A Standard CHO Cell Culture Protocol

A standard CHO cell culture protocol follows a broadly consistent arc: thaw a vial of cryopreserved cells into chemically defined medium, expand the culture through serial passages, transfer into a bioreactor under tightly controlled temperature, pH, and dissolved oxygen, feed the cells on a defined schedule, and harvest the product-containing fluid once the run is complete. Nearly 70% of recombinant therapeutic proteins on the market today come from Chinese hamster ovary cells, making this workflow one of the most consequential in modern pharmaceutical manufacturing.¹ But “standard” is relative. Each step involves choices that can shift product yield, product quality, and cell health in meaningful ways.

Thawing and Recovery

Every CHO culture begins with a frozen vial pulled from a cell bank stored in liquid nitrogen. The thawing step sounds simple, but getting it right sets the trajectory for everything that follows. A vial is warmed rapidly, usually in a 37 °C water bath, to minimize the time ice crystals spend in the damaging temperature range where they can grow large enough to puncture cell membranes. The thawed suspension is then diluted into pre-warmed medium to reduce the concentration of the cryoprotectant, typically dimethyl sulfoxide (DMSO), which is toxic to cells at culture temperatures. Several factors influence recovery: the cell concentration in the vial, the choice and concentration of cryoprotectant, and how quickly you get the cells into fresh medium all matter.²

Cell viability right after thaw is almost never 100%. Checking viability before moving forward is important, because if too many cells died during storage or thawing, the surviving population may not expand quickly enough or may behave differently from what your process expects. Most labs assess viability using trypan blue exclusion or an automated cell counter before seeding the first culture vessel.

What Goes Into the Medium

If you had walked into a mammalian cell culture lab in the 1980s, the flasks would have contained medium supplemented with fetal bovine serum, a complex and variable cocktail of growth factors, hormones, and proteins harvested from calf blood. The industry has since moved decisively away from serum. Today, CHO cells for therapeutic protein production are cultured in serum-free, chemically defined media, meaning every component is a known chemical at a known concentration.¹ The shift happened for several reasons: serum introduces batch-to-batch variability, carries a risk of viral or prion contamination, and makes regulatory filings harder.

A chemically defined medium typically includes amino acids, vitamins, trace metals, lipids, glucose, and buffering agents. Over decades of optimization, researchers identified which nutrients are critical, which metabolic byproducts are inhibitory, and how feeding strategies like fed-batch or perfusion can keep nutrient levels in a productive range while avoiding waste buildup.³ Some companies develop their own proprietary basal media through rounds of statistical screening and supplementation studies, which gives them more control over raw materials and more flexibility to tweak formulations during process development than buying a commercial off-the-shelf medium would.⁴

Passaging and Why Passage Number Matters

Between thaw and bioreactor, cells go through a seed-train expansion: you grow them in progressively larger vessels, splitting (passaging) the culture every few days to keep the density in a range where the cells are happy and dividing. The typical approach is to dilute a suspension culture to a target seeding density, let the cells grow for two to four days, then dilute again. Most CHO lines used in production are suspension-adapted, so they grow freely in liquid rather than stuck to a surface, which simplifies scale-up enormously.

One thing that surprises newcomers is how much cells can change over many passages. A study that tracked antibody-producing CHO cells in shake flasks for over 420 days found substantial differences in growth, productivity, and metabolism between populations cryopreserved at different time points. Growth performance and energy metabolism improved at higher passage numbers, and the cells showed signs of increased robustness based on stress-response proteins.⁵ That sounds like good news, but productivity can drift in unpredictable directions. This is why standard protocols specify a passage-number window: production runs pull from a working cell bank at a defined passage and never exceed a set number of additional doublings before the culture enters the bioreactor.

Selection Systems for High Producers

Before you have a culture to maintain, you need a cell line that actually makes your protein of interest at a useful level. The two dominant selection systems in CHO manufacturing are the dihydrofolate reductase (DHFR) system and the glutamine synthetase (GS) system. The DHFR system has been the historical workhorse, using a drug called methotrexate to apply selective pressure and amplify the gene encoding your product. The GS system uses a different inhibitor, methionine sulfoximine, and has gained ground because it generally requires a shorter timeline to generate stable, high-producing clones.⁶ In the GS approach, cells are grown in medium lacking glutamine; only cells expressing enough GS enzyme, linked to the gene for your target protein, survive and grow.⁷

