Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Angiogenesis plays a crucial role in many physiological and pathologic processes including tumor growth, rheumatoid arthritis, and diabetic retinopathies (Aiello et al., 1994; Fava et al., 1994; Ferrara, 1995). Of the growth factors implicated in tumor angiogenesis, vascular endothelial growth factor (VEGF) appears to be the most selective mitogen for endothelial cells. The growth of solid tumors is dependent on angiogenesis for the supply of nutrients and for the removal of metabolic waste products (Folkman, 1990, 1995). Elevated levels of VEGF have been reported in the tumor cytosol of breast (Brown et al., 1995; Gasparini et al., 1997), ovarian (Schlaeppi et al., 1996), and brain (Takano et al., 1996) tumors relative to surrounding nontumor tissue. Suppression of tumor growth (Kim et al., 1993; Asano et al., 1995; Warren et al., 1995) and resolution of retinal neovascularization (Adamis et al., 1996) have been demonstrated in animal models after administration of antibodies against VEGF.

Recombinant humanized monoclonal antibody (rhuMAb) VEGF is a recombinant humanized monoclonal antibody composed of the consensus human IgG₁ framework regions and antigen-binding, compliment-determining regions from a murine monoclonal antibody (A.4.6.1) (Presta et al., 1997). The monoclonal antibody blocks binding of VEGF to its receptors. Previous studies have shown that rhuMAb VEGF binds to primate VEGF and with lower affinity to rabbit VEGF but not to rat or mouse VEGF (N. Ferrara, unpublished data). To support clinical testing of rhuMAb VEGF, pharmacokinetic studies were conducted in mice, rats, and cynomolgus monkeys after i.v. and s.c. administration. The pharmacokinetic data in animals were used to predict the disposition of rhuMAb VEGF in humans using allometric scaling. In addition, the tissue distribution of ¹²⁵I-rhuMAb VEGF after i.v. administration was examined in rabbits to identify target organs and routes of elimination.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. rhuMAb VEGF, manufactured using recombinant DNA technology, was expressed in a genetically engineered Chinese hamster ovary cell line. The protein was available as a clear to slightly opalescent, sterile liquid at a concentration of 10 mg/ml.

Animal Husbandry. Female nude mice (19-25 g) were housed as a group. Male Sprague-Dawley rats (297-345 g) and New Zealand White rabbits (1.1-1.3 kg) were housed individually. Male cynomolgus monkeys (Macaca fascicularis), weighing between 3.0 and 3.9 kg, were acclimated for 3 to 4 weeks. Access to food and water was provided ad libitum to all animals. All studies were approved by the Institutional Animal Care and Use Committee and were performed in accordance with the guidelines of the American Association for Accreditation of Laboratory Animal Care.

Intravenous and s.c. Administration of rhuMAb VEGF to Rats and Mice. Mice (n = 30) were dosed with 9.3 mg/kg rhuMAb VEGF either as a single i.v. or s.c. injection. Intravenous injections were administered via a tail vein, and s.c. injections were administered in the flank. Serum was harvested on day 1 predose and at 5 min and 1, 2, 4, and 8 h, on day 2 at 0 and 8 h, and once on days 3, 4, 5, 6, 9, 12, and 15. At each sampling time two mice were sacrificed, blood was collected via cardiac puncture, and serum was harvested. Rats were randomized to three treatment groups (three animals/group). Animals received 0.66 or 10 mg/kg rhuMAb VEGF as a single i.v. bolus injection via a femoral vein catheter or 10 mg/kg as a single s.c. injection in the flank. Blood samples (0.4 ml) were collected at various times via a jugular vein cannula. Samples were collected before dosing on day 1, and at 5, 15, and 30 min and 1, 4, and 6 h postdose, on day 2 at 0 and 8 h, and once daily on days 3, 4, 5, 8, 9, 10, 12, and 15. Serum was harvested and stored at less than 60°C until assayed for rhuMAb VEGF concentrations.

