Cell Line Development for Biologics and Gene Therapy

Cell line development is the process of engineering a living cell to reliably produce a target protein or viral vector, then selecting and banking the single best-performing clone for large-scale manufacturing. It underpins virtually every biologic drug on the market, from monoclonal antibodies to gene therapies, and the choices made during development ripple forward into product quality, manufacturing cost, and regulatory approval. The process typically spans several months and involves picking a host cell, inserting the gene of interest, isolating individual clones, and proving that the chosen clone remains stable and productive over dozens of generations.

Why the Host Cell Matters So Much

The first decision in any cell line development campaign is which host cell to use. Chinese hamster ovary (CHO) cells dominate biopharmaceutical manufacturing for good reason: they grow well in suspension culture, they can be adapted to serum-free media, and regulators around the world have decades of experience reviewing products made in them. Other mammalian options include murine myeloma lines (NS0, Sp2/0) and, increasingly, human cell lines such as HEK293 and PER.C6. Human-derived lines are attractive because the sugar structures they attach to proteins more closely resemble those found on proteins in the human body, reducing the chance of triggering an immune response in patients.¹ Non-human mammalian cells can add sugar residues like galactose-α1,3-galactose and N-glycolylneuraminic acid that human cells do not produce, and these non-human sugars are potentially immunogenic.

The host cell’s influence goes beyond broad species differences. Even within a single product, the specific cell line used can dramatically change how sugars are added at individual sites on the protein. In one study comparing CHO and HEK cells producing the same antibody variant, sialylation at one glycan site dropped from roughly 82% in CHO to under 3% in HEK, while sialylation at other sites was comparable between the two.² These differences matter because sugar patterns influence how long a drug stays in the bloodstream, how effectively it triggers immune cells, and whether it causes unwanted side effects. Each host cell type expresses its own repertoire of glycosylation enzymes and sugar transporters, contributing to distinct and sometimes unpredictable glycan profiles.³

Getting the Gene Inside the Cell

Once a host cell is chosen, the gene encoding the desired protein needs to be inserted into the cell’s genome so that it stays there permanently and keeps being expressed through many rounds of cell division. The traditional approach uses plasmid DNA that integrates randomly into chromosomes. This works, but the randomness means every cell ends up with the gene in a different location and at a different copy number, leading to huge variation in how much protein each cell produces. Teams then have to screen thousands of clones to find the handful that are high producers.

Transposon-based systems have emerged as a more efficient alternative. Transposons are naturally occurring mobile DNA elements that have been repurposed as gene-delivery tools. Systems based on Sleeping Beauty, piggyBac, and Tol2 transposons can integrate foreign DNA into the host genome with high efficiency.⁴ In practical terms, CHO cell pools generated using transposon systems have produced roughly nine-fold more target protein than pools generated by conventional plasmid transfection, with cells carrying an average of ten to twelve integrated copies of the transgene.⁵ That kind of boost significantly reduces the number of clones that need to be screened downstream.

Another way to tilt the odds is to engineer the parent cell before transfection. Knocking out the glutamine synthetase (GS) gene in CHO cells, for example, increases selection stringency. When the gene encoding the product of interest is delivered alongside a functional GS gene, only cells that have genuinely integrated the transgene can survive in glutamine-free media. One study found that using GS-knockout parent cells improved bulk culture productivity two- to three-fold and achieved a six-fold improvement in the efficiency of identifying high-producing clones.⁶

Targeted Integration and Landing Pads

Random integration is giving way to more precise approaches. Targeted integration places the transgene at a pre-characterized genomic location known to support high and stable expression. One strategy involves building “landing pads” into the host cell genome: these are regions enriched with matrix attachment region (MAR) elements and containing specific recombination sites recognized by enzymes like BxB1 integrase. A cell line carrying these landing pads can then accept multiple genes of interest at defined locations simultaneously, which is especially useful for producing multi-chain proteins like antibodies.⁷ Because every clone has its transgene in the same chromosomal neighborhood, the variation between clones shrinks considerably, and the stability of expression tends to improve.⁸

CRISPR-Cas9 and related genome editing tools play an expanding role here, both for creating the landing pad host and for knocking out or modifying endogenous genes that affect product quality. But the landing pad concept is attractive because it separates the genome engineering from the product-specific campaign: once you have validated a host with good landing pads, you can drop different transgenes into the same site for product after product, saving months of development time.

Isolating a Single Clone

Regulatory agencies expect that a biopharmaceutical production cell line is derived from a single progenitor cell. This “clonal derivation” step is important because it ensures the cell bank is genetically uniform, which in turn supports consistent product quality across manufacturing campaigns.⁹ How you achieve single-cell cloning, however, has practical consequences for efficiency and timeline.

