V211D Mutation in MEK1 Causes Resistance to MEK Inhibitors in Colon Cancer

The RAS/RAF/MEK/ERK pathway is a key driver of tumor growth in human cancers. Recurrent genomic alterations in this pathway occur most commonly in the KRAS, NRAS, and BRAF genes and activate the MEK (MAP2K) kinases to constitutively activate downstream signaling. Thus, MEK represents a promising target for therapies directed against this pathway. Highly potent, allosteric MEK inhibitors that bind to MEK and keep it in a closed, inactive conformation are now clinically available. The MEK inhibitors trametinib, cobimetinib, and binimetinib are all FDA approved together with RAF inhibitors to treat BRAF^V600-mutant melanoma. In addition, MEK inhibitors as single agents have been shown to enhance radioiodine uptake in advanced thyroid cancer (1) and to cause regression of neurofibromas in patients with neurofibromatosis type 1 (2) and of BRAF-mutant pediatric low-grade gliomas (3). Dramatic clinical responses have been observed with MEK inhibitors in a small number of patients with MEK1 mutations, suggesting that MEK inhibitors may be an effective treatment in at least a subset of MEK1-mutant patients (4, 5). Although mechanisms of acquired resistance to RAF/MEK combinations have been extensively studied, mechanisms that limit the activity of MEK inhibitors in patients have yet to be defined.

A 39-year-old woman with a BRAF^K601E-mutant metastatic colon cancer that involved the chest, abdominal wall, distant lymph nodes, and bones was treated with combined binimetinib and panitumumab for 6 weeks in a phase Ib/II trial sponsored by Novartis Pharmaceuticals and then Array BioPharma (NCT01927341; Fig. 1A). BRAF^K601E is an activating, non-V600 BRAF mutation that is unresponsive to RAF inhibitors (6), unlike BRAF^V600 alterations. Patients with colorectal cancers harboring activating non-V600 BRAF mutants do not clinically respond to anti-EGFR antibodies (manuscript provisionally accepted). Reactivation of EGFR signaling has been shown to limit the clinical activity of ERK pathway inhibitors in colorectal cancers (7, 8). In this patient, the clinical trial provided the opportunity to treat with the MEK inhibitor binimetinib to target ERK activation with the addition of the anti-EGFR antibody panitumumab to overcome reactivation of EGFR signaling after ERK inhibition. At 6 weeks, imaging showed a stable chest wall mass and an increase in the periosteal reaction and extraosseus soft-tissue component anterior to the right femur, and she underwent palliative fixation of the right hip for persistent pain (Fig. 1B). Next-generation sequencing with MSK-IMPACT (9) of the right femur bone tissue, obtained while she was on treatment, revealed a new, subclonal MEK1^V211D mutation (Fig. 1C). The MEK1^V211D mutation was not identified in biopsy specimens collected either soon after diagnosis from the chest wall metastasis (0/824 reads) or immediately before starting this treatment from an abdominal wall nodule (0/870 reads). A section of the right femur tumor was implanted in a mouse to generate a patient-derived xenograft (PDX) model, and sequencing suggested enrichment of the MEK1^V211D variant allelic fraction in the growing PDX (Fig. 1C).

To determine whether the BRAF^K601E and MEK1^V211D mutations arose in the same population of tumor cells, we performed single-cell DNA sequencing without whole-genome amplification of the cell line generated from the PDX (CLR36). A total of 5,895 cells were sequenced (Supplementary Table S1). The BRAF^K601E and MEK1^V211D alterations were found to co-occur in 92% of all cells (n = 5,423) with a median variant allelic frequency (VAF) of 75% and 50%, respectively, very similar to the VAFs identified from the bulk sequencing of the PDX (Fig. 1D and E; Supplementary Fig. S1A–S1E; Supplementary Table S2). Single-cell sequencing identified the concurrent mutations in three populations: a major clone (n = 5,095 cells) heterozygous for both BRAF^K601E and MEK1^V211D variants, a subclone (n = 267 cells) homozygous for BRAF^K601E and heterozygous for MEK1^V211D, and a subclone (n = 61 cells) heterozygous for BRAF^K601E and homozygous for MEK1^V211D.

