PBPK (physiologically based pharmacokinetic) modelling is a mathematical technique used to predict the in vivo pharmacokinetics (PK) of a xenobiotic (such as a drug or other compound), using information relating to its ADME (absorption, distribution, metabolism and excretion) properties. This approach typically involves developing a multi-compartment model in which the compartments represent the major organs and tissues in the body, whilst the connections between compartments represent their interconnecting blood flows. Within the PBPK model, complex ADME processes are modelled by incorporating anatomical, physiological, biochemical and chemical information, including physiochemical data and in vitro ADME data.
As these models typically require in vitro data, their implementation within the drug discovery process normally only occurs during the lead optimisation phase or later. Methods used to predict PK directly from structure do exist, but often rely on QSAR-based methods which tend not to be as robust as PBPK models optimised directly on human clinical data.
chemPKTM is a new product on the market that can predict the key human PK properties AUC, Cmax and Tmax directly from chemical structure using a PBPK model optimised on human clinical PK data. This has the potential to revolutionise the drug discovery process by providing insight into the key pharmacokinetic parameters, before the compound has even been synthesised. Medicinal chemists and computational chemists now have a virtual screening tool that can assist in directing their chemistry and prioritising their future screening activities by taking into account the predicted exposure of their compounds in humans.
Daniel Mucs PhD, a mathematical modeller at Cyprotex, has played a major role in developing the chemPKTM workflow solution. An overview of the technology, as implemented as a set of nodes within the KNIME Analytics Platform, a workflow management system popular for cheminformatics, is presented by Daniel in the following video.