Type III interferons induce pyroptosis in gut epithelial cells and impair mucosal repair

Abstract

Tissue damage and repair are hallmarks of inflammation. Despite a wealth of information on the mechanisms that govern tissue damage, mechanistic insight into how inflammation affects repair is lacking. Here, we investigated how interferons influence tissue repair after damage to the intestinal mucosa. We found that type III, not type I or type II, interferons delay epithelial cell regeneration by inducing the upregulation of Z-DNA-binding protein 1 (ZBP1). Z-nucleic acids formed following intestinal damage are sensed by ZBP1, leading to caspase-8 activation and the cleavage of gasdermin C (GSDMC). Cleaved GSDMC drives epithelial cell death by pyroptosis and delays repair of the large or small intestine after colitis or irradiation, respectively. The type III interferon/ZBP1/caspase-8/GSDMC axis is also active in patients with inflammatory bowel disease (IBD). Our findings highlight the capacity of type III interferons to delay gut repair, which has implications for IBD patients or individuals exposed to radiation therapies.

Keywords: Z-nucleic acid; colitis; damage; gasdermin; inflammation; inflammatory bowel disease; interferons; intestinal epithelial cell; intestinal stem cell; irradiation; pattern recognition receptors; pyroptosis; repair.

FIGURE 1.. IFN-λ acts on IECs delaying tissue repair during colitis and irradiation damage.

(A-I) Colitis was induced as illustrated (A). Colon length (B) and histology images of (C) Ifnlr1^fl/fl or Vil^CRE Ifnlr1^fl/fl mice over time. WT (D, E) or WT and Vil^CRE Ifnlr1^fl/fl mice (F) were injected i.p. with rIFN-λ or saline (Veh) upon DSS withdrawal. Colon length (D, F), histological score, and histology images (E) on Day-10. WT or Ifnl2/3 dKO mice (G-I) were treated as in (D-F). Relative body weight (G) colon length (H) and histological score (left) with representative histology images (right) (I). (J-P) Mice of indicated genotypes received sublethal 11 Gy (K-N) or a partially lethal 14 Gy (O, P) partial body irradiation as illustrated (J), and were administered, or not, rIFN-λ or Veh. Intact crypts were counted at indicated times (L) or after 96h (K, M, N left: quantification; right: representative images) following irradiation in mice of the indicated genotypes. (O, P) Mice were irradiated and treated with rIFN-λ, or Veh and followed over time. Survival (O) and weight loss (P) curves of indicated mice and treatments after 14 Gy irradiation. Statistics: (*): comparison “WT + Veh” vs “Vil^CREIfnlr1^fl/fl + Veh” and (§): comparison “WT + Veh” vs “WT + rIFN-λ”. (B, G, P) Mean with SEM. (D-F, H, I, K-N) Box plots with median, range and interquartile range. Each dot represents one mouse. (O) Kaplan-Meier survival curve. N = 3–4 mice per group. Statistics: (B, F, G, L, P) Two-way ANOVA with Šídák (B, F, G), Dunnett (L, P), or Tukey (L, P) post hoc test. (D, E) Mann-Whitney test. (H, I, K, M, N) One-way ANOVA with Tukey (H, I) or Dunnet (K, M, N) post hoc test. (O) Log-rank (Mantel-Cox) test. ns= not significant (p > 0.05); * or § p < 0.05; ** or §§ p < 0.01; *** or §§§ p < 0.001; **** or §§§§ p < 0.0001. Scale bars: 500 μm (C, I) and 200 μm (E, K, N).

FIGURE 2.. IFN-λ dampens regenerative and proliferative transcriptional programs in IECs.