Review of more than 30,000 advanced tumors analyzed with MSK-IMPACT (http://cbioportal.mskcc.org) and more than 250 colorectal cancers in The Cancer Genome Atlas (TCGA; ref. 10) identified no cases with the MEK1^V211D mutation. Thus, this mutant rarely occurs in nature, and this finding emphasizes its emergence as the result of treatment exposure in this patient.

To characterize MEK1^V211D functionally, we first examined whether this mutant could activate ERK signaling as compared with wild-type (WT) MEK. In NIH-3T3 cells, expression of MEK1^V211D induced higher levels of pMEK and pERK than WT MEK1 does in both serum-containing and serum-starved conditions (Fig. 2A). However, serum starvation reduced ERK activation in both WT MEK1 and MEK1^V211D-mutant expressing cells. The decrease of MEK/ERK phosphorylation in the MEK1^V211D-expressing cells could be due to the inhibition of endogenous MEK proteins or the mutant MEK1. We tested whether the phosphorylation and kinase activity of MEK1^V211D is still regulated by upstream RAF kinase. We purified GST-tagged WT MEK1 and MEK1^V211D-mutant proteins and performed an in vitro kinase assay in the absence or presence of active BRAF kinase. Purified WT MEK1 was not phosphorylated in the absence of RAF kinase, nor could it phosphorylate ERK. Addition of activated BRAF kinase induced both the phosphorylation and kinase activity of WT MEK1. In contrast, we found MEK1^V211D was phosphorylated and could phosphorylate ERK in the absence of activated BRAF, suggesting this mutant has acquired RAF-independent phosphorylation and basal activity. The phosphorylation and kinase activity of MEK1^V211D could be further enhanced with the addition of activated BRAF kinase (Fig. 2B). In addition, under the same reaction conditions, MEK1^V211D was more effectively phosphorylated by RAF kinase than was the WT MEK protein. This could be responsible for its increased sensitivity to RAF-mediated kinase activation. Therefore, MEK1^V211D is among the class of RAF-regulated MEK1 mutants (11) that we recently defined as having autonomous kinase activity that can be further activated by RAF and more effectively transduces RAF activity downstream to ERK. In the patient, the MEK1^V211D mutant developed in a tumor with an activating BRAF^K601E mutation that signals independently of RAS (Supplementary Fig. S2A). The MEK1^V211D mutant would be expected to further activate signaling in the setting of activated BRAF, and thus amplify ERK signaling in this tumor.

We evaluated the effects of the V211D mutation on the interactions between MEK1 and its kinase and substrate (Supplementary Fig. S2B). We expressed either WT MEK1 or V211D MEK1 together with WT or K601E BRAF or with WT ERK1 or ERK2 in 293H cells and performed immunoprecipitation. The MEK1^V211D mutation did not affect MEK1 binding to WT or K601E BRAF. However, the binding of MEK1 to ERK was reduced by the MEK1^V211D mutation. This is likely due to the elevated kinase activity of the MEK1^V211D mutant versus WT MEK1.

To determine whether the MEK1^V211D allele affects sensitivity to allosteric MEK inhibitors, we tested the effects of binimetinib in NIH-3T3 cells expressing WT MEK1 or MEK1^V211D mutant. Binimetinib potently inhibited ERK activation at a dose of 0.1 μmol/L in vector-expressing parental cells and cells with ectopic expression of WT MEK1, whereas pERK remained unaffected by 3 μmol/L binimetinib in cells expressing MEK1^V211D (Fig. 3A). These data suggest that, in cells, MEK1^V211D-driven ERK activation is insensitive to binimetinib treatment. Silencing the expression of MEK1^V211D in CLR36 cells sensitized the cells to treatment with binimetinib (Supplementary Fig. S3A and S3B). Similarly, this MEK1 mutation also decreased the sensitivity of ERK signaling to another allosteric MEK inhibitor, cobimetinib (Fig. 3B). To understand the mechanism underlying this insensitivity to MEK inhibitors, we tested whether the V211D mutation causes resistance of MEK1 to these drugs in vitro. Purified GST-tagged WT MEK1 and MEK1^V211D-mutant proteins were incubated with increasing doses of MEK inhibitors and their kinase activity was assessed by an in vitro kinase assay using inactive ERK2 as substrate. Consistent with what we observed in cells, the activity of MEK1^V211D, reflected in pERK levels, remained unchanged following increasing doses of either binimetinib or cobimetinib, compared with potent inhibition of WT MEK1 activity by the above two inhibitors (Fig. 3C and D). These data suggest that MEK1^V211D is sufficient to cause resistance to multiple allosteric MEK inhibitors both in vitro and in cells.

