Profiling Microglia From Alzheimer's Disease Donors and Non-demented Elderly in Acute Human Postmortem Cortical Tissue

Abstract

Microglia are the tissue-resident macrophages of the central nervous system (CNS). Recent studies based on bulk and single-cell RNA sequencing in mice indicate high relevance of microglia with respect to risk genes and neuro-inflammation in Alzheimer's disease (AD). Here, we investigated microglia transcriptomes at bulk and single-cell levels in non-demented elderly and AD donors using acute human postmortem cortical brain samples. We identified seven human microglial subpopulations with heterogeneity in gene expression. Notably, gene expression profiles and subcluster composition of microglia did not differ between AD donors and non-demented elderly in bulk RNA sequencing nor in single-cell sequencing.

Keywords: Alzheimer’s disease; barcoded Smart-seq2; human; microglia; single-cell RNA sequencing.

Figure 1

Microglia gene expression profiling of four groups of donors from acute postmortem human brain tissue. (A) Tissue samples from GFS and LPS were classified into four experimental groups (CTR, CTR+, AD, MCI) based on Aβ and PHF-tau immunohistological stainings and the clinical report of the donor (scale bar = 50 μm). Microglia were mechanically isolated and collected by CD11B+CD45+ -based FACS sorting. Sorted microglia were used for bulk or single-cell sequencing (barcoded Smartseq2 and 10× Genomics sequencing techniques). (B) Heat map depicting log₂ CPM expression of cell type-specific markers in bulk-sorted microglia, as previously published in different data sets (Lake et al., ; Galatro et al., ; Zhong et al., 2018). Abbreviations: AD, clinically diagnosed Alzheimer’s Disease; AST, astrocytes; Aβ, amyloid beta; CPM, counts per million; CTR, Control; CTR+, Control with Aβ plaques and/or hyperphosphorylated tau; END, endothelial cells; EX_NEU, excitatory neurons; GFS, superior frontal gyrus; GRA, granulocytes; IN_NEU, inhibitory neurons; LPS, superior parietal lobe; MCI, mild cognitive impairment; MIC, microglia; NPC, Neural Progenitor Cells; OLI, oligodendrocytes; OPC, Oligodendrocyte Precursor Cells; PER, pericytes; PHF-tau, hyperphosphorylated tau; scRNAseq, single-cell RNA sequencing.

Figure 2

Transcriptomic analysis of microglia populations isolated from CTR, CTR+, and AD donors. (A) Principal component analysis (PCA) of RNA-sequencing data from acutely isolated microglia illustrating the effect of donor (each sample is indicated with a donor label), donor groups, sex, brain region, and age. (B) Differentially expressed genes between AD and CTR+ donors (likelihood ratio test, SIGLEC1 ***FDR = 0.002, CXCL10 *FDR = 0.02 and CXCR2 *FDR = 0.02). (C) Selected examples of expression levels (log₂ CPM) of disease-associated microglia genes in bulk microglia samples. (D) Heat map depicting expression of most abundant disease-associated microglia genes in bulk microglia samples from three donor groups in row z scores of log₂ CPM values. Donors are ordered by age (young to old) within the donor group. Abbreviations: AD, Alzheimer’s Disease; CPM, counts per million; CTR, Control; CTR+, Control with Aβ plaques and/or hyperphosphorylated tau; F, female; M, male; GFS, superior frontal gyrus; LPS, superior parietal lobe.

Figure 3

Single-cell expression profiling of microglia identified seven subsets of microglia but no AD-associated cluster using bc-Smart-seq2. (A) UMAP visualization and unsupervised clustering of microglia derived from AD (n = 4), MCI (n = 1), CTR+ (n = 5), and CTR (n = 4) donors. Colors represent cluster identity. (B) Cluster contribution normalized to donor. (C) Expression of cell type–specific markers for neurons (RBFOX3), oligodendrocytes (MOG), astrocytes (GFAP), circulating monocytes (CCR2), erythrocytes (HBA1), and microglia (ITGAX). (D) Dot plot visualizing scaled average expression and the percentage of cells expressing the indicated genes. (E) Average expression of the mouse DAM (Keren-Shaul et al., 2017), myeloid neurodegeneration-related (Friedman et al., 2018), and human AD-related single-nucleus (Mathys et al., 2019) gene sets per cell across clusters (multiple linear regression with Bonferroni correction, *indicate p-value < 0.001). Horizontal lines in the boxplots represent the mean; the lower and upper hinges represent the 25th and 75th percentiles. Abbreviations: AD, Alzheimer’s Disease; avg, average; CPM, counts per million; CTR, Control; CTR+, Control with Aβ plaques and/or hyperphosphorylated tau; DAM, disease-associated microglia; MCI, mild cognitive impairment; UMAP, uniform manifold approximation and projection.

Figure 4

Single-cell expression profiling of microglia from individual donors identified three clusters but no DAM-like cluster using 10× Genomics. (A) UMAP visualization and unsupervised clustering of microglia from AD donor 2018-135. (B–D) Average expression of the mouse DAM (B), myeloid neurodegeneration-related (C), and human AD-related single-nucleus (D) gene set expression per cell across clusters for AD donor 2018-135 (multiple linear regression with Bonferroni correction, *indicate p-value < 0.001). (E) UMAP visualization and unsupervised clustering of MCI donor 2019-010. (F–H) Average expression of the mouse DAM (F), myeloid neurodegeneration-related (G) and human AD-related single-nucleus (H) gene set expression per cell across clusters for MCI donor 2019-010. Horizontal lines in the boxplots represent the mean; the lower and upper hinges represent the 25th and 75th percentiles. Abbreviations: AD, Alzheimer’s Disease; avg, average; CPM, counts per million; CTR, Control; CTR+, Control with Aβ plaques and/or hyperphosphorylated tau; DAM, disease-associated microglia; MCI, mild cognitive impairment; UMAP, uniform manifold approximation and projection.