All vaccines have the same goal: disease prevention. But different vaccines reach this goal in different ways. That is, vaccine designs vary widely. Bacterial polysaccharides can be conjugated to immunogenic carrier proteins, or vaccines can be developed based on DNA, RNA, or proteins.

All these options reflect how the vaccine development toolbox keeps expanding. In many instances, vaccine development can incorporate genetic engineering techniques that make it possible to generate more stable and heat-resistant products. Such products are advantageous because they lend themselves to simplified manufacturing and distribution processes.

The latest advances in vaccine design and production were discussed at the 13th Annual Bioprocessing Summit, a hybrid event that was held last August and that accommodated virtual and in-person participants. (The latter convened in Boston.) The event’s most compelling presentations are revisited in this article, which offers expanded commentary from the original presenters.


Highly abundant and distinct from eukaryotic sugars, the surface-exposed capsular polysaccharides of some bacteria are excellent vaccine targets. But these polysaccharides are usually very poor immunogens unless they are conjugated to a protein. Five immunogenic carrier proteins are currently used: diphtheria toxoid; tetanus toxoid; cross-reacting material 197 (CRM197); Haemophilus protein D; and the outer membrane protein complex (OMPC) of serogroup B meningococcus.

The major foci for conjugate vaccines in the United States are Haemophilus influenzae type b, Streptococcus pneumoniae (pneumococcus), and Neisseria meningitidis.

Manufacturing conjugate vaccines is complex. First, the pathogen is grown in bulk and the target polysaccharide extracted. “This is a very delicate process,” stresses Christian M. Harding, PhD, co-founder and CEO, VaxNewMo. “If you alter the polysaccharide, you may generate an immune response to something that does not exist in nature.”

The carrier protein is produced in large-scale fermenters, purified, and subsequently linked to purified polysaccharides using a chemical conjugation process. Many required release controls ensure minimal residual free polysaccharide or unconjugated carrier protein.

“Each serotype requires its own extraction and conjugation process,” Harding notes. “When Pfizer makes Prevnar 20, it has to carry out these arduous and technically challenging processes for all 20 serotypes.”

VaxNewMo process diagram
VaxNewMo has developed a new way to generate polysaccharide/protein conjugate vaccines. The conventional approach (upper panel) comprises three GMP processes (polysaccharide extraction, carrier protein purification, and chemical conjugation). VaxNewMo’s approach (lower panel) utilizes a single GMP process for one-pot synthesis of conjugate vaccines. The company’s in vivo platform incorporates conjugating enzymes that covalently attach a polysaccharide to an acceptor protein, bypassing the need for chemical conjugation.

An alternative is in vivo bioconjugation. VaxNewMo has engineered laboratory-safe Escherichia coli to co-express the vaccine polysaccharide antigen as a lipid-linked precursor, a carrier protein containing engineered glycotags (small regions of amino acids that will be glycosylated), and a powerful conjugating enzyme.1

“Instead of three GMP processes, we have one,” Harding states. “It is a beautiful symphony.” Other nonessential E. coli polysaccharide pathways are deleted. Only the vaccine polysaccharide antigen is transferred to the carrier protein.

The bioconjugation technology was discovered by Mario F. Feldman, PhD, and colleagues at the Swiss Federal Institute of Technology (ETH) in Zurich. To commercialize the technology, ETH-Zurich spun-out GlycoVaxyn in 2004. (This company was acquired by GlaxoSmithKline in 2015.) In 2016, Feldman co-founded VaxNewMo, where he is currently the chief science officer. He is also an associate professor of molecular biology at the Washington University School of Medicine.

Today, multiple bioconjugated vaccines are in clinical trials. “Our conjugating enzyme is more universal and will transfer virtually any lipid-linked polysaccharide to the carrier protein,” Harding asserts. “Other conjugating enzymes have more limited substrate versatilities and will not transfer polysaccharides with glucose as the first sugar of the polysaccharide chain.

“Most pneumococcal and Klebsiella pneumoniae capsules as well as all group B Streptococcus bacteria have glucose as the first sugar. Our bioconjugation technology can be applied to these vaccine targets and other pathogenic bacteria.”

DNA vaccines

“In January 2020, it was clear with our experience with Middle East respiratory syndrome (MERS) that we could quickly come up with a design for a SARS-CoV-2 DNA vaccine,” says Kate Broderick, PhD, senior vice president, R&D, Inovio Pharmaceuticals.

Inovio scientists uploaded viral genetic sequence information for SARS-CoV-2 into the company’s proprietary AI-based algorithm. And after a few hours, they had the design for INO-4800 based on the full-length spike protein. Small lots were manufactured for laboratory and animal tests. “The vaccine was in the clinic within 83 days,” Broderick notes. Phase I and II trials are complete.2

DNA molecules are extremely stable, so DNA-based vaccines do not require a frozen cold chain. The INO-4800 formulation contains only DNA, water, and salt for stabilization. The final ready-to-use liquid can be stored at room temperature for over a year, and at 37°C for over one month. Manufacturing is a straightforward process involving the fermentation of E. coli.

