Volume 4, Issue 6, Pages 1090-1099 (September 2013) Concurrent MEK2 Mutation and BRAF Amplification Confer Resistance to BRAF and MEK Inhibitors in Melanoma  Jessie Villanueva, Jeffrey R. Infante, Clemens Krepler, Patricia Reyes-Uribe, Minu Samanta, Hsin-Yi Chen, Bin Li, Rolf K. Swoboda, Melissa Wilson, Adina Vultur, Mizuho Fukunaba-Kalabis, Bradley Wubbenhorst, Thomas Y. Chen, Qin Liu, Katrin Sproesser, Douglas J. DeMarini, Tona M. Gilmer, Anne-Marie Martin, Ronen Marmorstein, David C. Schultz, David W. Speicher, Giorgos C. Karakousis, Wei Xu, Ravi K. Amaravadi, Xiaowei Xu, Lynn M. Schuchter, Meenhard Herlyn, Katherine L. Nathanson  Cell Reports  Volume 4, Issue 6, Pages 1090-1099 (September 2013) DOI: 10.1016/j.celrep.2013.08.023 Copyright © 2013 The Authors Terms and Conditions

Cell Reports 2013 4, 1090-1099DOI: (10.1016/j.celrep.2013.08.023) Copyright © 2013 The Authors Terms and Conditions

Figure 1 A De Novo MEK2 Mutation in Trametinib-Resistant Melanoma (A) MRI images from a 62-year-old white male with metastatic melanoma, with involvement of the lungs, chest wall, heart, and liver prior to treatment with trametinib and at different times after treatment initiation. VLA, vertical long axis; SAX, short axis. (B) Sequenom iPLEX assay depicting nucleotide determination on pretreatment and at progression samples. Specific assay depicted is for MEK2-Q60P mutation. A, wild-type nucleotide; C, mutant nucleotide. (C) Fold amplification of BRAF in human samples determined by qPCR. Normal DNA control; FFPE, formalin-fixed paraffin-embedded sample; pre-tx-Tram, trametinib pretreatment; post-tx-Tram, progression on trametinib; pre-tx-Dabra, dabrafenib pretreatment. Error bars denote SD for four replicate PCRs. (D–G) Melanoma cells were treated with MEK (D and E) or BRAF (F and G) inhibitors for 72 hr. Cell viability was calculated relative to the untreated cells. Data are represented as mean ± SEM with n = 3. (H and I) Mel1617 parental and trametinib-resistant (MR) sublines were treated with trametinib (H) or PLX4720 (I) for 20 hr. Protein lysates were analyzed by immunoblotting. See also Figure S1. Cell Reports 2013 4, 1090-1099DOI: (10.1016/j.celrep.2013.08.023) Copyright © 2013 The Authors Terms and Conditions

Figure 2 A MEK2-Q60P Mutation Decreases Sensitivity to BRAF and MEK Inhibitors (A) Chromatogram of Sanger sequencing depicting the de novo Q60P mutation (c.179A > C, p.Q60P) identified in exon 2 of MAP2K2 in BRAF-V600E melanoma cells resistant to trametinib. (B) Structure of MEK1 (Protein Data Bank accession number 3EQC) bound to ATP and the MEK inhibitor selumetinib (AZD6244) highlighting the position of the MEK1-Q56P (MEK2-Q60P, red stick) mutation. Only the N-terminal kinase lobe of MEK1 is shown as a cartoon with the ATP and AZD6244 molecules shown as stick figures. The figure was generated with PyMol. (C) 293T cells were mock infected or infected with WT-MEK2 or mutant MEK2-Q60P. Cells were serum starved for 48 hr. Cell lysates were analyzed by immunoblotting. (D–G) Mel1617 parental cells were mock infected (MI) or infected with a lentivirus carrying wild-type MEK2 (WT-M2), or mutant MEK2-Q60P (M2-Q60P). (D and E) Immunoblotting analysis of cells treated with trametinib (Tram) or PLX4720 (PLX) for 20 hr. (F and G) Cell viability was determined by MTT assays and calculated relative to the untreated cells. Data are represented as mean ± SEM, n = 6. (H–K) Mel1617-MR cells were infected with lentiviral vectors expressing a nontargeting small hairpin RNA (shRNA) (shNT2) or MEK2 shRNA (shM2-1 and shM2-2). Cell lysates were analyzed by immunoblotting (H and I). Relative cell viability was determined by MTT assays (J and K). Data are represented as mean ± SEM, n = 5. See also Figure S2 and Table S1. Cell Reports 2013 4, 1090-1099DOI: (10.1016/j.celrep.2013.08.023) Copyright © 2013 The Authors Terms and Conditions

