Gene expression databases for physiologically based pharmacokinetic modeling of humans and animal species.

Abstract

In drug research, developing a sound understanding of the key mechanistic drivers of pharmacokinetics (PK) for new molecular entities is essential for human PK and dose predictions. Here, characterizing the absorption, distribution, metabolism, and excretion (ADME) processes is crucial for a mechanistic understanding of the drug-target and drug-body interactions. Sufficient knowledge on ADME processes enables reliable interspecies and human PK estimations beyond allometric scaling. The physiologically based PK (PBPK) modeling framework allows the explicit consideration of organ-specific ADME processes. The sum of all passive and active ADME processes results in the observed plasma PK. Gene expression information can be used as surrogate for protein abundance and activity within PBPK models. The absolute and relative expression of ADME genes can differ between species and strains. This is affecting both, the PK and pharmacodynamics and is therefore posing a challenge for the extrapolation from preclinical findings to humans. We developed an automated workflow that generates whole-body gene expression databases for humans and other species relevant in drug development, animal health, nutritional sciences, and toxicology. Solely, bulk RNA-seq data curated and provided by the Swiss Institute of Bioinformatics from healthy, normal, and untreated primary tissue samples were considered as an unbiased reference of normal gene expression. The databases are interoperable with the Open Systems Pharmacology Suite (PK-Sim and MoBi) and enable seamless access to a central source of curated cross-species gene expression data. This will increase data transparency, increase reliability and reproducibility of PBPK model simulations, and accelerate mechanistic PBPK model development in the future.