Differential modulation of voltage-gated sodium channels by nerve growth factor in three major subsets of TrkA-expressing nociceptors.

Nerve growth factor is an inflammatory mediator that induces long-lasting hyperalgesia, which can partially be attributed to nerve growth factor-induced sensitization of primary afferent nociceptors. It was shown that nerve growth factor increases the excitability of polymodal C-fibre nociceptors by modulating tetrodotoxin-sensitive and tetrodotoxin-resistant voltage-gated sodium channels, but hitherto only little is known about the effects of nerve growth factor on sodium currents in other nociceptor subtypes that express the nerve growth factor receptor TrkA. We previously characterized two reporter mouse lines that allow the unequivocal identification of two important subclasses of TrkA-expressing nociceptors - i.e. neuropeptide Y receptor type 2 (NPY2R+ ) Aδ-fibre nociceptors that mediate pinprick pain and nicotinic acetylcholine receptor alpha-3 subunit (CHRNA3+ ) silent nociceptors, which are the most abundant TrkA+ nociceptors in visceral organs and deep somatic tissues. Here, we utilized these mouse lines to investigate the expression patterns and the possible nerve growth factor-dependent modulation of sodium channels in these neurons using whole-cell patch-clamp recordings and quantitative real-time polymerase chain reaction. We demonstrate that NPY2R+ nociceptors, CHRNA3+ 'silent' nociceptors and polymodal C-fibre nociceptors express different combinations of sodium channel α- and β-subunits and accordingly exhibit functionally different sodium currents. Moreover, we demonstrate that nerve growth factor produces robust hyperpolarizing shifts in the half-activation voltage of tetrodotoxin-resistant currents in NPY2R+ nociceptors and polymodal C-fibre nociceptors and also shifts the half-activation of tetrodotoxin-sensitive currents in polymodal C-fibre nociceptors. In silent nociceptors, however, nerve growth factor solely increases the current density of the tetrodotoxin-resistant current but does not alter other sodium channel properties. Considering the different peripheral target tissues and the previously reported roles in different forms of pain of the nociceptor subpopulations that were examined here, our results suggest that nerve growth factor differentially contributes to the development visceral and cutaneous pain hypersensitivity and highlights the importance of developing different therapeutic strategies for different forms of pain.

Introduction

Nerve growth factor (NGF) is not only a neurotrophic factor that is important for the embryonic development of the sympathetic and sensory nervous system, but it is also a potent inflammatory mediator that plays a crucial role in the development and maintenance of chronic inflammatory pain.[1,2] Thus, the NGF levels are increased in inflamed tissues in several human pain disorders including interstitial cystitis,[3] irritable bowel syndrome,[4] chronic pancreatitis[5] and osteoarthritis,[6] and administration of NGF produces profound and long-lasting thermal and mechanical hyperalgesia in rodents and humans.[7-11] Accordingly, antibodies that block NGF signalling by sequestering endogenously produced NGF, have proven great efficacy in alleviating pain in rodents[12] and, most importantly, in humans suffering from the aforementioned pain disorders.[1,13]

The cellular and molecular basis of NGF-induced thermal hyperalgesia is well understood, but only little is known about the mechanism underlying mechanical hyperalgesia.[2] Thus, thermal hyperalgesia results from the sensitization of the heat and capsaicin-sensitive ion channel TRPV1 in polymodal C-fibre nociceptors.[14-16] Regarding mechanical hyperalgesia, both, central as well as peripheral sensitization, mechanisms have been proposed to be involved.[1,2] Central sensitization is thought to be, at least partially, mediated by an NGF-induced increase in brain-derived neurotrophic factor release from the central terminals of nociceptors, which causes a strengthening of synaptic transmission between nociceptors and second-order neurons in the spinal dorsal horn via modulation of post-synaptic N-methyl-D-aspartic acid receptors.[17]

Peripheral sensitization, that is increased sensitivity of nociceptive primary afferents to noxious mechanical stimuli, is thought to result from a combination of alterations of different functional properties of nociceptors. One possibility is that the yet unknown ion channels that mediate mechanotransduction in nociceptors – i.e. the conversion of noxious mechanical stimuli into electrical signals – are sensitized by NGF, such that a given mechanical stimulus causes a larger depolarization of the peripheral sensory endings and hence elicits more action potentials. Indeed, it was shown that mechanotransduction currents in small diameter nociceptors are sensitized by direct activation of protein kinase C or treatment with a mixture of the inflammatory mediators bradykinin, prostaglandin E2, histamine and serotonin, only when the cells are pre-exposed to NGF.[18,19] Moreover, we have recently shown that a subset of peptidergic nociceptors that are characterized by the expression of the alpha-3 subunit of the nicotinic acetylcholine receptor (CHRNA3) and that are normally completely insensitive to mechanical stimuli becomes mechanosensitive following treatment with NGF.[19] Considering that CHRNA3+ nociceptors account for almost half of all peptidergic nociceptors in visceral organs, muscles and joints,[19] it is tempting to speculate that the NGF-induced un-silencing of these afferents would greatly increase nociceptive input to projection neurons in the spinal dorsal horn and may thus significantly contribute to mechanical hyperalgesia.

In addition to mechanotransduction channels, voltage-gated sodium channels (VGSCs) are also major determinants of nociceptor sensitivity as they set the action potential threshold and the maximum firing frequency. Accordingly, modulation of VGSCs might also significantly contribute to mechanical hyperalgesia.[20,21] Nociceptors predominantly express four of the nine known isoforms of VGSC alpha subunits, namely, the tetrodotoxin-sensitive (TTX-S) isoforms Nav1.6 and Nav1.7 as well as the TTX-resistant (TTX-R) isoforms Nav1.8 and Nav1.9.[20,21] NGF was shown to modulate voltage-gated sodium currents via both transcriptional regulation as well as post-translational modifications.[1,2,20,21] Thus, NGF was shown to increase the current density and to cause a hyperpolarizing shift in the voltage dependence of activation of Nav1.8. The former effect appears to require p38 mitogen-activated protein (MAP) kinase,[22] whereas the latter is mediated by a signalling cascade that involves the second messenger ceramide and the atypical protein kinase PKMζ.[23,24] The increase in current density is commonly attributed to an NGF-induced upregulation of Nav1.8, but in fact there is only indirect evidence for this hypothesis and the reports regarding this matter are somewhat conflicting. Thus, while some studies show that NGF reverses axotomy-induced downregulation of Nav1.8[25,26] and that in-vivo NGF deprivation causes a downregulation of Nav1.8,[27] other found no differences in the expression of Nav1.8 in dorsal root ganglion (DRG) neurons isolated from animals in which hyperalgesia had been induced with NGF.[28,29] Nav1.7 expression was also shown to be upregulated by NGF[29] and there is indirect evidence suggesting that NGF might also shift the voltage dependence of activation of Nav1.7 to more negative potentials, because ERK1/2 MAP kinase, which is a downstream target of NGF-TrkA signalling, was shown to do so in heterologous systems.[30]

