PD-1 pathway regulates ILC2 metabolism and PD-1 agonist treatment ameliorates airway hyperreactivity

Abstract

Allergic asthma is a leading chronic disease associated with airway hyperreactivity (AHR). Type-2 innate lymphoid cells (ILC2s) are a potent source of T-helper 2 (Th2) cytokines that promote AHR and lung inflammation. As the programmed cell death protein-1 (PD-1) inhibitory axis regulates a variety of immune responses, here we investigate PD-1 function in pulmonary ILC2s during IL-33-induced airway inflammation. PD-1 limits the viability of ILC2s and downregulates their effector functions. Additionally, PD-1 deficiency shifts ILC2 metabolism toward glycolysis, glutaminolysis and methionine catabolism. PD-1 thus acts as a metabolic checkpoint in ILC2s, affecting cellular activation and proliferation. As the blockade of PD-1 exacerbates AHR, we also develop a human PD-1 agonist and show that it can ameliorate AHR and suppresses lung inflammation in a humanized mouse model. Together, these results highlight the importance of PD-1 agonistic treatment in allergic asthma and underscore its therapeutic potential.

Fig. 1. IL-33 induces PD-1 expression in pulmonary ILC2s.

a Pulmonary ILC2s were defined and/or sorted by a lack of lineage markers (CD3e, CD45R, Gr-1, CD11c, CD11b, Ter119, CD5, TCR-β, TCR-γδ, NK1.1, and FcɛRI) and expression of CD45, ST2, and CD127. b–j BALB/cByJ mice (WT) mice were challenged or not (naïve) with 0.5 μg of rm-IL-33 for 3 consecutive days. ILC2s from IL-33-challenged mice were defined as activated (aILC2s). b Representative histogram of the expression of PD-1, c PDL-1, and d PDL-2 in pulmonary ILC2s and corresponding quantification (right) presented as Mean Fluorescence Intensity (MFI); n = 5. e Gating strategy to define PD-L1 and PD-L2 live positive cells from f CD45⁺ and i CD45⁻ cells in IL-33 challenged mice. g, h Percentage of PD-L2⁺ and/or PD-L1⁺ cells from CD45⁺ and j CD45⁻ cells; n = 4. Data are representative of three independent experiments and are presented as means ± SEM (two-tailed Student’s t test, n.s. non-significant).

Fig. 2. PD-1 axis controls cytokine production by aILC2s and decreases their survival.

a–g; j–o ILC2s were sorted from WT and PD-1 KO mice after three intranasal challenges with 0.5 μg of rm-IL-33. Sorted cells were incubated with rm-IL-2 (10 ng mL⁻¹) and rm-IL-7 (10 ng mL⁻¹) for 24 h. a Volcano plot comparison (left) and heatmap representation (right) of total differentially regulated genes. Gene−specific analysis (GSA) algorithm was used to test for differential expression of genes (p-value < 0.05, n = 3). b Differentially expressed cytokine and cytokine receptor genes plotted as the normalized counts in WT compared to PD-1 KO aILC2s. Gray area indicates region of 1.5-fold-change cutoff or lower change in expression; n = 3. c Levels of IL-5, d IL-13, and e IL-9 quantified using LEGENDplex bead-based immunoassay; n = 4. f Levels of IL-5 and g IL-13 secreted by aILC2s in the presence of PD-L2 Fc or the isotype; n = 7. h, i ILC2s were sorted from naïve mice (nILC2s), cultured and stimulated in vitro with rm-IL-33 (20 ng mL⁻¹) for 48 h. PD-1 blocking antibody (αPD-1; 10 μg mL⁻¹) or isotype control were added in some conditions. h GATA-3 quantification presented as MFI in WT and PD-1 KO ILC2s; n = 4. i Levels of IL-5 quantified in ILC2 supernatants; n = 6. j Heatmap representation of significantly regulated pro- and anti-apoptotic genes. GSA algorithm was used to test for differential expression of genes (p-value < 0.05, n = 3). k Representative flow cytometry plots of AnnexinV DAPI staining and l corresponding quantification presented as the percentage of apoptotic and m dead ILC2s; n = 4. n Representative flow cytometry plot of Bcl-2 staining and o corresponding quantification presented as MFI in WT and PD-1 KO aILC2s; n = 4. Data are representative of three independent experiments and are presented as means ± SEM (two-tailed Student’s t test or one-way ANOVA).

