1 subtypes specifically involved in AP conduction along the axons within the vagus nerve, we took another approach: instead of stimulating the sensory afferent terminals within the esophagus, we electrically stimulated the peripheral end of the desheathed vagus nerve and used GCaMP6s imaging of vagal ganglionic neurons to quantify axonal AP conduction from the stimulating electrode on the vagus to the neuronal cell bodies (5 Hz for 20 s) (Fig. 10A). Retrograde tracing using esophageal injections of WGA-555 identified the esophageal afferents (Fig. 10A). Electrical stimulation evoked transient increases in GCaMP6s fluorescence in both capsaicin-insensitive (Fig. 10B) and capsaicin-sensitive neurons (Fig. 10E) innervating the esophagus, which were presumed to be esophageal tension mechanoreceptors and nociceptive fibers, respectively. Thirty-minute incubation with PF771 (3 µM) caused a significant reduction in electrically evoked GCaMP6s fluorescence in both capsaicin-insensitive (Fig. 10B, n = 12 neurons, P < 0.05) and capsaicin-sensitive (Fig. 10E, n = 31, P < 0.05) esophageal neurons. Similarly, 30-min incubation with ICA (10 µM) caused a significant reduction in electrically evoked GCaMP6s fluorescence in both capsaicin-insensitive (Fig. 10B, n = 35, P < 0.05) and capsaicin-sensitive (Fig. 10E, n = 21, P < 0.05) esophageal neurons. In both PF771 and ICA studies, a subsequent 15-min incubation with 300 nM TTX abolished electrically evoked GCaMP6s fluorescence in both capsaicin-insensitive and capsaicin-sensitive esophageal neurons (Fig. 10, B and E, P < 0.05). There was no difference between the % inhibition of electrically evoked GCaMP6s fluorescence by PF771 and ICA in either capsaicin-insensitive or capsaicin-sensitive neurons (Fig. 10, C and F, P > 0.05). Again, there was heterogeneity in the inhibitory effect of PF771 and ICA: ∼40% of neurons were completely inhibited, ∼30% of neurons were partially inhibited, and ∼30% neurons were unaffected (Fig. 10, D and G). These data indicate that AP conduction in esophageal neurons is completely dependent on TTX-sensitive Na_V1s. Furthermore, in some neurons AP conduction is solely dependent on Na_V1.7 (completely inhibited by PF771).

Figure 10.

The role of specific tetrodotoxin (TTX)-sensitive, voltage-gated sodium channels (Na_V1s) in mediating action potential (AP) conduction in esophageal vagal afferents. A: representative multiphoton images of vagal ganglia of Pirt-GCaMP6s mice retrogradely labeled from the esophagus with wheat germ agglutinin (WGA)-555 in response to electrical stimulation of the desheathed vagus nerve and ganglionic treatment with capsaicin (1 µM). Scale bar denotes 50 µm. GCaMP6s fluorescence evoked by vagal stimulation for WGA-labeled capsaicin-insensitive (B) and capsaicin-sensitive (E) neurons under control conditions and following treatment with PF771 (PF, 3 µM, left) or ICA (10 µM, right) and TTX (300 nM) (black lines denote means ± SE). CGaMP6s responses to vagal stimulation following Na_V1 blocker treatment normalized to control responses for WGA-labeled capsaicin-insensitive (C) and capsaicin-sensitive (F) neurons (black lines denote means ± SE). Percentage of WGA-labeled capsaicin-insensitive (D) and capsaicin-sensitive (G) neurons whose normalized responses to vagal stimulation following Na_V1 blocker treatment were <20% (blocked), between 20% and 80% (partially blocked), and >80% (no effect).

DISCUSSION

Our ex vivo single-fiber recordings from vagal neurons innervating the mouse esophagus indicate that there is a substantial population of afferents that are sensitive to low-threshold changes in distension pressure, whose robust responses are slowly adapting, reproducible, and saturate within the physiological range. These properties indicate that these afferents are mouse esophageal tension mechanoreceptors, which have been shown to display similar slowly adapting responses to low-threshold changes in circumferential stretch (7, 27, 28). Esophageal tension mechanoreceptors have been recorded in many other mammals including cat, dog, ferret, opossum, rat, and guinea pig (3–6, 8, 28) and are considered to play an important role in regulating esophageal motility including swallowing, secondary peristalsis, sphincter function, and vomiting (9–12). Based upon their saturation to physiological distension and lack of sensitivity to the canonical nociceptive stimulus capsaicin, esophageal tension mechanoreceptors are considered to be non-nociceptive (6, 8).

