How the Modern RSV Vaccine Is Actually Made

Every approved RSV vaccine starts with the same molecular trick: locking the virus’s fusion protein into an unstable shape it briefly holds before infecting a cell. That prefusion form of the F protein is what the human immune system responds to most powerfully, and figuring out how to freeze it in place, mass-produce it, and deliver it into a person’s arm is really the story of how RSV vaccines get made. The details vary depending on the platform, but the underlying logic is shared across protein subunit, mRNA, and viral vector approaches.

Why the Prefusion Shape Matters So Much

RSV carries two main surface proteins, called G and F. The F protein is responsible for fusing the virus’s outer membrane with a human cell, which is how the virus gets inside and starts replicating. Both G and F are targets for neutralizing antibodies, but F has drawn most of the vaccine attention because it is more conserved across RSV strains and because it provokes a particularly strong immune response.

The F protein exists in two very different three-dimensional shapes. Before it does its job, it sits in a spring-loaded “prefusion” conformation. Once it helps the virus fuse with a cell, it snaps irreversibly into a “postfusion” shape. Research showed that most of the neutralizing antibody activity in human blood is directed at the prefusion form, not the postfusion one.

¹ This was a major insight because earlier vaccine attempts had used F protein that had already flipped into the postfusion shape, which turned out to be a far weaker target for the immune system.

The problem is that the prefusion form is inherently unstable. Left alone, the protein tends to spring into its postfusion shape on its own. So making a vaccine around prefusion F required solving an engineering puzzle: how do you keep an unstable protein locked in the shape you want long enough to manufacture it, ship it, and inject it?

A Lesson From a 1960s Disaster

The urgency behind getting the antigen design right has deep roots. In the 1960s, researchers tried a formalin-inactivated RSV vaccine in children. Instead of protecting them, the vaccine made subsequent natural RSV infections dramatically worse. Children who received it were hospitalized at high rates, and two died. Decades of research eventually explained the mechanism: the chemical inactivation process altered the virus’s proteins, producing antibodies that failed to neutralize the live virus and priming a harmful inflammatory response in the lungs upon real infection.

That catastrophe stalled RSV vaccine development for decades. It also cemented a lesson that shapes every modern RSV vaccine: the antigen has to faithfully present the right surfaces to the immune system. A mangled or reshaped protein can do more harm than good. Stabilizing the prefusion F protein was the field’s answer to that problem, because it presents the exact molecular surfaces that the immune system naturally learns to attack during a real RSV infection.

Engineering a Stable Prefusion F Protein

The breakthrough came from structural biology. Once researchers solved the crystal structure of the F protein in its prefusion state, they could identify the weak spots where the molecule tends to rearrange. They then introduced targeted mutations to shore up those weak spots. The most well-known set of modifications is called DS-Cav1, which combines a designed disulfide bond (“DS”) that pins two parts of the protein together and cavity-filling mutations (“Cav1”) that pack hydrophobic amino acids into an internal void, preventing the protein from collapsing into postfusion form.

Subsequent work refined the approach further. Researchers optimized the DNA sequence encoding the protein to make cells produce more of it and added further stabilizing mutations. One group improved prefusion stability by adding both the disulfide and cavity-filling mutations while also modifying the genetic code to lower CpG content and boost expression levels.

⁴ More recently, a proline-scanning strategy identified additional stabilizing substitutions, producing a candidate called preF7P that achieved high expression yields in manufacturing cells while remaining locked in prefusion form.

⁵

Even with these stabilizing mutations, long-term storage at refrigerator temperatures can cause the protein to gradually shift toward intermediate structures. Researchers addressed this by screening for additional mutations that kept the protein tightly in prefusion form over extended periods, using antibody-binding assays as a quality check: if the stored protein still bound antibodies known to recognize only the prefusion shape, it had held its conformation.

Growing the Protein at Scale

For protein subunit vaccines, once you have the stabilized gene sequence, you need a cellular factory to produce the protein. The workhorse for this job is the Chinese Hamster Ovary cell, better known as the CHO cell. CHO cells have been used for decades to produce monoclonal antibodies and other therapeutic proteins, and they bring well-understood manufacturing infrastructure. The recently approved CHO-derived RSV subunit vaccines are a direct extension of this established production platform.

⁶

The basic process works like this: the gene encoding the stabilized prefusion F protein is inserted into CHO cells. These cells are grown in large bioreactors, essentially temperature-controlled steel tanks filled with nutrient-rich liquid. As the cells multiply, they continuously secrete the F protein into the surrounding growth medium. After a production run, the liquid is harvested and the protein is separated from the cell debris and growth medium through a series of purification steps.

Yields matter enormously for global vaccine supply. The preF7P candidate mentioned earlier achieved expression levels of roughly 10 grams per liter in clinical-grade CHO cells, a high figure that makes large-scale manufacturing more feasible.

⁵ For context, many recombinant proteins express at one to three grams per liter, so a tenfold yield improvement translates directly into lower costs and greater availability.

