Airway Innervation and Plasticity in Asthma

Abstract

Airway nerves represent a mechanistically and therapeutically important aspect that requires better highlighting in the context of diseases such as asthma. Altered structure and function (plasticity) of afferent and efferent airway innervation can contribute to airway diseases. We describe established anatomy, current understanding of how plasticity occurs, and contributions of plasticity to asthma, focusing on target-derived growth factors (neurotrophins). Perspectives toward novel treatment strategies and future research are provided.

Introduction

In all mammals, peripheral innervation serves critical roles tailored to the function of the target organ. Both the airways and lung parenchyma are densely, diffusely, and exquisitely innervated with particular targeting of airway epithelium, airway smooth muscle, glands, and vasculature. The long-established basic anatomy and broad functionality of airway innervation underlines the potential importance of nerves not only for normal airway structure/function but also disease pathophysiology that involves altered airway tone, mucus secretion, inflammation, and other critical aspects. However, the neuronal changes in and contributions to airway diseases including asthma have received far less attention than expected, largely due to the understandable focus on inflammatory aspects of the disease per se. Yet, there is increasing evidence for a number of novel mechanisms within airway nerves that contribute to multiple aspects of airway structure and function, including tone, secretion, modulation of inflammation, and fibrosis. Furthermore, the targets of innervation (epithelium, smooth muscle) produce factors that induce neuronal plasticity that in turn can influence airway structure/function. Indeed, it is likely that neuronal plasticity is a fundamental aspect of normal airway development and growth, and thus perturbations in pathways modulating plasticity contribute to airway diseases across the age spectrum. In this review, we focus on established and emerging aspects of target-derived neuronal growth factors (neurotrophins) and their roles in neuronal plasticity in the context of airway diseases from the developmental/perinatal period through adulthood, highlighting the potential for targeting these novel mechanisms in the context of therapy.

Airway Innervation

Elegant studies in multiple species have established the fundamental organization of respiratory innervation (reviewed in Refs. 24, 109). The respiratory tract is innervated via a network of afferent and efferent nerves, which travel via the vagus nerve. Airway nerves regulate many aspects of airway function, including breathing pattern and the Hering-Breuer inflation reflex, airway smooth muscle tone, mucus secretion, and cough (37, 195). Airway vagal nerve fibers can be divided into three major components: the primary afferent nerve fibers, the integrating centers in the brain, and the parasympathetic efferent nerve fibers (FIGURE 1) (177).

FIGURE 1.

Innervation of the airways

Afferent pathways that sense mechanical forces, inhaled irritants, and endogenous inflammatory mediators lead to the central nervous system via the vagus nerve (with some additional input from laryngeal and esophageal afferents). The major tracts are the nodose and jugular ganglia. Efferent parasympathetic pathways from the central nervous system (CNS) travel via the vagus to postganglionic neurons located within the airway wall that then innervate airway smooth muscle and submucosal glands to induce contraction and mucus secretion, respectively. Non-adrenergic, non-cholinergic (NANC) pathways are particularly involved in bronchodilation via vasoactive intestinal peptide (VIP) and nitric oxide (NO). Additionally, in some species, sympathetic innervation from the thoracic segments exists. Blue: efferent nerve fibers; red: afferent nerve fibers.

Airway Afferent Nerves

Afferent innervation is derived from either the nodose (inferior) or jugular (superior) vagal ganglia, with a small proportion coming from thoracic dorsal root ganglia (179). Afferent, sensory nerve fibers can be classified into various phenotypes, and there is no classification that takes all features into account. In general, afferent nerves are considered to include stretch-sensitive myelinated A fibers and unmyelinated C fibers (183). Most airway afferent nerves are unmyelinated C fibers present throughout the airways, with terminals in walls of blood vessels, on ASM, on epithelial cells, and in airway ganglia (140). A fibers conduct action potentials at a relatively high velocity and consist of rapidly adapting receptors (RARs) in the mucosal layer, which respond to the dynamic phase of inspiration, and slowly adapting receptors (SARs) in the ASM layer, which respond to maintained inflation (155, 178, 183). A fibers are also classified as mechanosensors, with activation dependent on the rate and depth of breathing. Furthermore, a group of A-delta fibers exists that is activated by punctate mechanical stimulation but not by stretch or breathing, and contributes to explosive cough (26). The slower C fibers are also classified as chemosensors and, in contrast to mechanosensors, are largely quiescent under normal conditions, being activated only following airway irritation or inflammation (123). They are also more broadly responsive to stimuli including heat or cold, changes in pH, mechanical forces, and mediators released upon tissue damage and inflammation, resulting in defensive reflexes including rapid shallow breathing, mucus secretion, bronchoconstriction, and cough (38, 183). Chemosensor activity is modulated via a variety of receptors and ion channels. Capsaicin is a classical C-fiber activator, inducing cough mediated via the ligand-gated ion channel transient receptor potential vanilloid type 1 (TRPV1) (108). In addition to TRP channels, serotonin and purinergic receptors, the latter activated by ATP, can depolarize the nerve directly. Decreased pH can also activate ligand-gated ion channels on C fibers. G-protein-coupled receptors for histamine, bradykinin, adenosine, prostanoid, opioid, cysteinyl leukotrienes, and neurokinins (NKs) are expressed, as are voltage-gated sodium and potassium channels (179).

