October 15, 2014 (Vol. 34, No. 18)
Angelo DePalma Ph.D. Writer GEN
The adoption of modern analytics, design of experiment approaches, robust scaledown models, and data analysis software has aided bioprocessors in addressing or solving most of the major engineering complexities related to bioprocess scaleup.
According to Susan Dexter, principal consultant at Latham Biopharma, the overriding issue right now is to preserve quality.
“Most companies prefer, and are better positioned, to operate at larger rather than smaller scale,” she says. “If process development has proceeded well, and you’re only fine-tuning—not making a lot of substitutions in reagents and materials—the real challenges are maintaining product quality attributes, and understanding how various parameters affect quality as the process scales.”
Dexter cautions against setting quality set points too narrowly, and having product fall out of specification as a result of scaleup. When narrow set points become difficult to maintain at larger scale, manufacturers might consider establishing the set points themselves as quality attributes, or changing those set points within the capabilities of analytics, which Dexter calls “the key” to understanding product quality.
Engineers will say that scales differ in fundamental mechanics—that pressure is higher at the bottom of a bioreactor, for example, or that gas and mass transfer, agitation efficiency, shear forces, and other parameters are difficult to duplicate between scales. True, but process parameters at 2,000 L need not be exactly identical to those at 3 L provided quality remains within specifications and yield is acceptable.
Dexter says bioprocessors must believe in their scaledown models against the backdrop of quality: Processes at 10 L that do not predict performance at 200 or 2,000 L cannot be considered scalable. “You don’t want to do process development at 2,000 L.”
Early Cell Work
Scaleup will proceed less than optimally if developers do not approach early cell work systematically, says Vincent Pizzi, upstream strategy leader at GE Healthcare: “If you prepare properly, you will ultimately obtain a stable, high-yielding cell line that you can use in production.” The process involves utilizing the appropriate screening tools; obtaining gene integration at a stable, productive region of the genome; selecting clones; measuring titers; and applying selective pressures on the cell line itself, which is often engineered through knockouts.
“You don’t want to leave any stones unturned when choosing clones because if it works out, this becomes your future manufacturing platform,” Pizzi adds. “The details you concentrate on early with clonal selection makes a huge difference at manufacturing scale.”
At this stage, developers should work on optimizing cell culture conditions including medium along with feed and uptake of oxygen and glucose. Initial pilot and manufacturing scale productivity can only be as good as what was achieved during development.
The near-universal adoption of serum-free, chemically defined media in drug development settings has reduced variability between scales considerably. On the negative side, these formulations are more appropriate for suspension cells that have been adapted to them. Developers still must transform attachment-dependent cell lines, which usually arrive in serum-containing media, to suspension culture media. This adaptation process can consume considerable time and could result in nonoptimal cell growth.
By contrast, clone screening is not the bottleneck it was a decade ago. “Much of the manual work has been automated,” Pizzi tells GEN. Companies may choose to produce more clones for evaluation, or to complete standard workflows more quickly or with less human intervention. He mentions flow cytometry as a tool for selecting highly productive clones after expansion. Automation around titer analysis through liquid chromatography has also become streamlined through modern UHPLC systems and automation of liquid handling, sample prep, and injection.
Another area of improvement has been transfection. Pizzi mentions electroporation, new transfection reagents, and improved methods as bright spots. GE has recently licensed technology for site-directed gene integration, which allows greater control over the location within the genome where the vector integrates. Pizzi calls these locations “landing pads.”
The benefit, he says, is a reduction in the number of clones required to produce cells with the right combination of viability, growth, and productivity. “That capability is still in the future, but we have great hopes that it will make these aspects of cell transfection and selection of high-expressing clones extremely efficient.”
In the age of process understanding, the ability to scale up and down is critical, according to Willem Kools, head of field marketing and biomanufacturing sciences network at EMD Millipore. Scalability is essential for all unit operations, including cell culture, membrane chromatography, and mixing. Vendors support customers with “proof statements” or validation documentation certifying that they have conducted and demonstrated scaleup for specific product lines. Kools cites documentation on leachables and extractables for single-use products as examples of essential, usable documentation.
Customers increasingly ask about transferring processes between single-use equipment and stainless steel. Within this context, mixing is apparently one of the simplest unit operations to execute but one of the more difficult to describe mathematically. EMD Millipore has recently introduced improved mixing-scaling understanding through its NovAseptic® Mixer and disposable Mobius Mix®.
Downstream, vendors must demonstrate that column-packing protocols are well understood so chromatography processes are scalable and robust. Anyone who has packed a column knows that achieving the same flow distribution at small and large scale can be challenging. “Customers are seeking assurance that large production columns may be packed consistently,” Kools says. “The key is reproducing process conditions experienced at manufacturing scale across all scales—and not simply for volumes or membrane surface areas, but for the entire process.”
