Cytokines reprogram airway sensory neurons in asthma

Nociceptor neurons play a crucial role in maintaining the body’s homeostasis by detecting and responding to potential dangers in the environment. However, this function can be detrimental during allergic reactions, since vagal nociceptors can contribute to immune cell infiltration, bronchial hypersensitivity, and mucus imbalance, in addition to causing pain and coughing. Despite this, the specific mechanisms by which nociceptors acquire pro-inflammatory characteristics during allergic reactions are not yet fully understood. In this study, we aimed to investigate the molecular profile of airway nociceptor neurons during allergic airway inflammation and identify the signals driving such reprogramming. Using retrograde tracing and lineage reporting, we identified a unique class of inflammatory vagal nociceptor neurons that exclusively innervate the airways. In the ovalbumin mouse model of airway inflammation, these neurons undergo significant reprogramming characterized by the upregulation of the NPY receptor Npy1r. A screening of cytokines and neurotrophins revealed that IL-1β, IL-13 and BDNF drive part of this reprogramming. IL-13 triggered Npy1r overexpression in nociceptors via the JAK/STAT6 pathway. In parallel, sympathetic neurons and macrophages release NPY in the bronchoalveolar fluid of asthmatic mice, which limits the excitability of nociceptor neurons. Single-cell RNA sequencing of lung immune cells has revealed that a cell-specific knockout of Npy1r in nociceptor neurons in asthmatic mice leads to an increase in airway inflammation mediated by T cells. Opposite findings were observed in asthmatic mice in which nociceptor neurons were chemically ablated. In summary, allergic airway inflammation reprograms airway nociceptor neurons to acquire a pro-inflammatory phenotype, while a compensatory mechanism involving NPY1R limits nociceptor neurons’ activity.

Next, we sought to test whether the airway nociceptor neuron transcriptome is impacted during allergic airway inflammation (AAI).Retrograde tracer-exposed nociceptor reporter mice (NaV1.8cre ::tdTomato fl/wt ) underwent the classic ovalbumin (OVA) model of asthma 18 .OVA-exposed mice showed significant airway inflammation characterized by eosinophilic infiltration in the bronchoalveolar fluid (Fig. 2A) but had a similar number of backlabeled airway nociceptor neurons (SF.4A).Simultaneously, we evaluate how mice exposed to OVA modulate the influx of immune cells into the BALF of allergic mice by using single-cell RNA sequencing on flow cytometrysorted lung CD45 + cells (Fig. 2B).The various cell types were identified using standard markers of lung immune cells, and include including B cells, T cells, Granulocytes (neutrophils and eosinophils), NK cells, Basophils, and Antigen Presenting Cells (APC, which include Macrophages and Dendritic cells) endothelial cells and alveolar macrophages (SF.4B-C).DEGs detailed in the ST. 2, indicate significant airway inflammation with drastic changes in lung immune cells gene expression profiles.

