Phenotypic distinctions between neural crest and placodal derived vagal C-fibres in mouse lungs


Corresponding author B. J. Undem: Johns Hopkins Allergy and Asthma Center, 5501 Hopkins Bayview Circle, Baltimore, MD 21224, USA.  Email:


Two major types of nociceptors have been described in dorsal root ganglia (DRGs). In comparison, little is known about the vagal nociceptor subtypes. The vagus nerves provide much of the capsaicin-sensitive nociceptive innervation to visceral tissues, and are likely to contribute to the overall pathophysiology of visceral inflammatory diseases. The cell bodies of these afferent nerves are located in the vagal sensory ganglia referred to as nodose and jugular ganglia. Neurons of the nodose ganglion are derived from the epibranchial placodes, whereas jugular ganglion neurons are derived from the neural crest. In the adult mouse, however, there is often only a single ganglionic structure situated alone in the vagus nerve. By employing Wnt1Cre/R26R mice, which express β-galactosidase only in neural crest derived neurons, we found that this single vagal sensory ganglion is a fused ganglion consisting of both neural crest neurons in the rostral portion and non-neural crest (nodose) neurons in the more central and caudal portions of the structure. Based on their activation and gene expression profiles, we identified two major vagal capsaicin-sensitive nociceptor phenotypes, which innervated a defined target, namely the lung in adult mice. One subtype is non-peptidergic, placodal in origin, expresses P2X2 and P2X3 receptors, responds to α,β-methylene ATP, and expresses TRKB, GFRα1 and RET. The other phenotype is derived from the cranial neural crest and does not express P2X2 receptors and fails to respond to α,β-methylene ATP. This population can be further subdivided into two phenotypes, a peptidergic TRKA+ and GFRα3+ subpopulation, and a non-peptidergic TRKB+ and GFRα1+ subpopulation. Consistent with their similar embryonic origin, the TRPV1 expressing neurons in the rostral dorsal root ganglia were more similar to jugular than nodose vagal neurons. The data support the hypothesis that vagal nociceptors innervating visceral tissues comprise at least two major subtypes. Due to distinctions in their gene expression profile, each type will respond to noxious or inflammatory conditions in their own unique manner.


dorsal root ganglion


GDNF family ligand receptor


nerve growth factor




ret proto-oncogene


tropomyosin-related kinase


Nociceptors derived from dorsal root ganglia (DRGs) are broadly subcategorized into two general subtypes, referred to as peptidergic and non-peptidergic based on expression of tachykinins and calcitonin gene-related peptide (CGRP) in the former type (Woolf & Ma, 2007). Isolectin B4 has proven useful in defining these nociceptive subtypes in that it stains somewhat selectively the non-peptidergic phenotype. These two subtypes innervate different regions of the CNS, and thus clearly subserve distinct sensations and functions in the organism (Molliver et al. 1995; Woolf & Ma, 2007).

At a more molecular level, recent studies are refining these two major nociceptive subtypes in dorsal root ganglia by evaluating the expression of transcription factors such as runt-related transcription factor 1 (RUNX1) and receptors for neurotrophic factors. The data from studies with mice support the hypothesis that RUNX1 is expressed embryologically in virtually all nociceptors (but seldom in non-nociceptors) (Chen et al. 2006; Kramer et al. 2006). Perinatally, however, RUNX1 expression is extinguished in a subset of nociceptors (Chen et al. 2006) and this coordinates a series of events that define this RUNX1 negative phenotype as peptidergic nociceptors. The RUNX1 expressing nociceptors not only fail to express neuropeptides but also differ from the peptidergic subtype in that they express ret proto-oncogene (RET), the key transmembrane signalling component of the glial cell derived neurotrophic factor (GDNF) family ligand receptors (GFRs). By contrast, rather than depending on RET/GFRs, the small peptidergic nociceptors in the DRG express tropomyosin-related kinase (TRK)A and depend on nerve growth factor (NGF) for survival (Yoshikawa et al. 2007).

There is little information on nociceptor subtypes in the vagal afferent system. Vagal afferent C-fibres innervating the respiratory tract are generally quiet during healthy respiration, but are activated by noxious stimuli and inflammatory mediators (Taylor-Clark & Undem, 2006). Inasmuch as they serve to provide the organ with a ‘sense of its own potential injury’, they fit Sherrington's definition of a nociceptor. These nociceptors do not typically evoke pain sensations; rather their activation can lead to coughing, sneezing, dyspnoeic sensations, as well as reflex secretion and bronchoconstriction (Coleridge & Coleridge, 1977; Coleridge et al. 1993). In most mammals studied thus far, C-fibres are categorically activated by capsaicin. Thus, transient receptor potential vanilloid 1 (TRPV1) expression is a useful molecular marker of vagal respiratory nociceptors.

An immediate distinction between vagal and spinal nociceptors in larger mammals, including guinea pigs, is that vagal neurons are situated in only two ganglia, termed nodose and jugular (supranodose). The jugular ganglion has a neural crest origin similar to the DRG, whereas the neurons comprising the nodose ganglion are derived from the epibranchial placodes (Baker & Schlosser, 2005). Based on electrophysiological, histological, and gene expression analysis we have previously demonstrated that the vagal nociceptors innervating the respiratory tract and oesophagus of guinea pigs can be broadly subdivided into two subtypes (Kwong et al. 2008). Similar to spinal nociceptors one subtype is peptidergic and the other largely non-peptidergic (Undem et al. 2004). The peptidergic neurons are derived from the jugular (neural crest) ganglia, whereas the non-peptidergic neurons were derived from the placodal ganglion.

The extent to which knowledge about vagal nociceptor subtypes can be inferred from information regarding subtypes in the mouse DRG is unknown as there have been no studies on vagal nociceptor subtypes in the mouse. In the present study we addressed the hypothesis that there are two general nociceptor phenotypes in the mouse vagus, and these subtypes can be differentiated based on expression of the P2X2 receptor, certain neurotrophic factor receptors and the sensory neuropeptide, substance P. The mouse offers the advantage of defining bona fide neural crest lineage neurons by using Wnt1Cre/R26R mice. We therefore also directly address the hypothesis that whether the neurons are placodal or neural crest in origin largely predicts the nociceptive phenotype in the adult animal. To keep other variables at a minimum, we focused on nociceptor subtypes innervating a single defined peripheral compartment, namely the mouse lung.


