Gene Editing Tackles Cell-Line Development

Exploiting the wide-ranging utility of cells in medical diagnostics, drug discovery, biomanufacturing, and basic science requires having the right cells. Cell-line development (CLD) is the process by which desirable phenotypes are created from cells that exhibit those properties in attenuated form, or not at all.

Traditional CLD involved the selection of cells tending, with each passage, toward a desirable trait or phenotype, with or without the addition of evolutionary pressures like the growth factors or culture conditions. Refinements in non-editing techniques, like single-cell cloning, gene or transgene overexpression, cell-line adaptation, differentiation, and immortalization, advanced the field to a degree and are still used occasionally, despite limitations in efficiency and scope, and high resource utilization.

Single-cell derivation methods, for example, involve establishing clonal populations, looking for desirable phenotypes, and repeating the process until cells with more desirable characteristics emerge. Overexpression methods involving the insertion of foreign genes are rapid but express genes in “always on” mode, without typical checks on activation.

Manipulation of cells destined for monoclonal antibody (mAb) production constitutes, arguably, the highest value-added sector for commercial cell lines. Hybridoma-based CLD approaches for this application often produced unstable cell lines, and the mAb products themselves induced human anti-mouse antibodies.

CRISPR, the “consensus” editing tool

Older gene-editing methods, including mutagenic screening and classical gene targeting, are still used for some applications but have mostly given way to modern tools like zinc finger nucleases (ZFN), transcription activator-like effector nucleases (TALENs), and clustered regularly interspaced short palindromic repeats (CRISPR).

ZFNs and TALENs were considered revolutionary at the time “but both are expensive to develop, difficult to handle, are time-consuming, and lack the necessary specificity,” says Daniel Gibson, chief technology officer at Codex, an artificial DNA company. “These drawbacks limit their widespread use, particularly for large scale, high-throughput studies.” The methods also suffer from a high incidence of nonspecific DNA cleavage and mosaicism, where a mutant allele is produced in only some of the cells.

CRISPR, which has overtaken TALEN and ZFN technology in recent years, is relatively straightforward but not without its challenges, says Elizabeth Turner Gillies Ph.D., Cell Systems Scientist at ATCC. “CRISPR requires introducing Cas9 into the cell along with a gene-targeting guide RNA, followed by validation of the enzyme’s gene-editing functionality.” To streamline the process, ATCC offers stable cell lines that constitutively express Cas9 nuclease. Recently, they have used CRISPR-Cas9 to improve the viral production capacity of Vero and MDCK cell lines, two workhorse lines for viral vaccine manufacturing.

Gene editing works best with cellular functions where a direct connection between a specific genotype and phenotype is known or can be discovered using a genetic screen. “If you know how a gene controls a function, you can change that function,” Gillies says. “The process becomes diabolically complex though, when hundreds of genes work together in a complex web of cause and effect to control a single function.”

One can make as many CRISPR gene modifications as time allows, but always with an eye on editing efficiency. “If efficiency is low or involves a more complex edit, investigators should edit one gene at a time, isolate the edited cells, then make the next edit. If edits are straightforward and the efficiency is high, cells may undergo several edits simultaneously.”

Gene editing allows the creation of new cell lines but our ability to mess with Mother Nature goes only so far. “There is always the potential, even with well-understood editing methods, to damage the cell, especially when editing multiple genes,” Gillies says.

To avoid these issues, she advises researchers to work with edits with thoroughly understood consequences such as phenotypes that occur naturally as cancer mutations, natural gene variants, disease mutations, or genetic modifications known to be nonlethal in other models.

“If your gene of interest is not well characterized you can always make an educated guess about which edits will bring about the desired effect,” Gillies adds. “Alternatively, you can edit several targets and screen for the one that does what you want. Note, however, that this approach can consume a lot of resources.”

A twist on phage display

In 2008 Daniel Gibson, then at the Craig Venter Institute, built the first living cell with an entirely artificial chromosome, and in the process introduced Gibson Assembly^®, a new paradigm in standard cloning methods. Gibson Assembly, now the core technology of Codex DNA, allows rapid cloning of multiple DNA fragments into vectors without the use of restriction enzymes.

“One of the main challenges in cell-line engineering remains the transfectability or transductabilty of cells,” Gibson tells Biocompare. “Selecting the most compatible technique for a cell line, from lipid-based transfection and electroporation to viral delivery and naked genetic transfer can mean the difference between progression and a stalled experiment.”

Phage display libraries are a high-throughput approach to CLD but their creation and validation are time-consuming. According to Codex DNA, preparation can be shortened from about 12 days to as few as five days, with minimal hands-on time, using Gibson Assembly and the company’s BioXp™, an automated platform for building libraries, fragments, and clones.

Operational considerations

The deeper one delves into automated or high-throughput laboratory processes, the easier it becomes to overlook operational factors that may be just as important as sheer numbers, for example consistency and degree of automation. Consistency provides assurance that results are trustworthy and comparable from day to day, lab to lab, user to user, and reagent batch to reagent batch. Lab automation flawlessly executes repetitive, monotonous, and error-prone operations, thereby reducing human error while freeing operators for more value-added tasks.

Integrated CLD workflows emply a wide variety of equipment, at minimum a temperature-controlled CO_{2 i}ncubator, an imaging system to study confluency or to find or characterize specific phenotypes, plus centrifuges, plate manipulation and storage, and liquid handlers. Even capping and uncapping freezer tubes require dedicated hardware.

Scheduling and timing become critical during projects involving the identification, passaging, expansion, and assaying of multiple cell lines. “You’ll often have an incubator full of cell culture plates that need to be manipulated at specific times,” says Casey Snodgrass, Ph.D., market segment leader for pharmaceutical sciences at Hamilton Company. “For a person to keep track of everything requires constant vigilance and significant hands-on activity. These processes can be inefficient when performed manually, and this is where automation can help to provide more walk-away time.”

Vendors have differentiated their business by focusing on incubators, plate handlers, readers, cell counters, or media, but the highly integrated nature of CLD demands collaboration. Hamilton, for example, offers the Microlab STAR and Microlab VANTAGE liquid handlers but works with BioTek Instruments for critical cell imaging capabilities and with others to design custom CLD workstations. “Whether you only need to monitor confluence, or if your work involves specific cell-based assays, vendors are happy to work with you to automate your cell culture workflow,” Snodgrass says. For example, Hamilton retains mechanical engineers on staff to accommodate special needs, for example for building an automated CLD system within a biosafety cabinet. “The correct approach is not simply to take one device and add it to a second device, but rather to try to understand synergies and interoperability.”

Given the wide choices in expression systems and editing methods, it seems reasonable to expect CLD to deliver the “ideal” cellular expression system at some point, but according to ATCC’s Gillies, this hope is unreasonable.

“The main reason we haven’t seen a one-size-fits-all cell line is because each protein product is different, and there are just too many variables to control,” she says. Gene edits that make a cell efficient at producing therapeutic antibodies won’t necessarily make them more efficient at producing viral proteins or live viral particles for vaccines. Also, since the relationship between genes and protein production capacity is complicated, in most cases there’s no one small change or series of changes to the genome that can provide universal and significant improvements in protein expression.”

Or, as Codex’s Daniel Gibson warns, “there is no perfect cellular expression system.”