Purpose: To evaluate the safety, pharmacokinetics, and antitumor activity of AMG 386, an investigational selective angiopoietin 1/2-neutralizing peptibody, in combination with FOLFOX-4 (F), carboplatin/paclitaxel (C/P), or docetaxel (D), in adult patients with advanced solid tumors.
Experimental Design: Three cohorts of patients (F, n = 6; C/P, n = 8; D, n = 12) received one full cycle of chemotherapy alone during the pretreatment phase, followed by administration of AMG 386 10 mg/kg i.v. weekly in combination with chemotherapy until disease progression or intolerance. Safety and tolerability, tumor response, pharmacokinetic profiles, and biomarkers were assessed.
Results: Twenty-six patients were enrolled; 22 received treatment with AMG 386. No dose-limiting toxicities or grade 3 or 4 adverse events related to AMG 386 were reported. The most common adverse events were diarrhea and hypomagnesemia (n = 3 each). One patient developed grade 2 hypertension and one had grade 1 subconjunctival eye hemorrhage. No neutralizing antibodies to AMG 386 were detected. There were no pharmacokinetic interactions between AMG 386 and F, C/P, or D. One patient receiving AMG 386 plus C/P for bladder cancer refractory to gemcitabine/cisplatin had a complete response at week 8. The remaining best tumor responses were partial response (n = 3, one from each cohort), stable disease ≥8 weeks (n = 13), and progressive disease (n = 1).
Conclusions: Weekly administration of AMG 386 in combination with three common chemotherapy regimens was well tolerated in patients with advanced solid tumors. No pharmacokinetic interactions between AMG 386 and any of the tested chemotherapy regimens were noted. Promising antitumor activity was observed with all three treatment combinations. Clin Cancer Res; 16(11); 3044–56. ©2010 AACR.
Therapies targeting vascular endothelial growth factor or its receptors in combination with standard chemotherapy have been shown to be effective in the treatment of some solid tumors but may have burdensome or even unacceptable toxicities. AMG 386 is an investigational first-in-class peptibody that inhibits tumor angiogenesis by blocking the interaction between angiopoietins 1 and 2 and their receptor, Tie2. The results from this phase 1 study show that administration of AMG 386 at 10 mg/kg i.v. once weekly in combination with FOLFOX-4, carboplatin/paclitaxel, or docetaxel at standard doses was well tolerated and showed evidence of antitumor activity. Adverse events generally associated with vascular endothelial growth factor pathway inhibition were uncommon. Pharmacokinetic results suggested that there was no drug-drug interaction potential between AMG 386 and the various chemotherapies. The data from this study show that blockade of the angiopoietin axis may provide a well-tolerated alternative to blocking angiogenesis.
Angiogenesis, which is required for tumor proliferation and metastasis, is tightly controlled by a balance between proangiogenic and antiangiogenic factors. When the balance shifts in favor of the proangiogenic factors, tumors initiate the development of a vasculature (1). Thus, targeted agents inhibiting such factors have been successfully used in the treatment of several types of cancers. Specifically, inhibitors targeting the vascular endothelial growth factor (receptor) [VEGF(R)] pathway have shown clinical efficacy (2–6) but may be associated with an increased risk for a number of adverse events that are considered class effects, such as hypertension, hemorrhagic and thromboembolic events, gastrointestinal perforations, and impaired wound healing (7).
Angiopoietin 1 and angiopoietin 2 (Ang1 and Ang2), which bind to the vascular receptor tyrosine kinase Tie2, are key cytokines that regulate, in distinct ways, proangiogenic pathways involved in the later stages of neovascularization. Although Ang1 seems to promote recruitment of pericytes to vascular tubes (8), thereby contributing to vessel stabilization and maturation, Ang2 is expressed predominantly at sites of vascular remodeling, where it acts as a vessel destabilizer. Depending on the signaling environment, Ang2 may have either antagonistic or agonistic properties (9, 10). Ang2 has been shown to be upregulated at sites of tumor angiogenesis (11–13) and Ang2 overexpression results in enhanced tumor angiogenesis and tumor growth in some experimental models of cancer (14). Increased Ang2 expression is a negative prognostic factor associated with poor survival in patients with acute myeloid leukemia (15) and breast cancer (13).