The data show that distension-evoked AP discharge in mouse esophageal tension mechanoreceptors was abolished by TTX. Based on the single-cell RT-PCR data, we conclude that the TTX-sensitive channels likely involved in mechanotransduction were Na_V1.1, Na_V1.2, Na_V1.6, and Na_V1.7. This complex array of TTX-sensitive Na_V1s in mouse esophageal mechanoreceptors is somewhat different from other identified vagal afferent populations: for example, guinea pig TRPV1-positive neurons innervating the esophagus also express Na_V1.3 (17), whereas guinea pig TRPV1-negative neurons innervating the trachea express Na_V1.3, Na_V1.6, and Na_V1.7 but lack Na_V1.1 and Na_V1.2 (26). As in these other studies, Na_V1.7 was most widely expressed in mouse esophageal mechanoreceptors.

Early patch clamp studies of dissociated vagal neurons suggested that TTX-sensitive, voltage-gated Na⁺ currents were described by a single component, and this was thought to be mediated by Na_V1.7 alone (29–32). However, further patch clamp studies in heterologous expression systems suggest subtle differences in biophysical properties between the TTX-sensitive currents mediated by Na_V1.1, Na_V1.2, Na_V1.3, Na_V1.6, and Na_V1.7, which impact their physiological and pathophysiological function in AP initiation, conduction, and electrical excitability (1, 33–35): for example, Na_V1.1 and Na_V1.6 have a faster recovery from inactivation than Na_V1.2 and Na_V1.3; Na_V1.7 has slow closed-state inactivation leading to the development of ramp currents; and Na_V1.6 may mediate persistent currents. However, it is important to state that recorded biophysical properties of Na_V1s are highly dependent on the system studied, and this demonstrates the substantial impact of modulators including auxiliary β subunits and posttranslational modifications (1, 36). The expression of β subunits and the manifestation of posttranslational modifications are largely unknown at the nerve terminal of esophageal tension mechanoreceptors.

The current study demonstrates that multiple TTX-sensitive Na_V1s contribute to AP initiation and conduction in mouse esophageal tension mechanoreceptors. Our conclusions are dependent on the selectivity of pharmacological inhibitors PF771, ICA, and LSN for the various TTX-sensitive Na_V1 subtypes that are expressed by the esophageal mechanoreceptors. It is important to note that, unlike TTX, these are state-dependent inhibitors, whose IC50s are typically determined in patch clamp studies whose protocols force the channel into a particular state (e.g., the inactivate state). Given that it is impossible to know the proportion of native channels in closed, open, and inactivate states in our ex vivo preparation, the relationship between published IC₅₀ and effective IC₅₀ is likely to be relative. PF771 is considered a selective Na_V1.7 inhibitor, with IC₅₀ of 8 and 11 nM at mouse and human Na_V1.7 (25). PF771 IC50s for other Na_V1s are >110 nM. Previous studies have shown that 3 µM PF771 is sufficient to block Na_V1.7-mediated AP discharge from nodose C-fibers innervating the guinea pig airways (26), and that increasing the concentration of PF771 from 3 to 10 µM did not increase the inhibition of distension-evoked activation of guinea pig esophageal C-fibers (17). Furthermore, 10 µM PF771 had no effect on Na_V1.7-independent AP conduction in guinea pig esophageal C-fibers (17). The selectivity of ICA and LSN for specific Na_V1s depends on the species. ICA blocks human Na_V1.1, Na_V1.2, and Na_V1.3 with IC₅₀ of 23, 240, and 13 nM, respectively, but has little effect on Na_V1.6 or Na_V1.7 (IC₅₀ > 10 µM) (37). Although ICA similarly inhibits mouse Na_V1.1 (IC₅₀ of 10 nM), it also inhibits Na_V1.6 (IC₅₀ of 3.7 nM) yet has limited affinity for Na_V1.2 (IC₅₀ of 553 nM) (38). LSN blocks human Na_V1.1, Na_V1.2, Na_V1.3, Na_V1.6, and Na_V1.7 with IC₅₀ of 8, 6, 460, 8, and 1 nM, respectively (19). Although LSN similarly inhibits mouse Na_V1.2 and Na_V1.6 (IC₅₀ of 4 and 1.6 nM, respectively), it has limited affinity for Na_V1.1 (IC₅₀ of 261 nM) (38). Thus, for the esophageal mechanoreceptor Na_V1s, at the concentrations used in this study, ICA and LSN would both be expected to block Na_V1.6, but ICA would also block Na_V1.1 and LSN would also block Na_V1.2 and Na_V1.7. Indeed, our previous study of AP conduction in mouse vagal efferent nerves innervating the airways demonstrated that ICA and LSN (at the same concentrations used here) blocked nonoverlapping Na_V1 subtypes (19).

