Type-1-to-type-2 transition of brain microvascular pericytes induced by cytokines and disease-associated proteins: Role in neuroinflammation and blood-brain barrier disruption

Abstract

While the concept of pericyte heterogeneity in the brain microvasculature is becoming more widely accepted, little is known about how they arise, or their functional contributions to the blood-brain barrier (BBB). We therefore set out to examine the distribution of subtypes of pericytes at the BBB and sought to elucidate some of their functional characteristics by examining their unique mRNA expression patterns. We demonstrate that type-1 pericytes (PC1) that are associated with young healthy brains and BBB homeostasis, can transition into type-2 pericytes (PC2) that are associated with disease and BBB breakdown, both in vitro and in vivo, in the presence of both endogenous and disease associated ligands. We identified PC1 and PC2 in single-cell RNA-sequencing from vascular enriched mouse brain and identified transcriptional differences between PC1 and PC2. PC2 showed increased expression of genes associated with phagocytosis and peripheral immune cell infiltration. On the contrary, PC1 displayed increased expression of genes involved in hedgehog signaling, which is known to promote tight junction formation at the BBB. Our data support the PC1-to-PC2 transition as an origin of PC diversity and suggest a functional role for PC1 in maintaining BBB homeostasis and PC2 in responding to pathological conditions.

Keywords: Amyloid beta; HIV-1 gp120; atherosclerosis; blood-brain barrier; single-cell RNA sequencing; smooth muscle actin.

Figure 1.

Factors known to disrupt BBB integrity induce PC1-to-PC2 transition in vitro and in vivo. Passage-4 pHBMVPCs treated with various concentrations of human recombinant IL-1B (a), TGF-B1 (b), and VEGF (c) all showed a significant increase in the Continued.percentage of aSMA stained (%SMA+, green) cells by ICC compared to the number of PDGFRB (red) positive DAPI nuclei (blue). Some treatments induced cell growth (green points on graph) as measured by a significant increase in the number of DAPI-labeled nuclei compared to the vehicle-treated well, or significant cell death (red points on graph) as measured by a significant decrease in the number of DAPI nuclei compared to the vehicle-treated well. Treatment with human recombinant TGF-B1 at 50 and 100 ng/mL resulted in too much cell death for these wells to be quantified (b). Representative image of a PC1 and PC2 cell in vivo with PDGFRB (red) and aSMA (green) (d). Both %PC2 (e) and fibrinogen extravasation (f) were significantly increased in the brains of mice injected icv with IL-1B, TGF-B1, and VEGF compared to saline-injected controls, which were not significantly different than non-surgery controls (nSx). There was a strong positive correlation between %PC2 and fibrinogen extravasation in each animal (g).

Figure 2.

HIV GP120 can induce PC1-to-PC2 transition in vitro and in vivo. Passage-4 pHBMVPCs treated with various concentrations of HIV GP120 showed a significant increase in the percent aSMA-stained (green) cells by ICC compared to the number of PDGFRB (red) positive DAPI nuclei (blue) (a). Graphical representation of the GFAP-GP120 transgenic (tg) mouse model (b). HIV GP120tg mice had a significantly higher %PC2 (c) and fibrinogen extravasation than non-transgenic littermates, but HIV GP120tg mice had a significantly higher %PC2 and fibrinogen extravasation than mice that did not express HIV GP120 (d). There was a significant positive correlation between %PC2 and fibrinogen extravasation in these animals (e).

Figure 3.

Amyloid beta induces PC1-to-PC2 transition in vitro and in vivo. Passage-4 pHBMVPCs treated with various concentrations of human amyloid beta 40 (AB40) or 42 (AB42) showed a significant increase in the percentage of aSMA-stained (green) cells by ICC compared to the number of PDGFRB (red) positive DAPI nuclei (blue) (a, b). Both %PC2 (c) and fibrinogen extravasation (d) were significantly increased in the brains of mice injected icv with AB40 or AB42 compared to saline injected controls, which were not significantly different than non-surgery controls. There was a strong positive correlation between %PC2 and fibrinogen extravasation in each animal (e).

Figure 4.

LDLR^−/− mice on DDC diet injected with AAV-PCSK9 have increased %PC2, which correlates with BBB breakdown and corresponds to trends in all measured models. Graphical representation of Stat4^flox LDLR^−/− mouse model (a). LDLR^−/− mice on DDC diet showed a significant increase in %PC2 (b) and fibrinogen extravasation (c) compared to chow fed controls, which were not significantly different than non-surgery controls. There was a strong positive correlation between %PC2 and fibrinogen extravasation in LDLR^−/− animals (d). There was a significant positive correlation between %PC2 and fibrinogen extravasation in all mice studied regardless of group (e).

Figure 5.

Single-cell RNA sequencing of vascularly enriched mouse brain fraction discovers two distinct pericyte populations. Graphical representation of single cell experiment (a). UMAP of all quality cell reads from young adult and aged adult vascularly enriched mouse brains (b). UMAP of vascular fraction sub-clustering containing ECs, SMCs, PCs, and fibroblasts (c). Heatmap of relative gene expression from genes used to identify clusters (d). UMAP of pericyte only sub-clustering (e). Volcano plot of differentially expressed genes between PC1 and PC2 (f). Representative UMAP of the RNA velocity-based pseudotime analysis results (g).

Figure 6.

Transcriptional differences between PC1 and PC2 IPA differentially regulated pathway analysis showing nodes upregulated in PC1 (orange), PC2 (blue), or both (grey) (a). Representative IF images of PDGFRB (blue), LAMP1 (green), and aSMA (red) in a PC1- (b) and PC2- (c) associated vessel shows higher LAMP1 in PC2-associated vessels (d). Graph of upregulated (positive) and downregulated (negative) pathways identified in PC1 (green) and PC2 (red) based on PANTHER gene ontology analysis supports findings from IPA (e). Heatmap of differentially regulated genes associated with the KEGG hedgehog pathway list in PC1, PC2, Cap EC1, and Cap EC2 cells (f). Representative IF images of PDGFRB (blue), CDON (green), and aSMA (red) in a PC1 (g) and PC2 (h) associated vessel shows higher CDON in PC1 associated vessels (i). A graphical representation of in situ validated functional difference between PC1 and PC2 (j).