Rentosertib phase 2a results are published in Nature Medicine

Background

Idiopathic pulmonary fibrosis (IPF) is a progressive lung disease in which scar tissue builds up in the lungs and lung function tends to decline over time. Patients are treated on top of a standard of care that can include antifibrotic drugs ^[1]. Drug discovery for the condition has usually started with a biological target chosen by researchers and then moved to the long search for a molecule that acts on it.

Rentosertib, formerly known as ISM001-055, took the opposite route: both the target and the compound came out of an AI-driven discovery platform. On 3 June 2025 Nature Medicine published the results of the drug's phase 2a trial ^[1].

The trial

The study was multicentre, double-blind, randomised and placebo-controlled, and ran at 21 sites in China, enrolling between 19 July 2023 and 11 June 2024 ^[1]. Of 128 screened patients, 71 were randomised in a 1:1:1:1 ratio to placebo, 30 mg once daily, 30 mg twice daily or 60 mg once daily, for 12 weeks, on top of standard of care ^[1]. Rentosertib is a TNIK inhibitor ^[1].

The primary endpoint was safety and tolerability rather than efficacy: the share of patients with at least one treatment-emergent adverse event ^[1].

Results

On the primary endpoint the arms were similar, with 70.6 to 83.3 percent of patients reporting at least one treatment-emergent adverse event ^[1].

The trial also measured forced vital capacity, the volume of air a patient can force out of the lungs after a full breath, a standard way of tracking lung function in fibrosis. At the highest dose, forced vital capacity rose by a mean 98.4 ml over the study, against a decline of 20.3 ml on placebo ^[1]. In the subgroup of patients who were not taking an antifibrotic, the 60 mg group gained 187.8 ml ^[1].

Blood levels of the drug settled quickly: steady state was reached by week 2, and the terminal half-life at week 12 was 10.9 to 12.0 hours ^[1].

Where the molecule came from

The paper describes the compound as the first reported instance of an AI platform discovering both a disease-associated target and a compound against that target ^[1]. The timeline given is short by the standards of conventional discovery: preclinical candidate nomination took 18 months, and phase 0/1 testing was completed in under 30 months from the start of target discovery ^[1]. Insilico Medicine, the company behind the programme, says the candidate was identified after screening and testing only 78 molecules ^[2].

Limitations and what followed

The authors point to three limits on what the results can show. The treatment arms were small, all participants were resident in China and of similar race, and follow-up was short ^[1]. The lung-function findings therefore come from a small, single-country study over 12 weeks, and the trial was set up chiefly to look at safety.

Since publication, the company's pipeline page has listed the TNIK inhibitor for IPF in phase 3 ^[3], the stage at which a drug is tested in larger groups of patients.

The trial matters less as a verdict on one molecule than as a worked example: a target and a drug candidate both proposed by software, then carried into a randomised, placebo-controlled patient study and published with its design and its limitations in view ^[1].