IL-33 controls IL-22-dependent antibacterial defense by modulating the microbiota

Abstract

IL-22 plays a critical role in defending against mucosal infections, but how IL-22 production is regulated is incompletely understood. Here, we show that mice lacking IL-33 or its receptor ST2 (IL-1RL1) were more resistant to Streptococcus pneumoniae lung infection than wild-type animals and that single-nucleotide polymorphisms in IL33 and IL1RL1 were associated with pneumococcal pneumonia in humans. The effect of IL-33 on S. pneumoniae infection was mediated by negative regulation of IL-22 production in innate lymphoid cells (ILCs) but independent of ILC2s as well as IL-4 and IL-13 signaling. Moreover, IL-33's influence on IL-22-dependent antibacterial defense was dependent on housing conditions of the mice and mediated by IL-33's modulatory effect on the gut microbiota. Collectively, we provide insight into the bidirectional crosstalk between the innate immune system and the microbiota. We conclude that both genetic and environmental factors influence the gut microbiota, thereby impacting the efficacy of antibacterial immune defense and susceptibility to pneumonia.

Keywords: IL-22; IL-33; Streptococcus pneumoniae; microbiota; pneumonia.

Fig. 1.

IL-33 deficiency enhances the resistance of mice to pneumococcal pneumonia. WT (n = 11) and Il33^−/− (n = 12) mice were intranasally infected with 5 × 10⁶ CFU/mouse of S. pneumoniae and bacterial loads in BALF (A) and blood (B) were determined 48 h postinfection, and temperature was measured at the indicated time points (C). (D–F) WT (n = 12) and Il1rl1^−/− (n = 12) were infected and bacterial loads in BALF (D) and blood (E) were determined 48 h postinfection, and temperature was measured (F). (G and H) Bone marrow chimeras (n = 12 per group) were infected and bacterial loads in BALF were determined 48 h postinfection (G), and temperature was measured (H). (A, B, D, E, and G) Data are shown as individual points. Lines represent the median and the dashed line the lower detection limit. The Wilcoxon rank-sum test or Kruskal–Wallis followed by Dunn’s post hoc test was used if more than two groups were compared. (C, F, and G) Data are shown as mean ± SD, Kruskal–Wallis followed by Dunn’s post hoc test; *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 2.

ILC2 as well as IL-4 and IL-13 signaling do not regulate early antibacterial defense during pneumococcal pneumonia. (A) WT and Il33^−/− mice were infected with S. pneumoniae and killed after 18 h, and pulmonary cells were analyzed by scRNAseq (pulmonary cells were pooled from n = 4 per group). Normalized expression levels of Il1rl1 were plotted in an UMAP embedding. (B) Gating strategy to quantify ILC2s by FACS. (C) Numbers and frequencies of pulmonary Lin^- CD90⁺ CD127⁺ GATA3⁺ ILC2s from the lungs of WT (n = 12) and Il33^−/− mice (n = 13) 18 h after infection. Bars represent mean + SD. (D and E) The lungs of 3 to 4 mice per group were digested at 12 h postinfection and sorted for ILC2s, and RNA was isolated and analyzed using bulk RNA sequencing. (D) Volcano plot displaying differentially expressed genes (DEGs) in ILC2s of WT and Il33^−/− animals. Significant DEGs with a log2FC threshold of ± 2 and an adjusted P value of < 0.05 are indicated in red. (E) GSEA analysis utilizing HALLMARK collection, normalized enrichment score (NES), and adjusted P-values (padj) are indicated. (F and G) Nmur1^iCre-eGFPId2^fl/fl mice (ILC2^cKO) (n = 9) and littermate controls (n = 10) were infected, and bacterial loads in BALF and blood were assessed 48 h post infection. (H and I) Il4ra^−/− (n = 7) and WT mice (n = 8) were infected, and bacterial loads in BALF (H) and blood (I) were assessed at 48 h postinfection. Data are shown as individual points. Lines represent the median and the dashed line the lower detection limit.

Fig. 3.

The protective effect of IL-33 deficiency in pneumococcal infection is mediated by enhanced production of IL-22. (A) WT (n = 24) and Il33^−/− (n = 24) mice were infected with S. pneumoniae and killed after 18 h, and levels of IL-22 in BALF were measured by the ELISA. (B) Representative gating strategy for analyzing IL-22⁺ lymphoid cells by FACS. (C and D) Frequency and absolute numbers of IL-22-producing lymphoid cells in the lungs of WT (n = 10) and Il33^−/− mice (n = 11) 18 h after infection. (E and F) WT (n = 16), Il33^−/− (n = 15), Il22^−/− (n = 15), and Il33^−/−Il22^−/− mice (n = 9) were infected for 48 h, and bacterial loads were determined in BALF (E) and blood (F). Data are shown as individual points. Bars represent mean + SD, lines represent the median, and the dashed line the lower detection limit. The Wilcoxon rank-sum test or Kruskal–Wallis test followed by Dunn’s post hoc test was used if more than two groups were compared, ns = P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Fig. 4.

Il33^−/− mice from different vivaria vary in their susceptibility to S. pneumoniae infection. WT and Il33^−/− mice from seven different animal vivaria were infected and killed after 48 h, and bacterial loads in BALF were assessed (n = 5 to 11 per group). Data are shown as individual points. Lines represent the median and the dashed line the lower detection limit. Wilcoxon rank-sum test, ns = P > 0.05, *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 5.

IL-33’s influence on antibacterial defense depends on its modulatory effect on the microbiota. (A and B) WT (n = 9) and Il33^−/− mice (n = 9) were microbiota depleted by oral treatment with antibiotics, subsequently infected with 1 × 10⁶ CFU/mouse, and killed after 48 h. Bacterial loads (A) and IL-22 levels (B) were assessed in BALF. (C and D) WT and Il33^−/− mice were cohoused for 4 wk and infected for 48 h. Bacterial loads (C) and IL-22 levels (D) were assessed in BALF. (A–D) Data are shown as individual points, lines represent the median, the dashed line the lower detection limit, and bars represent mean + SD. Wilcoxon rank-sum test, ns = P > 0.05. (E) Alpha diversity of microbiota samples from Il33^−/− animals with resistant (Il33^−/−R) and susceptible phenotype (Il33^−/−S) and corresponding controls (WT^ctrlR and WT^ctrlS) is shown. (F) Heatmap of shotgun-sequenced bacterial microbiota derived from resistant (n = 15) and susceptible Il33^−/− (n = 13), their respective controls (for WT^ctrlR n = 15 and WT^ctrlS n = 12), and cohoused WT (n = 11) and Il33^−/− mice (n = 10). Cliff’s delta was applied to visualize most differential effect sizes between Il33^−/−R and WT^ctrlR. (G) WT, Il33^−/−, and Il33^−/−Il22^−/− mice were treated orally with an antibiotic cocktail to deplete their own microbiota. Afterward, mice were transplanted with fecal samples derived from WT^ctrlR or Il33^−/−R animals. After a reconstitution time of 18 d, mice were intranasally infected with S. pneumoniae. Data are shown as individual points (n = 8 for WT^ctrlR > WT, n = 8 for Il33^−/−R > WT, n = 12 for WT^ctrlR > Il33^−/−, n = 12 for Il33^−/−R > Il33^−/− and n = 6 for Il33^−/−R > Il33^−/−Il22^−/−). Kruskal–Wallis test followed by Dunn’s post hoc test; ns = P > 0.05 and *P < 0.05.