TRAIL receptor agonist conatumumab with modified FOLFOX6 plus bevacizumab for first-line treatment of metastatic colorectal cancer

A randomized phase 1b/2 trial

Authors


  • We thank Marintan Spring for support with PK analysis. Kerri Hebard-Massey (Amgen Inc.) provided medical writing support.

Abstract

BACKGROUND

In patients with previously untreated metastatic colorectal cancer (mCRC), we conducted a phase 1b/randomized phase 2 trial to define the safety, tolerability, and efficacy of mFOLFOX6 plus bevacizumab (mFOLFOX6/bev) with conatumumab, an investigational, fully human monoclonal IgG1 antibody that specifically activates death receptor 5 (DR5).

METHODS

Twelve patients were enrolled in a phase 1b open-label dose-escalation trial of conatumumab with mFOLFOX6/bev; thereafter, 190 patients were randomized 1:1:1 to receive mFOLFOX6/bev in combination with 2 mg/kg conatumumab, 10 mg/kg conatumumab, or placebo. Therapy cycles were repeated every 2 weeks until disease progression or the occurrence of unacceptable toxicity.

RESULTS

In phase 1b, conatumumab with mFOLFOX6/bev was tolerated without apparent added toxicity over mFOLFOX6/bev alone. In phase 2, conatumumab with mFOLFOX6/bev did not confer a benefit in progression-free survival when compared with placebo with mFOLFOX6/bev. Toxicity was similar in all treatment arms. Following treatment, similar increases in circulating caspase-3 levels were observed in all arms.

CONCLUSIONS

Conatumumab with mFOLFOX6/bev did not offer improved efficacy over the same chemotherapy with placebo in first-line treatment of patients with mCRC. These data do not support further development of conatumumab in advanced CRC. Cancer 2013;119:4290–4298. © 2013 American Cancer Society.

INTRODUCTION

Despite introduction of several new systemic therapies for treatment of metastatic colorectal cancer (mCRC), the 5-year survival rate remains only 12%.[1-3] New therapies are needed to address this unmet patient need.

Selective activation of extrinsic apoptosis pathways in cancer cells represents a potential therapeutic target for systemic treatment of advanced malignancies.[4] Tumor necrosis factor–related apoptosis inducing ligand (TRAIL) is a member of the tumor necrosis factor superfamily. TRAIL can induce apoptosis by binding to death receptor 4 (DR4) or death receptor 5 (DR5) and activating downstream caspases, which ultimately leads to p53-independent cell death.[4]

Conatumumab is an investigational, fully human monoclonal IgG1 antibody that specifically binds DR5.[5] In preclinical studies, conatumumab induced apoptosis in a variety of cancer cell lines with little toxicity to normal cells and inhibited growth and potentiated the effects of irinotecan and 5-fluorouracil (5-FU) in CRC xenografts.[5]

In humans, DR5 is expressed in most CRCs.[6] In the first-in-human phase 1 trial of single-agent conatumumab, 1 patient with mCRC achieved a metabolic partial response, and 4 patients with mCRC had stable disease for >12 weeks.[7] Together, the preclinical and clinical evidence suggests that conatumumab combined with chemotherapy may be effective in patients with mCRC. We therefore conducted a phase 1b trial to define the maximum tolerated dose (MTD) for conatumumab and mFOLFOX6 plus bevacizumab (mFOLFOX6/bev) in patients with previously untreated mCRC followed by a randomized, placebo-controlled phase 2 trial of mFOLFOX6/bev ± conatumumab to assess treatment efficacy, safety, and tolerability. Exploratory pharmacodynamic and patient stratification biomarkers were also assessed.

MATERIALS AND METHODS

Patient Population

The patient population included patients ≥18 years with histologically or cytologically confirmed metastatic adenocarcinoma of the colon or rectum; measurable disease (per RECIST v1.0); Eastern Cooperative Oncology Group (ECOG) performance score of 0 or 1; and adequate hematological, renal, and hepatic function. Key exclusion criteria included history or known presence of central nervous system metastases or other malignancies; prior chemotherapy or other systemic therapy for metastatic disease; radiotherapy to >25% of bone marrow; or receipt of an investigational agent for treatment of CRC. Adjuvant chemotherapy was allowed if disease progression was documented ≥12 months after completion of adjuvant chemotherapy. The study was approved by local institutional review boards at each site, and written informed consent was obtained from all patients.