Afferent nerves play important roles in airway physiology. They can mediate local airway reflexes, involving release of peptide neurotransmitters, including the tachykinins substance P (SP) and neurokinin A (NKA), and calcitonin gene-related peptide (CGRP), which can induce airway smooth muscle contraction directly (188). Furthermore, these peptides are transported to both peripheral terminals and central terminals to help mediate the central reflex arch via the brain stem (178). Sensory fibers terminate predominantly in the middle and caudal nucleus tractus solitarius (NTS), a sensory integration of the dorsal medulla (91). Different NTS termination regions exist for distinct afferent nerve types, including SAR, RAR, and C fibers (91). The primary excitatory neurotransmitter in the central nervous system (CNS) is glutamate. Secondary neurons or interneurons in the CNS provide input to preganglionic airway neurons via efferent nerve fibers (105).

Airway Efferent Nerves

Efferent fibers include parasympathetic, sympathetic, and non-adrenergic non-cholinergic (NANC) nerves (24). Parasympathetic nerves derive primarily from the brain stem nucleus ambiguous, with small numbers from the dorsal motor nucleus of the vagus (albeit of unknown functional significance) (24, 105). Preganglionic nerves travel to the airway wall and synapse with airway intrinsic neurons or parasympathetic ganglia located directly near the bronchial tree (177). Postganglionic parasympathetic (motor) neurons have nonmyelinating Schwann cells (168), use ACh as neurotransmitter, and innervate ASM, vascular smooth muscle, and mucus glands throughout the bronchial tree (FIGURE 1) (105). Parasympathetic cholinergic nerves are the major regulators of ASM tone (22, 25) and are tonically active to establish a basal bronchomotor tone and mucus secretion, inhibited by ganglionic blockade (24, 62, 79). Ganglionic neurotransmission by ACh is mediated via nicotinic receptors, whereas functional effects on target organs are mediated via muscarinic receptors, particularly M₃ receptors (56). Although M₂ receptors represent the majority of muscarinic receptors on ASM, their role in contraction is negligible (51), and their role in functional antagonism of β₂ receptor is debated (113). More likely, they function as prejunctional auto-inhibitory receptors on neurons, limiting ACh release in parasympathetic ganglia and nerve terminals (114). Although it was long thought that neural innervation was confined to larger airways, increasing evidence suggests small airways are innervated to a similar extent, as elegantly shown by recent imaging studies (160, 194) and confirmed by functional experiments using electrical field stimulation in precision-cut lung slices (94, 106, 159).

Sympathetic postganglionic neurons use the catecholamines norepinephrine and epinephrine (185) (FIGURE 1). However, at least in humans, sympathetic nerves do not innervate ASM directly, and functional studies suggest a rather limited role for sympathetic nerves in regulating airway caliber, which is mostly achieved via circulating epinephrine inducing bronchodilation via β₂ receptors expressed on airway smooth muscle (2, 24).

Postganglionic NANC innervation involves inhibitory (iNANC) and excitatory (eNANC) pathways (FIGURE 1). These nerves use a variety of mediators, including nitric oxide (NO), and peptides such as vasoactive intestinal peptide (VIP), SP, and NKA. Both iNANC NO and VIP induce bronchodilation and are suggested to act as a brake against cholinergic bronchoconstriction (9, 167). It is very likely that NANC originate from a distinct population of preganglionic nerves in different locations of the nucleus ambiguous or from dorsal motor nuclei of the vagus (24). However, it is unlikely they regulate distinct aspects of airway physiology; rather, the NANC system “fine tunes” activities of the cholinergic nervous system. Regulation of airway physiology is a balance between afferent, sensory input and efferent motor output. This balance can be altered by ASM contraction in itself, which can either dampen the response via SARs, particularly after strong contractions, or evoke further contraction via reflexes, in particular after modest ASM contraction induced by, for example, bradykinin, PGE₂, or ATP (24, 140, 192).

The Role of Nerves in Asthma

Asthma is a common chronic lung disease, affecting over 300 million people worldwide (55). Many clinical symptoms of asthma are related to the nervous system and can at least partially be attributed to altered activation or regulation of airway nerves (24). Asthma is associated with coughing, sneezing, and wheezing, which are neural reflexes. Moreover, bronchoconstriction and mucus secretion are under cholinergic control, contributing to dyspnea in asthma (83, 84). Reversible airway obstruction is primarily due to increased cholinergic tone, confirmed by the fact that anticholinergics are effective bronchodilators in patients with moderate and severe asthma (24, 76, 77). In addition, in half of the asthmatic patients, an early asthmatic reaction in response to allergen exposure is followed by a late asthmatic response, associated with wheeze and bronchoconstriction (129). Interestingly, this late asthmatic response is dependent on sensory nerve activation via TRPA1 (149).

