Altered metabolism and DAM-signatures in female brains and microglia with aging

Abstract

Despite Alzheimer's disease (AD) disproportionately affecting women, the mechanisms remain elusive. In AD, microglia undergo 'metabolic reprogramming', which contributes to microglial dysfunction and AD pathology. However, how sex and age contribute to metabolic reprogramming in microglia is understudied. Here, we use metabolic imaging, transcriptomics, and metabolic assays to probe age- and sex-associated changes in brain and microglial metabolism. Glycolytic and oxidative metabolism in the whole brain was determined using Fluorescence Lifetime Imaging Microscopy (FLIM). Young female brains appeared less glycolytic than male brains, but with aging, the female brain became 'male-like.' Transcriptomic analysis revealed increased expression of disease-associated microglia (DAM) genes (e.g., ApoE, Trem2, LPL), and genes involved in glycolysis and oxidative metabolism in microglia from aged females compared to males. To determine whether estrogen can alter the expression of these genes, BV-2 microglia-like cell lines, which abundantly express DAM genes, were supplemented with 17β-estradiol (E2). E2 supplementation resulted in reduced expression of DAM genes, reduced lipid and cholesterol transport, and substrate-dependent changes in glycolysis and oxidative metabolism. Consistent with the notion that E2 may suppress DAM-associated factors, LPL activity was elevated in the brains of aged female mice. Similarly, DAM gene and protein expression was higher in monocyte-derived microglia-like (MDMi) cells derived from middle-aged females compared to age-matched males and was responsive to E2 supplementation. FLIM analysis of MDMi from young and middle-aged females revealed reduced oxidative metabolism and FAD+ with age. Overall, our findings show that altered metabolism defines age-associated changes in female microglia and suggest that estrogen may inhibit the expression and activity of DAM-associated factors, which may contribute to increased AD risk, especially in post-menopausal women.

Keywords: Aging; Alzheimer’s disease; Lipids; Lipoprotein lipase; Metabolism; Microglia; Sex differences.

Figure 1.. Fluorescence Lifetime Imaging Microscopy (FLIM) of young and old female brains.

A. Schematic representation of model and sample preparation. The M1 region of the cortex was imaged using FLIM for all brain sections. B. Schematic representation of FLIM. C. Glycolytic indices of young (16-week) and old (20-month), female and male brain cortices showing increased glycolytic index with aging in females, with equation for calculations of glycolytic index. D. Fluorescence Lifetime Imaging Redox Ratio (FLIRR) of young and old, female and male brain cortices indicating less oxidative metabolism in the cortex of females with aging but more in young females than male counterparts with the equation for FLIRR. No significant effects of aging were observed in males. Data points reflect biological and technical replicates. (* p < 0.05; *** p < 0.001; **** p < 0.0001) Created with BioRender.

Figure 2.. Differential gene expression of CD11b+ microglia isolated from 16-week and 20-month males and females reveals elevated metabolism and DAM signatures in aged females.

A. Volcano plot comparing expression between 16-week males and females with significantly changed genes in red. B. Volcano plot comparing expression in 20-month males and females with significantly changed genes in red. C. Bubble enrichment plot showing differentially altered pathways in 16-week males and females. D. Bubble enrichment plot showing differentially altered pathways in 20-month males and females. E. Heatmap comparing specific disease-associated microglia (DAM), immunometabolic, cholesterol transport, glycolysis, TCA cycle, oxidative phosphorylation (OXPHOS), and mitochondrial genes in 16-week and 20-month females and males. Here, a decrease in enrichment is a value between 0–1.

Figure 3.. Estrogen acts via ESR1 to alter the expression of DAM genes and lipid metabolism.

A. Schematic of experiments carried out in BV-2 microglia using ESR1 and ESR2 -specific agonists 4,4′,4″-(4-Propyl-[1H]-pyrazole-1,3,5-triyl) trisphenol (PPT) and diarylpropionitrile (DPN) respectively. B. qPCR gene expression data of TNF-α, IL-1β, LPL, Trem2, and ApoE in BV-2 microglia-like cells after 24 hours of drug exposure (estradiol (E2), or vehicle). Individual data points represent biological replicates. C. Lipoprotein lipase (LPL) activity in BV-2 microglia-like cells grown in either 1% FBS or 10% FBS exposed to E2 for 24 hours. Data points reflect biological and technical replicates. D. Cholesterol efflux from BV-2 cells exposed to E2, PPT or DPN 24 hours. Data points reflect biological and technical replicates. (# p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001) Created with BioRender.

Figure 4.. Lipoprotein Lipase (LPL) and Total Lipase Activity in aged Wild Type and 5XFAD Mice brains.

A. Extracellular and intracellular LPL activity in whole brain tissue of 9-month-old wild type (WT) and 5XFAD male and female mice. (* p < 0.05; ** p < 0.01; **** p < 0.0001).

Figure 5.. Metabolic effects of Estrogen on BV-2 microglia.

A. Glycolytic index of BV-2 microglia in 1% or 10% FBS containing media treated with equal volumes of media (control), ethanol (vehicle), or E2 dissolved in ethanol (E2). B. FLIRR of BV-2 microglia in 1% or 10% containing media treated with equal volumes of media (control), ethanol (vehicle), or E2 dissolved in ethanol (E2). C. Heatmap of glycolytic intermediates from metabolomics data. D. Heatmap of TCA cycle intermediates from metabolomics data.

Figure 6.. Metabolic effects of Estrogen on Human Microglia.

A. Representative image of monocyte-derived microglia-like (MDMi) cells stained with DAPI (blue) and Iba1(red). D. Normalized expression data from MDMi cells from an aged male and female donor, with female MDMi cells treated with E2 for 48 hours. C. Representative brightfield image of MDMi before FLIM analysis. D. FLIM analysis of MDMi derived from 25 yo female or 55 yo female donors after treatment with vehicle (0.000001% Ethanol) or E2 for 48 hours. E. Proteomic comparison of media proteins from 55 yo MDMi cultures, compared to 25 yo MDMi cultures. Pathways that are up-regulated with age. F. Proteomic comparison of media proteins from 55 yo MDMi cultures, compared to 55 yo MDMi cultures supplemented with E5 for 48 hours. Fold enrichment of pathways that are down-regulated following E2 supplementation. Individual data points reflect biological and technical replicates. (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001).