Opposing effects of bronchopulmonary C-fiber subtypes on cough in guinea pigs

Abstract

We have addressed the hypothesis that the opposing effects of bronchopulmonary C-fiber activation on cough are attributable to the activation of C-fiber subtypes. Coughing was evoked in anesthetized guinea pigs by citric acid (0.001–2 M) applied topically in 100-µl aliquots to the tracheal mucosa. In control preparations, citric acid evoked 10 ± 1 coughs cumulatively. Selective activation of the pulmonary C fibers arising from the nodose ganglia with either aerosols or continuous intravenous infusion of adenosine or the 5-HT₃ receptor-selective agonist 2-methyl-5-HT nearly abolished coughing evoked subsequently by topical citric acid challenge. Delivering adenosine or 2-methyl-5-HT directly to the tracheal mucosa (where few if any nodose C fibers terminate) was without effect on citric acid-evoked cough. These actions of pulmonary administration of adenosine and 2-methyl-5-HT were accompanied by an increase in respiratory rate, but it is unlikely that the change in respiratory pattern caused the decrease in coughing, as the rapidly adapting receptor stimulant histamine also produced a marked tachypnea but was without effect on cough. In awake guinea pigs, adenosine failed to evoke coughing but reduced coughing induced by the nonselective C-fiber stimulant capsaicin. We conclude that bronchopulmonary C-fiber subtypes in guinea pigs have opposing effects on cough, with airway C fibers arising from the jugular ganglia initiating and/or sensitizing the cough reflex and the intrapulmonary C fibers arising from the nodose ganglia actively inhibiting cough upon activation.

INTRODUCTION

The role of bronchopulmonary C fibers in cough has been the subject of considerable debate. Aerosols of capsaicin, acid, resiniferatoxin, anandamide, cinnamaldehyde, nicotine, and bradykinin evoke cough in awake humans or guinea pigs (3, 22, 28, 32, 35, 42, 43). Bronchopulmonary C fibers are activated by all of these agents, in some cases selectively (8, 12, 17, 29, 37, 38, 43, 44, 60). Activation of the ionotropic receptors transient receptor potential-vanilloid 1 (TRPV1) and transient receptor potential ankyrin 1 (TRPA1) is involved with the coughing evoked by many of these irritants, and bronchopulmonary C fibers preferentially express these ion channels (2–4, 8, 24, 27, 32, 37, 38, 41, 44, 52, 66, 68). Coughing induced by capsaicin and citric acid in guinea pigs is also reduced or abolished by neurokinin receptor antagonists, and neurokinins, like TRPV1 and TRPA1, are also preferentially expressed by C fibers in the guinea pig airways (5, 18, 26, 28, 40, 45, 48, 60, 69, 70). Together, these data argue strongly for an essential role of C fibers in cough.

In contrast to the results of cough studies in awake animals, however, C-fiber-selective stimulants have consistently failed to evoke coughing in anesthetized animals (12, 14, 64, 65). This is not because anesthesia prevents coughing evoked by all stimuli or prevents C-fiber activation. Anesthetized animals cough reliably to mechanical stimulation or acid applied topically to the laryngeal, tracheal, or bronchial mucosa (12, 31, 59, 64, 65, 72). Despite clear evidence for C-fiber activation in anesthetized animals and evidence for other C-fiber-dependent reflexes (7, 12, 14, 17, 29, 30, 45), thorough physiological analyses of the afferent nerves regulating the cough reflexes in anesthetized dogs, cats, and guinea pigs argue in favor of a role for myelinated “cough receptors” and argue equally strongly against a role for C fibers (12, 48, 64, 65, 72, 73).

The observations summarized above are not necessarily contradictory but suggest two separate afferent pathways regulating cough, with C-fiber-dependent cough more sensitive to general anesthesia than the C-fiber-independent cough reflex. However, what is not so readily explained has been the observation that C-fiber activation not only fails to evoke cough in anesthetized animals, it can also be acutely inhibitory of cough (64, 65). Anesthesia must thus accentuate an inhibitory signal for cough initiated by C-fiber stimulation or the typically more invasive preparations used for reflex studies in anesthetized animals may preferentially target an exclusively inhibitory pathway for C fibers in cough.

