Concomitant Administration of Bevacizumab,Irinotecan,5-Fluorouracil,and Leucovorin: Nonclinical Safety and Pharmacokinetics

Abstract

Bevacizumab (Avastin) is a humanized monoclonal antibody against vascular endothelial growth factor approved for use in combination with 5-fluorouracil (5-FU)-based chemotherapy for first-line treatment of metastatic colorectal cancer. The Saltz regimen (irinotecan/5-FU/leucovorin [LV]) is a first-line treatment for this indication. The objective of this study was to evaluate the safety of bevacizumab when administered concomitantly with the Saltz regimen to cynomolgus monkeys, and to determine if the pharmacokinetics of bevacizumab, irinotecan, SN38 (the active metabolite of irinotecan), or 5-FU were affected by combined administration. Male cynomolgus monkeys were intravenously administered the Saltz regimen (125 mg/m² irinotecan, 500 mg/m² 5-FU, 20 mg/m² LV) alone (n = 4) or concomitantly with 10 mg/kg bevacizumab (n = 5) on days 1 and 8. All animals survived to euthanasia on day 15. Adverse effects associated with the Saltz regimen included diarrhea and neutropenia. Macroscopically, two animals from each group had small thymus glands that correlated microscopically with lymphoid depletion. Myeloid hypoplasia and/or erythroid hyperplasia was observed in the sternal bone marrow of most animals. These effects were considered to be associated with the Saltz regimen; concomitant bevacizumab administration did not alter the severity of these findings. Irinotecan and 5-FU were observed to be rapidly eliminated (t _1/2 = 1 h and 0.5 h, respectively). Although the number of animals in each group was small and no statistical comparison between groups was performed, bevacizumab did not affect the disposition of either agent. These results indicate that bevacizumab can be safely administered in combination with the Saltz regimen without pharmacokinetic interaction.

Tumors grow and metastasize using angiogenesis to stimulate the development of new vasculature (Berlin 2002). Conventional therapies target neoplastic cells within a tumor. However, antiangiogenic therapy represents a promising new frontier of oncology.

Vascular endothelial growth factor (VEGF), and VEGF-A in particular, is an important target for antiangiogenic tumor therapy (Ferrara 2002; Jain 2002). It is a secreted endothelial-specific mitogen with potent proangiogenic properties and significant effects on vascular permeability (Connolly et al. 1989). The conversion of tumors to an angiogenic phenotype is triggered by an alteration in the balance of pro- and antiangiogenic factors; angiogenesis is favored, and expression of VEGF-A is substantially increased in most human tumors (Ferrara 1999). This “angiogenic switch” allows tumor neovascularization, growth, and metastasis (Hanahan and Folkman 1996). Preclinical studies of A.4.6.1, a murine anti-human VEGF monoclonal antibody, conducted in a number of human tumor xenografts in immunodeficient mice show that tumor vessel permeability is reduced, causing the vasculature to regress and suppressing the tumor growth rate (Gossmann et al. 2000). Data such as these led to the development of bevacizumab (Avastin, rhuMAb-VEGF; Genentech, South San Francisco, CA), a recombinant humanized anti-VEGF monoclonal antibody (IgG₁) with high affinity for human VEGF. Bevacizumab has been shown to bind to human, cynomolgus monkey, and rabbit VEGF; the cynomolgus monkey was chosen for the safety evaluation of bevacizumab because human and cynomolgus monkey VEGF are predicted to have identical amino acid sequences (Shima et al. 1996). Once-or twice-weekly intravenous bolus administration for up to 26 weeks in cynomolgus monkeys at doses up to 50 mg/kg/dose was generally well tolerated, with no dose limiting toxicity or production of anti-bevacizumab antibodies. Marked inhibition of physiologic angiogenesis was observed in the growth plate and ovary of treated monkeys, indicating that bevacizumab is a potent antiangiogenic agent (Ryan et al. 1999). Phase I clinical trials showed that bevacizumab was generally well tolerated, either as monotherapy (Gordon et al. 2001) or in combination with standard chemotherapeutic regimens (Margolin et al. 2001). A phase II trial of bevacizumab plus 5-FU/LV in patients with metastatic colorectal cancer increased survival and time to disease progression, with acceptable tolerability, compared with 5-FU/LV alone (Kabbinavar et al. 2003). Subsequent to this phase II study, several phase III trials were initiated. One study evaluated the efficacy and safety of bevacizumab (5 mg/kg every 2 weeks) in combination with 5-FU/LV and the topoisomerase I inhibitor, irinotecan, (Saltz (IFL) regimen) in metastatic colorectal cancer and showed that addition of bevacizumab to bolus irinotecan/5-FU/LV increased median survival (from 15.6 months to 20.3 months), objective response rate, and progression-free survival (Hurwitz et al. 2004). Bevacizumab was well tolerated and adverse effects were manageable. Bevacizumab was approved by the U.S. Food and Drug Administration (FDA) in February of 2004 (BLA 125085) for the treatment of metastatic colorectal cancer in combination with 5-FU–based chemotherapy. In addition, bevacizumab was subsequently approved in the European Union, Australia, Dominican Republic, Iceland, Israel, Mauritius, Norway, and Switzerland. The non-clinical study described in this report was initiated prior to the pivotal phase III clinical trial (Hurwitz et al. 2004) to evaluate the safety of concomitant bevacizumab and Saltz regimen treatment, and to evaluate the pharmacokinetics of the Saltz regimen when administered with bevacizumab.

