Sym015: a Highly Efficacious Antibody Mixture Against MET-Amplified Tumors

Published OnlineFirst July 5, 2017; DOI: 10.1158/1078-0432.CCR-17-0782

Cancer Therapy: Preclinical Clinical Cancer Research Sym015: A Highly Efficacious Antibody Mixture against MET-Amplified Tumors Thomas Tuxen Poulsen, Michael Monrad Grandal, Niels Jørgen Østergaard Skartved, Rikke Hald, Lene Alifrangis, Klaus Koefoed, Trine Lindsted, Camilla Frohlich,€ Sofie Ellebæk Pollmann, Karsten Wessel Eriksen, Anna Dahlman, Helle Jane Jacobsen, Thomas Bouquin, Mikkel Wandahl Pedersen, Ivan David Horak, Johan Lantto, and Michael Kragh

Abstract

Purpose: Activation of the receptor tyrosine kinase MET is analogue of emibetuzumab, a monoclonal IgG4 antibody against associated with poor clinical outcome in certain cancers. To target MET currently in clinical development. Sym015 also induced MET more effectively, we developed an antagonistic antibody antibody-dependent cellular cytotoxicity (ADCC) in vitro, suggest- mixture, Sym015, consisting of two humanized mAbs directed ing that secondary effector functions contribute to the efficacy of against nonoverlapping epitopes of MET. Sym015. Experimental Design/Results: We screened a large panel of Retrospectively, all responsive, high MET-expressing models well-annotated human cancer cell lines and identified a subset were scored as highly MET-amplified by in situ hybridization, with highly elevated MET expression. In particular, cell lines of pointing to MET amplification as a predictive biomarker for lung cancer and gastric cancer origin demonstrated high MET efficacy. Preclinical toxicology studies in monkeys showed that expression and activation, and Sym015 triggered degradation of Sym015 was well tolerated, with a pharmacokinetic profile sup- MET and significantly inhibited growth of these cell lines. Next, we porting administration of Sym015 every second or third week in tested Sym015 in patient- and cell line–derived xenograft models humans. with high MET expression and/or MET exon 14 skipping altera- Conclusions: The preclinical efficacy and safety data provide tions, and in models harboring MET amplification as a mecha- a clear rationale for the ongoing clinical studies of Sym015 nism of resistance to EGFR-targeting agents. Sym015 effectively in patients with MET-amplified tumors. Clin Cancer Res; 23(19); inhibited tumor growth in all these models and was superior to an 5923–35. 2017 AACR.

Introduction (NSCLC; refs. 3, 4). In addition, genetic alterations leading to MET exon 14 skipping have recently gained interest. MET exon 14 The receptor tyrosine kinase (RTK) MET, also known as c-MET encodes the juxtamembrane receptor domain containing the or hepatocyte growth factor (HGF) receptor (HGFR), is essential binding site for the ubiquitin ligase CBL, responsible for triggering for embryonic and liver development, wound healing, and tissue MET internalization and degradation. Exon 14 skipping thus regeneration. Lately, evidence has accumulated for the role of induces stabilization of MET on the cell surface, leading to elevated MET activation as an oncogenic driver in a number of enhanced and constitutively active MET signaling (5, 6). A shift malignancies (1). In a context-dependent manner, high activation to MET dependency as a mechanism of resistance to tyrosine of MET induces increased cell proliferation, survival, motility, kinase inhibitors (TKI) targeting other RTKs is frequently scattering, and angiogenesis and drives epithelial–mesenchymal observed, and up to 20% of EGFR-mutated NSCLCs develop transition and cell invasion (2). resistance by MET gene amplification in response to treatment There are multiple mechanisms for aberrant MET activation in with EGFR-targeting TKIs (7, 8). tumorigenesis, including overexpression of its ligand HGF, as well A number of MET-targeted TKIs and therapeutic antibodies as amplification, activating mutations, and other modifications of have entered clinical development, but so far, none have been the MET gene. In particular, MET gene amplification is observed in approved as MET-targeting agents (9). TKIs generally have limited a distinct subset of patients and is associated with poor prognosis target specificity, which may lead to pleiotropic effects and in malignancies such as gastric and non–small cell lung cancer unwanted toxicity. Furthermore, after long-term therapy with MET-targeting TKIs, subsets of quiescent tumor cells survive treatment, resulting in a rebound effect with aggressive tumor Symphogen A/S, Ballerup, Denmark. regrowth upon release of treatment pressure (10, 11). MET-target- Note: Supplementary data for this article are available at Clinical Cancer ing antibodies in clinical development include the one-armed Research Online (http://clincancerres.aacrjournals.org/). engineered antibody onartuzumab (MetMAb), which demon- Corresponding Author: Thomas Tuxen Poulsen, Symphogen A/S, Pederstrup- strated activity in early-stage clinical trials (12–15), but failed to þ þ vej 93, Ballerup DK-2750, Denmark. Phone: 45 4526-5050; Fax: 45 4526- improve survival when used in combination with erlotinib to treat 5060; E-mail: [email protected] NSCLC in a phase III study (16). More recently, other antibodies doi: 10.1158/1078-0432.CCR-17-0782 targeting MET have entered the clinic. These include ABT-700, 2017 American Association for Cancer Research. which has demonstrated activity in patients with MET-amplified

www.aacrjournals.org 5923

Poulsen et al.

mycoplasma-free, cultured per supplier recommendations, con- Translational Relevance tinuously evaluated for visual changes, and either used within In certain malignancies, such as gastric cancer and lung 6 months of purchase or profiled by cell authentication testing cancer, increased activation of the receptor tyrosine kinase (LGC Standards, Germany). Cells were discarded after 6 weeks of (RTK) MET is associated with poor clinical outcome and culture to minimize the risk of alterations. The cell line resistance to targeted therapies. A number of MET-targeting HCC827R1_Cet#3 was obtained by harvesting cells derived from mAbs are in clinical development, but so far, none have been cetuximab-resistant xenograft tumors of the erlotinib-resistant approved. In recent years, preclinical and early clinical studies cell line HCC827R1 (26). have demonstrated enhanced efficacy and antitumor activity of antibody mixtures directed against different RTK targets, Quantification of receptor protein levels in cell lines and leading to development of Sym015, an MET-targeting anti- xenografts body mixture comprising two antibodies directed against Samples were diluted to 0.2–1.0 mg/mL in a master mix contain- nonoverlapping epitopes. In the current study, we demon- ing fluorescent standards and reducing agent, and processed strate that Sym015 inhibits MET-dependent cancer cell and under standard conditions in a Sally/Sally Sue instrument in vitro in vivo tumor growth in multiple MET-dependent and (ProteinSimple, CA, USA). The following rabbit primary antibo- models, including models exhibiting resistance to EGFR- dies (from Cell Signaling Technology, MA, USA) diluted 1:50 were fi targeting agents. The promising preclinical ef cacy and safety used (clone number and phosphorylation site denoted in paren- data justify the ongoing clinical studies of Sym015 in patients theses): MET, pMET (D26; Y1234/Y1235), pMET (130H2; MET fi with -ampli ed tumors. Y1349), EGFR (C74B9), pEGFR (Y1068), and Pan-Actin.

