Immune checkpoint molecule Tim-3 regulates microglial function and the development of Alzheimer’s disease pathology

Analysis of bulk RNA-seq data (📁bulkRNAseq)

• Functions used in the analysis: 0.bulkRNAseq_functions.R
• In-house generated datasets
  • Dataset 1 Batch 1 (3-month-old mice) (1.bulkRNAseq_dataset1_batch1_3month.R): control phagocytosing vs. Havcr2^cKO phagocytosing vs. control non-phagocytosing vs. Havcr2^cKO non-phagocytosing microglia
  • Dataset 1 Batch 2 (7-month-old mice) (2.bulkRNAseq_dataset1_batch2_7month.R): control vs. Havcr2^cKO vs. 5xFAD vs. Havcr2^cKO;5xFAD microglia
  • Dataset 2 (1-month-old mice) (3.bulkRNAseq_dataset2_1month.R): Havcr2^flox/flox vs. Havcr2^cKO
  • Dataset 3 (4-month-old mice) (4.bulkRNAseq_dataset3_4month.R): control vs. Havcr2^cKO vs. 5xFAD vs. Havcr2^cKO;5xFAD microglia
• Public datasets
  • Lund et al. 2018 (PMID: 29662171) (5.bulkRNAseq_Tgfbr2cKO_vs_control_Lund_2018.R): Tgfbr2^cKO vs. control microglia
  • Krasemann et al. 2017 (PMID: 28930663) (6.bulkRNAseq_Clec7apos_vs_neg_Krasemann_2017.R): Clec7a⁺ vs. Clec7a^- microglia

Processing and annotating snRNA-seq data (📁snRNAseq)

• 0.snRNAseq_functions.R: functions, color palettes and gene signatures used in the analysis
• 1.snRNAseq_process.R: QC, processing the snRNA-seq data; combining and subclustering microglia and PVM clusters
• 2.snRNAseq_doublet_removal.R: doublet removal using DoubletFinder
• 3.snRNAseq_annotation.R: annotating the snRNA-seq clusters
• 4.snRNAseq_subcluster_MG.R: further subclustering microglia cells; splitting Havcr2^icKO 5XFAD nuclei in cluster 2 (DAM/MGnD) into subpopulations P1 and P2 based on MSigDB Hallmark TGFβ signaling signature score

Processing and annotating scRNA-seq data (📁scRNAseq)

• 0.scRNAseq_functions.R: functions, color palettes and gene signatures used in the analysis
• 1.scRNAseq_process.R: QC and processing of the scRNA-seq data
• 2.scRNAseq_subcluster.R: subclustering cluster 6, 7, 8 obtained in 1.
• 3.scRNAseq_DoubletFinder.R: doublet identification using DoubletFinder
• 4.scRNAseq_annotation.R: identify MGnD, Homeostasis, interferon-rich and cycling populations based on signature score; combining and reclustering the cells based on the population characteristics; finalizing annotations
• 5.scRNAseq_DGE_analysis.R: differential gene expression analysis (Wilcoxon) comparing Havcr2^icKO;5xFAD and 5xFAD cells in each cluster using the FindMarkers function

Processing and annotating public AD sc/snRNA-seq datasets (📁public_AD)

• 1.GSE140510_snRNAseq_7M.R: processing 7-month snRNA-seq dataset from GSE140510
• 2.GSE98969_scRNAseq_6M.R: processing 6-month scRNA-seq dataset from GSE98969

