Neuroimmune recognition and regulation in the respiratory system

Trigeminal nerve

The ophthalmic and maxillary nerves of the trigeminal nerve project into the nasal epithelium in the nasal cavity. The trigeminal nerve is a heterogeneous nerve. Its sensory neurons are unique, with most of the cell bodies located in the trigeminal ganglia and the rest in the midbrain trigeminal nucleus of the brain [23]. Stimuli are detected by free nerve endings in the nasal epithelium [24]. In the nasal epithelium, while certain lipophilic stimuli (e.g. peppermint, ammonia) can be detected by the free intraepithelial nerve endings of the trigeminal nerve through intrinsic TRP channels, including TRPV1 (capsaicin), TRPA1 (mustard oil, etc.) and TRPM8 (peppermint), the ability of free nerve endings to penetrate the barrier and contact stimuli is limited. Thus, the trigeminal nerve enhances its ability to detect stimuli by projecting onto specialised SCCs [25, 26]. This enhances the ability of the trigeminal nerve to detect various types and quantities of pathogens [27].

Olfactory nerve

The olfactory nerve is located mainly at the top of the nasal cavity. It is made up of pure sensory neurons and are bipolar neurons that express two olfactory receptors: odour receptors and trace amine-associated receptors [28]. These receptors closely integrate signals to sense >1 trillion molecular stimuli in the air [29], transmitting the signals through the olfactory bulb to the olfactory cortex. When a harmful or toxic gas is detected, this triggers an aversive emotional response that mediates avoidance and defensive behaviour. Multiple respiratory pathogens target olfactory neurons, so loss of smell has been observed in various respiratory diseases [30].

Glossopharyngeal nerve

The glossopharyngeal nerve primarily innervates the pharynx in the respiratory tract. It is also a heterogeneous nerve, with its sensory neurons being pseudounipolar neurons. The cell bodies of the sensory neurons are located in the petrosal ganglion, with one end of the neuron extending to the pharynx, transmitting recognition information to the paratrigeminal nucleus (Pa5) and the nucleus of the solitary tract (NTS) [31, 32]. The glossopharyngeal nerve is primarily composed of sensory neurons, capable of detecting the mucosa of the throat, blood pressure, eardrum and tonsils. It is considered to be an important component of neuroimmune surveillance of the oral cavity [33]. Recent studies have found that γ-aminobutyric acid receptor subunit α1 (GABRA1⁺) petrosal neurons located in the posterior region of the nasopharynx sense influenza virus infection, inducing sickness behaviour [15]. Additionally, the motor neurons of the glossopharyngeal nerve, with cell bodies located in the dorsal motor nucleus of the vagus, nucleus ambiguus and inferior salivatory nucleus [34] extend synapses to the periphery, regulating swallowing, saliva release and exerting parasympathetic functions [35]. However, it remains unclear how motor neurons of the glossopharyngeal nerve directly regulate peripheral immune responses. Nevertheless, they can influence the drainage and cleansing of the pharyngeal mucosa through the movement of pharyngeal muscles and secretion of salivary glands, thereby impacting mucosal immune responses. Additionally, the glossopharyngeal and vagus nerves are intimately related, thus it has the potential to regulate immune responses by affecting the vagus nerve.

Vagus nerve

The vagus nerve connects visceral tissues with the CNS, mediates internal sensory and physiological regulation in the body and responds to external stimuli [36]. The vagus nerve is the most extensive nerve that projects into the respiratory tract, from the pharynx to the alveoli. The vagus nerve is made up of sensory neurons and motor neurons that move together in the same bundle, with sensory neurons making up ∼80% of the total number of neurons. The cell bodies of sensory neurons are located in the nodose and jugular ganglia [37]. Ganglia contain not only neurons, but also glial cells, Schwann cells, endothelial cells and immune cells, which are collectively involved in signal recognition, axon regeneration and transmission [38, 39].

Vagal sensory neurons project into the respiratory tract and are primarily located in the right vagus nerve [40]. Vagal sensory neurons have a pseudounipolar structure. One end extends to the NTS or Pa5, while the other end extends to peripheral organs. Neurons projecting to the NTS are predominantly located in the nodose ganglia, and their axons are predominantly projected to the trachea, across the entire lung and onto the bronchial/bronchioles, vasculature/lymphatic, alveoli and PNECs [41, 42]. Sensory neurons projected to Pa5 originate from neurons located in the jugular ganglia, and their axons project mainly to the pharynx and trachea, and to a small extent to the bronchi and alveoli and vessels of various sizes [43]. Vagal sensory neurons projected to the trachea and lungs originate mainly from the nodose ganglia, and ∼65% of tracheal sensory neurons originate from the nodose ganglia, with a few from the jugular ganglia. Similarly, sensory neurons that project to the lungs also have their central projections in the NTS [22, 40, 42]. Most vagal sensory neurons are unmyelinated, slow-conduction C-fibres that can be activated by a variety of mechanical and chemical stimuli to enable the transmission of peripheral immune messages to the brain. Although vagus neurons make up only a small portion of the nervous system, vagus neurons exhibit a high degree of heterogeneity, with different subpopulations of neurons having different important functions and can be divided into ≥37 clusters [44]. Vagal neurons that control different tissues exhibit different transcriptome patterns, especially in neurons that control the lungs and oesophagus, where they show a distinct subset distribution [22], indicating their functional independence [45]. The cell bodies of vagus motor neurons are located in the dorsal motor nucleus of the vagus and nucleus ambiguus. Their axons project into the respiratory tract, and their nerve endings release acetylcholine (ACh) directly onto target cells or act on intrinsic neurons within the respiratory tract [46].

Dorsal root ganglia

Sensory neurons in DRGs are classified as pseudounipolar neurons. Neurons that project to the lungs are predominantly located in the DRG segment T1–T3, project to bronchial tubes and blood vessels >376 mm in diameter, and project to neuroepithelial bodies. The secretion of CCL2 by vascular endothelial cells is crucial for early neural recognition [47]. However, projections to alveoli and small blood vessels were not observed [43]. Further confirmation is required to determine if the DRG can be projected onto the trachea.

Sympathetic neurons

The sympathetic neurons projected to the respiratory tract mainly originate from the superior cervical ganglia, stellate ganglia and thoracic sympathetic chain ganglia in the T2–T4 segments [40]. These neurons exhibit different projection sites along the respiratory tract. Specifically, the superior cervical ganglion protrudes into the upper or middle part of the cervical trachea, and the stellate ganglion protrudes primarily into the inferior trachea, but also to the lungs. The thoracic sympathetic links of the T2–T4 segments extend into the lungs [40]. Similar to the fact that vagus nerve sensory neurons originate mainly on the right side, sympathetic neurons projected into the respiratory tract are also mainly distributed on the right side, accounting for ∼70% of the distribution [41].

Pulmonary parasympathetic inflammation reflex

Both sensory neurons and motor neurons contribute to immune regulation, which helps modulate the local inflammatory microenvironment during infection, promoting pathogen clearance. In particular, sensory neurons can transmit signals to the central nervous system, activate specific brain regions and trigger specific efferent nerve excitations. Currently, the theory of immune regulation of respiratory efferent nerves is the “pulmonary parasympathetic inflammatory reflex” [71]. The parasympathetic nervous system promotes physiological and pathological processes in the lungs by releasing active substances such as ACh, VIP, neuromedin U (NMU), etc. These neuropeptides and neurotransmitters act on a variety of cells, including macrophages, T-cells, T follicular helper cells, B-cells, red blood cells, type II alveolar epithelial cells and airway epithelial cells. ACh, in particular, can interact with nAChRs and muscarinic acetylcholine receptors (mAChRs) expressed on these cells. In addition, ACh in the lungs comes not only from the vagus nerve, but also from airway epithelial cells, and immune cells such as B-cells and T-cells can also produce ACh [11, 72, 73]. Neuroepithelial–immune interactions work together to promote ACh signalling in physiological and pathological processes in the lungs. Traditionally, activation of nAChR in the airways has been associated with inducing anti-inflammatory effects, while activation of mAChR has been thought to cause pro-inflammatory effects [74]. Regarding the role of α7nAChR, the common concept is that it is an important component of the cholinergic anti-inflammatory pathway. Activation of α7nAChR is thought to downregulate the production of pro-inflammatory factors such as tumour necrosis factor (TNF)-α. However, it was found that the effects of α7nAChR activation depended on the cell type and disease model [18–20, 75–82]. Activation of α7nAChR can lead to different regulatory roles in various cellular and disease models (table 1). For example, activation of α7nAChR upregulates the inflammatory response of megakaryocytes. Lung megakaryocytes produce ∼50% of platelets in the body and exhibit significant differences in gene expression compared to bone marrow megakaryocyte cells [83]. Lung megakaryocytes express fewer mature megakaryocyte markers, and their gene expression pattern resembles that of antigen-presenting cells. Lung megakaryocytes can induce the activation of CD4⁺ T-cells both in vitro and in vivo in a major histocompatibility complex II-dependent manner. Additionally, the immune phenotype can be altered based on the tissue environment, such as pathogen challenge and interleukin IL-33 [84]. This means that lung megakaryocytes act as early responders regulated by vagal-α7nAChR signalling (our unpublished data). In addition, α7nAChR is expressed in lung fibroblasts. Activation of α7nAChR in human fibroblasts enhances fibrosis genes (Acta2, Colla1) and promotes pulmonary fibrosis [20]. Vagus-α7nAChR signalling in different cells ensures the ability to respond to different stages of infection caused by various pathogens, coordinating the balance between pathogen recognition, clearance and tissue repair.

Bacterial infection

Currently, the process of neural recognition and its mediation of neuroimmune responses to various bacteria has been discovered. Here, we primarily discuss S. aureus, P. aeruginosa, Bacillus anthracis and M. tuberculosis. In the nasal epithelium (figure 4a), irritants can be detected at the free intraepithelial nerve endings of the trigeminal nerve. At the same time, the trigeminal nerve enhances its ability to detect irritants by projecting to specialised SCCs [27]. SCCs express multiple types of receptors, including the bitter taste receptor T2R, which can identify acyl-homoserine lactones produced by Gram-negative bacteria. SCCs also express choline O-acetyltransferase (CHAT) [85, 86], which enables signal transmission through acetylcholine (ACh), activating the trigeminal nerve. This activation leads to the release of CGRP and substance P through axon reflex, triggering neurogenic inflammation, which affects vascular endothelial cells and results in increased blood flow, vascular leakage and oedema. This process facilitates the recruitment of inflammatory leukocytes [58]. The arrival of leukocytes aids in clearing bacteria, thereby combating bacterial invasion.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4

