Clostridium difficile flagella induce a pro-inflammatory response in intestinal epithelium of mice in cooperation with toxins

Abstract

Clostridium difficile is the most important enteropathogen involved in gut nosocomial post-antibiotic infections. The emergence of hypervirulent strains has contributed to increased mortality and morbidity of CDI. The C. difficile toxins contribute directly to CDI-associated lesions of the gut, but other bacterial factors are needed for the bacteria to adhere and colonize the intestinal epithelium. The C. difficile flagella, which confer motility and chemotaxis for successful intestinal colonization, could play an additional role in bacterial pathogenesis by contributing to the inflammatory response of the host and mucosal injury. Indeed, by activating the TLR5, flagella can elicit activation of the MAPK and NF-κB cascades of cell signaling, leading to the secretion of pro-inflammatory cytokines. In the current study, we demonstrate, by using an animal model of CDI, a synergic effect of flagella and toxins in eliciting an inflammatory mucosal response. In this model, the absence of flagella dramatically decreases the degree of mucosal inflammation in mice and the sole presence of toxins without flagella was not enough to elicit epithelial lesions. These results highlight the important role of C. difficile flagella in eliciting mucosal lesions as long as the toxins exert their action on the epithelium.

Figure 1

C. difficile R20291 flagella are involved in caecal inflammation of mice. C57BL/6 mice (n = 10) were infected or not and caeca were collected at the clinical end point day 2 for histology. (A) Representative histology of caecum of healthy uninfected mice (control) with epithelium integrity; (B) R20291 WT-infected mice with ulcerative colitis, necrosis, desquamation, exudates and necrotic cells in the intestinal lumen, edema, inflammatory submucosal cell infiltration, loss of architecture of epithelium and presence of pseudomembranes in caecum; (C) R20291 fliC and (D) A⁻B⁻ mutant-infected mice with edema, focal desquamation of necrotic cells in the intestinal lumen and submucosal cell infiltrate; (E) R20291 motB mutant-infected mice with similar degree of mucosal lesion than R20291 WT-infected mice (200X magnification). Bar = 200 µm. (F) Inflammation histological scores for individual mice in each group. The horizontal lines represent the mean scores for each group of animals. (G) Individual evaluation criteria of intestinal inflammation. The bars represent the mean scores for each group of animals and standard deviations. *P < 0.01 compared to WT-infected mice.

Figure 2

C. difficile 630Δerm flagella are involved in caecal inflammation of mice. C57BL/6 mice (n = 10) were infected by oral gavage or not and caeca were collected at the clinical end point day 2 for histology. (A) Representative histology of caecum of healthy uninfected mice (control) with epithelium integrity; (B) 630Δerm WT-infected mice with moderate desquamation, exudates and necrotic cells in the intestinal lumen, edema, and inflammatory submucosa cell infiltration; (C) fliC and (D) A⁻B⁻ mutant-infected mice with normal mucosa (200X magnification). Bar = 200 µm. (E) Inflammation histological scores for individual mice in each group. The horizontal lines represent the mean scores for each group of animals. (F) Individual evaluation criteria of intestinal inflammation. The bars represent the mean scores for each group of animals and standard deviations. *P < 0.01 compared to WT-infected mice.

Figure 3

C. difficile flagella and caecal inflammation in C57BL/6 tlr5 ^−/− KO mice. C57BL/6 tlr5 ^−/− KO mice (n = 10) were infected by oral gavage or not and caeca were collected at the clinical end point day 2 for histology. (A) Representative histology of caecum of healthy uninfected mice (control) with epithelium integrity; (B) R20291 WT-infected mice with mild to moderate edema; (C) R20291 fliC mutant and (D) 630Δerm WT-infected mice with mild to moderate edema (200X magnification). Bar = 200 µm. (E) Inflammation histological scores for individual mice in each group. The horizontal lines represent the mean scores for each group of animals. (F) Individual evaluation criteria of intestinal inflammation. The bars represent the mean scores for each group of animals and standard deviations. *P < 0.01 compared to R20291 WT-infected mice.

Figure 4

C. difficile flagella-induced degradation of IκB-α. Caecal lysates were prepared as indicated in Material and Methods section and proteins were resolved by SDS-PAGE. (A) R20291 derivative-infected C57BL/6 mice; (B) 630Δerm derivative-infected C57BL/6 mice, and (C) R20291 derivative- or 630Δerm WT-infected C57BL/6 tlr5 ^−/− KO mice. Western blot were then performed using anti-IκB-α, actin antibodies. Western blot cropped pictures (full-length blots are in the Supplementary Information file) show the results of a representative experiment. The density of the bands was measured using Fusion software. Ratios IκB-α/actin were calculated and for the negative control (C-, uninfected mice), this ratio was normalized to 1. The ratio of the other samples was reported to the negative control. Results represent the mean (n = 10) ± standard deviations for each group of animals. *Statistically significant differences (P < 0.05) compared to the negative control.