June 1, 2014 (Vol. 34, No. 11)
Angelo DePalma Ph.D. Writer GEN
Bioprocessors have adopted numerous approaches to optimizing cell culture, from clone selection to process conditions to feed and media strategies. Finding high-producing cell lines has been a major focus of development programs for protein therapeutics.
At the recent BioProcess International European Summit in Prague, a number of presenters discussed various methodologies and options for optimizing cell culture operations.
According to Tim Ward, strategic marketing director at TAP Biosystems, now part of Sartorius Stedim Biotech, a clone’s specific productivity defines the size of the production facility and, ultimately, the drug’s profitability. Given the importance of high-dose monoclonal antibody therapies, bioprocessors have come to expect multigram-per-liter results from platform processes. “Achieving this goal rapidly and efficiently is essential,” says Ward.
Because traditional cell-line optimization is often costly and time-consuming, the industry has searched for tools to improve screening and selection. Initially, developers focused on automated clone screening (such as Genetix’ ClonePix) and microplate-based liquid-handling systems (such as the TAP Biosystems’ Cello). More recently, the focus has shifted to high-throughput microbioreactors.
The first such system, BioProcessors’ automated SimCell, used a custom-designed, multiple-chamber cassette system that ran hundreds of plates in parallel. Although several companies tested the system, SimCell’s size and complexity led to limited adoption. Smaller benchtop systems were also developed based on 24-well shaken plate designs.
Systems such as MicroReactor Technologies’ Micro24 (now sold through Pall) initially focused on microbial strain selection. Later they were tested for cell cultures despite lacking automated pH control, a shortcoming that was overcome through liquid base addition.
The first true microscale mimic of a benchtop bioreactor incorporating an impeller, gas sparging, and full pH control using CO2 and automated liquid base addition, was TAP Biosystems’ ambr. This system combined individual bioreactor control for 24 or 48 15-mL reactors in parallel, with the advantage of fully automated liquid handling for media, inoculate, feed, and sampling.
“Many reports note ambr’s scalability of process parameters, which results from the system’s ability to model all critical aspects of a benchtop stirred tank reactor—from one up to two hundred liters,” notes Ward.
Fully Closed Upstream System
Transferring cells from frozen vials to shake flasks during cell expansion is one of several open upstream steps that entail the risk of contamination. Expansion may take several days or weeks. A group at Merck Millipore headed by Aurore Lahille, new technology manager, has studied this problem and developed a completely closed upstream process, based on single-use bioreactors, that combines disposables with traditional process steps.
Lahille evaluated cell freezing and thawing in disposable bags as the run-up to inoculation of a 1,250 L bioreactor. She demonstrated feasibility through a trial employing seven different CHO cell lines using single-use bioreactors of the type and volume (3–200 L) commonly used during process development and as seeding or production bioreactors. Lahille also conducted several clinical-size runs at 200 L and 1,250 L to ensure a meaningful comparison, and compared glass and stainless-steel bioreactors ranging in size from 3.6 L to 1,250 L. “As a result, we have developed a fully closed USP process by coupling cell freezing in bags and disposable bioreactors up to production scale,” Lahille says.
The new process saves time by protecting cells at every stage, and by virtue of its larger working volumes compared with conventional amplification protocols. “We can freeze 50 times more cells in bags than in vials,” Lahille explains, “and this cuts amplification time significantly.”
Another advantage of interest to contract manufacturers or companies with robust pipelines is greatly improved facility utilization through the realization of fully disposable multiproduct suites.
The upstream process, which Merck Millipore has validated, is based on existing single-use bags and connectors from the company’s catalog. Lahille has now turned her attention toward a fully closed downstream process, for which she says new single-use technology may need to be invented.
Single-Use for Microbial Fermentation
“We view single-use bioreactors in cell culture as stable, almost mature,” says Ken Clapp, senior global product manager for Xcellerex bioreactors and fermenters at GE Healthcare. Almost in parallel, CHO cultures have become industrialized and highly optimized for cell densities and productivity.
As single-use equipment evolved, the issue of a practical size limit has practically disappeared. “No longer do customers ask for volumes of 5,000 L or more,” Clapp remarks. “With 2,000 L being offered by us and our main competitor, current production needs are addressed with one or more of those reactors.”
What does Clapp see down the road for mammalian culture? Mostly tools that facilitate efficiencies, such as perfusion culture, single-use heat exchangers, and large-diameter sterile connectors.
The focus now turns to microbial fermentation, which as a process is much more demanding than cell culture, especially with respect to power, heat removal, and mass transfer. Users of fermentation bioprocessing are now desirous, Clapp says, of the benefits universally expected from single-use, namely, lower capital costs, simpler infrastructure requirements, decreased risk of cross-contamination, lower water usage, and greater versatility.
