Fig 2: ITLN1 affects the inner mucus layer thickness in a microbiota-dependent manner. (A) Representative fluorescence images obtained after methacarn fixation and combined immunofluorescence (IF) and FISH of the distal colon from TgVil1-Itln1 mice and their respective wild-type littermates (WT) under SPF conditions. For each genotype, we present representative grayscale images depicting the nuclei (DAPI), the intestinal epithelial surface (?-actin IF = ?-actin), the intestinal microbiota (Eubacteria probes FISH = Eubacteria), and the mucus layer (Mucin-2 IF = MUC2) and merged image with pseudo coloring (DAPI = blue, ?-actin = gray, Eubacteria = magenta, MUC2 = green). The inner mucus layer was characterized by a well-organized stratified MUC2 lamellar appearance between the epithelium and the bacterial biomass in the merged image (between white arrowheads). Scale bars indicate 20 µm. (B) Inner mucus thickness was measured between white arrowheads in A as described by Earle et al. (2015). 288 independent measurements 50 µm apart were obtained from four different mice per genotype (n = 4). Mean and SD were determined from the average measurements for each mouse. Each dot represents the mean value of measurements per mouse. Error bars represent the SD. (C) Representative fluorescence images were obtained after methacarn fixation and combined IF and FISH of distal colon from Itln1-/- mice and their respective wild-type littermates (wt) under SPF conditions. For each genotype, we present representative fluorescence images depicting the nuclei (DAPI), the intestinal epithelial surface (?-actin IF = ?-actin), intestinal microbiota (Eubacteria probes FISH = Eubacteria), and the mucus layer (Mucin-2 IF = MUC2) and merged images with pseudo coloring (DAPI = blue, ?-actin = gray, Eubacteria = magenta, MUC2 = green). The inner mucus layer was characterized by a well-organized stratified MUC2 lamellar appearance between the epithelium and the bacterial biomass in the merged image (between white arrowheads). Scale bars indicate 20 µm. (D) Inner mucus thickness was measured between white arrowheads in C as described by Earle et al. (2015). 276–284 independent measurements 50 µm apart were obtained from four different mice per genotype (n = 4). Mean and SD were determined from the average measurements for each mouse. Each dot represents the mean value of measurements per mouse. Error bars represent the SD. (E) Representative fluorescence images obtained after methacarn fixation and combined IF and FISH of distal colon from GF wild-type, Itln1-/-, and TgVil1-Itln1 mice. For each genotype, we present representative fluorescence images depicting the nuclei (DAPI), the intestinal epithelial surface (?-actin IF = ?-actin), intestinal microbiota (Eubacteria probes FISH = Eubacteria), and the mucus layer (Mucin-2 IF = MUC2) and merged images with pseudo coloring (DAPI = blue, ?-actin = gray, Eubacteria = magenta, MUC2 = green). The inner mucus layer was characterized by a well-organized stratified MUC2 lamellar appearance above the epithelium (between white arrowheads). Scale bars indicate 20 µm. (F) Inner mucus thickness was measured between white arrowheads in E as described by Earle et al. (2015). 244–276 independent measurements 50 µm apart were obtained from four different mice per genotype (n = 4). Mean and SD were determined from the average measurements for each mouse. Each dot represents the mean value of measurements per mouse. Error bars represent the SD. (G) Representative fluorescence images obtained after methacarn fixation and combined IF and FISH of distal colon from GF wild-type, Itln1-/-, and TgVil1-Itln1 mice monocolonized with A. muciniphila. For each genotype, we present representative fluorescence images depicting the nuclei (DAPI), the intestinal epithelial surface (?-actin IF = ?-actin), A. muciniphila (A. muciniphila probe FISH = A. muciniphila), and the mucus layer (Mucin-2 IF = MUC2) and merged images with pseudo coloring (DAPI = blue, ?-actin = gray, A. muciniphila = red, mucus layer = green). The inner mucus layer was characterized by a well-organized stratified MUC2 lamellar appearance between the epithelium and the bacterial biomass in the merged image (between white arrowheads). Scale bars indicate 20 µm. (H) Inner mucus thickness was measured between white arrowheads in G as described by Earle et al. (2015). 244–276 independent measurements 50 µm apart were obtained from four different mice per genotype (n = 4). Mean and SD were determined from the average measurements for each mouse. Each dot represents the mean value of measurements per mouse. Error bars represent the SD. WT = wild-type littermate from TgVil1-Itln1 colony. wt = wild-type littermate from Itln1-/- colony. P values were calculated by unpaired T test (B and D) and one-way ANOVA corrected for multiple comparisons with Dunnet (F and H). *P < 0.05.

