Integration of PBPK/PD Models to Inform Compound Design and Prioritization in Early Stages of Drug Discovery

Physiology-based pharmacokinetic (PBPK) models are broadly applied in late phase drug discovery/ development. Traditionally, compounds are prioritized based on hierarchical filtering with predefined cut-offs for desirable range of various parameters. PBPK models, when combined with related pharmacodynamic assumptions, offer a valuable platform to integrate multiple parameters driving the in vivo PK profile required for target engagement. They also provide mechanistic interpretation of key drivers for the predicted profile to further help with compound design strategies. As such, these results enable compound prioritization in a holistic manner, focusing on multi-property optimization (MPO). This presentation will provide a brief overview of the structure and application of an internal PBPK model. Examples of successful application of this tool on a small molecule drug discovery program will be shared to illustrate its role in driving decisions to guide compound progression.

The optimal PK necessary for a drug candidate to achieve efficacy is highly dependent on the targeted PD, a relationship often not well characterized during the early stages of drug discovery. Relying on generic assumptions about PK and potency can misguide screening and compound design, leading to suboptimal ADME or molecular properties. This, in turn, may increase attrition rates and extend hit-to-lead and lead optimization timelines. The "PD before PK" approach, detailed in this presentation, can be applied in two ways: forward, to virtually screen compounds for potential effectiveness, and reverse, to identify pharmacology-specific PK drivers and the related potency-ADME space early in discovery. This strategy aims to enhance the probability of success and reduce clinical attrition. Potential applications of this approach will also be discussed.