AML1-ETO (AE) provides an initial impetus towards leukemogenesis and is necessary for disease maintenance. Much work has been done to identify the domains required for AE function using a variety of different model systems. Here, we determined which AE functional domains are necessary to establish long-term self-renewal of human CD34+ cord blood cells. Modified forms of AE with deletions of the conserved ETO functional domains NHR1, 2, or 3/4 (termed Δ1, Δ2, and AE9a, respectively) had significant potential to extend the life of transduced cells. By comparison, cells expressing a mutant with combined deletion of NHR2 and AE9a (9aΔ2), or point mutants lacking the ability to bind DNA or CBFβ, had lifespans similar to vector transduced control cells. Human in vitro cultures expressing the mutant AE proteins Δ1, Δ2, or AE9a maintained a substantial population of CD34+ progenitor cells with increased proliferative capacity and CFU-C re-plating activity. All long-term cultures established using AE deletion mutants consistently expressed 2 to 10 times higher levels of the surrogate protein surface marker Thy1 (correlating with AE mutant mRNA levels due to an IRES-Thy1 element) than what is seen for cultures expressing full length AE. A functional selection was clearly driving this process, since transduction frequencies and initial expression levels of Thy1 were comparable to that of full length AE. Increased mutant protein expression was also observed in cells that successfully engrafted immunodeficient mice, suggesting that this was likely a necessary factor for mutant protein function. Interestingly, AE9a cultures occasionally generated granulocytic sarcomas near the site of intrafemoral injection, and also showed enhanced levels of engraftment in vivo, mimicking to some degree the transforming properties evident in mouse blood cells. Using the previously defined AE9a mouse leukemia model, we transduced murine fetal liver cells and sorted cells based on expression levels of a GFP surrogate marker. Cells expressing high levels of AE9a performed better in re-plating assays compared to low expressing cells, again implying a selection process based on enhanced function in response to increased levels of mutant protein. In contrast, there was little difference in colony forming ability between cells expressing low and high levels of the full-length AE protein. Finally, by tracking GFP percentage in the peripheral blood throughout AE9a-induced murine AML development, we noted an obvious intensity shift leading up to disease endpoint, with all AE9a AML samples showing intense expression of the mutant protein. Together, these data indicate that increased expression levels of AE mutant proteins are an important determinant for promoting self-renewal (and leukemia in the case of AE9a) in model systems. Interestingly, recent studies have shown that t(8;21)+ AML with high and persistent levels of AE9a have a worse prognosis than t(8;21)+ AML with low AE9a levels (Jiao et al., Leukemia 23(9):1598-1604, 2009; Ommen et al., Eur J Haematol 84(2):128-132, 2009). We are currently analyzing the signaling cascades initiated by enhanced mutant protein expression in comparison to those initiated by full-length AE. Together, these data highlight the dynamic nature of the AE fusion protein and will serve significant value in defining the mechanisms of AE involvement in the self-renewal process using a highly relevant human model system.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.