P312: EMPLOYING WHOLE GENOME OPTICAL MAPPING TO CHARACTERIZE THE LANDSCAPE OF STRUCTURAL VARIANTS IN B-CELL PRECURSOR ACUTE LYMPHOBLASTIC LEUKEMIA

Abstract

Background: B-cell precursor acute lymphoblastic leukemia (BCP-ALL) is the most frequent pediatric cancer. The most common subtypes are ETV6::RUNX1+ BCP-ALL, which harbor a reciprocal translocation between chromosome 12 and 21 leading to a juxtaposition of ETV6 to RUNX1, and high hyperdiploid (HHD) BCP-ALL harboring 51-67 chromosomes. In large-scale genomic studies, the mutational landscape of somatic single nucleotide variants and indels as well as copy number gains and losses based on array CGH have been well described in both subtypes. However, a comprehensive analysis of somatic structural variants (SVs) is still lacking which may reveal significant new insights on tumor development, progression, and treatment options in BCP-ALL. Aims: We aimed to primarily employ the novel technology whole genome optical mapping (WGOM) to unmask the landscape of structural variants in BCP-ALL. Firstly, we determined the concordance of chromosomal abnormalities identified by WGOM in comparison to standard techniques. Superiorly, we explored the landscape of somatic aberrations in ETV6::RUNX1+ and HHD tumors to identify recurrent alterations and novel SVs affecting genes not yet related to BCP-ALL pathogenesis. Methods: We extracted ultra-high molecular weight (UHMW) DNA of 13 ETV6::RUNX1+ and 18 HHD BCP-ALL with a tumor cell content of 25-96% at diagnosis. Additionally, we isolated DNA of fibroblasts or remission material of each pediatric patient as matching non-tumor control. Informed consent was obtained from all participants. DNA was fluorescently labeled by a sequence specific enzyme and subsequently linearized and imaged using a Saphyr instrument (Bionano Genomics). Using this approach, we generated comprehensive datasets of 31 matched tumor/non-tumor pairs, allowing the identification of somatically acquired SVs >500bp. In addition, corresponding karyotype, fluorescence in situ hybridization and SNP-array data of the leukemia samples were integrated and compared to WGOM data. Results: We validated 95% of 233 events observed by standard techniques with WGOM. All hallmark alterations including ETV6::RUNX1 fusion and high hyperdiploidy were identified by WGOM. In addition, we detected a median of 11 (1-36) somatically acquired events in leukemic cells with ETV6::RUNX1 fusion, compared to 3 (1-10) events in HHD cases. WGOM identified recurrent somatic structural aberrations occurring in at least three BCP-ALL cases, including focal deletions affecting ETV6 (9/31), PAX5 (5/31), ARPP21 (3/31), CD200/BTLA (3/31) and RAG2 (3/31). We detected subtype specific recurrent focal deletions disrupting ATF7IP (7/13) and BTG1 (5/13) in ETV6::RUNX1+ subclones whereas deletion of the CDKN2A/B locus (3/18) was only observed in HHD BCP-ALL. Interestingly, we discovered novel recurrently deleted regions on 12q24.11 (GPN3, 4/31) and Xq25 (STAG2, 3/31) in all BCP-ALL cases as well as focal loss of 19q13.11 (UBA2, 3/13) specific for ETV6::RUNX1+ subclones using WGOM. Moreover, complex three-way translocations in ETV6::RUNX1+ tumors were discovered by WGOM, including t(6;8;12)(p12.3;q24.23;q24.21) that leads to the disruption of MED13L and HMGCLL1. Summary/Conclusion: Our study showed that WGOM detects SVs as translocations and aneuploidies in BCP-ALL as good as conventional methods with a 95% concordance. However, WGOM allows a higher resolution and more precise localization of chromosomal breakpoints compared to standard techniques. Hence, we identified novel recurrent somatically acquired SVs including a deletion on Xq25 affecting STAG2, which might play an important role in pediatric BCP-ALL leukemogenesis.