The American Society of Gene and Cell Therapy (ASGCT) meeting, held last week in Baltimore, MD, was chock full of sessions—both large and small. One of the highlights of the meeting was the George Stamatoyannopoulos Memorial Lecture, a plenary session that included multiple award announcements and several scientific talks.
In the session, Beverly Davidson, PhD, chief scientific strategy officer at the Children’s Hospital of Philadelphia (CHOP) Research Institute, highlighted the progress that her lab has made in identifying adeno-associated viruses (AAVs) for use in brain therapies. Davidson explained that she was thrilled to give the “Stam lecture” not only because he encouraged her to join ASGCT years ago but because Stamatoyannopoulos was a pioneer in considering the power of genetic medicine.
Davidson’s talk titled, “Unlocking the power of AAV for brain gene therapies” highlighted both recessive and dominant diseases, and she described a vignette of each all while focusing on the questions, “how do we get to the region of interest and how do we refine the genetic cargo?”
The work described originated in the late 1990’s over a beer. A conversation between Davidson and John (Jay) Chiorini, PhD, acting deputy scientific director in the adeno-associated virus biology section of the NIH/NIDCR, spurred her to think about the best way to identify the right AAV for the right targets.
The first story that she told had two children with Batten disease at the center—a lysosomal storage disease that has a clinical onset occurring between 18 months and 3 years of age followed by a rapid decline. Davidson’s idea, sparked by a Science paper she read, was to overexpress a gene so that it would offer cross correction. More specifically, she would introduce genetic material to a handful of cells and count on it spreading to neighboring cells, which would allow for widespread expression in the brain.
First, they tested the hypothesis in a mouse model of lysosomal storage disease. They then moved into a dog model of the disease—a dachshund with TPP1 deficiency. When Davidson’s group infused the gene into the CSF of the mutant puppies, the dogs moved in a manner similar to a wild type dog. Davidson concluded that AAV2.TPP1 improves neuronal ceroid lipofuscinosis 2 (NCL2) disease in dogs that lack TPP1.
In addition, this treatment could achieve steady state levels. This was important because Biomarin was developing a protein-based enzyme replacement therapy for the disease. But it only lasted for about 24–48 hours resulting in a treatment where the brain would be exposed to high levels of enzyme followed by none. Davidson asserted that a steady state would be an improvement.
Unfortunately, they saw AAV-induced dorsal root ganglia (DRG) toxicity in nonhuman primates (NHP) and needed to tailor the AAVs. Taking a team approach, and developing AAV libraries on multiple serotypes, they infused them into the brains of nonhuman primates and underwent rounds of enrichment.
A selected vector, AAV-Ep+, transduced the desired targets (both ependyma and parenchyma) throughout the CNS of all tested NHPs, showed low to no off-target transduction, and also low to no transduction in the heart or liver.
When AAV-Ep+ was loaded with TPP1, they saw large amounts of protein being produced (even a low dose resulted in improved survival). AAV-Ep+ worked better than other vectors such as AAV2.TPP1 and it also worked well in NHP.
Davidson’s group will continue to pursue this work and has two other similar programs. One of their main goals is to lower the dose. If they can do that, she concluded, it will improve safety and reduce the cost of goods.
Davidson then pivoted to describe their work on gain of function disorders like Huntington’s disease. The key to treating the disease, she said, is to reach the right regions. More specifically, they need to not only hit the basal ganglia but also connected cortical structures. “We need to target the entire structure,” she said.
The team generated capsid libraries, infused them into a network node, and underwent selections. The final selection went through two monkeys. This work was posted earlier this month as a preprint on bioRxiv titled, “Optimized AAV capsids for diseases of the basal ganglia show robust potency and distribution in adult nonhuman primates.”
One capsid, AAV-DB-3, hit the right sections and at doses 1.5 log lower in 10% of the volume. “This is an incredibly low dose in an NHP,” Davidson asserted, adding that “we have a lot of room to reach a broad therapeutic index.”
The goal is not only to get to the right region, but also the right cells within that region. With this vector, Davidson noted that, “we are hitting the right regions and the right neurons.” One example was the ability of the AAV-DB-3 vector to hit medium spiny neurons—one of the most vulnerable cells in the disease. Davidson’s data showed that AAV-DB-3 could transduce about 50% of medium spiny neurons.
The team will now use their platform to target specific cells in different diseases. To do that, they are doing split pool barcoding or AAV capsid screening at single cell resolution. Starting in the mouse retina, they put in a pool of just under 200 capsids, isolated single cells and single nuclei, and could identify cells from the relevant regions and cell types. Not only that, but AAV-DB-3 came out as the top ranked hit. They will now move forward into the NHP brain.
