The Effect of Growth Restriction on the Gut Microbiome in Mice

INTRODUCTION Early life growth restriction increases susceptibility to non‐communicable adult‐onset diseases. However, there is minimal mechanistic rationale for elevated disease risk. The gut microbiome assists in regulating host physiological processes, with emerging evidence suggesting microbiome dysregulation leads to increased risk of chronic disease. Thus, the purpose of this investigation was to determine differences in the gut microbiome of growth restricted mice as compared to non‐restricted mice. METHODS A cross‐fostering, protein‐restricted (8% protein) nutritive model in FVB mice was used to induce growth restriction during gestation (GUN; N=10 male,15 female) or lactation from days 1–21 of life (PUN; N=11 male, 12 female) along with a control group (CON, 20% protein, N=12 male, 11 female). At 21 postnatal days of age (PN21) all mice were weaned to a non‐restricted diet (20% protein), isolating undernutrition to early life windows. Fecal samples were collected weekly across the lifespan (until PN80) to determine longitudinal programming effects of growth restriction on the gut microbiome. Cecum samples were collected at PN21 and PN80. Microbiome analysis of fecal and cecum samples were conducted by first extracting the DNA using Qiagen Powersoil DNA extraction kits and conducting PCR amplification of the bacterial 16S rRNA gene amplicon profiles through the Qiita bioinformatics platform. Phenotypic growth markers of all three diet groups (CON, GUN, PUN) were analyzed using repeated measures ANOVA. Metabolomics data on the same samples is currently being analyzed and will be integrated with the microbiome dataset. RESULTS The fecal microbiome was strongly separated by treatment group using the Bray‐Curtis distance measure (PERMANOVA p = 0.001). A random forest machine learning classification validated this separation showing an overall out‐of‐bag error rate of 18.6% across the three groups in the whole dataset. The PUN group was distinctly separated from GUN and CON, with a group error rate of only 15.0%. Differences in the microbiome were not evident through analysis at the Phylum level (Firmicutes/Bacteroidetes ratio) but was instead driven by differential abundances of specific rare taxa. Bifidobacterium was significantly elevated in the PUN group (Kruskal‐Wallis test p=0.00017), beginning as early as PN21 and remaining elevated in the PUN group through the lifespan (PN80). CONCLUSIONS A protein‐restricted nutritive model induces strong imprinting on the structure of the murine microbiome. Changes in the PUN group are primarily driven by differences in low abundance taxa. Understanding the mechanisms that lead to the microbiome dysbiosis shown here could lead to the development of therapeutic interventions to limit the development of a stunted or immature gut microbiota. Support or Funding Information MSU Department of Biochemistry and Molecular Biology; MSU Department of Kinesiology