Vaccine Types and Platforms

A vaccine works by showing the immune system part or all of a virus under controlled conditions, so that the adaptive response (the antibody and T cell arms that carry lasting memory) is primed before the first natural infection. When the real virus arrives, it is met by a fast, high-quality recall response instead of one assembled from scratch over one to two weeks. Everything that distinguishes one vaccine from another follows from a single question: how is the viral antigen presented? The answer defines the platform, and the platform sets the balance between how strong and complete the immune response is and how safe the vaccine is to give. This article takes the platforms in turn, then draws out how the platform is matched to the virus, how the seasonal influenza vaccine is produced, and where the technology is heading.

How a vaccine protects

Protective immunity established before infection rests almost entirely on the adaptive immune system: B cells and their antibodies, and T cells. The innate system (above all the type I interferons) shapes how large and what kind of adaptive response is generated, but it carries no durable memory and is not the arm a vaccine is trying to build.

Antibody and T cells divide the work. Antibody is the only defence that acts on free virus particles before they enter a cell, so it is the chief mediator of resistance to reinfection. The most potent antibodies are neutralising: they block attachment, entry or uncoating, and they recognise folded, three-dimensional shapes on the virus surface rather than flat fragments of sequence. CD8 T cells (cytotoxic T lymphocytes) recognise and kill cells that are already infected, displaying viral peptides on major histocompatibility complex (MHC) class I. CD4 T cells provide the help that drives strong antibody and durable CD8 memory. A useful generalisation is that antibody prevents infection while CD8 T cells clear it once it has started, with considerable overlap between the two.

A measurable immune marker that predicts protection is a correlate of protection, and where one exists a new vaccine can be licensed on the correlate instead of a full disease-endpoint trial. The classic example is influenza, where a haemagglutination-inhibition antibody titre of about 1:40 is the accepted surrogate. Two further points govern vaccine design. Mucosal protection depends on secretory immunoglobulin A (IgA), which serum immunoglobulin G (IgG) reaches only weakly, so a vaccine meant to block infection at the respiratory or gut surface ideally engages the mucosa directly. And antibody quality improves with repeated exposure through affinity maturation in lymph nodes, which is why most non-living vaccines need a primary course of two or more doses to reach full, durable protection.

The platforms at a glance

The platforms can be arranged on a spectrum from whole replicating virus to a single synthetic gene.

Platform	What it is	Live?	Example viral vaccines	Key strength	Key limitation
Live attenuated	A weakened whole virus that still replicates	Yes	Measles, mumps, rubella, varicella, oral polio, yellow fever, rotavirus	Strong, broad, durable immunity from one or two doses	Unsafe in pregnancy and immunocompromise; cold chain; reversion
Inactivated	Whole virus killed chemically	No	Inactivated polio, hepatitis A, rabies, most influenza	No infection risk; stable; safe in pregnancy	Needs adjuvant and boosters; weak mucosal and CD8 response
Protein subunit	A purified or engineered viral protein	No	Recombinant influenza (Flublok), recombinant zoster	Defined, very safe; scalable	Often weakly immunogenic without a strong adjuvant
Virus-like particle	Self-assembling capsid proteins, no genome	No	Hepatitis B, human papillomavirus	Native-shaped epitopes without any nucleic acid	Complex, costly manufacture; no antigen amplification
Viral-vectored	A harmless carrier virus carrying the antigen gene	Vector-dependent	Ebola (Ad26 and rVSV), some COVID-19	Live-like immunity with non-live safety	Pre-existing immunity to the vector can blunt it
mRNA	Lipid-wrapped messenger RNA encoding the antigen	No	COVID-19 (Spikevax, Comirnaty)	Designed from sequence; very fast to make	Short antibody durability; cold storage; reactogenic
DNA	Plasmid DNA encoding the antigen	No	None licensed for humans	Stable, cheap, easy to engineer	Inefficient delivery into the cell nucleus

Live attenuated vaccines

A live attenuated vaccine is a whole virus weakened so that it still replicates in the recipient but no longer causes disease. Because it replicates, it amplifies its own antigen and presents it exactly as a natural infection would, engaging antibody, CD4 and CD8 responses together, and when given by mouth or nose it raises mucosal IgA. The result is strong, broad and durable immunity, often from a single dose and usually without an adjuvant.

Attenuation is achieved by forcing the virus to adapt away from humans: repeated passage in animal cells or eggs accumulates mutations across several genes that lower virulence for people while preserving the antigens. The named vaccine strains are part of the history of the field and are worth knowing.

