|Home | About | Journals | Submit | Contact Us | Français|
This study aimed to assess the safety and feasibility of administering volociximab, a chimeric monoclonal antibody that specifically binds to α5β1 integrin, and to determine the pharmacokinetics, pharmacodynamics, and preliminary evidence of antitumor activity.
Patients with advanced solid malignancies were treated with escalating doses of volociximab i.v. administered over 60 minutes. Blood samples were assayed to determine plasma pharmacokinetic parameters, detect human antichimeric antibody formation, and determine the saturation of α5β1 sites on peripheral blood monocytes.
Twenty-one patients received 223 infusions of volociximab at doses ranging from 0.5 to 15 mg/kg i.v. on days 1, 15, 22, 29, and 36; and weekly thereafter. Treatment was well tolerated, and dose-limiting toxicity was not identified over the range examined. Mild (grade 1 or 2), reversible fatigue was the principal toxicity of volociximab at the highest dose levels of 10 and 15 mg/kg. Nausea, fever, anorexia, headache, vomiting, and myalgias were mild and infrequent, and there was no hematologic toxicity. Volociximab had biexponential distribution; clearance was inversely related to increasing dose, and the half-life at 15 mg/kg was estimated as being 30 days. Three patients tested positive for anti-volociximab antibodies. Saturation of monocyte α5β1 integrin sites was dose-dependent up to 15 mg/kg. There was one minor response (renal, 7 months) and one durable stable disease (melanoma, 14 months).
Volociximab can be safely administered at 15 mg/kg i.v. per week. The absence of severe toxicities and preliminary activity at the highest dose level warrants further disease-directed studies.
Volociximab (M200, Eos 200-4) is a high-affinity IgG4 chimeric (82% human, 18% murine) monoclonal antibody (mAb) that specifically binds to α5β1 integrin. Integrins are a superfamily of widely expressed transmembrane glycoprotein receptors for extracellular matrix ligands, such as fibronectin, vitronectin, laminin, collagens, and other plasma membrane proteins, and function in the regulation of a broad variety of cellular processes, including embryogenesis, inflammation, bone metabolism, apoptosis, cell proliferation, angiogenesis, and tumor metastasis (1–3).
Integrins exist as noncovalent heterodimers comprising α and β subunits (4–6). Receptor diversity, function, and versatility in ligand binding is determined by the specific pairing of α and β subunits (7). The cytoplasmic tail of the β subunit links to the actin cytoskeleton and components of the focal adhesion plaque. The interaction with the focal adhesion plaque can lead to signaling, through different pathways, to influence cell survival, growth, and motility. Consequently, all of these protein associations allow cells to sense and respond to their extracellular environment (8).
Endothelial cell expression of the α5β1 integrin and the ligand fibronectin are both up-regulated during tumor angiogenesis (5, 9–11). The sites of the α5β1 integrin increase in expression and are more accessible in the vasculature during angiogenesis and tumor growth, which is in contrast to normal tissue vasculature (12, 13). Disruption of α5β1 integrin binding to fibronectin results in the inhibition of angiogenesis and the induction of apoptosis of activated endothelial cells (5, 14). In preclinical models, selective antagonists targeted to α5β1 integrin inhibit tumor growth (5, 14–16).
Relevant preclinical models for the mechanism of action and antitumor activity evaluation were selected based on the cross-reactivity of volociximab to the nonhuman α5β1 homologues. Volociximab and its parent mouse antibody, IIA1, do not cross-react with murine α5β1 integrin, but do cross-react and block the chicken and cynomolgus monkey target protein (5). Volociximab inhibited human umbilical vein endothelial cells from forming tube-like vessel structures in a three-dimensional fibrin matrix in vitro and was independent of the growth factor stimulus (16). These data suggest that the α5β1 signaling pathways are downstream of growth factor stimulation. Moreover, volociximab inhibited growth factor–stimulated human neonatal foreskin vascular growth when grafted into severe combined immunodeficient mice in vivo (5). In addition, volociximab inhibited vessel formation and human tumor xenograft growth in the chicken chorioallantoic membrane model in ovo (5). Finally, in a preclinical model of choroidal neovascularization in cynomolgus monkeys, volociximab was a potent inhibitor of angiogenesis (16).
