The t(8;21)(q22;q22) occurs frequently in acute myelogenous leukaemia and gives rise to the transcription factor fusion protein, RUNX1-RUNX1T1 (also known as AML1-ETO). To identify the genes dysregulated by the aberrant transcriptional activity of RUNX1-RUNX1T1, we used microarrays to determine the effect of this mutation on gene expression in human progenitor cells and during subsequent development. Gene signatures of these developmental subsets were very dissimilar indicating that effects of RUNX1-RUNX1T1 are highly context dependent. We focused on gene changes associated with the granulocytic lineage and identified a clinically relevant subset of these by comparison with 235 leukaemia patient transcriptional signatures. We confirmed the overexpression of a number of significant genes (Sox4, IL-17BR, CD200 and c-catenin). Further, we show that overexpression of CD200 and c-catenin is also associated with the inv(16) abnormality which like RUNX1-RUNX1T1 disrupts core binding factor activity. We investigated the functional significance of CD200 and c-catenin overexpression in normal human progenitor cells. The effect of IL17 on growth was also assessed. Individually, none of these changes were sufficient to recapitulate the effects of RUNX1-RUNX1T1 on normal development. These data provide the most comprehensive and pertinent assessment of the effect of RUNX1-RUNX1T1 on gene expression and demonstrate the highly context-dependent effects of this fusion gene.

No abstract

The t(8;21) is one of the most frequent translocations in AML occurring at a frequency of 10–15% and at 30–40% incidence in cases of AML-FAB subtype M2. This translocation results in expression of the chimaeric transcription factor RUNX1-RUNX1T1 (aka AML1-ETO). Using a human cord blood model, we have previously reported that expression of this fusion gene inhibits granulocytic differentiation, a feature which is characteristic of 8;21 leukaemias. This model therefore provides a means to probe the mechanism by which RUNX1-RUNX1T1 participates in the inhibition of differentiation and the promotion of self-renewal. Data increasingly suggest that RUNX1-RUNX1T1 induces the dysregulation of target genes critical for haematopoietic differentiation. Here we report the use of Affymetrix HG-U133A microarrays (22,283 probe sets) to determine the effect of RUNX1-RUNX1T1 on gene expression during the development of human blood cell progenitors (CD34+ cells). These cells were isolated by MACS and RUNX1-RUNX1T1 expression was achieved by retroviral infection of these cells using a vector co-expressing GFP. Transduced cells were subsequently cultured in IL-3, SCF and G/GM-CSF supplemented medium. Using this strategy we generated matched RUNX1-RUNX1T1 and control populations representing progenitor cells (day 3) and individual developmental subsets isolated after 6 days of culture. The absolute Genechip data were imported into Genespring software and filtered to remove all probe sets that changed by <1.5 fold between control and RUNX1-RUNX1T1 transduced cells and were not statistically significant using ANOVA or paired t-test (P<0.05). We identified 26 genes that were differentially expressed between control and RUNX1-RUNX1T1 transduced myeloblasts (days 3–6). To focus on genes relevant to leukaemogenesis we also compared them to AML t(8;21) diagnostic (n=11) and normal donor (n=9) samples obtained from patients who were treated in the UK NCRI-AML Trial. In each of these data sets expression of RUNX1-RUNX1T1 was associated with over-expression of γ-catenin (>2.6 fold), an observation confirmed at the protein level. Furthermore, t(8;21) patients also over-expressed γ-catenin compared to FAB M2 patients without cytogenetic abnormality. γ-Catenin (like β-catenin) is believed to be intimately linked to the control of differentiation and self-renewal of stem cells. We therefore over-expressed this protein in CD34+ cells to determine whether this was sufficient to account for the dysregulation of development observed in RUNX1-RUNXT1 cells. We found that, as with cells expressing RUNX1-RUNX1T1, over-expressing γ-catenin promoted retention of myeloid colony-forming ability, such that by the second week of culture, myeloid colony forming efficiency exceeded that of control cultures by 3 fold. (P<0.05). We are currently studying the effect of γ-catenin over-expression in bulk liquid culture in the presence of IL-3, SCF and G/GM-CSF. Under these conditions over-expression of γ-catenin does not appear to increase the long-term proliferation of these cells compared to control, however, morphologically an inhibition of granulocytic differentiation was observed. In summary, RUNX1-RUNX1T1 promotes γ-catenin expression and this in turn appears to participate in leukaemogenesis associated with FAB M2.