Enterocloster clostridioformis protects against Salmonella pathogenesis and modulates epithelial and mucosal immune function

Abstract

Background: Promoting resistance to enteric pathogen infection is a core function of the gut microbiota; however, many of the specific host-commensal interactions that mediate this protection remain uncharacterised. To address this knowledge gap, we monocolonised germ-free mice with mouse-derived commensal microbes to screen for microbiota-induced resistance to Salmonella Typhimurium infection.

Results: We identified Enterocloster clostridioformis as a protective species against S. Typhimurium infection. E. clostridioformis selectively upregulates resistin-like molecule β and cell cycle pathway expression at the level of caecal epithelial cells and increases T-regulatory cells in the underlying mucosal immune system, potentially contributing to reduced infection-induced pathology.

Conclusions: We highlight novel mechanisms of host-microbe interactions that can mediate microbiota-induced resistance to acute salmonellosis. In the backdrop of increasing antibiotic resistance, this study identifies novel potential avenues for further research into protective host responses against enteric infections and could lead to new therapeutic approaches. Video Abstract.

Keywords: Enterocloster clostridioformis; Escherichia coli; Acute salmonellosis; Host–commensal interactions; Microbiota; Mouse models; Pathogen resistance; Resistin-like molecule β; Screen; T-regulatory cells.

Fig. 1

Screened isolates of the Mouse Culture Collection. A Maximum-likelihood phylogenetic tree of the 276 isolates of the MCC in addition to the genome of Enterococcus_B lactis str. Com15 (formerly E. faecium). Distances represent BLOSUM45 matrix analysis of amino acid alignments of 120 bacterial core genes using the JTT + CAT model. Screened isolates are indicated in black boldface type. Inner colour bar indicates the taxonomic phylum of isolates. Filled black boxes illustrate isolates that successfully colonised mice after 14 days monocolonisation, while filled white boxes indicate species that were administered to mice but could not be recovered from faeces after 14 days. B Monocolonisation with commensal isolates prior to Salmonella Typhimurium infection. From left to right: species name, filled boxes demarcate isolates that were recovered from the faeces 14 days after inoculation, mouse weights at 1 dpi as a percentage of starting weights (0 dpi), mouse survival post-infection. Data shown are pooled from six screening experiments (n = 2–18). Wilcoxon tests comparing commensal isolate-monocolonised mice to PBS controls with Benjamini–Hochberg correction

Fig. 2

E. clostridioformis reduces severity of S. Typhimurium infection in germ-free mice. A Weight loss and B survival following infection with 10² CFU/mouse S.Tm. Weight loss data represent the mean ± SEM; means are not shown for timepoints after any mouse within the group was euthanised. Survival represents the day at which mice reached 80% of pre-infection weight. Kruskal–Wallis and log-rank tests were used for statistical analyses in A and B, respectively. Data shown are pooled from six experiments (n = 18). C Cumulative pathology scores for H&E-stained sections of the ileum, caecum, and colon of GF or E. clostridioformis-monocolonised mice at 1 dpi. Representative data shown are from one of two experimental repeats, mean ± SEM of cumulative pathology scores from individual mice (n = 2–4). D Representative sections of the caecum at 1 dpi (PBS, E. clostridioformis) or from noninfected GF mice (Mock). L, caecal lumen; Oe, submucosal oedema. Scale bars indicate 40 μm. E Breakdown of the caecum pathology scores at 1 dpi according to the six contributing subcomponents, mean ± SEM. Representative data shown are from one of two experimental repeats (n = 2–4)

Fig. 3

E. clostridioformis protects germ-free mice against S. Typhimurium infection independently of colonisation resistance. A S.Tm titres (CFU/g tissue) at 1 dpi in caecum luminal contents, mLN, liver, and spleen (left to right). Each data point represents a single mouse; Wilcoxon tests were used to assess statistical significance, and the Benjamini–Hochberg procedure were used to correct for multiple comparisons. Data shown are pooled from three experimental repeats (n = 5–13). B Titres of S.Tm following inoculation of filter-sterilised spent media (left) or stationary phase media (right) from commensal isolates. Data are pooled from 2 to 3 experimental repeats (n = 3–9). C Commensal and S.Tm titres at 24 h in LB broth under anaerobic conditions Pooled data are shown from two experiments (n = 6). D Growth of S.Tm in sterile-filtered spent media following addition of glucose. Data shown from one experiment (n = 3). E Growth of S.Tm following inoculation of caecal contents from GF or commensal-monocolonised mice. Representative data shown are from one of three experiments (n = 2–5)

Fig. 4

E. clostridioformis is closely associated with the caecal epithelium and induces expression of RELMβ. A Representative micrographs showing caecal (top row) and colonic (bottom row) epithelia from SPF mice and mice monocolonised with E. clostridioformis, E. coli, and A. faecis. Epithelial nuclei are stained with DAPI (blue), and bacteria are stained with a universal 16S rRNA gene FISH probe conjugated to FITC (green). Scale bars indicate 100 μm. B, C Epithelial proximity scores of the caecum (B) and colon (C). Scoring was performed by an independent, blinded investigator. Data are pooled from five experiments (n = 4–7) Wilcoxon tests with Benjamini–Hochberg correction. D Relative expression by qPCR of host defence genes in caecal epithelial cells isolated from GF, SPF, or E. clostridioformis-, E. coli-, or A. faecis-monocolonised mice. Data shown are pooled from six experiments (n = 4–14). Student’s t-tests with Benjamini–Hochberg correction

Fig. 5

E. clostridioformis and E. coli induce specific transcriptional signatures in caecal epithelial cells. Bulk RNA sequencing of caecal epithelial cells isolated from SPF mice and mice monocolonised with either E. clostridioformis or E. coli compared to GF controls (n = 6 from three experiments). A Venn diagram of significantly differentially expressed genes. B, C Volcano plots of genes from E. clostridioformis- and E. coli-monocolonised mice compared to GF controls. Genes highlighted in red or blue are significantly differentially expressed (FDR ≤ 0.05) with a log₂ fold change either ≥ 1.5 (red) or ≤ − 1.5 (blue). D Top 10 most significantly enriched GO terms following E. clostridioformis colonisation. E Heatmap showing genes from the cell cycle gene set that are differentially expressed following E. clostridioformis monocolonisation. F Top 10 most significantly enriched GO terms following E. coli colonisation. G Heatmap showing genes from the leukocyte activation gene set that are differentially expressed following E. coli monocolonisation

Fig. 6

E. clostridioformis increases the proportion of Tregs in the caecal epithelium, while E. coli increases γδ-T cells. Lymphocyte populations were quantified for different intestinal compartments using flow cytometry. Pooled data are shown from three experiments (n = 6–12). A Lymphocytes from the caecal intraepithelial compartment. B Lymphocytes from the caecal lamina propria. C Ratios of Foxp3 + CD4 + T-regulatory cells to Foxp3 − CD4 + T-effector cells from the caecal, colonic, and small intestinal intraepithelial compartments. Each data point represents a single mouse. Wilcoxon tests with Benjamini–Hochberg correction. Bars and error bars represent median ± median absolute deviation, respectively