Age-Dependent Susceptibility to Enteropathogenic Escherichia coli (EPEC) Infection in Mice

Abstract

Enteropathogenic Escherichia coli (EPEC) represents a major causative agent of infant diarrhea associated with significant morbidity and mortality in developing countries. Although studied extensively in vitro, the investigation of the host-pathogen interaction in vivo has been hampered by the lack of a suitable small animal model. Using RT-PCR and global transcriptome analysis, high throughput 16S rDNA sequencing as well as immunofluorescence and electron microscopy, we characterize the EPEC-host interaction following oral challenge of newborn mice. Spontaneous colonization of the small intestine and colon of neonate mice that lasted until weaning was observed. Intimate attachment to the epithelial plasma membrane and microcolony formation were visualized only in the presence of a functional bundle forming pili (BFP) and type III secretion system (T3SS). Similarly, a T3SS-dependent EPEC-induced innate immune response, mediated via MyD88, TLR5 and TLR9 led to the induction of a distinct set of genes in infected intestinal epithelial cells. Infection-induced alterations of the microbiota composition remained restricted to the postnatal period. Although EPEC colonized the adult intestine in the absence of a competing microbiota, no microcolonies were observed at the small intestinal epithelium. Here, we introduce the first suitable mouse infection model and describe an age-dependent, virulence factor-dependent attachment of EPEC to enterocytes in vivo.

Fig 1. Intestinal colonization following oral administration.

(A-B) Neonate and adult mice were orally infected with WT EPEC. Small intestine (A) and colon (B) tissues were collected 4 days p.i., homogenized and plated on LB agar plates supplemented with the appropriate antibiotic (n = 12-23/time point from at least 2 litters, median). (C-D) 1-day-old mice were orally infected with WT EPEC. Small intestine (C) and colon (D) tissues were collected at defined time points p.i., homogenized and plated on LB agar plates supplemented with the appropriate antibiotic (n = 9-16/time point from at least 3 litters; median). ANOVA with Dunnett’s post-test. **, p<0.01; ***, p<0.001.

Fig 2. T3SS- and BFP-dependent generation of epithelial microcolonies in vivo.

(A) Immunostaining and electron microscopy of small intestinal tissue sections collected 8 days p.i. from mice orally infected on the day of birth with WT EPEC. (Ai;vi-vii) 2d representation of microcolonies attached to the small intestinal epithelium. (Aii) enlarged 3d representation of the insert marked in (Ai) (EPEC, red; E-cadherin, white; wheat germ agglutinin (mucus), green; DAPI, blue, bar = 5μm). (Aiii) SEM of EPEC forming a microcolony (bar = 1 μm). (Aiv-v) TEM of EPEC attached to the epithelial plasma membrane (bar = 0.5 μm). (B) Number of microcolonies per small intestinal tissue section at defined time points p.i. (n = 9 from 3 mice per time point, mean ± SD). (C) 1-day-old mice were orally exposed to WT EPEC, escV or bfpA single mutants, a escV/bfpA double mutant or two commensal E. coli strains (#1, a murine commensal E. coli isolate; #2, E. coli Nissle). Small intestinal tissues were collected 4 days p.i., homogenized and plated on LB agar plates supplemented with streptomycin (WT), kanamycin (escV, bfpA, escV/bfpA mutants) or ampicillin (commensal E. coli strains) (n = 7–13 from at least 2 litters; median). (D) 1-day-old mice were orally co-infected with WT EPEC and either escV or bfpA mutants at a 1:1 ratio (total: 1–2×10⁵ CFU). Total small intestinal tissues were collected 8 days p.i., homogenized and plated on different LB agar plates supplemented with the appropriate antibiotic to discriminate WT EPEC from escV or bfpA mutants (n = 15–24 from at least 2 litters; data are represented in box and whisker plot format). (E) Immunostaining of small intestinal tissue sections collected 8 days p.i. from mice orally infected at birth with escV (i) or bfpA mutants (ii) (EPEC, red; E-cadherin, white; wheat germ agglutinin (mucus), green; DAPI, blue; bar = 5μm). (F) Number of microcolonies observed per small intestinal tissue section in animals infected with WT EPEC or with escV or bfpA mutants at 8 days p.i. (n = 9 from 3 mice per time point, mean ± SD). ANOVA with Dunnett’s post-test (C and F). ns, p>0.05; *, p<0.05; **, p<0.01; ***, p<0.001.

Fig 3. Characterization of the epithelial response to EPEC infection in vivo.

