A pharmaceutical compound designed by artificial intelligence has demonstrated the ability to reverse markers associated with biological aging in human patients, according to new peer-reviewed research. The study, published in Nature Biotechnology in September 2026, examined 42 patients suffering from idiopathic pulmonary fibrosis who were treated with Rentosertib.

Analysis of blood samples revealed that six independent aging clocks, which estimate biological age based on protein patterns, all predicted a younger age for the treated group compared to those on a placebo. The most significant reduction indicated a decrease of up to six years in biological age, though researchers emphasise this reflects protein markers rather than confirmed longevity.

Consistency across independent models

The core finding rests on the agreement between different algorithmic models. Each aging clock was developed by separate teams at institutions including Harvard, Oxford and Peking University, using distinct training data and features. Despite these differences, all six models registered a lower predicted biological age for patients receiving the drug.

Michael Levitt, a Nobel Laureate in Chemistry, pointed to this consistency as the most compelling evidence. He noted that the models do not share their underlying characteristics or training sets, making the convergent result statistically notable. However, he stopped short of declaring the drug a proven anti-aging treatment without further validation.

Disease context complicates interpretation

Significant caution is required when interpreting these results. The trial participants were not healthy volunteers but patients with a serious lung condition that causes scarring of lung tissue. It remains unclear whether the shift in protein markers represents genuine systemic rejuvenation or simply a reduction in the physiological stress caused by the disease.

Vadim Gladyshev from Harvard Medical School described the work as the first study to clearly show predicted biological age can be reduced. Yet he highlighted the small sample size of 42 patients and the lack of data on healthy individuals. Aging clocks are predictive tools, not direct measurements of time lived, and their reliability in this context is still being established.

Dosage effects diverge from lung function

Further analysis suggests the effect on aging markers may be partially independent of the drug's primary function. The dosage that provided the most benefit to lung function was 60 mg once daily. However, the dosage that most significantly lowered predicted biological age was 30 mg twice daily. This divergence implies the compound may interact with aging pathways separately from its effect on pulmonary fibrosis.

Researchers compared the protein profiles of treated patients against more than 55,000 profiles from the UK Biobank. The data showed Rentosertib pushed age-related protein changes in the opposite direction to typical aging trends. While promising, this observation relies on correlational data rather than a controlled measurement of lifespan extension.

Commercial race for AI therapeutics

Insilico Medicine, the Hong Kong-based company behind Rentosertib, is leveraging generative AI to accelerate drug discovery. The firm uses one system to identify disease-relevant proteins and another to generate molecular structures that target them. This process identified the protein TNIK as a target for both aging and lung fibrosis, reducing the timeline from target identification to candidate selection to approximately 18 months.

The commercial stakes are high. Pharmaceutical giant Eli Lilly has invested in Insilico to bring AI-developed medicines to market. As of March 2026, the company claims to have developed at least 28 drug candidates using generative AI, many of which are in clinical trials. Success with Rentosertib could validate the business model for AI-driven pharmacology across the sector.

Regulatory hurdles remain

Despite the positive signals, the drug faces standard regulatory pathways. Rentosertib is currently in a Phase III study, the final clinical trial phase before potential approval for treating idiopathic pulmonary fibrosis. Regulators in Europe and the United States will require robust evidence of safety and efficacy on specific disease endpoints before granting a licence.

Broader claims about anti-aging effects will face even higher scrutiny. European health policy frameworks, such as those managed by the European Commission health policy directorate, do not currently recognise aging as a treatable indication. Any marketing of the drug as an anti-aging therapy would require new regulatory precedents and extensive long-term data.

People mentioned

  • Michael Levitt

    Nobel Laureate in Chemistry, Stanford University

  • Vadim Gladyshev

    Professor, Harvard Medical School

  • Eric Topol

    Cardiologist, Scripps Research

  • Alex Zhavoronkov

    Founder and CEO, Insilico Medicine

Organisations

Insilico Medicine · Harvard Medical School · Nature Biotechnology · Eli Lilly