Clonal Competition of Cancer Cells During Acute Lymphoblastic Leukemia Treatment

The importance of cancer cell heterogeneity is increasingly recognized. However, the interaction of different cancer cells remains poorly understood, particularly in response to therapeutic treatment. Past studies used naturally occurring genetic mutations to examine the heterogeneity and evolutionary trajectories of cancer cells. As these mutations occur at different times, clones marked by these mutations cannot be directly compared to assess cellular interactions. In this study, we use engineered genetic barcodes to track patient derived acute lymphoid leukemia (ALL) cells in a novel mouse model that closely mimics the 'intensive' and 'maintenance' chemo-therapy regiments used in clinical treatment.

Intensive and maintenance chemotherapy are commonly combined together in ALL treatment. After the combination treatment, ALL patients who relapse early (<36 months from diagnosis) are particularly resistant to additional chemotherapy and have a very poor prognosis. In contrast, a fraction of ALL patients who relapse late (>36 months from diagnosis) have a much better prognosis, and remain chemo-sensitive. It is unclear why ALL patients who relapse late are chemo-sensitive while patients who relapse early are chemo-resistant. We hypothesize that low intensity maintenance therapy allows "beneficial competitor cells" to survive and outcompete chemo resistant cancer clones. These beneficial competitor cells expand slowly during the maintenance treatment. They are chemo sensitive and are easily eradicated by intensive therapy.

To test our hypothesis, we took primary human B-ALL cells and genetically barcoded them before transplantation into immune deficient NSG mice. We established a chemotherapy treatment protocol in mice that closely resembles the 'intensive' and 'maintenance' phase of ALL treatment currently used in clinic. In our experiment, mice were first treated with a four-week 'intensive' phase of chemotherapy consisting of Vincristine, Dexamethasone and PEG L-Asparaginase. This intensive phase was followed by a long-term 'maintenance' chemotherapy phase consisting of Methotrexate.

We found that 'intensive' therapy did not significantly change the clonal composition in mice. However, 'new' leukemia clones arose during 'maintenance' treatment. These clones might have been suppressed during the undisturbed progression of the disease, and they may play a crucial role in the relapse of the disease during the long-term 'maintenance' treatment.

To investigate the growth phenotype of chemotherapy surviving ALL clones, we used a competitive transplantation assay. We took surviving ALL clones from 'intensive' only and 'intensive + maintenance' treated xenografts and co-transplanted these cells with untreated ALL cells. As predicted by our hypothesis, we found that ALL clones that survived 'intensive' only treatment exhibited a more aggressive growth phenotype compared to untreated ALL clones. Conversely, ALL clones that survived 'intensive + maintenance' treatment grew more slowly than untreated clones. More importantly, the growth of the co-transplanted untreated ALL clones was significantly suppressed by the presence of ALL clones that had survived intensive treatment. In contrast, the growth of the co-transplanted untreated ALL clones was significantly accelerated in the presence of ALL clones that survive the 'intensive + maintenance' treatment, suggesting that new clones that expanded during the long-term maintenance treatment were outcompeted by other clones in the absence of maintenance treatment.

Together, our findings suggest that the clonal competition of cancer cells plays an important role in determining treatment outcome. In particular, the benefit of maintenance therapy to ALL patients may arise from cellular competition. Further characterization of the beneficial competitor cells may help to improve ALL treatment and may also help to improve the treatment of other types of cancer.