Interleukin-33 activates regulatory T cells to suppress innate γδ T cell responses in the lung

Abstract

Foxp3⁺ regulatory T (T_reg) cells expressing the interleukin (IL)-33 receptor ST2 mediate tissue repair in response to IL-33. Whether T_reg cells also respond to the alarmin IL-33 to regulate specific aspects of the immune response is not known. Here we describe an unexpected function of ST2⁺ T_reg cells in suppressing the innate immune response in the lung to environmental allergens without altering the adaptive immune response. Following allergen exposure, ST2⁺ T_reg cells were activated by IL-33 to suppress IL-17-producing γδ T cells. ST2 signaling in T_reg cells induced Ebi3, a component of the heterodimeric cytokine IL-35 that was required for T_reg cell-mediated suppression of γδ T cells. This response resulted in fewer eosinophil-attracting chemokines and reduced eosinophil recruitment into the lung, which was beneficial to the host in reducing allergen-induced inflammation. Thus, we define a fundamental role for ST2⁺ T_reg cells in the lung as a negative regulator of the early innate γδ T cell response to mucosal injury.

Extended Data Fig. 1. Profiling ST2-expressing T_reg and T_H cells in the lungs of HDM-treated mice.

a-d, Foxp3^YFPcre mice were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. a, Schematic of experimental design. b, NanoString analysis of gene expression comparison between flow sort-purified CD4⁺Foxp3⁻ T helper (T_H) cells expressing ST2 (T_H2) or not (ST2⁻ T_H) from lung parenchyma. Data are presented as volcano plot from three independent experiments of pooled mice (n = 7 mice per experiment). Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used to obtain the P values. c, Representative flow cytometry plots for IL-13 and IL-5 (top), and for IL-17A and IFN-γ (bottom) in T_H2 and ST2⁻ T_H cells from lung parenchyma. d, Representative histograms showing the expression of the indicated transcription factors in ST2⁺ T_reg cells (red) or ST2⁻ T_reg cells (blue) from lung parenchyma. Isotype control is shown in gray.

Extended Data Fig. 2. Subject characteristics.

Aeroallergen-allergic human subjects screened for eligibility with a full medical history, baseline spirometry, methacholine challenge, and allergen skin testing to confirm allergy to either cat dander or Dermatophagoides pteronyssinus)(DP). The threshold level of allergen sensitivity was determined by skin prick test titration using serial threefold dilutions of allergen extract.

Extended Data Fig. 3. IL-33 activates ST2⁺ T_reg cells in the lung after HDM exposure.

a-c, Wild-type (WT) and Il33^−/− mice were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 10. a, Representative flow cytometry for Foxp3 and ST2 in CD4⁺ T cells from lung parenchyma. b, Percentage of ST2⁺ T_reg cells, ST2⁻ T_reg cells, and T_H2 (Foxp3⁻ST2⁺) cells in the lung. c, Representative histograms showing the expression of the indicated surface markers in ST2⁺ T_reg and T_H2 cells from the lung parenchyma of WT (black) and Il33^−/− (gray) mice. Data represent one experiment with n = 6 mice per group of two independent experiments in panels (a-c). Unpaired two-tailed t test was used for statistical analysis in panel (b). Error bars denote mean ± s.d. P values are indicated in the figure.

Extended Data Fig. 4. T_reg cell-specific deletion of ST2 does not result in systemic alterations in the number or activation of T cells.

Thymus, spleen, cervical lymph nodes (cLN) and mesenteric lymph nodes (MLN) were harvested from naive Il1rl1^fl/flFoxp3^YFPcre mice, Il1rl1^fl/+Foxp3^YFPcre mice, and Il1rl1^+/+Foxp3^YFPcre (Foxp3^YFPcre) littermate control mice for flow cytometric analysis. a, Representative flow cytometry for Foxp3 and CD4 in CD4⁺ T cells (top) and percentage of Foxp3⁺ T_reg cells (bottom) from the indicated tissues. b, Representative flow cytometry for CD8 and CD4 in CD3⁺ T cells (top) and percentage of CD8⁺ T cells and CD4⁺ T cells (bottom) from the indicated tissues. c, Representative flow cytometry for CD44 and CD62L in CD4⁺ T cells (top) and percentage of CD44⁺CD4⁺ T cells (bottom) from the indicated tissues. Data represent one experiment (Foxp3^YFPcre n = 5; Il1rl1^fl/+Foxp3^YFPcre n = 3; Il1rl1^fl/flFoxp3^YFPcre n = 4 mice per group) of two independent experiments in panels (a-c). Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panels (a-c bottom). No statistical difference was found. Error bars denote mean ± s.d.

