X

AI-Designed Drug for Lung Disease Shows Signs of Turning Back Biological Clocks

Insilico Medicine built a drug from scratch with artificial intelligence. The compound, rentosertib, already demonstrated it could improve lung function in patients with idiopathic pulmonary fibrosis. Now fresh data from that same trial suggests something more striking: the drug appears to make patients’ blood proteins look younger.

Six independent aging clocks, built by separate research groups, all pointed the same way. Treated patients showed reductions in predicted biological age. Placebo patients did not. The effect peaked at about three to four years on several clocks, with one reading as much as six years younger after four weeks.

But. The patients suffered from a disease that accelerates aging-like processes in the lungs. Researchers caution they cannot yet disentangle whether the shift reflects true geroprotection or simply better lung health. Still, the consistency across clocks stands out.

The study, published in Nature Biotechnology, marks the first time multiple proteomic aging clocks have been tested head-to-head in a clinical drug trial. It also represents one of the earliest examples of a molecule both discovered and designed by AI showing potential effects on aging markers in humans.

Alex Zhavoronkov, founder and co-CEO of Insilico, called the findings a milestone. He will present them Tuesday at a Nature conference in Paris. “This is the first study that shows, very clearly, that predicted biological age can be reduced,” said Vadim Gladyshev, Harvard Medical School professor and developer of one of the clocks, in The New York Times.

The trial itself was small. Of 71 patients in the Phase 2a study for IPF, 42 provided blood samples for proteomic analysis. Researchers measured 2,841 proteins using an Olink panel at baseline and at weeks two, four, and 12. They then fed the data into six clocks developed by teams at Harvard, Oxford, Peking University, and Insilico itself. All six showed the treated groups trending younger while placebo held steady or worsened slightly.

Dosing mattered. The 30-milligram twice-daily arm produced the broadest agreement across clocks. The 60-milligram once-daily dose, which had shown stronger effects on lung function in the original trial, produced slightly smaller age reductions on some measures. Forced vital capacity, a key lung function metric that declines with age, improved in a dose-dependent manner alongside the proteomic shifts.

Insilico identified TNIK, the target protein, using its PandaOmics AI platform. The system flagged the gene as connected to multiple hallmarks of aging and to fibrosis. Generative AI through the company’s Chemistry42 platform then produced the small-molecule inhibitor in roughly 18 months. The entire process from target to candidate was faster than traditional methods.

From Target Discovery to Dual-Purpose Signals

This integration of AI at every stage sets rentosertib apart from earlier longevity compounds. Most prior candidates tested in humans were repurposed drugs like rapamycin or metformin. Here the target and molecule both originated from AI models trained on aging and disease data.

The clocks themselves rely on different training approaches. Some predict chronological age from protein patterns in large populations like the UK Biobank, which supplied more than 55,000 reference profiles. Others forecast mortality risk or organ-specific aging. Their agreement surprised even the authors. Different clocks often disagree on the same individual. Not this time.

Yet limitations remain clear. IPF compresses decades of aging processes into a few years. Improving lung scarring could shift systemic markers without addressing aging itself. The authors note in the paper that larger studies in healthier populations will be needed to separate the two. They also highlight that the biological age reduction appeared strongest at week four and partially attenuated by week 12.

Recent coverage underscores both the excitement and the caution. Bloomberg reported that patients on the drug reversed biological signs of aging while those on placebo saw little change. The Next Web emphasized that the authors directly acknowledge they cannot yet rule out the effect stemming from treated lungs.

Zhavoronkov has long argued that fibrosis drugs could serve as a proving ground for anti-aging therapies. The disease’s rapid progression makes signals easier to detect. Success there might translate to broader age-related conditions. Insilico has already advanced rentosertib into Phase 3 for IPF in China. Additional trials focused specifically on aging biomarkers could follow.

Other recent work adds context. A Nature article from earlier this week described how the weight-loss drug semaglutide slowed aging in female mice, improving physiological function and extending lifespan. That study, like this one, connects metabolic and inflammatory pathways to longevity. But human data remains sparse.

Meanwhile, autonomous AI agents scanning molecular profiles have identified hundreds of potential interventions that appear to reduce biological age in preclinical data, according to a 2025 preprint analyzed in recent discussions. These computational efforts increasingly guide where companies like Insilico look next.

The rentosertib results arrive as the biotech industry watches AI drug discovery mature. Insilico has nominated multiple candidates in 2026 alone and reported its first profitable half-year. Major pharmaceutical players have signed collaboration deals worth billions. The question now shifts from whether AI can generate molecules to whether those molecules can deliver meaningful clinical outcomes beyond their primary indication.

So far the data intrigues. Six clocks converging on the same signal in a real human trial represents a rare alignment in a noisy field. It does not prove people will live longer or even feel younger. It does suggest that tracking proteomic changes in standard disease trials could become routine. That alone might accelerate the search for compounds that address aging biology.

Investors, researchers, and regulators will scrutinize the Phase 3 IPF data closely. If the lung benefits hold and the proteomic signals persist, rentosertib could open a new chapter. One where AI not only speeds discovery but helps rewrite the timeline of human decline. The clocks are ticking. For once, they may be ticking slower.

Web & IT News Editor:

This website uses cookies.