S112 MLL FUSIONS AFFECT SPLICING TO INDUCE EXPLOITABLE METABOLIC DEPENDENCIES

Background: MLL fusions are aberrant transcription factors that induce highly aggressive and hard‐to‐treat leukemia with dismal prognosis. A better understanding of leukemia specific properties may reveal potential therapeutic intervention points. Aims: Known direct downstream targets of MLL‐ENL were scored for their importance for leukemic cell survival and genes with a positive readout were further investigated for their targeting potential. Methods: A shRNA screen interrogating known MLL‐ENL target genes identified the splice‐modulator PTBP1 (polypyrimidin tract binding protein 1) as rate limiting for proliferation of MLL‐ENL transformed murine cells and also in a human patient cell line. Results: Reduction of PTBP1 slowed down cell cycle without inducing apoptosis and it reduced the competitive fitness of knock‐down cells compared to controls. This was accompanied by a conspicuous phenotype characterized by reduced medium acidification during culture. Examination of the splicing pattern of PKM transcripts (pyruvate kinase muscle‐type), a known PTBP1 splicing‐target, revealed a shift towards the PKM1 isoform upon PTBP1 ablation. PKM1/PKM2 differ in the utilization of an alternate exon and the ratio of the isoforms is one of the major control points switching from anabolic (PKM2 high) to catabolic (PKM1 high) metabolism. Mechanistically PKM2 restricts the outflow of glycolytic products into oxidative phosphorylation enabling the provision of proliferating cells with necessary anabolic intermediates. PKM2 fosters glycolysis at the expense of oxidative phosphorylation thus reducing energy yields obtained from glucose (Warburg effect). As a consequence cells increase glucose uptake and recycle reduction equivalents through production of lactate. PTBP1 knock‐down reverted this phenotype, reduced glucose consumption and lowered lactate production compared to controls. Because MLL fusions obviously hardwire cells into an anabolic‐proliferative state that requires large amounts of building blocks we tested the effect of starvation mimicking drugs (SMD) affecting glucose utilization, amino acid synthesis and energy production on MLL transformed and on primary hematopoietic precursors (Kit+ fraction of bone marrow). Whereas the immediate impact of SMDs on highly proliferative normal and MLL‐ENL transformed cells was comparable, the long term effects after an intermittent treatment were dramatically different. In contrast to their normal counterpart MLL‐ENL transformed cells lost a significant portion of their colony forming potential upon treatment indicating substantial damage suffered during episodes of starvation. On a molecular level this was accompanied by drastically higher ROS (reactive oxygen species) and an altered p53 response in MLL transformed cells. Summary/Conclusion: The feasibility to exploit a differential sensitivity towards selective starvation for therapeutic purposes is currently explored and options for a possible translation into a clinically applicable strategy will be discussed.