Single-cell and chromatin accessibility profiling reveals regulatory programs of pathogenic Th2 cells in allergic asthma

Abstract

Lung pathogenic T helper type 2 (pTh2) cells are important in mediating allergic asthma, but fundamental questions remain regarding their heterogeneity and epigenetic regulation. Here we investigate immune regulation in allergic asthma by single-cell RNA sequencing in mice challenged with house dust mite, in the presence and absence of histone deacetylase 1 (HDAC1) function. Our analyses indicate two distinct highly proinflammatory subsets of lung pTh2 cells and pinpoint thymic stromal lymphopoietin (TSLP) and Tumour Necrosis Factor Receptor Superfamily (TNFRSF) members as important drivers to generate pTh2 cells in vitro. Using our in vitro model, we uncover how signalling via TSLP and a TNFRSF member shapes chromatin accessibility at the type 2 cytokine gene loci by modulating HDAC1 repressive function. In summary, we have generated insights into pTh2 cell biology and establish an in vitro model for investigating pTh2 cells that proves useful for discovering molecular mechanisms involved in pTh2-mediated allergic asthma.

Fig. 1. scRNA-seq analysis of lung CD4⁺ T cells uncovers the heterogeneity of pTh2 cells in response to HDM.

a Experimental design for scRNA-seq analysis of lung CD4⁺ T cells. Partly created in BioRender. Boucheron, N. (2025) https://BioRender.com/f94i654. After obtaining single-cell suspensions from the lungs of WT (HDAC1^f/f x CD4-Cre^-/-) and HDAC1-cKO (HDAC1^f/f x CD4-Cre^+/-) IL-13 tdTomato-reporter mice that were sensitised and challenged with PBS or HDM (as in Supplementary Fig. 1a), we sorted the following lung CD4⁺ T cells: naïve (TCRβ⁺CD4⁺CD62L⁺CD44^-IL-13^-), IL-13^- Th (TCRβ⁺CD4⁺CD62L^-CD44⁺IL-13^-), and IL-13⁺ Th (TCRβ⁺CD4⁺CD62L^-CD44⁺IL-13⁺) cells. A total of ten samples based on the two genotypes (WT or HDAC1-cKO) and experimental conditions (HDM or PBS) were obtained. Each sample was labelled with a unique hashtag oligonucleotide (HTO). All ten samples were pooled for single-cell RNA-sequencing (scRNA-seq) analysis, and each sample is a pool of cells from three different experimental animals per group. The two independent scRNA-seq experiments were integrated for the analyses. b Uniform manifold approximation and projection (UMAP) of the different lung CD4⁺ T cell clusters identified. c Heatmap shows the top 10 differentially expressed genes (DEGs) per cluster. d–f Comparing the transcriptional signatures of lung peTh2 and Th2 Trm cell subsets identified in the present study to that of published airway pTh2 cells. d, e Enrichment plots depicting the association of lung peTh2 cells (d) and Th2 Trm cell (e) with airway pTh2 cells. The list of all genes in peTh2 cells (cluster 2) and Th2 Trm cells (cluster 5) were compared with the list of DEGs (adjusted P-value < 0.05) in the airway pTh2 gene set. Adjusted P-values were calculated using Seurat’s default two-tailed Wilcoxon rank sum test with Bonferroni correction. For the enrichment plots in d, e P-values were calculated using fgsea’s adaptive multilevel splitting Monte Carlo scheme with Benjamini-Hochberg correction. f Venn diagram depicting the overlap between the DEGs (adjusted P value < 0.05; calculated using Seurat’s two-tailed Wilcoxon rank sum test with Bonferroni correction) in peTh2 and Th2 Trm cells from this study and the DEGs in the published airway pTh2 cells. g Violin plots of selected marker genes associated with lung pTh2 cells. WT wild type, HDAC1-cKO HDAC1-conditional knockout, PBS phosphate-buffered saline, HDM house dust mite, Th T helper, pTh2 pathogenic Th2, peTh2 pathogenic effector Th2, Th2 Trm, pathogenic Th2 Tissue resident memory, FDR false discovery rate, NES normalised enrichment score.

