Synergistic effects of APOE and sex on the gut microbiome of young EFAD transgenic mice
Alzheimer’s disease (AD) is a fatal neurodegenerative disease. APOE4 is the greatest genetic risk factor for AD, increasing risk up to 15-fold compared to the common APOE3. Importantly, female (♀) APOE4 carriers have a greater risk for developing AD and an increased rate of cognitive decline compared to male (♂) APOE4 carriers. While recent evidence demonstrates that AD, APOE genotype, and sex affect the gut microbiome (GM), how APOE genotype and sex interact to affect the GM in AD remains unknown.
This study analyzes the GM of 4-month (4 M) ♂ and ♀ E3FAD and E4FAD mice, transgenic mice that overproduce amyloid-β 42 (Aβ42) and express human APOE3+/+ or APOE4+/+. Fecal microbiotas were analyzed using high-throughput sequencing of 16S ribosomal RNA gene amplicons and clustered into operational taxonomic units (OTU). Microbial diversity of the EFAD GM was compared across APOE, sex and stratified by APOE + sex, resulting in 4-cohorts (♂E3FAD, ♀E3FAD, ♂E4FAD and ♀E4FAD). Permutational multivariate analysis of variance (PERMANOVA) evaluated differences in bacterial communities between cohorts and the effects of APOE + sex. Mann-Whitney tests and machine-learning algorithms identified differentially abundant taxa associated with APOE + sex.
Significant differences in the EFAD GM were associated with APOE genotype and sex. Stratification by APOE + sex revealed that APOE-associated differences were exhibited in ♂EFAD and ♀EFAD mice, and sex-associated differences were exhibited in E3FAD and E4FAD mice. Specifically, the relative abundance of bacteria from the genera Prevotella and Ruminococcus was significantly higher in ♀E4FAD compared to ♀E3FAD, while the relative abundance of Sutterella was significantly higher in ♂E4FAD compared to ♂E3FAD. Based on 29 OTUs identified by the machine-learning algorithms, heatmap analysis revealed significant clustering of ♀E4FAD separate from other cohorts.
The results demonstrate that the 4 M EFAD GM is modulated by APOE + sex. Importantly, the effect of APOE4 on the EFAD GM is modulated by sex, a pattern similar to the greater AD pathology associated with ♀E4FAD. While this study demonstrates the importance of interactive effects of APOE + sex on the GM in young AD transgenic mice, changes associated with the development of pathology remain to be defined.
KeywordsAlzheimer’s disease Gut microbiome APOE genotype Sex
Operational taxonomic units
Permutational multivariate analysis of variance
Short chain fatty acid
The gut microbiome (GM), the collective genome of gastrointestinal bacteria, is an integral component of human physiology [1, 2, 3, 4, 5]. Recent studies link dysbiotic GM profiles with neurological disorders, with multiple sclerosis the first identified [6, 7, 8, 9, 10, 11, 12]. While subsequent studies have linked dysbiosis with Alzheimer’s disease (AD) pathology [13, 14, 15, 16, 17, 18, 19, 20, 21, 22], the effects of AD risk factors, specifically APOE genotype, sex and their interaction, on the GM remain unclear.