MATERIALS AND METHODS

Young adult naïve male cynomolgus monkeys (Covance Research Products Inc., Alice, TX) approximately 2 to 3.5 years of age and weighing 2.6 to 3.4 kg were assigned to treatment groups (n = 4 to 5 animals per group) using a randomization scheme designed to achieve similar mean body weights across groups. Female animals were not included in this study as previous studies demonstrated that the pharmacokinetics of bevacizumab are not affected by gender. All procedures employed were in compliance with the Animal Welfare Act Regulations, 9 CFR 1–4, and were approved by the Institutional Animal Care and Use Committee of Covance Laboratories. Approximately 12 days before the start of dosing, all animals were surgically implanted with a Micro-Renathane (Braintree Scientific, MA) catheter via the right femoral vein into the inferior vena cava. Bevacizumab and the components of the Saltz regimen were administered by the same routes and sequence used clinically. On study days 1 and 8, animals were administered either the Saltz regimen (Irinotecan HCl Injection [Camptosar, Pharmacia & UpJohn, Kalamazoo, MI], 5-flurouracil (5-FU) [Adrucil, Pharmacia & Upjohn, Kalamazoo, MI], and leucovorin (LV) [Leucovorin Calcium for Injection, Immunex Corporation, Seattle, WA]) alone or in combination with bevacizumab (Avastin, Genentech Inc., South San Francisco, CA). The study design is summarized in Table 1. Irinotecan was administered first as a 90-min intravenous (IV) infusion by an infusion pump, followed by LV and 5-FU as bolus IV injections, followed by bevacizumab administered to assigned animals by bolus IV injection. A 3-ml flush of sterile saline was administered following each bolus dose. The bevacizumab dose (10 mg/kg) was selected to result in serum concentrations in cynomolgus monkeys similar to those reached in humans receiving 5 mg/kg bevacizumab every 2 weeks. The doses of 5-FU, LV, and irinotecan used in this study were identical to those used in phase III clinical trials. An additional three animals were administered the Saltz regimen, without bevacizumab, at a lower irinotecan dose of 100 mg/m² to assess the dose-response relationship of the toxicities associated with the Saltz regimen. Observations were similar to those observed in animals treated with 125 mg/m² irinotecan, and, therefore, are not described in this report.