An alternative manufacturing approach uses Vero cells grown on tiny bead-like surfaces called microcarriers inside stirred-tank bioreactors. This method has been explored for live-attenuated or vector-based RSV vaccine candidates and can be scaled using serum-free culture conditions, which simplifies regulatory approval and avoids animal-derived raw materials.

⁷

Purifying the Protein

Raw bioreactor harvest is a messy mixture. It contains the target protein along with host cell proteins, DNA fragments, lipids, and spent nutrients. Purification typically proceeds through multiple chromatography steps: affinity chromatography first grabs the target protein using a molecular “hook” that recognizes it specifically, pulling it away from everything else. Ion exchange chromatography then separates remaining impurities based on electrical charge, and size exclusion chromatography sorts molecules by size to polish the final product.

⁸ Each step reduces contaminants while preserving the protein’s prefusion shape. Quality control checks at each stage confirm that the F protein has not flipped to postfusion form during handling.

Covering Both RSV Subtypes

RSV comes in two major subtypes, A and B, and their F proteins differ enough that strong immunity against one subtype does not always fully neutralize the other. Some approved vaccines address this with a bivalent design, combining prefusion F proteins from both subtypes in a single shot. The Pfizer vaccine (Abrysvo), for example, delivers 120 micrograms of total prefusion F protein, split evenly between RSV-A and RSV-B components at 60 micrograms each.

⁹ This bivalent approach matters because RSV-A and RSV-B circulate simultaneously in most seasons and can cause equally severe disease.

¹⁰

Other vaccines use only an RSV-A prefusion F protein, betting that the cross-neutralization against RSV-B will be strong enough. The GSK vaccine (Arexvy) takes this approach, pairing its single-subtype antigen with an adjuvant to amplify the immune response. How each manufacturer decided on monovalent versus bivalent design comes down to their clinical trial data showing whether cross-protection was adequate or whether the second component was needed.

How Adjuvants Boost the Response

A purified protein on its own often does not provoke a strong enough immune response, especially in older adults whose immune systems are less responsive. Adjuvants are added to the formulation to amplify and shape that response. The GSK vaccine uses an adjuvant system called AS01E, which contains two immune-stimulating molecules embedded in liposomes (tiny fat bubbles). In a large trial of older adults, a single adjuvanted dose boosted neutralizing antibody titers against RSV-A roughly tenfold at one month after vaccination. Titers declined over time but remained about threefold above pre-vaccination levels a full year later.

¹¹

The Pfizer bivalent vaccine, by contrast, is unadjuvanted. It relies on the higher total antigen dose and the dual-subtype design to generate sufficient immunity. The trade-off between adjuvanted and unadjuvanted formulations involves balancing stronger immune activation against potential increases in injection-site reactions and systemic side effects. Both approaches have cleared regulatory review, which suggests the field has not settled on a single best strategy.

mRNA and Viral Vector Approaches

Not every RSV vaccine delivers a pre-made protein. Some use the body’s own cells as the factory. Moderna’s mRNA RSV vaccine, mRNA-1345, encodes the prefusion F protein in a strand of messenger RNA. That mRNA is wrapped in lipid nanoparticles, tiny fat-based delivery vehicles that protect the fragile RNA and shuttle it into cells after injection. Once inside, your cells read the mRNA instructions and produce the prefusion F protein themselves, which then gets displayed on cell surfaces and triggers an immune response.

¹²

The mRNA is codon-optimized, meaning its genetic sequence has been tweaked so human cells translate it efficiently. It also encodes an F protein with a deleted cytoplasmic tail, a modification that improves how the protein gets presented to the immune system. Manufacturing mRNA vaccines involves in vitro transcription rather than cell culture: enzymes copy the RNA from a DNA template in a test tube, and the resulting mRNA is then purified and mixed with lipids using microfluidic devices that force the two components together at precisely controlled flow rates to form uniform nanoparticles.

Viral vector vaccines take yet another approach. Johnson & Johnson developed Ad26.RSV.preF, which uses a replication-incompetent adenovirus 26 vector carrying the gene for prefusion F. The adenovirus infects your cells but cannot replicate; it simply delivers the F gene, and your cells produce the protein.

¹³ GSK explored a chimpanzee adenovirus vector (ChAd155-RSV) that encodes not just the F protein but also the nucleocapsid and antitermination proteins, aiming for a broader immune response.

¹⁴ Using a chimpanzee adenovirus sidesteps the problem of pre-existing immunity to human adenoviruses, which can blunt a vaccine’s effectiveness.

Maternal Vaccines and Protecting Newborns

RSV is most dangerous for very young infants, who cannot be vaccinated themselves in the first months of life. One solution is vaccinating pregnant women so that their antibodies cross the placenta and protect the baby after birth. The Pfizer bivalent vaccine is approved for this use during weeks 32 to 36 of pregnancy. Maternal vaccination induces strong neutralizing antibody responses, and these antibodies transfer efficiently to the fetus, providing passive protection during the highest-risk window after birth.

