It is unclear how cholinergic signaling regulates pulmonary immune responses during type 2-dominated airway inflammation. We have recently shown that ILC2s are significant producers of ACh following alarmin-mediated activation in the context of type 2 immunity against parasitic nematode infection and in response to A. alternata extract (4). During Alternaria-induced inflammation, serine protease activity releases IL-33 from airway epithelial cells (15), which activates many immune cell types, including ST2⁺ ILC2s. ILC2s are central mediators of type 2 inflammation, including eosinophilia (16), and we therefore explored whether ILC2-derived ACh was required for the generation of this response.

Administration of Alternaria to Rora ^Cre+ Chat ^LoxP animals resulted in increased numbers of CD45⁺ cells in the lungs 24 h post-exposure ( Figure 1A ). The numbers of eosinophils and neutrophils in the lung tissue at baseline were comparable between Rora ^Cre+ Chat ^LoxP, Chat ^LoxP, and C57BL/6J (WT) mice ( Figures 1B, C ). Exposure of Rora ^Cre+ ChatLoxP and Chat ^LoxP mice to Alternaria resulted in a comparable eosinophil influx to lung tissue ( Figure 1B ). Although Rora ^Cre+ Chat ^LoxP animals showed a trend toward increased numbers of neutrophils in the lungs, this was not significantly different from control groups at this timepoint ( Figure 1C ). Histologically, lung tissue appeared more inflamed in Alternaria-treated Rora ^Cre+ Chat ^LoxP and Chat ^LoxP mice relative to untreated controls ( Figure 1D ). In the airways, alveolar macrophages (AM) were the dominant cell type in WT control mice, whereas significantly increased eosinophils and neutrophils were observed in Alternaria exposed Rora ^Cre+ Chat ^LoxP and Chat ^LoxP genotypes ( Supplemental Figure 1 and Figures 1E, F ) as expected (14, 17). Airway eosinophil numbers were comparable between Alternaria-exposed genotypes ( Figure 1E ), but more neutrophils were observed in Rora ^Cre+ Chat ^LoxP mice ( Figure 1F ), indicating a role for ILC2-derived ACh in regulating neutrophil influx following exposure to Alternaria.

Analysis of transcript expression for chemokines involved in granulocyte trafficking revealed that eotaxin-1 (Ccl11) was unaffected in Rora ^Cre+ Chat ^LoxP and Chat ^LoxP genotypes relative to WT mice and was also unaffected by exposure to Alternaria ( Figure 1G ). Transcripts for the neutrophil-attractant chemokines CXCL1/CXCL2 were also similar in untreated Rora ^Cre+ Chat ^LoxP and Chat ^LoxP mice but were significantly elevated in Rora ^Cre+ Chat ^LoxP mice following exposure to Alternaria ( Figure 1G ).

Proinflammatory macrophages with a classical ‘M1’ profile are key producers of neutrophil chemoattractants including CXCL1 and CXCL2, whereas alternatively activated or ‘M2’ macrophages do not characteristically express these molecules (18, 19). Expression of the M1 marker Nos2 (inducible nitric oxide synthase, iNOS) was increased in the total lung tissue of Rora ^Cre+ Chat ^LoxP mice, while the M2 markers Mrc1 (mannose receptor C-type 1, CD206) and Arg1 (arginase 1) were downregulated following Alternaria exposure ( Figure 1H ). These data suggest that increased neutrophil infiltration following Alternaria exposure might result from an M1 macrophage bias in the absence of ILC2-derived ACh.