Vaccine	Strain	Origin
Yellow fever	17D	Theiler, by passage of the virulent Asibi strain in chick embryo tissue
Measles	Edmonston-derived	Enders and Peebles, isolated 1954
Mumps	Jeryl Lynn	Hilleman, isolated from his own daughter
Rubella	RA 27/3	Plotkin, human diploid cell passage
Varicella and zoster (live)	Oka	Takahashi
Oral polio	Sabin types 1 to 3	Naturally and passage-attenuated
Smallpox	Vaccinia and modified vaccinia Ankara	Derived from cowpox-related orthopoxvirus

The disadvantages all stem from the fact that the virus is alive. It can revert toward virulence: oral poliovirus vaccine (OPV) causes vaccine- associated paralysis in about one per two to three million doses, mostly after the first dose, and circulating vaccine-derived strains can seed outbreaks where coverage is low. A live virus can spread from the vaccinee to contacts, which boosts community immunity but raises a consent question. Most live viral vaccines are heat-labile and depend on an unbroken cold chain, a real constraint in warm climates. And, decisively for clinical practice, a replicating vaccine virus can cause disease in pregnancy or significant immunocompromise, so live vaccines are contraindicated in those groups.

Inactivated vaccines

An inactivated vaccine is a virulent virus killed so that it cannot replicate, usually with formalin or beta-propiolactone, then given as the whole killed particle or split into components. Inactivated poliovirus vaccine (IPV), hepatitis A vaccine, rabies vaccine and most seasonal influenza vaccines are of this type. Because nothing replicates, there is no infection risk and the vaccine is safe in pregnancy and immunocompromise, and the product is stable.

The cost of killing the virus is a weaker response. There is no internal antigen production, so the immune system sees the dose once; the response is antibody-dominated, with little mucosal IgA and poor CD8 T cell priming, and a primary course of two or three doses plus an adjuvant is usually needed. Inactivated polio vaccine illustrates the trade-off neatly: it reliably prevents paralytic disease through serum antibody but does not stop the virus replicating in the gut, so it protects the individual without fully interrupting transmission. Chemical inactivation can also damage antigens if done badly, a lesson learned from historical inactivated measles and respiratory syncytial virus (RSV) preparations that worsened later natural infection; the enhanced- disease story is taken up in the public-health article.

Subunit and virus-like particle vaccines

These platforms discard the whole virus and present only its protective protein. Most are made by recombinant DNA technology: the gene for a viral protein is inserted into a production cell (yeast, insect, mammalian or bacterial), which then manufactures the protein in bulk. Recombinant describes how the antigen is made, not what the finished vaccine looks like, and the technology reaches well beyond protein vaccines: viral-vector vaccines are genetically engineered viruses, and mRNA and DNA vaccines depend on the same molecular-cloning toolkit. Its importance is that a defined antigen can be produced without ever growing the pathogen, which makes vaccines possible against viruses that are dangerous, slow or impossible to culture, and yields a pure, consistent, scalable product.

A protein subunit vaccine presents the antigen as the purified protein itself. Recombinant haemagglutinin made in insect cells (Flublok) and the recombinant zoster vaccine (glycoprotein E) are the examples. Such vaccines are highly defined and very safe, but a single protein is often poorly immunogenic on its own and depends on a strong adjuvant.

A virus-like particle (VLP) is the same idea in particulate form. Viral capsid proteins are expressed so that they self-assemble into a particle that mimics the virus surface but contains no genome and cannot replicate or infect. A VLP is essentially a subunit displayed as an assembled, multimeric particle, and the immune system responds far more strongly to that ordered particle than to the equivalent free monomeric protein. The two great successes are both VLP vaccines: the recombinant hepatitis B vaccine (a 22-nanometre HBsAg particle, which replaced an earlier vaccine made from HBsAg purified from infected donors’ plasma) and the human papillomavirus (HPV) vaccine, built from the L1 capsid protein. HPV types 16 and 18 cause about 70% of cervical cancer; the bivalent and quadrivalent vaccines target these, and the nine-valent vaccine extends coverage to types responsible for roughly 90%. By preventing the chronic infection that drives liver cancer, the hepatitis B vaccine was also the first vaccine shown to prevent a human cancer. The limitations of VLPs are manufacturing cost and complexity, and the absence of any antigen amplification once injected.

Viral-vector vaccines

A viral-vector vaccine uses a harmless carrier virus, genetically engineered to carry the gene for the target antigen, to deliver that gene into the recipient’s cells, which then make the antigen themselves. This combines much of the immunogenicity of a live vaccine, including strong CD8 T cell responses, with the safety of a non-replicating product, and it needs no pathogen to be grown.

