November 1, 2016 (Vol. 36, No. 19)
Angelo DePalma Ph.D. Writer GEN
As Demand for Therapeutic Proteins Rises, Bioprocessors Have a Choice: Try to Do More of The Same, Or Try to Do Better
The global cell-line development market will keep growing, says MarketsandMarkets, which projects that cell-line development products—equipment, media, and reagents—will reach $4.0 billion in value by 2019, up from $2.2 billion in 2014, reflecting a compound annual growth rate of about 12.5%. This growth was attributed to increasing demand for monoclonal antibodies, rising vaccine production, and widening use of innovative technologies to develop cell lines.
Close to 50 monoclonal antibodies have been approved in Europe and the United States. The National Center for Biotechnology Information predicts that as many as 70 monoclonal antibodies will be on the market by 2020 (not counting follow-on products).
Besides interest in monoclonal antibodies, there is confidence that derivative products—biosimilars, antibody-drug-conjugates, and biobetters—will expedite cell-line improvements. The emergence of these derivatives underscores the need for optimization, particularly for strains that, unlike Chinese hamster ovary (CHO) cells, do not yet exist in optimized forms.
The cell-line development market is growing in size and advancing in sophistication to outfit several application areas: tissue engineering and regenerative medicine, toxicity testing, research, and discovery, and last but not least, bioproduction. For example, cell-line development technology is being used to enhance platforms that produce biotherapeutics such as bispecific antibodies.
In late 2015, Zymeworks enlisted ProBioGen for developing cell lines suitable for expression of Zymeworks’ bispecific antibody candidate. The companies arranged for Zymeworks’ Azymetric™ platform to incorporate ProBioGen’s GlymaxX® technology, which prevents fucosylation to N-linked glycans, which is believed to enhance ADCC. ProBioGen made use of its CHO expression system, which ProBioGen chief scientist Volker Sandig M.D., Ph.D., asserts is “ideally suited for the expression of biospecific antibodies with high purities and expression titers.”
Since bispecific antibodies bind to two targets simultaneously, any expression system for these molecules has a lot of work to do. Through Zymeworks’ Azymetric platform, bispecifics assemble spontaneously into a single molecule with two different Fab domains. The process, which allows manufacturing to proceed by means of conventional antibody processing, is easily adapted to rapidly screen target and sequence combinations.
Traditional Approach
Abzena provides development and manufacturing services—such as immunogenicity assessment, drug design, and antibody humanization—and takes a traditional approach to the development of cells for biomanufacturing. “We first confirm that we have the right molecule,” explains Simon Keen, Abzena’s group leader for cell-line development. “Then we generate constructs from our proprietary vector system and transfect cell lines of choice, usually CHO, NS0, or SP2/0.”
Keen’s group then uses semiautomated screening that selects high-producing cells that generate proteins with desirable critical quality attributes. Thanks to biosimilars, the NS0 and SP2/0 cell lines and other mouse myeloma lineages are once again popular expression systems. Many legacy products now coming off patent were produced in these lines, which differ in how they generate post-translational modifications, particularly glycosylation.
“Developers of biosimilars must therefore consider their choice of host cell,” Keen tells GEN. “To match a biosimilar to a protein that is already made in SP2/0, you can significantly de-risk the project by continuing to use SP2/0, as opposed to using CHO, even though CHO has seen huge advances. Mouse myeloma cells have a great deal of catching up to do with respect to more popular cell lines. However, manufacturers considering quality vs. productivity often trade quality for yield, and they act on this preference at the very earliest stages.”
The diminishing returns of chasing yield over quality becomes apparent at very high titers, where Keen says product heterogeneity becomes problematic. “Moreover, downstream purification has to deal with the thick, gloopy reactor harvest,” Keen notes. “That we’re close to the limit of what downstream can handle is just another reason to focus on quality.”