Intravenous and s.c. Administration of rhuMAb VEGF to Cynomolgus Monkeys. Animals were randomized into four groups (four animals/group). Monkeys received a single i.v. injection of 2, 10, or 50 mg/kg rhuMAb VEGF into the sapheneous vein or a 10 mg/kg s.c. injection in the dorsal cervical region. Blood samples (~1 ml) were collected from the femoral vein predose, 5-, and 30-min and 1-, 2-, 4-, 8-, 12-, 18-, 24-, and 36-h postdose, and daily on days 3 to 16 and on days 18, 21, 24, 27, and 30. Serum was harvested and stored at less than 60°C until analyzed. Additional serum samples were collected twice pretreatment and on day 30 for measurement of antibodies to rhuMAb VEGF.

Iodination of rhuMAb VEGF and rhuMAb E25 for Tissue Distribution. rhuMAb VEGF and a humanized isotypic control antibody directed against IgE, rhuMAb E25, were radiolabeled with sodium 125iodide by the lactoperoxidase method. Briefly, 20 µg of each antibody was labeled in sodium acetate buffer, pH 5.5, with lactoperoxidase (1 U/ml) and 2 mCi sodium ¹²⁵iodide. The reaction was initiated by the addition of 15 µl of H₂O₂ diluted 1:174,000. After a 5-min incubation at room temperature, an additional 15 µl of H₂O₂ was added, and the reaction was stopped 5 min later with the addition of 15 µl of N-acetyl-L-tyrosine (20 mM). The labeled proteins were separated from unincorporated sodium ¹²⁵iodide using a PD-10 size-exclusion column (Pharmacia, Uppsala, Sweden). SDS-polyacrylamide gel electrophoresis (PAGE) was performed on the radiolabeled materials to ensure that there were no degradation products present (data not shown). Radiolabeled rhuMAb VEGF and rhuMAb E25 were >95% trichloroacetic acid (TCA)-precipitable and had a specific activity of 118 and 91 µCi/µg, respectively.

Intravenous Administration of ¹²⁵I-rhuMAb VEGF and ¹²⁵I-rhuMAb E25 to Rabbits. Eight rabbits were assigned to one of two groups (four animals/group). Each group received a trace dose of either ¹²⁵I-rhuMAb VEGF or ¹²⁵I-rhuMAb E25. Rabbits received a single i.v. bolus dose of 464 to 652 µCi/kg of the labeled material via an ear vein. Sodium iodide (15 mg) was administered i.p. 48, 24, and 0.5 h before administering the radiolabeled material to minimize uptake of ¹²⁵I to the thyroid (Nakajo et al., 1983). Two rabbits from each group were sacrificed at 2 h and another two animals in each group were sacrificed 48 h after administration of the radiolabeled material. Blood was collected from the ear vein contralateral to the dosing ear at 2 h from all animals. In addition, blood was collected at 5, 8, 24, 32, and 48 h from rabbits sacrificed at 48 h. Plasma was harvested and stored frozen at less than 70°C until assayed for total and TCA-precipitable radioactivity.

Analysis of Rabbit Tissue Samples. Tissue sections were weighed and the total radioactivity per gram of tissue was measured in a gamma counter (Packard MinAxi Auto Gamma 5000 series). Partially frozen tissues were minced and homogenized in lysis buffer (PBS, 20 mM EDTA, 1% Triton X-100) using a probe-type tissue homogenizer (Tekmar Tissumizer). The tissue slurry was centrifuged (2000g for 20 min, 8°C) and the tissue homogenate supernatant was stored at less than 60°C until TCA precipitation.

Characterization of ¹²⁵I-rhuMAb VEGF and ¹²⁵I-rhuMAb E25 in Tissues by SDS-PAGE. To determine the stability and processing of ¹²⁵I-rhuMAb VEGF and ¹²⁵I-rhuMAb E25 in the tissues over time, 800 µl of tissue homogenate supernatant was immunoprecipitated with 33 µl of protein A-Sepharose slurry (P-3391; Sigma, St. Louis, MO). After centrifugation, the protein A-Sepharose pellet was resuspended in 40 µl of SDS sample loading buffer, boiled, and centrifuged. Thirty microliters of the resulting supernatant was subjected to SDS-PAGE on a 6% Tris-glycine gel (Novex, San Diego, CA). Plasma samples collected at specified times and urine samples collected at the time of sacrifice were also characterized by this procedure without immunoprecipitation.

