Gene Therapy · us
Delivering Gene Vectors More Precisely to Muscle: Dyno Announces Animal Data on AI-Designed Capsids
Two new AAV capsids showed better muscle delivery in mice and nonhuman primates, offering research insights into reducing gene therapy doses. But greater vector delivery and lower liver distribution remain a long way from demonstrating patient benefit and safety.
One challenge in gene therapy for muscle diseases is getting vectors carrying therapeutic genes to reach enough muscle while reducing the proportion that goes to other organs. Dyno Therapeutics announced that two adeno-associated virus (AAV) capsids designed with the help of artificial intelligence showed better muscle delivery in experiments in mice and nonhuman primates. These results offer insights for designing lower-dose treatments, but the evidence remains at the animal stage.
Megan Cramer presented these data on September 30 at the 31st World Muscle Society Congress in Hiroshima, Japan, in late-breaking research poster 101LBP. The company website lists the announcement date as October 5, while the same Business Wire release, republished by AOL and RICentral, was published on October 2. These versions originate from the company announcement and cannot be considered independent validation by multiple research teams.
The capsid is the protein shell that encloses AAV's genetic material and is an important factor in determining which cells the vector enters. Dyno said the two candidate capsids can interact with more than one cell-entry receptor, and their design considered muscle delivery, reduced liver distribution, and performance across species together. The specific role of AI here was to help design delivery tools that were then tested in animal experiments; the announcement did not disclose the scale of the model's training data or the complete design process.
The experiments used systemic administration at a dose of 5×10¹² vector genomes per kilogram of body weight (vg/kg), with MyoAAV-4E as the comparator. According to the company's data, both capsids showed approximately twice the vector distribution and transduction of the comparator in the skeletal muscles tested in mice. Tissue distribution describes where a vector reaches, while transduction reflects its ability to deliver genes into cells and produce expression. Both are measures of delivery efficiency, but neither can be directly equated with improvements in muscle strength or disease symptoms.
The nonhuman primate results showed different strengths: one capsid achieved approximately 20 times the skeletal muscle distribution and approximately 40 times the transduction of MyoAAV-4E; the other also showed better muscle tropism, with liver distribution approximately one-tenth that of the comparator. These figures come from different candidate capsids and cannot be combined to suggest that a single vector offers all these advantages simultaneously.
Higher muscle delivery efficiency could allow developers to achieve the required gene expression with fewer vectors. Dyno compared the dose used here with a muscle gene therapy dose of approximately 1×10¹⁴ vg/kg, stating that the former was only one-twentieth of the latter. However, this is a comparison of dose levels, not evidence from human trials demonstrating that doses can be reduced 20-fold. Lower liver distribution is also insufficient to establish a corresponding reduction in liver toxicity or overall treatment risk.
The public announcement did not provide animal numbers, complete statistical analyses, follow-up duration, or detailed toxicology results, nor did it present patient data. The next key question is whether these delivery advantages can be maintained when the capsids carry actual therapeutic genes and translate into lasting functional improvements and acceptable safety. Moving from animals to humans will still require further research and clinical trials to answer these questions.