Inflammation promotes stomach epithelial defense by stimulating the secretion of antimicrobial peptides in the mucus

Abstract

The mucus serves as a protective barrier in the gastrointestinal tract against microbial attacks. While its role extends beyond merely being a physical barrier, the extent of its active bactericidal properties remains unclear, and the mechanisms regulating these properties are not yet understood. We propose that inflammation induces epithelial cells to secrete antimicrobial peptides, transforming mucus into an active bactericidal agent. To investigate the properties of mucus, we previously developed mucosoid culture models that mimic the healthy human stomach epithelium. Similar to organoids, mucosoids are stem cell-driven cultures; however, the cells are cultivated on transwells at air-liquid interface. The epithelial cells of mucosoids form a polarized monolayer, allowing differentiation into all stomach lineages, including mucus-secreting cells. This setup facilitates the secretion and accumulation of mucus on the apical side of the mucosoids, enabling analysis of its bactericidal effects and protein composition, including antimicrobial peptides. Our findings show that TNFα, IL1β, and IFNγ induce the secretion of antimicrobials such as lactotransferrin, lipocalin2, complement component 3, and CXCL9 into the mucus. This antimicrobial-enriched mucus can partially eliminate Helicobacter pylori, a key stomach pathogen. The bactericidal activity depends on the concentration of each antimicrobial and their gene expression is higher in patients with inflammation and H.pylori-associated chronic gastritis. However, we also find that H. pylori infection can reduce the expression of antimicrobial encoding genes promoted by inflammation. These findings suggest that controlling antimicrobial secretion in the mucus is a critical component of epithelial immunity. However, pathogens like H. pylori can overcome these defenses and survive in the mucosa.

Keywords: Antimicrobial peptides; Epithelial defence; Helicobacter pylori; Inflammation; Innate immunity; Mucus; Stomach.

Graphical abstract

Figure 1.

Inflammation triggers the expression of genes encoding for antimicrobial peptides. Cells from three different mucosoid lines were treated with TNFα (5 ng/mL), IL1β (2.5 ng/mL), and IFNγ (10 ng/mL) combined or individually (as indicated), for 5 days before extracting RNA for transcriptomic profile by microarray. (a). Heatmap of the most relevant GSEA hallmarks (https://www.gsea-msigdb.org/gsea/msigdb/collections.jsp) relative to the comparison of the transcriptome of the treated vs. untreated mucosoids. NES= normalized enrichment score, *= false discovery rate < 0.0001. Red = upregulation, white = no change, blue = downregulation. (b). Enrichment plot of the most significant GSEA hallmark resulting from the comparison of the transcriptome of the mucosoids treated with TNFα, IL1β, and IFNγ, compared to the untreated samples. (c). Enrichment plot of the GSEA hallmark “antimicrobial peptides” (d). Heatmap of validated nf-κB target genes as reported in (https://www.bu.edu/nf-kb/gene-resources/target-genes/) (e). Heatmap of selected interferon-dependent chemokines, enzymes and transcription factors. (f). Heatmap of genes coding for antimicrobial peptides. Genes showing consistent upregulation in at least one of the conditions are outlined in black.

Figure 2.

Inflammation induce the secretion of AMPs in the mucus. (a). Mucus was removed from three different mucosoid lines after treatment with TNFα (5 ng/mL), IL1β (2.5 ng/mL), and IFNγ (10 ng/mL) combined or individually (as indicated) for 5 days. The total mucus secreted and accumulated was analyzed by mass spectrometry. (b). Heatmap of the abundance of the proteins in the mucus determined by label free quantification (LFQ). Only proteins with an altered expression of at least 10-fold in one of the conditions are represented. Outlined in black are those proteins with reported bactericidal activity. The other proteins are categorized by location or function. (c)(d)(e)(f) The Log₁₀LFQ values are plotted for LTF (lactotransferrin), LCN2 (lipocalin 2), C3 (complement component C3), CXCL9 (C-X-C motif chemokine ligand 9) the horizontal bar represent the median of tree individual measurements from the mucus produced by three different mucosoid lines (GAT23, 27, 29). Multiple comparison test was performed after one way ANOVA to assess if the differences between treated and untreated samples are significant: *p = <0.05, **p = <0.005, ***p = <0.0005, **** p= <0.00005. The relative abundance of other antimicrobials CXCL1 and CXCL3 did not pass the statistical test for any of the multiple comparisons. (g)(h)(i)(j) validation of expression of the antimicrobial was confirmed by Western blot using a further two samples of mucosoids (GAT28, GAT31) treated with the combination of pro-inflammatory cytokines as before. As housekeeping proteins are not yet available for the mucus, Red Ponceau was used as a loading control. (k)(l)(m)(n) Mucosoids were treated with a combination of pro-inflammatory cytokines as before, and immunofluorescence experiments carried out to localize the expression of intracellular LTF, LCN2, C3 and CXCL9. Stained whole mount mucosoids were imaged using confocal microscopy. Images were taken across the whole thickness of the monolayer (z-stacks) and the panels shows a top and a lateral projection. Scale bar = 5 µm.

