P2X2 receptors differentiate placodal vs. neural crest C-fiber phenotypes innervating guinea pig lungs and esophagus

Abstract

The lungs and esophagus are innervated by sensory neurons with somata in the nodose, jugular, and dorsal root ganglion. These sensory ganglia are derived from embryonic placode (nodose) and neural crest tissues (jugular and dorsal root ganglia; DRG). We addressed the hypothesis that the neuron's embryonic origin (e.g., placode vs. neural crest) plays a greater role in determining particular aspects of its phenotype than the environment in which it innervates (e.g., lungs vs. esophagus). This hypothesis was tested using a combination of extracellular and patch-clamp electrophysiology and single-cell RT-PCR from guinea pig neurons. Nodose, but not jugular C-fibers innervating the lungs and esophagus, responded to α,β-methylene ATP with action potential discharge that was sensitive to the P2X3 (P2X2/3) selective receptor antagonist A-317491. The somata of lung- and esophagus-specific sensory fibers were identified using retrograde tracing with a fluorescent dye. Esophageal- and lung-traced neurons from placodal tissue (nodose neurons) responded similarly to α,β-methylene ATP (30 μM) with a large sustained inward current, whereas in neurons derived from neural crest tissue (jugular and DRG neurons), the same dose of α,β-methylene ATP resulted in only a transient rapidly inactivating current or no detectable current. It has been shown previously that only activation of P2X2/3 heteromeric receptors produce sustained currents, whereas homomeric P2X3 receptor activation produces a rapidly inactivating current. Consistent with this, single-cell RT-PCR analysis revealed that the nodose ganglion neurons innervating the lungs and esophagus expressed mRNA for P2X2 and P2X3 subunits, whereas the vast majority of jugular and dorsal root ganglia innervating these tissues expressed only P2X3 mRNA with little to no P2X2 mRNA expression. We conclude that the responsiveness of C-fibers innervating the lungs and esophagus to ATP and other purinergic agonists is determined more by their embryonic origin than by the environment of the tissue they ultimately innervate.

the majority of nerve fibers in the vagus are unmyelinated C-fibers (1, 14). Vagal C-fibers innervating the respiratory tract have activation profiles consistent with nociceptors. They are normally quiet during normal respiration, but are activated by potentially noxious stimuli including capsaicin, acid, tissue inflammation, and excessive stretch. Activation of these fibers does not evoke the sensation of pain. Rather, they elicit a panoply of cardiopulmonary reflexes including changes in the rate and depth of breathing, increases in airway smooth muscle tone, mucus secretion, changes in heart rate, and vascular resistance (9, 20).

In the somatosensory system, at least two distinct sensory nociceptive C-fiber subtypes have been described based mainly on neuropeptide content and binding to the isolectin B4 (IB4) (22), and, more recently, on expression of the transcription factor Runx1 (38). The IB4-postive neurons are relatively devoid of neuropeptides, whereas the IB4-negative subtype expresses tachykinins and calcitonin gene-related peptide. These two fiber subtypes display distinct functional characteristics in vitro (34). Moreover, they may also serve distinct physiological functions based on the fact that their central terminals synapse on distinct layers of the dorsal horn of the spinal cord (12, 22).

Much less is known about vagal C-fiber subtypes innervating visceral tissues. In the 1970s, the Coleridges (8) concluded that two different types of vagal C-fibers innervate the respiratory tract and that they can be distinguished based on the vascular accessibility of the nerve endings. One C-fiber type responded immediately to capsaicin only when injected into the right atrium (pulmonary circulation); the other responded immediately only when injected into the left atrium or directly into the bronchial artery (systemic circulation). The former type was referred to as “pulmonary C-fiber,” whereas the latter was referred to as “bronchial C-fiber.” The pulmonary C-fibers were considered to be analogous to the J receptors elegantly described in the 1950s by Paintal (26–29) and are thought to terminate near the pulmonary capillaries. The bronchial C-fibers were thought to reflect those that terminated within the wall of the conducting airways. Despite observations that these two C-fiber subtypes were differentially stimulated by certain inflammatory mediators, the idea that they represent distinct phenotypes was not universally accepted (31).