Most of the above-mentioned studies that investigated the modulation of VGSCs by NGF in DRG cultures using whole-cell patch-clamp recordings have focused on polymodal C-fibre nociceptors that express the NGF receptor TrkA, because these neurons can readily be identified in cultures as small diameter cells that are not labelled by Isolectin B4 (IB4) and respond to the TRPV1 agonist capsaicin.[1,21] The TrkA receptor is, however, also expressed by other nociceptor subpopulations, but the lack of tools that would allow the unequivocal identification of these subpopulations in DRG cultures has hitherto precluded the thorough examination of the effects of NGF on sodium currents in these physiologically important nociceptor subtypes. We have recently generated a mouse line in which enhanced yellow fluorescent protein (EYFP)-tagged Channelrhodopsin 2 is exclusively expressed in sensory neurons that express the neuropeptide Y receptor type 2 (NPY2R). We showed that NPY2R+ neurons are myelinated TrkA+ Aδ-fibre nociceptors and demonstrated that these nociceptors are required for the detection of pinprick stimuli.[31] Moreover, as mentioned earlier, we have recently characterized a reporter mouse line in which green fluorescent protein (GFP) is expressed in CHRNA3+ neurons and demonstrated that these neurons constitute TrkA+ ‘silent’ C-fibre nociceptors.[19]

Here, we utilized the NPY2R and CHRNA3 reporter mouse lines to investigate the modulation of voltage-gated sodium currents by NGF in Aδ-fibre nociceptors and silent nociceptors in DRG cultures, using whole-cell patch-clamp recordings and quantitative real-time polymerase chain reaction (qRT-PCR).

Materials and methods

Animals

CHRNA3-EGFP mice, official name Tg(Chrna3-EGFP)BZ135Gsat/Mmnc (RRID:MMRRC_000243-UNC) and Npy2rCre/+, official name Tg(Npy2r-cre)SM19Gsat/Mmucd (RRID:MMRRC_036630-UCD) and ChR2loxP/+ animals (B6;129S-Gt(ROSA)26Sortm32(CAG-COP4* H134R/EYFP)Hze/J; RRID:IMSR_JAX:012569) were obtained from Jackson Laboratories. Mice were housed in the Interfacultary Biomedical Facility of Heidelberg University according to the institutional guidelines, and all animal experiments were carried out according to the German Animal Protection Law and with permission of the Regierungspraesidium Karlsruhe (T-57/16).

DRG cell culture

Twelve to eighteen weeks old female mice were killed by placing them in a CO2-filled chamber for 2 to 4 min followed by cervical dislocation. L2–L5 DRGs were collected in Ca2+ and Mg2+-free phosphate-buffered saline (PBS) and incubated in a mixture of Collagenase type I (2.0 μg/ml, Sigma) and Trypsin (1.5 μg/ml, Sigma) for 60 min at 37°C. Digested DRG’s were washed twice with growth medium (Dulbecco’s modified Eagle’s medium-F12 (Invitrogen) supplemented with L-glutamine (2 µM, Sigma), glucose (8 mg/ml, Sigma), penicillin (200 U/ml)–streptomycin (200 µg/ml) (both Life Technologies) 5% fetal horse serum (Life Technologies)), triturated with a fire-polished Pasteur pipette and plated in a droplet of growth medium on glass coverslips precoated with poly-L-lysine (20 µg/cm2, Sigma) and laminin (4 µg/cm2, Life Technologies). To allow neurons to adhere, the coverslips were kept for 3 to 4 h at 37°C in a humidified 5% incubator before being flooded with fresh growth medium. Cultures were used for patch-clamp experiments on the next day.

Immnuohistochemistry

DRGs were dissected in ice-cooled PBS, fixed with 4% paraformaldehyde for 30 min at 4°C and incubated overnight in 30% sucrose at 4°C. DRGs were then embedded in Tissue-Tek O.C.T compound and cut into 16 µm cryo-sections. After drying, sections were incubated in 50 mM Glycine for 20 min, washed twice with PBS with Tween 20 (PBST) (0.2%), blocked with PBST (0.2%) + 10% donkey serum + 1% bovine serum albumin and then incubated with primary antibodies for 1 h at room temperature (RT). Primary antibodies were diluted in PBST (0.2%) + 10% donkey serum. Sections were then washed four times with PBST (0.2%), subsequently incubated with secondary antibodies for 1 h at RT, washed with PBST four times, dried and mounted with fluorogel (Fluoprobes).

Antibodies

The following primary antibodies were used: rat anti-GFP (Nacalai tesque, #04404–84, 1:3000; RRID:AB_10013361), mouse anti-neurofilament heavy polypeptide (NEFH) (Sigma-Aldrich, N0142, 1:600, RRID:AB_477257), goat anti-TrkA (R&D, AF1056, 1:200, RRID:AB_2283049) and Isolectin GS-IB4-Alexa Fluor® 568 Conjugate (Life technologies, 3 µg/ml). Secondary antibodies were AlexaFluor-488 donkey anti-rat (Life technologies, AF21208, 1:500, RRID:AB_2535794), AlexaFluor-594 donkey anti-mouse (Life technologies, AF21203 1:500, RRID:AB_2535789) and AlexaFluor-633 donkey anti-goat (Life technologies, A21082, 1:500, RRID:AB_10562400).

