Genome-Wide CRISPR/Cas9 Screen Reveals That the Dcps Scavenger Decapping Enzyme Is Essential for AML Cell Survival

Genome-wide knockout screening employing CRISPR-Cas9 genome-editing is a powerful tool for functional genomics. However, identifying actionable targets for cancer therapy has been challenging due, in part, to the complex genetic background of cell lines used for screening. To overcome this obstacle, we generated two acute myeloid leukemia (AML) lines from mouse leukemia models based on activity of the leukemia oncogenes CALM/AF10 or MLL/AF9. Both lines exhibited a normal karyotype and intact Tp53 activity. Using these lines, we performed genome-wide CRISPR-Cas9 screening, followed by a second screen in vivo . We then selected genes meeting the following criteria: 1) they encoded a protein with an available inhibitor or 2) their germline mutation loss-of-function phenotype was relatively moderate based on the literature or the human exome-sequencing database. We excluded genes with a well-defined function in leukemogenesis as well as those encoding components of basal cellular machineries.

Among genes significantly depleted in our primary screen was the mRNA decapping enzyme scavenger (Dcps), which encodes a mRNA 5' cap binding protein implicated in mRNA decay. Read counts for each Dcps-targeted single-guide RNA (sgRNA) significantly decreased in vitro (AML cell lines) and in vivo (mouse AML model). A negative selection CRISPR-Cas9 mutagenesis scan of all Dcps coding exons revealed that the C-terminal Dcps domain, namely aa 230-240, plays a critical role in AML survival.

RG3039, an orally active quinazoline derivative, is a DCPS inhibitor that was originally developed to treat spinal muscular atrophy (SMA) and has been judged safe in a phase I trial in healthy volunteers (Van Meerbeke JP et al. Hum Mol Genet. 2013). We validated RG3039 binding to DCPS protein in AML cells via a cellular thermal shift assay (CETSA). We then assessed anti-leukemia effects of RG3039 by treating human AML lines with RG3039 and generating growth curves. That analysis showed that RG3039 had dose-dependent anti-proliferative effects. RG3039 treatment induced cell cycle arrest and apoptosis, revealed by EdU incorporation assay and Annexin V stain, respectively.

Since DCPS protein was predominantly nuclear in human primary AML cells, we hypothesized that DCPS primarily functions in that compartment rather than in the cytoplasmic mRNA 3' end decay pathway. To search for nuclear proteins potentially interacting with DCPS, we undertook immunoprecipitation with an anti-DCPS antibody of lysates of AML cells followed by mass spectrometry analysis. Among highly significant interactors, we identified components of pre-mRNA processing machineries including spliceosomes. To assess whether DCPS inhibition would impair pre-mRNA splicing, we treated AML cells with RG3039 and performed RNA-sequencing to determine potential transcriptome-wide splicing changes. Alternative 5' splice site selection was most frequently observed in RG3039-treated cells, and most aberrant splicing events involved the first exon. Bioinformatic prediction analysis revealed that approximately 40% of the aberrantly-spliced genes were NMD (nonsense-mediated mRNA decay)-sensitive. Gene Set Enrichment Analysis identified gene signatures representing a type-I interferon response, reminiscent of that observed in RNAi-treated cells, in RG3039-treated AML cells.

We next explored effects of DCPS deficiency on normal hematopoiesis in humans. To do so, we examined peripheral blood counts of three children harboring germline homozygous loss-of-function mutations as well as heterozygous relatives in a family reported previously (Ng et al. Hum Mol Genet 2015). They exhibited normal blood counts, indicating that DCPS is dispensable for steady-state hematopoiesis in humans.

Finally, to investigate potential anti-leukemia effects of RG3039 in vivo, we tested RG3039 efficacy using patient-derived xenograft (PDX) AML models established from three human AML lines. RG3039 exhibited anti-leukemic activity, as evidenced by the lower leukemia burden in PB and BM of RG3039-treated mice. They survived significantly longer than vehicle-treated mice, indicative of therapeutic efficacy of RG3039 monotherapy against AML in vivo.

In summary, our findings shed a new light on a pre-mRNA metabolic pathway regulated by DCPS. They also identify DCPS as a novel target for AML therapy and suggest potential "repurposing" of RG3039 as an anti-leukemia drug.