Fig 2: RNA-seq and analysis of mouse primary cortex ECs. (A) Expression of CD147 was detected before performing RNA-seq and the groups were based on genotype and age. (B) Hierarchical cluster analysis of all groups showed the reliability of our dataset. (C) Venn diagram of the overlap number of DEGs between comparisons. (D) Volcano plots showing the upregulated (red) and downregulated (blue) DEGs for each comparison. (E, F) Dot plots of GO terms enriched based on the DEGs for genotype (E) and age (F). The CD147 RNA expression level is presented relative to that of GAPDH, and the data are presented as the mean ± SEM of three independent experiments. ****p < 0.0001. DEGs were defined as genes with a fold change =1.5 and FDR-adjusted p < 0.05

Fig 3: Effects of CD147 deficiency on BBB integrity in vivo and in vitro. (A) WT and EC-KO mice at 8–12 weeks or 6–7 months of ages were given an i.v. injection of sodium fluorescein (Na-F), and the absorption of Na-F in the mouse brain was measured by a microplate reader at 485 nm, n = 5. (B) Brain water content of WT and EC-KO mice at 8–12 weeks or 6–7 months of age, n = 6. (C) Heatmap showing the 82 DEGs (see Table S1) from the following five KEGG pathways: Tight junction (ko 04530), Adherens junction (ko 04520), Focal adhesion (ko 04510), Cell adhesion molecules (CAMs; ko 04514) and Gap junction (ko 04540). All these enrichment pathways were p < 0.05 by hypergeometric tests. (D) Expression changes of selected genes (Ocln, Cldn5, Sparcl1) from the above dataset by RT-qPCR. Data are presented as the mean ± SEM of 3 independent experiments. (E) Western blot of DEGs of interest (OCLN, TJP1, SPARCL1, JAM-1). (F) Quantification of protein levels of molecules in (E) relative to GAPDH. (G) TEER measurement at 7 days after seeding BECs in the upper chamber via an EVOM Volt-Ohm resistance meter with STX2 chopstick electrodes. A schematic of the in vitro monocellular BBB model and TEER measurement is shown (upper right), n = 3 per group. (H) The paracellular permeability of the BBB model was assessed by 40 kDa FITC-dextran at different timepoints, including 0, 0.5, 1, 2, 4, 8, and 24 h. The fluorescence was measured using a Cytation Imaging Reader, n = 3 per group. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 compared to WT at the same ages; # p < 0.05 and ## p < 0.01 compared to 8–12 weeks with the same genotype

Fig 4: Disease-related transcriptomic changes in EC-KO. (A) IHC staining of Aß, CD45, Iba1, and CD147 in 18-month-old APP/PS1 mouse brain slices and peer comparisons. Scale bar: 100 µm (B) The expression of DEGs whose human orthologs are related to neurodegenerative and cerebrovascular diseases in GWAS datasets, expression is represented as FPKM. Disease abbreviations: ALS, amyotrophic lateral sclerosis; WMH burden/SVD, white matter hyperintensity burden and cerebral small vessel disease; MSA, multiple systems atrophy; AD, Alzheimer's disease; PD, Parkinson's disease; FTD, frontotemporal dementia; HD, Huntington's disease. There were 21 DEGs related to AD (p = 0.00266). (C) Visual PPI network of the AD-related dataset. The edges between nodes represent different protein–protein associations, including experimental determination (purple), interactions from current databases (blue), and coexpression (black). (D) Expression changes of selected genes (App, Psen2, Epha1, Cd2ap, Bin1, Mef2c, Scarb2) by RT-qPCR. Data are presented as the mean ± SEM of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. (E) Human AD brain differential expression versus the control (p < 0.05) served as the y-axis, and DEGs (fold change =2 with FDR-adjusted p < 0.05) from EC-KO mice versus WT mice served as the x-axis. The genes and their human orthologs were concordant in the first and third quadrants (light green), whereas they were discordant in the second and fourth quadrants (light red). Twenty-one DEGs were specific in the groups compared at 8–12 weeks of age (red dot, EC-KO versus WT), and 35 DEGs were specific in the groups compared at 6–7 months of age (green dot), whereas 11 DEGs were shared (blue dot) in both groups

Fig 5: Expression of CD147 on BECs. (A) Brain sections from mice were costained for CD147 (red), NeuN (neuron marker, green), GFAP (astrocyte marker, green), MBP (oligodendrocyte marker, green), Iba1 (microglial cell marker, green), CD31 (endothelial cell marker, green), and nuclei (blue) to evaluate CD147 expression on BECs. Scale bars: 100 µm. (B) H&E and IHC staining of CD147 showed brain features and human and mouse brain vessels stained positively for CD147 (arrowhead). Scale bars: 100 µm. (C) IHC staining of CD147 in 8–12-week-old and 6–7-month-old mouse brains, Scale bars: 100 µm. (D) RT-qPCR and (E) WB validated the dynamic CD147 expression in isolated primary BECs from mice between 8 and 12 weeks and 6–7 months of age. Data are presented as the mean ± SEM of n = 3 independent experiments. **p < 0.01