A functional cancer genomics screen identifies a druggable synthetic lethal interaction between MSH3 and PRKDC.

2014 
Here, we use a large-scale cell line–based approach to identify cancer cell–specific mutations that are associated with DNA-dependent protein kinase catalytic subunit (DNA-PKcs) dependence. For this purpose, we profiled the mutational landscape across 1,319 cancer-associated genes of 67 distinct cell lines and identified numerous genes involved in homologous recombination–mediated DNA repair, including BRCA1 , BRCA2 , ATM , PAXIP , and RAD50 , as being associated with non-oncogene addiction to DNA-PKcs. Mutations in the mismatch repair gene MSH3 , which have been reported to occur recurrently in numerous human cancer entities, emerged as the most significant predictors of DNA-PKcs addiction. Concordantly, DNA-PKcs inhibition robustly induced apoptosis in MSH3 -mutant cell lines in vitro and displayed remarkable single-agent efficacy against MSH3 -mutant tumors in vivo . Thus, we here identify a therapeutically actionable synthetic lethal interaction between MSH3 and the non-homologous end joining kinase DNA-PKcs. Our observations recommend DNA-PKcs inhibition as a therapeutic concept for the treatment of human cancers displaying homologous recombination defects. Significance: We associate mutations in the MSH3 gene, which are frequently detected in microsatellite-instable colon cancer (∼40%), with a therapeutic response to specific DNA-PKcs inhibitors. Because potent DNA-PKcs inhibitors are currently entering early clinical trials, we offer a novel opportunity to genetically stratify patients who may benefit from a DNA-PKcs–inhibitory therapy. Cancer Discov; 4(5); 592–605. ©2014 AACR . See related commentary by Hemann, [p. 516][1] This article is highlighted in the In This Issue feature, [p. 495][2] [1]: /lookup/volpage/4/516?iss=5 [2]: /lookup/volpage/4/495?iss=5
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    52
    References
    55
    Citations
    NaN
    KQI
    []