Role of Reactive Oxygen Species and TRP channels in the Cough Reflex

Abstract

The cough reflex is evoked by noxious stimuli in the airways. Although this reflex is essential for health, it can be triggered chronically in inflammatory and infectious airway disease. Neuronal transient receptor potential (TRP) channels such as ankyrin 1 (TRPA1) and vanilloid 1 (TRPV1) are polymodal receptors expressed on airway nociceptive afferent nerves. Reactive oxygen species (ROS) and other reactive compounds are associated with inflammation, from either NADPH oxidase or mitochondria. These reactive compounds cause activation and hyperexcitability of nociceptive afferents innervating the airways, and evidence suggests key contributions of TRPA1 and TRPV1.

Graphical abstract

Introduction

The cough reflex serves to clear the airways of obstruction, thus protecting the airways during breathing. The cough reflex is triggered by stimulation of afferent (sensory) nerve terminals within the airways, resulting in the central co-ordination of motor fibers innervating the glottis, chest wall and diaphragm. Appropriate activation of the cough reflex is protective, but inappropriate and persistent activation of cough is a serious health burden. Cough is the most common reason for seeking medical attention, with it accounting for ~40% of visits to primary health care physicians [1]. Present treatments for cough are largely ineffective [2–5]. Cough is clinical associated with infectious and inflammatory diseases [2, 4, 6, 7], which likely are the source of neuroactive agents that cause excessive/inappropriate activation of airway afferent nerves. Recent evidence suggests that oxidative stress can evoke profound effects on airway afferent nerves, particularly through the modulation of neuronal transient receptor potential (TRP) channels. Although mammals share similarities in their respiratory physiology and airway afferent pharmacology, the cough reflex is only reliably present in guinea pigs and larger mammals [8–10]. Rats and mice do not possess a recognizable cough reflex, although these mammals have proved useful in elucidating aspects of airway afferent neurophysiology.

Sensory afferent nerves involved in cough

The airways (larynx, trachea, bronchi, bronchioles and alveoli) are densely innervated by sensory/afferent fibers, whose cell bodies reside in the vagal ganglia (nodose and jugular) [11–13]. These nerves synapse with brainstem neurons in the medulla, and modulate breathing rhythms, autonomic function and defensive airway reflexes [14–17]. Vagal airway afferents are heterogeneous with respect to embryological source, anatomical distribution, structure, myelination, protein expression, connectivity and function. Most vagal neurons reside in the vagal nodose ganglia and are embryologically derived from the placodes; the remaining neurons reside in the vagal jugular ganglia and are derived from the neural crest [18, 19]. Nodose and jugular neurons are under differential neurotrophic control [20], which impacts their gene expression and function [21–23].

The vast majority of vagal airway afferents are unmyelinated C fibers that innervate the epithelial layer [10, 13, 24, 25]. These nerves are sensitive to a wide-range of noxious stimuli (e.g. heat, cold, non-isotonicity, low pH (<6), irritants and inflammatory mediators) and are often termed as ‘nociceptors’ [22, 26–30]. Airway nociceptive C fibers are synonymous with C fibers projected from the dorsal root ganglia (DRG) that innervate the skin and viscera and initiate painful sensations when activated by noxious stimuli. Nevertheless, activation of airway C fibers does not lead to pain; rather it can evoke dyspnea, cough, apnea, tachypnea, bronchospasm and bradycardia [24]. Airway C fibers are projected from both nodose and jugular ganglia [21–23]. Airway C fibers overwhelmingly express the capsaicin-sensitive TRP vanilloid 1 (TRPV1) channel [30–32], the cinnamaldehyde-sensitive TRP ankyrin 1 (TRPA1) channel (Fig. 1) [33, 34] and metabotropic receptors for bradykinin [22, 35] and trypsin [36]. Few airway vagal C fibers express the menthol-sensitive TRP melastatin 8 (TRPM8) channel [33], although its expression is more common in the larynx and nasal airways [37, 38]. Nodose C fibers also express P2X2/3 channels, whereas jugular C fibers express peptide neurotransmitters such as substance P [21–23].

Figure 1.

