Somatic Mutational Landscape of AML with Inv(16) and t(8;21) Identifies Two Distinct Patterns in Relapse Tumors

Acute myeloid leukemia (AML) is a heterogeneous disease with a wide prognostic spectrum ranging from poor to good depending upon the underlying mutations and/or cytogenetic abnormalities. Although AMLs with inv(16)/t(16:16) or t(8,21), collectively referred to as core binding factor leukemias (CBF-AMLs), are classified as prognostically favorable, such patients often succumb to their disease following relapse after an initial response to cytarabine/anthracyclin-based treatment regimens. Thus, to develop successful treatment strategies, it is critical to understand the mechanisms leading to disease relapse and target them with novel therapeutic approaches. To pursue this goal, we applied genomic approaches (whole exome sequencing and single nucleotide polymorphism arrays) on DNA from samples collected at sequential time points (i.e., diagnosis, complete remission and relapse) in seven patients with inv(16) and six patients with t(8;21). We identified mutations in several previously identified AML driver genes, such as KIT, FLT3, DNMT3A, EZH2, SMC1A, SMC3, WT1 and NRAS. Three relapse samples showed mosaicism for monosomy/disomy of the region of chromosome 3 containing GATA2. Overall, our data revealed two distinct profiles that support different mechanisms of relapse: 1) diagnosis and relapse blasts harbor the same driver gene mutations, indicating the intrinsic resistance of the major clones present at diagnosis to treatment regimen used; 2) diagnosis and relapse tumors have different driver gene mutations, indicating disease clonal evolution possibly through treatment selective pressure. Furthermore, our data has identified previously unreported putative driver genes for AML. Among these, we identified same somatic variant (R222G) in DHX15, an RNA helicase involved in splicing, in two patients at diagnosis. The variant was also detected at relapse in one of these patients. Functional validation of the mechanistic roles of wild type and mutated DHX15 in hematopoiesis and leukemogenesis, respectively, is ongoing in in vitro and in vivo models.