We recently sat down with Paul Wille, Director of Product Development at Abeona Therapeutics. He reviews his take on AAV as a modality and the field as a whole.

Paul, we’re glad to have you today. Can you tell us about your background and how you got into gene therapy?

I grew up in the Pacific Northwest and went to undergrad at Whitman College, a small school in Eastern Washington. Originally thinking I’d be a doctor, I slowly transitioned more and more to the research side of things and got a PhD, rather than an MD. After getting my PhD in virology at OHSU, I made my way to Case Western to work on HIV research. I worked in a couple of great labs as a postdoc and the common thread through all these years prior to transitioning to industry was really virology and viral vectors. I always had in the back of my mind the potential for how we can use different viruses to positively impact human health. That's where my interest had been, and I got lucky finding a company working on viral vectors, Abeona, which was just 20-something people when I joined. I've been here for the last six years.

I support a lot of the preclinical work as the director of product development. Our development team handles everything from concept to the clinic, and whatever aids in that transition. We wear many hats.

How does your group think about prioritizing indications to go after?

We’ve spent the last three years developing three ophthalmology indications, and prioritization hinges on balancing how to get the most out of the resources available on the preclinical side, finding an indication with a favorable competitive landscape, and having the potential for treating enough people to be commercially sustainable. So, although our current programs are in the rare disease space, we've had instances where projects were too rare to recoup research and development costs and advance to commercialization. I have been involved in those conversations where the scientific rationale might be compelling, but is the opportunity able to sustain a business and thus warrants pushing forward to the next set of experiments that cost X amount of money?

The field started with these monogenic disorders. Do you see an emphasis on ophthalmic indications, and if so, why is that?

The obvious limitation with AAV from the start has been effective delivery targets: where can you target delivery, and what diseases can you focus on as a result? AAV accumulates in the liver, so that’s an obvious target, where you could use an AAV for delivering a gene encoding an enzyme to take the place of an enzyme replacement therapy. And then you had ophthalmology. Because administration is local to the eye, you don't need a lot of vector to go in and affect something like a blinding disease. The biggest impact is going to be delivery of a single, small to medium sized gene due to the limited genome size that fits into an AAV capsid. Obviously, there's been a lot more creativity since then in targeting tissues and organs, but really, because of the coding capacity and the challenges of delivery, it's mostly stayed within that monogenic disease category for the time being.

In these past six plus years, you've lived through the pandemic-fueled boom. Could you take us back to the year after the pandemic started, when things kind of normalized a bit, and AAV really took off and give us a sense of what that was like?

Going into the pandemic, we were at the opposite end of the boom. We had really focused our resources on our lead program for epidermolysis bullosa, now called pz-cel, as well as our two AAV-based neurodegenerative clinical programs, for Sanfilippo syndromes MPS3A and MPS3B. We returned the IND for MPS3B to Nationwide Children's Hospital, and the MPS3A program was licensed to Ultragenyx, who is pushing that program forward. Both MPS programs required systemic delivery and consequently large amounts of virus for a single treatment. On the preclinical side, we were transitioning into ophthalmology, which gave us the ability to tackle diseases that are on the more common end of the rare disease spectrum and did not have the challenges associated with systemic delivery. You can use a much smaller amount of virus to treat in the eye versus systemically.

As we moved into mid-2020 and 2021, we were really getting into our proof-of-concept experiments on the preclinical side. I think because much of the fervor that was happening at the time was focused on new companies, we weren't necessarily going to capture a lot of that attention because we were already established. We weren’t coming out with this “new” flashy thing. I think a lot of what happened at that time was people realizing we can use these vectors for editing programs and tackle these more common diseases. So, the joke always is about somebody taking a good idea and suddenly having a hundred million dollars to spend and not actually knowing how to get from concept to clinic. So, I think in some ways we didn’t benefit as much as other similar companies from that goldrush because we already had our heads down working.

It sounds like there wasn't ultimately much difference in how you thought about programs and prioritized them during that boom time, at least compared to now.

I would say that's fair. I think the focus has really been on trying to move forward some of these AAV programs in a way that is not too capital-intensive since we're confident in the capsid that we're working with. And the fact that we focus on eye indications, as I said before, the dose you need is low, relatively speaking, so we don't have to squeeze every tiny bit of titer out of our production. This is the problem for AAV tackling broader or systemic indications where the dose is so much higher. We're just not concerned with that. We're using 1,000-fold or 10,000-fold less vector than some systemic programs.

