October 15, 2011 (Vol. 31, No. 18)

Angelo DePalma Ph.D. Writer GEN

Why There’s a Good Deal of Truth to the Dictum: “As Cells Go, So Goeth the Process”

Scaleup trends closely follow the larger bioprocessing industry. According to Christian Kaisermayer, Ph.D., project manager for cell culture at GE Healthcare Life Sciences the two significant drivers are process flexibility and complete solutions.

Underlying both trends are the rise in protein titers over the past decade and disposable bioprocessing.

Disposables play an obvious role in scaleup, i.e., that form factors for smaller single-use cell culture vessels (e.g. GE Healthcare’s Wave Bioreactor™ systems) can be identical to production-scale equipment, thus facilitating scaledown exercises. GE offers Wave systems as “segmented” bags for parallel experiments or screening, with working volumes in each compartment ranging in volume from 50–250 mL. Each chamber is controlled for temperature and headspace gas composition.

“The upstream capacity crunch discussed just a few years ago never occurred,” Dr. Kaisermayer says. “Accordingly, most products in development today will never see 10,000 to 20,000 liter bioreactors. Thus the focus on flexibility. Moreover, the sweet spot transition volume between disposable and stainless steel reactors has already surpassed 1,000 liters.”

With few exceptions today’s production engineers consider disposables at every level of process design, from benchtop to pilot scale and beyond. The more rapid setup and reduced cleaning and related validation associated with single-use bioprocessing improve equipment utilization—not just for final production but also during development.

“Personnel can now concentrate on their experiments rather than spending time on cleaning and sterilization,” explains Dr. Kaisermayer.

On the subject of integration and “complete solutions,” GE Healthcare’s ReadyToProcess™ product line covers all unit operations from media preparation and cell cultivation to protein purification. Customers can source stand-alone or complete single-use technologies for end-to-end processes from GE and even engage their Enterprise Solutions team to design, develop, and execute entire ready-to-use bioproduction facilities.

Moreover GE’s recent announcement that it will acquire media specialist PAA Laboratories further strengthens its upstream process offerings. Experts believe that most improvements in volumetric productivity—a key element in designing process scale—involve media and feed strategies.

Even in situations where development occurs in plastic, and production in stainless steel, the disconnect need not be problematic.

“Transferring a process from disposables to stainless can be controversial in terms of reactor engineering and fluid dynamics,” says Dr. Kaisermayer. “Some experts feel that cells should be shaken, others that cells should be stirred. It usually doesn’t matter because agitation intensity can be adapted to the robustness of the cell line.”

No Impellers Needed

PBS Biotech designs and manufactures single-use bioreactors with a twist. According to Daniel Giroux, vp of R&D, duplicating the agitation mechanisms of small reactors in large systems is problematic without subjecting cells to significant shear damage.

PBS’ U-shaped bioreactors employ a patented mixing technology, the Air-Wheel™, that exploits the fact that cell cultures must be oxygenated. Air-Wheel uses the buoyancy of rising gas bubbles to turn the Air-Wheel structure, thus avoiding the need for external actuation of the stirring mechanism.

The result, according to the company, is rapid, homogeneous mixing and a high mass transfer rate with extremely low shear stress on cells.

PBS offers scalability from three liters for lab work, and 80 liters for preclinical and clinical batches. In targeting full-scale GMP production, the company plans to produce a 500 L system later this year, and a 2,500 L bioreactor in 2012.

“Since the mixing mechanism is proportional to the size of the vessel, Air-Wheel works more efficiently as bioreactor volume increases,” explains Giroux, a definite benefit given that mixing is one of scaleup’s biggest challenges.

According to Giroux, traditional bioreactors were designed for steam sterilization—essentially as pressure vessels. “They were never optimized for mixing. Mammalian cell culture reactors were adapted from chemical mixing tanks, which used shaft-and-impeller mixing.”

As bioreactors grew in size, designers were more or less forced to retain the original form factor and stirring mechanism, including impeller design. “Which is why mixing and mass-transfer problems get worse as the size increases.”

Designers of traditional single-use reactors, he says, never fully appreciated that cleaning was unnecessary. Except for Wave, now part of GE, they mostly retained the shape and mechanical characteristics of traditional tanks. “But it becomes difficult to put impellers into the right locations.”

He claims that at three liters, glass or stainless steel reactors exhibit up to ten times as much shear as the PBS bioreactors. “Our shear goes up a bit at higher scale, but very little.” Ditto for mixing time. Time to 95% mixing is claimed to be less than one minute.

“If you have a tool where critical parameters are so similar across all scales, scaleup will be much easier.”

What about cells for which shear is not a problem? Many such lines exist, Giroux says, but for early-stage development lines, where shear sensitivity is unknown, knowing that shear will not be a problem across production scales is no small comfort.

Scaling Down

A major concern for process scaleup, according to Oren Beske, Ph.D., vp of laboratory services at Aragen Bioscience is having a robust scaled-down model.

“Whether it’s a transient transfection in HEK cells or stable expression in CHO cell lines, a scaledown model helps predict what is likely to occur at larger scale,” says Dr. Beske. “The presumption is, of course, that you will obtain similar cell growth characteristics, productivity, and product quality at large scale as was observed at small scale.”

Aragen is a CRO specializing in molecular biology, cell-line development, process development, and preclinical studies in rodents. Aragen’s scaledown models typically begin at about 20 mL, with the goal of predicting results up to 100 liters, but they occasionally perform work in microtiter formats as well. The factors that are typically controlled and evaluated at the smaller scale are surface area to volume ratio, mixing/agitation rates, feeds and feed schedules, and temperature.