Subtypes of bronchopulmonary C fibers have been described previously (17, 36, 54, 56, 70). In guinea pigs, C-fiber subtypes are differentiated based on their ganglionic origin (nodose vs. jugular ganglia), their sites of peripheral termination (extrapulmonary vs. intrapulmonary), expression of neurokinins, and responsiveness to adenosine, serotonin 5-HT₃, and ATP/P2X2/3 receptor agonists (15, 16, 41, 70). In relation to cough, we have shown previously that C-fiber activation in the large intrapulmonary and extrapulmonary airways of anesthetized guinea pigs with either capsaicin or bradykinin greatly enhanced sensitivity to subsequent tussive stimuli applied to the tracheal mucosa (47). This effect of C-fiber activation was reversed by centrally administered neurokinin receptor antagonists, a result that implicates jugular C fibers in cough initiation and sensitization. We have also shown that in awake guinea pigs, the nodose-selective stimulants adenosine and 2-methyl-5-HT fail to evoke coughing upon inhalation and that inhibiting action potential conduction in nodose afferents by selectively reducing expression of the voltage-gated sodium channel NaV1.7 in nodose neurons is without effect on capsaicin-evoked cough responses (53). These results are consistent with an excitatory role for jugular C fibers in cough, but no clear role for nodose C fibers. In the present study, we have set out to address the hypothesis that jugular C fibers promote coughing in guinea pigs while nodose C-fiber activation is acutely inhibitory to cough.

METHODS

The majority of the experiments described in this study were carried out using anesthetized male Hartley guinea pigs (200–400 g, pathogen-free, Harlan) and were approved by the Johns Hopkins Animal Care and Use Committee. Guinea pigs were anesthetized with 1.5 g/kg urethane ip, which produces a deep anesthesia lasting far longer than the maximum of 2 h of experimentation we have performed. We confirmed adequate anesthesia throughout the experiment by monitoring responses to pinching of a hindlimb. If deemed necessary (never in this set of experiments), as much as 0.5 mg/kg more urethane (dosed incrementally to effect) would have been provided to achieve stable anesthesia. When the experiments were completed, animals were asphyxiated with carbon dioxide and then exsanguinated.

Anesthetized, guinea pigs were secured supine on a warming pad. The extrathoracic trachea was cannulated at its caudal most end with a bent luer stub adaptor. The cannula was attached to a length of tubing that terminated inside a chamber that warmed and humidified inspired room air. Care was taken to preserve the innervation and vasculature of the trachea. A pressure transducer attached to a side port in the tracheal cannula monitored respiratory efforts, which were recorded digitally (Biopac data acquisition system, Santa Barbara, CA).

The segment of the extrathoracic trachea rostral to the tracheal cannula was opened lengthwise along its ventral most aspect and retracted bilaterally. Polyethylene (PE) tubing was threaded through the larynx and nasal cavity and out a nostril. Starting from the caudal end of the trachea, warmed, oxygenated Krebs bicarbonate buffer (composition in mM: 118 NaCl, 5.4 KCl, 1 NaHPO₄, 1.2 MgSO₄, 1.9 CaCl₂, 25 NaHCO₃, and 11.1 dextrose, pH 7.4) was superfused (3 ml/min) over the tracheal mucosa and removed at the rostral end of the trachea by attaching the PE tubing threaded through the upper airways to a gentle suction source. Formation of neuromodulatory prostanoids was prevented with 3 µM indomethacin added to the perfusing buffer. In some preparations, a jugular vein was cannulated with heparinized PE-60 tubing for intravenous drug challenges.

When the surgery was complete, animals were allowed to breathe spontaneously and without any further manipulations for 10 min. Thereafter, we evoked cough by applying citric acid (0.001–2M) topically to the tracheal mucosa, delivered in 100-µl aliquots at 1-min intervals (11). Coughing evoked from the trachea of anesthetized guinea pigs is attributable to the activation of Aδ “cough receptors,” with the acid activating these afferent nerves in a TRPV1-independent manner, presumably working through acid-sensing ion channels (11, 12, 27, 37, 48). Concentration-response curves were constructed in an ascending fashion. Cough was defined by visual confirmation of cough-like respiratory efforts and associated signature pressures, including a ≥200% increase in peak expiratory pressure preceded by an enhanced inspiratory effort, all occurring in <1 s. Each individual cough effort during any stimulus was counted as a cough; coughing bouts (defined here as 1 more coughing efforts) were not quantified.

