Exploiting MCL1 Dependency with Combination MEK + MCL1 Inhibitors Leads to Induction of Apoptosis and Tumor Regression in KRAS-Mutant Non–Small Cell Lung Cancer

Published OnlineFirst September 25, 2018; DOI: 10.1158/2159-8290.CD-18-0277

RESEARCH ARTICLE

Varuna Nangia1, Faria M. Siddiqui1, Sean Caenepeel2, Daria Timonina1, Samantha J. Bilton1, Nicole Phan1, Maria Gomez-Caraballo1, Hannah L. Archibald1, Chendi Li1, Cameron Fraser3, Diamanda Rigas2, Kristof Vajda1, Lorin A. Ferris1, Michael Lanuti4, Cameron D. Wright4, Kevin A. Raskin5, Daniel P. Cahill6, John H. Shin6, Colleen Keyes7, Lecia V. Sequist8,9, Zofia Piotrowska8,9, Anna F. Farago8,9, Christopher G. Azzoli8,9, Justin F. Gainor8,9, Kristopher A. Sarosiek3, Sean P. Brown10, Angela Coxon2, Cyril H. Benes1,9, Paul E. Hughes2, and Aaron N. Hata1,8,9

ABSTRACT BH3 mimetic drugs, which inhibit prosurvival BCL2 family proteins, have limited single-agent activity in solid tumor models. The potential of BH3 mimetics for these cancers may depend on their ability to potentiate the apoptotic response to chemotherapy and targeted therapies. Using a novel class of potent and selective MCL1 inhibitors, we demonstrate that concurrent MEK + MCL1 inhibition induces apoptosis and tumor regression in KRAS-mutant non–small cell lung cancer (NSCLC) models, which respond poorly to MEK inhibition alone. Susceptibility to BH3 mimetics that target either MCL1 or BCL-xL was determined by the differential binding of proapoptotic BCL2 proteins to MCL1 or BCL-xL, respectively. The efficacy of dual MEK+ MCL1 blockade was augmented by prior transient exposure to BCL-xL inhibitors, which promotes the binding of proapoptotic BCL2 proteins to MCL1. This suggests a novel strategy for integrating BH3 mimetics that target different BCL2 family proteins for KRAS-mutant NSCLC.

SIGNIFICANCE: Defining the molecular basis for MCL1 versus BCL-xL dependency will be essential for effective prioritization of BH3 mimetic combination therapies in the clinic. We discover a novel strategy for integrating BCL-xL and MCL1 inhibitors to drive and subsequently exploit apoptotic dependencies of KRAS-mutant NSCLCs treated with MEK inhibitors. Cancer Discov; 8(12); 1–16. ©2018 AACR.

See related commentary by Leber et al., p. 1511.

1Massachusetts General Hospital Cancer Center, Charlestown, Massa- Medicine, Harvard Medical School, Boston, Massachusetts. 10Department chusetts. 2Department of Oncology Research, Amgen, Thousand Oaks, of Medicinal Chemistry, Amgen, Thousand Oaks, California. 3 California. Department of Environmental Health, Harvard T. H. Chan Note: Supplementary data for this article are available at Cancer Discovery 4 School of Public Health, Boston, Massachusetts. Department of Surgery, Online (http://cancerdiscovery.aacrjournals.org/). Massachusetts General Hospital, Boston, Massachusetts. 5Department of Orthopaedics, Massachusetts General Hospital, Boston, Massachusetts. V. Nangia and F.M. Siddiqui contributed equally to this article. 6Department of Neurosurgery, Massachusetts General Hospital, Bos- Corresponding Author: Aaron N. Hata, Massachusetts General Hospital, ton, Massachusetts. 7Division of Pulmonary and Critical Care Medicine, 149 13th Street, Charlestown, MA 02129. Phone: 617-724-3442; Fax: Department of Medicine, Massachusetts General Hospital, Boston, Mas- 617-724-9648; E-mail: [email protected] 8 sachusetts. Division of Hematology Oncology, Department of Medicine, doi: 10.1158/2159-8290.CD-18-0277 Massachusetts General Hospital, Boston, Massachusetts. 9Department of ©2018 American Association for Cancer Research.

OF1 | CANCER DISCOVERY DECEMBER 2018 www.aacrjournals.org

Combination MEK + MCL1 Inhibitors for KRAS-Mutant NSCLC RESEARCH ARTICLE

INTRODUCTION (8, 9), although these have yet to be tested in the clinic. Thus, there remains an urgent need for new therapeutic strategies KRAS, a small GTPase that activates MAPK signaling, is that can target KRAS-mutant cancers. one of the most frequently mutated driver oncogenes (1). Several studies have shown that suppression of MAPK KRAS mutations, which are largely localized to residues G12, signaling, either by depletion of mutant KRAS or by phar- G13, and Q61, decrease either intrinsic and/or GAP-medi- macologic inhibition of downstream MEK1/2, is insufficient ated hydrolysis, causing constitutive activation of MAPK and to induce apoptosis in a significant number ofKRAS -mutant other downstream signaling pathways (2). Although effective cell lines (10–12). Therapeutic strategies that cotarget kinase molecular targeted therapies that inhibit oncogenic mutant signaling pathways and apoptotic regulators may increase kinases in the RAS–MAPK pathway have been developed (e.g., apoptosis and convert cytostatic responses into tumor regres- EGFR inhibitors for EGFR-mutant NSCLC, BRAF inhibi- sions (13). Activated kinase signaling pathways such as MAPK tors for BRAF-mutant melanoma), there currently are no (RAS/RAF/MEK/ERK) and PI3K/AKT converge on the BCL2 approved targeted therapies for KRAS-mutant cancers. KRAS protein family, which regulates the mitochondrial or intrinsic mutations are found in 20% to 25% of patients with non– apoptotic response (14). In cells with MAPK activation, ERK small cell lung cancer (NSCLC) and predict lack of response phosphorylation suppresses the proapoptotic BH3 protein to EGFR inhibitors (3). Attempts to target downstream BIM by targeting it for degradation (15, 16). MEK inhibition MAPK signaling with inhibitors of MEK1/2 have yielded causes BIM to accumulate (16); however, BIM can be neutral- disappointing results (4, 5), and strategies that simultane- ized by prosurvival BCL2 family members such as BCL-xL or ously target multiple signaling pathways have been limited by MCL1. Combining MEK inhibitors with the BH3 mimetic toxicity (6, 7). Most recently, a novel class of KRAS inhibitors navitoclax (ABT-263), which prevents the binding of BIM that covalently bind to the G12C mutant has been described to BCL2 and BCL-xL, led to greater apoptosis and tumor

DECEMBER 2018 CANCER DISCOVERY | OF2

RESEARCH ARTICLE Nangia et al. regression in KRAS experimental models compared with MEK AM-8621 (Supplementary Fig. S1A–S1D). AM-8621 is the inhibitors alone (11), and a clinical trial evaluating this combi- prototypical member of a novel class of spiromacrocyclic nation is currently ongoing (NCT02079740; www.clinicaltrials. MCL1 inhibitors, which also includes the tool compound gov). To date, these approaches have been limited to target- AM-4907 and the clinical compound AMG 176, which exhibit ing BCL2 and BCL-xL due to the lack of selective and potent potent and selective MCL1 inhibition in vitro and in vivo inhibitors that target other members of the BCL2 family. (19). Synergistic drug combinations were determined using MCL1 is frequently amplified in lung cancers (17), and the a Loewe excess additivity model (20). A number of potential development of potent and selective MCL1 inhibitors has long clinically relevant synergistic combinations were identified, been of interest. Recently, a novel MCL1 inhibitor, S63845, with including conventional cytotoxic chemotherapeutic agents, in vivo activity was reported (18). Significant activity was observed HSP90 inhibitors, and BH3 mimetics targeting BCL-xL/BCL2 in leukemia, myeloma, and lymphoma models, and several dif- (Fig. 1A; Supplementary Fig. S1E). ferent MCL1 inhibitors are currently in clinical development To prioritize clinically promising MCL1 inhibitor com- for these malignancies (NCT02992483, NCT02979366, and binations, we considered the spectrum of oncogenic driver NCT02675452; www.clinicaltrials.gov). Single-agent activity of mutations present in the cell lines. One of the most nota- S63845 was limited in solid tumor models including NSCLC ble genotype-associated synergies was between AM-8621 and and breast cancers; however, combining S63845 with relevant drugs that inhibit the MAPK pathway (MEK and BRAF inhibi- kinase inhibitors led to decreased cell viability of BRAF-, EGFR-, tors) in cell lines with oncogenic MAPK pathway activation and HER2-addicted cell lines in vitro, providing proof of prin- (Fig. 1B). Six of 15 cell lines showed statistically significant ciple that MCL1 inhibition, similar to BCL-xL inhibition, may synergy with a majority of the MAPK pathway inhibitors potentiate the response to kinase inhibitor–targeted therapies. (MEK and BRAF) included in the screen, and eight cell lines However, due to the lack of studies that directly compare analo- showed statistically significant synergy with a majority of gous combination strategies that target either MCL1 or BCL- MEK inhibitors. Of these eight cell lines, five harbored muta- xL, the optimal pairing of kinase inhibitors with BH3 mimetics tions known to cause MAPK pathway activation (A427, H23: that target different BCL2 family proteins in specific subsets of KRAS; H1395, G-361: BRAF; H1437: MEK1). Additionally, cancer remains undefined. two of the three other cell lines that did not harbor MAPK- Here, we assessed the activity of a novel class of potent activating mutations have previously been demonstrated to and selective spiromacrocyclic MCL1 inhibitors in combina- have functional MAPK pathway activation (HCC70: triple- tion with MEK inhibition in KRAS-mutant NSCLC models negative breast cancer with activated RAS gene signature, ref. and compared this to the parallel strategy of MEK + BCL-xL 21; G402: malignant rhabdoid tumor with hyperactivation of inhibition. Distinct but overlapping subsets of KRAS-mutant FGFR1, ref. 22). We next confirmed synergy between AM-8621 NSCLC models were more sensitive to MEK + MCL1 versus and PD032591 (MEK inhibitor) and vemurafenib (BRAF inhib- MEK + BCL-xL inhibition, which was determined by the itor), as well as the clinically relevant MEK inhibitor trametinib binding interactions between specific BCL2 family proteins. and BRAF inhibitor dabrafenib, in KRAS- and BRAF-mutant By altering the cellular localization and interactions between cells by testing the heterologous (A × B) combinations or BCL2 family proteins with transient exposure to BCL-xL corresponding self-crosses (A × A, B × B) for all possible dose inhibitors, KRAS-mutant NSCLC cells could be induced into combinations in a 10 × 10 complete combination matrix. an MCL1-dependent state with increased sensitivity to MEK + Strong synergy was again observed between AM-8621 and the MCL1 inhibition. These results provide a rationale for the MEK inhibitors PD0325901 and trametinib in A427 and H23 clinical evaluation of MEK + MCL1 inhibitors in KRAS- cells (KRAS-mutant NSCLC) and G-361 cells (BRAF-mutant mutant NSCLC and suggest a broader strategy for integrat- melanoma); synergy was not observed in the KRAS wild-type ing MCL1 and BCL-xL inhibitors to maximize efficacy of H661 (NSCLC) or DMS-273 (small-cell lung cancer) cell lines kinase inhibitor–targeted therapies. (Fig. 1C). Synergy was also observed between AM-8621 and the BRAF inhibitors dabrafenib and vemurafenib in G-361 cells RESULTS (Fig. 1C). No synergy was observed with the corresponding self- crosses. These results suggest that combining MEK + MCL1 Combination Drug Screen Identifies Synergistic inhibitors may be a promising combination strategy specifi- Activity between AM-8621 and MAPK Pathway cally for cancers with MAPK pathway activation. Inhibitors in Cell Lines with MAPK Pathway Activation KRAS-Mutant NSCLCs Are Sensitive to Recently developed potent and selective MCL1 inhibitors Combination MEK + MCL1 or BCL-xL Inhibitors have demonstrated limited single-agent activity against solid Recent studies have demonstrated that combining MEK tumor malignancies (18, 19), suggesting that drug combi- inhibitors with BH3 mimetics targeting BCL-xL/BCL2 (e.g., nations may be required to maximize the potential clinical navitoclax or ABT-263) synergistically induces apoptosis in benefit of this class of inhibitors for these cancers. In order KRAS-mutant cancer models (11, 12). Based on the results of to determine agents that synergistically enhance the activity the combination drug screen, we hypothesized that a parallel of MCL1 inhibition, 187 compounds targeting a diverse array strategy targeting both MEK and MCL1 might be effective of mechanisms (Supplementary Table S1) were screened in against KRAS-mutant cancers, particularly those that do not pairwise combinations with a novel potent and selective respond to MEK + BCL-xL inhibition. To assess the relative MCL1 inhibitor, AM-8621, against a panel of 15 diverse solid dependency of KRAS-mutant cell lines on MCL1 and BCL-xL, tumor cancer cell lines that had a range of sensitivities to we first tested whether AM-8621, navitoclax, or venetoclax