Animal experiments

All experiments were performed with approval from the Johns Hopkins Animal Use and Care Committee.

Retrograde labelling and cell dissociation

Bronchopulmonary afferent neurons of C57/BL6 mice (male, 6–8 weeks) were retrogradely labelled using DiI (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate; DiC18(3); Molecular Probes/Invitrogen, Eugene, OR, USA) solution (0.1%, 50 μl; dissolved in 10% DMSO and 90% normal saline) or aminostilbamidine methanesulfonate (Molecular Probes/Invitrogen; 0.05%, 50 μl, dissolved in normal saline). In comparison to the lipophilic DiI (absorption maximum: 549 ± 3 nm), the polar aminostilbamidine (absorption maximum: 386 ± 5 nm) was compatible with the detergent-containing washing buffers used for immunohistochemistry. For all other experiments (single-cell RT-PCR, calcium imaging), we used DiI as a neuronal tracer.

Under anaesthesia (2 mg ketamine and 0.2 mg xylazine i.p. per mouse), mice were orotracheally intubated, and DiI or aminostilbamidine was instilled into the tracheal lumen 5–9 days before an experiment.

After the animals were killed by CO2 asphyxiation, the jugular/nodose ganglia or DRGs (C1, T1–6) were dissected and cleared of adhering connective tissue. Isolated ganglia were incubated in an enzyme buffer (2 mg ml−1 collagenase type 1A and 2 mg ml−1 dispase II in Ca2+-, Mg2+-free Hanks’ balanced salt solution) for 30 min at 37°C. Neurons were dissociated by trituration with three glass Pasteur pipettes of decreasing tip pore size, then washed by centrifugation (three times at 1000 g for 2 min) and suspended in L-15 medium containing 10% fetal bovine serum (FBS). The cell suspension was transferred onto poly-d-lysine/laminin-coated coverslips. After the suspended neurons had adhered to the coverslips for 2 h, the neuron-attached coverslips were flooded with the L-15 medium (10% FBS) and used within 8 h.

Single-cell RT-PCR

First strand cDNA was synthesized from single lung-labelled jugular/nodose cells by using the SuperScript III CellsDirect cDNA Synthesis System (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's recommendations.

Cell picking Coverslips of retrogradelly labelled, dissociated neurons were constantly perfused by Locke's solution and identified by using fluorescence microscopy. Single cells were harvested into a glass-pipette (tip diameter 50–150 μm) pulled with a micropipette puller (Model P-87, Sutter Instruments Co., Novato, CA, USA) by applying negative pressure. The pipette tip was then broken in a PCR tube containing 1 μl RNAse Inhibitor (RNAseOUT, 2 U μl−1), immediately snap frozen and stored on dry ice. From one coverslip, one to four cells were collected. A sample of the bath solution from the vicinity of a labelled neuron was collected from each coverslip for no-template experiments (bath control).

RT-PCR Samples were defrosted, lysed (10 min at 75°C) and treated with DNAse I. Then, poly(dT) and random hexamer primers (Roche Applied Bioscience) were added. Twenty-five microlitres of the volume was reverse transcribed by adding SuperscriptIII RT for cDNA synthesis, whereas water was added to the remaining sample, which was used in the following as –RT control.

PCR Of each sample, 1.2 μl (cDNA, RNA control or bath control) was used for PCR amplification of mouse β-actin, TRPV1, P2X2, preprotachykinin A (PPT-A), RUNX1 and the diverse nerve factor receptors by the HotStar Taq Poymerase Kit (Qiagen) according to the manufacturer's recommendations in a final volume of 20 μl. After an initial activation step at 95°C for 15 min, cDNAs were amplified with custom-synthesized primers (Invitrogen) (Table 1) by 45 cycles (b-Actin and TRPV1) or 50 cycles (all other genes) of denaturation at 94°C for 30 s, annealing at 60°C for 30 s and extension at 72°C for 1 min followed by a final extension at 72°C for 10 min. Products were then visualized in ethidum bromide-stained 1.5% agarose gels.

Table 1.  Primer sequences for single cell RT-PCR
GenePrimerSequence (5′ to 3′)GenBankProduct length
NM_007393302 bp
NM_001001445285 bp
NM_153400241 bp
D 17584215/171/162/118 bp
NM_009050260 bp
NM_001080780357 bp
NM_001080780344 bp
NM_010279185 bp
NM_008115209 bp
NM_010280230 bp
NM_020014232 bp
AK081588216 bp
NM_008745151 bp
NM_008746213 bp
NM_009821184 bp
J01636433 bp
NM_182993230 bp
AF324864175 bp


The ganglia were fixed with 1% formalin in phosphate buffered saline (PBS, 4 h, 4°C), rinsed with PBS, cryoprotected overnight in PBS containing 18% sucrose, frozen and sectioned (10–12 μm). For GFRα3 staining, the sections were blocked with goat serum (10% in PBS with 1% BSA, 0.5% Tween 20, 2 h), and incubated overnight (4°C) with rabbit pAbs to GFRα3 (M-210, Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA; 7.5 μg ml−1), or normal rabbit IgG (R&D Systems, Minneapolis, MN, USA) as isotype control antibodies. For RET staining, sections were blocked with donkey serum and incubated overnight with goat pAbs to RET (AF482, R&D Systems; 2.5 μg ml−1) or normal goat IgG (R&D Systems) as isotype control antibodies.

Sections were then rinsed with PBS containing 0.3% Triton X-100 and 1% BSA and incubated with Alexa Fluor 647 goat anti-rabbit IgG (for GFRα3 staining), or Alexa Fluor Cy3 donkey anti-goat (for RET staining; both Molecular Probes/Invitrogen) for 2 h at room temperature. Washed slides were coverslipped with PBS (pH 8.6). The tissues were photographed (Q-Imaging Retiga EXi camera, BioVision, Exton, PA, USA), with an epifluorescence microscope (Olympus BX60, Olympus America, Inc., Melville, NY, USA) equipped with appropriate filter sets to allow separate visualization of Alexa 647, Alexa Cy3, and aminostilbamidine labelling using IPLab software (BioVision).