(A) Heat map showing the intestinal epithelial gene expression in small intestinal tissue of uninfected mice (uninf.) as well as at day 8 p.i. with WT EPEC or escV mutant (n = 4–6 from at least 2 litters). A selection of the most significantly up-regulated genes following WT EPEC infection are shown (p-value = 0.02; q-value = 0.22). (B) Clusters of orthologous group (COG) analysis of the genes shown in (A). (C-D) 1-day-old mice were orally infected with WT EPEC, escV mutant, bfpA mutant, escV/bfpA double mutant EPEC, a murine commensal E. coli strain or left untreated. IEC were isolated from the small intestinal tissue at 8 days p.i. and the expression levels of (C) saa3 and (D) cpn2 were determined by quantitative RT-PCR and normalized to the values obtained for the housekeeping gene hprt (n = 4–16 from at least 2 litters; median). Kruskal-Wallis ANOVA with Dunn’s multiple comparison post-test (C-D). ns, p>0.05; **, p<0.01; ***, p<0.001.

Fig 4. Analysis of bacterial and host factors required for the EPEC-induced epithelial response.

(A) 1-day-old mice were orally infected with WT EPEC, escV mutant, bfpA mutant, escV/bfpA double mutant EPEC, a murine commensal E. coli strain or left untreated. IEC were isolated 8 days p.i. and the expression level of RegIIIγ was determined and normalized to the values obtained for hprt (n = 4–16 from at least 2 litters; median). (B) Immunostaining for RegIIIγ in the distal part of the small intestine of 9-day-old untreated control animals (i) or mice infected at birth with WT (ii), escV mutant (iii), or bfpA mutant EPEC (iv) at 8 days p.i. (RegIIIγ, red; E-cadherin, white; DAPI, blue; bar = 50 μm). (C) Time kinetic of the epithelial RegIIIγ expression in infected neonates and age-matched controls. 1-day-old mice were orally infected with WT EPEC (black bars) or left untreated (white bars). IEC were isolated at 5, 9, 13, 17 and 21 days after birth and the expression of RegIIIγ was measured and normalized to hprt (n = 7–17 from at least 2 litters; geometric mean ± 95% confidence interval). (D) 1-day-old C57BL/6 WT, MyD88^-/-, TRIF^-/-, Tlr4^-/-, Tlr5^-/- and Tlr9^-/- mice were orally infected with WT EPEC (black bars) or left untreated (white bars) and the epithelial RegIIIγ expression was measured and normalized to hprt (n = 11–27 from at least 2 litters; geometric mean ± 95% confidence interval). Kruskal-Wallis ANOVA with Dunn’s multiple comparison post-test (A and C-D). ns, p>0.05; *, p<0.05; **, p<0.01; ***, p<0.001.

Fig 5. EPEC infection temporally alters the composition of the enteric microbiota.

Bacterial DNA was extracted from the small intestine and colon of newborn mice infected with WT EPEC at 8 and 20 days p.i. or untreated age-matched control animals and analyzed by 16S rDNA sequencing. (A) Relative abundance of the OTU 4425571 (Enterobacteriaceae, representing EPEC in red) in infected and uninfected animals. (B) PC1/PC2 plot illustrating the similarity of the microbiota within the different groups in the small intestine (left) and in the colon (right) (d8 uninfected, blue squares; d8 EPEC-infected, red squares; d20 uninfected, green triangles; d20 EPEC-infected, orange triangles). (C) Phylum level composition of the 8 different groups. Proteobacteria are represented in different shades of red, Bacteroides in different shades of yellow, Firmicutes in different shades of green and Tenericutes in blue.

Fig 6. The influence of the enteric microbiota and intestinal development on colonization and microcolony formation.

(A) 1-day-old conventional (conv.) and germ-free (GF) mice were orally infected with WT EPEC. Small intestinal and colon tissues were collected 20 days p.i., homogenized and plated on LB agar plates supplemented with the appropriate antibiotic (n = 6–12 from 1 litter for GF mice and 5 litters for conventional mice; median). (B) Conventional adult mice treated with streptomycin in drinking water (red dotted line) or left untreated (black line) as well as untreated adult germ-free mice (blue dotted line) were infected with WT EPEC. Streptomycin treatment was stopped 8 days p.i. (red arrow). Feces were collected at the indicated time point p.i., homogenized and plated on LB agar plates supplemented with the appropriate antibiotic (n = 3–33; median ± interquartile range). (C) Immunostaining of small intestinal tissue sections collected 8 days p.i. from adult streptomycin-treated (i) and germ-free mice (ii) orally infected with WT EPEC (EPEC, red; E-cadherin, white; wheat germ agglutinin (mucus), green; DAPI, blue; bar = 5μm). (D) Number of microcolonies observed per small intestinal tissue section of 1-day-old and non-treated, streptomycin-treated and GF adult animals infected with WT EPEC 8 days p.i. (n = 9 from 3 mice per time point, mean ± SD). (E) IEC were isolated from the small intestines of adult streptomycin-treated, WT EPEC or escV mutant-infected animals at 8 days p.i. and the expression level of RegIIIγ was measured and normalized to the values obtained for hprt (n = 5–10; median). Student’s t-test (A) and Kruskal-Wallis ANOVA with Dunn’s multiple comparison post-test (E). ns, p>0.05; ***, p<0.001.