OF3 | CANCER DISCOVERY DECEMBER 2018 www.aacrjournals.org

Combination MEK + MCL1 Inhibitors for KRAS-Mutant NSCLC RESEARCH ARTICLE

A C MAPK: Activated WT 53 BH3 (BCL2) HSP90 Topoisomerase MAPK (MEK) HDAC

Proteasome A427 (KRAS) H23 (KRAS) G-361 (BRAF) NCI-H611 DMS-273 Drug DrugA B PI3K/AKT/mTOR PD0325901 DNA damaging MEKi Trametinib

187 compound library 0510 15 Combination Number of cell lines Dabrafenib BRAFi Vemurafenib AM-8621 B PD0325901 MEKi Self-cross Trametinib Dabrafenib BRAFi Vemurafenib A427 DMS-273 NCI-H1395 NCI-H1437 NCI-H1568 NCI-H23 NCI-H661 CAMA-1 HCC1187 HCC70 SK-BR-3 CHL-1 G-361 Colo-320-HSR G-402 <6 >12 PD0325901 <6 Synergy score MEKi Pimasertib Synergy Selumetinib score GDC-0879 BRAFi Vemurafenib >12 **** #*# Lung Breast Melanoma Colon Kidney

Figure 1. Synergy between AM-8621 and MAPK pathway inhibitors in cell lines with MAPK pathway activation. A, AM-8621 synergizes with multiple classes of drugs in diverse cancer cell lines. The number of cell lines in the combination drug screen in which statistically significant synergy was observed for the majority of the compounds in a given drug category is shown. The top eight drug classes (which represent 53 of 187 drugs) with the most synergistic interactions are shown. Gray hatched bars represent specific drug subclass indicated in parentheses [BCL2-specific inhibitors and MEK inhibitors (MEKi)]. Synergy scores for individual compounds and cell lines are shown in Supplementary Fig. S1E. B, Synergy scores for AM-8621 combined with MEK and BRAF inhibitors across the cell line panel. Strong synergy is indicated in red. Asterisks denote cell lines with activating mutations in the MAPK pathway. Hash marks indicate cell lines with known functional activation of MAPK signaling. C, Heterologous and self-cross drug combinations confirm synergy between AM-8621 and clinically relevant MEK/BRAF inhibitors in cell lines with oncogenic mutations that lead to MAPK pathway activation.

(ABT-199, a selective BCL2 inhibitor) alone could decrease combination was observed in cell lines that were completely cell viability. Whereas most KRAS-mutant colorectal cancer insensitive to AM-8621 alone (e.g., A549 and LU-99A). This cell lines were more sensitive to navitoclax, subsets of KRAS- heterogeneity was reflected in the degree of caspase-3/7 activa- mutant NSCLC cell lines were partially sensitive to either tion and apoptosis induced by MEK inhibitor (MEKi) + BH3 AM-8621 or navitoclax, both, or neither (Supplementary Fig. mimetic combinations (Supplementary Fig. S2F; Fig. 2C) and S2A–S2C). Very little activity of venetoclax was observed in overall decrease in cell viability (Fig. 2D). The apoptotic effect either NSCLC or colorectal cancer cell lines. of AM-8621 in combination with trametinib was recapitulated We next sought to determine the degree of heterogeneity by siRNA-mediated MCL1 knockdown (Supplementary Fig. in the response of KRAS-mutant NSCLC cells to combined S2G), validating an on-target mechanism of MCL1 inhibition MEK + MCL1 inhibition. Strong synergy was observed with by AM-8621. We also confirmed that the effects of navitoclax the combination of AM-8621 and trametinib in KRAS-mutant were due primarily to inhibition of BCL-xL, as similar results NSCLC cell lines with partial sensitivity to AM-8621 alone were obtained using the BCL-xL–selective inhibitor A1331852 (e.g., H23 and H2030), and the combination of navitoclax (23) but not the BCL2-selective inhibitor venetoclax (Supple- and trametinib was synergistic in cell lines with partial sen- mentary Fig. S2H). These data demonstrate that inhibition of sitivity to navitoclax alone (e.g., H1734 and H1944; Fig. MCL1 or BCL-xL potentiates the effects of MEK inhibition in 2A and Supplementary Fig. S2D). Across the entire KRAS- distinct but overlapping subsets of KRAS-mutant NSCLC. mutant NSCLC and colorectal cancer cell line panels, navitoclax strongly enhanced the inhibitory effect of trametinib Sensitivity to MCL1 and BCL-xL Inhibition Is in almost all colorectal cell lines (Supplementary Fig. S2E), Determined by Specific BCL2 Family Protein whereas the response of KRAS-mutant NSCLC cell lines was Interactions heterogeneous, with subsets of cell lines exhibiting enhance- Given the potential heterogeneity of response of KRAS- ment with AM-8621, navitoclax, both, or neither (Fig. 2B). mutant NSCLCs to MEK inhibitor + BH3 mimetic combi- Notably, significant activity of the AM-8621+ trametinib nations, a more complete understanding of the underlying

DECEMBER 2018 CANCER DISCOVERY | OF4

RESEARCH ARTICLE Nangia et al.

A B D H2030 100 TRAM TRAM A427H2030H23H1792H1944H1734 TRAM + AM-8621 75 TRAM + navitoclax 0 AM-8621 50 20 Trametinib + Vehicle NavitoclaxAM-8621TramitinibTram navTram AM Navitoclax 0 40 + + 25 60 +Nav AM>na 0 A427 % Cell proliferation 0 1 80 10 100% <6 >12 0.1 100 +AM-8621 1,000 100 MEKi enhancement

Synergy score vN Trametinib (nmol/L) H23 TRAM H2030 H1734 KRAS-mutant NSCLC Trametinib Trametinib A427 H2030 H2122 0 H2030 H1792 H23

AM-8621 H1792 100 LU-65 A549 H2009

Growth inhibition Growth LU-99A 200 LU-99A Navitoclax A549 H1573 COR-L23 H1734 SW900 C H441 AM>navNav>AM DV-90 H1944 av>AM 100 H1944 H2347 75 Trametinib H1155 SW1573 AM-8621 Calu-1 50 Navitoclax SW1573 H1734 Tram + AM SK-LU-1 25 Tram + nav H358 H460 0 % Apoptosis (over baseline) % Apoptosis (over 100 50 100500 H23 A427 A549 Navitoclax AM-8621 H2030 H1792LU99AH1734H1944 SW1573 MEKi enhancement

Figure 2. KRAS-mutant NSCLC cell lines exhibit a heterogeneous pattern of synergy between MEK inhibitors and BH3 mimetics. A, Cell lines were treated with trametinib (TRAM) in combination with AM-8621 or navitoclax (nav) in a 8 × 6 dose matrix, and synergy was calculated according to the Loewe excess additivity model. B, Cell lines were treated with increasing concentrations of trametinib alone or in the presence of 1 μmol/L AM-8621 or navitoclax for 72 hours, and viability was determined. The enhancement of growth inhibition by the addition of AM-8621 or navitoclax to trametinib was determined by normalizing the viability after combination treatment to trametinib alone for each dose point (0 = no effect; 100% = complete cell killing). The bottom plot compares the relative enhancement achieved by the addition of AM-8621 (red) or navitoclax (blue) to trametinib for the entire KRAS NSCLC cell line panel. C, Combining trametinib (100 nmol/L) with AM-8621 (1 μmol/L) or navitoclax (1 μmol/L) leads to increased apoptosis as determined by annexin V staining. D, Effect of combining trametinib (100 nmol/L) with AM-8621 (1 μmol/L) or navitoclax (1 μmol/L) on cell viability. Cells were treated for 72 hours and stained with crystal violet.

apoptotic dependencies that determine drug sensitivity will predict differential drug sensitivity. In general, NSCLC cell be critical for selecting the most effective combination. Pro- lines with greater MCL1 tended to also have greater BCL-xL filing the response to AM-8621 across a large cancer cell line expression. A similar pattern was observed in KRAS-mutant panel demonstrated an inverse correlation between AM-8621 NSCLC tumors (Supplementary Fig. S3C and S3D). There- sensitivity and BCL-xL expression (19), in agreement with a fore, baseline expression levels of BCL-xL or MCL1 may recent study that assessed the sensitivity of S63845 in a more not be sufficiently robust to predict sensitivity to MEK+ limited set of hematologic cancer cell lines (18). Comparison MCL1 or BCL-xL inhibition for individual cell lines or of MCL1 and BCL-xL expression levels in NSCLC and colo- tumors. rectal cancer cell lines in the Cancer Cell Line Encyclopedia To determine the molecular basis for the differential sen- revealed that as a group, colorectal cancer cell lines exhibited sitivity to MEK + MCL1 versus BCL-xL inhibition, we exam- higher BCL-xL and lower MCL1 expression than NSCLC cell ined the impact of drug treatment on apoptotic signaling. lines (Supplementary Fig. S3A), consistent with our obser- As expected, inhibition of phospho-ERK by trametinib led vation that KRAS-mutant colorectal cancer cell lines were to accumulation of the BIM protein (Supplementary Fig. more sensitive to BCL-xL inhibition (Supplementary Fig. S3E). Treatment with AM-8621 or navitoclax interrupted S2). However, when examining individual cell lines within BIM:MCL1 and BIM:BCL-xL interactions, respectively (Sup- our panel of KRAS-mutant NSCLC cell lines, we did not plementary Fig. S3F), with a commensurate increase in BIM observe an obvious correlation between combination drug binding to the uninhibited prosurvival partner, consistent sensitivity and expression of MCL1 or BCL-xL mRNA, or with the ability of prosurvival BCL2 family proteins to buffer ratios of BIM to MCL1 or BCL-xL (Supplementary Fig. S3B). BIM liberated by BH3 mimetics. To compare the effects of We also did not observe subsets of KRAS-mutant NSCLC cell BIM induction on apoptotic priming in H2030 and H1734, lines with inverse coexpression of BCL-xL and MCL1 (e.g., we performed BH3 profiling (24) before and after treatment MCL1hi/BCL-xLlo or MCL1lo/BCL-xLhi) that might obviously with trametinib. Trametinib treatment increased overall

