How Are Monoclonal Antibodies Manufactured?

Manufacturing a monoclonal antibody is one of the most complex undertakings in modern pharmaceutical production, involving living cells that must be coaxed into producing a single, precisely shaped protein at industrial scale. The process spans months from frozen cell bank to filled vial, passing through bioreactor cultivation, multi-step purification, viral safety clearance, and formulation into a product stable enough to survive storage and shipping. Each stage carries its own engineering challenges, and recent years have brought major shifts in how the industry approaches yield, cost, and sustainability.

Building the Cell Factory

Nearly all commercial monoclonal antibodies are produced in Chinese hamster ovary (CHO) cells. These cells were adopted decades ago because they grow well in suspension culture, can be adapted to serum-free media, and attach human-like sugar modifications to the antibody protein. The starting point for any new antibody product is cell line development: engineers introduce the gene encoding the desired antibody into CHO cells and then screen thousands of resulting clones to find one that produces the antibody at high levels and remains genetically stable over many cell divisions.

Traditional methods insert the antibody gene at random spots in the cell’s genome, which means every clone behaves differently. Site-specific integration systems aim to solve this by placing the gene at a predetermined chromosomal location, producing more predictable results. One study evaluating such a system for a complex bispecific antibody found that even with a theoretically fixed insertion site, enough natural genetic variation arose among clones to yield meaningfully different production profiles, which actually helped identify high-performing candidates for a multi-chain molecule that would be difficult to produce otherwise.¹ Another group demonstrated that a site-specific clone could be carried through more than 120 generations and scaled to 12,000-liter production runs while maintaining consistent performance.²

Beyond gene insertion strategy, the choice of promoter (the genetic switch that tells the cell how hard to work) matters. Researchers have explored using a stress-responsive promoter from the Hspa5 gene, which keeps antibody production humming during the later, nutrient-depleted phase of culture when conventional promoters tend to fade. In fed-batch cultures, this approach maintained transcription rates through the final days and improved overall antibody yield compared to a standard promoter.³ A newer CHO-derived host called CHO-MK, paired with specially optimized media, pushed titers to roughly 12.6 grams per liter in just seven days of fed-batch culture, a pace that translates to about 1.8 grams per liter per day.⁴ Numbers like these were almost unthinkable a decade ago and represent a generational leap in how much antibody a single batch can deliver.

Fed-Batch Versus Perfusion Bioreactors

Once you have a productive cell line, the next question is how to grow it. The two dominant approaches are fed-batch and perfusion culture, and each has trade-offs that ripple through every downstream decision.

In fed-batch culture, cells are seeded in a bioreactor, fed concentrated nutrients over roughly two weeks, and harvested all at once when the culture is spent. This is the industry workhorse. Peak cell densities in fed-batch can reach around 49 million cells per milliliter, with antibody titers climbing to about 5 grams per liter over a 15- to 16-day run.⁵ The product accumulates in the broth and is harvested in a single large batch at the end.

Perfusion culture works differently. Fresh media flows continuously into the bioreactor while spent media carrying the antibody is continuously drawn off through a cell-retention device. Cells stay in the vessel, and the antibody is collected in a steady stream. A head-to-head comparison found that while fed-batch produced a higher final product concentration in the vessel (roughly 2.5 times higher than perfusion), the average daily productivity of perfusion was about 7.5 times greater.⁶ Perfusion also tends to produce a more uniform product because the antibody spends less time sitting in warm, enzyme-rich culture fluid. In 50-day perfusion runs, sustained harvest rates above one gram per liter per day have been demonstrated.⁷

The Scale-Up Problem

Moving from a small lab bioreactor to an industrial vessel of thousands of liters is one of the hardest steps in manufacturing. The core issue is hydrodynamic stress: as bioreactors get larger, the fluid dynamics change. Impeller blades spin to mix nutrients and oxygen, creating zones of intense shear near the impeller and calmer regions elsewhere. Cells cycling through those zones experience repeated bursts of mechanical force.

