A Human IPS Model Implicates Embryonic B-Myeloid Fate Restriction as Developmental Susceptibility to B Acute Lymphoblastic Leukemia-Associated ETV6-RUNX1

Abstract

ETV6-RUNX1 is associated with childhood acute B-lymphoblastic leukemia (cALL) functioning as a first-hit mutation that initiates a clinically silent pre-leukemia in utero. Because lineage commitment hierarchies differ between embryo and adult, and the impact of oncogenes is cell-context dependent, we hypothesized that the childhood affiliation of ETV6-RUNX1 cALL reflects its origins in a progenitor unique to embryonic life. We characterize the first emerging B cells in first-trimester human embryos, identifying a developmentally restricted CD19^-IL-7R⁺ progenitor compartment, which transitions from a myeloid to lymphoid program during ontogeny. This developmental series is recapitulated in differentiating human pluripotent stem cells (hPSCs), thereby providing a model for the initiation of cALL. Genome-engineered hPSCs expressing ETV6-RUNX1 from the endogenous ETV6 locus show expansion of the CD19^-IL-7R⁺ compartment, show a partial block in B lineage commitment, and produce proB cells with aberrant myeloid gene expression signatures and potential: features (collectively) consistent with a pre-leukemic state.

Keywords: B cell; CRISPR/Cas9; ETV6-RUNX1; acute lymphoblastic leukemia; genome engineering; human fetal lymphopoiesis; human pluripotent stem cells; in vitro B cell differentiation.

Figure 1

The Majority of Emerging B Cells in the Human Fetal Liver Express IL-7R (A) Major hematopoietic sites in the developing human embryo. Timescale shows days of gestation and embryonic development by Carnegie stage (CS) (Tavian and Peault, 2005). (B) CD19⁺ B cells from human fetal livers (FLs) at CS17 and CS21-22, cord blood (CB) and adult bone marrow (BM) were analyzed for surface expression of IL-7R. Viable cells were gated CD45⁺LIN⁻, further gating as indicated. Top panel: mean percentage CD19⁺ B cells of total CD45⁺ cells. Bottom panel: mean percentage of CD19⁺ B cells expressing IL-7R. (C) Emergence of CD19⁺ B cells as percent of total CD45⁺ cells (top panel) and percent CD19⁺ B cells expressing IL-7R (bottom panel) at different time points of development. Each dot represents one biological sample. Mean with range. (D) Quantitative gene-expression analysis of CD34⁺CD19⁺ proB cells (CS 20–21 FL and CB). Data are presented relative to GAPDH; mean ± SD, n = 2–3. See also Figure S1.

Figure 2

Identification of a CD19^–IL-7R⁺ Progenitor (A) Fluorescence-activated cell sorting (FACS) analysis of CS17 FL showing CD19⁻IL-7R⁺ progenitor co-expressing CD34 and CD45RA. Adult BM (right) is shown for comparison. Cells were gated CD45⁺LIN⁻CD19^–. Further gating as indicated. Mean percentage of total CD45⁺CD34⁺ cells. (B) Kinetics of IL-7R progenitor emergence during development as percentage of total CD45⁺CD34⁺ cells. Each dot represents one biological sample. Mean with range. Statistical comparison versus CB control. (C) B cell potential of IL-7R⁺KIT⁺ FL progenitors was tested by co-culture on MS-5 for 2 weeks. Percentage of CD19⁺ B cells of total CD45⁺ cells is shown (left) with representative FACS analysis (right). CD34⁺ cells were used as control. Mean ± SD, n = 2–3. (D) Agarose gel of IGH DJ rearrangements in IL-7R⁺ FL progenitor. Top: D_H7J_H recombination assay. Bottom: D_H1-6J_H recombination assay. Label for lanes 1–6 is shown to the right. Marker lanes (M). Predicted D_H7J_H recombination between 100 and 130 bp. Predicted D_H1-6J_H rearrangements between 110 and 290 bp, n = 1 (van Dongen et al., 2003). (E) GM and erythroid potential of IL-7R⁺KIT⁺ FL progenitor in semi-solid media. Data show distribution of colony types as percentage of plated cells versus CD34⁺ cells as controls. Mean ± SD, n = 2 per developmental stage. (F) Myeloid potential of IL-7R⁺KIT⁺ FL progenitor in liquid culture (20–40 cells/well). Graph shows percentage of wells expanded to >50 cells (left). Cytospin of macrophages generated from FL IL-7R⁺KIT⁺ progenitor (right). Mean ± SD, n = 3. See also Figure S2.

