RAG-Mediated Recombination Is The Predominant Driver Of Oncogenic Rearrangement In ETV6-RUNX1 Acute Lymphoblastic Leukemia

Approximately 25% of pediatric B-cell precursor ALL are characterized by the ETV6-RUNX1 fusion and are associated with a favorable prognosis. Monozygotic twin studies with concordant ALL and ‘backtracking’ studies using archived neonatal blood spots established that ETV6-RUNX1 is an initiating event arising prenatally in a committed B-cell progenitor. However, the fusion gene is not sufficient on its own to cause overt leukemia and a number of subsequent studies have now provided strong evidence that additional mutations are essential for the clinical development of ALL.

To obtain a detailed understanding of genetic events that drive this subtype of ALL, we carried out extensive genomic analysis in patients with ETV6-RUNX1 ALL. A total of 57 cases with diagnostic (leukemic) cell DNA paired with matched, remission samples as a source of constitutive DNA were used for exome sequencing (n=56) and low-depth wholegenome sequencing for structural variation (SV) analysis (n=51). Integrative analysis of exome and wholegenome data was performed and data were evaluated for recurrent gene mutations, copy number alterations and genomic rearrangements. Signatures of somatic mutation and structural variation were studied for insights into operative mutational processes and exposures that may drive ETV6-RUNX1pathogenesis.

Wholegenome profiling identified 524 somatic SVs (average:11, range 0-49) including 34 tandem duplications, 66 inversions, 106 intrachromosomal translocations and 317 deletions. Resolution of the breakpoint junction to the base-pair level was possible for 354 SVs and in 40% of these breakpoints, we identified a significant enrichment of V(D)J recombination signal sequences or motifs that closely approximate these (E-value=3.8x10^-116). Correlation with chromatin marks reveals a ten-fold genomewide enrichment of SVs (p<2.2x10^-16) at promoters and enhancers of genes actively transcribed in early B-lineage development. Overall, 83% of resolved rearrangements mapping to promoters and 76% of rearrangements mapping to enhancer regions were in close proximity to a V(D)J recombination sequence motif as opposed to an average of 29% of the rearrangements in other chromatin states. These findings suggest that RAG-mediated genomic rearrangement is the dominant mutational process driving SV in ETV6-RUNX1 ALL. Single-cell tracking of genomic rearrangements bearing the hallmarks of aberrant V(D)J targeting in two patients, shows that these can occur multiple times and in distinct clones. This suggests that this mechanism is not restricted to one founder cell but is rather continuous during leukemic evolution and likely to contribute to the branching clonal phylogenies observed in ETV6-RUNX1ALL at presentation.

Exome sequencing analysis confirmed 775 somatic substitutions and 16 indels across 715 protein coding genes. Oncogenic mutations in KRAS, NRAS, CTCF, DAXX, EZH2, and KDM6A are described as well as components of the cohesion complex including STAG2, SMC5 and SMC1A. Integrative analysis of whole-genome and exome sequencing data identified ATF7IP, MGA and STAG2 to be recurrently affected by both deletions or inactivating mutations and highlights their role as novel tumor suppressors in ETV6-RUNX1 ALL.

Sequence context analyses on the acquired point mutations identifies two distinct mutational signatures that are operative in ETV6-RUNX1 ALL, signature A characterized by C>T transitions at CpGs and signature B characterized by C>G and C>T at TpCs, contributing 56% and 32% of all substitutions respectively. Signature B is consistent with the reported preference of the APOBEC family of enzymes and is often observed in patients presenting with a hypermutator phenotype and dense clustering of mutations.

We report the identification of a novel spectrum of somatic mutations in ETV6-RUNX1 ALL and present the first detailed characterization of the genomic landscape of this ALL subtype. Our analysis shows that the critical secondary events to the ETV6-RUNX1 fusion gene formation are genomic rearrangements mostly driven by aberrant V(D)J recombination events targeting of actively transcribed genes during early B-Cell development. This represents a unique mechanism of genomic instability in cancer, that specifically targets the very genes required for normal B-Cell development and survival, which may in part explain the positive therapeutic response in this subtype of ALL.