The Hemogenic Competence of Endothelial Progenitors Is Restricted by Runx1 Silencing during Embryonic Development

Abstract

It is now well-established that hematopoietic stem cells (HSCs) and progenitor cells originate from a specialized subset of endothelium, termed hemogenic endothelium (HE), via an endothelial-to-hematopoietic transition. However, the molecular mechanisms determining which endothelial progenitors possess this hemogenic potential are currently unknown. Here, we investigated the changes in hemogenic potential in endothelial progenitors at the early stages of embryonic development. Using an ETV2::GFP reporter mouse to isolate emerging endothelial progenitors, we observed a dramatic decrease in hemogenic potential between embryonic day (E)7.5 and E8.5. At the molecular level, Runx1 is expressed at much lower levels in E8.5 intra-embryonic progenitors, while Bmi1 expression is increased. Remarkably, the ectopic expression of Runx1 in these progenitors fully restores their hemogenic potential, as does the suppression of BMI1 function. Altogether, our data demonstrate that hemogenic competency in recently specified endothelial progenitors is restrained through the active silencing of Runx1 expression.

Graphical abstract

Figure 1

ETV2::GFP-Expressing FLK1 Cells Have Different Hematopoietic Potentials at Different Developmental Stages (A) FACS analysis for the expression of endothelial and hematopoietic markers at the E7.5 and E8.5 developmental stages. EP: embryo proper; YS: yolk sac. (B) Schematic representation of the experimental design. (C) FACS analysis of the sorted GFP⁺FLK1⁺CD41⁻ cells after three days in HE culture conditions. (D) Clonogenic assay for hematopoietic progenitors of FLK1⁻GFP⁻CD41⁻, FLK1⁺GFP⁺CD41⁻, FLK1⁺GFP⁻CD41⁻, and FLK1⁺GFP⁺CD41⁺ populations after three days in culture. Error bars represent 1 SEM from three independent experiments. CFU, colony-forming units; GM, granulocyte-macrophage. (E) Representative images of hematopoietic colonies from semisolid clonogenic assays. Ery^P, primitive erythrocytes; Ery^D, definitive erythrocytes; Mac, macrophages, Mix, granulocyte-erythroid-monocyte-macrophage (GEMM)/granulocyte-erythroid-macrophage (GEM). See also Figure S1.

Figure 2

Live Imaging and Microarray Analysis Reveal Migratory Pattern and Differential Gene Expression of ETV2::GFP⁺ Cells at Different Developmental Stages (A) Still images from Movie S1 displaying the migratory path of some representative ETV2::GFP⁺ cells during the ex vivo culture of E7.5 embryos. Top: transmitted and GFP images of Z sections were stacked together to create a 3D reconstruction of the embryo. Bottom: tracking of ETV2::GFP⁺ cells through time. Color spots depict examples of some of the migrating ETV2::GFP⁺ cells. Scale bars, 50 μm, and time (t) is given in hr:min:sec. (B) Hierarchical clustering and heatmap for depicted genes in the GFP⁺FLK1⁺CD41⁻ cell population sorted from E7.5 and E8.5 (EP) ETV2::GFP embryos. (C) RT-qPCR for some of the differentially expressed genes found in the exon array. Three primer sets are used for detection of Runx1 isoforms: Runx1b, Runx1c, and Runx1b+c (Runx1all). Data are means of triplicate experiments ± SD; two-tailed Student’s t test. ^∗p < 0.05. (D) Exon array gene expression for hematopoietic transcription factors. Error bars indicate SD. (E) GSEA analysis depicting significantly enriched molecular processes in the gene set correlated with low gene expression in E8.5 compared to an E7.5 GFP⁺FLK1⁺CD41⁻ population. NES: normalized enrichment score; FDR: false discovery rate. See also Figure S2 and Movies S1, S2, S3, and S4.

Figure 3

Single-Cell Gene Expression Analysis Reveals Heterogeneity in ETV2::GFP-Expressing FLK1 Cells (A) The Fluidigm platform was used for single-cell gene expression analysis as described in the Supplemental Experimental Procedures. Unsupervised hierarchical clustering heatmap for depicted genes. (B) PCA for the cells analyzed shows the degree of segregation among the three populations. (C) Frequency of distribution of the expression level for each gene shown as Violin plots. The width of the violin plot represents the frequency of cells at that expression level. See also Figure S3.