Study Objectives

The primary phase 1 end point was the incidence of adverse events (AEs) and clinical laboratory abnormalities defined as dose-limiting toxicities (DLTs). The primary phase 2 end point was the estimation of progression-free survival (PFS). Secondary end points included overall survival (OS), objective response rate, incidence of AEs and clinical laboratory abnormalities, incidence of conatumumab antibody formation, relative dose intensity, and conatumumab pharmacokinetic (PK) parameters. Exploratory end points included biomarker analysis of caspase activation and polymorphisms in the low-affinity immunoglobulin gamma Fc region receptor III-A (FCGR3A).

Study Design and Treatment

Phase 1b was an open-label dose-escalation study conducted at 6 sites to determine the safety, tolerability, and MTD of conatumumab and mFOLFOX6/bev. Two doses of intravenous conatumumab (3 and 10 mg/kg) were evaluated in sequential dose levels on a once-every-2-weeks (Q2W) schedule. On day 1 of each treatment cycle, patients received a 2-hour infusion of 85 mg/m2 oxaliplatin with 200 mg/m2 leucovorin l-form or 400 mg/m2 racemic, then an intravenous bolus of 400 mg/m2 5-FU, then 5 mg/kg bevacizumab over 30 minutes, then conatumumab over 30 minutes, and then a 46- to 48-hour continuous infusion of 2400 mg/m2 5-FU. Treatment cycles were repeated until disease progression or the occurrence of unacceptable toxicity.

In phase 2, 2 and 10 mg/kg conatumumab was investigated because PK modeling predicted that Q2W 2 mg/kg conatumumab was expected to provide continuous exposure above the EC50 values observed in preclinical xenograft models,[5] antitumor activity was seen in the first-in-human study with a dose as low as 0.3 mg/kg,[7] and a wider range of exposures provided better assessment of exposure-response relationships.

Phase 2 was a randomized, double-blind, placebo-controlled study conducted at 48 sites to evaluate the safety and efficacy of conatumumab and mFOLFOX6/bev for first-line treatment of mCRC. One hundred and ninety patients were randomized 1:1:1 to receive mFOLFOX6/bev in combination with 2 mg/kg conatumumab, 10 mg/kg conatumumab, or placebo intravenously on a Q2W schedule. Drugs were administered as in phase 1, except administration of bevacizumab before mFOLFOX6 was allowed at the investigator's discretion. Patients were stratified by ECOG performance score (0 or 1) and serum lactate dehydogenase (≤ institutional upper limit of normal [ULN] vs >ULN) at the time of randomization. Patient disposition is shown in Figure 1.

Figure 1.

Phase 2 patient disposition.

Sample Size Selection

Phase 2 was designed to estimate the treatment effect (assessed by PFS) of adding conatumumab to mFOLFOX6/bev, rather than to test formal inferences. Assuming a median PFS of 9 months in the placebo arm, approximately 180 patients were needed to have 33% power to detect a 3-month improvement in PFS at a 5% level of significance when combining the 2 active arms.

Data Analysis

Analyses were conducted separately for phase 1b and phase 2. Summary statistics are provided for safety, efficacy, PK, antibody, and exploratory data. For continuous variables, mean and standard deviation were calculated. For categorical variables, frequency and percentage in each category were displayed.

The first patient was dosed on October 3, 2007. The primary analysis was performed when 120 patients had documented investigator-assessed disease progression or death (May 4, 2010). The final analysis was conducted when the study ended (September 1, 2011). At that time, 152 patients had documented, investigator-assessed disease progression or death, and 102 patients had died. PFS from the primary and final analyses was very similar. Thus, only results from the final analysis are presented.

Safety

Safety end points were DLTs and AE incidence. A DLT was defined as any grade ≥3 hematologic or nonhematologic toxicity related to conatumumab or to the combination of conatumumab and mFOLFOX6/bev. Exceptions included alopecia, lymphocytopenia, and anemia. AE severity was determined using the National Cancer Institute Common Toxicity Criteria for AEs v3.0.