AMG 386 is an investigational recombinant peptide-Fc fusion protein (“peptibody”) containing a peptide sequence that binds Ang1 and Ang2. AMG 386 reduces tumor angiogenesis by selectively neutralizing Ang1 and Ang2, thereby blocking their interaction with the Tie2 receptor (16). In a tumor xenograft mouse model, systemic treatment with AMG 386 inhibited tumor growth, with subsequent disappearance of all measurable tumors in some of the animals and evidence of an antiangiogenic mechanism of action (16). In a first-in-human dose escalation study in advanced solid tumors, treatment with single agent AMG 386 showed promising antitumor activity and was well tolerated, with no maximum tolerated dose reached and no reports of toxicities typically associated with VEGFR blockade, except for transient grade 1 or 2 hypertension in patients who had preexisting hypertension (17). The aim of the present study was to evaluate the safety, pharmacokinetic profile, and antitumor activity of AMG 386 administered at a weekly dose of 10 mg/kg in combination with commonly administered chemotherapy regimens given at standard doses in adult patients with advanced solid tumors.
Materials and Methods
Study design and eligibility criteria
This open-label phase 1b study conducted at two centers in the United States assessed the safety, tolerability, and pharmacokinetic profiles of AMG 386 in combination with each of three chemotherapy regimens that are typically administered in three of the most common tumor types (colorectal cancer, non–small-cell lung cancer, and breast cancer): FOLFOX-4 [oxaliplatin, leucovorin, and 5-fluorouracil (5-FU)], carboplatin/paclitaxel, and docetaxel. The study protocol was approved by the institutional review boards of the participating centers, and the study was conducted in accordance with the Declaration of Helsinki. Signed informed consent was obtained from all patients. The primary end point was the clinical safety profile of AMG 386. Secondary end points were to characterize the pharmacokinetic profiles of AMG 386 in combination with each of the three chemotherapy regimens and to assess anti–AMG 386 antibody formation and tumor response as measured by Response Evaluation Criteria in Solid Tumors (RECIST). Exploratory end points were to evaluate changes in serum levels of angiogenic cytokines.
Key eligibility criteria were age ≥18 years; pathologically documented advanced solid tumors refractory to standard therapy (or which was refused by the patient or was unavailable); life expectancy >3 months; measurable or evaluable disease per RECIST guidelines; Eastern Cooperative Oncology Group performance status of 0 to 2; adequate hematologic, hepatic, and renal functions; and candidacy for study chemotherapy regimens (FOLFOX-4, carboplatin/paclitaxel, or docetaxel), as deemed by the investigator. Patients were ineligible if they had received prior treatment with AMG 386 or other antiangiogenic agents; concurrent or prior treatment (within 1 week of enrollment) with an anticoagulant; tumor-directed antibody therapy within 6 weeks of study entry; anticancer therapy within 30 days of enrollment; or therapeutic or palliative radiation therapy within 3 weeks of study entry. Other key exclusion criteria were symptomatic or untreated central nervous system metastases requiring concurrent treatment; cardiovascular disease; myocardial infarction or arterial or venous thrombosis within 1 year of enrollment; uncontrolled hypertension; history of bleeding diathesis; or pulmonary hemorrhage or gross hemoptysis within 6 months of study entry.
Initially, six patients were enrolled concurrently into each of the three cohorts based on the chemotherapy regimen chosen by the investigator. Cohort expansion up to a total of nine patients per cohort was permitted if two of the initial six patients experienced a dose-limiting toxicity (DLT) or a serious adverse event deemed possibly related to AMG 386 by the investigator within the first 28 days of AMG 386 treatment.
Each patient received one full cycle of standard-dose chemotherapy before beginning treatment with AMG 386 (cycle 1). A cycle was defined as 2 weeks for the FOLFOX-4 regimen and 3 weeks for the carboplatin/paclitaxel and docetaxel regimens. The FOLFOX-4 regimen was administered every 2 weeks as follows: On day 1, patients received oxaliplatin 85 mg/m2 and leucovorin 200 mg/m2 i.v. infusion, followed by 5-FU 400 mg/m2 i.v. bolus, followed by 5-FU 600 mg/m2 i.v. continuous infusion. On day 2, patients received leucovorin 200 mg/m2 i.v. infusion, followed by 5-FU 400 mg/m2 i.v. bolus, followed by 5-FU 600 mg/m2 i.v. continuous infusion. Carboplatin/paclitaxel was administered at carboplatin area under the serum concentration-time curve (AUC) of 6 mg/mL · min and paclitaxel at 200 mg/m2 i.v. on day 1 of the 3-week cycle. Docetaxel was administered at 75 mg/m2 i.v. on day 1 of a 3-week cycle. Patients who experienced chemotherapy-related adverse events of grade 3 or worse according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) version 3.04 or patients who were unable to receive a second cycle of full-dose chemotherapy due to safety concerns were withdrawn from the study and new patients were enrolled.
