Model based approach to inform early clinical development for a biologic
Miren Zamacona, Kosmas Kretsos, Ruth Oliver
Modelling & Simulation, Global Exploratory Development, UCB, Slough, UK
Objectives: The knowledge from preclinical in vitro and in vivo preclinical models was integrated in a model-based approach in order to inform the design of early clinical trials for a biologic under development for an inflammatory condition. The objective of this analysis was to determine the therapeutic concentration range and subsequent dose range that was expected to result in clinical efficacy.
Methods: The following modeling & simulation steps were followed: a) PK/target turnover model: a population PK model was build from monkey data and human PK parameters were obtained by allometric scaling. In vitro properties of the biologic and relevant physiologic and pathophysiologic data from the literature were combined to build the turnover model. The PK/target turnover model was used to simulate different dose levels, regimens and routes of administration to predict the anticipated target occupancy in plasma and lymph; b) PKPD models were built from data obtained in preclinical models of disease for three pharmacodynamic endpoints. The PK model was used to derive the AUC as the exposure variable to be related to the pharmacodynamic endpoints. c) Translation of PKPD models to humans based on potency and PK and integration with the PK/target turnover model to determine the therapeutic dose regimen. PK/target turnover model simulations were performed with Berkeley Madonna version 8.3.14. and PKPD analysis were performed with NONMEM version 6.
Results: The PK/target turnover model simulations showed that sustained target engagement was achieved following an every two weeks administration regimen. The relationship between exposure and the three pharmacodynamic endpoints was best described using an inhibitory Emax model. Based on the potency estimates, the expected human doses resulting on 20 to 90% of maximum response were calculated. These results were combined with the PK/target turnover to estimate what target receptor occupancy was needed in order to observe pharmacological response.
Conclusions: The integration of disease and target knowledge, with in vitro and in vivo preclinical disease experiments into a model based approach allowed the prediction of the pharmacologically active concentration/dose level/regimen informing in this way the early clinical development plan.