Tracing the first hematopoietic stem cell generation in human embryo by single-cell RNA sequencing

Abstract

Tracing the emergence of the first hematopoietic stem cells (HSCs) in human embryos, particularly the scarce and transient precursors thereof, is so far challenging, largely due to the technical limitations and the material rarity. Here, using single-cell RNA sequencing, we constructed the first genome-scale gene expression landscape covering the entire course of endothelial-to-HSC transition during human embryogenesis. The transcriptomically defined HSC-primed hemogenic endothelial cells (HECs) were captured at Carnegie stage (CS) 12-14 in an unbiased way, showing an unambiguous feature of arterial endothelial cells (ECs) with the up-regulation of RUNX1, MYB and ANGPT1. Importantly, subcategorizing CD34⁺CD45^- ECs into a CD44⁺ population strikingly enriched HECs by over 10-fold. We further mapped the developmental path from arterial ECs via HSC-primed HECs to hematopoietic stem progenitor cells, and revealed a distinct expression pattern of genes that were transiently over-represented upon the hemogenic fate choice of arterial ECs, including EMCN, PROCR and RUNX1T1. We also uncovered another temporally and molecularly distinct intra-embryonic HEC population, which was detected mainly at earlier CS 10 and lacked the arterial feature. Finally, we revealed the cellular components of the putative aortic niche and potential cellular interactions acting on the HSC-primed HECs. The cellular and molecular programs that underlie the generation of the first HSCs from HECs in human embryos, together with the ability to distinguish the HSC-primed HECs from others, will shed light on the strategies for the production of clinically useful HSCs from pluripotent stem cells.

Fig. 1

Transcriptomic identification of different cell populations in human CS 13 dorsal aorta (DA) and the capture of HECs. a Identification of cell populations in CS 13 DA visualized by UMAP. Each dot represents one cell and colors represent cell clusters as indicated. b UMAP visualization of the expression of curated feature genes for the identification of cell clusters (CDH5, SPI1, PDGFRA and EPCAM). c Violin plots showing the expression of feature genes in each cell cluster. Colors represent cell clusters indicated in a. d Pie chart showing the percentages and absolute numbers of each cell cluster involved in CS 13 DA. e UMAP visualization of AEC, HEC and HC clusters, resulted from sub-dividing the cells in AEC and Hem clusters described in a as indicated in the lower right frame. f Heatmap showing the scaled expression of top 10 differentially expressed genes (DEGs) in AEC, HEC and HC clusters. g The major Gene Ontology biological process (GO:BP) terms in which DEGs are enriched for each cluster. h Dot plots showing the scaled expression level of top 10 significantly differentially expressed surface marker genes in AEC, HEC and HC. Note no specific marker for HEC and only nine genes for HC cluster met the criteria. Colors represent the scaled expression and size encodes the proportion of gene-expressing cells

Fig. 2

Capture and further analysis of HECs in CD44⁺ ECs from CS 12/13/14 embryos. a Sorting strategy of CD44⁺ EC and CD44⁻ EC. b UMAP with phenotypically different populations (left panel) and two transcriptionally distinct clusters (right panel) mapped on it. c UMAP plots displaying the expression of hematopoietic and endothelial genes. d Heatmaps showing the average expressions of top 10 differentially expressed surface markers and TFs between aEC and vEC clusters. e The aEC cluster is divided into two sub-clusters. The expression of their feature genes is shown on the violin plots to the right. f Dot plots showing top 10 DEGs in the two sub-clusters. g Enriched GO terms in CXCR4⁺ aEC and HEC, respectively. h Bar plots displaying top genes positively correlated with RUNX1 in aEC cluster. Genes related to ribosome biogenesis were removed from the gene list. i PCA plot showing expression of endothelial and arterial genes and representative genes from h in aEC cluster. HEC shares endothelial and arterial features with CXCR4⁺ aEC. Hematopoietic genes correlated with RUNX1 are enriched in HEC

