RUNX1-RUNX1T1 Induces Over-Expression of γ-Catenin in Human CD34⁺ Cells, Increasing Self Renewal and Impairing Granulocytic Differentiation.

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.