Temperature, pH, and Dissolved Oxygen

Once cells are in a bioreactor, the environment is controlled far more precisely than in a shake flask. The three most important physicochemical parameters are temperature, culture pH, and dissolved oxygen (DO). All three individually and interactively affect cell growth, viability, metabolic activity, and the quality of the protein you are producing.⁸

pH management is trickier than it sounds. Carbon dioxide accumulates as a metabolic byproduct, and if it is not stripped out efficiently, the culture acidifies. One study found that inefficient CO₂ removal was the primary factor driving pH drift, and that the interaction between agitation speed and headspace airflow rate determined how much CO₂ built up. After increasing both parameters, the team maintained pH between 6.95 and 7.1, resulting in a 51% increase in final product titer compared to the poorly controlled run.⁹ Others have demonstrated that pH can be controlled using only sparging gases, without liquid base addition, by adjusting air sparging in a feedback loop. The approach scaled cleanly from benchtop to 200-liter reactors.¹⁰

The Temperature Shift Strategy

A common tactic in CHO fed-batch processes is the biphasic temperature shift. You start the culture at 37 °C, the physiological norm, to maximize cell growth during the early phase. Once the cells reach a high density, you drop the temperature, often to around 30–34 °C. This slows cell division but tends to extend viability and boost the amount of protein each cell produces. In one well-characterized example, shifting from 37 °C to 33 °C did not reduce maximum cell density, kept viability above 80% for nearly four extra days compared to the constant-temperature control, and increased total protein production six-fold.¹¹

The magnitude of the benefit varies by cell line and product, and the optimal shift timing and target temperature need to be worked out empirically for each process.¹² Microcarrier-based cultures under hypothermic conditions have shown productivity increases of roughly three-fold, while suspension cultures under similar shifts have seen increases as high as eight-fold.¹³ The point is that temperature is not just a set-it-and-forget-it parameter; it is an active lever for process performance.

Dealing with Lactate and Ammonia

CHO cells produce two metabolic waste products that cause trouble if they accumulate: lactate and ammonia. Both arise naturally from the consumption of glucose and amino acids, especially glutamine. High ammonia levels reduce viable cell density and alter the metabolism of other amino acids, triggering increased accumulation of alanine and glutamate.¹⁴ Ammonia stress also changes the sugar decorations (glycosylation) on antibody products, which is a product-quality concern, not just a cell-health one.

Lactate’s effects are more nuanced. At high concentrations, lactate inhibits growth by about 25%, but it can slightly increase the amount of protein each cell makes. It also reshapes how cells use glucose and glutamine: glucose and glutamine consumption rates drop by 15–20%, while ammonia and alanine production drop sharply, by roughly 64% and 70% respectively.¹⁵ Some of lactate’s apparent growth inhibition comes from the osmolarity increase it causes rather than from lactate itself. Managing these waste products is a major reason why fed-batch and perfusion strategies, which dilute or remove spent medium, outperform simple batch culture.

Fed-Batch Versus Perfusion

The two dominant bioreactor operating modes for CHO production are fed-batch and perfusion. In fed-batch, you start with a defined volume of medium, add concentrated nutrient feeds on a schedule as the culture progresses, and harvest everything at the end of the run, typically after 10–14 days. In perfusion, fresh medium is continuously pumped into the reactor while spent medium (containing your product) is continuously removed through a cell-retention device, keeping the cells inside.

Fed-batch dominates commercial manufacturing, largely because it scales more easily. One head-to-head comparison found that while a perfusion bioreactor’s volumetric productivity per unit time was about 7.5 times higher than fed-batch, the product concentration in the fed-batch harvest was about 2.5 times higher. Perfusion produced a more consistent product with less intra-lot heterogeneity and lower exposure of the product to cell debris. But fed-batch was simpler to operate and scaled to larger volumes.¹⁶ A separate study at industrial pilot scale found that the overall annual output of a fed-batch process was about 18-fold higher than a perfusion system that was limited to 0.6 cubic meters, even though the perfusion reactor was more productive on a per-volume basis.¹⁷

Real-Time Process Monitoring

Traditional monitoring of a CHO culture involves pulling samples by hand and running them on offline analyzers for glucose, lactate, ammonia, cell density, and viability. That works, but it gives you a snapshot rather than a continuous picture, and every time you open the bioreactor to sample, you introduce a contamination risk. The industry has been moving toward in-line sensors that measure key analytes in real time without breaking sterility.

Raman spectroscopy has emerged as a leading technology for this. A Raman probe sits inside the bioreactor and measures the scattering of laser light by molecules in the culture, which can be correlated with concentrations of glucose, lactate, ammonium, viable cell density, and even antibody titer.¹⁸ One group developed a glucose biosensor integrated into an automated feeding system that maintained stable glucose levels and showed that continuous feeding at even very low glucose concentrations had no harmful effects on protein quantity or quality.¹⁹ These tools are not just about convenience. They tighten control over the culture environment, which directly affects the consistency and quality of the product.

