Nicotinic acetylcholine receptor agonist attenuates ILC2-dependent airway hyperreactivity

Allergic asthma is a complex and chronic inflammatory disorder that is associated with airway hyperreactivity (AHR) and driven by Th2 cytokine secretion. Type 2 innate lymphoid cells (ILC2s) produce large amounts of Th2 cytokines and contribute to the development of AHR. Here, we show that ILC2s express the α7-nicotinic acetylcholine receptor (α7nAChR), which is thought to have an anti-inflammatory role in several inflammatory diseases. We show that engagement of a specific agonist with α7nAChR on ILC2s reduces ILC2 effector function and represses ILC2-dependent AHR, while decreasing expression of ILC2 key transcription factor GATA-3 and critical inflammatory modulator NF-κB, and reducing phosphorylation of upstream kinase IKKα/β. Additionally, the specific α7nAChR agonist reduces cytokine production and AHR in a humanized ILC2 mouse model. Collectively, our data suggest that α7nAChR expressed by ILC2s is a potential therapeutic target for the treatment of ILC2-mediated asthma.

A sthma, which is a major worldwide health problem, is a chronic inflammatory disease of the airways with several phenotypes, comprised of both allergic and nonallergic asthma 1,2 . Allergic sensitization in which antigenpresenting cells (APCs) present allergens, followed by T-helper type 2 (Th2) cell skewing and eosinophilic inflammation, are essential for the development of allergic asthma. Obesity, ozone, viral infections, stress and air pollution are associated with nonallergic asthma, the pathogenesis of which involves the innate pathway rather than Th2 cell-mediated immunity [3][4][5] . Indeed, non-Th2 factors such as interferon-g, IL-17 and neutrophils are often found in the lungs of patients with severe non-atopic asthma 1,2 . Moreover, these allergic and non-allergic components may be present in individual patients to various degrees, leading to a complex immune milieu and disease heterogeneity 1,2 .
Innate lymphoid cells (ILCs) are a non-B cell, non-T-cell lymphocyte population in mucosal and lymphoid tissues that are not antigen specific, but respond rapidly to environment factors to induce various types of cytokines 6,7 . Among the ILCs, group 2 ILCs (ILC2s) are directly activated by innate signals from myeloid and epithelial-derived cytokines and alarmins, such as IL-25, IL-33 and proteases, without requiring further differentiation. Following activation, ILC2s produce robust amounts of Th2 cytokines IL-5 and IL-13 to promote eosinophilic inflammation and airway hyperreactivity (AHR); thus, they play an essential role in the pathogenesis of asthma 6,7 . The suggestion that ILC2s are critical for innate immunity activation in asthma is logical as influenza infection 5 and exposure to proteases and fungi 8,9 induce AHR by activating innate lymphoid cells. In addition, ILC2s participate in shaping and regulating adaptive immune responses 10 . ILC2-produced IL-5 and IL-13 also contribute to asthma development by respectively recruiting eosinophils in airways and inducing goblet cell mucus production. ILC2s can also directly stimulate a Th2 response in vitro 11 and facilitate an antigen-specific T-cell response during helminth infection 12 . This evidence suggests that ILC2s are deeply involved in the pathogenesis of asthma by modulating both innate and adaptive immune responses. Thus, regulating the function of ILC2s could be an ideal therapeutic strategy.
