EHA Library - The official digital education library of European Hematology Association (EHA)

RUNX1/ETO is a driver leukaemic fusion protein produced by the translocation t(8;21) and accounts for 10-12% of Acute Myeloid Leukaemia (AML) cases. RUNX1/ETO is required for maintaining the leukaemic phenotype influencing both leukaemic clonogenicity and growth. Currently the molecular mechanism(s) by which it regulates leukaemic self-renewal and propagation are just beginning to be unravelled. In our previous work, we characterised the core-transcriptional network driven by RUNX1/ETO required to maintain the t(8;21) AML phenotype (Ptasinska et al., 2014).

We have now functionally characterised the significance of individual members of this network for leukaemic self-renewal by performing targeted RNAi screens both in tissue culture and in a xenotransplant setting.

To that end, we intersected RNAseq, CHIPseq and microarray data and identified a set of 103 genes comprising both direct RUNX1/ETO target genes and genes potentially cooperating with such targets. For the RNAi screens, we used a Doxycycline (Dox)-inducible lentiviral RNAi library covering each gene of this set with at least 3 shRNAs. We transduced two t(8;21)-positive AML cell lines (Kasumi-1 and SKNO-1) with this library and performed parallel screens employing colony formation and long-term suspension culture assays in the in vitro arm, and intrafemoral transplantation of highly immunodeficient NSG mice for the in vivo screen. Both test series contained Dox and no Dox groups. RNA and genomic DNA were isolated from transduced cells and subjected to targeted Next Generation Sequencing. Changes in shRNA pool compositions were identified by comparison of the corresponding Dox and no Dox groups using DESEQ.

The analysis of the in vitro screen demonstrated that l RUNX1/ETO shRNA constructs rapidly disappeared upon induction under all in vitro conditions, confirming the central role of RUNX1/ETO in maintaining t(8;21) AML, while levels of non-targeting control shRNA genes did not change during the course of the experiments. Moreover, RUVBL1 (Pontin) shRNAs were depleted in agreement with our previous finding of a dependence of t(8;21) AML cells on this factor. In addition, shRNAs targeting KIT which in both cell lines carries the activating mutation N822K, disappeared. These combined results demonstrate the functionality of this targeted screen. Most importantly, the top hits of this in vitro screen identified a group of self-renewal genes including those encoding the ubiquitin ligase gene SKP2 and Cyclin D2 (CCND2).

The more stringent in vivo screen confirmed CCND2 as a RUNX1/ETO target gene relevant for leukaemic propagation. Notably, this screen also identified several new genes, which had not scored in the in vitro setting. This included several genes involved in ubiquitination (UBASH3B, UBE2L6), regulation of G proteins (GPRC5C, ARHGEF12, ARHGEF3), transcriptional control (ID1, KLF2, ERG) or glycolysis (SCL2A3, PFKP). Importantly, several shRNAs targeting genes such as NOTCH2, SLA or LAPTM5, which are repressed by RUNX/ETO, were also depleted in the in vivo screen emphasising our previous observation that leukaemic self-renewal and propagation is driven by a dynamic equilibrium between RUNX1 and RUNX1/ETO.

We are now further characterising the contribution of these different target genes to the leukaemic self-renewal and propagation both in vitro and in vivo. The ultimate aim of our studies is the identification of RUNX1/ETO dependent pathways, whose combined inhibition may phenocopy the loss of this hard-to-target initiator and driver of leukaemogenesis.