Development of a web-based predictive model of nanoparticle delivery to tumors by integrating physiologically-based pharmacokinetic modeling with artificial intelligence

University Of Florida, Gainesville FL

Investigators

Linked publications & trials

Abstract

Project Summary and Abstract In the field of nanomedicine, there are several challenges: (1) the delivery efficiency of nanoparticles (NPs) to tumor is very low (<1% injected dose); (2) it is unknown whether there is a sex difference in tumor delivery efficiency of NPs; (3) there is a lack of models that can predict distribution of NPs and the carried drug to tumor and major organs and extrapolate from animals to humans. To address these challenges, the objective of this renewal application is to develop a mechanistic, well-validated, and predictive generic physiologically based pharmacokinetic (PBPK) models for NPs and the carried drug in male and female tumor-bearing mice that is extrapolatable to rats. We hypothesize that tissue distribution and tumor delivery of NPs and the carried drug can be predicted with a well-trained PBPK model by using species-specific, sex-specific, and NP-specific parameters with machine learning and artificial intelligence (AI) approaches. Compared to the previous cycle of this R01 grant, this renewal is innovative in several aspects: (1) the first time to systemically analyze whether there is a sex difference in tumor delivery efficiency of NPs; (2) the first PBPK model for NPs in tumor-bearing animals to evaluate the cross-species extrapolation capability, which is critical to extrapolate animal results to humans; (3) the previous cycle of this grant only focused on NPs in tumor using a single-task learning model, but this renewal will comprehensively analyze distribution of NPs and the carried drug to both tumor and major organs using multi-task learning models. Three specific aims were designed. Aim 1: To develop a PBPK model for NPs and its carried drug in male and female tumor-bearing mice using traditional PBPK approaches. Aim 2: To develop a robust, validated and predictive AI-assisted generic PBPK model for NPs and its carried drug in male and female tumor-bearing mice by employing machine learning and AI approaches. Aim 3: To validate our PBPK model with new experimental data, extrapolate it to rats, and convert it to an AI-empowered web dashboard. In Aim 1, we will create a Nano-Drug Database by including new data published from 2021-2029, especially new data on the carried drugs of NPs, which were not included in previous studies. In Aim 2, the multi-task machine learning model enables to determine the impact of distribution to major organs on tumor delivery efficiency. In Aim 3, we will conduct original pharmacokinetic and tissue distribution experiments in both tumor-bearing mice and rats to evaluate and validate our model. This project is significant because it directly addresses the low tumor delivery efficiency of NPs, which is a critical barrier to progress in this field for >2 decades. This project has broad impacts because this project will create tangible nanomedicine database and AI-empowered PBPK model to answer a key question on the potential sex difference, improve our scientific knowledge on key determinants of tumor delivery efficiency of NPs, and provide a high- throughput screening tool to improve our technical ability to efficiently screen and design new nanomedicines, and help reduce animal experimentation and ultimately accelerate clinical translation of nanomedicines.