Phase I Dose-Escalation Study of Intravenous Aflibercept in Combination with Docetaxel in Patients with Advanced Solid Tumors

Vascular endothelial growth factor (VEGF) has become a major target for anticancer therapy as its overexpression is associated with increased tumor vascularity, progression, metastasis, and poor prognosis (1–3). Convincing clinical evidence in support of targeting VEGF was first shown with the humanized anti-VEGF monoclonal antibody, bevacizumab in several tumor types (4–6). The successful demonstration of the role of anti-VEGF therapy offers the perspective that the optimization of such approaches might deliver further improvements in efficacy.

Aflibercept (VEGF trap) is a fully human soluble decoy receptor protein composed of domain 2 of VEGF receptor-1 (VEGFR-1) and domain 3 of VEGF receptor-2 (VEGFR-2) fused to the Fc region of human immunoglobulin 1 (7). This construct gives aflibercept broad affinity for all ligands that bind to VEGFR-1 and VEGFR-2, including all isoforms of VEGF-A, VEGF-B, and also placental growth factor (PIGF; refs. 7, 8). Aflibercept has shown robust antitumor effects and antiangiogenic activity as single agent and in combination with chemotherapy in a variety of mouse models of cancer (7–12). Although treatments with docetaxel alone and in combination with aflibercept both initially regressed tumor size, the combination was significantly more effective at delaying long-term tumor regrowth (12).

Results from phase I study of aflibercept monotherapy (dosage range, 0.3–7 mg/kg every 2 weeks) in patients with advanced solid tumors showed that aflibercept was well tolerated and had evidence of clinical activity (13). Dose-limiting toxicities (DLT) were rectal ulceration and proteinuria encountered at the 7.0 mg/kg dose. Other VEGF blockade-specific toxicities included hypertension, proteinuria, and dysphonia. On the basis of these DLT observations and the increase in frequency and severity of adverse events (AE) at aflibercept doses greater than 4 mg/kg every 2 weeks, the recommended phase II dose of aflibercept (RP2D) as a single agent is 4 mg/kg intravenously every 2 weeks. In addition, this dose met the biologic hypothesis of achieving adequate drug concentration of free and bound drug levels.

In a recently published phase II study, in heavily pretreated patients with lung adenocarcinoma aflibercept monotherapy given at 4 mg/kg every 2 weeks showed a favorable safety profile and preliminary evidence of antitumor activity as monotherapy: overall response rate of 2% (95% CI: 0.2–7.2), prolonged disease stabilization [including 22% stable disease (SD) at 90 days], and median progression-free survival (PFS) of 12 weeks (14). Given the enhanced activity of aflibercept plus docetaxel in preclinical studies, the aim of the current investigation was to evaluate the safety of aflibercept plus docetaxel in patients with advanced solid tumors in anticipation of further evaluations of efficacy with this combination.

Patients aged 18 years or more were included if they had a histologically or cytologically confirmed solid malignancy that was metastatic or unresectable or for which no standard conventional therapy existed, but for which docetaxel treatment was considered appropriate. Other main inclusion criteria were Eastern Cooperative Oncology Group (ECOG) performance status 2 or less, adequate hematologic, hepatic and renal function; and resolution of any toxicity (except alopecia) from other anticancer treatments. Patients were excluded if they had squamous cell lung carcinoma; brain metastases; prior cumulative platinum chemotherapy exceeding 300 mg/m²; cumulative radiotherapy to more than 25% of the total bone marrow; chemotherapy, hormonal therapy, radiotherapy, or surgery during the previous 3 weeks; immunotherapy or cytokine therapy within the previous 6 weeks; or uncontrolled hypertension. All patients gave written informed consent. The study was approved by local ethics committees and conducted according to the Declaration of Helsinki.