Indeed, MEK1^V211D was previously implicated as a resistance allele to diarylamine MEK inhibitor (AZD6244 or CI-1040) in a random mutagenesis screen. On the basis of mapping the mutant allele within the three-dimensional structure of the full-length MEK1 kinase domain, Emery and colleagues suggested that the V211D mutation, situated directly within the arylamine-binding pocket, may cause resistance by direct interference with drug binding (12). To evaluate the structural effects of this mutant, structural models of MEK1 WT and MEK1^V211D were generated using template-based modeling and molecular dynamics simulations (Fig. 3E). MEK1^D211 residue forms a hydrogen bond to nearby MEK1 residues, which does not occur in WT MEK1 and results in displacement of D211 from the WT position by 7 ångströms. A zoomed-in image with cobimetinib shows that D211 MEK1 is pulled away from its WT position and faces away from the drug's binding site. The hydrophobic carbon atoms of V211 that interact with cobimetinib are lost, as D211 is not a hydrophobic residue. Our data indicate that V211D is a gatekeeper mutation for allosteric MEK inhibitors. Furthermore, our findings also suggest that MEK1^V211D displays enhanced kinase activity in addition to its effect in reducing drug binding, which promotes resistance to allosteric MEK inhibitors.

We previously reported that allosteric MEK inhibitor–insensitive MEK1 mutants which exhibit RAF-independent activity could be effectively treated by a selective ATP-competitive MEK inhibitor, MAP855, through direct interference with ATP binding (11). We thus hypothesized that MAP855 could also inhibit MEK1^V211D-driven ERK signaling by targeting its ATP site. We tested the activity of MAP855 in MEK1^V211D-expressing NIH-3T3 cells and found that this drug inhibited ERK activation driven by either WT MEK or MEK1^V211D at similar doses, although MEK1^V211D-expressing cells had higher initial phospho-ERK (Fig. 4A). Using the additional ATP-competitive MEK inhibitor BI-847325 (13), we confirmed that these results were not compound-specific and that WT and V211D MEK1 exhibited similar sensitivity to same type ATP-competitive MEK inhibitors (Supplementary Fig. S3C). Consistent with these findings, the kinase activity of MEK1^V211D was inhibited by MAP855 at equal potency compared with WT MEK1 in vitro (Fig. 4B). We tested whether the patient's tumor might be sensitive to MAP855 in the PDX model derived from the progressing right femur lesion, which produced tumors that continued to grow with binimetinib treatment either alone or in combination with the EGFR antibody cetuximab. In contrast, MAP855 treatment at a nontoxic dose led to around 30% tumor regression (Fig. 4C; Supplementary Fig. S4). Consistently, MAP855 potently inhibited ERK signaling and tumor proliferation (Ki-67) and induced the apoptosis marker cleaved capsase-3 in the PDX tumors, which were resistant to either binimetinib alone or combined binimetinib/cetuximab treatment (Fig. 4D and E). Taken together, our data suggest that ATP-competitive MEK inhibition represents a novel therapeutic strategy for tumors with acquired resistance to current allosteric MEK inhibitors.

Previous studies in cell line models have identified multiple mechanisms for acquired resistance to allosteric MEK inhibitors, including amplification of upstream oncogenic drivers of the ERK pathway in BRAF- or KRAS-mutant colorectal cancer cells (14, 15), or MEK1 mutations in both helix A and the allosteric binding pocket of MEK protein (12). In our recent work, we have shown that MEK1 mutations exhibit allele-specific mechanisms of ERK activation (11). A subset of MEK1 mutants acquire RAF-independent kinase activity. The degree of autonomous ERK activation varies across mutants and can be further enhanced by RAF activation (RAF-regulated mutants) or totally independent of RAF. In addition, we showed that the RAF-independent activities of MEK1 mutants reduced their sensitivity to current MEK inhibitors. However, the clinical relevance of the proposed resistance mechanisms from cell line models needs validation in tumor samples from patients treated with MEK inhibitors. Our study reports the first case of a patient with cancer who acquired a MEK1^V211D mutation in a progressing tumor. We further propose a strategy to overcome this resistance mechanism using a new class of ATP-competitive MEK inhibitor and demonstrate its efficacy in the PDX made from this patient's progressing tumor. Our results suggest that treatment with the ATP-competitive MEK inhibitor is a rational therapeutic strategy for patients whose tumors exhibit acquired resistance to allosteric MEK inhibitors.

