The Molecule Nobody Wanted: How mRNA Spent 60 Years Becoming a Vaccine

On December 8, 2020, a 90-year-old grandmother named Margaret Keenan rolled up her sleeve at a hospital in Coventry, England, and became the first person to receive an authorized mRNA vaccine outside a clinical trial.  To most of the world, the technology seemed to have appeared out of nowhere.  The virus behind COVID-19 had been identified less than a year earlier, and suddenly there were shots built on a kind of medicine that had never been approved for anything.

In fact, the vaccine in that syringe was the end of a story that began in 1960.  It ran through a Cambridge lab playing with soap-bubble-like blobs of fat, a graduate student’s late-night experiment in California, a Hungarian biochemist who was demoted for refusing to give up, and a chance meeting at a photocopier.  For most of those six decades, messenger RNA, or mRNA, was the molecule nobody wanted.  Here is how it went from a fragile lab curiosity to the fastest vaccine ever made, and a Nobel Prize.

A messenger that falls apart

Your DNA is a library of recipes that never leaves the nucleus of the cell.  To make a protein, the cell copies one recipe onto a short-lived strip of RNA and sends it out to the ribosomes, the machines that build proteins.  That disposable copy is messenger RNA.

The idea came up in April 1960, in a conversation in Cambridge, England, between Sydney Brenner, Francis Crick and the French biologist François Jacob.  That summer, Brenner and Jacob traveled to the California Institute of Technology and, with Matthew Meselson, ran the experiment that showed the messenger was real.  A team at Harvard that included James Watson reached the same conclusion, and in May 1961 the two groups published side by side in the journal Nature.

From the start, mRNA had one defining trait: it does not last.  Cells are full of enzymes that chop RNA up within minutes or hours, which is exactly what you want from a message that should stop once the job is done.  It is also why, for decades, few people imagined injecting it as a medicine.  Anything you put in a syringe would likely be destroyed long before it reached a cell.

Bubbles of fat in a Cambridge lab

The solution to that problem began somewhere completely different.  In 1961, at the Babraham Institute near Cambridge, a British researcher named Alec Bangham was testing a new electron microscope.  When he and a colleague looked at dried fat molecules called phospholipids mixed with water, they saw that the molecules had curled up into tiny hollow spheres.

Phospholipids have a head that likes water and two tails that avoid it.  In water, they line up tail to tail in a double layer, the same structure that forms the outer skin of every living cell, and that layer closes into a ball with a little water trapped inside.  Bangham and his colleagues described these spheres in a landmark 1965 paper.  An American physician, Gerald Weissmann, later named them liposomes, from the Greek for “fat body.”

Researchers soon realized that liposomes could act as tiny delivery trucks, wrapping a drug inside a shell the body might accept.  In 1978, two groups showed that liposomes could carry mRNA into cells, which then made the protein it coded for.  It was an early hint, but the methods were clumsy and the field did not take off.

A molecular stew in California

In late 1987, Robert Malone, a graduate student at the Salk Institute in San Diego, mixed strands of mRNA with droplets of a positively charged fat developed by the chemist Philip Felgner.  The fat helped, because mRNA carries a negative charge, just like the outer surface of a cell, so on its own it is pushed away.  Human cells bathed in the mixture absorbed the mRNA and started making protein from it.  The results, published in 1989, made researchers wonder whether RNA could be used as a drug.

A year later, Jon Wolff at the University of Wisconsin, working with Felgner, injected bare mRNA and DNA into the leg muscles of mice.  The muscle cells made the encoded proteins.  In 1993, a French team used liposomes packed with mRNA for an influenza protein to trigger an immune response in mice.  On paper, the concept of an mRNA vaccine was born.

Then it stalled.  Making mRNA in the lab was expensive.  It broke down easily.  And, most worrying, injected mRNA set off inflammation, as if the body had detected an invader.  Many companies decided DNA was the more practical molecule and moved on.

The scientist who would not let go

Katalin Karikó grew up in a small town in Hungary, in a house without running water or a refrigerator.  She earned a PhD in biochemistry at the University of Szeged and began working on RNA.  In 1985, her lab’s funding ran out, and she left for the United States with her husband and two-year-old daughter.  The family smuggled out the money from selling their car by sewing it into the girl’s teddy bear.

By 1989 she was at the University of Pennsylvania, convinced that mRNA could be used to make cells produce healing proteins.  Grant after grant was rejected.  In 1995, the university demoted her and cut her pay.  She stayed anyway, often relying on whoever would let her work in their lab.  She would never be granted tenure at Penn.

Her stubbornness is a familiar pattern in medicine.  Like the early champions of phage therapy, or the engineers behind the first implantable pacemakers, she kept pushing an idea that most of her field had quietly written off.

A chance meeting at the photocopier

In 1997, an immunologist named Drew Weissman arrived at Penn.  He wanted to build better vaccines, including one against HIV, and was interested in dendritic cells, the immune system’s scouts.  The two met while waiting at a photocopier, started talking about their work, and found that each had what the other needed.  Karikó could make mRNA. Weissman knew how to test what the immune system thought of it, and he had funding.

The problem they kept running into was the same one that had stalled the field.  When they gave lab-made mRNA to dendritic cells, the cells reacted as if they had found a virus, pouring out inflammatory signals.  In a living animal, that kind of reaction could make a treatment dangerous and could also stop the cells from making the protein at all.

The clue hiding in transfer RNA

The breakthrough came from a control experiment.  Cells contain other kinds of RNA, including transfer RNA, the small adapters that carry amino acids to the ribosome.  When Karikó and Weissman tested transfer RNA, it did not set off the alarm.  Why would the immune system ignore one kind of RNA and attack another?