Sym015 and control antibodies Sym015 comprises a balanced (1:1) mixture of two MET- tumors (17, 18), and ARGX-111, which has been glycoengineered targeting monoclonal antibodies (Hu9006 and Hu9338) gener- to enhance secondary effector functions and also demonstrated ated and characterized as described in detail elsewhere (ref. 25, efficacy in a single patient with MET gene amplification (19, 20). Grandal and colleagues, manuscript submitted). Briefly, mice One of the most advanced MET-targeting antibody therapeutics in were immunized with soluble, recombinant MET protein and clinical development is emibetuzumab (LY2875358), a human- MET-overexpressing human cancer cells. Splenic B cells were ized IgG4 reported to block ligand binding and induce receptor harvested and sorted to single cells, antibodies were isolated and internalization (21). Emibetuzumab has demonstrated prelimi- cloned using Symplex technology (27), humanized, and nary evidence of antitumor activity in combination with erlotinib expressed as full-length IgG1 in mammalian cells. A negative in patients with MET-positive tumors and is currently being control antibody targeting a nonhuman antigen was cloned and evaluated in phase II studies (22). expressed using the same approach. The Sym015 concentration As an alternative to mAbs, mixtures comprising more than used for in vitro and in vivo experiments refers to the total antibody one antibody have demonstrated strong potential as RTK- concentration (e.g., for 50 mg/mL Sym015, the concentration of targeting anticancer agents. A major therapeutic advantage of each constituent antibody is 25 mg/mL). Cetuximab (Erbitux) was antibody mixtures is their ability to orchestrate receptor inter- purchased from Merck KGaA, Germany. An analogue of the MET- nalization and degradation in a manner superior to mAbs, as targeting antibody emibetuzumab (LY2875358; ref. 21) was previously demonstrated for different targets in the human generated on the basis of publicly available sequence information EGFR family (23, 24). (28) by cloning and expressing the variable region sequences in In light of the role of MET as a therapeutic target in cancer and human IgG4 format in mammalian cells. Epitope mapping dem- the enhanced activity of antibody mixtures in targeting other onstrated that the analogue binds an epitope located in blade 2/3 RTKs, we developed the antibody mixture Sym015 comprising of the SEMA domain (data not shown), in accordance with the two antibodies directed against nonoverlapping epitopes on the previously published epitope for emibetuzumab (28). MET extracellular domain (ref. 25, Grandal and colleagues, man- fi uscript submitted). Here, we present data demonstrating ef cacy In vitro cell metabolic activity MET fi of Sym015 in ampli ed cancer models. Sym015 demon- Metabolic activity was measured by WST-1 assay (Roche Diag- strated superior anti-proliferative and tumor growth-inhibitory nostics) following DMSO, erlotinib (Selleckchem), or antibody effects in many cell lines and xenograft models, including models treatment for 96 hours as described previously (24). resistant to an analogue ofemibetuzumab. Furthermore, Sym015 was efficacious in models harboring MET exon 14 deletions and in In vivo models with acquired MET dependency in response to EGFR- xenograft studies – targeting agents. In combination with favorable safety and phar- Studies were performed using female mice. For cell line macokinetic (PK) profiles of Sym015 in cynomolgus monkeys, derived xenograft (CDX) models EBC-1 and Hs746T, 8- to nu/nu these results have enabled recent initiation of clinical trials of 9-week-old BALB/c and C.B-17 SCID mice were obtained Sym015 efficacy in patients with MET-amplified tumors. from Taconic A/S or Janvier Labs. Exponentially growing cells were harvested and resuspended in medium, and 5 106/1 107 Materials and Methods cells were injected subcutaneously. Studies in patient-derived xenograft (PDX) models were performed at Crown Bioscience Cell culture and reagents (models LU0858, LU1901, and LU2503 in 8- to 10-week-old Source, origin, supplier, and growth medium for each cell line BALB/c nu/nu mice), Crown Bioscience, UK/Precos (model are summarized in Supplementary Table S1. All cell lines were LU6425 in 5- to 8-week-old HsdOla:MF1-FOXN1nu mice), or

5924 Clin Cancer Res; 23(19) October 1, 2017 Clinical Cancer Research

Sym015 Inhibits MET-Ampliﬁed Tumors

Oncotest GmbH (model LXFA0526 in 4- to 6-week-old NMRI Bioscience). After 1 day, the growth medium was changed nu/nu mice) using subcutaneously serially transplanted tumor to assay medium containing HGF (200 ng/mL), Sym015 fragments. (50 mg/mL), Sym015 þ HGF, or a negative control antibody. Tumors were measured two to three times weekly by caliper, After 116 hours, the experiment was stopped, and the images were and tumor volume was calculated using the formula: 0.5 analyzed for cell confluence. length (width)2. At a predetermined tumor size, mice were randomized into groups for treatment. Antibodies were admin- In vivo pharmacokinetic and safety studies in cynomolgus istered intraperitoneally three times weekly for a maximum of monkeys fi 10 doses. Harvested tumors were xed in formalin and embedded Serum levels of Sym015 and the two antibodies constituting fi in vivo in paraf n or snap frozen. All studies were performed in Sym015 (Hu9006 and Hu9338) were measured at various time accordance with applicable national laws and regulations relating points after a single intravenous infusion of Sym015 at 15 mg/kg to the care and use of laboratory animals. to two male and two female cynomolgus monkeys. Sym015, Hu9006, and Hu9338 were quantified in monkey serum using IHC two different immunoassay methods on the Gyrolab xP platform IHC experiments were conducted as described previously (29) (Gyros AB), one detecting human IgG, that is, total Sym015, and by MicroMorph Histology Services using the following primary one separately detecting binding of the two antibodies to recom- antibodies: FITC-conjugated human anti-MET antibody cloned binant MET ectodomains uniquely specific for either Hu9006 or fi and produced at Symphogen (con rmed to bind an epitope not Hu9338. overlapping with Sym015, data not shown), rabbit anti-pMET In a regulated repeat-dose toxicity study, Sym015 was admin- (Y1003, R&D Systems), and mouse anti-Ki67 (Dako). Slides istered to cynomolgus monkeys (n ¼ 3/sex/group in the main incubated with anti-MET antibody were washed and incubated study and n ¼ 2/sex/group in the control group and the high-dose with rabbit anti-FITC antibody (Serotec/Bio-Rad), followed by group for recovery) once weekly for 4 weeks by intravenous incubation with horseradish peroxidase (HRP)-conjugated infusion at dose levels of 0, 15, 50, or 100 mg/kg. At the end of BrightVision anti-rabbit for MET/pMET or HRP anti-mouse 0 the study, kidney, liver, and skin tissues were collected, snap (Immunologic) for Ki67. Finally, 3,3 -diaminobenzidine (DAB) frozen, and used for IHC. The study was ethically approved, hematoxylin staining was performed. Assessment of general his- conducted in compliance with Good Laboratory Practice, and tology/staining intensity was blinded. Staining of MET and pMET included standard toxicology and toxicokinetic endpoints. was scored as negative (0), weak (þ), moderate (þþ), or strong (þþþ). Ki67 staining was scored as no positive cells (0), few (þ), Statistical analysis moderate (þþ), or high number of positive cells (þþþ). Statistical analyses were performed using the unpaired Student t test and P values below 0.05 were considered statistically Antibody-dependent cellular cytotoxicity significant. Cells loaded with 51Cr (PerkinElmer) for 1 hour were incubated with antibodies and freshly isolated peripheral blood mononu- clear cells for 4 hours. Radioactivity in the supernatant was Results measured as counts per minute with LumaPlates in a TopCount Sym015 inhibits proliferation of cell lines with elevated MET instrument (PerkinElmer). Specific lysis was calculated as a per- protein expression centage of maximum lysis (incubation with 1% Triton X-100) To search for MET-dependent in vitro models, we screened a with spontaneous lysis subtracted. large panel of cell lines derived from a broad range of malignancies and with varying cell morphology for expression of total MET Quantitative PCR–based quantification of MET amplification and phosphorylated MET by quantitative Simple Western anal- Genomic DNA was isolated using a QIAamp DNA Mini Kit ysis. As shown in Fig. 1A, a subset of cell lines of primarily lung (Qiagen). Quantitative PCR was carried out using a qBiomarker cancer and gastric cancer origin expressed very high levels of MET MET copy-number PCR assay (VPH107-0581564A) and a qBio- and phosphorylated MET protein, indicative of MET dependency, marker Multicopy Reference PCR assay (Qiagen). MET gene copy and therefore, the effect of Sym015 in these models was further number was determined by the DDCt method, and using DNA explored. from the cell line T47D and the PDX tumor LI0050 as diploid MET The antiproliferative effect of Sym015 was compared with references. treatment with an analogue of the MET-targeting mAb emibetuzumab (LY2875358, currently in phase II clinical trials) in three FISH scoring of MET amplification representative high MET-expressing gastric (OKAJIMA and FISH scoring was performed at Laboratory of Tumor Genetics, KATO-II) and lung (H1993) cancer cell lines. As shown in Fig. Radboud University Medical Center, Nijmegen, the Netherlands. 1B, Sym015 demonstrated similar or slightly better antiprolifera- Tissue slides were pretreated by microwaving in sodium citrate tive efficacy than the emibetuzumab analogue. pH 6.0, digested in pepsin for 15 minutes at 37C, denatured at As antibody mixtures against other RTKs, such as EGFR, HER2, 80C, and hybridized overnight at 37Cwithprobesdiluted1:10 and HER3, are known to effectively trigger target receptor inter- for MET (Vysis MET FISH Probe 06N05-020) and the chromosome nalization and degradation, the effect of Sym015 treatment on 7 centromere (Vysis CEP 7 FISH probe D7S522) both Abbott. total and phosphorylated MET protein levels was investigated and compared with that of the emibetuzumab analogue. As shown Primary hepatocytes in Fig. 1C, Sym015 treatment of cells for 24 hours caused effective Primary hepatocytes seeded in 96-well, collagen-coated plates degradation of MET protein and consequently inhibited MET were incubated in an IncuCyte ZOOM microscope (Essen phosphorylation in all three cell lines. Again, the effect of Sym015

www.aacrjournals.org Clin Cancer Res; 23(19) October 1, 2017 5925

Poulsen et al.

