Safety and Activity of the First-in-Class Sym004 Anti-EGFR Antibody Mixture in Patients with Refractory Colorectal Cancer

Metastatic colorectal cancer is expected to cause 50,000 deaths in the United States alone in 2015 (1). Multiple clinical trials have confirmed the benefit from anti-EGFR therapy in patients with wild-type–KRAS metastatic colorectal cancer (2–6), recently refined to the wild-type RAS (i.e., wild-type KRAS and NRAS) population (4). Preclinical findings point to conservation of EGFR dependency of tumors that progressed upon cetuximab or panitumumab therapy (7), even in the case of acquired mutations in RAS genes (8, 9), with EGFR ligands playing a central role in autocrine and paracrine resistance networks (10–12). Other mechanisms of acquired cetuximab resistance include mutations in the extracellular domain of EGFR and EGFR interactions with the receptor kinases HER3 (ERBB3) or MET (10, 13–16). These findings raise the question of whether more effective EGFR targeting, e.g., more potent receptor downmodulation, could overcome resistance to available anti-EGFR therapy (17).

Sym004 is a 1:1 mixture of two recombinant, human–mouse chimeric mAbs directed against nonoverlapping EGFR epitopes (mAb992 and mAb1024). A unique feature of Sym004 is its ability to mediate rapid EGFR internalization and subsequent degradation of internalized receptors via EGFR cross-linking (17, 18). Preclinical studies with Sym004 showed both superior induction of tumor regression as compared with other EGFR-targeting antibodies and antitumor activity in models of acquired cetuximab resistance (7, 17). Toxicology studies in cynomolgus monkeys showed a toxicity profile consistent with other drugs of the same class, including dermatologic and gastrointestinal events (19).

In this article, we first present novel preclinical experiments demonstrating enhanced activity of Sym004 over cetuximab in the presence of various EGFR ligands, providing a rationale for evaluating the clinical activity of Sym004 in an anti-EGFR–refractory metastatic colorectal cancer setting. Next, we present the safety, tolerability, and pharmacokinetic profile of escalating doses of Sym004 from the first-in-human phase I trial in patients with advanced epithelial tumors. Finally, we describe the pharmacodynamic effects and clinical activity of Sym004 from two dose-expansion cohorts in patients with wild-type–KRAS metastatic colorectal cancer and acquired resistance to approved anti-EGFR antibody treatment, supporting Sym004′s proof of mechanism and proof of concept.

A high-throughput cell line screen was performed to determine the activity of Sym004 in the presence of the EGFR ligands EGF, TGFα, and epiregulin (EREG) in a panel of 22 colorectal cancer cell lines (Fig. 1A and B). Approximately 60% of the tested cell lines (13/22) were sensitive to antibody targeting (defined by an activity area above 2). No clear differentiation of Sym004 and cetuximab was seen in the absence of ligands. In addition, no correlation between KRAS or BRAF mutational status and response to antibody targeting was observed, although the two cell lines harboring the BRAF^V600E mutation (HT29 and COLO205) were completely resistant. We did observe that the cells stimulated by EGFR ligands were those that also responded to EGFR inhibition, supporting the notion that EGFR dependency in colorectal cancer is largely driven by the various EGFR ligands.

Knowing that high-affinity EGFR ligands confer resistance to cetuximab in preclinical models (11, 12), we assessed whether Sym004, with its distinguished mechanism of action, had superior activity as compared with cetuximab in the presence of various EGFR ligands. Proliferation experiments of colorectal cancer cell lines in the presence of the high-affinity ligands EGF and TGFα as well as the low-affinity ligand EREG (Fig. 1C and D) clearly demonstrated that Sym004 outperformed cetuximab in the presence of the high-affinity ligands, whereas the difference was less pronounced in the presence of the low-affinity ligand. Mechanistically, Sym004 led to more pronounced EGFR degradation than cetuximab following 24 hours of treatment, with increased inhibition of ERK1/2 and S6R phosphorylation following ligand stimulation (Fig. 1E).

Our preclinical results thus provide evidence that Sym004 could overcome acquired resistance to cetuximab mediated by EGFR ligands and a rationale for clinical evaluation of Sym004 in this setting.