Cinnamaldehyde, the TRPA1 agonist, activates airway C fibers and causes cough. A, Dose-response curve of cinnamaldehyde and allyl isothiocyanate on mouse TRPA1-expressing CHO cells using FLIPR. Each datapoint represents an average of four to eight independent readings. Taken from [76]. B, Representative action potential discharge in mouse bronchopulmonary C fiber (recorded ex vivo) in response to cinnamaldehyde (CA) during perfusion with vehicle and after 15 min pretreatment with the TRP channel blocker ruthenium red (30 μM). Taken from [33]. C and D, Mean ± SEM coughs evoked by aerosolized cinnamaldehyde-induced cough in guinea pigs. C, Dose-response relationship. D, Effect of aerosolized TRPA1 inhibitor HC-030031 (HC, 0.3 mM) or TRPV1 inhibitor capsazepine (CPZ, 10 μM) on cough induced by cinnamaldehyde (10 mM) in guinea pig. *P < 0.05, significant difference versus vehicle (VEH, 5% ethanol and 3.5% Tween-80 in isotonic saline). # P < 0.05, significant difference to cinnamaldehyde without inhibitor. Taken from [103].

Stimulation of jugular C fibers causes cough in conscious animals, but not under ketamine anesthesia [39–44]. The most common tussive stimuli studied are capsaicin, citric acid and bradykinin. Capsaicin and extracellular acidification are direct stimulators of TRPV1 [32, 45] and activation of bradykinin B2 receptors has been shown to gate TRPV1 channels via second messenger systems [35, 46, 47], thus inhibition of TRPV1 reduces cough by these agents. In humans, inhalation of C fiber stimulants evokes cough and ‘urge-to-cough’ sensations [48, 49]. However, selective stimulation of nodose C fibers fails to evoke cough [44]. The mechanisms underlying this disparity are unknown, although recent evidence suggests that jugular and nodose C fibers modulate distinct brainstem networks [50, 51].

The airways are also innervated by myelinated afferents that are typically sensitive to mechanical forces but insensitive to noxious stimuli. In the lungs, heavily myelinated afferents (Aβ fibers) detect mechanical changes induced by breathing and feedback to brainstem circuits that control breathing [52, 53]. In the larynx, trachea and main bronchi, there is a subtype of partially myelinated nodose afferents (Aδ fibers) that innervate the smooth muscle layer and that are exquisitely sensitive to punctate mechanical force, low pH (<6.5) and hypotonicity [54, 55]. Activation of these nodose Aδ fibers evokes cough [39, 56], which is not abolished by ketamine anesthesia. Nodose Aδ fibers do not express TRPV1 or TRPA1 and are not sensitive to temperature, irritants and inflammatory mediators [54, 55, 57]. The receptor responsible for acid sensitivity in nodose Aδ fibers is unknown but is possibly a member of the Acid Sensing Ion Channel (ASIC) family.

Thus cough can be evoked by activation of two distinct vagal afferent groups innervating the airways: jugular C fibers and nodose Aδ fibers. Based upon their activation profile it is likely that Aδ fiber cough is evoked by aspiration and C fiber cough is evoked by endogenous inflammation/infection and by inhalation of irritants. Nevertheless, functional evidence suggests that these distinct pathways converge within the brainstem. Activation of airway C fibers augments Aδ fiber-evoked cough [58, 59]. Furthermore there is evidence that activation of TRPM8-expressing afferents innervating the larger airways can reduce the cough reflex [37, 38].

Sensitivity of airway afferent nerves to ROS

There are no published reports of the effect of H2O2 (or other ROS) on the cough reflex in vivo. Nevertheless, inhalation of H2O2 caused canonical reflex changes in respiratory rate in mice [60] and rats [61] consistent with reports that H2O2 activates airway nociceptive C fibers [62–64]. Furthermore, H2O2 also augmented changes in respiratory rate evoked by other noxious stimuli [65], consistent with data demonstrating that H2O2 causes hyperexcitability in airway nociceptive C fibers [64, 65]. Overall, the available data suggests that ROS have the potential to profoundly increase the activity of airway C fibers associated with cough.

ROS have the capacity to evoke these acute effects via the direct and indirect modulation of a wide range of cellular components, including ligand-gated ion channels, voltage-gated ion channels, organelle ion channels, transporters, enzymes, signaling molecules and lipids [66–73]. As such, H2O2-induced activation and hyperexcitability are likely complex interactions of multiple pathways, most of which have not been studied in airway nociceptive neurons. Some of these ROS-mediated effects may oppose others. For example oxidation of KCNQ channels, which are widely expressed in vagal sensory neurons [74], causes increased ‘M-currents’, which would be expected to decrease neuronal excitability [70]. Here, I shall focus on those mechanisms that have been directly studied in airway nociceptive nerves.