Figure 3 BRAF-V600E Overexpression and Mutant MEK2 Promote Drug Resistance (A) Targeted amplification of the BRAF locus on chromosome 7 demonstrated with aCGH in trametinib-resistant cells. (B) BRAF mRNA levels in Mel1617 and 451Lu sublines assessed by quantitative real-time PCR. Data are represented as mean ± SD (n = 6); p < 0.05. (C) BRAF protein levels were analyzed by immunoblotting in parental and resistant sublines. (D–G) Mel1617-MR cells were infected with vector control (pLKO.1) or BRAF shRNA. MAPK signaling was assessed by immunoblotting (D and F) and cell viability by MTT assays (E and G). Data are represented as mean ± SEM (n = 7). Data were analyzed using Wilcoxon rank sum test; p < 0.001 when comparing Mel1617-MR-pLK0.1 with Mel1617-MR expressing BRAF shRNA at ≥100 nM tram (E) or ≥5,000 nM PLX (G). (H and I) Mel1617 ectopically expressing WT-BRAF, BRAF-V600E or MEK-Q60P, and trametinib-resistant Mel1617 (MR) cells were treated with 0.1 μM of trametinib (H) or 1 μM PLX4720 (PLX; I). Cell viability was calculated relative to the untreated control. Data are represented as mean ± SEM (n = 6); p < 0.01 when comparing Q60P/V600E with Q60P or V600E and p > 0.05 when comparing Q60P/V600E with MR cells treated with PLX. See also Figure S3. Cell Reports 2013 4, 1090-1099DOI: (10.1016/j.celrep.2013.08.023) Copyright © 2013 The Authors Terms and Conditions

Figure 4 Targeting S6K Can Halt the Growth of Trametinib-Resistant Tumors In Vivo (A and B) Mice bearing tumors derived from Mel1617 (A) or Mel-1617-MR (B) sublines were treated with vehicle or trametinib (3 mg/kg po qd; n = 5/group). Tumor volumes were plotted over time (mean ± SD). Tumor growth trends were significantly different between treatment groups in Mel1617 (p < 0.001), but no significant difference was observed in Mel1617-MR (p = 0.11). (C) Mice bearing trametinib-resistant tumors were treated with dabrafenib (30 mg/kg), trametinib (3 mg/kg), or the combination of both drugs at the same doses. Dabrafenib-treated tumors grew more rapidly and animals were sacrificed at day 15. Mice were treated with trametinib or the combination for 22 days. There was no significant difference between treatment groups (p = 0.727). (D) Immunoblotting analysis of tumors isolated 4 hr after the last treatment. (E and F) Parental and MR sublines were treated with PLX4032 or trametinib for 22 hr. Cell lysates were analyzed by immunoblotting. (G) Mice bearing trametinib-resistant tumors were treated with vehicle, 458 (3 mg/kg) dabrafenib (Dab; 30 mg/kg) plus trametinib (Tram; 0.3 mg/kg), or a combination of three drugs at the same doses. Tumors in mice treated with vehicle or dabrafenib plus trametinib grew more rapidly, and animals were sacrificed at day 15. Mice were treated with 458 or the triple combination for 21 days. Tumor growth was significantly faster in the Dab + Tram group (p < 0.0001) and the 458 single-agent group (∗p = 0.005) compared with the triple combination (Dab + Tram + 458). Tumor growth rate of the Dab + Tram group was significantly faster than the group treated with 458 (p < 0.001). (H and I) A short-term culture derived from a CRPDX, WM3942, was treated with the indicated inhibitors for 24 hr and analyzed by immunoblotting (H) or 72 hr for MTT assays (I). Data are represented as mean ± SEM (n = 7). See also Figure S4. Cell Reports 2013 4, 1090-1099DOI: (10.1016/j.celrep.2013.08.023) Copyright © 2013 The Authors Terms and Conditions

