AI-assisted Drug Re-purposing for Human Liver Fibrosis

Liver fibrosis is a severe disease with few treatment options due to the poor quality of the available animal and in vitro models. To address this, human hepatic organoids were grown in microwells (i.e., microHOs) to provide a robust, live cell imaging platform for serially evaluating the anti-fibrotic effects of drug candidates. Here, we developed a multi-parameter image analysis workflow that enabled anti-fibrotic efficacy and drug toxicity to be serially assessed in microHOs, and endpoint analyses (flow cytometry and scRNA-Seq) were used to characterize drug effects on different cell types. Transcriptomic analysis revealed that the efficacy of five anti-fibrotic agents occurred via two different mechanisms: two inhibited TGF?-induced intracellular signaling while three altered TGF?-induced mesenchymal cell differentiation. microHO data facilitated drug re-purposing by enabling the effects of different drugs to be directly compared. This study provides a first demonstration that a hypothesis generating multi-agent AI system (AI co-scientist) can assist in re-purposing drugs for a disease with few treatment options. Two of three AI co-scientist-recommended drugs targeting epigenomic modifiers exhibited significant anti-fibrotic activity; and one is an FDA-approved drug (Vorinostat) that also promoted liver parenchymal cell regeneration. Hence, the use of AI co-scientist coupled with this enhanced microHO platform identified a new generation of liver fibrosis treatments that also promote liver regeneration. Overall design: scRNA-seq and scATAC-seq profiling microHO treated with TGFb1 and TGFb1 plus different anti-fibrosis drugs.