Between March 2010 and April 2012, a total of 62 patients were enrolled in the trial. Of these, 20 patients with solid epithelial tumors received escalating doses of Sym004, ranging from 0.4 to 12 mg/kg administered weekly. The remaining 42 patients had metastatic colorectal cancer and received either 9 or 12 mg/kg weekly in the dose-expansion phase and were evaluated for pharmacodynamic effects and response (Table 1). All 62 patients were assessed for toxicity.

Prior treatment with anti-EGFR antibodies was required for eligibility only in the dose-expansion cohorts. Fourteen patients had more than 1 line of anti-EGFR mAb treatment (Table 1), and the majority (90%) had at least 1 line of anti-EGFR mAb treatment in combination with chemotherapy. All patients had a documented response to previous anti-EGFR mAb treatment followed by disease progression, with a median time until administration of the first dose of Sym004 of 54 and 33 days for patients treated at 9 and 12 mg/kg, respectively. Of note, the median time to disease progression after termination of the latest anti-EGFR treatment was 8 days, and the median time from the last anti-EGFR mAb treatment until the first dose of Sym004 was 81 days.

A formal MTD was not reached, and the maximum administered dose, declared in a joint decision of the investigators, sponsor, and the independent data-monitoring committee, was set at 12 mg/kg. The decision to stop dose escalation was based on the observation of accumulating skin and serum electrolyte (hypomagnesemia) toxicities observed beyond the 4-week observation period for dose-limiting toxicities (DLT).

Adverse events of any grade were reported in 61 patients (98%; Table 2). Serious adverse events that investigators considered to be related to treatment occurred in 11 of 36 patients (31%; Supplementary Table S1). The most common drug-related adverse events of any grade were skin rash (69%), dry skin (45%), hypomagnesemia (53%), pruritus (39%), mucosal inflammation (31%), and diarrhea (27%; Supplementary Table S1).

Skin toxicity and hypomagnesemia events of grade 3 or higher were reported for 50% and 21% of the patients, respectively, whereas only 3% of the patients experienced diarrhea of grade 3 or higher (Table 2). Skin-related adverse events and hypomagnesemia were mainly controlled by supportive care, dose delays, and reductions as per protocol. Drug-related adverse events (rash, acneiform dermatitis, and infusion-related reaction) led to discontinuation in 3 patients in the dose-escalation part. In the dose-expansion cohorts, the predominant reason to discontinue was disease progression (88%), the median number of Sym004 infusions was 10 (range, 2–47), and the median relative dose intensity was 71% (Table 2). Treatment was delayed at least once in 86% of the patients, and the assigned dose was reduced at least once in 43% of the patients, predominantly due to grade 3 skin adverse events and in accordance with the protocol. Both interventions were observed more often in patients treated at 12 mg/kg (Table 2).

Infusion-related reactions were observed in 5 of 62 patients (8%). We observed only a single grade 3 infusion-related reaction in a patient from the 9 mg/kg dose-escalation cohort. This event led to the expansion of an already-implemented premedication scheme with antihistamines by adding antipyretics (for the first 4 infusions) and glucocorticoids (for the first 2 infusions) for the remaining patients.

Serum mAb levels increased in a dose-dependent manner from 0.4 to 12 mg/kg in patients of the dose-escalation cohort. Doses of 3 mg/kg or less were rapidly cleared and had a terminal serum half-life of 2 days or less (Supplementary Table S2 and Supplementary Fig. S1A). The relationship between doses and geometric mean area under the curve appeared to be linear for doses of 6 mg/kg and higher (Supplementary Table S2 and Supplementary Fig. S1B). The serum half-life of the doses chosen for cohort expansion, estimated by compilation of serum elimination curves, was 3 days after the first infusion and 4 to 5 days after the fourth infusion (Supplementary Table S2). In patients of the dose-expansion cohort who received uninterrupted weekly treatment, trough levels revealed a continuous drug exposure (Supplementary Fig. S1C). A single patient treated at 9 mg/kg showed a transient anti-Sym004 antibody response of low magnitude.