Analysis of IL-5 and IL-13 in bronchoalveolar lavage (BAL) fluid confirmed that dosing with Alternaria promoted a type 2 immune response. However, removal of the capacity of ILC2s to synthesize ACh did not impact the overall type 2 cytokine environment of the airways ( Figures 2A, B ). Expression of IL-5 and IL-13 ( Supplemental Figure 2A , Figures 2C–H ) and the activation markers ICOS or ST2 were not further altered in pulmonary ILC2s from Rora ^Cre+ Chat ^LoxP mice ( Figures 2I, J ). Lung CD4⁺ T cells were also analysed as they are also a potential source of type 2 cytokines, can express RORα and are known producers of ACh (20). However, as expected given the acute nature of the model, expression of IL-5 and IL-13 by total CD4⁺ T cells ( Supplemental Figures 2B–G ) was extremely limited, and no differences between the genotypes were observed. A small proportion of total CD4⁺ T cells displayed a Th2 phenotype (Gata3⁺ST2⁺) which was unaltered in the lungs of Alternaria-treated mice ( Supplemental Figures 2B, H ), but again there were no significant differences between Rora ^Cre+ Chat ^LoxP and Chat ^LoxP genotypes.

ILC2-derived ACh did not appear to greatly influence the onset of the type 2 immune response in the brief 24-hour timeframe of this model. However, ILC2s are only one cell type capable of ACh synthesis in the lung. We thus examined the effect of broader, non-selective depletion of ACh using a stable secretory acetylcholinesterase (AChE) from N. brasiliensis (21) and a catalytically inactive form of the enzyme generated by site-directed mutagenesis (22). The enzymatic activity of the preparations was confirmed by in-gel staining ( Supplemental Figure 3A ) and Ellman’s assay, which also confirmed that no inhibitors of AChE activity were present in the Alternaria extracts ( Supplemental Figure 3B ). Given our observations in Rora ^Cre+ Chat ^LoxP mice, we were particularly interested in examining the neutrophilic response following ACh depletion. Therefore, we used BALB/c mice, which have a dominant early neutrophilic response to Alternaria exposure compared with C57BL/6 mice, which are more eosinophil-dominant (17). The enzymes were co-administered intranasally with Alternaria, as shown in Supplemental Figure 4A .

Intranasal administration of Alternaria to mice resulted in a moderate influx of neutrophils ( Figures 3A–D ) and pronounced eosinophilia in the lungs and airways after 48 hours, as anticipated ( Figures 3E–G ) (15, 16). When Alternaria was co-administered with enzymatically active AChE, elevated numbers of neutrophils were observed in the lungs ( Figures 3A–D ), but eosinophilia was strikingly reduced in both sites ( Figures 3E–G ). Inactive AChE had no effect on eosinophil or neutrophil numbers, indicating that the effects of the active enzyme were due to hydrolysis of ACh ( Figures 3B–G ). Treatment with active AChE alone also resulted in significantly increased numbers of neutrophils ( Figures 3B–D ) and a small increase in eosinophils ( Figures 3E–G ) relative to PBS and inactive AChE baseline controls. Of note, eosinophil numbers in active AChE-treated mice at baseline were lower than those in Alternaria + PBS/inactive AChE-exposed tissues.

Granulocytes in the lungs also showed some phenotypic alterations following the depletion of ACh ( Figures 3H, I and Supplemental Figures 4B–G ). Eosinophils in the lung tissue, but not the airways, of Alternaria-treated mice demonstrated enhanced expression of Siglec-F, typical of recruited inflammatory cells (23), and this was further enhanced by co-administration of active but not inactive AChE ( Figures 3H, I ). Exposure to active AChE alone also slightly enhanced eosinophil Siglec-F expression, but to a lesser extent than following Alternaria exposure ( Figure 3H ). Depletion of ACh resulted in reduced CD11b (Integrin alpha M) expression on the surface of eosinophils and neutrophils ( Supplemental Figures 4B–E ), wheareas expression of Gr-1 (Ly6C/Ly6G antigen) on airway neutrophils was enhanced by either Alternaria or depletion of ACh ( Supplemental Figures 4F ).

Given the striking alterations to inflammatory cell influx caused by the depletion of ACh, we investigated whether this was accompanied by an altered chemokine environment. The level of CXCL1 in lung extracts was slightly increased by exposure to active AChE alone but was greatly enhanced by Alternaria + active AChE ( Figure 3J ). CXCL2 was not affected by active AChE alone but was considerably enhanced by co-administration of Alternaria and active AChE ( Figure 3K ). In contrast, although eotaxin-1 was only elevated following exposure to Alternaria, its maximal level was unaffected by ACh depletion ( Figure 3L ). In all cases, co-administration of an inactive enzyme with Alternaria had no significant effect on chemokine levels in the pulmonary tissues.

