This article is adapted from sponsored content originally published on https://www.genengnews.com.
Adeno-associated virus (AAV) has emerged as the most flexible and useful vector for gene therapy. With hundreds of gene therapy trials ongoing, safe and efficient production of AAV is absolutely critical.
In this roundtable discussion, four experts from academia and industry met to discuss challenges in refining and scaling up AAV production pipelines to meet the increasing demand for these indispensable viral vector delivery systems.
We thank Nicole Paulk, PhD, Assistant Professor, Biochemistry and Biophysics at the University of California, San Francisco, Mats Lundgren, PhD, Customer Applications Director at Cytiva, Uppsala, Sweden, Magnus Gustafsson, PhD, Head of Global Business Development at Biovian, Stockholm, Sweden, and Daozhan Yu, PhD, President and CEO at AAVnerGene, Maryland, for participating in this important conversation.
Q1. AAV has emerged as a leading vehicle for gene delivery. What are the opportunities and challenges that come with this viral vector?
Paulk: The biggest opportunities are to expand beyond what we are currently doing, using AAV as a genomic medicine delivery tool in rare and ultra-orphan monogenic indications. I’m beginning to hear more about expanding into polygenic disorders and disorders with more complex etiology. Expanding our treatment world to reach more patients and the challenges that come with it—that’s where the greatest opportunity lies. We still need to learn more about things like safe dosing at high doses when we’re trying to get to tissues throughout the body. Then there’s the ever-present problem—manufacturing: can we make enough to treat every patient on the planet if we were to go after something like diabetes or congestive heart failure?
Lundgren: The whole field is evolving. There are new serotypes being discovered or synthetic variants. There are more opportunities coming up, but manufacturing is difficult, especially at high doses.
Gustafsson: If we look at the approval of the first antibody 40 years ago, we were talking about mouse antibodies made from hybridoma cells. Today, we have transgenic mice expressing human antibodies. And somewhere in between, we have chimeric and humanized antibodies. With time, the mAb titers have gone up, and you can do so much more with these drugs. They have revolutionized the world. AAV has just started to make an impact. Today, transfection is probably the fastest way to make AAV, but it’s not the cheapest, the most robust, or the safest. There will be more technologies making an impact on manufacturing. Technology innovation will be the driver for this, not the drugs themselves.
Yu: We have so many opportunities, not only treating rare diseases with AAV but also many common diseases like diabetes, Alzheimer’s, aging, and even COVID-19. Replacing the gene of interest and the capsid, you can make a new drug. The potential is unlimited. Manufacturing will be a problem for a long time as so many gene therapy companies are emerging that need a lot of AAV. However, some gene therapy companies think that AAV is so safe that you can arbitrarily increase the dose. Last year, the FDA put a hold on several clinical trials, mostly related to manufacturing problems. Many are using very high doses. Several patients have died from high-dose AAV treatments. That poses a great challenge that may delay or even ruin the field. High-dose risk is a major problem. Our focus is not just to increase the production mass but to increase production efficiency.
Q2. Manufacturing large quantities of viral vectors remains a challenge. We hear of long wait times to get AAV. How is this manufacturing gap being addressed? What recent innovations in the AAV production pipeline help reduce production timelines?
Paulk: Right now, we largely address this by thinking in the short term because we’re desperate! We have patients waiting and folks who have programs to start. In the short term, you build more facilities with more bioreactors so that you can meet the demand. But folks are starting to think more in the long term. We’ve put the BAND-AID™ over the wound, but now what do we need to do in the long term? That’s going to require a wholesale reimagining of what it means to make AAV. That will likely require us to stop producing it like it’s a monoclonal antibody (mAb).
Lundgren: Every CMO and CDMO is trying to build up capacity, but that doesn’t mean all companies have the expertise and experience. It is important to select CDMOs that are credible and have the right expertise. We have a good example from Biovian here today. When it comes to the process, there will be tremendous improvements: producer cell lines where you will have constitutive expression, improvements in how you grow the cells, and in how you increase the amount of virus. Everyone is trying to increase the proportion of full capsids. Downstream, there will be improvements in purification—how can you purify the full capsids, formulate the material, stabilize it?
Gustafsson: This massive increase in demand is because there are more clinical trials. And that’s because AAV vectors work. We have higher dosing because companies choose systemic delivery over local administration and study larger patient populations. Different applications require vastly different amounts of vector. We need to have something to adjust for that massive difference. We also have a capacity crunch that is addressed by scaling up or scaling out—more reactors or bigger reactors. There is a financial crunch because someone has to pay for that. Massive bioreactors are just not feasible. You have a technology crunch. You need higher titer and more efficient vectors that are more specific for different tissues. And you have a time crunch. Cytiva has been instrumental for years in equipment for both mAbs and viral vectors. In technology development, academics are instrumental in giving us better tools.
Yu: A lot of technology is being developed. Take the AAV manufacturing developed by my partner Dr. Wang. It gives very high yield of AAV and can be easily scaled up. We have borrowed a lot of technology from the antibody manufacturer. However, AAV manufacturing is more complex than small- and large-molecule manufacturing. As former FDA Commissioner Scott Gottlieb said, in AAV gene therapy we will focus more on product manufacture and quality itself. That is different from traditional drug development. You can always run more bioreactors and increase the volume, but scalability is not that easy. Many companies don’t have adequate expertise. We are still using plasmid transfection protocols where cells tend to aggregate. Some companies are developing new transfection reagents that work well.
Q3. What are the challenges to a scalable start-to-finish workflow?
