Myeloid Progenitors of Defined Features Provide Permissiveness to Leukemic Transformation

The conventional view of oncogenesis posits that driver mutations, when present in appropriate cell types, initiate malignant transformation. At the apex of the differentiation hierarchy, stem cells, such as hematopoietic stem cells, have been suggested to be the cell-of-origin for a variety of hematopoietic malignancies. However, since the mutation bearing hematopoietic stem cells can differentiate to give rise to downstream progeny, it suggests the possibility that some of the downstream cells be the relevant cell-of-origin permitting malignant transformation. Given that cancer cell-of-origin has been a product of logic deduction evading rigorous experimentation, whether and how their unique cellular properties, if any, contribute to transformation remain undetermined.

To experimentally define the key properties of the cell-of-origin permissive to oncogenic transformation, we constructed a doxycycline (Dox) inducible MLL-AF9 knock-in mouse model, which allows precise temporal control of oncogene expression in desired hematopoietic cells, in vitro and in vivo . Whole body induction by Dox led to 100% lethality due to acute myeloid leukemia (AML) within 6-8 weeks, while the un-induced littermates remained healthy, confirming the inducibility and potency of the MLL-AF9 oncogene. To minimize the differences contributed by cellular differentiation stages and to focus solely on the contributions by specific cellular properties, we narrowed our analysis on a phenotypically homogeneous population, the granulocyte-macrophage progenitors (GMPs), which have been shown to be critically important in supporting leukemic development (Ugale et al ., 2014, Cell Reports; Ye et al ., 2015, Cell Stem Cell). We found that although MLL-AF9 expression can be induced in all GMPs, only a subset (~25%) of them is permissive to transformation. These results prompted us to investigate the underlying cellular and molecular properties that enable oncogenic transformation.

Inspired by our earlier work demonstrating that the cell fate plasticity of GMPs is related to their rapid cell cycle behavior (Guo et al ., 2014, Cell), we tested whether the rapid proliferation of GMPs underlies their permissiveness to oncogenic transformation. If so, transformation potential should be enriched in the faster-proliferating GMP subsets. To test this possibility, we designed single cell assays to investigate the transformation efficiencies of GMPs with regard to their cell cycle speeds. Specifically, the intrinsic cell cycle speed was measured within 24 hours of Dox addition, and transformation status of hundreds of individual GMPs was determined by methylcellulose colony formation assay. Within the first 24 hours of Dox addition, the cell cycle speed remained unchanged, reflecting the intrinsic cell cycle kinetics, rather than a proliferative response to oncogene expression. Our results show that the faster a GMP divides, the more likely it transforms in response to MLL-AF9 expression. Strikingly, most GMPs that can divide more than 3 times within 24 hours are transformed. These data indicate the fastest cycling GMPs provide the oncogene with a permissive cellular context to exert its function. Furthermore, prospectively isolated faster cycling GMPs enriched for transformed colony formation in vitro, and induced earlier lethal AML in vivo . Importantly, transient deceleration of GMPs for 48 hours by low-dose CDK4/6 inhibitor, Palbociclib, significantly reduced transformation efficiency by ~ 60%. Taken together, these results demonstrate that MLL-AF9 initiated transformation is dependent on specific cellular properties which are associated with a rapid cell cycle. Genomic analyses of the transformation-permissive GMPs, using non-permissive GMPs as controls, are being performed to define the permissiveness in molecular details, on both transcriptional and epigenetic levels.