Beginning on day 1 of week 1 of cycle 2, patients in each cohort received AMG 386 10 mg/kg i.v. once weekly in combination with one of the three chemotherapy regimens at the described doses and dose schedules. In the first-in-human monotherapy study, doses of 10 and 30 mg/kg were shown to be tolerable (17). Following a conservative approach, 10 mg/kg was chosen for the combination treatment with chemotherapy. This dose was modeled in preclinical toxicology studies to maintain AMG 386 concentrations in the serum that exceed the 90% effective concentration (EC90; Amgen, Inc., data on file). No AMG 386 dose escalations were permitted. Patients continued combination treatment until disease progression or intolerance occurred. Dose adjustments for both AMG 386 and chemotherapy regimens were at the discretion of the investigator and followed protocol-specified rules based on treatment-related toxicities in the preceding cycle.
Adverse events were monitored throughout the study and were classified according to the Medical Dictionary for Regulatory Activities5 and graded according to CTCAE, version 3.0. On day 1 of the pretreatment phase and on day 1 of each subsequent treatment week throughout the study period, a clinical evaluation, complete blood count, hematology tests, coagulation tests, chemistry panel, and urinalysis with microscopic examination were performed.
A DLT was defined as any AMG 386–related (as determined by the investigator) grade 3 or worse hematologic or nonhematologic toxicity that occurred during the initial 28 days of treatment, with some adverse events considered a DLT only if the following criteria were met: transient grade 3 infusion reactions lasting more than 2 hours; grade 3 fatigue persisting for more than 7 days; grade 3 or 4 nausea, vomiting, or diarrhea despite maximum supportive care; grade 3 or 4 neutropenia with fever >38.5°C; grade 4 neutropenia persisting for more than 7 days; grade 4 thrombocytopenia; grade 4 anemia; grade 3 or 4 peripheral neuropathy; or aspartate aminotransferase or alanine aminotransferase >10 times the upper limit of normal. Toxicities occurring during the pretreatment phase (before AMG 386 administration) were not considered DLTs.
Tumor response assessments
Tumor assessment was done using computed tomography or magnetic resonance imaging within 4 weeks before the start of the pretreatment phase (baseline) and every 8 weeks thereafter throughout the study. Tumor response was assessed by the investigator using RECIST 1.0 (18) and included all patients who received AMG 386 and who had an assessment of tumor burden at baseline and at one or more time points thereafter.
Serum samples for pharmacokinetic analysis of AMG 386 were collected predose on day 1; predose at weeks 2 and 3; predose and at the end of infusion at week 4; and 6, 24, 48, 96, and 168 (before the next dose) hours postdose at week 4; predose at week 8; every 8 weeks thereafter and at study exit. Plasma samples for pharmacokinetic analysis of chemotherapy components were collected on day 1 of the pretreatment phase at predose (pre-bolus for 5-FU), 1 hour after infusion start (preinfusion for 5-FU), and at the end of infusion. Additional samples were collected predose, 1 hour after infusion start, end of infusion, and 0.5, 1, 6, 24, and 48 (except leucovorin) hours after infusion at week 4 (carboplatin/paclitaxel and docetaxel cohorts) or week 5 [FOLFOX-4 cohort; 5-FU pharmacokinetic samples collected prebolus, preinfusion; at 1, 4, and 6 hours after infusion start; and at the end of infusion (24 hours)]; at weeks 6 and 7 (oxaliplatin pharmacokinetics); and at the end of the study.