Independent evidence of vagal bronchopulmonary C-fiber subtypes was obtained by considering the location of the cell body rather than the nerve endings (35). The cell bodies of ∼50% of the vagal C-fibers innervating the guinea pig lungs were found to reside in the nodose ganglion with the balance located in the jugular vagal sensory ganglia. These two ganglia are embryologically distinct, with neurons in the jugular ganglion derived from neural crest cells of the postotic hindbrain, whereas the nodose neurons arise from the epibranchial placodes (2). Both the nodose and jugular C-fibers were found to terminate within the lung compartment, but only jugular C-fibers terminated in the extrapulmonary airways (30, 35). The nodose and jugular C-fiber neurons could be readily distinguished pharmacologically. Both subtypes responded directly to capsaicin and bradykinin, but only the nodose C-fibers responded with action potential discharge to ATP and other purinergic P2X receptor-selective agonists, adenosine receptor (A1 and A2A) agonists, and 5-HT3 receptor agonists (5, 6, 35). In addition, the jugular C-fibers were more apt to express neuropeptides than the nodose C-fibers (35). Others studies have shown that the vagal sensory innervation of the respiratory tract of mice and rats is also derived from both nodose and jugular ganglion neurons (18, 32). The advantage of studying the guinea pig, however, is that unlike the adult mouse, where the two ganglia are often found fused into a nodose-jugular complex, and even the rat, where they are distinct but not completely separate structures, in the guinea pig the two ganglia form readily recognizable discrete structures (e.g., see Fig. 1 in Ref. 13).

Fig. 1.

α,β-methylene ATP (10 μM)-evoked nodose C-fiber stimulation in isolated innervated perfused guinea pig lung preparation is reversibly antagonized by A-317491, a selective P2X3 and P2X2/3 receptor antagonist. Inset: representative recording of action potential discharge elicited by α,β-methylene ATP. Bar indicates perfusion delivery of drug. *Statistically significant (P < 0.05) comparison with control response.

Action potential discharge in response to ATP also differentiates nodose from jugular C-fibers innervating the guinea pig esophagus. There are seven members of the P2X receptor family (17). Members of this family assemble as homomeric and heteromeric channels. Extensive studies have revealed that the P2X2 and P2X3 subunits are particularly relevant in peripheral sensory neurons (11). Receptor activation by ATP elicits a nonselective cationic conductance with a distinctive kinetics: activation of P2X3 homomeric channels display a rapidly inactivating conductance, whereas activation of P2X2/3 heteromeric channels display a non-inactivating conductance (11). In both lungs and esophagus, activation of nodose C-fibers by ATP is mimicked by the P2X3 (P2X2/3) selective agonist α,β-methylene ATP (3, 37). This agonist can also stimulate nodose-derived, low-threshold mechanosensitive A-fibers in the lungs and esophagus (3, 37).

The question arises as to whether these two phenotypes of vagal C-fibers reach the tissue as naive C-fibers and then are “instructed” to a defined phenotype based on cues emanating from their different tissue microenvironments, or, alternatively, whether their embryonic origin selects the phenotype before reaching the tissue of innervation. We have addressed this question here using two strategies. First, we compared the C-fiber phenotype, focusing specifically on purinergic receptor expression and function, of spinal (dorsal root ganglion; DRG) neurons innervating the intrapulmonary tissue with that of jugular and nodose ganglion lung-specific C-fiber neurons. The neurons in the more rostral DRGs are derived from the same postotic hindbrain neural crest structure as the jugular neurons. Therefore, if the embryonic environment selects the phenotype, we predict that these neurons will resemble jugular neurons more than nodose neurons. The second strategy is to compare the placodal (nodose) and neural crest (jugular and DRG) C-fiber neurons innervating a disparate tissue environment, namely the esophageal wall. We again predict that if the embryonic environment is more important than the tissue environment in selecting the phenotype then the C-fiber subtypes will be the same in the two distinct tissue compartments. On the basis of our results, we conclude that the purinergic responsiveness of C-fibers, as defined by action potential discharge, is limited to placodally derived C-fibers, regardless of the tissue they innervate, and that this responsiveness is due to the selective dual expression of P2X2 and P2X3 receptors compared with the neural crest C-fibers innervating these tissues that express only the P2X3 receptor subtype.

METHODS

Experimental procedures were approved by the Johns Hopkins University Animal Care and Use Committee.

