Abstract 4508: Predicting therapeutic monoclonal antibody efficacious human doses from mouse experimental tumors: a pharmacokinetic/pharmacodynamic approach

Abstract

Abstract Mechanism-based pharmacokinetic (PK)/pharmacodynamic (PD) modeling allows us to dissect system-specific and drug-specific effects, and has proven to be a powerful tool for rational human dose projections. While PK/PD modeling has been applied reasonably well to predict the efficacious dose levels of small molecule cytotoxic drugs and antibody-drug conjugates (ADCs), there is no systemic evaluation of methods in predicting human efficacious doses for monoclonal antibodies (mAbs) targeting the signal transduction pathways for oncology indications. We collected reported preclinical and clinical data from 4 mAbs targeting the signal transduction pathways (cetuximab (anti-EGFR), panitumumab(anti-EGFR), onartuzumab(anti-cMET), trastuzumab(anti-HER2)) and 1 anti-angiogenesis mAb (bevacizumab) for anti-cancer therapy. A series of mechanism-based tumor growth/tumor inhibition models were developed using published dose-ranging mouse xenograft PK and PD data. PK parameters for each mAb in tumor-bearing mice were first estimated and were subsequently used to drive the PD parameter estimation for observed tumor growth and growth inhibition using non-linear mixed effects modeling approach by NONMEM (version 7.2, Icon Solutions). The intrinsic tumor exponential and/or linear growth rates (Kg), the extrinsic drug-specific maximum tumor inhibition capacity (Kmax) and the serum drug concentrations needed to achieve half of Kmax (EC50) were obtained by weighted least squares estimation. For all 5 mAbs, the steady-state plasma drug levels in cancer patients associated with clinically efficacious doses were found to fall between the model-predicted mean EC90 and EC95 values for each mAb. Using model-based mouse EC90 and EC95 and human antibody systemic clearance, the predicted clinically efficacious doses for these 5 mAbs were within 2-fold of their efficacious doses demonstrated in cancer patients. Interestingly, using the calculated mouse tumorstatic concentrations based on PK/PD model parameters, the predicted human doses were in general more than two-folds lower compared with clinically efficacious doses. Although tumorstatic concentrations were used to predict the clinical doses of small molecule cytotoxic drugs and ADCs reasonably well, we found that tumoricidal concentrations represented by EC90 and EC95 appear to have a better predictive value for mAbs targeting the signal transduction pathways in oncology. In conclusion, a mechanism-based PK/PD modeling approach was developed to project human efficacious doses for several mAbs in the cancer clinic and it can be used to predict human doses for investigational mAbs in the same class. Citation Format: Songmao Zheng, Honghui Zhou, Weirong Wang. Predicting therapeutic monoclonal antibody efficacious human doses from mouse experimental tumors: a pharmacokinetic/pharmacodynamic approach. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4508. doi:10.1158/1538-7445.AM2015-4508