“Our data show that with the right design and selection of materials, a scalable product line of single-use fermenters is possible,” Clapp asserts. The typical performance and scale-up metrics from stainless-steel stirred tank reactors (such as kLa, oxygen transfer rate, and power per unit volume) are readily used to make correlations in support of wider industry adoption of disposables, particularly with respect to the challenges of microbial processes.
Materials of construction will play a significant role. Clapp notes that when various grades of stainless steel failed to meet the needs of animal cell culture, new ones were invented or adapted: “In those cases, we relied on a metallurgist to understand and guide selection. Now, we have an alphabet soup of polymer names, and polymer chemists are the new metallurgists.”
The Boehringer Ingelheim High Expression Technology (Bi-Hex®) is an established, high-expression technology platform. Boehringer naturally has optimized media for that platform, but what may not be generally known is that Bi-Hex media work quite nicely with other cell lines, including GS CHO, DG44 CHO, and CHOK1, according to Benedikt Greulich, Ph.D., associated director of process science at the company. “We’ve seen titer improvements of up to 3.5-fold,” says Dr. Greulich.
The power of this medium arises from how it was designed. Classically, media developers mix and match components, or media, themselves. Then the developers test for desirable characteristics in cells that have been grown in the media.
Dr. Greulich and colleagues departed from trial-and-error methodology, relying instead on gene sequencing and expression analysis, metabolomics, and DNA chips. With these tools they discovered deregulated genes and/or pathways, and identified media ingredients that corrected the situation and improved titers.
In one case, they achieved a 30% titer increase by adding lipids. In another experiment that employed metabolomics, they discovered extracellular metabolites that transformed lactic acid producers to lactic acid consumers.
Lactic-acid-consuming cells have a twofold greater citric acid flux as well as higher energy production within the mitochondria. By contrast, producers have an overall reduced carbon flux, disturbed energy production, and a reduction in pathways associated with increased biomass.
“Several of these metabolites were suitable as media components, including three that inhibited lactic acid production, and thereby caused a 17% titer increase,” Dr. Greulich asserts.
If the claims are correct, Bi-Hex and its optimized medium represent a new level of platforming. “It’s possible to obtain several grams per liter of product simply by applying these process conditions, without complex handling or feed addition schemes, and with a minimum of process development,” Dr. Greulich explains.
Biological Basis for Media Development
Optimizing cell culture processes is a balancing act, where the more that is known and controllable, the greater the opportunity to create a more predictable process. Among the critical variables are gas entrance kinetic energy and vessel liquid height.
Gas entrance kinetic energy is determined by the velocity of gas entering the bioreactor. All things being equal, it would seem that sparging at a higher gas volume per process volume over time would result in higher gas kinetic energy. Not so.
“Higher flow is only part of the issue,” says Christopher Brau, associate engineer for bioprocess production at Thermo Fisher Scientific. While kinetic energy may be directly increased and decreased in existing systems by raising or lowering the volume of gas delivered, in new systems one has the option of changing the size and quantity of the pores through which gas passes. Pore size, notes Brau, directly affects not only the velocity at which a given gas flows, but also how fast the velocity will increase or decrease relative to pore size.
Brau studied systems in which the velocity or kinetic energy of the gas entering the vessel was maintained below thresholds determined to create more uniform behavior. “Excessive kinetic energy in a sparge system risks damage to cells in addition to generating a wide bell curve of bubble sizes which has its own host of potential issues,” he explains.
Similarly, vessel liquid column height and mixer flow pattern determine the total area available for mass transfer between liquid bulk and sparged gas. When bioprocessors do not account for these factors and the same sparge design is used despite increasing liquid columns, the total area available for mass transfer skews the behavior of a sparge system, potentially quite unfavorably. “This often manifests as a combination of CO2 buildup, a gas gradient in the column, excess foam, high holdup volume, and higher cell entrainment in the foam layer, while reducing the uniformity of scale up behavior between vessels,” Brau elaborates.
Brau’s work holds special significance for kLa, the mass transfer coefficient (kL) multiplied by the area (a) available for mass transfer based on a simplified gas liquid film theory equation. Here, kLa is a combined term describing relative mass transfer efficiency for a given set of operating conditions.
Brau measured kLa for both O2 delivery and CO2 stripping. “The significance here compared to most of industry is that we look at and consider the tradeoff between the two processes, as well as the limitations that govern and drive the behavior in the first place,” Brau says.
An excessively efficient sparge will achieve the maximum possible oxygen delivery, but is likely to generate excessive foam and ensure CO2 buildup. “That is why you have to account for the ratio of performance a sparge delivers under given operating conditions, specifically the ratio of O2 delivery and CO2 stripping,” Brau explains. Going too far toward CO2 stripping causes the overall sparging efficiency to suffer.
“You have to strike a perfect balance to achieve an O2 delivery and CO2 stripping kLa that is harmonious for your process conditions,” concludes Brau. “In other words, [this is about] managing the quantity of O2 consumed and CO2 produced. Striking this balance means you are maintaining your cell culture in ideal conditions with [the smallest possible] gas flow rates, relatively speaking.”