The a7 nicotinic acetylcholine receptor (nAChR), mediates rapid excitatory synaptic transmission and is shown to be a potential therapeutic target in neuropsychiatric 13 , neurodegenerative 14 and inflammatory disease 15,16 . In pulmonary allergic inflammation, nicotine, an agonist for nAChR, attenuates Th2 cytokine, IgE and cysteinyl leukotriene levels, resulting in reduced allergic inflammation 17,18 . However, the specific cellular mechanisms by which nAChR activation regulates allergic inflammation are not clearly defined. Given the therapeutic potential of a7nAChR in neurological disorders and inflammatory diseases, a variety of a7-specific ligands have been developed by medicinal chemistry 19,20 and structural approaches [21][22][23] . A number of these a7nAChR compounds showed promising efficacy and have advanced to clinical trials 19 . For the initial proof-of-concept studies presented here, we select a leading compound, GTS-21 (also known as DMXBA), that is functionally known to act as an agonist with partial selectivity toward a7nAChR 24,25 , and characterize its role in ILC2-dependent AHR. Utilizing this agonist, we demonstrate a critical role for a7nAChR in regulating ILC2-mediated AHR and airway inflammation in a preclinical model of allergic asthma. Moreover, we validate the specificity of the a7nAChR agonist using a7nAChR-deficient mice. As expected, the a7nAChR agonist was impaired in suppressing ILC2-dependent AHR in a7nAChR-deficient mice. Reconstitution studies with alymphoid mice suggest that while recipients of wild-type ILC2s responded to suppressive activity of the agonist, recipients of the a7nAChR-deficient ILC2s did not show any alteration in AHR or eosinophilia after agonist treatment. These observations establish that a7nAChR expression by ILC2s is crucial for the anti-inflammatory role of the agonist, and our reagent acts specifically through a7nAChR. Administration of the a7nAChR agonist inhibits activation of key transcriptional factors, such as GATA-3 and NF-kB. Furthermore, a7nAChR agonist abrogates phosphorylation of IKKa/b, a critical kinase upstream of the NF-kB signalling pathway. Using a translational approach, we also show that engagement of a7nAChR results in decreased cytokine production in human ILC2s and a reduction in AHR in a humanized ILC2 model. Our findings provide insight into the regulation of ILC2s that could ultimately be used to generate new therapeutic approaches for ILC2-dependent asthma.

Results
a7nAChR agonists and a7nAChR expression by ILC2s. Because of the potential therapeutic application in neurological disorders and inflammation, a7-selective ligands have been intensively studied. Li et al. 21 solved the crystal structure of a receptor chimera constructed from the extracellular domain of a7nAChR and acetylcholine binding protein (AChBP), which represents the closest structural homologue of the native a7nAChR. We have also determined structures of a7nAChR bound by agonist and antagonist 21,26 , and performed structure-based screens of a7nAChR-specific compounds 22 . Our studies, together with those reported by others, have demonstrated that the crystal structure of the a7nAChR/AChBP chimera is a much better template for structure-based drug screening than AChBP 23 , which has been widely used in previous drug screening and design. These studies have identified a variety of potential ligands and allosteric modulators of a7nAChR, including many previously known a7-specific compounds. Of these, GTS-21 (DMXB-A) has been previously characterized as a neuronal nAChR ligand. GTS-21 binds to both the a4b2 and a7 subtypes 24,25,27,28 , but activates only the a7 subtype to a significant extent. In our structure-based drug screen, we selected 4-OH-GTS-21, the active form of GTS-21 in vivo, from the docking analyses. As shown in Fig. 1a, the overall structure is similar to that of 4-OH-GTS-21 bound to AChBP 29 , including the conformation and orientation of 4-OH-GTS-21 bound to the ligand pocket. On the other hand, our docking analyses reveal that a number of a7-specific residues located at loop C (Arg182 and Glu185) and loop F (Glu158 and Asp160), which are absent in AChBP, interact, or are poised to interact, with functional groups on 4-OH-GTS-21. These observations provide structural bases to guide further synthetic modification of GTS-21 to gain higher selectivity toward the a7 subtype and to modify the functional effects of similar ligands, thus expanding the therapeutic repertoire of agents based on the anabaseine scaffold.
The a7nAChR is present on macrophages 15 , T cells 30,31 and B cells 32 , and activation of this receptor mediates anti-inflammatory effects. However, there are no studies investigating a7nAChR expression on ILC2s, despite their essential role in the development of asthma. We identified ILC2s as a Lin -, CD25 þ , CD45 þ , CD90.2 þ , and ST2 þ population (Fig. 1b). As mentioned previously, ILC2s are activated by IL-33; however, they are not the only cells that express the IL-33 receptor. As GTS-21 binds to both a4 and a7 subtypes of nAChR, we further assessed mRNA levels of a7 and a4nAChR in various types of immune cells. Interestingly, a7nAChR expression was significantly higher in ILC2s as compared with other ST2 þ peripheral immune cells (Fig. 1c). Importantly, only a7nAChR in ILC2s, but not a4nAChR, was significantly present and up-regulated by in vitro IL-33 treatment (Fig. 1c,d). To confirm this finding at the protein level and also to investigate the effect of IL-25 stimulation, we treated mice with intranasal recombinant mouse (rm)-IL-33, (rm)-IL-25, or PBS as a negative control, for three consecutive days. As shown in Fig. 1e, we found, for the first time, that a7nAChR is expressed on ILC2s. Importantly, a7nAChR expression on ILC2s was significantly upregulated in mice treated with IL-25 or IL-33, when compared with the PBS control group. a7nAChR expression was additionally confirmed by cytometry using fluorochromeconjugated a-bungarotoxin, a nicotinic cholinergic blocker ( Supplementary Fig. 1). Meanwhile, a7nAChR did not alter expression of CD25, CD127 and ST2, known as IL-2R, IL-7R and IL-33R, respectively, and which are essential for development of immune cells (Supplementary Fig. 2).