A 200,000 Tissue origin Breast MET Central nervous system pMET (Y1234/1235) Endometrium 150,000 Esophagus Hematopoietic/lymphoid Large intestine Lung 100,000 Ovary Pancreas Prostate 50,000 Soft tissue Signal intensity Signal Stomach Upper aerodigestive tract Urinary tract 0 Tissue origin RD T84 H23 N87 HN5 TE-II GEO RT-4 KP-4 H596 H292 A549 H508 H358 H820 LoVo FaDu OE33 PL-21 HT-29 GP5-d RMG-I H1993 H1703 H1573 H1648 H2009 EBC-1 SNU-5 DU145 AU565 HCT15 BxPC3 SW403 SW948 SW837 SW620 RT-112 RL95-2 MKN74 U87MG LS1034 SKMG3 HPAF II LS174T ASPC-1 HS746T SKBR-3 KATO-II MKN-45 Okajima KATO-III HCC 202 HCC 827 CFPAC-1 HCC2218 SNU-C2A CAPAN-2 OVCAR-5 MOLM-13 BFTC-905 KYSE-520 Panc02.13 Panc 08.13 Cell lines B H1993 Kato-II 150 Okajima 150 150

100 100 100

50 50 50 Control mAb Metabolic activity Metabolic activity Metabolic activity Metabolic Emibetuzumab analogue % of untreated control untreated % of % of untreated control untreated % of

% untreated of control Sym015 0 0 0 0.01 0.1 1 10 100 0.01 0.1 1 10 100 0.01 0.1 1 10 100 Antibody concentration (µg/mL) Antibody concentration (µg/mL) Antibody concentration (µg/mL)

C Okajima H1993 Kato-II Neg. control 150 150 150 Emibetuzumab analogue Sym015 100 100 100

** ** ** * ** 50 50 ** 50 Signal intensity Signal Signal intensity Signal Signal intensity Signal % negative control of % negative control of % negative control of 0 0 0 MET pMET (Y1234/1235) MET pMET (Y1234/1235) MET pMET (Y1234/1235)

Figure 1. Cell lines exhibiting high MET expression and phosphorylation are sensitive to Sym015. A, Relative quantiﬁcation of total MET (blue bars) and phosphorylated MET (pMET, MET phosphorylated at Y1234/Y1235, green bars) in 64 cell lines, denoted as signal intensity calculated using Compass software (ProteinSimple). B, Dose–response curves for cellular metabolic activity in the MET-dependent cell lines Okajima, H1993, and Kato-II upon treatment with the emibetuzumab analogue (red) or Sym015 (green). Data, mean SEM. C, Relative quantiﬁcation of total MET and pMET (phosphorylated at Y1234/Y1235) in Okajima, H1993, and KATO-II cell lysates after treatment with Sym015 (green), emibetuzumab analogue (red), or negative control antibody (black) for 24 hours. Signal intensity was normalized to the corresponding actin signal as loading control and is presented as mean percentage of the negative control SEM. Experiments were performed in triplicate.

was similar to or slightly better than that of the emibetuzumab was maintained after treatment withdrawal. Interestingly, large analogue. tumors that outgrew emibetuzumab analogue treatment retained sensitivity to Sym015 (Fig. 2A). To investigate the Sym015 inhibits tumor growth of MET-dependent CDX and effects of lower doses of Sym015 on tumor growth, mice PDX models transplanted with a high MET-expressing PDX lung cancer The antitumor activity of Sym015 was further explored in vivo.A model (LXFA0526) were randomized into ﬁve groups that number of high MET-expressing CDX and PDX models were received 1, 5, 25, or 50 mg/kg Sym015 or 50 mg/kg emibetu- treated three times weekly with Sym015 or the emibetuzumab zumab analogue. As shown in Fig. 2B, Sym015 dosed at 25 or analogue, and tumor growth was monitored. Results of treatment 50 mg/kg effectively eliminated the tumors, whereas the emi- of the MET-dependent lung cancer CDX model EBC-1 are shown betuzumab analogue dosed at 50 mg/kg resulted in tumor in Fig. 2A. Although emibetuzumab analogue treatment caused growth inhibition similar to that with Sym015 dosed at transient tumor growth inhibition followed by regrowth, repeated 5 mg/kg. Furthermore, EBC-1 tumors were effectively inhibited dosing with Sym015 resulted in sustained tumor shrinkage, which at a Sym015 dose of 10 mg/kg (Supplementary Fig. S1).

5926 Clin Cancer Res; 23(19) October 1, 2017 Clinical Cancer Research

Sym015 Inhibits MET-Ampliﬁed Tumors

Sym015 comprises two mAbs in a balanced ratio. To investigate sensitive to the EGFR-targeted agents cetuximab and erlotinib and whether different ratios between the two antibodies affected only moderately sensitive to Sym015 and the emibetuzumab efficacy, antitumor activity of Sym015 mixtures comprising analogue, the cetuximab/erlotinib–resistant cell line demonstrat- skewed ratios of the two antibodies was evaluated in the EBC-1 ed increased sensitivity to MET targeting by the emibetuzumab model. No significant difference was observed between the inhib- analogue and Sym015 (Fig. 3B). To further explore the effect of itory effects of skewed and balanced antibody ratios on tumor Sym015 in the EGFR-resistant setting in vivo, two lung cancer PDX growth at two dose levels (Supplementary Fig. S2). models (LU0858 and LU1901) with activating EGFR tyrosine Sym015- and emibetuzumab analogue-treated tumors were kinase mutations and MET gene amplification were treated with also resected and evaluated for MET and phosphorylated MET Sym015 or the emibetuzumab analogue. As shown in Fig. 3C and protein levels by IHC and Simple Western analyses. Confirming D, Sym015 induced strong tumor growth inhibition in both in vitro findings, Sym015 was found to effectively reduce MET models, while only one of the models (LU1901) responded to and phosphorylated MET protein levels in the tumors by more emibetuzumab analogue treatment. than 50%, as quantified by Simple Western blotting (Fig. 2C–E). Furthermore, reduced proliferation, as indicated by a reduced MET-overexpressing Sym015-sensitive models harbor Ki67 score, was observed in Sym015-treated tumors (Fig. 2C high-level MET gene amplification and D). In the clinic, high MET expression scored by IHC and MET gene As mentioned previously, mutations in the MET gene resulting amplification scored by in situ hybridization or next-generation in exon 14 skipping cause stabilization of the receptor on the cell sequencing are commonly applied to stratify patients for enroll- surface and can drive MET dependency (5, 6). Therefore, Sym015 ment into clinical trials with MET-targeting antibody therapeutics, was evaluated and compared with treatment with the emibetu- such as emibetuzumab. For this reason, the MET gene amplifi- zumab analogue in models harboring exon 14 skipping altera- cation status of the high MET expressing Sym015-responsive cell tions plus high MET expression indicative of dependency. As lines and xenograft models was investigated. Protein levels of the shown in Fig. 2F and G, both the lung cancer PDX model LU2503 initially screened cell lines (Fig. 1A) were compared with the and the gastric cancer CDX model Hs746T were strongly inhibited published gene amplification status of the cell line (no gene copy- by Sym015. In contrast, treatment with the emibetuzumab ana- number information was available for H1993, ref. 31), and the logue resulted in initial tumor growth inhibition followed by high MET protein-expressing Sym015-responsive cell lines were regrowth of both models. When the Hs746T tumors that outgrew all found to harbor amplification of the MET gene, with a copy emibetuzumab analogue treatment were re-treated with a single number above 5 (Fig. 4A). Quantitative PCR was performed to dose of Sym015, a strong tumor-inhibitory response was score the MET gene amplification level of the responsive models observed, similar to the findings for EBC-1 tumors (Fig. 2A and and, as shown in Fig. 4B, all responsive models were highly MET G). In addition, Sym015 demonstrated an inhibitory effect in a amplified. This translated into a MET/CEP7 ratio much higher lung cancer CDX model with an exon 14 skipping mutation and than 2, as scored by FISH, for all responsive models (Fig. 4C; low MET expression (Supplementary Fig. S3). Supplementary Table S2), confirming previously published As secondary effector functions have been found to contribute observations (17). Collectively, these data point to MET gene to the mechanism of action of antibody mixtures targeting RTKs, amplification scoring of tumor tissue as an appropriate tool for the ability of Sym015 and the emibetuzumab analogue to induce selection of patients with MET-amplified tumors for inclusion in antibody-dependent cellular cytotoxicity (ADCC) was investigat- clinical trials of Sym015. This is in line with previous reports, ed in the EBC-1 and Hs746T cell lines in vitro. As shown in Fig. 2H, which point to MET amplification as a superior biomarker com- both treatments inhibited cell proliferation of both cell lines to pared with MET IHC score in predicting MET dependency and the same degree, while ADCC was only induced by Sym015. Lack treatment response (32). of effector function of the emibetuzumab analogue was expected, as the IgG4 isotype does not induce ADCC. Thus, secondary effector functions, such as ADCC, may in part explain the in vivo Preclinical pharmacokinetics and safety of Sym015 support superiority of Sym015 over the emibetuzumab analogue (Fig. 2A, first-in-human studies F, and G). The preclinical pharmacokinetics and safety of Sym015 were evaluated in cynomolgus monkeys, as the amino acid sequence of Sym015 is effective in models with MET-driven resistance to MET is evolutionarily conserved in humans and monkeys (99% anti-EGFR therapy sequence identity). The individual antibodies constituting Transition from EGFR to MET dependency by acquired MET Sym015 bind with similar high affinity to both human and gene amplification is a confirmed mechanism of resistance to cynomolgus MET (data not shown). EGFR-targeted agents in various malignancies, such as colorectal Two elimination phases were present in the pharmacokinetic cancer and lung cancer (7, 30). To model escape from EGFR- profile (Fig. 5A), indicating the presence of a saturable elimina- targeting therapy by switching to MET dependency in vitro, the tion pathway. For a 15 mg/kg dose, the half-life was approxi- EGFR-amplified and EGFR-dependent lung cancer cell line mately 11 days up to 2 weeks after dosing and 6 days thereafter. HCC827 was subjected to long-term exposure to the EGFR-tar- The two antibodies constituting Sym015 had similar pharmaco- geted therapies erlotinib and cetuximab, thereby generating resis- kinetic profiles. tant cell clones. As shown in Fig. 3A, one of these clones After four doses of 15 to 100 mg/kg Sym015, an approximate (HCC827R1_Cet#3) was found to exhibit elevated levels of MET twofold accumulation was observed (Fig. 5B). No departure from and phosphorylated MET protein, and reduced levels of EGFR and dose proportionality was observed. Prediction of human phar- phosphorylated EGFR, compared with the parental cell line. macokinetic parameters using one-species allometric scaling Furthermore, although the parental cell line HCC827 was highly (Supplementary Table S3; ref. 33) indicates that at Sym015 serum

