Epigenetic Activation of Pro-angiogenic Signaling Pathways in Human Endothelial Progenitors Increases Vasculogenesis

Abstract

Human endothelial colony-forming cells (ECFCs) represent a promising source of adult stem cells for vascular repair, yet their regenerative capacity is limited. Here, we set out to understand the molecular mechanism restricting the repair function of ECFCs. We found that key pro-angiogenic pathways are repressed in ECFCs due to the presence of bivalent (H3K27me3/H3K4me3) epigenetic marks, which decreases the cells' regenerative potential. Importantly, ex vivo treatment with a combination of epigenetic drugs that resolves bivalent marks toward the transcriptionally active H3K4me3 state leads to the simultaneous activation of multiple pro-angiogenic signaling pathways (VEGFR, CXCR4, WNT, NOTCH, SHH). This in turn results in improved capacity of ECFCs to form capillary-like networks in vitro and in vivo. Furthermore, restoration of perfusion is accelerated upon transplantation of drug-treated ECFCs in a model of hindlimb ischemia. Thus, ex vivo treatment with epigenetic drugs increases the vascular repair properties of ECFCs through transient activation of pro-angiogenic signaling pathways.

Keywords: ECFCs; EZH2; UTX; angiogenesis; bivalent genes; epigenetics; hindlimb ischemia; human endothelial progenitors; pro-angiogenic pathway; vasculogenesis.

Graphical abstract

Figure 1

Pharmacologic Treatment with a Combination of EZH2 and HDAC Inhibitors Enhances ECFC Migration, Capillary Network Formation, and Resistance to Serum Starvation-Induced Apoptosis (A) GSK-343 and panobinostat lead to a global decrease in H3K27me2/3 levels and an increase in H3/H4-acetyl levels, respectively, as analyzed by western blot. Molecular masses are indicated in kDa. Total histone H3 and H4 serve as loading controls. (B) GSK-343 and panobinostat do not affect the overall level of indicated proteins, as analyzed by western blot. Molecular masses are indicated in kDa. Tubulin serves as a loading control. (C) GSK-343 and panobinostat increase the kinetics of ECFC migration as measured by gap closure assay. Left: representative pictures of ECFCs stained with crystal violet after 10 hr migration (10× magnification; scale bar, 250 μm). Dashed circles represent the gap area prior to cell migration. Right: fraction of the gap area invaded by migrating cells at the indicated times. Average from thee independent experiments (n = 3) each performed in duplicate are shown ±SEM. (D) GSK-343 and panobinostat increase ECFC-mediated capillary-like structure formation on Matrigel. Left: representative pictures of capillary-like structures stained with calcein (2.5× magnification; scale bar, 1 mm). Right: total capillary length of the network is expressed as the mean percentage of control values corresponding to cells treated with vehicle ±SEM. Data shown from three independent experiments (n = 3) each performed in duplicate. (E) GSK-343 and panobinostat do not induce apoptosis in ECFCs as measured by fluorescence-activated cell sorting (FACS) after 7-AAD and annexin V staining. (F) GSK-343 and panobinostat reduce serum starvation-induced apoptosis in ECFCs as measured by FACS after 7-AAD and annexin V staining. ECFCs were grown in serum-containing EGM-2 medium or serum-depleted EBM-2 medium. (E and F) Data expressed as mean percentage of control values corresponding to cells treated with vehicle ±SEM from three independent experiments (n = 3), each performed in duplicate. (G) GSK-343 and panobinostat decrease ECFC proliferation as analyzed by bromodeoxyuridine (BrdU) incorporation. The percentage of cells incorporating BrdU is indicated ±SEM from four independent experiments (n = 4), each performed in duplicate. ^∗∗∗p < 0.001; ^∗∗p < 0.01; ^∗p < 0.05; ns, non-significant. See also Figures S1 and S2.

Figure 2

Pharmacologic Treatment of ECFCs with a Combination of EZH2 and HDAC Inhibitors Activates Multiple Pro-angiogenic Pathways Simultaneously (A) Global changes in gene expression upon drug treatment versus vehicle treatment analyzed by RNA-seq. Scatterplot shows upregulated (red) and downregulated (green) genes (n = 3). (B) Representative Gene Ontology (GO) Biological Process categories significantly enriched for up- and downregulated genes in ECFCs treated with GSK-343 and panobinostat. (C) Heatmaps of up- and downregulated genes that belong to the indicated signaling pathways as identified by RNA-seq analysis and validated by qRT-PCR (^∗). Data are expressed in log2(FPKM) (n = 3). The feedback arrow indicates a gene belonging to both receptor and target gene categories. (D) Validation of cell-signaling pathway activation upon drug treatment at the protein level. Left: western blot analysis of ECFC extracts shows an increase in global protein levels of VEGFR1 (FLT1), VEGFR2 (KDR), cleaved intracellular domain of NOTCH1 (NOTCH1-NICD), and CXCR4 upon drug treatment. Middle: western blot analysis of VEGFR2 immunoprecipitates (IP) shows an increase in phospho-VEGFR2 upon drug treatment. Right: western blot analysis of ECFC extracts shows an increase in phospho-AKT and phospho-ERK1/2 upon drug treatment. Tubulin is used as a loading control. Molecular masses are indicated in kDa. See also Figures S3 and S4 as well as Tables S1 and S2.