Several vendors offer scaled-down systems, or microbioreactors, that control even more variables. Manufacturers for such systems include Dasgip, Sartorius Stedim Biotech, HexaScreen, New Brunswick Scientific, TAP Biosystems, and Applikon.

“These products are wonderful tools for process development on late-stage clones and, in some cases, earlier clone screening,” Dr. Beske says.

Yet as a CRO, Aragen does not enjoy the luxury of multimonth or even multiweek experimentation.

“Those higher-end technologies are useful when you’re investigating and defining nuances of cell culture processes, to tweak and obtain the highest expression levels. However, many of our customers need two grams in two weeks. We can’t very well ask them to give us six months to come up with an optimized process,” notes Dr. Beske.

But he admits that Aragen is moving toward these types of scaledown technologies. “We’ve entertained acquiring microbioreactor technologies as growth of our cell culture process development business justifies it.”


Wave bags and stirred tank bioreactor based production methods are utilized at Aragen Bioscience for small- to medium-scale protein production. However, appropriate scaled-down model experiments must be performed in order to predict the yield and product quality of these methods.

Stem Cells and Biosimilars

Scaling up stem cell cultures presents a maze of challenges quite different from those facing the expansion of mammalian or insect production cells, bacteria, or yeast. The obvious distinction is that therapeutic stem cells are the product itself, not merely production vehicles.

Moreover the concept of “batch” is turned on its head. “With proposed applications involving one batch per patient in many cases, we’re talking much smaller-volume production runs,” points out Ruedi Stoffel, who heads Roche’s cellular analysis business.

Stoffel compares the step up in complexity from traditional biomanufacturing to stem cell culture to the differences between small molecule drugs and biotherapeutics. Nonbiotech drugs are quality-checked using conventional chemical analysis. Biomolecules are much tougher since their characterization involves much more complex chemistry and biological activity.

“With therapeutic cells, production variability can mean treatment failures. In biomanufacturing you always retain the option of analyzing the product before release, and potentially remediating any production variability inefficiencies,” according to Stoffel.

The most critical steps in biomanufacturing occur during cell-line development and afterward in designing a feed strategy. This is true to some degree for stem cells as well, as investigators must select precursors capable of differentiating into the desired cells, and devise a differentiation strategy based on “feeding” the proper growth factors and cytokines.

But at the comparable point where biomanufacturers stop caring about cells, therapeutic cell producers’ interests peak.

The cell systems and scaleup strategies also differ in terms of economics. Traditional biotech is characterized by high R&D outlays, low manufacturing costs, and high market value. By contrast stem and therapeutic cell production is expensive at all levels and stages, and that is unlikely to change despite advances in manufacturing automation and the use of disposable equipment.

Finally, producers of protein drugs have developed both process- and unit operation-level platform technologies for production and purification. For therapeutic cells, aside from harvest and washing, production is always a custom, labor-intensive activity. “The steps are similar from product to product, but the timing and exact formulation of growth factors will always be different from one type of cell to the next,” says Stoffel.

“Four years ago Roche launched the Cellavista system, a microtiter plate reader that characterizes cells using bright field and fluorescence imaging. This automated technique provides extremely high-quality images of cells. By reading the contents of 96 wells in less than four minutes, Cellavista provides significant time savings compared with conventional microscopy.”

A number of researchers and producers of therapeutic (and protein-producing) cells have adopted this analytical platform to monitor the effects of growth conditions. Most recently, Dutch researchers used it to characterize mouse embryonic stem cell renewal.

Scaleup can be trickier for biosimilars than for originator molecules. “Developers must fit through a very narrow window of similarity,” notes Patrick Daoust, a consultant with Viropro which specializes in biosimilars. The company is working on follow-on versions of rituxumab and several other products.

U.S. regulations on biosimilars are notoriously fuzzy. Various working groups grapple with what constitutes chemical similarity, similarity in biological activity, and the relative importance of the two. The boundary developers of biosimilars do not wish to cross is the point of dissimilarity—again, poorly defined—that will trigger new clinical studies. Here Hatch-Waxman rules (or whatever emerges as the biosimilars guidance documents) will be of little comfort.

With small molecule drugs you either have the chemical structure (and to a lesser extent secondary characteristics and formulation) or you don’t. With biologics, as the process changes so does the product. And that includes secondary characteristics, like post-translational modifications, which may or may not affect the product’s effectiveness or safety.

Since nobody knows for sure, most experts agree that U.S. regulators will cite noticeable changes in such secondary characteristics.

Daoust notes an example of a process that provided high yield but slight differences in glycosylation relative to the comparator molecule. The company decided to forego huge cost-of-goods advantages and stick with the process that was less efficient but more faithful to the original molecule’s chemical composition.

“It’s a tradeoff that all developers of biosimilars will have to face,” he says. Developers have to begin thinking of products from an analytic perspective, and work with process development on analysis from very early on, to avoid surprises during scaleup.” That includes, he says, when the biosimilar has therapeutic efficacy that is superior to the original molecule.


Roche’s Cellavista system is an automated microtiter plate reader that characterizes cells using bright field and fluorescence imaging.

Previous articleOvercoming Downstream Bottlenecks
Next articleTreating Spinocerebellar Ataxia Type 1 Using VEGF Ameliorates Pathology and Improves Motor Function