CHO cells are generally robust, but they have limits. Scale-down model experiments showed that when cells experienced sustained elevated shear stress, their antibody output per cell dropped by about a quarter. When the stress was made to oscillate, mimicking the repeated passage of cells through the turbulent impeller discharge zone of a large-scale reactor, productivity fell by roughly 30%.⁸ That is a meaningful loss at manufacturing scale, where even small percentage drops translate to kilograms of lost product.

Because you cannot easily experiment on a 10,000-liter reactor, engineers build small-scale models designed to recreate the shear environment of the big vessel. One approach uses computational fluid dynamics to map how shear rates relate to impeller tip speed across reactor sizes, from lab-scale vessels of a few liters up to 1,000-liter industrial tanks.⁹ Another recent strategy focused specifically on generating high shear stress in small reactors without introducing confounding effects like vortex formation, giving a cleaner picture of how cells would behave at manufacturing scale.¹⁰ Getting scale-up right means the antibody produced in a 2,000-liter or 12,000-liter vessel behaves the same as the material that was tested in clinical trials at bench scale.

Purification From Harvest to Bulk Drug

When the bioreactor run ends or harvest fluid is collected from a perfusion system, the antibody is mixed with a soup of host cell proteins, DNA, media components, and cellular debris. Turning that into pharmaceutical-grade material requires a multi-step purification process that typically begins with clarification and ends with a polishing train.

Clarification usually involves depth filtration to remove cells and large particles. As upstream titers and cell densities climb, this step has become more challenging: higher loads mean filters can clog faster, and host cell protein levels tend to increase. A recent quality-by-design study found that filter encapsulation design and lower feed flow rates improved both filtration capacity and host cell protein removal, while pore size mainly affected processing time rather than product quality.¹¹

After clarification, the antibody enters Protein A affinity chromatography, which is the near-universal capture step for antibodies. Protein A is a bacterial protein that binds tightly to the Fc region of most antibodies. The cell culture fluid passes through a column packed with Protein A resin, the antibody sticks, impurities wash through, and then a low-pH buffer releases the antibody in a highly concentrated, relatively pure form. This single step delivers high yield and effective clearance of host cell proteins, DNA, and aggregates.¹² The elution pH is a critical lever: it governs both the amount of antibody recovered and the proportion of intact monomers versus undesirable aggregates.¹³

One persistent headache with Protein A chromatography is ligand leaching: small amounts of the Protein A itself, and sometimes fragments of it, wash off the resin and contaminate the antibody pool. Fragments generated by proteinases in the cell culture fluid can be harder to remove in subsequent steps than intact Protein A.¹⁴ Managing this requires careful resin lifetime tracking and robust downstream polishing.

Polishing steps follow Protein A capture to remove remaining trace impurities, aggregates, and charge variants. A flow-through polishing approach combining activated carbon, anion-exchange, and cation-exchange chromatography in a connected, pool-less configuration achieved over 95% impurity clearance with over 95% antibody yield, matching the performance of traditional batch processing while streamlining the workflow.¹⁵

Viral Safety

Because monoclonal antibodies are produced in mammalian cells, viral contamination is a serious concern. CHO cells can harbor endogenous retrovirus-like particles, and adventitious viruses can enter through raw materials. The purification process therefore includes dedicated viral clearance steps, and regulators require manufacturers to demonstrate that these steps provide a cumulative margin of safety.

The standard toolkit includes low-pH viral inactivation (often performed right after Protein A elution, since the eluate is already at low pH), nanofiltration through virus-retentive filters, and chromatographic removal during polishing steps. Each step clears a different subset of viruses. One study examining multiple monoclonal antibodies found that low-pH treatment robustly inactivated retroviruses, nanofiltration reliably removed small parvoviruses, and flow-through anion-exchange chromatography removed retroviruses within established ranges of protein load and host cell protein content. Together, these steps provided a cumulative safety factor of about 10,000-fold over the expected level of endogenous retrovirus-like particles.¹⁶ The clearance was consistent across different antibody molecules, strengthening the case for using prior knowledge rather than repeating expensive virus-spiking experiments for every new product. Column chromatography steps throughout the process also contribute to overall viral removal, with Protein A itself providing a meaningful level of clearance.¹⁷

Formulation and Shelf Life

A purified antibody in solution is not yet a medicine. It needs a formulation: a carefully designed mixture of buffers, sugars, surfactants, and sometimes amino acids that keeps the protein stable, prevents aggregation, and ensures the product remains effective over its shelf life. This is where biophysics meets practical pharmacy.