So far, we have not identified any cases of the MEK1^V211D mutation in a review of more than 30,000 clinical specimens sequenced at Memorial Sloan Kettering Cancer Center (MSKCC) and in TCGA, suggesting that this mutation does not arise in the absence of therapy. In this patient, MEK1^V211D occurred in the setting of a BRAF^K601E-activating mutation upon drug treatment and the two alterations are in the same cells, validating our finding that RAF-regulated MEK1 mutants can co-occur with upstream alterations to amplify BRAF signaling. Interestingly, the MEK1^V211D mutation was first discovered in a screen of resistance mechanisms to MEK inhibitors in the background of BRAF^V600E melanoma cells (12). In the absence of drug, the resultant hyperactivation of ERK signaling may have led to a growth disadvantage in cells. This is also reflected by the low occurrence of the hyperactive RAF-independent MEK1 mutants (11). In clinical samples from this patient, the MEK1^V211D mutant was not detected, even with deep-tumor sequencing, prior to targeted therapy treatment and emerged as a resistance alteration to treatment. These data suggest that in this patient, the MEK1^V211D was likely present only in a rare subclone that was then selected with drug exposure or acquired rapidly after treatment.

The genomic background of mutant BRAF^K601E may have affected the resistance alteration seen in this case. Our group has recently shown different functional properties of allele-specific BRAF alterations and has classified BRAF mutants into three groups (16, 17). Class 1 BRAF mutants consist of BRAF^V600 alterations, are highly activating, and can signal as monomers independent of RAS. Class 2 BRAF alterations, such as K601E, are activating, but often less so than V600E, and signal as RAS-independent dimers. This case suggests that in tumors with less-activating alterations, such as non-V600 BRAF mutations, secondary mutations may develop to amplify ERK signaling, and these alterations may attenuate the effect of ERK pathway inhibitors.

Limitations of our study include that only 1 patient with resistance to MEK inhibitor treatment was studied and biopsy specimens were analyzed so multiregional samples for each metastatic site were not available. However, consistent with our finding that alterations that amplify BRAF signaling can confer resistance to MEK inhibitors in the clinic, MEK mutations were identified at resistance to MEK inhibitors in 2 patients with BRAF^V600E melanoma (12, 18). In the first patient, treated with selumetinib, post-progression tissue harbored MEK1^P124L, a RAF-dependent MEK1 mutant that amplifies ERK signaling from activated BRAF (11). The other patient was treated with trametinib and developed concurrent BRAF amplification and MEK2^Q60P, an alteration analogous to the RAF-regulated MEK1^Q56P mutation (11), at progression. Together, these data suggest that MEK alterations that increase ERK pathway activation represent a clinically relevant, recurrent mechanism of resistance to allosteric MEK inhibitors, and these alterations would still be sensitive to ATP-competitive MEK inhibitors (11).

In summary, we report and functionally characterize a mechanism of acquired resistance to MEK inhibitors in the clinic. We find in a patient with colon cancer that MEK1^V211D emerged with treatment and caused resistance by amplifying ERK activation and interfering with allosteric inhibitor drug binding. Our data suggest that this resistance to current MEK inhibitors could be overcome by a selective ATP-competitive inhibitor by its binding to a different site on the MEK protein.

The patient provided written informed consent to treatment in the clinical trial. Progression biopsies and collection of patient samples were conducted under appropriate Institutional Review Board protocols (#06-107, 14-019). DNA from pretreatment and disease-progression specimens were analyzed using MSK-IMPACT (Integrated Mutation Profiling of Actionable Cancer Targets), a targeted exome capture assay with deep-sequencing coverage. Target-specific probes for hybrid selection were designed as previously described to capture all protein-coding exons of more than 300 oncogenes, tumor suppressor genes, and components of pathways deemed actionable by targeted therapies (9). All studies were conducted in accordance with the Declaration of Helsinki.