Serum samples were analyzed for AMG 386 concentrations by MDS Pharma Services using a validated ELISA with a lower limit of quantification of 19.578 ng/mL. Analyses of sodium-heparinized plasma samples or serum samples for concentrations of chemotherapy agents were carried out using validated methods with the following lower limits of quantification: paclitaxel and docetaxel, 0.100 and 10.0 ng/mL, respectively (liquid chromatography-tandem mass spectrometry method; BASi Northwest Laboratory, Inc.); total and free platinum, 50 ng/mL (inductively coupled plasma mass spectrometry method; MDS Pharma Services); 5-FU, 5 ng/mL (liquid chromatography-tandem mass spectrometry method; Quest Pharmaceutical Services); leucovorin, 0.500 μg/mL for R-leucovorin and S-leucovorin and 0.100 μg/mL for (6S)-5-methyltetrahydrofolic acid (liquid chromatography-tandem mass spectrometry method; PPD, Inc.).
Noncompartmental analysis was done using WinNonlin Professional software on Citrix (version 5.1.1, Pharsight Corporation) to estimate the pharmacokinetic parameters from individual AMG 386 concentrations in serum or chemotherapeutic agent concentrations in plasma or serum.
For AMG 386, the following pharmacokinetic parameters were assessed: the terminal half-life [t1/2 = ln(2)/λz, where λz was the first-order terminal rate constant estimated via linear regression of the terminal log-linear decay phase]; the observed maximum concentration (Cmax) value after i.v. infusion, time (tmax) at which Cmax occurred, and the area under the serum concentration-time curve from time 0 to the dosing interval (AUC0-τ, with τ equal to 168 hours postdose), estimated using the linear/log trapezoidal method; serum clearance (CL), calculated as the mg/kg dose divided by AUC0-τ at the fourth dose; and the minimum observed concentration (Cmin) value for each patient, calculated by taking the mean of the trough concentrations at steady state. For chemotherapy, the Cmax and tmax values after i.v. infusion, AUC values from time 0 to the last quantifiable concentration or to infinity (AUC0-t and AUC0-inf, respectively), and CL were evaluated.
To assess the effect of AMG 386 administration on chemotherapy, the SAS PROC MIXED procedure (SAS for Windows, version 9.1, WIN_PRO platform, SAS Institute, Inc.) was used to calculate the geometric least square means (GLSM) and the ratio of GLSM between cycle 2 (treatment with AMG 386) and cycle 1 (treatment without AMG 386) for Cmax. For the calculation of GLSM, the least square means for the log-transformed Cmax values (logCmax) for cycle 2 and cycle 1 were obtained first, and then these values were converted back to the original scale. The calculation of GLSM ratios and 90% confidence interval (90% CI) was done as follows: First, the difference (and 90% CI) in the least square means between cycle 2 and cycle 1 for logCmax was estimated; then, the obtained numbers were converted back to their original scale. The analysis included all patients who received AMG 386 and chemotherapy and who had evaluable pharmacokinetic results.
Anti–AMG 386 antibody assays
Serum samples were collected before dosing at treatment weeks 1, 2, 4, and 8, every 4 weeks thereafter, and at study exit. Samples were probed for the presence of anti–AMG 386 binding antibodies using a validated acid-dissociation bridging electrochemiluminescent immunoassay [Meso-Scale Discovery (MSD)]. Briefly, anti–AMG 386 antibodies were detected by forming a bridged immune complex with biotinylated AMG 386 and AMG 386 conjugated to a ruthenium chelate complex. Samples positive for anti–AMG 386 antibodies in the immunoassay were further evaluated for neutralizing antibodies using a validated in vitro receptor binding assay. The assay measured antibodies capable of inhibiting the biological activity of AMG 386 (i.e., blockade of the interaction of angiopoietin with a soluble Tie2 receptor). Samples that were positive in both assays were considered positive for neutralizing antibodies; samples that were negative in the immunoassay were not tested in the receptor binding assay. Samples that were positive in the immunoassay but negative in the receptor binding assay were defined as positive for binding antibodies.
Blood samples were collected before dosing on days 1 (baseline), 2, 3, 8, 15, 22, 50, and 106 of the treatment phase and at study exit. Serum samples originally collected for pharmacokinetics on days 22 (6 hours after infusion), 23, 24, and 26 were also used.