Cytokines trigger gene expression changes in nociceptors.
Since the transcriptome changes induced by AAI are restricted to airway nociceptors and are virtually absent in other visceral nociceptors and NaV1.8 -cells, these variations are likely triggered by mediators in the airways detected by peripheral nerve endings.To identify these neuromodulators, we exposed (24h) cultured JNC neurons to various allergy driving cytokines, inflammatory lipids, and neurotrophins, and used transcription changes to Npy1r, Sting1, Bdnf, and Il6 as proxies for an AAI-like signature (Fig. 2D).This screening approach revealed that IL-4 and IL-13 triggered the overexpression of Npy1r (Fig. 3A), TNF-α and IL-1β induced Il6 (Fig. 3B), Bdnf was triggered by the neurotrophin itself (Fig. 3C), while Sting1 was induced by IL-1β, BDNF, TNF-α and IgE mixed with its cognate antigen ovalbumin (Fig. 3D).The other cytokines and neurotrophins tested did not impact the expression of the tested genes.
Cytokines and neurotrophins are known to initiate transcription programs in their target cells 51,52 .We then exposed cultures of JNC from NaV1.8 cre ::tdTomato fl/wt reporter mice to IL-13, IL-1β, BDNF and TNF-α (24h), before purification of the nociceptor neurons by flow cytometry (tdTomato + ) and RNA-sequencing.DEGs were triggered in the four different conditions (Fig. 3E-I, ST. 4), with the most drastic changes induced by IL-1β and BDNF.
To compare reprogramming induced by AAI and the tested cytokines, a gene set enrichment analysis (GSEA) was performed using the most significantly overexpressed genes (FDR<0.2) in all tested conditions.These data revealed that the gene sets elevated in cultured JNC neurons exposed to IL-1β, IL-13, and BDNF, as well as the ones previously identified in injured neurons 53 , were enriched in JNC airway nociceptors of AAI mice (Fig. 3J, ST. 4).IL-13 signature showed the strongest enrichment (NES=2.0).Specifically, 11 DEGs identified in AAI neurons were then also induced in vitro by BDNF, and 9 genes induced by IL-1β, among which 4 were induced by these two proteins (Fig. 3K).IL-13 induced 2 genes identified in AAI, Npy1r and Serpina3i (Fig. 3K).Morphologically, BDNF influenced neurite outgrowth in cultured JNC nociceptors, whereas the tested cytokines did not have any effect (Fig. 3L).These results highlight the similarities and specificities between transcription programs inducible in nociceptor neurons, and suggest that IL-1β, BDNF and IL-13 signaling pathways are complementary to induce the AAI signature in our mouse model of asthma.IL-13 reprograms nociceptors through JAK/STAT6.While IL-1β and BDNF effects on nociceptors has been thoroughly investigated 6,50,[54][55][56][57] , the literature regarding IL-13 is scarcer 14,58 , and this interaction was not yet reported in vagal sensory neurons.We then aimed to delineate the mechanisms by which IL-13, a type 2 specific cytokine, activates nociceptor neurons and changes their transcriptome.IL-13 (24h) significantly affected the expression of 47 genes, including Npy1r (Fig. 3B, C, Fig. 4A, ST. 5).Since both IL-4 and IL-13 and had comparable effects in vitro (Fig. 3A), we used our transcriptomic data to identify which IL-13 and IL-4 receptors are expressed by nociceptor neurons and found that only the IL4RII subunits Il4ra and Il13ra1 are detected in these cells (SF.5A, ST. 6).Il13ra1 expression was higher in nociceptors than in Nav1.8 -cells (SF.5A) and was also found to be expressed in the inflammatory airway nociceptor cluster NN8 (Fig. 1E; SF. 2I), along with the signaling mediator Stat6 (SF.2J).
Since IL4RII is a shared receptor for IL-13 and IL-4 that signals via JAK1/2 and STAT6 to regulate immune cells' transcription 59 , we tested whether a similar mechanism was at play in nociceptor neurons.We found that IL-13 (30min) triggered STAT6 phosphorylation in JNC and DRG neuron cultures (Fig. 4B-D).In addition, IL-13-mediated induction of Npy1r was prevented by the STAT6 inhibitor AS1517499 as well as by the JAK1/2 inhibitor ruxolitinib (Fig. 4E).In the tested conditions, IL-13 did not induce calcium flux in nociceptor neurons (SF.5B).The effect of IL-13 on Npy1r expression in JNC nociceptors was also observed in cultured DRG neurons (SF.5C).To see if this pathway is responsible for Npy1r overexpression in vivo, we treated AAI mice with intranasal instillation of AS1517499 (150µg/50µL).AS1517499 prevented Npy1r induction in nociceptor neurons without affecting Il6 overexpression, confirming that the AAI molecular profile depends in part on pSTAT6 activity (Fig. 4F).
Sympathetic neurons and APCs produce NPY in the airways during asthma.
Npy1r thus appears as a major marker of AAI inflammation in nociceptor neurons.Since the optogenetic activation of NPY1R + JNC neurons trigger expiratory reflexes in mice 1 and NPY1R activation impacts pain and itching reflexes [60][61][62][63][64][65][66][67][68] , we sought to test whether Neuropeptide Y (NPY)/NPY1R interaction could occur in the lung.Along with airway inflammation (Fig. 5A-B), allergen challenges (day 18) significantly increased the release of NPY in BALF (Fig. 5C) and serum (Fig. 5D).Of note, NPY levels were normal upon allergen sensitization (day 1 and 7), and while immune cells were still elevated, the neuropeptide concentration returned to baseline during the resolution phase (day 21; Fig. 5C-D).A similar pattern was observed for BALF IL-13 level (Fig. 5C-D).Concomitantly, Npy1r overexpression in JNC nociceptors also transiently peaked upon allergen challenges (Fig. 5E).We then tested whether these changes were specific to OVA-induced airway inflammation model and found a similar rise in BALF NPY in the house dust mite (HDM) model of asthma (Fig. 5F).
Using lung cryosections and immunostainings, we observed a strong and specific expression of NPY in nerve fibers often located around the bronchi (Fig. 6A).Staining of lung innervating peripheral neuron ganglia showed that 40% of sympathetic neurons in the stellate ganglia (SG) express NPY (Fig. 6B-D) which is in sharp contrast with virtually no expression in the JNC neurons (Fig. 6B-D).Using triple-labeling, we found that NPY-expressing neurons around the bronchi colocalized with the sympathetic neuron marker tyrosine hydroxylase (TH; Fig. 6E-F).While in proximity, the NPY-expressing neuron fibers were mostly distinct from NaV1.8 positive nociceptor nerve endings (Fig. 6A, E, F).We confirmed that NPY positive nerve fibers in the lung were of sympathetic and not sensory origin by performing a chemical sympathectomy with 6-OHDA (Fig. 6G-H).6-OHDA treatment completely abolished NPY and TH expression in the lung but did not affect the presence of CGRP + nociceptor neurons (Fig. 6G-H).
While it might seem logical to conclude that sensory and sympathetic fibers interact via NPY/NPY1R due to their close proximity, 6-OHDA-induced sympathectomy surprisingly did not affect NPY release in the airways of asthmatic mice (Fig. 6I).However, in the AAI model, NOD scid gamma (NSG) immunodeficient mice exhibited no increase in BALF NPY levels in the airways (Fig. 6J-K), suggesting that immune cells might also be a source of NPY.Moreover, both OVA sensitization and challenge were necessary to trigger NPY release in the airways (Fig. 6K), confirming that the full adaptive immune response to the allergen is essential.Supporting this hypothesis, we observed an increase in Npy1r RNA expression in the entire lung from mice with AAI (Fig. 6L), indicating NPY production by non-neuronal cells.Through our single-cell RNA-seq of AAI lung immune cells (Fig. 2B), we identified Npy expression exclusively in a subtype of antigen-presenting cells (APC) (Fig. 7M-N), which bear M2 macrophages markers such as Fcgr1, Ccr5, Cd63, Mrc1, Msr1, Ccl24 (SF.6A).This finding aligns with a prior study that reported NPY expression in phagocytes in a mouse influenza model 69 .We conclude that, although sympathetic neurons might interact with sensory neurons due to their proximity, the primary drivers of this interaction during AAI are likely NPY-producing M2 macrophages.