Patient-derived tumor models were generated by mincing about 1 g of tumor tissue, mixing it with Matrigel (50%), and implanting subcutaneously into NSG (NOD/SCID gamma) mice (Institutional Review Board protocols 06-107, 14-091). The PDX generated was sequenced to confirm the genomic alterations present. A cell line was generated from the PDX by growing about 1 g of tumor tissue from the PDX in McCoy media.

The cell line generated from the PDX was subjected to single-cell sequencing (please see Supplementary Methods for full details). A total of 250,000 cells were used for the barcoding run. The droplet workflow for genomic DNA amplification and barcoding was done as described previously (19). Libraries were analyzed on a DNA 1000 assay chip with a Bioanalyzer (Agilent Technologies) and sequenced on a Illumina HiSeq 2500 instrument (Illumina, Inc.). Sequence data were analyzed using the proprietary software provided by Mission Bio (19).

NIH-3T3 and Phoenix-AMPHO cells were purchased from ATCC between 2013 and 2015. Cells were grown in DMEM with glutamine, antibiotics, and 10% FBS. Cell lines were validated by short tandem repeat profiling at the Integrated Genomics Operation of MSKCC and screened for Mycoplasma using MycoAlert Plus Mycoplasma Detection Kit from Lonza.

NIH-3T3 cells were used to construct stable lines with inducible expression of mutant MEK1s to study MEK1 mutant–driven ERK signaling and their responses to different types of MEK inhibitors. Cell lines were used within 3 months of passages post-receipt for the above experiments.

Western blot, immunoprecipitation, and in vitro kinase assays were performed as described previously (11). The following antibodies were used: anti-p217/p221-MEK1/2 (pMEK1/2; #9154), anti-p202/p204-ERK1/2 (pERK1/2; #4370), anti-MEK1/2 (#4694), anti-ERK1/2 (#4696), anti-p380-p90RSK(pRSK; #9341), GAPDH (#2118) from Cell Signaling Technology, anti-V5 (R960-25) from Thermo Fisher Scientific, and anti-BRAF (sc-5284) and anti-cyclin D1 (M-20) from Santa Cruz Biotechnology.

The MEK1 gene was subcloned into pGEX-6-P1 (Addgene) for in vitro protein purification. Plasmids TTIGFP-MLUEX and pMSCV-rtTA3-PGK-Hygro for inducible gene expression were provided by Scott Lowe's laboratory at MSKCC. The MEK1 gene was subcloned into TTIGFP-MLUEX vector harboring the Tet-responsive promoter. Mutations were introduced by using the site-directed Mutagenesis Kit (Stratagene).

Binimetinib and cobimetinib were obtained from Selleckchem. MAP855((1-((3S,4S)-4-(8-(2-chloro-4-(pyrimidin-2-yloxy)phenyl)-7-fluoro-2-methyl-1H-imidazo[4,5-c]quinolin-1-yl)-3-fluoropiperidin-1-yl)-2-hydroxyethanone)) (11) was obtained from Novartis (compound No. 1, WO2015022662). These drugs were dissolved in DMSO to yield 10 mmol/L stock and stored at −20°C. Cetuximab for in vivo experiments was purchased from the hospital pharmacy.

Retroviruses encoding rtTA or MEK1 were packaged in Phoenix-AMPHO cells. The supernatant containing virus was filtered with 0.45 μm polyvinylidene difluoride membrane. The target cells were infected with virus for 8 hours. Forty-eight hours later, cells were selected in medium containing puromycin (2 μg/mL) or hygromycin (100 μg/mL) for 3 days. The positive infected cell populations were further sorted using GFP as a marker after overnight exposure to 1 μg/mL doxycycline. GFP-positive cells were then cultured and expanded in medium with doxycycline along with antibiotics.

Human WT MEK1, as well as the V211D mutant used in this study, were subcloned into pGEX-6P-1, expressed as glutathione-S-transferase fusions and purified by Pierce Glutathione Agarose (Thermo Fisher Scientific).