The answer was in the letters.  RNA is written in four chemical bases, abbreviated A, U, G and C. In our own cells, many of those bases carry small chemical decorations, and transfer RNA is especially heavily decorated.  RNA made in a test tube has none.  Karikó and Weissman suspected that the immune system uses those decorations to tell “self” RNA from foreign RNA, such as a virus’s.

So they made mRNA with modified bases, one version at a time.  The most dramatic result came from swapping uridine, the U, for a natural cousin called pseudouridine.  The inflammatory response almost disappeared.  Nature and Science both turned the paper down.  It was published in 2005 in the journal Immunity and attracted little attention.  Follow-up studies in 2008 and 2010 showed something just as important: modified mRNA did not just stay quiet, it also produced far more protein.

Building a better bubble

Quieting the mRNA was half the job.  It still needed a way into cells.  The answer came from a separate line of work led by Pieter Cullis, a biochemist at the University of British Columbia in Vancouver, and the companies he co-founded.  Instead of the permanently charged fats of the 1980s, which were harsh on cells, his teams developed “ionizable” lipids that are nearly neutral in the bloodstream but turn positive in the acidic pouches inside a cell, which helps release their cargo.

Modern lipid nanoparticles mix four ingredients: an ionizable lipid that grips the RNA, cholesterol and a helper lipid that give the particle its structure, and a coating of PEG-lipid that keeps particles from clumping.  In 2018, that design reached patients for the first time in Onpattro, a drug that uses a different kind of RNA to silence a faulty gene in the liver.  Weissman’s lab tested modified mRNA inside lipid nanoparticles from Acuitas Therapeutics, a Vancouver company co-founded by Cullis, and found that cells kept making the protein for days.

These particles are still not perfect.  They can stir up inflammation of their own, which is why researchers are now redesigning the fats themselves, as in a new kind of lipid nanoparticle that flips its electrical charge to keep the immune system calmer.

The companies that bet on it

Slowly, the business world caught up.  In Germany, the physician-scientists Uğur Şahin and Özlem Türeci founded BioNTech in 2008 to develop mRNA treatments for cancer.  In 2010, a group of scientists and investors in Boston founded Moderna, a name built from “modified RNA,” after a Harvard researcher used modified mRNA to reprogram human cells.  Both eventually licensed Karikó and Weissman’s modified-RNA patents, which Penn had earlier licensed to a small company for a few hundred thousand dollars.  In 2013, Karikó left Penn to join BioNTech.

The U.S. military’s research agency, DARPA, also began funding work on RNA vaccines that could be made quickly against new threats.  In 2013, the first human trial of an mRNA vaccine against an infectious disease began.  It targeted rabies. Over the next few years came trials against influenza, Zika and other viruses.  None had reached the market when 2020 began.

Sixty-three days

On January 11, 2020, scientists in China and Australia posted the genetic sequence of the new coronavirus online.  Within days, researchers at Moderna and the U.S. National Institutes of Health had designed their vaccine on a computer.  They did not need the virus itself, just its code.  On March 16, a volunteer in Seattle received the first dose in a clinical trial, about 63 days after the sequence was released.  BioNTech, which soon partnered with Pfizer, moved just as fast.

Both vaccines carried instructions for the virus’s spike protein, with two small tweaks, borrowed from earlier work on the related MERS virus at the NIH and the University of Texas at Austin, that hold the spike in the shape antibodies need to recognize.  Both used a refined version of Karikó and Weissman’s trick, replacing every U with N1-methylpseudouridine.  And both were packaged in lipid nanoparticles.  In November 2020, the companies reported that their vaccines were about 95 percent effective at preventing symptomatic COVID-19 in large trials.

Modified bases mattered.  A third mRNA vaccine, from the German company CureVac, used unmodified mRNA and in 2021 reported only about 48 percent effectiveness, a result many experts linked partly to that design choice.

Regulators in the United Kingdom authorized the Pfizer-BioNTech vaccine on December 2, 2020, and Margaret Keenan got her shot six days later.  The U.S. Food and Drug Administration authorized the Pfizer-BioNTech vaccine on December 11 and Moderna’s a week later.  The fragile nature of mRNA still showed.  The Pfizer-BioNTech vaccine originally had to be shipped at about minus 70 degrees Celsius (minus 94 degrees Fahrenheit), which sent hospitals scrambling for ultracold freezers.

A Nobel Prize, and what comes next

On October 2, 2023, Karikó and Weissman were awarded 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 Nobel committee noted that the vaccines, together with others made using different methods, had been given more than 13 billion times worldwide and saved millions of lives.

The technology has kept moving.  In 2024, the U.S. approved its first mRNA vaccine for a disease other than COVID-19, Moderna’s shot against respiratory syncytial virus (RSV) for older adults.  Flu and combined flu-and-COVID vaccines have followed in some countries.  The biggest hopes may lie in cancer.  BioNTech was founded to make personalized cancer vaccines, and trials are now testing shots designed around the mutations in each patient’s own tumor.  Beyond vaccines, researchers want to use mRNA to replace missing proteins in rare genetic diseases, an idea that brings Karikó’s original dream full circle.

It is tempting to call the mRNA vaccines an overnight success.  They were anything but.  They needed the discovery of the messenger in 1961, Bangham’s fat bubbles, a stew of lipids and RNA in 1987, a clue from transfer RNA, years of lipid chemistry in Vancouver and companies willing to bet on an unproven idea.  Most of all, they needed a few people who kept working on a molecule nobody wanted, long after it would have been easier to stop.

Explore the molecules of life, and the race for a vaccine