Recurrence of acute allergic asthma depends on the role of ILC2 driven by Il1rl1 signaling

Abstract

Background

Asthma is a chronic inflammatory airway disease characterized by recurrent episodes that significantly impair disease control and reduce patients’ quality of life. Despite its clinical importance, the mechanisms underlying asthma relapse remain poorly understood, and effective strategies to prevent exacerbations are still lacking.

Methods

An acute allergic asthma relapse mouse model was established using ovalbumin sensitization and challenge. Single-cell transcriptomics was employed to investigate the cellular and molecular mechanisms driving asthma relapse. Flow cytometry and gene knockout experiments were conducted to validate the findings.

Results

We successfully established an acute allergic asthma relapse mouse model. Single-cell transcriptomic analysis revealed that T cells and type 2 innate lymphoid cells (ILC2s) are pivotal during asthma relapse, serving as the primary cellular sources of type 2 inflammatory cytokines. Further subcluster analysis identified T-cell subcluster 4 and ILC2 subcluster 0 as the predominant contributors to type 2 cytokine production. Complex intercellular communication networks were observed, with macrophages, natural killer (NK) cells, and dendritic cells functioning as central signaling hubs. Pseudo-time trajectory analysis highlighted the critical role of ILC2s and the Il1rl1 signaling pathway in asthma relapse. These findings were corroborated by flow cytometry. Il1rl1-deficient mice displayed similar pulmonary inflammation to wild-type mice during the initial asthma episode; however, asthma relapse was significantly attenuated. Mechanistically, Il1rl1 deficiency resulted in a substantial reduction in both the number and functional capacity of ILC2s.

Conclusion

The recurrence of acute allergic asthma is driven, at least in part, by ILC2s through Il1rl1 signaling. Genetic ablation of Il1rl1 significantly suppresses asthma relapse, suggesting that targeting Il1rl1 may represent a novel therapeutic strategy for preventing asthma exacerbations.

Background

Allergic asthma is a chronic inflammatory disease of the airways that affected an estimated 262.4 million people globally in 2019 [1]. Despite significant declines in asthma-related mortality over the past few decades, largely due to standardized management strategies and the development of novel therapeutic drugs, the global burden of asthma remains substantial. This is particularly evident in regions with lower socioeconomic indices, where access to consistent treatment may be limited [1]. Consequently, asthma prevention and control continue to represent a critical global public health challenge.

A major obstacle in the management of asthma is its recurrent nature. Patients diagnosed with asthma typically require long-term maintenance therapy, predominantly with inhaled corticosteroids, to prevent exacerbations and achieve symptom control [2]. In theory, standardized treatments could enable up to 80% of asthma patients to maintain good disease control. However, real-world data reveal that fewer than 60% of patients achieve this goal. This discrepancy underscores the limitation of current therapies, which focus on symptom management rather than addressing the underlying mechanisms of recurrence. Hence, understanding the mechanisms driving asthma relapse and identifying strategies to prevent it are critical research priorities.

In recent years, asthma mouse models have been used to investigate the issue of asthma recurrence [3,4,5]. Recent investigations suggest that lung-resident memory T cells are central contributors to the recurrence of asthma. These cells can mount rapid immune responses upon re-exposure to allergens, driving the inflammatory processes that characterize asthma relapse. Importantly, experimental evidence indicates that blocking the activity of resident memory T cells significantly mitigates asthma recurrence [5]. However, it remains unclear whether other immune cell types are also involved in the pathological mechanisms underlying asthma memory.

Among the immune cells implicated in asthma pathogenesis, group 2 innate lymphoid cells (ILC2s) have garnered considerable attention for their role in type 2 inflammation [6, 7]. ILC2s are present in the lungs and can also be recruited from peripheral tissues during inflammation [8, 9]. These cells are activated by epithelial cell-derived alarmins, including IL-33, thymic stromal lymphopoietin (TSLP), and IL-25, leading to the production of large amounts of type 2 cytokines such as IL-4, IL-5, and IL-13 [10,11,12,13]. This cytokine cascade promotes airway inflammation and hyperresponsiveness, hallmarks of asthma. However, whether ILC2s contribute to inflammatory memory and asthma recurrence remains uncertain. Emerging studies suggest that ILC2s may exhibit memory-like properties, akin to those of adaptive immune cells like T cells [14,15,16]. Based on this, we hypothesize that ILC2s could play a pivotal role in the recurrence of allergic asthma.

To test this hypothesis, we developed a mouse model of acute allergic asthma recurrence, which recapitulated the key features of asthma relapse. Using single-cell transcriptomics, we analyzed the dynamic changes in immune cell populations within the lungs during the recurrence process. Our results highlight the significant role of ILC2s in asthma relapse and identify the critical involvement of Il1rl1 signaling in this process. These findings were further validated through flow cytometry and experiments using gene knockout mice. Together, our study provides novel insights into the immunological mechanisms of asthma recurrence and suggests potential targets for improved therapeutic strategies.

Materials and methods

Animals and reagents

Female wild-type C57BL/6 mice were purchased from Sibeifu Biotech Inc. (Beijing, China). Il1rl1 knockout (C57BL/6 background) mice were obtained from Shanghai Model Organisms (Shanghai, China). All mice were housed under specific pathogen-free (SPF) conditions in the Department of Laboratory Animal Sciences at Guangzhou Medical University (Guangzhou, China). The animals were maintained on a 12-hour light/dark cycle with unrestricted access to food and water. The number of mice in each group was determined to be 4–6 according to different experiments. All experimental procedures were approved by the Institutional Animal Care and Use Committee of Guangzhou Medical University (Approval No. 2022327).

