Pharmacological Induction of RAS-GTP Confers RAF Inhibitor Sensitivity in KRAS Mutant Tumors

•NRAS but not KRAS mutant tumors are more sensitive to RAF kinase inhibition•Combined RAF-MEK or RAF-PI3K inhibition exhibits synergy in RAS mutant cancers•KRAS-G13D mutant tumors exhibit higher synergy with combined RAF-MEK inhibition•Drug-induced RAF dimers and RAS-GTP levels confer sensitivity to RAF inhibition Targeting KRAS mutant tumors through inhibition of individual downstream pathways has had limited clinical success. Here we report that RAF inhibitors exhibit little efficacy in KRAS mutant tumors. In combination drug screens, MEK and PI3K inhibitors synergized with pan-RAF inhibitors through an RAS-GTP-dependent mechanism. Broad cell line profiling with RAF/MEK inhibitor combinations revealed synergistic efficacy in KRAS mutant and wild-type tumors, with KRASG13D mutants exhibiting greater synergy versus KRASG12 mutant tumors. Mechanistic studies demonstrate that MEK inhibition induced RAS-GTP levels, RAF dimerization and RAF kinase activity resulting in MEK phosphorylation in synergistic tumor lines regardless of KRAS status. Taken together, our studies uncover a strategy to rewire KRAS mutant tumors to confer sensitivity to RAF kinase inhibition. Targeting KRAS mutant tumors through inhibition of individual downstream pathways has had limited clinical success. Here we report that RAF inhibitors exhibit little efficacy in KRAS mutant tumors. In combination drug screens, MEK and PI3K inhibitors synergized with pan-RAF inhibitors through an RAS-GTP-dependent mechanism. Broad cell line profiling with RAF/MEK inhibitor combinations revealed synergistic efficacy in KRAS mutant and wild-type tumors, with KRASG13D mutants exhibiting greater synergy versus KRASG12 mutant tumors. Mechanistic studies demonstrate that MEK inhibition induced RAS-GTP levels, RAF dimerization and RAF kinase activity resulting in MEK phosphorylation in synergistic tumor lines regardless of KRAS status. Taken together, our studies uncover a strategy to rewire KRAS mutant tumors to confer sensitivity to RAF kinase inhibition. This study delineates the mechanistic principles of treating RAS mutant tumors with pan-RAF inhibitors. We find that NRAS mutant tumor cells are more sensitive to pan-RAF inhibitors compared with KRAS mutant tumors. Treatment with MEK or PI3K inhibitors sensitizes subsets of KRAS mutant and WT tumors to pan-RAF inhibitors. In the context of MEK inhibition, this sensitization is a result of increased RAS-GTP levels and BRAF-CRAF dimerization, resulting in increased RAF kinase activity. Moreover, tumors harboring KRASG13D mutations exhibit greater combination synergy compared with KRASG12 mutants, providing a potential predictive clinical biomarker. Our study thus uncovers a “chemical synthetic lethal” strategy to rewire KRAS mutant tumors through pharmacological induction of RAF kinase activity. RAS mutations are the most frequently observed oncogenic mutations, found in almost 20% of all cancers (Cerami et al., 2012Cerami E. Gao J. Dogrusoz U. Gross B.E. Sumer S.O. Aksoy B.A. Jacobsen A. Byrne C.J. Heuer M.L. Larsson E. et al.The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data.Cancer Discov. 2012; 2: 401-404Crossref PubMed Scopus (9406) Google Scholar, Gao et al., 2013Gao J. Aksoy B.A. Dogrusoz U. Dresdner G. Gross B. Sumer S.O. Sun Y. Jacobsen A. Sinha R. Larsson E. et al.Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal.Sci. Signal. 