Electrophysiology

Whole-cell patch-clamp recordings were made at RT (20–24°C). Patch pipettes with a tip resistance of 0.8 to 1.5 MΩ were pulled (Flaming-Brown puller, Sutter Instruments, Novato, CA, USA) from borosilicate glass capillaries (BF150-86–10, Sutter Instrument), filled with a solution consisting of 120 mM CsCl, 10 mM NaCl, 20 mM TEA-Cl, 1 mM 4-aminopyridine, 1 mM EGTA, 10 mM HEPES, 2 mM guanosine 5′-triphosphate (GTP) and 2 mM adenosine 5′-triphosphate (ATP) adjusted to pH 7.3 with CsOH. The bathing solution contained 144 mM NaCl, 2 mM CaCl2, 1 mM MgCl2, 4 mM glucose, 10 mM HEPES and 100 µM CdCl2 adjusted to pH 7.4 with NaOH. TTX and capsaicin were applied with a gravity-driven multi-barrel perfusion system (Valvelink8.2, Automate Scientific). All recordings were made using an EPC-10 amplifier (HEKA, Lambrecht, Germany) in combination with Patchmaster© and Fitmaster© software (HEKA). Pipette and membrane capacitance were compensated using the auto function of Patchmaster, and series resistance was compensated by >70% to minimize voltage errors. Using this approach, the maximal voltage errors shown in Table 1 for TTX-S and TTX-R currents were achieved.

Currents were sampled at 20 kHz, filtered at 2.9 kHz and linear leak currents were subtracted using a P/4 pulse protocol. Whole-cell sodium currents were elicited by a series of 40 ms step depolarizations from a holding potential of −120 mV to test potentials ranging from −70 mV to +60 mV in increments of 5 mV at an interval of 1 s. The TTX-S currents were isolated by subtracting the current recorded in the presence of 300 nM TTX (TTX-R current) from the total current recorded in the absence of TTX. For conductance–voltage curves, Na+ conductance (G) was calculated was calculated using the following equation: G = INa/(V-ENa), where INa is the peak current amplitude at the test potential V, and ENa is the reversal potential of sodium (+62.9 mV), which was calculated with the Nernst equation. To determine the voltage at which half of the sodium channels are activated (V1/2), G/Gmax curves were fit with a Boltzmann equation: G/Gmax = 1/[1+exp(V1/2–V/k)], where Gmax is the maximal conductance, V is the test potential and k is the slope factor.

Fura2-Ca2+ imaging

To identify peptidergic C-fibre nociceptors – i.e. IB4− neurons that respond to the TRPV1 agonist capsaicin – for the collection of cell samples for qRT-PCR, FURA-2-Ca2+ imaging was used. To this end, cultured DRG neurons loaded with FURA-2 (25 µM) and labelled with 1.5 µg/ml Isolectin GS-IB4-Alexa Fluor 568 conjugate at 37°C for 25 min and then washed with PBS for 10 min. Neurons were placed in a chamber containing extracellular buffer consisting of 140 mM NaCl, 1 mM MgCl2, 2 mM CaCl2, 4 mM KCl, 4 mM glucose and 10 mM HEPES adjusted to pH 7.4. Cells were illuminated alternately at 340 nm and 380 nm (100 ms exposure time Polychrome V, Visitron Systems) at a sampling frequency of 2 Hz, and Fura-2 fluorescence images were acquired with a ProgRes MFCool charge-coupled device (CCD) camera (Jenoptik, Germany). The Polychrome V and the ProgRes MFCool CCD camera were controlled by the imaging extension of the PatchMaster software (HEKE Elektronik). Fura-2 ratio images were generated with the ImageJ software. Neurons with a cell soma diameter <25 µm that were not labelled by IB4 and exhibited an increase in the Fura-2 signal in response to capsaicin were considered as peptidergic polymodal C-fibre nociceptors. To ensure that we did not collect silent nociceptors, some of which are also small and respond to capsaicin, all IB4−/caps+ neurons were collected from cultures prepared from CHRNA3EGFP mice and the lack of a GFP signal was used as an additional selection criterion.

Cell sample collection and qRT-PCR

mRNA expression levels of VGSCs and beta subunits were determined as follows. From each nociceptor subpopulation (CHRNA3+, NPY2R+ or IB−/caps+) and each treatment condition (control and NGF), four cell samples (20 cell per sample) were collected from DRG cultures by aspirating the cells into a patch-clamp pipette with a tip diameter of 25 µm, filled with 2 to 4 µl PBS containing 4 U/µl RNAseOUT (Thermofisher). cDNA synthesis was carried out directly on the sample using the Power SYBR® Green Cells-to-CT™ Kit (Life Technologies) according to the manufacturers’ instructions. qRT-PCR reactions were set up using FastStart Essential DNA Green Master (Roche) by adding 4 µL of the obtained cDNA as template, and the primer pairs mentioned in Table 2 were used at a concentration of 250 nM.qRT-PCR reactions were performed in a LightCycler 96 (Roche) with a thermal cycler profile as follows: 10 min preincubation step at 95°C followed by 40 cycles of PCR with a 10 s denaturing cycle at 95°C, followed by 10 s of annealing at 60°C and 10 s extension at 72°C.

Table 2.

List of primers used for qRT-PCR.

Nav1.6_ FWD	5′-CAGCAGGTCGAGAAATGTCTTTC-3′
Nav1.6_ REV	5′- CTCAAGCATTTCTGCCATTTCC-3′
Nav1.7_ FWD	5′-GTGGGCGAATTCACCTTCCT-3′
Nav1.7_ REV	5′-TCGAAGAGCTGAAACATTGCCTA-3′
Nav1.8_FWD	5′-ATAAGGGGCCAGGACATCTTC-3′
Nav1.8_REV	5′-AAACACGAAGCCCTGGTACT-3′
Nav1.9_FWD	5′-TGAGGCAACACTACTTCACCAATG-3′
Nav1.9_REV	5′-AGCCAGAAACCAAGGTACTAATGATG-3′
Beta1_ FWD	5′-TGCATGATGGGTGAAGCAAT-3′
Beta1_ REV	5′-GCCTATGTGCAAATCAGCAAAA-3′
Beta2_ FWD	5′-ATACCCGCCTGCCCTGTACC-3′
Beta2_ REV	5′-TCTCCAAACCGCTCCAGCTTC-3′
Beta3_ FWD	5′-TATAAATGAGAAAGCCCTGGTTCTG-3′
Beta3_ REV	5′-GCCTAACCTAGTGCTCGACACA-3′
Beta4_ FWD	5′-GGATGCCCAAATGAGGGATT-3′
Beta4_ REV	5′-CCAGAGGACTAAACCATGCATAAGA-3′
GAPDH- FWD	5′-GCATGGCCTTCCGTGTTC-3′
GAPDH-REV	5′-GTAGCCCAAGATGCCCTTCA-3′

GAPDH: Glyceraldehyde 3-phosphate dehydrogenase.