Ovalbumin (OVA, grade V; Solarbio, China) and methacholine were purchased from Sigma-Aldrich (United States). Aluminum hydroxide adjuvant was sourced from Bioss Inc. (China). Collagenase D and DNase I were obtained from Roche Diagnostics (United States).

OVA sensitization, challenge, rest, and Re-challenge protocol

A murine model of asthma was established in both wild-type and Il1rl1 knockout mice using a protocol based on ovalbumin (OVA) sensitization and challenge, as previously described [17]. For sensitization, each mouse received an intraperitoneal injection of 50 µg ovalbumin (OVA) dissolved in 200 µL phosphate-buffered saline (PBS) mixed with 50 µL aluminum hydroxide adjuvant. This was followed by intranasal administration of 100 µg OVA (in 20 µL PBS) under isoflurane anesthesia to induce the challenge phase. After successfully establishing an acute asthma model, the mice were left undisturbed for a rest phase of 6 to 8 weeks, as described in prior studies [4]. During this period, the mice were maintained under routine housing conditions with standard feeding and watering, without additional interventions. Following the rest phase, the mice entered the re-challenge phase. During this phase, they were administered 100 µg OVA (in 20 µL PBS) intranasally once daily for three consecutive days. In addition, in order to verify that the recurrence of asthma in mice is allergen-specific, we used HDM (containing 5ug HDM, dissolved in 20uL PBS) to perform nasal provocation on another group of memory mice. After the re-challenge, various assessments were performed to evaluate key asthma-related parameters, including lung inflammation, airway mucus production, and other markers of asthma pathology. We called the mice that developed asthma after OVA sensitization and challenge as the “Asthma” group, and the mice that were sensitized with OVA but nasally administered with PBS as the “Control” group. The asthmatic mice that were left alone for 6–8 weeks without any treatment became the “Memory” group, and the mice in the memory group that were challenged with OVA nasally again were called the “Re-challenge (ReCh)” group.

Measurement of airway hyperresponsiveness

Airway hyperresponsiveness (AHR) was assessed using a Buxco FinePointe modular system with an invasive setup. Pulmonary resistance (RL) values were measured as an indicator of airway resistance. Mice were anesthetized, and their tracheas were exposed via a transverse incision for intubation. The mice were then placed in a supine position within a body plethysmograph chamber connected to a ventilator. Airway responses to increasing doses of methacholine (0, 3.125, 6.25, 12.5, 25, and 50 mg/ml) were recorded.

Total and differential cell counts in Bronchoalveolar lavage fluid (BALF)

Bronchoalveolar lavage fluid (BALF) was collected by inserting a 24G venipuncture needle into the trachea. The lungs were perfused twice with 600 µl of phosphate-buffered saline (PBS). The collected BALF was centrifuged at 500 × g for 5 min at 4 °C to separate cells. The resulting cell pellet was resuspended in 500 µl of pre-cooled PBS for total cell counts.

For differential cell counts, cytospin slides were prepared using the Cytospin 4 centrifuge (Epredia, UK). Slides were stained with hematoxylin and eosin (H&E), and at least 300 cells per sample were counted under a microscope to determine the proportions of different cell types.

Tissue harvest, processing, histological analysis and flow cytometry

Lungs were perfused via the right ventricle with 5 mL of pre-cooled phosphate-buffered saline (PBS) to remove blood. The lungs were carefully excised, separating them from the heart, thymus, and bronchial lymph nodes. The right lung lobe was fixed in 4% formaldehyde, dehydrated, and embedded in paraffin. Section (5 μm thick) were prepared using a rotary microtome, mounted on glass slides, and stained with hematoxylin-eosin (H&E) and periodic acid Schiff (PAS). H&E-stained sections were analyzed to assess lung inflammation. A pathologist, blinded to the experimental groups, assigned an inflammation score based on peri-bronchial and perivascular inflammatory cell infiltration using the following criteria. 0 = no inflammatory cell infiltration; 1 = Few inflammatory cells; 2 = a ring of inflammatory cells one cell layer deep; 3 = a ring of inflammatory cells two to four cells deep; 4 = a ring of inflammatory cells of more than four cells deep. Goblet cell hyperplasia was assessed as the percentage of PAS-positive cells (goblet cells) in airway epithelial cells in PAS-stained Sect. 0: Goblet cells ≤ 5% goblet cells; 1 = 5–25%; 2 = 25–50%; 3 = 50–75%; 4: Goblet cells ≥ 75%. For each lung, the sum of central and peripheral airway scores was divided by the number of airways examined across 10–25 consecutive fields per slide. Representative regions of tissue were selected for imaging and documentation.

For flow cytometry analysis, the left lung lobe was minced into small pieces and digested in Hank’s Balanced Salt Solution (HBSS, Gibco) containing 2% fetal calf serum (FCS, DIB), 25 mM Hepes (Solarbio), 1 mg/mL collagenase D (Roche), and 0.1 mg/mL DNase I (Roche) at 37 °C for 45 min. The digested tissue was passed through a 70-µm cell strainer (Biologix) to obtain single-cell suspensions. Red blood cells were lysed with ammonium-chloride-potassium lysis buffer (Biosharp). For intracellular cytokine staining, cells were stimulated with 1 ml of RPMI 1640 medium (containing 10% FCS) containing Leukocyte Activation Cocktail with GolgiPlug (2 µL per 10⁶ cells/mL, BD Biosciences) for 4–6 hours in a 37 °C humidified CO₂ incubator. After staining surface antigens, cells were fixed and permeabilized using the BD Cytofix/Cytoperm Kit, followed by intracellular cytokine staining. For T cells panel, the following antibodies were utilized: BUV396-CD45 (BD), FITC anti-mouse CD3ε (BD), BV 510 anti-mouse CD4 (BD), BV 650 anti-mouse CD69 (BD), BV786 anti-mouse IFN-γ(BD), APC anti-mouse IL-5 (BD), PE anti-mouse IL-4 (Biolegend), BV 421 anti-mouse IL-9 (BD). For ILC2 panel, the following antibodies were utilized: BUV396 Rat anti-mouse CD45 (BD), FITC anti-mouse Lineage Cocktail (anti-mouse CD3ε, anti-mouse Ly-6G/Ly-6 C (BD), anti-mouse CD11b (BD), anti-mouse CD45R/B220 (BD), BV605 anti-mouse KLRG1 (BD), PE-Cy7 anti-mouse SCA-1 (BD), APC anti-mouse IL-5, (BD) PE anti-mouse IL-4 (Biolegend), BV 421 anti-mouse IL-9 (BD). ILC2 were defined as CD45⁺Lineage⁻Sca1⁺KLRG1⁺ cells [18]. Samples were acquired and analyzed using an LSRFortessa flow cytometer (BD Biosciences), and data were processed using FlowJo software (version 10). Cytokine-positive populations were gated based on unstimulated or naïve cell controls.