X-Gal staining in Wnt1Cre/R26R mice

The Wnt1 gene is uniformly, albeit briefly, expressed in early migratory neural crest cells at all axial levels. The expression is extinguished as the cells migrate away from the neural tube. The Wnt1Cre/R26R mice have a transgene expressing Cre recombinase under the Wnt1 promoter and enhancer. The second component is a reporter gene referred to as R26R that expresses β-galactosidase. In these mice the neural crest lineage neurons (but not the placodal lineage) can be observed by their expression of β-galactosidase; for more details see Jiang et al. (2000). The enzyme activity of β-galactosidase in vagal ganglia of Wnt1Cre/R26R mice was visualized by using the Marker Gene β-galactosidase staining kit according the manufacturer's recommendations (Marker Gene Technologies Inc., Eugene, OR, USA).

Ca2+ imaging

The intracellular [Ca2+]free measurements were performed in dissociated jugular/nodose neurons fromWnt1/R26R mice (n= 3). The coverslips were loaded with Fura-2 AM (8 μm) in L-15 medium containing 20% FBS and incubated for 40 min at 37°C. The coverslip was placed in a custom-built chamber (600 μl bath volume) that was superfused with Locke's solution (at 35°C) for 20 min before the experiment by an infusion pump (8 ml min−1). Changes in intracellular [Ca2+]free were measured by digital microscopy (Universal; Carl Zeiss, Inc., Thornwood, NY, USA) equipped with in-house equipment for ratiometric recording of single cells. A field of cells was monitored by sequential dual excitation, 352 and 380 nm, and the analysis of the image ratios used methods previously described (MacGlashan, 1989). The ratio images were acquired every 6 s. Superfused buffer was stopped 20 s prior to drug applications. In each experiment, the cells on the coverslip were exposed to α,β-methylene ATP (10 μm), capsaicin (1 μm) and KCl (75 mm) for 1 min each. The KCl was used as an indicator of voltage-sensitive cells. Between each stimulus, the cells were washed with fresh buffer for at least 3 min for the cells to recover prior to the addition of the second stimulus. After the last stimulation, cells were fixed in glutaraldehyde and were stained in situ with the Marker Gene β-galactosidase staining kit as described above. Each set of images for the Ca2+ measurements also included a brightfield image of the field of cells under study before and after the X-Gal staining. Neural crest derived cells were identified based on their dark appearance. Only cells that had an average diameter (long and short axis) of over 15 μm were analysed. Those cells that failed to respond to capsaicin with a rapid rise in Ca2+ were considered non-nociceptive and thus were not considered for statistical analysis.

Extracellular recording

The animals were killed by CO2 asphyxiation followed by exsanguination. The innervated isolated trachea-lung preparation was prepared as previously described (Kollarik et al. 2003). Briefly, the airways and lungs with their intact right-side extrinsic innervation (vagus nerve including jugular/nodose ganglia) were taken and placed in a dissecting dish containing Krebs bicarbonate buffer solution composed of (mm): 118 NaCl, 5.4 KCl, 1.0 NaH2PO4, 1.2 MgSO4, 1.9 CaCl2, 25.0 NaHCO3 and 11.1 dextrose, and equilibrated with 95% O2 and 5% CO2 (pH 7.2–7.4). Connective tissue was trimmed away leaving the larynx, trachea, main bronchi and lungs with their intact nerves. The airways were then pinned to the larger compartment of a custom-built two-compartment recording chamber which was lined with silicone elastomer (Sylgard 184 Silicone Elastomer Kit, Dow Corning, Midland, MI, USA). The left jugular/nodose ganglion was gently pulled into the adjacent compartment of the chamber through a small hole and pinned. Both compartments were separately superfused with Krebs solution, which was warmed by a warming jacket (39–42°C) to keep airway tissues and ganglia at 37°C. A sharp glass electrode was pulled by a Flaming Brown micropipette puller (P-87; Sutter Instrument Co.) and filled with 3 m NaCl solution. The electrode was gently inserted into the jugular/nodose ganglion so as to be placed near the cell bodies. The recorded action potentials were amplified (Microelectrode AC amplifier 1800; A-M Systems, Everett, WA, USA), filtered (0.3 kHz of low cut-off and 1 kHz of high cut-off), and monitored on an oscilloscope (TDS340; Tektronix, Beaverton, OR, USA) and a chart recorder (TA240; Gould, Valley View, OH, USA). The scaled output from the amplifier was captured and analysed by an Apple Macintosh computer using NerveOfIt software (Phocis, Baltimore, MD, USA). For measuring conduction velocity, an electrical stimulation (S44; Grass Instruments, Quincy, MA,USA) was applied on the core of the receptive field. The conduction velocity was calculated by dividing the distance along the nerve pathway by the time delay between the shock artifact and the action potential evoked by electrical stimulation. If a C-fibre (<1 m s−1) was found, the recording was started. One millilitre of vehicle, α,β-methylene ATP (10 μm) or capsaicin (1 μm) was intratracheally applied for 10 s. We have found that this technique does not record action potentials in ‘through fibres’. For example, positioning the electrode away from cell bodies on the vagus itself fails to record action potential stimulated by receptive fields, or by electrical nerve stimulation.

Statistical analysis

There was little (<1 Hz) or no background activity in the extracellular recordings. In all experiments, a single unit was recorded. The action potential discharge evoked by vehicle, α,β-methylene ATP and capsaicin stimulation was quantified off-line and segregated into consecutive 1 s bins. The response was considered to be terminated when the number of spikes in the bins declined to <2 × baseline. The total number of action potentials recorded following vehicle, α,β-methylene ATP and capsaicin application was counted. The peak frequency evoked by a stimulus was quantified as the maximum number of action potentials that occurred within any 1 s bin.

