The manufacture of therapeutic proteins, particularly monoclonal antibodies, has evolved into reliable, robust protocols characterized by platform processes and standardized unit operations. These production methods, and their attendant efficiencies, have been disruptive in the traditional sense but took decades before they were universally accepted. For example, single-use processing took years to reach its current status, and has perhaps reached an adoption steady state. Similarly, the roughly tenfold improvement in titers for fed-batch CHO (Chinese hamster ovary) cell cultures—a significant driver for disposable manufacturing—has been transformative, but took decades to achieve.
Within this context, therapeutic biotechnology is constantly searching for unit operations that outperform standard, tried-and-true unit ops. Beginning at the beginning, with expression systems, and moving along the manufacturing value chain through culture, capture, polishing, and formulation, processes not incorporating CHO, fed-batch, protein A, and anion exchange exist and out-perform older methods in certain situations, at least on paper. Whether companies ever adopt them is another matter.
We discussed these topics with a team from GE Healthcare, which is as responsible as any company for the status of biomanufacturing in 2019. Our experts are Peggy Lio, Global Cell Culture Services Leader; Peter Guterstam, Product Manager, Chromatography and Membrane Adsorbers; Gustav Rodrigo, Senior Scientist, Non-Protein A Capture; and Sofie Stille, General Manager BioProcess Downstream Resins.
Taking the platform approach to monoclonal antibody production as a reference point, and noting that steps in such a process are typically discrete, separate, and of the batch type, the first promising innovation to consider is continuous cell culture or perfusion culture, an area in which GE is among the commercial leaders.
What is the status of continuous cell culture today? Is it a growing option, or has it settled into a steady state with regard to utilization?
Peggy Lio: Perfusion’s main advantage is cell density, and its attractiveness is a consequence mostly for processes that are low-producing. A typical fed-batch process might reach 20 million cells per milliliter, while continuous cell culture easily reaches 80 million cells. This translates directly to a higher volumetric productivity (protein per volume per unit time), and a compression of time-to-market or time-to-failure. Perfusion systems have a smaller physical presence than batch cultures, which means processes can occupy a smaller area within a manufacturing suite. Despite these advantages, and ongoing interest both from users and vendors, continuous cell culture is not quite yet ready for large-scale biomanufacturing. Its integration into conventional large-scale processes, has been limited to seed trains, where perfusion allows skipping intermediate scaleup volumes. Many challenges exist, however. Users are concerned about the potential effect of continuous cell culture on product quality. The issue remains as well regarding how to define “batch” from a regulatory and scientific perspective, since the whole point is to avoid a physical batch.
Peter Guterstam: That being said, perfusion cell culture is very attractive for new classes of biotherapeutics, not necessarily proteins. Perfusion cultures are well-tailored for non-protein products such as therapeutic exosomes, and gene therapy. In addition contract manufacturers have been setting themselves up for large-scale perfusion for more traditional products as well as non-monoclonal therapeutic proteins. Perfusion will also be attractive for companies developing proteins that are difficult to express or manufacture, that are not as easy to work with. Having the experience of using perfusion with CHO and mAbs will be invaluable in applying that approach toward obtaining higher yields for non-traditional products.
Should biomanufacturers be considering other upstream alternatives, perhaps new expression systems?
Peggy Lio: CHO will remain the workhorse expression system for mAbs for a long time. Developers will turn to CHO, even for non-mAb proteins, because there aren’t many widely industry accepted, reliable, and robust expression system alternatives. Instead there is a renewed interest in cell-line development platforms enabling targeted integration of the gene of interest, rather than traditional random gene integration. Right now a typical time for cell-line development is about four months, which is actually pretty short, but targeted integration reduces the number of clones you need to screen to identify high and stable producers. With so many more molecules entering development and clinical studies, there is a need to find those high-producing “needles in haystacks” in shorter development timelines with new enabling platforms.
Peter Guterstam: With so many more molecules entering development and clinical studies, including biosimilars, no one has the resources to look for those “needles in haystacks” in an unstructured manner.
Sofie Stille: There is one caveat, though, generally, regarding upstream efficiency. If bioprocessors want to rely on existing cell culture and expression systems, they may need to accept a lower titer to gain advantages in time and efficiency. A new expression system is a nice idea, but it takes a very long time even to adopt even a new CHO line. Existing, approved lines are always less risky because they’ve already been through regulatory scrutiny.
Downstream purification presents greater opportunities for process innovation. Of capture, filtration, and chromatography steps, which are the ripest for innovation?
Gustav Rodrigo: There are only one or two processes I know of where protein A is not used to capture immunoglobulins. Cation exchange is cheaper, and at one time had a reputation for higher capacity, but this has been rendered moot by newer, higher-capacity protein A resins. Additionally, cation exchange capture must be developed for each product, whereas protein A is robust and provides amazing purity in just one step with minimal development.
But protein A doesn’t necessarily require a particle-based matrix to be effective. In 2017, GE Healthcare acquired Puridify, which specializes in membrane technology. We’re looking at protein A in membrane format.
Precipitation has been considered a possible alternative to a formal capture step, but manufacturers aren’t adopting it because, like cationic exchange, it requires considerable development, and optimization for every product. Companies just don’t have the time for that.
Sofie Stille: The challenge for suppliers, given the diversity of antibodies and emerging therapeutic proteins, will be to identify novel affinity capture modalities, particularly for situations where protein A is unsuited. Our goal is to discover purification platforms as efficient as protein A for those molecules.
Peter Guterstam: Protein A’s affinity for immunoglobulins has evolved over millions of years, and will be very difficult to duplicate.
But don’t membranes have very low binding capacity compared with resins?
Gustav Rodrigo: Yes, but Puridify found that you can cycle protein A membranes much more frequently than columns. In the same time it takes to run just a few cycles on a resin you can cycle a membrane hundred of times, and purify the same amount of protein or achieve the same volumetric productivity with a unit that’s much smaller than a column. We expect, furthermore, that the capacity of membranes will continue improving.
Peter Guterstam: Cycling presents opportunities to improve productivity but it will be crucial to have the capture step eluate at the right concentration. In addition, the design of the membrane itself, the housing, must be optimized to keep product concentrations as high as possible before moving to polishing steps.
Gustav Rodrigo: One potential bonus from using a membrane-based medium is that it enables single-use capture, which is not economically feasible today due to the expense of protein A resin. A capture membrane will contain so little resin, companies will be able to throw them away and not worry about cleaning or cross-contamination after a single upstream batch is processed.
Conclusion
GE Healthcare has been at the center of bioprocess innovation since the company was known by another name. With an obvious commercial interest in replacing expensive unit operations with less costly, more efficient ones, GE is nevertheless cautious about the prospects of alternatives to platform processes and unit ops. GE’s experts noted that healthcare generally, and biomanufacturing specifically, are conservative, highly regulated industries, which are simultaneously highly innovative (in terms of products, at any rate) and fiercely competitive. So, regardless of how exciting an expression system or purification step looks on paper, it is always considered in light of the overall process, the marketplace, and its inherent regulatory risk.