Lung immune cell purification

Lung immune cells were isolated for single-cell RNA sequencing (scRNA-seq) from three experimental groups of mice: control group of wild-type mice (n = 3), rest group (n = 4, harvested at the end of the rest phase), and re-challenge group (n = 4, harvested following OVA re-challenge). Mice were euthanized via intraperitoneal injection of an overdose of pentobarbital. Following euthanasia, thoracotomy was performed, and the pulmonary circulation was perfused with 37 °C PBS to remove intravascular cells. Lung tissues were digested to generate single-cell suspensions as described previously. Briefly, digested tissues were passed through a 70-µm cell strainer, and red blood cells were lysed using ammonium-chloride-potassium (ACK) lysis buffer. The resulting cells were washed with ice-cold 1% fetal calf serum (FCS) in PBS. To block non-specific Fc receptor binding, cells were incubated with anti-mouse CD16/32 antibody (1:100 dilution). Following this, cells were stained with anti-mouse CD45 antibody (1:100 dilution) to label immune cells. After staining, cells were washed twice with ice-cold 1% FCS/PBS and subsequently stained with Fixable Viability Dye 7-AAD (1:50 dilution) to exclude dead cells. Lung immune cells were sorted using a BD FACSAria III cell sorter. The sorting process yielded highly purified populations of CD45⁺ viable immune cells, with a purity exceeding 90%, which were used for downstream single-cell RNA sequencing.

Droplet-based single-cell RNA sequencing (scRNA-seq)

Immediately following cell sorting, CD45⁺7-AAD⁻ single cells were processed using the 10x Chromium system (10x Genomics). Library preparation was performed by Accuramed Technology (Shanghai) Limited following the Chromium Single-Cell 3’ Reagent Kits v3.1 protocol (10x Genomics). Sequencing was conducted on an Illumina NOVASeq 6000 platform. Data processing and quality control were performed using the Cell Ranger software package (version 7.1.0; 10x Genomics). Preliminary analysis using Cell Ranger revealed 15,407 cell barcodes with a median of 1715 genes per cell in the control group, 14,867 cell barcodes with a median of 1779 genes per cell in the rest group, and 15,124 cell barcodes with a median of 1698 genes per cell in the re-challenge group. These high-quality data served as the basis for downstream analyses of cellular transcriptional profiles and immune dynamics across experimental groups.

Bioinformatic analysis of scRNA-seq data

Gene expression matrices generated by the 10x Cell Ranger pipeline (aggregate option) were analyzed using the R package Seurat [19] (version 4.3.0) with default parameters. Filtered expression matrices for each sample were loaded into R as Seurat objects. Cells expressing fewer than 200 genes and genes detected in fewer than three cells were excluded. Additionally, cells with more than 20% of their unique molecular identifiers (UMIs) mapped to mitochondrial genes were removed to eliminate low-quality cells. Potential doublets were identified and filtered using the DoubletFinder R package for each dataset independently [20].

Normalization of gene expression data was performed by scaling each cell’s total expression (default scaling factor: 10,000) and applying a logarithmic transformation via Seurat’s integrated normalization function. Clustering analysis followed standard Seurat workflows, using a resolution of 1.2 to identify clusters. Clusters were visualized using Uniform Manifold Approximation and Projection (UMAP) based on principal components derived from principal component analysis (PCA). Average gene expression matrices for each cluster were computed, and differential expression analysis was performed with Seurat’s FindAllMarkers function (parameters: only.pos = FALSE, min.pct = 0.2, thresh.use = 0.2) to identify significant marker genes for each cluster.

Cell clusters were annotated using SingleR [21], an unbiased cell-type identification tool, cross-referenced with the reference dataset from the Immunological Genome Project (ImmGen). Spearman correlation analysis was used to compare variable genes in the dataset with those in the reference dataset. Correlation coefficients were aggregated for each cell type to derive a unified correlation value per cell type per single cell.

For visual representation, UMAP feature plots were generated for selected marker genes using Seurat’s FeaturePlot function. These visualizations highlighted the spatial expression patterns of key markers across clusters.

Cell-cell communication analysis

To investigate intercellular communication within lung tissue, we employed CellChat (version 1.5.1) [22], a computational framework leveraging a comprehensive database of ligand-receptor interactions. A CellChat object was generated from the Seurat object using the “CellChat” function. Over-expressed ligands, receptors, and interactions specific to different experimental groups were identified using the “Identify Over Expressed Genes” function. Communication probabilities and cell-cell communication networks were inferred using the “CommunProb” and “computeCommunProb Pathway” functions, which estimate interaction probabilities based on the average gene expression within each experimental group. Interactions with a p-value less than 0.05 were considered statistically significant. Cell-cell communication networks were visualized using various built-in visualization tools provided by the CellChat package, allowing us to map and interpret interaction pathways between distinct cell populations.

