Studies of congenital genetic aberrations behind childhood leukemia

Abstract

Acute lymphoblastic leukemia (ALL) is the most common cancer in childhood, and most frequently (85%) of B-cell precursor type (BCP-ALL). Acquired chromosomal rearrangements or aneuploidies are the recurrent, often prenatal, initiators of BCP-ALL. These aberrations define distinct molecular subtypes that are associated with differences in prognosis and used to guide treatment. Initiating variants are disease driving, but secondary variants are required to drive progression to overt disease. Although constitutional predisposing variants are found in an increasing share of cases (10-18%), BCP-ALL etiology remains largely unknown. Recent studies have suggested that exposure to common infections may modulate progression of BCP-ALL.

The aim of this thesis was to identify, asses and quantify congenital genetic aberrations behind childhood BCP-ALL predisposition and initiation, as well as to characterize subsequent clonal evolution and identify drivers of progression to overt disease. To this end, we performed whole genome sequencing (WGS) to identify constitutional BCP-ALL-predisposing variants. In paper I, we reported familial predisposition mediated by a constitutional t(12;14), where haploinsufficiency of the powerful transcription factor ETV6 was suggested to cause predisposition. In paper III, monozygotic twins with concordant BCP-ALL shared a constitutional, maternally inherited, novel variant in NF1, predicted to be highly damaging. As none of the carriers has any clinical sign of the cancer syndrome neurofibromatosis type 1 (NF1), we classified the variant to be of unknown significance (VUS), but speculated its possible BCP-ALL-predisposing effect.

We developed a sensitive and quantitative method for backtracking BCP-ALL to pre-leukemic clones (paper II), applying chip dPCR in combination with WGS to analyze DNA from neonatal dried blood spots. In paper II, only one case of BCP-ALL, diagnosed at age 1 month, had detectable copy numbers of genomic breakpoint sequence at birth. Failed detection in the remaining six cases was suggested to be caused by technical and sample related limitations, and less frequently postnatal initiation. In paper III, WGS identified a shared somatic complex rearrangement, generating ETV6- RUNX1, in the BCP-ALLs of monozygotic twins. Detection at birth by dPCR failed, but identical breakpoint sequences confirmed its prenatal origin. Surprisingly, a shared (prenatal) deletion in UBA2 was found to precede the complex rearrangement, persisting after several years in remission.

Clonal evolution of concordant BCP-ALLs was characterized in paper III, detecting shared and unique overlapping secondary putative driver variants, supporting independent although convergent clonal evolution. In paper I, 7-10 secondary putative driver variants, in genes recurrently targeted in childhood ALL, were identified in BCP-ALLs with ETV6-mediated predisposition. This further supported that secondary drivers are required for progression, although phylogenetics of somatic events in ETV6-predisposed cases remains to be delineated. In paper IV, we assessed a Swedish population-based cohort of 1380 BCP-ALL cases and used GARIMAX to demonstrate informative seasonal variation in onset and interpreted peak onset to fall in August. Four explanatory models, related to exposure to common infections as a driver of final progression to overt disease, were suggested. The likelihood of each model depends on still unknown induction time of childhood BCP- ALL.

Together, these studies add to our understanding of; congenital susceptibility to BCP-ALL through constitutional predisposing variants and prenatally initiated pre-leukemic clones, progression to overt disease through somatic clonal evolution and the genetic and environmental drivers of this process.