The traditional approach is limiting dilution: cells are serially diluted into multi-well plates at a statistical density where most wells should contain zero or one cell. It is simple but wasteful, with single-cell efficiency around 13% for CHO lines. Flow cytometry-based cell sorting improves this to roughly 64–67%, but the more recently developed single-cell printers push single-cell deposition efficiency above 80%. Perhaps more important than deposition is outgrowth: the fraction of single cells that survive and expand into usable colonies. For some CHO lines, single-cell printers achieved outgrowth rates over 50%, compared with around 3% for flow cytometry and 10% for limiting dilution.¹⁰ The outgrowth advantage is cell-line-dependent, though: for a different CHO line tested in the same study, flow cytometry and single-cell printing performed similarly. Knowing which cloning method suits your specific host matters more than assuming the newest technology is always best.

Screening Clones at Small Scale

After cloning, you might have hundreds of candidates, each needing evaluation for productivity, growth, and product quality. Running each one in a full bench-top bioreactor would be prohibitively slow and expensive. Miniaturized, automated bioreactor platforms have filled this gap. The Ambr 250 system, for example, runs disposable vessels at a scale of a few hundred milliliters but generates cell growth, productivity, metabolite, and product quality profiles comparable to those seen in five-liter and even thousand-liter bioreactors.¹¹ Even smaller systems, operating at 15 milliliters, have shown that while their physical mixing characteristics differ substantially from large-scale stirred tanks, the controlled environment they provide yields biological performance similar to bench-top bioreactors, unlike shake flasks, which often give misleading results.¹² These automated platforms let development teams evaluate dozens of clones in parallel, compressing what used to be months of sequential bench-top work into weeks.

Keeping Clones Stable Over Time

Finding a high-producing clone is only half the battle. The clone needs to maintain its productivity across the sixty to eighty generations it will go through between the master cell bank and the end of a manufacturing campaign. Production instability is one of the most frustrating problems in cell line development, and a significant cause is epigenetic silencing of the transgene promoter. The commonly used human cytomegalovirus (hCMV) promoter contains CpG sites that can become methylated over time, effectively turning down or shutting off the gene it is supposed to drive. Researchers have identified specific CpG dinucleotides within this promoter that are frequently methylated in unstable CHO cell lines.¹³ High transgene copy number also correlates with instability, likely because some of the integration sites are in chromosomal regions prone to silencing.

Stability testing is therefore a non-negotiable part of cell line development. Clones are typically passaged continuously for the expected production lifespan, with samples taken at intervals to measure productivity and compare product quality to the starting material. Emerging work suggests that machine learning models trained on epigenetic data could eventually predict which clones will remain stable, potentially shortening this testing phase.¹⁴ These models are still at an early stage, but the idea of flagging unstable clones before investing months in stability studies is compelling.

Serum-Free Media and Culture Adaptation

Modern biopharmaceutical production requires serum-free media. Animal-derived serum introduces variability between batches, carries a risk of contamination with viruses or prions, and makes regulatory filing more complicated. Adapting cells to grow without serum is therefore a standard step in cell line development. For CHO, HEK293, Vero, BHK-21, and MDCK cells, successful serum-free suspension culture has been achieved through direct or gradual adaptation, medium optimization, and sometimes genetic modification.¹⁵

The adaptation itself is not trivial. When HEK293 cells were moved from serum-containing to serum-free medium, the process took about a month, during which cell viability stayed above 90%. But the metabolic profile shifted meaningfully: the biggest metabolic differences correlated with the switch between adherent and suspension growth modes, followed by the change in medium composition.¹⁶ When serum-free-adapted suspension cells were returned to serum-containing medium, their metabolic profiles reverted immediately, showing that the extracellular environment, not permanent genetic changes, was driving most of the metabolic shifts. This is reassuring for consistency but also means that media composition needs to be tightly controlled throughout manufacturing.

Tuning Glycosylation During Manufacturing

Because the sugar structures on a biologic drug directly influence its safety and efficacy, controlling glycosylation has become a field unto itself. Several strategies exist for steering glycan profiles toward a desired outcome. Cell line engineering approaches involve knocking out or overexpressing specific glycosylation enzymes inside the production cell. Media-based approaches add supplements like manganese, galactose, or uridine during fermentation to push metabolic pathways in the desired direction.¹⁷ In vitro glycoengineering, carried out after the protein has been purified, offers yet another layer of control.¹⁸

In practice, most teams use a combination of genetic and media-based levers. For example, knocking out the gene for fucosyltransferase in CHO cells produces antibodies with reduced fucosylation, which enhances their ability to recruit natural killer cells against tumor targets. This kind of glycoengineering has already reached the market: several approved antibodies are manufactured in glyco-engineered cell lines designed to boost their cancer-killing activity.¹⁹

Adapting Cell Line Development for Perfusion Processes

The biopharmaceutical industry is shifting toward intensified manufacturing, where cells are grown at higher densities in perfusion bioreactors that continuously feed fresh medium and remove spent medium. This can deliver more product from the same equipment, but it exposes a mismatch in how cell lines are traditionally developed. Standard cell line development campaigns select clones that produce the highest total titer in a fed-batch format, where cultures run for about two weeks and then stop. These clones are optimized for that pattern: fast growth early on, then a shift to high productivity as nutrients deplete. In a perfusion process, you need cells that maintain both steady growth and high per-cell productivity simultaneously, sometimes for weeks on end.²⁰ A clone that was top-ranked in a fed-batch screen may underperform in perfusion, and vice versa. Teams developing for intensified processes increasingly screen clones under perfusion-like conditions from the outset rather than trying to adapt fed-batch winners after the fact.