DNA-based vaccines generate quite specific and robust T-cell responses. “We believe those T-cell responses are key to mitigating the effects of the mutational changes of the virus,” Broderick relates. “There are no increasing side effects with more frequent administration. INO-4800 ticks all of the boxes that are needed for a global booster.”

INO-4800 has been tested against the alpha, beta, gamma, and delta variants, and it is currently being tested against lambda and mu. “So far, we have been able to generate neutralizing antibodies against all of them,” Broderick asserts. “In addition, our Cellectra device uses smart technology to ensure the vaccine is delivered in a tolerable and simple procedure.

She explains that the device “essentially pushes the DNA molecules into the cells in the right place to elicit the best type of immune response.” The device, which is battery operated, works like a laboratory electroporation device; however, it works with electrical parameters that are quite different for those used in human tissue applications. Robustly engineered, it employs a single-use disposable tip.

“Proactively, we need to have libraries of vaccines for different diseases, so that when they re-emerge, we are not scrambling to catch up,” Broderick declares. “A plug-and-play portfolio and a booster agnostic to the priming vaccine are absolutely key.”

INO-4800 is just one of a dozen DNA-based products in Inovio’s pipeline. Like INO-4800, one of these products, a therapeutic vaccine for diseases associated with human papillomavirus, is in a Phase III trial.

Molecular clamps

Many viral surface proteins are large, complex trimers that can exist in pre- and post-fusion forms. “Theoretically, if you can lock the molecule into the pre-fusion compact form, you have a better chance of maintaining the relevant epitopes for a good vaccine,” says Trent Munro, PhD, director of the National Biologics Facility and program director of the Rapid Response Vaccine Pipeline, Australian Institute for Bioengineering and Nanotechnology, the University of Queensland.

But when the molecule is in the compact, pre-fusion form, there is a tendency for the molecule to spring open and assume the more energetically favorable post-fusion form, a form in which epitopes may be obscured. One way to lock in the pre-fusion form is to engineer substitutions into the sequence, such as prolines. This approach, which may work well for COVID-19 vaccines, builds on elegant MERS research.

“But in a way, this was luck,” Munro admits. “This approach may have been more challenging for another sequence. We put our molecular clamp onto the SARS-CoV-2 spike protein. This trimerization domain lets us produce the vaccine recombinantly in CHO cells in a locked pre-fusion conformation.” The trimerization domain makes the protein very stable, especially to heat stress.

“The clamp, a plug-and-play domain, can be genetically fused onto any trimeric fusion protein even if little is known about the protein,” Munro continues. “In preclinical work, it worked well for pre-fusion proteins for Ebola, Nipah, SARS, MERS, influenza, and respiratory syncytial viruses. The version we used contained two small domains of the GP41 protein from HIV joined by a short linker. It was a solid trimerization domain. We also had a great antibody to create an affinity resin for future rapid manufacturing.”

fusion proteins
Many subunit vaccines are based on fusion proteins that stud viral capsids. In their pre-fusion form, these proteins may elicit a strong immune response. In their post-fusion form, however, they may effectively shield their antigenic sites. To “lock in” pre-fusion forms in candidate vaccines, researchers at the University of Queensland have developed a molecular clamp. The researchers have engineered it into the SARS-CoV-2 spike, a trimeric protein, to generate a new vaccine candidate that can be produced recombinantly in CHO cells in a locked pre-fusion conformation.

In partnership with Cytiva, Munro’s team developed a custom immuno-affinity resin that provided a high degree of purity in one step, simplifying the downstream production process. “To prove it could be manufactured efficiently at scale, we partnered with CSL, an Australian vaccine and biologics manufacturer,” Munro says. “The process worked successfully.”

Biologically, the clamp functioned as expected and generated robust protection against disease in animal studies. “However, our clinical trial cohorts also made a low immune response against the clamp,” Munro points out.3 “Any HIV screening test that contained GP41 created a cross reactivity issue.”

“Protein-based vaccines seem to have a better tolerability profile than nucleic acid and viral vector vaccines, but we have to verify the biology,” Munro adds. “Our next-generation 2.0 clamp is in preclinical testing, and we hope to take it to the clinic early next year.”

Sterile filtration

Sterile filtration using 0.2-µm-rated filters at the end of the vaccine purification process ensures bacterial removal. Yet this step can result in significant yield losses because of the size of some viral vaccines. The alternative approach, aseptic processing, significantly increases manufacturing costs.

“Sterile filtration is challenging with large viral vaccines,” says Andrew Zydney, PhD, Bayard D. Kunkle Chair and professor of chemical engineering, Pennsylvania State University. “We collaborated with Merck on a live attenuated cytomegalovirus
vaccine which has particles from 100 to more than 400 nm in size.”

sterile filtration chart
Researchers led by Andrew Zydney, PhD, at Pennsylvania State University have been working to improve sterile filtration as a means of vaccine purification. For example, they have developed a fluorescent nanoparticle model suspension that can mimic a live attenuated cytomegalovirus vaccine in terms of size distribution and interactions with sterile filtration membranes. The researchers say that the model can simplify yield determination.

