Continuous upstream processing (perfusion) is not a new concept in the bioprocessing industry. Genzyme, Bayer, Centocor, and other companies have been implementing perfusion processes for many years. However, interest is now growing for extending this concept to downstream operations to create fully integrated continuous processing. During the past year, Genzyme has presented on and published about its advancement toward the development of an integrated continuous system (1). The company has completed proof-of-principle development at laboratory scale with different molecules, and the results, says Konstantin Konstantinov (vice president, commercial process development) have been “highly successful and very exciting.” I spoke with him about the technologies involved in a completely integrated continuous process and how it may affect the design of future facilities.
BPI: What are the first steps toward developing a continuous process?
KK: Our starting point is to focus on the process objectives first and then think about the facility. The process is the driver. We want to design a process that is universal, flexible, and very standardized. At the end of the day, we want to have a platform that can be used for any therapeutic protein. This is generally what happens in many industries when they mature: They come up with one harmonized process, often referred to as “the dominant design.” This is what Genzyme is trying to develop — the dominant design of the future.
BPI: How does a continuous upstream process affect downstream?
KK: The downstream conversion into continuous is more interesting than the upstream. I started working on perfusion processes more than 25 years ago, so it is not something that is very unique or that the biotechnology industry recently invented. It is improving, however, because we are using better cell lines, better media, and better cell-retention devices that allow us to operate at very high cell densities.
The newer concepts are in continuous downstream operation. Genzyme has converted the first step (capture) into continuous processing. It is directly integrated with the perfusion cell culture process without any equipment between the two operations. So the harvest hold tanks, microfiltration, and centrifuges, are all removed.
Downstream steps are integrated with the upstream steps; all the flow rates are adjusted to be the same. Integration requires that harvest coming out from the perfusion bioreactor is continuously loaded onto the capture columns directly. It should be compatible with the pH, osmolality, and other parameters without many adjustments. We try to ensure that cycle times and hold times are minimized or completely eliminated for rapid processing, which is especially beneficial for proteins that are sensitive to degradation during any hold steps.
BPI: Continuous processing has been used in other industries. What can biomanufacturers take as “lessons learned” from those industries?
KK: It’s a lot, actually. I mentioned that biotech companies have worked on continuous upstream quite a bit. But Genzyme’s understanding and vision of integrated continuous processing has significantly evolved over the past three years or so. There are many industries that started as batch, but over the years they have converted to continuous (for example, steel, petrochemical, glass, paper, food, some chemical and pharmaceutical). If you read books and articles related to this, you will discover that despite the diversity of these industries, the driving forces to convert to continuous are always the same. Those industries make changes for same reasons that are completely applicable to biotechnology — with the exception of one in the petrochemical and chemical industries. which cannot include hold steps because some intermediates are highly explosive or toxic. All other business drivers are exactly the same.
BPI: How is a batch defined in a continuous process?
KK: That question is always asked when we give presentations. The answer is very simple: As long as your processes and logistics are well-established, validated, and documented, and you supported your process with solid product-quality data, you can define a batch any way you want. It can be time-based, it can be product-volume based, or a combination of both
BPI: How do equipment changes needed for converting to continuous processing affect facility design?
KK: Continuous upstream processing is being done in traditional bioreactors, but of much smaller size compared with batch processing. We are considering small disposable reactors of several hundred liters. For perfusion, you don’t really need exotic upstream equipment beyond robust cell-retention devices.
There is a saying in the chemical industry or other industries relating to continuous processing “think continuous, think small.” When converting to continuous, the volumetric productivity of the process dramatically increases. So, you reduce the size and volume of equipment dramatically as well. You can operate with very small tanks. For example, the size of a column for a continuous chromatography capture step is greatly reduced. A column for a typical batch process is usually more than 100 L of volume, but if you convert it into continuous process with the same capacity and volumetric productivity, the column volume decreases 50- to 100-fold. Thus, the footprint for the equipment is greatly reduced, which makes the facility smaller. A facility running an integrated continuous process will be significantly smaller than a large-scale batch facility, thus reducing capital expenses.
BPI: How is scale up conducted for a continuous process?