The APOE4 genotype is the greatest genetic risk factor for AD, increasing risk up to 15-fold compared to the more common APOE3 genotype [23, 24]. Apolipoprotein E (apoE) is a member of the apolipoprotein family, the protein components of lipoproteins. Both humans and AD transgenic (−Tg) mice with APOE4 exhibit an increase in amyloid-β (Aβ) peptide accumulation, both as amyloid plaques, a hallmark of the disease, and small soluble aggregates. Thus, one explanation for the APOE4-associated AD risk is a loss of function in Aβ clearance. Tran and colleagues demonstrated significant differences between the GM of human APOE3 and APOE4 carriers, as well as differences between the GM of APOE3 and APOE4 targeted replacement (−TR) mice . These differences were attributed to a loss of apoE4 function in lipid homeostasis, as APOE4 is associated with higher levels of cholesterol, triglycerides and low-density lipoproteins compared to APOE3 [26, 27, 28, 29], changes that significantly affect the GM [30, 31, 32, 33, 34, 35, 36, 37]. Sex is another risk factor for AD as females (♀) exhibit almost two-fold greater lifetime AD risk compared to males (♂) . Additionally, sex plays an important role in the GM as the bacterial composition and metabolic function differ significantly between ♂ and ♀ [37, 39, 40, 41, 42, 43, 44, 45, 46]. Importantly, ♀APOE4 carriers have a greater lifetime risk for developing AD, an increased rate of cognitive decline and an accelerated accumulation of Aβ compared to ♂APOE4 carriers [47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61]. While the underlying mechanism is unclear, evidence suggests this interaction modulates the GM.
EFAD-Tg mice  overexpress Aβ42 via five familial AD (FAD) mutations  and express h-APOE3 or APOE4, allowing for the study of the interaction among AD risk factors [64, 65, 66]. EFAD mice expressing the APOE4+/+ genotype (E4FAD), compared to E3FAD mice, exhibit increased behavioral deficits, Aβ deposition and neuroinflammation. Importantly, these differences are reproduced in ♀ vs ♂EFAD mice, resulting in 4 pathologically-distinct cohorts when the EFAD mice are stratified by APOE + sex (♀E4FAD > ♂E4FAD = ♀E3FAD > ♂E3FAD), a phenotype that develops with age [65, 66]. For this study, we focused on 4 M EFAD mice to evaluate the interactive effects of APOE + sex on the GM at an age prior to, or early in, the development of pathology. Microbial analysis of fecal samples demonstrated that APOE + sex have a significant effect on the GM at various taxonomic levels.
As previously described, the EFAD (5xFAD+/−/APOE+/+) mice are homozygous for APOE2, APOE3, or APOE4 and heterozygous for the 5x familial AD (5xFAD) mutations [62, 63]. Although APOE2 is considered neuroprotective, 100% of APOE2+/+ mice have type III hyperlipoproteinemia, compared to only 15% of human ε2/2 carriers [67, 68, 69]; thus, E2FAD mice were excluded from the current study. At 4 M, fecal samples were obtained from the 4 cohorts (9 ♂E3FAD, 8 ♂E4FAD, 19 ♀E3FAD, 12 ♀E4FAD) by individually placing mice in clean disposable Styrofoam cups. Feces were flash frozen and stored at − 80 °C until DNA isolation.
Fecal DNA was isolated using a PowerSoil DNA isolation kit (Mo Bio Laboratories) and DNA concentrations determined by UV absorbance (Nanodrop, ThermoFisher). The V4 variable region of 16S ribosomal RNA gene was PCR-amplified using target-specific primers containing bar codes and linker sequences . PCR reaction conditions included an initial denaturation step of 30 s (s) at 98 °C, followed by 28 cycles of 10s at 98 °C, 15 s at 60 °C, 30s at 72 °C, and a final elongation step of 7 min at 72 °C. The PCR master mix (20 μl volume) contained 100 ng of DNA template, 0.5 μM forward and reverse primers, Phusion Hot Start DNA polymerase and high-fidelity buffer (New England Biolabs), dNTPs and sterile water. Results were checked by polyacrylamide gel electrophoresis and samples pooled in equimolar ratio. The samples were sequenced on an Illumina MiSeq sequencer at the University of Kentucky Advanced Genetic Technologies Center, with sequence merging, trimming, chimera removal, clustering and annotation performed using the software package QIIME . The Greengenes database was implemented for Operational Taxonomic Unit (OTU) annotation at a threshold of 97% sequence similarity . To avoid effects of uneven sequencing depth , datasets were rarified to 3000 sequences/sample prior to analysis. For statistical analyses, OTUs with a frequency below 0.1% across the dataset were removed .