Efficacy

Time-to-event end points were summarized using the Kaplan-Meier method among patients enrolled in the phase 2 study. PFS was calculated as the number of days from randomization to the first assessment of disease progression (per RECIST v1.0 or clinical progression) or death due to any cause. Stratified Cox regression models were used for analysis of PFS and OS, with patients analyzed as randomized. Stratified log-rank tests and Tarone trend tests were used to determine the strength of any conatumumab dose response.

Tumor response was assessed by the investigator per RECIST v1.0. Imaging to evaluate tumor response was performed every 8 weeks (±7 days). The proportion of patients achieving an objective response was calculated, and the presence of a dose response was tested using a Cochran-Armitage trend test.

Conatumumab antibodies

Serum samples for anticonatumumab antibody assessments were collected before administration on day 1 of cycles 1 and 5 and every 12 cycles thereafter and at the end-of-study visit.

Pharmacokinetics

Blood samples were collected at all phase 1 and selected phase 2 sites to measure serum levels of conatumumab, bevacizumab, oxaliplatin, and 5-FU. Conatumumab serum concentrations were measured before infusions on day 1 of cycles 1, 2, 3, and 5 and every 8 cycles thereafter; within 5 minutes before the end of the conatumumab infusion on cycles 1, 3, and 5 and every 8 dosing cycles thereafter; 48 (±4) hours after the start of the conatumumab infusion on cycles 1 and 3; and at the end-of-study visit. Bevacizumab, oxaliplatin, and 5-FU concentrations were measured before each infusion and within 5 minutes before the end of the infusion during cycles 1 and 3.

As previously described,[8] serum concentrations of conatumumab and bevacizumab were measured using a validated enzyme-linked immunosorbent assay, and plasma 5-FU and oxaliplatin concentrations were estimated by validated liquid chromatography–mass spectrometry assays. PK parameters were summarized using WinNonlin Enterprise software, version 5.1.1 (Pharsight Corp., Mountain View, CA).

Biomarkers

At all phase 1 sites and select phase 2 sites, blood samples were collected before mFOLFOX6/bev infusions on day 1 and 48 (±4) hours after the start of the mFOLFOX6/bev infusions on day 3 of cycles 1 and 3. DNA from blood samples was used for FCGR3A F158V single-nucleotide polymorphism genotyping by allelic discrimination assay. FV and VV genotypes were combined for PFS and OS analyses because of the low number of patients with the VV polymorphism and overall small sample size.[9] Active caspase 3/7, caspase-cleaved cytokeratins, and circulating DNA fragments were assayed as previously described.[10]

RESULTS

Patients

In phase 1b, 16 patients were screened, 6 patients received 3 mg/kg conatumumab with mFOLFOX6/bev, and 6 patients received 10 mg/kg conatumumab with mFOLFOX6/bev. Patient demographics and baseline disease characteristics are summarized in Table 1.

Table 1. Phase 1b Baseline Patient Demographics and Disease Characteristics
 3 mg/kg Conatumumab (n  =  6)10 mg/kg Conatumumab (n  =  6)All Patients (n  =  12
  1. Abbreviation: ECOG, Eastern Cooperative Oncology Group.

Sex, n (%)   
Male1 (17)3 (50)4 (33)
Female5 (83)3 (50)8 (67)
Race/ethnicity, n (%)   
White/Caucasian4 (67)6 (100)10 (83)
Black/African American2 (33)0 (0)2 (17)
Age (y) at enrollment   
Mean (SD)57.7 (8.9)49.0 (9.1)53.3 (9.7)
Primary tumor type, n (%)   
Colon2 (33)4 (67)6 (50)
Colorectal, nonspecified2 (33)1 (17)3 (25)
Rectal2 (33)1 (17)3 (25)
Months since metastatic diagnosis   
Mean (SD)1.07 (1.01)1.17 (0.42)1.12 (0.74)
Disease stage at time of enrollment, n (%)   
Stage IV6 (100)6 (100)12 (100)
ECOG performance score at screening, n (%)   
05 (83)3 (50)8 (67)
11 (17)3 (50)4 (33)

