An overview of the sensory receptors regulating cough

Two classic types of low threshold mechanosensors have been described in the intrapulmonary airways of a number of mammalian species, namely the rapidly adapting receptors (RARs) and slowly adapting receptors (SARs) [8, 9, 16‐20]. However, when comparing only a limited number of phenotypic traits RARs and SARs may appear indistinguishable (Table 1). Thus, RARs and SARs both originate in the nodose ganglia, terminate in the intrapulmonary airways and lung parenchyma, conduct action potentials in the Aβ-range (10–20 m/s) and are sensitive to many mechanical stimuli, including changes in lung volume, airway smooth muscle constriction and airway wall oedema [9, 12, 17‐21]. Accordingly, RARs and SARs may both display activity when the lungs are inflated [9, 16‐19]. RARs and SARs are also both generally insensitive to a wide range of chemical stimuli, unless the stimulus evokes coincidental changes in airway smooth muscle tone, mucus secretion or airway wall volume [8, 17, 19].

Nevertheless, RARs and SARs can be differentiated by comparing their individual mechanical activation profiles, mechanical adaptation properties, central termination patterns and the reflexes that each precipitate (Table 1). Thus, RARs may be activated during both inflation and deflation of the lungs (including lung collapse) [9, 17]. SARs, on the other hand, display activity during tidal inspirations, peaking just prior to the initiation of expiration [9, 16]. As their names suggest, RARs display rapid adaptation (i.e., a rapid reduction in the number of action potentials) during sustained lung inflations, whereas SARs adapt slowly to this stimulus [9, 17]. It is important to note, however, that this rapid adaptation shown by RARs during sustained lung inflations is unlikely an electrophysiological property of the nerve terminal but rather relates to the nature of the stimulus. RARs typically adapt slowly to other types of mechanical stimuli, including dynamic lung inflations, bronchospasm and lung collapse [12, 19]. Finally, activation of RARs evokes tachypnea and airway smooth muscle constriction, whereas SARs are likely the primary afferent fibers involved in the Hering-Breuer reflex, which terminates inspiration and initiates expiration when the lungs are adequately inflated [16, 17]. SAR activation also inhibits cholinergic drive to the airway smooth muscle, resulting in a reduction in airway tone [8]. The different reflexes that are evoked by these afferent nerve subtypes likely reflect the distinct brainstem neurons innervated by RARs and SARs [reviewed in 22].

A third type of low threshold mechanosensor has been described in the guinea pig airways [12]. These fibers also originate from the nodose ganglia, but are primary located in the extrapulmonary airways (larynx, trachea and large bronchi) and are quite distinct to RARs and SARs (Figure 2; Table 1). Extrapulmonary low threshold mechanosensors are exquisitely sensitive to punctate mechanical stimuli (such as touch) but are insensitive to physiologically-relevant tissue stretching, changes in luminal pressure or airway smooth muscle constriction [12]. Extrapulmonary low threshold mechanosensors are also readily differentiated from their intrapulmonary counterparts by a much slower conduction velocity (~5 m/sec, Aδ-range) and a lack of sensitivity to the purinergic agonist ATP [12]. During sustained punctate mechanical stimulation, extrapulmonary mechanosensors display rapid adaptation, although again this likely reflects some property of the mechanics of the stimulus in relation to tissue surrounding the nerve terminal rather than reflecting electrophysiological adaptation [23]. Circumstantial evidence suggests that analogous fibers may be present in the extrapulmonary airways of cats, dogs and humans [2, 24‐30]. It is presently unknown whether this mechanosensor subtype is activated during normal breathing.