Airway hyperresponsiveness (AHR) is another characteristic feature of asthma, which is defined by an exaggerated response of the airways to specific (e.g., histamine) and non-specific (e.g., cold air) stimuli resulting in bronchoconstriction (144). Although airway inflammation has long been considered one of the main drivers of AHR, several lines of evidence suggest that nerve activity and interactions between inflammatory cells and airway nerves are key aspects in development of AHR (188). AHR persists in the absence of inflammation, and novel treatment strategies targeting IgE or specific cytokines do not impact AHR in patients with asthma, suggesting that other mechanisms contribute (27, 102). Several studies suggest that, in fact, airway nerves drive AHR, whereas anticholinergics either fully or partially inhibit AHR in response to a variety of stimuli (27, 105). For example, AHR in response to histamine is inhibited by anticholinergic treatment in ovalbumin-challenged guinea pigs (156). Furthermore, vagotomy inhibits ovalbumin-induced AHR in multiple species (39, 101, 111). Interestingly, vagotomy inhibits methacholine-induced airway contraction in ovalbumin-challenged mice but not in controls, and AHR ex vivo in tracheal strips is observed in response to EFS and not to methacholine, highlighting altered neural activity in asthma (72, 111).

Muscarinic receptors are expressed not only on ASM cells but also on fibroblasts, epithelial cells, and inflammatory cells (83). Thus, beyond contractility, neurally derived ACh may promote airway inflammation and remodeling, including increased ASM mass and excessive extracellular matrix deposition, as observed in asthma (59, 83–85, 130, 187). This has been confirmed by in vivo animal studies, including studies using anticholinergic treatment and muscarinic M₃ receptor knockout animals (81–83, 131). A cholinergic anti-inflammatory pathway, mediated via nicotinic ACh receptors, has also been described (188, 198). The functional relevance of ACh-mediated effects on inflammation and remodeling has yet to be proven in asthma patients (83).

Neuronal Plasticity in Asthma

Neuronal plasticity is the ability of the nervous system to change its phenotype in response to intrinsic and extrinsic stimuli by reorganizing its structure, function, and connections (41). Such changes can outlast the initiating stimulus for days, weeks, or even years, and may thereby become permanent, in line with other structural changes. Plasticity or remodeling of airway nerves may in fact drive many of the symptoms observed in asthma, including AHR. Although this aspect of neuronal contributions has not gained much attention in airway diseases, neuronal plasticity is a well-known feature in other diseases, particularly in the CNS, and is considered a therapeutically targetable mechanism (41, 69, 75, 90, 150). Nonetheless, there is evidence for increased neuronal reactivity or plasticity in asthma. In the airways, plasticity occurs at all levels of the nervous system, including changes in afferent nerves, the integrating centers in the brain, and efferent nerves (25, 166, 178). Plasticity may be subdivided into synaptic plasticity (changes in connections between neurons) and non-synaptic plasticity with changes in neurons themselves. There is acute sensitization of airway nerves and chronic sensitization, or remodeling, of nerves (110).

Although several mechanisms likely contribute to neuronal plasticity, such factors could be simplistically divided into immune cell/inflammation-driven pathways, such as eosinophils and mast cells (185) (briefly discussed in this review), apropos to the role of inflammation in asthma, and target cell-derived pathways that induce and promote nerve growth as well as altered function, with neurotrophins (the focus of this review) being a classic aspect.

Plasticity of Afferent Nerves

Increased peptide synthesis, increased excitability, or lowering of activation threshold, and enhanced transmission are different mechanisms underlying acute changes in activity of afferent sensory nerves in asthma (184, 185). Epithelial damage exposes afferent sensory nerve endings in the subepithelial layer to the airway lumen (93). Furthermore, inflammatory mediators, including histamine, tachykinins, prostaglandins, and thromboxane A₂, stimulate sensory nerve fibers directly (179, 186). In rat and guinea pig models of asthma, ovalbumin sensitization induces acute hypersensitivity of pulmonary C fibers in response to several stimuli, including capsaicin (10, 92), further enhanced by ovalbumin challenge (200), thereby increasing afferent output. In addition, the force needed to induce action potentials in mechanosensitive A fibers decreases substantially after ovalbumin sensitization in vivo and subsequent antigen exposure in vitro (153). Increased excitability of C fibers from dogs in response to inflation or capsaicin occurs after histamine pretreatment (99). Moreover, excitability of guinea pig nodose ganglion neurons is enhanced after ovalbumin sensitization (180). Eosinophils and mast cells colocalize to airway nerves (185), and eosinophil-derived mediators such as major basic protein increase capsaicin-evoked activity of pulmonary sensory neurons (58). Thus multiple mechanisms exist for plasticity of afferent innervation in the context of asthma (FIGURE 2).