The pharmacotoxicologic profile of volociximab was evaluated in cynomolgus monkeys. In the multidose studies, doses of up to 50 mg/kg were shown not to induce mortality or morbidity. Histomorphologically, there were no gross or microscopic findings attributable to treatment, and the no observable adverse effect level was established at the highest dose tested of 50 mg/kg. In cynomolgus monkeys, nonlinear pharmacokinetic parameters were evident with clearance (CL) declining as the dose increased, thereby implying a saturation pharmacokinetic profile (data on file, PDL Biopharma).
The rationale supporting the initiation of a clinical development program with volociximab include the novel mechanism of antiangiogenesis action, the differential expression of the target in tumor vessels compared with normal tissues, and the impressive antiangiogenic activity in preclinical models of angiogenesis.
The principal objectives of this phase I study were to: (a) assess the feasibility of i.v. volociximab administration, (b) evaluate the safety profile, (c) characterize the pharmacokinetic and pharmacodynamic behavior, and (d) describe the preliminary anticancer activity of volociximab in patients with advanced solid malignancies.
Patients with pathologically confirmed solid malignancies refractory to standard therapy, or for whom no standard therapy existed were eligible. The study entry criteria also included: age ≥18 y; life expectancy ≥12 wk; an Eastern Cooperative Oncology Group performance status of =2; previous systemic therapy ≥4 wk (6 wk for prior mitomycin C or a nitrosourea); hemoglobin ≥9 g/dL; absolute neutrophil count ≥1,500/μL; platelet count ≥100,000/μL; bilirubin ≤1.5 mg/dL; aspartate serum transferase, alanine serum transferase, and alkaline phosphatase ≤3 times the upper limit of normal or ≤5 times upper limit of normal if hepatic metastases present; serum creatinine ≤2.0 mg/dL; measurable disease; no evidence of brain metastases; no evidence of HIV, hepatitis B or C seropositivity; and no coexisting severe medical conditions or anticoagulation therapy. Patients were excluded if they had a history of hypersensitivity reactions to other mAb therapies or had a positive test for anti-volociximab antibodies before study enrollment. Pregnant women or patients with a history of any bleeding disorder, thromboembolic events, or major surgical procedure within 4 wk of treatment were also excluded. Patients gave written informed consent according to federal and institutional guidelines before any study procedures.
Volociximab was administered i.v. over 60 min at dose levels of 0.5 to 15 mg/kg. Patients initially received five infusions of volociximab. The first two doses were separated by a 2-wk period to allow for single-dose pharmacokinetic sampling, whereas the remaining three doses were given weekly. Patients who obtained clinical benefit were eligible to continue therapy; consisting of eight weekly doses of volociximab per course for up to seven additional courses of therapy. An accelerated dose-escalation design was initially used, allowing advancement of single-patient cohorts to the next higher dose level if no ≥grade 2 toxicity occurred. In the event that ≥grade 2 toxicity was reported, a more conservative dose-escalation scheme was to be used. If one patient experienced a dose-limiting toxicity (DLT), the cohort was expanded to six patients. The maximum tolerated dose was defined during course 1 as the highest dose at which <2 of 6 patients experienced a DLT. Toxicities were graded according to the National Cancer Institute’s Common Toxicity Criteria, Version 3.0. DLT was defined as any grade 3 or 4 adverse event (including grade ≥3 nausea or vomiting despite optimal antiemetics).
Volociximab was supplied in 150-mg (20-mL) single-use vials by PDL Biopharma, Inc. The composition of each vial was 10 mg/mL of volociximab, 25 mmol/L citrate, 150 mmol/L sodium chloride, and 0.05% polysorbate (Tween) 80 (pH 6.5). Each dose was administered in a fixed total volume of 120 mL, by adding sodium chloride for injection, U.S. Pharmacopeia (0.9%).