Chemically-sensitive airway afferent fibers are found throughout the airways and lungs and are generally quiescent in the normal airways, becoming recruited during airways inflammation or irritation. Airway chemosensors are derived from both the nodose and jugular vagal ganglia, as well as from the dorsal root ganglia [13‐15]. As described above, chemosensors are typically defined by the ability of a variety of chemicals to directly activate the nerve terminal (i.e., not secondarily to structural alterations within the tissue; Table 2). However, care needs to be taken when differentiating an airway chemosensor form other airway afferent nerve subtypes. For example, often airway chemosensors are stereotypically defined by their responsiveness to the irritant chemical capsaicin and, hence, the expression of the capsaicin receptor (TRPV1). This definition, however, is not strictly accurate, as at least some species possess capsaicin-insensitive, TRPV1-negative chemosensors [31]. Alternatively, it may be assumed that all airway chemosensors are C-fiber type nociceptors. This is also incorrect, as airway (and other visceral) chemosensors that conduct action potentials in the Aδ-fiber range have been identified (analogous to somatic Aδ-nociceptors) [13, 32, 33]. Furthermore, due to the overwhelming number of studies conducted in guinea pigs, chemically-sensitive fibers are often presumed to express tachykinins (substance P and/ or neurokinin A) (Fig 2). Guinea pigs are perhaps unique amongst mammals and express a high density of tachykinin-containing airway C-fibers, especially in their extrapulmonary airways [34‐36]. Indeed, in the airways of most mammalian species (and in the guinea pig intrapulmonary airways) the majority of C-fiber chemosensors do not express tachykinins [35, 36]. Given these reasons, airway chemosensors are sometimes thought of as high threshold mechanosensors. Within this group are fibers that are not readily excited by mechanical stimulation (bronchoconstriction, lung inflations light touch, etc), but can be activated using severe mechanical manipulations (lung hyperinflation, forceful punctate stimuli etc) and one or more chemical stimuli (capsaicin, bradykinin, adenosine etc).

RARs have long been presumed to be the primary afferent nerve fibers that evoke defensive cough in the airways [1, 4, 5, 39]. Indeed, it has been proposed that coughing can be initiated following the activation of RARs by airway smooth muscle constriction, mucous accumulation, mechanical irritation and even capsaicin and bradykinin application (due to the resulting airway obstruction) [1, 4, 17]. However, several observations argue against the role of classic RARs as the primary cough-provoking afferent fibers. For example, many stimuli that produce robust activation of RARs (e.g. thromboxane, leukotriene C₄ (LTC₄), histamine, neurokinins, methacholine) are ineffective or only modestly effective at evoking cough [17, 28, 40‐42]. Moreover, in some coughing species (e.g., guinea pigs) many RARs are spontaneously active throughout the respiratory cycle and yet cough is only induced in response to very specific stimuli [8, 12, 14, 19].

Evidence also supports a role of airway chemosensitive nerve fibers in the cough reflex. For example, stimuli that are known to activate airway chemosensors, such as capsaicin, bradykinin and citric acid, are amongst the most potent tussigenic agents in conscious animals and humans [12, 26, 38, 43, 44]. However, capsaicin and bradykinin do not evoke cough in anesthetized animals or humans, even though cough can be evoked in these same animals by mechanically probing the airway mucosa [12, 25, 27]. In fact, in anesthetized animals acute capsaicin challenge has been shown to inhibit breathing and, as a consequence, inhibit cough evoked by mechanical stimulation of the airways [12, 25, 27]. These conflicting observations have lead to suggestions that in conscious animals cough-evoked by chemosensor stimuli relies on cortical processing of the stimulus, in which the activation of a subset of airway chemosensors generate the conscious perception of airway irritation and promote the urge to cough [3]. Indeed, it is interesting that capsaicin-evoked cough can be consciously suppressed in human subjects [45]. If this hypothesis is correct, then chemosensor-mediated cough may not strictly be reflexive in nature. Rather, the perception of airway irritation may induce the conscious/ voluntary decision to cough. The true respiratory reflex response that is evoked by airway chemosensor stimulation may in fact be rapid inhibition of respiratory activity, which is observed during anesthesia and perhaps over-ridden (unless the reflex is robustly activated) by voluntary control in the conscious state.

