Abstract
Abstract 781
While serving as an initiating event in the disease etiology, chromosomal translocation alone may not be sufficient to drive a particular phenotype of leukemia. Acute leukemia characterized by rearrangements of the histone methyltransferase gene MLL (mixed lineage leukemia) requires additional mutations to develop the full blown malignancy, yet the molecular basis of cooperating events remains under-studied. We reasoned that monozygotic (MZ) twin pairs discordant for human leukemia are well-matched for both inherited genetic background and tissue-specific events, and thereby somatic mutations that arise in the disease twin may play a more prominent role in leukemogenesis. Using whole genome sequencing in a pair of 3-year old monozygotic twins discordant for MLL leukemia, we identified a MLL-NRIP3 fusion gene and mutations in histone H3 lysine 36 methyltransferase SETD2 in the leukemia twin. Retrovirus-mediated ectopic expression of MLL-NRIP3 in mouse hematopoietic cells was able to induce the same type of myeloid leukemia as the patient's in a transplant mouse model. A relative delay in disease onset suggested the occurrence of cooperating events in addition to the initial hit of MLL-NRIP3, in the development of induced leukemia. SETD2 mutations were recurrent (5.4%) in 241 acute leukemia patients, particularly in those with MLL-rearranged myeloid leukemia (22.2%). The identified SETD2 mutations are loss-of-function in nature, characterized by biallelic and truncating mutations, and accompanied by a global loss of trimethylation of H3K36 (H3K36me3) in the patient leukemic blasts. These data suggest that SETD2 acts as tumor suppressor gene in leukemia development. Functionally, transfection of SETD2 shRNA in MLL-AF9 knockin bone marrow cells resulted in decreased levels of SETD2 expression and H3K36 trimethylation. Notably, SETD2 knockdown significantly accelerated the development of MLL-AF9 leukemia in the mouse bone marrow transplantation experiment. Moreover, SETD2 knockdown yielded a significantly higher number of total colonies through the second and third rounds of serial replating of colony-forming cell (CFC) assay, suggesting that loss of SETD2 increased the self-renewal and proliferation potential of MLL leukemia-initiating cells. Finally, we showed that SETD2 deficiency was able to activate gene expression in MAPK, Jak-STAT and mTOR signaling, and dysregulates multiple metabolic and DNA repair pathways that are known to directly contribute to leukemogenesis. This comprehensive study provides compelling evidence for SETD2 as a novel tumor suppressor for leukemia, and suggests that the disruption of distinct histone modifying enzymes, MLL and SETD2, synergistically promotes the development of human leukemia. In addition, our study illustrates that whole-genome sequencing of phenotypically discordant monozygotic twins provides an effective approach, in combination with mutational analysis in patients and functional assays using experimental models, to uncover disease-causal genes.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.