Transient Receptor Potential Ankyrin 1 (TRPA1)

In the airways, TRPA1 is a non-selective cation channel that is expressed on the majority of nociceptive C fibers [33, 34]. Activation of TRPA1 leads to neuronal depolarization and action potential discharge in the afferent fiber. TRPA1 is a polymodal channel that is directly stimulated by a wide range of stimuli: including noxious cold, cysteine-modifying reactive electrophiles, intracellular Ca2+, 2-APB, thymol, menthol and d9-tetrahydrocannabinol [75–80].

The two major sources of cysteine-modifying reactive electrophiles for airway afferent TRPA1 channels are pollutants and oxidative stress [81]. TRPA1, expressed in either heterologous systems or in nociceptive neurons, is activated by pollutants such as acrolein, ozone, diisocyanates, isothiocyanates, formaldehyde and crotonaldehyde [77, 82–86]. Similarly, TRPA1 is activated by ROS such as superoxide and H2O2 (Fig. 2), by downstream products of lipid peroxidation such as 4-hydroxynonenal, 4-oxononenal, 9-nitrooleate and by dehydrated prostanoids [34, 60, 87–93]. In fact, all cysteine-modifying agents that can access the cytosol can activate TRPA1, suggesting intracellular targets on the channel [79]. Point mutations of select Cys residues in TRPA1 have demonstrated reduced electrophile-induced activation, while retaining functional responses to non-electrophilic activators. However, there is some disparity with C414 and C421 being important in mouse TRPA1 responses [79] and C621, C641 and C665 being important for human TRPA1 responses [78, 91]. Chemically, unsaturated α,β carbonyl compounds such as acrolein, 4-hydroxynonenal, 15d-Δ12 prostaglandin J2 and cinnamaldehyde, cause irreversible adduction of Cys. ROS-mediated TRPA1 activation is reversible [90] and is reported to coincide with the formation of disulfide bonds within TRPA1 [94].

Figure 2.

H2O2 activates nociceptive nerves via TRPA1. A, Requirement of covalent agonist acceptor sites for TRPA1 activation by H₂O₂. [Ca²⁺]_ichanges were compared between HEK293t cells expressing human TRPA1 wildtype channels (black lines) and cells expressing TRPA1 channels with mutated interaction sites (C619, C639, C663, and K708; denoted 3CK; grey lines). Values denote percent maximal response to carvacrol, nonreactive TRPA1 agonist (n = 60 cells/trace). B, Activation of Ca²⁺influx by H2O2 into DRG neurons plotted against time. Average [Ca²⁺]_i concentration of neurons activated by application of H₂O₂ followed by mustard oil (a.k.a. allyl isothiocyanate), capsaicin, and 65 mM KCl. Thick and thin lines denote mean and ± SEM, respectively. N = 189 Trpa1^+/+ neurons (black line) and 146 Trpa1^−/− neurons (grey lines). Taken from [60]. C, Representative action potential discharge in rat bronchopulmonary C fiber (recorded in vivo) to intravenous capsaicin (1 μg/kg, 0.1 ml, left), and aerosolized H₂O₂ (0.4%, right). Also shown: tracheal pressure (Ptr), and arterial blood pressure (ABP). D, Effect of TRPV1 inhibitor capsazepine (CPZ, 3 mg/kg), TRPA1 inhibitor HC-030031 (3 mg/kg) and P2X inhibitor iso-PPADS (15mg/kg) on action potential discharge by bronchopulmonary C fibers to 0.4% H2O2. Taken from [62].

TRPA1 activation has also been demonstrated following the activation of other neuronal receptors. In particular, stimulation of Gq coupled receptors such as Bradykinin B2, protease-activated receptor (PAR) 2, the bile receptor TGR5 and two members of the Mas-related G protein-coupled receptor family, MrgprA3 and MrgprC11, causes activation of TRPA1 in nociceptive neurons and heterologous systems [76, 82, 95–99]. In most cases, TRPA1 is activated downstream of phospholipase C activation; and diacylglycerol, cytosolic Ca2+, PKC and the hydrolysis of PIP2 have all been implicated. Recent studies have identified TRPA1 as a downstream target of toll-like receptor signaling [100–102]. TLR7 was shown to interact directly with TRPA1 in excised patch recordings [100], whereas LPS was shown to directly stimulate TRPA1 independently of TLR4 [101].

