AI-designed drug also slowed biological aging, study finds

Six separate ageing clocks, run on the same 43 patients, all pointed in the same direction: younger predicted biological age after 12 weeks on an experimental drug.

The drug, rentosertib, was built by Insilico Medicine to treat idiopathic pulmonary fibrosis, a scarring lung disease with no cure. Its anti-ageing signal showed up almost as a side finding.

Rentosertib's molecular structure was generated using Insilico's Chemistry42 platform, the same category of AI technique behind image generators and chatbots.

Ageing clocks work differently: they estimate biological age from measurements like protein levels or DNA methylation patterns, and different clocks can disagree sharply on the same patient. In this trial, published in Nature Biotechnology, all six agreed in the same direction after treatment.

Idiopathic pulmonary fibrosis causes lung tissue to scar and thicken to the point that oxygen cannot be exchanged effectively from the lungs to the bloodstream, and death will ensue in just a few years after the diagnosis. Rentosertib inhibits a protein known as TNIK, which is a hub gene associated with both fibrosis and ageing that was found through Insilico’s PandaOmics.

Phase IIa showed a statistically significant increase in lung capacity, and the company has progressed it to Phase III for about 320 patients in China.

One of the ageing clocks involved in the study was developed by Vadim Gladyshev from Harvard Medical School. He described this study as the most convincing evidence so far on the reduction of predicted biological age.

Gladyshev also emphasised that the number of participants was small and that ageing clocks are not perfect predictors. Only IPF patients were tested in this study. So far no healthy individuals have been tested with rentosertib, so its effects on a healthy organism are unknown.

Insilico is already treating IPF as a "gateway indication", with plans to test TNIK-targeting drugs against other age-related conditions. Tel Aviv University longevity researcher Evelyne Bischop and other scientists have pointed to this study's design, pairing an AI-discovered drug with AI-built ageing clocks, as a template other longevity trials are likely to follow.