Translational Pharmacokinetic/Pharmacodynamic Modeling of Targeted mRNA-LNP Therapeutics

Abstract ID 98013 Poster Board 435 The use of nucleic acids has grown exponentially in recent years following approvals of siRNA therapeutics (2018) and the global deployment of mRNA vaccines for COVID-19 (2021). A major challenge in accurately predicting safe and efficacious doses of these therapeutics is the significant spatiotemporal disconnect that can be observed between blood/tissue pharmacokinetics and pharmacodynamics. This is further complicated by the relative paucity of pharmacokinetic data in the literature, particularly for mRNA-based therapeutics, which are often tracked in preclinical models by tracing transgene expression, a marker of pharmacodynamic effect that does not always correlate with the site of mRNA-LNP delivery. To bridge this gap, we have developed a translationally-relevant pharmacokinetic/pharmacodynamic model for intravenously injected mRNA-LNP. This was accomplished through the use of a physiologically-based pharmacokinetic model platform that includes key sites of nanoparticle uptake and clearance (lungs, liver, spleen, kidneys, heart) and linked the kinetics of transgene product expression (luciferase) to the intracellular degradation kinetics of LNP. Through this framework, accurate characterization of the blood and tissue kinetics of untargeted and endothelial targeted mRNA-LNP were made following IV injection in mice. The model was also able to capture the kinetics of luciferase expression in all tissues as measured by both an ex vivo luciferase assay as well as bioluminescent imaging. To evaluate the predictive capacity of the model, we performed simulations of the biodistribution and transgene expression pattern of mRNA-LNP targeted to a distinct molecule, intercellular adhesion molecule 1 (ICAM). With no additional parameter estimation, the model was able to predict biodistrbution with a mean % prediction error (%PE) of 37.0% and transgene expression with a mean %PE of 17.7%. This suggests that the model platform is useful in extrapolating predictions to other target molecules. We are currently evaluating the performance of this model for additional targets, as well as for its utility in predicting PK/PD in higher species. This work was supported by R00HL153696.