A complete medical history, physical examination, and routine laboratory studies were done before treatment and weekly thereafter. Routine laboratory studies included a complete blood count, differential white blood count, serum chemistry including electrolytes, blood urea nitrogen, creatinine, uric acid, glucose, alkaline phosphatase, alanine serum transferase, aspartate serum transferase, total bilirubin, calcium, amylase, and urinalysis. Pretreatment assessments also included prothrombin time, partial thromboplastin time, an electrocardiogram, relevant radiologic studies, and tumor markers. Radiologic studies for disease status were done after the initial five infusions of volociximab and then repeated for 8 wk. Response was assessed by Response Evaluation Criteria in Solid Tumors guidelines.
Serum samples for the determination of volociximab concentrations were obtained immediately before and at the end of the infusion, and at 4, 24, 48, and 168 h after completion of the first dose, and immediately before, at the end of the infusion, and 4 h after each infusion on days 15, 22, 29, and 36; as well as on days 38, 43, and 81 after discontinuing study drug.
The concentration of volociximab was determined using a validated ELISA. Briefly, a fusion protein containing the extracellular domains of human α5β1 heterodimer and human Fc domain was used as the capture reagent for volociximab. Assay calibrators were prepared by spiking pooled normal human serum with volociximab at 0, 25, 50, 100, 250, 500, 1,000, and 2,500 ng/mL, then aliquoted and frozen. Quality control samples were prepared similarly at 0, 100, 400 and 1,000 ng/mL. Calibrators, controls, and validation samples were incubated on fusion protein-coated plates. Detection was accomplished using a biotin-conjugated mouse antihuman IgG4 and streptavidin-horseradish peroxidase conjugate. Tetramethylbenzidine was used as substrate and absorbance was read at 650 nm. A 4-parameter logistic calibration curve was plotted using the corrected absorbance values (mean absorbance of zero calibrator subtracted from mean absorbance of a calibrator). Quality control and sample absorbance values were similarly corrected using the negative quality control and prestudy baseline samples, respectively. Concentrations of quality control and study samples were determined by interpolation from the calibration curve. The validated quantitative range of the assay was established as 100 ng/mL to 1,500 ng/mL in human serum.
The maximum observed concentrations (Cmax) were determined by inspection of plasma concentration-time data, and the area under the curve up to last measurable time points (AUC0-last) was calculated using the log trapezoidal rule (WinNonlin Professional 4.0, Pharsight, Inc.). Pharmacokinetic parameters were obtained from the individual post hoc estimates of a population pharmacokinetic model developed using the Monte Carlo parametric expectation method implemented in SADAPT software (17). A two-compartment model with constant rate infusion and parameterized in terms of CL, central volume of distribution (V1), intercompartmental clearance (Q), and peripheral volume of distribution (V2), incorporating between-patient variability with respect to CL, V1, Q, and V2 was used to describe the pharmacokinetics of volociximab. A plasma pharmacokinetic sample with below-quantification level was handled as a fixed-point censored concentration and included in the model analysis according to the method proposed by Beal et al. (18). Dose proportionality was assessed by evaluating dose independence of post hoc estimates of CL, V1, Q, and V2 using a power model [Ln(PK) = β0 + β1*Ln(dose), where PK is the pharmacokinetic parameter, and β0 and β1 are the intercept and slope or power model factor of the linear regression line, respectively] in a manner analogous to the method proposed by Smith et al; the 90% confidence intervals of β1 with range of −0.07 to 0.07 (which corresponds to a 90% confidence interval for the ratio of parameter at the highest dose to that of the lowest dose being contained within the range of 0.8 to 1.25) were used as the criteria for dose independence of each parameter over the entire range of doses administered in this study (19).
Serum samples for anti-volociximab antibodies were obtained from subjects before treatment, before the second dose, and at 45 d after the last dose of volociximab (study exit). In addition, selected serum samples collected before dosing at days 22, 29, 36, 38, and 43 from three subjects at the two lowest doses were tested for the presence of antidrug antibodies. Samples that tested positive for antibodies to volociximab were subsequently tested for neutralizing antibodies to volociximab.
