A paper published on 14 July 2026 in Nature Communications describes an engineered enzyme that reduced a specific chemical marker of tissue aging in donated human skin to levels below what is typically found in the skin of a 31-year-old. In a separate experiment using the abdominal aorta of a 75-year-old donor, the same enzyme cut the marker by more than seventy per cent overnight.
The paper, titled “Reversal of protein chemical aging by enzymatic deglycation,” comes from a team led by Aaron Cravens at Revel Pharmaceuticals, a private biotech firm, working with researchers at Calico Life Sciences and at the University of Colorado Anschutz Medical Campus. This is one study, not settled consensus, and it was conducted entirely on tissue removed from the body: thin sections of donor skin, an isolated donor aorta, and purified proteins in a dish. No animal model and no living patient has been treated with this enzyme.
What the enzyme actually targets
The damage in question is a compound called Nε-carboxymethyl-lysine, usually abbreviated CML. It belongs to a family of molecules known as advanced glycation end-products, or AGEs, which form when sugars in the body react with proteins in a slow, unregulated process chemists sometimes compare to the browning of bread in a toaster: the same non-enzymatic reaction, the Maillard reaction, that browns a crust also builds up, gradually and irreversibly, inside collagen, artery walls, and the lens of the eye. Unlike most cellular damage, the body has no dedicated repair system for it. Once an AGE cross-link forms, it has, until now, simply stayed.
Revel’s researchers built an enzyme, which they call CMLase, by taking a bacterial glycine oxidase found in soil bacteria and running it through directed evolution: a laboratory process of introducing random mutations, testing the results, and repeating the cycle across generations of the enzyme. The paper describes screening more than 500 million variants to arrive at a version capable of recognising and breaking down CML specifically, without the broader collateral damage a less selective compound might cause.
What the two figures actually measure
It matters that the “31-year-old” figure and the “seventy per cent” figure come from two different tissues and two different donors, not one. In the skin experiment, CML staining across both the epidermis and dermis fell by more than fifty-five per cent after treatment, bringing the treated tissue below the CML levels ordinarily seen in skin from a 31-year-old. The exact age of the donor whose skin was treated is not specified in the paper beyond a general range spanning several decades. In the artery experiment, the tissue came specifically from a 75-year-old donor, and CML levels in the aorta fell by more than seventy per cent after overnight treatment. A third experiment, on lens tissue from a 64-year-old donor, showed reductions of 45 to 78 per cent depending on the measurement method used. Reporting these as belonging to one donor, or one tissue, would overstate what the paper shows.
The authors also tested the enzyme against a broader set of proteins in isolation. Across the panel of modification sites they examined, the large majority showed some reduction in CML, and several showed reductions greater than ninety per cent. Reduction levels varied considerably by protein, from around fifty per cent in one common blood protein to nearly ninety-seven per cent in casein, a milk protein often used as a laboratory reference.
Why this field has reason to be cautious
The idea of chemically breaking AGE cross-links is not new, and its history is not encouraging. Alagebrium, a compound developed by Alteon Inc. under the name ALT-711 in the early 2000s, was designed to do something similar and showed early promise in small trials, including improvements in how well the heart’s left ventricle filled with blood between beats. Alteon’s later trial programme was never completed, and the company’s finances collapsed before the drug reached approval. Cravens and his co-authors cite this history directly in the paper, which is itself notable: it would have been easy to present CMLase as an unprecedented advance rather than the latest attempt at a problem with a documented record of failure.
The authors are equally direct about what remains unknown. Whether removing CML restores any physical property of aged tissue, such as the stiffness of an artery wall, has not been tested. Whether the enzyme would penetrate an intact, living organ the way it penetrated a thin tissue section or an isolated protein sample is untested. CML is also only one AGE among several; glucosepane, considered the dominant cross-linking AGE in ageing collagen, is not a target of this enzyme at all. And because the enzyme originates in bacteria, the paper flags immunogenicity, the risk that a person’s immune system would react against it on repeated dosing, as an open question rather than a solved one.
A finding, not a treatment
Revel Pharmaceuticals is a privately held company that has previously reported a $3.8 million grant from the US National Institutes of Health and a $12 million seed funding round. It is publishing on its own drug candidate, which is a normal and disclosed part of how early biotech research reaches print, but it is also a reason to read the framing of the result with some care. No clinical timeline appears in the paper. Outside comment on the work so far has been measured: pharmacologist James Galligan, quoted in coverage of the paper, described the approach as “pretty bold” rather than settled.
What the paper adds to the record is narrower and more interesting than a headline about reversed ageing would suggest. For four decades, CML accumulation has been treated in the biochemistry literature as one of the fixed costs of getting older, a form of damage the body simply cannot undo. This result does not overturn that literature. It gives one research group, working on donated tissue outside the body, a reason to ask whether the assumption of permanence was ever really tested as thoroughly as it was assumed.