For further analysis, we used the R package “NicheNet” to predict the regulatory effects of ligands secreted by sender cells on downstream target genes in receiver cells [23]. NicheNet integrates gene expression data with knowledge of signaling pathways and gene regulatory networks to identify potential signaling interactions and their downstream impacts.

Trajectory inference

Trajectory analysis was performed using the R package Monocle (version 2.14.0) [24] to delineate the developmental trajectory of innate lymphoid cells type 2 (ILC2s). Highly variable genes (HVGs) were selected based on the intersection of differentially expressed genes identified from single-cell RNA sequencing (scRNA-seq) and bulk RNA-seq datasets. Statistical models for the data were constructed using the “estimateSizeFactors” and “estimateDispersions” functions. Dimensionality reduction was carried out with the “reduceDimension” function, after which cells were ordered along a pseudotime trajectory using the “orderCells” function. This approach enabled the temporal mapping of ILC2 development and differentiation states.

Statistical analysis

All data are presented as mean ± standard error of the mean (SEM). Statistical analyses were conducted using GraphPad Prism (version 9.4.1; GraphPad Software). One-way or two-way analysis of variance (ANOVA) was used to evaluate statistical significance, followed by Tukey’s post hoc test for pairwise comparisons, as detailed in the figure legends. A p-value of less than 0.05 was considered statistically significant.

Results

Significant allergen-specific inflammatory memory in a mouse model of acute allergic asthma

To investigate inflammatory memory in allergic asthma, we established a mouse model of asthma recurrence (Fig. 1A). First, an acute allergic asthma model was generated using a standardized protocol. Mice were sensitized with intraperitoneal ovalbumin (OVA) injections, followed by intranasal challenge with OVA under light isoflurane anesthesia. This procedure induced acute asthma, characterized by pronounced airway inflammation, increased airway mucus production, and significant eosinophil infiltration in the bronchoalveolar lavage fluid (BALF) (Fig. 1B) and airway resistance also increased significantly (Figure S1). After inducing acute asthma, the mice were left untreated for a 6-8-week rest period. During this time, pulmonary inflammation subsided, airway resistance returned to baseline levels, and eosinophil counts in the BALF decreased significantly. We termed these mice, which exhibited resolution of airway inflammation and eosinophil infiltration, as “asthma memory” mice.

Following the rest period, “asthma memory” mice were re-challenged with three intranasal doses of OVA. This re-exposure triggered robust asthmatic responses, including exacerbated pulmonary inflammation, increased mucus production, a marked rise in eosinophil infiltration in the BALF (Fig. 1) and airway resistance also increased significantly (Figure S1). In contrast, when these OVA-sensitized “asthma memory” mice were challenged with house dust mite (HDM) extracts, no notable pulmonary inflammation was observed (Figure S2A). These results demonstrate that the inflammatory memory in this mouse model is allergen-specific. Overall, our findings indicate that this acute allergic asthma mouse model exhibits a strong, allergen-specific inflammatory memory, which can be reactivated by the original sensitizing allergen (Fig. 2A).

To evaluate the persistence of inflammatory memory in acute asthmatic mice, we extended the rest period to 6 and 9 months without additional treatment. Upon re-challenge with OVA after these rest periods, the mice continued to exhibit significant pulmonary inflammatory cell infiltration (Figures S2B and S2C). However, the severity of pulmonary inflammation decreased notably as the rest period lengthened.

Interestingly, while inflammatory cell infiltration was still evident after a 9-month rest period, airway mucus production was no longer observed upon re-challenge. These findings indicate that the inflammatory memory in this acute asthmatic mouse model can persist for at least 9 months but diminishes progressively over time.

T cells and ILC2s are key effector cells in asthma recurrence

Single-cell transcriptomics enables high-throughput analysis of gene expression at the single-cell level, uncovering cellular heterogeneity and identifying functionally distinct subpopulations. This technology is particularly valuable for investigating the mechanisms of complex diseases such as asthma. To explore the cellular mechanisms underlying acute allergic asthma recurrence, we performed single-cell transcriptomic sequencing and bioinformatic analysis on CD45⁺ immune cells isolated from mouse lung tissues at different stages of asthma recurrence (Figure S3).

A total of 142,641 high-quality immune cells passed quality control filtering and were included in the analysis (Control = 40,036; Memory = 49,053; Re-challenge = 53,552). UMAP clustering identified 21 distinct cell clusters, each annotated based on established marker genes (Fig. 2A). These clusters exhibited unique gene expression profiles, reflecting their functional diversity (Fig. 2B).

We analyzed the expression of inflammatory cytokines in pulmonary CD45⁺ immune cells across the different groups. Th1-type cytokines and their associated transcription factors were downregulated in the re-challenge group compared to both the memory and control groups (Fig. 2C). In contrast, type 2 inflammatory cytokines (IL-4, IL-5, and IL-13) were significantly upregulated in the re-challenge group (Fig. 2D).

Among CD45⁺ immune cell subsets, macrophages, neutrophils, and Th17 cells were identified as the primary sources of Th1-type cytokines and transcription factors (Fig. 2E). Conversely, Th2 cells, ILC2s (type 2 innate lymphoid cells), and CD4⁺ T cells were the predominant sources of Th2-type cytokines (Fig. 2F). Interestingly, while IL-9 has been implicated in chronic asthma recurrence models, our analysis revealed no significant IL-9 expression in the lung tissues of any group (Figure S4).