As previous studies have indicated that endothelial cell adhesion molecules could be downregulated in response to cholinergic agonists during inflammation (24, 25), we also assessed whether hydrolysis of airway ACh affected their expression. Administration of Alternaria induced pronounced upregulation of E-Selectin (Sele) ( Figure 3M ) and modest increases in Intercellular adhesion molecule 1 (Icam1) ( Figure 3N ) and Vascular cell adhesion molecule 1 (Vcam1) ( Figure 3O ) transcripts in lung tissue, but these were not further affected by co-administration of either active or inactive AChE. A small increase in Sele was also observed when active AChE was administered alone, in the absence of Alternaria ( Figure 3M ).

As anticipated, administration of Alternaria resulted in increased production of IL-5 and IL-13 by total unseparated cells from lung tissue, but this was significantly reduced following non-selective depletion of airway ACh ( Figures 4A, B ). Active AChE treatment alone did not alter total IL-5 production, although IL-13 was increased ( Figure 4B ). Although ILC2-derived ACh did not appear to impact on these changes ( Figure 2 ), we reasoned that ILC2 activity may still be dependent directly or indirectly on cholinergic signaling and therefore investigated this in more detail. Expression of IL-5 and IL-13 was increased by pulmonary ILC2s in response to Alternaria, with a significantly higher proportion of ILC2s expressing the cytokines and higher per-cell expression levels ( Figures 4C–H ). This was strongly suppressed by active but not inactive AChE ( Figures 4C–H ). Conversely, treatment with active AChE at baseline had the opposite effect, increasing IL-5 and IL-13 production by lung ILC2s relative to PBS and inactive AChE-treated controls ( Figure 4C–H ). Cytokine production by ILC2s reflected the activation status of the cells, as measured by ST2 and ICOS expression, which were enhanced in Alternaria-treated mice alone or with inactive AChE, and in active AChE only treated mice, but restricted following depletion of ACh with active enzyme in the context of Alternaria exposure ( Figures 4I, J ).

Both non-selective depletion of ACh and loss of ILC2-specific capacity to synthesize ACh enhanced neutrophil chemoattractants and cellular influx into the lungs following exposure to Alternaria, indicating that neutrophil trafficking is sensitive to suppression by cholinergic signaling. We therefore examined the M1/M2 phenotypes of pulmonary macrophages following enzymatic depletion of ACh. We analyzed three macrophage populations: monocyte-derived airway macrophages (Mo-AMΦ), tissue resident alveolar macrophages (TR-AMΦ), and interstitial macrophages (IMΦ) in lung tissue ( Supplemental Figures 5A–E ).

Mice exposed to Alternaria alone or Alternaria + inactive AChE showed enhanced frequencies and numbers of Mo-AMΦ ( Figures 5A–B ), no alterations in overall numbers of TR-AMΦ, but reduced frequencies of these cells due to greatly enhanced neutrophil and eosinophil influx, ( Figures 5C, D ), and increased frequencies and numbers of total IMΦ in the lung ( Figures 5E, F ). Alternaria + active AChE treatment did not alter total Mo-AMΦ ( Figures 5A, B ) but reduced total TR-AMΦ ( Figures 5C, D ) and resulted in a non-significant trend toward increased lung IMΦ ( Figures 5E, F ). Treatment with inactive AChE alone did not alter macrophage populations relative to PBS only treated controls, however active AChE alone resulted in significantly more Mo-AMΦ ( Figure 5B ), less TR-AMΦ ( Figure 5D ) and more IMΦ ( Figure 5E ). Notably, Alternaria + active AChE treatment resulted in significantly less TR-AMΦ ( Figure 5D ) but more IMΦ ( Figure 5E ), compared to the active AChE alone.

