Every vaccine before 2020 worked roughly the same way: you put the thing you want the immune system to recognise into the body. A killed virus, a weakened virus, or a purified protein grown in a factory. The immune system meets it, learns it, and remembers.
An mRNA vaccine does something stranger. It puts in no antigen at all. It delivers a short, chemically modified instruction — a message written in the same language your cells use every second of every day — and your own cells build the target protein themselves. Then they break the message down and get rid of it.
That single design change explains almost everything people find confusing about these vaccines: why they were designed in days rather than years, why they needed freezers, why they cannot touch your DNA, and why the same platform has now produced an approved flu vaccine and a cancer therapy that passed a Phase 3 trial three days ago. Here is what actually happens.
Editor's note: This is an educational explainer, not medical advice. Decisions about any vaccine or cancer therapy belong with a qualified clinician who knows your history. The mechanism described below is well established; the sections on cancer and policy describe active, fast-moving areas and are sourced at the end.
Shipping the Recipe, Not the Meal
The useful analogy is a kitchen. A traditional vaccine ships you a finished dish — grown, cooked, inactivated, purified, bottled. An mRNA vaccine ships a recipe card and trusts that you already own a kitchen.
You do. Every cell in your body runs the same three-step process constantly: DNA is transcribed into messenger RNA, and ribosomes read that mRNA to assemble proteins. Messenger RNA is the courier in the middle. It is designed by evolution to be temporary — a working note, not an archive.
An mRNA vaccine exploits that existing machinery. It supplies one synthetic courier carrying instructions for a single protein: the SARS-CoV-2 spike protein, an RSV surface protein, an influenza protein, or — as we'll come to — a set of mutated proteins unique to one person's tumour.
| Inactivated virus | Protein subunit | mRNA | |
|---|---|---|---|
| What is injected | The pathogen, killed | The finished antigen protein | Instructions for the antigen |
| Where the antigen is made | In a production facility, before bottling | In a production facility, before bottling | Inside your own cells, after injection |
| To retool for a new target | Grow and inactivate a new pathogen | Develop and purify a new protein | Change the sequence; the process is unchanged |
That last row is the commercially and strategically important one, and we'll return to it.
What Actually Happens After the Injection
The sequence is short and it ends in disposal:
- The injection lands in muscle tissue. The mRNA is not floating free — it is packaged inside a microscopic fat bubble, of which more below.
- Cells take the package in. The bubble is absorbed, and the mRNA is released into the cytoplasm — the fluid outside the nucleus, where protein-making happens.
- A ribosome reads it. The cell's own protein factory latches on and translates the sequence exactly as it would translate one of your own messages. It cannot tell the difference and does not need to.
- The target protein is built. One harmless fragment — not a virus, not anything that can replicate.
- The protein is displayed. Fragments end up on the cell surface and in the surrounding tissue, where immune cells encounter them.
- The immune system responds. It generates antibodies and trains T cells against a threat that was never actually present.
Then the message is degraded. The CDC's own description is blunt on this point: after the protein is made, "our cells break down the mRNA and remove it, leaving the body as waste." The mRNA is a consumable, not a deposit.
Why It Cannot Change Your DNA
This is the question people actually want answered, so it deserves a direct answer rather than reassurance.
Your DNA is stored in the nucleus, a membrane-bound compartment with controlled entry. Vaccine mRNA is delivered into the cytoplasm and read there. It has no mechanism for entering the nucleus and, as the CDC states plainly, the material "never enters the nucleus of your cells, which is where your DNA is kept."
There is a second, quieter reason worth knowing. DNA and RNA are different molecules, and the cell's normal flow of information runs one way: DNA to RNA to protein. Turning an RNA message back into DNA and inserting it into a chromosome is not a step in that pipeline; it requires machinery the vaccine does not supply and a location the mRNA does not reach.
How long does it stay? Most of it is gone within days — mRNA is chemically fragile by nature, which is the same property that forces the cold chain. Some studies using highly sensitive assays have detected trace amounts of vaccine mRNA in the draining lymph nodes for longer than that, and the exact clearance timeline is still being characterised in the literature. That is a legitimate open research question. It is a separate question from nuclear entry, and it does not change the DNA answer.
The Twenty-Year Problem Nobody Could Solve
The idea of using mRNA as a medicine is old. The reason it took decades to arrive is that, for a long time, it simply did not work in a body.
Synthetic RNA injected into an animal triggered a fierce innate immune reaction. Your cells carry sensors — Toll-like receptors among them — whose job is to detect foreign RNA, because foreign RNA usually means a virus. Those sensors fired, inflammation followed, the mRNA was destroyed, and barely any of the intended protein was made. It was the worst of both outcomes: too inflammatory to be safe, too feeble to be useful.