Figure S1 Resistance to BRAF and MEK Inhibitors Is Coupled to MAPK Reactivation, Related to Figure 1 (A) Immunohistochemical analysis of paired BRAF mutant melanoma tumor biopsies at pretreatment, day 15, and progression. The patient was administered a loading-dose regimen of 6/6/2 mg QD of trametinib. Progression biopsy was obtained a month after trametinib was discontinued. (B–E) 451Lu parental and 451Lu-MR trametinib-resistant sublines were treated with the indicated concentrations of MEK (B and C) or BRAF (D and E) inhibitors for 72h. Cell viability was assessed by MTT assays and calculated relative to the DMSO (vehicle control)-treated cells. Data are represented as mean ± SEM (n = 6). (F) Mel1617 parental and Mel1617-MR trametinib resistant sublines were treated with the indicated concentrations of paclitaxel (top) or carboplatin (bottom) for 72h. Cell viability was determined by MTT assays. Data are represented as mean ± SEM (n = 6). (G) Mel1617 parental and Mel1617-MR trametinib resistant cells were treated with the indicated doses of AZD6244 or dabrafenib for 20h. Equal amounts of total proteins were resolved by SDS-PAGE and analyzed by immunoblotting using the indicated antibodies. (H–K) 451Lu parental and isogenic trametinib-resistant (MR) sublines were treated with increasing concentrations of the indicated MEK (H and I) or BRAF (J and K) inhibitors for 20h. Equal amounts of protein lysates were resolved by SDS-PAGE and analyzed by immunoblotting. Cell Reports 2013 4, 1090-1099DOI: (10.1016/j.celrep.2013.08.023) Copyright © 2013 The Authors Terms and Conditions

Figure S2 A De Novo MEK2-Q60P Mutation Decreases Sensitivity to MEK and BRAF Inhibitors, Related to Figure 2 (A) A375P BRAF-V600E melanoma cells were transduced with wild-type MEK2 (WT) or MEK2-Q60P. Transduced cells were treated with trametinib or dabrafenib for 72h. MEK2 wild-type transduced cells are more sensitive to trametinib (top) (IC50 = 4 ± 0.2 nM) and dabrafenib (bottom) (IC50 = 24 ± 4 nM) than cells expressing MEK2-Q60P (trametinib IC50 = 16 ± 2 nM; dabrafenib IC50 = 110 ± 17 nM). (B–E) Mel1617-MR cells were infected with lentiviral vectors expressing a nontargeting control shRNA (shNT2) or MEK2 shRNA (shM2-1, shM2-2). Infected cells were sorted by GFP expression. (B and C) Cell lines were treated with DMSO, the indicated doses of AZD6244, or dabrafenib for 20h. Cell lysates were analyzed by immunoblotting. (D and E) Mel1617 parental and Mel1617-MR infected cell lines were treated with increasing doses of AZD6244 or dabrafenib for 72h. Cell viability was calculated relative to untreated cells. Representative results of three independent experiments are shown. Data are represented as mean ± SEM, n = 5. (F–M) Mel1617-MR cells were infected with lentiviral vectors expressing a nontargeting control shRNA (shNT2) or MEK1 shRNA (shM1-3, shM1-3/1). Infected cells were sorted by GFP expression. Cell lines were treated with DMSO or the indicated doses of trametinib, PLX4720, AZD6244, or dabrafenib for 20h. Cell lysates were analyzed by immunoblotting (F–I). For cell viability assays Mel1617 parental and Mel1617-MR infected cell lines were treated with increasing doses of the indicated compounds for 72h (J–M). Cell viability was calculated relative to untreated cells. Representative results of three independent experiments are shown. Data are represented as mean ± SEM, n = 5. (N–Q) Mel1617 cells were infected with lentiviral vectors expressing empty vector, pLKO.1, MEK2 shRNA (shM2-1, shM2-2) or MEK1 shRNA (shM1-1, shM1-3). Cell lysates were analyzed by immunoblotting (N and O). Mel1617 parental cells were treated with increasing doses of trametinib (P and Q) for 72h. Cell viability was calculated relative to untreated cells. Data are represented as mean ± SEM, n = 7. Data shown in panels P and Q are from one representative experiment; the same Mel1617-pLKO.1 control is depicted in both panels. Cell Reports 2013 4, 1090-1099DOI: (10.1016/j.celrep.2013.08.023) Copyright © 2013 The Authors Terms and Conditions

