Retroviral Mediated Transfer of MLL Fusion Genes into Human CD34+ Cord Blood Cells Supports the Establishment of Long-Term Myeloid Cultures with Growth Rates, Immunophenotypic and Transcriptional Features Distinct from Those of AML1-ETO Transduced Cultures.

Translocations involving the MLL gene, primarily t(9;11) and t(10;11) together with rearrangements affecting the core binding factor (CBF) genes, t(8;21) and inv(16) comprise the most frequent cytogenetic abnormalities in acute myeloid leukemia (AML). Although all of these rearrangements generate chimeric transcription factors (MLL-AF9, MLL-AF10, AML1-ETO and CBFβ-MYH11) clinicopathologic features and transcriptional profiles clearly distinguish MLL-rearranged from CBF-rearranged AML. To understand how these distinct subgroups of AML arise, we have developed a model for studying the effects of MLL and CBF fusion proteins on the growth, survival and differentiation human myeloid progenitors in vitro. Using retroviral mediated gene transfer, we transduced CD34 selected normal human cord blood (CB) cells with vector (MIEG3) alone or with vectors expressing MLL-AF9, MLL-AF10 or AML1-ETO fusion genes. Whereas CB transduced with MIEG3 proliferated in liquid culture for 6 to 8 weeks, MLL-AF9, MLL-AF10 and AML1-ETO transduced cells have continued to proliferate continuously in culture for more than 16 weeks without any sign of crisis. At any point after transduction, CB expressing MLL-AF9 or MLL-AF10 exhibited a faster rate of growth as compared with AML1-ETO or MIEG3 transduced CB. This difference in growth rate was associated with a reduced frequency of spontaneous apoptosis by annexin staining in the MLL cultures, as compared with the AML1-ETO cultures, but no difference in the fraction of cells in S-phase. MLL-AF9 and MLL-AF10 transduced CB also exhibited evidence of early myeloid maturation arrest based on morphology and surface antigen expression. However, while AML1-ETO transduced cells continue to express CD34 throughout their time in culture, MLL cultures lose expression of this stem cell-associated antigen and acquire expression of c-Kit and CD33, neither of which is expressed in AML1-ETO cultures. Also, unlike AML1-ETO transduced cells, CB transduced with with MLL fusions retain serial clonogenicity for 3 or more rounds of plating in methylcellulose assays. We used quantitative realtime RT-PCR to measure expression of 3 genes that are differentially expressed in patients with MLL or CBF gene fusions based on published microarray data. While expression of SPARC increased over time in MIEG3 cultures or remained stable in CB transduced with AML1-ETO, it decreased to nearly undetectable levels in MLL-AF9 transduced cultures. In contrast, expression of both BMI-1 and HOXA9 increased in the MLL-AF9 cultures and decreased in the MIEG3 and AML1-ETO cultures. The transcriptional changes in our long-term cultures mirror the gene expression differences that have been observed in AML associated with MLL or CBF fusions and suggest that this will be a useful model to study how chimeric transcription factors contribute to myeloid leukemogenesis. Interestingly, CB transduced with a mutated MLL-AF10 (MA10ΔLZ) lacking the leucine zipper domain required for transformation of primary murine myeloid progenitors did not differ, in terms of growth or differentation, from MIEG3 transduced cells. This suggests that the effects of MLL-AF9 and MLL-AF10 on normal CB may reflect early events in myeloid leukemogenesis. The in vivo leukemogenic potential of MLL fusion transduced CB is currently being evaluated in NOD/SCID mice.