Fig. 3. The lack of PD-1 signaling upregulates glycolysis in aILC2s.

a–i ILC2s were sorted from WT and PD-1 KO mice after three intranasal challenges with 0.5 μg of rm-IL-33. Sorted cells were incubated with rm-IL-2 (10 ng mL⁻¹) and rm-IL-7 (10 ng mL⁻¹) for 24 h. a Heat map representation of differentially regulated genes involved in the metabolism of aILC2 from WT and PD-1 KO mice. GSA algorithm was used to test for differential expression of genes (p-value < 0.05, n = 3). b Representative histogram (left) of 2-NBDG uptake and the corresponding quantification (right) presented as MFI in FACS-sorted WT and PD-1 KO aILC2s; n = 4. c–f Relative levels of metabolites in the glycolysis pathway analyzed using an LC-MS/MS system. g Measurement of the Oxygen consumption rate (OCR) under basal conditions and in response to indicated drugs. h Cell energy phenotype presented as OCR against extracellular acidification rate (ECAR). i Glycolytic capacity calculated as the difference between maximal and basal ECAR. Data are presented as means ± SEM (n = 3; two-tailed Student’s t test).

Fig. 4. Lack of PD-1 enhances methionine and glutamine catabolism in pulmonary aILC2s.

a–l ILC2s were sorted from WT and PD-1 KO mice after three intranasal challenges with 0.5 μg of rm-IL-33. Sorted cells were incubated with rm-IL-2 (10 ng mL⁻¹) and rm-IL-7 (10 ng mL⁻¹) for 24 h. a Volcano plot comparison of the relative levels of cellular metabolites, analyzed using an LC-MS/MS system. Color annotation were attributed to differentially abundant metabolites according to their classification (2-fold change cutoff, q-value ≥ 0.1; n = 3). b Methionine catabolism pathway showing measured metabolites in green. c Relative levels of measured methionine catabolism metabolites: SAM, d SAH, e Glutathione, and f Taurine in WT and PD-1 KO aILC2s. g Glutaminolysis pathway showing measured metabolites in green. h Relative levels of measured glutaminolysis metabolites: N-Methyl glutamate, i Glutamic acid, jN-Acetyl Glutamic acid, k Ornithine, and l GABA in WT and PD-1 KO aILC2s. Data are presented as means ± SEM (n = 3; two-tailed Student’s t test).

Fig. 5. PD-1 controls aILC2 proliferation through metabolic regulation.

a–f ILC2s were sorted from WT and PD-1 KO mice after three intranasal challenges with 0.5 μg of rm-IL-33. Sorted cells were incubated with rm-IL-2 (10 ng mL⁻¹) and rm-IL-7 (10 ng mL⁻¹) for 24 h. a Heat map representation of differentially regulated transcription factors in FACS-sorted aILC2s from WT and PD-1 KO mice lungs. GSA algorithm was used to test for differential expression of genes (p-value < 0.05, n = 3). b–f Sorted aILC2s were cultured in vitro for 24 h with 2-DG (0.5 mM) or CYL (20 mM) in some conditions. b Representative flow cytometry plots of GATA-3 and c intranuclear Ki67 with e, f the respective corresponding quantification presented as MFI in live WT and PD-1 KO aILC2s. g–k WT and PD-1 KO mice were intranasally challenged for 3 consecutive days with 0.5 µg rm-IL-33. On day 2 and 3, mice were intraperitoneally (i.p.) injected with 2-DG (500 mg kg⁻¹). On day 4, mice were euthanized. h Representative flow cytometry plots of ILC2s gated as CD127⁺ ST-2⁺ from CD45⁺ lineage⁻ lung cells and i the corresponding quantification presented as the absolute number of ILC2s. j Quantification of GATA-3 and k intranuclear Ki67 presented as MFI in WT and PD-1 KO aILC2s. Data are representative of at least two independent experiments and are presented as means ± SEM (n = 4; two-tailed Student’s t-test or one-way ANOVA). Mouse image provided with permission from Servier Medical Art.