Vil^CRE Ifnlr1^fl/fl or WT mice received sublethal partial body irradiation. (A-F) Targeted transcriptomics on small intestinal crypts after 96h. N = 3 mice per group. Dot plots of Gene Ontology (GO) terms enriched in differentially expressed genes (DEGs) overexpressed in WT (A) or Vil^CRE Ifnlr1^fl/fl (B), and Volcano plot depicting DEGs (C) (red: P< 0.005; blue: P< 0.005; green: P> 0.005; gray: P> 0.005; dotted lines are set at 2). P.adj: adjusted p-value. GSEA enrichment plots (D-F) of indicated gene sets. padj: adjusted p-value, NES: Normalized enrichment score. (G-L) Single-cell RNA sequencing on small intestinal crypt cells (IECs) 72h after irradiation. UMAP of IECs after removal of mature enterocytes, colored by cell type (G) or genotype (I). ISC: Intestinal stem cells, Imm.: Immature, TA: Transit amplifying. Heatmap of top 5 marker genes (H). Quantification (J) of total cells (top) and relative enrichment as indicated by Pearson residuals from Chi² analysis (bottom). Dot colors/size correspond to Pearson residuals. Violin plots (K) of interferon-stimulated gene (ISG) module score. Statistics: Kruskal-Wallis test. Dot plots (L) of Gsdmc2–4 expression.

FIGURE 3.. IFN-λ-dependent activation of GSDMC delays intestinal repair upon irradiation- or colitis-driven damage.

(A, B) Overexpression of full length human GSDMC (hGSDMC FL), the N-terminal domain of human GSDMC (hGSDMC NT), or empty vector control (EV) in HEK293T cells. Cell viability and LDH release 24h after transfection (A). Sytox Orange uptake (B). (C, D) Small intestinal crypt cells were isolated from Vil^CRE Ifnlr1^fl/fl or WT mice that received sublethal partial body irradiation. Immunoblot of indicated proteins (C) with densitometry quantification (D). Red statistics: Comparison to non-irradiated control of the same genotype. (E-K) Mice were administered DSS and monitored over time. (E-G) Colon IECs were isolated at indicated days (E) or on Day-10 (G). Representative immunoblot (E) and quantification of band intensities (F). Red statistics: Comparison to Day-0 control of the same genotype. (H-M) Relative body weight (H, K), colon length (I, K), histology images (J, left) and histological score (J, right; K) at Day-10 of mice of indicated genotypes. (L, M) DSS-treated mice were injected with Z-IETD or vehicle (Veh) i.v. from Day-7. Experimental setup (L) and relative body weight, colon length and histological score on Day-12 (M). (A, B) Mean with SEM of 3 independent experiments. (C, E, G) Each lane represents one mouse. Representative data of 2 (E) or 3 (C, G) independent experiments. (H-K, M) Box plots with median, range and interquartile range. Each dot represents one mouse. Statistics: (A, B) One way ANOVA with Dunnett correction for multiple comparisons. (B) Statistics were performed on area under curve (AUC). (D, F) Two-way ANOVA with Šídák (D) or Tukey (F) correction for multiple comparisons. (H-K, M) Unpaired t-test. ns= not significant (p > 0.05); *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

FIGURE 4.. GSDMCs remodel the cellular landscape of the colon during colitis.

Vil^CREGsdmc1–4^fl/fl or Gsdmc1–4^fl/fl littermates were administered DSS for 7 days or left untreated. On Day-10 epithelial and lamina propria cells were isolated and subjected to scRNA-seq. (A, B) UMAP of all cells (A) and across genotypes and treatment (B). (C) Frequency of epithelial, immune or stromal cells. (D, E) UMAP after re-clustering of all epithelial cells (D) and across genotype and treatment (E). (F, G) Frequency of epithelial subpopulations among all epithelial cells. (H) Gene set variation analysis (GSVA) of fetal gene score as described in. n.d.= not determined (insufficient cell number). (I) Dot plots of indicated genes across epithelial subpopulations. Dot color: Level of expression. Dot size: Percentage of expression within the population. (C, F, G, H) Mean with SEM. Statistics: ns= not significant (p > 0.05); *p < 0.05; **p < 0.01, *** P<0.001, **** p<0.0001. Two-way ANOVA with Fisher’s LSD test. Each dot corresponds to one mouse.

Figure 5.. The IFN-λ/ZBP1/Casp-8/GSDMC axis mediates IEC pyroptosis in response to damage.