Extracellular Recording of Bronchopulmonary Vagal Afferent Nerves

The method for the extracellular recording from the vagal sensory neurons projecting to the guinea pig lungs has been described in detail previously (3). Briefly, male Hartley guinea pigs (100–200 g) were killed with CO₂ and exsanguination. The blood from the pulmonary circulation was washed out by in situ perfusion with Krebs bicarbonate solution (KBS), composed of (in mM) 118 NaCl, 5.4 KCl, 1.0 NaH₂PO₄, 1.2 MgSO₄, 1.9 CaCl₂, 25.0 NaHCO₃, and 11.1 dextrose, gassed with 95% O₂-5% CO₂ (pH 7.4). The buffer contained 3 μM indomethacin to reduce the indirect influence of tissue prostanoids on sensory fiber activity. Trachea and right lungs with intact right-side extrinsic vagal innervation including right jugular and nodose ganglia were dissected and placed in a two-compartment tissue bath. The right nodose and jugular ganglia along with rostral vagus nerve were placed in one compartment and lung and trachea in the second compartment. The two compartments were separately superfused with KBS (6 ml/min, 37°C). In addition, the pulmonary artery and trachea were cannulated with PE tubing and continuously perfused with KBS (4 ml/min and 2 ml/min, respectively). The tracheal perfusion pressure reflecting the airway smooth muscle contraction was measured with a pressure transducer (P23AA; Statham, Hato Rey, Puerto Rico). Before the perfusion, 10 punctures with a 26-gauge needle were made through the surface of the lung. The perfusing buffer solution therefore exits the lungs via both these puncture ports as well as via the pulmonary veins.

The glass microelectrode recording electrode was manipulated into the nodose or jugular ganglion. A mechanosensitive receptive field was identified when the mechanical stimulus (Von Frey hair, 1,800–3,000 mN) bluntly applied to the lung surface evoked a burst of action potentials. Once a mechanosensitive receptive field was identified, a brief (<1 ms) electrical stimulus was delivered by a small concentric electrode positioned over this discrete mechanosensitive region to determine the conduction velocity of the fiber.

The drugs were diluted in KBS and infused simultaneously into both the tracheal and pulmonary artery circuit at a rate of 50 μl/s. The agonist α,β-methylene ATP was infused through each route in a volume of 1 ml. The selective P2X receptor antagonist (1 or 10 μM) was infused to the lungs simultaneously via tracheal and pulmonary artery perfusion 30 min before agonist. All the activity evoked by a given concentration of agonist was recorded in 1-s bins and analyzed offline.

Identification of Ganglion Neurons Innervating Organ-Specific Compartments

Seven animals were anesthetized with ketamine (50 mg/ml) and xylazine (2.5 mg/ml) dissolved in PBS. Supplemental anesthesia was given as needed to abolish the hind paw-pinch reflex. A fluorescent retrograde tracer, DiI (2% in DMSO, then diluted 1:10 in sterile saline; Invitrogen, Carlsbad, CA), was used for identification of jugular, nodose, and DRG neurons innervating organ-specific compartments as previously described (35, 37). Briefly, for tracing of lung- and airway-specific neurons, DiI was injected into the lumen of the trachea through a tracheal incision. For labeling of esophagus-specific neurons, the esophagus was surgically exposed, and DiI was injected into two to five sites up to the total volume 20 μl. All animals were closely monitored on an hourly basis for several hours postoperatively and twice daily thereafter. The animals were allowed to recover for ∼2 wk for sufficient labeling of cell bodies in the various ganglia. Any animal that displayed behaviors indicating excessive pain or infection was euthanized immediately. Postmortem showed that the dye injected into the trachea remained delimited to respiratory tissues, and the dye injected into the esophagus remained in the esophagus.

Cell Culture

The cell culture is a modification of the protocol performed previously (19). Briefly, animals were euthanized using CO₂ asphyxiation. The jugular, nodose, or dorsal root ganglia (T1 to T4) were separately treated with an enzymatic cocktail of dispase (2 mg/ml) and collagenase (2 mg/ml) dissolved in HBSS without calcium or magnesium. The cells were gently triturated using fire-polished Pasteur pipettes. The cells were washed in L-15 media supplemented with 10% fetal bovine serum and centrifuged. To increase the density of culture, the pellet was resuspended in a small volume of L-15 media supplemented with 10% fetal bovine serum, pipetted (25 μl) onto the center coverglass treated with poly-d-lysine (0.1 mg/ml) and laminin (5 μg/ml), and incubated at 37°C for 2 h. The cells were flooded with additional L-15 supplemented with 10% fetal bovine serum and incubated at 37°C. Recordings were performed between 12 and 24 h postculture. All cell culture materials were obtained from Invitrogen.