The intracellular [Ca2+]free of vagal mouse neurons was calculated as previously described using the Tsien calculation (Taylor-Clark et al. 2008). If a cell lacked a robust response to capsaicin (1 μm) or KCl (75 mm), or had an averaged diameter (long and short axis) of less than 15 μm, it was not included in the analysis. A cell was considered as α,β-methylene, capsaicin or KCl positive if the drug-induced peak increase was greater than two standard deviations above the mean baseline intracellular [Ca2+]free.

Student's paired or unpaired t test or ANOVA was used when appropriate. P < 0.05 was considered statistically significant. All data are expressed as means ±s.e.m.

Drug preparations and applications

Capsaicin was dissolved in ethanol, and α,β-methylene ATP and KCl were dissolved in H2O, respectively, and diluted with the appropriate buffer to the final concentrations. The final ethanol concentration was 0.1% for calcium measurements and 0.01% for extracellular recordings. Capsaicin was purchased from Sigma-Aldrich (St Louis, MO, USA). Fura-2 AM was purchased from Molecular Probes/Invitrogen. L-15 and Hanks’ balanced salt solution (HBSS) were purchased from Gibco/Invitrogen (Carlsbad, CA, USA).


Anatomy of the mouse jugular nodose complex

Vagal sensory neurons are situated in the nodose (placode derived) and jugular ganglion (neural crest derived). These disparate ganglia are very evident in guinea pigs and larger mammals, but there has been some issue of their location in mice. We carried out dissections of the vagus and accompanying ganglia in adult mice, and found that indeed, compared to the guinea pig, there is often no discrete nodose and jugular ganglion on the vagus nerve of the mouse, and only occasionally a minimal ganglionic structure is observed rostral to the nodose ganglion (Fig. 1).

Figure 1.

Dissection of the vagal ganglia in guinea pigs and mice
The right vagus and accompanying ganglia were dissected in guinea pigs (left photograph) and mice (right photograph). Compared to the guinea pig in which the nodose and the jugular ganglion are clearly separated, mice possess a fused jugular nodose complex.

The nodose ganglion in the mouse often appears somewhat elongated, so we hypothesized it may represent a fused nodose–jugular complex with the rostral pole more jugular-like, and the body and caudal pole, nodose-like. To address this we obtained mice in which the premigratory neural crest neurons expressed β-galactosidase encoded by the E.coli lacZ gene (Wnt1Cre/R26R mice) (Jiang et al. 2000; Makita et al. 2008). When the vagal ganglia of these mice (n= 16) were evaluated it was clear that neural crest derived cells formed the more rostral aspect of the ganglion whereas the caudal pole of the ganglion comprised non-neural crest (placodal) neurons. In several ganglia, a significant number of neural crest neurons were observed down to the central body of the ganglion (Fig. 2). In about 1 in 3 mice a jugular ganglion, entirely consisting of neural crest derived cells, can be recognized as the top half of an hourglass like structure on the vagus.

Figure 2.

X-Gal staining of the jugular nodose complex in mice
In order to evaluate the localization of neural crest derived and placodal neurons within the jugular–nodose complex, X-Gal staining was performed in Wnt1Cre/R26R mice. Neural crest derived cells formed the rostral pole of the ganglion whereas the body and caudal pole of the ganglion comprised the placodal cells (A–D). A jugular ganglion, entirely consisting of neural crest derived cells, can be recognized as the top half of an hourglass like structure on the vagus in some ganglia (C and D).

Calcium imaging

In a previous study with guinea pigs we noted that the nodose C-fibre terminals in the respiratory tract responded strongly with action potential discharge to the P2X-2/3 selective agonist α,β-methylene ATP, whereas the jugular neurons were categorically unresponsive. We also noted at the level of the cell body that the nodose neurons responded to ATP or α,β-methylene ATP with a large persistent inward current that was explained by the presence of both P2X2 and P2X3 receptors. Jugular neurons expressed P2X3 but not P2X2 receptors and responded to ATP with a very transient inward current, which is consistent with the current signature of P2X3 homomeric receptors (Kwong et al. 2008).

We evaluated the α,β-methylene ATP responsiveness in mouse vagal ganglion neurons. Using the Wnt1Cre/R26R mice, we dissociated the vagal ganglion complex (8 ganglia from 4 mice) and evaluated the calcium rise in β-galactosidase+ neural crest neurons and in β-galactosidase placodal neurons. About 58% (47/81) β-galactosidase+ cells and 82% (115/140) β-galactosidase cells responded to capsaicin. We focused on capsaicin-sensitive neurons as these are presumably C-fibre neurons. We found that the α,β-methylene ATP was selective for activation of the placodal phenotype (Fig. 3). These data support the hypothesis that, as in guinea pigs, mouse placodal C-fibres can be distinguished from neural crest derived cells based on purinergic responsiveness.

Figure 3.

α,β-mATP increases intracellular [Ca2+]free selectively in placodal C-fibres
The effect of α,β-mATP (10 μm) on intracellular [Ca2+]free was investigated in dissociated vagal ganglion in capsaicin (1 μm)-sensitive C-fibres from Wnt1Cre/R26R mice. A calcium rise in response α,β-methylene ATP was found in the vast majority of placodal (X-Gal) cells (black squares; 97/116 C-fibres). Only 1/47 neural crest derived cells (X-Gal+) neurons were activated by α,β-methylene ATP (grey circles). Data are shown as means ±s.e.m.

Extracellular recordings

To evaluate nociceptive subtypes within a single defined peripheral compartment we set up the isolated vagally innervated mouse lung preparation, and evaluated the function of single vagal afferent nerves using extracellular recording (Fig. 4). We evaluated 116 vagal afferent fibres with a receptive field in the lung (one fibre per mouse). Among these 116 neurons 89 (77%) were capsaicin sensitive and conducted action potentials at <0.6 m s−1. The remainders were capsaicin insensitive, and conducted action potentials 0.7–5 m s−1. From this point, we focused our attention on the capsaicin-sensitive (presumably TRPV1 expressing) neurons.

Figure 4.