Lundgren: Transient transfection is still very good, and it gives you high-quality material, but the problem is scaling up. Other systems are easier to scale up. The adenovirus system, for example, is brilliant when you can work with suspension cells. You can infect them with the adenovirus, and the adenovirus can then be removed in the downstream process. Producer cell lines will be coming, although it’s not as easy as it sounds. Since a lot of innovation is coming from academia, the technologies are inherently small scale. And some steps can be difficult to scale. For example, ultracentrifugation is difficult to scale up because you need to have enormous centrifuges, and it’s difficult to clean them. So, the trend now is to move away from ultracentrifugation and more into filtration and chromatography technologies. And there’s another reason for that—the safety aspects—to ensure that you can have a closed system, so you don’t have any cross-contamination. The way forward is closed and scalable systems and trying to increase titers, the purity, and the potency of the material.
Gustafsson: We face this problem at Biovian. We have clients who start small scale and then want to go up to larger scale. As a CMO, we do what our clients tell us to do, but if we have a chance to develop a process, we try to go for suspension cultivation because it’s more scalable. We have selected Cytiva to be our provider. We have their scalable Xcellerex™ system, and we use their ÄKTA™ system for purification, which is also scalable. If you go for triple transfection, then you also must scale up plasmid manufacture. That becomes the bottleneck. You need huge quantities of plasmids to make virus. You could scale up with stable cell lines and use viral vectors to transduct your AAV. Purification is very important. It would be nice to have a disposable ultracentrifugation system that is serotype nonspecific. But it doesn’t exist yet. So, we go for chromatography, which is scalable.
Q4. Cost remains a pain point, particularly from the industrial point of view. How can AAV production be made more cost effective while maintaining rigorous safety standards?
Paulk: One of the main reasons AAV costs so much is because we dose so much. The best way to reduce price in the short term is to just reduce doses. To enable that, you’re going to need innovations around how to make these more potent so that you can give a lower dose. Beyond that, is the need to make an entirely new tissue culture media that costs 1/10,000th of what it costs now. This is the realm of much larger innovations that might take many years. In the short term, we can do things around finding capsids that produce slightly higher titers, finding methods to recover and retain more so as not to lose so much during the purification process, finding ways to have more full capsids from the beginning. In the long term, we will need a wholesale reimagining around how you can make AAV in a cell-free system, without plasmids, bioreactors, cells, or tissue culture media. Can you make this in a veritable test tube?
Lundgren: The biology will have a great impact if you can find something that is more potent. Then the costs come down. But in terms of processing, we can look at process intensification. For example, you can grow cells with perfusion systems, so you have more cells in the reactor and higher productivity per volume. We can try to develop platform-type purification processes that will not need to be adapted every time you switch to another virus. Supplies—how can we make cheaper media, better bioreactors? Trying to find scalable platform processes that are easy to implement while maintaining safety and efficacy is important. In the long term, you could go for cell-free production. Even though they are very small, these viruses are still much more complex than mAbs.
Gustafsson: We need to look at the platform with a holistic approach. There will not be one answer to fix everything. You have probably five things to address: the vector itself, the plasmids, cell lines, purification, and manufacturing systems. Today, we use plasmids as they are, but you can use nanoplasmids, minicircles, or rolling circle amplification. The plasmid has to be tailor-made both for the E. coli where it’s expressed and for the cell where it’s used. Or you can skip the plasmids and use transduction by a virus, such as baculovirus. You can use seed stocks. When it comes to cell lines, there are some 160 different HEK293 cell lines. You have to select one that is better than others. You have other cell lines too: CAP™ cell lines, PER.C6®, HeLa. There are stable cell lines and cell lines that claim they can export or secrete AAV out of the cell. If that technology is developed, you can have continuous production of AAV. We’re looking into suspension and adherent manufacturing systems, but a system for a continuous process that does not lyse cells is most important. Better, more, and serotype-nonspecific ligands or ion-exchange chromatography that can fish out full from empty capsids would be wonderful. Analytical tools to look at your process development and analyze the outcome are important. High-throughput electron microscopy with AI to make conclusions from your analysis would be wonderful.
Paulk: We assume that the list price for an AAV is based off production and process development. But there are also other costs. As an AAV gene therapy company, you have staff and clinical trial costs that get baked into what you choose to list. To reduce your list price, you can bring those costs down. All the work you put in up front matters. If you end up getting a clinical or preclinical hold—that’s time and cost you have to recoup. The more time you invest up front, either while you’re still in academia or right as you’re spinning it out, into determining the backbone, promoter, and inverted terminal repeats (ITRs) you want to use, the better.
Q5. How do you see academia and industry continuing to develop their communication and collaboration with regard to AAV production?
Paulk: We’re seeing more industry money coming into academia. But the danger is when you get money from industry there are often strings attached, where you either implicitly or explicitly feel the need to produce for that company with that money. To ensure that academics still have the freedom to research in any way they wish, to publish the results whether they’re positive or negative for your company, and to present that data—to make sure that we remove those conflicts of interest from it so that we don’t poison our own drinking water supply—is important.
Lundgren: It’s not good if you get money from us and we ask you to do contract research. Collaborations around scientific discussions about your pain points, what you foresee, could be good. Funding from industry is dangerous if you rely too much on it. We have collaborations with research institutes on manufacturing technologies—that works because we are not funding them but are working together. We can perhaps do large-scale productions for them to show their molecule works. As an industry, we can never solve biology questions. Analytics is an area where industry and academia could collaborate, but it is important to have the free academic spirit for innovations.
Gustafsson: It would be wonderful to have an academic/commercial initiative with open-source viral vector platform process development. Something like Linux, developed as an open-source student project. We have seen such collaborations in sequencing of the human genome. We are all fighting initial costs. It would be wonderful to have something to build on that is available for everyone.