NPY1R blunts vagal nociceptor excitability.
NPY1R is a Gi-coupled receptor 70 and the action of NPY on DRG nociceptor neurons promotes either analgesia or noxious hypersensitivity [60][61][62][63][64][65][66][67][68] .Thus, A significant body of literature shows that NPY-NPY1R signaling contributes to the pathology of bronchial asthma and airway hyperresponsiveness in mouse models of allergic airway inflammation 71,72 .Additionally, NPY polymorphisms have been linked to an increased risk of asthma in overweight individuals 73 .Given that both Npy1r and NPY levels were elevated during allergic airway inflammation, we set out to address how NPY-NPY1R modulates cultured JNC nociceptor neuron sensitivity.To do so, we used whole cell patch clamp recording of NPY1R + JNC nociceptor neurons (NPY1R cre ::tdTomato fl/wt ).We measured electrical changes in response to the NPY1R-specific agonist Leu 31 Pro 34 NPY.We observed a significant reduction in the nociceptor neurons' excitability (Fig. 7A-B, SF. 7A-C), as measured by a reduced number of action potentials in response to current injection.We also measured a reduced level of intracellular cAMP in cultured JNC neurons exposed to Leu 31 Pro 34 NPY (Fig. 7C).Both effects indicate a direct engagement of NPY1R on nociceptor neurons and a strong reduction in the electrical activity and response capacity of these neurons.
To comprehensively understand the impact of NPY1R expression on sensory neurons on the asthma phenotype, we generated mice with a conditional knockout of this receptor in nociceptors (NaV1.8cre ::Npy1r fl/fl ).First, we used flow cytometry to immunophenotype the BALF of OVA-exposed NaV1.8 cre ::Npy1r fl/fl and its littermate control, and found similar immune cell numbers across both groups (SF.8A-B).Next, we used single-cell RNA sequencing (scRNA-seq) along with a pseudo-bulk and DESeq2 analysis approach to map the entire airway immune cell landscape in mice that underwent the asthma induction protocol.In the absence of NPY1R, we observed changes in T-cell populations, including 17 differentially expressed genes (DEGs) and the modulation of 160 Gene Ontology (GO) gene sets (Fig. 7D-F, ST. 7-9).Looking at T-cell subpopulations in detail, changes were primarily observed in the Rag1+ immature T-cell 1 (iT-cell-1) cluster (Fig. 7G-I, SF. 8C-G, ST. 10).We noted an increased proportion of these cells in the conditional knockout (NaV1.8cre ::Npy1r fl/fl ) mice compared to their littermate control animals.Most notably, the chemical ablation (RTX-exposed mice) of nociceptor neurons had the opposite effect, completely eliminating the emergence of this T-cell cluster in the BALF of asthmatic mice (Fig 7J-K).
DISCUSSION.Nociceptor neurons shape host defense at mucosal barriers.They do so by detecting environmental danger, triggering an avoidance response, and by tuning immune responses.In the context of allergy, nociceptor neurons were found to amplify dermatitis 14,19,20,74 , conjunctivitis 28 and airway inflammation 18,30,47,48,75,76 .In the lungs, these responses range from coughing and bronchoconstriction to mucus secretion and depending on the context, amplifying, or taming immunity.Such broad responses are made possible by the highly heterogenous nature of nociceptor neurons.
Airway nociceptors.The exact neuronal subset involved in these responses remained to be defined.Zhao and colleagues 7 posit that the variety of organs innervated by the vagus nerve explains in part the need for JNC sensory neuron heterogeneity.As such, airway sensory neurons 8,9,11,46 were classified as i) low-threshold stretch-sensitive neurons (essential to the respiratory cycle); ii) mechanoreceptors (sensitive to punctate mechanical stimuli); and iii) high threshold thermosensitive and chemosensitive nociceptors (recruited in response to tissue injury, inflammation, noxious chemicals or temperatures).
Using a combination of lineage reporters, retrograde tracing and transcriptomic analysis, we revealed that JNC airway nociceptor neurons have a unique gene signature segregated from that of other visceral nociceptors.We identified a new class of Kcng1-expressing inflammatory nociceptors (Trpa1 + , Trpv1 + , Il6 + , Npy1r + , Il13ra1 + ) that exclusively innervate the airways (NN8).This confirms the assumption of Kupari and colleagues 2 that this neuron subtype (which they labeled as NG14) consists of pulmonary afferent unmyelinated neurons.Additionally, we found that the airways are preferentially innervated by a neuronal subset (Kcnv1 + , Piezo1 + , Piezo2 + ) reminiscent of cough mechanoreceptors (NN7, NG3 in Kupari et al.'s study) as well as by a subset of polymodal nociceptors (NN2), and an NaV1.8 low population possibly belonging to the low-threshold stretchsensitive neurons 10,11 .These neuronal subtypes and their markers are listed in supplementary table 1. AAI reprogramming.We identified a drastic reprogramming of airway nociceptor neurons in response to allergic airway inflammation.Interestingly, the gene signature that we identified largely overlap with the one typically observed in injured nociceptor neurons.We can hypothesize that inflammation induces neuronal damage in the airways, which would explain such similarities.These changes are also reminiscent of those observed in LPS model of airway inflammation 77 .
While immune and glial cell activation or infiltration is sometimes reported in the spinal cord or DRG following peripheral inflammation or nerve injury 66,[78][79][80][81][82] , this was not the case in the JNC of AAI mice.This implies that the neuro-immune interactions occur at the peripheral nerve ending level rather than in the ganglia.Additionally, the nociceptor neurons innervating other organs were shielded from the transcriptional changes we observed in airway neurons, which also involves that those changes are triggered by signals originating from the nerve ending in the airways.
Cytokines reprograms nociceptor neurons.As we found that AAI reprogrammed airway nociceptor neurons' transcriptome, we tested how various cytokines impact JNC neuron expression profiles.Interestingly, nociceptor neurons showed different gene signatures when exposed to IL-13/IL-4, IL-1β, TNF-α, or BDNF.Thus, it appears that the combination of signaling pathways induced by nerve ending damage, inflammatory cytokines and neurotrophins add up in vivo and result in the nociceptors AAI signature.These findings are also indicative of nociceptor neurons' plasticity to various inflammatory conditions and subsequent contextdependent neuro-immune responses 34 .
IL-13 mimics some of the transcriptional changes observed during AAI, an effect that involves STAT6 phosphorylation and subsequent regulation of gene expression.In physiology, our findings suggest that IL-4/IL-13 released by airway TH2 and ILC2 cells are locally sensed by IL4RII-expressing vagal nerve endings.In turn, the intracellular signals are likely to be retrogradely transported to the soma to generate transcriptional changes.Such retrograde transport has been reported in nociceptors for STAT3 [83][84][85][86] and CREB 87,88 , and thus may also occur for STAT6 (SF.7).
IL-4 and IL-13 were previously found to induce calcium flux in dorsal root ganglia nociceptor neurons and to trigger itching 14,58 .While we did not observe direct calcium flux in JNC nociceptor neurons exposed to IL-13, our data converge regarding the functional expression of IL4RII and JAK1/2 activation in nociceptor neurons 58,89 .IL-13/IL-4/JAK/STAT6 is a key signaling pathway essential to type 2 inflammation and allergies 90,91 , and strategies to target it have proven effective to treat atopic dermatitis, asthma and to prevent itching 92 .We can reason that the sensory relief observed in these patients may, in part, be due to the silencing of this pathway in nociceptor neurons.
NPY, NPY1R, pain and allergy.Pain warns the organism of environmental dangers.Endogenous neuromodulatory mediators can either increase or decrease the organism's perception.In the context of pain, the impact of NPY and NPY1R remains controversial.Nevertheless, a consensus has emerged as for NPY1R expression and antinociceptive effect in the central nervous system [60][61][62][63]65 .
The effect on primary afferent nociceptor neurons is less established and NPY exerts a complex influence on pain sensitivity and neuropeptide release.This duality is likely explained by nociceptor neurons' co-expression of NPY1R and NPY2R, with the former dampening pain and inflammation while the latter exacerbates it 64,[66][67][68] .Building on these findings, we presented the first set of data suggesting the impact of NPY-NPY1R on vagal neuron sensitivity.We discovered that NPY-NPY1R decreased JNC nociceptor neuron activity by decreasing the levels of cAMP and reducing action potential firing in response to electrical stimulation.cAMP has long been recognized as an intracellular messenger promoting nociceptor sensitization [93][94][95][96][97][98] , an effect mediated in part by PKA-induced phosphorylation of NaV1.8 channels 99 .NPY also fulfills numerous physiological functions, ranging from the control of hunger/feeding to energy homeostasis 100,101 , vasoconstriction 102 and immunomodulation 103 .As we found to be the case in our OVA-and HDM-challenged mice, other studies showed elevated NPY levels in various rodent models of lung inflammation 71,72,[104][105][106] as well as in the plasma of elderly asthma patients 107 .Intriguingly, most of the NPY released appeared to originate from macrophages, rather than from sympathetic neurons.
Our single cell RNA sequencing data now reveal that NPY1R-expressing nociceptor amplify, the regulation and function of T cells, with implications for asthma, allergies, and broader inflammatory processes.Arpp21 and Stmn1 influence RNA splicing and microtubule dynamics, respectively, affecting T cell activation and migration crucial for allergic responses.Dntt and Rag1 are pivotal in the development of diverse T-cell receptor repertoires, directly impacting sensitivity to allergens and the immune response's adaptability.Jchain, though more closely associated with B cells, contributes to mucosal immunity relevant in asthma by mediating IgA and IgM transport.Histone genes like Hist1h1b, Hist1h1a, Hist1h2ap, and Hist1h2ae regulate gene expression through chromatin remodeling, influencing cytokine production in allergic inflammation.Cell cycle regulators such as Mki67, Cdca3, Ccna2, Ube2c, and Top2a affect T cell proliferation, a key event in the immune response to allergens.Marcks' role in cell signaling and migration is critical for T cell trafficking to inflammation sites, while Areg is involved in tissue repair and remodeling in asthma.Lastly, Syt13's potential impact on cytokine release through vesicular trafficking highlights the complex network of genes influencing T cell behavior and the pathophysiology of asthma, allergies, and inflammation, offering insights into potential therapeutic targets for these conditions.This supports the hypothesis that NPY1R reduces airway nociceptor activity, thereby diminishing their impact on the airway immune response in asthma.Knocking out NPY1R disrupts this regulatory pathway, leading to an outcome opposite to that of neuronal inactivation by resiniferatoxin.NPY was also reported to increase methacholine-induced bronchoconstriction 71,104 while NPY1R + vagal neurons were shown to trigger expiratory reflexes 1 .It would thus be of interest to assess whether its expression on vagal neurons is involved in cough and bronchoconstriction.Of note, NPY1R was shown to drive immune cell infiltration in the airways 71 .