In phase 2, 236 patients were screened; 190 patients were randomized 1:1:1 to 2 mg/kg conatumumab/mFOLFOX6/bev (n = 64), 10 mg/kg conatumumab/mFOLFOX6/bev (n = 62), or placebo/mFOLFOX6/bev (n = 64). Baseline patient demographics and disease characteristics were generally well balanced across treatment arms with slightly more women in the conatumumab arms than in the placebo arm (Table 2). Disease progression (44%, 58%, and 47%), partial withdrawal of consent (11%, 11%, and 9%), and AEs (16%, 10%, and 6%) were major reasons for discontinuation of treatment in the 2 mg/kg conatumumab, 10 mg/kg conatumumab, and placebo arms, respectively.

Table 2. Phase 2 Baseline Demographics and Disease Characteristics
 2 mg/kg Conatumumab (n  =  64)10 mg/kg Conatumumab (n  =  62)Placebo (n  =  64)
  1. Abbreviations: ECOG, Eastern Cooperative Oncology Group; LDH, lactate dehydrogenase; ULN, upper limit of normal.

Sex, n (%)   
Male40 (63)35 (56)44 (69)
Female24 (38)27 (44)20 (31)
Race/ethnicity, n (%)   
White/Caucasian55 (86)46 (74)48 (75)
Black/African American6 (9)6 (10)8 (13)
Hispanic/Latino2 (3)7 (11)5 (8)
Asian1 (2)2 (3)3 (5)
American Indian/Alaska Native0 (0)1 (2)0 (0)
Age (y) at randomization   
Mean (SD)58.8 (11.5)57.1 (12.0)59.6 (12.6)
Primary tumor type, n (%)   
Colon49 (77)48 (77)51 (80)
Colorectal, nonspecified2 (3)4 (6)3 (5)
Rectal13 (20)10 (16)10 (16)
ECOG performance status   
032 (50)31 (50)34 (53)
132 (50)31 (50)30 (47)
Prior chemotherapy, n (%)   
Yes11 (17)8 (13)9 (14)
No53 (83)54 (87)55 (86)
LDH, n (%)   
≤ULN33 (52)33 (53)34 (53)
>ULN31 (48)29 (47)30 (47)

Safety

Phase 1b adverse events

Grade 3 or 4 treatment-emergent AEs were seen in 5 patients (83%) in the 3 mg/kg conatumumab arm and in 6 patients (100%) in the 10 mg/kg conatumumab arm. Common grades 3 and 4 treatment-emergent AEs are listed in Table 3. No patients experienced grade 5 AEs.

Table 3. Grades 3/4 Treatment-Emergent Adverse Events With Incidence ≥10% in Phase 1b
Adverse Event3 mg/kg Conatumumab (n  =  6)10 mg/kg Conatumumab (n  =  6)
  1. No grade 5 adverse events were observed.

Febrile neutropenia, n (%)  
Grade 31 (17)1 (17)
Grade 40 (0)0 (0)
Neutropenia, n (%)  
Grade 31 (17)0 (0)
Grade 41 (17)1 (17)
Diarrhea, n (%)  
Grade 31 (17)1 (17)
Grade 40 (0)0 (0)
Peripheral neuropathy, n (%)  
Grade 31 (17)2 (33)
Grade 40 (0)0 (0)
Pulmonary embolism, n (%)  
Grade 30 (0)0 (0)
Grade 40 (0)2 (33)
Palmar-plantar erythrodysesthesia syndrome, n (%)  
Grade 30 (0)2 (33)
Grade 40 (0)0 (0)
Deep vein thrombosis, n (%)  
Grade 32 (33)0 (0)
Grade 40 (0)0 (0)

No patients in the 3 mg/kg conatumumab arm experienced a DLT. Two patients in the 10 mg/kg conatumumab arm experienced grade 4 pulmonary embolisms, which were not thought to be related to study medication. Thus, the MTD was not reached.