Airway C fibers are activated by cinnamaldehyde, allyl isothiocyanate (AITC), ozone, 4-oxononenal and 9-nitrooleate [33, 34, 57, 85, 93], and these responses are abolished by genetic knockout or pharmacological inhibition of TRPA1. Inhalation of TRPA1 agonists causes cough in conscious guinea pigs and humans [57, 103, 104]. Furthermore bradykinin evokes cough in conscious guinea pigs [39, 105], which is inhibited by TRPA1 inhibitors [97]. The role of TRPA1 in H2O2-mediated nociceptor activation is concentration dependent: at concentrations lower than 10mM, H2O2 responses were co-localized to TRPA1-expressing neurons [60, 64, 89, 90] and were abolished by TRPA1 knockout [60, 90]; at higher concentrations (up to 120mM), inhibition of TRPA1 has only a partial inhibitory effect [62], suggesting the involvement of other pathways. Thus TRPA1 is a likely mechanism involved in cough associated with inflammation, oxidative stress and air pollution.

Other neuronal proteins

Although TRPA1 is the main mechanism through which ROS and downstream products of lipid peroxidation activate airway C fibers, both TRPV1 and P2X receptors have also been implicated. In heterologous systems, TRPV1 is activated by some electrophilic cysteine-modifiers including allicin and 4-oxononenal [34, 106, 107], albeit at concentrations higher than those capable of TRPA1 activation. Knockout of TRPV1 partially reduced nocifensive behaviors evoked by allicin [107]. However, no evidence was found for a role of TRPV1 in the 4-oxononenal-mediated activation of airway C fibers [34]. Oxidation of select TRPV1 cysteines can cause channel activation [108, 109], and action potential discharge in airway C fibers evoked by high concentrations of H2O2 were reduced by capsazepine [62, 63], a TRPV1 inhibitor. Thus TRPV1 is a potential mediator of ROS-stimulation of cough-associated C fibers.

Inhibition of P2X channels and the scavenging of ATP have been shown to reduce action potential discharge in airway C fibers evoked by high concentrations of H2O2 [62, 63]. Although H2O2 may directly modulate P2X channel function [110], it is likely that such high concentrations of H2O2 cause cellular damage leading to the release of cytosolic ATP [62]. Regardless of the mechanism involved, it is unlikely that P2X channels contribute significantly to oxidative stress-mediated cough. Only C fibers from the nodose ganglion express the functional heteromeric channel P2X2/3 [21, 23], and stimulation of nodose C fibers fails to evoke cough [44]. Whereas the cough-associated jugular C fibers only express the rapidly inactivating P2X3 homomeric channel [21, 23], whose stimulation by ATP analogs is unable to evoke significant nerve discharge [22].

In addition to causing C fiber activation, H2O2 also causes hyperexcitability in airway nociceptive C fibers [64, 65]. Scavenging of ROS prevented this response. The role of TRPV1 and TRPA1 in oxidant-induced hyperexcitability of airway nociceptors is unclear: in one study inhibition of either TRPV1 or TRPA1 partially diminished the H2O2-induced augmentation of α,β methylene ATP (P2X2/3 selective agonist)-induced apneic reflexes [65]; however we found that oxidant-induced hyperexcitability of airway nociceptors did not correlate with TRPA1 expression nor was it reduced by either inhibition or knockout of TRPV1 [64]. Instead, we found that oxidant-induced airway C fiber hyperexcitability was prevented by inhibition of protein kinase C using BIM I (but not by the inactive analog BIM V) [64]. Activation of PKC causes its translocation to the plasma membrane, and consistent with this, H2O2 caused acute translocation of PKC to the plasma membrane in dissociated vagal neurons. The mechanism underlying PKC-mediated oxidant-induced nociceptor hyperexcitability is presently unclear, although direct stimulation of PKC using phorbel-12-myristate-13-acetate has been shown repeatedly to induce temporary nociceptor hyperexcitability [64, 111–114], likely via its phosphorylation of voltage-gated Na+ channels, which increases their voltage-sensitivity and currents. Nociceptive neurons express multiple PKC isoforms [111], many of which are known to be activated by ROS [72, 115, 116].