We attempted to modulate citric acid-induced cough by activating nodose C fibers with adenosine, 2-methyl-5-HT, and bradykinin. We also studied the effects of the rapidly adapting receptor-selective stimulus histamine on citric acid-induced cough. The doses of each agonist used in these experiments were selected based on the results of previous studies indicating an excitatory effect on afferent nerves in vitro or reflexes studied in vivo (12, 14–16). Adenosine (0.3 mg·kg⁻¹·min⁻¹), 2-methyl-5-HT (0.4 mg·kg⁻¹·min⁻¹), bradykinin (1 nmol·kg⁻¹·min⁻¹), histamine (1 µg/min), or vehicle (saline; 100 µl/min) were administered by continuous intravenous infusion as described elsewhere (14). In the studies with bradykinin, animals were first pretreated with the angiotensin-converting enzyme inhibitor captopril (1 mg/kg iv) to limit bradykinin metabolism in the lung. When the effects of these agents on respiration reached equilibrium, citric acid concentration-response curves were constructed as described above. The effects of adenosine (0.1 µM) and 2-methyl-5-HT (10 µM) added directly to the tracheal perfusate on citric acid-induced cough were also evaluated, as were the effects of inhalation of adenosine (2 mg/ml) and 2-methyl-5-HT (1 mg/ml) on cough. Aerosolized inhalation challenges were delivered using a nebulizer connected in series with a pump delivering air to the chamber from which the guinea pigs inhaled warmed and humidified air. As with the intravenous infusion experiments, adenosine and 2-methyl-5-HT inhalation continued until their effects on respiratory rate reached equilibrium, and then the citric acid concentration-response curves were carried out. Vehicle control experiments with aerosolized vehicle were carried out in parallel.

To further characterize the effects of C-fiber-selective stimuli on cough and on respiratory rate, we quantified changes in breathing pattern and the number of coughs evoked during 10 min of continuous perfusion of the trachea with capsaicin (10 µM), adenosine (0.1 µM), or 2-methyl-5-HT (10 µM) and in response to bolus intravenous administration of capsaicin (2 µg/kg) and 2-methyl-5-HT (1 mg/kg). These doses were selected based on the results from previous or preliminary studies (12, 14–16).

To evaluate the time course of the effects of nodose C-fiber-dependent modulation of cough, we evoked cough electrically at 2-min intervals before during and after a continuous infusion of adenosine (0.3 mg·kg⁻¹·min⁻¹) or vehicle. As described previously (12, 13, 47, 48), electrically evoked cough was induced using a bipolar platinum electrode placed on the tracheal mucosa and receiving trains of square pulses of supraoptimal intensity (16 Hz, 10-s trains, 1-ms pulse duration, 8 V) at 2-min intervals. Finally, to evaluate the relative ability of C-fiber subtypes to initiate coughing and the interaction between these subtypes during cough, we studied the effects of capsaicin (which activates both jugular and nodose C fibers) and adenosine (a nodose C-fiber-selective stimulant) using a conscious cough model. There were two treatment groups for this study. One group of guinea pigs was challenged with aerosolized adenosine (10 mM) for 10 min and then, 10 min later, challenged again with aerosolized capsaicin (3–10 µM) combined with adenosine. The second group of animals was initially challenged with the vehicle for adenosine (saline), followed by capsaicin dissolved in saline. The number of coughs evoked by each of these aerosol challenges was counted. Coughing was monitored using an EMKA system configured specifically for cough, with coughs identified by chamber pressure recordings and sound and visually by the experimenter (EMKA Technologies, Falls Church, VA).

Statistical analyses.

A nonpaired, parallel group experimental design had to be employed. Results are presented as a means ± SE of n experiments where n is a single animal. Differences among group means were assessed by one-way ANOVA and Scheffé F-test for unplanned comparisons. P < 0.05 was considered statistically significant. Rarely (<10% animals), guinea pigs did not cough during surgery or in response to citric acid and had basal respiratory rates of ≤45 breaths/min. These animals were excluded from subsequent analyses.

Reagents.

Adenosine, 2-methyl-5-HT, capsaicin, histamine, indomethacin, captopril, bradykinin, and citric acid were all purchased from Sigma (St. Louis, MO). Capsaicin and indomethacin were prepared as stock solutions in ethanol. All other compounds were prepared as stock solutions in either water or saline and diluted in saline just before experimentation.

RESULTS

Respiratory reflexes evoked by C-fiber subtype-selective stimulants.

Baseline respiratory rate averaged 68 ± 1 breaths/min in control animals (n = 44). Consistent with previous studies (11–14, 48), capsaicin applied topically to the tracheal mucosa initiated apneas and respiratory slowing while adenosine and 2-methyl-5-HT were without effect when administered selectively to the trachea (Fig. 1). In contrast, intravenously administered capsaicin evoked a short-lived tachypnea followed by an apnea and a sustained precipitous fall in respiratory rate. The initial tachypnea was mimicked by intravenous administration of either adenosine or 2-methyl-5-HT (Fig. 2). None of these challenges evoked coughing in anesthetized guinea pigs.

