Abstract
Background
Patients with chronic myeloid leukemia (CML) who develop blast crisis are resistant to TKI therapy. A key focus in CML research is the identification of genetic factors that promote blast crisis and TKI resistance. By using an integrative genomic approach we recently reported frequent structural variation in CML patients, particularly at lymphoid blast crisis (LBC) (Blood, 2018). Developing lymphocytes are uniquely equipped with a molecular toolkit capable of programmed DNA damage and structural variant formation; the V(D)J recombination pathway. Recombination activating genes (RAG1 and RAG2) are involved in cleavage and recombination of immunoglobulin genes to provide diversity in antibodies and T cell receptors. Off target RAG activity is reported to occur in lymphoid malignancies and cause oncogenic structural rearrangements. However, RAG expression and the extent of RAG mediated structural variation in CML are largely uncharacterized.
Aim
To elucidate the role of RAG mediated recombination as a source of oncogenic structural rearrangement in CML.
Methods
In a study of samples of 49 patients at chronic phase (CP) diagnosis (Dx), 20 at LBC and 19 at myeloid blast crisis (MBC), we performed whole exome sequencing and/or RNAseq. Bioinformatic analyses included fusion detection (Manta & STAR), gene expression analysis (EdgeR), and copy number variation analysis (in house developed tool). Unsupervised motif detection of sequences surrounding breakpoints was performed with MEME, and fusions were visualized with Jcircos. To identify off target RAG mediated mutation we interrogated the breakpoints of structural variants, excluding those associated with the Philadelphia translocation and those solely involving antigen receptor gene rearrangement.
Results
33 structural variants were identified in 22 patients with samples at Dx and/or blast crisis involving genes regularly mutated in hematologic malignancy such as MLL, MECOM, RUNX1 and IKZF1. Differential expression analysis between patients at Dx, MBC and LBC revealed >1000 genes that were differentially expressed, P<0.001. Amongst the most upregulated genes at LBC were RAG1, RAG2 and DNTT, reflected in the most enriched gene ontology: V(D)J recombination. V(D)J recombination genes were sufficient alone in stratifying LBC samples from CP Dx and MBC [Figure 1A]. Furthermore, RAG expression was detectably elevated at Dx in most patients who subsequently developed LBC [Figure 1B].
To identify off target RAG mediated structural variation, we interrogated the sequence surrounding breakpoints of the 33 identified structural variants (66 breakpoints) that were separated into groups of high (n=40) and low (n=26) RAG expression. The canonical heptamer CACAGTG recombination signal sequence (RSS) was enriched only at breakpoints in the high RAG group with sequences from 31/40 breakpoints making up the motif [Figure 1C]. The comparable group of low RAG breakpoints were not enriched for the motif. Another indicator of RAG involvement is the addition of non-template nucleotides at the breakpoint, consistent with DNTT nucleotidyl-transferase activity. This was detected almost exclusively at breakpoints in patients with high RAG, 16/20 structural variants, compared to 1/13 with low RAG.
There was an association between high RAG expression and structural variants bearing the RAG RSS motif at breakpoints with nucleotide additions. The most frequently observed of these were intragenic deletions of IKZF1 in 7 patients, including 3 at Dx with subsequent LBC, 3 at LBC and 1 at MBC. A number of novel structural variants also had the RAG recombination signature: IKZF1-IGK fusion, RUNX1 deletion and recurrent HBS1L-MYB intergenic locus deletion.
RAG mediated structural variation was further investigated on a genomic level using putative low level fusions identified from RNAseq. By assessing the immunoglobulin/T cell receptor genes, as known sites of RAG activity, we found enrichment of interchromosomal fusions involving these genes in samples with high RAG, compared to Dx and MBC with low RAG [Figure 1D]. An agnostic motif enrichment search at sequences surrounding putative fusions showed enrichment of a motif resembling the recombination signal sequence.
Conclusion
Our results implicate RAG as a driver of structural variation and the formation of novel oncogenic rearrangements in CML.
Figure 1. Figure 1.
Disclosures
Hughes: Novartis: Honoraria, Membership on an entity's Board of Directors or advisory committees; Incyte: Honoraria, Membership on an entity's Board of Directors or advisory committees; BMS: Honoraria, Membership on an entity's Board of Directors or advisory committees. Branford:Novartis: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau; Qiagen: Honoraria, Membership on an entity's Board of Directors or advisory committees; BMS: Honoraria, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Cepheid: Honoraria.
Development of a potent high-affinity human therapeutic antibody via novel application of Recombination Signal Sequence (RSS)-based Affinity Maturation