Fig. 1

• Fig. 1a (Fig.1a_microarray_Thion_2018.R): Heatmap visualization of the developmental alterations of gene expressions of immune checkpoints and TGFβ pathway-related molecules in microglia in a published microarray dataset (Thion et al. 2018, PMID: 29275859)
• Fig. 1b (Fig.1b_GSE127449_P9_P28_scRNAseq.R): Dotplot visualization of gene expression at postnatal stages (P9 and P28) in a published scRNA-seq dataset (He et al. 2018, PMID: 34982959; GSE127449)
• Fig. 1h (Fig.1h_1month_heatmap_DEGs.R): Heatmap visualization of DEGs (FDR < 0.05) of 1-month-old Havcr2^cKO mice (n = 4 (3 males, 1 female)) compared to Havcr2^flox/flox mice (n = 5 (4 males, 1 female)). Rows represent genes and columns are biological replicates. Gene expression is row-normalized across replicates.
• Fig. 1i (Fig.1i_1month_volcanoplot.R): Volcano plot of differential gene expression analysis performed by DESeq2 comparing 1-month-old Havcr2^cKO to Havcr2^flox/flox mice
• Fig. 1j-l (Fig.1j_1l_1month_signature_score.R): Boxplot visualization of the the scores of MGnD (j), homeostasis (k), and KEGG phagosome (l) signatures in 1-month-old Havcr2^cKO compared to Havcr2^flox/flox mice. The Y-axis represents log2 transformed average TPM, and significance was calculated using the Student's two-tailed t-test.

Fig. 2

• Fig. 2a (Fig.2a_1month_Tgfbr2_circos_overlap.R): Circos plot comparison of up- and down-regulated DEGs in (1) Havcr2^cKO vs. control microglia from 1-month-old mice (2) Tgfbr2^cKO vs. control microglia
• Fig. 2b (Fig.2b_1month_Tgfbr2_scatter_plot.R): Scatter plot of genes based on expressional difference represented by log2-transformed fold changes in Havcr2^cKO (X-axis) and Tgfbr2^cKO (Y-axis) compared to control microglia
• Fig. 2e (Fig.2e_1month_DEG_homer_motif.Rmd): Transcription factor motif enrichment analysis using DEGs comparing Havcr2^cKO to control microglia from 1-month-old mice

Fig. 3

• Fig. 3g (Fig.3g_3month_heatmap_overlapping_DEGs.R): Heatmap visualization of top DEGs shared by at least two of the following three comparisons (3-month-old mice) (1) control phagocytosing versus control non-phagocytosing microglia (2) Havcr2^cKO phagocytosing versus control phagocytosing microglia (3) Havcr2^cKO non-phagocytosing versus control non-phagocytosing microglia
• Fig. 3k (Fig.3k_3month_nonphago_DEGs_pathway.R): Pathway analysis of DEGs in non-phagocytosing microglia from 3-month-old Havcr2^cKO mice compared to control mice. Disease pathways (pathways under section 6.1-6.10 from https://www.kegg.jp/kegg/pathway.html) and ribosomal genes were excluded from the analysis.
• Fig. 3l (Fig.3l_3month_nonphago_vs_Tgfbr2_circos_overlap.R): Circos plot comparison of up- and down-regulated DEGs from 3-month-old mice in below comparisons (1) Havcr2^cKO non-phagocytosing vs. control non-phagocytosing (2) control phagocytosing vs. control non-phagocytosing microglia

Fig. 4

• Fig. 4a (Fig.4a_overlapping_DEGs_pairwise.R): Number of overlapped genes and permutation test p-values between each pair of DEGs up- and down-regulated in Havcr2^cKO, phagocytosing, Tgfbr2^cKO, and Clec7a⁺ microglia compared to control microglia
• Fig. 4b (Fig.4b_overlapping_DEGs_corr_hetamap.R): Heatmap visualization of correlation between vectors of the expression levels of DEGs of Havcr2^cKO microglia compared to control microglia, as well as the DEGs shared by all three other comparisons (phagocytosing, Tgfbr2^cKO, Clec7a⁺ microglia compared to their corresponding control groups).

