June 15, 2009 (Vol. 29, No. 12)
Angelo DePalma Ph.D. Writer GEN
Both Factors Need to Be Maximized and Balanced to Achieve the Best Results
Optimizing mammalian cell culture for volumetric productivity involves working on different levels at different stages of the product’s lifecycle: cell line genetic engineering, clone selection, nutrition (media, supplements, feed), and the process itself are all reasonable places to intervene. The first two activities fall under the category of nature since they circumscribe innate capabilities, whereas the latter involve nurture.
Ten years ago one might have described optimization efforts as holistic since the goal of cell optimization is overall improvement, not necessarily achieving the highest specific productivity per cell (through nature) or cell density (nurture). Developers often find that they must play one factor against others to achieve the right mix. “For constitutive expression systems, devoting resources toward expression may divert them from cell growth, which can result in overall lower productivity,” says James Blackwell, Ph.D., senior consultant with Bioprocess Technology Consultants. Similarly, lowering the temperature of a cell culture from optimal may slow cell division but improve volumetric productivity by directing cells toward production rather than reproduction.
Thomas Seewoester, Ph.D., director of process development at Amgen, notes that cell culture process yields are defined by total viable cell mass and cell-specific productivity. “To optimize a process, both factors need to be maximized and sometimes balanced against each other,” he says.
Large companies like Amgen possess considerable in-house optimization skills to cover cell engineering, clone selection, media and feed compositions, and process parameters. Its experience with a large number of similar molecules has enabled Amgen to establish process platforms that avoid extensive optimization at each level and each molecule early in process development. Later on, each product receives targeted media- and process-level optimization to achieve productive, cost-effective, robust processes. Still, the company works with external collaborators and partners to explore new technologies and novel approaches to cell process optimization.
Cells vs. Microorganisms
Nurture tends to provide greater benefits compared with nature for mammalian cells, while the opposite is true for microbial cultures. Thus, for mammalian cells, increasing biomass is the key to improving volumetric productivity, while for yeast the answer is genetic engineering. “It is possible to engineer characteristics into yeast in ways that are not possible with mammalian cells,” notes David Robinson, Ph.D., vp for bioprocess R&D at Merck Research Laboratories.
Yeast possess much simpler genomes than do mammalian cells—approximately 6,000 genes versus 20,000. “Plus, the number of tools to do genetics in yeast is higher,” Dr. Robinson adds. Merck owns the rights to yeast-based protein expression technology based on Pichia pastoris through its acquisition of GlycoFi in 2006.
Merck is excited about Pichia because this expression system allows developers to control glycosylation patterns in proteins where this is significant. Glycosylation and other post-translational modifications are among the crucial issues facing developers of biosimilars. “Control over glycosylation provides us with the ability to do product engineering as opposed to simply productivity enhancements,” says Dr. Robinson, who uses the term “bio-better” to describe proteins whose glycosylation patterns provide superior properties such as circulating half-life, tissue targeting, and biological activity. “It has the potential for therapeutic protein development that has never before existed.”
Merck also employs Lonza Biologics’ glutamine synthetase technology, GS System™, which uses a viral promoter and selection via glutamine metabolism to provide stable, high-yielding mammalian cell lines. GS System is used by more than 85 biotechnology companies. Five approved products, including Zenapax® (Roche) and Synagis® (MedImmune), use the technology, which allows developers to reach the 50 pg/cell/day production level, according to Lonza Biologics. From there, Merck scientists increase biomass through standard fed-batch techniques.
According to Dr. Robinson, 50 pg/cell/day is a kind of soft practical upper limit for cultured monoclonal antibody-producing cells. Some produce more, but those tend to be outliers. Dr. Blackwell notes that secreted antibody levels from B cells in response to a pathogen challenge, on the order of 100 pg/cell/day, may suggest a true biological ceiling for antibody production in mammalian cells.
A Good Pedigree
While nurture has been responsible for most of the recent improvements in volumetric productivity, starting with a high-producing cell line is of utmost importance. Poorly producing cells can be improved 100-fold and still not be up to production standards. A high producer for which a three- or fourfold improvement is achieved through media/feed/supplement strategies, and another doubling or so during process development, can make a huge difference in subsequent cost of goods and capital expenditures.
“We have never seen a cell line that makes less than 100 mg/L be improved to the grams per liter range using process optimization,” says Simin Zaidi, manager of process development at Avid Bioservices, “especially given schedule restraints most projects have.”
She cites as an example a CHO cell line producing an IgG at 70 mg/L that was improved to 350 mg/L through an extensive optimization effort. Another cell line that began at 400 mg/L with process optimization easily reached 1g/L with only minimal work. “And there is a good chance that this could be improved even further with additional process-development work.”
In addition to production services, Avid provides optimization competencies that fall under both nature and nurture categories. Avid’s cell-line development group optimizes cells, then passes them off to process development, which “works magic with media and feeds,” according to cell line development manager Jeanette Doerr, Ph.D.