Katalin Karikó and Drew Weissman solved it by changing the chemistry of the letters. In work published in Immunity in 2005, they showed that swapping one of RNA's building blocks for a naturally occurring modified version stopped the immune sensors from reacting. A 2008 follow-up in Molecular Therapy showed the modified mRNA also produced substantially more protein — the reaction that had been destroying the message had also been suppressing translation.
Both licensed COVID-19 mRNA vaccines rest on this. In each, every uridine is replaced with N1-methylpseudouridine, a modified nucleoside that keeps the message legible to ribosomes while making it nearly invisible to the sensors looking for viral RNA.
The reception at the time is worth recording. The 2005 manuscript was rejected by both Nature and Science, with reviewers judging the work "not novel" and "not of interest to the broad readership." In October 2023, Karikó and Weissman received the Nobel Prize in Physiology or Medicine "for their discoveries concerning nucleoside base modifications that enabled the development of effective mRNA vaccines against COVID-19."
The Other Half of the Invention
The modified mRNA still had a delivery problem. Naked RNA injected into muscle is degraded almost immediately by enzymes, and it cannot cross a cell membrane on its own — both molecules carry a negative charge and repel each other.
The answer is the lipid nanoparticle, and it is genuinely half the invention. It is a fat droplet, tens of nanometres across, built from four components working together: an ionizable lipid that grabs the RNA and helps release it inside the cell, a phospholipid and cholesterol that give the particle a stable structure, and a PEG-lipid on the outside that stops the particles clumping together.
This is also where the freezers came from. The dominant cause of instability in an mRNA-lipid nanoparticle product is hydrolysis of the mRNA itself — the message breaking apart in water. Comirnaty originally shipped at −80 °C to −60 °C for exactly that reason, and storage requirements have relaxed since as formulations improved. It was never a sign that something exotic was going on. It was a chemically fragile molecule being kept cold, for the same reason you freeze fish rather than leaving it on the counter.
Why the Platform Is So Fast
Here is the number that changed how governments think about pandemic preparedness.
The genetic sequence for Moderna's COVID-19 vaccine candidate was finalised with the NIH's Vaccine Research Center on 13 January 2020. The first human being received a dose in the NIH-led Phase 1 trial on 16 March 2020. Sixty-three days.
That was not a shortcut through safety testing; the trials that followed ran their full course. It was possible because nothing in the manufacturing process is virus-specific. There was no new pathogen to grow, no new cell culture to validate, no new protein to purify. A DNA template is printed from a digital sequence, transcribed into mRNA, wrapped in lipid, and filled. Swap the sequence and the machinery is unchanged.
That is what people mean by calling mRNA a platform. It is the same reason CRISPR gene editing became consequential so quickly: both technologies reduced a hard biological problem to a programmable one, where the expensive, slow part gets built once and the target becomes an input. Our biology coverage keeps returning to this pattern, because it is the through-line of the last decade in the life sciences.
What the Trials Actually Showed
Two Phase 3 results published in December 2020 established the platform clinically.
- BNT162b2 (Pfizer-BioNTech): 95% efficacy against symptomatic COVID-19, based on 8 cases among vaccine recipients versus 162 among placebo recipients across 36,523 participants with no evidence of prior infection.
- mRNA-1273 (Moderna): 94.1% efficacy in the 30,420-participant COVE trial across 99 US sites — 11 cases versus 185. All 30 cases of severe COVID-19 in that analysis occurred in the placebo group.
Two honest caveats belong next to those numbers. First, they were measured against the original virus, in a mostly unexposed population, weeks after a second dose — protection against infection by later variants was considerably lower, which is a fact about a mutating virus and waning antibody levels rather than a failure of the platform. Second, protection against severe disease has consistently held up far better than protection against catching it at all. Conflating the two is the single most common error in arguments about these vaccines in both directions.
Comirnaty became the first fully FDA-approved COVID-19 vaccine on 23 August 2021; Spikevax followed on 31 January 2022.
The Trade-offs
No platform is free. The honest ledger:
| Strength | The cost that comes with it |
|---|---|
| Retooling needs only a new sequence | The product is a fragile molecule, so the cold chain is harder and more expensive than for a protein vaccine |
| Very high protein expression from a small dose | Reactogenicity is real — sore arms, fatigue, headache and muscle ache are common in the day or two after a dose |
| No pathogen has to be grown or handled | Manufacturing depends on specialised lipids and equipment, which concentrates supply in relatively few facilities |
| Antigen is made inside your own cells, which recruits strong T-cell responses | Antibody levels wane, so protection against mild infection is time-limited even when protection against severe disease persists |
The reactogenicity point deserves its due, because it is where honest reporting and public perception diverge most. When the FDA approved Moderna's mRNA influenza vaccine this month, the trial data showed solicited adverse events were more frequent than with existing flu vaccines — injection-site pain, fatigue, headache and muscle ache — while being mostly mild to moderate and short-lived. Stronger immune stimulation and a rougher next morning are not separate phenomena.