Antibodies are large, fragile molecules prone to unfolding and clumping when exposed to heat, freezing, shaking, or the wrong pH. Studies of multiple clinical-stage antibodies show that even molecules with quite different thermal stability profiles can be formulated into pharmaceutically stable products, as long as the formulation is designed to boost the unfolding temperature of the antigen-binding region.¹⁸ The choice of excipients matters in specific ways: nonionic additives like sugars and polysorbate surfactants promote repulsive interactions between antibody molecules in solution, keeping them from sticking together, while ionic additives such as salts can increase short-range attraction at higher concentrations.¹⁹

Most antibody drug substances are stored frozen before final formulation and fill. The industry standard has been storage at −80 °C, but recent work found that −40 °C preserved stability just as effectively for a monoclonal antibody across multiple formulations, regardless of the glass-transition temperature. Interestingly, formulations containing more histidine had a lower glass-transition temperature yet were more stable, contradicting the intuition that staying far above that threshold is always better. For companies, this is meaningful: −40 °C freezers are cheaper to buy and run than −80 °C units, and supply chain logistics become simpler.

For products that need to be dried into a solid form, lyophilization (freeze-drying) is the traditional method but is slow and exposes the protein to temperature stress. A newer technology called Microglassification uses solvent extraction at ambient temperature to dehydrate the formulation rapidly. Side-by-side testing with lyophilization showed similar chemical stability, monomer content, charge variant profiles, and antigen binding for a monoclonal antibody processed by either method.²⁰

High-Concentration Formulations for Self-Injection

Many patients and healthcare systems would prefer subcutaneous injection over intravenous infusion. A subcutaneous shot can be self-administered at home in minutes, whereas an IV infusion typically requires a clinic visit lasting an hour or more. But the volume you can push under the skin in a single injection is limited, so delivering a therapeutic dose often means concentrating the antibody to extreme levels.

Current commercial formulations top out at roughly 150 milligrams per milliliter, and that ceiling exists for a reason: as antibody concentration rises, viscosity skyrockets.²¹ A solution too thick to push through a thin needle is useless for self-injection. Protein-protein interactions drive this thickening, and each antibody behaves differently depending on its surface charge, hydrophobicity, and shape.²² Formulation scientists have explored adding ionic liquids and other novel excipients to break up these interactions and bring viscosity down, potentially pushing concentrations well above the current commercial ceiling. Beyond viscosity, challenges include maintaining stability at high concentration, ensuring adequate bioavailability after subcutaneous injection, and managing immunogenicity risk.²³

Continuous Manufacturing and Real-Time Monitoring

Traditional antibody manufacturing runs in discrete batches: grow cells, harvest, purify, formulate, fill. Continuous manufacturing aims to link these steps into a single, unbroken flow, which could shrink the facility footprint and reduce costs. An economic and sustainability analysis estimated that compared to a best-in-class fed-batch process achieving 15 grams per liter, an end-to-end continuous process could cut annual production costs by up to about 23%, facility footprint by roughly half, plastic waste by close to 57%, and carbon emissions by about 54% in a multi-product facility.²⁴

Yet the industry has not rushed to adopt fully continuous processes. A recent analysis argued that rather than replacing all batch equipment wholesale, manufacturers should selectively adopt continuous technologies where they deliver clear value, such as perfusion bioreactors or continuous chromatography, while keeping proven batch steps elsewhere.²⁵ This hybrid approach avoids the risk and capital cost of a complete process overhaul while still capturing much of the efficiency gain.