www.aacrjournals.org Clin Cancer Res; 23(19) October 1, 2017 5927

Downloaded from clincancerres.aacrjournals.org on September 23, 2021. © 2017 American Association for Cancer Research. 5928 lnCne e;2(9 coe ,2017 1, October 23(19) Res; Cancer Clin ole tal. et Poulsen Downloaded from Emibetuzumab H

Metabolic activity analogue Sym015 C

% of untreated control Vehicle 100 150 50 .0 .101110 1 0.1 0.01 0.001 0 Antibody concentration (µg/mL) Antibody concentration A Emibetuzumab analogue Sym015 Control mAb clincancerres.aacrjournals.org 3 F Tumor volume (mm3) Tumor volume (mm ) Published OnlineFirstJuly5,2017;DOI:10.1158/1078-0432.CCR-17-0782 1,000 1,000 1,500 2,000 EBC-1 100 200 300 400 500 600 700 800 900 500 0 0 02 01 02 03 40 35 30 25 20 15 10 5 0 MET Treatment period LU2503,MET exon 14del Days after first treatment first after Days 100 µm treatment Primary period 55 Days after cell inoculation cell after Days ADCC (% of maximum lysis) 10 15 .0100100 . 10 1 0.1 0.01 0.001 0.0001 0 5 Antibody concentration (µg/mL) Antibody concentration EBC-1 07 analogue Emibetuzumab Sym015 Vehicle pMET EBC-1 on September 23, 2021. © 2017American Association for Cancer Secondary treatment 51 Research. treatment Sym015 Secondary Sym015 Emibetuzumab analogue Vehicle G Tumor volume (mm3) period 100 200 300 400 500 600 700 800 0 04 treatment Secondary Sym015 Emibetuzumab analogue Sym015 Vehicle

Metabolic activity exonHs746T, MET 14del 0

% of untreated control 0 100 150 50 .0 .101110 1 0.1 0.01 0.001 0 Antibody concentration (µg/mL) Days afterDays cellinoculation Ki67 B Tumor volume (mm3) 1,000 1,500 2,000 Hs746T 500 treatment 0 Primary 0 period Retreated withasingledoseofSym015 0 D E Emibetuzumab analogue (50 mg/kg) (50 Emibetuzumab analogue mg/kg) (1 Sym015 mg/kg) (5 Sym015 mg/kg) (25 Sym015 Vehicle Signal intensity mg/kg) (50 Sym015 analogue group

Emibetuzumab % of vehicle control IHC Score 100 150 E 01 025 20 15 10 5

Emib Vehicle 0 1 2 3

50 m Ve 0 ib h etuz Days after first treatment e t ic uzumable umab Treatment period Total MET Total MET ns Sy Sy LXFA0526 m ADCC % of maximum lysis m0 0 10 15 .0100100 . 10 1 0.1 0.01 0.001 0.0001 15 ***

0 5 1 E 5 E mi Ve m Vehicl i Antibody concentration (µg/mL) b hi be (Y1234/35)

e (Y1234/35) tuz c tu le z e pMET u pMET um ma S *** a ym b Sym b 0 0 1 *** 15 Emib 5 lnclCne Research Cancer Clinical Hs746T Ve E mibe V hi e e h t c ic uz le t le

(Y1349) uz u m pMET um Ki67 Sym015a ** b S ab ym015 *** Published OnlineFirst July 5, 2017; DOI: 10.1158/1078-0432.CCR-17-0782

Sym015 Inhibits MET-Ampliﬁed Tumors

concentrations above the level for target saturation, the half-life of (23, 24). Therefore, aiming to identify a potent anti-MET antibody Sym015 in humans may approach the 3 weeks observed for mixture with an efficacy profile superior to that of mAbs, we endogenous IgGs (34). screened a large panel of antibody mixtures for their antiproli- Sym015 was well tolerated in cynomolgus monkeys. There ferative activity in vitro and in vivo. Sym015, consisting of two were no significant toxicologic observations or changes in hema- humanized IgG1 antibodies, directed against nonoverlapping tology, blood chemistry, or urinalysis during or after the once- epitopes of the MET ectodomain, was identified as the most weekly administration of Sym015 at 15, 50, or 100 mg/kg for potent and efficacious antibody mixture (ref. 25, Grandal and 4 weeks. No late-onset toxicity was observed during the 6-week colleagues, manuscript submitted). dosing-free recovery period, and 100 mg/kg was established as the To identify MET-dependent models for preclinical develop- no observed adverse effect level (NOAEL) in this study. ment of Sym015, and predict clinical indications that may poten- At the time of study termination 1 week after the fourth tially benefit from MET-targeting therapy, we screened a large infusion, Sym015-treated animals showed a reduction in mem- panel of well-annotated human cancer cell lines for MET protein branous MET expression in liver, kidney, and skin compared with expression. In particular, cell lines of lung and gastric cancer origin the control group (Fig. 5C). demonstrated high MET protein expression and activation and mAbs targeting MET have previously been reported to act sensitivity to Sym015 (Fig. 1A and C), and the Sym015-responsive agonistically and to increase cell motility (35). The effect of models were found to harbor MET amplification. Sym015 on a proliferative index measured by Ki67 staining was Furthermore, Sym015 demonstrated robust in vivo efficacy in investigated in kidney and liver samples from cynomolgus mon- initially EGFR-dependent NSCLC models that had escaped keys exposed to Sym015. As shown in Fig. 5C, no increase in Ki67 EGFR-targeting agents by acquiring MET gene amplification, staining was observed in these tissues. Furthermore, the effect of resulting in a shift from EGFR dependency to MET dependency Sym015 on the proliferation of nonimmortalized human hepa- (Fig. 3). Interestingly, the HCC827 cell line, which harbors tocytes was investigated in vitro. As shown in Fig. 5D, although the EGFR amplification and EGFR exon 19 deletion and exhibits MET ligand HGF induced hepatocyte proliferation, no effect was elevated MET protein levels and phosphorylation (Figs. 1A observed upon treatment with Sym015, and Sym015 neutralized and 4B), was only slightly sensitive to MET inhibition the stimulating effect of HGF. In combination with data demon- (Fig. 3B). It was previously reported that elevated MET activa- strating that Sym015 does not induce scattering of tumor cells tion in NSCLC cell lines harboring EGFR alterations (including or stimulate growth of human umbilical vein endothelial HCC827) stimulates cell invasion rather than survival (36), cells (ref. 25, Grandal and colleagues, manuscript submitted), which can explain the limited antiproliferative effect of these findings suggest that Sym015 does not stimulate cell Sym015. Both tested Sym015-responsive PDX models with proliferation. EGFR-activating mutations (LU0858 and LU1901) were previously found to be resistant to the EGFR-targeting TKIs fi Discussion ge tinibanderlotinib,andtocetuximab(CrownBioscience, personal communication). Resistance to EGFR-targeting TKIs Aberrant MET receptor activation and gene amplification are due to acquired MET amplification, potentially driving activa- associated with poor prognosis in different cancers, suggesting tion of HER3, is reported in up to 20% of NSCLCs (7, 8), and that MET-targeted therapies may benefit patients with MET gene– resistant patients may potentially benefit from Sym015 amplified tumors. A number of mAbs targeting MET are in clinical treatment. development, and some have demonstrated antitumor activity, Depending on the detection method, varying frequencies of but so far, no antibody targeting MET has been approved. MET amplification were reported in solid tumor malignancies. Previously, mixtures of antibodies binding to nonoverlapping FISH is the most commonly used method for detecting MET epitopes of RTKs have been found to induce target receptor amplification, as this methodology can distinguish between internalization and degradation in a manner superior to mAbs polysomy and true gene amplification (37). In a recent study,