Like most early-stage development services companies, Avid uses the standard limiting dilution method for generating clones. Recently the company adopted the ClonePix™ FL cell screening and selection platform from Genetix. Where the dilution method takes about eight labor-intensive weeks to screen 1,000 clones, ClonePix processes 10 times as many cells in about three weeks, according to the company. The instrument provides a visual image, based on fluorescence detection, of the highest-producing colonies, and ultimately individual cells. ClonePix also confirms that a colony is truly clonal, whereas limiting dilution requires examining every single cell.
Another interesting cytometry system for clone and cell line selection is LEAP, a microplate-based cytometry system from Cyntellect that CEO Fred Koller, Ph.D., says satisfies requirements for both nature and nurture. LEAP identifies the highest-producing cells from either genetically heterogeneous collections of cells or, under varying culture conditions, for a stable clone.
LEAP generates feature-rich images of cells in a closed environment, which flow cytometry and bead-based selection methods do not, Dr. Koller says. Instead of picking the most promising cells, LEAP leaves them in place and obliterates the undesirable cells through laser irradiation. This assures that colonies arise from a single cell.
“The cells you choose haven’t been taken anywhere or otherwise manipulated,” notes Dr. Koller. “They are unaware of what has happened to neighboring cells.” Excessive manipulation, he says, can change a cell’s gene expression, particularly for adherent cells. “It’s particularly important not to disturb cells during measurement periods.” Once cells are selected, they may be removed and cultured within larger vessels.
Nutritional and Process Strategies
Once a cell line has been engineered and/or set by cloning, nature’s input ends; the nurture part begins in earnest immediately and continues at some level for the remainder of the product’s life cycle.
On the process side of the continuum, Avid achieves yield improvements by optimizing physical setpoints such as pH, osmolality, and dissolved oxygen in bioreactors while media and feed strategy experiments are carried out at small scale.
Avid is familiar with most commercial media, but sometimes clients ask the company to explore specific media and feeds that worked well in early development. Zaidi calls such projects “informed scouting.” Her group performs small-scale experiments and applies software statistical packages to assure that the project proceeds “in the right direction.”
Numerous companies now offer media and feed development services. One of the more comprehensive is BD Biosciences’ BD AutoNutrientSM Media Design Service (AMDS).
AMDS optimizes chemically defined media and supplementation ingredients for cells that have already been optimized for innate protein production. An AMDS project takes between 3 and 12 months.
Although BD does not perform cell-line engineering, it recognizes the importance of nature in achieving optimal productivity. “We receive all types of cell lines, both low- and high-producers,” says Jon Wannlund, Ph.D., R&D director. Two daughter lines, selected from different clones, can have radically different productivity profiles. With some, no amount of nutritional supplementation or feeding strategies can improve productivity. But many, particularly those that express above approximately 1 µg/mL, can, through media strategies, become high producers.
In one AMDS project two years ago, nutritional optimization provided a 1.7-fold increase in IgG production; peptone supplementation improved this by another 2.1-fold, while feed strategies improved productivity by an additional 30%, for an overall 4.5-fold increase in productivity. Another project saw a fourfold improvement from media and supplementation alone (no feed work was done on these cells), to about 3.5 g/L. The client upped productivity to 6 g/L through process optimization.
But as the advertisement says, “mileage may vary.” The highest-producing cells tend to benefit most from nurturing strategies, but some lines do not respond at all, and those showing the greatest improvement or the highest productivity in the end may not be those with the highest output before media/feed strategies are applied.
A standard approach to the in-process nurture strategies for cell-line optimization is to monitor fermentation for a relatively narrow panel of analytes such as pH or dissolved oxygen. This technique is related to process analytics.
Similarly, metabolomics—the emerging omics discipline that quantifies metabolites and other small molecules—may serve as an indirect measure (as opposed to measuring protein concentration) of cell health and productivity. “Metabolomics tells you which nutrients are being used, which metabolites are getting into the cell, and what types of waste products are being produced,” notes Denise Sonntag, Ph.D., senior scientist at Biocrates Life Sciences.
The metabolome of a mammalian cell is typically large—upward of 3,000 compounds. Specialists like to mention that this is far fewer than the number of relevant genes or proteins, which is true, but concentration-dynamic ranges of small molecules of interest can be huge, a fact that complicates quantification. Nevertheless, the analysis can focus on the disposition of specific nutrients and cofactors, or take in a subset of biogenic amines, amino acids (and their metabolites), lipids, fatty acids, or energy-related metabolites. Profiles of a predefined, select subset of the metabolome give a precise snapshot of what is occurring in the cell at a specific time, for example during growth or expression phases.
Considering that cell-based biomanufacturing seeks to produce proteins that are foreign to the expression systems, in bioreactors instead of intact organisms, it is no wonder so much effort is expended on optimization. One may argue the relative merits of hard-wiring cells for productivity, or coaxing them through media/feed strategies later on, but it is clear that success depends on both strategies.
“Nature and nurture need to come together,” opines Dr. Seewoester. But there are limits to what may be achieved with poorly producing cells. “Without natural expression capabilities, nurturing through media and process conditions will be limited.” On the other hand, every cell can be improved through selection and engineering. “We have not seen one example of a wild-type host cell in the mammalian or microbial world that can’t be effectively and significantly improved through targeted optimization on the genetic or physiological level,” he says.