Fig 2: Youth-associated TIMP2 alters microglia in aged mice.(a) Schematic of experimental timeline of microglial isolation from 20-month-old WT mice, followed by incubation with TIMP2 or vehicle. (b) Representative images from cultures of WT primary microglia stained with anti-IBA1 and anti-CD68 antibodies by ICC (arrowheads indicate IBA1+CD68+ cells; scale bar = 100 μm), with corresponding (c) quantification of the number of CD68+ cells normalized to total cell number (N = 8 mice per treatment group, 2 independent experiments combined; sex-matched; mean ± SEM). (d) Schematic diagram showing paradigm of systemic treatment with recombinant TIMP2 or vehicle in 20-month-old WT mice, followed by downstream analyses. (e) Representative confocal images of DG from 20-month-old WT mice treated systemically with TIMP2 or vehicle, stained with anti-IBA1 and anti-CD68 antibodies (scale bar = 100 μm; white arrowheads indicate IBA1+CD68+ cells), with corresponding (f) quantification of number of IBA1+CD68+ cells normalized to DG area (N = 9 male mice per group; mean ± SEM). (g) Uniform Manifold Approximation and Projection (UMAP) plot of hippocampal microglial nuclei (2,831) isolated from 20-month-old WT mice treated systemically with vehicle or TIMP2 (N = 4 male mice per group combined as 2 samples per group). (h) Proportion of nuclei across microglial subclusters for each treatment group with corresponding (i) Significant canonical pathways from IPA for marker genes significantly upregulated in subcluster 2, (j) subcluster 5, and (k) subcluster 6. (l) IPA upstream regulator network based on subcluster 6 marker genes, with predicted relationships for target genes. (m) Violin plots of normalized gene expression values from selected top DEGs; *P < 0.0001, from DESeq2 analysis. (n) Significant canonical pathways from IPA on microglial DEGs (P-valueadj.< 0.05) from TIMP2-treated vs. vehicle-treated aged mice. (o) Dot plot from Gene Set Enrichment Analysis (GSEA) performed on ranked gene lists derived from differential expression analysis of microglia following TIMP2 treatment. Significantly enriched gene sets derived from published transcriptomic states are shown with corresponding Normalized Enrichment Score (NES), with point size representing number of overlapping genes and color indicating P-valueadj. (p) Representative confocal images (left) of DG from 20-month-old WT mice treated systemically with TIMP2 or vehicle, stained with anti-IBA1, anti-CD68, and anti-Vglut1 antibodies (scale bar = 100 μm), with Imaris-based reconstruction (right) of insetted microglial volume (red), containing engulfed VGLUT1+ puncta (purple with yellow outlines indicated by white arrowheads) within microglial lysosomes (green) (scale bar = 50 μm). (q) Quantification of the volume of engulfed VGLUT1+ signal within CD68+ volume, normalized to microglia volume. Values represent N = 5 mice per group; mean ± SEM). *P < 0.05, **P < 0.01; Student’s t test.

Fig 3: TIMP2 is released by microglia and alters microglial response to myelin.(a) Schematic diagram showing timeline of primary microglia culturing from adult mice and media assay for TIMP2 immunoblotting experiments. (b) TIMP2 immunoblotting (representative) from conditioned serum-free media from primary microglia isolated from 2-month-old WT mice, including Ponceau stain (bottom), with corresponding (c) TIMP2 band intensity quantification (N = 4 mice per timepoint; 4 independent experiments combined, with a full timecourse for each experiment; sex-matched; mean ± SEM). (d) TIMP2 immunoblotting (representative) from conditioned serum-free media from primary microglia isolated from 2-month-old WT mice at baseline (0 hours with no treatment) or 12 hours following treatment with either PBS, zymosan (10 mg/mL), myelin (10 mg/mL), or LPS (100 ng/mL), with corresponding (e) TIMP2 band intensity quantification (N = 3–4 mice per treatment; 4 independent experiments combined, with all treatments representing an experiment; sex-matched; mean ± SEM). (f) Schematic diagram of treatment of adult primary microglia with pHrodo-labeled myelin and subsequent Incucyte imaging to assess uptake. (g) Timecourse from an experiment with total integrated density (red object mean intensity [RCU, red, x average phase object area (cell, μm2)) of pHrodo-labeled myelin signal in primary microglia cultured from tamoxifen-treated 5–7-month-old Cx3cr1CreERT2/+;Timp2fl/fl and Timp2fl/fl control mice, with corresponding (h) uptake quantification using area under the curve (AUC; N = 6–7 mice per group; 2 independent experiments combined; sex-matched; mean ± SEM). (i) Schematic of experimental design to assess clearance of pHrodo-labeled myelin in Cx3cr1CreERT2/+;Timp2fl/fl and Timp2fl/fl control mice. (j) Timecourse from an experiment with total integrated density (RCU [red calibrated units]) x average phase object area (cell, μm2)) of pHrodo-labeled myelin signal in primary microglia cultured from 5–7-month-old tamoxifen-treated Cx3cr1CreERT2/+;Timp2fl/fl and Timp2fl/fl control mice, with corresponding (k) clearance rate quantification (from slopes of regressions), represented relative to Timp2fl/fl control (N = 5–7 mice per group; 2 independent experiments combined; sex-matched; mean ± SEM). (l) Schematic diagram of neuron-specific TIMP2 deletion using Timp2fl/fl and SynCre/+;Timp2fl/fl mice for confocal imaging experiments. (m) Representative confocal images of DG from 2–3-month-old Timp2fl/fl and SynCre/+;Timp2fl/fl littermates stained with anti-IBA1 and anti-CD68 antibodies (left; scale bar = 100 μm) and magnified inset (right; scale bar = 50 μm; white arrowheads indicate IBA1+CD68+ cells) with corresponding (n) quantification of IBA1+CD68+ cell number normalized to DG area (N = 8–9 female mice per group; mean ± SEM). *P < 0.05, **P < 0.01; One-way ANOVA with Tukey’s post hoc test (c); One-way ANOVA with Dunnett’s post hoc test (e); Student’s t test (h,k,n).

