New Tools Sculpt Tailored AAVs

More than a hundred clinical trials have used adeno-associated viruses (AAVs) to deliver gene therapies. At least eight viral-vector gene therapies have already garnered FDA approval, including ZOLGENSMA, a one-time gene therapy for spinal muscular atrophy that uses AAVs to deliver a new copy of a gene to infected cells.¹

Most clinical trials, though, ultimately fail and will never lead to an FDA-approved drug. That failure isn’t always due to problems with the gene therapy itself—often, it's a byproduct of the virus infecting the wrong tissues, or triggering an immune response due to the high doses required. In recent months, however, several studies have reported revamped tools to produce AAVs with greater tissue specificity, larger volumes, or reduced immune responses.

Revamped history

Discovered in 1965, AAVs are so often used in gene therapies because they replicate poorly—or not at all—in human cells.² These single-stranded DNA viruses, though, can only package about 4,700 bases of DNA, which limits the types of genetic therapies they can deliver.³

A dozen naturally occurring AAV variants have been discovered thus far, and each infects different types of cells. A common variant, called AAV9, is often used for gene therapies because it has high transduction efficiencies, but unfortunately has wide distribution throughout the body, ultimately ending up in the liver, brain, and many other places.³ These off-target effects necessitate high doses for gene therapies, though, and at least three recent clinical trials injected 200 trillion viruses per kilogram; the resulting toxicity caused the deaths of several children.⁴

To design AAVs with a higher specificity toward specific organs, a recent study created thousands of viral variants at once by randomly shuffling virus-encoding genes. These random variants were injected into mice and, a week later, the heart, lungs, and other organs were extracted. PCR was used to retrieve highly enriched viruses from each tissue.⁵ This process was repeated three times to gradually select those viruses that traveled to specific tissues. Two variants, called AAVS1 and AAVS10, were highly specific to muscle, but still generated an immune response commensurate to that induced by AAV9.

While this method can identify tissue-specific AAVs, it isn’t a panacea for making variants that can carry larger genetic payloads. The small volumes of many AAVs have already proven problematic in gene therapies for Duchenne muscular dystrophy, hemophilia A, and Stargardt disease, all of which require large genes to be delivered to affected cells.

Bring the payload

Strategies to deliver larger lengths of DNA in AAVs generally rely on two strategies: Deliver different parts of the gene therapy in separate viruses, or engineer capsids to be physically larger.

Several groups, in recent years, have shown that genes with attached intein sequences—protein segments that can form peptide bonds with complementary inteins—can be used to express large proteins from AAVs.⁶ Although this strategy has successfully been used to express a 290 kDa protein in mice, efficiencies are low because viruses carrying each part of the final-length protein must infect the same cell, and then the inteins must connect.⁷

A better strategy is to make physically larger AAVs. Wild-type variants are small icosahedrals measuring 25 nanometers in diameter, but recent research shows that the number of subunits used to build AAVs can be expanded to create AAVs with diameters up to 70 nanometers.⁸ By modifying the interactions between subunits and engineering cells to build the new capsids, this method worked for multiple AAV serotypes, including AAV9, AAV2, and AAV5.

These approaches, again, say nothing of immune responses. When viruses are injected into a mouse, monkey, or human, antibodies bind to the capsid and trigger a response, thus making it more difficult for repeated doses and long-term efficacy. Reducing B-cell-mediated immunity is key to improving AAV re-administration, and depleting B cells prior to redelivering an AAV enabled “systemic AAV re-administration in mice,” according to a recent study, and “the presence of any anti-AAV IgM or IgG prevents AAV transduction.”⁹ Rational engineering of AAVs could possibly be used to sculpt variants that avoid IgG or IgM antibodies.

