Supplementary Figure 6: Single-microglial-nuclei RNA-sequencing confirms the distinct molecular signatures of cbMg and stMg.

(a) Schematic showing microglia-TRAP, microglial cell isolation by FACS and microglial nuclei isolation performed in parallel. (b) Representative flow cytometry data illustrates isolation of microglial nuclei singlets from the striatum and cerebellum of Cx3cr1^CreErt2/+;Eef1a1^{LSLeGFPL10a/+} mice. Left: x-axis: DyeCycle Ruby to quantify the number of nuclei, y-axis: GFP. Gate for singlets of nuclei is depicted. Right: x-axis: GFP signal, y-axis: Pacific Blue signal to check for autofluorescence. Gate for GFP-positive singlets. Percentage of the populations gated compared to total scatter is shown. Quantification of percentages of low DyeCycle Ruby events (indicating nuclei singlets) in the striatum (mean=82.70, SEM=3.821) and cerebellum (mean=82.47, SEM= 2.682) p=0.9551, F=2.03, t₂=0.0635. Quantification of percentages of GFP+ events (indicating microglial nuclei singlets) in the striatum (mean=3.367, SEM=0.5925) and cerebellum (mean=0.2, SEM= 0.0577) p=0.0275, F=105.3, t₂=5.903. Bar graphs with individual data points show mean ± SEM. *P ≤ 0.05, T-tests were two-tailed paired (n=3 independent experiments). (c) Linear (Pearson) correlation of RNA obtained by nuclear isolation and TRAP. Representative scatter plots of genic expression (cpm) with correlation values from cbMg and stMg (n=1 replicate). Scatter plots of genic expression were reproduced with similar results for all replicates. (d) The variation in the expression levels (z-scored log2 RPKM) of 40 selected immediate early and proinflammatory genes. Left: heatmap with hierarchical clustering distances; right: box-and-whisker plots (Tukey) for TRAP, sorted microglial nuclei, and sorted microglial cells. TRAP: min=-5.775, 25%=-2.195, median=-1.36, 75%=-0.86, max=0.127; nuclei: min=-5.97, 25%=-3.088, median=-2.458, 75%=-1.29, max=-0.5081; cells: min=2.038, 25%=3.123, median=3.707, 75%=4.97, max=7.555; p < 0.0001, F=271.3, KWS=82.65, one-way ANOVA (Kruskal-Wallis test) with Dunn’s Multiple Comparison post-hoc analysis to compare all pairs of columns (n=4/method). ***p ≤ 0.001. (e) Linear (Pearson) correlation among individual nuclei samples and between total nuclear RNA, TRAP and microglial cell RNA. Representative scatter plots of genic expression (cpm) with correlation values from cbMg and stMg (n=1 replicate). Scatter plots of genic expression were reproduced with similar results for all replicates. (f) Number of counts per gene (cpm, top) and number of genes (bottom) obtained from each of 34 stMg (purple) and 29 cbMg (orange) single nuclei after quality control filtering. (g) Modified MA plot for regional contrast provides overview of transcriptional landscape with percentage of cells expressing the gene (x-axis) and log2 (fold change) of expression (y-axis) from n=34 stMg/29 cbMg single nuclei (edgeR-QL from 4 Cx3cr1^CreErt2/+;Eef1a1^{LSLeGFPL10a/+} mice). stMg- (purple), cbMg-enriched (orange) genes are shown. Genes in green are equally enriched. Scatter plot shows -log10 p of the GO term enrichment for the stMg-DEGs (y-axis) vs cbMg-DEGs (x-axis). stMg-enriched (purple), cbMg-enriched (orange), equally enriched (green) GOs are shown. Selected GOs are named. Dotted lines: p = 0.001. (h) Scatter plot compares the log2 (fold change) by TRAP (x-axis) vs by microglial single-nuclei sequencing (y-axis) for the 297 cbMg- and 734 stMg-enriched genes identified by TRAP. Genes that are enriched in cbMg (orange) and stMg (purple) by both methods are shown. GO annotations (Enrichr) enriched for 314 stMg-enriched and 82 cbMg-enriched genes by both methods. y-axis: -log10 p. Dotted lines: p=0.05. Pie charts show the GO-based categories of 314 stMg-enriched genes (top) and 82 cbMg-enriched genes (bottom).