Developers of protein therapeutics therefore increasingly turn to engineering to incorporate qualities that cells currently lack such as specific glycosylation. In one project for example, Abzena co-transfected the sialyl transferase gene in addition to the recombinant protein sequence to raise levels of terminal sialylation to improve protein pharmacokinetics.
Continuous Improvement
MilliporeSigma’s involvement in cell-line engineering can be traced back to TargeTron, a bacterial gene-editing technology launched in 2004 by Sigma-Aldrich, which was acquired by MilliporeSigma in 2015. TargeTron, says MilliporeSigma, was the first application of genome editing to improve cell lines. MilliporeSigma is also realizing benefits of another Sigma-Aldrich initiative, the purchase in 2006 of exclusive rights for zinc finger nuclease (ZFN) technology for manipulating mammalian cell lines.
The ZFN technology enabled Sigma-Aldrich to explore genetic modifications to improve productivity, safety, and usability of host cell lines. Eventually, these efforts led to the CHOZN® platform host cell line, which is suitable for rapid, robust production of biologics.
CHOZN began with a CHO cell line adapted to robust growth in chemically defined media, and engineered through ZFN editing to lack a functional copy of the glutamine synthetase (GS) gene. “That provided a very stringent selection mechanism whereby cells would not grow unless they received the GS gene in addition to the gene coding for the protein of interest,” says Kevin Kayser, Ph.D., senior director of upstream development at MilliporeSigma. The cell line and associated media and feeds are used to isolate clonal cell lines producing a biologic of interest.
Since it was launched in 2012, CHOZN has incorporated several improvements. Most recently, it gained an enhanced ability to resist viral infections. “Minute Virus of Mouse (MMV) has been responsible for multiple contaminations,” notes Dr. Kayser. “It is one of the few viruses that propagates in chemically defined media.” CHO cells are particularly vulnerable.
MilliporeSigma identified a genetic modification that blocks viral uptake into the cells, rendering the modified cells impervious to the threat of the virus. The ZFNs targeting this gene are available as part of the Centinel™ Intelligent Virus Defense system, which goes commercial in 2017.
ZFNs can generate cells that produce more human-like therapeutic glycoproteins. Because genes that modulate post-translational modifications can differ between CHO and human cells, CHO-produced proteins may be recognized as foreign in patients, an eventuality that gives rise to immunogenic responses or reductions in protein half-life. Two genes that lead to these nonhuman modifications have been removed via ZFN-mediated engineering, creating a cell line that produces glycoproteins with more desirable glycosylation patterns.
MilliporeSigma employs ZFN technology to various other host cell lines through specific projects on behalf of customers. In these projects, modifications can be made to fix problems specific to an already established host cell line such as the removal of an endogenous protein that co-purifies with the biologic of interest.
Scientists at MilliporeSigma’s laboratory in St. Louis use RNA interference, zinc finger nuclease, and CRISPR technologies to develop improved CHO expression systems.
Doing What CHO Cannot Do
The productivity of CHO-based biomanufacturing for recombinant proteins has risen almost 1,000-fold, according to Jorg Thommes, Ph.D., senior vice president for technical development at Biogen. “Current CHO manufacturing,” he states, “can currently serve patient populations in the indications we currently treat.”
Enlarging the physical footprints of existing manufacturing facilities could serve the needs of expanded patient populations. Simply doing more of the same, however, may not be the best way to boost output. “To expand to much larger patient populations, from hundreds of thousands for one indication to tens of millions, exploring alternatives to conventional antibody manufacturing is worthwhile,” suggests Dr. Thommes.
The same philosophy applies to expanding access to biopharmaceuticals to worldwide markets. “We do not know whether alternate systems really offer a step change in productivity over today’s CHO potential,” admits Dr. Thommes. “But in the spirit of investing into the future of biotherapeutics, the Bill & Melinda Gates Foundation and Biogen found this an important question to answer.”
Dr. Thommes refers to the August 16, 2016, announcement that the Gates Foundation was partly funding Biogen’s search for alternative expression systems. Biogen hopes to “identify new methods to significantly increase the amount of antibody-based therapies that can be produced compared to today’s methods.”