Throughout the duration of the study, all animals were observed four times daily (i.e., in the early AM, late AM, early PM, and late PM); abnormal findings or an indication of normal were recorded. As severe diarrhea is the dose-limiting toxicity associated with the administration of irinotecan (Wiseman and Markham 1996), the incidence and consistency of feces were recorded for each animal during these observation intervals. The weekly incidence of liquid, mucoid, or nonformed feces was calculated for each group and divided by the number of animals per group to obtain the average incidence per animal per week. Qualitative food consumption observations were made once daily. Individual animal body weights were recorded weekly before initiation of treatment and on days 1, 7, and 14. Blood and urine samples were collected from each animal, after an overnight fast, for routine hematology (including coagulation), clinical chemistry, urinalysis, and urine chemistry before the start of treatment, before treatment on day 8 (hematology and clinical chemistry only) and on day 15 before necropsy. A complete necropsy was performed on all animals on day 15; specific tissues were weighed and organ-to-organ body weight percentages and organ-to-brain weight ratios were calculated. With the exception of the ocular tissues, all histopathological samples for microscopic analysis were preserved in 10% neutral-buffered formalin, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. Bone marrow smears from the sternum were prepared and stained with Wright’s stain.

Whole blood (∼1 ml) for pharmacokinetic analysis was collected from the femoral vein of all animals on study days 1 and 8. Samples for plasma irinotecan and 5-FU analyses were collected into sodium heparin tubes; samples for bevacizumab analysis were allowed to clot. Irinotecan and SN38 (the active decarboxylated metabolite of irinotecan) plasma concentrations were measured using validated high-performance liquid chromatography (HPLC) with fluorescence detection (Cedra Corporation, Austin, TX). Irinotecan and SN38 were extracted from plasma into an organic solvent mixture. Following centrifugation, the organic layer was transferred to a fresh tube and evaporated to dryness. An aliquot of the reconstituted extract was analyzed by HPLC with fluorescence detection. Quantitation was performed using separate weighted (1/x ²) linear least squares regression analyses generated from combined calibration standards prepared immediately prior to each run. The assay lower limit of quantitation (LLOQ) was 5.0 ng/ml for both analytes. SN38 pharmacokinetic parameters are not reported because concentrations were generally below the LLOQ. 5-FU plasma concentrations were measured using validated HPLC with ultraviolet absorbance detection (PPD Development Inc., Middleton, WI). Depending on the dilution, the LLOQ was either 20 or 40 ng/ml. Serum bevacizumab concentrations were determined using a qualified enzyme-linked immunosorbent assay (ELISA) that used VEGF as the capture agent and a goat anti-human anti-bevacizumab antibody (conjugated with horseradish peroxidase) as the secondary antibody (Lin et al. 1999). The assay had a LLOQ of 3.9 ng/ml in serum. The toxicities associated with the Saltz regimen are not commonly attributed to leucovorin (Almonti et al. 2004; Hillman 2001; Saltz et al. 2000), therefore leucovorin concentrations were not measured in this study.

Statistical Analysis

Clinical pathology and organ weight data collected prior to, on, or after the first day of treatment were subject to statistical analysis. Comparisons were evaluated at the 5%, two-tailed probability level. Levene’s test (Levene 1960) was performed to test for variance homogeneity. In case of heterogeneity (p ≤ 0.05), the data were transformed. One-way analysis of variance (ANOVA) was used to analyze clinical pathology values and organ weight data. ANOVA was performed on the homogeneous or transformed data. If the ANOVA was significant, Dunnett’s multiple comparison t test (Dunnett 1964) was used for pairwise comparisons between groups. If the ANOVA was significant, Dunnett’s t test was used for pairwise comparisons between groups.

Pharmacokinetic Analysis

Pharmacokinetic parameters for irinotecan and 5-FU were estimated for each animal by a two-stage approach using a one-compartment infusion or bolus injection model and data from both doses were fitted simultaneously using WinNonlin Pro, version 3.2, software (Pharsight Corp., CA). Bevacizumab concentrations were summarized by timepoint but no pharmacokinetic analysis was performed. Values at LLOQ were not included. Due to the relatively small sample size, no comparative statistics were conducted on pharmacokinetic data.