Consistent with the expression of only Na_V1.1, Na_V1.2, Na_V1.6, and Na_V1.7 in esophageal mechanoreceptors, the AP discharge from these fibers was eliminated at 30 mmHg by the LSN and ICA combination. Although significant, this inhibition was more variable at lower pressures, even though these responses were entirely TTX-dependent. This could be due to the state-dependent affinity of these inhibitors for the Na_V1 inactive state (37), which channels are unlikely to be in during low stimulus intensities in native tissue. Nevertheless, we also saw preferential block of the higher stimuli with a low concentration of TTX (30 nM), which binds to Na_V1s in a state-independent manner. It is likely therefore that the decreased inhibition at low stimulus intensities by these inhibitors reflects some other biophysical property of the afferent terminal in converting the kinetics of the graded potential into APs. Indeed, we found a trend for greater relative inhibition of AP discharge for all inhibitors (PF771, ICA, and LSN) at 30 mmHg than the lower stimulus intensities. This tendency precludes any meaningful investigation of stimulus intensity-dependent recruitment of Na_V1s, which we had previously observed in distension-induced responses in C-fibers innervating the guinea pig esophagus (17). In those fibers, Na_V1.7 contributed preferentially to responses to low stimulus distension, whereas Na_V1.1, Na_V1.2, and Na_V1.3 only contributed to responses to high stimulus distension.

Inhibition of mechanoreceptor responses to esophageal distension to 30 mmHg by PF771 alone was significant yet limited, indicating that while Na_V1.7 contributed, it was not required for most of the response. Similarly, inhibition of responses to 30 mmHg by ICA alone was significant yet quite variable, indicating that either one or both of Na_V1.1 and Na_V1.6 contributed, but that Na_V1.2 and Na_V1.7 were sufficient for much of the AP discharge. Inhibition of responses to 30 mmHg by LSN alone was substantial, indicating that some combination of Na_V1.2, Na_V1.6, and Na_V1.7 contributed. Furthermore, the residual response in the presence of LSN suggested that Na_V1.1 contributed. As expected, due to their nonoverlapping Na_V1 inhibition, adding ICA to PF771 or adding PF771 to ICA caused greater inhibition than either agent alone. In some fibers, the response was completely abolished by the PF771 and ICA combination, but in ∼50% of the fibers the residual response suggested that Na_V1.2 contributed. Thus, we provide direct evidence that Na_V1.1, Na_V1.2, and Na_V1.7 all contribute to distension-evoked AP discharge in esophageal mechanoreceptors. Interestingly, the inhibition evoked specifically by ICA (blocks Na_V1.1 and Na_V1.6) was significantly increased by the presence of LSN (blocks Na_V1.2, Na_V1.6, and Na_V1.7) but not by the presence of PF771 (blocks Na_V1.7) (see Fig. 9B). This suggests redundancy of recruitment for Na_V1.1 and Na_V1.2 in AP initiation. There was a trend for the inhibition evoked specifically by PF771 to be increased by the presence of ICA, suggesting further redundancy between Na_V1s but this did not reach significance.

It is not possible using these agents alone to definitively ascribe a role for Na_V1.6 in these responses. However, ∼45% of these neurons lack Na_V1.1 and almost all lack Na_V1.3. Thus, the only ICA-sensitive channel in this neuronal subset would be Na_V1.6. Given that under some conditions, ICA has a direct inhibitory role on >80% of fibers, this would suggest that Na_V1.6 also contributes to AP discharge.