Levels of placental growth factor (PlGF), VEGF-A, VEGFR-1 (and basic fibroblast growth factor; data not shown) were quantified using a 4-plex sandwich immunoassay with electrochemiluminescent detection [Meso-Scale Discoveries (MSD)] following the manufacturer's instructions. Similarly, a 2-plex MSD assay was used to measure levels of VEGFR-2 (and Kit; data not shown). Briefly, samples were applied to 96-well microplates containing monoclonal capture antibodies spotted onto working electrodes. After incubation and washing steps, ruthenium-linked detection antibodies specific for each analyte were added and electrochemiluminescence was measured after application of a current. A standard curve run with every plate was used to quantify each analyte. Each sample was measured in triplicate. Soluble vascular cell adhesion molecule-1 (sVCAM-1), soluble intercellular adhesion molecule-1 (sICAM-1), and Tie2 were quantified using specific ELISA kits (R&D Systems) following the manufacturer's instructions, and a standard curve was run in parallel. To achieve a normal distribution, all data were log-transformed before statistical testing. All patients who received AMG 386 and who provided blood samples were included in the biomarker analysis. The relationship between the change in biomarker concentration at various treatment times was evaluated using a regression ANOVA model. Separate models were fit for each combination of the biomarkers, chemotherapy cohort, and sample type (serum versus plasma). P values were not corrected for multiple comparisons.
The safety analysis set included all treated patients who received at least one dose of AMG 386. All data were summarized descriptively. No formal comparisons between cohorts were done. The data were analyzed using SAS software version 8.2 (SAS Institute). The statistical analysis was completed on May 30, 2008.
Patient disposition, demographics, and baseline characteristics
From March 2006 to January 2007, a total of 26 patients were enrolled in the study, 4 of whom experienced chemotherapy-related dose-limiting adverse events during cycle 1 when no AMG 386 was given and therefore discontinued the study. The remaining 22 patients received at least one dose of AMG 386: 6 patients were enrolled in the FOLFOX-4 cohort, 7 in the carboplatin/paclitaxel cohort, and 9 in the docetaxel cohort. Table 1 summarizes the demographic and baseline characteristics. Patients were mainly Caucasian (73%); the median age of the patient population was 59 years. Across treatment groups, most patients had an Eastern Cooperative Oncology Group performance status of 0 or 1. All patients had disease stage IV and had prior treatment for cancer. The most common tumor types were breast (n = 6; 27%) and prostate (n = 3; 14%). At the time of the data analysis, two patients were still receiving treatment (both in the docetaxel arm; one patient with breast cancer and one with prostate cancer). Twenty patients discontinued treatment for the following reasons: disease progression (n = 5), adverse event [jugular vein thrombosis (n = 1; FOLFOX-4 cohort), bacterial sepsis (n = 1; carboplatin/paclitaxel cohort), eye hemorrhage and pulmonary embolism (n = 1 each; docetaxel cohort)], withdrawal of consent (n = 3), death (n = 2), and other (n = 6).
Patients received a median of 14 once-weekly doses of AMG 386 (range, 1-62+). Patients in the FOLFOX-4 cohort received a median of 6.5, 13, and 26 cycles of oxaliplatin, leucovorin, and 5-FU, respectively. Patients in the carboplatin/paclitaxel and docetaxel cohorts each received a median of 6 cycles of chemotherapy agents.
Dose-limiting toxicities and adverse events
No DLTs were observed in either of the cohorts. Therefore, cohort expansion was not necessary per protocol-specified criteria. The carboplatin/paclitaxel and docetaxel cohorts enrolled one and three more patients, respectively, to replace those who discontinued the study before completing 28 days of treatment with AMG 386 without experiencing a DLT.
All 22 patients had at least one adverse event. Of these, 11 patients had grade 3 events [mostly neutropenia (n = 6) and dyspnea (n = 3)] and 3 patients experienced a grade 4 event (all reversible neutropenia; one in each cohort). Eight patients had serious events, including grade 3 pulmonary embolism (n = 1; docetaxel cohort) and grade 3 jugular vein thrombosis (n = 1; docetaxel cohort); both events were deemed possibly related to AMG 386 treatment. One patient had grade 2 hypertension (docetaxel cohort). This patient had a history of hypertension and had received antihypertensive medication before enrollment in the study. Two deaths occurred on study [hepatic failure (carboplatin/paclitaxel cohort) and respiratory failure (docetaxel cohort)]; neither was considered to be related to AMG 386 treatment.
