Cathelicidin regulates goblet cell mucus secretion and mucus-associated proteins in Citrobacter rodentium-induced colitis

Abstract

Colonic goblet cells play a crucial role in mucosal defense by secreting Muc2 mucin and other proteins that entrap and expel enteropathogens. However, the role of innate effectors in the gut like cathelicidin peptides in regulating the mucus barrier during infections remains unclear. In this study, we used cathelicidin-deficient (Camp^-/-) littermates, colonoids, and human LS174T goblet-like cells to investigate how cathelicidin modulates goblet cell function and mucosal defense against attaching/effacing enteropathogen Citrobacter rodentium. We showed that increased fecal shedding and epithelial colonization by C. rodentium in Camp^-/- littermates was accompanied by impaired mucus secretion and higher retention of mucin granules and trefoil factor 3 (Tff3) in bloated colonic goblet cells. Reduction in mucus secretion by goblet cells was accompanied by reduced reactive oxygen species (ROS) production during C. rodentium infection in Camp^-/- as compared to Camp^+/+ littermate controls. In LS174T goblet-like cells, human cathelicidin LL-37 stimulated the secretion of TFF3 and resistin-like molecule β (RELMβ) in a ROS-dependent manner. These findings reveal that cathelicidin regulates goblet cell mucus and mucus-associated protein secretion through a ROS-mediated mechanism critical for bacterial clearance and maintenance of gut homeostasis.

Keywords: Colonic epithelium; cathelicidin; citrobacter rodentium; goblet cells; mucus; reactive oxygen species.

Figure 1.

Camp^-/- mice are more susceptible to C. rodentium colonization. Camp^+/+ and Camp^-/- littermates were infected orally with C. rodentium (5 ×10⁸ CFU) or control PBS. (A) fecal shedding of C. rodentium was assessed by microbiological culture at the indicated days post-infection (n = 6–9 mice/group). (B-C) dissemination of bioluminescent C. rodentium with in vivo imaging expressed as arbitrary units (AU) of luminescence in (B) the whole body and (C) ex vivo excised colons (n = 5–7 mice/group; 7- dpi). (D-E) colonization of C. rodentium on the colonic epithelium assessed by (D) immunostaining with an anti-LPS antibody (green) (scale bar: 25 µm) and (E) RNAscope staining for C. rodentium virulence factor espB mRNA (scale bar 50 µm). DAPI (blue) was used as a nuclear marker. Representative images of 3–4 mice/group for 5 fields of view are shown. (F) bacterial adherence and pedestal formation on epithelial cells showed by transmission electron microscopy in distal colonic sections. Scale bar 500 nm. Representative images from 3 mice, 4–5 fields per sample. Data are shown as means ± SEM. *p < 0.05 (two-tailed Student’s t-test or one-way ANOVA post hoc Tukey’s test for multiple comparisons) was considered significant.

Figure 2.

Camp^+/+ and Camp^-/- display similar colitis during C. rodentium infection. Camp^+/+ and Camp^-/- littermates were infected orally with C. rodentium (5 ×10⁸ CFU) or PBS. (A) Colon length was determined at 7- and 14- dpi with C. rodentium (n = 5–6/group). (B) Microphotographs of distal colons at the indicated days post-infection and histopathological scores (hematoxylin & eosin) (n = 6–8 mice/group). Scale bar 50 µm. (C) Flow cytometric analysis of colonic lamina propria immune cells showing the percentage of neutrophils and macrophages from total CD45⁺ cells at C. rodentium infection peak (7- dpi). (D) Ingenuity canonical bulk-RNA-seq pathways in Camp^+/+ and Camp^-/- colons at C. rodentium infection peak (7- dpi). The bar graph shows significant pathways with positive and negative Fold changes, indicating upregulation or downregulation of various signalling pathways and detoxification processes. (E) Gene relative abundance of Rnase6, an epithelial-derived antimicrobial peptide member of the RNase a superfamily. Data are shown as means ± SEM. *p < 0.05 and ** p < 0.01 (two-tailed Student’s t-test or one-way ANOVA post hoc Tukey’s test for multiple comparisons) were considered significant.

Figure 3.

