The chemokine ligand 2 (CCL2) promotes angiogenesis, tumor proliferation, migration, and metastasis. Carlumab is a human IgG1κ monoclonal antibody with high CCL2 binding affinity. Pharmacokinetic/pharmacodynamic data from 21 cancer patients with refractory tumors were analyzed. The PK/PD model characterized the temporal relationships between serum concentrations of carlumab, free CCL2, and the carlumab-CCL2 complex. Dose-dependent increases in total CCL2 concentrations were observed and were consistent with shifting free CCL2. Free CCL2 declined rapidly after the initial carlumab infusion, returned to baseline within 7 days, and increased to levels greater than baseline following subsequent doses. Mean predicted half-lives of carlumab and carlumab-CCL2 complex were approximately 2.4 days and approximately 1 hour for free CCL2. The mean dissociation constant (KD ), 2.4 nM, was substantially higher than predicted by in vitro experiments, and model-based simulation revealed this was the major factor hindering the suppression of free CCL2 at clinically viable doses.

Timing of the anti-angiogenic agent with respect to the chemotherapeutic agent may be crucial in determining the success of combination therapy in cancer. We investigated the effects of sequential therapy with the potent VEGF inhibitor, aflibercept, and doxorubicin (DOX) in preclinical acute myeloid leukemia (AML) models. Mice were engrafted with human HL-60 and HEL-luciferase leukemia cells via S.C. and/or I.V. injection and treated with two to three doses of aflibercept (5-25 mg/kg) up to 3-7 days prior to doxorubicin (30 mg/kg) administration. Leukemia growth was determined by local tumor measurements (days 0-16) and systemic bioluminescent imaging (days 0-28) in animals receiving DOX (3 mg/kg) with or without aflibercept. A PK/PD model was developed to characterize how prior administration of aflibercept altered intratumoral DOX uptake. DOX concentration-time profiles were described using a four-compartment PK model with linear elimination. We determined that intratumoral DOX concentrations were 6-fold higher in the aflibercept plus DOX treatment group versus DOX alone in association with increased drug uptake rates (from 0.125 to 0.471 ml/h/kg) into tumor without affecting drug efflux. PD modeling demonstrated that the observed growth retardation was mainly due to the combination of DOX plus TRAP group; 0.00794 vs. 0.0043 h(-1). This PK/PD modeling approach in leukemia enabled us to predict the effects of dosing frequency and sequence for the combination of anti-VEGF and cytotoxic agents on AML growth in both xenograft and marrow, and may be useful in the design of future rational combinatorial dosing regimens in hematological malignancies.

2594 Background: The anthracycline doxorubicin (DOX) is widely used to treat breast cancer. Doxorubicin is associated with pharmacokinetic and pharmacodynamic variability and despite its use for several years there is limited understanding behind it. Hepatic carbonyl reductases (CBR1 and CBR3) catalyze the reduction of DOX into its main circulating C-13 metabolite doxorubicinol (DOXOL). Polymorphisms in CBR1 and CBR3 influence synthesis of DOXOL, and could potentially play a role in the pharmacokinetic (PK) variability seen with doxorubicin treatment. In this study, we examined the influence of genetic polymorphisms in CBR1 and CBR3 on DOX and DOXOL PK. Methods: DOX was administered IV to 79 breast cancer patients at 60 mg/m2. Population PK modeling was performed on the parent concentration-time profiles with the following patient factors: [BSA (1.4-2.6 m2), weight (40-140 kg), age (25-75), race (78% white), CBR1 rs9024 (78% wild-type), CBR3 V244M (47% wild-type), CBR3 C4Y (20.5% wild-type); followed by model validation. Noncompartmental analysis (NCA) was performed for both DOX and DOXOL and the metabolic ratio was calculated as AUCDOXOL0-24:AUCDOX0-24. Results: A two-compartment model was used to describe DOX PK. Mean predicted (%SEM) clearance (CL), plasma volume (Vp), tissue volume and distribution clearance were 28.1 L/hr (7.72), 22.5 L (3.80), 257 L (13.8) and 13.6L (21.8), respectively. Interpatient variability on CL and Vp were 22.1% and 12.6%, respectively. BSA was found to be a significant predictor of the interpatient variability on CL. No other patient factors were found to be significant on parent drug PK. The metabolic ratio, assessing the conversion of DOX to DOXOL, was stratified by different polymorphisms of CBR1 and CBR3. There were no significant differences in metabolic ratio due to CBR1 and CBR3 genotypes. Conclusions: A PK model was developed that was able to characterize DOX pharmacokinetics. CBR1 and CBR3 polymorphisms were tested as covariates but were not found to be significant contributors to the variable pharmacokinetic profiles of DOX and DOXOL.

e13505 Background: Paclitaxel (PTX) is a widely used potent anticancer drug that blocks mitosis by stabilization of microtubules. Previous attempts for oral administration of PTX have been unsuccessful at least partly due to its affinity for the efflux pump, P-gp. Oraxol is an oral formulation of PTX combined with a novel potent P-gp inhibitor, HM30181A, to maximize absorption of paclitaxel without itself being systemically absorbed. A population PK/PD model was developed to characterize the temporal relationship between PTX exposure and neutropenia, and subsequently model-based simulations were utilized to determine the optimal dosing strategy for oraxol. Methods: Oraxol was administered to 68 (51M/17F) patients with various malignancies. Oraxol dose levels ranged from 60 to 420 mg/m2 QW, and 90-150 QWx2 for 3/4 weeks. Population PKPD modeling was performed with the following patient factors: [BSA (1.29-2.15 m2), ECOG PS scores (97% pts ≤ 1), age (36-81.4 yrs), CrCL (39.6-132.8 ml/min), ALB (2.3-4.8 mg/dL), ALT (6-61 IU/L), TBIL (0.2-1.6 mg/dL), and smoking status (NS:57%)]; followed by model validation. Simulations assessed the influence of dose and schedule on the time course and the extent of neutropenia. Results: A precursor-dependent indirect PKPD response model, including three transit compartments was employed to describe the time course of ANC after oraxol administration. Mean predicted (%SEM) baseline ANC and MTT of neutrophils in plasma were 3.6 (3.9) x109 cells/L and 6.3 (4.2) days. The magnitude of the inhibition on marrow cells was 3.3% per 100 ng/mL of PTX. Model validation results showed excellent concordance with the observed and simulated ANC profiles at the 150 mg/m2 dose level. Simulations showed that severity of neutropenia was dose- and schedule-dependent. The optimal dosing scenarios for oraxol 150 mg/m2QDx3 and QDx5 showed the incidence of grade ³3 neutropenia to be 53.8% and 58%, respectively. Simulated grade 4 neutropenia was 25 and 34.6% for the same regimens; similar to IV PTX. Conclusions: A PKPD model quantifying the neutropenic effect of oraxol has been developed. The model predicts that the QDx3 or QDx5 oraxol dosing 3/4 weeks may optimize efficacy with manageable toxicity.