a7nAChR agonist suppresses cytokine production in ILC2s. Nicotine is known as a major constituent of cigarette smoke, which causes impairment of lung function and exacerbation of asthma. Interestingly, nicotine administration attenuates production of Th2 cytokines and leukotrienes in preclinical models of asthma 17 . Nicotine is an agonist for a variety of pentameric nAChRs made up of different combinations of the sixteen nicotinic receptor subunits, thus it lacks specificity for a7nAChR 21,33 . Therefore, using GTS-21, an agonist specific for a7nAChR, we sought to determine whether engagement of a7nAChR would alter ILC2s' function. To address this question quantitatively, pulmonary ILC2s were isolated and cultured with increasing doses (2.5, 10 and 50 mg ml À 1 ) of a7nAChR agonist in the presence of rm-IL-33, rm-IL-2 and rm-IL-7. Our results show dose-dependent anti-inflammatory effects of a7nAChR agonist on ILC2s (Fig. 2). These effects were independent of cell viability except at the highest tested dose of the agonist (Fig. 2a,b, right panels). To verify that the actions of the agonist were not due to a reduction in the number of ILC2s, we administered IL-33 to mice in vivo with or without agonist treatment, and then quantified the number of IL-5 þ and IL-13 þ ILC2s within a determined number of ILC2s (Fig. 2c). Similarly to IL-33, IL-25 increased IL-5 and IL-13 secretion in ILC2s in vitro, though this effect was less robust. IL-25-induced cytokine secretion was also susceptible to suppression by the agonist (Supplementary Fig. 3).
As reported previously [6][7][8][9]11,12,34,35 , the IL-5 and IL-13 cytokines produced by activated ILC2s are essential for eosinophilic inflammation and AHR development. We investigated whether the attenuated ILC2 function by a7nAChR stimulation in vitro could result in inhibition of ILC2-mediated AHR and allergic inflammation. Rag2 deficient mice (devoid of T and B cells) were given intranasal (i.n.) rm-IL-33, with or without a7nAChR agonist for three consecutive days (Fig. 3a). As IL-33 administration specifically induces ILC2s, thereby causing AHR, with this model we can readily explore the effect of an a7nAChR agonist in ILC2-mediated AHR. One day after the last challenge, lung function was evaluated by direct measurements of lung resistance (R L ) and dynamic compliance (C dyn ), as described in Methods section. We found that stimulating a7nAChR significantly reduced levels of R L and of C dyn in response to IL-33, as compared with PBS (Fig. 3b,c). Bronchoalveolar lavage fluid (BALF) analyses also showed decreased eosinophilic infiltration, as well as total cell counts, in Rag2 À / À mice treated with the a7nAChR agonist (Fig. 3d). Histological analyses revealed that a7nAChR stimulation prevented airway wall thickness and infiltrated cells (Fig. 3e,f). As shown in Fig. 3g, treatment with GTS-21 also significantly suppressed the frequency and absolute number of lung ILC2s. Furthermore, the intracellular cytokine assay also revealed significantly decreased levels of IL-5 and IL-13 producing lung ILC2s in a7nAChR agonist treated mice compared with untreated mice (Fig. 3h). These findings concur with the reduction of AHR and eosinophil counts in BALF ( Fig. 3b-d). Thus, a7nAChR agonist represses IL-5 and IL-13 production, eosinophil recruitment, and ILC2-dependent AHR.
a7nAChR on ILC2s is critical for GTS-21 action in vivo. To assess whether the previously observed effects of GTS-21 are due to its actions on the a7nAChR, we challenged WT and a7nAChR-deficient mice with or without rm-IL-33, and with or without a7nAChR agonist, for three consecutive days (Fig. 4a,b). As expected, in the WT mice, the agonist repressed IL-33-induced AHR and eosinophilic infiltration. However, in the absence of a7nAChR, the agonist affected neither AHR nor eosinophilia. Taken together, these results indicate that engagement of a7nAChR ameliorates ILC2-mediated AHR and allergic inflammation.