Fig. 2. Comparison of lung ST2⁺ Th subsets.

a Venn diagram depicting the overlapping DEGs (adjusted P-value < 0.05) for the different lung ST2⁺ Th subsets (peTh2, Th2 Trm, Treg/Th2 and Treg) we have identified. Adjusted P-values were calculated using Seurat’s default two-tailed Wilcoxon rank sum test with Bonferroni correction. b Dot plot showing the expression of selected genes. The size of the dot represents the percentage of cells expressing the indicated gene per cluster and the colour intensity indicates the scaled average expression level of the gene. c Violin plots of selected surface markers to distinguish lung ST2⁺ Th subsets^. For c two independent experiments with 10 samples each were performed, and each sample is a pool of cells from three different experimental animals per group. The two independent scRNA-seq data were integrated for the analysis. For the inset box plots, the horizontal line represents the median, the hinges denote the first and third quartiles, and the whiskers denote the minimum and maximum values. peTh2, pathogenic effector Th2, Th2 Trm pathogenic Th2 Tissue-resident memory; Treg, regulatory T cell.

Fig. 3. Flow cytometric characterisation of lung pTh2 subsets in response to HDM.

a–l Flow cytometry characterisation of lung ST2⁺ Th cells in mice sensitised and challenged with PBS or HDM (as in Supplementary Fig. 1a). a–c Confirmation of ST2 expression by lung KLRG1⁺ and CD27⁺ Th cells. a Representative plots showing all lung ST2⁺ Th cells (gated on TCRβ⁺CD4⁺CD44⁺). b Representative plots showing ST2 against KLRG1 expression on Th cells gated on ST2 as depicted in (a) (top) and IL-13 expression by ST2⁺KLRG1⁺ Th cells (bottom). c Representative flow cytometry plots showing ST2 against CD27 expression on Th cells gated on ST2 as depicted in (a) (top) and IL-13 expression by ST2⁺CD27⁺ Th cells (bottom). d Representative plots of FoxP3 and GATA3 expression in lung Th cells (gated on TCRβ⁺CD4⁺CD44⁺ST2⁺KLRG1⁺). e Representative plots of FoxP3 and GATA3 expression in lung Th cells (gated on TCRβ⁺CD4⁺CD44⁺ST2⁺CD27⁺). f, g Defining the distinct subsets of lung ST2⁺ Th cells using CD27 and KLRG1 as markers. f Representative plots showing CD27 and KLRG1 expression by lung Th cells (gated on TCRβ⁺CD4⁺CD44⁺ST2⁺). g Histograms showing the expression of FoxP3 (top) and IL-13 (bottom) by the different ST2⁺ Th subsets in mice exposed to HDM. h–l Gating strategy to define lung peTh2 and Th2 Trm cells. h Flow cytometry plots of CD69 expression (CD69^low and CD69^high) by CD27^-KLRG1^- Th cells in (f). i Flow cytometry plots showing PD1 expression by CD69^low cells in (h), representing peTh2 cells. j IL-13 expression by the peTh2 cells in (i). k Flow cytometry plots showing PD1 expression by CD69^high cells in (h), representing Th2 Trm cells. l IL-13 expression by the Th2 Trm cells in (k). m The schematic summarises the proposed surface markers to distinguish lung peTh2 cells (TCRβ⁺CD4⁺CD44⁺ST2⁺CD27^-KLRG1^-CD69^lowPD1^high) and Th2 Trm cells (TCRβ⁺CD4⁺CD44⁺ST2⁺CD27^-KLRG1^-CD69^highPD1^high) from non-pTh2 ST2⁺ Th cells (ST2⁺FoxP3⁺ Th cells: ST2⁺CD27⁺KLRG1^-, ST2⁺CD27⁺KLRG1⁺, ST2⁺CD27^-KLRG1⁺). Created in BioRender. Khan, M. (2025) https://BioRender.com/w45h785. Data are representative of three independent experiments.