The pharmacodynamic response in tumor biopsies of patients with metastatic colorectal cancer, obtained before the first and the planned fifth infusion, showed a statistically significant durable downmodulation of EGFR and decrease in Ki67 expression (Fig. 2A), providing clinical proof of mechanism of Sym004. Internal validation for this observation was obtained from skin biopsies obtained on the identical schedule (Fig. 2B). To confirm the molecular status of the metastatic colorectal cancer of patients enrolled to the study, we conducted next-generation sequencing (NGS) of baseline biopsies from metastatic lesions. In 20 samples with enough neoplastic tissue available for analysis, we found 3 mutations of KRAS, 2 of PIK3CA, and 1 of EGFR (Table 3). MET gene FISH analysis revealed 6 cases of polysomy among 20 baseline samples, but no MET amplification (Table 3).

Of 42 patients with metastatic colorectal cancer resistant to EGFR treatment enrolled in the dose-expansion cohorts, 39 had imaging data available for central radiological evaluation. Five patients (13%) achieved a partial response of target lesions (Fig. 3). Overall, 17 patients (44%) had some degree of tumor shrinkage during Sym004 therapy, of which 4 (33%) of 12 and 13 (48%) of 27 patients had received 9 and 12 mg/kg, respectively (Fig. 3). The overall disease control rate (partial response and stable disease) was 67% (58% and 70% for patients on 9 and 12 mg/kg, respectively). In addition, Fig. 3 shows treatment duration and time points of first onset of grade >2 skin rash, confirming that this toxicity occurred early, did not lead to significant treatment discontinuation, and was not indicative of predicting response. Overall, median progression-free survival (PFS) was 3.3 (95% confidence interval, 2.6–4.9) months, and individual PFS intervals are shown in Fig. 3. At the time of data cutoff, 41 of 42 patients (98%) had died due to progression of cancer.

In an exploratory analysis of potential associations between molecular profile and antitumor response, we found that all patients experiencing partial response to Sym004 therapy with individual PFS intervals of 5.0 to 6.7 months had tumors with wild-type status of KRAS, NRAS, and BRAF and lacked MET amplification (Table 3). One patient experiencing a partial response to Sym004 had a baseline EGFR^S492R mutation, predictive of resistance to cetuximab (14, 15). The 3 patients with KRAS mutations at baseline had stable disease with PFS figures of 2.6, 4.8, and 7.1 months (Table 3).

In this trial, the anti-EGFR mAb mixture Sym004 induced clinically meaningful rates of partial response (13%) or minor tumor regression (44%) of target lesions in patients with metastatic colorectal cancer with acquired resistance to anti-EGFR mAb treatment. To the best of our knowledge, this is the first trial that demonstrates antitumor activity of a mixture of two mAbs targeting nonoverlapping eptitopes of the same receptor, with encouraging results that support targeted therapy alone in a very refractory patient population. Acknowledging the limited number of patients treated with Sym004, it remains noteworthy that the response rate seen in this trial was similar to that assessed in previous phase III studies with cetuximab and panitumumab monotherapy in EGFR treatment–naïve patients with metastatic colorectal cancer (5, 6). In addition, clinical trials with cetuximab and panitumumab monotherapy have shown no clinically relevant activity in patients resistant to previous anti-EGFR mAb treatment (20, 21).

It is noteworthy that some metastatic colorectal cancer tumors appear to remain “EGFR-addicted” despite progression on anti-EGFR treatment, which is supported by our preclinical and emerging clinical data (7). The clinical antitumor activity seen with Sym004 is in line with the superior antitumor activity of Sym004 over cetuximab in preclinical animal models of acquired cetuximab resistance (7, 17). In addition, we showed here that Sym004 is highly efficient at blocking colorectal cancer cell line growth in the presence of the high-affinity ligands EGF and TGFα, factors known to be both upregulated in response to anti-EGFR antibody treatment and determinants of EGFR inhibitor resistance (13). We further hypothesize that the pharmacodynamic effects of Sym004, with sustained decrease in EGFR expression in tumor biopsies, relate to the ability of the mAbs to cross-link the EGFR, causing internalization and subsequent degradation of the antibody–receptor complexes (17). Thus, effective and broad blockade of ligand–receptor interaction together with receptor downmodulation may be responsible for the promising antitumor activity observed in this trial, a hypothesis that might also be translated to the setting of upfront anti-EGFR mAb treatment.

