#apaperaday: Semirational bioengineering of AAV vectors with increased potency and specificity for systemic gene therapy of muscle disorders

In today’s #apaperaday, Prof. Aartsma-Rus reads and comments on the paper titled: Semirational bioengineering of AAV vectors with increased potency and specificity for systemic gene therapy of muscle disorders

Paper on engineering AAV for improved muscle delivery by El Andari from Grimmlab in science advances. Doi: 10.1126/sciadv.abn4704

Adeno-associated viral vectors (AAV) can deliver genetic material to a variety of tissues. Improving tissue specificity is warranted as it would allow use of lower doses and pose less safety concerns for off target tissues (mostly liver).

Authors use a barcode system that allows injecting a mix of different modified AAVs & then studying which type goes to which tissue using a barcode system. This was used earlier by the group to identify AAV Myo, an AAV with improved homing to muscle and reduced uptake in liver.

Here authors further bioengineered AAVs to try and improve their AAV Myo. First they made a library of AAVs containing mixes of AAV 1, 6, 8, 9, or also AAVP01. They selected the 32 capsids that had best muscle delivery, compared to low liver delivery.

Adding also control capsids, they injected a mix of 36 capsids in a mouse and studied delivery to muscle and other tissues in more detail. From this they selected the 2 best capsids, which however, showed lower efficiency to deliver to muscle than the original AAV Myo.

Authors identified a specific peptide in AAV Myo likely responsible for its improved muscle uptake and added that peptide to the other two capsids, generating AAV Myo2 and AAV Myo3. Indeed this resulted in increased muscle delivery compared to AAV Myo

It is not only about delivery to muscle and not liver. Authors also studied other aspects, showing the 2 AAV Myos could be produced and purified with efficiencies similar to AAV9. They also performed 3D modelling of the new capsids to study their binding to receptors.

They showed that the AAV Myos were recognized by antibodies recognizing AAV9. In other words, individuals with preexisting immunity to AAV9 likely are not eligible to be treated with AAV Myo2 or AAV Myo3. Finally authors tested the 2 new AAV capsids in animal models

They delivered transgenes to an Mtm1 knockout mouse (for X-linked myotubular myopathy), showing increased survival, improved strength and weight and improved muscle pathology. In this model AAV Myo3 was more efficient to deliver the transgene than AAV9.

They also showed the new capsids could deliver micro-dystrophin to the mdx mouse. Here the delivery and efficiency was comparable to AAV9. It is important that authors confirm the capsids can also deliver in a disease model as muscle pathology can improve or reduce uptake.

Authors discuss that using mice to identify optimal capsids is a risk, as what works in mouse will not necessarily work or be optimal in human muscle. However, they argue that the same holds for pig and monkey and ideally capsids are selected that work across species.

The paper contains a LOT of work and is a very complete story. It will be interesting to see how this develops further. The current challenge with AAV gene therapy for muscle is that the doses needed to get good delivery to muscle are toxic to the liver in some patients.

If the dose could be reduced by a factor of 5 or 10 with the same efficiency to muscle delivery that would make this approach a lot more tolerable – plus make manufacturing a lot easier as 5-10 times more patients could be treated with the same batch.

Looking forward to further work from this group and others who are tinkering with AAV to try and improve muscle delivery.