FIGURE 2.

Neuronal plasticity in asthma

On the afferent limb, inflammation (characteristic of asthma) can lead to enhanced histamine, tachykinins (TKs), thromboxane (TXA), prostaglandins (PGs), and other mediators that influence nerve function. Asthma is associated with increased density of sensory nerve ending (chronic neuronal remodeling) and an overall afferent output to the CNS. The nodose ganglia can also express increased substance P and calcitonic gene-related peptide (CGRP). The CNS itself can undergo plasticity with increased brain stem N-methyl-D-aspartate (NMDA). On the efferent limb, there is increased output to the parasympathetic ganglia, which leads to enhanced acetylcholine release and an increase in cholinergic tone. Neuronal remodeling again involves enhanced cholinergic nerve density. Dysfunction of NANC nerves leads to a cholinergic rather than nitrinergic phenotype. Overall, these changes (left) contribute to the many symptoms and pathognomonic features of asthma (right).

In humans, levels of tachykinins are increased in bronchoalveolar lavage fluid, sputum, and plasma of asthmatics (28, 126, 174). Additionally, increased SP immunoreactive nerves are observed particularly surrounding vessels and glands, with greater branching and increased length compared with non-asthmatic controls, suggesting nerve outgrowth that is likely chronic and persistent (134). This was confirmed recently in a larger, well-characterized patient population and was shown to be correlated to eosinophil number (46). Animal studies have found increased expression of SP, NKA, and CGRP specifically in nodose ganglia in response to allergen exposure (50). Interestingly, there is an increased percentage of SP, NKA, and CGRP immunoreactive neurons in the nodose ganglia, indicating that neurons start synthesizing these peptides in response to allergen exposure and potentially undergo a phenotypic switch (45, 50, 98). This phenomenon has been elegantly shown in A fibers that do not respond to noxious stimuli or tissue damage, express TRPV1 channels, or produce neuropeptides under normal conditions. However, non-nociceptive A fibers become responsive to capsaicin after chronic (but not acute) ovalbumin challenge (199). Such changes are accompanied by increased production of SP and CGRP as well as TRPV1 expression (34, 100, 120, 199).

TRPV1-expressing neurons play a critical role in AHR, as suggested by prevention of AHR by genetic silencing of these neurons (176). Interestingly, capsaicin and optogenetic activation of TRPV1-expressing neurons potentiates AHR, but only in ovalbumin-challenged animals, supporting the hypothesis that neuronal plasticity underpins AHR (176). Neuronal plasticity also contributes to the cough response, with capsaicin inducing cough at lower doses and with greater maximal responses in asthma patients compared with healthy controls (157). Interestingly, this is independent of the level of inflammation or AHR, and is more pronounced in non-atopic patients (157). Furthermore, a recent study demonstrated that patients with respiratory diseases including asthma and COPD exhibit different patterns of cough responses compared with healthy controls, likely mediated via altered expression and functionality of TRPV1 channels (8). Another member of the TRP family that plays a critical role in asthma and neuronal plasticity is the TRPA1 channel (57), which is directly activated by inflammatory mediators and ozone (124, 169, 170). Knockout or inhibition of TRPA1 channels inhibits neuropeptide release and AHR in ovalbumin-challenged mice (23). Channels essential for action potential formation, including sodium and potassium channels, may also undergo changes in the presence of inflammatory mediators, further contributing to increased excitability and thereby afferent nerve activation (110, 185).

Overall, there is increased reactivity of afferent nerves in asthma, with increased NK release, resulting in an increase in the local reflex arc and neurogenic inflammation, as well as increased action potentials to the NTS in the CNS (FIGURE 2).