rhuMAb VEGF and anti-rhuMAb VEGF Antibody Enzyme-Linked Immunosorbent Assays (ELISAs). Serum samples from mice, rats, and cynomolgus monkeys were analyzed for the immunoreactive rhuMAb VEGF concentration by an ELISA. ELISA plates were coated with 0.25 µg/ml recombinant VEGF₁₆₅ in 50 mM sodium carbonate buffer, pH 9.6, at 4°C, incubated overnight, and then blocked with 0.5% BSA in 8 mM Na₂HPO₄, 1.5 mM KH₂PO₄, 2.7 mM KCl, and 137 mM NaCl, pH 7.2 (PBS) at room temperature for 1 h. Standards and four 2-fold serial dilutions of serum samples were prepared in PBS containing 0.5% BSA and 0.05% polysorbate 20, and were incubated on the plate for 1 h. Standards in 10% mouse or rat serum were used for assaying pretreatment mouse or rat samples at 1:10 dilution. Bound rhuMAb VEGF was detected using horseradish peroxidase (HRP)-labeled mouse anti-human IgG (Fc) (Jackson ImmunoResearch, West Grove, PA) for mouse and rat samples and goat anti-human IgG (Fc) for cynomolgus samples. The substrate used was 3,3',5,5'-tetramethyl benzidine (Kirgaard & Perry Laboratories, Gaithersburg, MD). Absorbance was read at 450 nm on a V_max plate reader (Molecular Devices, Menlo Park, CA). A standard curve was fit using nonlinear regression. The detection limit of this assay was 7.8 ng/ml rhuMAb VEGF in mouse and rat serum and 3.9 ng/ml in cynomolgus monkey serum.

Pharmacokinetic Analysis. The immunoreactive concentration versus time data from mice, rats, and cynomolgus monkeys were analyzed using a two-compartment model for the i.v. doses and a one-compartment model for s.c. dosing (WinNonlin, Apex, NC). The clearance, initial-, and steady-state volumes of distribution, half-lives, mean residence time, and bioavailability were estimated using standard methods (Gibaldi and Perrier, 1982).

Allometric Scaling. Pharmacokinetic data from mice, rats, and cynomolgus monkeys after i.v. administration were used to predict the pharmacokinetics of rhuMAb VEGF in humans using allometric scaling methods (Boxenbaum, 1982; Mordenti, 1986; Chappell and Mordenti, 1991). Pharmacokinetic data can be scaled using an empirical power function of the species body weight

(1)

where Y is the pharmacokinetic parameter of interest (e.g., clearance), W is the body weight of the species, a is the allometric coefficient, and b is the allometric exponent. The allometric coefficient assumes the value of the pharmacokinetic parameter when the body weight is unity.

(2)

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Pharmacokinetics of rhuMAb VEGF in Mice and Rats. The pharmacokinetics of rhuMAb VEGF in mice and rats are summarized in Table 1. After i.v. bolus administration, rhuMAb VEGF concentrations were cleared from the serum in a biphasic manner with an initial half-life of 1.2 h in mice and approximately 7 h in rats (Fig. 1). The terminal half-life was 1 to 2 weeks and was dominant, accounting for >90% of the total area under the curve (AUC). After an i.v. dose of approximately 10 mg/kg rhuMAb VEGF, the clearance was 15.7 ml/day/kg in mice and 4.83 ml/day/kg in rats. In rats, a dose-dependent clearance was noted. At the 0.66-mg/kg dose, the clearance in rats was 8.37 ml/day/kg. After s.c. administration, peak concentrations of rhuMAb VEGF in the serum were approximately half those noted after i.v. administration of a similar dose. In mice, complete bioavailability was estimated after s.c. administration, whereas in the rat comparison of the AUC after the 10-mg/kg s.c. and i.v. dose indicated a bioavailability of 69%.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   rhuMAb VEGF serum concentration versus time profile after i.v. and s.c. administration in mice (A) and rats (B). For mice, each symbol represents an individual animal. For rats, mean ± S.D. of three animals is presented. The curves represent the simulated rhuMAb VEGF concentration based on the mean pharmacokinetic parameters.