Figure 3

Pharmacologic Treatment of ECFCs with a Combination of EZH2 and HDAC Inhibitors Increases Gene Expression through H3K27me3 Demethylation and Histone H3 Acetylation (A) Transcript levels of indicated genes were measured by qRT-PCR following treatment of ECFCs with vehicle, GSK-343, panobinostat, or GSK-343-panobinostat combination, as indicated. qRT-PCR values are expressed as mean percentages of vehicle-treated cells ±SEM with B2M serving as internal control. (n ≥ 3). ^∗∗∗p < 0.001; ^∗p < 0.05; ns, non-significant. (B) RNA polymerase II (pol II) binding to indicated gene promoters was assessed by ChIP-qPCR. (C–E) Enrichment of H3K27me3 (C), H3K4me3 (D), and H3 acetyl (E) histone marks to indicated gene promoters was assessed by ChIP-qPCR. (B–E) ChIPs were performed using indicated antibodies (Ab) or IgG as a negative control (Mock). qPCR values are expressed as mean fractions of input ±SD. See also Figure S5.

Figure 4

Co-binding of Histone H3K27 Methyltransferase/Demethylase and Histone Acetyltransferase/Deacetylase to Gene Promoters (A–D) Binding of the H3K27 methyltransferase EZH2 (A), the H3K27 demethylase UTX (B), the histone deacetylase HDAC1 (C), and the histone acetyltransferase p300 (D) to indicated gene promoters was measured by ChIP-qPCR. ChIPs were performed using indicated antibodies (Ab) or IgG as a negative control (Mock). qPCR values are expressed as mean fractions of input ±SD. (E) Model representing co-binding of antagonistic enzymes on gene promoters such that transcription is maintained at low levels (poised gene). Combined inhibition of the repressive enzymes EZH2 and HDAC1 displaces the equilibrium leading to transcriptional activation.

Figure 5

Gene Activation upon Treatment with a Combination of HDAC and EZH2 Inhibitors Is Reversible (A) Transcript levels of indicated genes were measured by qRT-PCR at the indicated times after removal of the GSK-343/panobinostat drug combination. qRT-PCR values are expressed as mean percentages of vehicle-treated cells ±SEM with B2M serving as internal control. (n = 3). (B and C) Enrichment of H3K27me3 (B) and H3 acetyl (C) histone marks is measured by ChIP-qPCR at the indicated times after removal of the GSK-343/panobinostat drug combination. qPCR values are expressed as mean fractions of input ±SD. ^∗∗∗p < 0.001; ^∗∗p < 0.01; ^∗p < 0.05; ns, non-significant.

Figure 6

Pharmacologic Treatment with a Combination of EZH2 and HDAC Inhibitors Accelerates Vasculogenesis and Blood-Flow Recovery by Transplanted ECFCs in a Mouse Model of Hindlimb Ischemia (A) H&E staining (left) and anti-human (h) CD31 staining (right) identify ECFCs that have formed micro-capillaries in collagen gels after 3 weeks of implantation in mice (scale bar, 200 μm). Numbers of fully closed ECFC-derived hCD31+ capillaries are indicated per measured explanted area as mean values ± SEM (n = 6 mice per group). (B) Laser Doppler perfusion imaging was used to measure blood flow pre-operatively (pre-op), immediately after induction of hindlimb ischemia (day 0) and at the indicated time points after injection of ECFCs pretreated with vehicle or with the drug combination. Left: representative laser Doppler images (Isch., Ischemic leg; NI, non-ischemic leg). Right: perfusion ratios (Isch./NI) are indicated as mean values ± SEM (ECFCs: days 0–14, n = 7 mice; days 21–28, n = 3 mice. ECFCs treated: days 0–14, n = 8 mice; days 21–28, n = 4 mice). (C) EZH2 and HDAC inhibitors upregulate multiple transduction pathways in ECFCs. Schematic representation of cell-signaling molecules and processes that are upregulated (red) or downregulated (green) upon treatment of ECFCs with the GSK-343 and panobinostat combination. ^∗∗p < 0.01; ^∗p < 0.05. See also Figure S6.