OF5 | CANCER DISCOVERY DECEMBER 2018 www.aacrjournals.org

Combination MEK + MCL1 Inhibitors for KRAS-Mutant NSCLC RESEARCH ARTICLE

A 60 P = 0.03 B P < 0.001 IP: MCL1 BCL-xL IP: MCL1 BCL-xL IP: MCL1 BCL-xL 40 H2030 lg In IP SIPS lg In IP SIPS lg In IP SIPS H1734 MCL1 20 BCL-xL BIM priming (MOMP) 0 % Change in apoptotic BID MS1 HRK PUMA NOXA Overall MCL1 BCL-xL BAX priming Dependency BAK A427 H23 H2030 C Cytosol Mitochondria Trametinib Navitoclax AM-8621 MCL1 BCL-xL MCL1 BCL-xL BCL-xL IP: IP:

Ig In IP SSIP SSIg In IPIP A427 H23 H2030 H1944 H1734 MCL1 MCL1 BIM BIM

H2030 BCL-xL COXIV PUMA BIM Tubulin NOXA PUMA BAX BCL-xL NOXA MCL1 BAX BAK BIM BAK Relative binding H1734 COXIV H1944 H1734 MCL1 BCL-xL Tubulin

VehicleAM-8621NavitoclaxTrametinibTramTram + AM + Vehiclenav AM-8621NavitoclaxTrametinibTram + AMTram + nav Vehicle AM-8621 Navitoclax Trametinib Tram + nav Tram + nav IP: Ig control M BMBMBMBMBMB In smsb In smsb In smsb In smsb In smsb In smsb MCL1

BCL-xL H2030 BIM

MCL1

BCL-xL

H1734 BIM

BAX

Figure 3. Mitochondrial BCL2 family interactions determine sensitivity profile to BH3 mimetics. A, The change in apoptotic priming after treatment with 100 nmol/L trametinib for 16 hours was determined by BH3 profiling, which assesses mitochondria depolarization (MOMP) after treatment with BH3 peptides. Total apoptotic priming is determined by BID peptide, whereas MS1 and HRK peptides are specific for MCL1 and BCL-xL dependency, respectively. Data are the mean and standard error of four independent experiments. B, Direct binding interactions between BCL2 family proteins. MCL1 and BCL-xL were immunoprecipitated, and interacting proteins were determined by western blotting (Ig, immunoglobulin control; In, total cell lysate input; IP, immunoprecipitated fraction; S, supernatant). Right plot summarizes binding of proapoptotic BCL2 family proteins to MCL1 versus BCL-xL. C, Subcellular localization of BCL2 family proteins after 24-hour drug treatment. D, Effect of drug treatment on binding interactions between BCL2 family proteins. Cells were treated with 100 nmol/L trametinib (tram), 1 μmol/L AM-8621 (AM), and/or 1 μmol/L navitoclax (nav) as indicated for 24 hours (M, MCL1 immunoprecipitation; B, BCL-xL immunoprecipitation; In, Input; sm, MCL1 supernatant; sb, BCL-xL supernatant). apoptotic priming (BID peptide) in both cell lines, indicat- baseline, MCL1 was localized primarily to the mitochondria ing that proapoptotic signaling was induced by the treat- in most cell lines, whereas BCL-xL was variably distributed ment (Fig. 3A). Notably, in H2030 cells, this corresponded between the mitochondria and cytosol (Supplementary Fig. to increased MCL1 dependence, whereas H1734 cells became S3G). BIM was exclusively localized to the mitochondrial frac- more BCL-xL dependent, in agreement with the combination tion, indicating that the critical BIM:MCL1 and BIM:BCL-xL treatment data (Fig. 2). interactions occur at the mitochondria. To directly compare We next examined whether the differences in apoptotic the amount of BIM bound to MCL1 relative to BCL-xL, we dependency result from differential binding of BIM and performed immunoprecipitation of MCL1 and BCL-xL and other proapoptotic BCL2 proteins to MCL1 or BCL-xL. At assessed for binding of proapoptotic BCL2 family members.

DECEMBER 2018 CANCER DISCOVERY | OF6

RESEARCH ARTICLE Nangia et al.

In A427, H23, and H2030 cells, which are MCL1 dependent, combination of trametinib + navitoclax showing little activity BIM was preferentially bound to MCL1 at baseline (Fig. 3B). beyond that of trametinib alone (Supplementary Fig. S4D). In contrast, BIM was bound to BCL-xL as well as MCL1 in To evaluate the efficacy of MEK+ MCL1 inhibition in addi- BCL-xL–dependent H1944 and H1734 cells. Additionally, tional clinically relevant in vivo models, we generated primary we observed that BAK was bound to MCL1 in H23 and patient-derived xenograft (PDX) mouse tumor models from H2030 cells, whereas BAX was bound to BCL-xL in H1944 KRAS-mutant adenocarcinomas (Supplementary Table S2) and H1734 cells. Treatment of H2030 cells with trametinib that represented a variety of KRAS point mutations as well led to additional accumulation of BIM at the mitochondria as co-occurring mutations. Mice bearing PDX tumors were (Fig. 3C) and increased binding of BIM to MCL1, explaining treated with trametinib or the combination of trametinib + the increase in MCL1 priming revealed by BH3 profiling AM-4907 or trametinib + AMG 176 (the clinical analogue) for (Fig. 3D). Treatment with combined trametinib + AM-8621 2 weeks and evaluated for tumor regression. Whereas tumor prevented the accumulation of BIM on MCL1 and led to regression was observed with single-agent trametinib in only a decrease in mitochondrial BIM, suggesting that other 1 of 8 models (MGH6024, <30%), combination treatment prosurvival BCL2 family proteins such as BCL-xL are insuf- led to average tumor regression in 6 of 8 models, with four ficient to neutralize and stabilize BIM once it is liberated achieving greater than 30% average tumor volume reduction from MCL1. In H1734 cells, trametinib treatment led to (Fig. 4C; Supplementary Fig. S4E). Additionally, for 6 of 7 BIM accumulation on BCL-xL as well as MCL1; cotreatment evaluable models, the combination of trametinib + AM-4907 with ABT263 + trametinib displaced both BIM and BAX had statistically greater antitumor activity than trametinib from BCL-xL (Fig. 3D). Thus, the sensitivity of H2030 and alone. In contrast, the combination of trametinib + navito- H1734 cells to combined MEK + BH3 mimetics appears to clax showed minimal activity beyond trametinib alone in 4 of be determined by a primary dependency on MCL1 or BCL- 4 models tested (Supplementary Fig. S4E and S4F). Consist- xL, respectively, to neutralize multiple proapoptotic BCL2 ent with these results, coimmunoprecipitation experiments family members. Together, these results point to the com- confirmed that BIM was preferentially bound to MCL1 in plexity and diversity of interactions between BCL2 family MGH1070 and MGH1089-1 tumors (Supplementary Fig. proteins that determine sensitivity to MEKi + BH3 mimetic S4G). Together, these results demonstrate in vivo efficacy of combinations. combined inhibition of MEK and MCL1 and suggest that this may be a promising therapeutic approach for treatment Combination MEK + MCL1 Inhibition Leads of KRAS-mutant NSCLC in the clinic. to KRAS-Mutant NSCLC Tumor Regression To evaluate the in vivo activity of MEK + MCL1 inhibi- Prior BCL-xL Inhibition Increases MCL1 tion, we generated H2030 and A549 subcutaneous xenograft Dependence and Increases Sensitivity tumors in immunocompromised mice. Consistent with prior to MCL1 Inhibitors studies of KRAS-mutant xenograft tumors (25, 26), treatment The development of selective MCL1 inhibitors with potent with trametinib suppressed phospho-ERK signaling and led in vivo activity now provides the opportunity to rationally to modest tumor growth inhibition in both models (Fig. design therapies that can specifically target each of the major 4A and B; Supplementary Fig. S4A and S4B). To investigate prosurvival BCL2 family proteins that protect cells during the effects of MCL1 inhibition in vivo, we used AM-4907, a tyrosine kinase inhibitor treatment (14). However, our results structural analogue of AM-8621 with improved oral bioavail- suggest the selection of the most appropriate BH3 mimetic ability (Supplementary Fig. S1A; ref. 19). AM-4907 modestly may be hampered by the lack of clinically feasible biomark- slowed the growth of H2030 xenograft tumors, but had little ers (e.g., gene mutations, RNA, or protein expression levels) activity against A549 xenograft tumors, largely mirroring the that can robustly predict whether tumors are MCL1 or BCL- response of the respective cell lines to AM-8621 (Supplemen- xL dependent. To circumvent this problem, we investigated tary Fig. S2B). Combination treatment with trametinib and whether intermittent treatment with alternating MCL1 and AM-4907 led to activation of cleaved caspase-3 (Fig. 4B) and BCL-xL inhibitors might represent a practical strategy to caused regression of both A549 and H2030 xenograft tumors maximize treatment efficacy without the need to prospectively (Fig. 4A). Similar results were also observed in A427 xenograft determine apoptotic dependencies. We treated H2030 cells tumors (Supplementary Fig. S4C). The differential sensitivity with AM-8621 or navitoclax in combination with trametinib of H2030 xenograft tumors to BH3 mimetic combinations for 4 days, followed by drug washout for 3 days (for detailed was consistent with the in vitro drug sensitivity, with the treatment schemas, see Supplementary Table S3). After washout,

Figure 4. Combined MEK and MCL1 inhibition leads to tumor regression in vivo. A, The combination of trametinib (3 mg/kg daily) + AM-4907 (100 mg/kg daily) leads to tumor regression of A549 and H2030 subcutaneous xenograft tumors. A549: each treatment group N = 7 animals. H2030: each treatment group N = 7, except control N = 5. Data shown are mean and standard error, with asterisks denoting a statistical significance difference (P < 0.05) between combination treatment and trametinib alone (as determined by multiple t test with Holm–Sidak correction for multiple comparisons). B, Trametinib (TRAM) + AM-4907 (AM) treatment causes suppression of phospho-ERK and activation of cleaved caspase-3 (CC3) in A549 xenograft tumors. Scale bar, 50 mm. Right, CC3-positive cells were counted in three high power fields (HPF) for two independent xenograft tumors for each drug treatment. C, Combined MEK + MCL1 inhibition leads to tumor regression of KRAS-mutant NSCLC PDX tumors. Mice were treated with trametinib (3 mg/kg; diamonds) alone or in combination with AM-4907 (100 mg/kg; solid circles) or AMG 176 (50 mg/kg; open circles). Asterisks indicate a statistically significant P( < 0.05) decrease in combination treatment compared with trametinib alone. Right, number of individual animals treated with each drug.

