The bone marrow microenvironment of pre-B acute lymphoblastic leukemia at single-cell resolution

Abstract

The bone marrow microenvironment (BMM) plays a key role in leukemia progression, but its molecular complexity in pre-B cell acute lymphoblastic leukemia (B-ALL), the most common cancer in children, remains poorly understood. To gain further insight, we used single-cell RNA sequencing to characterize the kinetics of the murine BMM during B-ALL progression. Normal pro- and pre-B cells were found to be the most affected at the earliest stages of disease and this was associated with changes in expression of genes regulated by the AP1-transcription factor complex and regulatory factors NELFE, MYC and BCL11A. Granulocyte-macrophage progenitors show reduced expression of the tumor suppressor long non-coding RNA Neat1 and disruptions in the rate of transcription. Intercellular communication networks revealed monocyte-dendritic precursors to be consistently active during B-ALL progression, with enriched processes including cytokine-mediated signaling pathway, neutrophil-mediated immunity and regulation of cell migration and proliferation. In addition, we confirmed that the hematopoietic stem and progenitor cell compartment was perturbed during leukemogenesis. These findings extend our understanding of the complexity of changes and molecular interactions among the normal cells of the BMM during B-ALL progression.

Figure 1

Inferred cell type clusters used in downstream analyses. t-SNE dimensionality reduction plots show all samples (a) and individual samples at each time point (b). Cells were clustered using Seurat and annotated by SingleR. ST-HSC & MLP = short term repopulating cells and multilineage progenitors, GMP = granulocyte–macrophage progenitors, MDP = monocyte-dendritic precursors.

Figure 2

Number of differentially expressed genes by cell type and direction of change (↓ = underexpressed and ↑ = overexpressed) during disease progression. Results are adjusted for multiple testing (Bonferroni correction) and genes have an absolute log fold change of at least 0.4. ST-HSC & MLP = short term repopulating cells and multilineage progenitors, GMP = granulocyte–macrophage progenitors, MDP = monocyte-dendritic precursors.

Figure 3

Intercellular communication networks of known ligand/receptor pairs. Edge widths correspond to the number of links between cell pairs or within the same cell type. A link is one-way directional and is defined when a cell type expresses the ligand and another cell type (or the same cell type) expresses the corresponding receptor. ST-HSC & MLP = short term repopulating cells and multilineage progenitors, GMP = granulocyte–macrophage progenitors, MDP = monocyte-dendritic precursors.

Figure 4

RNA velocity estimates for each cell type over time. Estimates (arrows) are plotted on a Uniform Manifold Approximation and Projection (UMAP) dimensionality reduction. The length of the arrows represents the transcription rate (velocity) and the direction of the arrows points to the inferred future state of the cell based on other cells present on the UMAP plot. Transparent points are those cells that were filtered out as described in “Results” (cell types in the bone marrow microenvironment). ST-HSC & MLP = short term repopulating cells and multilineage progenitors, GMP = granulocyte–macrophage progenitors, MDP = monocyte-dendritic precursors.

Figure 5

B-ALL alters the HSPC compartment in the bone marrow of leukemia-bearing mice. (a) Percentage (left) and number (right) of lineage negative (lin⁻) cells. (b) Percentage of LSK cells and myeloid progenitors. (c) Percentage of HSCs. (d) Percentage of multipotent progenitors. (e) Percentage of CMPs, GMPs, MEPs and CLPs. (f) Red blood cell count, hemoglobin, hematocrit and platelet count in the peripheral blood of leukemia-bearing mice (n = 3–8). (a–e) n = 5–7. Throughout, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Error bars represent mean ± SEM.