Changes in the fecal microbiota during C. rodentium infection. (A) Bar plot displaying the relative abundance of bacterial taxa at the phylum level across different samples. Each bar represents a sample, and colours represent different bacterial species. (B) Alpha diversity analysis based on the Shannon diversity index. Each dot represents a sample. *p < 0.05 (two-tailed Student’s t-test or one-way ANOVA post hoc Tukey’s test for multiple comparisons) was considered significant. (C) Differential abundance analysis comparing the relative abundance between Camp^-/- and Camp^+/+ groups at 7- dpi, displayed as Log₂ Fold change (Log₂FC).

Figure 4.

Camp^-/- littermates display dysfunctional goblet cell mucus secretion. (A-B) identification and quantification of goblet cells by Alcian blue and PAS staining in distal colons of (A) Camp^+/+ and Camp^-/- littermates and (B) Camp^fl/fl and Villin^CreCamp^fl/fl infected orally with C. rodentium (5 ×10⁸ CFU) or control PBS. Goblet cells were quantified at C. rodentium infection peak (7- dpi) as the number of Alcian blue full-filled goblet cells per crypt (n 4–5 mice/group, with at least 6 crypts counted per mouse). Scale bar 20 µm. (C) transmission electron microscopy of goblet cells in the colon and quantification of mucin granules per goblet cell in controls (n = 3 mice/group, with at least 5 goblet cells imaged per mouse). Scale bar 2 µm. (D) FISH of uninfected colons using a pan-bacterial probe (EUB338-red) and DAPI (blue) to visualize the separation of the microbiota and epithelium (n = 3 mice/group with at least 13 measurements taken at random locations along the colonic cross-section). Scale bar 50 µm. Data are shown as individual measurements. (E) immunofluorescence Ki67 staining for epithelial proliferation (red). Three mice per group for 5 different fields of view were captured. The average numbers of Ki67⁺ cells were measured using ImageJ. Scale bar 10 µm. Data are shown as means ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 (two-tailed Student’s t-test or one-way ANOVA post hoc Tukey’s test for multiple comparisons) were considered significant.

Figure 5.

Camp^-/- littermates have an altered colonic mucus barrier. Camp^+/+ and Camp^-/- littermates were infected orally with C. rodentium (5 ×10⁸ CFU) or control PBS, and distal colons were collected and analyzed at infection peak (7- dpi). (A) RNAscope assay staining for Muc2 (green) and Tff3 (red) mRNAs (n = 3 mice per group for a total of 5 different fields of view). Scale bar 50 µm. Scored was adapted from the ACD RNAscope manual. (B) immunofluorescence glycan staining for fucose (UEA-I; green) and N-acetylglucosamine sialic acid (WGA; red). Three mice per group for 5 different fields of view were captured, and the percent positive areas were measured using ImageJ. Scale bar 25 µm. (C) confocal microphotographs of colonoid spheroids derived from Camp^+/+ and Camp^-/- littermates stained for fucose (UEA-I⁺; green) and sialic acid (WGA⁺; red) glycosylated mucin using DAPI to stain nuclei (representative images of n = 3 individual imaging experiments shown). Scale bar 50 µm. Data are shown as means ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 (two-tailed Student’s t-test or one-way ANOVA post hoc Tukey’s test for multiple comparisons) were considered significant.

Figure 6.

LL-37 stimulates the secretion of TFF3 and RELMβ from colonic goblet-like cells. Human colonic epithelial-like goblet LS174T cells were stimulated with human cathelicidin LL-37 or a scrambled peptide (scLL-37) (4.5 µM; 2 h). (A-B) quantification of secreted TFF3 in the supernatant of LS174T cells stimulated with (A) the indicated concentrations of LL-37 or scLL-37 for 2 h and (B) a fixed dose of LL-37 and scLL-37 (4.5 µM) for 12 and 24 h. (C) secreted RELMβ in the supernatant of LS174T cells stimulated with LL-37 or scLL-37 (4.5 µM;) for 2 h, 12 h and 24 h. (D) TFF3 and RETNLB gene expression in LS174T cells stimulated with scLL-37 or LL-37 (4.5 µM) for 2 h. (E) quantification of secreted TFF3 in LS174T cells following silencing of the CAMP gene (shLL-37) or vector (control, sh(-)) cells stimulated with increasing concentrations of LL-37. For all experiments, n = 4–6. Data are shown as means ± SEM. *p < 0.05 (two-tailed Student’s t-test or one-way ANOVA post hoc Tukey’s test for multiple comparisons) was considered significant.