From Infections to Cancer
The platform's second act is the more interesting one, and it is not really a vaccine in the everyday sense.
A cancer therapy of this type does not prevent disease; it treats existing disease by teaching the immune system to hunt something already present. The targets are neoantigens — what the National Cancer Institute describes as "abnormal proteins, or neoantigens, produced by cancer cells." Because those proteins are not found on healthy cells, they are targets the immune system can be pointed at without collateral damage.
The process is genuinely bespoke:
- A sample of the patient's tumour is sequenced, and its mutations are catalogued.
- Computer algorithms rank which of those mutations produce fragments most likely to be visibly presented to T cells. This is a prediction problem, and the tools that do it are now largely neural networks trained on which peptides are actually presented — the same broad family of methods covered in our guide to machine learning.
- The top-ranked candidates are encoded into a single mRNA molecule and manufactured. The NCI puts the turnaround at one to two months after tissue collection.
On 19 August 2026, that approach cleared its highest bar yet. Merck and Moderna announced that the Phase 3 INTerpath-001 trial met its primary endpoint of recurrence-free survival and its key secondary endpoint of distant metastasis-free survival. The trial randomised 1,137 patients with completely resected stage IIB–IV melanoma to receive either intismeran autogene — an individualised therapy encoding up to 34 neoantigens — plus pembrolizumab, or pembrolizumab alone.
Two pieces of restraint are warranted. The companies released topline results only: they confirmed the endpoints were met without publishing hazard ratios or survival curves, and said they will present the data at a medical meeting and engage regulators. Until that data appears, the size of the benefit is unknown. And the earlier Phase 2b trial that motivated this one, KEYNOTE-942, reported a 49% reduction in the risk of recurrence or death (HR 0.51; 95% CI 0.294–0.887) — a wide confidence interval from a small randomised study, which is precisely why a Phase 3 was needed. As of the NCI's guidance, no mRNA cancer vaccine has yet been approved by the FDA.
Still, the direction is clear, and it is the platform argument made concrete: the same manufacturing line, pointed at a target that exists in exactly one patient.
The 2026 Picture: Scientific Expansion, Political Retreat
Two things are happening at once, and following only one of them gives a badly distorted view.
The science is broadening. Moderna's RSV vaccine mRESVIA was approved on 31 May 2024 on the strength of the ConquerRSV trial in roughly 37,000 adults across 22 countries. On 5 August 2026, the FDA approved mFLUSIVA (mRNA-1010) for adults aged 50 and over — the first mRNA-based influenza vaccine ever licensed — following a Phase 3 trial of 40,805 adults in 11 countries that showed 26.6% relative efficacy against a standard-dose comparator, with the indication for adults 65 and over granted under accelerated approval. Two weeks later came the melanoma result.
US federal support has moved the other way. In early August 2025, the Department of Health and Human Services terminated 22 BARDA-funded mRNA vaccine development projects worth close to $500 million, including contracts with Moderna, Pfizer and CSL Seqirus. Secretary Robert F. Kennedy Jr. stated that the data show these vaccines "fail to protect effectively against upper respiratory infections like COVID and flu," and the department said it would redirect funding toward other vaccine platforms. Non-vaccine uses of mRNA within the department were not affected.
The decision drew sharp criticism from people who had run the relevant programmes. Rick Bright, a former BARDA director, said that "by dismantling that platform, we're crippling our front-line defense, just ahead of unknown biological threats." Paul Offit of Children's Hospital of Philadelphia called it "a giant step backward for science."
The distinction that matters for a reader: a funding decision is not an experimental result. The clinical record and the appropriations record are two different things, and in 2026 they are pointing in opposite directions.
Common Misconceptions
- "It was rushed." The deployment was fast. The underlying science was not — the enabling discovery dates to 2005 and spent fifteen years being refined before anyone had heard of SARS-CoV-2.
- "It's gene therapy." Gene therapy alters or replaces genes. An mRNA vaccine delivers a temporary message that never reaches your genes. CRISPR edits DNA; this does not.
- "The spike protein stays in your body." The mRNA that encodes it is degraded, and the protein along with it. Detecting trace residue with a sensitive assay is not the same as ongoing production.
- "mRNA vaccines don't work because people still got COVID." Efficacy against mild infection and efficacy against severe disease are different measurements. The trials measured the first and found the second held up better.
