(T1530-10-63) Prediction of Human PK for Multispecific Antibody TAVO412 Using Non-linear Population PK Model and Allometric Scaling

Purpose: TAVO412 is a multispecific anti-EGFR (dual epitopes), anti-cMET, and anti-VEGF antibody with enhanced Fc functions and tumor microenvironment low pH enhanced target binding. This molecule is currently in Phase 1 clinical trial for difficult-to-treat solid tumors. Potential indications include non-small and small cell lung, gastric, pancreatic, triple negative breast, and other tumors. TAVO412 demonstrated target mediated drug disposition (TMDD) in preclinical models. The purpose of this study was to predict human PK. Results were used to set up the starting dose and dosing range in the first-in-human, Phase 1 study.The prediction is validated against preliminary clinical PK data.

Methods: Pharmacokinetic studies of TAVO412 were conducted in the cynomolgus monkey by single or repeated intravenous injections at doses ranging from 5 to 60 mg/kg. Serum samples were collected for PK analyses. Serum drug concentrations were analyzed using ELISA assays. A two-compartment with parallel Michaelis-Menten non-linear clearance population PK model (by ADAT 5 V5.0.63) was used to simultaneously fit the animal PK data obtained at all dose levels and different regimens. Human PK was predicted based on the model and using allometric scaling with an exponent of 0.85 for the CL and Q, 1.0 for volume of distribution and Km, and 0.75 for Vmax. TAVO412 drug concentration profiles of a typical patient at proposed dose level were simulated and compared with the drug concentrations at efficacious doses in mouse models and highest non-severe toxic dose in cynomolgus monkeys. A Phase 1 clinical trial of TAVO412 is conducted in patients with solid tumors in four medical centers in the US. The predicted PK profile is used to compare with preliminary clinical PK results.

Results: Following single and weekly repeated IV administrations in cynomolgus monkeys, TAVO412 exhibited dose dependent non-linear PK, which had been seen with other anti-EGFR and anti-cMet molecules. After the first dose administration, the Cmax values increased in an approximately dose proportional manner. However, the AUC increased in a higher than dose proportional manner from 5 mg/kg to 30 mg/kg and in an approximately dose proportional manner from 30 mg/kg to 60 mg/kg. At lower doses, TAVO412 exhibited faster clearance. All animal PK data were well described by the two compartment, parallel non-linear clearance population model. Resulted model parameters are shown in the table below, and the goodness of fitting is plotted in the figure. Predicted human PK data were simulated based on the model and compared to preclinical efficacious concentrations and the Cmax at the HNSTD to set the starting dose and dose range. Preliminary TAVO412 first-in-patient Phase 1 PK results will be reported and used for a cross evaluation of the predicted values.

Conclusion: TAVO412 showed non-linear PK profiles in the preclinical models. A two compartment and parallel non-linear clearance population PK model described the preclinical data very well. Clinical PK was predicted using the allometric scaling approach. The predictability of this model will be evaluated by clinical data generated by the ongoing Phase 1 study.

Table. Model parameters of monkey PK

Figure. Plots of Goodness of Fitting. Conditional Model Predicted versus Observed Data (left) and Population Model Prediction versus Observed Data (right)