Figure 2. Sym015 effectively inhibits tumor growth of MET-dependent CDX and PDX models. A, Effect of Sym015 and emibetuzumab analogue treatment on tumor growth in the MET-dependent CDX model EBC-1. Primary treatment period denotes initial treatment with Sym015 (green) or the emibetuzumab analogue (red), administered three times weekly at 50 mg/kg for a total of 10 doses. Secondary treatment period denotes re-treatment of mice carrying emibetuzumab analogue-resistant tumors with Sym015 (blue) three times weekly at 50 mg/kg. Data are presented as mean tumor volume (n ¼ 10 mice per group) SEM for each time point. B, Efficacy of the emibetuzumab analogue (red) administered at 50 mg/kg and dose titration of Sym015 (green) administered at different doses in the high MET-expressing lung cancer PDX model LXFA0526. Data are presented as mean tumor volume (n ¼ 7 mice per group) SEM for each time point. C, Representative images of IHC staining of MET, pMET (phosphorylated at Y1234/Y1235), and Ki67 protein in vehicle control-, Sym015-, and emibetuzumab analogue-treated EBC-1 tumor sections. D, Scoring of MET, pMET, and Ki67 IHC staining in Sym015-, emibetuzumab analogue-, and vehicle control- treated tumor sections (n ¼ 3 tumors analyzed per group). Black dots, individual tumors; horizontal lines, mean scores. E, Quantification of MET and pMET (phosphorylated at Y1234/Y1235 or Y1349) in tumor lysates by Simple Western analysis (n ¼ 5–8 tumors analyzed per group). Signal intensity was normalized to the corresponding actin signal as loading control and is presented as percentage of the signal in vehicle control-treated tumors. Experiments were performed in triplicate, and data are presented as the mean SEM. Statistical significance of differences between the treatment and control groups was calculated: , P < 0.001; , P < 0.01; ns, not significant. F and G, Effect of Sym015 and emibetuzumab analogue treatment on tumor growth in the PDX and CDX models LU2503 (F, n ¼ 5 mice per group) and Hs746T (G, n ¼ 8 mice per group), which harbor MET exon 14 deletions. Primary treatment period denotes initial treatment with Sym015 (green) or the emibetuzumab analogue (red), dosed three times weekly at 50 mg/kg for a total of 10 doses. The dotted line in G denotes single dosing of mice carrying emibetuzumab analogue-resistant tumors with Sym015. Data are presented as mean tumor volume SEM for each time point. H, Dose–response curves showing the effect of emibetuzumab analogue (red) and Sym015 (green) treatment on metabolic activity and ADCC in the cell lines EBC-1 and Hs746T. Data, mean SEM.

www.aacrjournals.org Clin Cancer Res; 23(19) October 1, 2017 5929

Poulsen et al.

A B HCC827 125 Control mAb 200 MET pMET EGFR pEGFR 100 DMSO Erlotinib ** 75 Cetuximab 150 *** Emibetuzumab analogue 50 Sym015

25 100 Metabolic activity % of untreated control 0 0.01 0.1 1 10 100 1,000 Drug concentration (nmol/L) % of HCC827 50 Signal intensity *** HCC827R1_Cet#3 ** 125 0 100

HCC827 HCC827 HCC827 HCC827 50

25 Metabolic activity % of untreated control 0 HCC827R1_cet#3 HCC827R1_cet#3 HCC827R1_cet#3 HCC827R1_cet#3 0.01 0.1 1 10 100 1,000 Drug concentration (nmol/L) G719A C LU0858, EGFRL858R D LU1901, EGFR 2,500 2,500 Treatment period Vehicle Treatment period Sym015 Emibetuzumab analogue

2,000 2,000 ) ) 3 3

1,500 1,500

1,000 1,000 Tumor volume (mm volume Tumor Tumor volume (mm

500 500

0 0 0 5 10 15 20 25 30 35 40 45 0 5 10 15 20 25 Days after first treatment Days after first treatment

Figure 3. Sym015 is effective in the anti-EGFR–resistant setting. A, Comparison of total MET, phosphorylated MET (pMET, phosphorylated at Y1234/Y1235), and phosphorylated EGFR (pEGFR, phosphorylated at Y1068) protein levels in the HCC827 and HCC827R1_Cet#3 cell lines by Simple Western analysis. Experiments were performed in triplicate, and data are presented as mean signal intensity (percentage of the HCC827 signal) SEM. Statistical signiﬁcance of differences between the treatment and control groups: , P < 0.001; , P < 0.01. B, Dose–response curves for HCC827 and HCC827ER1_CET#3 cellular metabolic activity upon treatment with negative control antibody, DMSO negative control, erlotinib, cetuximab, emibetuzumab analogue, or Sym015. Data are presented as the mean SEM. C and D, Effect of Sym015 and emibetuzumab analogue treatment on tumor growth in the MET-dependent patient-derived models LU0858 (C) and LU1901 (D), which harbor the EGFR tyrosine kinase–activating mutations L858R and G719A, respectively. Primary treatment period denotes initial treatment with Sym015 (green) or the emibetuzumab analogue (red), injected three times weekly for a total of 10 doses. Data are presented as mean tumor volume (n ¼ 5 mice/group) SEM for each time point.

MET gene amplification (defined as a MET/CEP7 ratio >2by frequencies above 10% for certain tumors, albeit based on a FISH) was detected in 29 of 1,115 patients with advanced solid limited number of patients (38). In other studies, MET gene tumor malignancies (2.6%; ref. 38). Tumor types for which MET amplification has been shown to occur in 3% to 4% of NSCLCs gene amplification was detected include adrenocortical, renal cell, (32, 39–41) and in 4% to 6% of gastric cancers (22, 42, 43), as gastroesophageal, breast, ovarian, colorectal, biliary, bladder, determined by FISH. A number of studies also demonstrated salivary gland, and lung carcinomas, as well as malignant mela- some degree of correlation between MET expression evaluated by noma. This recent study also reported MET gene amplification IHC and MET amplification evaluated by FISH (32, 44).