Statistics

All statistical analyses were made using GraphPad Prism 7. The statistical tests that were used are mentioned in the figure legends and in the main text. N-numbers for each experiment – i.e. numbers of biological replicates – are provided in the figure or in the figure legend. To select the appropriate test, all data sets were tested for Gaussian distribution using the Kolmogorov–Smirnov test.

Results

NGF exerts its effects via activation of the TrkA receptor, which is expressed in various functionally different subtypes of nociceptors including polymodal C-fibre nociceptors, mechanically insensitive ‘silent’ C-fibre nociceptors as well as myelinated Aδ-fibre nociceptors. Most of our current knowledge about the modulation of VGSCs by NGF is, however, limited to small diameter C-fibre nociceptors. Here, we have utilized two recently characterized reporter mouse lines that allow the unequivocal identification of silent nociceptors and Aδ-fibre nociceptors in DRG cultures in order to compare voltage-gated sodium currents and their possible modulation by NGF in all three subpopulations of TrkA+ nociceptors. Since there is strong evidence for sex- and age-dependent differences in pain signalling both in mice and humans,[32,33] we only examined sensory neurons from fully matured (12–18 weeks old) female mice to ensure that these factors do not confound our results.

The properties of VGSCs differ between the three major subpopulations of TrkA+ nociceptors

To identify silent nociceptors, we prepared DRG cultures from CHRNA3-EGFP mice in which EGFP is exclusively expressed in this subset of nociceptors, which are unmyelinated as indicated by the lack of NEFH expression, have small-to-medium cell soma diameters and express the NGF receptor TrkA (Figure 1(a)).[19] Aδ-fibre nociceptors were visualized using NPY2RChR2 mice in which EYFP-tagged channelrhodopsin 2 expression is confined to myelinated TrkA+ neurons with large cell bodies (Figure 1(b)), as previously described.[31] Finally, polymodal C-fibre nociceptors were identified in cultures from CHRNA3-EGFP mice as EGFP−/IB4− small diameter neurons (< 25 µm) that were sensitive to capsaicin (Figure 1(c)).

Figure 1.

Differences in TTX-S sodium currents in three major subsets of TrkA-expressing neurons. (a) Immunohistochemistry of a DRG section showing that CHRNA3+ neurons co-express the TrkA receptor but not NEFH. (b) Immunohistochemistry showing that NPY2R+ neurons express both TrkA and NEFH. (c) Representative example traces of TTX-S currents (bottom traces) elicited by the indicated pulse protocol (top trace). The rightmost trace shows a typical response of a small diameter IB4− neuron to 500 nM capsaicin. (d) Current–voltage relationship of the TTX-S currents (mean current density ± SEM) in the indicated cell types. (e) Comparison of the current densities of the maximal TTX-S current in the indicated cell types using one-way ANOVA and Holm–Sidak’s multiple comparison test. Note that the current densities of the TTX-S current were significantly bigger in NPY2R+ neurons as compared to the other two nociceptor subtypes. (f) Boltzmann fits of the conductance–voltage plots showing that TTX-S currents in NPY2R+ and CHRNA+ activate at more negative potentials compared to TTX-S currents in IB4− neurons. (g) Comparison of the V1/2 ± SEM values obtained by the Boltzmann fits shown in (f). Boltzmann fits (V1/2 values) were compared pairwise using the extra sum-of-squares F-test. ****P < 0.0001. TTX: tetrodotoxin; IB4: Isolectin B4; NPY2R: neuropeptide Y receptor type 2; NEFH: neurofilament heavy polypeptide; CHRNA3: nicotinic acetylcholine receptor alpha-3 subunit.

Whole-cell sodium currents elicited by a series of 40 ms step depolarizations from −120 mV to test potentials ranging from −70 mV to +60 mV in increments of 5 mV were recorded before and after application of 300 nM TTX. The TTX-S currents were isolated by subtracting the TTX-R current from the total current recorded in the absence of TTX. The current–voltage relationship showed that TTX-S currents in CHRNA3+ neurons and NPY2R+ neurons activated at more hyperpolarized potentials than TTX-S currents in IB4−/caps+ nociceptors (Figure 1(a)) and revealed differences in current densities between the three nociceptor subpopulations (Figure 1(e)). Thus, the maximum current densities of the TTX-S currents in NPY2R+ neurons were significantly bigger than the TTX-S current densities in the other two populations (NPY2R+ Imax-TTX-S = −1052 ± 123.8 pA/pF; CHRNA3+: Imax-TTX-S = 450.2 ± 112.4 pA/pF; IB4−/caps+: Imax-TTX-S = −609.2 ± 71.1 pA/pF; Figure 1(e)). Boltzmann fits of the conductance–voltage relations shown in Figure 1(f) revealed that the voltage for half-maximal activation (V1/2) of TTX-S currents was significantly more negative in CHRNA3+ (V1/2 = −26.74 ± 0.51 mV) and NPY2R+ (V1/2 = −30.29 ± 0.37 mV) neurons as compared to IB4−/caps+ nociceptors (V1/2 = −14.22 ± 0.82 mV) (Figure 1(g)). The TTX-R currents in the three nociceptor subpopulations (Figure 2(a)) also differed from one another in several aspects. Thus, the current densities of TTX-R currents were significantly bigger in NPY2R+ nociceptors (−272.7 ± 30.4 pA/pF) as compared to CHRNA3+ neurons (−188.4 ± 21.7 pA/pF) and IB4−/caps+ nociceptors (−140.0 ± 24.3 pA/pF) (Figure 2(b) and (c)). Moreover, TTX-R currents in CHRNA3+ neurons were activated at significantly more negative potentials (V1/2 = −16.8 ± 0.5 mV) as compared to TTX-R currents in NPY2R+ neurons, which had a V1/2 of −7.7 ± 1.0 mV, and IB4−/caps+ nociceptor with a V1/2 of −4.3 ± 0.6 mV (Figure 2(d) and (e)).

Figure 2.

Differences in TTX-R sodium currents in three major subsets of TrkA-expressing neurons. (a) Representative example traces of TTX-R currents (bottom traces) elicited by the indicated pulse protocol (top trace). (b) Current–voltage relationship of the TTX-R currents (mean current densities ± SEM) in the indicated cell types. (c) Comparison of the current densities of the maximal TTX-R current in the indicated cell types using one-way ANOVA and Holm-Sidak’s multiple comparison test. Note that the current densities of the TTX-R current in NPY2R+ neurons are significantly bigger than in the other two subtypes. (d) Boltzmann fits of the conductance–voltage plots showing that TTX-R currents in NPY2R+ and CHRNA+ activate at more negative potentials compared to TTX-R currents in IB4− neurons. (g) Comparison of the V1/2 ± SEM values obtained by the Boltzmann fits shown in (f). Boltzmann fits (V1/2 values) were compared pairwise using the extra sum-of-squares F-test. ****P < 0.0001. TTX: tetrodotoxin; IB4: Isolectin B4; NPY2R: neuropeptide Y receptor type 2; CHRNA3: nicotinic acetylcholine receptor alpha-3 subunit.