Patch-Clamp Electrophysiology

Voltage clamp recordings were made using standard whole cell patch-clamp technique (lung-specific neurons) or perforated patch technique (esophagus-specific neurons). A total of 38 lung-specific neurons harvested from 18 animals was analyzed. A total of 36 esophagus-specific neurons from 19 animals was analyzed. Patch pipettes (1–3 MΩ) were fabricated from borosilicate capillary tube with filament (Sutter Instrument, Novato, CA). The pipette solution was composed of (in mM) 100 CsCl, 40 tetraethylammonium-Cl, 10 NaCl, 1 CaCl₂, 10 MgCl₂, 10 HEPES, 11 EGTA, and 2 Mg-ATP, pH 7.2, 334 mosM. Gramicidin (1 μg/ml) was dissolved in the pipette solution for the perforated patch experiments. The bath solution was Locke solution, which was composed of (in mM): 136 NaCl, 5.6 KCl, 2.2 CaCl₂, 1.2 MgCl₂, 14.3 NaHCO₃, 1.2 sodium phosphate, and 10 dextrose, pH 7.4, 336 mosM, gassed with a mixture of 95% O₂ and 5% CO₂. Electrophysiological recordings were made using a 700A Multiclamp amplifier and Digitata 1320A (Molecular Devices, Sunnyvale, CA). Voltage clamp signals were sampled at 50 kHz and filtered at 2 or 10 kHz. Membrane currents were normalized to the cell capacitance and expressed as current density (pA/pF).

Recordings were made at room temperature (lung-specific neurons) or at 35–37°C using an inline heating system (esophagus-specific neurons). A gravity-feed perfusion system was used to change the bath solution and to deliver drugs (∼8 ml/min). A complete solution change in the recording chamber was estimated to be <1 min. Cells were categorized as sensitive to capsaicin if the evoked current magnitude was >−2 pA/pF (−100 pA for an average cell) and displayed an appropriate time course for onset and offset of the response.

Single-Cell RT-PCR

Experiments were performed from lung- and esophagus-specific neurons identified by retrograde DiI labeling as described and published previously (19, 23). An average of 3–10 lung-specific or esophagus-specific neurons were harvested from a total of 7 lung- or esophagus-labeled animals. Individual fluorescent neurons from nodose, jugular, or DRG (or perfusion buffer from the vicinity of these cells as a bath control) were collected in microcentrifuge tubes that contained 1 μl of RNase OUT (Invitrogen) and immediately placed on dry ice. Cells were processed using SuperScript III CellsDirect cDNA Synthesis Kit (Invitrogen) according to the manufacturer's instructions. For first-strand synthesis, 1 μl of oligo(dT) and 1 μl of random hexamers (Roche Applied Science, Indianapolis, IN) were used for each reverse transcriptase (RT) reaction. RT and no-RT reactions were obtained from each cell. First-strand reaction conditions were 50°C for 50 min followed by 85°C for 5 min. The PCR reaction mix (final volume of 20 μl) contained 0.5 units of HotStar Taq Polymerase (Qiagen, Valencia, CA) supplemented with 2.5 mM MgCl₂, PCR buffer, dNTPs, custom synthesized primers for β-actin (sense: 5′-tggctacagtttcaccacca-3′, antisense: 5′-ggaaggagggctggaaga-3′, GenBank/EMBL/DDBJ acc. no. AF508792, calculated product length 212 bp); TRPV1 (sense: 5′-gtcgtcgcctattgattgct-3′, antisense: 5′-agagcttgagtggcttctcg-3′ or sense: 5′-cagagagccatcaccatcct-3′, antisense: 5′-gggaccagggcaaagttc-3′, GenBank/EMBL/DDBJ acc. no. AY729017, calculated product length 167 or 284 bp, respectively); P2X2 (sense 5′-gctgctcatcctgctctacttt-3′, antisense 5′-gggcttcacatactcctccac-3′, GenBank/EMBL/DDBJ acc. no. AF053327/AF053328/AF053329, calculated product length 157 bp); or P2X3 (sense 5′-ttcccctggctacaacttca-3′, antisense 5′-cagtcacctcctcaaacttcct-3′, GenBank/EMBL/DDBJ acc. no. NM145526, calculated product length 266 bp), and 3 μl of template (cDNA, no-RT control or bath control). As a positive control, cDNA from a whole trigeminal ganglion (isolated by RNeasy Plus Mini Kit, Qiagen) and reverse transcription by OmniScript RT Kit (Qiagen) was used. Primers were purchased from Invitrogen. PCR reaction conditions were initial activation at 95°C for 15 min, denaturation at 94°C for 30 s, annealing at 60°C for 30 s, extension at 72°C for 60 s for 45 cycles, and final extension at 72°C for 10 min. Control experiments testing the amplification of the no-RT of each individual neuron (or bath control) by using either one or multiple primer pairs did not produce a specific product. Electrophoresis of PCR products was run on 1.5% agarose gels.

