A common abnormality in AML (observed at a frequency of ~12%) is the t(8;21)(q22;q22) encoding the RUNX1-ETO fusion protein. Previously we showed that RUNX1-ETO can block the differentiation of normal human primary CD34+ hematopoietic stem progenitor cells (HSPC) and promote their self-renewal (consistent with its putative initiating role; Tonks et al 2003, 2004). To identify the genes that may mediate this phenotype, transcriptome analysis was performed (Tonks et al 2007). We now report that RUNX1-ETO significantly downregulates the expression of RUNX3 (2.1±0.6 fold in CD34+ HSPC; p<0.05). Furthermore, we also show that RUNX3 mRNA is significantly downregulated in t(8;21) patients compared to normal human bone marrow by 4 fold (p<0.001). Interestingly, in non-t(8;21) AML, RUNX3 expression is overexpressed compared with normal HSC, suggesting that both down and upregulation of RUNX3 may contribute to the pathogenesis of AML. To investigate this further, we examined the impact of RUNX3 expression on normal human hematopoiesis.

Initially, we investigated the role of RUNX3 in erythropoiesis (which is severely disrupted by RUNX1-ETO). We mimicked the effects of RUNX1-ETO through shRNA-mediated knockdown (KD) of RUNX3 in CD34+ HSPC which co-expressed GFP. Following infection, GFP+ CD13low cells were isolated by FACS to yield a population of cells enriched for erythroid progenitors. Under clonal conditions, RUNX3-KD cells formed significantly less erythroid colonies than control (1.9±0.7 fold, p<0.05), matching the impairment seen in erythroid progenitors expressing RUNX1-ETO. To examine this phenotype in more detail, we studied the effect of RUNX3-KD on the EPO-independent and EPO-dependent stages of erythroid development. RUNX3-KD significantly inhibited growth during the EPO-independent development (suppressed by 1.7±0.1 fold; p<0.001) and showed impaired differentiation with reduced CD36 expression (1.5±0.3 fold, p<0.01), retention of HLA-DR (1.2±0.1 fold, p<0.01) and CD34 (1.2±0.1 fold, p<0.05). These cells showed partial response to EPO but again growth response was suppressed (2±0.8 fold compared to control). Developmentally, RUNX3-KD cells failed to upregulate GlyA to the same levels as controls (suppressed by 1.3±0.1 fold p<0.05), suggesting that the inhibition of differentiation persisted in the EPO-dependent phase of development. In summary, KD of RUNX3 in HSPC leads to an impairment in erythropoiesis similar to the phenotype induced by RUNX1-ETO, which suggests that suppression of RUNX3 by this fusion protein contributes to the RUNX1-ETO phenotype.

We next assessed the effect of RUNX3 overexpression on erythroid development as above, this time using retrovirus co-expressing RUNX3 and DsRed. Interestingly, RUNX3 overexpression also suppressed erythroid colony formation by 1.7±0.7 fold (p<0.01) compared to control, however, in a serial replating strategy, RUNX3 cells were able to form 2.7±0.6 fold more erythroid colonies than control, which suggests that RUNX3 promotes self-renewal. In the EPO-independent phase of development RUNX3 overexpression significantly inhibited the growth of erythroid progenitors by 2.7±1.3 fold (p<0.05), possibly arising from reduced erythroid commitment (CD36+CD13neg reduced by 1.5±0.2 fold, p<0.01). Erythroid committed cells showed a greater proliferative response to EPO (4.6±2.8 fold greater than controls) and showed phenotypic evidence of inhibition of development: decreased expression of GlyA (suppressed by 1.3±0.1 fold, p<0.01), retention of CD36 (1.4±0.2 fold, p<0.05) and 20% increased cell size (p<0.05) compared to controls. Morphologically, these cells appeared to be blocked in an early/intermediate stage of development. These data imply that RUNX3 overexpression impairs erythroid differentiation and blocks terminal differentiation.

Overall, these results show that both the overexpression and KD of RUNX3 in normal HSPC leads to a block in erythroid differentiation. Further, our preliminary data suggests that ectopic expression of RUNX3 in CD34+ HSPC selectively inhibits granulocytic development. These experimental models indicate that the level of RUNX3 expression is an important regulator of normal human hematopoietic development. We are currently determining the underlying mechanisms by which RUNX3 perturbs normal hematopoiesis, which ultimately may lead to new therapies for AML patients.


Knapper:Daiichi Sankyo: Honoraria; Jazz: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Novartis: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Pfizer: Membership on an entity's Board of Directors or advisory committees; Tolero: Membership on an entity's Board of Directors or advisory committees.

Author notes


Asterisk with author names denotes non-ASH members.