To test if the observed differences in current densities and in half-activation voltages of TTX-S and TTX-R currents (Figures 1 and 2) result from differences in the expression levels of certain VGSC α- and β-subunits, we next quantified the mRNA levels of Nav1.6, Nav1.7, Nav1.8 and Nav1.9, which are the most abundant α-subunit isoforms in DRG neurons, as well as the mRNA levels of the β1, β2, β3 and β4 subunits using qRT-PCR. This analysis revealed that NPY2R+ neurons express Nav1.6 and the β-subunits β1 and β4 at significantly higher levels than IB4−/caps+ neurons (Figure 3(a)). Moreover, we observed a trend towards higher expression levels of Nav1.7, Nav1.8 and β2 in NPY2R+ neurons as compared to IB4−/caps+ neurons, but these differences were not statistically significant (Figure 3(a)). CHRNA3+ neurons exhibited a similar sodium channel expression profile as NPY2R+ neurons, with higher levels of Nav1.6, Nav1.7, Nav1.8, β2 and β4 than IB4−/caps+ neurons, but none of the differences was statistically significant.

Figure 3.

Comparison of the expression patterns of sodium channel α and β subunits. (a) Bars show the mean ± SEM mRNA expression levels of the indicated sodium channel α- and β-subunits normalized to the expression levels of the housekeeping gene GAPDH in the same samples. The expression levels of the individual samples are indicated by the symbols. The mean expression levels were compared using one-way ANOVA and Holm–Sidak’s multiple comparison test and where significant differences were observed, the exact P values are provided in the graph. TTX: tetrodotoxin; GAPDH: Glyceraldehyde 3-phosphate dehydrogenase; IB4: Isolectin B4; NPY2R: neuropeptide Y receptor type 2; CHRNA3: nicotinic acetylcholine receptor alpha-3 subunit.

Taken together, our electrophysiological characterization together with the results of the qRT-PCR indicate that each of the three major subsets of TrkA+ nociceptors expresses a different combination of sodium channel α- and β-subunits, which results in the generation of functionally different sodium currents in each of the three subpopulations.

NGF differentially modulates VGSC in different subsets of TrkA+ nociceptors

We next asked if the TTX-S and TTX-R currents are differentially modulated by NGF in the three nociceptor subpopulations. To this end, 100 ng/ml NGF was added to the growth medium at the time of cell plating, and sodium currents were recorded on the next day – i.e. usually between 20 and 28 h after the addition of NGF. NGF treatment did not alter any of the examined properties of TTX-S currents in CHRNA3+ neurons (Figure 4(a) to (d)). However, NGF did induce a significant increase in the TTX-R current density in these neurons, from −188.4 ± 21.7 pA/pF under control conditions to −291.3 ± 22.2 pA/pF after NGF treatment (Figure 4(e) to (f)). The half-activation voltage of TTX-R currents was only minimally affected (Figure 4(g) to (h)). Surprisingly, the effects of NGF on NPY2R+ neurons were completely different. Thus, NGF treatment caused a small but significant reduction in the TTX-S current densities (CTL: −1052 ± 123.8 pA/pF vs. NGF: −694.3 ± 73.2 pA/pF; Figure 5(a) and (b)); the voltage dependence of activation of TTX-S current was not altered (TTX-S-CTL-V1/2 = −30.3 ± 0.4 mV vs. TTX-S-NGF-V1/2 = −29.4 ± 0.7 mV; Figure 5(c) and (d)). Unlike in CHRNA3+ neurons, NGF treatment did not alter the TTX-R current densities in NPY2R+ (Figure 5(e) and (f)) but instead shifted the half-activation voltage by ∼7 mV to more negative potentials (TTX-R-CTL-V1/2 = −8.7 ± 0.7 mV vs. TTX-R-NGF-V1/2 = −15.1 ± 1.0 mV; Figure 5(g) and (h)). The sodium currents in IB4−/caps+ neurons were modulated in a similar manner as those in NPY2R+ neurons. Thus, the TTX-S current densities in IB4−/caps+ neurons were slightly, though not significantly, reduced by NGF treatment (Figure 6(a) and (b)). Interestingly, in contrast to NPY2R+ and CHRNA3+ neurons, NGF treatment significantly shifted the voltage dependence of activation to more negative potentials in IB4−/caps+ neurons (TTX-S-CTL-V1/2 = −13.9 ± 0.7 mV vs. TTX-S-NGF-V1/2 = −19.6 ± 1.3 mV; Figure 6(c) and (d)). TTX-R current densities were also slightly, but not significantly, increased (Figure 6(e) and (f)) and, as in NPY2R+ neurons, NGF shifted the V1/2 of the TTX-R currents in IB4−/caps+ neurons by ∼12 mV to more negative potentials (TTX-R-CTL-V1/2 = −3.9 ± 0.5 mV vs. TTX-R-NGF-V1/2 = −16.4 ± 1.3 mV; Figure 6(g) and (h)). We next asked if the effects of NGF on sodium current properties could, at least partially, be attributed to changes in the expression levels of sodium channel α- and β-subunits. To this end, we compared the mRNA levels of Nav1.6–1.9 and β1–4 in cells that were cultured for 24 h in normal growth medium with cells that were cultured in the presence of 100 ng/ml NGF for the same period of time. Surprisingly, the only noteworthy change that we could observe was a more than two-fold down-regulation of the mRNA levels of Nav1.8 in IB4−/caps+ neurons, which, however, was not statistically significant (Figure 7(a) to (c)). All other sodium channel α- and β-subunits appeared to be expressed at the same level in the absence and presence of NGF.

Figure 4.