Figure 4

Expression Pattern of RUNX1b and ETV2 in Developing Embryos (A and B) Transversal section of Etv2::gfp Runx1b::rfp E7.5 (A) and E8.5 (B) embryos stained for ETV2::GFP (green), Runx1b::RFP (red), and FLK1 (white). Scale bars, 50 μm. (C) Whole-mount imaging for ETV2::GFP RUNX1b::RFP and CD31 for E7.5-E10.5 developmental stages. Pr: Proximal; D: Distal; A: Anterior; P: Posterior; BI: blood island; H: head; ISV: intersomitic vessel; da: dorsal aorta; ec: endocardium; and P-Sp: para-aortic splanchnopleura region. See also Figure S4 and Movies S5 and S6.

Figure 5

BMP4-pSMAD1/5 Activity in E7.5 Underlies the High Hematopoietic Potential of the ETV2::GFP Cells (A) E7.5 ETV2::GFP embryos were stained for GFP (green) and BMP4 (red). BI: blood island; Al: allantois. (B) Serial section of the same E7.5 embryo used in (A) was used for staining with GFP and pSMAD1/5. (C) E8.5 ETV2::GFP embryos were stained for GFP (green) and BMP4 (red). (D) E8.5 ETV2::GFP embryo stained for pSMAD1/5 (red). Scale bars, 50 μm for whole-embryo images and 15 μm for magnified images.

Figure 6

RUNX1b Ectopic Expression in E8.5 ETV2::GFP⁺FLK1⁺ Cells Enhances Their Hematopoietic Potential (A) Schematic representation of the experimental design. (B) Time-lapse imaging during the three-day OP9 culture of cells transduced with control or Runx1 lentivirus. White arrows indicate endothelial clusters; red arrows indicate hematopoietic cells. (C) Clonogenic assay for hematopoietic progenitors of control and Runx1b transduced cells after three days of culture on OP9 stroma layer. Left graph shows the number of hematopoietic colonies formed for each type of hematopoietic CFU. Right graph shows the overall change in CFU number per treatment. Error bars represent 1 SD from three independent experiments. ^∗p < 0.05. Mix, GEMM/GEM; GM: granulocyte-macrophage; Mac: macrophages; Ery^P: primitive erythrocytes; and Ery^D: definitive erythrocytes. (D) Examples of hematopoietic colonies at day eight of replating in semisolid clonogenic assay. Scale bars represent 500 μm. (E) Left: limiting dilution assay (LDA) of hemogenic potential of FLK1⁺GFP⁺CD41⁻-sorted cells derived from E8.5 ETV2::GFP embryos and infected with either control or Runx1 lentivirus before plating on OP9 stoma. Right: example of hematopoietic cells emerging in OP9 cultures indicative of hemogenic endothelial potential in LDA. Data shown are representative of two independent experiments. See also Figure S5 and Movie S7.

Figure 7

Inhibition of BMI1 De-represses Runx1 Expression and Confers Hemogenic Potential to E8.5 ETV2::GFP Endothelial Progenitors (A) Single-cell gene expression analysis for Bmi1 in E7.5 and E8.5 (EP) ETV2::GFP⁺FLK1⁺CD41⁻ progenitors. Error bars represent 1 SEM for at least 54 single cells. ^∗∗p ≤ 0.001. (B) Western blot analysis for ubiquitinylated H2AK119, a downstream target of the PRC1 complex. FLK1⁺ sorted cells from day-3 embryoid body cultures were treated with 25 μM PRT4165 inhibitor or DMSO for 3 days prior to protein isolation. (C) Schematic representation of the experimental design. (D) RT-qPCR for E8.5 ETV2::GFP⁺FLK1⁺CD41⁻ cells cultured in HE conditions and treated with DMSO or 25 μM PRT4165 for 3 days prior to RNA extraction. Error bars represent 1 SEM for at least five experiments (expect for BMP4, n = 4). ^∗p < 0.05. (E) FACS analysis of day-8 hematopoietic culture for the expression of hematopoietic markers. (F) Emergence of hematopoietic colonies after 8 days in hematopoietic culture conditions. Scale bars, 500 μm. (G) Representative May-Giemsa staining from PRT4165-treated cells after 8 days in hematopoietic culture conditions. Gr: granulocytes; Mac: macrophages; Mk: megakaryocyte; P: progenitor, and scale bars, 20 μm. See also Figure S6.