Statistics

Data were analyzed using one-way ANOVA. Pairwise comparisons were made using Tukey post hoc tests. P values of <0.05 were considered significant.

RESULTS

Lung-Specific Sensory Neurons

Nerve ending responses.

We have previously reported the P2X3 (P2X2/3)-selective purinergic agonist α,β-methylene ATP stimulated action potential discharge in 20 of 20 guinea pig lung C-fibers arising from the nodose ganglion, but failed to evoke action potential discharge in jugular C-fibers in the lung (0 of 7 fibers tested) (35). To provide further evidence that the effect of α,β-methylene ATP on nodose C-fibers in the lung involved P2X3 or P2X2/3 receptor activation, we evaluated the response before and after addition of a selective P2X3, P2X2/3 receptor antagonist A-317491 (15). At 10 μM, α,β-methylene ATP evoked 125 ± 23 action potentials in every nodose lung C-fiber tested. This was reversibly inhibited by nearly 70% with the selective P2X3 receptor antagonist A-317491 (10 μM) (Fig. 1).

P2X currents in lung-specific capsaicin-sensitive neurons.

Bath application of α,β-methylene ATP caused a large slowly inactivating inward current in 10 of 10 lung-specific (retrogradely labeled with DiI in the lungs), capsaicin-sensitive nodose neurons (Fig. 2). Consistent with its lack of effect at the jugular C-fiber nerve terminals, α,β-methylene ATP did not evoke large slowly inactivating currents in lung-specific, capsaicin-sensitive jugular neurons (n = 14). Although large inward currents were absent in jugular neurons, in six jugular neurons, a much smaller, rapidly inactivating current was observed (in Fig. 2, note difference in scale between jugular and nodose neurons).

Fig. 2.

A: inward currents were elicited by α,β-methylene ATP (30 μM) in lung-labeled nodose ganglion (left), jugular ganglion (middle), and dorsal root ganglion (right) neurons. Neurons were capsaicin sensitive. Standard ruptured patch whole cell recordings were made in Locke solution. Recordings were made at room temperature. Holding potential was −60 mV. Bar indicates duration of α,β-methylene ATP application. B: comparison of peak α,β-methylene ATP-elicited current density revealed a significant overall difference among the 3 populations of neurons (1-way ANOVA, P < 0.001). Pairwise comparisons (Tukey) revealed significant differences between nodose (n = 10) and jugular (n = 14) populations (P < 0.05) and between nodose and dorsal root (n = 7) populations (P < 0.05). *Statistically significant (P < 0.05) comparison with nodose ganglion neurons.

DiI injected into the lungs also labeled a small population (1–2%) of neurons in the thoracic T1 to T4 DRGs. Similar to the jugular neurons, the lung-specific, capsaicin-sensitive DRG neurons failed to respond to α,β-methylene ATP with a large slowly inactivating inward current (n = 7, from 2 guinea pigs). Also, like the lung-labeled jugular neurons, three of the seven DRG neurons responded to α,β-methylene ATP with a very small rapidly inactivating current (Fig. 2).

Single neuron RT-PCR analysis of P2X receptor expression.

