Characterization of the pathophysiological determinants of diarrheagenic Escherichia coli infection using a challenge model in healthy adults

Abstract

An experimental human challenge model with an attenuated diarrheagenic Escherichia coli (E. coli) strain has been used in food intervention studies aimed to increase resistance to E. coli infection. This study was designed to refine and expand this challenge model. In a double-blind study, healthy male subjects were orally challenged with 1E10 or 5E10 colony-forming units (CFU) of E. coli strain E1392/75-2A. Three weeks later, subjects were rechallenged with 1E10 CFU of E. coli. Before and after both challenges, clinical symptoms and infection- and immune-related biomarkers were analyzed. Subset analysis was performed on clinically high- and low-responders. Regardless of inoculation dose, the first challenge induced clinical symptoms for 2-3 days. In blood, neutrophils, CRP, CXCL10, and CFA/II-specific IgG were induced, and in feces calprotectin and CFA/II-specific IgA. Despite clinical differences between high- and low-responders, infection and immune biomarkers did not differ. The first inoculation induced protection at the second challenge, with a minor clinical response, and no change in biomarkers. The refined study design resulted in a larger dynamic range of symptoms, and identification of biomarkers induced by a challenge with the attenuated E. coli strain E1392/75-2A, which is of value for future intervention studies. Addition of a second inoculation allows to study the protective response induced by a primary infection.Clinicaltrials.gov registration: NCT02541695 (04/09/2015).

Conflict of interest statement

AM, KV, JB were employed by Société des Produits Nestlé SA, Switzerland. EvH, JS, AK, EL-vdB, StB received an unrestricted grant from Société des Produits Nestlé SA, Switzerland, to execute the study. All other authors declare no competing interests.

Figures

Figure 1

Effect of the dose of oral E. coli E1392/75-2A challenge on clinical parameters. Percentage and total fecal wet weight (A,B), stool consistency and stool frequency (C,D) and gastrointestinal disease-related symptoms and quality of life (E,F) were measured before and after infection with E. coli (strain E1392/75-2A) on study day 14 with standard dose 1E10 CFU (n = 20; black circles/bars) or high dose 5E10 CFU (n = 22; grey open circles/bars). On study day 35, all subjects received a second inoculation of 1E10 CFU of E. coli (n = 42). Results are expressed as means ± SEM. Timepoint at which significance is reached as compared to baseline day 13 (within the first infection period) or day 34 (within the second infection period) is indicated by *, with p < 0.05 within the specified group (repeated measures ANOVA ; black *: within standard dose; grey *: within high dose). Significant differences in the overall response between standard and high dose (two-way ANOVA) are indicated separately for the first (a, days 11–17) and second (b, days 32–38) infection period; ns not significant.

Figure 2

Effect of oral E. coli E1392/75-2A challenge in high- and low-responders on clinical parameters. Percentage and total fecal wet weight (A,B), stool consistency and stool frequency (C,D) and gastrointestinal disease-related symptoms and quality of life (E,F) were measured before and after infection with E. coli on study day 14 (strain E1392/75-2A; dose 1E10 CFU or 5E10 CFU). On study day 35, all subjects received a second inoculation of 1E10 CFU of E. coli (n = 42). Results are shown for a subset of high-responders (n = 12; black circles/bars) and low-responders (n = 12; grey diamonds/bars), selected based on ranking of a combination of clinical outcomes, and are expressed as means ± SEM. Timepoint at which significance is reached as compared to baseline day 13 (within the first infection period) or day 34 (within the second infection period) is indicated by *, with p < 0.05 within the specified group (repeated measures ANOVA; black *: within high-responders; grey *: within low-responders). Significant differences in the overall response between high- and low-responders (two-way ANOVA) are indicated separately for the first (a, day 11–17) and second (b, day 32–38) infection period; ns not significant.

Figure 3

Effect of oral E. coli E1392/75-2A challenge on local fecal inflammatory and immune biomarkers. In fecal samples of high- and low-responders (n = 12 each), levels of calprotectin (A), β-defensin 2 (B), total sIgA (C) and fecal CFA/II-specific IgA (D) were analyzed. On day 14, subjects were orally infected with E. coli strain E1392/75-2A (dose 1E10 CFU or 5E10 CFU). On study day 35, all subjects received a second inoculation of 1E10 CFU of E. coli. Results are shown as means ± SEM. Timepoint at which significance is reached as compared to baseline day 13 (within the first infection period) or day 34 (within the second infection period) is indicated by *, with p < 0.05 within the specified group (repeated measures ANOVA; black *: within high-responders; grey *: within low-responders); #p < 0.05 at day 36 compared to baseline day 13. Significant differences in the overall response between high- and low-responders (two-way ANOVA) are indicated separately for the first (a, days 13–34) and second (b, days 34–42) infection period; ns not significant.

Figure 4

Effect of oral E. coli E1392/75-2A challenge on systemic inflammatory and immune biomarkers. In blood of high- and low-responders (n = 12 each), lymphocyte, neutrophil and monocyte counts were measured (A–C). In serum or plasma, levels of C-reactive protein, IP-10 (CXCL10) and CFA/II-specific IgG were measured (D–F). On day 14, subjects were orally infected with E. coli strain E1392/75-2A (dose 1E10 CFU or 5E10 CFU). On study day 35, all subjects received a second inoculation of 1E10 CFU of E. coli. Results are shown as means ± SEM. Timepoint at which significance is reached as compared to baseline day 10 (within the first infection period) or day 28/31 (within the second infection period) is indicated by *, with p < 0.05 within the specified group (repeated measures ANOVA; black *within high-responders; grey *within low-responders). Significant differences in the overall response between high- and low-responders (two-way ANOVA) are indicated separately for the first (a, days 10–31) and second (b, days 28/31–49) infection period; ns not significant.

Figure 5

Ex vivo cytokine and chemokine production in whole blood before and after E. coli challenge. Ex vivo stimulation of whole blood cells was performed in a subset of subjects (n = 6 low-responders, closed circles, and n = 6 high-responders, open circles). Subjects were orally infected with a live attenuated E. coli strain E1392/75-2A on study day 14. Whole blood stimulation was performed on study day 10 and 17, using Null TruCulture tubes (medium only). No difference was observed between low- and high-responders, therefore the data of both subgroups were combined in statistical analysis. From the total panel of cytokines and chemokines that were analysed, the figures are shown for those that changed (borderline) significantly after challenge (Wilcoxon Signed Rank test).

Figure 6

Flow diagram of study subjects.