Fig. 6. PD-1 expression on ILC2s ameliorates AHR and lung inflammation.

a–f WT and PD-1 KO mice were challenged intranasally for 3 consecutive days with 0.5 µg rm-IL-33. On day 4, AHR and lung inflammation were assessed. b Lung resistance and c dynamic compliance measured in restrained tracheostomized mechanically ventilated mice exposed to increasing concentrations of methacholine; n = 4. d Total number of eosinophils in BAL gated as CD45⁺ SiglecF⁺ CD11c⁻; n = 7. e Total number of pulmonary ILC2s gated as lineage⁻ CD45⁺ ST2⁺ CD127⁺; n = 7. f Total number of IL-5⁺ and IL-13⁺ ILC2s identified by intracellular staining; n = 4. g–kRag2^−/− mice received intraperitoneal injection (i.p.) of anti-PD-1 blocking antibody (500 μg) or isotype control at day 1. Then mice were challenged intranasally on day 2–4 with 0.5 µg rm-IL-33. On day 5, AHR and lung inflammation were assessed. h Lung resistance and i dynamic compliance in response to increasing concentrations of methacholine; n = 4. j Total number of eosinophils in BAL and k total number of pulmonary ILC2s assessed by flow cytometry; n = 7. Data are representative of at least two independent experiments and are presented as means ± SEM (two-tailed Student’s t test or one-way ANOVA). Mouse image provided with permission from Servier Medical Art.

Fig. 7. PD-1 is inducible on human ILC2s and PD-1 agonist suppresses ILC2-mediated AHR.

a Human ILC2s were FACS-sorted from PBMCs as CD45⁺ lineage⁻ CD127⁺ CRTH2⁺. Purity was always ≥95%. b–d Cells were cultured (5 × 10⁴ per mL) in the presence of rh-IL-2 (10 ng mL⁻¹) and rh-IL-7 (20 ng mL⁻¹), with or without rh-IL-33 (20 ng mL⁻¹), PD-1 agonist (25 μg mL⁻¹), or corresponding isotype for 48 h. b Representative histogram of the expression of PD-1 in human ILC2s and corresponding quantification (right) presented as MFI; n = 4. c Levels of IL-5 and d IL-13 were quantified in culture supernatants from isotype- and PD-1 agonist-treated human ILC2s; n = 7. e–m In vitro cultured human ILC2s were adoptively transferred (intravenously) into Rag2^−/− Il2rg^−/− mice. At day 1, mice received an intraperitoneal injection of PD-1 agonist or control isotype (500μg). Then, mice were intranasally challenged with rh-IL-33 or PBS and injected with PD-1 agonist or isotype (250 μg) on days 2–4. Measurement of lung function and inflammation followed on day 5. f Lung resistance and g dynamic compliance measured in restrained tracheostomized mechanically ventilated mice exposed to increasing concentrations of methacholine; n = 4. h, i Total number of human ILC2s in lungs gated as CRTH2⁺ CD127⁺; n = 5. j Total number of eosinophils in BAL gated as CD45⁺ SiglecF⁺ CD11c⁻; n = 5. k Lung histology (Scale bars, 50 mm); representative of two independent experiments. l Quantification of airway epithelium thickness and m infiltrating cells; n = 6. Data are representative of at least two independent experiments and are presented as means ± SEM (two-tailed Student’s t test or one-way ANOVA). Mouse and human images provided with permission from Servier Medical Art.