Fig. 1.

C-fiber stimuli including C-fiber subtype-selective stimuli delivered directly to the tracheal mucosa have wide-ranging effects on breathing and cough. A: as shown previously (12, 14), capsaicin (10 µM) applied topically to the tracheal mucosa induces an apnea that terminates with an augmented breath but no coughing. B: citric acid (0.1 M), which activates the Aδ-cough receptors in addition to the jugular C fibers innervating the trachea, evokes coughing almost immediately upon topical application to the trachea and a transient slowing of respiratory rate. The traces in A and B span 5 s of continuous recordings. The gray arrows in A and B denote the onset of continuous perfusion of capsaicin and the time at which citric acid was applied topically to the tracheal mucosa, respectively. C: means ± SE change in respiratory rate evoked by 10 µM capsaicin or 0.1 µM of the nodose C-fiber-selective stimulant adenosine (16). A lack of effect on breathing like that seen with adenosine was observed with tracheal application of 10 µM 2-methyl-5-HT (not shown). D: the total number of coughs evoked by citric acid (0.001–2M), capsaicin (10 µM), adenosine (0.1 µM), and 2-methyl-5-HT (10 µM). Citric acid was applied topically in 100-µl aliquots in ascending concentrations at 1-min intervals, while capsaicin, adenosine, and 2-methyl-5-HT were delivered continuously via the tracheal perfusate. The results are presented as the means ± SE of 3–40 experiments.

Fig. 2.

These representative traces show typical changes in breathing pattern following intravenous challenge with C-fiber-selective stimulants. A: capsaicin evokes an initial tachypnea followed by a sustained apnea that terminates with an augmented breath and a gradual recovery of basal respiratory rate. B: by contrast, continuous infusion of adenosine induces a slowly developing but sustained tachypnea. Two, 10-s epochs of the breathing pattern at the outset and during the peak response to adenosine are shown. Comparable increases in respiratory rate were evoked by continuous infusion of 2-methyl-5-HT (not shown). These traces are from different animals and are representative of 6 experiments. C: means ± SE changes in respiratory rate in response to bolus infusions of capsaicin or 2-methyl-5-HT or continuous infusion of adenosine (n = 3–6).

Effects of adenosine and 5-HT₃ receptor activation on citric acid-evoked cough.

Citric acid (0.001–2 M) applied topically to the tracheal mucosa in 100-µl aliquots at 1-min intervals evoked coughing, typically 1–2 coughs/challenge with concentrations ≥0.03 M and 10 ± 1 coughs cumulatively in control preparations (Fig. 1). Occasionally, a brief cessation of respiration (or apnea) occurred following acid challenge. These apneas only occurred after cough was evoked. Also occasionally, citric acid challenges failed to evoke coughing but instead induced augmented breaths. The coughs evoked could be differentiated from augmented breaths, both having an enhanced inspiratory and expiratory component, but the expiratory component during augmented breaths was roughly symmetric or smaller in magnitude to the inspiratory component, producing expiratory pressures just two to three times that associated with tidal expiration at eupnea. The expiratory component of cough created pressures that were as much as 10 times the pressures associated with a tidal expiration. The shorter time course of the cough cycle also differentiated coughing from augmented breaths (Fig. 2).

Adenosine administered by continuous intravenous infusion evoked tachypnea but did not evoke coughing in any of the anesthetized animals used in this study. The tachypnea induced by intravenous adenosine infusion was quite pronounced (50–120% increase in respiratory rate) and was due largely to a decrease in preinspiratory pause (Fig. 2). When the effects of adenosine on respiratory rate reached equilibrium, citric acid concentration-response curves were constructed. Remarkably, during adenosine infusion, citric acid was essentially ineffective at evoking cough (Fig. 3). In fact, in most instances, citric acid challenges failed to even slightly alter respiratory pattern, much less cause cough.

Fig. 3.

Continuous infusion (100 µl/min) of adenosine (0.3 mg·kg⁻¹·min⁻¹) inhibits citric acid-evoked coughing in anesthetized guinea pigs. A: citric acid (0.001–2 M) applied topically to the tracheal mucosa reliably evokes coughing (*) and a transient decrease in respiratory rate in anesthetized guinea pigs. B: during adenosine infusion, citric acid challenge typically failed to evoke coughing or respiratory slowing but instead induced modest augmented breaths, or no effect. Responses to 0.1 M citric acid challenges are illustrated in the traces. C: the effects of continuous infusion of adenosine or the vehicle (saline) for adenosine on citric acid-evoked coughing are presented as the means ± SE of 6 experiments. *P < 0.05, statistically significant difference relative to the control value.