Fig. 4. Loss of HDAC1 augments the pathogenicity of pTh2 subsets and Th2 Trm cell generation.

a UMAP of lung CD4⁺ T cell clusters (left), and UMAP of lung CD4⁺ T cell coloured by origin (sorted samples labelled with unique HTOs), representing cells from HDAC1-cKO mice exposed to either HDM or PBS, and cells from WT mice exposed to either HDM or PBS (right). b Frequencies of IL-13^- Th and IL-13⁺ Th cells from WT and HDAC1-cKO mice exposed to HDM in the peTh2 (top) and Th2 Trm (bottom) clusters as in (a). c, d Violin plots of selected pathogenic marker genes in peTh2 cells (c) and Th2 Trm cells (d) from WT and HDAC1-cKO mice. e Violin plots of marker genes associated with tissue residency. f Representative flow cytometry plots of lung ST2⁺CD69^low and ST2⁺CD69^high cells from WT and HDAC1-cKO mice sensitised and challenged to PBS or HDM. Cells in f are gated on TCRβ⁺CD4⁺CD44⁺ST2⁺CD27^-KLRG1^-. g, h Graphs show the frequency (g) and total number (h) of ST2⁺CD69^high Th cells in (f). For f–h, data are pooled from three independent experiments (n = 8 for WT; n = 13 for HDAC1-cKO). For c and d, two-tailed P-values were generated using the stat_compare_means (Wilcoxon test) function of the ggpubr package. For the inset box plots, the horizontal line represents the median, the hinges denote the first and third quartiles, and the whiskers denote the minimum and maximum values. For g and h data are presented as the mean ± SEM, and statistical analysis was performed using a two-tailed Mann-Whitney U test. Source data (g, h) are provided as a Source Data file.

Fig. 5. Lung pTh2 subsets exhibit shared and distinct transcriptional signatures.

Volcano plots showing a comparison between peTh2 cells and all other clusters (a), Th2 Trm cells and all other clusters (b), and peTh2 cells and Th2 Trm cells (c). The vertical and horizontal lines indicate average Log2 Fold change of ≤ −0.58 and ≥ 0.58, and adjusted P-value < 0.05, respectively. Adjusted P-values were calculated using Seurat’s default two-tailed Wilcoxon rank sum test with Bonferroni correction. d UMAP plots of CD4⁺ T cells colour-coded based on the log-normalised expression levels of selected genes. e GSEA of HALLMARK pathways upregulated (unadjusted P-value < 0.05) in peTh2 cells as compared with all other clusters. f GSEA of HALLMARK pathways upregulated (unadjusted P-value < 0.05) in Th2 Trm cells as compared with all other clusters. g Enrichment plots of the IL-2/STAT5 (left) and TNF/NFκB (right) pathways in peTh2 cells (Fig. 5e). For (e–g), P-values were calculated using fgsea’s adaptive multilevel splitting Monte Carlo scheme with Benjamini-Hochberg correction. h Heatmap shows the leading-edge genes for the pathways in (g). i UMAP plots of CD4⁺ T cells colour-coded based on the log-normalised expression levels of selected genes. IL-2 Interleukin-2, STAT5 signal transducer and activator of transcription factor 5, TNF tumour necrosis factor alpha, NFKB nuclear factor kappa B, EMT epithelial-mesenchymal transition, UV ultraviolet, FDR false discovery rate, NES normalised enrichment score.