KRAS mutations are well-established predictors of clinical resistance to anti-EGFR treatment in metastatic colorectal cancer (4, 22), although this association is not strong in preclinical models (23–25). Our initial clinical development program focused on metastatic colorectal cancer with wild-type KRAS (determined on archived tumor samples), anticipating that Sym004 could overcome an acquired nongenetic mechanism of resistance to anti-EGFR treatment, even though we did see some activity of Sym004 in colorectal cancer cell lines with mutated KRAS in the presence of ligands. In an exploratory analysis, we investigated molecular events potentially associated with response and resistance to Sym004. The results must be viewed with caution, considering that baseline tumor biopsies and paired sets of biopsies with sufficient tissue for NGS were available from only 43% and 17% of the 42 enrolled patients, respectively. Although partial responses were enriched in patients without genetic mechanisms of EGFR inhibitor resistance (KRAS, NRAS, BRAF, MET), we did see antitumor activity in a patient with acquired EGFR^S492R mutation. The activity of Sym004 mAbs in preclinical models of acquired EGFR extracellular domain mutations will be further explored. Finally, with regard to the potential clinical benefit of Sym004 in the setting of acquired RAS mutations, it is important to emphasize that the clinical significance of these genomic events in the anti-EGFR refractory setting remains poorly understood. Nongenetic resistance mechanisms, such as upregulation of EGFR ligands in the tumor microenvironment and cross-talk between EGFR and other receptors, most likely coexist with RAS mutations (11). The effect of Sym004 on different disease subclones selected under treatment pressure needs additional investigation.

This trial has not expanded the spectrum of known adverse events of anti-EGFR drugs. Furthermore, consistent with the experience from approved anti-EGFR antibodies, Sym004-related grade 3 skin toxicity and hypomagnesemia could be managed without treatment discontinuation in both dose-expansion cohorts. The observed grade ≥3 event rate may be related to pronounced EGFR downmodulation and/or the re-exposure to anti-EGFR mAb treatment. In addition, a propensity to develop high-grade skin toxicities might be explained by the inclusion of patients who had previously had a clinical benefit from anti-EGFR mAbs, which is typically linked to an experience of high-grade skin toxicity.

In this trial, we validated the EGFR pathway as an important target for therapeutic intervention in wild-type–KRAS refractory metastatic colorectal cancer even beyond cetuximab/panitumumab. The disease control achieved with Sym004 in 26 of 39 patients in the acquired resistance setting forms a robust clinical basis to be tested in future trials. Such trials are currently under way to further investigate an optimal dose and dosing schedule in patients with metastatic colorectal cancer (NCT02083653 and NCT01117428).

Anti-EGFR mAb-sensitive (GEO, H508, SW403, LS174T, T84, SW1463, SW837, SW948, SNUC2A, CACO2, COLO678, GP5D, SW480) and mAb-resistant (LOVO, HCT116, DLD1, H716, SW620, LS1034, HT29, COLO205, HCT15) colorectal cancer cell lines were used for the experiments. The cell lines, with the exception of GEO, were obtained from the ATCC. The colon cancer cell line GEO has been described previously (26) and was a generous gift from Professor Douglas Boyd (MD Anderson Cancer Center, Houston, TX). All cell lines were Mycoplasma free, cultured according to the suppliers' recommendations, and used within 6 months of resuscitation. The cell lines were obtained in the period from 2010 to 2014. Cell lines from the ATCC are routinely tested for authenticity using short tandem repeat profiling.

A standard 4-day WST-1 viability assay (Roche Diagnostics) was used to measure growth and growth inhibition following treatment with mAbs and mAb mixtures and was performed as previously described (17). The number of viable cells was calculated as percentage of untreated control. All antibodies were stored individually at −80°C or at 4°C for shorter periods. Antibody mixtures were generated before performing experiments, mixed in ratios of 1:1 (w/w), and immediately added to experimental wells.

