KRAS is one of the most frequently mutated oncogenes in human cancer. Despite substantial efforts, no clinically applicable strategy has yet been developed to effectively treat KRAS-mutant tumors. Here, we perform a cell-line-based screen and identify strong synergistic interactions between cell-cycle checkpoint-abrogating Chk1- and MK2 inhibitors, specifically in KRAS- and BRAF-driven cells. Mechanistically, we show that KRAS-mutant cancer displays intrinsic genotoxic stress, leading to tonic Chk1- and MK2 activity. We demonstrate that simultaneous Chk1- and MK2 inhibition leads to mitotic catastrophe in KRAS-mutant cells. This actionable synergistic interaction is validated using xenograft models, as well as distinct Kras- or Braf-driven autochthonous murine cancer models. Lastly, we show that combined checkpoint inhibition induces apoptotic cell death in KRAS- or BRAF-mutant tumor cells directly isolated from patients. These results strongly recommend simultaneous Chk1- and MK2 inhibition as a therapeutic strategy for the treatment of KRAS- or BRAF-driven cancers.

Cell 162, 146-159; July 2, 2015) Our paper presented a new algorithm, named PreCISE, designed to identify synergistic drug interactions that are effective at killing cancer cells harboring specific driver mutations. Using this platform and a cell-line-based screen, we identified a synergistic drug interaction between Chk1-and MK2 inhibitors in KRASor BRAF-driven cells, and that combination of therapy focused on these two kinase inhibitors is effective at inducing cell death of KRASand BRAF mutant tumors in vivo.While reviewing the paper after publication, we noticed that we had included two erroneous duplications of western blot loading control bands in the final version of Figure 2E. This figure shows that simultaneous inhibition of Chk1 and MK2 induces genotoxic stress and apoptosis in several KRAS-driven cancer cell lines and respective controls. The incorrect loading controls are presented for the HSP27 blot for the H1703 cell line, as well as for the CDC25B blot for the H1437 cell line. The errors occurred when we were copying each blot to construct the final figure. We recovered the original autoradiographs of these experiments and now provide a new version of the figure containing the correct loading controls. The correct data supports the original interpretation of the experiment, and the conclusions of the paper remain unchanged. In addition, we observed a typo in Figure S3A, showing the distribution of all cell lines used in the initial screen of our paper. In the pie chart, the slice of the pie in purple representing ''lung squamous'' cell lines was incorrectly labeled with n = 18. The correct value is n = 3, as depicted in the figure legend and in the main text. Both figures are now corrected online. We regret not being able to identify these errors before and sincerely apologize for any inconvenience they may have caused.