A characterization of macrophage polarization phenotypes revealed that Alternaria exposure combined with ACh depletion reduced the numbers expressing mannose receptor c-type 1 (CD206) ( Figures 5G–L ), a specific indicator of the M2 phenotype (26). Conversely, active AChE treatment alone showed a tendency towards increased numbers of M2 Mo-AMΦ ( Figure 5H ) and IMΦ ( Figure 5L ) relative to PBS and inactive AChE-treated controls. To investigate whether this was correlated with enhancement of the M1 phenotype, we gated on CD206^-MHCII^hi macrophages ( Supplemental Figure 5E ). Active AChE alone, and Alternaria treatment + PBS or inactive AChE increased numbers of M1 Mo-AMΦ, while active AChE treatment and Alternaria co-exposure decreased M1 Mo-AM and TR-AMΦ relative to Alternaria exposed controls ( Figures 5M–P ). However, M1-like IMΦ were increased significantly in Alternaria + active AChE treated mice, relative to increases observed in Alternaria + PBS and inactive AChE controls as well as active AChE only treated animals ( Figures 5Q, R ).

Cumulatively, these data indicate that cholinergic signaling is an important and complex regulator of immune responses, both at baseline during homeostasis and during fungal allergen-induced, allergic inflammation in the murine lung.

This study was approved by the Animal Welfare Ethical Review Board at Imperial College London and was licensed by and performed under the UK Home Office Animals (Scientific Procedures) Act Personal Project Licence number 70/8193: ‘Immunomodulation by helminth parasites’. C57BL/6J and BALB/c mice, aged 6–8 weeks old, were purchased from Charles River. The Rora ^Cre+ Chat ^Loxp mice used in this study were generated as previously described (4).

Extracts of A. alternata were purchased as lyophilized protein extracts from Greer Laboratories (USA). Mice were anesthetized with aerosolized isofluorane before intranasal administration with 50 μg of Alternaria in a final volume of 50 µl of phosphate buffered saline without Ca²⁺ or Mg²⁺ (PBS, Sigma). Mice were exposed to a single dose of Alternaria for 24 or 48 h as indicated. Control animals were dosed with 50 μl of PBS on the same schedule. For co-administration of active or inactive AChE, 20 μg of either enzyme was mixed with 50 μg of Alternaria in PBS and the volume adjusted to 50 µl for the first dose, and 20 μg of the enzyme alone was administered in 50 μl of PBS for the second dose.

N. brasiliensis AChE B was expressed in Pichia pastoris as a secreted protein and purified from culture supernatants as previously described (21). An enzymatically inactive form of the enzyme was generated via site-directed mutagenesis, changing the active site serine residue Ser-193 (Ser-200 in Torpedo AChE) to alanine (S193A), using the Quickchange site-directed mutagenesis kit (Stratagene) as previously described (22). The mutation was confirmed by sequencing before expression in P. pastoris, and purification confirmed that the enzyme was catalytically inactive. Proteins were passed through endotoxin removal columns (Pierce) and endotoxin removal was confirmed using a LAL Chromogenic Endotoxin Quantitation Kit (Pierce). Protein concentrations were determined using the Pierce Coomassie Plus (Bradford) Assay Kit (Thermo Scientific). AChE activity was determined by the method of Ellman with 1 mM acetylthiocholine iodide as the substrate in the presence of 1 mM 5,5’-dithiobis(2-nitrobenzoic acid) (DTNB) in 100 mM sodium phosphate pH 7.0 at 20°C. The reaction was monitored by measuring the absorbance at 412 nm, and the hydrolysis of acetylthiocholine iodide was calculated from the extinction coefficient of DTNB (53). One unit of AChE was defined as 1 μmol of substrate hydrolyzed per min at 20°C.

Purified recombinant active and inactive enzymes were resolved by SDS-PAGE on 10% polyacrylamide gels followed by staining with Coomassie brilliant blue. The same preparations were resolved under non-denaturing conditions by electrophoresis in 7.5% polyacrylamide gels in Tris-Borate-EDTA buffer pH 8.0, and enzyme activity was assayed using the method of Karnovsky and Roots (54).