RESULTS

All animals survived to scheduled necropsy on day 15. Prior to euthanasia on day 15, one animal in the Saltz regimen group was observed to be in poor health (periorbital swelling, hunched posture, hypoactivity), and the condition of the animal was considered to be related to treatment with the antineoplastic agents. Diarrhea, a known side effect of irinotecan, was observed consistently in all animals in both groups during the active treatment period and was attributed to the administration of irinotecan. Administration of bevacizumab had no effect on the incidence (Table 2) or consistency of fecal events when compared to the fecal findings in the animals administered the Saltz regimen alone. With the exception of one animal in the Saltz plus bevacizumab group, all animals gained no weight or lost between 0.1 and 0.4 kg over the 2-week treatment period. This weight loss was consistent with observations of diarrhea and low food consumption. Low food consumption was observed in all animals, with the exception of one animal in the Saltz plus bevacizumab group, with no apparent differences between groups. Low food consumption was considered to be related to the Saltz regimen.

Administration of the Saltz regimen was associated with a decrease in white blood cell count characterized by decreased numbers of neutrophils, lymphocytes, monocytes, and eosinophils. The decrease in white blood cell count appeared cumulative with administration of the Saltz regimen, with mean leukocyte counts (± SD) being lower on day 15 (3.2 × 10³ /μl [1.49]) than day 8 (5.8 × 10³ /μl [2.34]) and baseline (9.1 × 10³ /μl [3.6]) in animals receiving the Saltz regimen alone. The administration of bevacizumab did not alter the magnitude of these changes (mean leukocyte counts [± SD] on day 15: 4.3 × 10³ /μl [3.46]; day 8: 6.1 × 10³ /μl [2.43]; baseline: 11.8 × 10³ /μl [2.46] in animals receiving the Saltz regimen plus bevacizumab). Individual animals in each group exhibited other changes likely related to treatment with the Saltz regimen, including decreased platelet count, cholesterol, and potassium. Decreases in red blood cell count, hemoglobin, hematocrit, and serum proteins were also observed in all animals and were attributed, in part, to multiple blood sampling for pharmacokinetics; however, administration of the Saltz regimen was considered to have exacerbated the decreases. Administration of bevacizumab did not impact the magnitude of these effects.

Treatment-related macroscopic effects were limited to small thymus in two animals administered the Saltz regimen alone and two animals administered the Saltz regimen plus bevacizumab; these observations correlated with histologic findings of slight to moderately severe lymphoid depletion in the thymus of all animals in both dose groups. With the exception of one animal administered the Saltz regimen alone, myeloid hypoplasia and/or erythroid hyperplasia was observed in the sternal bone marrow in all animals. These changes were considered related to the Saltz regimen, and administration of bevacizumab did not alter the magnitude of these effects.

The plasma concentration-time profiles for irinotecan and 5-FU following intravenous administration on days 1 and 8 are presented in Figures 1 and 2. The kinetics of irinotecan and 5-FU were characterized by rapid plasma clearance with a half-life of approximately 1 h and 0.5 h, respectively (Tables 3 and 4). Concentrations of the active metabolite of irinotecan, SN38, were quantifiable in most animals up to 15 min following completion of the irinotecan infusion only, and ranged from 5.0 to 12.4 ng/ml. Mean (± SD) bevacizumab concentrations on days 2 (24 h after the first dose), 8 (7 days after the first dose) and 9 (24 h after the second dose) were 125 (6.87), 61.8 (13.6), and 197 (54.1) μg/ml, respectively.

DISCUSSION

Antiangiogenic therapy offers several potential advantages as an approach to cancer treatment, particularly physical accessibility and genetic stability of target cells (Jain 2002). VEGF plays a crucial role in angiogenesis and in pathological processes such as tumor growth, rheumatoid arthritis, and ocular neovascularization. A number of antiangiogenic drugs are in development. Agents such as thalidomide have a broad spectrum of immunomodulatory effects that include antiangiogenic activity. Small molecules such as PTK-787 (Drevs 2003) are designed to target protein tyrosine kinase receptors, although selectivity in targeting a subset of the approximately 550 protein kinases expressed in mammalian cells may present a challenge. The humanized monoclonal antibody bevacizumab is a highly selective inhibitor of VEGF. Its selectivity should limit the potential for nonspecific adverse events and has demonstrated a significant clinical benefit with no overlapping toxicity with the Saltz regimen in patients with metastatic colorectal cancer (Hurwitz et al. 2004).