α,β-mATP stimulation of vagal sensory nerves elicits action potential discharge in a subpopulation of capsaicin-sensitive bronchopulmonary C-fibres
The action potential discharge in response to α,β-mATP and capsaicin stimulation was evaluated in bronchopulmonary vagal afferent nerves by extracellular recording. 89/116 neurons (77%) were capsaicin sensitive and conducted action potentials at <0.6 m s−1. We noted that among the capsaicin-sensitive population (n= 89), 71% responded with action potential discharge to α,β-methylene ATP; these are likely to be the placodal C-fibres. The remainder of capsaicin-sensitive neurons responded strongly to bradykinin and capsaicin, but were unresponsive to α,β-methylene ATP; these are the neural crest derived C-fibres

With a random search protocol, we noted that among the capsaicin-sensitive population (n= 89) 71% responded with action potential discharge to α,β-methylene ATP; based on our previous studies (Kwong et al. 2008), and the studies described above, these are likely to be the placodal C-fibres. The remainder of capsaicin-sensitive neurons responded strongly to bradykinin and capsaicin, but were unresponsive to α,β-methylene ATP; these are the neural crest derived C-fibres. To further test this assumption, in five experiments the recording electrode was limited to the very rostral aspect of the ganglion that we noted comprised neural crest neurons (see Fig. 2). All 5/5 fibres that were recorded from neurons situated in this part of the ganglion, were capsaicin and bradykininsensitive, but α,β-methylene ATP insensitive (i.e. jugular neuronal crest fibres).

Gene expression in lung-labelled neurons

We evaluated the neurons in vagal ganglion retrogradely labelled by dye injection into the lungs. The labelled neurons were found predominantly in the body of the ganglion (nodose compartment) but also in the rostral aspect of the ganglion (jugular compartment). The ganglion was dissociated and the lung-labelled neurons were evaluated at a single neuron level for expression of various genes. Among all lung-labelled neurons (n= 208 from 18 mice), we noted that 70.2% expressed TRPV1. This is roughly consistent with our studies of nerve ending responsiveness to capsaicin (see Fig. 4). We focused our attention on gene expression in this population of TRPV1 expressing neurons.

Based on the results shown in Figs 3 and 4 we hypothesized that P2X receptor expression might serve as a convenient molecular marker for jugular vs. nodose neurons, or neural crest derived vs. placodal neurons, respectively, as has been previously shown in guinea pigs (Kwong et al. 2008). In order to support the hypothesis that absence of P2X2 expression is a predictive marker for neural crest derived C-fibres, we evaluated P2X2 and lacZ gene expression in 66 TRPV1 positive jugular/nodose cells of Wnt1Cre/R26R transgenic mice (n= 2). As expected, >90% of neurons were P2X2+ (61/66) and a small number of cells was P2X2 (5/66). However, all 5 TRPV1+ neurons lacking P2X2 expression were found to express lacZ gene, indicating that the absence of P2X2 expression can be used to define neural crest derived C-fibres in dissociated jugular nodose cells by single cell RT-PCR (Fig. 5).

Figure 5.

Neural crest derived C-fibres lack P2X2 mRNA expression
Single cell RT-PCR was performed in vagal C-fibres from Wnt1Cre/R26R mice. Five out of 66 TRPV1+ cells did not express P2X2. All 5 cells were lacZ+ and therefore considered to be of neural crest origin. Further analysis revealed that this cell population consists of two different phenotypes as revealed by TRK receptor mRNA expression.

Based on these results we deduced that the TRPV1 expressing population could be subdivided into those expressing P2X2 (presumably placodal) vs. those that failed to express P2X2 (presumably neural crest) neurons (Fig. 6). Again, consistent with our functional data, the majority of lung specific TRPV1 expressing neurons were placodal in that they expressed P2X2 (87%, 101/116 TRPV1+).

Figure 6.

Characterization of distinct bronchopulmonary C-fibre subsets by single cell RT-PCR
Since the lack of P2X2 receptors was specific for neural crest derived neurons, lung-labelled TRPV1+ vagal neurons were divided in two groups according to their P2X2 expression. In each individual cell, the expression of PPT-A gene (preprotachykinin A), glial cell derived neurotrophic family ligand receptors (GFLRs) and neurotrophin receptors (NTRs) were analysed. Up to 82 placodal cells and 15 neural crest derived cells were analysed per individual target. The figure shows representative results from 15 individual placodal cells and all neural crest neurons. In order to facilitate the understanding of the figure, the neural crest cells were sorted according to their TRKA expression. P, positive control (cDNA from whole JNG); N, –RT negative control; B, bath control; *not done

Gene expression in P2X2-negative ‘jugular’ neurons Previous studies have indicated that spinal (neural crest) nociceptors can be subdivided into a peptidergic TRKA+ group and a non-peptidergic TRKA subpopulation (Woolf & Ma, 2007). The occurrence of the two different neural crest C-fibre subpopulations was confirmed in vagal TRPV1+lacZ+P2X2 cells. Three of five cells obtained from two mice expressed TRKA but not TRKB, and 2/5 cells were TRKB+ but did not express TRKA (Fig. 5).

Next, we evaluated gene expression in target-specific cranial neural crest cells, namely the lung-specific nociceptors, which were stained by intratracheal application of DiI. As mentioned above, the neural crest derived cells represent a relatively small population of capsaicin-sensitive (or TRPV1 expressing) neurons innervating the lungs. Nevertheless we were able to evaluate gene expression in 15 TRPV1+P2X2 neurons obtained from eight animals (Fig. 6), and noted that both TRKA+ (6/15) and TRKA neurons were present (9/15). TRKB was expressed in 5 of the 9 TRKA neurons leaving only four neurons that failed to express either TRKA or TRKB. None of the neurons expressed TRKC (data not shown). We next evaluated whether expression of preprotachykinin-A (PPT-A) mRNA correlated with TRKA (as it does in DRG neurons). Indeed 5/6 TRKA+ cell were PPT-A+, whereas only 3 of 8 TRKA cells expressed PPT-A (Fig. 6).