- "An mRNA cancer vaccine will stop me getting cancer." These are treatments for existing, already-diagnosed disease, built from that specific tumour. They are not preventive.
The Bottom Line
An mRNA vaccine is a delivery of information, not of substance. A fragile chemical message, chemically modified so your immune sensors ignore it, wrapped in a lipid bubble so it survives long enough to be read, is translated once by your own ribosomes into a protein your immune system then learns. Then it is gone.
Everything else follows from that. The speed follows from it, because only the sequence needs to change. The freezers follow from it, because messages are fragile. The reactogenicity follows from it, because strong expression means a strong response. And the expansion into RSV, influenza and now individualised cancer therapy follows from it too — the platform was never about one virus.
The most interesting fact about mRNA in 2026 is not the COVID vaccine at all. It is that a manufacturing process built for a pandemic turned out to be a general-purpose way of telling the immune system what to look for. Follow that thread through our medicine research coverage and it keeps going.
Frequently Asked Questions
Do mRNA vaccines change your DNA?
No. Your DNA sits inside the cell nucleus, and vaccine mRNA is delivered into the cytoplasm outside it. The CDC states that the material never enters the nucleus where DNA is kept. There is also no step in normal cellular biology that converts an mRNA message back into DNA and inserts it into a chromosome — the flow of information runs DNA to RNA to protein, not the reverse. The mRNA is read and then destroyed.
How long does the mRNA stay in my body?
Most of it is gone within days. Messenger RNA is chemically unstable by design — that is the same property that forced the original ultra-cold storage. Studies using very sensitive detection methods have found trace amounts persisting in the draining lymph nodes for longer, and the precise clearance timeline is still an active research topic. Detecting residue is not evidence of continued protein production, and it has no bearing on the DNA question.
Why did the first mRNA vaccines need such extreme cold storage?
Because the mRNA breaks down in water. Hydrolysis of the message is the main driver of instability in an mRNA-lipid nanoparticle product, so Comirnaty originally shipped at −80 °C to −60 °C to slow that chemistry down. Nothing about the requirement was mysterious; storage conditions have since relaxed as formulations improved. It was fragility, not exoticism.
Is an mRNA cancer therapy the same thing as a COVID vaccine?
Only in the delivery mechanism. A COVID vaccine is preventive and identical for everyone. An individualised cancer therapy is a treatment for disease that already exists, and it is manufactured from that one patient's tumour after sequencing it and predicting which mutated proteins the immune system can see. The NCI puts production at one to two months per patient. No mRNA cancer therapy has been FDA-approved yet.
Why do mRNA shots often feel rougher than a flu shot?
Because they generate strong protein expression from a small dose, which produces a correspondingly strong immune response. Sore arm, fatigue, headache and muscle ache in the following day or two are common. When the FDA approved Moderna's mRNA flu vaccine in August 2026, the trial data showed these effects were more frequent than with existing flu vaccines, though mostly mild to moderate and short-lived. The discomfort and the immune response are the same event.
Was the COVID vaccine really developed in a year?
The deployment was. The enabling science was not. Karikó and Weissman's nucleoside-modification work was published in 2005 and rejected by two major journals before that; it took another fifteen years of refinement in lipid delivery and manufacturing before anything was ready to scale. What 2020 demonstrated was that once a platform exists, retargeting it is fast — 63 days from sequence to first human dose.
Can the same platform be used for the next pandemic?
Technically, yes — that is the strongest argument for it, and the reason the 63-day figure attracted so much attention. Practically, it depends on sustained manufacturing capacity and funding. The US terminated 22 BARDA-funded mRNA vaccine development projects in August 2025, which former officials publicly argued would weaken exactly that capability. The technical readiness and the institutional readiness are separate questions.
Sources
- The Nobel Prize in Physiology or Medicine 2023: advanced information — The Royal Swedish Academy of Sciences
- Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine — Polack et al., New England Journal of Medicine (2020)
- Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine — Baden et al., New England Journal of Medicine (2020)
- mRNA-lipid nanoparticle COVID-19 vaccines: structure and stability — Schoenmaker et al., International Journal of Pharmaceutics (2021)
- Can mRNA Vaccines Help Treat Cancer? — National Cancer Institute
- Phase 3 INTerpath-001 trial of intismeran autogene plus KEYTRUDA met endpoints of RFS and DMFS in resected stage IIB–IV melanoma — Merck and Moderna (19 August 2026)
- FDA approves Moderna's mRNA flu vaccine — Healthcare Dive (August 2026)
- HHS abandons mRNA vaccine research — BioPharma Dive (August 2025)