For example, it was recently reported that at large scale, virus filters might lose their ability to clear viruses when they are de-pressurized. This effect may not be noticed at small scale, so accounting for it in scaledown virus clearance validation is difficult. Nevertheless, virus filter vendors need to create an appropriate model to prove their products are capable of reducing virus load. EMD Millipore has performed this study for its Viresolve Pro® virus filters, and shown that the effect of depressurization is insignificant.
Attachment-dependent cell cultures occur at scales much smaller than those for antibody production, yet scale-up issues remain for vaccine and cell therapy companies that rely on T-flasks and microcarriers.
While cultures of anchored cells may be smaller in volume than most CHO cultures, as the culture containers multiply, their maintenance becomes cumbersome and labor-intensive. “That is why we often collaborate with automation suppliers,” says Ken Ludwig, business director at Corning. “It’s difficult to move large 40-layer stacked vessels like CellSTACK® when they’re full of media.” CellSTACK is Corning’s stackable—and eminently scalable—system comprised of multiple flat culture vessels, each with 636 cm3 of culture area. For even greater culture area Corning offers CellCube®, a compact, perfusion bioreactor.
Automation eliminates the drudgery of manually replenishing media, moving cultureware without spilling, and associated human error. Many vaccine and therapy providers use off-the-shelf automation such as Thermo Fisher’s Automatic Cell Factory Manipulator System and TAP’s Cellmate, whereas other providers prefer customized production automation, such as that provided by Invetech.
Stacked plate vessel products mimic culture conditions in T-flasks and roller bottles, where cells grow on a modified surface. Alternately, cells may also be grown on microcarriers suspended in medium. Instead of growing on 2D surfaces, the cells grow on the surface of 3D beads. Companies like Corning continue to innovate in this area. Last year, the company introduced four new microcarriers with surface treatments that improve cell attachment and viability.
“Whereas suspension cell culture suppliers talk about volumes, we usually talk about square centimeters of culture space. The more beads you add, the more space is available for cell culture,” Ludwig says. “We think of it as scaling out vs. scaling up.”
Vaccine Growth and Harvest
Adherent cell cultures for producing recombinant proteins, say, for vaccines, do not require very large quantities of product. Since cells are destroyed after they secrete the protein of interest, their health status at the end of culturing is not a concern. Cultures for cell therapy, including stem cells, require that cells be removed from culture intact and viable. Harvesting is accomplished with proteases, but conditions must be mild enough that cells and their surface receptors are not harmed.
Vaccine producers historically used roller bottles or multitray systems to grow cells from which they harvested viruses, for example, poliovirus. That production process is eminently scalable: 250 bottles present no particular engineering problems compared with two bottles. That worked fine until the explosion in vaccine demand since the 1980s. “People began to wonder how they could eliminate bottles,” says José Castillo, global director, cell culture technologies at Pall.
Top vaccine makers, including GlaxoSmithKline, Sanofi, Pfizer, and Boehringer Ingelheim, turned to microcarriers, but scaling to manufacturing volumes and achieving reliable, consistent cultures was challenging, Castillo says. Even if some large companies succeeded, for most mid-size and small companies, which lacked sufficient resources, the “transfer of viral processes from roller bottles to microcarriers in a stirred-tank bioreactor was the bottleneck.”
Pall met the challenge by introducing the Integrity® iCELLis Bioreactor, which supplies previously unachievable surface areas per volume—up to 20 m2/L. Cell growth surfaces consisted of sponge-like, polyester microfibers. “A 25 L bioreactor can now provide the same surface area as 6,000 roller bottles,” Castillo tells GEN. “And the cells ‘feel’ like they’re in a bottle.”
A second aspect of scaleup close to Castillo’s expertise involves stem cells used for allogenic cell therapy. Culture and scaleup of stem cells involves one of three systems: microcarriers, 3D scaffolds, or multitrays. Multitrays scale similarly to roller bottles; however, for very large scaleup, they become cumbersome, difficult to move and manipulate. Since every tray is an individual culture unit, the risk of contamination and error in detaching cells from culture surfaces is multiplied. “Plus the process takes a lot of time, and that alone may affect cell quality and ultimately cost,” Castillo explains.
Pall’s solution is a culture system that reproduces conditions within a multitray bioreactor, but which is closed, similar to a bioreactor with shelves. The product is composed of up to 200 layers of rigid polystyrene on which cells settle, adhere, spread, and grow. “The cells experience the same microenvironment as they do in the reference multitray system,” Castillo says.