Mutational mechanisms in childhood acute lymphoblastic leukemia with multiple relapses
Improvements in the treatment of pediatric acute lymphoblastic leukemia (ALL) have resulted in high overall survival rates, now approaching 90%. However, relapse still represents the most common cause of therapy failure in pediatric ALL, and better understanding of mechanisms driving relapse could aid further improvements in outcome. We recently found that pediatric ALL relapses have increased mutational loads, sometimes resulting in a hypermutation phenomenon, which appear to occur across childhood ALL subtypes, particularly in second relapses, and can be caused by different mutational mechanisms, including abberant AID/APOBEC activity and mismatch repair deficiency.
Mutations arise in single cells which subsequently grow out into (sub)clones and become detectable. Since ALL is a multiclonal disease, different mechanisms can be active simultenously or subsequently, which arise endogenously or be induced by therapy. We performed whole genome sequencing of 33 samples (diagnosis, relapses and complete remissions) from six leukemia patients and clustered the mutations that followed the same dynamics over time. These mutational clusters were used to identify endogenous and/or external mutational mechanisms in different subclones within each patient. By analyzing mutation types in their immediate context, which reveal typical mutational signatures, we determined active mutational mechanisms.
We defined up to six clusters per patient, with an average cluster size of 1,067 mutations [63-5.767] . We idenified at least five mutational mechanisms, including abberant activity of AID/APOBEC deaminases, and an incompletely understood mutational pattern associated with UV exposure previously also reported by others in ALL2. In four patients we observed mutational mechanisms only becoming active in rising clones at first or second relapse. In two patients we identified a relapse-specific mutational signature that was caused by thiopurine treatment. In one Down-syndrome patient who developed two relapses, we identified at least two different mutational mechanisms that were active at different stages of disease. One of these mechanisms (AID/APOBEC) was active in a falling clone at diagnosis and in rising subclones at first and second relapse, suggesting that this aberrant AID/APOBEC activity gave rise to new clones during the course of disease.
We have shown dynamic clonal evolutionary trajectories driven by multiple mutational mechanisms in children with ALL and mutliple relapses. Understanding of the causes and consequences of the mutational mechanisms driving relapse in ALL may provide opportunities for treatment.