Artificial intelligence-integrated mechanistic modeling for rational design of nanoparticles to improve organ targeting and safety

PROJECT SUMMARY. Nanoparticles (NPs) hold great promise as targeted drug delivery systems but tailoring their pharmacokinetics (PK) to specifically target regions of interest remains a challenge. This limits the clinical translation of NPs due to poor efficacy and safety concerns associated with off-target

accumulation of NP-based formulations. Due to the interactions of NPs with biological components, driven by their structural properties, customizing the pharmacokinetics (PK) of NPs requires a quantitative understanding of the effect of NP structural properties on their whole-body biodistribution, which in turn also governs their safety profile. Therefore, to enable rational design

of NPs to achieve organ targeting and safety, we propose to leverage artificial intelligence to develop a toxicology-integrated physiologically-based pharmacokinetic model (PBPK-Tox) capable of accurately predicting the whole-body exposure and safety of novel nanomaterials, based solely on their structural properties, dose, and route of administration. For this, we will (1)

develop the PBPK-Tox model based on diverse datasets from literature, (2) establish the quantitative relationship between NP properties, exposure, and toxicity, and (3) experimentally test the model predictions of rational design to target one or more organs. Our proposed modeling framework will enable efficient preclinical development of novel nanomaterials (and accelerate

their clinical translation) by providing rational design guidelines through in-depth computational investigation of biological and physicochemical variability on biodistribution and safety of NPs.