Extended Data Fig. 5. Immune response in the lung to HDM in mice with a T_reg cell-specific deletion of ST2.

a-c, Il1rl1^fl/flFoxp3^YFPcre mice and Foxp3^YFPcre littermate controls were left untreated (Naive) or were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. a,b, Representative flow cytometry for IL-13 and IL-5 (a) and for IFN-γ and IL-17A (b) in CD4⁺Foxp3⁻ (T_H) cells from lung parenchyma. Data represent one experiment with n = 5 mice per group of two independent experiments in panels (a,b). c, Lung explants from Naive and HDM-treated mice were re-stimulated ex vivo with HDM and the indicated cytokines measured in culture supernatants 72 h later by ELISA. Data represent one experiment with n = 4 mice per group of two independent experiments. Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis. Error bars denote mean ± s.d. P values are indicated in the figure.

Extended Data Fig. 6. Innate lymphocyte response in the lung to HDM.

a-g, Il1rl1^fl/flFoxp3^YFPcre mice and Foxp3^YFPcre littermate controls were left untreated (Naive) or were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. a, Gating strategy for flow cytometric analysis of innate lymphocytes in lung parenchyma. b-d, Percentage of ILC1 (T-bet⁺), ILC2 (ST2⁺), and ILC3 (ST2⁻T-bet⁻) of CD3⁻Lin⁻CD127⁺ cells (b), percentage of natural killer (NK) cells of CD3⁻TCRαβ⁻ cells (c), and percentage of NKT cells of CD3⁺TCRγδ⁻ cells (d) in the lung parenchyma. e-g, Percentage of lung ILCs (e), NK (f), and NKT cells (g) expressing IL-13 and IL-5, IFN-γ, or IL-17A. Data represent one experiment with n = 5 mice per group of three independent experiments in panels (b-g). Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panels (b-g). No statistical difference was found. Error bars denote mean ± s.d.

Extended Data Fig. 7. Lung and airway inflammatory response to HDM and A. alternata in mice with a T_reg cell-specific deletion of ST2.

a,b, Il1rl1^fl/flFoxp3^YFPcre mice and Foxp3^YFPcre littermate controls were left untreated (Naive) or were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. a, Gating strategy for flow cytometric analysis of eosinophils (CD11c⁻Siglec-F⁺), neutrophils (Siglec-F⁻CD11b⁺Ly6G⁺), and Ly6C⁺ inflammatory monocytes (Ly6C⁺ iMo, Siglec-F⁻CD11b⁺Ly6C⁺) in the lung parenchyma of HDM-treated mice. b, Quantification of cell differential counts in BAL. Mn - Mononuclear cells; Ne - Neutrophils; Eo - Eosinophils. Data represent one experiment (Foxp3^YFPcre Naive n = 3; Foxp3^YFPcre HDM n = 5; Il1rl1^fl/flFoxp3^YFPcre Naive n = 3; Il1rl1^fl/flFoxp3^YFPcre HDM n = 5 mice per group) of two independent experiments. Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis. c, Il1rl1^fl/flFoxp3^YFPcre mice and Foxp3^YFPcre littermate controls were treated with A. alternata i.n. on days 0, 1 and 17–19 and analyzed on day 20. Gating strategy for flow cytometric analysis of eosinophils and neutrophils in the lung parenchyma of A. alternata-treated mice. Error bars denote mean ± s.d. P values are indicated in the figure.