Examples of neuroimmune interactions in the respiratory tract. a) Bacterial infection: sensory neurons act as second-order neurons that receive input from respiratory epithelial cells, such as solitary chemosensory cells (SCCs). SCCs sense acyl-homoserine lactones (AHLs) [27, 87] produced by Gram-negative bacteria through bitter taste receptors (T2R). The release of acetylcholine (ACh) from SCCs activates trigeminal sensory neurons, triggering axonal reflexes, leading to the release of calcitonin gene-related peptide (CGRP) and substance P (SP) from peripheral nerve endings and neurogenic inflammation. Nerves can sense bacterial infections in a variety of ways. One way to do this is to identify pathogens directly through neurons. For example, transient receptor potential (TRP)V1⁺ vagus sensory neurons directly recognise Staphylococcus aureus, resulting in the release of CGRP from neurons, which acts on locally infected neutrophils and γδT-cells, inhibiting their inflammatory response. In addition, sensory neurons act as second-order neurons that receive input from respiratory epithelial cells, such as brush cells. Brush cells sense bitter-tasting quorum-sensing molecules (QSM) produced by Pseudomonas aeruginosa through TAS2R taste receptors. Similar to SCCs, the release of ACh from brush cells activates sensory neurons, triggering axonal reflexes, leading to the release of CGRP and SP from peripheral nerve endings and neurogenic inflammation. b) Viral infection: severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) can directly infect olfactory sensory neurons, resulting in olfactory epithelial destruction, leading to anosmia. In addition, SARS-CoV-2 can enter the brain through the olfactory nerve, triggering the activation of microglia and recruiting CD8 T-cells. This activation causes microglia and CD8 cells to aggregate, forming nodules that cause damage to brain tissue. In the case of influenza virus infection, γ-aminobutyric acid receptor subunit α1 (GABRA1⁺) petrosal neurons located in the posterior region of the nasopharynx expressing prostaglandin E2 (PGE2) receptor 3 (EP3) can sense pathogen-induced PGE2, thereby inducing disease behaviour. c) Allergic airway inflammation: multiple types of nerves are involved in regulating allergic airway inflammation. For the vagus nerve, TRPV1⁺ vagal neurons can sense IgE produced by the body's immune response through FcεR1 receptors. They release SP, promoting the inflammatory response of type 2 helper (Th2) cells and enhancing the secretion of mucus by goblet cells. Simultaneously, Mrgprc11⁺ jugular neurons of the vagus nerve activate parasympathetic neurons, leading to the release of ACh, which mediates bronchoconstriction, inducing airway hyperreactivity. The parasympathetic ganglia can also release neuromedin U (NMU), which acts on NMUR1 receptors on group 2 innate lymphoid cells (ILC2), activating them and amplifying allergic inflammation. The parasympathetic nervous system can also negatively regulate ILC2 cells by releasing ACh, which acts on α7 nicotinic acetylcholine receptors (a7nAChR) on ILC2 cells, inhibiting their transformation into inflammatory ILC2 cells. ILC2 cells play a crucial role in the neuroimmune interaction of airway allergic inflammation and can be regulated by various neurons. Regarding sensory neurons, they sense allergic inflammation and release vasoactive intestinal peptide (VIP), which acts on VPAC2 receptors on ILC2 cells, promoting the release of interleukin (IL)-5 and enhancing the inflammatory response of eosinophils. Activated eosinophils can increase airway epithelial nerve density, further leading to airway hyperreactivity. The sympathetic nervous system similarly regulates ILC2 cells by releasing dopamine and norepinephrine, which act on the D1 dopamine receptor (DRD1) and β₂-adrenergic receptors (AR) on ILC2 cells, respectively, negatively modulating their response. The sympathetic nervous system can also release norepinephrine, which acts on β₂-ARs on airway smooth muscle cells (ASMs), mediating airway dilation. Additionally, oesophagus intrinsic neurons can mediate airway dilation by releasing VIP and nitric oxide (NO). Not limited to neurons, ILC2 cells also receive regulation from pulmonary neuroendocrine cells (PNECs) that secrete CGRP and γ-aminobutyric acid (GABA), promoting the release of IL-5 and IL-3, leading to a robust inflammatory response. d) Acute lung injury. The parasympathetic nervous system plays a vital role in this process. ACh acts on the α7nAChR receptor of type 2 alveolar cells (AT2), promotes the differentiation of AT2 cells into AT1 cells, and mediates the repair of alveolar epithelial damage. This process depends on the vagus nerve. IAV: influenza virus; H: hydrogen; P2Y purinoceptor 1; OVA: ovalbumin; DC: dendritic cell; EET: eosinophil extracellular traps; NE: norepinephrine; CHAT: choline O-acetyltransferase; M3: M3 muscarinic ACh receptor; DRG: dorsal root ganglion; CNS: central nervous system.

S. aureus can directly activate TRPV1⁺ nociceptive neurons projecting to the respiratory tract (figure 4a), releasing CGRP neuropeptides that act directly on neutrophils [87], inhibiting their recruitment and surveillance. Additionally, it alters the population of γδT-cells in the lungs and suppresses the immune response to S. aureus [17]. The role of CGRP here is different to that described earlier in triggering neurogenic inflammation. In the early stages of infection, neuronal release of CGRP acts on endothelial cells, alters vascular permeability, recruits neutrophils and amplifies inflammation, thereby promoting bacterial clearance [17]. In the later stages of infection, when there are large numbers of immune cells at the site of infection, CGRP mainly targets immune cells, such as neutrophils, to inhibit their inflammatory response. This will attenuate the inflammatory response in the lungs caused by streptococci, thereby facilitating tissue repair [87]. By precisely regulating the immune response according to the local immune environment at different stages of infection, the balance between bacterial clearance and tissue repair is maintained [88].

Regarding P. aeruginosa, nociceptive neurons, through their projections to the trachea's brush cells, sense bitter-tasting quorum-sensing molecules from P. aeruginosa (figure 4a). When brush cells are activated, they release ACh, which activates nociceptive neurons. Through axonal reflexes, this activation leads to the release of CGRP and substance P, inducing neurogenic inflammation and activating the innate immune system [59]. Substance P can interact with the mast cell receptor (MRGPRB2; Mas-related G-protein coupled receptor member B2) to promote degranulation and the release of inflammatory factors, further amplifying the inflammatory response [89].

