Excess vascular endothelial growth factor-A disrupts pericyte recruitment during blood vessel formation

Abstract

Pericyte investment into new blood vessels is essential for vascular development such that mis-regulation within this phase of vessel formation can contribute to numerous pathologies including arteriovenous and cerebrovascular malformations. It is critical therefore to illuminate how angiogenic signaling pathways intersect to regulate pericyte migration and investment. Here, we disrupted vascular endothelial growth factor-A (VEGF-A) signaling in ex vivo and in vitro models of sprouting angiogenesis, and found pericyte coverage to be compromised during VEGF-A perturbations. Pericytes had little to no expression of VEGF receptors, suggesting VEGF-A signaling defects affect endothelial cells directly but pericytes indirectly. Live imaging of ex vivo angiogenesis in mouse embryonic skin revealed limited pericyte migration during exposure to exogenous VEGF-A. During VEGF-A gain-of-function conditions, pericytes and endothelial cells displayed abnormal transcriptional changes within the platelet-derived growth factor-B (PDGF-B) and Notch pathways. To further test potential crosstalk between these pathways in pericytes, we stimulated embryonic pericytes with Notch ligands Delta-like 4 (Dll4) and Jagged-1 (Jag1) and found induction of Notch pathway activity but no changes in PDGF Receptor-β (Pdgfrβ) expression. In contrast, PDGFRβ protein levels decreased with mis-regulated VEGF-A activity, observed in the effects on full-length PDGFRβ and a truncated PDGFRβ isoform generated by proteolytic cleavage or potentially by mRNA splicing. Overall, these observations support a model in which, during the initial stages of vascular development, pericyte distribution and coverage are indirectly affected by endothelial cell VEGF-A signaling and the downstream regulation of PDGF-B-PDGFRβ dynamics, without substantial involvement of pericyte Notch signaling during these early stages.

Keywords: Angiogenesis; Development; Mouse embryonic stem cells; Pericyte; VEGF-A.

Fig. 1

Genetic loss of Flt1 impairs Ng2+ pericyte coverage of ESC-derived blood vessels. A Representative images of WT (i-iv) and Flt1^−/− (v-viii) Day 9 ESC-derived blood vessels labeled for endothelial cells (Pecam1: i and v, green in iv and viii), pericytes (Ng2: ii and vi, red in iv and viii), and cell nuclei (DAPI: iii and vii, blue in iv and viii). Scale bars, 50 μm. B Average percentages of Ng2+ pericyte coverage on Days 9 (n=6 of biological replicates) and 10 (n=10 of biological replicates) ESC derived vessels for WT (black bars) and Flt1^−/− (white bars) conditions. Values are averages + Standard Error of the Mean (SEM). *P≤0.05 vs. WT at the same time point

Fig. 2

Flt1^−/− ESC-derived vessels display defects in pericyte distribution. A Schematic of ESC-derived vasculature with specific morphological locations denoted with dotted boxes: 1- vessel stalks, 2- branch points, and 3- thick areas. B Average number of Ng2+ pericytes at the indicated vessel locations within Day 9 WT (black bars, n=23 cells) and Flt1^−/− (white bars, n=28 cells) ESC-derived vasculature. Values are averages + SEM. *P≤0.05 vs. WT at the same vessel location. C Percent distribution of pericytes at each vessel location (stalk: blue, branch point: red, thick area: gray) for Day 9 WT and Flt1^−/− vessels (WT: n=158 cells, and Flt1^−/−: n=127 cells). *P=0.0002, Chi-square test of WT and Flt1^−/− distributions. D Schematic of approach to quantifying pericyte density, specifically pericytes within a 50 μm radius of one another (dashed arrow and circle). E Average number of Ng2+ pericytes within a 50 μm radius of one another on Day 8–10 WT (black bars, Day 8: n=7, Day 9: n=23, Day 10: n=56) and Flt1^−/− (white bars, Day 8: n=18, Day 9: n=28, Day 10: n=53) ESC-derived vessels. Values are averages + SEM. *P≤0.05 vs. WT at Day 9 and Day 10, and **P≤0.05 for Day 8 WT vs. Day 10 WT.