Detailed analysis of T cells revealed 11 distinct subsets, with subcluster 4 exhibiting the highest levels of type 2 inflammatory cytokines. Notably, expression in this subcluster was significantly elevated in the re-challenge group (Figs. 3A, C, E, G). Similarly, ILC2s were categorized into 4 subclusters, and subcluster 0 demonstrated the highest expression of type 2 cytokines, which were markedly increased in the re-challenge group (Figs. 3B, D, F, H). These findings suggest that subcluster 4 of T cells and subcluster 0 of ILC2s are critical effector populations driving asthma recurrence.

Flow cytometry was used to validate these findings. The results confirmed a significant increase in Th2 cells in lung tissues of the re-challenge group, while Th1 and Th9 cell populations did not show corresponding increases (Figure S5). Similarly, ILC2s were significantly elevated in the re-challenge group, with marked upregulation of IL-4 and IL-5. Consistent with the transcriptomic data, IL-9 expression remained negligible across all groups (Figure S6).

Fig. 1

Establishment and validation of a mouse model of acute allergic asthma recurrence. (A) Schematic representation of the experimental design for generating the acute asthma recurrence mouse model. Mice were sensitized to ovalbumin (OVA) via intraperitoneal injection (i.p.) once a week for three consecutive weeks. An acute allergic asthma model was established by challenging the mice with daily intranasal OVA under anesthesia for seven consecutive days. Following this, the asthmatic mice were left untreated for a 6-8-week rest period to generate “memory” asthmatic mice. Recurrence of asthma was induced by re-challenging the “memory” asthmatic mice with three intranasal doses of OVA. (B) Representative images of lung tissue sections stained with hematoxylin and eosin (H&E) and periodic acid-Schiff (PAS) for mucus, as well as H&E-stained bronchoalveolar lavage fluid (BALF), from the control, acute allergic asthma, memory, and re-challenge groups. (C) Quantification of lung inflammation (H&E scores) and mucus production (PAS scores) in paraffin-embedded lung sections from the different groups of mice (n = 6). (D) Total cell counts and eosinophil percentages in BALF from the control, acute allergic asthma, memory, and re-challenge groups (n = 6). Data are presented as mean ± SEM and represent results from two independent experiments. Statistical significance was assessed using one-way analysis of variance (ANOVA) with Tukey’s post hoc test for (C) and (D). Significance levels: *P < 0.05, **P < 0.01, ****P < 0.0001; ns, not significant. BALF, bronchoalveolar lavage fluid

Fig. 2

Immune cell landscape in the lungs of acute asthma recurrence mouse models revealed by scRNA-seq. (A) UMAP visualization showing 21 distinct immune cell clusters identified from scRNA-seq analysis of CD45⁺ immune cells isolated from lung tissues. (B) Violin plot displaying the expression levels of canonical immune cell surface markers across CD45⁺ immune cells, correlating these markers to their respective clusters (clusters 0–21). (C) Dot plot illustrating the expression levels of Th1-associated inflammatory cytokines across different experimental groups. (D) Dot plot illustrating the expression levels of Th2-associated inflammatory cytokines across different experimental groups. (E) Dot plot showing Th1-associated inflammatory cytokine expression across various CD45⁺ immune cell clusters in different experimental groups. (F) Dot plot showing Th2-associated inflammatory cytokine expression across various CD45⁺ immune cell clusters in different experimental groups. The 21 immune cell clusters identified were annotated based on established marker genes and biological functions as follows: Cluster 0: Pro B cells, 1: Pre T cells, 2: NK cells, 3: M2-Alveolar macrophages, 4: CD4⁺T cells, 5: CD8⁺T cells, 6:TR-Alveolar macrophages, 7: Neutrophils, 8: Mature B cells, 9: cDC1, 10: Th17 cells, 11: Plasma cells, 12: Clqc⁺ macrophages, 13: Th2 cells, 14: cDC2, 15: ILC2, 16: Proliferative Plasma cells, 17: Proliferative TR-Alveolar macrophages, 18: pDC, 19: Inflammatory macrophages, 20: Neutrophil-like macrophages

Fig. 3

Subcluster analysis of T cells and ILC2s in the lungs of acute asthma recurrence mouse models. (A) UMAP visualization showing 11 distinct T cell subclusters identified in CD45⁺ immune cells from lung tissues. (B) UMAP visualization showing 4 distinct ILC2 subclusters identified in CD45⁺ immune cells from lung tissues. (C) Dot plot displaying Il4 expression across the 11 T cell subclusters in different experimental groups (control, memory, and re-challenge). (D) Dot plot displaying Il4 expression across the 4 ILC2 subclusters in different experimental groups. (E) Dot plot displaying Il5 expression across the 11 T cell subclusters in different experimental groups. (F) Dot plot displaying Il5 expression across the 4 ILC2 subclusters in different experimental groups. (G) Dot plot displaying Il13 expression across the 11 T cell subclusters in different experimental groups. (H) Dot plot displaying Il13 expression across the 4 ILC2 subclusters in different experimental groups

Cell communication analysis reveals complex intercellular interactions during asthma recurrence

To investigate intercellular communication associated with asthma recurrence, we used the CellChat tool to analyze potential ligand-receptor interactions based on high-resolution single-cell RNA sequencing data. The analysis revealed that macrophages, natural killer (NK) cells, and dendritic cells (DCs) served as central “hubs” of cell communication in both the memory and re-challenge groups, indicating their pivotal roles in intercellular signaling networks (Fig. 4A). In the re-challenge group, specific shifts in cell communication dynamics were observed. B cells, ILC2s, and DCs emerged as the primary signal-receiving cell types, while T cells, ILC2s, and DCs were the predominant signal-sending cell types. Additionally, interactions between B cells and ILC2s, T cells and macrophages, and ILC2s and DCs were significantly enhanced in the re-challenge group, suggesting that these communication pathways are closely associated with asthma recurrence. Conversely, the communication signals involving NK cells and neutrophils with other cell types were notably diminished in the re-challenge group compared to the memory group (Fig. 4B). These findings highlight the dynamic changes in intercellular communication that accompany asthma recurrence and suggest specific cell types and pathways that could serve as therapeutic targets.