All 15 of the lung-specific neural crest neurons expressed RET, but the expression of GDNF family receptor subtypes was selective for TRKA+vs. TRKA neurons. The GDNF receptor, GFRα1, was selectively expressed in the non-peptidergic TRKA neurons (8/9 neurons were GFRα1+), and only rarely expressed in TRKA+ neurons (1/6). By contrast, GFRα3, the receptor for artemin, was expressed selectively in the peptidergic TRKA+ subtype. All of the TRKA+ neurons expressed GFRα3, whereas only 1 of 9 TRKA neurons expressed this receptor. GFRα2 (2 of 15) and GFRα4 (2of 15) were rarely expressed in these neural crest nociceptive neurons (data not shown).

Gene expression in placodal P2X2-positive ‘nodose’ neurons As mentioned the majority of vagal sensory neurons innervating the mouse lungs are placodal in nature, and thus we were able to evaluate gene expression in a larger number (up to n= 82) of lung-specific TRPV1+ placodal neurons than neural crest neurons obtained from 18 mice.

By contrast to the neural crest neurons, nearly all of the TRPV1+ placodal ‘nodose’ neurons expressed TRKB (72/82, 87.2%), whereas TRKA was relatively rarely expressed (17/82, 20.7%) (Fig. 6). Comparable with the neural crest neurons, TRKC was rarely expressed in the TRPV1+ placodal neurons (7/73, 9.6%) (data not shown).

PPT-A was expressed in <10% (5/51) of lung-specific C-fibre placodal neurons. These data show that in the adult mouse the majority of tachykinergic C-fibres innervating the lungs are by and large neural crest and not placodal in nature, consistent with reports in other species (Katz & Karten, 1980; Lundberg et al. 1983; Springall et al. 1987; Kummer et al. 1992; Undem et al. 2004).

As with the neural crest population of neurons, we found that that RET mRNA was uniformly expressed by TRPV1 expressing lung labelled placodal neurons (80/80 neurons, 100%). Single cell RT-PCR with primers that were able to discriminate between RET9 and RET51 isoforms showed that most of the cells co-expressed both variants (data not shown; RET9 was expressed in 20/21 cells (95%); RET51 was expressed in 18/21 cells (85.7%)). Thus, RET is expressed indiscriminately in vagal sensory neurons innervating adult mouse lungs.

This conclusion based on gene expression was supported by immunohistochemical data revealing that indeed virtually all vagal ganglion neurons were RET+ (Fig. 7).

Figure 7.

RET expression in mouse jugular nodose complex
RET expression was tested in lung-labelled jugular nodose ganglia sections of male C57BL/6 mice. Almost all neurons expressed RET. A, the left section was stained with a RET mAbs (red fluorescence) and the right section was incubated with the corresponding isotype control. Vagal nerve cells projecting to the airways were retrogradely stained by intratracheal application of aminostilbamidine (green fluorescent cells). B, higher magnification reveals the presence of RET immunoreactivity in lung-labelled neurons.

The GDNF receptor GFRα1 was expressed in nearly all placodal nociceptors innervating the lungs (53/60, 88%). The artemin receptor GFRα3 was not expressed in placodal C-fibre neurons (0/51). Thus, this receptor is expressed selectively in neural crest C-fibre neurons (Fig. 6). To support the conclusion that GFRα3 is selectively expressed only in vagal ganglion neurons of neural crest origin, we evaluated GFRα3 immunohistologically in ganglia isolated from the Wnt1Cre/R266R mice. As predicted from our gene expression studies, neurons stained with anti-GFRα3 antibodies were limited to the small minority of neural crest neurons with the majority of lung labelled neurons being placodal neurons that are negative for GFRα3 (Fig. 8). GFRα2 mRNA was expressed in 15/56 (27%) and GFRα4 mRNA was expressed in 28/56 (50%) of the neurons (data not shown).

Figure 8.

GFRα3 expression in lung-labelled neurons
GFRα3 expression was evaluated by immunohostistochemistry in lung-labelled vagal ganglia of Wnt1Cre/R26R mice. GFRα3 expression was selectively expressed in neural crest derived (X-Gal+) neurons. B, arrowheads indicate lung-labelled GFRα3+ neural crest derived neurons.

RUNX1 expression In DRG neurons, RUNX1 expression, as with RET expression, was found to be selectively expressed in the non-peptidergic subtype of nociceptor, at least in early post-natal life (Chen et al. 2006; Woolf & Ma, 2007; Yoshikawa et al. 2007). We addressed the hypothesis that the placodal subtype of vagal nociceptor would preferentially express RUNX1. RUNX1 mRNA was expressed in the majority of TRPV1 expressing lung-labelled neurons of both placodal (24/37) and neural crest (9/15) origin. RUNX1 expression was not selectively associated with other genes we evaluated (PPT-A, TRK receptors, or GFLR receptors) (Fig. 6).

Glutamate tansporter expression Different sensory nerve phenotypes have been shown to express different glutamate transporters (Oliveira et al. 2003; Morris et al. 2005; Brumovsky et al. 2007; Mazzone & McGovern, 2008). We evaluated 28 lung specific vagal TRPV1+ neurons for expression of VGluT1 and VGluT2. Both VGluT1 and VGluT2 were expressed in the total vagal ganglion. However, none of the 28 TRPV1+ neurons expressed VGluT1 (data not shown), whereas each of the 28 neurons expressed VGluT2 mRNA. The 28 neurons were of both placodal and neural crest origin (Fig. 6). VGluT1 was expressed in 2 of 7 lung-specific neurons that were TRPV1-negative (data not shown).

Dorsal root ganglion neurons Dinh et al. (2004) noted that a portion of DRG neurons provide TRPV1+ axons to the mouse lungs. These DRG neurons arise from the same region of postotic hindbrain that also gives rise to vagal jugular neurons (Baker & Schlosser, 2005). Therefore, we predicted that the gene expression in the TRPV1+ DRG neurons should be similar to the jugular neurons.

In comparison to the percentage of TRPV1+ cells in vagal ganglia (70.2% from 208 neurons), TRPV1 expression was lower in T1–T6 DRG (51.1% from 45 neurons obtained from 6 mice) and in C1 DRG (25.5% from 98 neurons obtained from 5 mice). Since no obvious differences in the gene expression profile from TRPV1+ neurons could be observed in T1–T6 (n= 25) and C1 DRG (n= 23), the data were pooled.