2013; 6: pl1Crossref PubMed Scopus (8709) Google Scholar). Despite their prevalence, effective treatment paradigms continue to be an area of unmet need. RAS is a small GTPase that toggles from an inactive guanosine diphosphate (GDP)-bound state to an active guanosine triphosphate (GTP)-bound state, where it mediates protein-protein interactions with effector proteins in several downstream pathways including the mitogen-activated protein kinase (MAPK), phosphatidylinositol 3-kinase (PI3K), and RAL-GDS pathways (Stephen et al., 2014Stephen A.G. Esposito D. Bagni R.K. McCormick F. Dragging ras back in the ring.Cancer Cell. 2014; 25: 272-281Abstract Full Text Full Text PDF PubMed Scopus (593) Google Scholar). Mutations in RAS render the protein in a constitutively active GTP-bound state where it can bind to the RAS-binding domain (RBD) of effectors such as RAF (with three isoforms, ARAF, BRAF, and CRAF) and PI3K. In the context of RAF, interaction with RAS-GTP facilitates membrane localization, RAF dimerization and phosphorylation, resulting in induction of RAF kinase activity and phosphorylation of its substrate MEK1/2 (Wellbrock et al., 2004Wellbrock C. Karasarides M. Marais R. The RAF proteins take centre stage.Nat. Rev. Mol. Cell Biol. 2004; 5: 875-885Crossref PubMed Scopus (929) Google Scholar). Activation of the MAPK pathway results in increased cell growth and proliferation. Once activated, the duration and amplitude of MAPK signaling is regulated through a network of negative feedback loops mediated by ERK, the terminal kinase in the signaling cascade. ERK suppresses pathway signaling either directly through phosphorylation of upstream components, such as SOS or CRAF, or transcriptionally through the regulation of expression of MAPK phosphatases (DUSP family members) and the Sprouty family of proteins (Dougherty et al., 2005Dougherty M.K. Muller J. Ritt D.A. Zhou M. Zhou X.Z. Copeland T.D. Conrads T.P. Veenstra T.D. Lu K.P. Morrison D.K. Regulation of Raf-1 by direct feedback phosphorylation.Mol. Cell. 2005; 17: 215-224Abstract Full Text Full Text PDF PubMed Scopus (463) Google Scholar, Owens and Keyse, 2007Owens D.M. Keyse S.M. Differential regulation of MAP kinase signalling by dual-specificity protein phosphatases.Oncogene. 2007; 26: 3203-3213Crossref PubMed Scopus (635) Google Scholar, Pratilas et al., 2009Pratilas C.A. Taylor B.S. Ye Q. Viale A. Sander C. Solit D.B. Rosen N. (V600E)BRAF is associated with disabled feedback inhibition of RAF-MEK signaling and elevated transcriptional output of the pathway.Proc. Natl. Acad. Sci. USA. 2009; 106: 4519-4524Crossref PubMed Scopus (449) Google Scholar). Analogous to RAF kinase activation, PI3K also interacts with RAS-GTP through its RBD and is activated at the membrane (Pacold et al., 2000Pacold M.E. Suire S. Perisic O. Lara-Gonzalez S. Davis C.T. Walker E.H. Hawkins P.T. Stephens L. Eccleston J.F. Williams R.L. Crystal structure and functional analysis of Ras binding to its effector phosphoinositide 3-kinase gamma.Cell. 2000; 103: 931-943Abstract Full Text Full Text PDF PubMed Google Scholar). This results in phosphorylation of phosphatidylinositol 4,5-bisphosphate (PIP2) to PIP3, which then recruits AKT to the membrane via its pleckstrin homology domain and subsequently activates its downstream components resulting in enhancement of cell growth and survival (Fruman et al., 2017Fruman D.A. Chiu H. Hopkins B.D. Bagrodia S. Cantley L.C. Abraham R.T. The PI3K pathway in human disease.Cell. 2017; 170: 605-635Abstract Full Text Full Text PDF PubMed Scopus (1178) Google Scholar, Zhao and Vogt, 2008Zhao L. Vogt P.K. Class I PI3K in oncogenic cellular transformation.Oncogene. 2008; 27: 5486-5496Crossref PubMed Scopus (459) Google Scholar). Mutations within BRAF and PI3K have been observed in subsets of cancers, with PI3K mutations enriched for in breast cancers (Baselga, 2011Baselga J. Targeting the phosphoinositide-3 (PI3) kinase pathway in breast cancer.Oncologist. 