We noted that the responses of some mechanoreceptors to distension began to adapt following treatment with Na_V1 inhibitors. Despite a trend for greater inhibition of the 2nd 10-s period of the distension compared with the 1st 10-s period (particularly for the 30 mmHg distension), this was only significant for the responses of fibers to 30 mmHg following treatment with LSN. This suggests that the Na_V1 subunits responsible for the AP discharge in the presence of LSN (likely to be Na_V1.1) may be less able to maintain repetitive APs than LSN-sensitive Na_V1s during the high-intensity generator potential evoked by 30 mmHg stimulus. However, Na_V1.1 has been shown in patch clamp studies in heterologous systems to have faster recovery from inactivation and a larger window current than other Na_V1s (34), both of which would enable repetitive AP firing. As such it is possible that the reduced ability to maintain distension-evoked AP discharge during LSN treatment is because there is only limited number of Na_V1.1 channels available for activation in the terminal. However, as mentioned earlier, Na_V1s’ biophysical properties in the afferent terminal/axon are likely modified by unidentified auxiliary β subunits and posttranslational modifications (1, 36). Furthermore, repetitive AP firing to current steps in patch clamp studies of dissociated A-type vagal neurons has been shown to be dependent on currents within the afterhyperpolarization as well as TTX-sensitive currents (30). How these voltage-dependent currents respond to the generator potential evoked by mechanosensitive channels in the afferent nerve terminal is therefore unclear. More work is needed to understand the impact of Na_V1 blockers on mechanoreceptor adaptation.

Previous studies have shown that AP initiation and conduction in vagal afferents innervating the guinea pig esophagus and airways may be mediated by distinct Na_V1s (17, 26). Here, we found that AP conduction in both tension mechanoreceptors and capsaicin-sensitive nociceptors was completely dependent on TTX-sensitive channels, similar to previous vagal afferent studies (17, 26, 39). Despite the limited inhibition of AP initiation from esophageal mechanoreceptors by PF771 and ICA, these inhibitors caused substantial inhibition of AP conduction in desheathed vagal esophageal afferent fibers. Thus, in some afferents AP conduction was completely dependent on Na_V1.7 despite the presence of other TTX-sensitive Na_V1s in >90% of neurons. In many neurons, however, Na_V1.7 function was not sufficient for AP conduction, as ICA completely prevented conduction in ∼40% of tension mechanoreceptors. Na_V1.7 activity has previously been shown to be necessary and sufficient for all AP conduction in vagal afferents innervating the guinea pig airways (26), whereas Na_V1.7 was only necessary for conduction in 50% of C-fibers innervating the guinea pig esophagus (17).

In summary, multiple Na_V1 α subtypes are expressed in tension mechanoreceptors innervating the mouse esophagus and we have provided evidence of a complex contribution of Na_V1.1, Na_V1.2, and Na_V1.7 (and likely Na_V1.6) in both AP initiation and conduction. It is likely that there is some redundancy in the function of these Na_V1s in AP initiation, but we have little evidence that their recruitment is dependent on stimulus intensity. Instead, our data indicate that all inhibitors tested are more effective in blocking responses to high stimulus intensities, but this is unlikely to be solely due to state-dependent inhibition. There appears to be less redundancy in the function of Na_V1s in AP conduction.

Pharmacological inhibition of Na_V1s is a rational approach for targeting multiple pathological conditions involving excitable cells (1, 40). For example, pain and other sensory disorders involving inappropriate electrical activity in afferent nerves could be reduced by blocking Na_V1 function (2). Targeted administration of local anesthetics such as lidocaine blocks multiple Na_V1s, reducing regional pain (41). However, systemic use of Na_V1 inhibitors is hampered by the selectivity of the agent for specific Na_V1s and the complex role that each specific channel plays in the physiology of excitable systems. New selective Na_V1 inhibitors are being developed (25, 37, 42). Here, we show that a selective inhibitor of a single Na_V1 would likely have minimal effect on the physiological activity of esophageal tension mechanoreceptors. As such, physiological afferent regulation of esophageal motility is unlikely to be reduced by systemic selective Na_V1 therapies for other neurological conditions (e.g., pain). Furthermore, our data suggest that selective inhibition of individual Na_V1 channels is unlikely to be an effective approach for the inhibition of dysfunctional esophageal tension mechanoreceptor activity, which is thought to contribute to ineffective esophageal motility and lower esophageal sphincter function associated with spastic disorders of the esophagus and noncardiac chest pain (12–14).

GRANTS

This work was funded by the National Institutes of Health Stimulating Peripheral Activity to Relieve Conditions (SPARC) Grant U01DK116311 (to M. Kollarik and Thomas E. Taylor-Clark).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.K. and T.E.T.-C. conceived and designed research; S.H., M.J.P., and N.P. performed experiments; S.H., M.J.P., N.P., M.K., and T.E.T.-C. analyzed data; S.H., M.J.P., N.P., and T.E.T.-C. interpreted results of experiments; S.H., M.J.P., N.P., and T.E.T.-C. prepared figures; M.K. and T.E.T.-C. drafted manuscript; S.H., M.J.P., N.P., and T.E.T.-C. edited and revised manuscript; S.H., M.J.P., N.P., and T.E.T.-C. approved final version of manuscript.