Vectors come in two kinds. Replication-defective vectors are crippled so they deliver their gene but cannot reproduce; the adenovirus vectors are the leading example, including the Ad26 vector used in an Ebola vaccine and in one COVID-19 vaccine, and the chimpanzee adenovirus vector ChAdOx1. Their main weakness is pre-existing immunity to the vector itself: antibody from past exposure to common human adenoviruses can neutralise the carrier before it delivers its cargo, which is why rare or animal-derived serotypes are chosen. Replication-competent vectors are live attenuated viruses carrying the extra gene; the recombinant vesicular stomatitis virus (rVSV) Ebola vaccine is the outstanding success, and the yellow fever 17D backbone has been used to carry the antigens of Japanese encephalitis and dengue viruses. Vectors are also used in heterologous prime-boost regimens, where a different vector or platform is given for the priming and boosting doses to broaden the response and sidestep anti-vector immunity.

Nucleic acid vaccines: mRNA and DNA

Nucleic acid vaccines take the logic one step further: instead of delivering a protein or a whole carrier virus, they deliver the genetic instructions, built with the same recombinant DNA technology, and let the host cell manufacture the antigen.

Messenger RNA (mRNA) vaccines package the antigen-encoding mRNA inside a lipid nanoparticle (LNP) that both protects the fragile RNA and acts as a built-in adjuvant. Two advances made the platform practical: replacing the normal base uridine with pseudouridine (the work of Karikó and Weissman), which stops the innate system from destroying the RNA and greatly increases how much protein is made, and stabilising the encoded surface protein in its pre-fusion shape so the antibodies raised are neutralising. The COVID-19 mRNA vaccines proved the approach at scale and demonstrated its defining advantage: a vaccine was designed, manufactured and authorised in under eleven months from the moment the viral sequence was published, far faster than any previous vaccine. The trade-offs are short-lived antibody responses that require boosting, a demanding cold chain, and more injection-site and systemic reactions than older vaccines. The platform is now moving beyond COVID-19, with mRNA vaccines against influenza, cytomegalovirus (CMV) and respiratory syncytial virus in clinical trials.

DNA vaccines deliver the antigen gene on a plasmid. They are stable, cheap and easy to engineer, and they induce both antibody and T cell responses, but the plasmid must reach the cell nucleus to be read, and this inefficiency has kept protective responses lower in humans than in small animals. No DNA vaccine is yet licensed for human use, though delivery methods such as electroporation are narrowing the gap.

Adjuvants

An adjuvant is a substance added to a non-living vaccine to strengthen and shape the immune response, compensating for the antigen production, native structure and innate stimulation that a live vaccine provides for free. Adjuvants act mainly by triggering innate sensors and recruiting and activating the antigen-presenting cells that start the adaptive response, and the choice of adjuvant can bias immunity toward antibody or toward cell-mediated responses.

A small number are licensed for viral vaccines, and each is worth recognising. Aluminium salts (alum) are the oldest and most widely used, working largely as a depot and innate trigger; they are in the inactivated vaccines and the hepatitis B and HPV products. AS04, which combines alum with monophosphoryl lipid A (MPL, a toll-like receptor 4 agonist), was the first adjuvant to add a defined innate ligand and is used in one HPV vaccine. MF59, a squalene oil-in-water emulsion, broadens influenza antibody responses and is dose-sparing. AS01B, a liposomal combination of MPL with the saponin QS-21, gives the recombinant zoster vaccine its high, durable efficacy. CpG 1018, a synthetic toll-like receptor 9 agonist, is used in a hepatitis B vaccine to improve response in adults who respond poorly to the older product.

Matching the platform to the virus

No platform is best for every virus; the right choice falls out of what the vaccine has to achieve. Where strong, durable, mucosal immunity is wanted and the recipients are healthy, a live attenuated vaccine is hard to beat, which is why measles, oral polio and rotavirus vaccines are live. Where the target population includes pregnant women and immunocompromised people, a non-living platform is essential, so a live product is avoided or a non-live alternative is developed (the recombinant zoster vaccine replacing the live one is the clearest case). Where the protective antigen is a single well-defined surface protein, a subunit or VLP vaccine gives a very clean, safe product, as for hepatitis B and HPV. Where speed matters, against a pandemic or a rapidly drifting virus, the gene-based platforms win because a new construct can be made from sequence within weeks. And where a strong T cell response is needed, the platforms that make antigen inside the cell (live, viral-vectored and nucleic acid) have the advantage over killed and subunit vaccines.

The single most important practical consequence is the contraindication that follows from a vaccine being alive.