Plasticity of the Central Nervous System

There is growing evidence that, like neurons in other parts of the CNS, NTS neurons also undergo plasticity (central sensitization), thus modulating the output to efferent nerves of the airways (11). Plasticity of afferent neurons may induce plasticity of NTS neurons, for example, in response to increased firing over time. In addition, inflammatory mediators may be retrogradely transported to the CNS, whereas NTS neurons also express mediators such as substance P. Substance P can modulate release of the primary neurotransmitter glutamate, thereby enhancing neurotransmission to secondary neurons (151). In addition, NMDA receptors expressed in NTS are well-known for CNS plasticity (11). NMDA receptors contribute little to synaptic transmission under normal conditions but can be activated as a consequence of increased neuropeptide or excitatory amino acid release (11, 182). Allergen exposure can alter the intrinsic excitability and activity of NTS neurons, as shown in brain slices of house dust mite-exposed monkeys (32). Furthermore, in sensitized rats, intratracheal capsaicin increases firing activity and duration in NTS neurons (165). In addition, there is evidence for central plasticity in the cough response, shown to be mediated via NK1 receptors (119). Substance P injection in the NTS augments effects of systemic capsaicin on tracheal pressure (119) and reduces the electrical threshold for cough (108). Furthermore, patients with cough hypersensitivity have increased sensitivity toward inhaled capsaicin associated with altered brain activity in the NTS area (3). In addition, changes in the hippocampus, including neurogenesis, occur in response to grass pollen exposure, suggesting that allergies may also affect cognitive functions by CNS alterations (87). Overall, the CNS, and in particular the NTS, is subject to plasticity in response to allergen exposure. It is likely many other forms of plasticity exist, but central plasticity is a thoroughly underexplored area.

Plasticity of Efferent Nerves

Efferent neurons and parasympathetic ganglia also undergo plasticity in asthma (FIGURE 2), and cholinergic output can be increased as a result. Parasympathetic ganglia are subject to acute changes in response to ovalbumin exposure, including enhanced excitability and reduced filtering capacity. Ganglia from sensitized guinea pigs exposed to ovalbumin show decreased membrane potential and increased action potential generation (74, 122) mediated via mast cell mediators and tachykinins (121, 177, 188). The resultant increase in synaptic efficacy increases the output to postganglionic parasympathetic nerves (182). Indeed, allergens potentiate postganglionic parasympathetic ACh release, resulting in enhanced bronchoconstriction (182). In addition to facilitated neurotransmission, increased cholinergic nerve density can play a role, as evidenced by our recent findings in adult mice (Kistemaker LEM, unpublished observations), and the fact that there is an increase in the pan-neuronal markers neurofilament and class III β-tubulin (TuJ1) (5). M₂ receptors and eosinophils play important roles in efferent nerve plasticity. Auto-inhibitory prejunctional M₂ receptors on parasympathetic ganglia, which limit acetylcholine release under healthy conditions, are dysfunctional in asthmatic airways (53, 171), as shown in animal models of allergen exposure, viral infection, and ozone exposure (96, 154, 171), and also in patients with asthma (56, 115). M₂-receptor dysfunction is mediated via major basic protein secreted by eosinophils recruited to airway nerves in asthma and acts as an allosteric antagonist for the M₂ receptor (40). An antibody against major basic protein prevents M₂ receptor dysfunction in guinea pigs (48). Furthermore, ACh content in the synaptic cleft is altered not only due to dysfunctional M₂ receptors but also to reduced cholinesterase activity, as found in ragweed pollen-sensitized dogs (116). Eosinophils mediate acute and chronic plasticity of cholinergic nerves in addition to M₂-receptor dysfunction. In vitro, eosinophils can inhibit apoptosis of cholinergic nerves induced by cytokines such as IL-1β, TNF-α, and IFN-γ (118), an effect mediated by major basic protein (117). In addition, eosinophils, more specifically, eosinophil peroxidase, upregulate the Ach-synthesizing enzyme ChAT and the transporter VAChT (47, 118), leading to increased ACh levels (158).

Plasticity of the NANC System

Plasticity of the NANC system is also observed in asthma, which may further contribute to increased bronchoconstriction. Allergen exposure in guinea pigs leads to isotype switching such that inhibitory iNANC nerves, which do not use ACh as a neurotransmitter, become cholinergic (136) (FIGURE 2). Moreover, VIP and NO production have been found to be reduced in asthma (24). Allergen challenge in patients with atopic asthma impaired the release of NO and downregulated eNOS (152), and no VIP-positive nerves have been seen in cases of fatal asthma (133). Furthermore, plasma VIP levels are lower in asthma patients during exacerbations compared with healthy controls (28). Separately, increased arginase expression in asthma contributes to reduced neuronal NO expression (107). However, further research is needed in this area, since there are also reports demonstrating no changes in VIP and NO levels, contrasting with functional studies showing tracheal segments of ragweed-challenged rabbits with reduced iNANC response (49) and bronchi from ovalbumin-challenged guinea pigs with increased eNANC response (71). Overall, it is very likely that plasticity of afferent nerves, together with changes in efferent nerves and the CNS, contribute to asthma.

Plasticity Early in Life

Development of Airway Innervation

The respiratory tract develops from the foregut. During embryogenesis, the lung bud territory is defined, and the foregut is divided into trachea and esophagus. As the primary pair of lung buds form, elongate, and branch in generating the entire bronchial tree (21), these lung buds already express ASM while the airway nerves start to develop (173, 191), with extensive innervation of the tracheobronchial tree and ganglia formation occurring early in gestation (pseudoglandular phase, first trimester) (164, 191). This fine fiber network originating from two large nerve trunks in close proximity to the airway wall covering the ASM layer is already functional in fetal porcine lungs and promotes ASM contraction in response to ACh or EFS to an extent comparable to that observed in later stages (14). Thus functional innervation appears to be an early aspect of lung development, but what these functions are per se in the context of embryonic lung structure and its growth is not entirely understood, particularly in humans. Denervation of whole lung explants does halt airway branching (15, 54, 88), but the contributions of specific neuronal pathways, signaling mechanisms, or the target cell types involved remain to be established.

