Single AAV Injection Extends Lifespan and Multi-Organ Health in Aged Mice

The longevity headline of the month is that one shot of a gene therapy added a fifth to the lives of old mice. The number is real. The framing is not quite what the coverage suggests, and the reason is worth more than the number.

A single intramuscular injection of an AAV vector carrying the gene for fibroblast growth factor 21 raised life expectancy by 20.54% in aged mice and preserved function across liver, kidney, heart, skeletal muscle and brain, in a 27-month study led by Professor Fátima Bosch at the Center for Animal Biotechnology and Gene Therapy at the Universitat Autònoma de Barcelona. It ran in Molecular Therapy as Jimenez et al., DOI 10.1016/j.ymthe.2026.05.025, open access.

It was published in mid-June. The reason it surfaced again in August is that Kriya Therapeutics issued a press release on 5 August announcing the publication. Kriya is a clinical-stage biopharmaceutical company. It already has this vector in humans. Not for aging.

The therapy already has a human program, and it is a liver drug

Kriya's candidate is KRIYA-497, a one-time intramuscular AAV gene therapy expressing native FGF21, in development for MASH, specifically later-stage disease with liver-predominant pathology at F3 and compensated F4. It is registered on ClinicalTrials.gov as NCT07732400, also identified as VV-14303, an open-label Phase 1/2 study in adults with MASH. As of early August it had not reported results.

This changes the shape of the story. The aging result is not a preclinical finding waiting for someone to build a translation path. The translation path exists, is funded, is in patients, and is pointed at a liver indication where the benefit-risk case is legible to a regulator. The aging claim is a halo on it.

That distinction matters because a Phase 1/2 safety study in advanced fibrosis patients is not a longevity trial and will not become one by proximity. It will generate exactly the human data the mouse work cannot: immunogenicity against the vector, durability of muscle-driven expression in a person rather than a mouse, and dose tolerance. Those readouts are the real signal. The 20.54% is the thing that gets written about.

Worth noting on disclosure: Bosch sits on Kriya's scientific advisory board, an affiliation stated in the company's own 2024 release on the earlier MASH work. Nothing improper about it. It is simply context that most of the aggregated coverage of the aging paper omits.

What the study actually did

The design deserves the credit it has been given, and more precision than it has received.

Treatment started after decline. Most longevity mouse work loads the dice, dosing young animals or using strains engineered to overexpress the target from birth. This study treated mice at 13, 19 and 22 months of age, after age-related deterioration had set in. That maps far more closely to the clinical reality of treating older adults than an intervention given before anything has gone wrong.

The vector and dose are specific and translatable-adjacent. Animals received 3 × 10¹¹ viral genomes of an AAV serotype 1 vector encoding FGF21, delivered intramuscularly. Serotype and dose are not trivia here. They are the entire question of whether this scales, since AAV1 pre-existing immunity in humans and the vg/kg arithmetic of dosing a 70kg adult rather than a 30g mouse are where muscle-directed gene therapy programs typically run into trouble.

The flagship cohort was metabolically sick, not merely old. The lead experiment used overweight, insulin-resistant 13-month-old chow-fed males. Control animals given a null vector kept gaining weight while treated animals lost it and converged on the body weight of young mice. This complicates the reading. FGF21 is a metabolic regulator and does exactly this to metabolic dysfunction. Some portion of what is being scored as anti-aging is a well-characterized obesity and insulin-sensitivity effect, measured against controls who were getting fatter for the duration.

The lifespan number may be male-only. The paper reports on male and female animals, and coverage has described the 20.54% figure as spanning both. But Lifespan.io's read of the paper is that most experiments including the lifespan arm used only males, with females showing improvement in several domains including cognition. Anyone quoting the headline number should establish which cohort produced it.

The organ report card, and one genuinely good safety finding

Body weight and adiposity fell. Insulin sensitivity, glucose homeostasis and energy expenditure improved. Kidney function held. The cardiac and renal fibrosis and amyloidosis expected at that age did not appear. Muscle strength and endurance were maintained and treated animals outperformed controls on cognitive testing. Transcriptomic analysis points upstream: better mitochondrial function, restored proteostasis, dampened inflammatory and fibrotic signaling, AMPK activation across tissues.

The most interesting result is one that most coverage skipped. FGF21 has a known association with bone loss, and earlier transgenic FGF21 mice had low bone mass. Here, after more than a year of elevated FGF21, bone formation and resorption markers were unchanged. The authors attribute the difference to adult-onset, muscle-restricted expression rather than lifelong systemic overexpression.

That is the strongest evidence in the paper for the argument the study is really making, which is not about the molecule. FGF21 biology has been worked over for years. What is new is the delivery model: muscle as a long-term production depot, one administration, continuous native protein rather than repeat dosing of an analog. Kriya's own rationale for the design says as much, citing steady expression as a way to limit the peak-concentration gastrointestinal effects that have dogged injected FGF21 analogs.

What the mouse data cannot tell you, and what already exists

The gap the draft coverage leaves open is smaller than it looks, because large-animal data exists. Kriya's 2024 MASH work reported durable expression and biological activity in metabolic tissues in dogs. No non-human primate data appears in the available summaries, but "no data beyond mice" is not the correct description of this program.

What remains genuinely unknown sits in humans, and is precisely what NCT07732400 is built to find out. AAV brings pre-existing immunity in a substantial fraction of the population and neutralizing responses that can blunt or block expression, none of which a mouse colony reveals. Durability of muscle FGF21 production across a human lifespan, rather than 27 mouse months, is untested. And a therapy that is irreversible by design carries a different risk calculus in a healthy 60-year-old than in an F4 cirrhotic with limited alternatives, which is one reason the first indication is the one it is.

Bosch has said the results position the approach as a strategy "to promote healthy aging." Fair enough as a research direction. The operative word is strategy, not therapy.

What to watch

Not another rodent cohort. Three things - The Phase 1/2 safety and immunogenicity readout from the MASH study, which will be the first evidence of whether muscle-directed AAV1 delivers durable FGF21 expression in a person against real seroprevalence. Whether expression holds past the first year in humans, since the entire one-shot proposition depends on it. And whether anyone runs the aging question as an actual trial with an aging endpoint, rather than inferring it from a liver program.

Until the first of those reports, the honest description is a strong preclinical result attached to a clinical program that is testing something else. That is considerably more than most longevity headlines have behind them. It is also not the same thing as a therapy for aging.