Mini Review: Neural Mechanisms Underlying Airway Hyperresponsiveness

Abstract

Neural changes underly hyperresponsiveness in asthma and other airway diseases. Afferent sensory nerves, nerves within the brainstem, and efferent parasympathetic nerves all contribute to airway hyperresponsiveness. Inflammation plays a critical role in these nerve changes. Chronic inflammation and pre-natal exposures lead to increased airway innervation and structural changes. Acute inflammation leads to shifts in neurotransmitter expression of afferent nerves and dysfunction of M₂ muscarinic receptors on efferent nerve endings. Eosinophils and macrophages drive these changes through release of inflammatory mediators. Novel tools, including optogenetics, two photon microscopy, and optical clearing and whole mount microscopy, allow for improved studies of the structure and function of airway nerves and airway hyperresponsiveness.

Introduction

Airway hyperresponsiveness, defined as increased bronchoconstriction (constriction of airways) in response to an inhaled agonist [1], is characteristic of asthma, and is often accompanied by other exaggerated airway responses including cough, mucus production, and increased microvascular leakage. Airway hyperresponsiveness is seen in many animal models of asthma and asthma exacerbation, including antigen sensitization and challenge, exposure to ozone or to organophosphate pesticides, virus infection, and obesity. In all these animal models airway hyperresponsiveness is mediated, in whole or part, by changes in neural control of the airways.

Neural changes underlying hyperresponsiveness could include changes in 1) afferent (sensory) nerves, 2) neurons within the central nervous system, and 3) efferent nerves, including parasympathetic cholinergic nerves and nitrergic bronchodilatory nerves. Although sympathetic nerves are present in lungs, they do not significantly affect airway tone in humans [2]. Reflex bronchoconstriction is initiated when an external chemical or mechanical stimulus depolarizes sensory nerves, which signal through the central nervous system to activate efferent parasympathetic nerves, which in turn release acetylcholine onto airway smooth muscle to cause contraction. In this review, we will discuss each of these airway nerve pathways and how they may contribute to airway hyperresponsiveness.

Afferent sensory nerves

Airway sensory nerves possessing nociceptive qualities are pseudounipolar Aδ and C fibers that travel in the vagus to supply airway epithelium, submucosa, alveoli, and smooth muscle (Figure 1A, C). Afferents from the trachea arise from cell bodies in the jugular ganglia and are derived from neural crest cells, whereas sensory nerves from the lungs and distal airways arise from cell bodies in the nodose ganglia and are derived from epibranchial placodes [3]. Sensory nerve endings are dispersed through the trachea and primary bronchi epithelial layer, with increased density of nerve endings around airway branch points, in the proximal trachea, and on the dorsal aspect of the airways [4].

Figure 1. Neural mechanisms of airway hyperresponsiveness.

(A) Optically-cleared whole mouse lung labeled with pan-neuronal marker PGP9.5 (white), showing nerves in the trachea (top), esophagus (bottom middle), and branching through the airways. (B) Parasympathetic ganglion located within the trachea. (C) Sensory nerve endings within the mouse lung surrounding an airway (black oval). (D-E) Computer model of nerves within epithelium of human bronchial biopsies from healthy patients (D) and patients with asthma (E), with nerve model in green and nerve branch points marked with a red dot (Imaris software). (F-G) Quantification of nerve length and branch points, showing increases in patients with asthma. *p < 0.05. (H-K) Eosinophils closely associate with airway nerves in asthma. (H-I) Airways of patients who died from fatal asthma, showing eosinophils (major basic protein antibody, pink) localizing around nerve bundles (PGP9.5 antibody, black) (H) and a parasympathetic ganglion (I). (J) Guinea pig airway stained with hematoxylin and eosin. Cross section of nerve bundle center, eosinophils in pink. (K) House dust mite treated mouse airway showing parasympathetic ganglia (PGP9.5, white) with associated eosinophils (GFP expressed under the eosinophil peroxidase promoter, magenta).

Figure A: Reprinted with permission of the American Thoracic Society. Copyright © 2020 American Thoracic Society. Cite: Scott, G. D., Blum, E. D., Fryer, A. D., & Jacoby, D. B. (2014). Tissue optical clearing, three-dimensional imaging, and computer morphometry in whole mouse lungs and human airways. American Journal of Respiratory Cell and Molecular Biology, 51(1), 43–55. The American Journal of Respiratory and Critical Care Medicine is an official journal of the American Thoracic Society.