5930 Clin Cancer Res; 23(19) October 1, 2017 Clinical Cancer Research

Sym015 Inhibits MET-Ampliﬁed Tumors

A 25 25 EBC-1 EBC-1 SNU-5 SNU-5 H1573 20 20 H1573

15 15 Hs746T MKN-45 Hs746T MKN-45 10 10 OE33 OE33

5 5 HCC827 HCC827 MET Gene copy-number 0 MET Gene copy-number 0 0 0 Figure 4. 20,000 40,000 60,000 80,000 50,000 High MET-expressing Sym015-sensitive 100,000 100,000 150,000 200,000 models harbor high copy-number MET MET Protein signal pMET Protein signal gene ampliﬁcation. A, Correlation of MET protein signal quantiﬁed by Simple B Cell line Western and MET gene copy-number 50 (when available, obtained from ref. 31) PDX Model for the cell lines listed in Fig. 1A. High 40 MET-expressing cell lines sensitive to Sym015 are highlighted in green. Gene 30 copy-number information was not available for the H1993 cell line. B, MET copy-number analysis by quantitative 20 PCR in cell lines and patient-derived models. Data are presented as the mean 10 SD, and samples are ranked according to MET copy-number. C, FISH MET Gene copy-number 0 of MET (red) and the chromosome 7 1 3 II 6 5 3 3 5 7 58 centromere (CEP7, green) in tumor 99 U-5 E3 190 08 O I0801 sections from models LI0801, LU2503, U H1 A052 U SN EBC-1 _cet#3 L L Okajima KATO- Hs746TF MKN-4 LU250 L 1 LU642 HCC82 LU0858, and OE33. R LX 7

C HCC82 MET/CEP7

LI0801 LU2503 OE33 (Nonampliﬁed control)

Recent studies investigated the interplay between MET depen- As mentioned previously, a number of antibodies targeting dency, MET gene amplification status, and other oncogenic dri- MET are currently being evaluated in clinical trials (13, 14, 16–22), vers. In lung cancer, KRAS and NRAS mutations were reported to but none of these agents are commercially available. Here, we confer resistance to MET-targeting TKIs in a preclinical study (45). chose to compare Sym015 with a cloned analogue of emibetu- Oncogenic driver mutation profiling of the cell line models used zumab, one of the most clinically advanced MET-targeting ther- in the current study demonstrated a lack of other oncogenic apeutic antibodies. Sym015 and the emibetuzumab analogue alterations, such as RAS, RAF, MYC, and PIK3CA mutations, in demonstrated similar antiproliferative efficacies in most cell lines the MET amplified and high MET-expressing models that were tested (Figs. 1B and 2H), while Sym015 was superior in most of responsive to Sym015 (Supplementary Fig. S4). Furthermore, a the in vivo models (Figs. 2 and 3). In dose–titration studies, recently published analysis of MET gene amplification and coex- Sym015 was active at therapeutically relevant dose levels of 5 to istence of mutations/alterations in other oncogenic drivers, 10 mg/kg in three highly MET gene-amplified xenograft models including EGFR, KRAS, NRAS, and PIK3CA, showed that tumors (Fig. 2B; Supplementary Fig. S1, and data not shown), and 5 mg/ with a MET/CEP7 ratio 5 never exhibited oncogenic overlap. kg Sym015 was as effective as 50 mg/kg emibetuzumab analogue. Although rare, all such patients treated with the MET-targeting TKI This superiority of Sym015 over the emibetuzumab analogue crizotinib demonstrated high response rates. Furthermore, a clin- could in part be due to secondary effector functions, such as ical response to crizotinib was observed in cases with MET/CEP7 ADCC and complement-dependent cytotoxicity (CDC), which ratio 1.8 and no overlapping oncogenic driver alterations (46). are well-known modes of action for antibody therapeutics. The

www.aacrjournals.org Clin Cancer Res; 23(19) October 1, 2017 5931

Poulsen et al.

1,000 A Sym015 Hu9006 Hu9338 100

10 T T Sym015 ½: 11 Days Sym015 ½: 6 Days

Serum concentration (µg/mL) concentration Serum 0 10203040Figure 5. Time (days) Pharmacokinetics and pharmacodynamics of Sym015. A, Mean 10,000 B 100 mg/kg (SD) serum concentration–time 50 mg/kg proﬁles for Sym015 and the individual 15 mg/kg antibodies (Hu9006 and Hu9338) constituting Sym015. Serum levels of Sym015 and the individual antibodies 1,000 were measured at the indicated time

(µg/mL) points following a single infusion of Sym015 at 15 mg/kg in male (n ¼ 2) and female monkeys (n ¼ 2). The range of elimination half-lives of Sym015 was Serum concentration 100 calculated in individual animals using the 1357 1357linear slope between 3 and 15 days Time after first infusion Time after fourth infusion postdosing and between 15 and 36 days (days) (days) postdosing (represented by the vertical dotted lines). B, Mean (SD) serum – ﬁ C MET Ki67 concentration time pro les after the ﬁrst and fourth once-weekly doses of Control Sym015 Control Sym015 Sym015 at the indicated dose levels in male and female monkeys (n ¼ 3or5per sex and dose group). For all dose levels, the mean accumulation ratio for the

Liver maximum serum concentration (Cmax) ranged from 1.5 to 2.1, and the mean accumulation ratio for AUC for the dosing interval (AUCt) ranged from 2.0 to 2.6. C, Representative IHC staining of MET and Ki67 proteins in the indicated Kidney tissues from monkeys given four once- weekly infusions of vehicle control or Sym015 at 15 mg/kg. D, Percentage cell conﬂuence relative to treatment start after primary hepatocytes were treated for 116 hours with Sym015 (50 mg/mL), Skin HGF (200 ng/mL), or the combination as indicated. Data, mean SEM. Data are representative of three independent experiments. D 250 ** 200

150

100 Confluency 50

% of treatment start 0

d te a l Ab HGF e o Sym015 15 + HGF Untr contr 0 eg. N Sym

antibodies comprising Sym015 are of the IgG1 isotype and are to generate ADCC/CDC–silenced variants reduces antitumor efﬁ- fully capable of inducing effector functions, whereas the IgG4 cacy in vivo, further supporting the contribution of ADCC to backbone of emibetuzumab does not (Fig. 2H). Furthermore, Sym015 antitumor activity (ref. 25, Grandal and colleagues, mutation of the Fc portions of the Sym015 constituent antibodies manuscript submitted). Antibody mixtures directed against

5932 Clin Cancer Res; 23(19) October 1, 2017 Clinical Cancer Research

Sym015 Inhibits MET-Ampliﬁed Tumors

various RTK targets have demonstrated increased ADCC and CDC dose proportionality was observed for once-weekly dosing at 15 activity over mAbs, possibly due to the nonoverlapping binding of to 100 mg/kg, indicating that this regimen is sufficient to saturate the multiple antibodies resulting in a higher density of antibodies the target-mediated disposition mechanism. In humans, the pre- on the tumor cells (24, 29, 47). dicted t1/2 above the target saturation limit is in the range of 14 to Sym015 also demonstrated efficacy in models harboring MET 24 days (Supplementary Table S3), which indicates that the t1/2 exon 14 skipping alterations (Fig. 2F and G). It could be specu- may approach the 21 days seen for endogenous IgG (34). Hence, lated that ADCC also contributes to Sym015 efficacy in these sustained clinical exposure to Sym015 might be achieved with models, due to slowed internalization of MET exposing the dosing every 2 to 3 weeks. receptor to increased binding of Sym015. Patients with tumors In summary, the hypothesis generating preclinical results pre- harboring exon 14 skipping alterations constitute a small subset sented here indicates that Sym015 is safe and highly efficacious in accounting for 1% to 3% of lung cancers (48, 49), and recent targeting MET-amplified tumors and thus supports efforts to studies point to MET exon 14 deletion as a driver independent of clinically validate MET amplification as a predictive biomarker MET amplification (50, 51). A tumor-inhibiting effect of Sym015 for response in the ongoing trial of Sym015 in patients with was observed in a nonamplified lung cancer model with MET advanced MET-amplified solid tumors (ClinicalTrials.gov identi- exon 14 deletion (Supplementary Fig. S3). Although limited, fier NCT02648724). these data suggest that patients harboring MET exon 14 skipping MET fi fi alterations without ampli cation could bene t from treat- Disclosure of Potential Conflicts of Interest ment with Sym015. I.D. Horak is the chief scientificofficer/chief medical officer at Symphogen Some years ago, an antibody mixture targeting MET (denoted A/S. No potential conflicts of interest were disclosed by the other authors. R13/R28) was reported to demonstrate preclinical efficacy in colorectal cancer models with high MET expression in vivo Authors' Contributions (52). However, R13/R28 has not progressed to clinical trials. A Conception and design: T.T. Poulsen, M.M. Grandal, C. Frohlich,€ T. Bouquin, head-to-head comparison between Sym015 and an analogue of M.W. Pedersen, I.D. Horak, J. Lantto, M. Kragh MET fi in R13/R28 in -ampli ed cells demonstrated similar or better Development of methodology: T.T. Poulsen, M.M. Grandal, R. Hald, vitro efficacy of Sym015 (data not shown). T. Lindsted, A. Dahlman, H.J. Jacobsen, M.W. Pedersen Sym015 contains a balanced ratio of its two constituent anti- Acquisition of data (provided animals, acquired and managed patients, bodies, but different ratios of the single antibodies were also provided facilities, etc.): T.T. Poulsen, M.M. Grandal, K. Koefoed, T. Lindsted, € evaluated in vivo to test whether a nonbalanced ratio of the C. Frohlich, S.E. Pollmann, K.W. Eriksen, A. Dahlman, T. Bouquin, fi M.W. Pedersen, M. Kragh antibody constituents would be more ef cacious. Interestingly, Analysis and interpretation of data (e.g., statistical analysis, biostatistics, fi signi cantly skewed ratios of the two antibodies yielded antitu- computational analysis): T.T. Poulsen, M.M. Grandal, N.J.Ø. Skartved, R. Hald, mor efficacy comparable with that of Sym015 (Supplementary L. Alifrangis, K. Koefoed, T. Lindsted, C. Frohlich,€ S.E. Pollmann, K.W. Eriksen, Fig. S2), and in combination with similar pharmacokinetic pro- A. Dahlman, T. Bouquin, M.W. Pedersen files of the two individual Sym015 antibody constituents Writing, review, and/or revision of the manuscript: T.T. Poulsen, (Fig. 5A), the results support the use of a balanced antibody M.M. Grandal, R. Hald, L. Alifrangis, K. Koefoed, T. Lindsted, S.E. Pollmann, fi K.W. Eriksen, A. Dahlman, T. Bouquin, M.W. Pedersen, I.D. Horak, J. Lantto, mixture. Of note, this robustness in ef cacy at skewed antibody M. Kragh fi fi ratios con rms previous ndings for the EGFR-targeting antibody Administrative, technical, or material support (i.e., reporting or organizing mixture Sym004 (24). data, constructing databases): T.T. Poulsen The repeat-dose toxicology study in monkeys showed that Study supervision: T.T. Poulsen, H.J. Jacobsen, M.W. Pedersen, I.D. Horak, once-weekly administration of Sym015 was well tolerated, with J. Lantto, M. Kragh a NOAEL of 100 mg/kg, the highest dose tested. Tolerability and Other (toxicology): N.J.Ø. Skartved absence of significant toxicity when targeting MET in monkeys have also been reported for emibetuzumab (21). The biphasic Acknowledgments pharmacokinetic profile in cynomolgus monkeys indicates the The authors thank Stephen Gilliver (TFSCRO, Sweden) for proofreading the presence of a saturable elimination pathway. As seen for mAbs manuscript. The costs of publication of this article were defrayed in part by the payment of that bind to membrane-bound targets, this is most likely due page charges. This article must therefore be hereby marked advertisement in to target-mediated disposition (34) of the Sym015 antibodies accordance with 18 U.S.C. Section 1734 solely to indicate this fact. following binding to MET. For serum concentrations above t the saturation limit, a 1/2 of 11 days (Fig. 5A; Supplementary Received March 20, 2017; revised June 1, 2017; accepted June 29, 2017; Table S3) was observed in monkeys. In the 4-week toxicity study, published OnlineFirst July 5, 2017.

