Acute B lymphoblastic leukaemia-propagating cells are present at high frequency in diverse lymphoblast populations

Abstract

Leukaemia-propagating cells are more frequent in high-risk acute B lymphoblastic leukaemia than in many malignancies that follow a hierarchical cancer stem cell model. It is unclear whether this characteristic can be more universally applied to patients from non-'high-risk' sub-groups and across a broad range of cellular immunophenotypes. Here, we demonstrate in a wide range of primary patient samples and patient samples previously passaged through mice that leukaemia-propagating cells are found in all populations defined by high or low expression of the lymphoid differentiation markers CD10, CD20 or CD34. The frequency of leukaemia-propagating cells and their engraftment kinetics do not differ between these populations. Transcriptomic analysis of CD34(high) and CD34(low) blasts establishes their difference and their similarity to comparable normal progenitors at different stages of B-cell development. However, consistent with the functional similarity of these populations, expression signatures characteristic of leukaemia propagating cells in acute myeloid leukaemia fail to distinguish between the different populations. Together, these findings suggest that there is no stem cell hierarchy in acute B lymphoblastic leukaemia.

Figure 1. B-cell differentiation markers and engraftment potential

1. Expression of the cell surface antigens CD10, CD19, CD20 and CD34 during normal human B-cell development (van Zelm et al, 2005).

2. Percentage of engrafting primary patient and primograft samples for the different populations (CD10^low/high, CD20^low/high, CD34^low/high). Black bars provide the data for primary and white bars for primograft samples.

3. Weighted mean of engraftment in mice transplanted with at least 1000 blasts sorted for CD10, CD20 or CD34 expression. The analysis takes into account that different numbers of mice were transplanted with cells from different samples. For data of engraftment of mice transplanted with cells sorted for expression of CD10 and CD20, see also Supporting Information Tables S1 and S2. The CD34 data are taken from the limiting dilutions experiments (Table 2) and include only the mice transplanted with 1000 cells.

Figure 2. Phenotype of the re-established leukaemia after transplantation of purified B-ALL blast populations

1. Engraftment of primografted CD10^low and CD10^high cells (patient L4951).

2. Engraftment of primografted CD20^low and CD20^high cells (patient EMCR1).

3. Engraftment of primografted CD34^low and CD34^high cells (patient EMCR1).

4. Engraftment of primary CD34^low and CD34^high cells (patient L784).

Figure 3. Self-renewal gene expression in CD34^high and CD34^low B-ALL blasts

1. Expression of TERT in CD34^high and CD34^low umbilical cord blood cells and leukaemic blasts. Umbilical cord blood shows higher expression of TERT in immature CD34^high cells whilst no difference is seen between immature CD34^high and CD34^low blasts. TERT expression in blasts is comparable with, or higher than, that in CD34^high umbilical cord blood progenitor cells. Expression values are 2^-ΔCt using GAPDH as the reference gene. For purity of cells sorted for CD34 expression, see Supporting Information Fig S6.

2. The PCA using a haematopoietic “self-renewal” signature containing 59 genes (213 probes) (Kim et al, 2009) also failed to separate CD34^high and CD34^low B-ALL populations (Supporting Information Fig S4). Here, we show the heatmap of the expression of 41 selected probes from this “self-renewal” signature.

3. PCA plot generated using an AML “stem cell” signature (Eppert et al, 2011) that separates Lin-CD34^highCD38^low blasts (dark purple symbols) from more mature CD34^low (pink symbols) blasts in AML (Gentles et al, 2010) but fails to detect any significant differences in “self-renewal” gene expression amongst the leukaemic subpopulations (black & white symbols) in B-ALL.

Figure 4. Engraftment kinetics of purified B-ALL subpopulation as assessed by sequential bone marrow punctures

Mice from all limiting dilution experiments were pooled, as the number of mice in each individual transplantation experiment was too low for the analysis. Data are mean percentage of engraftment with bars showing standard error of the mean.

1. A,B. Engraftment kinetics of CD10^high and CD10^low cells after transplantation of 100 (n = 4) or 1000–2000 blasts (n ≥ 4).

2. C,D. Engraftment kinetics of CD20^high and CD20^low cells after transplantation of 100–300 (n ≥ 8) or 500–3000 blasts (n ≥ 21).

3. E,F. Engraftment kinetics of CD34^high and CD34^low cells after transplantation of 100 (n ≥ 6, last points n = 2 for CD34^high and n = 3 for CD34^low) or 1000 blasts (n ≥ 9).

Figure 5. Models for regulation of self-renewal, proliferation and differentiation in normal haematopoietic stem cells and acute myeloid leukaemia as compared with acute lymphoblastic leukaemia

1. In normal haematopoietic stem cells and AML, self-renewal is tightly regulated by a differentiation-dependent loss of self-renewal.

2. In B-ALL, clonal expansion is tightly regulated by the dependence of positive survival and proliferation signals. Without those lymphoid blasts undergo apoptosis.