2011; 16: 12-19Crossref PubMed Scopus (215) Google Scholar) and BRAF mutations enriched for in melanoma and thyroid cancers (Cancer Genome Atlas Research Network, 2014Cancer Genome Atlas Research Network Integrated genomic characterization of papillary thyroid carcinoma.Cell. 2014; 159: 676-690Abstract Full Text Full Text PDF PubMed Scopus (1823) Google Scholar, Foster et al., 2016Foster S.A. Klijn C. Malek S. Tissue specific mutations in BRAF and EGFR necessitate unique therapeutic approaches.Trends Cancer. 2016; 2: 699-701Abstract Full Text Full Text PDF PubMed Scopus (3) Google Scholar, Hodis et al., 2012Hodis E. Watson I.R. Kryukov G.V. Arold S.T. Imielinski M. Theurillat J.P. Nickerson E. Auclair D. Li L. Place C. et al.A landscape of driver mutations in melanoma.Cell. 2012; 150: 251-263Abstract Full Text Full Text PDF PubMed Scopus (1845) Google Scholar). Interestingly, BRAF mutations are mutually exclusive with KRAS mutations, whereas PI3K mutations can co-occur with KRAS mutations, suggesting that KRAS and the RAF family members may have overlapping functions. Further, there continues to be accumulating evidence that RAF, specifically the CRAF isoform, plays a key role in RAS mutant tumors (Blasco et al., 2011Blasco R.B. Francoz S. Santamaria D. Canamero M. Dubus P. Charron J. Baccarini M. Barbacid M. c-Raf, but not B-Raf, is essential for development of K-Ras oncogene-driven non-small cell lung carcinoma.Cancer Cell. 2011; 19: 652-663Abstract Full Text Full Text PDF PubMed Scopus (224) Google Scholar, Cisowski et al., 2016Cisowski J. Sayin V.I. Liu M. Karlsson C. Bergo M.O. Oncogene-induced senescence underlies the mutual exclusive nature of oncogenic KRAS and BRAF.Oncogene. 2016; 35: 1328-1333Crossref PubMed Scopus (58) Google Scholar, Karreth et al., 2011Karreth F.A. Frese K.K. DeNicola G.M. Baccarini M. Tuveson D.A. C-Raf is required for the initiation of lung cancer by K-Ras(G12D).Cancer Discov. 2011; 1: 128-136Crossref PubMed Scopus (106) Google Scholar, Sanclemente et al., 2018Sanclemente M. Francoz S. Esteban-Burgos L. Bousquet-Mur E. Djurec M. Lopez-Casas P.P. Hidalgo M. Guerra C. Drosten M. Musteanu M. Barbacid M. c-RAF ablation induces regression of advanced Kras/Trp53 mutant lung adenocarcinomas by a mechanism independent of MAPK signaling.Cancer Cell. 2018; 33: 217-228.e4Abstract Full Text Full Text PDF PubMed Scopus (63) Google Scholar). Based on these findings, there has been intense focus on targeting the MAPK pathway as a strategy to treat KRAS mutant tumors either through combinations of agents targeting the MAPK pathway or dual inhibition of the MAPK and PI3K pathways. To date, there are three clinically approved kinase inhibitors targeting BRAFV600, vemurafenib, dabrafenib, and encorafenib, and three approved inhibitors targeting MEK, cobimetinib, trametinib, and binimetinib. These inhibitors are approved for the treatment of BRAFV600 mutant metastatic melanomas and not KRAS mutant tumors. Importantly, BRAFV600 inhibitors have been shown to activate rather than inhibit the MAPK pathway (termed “paradoxical activation”) in KRAS mutant tumors, limiting the use of this class of drugs to BRAFV600 mutant tumors (Hatzivassiliou et al., 2010Hatzivassiliou G. Song K. Yen I. Brandhuber B.J. Anderson D.J. Alvarado R. Ludlam M.J. Stokoe D. Gloor S.L. Vigers G. et al.RAF inhibitors prime wild-type RAF to activate the MAPK pathway and enhance growth.Nature. 2010; 464: 431-435Crossref PubMed Scopus (1281) Google Scholar, Poulikakos et al., 2010Poulikakos P.I. Zhang C. Bollag G. Shokat K.M. Rosen N. RAF inhibitors transactivate RAF dimers and ERK signalling in cells with wild-type BRAF.Nature. 