How vaccines actually work, and what the different platforms do

Every vaccine, whatever its technology, does one thing: it shows the immune system a distinctive piece of a pathogen — an antigen — under safe conditions, so that the adaptive immune system can build a memory of it. A review in Nature Reviews Immunology describes the core mechanism plainly. The immune response to a vaccine generates antibody-producing B cells and T cells specific to the antigen, and, crucially, long-lived memory cells; if the real pathogen arrives later, that memory lets the body respond faster and harder than it could on a first, naïve encounter [s1]. The point of a vaccine is to buy that first encounter without paying the price of the disease.

What separates the platforms is not the goal but the method of presentation — how they get the antigen in front of the immune system, and how closely they mimic a natural infection. Each choice trades off immune strength, safety, manufacturing speed, and stability.

Live-attenuated: the real thing, weakened

Live-attenuated vaccines use a whole pathogen that has been weakened so it replicates poorly and does not cause disease in healthy people [s1]. Because it briefly replicates, it presents the immune system with essentially the full natural target and tends to generate strong, durable immunity, often from few doses — the measles, mumps and rubella (MMR), yellow fever, and oral polio vaccines work this way [s1]. The trade-off is the flip side of the same feature: a live vaccine can cause disease in people with severely weakened immune systems and is generally avoided in that group, and it needs careful handling to stay viable [s1].

Inactivated and subunit: showing a fragment

If a whole live pathogen is one end of the spectrum, showing only a piece of it is the other. Inactivated vaccines use a pathogen that has been killed so it cannot replicate at all; subunit, recombinant and conjugate vaccines go further and use only selected parts — a purified protein, a surface sugar, or a protein made in the lab rather than grown from the pathogen [s1]. These are very safe and stable, which is why they dominate routine schedules: inactivated polio, hepatitis A and B, the HPV vaccine, and the pneumococcal and Hib conjugate vaccines are all of this family [s1]. The cost is that a fragment is less immunogenic than a live infection, so these vaccines often need an adjuvant — an ingredient that boosts the immune response — and booster doses to reach and maintain protection [s1]. Toxoid vaccines, such as tetanus and diphtheria, are a related idea: they present an inactivated bacterial toxin so the body learns to neutralise it [s1].

Viral vector: a courier virus

Viral-vector vaccines take the gene for a target antigen and insert it into a harmless or weakened carrier virus, which delivers those instructions into cells so the body manufactures the antigen itself and mounts a response to it [s1]. The approach reached the public at scale with some COVID-19 vaccines and with the rVSV-based Ebola vaccine [s1]. Its limitations include the possibility that prior immunity to the carrier virus blunts the response [s1].

mRNA: instructions, not the antigen

The newest platform to reach billions of people delivers the genetic instructions for an antigen directly, as messenger RNA wrapped in lipid nanoparticles [s1]. The particles are taken up by cells, which read the mRNA and briefly produce the target protein — for the COVID-19 vaccines, the coronavirus spike — before the mRNA is degraded [s1]. The mRNA does not enter the cell nucleus and is not integrated into DNA; it is transient by design [s1]. Its advantages are speed and flexibility of manufacturing, since making a new mRNA is largely a matter of changing the sequence [s1].

The efficacy data from the pivotal trials were what made the platform's arrival so visible. In its phase 3 trial, the BNT162b2 (Pfizer-BioNTech) vaccine was 95% effective at preventing symptomatic COVID-19 relative to placebo [s2]. The mRNA-1273 (Moderna) vaccine was 94.1% effective in its own phase 3 trial [s3]. Those figures describe protection against symptomatic disease in those trials at that time, against the variants then circulating; efficacy against later variants and over longer follow-up is a separate and much-studied question, and neither number should be read as a permanent property of the vaccine.

Why the range exists

The reason there is no single best platform is that the requirements differ by disease and setting. A vaccine for a stable childhood target may favour the safety and shelf life of a subunit product; a response to a fast-moving new pathogen may favour the speed of mRNA; a disease where durable immunity from few doses matters most may favour a live-attenuated approach [s1]. The immunological destination — trained, remembered recognition of a specific threat — is the same. The engineering of how to get there is what the platform names describe.