Figure 3

The IL-7R Progenitor Transitions from a Myeloid to a Lymphoid Gene Expression Signature during Early Embryonic Development (A) Single-cell qPCR analysis of IL-7R⁺KIT⁺ progenitor from different developmental stages; FL CS17 (left), FL CS20 (middle), and adult BM (right). Each column represents a single cell, colored by cycle threshold (CT) value. Only cells expressing GAPDH are shown. Gene sets: red, myeloid; blue, lymphoid; green, B lineage related; 44–60 cells and n = 3 per developmental stage. (B) Agarose gel of D_H7J_H rearrangement in myeloid cells derived from myeloid liquid culture of CS20 IL-7R⁺ progenitors (also sorted positive for KIT), n = 1; marker lanes (M) 1 kb+ (Invitrogen). Predicted D_H7J_H recombination band 100–130 bp (van Dongen et al., 2003). (C) Co-expression of lineage-associated genes in IL-7R⁺KIT⁺ progenitor based on single-cell qPCR. Cells are considered myeloid (GM), lymphoid (L), or B-primed (B) if expressing two genes representative of that lineage, as defined in Figure 3A. Co-expression of two or more of each of the GM and L and or B genes are labeled GM/L and GM/B, respectively. Only cells expressing GAPDH and IL7R were analyzed. (D) PCA from single-cell qPCR data of IL-7R⁺KIT⁺ progenitor (triangles) and proB cells (squares) from FL (CS17 [progenitor only] and CS20), CB, and adult BM. Each dot represents a single cell. Right graphs show the direction and magnitude of eigenvectors contributing to separation for lineage-associated and most differentially expressed genes, respectively. Thirty-six genes were used for the PCA, n = 2–3 per developmental stage. Ellipses represent region of 90% confidence for each cellular compartment.

Figure 4

Lymphoid Commitment in hPSC Culture Recapitulates Human Fetal B Lymphopoiesis (A) hPSC in vitro B cell differentiation protocol. hPSCs are harvested onto a layer of overgrown OP9 and co-cultured for 10 days. After CD34 enrichment, cells were either analyzed (D10) or co-cultured for an additional 21 days on MS-5 in the presence of lymphoid cytokines as indicated (D31) (Carpenter et al., 2011). (B) Differentiated MIFF3 hPSCs were analyzed for CD19⁺ B cells (top) and CD19^–IL-7R⁺ progenitor (bottom) at D10 (left) and D31 (right). Cells were gated CD45⁺ or CD45⁺IGM⁻; further gating as indicated. Mean percentages of CD45⁺ cells, n = 5–7. (C) qPCR of D31 hPSCs. CD34⁺CD45RA⁻, IL-7R⁺ progenitor, and CD34⁺CD19⁺ proB cells were analyzed for expression of lymphoid (IL7R, RAG2, EBF1, and PAX5), megakaryocytic/erythroid (GATA1), and fetal-specific (LIN28B) genes. Data are normalized to GAPDH, mean ± SD, n = 2–3. (D) PCA of RNA-seq data. Left: PCA based on primary cells only; LIN⁻CD34⁺CD45RA⁻, IL-7R⁺ progenitor, and proB cells from adult BM (n = 3) and FL (CS21-22, n = 2). Right: hPSC-derived cells and the early CS17 FL IL-7R⁺ progenitor (outlier) were placed on the PCA calculation shown to the left. hPSCs (MIFF3 and H1) are from D31 (CD34⁺CD45RA⁻, IL-7R⁺ progenitor, and proB cells) of differentiation. PCA1 and PCA2 are shown and each dot represents one sample. The ellipses show the 90% density function for that cell type. (E) Single-cell qPCR analysis of IL-7R⁺ progenitor from D10 and D31 of differentiation (MIFF3). Each column represents a single cell. Colored by CT value. Only cells expressing GAPDH are shown. Genes in red are myeloid, blue are lymphoid, and green are B cell genes. A total of 35–58 cells per time point, n = 2/population. (F) PCA calculated from single-cell qPCR data from the human primary cells (Figure 3D) and differentiated hPSCs (MIFF3 and H1). hPSCs are from D10 (IL-7R⁺ progenitor only) and D31 (IL-7R⁺ progenitor and proB cells). Each dot represents a single cell (only hPSCs shown). Highlighted ellipses represent primary human cell clusters, 36 genes were used for the PCA, n = 2–3. See also Figure S3.