Sensory neurons are involved in the spread of bacteria. For example, M. tuberculosis can cause pulmonary tuberculosis, with persistent cough being a major symptom. It was previously believed that M. tuberculosis primarily spreads through coughing, but recent research suggests that tidal breathing is also a potentially significant transmission route [90]. In the pathway of coughing transmission, peripheral sensory neurons recognise M. tuberculosis, triggering the cough reflex, which may potentially facilitate transmission. Studies have found that lipid components of the M. tuberculosis cell wall lipoglycan-1 (SL-1) and organic extracts composed of lipophilic molecules derived from M. tuberculosis activate neurons of the vagus nerve and dorsal root ganglia in vitro, rapidly increasing intracellular calcium concentration [91]. The purinergic ion channel family (P2X) plays an important role in cough [4, 92]. However, receptors on sensory neurons responsible for recognising M. tuberculosis SL-1 have not been fully elucidated.

Viral infection

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) can infect SCCs expressing angiotensin-converting enzyme (ACE)2 in the nasal cavity [93], subsequently activating trigeminal sensory neurons. This activation triggers the release of neuropeptides such as CGRP and substance P through axonal reflexes, inducing neurogenic inflammation. Similarly, SARS-CoV-2 can directly infect nerves, damaging olfactory nerves and causing loss of smell.

The olfactory nerve also provides a direct pathway for SARS-CoV-2 to enter the brain (figure 4b) [94–96]. Signals of SARS-CoV-2 nucleoprotein have been detected in the glossopharyngeal nerve, vagus nerve and isolated cells of the brainstem [97]. Additionally, SARS-CoV-2 spike protein-positive signals have been observed in the perivascular region [16]. From the perivascular region, SARS-CoV-2 breaches the blood–brain barrier via the haematogenous pathway, triggering the activation and recruitment of microglia and CD8⁺ T-cells. The aggregation effect of recruited CD8⁺ T-cells and microglia leads to the formation of a large number of nodules in the perivasculature [16]. A brain organoid experiment found that SARS-CoV-2 infection leads to the formation of multicellular syncytia in brain organoids, with the SARS-CoV-2 spike protein causing neuron–neuron, neuron–glia and glia–glia fusion, resulting in the diffusion of large molecules and organelles and severe impairment of neuronal activity [98]. Additionally, α7nAChR is also a target of the SARS-CoV-2 spike protein, and the peptide from the neurotoxin-like region of SARS-CoV-2 both enhances and inhibits α7nAChR function, depending on concentration differences, leading to cholinergic dysfunction in coronavirus disease 2019 (COVID-19) and disruption of anti-inflammatory effects [99]. Cholinergic signalling can inhibit SARS-CoV-2 infection in lung epithelial cells. Activation of α7nAChR inhibits ACE2 expression in lung epithelial cells, thereby inhibiting the entry of the S protein. Activation of α7nAChR also reduces the inflammatory response of SARS-CoV-2-infected lung epithelial cells [77]. Furthermore, SARS-CoV-2 also targets dopaminergic neurons, inducing their senescence. A significant reduction in the number of neuromelanin⁺ and tyrosine-hydroxylase⁺ dopaminergic neurons and fibres in the substantia nigra tissue was observed in a cohort of severe COVID-19 patients [100]. SARS-CoV-2 inflicts damage on the CNS, leading to functional disorders in some patients even after apparent recovery from COVID-19, including fatigue, dyspnoea, hyposmia/anosmia, hypogeusia/ageusia, memory/cognitive impairment, sleep disturbances and pain syndromes. This clinical manifestation is referred to as “long COVID”. Additionally, a retrospective cohort study conducted 1 year after COVID-19 revealed a significantly increased risk of new-onset Alzheimer's disease in patients with a history of COVID-19 [101].

For the influenza virus (IAV), researchers have identified a specific neuron subtype expressing GABRA1 in the glossopharyngeal nerve. These neurons project nerve fibres to the nasopharynx (figure 4b), and following influenza infection, an increased expression of cyclooxygenase (COX)-2) is observed in the nasopharynx. Secreted PGE2 can be sensed by the GABRA1⁺ neurons of the glossopharyngeal nerve through the PGE2 receptor 3 (EP3), thereby inducing influenza disease behaviour [15]. By ablating GABRA1 neurons or selectively deleting EP3 receptors in these neurons, symptoms of reduced food intake, reduced water intake and impaired exercise capacity during early influenza infection can be eliminated, and the survival rate of mice can be improved. Cholinergic signalling also regulates PGE2 synthesis, activates α7nAChR and selectively upregulates COX-2 and PGE2 expression in rat microglia [102]. In addition, the sympathetic nervous system plays a key role in regulating influenza infection. Sympathetic nerves enhance influenza virus-induced lung inflammation, enhance inflammatory influx of monocytes, neutrophils, and natural killer cells by activating α-adrenergic receptors, and promote the production of pro-inflammatory cytokines and chemokines in bronchoalveolar lavage fluid. This exacerbates the pathogenesis of IAV [103].