Sources of neuronal ROS

ROS are produced during inflammation, particularly downstream of NADPH oxidase activation in infiltrating immune cells [117, 118]. ROS may evoke responses through direct interactions with neuronal proteins or indirectly via electrophiles such as acrolein, 4-hydroxynonenal and 4-oxononenal downstream of lipid peroxidation. However, there is evidence that ROS can be produced within nociceptive neurons. NADPH oxidase subunits have been demonstrated in nociceptive neurons [119–121], with evidence suggesting that their activation contributes to neuronal excitability.

The mitochondrial electron transport chain (ETC) is another potent source of ROS [122], particularly during the early stages of apoptotic cascades. However it is becoming increasingly clear that mitochondrial ROS play critical roles in physiological signaling downstream of multiple pathways [123], including hypoxia [124, 125], TNFα [126], neurotrophins via p75NTR [127], Toll-like Receptors [128, 129] and TGFβ [130]. Importantly, afferent terminals innervating the peripheral tissue, including the airways, are densely packed with mitochondria [131–133]. Thus both mitochondria and afferent receptors are co-localized. Antimycin, an inhibitor of the ETC complex III, causes significant mitochondrial ROS production [122]. Consistent with the hypothesis that mitochondrial ROS could modulate airway nociceptor activity, antimycin caused both activation and hyperexcitability of airway C fibers [64, 134]. Inhibition of complex III will also depolarize the mitochondrial membrane potential, thus disrupting ATP production. Selective inhibition of mitochondrial ATP production using oligomycin (ATP synthase inhibitor) failed to alter nociceptor activity, suggesting that mitochondrial ATP levels were not acutely critical to the antimycin-induced responses. Antimycin-induced nociceptor activation was substantially reduced by knockout or selective inhibition of TRPA1, although TRPV1 was also implicated to a lesser extent [134]. In addition, antimycin activated heterologously expressed TRPA1 and this was inhibited by a combination of tempol and MnTMPyP, indicating the role of ROS. Antimycin-induced hyperexcitability was only observed in nociceptive afferent fibers innervating the airways [64]. Despite its correlation with TRPV1 expression, Antimycin-induced hyperexcitability was not reduced by knockout or selective inhibition of TRPV1. Instead, mitochondrial ROS caused airway C fiber hyperexcitability via the activation (and translocation) of PKC [64].

Interactions of TRPA1 and TRPV1

TRPA1 and TRPV1 are co-expressed in many nociceptive neurons [33, 34, 77, 82] including jugular C fibers innervating the airways [57]. Numerous studies have investigated the potential interaction between the two channels. Co-stimulation of TRPA1 and TRPV1 produced synergistic increases in C fiber action potential firing and synergistic increases in apneic responses [135]. Such synergy was recapitulated in whole cell patch clamp studies of dissociated vagal nociceptors and was absent in Ca2+-free conditions [136]. This synergy was specific to TRPA1/TRPV1 interactions, and was not observed with simultaneous treatments with activators of P2X2/3 and 5HT3 channels, suggesting specific interactions (physical or functional) between TRPA1 and TRPV1. Pretreatment with TRPA1 agonists also increases subsequent TRPV1 responses in nociceptive neurons and heterologous systems via Ca2+ and PKA signaling [137]. There is also evidence in DRG nociceptors of a complex physical interaction between TRPA1 and TRPV1 subunits under the regulation of Tmem100 [138], although whether this happens in airway C fibers is unknown.

Potential role of ROS in chronic plasticity

Previous studies have highlighted the possibility that airway sensory nerve protein expression is plastic and can be chronically modulated by pathways associated with inflammation and infection [139–144]. These changes in ion channel and neuropeptide expression are hypothesized to induce long lasting changes in the function of airway afferents associated with cough [10, 145]. A direct role of ROS in chronic plasticity has not been studied. Nevertheless, studies in other neuronal subtypes suggest that ROS have the potential to initiate long-term changes in transcription/translation. For example, ROS are stimulators of nuclear factor erythroid 2-related factor 2 (Nrf2) [146], hypoxia-inducible factor 1α (HIF1α) [125, 147], myocyte enhancer factor 2 (MEF2) [148], and neurotrophin receptors [127, 149]. Alternatively, oxidative stress can activate NFAT and ERK1/2 indirectly via modulation of mitochondrial Ca²⁺ handling [150, 151]. How ROS impacts signaling of these transcription factors in cough afferents is yet to be studied.

