November 1, 2017 (Vol. 37, No. 19)
Angelo DePalma Ph.D. Writer GEN
Advances in Gene-Editing Techniques Can Help Create Pools of High-Expressing Cell Lines
After selecting an expression species, cell line, and conducting any organism- or gene-level engineering, manufacturers of therapeutic proteins entrust their productivity to media and feeds—the defining factors that nurture the best in cells.
In 2015, world-renowned cell-culture expert Professor Florian Wurm, Dr. rer. nat., of the Swiss Federal Institute of Technology (Lausanne) and cofounder of ExcellGene, told the author that media and feed were responsible for most of the improvements in monoclonal antibody yield in CHO cells.
Since 2014, Dr. Wurm has doubled down on that message. “Media and feed will eventually bring volumetric productivities into the 10–20-g/L range from fed-batch processes. We are still today using just 3–4% of bioreactor volumes for cells, and 20–30% for microbial cultures.” For mammalian cultures, that translates to 15–20 million cells per mL. Dr. Wurm predicts that “we will eventually reach 50 million or more cells per milliliter, keep them viable for two weeks or more, and thus push volumetric yields into those high ranges.”
On the other hand, according to Dr. Wurm, gene editing will have little direct impact on volumetric yield, outside of specific improvements like glycosylation and improved processability. For example, designer proteins, which pose very specific challenges to the endoplasmic reticulum and Golgi processing machinery, might also be addressed eventually by genome engineering.
Generally, CHO cells remain a mystery and will continue to be that way because of their rapidly evolving genetic selections right inside of bioreactors. This will continue to be the reason for the need to individualize developmental efforts for each protein.
Dr. Wurm recently updated his position on CHO cultures through a review1 of the literature on the genomic instability of CHO cells. One take-home lesson from that paper is that the genetic diversity of CHO cells—and how it affects product yields and quality—has been underappreciated.
Vector + Expression = Success
Earlier this year, ATUM (formerly DNA2.0) and Horizon Discovery entered a cross-license agreement for Horizon’s CHO Source platform and ATUM’s vector technology. ATUM will use CHO Source, which includes the glutamine synthetase (GS) knockout CHO-K1 for CHO cells, with its Leap-In® transposase technology for cell-line development services. Horizon now holds an exclusive license to an ATUM vector suite for its CHO Source platform. Together, these technologies will enable expression of complex biologics for both customers’ clients.
“Horizon built strains for biopharmaceutical production, and we have a way to express recombinant proteins through our vectors and transposons,” explains Claes Gustafsson, Ph.D., ATUM’s CEO. “The cross-license gives our customers access to both technologies.”
ATUM’s technology is based on transposons, or jumping genes, discovered by Nobel Laureate Barbara McClintock in maize.
Gene-editing methods like CRISPR are great at removing genes with high precision, but not so good at knocking things in. Therefore, Dr. Gustafsson looked into GenBank, the NIH gene sequence repository that he calls “a genomic candy store” for answers. There he found hundreds of transposons not protected by patent, many with superior gene-inserting capability. “We cleaned up the corresponding transposases and engineered them for robustness and efficiency. The result was our Leap-In technology, which allows [one] to create pools of high-expressing cell lines in about two weeks.”
In complementary fashion, Horizon’s CHO Source increases selection stringency compared with chemical inhibition platforms based on methionine sulphoximine (MSX) and methotrexate, resulting in more rapid identification of clones that express biotherapeutic products.
Several companies have licensed ATUM’s transposase-based tool for multi-gram/L, stable, protein-expressing cell pools.
The CHOZN People
GEN has previously published on MilliporeSigma’s CHOZN® GD−/− cell line, a platform for CHO protein expression based on zinc finger nuclease (ZFN) technology. CHOZN works by eliminating the CHO cells’ endogenous glutamine synthetase (GS). The presence of GS renders the cells incapable of synthesizing L-glutamine, an essential nutrient for cell growth.
Building on CHOZN’s success, MilliporeSigma continues to evaluate and develop new expression technologies, as explained by Andrew Bulpin, Ph.D., head of process solutions. “We are investigating glycoengineering for [the] generation of host-cell lines expressing safer and/or more effective therapeutic proteins. Additionally, we have active projects to reduce cell-line development timelines, while enhancing productivities and clonal stability.”
One of these projects seeks to improve transgene expression levels and long-term stability through gene-regulatory elements and expression-vector design. Another seeks to develop a targeted gene-integration platform enabling transgene expression from well-characterized, stable hot spots within the CHOZN genome.
Even though several gene-editing technologies are available, MilliporeSigma selected ZFN because patent issues complicate the use of TALEN and CRISPR. The use of ZFN is not associated with patent restrictions, allowing the company the option to easily license CHOZN products to other vendors.
“Like all MilliporeSigma ZFNs, those used in CHOZN were manufactured through processes developed with our partner, Sangamo Biosciences, whose ZFN platform has the longest track record in human clinical applications and more safety data than any other genome-editing platform,” Dr. Bulpin tells GEN. “We are additionally exploring opportunities to apply our knowledge of biomanufacturing systems and cell-line engineering to adjacent markets, such as gene therapy and recombinant vaccine expression systems.”
Because of the complexity of protein expression in non-native organisms, and their oft-noted genetic instability, optimizing CHO cells for productivity is an ongoing concern.
