Insilico Medicine's AI Drug for IPF Advances to Phase III Trials

Insilico Medicine is advancing its AI-identified drug, rentosertib, to Phase III clinical trials for treating idiopathic pulmonary fibrosis (IPF). This progression marks a significant milestone in the computational drug discovery sector, moving beyond initial safety evaluations to late-stage efficacy validation. The trial, initiated across 22 clinical sites in China, aims to address the severe lung tissue scarring associated with IPF.

Clinical Trial Details for Rentosertib

The Phase III trial involves a randomized study of 71 patients, separated into placebo and active treatment groups. Participants received either 30 mg or 60 mg of rentosertib daily for a 12-week observation period. Results showed that patients on the 60 mg regimen experienced a mean forced vital capacity gain of +98.4 mL, while the placebo group recorded a capacity loss of -20.3 mL.

The safety profile of rentosertib remained manageable, with adverse events aligning with expected baseline rates across all groups. The U.S. Food and Drug Administration (FDA) granted the drug 'Orphan Drug Designation' in February 2023, underscoring its potential in treating IPF.

AI-Driven Drug Discovery and Target Prioritization

Insilico Medicine's development process utilizes its proprietary platform, Pharma.AI, which segments into distinct engines for biological and chemical engineering tasks. The PandaOmics engine executes the initial target discovery phase by analyzing vast biological datasets, including genomics and clinical trial outcomes. It identified TNIK as the primary target for IPF intervention, bypassing traditional receptor tyrosine kinase pathways targeted by existing antifibrotic medications.

Feng Ren, PhD, Co-CEO and Chief Scientific Officer of Insilico Medicine, stated, “IPF is one of the clearest clinical examples of an age-related disease in which fibrosis, chronic inflammation, extracellular matrix remodeling, and cellular senescence intersect.” This approach connects TNIK to fibrotic and inflammatory disease mechanisms, leading to the creation of rentosertib.

Generative Chemistry and Preclinical Testing

Following target selection, Insilico's Chemistry42 engine employs generative molecular design, moving away from conventional high-throughput screening. Instead of searching existing compound libraries, it uses Generative Tensorial Reinforcement Learning to create molecules that align with the target protein pocket. This method significantly reduced the timeline for preclinical candidate nomination to 18 months.

The generative chemistry phase produced 79 potential molecules, with the 55th iteration selected for preclinical testing. This innovative approach aims to streamline drug development processes, as outlined in the company's 2019 GENTRL methodology publication in Nature Biotechnology.

Clinical assessments integrate complex proteomic analyses to validate predicted biological interactions. Insilico Medicine employs various proteomic aging-clock frameworks within the IPF trial, capturing key exploratory geroscience readouts.