Fig. 6: snRNA-seq data

• Fig. 6a (Fig.6a_snRNAseq_microglia_UMAP.R): UMAP visualization of microglia snRNA-seq clusters split by genotype
• Fig. 6b (Fig.6b_snRNAseq_microglia_HallmarkTGFB_score.R): Probability density curves of the signature score for the Hallmark TGFβ pathway in the 5xFAD (purple) and Havcr2^icKO:5xFAD (yellow) phenotype. Individual cells of each genotype are represented by the bars on top.
• Fig. 6c-e: Projection of P1(magenta) and P2 (green) marker genes on the volcano plot of DEGs in different microglial perturbations: (c) Phagocytosis assay, (d) MGnD, (e) Tgfbr2^cKO. Only marker genes which are also significant in each perturbation are indicated. i, Spearman correlation and significance of log2 fold changes of all expressed genes in each perturbation condition (f-h) with those in P1 vs P2.
• Fig. 6f (Fig.6f_snRNAseq_microglia_dotplot_inflame_phago.R): Dotplot representation of P1 and P2 marker genes that are significant DEGs in perturbations (f-h) and known anti-inflammatory, pro-inflammatory and phagocytic properties
• Fig. 6g-h (Fig.6g_6h_snRNAseq_microglia_signature_score_cluster2.R): Violin plot of alternate macrophage (l) and phagocytic (m) signature genes among genotypes in cluster 2/MGnD/DAM. Significances are computed using one-way ANOVA with Tukey’s HSD correction.

Fig. 7: scRNA-seq data

• Fig. 7a (Fig.7a_scRNAseq_UMAP.R): UMAP visualization of microglia scRNA-seq clusters split by genotype
• Fig. 7b (Fig.7b_scRNAseq_signature_score_hedgesg.R): Heatmap visualization of the Hedges' g effect size (color) and significance (text) comparing the signature scores in Havcr2^icKO;5xFAD and 5xFAD cells in each microglia cluster. Significance is computed by t-test and adjusted using the Benjamini-Hochberg procedures. Signatures include:
  • cGAS-STING signature: Top 100 up- and down-regulated DEGs ordered by p-value comparing Cgas^WT/R241E to Cgas^WT/WT mice from Gulen et al. 2023 (PMID: 37532932)
  • MSigDB Hallmark interferon alpha signaling
  • MSigDB Hallmark TNFA signaling and KEGG TNF signaling pathways
  • KEGG phagosome and GO phagocytosis pathways, excluding MHC genes
  • KEGG lysosome pathway
  • MSigDB Hallmark PI3K/AKT/mTOR signaling and KEGG PI3K-Akt signaling pathways
  • MSigDB Hallmark MTORC1 signaling pathway
  • MSigDB Hallmark hypoxia signaling pathway
• Fig. 7c (Fig.7c_scRNAseq_signature_score_ttest.R): Heatmap visualization of T-test results comparing the scores of microglia population markers from Ellwanger et al. 2021 (PMID: 33446504) in Havcr2^icKO;5xFAD and 5xFAD cells in each microglia cluster. T-statistics are displayed in text (black if FDR < 0.05; white if FDR $\geq$ 0.05). Colors represent whether the average signature scores are higher in Havcr2^icKO;5xFAD or 5xFAD (pink: higher in Havcr2^icKO;5xFAD; blue: higher in 5xFAD).
• Fig. 7d (Fig.7d_scRNAseq_heatmap_selected_DEGs.R): Heatmap visualization of log2 fold-change and significances of selected DEGs comparing Havcr2^icKO;5xFAD to 5xFAD in each cluster.

Extended Data Fig. 1

• Extended Data Fig. 1e-f (Fig.s1e_s1f_l5_mousebrain_scRNAseq.R): analysis of a public scRNA-seq dataset of mouse nervous system (Zeisel et al. 2018, PMID: 30096314)
  • Extended Data Fig. 1e: The expression of immune checkpoint genes, including Havcr2, Lag3, and Vsir in mouse microglia and other cell populations in the central and peripheral nervous system
  • Extended Data Fig. 1f: Dissociation signature score in microglia and perivascular macrophages
• Extended Data Fig. 1g (Fig.s1g_humanbrain_scRNAseq.R): The expression of immune checkpoint genes and TGFβ pathway-related genes in human microglia and other cell populations in the brain (Gaublomme et al. 2019, PMID: 31266958)
• Extended Data Fig. 1i (Fig.s1i_microarray_Thion_2018.R): Heatmap visualization of the developmental alterations of the expressions of MGnD and homeostasis associated genes in microglia in a published microarray dataset (Thion et al. 2018, PMID: 29275859)