Figure C: Reprinted with permission from Scott, G. D. (2012). Sensory Neuroplasticity in Asthma. Oregon Health & Science University.[95]

Figures D-G reprinted with permission from Drake, M. G., Scott, G. D., Blum, E. D., Lebold, K. M., Nie, Z., Lee, J. J., Fryer, A. D., Costello, R. W., & Jacoby, D. B. (2018). Eosinophils increase airway sensory nerve density in mice and in human asthma. Science Translational Medicine, 10(457), eaar8477. Figures H-J: Reprinted with permission from Costello, R. W., Schofield, B. H., Kephart, G. M., Gleich, G. J., Jacoby, D. B., & Fryer, A. D. (1997). Localization of eosinophils to airway nerves and effect on neuronal M2 muscarinic receptor function. American Journal of Physiology - Lung Cellular and Molecular Physiology, 273(1 17–1).

The physiological role of sensory neurons is to respond to chemical and mechanical stimuli in the lungs, such as pulmonary stretch or inhaled irritants. The sensitivity of these nerves increases in asthma models: antigen exposure in sensitized animals alters the intrinsic, electrophysiological properties of nodose ganglion neurons. Ovalbumin-exposed guinea pig nodose neurons react to antigen with increased resting membrane potential, making it easier for these nerves to fire [5].

Airway sensory nerve morphology also changes in response to long-term antigen challenge and resulting long-term inflammation. The long term effect of inflammation on increased sensory nerve density and airway hyperresponsiveness has been shown in humans and animal models of asthma [6]. Bronchoscopic biopsies of human airways from patients with severe eosinophilic asthma and non-asthmatic controls were labeled with antibodies against PGP9.5, a general marker for neurons, and whole mount tissues were optically cleared and imaged in three dimensions. Computer modeling and quantification of airway nerves revealed increased epithelial nerve length and branching in patients with moderate to severe asthma (Figure 1D-G), suggesting that sensory nerve overgrowth may be a mechanism underlying airway hyperresponsiveness. Experiments in transgenic mice that express IL-5 in airway epithelium, causing airway eosinophilia, revealed similar increases in sensory innervation.

Prenatal factors may also contribute to developmental changes favoring hyperresponsiveness. To study this [7], female transgenic mice that overexpress IL-5 were bred with wild type males, and gave birth to offspring both with and without increased airway IL-5. Genetically wild type offspring of this cross had markedly increased sensory innervation and airway hyperresponsiveness. The amniotic fluid surrounding the wild-type fetuses had high levels of IL-5, resulting in fetal eosinophilia. Fetal exposure to IL-5 permanently altered neural supply to the lung, as these wildtype offspring exhibited hyperinnervation and hyperresponsiveness into adulthood, even though neither IL-5 nor eosinophilia were present after birth. The central role of fetal eosinophils in airway hyperinnervation was seen in subsequent experiments where IL-5 transgenic mice were crossed with mice congenitally devoid of eosinophils. In these studies, IL-5 was still elevated in the aminiotic fluid but the fetuses could not develop eosinophilia, and they did not develop hyperinnervation or hyperresponsiveness as adult mice [7].

In addition to increased nerve growth and changed electrophysiological properties, increases in expression of neuropeptides, such as substance P, occur in the setting of allergic inflammation and likely also play a role in airway hyperresponsiveness. The same study that found increased airway sensory innervation in biopsies from humans with asthma also found an increase in the percentage of nerves expressing substance P [6]. Substance P containing nerve endings are located most densely in the epithelium of the ventral aspect of the airways and do not tend to cluster around airway branch points [4]. This sets them apart from the majority of sensory nerve endings and suggests they may serve a distinct role in control of airways. Peptidergic sensory nerves have long been studied in the context of airway hyperresponsiveness. Antigen challenge increases mRNA and protein expression of substance P, calcitonin gene related peptide (CGRP), and neurokinin A in sensory nerves 24 hours later [8]. Changes in neuropeptide expression are due in part to new expression in neurons that did not previously produce tachykinins, and does not merely reflect increased production in existing peptidergic neurons [9]. Furthermore, neurons in the nodose and jugular ganglia that newly express tachykinins in response to antigen challenge are larger in diameter than the peptidergic neurons in controls (>20μm), suggesting they belong to specific subset of neurons [10], specifically those expressing rapidly adapting stretch receptors [11].