Dr. Thommes acknowledges that the project may prove current CHO-based production to be the best after all, but he is also open to other possibilities: “What if we could use yeast, fungi, or algae as more productive expression systems and still produce high-quality proteins in much larger quantities without expanding facilities? We believe this may be possible, and could unlock even greater potential to improve human health worldwide.”
Biogen plans to test expression of a specific antibody head-to-head in CHO against eight alternative expression systems. “We have substantial laboratory-, pilot-, and manufacturing-scale experience with this molecule,” indicates Dr. Thommes. “So it will be a true direct comparison with conventional manufacturing.”
Initially, Biogen will emphasize cellular engineering; later, the company will employ upstream and downstream process development. Ultimately, Biogen intends to improve both yield and quality. The company, Dr. Thommes anticipates, will release results in 12 to 18 months.
In August 2016, Biogen entered a partnership with Amyris to use that company’s Automated Strain Engineering (ASE) platform to generate alternatives to mammalian cell expression systems for therapeutic proteins. Joel Cherry, Ph.D., president of R&D at Amyris, uses the term “tractability” to describe a broad-based evaluation of microorganisms for the yield, quality, and therapeutic efficacy of their protein products. He was not at liberty to disclose which organisms were under consideration.
Cell-line optimization occurs mainly through manipulating media, feeds, and process conditions; adaptation followed by selection; or genetic engineering. Amyris’s approach relies most strongly on engineering.
“ASE was first developed in yeast for constructing and inserting DNA into the yeast genome to optimize small molecule production” Dr. Cherry explains. The technique can engineer thousands of yeast strains in three weeks with DNA introduced at multiple different loci simultaneously. “It will now be used to optimize therapeutic protein production in commercially robust cell lines.”
Not a Panacea
A Bayer biotechnology plant in Berkeley, CA, that has produced recombinant Factor VIII since 1993, recently received approval from the U.S. Food and Drug Administration for a new, improved Factor VIII process. The company expects to take this process live in March 2017.
Factor VIII is an extremely large, complex protein with global sales rivaling its molecular weight (if one may make such comparisons). “It’s the largest and most glycosylated commercial recombinant protein,” says Paul Wu, Ph.D., director of upstream development at Bayer Biological Development. Most glycosylated antibodies have one to two N-linked glycosylation sites. Factor VIII has 25 N-linked and 10 O-linked glycans.
Bayer uses a baby hamster kidney (BHK) cell line to produce Factor VIII through a perfusion cell culture process. To assist with proper protein folding and improve cell viability, Dr. Wu’s group has introduced a heat shock protein along with the gene for Factor VIII.
“Production of heat shock proteins rises when cells undergo heat or other stresses,” Wu explains. “The heat shock proteins facilitate protein folding and reduce apoptosis. BHK cells don’t normally express this protein, or produce it at very low quantities.”
Compared with the size of Factor VIII, heat shock protein is tiny, so the additional burden on expression cells is not that great, according to Dr. Wu.
The new approach, which not only incorporates heat shock protein but also benefits from process improvements, has significantly increased cell-line productivity. “We also saw pharmacokinetic improvements in the final product, as determined in clinical studies, including data from the LEOPOLD (Long-Term Efficacy Open-Label Program in Severe Hemophilia A Disease) clinical trials,” asserts Dr. Wu.
Nevertheless, Dr. Wu cautions against extrapolating Bayer’s promising results to other proteins: “We know that academic groups are interested in heat shock protein. However, for most therapeutic proteins, which are much simpler than Factor VIII, this type of enhancement may not needed.”
At Bayer’s manufacturing facility in Berkeley, CA, technicians inspect vials of Kogenate FS®, the company’s rFVIII treatment for hemophilia. Before they can approve the product for packaging, the technicians must confirm that it is free of potential defects, more than 25 of them, and do so in 12 seconds or less.
