First AI-Discovered Drug With Anti-Aging Rationale Enters Landmark Phase 3 Trial for Pulmonary Fibrosis

The pharmaceutical industry has crossed a historic threshold. For the first time, a medication in which both the biological target and the molecular structure were discovered entirely by artificial intelligence has entered a Phase 3 clinical trial.[1]

The drug, known as rentosertib (formerly ISM001-055), was developed by the clinical-stage biotechnology company Insilico Medicine. It is designed to treat idiopathic pulmonary fibrosis (IPF), a devastating and progressive lung disease that currently has no cure.[1][2]

Reaching Phase 3—the final, large-scale human testing stage before regulatory approval—marks a graduation point for the AI drug discovery sector. The field has spent years raising billions of dollars on the promise that machine learning could revolutionize how medicines are made; now, that promise faces its ultimate clinical test.

To understand the significance of rentosertib, one must first understand the severity of idiopathic pulmonary fibrosis. IPF is a chronic condition characterized by the progressive scarring, or fibrosis, of lung tissue.[2]

As this scar tissue accumulates, the lungs become stiff, making it increasingly difficult for patients to breathe and absorb oxygen. The disease disproportionately affects older adults and carries a grim prognosis, with a median survival rate of just three to four years after diagnosis.

Currently, approximately 5 million people worldwide live with IPF. While existing antifibrotic medications can slow the decline in lung function, they cannot halt or reverse the scarring process, leaving a massive unmet medical need for disease-modifying therapies.[1]

This is where Insilico Medicine's AI platform intervened. Rather than starting with a known biological target and screening thousands of existing chemical compounds—the traditional pharmaceutical approach—the company utilized a "biology-first, aging-informed" AI workflow.[1]

The process began with PandaOmics, Insilico's AI-powered biology engine. The system ingested massive datasets of omics and clinical data related to tissue fibrosis, aging, and disease pathways. It identified 20 potential targets before prioritizing a novel intracellular protein called TNIK (Traf2- and Nck-interacting kinase).[1][2]

TNIK is a kinase that the AI linked directly to the signaling pathways driving lung scarring, chronic inflammation, and cellular senescence—the biological hallmarks of aging. By targeting a mechanism so closely tied to the aging process itself, the researchers hypothesized they could address the root cause of the fibrotic cascade.[1]

Once TNIK was identified as the target, Insilico deployed Chemistry42, its generative chemistry platform. Using deep learning algorithms, Chemistry42 "imagined" and generated entirely new molecular structures designed specifically to inhibit the TNIK protein.[1][2]

The result was rentosertib, a first-in-class oral small-molecule inhibitor. The entire discovery process—from target identification to the nomination of a preclinical candidate—was completed in under 18 months, a fraction of the time and cost typically required in traditional drug development.[2]

The underlying claim of AI drug discovery is not just speed, but improved efficacy. The evidence supporting rentosertib's progression comes from a randomized, double-blind Phase IIa clinical trial, the results of which were published in the journal Nature Medicine in 2025.[2]

The GENESIS-IPF trial enrolled 71 patients across 22 sites. The study successfully met its primary endpoints for safety and tolerability, showing that the AI-generated molecule did not introduce unexpected toxicities.

More importantly, the trial provided an early efficacy signal. Patients receiving a 60 mg once-daily dose of rentosertib demonstrated a mean improvement in forced vital capacity (FVC)—a standard measure of lung function—of +98.4 milliliters at 12 weeks. In contrast, the placebo group experienced a decline of -20.3 milliliters.

These mid-stage results provided the confidence to advance to Phase 3. The newly initiated pivotal trial is a prospective, randomized, double-blind, placebo-controlled study that will enroll 320 patients with IPF across 47 centers in China.[1]

Participants will receive once-daily rentosertib over a 52-week period. The primary endpoint will measure the annual rate of decline in forced vital capacity, mirroring the rigorous registrational standards required by regulatory agencies. The trial is being led by prominent respiratory experts, including Professor Zuojun Xu and Academician Nanshan Zhong.[1][2]

Despite the optimism, significant uncertainty remains. Phase 3 trials are notoriously the graveyard of pharmaceutical development, where approximately 40% to 50% of drugs that succeed in Phase 2 ultimately fail when tested in larger, more diverse populations over longer periods.

Industry analysts note that while AI has undeniably transformed the "front end" of drug discovery—finding targets and designing molecules—it has yet to prove it can meaningfully reduce the Phase 3 efficacy failure rate. Other prominent AI drug companies have previously seen lead candidates stumble in mid-to-late-stage trials when early efficacy signals failed to hold.

For patients with IPF, rentosertib represents a beacon of hope for a disease-modifying treatment that could potentially reverse lung scarring. For the biotechnology sector, the next 52 weeks will serve as a definitive referendum on whether artificial intelligence can truly engineer better medicines.[2]