The present study reports the safety and pharmacokinetics of bevacizumab in combination with the Saltz regimen in cynomolgus monkeys. It was performed prior to initiation of large phase III trials to provide assurance of the safety of concomitant administration of bevacizumab with the Saltz regimen. The cynomolgus monkey was used as the test species because of the similarity in the human and cynomolgus monkey VEGF protein sequences (Shima et al. 1996), and the ability of bevacizumab to bind to cynomolgus monkey VEGF. The results of this study show that bevacizumab did not exacerbate the severity of the toxicities attributed to treatment with the Saltz regimen alone. The toxicities observed in this study were those commonly associated with the cytotoxic components of the Saltz regimen, particularly diarrhea and hematologic disturbances (e.g., leukopenia/neutropenia). Previous studies assessing the safety of multiple doses of bevacizumab alone in cynomolgus monkeys reported physeal dysplasia, subchondral bony plate formation, and inhibition of vascular invasion of the growth plate following 4 or 13 weeks of dosing (Ryan et al.1999), providing further support that the toxicities reported in this study were primarily related to the Saltz regimen. The findings of the present study correspond with observations in humans; clinical experience to date suggests that bevacizumab has a toxicity profile distinct from that of chemotherapeutic agents (Hurwitz et al. 2004; Kabbinavar et al. 2003).

The pharmacokinetics of irinotecan and 5-FU were characterized by rapid clearance with a half-life of approximately 1 h and 0.5 h, respectively. Irinotecan is rapidly converted to the active metabolite SN38. Unfortunately, in this study SN38 data were mostly below the LLOQ, precluding pharmacokinetic analysis. The determination of the pharmacokinetics of irinotecan is challenging because both the parent drug and active metabolite occur as active lactone and inactive carboxylate forms and significant variability in metabolism is observed (Blaney et al. 1998; Inaba et al. 1998). Concomitant administration of bevacizumab did not appear to influence the disposition of either irinotecan or 5-FU. Indeed, an interaction between bevacizumab and irinotecan and/or 5-FU is unlikely because IgGs are cleared through F_c/F_cRn systems, whereas irinotecan and 5-FU are primarily enzymatically transformed in the liver (Diasio and Harris 1989; Ghetie and Ward 2000; Horowitz, Wadlen, and Wiernick 1997; Trang 1992). The phase III clinical trial of colorectal cancer patients treated with irinotecan/5-FU/leucovorin (bolus-IFL), with or without bevacizumab, characterized irinotecan and SN38 disposition in a small pharmacokinetic substudy. Irinotecan concentrations were similar in patients receiving bolus-IFL alone or in combination with bevacizumab. The concentrations of SN38 were, on average, 33% higher in patients receiving bolus-IFL in combination with bevacizumab when compared with bolus-IFL alone. Due to high interpatient variability and limited sampling, the extent of the increase in SN38 levels in patients receiving concurrent irinotecan and bevacizumab is uncertain. In contrast, in cynomolgus monkeys, characterization of the disposition of SN38 was limited and will require additional studies.

The pharmacokinetics of bevacizumab were rigorously investigated in a previous study (Gaudreault et al. 2002) in cynomolgus monkeys. Although the sampling was too limited to fully characterize bevacizumab, the results obtained in this study were consistent with previous data demonstrating that bevacizumab exhibited multicompartmental pharmacokinetics with an alpha half-life of <24 h and a beta half-life of 1 to 2 weeks (Lin et al. 1999), and suggest that bevacizumab pharmacokinetics are not altered by concomitant administration of chemotherapy.

In conclusion, this safety and pharmacokinetic study in male cynomolgus monkeys demonstrated that bevacizumab did not alter the magnitude of the observed adverse effects, all of which were attributed to components of the Saltz regimen. In addition, there was no evidence that bevacizumab affected the pharmacokinetics of 5-FU or irinotecan.