In mouse and chick embryos, airway neurons are derived from neural crest cells (nodose sensory neurons derive from epibranchial placodes) (21, 95, 125, 173). Intrinsic neurons, whose cell bodies are within trachea and main bronchi, require distinct neurogenic signals compared with extrinsic neurons, whose cells bodies are located elsewhere. Intrinsic neurons depend on glial cell-derived neurotrophic factor (GDNF) family and neurturin signaling, whereas extrinsic neurons depend on the nerve growth factor (NGF) family, in particular brain-derived neurotrophic factor (BDNF) signaling (4, 54, 95, 148). As discussed further, these neurotrophins appear to serve critical roles in the airway throughout life.

Plasticity During Development

There is increasing evidence that exposure to allergens early in life has marked influences on airway nerves and their function later in life. Alterations in airway nerve structure might start very early during so-called “critical periods” (67). Early life exposure of monkeys to ozone affects NTS neurons. Repeated episodic ozone exposure increases neuronal excitability and output, but surprisingly decreases responsiveness to sensory nerve stimulation (31). Furthermore, exposure of infant rhesus monkeys to ozone and house dust mites results in changes in airway innervation in the epithelial region, as shown by morphological analyses (73, 97). Interestingly, in the short term, there is a decrease in PGP9.5-positive nerves in intrapulmonary airways, but repeated episodes of allergen and/or ozone exposure over months increases epithelial nerve density (73, 97). Postnatal exposure of ozone in rats results in increased sensory innervation but only after very early ozone exposure, again highlighting a critical window for such neuronal plasticity (67).

Efferent nerves may also undergo remodeling after early exposure. Exposure of mouse neonates to ovalbumin leads to increased VAChT, without affecting the level of the sensory marker CGRP (5, 137, 138). Early ovalbumin exposure enhances ASM innervation, associated with AHR in adulthood, highlighting long-term changes (5). Furthermore, the response to EFS in lung slices is increased after neonatal ovalbumin exposure, whereas M₃ antagonism prevents AHR (138). Early life exposure likely also effects the NANC systems, as suggested by decreased iNANC responses in young rabbits exposed to ragweed (36). Respiratory syncytial virus (RSV) infection early in life is also thought to contribute to neuronal remodeling and thereby to AHR and asthma later in life (35, 143). Thus exposure to allergens or environmental stress early in life leads to changes in airway innervation with long-lasting effects in the context of asthma pathophysiology. However, the instigating mechanisms within neurons or resident airway cells early in life or the factors that facilitate long-term effects of early insults remain to be determined.

Neurotrophins as Mediators of Neuronal Plasticity

Neurotrophins are factors that promote CNS neuronal growth and differentiation but may play an important part in airway neuronal plasticity (18, 145). The neurotrophin family comprises nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin 3 (NT3), and NT4. These factors act on tropomyosin-related tyrosine kinase (Trk) receptors expressed on nerve terminals, respectively, TrkA (NGF), TrkB (BDNF and NT4), and TrkC (NT3) (68). In addition, all neurotrophins can interact with the low-affinity neurotrophin receptor p75. Neurotrophins can be released from different cell types in the airways, including neurons, epithelium, ASM, lymphocytes, macrophages, and mast cells (18). Importantly, neurotrophin expression now has been shown to be increased in asthma (12, 104, 132, 189, 190).

The most widely studied neurotrophin is NGF. In the context of asthma, several studies point to a role for NGF in neuronal plasticity and AHR. This can be mediated via inflammatory cells, given that NGF inhibits eosinophil apoptosis while NGF release from eosinophils of allergic subjects is increased, along with increased expression of neurotrophin receptors (147). Treatment of ovalbumin-challenged mice with an anti-NGF antibody reduces AHR (16, 139). Silencing NGF attenuates AHR (33), whereas NGF overexpression aggravates it (19, 139). NGF overexpression increases nerve density, in particular of tachykinin-expressing sensory fibers (63), which play a role in mediating NGF effects on nerves (42). Furthermore, NGF increases substance P expression in airway nodose neurons and a phenotype switch of mechanical sensitive neurons (44, 66, 181). In vitro, NGF induces remodeling of airway parasympathetic ganglia (61) and nicotinic transmission in parasympathetic neurons (193) with enhanced dendritic growth and sprouting (61). NGF may also be involved in neuronal plasticity induced early in life (65), but there is currently much less information on this topic. Nonetheless, NGF likely plays an important role in neuronal aspects of asthma (FIGURE 3).