P2X2 and P2X3 receptors have been found to account for the electrophysiological purinergic responses in primary afferent neurons (7). Lung-labeled neurons were obtained from three animals. Those neurons that expressed TRPV1 mRNA (presumably capsaicin-sensitive neurons) were further evaluated for expression of P2X2 and P2X3 receptors at the level of the single identified neuron. All 24 lung-labeled TRPV1-positive nodose neurons evaluated expressed both P2X3 and P2X2 mRNA (Fig. 3).

Fig. 3.

Single-cell RT-PCR analysis of 24 lung-specific nodose neurons (A), 22 lung-specific jugular neurons (B), and 13 lung-specific dorsal root ganglion (DRG) neurons (C). Message for P2X2 and P2X3 was analyzed in TRPV1-positive neurons. β-actin was used as a positive control. Top, middle, and bottom bands of the ladder represent 300, 200, and 100 bp, respectively. Lane + indicates positive control from whole trigeminal ganglion. Lane - represents no-RT control. Lane B indicates a representative bath control.

P2X3 mRNA was also expressed in 18 of 22 TRPV1-positive lung-labeled jugular neurons. In contrast to the nodose neurons, however, only 3 of 22 jugular neurons also expressed P2X2 (Fig. 3). Similar to the jugular neurons, 11 of 13 TRPV1-positive lung-specific DRG neurons expressed P2X3 mRNA, but none of the 13 expressed P2X2 mRNA (Fig. 3). These data would indicate that nodose neurons are apt to express P2X2/3 heteromeric receptors, whereas the jugular and DRG neurons will likely express mainly homomeric P2X3 receptors.

Esophagus-Specific Jugular, Nodose, and DRG C-Fiber Neurons

We evaluated the phenotypes of capsaicin-sensitive neurons innervating the esophagus, focusing specifically on the response to α,β-methylene ATP. We previously noted that 13 of 16 nodose C-fiber in the esophagus responded to α,β-methylene ATP with an action potential discharge that averaged nearly 10 Hz, whereas 0 of 8 jugular C-fibers responded to the P2X3 agonist (37). As with the lungs, dye injected specifically into the wall of the esophagus backlabeled neurons in the nodose, jugular, and rostral DRGs (n = 4 animals). Whole cell patch-clamp recordings of esophageal-specific nodose neurons responded to α,β-methylene ATP in the same fashion as lung-specific nodose neurons with a large slowly inactivating current (n = 16 neurons; Fig. 4). The esophageal-labeled neurons in the jugular and the DRG (T1 to T4) failed to respond with a large slowly inactivating current; however, as observed with lung-specific neurons, some (3 of 10 jugular and 4 of 10 DRG neurons) responded with the small rapidly inactivating current that was also observed in lung-labeled neurons.

Fig. 4.

A: inward currents were elicited by α,β-methylene ATP (10 μM) in esophagus-labeled nodose ganglion (left), jugular ganglion (middle), and DRG (right) neurons. Neurons were capsaicin sensitive. Perforated patch whole cell recordings were made in Locke solution. Holding potential was −60 mV. Recordings were made at 35–36°C. Bar indicates duration of α,β-methylene ATP application. B: comparison of peak α,β-methylene ATP-elicited current density revealed a significant overall difference among the 3 populations of neurons (1-way ANOVA, P < 0.001). Pairwise comparisons (Tukey) revealed significant differences between nodose (n = 16) and jugular (n = 10) populations (P < 0.05) and between nodose and dorsal root (n = 10) populations (P < 0.05). *Statistically significant (P < 0.05) comparison with nodose ganglion neurons.

Single neuron RT-PCR analysis of P2X receptor expression.

Esophagus-specific TRPV1-positive neurons were evaluated for coexpression of P2X2 and P2X3 mRNA (Fig. 5). All nine TRPV1-positive nodose ganglion neurons expressed P2X3 mRNA, and eight of these neurons also expressed P2X2. All seven TRPV1-positive jugular ganglion neurons innervating the esophagus expressed P2X3, but only one of these neurons coexpressed P2X2. Likewise, of the six esophagus-specific TRPV1-expressing DRG neurons evaluated, all expressed P2X3, but P2X2 mRNA was not detected in any of the esophagus-specific DRG neurons. No products were detected in the no-RT or bath controls (not shown). Overall, the pattern of mRNA expression in esophagus-specific neurons is in full agreement with that in the lung-specific neurons.

