Delivery is the largest challenge in the ongoing development of gene therapy: how to put enough of a vector into the target tissues without sending too much of it elsewhere in the body, particularly the liver, which is where much of every injected substance tends to end up. This is a big issue for systemic administration of gene therapies intended to affect much of the body, given severe side-effect and deaths that have occurred in human trials at high doses of viral vectors. A greater ability to target specific tissues means that a lower dose can be used, and thus off-target effects produced by the vector itself are minimized. The AAV approach noted here is an order of magnitude better than the standard serotypes when it comes to preferentially targeting muscle tissue. That seems to me a big deal, enough to enable systemic delivery of AAV-based therapies at doses far below those at which toxicity and deaths have occurred in human trials.
Recombinant adeno-associated viruses (rAAVs) are the most commonly used vehicles for in vivo gene replacement therapy and gene editing in preclinical and clinical studies, yet selective transduction of specific tissues after systemic delivery remains a challenge. Recombinant AAVs generated using naturally occurring capsids are predominantly sequestered in the liver after systemic injection. This sequestration limits the efficiency of transduction in other organs and poses a particular challenge for gene delivery to skeletal muscle. Because muscle comprises up to 40% of total body mass, achieving therapeutic thresholds in muscle with natural capsid variants requires extremely high virus doses (~2E+14 vg/kg), which creates a formidable hurdle for vector manufacturing and can result in therapy-limiting toxicity, as observed in some recent clinical trials.
Here, we developed the DELIVER (directed evolution of AAV capsids leveraging in vivo expression of transgene RNA) strategy to combine diverse capsid library generation with stringent transcript-based in vivo selection and to enable directed evolution followed by identification of functional capsid variants in any tissue of interest and any animal model. We apply DELIVER to develop muscle-tropic capsids (MyoAAV) in mice and non-human primates (NHPs) and compare our results to AAV9 and AAVrh74, both of which are naturally occurring AAV capsids currently used in gene replacement trials for Duchenne muscular dystrophy (DMD).
Quantification of in different skeletal muscles of male and female C57BL/6J mice revealed 10 to 29 times higher transgene expression in muscles of MyoAAV-injected compared to AAV9-injected mice. Expression was 6.3 times higher in the heart and 2.8 times lower in the liver of MyoAAV injected animals. Notably, improved transduction efficiency by MyoAAV was restricted to striated muscle tissues, and this engineered capsid variant transduced the lung, kidney, spleen, and brain of injected animals with similar or lower efficiency compared to AAV9. We anticipate that adoption of DELIVER to additional tissue and organ systems will have a far-reaching impact in accelerating the development and translation of gene therapy and other genomic medicine approaches for a variety of human diseases.