Further analysis of cell communication identified significant changes in ligand-receptor interactions in the re-challenge group compared to the memory group (Fig. 4C). Among the upregulated signaling pathways, ligands such as Cdh1, Ccl3, and H2-Q7 were predominantly enriched. Conversely, the downregulated pathways were primarily associated with ligands including Pros1, Ccl6, Ptprc, and Cxcl2.

Scatter plots provided a detailed view of the primary signal-sending and signal-receiving cell types in the memory and re-challenge groups. Macrophages were identified as the major signal-sending cells in both groups, but the interaction strength between cells remained relatively stable in the re-challenge group compared to the memory group. In contrast, ILC2s exhibited significantly increased activity as signal-sending cells in the re-challenge group, highlighting their enhanced role in asthma recurrence (Fig. 4D). Additionally, pathway-level analysis revealed that various asthma-related signaling pathways between different cell types were significantly upregulated in the re-challenge group. These findings are consistent with earlier observations, underscoring the dynamic intercellular communication changes that occur during asthma recurrence (Fig. 4E).

Overall, this analysis provides valuable insights into the intercellular interactions among immune cells during asthma recurrence. The identification of key ligand-receptor pairs and signaling pathways lays the groundwork for future studies aimed at elucidating the molecular mechanisms of asthma recurrence and identifying potential therapeutic targets.

Pseudotime analysis of ILC2s highlights the critical role of Il1rl1 signaling in asthma recurrence

Building on our earlier findings, which highlighted the pivotal role of ILC2s in asthma recurrence, we performed pseudotime analysis to examine the dynamic gene expression profiles of ILC2s during this process. A pseudotime developmental tree was constructed, revealing a distinct branching point along the trajectory (Fig. 5A). As ILC2s progressed along the pseudotime trajectory, their states transitioned into three distinct cell states, reflecting functional changes over time (Fig. 5B). The Il1rl1 gene, encoding the ST2 receptor for IL-33, was identified as a key molecule with increasing expression along the pseudotime trajectory. This upregulation suggests a strong association between Il1rl1 signaling and asthma recurrence (Fig. 5C). Further analysis revealed that Il1rl1 expression was highest in subcluster 4 of T cells during the re-challenge phase in asthmatic mice (Fig. 5D). Similarly, Il1rl1 expression was markedly elevated in subcluster 0 of ILC2s during the re-challenge phase (Fig. 5E). These findings underscore the critical role of Il1rl1 in the pathophysiology of asthma recurrence. The dynamic changes in Il1rl1 expression across specific T cell and ILC2 subclusters highlight its potential as a therapeutic target for managing asthma exacerbations.

To validate the role of Il1rl1 in asthma recurrence, we performed experiments using Il1rl1-deficient mice (Fig. 6). The absence of Il1rl1 did not affect the initial development of asthmatic inflammation. Both wild-type and Il1rl1-deficient mice exhibited comparable levels of pulmonary inflammation and eosinophil infiltration in BALF following initial allergen sensitization and challenge (Fig. 6). Similarly, after a rest period of 6–8 weeks, the asthmatic inflammation subsided in both groups, characterized by reduced airway inflammation and normalization of BALF eosinophil counts. Upon re-challenge following the rest period, the lungs of Il1rl1-deficient mice exhibited significant inflammation, including increased eosinophil infiltration in BALF, airway inflammation, and mucus secretion. However, compared to wild-type mice, Il1rl1-deficient mice showed markedly attenuated pulmonary inflammation after re-challenge. This attenuation was evidenced by a significantly lower total cell count in the BALF, a reduced proportion of eosinophils in BALF, and lower airway inflammation and mucus production scores in lung tissue.

Further analysis revealed a substantial reduction in the number of pulmonary ILC2s (type 2 innate lymphoid cells) in Il1rl1-deficient mice compared to wild-type mice following re-challenge (Fig. 7). Additionally, the proportion of ILC2s producing IL-4 and IL-5 was significantly decreased in Il1rl1-deficient mice, further supporting the critical role of Il1rl1in driving type 2 inflammation during asthma recurrence.

In addition, following rechallenge, the number of CD4⁺ T cells in lung tissues of Il1rl1-deficient mice was significantly reduced compared to wild-type mice, and the expression levels of IL-4 and IL-5 in CD4⁺ T cells were also markedly decreased (Fig. S7).

These findings confirm that while Il1rl1 is not essential for the initial development of asthmatic inflammation, it plays a crucial role in exacerbating pulmonary inflammation during asthma recurrence. The observed reductions in pulmonary ILC2 and CD4⁺ T cells numbers and cytokine production in Il1rl1-deficient mice highlight the pivotal role of Il1rl1 signaling in asthma relapse mechanisms.