For isolation of bronchoalveolar cells, the lungs were lavaged twice via the trachea in 2 ml of PBS with 0.2% BSA and 2 mM EDTA. Erythrocytes were lysed, leukocytes resuspended and counted. For lung single cell suspensions, lungs were perfused with PBS via injection into the heart before harvest, and lung leukocytes were isolated by dicing lung tissue and digesting for 1 h at 37°C with 300 U ml⁻¹ of collagenase-II (Gibco) + 150 mg ml⁻¹ of DNAse I (Sigma) in PBS without Mg²⁺ or Ca²⁺, followed by sample mashing through a 70 µm cell strainer into single-cell suspension, followed by red blood cell lysis.

Single cell suspensions were stained with fixable viability dyes (Invitrogen), then treated with rat anti-mouse CD32/CD16 (FcBlock, BD Biosciences), washed, and stained for extracellular markers using fluorophore conjugated monoclonal antibodies (eBioscience, Miltenyi Biotec, or Biolegend). For intracellular staining, cells were fixed for 30 min at room temperature, then permeabilized using the FoxP3/transcription factor staining buffer kit (eBioscience) and stained with fluorochrome-conjugated antibodies. Unstained samples and fluorescence minus one control were used as appropriate. Samples were analysed on a BD LSR Fortessa™ analyzer.

Unless otherwise stated, leucocyte populations were identified by flow cytometry by gating live cells, followed by single cell and CD45⁺ gating, and then using the following markers: ILC2: CD45⁺Lineage^-CD127⁺GATA3^hiICOS⁺ST2⁺; CD4⁺ T cells: CD3⁺CD4⁺; neutrophils: CD11b⁺Siglec-F^-GR-1^hiCD11c^−/lo; eosinophils: CD11b⁺Siglec-F⁺Gr-1^−/loCD11c⁻; tissue resident alveolar macrophages (TR-AMΦ): CD11b⁻F4/80⁺Siglec-F⁺CD11c⁺Gr-1⁻; monocyte-derived alveolar macrophages (MO-AMΦ): CD11b⁺F4/80⁺Siglec-F⁺CD11c⁺Gr-1⁻; interstitial lung macrophages (IMΦ): CD11b⁺F4/80⁺Siglec-F^-CD11c^−/+Gr-1⁻. CD206 was used as a marker of alternatively activated M2 macrophage subsets. Absence of CD206 expression (CD206⁻) and positive expression of MHCII (MHCII^hi) were used to identify M1 macrophage subsets. The lineage panel consisted of antibodies against CD3, CD4, CD8, CD5, B220, CD19, TER119, CD49b, FcϵRI, CD11c, and Gr-1 (Ly6C/Ly6G). Scatter profiling of gated myeloid populations was also used to validate identity, as shown in Supplemental Figure 1 .

Single cell suspensions were diluted to 5× 10⁶ cells ml⁻¹ in cDMEM (DMEM + 10% FCS, + 2 mM L-glutamine + 100 U ml⁻¹ penicillin + 100 ug ml⁻¹ streptomycin) and either stimulated for 4 h at 37°C/5% CO2 with 1 ug ml⁻¹ PMA/100 ng ml⁻¹ ionomycin with 1× brefeldin-A (GolgiPlug, BD Biosciences) + 1 μM monensin ((Sigma) or left unstimulated (golgi inhibitors alone). Samples were stained, fixed, and permeabilized as described, and intracellular staining for Fc receptor blocking followed by fluorophore conjugated mAbs against IL-5 and IL-13 was performed in permeabilization buffer.

Total lung tissue was homogenized using a Tissuelyser II (Qiagen). Total RNA was extracted by TRIzol/chloroform phase-separation, DNAse-1 treated, then 1 μg of RNA was reverse transcribed using the iScript cDNA synthesis kit (Biorad). RT-qPCR reactions were carried out using the PowerUp SYBR Green Mix (ThermoFisher) in an ABI 7500 Fast Real-time PCR thermocycler (Applied Biosystems). RT-qPCR reactions were run in triplicate, with no template and no RT controls. Relative expression of Ccl11, Cxcl1, and Cxcl2 was calculated by the comparative cycle threshold (Ct) method (2^−ΔΔCT) using Actb, Hprt, and Gapdh as reference genes. Eef2 and Ppia were used as the reference genes for calculating Icam1, Vcam1, and Sele expression. Primer sequences used for RT-qPCR were as follows:

Nos2: F: 5’- CCGGCAAACCCAAGGTCTAC-3’, R: 5’- CTGCTCCTCGCTCAAGTTCA-3’.