Conclusion.
In summary, our data revealed a new class of vagal airway specific nociceptors that acquire an inflammatory gene signature during allergic inflammation or when stimulated with IL-13.Npy1r overexpression was induced by IL-13 in a JAK/STAT6-dependent manner.During allergic airway inflammation, sympathetic nerve fibers and M2 macrophages release NPY, which subsequently decrease nociceptor neurons' activity (SF.9) and impact the levels of allergic airway inflammation.Future work will reveal whether targeting vagal NPY1R constitutes a relevant therapeutic target to quell asthma-induced bronchoconstriction and cough.(A) Diagram depicting the retrotracing of the airway-innervating nociceptor neurons.To identify airwayinnervating nociceptor neurons, naive 8-week-old male and female nociceptor neurons reporter (NaV1.8cre ::tdTomato fl/wt ) mice were injected intranasally with the retrograde tracer DiD' (200 µM).Fourteen days later, the mice were euthanized and their JNC ganglia isolated and dissociated.Airway-innervating nociceptor neurons (NaV1.8+ DiD + ), visceral nociceptors (NaV1.8+ DiD -) and NaV1.8 -cells were purified by flow cytometry and their RNA sequenced (A).
(B) Heatmap displaying the z-score of the 50 top differentially expressed genes between visceral and airway nociceptors.
(E) 8-week-old male and female C57BL6 mice were injected intranasally with the retrograde tracer DiD' (200µM).Fourteen days later, the mice were euthanized and their JNC ganglia isolated and dissociated.JNC neurons were cultured (16h) and responsiveness to noxious stimuli was assessed.While responsiveness to the TRPV1 agonist capsaicin (300 nM) was stable between the two groups, the proportion of neurons responsive to the TRPA1 agonist JT010 (50 µM) was higher in airway-innervating neurons (DiD + ; D).
(F) Naive 8-week-old male and female NPY1R reporter (NPY1R cre ::tdTomato fl/wt ) mice were injected intranasally with the retrograde tracer DID' (200 µM).Fourteen days later, the mice were euthanized and their JNC ganglia isolated and dissociated.JNC neurons were cultured (16h), and neurons defined by KCl calcium responses.Using fluorescent imaging, we found that NPY1R-expressing neurons (tdTomato + ) are more frequent in the airway-innervating population (DiD + ; E).
(G) UMAP of Slc17a6 + (VGLUT2) JNC neurons from single-cell RNA sequencing revealed heterogeneous neuronal subsets.Gene set enrichment analysis was performed to address which neuron subtype preferentially innervated the airways, with normalized enrichment score (NES) indicated in blue.The neuronal cluster NN8 expresses several neuro-inflammatory markers (Il6, Kcng1, Npy1r, Trpa1, Trpv1) and exclusively innervates the airways (NES=2.3).Experimental details were defined in Prescott et al. and the bioinformatic analysis is described in the method section (F).(A-B) 8-week-old male and female nociceptor neuron reporter (Nav1.8cre ::tdTomato fl/wt ) mice underwent the ovalbumin mouse models of asthma.Allergic inflammation was induced in mice by an initial sensitization to ovalbumin (OVA) (i.p. day 0 and 7) followed by inhaled OVA challenges (days 14-17).Analysis of Bronchoalveolar lavage fluid showed significant airway inflammation characterized by a significant eosinophilic (CD45 + CD11C low SiglecF Hi ) infiltration (A).In other groups of mice, lungs were collected, and CD45 + cells were purified using FACS and subsequently analyzed through single-cell RNA sequencing.UMAPs demonstrated a change in the number and polarization of virtually all types of immune cells in the BALF of ovalbumin-exposed mice, including B cells, T cells, Granulocytes, NK cells, Basophils, and Antigen Presenting Cells (APC, which include Macrophages and Dendritic cells).(B).
(J) GSEA analyses were performed to compare gene signatures (i.e., overexpressed DEGs) induced by AAI, nerve injury 53 , BDNF, IL-1β, IL-13 and TNF-α to their respective whole datasets.The gene signatures induced by nerve injury, IL-13, IL-1β and BDNF were enriched in JNC neurons from mice with AAI (J).
(K) The common DEG induced by cytokines and AAI are depicted in a Venn diagram.IL-1β and BDNF signatures had similarities, while IL-13 induced two specific AAI genes Npy1r and Serpina3i (K).