Beyond Antibodies and Into Gene Therapy

Cell line development is no longer just about producing soluble proteins. The explosion of gene therapy has created enormous demand for stable cell lines that produce viral vectors. Adeno-associated virus (AAV) vectors, the workhorse of in vivo gene therapy, are still largely made by transient transfection: large quantities of plasmid DNA are pushed into cells fresh for each production batch. This is expensive and hard to scale. The field is moving toward stable packaging and producer cell lines, typically based on HEK293 or HeLa, where the viral genes are already integrated into the genome and can be induced on demand.²¹ The same logic applies to lentiviral vectors used in ex vivo gene therapy and CAR-T cell manufacturing. Stable suspension producer cell lines have now been generated for serum-free lentivirus production, with monoclonal lines achieving infectious titers near 10⁸ transducing units per milliliter and maintaining productivity above 10⁷ units per milliliter per day for nearly a month in a perfusion bioreactor.²²

Building these vector producer lines borrows many of the same principles used for antibody cell lines: host selection, transgene integration, single-cell cloning, stability testing, and serum-free adaptation. But the biology is trickier because the cells have to produce complex multi-component viral particles rather than a single secreted protein, and the viral components can be toxic to the producing cell. Engineering inducible expression systems, where viral gene expression is turned on only when production is needed, helps manage that toxicity.

Regulatory Expectations for Cell Banks

A finished cell line becomes a master cell bank: a large number of vials frozen under controlled conditions, each containing the same clone. Regulatory agencies require extensive characterization of this bank, including identity testing to confirm the species and cell line, purity testing to rule out contamination by bacteria, fungi, mycoplasma, and adventitious viruses, and stability testing to confirm that cells recovered from the bank perform consistently over their intended production lifespan.²³ For newer cell types like induced pluripotent stem cells, regulators are still working out the details of what minimum testing should look like, but the general framework of identity, purity, and stability applies regardless of cell type.

Clonal derivation fits into this regulatory picture as a quality control tool. Demonstrating that the production cell line descends from a single cell gives regulators confidence that the cell population is genetically homogeneous, reducing the risk that a subpopulation of cells with different characteristics will emerge during manufacturing and alter product quality.

Cell Lines for Cultured Meat

One of the more unexpected applications of cell line development is in food production. Cultivated meat requires enormous quantities of muscle and fat cells grown in bioreactors, and primary cells taken from animal biopsies stop dividing after a limited number of passages. To make large-scale production feasible, researchers are immortalizing livestock satellite cells, the muscle stem cells responsible for regeneration, by overexpressing genes like telomerase (TERT) and cyclin-dependent kinase 4 (CDK4). Immortalized bovine satellite cells developed this way have achieved over 120 population doublings while retaining the ability to differentiate into muscle fibers.²⁴

The same strategy has been extended to other species. Tibetan pig satellite cells immortalized with TERT, CDK4, and survivin reached over 40 passages across 160 days, maintained myogenic differentiation capacity, and showed no signs of tumorigenicity in soft agar assays.²⁵ Ovine satellite cells have been immortalized with CDK4 and TERT as well, achieving stable proliferation with doubling times of 14 to 22 hours through 67 passages.²⁶ The big challenge unique to food applications is consumer and regulatory acceptance of genetically modified cell lines in a product people eat, a question that the pharmaceutical world largely does not face.

Insect Cells and Other Non-Mammalian Hosts

Not every product needs a mammalian cell factory. Insect cells, particularly the Sf9 line derived from fall armyworm, have long been used with baculovirus expression systems to produce recombinant proteins and virus-like particles for vaccines. Baculovirus-free approaches using plasmid transfection of insect cells are gaining ground because they simplify purification and enable the development of stable monoclonal insect cell lines that continuously produce the target protein or particle.²⁷ Insect cells grow quickly, tolerate high cell densities, and are easier to maintain than mammalian cells, but their glycosylation machinery is simpler and does not produce the complex sugar structures needed for many therapeutic antibodies. They fill a genuine niche for vaccines and structural biology, though, and cell line development principles like clone screening, stability testing, and serum-free adaptation apply just as much here as in the CHO world.

Engineering Cells to Resist Death

A production run in a bioreactor is a hostile environment: nutrients deplete, waste products accumulate, and dissolved oxygen fluctuates. Cells respond to these stresses by activating programmed death pathways. One engineering strategy to counteract this is overexpressing anti-apoptotic proteins from the Bcl-2 family, which block key steps in the cell’s suicide program. This approach has been shown in multiple cell types to extend culture viability and, by keeping more cells alive and productive during the late stages of a batch, increase total product yield.²⁸ Anti-apoptotic engineering is not universally adopted, partly because regulators view additional genetic modifications with scrutiny, and partly because improvements in media formulation and process control have independently reduced cell death in modern manufacturing. Still, for particularly challenging molecules or intensified processes that push cells to their limits, giving the cells built-in resistance to death remains a useful tool.