MLL-PTD Causes Hypomorph Condition of CBF Complex (RUNX1/CBFβ) and Predisposes the Abnormal Hematopoietic Stem and Progenitor Cells (HSPCs) to Clonal Expansion

Abstract 2801

A relatively common mechanism through which Mixed-Lineage Leukemia (MLL) gene is disrupted in myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML) is a partial tandem duplication of its coding sequence (MLL-PTD), which duplicates the MLL N-terminal region and creates a larger in-frame oncogenic protein. We have previously shown that MLL interacts with and stabilizes the core binding factor (CBF) complex, a heterodimeric transcription factor complex composed of a DNA-binding subunit RUNX1 (CBFa) and a non-DNA-binding subunit CBFβ. MLL is required for CBF dependent regulation of the downstream target gene PU.1¹. In MDS and AML, RUNX1 is frequently disrupted by mutations, and the presence of MLL-PTD and/or RUNX1 mutations may negatively impact on clinical outcome. However, little is known how mutated MLL and RUNX1 contribute to MDS-associated ineffective hematopoiesis, clonal expansion and transformation into AML. Interestingly, MDS patients with either mutant RUNX1 or MLL-PTD tend to develop secondary AML harboring both RUNX1 and MLL mutations, suggesting a leukemogenic cooperation between the two genes through converging mechanisms. We hypothesized that the MLL-CBF nuclear signaling pathway maintains stem cell fitness, and the MDS-associated mutations co-opt this pathway leading to ineffective hematopoiesis, clonal expansion and transformation into AML. To address these questions, MLL-PTD+ and MLL-PTD(−) AML cell lines were compared. RUNX1 and CBFβ proteins were expressed at low levels in all 4 MLL-PTD+ AML cell lines (OCI-AML2, EOL, KOMP88, MUTZ11) compared with MLL-PTD(−) AML cell lines (OCI-AML5, Kasumi1, HL60, KG1a). In contrast to wild type MLL, which stabilized the CBF complex, MLL-PTD downregulated RUNX1 and CBFβ proteins in co-expression assays in 293T cells. This finding was further confirmed using the Mll^PTD/WT knock-in mouse model. In Mll^PTD/WT mice, the abnormal hematopoietic stem and progenitor cells (HSPCs) showed decreased expression levels of Runx1 (50%) and Cbfβ (30%) proteins and in turn PU.1 (30%) and its target gene, Mcl-1 (20%) compared with WT HSPCs. Consistent with this finding, HSPCs from Mll^PTD/WT mice had increased spontaneous apoptosis compared with WT HSPCs [(Lin−/cKit+/Sca-1+ (LSK), Annexin V positivity: 44% ±7.5% vs 25%±4.5%, p<0.05)]. At steady state, the bone marrow LSK and SLAM+LSK (CD150+/CD48−/LSK) populations were reduced by 50% (p<0.01) and 20% (p<0.01) respectively in Mll^PTD/WT mice compared with WT controls. Yet, when exposed to stress conditions (5-FU treatment), the HSPCs from Mll^PTD/WT mice exhibited rapid proliferation, reduced apoptosis and increased survival signaling with upregulated Bcl-xL compared with WT controls. Strikingly, HSPCs from Mll^PTD/WT mice exhibited enhanced expansion in serial bone marrow transplantation assays compared to those from WT controls (percentage of donor-derived cells: 91.4% ±4.9% vs 38.3% ±3.3% in tertiary, p<0.01; 62.5%±3.7% vs 27.9%±2.5% in quaternary transplantation, p<0.01). The enhanced expansion of Mll^PTD/WT HSPCs in recipient mice was similar to that of Runx1^+/− HSPCs (percentage of donor-derived cells: 79.4%±2.4% in tertiary; 44.5%±2.6% in quaternary transplantation). In experiments using double heterozygous HSPC cells (Mll^PTD/WT/Runx1^+/−), the enhanced expansion in recipient mice was even more pronounced (percentage of donor-derived cells: 89.1%±2.1% in tertiary; 79%±9.4% in quaternary; 64%±8.3% in 5^th transplantation, p<0.01; table 1). These double heterozygous HSPC have even lower levels of Runx1 and Cbfb (20% and 10% respectively of the WT level). In a competitive bone marrow transplantation assay (at 1:1 ratio), the HSPCs of Mll^PTD/WT also show advantage in expansion compared with those from WT controls (90.1%±1.9% vs 54.6%±3.6% 16 weeks after transplantation, p<0.01; table 1). These data indicate that HSPCs of Mll^PTD/WT and Runx1^+/− have similar enhanced expansion activity compared to wild type controls, while HSPCs of Mll^PTD/WT/Runx1^+/− have even greater expansion activity compared to single mutant heterozygous HSPCs. This finding could partially explain the advantage of those HSPCs with both MLL-PTD and RUNX1 mutations in s-AML and de novo AML. Our findings provide a mechanistic model for development and maintenance of MLL-PTD- and RUNX1-related MDS and AML and provide novel targets for developing new therapies for these diseases.

View large

View Large