Phase 2 adverse events

Grade 3 or greater treatment-emergent AEs were seen in 81% of patients in the 2 mg/kg conatumumab arm (n = 51), 84% of patients in the 10 mg/kg conatumumab arm (n = 52), and 87% of patients in the placebo arm (n = 54). AEs were relatively evenly distributed across treatment arms, with the exception of grades 3 and 4 hypokalemia, which occurred more frequently in the conatumumab arms than in the placebo arm (Table 4). Overall, 4 patients (2%) had a grade 5 treatment-emergent AE; 1 patient in the 2 mg/kg conatumumab arm had a malignant neoplasm and died within 30 days of the last study treatment because of progressive disease; 1 patient in the 10 mg/kg conatumumab arm had respiratory failure; and in the placebo arm, 1 patient had a cerebral hemorrhage, and 1 had pneumonia.

Table 4. Grades 3/4 Treatment-Emergent Adverse Events With an Incidence ≥10% in Phase 2
Adverse Event2 mg/kg Conatumumab (n  =  63)10 mg/kg Conatumumab (n  =  62)Placebo (n  =  63)
  1. Four patients (2%) had a grade 5 treatment-emergent adverse event.

Abdominal pain, n (%)   
Grade 37 (11)5 (8)6 (10)
Grade 41 (2)0 (0)0 (0)
Decreased appetite, n (%)   
Grade 36 (10)1 (2)4 (6)
Grade 40 (0)0 (0)0 (0)
Deep vein thrombosis, n (%)   
Grade 36 (10)3 (5)5 (8)
Grade 40 (0)1 (2)0 (0)
Diarrhea, n (%)   
Grade 37 (11)10 (16)9 (14)
Grade 40 (0)3 (5)0 (0)
Fatigue, n (%)   
Grade 37 (11)7 (11)8 (13)
Grade 40 (0)1 (2)0 (0)
Hypokalemia, n (%)   
Grade 33 (5)9 (15)0 (0)
Grade 41 (2)0 (0)0 (0)
Peripheral neuropathy, n (%)   
Grade 38 (13)8 (13)11 (17)
Grade 40 (0)0 (0)0 (0)
Neutropenia, n (%)   
Grade 312 (19)15 (24)12 (19)
Grade 46 (10)9 (15)6 (10)
Pulmonary embolism, n (%)   
Grade 30 (0)3 (5)0 (0)
Grade 44 (6)4 (6)6 (10)

Anticonatumumab antibodies

One patient in phase 2 tested positive for preexisting anticonatumumab antibodies on study day 1. All subsequent immunoassay test results for this patient were negative. No other patients developed anticonatumumab antibodies.

Efficacy

Progression-free survival

PFS was not significantly different among the conatumumab and placebo arms. Median PFS (95% CI) was 9.5 months (8.5-14.0 months) in the 2 mg/kg conatumumab arm, 10.6 months (9.0-13.0 months) in the 10 mg/kg conatumumab arm, and 11.0 months (9.1-13.1 months) in the placebo arm. Compared with placebo, the stratified HR (95% CI) was 0.90 (0.60-1.35) for the 2 mg/kg conatumumab arm (P = .59 by stratified log-rank test) and 0.95 (0.64-1.40) for the 10 mg/kg conatumumab arm (P = .80 by stratified log-rank test). Kaplan-Meier estimates of PFS are shown in Figure 2.

Figure 2.

Kaplan-Meier plot of progression-free survival.

Overall survival

OS was not significantly different among the conatumumab and placebo arms. Median OS (95% CI) was 22.6 months (17.5, not estimable [NE] months) in the 2 mg/kg conatumumab arm, 29.0 months (18.6, NE months) in the 10 mg/kg conatumumab arm, and 23.8 months (17.7, NE months) in the placebo arm. Compared with placebo, the stratified HR (95% CI) was 0.93 (0.58-1.49) in the 2 mg/kg conatumumab arm (P = .80 by stratified log-rank test) and 0.80 (0.49-1.31) in the 10 mg/kg conatumumab arm (P = .30 by stratified log-rank test). Kaplan-Meier estimates of OS are shown in Figure 3.

Figure 3.

Kaplan-Meier plot of overall survival.

Objective response

Of 190 patients randomized in phase 2, 185 (97%) had measurable disease at baseline. Objective response was similar across treatment arms; 30 (49%), 32 (52%), and 31 (50%) patients had an objective response in the 2 mg/kg conatumumab, 10 mg/kg conatumumab, and placebo arms, respectively.