Arg1: F: 5’-AAAGGCCGATTCACCTGAGC -3’, R: 5’- CTGAAAGGAGCCCTGTCTTGTA-3’.

Mrc1: F: 5’-GGAGGGTGCGGTACACTAAC-3’, R: 5’-TCAGTAGCAGGGATTTCGTCTG-3’.

Ccl11: >F: 5’-TGGCTCACCCAGGCTCCATC-3’, R: 5’-TCTCTTTGCCCAACCTGGTCTT-3’.

Cxcl1: >F: 5’-ACCCAAACCGAAGTCATAGCCA -3’, R: 5’-TCAGAAGCCAGCGTTCACCA-3’.

Cxcl2: >F: 5’-TCCAAAAGATACTGAACAAAGGCAA-3’, R: 5’-ATCAGGTACGATCCAGGCTTCC-3’.

Icam1: F: 5’-AGCTCGGAGGATCACAAACG-3’, R: 5’- TCCAGCCGAGGACCATACAG-3’.

Sele: F: 5’-CCCAGTGCTTCTGGACCTTT-3’, R: 5’- TTCACAGCTGAACACGTGGG-3’.

Vcam1: F: 5’-CGACCTTCATCCCCACCATT-3’, R: 5’-GGGGGCAACGTTGACATAAAG-3’.

Eef2: F: 5’-CCCCAACATTCTCACCGACA-3’, R: 5’-AGAGAGCGCCCTCCTTAGTA-3’

Ppia: F: 5’- GCATACAGGTCCTGGCATCT-3’, R: 5’- ATGCTTGCCATCCAGCCATT-3’

Gapdh: >F: 5’-GTCATCCCAGAGCTGAACGG-3’, R: 5’-TACTTGGCAGGTTTCTCCAGG-3’.

Actb: >F: 5’-TTCCTTCTTGGGTATGGAATCCT-3’, R: 5’-TTTACGGATGTCAACGTCACAC-3’.

Hprt: F: 5’-ACAGGCCAGACTTTGTTGGA-3’, R: 5’-ACTTGCGCTCATCTTAGGCT-3’.

Protein lysates were made by snap freezing the small right lung lobe directly after tissue harvest. Harvested lung samples were stored at −80°C and homogenised in PBS with an electric homogeniser. Samples were centrifuged and the supernatant collected. Samples were diluted to a final concentration of 50 mg of lung tissue ml⁻¹ for use. Total lung cells were cultured at a density of 5 × 10^6/ml for 12 h with 12.5 ng/ml PMA and 125 ng/ml Ionomycin, before collecting culture supernatants and centrifuging to clear residual cells, followed by freezing at −20°C until analysis. Proteins were quantified using commercial ELISA kits according to the instructions of the manufacturer (Duoset, R&D). For CXCL1 and CXCL2 detection, results were obtained on a Tecan Sunrise 96-well Microplate reader F039300 with Tecan Magellan Anlaysis Software V7.2 software. For Eotaxin-1 (Ccl11), IL-5 and IL-13, results were obtained on a FLUOStar Omega microplate reader.

Lung tissue was harvested and fixed in 4% paraformaldehyde overnight at 4°C. Tissue was paraffin embedded, sectioned and stained with hematoxylin and eosin (H&E) according to standard techniques.

Flow cytometry data was analyzed using FlowJo software (Treestar). Graphs and statistical tests were carried out using Graphpad prism software (Graphpad). The normality of data distribution was analyzed by the Shapiro–Wilk test. Parametric data were analyzed by Welch’s t-test, non-parametric data were analyzed by Mann–Whitney-U test. Data represent mean ± SEM unless otherwise stated. Statistical significance between groups is indicated as *p <0.05, **p <0.01, ***p <0.001, ****p <0.0001, n.s = non-significant difference (p >0.05).