All adverse events considered to be related to AMG 386 treatment were grade ≤2 (10 of 22 patients; 45%) and were mostly diarrhea and hypomagnesemia (Table 2). Two patients (docetaxel cohort) had adverse events of interest that have previously been associated with angiogenesis inhibitors targeting the VEGF pathway: one patient had grade 2 hypertension and one patient had grade 1 subconjunctival eye hemorrhage. No AMG 386–related gastrointestinal perforations or events of impaired wound healing were observed during the study (the events of pulmonary embolism and jugular vein thrombosis were possibly related to AMG 386). Fourteen patients (64%) had proteinuria (by dipstick) of level 1+ or higher; two patients (one each in the carboplatin/paclitaxel and docetaxel cohorts) had proteinuria of level 3+. All shifts from baseline in urine protein were 1 or 2 grades.
Twenty patients developed one or more grade 3 laboratory toxicities, most frequently lymphopenia (n = 11), neutropenia (n = 11), leukopenia (n = 8), and hyperglycemia (n = 5). Five patients had reversible grade 4 toxicities, most commonly low total neutrophil count (n = 3).
Anti–AMG 386 antibodies
Twenty-one patients provided evaluable postdose samples for antibody testing. One patient developed nonneutralizing binding antibodies to AMG 386 during treatment, and one had preexisting, nonneutralizing binding antibodies (samples from week 1); samples collected during AMG 386 treatment tested negative. The presence of anti–AMG 386 binding antibodies had no apparent effect on the AMG 386 concentration in the serum or on the pharmacokinetic parameters of any of the patients who tested positive for binding antibodies (data not shown). All 21 patients tested negative for anti–AMG 386 antibodies at study exit.
Pharmacokinetic data for AMG 386 in combination with chemotherapy compared with data for AMG 386 monotherapy treatment obtained in the phase 1 study (17) are shown in Fig. 1. After four once-weekly infusions of AMG 386, Cmax and AUC0-τ seemed to be similar across chemotherapy cohorts, with average values of 219 μg/mL and 8,120 h·μg/mL, respectively. Likewise, the mean CL also seemed to be similar across chemotherapy cohorts, with an overall mean of 1.08 mL/h/kg. The mean terminal t1/2 ranged from 80.8 to 99.0 hours (3.4-4.1 days). AMG 386 seemed to reach steady state after four weekly i.v. infusions. As shown in Fig. 1, these data are similar to what has been reported in the AMG 386 phase 1 single-agent study.
Pharmacokinetic parameter values for the chemotherapy agents are summarized in Table 3. Cmax values for carboplatin and paclitaxel were comparable in the same patients who first received chemotherapy alone during the pretreatment period (cycle 1) and then later in combination with 10 mg/kg AMG 386. The ratio of the GLSM point estimates (GLSMratio) of total platinum Cmax values for oxaliplatin was slightly greater than 1 (GLSMratio, 1.13; 90% CI, 1.05-1.22). Cmax values for docetaxel were slightly higher during the cycle 2 (week 4) assessment (mean GLSMratio, 1.37; 90% CI, 0.94-1.99) in the same patients who first received chemotherapy alone (cycle 1) and then in combination with 10 mg/kg AMG 386; however, the 90% CI includes 1.0. Cmax values for 5-FU coadministered as FOLFOX-4 with 10 mg/kg AMG 386 could not be compared with 5-FU Cmax values administered without AMG 386 because 5-FU alone pharmacokinetic parameters were not available for any patients. Leucovorin Cmax values in the same patients receiving FOLFOX-4 alone (cycle 1) and then in combination with 10 mg/kg AMG 386 were generally comparable between most patients (data not shown).
Tumor response was evaluable for 18 patients (FOLFOX-4, n = 5; carboplatin/paclitaxel, n = 6; docetaxel, n = 7); 4 patients had discontinued the study without receiving a follow-up scan. One patient with bladder cancer (carboplatin/paclitaxel cohort) achieved a complete response at week 8 of the treatment phase for 10 weeks (prior cancer treatment: one line of chemotherapy). Three patients achieved a partial response: one patient with pancreatic cancer (FOLFOX-4 cohort; prior cancer treatment: two lines of chemotherapy); one patient with prostate cancer (docetaxel cohort; prior cancer treatments: hormonal and radiation therapy); and one patient with breast cancer (carboplatin/paclitaxel cohort; prior cancer treatments: four lines of chemotherapy and radiation therapy). At the time of data analysis, the patient with pancreatic cancer had discontinued the study while the patient with prostate cancer was still receiving treatment with AMG 386 for 64+ weeks. Thirteen patients had a best response of stable disease (FOLFOX-4, n = 3; carboplatin/paclitaxel, n = 4; docetaxel, n = 6) and one had progressive disease. Fourteen patients showed some reduction from baseline in tumor measurement, with 12 patients showing a reduction >10% (Fig. 2).
