A novel mouse model of Campylobacter jejuni enteropathy and diarrhea

Abstract

Campylobacter infections are among the leading bacterial causes of diarrhea and of 'environmental enteropathy' (EE) and growth failure worldwide. However, the lack of an inexpensive small animal model of enteric disease with Campylobacter has been a major limitation for understanding its pathogenesis, interventions or vaccine development. We describe a robust standard mouse model that can exhibit reproducible bloody diarrhea or growth failure, depending on the zinc or protein deficient diet and on antibiotic alteration of normal microbiota prior to infection. Zinc deficiency and the use of antibiotics create a niche for Campylobacter infection to establish by narrowing the metabolic flexibility of these mice for pathogen clearance and by promoting intestinal and systemic inflammation. Several biomarkers and intestinal pathology in this model also mimic those seen in human disease. This model provides a novel tool to test specific hypotheses regarding disease pathogenesis as well as vaccine development that is currently in progress.

Fig 1. Antibiotics are required for C. jejuni colonization.

Panel A. Weights following C. jejuni infection in house chow (HC) fed C57Bl/6J male mice. Panel B. qPCR detection of C. jejuni (cadF) following infection in house chow (HC) fed C57Bl/6J male mice. *, Significantly greater than no abx; P<0.05. Panel C. Myeloperoxidase (MPO) levels in mouse stool following C. jejuni infection. *, Significantly greater than all other groups; P<0.05. (N = 5/group).

Fig 2. Prolonged weight loss following Campylobacter infection.

Panel A. Mice fed either HC, or dPD had transient weight loss following infection, but dZD-fed infected mice had significant weight loss (* dZD infected vs uninfected days 2–14 post infection; P<0.001). While mice on dPD showed weight loss with no diarrhea, mice on HC had non-bloody soft stools on days 1–3 post infection, and mice on dZD had persistent bloody diarrhea on days 2–11 post infection. Panel B. Increasing Campylobacter detected in stool for the duration of the experiment (* dZD infected vs HC or dPD infected days 7&11 post infection; P<0.0001). Panel C. Images of stool following Campylobacter infection. Stool collected from HC (images from day 1 and 3 post infection) and dZD fed mice progressed from soft to bloody diarrhea by day 3 post infection (images from day 3 and 7 post infection). (N = 8/group).

Fig 3. Inflammatory biomarkers following Campylobacter infection.

The inflammatory biomarkers myeloperoxidase (MPO) and lipocalin-2 (LCN-2) were measured in cecal contents of antibiotic pretreated mice. LCN-2 was significantly elevated in dZD C. jejuni infected mice on both day 2 and 14 post infection, * dZD infected vs dZD uninfected; P<0.05. MPO was also significantly increased in dZD-fed infected mice on both day 2 and 14 post infection, ** dZD infected vs dZD uninfected; P<0.01.

Fig 4. Intestinal morphometry.

Intestinal sections from dZD fed mice stained for H&E or PAS. Panel A: ileum. Panel B: colon.

Fig 5. A non-encapsulated mutant of C. jejuni is defective in weight loss, diarrhea, and shedding of organism in stool. Panel A.

Weights following infection with either wildtype or 81–176 kpsM in dZD C57Bl/6J male mice. * WT vs kpsM, P<0.05. Diarrhea noted by D. Panel B. qPCR detection of Campylobacter jejuni (cadF) following infection in dZD mice. * WT vs kpsM, P<0.05. (N = 4/group).

Fig 6. Metabolic perturbations induced by Campylobacter infection.

Heatmap of the significant metabolic perturbations induced by C.jejuni infection identified by OPLS-DA models. Metabolic shifts are represented as correlation coefficients (r) of infected mice at days 4, 5 and 9 post infection versus diet and age-matched uninfected animals. Red and blue colors indicate increased or decreased excretion of metabolites following C.jejuni challenge respectively. Abbreviations: 2-OIV, 2-oxoisovalerate; 2-OIC, 2-oxoisocaproate; 2-MOV, 3-methyl-2-oxovalerate; 2-PY, N-methyl-2-pyridone-5-carboxamide; 4-CG; 4-cresol glucuronide; 4-CS, 4-cresyl sulfate; 4-HPA, 4-hydroxyphenylacetate; 4-HPPA, 4-Hydroxyphenylpyruvate; 4-PY, N-methyl-4-pyridone-3-carboxamide; DMA, dimethylamine; GAA guanidinoacetate; NAG, N-acetyl glutamine; NMND, N-methylnicotinamide; PAG, phenylacetylglycine; TMA, trimethylamine; TMAO, trimethylamine-N-oxide.

Fig 7. Correlations between urinary metabolites and inflammatory biomarkers.

Spearman’s correlation heatmap between the urinary metabolites identified in the OPLS-DA models and the levels of inflammatory biomarkers LCN-2 and MPO on day 14 post infection. Only significant correlations following P value adjustment are shown (Benjamini-Hochberg for 5% false discovery rate).

Fig 8. Summary of diversity and composition of the fecal microbiota of infected and uninfected mice on each diet.

A) Richness (number of unique amplicon sequence variants (ASVs)) in each sample. B) Evenness (inverse of the Simpson Index) for each sample. C) Non-metric multidimensional scaling (NMDS) of the bray-curtis dissimilarity of each sample. NMDS was performed using all samples shown in the three panels, then samples were plotted separately by diet to improve visualization.

Fig 9. Heatmap of differentially-abundant ASVs.

For each time point within each diet, microbiota were compared between infected and uninfected samples. Differentially-abundant ASVs from any comparison are shown in the heatmap, and the color of each cell indicates the log2 fold-change calculated with DESeq2. For each condition, only ASVs with a multiple testing-corrected p < 0.05 are shown (otherwise, cell is white). ASV labels include assigned family, genus and an ASV number used to differentiate multiple ASVs with identical genus-level assignment.