Longevity, genes,
and disease-modifying AI.

Bigger than CRISPR? A guide to the latest genome editors (Nature; Heidi Ledford, 28 Sep 2026).

Nature’s technology feature maps tools that aim to install whole genes or large DNA segments, including PASTE and integrases, CRISPR-associated transposases (evoCAST), recombineering, and R2 retrotransposons, and explains why delivery and complexity still bottleneck clinical use. Cas9, base editors, and prime editors remain the clinical workhorses for small edits. Large-gene therapies that use one construct for many alleles are still mostly preclinical.

Caveat: This is not a claim that aging or disease is cured.

Life Biosciences ER-100, a first-in-human Phase 1 OSK gene therapy for optic neuropathies (first patient dosed June 2026; registry shows recruiting).

This delivers OCT4, SOX2, and KLF4 (OSK) by AAV, with oral doxycycline for about 8 weeks, and tests safety and tolerability in open-angle glaucoma and NAION (NCT07290244, about 18 patients planned). The sponsor says the first participant was dosed on June 9, 2026.

Caveat: This is a Phase 1 eye safety study. It is not whole-body age reversal, not an approved longevity product, and no human efficacy results had been posted as of a late-August 2026 review.

MINT, retargeted Bxb1 serine integrases for one-step large DNA integration at AAVS1 and TRAC (Nature Biotechnology, published 29 Jun 2026).

Sangamo’s modular integrase (MINT) system retargets Bxb1 without a pre-installed attachment site. It reported about 29% integration at AAVS1 and about 35% at TRAC in K562 cells, and about 29% GFP integration at TRAC in primary human T cells when paired with activity-boosting mutants and zinc-finger domains.

Caveat: This is a preclinical genome-engineering paper, not a clinical efficacy result.

enIS621-tebRNA, an engineered IS621 bridge recombinase that inserts kilobase-scale DNA in human cells, including CD19-CAR and Factor IX cassettes (Nature Communications; published 6 Jun 2026, version of record 24 Jul 2026).

Rational engineering of the protein and the bridge RNA yields enIS621-tebRNA, which makes scarless, site-specific insertions (up to about 10 kb tested, with efficiency varying by locus and cargo). Functional demos include CD19-CAR insertion in Jurkat cells with an IL-2 response, and human F9 insertion in Huh-7 cells with Factor IX secretion.

Caveat: This is preclinical, and off-target integration risk is still being characterized. It is not an approved therapy.

Research AI is not clinical cure claims.

A Nature technology survey, a Phase 1 eye-safety reprogramming trial with no efficacy results yet, and two preclinical large-gene insertion platforms are different evidence classes. Label them so a headline never collapses into “AI cured aging” or “gene therapy reversed glaucoma.”