Recently I have been fortunate enough to chat to a quite a number of customers about their AAV manufacturing process. For me, the most striking observation is how guarded everyone is about their data. The competitive pressure is palpable. No one is willing to give away any secrets. Everyone is protecting their competitive edge. This makes sense though — there are hundreds if not thousands of AAV-based gene therapies under development. This means that there are often multiple companies developing treatment for an individual disease. The competition is intense, especially when many AAV-based gene therapies are for rare diseases. In some cases, the potential patient population may be so small that the therapy that is released first may be used to treat the vast majority, leaving little room for any alternatives that may come later. With this possibility, the desire to accelerate process development is always in the background.
The upshot of the secrecy is that it is relatively difficult to directly discover what all the different companies are doing. Direct questions often lead to head shaking or blank stares. I have been trying to take a different approach. I will show a slide or some data and ask if that “resonates” or “looks familiar.” Then, perhaps, I can trigger some more discussion. Using this approach it seems clear that most people I have chatted with are heading toward a platform process for AAV purification. This downstream process includes clarification, concentration, affinity chromatography, ion exchange chromatography, further concentration, and filtration. This is a great place to be, as it looks not too dissimilar from a monoclonal antibody (mAb) purification process. And that is probably not a surprise— for many reasons, including the success and platform-ability of mAb purification and the simple fact that many scientists who worked on mAb purification now lead AAV purification process development. This route brings with it some of the same challenges that faced the early developers of mAb therapies, where technology choices that work well at lab scale are found wanting for larger scale GMP manufacture.
Ultracentrifugation – the origins of AAV purification
The traditional method for AAV purification in a laboratory setting is ultracentrifugation using density gradients. This enabled an effective purification in a single step. This is ideal for academia, where a lot of AAV based therapies have originated, and for very early-stage drug discovery and development. The downside being that ultracentrifugation is not readily scalable, making it only suitable for the production of relatively small amounts of AAV. This small-scale production may be sufficient for a rare disease requiring a relatively low annual yield. One such disease is retinal dystrophy where there are perhaps fewer than 100 patients a year and, as injections of AAV can be direct to the eye, the effective dose is far lower than for a more prevalent disease. An example in this instance would be the drug Luxturna. It is estimated that perhaps only ~1x1013 AAV per year is required and this can probably be generated in just 1 L of cell culture media. However, as AAV is targeted to more mainstream applications, such as muscular dystrophy or cancers, the patient population will be much larger, in the tens of thousands per year and the dose will have to be perhaps a million times larger per patient, requiring in the order of millions of liters of upstream production. This is clearly not feasible to purify via an ultracentrifugation method.
To address the lack of scalability associated with ultracentrifugation, a closer look at alternative methods for purification such as direct flow filtration (DFF), tangential flow filtration (TFF) and chromatography open the door to larger scale manufacturing possibilities.