Transcriptional modulation of intestinal innate defense/inflammation genes by preterm infant microbiota in a humanized gnotobiotic mouse model

Abstract

Background and aims: It is known that postnatal functional maturation of the small intestine is facilitated by microbial colonization of the gut. Preterm infants exhibit defects in gut maturation, weak innate immunity against intestinal infection and increased susceptibility to inflammatory disorders, all of which may be related to the inappropriate microbial colonization of their immature intestines. The earliest microbes to colonize the preterm infant gut encounter a naïve, immature intestine. Thus this earliest microbiota potentially has the greatest opportunity to fundamentally influence intestinal development and immune function. The aim of this study was to characterize the effect of early microbial colonization on global gene expression in the distal small intestine during postnatal gut development.

Methods: Gnotobiotic mouse models with experimental colonization by early (prior to two weeks of life) intestinal microbiota from preterm human infants were utilized. Microarray analysis was used to assess global gene expression in the intestinal epithelium.

Results and conclusion: Multiple intestinal genes involved in metabolism, cell cycle regulation, cell-cell or cell-extracellular matrix communication, and immune function are developmental- and intestinal microbiota- regulated. Using a humanized gnotobiotic mouse model, we demonstrate that certain early preterm infant microbiota from prior to 2 weeks of life specifically induce increased NF-κB activation and a phenotype of increased inflammation whereas other preterm microbiota specifically induce decreased NF-κB activation. These fundamental differences correlate with altered clinical outcomes and suggest the existence of optimal early microbial communities to improve health outcomes.

Fig 1. MetaCore gene enrichment analysis of ileum microarray data representing development- and microbial colonization- effects in GF and SPF mice.

Differentially expressed genes (fold change ≥ ±2) in SPF and GF ileum from preweaned (n = 2) and postweaned (n = 2) mice were subjected to gene enrichment analysis using GeneGo MetaCore software. Panel A: Development dependent genes; top 10 overrepresented gene ontology (GO) biological processes in GF and SPF ileum (post vs preweaned). Panel B: Microbiota dependent genes; top 10 overrepresented gene ontology (GO) biological processes in pre and postweaned ileum (SPF vs. GF).

Fig 2. Comparison of bacterial communities following transfaunation of germ free mice with M_PI fecal lysates.

A: The relative abundance of different bacterial phyla in input M_PI-L,-H fecal samples and from M_PI-L, H and SPF preweaned mouse pups is shown. B: PCoA analyses of the M_PI-L and M_PI-H in original human infants, transfaunated dams and respective pups based on Bray Curtis dissimilarities among samples given presence/absence of major taxa present in at least one sample. Axis 1 explained 27.5% of variation and axis 2, 18.1%.

Fig 3. MetaCore gene enrichment analysis of ileum microarray data representing microbiota effects in preweaned M_PI-L and M_PI-H-gnotobiotic mice.

M_PI-L and M_PI-H induced differentially expressed genes were analyzed by GeneGO MetaCore software. Panel A: Top 10 overrepresented GO biological processes in preweaned M_PI-L, H vs GF mice. Panel B: Top 10 scored overrepresented networks in preweaned M_PI-L, H vs GF mice. Panel C: Comparison of the differentially expressed genes in M_PIL_GF and M_PI-H_GF presented in inflammatory response-related networks using the IPA program. Orange line: activation; Blue line: inhibition; Grey: not present.

Fig 4. Differential expression of innate immune/inflammatory response-related genes in M_PI-L,-H by qRT-PCR and immunohistochemistry.

A: Ileum RNAs were extracted from preweaned GF (n = 3), M_PI-L (n = 5) and M_PI-H (n = 5) gnotobiotic mice. M_PI-dependent gene expression levels were normalized to GAPDH and expressed as fold change over GF (designated as 1). B: The NF-κB-target gene VCAM1 was analyzed by immunohistochemical staining of 5 micron paraffin-embedded distal ileum tissue from preweaned GF, M_PI-L,-H-gnotobiotic and SPF mice (n = 3/per group) using anti- VCAM1 respectively. Representative areas are shown (magnification 400x), black arrows indicate VCAM-1 expression in ileal epithelium and endothelial cells. C: MCP-1 expression was analyzed by immunohistochemical staining of 5 micron paraffin-embedded distal ileum tissue from 3wk-old GF,M_PI-L,-H and SPF mice (n = 3/per group) using anti- MCP-1 respectively. Representative areas (n = 3/per group) are shown (magnification 400x), black arrows indicate MCP-1 expression in ileal epithelium.

Fig 5. M_PI effects on systemic immune and inflammatory cytokine expressions by multiplex immunoassay.

Multiplex immunoassay of immune and inflammatory cytokine expressions in serum collected from preweaned GF and M_PI-L, H and SPF. The levels of these cytokines are presented as mean ± SEM. One-way ANOVA with post-hoc Tukey’s honest significance (HSD) test was used to compare the groups.

Fig 6. NF-κB activation in M_PI-L, -H and SPF GF ileum mucosa by immunohistochemistry.

A: NF-κB activation was analyzed by immunohistochemical staining of 5 micron paraffin-embedded distal ileum tissue from GF, M_PI-L, H and SPF mouse. Representative areas (n = 3/per group) are shown (magnification 400x), black arrows indicate nuclear translocation of the phosphorylated NF-κB p65 subunit. B: Nuclear translocation of pp65 was quantified by ImmunoRatio tool from Image J software and presented as % of pp65 positive nuclei/total nuclei. C: Growth curve of M_PI-L (n = 20) and M_PI-H (n = 15) from birth to weaning. One-way ANOVA with post-hoc Tukey’s HSD test was used to compare the groups.