Brain and lungs

Recent efforts have also focused on crafting AAV variants with higher specificity to specific parts of the body, using similar in vivo screening approaches. Two recent papers report AAVs that are ultra-specific for the brain or lung epithelium. AAV9 naturally targets the central or peripheral nervous system (PNS), albeit with low efficiencies. There are several AAVs available to target the CNS in mice, but they are limited to specific animal strains because of variations in their blood-brain barriers.¹⁰

By inserting random barcodes between positions 588 and 589 of AAV9 capsids, and then injecting the resulting library of AAV variants into adult mice and running PCR on organs after two weeks, a recent study found more than 6,000 AAVs that “showed a bias toward one or more of the PNS tissues,” according to the study. Two variants ended up in parts of the brain that are particularly hard to reach for AAV9, including the cortex, thalamus, cerebellum, and brainstem. And one variant had a 5.5-fold increase in neuronal transduction compared to AAV9 in marmoset monkeys.¹¹

To design AAVs that travel specifically to the lungs, a recent study randomly modified amino acid residues exposed on the viral surface—sites 452 to 458—and used a similar in vivo screening approach. One of the engineered variants had an 18-fold improvement in lung transduction in mice compared to normal AAV9.¹²

References

1. U.S. Food & Drug Administration. (2022, September 19). Approved Cellular and Gene Therapy Products. FDA. Retrieved November 16, 2022, from https://www.fda.gov/vaccines-blood-biologics/cellular-gene-therapy-products/approved-cellular-and-gene-therapy-products

2. Hastie, E., & Samulski, R. J. (n.d.). Adeno-Associated Virus at 50: A Golden Anniversary of Discovery, Research, and Gene Therapy Success—A Personal Perspective. Human Gene Therapy, 26(5), 257-265. https://doi.org/10.1089%2Fhum.2015.025

3. Adachi, K. (2015). Capacity of Viral Genome Packaging and Internal Volumes of AAV Viral Particles. Molecular Therapy, 23(Supp 1), S17. https://doi.org/10.1016/S1525-0016(16)33642-5

4. Wilson, J. M., & Flotte, T. R. (2020). Moving Forward After Two Deaths in a Gene Therapy Trial of Myotubular Myopathy. Human Gene Therapy, 31(13-14). https://doi.org/10.1089/hum.2020.182

5. Andari, J. E. (n.d.). Semirational bioengineering of AAV vectors with increased potency and specificity for systemic gene therapy of muscle disorders. Science Advances, 8(38). https://doi.org/10.1126/sciadv.abn4704

6. Villiger, L. (2018). Treatment of a metabolic liver disease by in vivo genome base editing in adult mice. Nature Medicine, 24(10), 1519-1525. https://doi.org/10.1038/s41591-018-0209-1

7. Tornabene, P. (2019). Intein-mediated protein trans-splicing expands adeno-associated virus transfer capacity in the retina. Science Translational Medicine, 11(492). https://doi.org/10.1126/scitranslmed.aav4523

8. Ding, X., & Gradinaru, V. (2020). Structure-Guided Rational Design of Adeno-Associated Viral Capsids with Expanded Sizes. Molecular Therapy. 

9. Chen, M. (2022). Immune profiling of adeno-associated virus response identifies B cell-specific targets that enable vector re-administration in mice. Gene Therapy. https://doi.org/10.1038/s41434-022-00371-0

10. Deverman, B. E. (2016). Cre-dependent selection yields AAV variants for widespread gene transfer to the adult brain. Nature Biotechnology, 34, 204-209. https://doi.org/10.1038/nbt.3440

11. Chen, X. (2022). Engineered AAVs for non-invasive gene delivery to rodent and non-human primate nervous systems. Neuron, 110(14), 2242-2257. https://doi.org/10.1016/j.neuron.2022.05.003

12. Goertsen, D. (2022). Targeting the lung epithelium after intravenous delivery by directed evolution of underexplored sites on the AAV capsid. Methods & Clinical Development, 331-342. https://doi.org/10.1016/j.omtm.2022.07.010