What happened?

AI-driven pharmaceutical company Insilico Medicine reports signs that one of its AI-designed drug candidates reverses markers of biological ageing. The study was published in Nature Biotechnology.

The drug — rentosertib — was not developed for ageing but for idiopathic pulmonary fibrosis (IPF), a disease that turns lung tissue into scar tissue. A Phase IIa trial last year with 42 patients showed improved lung function, and blood samples were collected during the same study. The new analysis rests on that data.

Six independent "ageing clocks"

Ageing clocks are AI models that predict a body's biological age from blood proteins. The researchers ran six of them at once on the patient data, built separately by teams at Harvard, Oxford, Beijing and Insilico itself.

  • All six predicted a lower biological age for treated patients than for those on placebo.
  • The strongest effect was a drop of three to four years by week four; on one clock the gap reached six years.
  • No notable change appeared in the placebo group.
  • The clearest result came in the group taking 30 milligrams twice daily.

Nobel laureate in chemistry Michael Levitt puts the emphasis somewhere other than effect size: "What convinces me is not the size of the effect but the agreement, because these models share neither their features nor their training data."

But what does this not mean?

The result does not mean patients measurably got younger. It only shows that their blood protein patterns shifted in ways the models read as a younger biological age.

The values might look younger simply because the lungs work better and the body is under less overall strain. There is one sign against that: the dose that helped the lungs most (60 mg once daily) differs from the one that cut predicted biological age most (30 mg twice daily), suggesting an effect at least partly independent of lung function.

The researchers also compared treated patients' blood proteins against more than 55,000 profiles in the UK Biobank, a database tracking how proteins typically change with age. According to the company, rentosertib reversed exactly those changes. These are correlations, not proof.

What do experts say?

Cardiologist Eric Topol finds the drug "encouraging" but adds: "We do not yet have a definitive trial to make the final judgment." Vadim Gladyshev of Harvard Medical School points to the small sample size and to the fact that biological clocks are not always reliable — while still calling it "the first study that shows, very clearly, that predicted biological age can be reduced." What is missing is a trial in healthy people; the results may apply only to patients with the lung disease.

What exactly did the AI do here?

Insilico used two separate systems: one combs through health data and scientific literature for disease-relevant proteins, while the other analyses their structure and generates molecules that can bind to them. Chief executive Alex Zhavoronkov likens it to studying the structure of a lock and producing a new key that fits.

The protein TNIK was chosen as the target. Going from target to drug candidate took about 18 months. Rentosertib is now in Phase III for IPF — the final clinical stage before potential approval.

What's next?

Insilico is testing the drug in a larger group: a study planned across 47 centres in China with 320 patients followed for 52 weeks. Founded in Hong Kong in 2014, the company says it had developed at least 28 drug candidates with generative AI as of March 2026; pharma giant Eli Lilly recently invested in it.