OF7 | CANCER DISCOVERY DECEMBER 2018 www.aacrjournals.org

Combination MEK + MCL1 Inhibitors for KRAS-Mutant NSCLC RESEARCH ARTICLE

A A549 H2030 Control 150 150 AM-4907 100 100 Trametinib AM-4907 + trametinib 50 50

0 0 P < 0.05 P < 0.05

−50 * −50 * * * * * * * *

% Change in tumor volume * −100 −100 0714 21 28 0714 21 28 Time (days) Time (days)

Cleaved B H&E Phospho-ERK caspase-3

Control

P < 0.01 25

20 N.S. Trametinib 15 N.S. 10 cells (per HPF) + 5 CC3 AM-4907 0

Control TRAM AM-4907 TRAM + AM Trametinib + AM-4907

C TRAM TRAM 300 + + TRAM AM-4907 AMG 176

200 MGH6024 2 2 2 MGH1112 2 3 2 MGH1091 3 3 3 100 MGH1089 2 2 2 MGH1088 3 2 2 MGH1070 3 4 - 0

% Change in tumor volume MGH1065 4 6 - MGH1040 1 3 - −100 *** *** Trametinib Trametinib + MCL1i

DECEMBER 2018 CANCER DISCOVERY | OF8

RESEARCH ARTICLE Nangia et al. cells were rechallenged with either the same or opposite drug script levels in H2030, H2009, H1734, and H1944 cells that combination, for up to three rounds of treatment. Unex- could explain the sensitization (Supplementary Fig. S5E). We pectedly, we observed that after exposure to navitoclax + next examined the binding interactions and subcellular locali- trametinib, cells became exquisitely sensitive to AM-8621 + zation of BCL2 family proteins. In H2030 cells, in which the trametinib (Supplementary Fig. S5A). To determine whether majority of BIM is already bound to MCL1 cells at baseline, “pretreating” cells with navitoclax might sensitize to MCL1 navitoclax or A1331852 pretreatment caused a further modest inhibition, we selected cell lines with varied MCL1 dependen- increase in the relative amount of BIM bound to MCL1 which cies and treated them sequentially with navitoclax, drug wash- persisted even after drug washout (Supplementary Fig. S5F). out, and finally AM-8621+ trametinib. Although navitoclax An even more pronounced increase in BIM bound to MCL1 pretreatment had little impact on the sensitivity to trametinib was observed in H2009, H1944, and H1734, which exhibit less alone, cell lines became more sensitive to the AM-8621 + MCL1 dependence at baseline but become sensitive to MEK + trametinib combination, as indicated by the degree of separa- MCL1 inhibition after navitoclax or A1331852 pretreatment tion (ΔAUC) between the trametinib-only and trametinib + (Fig. 5F). In contrast, there was little or no increase in BIM AM-8621 dose–response curves (Fig. 5A–C). Conversely, pre- bound to MCL1 after navitoclax or A1331852 pretreatment in treatment with AM-8621 did not sensitize cells to combined cells that did not sensitize (H358 and H460). Conversely, there navitoclax + trametinib (Fig. 5B and C). The increased sensi- was no increase in BIM bound to BCL-xL after pretreatment tivity was observed after a range of navitoclax pretreatment with AM-8621 (Supplementary Fig. S5G), consistent with the and drug washout periods (Supplementary Fig. S5B) and lack of sensitization to BCL-xL inhibition with the reciprocal corresponded with increased apoptosis (Fig. 5D). pretreatment strategy (Fig. 5B and C). We also observed addi- We next examined whether navitoclax pretreatment causes tional changes in the subcellular localization of BCL2 family KRAS-mutant NSCLC cells to become sensitive to AM-8621 proteins, with an increase in MCL1 and BIM and/or decrease in alone. Whereas H2030 cells, which exhibit baseline sensitiv- BCL-xL protein at the mitochondria (Supplementary Fig. S5H ity to MCL1 inhibition, became more sensitive to AM-8621, and S5I), all consistent with increased dependence on MCL1. cell lines with little or no baseline MCL1 dependence (A549, To understand why the effects of navitoclax persist despite H1944, and H1734) did not become sensitive to single- drug washout, we quantified the amount of intracellular navi- agent AM-8621 (Supplementary Fig. S5C and S5D). Thus, in toclax after drug washout. In both H2030 and H1734 cells, the majority of cells, an additional apoptotic stimulus such approximately 50% of intracellular navitoclax was retained as BIM induction is still required to trigger an apoptotic 24 hours after washout, and this remained stable for up to response, suggesting that sequential application of BCL- 96 hours (Supplementary Fig. S5J). In contrast, AM-8621 was xL followed by MCL1 inhibitors might selectively sensitize depleted from cells immediately upon removal of drug from the KRAS-mutant NSCLC cells to MEK inhibition. To test this, culture media (Supplementary Fig. S5K), again consistent with we compared the response of KRAS-mutant NSCLC lines its inability to sensitize cells to BCL-xL inhibition. All together, with non–KRAS-mutant lung adenocarcinoma, squamous these results suggest that transient exposure of KRAS-mutant cell carcinoma, and noncancer cell lines. Navitoclax pre- NSCLC cells to BCL-xL inhibitors leads to sustained loading treatment increased the sensitivity of the majority of KRAS- of BIM on MCL1, increasing MCL1 dependence and thus pois- mutant NSCLC cell lines to AM-8621 + trametinib (Fig. 5E), ing cells to more readily undergo apoptosis upon subsequent with the exception of KRAS-mutant NSCLC cells that were treatment with combination MEK + MCL1 inhibitors. Indeed, completely insensitive to both MEK + MCL1 and MEK + navitoclax pretreated H2030 and H1734 cells underwent rapid BCL-xL inhibition at baseline (H460 and H358). Of note, mitochondrial depolarization and cytochrome C release after cell lines that were sensitive to MEK + BCL-xL inhibition but treatment with trametinib + AM-8621 compared with vehicle- not MEK + MCL1 inhibition at baseline (H1734 and H441) pretreated cells (Fig. 5G; Supplementary Fig. S5L). became sensitive to MEK + MCL1 inhibition after navitoclax pretreatment. We did not observe sensitization of non–KRAS- Pretreatment with BCL-xL Inhibitors mutant cells including KRAS wild-type lung cancer cells and May Permit Alternative MEK + MCL1 Inhibitor cancer-associated fibroblast cells, consistent with the notion Dosing Strategies that this approach selectively targets KRAS-mutant cells that The use of MEK inhibitors in the clinic is associated with have MAPK-mediated suppression of BIM. Finally, to verify significant toxicity in many patients (27), and this may limit that the effect of navitoclax pretreatment was mediated by the promise of MEK-based combination therapies. Alterna- BCL-xL, we tested the BCL-xL–selective inhibitor A1331852 tive drug administration schedules such as intermittent dos- and observed even more pronounced sensitization in KRAS- ing have the potential to minimize toxicity, but may come mutant NSCLC cells, but not in non–KRAS-mutant cell lines with the cost of decreased efficacy. We hypothesized that (Fig. 5E). Overall, these results suggest that sequential appli- increasing the MCL1 inhibitor sensitivity of KRAS-mutant cation of BH3 mimetics can be used to selectively sensitize NSCLC cells by pretreatment with BCL-xL inhibitors might KRAS-mutant NSCLC cells to MEK inhibition. offset decreased activity of intermittent MEK + MCL1 inhi- To determine the mechanism of increased MEK + MCL1 bition. To test this, we monitored cell viability of KRAS- inhibitor sensitivity of KRAS-mutant cells after transient BCL- mutant NSCLC cell lines during treatment with continuous xL inhibitor treatment, we first examined whether pretreatment or intermittent trametinib + AM-8621 (3 days on/4 days with navitoclax induced changes in the expression of BCL2 off), after either vehicle or navitoclax pretreatment. We also family genes. After both navitoclax treatment and washout, assessed patient-derived cell lines that corresponded to PDX we did not observe any consistent changes in the RNA tran- models used for the in vivo studies (independently generated

OF9 | CANCER DISCOVERY DECEMBER 2018 www.aacrjournals.org

Combination MEK + MCL1 Inhibitors for KRAS-Mutant NSCLC RESEARCH ARTICLE

Navitoclax Trametinib ABVeh pre Tram Sensitization MCL1 inhibition Sensitization BCL-xL inhibition E A1331852 AM-8621 Veh pre Tram+AM ∆AUC Nav pre Tram Nav pre Tram+AM Veh pre Veh pre

H2030 H2030 100 Nav pre AM pre 75 H1792 50 25 H2009 0 600 600 % Cell proliferation Veh pre Veh pre 00.1 110 100 1,000 Nav pre AM pre Trametinib (nmol/L) 400 400 hibit baseline C A549 C H2009 ∆ AU

100 ∆ AU 200 200 75 H441

0 0 MEKi + BH3 sensitivity 50 Cell lines that ex A549 A549 25 H2030 H2009 H1944 H1734 H2030 H2009 H1944 H1734 H1734 0 KRAS -mutant NSCLC % Cell proliferation 00.1 110 100 1,000 Trametinib (nmol/L) H1944 C P < 0.01 D H2030 H1734 H1734 200 P < 0.001 P < 0.001 3 100 80 50 ct P < 0.01 40 VEH H460 75 100 60 Tram 30 AM-8621 50 40 20 Tram + AM 0 sensitivity H358 25 20 % Apoptosis 10 no MEKi + BH 0 % Cell proliferation −100 0 0 0 0.1 110 100 1,000 MGH1069 (increase in ∆ AUC) Trametinib (nmol/L) Pretreatment ef fe (squamous) No pre No pre −200 Nav pre Nav pre MGH1072 (adeno) F IP:BMCL1 CL-xL Input IP:BMCL1 CL-xL Input MGH1147 (fibroblast)

Veh Nav A133 Veh Nav A133 Veh Nav A133 Veh Nav A133 Veh Nav A133 Veh Nav A133 Non– KRAS -mutant MGH1140 MCL1 MCL1 (fibroblast) BCL-xL BCL-xL H358 H2009 BIM BIM G MCL1 MCL1 1.00 BCL-xL BCL-xL

H460 0.75 H1944 BIM BIM H2030H1734 0.50 MCL1 Vehicle 0.25 BCL-xL Navitoclax H1734 Cytosolic cytochrome C 0.00 BIM A1331852 0204060 360 Time (minutes)

