Escherichia coli isolates from inflammatory bowel diseases patients survive in macrophages and activate NLRP3 inflammasome

Abstract

Crohn's disease (CD) is a multifactorial pathology associated with the presence of adherent-invasive Escherichia coli (AIEC) and NLRP3 polymorphic variants. The presence of intracellular E. coli in other intestinal pathologies (OIP) and the role of NLRP3-inflammasome in the immune response activated by these bacteria have not been investigated. In this study, we sought to characterize intracellular strains isolated from patients with CD, ulcerative colitis (UC) and OIP, and analyze NLRP3-inflammasome role in the immune response and bactericidal activity induced in macrophages exposed to invasive bacteria. For this, intracellular E. coli isolation from ileal biopsies, using gentamicin-protection assay, revealed a prevalence and CFU/biopsy of E. coli higher in biopsies from CD, UC and OIP patients than in controls. To characterize bacterial isolates, pulsed-field gel electrophoresis (PFGE) patterns, virulence genes, serogroup and phylogenetic group were analyzed. We found out that bacteria isolated from a given patient were closely related and shared virulence factors; however, strains from different patients were genetically heterogeneous. AIEC characteristics in isolated strains, such as invasive and replicative properties, were assessed in epithelial cells and macrophages, respectively. Some strains from CD and UC demonstrated AIEC properties, but not strains from OIP. Furthermore, the role of NLRP3 in pro-inflammatory cytokines production and bacterial elimination was determined in macrophages. E. coli strains induced IL-1β through NLRP3-dependent mechanism; however, their elimination by macrophages was independent of NLRP3. Invasiveness of intracellular E. coli strains into the intestinal mucosa and IL-1β production may contribute to CD and UC pathogenesis.

Keywords: Crohn's disease; E. coli; IL-1β; Inflammasome; NLRP3.

Copyright © 2014 Elsevier GmbH. All rights reserved.

Figures

Figure 1. Intracellular E. coli in intestinal biopsies from CD, UC and OIP patientes

Ileal biopsies from patients with CD (n=34), UC (n=57), OIP (n=17), and controls (n=22) were incubated with gentamicin for 1h, and homogenates were then plated in MacConkey agar. Graph shows the bacterial content represented as colony-forming units (CFU) of identified intracellular E. coli in each biopsy and mean bacteria in each group (Kruskal-Wallis two-tailed test, Dunn’s Multiple Comparison test was used, significance level set at <0.05). Prevalence of putative invasive E. coli is expressed as the percentage of patients with the bacteria in each group; Fisher’s exact test with a two-sided p-value was used to evaluate statistical significance as compared to the control group (p<0.05, CD p**=0.0023, UC p***=0.0005, OIP p***=0.0003).

Figure 2. Genetic variability among E. coli isolates from patients with CD, UC and OIP

Dendrogram based on pulsed-field gel electrophoresis (PFGE) of E. coli strain DNA digested with XbaI. PFGE analysis identified two groups of genetically-related strain clusters (I and II). Cluster A harboured the greatest number of strains and was in turn divided into I.1 (I.1.1 and I.1.2) and I.2 (I.2.1 and I.2.2). The right side of the figure shows the strain name (patient code-clone), virulence genes identified in each strain by PCR, phylogenetic group and serogroup. E. coli isolates were compared to HS commensal bacteria, and NRG857c and HM605 reference AIEC strains.

Figure 3. Invasion and adhesion of isolated E. coli strains in epithelial cells

(A) To evaluate invasion, Caco-2 cells were infected with representative E. coli strains isolated patient-derived tissue with a MOI=10 for 3h and then incubated with gentamicin for 1h. Graph shows the percentage of inoculum surviving gentamicin treatment. Graph represents mean and standard errors of three independent experiments in duplicate. (B) The capacity of isolated E. coli to adhere to Caco-2 cells was evaluated using a MOI=10 for 30 min. Graph shows the percentage of adhered bacteria to epithelial cells related to the initial inoculum. Data represents mean and standard errors of three independent experiments in duplicate.

Figure 4. IL-1β production by macrophages infected with isolated E. coli depends on the NLRP3 inflammasome

(A) Bone marrow-derived macrophages from Nlrp3−/−,Asc−/− or WT mice cells were primed with LPS for 4 h. After that, cells were infected with E. coli isolated from CD (CD1-a, CD2-a, CD6-b, CD6-r, CD8-a) patients, control (C7-a) or HS strain (MOI=10) for 1 h and then treated with gentamicin. Twenty-four h.p.i, supernatants were analysed for IL-1β TNF-α and (C) IL-6 secretion by ELISA. (D) As an indicator of caspase-1 activation, content of caspase-1 p20 subunit was determined in WT or Nlrp3−/− macrophages primed with LPS for 4 h, infected with CD2-a or NRG857c strains (MOI=10) for two h and then treated with gentamicin overnight. Extracts were prepared from cell and supernatant, and immunobloted using a caspase-1 antibody. Arrows denote procaspase-1 and its processed p20 subunit. Graph bar shows content of caspase-1 p20 subunit related to procaspase-1 in each condition expressed as the fold induction in infected to non-infected cells. Graph represents mean and standard errors of three independent experiments (E) Bacterial clearance ability of Nlrp3−/− or WT macrophages infected with CD2-a or HS strains was evaluated at 3, 24 and 48 h.p.i. Graph shows the ratio of gentamicin-resistant bacteria to initial uptake (3 h.p.i). Graph represents mean and standard errors of three independent experiments, performed in duplicate. No significant differences were found in bacterial clearance between Nlrp3−/− and WT macrophages, infected with the same strain, using Fisher’s exact test with a two-sided p-value.