An AI-designed drug, rentosertib, from Insilico Medicine improved lung function by 98.4 mL in a 71-patient idiopathic pulmonary fibrosis study published June 3 in Nature Medicine, marking what may be the first peer-reviewed phase 2a result for a molecule generated, with its target discovered, entirely by generative AI.
In an email, Insilico CEO Alex Zhavoronkov, Ph.D. said the paper on the study, which was published in Nature Medicine, was “very important.”
Insilico sees the phase 2a data as a steppingstone toward ‘pharmaceutical super-intelligence’
Zhavoronkov said Insilico publishes about “one paper every week,” many with preclinical or clinical data, and has “nominated 22 developmental candidates and 10 reached the clinical stage.” He called the rentosertib study “special,” noting it is the company’s first program to run “from novel target to novel molecule to Phase IIa complete.”
He noted in the same message that rentosertib is part of Insilico’s broader goal to achieve “pharmaceutical super-intelligence.” While rentosertib showed promise, with patients on the 60 mg dose gaining a mean 98.4 mL in forced vital capacity versus a 20.3 mL decline on placebo, and a subgroup not taking standard antifibrotics improving by 187.8 mL—there were notable safety signals. Seven patients discontinued owing to “liver injury or dysfunction,” with four of those cases occurring in patients concurrently taking nintedanib antifibrotic therapy. Only 12 of 18 patients (67%) completed the high-dose regimen compared to 88% in the placebo group.
From the Nature Medicine paper: Figure 2a illustrates the changes in Forced Vital Capacity (FVC).
Early signal for AI-generated TNIK drug
Previous analyses show AI-originated compounds clear phase 2 at roughly the same 40% rate as conventional drugs, with early high-profile failures including Exscientia’s discontinued cancer treatment EXS-21546 and BenevolentAI’s dermatitis drug BEN-2293. The authors acknowledge “the small cohort size of each arm, the geographical and demographic homogeneity of the participants (all were residents of China of similar race) and a short period of follow-up” as key limitations, but note that “despite the short duration and the number of withdrawals from the trial across all arms,” the results remain “encouraging for additional study.”
Study authors wrote: “Treatment-related serious adverse event rates were low and comparable across treatment groups, with the most common events leading to treatment discontinuation related to liver toxicity or diarrhea.”
The randomized, double-blind trial tested three doses of rentosertib against placebo over 12 weeks, with patients receiving 60 mg once daily showing a mean forced vital capacity gain of +98.4 mL compared to a -20.3 mL decline in the control group. The compound targets TNIK, a kinase that Insilico’s algorithms identified as a novel regulator of lung fibrosis pathways.
Currently, there are only two FDA-approved therapies for idiopathic pulmonary fibrosis: Pirfenidone (Esbriet) (approved in 2014) and Nintedanib (Ofev), which was approved that same year.
Preclinical and phase I data on rentosertib was published in Nature Biotechnology last year.
Filed Under: Pulmonology