Data are shown as mean ± S.E.M (A-D, L), as a volcano plot displaying DESeq2 normalized count fold change and nominal p-values for each gene (E-H) as number of DEGs (I), as heatmap displaying normalized enrichment score (NES) (J) or as a Venn diagram (K). N are as follows: A-D: n=3-4 cultures from different mice per group, E-H: n=3-4 cultures from different mice per group, L: n=6 culture dishes per group. P-values were determined by one-way ANOVA with post hoc Dunnett's (A-D, L) or DESeq2 analysis (E-H), or GSEA analysis (J). P-values
(F) 8-week-old female nociceptor neuron reporter (NaV1.8cre ::tdTomato fl/wt ) mice underwent the ovalbumin mouse models of asthma.Allergic airway inflammation was induced in mice by an initial sensitization to ovalbumin (OVA) (i.p. day 0 and 7) followed by intranasal instillation with OVA mixed with STAT6 inhibitor AS1517499 (150ug/50ul) or vehicle (days 14-17).One day after the last allergen challenge, mice were sacrificed, and JNC nociceptor neurons (tdTomato + ) purified by flow cytometry for RNA extraction and RT-qPCR.OVA-exposed mice nociceptors showed increased expression of Npy1r and Il6.Npy1r overexpression was prevented by AS1517499 (F).(A) Schematic of the AAI protocol.8-week-old female C57BL6 mice underwent the ovalbumin mouse models of asthma.Allergic airway inflammation was induced in mice by an initial sensitization to ovalbumin (OVA) (i.p. day 0 and 7) followed by inhaled OVA challenges (i.n.days 15-17).

