How Can We Achieve The Next Generation Of AAV Vectors?

By Anna Rose Welch, Editorial & Community Director, Advancing RNA

A few months ago, Endpoints News put together a webinar on the past, present, and future of AAV vectors. AAV vectors may be the most commonly used vector today in C&G therapies, but there are a variety of limitations holding AAV-based therapies back from successfully reaching patient populations. If you have an hour, be sure to watch the webinar itself because it really was a wonderfully informative 60 minutes. However, for those of you that want the basic gist, I’ve distilled what I found to be the most insightful themes to emerge from this discussion in the form of a Q&A. (Please note, these are my comments and questions, and mine alone — they are a far cry from the professional ones aired during the webinar.)

Alright, spill the tea on the biggest issues holding back AAV-based therapies today.

On the whole, the AAV gene therapy space currently remains limited to monogenic diseases as opposed to polygenic or non-genetic diseases because of ongoing deliverability limitations.

As speaker Dr. Nicole Paulk, UCSF AAV gene therapy professor, nicely highlighted, we keep finding ourselves stumped by what is turning into a timeless question, “Can we actually deliver the virus to where we want it to go?” Currently, the brain and the eye are the most beloved tissues because you don’t need high doses/high potency or transduction rates to see a clinical impact. On the other hand, the kidney remains an alluring but currently unreachable organ for AAV vectors because the capsids we have today aren’t capable of transducing the kidney — at least not using our current go-to method of IV administration. 

Well, that’s a mood.

It sure is. And that’s not all — most of you are also likely struggling with the second biggest hurdle Paulk called out: manufacturing enough of the vector to even be able to dream of tackling more populous diseases like diabetes. (Though, that certainly isn’t keeping some companies from trying to develop C&G therapies for these chronic diseases.)

Oh, just ask me about my full-empty capsid ratio and how the heck we’re supposed to deal with it.

LOL. Analytical characterization and purification are hard — and talk about losing lots of vector.  

What’s the light at the end of the tunnel here?  

Well, as Dr. Paulk explained, there are a lot of ways being explored to address some AAV-centric problems. Here are a few:

• Paulk’s lab is looking into the concept of “universal payloads.” We currently need to “personalize” each vector for the individual disease. In other words, we can’t use the same vector to treat two different diseases because the genes used and the treatment pathways would be different. So, if we could essentially reuse vectors and their payloads to treat different diseases, that would really bring the space closer to the cheaper, broader manufacturability of mAbs.  
• There’s lots of work being done to innovate on the different components within the genome, including the promotor, the enhancer, and the inverted terminal repeats (ITRs). Paulk’s work in the sequencing of plasmids from academic labs, industry, and those that are sold commercially by suppliers have revealed quite a bit of differentiation in terms of the ITRs’ sequences, lengths, and structures. She has yet to see a plasmid that contains the wild-type AAV sequence, the ITRs of which comprise 145 base pairs and a “flip-and-flop orientation.”
• Last but far from least, engineering new capsids for harder to transduce targets is a big goal in the industry.

Tell me more about the capsid engineering quest. Where are we today?

Well, as luck would have it, this panel featured the CEO of Dyno Therapeutics, a company dedicated to finding the “perfect capsid.”

Well, that sounds like a sexier, more scientific version of Cinderella.

It really does, though.

To be clear, the goal isn’t necessarily to find “the perfect capsid.” More realistically, Dyno’s CEO Eric Kelsic’s goal is to create “enough” (e.g., 10-50 different) capsids that are more specifically targeted or that can transduce organs that remain bucket-list for C&G therapies today — the kidney, lungs, or heart.

As Kelsic explained, a majority of the industry is relying on “natural capsids,” the sequence of which has been sourced from a natural virus. But we didn’t pick these natural capsids because they were the absolute best solution. They were kind of the low hanging fruit; they were what was readily available to meet the industry’s burgeoning needs back in the day.

We’re still pretty far away from the fancier fruit at the top of the tree, though.  

Right — this is another one of those timeless issues in the AAV space. For one, we still struggle to predict how a change to any protein will impact its functional profile. AAV is a particularly enormous protein, and we’re still behind in understanding its biological profile as a stand-alone entity, let alone how it interacts with cells and human cell receptors in the body.

So, a lot of people have been relying on manually mutating the sequence to see if they can fortuitously find an improved variant. Not only is this rare (“finding a needle in a haystack,” as Kelsic called it), but it’s also hard to find an improved variant that won’t muck up any or all of the critical “-ilities:” tolerability, specificity, deliverability, and manufacturability.

Do we have a saving grace here?

New technology is key. We’re living in the age of next-gen DNA synthesis and sequencing, which can more efficiently tell us how changes in sequences will impact function. This technology can also help us determine more clearly what tissues or cell types different sequences of AAV are drawn to.

Automation of the analysis of these experiments could be another potential boon to help researchers garner greater insights into the different biological patterns in the data. In turn this could help determine which AAV sequence will be best for your disease area/targeted cells and, in the long-term, enable the automation of different capsid designs.  

What’s the end-game here for AAV manufacturing if we achieve the “perfect capsid(s)?

Dosage is a huge factor behind the current high cost of goods. So, as Kelsic reasons, if a better capsid improves delivery efficiency, that opens up a whole lot of magical possibilities. For one, you can treat more patients suffering from a larger array of diseases because you can reach new targets. If it’s more efficiently delivered to the target, you can also reduce the dosage volume needed. This would mean you don’t have to manufacture as much vector in the long run, nor would you have to administer as much to the patients, bolstering the safety profile overall.

Anything else?

The natural AAVs many are relying upon have the reputation of being a bit flighty and getting distracted once they’re in the body. As Kelsic explained, while some may end up at their final, desired destination (e.g., the CNS), a good amount of the vector will also end up at the liver or be filtered out through the circulation. So, engineering better capsids can also mean that the vector can be “trained” or “detargeted” from going towards the unwanted locations (i.e., the shiny things). This will mean greater efficacy and fewer off-target effects/safety risks.

So, besides better capsids, what can we expect (or hope for) from AAV in the future?

There are three things:

• Expect more work being done on improving potency of AAV vectors. Beyond working on the basic process development tools, could new methods of administration or engineering capsids to be more potent reduce our reliance on systemic IV administration?
• Ongoing biological learnings will be critical to improving AAV potency/delivery efficiency. For example, there is a 10-fold (or more) difference in per particle infectivity of natural AAVs vs. the gene therapy industry standard rAAVs. Those pesky empty-full ratios plaguing large-scale manufacturing today certainly don’t do a gene therapy any favors. Kelsic asked the perfect question: “Can we learn what the natural viruses are doing that makes them have a nearly perfect particle infectivity ratio?... There’s a lot of value in basic research to try to understand what we can from nature then apply that to what we’re doing therapeutically.”
• “There’s enough sand in the sandbox for everybody to play with.” I loved that quote from Paulk who was emphasizing just how many other viruses beyond adenoviruses, AAVs, and lentiviruses are being explored for future therapies — both in and outside of academia. Naturally, there will be indications that are best treated using each specific type of vector. But looking beyond the vectors — both viral AND nonviral — there are upwards of 8,000 monogenic diseases to pursue, and that’s not even including polygenic diseases.