RUNX1 Mutations Lead to a Myeloid Differentiation Block in Association with Enhanced Proliferation By Altering the RUNX1 Transcriptional Program

RUNX1 (AML1) is a transcription factor critically involved in normal haematopoiesis. Inactivating RUNX1 mutations have been frequently described in a variety of myeloid neoplasms, including high-risk myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML). Here, we aimed to functionally and molecularly define the actions of a dominant negative mutant.

Overexpression of the RUNX1 mutant S291fs300X (RUNX1 mt) in cord blood (CB) CD34⁺ cells caused a decline in erythroid colony formation (p= 0.01) which was associated with an immature morphology and an increased number of Annexin V positive erythroid cells. The negative phenotype coincided with a strong reduction in GATA-1 and GATA-2 expression in RUNX1mt CD34⁺ CB cells after 24-36 hours transduction. On the contrary, the CFU-GM colonies showed a greater than 3-fold enhanced replating capacity compared to control. It appeared that the replating potential was predominantly present in the CD34⁺CD38^- and CD14^-CD15^- cell population.

Upon long-term MS5 stromal co-cultures of transduced RUNX1 mt CB CD34⁺ cells a rather homogenous cell population was obtained after 8-10 weeks that localized underneath the stromal layer. This cell population was phenotypically defined by CD34⁺/CD38^-. The interactions with the stroma appeared to be important to limit proliferation and retain quiescence. Suspension cultures with myeloid growth factors (IL-3, SCF) of RUNX1 mt CB CD34⁺ cells provided a similar cell population after 10 weeks of culture. These cells were growth factor-dependent and were phenotypically defined by CD34⁺/CD38⁺/CD33⁺/CD45RA⁺/CD123⁺, resembling a GMP phenotype which could be propagated for approximately 20 weeks in suspension. Karyotype analyses of this cell population demonstrated no abnormalities while integration site analysis showed multiple different integration sites that differed between individual experiments, suggesting that the myeloid differentiation block was related to the RUNX1 mutation and not to aspecific integration site effects. Comparable results were obtained with adult normal bone marrow CD34⁺ cells or peripheral blood CD34⁺ cells transduced with the RUNX1 mt.

Morphological analysis after 8-10 weeks of culture showed monoblastic cells with limited macrophage differentiation but without signs of granulocytic differentiation, even in the presence of G-CSF. This might might be due to downregulation of C/EBPα, one of the key targets of RUNX1. Therefore, week 10 RUNX1 mt CB CD34⁺ cells were transduced with a retroviral C/EBPα-ER overexpression vector. Re-expression of C/EBPα by tamoxifen induction resulted in a reduction in cell proliferation, decline of undifferentiated blasts and an increase in CD15 expression.

In our goal to asses RUNX1 mt target binding and gene expression changes at early time points we used induced pluripotent stem cells (iPS cells) with conditional expression of RUNX1 mt. Expression of RUNX1 mt during early differentiation leads to a differentiation block with the maintenance of CD34⁺ cells. Genome-wide binding analysis of these RUNX1 mt cells revealed RUNX1 occupancy at similar regions as observed in primary AMLs with a RUNX1 mutation. These binding sites were enriched for RUNX and ETS motifs and associated with genes important for differentiation and chromatin regulation. Moreover, expression analysis revealed a severely altered RUNX1 transcription program in the presence of the RUNX1 mt.

Together these results reveal that the RUNX1 mutant S291fs300X, as a single oncogenic hit, induces cell proliferation and a myeloid differentiation block by altering the RUNX1 gene program.