Data are shown as a heatmap displaying the z-score of DESeq2 normalized counts (A), as Western blots (B-C), as mean ± S.E.M (D, F), or as box (25th-75th percentile) and whisker (min-to-max) plot (E). N are as follows:
(B-D) 8-week-old female C57BL6 mice underwent the ovalbumin mouse models of asthma.BALF, serum and lung were harvested at different time points (days 0, 14, 18, and 21) and the levels of inflammatory mediators were analyzed by ELISAs and qPCR.OVA-exposed mice showed significant airway inflammation characterized by leukocytes (CD45 + ) and eosinophil (CD45 + CD11C low SiglecF Hi ) infiltration on day 18 (F).Along with this rise in airway inflammation, we found an increase in BALF (G) and serum (H) NPY, while lung Npy expression was also increased (I).
(E) JNC ganglia were harvested from OVA-exposed nociceptor reporter mice (TRPV1 cre ::tdTomato fl/wt ) at different time points, TRPV1 + neurons were purified (tdTomato + ) by flow cytometry, and changes in transcript expression were measured by qPCR.At the peak of inflammation (day 18), we found a transient increase in Npy1r expression (J).
(F) To induce another model of allergic airway inflammation, 8-weeks-old female C57BL6 mice were challenged (day 1-5 and 8-10) with house dust mite (HDM; 20μg/50μL, i.n.).The mice were sacrificed on day 11, their BALF harvested, and cell free supernatant analyzed by ELISA.HDM-exposed mice showed a significant increase in BALF NPY level (F).(A-F) Lung (A), jugular nodose complex ganglia (JNC; B) and stellate ganglia (SG; C) were harvested from naïve 8-weeks-old male and female nociceptor neuron reporter (NaV1.8cre ::tdTomato fl/wt ) mice.The tissues were cryosectioned, and the source of NPY assessed by immunofluorescence.In the lung, NPY (green) and NaV1.8 (tdTomato, red) were expressed in nerve fibers around the bronchi (A).NPY was not expressed in the JNC (B, D).In the stellate ganglia (SG), NPY was strongly expressed in TH + sympathetic neurons (C, D).PGP9.5 (white) was used to define JNC and SG neurons (B, C, D).While sympathetic and sensory fibers were often found in proximity, NPY (green) mostly colocalized with the sympathetic neuron marker tyrosine hydroxylase (TH; blue) rather than with NaV1.8 (tdTomato; red) nociceptor fibers (E-F).NPY and TH were absent in lung cryosections from mice with chemical sympathectomy induced by OHDA (G-H), while sensory nerve fibers (CGRP+) were not affected (G-H).

Data are shown as experiment schematics (A), or as mean ± S.E.M (B-F). N are as follows
(I) NPY is still elevated in bronchoalveolar lavage fluid of AAI mice with chemical sympathectomy (I).
(J-K) Immunodeficient mice (NSG) showed absence of immune cell infiltration in the airways following OVA protocol (J).While NPY was released in the BALF of wild type mice with AAI, it was absent in the airways of immunodeficient mice (NSG) (K).Additionally, OVA challenge without previous sensitization neither induced immune response (J) nor NPY release in the airways (K).
(L-N) NPY was elevated in the RNA of mice at the peak of inflammation of the AAI protocol, (day18), and persisted 3 days later (day 21) (L).Single-cell RNA sequencing of FACS-purified CD45 + cells revealed selective expression of Npy in a subset of antigen-presenting cells (APC) and macrophages within the bronchoalveolar lavage fluid (BALF) of ovalbumin-exposed mice (M-N).(A-C) 8-week-old male and female NPY1R reporter (NPY1R cre ::tdTomato fl/wt ) mice were sacrificed and their JNC neurons harvested and cultured (16 hours).Whole cell patch clamp electrophysiology was performed on the NPY1R + nociceptor neurons.A current clamp was applied while the neurons' membrane potential was recorded before and after exposing (10 min) the cell to Leu 31 Pro 34 NPY (250 nM) or its vehicle.The number of action potentials was counted for each current stimulation, and the areas under the curve was calculated and plotted (A).A representative trace of neuronal response to 315pA stimulation before and after Leu 31 Pro 34 NPY is displayed in (B).While the vehicle had little to no effect on neuronal excitability, Leu 31 Pro 34 NPY reduced the number of action potentials in response to current stimulation in NPY1R + (tdTomato) neurons (A, B).

Data are shown as immunostained tissue, scale bar 50 μm (A-C, E, G), or as mean ± S.E.M (D, F, H-L), or Seurat normalized gene expression (M, N). N are as follows: D: n=4-5 mice per group, F: n=12 field of views from 4 different mice, D: n=4-5 mice per group, I: n=4-10 mice per group, J-K: n=4-10 mice per group, L: n=6-7 mice per group. P-values were determined by a two-sided unpaired Student's t-test (D, F, H) or by oneway ANOVA with post hoc Dunnett's (I, J, K). P-values are shown in the
(C) 8-week-old C57BL6 male and female mice were sacrificed and their JNC neurons harvested and cultured (16 hours).The neurons were exposed to Leu 31 Pro 34 NPY (250 nM) or vehicle for 30 minutes in presence of phosphodiesterase inhibitors.The cells were then lysed, and the cAMP concentration assessed by enzymatic assay.Leu 31 Pro 34 NPY significantly reduced cAMP concentration (C).
(D-F) Eight-week-old female mice with a nociceptor neuron-specific NPY1R conditional knockout (NaV1.8cre ::NPY1R fl/fl ) and their littermate controls (NaV1.8wt ::NPY1Rf l/fl ) were subjected to the ovalbumininduced asthma model.On day 18, lungs were harvested, and CD45 + cells were purified by flow cytometry before undergoing analysis by single-cell RNA sequencing.Gene expression was averaged for each cell type in each replicate to conduct pseudobulk analysis using DESeq2 (D) and gene set enrichment analysis (E).NaV1.8 cre ::NPY1R fl/fl mice exhibited exacerbated allergic airway inflammation, with T-cells displaying the most significant differences in terms of the number of differentially expressed genes (D; False Discovery Rate < 0.2) and enriched Gene Ontology gene sets (E; False Discovery Rate < 0. 8 and 9.