Extended Data Fig. 8. Decreased lung CCL11 and CCL24 and allergic pulmonary inflammation in γδ T cell-deficient mice.

a-c, TCRγδ-deficient (Tcrd^−/−) and WT mice were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. a, Lung Ccl11 and Ccl24 mRNA levels relative to β2M determined by RT-qPCR. Data represent one experiment with n = 4 mice per group of two independent experiments. b, Number of eosinophils (CD11c⁻Siglec-F⁺) and neutrophils (Siglec-F⁻CD11b⁺Ly6G⁺) in lung parenchyma. c, Quantification of cell differential counts in BAL. Mn - Mononuclear cells; Ne - Neutrophils; Eo - Eosinophils. Data represent one experiment (WT n = 5; Tcrd^−/− n = 6 mice per group) of two independent experiments in panels (b,c). Unpaired two-tailed t test was used for statistical analysis in panels (a,b) and unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panel (c). Error bars denote mean ± s.d. P values are indicated in the figure.

Extended Data Fig. 9. ST2⁺ T_reg cell-derived Ebi3 suppresses γδ T cell responses in the lung to influenza infection.

a-d, Foxp3^YFPcre, Il1rl1^fl/flFoxp3^YFPcre and Ebi3^fl/flFoxp3^YFPcre mice were infected with PR8 influenza and lungs were harvested for flow cytometry analysis in the indicated time points. a, Representative flow cytometry for IL-17A and IFN-γ in γδ T cells from lung parenchyma on day 11 post-infection. b, Number of IL-17A⁺ γδ T cells in the lung over the course of infection. Data pooled from two independent experiments (day 0: Foxp3^YFPcre n = 5 and Il1rl1^fl/flFoxp3^YFPcre n = 4; day 4: Foxp3^YFPcre n = 18 and Il1rl1^fl/flFoxp3^YFPcre n = 16; day 7: Foxp3^YFPcre n = 10 and Il1rl1^fl/flFoxp3^YFPcre n = 12; day 11: Foxp3^YFPcre n = 9 and Il1rl1^fl/flFoxp3^YFPcre n = 9 mice from the two experiments in each time point). Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panel (b). c, Representative flow cytometry for TCRαβ and TCRγδ in CD3⁺ T cells (top) and for IL-17A and IFN-γ in γδ T cells (bottom) from lung parenchyma on day 7 post-infection. d, Number of lung IL-17A⁺ γδ T cells. Data pooled from two independent experiments (Foxp3^YFPcre n = 10; Il1rl1^fl/flFoxp3^YFPcre n = 8; Ebi3^fl/flFoxp3^YFPcre n = 10 mice from the two experiments) in panel (d). Ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for statistical analysis in panel (d). Error bars denote mean ± s.d. P values are indicated in the figure.

Figure 1.. ST2⁺ T_reg cells expand rapidly in the lung and airways of mice and humans after aeroallergen exposure.

a-d, Foxp3^YFPcre mice were left untreated (Naive) or were treated with HDM i.n. on days 0 and 7–11 and analyzed on days 10 and/or 14. a, Gating strategy for flow cytometric analysis of CD4⁺ T cells in the lung vasculature that were labeled with anti-CD45 antibody given intravenously (CD45 i.v. labeled cells) or in the lung parenchyma (CD45 i.v. unlabeled cells). b,c, Representative flow cytometry for Foxp3 and ST2 in CD4⁺ T cells (b) and numbers of ST2⁺ T_reg cells (red), ST2⁻ T_reg cells (blue) and T_H2 (Foxp3⁻ST2⁺) cells (green) (c) in the lung vasculature, lung parenchyma, and bronchoalveolar lavage (BAL). Data represent one experiment (day 0 n = 3 mice; day 10 n = 4 mice; day 14 n = 4 mice) of two independent experiments in panels (b,c). Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panel (c). P values between ST2⁺ and ST2⁻ T_reg cells are shown. d,e, HDM or cat dander allergic human subjects (n = 11) underwent bronchoscopic segmental allergen challenge (SAC) with HDM or cat dander, respectively, and BAL and blood were collected before (Pre-SAC) and 24 h after segmental allergen (Post-SAC) or diluent (Post-Diluent) challenge. d, Representative flow cytometry for Foxp3 and ST2 in CD4⁺ T cells (left) and percentage of ST2⁺ T_reg cells (right) from BAL (orange) and blood (black) pre- and post-SAC. Each symbol represents one subject. Repeated measures (RM) two-way ANOVA with Sidak’s multiple comparisons test was used for statistical analysis. e, Number of ST2⁺ T_reg cells recovered from pre-SAC, Post-Diluent and Post-SAC BAL. Each symbol represents one subject. RM one-way ANOVA with Tukey’s multiple comparisons test was used for statistical analysis. Error bars denote mean ± s.d. P values are indicated in the figure.