Figure 5. Transient BCL-xL inhibition alters BCL2 family interactions and increases sensitivity to MCL1 inhibition. A, Navitoclax (Nav) pretreatment increases sensitivity to combination MEK + MCL1 inhibition. Cells were pretreated with navitoclax for 48 hours followed by drug washout for 48 hours and then treated with trametinib (Tram) ± 1 μmol/L AM-8621 (AM) for 72 hours (blue boxes, navitoclax; red boxes, AMG-8621; open boxes, vehicle; line, drug washout period). For detailed treatment schemas, see Supplementary Table S3. B, Navitoclax pretreatment increases sensitivity to combined MEK + MCL1 inhibition; however, the reciprocal strategy of AM-8621 pretreatment does not increase sensitivity to MEK + BCL-xL inhibition. ΔAUC cor- responds to the shaded area between trametinib-only (black) and trametinib + BH3 mimetic (colored) curves and is a measure of the sensitization effect. C, Comparison of the increase in sensitization observed with the reciprocal pretreatment strategies. Data are replotted from B, and each dot represents the increase in MCL1 sensitization after pretreatment with navitoclax (red) or BCL-xL sensitization after pretreatment with AM-8621 (blue) for each cell line. D, Cells were pretreated with 1 μmol/L navitoclax for 48 hours followed by drug washout for 48 hours and then treated with 100 nmol/L trametinib, 1 μmol/L navitoclax, 1 μmol/L AM-8621, or combination for 72 hours. Apoptosis was determined by propidium iodide (PI)/annexin staining. E, Navitoclax or A1331852 pretreatment sensitizes KRAS-mutant NSCLC cells that exhibit baseline sensitivity to either MEK + MCL1 or MEK + BCL-xL inhibition, but not KRAS-mutant NSCLC cells without baseline sensitivity or non–KRAS-mutant cells. F, Navitoclax or A1331852 pretreatment leads to increased BIM bound to MCL1 in cells that exhibit sensitization (H2009, H1944, and H1734), but not in cells that do not sensitize (H358 and H460). G, Navitoclax pretreated cells undergo rapid mitochondrial depolarization and release of cytochrome C upon treatment with trametinib + AM-8621. from the clinical tumor samples, except for MGH1070x, AMG 176 combinations compared with vehicle-pretreated which was derived from the established mouse PDX tumor). tumors (Fig. 6B). Thus, transient treatment of KRAS-mutant Intermittent trametinib + AM-8621 was slightly less effec- NSCLCs with BCL-xL inhibitors increases sensitivity to com- tive than continuous treatment (P = 0.055) for reducing bined MEK + MCL1 inhibition in vitro and in vivo and may cell viability (Fig. 6A). However, when cells were pretreated enable intermittent dosing strategies with preservation of with navitoclax, intermittent trametinib + AM-8621 led to antitumor efficacy. dramatically reduced cell viability in all cell lines. Finally, we tested whether BCL-xL inhibitor pretreatment sensitizes KRAS-mutant xenograft tumors to MEK + MCL1 inhibition DISCUSSION in vivo. H2030 xenografts that were pretreated with navito- The development of MCL1 inhibitors with potent and clax or A1331852 experienced deeper and more rapid tumor selective in vivo activity represents a major advance in our abil- regressions upon treatment with trametinib + AM-4907 or ity to individually target the three major prosurvival BCL2

DECEMBER 2018 CANCER DISCOVERY | OF10

RESEARCH ARTICLE Nangia et al.

A A C P < 0.01 B Pretreatment 4 P < 0.01 – + P = 0.055 4 4 2 0 2 2 0 0 0 –25 –2 –2 –2 volume –4 –4 FC in cell viability –4

2 –50 FC in cell viability –6 FC in cell viability –6 2 2 –6 Log –8 –8 % Change in tumor Log Log –8 4710 13 4710 13 –75 DCBA P 0.01 Time (days) Time (days) < BD Week 1 Week 2 LU-99A Trametinib Navitoclax 4 4 H2030 BCL-xLi MCL1i + MEKi 2 AM-8621 2 H1792 Pre Combo A 0 0 H1734 Veh AM-4907 –2 –2 MGH1088 B Veh AMG 176 –4 –4 MGH1070x Nav AM-4907 FC in cell viability FC in cell viability –6 C –6 2 MGH1089-2

2 Nav AMG 176 –8 –8 MGH1089-1 D A133 AMG 176 Log Log 4710 13 4710 13 Time (days) Time (days)

Figure 6. Pretreatment with BCL-xL inhibitors increases efficacy of combined MEK+ MCL1 inhibition. A, KRAS-mutant NSCLC cell lines were treated with vehicle (“A”), continuous trametinib + AM-8621 (“B”), intermittent (3 days on/4 days off) trametinib + AM-8621 (“C”), or navitoclax followed by intermittent trametinib + AM-8621 (“D”), and cell viability was determined at indicated time points in replicate plates. Data are normalized to the cell viability at the time of initial trametinib + AM-8621 treatment. Right, aggregate data for the final time point. FC, fold change. B, Mice bearing H2030 subcutaneous xenograft tumors were pretreated with navitoclax (100 mg/kg daily) or A1331852 (25 mg/kg twice daily) for 4 days. After 3 days of drug holiday, mice were treated with trametinib (3 mg/kg) plus AM-4907 (100 mg/kg daily) or AMG 176 (50 mg/kg daily) for 5 days. Change in tumor volume at the end of 2 weeks relative to tumor size at the start of MEK + MCL1 inhibitor treatment is shown. family proteins that regulate the apoptotic response: MCL1, reinforce the notion that therapies that target oncogenic BCL2, and BCL-xL. Although BH3 mimetics that target these dependencies and induce proapoptotic BCL2 proteins (e.g., three proteins have demonstrated activity against various induction of BIM upon MEK inhibition) drive dependency on hematologic malignancies (18, 19, 28), emerging data suggest BCL-xL or MCL1 that can be therapeutically exploited with that solid tumor malignancies such as NSCLC and breast BH3 mimetics. Combining trametinib with the orally bioavail- cancer will be largely refractory to these drugs when used able analogue AM-4907 or the clinical compound AMG 176 alone. For these cancers, the promise of BH3 mimetics lies induced tumor regressions in KRAS-mutant NSCLC models in their ability to potentiate the apoptotic response of other in vivo, providing a strong rationale for the clinical investiga- therapeutic modalities such as targeted therapies or cytotoxic tion of this combination for KRAS-mutant NSCLC. chemotherapy. In order to move beyond proof-of-concept studies and A number of studies have demonstrated that the BCL2/ define specific strategies for effectively deploying BH3 mimet- BCL-xL inhibitors ABT-737 and ABT-263 (navitoclax) can ics in the clinic, a more precise understanding of the specific synergize with targeted therapies such as EGFR inhibitors in functional apoptotic dependencies of different cancer subsets NSCLC (29–31) and MEK inhibitors in BRAF melanoma (32) is needed. Now that potent and selective inhibitors that tar- and KRAS-mutant cancers (11). Similarly, two recent studies get MCL1, BCL2, and BCL-xL are available, it is possible to demonstrated that the novel MCL1 inhibitor S63845 syn- directly compare parallel approaches targeting each of these ergizes with targeted therapies and chemotherapy in breast proteins. In this study, comparing the response of KRAS- cancer, melanoma, and lung cancer models (18, 33). Our mutant NSCLC and colorectal cancer cell lines with dual combination drug screen similarly identified strong syner- MEK + MCL1 or BCL-xL blockade yielded several important gistic associations between the novel spiromacrocyclic MCL1 insights. First, we found that colorectal cancer cell lines are inhibitor AM-8621 and a number of clinically relevant com- more uniformly BCL-xL–dependent than NSCLC cell lines. pounds that may warrant further investigation. In this study, This finding is consistent with the observation that colorec- we focused on synergy between AM-8621 and MEK inhibitors tal cancer cell lines as a group have higher BCL-xL and lower in KRAS-mutant lung cancers because of the clear association MCL1 expression levels, which would be expected to result with an oncogenic driver mutation and the lack of effective in increased BCL-xL dependency. Whether this results from targeted therapy options for these patients. In a subset of differences in co-occurring mutations or the influence of cell KRAS-mutant NSCLC cell lines, combining the MEK inhibi- lineage remains to be determined; however, it suggests that tor trametinib with AM-8621 led to robust activation of apop- tissue of origin may play an influential role in determining tosis and potent suppression of cell viability. These results apoptotic dependencies independent of an oncogenic driver.

OF11 | CANCER DISCOVERY DECEMBER 2018 www.aacrjournals.org

Combination MEK + MCL1 Inhibitors for KRAS-Mutant NSCLC RESEARCH ARTICLE

Second, our results suggest that the specific apoptotic in BCL2 family protein interactions persist even after drug dependencies of NSCLCs may be heterogeneous, with dif- washout is likely facilitated by the long intracellular half-life ferent KRAS-mutant NSCLC cell lines being dependent on of navitoclax. Although our studies did not directly deter- MCL1, BCL-xL, both, or neither. Developing reliable and mine whether navitoclax remains engaged in the BCL-xL robust biomarkers that can predict these apoptotic depend- binding pocket after washout, we did not observe restoration encies and guide selection of BH3 mimetics will be crucial of BIM:BCL-xL interactions in H1734 cells after washout to guiding patient selection in the clinic. Many studies have (Supplementary Fig. S5F). Interestingly, BCL-xL retrotrans- demonstrated an inverse correlation between the sensitivity location from mitochondria to cytosol has been previously of BH3 mimetics and the expression level of the nontargeted reported to protect cells from apoptosis by helping to recycle prosurvival BCL2 family protein. For instance, high expres- mitochondrial BAX back to its inactive cytoplasmic pool, and sion levels of MCL1 correlate with decreased sensitivity to disruption of this process by ABT-737 increases both BCL-xL ABT-737/navitoclax (34–36); conversely, sensitivity to S63845 retrotranslocation and mitochondrial BAX localization (39). and AM-8621 inversely correlates with BCL-xL expression In our models, we did not observe any changes in the cytosolic– levels (18, 19, 33). However, it is not clear if these correla- mitochondrial distribution of BAX after navitoclax pretreat- tions are sufficiently robust to enable MCL1 and BCL-xL ment, suggesting that this mechanism does not account for expression levels to be used to predict drug sensitivity of the observed apoptotic sensitization. individual tumors within cancer subclasses. For our KRAS- Although this study focused specifically onKRAS- mutant mutant NSCLC cell line panel, we found that expression lev- NSCLC, the strategy of alternating BH3 mimetics that selec- els of BCL2 family members were insufficient for predicting tively target MCL1 and BCL-xL may be relevant to other response to combination MEK + MCL1 versus MEK + BCL- oncogene-addicted cancers in which BH3-only proteins, such xL inhibition. Coimmunoprecipitation experiments revealed as BIM, can be induced by targeted therapies (40). It is note- notable differences in the binding interactions between proa- worthy that sensitization after BCL-xL inhibitor pretreatment poptotic BCL2 proteins and MCL1 or BCL-xL in cell lines occurred only in cell lines that exhibited at least partial sen- that had similar levels of MCL1 and BCL-xL expression, sitivity to either MEK + MCL1 or MEK + BCL-xL inhibition. and this corresponded to sensitivity to MCL1 or BCL-xL KRAS-mutant NSCLC cells that were completely insensitive to inhibition. This indicates that gene or protein expression both trametinib + BH3 mimetic combinations did not become levels alone incompletely describe the functional apoptotic more sensitive MEK + MCL1 inhibitors after exposure to dependencies that determine response to BH3 mimetics. BH3 BCL-xL inhibitors, nor did KRAS wild-type lung cancer cells profiling has recently emerged as a promising method for or stromal fibroblasts, which are not driven by oncogenic determining the apoptotic priming state of cancer cells (24, MAPK pathway activation. This suggests that matching the 37, 38), and in this study it differentiated between cells that targeted therapy to oncogenic driver remains a critical deter- were sensitive to trametinib + AM-8621 versus trametinib + minant of selectivity, with BH3 mimetics acting as a poten- navitoclax. It remains to be seen whether BH3 profiling can tiator of the apoptotic response. Thus, these results provide be effectively applied to solid-tumor malignancies for which a rationale for further investigation of intercalated and inter- routine diagnostic core biopsy tissue samples are scant. mittent dosing strategies that may maximize the efficacy Given these challenges, therapeutic strategies that incor- of BH3 mimetic–targeted therapy combinations in diverse porate both BCL-xL and MCL1 inhibitors might obviate cancer types. the need to define specific apoptotic dependencies. In our combination drug screen, the combination of AM-8621 with METHODS ABT-737 or navitoclax was universally potent, with marked synergy observed against every cell line, suggesting that Cell Lines, Antibodies, and Reagents simultaneous administration might be associated with unac- KRAS-mutant NSCLC and colorectal cancer cell lines were obtained ceptable toxicity. This motivated us to explore intercalating from the Center for Molecular Therapeutics at the Massachusetts intermittent AM-8621 or navitoclax in combination with General Hospital (MGH) Cancer Center, which performs routine SNP and short tandem repeat (STR) authentication. Additionally, cell lines trametinib, a strategy that might also minimize the throm- underwent STR validation at the initiation of the project (Biosynthe- bocytopenia associated with navitoclax. Unexpectedly, we sis, Inc.). Cell lines were routinely tested for Mycoplasma during experi- observed that transient exposure to BCL-xL inhibitors caused mental use. Cell lines were maintained in RPMI supplemented with cells to become more sensitive to trametinib + AM-8621, 5% FBS except A427, SW1573, H2009, H1573, GP5d, LS174T, LS1034, including cells that exhibited no sensitivity to MCL1 inhi- LoVo, SW1116, SW837, and T84 cells, which were maintained in bition at baseline. On a molecular level, pretreatment with DMEM/F12 supplemented with 5% FBS. Patient-derived NSCLC and navitoclax or A1331852 caused displacement of BIM from cancer-associated fibroblast cell lines were established in our labora- BCL-xL, retrotranslocation of BCL-xL from the mitochondria tory from surgical resections, core-needle biopsies, or pleural effu- to the cytosol and, in some cases, an increase in mitochon- sion samples as previously described (41), with the exception of the drial MCL1. The net result of these changes was an increase MGH1070x cell line, which was derived from a primary mouse PDX model. All patients signed informed consent to participate in a Dana- in the amount of BIM bound to MCL1 at the mitochondria, Farber/Harvard Cancer Center Institutional Review Board–approved poising the cell for rapid mitochondrial depolarization and protocol, giving permission for research to be performed on their release of cytochrome C upon treatment with trametinib + samples. Clinically observed KRAS mutations (determined by MGH AM-8621. In contrast, cell lines that did not sensitize after SNaPshot NGS genotyping panel) were verified in established cell navitoclax or A1331852 pretreatment exhibited very little lines. Established patient-derived cell lines were maintained in RPMI + increase in BIM bound to MCL1. The fact that the changes 10% FBS. For cell culture and in vivo studies, trametinib, navitoclax