Pharmacokinetics of rhuMAb VEGF in Cynomolgus Monkeys. The mean rhuMAb VEGF serum concentration versus time data is shown in Fig. 2. Table 2 summarizes the pharmacokinetics of rhuMAb VEGF after i.v. and s.c. administration in cynomolgus monkeys. After i.v. bolus injection, an approximately dose-proportional increase in the peak concentration was noted between 2 and 50 mg/kg. The clearance of rhuMAb VEGF after i.v. dosing ranged from 4.81 to 5.59 ml/day/kg and did not depend on dose. rhuMAb VEGF was cleared from the serum in a biphasic manner, with an initial half-life of 11 to 26 h and a terminal half-life of 1 to 2 weeks. The terminal phase was dominant, comprising >89% of the total AUC. The clearance was approximately 5 ml/day/kg and the steady-state volume of distribution was ~70 ml/kg.

View larger version (25K): [in this window] [in a new window]   Fig. 2.   rhuMAb VEGF serum concentration versus time profile after i.v. and s.c. administration in cynomolgus monkeys. Each symbol is the mean ± S.D. of four animals. The curves represent the simulated rhuMAb VEGF concentration based on the mean pharmacokinetic parameters.

Allometric Scaling. The results of the allometric regression are tabulated in Table 3 and shown graphically in Fig. 3. The predicted pharmacokinetics of rhuMAb VEGF in humans are listed in Table 3. Based on allometric scaling, it is predicted that rhuMAb VEGF has a clearance of 2.4 ml/day/kg and an elimination half-life of 12 days. Also presented in Table 3 are the observed clearance and terminal half-life of rhuMAb VEGF from a phase I study in cancer patients (Gordon et al., 1998).

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Allometric scaling of rhuMAb VEGF clearance, initial- and steady-state volume of distribution, and elimination half-life versus species body weight. Each line represents the unweighted least-squares regression of the log-transformed data.

Tissue Distribution of ¹²⁵I-rhuMAb VEGF in Rabbits. Radioactivity in the plasma samples was >76% TCA precipitable in all samples (data not shown). Plasma TCA-precipitable radioactivity decreased by nearly 2.5-fold between 2 and 48 h for both ¹²⁵I-rhuMAb VEGF and ¹²⁵I-rhuMAb E25 (Fig. 4). Plasma concentration versus time profiles suggest that ¹²⁵I-rhuMAb VEGF and ¹²⁵I-rhuMAb E25 have similar pharmacokinetics for the first 48 h after i.v. administration in rabbits. Less than 10% of the radioactivity in the urine at both the 2- and 48-h time points was TCA-precipitable.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   TCA-precipitable radioactivity in plasma after i.v. administration of 125I-rhuMAb VEGF or 125I-rhuMAb E25. Data for individual animals are presented.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Quantitation of TCA-precipitable radioactivity in tissues 2 and 48 h after i.v. dosing of 125I-rhuMAb VEGF or 125I-rhuMAb E25. Values are represented as a percentage of administered dose that is TCA-precipitable per gram of wet tissue. Data for individual animals are presented.

Characterization of Radioactivity In Vivo. Electrophoretic analysis of ¹²⁵I-rhuMAb VEGF and ¹²⁵I-rhuMAb E25 plasma samples (Fig. 6) indicated that the administered drug remained as a single band with minimal degradation products of lower molecular weight for as long as 48 h. This suggests that the TCA-precipitable radioactivity detected in the 2- and 48-h plasma samples represents mostly intact ¹²⁵I-rhuMAb VEGF and ¹²⁵I-rhuMAb E25. SDS-PAGE of urine samples failed to reveal any trace of labeled material at 2 and 48 h.

View larger version (39K): [in this window] [in a new window]   Fig. 6.   SDS-PAGE of plasma and urine samples after administration of 125I-rhuMAb VEGF (A) and 125I-rhuMAb E25 (B). Plasma samples obtained at various sampling times and urine obtained at animal sacrifice were subjected to SDS-gel electrophoresis. Molecular mass standards with sizes in kDa are shown on the left.