Allergic airway inflammation

Neuroimmune interactions in allergic airway inflammation have been studied extensively, and various types of nerves are collectively involved in the recognition and immunomodulation of allergic airway inflammation. Among them, TRPV1⁺ vagus sensory neurons sense IgE through the expression of FcεR1 (figure 4c). Subsequently, they release substance P via axonal reflex, which amplifies the influx and polarisation of type 2 helper (Th2) cells in the airways [53]. Mast cells link IgE immune sensing to mediate antigen-avoidance behaviour [104, 105]. Allergen recognition signals from vagal sensory neurons are transmitted to the brain, activating the NTS. This process is associated with mast cells and IL-4. As a result, Dbh⁺ neurons in the nucleus ambiguus are activated, resulting in enhanced hyperresponsiveness [106]. Mast cells exclusively interact with bronchial nerves. Based on chymase expression, mast cells are classified into two subtypes: mast cells containing only tryptase (MC_T) and those containing both tryptase and chymase (MC_TC) [107]. Under physiological conditions, MC_T are anatomically associated with nerves and play a crucial role in axon reflex. However, the interaction between MC_TC and neurons is not observed in the physiological state. During allergic airway inflammation induced by house dust mite, mast cells migrate from the systemic circulation to the airways and lungs, yet there is no increase in the interaction between MC_T and neurons. Nevertheless, the interaction between MC_TC and nerves does occur [5]. However, the specific interactions between these two different subtypes of MCs and neurons, as well as their roles in allergic airway inflammation, remain unclear. Therefore, elucidating the differences in gene expression, as well as the expression of neural receptors between these two cell types, will contribute to a better understanding of the neural regulation mechanisms underlying allergic airway inflammation.

ILC2 cells, airway smooth muscle cells and goblet cells are important targets for neuromodulation. Goblet cells can receive substance P released by sensory neurons [108], ACh released from parasympathetic neurons and γ-aminobutyric acid (GABA) released from PNECs [8]. These neuropeptides and neurotransmitters promote goblet cell proliferation and Muc5AC mucus imbalance, leading to allergic airway inflammation (figure 4c).

ILC2 cells play a crucial role in type 2 inflammation, producing type 2 cytokines such as IL-4, IL-5 and IL-13, that mediate a cascade involving eosinophils and Th2 cells (figure 4c). Multiple types of nerves and neurotransmitters can regulate the function and activity of ILC2. For the parasympathetic nervous system, its regulatory effect on ILC2 is diverse. For example, it releases ACh, which acts on the a7nAChR of ILC2 cells, reducing the amount of ILC2 in the lungs and inhibiting their transition to inflammatory ILC2 [19, 109]. Meanwhile, it releases NMU, which acts on Nmur1, activating ILC2 and amplifying allergic inflammation [110]. For the other neuropeptide, neuromedin B (NMB), its stimulation inhibits ILC2 responses, and this effect depends on basophils. Basophils directly enhance the expression of NMB receptors on ILC2 [111]. Nasal trigeminal sensory neurons release NMB to mediate sneezing signalling, playing a crucial role in allergic rhinitis and viral respiratory infections [112].

The sympathetic nerve releases dopamine, which acts on the D1 dopamine receptor (DRD1) of ILC2 cells, impairing the mitochondrial oxidative phosphorylation pathway in ILC2s, thereby inhibiting ILC2 activation and airway inflammation [113]. However, the modulation of dopaminergic signalling by the sympathetic nervous system to allergic airway inflammation is age-dependent and is affected by ongoing changes in the innervation of the immune system and surrounding tissues after birth. In the lungs of young mice, dopaminergic signalling promotes the differentiation of Th2 cells through DRD4. These findings reveal an age-related neural mechanism that predisposes young individuals to allergic inflammation [114]. Furthermore, the sympathetic nervous system can also release norepinephrine that acts on the β₂-adrenergic receptor of ILC2 cells, inhibiting cell proliferation and activity. This subsequently suppresses ILC2-dependent type 2 inflammation [115].

Sensory neurons sense allergens, release VIP and stimulate ILC2 cells through VPAC2 receptors to induce the release of IL-5, thereby increasing the number of eosinophils in tissues [116]. Eosinophils in asthma can increase airway epithelial nerve density [117]. Disruption of normal neuromodulation leads to airway narrowing, remodelling and inflammation in asthma and airway hyperresponsiveness. Additionally, during allergic inflammation, there is an increase in the proportion of CGRP⁺ neurons in the vagal ganglia, accompanied by an elevated number of dendritic cells [118, 119]. However, how these dendritic cells and CGRP⁺ neurons regulate allergic airway inflammation remains unknown. Dendritic cells are the most potent antigen-presenting cells, and specific subsets of dendritic cells play a crucial role in initiating and sustaining Th2 immune responses. However, certain dendritic cells may also suppress asthma responses [120]. Therefore, clarifying the role of dendritic cell subsets in neuroimmune dysregulation is essential for understanding the occurrence and progression of allergic airway inflammation. Additionally, CGRP has been found to exert opposing regulatory effects on ILC2 [8, 121]. However, it is conceivable that they may influence allergic inflammation by altering the immune microenvironment of the vagal ganglia, thereby modifying the recognition function of sensory neurons. ILC2 cells respond to neuropeptides and neurotransmitters released by PNECs. In response to allergens and eosinophil extracellular traps formed in the airways [122], PNECs can sense and produce CGRP and GABA [8], which stimulate mucus production in goblet cells. Simultaneously, CGRP stimulates ILC2 cells to produce cytokines, leading to the recruitment of downstream immune cells including Th2 cells and eosinophils [8]. However, the mechanism of CGRP on ILC2 is still under further investigation. The spinal sensory JAK1 and the vagal βCGRP axis have been identified to suppress ILC2 responses and allergic lung inflammation [88]. Additionally, CGRP negatively regulates IL-33-induced airway inflammation and alarmin-driven ILC2 responses [123]. PNECs are innervated by various types of neurons and serve as airway monitoring points, primarily at airway bifurcations, where they detect oxygen, carbon dioxide, hydrogen ions, nicotine and mechanical stimuli [124]. However, the mechanisms of signalling between PNECs and neurons and how neurons regulate PNECs are not fully understood.