Decreased breakdown of substance P is a separate, complementary mechanism leading to airway hyperresponsiveness. Neutral endopeptidase is the primary enzyme responsible for cleaving substance P in the airway. Antigen challenge [12], virus infection [13,14], cigarette smoke [15], and exposure to some toxic chemicals [16], all cause airway hyperresponsiveness to substance P by decreasing neutral endopeptidase activity in airway epithelium. Conversely, neutral endopeptidase is increased in airway epithelium of asthma patients on long term steroids, and the resulting increased breakdown of substance P may contribute to the beneficial effects of steroid treatment [17].

Subtypes of sensory nerves that contribute to airway hyperresponsiveness continue to be identified. A recently published atlas of vagal sensory neurons [18] identified a large variety of nerve subtypes within nodose and jugular ganglia using single-cell RNA sequencing and transcriptomic analysis. The resulting atlas includes the complete transcriptional profiles of eighteen transcriptionally distinct subtypes of sensory nerves within nodose ganglia, including several subtypes of pulmonary chemosensors that express a variety of receptors and may be uniquely primed to sense inflammatory signaling. This represents a significant step forward for the field: in the past, studies have used a single receptor to identify a subset of nerves that may play a role in airway hyperresponsiveness. This approach can be quite illuminating, as in the case of G protein coupled receptor MgprC11. MgprC11 is known for its role mediating itch in dermal dorsal root ganglia neurons, and was recently identified on jugular airway neurons and used to identify a novel subgroup that contributes to cholinergic bronchoconstriction and airway hyperresponsiveness [19]. However, some receptors span multiple subgroups, and may play multiple roles. Patients with asthma have increased sensitivity to TRPV1 receptor agonists [20], and TRPV1 knockdown attenuates airway hyperresponsiveness in mice [21]. Conversely, TRPA1 receptor agonists are more effective inhibitors of bronchoconstriction in two different animal models of allergic asthma [22]. Yet the study by Kupari et al. demonstrated that multiple subgroups of airway nerves express both receptors [18]. This contradiction demonstrates the pitfalls of categorizing cellular subtypes by one receptor only, and highlights the potential benefits of a well-defined transcriptional atlas and phylogeny to understand sensory nerve contributions to airway hyperresponsiveness.

Central nervous system

Neurons in the brainstem integrate signals from airway sensory neurons and deliver them to preganglionic parasympathetic neurons that control airway tone. Cell bodies for sensory nerves supplying lungs and airways are primarily in nodose and jugular ganglia (together called the vagal ganglia) at the base of the skull, with a smaller number in upper thoracic dorsal root ganglia [23]. Anterograde tracing studies have mapped projections of nodose neurons to second order neurons in the nucleus of the solitary tract (NTS) in the medulla, while jugular neurons project primarily to the medullary paratrigeminal nucleus [24,25]. Studies in ferrets suggest that sensory neurons release glutamate onto AMPA receptors to activate central neurons, and that blocking this activation prevents reflex tracheal bronchoconstriction [26].

Several studies have examined the changes to NTS neuron excitation in asthma and inflammation. Neurons of the NTS have increased firing activity after antigen exposure in rats [27] and in monkeys after extended exposure to antigen [28]. A study of secondhand smoke exposure in young guinea pigs found an increase in evoked excitatory postsynaptic currents onto second order neurons in the NTS, which was reversed by blocking NK1 receptors, implicating a role for endogenous substance P [29]. A different study performed in rhesus monkeys found changes in the expression of histamine H3 receptors in NTS neurons may also contribute to increased central nervous system hyperresponsiveness. Histamine H3 receptors modulate the release of other neurotransmitters and have wide-ranging effects on synaptic transmission, and these receptors were found to be downregulated in young rhesus monkeys exposed over months to inhaled antigens and ozone [30].