Figure 2.. ST2⁺ T_reg cells respond to IL-33 but not directly to allergen peptide.

a, Foxp3^YFPcre mice were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. Representative flow cytometry for MHC class II I-A^b:Der p 1_117–127 (I-A(b):Derp1) tetramer (left) and percentage of tetramer-specific CD4⁺ T cell populations (right) from lungs after tetramer-based cell enrichment. Data pooled from seven independent experiments. Each symbol represents a pool of n = 2 mice per experiment. Ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for statistical analysis in panel (a, right). b, Representative flow cytometry for ST2 on CD45.1 and CD45.2 CD4⁺Foxp3⁺ (T_reg) and CD4⁺Foxp3⁻ (T_H) cells in the lung parenchyma (left) and percentage of lung ST2⁺ T_reg cells and ST2⁺Foxp3⁻ (T_H2) cells over total CD4⁺Foxp3⁺ T_reg and CD4⁺Foxp3⁻ T_H cells, respectively (right). Data pooled from three independent experiments (n = 8 parabiontic pairs from the three experiments). Each symbol represents one mouse. Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panel (b, right). c-e, Wild-type (WT) and Il33^−/− mice were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 10. Mean fluorescence intensity (MFI) of ST2 (c), CD69 (d), and CD44 (e) on ST2⁺ T_reg and T_H2 cells isolated from lung parenchyma. Data represent one experiment with n = 6 mice per group of two independent experiments in panels (c-e). Unpaired two-tailed t test was used for statistical analysis in panels (c-e). Error bars denote mean ± s.d. P values are indicated in the figure.

Figure 3.. T_reg cell-specific deletion of ST2 increases γδ T cells in the lung in response to HDM.

a-g, Il1rl1^fl/flFoxp3^YFPcre mice and Foxp3^YFPcre littermate controls were left untreated (Naive) or were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. a, Representative flow cytometry for Foxp3 and ST2 in CD4⁺ T cells from lung parenchyma. b,c, Number of total T_reg (b) and T_H2 (Foxp3⁻ST2⁺) cells (c) in the indicated groups. d, Percentage of lung CD4⁺Foxp3⁻ T helper (T_H) cells expressing IL-5, IL-13, IFN-γ, or IL-17A. Data represent one experiment with n = 5 mice per group of two independent experiments in panels (b-d). Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panels (b-d). No statistical difference was found. e, Lung explants from HDM-treated mice were re-stimulated ex vivo with HDM and the indicated cytokines measured in culture supernatants 72 h later by ELISA. Data are presented as heatmap from one experiment with n = 4 mice per group of two independent experiments. Unpaired two-tailed t test was used for statistical analysis. *** P = 0.000897. f, Number of different innate lymphocytes in the lung parenchyma. Data represent one experiment with n = 5 mice per group of three independent experiments. Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis. g, Representative confocal microscopy images of γδ T cells identified with an anti-TCRγδ mAb (left), and quantification of total γδ T cells/400 μm² area (right) in lung sections from three HDM-treated Il1rl1^fl/flFoxp3^YFPcre mice (n = 7 lung sections) and three HDM-treated Foxp3^YFPcre littermate controls (n = 8 lung sections). Each symbol represents the analysis of one lung section. Scale bars, 100 μm. Unpaired two-tailed t test was used for statistical analysis in panel (g, right). Error bars denote mean ± s.d. P values are indicated in the figure or otherwise indicated in the legend.