Modulation of TTX-S and TTX-R currents by NGF in silent nociceptors. (a) Current–voltage relationship of the TTX-S currents (mean current densities ± SEM) in CHRNA3+ neurons cultured for 24 h in normal growth medium (CTL; green circles) and 24 h in the presence of 100 ng/ml NGF (NGF, half-filled circles). (b) Comparison of the current densities of the maximal TTX-S current in control conditions (CTL, green bars) and in the presence of NGF (hatched bars) using Student’s t-test. (c) Boltzmann fits of the conductance–voltage plots showing that NGF does not alter the voltage dependence of activation of TTX-S currents in CHRNA+ neurons. (d) Comparison of the V1/2 ± SEM values obtained by the Boltzmann fits shown in (c). Boltzmann fits (V1/2 values) were compared using the extra sum-of-squares F-test. (e) Current–voltage relationship of the TTX-R currents (mean current densities ± SEM) in CHRNA3+ neurons. (f) Comparison of the current densities of the maximal TTX-R current using Student’s t-test showing that NGF significantly increases the TTX-R current in CHRNA3+ neurons (g) Boltzmann fits of the conductance–voltage plots showing that NGF does not alter the voltage dependence of activation of TTX-R currents in CHRNA+ neurons. (h) Comparison of the V1/2 ± SEM values in the absence and presence of NGF obtained by the Boltzmann fits shown in (g). Boltzmann fits (V1/2 values) were compared using the extra sum-of-squares F-test. N-numbers were same in (a) to (h) and are provided below the bars in (b), (d), (f), and (h). The exact P values of the statistical comparisons are also indicated below the bar graphs. TTX: tetrodotoxin; CTL: control; NGF: nerve growth factor; CHRNA3: nicotinic acetylcholine receptor alpha-3 subunit.

Figure 5.

Modulation of TTX-S and TTX-R currents by NGF in NPY2R+ Aδ-nociceptors. (a) Current–voltage relationship of the TTX-S currents (mean current densities ± SEM) in NPY2R+ neurons cultured for 24 h in normal growth medium (CTL; orange circles) and 24 h in the presence of 100 ng/ml NGF (NGF, half-filled circles). (b) Comparison of the current densities of the maximal TTX-S currents in control conditions (CTL, orange bars) and in the presence of NGF (hatched bars) using Student’s t-test showing that NGF causes a small but significant reduction in TTX-S current density. (c) Boltzmann fits of the conductance–voltage plots showing that NGF does not alter the voltage dependence of activation of TTX-S currents in NPY2R+ neurons. (d) Comparison of the V1/2 ± SEM values in the absence and presence of NGF obtained by the Boltzmann fits shown in (c) using the extra sum-of-squares F-test. (e) Current–voltage relationship of the TTX-R currents (mean current densities ± SEM) in NPY2R+ neurons. (f) Comparison of the current densities of the maximal TTX-R current using Student’s t-test. (g) Boltzmann fits of the conductance–voltage plots showing that NGF shifts the voltage dependence of activation of TTX-R currents in NPY2R+ neurons to more hyperpolarized potentials. (h) Comparison of the V1/2 ± SEM values obtained by the Boltzmann fits shown in (g) using the extra sum-of-squares F-test showing that NGF treatment shifts the half-activation voltage by ∼7 mV from −8.7 ± 0.7 mV to −15.1 ± 1.0 mV. N-numbers were same in (a) to (h) and are provided below the bars in (b), (d), (f), and (h). The P values of the statistical comparisons are also indicated below the bar graphs. TTX: tetrodotoxin; CTL: control; NGF: nerve growth factor; NPY2R: neuropeptide Y receptor type 2.

Figure 6.

Modulation of TTX-S and TTX-R currents by NGF in polymodal C-fibre nociceptors. (a) Current–voltage relationship of the TTX-S currents (mean current densities ± SEM) in small diameter IB4− capsaicin-sensitive (caps+) neurons cultured for 24 h in normal growth medium (CTL; red circles) and 24 h in the presence of 100 ng/ml NGF (NGF, half-filled circles). (b) Comparison of the current densities of the maximal TTX-S current using Student’s t-test showing that NGF does not alter TTX-S current density in IB4−/caps+ neurons. (c) Boltzmann fits of the conductance–voltage plots showing that NGF shifts the voltage dependence of activation of TTX-S currents in IB4−/caps+ neurons to more hyperpolarized potentials. (d) Comparison of the V1/2 ± SEM values obtained by the Boltzmann fits shown in (c) using the extra sum-of-squares F-test. Note that NGF treatment shifts the half-activation voltage by ∼6 mV from −13.9 ± 0.7 mV to −19.6 ± 1.3 mV. (e) Current–voltage relationship of the TTX-R currents (mean current densities ± SEM) in the same cells shown in (a) to (d). (f) Comparison of the current densities of the maximal TTX-R current using Student’s t-test. (g) Boltzmann fits of the conductance–voltage plots showing that NGF shifts the voltage dependence of activation of TTX-R currents in IB4−/caps+ neurons to more hyperpolarized potentials. (h) Comparison of the V1/2 ± SEM values obtained by the Boltzmann fits shown in (g) using the extra sum-of-squares F-test showing that NGF treatment shifts the half-activation voltage by ∼12 mV from −3.9 ± 0.5 mV to −16.4 ± 1.3 mV. N-numbers were same in (a) to (h) and are provided below the bars in (b), (d), (f), and (h). The P values of the statistical comparisons are also indicated below the bar graphs. TTX: tetrodotoxin; CTL: control; NGF: nerve growth factor; IB4: Isolectin B4.

Figure 7.

NGF does not alter the expression levels of sodium channel α- and β-subunits. Comparisons of the mean ± SEM mRNA expression levels of the indicated sodium channel α- and β-subunits normalized to the expression levels of the housekeeping gene GAPDH in CHRNA3+ (a), NPY2R+ (b) and IB4−/caps+ (c) neurons cultured for 24 h in the absence (CTL) and neurons cultured for 24 h in the presence of 100 ng/ml NGF (NGF). The expression levels of the individual samples are indicated by the symbols. The expression levels were compared using Student’s t-test, but no significant changes in mRNA expression were detected. GAPDH: Glyceraldehyde 3-phosphate dehydrogenase; CTL: control; NGF: nerve growth factor.

Taken together, our results show that NGF differentially modulates voltage-gated sodium currents in different subpopulations of TrkA+ neurons and suggest that none of these effects is mediated by changes in the expression levels sodium channel α- or β-subunits.

Discussion

The goal of this study was to examine the properties of voltage-gated sodium currents in two recently identified subpopulations of TrkA+ nociceptors, namely, CHRNA3+ ‘silent’ C-fibre nociceptors[19] and NPY2R+ Aδ-fibre nociceptors[31] and to compare them with the well-characterized sodium channel properties of TrkA+ polymodal C-fibre nociceptors. Moreover, we asked if the inflammatory mediator NGF exerts different effects on VGSCs in these three nociceptor subpopulations. Indeed, we observed significant differences in the properties of voltage-gated sodium currents and found that NGF differentially modulates sodium currents in the three cell types.