Figure 7.

Alterations in Fcgbp in canonical Camp^-/- goblet cells. Camp^+/+ and Camp^-/- littermates were infected orally with C. rodentium (5 ×10⁸ CFU) or control PBS, and distal colons were collected and analyzed at C. rodentium infection peak (7- dpi). (A) gene abundance of goblet cell canonical Fcgbp and Clca1, Agr2, and non-canonical Aqp8, Hes1, Dmbt1, Gsdmc4 makers in Camp^+/+ and Camp^-/- littermates infected with C. rodentium based on bulk RNA analysis. (B) RNAscope assay staining for Fcgbp (red) mRNAs (n = 3 mice per group for 5 fields of view). Scale bar 50 µm. Scoring was adapted from the ACD RNAscope manual. (C) Western blotting analysis for Fcgbp protein in whole colons (n = 3–4 mice per group). Quantification and finding the ratio of Fcgbp over β-actin was done through ImageJ. For all experiments, n = 3–6. Data are shown as means ± SEM. *p < 0.05 (two-tailed Student’s t-test or one-way ANOVA post hoc Tukey’s test for multiple comparisons) was considered significant.

Figure 8.

Camp^-/- littermates show distinctive transcriptional profiling associated with impaired reactive oxygen species (ROS) synthesis. (A) Ingenuity canonical pathways in Camp^+/+ and Camp^-/- littermates at C. rodentium infection peak (7- dpi). The bar graph shows significant pathways with positive and negative Fold changes, indicating upregulation or downregulation of various signalling pathways and detoxification processes. The inset bar graph highlights the Fold change in the expression of Gpx2 and Nox4 genes related to ROS detoxification in Camp^+/+ and Camp^-/- littermates. (B) RNAscope assay staining for Duox 2 (red) mRNAs (n = 3 mice per group for 5 fields of view). Scale bar 50 µm. Scoring was adapted from the ACD RNAscope manual. (C) ROS levels measured by DCFDA fluorescence in colonic epithelial cells of Camp^+/+ and Camp^-/- littermates. For all experiments, n = 3–6. Data are shown as means ± SEM. ** p < 0.01 (two-tailed Student’s t-test or one-way ANOVA post hoc Tukey’s test for multiple comparisons) was considered significant.

Figure 9.

LL-37-induced TFF3 secretion requires reactive oxygen species (ROS) production. (A) ROS production in human LS174T colonic goblet cells and WT murine bone marrow-derived macrophages stimulated with LL-37, scramble peptide (scLL-37) (4.5 µM) or the positive control, hydrogen peroxide (H₂0₂; 100 µM) for 30 min. Cells were exposed to a general oxidative stress indicator (CM-H2DCFDA), and ROS production was measured using a fluorescent plate reader. Data are represented as a % increase relative to unstimulated cells ( = 8 for LS174T cells and n = 3 for murine BMMs). (B) MUC2 gene expression in LS174T cells stimulated with scLL-37 or LL-37 (4.5 µM) for 2 h. (C-E) quantification of secreted TFF3 in LS174T cells pretreated (30 min) with (C) an NADPH oxidase inhibitor (diphenyleneidodonium, DPI), (D) an endocytosis inhibitor (dynasore), (E) a scavenger for intracellular ROS (N-acetyl-l-cysteine, NAC), and (F) a mitochondria-specific ROS scavenger (Mito-TEMPO) followed by stimulation with LL-37 (4.5 µM; 2 h) (n = 4–6/group). Data are shown as means ± SEM. *p < 0.05 (one-way ANOVA post hoc Tukey’s test for multiple comparisons) was considered significant.

Figure 10.

A theoretical scheme of functions elicited by cathelicidins to promote mucus secretion in colonic goblet cells during infection. During intestinal infection with attaching/effacing citrobacter rodentium enteropathogen, cathelicidin released by non-epithelial cells, likely derived from infiltration neutrophils, activates NADPH oxidases and the production of reactive oxygen species (ROS). This ROS promotes the release of mucin granules and goblet cell-associated peptides, trefoil factor 3 (TFF3) and resistin-like molecule β (RELMβ). This enhancement in the colonic mucin barrier is implicated in lesser bacterial shedding at the peak of infection.