Faecalibacterium prausnitzii promotes intestinal epithelial IL-18 production through activation of the HIF1α pathway

Abstract

Introduction: Intestinal epithelial cells produce interleukin-18 (IL-18), a key factor in promoting epithelial barrier integrity. Here, we analyzed the potential role of gut bacteria and the hypoxia-inducible factor 1α (HIF1α) pathway in regulating mucosal IL18 expression in inflammatory bowel disease (IBD).

Methods: Mucosal samples from patients with IBD (n = 760) were analyzed for bacterial composition, IL18 levels and HIF1α pathway activation. Wild-type Caco-2 and CRISPR/Cas9-engineered Caco-2-HIF1A-null cells were cocultured with Faecalibacterium prausnitzii in a "Human oxygen-Bacteria anaerobic" in vitro system and analyzed by RNA sequencing.

Results: Mucosal IL18 mRNA levels correlated positively with the abundance of mucosal-associated butyrate-producing bacteria, in particular F. prausnitzii, and with HIF1α pathway activation in patients with IBD. HIF1α-mediated expression of IL18, either by a pharmacological agonist (dimethyloxallyl glycine) or F. prausnitzii, was abrogated in Caco-2-HIF1A-null cells.

Conclusion: Butyrate-producing gut bacteria like F. prausnitzii regulate mucosal IL18 expression in a HIF1α-dependent manner that may aid in mucosal healing in IBD.

Keywords: Faecalibacterium; HIF1α; HoxBan in vitro coculture system; IBD; IL-18; epithelial-bacteria interaction; intestine.

Figure 1

IL18 expression levels are significantly associated with abundances of F. prausnitzii in human intestinal mucosa. F. prausnitzii abundance is positively correlated to IL18 expression in intestinal mucosa. Patient cohort study performed in N = 760 biopsies.

Figure 2

IL18 expression strongly associates to HIF1α activation capacity in the human intestinal mucosa. (A) Gene expression levels of IL18 in intestinal mucosa of IBD patients, divided by location (colon and ileum). (B) HIF1α activation capacity (calculated by HIF1α scores) in intestinal mucosa of IBD patients. (C) Spearman correlation analysis shows significant positive association between IL18 and HIF1α scores in intestinal mucosa of IBD patients. Patient cohort study performed in n = 760 samples (biopsies), segregated by their inflammatory statues and biopsy location as colonic inflamed (n = 195), colonic non-inflamed (n = 307), ileal inflamed (n = 78) and ileal non-inflamed (n = 180) biopsies. Data presented as box and whiskers (min to max); ****p < 0.0001.

Figure 3

Stable knockout of HIF1A in Caco-2 cells leads to decreased HIF1α signaling upon pathway activation. (A) Time-dependent HIF1α responsive element (HRE)-luciferase assay over 48 h treatment with 1 mM DMOG. Gene expression levels of (B) HIF1A and the HIF1α pathway targets (C) PGK1, (D) EGLN3, and (E) IL18 in Caco-2-HIF1A-null and control cells treated with 1 or 5 mM DMOG for 24 h. Data presented as mean ± SD. All experiments were performed with n = 3; *p < 0.05, **p < 0.01, ****p < 0.0001.

Figure 4

Differential gene expression (DGE) analysis of Caco-2-HIF1A-null upon coculture with F. prausnitzii, compared to Caco-2 empty vector controls. (A) Schematic representation of the HoxBan coculture system, including experiment layout. Caco-2-HIF1A-null cells, or appropriate control (iCas9 empty vector), were cultured in the HoxBan system for 18 h in the absence or presence of F. prausnitzii inoculum (in figure, mono and co, respectively). Volcano plots showing differential gene expression analysis comparing Caco-2-HIF1A-null cells and Caco-2-empty vector in monoculture (B; purple and pink dots represent downregulated and upregulated genes, respectively) and coculture with F. prausnitzii (C; green and blue dots represent downregulated and upregulated genes, respectively). All dots display significant observations with p < 0.05. Veen diagrams were used to discover uniquely (D) upregulated and (E) downregulated DGE (numbers underlined inside blue and green diagrams, respectively) in Caco-2-HIF1A-null cultured with F. prausnitzii, compared to Caco-2-empty vector in same condition. All experiments performed in n = 3.

Figure 5

RNA sequencing analysis reveals that HIF1A ablation prevents upregulation of IL18 by F. prausnitzii in intestinal epithelial cells. (A,B) HIF1α and HIF2α scores shows a decrease in HIF1α, but not HIF2α, activation capacity in Caco-2-HIF1A-null, compared to Caco-2-empty vector control in coculture with F. prausnitzii. (C) Normalized gene expression counts of IL18 in Caco-2-HIF1A-null cells, compared to Caco-2-empty vector in monoculture and coculture with F. prausnitzii. (D) Spearman correlation analysis between IL18 gene expression and HIF1α scores on Caco-2 cells in the HoxBan system. All experiments performed in n = 3 (ns = not significant, **p < 0.001 and ***p < 0.0001).