Airway smooth muscle cells play a key role in the pathogenesis of asthma. These cells regulate muscle tone, contraction, and relaxation which regulates the calibre of the airways. Airway smooth muscle cells are involved in airway hyperresponsiveness and regulates airway stenosis, remodelling and inflammation in asthma [125]. Various types of neurons are involved in controlling the contraction and relaxation (figure 4c). The parasympathetic nerves play a crucial role in bronchoconstriction, a process that relies on the activation of MRGPRC11⁺ jugular neurons. These neurons relay signals to the brain, causing parasympathetic to release ACh, which acts on M3 mAChRs on the airway smooth muscle cells, resulting in bronchoconstriction and airway hyperresponsiveness [126]. However, the molecules that activate MRGPRC11⁺ jugular neurons are not fully understood. Conversely, the sympathetic nerves release norepinephrine, which acts on β₂-adrenoceptors on the airway smooth muscle cells to induce airway relaxation [127]. However, dysregulation of norepinephrine signalling leads to hyperstimulation of B-cell β₂-adrenoceptors, which promotes an increase in IgE, enhances Th2-dependent IgE responses and ultimately promotes the development of lung inflammation [128]. In addition, oesophageal muscular and intestinal neurons mediate airway relaxation through the release of NO and VIP. These different neurons release different neuropeptides and neurotransmitters to coordinate the regulation of airway contraction and relaxation and maintain the physiological stability of the airways.

Acute lung injury

The basic pathological features of acute lung injury (ALI) include excessive accumulation of inflammatory cytokines and immune cells, as well as damage to the alveolar epithelium. Reducing the inflammatory response and enhancing alveolar epithelial regeneration are two key strategies to promote lung repair in ALI. The parasympathetic nerve plays an important role in this regard, as it can utilise the cholinergic anti-inflammatory pathway to reduce systemic levels of ALI-induced inflammatory factors. By promoting the phosphorylation of AKT1 in α7nAChR⁺CD11b⁺ cells, this mechanism stabilises these cells within the spleen, reducing their efflux and recruitment from the spleen to the lungs, thereby attenuating Escherichia coli and LPS-induced acute lung inflammation [81]. The vagal-α7nAChR signal also regulates α7nAChR⁺Sca1⁺ bone marrow mononuclear cells transplanted into the lungs, attenuating lung damage [79]. The vagus nerve promotes alveolar epithelial regeneration by activating α7nAChR on type 2 alveolar cells (AT2) to promote AT2 cell proliferation (figure 4d). This is followed by differentiation into AT1 cells, which mediate alveolar regeneration [18]. A substantial number of epithelial cells also express CHAT [72], and the neuropeptide Y receptor type 2 (Npy2r⁺) vagal neurons can project to these epithelial cells [1]. Whether Npy2r⁺ vagal sensory neurons can stimulate these cells to secrete ACh and promote lung repair requires further investigation.

The interaction between pathogens within the respiratory tract

Infection by an initial pathogen can either enhance or mitigate the infection of a subsequent pathogen, leading to either a synergistic or antagonistic interaction [129]. Pre-infection with IAV has been shown to increase ACE2 expression, and compared to mice infected with SARS-CoV-2 alone, those co-infected with IAV exhibit increased viral load and more severe lung damage. As previously mentioned, the nervous system can sense IAV and mediate systemic sickness responses [15]. Cholinergic α7nAChR signalling promotes IAV replication in AT2 epithelial cells (table 1), and reduces the expression of ACE2 in airway epithelium, suppressing SARS-CoV-2 infection in the lung [77]. Additionally, the lung harbours diverse microbial communities, among which pathogenic bacteria may be latent. ∼23.0% of the global population has latent M. tuberculosis infection [130, 131], and recent evidence suggests that COVID-19 promotes the progression of latent tuberculosis infection to active tuberculosis [132]. However, not all co-infections affect the course of SARS-CoV-2, such as HIV, human respiratory syncytial (RSV) virus, human parainfluenza virus, or human rhinovirus 3 [133, 134]. Neurons can recognise and regulate individual pathogens, but how to regulate the co-infection of pathogens remains unclear. Research on this topic is of significant importance for the treatment of secondary infections, and it may involve factors such as different activation sequences of neurons, interactions between neurons and changes in neuronal activity, as well as alterations in the CNS and respiratory tract states.

Inflammation in the CNS induced by respiratory disturbance

The physiological alterations in the respiratory tract, provoked by immuno-inflammation, can exert profound influences on an individual's behavioural patterns and psychological states. Peripheral diseases, such as inflammation of the respiratory tract, can remotely cause damage to the CNS [135]. Several respiratory pathogens can cause CNS symptoms either directly or indirectly [136]. Pathogens may invade the CNS directly by crossing the blood–brain barrier or enter through neural transmission pathways. The indirect pathways affecting the CNS include misdirected host immune responses and neural reflexes that sense the respiratory inflammatory responses.

Pathogens use neural pathways that project to the respiratory system as portals into the brain, especially those that project to the nasal cavity, such as the olfactory and trigeminal nerves [137]. Regarding the invasion of the brain by SARS-CoV-2, as discussed earlier, COVID-19 has been shown to increase the risk of newly diagnosed Alzheimer's disease [101]. Some herpes viruses infect the olfactory epithelium to reach the CNS. Several subtypes of influenza A viruses, such as H5N1, WSN/33 and H1N1 invade the CNS through the olfactory nerve. The Sendai parainfluenza virus strain infects olfactory sensory neurons in mice and subsequently spreads to the glomeruli of the olfactory bulb. Additionally, amoebae, including Naegleria fowleri [138] and Balamuthia mandrillaris [139, 140], penetrate the olfactory epithelium and cribriform plate to access the brain. These pathogens enter the brain directly and cause an immune response in the CNS, leading to synaptic damage, memory impairment, dizziness, headaches, impaired consciousness, acute cerebrovascular disease, seizures and other CNS manifestations [141].