Neurons in the NTS project, directly or indirectly, onto airway-related vagal preganglionic neurons in the nucleus ambiguus, and to a lesser extent in the dorsal vagal nucleus [31]. These parasympathetic preganglionic neurons are regulated by extensive networks of neurons in the brainstem, including sympathetic regulation from the locus coeruleus [32]. Hyperresponsiveness in this region is regulated in part by expression of alpha 2A adrenergic receptors on the preganglionic neurons: antigen-challenged ferrets were found to have decreased alpha 2A receptors and decreased cholinergic output to the airways compared to controls [33]. This central nervous system hyperresponsiveness may have important clinical implications for secondhand smoke exposure, early life allergen exposure, and potentially understanding nocturnal asthma [34].

Efferent autonomic nerves

Parasympathetic nerves

Preganglionic parasympathetic nerves travel in the vagus to synapse with postganglionic parasympathetic nerve cells located in ganglia imbedded within the trachea and primary bronchi (Figure 1B). Postganglionic nerves supply smooth muscle and airway mucus glands. Parasympathetic nerves control airway tone and mediate bronchoconstriction via release of acetylcholine onto M₃ muscarinic receptors [35,36]. Acetylcholine release is normally limited by inhibitory M₂ muscarinic receptors on postganglionic parasympathetic nerves [37]. These neuronal M₂ receptors are dysfunctional in patients with asthma [38,39], and are also dysfunctional in every animal model of airway hyperresponsiveness tested, including antigen challenge, virus infection, exposure to ozone, exposure to organophosphate pesticides, and obesity [40–47]. In all these models, loss of neuronal M₂ muscarinic receptor function eliminates the negative feedback control they normally provide and significantly increases acetylcholine release and vagally mediated bronchoconstriction. Mechanisms leading to decreased M₂ function vary among these models, underscoring the heterogeneity of pathways involved in airway hyperresponsiveness.

In antigen challenge models, animals are sensitized to an antigen, such as ovalbumin or house dust mite, and subsequently challenged by inhalation with the same antigen, resulting in eosinophilic inflammation and airway hyperresponsiveness. Antigen induced hyperresponsiveness is blocked by severing the vagus nerves or administering atropine, demonstrating the central role of reflex bronchoconstriction [48–51]. M₂ muscarinic receptors are dysfunctional in antigen challenged animals [40]. Loss of M₂ function requires eosinophils [52], and eosinophil major basic protein [53], the principal pre-formed protein in eosinophil granules. Eosinophil major basic protein is an endogenous antagonist for M₂ muscarinic receptors, as demonstrated in radioligand binding studies [54]. In vivo, M₂ dysfunction is reversed acutely by agents, including heparin, that bind major basic protein [55], while antibodies against eosinophil major basic protein prevent antigen-induced M₂ receptor dysfunction and airway hyperresponsiveness in guinea pigs [53]. Thus, in antigen challenged guinea pigs, acute airway hyperresponsiveness is mediated by loss of M₂ receptor function due to the presence of an endogenous antagonist, eosinophil major basic protein.

In both antigen challenged animals and humans with fatal asthma, more eosinophils are associated with airway nerves than are found anywhere else in the airway [56] (Figure 1H-K). Eosinophils are actively recruited to airway nerves, which release eotaxin, via CCR3 receptors [57]. Eosinophil adhesion to neurons is associated with degranulation and release of eosinophil major basic protein, which binds to M₂ receptors, increasing acetylcholine release and bronchoconstriction [58].

In addition to eosinophil mediated receptor blockade, other mechanisms can also impair M₂ receptor control of acetylcholine release. Viral infection is a major cause of asthma exacerbations [59–61]. In the setting of viral infection, M₂ receptor dysfunction is due to both a direct, leukocyte-independent effect on M₂ function and an indirect, leukocyte-dependent effect [41,62]. Macrophage activation may be particularly important in virus induced M₂ receptor dysfunction [63], an effect that likely involves production of both TNF-⍺ and IL-1β, and may be due to decreased expression of M₂ receptors on parasympathetic neurons [64].