Given that a variety of cells participate in allergic inflammation, we wanted to investigate the effect of the agonist specifically on ILC2s. Rag2 À / À GC À / À mice lack not only B and T cells, but NK cells and ILC2s as well. Using methods previously described by our laboratory 36 , we adoptively transferred WT or a7nAChR-deficient ILC2s into these Rag2 À / À GC À / À mice, and then treated them with IL-33, with or without a7nAChR agonist. As expected, we observed a decrease of AHR and of eosinophil recruitment in mice injected with WT ILC2s in response to a7nAChR treatment ( Fig. 4c-e). However, in the absence of a7nAChR expression on ILC2s, the agonist affected neither AHR nor eosinophilia. Taken together, these results indicate that engagement of a7nAChR ameliorates ILC2-mediated AHR and allergic inflammation. These results, which are consistent with those seen in WT and a7nAChRdeficient mice, demonstrate that the effects we observed are due to ILC2s as opposed to other cells, such as structural cells, which express a7nAChR. a7nAChR agonist attenuates GATA3 and NF-jB expression.
To characterize the mechanism enabling a7nAChR agonist to attenuate ILC2-mediated AHR, we first investigated whether engagement of a7nAChR affects the development or maintenance of pulmonary ILC2s. However, we observed a significantly decreased proliferation rate of pulmonary ILC2s upon a7nAChR agonist treatment by evaluating Ki-67 levels (Fig. 5a). We also evaluated the expression of the transcription factor GATA binding protein-3 (GATA-3), which is essential for the development and maintenance of ILC2s 37-39 . We found that a7nAChR agonist treatment significantly inhibits GATA-3 transcription in ILC2s (Fig. 5b). This decrease in GATA-3 expression could be involved in the repressed proliferation and function of ILC2s in response to a7nAChR agonist. Next, to investigate the molecular mechanism of a7nAChR agonist-mediated anti-inflammatory effects, we analysed the gene expression profile of ILC2s with or without a7nAChR agonist treatment by NanoString technology as described in Methods  section. Evaluated genes were categorized and displayed in two panels; (1) genes mainly involved in the IL-33/IL-25 signalling pathway, and (2) genes associated with IL-2/IL-7 signalling pathway (Fig. 5c). As demonstrated in the IL-33/IL-25 signalling pathway, a7nAChR-mediated cholinergic activity inhibited expression of GATA-3 as well as STAT-6, NF-kB, IL-5, IL-9 and IL-13. Meanwhile, STAT-5a, STAT-5b and IL-2 receptors in the IL-2/IL-7 signalling panel were relatively unaffected. We also demonstrated that a7nAChR agonist did not activate factors associated with apoptosis, such as Casp3, Casp8, and Bcl-2. We further confirmed by flow cytometry that a7nAChR agonist reduced NF-kB p65 expression in ILC2s (Fig. 5d). To better establish the underlying mechanisms, we explored the signalling pathway upstream of NF-kB by measuring the activated form of IKKa/b, which is phosphorylated on serine residues 176 and 180. Consistent with our previous results shown in Fig. 5d, a7nAChR agonist reduced phosphorylated IKKa/b expression in ILC2s (Fig. 5e,f). NF-kB and STAT-6 are both critical for GATA-3 transcription. The inhibition of GATA-3 expression results in a down-regulation of IL-5, IL-9 and IL-13 secretion from ILC2s. Strikingly, these results demonstrate that a7nAChR stimulation likely reduces maintenance and development of ILC2s by inhibiting GATA-3 transcription, resulting in marked anti-inflammatory effects in the development of ILC2-mediated AHR and allergic inflammation.