Adenosine has been shown to be a selective stimulant of nodose C fibers in guinea pigs (16). Nodose C fibers terminate almost exclusively in the lung, with very few terminating in the mainstem bronchi, trachea, or larynx (12, 70). Consistent with a role for nodose C fibers in the inhibition of cough induced by adenosine, intravenous infusion of the serotonin 5-HT₃ receptor-selective agonist 2-methyl-5-HT, which like adenosine selectively activates nodose C fibers in the guinea pig airways and lung (15), mimicked the effects of adenosine on citric acid-evoked coughing (Fig. 4). Inhalation of either adenosine or 2-methyl-5-HT also nearly abolished citric acid-evoked cough. By contrast, no effect on citric acid-induced cough was apparent when perfusing the tracheal segment from which cough was evoked with either 0.1 µM adenosine or 10 µM 2-methyl-5-HT (Fig. 5).

Fig. 4.

Continuous infusion (100 µl/min) of 2-methyl-5-HT (0.4 mg·kg⁻¹·min⁻¹) inhibits citric acid-evoked coughing in anesthetized guinea pigs. The data are presented as the means ± SE of 6 experiments. *P < 0.05, statistically significant difference relative to the control value.

Fig. 5.

The inhibitory effects of the nodose C-fiber-selective stimulants adenosine and 2-methyl-5-HT on citric acid-evoked coughing depend on their route of administration. Continuously perfusing adenosine or 2-methyl-5-HT over the tracheal mucosa, with the agonists added directly to the tracheal perfusate targeted by citric acid, is without effect on cough responses. Adenosine (0.1 µM), 2-methyl-5-HT (10 µM), or their vehicles (saline) were added to the tracheal perfusate 10 min before and throughout tracheal citric acid challenges (0.001–2 M). In contrast to the effects of tracheal challenge, inhalation of adenosine (2 mg/ml; n = 6) or 2-methyl-5-HT (1 mg/ml; n = 6) nearly abolished citric acid-evoked coughing. Neither adenosine nor 2-methyl-5-HT, nor their vehicles, evoked much if any change (<10% increase in respiratory rate) in respiratory pattern before generating the citric acid concentration-response curves. The results are presented as means ± SE number of cumulative coughs evoked during a citric acid concentration-response curve in 3–6 experiments. *P < 0.05, statistically significant difference relative to the control value.

Effect of bradykinin and histamine on citric acid-evoked coughing.

In guinea pigs first pretreated with the angiotensin-converting enzyme inhibitor captopril, intravenous bradykinin infusion attenuated citric acid-evoked coughing (Fig. 6). As with adenosine and 2-methyl-5-HT infusion, continuous bradykinin infusion induced tachypnea and nearly abolished citric acid-evoked coughing. This was a surprising observation, given our previous studies showing that bradykinin inhalation or superfusion over the tracheal mucosa (targeting jugular C fibers) acutely sensitized guinea pigs to subsequent tussive stimulation (47). However, in addition to activating jugular C fibers, which we think induce and/or facilitate coughing, bradykinin also activates nodose C fibers, which terminate preferentially in the intrapulmonary airways and lungs and perhaps extensively in the vasculature targeted by intravenous challenges (17, 70). It is unlikely that the inhibitory effects of adenosine, 2-methyl-5-HT, and bradykinin on cough occur secondary to the tachypnea they evoke. The rapidly adapting receptor stimulant histamine also evokes tachypnea (present study; Refs. 12, 14), but even with continuous infusion and a sustained increase respiratory rate, histamine was without effect on cough (Fig. 6). By contrast, neither adenosine nor 2-methyl-5-HT inhalation substantially altered respiratory rate but nearly abolished citric acid-evoked coughing. Like adenosine and 2-methyl-5-HT, neither bradykinin nor histamine infusion-evoked coughing in any of the anesthetized animals used in this study.

Fig. 6.

Effects of continuous infusion of bradykinin and histamine on respiratory rate and citric acid-evoked coughing in anesthetized guinea pigs. Bradykinin (1 nmol·kg⁻¹·min⁻¹), histamine (1 µg·kg⁻¹·min⁻¹), or their vehicles were continuously infused via intravenous cannula until their effects on respiration reached equilibrium. Thereafter, tracheal citric acid (0.001–2 M) concentrations response curves were constructed. The effects of bradykinin (A and B) and histamine (C and D) on respiratory rate and citric acid-evoked coughing are presented as a means ± SE of 5 experiments. *P < 0.05, statistically significant difference relative to the control value.