In summary, we made the following observations:

TTX-S and TTX-R sodium currents in CHRNA3+ and NPY2R+ neurons are activated at more negative potentials than TTX-S currents in IB4−/caps+ neurons, which correlates with the different expression levels of β1 and β4 subunits in these cell types.

NPY2R+ neurons have significantly higher TTX-R current densities than CHNRA3+ and IB4−/caps+ neurons.

We observed that NGF shifts the half-activation voltage of TTX-R currents to more negative potentials in NPY2R+ and IB4−/caps+ neurons, but not in CHRNA3+ neurons, whereas it increased current densities of TTX-R current only in CHRNA3+ and to a smaller extent in IB4−/caps+ neurons.

We demonstrate that NGF shifts the half-activation voltage of TTX-S currents to more negative potentials in IB4−/caps+ neurons but not in CHRNA3+ and NPY2R+ neurons, whereas it reduces the current density of TTX-S currents in NPY2R+ neurons and to a small though not significant extent in IB4−/caps+ neurons.

Differences in the expression levels of sodium channel α- and β-subunits account for the functional differences in sodium current in the three nociceptor subpopulation

One of the key findings of this study was that TTX-S currents in CHRNA3+ and NPY2R+ nociceptors are activated at more negative potentials than TTX-S currents in IB4−/caps+ nociceptors (Figure 1(c) to (g)). One possible explanation for these differences is that the former two subpopulations express significant amounts of Nav1.6 and the β4-subunit (Figure 3). Nav1.6 has been shown to be activated at rather negative potentials, though it should be noted that the reported V1/2 values of Nav1.6 in DRG neurons vary between −35.9 mV and −18.7 mV in different studies,[34,35] and the β4-subunit was previously shown to shift the half-activation voltage of Nav1.6 by ∼8 mV to hyperpolarized potentials in heterologous systems.[36] Hence, the co-expression of these two proteins is most likely the reason for the extremely hyperpolarized half-activation voltages of TTX-S current in NPY2R+ and CHRNA3+ neurons. Our observation that NPY2R+ neurons, which are Aδ-fibre nociceptors, express high levels of Nav1.6 is consistent with previous reports showing that myelinated Aδ- and Aβ-fibres but not C-fibre nociceptors express Nav1.6.[37,38] We are not aware of any study that has examined the expression of Nav1.6 in silent nociceptors, but recently Feng et al.[39] demonstrated Nav1.6 expression in a significant proportion of Nav1.8-expressing colorectal afferents. Considering that CHRNA3+ afferents account for approximately one half of all nociceptive peptidergic colorectal afferents,[19] it is tempting to speculate that these fibres were also included in the population labelled by Feng et al.[39] For Nav1.7, V1/2 values around −25 mV have previously been reported[35,40] and there is no evidence suggesting that β-subunits alter the voltage dependence of Nav1.7 activation.[40-42] However, phosphorylation by ERK1/2 was shown to shift the V1/2 of Nav1.7 to more negative potentials.[30] Hence, another possible explanation for the hyperpolarized half-activation voltages is that the basal activity of ERK1/2 is higher in NPY2R+ and CHRNA3+ neurons.

Another important observation of our study was that the TTX-R currents in NPY2R+ neurons had significantly higher current densities than CHRNA3+ neurons and IB4−/caps+ neurons (Figure 2(c)). Since NPY2R+ neurons only express small amounts of Nav1.9 (Figure 3), most of the TTX-R current is probably carried by Nav1.8. The expression levels of Nav1.8 did not significantly differ between the three populations (Figure 3) and thus we propose that the high TTX-R current densities in NPY2R+ neurons result from the significantly higher levels of the β1-subunit (Figure 3), which has indeed been shown to increase Nav1.8 mediated currents in heterologous systems.[36] We further observed that TTX-R currents in CHRNA3+ neurons had more hyperpolarized half-activation voltages than TTX-R currents in both NPY2R+ and IB4−/caps+ nociceptors (Figure 2). The hyperpolarized half-activation voltage of TTX-R currents in CHRNA3+ is consistent with the high expression levels of the β4-subunit, which was shown to shift the V1/2 of Nav1.8 to hyperpolarized potentials by more than 15 mV,[36] which is comparable to the difference between CHRNA3+ and IB4−/caps+ positive observed here (Figure 2(d) and (e)). Cell-to-cell variability in the properties of TTX-R sodium currents has previously been reported,[43] and, moreover, it was shown that modulation of Nav channel function by β-subunits can depend on the cellular background. The β1-subunit, for example, was shown to increase the rate of open-channel inactivation and to shift steady-state inactivation of Nav1.3 in Xenopus oocytes but not in HE293 cells.[44] Sodium channel function is also altered by posttranslational modifications, e.g. by protein kinases, glycosylation and ubiquitylation.[20,21] Hence, one possible explanation for the cell type-specific effects of β-subunits is that different posttranslational modifications occur in different cell types, which could also explain why TTX-R currents in NPY2R+ neurons have half-activation voltages similar to TTX-R currents in IB4−/caps+ neurons (Figure 2(d) and (e)), despite expressing significantly higher levels of the β4 subunit (Figure 3). However, considering the variety of possible post-translational modification of VGSCs that had previously been described in other cells types,[20,21] we did not address this question here, as this would have been beyond the scope of this study.