Open questions and challenges

TRPA1 and TRPV1 are expressed on airway C fibers and their activation leads to cough. Furthermore, there is sufficient evidence to suggest that inflammation in the airways is associated with oxidative stress, and that ROS and associated products can activate TRPA1, TRPV1 and other neuronal targets. However there are substantial gaps in our knowledge of the role of TRP channels and oxidative stress in chronic cough.

Firstly, most in vivo animal studies have thus far tested pharmacological treatments on cough evoked by acute treatment with jugular C fiber or nodose Aδ fiber stimulants. For example both TRPA1 and TRPV1 have been implicated in bradykinin-evoked cough in conscious guinea pigs. However, it is likely that these models do not adequately resemble clinical presentation of cough, where coughs are spontaneously evoked during or following airway inflammation (allergic or viral or a combination). As such we have little evidence to suggest that TRP channels actually contribute to aberrant cough signaling. A role for nociceptive afferents and TRPA1 itself has been shown in the airway hyperreactivity associated with ovalbumin-induced allergic airway disease [152–154], although the link between aberrant afferent signaling and aberrant bronchial resistance, presumably dependent on parasympathetic reflexes, has not been clarified.

Secondly, the effect of chronic inflammation in clinical populations on airway afferent function is not well understood. Inflammation is associated with increased neurotrophin signaling [155, 156], which can alter expression of neuropeptides and receptor channels [20, 139], thus potentially altering afferent functionality. The effect of neurotrophin signaling on afferent mitochondrial function and oxidative stress is unknown. The mechanisms underlying subacute/chronic changes in afferent excitability in human airway disease [157, 158] have not been identified.

Thirdly, there is little definitive proof that afferent terminals undergo oxidative stress during inflammatory/infectious disease states. Partly this due to the technical challenges associated with measuring ROS in micrometer-wide structures within the complex three-dimensional structure of the airways. Furthermore, the therapeutic use of antioxidants have shown little clinical effectiveness in the treatment of chronic cough [159, 160]. It is possible that glutathione mimetics such as n-acetylcysteine could fail to block TRPA1 activation by electrophiles due to kinetic considerations. Indeed, micromolar electrophiles have been consistently shown to activate TRPA1 despite the presence of millimolar glutathione within the cytosol.

These are some of the critical gaps of our knowledge. It is perhaps fair to say that oxidative stress and TRP signaling can evoke cough, but whether they are critical to the clinical presentation of cough is yet to be determined.

Figure 3.

Antimycin A, mitochondrial complex III inhibitor, causes activation and hyperexcitability in bronchopulmonary C fibers. A, representative trace showing action potential discharge to antimycin A (20 μM) in an individual TRPA1-expressing mouse bronchopulmonary C fiber. B, Mean ± SEM peak discharge of TRPA1-expressing bronchopulmonary C fibers to antimycin A with and without pretreatment with TRPA1 inhibitor HC-030031 (30 μM). * Significant difference (p<0.05). Adapted from [134]. C, Representative traces of action potential discharge evoked by 10-second challenge with α,β-methylene ATP (P2X_2/3 agonist, 30 μM) in a mouse bronchopulmonary C fiber before (control) and 10 minutes after treatment with antimycin A (20 μM). D, mean ± SEM action potential discharge response to 2^nd application of α,β mATP (30 μM) normalized to response to 1^st application of α,β mATP prior to either vehicle (white bar) or antimycin A (20 μM, black bars) in mouse nociceptive bronchopulmonary C fibers. The role of TRPV1 was determined using TRPV1 antagonist iodoresiniferatoxin (IRTX, 1μM). The role of PKC was determined using PKC inhibitor BIM I (1 μM) and the inactive analog BIM V (1 μM). * Significant difference from vehicle (p<0.05). # Significant difference from antimycin A (p<0.05). Adapted from [64].

Highlights.

• Cough is triggered by the activation of specific airway afferent subtypes.

• ROS cause activation of airway C fibers via TRPA1.

• ROS increase the excitability of airway C fibers through multiple mechanisms