This study (TCD6120) comprises 4 independent cohorts investigating aflibercept in combination with different backbone chemotherapies administered every 3 weeks: aflibercept + docetaxel (75 mg/m²), aflibercept + docetaxel (75 mg/m²) + cisplatin (75 mg/m²), aflibercept + docetaxel (100 mg/m²), and aflibercept + pemetrexed (500 mg/m²). The first cohort is reported here; the other cohorts will be reported separately. This cohort comprised a dose-escalation study conducted in 2 centers. In the dose-escalation phase, patients received aflibercept (sanofi-aventis and Regeneron Pharmaceuticals) by i.v. infusion more than 1 hour on day 1 at a starting dose of 2 mg/kg, with enrollment into the next dose level (4, 5, 6, 7, and 9 mg/kg) based on DLT observation during the first cycle (i.e., 3-week period). A dose escalation to the subsequent dose level was made if no DLT was seen in a cohort of 3 patients. In addition, demonstration of acceptable safety for each dose level in a parallel study of single-agent aflibercept was a prerequisite for enrolling patients into the corresponding dose level in the present combination study. If DLT was observed in 1 patient, the cohort was expanded to 6 patients. The RP2D was defined as the highest dose at which 2 of 3 to 6 patients experienced DLT (defined below). Once RP2D had been determined, patients were enrolled in an expansion cohort to further evaluate safety and activity at the RP2D of the combination.

All patients received an i.v. infusion of docetaxel (75 mg/m²) more than 1 hour on day 1 immediately after the aflibercept i.v. infusion. Dexamethasone (8 mg) was given before and after docetaxel administration. Prophylactic use of hematopoietic growth factors was not permitted in the first cycle. Cycles were repeated every 3 weeks until Response Evaluation Criteria in Solid Tumors (RECIST, version 1.0)-defined disease progression (15), unacceptable toxicity, or withdrawal of patient's consent.

Primary objectives were to evaluate DLTs during the first cycle and determine the RP2D of aflibercept in combination with docetaxel. Secondary objectives were to assess safety, preliminary antitumor activity, pharmacokinetics, and immunogenicity during the entire study period for each given patient.

Safety was assessed based on AEs, laboratory data, vital signs, electrocardiograms, and physical examinations collected for each patient until 60 days after the end of study treatment. Toxicities were graded with the National Cancer Institute Common Terminology Criteria for Adverse Events (version 3.0). Hematologic DLTs were grade 3/4 neutropenia complicated by fever of 38.5°C or more or infection, grade 4 neutropenia for 7 days or more, grade 4 thrombocytopenia or grade 3 thrombocytopenia with hemorrhage. Nonhematologic DLTs were defined as grade 3/4 nonhematologic toxicities, except fatigue, anorexia, nausea, vomiting, diarrhea, serum alkaline phosphatase increase, or grade 3 events that were not clinically relevant or unrelated to study treatment in the investigator's opinion. Nonhematologic DLTs also included blood pressure of 150/100 mmHg or more (or ≥180/90 mmHg in patients with prior history of systolic hypertension) despite 4 weeks' medical management; urine protein more than 3.5 g/24 h that did not recover to less than 2.0 g/24 h within 2 weeks as assessed by spot urinalysis, unless otherwise warranted; and symptomatic arterial thromboembolic events. Safety evaluations were carried out on all patients exposed to at least 1 dose of study.

Tumor response was assessed according to RECIST using magnetic resonance imaging (MRI) or computed tomography (CT) on day 21 (± 4 days) of every other even-numbered cycle, at the end of the study treatment, or if disease progression was suspected. The patients evaluable for antitumor activity were those who underwent baseline assessment, received 2 or more cycles of study treatment, and had 1 or more postbaseline CT or MRI scan unless early progression or death occurred.

Pharmacokinetic analyses were conducted on the safety-evaluable population. Blood samples were obtained before first drug administration and then at 1, 2, 4, 8, 24, and 48 hours after the start of the aflibercept infusion on day 1 and on days 7 and 14 during cycle 1 for the assessment of free aflibercept and VEGF: aflibercept complex (VEGF-bound aflibercept). Samples were collected before dosing for subsequent cycles and at the end of treatment. Levels of free aflibercept and VEGF-bound aflibercept were measured in plasma by direct ELISA, with limit of quantification (LOQ) of 15.6 and 31.5 ng/mL, respectively. Pharmacokinetic parameters of free and VEGF-bound aflibercept were assessed by noncompartmental analysis (WinNonlin Professional, version 5.2.1, PharSight). Pharmacokinetic parameters were plasma concentrations before starting infusion at each treatment cycle (C_trough), maximum plasma concentration (C_max), time to reach C_max (t_max), and area under the concentration versus time curve from time 0 to t_last (AUC_last, for VEGF-bound aflibercept only). For free aflibercept only, AUC from time 0 to infinity (AUC_0–∞), elimination half-life (t_½) and clearance (CL) were also assessed.