Consistent with our prediction, the vast majority (87.5%) of these neurons did not express P2X2. Also consistent with what we observed in the jugular ganglion neurons, the TRPV1+ neurons expressed TRKA (35/48 neurons). The vast majority (92%) of these neurons also expressed RET, and at least one GFR receptor (Table 2).

Table 2.  Gene expression in TRPV1+ neurons in neurons isolated from the cervical and T1-T6 DRGS of the adult mouse
GeneNeuronsPercentage of positve neurons
  1. Neurons were isolated from C1 and from T1–T6 DRG of adult mice. The neurons were first screened for expression of TRPV1. A total of 48 TRPV1 expressing neurons were identified obtained from 8 mice.

TRKA and RET35/4873%


This study demonstrates that in the adult mouse, capsaicin-sensitive vagal C-fibres innervating a single organ (lungs) comprise two major subtypes, based on embryological origin. One subtype is neural crest in origin, the other placodal. The data also support our preliminary hypothesis that the capsaicin-sensitive neural crest vagal C-fibres can be further subdivided into two phenotypes based on expression of TRKA receptors. The main three nociceptive C-fibre subtypes differ in activation profile and expression of neuropeptide and growth factor receptor genes. They therefore likely subserve disparate sensations and reflexes, and will respond to noxious or inflammatory conditions in their own unique manner. The recognition of distinct vagal nociceptive phenotypes is essential in the development of our understanding of the role these nerves play in respiratory physiology and in inflammatory airway diseases such as asthma, chronic obstructive pulmonary disease, and chronic bronchitis.

In guinea pigs and larger mammals, the cell bodies of vagal sensory neurons form two distinct ganglia referred to as the nodose and jugular (or supranodose) ganglion. The neural crest derived cells form a jugular ganglion, which is clearly separated from the nodose ganglion that comprises neurons from the epibranchial placode (Yntema, 1942; D’Amico-Martel & Noden, 1983). We previously noted in guinea pig airways and oesophagus that P2X2 expression and responsiveness to α,β-mATP is a useful molecular marker of nodose (placodal) vagal sensory neurons (Kwong et al. 2008). In the mouse, the vagal sensory neurons most often formed a single elongated ganglion (see Fig. 1). In our studies of action potential discharge of capsaicin-sensitive C-fibres in the lungs, recorded with an extracellular electrode positioned in this vagal sensory ganglion, we noted that C-fibres could be subdivided into ATP sensitive (nodose-like) and ATP insensitive (jugular-like) cells. This predicts that the vagal ‘nodose’ ganglion in the mouse is actually a combination of neural crest and placodal neurons.

The hypothesis that the single ganglion structure in the mouse vagus is a combination of nodose (placodal) and jugular (neural crest) neurons is directly supported by our data using Wnt1Cre/R26R mice. These mice express β-galactosidase only in premigratory neural crest derived neurons. The neural crest neurons were consistently found within the single vagal ganglion structure; moreover, consistent with the predictions from the nerve endings studies, only the non-neural crest neurons (placodal neurons) responded to α,β-mATP with large increases in intracellular calcium. In a minority of mice we noted that the nodose and jugular neurons were separated with the later forming a very small swelling distinct from the nodose ganglion. Most often, however, the neural crest neurons formed the rostral aspect of a single ganglion structure with a significant minority of these neurons as far caudally as the ganglion centre. Thus, one must be cautious when evaluating neurons from these ganglia and assuming that they are all ‘nodose’ neurons. Consistent with the hypothesis that the ATP-unresponsive C-fibres terminating in the mouse lungs are jugular in nature was the observation that when the electrode was positioned in the rostral aspect of the vagal ganglion complex, the capsaicin-sensitive C-fibre under study categorically failed to respond with action potential discharge to α,β-mATP. In guinea pigs and rats, the number of C-fibre neurons innervating the respiratory tract is similar to the number of nodose neurons (Springall et al. 1987; Ricco et al. 1996; Yu et al. 2005). In the present study we found that in the mouse the nodose C-fibre neurons far outnumbered the jugular neurons. This may represent a species difference. It should be kept in mind that the neural crest C-fibres in the lung comprise not only jugular neurons, but also DRG neurons. Thus the neural crest derived C-fibres innervating the respiratory tract and oesophagus may actually outnumber the nodose C-fibres.

In guinea pigs and rats the sensory neurokinins in the lungs are preferentially localized to jugular (neural crest) ganglion C-fibres. Consistent with this, we found that the preprotachykinin gene, PPTA, was selectively expressed in the neural crest subpopulation of lung-labelled neurons. However, not all neural crest C-fibres were neurokinin-positive. The PPT-A gene was predominantly expressed in the TRKA+ subset of jugular neurons. This is consistent with findings in the somatosensory system where PPT-A is also expressed in the majority of TRKA+ nociceptors (Kashiba et al. 1996).

Vagal and spinal nociceptors are neurons that exhibit plasticity. They respond to local insults and inflammation with respect to both excitability changes and action potential discharge, but also via alterations in gene expression. For example at sites of airway inflammation induced by either allergen or viral exposure, there are phenotypic switches in the neuropeptide expression such that even capsaicin-insensitive placodal A-fibre neurons become peptidergic (Hunter et al. 2000; Carr et al. 2002; Chuaychoo et al. 2005; Dinh et al. 2005). Neurotrophic factor molecules produced at sites of inflammation are likely to be responsible for these types of phenotypic changes. It is therefore important to understand the nature of neurotrophic receptor expression in the subsets of adult vagal neurons. With respect to the neurotrophin receptors we noted that the placodal C-fibres are TRKB expressing neurons, whereas about 50% of the neural crest C-fibres were TRKA expressing neurons. Relatively few C-fibre placodal neurons expressed TRKC. This is basically consistent with pre- and perinatal studies of nodose and jugular neurons, and indicates that in the adult animals BDNF will be more apt to modulate the function of placodal neurons, whereas many neural crest neurons are more apt to be under the control of NGF.