The immunogenicity of volociximab was determined using a double-antigen bridging ELISA. Briefly, ELISA plates were coated with volociximab antibody to capture any anti-volociximab–specific human antibodies from the serum. Bound human anti-volociximab antibody was then detected by its bridging of biotinylated volociximab followed with streptavidin-horseradish peroxidase conjugate. Tetramethylbenzidine was used as substrate and absorbance was read at 450 nm after the enzymatic reaction was stopped by the addition of H2SO4. A murine IgG1 mAb specific for volociximab (anti-id) served as a positive control, and the assay sensitivity was 5 ng/mL anti-id equivalent.
To determine the saturation of α5β1 integrin sites by volociximab on normal circulating monocytes, samples of peripheral blood were taken before the first dose, on days 2 and 8, just before each subsequent dose, and on day 43. The saturation of α5β1 sites was determined by flow cytometry using CD14 antibodies to identify monocytes and adding labeled antihuman IgG4 to quantify bound volociximab on their surface. Unoccupied α5β1 integrin sites were detected by adding labeled murine antihuman α5β1 antibody (IIA1).
Twenty-one patients, whose pertinent demographic characteristics are displayed in Table 1, received a total of 223 infusions of volociximab at doses ranging from 0.5 to 15 mg/kg. The total number of new patients treated and the number of courses at each dose level, as well as the overall dose escalation scheme, are shown in Table 2. All patients enrolled in the study received the initial five infusions of volociximab, except one patient who discontinued therapy due to disease progression before receiving the fifth infusion of study drug. Six patients, who obtained clinical benefit, continued treatment consisting of eight weekly doses of volociximab per course, for up to seven additional courses of therapy. During this extension phase, the median number of courses administered per patient was 2 (range, 1–7).
Table 3 summarizes the adverse events at least possibly related to volociximab occurring in any course of treatment. No DLT was observed and no dose reductions were required. The most common adverse events noted were low-grade constitutional symptoms (fatigue, anorexia, fever, arthralgias, and myalgias), gastrointestinal symptoms (nausea and vomiting), headache, edema, and hypertension. Fatigue and myalgias were more common in the higher-dose groups. Routine premedication was not used, and of the 223 infusions, only 1 was interrupted due to an infusion reaction. The patient experienced facial warmth and dysesthesia 3 minutes after the start of the day 15 infusion. The volociximab infusion was interrupted; acetaminophen, diphenhydramine, ranitidine, and dexamethasone were administered; and the infusion was resumed and completed without further signs of a reaction. The patient tolerated subsequent infusions with the use of premedication. There was neither hematologic toxicity nor infectious complications.
Twenty-one patients underwent plasma sampling in the first course for pharmacokinetic analysis. Mean plasma volociximab concentrations versus time curves are shown in Fig. 1. The pharmacokinetic estimates of volociximab by compartmental modeling at each dose level are listed in Table 4. Interpatient variation for the CL, V2, Q, and volume of distribution at steady state (Vss) of volociximab was large. The CL of volociximab ranged from 0.9 to 141.6 mL/h and appeared to decrease with increasing dose from a mean of 141.6 mL/h following the 0.5 mg/kg dose to 9.7 mL/h following the 15 mg/kg dose. The V2 ranged from 826.3 to 12,561 mL, whereas the Vss ranged from 2,770.9 to 15,641.8 mL. At these doses, the V1 approximated the plasma volume and did not change with dose, with means ranging from 2,275.3 to 4,015.1 mL. The mean terminal half-life (t 1/2) of volociximab was 789.9 hours (32.9 days) for the 10 mg/kg dose and 730 hours (30.4 days) for the 15 mg/kg dose. Too few washout pharmacokinetic samples were collected and several samples in the terminal phase of the lower dose groups had values below assay quantitative limit. Therefore, only half-life estimates in the higher-dose groups (10 and 15 mg/kg) can be reported with confidence.