Figure 5

An hIPSC Model of ETV6-RUNX1 Shows a Block in B Lineage Commitment (A) Genome engineering strategy. A constitutive knockin cassette, encoding a splice acceptor (SA), cDNA for RUNX1 exons II–VIII V5 tagged at the C terminus, and linked to the mVenus fluorescent reporter by a self-cleaving furin/T2A peptide is inserted by CRISPR-directed homologous recombination toward the 5′ end of ETV6 intron V. The cassette is flanked by LoxP sites and includes a 3′ triple stop/poly(A) tail followed by an FLP floxable NeoR⁺ selection cassette. (B) Control MIFF3 (top panel), ETV6-RUNX1 hPSCs (middle panel) and reverted ETV6-RUNX1 hPSCs (bottom panel) were cultured according to Figure 4A and analyzed by flow cytometry for Venus reporter (left), proB (CD34⁺CD19⁺) and preB (CD34⁻CD19⁺) cells (middle), and IL-7R⁺ progenitor (right). Viable cells were gated as indicated. (C) Frequencies of IL-7R⁺ progenitor, proB, and preB cells in MIFF3 cells compared with ETV6-RUNX1 iPSCs and reverted ETV6-RUNX1 hPSCs analyzed at D27-31. Percentage of Venus⁻ (MIFF3/reverted clone) or Venus⁺ (ETV6-RUNX1 hIPSC) CD45⁺ blood cells. Each dot represents one replicate. Only samples with 8,000 cells or more were considered. Mean ± SD. Statistics were performed using MIFF3 as control group, ns, not significant. (D) Schematic drawing of in vitro competitive assay. Equal numbers of ETV6-RUNX1 and wild-type MIFF3-derived CD34⁺ cells from D10 of OP9 co-culture were seeded onto MS5 in B differentiating conditions. The proportions of Venus⁺ and Venus⁻ preB cell and IL-7R progenitors were analyzed by FACS at D29-31. (E) Differentiated MIFF3 and ETV6-RUNX1⁺ hPSCs, grown in competition, were analyzed for CD34⁻CD19⁺ B cells (top) and CD19⁻IL-7R⁺ progenitor (bottom) at D29–D31. Cells were gated CD45⁺; further gating as indicated. FACS plot of representative sample annotated with mean percentages values. Summary bar charts show mean ± SD, ^∗p < 0.05, n = 4. See also Figure S4.

Figure 6

ETV6-RUNX1 “proB” Cells Exhibit Multilineage Priming and Potential (A) RNA-seq data from ETV6-RUNX1-expressing iPSCs overlaid (haloed) on the existing PCA map calculated from primary samples (Figure 4D). Control hPSC-derived data (haloed) are shown for comparison (left). PCA also including reverted ETV6-RUNX1 IPS (pink) (right). PC1 and PC2 are shown; each dot represents one sample. The ellipses show the 90% density function for that cell type. Two different ETV6-RUNX1 cell lines (no. 2.1 and no. 2.8) used in thre experiments. Reverted clone, n = 2, in one experiment. (B) Comparison of gene expression in MIFF3/reverted ETV6-RUNX1 (RC) (green) and ETV6-RUNX1 hIPSC (purple) proB cells. Each dot represents one sample. Bars show mean fragments per kilobase of transcript per million fragments mapped (FPKM) value ± SD, n = 4–5. (C) Gene set enrichment analysis (GSEA) of STAT5A target gene expression comparing ETV6-RUNX1 hIPSC (red) with control proB cells at D31 of differentiation. NES, normalized enrichment score. FDR, false discovery rate. (D) GSEA as in (C). Lineage affiliations of gene sets used are indicated above (Laurenti et al., 2013). (E) Single-cell qPCR data of control MIFF3 and ETV6-RUNX1⁺ proB cells from D31. Each column represents a single cell. Colored by CT value. Genes labeled in red are myeloid, blue are lymphoid, and green are B cell genes. Only cells expressing GADPH and ETV6-RUNX1 are shown. A total of 37–56 cells investigated per population, n = 2–3. (F) Survival of MIFF3/reverted ETV6-RUNX1 (RC) (green) and ETV6-RUNX1 Venus (V)⁺ hIPSC (purple)-derived single proB cells (CD34⁺CD19⁺) grown in liquid culture supplemented with myeloid cytokines (percentage of wells with ≥3 cells at 14 days). Mean ± SD, n = ETV6-RUNX1 = 4; MIFF3/reverted = 6, in two experiments. (G) Myeloid differentiation potential MIFF3/reverted ETV6-RUNX1 (RC) (green) and ETV6-RUNX1 Venus (V)⁺ (purple)-derived single proB cells (CD34⁺CD19⁺) grown in liquid culture supplemented with myeloid cytokines (percentage of wells with ≥20 cells at 14 days). Mean ± SD, n = ETV6-RUNX1 = 4; MIFF3/reverted = 6, in two experiments. (H) Cytospin of macrophages generated from ETV6-RUNX1⁺ proB cells. (I) Agarose gel of D_H7J_H rearrangements in myeloid cells derived from liquid culture of ETV6-RUNX1 proB cells (clones no. 2.1 and no. 2.8) and primitive ETV6-RUNX1⁺ CD34⁺ cells. Marker lanes (M) 1 kb+ (Invitrogen). D_H7J_H recombination predicted band 100–130 bp (van Dongen et al., 2003). See also Figure S5.