Cells were seeded in media supplemented with 2% FBS and allowed to adhere overnight. Cells were treated with antibodies for 24 hours and then with ligands for 15 minutes. Following ligand stimulation, cells were lysed in RIPA buffer, and immunoblot analyses were performed using 10 μg of total protein. The ligands EGF, TGFα, EREG, and heregulin were all obtained from R&D Systems. Immunoblotting was conducted according to the antibody manufacturers' recommendations. Anti-HER2 (1:1,000), anti-pHER2 (Tyr1221; 1:1,000), anti-EGFR (1:5,000), anti-pEGFR (Tyr1068; 1:1,000), anti-HER3 (1:1,000), anti-pHER3 (Tyr1289; 1:200), anti-ERK1/2 (1:1,000), anti-pERK1/2 (Thr202/Tyr204; 1:1,000), anti-pAKT (Ser473; 1:1,000), anti-AKT (1:1,000), anti-pS6R (Ser235/236; 1:1,000), and anti–β-actin (1:2,000) antibodies were from Cell Signaling Technology. Bands were visualized using IRDye 800CW Secondary Antibodies (LI-COR Biosciences).

This trial was conducted as a multicenter phase I trial with dose escalation in patients with recurrent advanced solid tumors and dose expansion in patients with metastatic colorectal cancer and acquired EGFR inhibitor resistance (Trial registration ID: NCT01117428). The trial protocol was approved by the Institutional Review Boards/ethics committees at the participating centers, and the trial was conducted in accordance with the Declaration of Helsinki. All patients provided written informed consent. The trial was designed by senior academic authors and the sponsor, Symphogen A/S. Trial medications were provided by the sponsor. R. Dienstmann, S. Braun, and J. Tabernero wrote all drafts of the article, with editorial support provided by a medical writer and funded by the sponsor.

Inclusion criteria included: age of 18 years or older; Eastern Cooperative Oncology Group performance status of 2 or better; and adequate hematologic, hepatic, and renal function. Patients in the dose-escalation cohort had to have advanced solid tumors refractory to available standard treatment. Patients in the dose-expansion cohorts had to have KRAS–wild-type metastatic colorectal cancer with documented radiological response to an anti-EGFR mAb-containing regimen. Response was defined as complete response, partial response, or stable disease for at least 16 weeks. Patients also had to have documented disease progression (verified by computerized tomography or MRI) during or within 6 months after cessation of primary anti-EGFR mAb treatment, with the last anti-EGFR dose no more than 6 months before first Sym004 infusion. Exclusion criteria for all cohorts included symptomatic central nervous system metastases and antitumor treatment within 4 weeks of first planned Sym004 infusion, or within 12 weeks for experimental vaccines.

Sym004 was administered as an intravenous infusion at a maximum rate of 10 mg/min once every week. A premedication schedule included a glucocorticoid (before the first 2 infusions), an antihistamine, and an antipyretic (before the first 4 infusions).

The starting dose of the dose-escalation cohort was derived based on a recently published toxicology study in cynomolgus monkeys (19). Patients were enrolled sequentially, and cohorts received weekly doses of 0.4, 0.75, 1.5, 3, 6, 9, or 12 mg/kg; 1 patient in the 2 first cohorts followed by 3 to 6 patients in subsequent cohorts according to the traditional 3+3 design. Dose escalation proceeded if no DLTs were observed during the first 4 weeks. DLTs were defined as treatment-related grade ≥3 hematologic and nonhematologic toxicities (adapted from NCI CTCAE, version 4.02) observed during the first 4 weeks of treatment (corresponding to a minimum of 3 and a maximum of 4 infusions) and within 48 hours of the most recent infusion (allowing at most 30 days between first and most recent infusions). Exceptions were grade 3 fatigue improving in less than 2 weeks, grade 3 EGFR inhibition–associated skin toxicity (if improved to grade ≤2 within 2 weeks from onset), and grade 3 diarrhea (of less than 2 days' duration). All adverse events were coded with the use of the Medical Dictionary for Regulatory Activities.