DECEMBER 2018 CANCER DISCOVERY | OF12

RESEARCH ARTICLE Nangia et al.

(ABT-263), venetoclax (ABT-199), and A1331852 were purchased from The fractional inhibition for each component agent and combina- Selleckchem. AM-8621, AM-4907, and AMG 176 were provided by tion point in the matrix is calculated relative to the median of all Amgen. For cell culture studies, drugs were dissolved in DMSO to a vehicle-treated control wells. The synergy score equation integrates final concentration of 10 mmol/L and stored at −20°C; unless other- the experimentally observed activity volume at each point in the wise specified, 1μ mol/L concentration was used for in vitro cell culture matrix in excess of a model surface numerically derived from the experiments. For western blotting, the following antibodies were used: activity of the component agents using the Loewe model for additiv- MCL1, BCL-xL, BIM, BAK, PUMA, Cytochrome C, COX-IV, HDAC1, ity. Additional terms in the synergy score equation (above) are used b-tubulin, pERK1/2 T202/204, ERK1/2, and b-actin (Cell Signaling to normalize for various dilution factors used for individual agents Technology); NOXA and BAX (Santa Cruz Biotechnology). and to allow for comparison of synergy scores across an entire experiment. The inclusion of positive inhibition gating or an Idata multi- Combination Drug Screen plier removes noise near the zero-effect level and biases results for Viability Assay for Combination Screen. The AM-8621 combina- synergistic interactions that occur at high activity levels. tion screen was performed at Zalicus. In brief, cells were seeded in growth media in black 1,536-well tissue culture–treated plates at Self-Cross–Based Combination Screen Analysis. To objectively estab- optimized seeding densities to ensure log phase growth throughout lish hit criteria for the AM-8621 combination screen, 40 compounds the compound treatment duration. Cells were incubated at 37°C and from the combination library were selected to be self-crossed across the 15 cell line panel as a means to empirically determine a baseline 5% CO2 for 24 hours before treatment. At the time of treatment, a set of assay plates (which do not receive treatment) was collected, and additive, nonsynergistic response. The identity of the 40 self-cross ATP levels were measured via ATPLite viability assay (PerkinElmer). compounds was determined by selecting compounds with a variety These Tzero (T0) plates were read using ultrasensitive luminescence of maximum response values and single-agent dose–response steep- on Envision Plate Readers. Treated assay plates were incubated with ness. Those drug combinations, which yielded effect levels that sta- compound for 72 hours (see “Combination Screen Design” below). tistically superseded those baseline additivity values, were considered After 72 hours, plates were developed for endpoint analysis using synergistic. ATPLite. All data points were collected via automated processes, The synergy score measure was used for the self-cross analysis. quality controlled, and analyzed using Chalice software (Zalicus). Synergy scores of self-crosses are expected to be additive by defi- Assay plates were accepted if they passed the following quality con- nition and therefore maintain a synergy score of zero. However, trol standards: relative luciferase values were consistent throughout although some self-cross synergy scores are near zero, many are the entire experiment, Z-factor scores were greater than 0.6, and greater, suggesting that experimental noise or nonoptimal curve fit- untreated/vehicle controls behaved consistently on the plate. ting of the single-agent dose responses is contributing to the slight Growth inhibition (GI) values were calculated as a measure of cell perturbations in the score. Given the potential differences in cell viability. The cell viability of vehicle is measured at the time of dosing line sensitivity to AM-8621 combination activities, we used a cell (T0) and after 72 hours (T72). A GI reading of 0% represents no growth line–centric strategy for the self-cross–based combination screen inhibition—cells treated with compound and T72 vehicle signals are analysis, focusing on self-cross behavior in individual cell lines ver- equivalent. A GI of 100% represents complete growth inhibition—cells sus global review of the cell line panel activity. Combinations where treated with compound and T0 vehicle signals are equivalent. A GI of the synergy score was greater than the mean self-cross plus two 200% represents complete death of all cells in the culture well. GI was standard deviations (2σ) were considered candidate synergies at the calculated by applying the following test and equation: 95% confidence levels. A total of 960 hits were identified atσ 2 above the mean (32% hit rate).

 TV− 0  Confirmatory High-Resolution Synergy Experiments. For confirma- If TV<−0 :100* 1   V0  tory synergy studies reported in Fig. 1C, similar methodologies to those described for the AM-8621 combination screen were used  TV− 0  with the following exceptions: (i) studies were performed in 384- If TV≥−0 :100* 1   VV− 0  well plates and (ii) combinations and self-cross experiments were performed in 10 × 10 complete optimized matrices. All experiments where T is the signal measure for a test article or combination, V were performed in duplicate. is the vehicle-treated control measure, and V0 is the vehicle control measure at time zero. This formula is derived from the GI calcula- tion used in the National Cancer Institute’s NCI-60 high-through- Cell Proliferation and Viability Assays put screen. For cell proliferation dose–response assays, cell lines were seeded into 96-well plates 24 hours before the addition of drug. Cell prolif- Combination Screen Design. Combination analysis data were eration was determined by CellTiter-Glo assay (Promega) 72 hours collected in a 7 × 7 checkerboard optimized matrix for 15 cell after adding drug according to the manufacturer’s protocol. For lines screened in the 1,536-well format. One hundred eighty-seven short-term time-course experiments, plates were drugged in identi- enhancer compounds were combined with AM-8621 across the 15 cal fashion at indicated time points; all plates in an experiment were cell line panel. In addition, 40 compounds were combined in self- developed with CellTiter-Glo simultaneously. Cell viability studies cross analysis for each cell line. The starting concentrations and fold were performed using crystal violet staining. Cells were fixed with dilution for AM-8621 and enhancer compounds were optimized to glutaraldehyde, washed with water, and stained with 0.1% crystal best fit the individual single-agent dose–response curves. violet for 30 minutes. After imaging, staining was quantified by using 10% acetic acid to extract the stain and absorbance read at 590 Synergy Score Analysis. To measure combination effects in excess nm. Long-term cell viability studies were performed using RealTime- of Loewe additivity, a scalar measure was used to characterize the Glo assay (Promega) according to the manufacturer’s protocol. Cell strength of synergistic interaction termed the “synergy score.” The viability was assessed at the indicated time points, and fresh media synergy score is calculated as: were replaced immediately afterward. For navitoclax pretreatment studies, we observed similar sensitization over a range of pretreat-

Synergy score = log fX log fY ∑max(0, Idata) (Idata - ILoewe) ment and washout time periods (48–72 hours); for consistency,