Continuous processes demand continuous monitoring. Process analytical technology tools, from in-line spectroscopic probes to automated on-line chromatography, allow real-time tracking of product quality. On-line ion-exchange chromatography, for example, can detect shifts in antibody charge variants during both fermentation and purification, giving operators the information to intervene before a batch goes out of specification.²⁶ Challenges remain in areas like real-time viral detection and applying deep learning to process monitoring at commercial scale.²⁷

Alternative Production Platforms

CHO cells dominate commercial antibody production, but they are not the only option. Plant-based systems and cell-free synthesis platforms have carved out niches, especially for rapid response scenarios and cost-sensitive markets.

Plant molecular farming uses plants, most commonly Nicotiana benthamiana (a relative of tobacco), as living bioreactors. The antibody gene is introduced via a bacterial vector, and the plant’s own cellular machinery produces the protein. Transient expression in N. benthamiana can yield antibodies at the gram-per-kilogram level within weeks, far faster than the months needed to establish a stable CHO cell line.²⁸ Advances in glycoengineering have addressed the original concern that plant-made antibodies would carry non-human sugar modifications, and GMP-level plant production facilities now exist. The technology gained visibility during the Ebola outbreak, when the experimental antibody cocktail ZMapp was produced in tobacco plants.

Cell-free protein synthesis takes a more radical approach: instead of growing cells, you use a cell extract containing the translational machinery (ribosomes, enzymes, energy molecules) and add the DNA template directly. A platform based on CHO cell extract produced active, intact antibody at over 100 milligrams per liter and delivered results within two days, compared to at least a week for transient transfection in living cells.²⁹ A more recent workflow combined cell-free DNA template generation with cell-free synthesis and binding measurements to compress the antibody discovery cycle from weeks to hours.³⁰ Cell-free systems are unlikely to replace CHO for commercial-scale production anytime soon, but they are becoming valuable for screening large panels of candidate antibodies early in development, letting researchers identify winners before committing to the long process of cell line establishment.

Biosimilars and Acceptable Variability

As patents on blockbuster antibodies expire, biosimilar manufacturers must produce a version of the originator molecule that is highly similar but cannot be identical, because biological manufacturing inherently introduces variability. Monoclonal antibodies carry post-translational modifications, especially glycosylation patterns, that differ subtly depending on the cell line, culture conditions, and purification process. The regulatory question is: how much variation is acceptable?

A detailed mass-spectrometric comparison of rituximab biosimilars and the reference product found that the greatest glycan-related differences actually appeared between different biosimilars, rather than between any single biosimilar and the originator. Differences in glycoform distribution and afucosylated glycan levels correlated with slight variations in a laboratory measure of immune cell-killing activity. Yet overall, heterogeneity between each biosimilar and the reference product was relatively small, and because all the products analyzed were commercially approved, the observed variations represent real-world analytical variability that does not affect clinical performance.³¹ For patients, this is reassuring: the manufacturing differences between approved biosimilars fall within a window that regulators and analytical chemists have determined to be clinically meaningless.

Single-Use Equipment and Its Environmental Trade-Offs

Walk into a modern antibody manufacturing suite and you will see plastic everywhere: disposable bioreactor bags, tubing sets, filter capsules, and storage containers. Single-use technologies gained popularity because they eliminate the need for cleaning and sterilization between batches, reduce cross-contamination risk, and offer flexibility for facilities that produce multiple products. For years, the assumption was that single-use systems were also greener than traditional stainless-steel equipment because they save water and energy on cleaning cycles.

That assumption is now being challenged. A detailed bottom-up analysis of the carbon footprint of single-use technologies at the 2,000-liter scale found that CO₂ emissions from disposable equipment were significantly higher than previously estimated. Filters and bags emerged as the main emission hotspots. An operationally optimized hybrid facility, combining single-use components where they add flexibility with stainless-steel equipment where durability matters, significantly reduced both plastic waste and associated emissions compared to a fully single-use design.³² As the industry faces growing pressure to account for its environmental impact, hybrid facility designs are likely to become the standard rather than the exception.