The dose-dependency of volociximab was assessed using the power model analysis. The estimate of power model factor β1 for CL was −0.70 (90% confidence interval, −0.23 to −1.12) indicating that the CL was not proportional to the dose. The power factor for V1, Q, and V2 were −0.07 (90% confidence interval, −0.18 to 0.04), 0.29 (−0.19 to 0.77), and −0.03 (−0.32 to 0.26), respectively, suggesting that these parameters were independent of dose.
Three of the 21 patients (14.3%) tested positive for human antichimeric antibodies. All three subjects also tested positive for neutralizing antibodies.
Volociximab seemed to bind and saturate the free α5β1 integrin on monocytes in a dose-dependent manner. Monocytes were not saturated during the treatment period in subjects dosed with <1 mg/kg of volociximab, whereas monocyte saturation was achieved in the 2.5-, 5-, 10-, and 15-mg/kg dose groups. However, compared with the subjects in the 2.5-, 5-, and 10-mg/kg dose groups, the subjects in the 15-mg/kg dose group rapidly achieved, and consistently maintained, monocyte saturation throughout the treatment period. The mean percent free monocyte α5β1 integrin receptor levels for subjects in the 15-mg/kg dose group were consistently below 5% throughout the treatment period (1.9, 2.5, 3.0, 1.3, 2.3, 1.7, and 1.5 for days 2, 8, 15, 22, 29, 36, and 43, respectively). The mean percent free α5β1 integrin on monocytes on the last measurement day (day 43) were 62.4, 1.7, 1.3, 0.9, and 1.5 for the 1-, 2.5-, 5-, 10-, and 15-mg/kg dose groups, respectively.
A patient with metastatic renal cell carcinoma, refractory to sunitinib, had a minor response. Five patients, who had malignancies with documented progressive disease before study entry, experienced durable stable disease that lasted for at least 4 months. One of these patients who had metastatic melanoma with hepatic metastases had stable disease for 14 months.
Tumor angiogenesis is a complex multistep process involving diverse processes that include inciting triggers associated with tumor proliferation and hypoxic stimuli, growth factor stimulation of receptors, endothelial cell proliferation, and both cell-cell and cell-extracellular matrix interactions. The α integrins are critical proteins that are involved in the regulation, proliferation, survival, and signaling functions of normal and tumor-associated endothelial cells and are distinct from vascular endothelial growth factor signaling pathway. Furthermore, integrins act in concert with vascular endothelial growth factor to regulate, orchestrate, and amplify angiogenesis through signaling events involved in endothelial cell migration, proliferation, invasion, and anoikis resistance (20, 21). Integrin-mediated signal transduction pathways regulate the endothelial cell cytoskeleton to coordinate the complex process of vascular morphogenesis and organize into multicellular tubes with functional lumens (5, 22). Fibronectin and its major receptor, α5β1 integrin, have a central role in this successful vascular development and therefore represent an attractive target for antiangiogenesis therapy in malignancies. Volociximab represents the first mAb in clinical development that specifically binds α5β1 integrin (16).
The starting dose for this phase I study, 0.5 mg/kg, represented a very conservative dose and was based on the absence of homology of the target in humans from the preclinical models used for toxicology. In the current study no evidence of a DLT was observed despite a 30-fold increase in the doses evaluated, confirming the feasibility and safety of the dose range tested. A constellation of mild or modest non-hematologic adverse events were noted, which included fatigue, nausea, anorexia, fever, arthralgias, myalgias, and headache. Albeit α5β1 integrin is expressed on normal peripheral blood monocytes, there were no clinically apparent infectious complications associated with volociximab (23).
Although volociximab is a chimeric antibody, hypersensitivity reactions were rare. Human antichimeric antibodies were detected in three patients, and only one patient experienced a modest hypersensitivity reaction. Based on these findings, routine premedication is unnecessary.