Human metapneumovirus (hMPV) is the second leading cause of bronchiolitis, following RSV. Discovered in 2001, hMPV is a widespread and significant respiratory pathogen, and it is associated with a range of CNS diseases, from seizures to encephalitis [142]. Children infected with hMPV are more likely to experience seizures compared to those infected with RSV [142]. Autopsies of encephalitis patients have detected hMPV in both lung and brain tissues, suggesting it as a potential pathogen for acute encephalitis. However, the virus is not always detectable in cerebrospinal fluid, with one case report indicating its presence in cerebrospinal fluid may indicate severe encephalitis. In the early stages of mild encephalitis, hMPV is detectable in cerebrospinal fluid and nasal wash samples. Surprisingly, when clinical deterioration occurs, hMPV is detected in cerebrospinal fluid [143]. The mechanism by which hMPV enters the brain remains unclear, but possible pathways include neural pathways. hMPV undergoes biphasic replication in respiratory epithelial cells before migrating to neurons innervating the lungs, where it persists latently, undetectable in epithelial cells [144]. Currently, effective treatments for this virus are lacking. Exploring the process of infection in respiratory epithelial cells, the transition to neurons, and the latent mechanisms in the peripheral and CNS will provide new insights and treatment options for bronchiolitis, seizures, encephalitis and other related diseases.

In addition, pathogens that infect the respiratory tract can also break through the blood–brain barrier and blood–cerebrospinal fluid barrier [145]. For example, Streptococcus pneumoniae can enter the brain through the haematogenous route [146]. Human RSV (hRSV) can alter the permeability of the blood–brain barrier, allowing immune cells to enter the CNS, triggering inflammation of the CNS, and detecting the presence of hRSV in the cerebral cortex [147]. hRSV was detected in endothelial cells, neurons, microglia and astrocytes. Notably, astrocytes exhibited a higher frequency of infection, accompanied by an increased expression of glial fibrillary acidic protein. Additionally, there was a significant increase in the infiltration of monocytes, B-lymphocytes, and CD8⁺ T-cells within the CNS, potentially leading to behavioural and cognitive alterations.

Peripheral infection activates the immune system, chemoattracting immune cells and releasing inflammatory cytokines. These cells and cytokines not only affect peripheral tissues, but also act on the CNS, triggering neuroinflammation or altering the balance of neurotransmitters, further affecting brain function. For instance, autoimmune T-cells play a pivotal role in autoimmune diseases of CNS. Autoimmune T-cells mature in the lungs before migrating to the CNS to cause autoimmune diseases. After local stimulation of the lungs, these cells undergo vigorous proliferation and migrate to the CNS, subsequently inducing paralytic disease [148]. The lung microbiome can also regulate autoimmune diseases within the CNS. However, the pulmonary microbiome does not affect the activation of lung T-cells or their migration to the CNS. Instead, it enhances the CNS microglia signalling through interaction with LPS, promoting immune responses within the CNS, thereby increasing susceptibility to the CNS autoimmune diseases [149].

Respiratory tract inflammation also transmits signals to the CNS through neural fibres, mediating CNS symptoms. For example, in anxiety disorders, the glossopharyngeal and vagus nerves sense inflammation signals from the pharynx and transmit them to norepinephrinergic neurons in the nucleus of the solitary tract (NTS_NE), which project to the ventral bed nucleus of the stria terminalis, eliciting anxiety behaviour [150]. Influenza infection can result in numerous neurological complications, typically observed in children, including coma, seizures, paralysis and oculomotor palsies [135]. As mentioned earlier, GABRA1 neurons of the glossopharyngeal nerve also project to the pharynx, detecting locally produced prostaglandins and mediating systemic illness response to respiratory viral infections [15]. Investigating whether neurological symptoms are similarly mediated by this GABRA1 neuronal circuit and exploring the corresponding brain areas of injury will have a positive impact on the treatment of influenza.

Neuropsychological status can also affect immune responses. In the absence of immune challenge, psychosocial challenges induce lung inflammation and increases levels of the inflammatory cytokine IL-1β, as well as levels of chemokines such as keratinocyte-derived chemokine/CXCL1, macrophage inflammatory protein-2/CXCL2 and monocyte chemoattractant protein-1/CCL2. At the same time, the expression of adhesion molecules in the lungs is increased, recruiting neutrophils and monocytes into the lungs [151].

Environmental/occupational exposures

CNS diseases caused by environmental or occupational exposures are not limited to pathogens. These exposures can affect the immune response. Environmental exposures, primarily smoking and air pollution, have been studied extensively. A prospective cohort study involving >500 000 individuals in China systematically analysed the relationship between smoking and the incidence of >470 diseases. It was found that smoking was significantly associated with the incidence of 56 specific diseases, including cardiovascular diseases, respiratory diseases, cancer, digestive system diseases and others.