Vaccine properties vary slightly from lot to lot, and with storage and age. To eliminate as much variability as possible during the study, Zydney’s team developed a model nanoparticle suspension that had approximately the same size distribution as the vaccine and similar interactions with the membrane. The fluorescent nanoparticles also simplified yield determination.4

“It is a complementary approach,” Zydney explains. “Studies performed with the nanoparticles were confirmed with observations from the vaccines.”

The initial focus was to understand the different performance characteristics of sterile filters all rated as 0.2 µm. Results showed variability in pore size. Some pores were smaller than rated, and others were larger. The larger ones let the vaccine pass more easily. In addition, pore sizes could be relatively homogenous throughout the depth of the filter, or they could vary. When they varied, they were larger at the entrance and smaller at the exit.

“A difference in pore size does not impact filtering a 10-nm antibody, but if you are filtering a virus that is between 180 and 240 nm, it can make a big difference in yield,” Zydney points out. “If you are having difficulty, consider trying some of the filters that have slightly larger pore sizes.”

The study also demonstrated the importance of prefilters to remove some of the larger particles and to thereby retard fouling of the sterile filter. In addition, a critical flow rate appeared to determine performance.

“Typically, when the flow rates are high, the capacity goes down, and you are unable to process as much volume per filter,” Zydney explains. “The increase in yield we saw was sufficiently large to suggest operating at the transition point. You may go through a few more filters, but an increased yield outweighs the cost.”

Ramping up capacity

“Vaccine development, approval, manufacturing, and distribution all pose complex challenges from a technical and political perspective,” says Andrew Sinclair, president and founder, Biopharm Services. These challenges were all evident during the initial COVID-19 response, when organizations began leveraging prior R&D investments to rapidly develop a range of vaccines.

Repurposing existing manufacturing facilities quickly established capacity. Yet more is needed since only about 35% of the global population is vaccinated. The scale of these facilities makes modular technologies feasible, such as those from G-CON Manufacturing, to deliver capacity locally.

The vaccines use a wide range of critical raw materials and consumables, supplies of which were initially unpredictable. This necessitated rapid increases in output of suppliers’ individual product lines primarily by extending production rates and operating hours, in addition to building new infrastructure.

Lead times of consumables have increased dramatically, hampering important projects for other diseases and for countries relying on imports. Suppliers need to examine regional distribution of manufacturing to develop more resilience in the supply chain.

Many recent analyses from vaccine and therapeutics developers have stated the importance of using a scalable development and manufacturing platform when producing a vaccine to fight a pandemic, Sinclair states. The ongoing use of manufacturing platforms with standard and/or interchangeable consumables supports agile manufacturing and helps alleviate supply chain issues for raw materials and consumables.

The advances being made in agile manufacturing of monoclonal antibodies using facility- and process-level advancements should be transferred to vaccine manufacturing. Doing so would allow multiproduct facilities to respond to changes in demand, such as those precipitated by local epidemics.

“It is important to invest more into the plug-and-play development approach to vaccines,” Sinclair advises. “Manufacturing needs to be more evenly distributed globally, and countries need to examine their pandemic response, such as strategies to use manufacturing capacity when there is no pandemic, and have it switch use and/or be capable of quickly adding capacity when one happens.”


1. Harding CM, Nasr MA, Scott NE, et al. A platform for glycoengineering a polyvalent pneumococcal bioconjugate vaccine using E. coli as a host. Nat. Commun. 2019; 10(1):891. DOI: 10.1038/s41467-019-08869-9.
2. Mammen MP, Tebas P, Agnes, J, et al. Safety and immunogenicity of INO-4800 DNA vaccine against SARS-CoV-2: A preliminary report of a randomized, blinded, placebo-controlled, Phase 2 clinical trial in adults at high risk of viral exposure. medRxiv preprint. May 7, 2021. DOI: 10.1101/2021.05.07.21256652.
3. Chappell KJ, Li Z, Wijesundara DK, et al. Safety and immunogenicity of an MF59-adjuvanted spike glycoprotein-clamp vaccine for SARS-CoV-2: A randomized, double-blind, placebo-controlled, phase I trial. Lancet Infect. Dis. 2021; 21(10): 1383–1394. DOI: 10.1016/S1473-3099(21)00200-0.
4. Taylor N, Ma W, Kristopeit A, et al. Evaluation of a sterile filtration process for viral vaccines using a model nanoparticle suspension. Biotechnol. Bioeng. 2021; 118(1): 106–115. DOI: 10.1002/bit.27554.

Previous articleData Hoarding Poses an Existential Threat to Biopharma
Next articleCOVID-19 Shakes Up Vaccine Production