(A) ALI organoid cultures were treated with rIFN-λ, rIFN-β, or rIFN-γ for 3 days during indicated culture phases and cell death (Zombie, left), caspase 8 activation (CL-Casp8, middle), and ZBP1 (right) expression were assessed by immunofluorescence. Frequency of positive cells is depicted. (B-J) ALI organoids were treated with rIFN-λ during homeostasis (C, F), damage (B, D, G, I, J), or recovery (B, E, H) cell death (Zombie, left), caspase 8 activation (CL-Casp8, middle), and Ki67 (right) expression were assessed by immunofluorescence and frequency of positive cells is depicted. Representative immunofluorescence images (B). Organoids were treated with Z-VAD-FMK (Z-VAD, I), or Z-IETD-FMK (Z-IETD, J) or no inhibitor (NT) during rIFN-λ treatment. (K) Mouse small intestinal spheroids were treated with rIFN-λ in the presence of propidium iodide (PI) and imaged at indicated time points. Left: Representative images. Right: Quantification. (A, C-K) Mean with SEM of 3 experiments. Statistics: (A, C-K) Two-way ANOVA with Šídák (A, C-J) or Tukey (K) post hoc tests. ns= not significant (p > 0.05); *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Scale bar: 20 μm (B), 50 μm (K).

Figure 6.. Z-nucleic acids control the activation of a ZBP1/Casp-8/GSDMC complex induced by IFN-λ.

(A) ALI organoids were treated with rIFN-λ for 1 day and frequency of Z-NA⁺ cells was assessed by immunofluorescence. CBL0137 was used as positive control. (B, C) Quantification of Z-NA⁺ cell frequency (B) and representative images of colon sections (C) on Day-10 of WT mice administered indicated doses of DSS for 7 days or left untreated. (D, E) WT mice received sublethal partial body irradiation or were left untreated. Quantification of Z-NA⁺ cells 48h or 72h after irradiation (D) and representative images (E) in small intestine sections. (F-H) WT and Zbp1^{ΔZα2/ΔZα2} mice were administered DSS for 7 days. Relative body weight over time (F), Colon length (G) and histology score (H) on Day-12. (I, K) ALI organoids of the indicated genetic backgrounds were treated with rIFN-λ for 3 days during the recovery phase. Frequency of Zombie⁺ cells (I) and CL-CASP8⁺ cells (I, K). (J) WT mice were administered (+), or not (−), DSS for 7 days and on Day-10 RIPK1 was immunoprecipitated (IP) from colon IEC protein extracts and the indicated proteins were analyzed. (A, F, I, K) Mean and SEM of 2–4 independent experiments. (B, D, G, H) Box plots with median, range and interquartile range. Each dot represents one mouse. (C, E) Scale bars indicate 100 μm. Statistics: (B, D) One-way ANOVA with Dunnett post hoc test. (G, H) Unpaired t-test. (F, I, K) Two-way ANOVA with Šídák post hoc test. ns= not significant (p > 0.05); *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 7.. The IFN-λ/ZBP1/Casp-8/GSDMC axis is active in IBD patients.

(A-I) Bulk RNA sequencing (A-D), cytokine levels (H, G), and protein expression by immunoblot (E, F, I) in cell lysates of human colon biopsies of healthy controls (HC), and Crohn’s disease (CD), or ulcerative colitis (UC) patients. Principal component analysis (A), and differential gene expression analysis between IBD patients (CD and UC) and HC transcriptome (B-D) was performed. Bubble plot depicting results of Gene set enrichment analysis (GSEA) of Top 20 pathways by absolute normalized enrichment score (NES) (B); positive NES values: enrichment in IBD patients; negative NES values: enrichment in HC. Enrichment plot of indicated signatures in IBD patients (CD and UC) compared to HC (C) and Volcano plot depicting DEGs between HC and IBD patients (D). Red: P< 0.05; blue: P< 0.05; green: P> 0.05; gray: P> 0.05; dotted lines are set at 1. P.adj: adjusted p-value. Representative immunoblots (E), densitometry quantification normalized to HC (F), and multiplex analysis of cytokines (G). Heatmap of Pearson correlation coefficient (Pearson r) (H) or dot plots (I) for indicated cytokine and relative ZBP1, CASP8 p43/41, and GSMDCp20 protein expression in CD patients. (F, G) Each dot represents one sample. (F, G) Box plots with median, range and interquartile range. Statistics: (F, G) One-way ANOVA. (H, I) Pearson correlation coefficients (R) (H), or R² with two-tailed p value (I). ns= not significant (p > 0.05); *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.