Because the angiopoietin/Tie2 axis remains yet to be explored in detail, selecting a panel of suitable biomarkers for this pathway is challenging. Therefore, we have focused on signaling molecules that are generally known to be involved in angiogenesis and on adhesion molecules thought to be involved in endothelial remodeling. Samples for biomarker analysis were available from 22 patients. Across all combination regimens, a statistically significant increase in mean PlGF concentration (1.7-fold change from baseline; P < 0.0001) was noted after 3 days of treatment (Fig. 3A). By day 8, the mean PlGF concentration decreased toward baseline. Likewise, on day 22, the PlGF concentration was significantly higher than at baseline (1.7-fold change; P < 0.0001) as early as 6 hours after infusion before returning to baseline levels 4 days later (day 26). Mean fold PlGF concentration changes from baseline over time were similar when chemotherapy cohorts were analyzed and compared individually (Fig. 3B).
Concentrations of sVCAM-1 across all chemotherapy cohorts increased after AMG 386 administration, showing an initial peak at day 8 (1.4-fold change from baseline; P < 0.0001) before decreasing toward baseline (Fig. 3C). On day 22, 6 hours after AMG 386 infusion, the sVCAM-1 concentration was again significantly greater than at baseline, reaching a maximum at day 26 (1.7-fold change from baseline; P < 0.0001) before decreasing. Changes in mean sVCAM-1 concentration over time were similar comparing individual chemotherapy cohorts (Fig. 3D). No changes in levels of sICAM-1 were observed with AMG 386 treatment.
Across all chemotherapy cohorts, mean VEGF-A and VEGFR-1 levels showed no appreciable changes with AMG 386 treatment over time (data not shown). Mean concentrations of VEGFR-2 slowly decreased to below baseline levels (1.16-fold decrease, P = 0.006) up to day 22, with a rapid return up to baseline again after AMG 386 administration on day 22. Results were similar when individual chemotherapy cohorts were analyzed separately (data not shown). Levels of Tie2 did not change with AMG 386 treatment (data not shown).
The present study showed that AMG 386 administered at 10 mg/kg once weekly was well tolerated in combination with three commonly used chemotherapy regimens administered at their standard dose intensities to patients with advanced cancer. During treatment, no DLTs occurred and all AMG 386–related adverse events observed were mild to moderate (grade 1 or 2) and manageable. Overall, adverse events and laboratory toxicity profiles were comparable to those anticipated with administration of chemotherapy alone and similar to what has been described for AMG 386 monotherapy (17). AMG 386 in combination with chemotherapy showed evidence of antitumor activity.
Four patients had adverse events commonly associated with VEGF(R) pathway inhibition (7), including thromboembolic events. In two of those patients, the adverse events were considered to be related to AMG 386 treatment (grade 1 subconjunctival eye hemorrhage and grade 2 hypertension in a patient who had preexisting hypertension at enrollment), whereas they were deemed possibly related in two patients (grade 3 pulmonary embolism and jugular vein thrombosis). The data suggest that AMG 386 in combination with FOLFOX-4, carboplatin/paclitaxel, or docetaxel has a distinct toxicity profile that is consistent with results from the first-in-human study showing a similar adverse event summary when AMG 386 was administered as a monotherapy (17). These results are encouraging considering that combinations of targeted therapies with cytotoxic chemotherapy regimens at standard doses could result in unacceptable toxicity in some patient populations. Clearly, larger studies are needed to more thoroughly investigate the clinical safety and efficacy profile of agents such as AMG 386 in combination with chemotherapy. No neutralizing antibodies were detected. Whereas the presence of binding antibodies in one patient had no apparent effect on AMG 386 concentrations in the serum, its effect on pharmacokinetic parameters is currently unknown.