2). The average expression of all T-cell DEGs is displayed in a heatmap (F). Complete DESeq2 analysis (D) and GO gene sets (E) are detailed in Supplementary Tables
(G-K) The T cells isolated from the single-cell RNA-seq dataset of AAI-affected lungs were subclustered (G).In the UMAP projections, the cluster named immature T cells 1 (iTcells 1) appears distinct between the NPY1R conditional knockout mice (NaV1.8cre ::NPY1R fl/fl ) and the littermate control mice (NaV1.8wt ::NPY1R fl/fl ) (G).Quantification reveals a significantly higher number of these cells in the NaV1.8 cre ::NPY1R fl/fl mice (H).In contrast, iTcells 1 were less abundant in nociceptor neurons ablated (RTX-treated) mice compared to vehicle control (I-J).

Data are shown as traces of membrane potential for individual neurons (B), mean ± S.E.M (A, C), or UMAP (D). N are as follows: A: n=10 vehicle treated neurons and 9 Leu 31 Pro 34 NPY treated neurons, C: n=19 culture wells, D-I: n=3 pools of 2 mice per group, J-K: n=2 pools of 2 mice per group. P-values were determined by a two-sided unpaired Student's t-test (A, C, I, K), Deseq2 analysis (D), GSEA analysis (E). P-values are shown in the figure.
Supplementary Figure 1.Airway vagal nociceptor neurons have a unique transcriptome.
(C) Naive 8-week-old male and female NaV1.8 cre ::tdTomato fl/wt mice were injected intranasally with the retrograde tracer DiD' (200 µM).Fourteen days later, the mice were euthanized and JNC, thoracic DRG, and TG ganglia isolated, dissociated, and imaged with a fluorescence microscope.DiD' retrotracer was detected in JNC nociceptor neurons but virtually absent in other ganglia (C).
(D-E) Naive 8-week-old male and female C57BL6 mice were injected intranasally with the retrograde tracer DiD' (200 µM).Fourteen days later, the mice were euthanized and their JNC ganglia isolated and dissociated.JNC neurons were cultured (16 h) and calcium responsiveness to noxious stimuli was assessed.While the average neuronal responsiveness to the TRPV1 agonist capsaicin (300 nM) was stable between the two groups (D), the calcium flux induced by the TRPA1 agonist JT010 (50 µM) was higher in airway-innervating nociceptor neurons (E).
For voltage clamp experiments, the neurons were clamped at -60mV.Neuronal currents were recorded before and after drug exposure from a series of depolarization steps ranging from -120 to +70mV in 10 mV intervals.For the current clamp AP protocol, neurons were injected with a series of 35 pA current steps (1 s duration) from 0 to 630 pA.These voltage and current clamp experiments were performed before and 10 mins postaddition with a pipette of external solution (vehicle) or Leu 31 Pro 34 NPY (250 nM; Tocris, #1176).Current amplitudes and APs were quantified using Clampfit 10.7 (Molecular Devices).Since baseline activity varied among the JNC neurons, the activity of the neurons after application of vehicle or Leu 31 Pro 34 NPY was normalized to baseline activity of the same cell.
Whole-lung RNA extraction: Whole lungs were minced with a razor blade, and about a quarter of the minced lung was mixed with 500uL Trizol (Thermofisher #15596018) and stored at -80° for subsequent RNA extraction.RNA was separated from protein and DNA by mixing 500 μL of sample in Trizol with 100 μL chloroform before ultracentrifugation (15 min, 16000g, 4°).The upper phase was mixed with a half volume of 100% isopropanol, transferred to a purification column, and the RNA was then purified using the kit E.Z.N.A.Total RNA Kit I (VWR, #CA101319) following manufacturer's instructions.
JNC Neurons RNA extraction: RNA was separated from protein and DNA by mixing 500 μL sample in Trizol with 100 chloroform before ultracentrifugation (15 min, 16000g, 4°).The upper phase was mixed with a half-volume of 100° isopropanol, transferred to a purification column, and the RNA purified using the kit PureLink RNA Micro Scale (Thermofisher, #12183016) following manufacturer's instructions.
Immunofluorescence analysis.Using ImageJ or Nikon Elements, circular ROI were manually defined around each neuron based on PGP9.5 fluorescence for JNC and SG slices.For confocal pictures of lung slices, an orthogonal projection of the z-stack was made, then a threshold-based method was used to define the area of nerve fibers and immune cells for each marker as well as their overlapping areas.Cultured cells were analyzed with a threshold-based method to define neurons based on tdTomato or PGP9.5 expression.Average fluorescence in ROI was exported and all further analysis were performed in Microsoft Excel.When spillover between colors occurred, a compensation was applied using single-stain controls.
RNA sequencing.RNA libraries preparation and sequencing were carried out at the genomic platform of the Institut de Recherche en Cancérologie et en Immunologie (IRIC).Briefly, RNA quality was assessed using a Bioanalyzer (Agilent), and all preparations had an RIN>7.5.Libraries were prepared using the KAPA mRNA HyperPrep Kit (KapaBiosystems #KR1352).All barcoded samples were then sequenced with a Nextseq500 (Illumina) with 75-cycle single-end read.
Genome alignment and differential expression analysis were carried out (IRIC genomic platform).Sequences were trimmed for sequencing adapters and low quality 3' bases using Trimmomatic version 0.35 and aligned to the reference mouse genome version GRCm38 (gene annotation from Gencode version M25, based on Ensembl 100) using STAR version 2.7.1.Gene expressions were obtained from STAR as readcounts and computed using RSEM to obtain normalized gene and transcript level expression in FPKM.Differential expression analysis for the various comparisons of interest were made using DESeq2 182 and STAR readcounts.Further analysis and plots were made using RStudio or Microsoft Excel.Genes were considered as differentially expressed if their adjusted p-value (FDR) was less than 0.2.
In-silico analysis of JNC neuron single-cell transcriptome: Prescott et al. 1 generated single-cell sequencing data for jugular-nodose ganglia cells from 40 mice using the 10X Genomics platform.The data was downloaded from the NCBI Gene Expression Omnibus (GSE145216) and analyzed using Seurat.Neuronal cells were selected based on Slc17a6 (VGLUT2) expression (raw count ≥ 2).A standard workflow was used for quality control, preprocessing, normalization, and clustering (resolution = 0.2, PCs = 1:30).Phox2b and Prdm12 were used to identify nodose and jugular groups, while nociceptor neurons and lowthreshold sensory neurons were defined based on their expression of Scn10a and Scn1a 2 .Xhao et al. 7 generated single-cell projection sequencing data for jugular-nodose ganglia cells from mice in which barcode expressing viruses were used to retrogradely label neurons innervating various internal organs.The data was downloaded from the NCBI Gene Expression Omnibus (GSE192987) and analyzed using Seurat.Neurons expressing at least one barcode (raw count ≥ 2) were selected.A standard workflow was used for quality control, preprocessing, normalization, and clustering (resolution = 0.75, PCs = 1:20).GSEA analysis: Gene set enrichment analysis 183 was performed to compare similarities between different sequencing results together using the GSEA software.To assess the preferential innervation of airways by scRNAseq neuron clusters, GSEA analysis was performed using the cluster markers identified in Prescott et al.'s dataset as genesets, and the DESeq2 counts of airway and visceral neurons as gene expression dataset.To compare AAI, BDNF, IL-1b, IL-13, TNF-a, and nerve injury sequencing data, gene signatures were constituted by selecting overexpressed genes (FDR<0.2) induced in each condition.For the nerve injury signature, RNAseq data from Cobos et al 53 was downloaded from NCBI Gene Expression Omnibus (GSE102937), reanalyzed by DESeq2 (supplementary table 3), and overexpressed genes (FDR<0.2) we selected.The enrichment of each gene signatures in each DESeq2 count datasets were measured by GSEA.
Lung single-cell RNA-sequencing analysis.Reads were mapped to the mouse reference genome using Cell Ranger (Chromium) and analyzed using Seurat.Low RNA and dead cells were excluded (nFeature_RNA > 100 & percent.mt< 25).Total cells were clustered (dims = 1:20 and resolution = 0.5).21 clusters were identified, and were assigned using common markers and the Tabula Muris reference.Markers used included: Cd19, S100a8, Cd3e, Klrk1, Cd200r3, Cd68, Itgax, Lgals3, H2-Ab1.Clusters were grouped as large cell types (B cells, T cells, NK cells, antigen presenting cells (APC), Granulocytes, Basophils, Alveolar macrophages, and endothelial cells) to compared expression between different conditions.To compare naïve and AAI conditions, a single library was prepared and the inflammation markers were defined using Seurat function FindMarkers.To compare gene expression between NPY1R conditional Knock-Out and littermate controls, 3 biological replicates per group were prepared.To maximise statistical robustness, a pseudobulk analysis was then used to compare groups.Gene expression was averaged for each cell type and for each sample, and Deseq2 was used to identify differentially expressed genes, while GSEA was used to define gene ontology gene sets.Since a very low number of Basophils, Alveolar macrophages, and endothelial cells were sequenced, those were not used for the pseudobulk analysis.To focus the analysis on T cells specifically, T cells were subsetted, and Seurat functions were used to define the different groups of t cells (dims =1:40, resolution =, 1).Since V(D)J genes were driving excessive clustering of the T cells, those were not taken into account for the clustering.T cell clusters were defined based on various common markers including: Cd3e, Cd4, Cd8, Cd8b1, Il1rl1, Il13, Il17a, Trdc, Rag1, and Klrc1 (SF.6C).Data availability.Bulk and single-cell RNA-sequencing raw and processed data have been deposited in the NCBI's gene expression omnibus (GSE223355; single cell number pending).Processed data can also be accessed in the supplementary tables 1-10.Additional information and raw data are available from the lead contact upon reasonable request.