It has been previously reported that Alternaria alternata can induce AHR 40 . We further explored the anti-inflammatory effects of a7nAChR agonist in ILC2-mediated AHR with this clinically relevant allergen. Rag2 À / À mice were i.n. administered A. alternata extract with or without a7nAChR agonist for 4 consecutive days (Fig. 6a). As expected, a7nAChR agonist treated mice did not develop AHR (Fig. 6b). Accordingly, the number of eosinophils in BALF and lung ILC2s were increased in Alternaria-treated mice. Meanwhile, a7nAChR agonist treatment abolished eosinophils and ILC2s in the lung (Fig. 6c). These results suggest a7nAChR agonist treatment can attenuate ILC2-mediated AHR in response to other allergens besides IL-33.

Discussion
In the present study, using a leading compound with good selectivity and partial agonist activity towards a7nAChR, we examined the involvement of a7nAChR in the pathogenesis of asthma to evaluate its potential as a therapeutic target. We demonstrated, for the first time, that a7nAChR is expressed on ILC2s and upregulated after engagement of the ST2 receptor by IL-33. By inhibiting GATA-3 transcription in ILC2, administration of a7nAChR agonist significantly attenuated the development and function of ILC2s, abolishing both AHR and allergic inflammation. Importantly, we showed this agonist also ameliorated human ILC2-mediated AHR, indicating its potential as a novel therapeutic molecule for asthma patients. a7nAChR is a neuronal subtype of nAChR composed of a homopentamer of a7 subunits that mediates pre-and post-synaptic excitation. It is known that a7nAChR is located not only in the brain but also in the periphery, including immune cells such as effector T cells 30,31 , B cells 32 , Tregs 41 and macrophages 15 . Using an improved template based on crystal structures, we performed a virtual screen for a7nAChR ligands and selected a leading compound, GTS-21, that has been previously characterized as a functional agonist of a7nAChR. Although GTS-21 had previously been shown to have anti-inflammatory effects 42,43 , its role in ILC2-dependent AHR has not been examined to our knowledge. In line with previous studies, we examined a newly identified lymphocyte population of ILC2s and showed that ILC2s expressed a7nAChR. Importantly, activated ILC2s upregulated a7nAChR expression more so than other immune cells, indicating that asthma patients having activated ILC2s might benefit greatly from a nicotinic agonist compound such as GTS-21.
Nicotine is just one of the over 4,000 chemical constituents in tobacco smoke. Exposure to cigarette smoking causes impaired lung function and increases the risk of developing asthma 44,45 . In asthma, patients who smoke have more symptoms and exacerbations than non-smokers 46 . They also have increased risks of hospitalization and mortality 47 . On the other hand, epidemiological studies have indicated inverse correlations between smoking and incidence of allergic diseases. It is well known that the incidence of hypersensitive pneumonitides such as farmer's lung and bird fancier's lung is lower in the current smoker population than that in non-smokers [48][49][50][51] . Furthermore, other clinical studies have observed that asthma incidence could be higher in former smoker populations compared with active smokers 44,52 . In that longitudinal study, the observed increased asthma risks in former smokers were explained by the fact that, in some cases, asthma was self-reported or the supposition that people tend to quit smoking in response to respiratory symptoms. A cohort study also showed that development of allergic sensitization is negatively associated with sustained smoking 53 . In accordance with these clinical findings, nicotine, a major constituent of cigarette smoke and a ligand for nAChR, has shown anti-inflammatory effects in various diseases 31,54,55 . In asthma, nicotine attenuated HDM-induced allergic lung inflammation together with suppression of Th2 cytokines, although the underlying cellular mechanism remains unknown and AHR was not suppressed 17 . Moreover Kearley et al. 56 recently demonstrated that cigarette smoke has suppressive effects on IL-5 and IL-13 production by ILC2s. In preclinical models of asthma, we clearly demonstrated engagement of a7nAChR opposes the development of AHR and allergic inflammation through an ILC2-mediated mechanism. The differences in outcome likely resulted from the experimental design; our group administered a specific compound for a7nAChR together with rm-IL-33 or A. alternata to mice, whilst Sopori et al. examined nicotine itself and HDM-induced AHR 21,33 . To establish that our previous observations were dependent of a7nAChR expression on ILC2s, we demonstrated that the a7nAChR agonist had no effect on IL-33-challenged mice that were deficient for a7nAChR on ILC2s. Collectively, our data can explain the cellular mechanisms behind the clinical anti-inflammatory observations.