However, these epidemiological data also indicate that smoking reduces the incidence and/or severity of seven diseases (Parkinson's disease, other disorders of the conjunctiva, varicose veins, bronchiectasis, inguinal hernia, other arthrosis, and gonarthrosis) [152]. Here, the focus is on Parkinson's disease, and the aforementioned effects may be related to the impact of smoking components such as nicotine on the CNS. Nicotine, as an agonist, binds to nAChR and may exert anti-inflammatory effects through the pulmonary parasympathetic neural reflex or the cholinergic anti-inflammatory pathway, slowing the progression of Parkinson's disease. Studies have found that using auricular vagus nerve stimulation can alleviate neurodegenerative damage to dopaminergic neurons in Parkinson's disease, increasing the expression of tyrosine hydroxylase and α7nAChRs, while reducing the levels of inflammatory cytokines TNF-α and IL-1β. Additionally, it increases the number of regulatory T-cells while simultaneously decreasing Th17 cells [153]. Research on electronic cigarettes has also found adverse effects on blood–brain barrier function, downregulating the expression of vascular markers, enhancing vascular permeability, and increasing leukocyte–endothelial cell interaction and inflammation [154]. Therefore, quitting smoking is highly necessary, especially before the onset of the disease.

Air pollution not only regulates the infection of respiratory pathogens. For instance, in mice, exposure to particulate matter with aerodynamic diameter <2.5 µm can exacerbate influenza virus infection by compromising innate immune responses [155]. Air pollution also plays a role in the manifestation of brain diseases. Particulate matter from urban air pollution may also translocate to the brain. Additionally, there are indirect pathways affecting the brain. Microglia in the CNS play a crucial role in the process of nerve damage. Studies on automobile exhaust fumes found that exposing mice to diesel exhaust activates microglia through MAC1, inducing NOX2 activation, subsequently leading to production of reactive oxygen species [156]. Furthermore, microglia contribute to dopamine neurotoxicity in the brain [157], inducing a strong inflammatory response in the brain, with the midbrain showing the most pronounced inflammatory response. The striatum also exhibits a strong inflammatory response, as mice exposed to low concentrations of diesel exhaust during pregnancy showed decreased neurotransmitter turnover (an index of neuronal activity) of dopamine and noradrenaline in various regions of the CNS, including the striatum [158]. Additionally, α7nAChR are expressed in microglia [159] and may be considered as a target for treating microglia-mediated neuroinflammation.

Bioelectronic medicine

Electrochemical signals mediate signal transmission within the nervous system as well as between the nervous and immune systems, playing a role in neuroimmune interaction. Bioelectronic medicine harnesses and mimics neural bioelectrical signals associated with immune recognition and regulation. These therapeutic strategies include deep brain stimulators, spinal cord stimulators, vagus nerve stimulators (VNS) [160], sacral nerve stimulators, electroacupuncture and transcutaneous electrical nerve stimulation, among others. There are also instruments for neural signal recording, such as vagal electrical signals and sciatic nerve electrical signals. In addition, there are devices that combine stimuli and signal harvesting, such as brain–computer interfaces. A clinical trial involving 25 patients with asthma showed that transcutaneous VNS was associated with improvements in forced expiratory volume and perception of dyspnoea [161]. A clinical study involving two cases of transcutaneous VNS was conducted in COVID-19 patients, resulting in accelerated symptom recovery [162].

With the continuous development and in-depth research of bioelectronic technology, bioelectronic medicine can activate or inhibit disease-associated neurons, while simultaneously monitoring and collecting real-time respiratory and neural data from patients, followed by analysis through artificial intelligence. The collected data are utilised to refine electrical stimulation parameters, including the pH value [163], while also accounting for diverse individual responses to such stimulation, thereby enabling the customisation of bioelectronic medicine. Therefore, by recording signals in real time, analysing and adjusting the parameters of electrical stimulation, and determining treatment end-points, we can achieve more precise and personalised medical interventions, forming a closed-loop bioelectronic treatment.

Meditative breathing

Breathing can be classified as voluntary breathing and involuntary breathing. Involuntary breathing is regulated and controlled by a group of neurons in the brainstem, including the NTS, nucleus ambiguus, Bötzinger complex, pre-Bötzinger complex, post-inspiratory complex and locus coeruleus [164]. Meditation breathing is a voluntary breathing technique that improves emotions and releases stress through slow, deep breathing and focused attention [165]. It is characterised by reduced oxygen consumption, slowed heart rate, lowered blood pressure and increased θ-wave amplitude upon electroencephalogram [166]. Other related meditative movements include qigong, tai chi and yoga. Exogenous stimuli (such as drugs or bioelectronic medicine) can regulate the autonomic nervous system (ANS) and thus modulate the body's immune response. Meditation breathing has also been found to regulate the ANS, activating the sympathetic nervous system and mediating the release of catecholamines/cortisol. It reduces the levels of the pro-inflammatory mediators TNF-α, IL-6 and IL-8 induced by endotoxins, and increases plasma levels of IL-10 [167, 168]. Moreover, meditation breathing affects the brain, as evidenced by changes in scalp topographies and source analyses (variable resolution electromagnetic tomography), indicating relevant changes in neural sources in the left medial and lateral occipitotemporal areas, as well as the right lateral occipitotemporal and inferior temporal areas [169]. Additionally, synchronisation between respiration and locus coeruleus activity has been observed through functional magnetic resonance imaging and pupil dilation [170]. Meditation breathing alters nasal airflow, which is recognised by olfactory sensory neurons, triggering responses in the olfactory bulb [171]. The neurons (mitral cells) in the olfactory bulb can directly sense cardiovascular pressure through mechanosensitive ion channels (piezo), thereby regulating the activity of central neurons [172]. This may be part of the mechanism through which meditation breathing regulates physiological functions.

Meditation breathing has been applied as an adjunct therapy in clinical trials for pulmonary diseases such as COPD, COVID-19 and asthma [41, 173–175]. It has shown potential and positive effects on diseases in some studies. Therefore, further exploration of the mechanisms of meditation breathing in treating diseases and its use as preventive and adjunctive therapy is warranted.