Intestinal microbiota composition of interleukin-10 deficient C57BL/6J mice and susceptibility to Helicobacter hepaticus-induced colitis

Abstract

The mouse pathobiont Helicobacter hepaticus can induce typhlocolitis in interleukin-10-deficient mice, and H. hepaticus infection of immunodeficient mice is widely used as a model to study the role of pathogens and commensal bacteria in the pathogenesis of inflammatory bowel disease. C57BL/6J Il10(-/-) mice kept under specific pathogen-free conditions in two different facilities (MHH and MIT), displayed strong differences with respect to their susceptibilities to H. hepaticus-induced intestinal pathology. Mice at MIT developed robust typhlocolitis after infection with H. hepaticus, while mice at MHH developed no significant pathology after infection with the same H. hepaticus strain. We hypothesized that the intestinal microbiota might be responsible for these differences and therefore performed high resolution analysis of the intestinal microbiota composition in uninfected mice from the two facilities by deep sequencing of partial 16S rRNA amplicons. The microbiota composition differed markedly between mice from both facilities. Significant differences were also detected between two groups of MHH mice born in different years. Of the 119 operational taxonomic units (OTUs) that occurred in at least half the cecum or colon samples of at least one mouse group, 24 were only found in MIT mice, and another 13 OTUs could only be found in MHH samples. While most of the MHH-specific OTUs could only be identified to class or family level, the MIT-specific set contained OTUs identified to genus or species level, including the opportunistic pathogen, Bilophila wadsworthia. The susceptibility to H. hepaticus-induced colitis differed considerably between Il10(-/-) mice originating from the two institutions. This was associated with significant differences in microbiota composition, highlighting the importance of characterizing the intestinal microbiome when studying murine models of IBD.

Figure 1. Effects of H. hepaticus infection on MIT vs. MHH mice.

Intestinal inflammation induced by H. hepaticus infection in C57BL/6J Il10^−/− mice reared in MIT and MHH SPF facilities. A, histological sections (H&E stain); B, Histological sections of cecum were scored separately for inflammation, epithelial defects, edema, crypt atrophy, hyperplasia and dysplasia on a scale of 0 to 4. The scores from all the different parameters were added and represented as Typhlitis Index scores as described in Burich et al. . P values according to nonparametric Mann-Whitney test, **<0.005.

Figure 2. Rarefaction curves.

A–E, rarefaction curves for individual samples (n = 8 mice), grouped by sample set. F, rarefaction curves for combined data for each sample set (n = 5 sets), including 95% confidence intervals. Lci, lower bound of confidence interval; hci, higher bound of confidence interval. Each sample set consist of either the cecum or the colon samples for one batch of 8 mice. All curves generated after subsampling to 1227 sequences per sample.

Figure 3. Distribution of OTUs occurring in at least 4 samples of the same set.

A, cecum samples, B, colon samples.

Figure 4. Principal coordinates analysis (PCoA) of microbiota samples.

Based on Jaccard index distances measuring dissimilarity of OTU-level microbiota composition between individual samples. Axis labels including fractions of variance explained.

Figure 5. Phylum-level composition of microbiota.

A. Fraction of sequence counts in each of the individual samples, color coded by sample group. Grey boxes indicate significant differences between sample groups (unequal variances t-Test, p<0.05): a, significantly different between MIT and MHH2009 cecum samples; b, significantly different between MIT and MHH2011 cecum samples; c, significantly different between MHH2009 and MHH2011 cecum samples; d, significantly different between MIT and MHH2009 colon samples. B. Relative abundance of phyla in each of the sample groups, including standard deviation (n = 8).

Figure 6. Distribution of OTUs among samples.

Colors encode absolute OTU counts. Information on OTU classification according to RDP classifier. Numbers 1–8 are mouse identifiers. Within one mouse batch, equal numbers refer to cecum and colon samples of the same animal. Heatmap generated after subsampling to 1227 sequences per sample. The numbers on the color key correspond to untransformed OTU abundances.