Beyond Quiescent and Active: Intermediate Microglial Transcriptomic States in a Mouse Model of Down Syndrome

Abstract

Research on microglia in Down syndrome (DS) has shown that microglial activation, increased inflammatory gene expression, and oxidative stress occur at different ages in DS brains. However, most studies resulted in simplistic definitions of microglia as quiescent or active, ignoring potential intermediate states. Indeed, recent work on microglial cells in young DS brains indicated that those evolve through different intermediate activation phenotypes before reaching a fully activated state. Here we used single nucleus RNA sequencing, to examine how trisomy affects microglial states in the Ts65Dn mouse model of DS. Despite no substantial changes in the proportion of glial populations, differential expression analysis revealed cell type-specific gene expression changes, most notably in astroglia, microglia, and oligodendroglia. Focusing on microglia, we identified differential expression of genes associated with different microglial states, including disease-associated microglia (DAMs), activated response microglia (ARMs), and human Alzheimer's disease microglia (HAMs), in trisomic microglia. Furthermore, pseudotime analysis reveals a unique reactivity profile in Ts65Dn microglia, with fewer in a homeostatic state and more in an intermediate aberrantly reactive state than in euploid microglia. This comprehensive understanding of microglial transcriptional dynamics sheds light on potential pathogenetic mechanisms but also possible avenues for therapy for neurodevelopmental disorders.

Keywords: Down syndrome; disease-associated microglia; microglia.

Figure 1

Single-nucleus sequencing and cell-type identification. (A) Overview of the experimental approach with a schematic showing the NeuN-negative cell sorting (see Section 4). Hippocampus of WT and Ts65Dn were collected and the nuclei suspension was prepared using enzymatic digestion and mechanical dissociation. Nuclei were incubated with anti-NeuN-Alexa647 and NeuN-negative nuclei were selected by FANS and subsequently sequenced using 10x sequencing. (B) All hippocampus single nuclei embedded in UMAP, displaying cell clusters in different colors. Each dot represents a single nucleus. (C) Mapping of microglia (Cx3cr1), astroglia (Aqp4) and oligodendroglia (Mbp) markers. (D) Dot plot showing enrichment of canonical markers of microglia, oligodendroglia, astroglia, endothelial cells, and neurons. (E) Proportion of the main cell populations both in WT and Ts65Dn hippocampus. FANS: fluorescent activated nuclear sorting.

Figure 2

Transcriptionally different cell populations within the Ts65Dn hippocampus. (A) Distribution of DEGs along the mouse chromosomes. (B) Volcano plot of DEGs colored by each major glial type. (C) Strip plot showing the average log fold change of the different DEGs colored by glial cell type. (D) Biological pathways enriched for DEGs identified in each major glial cell type. DEGs differentially expressed genes.

Figure 3

Transcriptional shift in trisomic microglial cells. (A) UMAP plot showing the microglia cluster both in WT (blue) and trisomic (red) mice. (B) UMAP plot showing the different clusters of microglia both in Ts65Dn (left) and WT mice (right). (C) Mapping of the Aif marker (encoding for Iba1) both in Ts65Dn (left) and WT (right) mice. (D) Pseudotime showing the relative positioning of cells along the trajectory quantifying the relative reactivity of microglial cells. (E) Box plot showing the pseudotime score of different microglial cells according to the top 50 DEG genes of both the homeostatic microglia cluster and reactive microglial cluster in Ts65Dn (red and WT mice (blue) mice. p-value = 2.26 × 10⁻¹⁶. (F) Relative abundance or density of microglial cells alongside the different pseudotime scores in both Ts65Dn (red) and WT (blue) hippocampus. (G) Venn diagram showing overlap between the DEGs in microglia and DAM signature. DEGs differentially expressed genes, DAM disease-associated microglia.