¹⁵

Timing turns out to be critical. Antibody transfer across the placenta is not instantaneous; it takes weeks for antibodies to accumulate in the fetal bloodstream. A study examining cord-to-maternal antibody transfer ratios found that vaccination two to three weeks before delivery resulted in significantly lower transfer efficiency compared to vaccination five or more weeks before delivery. The transfer ratio jumped from about 0.76 when vaccinated two to three weeks out to about 1.43 when vaccinated more than six weeks before delivery.

¹⁶ This means vaccinating earlier in the approved window gives the baby substantially more protection.

Monoclonal Antibodies as an Alternative Delivery

Instead of training the immune system to make its own antibodies, monoclonal antibody products deliver ready-made antibodies directly. Nirsevimab (brand name Beyfortus) is a lab-engineered antibody that targets the same prefusion F protein site that vaccines aim to stimulate responses against. It binds a highly conserved spot called antigenic site Ø, blocking the virus from fusing with host cells. The antibody was engineered with three amino acid substitutions in its Fc region (known as the YTE modification) that extend its half-life to an average of 71 days, long enough that a single injection protects an infant through an entire RSV season.

¹⁷

Manufacturing monoclonal antibodies is its own industrial process, using CHO cells much like subunit vaccines but with different purification requirements and far higher doses per patient. The cost implications are significant, which feeds directly into decisions about whether to deploy maternal vaccination or infant monoclonal antibodies in different settings.

Measuring Whether It Works

One challenge facing RSV vaccine manufacturers is the lack of a universally accepted threshold that defines “protective” antibody levels. Researchers have been working to establish correlates of protection, meaning measurable immune markers that reliably predict whether someone is protected from disease. The evidence has been inconsistent historically, with fewer than a dozen studies investigating the relationship between antibody levels at birth and RSV outcomes in infants, and their findings have not always agreed.

¹⁸

More recent clinical trial data has been more encouraging. An analysis of Moderna’s mRNA-1345 trial found that neutralizing antibody levels and prefusion F-binding IgG measured about a month after vaccination met formal statistical criteria as surrogate endpoints for vaccine efficacy, with RSV-A neutralizing antibodies showing the strongest association.

¹⁹ Similar findings emerged from the adenovirus vector vaccine program, where both prefusion-binding IgG and neutralizing antibodies measured shortly after vaccination were significantly associated with reduced RSV illness.

²⁰ Establishing these correlates matters not just for understanding current vaccines but for streamlining the approval of future ones, since regulators could potentially accept immune-marker data instead of requiring massive efficacy trials every time.

Can the Virus Evolve Around These Vaccines?

Because all current RSV vaccines and monoclonal antibodies target the F protein, a reasonable worry is whether the virus can mutate its way out of protection. Research tracking natural RSV F protein evolution found that mutations do arise that allow escape from individual monoclonal antibodies. Deep mutational scanning of the nirsevimab binding site revealed numerous mutations, at more than a dozen amino acid positions, that reduce the antibody’s ability to neutralize the virus.

²¹

The reassuring finding is that these same mutations do not strongly erode neutralization by polyclonal sera, the mixture of many different antibodies that a vaccinated or previously infected person carries. Vaccines generate antibodies against multiple sites on the F protein simultaneously, so a mutation that dodges one antibody still gets caught by others. This is one reason why vaccines may be more durable against viral evolution than single monoclonal antibodies, and it is also why post-market surveillance of circulating RSV strains remains important.

Getting Vaccines to the Places That Need Them Most

RSV kills the most children in low- and middle-income countries, but deploying vaccines there faces practical obstacles that have nothing to do with molecular biology. Cold chain infrastructure is limited in many regions, seasonal RSV patterns are less predictable in tropical climates, and the timing requirements for maternal vaccination clash with irregular antenatal care schedules in some communities.

²²

Cost-effectiveness analyses in countries like Nepal have found that both maternal vaccination and infant monoclonal antibodies fall well below standard thresholds for cost-effective health interventions, with the maternal vaccine costing roughly $387 per disability-adjusted life year averted and the monoclonal antibody about $486.

²³ Those are favorable numbers on paper. But favorable cost-effectiveness does not automatically translate into funded programs, donated doses, or functioning cold chains in rural clinics. The gap between a vaccine that exists and a vaccine that reaches infants in sub-Saharan Africa or South Asia remains one of the hardest problems in global health, and RSV vaccines are no exception.

Post-Market Safety Signals

Both approved RSV vaccines for older adults have been flagged for a potential association with Guillain-Barré syndrome, a rare autoimmune condition affecting the peripheral nerves. Post-marketing surveillance through the U.S. Vaccine Adverse Event Reporting System found that the median time from vaccination to GBS onset was about 9 to 10 days for both vaccines, with no significant difference in reporting rates between the two products.

²⁴ The signal is rare enough that it did not prevent approval, but it is being tracked closely and has influenced recommendations about who should receive the vaccine and whether repeat dosing is advisable. For anyone weighing the decision, the risk of GBS appears to be on the order of a few cases per million doses, comparable to the background rate associated with some other vaccines.