Plasma samples for docetaxel analysis were obtained pretreatment and at various time points on day 1; plasma levels were measured by electrospray liquid chromatography mass spectrometry (LOQ; 1.0 ng/mL). CL and AUC for docetaxel were estimated on day 1 of cycle 1 by a Bayesian estimation method.

For measurement of anti-aflibercept antibodies, serum samples were to be obtained before the first dose of study drug, predose on day 1 of each odd-numbered cycle during the study treatment period, upon study withdrawal, and 3 months after the last aflibercept administration, for all treated patients. Detection of anti-aflibercept antibodies was carried out with a validated quantitative ELISA (LOQ, 238.4 ng/mL).

Fifty-four patients were enrolled and treated; 34 patients in the dose-escalation phase and 20 patients in the expansion phase. Patient characteristics at baseline are presented in Table 1. The median age was 56.0 (range, 37–73) years. Most patients had advanced breast (24.1%) or colorectal cancer (16.7%). Among the 52 (96.3%) patients with prior chemotherapy, 23 (42.6%) had previously received taxanes. Of these patients, 9 received paclitaxel and 5 received docetaxel all in the metastatic setting. A further 9 patients received both paclitaxel and docetaxel, 7 in the metastatic setting with a further 2 patients receiving paclitaxel in the metastatic setting and docetaxel in the adjuvant setting. Three of these patients were considered taxane refractory, that is, progressed during taxane treatment (1 patient received 2 mg/kg and 2 patients received 6 mg/kg of aflibercept).

A summary of DLT and patient enrollment is shown in Table 2. At the first aflibercept dose level of 2 mg/kg, 1 patient out of 3 experienced a DLT of neutropenic lung infection, which was considered related to docetaxel. Four additional patients were enrolled at this dose level, with no DLTs observed. The dose escalation continued at 4, 5, and 6 mg/kg without observation of any DLT in the total of 13 patients treated at these 3 dose levels. Three patients then received aflibercept (7 mg/kg); 1 patient with preexisting grade 1 dysphonia at baseline experienced an exacerbation of this symptom at cycle 1 (grade 3). Among the 2 additional patients enrolled at 7 mg/kg, 1 patient experienced hypertension that did not meet DLT criteria, but was considered in the RP2D decision as it was difficult to control compared with hypertension observed at lower dose levels. Similarly, a progressive increase in blood pressure was observed in 1 of 3 patients treated with aflibercept (9 mg/kg). On the basis of those 2 observations of hypertension, the 7 and 9 mg/kg dose levels were not to be recommended for further exploration for phase II dose, then 6 additional patients were enrolled at the immediate lower aflibercept dose level of 6 mg/kg. No DLT was observed.

Therefore, aflibercept (6 mg/kg) in combination with docetaxel (75 mg/m²) administered every 3 weeks was defined as the RP2D. Twenty additional patients were treated with 6 mg/kg in an expansion phase.

All 54 patients experienced 1 or more AE. The most common relevant nonhematologic AEs across all doses were epistaxis (83.3%), stomatitis (75.9%), and dysphonia (68.5%; Table 3). The most frequent grade 3/4 nonhematologic AEs were hypertension (18.5%), stomatitis (16.7%), and fatigue (24.1%).

Most patients experienced hematologic toxicity, with no apparent dose effect (Table 3). The main grade 3/4 hematologic AEs were neutropenia (85.2%) and leukopenia (74.1%). There were 10 patients with complicated neutropenia observed across dose levels: 7 patients had febrile neutropenia and 3 patients had neutropenic infection.

Grade 3/4 liver function abnormalities were observed in 1 patient in the 5 mg/kg group (alanine aminotransferase increase) and in 1 patient in the 6 mg/kg group (alkaline phosphatase increase) and were attributed to underlying disease. In addition, grade 3/4 creatinine CL decrease was observed in 1 patient treated at the 6 mg/kg dose level.