Fig 3: Baseline characterization of the microbiota of Itln1-/- and TgVil-Itln1 mice. (A) Stacked bars represent the aggregated total community composition at the phyla level in the large intestinal lumen of Itln1-/- mice and their wild-type littermates (wt; n = 14) and TgVil1-Itln1 mice and their wild-type littermates (WT; n = 13 or 14). (B) Stacked bars represent the aggregated total community composition at the phyla level in the colonic mucosa of Itln1-/- mice and their wild-type littermates (wt; n = 14) and TgVil1-Itln1 mice compared to wild-type littermates (WT; n = 12). (C) Stacked bars represent the aggregated total community composition at the family level in the large intestinal lumen of Itln1-/- mice and their wild-type littermates (wt; n = 14) and TgVil1-Itln1 mice compared to wild-type littermates (WT; n = 13 or 14). (D) Stacked bars represent the aggregated total community composition at the family level in the colonic mucosa of Itln1-/- mice and their wild-type littermates (wt; n = 14) and TgVil1-Itln1 mice compared to wild-type littermates (WT; n = 12). (E) Chao1 index for Itln1-/- mice compared to wild-type littermates (wt) in the large intestinal lumen (n = 14). (F) Chao1 index for Itln1-/- mice compared to wild-type littermates (wt) in the colonic mucosa (n = 14). (G) Shannon index for Itln1-/- mice compared to wild-type littermates (wt) in the large intestinal lumen (n = 14). (H) Shannon index for Itln1-/- mice compared to wild-type littermates (wt) in the colonic mucosa (n = 14). (I) Beta diversity of the microbiota composition of the large intestinal lumen by Bray–Curtis dissimilarity in Itln1-/- mice compared to wild-type littermates (wt; n = 14). Symbols represent individual mice. (J) Beta diversity of the microbiota composition of the colonic mucosa by Bray–Curtis dissimilarity in Itln1-/- mice compared to wild-type littermates (wt; n = 14). Symbols represent individual mice. (K) Chao1 index for TgVil1-Itln1 mice compared to wild-type littermates (WT) in the large intestinal lumen (n = 13 or 14). (L) Chao1 index for TgVil1-Itln1 mice compared to wild-type littermates (WT) in the colonic mucosa (n = 12). (M) Shannon index for TgVil1-Itln1 mice compared to wild-type littermates (WT) in the large intestinal lumen (n = 13 or 14). (N) Shannon index for TgVil1-Itln1 mice compared to wild-type littermates (WT) in the colonic mucosa (n = 12). (O) Beta diversity of the microbiota composition of the large intestinal lumen by Bray–Curtis dissimilarity in TgVil1-Itln1 mice compared to wild-type littermates (WT; n = 13 or 14). Symbols represent individual mice. (P) Beta diversity of the microbiota composition of the colonic mucosa by Bray–Curtis dissimilarity in TgVil1-Itln1 mice compared to wild-type littermates (WT; n = 12). Symbols represent individual mice. (Q) Bacteria gating strategy for ITLN1-seq. Stool bacteria were identified as SYBRgreen high (SYBRgreenhi) particles in SPF mice that were not present in GF mice. In the SYBRgreenhi fraction, we quantified the fraction of ITLN1(+) bacteria identified with a PE-conjugated ITLN1 antibody, as depicted in Fig. 2 B. P values for the alpha diversity (Chao1 and Shannon index) were calculated using a linear mixed model by regressing the alpha diversity value against the “genotype” with gender as a fixed effect and litter as a random effect (ns: P > 0.05). CAP1 and CAP2 are the first two axes from the constrained analysis of principal coordinates with the respective amount of variation in Bray–Curtis dissimilarity explained between brackets. For Bray–Curtis dissimilarity, P obtained by the anova.cca test with respect to the genotype with 10,000 permutations (*P < 0.05). WT = wild-type littermate from TgVil1-Itln1 colony. wt = wild-type littermate from Itln1-/- colony.

