Murine Gut Microbiome Association With APOE Alleles

Abstract

Background: Since APOE alleles represent the most impactful genetic risk factors for Alzheimer's disease (AD), their differential mechanism(s) of action are under intense scrutiny. APOE4 is robustly associated with increased AD risk compared to the neutral APOE3 and protective APOE2. APOE alleles have also been associated with differential inflammation and gastrointestinal recovery after insult in human and murine studies, leading us to hypothesize that APOE alleles impact the gut microbiome. Methods: To assess this hypothesis, we compared 16S ribosomal RNA gene amplicon-based microbiome profiles in a cohort of mice that were homozygous for APOE2, APOE3, or APOE4, and included both males and females as well as carriers and non-carriers of five familial AD (5xFAD) mutations. Fecal samples were analyzed from mice at 4 and 6 months of age. APOE genotype, as well as sex and 5xFAD status, was then tested for influence on alpha diversity (Shannon H index) and beta diversity (principal coordinate analyses and PERMANOVA). A Random Forest analysis was used to identify features that predicted APOE, sex and 5xFAD status. Results: The richness and evenness (alpha diversity) of the fecal microbiome was not robustly associated with APOE genotype, 5xFAD status or sex. In contrast, microbial community composition (beta-diversity) was consistently and strongly associated with APOE genotype. The association between beta-diversity and sex or 5xFAD status was less consistent and more modest. Comparison of the differences underlying APOE effects showed that the relative abundance of multiple bacterial taxa was significantly different as a function of APOE genotype. Conclusions: The structure of the gut microbiome was strongly and significantly associated with APOE alleles in this murine model. Further evaluation of these findings in humans, as well as studies evaluating the impact of the APOE-associated microbiota on AD-relevant phenotypes in murine models, will be necessary to determine if alterations in the gut microbiome represent a novel mechanism whereby APOE genotype impacts AD.

Keywords: APOE; Alzheimer's; cladogram; microbiome; resistant starch.

Figure 1

Alpha-diversity at the ASV level as a function of APOE, sex and 5XFAD status at 4 (A) and 6 (B) months of age as assessed using the Shannon H index. These samples from 4 to 6 months were rarified to 5,364 and 4,020 ASVs, respectively.

Figure 2

PCoA of fecal microbiome profiles in 4-month-old (A–C) and 6-month-old (D–F) mice shows a separation as a function of APOE relative to sex or 5xFAD status. Ellipses represent 95% confidence.

Figure 3

APOE genotype is associated with multiple bacteria. A LefSe analysis shows that the effects of APOE genotype are fairly broad across the microbiome (A). Quantification of representative bacterial families with significant differences with APOE in mice at both 4 and 6 months of age are shown (B,C). The p-values have been corrected by using an FDR approach.

Figure 4

Random Forest analyses find that microbiome features accurately predict APOE genotype relative to sex or 5xFAD status (box plots). Features that predict APOE genotype, sex and 5xFAD status are shown the features plots below. Note the differences in the abscissa scale for the graphs of APOE vs. sex and 5xFAD status, which reflects the greater prediction accuracy of the bacteria for APOE genotype.