NGF differentially modulates sodium currents in different subsets of TrkA+ nociceptors

The second major goal of our study was to test if NGF, which has previously been shown to modulate sodium currents in small diameter IB4− capsaicin-sensitive nociceptors, also modulates sodium currents in silent nociceptors (CHRNA3+) and Aδ-fibre nociceptors (NPY2R+). Specifically, we focused on two parameters – i.e. current density and voltage dependence of activation. Regarding current densities, our data show that NGF induces a significant increase in TTX-R currents in CHRNA3+ neurons (Figure 4(f)) but not in NPY2R+ neurons (Figure 5(f)), while it causes a small but significant decrease in TTX-S current density in NPY2R+ neurons (Figure 5(b)) without affecting TTX-S currents in CHRNA3+ nociceptors (Figure 4(b)). As mentioned earlier, changes in the current densities might result from changes in the expression levels of α-subunits and/or β-subunits, some of which are well known to enhance membrane trafficking of sodium channels.[20,41] However, unlike previously reported for small diameter neuron in long-term DRG cultures (7 div),[45] we did not observe an up-regulation of Nav1.8 in CHRNA3+ neurons (Figure 7(a)) after 24 h NGF treatment. Moreover, we neither detected a down-regulation of Nav1.7 in NPY2R+ neurons (Figure 7(b)) nor any significant changes in β-subunit expression in these two nociceptor subpopulations, suggesting that the changes in current densities are probably induced by post-translational modifications rather than transcriptional changes. In fact, downstream effectors of NGF-TrkA signalling, including p38 MAP kinase and ceramide, have previously been shown to increase Nav1.8 currents in small diameter IB4−/caps+ neurons, in case of p38 by facilitating membrane trafficking.[22,23,46] We also observed the previously described NGF-induced increase in Nav1.8 amplitudes in small diameter IB4−/caps+ neurons, though in our hands, this effect was not statistically significant, which may be attributed to the rather small samples size of these cells in our study (Figure 6(e) and (f)). The reduction in TTX-S current density in NPY2R+ neurons was significant, but rather small (Figure 5(b)), and was not reflected in a downregulation of TTX-S sodium channel α- or β-subunits (Figure 7(b)). However, the cell capacitance of NPY2R+ neurons increased from 30.5 ± 1.7 pF to 46.2 ± 5.3 pF after NGF treatment, which provides an alternative explanation for this effect.

The most robust effects of NGF on sodium currents that we observed were the hyperpolarizing shifts of the half-activation voltages of TTX-R currents in NPY2R+ (Figure 5(g) and (h)) and IB4−/caps+ neurons (Figure 6(g) and (h)) and the hyperpolarizing shift in V1/2 of TTX-S currents in IB4−/caps+ neurons (Figure 6(c) and (d)). It is difficult to unequivocally attribute changes in TTX-S and TTX-R currents to changes in the functional properties of a single sodium channel, because these are composite currents generated from multiple channels.[47] This is particularly true for the TTX-S currents in CHRNA3+ and NPY2R+ neurons, which express both Nav1.6 and Nav1.7 at considerable levels (Figure 3). The TTX-R currents are, however, most likely predominantly generated by Nav1.8, which is expressed at much higher levels than Nav1.9 in all three studied nociceptor populations (Figure 3). An NGF-induced hyperpolarizing shift in the voltage dependence of TTX-R sodium current, which is thought to be mediated by second messenger ceramide-induced activation of PKCζ, has previously been reported for small diameter IB4−/caps+ neurons,[23,24] but it has never been tested if this effect also occurs in other TrkA+ nociceptors. We did not detect any significant changes in the expression of sodium channel α-subunits and β-subunits (Figure 7(b)), which can also change the voltage dependence of activation.[20,41] Hence, it is tempting to speculate that the NGF-induced hyperpolarizing shift in the V1/2 of TTX-R currents in NPY2R+ Aδ-nociceptors may be mediated by the same mechanisms that was previously described for IB4−/caps+ neurons.[23,24] Since NGF is known to activate ERK1/2 MAP kinases[1,48] and phosphorylation of Nav1.7 by ERK1/2 was shown to cause a hyperpolarizing shift of channel activation,[30] it has been hypothesized that NGF shifts the V1/2 of TTX-S currents to more negative potentials, but actually this hypothesis has never been directly tested. Here, we demonstrate that NGF indeed shifts the voltage dependence of the TTX-S current in small diameter IB4−/caps+ neurons as previously proposed but not in myelinated NPY2R+ Aδ-fibre nociceptors or CHRNA3+ silent nociceptors. The observation that NGF did not alter the voltage dependence of activation of TTX-S currents in NPY2R+ and CHRNA3+ neurons was quite unexpected considering that ERK1/2, which supposedly mediates this effect in IB4−/caps+ neurons, is also activated by NGF in CHRNA3+ neurons, as we have recently shown,[19] and probably also in NPY2R+ neurons. While we can only speculate about the possible reasons for the lack of TTX-S sodium current modulation in CHRNA3+ neurons, our results clearly demonstrate that NGF exerts different effects on different nociceptor subpopulations, even though these subpopulations appear to express the same downstream effector proteins.

Conclusions

Previous studies have shown that each of the three subpopulations of nociceptors examined here is important for different forms of pain. Thus, NPY2R+ Aδ-fibre are required for pinprick pain,[31] whereas IB4−/caps+ C-fibre nociceptors signal heat pain and detect painful pinch stimuli,[16,49] and CHRNA3+ silent nociceptors are the most abundant TrkA+ nociceptors innervating visceral organs and deep somatic tissues.[19] Hence, our observation that NGF exerts different effects on sodium currents in Aδ-fibre nociceptors and polymodal C-fibre nociceptors as compared to silent nociceptors indicates that visceral and cutaneous pain may be mediated by fundamentally different mechanisms and thus highlights the importance of developing different therapeutic strategies for different forms of pain. We have not examined which downstream effectors of TrkA mediate the different effects of NGF on sodium currents in CHRNA3+ and NPY2R+ neurons as this would have been beyond the scope of this study, but our work provides a solid basis for future studies that aim at addressing this important question.

Table 1.

Summary of patch-clamp recording parameters.

Cell population	Access resistance (mean ± SEM)	Compensation (%)	Maximum voltage error TTX-S (mV)	Maximum voltage error TTX-R (mV)
CHRNA3+ (ctl)	3.1 ± 0.37	80.1 ± 2.4	6.7 ± 1.6	3.3 ± 0.8
CHRNA3+ (NGF)	2.9 ± 0.39	78.9 ± 2.1	9.6 ± 1.7	5.3 ± 1.6
NPY2R+ (ctl)	2.9 ± 0.22	87.8 ± 0.7	8.8 ± 1.4	2.6 ± 0.3
NPY2R+ (NGF)	2.5 ± 0.27	87.9 ± 0.5	7.1 ± 1.2	4.0 ± 0.7
IB4−/caps+ (ctl)	4.2 ± 1.21	76.2 ± 0.9	4.7 ± 1.3	2.2 ± 0.3
IB4−/caps+ (NGF)	3.9 ± 0.82	81.4 ± 3.1	4.1 ± 1.6	1.6 ± 0.4

NGF: nerve growth factor; IB4: Isolectin B4; NPY2R: neuropeptide Y receptor type 2; CTL: control.