KK: To borrow a term from the chemical industry, we “number up” — add more parallel small production lines instead of scaling upward. One important difference, however, between the two industries is the limit to how small the reactors can be. One trend in the chemical industry that has shown to be very beneficial, is the use of microreactors. You can produce very large volumes using several parallel microreactors, because those reactors can dramatically increase temperature and pressure. That accelerates chemical reactions, so chemical companies can reduce volumes much more than we can. We can run tanks of a few hundred liters, but we cannot reduce the volume to microscale. The temperature and pressure of our operations cannot increase too much because cells and the proteins must be kept under proper conditions.
BPI: How does the biologics industry’s needs for greater flexibility play into process design?
KK: Flexibility is becoming a major business driving force. It’s important because so much is going on in the business these days. It’s a competitive industry, and it’s often hard to predict long-term market demands. Demand depends on the drug, competition, government regulations, market for biosimilars, and so on.
So, how do you build a facility that can react quickly and flexibly to market demand? A flexible facili
ty should handle a large variety of therapeutic proteins, both stable and less stable (biochemically), that are required in large or small volume. In the past, the industry had different technological platforms to handle different cases. Genzyme is trying to have one platform that can handle all types of protein drugs.
A manufacturing facility that provides this flexibility, would have a couple of parallel production lines, but not large in volume. Continuous operation has two dimensions of flexibility: As I mentioned, the first one is that you can add or remove production lines without significant cost penalties. The other is that you can time the length of your production run. For example, if demand is smaller, you can run the process for two or three weeks. When demand increases, you can run the process for three months or longer. Such a facility has flexibility in product volume. If you want to produce a couple of hundred kilograms of product, you may dedicate two or three parallel lines to one product. One line can be dedicated to a small-volume product, and yet another line, for example, can be dedicated to another small-volume product.
Such a facility would be a closed-system design — the “facility of the future” definitely is a multiproduct facility. One interesting and useful concept being discussed in the industry is the concept of a multipurpose facility. As I said, the reactors for the whole production line are small, so you don’t do a lot of scale-up. The scale in development, the scale in clinical manufacturing, and the scale in commercial manufacturing are the same. In late-stage development we can run a reactor of a few hundred liters, which is the size of reactor needed for clinical manufacturing and commercial manufacturing. Ideally, in one facility, you can produce clinical material because the equipment is the same, and the scale is the same; you can produce commercial material and product multiple products. Such ideas are gravitating around the same concepts of continuous, closed-system operation. Together with disposable technology, this provides some powerful opportunities.
BPI: How does this all affect facility design?
KK: You may have heard the term ballroom-like facility — that is what we think about the facility of the future. It’s not something that is chopped into small rooms or using rooms to segregate and separate everything. Because the process is a closed system, you can keep your processing equipment in one large “ballroom.” This facility design is more likely to provide mobility than are current “stainless-steel” facility designs because the footprint of continuous equipment instrumentation is much smaller and portable. Therefore, you can move the equipment around and reconfigure the facility and add or remove lines without breaking walls and breaking the roof down to bring in new tanks. Disposable reactors of a few hundred liters definitely provide a lot of flexibility in terms of mobility. The same is true for the downstream skids. If they are much smaller than what we currently have, it is easier to modify a facility according to current needs.
BPI: What is needed to promote the adaptation of continuous processing?
KK: The industry has to continue to accumulate experience, especially downstream. We have plenty of experience in upstream processing, but the difference is that now it is an integrated system. The biotechnology community should continue to work actively on the concept of closed systems. That is essential because once you demonstrate and validate your system as a closed system, it may allow you to simplify your environmental control requirements for your facility, and you can conduct multiproduct operations as well.
One key requirement of continuous processing is good process control to keep operations at steady state. You have to have appropriate process analytical technology (PAT). As we start designing the process and the facility of the future, the arsenal of reliable PAT tools should grow. For example in our system, the downstream continuous capture step uses a very promising UV-based PAT that works completely automatically (described in Reference 1) over extended time periods.
Maribel Rios is managing editor for BioProcess International; firstname.lastname@example.org.