Fig. 5.

Single-cell RT-PCR analysis of 9 esophagus-specific nodose neurons (A), 7 esophagus-specific jugular neurons (B), and 6 esophagus-specific DRG neurons (C). Message for P2X2 and P2X3 was analyzed in TRPV1-positive neurons. β-actin was used as a positive control. Expected sizes for the β-actin, TRPV1, P2X2, and P2X3 products were 212, 284, 157, and 266 bp, respectively.

DISCUSSION

The data presented here provide direct evidence, at the level of purinergic receptor gene expression, of two distinct phenotypes of capsaicin-sensitive C-fibers innervating guinea pig lungs and esophagus. The data support the hypothesis that the two phenotypes are based less on the location and nature of the visceral tissues the nerve terminals innervate than on the embryological origin of the ganglion in which their cell bodies are situated. One type of C-fiber has its cell body in the neural crest-derived vagal jugular ganglia or rostral DRG; the other type has its cell body in the vagal nodose ganglia, a structure derived embryologically from the epibranchial placodes. Virtually all C-fibers innervating the lungs and esophagus expressed purinergic P2X3 receptors. This was somewhat surprising inasmuch as only the nodose-derived C-fiber terminals respond to ATP with action potential discharge. This is likely explained by the finding that only the nodose-derived neurons also expressed P2X2 receptors and thus are capable of forming functional P2X2/3 heteromeric receptors.

In several cranial nerves, the cell bodies of sensory neurons are spatially segregated into a neural crest-derived proximal “root” ganglion and a placodal-derived distal ganglion (2). In the vagus nerve, these ganglia are referred to as jugular and nodose, respectively. In mice and rats, these ganglia may fuse to form a nodose-jugular complex. In guinea pigs and larger mammals, these ganglia are spatially distinct entities (e.g., see Fig. 1 in Ref. 13). We previously reported that guinea pig lung C-fibers arising from nodose ganglia could be differentiated from the jugular C-fibers in the lungs by their neuropeptide content and by their responsiveness to purinergic receptor activation (35). The terminals of nodose C-fibers responded with action potential discharge to ATP and the P2X3, P2X2/3 selective agonist α,β-methylene ATP, whereas the jugular C-fibers failed to respond. On the other hand, the jugular C-fiber neurons nearly uniformly expressed tachykinins, whereas relatively fewer nodose C-fiber neurons expressed these neuropeptides.

There are seven members of the purinergic P2X family of receptors, P2X1 to P2X7 (17). The putative receptor subunit stoichiometry for the ion channel is thought to be homo- and heterotrimers (16, 24, 33). A wealth of information exists regarding the expression of purinergic receptors in peripheral ganglia (11). In nodose and dorsal root ganglia, P2X2 and P2X3 are the purinergic receptor subunits responsible for neuron activation (7). Here we confirm the distinction between nodose and jugular lung C-fiber terminals with respect to activation by α,β-methylene ATP and go on to show that this response is strongly inhibited by A-317491. This drug is a selective antagonist of P2X3 and P2X2/3 receptors (15).

Extensive electrophysiological studies have revealed that both homomeric P2X3 receptors and P2X2/3 receptors can be stimulated by α,β-methylene ATP, but the kinetics of the ionic current is markedly different (10, 17, 21). The homomeric P2X3 response is a transient, very rapidly inactivating current, whereas the P2X2/3 current is a long-lasting, slowly inactivating current. In our patch-clamp studies, we found, as predicted from the single neuron RT-PCR analysis of P2X receptor mRNA expression, that the capsaicin-sensitive nodose neurons innervating the lungs responded to α,β-methylene ATP with large, slowly inactivating P2X2/3-like current, whereas the jugular neurons failed to respond to α,β-methylene ATP, or responded only with a very transient P2X3-like current. That nodose neurons express P2X2/3 heteromeric receptors is consistent with both immunohistochemical and electrophysiological studies showing coexpression of P2X2 and P2X3 in a substantial population of nodose ganglion (36). It should be noted that the conditions were not optimized to study the rapidly inactivating P2X3-like current, and therefore the peak currents observed may be underestimated. Nevertheless, the data from the ex vivo vagus innervated lung preparation indicate that if the transient P2X3 current observed in capsaicin-sensitive jugular cell bodies exists in the terminals in the lungs and esophagus, it is not sufficient to evoke action potential discharge. One might speculate that the homomeric P2X3 receptors in these neurons may play a more important role at the central synapse than at the peripheral terminals.