Following dose escalation, dose-expansion cohorts with weekly doses of 9 or 12 mg/kg were conducted to obtain additional safety, pharmacokinetic, and pharmacodynamic data and to explore antitumor activity in patients with metastatic colorectal cancer. Doses were adjusted (paused, decreased, or discontinued) in response to treatment-emergent toxicities. All patients remained on treatment until disease progression, unacceptable toxicities, or withdrawal of consent. Toxicities were assessed through clinical examination and laboratory assessments and were graded by NCI CTCAE version 4.02. Safety evaluations were conducted for all treated patients at baseline and at regular intervals, including assessment of the onset and degree of skin toxicities.

For pharmacokinetic analyses, we collected serial blood samples to measure serum levels of the individual Sym004 antibodies using ELISAs. Blood samples were taken before and after all Sym004 infusions in the dose-escalation and dose-expansion cohorts. In addition, postinfusion pharmacokinetic profile samples were taken 1, 2, 4, 8, 24, and 48 hours after the first infusion and 1, 2, 4, 8, and 24 hours after the fourth infusion. For immunogenicity, a validated double antigen ELISA with immobilized Sym004 and horseradish peroxidase–labeled Sym004 in solution was used for detection of anti-Sym004 antibodies before the first and planned fifth infusions (27). The lower limit of detection was 150 ng/mL using rabbit anti-Sym004 IgG from hyperimmunized animals as a positive control for assay performance. Serum concentrations of Sym004 were calculated as the sum of the concentration of individual antibodies mAb992 and mAb1024 following infusion of Sym004, using two validated competitive ELISA-labeled anti-idiotypic antibody fragments specific for mAb992 or mAb1024 (AbD Serotec). The lower limit of quantification for the individual antibodies mAb992 and mAb1024 was 0.25 μg/mL.

For the pharmacodynamic analyses, serial skin and tumor biopsies [before the first (baseline) and planned fifth Sym004 infusion (week 4)] were obtained from patients enrolled in the dose-expansion cohorts after the patients had provided signed informed consent and investigators had assessed the risk before the procedure. Formalin-fixed and paraffin-embedded skin and tumor biopsy samples were sent to Covance Laboratories Ltd. and Pathology Diagnostics Ltd. for immunohistochemistry analysis and evaluation of EGFR expression (0.86 µg/mL anti-EGFR mAb; Dako M7239) and proliferation (0.92 µg/mL anti-Ki67 mAb; Dako M7240). A histopathologic score (H-score) with a range of 0 to 300 was calculated according to the following formula: (3 × percentage of cells with strong staining) + (2 × percentage of cells with moderate staining) + (1 × percentage of cells with weak staining).

Tumor samples were assessed for tissue content and content of neoplastic cells by a pathologist, and then analyzed for targeted mutation and gene copy-number profile (Department of Pathology, Radboud University Medical Center, Nijmegen, the Netherlands). Using NGS (IonTorrent platform), frequently mutated positions in AKT1 (reference sequence NM_005163.2; exon 3), BRAF (reference sequence NM_004333.4; exon 15), EGFR (reference sequence NM_005228.3; exons 12, 18–21), ERBB2 (reference sequence NM_004448.2; exon 20), KRAS (reference sequence NM_004985.3; exons 2–4), NRAS (reference sequence NM_002524.3; exons 2–4), and PIK3CA (reference sequence NM_006218.2; exons 10, 21) were assessed. MET copy-number changes were analyzed with FISH. Amplification was defined as MET gene(G)-to-copy number (CN) control probe (centromere chromosome 7) ratio (G:CN) of greater than 2.2 scored in >20 tumor nuclei. Equivocal G:CN ratios of <2.2 were considered negative for amplification, as was polysomy, i.e., increasing numbers of chromosomes carrying a single copy of the MET gene.

Tumors were evaluated by computerized tomography or MRI per RECIST version 1.1 (28) against a baseline assessment performed within 3 weeks before the first dose. Reassessments were performed after the first 6 weeks of treatment and thereafter every 8 weeks. Tumor responses were evaluated both locally (by investigators to enable treatment allocation) and centrally (for independent review and primary analysis).