Figure 4.. IL-33–ST2 signaling in T_reg cells restrains γδ T cell activation and function.

a-f, Il1rl1^fl/flFoxp3^YFPcre mice and Foxp3^YFPcre littermate controls were left untreated (Naive) or were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. a, Representative images of lung sections stained for CD4 (red), YFP (Foxp3, green) and TCRγδ (blue). Data are representative images of lungs from three HDM-treated mice of each strain. White arrowhead indicates the T_reg – γδ T cell interaction. Scale bars, 100 μm. b, Quantification of the distance between each T_reg and γδ T cells (T_reg – γδ T cell interaction for Foxp3^YFPcre n = 101 and for Il1rl1^fl/flFoxp3^YFPcre n = 475) in eight different lung sections analyzed from three HDM-treated mice of each strain. Unpaired two-tailed t test was used for statistical analysis. c, Percentage of T_reg cells within the indicated distance (μm) of each γδ T cell. Each symbol represents pooled data from analysis of one lung section (n = 8 lung sections) from three HDM-treated mice of each strain. Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis. d, Representative histograms showing CD69 and ICOS expression on CD3⁺TCRγδ⁺ (γδ T) cells from the lung parenchyma of Foxp3^YFPcre (black) or Il1rl1^fl/flFoxp3^YFPcre mice (blue). Isotype control is shown in gray. Numbers within histogram are mean ± s.d. of fluorescence intensity (MFI) for each markers of the indicated strain. Data represent one experiment with n = 3 mice per group of two independent experiments. Unpaired two-tailed t test was used for statistical analysis. **P=0.0018 and ***P=0.00012. e,f, Representative flow cytometry for IL-17A and IFN-γ in γδ T cells (e) and numbers of IL-17A⁺ γδ T cells (f) from the lung parenchyma. Data represent one experiment with n = 5 mice per group of three independent experiments in panels (e,f). Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panel (f). g-j, WT and Il33^−/− mice were treated with HDM i.n. on days 0 and 7–9 and analyzed on day 10. g,h, Representative histogram (g) and MFI (h) of CD69 expression in lung γδ T cells. i,j, Representative flow cytometry for IL-17A and IFN-γ in γδ T cells (i) and number of IL-17A⁺ γδ T cells (j) from lung parenchyma. Data represent one experiment with n = 6 mice per group of two independent experiments in panels (g-j). Unpaired two-tailed t test was used for statistical analysis in panels (h,j). k-n, HDM-exposed Foxp3^YFPcre mice were treated daily with soluble ST2 IgG-Fc fusion protein (sST2-Fc) to block IL-33 signaling or isotype control (IgG2a) i.v. during the allergen challenges (day 7–9) and analyzed on day 10. k,l, Representative histogram (k) and MFI (l) of CD69 expression in lung γδ T cells. m,n, Representative flow cytometry for IL-17A and IFN-γ in γδ T cells (m) and number of IL-17A⁺ γδ T cells (n) from lung parenchyma. Data represent one experiment with n = 5 mice per group of two independent experiments in panels (k-n). Unpaired two-tailed t test was used for statistical analysis in panels (l,n). Error bars denote mean ± s.d. P values are indicated in the figure or otherwise indicated in the legend.

Figure 5.. γδ T cells mediate exacerbated allergic pulmonary inflammation in mice lacking ST2 in T_reg cells.