References 1. Van Andel Institute. HGF/SF - MET and cancer online table. Available from: 5. Kong-Beltran M, Seshagiri S, Zha J, Zhu W, Bhawe K, Mendoza N, et al. https://resources.vai.org/Met/Index.aspx. Somatic mutations lead to an oncogenic deletion of Met in lung cancer. 2. Gherardi E, Birchmeier W, Birchmeier C, Vande Woude G. Targeting MET in Cancer Res 2006;66:283–9. cancer: rationale and progress. Nat Rev Cancer 2012;12:89–103. 6. Lee J, Ou S-HI, Lee JM, Kim HC, Hong M, Kim SY, et al. Gastrointestinal 3. Kawakami H, Okamoto I, Arao T, Okamoto W, Matsumoto K, Taniguchi H, malignancies harbor actionable MET exon 14 deletions. Oncotarget et al. MET amplification as a potential therapeutic target in gastric cancer. 2015;6:28211–22. Oncotarget 2013;4:9–17. 7. Engelman JA, Zejnullahu K, Mitsudomi T, Song Y, Hyland C, Park JO, et al. 4. Dimou A, Non L, Chae YK, Tester WJ, Syrigos KN. MET gene copy number MET amplification leads to gefitinib resistance in lung cancer by activating predicts worse overall survival in patients with non-small cell lung cancer ERBB3 signaling. Science 2007;316:1039–43. (NSCLC); a systematic review and meta-analysis. PLoS One 2014;9: 8. Bean J, Brennan C, Shih J, Riely G, Viale A, Wang L, et al. MET amplification e107677. occurs with or without T790M mutations in EGFR mutant lung tumors

www.aacrjournals.org Clin Cancer Res; 23(19) October 1, 2017 5933

Poulsen et al.