2010; 464: 427-430Crossref PubMed Scopus (1397) Google Scholar). Paradoxical activation is largely dependent on the mode of inhibitor binding to the RAF kinase, and molecules have been classified as type I, 1.5, or II binders. Type 1.5 inhibitors, dabrafenib and vemurafenib, induce an outward shift of the αC helix within the kinase domain of BRAF (Karoulia et al., 2016Karoulia Z. Wu Y. Ahmed T.A. Xin Q. Bollard J. Krepler C. Wu X. Zhang C. Bollag G. Herlyn M. et al.An integrated model of RAF inhibitor action predicts inhibitor activity against oncogenic BRAF signaling.Cancer Cell. 2016; 30: 501-503Abstract Full Text Full Text PDF PubMed Scopus (18) Google Scholar). Because of this binding mode, they bind to BRAF monomers and robustly inhibit BRAFV600E, which can signal as monomers, but induce paradoxical activation in RAS mutant tumors. More recently a type 1.5 BRAFV600E inhibitor has been reported, PLX-8394, that binds to BRAFV600E effectively but does not promote paradoxical activation (Zhang et al., 2015Zhang C. Spevak W. Zhang Y. Burton E.A. Ma Y. Habets G. Zhang J. Lin J. Ewing T. Matusow B. et al.RAF inhibitors that evade paradoxical MAPK pathway activation.Nature. 2015; 526: 583-586Crossref PubMed Scopus (250) Google Scholar). In contrast, other classes of RAF inhibitors, such as AZ-628 or LY3009120, are classified as type II binders that leave the αC helix in the “in” position. These inhibitors can inhibit RAF monomers and dimers (Peng et al., 2015Peng S.B. Henry J.R. Kaufman M.D. Lu W.P. Smith B.D. Vogeti S. Rutkoski T.J. Wise S. Chun L. Zhang Y. et al.Inhibition of RAF isoforms and active dimers by LY3009120 leads to anti-tumor activities in RAS or BRAF mutant cancers.Cancer Cell. 2015; 28: 384-398Abstract Full Text Full Text PDF PubMed Scopus (193) Google Scholar) and have been reported to have some efficacy in RAS mutant as well as BRAFV600E mutant tumors while exhibiting minimal paradoxical activation. Despite multiple trials aimed at treating RAS mutant tumors with MEK inhibitors, there has been little efficacy observed clinically, and there are still no clinically approved MEK inhibitors for RAS mutant tumors. Combinations of MEK and PI3K inhibitors have not been tolerated in early clinical trials, precluding the development of these combinations despite the compelling pre-clinical evidence for efficacy in KRAS mutant tumors (Engelman et al., 2008Engelman J.A. Chen L. Tan X. Crosby K. Guimaraes A.R. Upadhyay R. Maira M. McNamara K. Perera S.A. Song Y. et al.Effective use of PI3K and MEK inhibitors to treat mutant Kras G12D and PIK3CA H1047R murine lung cancers.Nat. Med. 2008; 14: 1351-1356Crossref PubMed Scopus (1143) Google Scholar, Hoeflich et al., 2012Hoeflich K.P. Merchant M. Orr C. Chan J. Den Otter D. Berry L. Kasman I. Koeppen H. Rice K. Yang N.Y. et al.Intermittent administration of MEK inhibitor GDC-0973 plus PI3K inhibitor GDC-0941 triggers robust apoptosis and tumor growth inhibition.Cancer Res. 2012; 72: 210-219Crossref PubMed Scopus (204) Google Scholar). Furthermore, studies with various MEK inhibitors have shown that, due to inhibition of negative feedback loops, RAS mutant tumors become resistant to MEK inhibitors over time, resulting in rebound of MAPK signaling (Hatzivassiliou et al., 2013Hatzivassiliou G. Haling J.R. Chen H. Song K. Price S. Heald R. Hewitt J.F. Zak M. Peck A. Orr C. et al.Mechanism of MEK inhibition determines efficacy in mutant KRAS- versus BRAF-driven cancers.Nature. 