Time course of cough inhibition and effects on cough in conscious guinea pigs.

The inhibitory effects of adenosine on cough require sustained administration and are reversible. Thus coughing was evoked repetitively by electrical stimulation of the tracheal mucosa (16 Hz, 10-s train, 1-ms pulse duration, 8 V) at 2-min intervals before, during, and after a continuous infusion (100 µl l/min for 8 min) of saline (n = 3) or adenosine (0.3 mg·kg⁻¹·min⁻¹; n = 4). In animals receiving a saline infusion, these electrical challenges reliably evoked one to two coughs in each animal at each 2-min interval over a 20-min recording period. After 6 min of adenosine infusion, however, coughing was completely abolished and remained completely inhibited at the next 2-min interval. Upon stopping the infusions, coughing was eventually observed in all four animals challenged with adenosine and within 4 min was no different than that evoked in the animals receiving a saline infusion (Fig. 7).

Fig. 7.

Intravenous adenosine-induced cough suppression is relatively rapid in both onset and reversal. Coughs were evoked repetitively at 2-min intervals by electrically stimulating (16 Hz, 10-s train, 8 V, 1-ms pulse duration) the tracheal mucosa of anesthetized guinea pigs. Control guinea pigs (n = 3) received a continuous 8-min infusion of saline (100 µl/min) while a second group of animals (n = 4) received adenosine (0.3 mg·kg⁻¹·min⁻¹) during the 8-min infusion. Coughs were evoked reliably at each 2-min interval in control animals over the 20-min stimulation period, while cough was completely abolished within 4 min of starting adenosine infusion but returned to control levels within 4 min of terminating the adenosine challenge. The results are presented as a means ± SE of 3–4 experiments. *P < 0.05, statistically significant difference relative to the control value.

Previous studies documenting the inhibitory effects of C-fiber activation on cough in cats and dogs were carried out only in anesthetized animals (64, 65). The results summarized above suggest a comparable inhibitory effect on cough in anesthetized guinea pigs and implicate nodose C fibers as the drivers of these inhibitory effects. Importantly, however, we have now shown that the ability of nodose C-fiber-selective stimuli to inhibit coughing is also apparent in awake guinea pigs, even when the tussive stimulus is the C-fiber-selective stimulant capsaicin. Consistent with our previous studies (53), continuous inhalation of adenosine or its vehicle (saline) evoked little or no coughing. These two groups of animals were subsequently challenged with capsaicin (3–10 µM). The capsaicin was dissolved in saline in the animals first receiving a challenge with the vehicle for adenosine, while the capsaicin was dissolved with adenosine in the adenosine challenged guinea pigs. Adenosine markedly inhibited capsaicin-evoked coughing (Fig. 8).

Fig. 8.

Adenosine inhalation fails to initiate coughing but inhibits cough evoked subsequently by capsaicin inhalation in awake guinea pigs. Adenosine (10 mM) or its vehicle (not shown) was nebulized into a recording chamber containing a freely breathing awake guinea pig. Neither aerosol-evoked coughing. Subsequently, adenosine or its vehicle was added to solutions containing capsaicin (3–10 µM), which were then nebulized to study the cough reflex. None of the challenges altered breathing patterns in these conscious animals, but adenosine prevented coughing evoked by capsaicin inhalation. The results are presented as a means ± SE of 10 experiments. *P < 0.05, statistically significant difference relative to the control value.

DISCUSSION

The results of this and previous studies suggest that bronchopulmonary C-fiber subtypes have opposing effects on cough. Aerosolized stimuli that activate jugular C fibers (capsaicin, bradykinin, citric acid, resiniferatoxin, anandamide, nicotine, and TRPA1 agonists) all evoke coughing in awake guinea pigs and in human subjects. Coughing evoked by these agents can be prevented by TRPV1 and/or TRPA1 blockade, and these ion channels are preferentially expressed in C fibers (2–4, 8, 24, 32, 41, 52, 66, 68). In guinea pigs, capsaicin, citric acid, and bradykinin-evoked coughing is also inhibited by neurokinin receptor antagonists (5, 18, 26, 28, 69), and neurokinins are expressed by jugular but not nodose C fibers (40, 45, 48, 60, 70). These results strongly implicate jugular C fibers in cough.

