Send to printer »

Feature Articles : Nov 15, 2012 (Vol. 32, No. 20)

Optimizing Cell Lines Improves Production—Really!

  • Angelo DePalma, Ph.D.

Of all the activities that constitute a bioprocessing operation, a good argument could be made that the most critical step involves optimizing cell-line development.

That was one of the main messages delivered at CHI’s “Bioprocessing Summit”, in the late summer.

Jesús Zurdo, Ph.D., head of innovation for biopharma development at Lonza Biologics, discussed the challenges and advantages of various strategies for minimizing high attrition rates related to “dwindling R&D productivity and spiraling development costs,” and the resulting cost pressures on process development for microbial fermentation and mammalian cell culture.

According to Dr. Zurdo’s model, risk may enter the picture anywhere during development, including preclinical studies, human testing, or manufacturing.

“Because we get involved early in the development process, we are often the first to notice issues in the form of low yield, aggregation, low chemical stability, immunogenicity, and immunotoxicology,” he said.

Dr. Zurdo’s approach to assessing “developability” involves in silico computational methods to predict productivity, aggregation, stability, and immunogenicity. This helps investigators select, from a collection of potential candidates, a molecule optimized with respect to these properties.

Next he examines, through in vitro and ex vivo assays on cultured human cells, immunogenicity-related events such as T- and B-cell activation and cytokine secretion. The objective of this two-pronged is not to characterize the molecule fully, but to select optimal candidates with low risk of stability and immunogenicity issues.

“Once you are comfortable with biological activity, you can select the lead molecule based on all these assays,” he added.

Computational Capabilities

Most biopharmaceutical developers employ some of these activities already.

“They attempt to bring research and development teams closer together to differing degrees. But our computational approach, which reduces cost while increasing throughput, allows investigators to explore things they could not otherwise explore.”

Lonza’s other innovation consists of integrating stability and immunogenicity studies. “Only a few companies have achieved this, and it’s unheard of in the CMO space,” claimed Dr. Zurdo.

Identifying and analyzing for key quality attributes early in development, rather than trying to solve them through process tweaks, requires a higher initial commitment but its reward is lower risk and, ultimately, greater flexibility in designing the eventual manufacturing process.

Delivery and formulation are two other factors that Dr. Zurdo believes companies should be thinking about earlier rather than later. The later these critical development activities begin, the more constrained the development space will be with respect to concentration and administration route.

Biopharm is perhaps the only industry, according to Dr. Zurdo, that generates prototypes—candidate molecules—through a fully developed commercial manufacturing process. “And this affects development cost and time,” explained Dr. Zurdo.

Industry must continue to seek ways to create “prototypes” that are safe (through early developability and immunogenicity testing), but which also facilitate the rapid transitioning of candidate molecules into the clinic. Only after a molecule has shown promise in early-stage clinical phases should the all-out effort be placed on a Phase III or production-worthy process.

Parallel, Multimolecule Strategy

Through her presentation, Marguerite Campbell, a scientist in biologics research at Janssen Research & Development, unveiled a unique approach to reducing biotherapeutic attrition. According to Campbell, molecules continue to fail during preclinical development despite great strides in molecule and cell development. Her company’s technique for minimizing early-stage attrition involves front-loading early development by introducing as many as five different cell lines that express an equal number of candidate molecules.

This parallel-development approach requires analytical, purification, and formulation to become involved early in the process on multiple molecules.

“Once the preclinical enabling stage begins we continue to engage all stakeholders,” Campbell said. “Their involvement throughout development allows us to create shorter learning cycles, with go/no-go decision points as we build molecules and cell lines to fit our manufacturing platform.”

The objective is to test and learn as much as possible about multiple candidates before committing to one molecule. “It’s all about options.”

According to Campbell the parallel approach is less risky than committing to five different projects for five different therapeutic targets or indications. She presented a case study of an oncology drug. At initiation, her group had two basic molecules with differing mechanisms of action: one, thought to be the most promising, with two variants and a back-up protein with one variant.

The first group of molecules failed due to aggregation problems, which under a more traditional development paradigm might have ended the project then and there. The second variant of the backup molecule did not exhibit the aggregation problem and showed high activity and the potential for market differentiation.

So in the end, the molecule that may have been the fifth-most-promising was the one that moved forward.

Designing for PTMs

Anne B. Tolstrup, Ph.D., director for cell culture II process science at Boehringer Ingelheim Pharma, spoke on tailoring product quality through cell-line development and process optimization.

Quality attributes vary depending on the molecule’s desired effector function. For example, anticancer antibodies should exhibit high antibody-dependent cellular cytotoxicity (ADCC), while biosimilars must closely match the physico-chemical profile and activity of originator molecules.

Dr. Tolstrup’s production cell, the proprietary Bi-Hex® (Boehringer Ingelheim high-expression) platform, consists of DG44 CHO cells and other elements enabling selection of cells for desired quality attributes.

Bi-Hex cells include two subclones that differ with respect to the number of galactose residues present in the N-linked sugar on Asn297 in the constant domain of the heavy chain. By manipulating and optimizing these characteristics and others related to glycosylation, Dr. Tolstrup can finetune safety and efficacy-related properties, like ADCC and complement-dependent cytotoxicity for anticancer biologics.