The GDNF family ligands include GDNF, neurturin, artemin, and persephin. These growth factors activate GDNF family ligand receptors (GFRs) GFRα1, GFRα2, GFRα3 and GFRα4, respectively (Bespalov & Saarma, 2007). Our findings indicate that GDNF will most likely modulate the function of placodal C-fibre neurons in the lungs, as well as the TRKA-negative neural crest C-fibres, in that these neurons express GFRα1 receptor. The GFRα3 receptor agonist artemin, on the other hand, is unlikely to influence placodal C-fibres but may modulate the peptidergic TRKA-positive neural crest C-fibre population in the lungs. Although our studies focused mainly on lung-specific neurons, this idea may be relevant to other visceral organs. Nodose neurons labelled from the mouse pancreas do not express GFRα3, whereas this receptor is expressed in the majority of pancreas specific DRG neurons (neural crest neurons) (Fasanella et al. 2008).

RET functions as an important co-receptor for the GFRs. Both in our gene expression analysis and in our immunohistochemical analysis we found RET to be uniformally expressed in the adult vagal sensory neurons. This is different from observations in the perinatal DRG neurons, where RET was found to be selectively expressed in the non-peptidergic, non-TRKA subtype of nociceptors (Woolf & Ma, 2007). The expression of the transcription factor RUNX1 is also a marker of the non-peptidergic subtype of nociceptor in DRGs (Chen et al. 2006). We found, however, that RUNX1 was expressed in about 50% of the lung-specific vagal noceptors, but it was not apparently coordinated with RET, TRK, or PPT-A expression.

The vagal jugular ganglion neurons emigrate from the postotic hindbrain at the level of the first seven somites. This region also gives rise to the neuron in the more rostral DRG (Baker & Schlosser, 2005). If the nociceptor phenotype is explained largely by embryonic origin, we would predict that the capsaicin-sensitive neurons in the DRGs innervating the lungs would be similar to the jugular neurons. Although we did not investigate lung-specific DRG neurons, the TRPV1 expressing neurons in the rostral DRGs are similar to jugular neurons in their lack of P2X2 expression, and in their expression of TRKA along with RET and in some neurons GFRα3. As we observed in the vagal ganglia, nearly all the TRPV1-expressing neurons comprising the rostral DRGs in the adult mice, including those that express PPT-A, express both TRK receptors and RET/GFRα receptors. This is similar to adult human DRG where a substantial percentage of TRKA expressing neurons co-express RET (Josephson et al. 2001). Therefore the idea, based largely on elegant perinatal studies of lumbar DRG neurons, that a nociceptor can be categorically segregated into TRK expressing or RET/GFRα expressing subtypes should be cautiously extrapolated to DRG neurons of the adult animal in general.

Glutamate transporters have been shown to be differentially expressed in sensory neurons of different phenotypes (Oliveira et al. 2003; Morris et al. 2005; Brumovsky et al. 2007; Mazzone & McGovern, 2008). There is little information on differential VGluT expression in mouse vagal sensory ganglia. Generally, in the DRG, VGluT1 has been shown to be expressed in large diameter non-nociceptive neurons, whereas VGluT2 has been found to be expressed in smaller diameter, presumed nociceptive neurons. There is some controversy about whether VGluTs may be absent from most small diameter peptidergic neurons in the DRG. In guinea pigs and mice, Morris et al. (2005) concluded that most peptidergic DRG neurons lack VGluT expression. Others have noted that nearly all CGRP containing and IB4-binding mouse DRG neurons are immunoreactive for VGluT2 (Brumovsky et al. 2007). We show that VGluT2, but not VGluT1 mRNA was expressed in virtually all TRPV1-expressing lung-specific neurons irrespective of their placodal vs. neural crest origin, including neurons expressing PPT-A. VGluT1 was found to be expressed only in those neurons not expressing TRPV1 (presumably non-nociceptor neurons). Our findings are consistent with histological studies of guinea pig airway-specific vagal neurons that show that VGluT1 identifies a large diameter population of neurons, whereas VGluT2 identifies smaller diameter neurons (Mazzone & McGovern, 2008). Tong et al. (2001) likewise noted that the majority of nodose and DRG neurons labelled from the rat stomach have also been found to express VGluT2. Therefore, VGluT1 and VGluT2 expression may differentially be expressed in vagal nociceptive vs. non-nociceptive neurons, but they are not differentially expressed in subsets of TRPV1 expressing visceral nociceptors.

The presence of different C-fibre subpopulations innervating the lung was first hypothesized by the Coleridges. In dogs, the capsaicin-sensitive vagal C-fibres could be subdivided into ‘pulmonary’ and a ‘bronchial’ subtypes based on different latencies of responses to capsaicin injection to the pulmonary or the bronchial circulation, and different sensitivities to phenyl diguanide, bradykinin and ozone (Coleridge & Coleridge, 1977; Kaufman et al. 1980; Coleridge et al. 1993). In guinea pigs we had noted that the neural crest, but not placodal, C-fibres innervate large extrapulmonary airways which would be consistent with the Colerdidges’‘bronchial’ C-fibre subtype. The nodose C-fibres deeper in the lung tissue may comprise ‘pulmonary C-fibres’. It is tempting to speculate, therefore, that the respiratory C-fibres accessible to stimuli in the external environment (‘extero-receptors’ that terminate in the epithelium of the larger airways) are mainly neural crest in nature, whereas those that terminate deeper within the lung tissue (‘intero-receptors’) are more placodal in nature. In any event, the sensations and reflexes that are evoked upon noxious stimuli will likely depend on the nature of the C-fibre subtype activated. Our data revealing the selective expression of various neurotrophic factor receptors indicate that the type of phenotypic neuromodulation that occurs in inflammatory airway diseases will depend on the subtype and location of the C-fibre terminals.


Author contributions

C.N. and B.J.U. were responsible for the conception and design of the experiments, and the drafting of the article; C.N., T.E.T.C., A.M., F.R., R.N., W.B. and B.J.U. collected, analysed and interpreted the data. All authors approved the final version of the manuscript.


This work was funded by The National Institutes of Health (NIH) and the German Research Foundation (DFG NA 836/2-1). The authors thank Mrs Sonya Meeker and Ms Silke Wiegand for excellent technical assistance and D. D. Ginty and T. Makita for providing the Wnt1Cre/R26R mice.