In contrast to the effects of capsaicin, citric acid, and bradykinin and the other stimuli listed above, nodose C-fiber-selective stimulants such as adenosine and 5-HT₃ receptor agonists have been consistently ineffective at evoking cough in any species (25, 33, 53, 64, 65). On the contrary, these agents inhibit coughing. In the present study, we found that adenosine and 2-methyl-5-HT inhibited coughing evoked by three different stimuli (acid and electrical stimulation delivered to the tracheal mucosa of anesthetized guinea pigs, capsaicin aerosols delivered to awake guinea pigs), when administered by each of two different routes (aerosol, intravenous) and in both awake and anesthetized guinea pigs. Comparable inhibitory effects of nodose C-fiber stimulants on cough have been reported in cats and dogs (64, 65). In humans, neither adenosine nor serotonin evoke much if any coughing upon inhalation or intravenous administration, and acutely at least serotonin suppresses evoked cough responses in humans (9, 23, 63).

It is unclear whether the C-fiber subtypes with opposing effects on cough in species other than the guinea pig also arise from distinct vagal ganglia. It is of interest, however, that the auricular branch of the vagus nerves (Arnolds nerve), which carry afferents that terminate in the ear, can initiate coughing when stimulated in humans (67). The cell bodies for the afferents carried by Arnolds nerve are situated in the jugular ganglia (55). In cats, jugular ganglia neurones account for ~15% of the axons in the cervical vagus nerve, but their peripheral terminations have not been determined (21). In mice as in guinea pigs, jugular C-fiber subtypes can be differentiated from nodose C fibers by their embryologic origin, their insensitivity to P2X receptor agonists, and their expression of neuropeptides (36, 54, 71). Now that suitable methods for studying cough in mice have been described (75), it will be of interest to see whether jugular or nodose C-fiber-selective stimuli are responsible for cough in this species.

Jugular C fibers are found throughout the airways including the larynx, trachea, and bronchi. The larynx and the conducting airways are the first to encounter inhaled tussive stimuli and are a primary site of deposition of aerosolized substances. The inhibitory nodose C fibers preferentially terminate in the peripheral lung and may be even more associated with the pulmonary vasculature than the intrapulmonary airways. Based on these patterns of peripheral termination, inhaled C-fiber stimulants would thus first and primarily activate jugular C fibers. Perhaps consistent with the latter assertion, when given intravenously, capsaicin typically fails to evoke coughing in either awake humans or awake nonhuman primates but does initiate respiratory reflexes and sensations (19, 74). In the present study, we observed that the nonselective C-fiber stimulant bradykinin, which evokes or enhances cough responsiveness when applied topically to the large airways (28, 47), profoundly inhibits evoked coughing when administered intravenously in anesthetized guinea pigs.

Implicit in the observation that C-fiber subtypes have opposing actions on cough is that they also have unique neurochemistry and/or distinct sites of termination in the nucleus tractus solitarius (nTS). As for distinct neurochemistry, nearly all airway and lung jugular C fibers express neuropeptides in guinea pigs, with only a minority of nodose C fibers labeled immunohistochemically for substance P and/or calcitonin gene-related peptide (40, 45, 48, 60, 70). The neuropeptide-containing C fibers innervating murine (54, 71) and rat (10, 49) airways also have their cell bodies in the jugular ganglia. Regarding distinct nTS termination sites, most functional studies have focused on the nodose-type C fibers, or more specifically, those intrapulmonary C fibers activated by intravenous administration of 5-HT₃ receptor agonists (7, 39, 57). These C fibers terminate centrally in the commissural subnucleus of nTS. Central termination sites of airway jugular C fibers were until recently poorly defined but recent studies implicate the paratrigeminal nucleus and not nTS as a site of their termination (20, 50).

The functional studies described in the present study and the published evidence for divergent central nervous system terminations and distinct neurochemistry are supportive of the hypothesis that nodose and jugular C fibers may have opposing effects on cough. However, contradictory evidence for central divergence must be acknowledged. Substance P can be colocalized with glutamate in some nTS nerve terminals (61), and as discussed above, only jugular C fibers reliably express these neuropeptides in otherwise healthy animals. Neurokinin-containing jugular afferents innervating the airways can be anterogradely labeled from the nTS in guinea pigs (45). Functional responses evoked by capsaicin microinjection into nTS (microinjection locations confirmed at autopsy), including enhanced cough responsiveness, are effectively abolished by prior neurokinin receptor antagonist nTS microinjections (46, 47). This suggests that neurochemical characterization or more likely, central nervous system terminal mapping, of C-fiber subtypes may be incomplete or incorrect. Regarding neurochemistry, Morris et al. (51) reported that contrary to widely held beliefs regarding cotransmission of glutamate and neuropeptides, they found little evidence that neuropeptide-containing sensory nerve terminals utilize glutamate as a cotransmitter. However, when studying either C-fiber-selective stimuli in conscious animals (62) or cough receptor-dependent cough in anesthetized animals (13), we see a prominent role for NMDA type glutamate receptors in the initiation of cough.