Extended Data Fig. 3

• Extended Data Fig. 3a (Fig.s3a_circo_overlapping_DEGs.R): Circos plot comparison of DEGs up- and down-regulated in Havcr2^cKO, phagocytosing, Tgfbr2^cKO, and Clec7a⁺ microglia compared to control microglia. Only islands with DEGs shared by at least 3 of the 4 comparisons were included. Permutation test p-values were displayed as ns P > 0.025, *P ≤ 0.025, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, and *****P ≤ 0.00001.
• Extended Data Fig. 3b (Fig.s3b_heatmap_overlapping_DEGs.R): Heatmap visualization of the expressions of DEGs shared by at least 3 of the 4 aforementioned comparisons.

Extended Data Fig. 7

• Extended Data Fig. 7a-b (Fig.s12a_4month_F_tim3cKO_5XFAD_circos_overlap.R, Fig.s12b_7month_F_tim3cKO_5XFAD_circos_overlap.R): Circos plot comparison of up- and down-regulated DEGs from 4- (a) and 7-month-old (b) female mice in the below comparisons (1) Havcr2^cKO vs. control microglia (2) 5xFAD vs. control microglia
• Extended Data Fig. 7c-d (Fig.s7c_s7d_4month_7month_boxplot_Havcr2.R): Boxplot of Havcr2 expression (log2 TPM) in 4- (c) and 7-month-old bulk RNA-seq male and female mice
• Extended Data Fig. 7h-j (Fig.s7h_s7j.snRNAseq_microglia_pct_barplot.R): Barplot visualization of proportions (Y-axis) of individual mouse (h), and genotype (i) per cluster (X-axis); and cluster per mouse (X-axis) (j)
• Fig. 7k (Fig.7k_snRNAseq_microglia_top5markers.R): Dotplot visualization of top 5 genes distinguishing each cell cluster from Fig. 6a
• Extended Data Fig. 7f-g (Fig.s7f_s7g_snRNAseq_allcells_UMAP_dotplot.R): snRNA-seq UMAP of all cells colored by annotations (f) and Dotplot of Havcr2 and Cx3cr1 expression in each cell type, split by genotype (g)
• Extended Data Fig. 7l-m (Fig.s7l_s7m.snRNAseq_microglia_signature_score.R): Violin plot of the MGnD (l) and homeostasis (m) signature score (Y-axis) across microglia clusters (X-axis) and genotypes.
• Extended Data Fig. 7ni (Fig.s7n.snRNAseq_microglia_proportion_test.R): Proportions of microglial cluster 0-4 in 5xFAD and Havcr2^icKO:5xFAD (n = 3 for each).
• Extended Data Fig. 7o (Fig.s7o.snRNAseq_microglia_signature_score_cluster0.R): Violin plot of scores for Tgfbr2^cKO signature genes among genotypes in cluster 0. Significances are computed using Wilcoxon test.

Extended Data Fig. 8: snRNA-seq data

• Extended Data Fig. 8a-b (Fig.s8a_s8b_snRNAseq_microglia_P1_P2_heatmap.R): Heatmap of top DEGs distinguishing P1 (a) and P2 (b) subpopulations in snRNA-seq cluster 2/MGnD/DAM
• Extended Data Fig. 8c-d (Fig.s8c_s8d_snRNAseq_microglia_P1_P2_featureplot.R): UMAP visualization of microglial snRNA-seq clusters colored by P1 (c) and P2 (d) signature scores split by genotype

