Population PK/PD modelling of functional receptor occupancy in a first-time-in human study
C. Chen (1), A. MacDonald (1)
(1) Clinical Pharmacology Modelling and Simulation, GlaxoSmithKline, Greenford, Middlesex, United Kingdom
Objectives: To characterise the pharmacokinetic-functional receptor occupancy relationship of a novel anti-inflammatory in healthy volunteers in a single ascending dose first-time-in-human study. Receptor occupancy was estimated by reduction in ex vivo stimulation of a membrane protein marker on a specific leukocyte sub-type in whole blood. The purpose of the modelling effort was to confirm in vitro predictions from the same assay and establish the concentration-effect relationship for dose prediction in subsequent studies. In addition, we wanted to estimate the most sensitive ligand concentration in the assay for future studies with the molecules in this target class.
Methods: A randomised, single blind, placebo-controlled, cross-over single dose study was conducted in healthy subjects over the dose-range 2-1100 mg. In addition, effects of food, age and gender were studied. The population PK/PD analysis was performed using NONMEM V utilising all data obtained in the study. A two-step sequential PK/PD methodology was employed. Various compartmental pharmacokinetic models were fitted to the population data. A physiological, competitive inhibition pharmacodynamic model was fitted to the receptor occupancy data, accounting for relationship between stimulating ligand and drug concentration and effect.
Results: A two-compartment PK model including an absorption time lag, mixed first and zero order absorption, plus dose-dependent non-linear functions for bioavailability and absorption rate was required to adequately describe the data across the full dose range. The pharmacodynamic model adequately described the PD time course. The apparent KI for inhibition of protein marker expression was 69 ng/mL. Inter-occasion variability was apparent in baseline protein levels and Km for the stimulating ligand (typical value range 17-36 nM). The adequacy of the model was confirmed by steady state data. The ex-vivo receptor occupancy data concurred with the pre-clinically determined in vitro results.
Conclusion: A mechanistic PK/PD model of target receptor occupancy of a novel anti-inflammatory was developed from data obtained in healthy volunteers in a first-time-in-human study. The model has been successfully used for simulating subsequent phase I clinical study scenarios. In vitro pharmacodynamic data from different patient populations can be incorporated into the model to estimate patient-healthy subject PK/PD differences a priori. In addition, the identification of the most sensitive ligand concentration led to an appreciable simplification of the assay.