2013; 501: 232-236Crossref PubMed Scopus (220) Google Scholar, Lito et al., 2014Lito P. Saborowski A. Yue J. Solomon M. Joseph E. Gadal S. Saborowski M. Kastenhuber E. Fellmann C. Ohara K. et al.Disruption of CRAF-mediated MEK activation is required for effective MEK inhibition in KRAS mutant tumors.Cancer Cell. 2014; 25: 697-710Abstract Full Text Full Text PDF PubMed Scopus (197) Google Scholar). Pre-clinical data combining MEK with ERK inhibitors have shown promising efficacy in KRAS-driven genetically induced mouse models of lung and pancreatic cancer (Merchant et al., 2017Merchant M. Moffat J. Schaefer G. Chan J. Wang X. Orr C. Cheng J. Hunsaker T. Shao L. Wang S.J. et al.Combined MEK and ERK inhibition overcomes therapy-mediated pathway reactivation in RAS mutant tumors.PLoS One. 2017; 12: e0185862Crossref PubMed Scopus (53) Google Scholar), but, in a phase 1 combination clinical trial, a tolerable regime could not be established (Weekes et al., 2017Weekes, C.D., Lockhard, A.C., LoRusso, P., Murray, E.R., Park, E., Tagen, M., Mueller, L., Dokainish, H., Shapiro, G.I., and Burris, H.A. (2017). A Phase Ib study to evaluate the MEK inhibitor cobimetinib in combination with the ERK1/2 inhibitor GDC-0994 in patients with advanced solid tumors[abstract]. Paper Presented at: AACR Annual Meeting (Washington, D.C., Cancer Research).Google Scholar). Furthermore, treatment of KRAS mutant cancer cell lines with RAF inhibitors in combination with MEK inhibitors was found to be efficacious in vitro, but whether this combination results in strong synergistic anti-tumor efficacy in vivo or in patients is unknown (Lamba et al., 2014Lamba S. Russo M. Sun C. Lazzari L. Cancelliere C. Grernrum W. Lieftink C. Bernards R. Di Nicolantonio F. Bardelli A. RAF suppression synergizes with MEK inhibition in KRAS mutant cancer cells.Cell Rep. 2014; 8: 1475-1483Abstract Full Text Full Text PDF PubMed Scopus (92) Google Scholar, Whittaker et al., 2015Whittaker S.R. Cowley G.S. Wagner S. Luo F. Root D.E. Garraway L.A. Combined Pan-RAF and MEK inhibition overcomes multiple resistance mechanisms to selective RAF inhibitors.Mol. Cancer Ther. 2015; 14: 2700-2711Crossref PubMed Scopus (50) Google Scholar). Importantly, it is unclear how RAS mutational status relates to response to the combination of RAF and MEK inhibition. Identifying subsets of RAS/MAPK-dysregulated tumors that might be more responsive to the combination of RAF and MEK inhibitors is of particular importance since treating patients with tumors that are predicted to be responsive to the combination of agents will be required to achieve tumor responses at tolerable combination clinical doses. To this end, the current clinical use of BRAFV600 and MEK inhibitors for the treatment of BRAFV600 mutant metastatic melanomas have identified safe and tolerable combination doses in defined patient populations, suggesting that there may be specific opportunities that could be explored for combining other classes of RAF inhibitors with MEK inhibitors. We assessed single-agent activity of a panel of RAF inhibitors in KRAS mutant versus BRAFV600E cell lines with type 1.5 and type II RAF inhibitors. The clinically approved vemurafenib and dabrafenib, as well as the pre-clinical “paradox breaker” PLX-8394 type 1.5 inhibitors, showed activity in BRAFV600E mutant cell lines but not in KRAS mutant cell lines (Figure 1A). As expected, vemurafenib and dabrafenib, but not PLX-8394, led to paradoxical activation of downstream phosphorylation of MEK (abbreviated pMEK) in KRAS mutant tumors (Figure S1A). In contrast to the type 1.5 inhibitors, type II inhibitors AZ-628 and LY3009120 show better inhibition of KRAS mutant cell lines. However, their potency in KRAS mutant cell lines was far weaker compared with the activity observed in the BRAFV600E mutant background (Figure 1A). Consistent with this activity, type II inhibitors, as well as PLX-8394, do not induce paradoxical activation (Figure S1A) (Zhang et al., 2015Zhang C. Spevak W. Zhang Y. Burton E.A. Ma Y. Habets G. Zhang J. Lin J. Ewing T. Matusow B. et al.RAF inhibitors that evade paradoxical MAPK pathway activation.Nature. 