Six aging clocks read lower biological age in an AI drug trial

AI-generated illustrationAI-generated illustration
Insilico Medicine reported on 7 September 2026 that Nature Biotechnology had published the first head-to-head clinical comparison of several proteomic aging clocks applied to one drug intervention. The samples came from a Phase 2a trial of rentosertib, formerly ISM001-055, in idiopathic pulmonary fibrosis, run in China across 2023 and 2024. Forty-two of the trial’s 71 patients consented to proteomic profiling, giving blood at baseline and at weeks 2, 4 and 12, across 2,841 proteins.
Six independently developed clocks — ProtAge, two OrganAge variants, PAC, ipfP3GPT and PAOPAC — all moved in the same direction. By Insilico’s account the peak effect landed at week 4 in the 30 mg twice-daily arm, at roughly three to four years of biological age reversed and up to six years on one clock. Lung function moved too: forced vital capacity rose 98.4 mL in the 60 mg once-daily arm while the placebo group declined 20.3 mL.
Rentosertib inhibits TNIK, a target Insilico selected with its PandaOmics platform and a molecule its Chemistry42 system generated. The drug has entered Phase 3 in China.
A proteomic clock is a model of aging, not an outcome of it. This trial was designed to test a lung disease in 71 people, and the aging read-out is a secondary analysis of a subset.