In his 1972 Nobel Prize acceptance speech, American biochemist Christian Anfinsen laid out a vision: One day it would be possible, he said, to predict the 3D structure of any protein merely from its sequence of amino acid building blocks. With hundreds of thousands of proteins in the human body alone, such an advance would have vast applications, offering insights into basic biology and revealing promising new drug targets. Now, after nearly 50 years, researchers have shown that artificial intelligence (AI)-driven software can churn out accurate protein structures by the thousands—an advance that realizes Anfinsen’s dream and is Science’s 2021 Breakthrough of the Year.
Protein structures could once be determined only through painstaking lab analyses. But they can now be calculated, quickly, for tens of thousands of proteins, and for complexes of interacting proteins. “This is a sea change for structural biology,” says Gaetano Montelione, a structural biologist at Rensselaer Polytechnic Institute. David Baker, a University of Washington, Seattle, computational biochemist who led one of the prediction projects, adds that with the bounty of readily available structures, “All areas of computational and molecular biology will be transformed.” Read the full story…
(Illustration) V. Altounian/Science; (Data) I. R. Humphreys et al., Science 374, eabm4805 (2021); DOI: 10.1126/science.abm4805
2021 Breakthrough of the Year
Protein structures for all
AI-powered predictions show proteins finding their shapes
In his 1972 Nobel Prize acceptance speech, American biochemist Christian Anfinsen laid out a vision: One day it would be possible, he said, to predict the 3D structure of any protein merely from its sequence of amino acid building blocks. With hundreds of thousands of proteins in the human body alone, such an advance would have vast applications, offering insights into basic biology and revealing promising new drug targets. Now, after nearly 50 years, researchers have shown that artificial intelligence (AI)-driven software can churn out accurate protein structures by the thousands—an advance that realizes Anfinsen’s dream and is Science’s 2021 Breakthrough of the Year.
Protein structures could once be determined only through painstaking lab analyses. But they can now be calculated, quickly, for tens of thousands of proteins, and for complexes of interacting proteins. “This is a sea change for structural biology,” says Gaetano Montelione, a structural biologist at Rensselaer Polytechnic Institute. David Baker, a University of Washington, Seattle, computational biochemist who led one of the prediction projects, adds that with the bounty of readily available structures, “All areas of computational and molecular biology will be transformed.”
Proteins are biology’s workhorses. They contract our muscles, convert food into cellular energy, ferry oxygen in our blood, and fight microbial invaders. Yet despite their varied talents, all proteins start out with the same basic form: a linear chain of up to 20 different kinds of amino acids, strung together in a sequence encoded in our DNA. After being assembled in cellular factories called ribosomes, each chain folds into a unique, exquisitely complex 3D shape. Those shapes, which determine how proteins interact with other molecules, define their roles in the cell.
Work by Anfinsen and others suggested interactions between amino acids pull proteins into their final shapes. But given the sheer number of possible interactions between each individual link in the chain and all the others, even modest-size proteins could assume an astronomical number of possible shapes. In 1969, American molecular biologist Cyrus Levinthal calculated that it would take longer than the age of a universe for a protein chain to cycle through them one by one—even at a furious pace. But in nature, each protein reliably folds up into just one distinctive shape, usually in the blink of an eye.
In the 1950s, researchers started to map proteins’ 3D structures by analyzing how x-rays ricocheted off the molecules’ atoms. This technique, known as x-ray crystallography, soon became the leading approach; today, the field’s central repository, the Protein Data Bank, contains some 185,000 experimentally solved structures. But mapping structures can take years—and cost hundreds of thousands of dollars per protein. To speed the process, scientists started to create computer models in the 1970s to predict how a given protein would fold.
At first, that was possible only for small proteins or short segments of larger ones. By 1994, however, computer models had grown sophisticated enough to launch the biennial Critical Assessment of protein Structure Prediction (CASP) competition. Organizers gave modelers the amino acid sequences of dozens of proteins. At the end of the event, the modelers’ results were compared with the latest experimental data from x-ray crystallography and emerging techniques such as nuclear magnetic resonance spectroscopy and cryo–electron microscopy (cryo-EM). Scores above 90 were considered on par with experimentally solved structures.
Early results were humbling, with median scores below 60. But over time, the modelers learned tricks to improve their calculations. For example, stretches of amino acids shared by two proteins often fold similarly. If a protein with an unknown structure shares, say, 50% of its amino acid sequence with a protein that does have a known structure, the latter can serve as a “template” to guide the computer models.
Another major insight came from evolution. Investigators realized that if one amino acid changed in a protein shared by closely related organisms, like chimpanzees and humans, amino acids located nearby in the folded molecule would have to change, too, to preserve the protein’s shape and function. That means investigators can narrow down a protein’s shape by looking for amino acids that coevolve: Even if they are far apart on the unfolded chain, they are likely neighbors in the final 3D structure.
I never thought I’d see this in my lifetime.
- John Moult
- University of Maryland, Shady Grove
By 2018, the modelers were often scoring in the mid-70s. Then, AlphaFold, an AI-driven software program, entered the scene. The program, developed by Google sister company DeepMind, trains itself on databases of experimentally solved structures. In its first competition, its median score was close to 80, and it won 43 of 90 matches against other algorithms. In 2020, its successor, AlphaFold2, shone even brighter. Powered by a network of 182 processors optimized for machine learning, AlphaFold2 rang up a median score of 92.4—on par with experimental techniques.