Figure 3.

The bactericidal activity of antimicrobial peptides found in the mucus. ((a) A suspension of 10⁵ CFU of kanamycin resistant H.pylori was incubated for 1 h and 2 h in the presence of different concentrations (5, 10, or 30 mg/mL) of AMPs. The suspension was plated in kanamycin plates and the number of colonies was counted relative to PBS to assess the bactericidal range of each antimicrobial. (b). Bactericidal activity of Lactotransferrin using the bioactive peptide lactoferricin (LFcin), the antimicrobial peptide derived from the hydrolysis of lactotransferrin. (c). Bactericidal activity of LCN2. (d). Bactericidal activity of C3a, the peptide derived from complement component 3. (e). Bactericidal activity of CXCL9. Multiple comparison test was performed after a two way ANOVA to assess if the difference in bactericidal activity, for each concentration of AMP, between different time points, is statistically significant. p values comparing t = 2 h and t = 0 are reported. ns= non significant. *p = <0.05, **p = <0.005, ***p = <0.0005, ****p = <0.00005. (f). Mucosoids were treated with TNFα (5 ng/mL), IL1β (2.5 ng/mL), and IFNγ (10 ng/mL) for 5 days, with a fresh supply of cytokines added on the third day. Mucus samples from mucosoids derived from 3 different patients was incubated with 10⁵ CFU of a kanamycin resistant isogenic P12 strain for 2 hours before plating and assessing the bactericidal activity by counting the percentage of colonies relative to PBS. Paired samples t-test was used to assess if the difference between the bactericidal activity of the two groups was significant. ***= p < 0.0005. (g) Mucosoids were treated with a combination of pro-inflammatory cytokines as before and infected on the second day with H.pylori at MOI 100 for 3 days. Mucosoids were washed with PBS to remove the dead cells, they were stained with DAPI and imaged using confocal microscopy to count the remaining cells after infection. Each dot represents the density of cells per area of imaging (0.18 mm²) and paired samples t-test was used to assess if the difference between the groups is significant *p=<0.05.

Figure 4.

The expression of AMPs in biopsy samples from patients with chronic gastritis. RNA from 16 biopsy samples from H.pylori positive chronic gastritis (GC) cases and 8 from normal-looking stomach (N) was retrotranscribed into cDNA for the detection of AMP genes and of genes encoding for pro-inflammatory cytokines. (a)–(g). The Log10 relative expression of LTF, LCN2, C3, CXCL9, TNFA, IL1B and IFNG is reported. The significance of the difference between the groups was calculated using an unpaired t-test. *p=<0.05, **p=<0.005, ***p=<0.0005, ****p=<0.00005. (h). The Log10 fold induction value of each gene detected in the chronic gastritis biopsies was used to compute a pearson R linear correlation heatmap. 1 = highest correlation, 0 = no correlation, −1=highest negative correlation.

Figure 5.

H. pylori downregulates the expression of antimicrobials of inflamed epithelial cells (a). Three different mucosoid lines were treated by four different protocols, according to the schematic. Upper: non-treated (NT); second row: H. pylori infection carried out at day 2 at MOI 100; third row: treatment with pro-inflammatory cytokines (TNFα, 5 ng/mL; IL1β, 2.5 ng/mL; IFNγ 10 ng/mL) from day 0; lower: treatment with pro-inflammatory cytokines at the same concentrations as previously from day 0, following H. pylori infection at MOI 100 at day 2. At day 5, RNA was extracted and the abundance of transcripts was measured relative to non-treated controls. Due to the large differences between the expression of genes under different treatments, results are expressed in Log10-fold change. (b). The expression of CXCL8 does not change when H. pylori infection follows the treatment with pro-inflammatory cytokines. (c)(d)(e) The expression of LTF, C3 and LCN2 is downregulated when H. pylori infection is carried out in addition to the treatment with pro-inflammatory cytokines. (f) CXCL9 expression does not significantly change when H. pylori infection is in addition to to the treatment with pro-inflammatory cytokines.