OF13 | CANCER DISCOVERY DECEMBER 2018 www.aacrjournals.org

Combination MEK + MCL1 Inhibitors for KRAS-Mutant NSCLC RESEARCH ARTICLE

48-hour pretreatment/48-hour washout was used unless otherwise was added to each plate followed by incubation at room temperature indicated. with shaking for 20 minutes. Samples were then spun down to col- lect supernatant, which was dried down and reconstituted in equal PI/Annexin Apoptosis Assay parts acetonitrile and water followed by analysis on an API 5500 mass spectrophotometer. Cells were seeded in triplicate at low density 24 hours prior to drug addition. Seventy-two hours after adding drugs, floating and adher- Mouse Xenograft Assays and Drug Treatment ent cells were collected and stained with propidium iodide (PI) and Cy5-Annexin V (BD Biosciences) and analyzed by flow cytometry. The All mouse studies were conducted through Institutional Animal annexin-positive apoptotic cell fraction was analyzed using FlowJo Care and Use Committee–approved animal protocols in accordance software. with institutional guidelines. KRAS-mutant NSCLC PDX models were generated from surgical resections, core-needle biopsies, or pleu- Caspase 3/7 Apoptosis Assay ral effusion samples by subcutaneous implantation into NSG mice (Jackson Labs). Subcutaneous tumors were serially passaged twice to Cells were seeded in triplicate at low density 24 hours prior to fully establish each model. Clinically observed KRAS mutations were drug addition. Forty-eight hours after adding drug, caspase cleavage verified in each established model. For drug studies, PDX tumors was quantified using ApoTox-Glo Triplex Assay (Promega). Rela- were directly implanted subcutaneously into NSG or athymic nude tive caspase cleavage was determined as follows: (luminescence)/ (NE/Nu) mice and allowed to grow to 250 to 400 mm3. For H2030 [400Ex/505Em (viability)]. and A549 xenograft studies, cell line suspensions were prepared in 1:1 matrigel:PBS, and 5 × 106 cells were injected unilaterally into the Subcellular Fractionation Assay subcutaneous space on the flanks of athymic nude (Nu/Nu) mice Cells were incubated with drug for designated time periods. Fol- and allowed to grow to approximately 350 mm3. Tumors were meas- lowing incubations, whole-cell, cytosolic, mitochondrial, and nuclear ured with electronic calipers, and the tumor volume was calculated fractions were sequentially isolated using a cell fractionation kit according to the formula V = 0.52 × L × W2. Mice with established (Abcam; catalog number 109719) according to the manufacturer’s tumors were randomized to drug treatment groups using covariate- protocol, with minor modifications. Purity of fractions was deter- adaptive randomization to minimize differences in baseline tumor mined by western blot analysis using specific nuclear, cytosolic, and volumes. Tumor volumes at the start of treatment (mean ± SD): A549 mitochondrial markers. control 372 ± 28 mm3, A549 trametinib 382 ± mm3, A549 AM-4907 374 ± 22 mm3, A549 trametinib + AM-4907 382 ± 11 mm3; H2030 Immunoprecipitation Assay control 315 ± 42 mm3, H2030 trametinib 318 ± 32 mm3, H2030 Immunoblotting was performed per antibody manufacturer’s speci- AM-4907 318 ± 60 mm3, H2030 trametinib + AM-4907 ± mm3; fications. Cells were lysed using Tris Lysis Buffer and Protease inhibitor H2030 navitoclax 227 ± 55 mm3, H2030 trametinib + navitoclax 283 ± cocktail (Meso Scale Diagnostics) and immunoprecipitated overnight at 66 mm3; A427 control 201 ± 20 mm3, A427 trametinib 318 ± 16 4°C with Protein A/G Agarose beads (Thermo Fisher) and the following mm3, A427 AM-4907 312 ± 17 mm3, A427 trametinib + AM-4907 antibodies: anti-MCL1 (BD Pharmingen), anti-BCL-xL (EMD Milli- 305 ± 19 mm3. Trametinib was dissolved in 0.5% HPMC/0.2% Tween pore), and anti-mouse IgG1 isotype control (Cell Signaling Technology). 80 (pH 8.0) and administered by oral gavage daily at 3 mg/kg, 6 days The immunoprecipitate, the supernatant, and a sample of the initial per week. AM-4907 and AMG 176 were dissolved in 25% hydroxypropyl- whole-cell lysate for each condition were then analyzed by western blot. beta-cyclodextrin (pH8.0) and administered by oral gavage daily at 100 mg/kg and 50 mg/kg, respectively. Navitoclax was dissolved in siRNA Knockdown Studies 60% Phosal 50/30% polyethylene glycol (PEG) 400/10% ethanol and administered by oral gavage daily at 100 mg/kg. A13318752 was dis- Cells were plated in Opti-MEM media (Invitrogen), transfected with solved in 60% Phosal 50/27.5% PEG 400/10% ethanol/2.5% DMSO 50 nmol/L siRNAs specific for MCL1 (Thermo Fisher, 4392420; Qia- and administered by oral gavage at 25 mg/kg twice daily. gen, S10218120S) using HiPerfect. Twenty-four hours after transfection, cells were seeded for experiments. Cell lysates for western blotting IHC of MCL1 were collected 48 hours after transfection. For caspase 3/7 Histology was performed by HistoWiz Inc. (histowiz.com) using activity assays, drugs were added starting 48 hours after transfection. standard operating procedures and fully automated workflow. Forma- lin-fixed paraffin-embeddedKRAS -mutant NSCLC xenograft tumors BH3 Profiling Assay were sectioned at 4 μm. IHC was performed on a Bond Rx autostainer BH3 profiling was performed as previously described (38). Briefly, (Leica Biosystems) with enzyme treatment (1:1,000) using standard cells were treated with trametinib at 100 nmol/L for 16 hours, protocols. Antibodies used were phospho-ERK (1:100; #4307, Cell and JC1-based BH3 profiling was performed on untreated versus Signaling Technology) and cleaved caspase-3 (1:300; #9661, Cell Sig trametinib-treated cells. Change in apoptotic priming was defined as naling Technology). Bond Polymer Refine Detection (Leica Biosystems) the difference in the extent of mitochondrial depolarization induced was used according to the manufacturer’s protocol. Sections were by BID, MS1, and HRK peptides in treated versus untreated cells. then counterstained with hematoxylin, dehydrated, and film cover- slipped using a Tissue-Tek Prisma and Coverslipper (Sakura). Navitoclax Washout Studies H2030 and H1734 cells were seeded in 10-cm dishes at a density of Confocal Fluorescence Microscopy 5 × 105 cells/dish in 10 mL growth media containing 5% FBS. Plates Cells were seeded at low density on a Lab-Tek Chamber Slide were incubated overnight at 37°C and 5% CO2 to allow cells to adhere System 48 hours prior to drug treatment. Cells were treated with to the bottom of the plates. Cells were then treated with 1 μmol/L 1 μmol/L A1331852 or vehicle for 48 hours, washed with PBS, and then navitoclax or AM-8621 for 72 hours, followed by washout and replace- cultured for 48 hours in media without drug. Cells were incubated ment with 10 mL of compound-free growth media. At indicated time with MitoTracker Red CMXRos (Thermo Fisher) for 30 minutes, points post washout, media were aspirated from plates followed by fixed with 1% paraformaldehyde, and permeabilized with 0.2% Titron two washes with 5 mL PBS. Plates were then stored at −80°C. To pre- X-100. Cells were then blocked with 5% NGS/5% BSA for 1 hour and pare for LC/MS analysis, plates were thawed, and 5 mL of acetonitrile incubated with 1:500 anti-BCL-xL rabbit mAb (54H6, Cell Signaling

DECEMBER 2018 CANCER DISCOVERY | OF14

RESEARCH ARTICLE Nangia et al.

Technology) overnight at 4°C. Cells were washed with 0.05% Tween Piece of the Solution. This project was also supported by a Stand 20 and then incubated with 1:500 anti-rabbit IgG Alexa Fluor 488 Up To Cancer–American Cancer Society Lung Cancer Dream Team conjugate (Cell Signaling Technology) for 1 hour at room tempera- Translational Research Grant (SU2C-AACR-DT17-15). Stand Up To ture. Finally, cells were stained with 1:1,000 DAPI for 10 minutes. Cancer (SU2C) is a division of the Entertainment Industry Founda- Slides were imaged the next day on a Zeiss LSM 710 inverted laser tion. Research grants are administered by the American Association scanning confocal microscope (excitation/emission: 495/519 nm for for Cancer Research, the Scientific Partner of SU2C. BCL-xL and 578/608 nm for MitoTracker) at 63× oil magnification. The costs of publication of this article were defrayed in part by Data and Statistical Analysis the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 Data were analyzed using GraphPad Prism software (GraphPad solely to indicate this fact. Software). Unless otherwise specified, data displayed are mean and standard error. Pair-wise comparisons between groups (e.g., experi- Received April 9, 2018; revised August 30, 2018; accepted mental vs. control) were made using paired or unpaired t tests as September 24, 2018; published first September 25, 2018. appropriate. For xenograft tumor measurements, individual time points were compared using multiple t tests with Holm–Sidak correction for multiple comparisons. Unless otherwise indicated, P values below 0.05 were considered to be statistically significant. REFERENCES . 1 Prior IA, Lewis PD, Mattos C. A comprehensive survey of Ras muta- Disclosure of Potential Conflicts of Interest tions in cancer. Cancer Res 2012;72:2457–67. S. Caenepeel is a senior scientist at Amgen Inc. and has ownership 2. Simanshu DK, Nissley DV, McCormick F. RAS proteins and their interest (including stock, patents, etc.) in the same. L.V. Sequist is regulators in human disease. Cell 2017;170:17–33. a consultant/advisory board member for AstraZeneca, Genentech, 3. Pao W, Miller VA, Politi KA, Riely GJ, Somwar R, Zakowski MF, et al. Blueprint Medicines, Pfizer, Novartis, Merrimack, and Boehringer Acquired resistance of lung adenocarcinomas to gefitinib or erlotinib is associated with a second mutation in the EGFR kinase domain. Ingelheim. Z. Piotrowska is a consultant/advisory board member for PLoS Med 2005;2:e73. AstraZeneca, Novartis, Takeda/Ariad, Guardant Health, and AbbVie. 4. Janne PA, van den Heuvel MM, Barlesi F, Cobo M, Mazieres J, Crino L, J.F. Gainor is a consultant/advisory board member for Novartis, et al. Selumetinib plus docetaxel compared with docetaxel alone and Bristol-Myers Squibb, Incyte, Takeda, Pfizer, Roche, Amgen, Agios, progression-free survival in patients with KRAS-mutant advanced Jounce, Oncorus, and Regeneron. S.P. Brown has ownership interest non-small cell lung cancer: the SELECT-1 randomized clinical trial. (including stock, patents, etc.) in Amgen. A. Coxon has ownership JAMA 2017;317:1844–53. interest (including stock, patents, etc.) in Amgen. P.E. Hughes has 5. Blumenschein GR Jr, Smit EF, Planchard D, Kim DW, Cadranel J, De ownership interest (including stock, patents, etc.) in Amgen Inc. Pas T, et al. A randomized phase II study of the MEK1/MEK2 inhibi- A.N. Hata reports receiving commercial research grants from Amgen, tor trametinib (GSK1120212) compared with docetaxel in KRAS- Novartis, and Relay Therapeutics. No potential conflicts of interest mutant advanced non-small-cell lung cancer (NSCLC). Ann Oncol were disclosed by the other authors. 2015;26:894–901. 6. Bedard PL, Tabernero J, Janku F, Wainberg ZA, Paz-Ares L, Van- Authors’ Contributions steenkiste J, et al. A phase Ib dose-escalation study of the oral Conception and design: F.M. Siddiqui, S. Caenepeel, D. Rigas, pan-PI3K inhibitor buparlisib (BKM120) in combination with the C.G. Azzoli, S.P. Brown, A. Coxon, P.E. Hughes, A.N. Hata oral MEK1/2 inhibitor trametinib (GSK1120212) in patients with selected advanced solid tumors. Clin Cancer Res 2015;21:730–8. Development of methodology: V. Nangia, F.M. Siddiqui, S. Caenepeel, 7. Shimizu T, Tolcher AW, Papadopoulos KP, Beeram M, Rasco DW, D. Timonina, D. Rigas, C.G. Azzoli, S.P. Brown, P.E. Hughes, A.N. Hata Smith LS, et al. The clinical effect of the dual-targeting strategy Acquisition of data (provided animals, acquired and man- involving PI3K/AKT/mTOR and RAS/MEK/ERK pathways in aged patients, provided facilities, etc.): V. Nangia, S. Caenepeel, patients with advanced cancer. Clin Cancer Res 2012;18:2316–25. D. Timonina, S.J. Bilton, M. Gomez-Caraballo, H.L. Archibald, C. Li, 8. Ostrem JM, Peters U, Sos ML, Wells JA, Shokat KM. K-Ras(G12C) C. Fraser, D. Rigas, K. Vajda, L.A. Ferris, C.D. Wright, K.A. Raskin, inhibitors allosterically control GTP affinity and effector interac- D.P. Cahill, J.H. Shin, C. Keyes, L.V. Sequist, Z. Piotrowska, A.F. Farago, tions. Nature 2013;503:548–51. C.G. Azzoli, J.F. Gainor, K.A. Sarosiek, A.N. Hata 9. Lito P, Solomon M, Li LS, Hansen R, Rosen N. Allele-specific inhibi- Analysis and interpretation of data (e.g., statistical analysis, tors inactivate mutant KRAS G12C by a trapping mechanism. Science biostatistics, computational analysis): V. Nangia, F.M. Siddiqui, 2016;351:604–8. S. Caenepeel, N. Phan, D. Rigas, L.A. Ferris, M. Lanuti, C.G. Azzoli, 10. Singh A, Greninger P, Rhodes D, Koopman L, Violette S, Bardeesy N, K.A. Sarosiek, S.P. Brown, P.E. Hughes, A.N. Hata et al. A gene expression signature associated with “K-Ras addiction” Writing, review, and/or revision of the manuscript: V. Nangia, reveals regulators of EMT and tumor cell survival. Cancer Cell 2009;15: F.M. Siddiqui, S. Caenepeel, D. Timonina, C. Li, M. Lanuti, C.D. 489–500. Wright, J.H. Shin, L.V. Sequist, Z. Piotrowska, C.G. Azzoli, J.F. Gainor, 11. Corcoran RB, Cheng KA, Hata AN, Faber AC, Ebi H, Coffee EM, et al. K.A. Sarosiek, A. Coxon, C.H. Benes, P.E. Hughes, A.N. Hata Synthetic lethal interaction of combined BCL-XL and MEK inhibi- Administrative, technical, or material support (i.e., reporting or tion promotes tumor regressions in KRAS mutant cancer models. organizing data, constructing databases): V. Nangia, D. Timonina, Cancer Cell 2013;23:121–8. N. Phan, L.A. Ferris, C.G. Azzoli, K.A. Sarosiek, C.H. Benes 12. Tan N, Wong M, Nannini MA, Hong R, Lee LB, Price S, et al. Bcl-2/ Bcl-xL inhibition increases the efficacy of MEK inhibition alone and Study supervision: V. Nangia, S. Caenepeel, A.N. Hata in combination with PI3 kinase inhibition in lung and pancreatic Other (compound synthesis): S.P. Brown tumor models. Mol Cancer Ther 2013;12:853–64. Acknowledgments 13. Cragg MS, Harris C, Strasser A, Scott CL. Unleashing the power of inhibitors of oncogenic kinases through BH3 mimetics. Nat Rev We thank members of the Hata Lab for helpful discussions and Cancer 2009;9:321–6. feedback. This study was funded by support from the Conquer 14. Hata AN, Engelman JA, Faber AC. The BCL2 family: key mediators Cancer Foundation of ASCO (A. Hata), Uniting Against Lung Cancer of the apoptotic response to targeted anticancer therapeutics. Cancer (A. Hata), NIHR01CA14059408 (C. Benes), Lungstrong, and Be a Discov 2015;5:475–87.

