We used whole exome and RNA sequencing of chronic myeloid leukaemia (CML) patients at diagnosis (n=49), myeloid blast crisis (MBC, n=19) and lymphoid blast crisis (LBC, n=20) to investigate mutagenic drivers associated with leukaemia progression (Thomson et al., 2020, Leukemia; Branford et al., 2018, Blood). In addition to the hallmark BCR-ABL1 fusion gene, patients at blast crisis had frequent structural variation, particularly large genomic deletions at LBC. Analysis of mutational burden and mutational signature, as defined by single nucleotide mutations, showed little difference between diagnosis to blast crisis. Also, the reported B-cell mutator AID (activation-induced cytosine deaminase) was not active.
We demonstrate that the predominant mutagenic driver at LBC is off-target V(D)J recombination; the molecular process required for antibody diversification but capable of mutagenic genomic rearrangement. Extreme up-regulation of the core components V(D)J recombination (including RAG1 and DNTT) was accompanied by promiscuous antigen receptor rearrangement and genomic deletions with molecular hallmarks of V(D)J recombination. Deletions included key drivers of haematological malignancy, RUNX1, PAX5, IKAROS, CDKN2A/B and RB1, and were present exclusively in patients with high RAG1/DNTT expression. We propose that by mutating these same genes that control B-cell differentiation and restrain RAG activity, off-target V(D)J recombination becomes central to the mutational process at LBC.
Furthermore, we find that RAG or DNTT expression at diagnosis predict subsequent blast crisis. To incorporate these findings into a diagnostic tool we are developing a diagnostic RNAseq panel and validating it with a further ~300 patient samples. This would bring a molecular pathology arsenal to clinical management of CML, capable of combined detection of single nucleotide variants (SNV’s), large deletions, fusion genes and biomarker gene expression.