On that point about manufacturing, how did you make the initial decision to build it out internally as opposed to outsourcing it?

I think we're now in a unique position here in terms of how this all worked out over time. We built a relatively small facility here in Cleveland, and initially, we were using it for both pz-cel and for the AAV programs. We were making 200-liter bioreactor scale batches of AAV for the Sanfilippo programs. And as we got to the point where we needed to move to commercial readiness for pz-cel, it was clear, based on how we built our facility, that it was going to need to be dedicated to a single product. That facility is now entirely dedicated to pz-cel, which has required us to outsource manufacturing for our AAV-based ophthalmology programs. So, we have all the know-how, we have the in-house experience with up to 200-liter bioreactor scale for AAV, and with that sort of experience it will obviously be very tempting to bring those programs back in-house at some point, but I think that all depends on the economics of it as we move down the road. There are pluses and minuses on both sides, so I think that's not a simple decision - you really have to look into all tradeoffs.

Speaking more into the hurdles of manufacturing AAV, this is something that's been in the spotlight. How did you view the difficulty of setting that production space up, at least initially?

It's not necessarily terribly complicated in terms of the space. For a suspension process, you need some bioreactors, and depending on what scale you need, you might need a couple of different sizes of reactors to scale up. And then it's all about whatever downstream purification methods you've chosen and what you physically need there.

I think a lot of the kinks in terms of workflow have been worked out. The challenges I think are now around scale-up to the enormous quantities that people are talking about, and that presents a lot of technical challenges. You don't want to have some small mistake screw up a 5,000-liter bioreactor. That would be millions of dollars in reagents for a single run. Even if one sources the most basic raw materials, those runs are going to be incredibly expensive. So, there’s obviously many layers of controls that go into even a single run.

Most everyone working in suspension culture now using a triple transfection or a double transfection of plasmids is working on some version of HEK293 suspension cells, and we are as well. And for us, we did the adaptation from an adherent line into a suspension line. Successful adaptation comes down to stepwise monitoring of titers and other quality attributes to see that you're actually doing the adaptation in a way that maintains good production. We went to a serum-free system as well, which is not uncommon since that tends to be the state of the art and fits well within regulatory guidelines. We were able to accomplish the transition using widely available media. I think that some of the challenges in the transition for others are probably occurring at that stage where they're looking to produce every tiny bit of vector they can get out of the cells.

You are actually going serum-free today. Do you think it's just that issue of titer and maximizing that, or is it spending the time to figure out a new media?

I think you have to accept a potential hit in titer in order to avoid the regulatory challenges in keeping a serum-based process. I think that if you can have a completely animal product-free process from start to finish, then that is your lowest risk strategy moving forward, and that you're just going to have to do the optimization work. It's better to start from there and do optimization than to deal with the regulatory challenges and the strong desire of the agencies to move to animal-free.

Turning to downstream purification and QC, how do you view the challenge of discriminating empty vs. full capsids?

Historically, distinguishing between empty and full capsids has been challenging. It's not just about obtaining pure fills; that's a chemistry problem we've largely solved. The real challenge lies in optimizing the chemistry to achieve the necessary separations for purifying full capsids, especially when using chromatography. It boils down to the effectiveness of the upstream process in producing full capsids compared to empty or partial ones, and the subsequent downstream impact on yields. Initially, many were reluctant to take a substantial hit in yield. This significantly affects production estimations and necessitates scaling up to meet expected yields. As for the clinical implications of impurities, it varies for each program. Liver toxicities, for instance, can be directly linked to impurity levels, with higher impurities correlating with worse toxicity. Hence, ensuring a pure product becomes imperative, especially for systemic administrations. However, scaling up to produce a purer product can significantly inflate production costs, making it harder to justify financially. This poses a considerable challenge, particularly when trying to balance purity requirements with production costs.

How do groups determine acceptable purity levels for administration?

Determining acceptable purity levels is not straightforward and often requires empirical data. Preclinical studies comparing different purity levels are seldom greenlit unless there's a compelling reason to do so. Thus, decisions on purity levels often rely on a balance between achieving maximum purity without excessively inflating production costs. In clinical trials, purity requirements may vary widely until regulatory bodies establish minimum purity standards, which might be challenging given the varied toxicities depending on the administration route and indication. For instance, in ophthalmology, where we're exploring para-retinal delivery for certain indications, the delivery method can influence toxicity profiles, making it crucial to tailor purity standards accordingly.