Fig. 4

CellChat analysis of cell-cell interactions during asthma recurrence. (A) Circle plot illustrating the number of ligand-receptor (L-R) interactions between seven major cell populations in the memory and re-challenge groups. The strength of interactions between cell population pairs is indicated by the number of L-R pairs, with edge width proportional to the interaction strength. The numbers of L-R pairs are labeled on the edges. (B) Heatmap comparing the number of pairwise cell interactions between the memory and re-challenge groups. The top bar plot represents the sum of incoming signals for each cell cluster, while the right bar plot represents the sum of outgoing signals. Red and blue color bars indicate an increase and decrease in the number of interactions in the re-challenge group compared to the memory group, respectively. (C) Word cloud visualizing the upregulated and downregulated ligands in the re-challenge group. Word size reflects the degree of enrichment for each ligand in the re-challenge group. (D) Scatter plot identifying the major signal-sending (source) and signal-receiving (target) cell populations in the memory and re-challenge groups. Cells on the right side of the x-axis represent the primary signal senders, while those on the left represent the main signal receivers. Circle size indicates the number of cell communication signals. (E) Comparison of significantly upregulated asthma-associated ligand-receptor pairs between the memory and re-challenge groups. The x-axis represents the cell types involved in the communication, with red font denoting memory group cell types and blue font representing re-challenge group cell types. The y-axis lists the ligand-receptor pairs. Circle color indicates the communication probability, with darker red representing higher probabilities and darker blue representing lower probabilities. Circle size reflects statistical significance, with small circles representing 0.01 < P < 0.05 and large circles representing P < 0.01

Fig. 5

Expression of Il1rl1 at different stages in the recurrent asthma model. (A) Pseudotime trajectory analysis of ILC2 development. Each point represents a single cell, with darker colors indicating smaller pseudotime values, corresponding to earlier stages in the trajectory (closer to the root node). (B) Cluster distribution of ILC2s along the pseudotime developmental tree. Different colors represent distinct cell states, highlighting the functional diversity of ILC2s during development. (C) Expression dynamics of Il1rl1 along the pseudotime trajectory. Each point corresponds to a single cell, with different colors representing distinct clusters. The trajectory shows increasing Il1rl1 expression over time. (D) Expression levels of Il1rl1 in different T cell subclusters across experimental groups (control, memory, and re-challenge). (E) Expression levels of Il1rl1 in different ILC2 subclusters across experimental groups, showing significant upregulation in specific subclusters during asthma recurrence

Fig. 6

Effects of Il1rl1 gene knockout on asthma recurrence in mice. (A-D) H&E staining of BALF from Il1rl1 gene knockout mice in the control group (A), asthma group (B), memory group (C), and re-challenge group (D). (E-H) H&E staining of lung tissue sections from Il1rl1 gene knockout mice in the control group (E), asthma group (F), memory group (G), and re-challenge group (H), showing differences in airway inflammation across groups. (I-L) PAS staining of lung tissue sections from Il1rl1 gene knockout mice in the control group (I), asthma group (J), memory group (K), and re-challenge group (L), highlighting mucus production. (M-P) Quantification of BALF cell counts (M, n = 4), the proportion of eosinophils in BALF (N, n = 4), airway inflammation scores (O, n = 6), and airway mucus scores (P, n = 4) across different experimental groups of Il1rl1 gene knockout mice. (Q-T) Comparative analysis between Il1rl1 gene knockout and wild-type mice during asthma re-challenge. Metrics include BALF cell count (Q, n = 5), proportion of eosinophils in BALF (R, n = 6), airway inflammation scores (S, n = 6), and airway mucus scores (T, n = 6)

Fig. 7

Flow cytometry analysis of ILC2 populations and cytokine expression during the re-challenge phase in wild-type and Il1rl1 gene knockout mice. (A) Representative gating strategy for identifying ILC2 cells by flow cytometry. (B) Comparison of the proportion of ILC2 cells among CD45⁺LIN⁻ cells in wild-type and Il1rl1 gene knockout mice during the re-challenge phase (n = 6). (C) Representative flow cytometry plot showing IL-4 expression by ILC2 cells. (D) Quantitative comparison of IL-4 expression by ILC2 cells between wild-type and Il1rl1 gene knockout mice (n = 6). (E) Representative flow cytometry plot showing IL-5 expression by ILC2 cells. (F) Quantitative comparison of IL-5 expression by ILC2 cells between wild-type and Il1rl1 gene knockout mice (n = 6)

Discussion

Our study demonstrates that the acute asthma mouse model exhibits significant features of inflammatory memory, with this asthma-related inflammatory memory persisting for up to nine months. During the recurrence of asthmatic inflammation, both T cells and ILC2 play crucial roles as the primary sources of type 2 inflammatory cytokines. Notably, subcluster 4 of T cells and subcluster 0 of ILC2s may serve pivotal functions in the process of asthma relapse. A complex network of intercellular communication exists during the recurrence of asthmatic inflammation. Specifically, macrophages, NK cells, and DCs act as “hubs” in intercellular signaling. B cells, ILC2, and DCs are the main signal-receiving cell types, while T cells, ILC2s, and DCs are the primary signal-sending cell types. Pseudotime analysis and animal experiments using Il1rl1 gene knockout mice have demonstrated that Il1rl1 plays a critical role in the recurrence of acute asthma.

Asthma is a chronic inflammatory disease of the airways that is prone to recurrent attacks. Our mouse model demonstrated that acute asthmatic mice, even after being left undisturbed for up to nine months, could still develop asthma-like pulmonary inflammation upon re-exposure to allergens. This suggests the presence of a persistent “memory” in the airways of these mice. Using single-cell transcriptomics technology, we identified that CD45⁺ immune cells in mouse lung tissue can be divided into 21 different subpopulations. Upon re-induction of asthmatic inflammation, the expression of classical type 2 inflammatory cytokines (IL-4, IL-5, IL-13) was significantly elevated. These asthma-associated cytokines mainly originated from T cell subpopulations and ILC2 (type 2 innate lymphoid cell) subpopulations. Previous studies using chronic asthma mouse models have demonstrated that IL-9 derived from tissue-resident memory CD4⁺ T cells plays a critical role in asthma recurrence, and blocking the IL-9 signaling pathway can significantly inhibit relapse [4]. In contrast, IL-9 was not detected in our acute asthma model, suggesting significant differences in the mechanisms of recurrence between acute and chronic asthma. In another study that established a recurrence model using an acute asthma mouse model, researchers successfully intervened in asthma relapse by using anti-CD3 antibodies [5]. These studies have all focused on CD4⁺ T cells. However, it remains unclear whether ILC2s, which also play an important role in type 2 inflammation, have a significant role in asthma recurrence. The role of T cells in asthma recurrence is attributed to their memory characteristics [25,26,27]. ILC2s are part of the innate immune system and were traditionally thought not to possess memory features. However, increasing evidence suggests that ILC2s may also exhibit memory functions similar to those of T cells [16, 28,29,30]. Furthermore, interactions between ILC2s and T cells may promote asthma recurrence [31]. Our study revealed that during asthma recurrence, the number of ILC2s increased significantly, and the production of type 2 inflammatory cytokines by ILC2s was also markedly elevated. These findings demonstrate the important role of ILC2s in an acute asthma recurrence model.

