Insilico's AI-designed fibrosis drug rentosertib enters Phase III in China
· TexaHub · Biotech & Med-Tech, AI in Industry, Artificial Intelligence
Insilico Medicine has opened a 320-patient Phase III trial of rentosertib, testing whether an AI-found target and AI-generated molecule can hold up in late-stage development.
What was announced
On 7 July, Insilico Medicine announced the initiation of a Phase III clinical trial of rentosertib, an oral small-molecule inhibitor of the kinase TNIK that the company is developing for idiopathic pulmonary fibrosis (IPF) [1]. The compound, previously known as ISM001-055 and INS018_055, was discovered and designed through the company's Pharma.AI platform, and Insilico describes it as potentially first-in-class [1]. Rentosertib is the first Insilico asset to enter Phase III [1].
Trade coverage the following day framed the step as the most advanced clinical milestone yet for Insilico's AI-enabled discovery platform [4]. Insilico calls it "a major late-stage milestone for AI-driven drug discovery": a medicine whose target was identified with AI and whose chemical structure was designed with generative AI [1]. The release stops short of claiming it is the first such drug in Phase III; any such claim would depend on how strictly "AI-designed" is defined.
Trial design
The study is prospective, randomised, double-blind, placebo-controlled and parallel-group, with once-daily rentosertib given for 52 weeks [1]. It plans to enrol 320 patients at 47 centres in China [1]. The primary endpoint is the annual rate of decline in forced vital capacity (FVC) over 52 weeks, and the key secondary endpoint is time to first disease progression event [1]. One trade report described eligibility as patients over 40 with an IPF diagnosis and restricted lung volume, and said a long-term open-label extension will be available after the trial [3].
The leading principal investigator is Professor Zuojun Xu of Peking Union Medical College Hospital, with Academician Nanshan Zhong and Professor Chang Chen of Shanghai Pulmonary Hospital as co-leading investigators [1][2]. Coverage of the expected timeline is less consistent. Clinical Research News reported a schedule running to last patient enrolled in November 2028, initial results in December 2029, a potential new drug application in March 2030 and a target approval date of September 2030, but the same schedule lists first patient enrolment as September 2025, which does not fit a trial announced this month [3]. Until Insilico publishes an updated timetable, those dates are best treated as indicative.
The trial is being run entirely in China, as was the 71-patient, 22-site Phase IIa study [1]. That should help recruitment, but the US Food and Drug Administration, which granted rentosertib Orphan Drug Designation in February 2023, would likely want to see how results translate to other patient populations before any US filing [1].
The evidence so far
The Phase III rests on a 12-week Phase IIa study published in Nature Medicine in June 2025 [5]. That trial randomised 71 patients to 30 mg once daily (18), 30 mg twice daily (18), 60 mg once daily (18) or placebo (17) [5]. Its primary endpoint was safety: the share of patients with at least one treatment-emergent adverse event was 72.2%, 83.3% and 83.3% in the three rentosertib arms and 70.6% on placebo [5]. The authors reported that the most common events leading to discontinuation were related to liver toxicity or diarrhoea [5].
On efficacy, the 60 mg once-daily group showed a mean FVC change of +98.4 ml, with a 95% confidence interval of 10.9 to 185.9, against -20.3 ml (95% CI -116.1 to 75.6) on placebo [5]. The overlapping intervals and small arms are why the authors themselves concluded only that the drug warranted larger, longer trials [5]. Insilico's release describes a dose-dependent efficacy signal and exploratory biomarkers consistent with anti-fibrotic and anti-inflammatory effects [1].
The disease context explains the interest. Insilico cites around 5 million people affected by IPF worldwide and a median survival of three to four years, and notes that current therapies slow progression without reversing it [1]. Chief executive Alex Zhavoronkov told Clinical Research News: "If we outperform the standard of care and show a disease modifying effect with this drug, it's going to be the first and only of its kind" [3].
What it tests about AI discovery
According to Insilico, TNIK was ranked as a top candidate by its PandaOmics target-discovery engine, which integrates multi-omics data, biological networks and causal inference, and the molecule was then generated with its Chemistry42 chemistry platform [1][2]. The company reports 31 preclinical candidate nominations, 13 IND clearances and eight programmes in Phase I, and claims an average of 12 to 18 months to nominate a preclinical candidate against a traditional 2.5 to 4 years [1]. Zhavoronkov gave a range of 9 to 18 months for the same step [3].
Speed to candidate is the part of the pipeline where AI tools have the clearest record. A Phase III readout tests something harder: whether an AI-prioritised target is biologically right in a large population over a full year. A positive result would be the first late-stage evidence that a target nominated by such software can change a disease course; a negative one would not by itself discredit the tools, but it would sharpen questions about target selection, which no current platform can fully check before human trials.
The wider field is well funded but mostly earlier-stage. In May, Alphabet's Isomorphic Labs raised a $2.1 billion Series B led by Thrive Capital to move its in-house pipeline into the clinic, without giving a public timeline [6].
The regulatory backdrop
How a drug was discovered does not change the evidential bar for approval; the FDA's draft guidance on AI in drug development focuses on the credibility of AI models used to support regulatory decisions on safety, effectiveness and quality [7]. When it issued that draft in January 2025, the agency said it had reviewed more than 500 drug and biological product submissions with AI components since 2016 [7]. The guidance remains a draft [7].
For now, the practical questions are operational ones: how quickly 47 Chinese sites can recruit 320 patients, and whether the liver and gastrointestinal signals seen in 12 weeks stay manageable over 52. Insilico has not yet reported first patient dosing in the Phase III.
Sources
[1] Insilico Initiates Phase III Clinical Trial for Rentosertib, Its AI-Empowered TNIK Inhibitor for Idiopathic Pulmonary Fibrosis — Insilico Medicine via PR Newswire, 7 Jul 2026
[2] Insilico Medicine launches Phase III trial of AI-designed Rentosertib drug — Drug Target Review, 8 Jul 2026
[3] Insilico Medicine Launches Phase III Trial for AI-Developed Drug — Clinical Research News, 8 Jul 2026
[4] Insilico starts Phase III trial of AI-discovered IPF therapy — BioSpectrum Asia, 8 Jul 2026
[5] A generative AI-discovered TNIK inhibitor for idiopathic pulmonary fibrosis: a randomized phase 2a trial — Nature Medicine, 3 Jun 2025
[6] Alphabet's AI biotech Isomorphic Labs bags $2.1B series B to fuel next-gen drug design model — Fierce Biotech, 12 May 2026
[7] FDA Proposes Framework to Advance Credibility of AI Models Used for Drug and Biological Product Submissions — US Food and Drug Administration, 6 Jan 2025