Glycosylation and Why Culture Conditions Shape Product Quality

CHO cells are the preferred host for therapeutic antibodies partly because they attach sugar chains (glycans) to proteins in a pattern similar to human cells. But the specific glycan profile on an antibody is not hardcoded by the gene encoding it. It shifts with culture conditions. The cell culture medium used in bioreactors influences both the yield and quality attributes of monoclonal antibodies.²⁰

Osmolality and culture duration are two variables with documented effects. In one study, combining high osmolality in both basal and feed media with an extended run time pushed the level of a specific glycan species, high-mannose 5, from about 12% to 28%, a more-than-two-fold increase. The same pattern held across four additional CHO cell lines. Supplementing with manganese chloride reversed the trend and reduced high-mannose levels.²¹ This matters because glycosylation affects how an antibody drug behaves in the body: its half-life, its ability to recruit immune cells, and even whether regulators consider a biosimilar close enough to its reference product.

Harvesting the Culture

Once the production run ends, the bioreactor contents are a murky soup of cells, cell debris, soluble host-cell proteins, DNA, metabolic waste, and your target protein. Clarification is the process of separating the product from everything else. The traditional sequence is centrifugation followed by depth filtration and then membrane filtration. Depth filtration is a popular choice for harvesting CHO cultures and can also remove some soluble impurities at this early stage.²²

High-titer processes, which can now exceed 10 g/L of antibody, push against the capacity limits of centrifuges and filters. The sheer volume of solids and sub-micron debris in these cultures can overwhelm a continuous centrifuge or clog the filtration train. One creative workaround is polymer-based flocculation: adding a positively charged polymer that clumps together the negatively charged cells and debris, creating larger particles that settle faster and are easier to filter out.²³

Guarding Against Mycoplasma

Contamination is the nightmare scenario in any cell culture facility, and mycoplasma is the most insidious form. These tiny bacteria lack a cell wall, which means they pass through the standard 0.2-micron filters used in bioprocessing. They can survive in CHO cultures with only subtle changes to process readings, making them hard to detect before they have done serious damage.²⁴ A contamination event disrupts cell growth, nutrient consumption, and oxygen use, and significantly degrades the quality of the antibody product.

Early warning signs from bioreactor data may include unusual shifts in dissolved oxygen demand, pH control behavior, ammonia levels, and arginine consumption.²⁵ Formal testing is still essential. Traditional culture-based mycoplasma assays take weeks to return results. PCR-based assays that can detect as few as one colony-forming unit per milliliter in CHO production cultures offer a faster alternative and are increasingly accepted by the industry.²⁶

CRISPR and the Engineering of Better Host Cells

The traditional approach to improving a CHO production process focuses on medium, feeds, and bioreactor conditions. Increasingly, though, the cells themselves are being redesigned. CRISPR-based gene editing has opened up a wide front of CHO engineering possibilities, from tuning glycosylation pathways to deleting genes that encode problematic host-cell proteins, to creating antibiotic-free selection systems that simplify regulatory submissions.²⁷

One striking example: researchers knocked out key genes in the apoptosis (programmed cell death) pathway using CRISPR-Cas9. Triple-knockout cells, with three apoptotic nodes disabled simultaneously, showed the highest viability, faster doubling times, and greater resistance to apoptosis. These cells outperformed wild-type and single- or double-knockout lines in producing a recombinant test protein.²⁸ Others have developed protocols to knock out up to ten genes in a single round of engineering, screening hundreds of clones to confirm all deletions at both DNA and protein levels.²⁹ The trend is toward purpose-built chassis cells, host lines pre-engineered with multiple beneficial modifications, rather than one-off edits for each new product.

Scale-Up Headaches and Single-Use Equipment

A process that works perfectly in a 3-liter benchtop bioreactor does not always translate to a 500-liter or 2,000-liter production vessel. Hydrodynamic stress, the mechanical forces exerted on cells by impeller rotation, sparging, and fluid flow, changes as bioreactors get bigger. Scale-down models that replicate the shear forces present at different locations inside a large stirred-tank reactor have been developed to help engineers predict how cells will respond before committing to an expensive large-scale run.³⁰

Single-use bioreactors, essentially large plastic bags inside a rigid support shell, have gained wide adoption because they eliminate the need for cleaning and sterilization between runs. But they introduce their own risks. In one recent case, a company scaling up production of a bispecific antibody saw slow cell growth and reduced titer at the 500-liter scale. The culprit turned out to be a chemical leaching from the polyethylene film of the single-use bag: a breakdown product of an antioxidant additive in the plastic. Switching to a different film chemistry restored normal performance.³¹ The lesson is that even the container your cells grow in is a process variable, and leachable testing should be part of every scale-up campaign, not an afterthought.