Relative Dose Intensity

In phase 2, mean relative dose intensity (calculated as the ratio of actual cumulative dose to protocol-specified cumulative dose over the specified period) for conatumumab was similar between the 2 mg/kg and 10 mg/kg conatumumab arms (94% vs 92%). Mean relative dose intensities for the 5-FU bolus (83% vs 85% vs 88%), 5-FU infusion (81% vs 83% vs 88%), leucovorin (86% vs 85% vs 86%), oxaliplatin (83% vs 84% vs 87%), and bevacizumab (88% vs 89% vs 91%) were also similar among the 2 mg/kg conatumumab arm, the 10 mg/kg conatumumab arm, and the placebo arm, respectively.

Pharmacokinetics

Conatumumab exposure increased approximately dose proportionally after infusions of 2, 3, or 10 mg/kg conatumumab plus mFOLFOX6/bev. Following 2 cycles of treatment with 10 mg/kg conatumumab, the mean serum trough and maximum observed concentrations were 61.7 and 249 mg/mL, respectively. Conatumumab did not have a significant effect on bevacizumab or oxaliplatin PK.

Biomarker Analysis

We examined whether FCGR3A F158V polymorphisms, which result in high- and low-affinity forms of FCGR3A, could influence outcome in patients who received conatumumab.[11, 12] Of 156 patients genotyped for the FCGR3A F158V polymorphisms, 72 patients had the FF genotype, 64 patients had the FV genotype, and 20 patients had the VV genotype. Relatively equal numbers of patients with each polymorphism were found in each arm. In FV/VV patients, no significant differences in PFS or OS were observed among patients treated with conatumumab versus placebo, although improvement in PFS approached significance (Table 5).

Table 5. PFS and OS by FCGR3A Polymorphism
 ConatumumabPlacebo
FF (n  =  48)FV or VV (n  =  57)FF (n  =  24)FV or VV (n  =  27)
  1. FF and FV, polymorphisms in the immunoglobulin gamma Fc region receptor III-A (FCGR3A) gene; PFS, progress-free survival; CI, confidence interval; HR, hazard ratio; OS, overall survival; NE, not estimable.

PFS    
Median PFS (mo)10.211.013.29.5
95% CI9.2, 15.98.7, 14.09.7, 23.86.1, 12.8
Unstratified HR1.500.60  
95% CI0.84, 2.700.35, 1.02  
OS    
Median OS (mo)30.324.324.224.2
95% CI17.5, NE20.1, NE19.0, NE16.9, NE
Unstratified HR1.310.92  
95% CI0.64, 2.660.48, 1.75  

Changes in serum levels of several cell death biomarkers, including caspase 3/7, gDNA, caspase-cleaved cytokeratin-18, and total soluble cytokeratin-18, were evaluated at baseline and on day 3 of cycle 1. Although statistically significant increases in caspase 3/7, gDNA, caspase-cleaved cytokeratin-18, and total soluble cytokeratin-18 were observed across all groups, no differences were seen among arms (Table 6).

Table 6. Change From Baseline for Biomarkers of Cell Death
Biomarkera2 mg/kg Conatumumab (n  =  64)10 mg/kg Conatumumab (n  =  62)Placebo (n  =  64)
  1. a

    Measured at cycle 1 day 3 relative to cycle 1 day 1 predose.

  2. P values determined from the Wilcoxon rank sum test on log-fold change from baseline.

Caspase 3/7   
n151215
Mean fold change over baseline1.331.371.27
Log SD0.260.370.28
P value.0012.0161.0026
gDNA   
n171716
Mean fold change over baseline1.491.591.79
Log SD0.520.590.55
P value.015.015.0003
Caspase-cleaved cytokeratin-18   
n171716
Mean fold change over baseline1.391.261.37
Log SD0.210.330.39
P value.0001.011.0131
Total Soluble cytokeratin-18   
n171716
Mean fold change over baseline1.271.351.38
Log SD0.170.220.37
P value< .0001.0003.0006

DISCUSSION

In this multicenter, phase 1b/randomized phase 2 trial, the addition of conatumumab to standard first-line chemotherapy for mCRC (mFOLFOX6/bev) did not confer a PFS benefit compared with mFOLFOX6/bev alone. In other recent phase 2 trials, conatumumab also did not significantly improve PFS in combination with panitumumab[13] or FOLFIRI in patients with mCRC.[14] The reasons for the lack of effect are unclear, as preclinical studies have shown that conatumumab can inhibit colorectal tumor growth and potentiate the effects of 5-FU in preclinical xenograft models.[5] In preclinical studies, conatumumab could also potentiate the effect of irinotecan.[5] Further research could be considered to determine the efficacy of conatumumab in combination with other regimens.