2015; 526: 583-586Crossref PubMed Scopus (250) Google Scholar). To expand upon these results, we screened three type II RAF inhibitors (LY-3009120, MNL-2480, and AZ-628) and PLX-8394 in a panel of 161 lung, skin, and colorectal cell lines (Figures 1B and S1B; Table S1). While there was a modest increase in activity of these inhibitors in RAS mutant lines compared with BRAF/RAS wild-type (WT) lines, BRAFV600E cell lines were the most sensitive (Figure 1B). We then analyzed a larger panel of 322 cell lines treated with AZ-628 and found that NRASQ61 mutant cell lines exhibited greater sensitivity to AZ-628 than KRAS mutant lines (Figure 1C; Table S2). In melanoma lines, sensitivity of NRASQ61 mutant lines was similar to that observed for BRAFV600E tumors (Figure S1C). These data suggest that lack of paradoxical activation is insufficient to portend sensitivity in the RAS mutant background, and that KRAS mutant tumors, unlike NRAS mutant tumors, are far less sensitive to RAF kinase inhibition. To determine whether a second pharmacological agent could sensitize KRAS mutant cell lines to RAF kinase inhibition, we screened a library consisting of 430 small-molecule tool compounds in combination with AZ-628 in the A549 KRAS mutant lung cancer cell line (Figure 1D; Table S3). The top hit from this screen was the MEK inhibitor cobimetinib. AZ-628 combined well with other MAPK pathway inhibitors, including four distinct MEK inhibitors and two ERK inhibitors in the top 20 hits. Other notable hits included several microtubule inhibitors as well as PI3K inhibitors pictilisib (GDC-0941) and taselisib (GDC-0032). To validate the hits from the primary screen, we re-screened the top hits in the HCT 116 (KRASG13D/PIK3CA mutant) cell line, which confirmed MEK inhibitors as the strongest agents that combine with type II RAF inhibitor AZ-628 (Table S3). While the top hit from the screen was an MEK inhibitor, not all MEK inhibitors were equally synergistic. Cobimetinib, pimasertib, refametinib, and PD901 scored highest among the compounds tested, whereas trametinib and GDC-0623 did not show synergy even though they are both potent MEK inhibitors (Figure S1D). We previously reported that MEK inhibitors have differential mechanisms of action depending on their ability to trap an inactive RAF-MEK complex (Hatzivassiliou et al., 2013Hatzivassiliou G. Haling J.R. Chen H. Song K. Price S. Heald R. Hewitt J.F. Zak M. Peck A. Orr C. et al.Mechanism of MEK inhibition determines efficacy in mutant KRAS- versus BRAF-driven cancers.Nature. 2013; 501: 232-236Crossref PubMed Scopus (220) Google Scholar). This drug-stabilized complex prevents RAF from phosphorylating MEK. We tested several MEK inhibitors with varying molecular mechanisms, and we found that MEK inhibitors that trap the inactive RAF-MEK complex, and hence do not induce pMEK (trametinib, GDC-0623, G-573, and CH-6766) (Figure 1E), are also less synergistic with AZ-628 in a full-dose Bliss matrix analysis (Figure 1F). We next investigated whether RAF inhibitors with different modes of binding exhibit similar synergistic activity with cobimetinib. We found that, while type II inhibitors readily combine with cobimetinib, type 1.5 inhibitors, including the paradox breaker PLX-8394, do not synergize with MEK inhibitors in KRAS mutant cell lines (Figure 2A). This suggests that paradoxical activation is not the primary reason type 1.5 RAF inhibitors do not synergize with MEK inhibitors in this setting. To test if these changes in sensitivity associate with changes in MAPK pathway signaling, we treated A549 and HCT 116 cells with increasing concentrations of vemurafenib and AZ-628 and a fixed