The primary endpoint of this trial was safety and tolerability; secondary endpoints included antitumor activity, and pharmacokinetic and pharmacodynamic effects (cutoff date, October 29, 2014). We evaluated baseline characteristics and adverse events in all 62 patients (full analysis set) who were treated with Sym004. For the assessments of pharmacokinetic and pharmacodynamic data, only the patients with available samples were considered. Descriptive statistics were provided for demographic, safety, exposure, antitumor, pharmacokinetic, and pharmacodynamic data. Categorical data were summarized by frequency and percentages; continuous data were summarized by median and range, or geometric mean and coefficient of variation for pharmacokinetics. H-scores of skin and tumor biopsies were presented using box plots. Mean pharmacokinetic profiles and geometric mean plots of AUC_{0–168 h} by dose (after first dose) were presented for the dose-escalation cohort, and dose-normalized C_trough values over time until dose reduction or interruption were presented for the dose-expansion cohorts.

R. Dienstmann has received speakers' bureau honoraria from Symphogen. A. Patnaik is a consultant/advisory board member for Symphogen. E. Van Cutsem has received research funding from Amgen, Bayer, Boehringer, Lilly, Merck Serono, Novartis, Roche, and Sanofi. A.W. Tolcher is a consultant/advisory board member for AbbVie, Adnexus, Ambit, AP Pharma, Aragon, Ariad, ArQule, Asana, Astex, Astellas, Avid, Bayer, Bind, BioMed Valley Discoveries, Blend Therapeutics, Bristol-Myers-Squibb, Celator, Clovis, Curis, Daiichi Sankyo, Dicerna, Eisai, Emergent Product Development, Endo, Five Prime, Galapagos, Heron, Janssen, Lilly, MedImmune, Merck, Sharp & Dohme, Mersana, Merus, Micromet, Nanobiotix, Nektar, Neumedicines, Novartis, OncoGenex, Onyx, Otsuka, Pfizer, Pharmacyclics, Pierre Fabre, ProNai, Proximagen, Sanofi-Aventis, Santaris, Symphogen, Vaccinex, Valent Technologies, and Zyngenia. J. Tabernero is a consultant/advisory board member for Amgen, Celgene, Chugai, Imclone, Lilly, Merck KGaA, Millennium, Novartis, Roche, Sanofi, Symphogen, and Taiho. No potential conflicts of interest were disclosed by the other authors.

The authors were fully responsible for all content and editorial decisions, were involved at all stages of manuscript development, and have approved the final version.

Conception and design: R. Dienstmann, A. Patnaik, M.W. Pedersen, I.D. Horak, E. Van Cutsem, A.W. Tolcher, J. Tabernero

Development of methodology: R. Dienstmann, M. Benavent, N.J.Ø. Skartved, R. Hald, M.W. Pedersen, M. Kragh, E. Van Cutsem, A.W. Tolcher, J. Tabernero

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Dienstmann, A. Patnaik, R. Garcia-Carbonero, A. Cervantes, M. Benavent, S. Roselló, B.B.J. Tops, R.S. van der Post, G. Argilés, M.W. Pedersen, M. Kragh, E. Van Cutsem, A.W. Tolcher, J. Tabernero

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Dienstmann, A. Patnaik, R. Garcia-Carbonero, R.S. van der Post, U.H. Hansen, M.W. Pedersen, S. Braun, E. Van Cutsem, A.W. Tolcher, J. Tabernero

Writing, review, and/or revision of the manuscript: R. Dienstmann, A. Patnaik, R. Garcia-Carbonero, A. Cervantes, S. Roselló, R.S. van der Post, G. Argilés, N.J.Ø. Skartved, U.H. Hansen, R. Hald, M.W. Pedersen, M. Kragh, I.D. Horak, S. Braun, E. Van Cutsem, A.W. Tolcher, J. Tabernero

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B.B.J. Tops, G. Argilés, S. Braun, J. Tabernero

Study supervision: R. Dienstmann, U.H. Hansen, S. Braun, A.W. Tolcher, J. Tabernero

The authors thank P. Pamperin, Symphogen A/S, and H. Liedman, TFS, for assistance in the preparation and editing of this article, and M. Düring, BioStata ApS, for statistical assistance. European sites thank the European Union Seventh Framework Programme, grant 259015 (COLTHERES), for supporting translational research in colorectal cancer.

This study was funded by Symphogen A/S and Merck KGaA (ClinicalTrials.gov trial registration ID: NCT01117428).

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