a-g, Mouse strains indicated in different colors were left untreated (Naive) or were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. a, Histopathology of lung sections stained for H&E and periodic acid-Schiff (PAS). Scale bars, 100 μm. b, Inflammation score. c, Mucus score (left) and RT-qPCR analysis of Muc5ac mRNA levels relative to β2 microglobulin (β2M) in total lung (right). Data represent one experiment with n = 5 mice per group in panels (a,b,c left) and one experiment (Foxp3^YFPcre Naive n = 4; Foxp3^YFPcre HDM n = 6; Il1rl1^fl/flFoxp3^YFPcre Naive n = 3; Il1rl1^fl/flFoxp3^YFPcre HDM n = 6) in panel (c right) of two independent experiments. d, Number of eosinophils (CD11c⁻Siglec-F⁺), neutrophils (Siglec-F⁻CD11b⁺Ly6G⁺) and Ly6C⁺ inflammatory monocytes (Siglec-F⁻CD11b⁺Ly6C⁺) in lung parenchyma. Data represent one experiment with n = 5 mice per group of three independent experiments. Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panels (b-d). e, Representative flow cytometry for CD4 and TCRγδ in CD3⁺ T cells from lung parenchyma of HDM-treated mice. f, Number of γδ T cells from lung parenchyma of HDM-treated mice. g, Number of eosinophils and neutrophils in lung parenchyma. Data pooled from three independent experiments (Foxp3^YFPcre n = 11; Il1rl1^fl/flFoxp3^YFPcre n = 12; Il1rl1^fl/flFoxp3^YFPcreTcrd^−/− n = 11 mice from the three experiments) in panels (f,g). Ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for statistical analysis in panels (f,g). h-k, Il1rl1^fl/flFoxp3^YFPcre mice and Foxp3^YFPcre littermate controls were treated with A. alternata i.n. on days 0, 1 and 17–19 and analyzed on day 20. h, Representative flow cytometry for CD4 and TCRγδ in CD3⁺ T cells (top) and for IL-17A and IFN-γ in γδ T cells (bottom) from lung parenchyma. i, Number of lung IL-17A⁺ γδ T cells. j, Representative histogram of CD69 expression in lung γδ T cells. Numbers within histogram are mean ± s.d. of fluorescence intensity of CD69 for the indicated strain. ***P = 0.00086. k, Number of eosinophils and neutrophils in lung parenchyma. Data pooled from two independent experiments (Foxp3^YFPcre n = 13; Il1rl1^fl/flFoxp3^YFPcre n = 7 mice from the two experiments) in panels (i,k) or represent one experiment (Foxp3^YFPcre n = 7; Il1rl1^fl/flFoxp3^YFPcre n = 4 mice) of two independent experiments in panel (j). Unpaired two-tailed t test was used for statistical analysis in panels (i-k). Error bars denote mean ± s.d. P values are indicated in the figure or otherwise indicated in the legend.

Figure 6.. IL-17 blockade decreases eosinophil-attracting chemokine production and eosinophil recruitment into the lung.

a,b, Il1rl1^fl/flFoxp3^YFPcre mice and Foxp3^YFPcre littermate controls were left untreated (Naive) or were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. a, Levels of the indicated lung chemokine mRNA relative to β2 microglobulin (β2M) determined by RT-qPCR. Data pooled from two independent experiments (Foxp3^YFPcre Naive n = 6; Foxp3^YFPcre HDM n = 6; Il1rl1^fl/flFoxp3^YFPcre Naive n = 6; Il1rl1^fl/flFoxp3^YFPcre HDM n = 8 mice from the two experiments). b, Levels of CCL11 and CCL24 protein in lung homogenates. Data represent one experiment (Foxp3^YFPcre Naive n = 3; Foxp3^YFPcre HDM n = 4; Il1rl1^fl/flFoxp3^YFPcre Naive n = 3; Il1rl1^fl/flFoxp3^YFPcre HDM n = 5 mice) of two independent experiments. Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panels (a,b). c-e, HDM-exposed Il1rl1^fl/flFoxp3^YFPcre mice were treated daily with anti-IL-17A blocking mAb or isotype control (IgG2a) i.n. during the allergen challenges (day 7–11) and analyzed on day 14. c, Lung Ccl11 and Ccl24 mRNA levels relative to β2M determined by RT-qPCR. d, Number of eosinophils (CD11c⁻Siglec-F⁺) and neutrophils (Siglec-F⁻CD11b⁺Ly6G⁺) in lung parenchyma. e, Quantification of cell differential counts in BAL. Mn - Mononuclear cells; Ne - Neutrophils; Eo - Eosinophils. Data pooled from two independent experiments (IgG2a n = 8; anti-IL-17A n = 9 mice from the two experiments) in panels (c-e). Unpaired two-tailed t test was used for statistical analysis in panels (c,d) and unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panel (e). Error bars denote mean ± s.d. P values are indicated in the figure.

Figure 7.. ST2⁺ T_reg cell-derived Ebi3 is required to suppress γδ T cell IL-17 responses in the lung.