Vaccine	Live?	Pregnancy	Significant immunocompromise	Other key caution
Measles, mumps, rubella (MMR)	Yes	Contraindicated	Contraindicated	Defer after immunoglobulin; gelatin or neomycin allergy
Varicella and live zoster	Yes	Contraindicated	Contraindicated	Avoid salicylates after varicella vaccine
Recombinant zoster	No	Permitted	Permitted (preferred)	Reactogenic
Yellow fever (17D)	Yes	Avoid (use only if unavoidable)	Contraindicated	Egg allergy; first dose carries rare serious reactions
Oral polio (OPV)	Yes	Avoid	Contraindicated	Sheds in stool; use inactivated polio instead
Rotavirus (oral)	Yes	Not applicable	Caution or avoid	Age window; prior intussusception
Live attenuated (intranasal) influenza	Yes	Contraindicated	Contraindicated	Use inactivated influenza instead
Inactivated influenza	No	Recommended	Permitted	Reduced response if very immunosuppressed
Hepatitis B, HPV	No	Permitted (HPV usually deferred)	Permitted	None major

The reason a live vaccine is dangerous in these groups is that the attenuated virus still replicates, and a fetus or a person without competent T cells may be unable to control even a weakened virus, allowing disseminated vaccine-strain disease. Live vaccines should be given at least four weeks before planned immunosuppression, and household contacts of immunocompromised patients can be protected by vaccinating those around them.

Producing the seasonal influenza vaccine

Influenza is the standing example of a vaccine that must be remade every year, because the virus continually changes its surface haemagglutinin (HA) and neuraminidase (NA). Twice a year, once for each hemisphere, the World Health Organization (WHO) reviews global surveillance data and recommends the strains the coming season’s vaccine should contain, typically an A/H1N1, an A/H3N2 and one or two B lineages.

The conventional vaccine is grown in embryonated hen eggs, the method in use since the 1940s. Because wild influenza strains often grow poorly in eggs, a high-yield reassortant is made by combining the surface genes of the recommended strain with the internal genes of a laboratory strain that grows well. The reliance on eggs has two drawbacks. It is slow, requiring a large egg supply months ahead, which limits how fast the vaccine can be updated; and the virus can acquire egg-adaptation mutations in haemagglutinin as it grows, subtly changing the antigen so that the vaccine matches the circulating virus less well. Two newer methods avoid eggs: cell-based production grows the virus in cultured mammalian cells, and recombinant production makes the haemagglutinin protein directly in insect cells, removing egg adaptation altogether. The mRNA platform represents the next step, allowing a strain change to be implemented in weeks rather than months, which is the central rationale for developing mRNA influenza vaccines.

New and future vaccine technology

Several advances are reshaping what vaccines can do. Structure-based antigen design uses the atomic structure of a viral protein to engineer a better immunogen: stabilising the RSV fusion protein in its pre-fusion shape turned a decades-long failure into a licensed vaccine, and displaying conserved fragments of haemagglutinin on a protein nanoparticle is one route toward a broadly protective influenza vaccine. Computational tools, including protein-structure prediction and consensus-antigen design, increasingly let immunogens be built to order rather than found by trial. Self-amplifying RNA vaccines extend the mRNA idea by encoding a replicase so that less RNA is needed for the same effect. Monoclonal antibodies, which deliver ready-made protection rather than training the immune system, increasingly sit alongside vaccines, and are treated in the passive-immunisation article.

Two harder problems remain active frontiers. Therapeutic vaccines aim to treat established infection rather than prevent it, with the most progress against HPV-associated precancer; success is harder for viruses that persist or damage immunity. And several major viruses still have no licensed vaccine: CMV, herpes simplex virus (HSV), human immunodeficiency virus (HIV) and hepatitis C virus (HCV) have each resisted decades of effort, for reasons ranging from immune evasion and latency to extreme antigenic variability, and each remains a leading target of current vaccine research.

References and recommended reading

• Crowe JE Jr. Immunization Against Viral Diseases. In: Fields Virology, 7th edition (Fundamentals), Chapter 15. Wolters Kluwer; 2023. The current, comprehensive reference for vaccine immunology, platforms and rational design.
• Burrell CJ, Howard CR, Murphy FA. Vaccines and Vaccination. In: Fenner and White’s Medical Virology, 5th edition, Chapter 11. Academic Press / Elsevier; 2017. Concise foundational account of vaccine types, adjuvants and contraindications.
• Ledgerwood JE, Graham BS. Immunization Against Viral Diseases. In: Richman DD, Whitley RJ, Hayden FG, editors. Clinical Virology, 4th edition, Chapter 17. ASM Press; 2016. Foundational reference for correlates of protection and the principles of vaccination.