Extended Data Fig. 9: snRNA-seq data

• Extended Data Fig. 9a-h (Fig.s9a_s9h.snRNAseq_microglia_BAM_PVM_monocyte_signature_score.R): Violin plot of border-associated macrophage (BAM, a), perivascular macrophage (PVM, b), monocyte (c-d) and microglia (e-h) signature scores across microglia clusters (X-axis) and genotypes.
• Extended Data Fig. 9i-j (Fig.s9i_s9j_snRNAseq_microglia_P1_P2_score.R): Violin plot of signature scores (Y-axis) identifying Havcr2^icKO:5xFAD populations P1 (i) and P2 (j) across all genotypes (color) and microglia clusters

Extended Data Fig. 10: snRNA-seq data

• Extended Data Fig. 10a (Fig.s10a_snRNAseq_microglia_DAM_pseudobulk_corr.R): Boxplot of pairwise spearman correlation coefficient (Y-axis) of whole gene expression pseudobulk microglial profiles between conditions (X-axis). Each point represents the spearman correlation coefficient between profiles of two biological replicates. Significance is computed by t-test.
• Extended Data Fig. 10b (Fig.s10b_snRNAseq_microglia_human_MG_markers.R): Twelve human microglial signatures curated using samples from patients with AD (Sun et al. 2023, PMID: 37774678). These signature scores were plotted for cluster 2 in our snRNAseq data. Significance were computed using one-way ANOVA with Tukey's HSD correction.
• Extended Data Fig. 10c-d (Fig.s10c_s10d_snRNAseq_microglia_P1_P2_vs_Hallmark_TGFB_sub2.R): Scatter plot of P1 (Y-axis, c) or P2 (Y-axis, d), and Hallmark TGFB signature scores among 5xFAD and Havcr2^icKO;5xFAD microglia in current study
• Extended Data Fig. 10e-g (Fig.s10e_s10g_public_snRNAseq_P1_P2_vs_Hallmark_TGFB_sub2.R): Scatterplot of P1 (Y-axis) or P2 (Y-axis), and Hallmark TGFB signature scores among 5xFAD microglia in MGnD clusters in public datasets (GSE98969 cluster 2 (e), GSE98969 cluster 3 (f) , GSE140510 cluster 1 (g))

Extended Data Fig. 11: scRNA-seq data

• Extended Data Fig. 11a (Fig.s11a_scRNAseq_Clec7a_IFNR_signature.R): UMAP of the scRNA-seq data colored by MGnD, Homeostatis, and interferon signature scores
• Extended Data Fig. 11b (Fig.s11b_scRNAseq_dotplot_top10markers_by_log2FC.R): Dotplot visualization of top 10 genes distinguishing each cell cluster
• Extended Data Fig. 11c (Fig.s11c_scRNAseq_perc_barplot_by_genotype.R): Barplot visualization of proportions (Y-axis) of genotype per cluster
• Extended Data Fig. 11d-e (Fig.s11d_s11e_scRNAseq_scatter_IFNDAMvsHMG_preMGnD.R): Scatter plot of genes based on expressional difference represented by log2-transformed fold changes of
  • IFN_DAM compared to IFN_HMG cluster in the scRNA-seq data (X-axis) and
  • late IFN-responsive pre-MGnD compared to early IFN-responsive pre-MGnD clusters from male (d) and female (e) mice defined in Yin et al. 2023 (PMID: 37291336)

Extended Data Fig. 12: scRNA-seq data

• Extended Data Fig. 12a-b (Fig.s12a_s12b_scRNAseq_scatter_Havcr2iKO5xFADvs5XFAD_preMGnD.R): Scatter plot of genes based on expressional difference represented by log2-transformed fold changes of
  • Havcr2^icKO;5xFAD vs 5xFAD in cluster HMG_0 (a) and DAM_0 (b)
  • Clec7a^- vs Clec7a⁺ (Y-axis)
• Extended Data Fig. 12c-d (Fig.s12c_s12d_scRNAseq_volcano_Havcr2icKO5xFADvs5xFAD.R): volcano plot of differential gene expression analysis comparing Havcr2^icKO;5xFAD and 5xFAD in cluster HMG_0 (c) and DAM_0 (d)