AI-Designed Drug Shows Lung Function Gains in Idiopathic Pulmonary Fibrosis Trial
Rentosertib, a first-in-class TNIK inhibitor created by generative AI, improved forced vital capacity in IPF patients in a phase 2a trial.
Summary
Rentosertib (ISM001-055), a small-molecule TNIK inhibitor discovered and designed using generative AI by Insilico Medicine, completed a phase 2a randomized, double-blind, placebo-controlled trial in 71 IPF patients across 16 Chinese centers. Patients received 30 mg once daily, 30 mg twice daily, 60 mg once daily, or placebo for 12 weeks. The primary safety endpoint was met with adverse event rates comparable across all groups. Crucially, the 60 mg once-daily group showed a mean FVC increase of +98.4 ml versus a −20.3 ml decline in placebo — a clinically meaningful signal. The trial marks a milestone as one of the first AI-generated molecules to demonstrate efficacy signals in a controlled human trial, with the entire journey from target discovery to phase 2a completed in roughly four years.
Detailed Summary
Idiopathic pulmonary fibrosis is a relentlessly progressive, age-related lung disease with a median survival of only 2–4 years after diagnosis. Current standard-of-care drugs — nintedanib and pirfenidone — slow decline but do not restore lung function or improve survival meaningfully, underscoring the urgent need for novel therapeutic approaches.
Researchers at Insilico Medicine used generative AI platforms to identify TNIK (Traf2- and Nck-interacting kinase) as a novel, first-in-class IPF target and then design rentosertib, a potent small-molecule TNIK inhibitor, entirely de novo. Preclinical candidate nomination took only 18 months, and phase 0/1 testing was completed within 30 months of initiating target discovery — a pace far faster than conventional drug development timelines averaging 10–15 years and $2–3 billion.
In this phase 2a trial (NCT05938920), 71 patients with confirmed IPF were randomized across 16 Chinese medical centers to 30 mg QD (n=18), 30 mg BID (n=18), 60 mg QD (n=18), or placebo (n=17) for 12 weeks. The primary endpoint — percentage of patients experiencing at least one treatment-emergent adverse event — was 72.2%, 83.3%, 83.3%, and 70.6% respectively, with no statistically significant differences between arms. Serious treatment-related adverse events were low across all groups; the most common events leading to discontinuation involved liver enzyme elevations and diarrhea.
For secondary endpoints, the 60 mg QD arm demonstrated a mean FVC change of +98.4 ml (95% CI: 10.9 to 185.9) compared to −20.3 ml (95% CI: −116.1 to 75.6) in the placebo group — a difference of approximately 118 ml that reached statistical and clinical significance. Additional secondary measures including DLCO, FEV1, 6-minute walk distance, Leicester Cough Questionnaire scores, and rates of acute IPF exacerbations were also assessed, along with detailed pharmacokinetic profiling. Pharmacokinetic parameters (Cmax, Ctrough, tmax, AUC, t1/2) supported dose-proportional exposure across regimens.
These results establish rentosertib as safe and well tolerated at all tested doses and provide the first human clinical evidence that TNIK inhibition can yield meaningful functional benefits in IPF. The authors call for larger, longer-duration phase 3 trials to confirm these findings. Notably, this study represents one of the most advanced clinical readouts for a fully AI-generated drug candidate, potentially validating generative AI as a transformative force in drug discovery.
Key Findings
- 60 mg QD rentosertib produced a mean FVC increase of +98.4 ml vs −20.3 ml decline in placebo over 12 weeks.
- Adverse event rates were similar across all arms (~70–83%), with no new safety signals identified.
- AI-driven target discovery and drug design compressed the timeline from target ID to phase 2a to roughly 4 years.
- TNIK is confirmed as a viable first-in-class antifibrotic target validated in a human clinical trial.
- Liver enzyme elevations and diarrhea were the most common reasons for treatment discontinuation.
Methodology
Phase 2a multicenter, double-blind, randomized, placebo-controlled trial across 16 Chinese centers enrolling 71 IPF patients into four arms (30 mg QD, 30 mg BID, 60 mg QD, placebo) for 12 weeks. Primary endpoint was treatment-emergent adverse event incidence; secondary endpoints included FVC, DLCO, FEV1, 6-minute walk distance, Leicester Cough Questionnaire, and detailed pharmacokinetics.
Study Limitations
The trial was small (n=71) and short (12 weeks), limiting conclusions about durability of benefit, survival impact, or long-term safety. The study was conducted exclusively in China, which may limit generalizability. No active comparator arm (e.g., nintedanib or pirfenidone) was included, making it impossible to benchmark rentosertib against current standard of care.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.