Environmental pollutants such as ozone and organophosphate pesticides are associated with asthma prevalence and contribute to exacerbations. Ozone induces airway hyperresponsiveness and M₂ dysfunction through release of eosinophil major basic protein [42,44] and also through release of reactive oxygen species from airway epithelium, and activation of the mitogen activated protein kinase (MAPK) pathway [65]. Organophosphate pesticides, including chlorpyrifos and parathion, also cause airway hyperresponsiveness by decreasing M₂ receptor function [66]. Organophosphates at exposure levels below those required to inhibit acetylcholinesterase potentiate airway hyperresponsiveness by driving lung macrophages to release TNF-⍺ [67,68], which may decrease M₂ expression similar to viral infection [63,64]. Obesity induced airway hyperresponsiveness appears to have a unique mechanism that involves an interaction between increased insulin (released in insulin resistant obesity) and M₂ muscarinic receptors [46].

In all these models, loss of M₂ receptor function and increased acetylcholine release is a common endpoint, arrived at via different inflammatory mechanisms, often involving either eosinophils or macrophages. Notably, however, mechanisms for M₂ dysfunction converge on eosinophils if animals are sensitized to an antigen before exposure to a subsequent environmental insult. In antigen sensitized animals, airway eosinophils can be activated by viral infection, ozone inhalation, or exposure to pesticides, all causing release of major basic protein and blockade of M₂ receptors [42–45,47].

In addition to these acute functional changes in the control of acetylcholine release, there is also evidence for increased cholinergic innervation in chronic inflammatory asthma. Studies of human bronchial biopsies demonstrated increased cholinergic fibers in asthmatic subjects [69], and this was associated with increased brain-derived neurotrophic factor and its principal receptor, TrkB. Likewise, airway hyperresponsiveness after antigen challenge is attenuated in mice lacking TrkB receptors [70]. Early exposure to antigen in mice (postnatal days 5–20) leads to increased innervation of smooth muscle mediated by neurotrophin-4 and TrkB signaling [71]. Thus, early or chronic exposure to antigens increases cholinergic input to smooth muscle and contributes to airway hyperresponsiveness.

Nitrergic and peptidergic bronchodilatory nerves

In human airways, bronchodilation is mediated by neuropeptides, such as vasoactive intestinal peptide (VIP), and by nitric oxide. VIP and other related peptides promote relaxation in human bronchi [72], and expression of VIP may be reduced in patients with asthma [73]. However, while blocking endogenous VIP potentiates nerve-mediated bronchoconstriction in guinea pigs [74], it does not in human airways [75]. In human airways, the principal bronchodilatory mediator is nitric oxide, which is produced by nerves, airway epithelium, and endothelial cells [76]. Nitrergic nerves arise from the myenteric plexus of the esophagus and project ventrally to the trachealis muscle [77]. Blocking nitric oxide production potentiates nerve-mediated contraction in human airways [75], and potentiates airway hyperresponsiveness in patients with mild asthma to the same degree as inhaled antigen [78]. Antigen sensitization and challenge decreases nitric oxide mediated bronchodilation [79–81]. This is due at least in part to increased arginase [82], which breaks down L-arginine, the substrate for nitric oxide synthase, and decreases nitric oxide production. Arginase isozymes are increased in patients with asthma [83]. Administration of L-arginine and other substrates for nitric oxide production after antigen challenge leads to recovery of airway relaxation in guinea pigs [84].

New methods to study airway nerves: Optogenetics and Imaging

Much has been learned using traditional methods to study airway nerve activation, either by direct electrical stimulation of airway nerves in vivo or by activating nerves in airway segments in vitro with electrical field stimulation. Varying stimulus parameters gives some degree of selectivity to which nerves are activated [85], but that selectivity is limited. Selectivity in activating neurons can be greatly enhanced by using optogenetics, a method wherein light-sensitive cation channels, known as channelrhodopsins, are incorporated into the mouse genome and expressed in specific neurons, which are then able to be activated by exogenous light. We demonstrated the use of this novel technique in the airways, by expressing channelrhodopsin 2 specifically in parasympathetic nerves using a choline acetyltransferase promoter [86]. Thirty second bursts of blue light at 20Hz caused bronchoconstriction that was enhanced by the acetylcholinesterase inhibitor physostigmine and blocked by the muscarinic antagonist atropine. The magnitude of bronchoconstriction correlated with the proportion of parasympathetic neurons expressing channelrhodopsin. Targeting specific nerve subtypes for expression of channelrhodopsin allows the study of these specific nerves in control of bronchoconstriction. Though a relatively new technique, optogenetics has already been used to identify subpopulations of sensory nerves that regulate patterns of breathing [87], and to identify a subpopulation of sensory nerves expressing TRPV1 and sphingosine-1-phosphate receptor 3 (S1PR3) that are important contributors to airway hyperresponsiveness in antigen challenged mice [88]. Currently these techniques are used in anesthetized, ventilated animals, but advances in the field of miniaturized implantable light diodes may make it possible to expand this technology and investigate airway nerve function in awake animals in the future.