In vitro kinase assays were conducted in the presence of 200 μmol/L ATP, at 30°C for 15 minutes. Briefly, GST-MEK1 or mutants were incubated in the absence or presence of active BRAF^V600E protein (Upstate). Changes in MEK1 phosphorylation were estimated by immunoblotting for pMEK. To test the kinase activity of WT or mutant MEK1 protein, recombinant-inactive ERK2 protein (GenWay Biotech) was used as a substrate and the reaction was terminated with the addition of 1X SDS loading buffer and boiling. Kinase activity was estimated by immunoblotting for pERK.

The patient-derived tumor was implanted as subcutaneous xenografts into 6-week-old NSG mice (Jackson Laboratories), and treatments started when tumors reached 100 mm³ volumes. Mice (5/group) were randomized to each treatment arm and observed daily throughout the treatment period for signs of morbidity/mortality. Body weights were recorded twice weekly. Tumors were measured twice weekly using calipers, and volume was calculated using the formula length × width² × 0.52. All studies were performed in compliance with institutional guidelines under an Institutional Animal Care and Use Committee–approved protocol. Investigators were not blinded when assessing the outcome of in vivo experiments.

The structure of MEK1 WT was generated using I-Tasser (v5.1; ref. 20) with a published MEK1 structure PDB ID:5KKR (21) as a template and stabilized by 10 ns of molecular dynamics simulation using GROMACS (v5.1.4; ref. 22). To generate the mutant structure, UCSF Chimera (v1.12) was used to mutate V211D, before simulating for additional 10 ns to restabilize. See Supplementary Methods for complete details.

Results are mean values ± SDs. All cellular experiments were repeated at least three times.

J.S. Reis-Filho is a consultant at Goldman Sachs and is a consultant/advisory board member for Volition RX, Paige.AI, Roche Diagnostics, Genentech, Novartis, and Invicro. N. Rosen reports receiving commercial research support from Chugai, has ownership interest (including stock, patents, etc.) in BeiGene, Kura Oncology, and Araxes, and is a consultant/advisory board member for Chugai, BeiGene, AstraZeneca, Array BioPharma, Novartis BioMed, Revolution Medicines, and Boehringer Ingelheim. R. Yaeger reports receiving commercial research support from Array BioPharma, Genentech, GlaxoSmithKline, and Novartis Pharmaceuticals and is a consultant/advisory board member for Array BioPharma. No potential conflicts of interest were disclosed by the other authors.

Conception and design: Y. Gao, J.F. Hechtman, N. Rosen, Z. Yao, R. Yaeger

Development of methodology: Y. Gao, A. da Cruz Paula, A.N. Gorelick, B. Weigelt, H. Zhao, J.S. Reis-Filho, Z. Yao

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Gao, A. Maria, N. Na, J. Carson, B.S. Taylor, H. Zhao, J.S. Reis-Filho, E. de Stanchina, Z. Yao, R. Yaeger

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Gao, A. Maria, N. Na, A. da Cruz Paula, A.N. Gorelick, B. Weigelt, B.S. Taylor, J.S. Reis-Filho, N. Rosen, Z. Yao, R. Yaeger

Writing, review, and/or revision of the manuscript: Y. Gao, A. Maria, J.F. Hechtman, J. Carson, R.A. Lefkowitz, B. Weigelt, B.S. Taylor, J.S. Reis-Filho, N. Rosen, Z. Yao, R. Yaeger

Study supervision: Z. Yao

This study was supported by the Byrne Fund (R. Yaeger), the Sarah Jenkins Fund and Breast Cancer Research Foundation (J.S. Reis-Filho), and the NIH R01 CA233736 (N. Rosen and R. Yaeger), U54 OD020355 (E. de Stanchina), R01 CA204749 (B.S. Taylor), and Cancer Center Core Grant P30 CA 008748. This research is the responsibility of the authors and does not necessarily represent the official views of the NIH. This work was also supported by a Stand Up To Cancer Colorectal Dream Team Translational Research Grant (SU2C-AACR-DT22-17). Stand Up To Cancer (SU2C) is a division of the Entertainment Industry Foundation. Research grants are administered by the American Association for Cancer Research, the Scientific Partner of SU2C.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.