ILC2s are recognized as pivotal mediators in allergic inflammation, serving as major producers of interleukin (IL)-5 and IL-13 in murine models of allergic asthma [32]. The function of ILC2s in allergic conditions, including asthma and atopic dermatitis, is critically dependent on the IL-33/ST2 signaling pathway [33,34,35]. IL-33 exerts its biological effects by binding its receptor ST2, encoded by the Il1rl1 gene [36]. The influence of the IL-33/ST2 axis extends beyond ILC2s; IL-33 derived from macrophages, for instance, exacerbates IgE-mediated airway inflammation via mechanisms involving both macrophages and CD4⁺ T cells, while also promoting airway remodeling through macrophage activation [37]. Moreover, studies using a mouse model relevant to human pollinosis demonstrate that the IL-33/ST2 pathway enhances the ability of dendritic cells to drive Th2 immune responses, consequently aggravating allergic inflammation [38]. Studies by Christina et al. have shown that the persistence of chronic asthma in mice is inextricably linked to the interaction between ILC2 and IL-33 [39]. In our research, we found that the expression levels of the Il1rl1 gene progressively increased during the recurrence of acute asthma in mice. However, in Il1rl1 gene-knockout mice, although the initial induction of asthmatic inflammation remained evident, the degree of asthma recurrence upon re-exposure was significantly reduced compared to wild-type mice. Further flow cytometry analyses demonstrated that the knockout of the Il1rl1 gene may alleviate asthma recurrence by affecting the number and function of ILC2. This finding offers new insights for the future prevention and treatment of asthma.

In our study, type 2 innate lymphoid cells (ILC2s) were divided into four distinct subpopulations through single-cell transcriptomic sequencing, among which subpopulation 0 was the primary source of type 2 inflammatory cytokines. This group of cells plays a crucial role in the recurrence of asthma. Unfortunately, we were unable to conduct an in-depth investigation of this cell population, but it is anticipated that this subpopulation may be highly dependent on Il1rl1 signaling. Previous studies have also reported on different ILC2 subpopulations. For instance, research by Clara Wenjing Xia and colleagues demonstrated that ILC2s exhibit diverse subpopulations during tumor progression [40]. Similarly, in asthma research, there are studies on the functions of different ILC2 subpopulations. Yi Jia and collaborators conducted a correlation study between IL-13⁺ ILC2s and asthma control and treatment response, showing that this subpopulation is positively correlated with patients’ asthma control status and is more resistant to glucocorticoid therapy in humans compared to Th2 cells [41]. In our research, subpopulation 0 of ILC2s expressed higher levels of type 2 inflammatory cytokines compared to other subpopulations. It is possible that this subpopulation of ILC2s also possesses memory characteristics, enabling it to rapidly express type 2 inflammatory cytokines in rechallenged mice, thereby promoting asthma recurrence. However, further in-depth studies are needed to confirm this hypothesis.

Our study has several limitations that should be acknowledged. First, while the ovalbumin (OVA)-induced asthma model demonstrates excellent reproducibility in establishing T-helper 2 (Th2) cell-driven eosinophilic airway inflammation - a feature well-aligned with our investigative focus on acute allergic asthma mechanisms - it fails to recapitulate the complexity of chronic, neutrophil-predominant phenotypes frequently observed in severe or steroid-resistant human asthma cases. This discrepancy arises from the fundamental immunological differences between OVA challenge and exposure to clinically relevant aeroallergens such as house dust mite (HDM) extracts. Second, our experimental design prioritized investigating asthma recurrence mechanisms, consequently omitting single-cell sequencing analysis of pulmonary immune populations during the acute asthma phase. This methodological constraint precludes direct comparison of immune cell dynamics between acute and recurrent disease stages, potentially overlooking critical transitional mechanisms in asthma pathogenesis. Third, although pharmacological agents targeting the ST2/IL-33 pathway are currently under clinical development for asthma, chronic obstructive pulmonary disease (COPD), and atopic dermatitis, our findings in the context of asthma recurrence remain preliminary. The therapeutic implications of ST2/IL-33 modulation in recurrent asthma require further validation through both mechanistic studies and clinical trials targeting this specific disease manifestation.

Notwithstanding the aforementioned limitations, our findings provide novel mechanistic insights into asthma recurrence pathophysiology. We demonstrate that acute allergic asthma establishes a recurrence-prone state in murine models, with group 2 innate lymphoid cells (ILC2s) emerging as a pathogenic hub driving relapse processes. The significant attenuation of recurrent inflammation through Il1rl1 gene knockout establishes the ST2/IL-33 axis as a central regulator of asthma recrudescence. These results not only delineate the cellular and molecular architecture of asthma recurrence but also identify potential therapeutic targets for intercepting recurrent exacerbations in allergic airway diseases.