FIGURE 3.

Neurotrophins in asthma

The family of neurotrophins, released by nerves as well as airway resident cells, works through specific tropomyosin related kinase (Trk) receptors in mediating multiple, complementary effects on both afferent and efferent limbs of airway innervation. Nerve growth factor (NGF), working via TrkA, is enhanced by inflammatory mediators. NGF can blunt eosinophil apoptosis (thus sustaining inflammation) and, as a growth factor, promote sensory neuronal remodeling (increase nerve density). NGF also increases release of neurotransmitters/neuromodulators. Brain-derived neurotropic factor (BDNF) can be released by nerves as well as by smooth muscle, and is elevated in inflammation/asthma. BDNF, via TrkB, promotes efferent cholinergic output and sensory TRPV1 activity. Neurotrophin-4 (NT4), also via TrkB, enhances mast cell activation, whereas NT-3 via TrkC promotes a phenotypic switch of NANC nerves. Next to these effects, it is very likely that the effects of neurotrophins in the airways might be broader, but this is a largely unexplored area.

Of particular interest for asthma pathogenesis is BDNF (127). Expression of BDNF and its receptor TrkB is relatively high in human airways compared with other neurotrophins, and there are genetic associations of the BDNF gene with asthma, highlighting its relevance (70). Furthermore, increased levels of BDNF in platelets is associated with airflow limitation and AHR in asthma (104). Interestingly, anti-BDNF treatment of ovalbumin-sensitized mice prevents airway obstruction in vivo (17). Moreover, BDNF treatment induces AHR in response to EFS assessed in tracheal smooth muscle segments, whereas anti-BDNF treatment inhibits ovalbumin-induced AHR (17). Interestingly, some studies have found no differences in methacholine response or inflammation, supporting the idea that BDNF purely induces neuronal changes in affecting airway contraction (17), whereas others have reported BDNF effects on inflammation and direct changes in ASM contractility (20, 52, 145). Regardless of what cells are targeted, BDNF can induce a phenotype shift in neurons from nodose ganglia, which express TrkB receptors, and start expressing TRPV1 channels after BDNF exposure comparable to that observed after ovalbumin (100). Furthermore, chronic BDNF exposure in vagal ganglion neurons enhances the sensitivity to capsaicin (196) (FIGURE 3).

Genetic modification of TrkB or p75 receptor also alters neuronal innervation and airway responses. In mice with a dysfunctional TrkB receptor involving a point mutation in the kinase domain, ovalbumin-induced increases in nerve density and AHR are reduced (5). Similar effects are observed after knockout of the neurotrophin NT4, which also signals via TrkB (5). NT4 derived from mast cells can be a major driver of allergen-induced changes in nerve density (137). In NT4 knockout mice, a strong reduction in sensory innervation of neuro-epithelial bodies is seen, which may impact mucus hypersecretion (7, 135). NT3 may also contribute to airway neuronal plasticity, since NT3, but not NGF or BDNF, is able to induce a shift of NANC neurons toward cholinergic phenotype (136). Finally, p75 knockout prevents development of AHR after ovalbumin challenge, despite higher levels of NGF in the lavage fluid (172) broadly highlighting the importance of neurotrophins.

Therapeutic Significance of Neuronal Plasticity

There is now strong evidence that, in asthma, neuronal plasticity underlies symptoms including AHR. Indeed, in this context, asthma could be considered a disease of the nervous system, i.e., the concept of neurogenic asthma proposed decades ago. It is interesting to speculate whether altered neuronal regulation is a cause or consequence of asthma. The available data would suggest the latter, although causal studies are lacking at this stage.

Development of novel therapies for asthma has largely focused on anti-inflammatory therapies, including antibodies against IgE, IL-5, and IL-13, with varying success depending on asthma phenotypes (6, 112). Now that we have better understanding of the extensive changes in neuronal regulation in asthma, it seems appropriate to focus therapies against pathways involved in neuronal dysregulation. This can in part be achieved by current therapies. Here, anticholinergics, muscarinic receptor antagonists that inhibit effects of the parasympathetic system, have certainly led the way, underscoring the relevance of increased cholinergic tone in asthma (83). Beyond alleviation of acute symptomatology, long-acting anticholinergics are effective in improving lung function and reducing exacerbations in patients with moderate and severe disease (76, 77). Interestingly, the treatment effect is comparable over a broad population and independent of typical baseline characteristics including reversibility (78). Nonetheless, there is room for improvement for anti-cholinergic therapies beyond bronchodilation and reduction of mucus secretion, since new data demonstrate the role of cholinergic pathways in airway remodeling and inflammation (83). Furthermore, cholinergic stimulation likely has long-term effects on target cells, including ASM, in terms of secretion of inflammatory mediators, growth factors, and even ECM components, making anti-cholinergics an attractive target for the foreseeable future.