Fig. 3

ESC-derived pericytes display little to no Flt1 promoter activity or gene expression. A Representative images of Ng2+ pericytes (Ng2: i, red in iv and v) and Flt1 promoter activity as indicated by β-galactosidase (β-gal: ii, blue in iv and v) production from the Flt1:LacZ gene. Cell nuclei are labeled by DAPI (iii, white in v). Scale bar, 5 μm. B Fold change in Flt1 expression between endothelial cells (yellow bar) and pericytes (purple bar) enriched from WT ESC-derived vessels. Values are averages + SEM, n=4 biological replicates. *P≤0.05

Fig. 4

Exogenous VEGF-A disrupts pericyte migration and limits pericyte distribution on developing embryonic blood vessels. A Representative sequential images from movies of vehicle control- (i-iii) and VEGF-A- (iv-vi) treated embryonic skin vessels in which endogenous pericytes (Ng2-DsRed+ and arrows, left column and red in right column) migrated along sprouting endothelial cells (Flk1-eGFP+ and arrowheads, middle column and green in right column). Time in upper right corner, hours:minutes (hh:mm). Scale bars, 100 μm. B Average percent of pericytes migrating on sprouts with directional persistence (tan bars) or static movement (dark blue bars) in control (n=23 movies) and VEGF-A-treated (n=15 movies) embryonic vessels. Values are averages + SEM. *P≤0.05. C Schematic of approach to quantifying pericyte density, specifically pericytes on sprouting endothelial cells within a 50 μm radius of one another (dashed arrow and circle). D Average number of Ng2-DsRed+ pericytes within a 50 μm radius of one another in the first (black bars) and last (red bars) frames of movies from control (n=6 biological replicates) and VEGF-A-treated (n=4 biological replicates) embryonic vessels. Values are averages + SEM. *P≤0.05 vs. First Frames of VEGF-A-treated group

Fig. 5

Loss of Flt1 disrupts transcriptional regulation within the Notch and PDGF-B pathways. A Fold change in Notch pathway gene expression between WT (dark blue bars) and Flt1^−/− (yellow bars) endothelial cells enriched from ESC-derived vessels. Values are averages + SEM, n=4–8 biological replicates per gene. *P≤0.05 vs. WT. B Fold change in Notch pathway gene expression between WT (blue bars) and Flt1^−/− (purple bars) pericytes enriched from ESC-derived vessels. Values are averages + SEM, n=4–8 biological replicates per gene. C Fold change in PDGF-B pathway gene expression between WT (dark blue bars) and Flt1^−/− (yellow bars) endothelial cells enriched from ESCderived vessels. Values are averages + SEM, n=4–8 biological replicates per gene. *P≤0.05 vs. WT. D Fold change in Pdgfrβ expression between WT (blue bars) and Flt1^−/− (purple bars) pericytes enriched from ESC-derived vessels. Values are averages + SEM, n=4–8 biological replicates per gene. *P≤0.05 vs. WT

Fig. 6

Flt1^−/− pericytes produce less PDGFRβ than WT pericytes, and pericytes in both backgrounds produce a truncated PDGFRβ isoform. A Representative images of Western Blots of recombinant PDGFRβ protein (top, antibody validation) and lysates from ESCderived pericytes (bottom). Full-length (~180 kDa) and truncated (~60 kDa) isoforms of PDGFRβ were detected in WT and Flt1^−/− pericyte lysates, and GAPDH (loading control) was used to normalize and compare relative amounts of full-length PDGFRβ (0.29) and this truncated isoform (0.34) between groups. n=4 biological replicates. B Map of Pdgfrβ exons with important features denoted as well as arrows indicating the location of forward and reverse primers used to identify potential mRNA splice variants. C Representative image of Pdgfrβ amplicons separated on an agarose gel. Number ranges indicate PCR products within indicated primer sets. * indicates the presence of a potential mRNA splice variant.

Fig. 7

Immobilized Dll4, but not Jag1, induces changes in embryonic pericyte gene expression, but neither ligand alters Pdgfrβ expression. A Schematic of experimental setup for coating plates with Notch or control ligands, deriving Ng2-DsRed+ embryonic pericytes, and cultures these cells on Notch ligands to measure changes in gene expression. B Fold change in Pdgfrβ and Notch pathway gene expression between embryonic pericytes cultured under control conditions (black bars), exposed to Fc-γ and blocking serum only (white bars), on Fc-γ with Human Fc (gray bars), on Fc-γ with Dll4:Fc (red bars), or on Fc-γ with Jag1:Fc (light blue bars). Values are averages + SEM, n=5 biological replicates. *P≤0.05 vs. control conditions for specified gene target.