dose of cobimetinib. Only AZ-628 in combination with cobimetinib effectively inhibited MAPK pathway output at the level of phosphorylated ERK (pERK) and phosphorylated RSK (pRSK) (Figure 2B). Pathway inhibition was restricted to the MAPK pathway as there was little observed impact on pAKT levels. Using RNA sequencing we examined canonical downstream transcriptional targets of MAPK signaling, DUSP6 and SPRY4, 6 hr after treatment with 0.1 μM AZ-628, 0.1 μM cobimetinib, or their combination, in four KRAS mutant lung cancer cell lines (Figure 2C). We observed greater than additive inhibition of MAPK target genes for the combination of AZ-628 and cobimetinib when compared with the sum of their effect as single agents (p < 0.01, moderated t test, Figure S2A). The combination also resulted in a significant induction of cell apoptosis after 48 hr of treatment as demonstrated by increased levels of cleaved PARP (Figure 2D) and a notable increase in the sub-G1 and G1 population as assessed by flow cytometry for bromodeoxyuridine-fluorescein isothiocyanate and propidium iodide (Figure 2E). To test whether these combination effects resulted in durable suppression of cell growth, we tested the RAF/MEK inhibitor combination in long-term colony formation assays. We observed a marked synergy between cobimetinib in combination with AZ-628 or LY3009120, but not with vemurafenib or PLX-8394 (Figure S2B). We next tested the type II RAF inhibitors LY3009120 and AZ-628 in combination with cobimetinib in NCI-H2122 lung (AZ-628 + cobimetinib) and HCT116 colon (LY3009120 + cobimetinib) xenograft tumor models. Consistent with the in vitro data (Figure S2C), our in vivo data indicate a robust combination effect of either inhibitor together with the MEK inhibitor when compared with the efficacy of either molecule alone (Figure 2F). While LY3009120 or cobimetinib as single agents modestly inhibited tumor growth, the combination was capable of arresting tumor cell growth in the NCI-H2122 model and effectively regressing the HCT 116 tumors. The combination was well tolerated in the mice, resulting in little or minimal changes in body weight post-treatment (Figure S2D). To assess the impact of the combination on MAPK signaling, tumor samples were collected 4 days post-treatment at various time points. Quantification of MAPK pathway signaling markers demonstrated better inhibition of pERK and pRSK, resulting in deeper suppression of downstream MAPK target genes DUSP6 and SPRY4 by the combination of LY3009120 and cobimetinib than either of the inhibitors alone at all time points tested (Figure S2E). Plasma and tumor drug concentrations confirmed that the improved inhibitory activity was not due to increased drug exposure but to a drug combination effect (Figure S2F). Taken together, the combination of RAF and MEK inhibition exhibited significant combination efficacy in vivo. To study the mechanism of synergy observed between cobimetinib and the type II RAF inhibitor AZ-628, we treated a panel of KRAS mutant and WT cell lines with cobimetinib for 24 hr and examined effects on MAPK pathway signaling. Following treatment with cobimetinib, we observed an increase in the pMEK levels in the KRAS mutant cells but not the KRAS WT cells (Figure 3A, input). In KRAS mutant cell lines treated with cobimetinib, BRAF strongly co-immunoprecipitated with CRAF (Figure 3A IP CRAF) and exhibited robust kinase activity as observed in an in vitro kinase assay (Figure 3A, kinase assay). This was not observed in the KRAS WT or BRAFV600E cell lines tested. The KRAS mutant lines also showed elevated RAS-GTP levels at baseline and upon cobimetinib treatment, likely explaining the induction of RAF dimers and RAF kinase activation (Figure 