Fig. 6. Co-stimulation of GITR and TSLPR drives in vitro differentiation of pTh2 cells.

a Schematic of in vitro differentiation of Th2 cells (left) and pTh2 cells (right). Naïve CD4⁺ T cells (TCRβ⁺CD4⁺CD62L⁺CD44^-) from WT and HDAC1-cKO mice were activated with anti-CD3 and anti-CD28 in the presence of Th2-promoting conditions (IL-4, IL-2, anti-IFN-γ, and anti-TGF-β; collectively referred as IL-4), or Th2-promoting conditions plus TSLP alone, anti-GITR antibody (DTA-1) alone, or both TSLP and DTA-1, then cultured for 5 days. On day 5, cells were restimulated with PMA and ionomycin in the presence of GolgiStop and GolgiPlug for 4 h, and cytokine analyses were performed by flow cytometry. b–g Flow cytometric analysis of cells differentiated under conditions indicated in (a). b Representative plots showing the expression of IL-4 and IL-13 in WT cells (top) and HDAC1-cKO cells (bottom). c Graph shows the frequency of IL-4 and IL-13 co-expressing cells in (b). d Representative plots showing the expression of IL-5 and IL-13 in WT cells (top) and HDAC1-cKO cells (bottom). e Graph shows the frequency of IL-5 and IL-13 co-expressing cells in (d). f Histograms depicting the expression of GATA3 in WT cells (blue) and HDAC1-cKO cells (red). g Graph shows the frequency of GATA3 expressing cells in (f). For c and e, the exact P-values are (IL-4 + TSLP = 0.0059; IL-4 + DTA-1 = 0.0707; IL-4 + TSLP + DTA-1 = < 0.0001) and (IL-4 + TSLP = 0.0151; IL-4 + DTA-1 = 0.0382; IL-4 + TSLP + DTA-1 = <0.0001), respectively. Th17 cells were used as controls for gating. Data are pooled from five independent experiments and presented as the mean ± SEM. Each symbol represents one mouse. Statistical analysis was performed using a Two-way ANOVA with Tukey’s multiple comparisons test. ^*P < 0.05, ^**P < 0.01, ^****P < 0.0001. The schematics in (a) were created using Illustrator. TSLP, thymic stromal lymphopoietin; GITR, glucocorticoid-induced TNFR-related protein. Source data (c, e, g) are provided as a Source Data file.

Fig. 7. Combined transcriptome and proteome profiling reveals a shared signature between in vitro generated pTh2 and lung peTh2 cells.

a Schematic of Bulk RNA-seq of Th2 and pTh2 cells from WT and HDAC1-cKO after 72 h in culture. b Volcano plot of DEGs (adjusted P-value < 0.1) between WT pTh2 and WT Th2 cells. c Volcano plot of DEGs between HDAC1-cKO pTh2 cells and HDAC1-cKO Th2 cells (adjusted P-value < 0.1). For (b, c), two-tailed P-values are based on DESeq2’s Wald test and adjusted using the Bioconductor Independent Hypothesis Weighting package. d–f GSEA of in vitro generated pTh2 cells to lung peTh2 cells. d Enrichment plot showing a comparison of WT pTh2 cells and WT Th2 cells to lung peTh2 cells. e Enrichment plot showing comparison of HDAC1-cKO pTh2 cells and HDAC1-cKO Th2 cells to lung peTh2 cells. f Enrichment plot showing comparison of HDAC1-cKO Th2 cells and WT Th2 cells to lung peTh2 cells. DEGs (adjusted P-value < 0.1; two-tailed P-values obtained by DESeq2’s Wald test and adjusted using the Bioconductor Independent Hypothesis Weighting package) between WT pTh2 vs WT Th2, HDAC1-cKO pTh2 vs HDAC1-cKO Th2, and HDAC1-cKO Th2 vs WT Th2 (Supplementary Data 7) were used to compare with lung peTh2 gene set (DEGs; adjusted P-value < 0.05 based on Seurat’s two-tailed Wilcoxon rank sum test with Bonferroni correction; Supplementary Data 2). For the enrichment plots in (d–f), P-values were calculated using fgsea’s adaptive multilevel splitting Monte Carlo scheme with Benjamini-Hochberg correction. g Schematic of proteomics analysis of in vitro generated Th2 and pTh2 cells from WT and HDAC1-cKO as in (a). Volcano plots depicting proteomics comparison between WT pTh2 and WT Th2 cells (h) and HDAC1-cKO pTh2 and HDAC1-cKO Th2 cells (i). For (h, i) two-tailed P-values were obtained by Limma moderated t-test with Benjamini-Hochberg correction. j Heatmap showing normalised reporter ion intensity values of selected proteins. For (b, c, h, i) the vertical and horizontal lines indicate Log2 Fold change of ≤−0.58 and ≥0.58, and adjusted P-value < 0.05, respectively. Transcriptomic and proteomic data are from three and four independent experiments, respectively. One WT pTh2 sample was excluded from the proteomics analyses due to poor sample quality.