Figure S3 Analysis of Potential Mechanisms of Resistance in Trametinib-Resistant Cells, Related to Figure 3 (A) Mel-1617 parental and trametinib-resistant sublines were treated with DMSO, 0.1 μM dabrafenib or 0.1 μM trametinib for 24 hr. RTK arrays were probed with whole cell lysates. Target proteins were visualized with fluorescent-based detection and fluorescent intensity was normalized to DMSO-treated parental cells. Data represent mean value ± SD (n = 2). (B) DNA isolated from parental and trametinib-resistant sublines was analyzed by aCGH. No difference in copy number of MAP3K8/COT (top) and NF1 (bottom) are seen in the parental and trametinib-resistant sublines (C) Cell lysates from parental and trametinib-resistant cells were analyzed for immunoblotting with the indicated antibodies. All sublines expressed the tumor suppressor PTEN and express similar levels of MAP3K8/COT. Although additional bands (<70KD) were detected with a monoclonal BRAF antibody, these bands were present in cell lysates from parental (sensitive) and resistant sublines. (D) Mel1617 ectopically expressing WT-BRAF, BRAF-V600E or MEK-Q60P, and trametinib-resistant Mel1617 (MR) cells were treated with 1 μM PLX4720 (PLX). Cell lysates were analyzed by immunoblotting with the indicated antibodies. Cell Reports 2013 4, 1090-1099DOI: (10.1016/j.celrep.2013.08.023) Copyright © 2013 The Authors Terms and Conditions

Figure S4 Resistance to the Combination of BRAF and MEK Inhibitors Is Linked to Sustained Phosphorylation of S6K, Related to Figure 4 (A and B) Mel1617 parental cells ectopically expressing mutant MEK2-Q60P at low (A) or high (B) levels were grown as tumor xenografts in NSG mice. Once tumors were > 200 mm3 mice were randomized into two groups and treated orally with 3 mg/kg trametinib once a day. Tumor volume was measured with calipers. (C and D) Tumor volume of 451Lu parental (C) and 451Lu-MR drug resistant (D) xenografts treated with vehicle or trametinib (3 mg/kg po qd) (mean ± SEM, n = 5). A likelihood ratio testing nested model showed that the trends of relative tumor growth are significantly different between the two sublines (p = 0.002). A mixed-effect model analysis indicated that tumor growth is significantly slower after treatment with trametinib in the parental cells (p = 0.033) compared with the resistant cells. Wilcoxon rank sum test shows that the tumor weight at the end of experiment is significant lower in the parental cells treated with trametinib compared with vehicle control (p = 0.008), whereas there is no significant difference in tumor weight at the end of experiment in resistant cells treated with trametinib compared with vehicle control (p = 0.117). (E) Total proteins were extracted from xenograft tumors derived from Mel1617 parental, Mel1617-MR (MR), or Mel1617 cells ectopically expressing low (Q60P-low) or high (Q60P-high) levels of mutant MEK2 treated with vehicle (Veh) or trametinib (Tram; 3 mg/kg po qd). Mice were sacrificed 4h after receiving the last dose of trametinib. The effect of trametinib on the MAPK pathway was assessed by immunoblotting with anti-pERK antibodies. BRAF levels are also shown and histone H3 (H3) was used as protein loading control. (F) Tumors derived from 451Lu or 451Lu-MR cells were extracted 4h after the last dose and total protein lysates analyzed by immunoblotting with the indicated antibodies. (G and H) Mel1617-MR cells (G) or A375 cells expressing MEK2-Q60P (H) were treated with dabrafenib (Dabr), trametinib (Tram), or the combination of both drugs (Comb). Total cell lysates were analyzed by immunoblotting. (I–K) Mel1617-MR (I), Mel1617 expressing MEK2-Q60P (J), and Mel1617-MR cells expressing MEK2 shRNA (shM2-2; K) were treated with single agent trametinib or dabrafenib or the combination of both drugs at the indicated doses. Cell viability was assessed by MTT assays after 72h of drug treatment and calculated relative to untreated controls. Data are represented as mean ± SEM, n = 7. (L and M) 451Lu parental and isogenic trametinib resistant (MR) sublines were treated with increasing concentrations of trametinib (L) or PLX4720 (M) for 20h. Equal amounts of protein lysates were resolved by SDS-PAGE and analyzed by immunoblotting. (N) Mel1617-MR cells were treated with the indicated drugs (0.1 μM) for 24h. Cells were collected, lysed and analyzed by immunoblotting with the indicated antibodies. (O) Sequenom iPLEX assay depicting MEK2 nucleotide determination. DNA was isolated from a Xenograft tumor derived from a progression biopsy of a patient treated with the combination of dabrafenib and trametinib. Specific assay depicted is for MEK2-Q60P mutation. “A” = wild-type nucleotide; “C” = mutant nucleotide. Cell Reports 2013 4, 1090-1099DOI: (10.1016/j.celrep.2013.08.023) Copyright © 2013 The Authors Terms and Conditions