Fig. 3

The developmental path from arterial ECs via HSC-primed HECs to HSPCs in the AGM region. a Sorting strategy of CD235a⁻CD45⁺CD34⁺ hematopoietic progenitors in CS 15 dorsal aorta. b Identities of five cell populations in CS 15 dorsal aorta visualized by UMAP. c Violin plots showing the expression of feature genes in each cell cluster. d Heatmap showing the average expressions of top 10 DEGs expressed in HSPC (HSPC1/2/3), Myeloid progenitor and Lymphoid progenitor clusters. e Heatmap showing the average expressions of top 10 DEGs in HSPC1 (GJA5⁺ HSPC), HSPC2 (Cycling HSPC) and HSPC3 (GFI1B⁺ HSPC) clusters. f PCA plot of vEC, CXCR4⁺ aEC, HEC, GJA5⁺ HSPC, Cycling HSPC and GFI1B⁺ HSPC. g Dot plots showing the scaled expression level of feature genes in the indicated clusters. Expression of four Notch signaling pathway genes is shown at the bottom. h Trajectory analysis by Monocle 2 combining two aEC sub-clusters (CXCR4⁺ aEC and HEC) from CS 12/13/14 with three HSPC clusters (GJA5⁺ HSPC, Cycling HSPC and GFI1B⁺ HSPC) from CS 15 indicates the developmental path from arterial ECs to HSPCs. Dynamic changes of proportion of clusters are shown on the bottom. i Four distinct gene expression patterns along the pseudotime axis inferred by Monocle 2. j Expression of representative genes of each pattern along pseudotime axis inferred by Monocle 2

Fig. 4

Different features and origins of the early and late HECs. a UMAP showing early endothelial and hematopoietic populations in CS 10–11 of human embryos. Early HEC lies more closely to endothelial cell populations. b Expression patterns of typical endothelial and hematopoietic genes in early populations (CS 10 and CS 11). c Heatmap showing the average expressions of top 5 DEGs enriched in distinct populations. d The expression of top 10 DEGs in early and late HEC. e GO:BP terms (left) and pathways (right) enriched in early HEC (upper) and late HEC (lower). f UMAP of endothelial and hematopoietic cells from early (CS 10 and CS 11) and late (CS 13) stages with different populations mapped on it. Two distinct HEC clusters at high magnification are shown to the lower right. The distribution of two HEC clusters and endothelial (EC) and hematopoietic (HC) populations on UMAP is shown as schematic to the upper right. g UMAP with the expressions of indicated genes mapped on it

Fig. 5

Computational analysis of the heterologous cellular interactions for the development of HSC-primed HECs. a The schematic diagram indicates the heterologous cellular interactions of ligand–receptor pairs between HEC and other clusters including EC, AEC, epithelial and four mesenchymal cell clusters. The direction of arrows indicates from the ligands to the corresponding receptors. Different colors of arrows indicate different databases for ligand–receptor analysis. The thickness of arrows indicates the relative number of ligand–receptor pairs detected. b Frequency of ligand–receptor genes involved in KEGG pathways. c Heatmap showing the average expression of molecules for each ligand–receptor pair in distinct stromal clusters when coupled with HEC. The hierarchical clustering result of cell populations (columns) and ligand–receptor pairs (rows) is shown. d The expression of ligand (blue) and its receptor (red) genes of indicated pairs. Dashed line frames indicate the populations in which the expression of the given gene met the criteria. e Dot plots showing the ligand–receptor pairs between EC/AEC as ligand-expressing cell and HEC as receptor-expressing cell

Fig. 6

The schematic showing the endothelial-to-hematopoietic transition (EHT) process in human early embryos. Different cell populations including two types of HECs, developmental stages, and cell type-specific marker genes along the endothelial-to-hematopoietic transition process identified in this study are shown. EC endothelial cells, HEC hemogenic endothelial cells, HSPC hematopoietic stem progenitor cells