Fig. 8. The p38 MAPK pathway regulates IL-5 and IL-13 expression in pTh2 cells.

a–f Impact of targeting AP-1 or p38 MAPK pathways on pTh2 cells generated in vitro. Th2 cells alone, pTh2 cells alone, pTh2 cells treated with an AP-1 inhibitor (T-5224; 10 µM), and pTh2 cells treated with a p38 MAPK inhibitor (SB 203580; 10 µM) were cultured for five days under Th2 and pTh2-promoting conditions (Fig. 7a). On day 5, cells were restimulated with PMA and ionomycin in the presence of GolgiStop and GolgiPlug for 4 h before cytokine analyses by flow cytometry. Flow cytometric analysis of WT cells. a Representative flow cytometry plots showing IL-5 and IL-13 expression under indicated conditions. b Graph shows the frequency of IL-5 and IL-13 co-expressing cells in (a). c Graphs showing the frequencies of IL-4, IL-5, IL-13, and GATA3 single-expressing cells. d–f, Flow cytometric analysis of HDAC1-cKO cells. d Representative flow cytometry plots showing IL-5 and IL-13 co-expression. e Graph shows the frequency of IL-5 and IL-13 co-expressing cells in (d). f Graphs showing frequencies of IL-4, IL-5, IL-13, and GATA3 single-expressing cells. Data are pooled from three independent experiments and presented as the mean ± SEM. Each symbol represents one mouse. Statistical analysis was performed using a one-way ANOVA with Tukey’s multiple comparisons test. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. AP-1, activator protein-1; p38 MAPK, p38 mitogen-activated protein kinase. Source data (b, c, e, f) are provided as a Source Data file.

Fig. 9. Type 2 cytokine expression is differentially regulated at chromatin level in pathogenic versus classical Th2 cells.

a Heatmap representing normalised ATAC-seq peaks of Th2 cytokines from WT and HDAC1-cKO Th2 and pTh2 cells generated in vitro after 24 or 48 h of culture. b UMAP of ATAC-seq profiles of WT and HDAC1-cKO Th2 and pTh2 cells that were cultured for 24 (left panel) or 48 h (right panel). Clustering of WT Th2 and pTh2 are highlighted with dashed circles at 48 h. c ATAC-seq signal tracks at the Th2 cytokine locus in WT and HDAC1-cKO Th2 and pTh2 cells generated in vitro after 48 h of culture. The average signals from three replicates are shown as one track. Statistically significant differences are highlighted as grey bars. Relative positions of the integration sites to the known DNase I hypersensitive sites (squares) and conserved noncoding regions/elements/control regions (arrows) are shown. HSS and LCR were coloured for a better visualisation. d Volcano plot showing a comparison between peTh2 cells (cluster 2) from scRNA-seq analysis versus all other effector (“non-naïve”) clusters. Major Th2 related Transcription factors are shown in red and cytokines in blue. e Volcano plot of chromVAR inferred TF binding sites between pTh2 and Th2 cells generated in vitro. GATA3 and a prominent AP-1 TF node were coloured to improve visualisation. f Schematic representation of the working model. Created in BioRender. Boucheron, N. (2025) https://BioRender.com/x77f144. In summary, conventional Th2 cells have a reduced chromatin accessibility, with high HDAC1 activity and low Il4 and Il13 expression. In contrast, pTh2 cells have enhanced chromatin accessibility, with a reduced/overridden HDAC1 activity coupled with high Il4, Il13, and Il5 expression.