Reporting last year in Cell Systems,2 a group at the University of California, San Diego headed by Professor Nathan Lewis, Ph.D., described a genome-scale model of CHO cell metabolism which, through identifying pathways involved in protein production, predicts improvements in protein yields under various conditions. In this study, Dr. Lewis focused on two reported conditions: lower culture temperature (which in some bioprocesses improves folding and secretion) and the addition of sodium butyrate.
“Sodium butyrate changes the chromatin state in CHO cells, thereby increasing protein production, but this effect is not completely understood mechanistically,” Dr. Lewis says. “Butyrate presumably opens up chromatin and allows transcription factors to activate suppressed genes.” Its ability to increase protein production has been has been studied (and described) extensively.
According to Lewis’ model, sodium butyrate and low temperature should boost protein production, but only at the expense of cell growth, which is an unwelcome tradeoff.
“The resources demanded by higher production levels must come from somewhere,” Dr. Lewis continues. “Cell growth competes for resources with protein-making. No matter what you do, you must deal with this tradeoff.” Bioprocess engineers face this conundrum whenever they apply strategies that should improve production. But according to Lewis’s model, these treatments achieve less than 10% of the possible improvement to protein production, “because resources are being used for something else.”
Their model, however, reported much higher improvements, up to 50% of the possible amount, through overexpression of secretory pathway proteins that facilitate protein synthesis in the endoplasmic reticulum. Dr. Lewis describes these folding- and secretion-enhancing genetic changes as manipulations of the cell’s “QA/QC” apparatus. His model predicts that changes positively affecting these cellular activities could increase protein production threefold compared with standard treatments like butyrate addition.
Dr. Lewis is further developing additional resources for CHO cell optimization. For next-generation gene editing in CHO, specifically, Lewis is working on a wide array of engineered cells with “clean” lines, sans undesirable contaminants.
Earlier this year, Selexis launched a novel method for screening and selecting cells with optimized expression levels. The product, SUREselect™, is designed for cell lines expressing difficult-to-express (DTE) proteins.
Selecting stable, high-producing mammalian cell lines is generally a roadblock in biopharmaceutical development, but particularly severe for cells expressing DTEs.
“Finding a high-producing manufacturing cell line after transfection is a numbers game,” says Igor Fisch, Ph.D., CEO of Selexis. After transcribing the recombinant protein gene into RNA, the RNA is translated to a protein, which must be successfully navigate the cell’s secretory pathway. DTE proteins may experience secretion bottlenecks anywhere along the way. After transfection, only a tiny fraction of transfected cells will possess the cellular machinery requisite for commercial manufacturing.
“Traditional selection methods for isolating high-secretion cell lines are somewhat ‘leaky,’ allowing a number of low-to-no-expressing cell lines to pass through the selection process, and making it much harder to find the needle (the high-expressing cell clones) in the haystack (all the non- or low-expressing clones),” Dr. Fisch adds.
Based on company-expression of a vitamin B5 transporter, SUREselect exploits the strict dependence of CHO cells on B5 for energy production, a process that, compared with wild-type cells, is far more selective. As a result, SUREselect clones express robustly. When combined with standard antibiotic selection, the platform uncovers cell clones that are metabolically suited to produce DTE proteins.
“We achieve a fourfold increase in production for a range of difficult-to-express proteins, including some molecules that were toxic to CHO expression cells,” says Dr. Fisch, continuing his analogy. “By exerting strong selection pressure, the haystack becomes smaller and the needles more easily identified.”
The same selection advantage is attainable for not-so-difficult-to-express proteins and, in these cases, the company claims they can improve productivity for some mAbs to 9 g/L.
Selexis has used SUREselect to produce an IFN-beta-producing cell line that was inaccessible due to the molecule’s cytotoxicity, and at levels that are five times higher than what is typically achieved through antibiotic selection alone.
Lonza has continuously improved its GS Gene Expression System® since the platform’s debut in 1992. The most recent version of which is the company’s GS Xceed® Gene Expression System. It generates high titers in a chemically defined, animal component–free system that does not require MSX selection. The GS Xceed platform typically achieves titers of up to 6g/L using Lonza’s proprietary media and feed system and exceptionally 10 grams/liter have been reported.
The platform uses the proprietary CHOK1SV GS-KO™ (GS KnockOut™) host cell line. This host cell is derived from Lonza’s suspension-adapted CHOK1SV™ line from which both alleles for GS have been knocked out, thus requiring supplemental glutamine.
Andy Racher, Ph.D., associate director for future technologies at Lonza, says that optimization strategies based on glutamine (and for that matter methotrexate) selection “may not have reached the boundaries of what is possible in terms of productivity.”
Thirty-five licensed protein therapeutics are manufactured using Lonza’s GS System® and about 500 molecules currently in clinical trials use the GS System.
“Depending on the product type (e.g., bispecifics),” Dr. Racher says. “You’ll sometimes see large differences in product titers. But at least you can get enough protein for tox studies, and have a line that may be further optimized.” The advent of gene editing, including knockins, will eventually provide greater understanding of CHO pathways and how to manipulate them. “It may become a lot easier to create new CHO hosts capable of producing next-generation biomolecules. I’m not saying it will be a trivial exercise, but it should be a lot easier than it was in the past.”
1. F.M. Wurm and M.J. Wurm, “Cloning of CHO Cells, Productivity, and Genetic Stability—A Discussion,” Processes 5(2), 20 (2017).
2. H. Hefzi et al., “A Consensus Genome-Scale Reconstruction of Chinese Hamster Ovary Cell Metabolism, Cell Sys. 3(5), 434–443 (2016).