Statistics.
No data were excluded.P values ≤ 0.05 were considered statistically significant.One-way ANOVA, two-way ANOVA, and Student t-tests were performed using GraphPad Prism.DESeq2 and Seurat analysis and statistics were performed using RStudio.
Replicates.Replicates () are described in the figure legends and represent the number of animals for in vivo data.For in vitro data, replicates can either be culture wells or dishes, animals, fields-of-view (microscopy), or neurons (patch-clamp), but always include different preparations from different animals to ensure biological reproducibility.

Figure 1 .
Figure 1.Airway vagal nociceptor neurons have a unique molecular profile.
are shown in the figure or indicated by * for p ≤ 0.05; ** for p ≤ 0.01; *** for p ≤ 0.001.
A: n=3 cultures from different mice per group, B-D: n=1 JNC culture and n=3 DRG cultures from different mice, E: n=5 cultures from different mice per group, F: n=7-8 mice per group.P-values were determined by one-way ANOVA with post hoc Dunnett's (D-F).P-values are shown in the figure or indicated by * for p ≤ 0.05; ** for p ≤ 0.01; *** for p ≤ 0.001.

Figure 5 .
Figure 5. NPY is released in airways during allergic airway inflammation.

Figure 6 .
Figure 6.NPY is expressed by sympathetic neurons and inflammatory macrophages in the lung.