OF15 | CANCER DISCOVERY DECEMBER 2018 www.aacrjournals.org

Combination MEK + MCL1 Inhibitors for KRAS-Mutant NSCLC RESEARCH ARTICLE

15. Luciano F, Jacquel A, Colosetti P, Herrant M, Cagnol S, Pages G, et al. 29. Cragg MS, Kuroda J, Puthalakath H, Huang DC, Strasser A. Gefitinib- Phosphorylation of Bim-EL by Erk1/2 on serine 69 promotes its deg- induced killing of NSCLC cell lines expressing mutant EGFR requires radation via the proteasome pathway and regulates its proapoptotic BIM and can be enhanced by BH3 mimetics. PLoS Med 2007;4: function. Oncogene 2003;22:6785–93. 1681–89; discussion 90. 16. Meng J, Fang B, Liao Y, Chresta CM, Smith PD, Roth JA. Apoptosis 30. Gong Y, Somwar R, Politi K, Balak M, Chmielecki J, Jiang X, et al. induction by MEK inhibition in human lung cancer cells is mediated Induction of BIM is essential for apoptosis triggered by EGFR kinase by Bim. PLoS One 2010;5:e13026. inhibitors in mutant EGFR-dependent lung adenocarcinomas. PLoS 17. Beroukhim R, Mermel CH, Porter D, Wei G, Raychaudhuri S, Dono- Med 2007;4:1655–68. van J, et al. The landscape of somatic copy-number alteration across 31. Hata AN, Niederst MJ, Archibald HL, Gomez-Caraballo M, Siddiqui human cancers. Nature 2010;463:899–905. FM, Mulvey HE, et al. Tumor cells can follow distinct evolutionary 18. Kotschy A, Szlavik Z, Murray J, Davidson J, Maragno AL, Le Toume- paths to become resistant to epidermal growth factor receptor inhibi- lin-Braizat G, et al. The MCL1 inhibitor S63845 is tolerable and effec- tion. Nat Med 2016;22:262–9. tive in diverse cancer models. Nature 2016;538:477–82. 32. Cragg MS, Jansen ES, Cook M, Harris C, Strasser A, Scott CL. Treat- 19. Caenepeel S, Brown SP, Belmontes B, Moody G, Keegan KS, Chui D, ment of B-RAF mutant human tumor cells with a MEK inhibitor et al. AMG 176, a selective MCL1 inhibitor, is effective in hematologic requires Bim and is enhanced by a BH3 mimetic. J Clin Invest 2008;118: cancer models alone and in combination with established therapies. 3651–9. Cancer Discov 2018;8:1582–97. 33. Merino D, Whittle JR, Vaillant F, Serrano A, Gong JN, Giner G, 20. Lehar J, Krueger AS, Avery W, Heilbut AM, Johansen LM, Price ER, et al. Synergistic action of the MCL-1 inhibitor S63845 with current et al. Synergistic drug combinations tend to improve therapeutically therapies in preclinical models of triple-negative and HER2-amplified relevant selectivity. Nat Biotechnol 2009;27:659–66. breast cancer. Sci Transl Med 2017;9. pii: eaam7049. doi: 10.1126/ 21. Hoeflich KP, O’Brien C, Boyd Z, Cavet G, Guerrero S, Jung K, et al. scitranslmed.aam7049. In vivo antitumor activity of MEK and phosphatidylinositol 3-kinase 34. Konopleva M, Contractor R, Tsao T, Samudio I, Ruvolo PP, Kitada inhibitors in basal-like breast cancer models. Clin Cancer Res 2009; S, et al. Mechanisms of apoptosis sensitivity and resistance to the 15:4649–64. BH3 mimetic ABT-737 in acute myeloid leukemia. Cancer Cell 2006; 22. Wohrle S, Weiss A, Ito M, Kauffmann A, Murakami M, Jagani Z, et al. 10:375–88. Fibroblast growth factor receptors as novel therapeutic targets in 35. van Delft MF, Wei AH, Mason KD, Vandenberg CJ, Chen L, Czabotar SNF5-deleted malignant rhabdoid tumors. PLoS One 2013;8:e77652. PE, et al. The BH3 mimetic ABT-737 targets selective Bcl-2 proteins 23. Leverson JD, Phillips DC, Mitten MJ, Boghaert ER, Diaz D, Tahir SK, and efficiently induces apoptosis via Bak/Bax if Mcl-1 is neutralized. et al. Exploiting selective BCL-2 family inhibitors to dissect cell sur- Cancer Cell 2006;10:389–99. vival dependencies and define improved strategies for cancer therapy. 36. Lin X, Morgan-Lappe S, Huang X, Li L, Zakula DM, Vernetti LA, et al. Sci Transl Med 2015;7:279ra40. ‘Seed’ analysis of off-target siRNAs reveals an essential role of Mcl-1 24. Ni Chonghaile T, Sarosiek KA, Vo TT, Ryan JA, Tammareddi A, in resistance to the small-molecule Bcl-2/Bcl-XL inhibitor ABT-737. Moore Vdel G, et al. Pretreatment mitochondrial priming correlates Oncogene 2007;26:3972–9. with clinical response to cytotoxic chemotherapy. Science 2011;334: 37. Deng J, Carlson N, Takeyama K, Dal Cin P, Shipp M, Letai A. BH3 1129–33. profiling identifies three distinct classes of apoptotic blocks to pre- 25. Gilmartin AG, Bleam MR, Groy A, Moss KG, Minthorn EA, Kulkarni dict response to ABT-737 and conventional chemotherapeutic agents. SG, et al. GSK1120212 (JTP-74057) is an inhibitor of MEK activ- Cancer Cell 2007;12:171–85. ity and activation with favorable pharmacokinetic properties for 38. Montero J, Sarosiek KA, DeAngelo JD, Maertens O, Ryan J, Ercan D, sustained in vivo pathway inhibition. Clin Cancer Res 2011;17: et al. Drug-induced death signaling strategy rapidly predicts cancer 989–1000. response to chemotherapy. Cell 2015;160:977–89. 26. Sun C, Hobor S, Bertotti A, Zecchin D, Huang S, Galimi F, et al. 39. Edlich F, Banerjee S, Suzuki M, Cleland MM, Arnoult D, Wang C, Intrinsic resistance to MEK inhibition in KRAS mutant lung and et al. Bcl-x(L) retrotranslocates Bax from the mitochondria into the colon cancer through transcriptional induction of ERBB3. Cell Rep cytosol. Cell 2011;145:104–16. 2014;7:86–93. 40. Faber AC, Corcoran RB, Ebi H, Sequist LV, Waltman BA, Chung E, 27. Flaherty KT, Robert C, Hersey P, Nathan P, Garbe C, Milhem M, et al. et al. BIM expression in treatment naïve cancers predicts responsive- Improved survival with MEK inhibition in BRAF-mutated mela- ness to kinase inhibitors. Cancer Discov 2011;1:352–65. noma. N Engl J Med 2012;367:107–14. 41. Crystal AS, Shaw AT, Sequist LV, Friboulet L, Niederst MJ, Locker- 28. Roberts AW, Davids MS, Pagel JM, Kahl BS, Puvvada SD, Gerecitano man EL, et al. Patient-derived models of acquired resistance can JF, et al. Targeting BCL2 with venetoclax in relapsed chronic lympho- identify effective drug combinations for cancer. Science 2014;346: cytic leukemia. N Engl J Med 2016;374:311–22. 1480–6.

DECEMBER 2018 CANCER DISCOVERY | OF16

Exploiting MCL1 Dependency with Combination MEK + MCL1 Inhibitors Leads to Induction of Apoptosis and Tumor Regression in KRAS-Mutant Non−Small Cell Lung Cancer

Varuna Nangia, Faria M. Siddiqui, Sean Caenepeel, et al.

Cancer Discov Published OnlineFirst September 25, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-18-0277

Supplementary Access the most recent supplemental material at: Material http://cancerdiscovery.aacrjournals.org/content/suppl/2018/09/25/2159-8290.CD-18-0277.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerdiscovery.aacrjournals.org/content/early/2018/11/14/2159-8290.CD-18-0277. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.