View larger version (58K): [in this window] [in a new window]   Fig. 7.   SDS-PAGE of tissue homogenate supernatants after administration of 125I-rhuMAb VEGF. A and B show the 2- and 48-h samples for rhuMAb VEGF, respectively. Molecular mass standards with sizes in kDa are shown on the left.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Tumor metastasis and tumor growth beyond a microscopic size depends on the formation of an adequate blood supply (Folkman, 1990). Although several growth factors have been reported to provide the angiogenic signal necessary for the development of a blood supply, VEGF appears the most specific for endothelial cells. Several studies with tumor models in animals have demonstrated a suppression in tumor growth after treatment with a monoclonal antibody against VEGF (Kim et al., 1993; Asano et al., 1995; Warren et al., 1995). There is increasing interest in the use of antiangiogenic therapy for the treatment of cancer. Antiangiogenic therapy potentially could result in fewer side effects than current treatment or could be combined with chemotherapy, possibly allowing dose reduction of the chemotherapy. To support investigation of rhuMAb VEGF in humans, the pharmacokinetics of rhuMAb VEGF were investigated after i.v. and s.c. dosing in mice, rats, and cynomolgus monkeys. In addition, the tissue distribution of ¹²⁵I-rhuMAb VEGF after i.v. administration was examined in the rabbit.

rhuMAb VEGF exhibited multicompartmental pharmacokinetics, consistent with several other IgG antibodies (Matsuzawa et al., 1992; Fox et al., 1996). The clearance of rhuMAb VEGF was comparable to the clearance of other humanized monoclonal antibodies in animals. Davis et al. (1995) reported that RSHZ19, a respiratory syncytial virus-specific reshaped human monoclonal antibody (IgG₁ framework), had a clearance of ~7.2 ml/day/kg in rats and ~3.4 ml/day/kg in cynomolgus macaques. Similar clearance values for humanized antibodies have been noted by others (Hakimi et al., 1991; Fox et al., 1996). The initial volume of distribution of rhuMAb VEGF was smaller than serum volume in all species studied (Davies and Morris, 1993), suggesting limited extravascular distribution. The distribution of rhuMAb E25 similarly was limited to the vascular space. Complete bioavailability of rhuMAb VEGF was noted after s.c. administration, and peak concentrations were approximately half those after i.v. bolus dosing of a comparable dose. No anti-rhuMAb VEGF IgG-type antibodies were noted in the serum after a single administration of rhuMAb VEGF to cynomolgus monkeys.

Table 3 presents a comparison of the predicted and observed pharmacokinetics of rhuMAb VEGF in humans. Based on allometry, a clearance of 2.4 ml/day/kg was predicted, comparable to the value (4.3 ml/day/kg) noted in cancer patients (Gordon et al., 1998). Allometric scaling has been used successfully to predict the pharmacokinetics of new chemical entities (Chappell and Mordenti, 1991) and has been applied to a limited extent to recombinant proteins, including antibodies (Mordenti et al., 1991; Lave et al., 1995). Allometric scaling has been applied successfully with other antibodies. The pharmacokinetics of CD4-IgG in humans were predictable using data obtained in animals (Mordenti et al., 1991). The clearance predicted from allometric scaling was 2.6 ml/min compared to an observed value of 2.62 ml/min in humans (Mordenti et al., 1991). Similarly, the biodistribution of monoclonal antibodies has been scaled from mice to humans using a physiological-based pharmacokinetic model (Baxter et al., 1995).

Assumptions underlying allometric scaling include the absence of species-specific clearance mechanisms and the absence of nonlinearities in the disposition. It is possible that the total body clearance of rhuMAb VEGF in humans may be the composite of the clearance of free rhuMAb VEGF and of rhuMAb VEGF-VEGF complexes. Increased levels of VEGF have been reported in several cancers in humans (Ferrari and Scagliotti, 1996; Yamamoto et al., 1996; Salven et al., 1997). Ferrari and Scagliotti (1996) noted that in patients with non-small cell lung cancer, the serum VEGF levels ranged from 38.96 to 4275 pg/ml compared with 60.01 ± 96.22 pg/ml in controls. Salven et al. (1997) noted similar increases in the median VEGF levels in their study (normals, 15 pg/ml; cancer patients, 197 pg/ml). The pharmacokinetics of other humanized monoclonal antibodies have been reported to depend on the concentration of the ligand (Froehlich et al., 1995; Baselga et al., 1996). The observed clearance of rhuMAb VEGF in cancer patients may depend on the relative molar ratios of rhuMAb VEGF and VEGF in the circulation. Thus, a predicted clearance within 2-fold of the observed value seems reasonable.