We demonstrated that a7nAChR activation altered ILC2 function in response to both exogenous IL-33 and A. alternata. Administration of either IL-33 and IL-25 activates ILC2s to induce AHR independently of the adaptive immune system 5,6,57 , and IL-33 was reported to be more potent than IL-25 in this regard 57 . A. alternata exposure also activates ILC2s and causes steroid-resistant AHR associated with elevated IL-33 in vivo 40 . ILC2s rapidly produce huge amounts of IL5 and IL-13 in response to stimuli. A recent study showed that ILC2-derived IL-13 is capable of inducing differentiation among Th2 cells 35 . Based on these reports, regulating ILC2s could potentially control allergic inflammation arising from not only the innate, but also adaptive, immune pathway. By targeting ILC2 function, a7nAChR agonist treatments could prove remarkably therapeutic for various allergic diseases.
To explore the mechanisms underlying the anti-inflammatory effects of the a7nAChR agonist, we initially demonstrated the engagement of a7nAChR caused reduced number of ILC2s in the lung, suggesting involvement of the cholinergic signal in cell fate and maintenance. We next found that a7nAChR agonist significantly suppressed Ki67 expression, a cellular marker for proliferation in ILC2s. In contrast, expression of anti-apoptotic factor Bcl-2 on ILC2s was unaffected by a7nAChR stimulation. In accordance with these in vivo data, viability of ILC2s in culture and results from the NanoString assay were comparable. These results suggest that cholinergic signal transmission regulates proliferation, but not cell death, in ILC2s. Our findings are consistent with the cholinergic underpinnings of anti-inflammatory mechanisms in other diseases of inflammation, such as autoimmune arthritis and experimental autoimmune encephalomyelitis 31,54,55 .
On one hand, Dowling et al. 58 also described that nicotine could inhibit the NF-kB pathway in an a7nAChR-dependent manner. We demonstrated by cytometry that in response to the agonist, the expression of the activated NF-kB p65 subunit was reduced. Similarly we also observed a decrease in the expression at the mRNA level of NF-kB1. However, with the Nanostring Technology we also noticed that RelA expression was unaffected. This discrepancy could be explained by the fact that NF-kB is mainly regulated at the post-translational level by phosphorylation. Phosphorylation of the NF-kB p65 subunit on certain residues plays a key role in regulating NF-kB activation and function. We also validate our results by assessing IKKa/b, upstream of NF-kB and observed a significant reduction in IKKa/b after agonist treatment.
On the other hand, in cancer immunity, nicotine administration enhances tumour growth by promoting cell proliferation and suppressing apoptosis [59][60][61][62] . Strikingly, these results suggest that cholinergic signal is involved in the pathogenesis of various diseases, with a distinct role in each disease.
In addition to attenuated proliferation, a7nAChR agonist also inhibited GATA-3 expression in ILC2s. GATA-3, a double zinc-finger transcription factor, is required for the development of Th2 cells and important for the production of IL-5 and IL-13 (refs 63-65). Recently, GATA-3 was also reported to be essential for differentiation and maintenance of ILC2s, and their production of IL-5 and IL- 13 (refs 37-39). Moreover, higher expression of GATA-3 is associated with an increase in ILC2-derived IL-13 (ref. 66). Collectively, these results indicate an underlying mechanism for the anti-inflammatory role of a7nAChR agonist in asthma: modulating GATA-3 expression and proliferation in ILC2s, which subsequently attenuates Th2 cytokine production from ILC2s, preventing the development of AHR and allergic inflammation.
We further investigated whether this cholinergic antiinflammatory effect was relevant in a humanized mice model, in which human ILC2s were adoptively transferred to Rag À / À IL2rg À / À mice and administered IL-33 to induce AHR. This unique system allows one to directly evaluate human ILC2-mediated AHR 36 . In accordance with the effects seen in murine ILC2s, a7nAChR agonist significantly suppressed human ILC2 function and dampened human ILC2-mediated AHR. Taken in their entirety, our results suggest that enhancing cholinergic signal might be an effective therapeutic strategy for patients with ILC2-mediated asthma.
