June 1, 2015 (Vol. 35, No. 11)
Angelo DePalma Ph.D. Writer GEN
Solving Bottlenecks May Require More Efficient Scheduling and Digging Deeper into the Workings of Unit Operations
Originally conceived as being analogous to what chemists would call a slow step in a chemical reaction, bioprocess bottlenecks are still primarily viewed as any unit operation that slows down the overall process. In practical terms, the definition may be expanded to operations that are for whatever reason difficult, risky, or expensive.
Bottlenecks occurring as a result of inefficient resource usage may often be alleviated through more efficient scheduling. Solving other bottlenecks may require process engineers to dig more deeply into the workings of unit operations.
Facility versus Process Bottlenecks
Guenter Jagschies, Ph.D., senior director of strategic customer relations at GE Healthcare Life Sciences, categorizes bottlenecks as facility-related and process-related. The former refers to issues originating from the size or capacity of installed equipment, for example, in tankage for buffer or intermediate product storage. “These volumes typically increase when the product titer increases,” Dr. Jagschies says, “and more mass needs to be bound to larger columns, leading to more buffer consumption and large intermediate product eluate.”
Process-related bottlenecks are based on issues in the flow of product through the process steps. For example, a process-related bottleneck may occur when a step has significantly lower batch handling capability than the previous one, typically either with regard to either volume or the product mass to be handled in a given time.
“Again, increased titers can be a cause for mass bottlenecks, as in a centrifuge’s biomass limitations and/or filter cascades in the initial post-bioreactor product harvest,” Dr. Jagschies continues. Highly diluted feedstreams, moreover, can cause volume-handling restrictions. This occurs when a process is limited to low-productivity cell lines.
One remedy for buffer-volume-related issues has been “in-line conditioning” (ILC). In such systems, buffers are created by a machine controlled by algorithms that accurately produce most if not all buffers applicable to a process from the buffer’s individual chemical ingredients rather than from conventional buffer concentrates. ILC simplifies buffer making and makes the process more robust. Up to 70% of the labor effort and the storage space can be saved as the buffer is delivered directly to a chromatography skid, for example.
Dr. Jagschies notes that centrifuges, which may experience issues from the increased biomass of high-titer bioreactors, may be entirely avoided in processes of up to 1,000–2,000 L bioreactor volumes because the harvesting process may be economically performed with just filters, for example, single-pass tangential flow filtration (TFF) modules. “However, the harvesting process is one of the concerns in a modern highly productive process where solutions are not yet widely established or developed to platform status,” he explains. “A candidate technology for aiding filtration is flocculation using various options of salts or cationic/anionic polymers, followed by fast settling of the flocculate.”
Jonathan Haigh, Ph.D., head of R&D downstream processing operations at Fujifilm Diosynth Biotechnologies, notes that early downstream processing provides unique opportunities for process intensification: “Rising upstream titers have forced companies to adjust downstream operations to maintain productivity and throughput.”
Using two homegrown technologies, Fujifilm has streamlined the early operations related to protein capture. One strategy involved dilution; the other, concentration.
Managing volumes and types of buffers, and their deployment times for every downstream unit operation requires scheduling and floor space. To reduce these requirements, Fujifilm has introduced a buffer-on-demand approach that reconstitutes concentrated buffers at the point of use.
In-line buffer dilution becomes necessary as downstream operations attempt to keep up with upstream productivity. Columns must be reused and recycled more frequently than in the past, with each pass requiring higher volumes of buffer.
“Making up all those buffers beforehand would entail a significant footprint and resources, explains Dr. Haigh. “Now we can utilize more of our manufacturing asset for equipment or load material.”
The second technology grew from necessity—a molecule that was sensitive to upstream process conditions and column load time. Concentration reduces load time, thereby reducing product loss due to instability.
“Buffer management can be quite challenging,” Dr. Haigh emphasizes. A typical downstream process employs three columns, several filtration steps, and buffer exchanges. “The number and volumes of buffers is vast. It’s not always an option to generate all of them up front at the beginning of the campaign. Generating concentrated stocks ahead of the process, and diluting at the point of use gives us a much better grasp of buffer management.”
Fujifilm uses software and equipment developed in-house that provides accurate buffer formulation and reconstitution. “This way we don’t spend nearly as much time on the shop floor titrating buffers for desired pH and conductivity,” Dr. Haigh adds.