Precisely how nodose C-fiber activation inhibits evoked coughing is unclear. There is, however, precedence for cough suppression during active afferent nerve stimulation, including the activation of arterial baroreceptors (59), pulmonary stretch receptors (13), and TRPM8-expressing nasal afferents (58). The latter two interventions can profoundly reduce respiratory rate in anesthetized animals and so perhaps inhibitory effects on respiratory network regulation alter cough sensitivity (6). If so, however, it seems surprising that in anesthetized guinea pigs, activating the jugular C-fiber subtypes promoting cough tends to acutely decrease respiratory rate whereas selectively activating the nodose C fibers inhibits cough and produces a coincident tachypnea (present study; Refs, 12, 14). We also found that the rapidly adapting receptor-selective stimulant histamine induces tachypnea in anesthetized guinea pigs and yet had no effect on cough (but enhances coughing in anesthetized dogs and awake guinea pigs; Refs. 31, 34). The extent to which nasal TRPM8 stimulation and nodose C-fiber stimulation suppress evoked cough is comparable (present study; Ref. 58) and yet they had directly opposite but comparably profound effects on breathing pattern. As shown here, we found that intravenous and inhaled adenosine and 2-methyl-5-HT produced comparable inhibition of cough but very different effects on breathing (with the differential effects on respiration likely attributable to a different intensity of afferent activation). We have also previously reported no correlation between cough sensitivity or cough number and basal respiratory rate in anesthetized guinea pigs (11). Together, the data imply a surprising and almost complete dissociation between cough sensitivity and respiratory drive.

Perspectives and Significance

A role for jugular C fibers in cough is consistent with the prominent role of C fibers and the large airways in cough as well as the termination patterns of jugular C fibers in the epithelium of these airways. We speculate that nodose C fibers are preferentially associated with the pulmonary vasculature, and we further speculate that their activation initiates the tachypnea/ventilatory compensation in response to exogenously administered adenosine or 5-HT or endogenous formation or release of adenosine, ATP and serotonin during pulmonary embolism and/or pulmonary vascular congestion (17). However, the active inhibition of cough induced by nodose C-fiber activation is difficult to explain (present study; Refs. 64, 65). Comparable cough suppression during counterirritant stimulation has been described in studies of TRPM8 activation in the nasal mucosa (58), and it is noteworthy that the profound cough suppression associated with P2X2/3 receptor blockade is invariably accompanied by a counterirritant-like effect on taste transduction (1). As to how the data relate to cough in disease, any condition or stimulus that favors the activation jugular C fibers over nodose C fibers might precipitate an excessive cough response. Discovering a jugular C-fiber-selective stimulant or a nodose C-fiber-selective inhibitor would help further address this hypothesis. Given the evidence that pulmonary C-fiber activation might initiate the undesirable sensation of dyspnea (9), however, it is difficult to imagine a therapeutic strategy for cough in which pulmonary/nodose C fibers are selectively activated. Finally, we would also emphasize the point that the observations summarized here and in the studies by Tatar and colleagues (64, 65) were all generated in otherwise healthy animals. Pathophysiological processes (pneumonia, altitude sickness, pulmonary edema, pulmonary embolism) associated with the more peripheral airways where nodose C fibers predominate are often associated with cough. These diseases may induce plasticity in central and peripheral encoding mechanisms for cough and may thus fundamentally change the interactions between afferent nerve subtypes and the reflexes associated with their activation.

GRANTS

The research was funded by National Heart, Lung, and Blood Institute Grant HL-083192.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

Y.-L.C., N.M., and B.J.C. conceived and designed research; Y.-L.C. and N.M. performed experiments; Y.-L.C., N.M., and B.J.C. analyzed data; Y.-L.C., N.M., and B.J.C. interpreted results of experiments; Y.-L.C., N.M., and B.J.C. prepared figures; Y.-L.C. and B.J.C. drafted manuscript; Y.-L.C. and B.J.C. edited and revised manuscript; Y.-L.C., N.M., and B.J.C. approved final version of manuscript.