Development of the AI-driven model for anti-SUD drug development based on neuronal plasticity

点击翻译按钮获取中文摘要

PROJECT SUMMARY Rates of stimulant use, both illicit (e.g. methamphetamine and cocaine) and prescription (e.g. Adderall) are surging on the heels of the opioid epidemic, worsened by the isolation associated with the COVID-19 pandemic. Unlike opioid use disorder, there are currently no FDA-approved medications for the treatment of stimulant use disorder (StUD), leaving only abstinence support and behavioral modification therapies. These are costly treatments with poor efficacy, as evidenced by the high rate of relapse (60-90%). Stimulant use disorder is clearly an unmet need. VeriSIM Life is an innovative company developing and utilizing artificial intelligence (AI) and machine learning (ML) technologies for faster drug discovery and development. Our patented BIOiSIM platform enables prediction of small molecule pharmacokinetics and pharmacodynamics via an AI/ML-parameterized whole-body physiologically based modelling. It is designed to accurately predict the clinical value of investigational drugs before human trials. The full-stack AI- enabled bio-simulation models significantly reduce the number of animal tests required for advancing therapeutics through the pipeline, accelerating drug development and markedly increasing return on investment. The primary goal of this SBIR application is to develop, validate and utilize a reliable and accurate AI-driven tool incorporated with the core BIOiSIM platform to accelerate discovery and development of pharmaceuticals intended for the mitigation of StUD. The current Phase I proposal will begin to address this goal through two complementary approaches. Development of the AI-driven pharmacokinetic modeling specific to CNS drug distribution with the following screening of virtual compound libraries will be performed in Aim 1. This will be supported by proof-of-concept validation of a subset of compounds with in vivo pharmacokinetics. Aim 2 will focus on development of pharmacodynamics prediction modeling for stimulant use disorder drug discovery and development. The focus will be on potential therapeutic targets that modulate neuronal plasticity, identified through an in-depth analysis of the preclinical and clinical literature. This will be supported by proof-of-concept validation of a subset of compounds with preclinical intravenous self-administration studies using cocaine and methamphetamine. The proposed AI/ML-driven approach is expected to markedly accelerate the development process for new stimulant use disorder medications by directing efforts to compounds with optimal ADME and pharmacodynamic properties. Such an approach will create an avenue for fast-track development of affordable and efficacious therapeutics for StUD among other indications.