Histological analysis of airway nerves complements the functional, pharmacological, and electrophysiological study of neural control of the airways. Histological work has traditionally relied on the study of thin tissue sections, which generated important data, but could not fully capture the extraordinary, complex architecture of airway nerves. Fine nerves that wind through airway smooth muscle were often lost in thin sections, which led to contradictory findings between studies. These limitations have been overcome through advances in tissue clearing - the process of rendering tissues transparent through the removal of lipids or equilibration in an immersion medium with refractive index matched to the tissue components. Today tissue clearing solutions are easy to use, inexpensive, generate excellent transparency in lung tissue and are compatible with both endogenous fluorescence and antibody-based immunolabeling [89]. The application of tissue clearing to airway whole mounts combined with optical sectioning in confocal microscopy allows for accurate analysis of full thickness mouse airways and unsectioned bronchoscopic biopsies from humans, with all nerves intact [4,86,90]. Detailed examination of anatomical structures in three dimensions has already allowed for the discovery of new nerve populations, including a subset of myelinated afferent fibers in humans that do not express the peripheral nerve marker protein gene product 9.5 (PGP9.5) [91], and for accurate quantification of nerves in disease models, where increased sensory innervation has recently been shown in humans with eosinophilic asthma [6] and with chronic cough [92].

The study of airway sensory nerves has also benefited from advances in two photon microscopy and calcium indicators, which have been combined with sophisticated ex vivo tissue preparations [93] to study subgroups of airway C-fibers in different conditions. Patil et al. [94] provides a detailed account of these experiments. Using this technique, they investigated the response of airway sensory nerves to sphingosine-1-phosphate, a signaling lipid elevated in asthmatic bronchoalveolar lavage fluid. They demonstrate that sphingosine-1-phosphate relies on S1PR3 specifically for activation of mouse vagal C-fibers, which aligns with data from single-cell RNA transcriptome studies that found S1PR3 on pulmonary chemosensitive sensory nerves [18]. In the end, a combination of molecular, functional, and imaging studies will all be necessary to understand the critical role of nerves in the lung, and how changes in nerve architecture, neurotransmitters and receptor expression all contribute to airway hyperresponsiveness.

Conclusions

Neural mechanisms underlying airway hyperresponsiveness are varied and affect sensory nerves, central nerves, parasympathetic cholinergic nerves, peptidergic nerves, and nitrergic bronchodilatory nerves in different ways and at different times during development. Some themes emerge from existing research. First, changes in sensory and parasympathetic nerve structure and function are central to airway hyperresponsiveness. Acute airway inflammation, a result of many different inhaled or infectious insults to the lungs, leads to hyperresponsiveness that is mediated through loss of neuronal M₂ receptor function and increased expression of neuropeptides. Structural changes that increase innervation are seen in models of both chronic inflammation (such as antigen challenge) and in pre and early post-natal exposures. Finally, inflammation in the form of eosinophils, macrophages, and inflammatory cytokines is a key mediator in inducing nerve changes. We continue to gain new insights into the complexities of airway nerves, including the roles that specific neurotransmitters, receptors, and neurotrophins play in shaping and adapting these systems. A more complete understanding of the mechanisms underlying nerve mediated hyperresponsiveness will reveal new therapeutic strategies and drug targets for patients with airway disease.

Highlights.

• Neural changes underly hyperresponsiveness in asthma and other airway diseases

• Afferent sensory nerves, nerves within the brainstem, and efferent parasympathetic nerves all contribute to airway hyperresponsiveness

• Inflammation is a key mediator in inducing nerve changes