For biosimilars, the goal is to replicate the originator efficacy and safety while preserving physico-chemical properties.

In addition to the Bi-Hex subclones varying in glycosylation pattern, Boehringer Ingelheim has in-licensed GlymaxX®, a glycosylation-optimization technology, from ProBioGen. Glymax enhances ADCC and is suitable for development of biobetters and novel antibodies with enhanced anticancer activity, according to Dr. Tolstrup.

“There’s a significant opportunity to finetune attributes like ADCC through process and media development as well,” she noted. “After clone selection, we typically do a round of media optimization, where we adjust parameters like the concentrations of essential nutrients in the medium and adjust the feeds.

“Then, in a second round of DOE, we fine-tune process parameters like pH, gassing, temperature, and agitation. We’ve shown examples where ADCC can vary from 40 percent to almost 400 percent for the same clone taken through different media or process-development schemes.”

Bi-Hex integrates novel vectors, serum-free transfection, optimized host cells, serum-free flow-based screening, and chemically defined production media. Early assessment of phenotypic stability and product quality attributes is done to generate cells that are stable and highly productive in fed-batch cultures.

Bi-Hex also involves the traditional exercises of optimization of cells, media, and process, but uses a highly integrated and interdiciplinary approach and relies heavily on automation as well as on design-of-experiment approaches, continued Dr. Tolstrup.

Transient Expression in CHO

James Brady, Ph.D., director of technical applications at MaxCyte, discussed the potential for fully scalable transient gene expression in CHO cells.

CHO cells’ scalability, adaptability, and human-compatible glycosylation has made them the workhorse cells for monoclonal antibody expression, but they are difficult to transfect transiently. Transient transfection’s advantages are rapid expression of gram quantities of protein for characterization and preclinical studies.

Conventional transient expression occurs in HEK cells, which are more amenable than CHO to lipid-based transfection methods. However, expression in CHO is highly desirable to maintain consistency between early- and late-stage development.

Dr. Brady demonstrated that a flow electroporation system, MaxCyte® STX™, can transfect 1010 cells with 95% efficiency and viability in a single run. “With our system, customers can do all their up-front work, and characterization studies, in the same cell background that they will ultimately use in biomanufacturing,” Dr. Brady noted. “They can do everything in CHO cells.”

This raises the question of potential differences between the products of transient and stable transfection, even within the same expression system.

“A number of our customers have compared proteins they’ve produced with transient expression, with our technology, to material from a stable cell line, and found no detectable differences in intactness, aggregation, charge, or post-translational modifications,” he pointed out.

MaxCyte licensed electroporation technology 14 years ago from Harvard University, where it was used to research experimental treatments for blood disorders. MaxCyte STX employs electrical pulses to create temporary openings in cells, through which the plasmid enters and performs its magic in the nucleus.

The MaxCyte instrument is sold with protocols to enhance the large-scale transfection of more than 50 cell types, including HEX and cells that are refractory to transient transfection.

Case for Insect Cells

Donald L. Jarvis, Ph.D., professor of molecular biology at the University of Wyoming, made a good case for the baculovirus-insect cell expression system to produce proteins with eukaryotic post-translational modifications, particularly glycosylation.

Insect cell lines, unfortunately, express relatively primitive glycosylation pathways that lack sialylation. Sialic acid groups impart useful pharmacokinetic properties on therapeutic proteins. “Molecules require sialic acid residues to survive the human circulatory system,” explained Dr. Jarvis.

His approach was to isolate genes for the missing glycosylation functions and insert them into insect cells, a technique he calls “strain engineering for glycosylation” or “glycoengineering.”

Ironically this approach is easier in practice than in theory.

“There are many steps in glycosylation pathways. When first thought of incorporating all those genes, of knocking in multiple functions, it seemed mind-boggling,” continued Dr. Jarvis.

That perceived hurdle delayed efforts to glycoengineer insect cells for several years, he concedes. “But when we tried it, we found it to be straightforward.”

His group has efficiently inserted as many as nine unlinked markers into insect cells by simple transfection and co-selection.

“We don’t have to screen more than 40 or 50 clones to find those that express all of these unlinked markers. Insect cell genomes just seem to soak up plasmid DNA,” he said.

In theory Dr. Jarvis could simultaneously transfect for product proteins as well, but he prefers to achieve that with standard baculovirus agents.

When asked “Why not simply use CHO cells, which already sialylate and are known to be robust expression systems,” Dr. Jarvis admits feeling “backed into a corner” when forced into a comparative analysis of expression systems.

“You need a variety of different expression systems in biotech because proteins will be expressed more successfully in one system than another. Biotechnology companies always talk about having four or five platforms in place—E. coli, baculovirus, Pichia, and mammalian cells—to cover their bases.”

Insect cells also provide a safety factor, in that adventitious insect viruses would not infect humans.

Perhaps more controversial is Dr. Jarvis’ contention that on a unit volume and time basis, insect cell culture is less expensive than CHO, and the regulatory issues have become moot. Two approved drugs produced in insect cells have been approved in the U.S: the human papillomavirus vaccine Cerevex, and the prostate cancer immunomodulatory agent Provenge.