Recent advances in single cell analysis have demonstrated genetic heterogeneity amongst cells from the same tumour population. The propagating compartment in acute lymphoblastic leukaemia (ALL) has been shown to undergo clonal evolution, leading to the co-existence of subclones with different copy number alterations or mutations. Certain subclones may respond differently to selective pressures such as therapy, resulting in clonal selection and relapse. However, little is known about how this apparent genetic heterogeneity translates to functional heterogeneity amongst the bulk population of ALL blasts, particularly under non-selective conditions. Although it is thought that a high proportion of blasts have the potential to engraft immunocompromised mice, it is unclear whether in practice the leukaemia is dominated by smaller numbers of clones and how this complexity changes over serial transplants or in selective environments.

To address these questions, we have used cellular barcoding as a fate mapping tool to track the engraftment and evolution of individual clones within a bulk tumour population. Cellular barcoding involves labelling cells with a random genetic barcode which is delivered using a lentiviral vector, so is both stable and heritable. We created a barcode library with a complexity in excess of 100,000 and used it to label ALL primograft leukaemia samples, which were subsequently transplanted serially into NOD/LtSz-scid IL-2Rγ null (NSG) mice.

Analysis of barcode composition in 3 primograft samples revealed that founder clones were highly abundant. Engraftment frequencies were estimated at between 1 in 2 to 1 in 14 (L4951, t(9;22)); 1 in 4 to 1 in 6 (L707, t(17;19)) and 1 in 19 (P929, t(4;11)). Following transplant of high cell numbers (10,000 barcoded), spleen samples contained high numbers of detectable barcodes (600-900). These contributed to the bulk population at similar frequencies, with no single barcode composing more than 5%. These clonal compositions were maintained in secondary recipients. This lack of dominance is suggestive of functional homogeneity within the propagating cell population under steady state conditions. In support of this, limiting dilution analysis revealed that all clones in the spleen have the capacity to engraft the leukaemia in the CNS and femoral bone marrow. We obtained several cases of near identical barcode composition in all sites, including one where a single barcode comprised the leukaemia in all organs.

Although all subclones had the capability to reconstitute the leukaemia in all organs, we also found evidence of clonal dominance under selective conditions. Small numbers of barcodes were often dominant in femur and CNS compartments, with single barcodes comprising up to 80% of the tumour population. This occurred more frequently at high transplant doses and was not found in the spleens, so may be due to competitive pressures in restrictive microenvironments such as the femur. These barcodes did not become dominant when samples were re-transplanted into secondary recipients, suggesting the dominance was due to stochastic founder effects as opposed to intrinsic properties of these clones. In contrast, selection of clones with increased fitness was observed under dexamethasone treatment in mice transplanted with sample L707. Treated mice had a substantial reduction in clonal complexity in spleen samples, with single barcodes comprising up to 85% of the tumour. For samples with lower initial complexity, dominant barcodes were the same in multiple mice, suggesting the existence of pre-existing clones with dexamethasone resistance.

In summary, we have demonstrated that many founder clones engraft and maintain the leukaemia in primary and secondary mice. This abundance of founder cells does not result in functional heterogeneity under constant selective conditions. Instead, the propagating cells are functionally homogenous, being able to contribute equally to leukaemia in the spleen and having the capability to reconstitute leukaemia in other organs. This homogeneity may reflect the typically low genomic instability in ALL. Clonal dominance only occurs under selective pressures, either stochastically through founder effects in potentially restrictive microenvironments, or selection in response to drug treatment.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution