On Mar. 11, 2026, taking a step towards developing genetic therapies at scale, researchers from Mass General Brigham and colleagues describe a novel, non-toxic approach for inserting gene-sized DNA in a study published in Nature.

Genome editing-based therapies typically aim to treat disease by correcting underlying genetic mutations in patient’s cells. However, most genetic disorders are caused by dozens or even thousands of unique mutations spread across a gene; this diversity results in challenges of scale when it comes to developing customized therapies for many individuals. A more universal strategy could involve the precise insertion of an entire copy of a healthy gene into a specific location in the genome, which could, in principle, provide a single, widely applicable treatment for all patients irrespective of their mutation within that gene.

Most approaches to insert large corrective DNA sequences into the genome utilize recombinase enzymes that rely on double-stranded DNA (dsDNA) molecules carrying the intended DNA cargo, which can trigger strong immune reactions that cause toxicity and ultimately limits how much treatment can be given, especially when therapies are delivered directly into the body. Although viruses can act as vectors to shuttle the DNA cargo into the nucleus, their use is subject to several caveats in terms of cost, safety, and other factors. Developing a virus-free targeted DNA integration method has therefore been a long-standing goal for the genome editing field.

The team recognized that circles composed of only single DNA strands were largely able to evade immune detection. However, this apparent solution resulted in a different problem: Recombinase enzymes evolved to recognize DNA molecules with two strands, making them incompatible with single-stranded DNA. The researchers then investigated ways that natural systems had overcome similar problems, and realized that bacteria and bacteriophages had developed tricks to insert single-stranded DNA into a dsDNA host genome using recombinases.

The team designed a different configuration using a DNA circle composed largely of only a single strand that could remain “stealth” to innate immune sensors, and a short region of two DNA strands to enable compatibility with recombinases. This dsDNA region could be designed to be long enough to allow the recombinase to function, but short enough to avoid immune detection.

The team then demonstrated that their new approach, called INSTALL, could enable non-toxic DNA integration in multiple human cell types. Furthermore, when performing in vivo experiments in mice by delivering the DNA via lipid nanoparticles (LNPs), INSTALL could successfully and safely insert large genetic payloads in the livers of mice, compared to fatal immune reactions observed when conventional dsDNA molecules were used.

The authors are optimistic that continued efforts to improve the DNA cargo, alongside recombinase enzyme engineering, will result in substantial advances in large sequence insertion technologies in the near future.