The Shannon H α-diversity index was used to assess bacterial richness and evenness. The interaction between APOE + sex in α-diversity measures was evaluated using a mixed effects model, similar to a two-way analysis of variance (ANOVA), that analyzes repeated measures with missing values. This analysis was performed in the software package GraphPad Prism (version 8.2.0). For β-diversity, permutational ANOVA (PERMANOVA) was used to compare microbial community structure within and among the EFAD cohorts based on Bray-Curtis dissimilarity [74, 75]. Pair-wise PERMANOVA was used to assess the effect of the interaction among universal biological variables on the microbiome composition . Principal coordinate analysis plots (PCoA; Bray-Curtis distances) with 95% confidence ellipses were used to visualize microbial communities [75, 77, 78]. The Mann-Whitney U (MWU) test under the Monte Carlo simulation, corrected with Benjamini-Hochberg False Discovery Rate (p < 0.05), was used to identify differentially abundant taxa associated with APOE + sex at the taxonomic level of genus. The Random Forest based Boruta algorithm was used to determine OTUs significant in distinguishing samples by APOE + sex compared to randomly generated probes or “shadow scores” . Heatmaps were generated using the R package, “pheatmaps”, calculating the Euclidean distance among cohorts.
Results and discussion
Mouse fecal microbial community structure was analyzed using high-throughput sequencing of 16S rRNA gene amplicons, followed by sequence clustering (97% similarity) into a total of 2063 OTUs. No significant difference in α-diversity (Shannon H index) was observed between E3FAD and E4FAD mice (p = 0.975; Additional file 1: Figure S1A) or between ♂EFAD and ♀EFAD (p = 0.949; Additional file 1: Figure S1B). In comparing across cohorts stratified by APOE + sex, Shannon H indices were significantly higher in ♂E4FAD and ♀E3FAD, compared to ♂E3FAD and ♀E4FAD (p < 0.05; Additional file 1: Figure S1C). Additionally, the interaction of APOE + sex significantly modulated α-diversity measures (p < 0.05; Additional file 1: Figure S1C), suggesting that analyses by APOE genotype or sex alone will mask effects on microbial community structure.
Compared to ♀E3FAD mice, ♀E4FAD mice exhibited a lower relative abundance of bacterial genera associated with short chain fatty acid (SCFA) production, including Prevotella and Ruminococcus [84, 85, 86, 87, 88, 89]. The GM is crucial for the production of SCFAs that, while the underlying mechanism is not completely understood, serve as energy sources for intestinal epithelial cells, regulators of plasma lipid levels, and modulators of immune cells [90, 91, 92, 93, 94, 95]. The current results suggest a metabolic dysfunction in the ♀E4FAD GM. However, metabolomic and metagenomic analyses will be required to interpret accurately the interactive effects of APOE + sex on the metabolic function of the EFAD GM.
This short report demonstrates: 1) the EFAD GM is modulated by APOE + sex, 2) the synergistic effects of ♀sex and APOE4 genotype yield a specific GM profile in ♀E4FAD mice, and 3) clustering samples by only APOE genotype or sex masks the interactive effects of APOE + sex on the EFAD GM. Notably, these findings are consistent with AD readouts from EFAD mice varying in severity of pathology by APOE + sex, including behavioral deficits, Aβ deposition and neuroinflammation greatest in ♀E4FAD > ♂E4FAD = ♀E3FAD > ♂E3FAD [65, 66]. Therefore, the GM would potentially serve as an AD readout, reflecting the interaction between APOE + sex. Although the use of 16S rRNA sequencing has more limited taxonomic resolution than shotgun metagenome sequencing , 16S rRNA sequencing is sufficiently robust to identify significant effects on the GM. This study demonstrates the importance of stratifying the EFAD population by APOE + sex to better understand the relationship between AD and the GM. Future studies will examine the composition and metabolic function of the GM throughout the development of EFAD pathology through the use of metagenomic and metabolomic analyses. In conclusion, the interactive effects of APOE + sex on AD play an important role in modulating the GM composition, and the current report is the first step in identifying and understanding these effects.