AEs associated with VEGF blockade were observed at all dose levels and included epistaxis (83.3% patients), proteinuria (68.5% patients), dysphonia (68.5% patients), and hypertension (53.7% patients). High blood pressure occurred predominantly in cycle 1 and was well controlled with 1 or 2 antihypertensive therapies. An episode of high blood pressure lasted 1 to 2 days in 65% of cases. Proteinuria, generally measured by spot urinalysis, was observed across the dose levels (2 mg/kg, 71.4%; 4, 5, and 9 mg/kg, 66.7%; 6 mg/kg, 73.3%; 7 mg/kg, 40.0%). Grade 3/4 proteinuria was reported in 1 patient (1 of 6) treated with 5 mg/kg and 2 patients (2 of 30) who received 6 mg/kg. Few arterial or venous thromboembolic events were reported (3 and 2 patients, respectively). No patient experienced a pulmonary embolism, and no cases of gastrointestinal perforations were reported.

Two patients died within 30 days from last study treatment administration, both were caused by disease progression. Ten patients had AEs leading to study treatment withdrawal due to proteinuria (n = 3), thrombosis (n = 2), infections (n = 2), and one each of myalgia, hypersensitivity, and left ventricular dysfunction. Twenty-eight patients had 1 or more dose reduction of docetaxel due to AEs, commonly due to palmar–plantar erythrodysesthesia syndrome, stomatitis, nail disorder, and fatigue.

Patients received a total of 384 infusions of aflibercept at doses ranging from 2 to 9 mg/kg. The median relative dose intensities (RDI) for aflibercept was 0.94 or more at all dose levels except with 7 mg/kg (RDI = 0.92 due to cycle delays) and with 9 mg/kg group (RDI = 0.64 due to cycle delays and dose reductions). The median RDI of docetaxel was lower in the 4 mg/kg group (RDI = 0.58) due to the higher rate of premature discontinuation or cycle delays due to AEs (mainly proteinuria) and lower with 9 mg/kg (RDI = 0.60) due to dose omissions/reductions compared with the other dose groups (RDI ≥ 0.83).

Reasons for treatment discontinuation included disease progression (n = 43), AEs (n = 10), and withdrawal of patient consent (n = 1).

Objective responses were assessed in all 54 patients. A total of 7 partial responses (PR) across all dose levels were observed. PRs occurred in 3 patients with breast cancer (at 2, 4, and 6 mg/kg), and 1 patient each with thymoma, uterus cancer, colon cancer (all at 6 mg/kg), and thyroid cancer (at 9 mg/kg). The longest duration of response was 14.95 months in a patient with undifferentiated thyroid adenocarcinoma previously treated with surgery and adjuvant doxorubicin and cisplatin who received 21 cycles of aflibercept and docetaxel. SD was observed in 32 patients including 18 patients who had a stabilization of the disease for more than 3 months and 7 patients who had SD for more than 10 months.

Free aflibercept exposure increased with increasing dose (Fig. 1), VEGF-bound aflibercept concentrations reached a plateau by cycle 3 at all doses. By cycle 3, the mean free to VEGF-bound aflibercept ratio was higher than 1 at doses above 4 mg/kg (Table 4 and Supplementary Fig. S1) suggesting available circulating VEGF is blocked by aflibercept at doses greater than 4 mg/kg given every 3 weeks. The mean (coefficient of variation, CV%) pharmacokinetic parameters for free and VEGF-bound aflibercept at cycle 1 are summarized in Table 5. From 2 to 9 mg/kg (4.5-fold increase in dose), free aflibercept AUC_0–∞ increased slightly more than expected by dose proportionality [6.9-fold (90% CI: 5.1–9.3) increase]. This supradose proportionality mainly resulted from an increase in AUC from 2 to 4 mg/kg. In contrast, AUC_last of VEGF-bound aflibercept remained constant [1.23-fold (90% CI: 1.04–1.45) increase]. CL of free aflibercept slightly decreased between 2 and 4 mg/kg (1.08–0.814 L/d) then remained constant up to 9 mg/kg (range, 0.814–0.960 L/d). The variability of free and VEGF-bound aflibercept C_max and AUCs was low to moderate. Aflibercept given from 2 to 9 mg/kg every 3 weeks did not affect the pharmacokinetic profile of docetaxel.