3A, RBD pull-down assay). The HCT 116 cell line showed less of an effect on RAS-GTP levels potentially due to elevated RAS-GTP levels due to the KRASG13D mutation (described in more detail below). Both the formation of the BRAF-CRAF heterodimer as well as the increase in kinase activity of the heterodimer were dose dependent upon addition of cobimetinib, occurring at 50–100 nM (Figures S3A and S3B), dose levels well below single-agent efficacy. We next examined the ability of RAF inhibitors to inhibit cobimetinib-induced RAF dimers. Cobimetinib-induced RAF dimers were immunoprecipitated from A549 cells, and then treated with AZ-628, LY3009120, or vemurafenib. Only the type II inhibitors, which can bind RAF dimers, were able to inhibit kinase activity shown by the decreased levels of pMEK (Figure 3B), consistent with the observed synergy observed in this cell line (Figure 2A). Taken together, these results demonstrate that, while 50–100 nM cobimetinib has little effect on cell growth on its own, it is sufficient to induce RAF dimerization and pathway reactivation enabling robust synergy with AZ-628 in 8-day clonogenic assays (Figure S2A). To test if mutant KRAS is required for this activity, we utilized siRNA targeting KRAS, and observed its effects on MAPK signaling in the presence or absence of cobimetinib. In each KRAS mutant cell line tested, KRAS knockdown dampened the cobimetinib-induced levels of pMEK as well as RAF dimer formation and in vitro kinase activity (Figure 3C). Collectively, these data indicate that KRAS plays an essential role in mediating reactivation of the MAPK pathway due to increased RAS-GTP levels following MEK inhibitor treatment. To determine whether particular genotypes are more sensitive to the combination of type II RAF inhibitors with MEK inhibitors, we screened 322 cell lines from indications in which RAS mutations are frequently found (pancreas, lung, colorectal, skin, ovary, and blood) with AZ-628 (starting at 20 μM) and cobimetinib (starting at 1 μM), and a co-dilution of both compounds (Table S4). We observed significantly higher synergy scores in RAS mutant cell lines compared with either BRAFV600 mutants or WT (non-RAS/BRAFV600 mutants) cell lines (Figure 4A) (p < 0.001, two-sided t test). Since the combination of MEK and ERK inhibitors was previously described as a potential treatment for RAS mutant cancer (Merchant et al., 2017Merchant M. Moffat J. Schaefer G. Chan J. Wang X. Orr C. Cheng J. Hunsaker T. Shao L. Wang S.J. et al.Combined MEK and ERK inhibition overcomes therapy-mediated pathway reactivation in RAS mutant tumors.PLoS One. 2017; 12: e0185862Crossref PubMed Scopus (53) Google Scholar) we also screened the combination of GDC-0994 (an ERK inhibitor) and cobimetinib. Surprisingly, this combination showed less synergy across tissue types and no association with the RAS/BRAF genotype (Figures 4A and S4A; Table S5). We determined full-dose matrix Bliss scores for the cobimetinib/GDC-0994 combination in A549 and NCI-H2122 cell lines and observed reduced synergy with the MEK/ERK combination compared with the RAF/MEK combination (Figure S4B). Using RNA sequencing we also examined canonical downstream transcriptional targets of MAPK signaling (DUSP6 and SPRY4) 6 hr after treatment with 0.1 μM GDC-0994, 0.1 μM cobimetinib, or the combination, in four KRAS mutant lung cancer cell lines (Figure S2A). In contrast to the RAF/MEK combination we did not observe significant differences in MAPK target gene expression with the MEK/ERK combination compared with the single agents (Figure S2A). To understand the greater synergistic effect of the RAF/MEK combination versus the MEK/ERK combination, we compared the reactivation of the MAPK pathway upon t