Acute Lymphoblastic Leukemia (ALL) is the most common malignancy in children. Altogether 90% of pediatric ALL patients achieve a complete hematologic remission with current high dose combination chemotherapy and 80% of them remain leukemia free. However, the outcome for patients showing refractory disease or those whose leukemia relapses after an initial transient response remains disappointingly poor with cure rates of less than 40%.

To investigate genetic drivers of relapse and resistance and explore the specific roles of clonal evolution in disease progression and relapse here we performed whole-exome sequence analysis of matched diagnosis, germline (remission) and relapse DNA samples in a panel of 55 pediatric ALL patients including 33 T-cell ALLs and 22 B-cell precursor ALLs. These analyses identified an average of 9 mutations present in diagnostic samples and 17 mutations in relapsed leukemia DNAs. Phylogenetic tree analysis for each of the 48 cases with optimal variant call parameters analyzing their clonal evolution dynamics during disease progression, combined with whole genome sequencing of targeted samples with low exonic mutation input, showed that branched evolution in which relapse clones contain some, but not all genetic lesions present in the major clone at diagnosis as the primary mechanism driving tumor progression and relapse present in 45/48 (94%) cases. In addition, and consistent with previous reports we identified the presence of chemotherapy associated mutations in NT5C2 (10/55), TP53 (3/55), CREBBP (4/55) and the NR3C1 glucocorticoid receptor gene (2/55). However, and most strikingly, 23/27 (85%) recurrently mutated genes in this series with mutations preferentially selected or retained at the time of relapse (mutation never lost in the relapse clone) were not implicated in relapse ALL before (HTR3A, MED12, USP9X, CACNA1H, ODZ3, AACS, SAMD4A, ANO5, PAPPA, NAALADL2, HIST3H2A, FZD7, TBX15, NEB, GREB1L, PLXNA4, SGK223, TSC1, PTPRG, FGF10, SYCP2, TRPM3 and EYS). A branched pattern of genetic evolution and the presence of recurrent mutations selected at relapse support that chemotherapy imposes a strong Darwinian genetic selection in leukemic cell populations.

In this context it is worth noting that RAS-MAPK pathway activating mutations in NRAS, KRAS and PTPN11 were present in 24/55 (44%) cases in our series. Interestingly, some leukemias showed retention or emergence of RAS mutant clones at relapse, while in others, RAS mutant clones present at diagnosis were replaced by RAS wild type populations, supporting a role for both positive and negative selection evolutionary pressures in clonal evolution of RAS-mutant leukemia. Most notably, and in agreement with this hypothesis, inducible expression of mutant KRAS in human ALL lines demonstrate that oncogenic KRAS G12D induces methotrexate resistance, but also improves leukemia response to vincristine; a phenotype perfectly recapitulated in a isogenic ALL leukemia model generated from a conditional inducible Kras G12D knockin mice. Mechanistically, KRAS G12 expression induces MAPK dependent abrogation of methotrexate induced apoptosis. Moreover, Kras mutant tumors show enhanced G2/M cell cycle arrest and apoptosis upon spindle poisoning with vincristine, a phenotype linked with increased PLK phosphorylation and transcriptional down-regulation of mitotic genes. Finally clonal competition assays demonstrate that the differential response to methotrexate and vincristine in isogenic Kras wild type and Kras mutant ALL cells results in clonal dominance of Kras G12D populations in cultures treated with methotrexate, while Kras wild type cells are selected the context of vincristine treatment. In all these results show novel insight on the genetics and mechanisms of clonal selection, disease progression and relapse in ALL and demonstrate a previously unrecognized dual role of RAS mutations in chemotherapy response.

Disclosures

Loh:Abbvie: Research Funding; Bristol Myers Squibb: Membership on an entity's Board of Directors or advisory committees.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution