Genomic Determinants of Therapy Response in ETV6::RUNX1 Leukemia

Background: Acute lymphoblastic leukemia (ALL) is the most common cancer in children, with about 3500 children diagnosed yearly in Europe. A quarter of cases harbor the ETV6-RUNX1 (E/R) fusion gene. E/R leukemias are usually classified into low risk group. However, a fraction of patients still encounters disease recurrence. The increased relapse risk has been linked with a non-optimal initial therapy response as measured by minimal residual disease (MRD). Aims: The aim of this study is to identify genetic differences between E/R cases stratified by MRD and to identify potential mediators of poor therapy response. Methods: We analyzed the genetic landscape of E/R leukemia in a discovery cohort of 35 patients treated with the NOPHO2008 protocol using whole genome sequencing (WGS). Targeted sequencing was included in validation, including additional cohorts. Patients were classified into three categories (slow, intermediate or fast responder) based on MRD at the end of induction therapy (day 29). Variant callers available in the Balsamic workflow were used to detect DNA alterations and subclonal events analyzed using Battenberg and FastClone tools. Results were compared against published recurrent ALL mutations, bone marrow single cell RNA-sequencing and genome-wide CRISPR-screens in the E/R+ cell line REH, with vincristine, cytarabine, methotrexate, L-asparaginase, maphosamide, daunorubicin and 6-mercaptopurine sensitivity. Nalm-6 screen data was used for evaluating glucocorticoid response. Results: We found comparable number of single nucleotide variants (SNVs), indels and structural variants (SVs) between the MRD-categories. However, based on mutation signature analysis, a significant negative correlation of MRD-level at day 15 was found with Signature 2 (APOBEC), while day 29 MRD correlated positively with Signature 3 (failure of DNA double-strand repair). Interestingly, the signature related to AID/APOBEC family has been reported with oncogenic role in T-ALLs, where APOBEC3 family genes display high expression during thymocyte development. We used single cell RNA-seq profiles from healthy bone marrow to study the relevance of APOBEC3 expression during precursor B-cell development, and showed that APOBEC3B is highly expressed in the dividing hematopoietic precursor, pro-B and especially pre-B cell stages. Well-known ALL-related driver genes did not differ between MRD categories, with the exception of KRAS. Instead of activating mutations, we found prevalent deletions in the fast responder category. To focus specifically on drug response modulating genes, we focused on genes that significantly increased or decreased sensitivity in genome-wide CRISPR screens (top 5% screen hits). Overall, each case harbored multiple DNA alterations in these genes. Moreover, fast responder cases had increased number of hits in sensitizing genes. Summary/Conclusion: Combining MRD analysis with WGS studies carries the potential to reveal crucial information about sensitizing and resistance mutations and underlying cellular mechanisms. The found DNA variants that associate with MRD-status can be utilized in the design and implementation of personalized therapy in E/R+ ALL.