with acquired resistance to gefitinib or erlotinib. Proc Natl Acad Sci U S A 27. Meijer PJ, Andersen PS, Haahr Hansen M, Steinaa L, Jensen A, Lantto J, et al. 2007;104:20932–7. Isolation of human antibody repertoires with preservation of the natural 9. Ariyawutyakorn W, Saichaemchan S, Varella-Garcia M. Understanding and heavy and light chain pairing. J Mol Biol 2006;358:764–72. Targeting MET Signaling in Solid Tumors - Are We There Yet? J Cancer 28. Davies J, Liu L, Lu J, Vaillancourt PE, Wortinger MA, Zeng W, inventors; Eli 2016;7:633–49. Lilly And Company, assignee. c-MET antibodies. Patent WO2010059654 10. Pupo E, Ducano N, Lupo B, Vigna E, Avanzato D, Perera T, et al. Rebound A12010 May 27. effects caused by withdrawal of MET kinase inhibitor are quenched by a 29. Jacobsen HJ, Poulsen TT, Dahlman A, Kjaer I, Koefoed K, Sen JW, et al. Pan- MET therapeutic antibody. Cancer Res 2016;76:5019–29. HER, an antibody mixture simultaneously targeting EGFR, HER2 and 11. Bertotti A, Burbridge MF, Gastaldi S, Galimi F, Torti D, Medico E, et al. Only HER3 effectively overcomes tumor heterogeneity and plasticity. Clin a subset of Met-activated pathways are required to sustain oncogene Cancer Res 2015;21:4110–22. addiction. Sci Signal 2009;2:er11. 30. Bardelli A, Corso S, Bertotti A, Hobor S, Valtorta E, Siravegna G, et al. 12. Catenacci DVT, Henderson L, Xiao S-Y, Patel P, Yauch RL, Hegde P, et al. Amplification of the MET receptor drives resistance to anti-EGFR therapies Durable complete response of metastatic gastric cancer with anti-Met in colorectal cancer. Cancer Discov 2013;3:658–73. therapy followed by resistance at recurrence. Cancer Discov 2011;1: 31. Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S, 573–9. et al. The Cancer Cell Line Encyclopedia enables predictive modelling of 13. Jin H, Yang R, Zheng Z, Romero M, Ross J, Bou-Reslan H, et al. MetMAb, the anticancer drug sensitivity. Nature 2012;483:603–7. one-armed 5D5 anti-c-Met antibody, inhibits orthotopic pancreatic tumor 32. Schildhaus HU, Schultheis AM, Ruuschoff€ J, Binot E, Merkelbach-Bruse growth and improves survival. Cancer Res 2008;68:4360–8. S, Fassunke J, et al. Met amplification status in therapy-na€ve adeno- 14. Spigel DR, Ervin TJ, Ramlau RA, Daniel DB, Goldschmidt JH, Blu- And squamous cell carcinomas of the lung. Clin Cancer Res 2015;21: menschein GR, et al. Randomized phase II trial of onartuzumab in 907–15. combination with erlotinib in patients with advanced non-small-cell lung 33. Wang J, Iyer S, Fielder PJ, Davis JD, Deng R. Projecting human pharma- cancer. J Clin Oncol 2013;31:4105–14. cokinetics of monoclonal antibodies from nonclinical data: comparative 15. Yoh K, Doi T, Ohmatsu H, Kojima T, Takahashi H, Zenke Y, et al. A phase I evaluation of prediction approaches in early drug development. Biopharm dose-escalation study of LY2875358, a bivalent MET antibody, given as Drug Dispos 2016;37:51–65. monotherapy or in combination with erlotinib or gefitinib in Japanese 34. Dirks NL, Meibohm B. Population pharmacokinetics of therapeutic mono- patients with advanced malignancies. Invest New Drugs 2016;34:584–95. clonal antibodies. Clin Pharmacokinet 2010;49:633–59. 16. Spigel DR, Edelman MJ, O'Byrne K, Paz-Ares L, Shames DS, Yu W, et al. 35. Prat M, Crepaldi T, Pennacchietti S, Bussolino F, Comoglio PM. Agonistic Onartuzumab plus erlotinib versus erlotinib in previously treated stage IIIb monoclonal antibodies against the Met receptor dissect the biological or IV NSCLC: Results from the pivotal phase III randomized, multicenter, responses to HGF. J Cell Sci 1998;111:237–47. placebo-controlled METLung (OAM4971g) global trial. J Clin Oncol 36. Xu L, Nilsson MB, Saintigny P, Cascone T, Herynk MH, Du Z, et al. 2014;32:8000. Epidermal growth factor receptor regulates MET levels and invasiveness 17. Wang J, Goetsch L, Tucker L, Zhang Q, Gonzalez A, Vaidya KS, et al. Anti-c- through hypoxia-inducible factor-1alpha in non-small cell lung cancer Met monoclonal antibody ABT-700 breaks oncogene addiction in tumors cells. Oncogene 2010;29:2616–27. with MET amplification. BMC Cancer 2016;16:105. 37. Kawakami H, Okamoto I, Okamoto W, Tanizaki J, Nakagawa K, Nishio K. 18. Strickler JH, LoRusso P, Yen C-J, Lin C-C, Kang Y-K, Kaminker P, et al. Phase Targeting MET amplification as a new oncogenic driver. Cancers 2014;6: 1, open-label, dose-escalation, and expansion study of ABT-700, an anti-C- 1540–52. met antibody, in patients (pts) with advanced solid tumors. J Clin Oncol 38. Jardim DLF, Tang C, Gagliato DDM, Falchook GS, Hess K, Janku F, et al. 2014;32:2507. Analysis of 1,115 patients tested for MET amplifi cation and therapy 19. Hultberg A, Morello V, Huyghe L, De Jonge N, Blanchetot C, Hanssens V, response in the MD anderson phase I clinic. Clin Cancer Res 2014;20: et al. Depleting MET-expressing tumor cells by ADCC provides a thera- 6336–45. peutic advantage over inhibiting HGF/MET signaling. Cancer Res 39. Cappuzzo F, Marchetti A, Skokan M, Rossi E, Gajapathy S, Felicioni L, et al. 2015;75:3373–83. Increased MET gene copy number negatively affects survival of surgically 20. Aftimos PG, Barthelemy P, Rolfo CD, Hanssens V, De Jonge N, Silence K, resected non-small-cell lung cancer patients. J Clin Oncol 2009;27: et al. A phase I, first-in-human study of argx-111, a monoclonal antibody 1667–74. targeting c-met in patients with solid tumors. J Clin Oncol 2015;33:2580. 40. Go H, Jeon YK, Park HJ, Sung S-W, Seo J-W, Chung DH. High MET gene 21. Liu L, Zeng W, Wortinger MA, Yan SB, Cornwell PD, Peek VL, et al. copy number leads to shorter survival in patients with non-small cell lung LY2875358, a neutralizing and internalizing anti-MET bivalent antibody, cancer. J Thorac Oncol 2010;5:305–13. inhibits HGF-dependent and HGF-independent MET activation and tumor 41. Okamoto I, Sakai K, Morita S, Yoshioka H, Kaneda H, Takeda K, et al. growth. Clin Cancer Res 2014;20:6059–70. Multiplex genomic profiling of non-small cell lung cancers from the LETS 22. Goldman JW, Rosen LS, Algazi AP, Turner PK, Wacheck V, Tuttle J, et al. phase III trial of first-line S-1/carboplatin versus paclitaxel/carboplatin: First-in-human dose escalation study of LY2875358 (LY), a bivalent MET results of a West Japan Oncology Group study. Oncotarget 2014;5: antibody, as monotherapy and in combination with erlotinib (E) in 2293–304. patients with advanced cancer. J Clin Oncol 2013;31:8093. 42. Catenacci DVT, Tang R, Oliner KS, Ang A, Loberg RD, O'Day E, et al. MET as 23. Pedersen MW, Jacobsen HJ, Koefoed K, Dahlman A, Kjær I, Poulsen TT, a prognostic biomarker of survival in a large cohort of patients with et al. Targeting three distinct HER2 domains with a recombinant antibody gastroesophageal cancer (GEC). ASCO Meet Abstr 2015;33:4034. mixture overcomes trastuzumab resistance. Mol Cancer Ther 2015;14: 43. Xu R, Qiu M, Zhou Y, Wang D-S, Zhang D, Wang F, et al. Evaluation 670–80. of tumor MET protein expression, MET gene amplification, and 24. Pedersen MW, Jacobsen HJ, Koefoed K, Hey A, Pyke C, Haurum JS, et al. HER2 expression in Chinese patients with advanced gastric or Sym004: a novel synergistic anti-epidermal growth factor receptor anti- gastroesophageal junction (G/GEJ) cancer. ASCO Meet Abstr 2015;33: body mixture with superior anticancer efficacy. Cancer Res 2010;70: 4048. 588–97. 44. Gavine PR, Ren Y, Han L, Lv J, Fan S, Zhang W, et al. Volitinib, a potent and 25. Grandal MM, Poulsen TT, Kofoed K, Eriksen KW, Dahlman A, Conrotto P, highly selective c-Met inhibitor, effectively blocks c-Met signaling and et al. A novel synergistic antibody pair targeting non-overlapping epitopes growth in c-MET amplified gastric cancer patient-derived tumor xenograft of MET effectively inhibits MET-driven cancer models. In: Proceedings of models. Mol Oncol 2015;9:323–33. the 107th Annual Meeting of the American Association for Cancer 45. Leiser D, Medova M, Mikami K, Nisa L, Stroka D, Blaukat A, et al. KRAS and Research; 2016 Apr 16–20; New Orleans, LA. Philadelphia, PA: AACR; HRAS mutations confer resistance to MET targeting in preclinical models of 2016. Abstract nr 1218. MET-expressing tumor cells. Mol Oncol 2015;9:1434–46. 26. Ohashi K, Sequist LV, Arcila ME, Moran T, Chmielecki J, Lin Y-L, et al. Lung 46. Noonan SA, Berry L, Lu X, Gao D, Baron AE, Chesnut P, et al. Identifying the cancers with acquired resistance to EGFR inhibitors occasionally harbor appropriate FISH criteria for defining MET copy number driven lung BRAF gene mutations but lack mutations in KRAS, NRAS, or MEK1. Proc adenocarcinoma through oncogene overlap analysis. J Thorac Oncol Natl Acad Sci U S A 2012;109:E2127–33. 2016;11:1293–304.

5934 Clin Cancer Res; 23(19) October 1, 2017 Clinical Cancer Research

Sym015 Inhibits MET-Ampliﬁed Tumors

47. Toth G, Szoo€r A, Simon L, Yarden Y, Szollo€ si J, Vereb G. The combination 50. Paik PK, Drilon A, Fan P-D, Yu H, Rekhtman N, Ginsberg MS, et al. of trastuzumab and pertuzumab administered at approved doses may Response to MET inhibitors in patients with stage IV lung adenocarcinomas delay development of trastuzumab resistance by additively enhancing harboring MET mutations causing exon 14 skipping. Cancer Discov – antibody-dependent cell-mediated cytotoxicity. MAbs 2016;8:1361–70. 2015;5:842 9. 48. Schrock AB, Frampton GM, Suh J, Chalmers ZR, Rosenzweig M, Erlich RL, 51. Drilon A, Cappuzzo F, Ou S-HI, Camidge DR. Targeting MET in Lung – et al. Characterization of 298 lung cancer patients harboring MET exon 14 Cancer: Will Expectations Finally Be MET? J Thorac Oncol 2017;12:15 26. skipping (METex14) alterations. J Thorac Oncol 2016;11:1493–502. 52. van der Horst EH, Chinn L, Wang M, Velilla T, Tran H, Madrona Y, et al. 49. Zheng D, Wang R, Ye T, Yu S, Hu H, Shen X, et al. MET exon 14 skipping Discovery of fully human anti-MET monoclonal antibodies with antitu- deﬁnes a unique molecular class of non-small cell lung cancer. Oncotarget mor activity against colon cancer tumor models in vivo. Neoplasia – 2016;7:41691–702. 2009;11:355 64.

www.aacrjournals.org Clin Cancer Res; 23(19) October 1, 2017 5935

Sym015: A Highly Efficacious Antibody Mixture against MET -Amplified Tumors

Thomas Tuxen Poulsen, Michael Monrad Grandal, Niels Jørgen Østergaard Skartved, et al.

Clin Cancer Res 2017;23:5923-5935. Published OnlineFirst July 5, 2017.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-17-0782

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2017/07/04/1078-0432.CCR-17-0782.DC1

Cited articles This article cites 49 articles, 19 of which you can access for free at: http://clincancerres.aacrjournals.org/content/23/19/5923.full#ref-list-1

Citing articles This article has been cited by 4 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/23/19/5923.full#related-urls

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 Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/23/19/5923. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.