Although the cadre of factors involved in asthma is ever-growing, it is likely that many of these factors, such as inflammatory mediators and growth factors, directly or indirectly influence neuronal plasticity. Here, neurotrophins certainly appear to be an important subset. Currently, there are no studies investigating the effect of neurotrophin antagonism in asthma. On the other hand, there is evidence that levels of BDNF and other neurotrophins are influenced by current therapies. Inhaled corticosteroids suppress neurotrophin levels (104, 128), whereas long-acting β-agonists increase BDNF (103). The effect of anticholinergics on neurotrophin levels in asthma is unknown.

To specifically target sensory nerves, NK1 and NK2 antagonists have been studied in the past as mediators of neurogenic inflammation, but the results have been disappointing (13). Novel treatment options include antagonists of channels and receptors expressed on sensory neurons, including TRPA1, TRPV1, and P2X3 antagonists. These compounds have been identified as promising targets for cough, with ongoing clinical studies, and may also be relevant for asthma (1, 80, 197), as suggested by in vivo animal data (43). Furthermore, channel inhibitors targeting the neuronal voltage-gated sodium channel 1.7, recently shown to be the primary sodium channel responsible for neurotransmission in parasympathetic ganglia, are of interest (89).

Beyond pharmacological approaches, interventional strategies such as bronchial thermoplasty are of increasing importance. During this bronchoscopic procedure, radio-frequency energy is delivered to the airway wall to ablate ASM. This results in an improved quality of life and reduced exacerbation frequency in patients with severe asthma (30). However, the relationship between local ASM ablation and broad lung/system-wide effects has remained difficult to reconcile. Interestingly, bronchial thermoplasty also ablates airway nerves (146), which may in fact be the mechanism underlying the observed broad effects. Another potential bronchoscopic intervention to address increased nerve activity in asthma is targeted lung denervation (TLD). Akin to bronchial thermoplasty, radio-frequency energy is delivered to the airway, in this case designed to specifically target nerves (163). Concomitant cooling during TLD is used to protect other structures. Complete deletion of vagal nerve tracts has been confirmed in animal studies (64) and may also occur in bronchial thermoplasty. However sensory neurons do show a stronger regenerative capacity compared with cholinergic neurons (142). Initial studies in patients with COPD are promising and suggest that TLD improves lung function and quality of life, and reduces airway inflammation (86, 163), and the first data from the sham-controlled study support a similar conclusion (162). Clinical studies in patients with asthma are currently ongoing.

Conclusions and Future Directions

Asthma is a chronic airway disease in which airway nerves play a central role. Airways are densely innervated, and there is compelling evidence that the nervous system is plastic and hyperactive in asthma. Most of the symptoms in asthma are secondary to changes in the nervous system, either peripheral or central. These alterations can occur already early in life and comprise acute and chronic, long-lasting changes. Future research into the exact mechanisms of neuronal plasticity in the airways is warranted and can contribute to novel treatment strategies for asthma, which are potentially more successful than anti-inflammatory therapies. In this regard, a number of answered issues should help drive future research:

• When do initiating events of asthma occur? With airway innervation starting early in the embryonic period, genetic and environmental factors that alter fetal airway branching and contractility hold substantial potential to influence growth and functionality of airway nerves, only exacerbated by postnatal factors.

• Are certain asthma phenotypes more likely to show neuronal plasticity pointing to therapies targeting this aspect in specific patient populations? Here, the fact that the anticholinergic tiotropium improves lung function independent of baseline characteristics including degree of airway obstruction, reversibility, or blood eosinophils (29, 78) and is thus effective across a broad asthma population might suggest this not to be the case. However, data from the TALC study does suggest the tiotropium response to be associated with degree of airflow limitation and reversibility (141). Thus it may be appropriate to identify and classify patients with altered airway innervation in the context of individualized therapy, although it remains unclear whether functional readouts such as AHR or reversibility with anticholinergics (or other nerve-targeting drugs) would be sufficient to demonstrate neuronal plasticity per se, without access to anatomic samples or advanced intraluminal imaging, particularly of small airways.

• Is the role of airway innervation and/or neuronal plasticity consistent in asthma across the age spectrum? Here, does neuronal plasticity early in life contribute to lifelong asthma, and how do such contributions change as a consequence of repeated, heterogeneous stimuli during the lifespan? In this regard, is new onset asthma later in life equally a matter of neuronal dysfunction, and does neuronal plasticity play a role in asthma of the elderly (161), a condition that is increasing in prevalence and is more resistant to current therapies?

• There is substantial evidence for sex differences in asthma across the lifespan, classically with greater asthma in boys and increased asthma in pre-menopausal women (60, 175). However, the contributions of genetics versus sex hormones and their specific effects on airway innervation or plasticity have not been explored.