a-c, Foxp3^YFPcre mice were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. a, NanoString analysis of comparative gene expression between flow sort-purified ST2⁺ and ST2⁻ T_reg cell populations isolated from lung parenchyma. Data are presented as volcano plot from three independent experiments of pooled mice (n = 7 per experiment). Unpaired two-taile t test with Holm-Sidak correction for multiple comparisons was used to obtain the P values. b, RT-qPCR analysis of Ebi3, Il12a (IL-12p35) and Il27 (IL-27p28) in flow sort-purified ST2⁺ and ST2⁻ T_reg cell populations from lung parenchyma. Data pooled from four independent experiments. Each symbol represents a pool of n = 7 mice per experiment. Unpaired two-tailed t test was used for statistical analysis. c, Representative histogram of IL-17RB (IL-25R) expression on ST2⁺ T_reg cells (red) and ST2⁻ T_reg cells (blue) from lung parenchyma. Isotype control is shown in gray. Data represent one experiment with n = 3 mice per group of two independent experiments. d, Flow sort-purified splenic ST2⁺ T_reg cells from naive Foxp3^YFPcre mice were stimulated in vitro with the indicated stimuli and RT-qPCR analysis of Ebi3 and Il12a performed 72 h later. Data are triplicates of pooled mice (n = 4) and represent one of two independent experiments. Ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for statistical analysis. e, Flow sort-purified splenic T_reg cells from naive Foxp3^YFPcre mice (WT T_reg), Il1rl1^fl/flFoxp3^YFPcre mice (Il1rl1^−/− T_reg), or from Ebi3^fl/flFoxp3^YFPcre mice (Ebi3^−/− T_reg) were stimulated in vitro as indicated and Ebi3 and IL-12p35 levels determined by ELISA 72 h later. Data are triplicates of pooled mice (n = 3) and represent one of two independent experiments. Unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panel (e). f-i, Foxp3^YFPcre, Il1rl1^fl/flFoxp3^YFPcre and Ebi3^fl/flFoxp3^YFPcre mice were treated with HDM i.n. on days 0 and 7–11 and analyzed on day 14. f, Representative flow cytometry for IL-17A and IFN-γ in γδ T cells from lung parenchyma. g, Number of lung IL-17A⁺ γδ T cells. h, Number of eosinophils (CD11c⁻Siglec-F⁺) and neutrophils (Siglec-F⁻CD11b⁺Ly6G⁺) in lung parenchyma. i, Quantification of cell differential counts in BAL. Mn - Mononuclear cells; Ne - Neutrophils; Eo - Eosinophils. Data pooled from two independent experiments (Foxp3^YFPcre n = 7; Il1rl1^fl/flFoxp3^YFPcre n = 9; Ebi3^fl/flFoxp3^YFPcre n = 7 mice from the two experiments) in panels (g-i). Ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for statistical analysis in panels (g,h) and unpaired two-tailed t test with Holm-Sidak correction for multiple comparisons was used for statistical analysis in panel (i). Error bars denote mean ± s.d. P values are indicated in the figure.

Figure 8.. IL-35 suppresses γδ T cell IL-17 responses in the lung induced by HDM.

a, RT-qPCR analysis of Il6st (left) and Il12rb2 (right), the two chains of the IL-35 receptor, in flow sort-purified lymphocytes from lung parenchyma of HDM-treated Foxp3^YFPcre mice. Data pooled from three independent experiments. Each symbol represents a pool of n = 4 mice per experiment. b, Flow sort-purified lung γδ T cells from HDM-treated Foxp3^YFPcre mice were stimulated in vitro as indicated and IL-17A protein levels determined by ELISA 72 h later. Data are triplicates of a pool of n = 4 mice and represent one of two independent experiments. Ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for statistical analysis in panels (a,b). c-f, HDM-exposed Foxp3^YFPcre mice were treated with mAbs that specifically neutralize IL-35 or IL-27, or with a mAb isotype control (IgG2b) i.p. on days 6 and 11 and analyzed on day 14. c, Representative flow cytometry for CD4 and TCRγδ in CD3⁺ T cells (top) and for IL-17A and IFN-γ in γδ T cells (bottom) from lung parenchyma. d, Number of total lung γδ T cells. e, Number of lung IL-17A⁺ γδ T cells. f, Number of eosinophils (CD11c⁻Siglec-F⁺) and neutrophils (Siglec-F⁻CD11b⁺Ly6G⁺) in lung parenchyma. Data pooled from two independent experiments (IgG2b n = 8; anti-IL-35 n = 10; anti-IL-27 n = 7 mice from the two experiments) in panels (d-f). Ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for statistical analysis in panels (d-f). Error bars denote mean ± s.d. P values are indicated in the figure.