Duplex Sequencing Uncovers Recurrent Low-frequency Cancer-associated Mutations in Infant and Childhood KMT2A-rearranged Acute Leukemia

Abstract

Infant acute lymphoblastic leukemia (ALL) with KMT2A -gene rearrangements ( KMT2A -r) have few mutations and a poor prognosis. To uncover mutations that are below the detection of standard next-generation sequencing (NGS), a combination of targeted duplex sequencing and NGS was applied on 20 infants and 7 children with KMT2A -r ALL, 5 longitudinal and 6 paired relapse samples. Of identified nonsynonymous mutations, 87 had been previously implicated in cancer and targeted genes recurrently altered in KMT2A -r leukemia and included mutations in KRAS , NRAS , FLT3 , TP53 , PIK3CA , PAX5 , PIK3R1 , and PTPN11 , with infants having fewer such mutations. Of identified cancer-associated mutations, 62% were below the resolution of standard NGS. Only 33 of 87 mutations exceeded 2% of cellular prevalence and most-targeted PI3K/RAS genes (31/33) and typically KRAS/NRAS . Five patients only had low-frequency PI3K/RAS mutations without a higher-frequency signaling mutation. Further, drug-resistant clones with FLT3 D835H or NRAS G13D/G12S mutations that comprised only 0.06% to 0.34% of diagnostic cells, expanded at relapse. Finally, in longitudinal samples, the relapse clone persisted as a minor subclone from diagnosis and through treatment before expanding during the last month of disease. Together, we demonstrate that infant and childhood KMT2A -r ALL harbor low-frequency cancer-associated mutations, implying a vast subclonal genetic landscape.

Real-Time Characterization of Clonal Fate Decisions in Complex Leukemia Samples by Fluorescent Genetic Barcoding

“…Despite the unclear implications of clonal evolution for our experiments, leukemia is generally composed of multiple competing subclones, which form a cellular reservoir for relapse [ 49 ]. Recent publications showed that minor relapse-driving subclones with increased drug resistance can already be present in diagnostic samples, but their awakening in response to therapy remains enigmatic [ 50 , 51 ]. Our color-coded leukemia model thus provides an interesting approach to test the consequences of targeted clone depletion on the reactivation of dormant subclones and the underlying “switchboard” signals with single cell resolution that may have therapeutic potential [ 52 , 53 ].…”