Conatumumab with mFOLFOX6/bev had a manageable toxicity profile, and treatment-emergent AEs were similar among arms. A greater incidence of grade 3 hypokalemia was seen in patients receiving either 2 mg/kg conatumumab (5%) or 10 mg/kg conatumumab (15%) relative to placebo (0%). The mechanism underlying hypokalemia is unknown. As there was no difference in gastrointestinal toxicity among arms, the higher rate of hypokalemia may indicate renal potassium wasting, but this would have to be confirmed by dedicated studies.

Several factors may have contributed to the lack of efficacy in this study. Preclinical studies have shown that conatumumab must be cross-linked to induce apoptosis.[5, 11, 12, 15] However, it is unknown whether conatumumab becomes cross-linked in human cancer patients. A polymorphism in the FCGR3A gene leads to high- and low-affinity forms of FCGR3A, with the VV genotype having higher affinity for IgG1 than the FF genotype. Previous studies have suggested that patients with higher-affinity forms of FCGR3A may have an improved response to conatumumab.[16, 17] In this study, a trend toward improved PFS with conatumumab treatment was observed among patients with high-affinity forms of FCGR3A, although this did not reach statistical significance. In contrast to the apparent interaction for PFS, FCGR3A genotype did not modify the effect of conatumumab on OS, and no clear OS benefit for conatumumab was observed for any genotype.

In addition, conatumumab induces apoptosis through activation of downstream caspases. In this study, levels of activated caspase-3 were elevated relative to baseline in both the 2 and 10 mg/kg conatumumab arms, consistent with the results of the first-in-human trial.[7] However, a similar increase over baseline was also seen in the placebo arm, indicating that conatumumab did not induce additional caspase activation over mFOLFOX/bev alone.

Together, these data suggest that weak activation of the TRAIL receptor may contribute to the lack of efficacy seen in this study. Stronger activation may be seen in patients with the VV polymorphism in the FCGR3A gene, contributing to the trend toward improved PFS observed in these patients. In addition, other components of the extrinsic apoptotic pathway, such as cFLIP and caspase 8, can negatively impact TRAIL-induced apoptosis[18] and may also have impacted this trial. These factors may act as predictive biomarkers for response to conatumumab and other TRAIL receptor ligands.

Multiple TRAIL receptor agonists are currently in clinical trials for non–small cell lung cancer, lymphoma, soft-tissue carcinoma, CRC, pancreatic cancer, ovarian cancer, hepatocellular carcinoma, multiple myeloma, and other advanced tumors. These agents have generally been well tolerated, but the efficacy results available so far have been disappointing.[7, 8, 13, 14, 19-29] Similarly, we have shown that conatumumab with mFOLFOX6/bev did not demonstrate improved efficacy over placebo in first-line treatment of patients with mCRC. Overall, the data do not support further development of conatumumab in advanced CRC. Identification of biomarkers and greater understanding of the extrinsic apoptosis pathway in cancer treatment may help to guide the development of therapies that target this pathway.

FUNDING SOURCES

This study was funded by Amgen Inc (Clinicaltrials.gov number NCT00625651).

CONFLICT OF INTEREST DISCLOSURES

Yong-jiang Hei, Yang Pan, Vincent Haddad, Cheng-Pang Hsu, and Antony Sabin are Amgen Inc. employees and stockholders. Francesco Galimi was an Amgen Inc employee at the time this study was conducted. Charles S. Fuchs has been a consultant for Amgen Inc. Mark F. Kozloff has been a consultant for and on a speakers bureau for Amgen Inc. Lee Schwartzberg served on an Amgen Inc advisory board. All remaining authors have declared no conflicts of interest.

Ancillary