The pharmacokinetic parameters of AMG 386 when coadministered with the three tested chemotherapy regimens were similar to those observed with AMG 386 in the phase 1 single-agent study (17), suggesting that the pharmacokinetics of AMG 386 was not significantly affected by the concurrent administration of either chemotherapy treatment. Likewise, a comparison of pharmacokinetic parameters between chemotherapy alone (cycle 1) and combination therapy with AMG 386 (subsequent cycles) suggests that coadministration of AMG 386 had no significant effect on the pharmacokinetics of carboplatin/paclitaxel or oxaliplatin (coadministered as part of the FOLFOX-4 regimen). However, these results need to be interpreted with caution. First, the pharmacokinetics of 5-FU was difficult to evaluate due to limited concentration-time profiles for cycles 1 and 2. Based on the results that were obtained, AMG 386 at 10 mg/kg did not seem to affect 5-FU exposure administered as FOLFOX-4 chemotherapy. Second, although pharmacokinetic results for docetaxel with and without AMG 386 were similar, the effect of AMG 386 coadministration on docetaxel exposure cannot be fully evaluated because docetaxel infusion duration varied and may have affected docetaxel Cmax values (data not shown).
Five potential biomarkers of angiogenesis were analyzed in the current study in an attempt to measure the biological activity of AMG 386. Of all biomarkers tested, only PlGF and VCAM-1 displayed a sustained change from baseline throughout the study. PlGF increased concomitantly with AMG 386 administration. This increase was similar in all three chemotherapy cohorts. Like VEGF, PlGF (a VEGF homologue) promotes pathologic angiogenesis (19). We hypothesize that inhibiting anywhere along the angiogenic pathways could affect either of these signaling molecules. Specifically, blocking downstream pathways could result in compensatory (increased) upstream proangiogenic signaling. In fact, upregulated PlGF seems to amplify the responsiveness of endothelial cells to VEGF (19). Increases in PlGF have been reported with other antiangiogenic agents such as sunitinib (20–22), bevacizumab (23), and motesanib (24). VCAM-1 is involved in vascular remodeling, and variations in this biomarker may be indicative of a biological response to changes in the vascular endothelium.
In conclusion, the present study showed that AMG 386, an investigational inhibitor of the Tie2 pathway, in combination with full doses of FOLFOX-4, carboplatin/paclitaxel, or docetaxel chemotherapy, was well tolerated and showed evidence of antitumor activity in adult patients with advanced solid tumors. The data from this study support larger clinical studies of AMG 386 in combination with standard chemotherapies.
Disclosure of Potential Conflicts of Interest
A.C. Mita: honoraria from speakers' bureau, Genentech; C.H. Takimoto: clinical study support, Amgen; A. Tolcher: extensive consultant agreements; E. Rasmussen, Z.D. Zhong, M.B. Bass: employment and ownership interest (stock), Amgen; Y-N. Sun, employment and ownership interest (stock), Amgen; N. Le, employment (at the time the study was conducted) and ownership interest (stock), Amgen P. LoRusso: commercial research grant, Amgen; honoraria, Abraxis, Bristol-Myers Squibb, Genentech, GlaxoSmithKline, National Cancer Institute, Novartis, Sanofi, Syndax, Takeda; scientific advisory board, Abraxis, AstraZeneca, Bristol-Myers Squibb, Eisai, Genentech, GlaxoSmithKline, Novartis, Pfizer, Sanofi, Takeda; speaker, Genentech, GlaxoSmithKline, Sanofi; scientific review committee, National Cancer Institute; data safety monitoring board, Syndax.
We thank Jennifer Klem, Ph.D. (funded by Amgen) and Beate Quednau, Ph.D. (Amgen) for editorial support (assistance in drafting the manuscript); Teresa Wong, B.S., Cindy Wake, B.S., Mario Bejarano, B.S., and Rebeca Melara, M.S. (Amgen) for pharmacokinetic sample and data analyses; Gail Silverman (Amgen) for biological sample management; Michael Davis, Ren Xu, Susan Cottrell, Kim Hamic, Chang-Pin Huang, Huan-Mei Khoo, Lisa Kivman, and Matthew Peach (Amgen) for biomarker assay support; and Steve Dinnogen, Susan Jacques, Arunan Kaliyaperumal, Ph.D., and Michael Moxness, Ph.D. (Amgen) for antibody assay support.
Grant Support: Amgen, Inc.
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.
↵5MedDRA; version 6.1 (2008) International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH). MedDRA/MSSO resources page. Available at: http://www.meddramsso.com/MSSOWeb/index.htm Updated October, 2003. Accessed January 27, 2009.
- Received December 31, 2009.
- Revision received March 24, 2010.
- Accepted April 4, 2010.