All samples were negative for anti-aflibercept antibodies (<LOQ), with the exception of 4 samples in the 6 mg/kg dose group. Two of these 4 samples exhibited a drug-specific response, but with low levels of anti-aflibercept antibodies (∼1,000 mg/mL) which did not seem to affect free aflibercept levels in these patients.

Here, we showed that aflibercept plus docetaxel was generally well tolerated, with toxicities consistent with those reported for docetaxel treatment (16–18) and in line with data from clinical studies of aflibercept monotherapy (13, 14) and other anti-VEGF agents (4–6). AEs associated with VEGF blockade were consistent with prior studies of anti-VEGF agents, included epistaxis, proteinuria, dysphonia, and hypertension. Severe hypertension appeared more common with aflibercept at doses of 7 mg/kg or above; otherwise, VEGF blockade-related AEs were reversible and generally manageable. Two samples specific for anti-aflibercept antibodies were documented, but an effect on free aflibercept levels was not apparent.

During the study, antitumor activity was observed, with 7 patients experiencing PRs and 32 with SD across all dose levels and in different tumor types. Of the 30 patients who received aflibercept (6 mg/kg), PRs were recorded in 4 patients (thymoma, breast, colon, and endometrium) and over half had SD, which lasted more than 3 months in more than one-third of these patients. In the pharmacokinetic analysis, the free to bound aflibercept ratio was more than 1 at the 6 mg/kg dose, suggesting that there is sufficient free aflibercept during the dosing period to be biologically active. In addition, the pharmacokinetics of aflibercept with docetaxel were similar to those observed with each agent given as monotherapy (13, 18, 19). On the basis of DLTs, an acceptable safety profile, pharmacokinetics, and evidence of durable antitumor activity, aflibercept (6 mg/kg) with docetaxel (75 mg/m²) every 3 weeks was selected as the RP2D for further evaluation.

Combining an agent that inhibits VEGF activity with docetaxel has been shown successful in the first-line therapy of patients with metastatic breast cancer (6). In a phase III study, bevacizumab (15 mg/kg) with docetaxel (100 mg/m²) every 3 weeks compared with placebo plus docetaxel failed to show overall survival benefit but showed superior median PFS (10.0 vs. 8.1 months; HR, 0.67; P <0.001) and increased response rates (64% vs. 46%; P <0.001), with little impact on the known toxicity profile of docetaxel. Compared with bevacizumab, aflibercept binds VEGF with an approximately 800-fold higher affinity (dissociation constants of ∼1 pmol/L vs. 0.1–10 nmol/L, respectively), (7, 20) and binds other ligands of VEGFR-1 including proangiogenic PlGFs, (7) and thus may represent a promising new antiangiogenesis treatment. A phase I/II study is ongoing to further investigate the antitumor activity and safety of aflibercept in combination with docetaxel in patients with persistent or recurrent ovarian epithelial cancer, primary peritoneal cancer, or fallopian tube cancer (21).

As a result of this study, aflibercept in combination with docetaxel was tolerated, with evidence of antitumor activity in patients with advanced solid malignancies. Based on these findings, aflibercept at 6 mg/kg plus docetaxel at 75 mg/m² every 3 weeks was selected for phase III evaluation in non–small cell lung cancer and metastatic hormone–resistant prostate cancer.

G. Freyer declared minor honoraria and consultant/advisory board fees from sanofi-aventis; P. Fumoleau declared minor consultant/advisory board fees from GSK, Johnson and Johnson, and sanofi-aventis; S. Assadourian, K. Soussan-Lazard, and S. Ziti-Ljajic are employees of sanofi-aventis.

The authors received editorial support in the preparation of this manuscript provided by Cancer Communications and Consultancy Ltd., funded by sanofi aventis (TCD6120). The authors are fully responsible for the content and editorial decisions for this manuscript.

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.