Fig 4: Baseline characterization of LP leukocytes in the Itln1-/- and TgVil-Itln1 mice. (A) Quantification of leukocytes in colonic LP of Itln1-/- mice compared to wt littermates (n = 11–19). Data were compiled from three independent experiments for the left and right panels and two independent experiments for the middle panel. Symbols represent individual mice. (B) Quantification of leukocytes in colonic LP of TgVil1-Itln1 mice compared to WT littermates (n = 6–11). Data were compiled from two independent experiments for the left and middle panels. Symbols represent individual mice. (C) Gating strategy for LP neutrophils and eosinophils in A and B. (D) Gating strategy for LP CD4+ T cells, CD8+ T cells, and B cells in A and B. (E) Gating strategy for innate lymphocytes (ILC) 1, 2, and 3 in A and B. Lin = CD3+, CD5+,CD19+, and LY6G+. (F) Gating strategy for natural killer cells (NK) in A and B. Lin = CD3+, CD5+, CD19+, and LY6G+. (G) Gating strategies for LP dendritic cells (DC) in A and B. Lin = CD3+, NK1.1+, CD19+, Ly6G+, and SiglecF+. (H) Gating strategies for LP macrophages in A and B. P1, P2, P3, and P5 macrophages correspond to macrophage subpopulations as described in Tamoutounour et al. (2012). Lin = CD3+, NK1.1+, CD19+, Ly6G+, and SiglecF+.

Fig 5: ITLN1 is increased in response to ER stress in intestinal epithelial cells. (A) ITLN1 expression level from bulk RNA-seq performed in colonic biopsies of healthy controls, patients with treatment-naive UC with colitis, or in remission from Taman et al. (2017) (n = 14–16). Boxes extend from the 25th to 75th percentile and whiskers from minimum to maximum value, and the line in the middle is the median. (B) Correlation between ITLN1 and UPR hallmark gene expression by bulk RNA-seq in Taman et al. (2017) (n = 44). Symbols represent individual human subjects. (C) Correlation between UPR hallmark genes and ITLN1 expression in goblet cells by scRNA-seq from Parikh et al. (2019) (n = 1,198 cells). Symbols represent individual epithelial cells. (D) ITLN staining in crypts from patients with CD that have GRP78 negative (-) or positive (+) staining of their crypts (n = 9–13). Left panel: Representative pictures from small intestine biopsies obtained from GRP78(-) patients and GRP78(+) patients (Deuring et al., 2014). Scale bars indicate 20 µm. Black arrows point to Paneth cells in the crypt. Right panel: Bars represent arithmetic means. Symbols represent individual human subjects. (E) Quantification of Itln1 transcripts by qPCR in mouse small intestinal organoids in the presence or absence of tunicamycin (TM; n = 6). Symbols represent individual biological replicates. Bars represent arithmetic means. Data were compiled from two independent experiments. (F) Quantification of Itln1 transcripts by qPCR in mouse small intestinal organoids after TM treatment alone or in the presence of 4µ8c, GSK2606414, or PF-429242 (n = 6–9). Symbols represent individual biological replicates. Bars represent arithmetic means. Data were compiled from two to three independent experiments. (G) Volcano plot showing log2-transformed fold-change of gene expression in crypts obtained by laser capture microscopy from GF Xbp1?IEC mice compared with crypts obtained by laser capture microscopy from GF Xbp1fl/fl controls (n = 3–4). Symbols represent individual genes. (H) Itln1 transcripts by qPCR in intestinal organoids from wild-type and Atf6tg mice at baseline and after treatment with TM (n = 8). Symbols represent individual biological replicates. P values were calculated by Wald-test and corrected for multiple testing by the method of Benjamini and Hochberg (A); generalized linear model (B); generalized negative binomial linear model (C); unpaired T test (D, E, and H); one-way ANOVA corrected for multiple comparisons with Dunnet (F); and two-stage step-up method of Benjamini, Krieger, and Yekutieli to control the FDR (G). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.