On September 7, 2026, *Nature Biotechnology* published a joint study by Insilico Medicine, Harvard Medical School, and collaborators analyzing pre-existing serum proteomics data from 42 patients with idiopathic pulmonary fibrosis (IPF) collected during a 12-week Phase IIa trial. Six independently developed proteomic aging clocks—ProtAge, OrganAge (two variants), PAC, ipfP3GPT, and PAOPAC—unanimously indicated biological age reversal in the treatment group. The peak effect occurred at week 4 in the 30 mg twice-daily arm: 3–4 years of predicted biological age decline, with one clock showing up to nearly 6 years.
The signal comes from Rentosertib (ISM001-055)—the first small molecule with an AI-discovered target and AI-designed structure to reach Phase III trials. The TNIK (TRAF2- and NCK-interacting kinase) target was identified by Insilico's PandaOmics platform at the intersection of fibrosis, inflammation, and aging hallmarks; the molecule was generated by Chemistry42—not repurposed from an existing drug library.
Clinical background: the GENESIS-IPF Phase IIa trial
The Phase IIa trial (NCT05938920) enrolled 71 IPF patients across 22 centers in China:
- Primary endpoint (safety): adverse events 72.2%–83.3% in treatment arms vs 70.6% on placebo—comparable across groups;
- Lung function: the 60 mg once-daily arm showed mean FVC improvement of +98.4 mL vs −20.3 mL on placebo—a meaningful signal in a disease where functional gains are rare.
- Dose with maximal lung function improvement: 60 mg once daily (+98.4 mL FVC);
- Dose with maximal aging-clock reversal: 30 mg twice daily (3–4 years; up to ~6 years on ipfP3GPT).
- Suppression of core senescence drivers including EREG, ESM1, IGFBP4, ITGA2, MMP10, MMP13, SPP1;
- Downregulation of RTK–PI3K and RAS–ERK pathways linked to accelerated aging;
- Modulation of antioxidant and cholesterol metabolism.
- Prospectively incorporate aging/senescence biomarkers as exploratory endpoints in disease trials;
- Pursue biomarker qualification via the FDA Biomarker Qualification Program and the FDA–NIH BEST framework;
- Potentially accelerate geroprotective drug discovery by years or decades compared with drug-repurposing approaches.
- Disease vs aging effects not fully separable: 42 IPF patients cannot cleanly isolate treating-the-lung effects from systemic rejuvenation. A healthy-volunteer trial—not yet announced—is the definitive experiment, as Nobel laureate Michael Levitt emphasized;
- Clocks were not pre-specified endpoints: aging-clock models were not designated as primary endpoints before the trial began;
- Senomorphic, not senolytic: the drug suppresses senescent-cell signaling without clearing the cells, so durability after discontinuation remains unknown;
- The week-12 plateau is unexplained.
- H1 2026 revenue ≈ US$106 million, up 287% year-over-year;
- Adjusted net profit > US$51 million—its first profitable half-year as a listed company;
- ~US$7.3 billion in 2026 deal total value (~US$11 billion cumulative since 2021), with partners including Eli Lilly, Servier, Takeda, SK Biopharmaceuticals, Qilu Pharmaceutical, Hygtia Therapeutics, CMS, and Tenacia;
- 9 development candidates nominated in 9 months in 2026, a company record.
- AI-designed drugs can act beyond a single intended indication;
- Aging endpoints can be embedded prospectively in routine disease trials;
- Dose dissociation is a direct tool for distinguishing "healthier, therefore younger" from genuinely younger.
The Phase IIa results were published in *Nature Medicine* on June 3, 2025. The new *Nature Biotechnology* paper reanalyzed the trial's pre-existing serum proteomics: longitudinal data from 42 participants covering 2,841 proteins, run through the six independent clocks—all pointing toward predicted biological age decline in the treatment group.
Dose dissociation: why "patients got healthier, so they looked younger" is questionable
The most striking finding:
These doses are different. If the aging signal were merely a downstream consequence of improved lung function, the two dose effects should overlap. The dissociation suggests Rentosertib's geroprotective activity is, at least partly, independent of its anti-fibrotic effect.
The team also cross-validated against 55,319 UK Biobank plasma proteomic profiles: Rentosertib directly reversed several canonical age-related protein expression trajectories.
Mechanism: a senomorphic agent
Mechanistic data support a senomorphic profile (suppressing senescent-cell signaling rather than killing senescent cells):
The effect plateaued by week 12—a dynamic the researchers explicitly note remains unexplained.
A reusable framework: aging endpoints in routine disease trials
Beyond the drug itself, the paper's key contribution is a development template other teams can reuse:
Data are deposited at the China National Center for Bioinformation (accession OMIX008341), and the analysis pipeline is released as open-source Python libraries on GitHub.
Limitations
Context: a financially established AI pharma company
Insilico Medicine's backdrop is notable:
Why this matters
Compared with prior AI-in-drug-development milestones (AlphaFold-style structure prediction, virtual cells like MIT's MAP and Arc Institute's State models), Rentosertib represents a new category: an AI-discovered, AI-designed first-in-class molecule showing consistent aging-clock signal in clinical trial data. The implications:
References
1. *Nature Biotechnology*: Insilico's AI-Driven IPF Candidate Rentosertib Shows Potential for Biological Age Reversal, as Assessed by Six Proteomic Aging Clocks, 2026-09-07 2. *Nature Medicine*: GENESIS-IPF Phase IIa primary readout (60 mg QD FVC +98.4 mL), 2025-06-03 3. Insilico Medicine H1 2026 financial disclosures and pipeline milestones 4. ClinicalTrials.gov: NCT05938920 (GENESIS-IPF) 5. CNCB-OMIX data accession: OMIX008341