A UT MD Anderson team reports skeletal muscle targeting and repeat dose safety in non human primates, building the structural case for a redosable alternative to single shot DMD gene therapy.
A non-viral delivery vehicle built from engineered extracellular vesicles has carried the full-length Duchenne muscular dystrophy gene into skeletal muscle, restored dystrophin expression in a mouse model of the disease, and cleared a basic safety bar in non-human primates, according to a Nature Biomedical Engineering paper published today from a team led by Betty Kim, a neurosurgeon and researcher at the University of Texas MD Anderson Cancer Center. The work is preclinical. It also lands at a moment when the field's viral-vector alternatives are running into the structural limits that have defined DMD drug development for a decade.
The mechanistic pitch is the full-length-versus-truncated-dystrophin distinction. DMD is caused by mutations in the largest known human gene, a coding sequence too large for adeno-associated virus (AAV) vectors, the delivery vehicle behind the only FDA-approved gene therapies for the disease. AAV's packaging ceiling forces developers to ship a shortened micro-dystrophin or skip individual exons, leaving patients with a partial protein. The Kim lab's platform instead packages the complete dystrophin mRNA inside skeletal-muscle-targeted extracellular vesicles, which the team calls DMD t-EVs, and releases the payload after a systemic intravenous injection, per the paper.
In the mdx mouse model, the vesicles restored wild-type dystrophin translation and, in the investigators' own framing as carried by GEN's coverage of the study, "dramatically" improved muscle strength, endurance, and function. The non-human primate arm is the genuinely new contribution: the same systemic dosing produced skeletal-muscle targeting without the immune reactions and toxicities that have complicated AAV-based DMD gene therapies, according to the Nature Biomedical Engineering report, and the effect held up across repeated doses.
That repeat-dose profile is the structural pitch against AAV. Viral gene therapies for DMD are effectively single-shot interventions: redosing is limited by neutralizing antibodies against the capsid, and the first generation of approved products has already accumulated the kind of safety record, including fatal adverse events, that has put pressure on the modality. GEN notes that these limitations "have resulted in the removal of at least one Food and Drug Administration-approved gene therapy from the market," without naming the product. The most likely referent is Sarepta's Elevidys (delandistrogene moxeparvovec), which saw major label and access actions in 2025, though the source does not explicitly draw that link and the specific product should be confirmed independently before being treated as established fact.
The Kim team is explicit about what it has not yet shown. Cardiac-muscle targeting remains unresolved, a critical gap because cardiac involvement drives most DMD mortality. Full clinical toxicology, dose-response in larger species, and the manufacturing realities of scaling engineered vesicles for chronic, repeat dosing are also open. Kim frames the vesicles as a "platform- and disease-agnostic" protein-restoration modality that could extend to cancer, neurodegeneration, and fibrosis, in remarks carried by GEN. That is an investigator's reach statement rather than a demonstrated result, and it should be read as such.
What to watch next: an IND-enabling toxicology package that includes cardiac tissue, comparator data against the current AAV micro-dystrophin standard, and any sign of an industry partner willing to bankroll the scale-up. The mechanistic case is the strongest thing this paper offers. Whether engineered vesicles can carry it from a mouse and a small cohort of primates into a clinic is the question the DMD field has been waiting on for years.