DAILY SERIES · MEDICAL INNOVATIONS
Rentosertib: The First Drug Designed by AI
No person invented this drug: an artificial intelligence designed it — and it has already been tested in patients. The AI pointed to a target almost no one was watching (the TNIK protein) and drew the molecule in about 18 months. In this video: the 71-patient trial, the biological-age clocks that read 3–4 years lower, and the honest fine print almost nobody tells you.
Educational content prepared with official health sources. Professional medical review pending.
No personalized diagnoses: consult your doctor.
📅 Published October 22, 2026 · 2-minute video · Full transcript included
Watch the direct video (without leaving this page) →
A different birth
The usual path is slow: someone picks a target in the body, tests thousands of molecules, and almost all of them fail. Here, the AI did both hard parts: it read data from diseased lungs and from aging itself, pointed to a brand-new target — a protein called TNIK — and drew a molecule made to measure to block it: rentosertib. From the target to the molecule, about 18 months went by, when the traditional path takes years of trial and error (Ren et al., Nature Biotechnology, 2025).
The disease and the trial
The target was a hard disease: idiopathic pulmonary fibrosis. The lung slowly hardens, scarred from the inside, and breathing gets harder every day. The phase 2a trial (Xu et al., Nature Medicine, 2025) enrolled 71 patients for 12 weeks, with a comparison group that got a placebo. The first thing measured was safety, and it held up. And as a secondary finding, at the highest dose breathing capacity (forced vital capacity) improved +98.4 mL on average, while on placebo it worsened −20.3 mL.
The clocks that read younger
Then came the part that made all the noise. Blood from 42 of those patients was run through six proteomic biological-age clocks: programs that estimate the body's age by measuring 2,841 proteins. All six read a lower predicted biological age in the treated patients, peaking at week 4: at their best moment, 3–4 years lower on the chronological clocks (Zhavoronkov et al., Nature Biotechnology, 2026).
The honest fine print — and it's big
This was an exploratory analysis, with few patients and only a few weeks. Not even the authors can fully separate whether the clock went down because the body got younger, or simply because the lung got better. And several of the authors work for the company that owns the drug: it's disclosed in the paper, and it matters. To be clear: there is no youth pill today. What does exist, for the first time, is a drug that a machine designed from start to finish, and that made it all the way to being tested in people.
Ask this at your next appointment
"Doctor, what real advances exist for my case, and which ones are still experiments?"
Asking that, calmly, is also taking care of yourself.
Frequently asked questions
Can I get rentosertib today?
No. It's an experimental phase 2a drug: it's only used inside controlled clinical trials. Anyone offering it outside a study is outside both the law and the science.
Does rentosertib make you younger?
Not proven. The proteomic clocks read 3–4 years lower in an exploratory analysis of 42 patients, and the data can't separate rejuvenation from lung improvement. It's an interesting signal, not a conclusion.
What is biological age?
The age your biomarkers suggest, as opposed to the calendar. Proteomic clocks estimate it from thousands of blood proteins; they're evolving research tools, not a verdict on your body.
Will AI design every drug now?
This case shows AI can speed up discovery (18 months from target to molecule), but testing in people remains human, slow, and necessary: rentosertib still has to clear larger phases before any approval.
Referencias
- Ren F, Aliper A, Chen J, et al. A small-molecule TNIK inhibitor targets fibrosis in preclinical and clinical models. Nature Biotechnology 2025;43(1):63–75. doi:10.1038/s41587-024-02143-0 · PMID: 38459338 · El blanco TNIK se identificó con un enfoque predictivo de IA y la molécula se generó con metodología de IA: unos 18 meses del descubrimiento del blanco a la nominación del candidato preclínico.
- Xu Z, et al. A generative AI-discovered TNIK inhibitor for idiopathic pulmonary fibrosis: a randomised phase 2a trial. Nature Medicine 2025;31(8):2602–2610. doi:10.1038/s41591-025-03743-2 · PMID: 40461817 · Registro del ensayo: NCT05938920 · 71 pacientes, 12 semanas; objetivo primario de seguridad cumplido; FVC +98.4 mL con 60 mg frente a −20.3 mL con placebo (resultado secundario).
- Zhavoronkov A, et al. Análisis con relojes proteómicos de edad biológica en los participantes del ensayo de rentosertib. Nature Biotechnology 2026 (publicado en línea el 7 de septiembre de 2026). doi:10.1038/s41587-026-03286-y · 42 pacientes con proteómica completa, 2,841 proteínas, seis relojes; los seis marcaron reducción de la edad biológica predicha en los tratados, con pico en la semana 4 (3–4 años menos en los relojes cronológicos en su mejor momento). Análisis exploratorio.
Cómo leer el ruido: que un reloj proteómico marque menos NO demuestra rejuvenecimiento. El análisis fue exploratorio, con pocos pacientes y pocas semanas, y ni los autores pueden separar si la señal viene de un cuerpo más joven o de un pulmón que mejoró. Además, varios autores trabajan en la empresa dueña del medicamento — está declarado en el paper, y también se pondera aquí.
Full video transcript
This drug was not invented by a person. It was designed by artificial intelligence. And it has already been tested in real patients.
Usually, a new drug takes years to be born: someone picks a target in the body, tests thousands of molecules, and almost all of them fail. Here, the AI did both hard parts: it read data from diseased lungs and from aging itself, and it pointed to a target almost no one was watching, a protein called TNIK. Then it drew a molecule made to measure to block it: rentosertib. From the target to the molecule, about 18 months went by.
And the target was a hard disease, idiopathic pulmonary fibrosis: the lung slowly hardens, scarred from the inside, and breathing gets harder every day. In 2025, the patient trial was published in Nature Medicine: 71 people, 12 weeks, with a comparison group that got a placebo. The first thing measured was safety, and it held up. And at the highest dose, breathing capacity improved on average, while on placebo it got worse.
Then came the part that made all the noise. The blood of those patients was run through six biological age clocks: programs that estimate the body's age by measuring thousands of proteins. At their best moment, the clocks read three to four years younger.
And here is the honest catch, and it is a big one: this was an exploratory analysis, with few patients and only a few weeks. Not even the authors can fully separate whether the clock went down because the body got younger, or simply because the lung got better. And several of the authors work for the company that owns the drug; that is disclosed, and it matters.
There is no youth pill today. But there is, for the first time, a drug that a machine designed from start to finish, and that made it all the way to being tested in people. Save this question for your next appointment: "Doctor, what real advances exist for my case, and which ones are still experiments?" Asking that, calmly, is also taking care of yourself.
🎬 Video from Red de Salud Hispana's daily series, Medical Innovations line. Also on YouTube, Facebook, and Instagram.
This opens a new line: medical innovations
Real scientific advances, explained plainly and with the study cited: no invented miracles, no inflated headlines. Every episode of the series — the doctor, wounds, veins, spine, and innovations — lives in one place.
See all educational videos →