Rabbits were selected for the tissue distribution studies because rhuMAb VEGF does not bind to rat or mouse VEGF but does bind to rabbit VEGF (purified from a rabbit tumor expressing high levels of VEGF) with approximately 5-fold lower affinity compared with human VEGF (N. Ferrara, unpublished data). The isotypic control humanized monoclonal antibody, rhuMAb E25, is directed against human IgE and does not bind rabbit IgE. rhuMAb E25 has the same IgG₁ construct as rhuMAb VEGF, and previous studies have shown rhuMAb E25 exhibits nonspecific clearance and distribution in cynomolgus monkeys (Fox et al., 1996). Both rhuMAb VEGF and rhuMAb E25 exhibited similar TCA-precipitable radioactivity versus time curves over the 48-h observation period in rabbits (Fig. 4), suggesting that they have a similar systemic disposition in animals.

To compare the distribution of ¹²⁵I-rhuMAb VEGF and ¹²⁵I-rhuMAb E25, the ratio of TCA-precipitable radioactivity per gram of wet tissue for rhuMAb VEGF or rhuMAb E25 at 2 and 48 h in various tissues was determined (Fig. 5). Overall, no significant differences in the distribution and clearance of rhuMAb VEGF and rhuMAb E25 were evident, suggesting that rhuMAb VEGF is cleared from the circulation via a general antibody-clearance mechanism. A potential difference in the distribution of rhuMAb VEGF compared with rhuMAb E25 was the relative higher radioactivity noted at 48 h in the heart, testes, bladder, and kidney for rhuMAb VEGF compared with rhuMAb E25 (Fig. 6). The heart, testes, and kidney all have been reported to express VEGF (Jakeman et al., 1992). The radioactivity noted in these organs may reflect binding of rhuMAb VEGF to VEGF. Since a group receiving radiolabeled rhuMAb VEGF in addition to excess unlabelled antibody was not included in the present study, it is unclear whether the radioactivity noted in the heart, testes, and kidney reflects specific binding of rhuMAb VEGF in these tissues.

Since rhuMAb E25 does not bind to nonprimate IgE, the distribution pattern noted in the present study for rhuMAb E25 is representative of the distribution of a generic humanized IgG₁ monoclonal antibody. The distribution of rhuMAb E25 in rabbits is generally similar to that reported previously by Fox et al. (1996) in cynomolgus monkeys, where the largest percentage of dose was noted in the liver and kidney excluding the plasma compartment. In the present study, in rabbits some distribution of rhuMAb E25 also was seen in other highly perfused organs such as the testes, spleen, and lung. Although every effort was made to fully perfuse the animals with PBS, the radioactivity noted in these organs may be at least partly reflective of residual blood in the organs. The distribution pattern for rhuMAb VEGF was similar to that for rhuMAb E25. Radioactivity was noted primarily in plasma, with minimal amounts noted in other tissues. These results with rhuMAb VEGF are similar to those reported for other human antibodies. Arizono et al. (1994) studied the distribution of ¹²⁵I-TI-23, a human monoclonal antibody against cytomegalovirus in rats, and noted a similar tissue distribution pattern. High levels of radioactivity were noted in organs that are highly perfused with blood, such as liver, kidney, and lung.

We have reported the pharmacokinetics of rhuMAb VEGF in mice, rats, and cynomolgus monkeys after i.v. and s.c. administration. At the doses studied, rhuMAb VEGF has a nonspecific clearance and tissue distribution that is consistent with the disposition of a general monoclonal antibody in animals. These tissue distribution data in animals and the predictable pharmacokinetics support further investigation of rhuMAb VEGF in humans for the treatment of cancer.