At the dose range examined in this study, volociximab showed nonlinear pharmacokinetic parameters consistent with a target-saturation effect. The estimates of volociximab CL declined with increasing doses. The target, α5β1 integrin, is expressed on normal peripheral blood monocytes, on the extracellular membrane of some tumor cells, as well as exposed within neovasculature (23). As such, the process of binding and saturation of target sites seems to account for the increased CL at lower doses (24). Consistent with these findings, the saturation of free α5β1 integrin sites on monocytes by volociximab was dose-dependent and was uniformly diminished to a mean value of <2% at the highest dose of 15 mg/kg. These results, taken together, may have important implications. Dose-reductions from the highest dose determined in this study may result in increased CL of volociximab, as well as incomplete saturation of all α5β1 integrin target sites.
The study was initially designed to estimate the pharmacokinetic parameters of volociximab by fitting individual data with a uniform sampling regimen for all subjects. However, a few subjects missed the last sample collection on days 81 as a result of disease progression or enrollment into the extension study. In addition, two subjects at the lower doses had insufficient samples to support the fitting of the individual data for pharmacokinetic parameter estimation. Therefore, data from all patients were pooled and analyzed using the mixed-effect modeling approach (17, 24). An attempt to fit the two-compartment pharmacokinetic model with both linear and nonlinear elimination to the data was unsuccessful, and the parameters associated with the nonlinear elimination were poorly estimated. Therefore, the results with a two-compartment linear pharmacokinetic model were reported. The final estimate value of V1 of volociximab closely resembles the plasma volume in humans, as is expected for a high molecular weight protein and consistent with the values reported for other therapeutic chimeric mAb (25, 26). The mean terminal t1/2 of volociximab was ~30 days for subjects in the 10- and 15-mg/kg dose group and is consistent with the t1/2 of natural IgG4 in humans (18). The dose-dependent CL values suggest that the nonlinear pharmacokinetic of volociximab is probably due to a saturated receptor-mediated CL mechanism; and also due to a human antichimeric antibody–mediated clearance pathway, because the three patients with high CL values in the lower-dose groups tested positive for human antichimeric antibodies (17).
Preliminary but encouraging activity was seen in two patients in the 15-mg/kg group. One patient with renal cell carcinoma refractory to sunitinib had a reduction in the size of pulmonary lesions and remained on study for 7 months, and a patient with visceral metastases from melanoma had durable stable disease for 14 months. The appearance of durable stable disease is intriguing. In contrast to most tumor cells, α5β1 integrin is expressed on the surface of some melanoma tumor cells (27, 28). The durable stable disease may be, at least in part, the result of direct action of volociximab on tumor cells, in addition to the intended antiangiogenesis effect.
In conclusion, volociximab can be safely administered to patients with advanced malignancies with doses of up to 15 mg/kg weekly for prolonged periods. To fully evaluate the role of volociximab in anticancer therapy, either as a single agent, in combination with chemotherapy, or with other antiangiogenesis inhibitors, additional disease-directed studies should be undertaken.
Tumor angiogenesis is a multistep process. This is the first phase I study report of a chimeric IgG4 antibody (volociximab) targeting specifically α5β1 integrin, a member of the integrin superfamily consisting of trans-membrane glycoprotein receptors for extracellular matrixligand. Volociximab could be safely administered at doses up to 15 mg/kg weekly, demonstrated a saturation pharmacokinetic profile, and had preliminary evidence of antitumor activity in one patient with renal cell carcinoma. Pharmacodynamic evaluation demonstrated that the 15 mg/kg dose level saturated α5β1 integrin sites on a normal tissue surrogate (peripheral blood mononuclear cells). In conclusion, targeting one aspect of the multistep process of angiogenesis, α5β1 integrin by volociximab is a feasible and safe approach.
Grant support: PDL Biopharma, Inc.
Presented in part at the European Organisation for Research and Treatment of Cancer-National Cancer Institute-AACR International Conference on Molecular Targets and Cancer Therapeutics, Geneva, Switzerland, September 28–October 1, 2004.
Disclosure of Potential Conflicts of Interest
S. Yazji is employed by PDL Biopharma; C. Ng is employed by Bristol Myers Squibb.