The recent characterization of an extrapulmonary low threshold mechanosensor in the guinea pig airways (distinct to classic intrapulmonary RARs and SARs) may provide some important insights into the identity of the primary cough-provoking afferent nerve fiber. As described above, these fibers are found within the wall of the larynx, trachea and mainstem bronchi and are functionally differentiated from RARs and SARs by their sensitivity to light punctuate mechanical stimulation, but not to tissue stretch, bronchospasm, ATP and positive/ negative luminal pressures within the physiological range [12]. In addition to touch-like sensitivity, extrapulmonary mechanosensors are also activated by rapid changes in pH (e.g., as might be expected to occur following aspiration of gastric contents) [11, 12]. Mechanical irritation and changes in pH are both stimuli that readily evoke cough in conscious and anesthetized animals and humans [11‐13, 26, 46]. This sensitivity profile, their apparent ideal location for airway defense (i.e., in the large airways), the absence of this fiber subtype in species that do not cough (e.g. rats and mice) and several other anatomical and functional observations makes these extrapulmonary low threshold mechanosensors a likely candidate for the primary afferent nerve subtype that evokes reflex defensive coughing in guinea pigs. Accordingly, the term 'cough receptor' has been reintroduced to describe this guinea pig afferent nerve fiber subtype [3, 12, 47‐49].

The identification of a unique afferent fiber subtype involved in generating cough from the guinea pig airways has generated much discussion within the field of cough research. For example, although these extrapulmonary fibers are easily distinguished from classic RARs and SARs in guinea pigs, it is unclear whether analogous fibers exist in the large airways of other species. The observation that the cough reflex can be readily evoked by light touch of the larynx, trachea or mainstem bronchi but not by bronchoconstricting agents, in dogs, cats and humans, provides circumstantial evidence that similar fibers may exist in these species [2, 24‐30]. It is also presently not known whether cough is the only reflex event initiated by this fiber type, nor is it certain that other fiber types can not produce coughing under some circumstances. However, this fiber type is the only sensory nerve in the guinea pig airways that once activated can initiate cough in both conscious and anesthetized animals [12]. Nevertheless, careful experimentation is required to adequately address these issues.

The appropriateness of employing the term 'cough receptor' to describe the guinea pig extrapulmonary low threshold mechanosensor has also been questioned. Although physiologists commonly describe sensory nerve fiber types as 'receptors' (e.g., muscle stretch receptors, tension receptors, RARs and SARs etc), the term 'receptor' can equally be applied to describe a pharmacological entity (e.g. a G-protein coupled receptor or a ligand-gated ion channel). Given the latter definition, and the observation that capsaicin is one of the most tussigenic stimuli available in conscious animals and humans, it is not surprising that TRPV1 (i.e., the capsaicin receptor) has been identified as a possible 'cough receptor' in guinea pigs and humans [50]. With this approach, any protein responsible for the transduction of a mechanical or chemical stimulus into electrical activity in the sensory nerve terminal (leading to cough) is a pharmacological cough receptor, and therefore a given sensory nerve is likely to have many different cough receptors. However, by defining a protein as a cough receptor it implies that this protein is therefore involved in the cough reflex irrespective of the cell type, tissue or species in which it is expressed. Using the example of TRPV1, it is unlikely that all TRPV1-expressing cells in the airways, and (perhaps with the exception of some nasal and esophageal afferent neurons) improbable that any TRPV1-expressing cells in other tissues or organs are involved in the cough reflex. Furthermore, species such as rats and mice lack the cough reflex, despite possessing numerous TRPV1-positive and capsaicin-sensitive airway afferent nerves [31, 51, 52].

Although these issues may seem an argument of semantics, they highlight the problems associated with the lack of any standard and widely accepted nomenclature system for defining terms and concepts employed by the field. Given that a 'cough receptor' was defined in the first instance as a putative afferent nerve subtype that evokes cough (and was not obviously intended to be employed to describe a pharmacological entity) [49], it therefore seems appropriate to define the guinea pig extrapulmonary low threshold mechanosensitive afferent nerve subtype as the only cough receptor identified to date.