ML and JMW analyzed the data and wrote the manuscript, IP and SE purified DNA and performed PCR, JMW and SJG performed statistical analyses, JY bred the mice, SJG and SE contributed to data analysis and manuscript editing. All authors read and approved the final manuscript.
The authors acknowledge funding from the NIH (R56-AG057589 and P01-AG030128) and the BrightFocus Foundation (A2014210S) to the Estus lab. The LaDu lab funding includes funds from NIH/NIA (R01 AG058068, R01 AG057008), institutional funds from the College of Medicine at the University of Illinois, Chicago, and philanthropic contributions. JMW was supported by an administrative supplement to R01 AG057008 and by the NSF Bridge to the Doctorate Fellowship.
Animal studies were performed in compliance with IACUC (Institutional Animal Care and Use Committee) at University of Illinois-Chicago.
Consent for publication
All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
- 16.Fox M, Knorr DA, Haptonstall KM. Alzheimer's disease and symbiotic microbiota: an evolutionary medicine perspective. Ann N Y Acad Sci. 2019;1449(1):3–24.Google Scholar
- 47.Farrer LA, Cupples LA, Haines JL. Hyman B, Kukull WA, Mayeux R, et al. effects of age, sex, and ethnicity on the association between apolipoprotein E genotype and Alzheimer disease. A meta-analysis. APOE and Alzheimer disease Meta analysis consortium. JAMA. 1997;278(16):1349–56.PubMedCrossRefPubMedCentralGoogle Scholar
- 54.Molero AE, Pino-Ramirez G, Maestre GE. Modulation by age and gender of risk for Alzheimer's disease and vascular dementia associated with the apolipoprotein E-epsilon4 allele in Latin Americans: findings from the Maracaibo aging study. Neurosci Lett. 2001;307(1):5–8.PubMedCrossRefPubMedCentralGoogle Scholar
- 63.Oakley H, Cole SL, Logan S, Maus E, Shao P, Craft J, et al. Intraneuronal beta-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer's disease mutations: potential factors in amyloid plaque formation. J Neurosci. 2006;26(40):10129–40.PubMedPubMedCentralCrossRefGoogle Scholar
- 64.Tai LM, Ghura S, Koster KP, Liakaite V, Maienschein-Cline M, Kanabar P, et al. APOE-modulated Abeta-induced neuroinflammation in Alzheimer's disease: current landscape, novel data and future perspective. J Neurochem. 2015;133(4):465–88.Google Scholar
- 66.Balu D, Karstens AJ, Loukenas E, Maldonado Weng J, Valencia-Olvera AC, LaDu MJ. The role of APOE in transgenic mouse models of AD. Neurosci Lett. 2019;134285. https://doi.org/10.1016/j.neulet.2019.134285. Available from: http://www.sciencedirect.com/science/article/pii/S030439401930357X PubMedCrossRefPubMedCentralGoogle Scholar
- 74.Clarke KR, Warwick RM. Change in Marine Communities: An Approach to Statistical Analysis and Interpretation. Plymouth: Primer-E Ltd; 2001.Google Scholar
- 78.Hammer Ø, Harper DAT, Ryan PD. Past: Paleontological Statistics Software Package for Education and Data Analysis. Palaeontol Electron. 2001;4(1):9.Google Scholar
- 95.Dopkins N, Nagarkatti PS, Nagarkatti M. The role of gut microbiome and associated metabolome in the regulation of neuroinflammation in multiple sclerosis and its implications in attenuating chronic inflammation in other inflammatory and autoimmune disorders. Immunology. 2018;154(2):178–85.PubMedPubMedCentralCrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.