In conclusion, this is the first report to reveal that the engagement of a7nAChR on ILC2s suppresses AHR. Therefore, our results suggest a protective role of cholinergic signalling in the pathogenesis of asthma, and present a7nAChR agonists as novel therapeutic candidates for controlling ILC2-mediated inflammatory lung diseases.

Methods
Mice. Female BALB/cByJ, RAG2 deficient (C.B6(Cg)-Rag2 tm1.1Cgn /J) mice, RAG2 GC deficient (C; 129S4-Rag2 tm1.1Flv IL2rg tm1.1Flv /J) mice and a7nAChR deficient (B6.129S7-Chrna7 tm1Bay /J) (6-8 weeks old) were purchased from Jackson Laboratory (Bar Harbor, ME). Rag2-deficient and Rag2 GC-deficient mice were bred in our facility at the Keck School of Medicine, University of Southern California (USC). Animal studies were approved by the USC Institutional Animal Care and Use Committee and conducted in accordance with the USC Department of Animal Resources' guidelines. All human studies were approved by USC institutional review board and conducted according to the principles of the Declaration of Helsinki. Participants gave written informed consent to before their inclusion in the study, and were identified by number.
Crystallization and structure determination of agonist. The structure of the agonist 4-OH-GTS-21 was docked into the ligand site of the crystal structure of the a7nAChR chimera 21 using our previously published computation docking procedures 22 . The figure is made by PyMOL (The PyMOL Molecular Graphics System, v 1.8, Schrodinger, LLC).
Measurement of airway hyperreactivity. Mice were i.n. administered carrier-free recombinant mouse IL-33 (BioLegend, San Diego, CA, 0.5 mg per mouse in 50 ml) with or without 125 ml of a7nAChR agonist (kindly provided by Lin Chen) on 3 consecutive days. Mice were i.n. administered carrier-free recombinant mouse IL-25 (BioLegend, San Diego, CA, 5 mg per mouse in 50 ml) with or without 125 ml of a7nAChR agonist on three consecutive days. For Alternaria alternata experiments, mice were i.n. administered A. alternata (Greer Labs, Lenoir, NC, 100 mg per mouse in 50 ml) with or without 125 mg a7nAChR agonist for 4 consecutive days. One day after the last challenge, mice were anesthetized using i.p. injection of ketamine (10 mg ml À 1 ) and xylazine (1 mg ml À 1 ). Measurements of airway resistance and dynamic compliance were conducted with the Fine Pointe RC System (Buxco Research Systems, Wilmington, NC), in which mice were mechanically ventilated using a modified version as previously described 36,67 . Mice were sequentially challenged with aerosolized PBS (baseline), followed by increasing doses of methacholine. Maximum lung resistance (R L ) and minimum compliance (C dyn ) values were recorded during a 3-min period after each methacholine challenge.
BD Cytofix Fixation Buffer and BD Phosflow Perm Buffer III were purchased from BD Biosciences (San Jose, CA). Flow cytometry was carried out on the FACSCanto II and FACSARIA III (BD Biosciences) and the data were analysed with FlowJo version 8.6 software (TreeStar, Ashland, Oregon).
Intracellular staining. Intracellular staining was performed using BD Cytofix/ Cytoperm kit (BD Bioscience, San Jose, CA) according to the manufacturer's instructions. For analysis of GATA3 and Ki-67 expression, freshly isolated cells were fixed and permeabilized using Fixation/Permeabilization buffer kit (eBioscience) according to the manufacturer's instructions and as previously described 70 .
Gene expression analysis with NanoString nCounter technology. The difference in the abundance of scripts between ILC2s purified from rm-IL33 administered mice with or without a7nAChR agonist were analysed with NanoString nCounter technology. Heat plots were generated with R statistical software.
Statistical analysis. A student t-test was used for comparisons between each group. P values of o0.05 were considered significant. All data are expressed as the mean±s.e.m.
Illustrations. In Figs 3,4,6 and 7 and thumbnail picture we used vector elements from Servier Medical Art PowerPoints under a Creative Commons License.
Data availability. The authors declare that the data supporting the findings of this study are available within the article and its Supplementary Information files.