The company is also evaluating third-party technology for dilution, as well as options for sourcing raw materials as concentrates or in-house formulation. “It’s something we have flexibility with,” Haigh asserts. “We’re not tied to one route or another. We’re considering both processing and regulatory factors.”
Feedstock concentration by ultrafiltration was also developed in-house, to address specific needs for a labile product for which time between initial generation and column loading was critical. Fujifilm, which has obtained good results for this technology at pilot scale, is investigating it at large scale. Thus far, volume and load time have been reduced approximately fourfold.
According to Dr. Haigh, the concentration technology may pay off even for molecules that are not particularly labile. “We’re still evaluating it from a cost-of-goods perspective,” he remarks. “We expect to understand those factors in the near future.”
Cost versus Quality Drivers
Due to rising titers, capture chromatography is still the most serious and costly bottleneck for many bioprocessors, says Lynne Frick, director of sales, continuous bioprocessing, Pall Life Sciences. She adds that high titers put a great deal of pressure on downstream operations during manufacturing.
For low-titer processes, single-pass TFF reduces process volumes while adding efficiency during capture. The idea of combining simulated moving bed chromatography with continuous viral inactivation has Frick’s support.
The combination could be especially suited to a single-use product contact flow path. “Even though no one has achieved it under GMP yet, innovation continues to happen,” Frick observes. Moreover, the combination should eventually allow process fluid to flow directly from Protein A capture to viral inactivation and on to the rest of process.
Continuous processing is another way to relieve the capture bottleneck, particularly for monoclonal antibodies. “Continuous chromatography is great in and of itself as a unit operation substitution, for example, to reduce chromatographic media use in this single step,” Frick notes. “However, its true promise is integrating with other steps to improve efficiency around the bottleneck.”
High-titer-platform monoclonal antibody processes constitute 40% of today’s therapeutic bioprocessing market. Such processes are where continuous processing will be adopted first.
At the other end of the spectrum, the purification of labile molecules (for example, enzyme-replacement therapies) will also benefit from continuous processing, for different reasons. “Many of these are manufactured in low-titer perfusion processes where you just want to get the molecule through the process as quickly as possible to enhance product quality,” Frick explains.
Continuous chromatography has the potential to reduce resin bed volumes during capture by 80%—a tremendous savings given that Protein A is the dominant cost driver for downstream processing of monoclonal antibodies. But Frick believes that the ultimate driver for continuous processing will be quality, not cost. The FDA has placed great emphasis on quality and its great enabler, process control. “And batch processes are innately more difficult to control than continuous processes,” Frick concludes.
Bioprocessors find themselves at a crossroads, says Surendra Balekai, senior global product manager, Thermo Fisher Scientific. Cell culture product titers and cell densities have risen dramatically and continue to rise. Balekai believes, however, that the adoption of upstream continuous processing may reintroduce capacity mismatches between upstream and downstream operations. Continuous processing in the form of perfusion cell culture is significantly ahead of continuous purification in terms of robustness and general adoption.
Another “lag” area exists for single-use products. Although they are now widely adopted upstream, downstream applications are rare outside of filtration, membrane chromatography, buffer prep, and hold operations. Downstream purification is chromatography-intensive, and resins are still too expensive to discard after one use.
Moreover, disposable sensing technology could stand improvement. “Single-use sensors for pH, conductivity, dissolved oxygen, etc. exist, but there’s a general lack of standardization,” Balekai complains. “The perception is that single-use sensors are not as reliable as they could be. Bioprocessors will often use them side-by-side with multiuse sensors for peace of mind.”
Single-use sensors are built into disposable process containers, which together are gamma-irradiated and delivered as a sterile system. According to Balekai, single-use sensors currently available have been known to drift after few days of use. Several new technologies in single-use sensing are under evaluation.
Reducing Media Consumption
A team at Bayer Healthcare led by Venkatesh Srinivasan, Ph.D., director, manufacturing sciences, has shown that bioprocess production efficiency may be increased by reducing media consumption through an intelligent perfusion rate reduction strategy that also preserves product quality. The validation strategy was designed to leverage qualified scale-down bioreactors to minimally utilize production capacity for process validation.
The group justifies this approach by demonstrating consistency from design through performance. They used a scaled-down, mammalian cell culture perfusion bioreactor in a case study and compared results with those from a commercial-scale system.
“Acceptance of this new approach—the validation of process changes through a scaled-down approach—could greatly facilitate implementation of important process improvements, increase production capacity, and improve efficiency while minimizing changes to infrastructure,” noted Dr. Srinivasan.