Fig 4: TIMP2 regulates microglial state and brain extracellular composition in vivo.(a) Representative confocal images of dentate gyrus (DG) of 3–4-month-old TIMP2 KO and WT mice stained with anti-IBA1 and anti-CD68 antibodies (left; scale bar = 100 μm) and magnified inset (right; scale bar = 50 μm) (N = 9–13 mice per group; sex-matched; arrowheads indicate IBA1+CD68+ cells) with corresponding (b) quantification of thresholded DG area covered by IBA1+ staining and (c) the number of IBA1+CD68+ cells normalized to DG area; mean ± SEM. (d) Representative confocal images of dentate gyrus (DG) of 6–7-month-old TIMP2 KO and WT mice stained with anti-IBA1 and anti-CD68 antibodies (left; scale bar = 100 μm) and magnified inset (right; scale bar = 50 μm) (N = 5–8 mice per group; female mice; arrowheads indicate IBA1+CD68+ cells) with corresponding (e) quantification of thresholded DG area covered by IBA1+ staining and (f) the number of IBA1+CD68+ cells normalized to DG area; mean ± SEM. (g) Schematic diagram of the high molecular-weight cut-off (HMWCO,1-MDa) in vivo microdialysis method to dialyze hippocampal ISF proteins and the corresponding (h) timeline of LPS treatment paradigm in mice undergoing microdialysis (N = 4; 6–7-month-old male 5XFAD mice) with (i) brain ISF protein measurements, represented as z-scored NPX values for significant proteins (paired t-test; *P < 0.05, **P < 0.01, ***P < 0.001; median with quartiles). (j) Timeline of HMWCO in vivo microdialysis experiments to evaluate steady state levels of hippocampal ISF proteins in 2–3-month-old TIMP2 WT and KO mice (N = 8–9 mice per group; female mice) with corresponding (k) brain ISF protein measurements, represented as z-scored NPX values for significant proteins (unpaired t-test with Welch’s correction; *P < 0.05, **P < 0.01; median with quartiles). Student’s t test used in panels (b-c) and (e-f); *P < 0.05.

Fig 5: TIMP2 is expressed in microglia and its deletion leads to phenotypes associated with increased activation.(a) Schematic diagram of isolation, culturing, and assays performed on TIMP2 KO and WT primary microglia. (b) TIMP2 immunoblotting from primary microglia lysates isolated from TIMP2 WT and KO littermates (N = 4 samples per genotype, comprised of 5–6 pups (P5-P8) pooled per sample). (c) Representative images from cultures of TIMP2 KO and WT primary microglia (N = 3 pups (P6-P7) pooled per genotype; repeated in 2 additional experiments) stained with anti-TIMP2 and anti-IBA1 antibodies by ICC (scale bar = 100 μm). (d) Gene expression (transcript per million) values for Timp2 from bulk RNAseq experiments from TIMP2 KO and WT primary microglia (N = 7 samples per genotype, with each sample representing a pool of 5 pups (P6-P7); sex-matched; mean ± SEM) with subsequent (e) over-representation analysis of GO terms (Padj < 0.05) using upregulated and downregulated DEGs in TIMP2 KO vs. WT primary microglia. (f) Representative images from WT primary microglia cultures stained by ICC with DAPI, as well as anti-TIMP2 and anti-CD68 antibodies for co-localization analyses (N = 8 mice, 5–6-month-old mice; sex-matched; 2 independent experiments combined; scale bar = 100 μm), with (g) quantification of the fraction of TIMP2+ cells that are TIMP2+CD68+; violin plot of median with quartiles.