The question arises as to whether the two C-fiber phenotypes are instructed by signals they received from the different compartments their terminals may innervate within the lungs or whether they are “preselected” and arrive at the tissue already as distinct phenotypes. The jugular neurons are derived from neural crest tissue in the region of the postotic hindbrain, whereas the nodose neurons are derived from the most caudal epibranchial placodes (2). This likely has ramifications with respect to nerve phenotype in development. The suite of transcription factors coordinating neurogenesis and development of sensory neurons in these two ganglia is only partially understood. It suffices to say that among the overlap of transcription factors, many are expressed selectively in placodal neurons, whereas others are expressed selectively in sensory neurons of neural crest origin (see Ref. 2).

Although both jugular and nodose C-fibers studied here innervated the lungs, it is likely that there may be substantive differences in the compartments within the lungs that the endings terminate. In fact, the jugular C-fibers innervate the guinea pig major conducting airways (i.e., trachea and main bronchus), whereas the nodose C-fibers do not (35). It is therefore possible that cues from these respective tissue compartments innervated may direct the C-fiber phenotype with respect to responsiveness to purinergic agonists and neuropeptide content. The neurons within the rostral DRGs are derived from the same postotic hindbrain neural crest as the jugular ganglia (2) and presumably would depend on a similar suite of transcriptional regulators for neurogenesis and development as jugular neurons. If it is the embryologic origin that is the most important factor in determining the characteristics of the C-fibers studied here, we would therefore predict that the C-fibers innervating the lungs that arise from the rostral DRGs would be more similar to jugular neurons than nodose neurons. Indeed, with respect to P2X receptor expression and function, the lung-specific DRGs seemed essentially indistinguishable from lung-specific jugular neurons. Recently, it has been found that the vast majority of small-diameter DRG neurons innervating guinea pig lungs expresses tachykinins (25). This too puts them in line with the neuropeptide-containing jugular C-fiber phenotype and differentiates them from most lung nodose neurons (35).

Additional evidence that would seem to indicate that ganglionic origin is more important than tissue of innervation with respect to directing the C-fiber phenotype comes from our findings with neurons innervating the wall of the esophagus. The tissue environment of the esophageal wall is distinct from the lungs, yet, at least with respect to their purinergic receptor pharmacology, the esophageal nodose neurons were very similar to the intrapulmonary nodose neurons, and the phenotype of the esophageal jugular/DRG neurons was basically the same as the intrapulmonary jugular/DRG neurons. Although this study focused on the functional analysis at the level of the esophageal-specific sensory cell soma, we have previously reported that, at least with respect to P2X sensitivity, the terminals of nodose C-fibers, but not jugular C-fibers, in the guinea pig esophagus responded to ATP and α,β-methylene ATP with action potential discharge (37). In addition, it has been shown that the neuropeptide-containing C-fibers innervating the esophagus are predominately jugular C-fibers (37). It should be noted that although recording conditions were different between the recordings of lung-specific (ruptured patch at room temperature) and esophagus-specific neurons (perforated patch at body temperature), similar results were found in each group. We can therefore conclude that with respect to neuropeptide content as well as purinergic receptor expression, the placodal and neural crest C-fiber phenotypes, in general, are similar between lungs and esophagus.

The hypothesis favored here that the two purinergic receptor phenotypes of vagal C-fibers innervating lungs and esophagus is instructed by signals early in development finds support from the observations of Cheung and Burnstock (4). These investigators noted that in rats, P2X3 is expressed in various sensory ganglion neurons as early as embryonic day 12.5 and is coexpressed with P2X2 in some rat nodose neurons as early as embryonic day 16.5. The type of modulation of vagal C-fiber phenotype that is caused by tissue-derived cues in inflammatory airway disease will likely depend on whether the C-fiber in question is neural crest or placodally derived.

GRANTS

This work was supported by grants from the National Institutes of Health.