Distinct Gene Expression Patterns Drive Proliferation of Dominant-Negative RUNX1 Immortalized HSPCs

The transcription factor RUNX1 is a master regulator of normal hematopoiesis and is involved in cell fate decisions. RUNX1 mutations have been shown to contribute to the development of myeloid neoplasms, and in myelodysplastic syndromes (MDS) it is one of the most frequently mutated genes. Such mutations lead to RUNX1 proteins that lack transactivation activity or DNA-binding ability resulting in a loss of its tumor suppressor function. The dominant-negative short isoform RUNX1a resembles truncated RUNX1 mutants and inhibits the function of the full-length RUNX1 proteins. Additionally, a recently published study identified overexpression of RUNX1a but not full-length RUNX1 in CD34⁺-cells from patients with myelodysplastic/myeloproliferative disease, which increased with disease progression. This strongly suggests that truncated RUNX1 plays a pivotal role in myelodysplastic disease. However, the precise molecular functions of mutating RUNX1, particularly with respect to the identity of RUNX1 target genes conferring its tumor suppressor function, remain unclear.

Previously, our group reported that overexpression of RUNX1a immortalized murine hematopoietic stem and progenitor cells (HSPCs) in vitro. Immunophenotyping of these cells confirmed the expansion of an immature subpopulation defined as Lin^- Sca1⁺ Kit⁺ (LSK). This phenotype was reversed upon turning RUNX1a-expression off and led to a loss of Sca1 expression (Lin^- Kit⁺, LK). To further understand the molecular consequences of RUNX1a overexpression we sorted the LK cells and LSK cells before and 36h after RUNX1a-expression was turned off. Next, we performed microarray analysis to assess differential gene expression in these different subpopulations. Gene set enrichment analysis (GSEA) identified upregulation of genes highly expressed in hematopoietic stem cells (HSC) and leukemic stem cells (LSCs) in RUNX1a-expressing LSK cells compared to those LSK cells in which RUNX1a expression was turned off. Conversely, a gene signature associated with stemness and self-renewal was lost in LK-cells when RUNX1a expression was turned off. Among the eleven leading edge genes, we found genes implicated in leukemogenesis, stem cell regulation, or both such as Erg, Meis1 and Bcl11a.

To further understand the role of RUNX1a in vivo we transplanted C57Bl6 mice (n=29) with HSPCs expressing RUNX1a in a competitive reconstitution setting. Consistent with the immortalization of HSPCs in vitro, RUNX1a-overexpressing HSPCs expanded in the bone marrow of transplanted mice. We observed significantly higher frequencies of LK (2.9-fold) and LSK cells (5-fold) in the RUNX1a-expressing bone marrow cells compared to transplanted control mice. High frequencies of RUNX1a-expressing cells in the bone marrow were associated with lower frequencies of RUNX1a-expressing cells in the peripheral blood indicating a differentiation block. In addition, we found that 85% of the RUNX1a-expressing cells were committed to the myeloid lineage (CD11b⁺/Ly6G⁺) at the expense of the lymphoid lineage (B220 and CD3e). Moreover, RUNX1a expression led to an increased percentage (65%) of immature erythroblasts (Ter119^-) in the bone marrow compared to control cells (55%).

In summary, we have demonstrated that RUNX1a overexpression immortalized HSPCs by upregulation of genes involved in leukemogenesis, stemness and self-renewal. In vivo such HSPCs showed a competitive advantage that was associated with a block of differentiation. Our study, particularly the gene expression analysis, provides novel insights into genetic drivers contributing to the development of myeloid malignancies in patients with RUNX1 mutations.