We use a much lower dose when we use para-retinal administration compared to intravitreal, about a tenfold lower dose. And in preclinical experiments that we've done, this has been associated with significantly less toxicity. One, because it's a smaller dose, but two, because we're just pushing it right down onto the retina. It's not coming out of the eye.

Right now, there’s a bit of a crossroads in the field - what do you see as AAV's advantage over mRNA-based approaches?

I think that the advantage, at this point, is having real clinical data as a gene therapy. Delivering something that is going to have benefit depending on the target cells, AAV has shown that in the eye and brain, for example. So, if you want to do that, I think AAV is going to continue to have a huge advantage there.

I don’t think that we know the clearest advantage for mRNA yet. The most obvious to me at this point is delivery of editing machinery because if you're just trying to do a quick hit and you have done the work to show that you're getting to target cells without significant toxicities, that's going to be a pretty significant benefit over AAV, which is going to stay in those cells depending on what cells you're targeting. In terms of toxicity, depending on the cell type it could be better with lipid nanoparticles versus AAV. The redosability seen with LNPs should also be a pretty significant advantage. But if you're just trying to deliver a gene that's stably expressed in a cell that's non-replicating, I think AAV is a better bet today. The mRNA space is definitely an exciting area at the moment, though.

What does AAV need to push forward to pick up speed again?

I won't say AI, which is what everyone says; instead I'll be more specific and say machine learning. I think the potential to more effectively use optimization resources through a combination of automation and machine learning is something that is going to massively benefit any vector technology. The problem is that with the natural capsids we have run into too many limitations in terms of delivery. There will certainly be lots of things that we can use them for, but if we want to continue to refine delivery, then we're going to have to do what some other companies are doing in terms of their machine learning development of AAV.

When do we see the first successful commercial launch?

I think we have seen some successful commercial launches; patients with SMA are getting Zolgensma and some LCA patients are receiving Luxturna, for example. Success is relative in the rare disease space. I think to add to the question: how do we maximize the benefit of a gene therapy commercial launch for all stakeholders? I think we’ll see several more approvals this year, but I think that what we really need to see is the combination of a genuine unmet need and a cleaner toxicity profile in a more common disease. I think certain past launches have been hampered by efficacy that just isn’t quite where you would want to see it. There’s enough skepticism such that not everyone affected is running out and getting it immediately. I think as you move into spaces where there is tremendous unmet need and you have products that have very clear-cut efficacy, that's when you're going to see that massive uptake immediately. Products like that will have a strong launch I’d think. So long as the previously mentioned manufacturing needs are met.

Can AAV move more broadly beyond monogenic disorders, for example, delivering CRISPR factors?

Yeah, I think that there's no doubt that it can. I think that there are two big challenges. The first is when people start talking about more common diseases. I think those of us who have seen the production limitations at this point are a bit skeptical for now. You're not going to go out and treat thousands or tens of thousands, to certainly millions of patients with an AAV therapy unless it is going to be very, very low dose right now. There's not a lot of diseases where that's true. Again, maybe some ocular indications would be possible. But anything that's systemic, absolutely not. That manufacturing challenge is still very significant at this point.

The second challenge is whether new technologies that are allowing us to move beyond gene augmentation like CRISPR/Cas or base and prime editing can be adapted to fit into a single AAV, and whether AAV makes sense as a delivery vehicle depending on the target and application. As we get better at manufacturing very large quantities of AAV there might be some appeal there, especially if we see more limitations or toxicities from LNPs that heretofore have not been uncovered.

What are you excited most about within AAV or across the field?

I think what I'm looking forward to the most, especially in the next couple of years, is the approvals that are coming along at a much greater rate and how they will demonstrate that a lot of the limitations that we've seen with AAV therapies have been overcome and that there are going to be a great deal of clinical and commercial applications for AAV. I think the most exciting thing is just seeing these currently developed programs get past that hump and get an approval and have commercial success.

Yes, there’s a lot of patients out there who can ultimately benefit from these therapies. Thanks, Paul for joining us today.

Thank you.