“I never thought I’d see this in my lifetime,” John Moult, a structural biologist at the University of Maryland, Shady Grove, and CASP co-founder, said at the time.
This year, AI predictions shifted into overdrive. In mid-July, Baker and his colleagues reported that their AI program RoseTTAFold had solved the structures of hundreds of proteins, all from a class of common drug targets. A week later, DeepMind scientists reported they had done the same for 350,000 proteins found in the human body—44% of all known human proteins. In coming months, they expect their database will grow to 100 million proteins across all species, nearly half the total number believed to exist.
The next step is to predict which of those proteins work together and how they interact. DeepMind is already doing just that. In an October preprint, its scientists unveiled 4433 protein-protein complexes, revealing which proteins bind to one another—and how. In November, RoseTTAFold added another 912 complexes to the tally.
Code for AlphaFold2 and RoseTTAFold is now publicly available, helping other scientists jump into the game. In November, researchers in Germany and the United States used AlphaFold2 and cryo-EM to map the structure of the nuclear pore complex, an assembly of 30 different proteins that controls access to the cell nucleus. In August, Chinese researchers used AlphaFold2 to map the structures for nearly 200 proteins that bind to DNA, which could be involved in everything from DNA repair to gene expression. Last month, Google’s parent company, Alphabet, launched a new venture that will use predicted protein structures to design new drug candidates. And Baker’s team is using its software to dream up novel protein sequences that will fold into stable structures, an advance that could lead to new antivirals and catalysts.
Even now, scientists studying SARS-CoV-2 are using AlphaFold2 to model the effect of mutations in the Omicron variant’s spike protein. By inserting larger amino acids into the protein, the mutations have changed its shape—perhaps enough to keep antibodies from binding to it and neutralizing the virus.
Much work remains. Protein structures aren’t static; they bend and twist as they do their jobs, and modeling those changes remains a challenge. And it’s still a daunting task to visualize most of the large, multiprotein complexes that carry out myriad jobs in cells. But this year’s explosion of AI-driven advances offers a view of the dance of life as never seen before, a panorama that will forever change biology and medicine.
Runners-up
DNA from fossils has transformed the study of human and animal evolution, revealing unknown relationships, tracing early migrations, and exposing ancient interspecies mating. Yet for humans, the entire field depends on just 23 archaic genomes, 18 of them from Neanderthals. Recently, scientists unlocked a much larger trove of ancient DNA: from the soil of cave floors. This year, for the first time, cave dirt yielded DNA once housed in the nucleus of human cells, and researchers used such “dirt DNA” to reconstruct the identity of cave dwellers around the world.
The new work borrows from the study of environmental DNA from living species. To find out which organisms inhabit lakes, forests, and other places, scientists collect the free-floating DNA they shed into air, water, and soil. By 2003, evolutionary geneticists showed discarded DNA could persist for thousands of years. It was used by researchers in 2015 to help reconstruct entire ancient ecosystems, even in the absence of fossils.
But much of that DNA comes from mitochondria, the cell’s power plants, which store tiny snippets of genetic material. Thanks to new techniques, scientists can now comb ancient soils for nuclear DNA, which carries the full instructions for life.
This year, scientists used nuclear DNA to chart the human and animal occupation of three caves. In Spain’s Estatuas Cave, nuclear DNA revealed the genetic identity and sex of humans who lived there 80,000 to 113,000 years ago, and suggested one lineage of Neanderthals replaced several others after a glacial period that ended 100,000 years ago. In 25,000-year-old soil from Georgia’s Satsurblia Cave, scientists found a female human genome from a previously unknown line of Neanderthals, along with the genetic traces of a bison and a now-extinct wolf. And by comparing 12,000-year-old black bear DNA from Mexico’s Chiquihuite Cave with that of modern bears, scientists discovered that after the last ice age, the cave bears’ descendants migrated as far north as Alaska.
Techniques for extracting and sequencing nuclear DNA from ancient soils are still improving. As they do, researchers hope to answer even more questions about the rise and fall of ancient species.
Is fusion energy about to overcome its reputation as a field that promises the stars but never delivers? In an August result that surprised the researchers themselves, the U.S. National Ignition Facility (NIF) produced a fusion reaction that came tantalizingly close to reaching official “breakeven,” the point at which a reaction produces more energy than the laser energy needed to kindle it.
Fusion, which powers the Sun and other stars, has long been seen as a solution to Earth’s energy problems. But achieving the pressures and temperatures required—10 times as hot as the Sun’s core—is notoriously difficult. Many efforts cage a superhot plasma in a magnetic field; NIF uses a pulse from the world’s highest energy laser to compress a peppercorn-size capsule of the hydrogen isotopes deuterium and tritium. Earlier this year, that method generated 170 kilojoules of fusion energy per shot—far short of the laser input of 1.9 megajoules.
But in an 8 August shot, that yield surged to 1.35 megajoules. Researchers think it’s the result of a burning plasma, meaning the fusion reaction generated enough heat to spread through the compressed fuel like a flame. The result has not been scrutinized by a peer-reviewed journal, but it was presented at the American Physical Society’s Division of Plasma Physics conference in November.
Now, the team is trying to understand the shot’s high yield and figure out how to tweak starting conditions to do even better, by using larger or smoother fuel capsules, more even layers of frozen fuel, or higher quality laser pulses. They’re also making efforts to replicate the shot: One attempt in October reached 430 kilojoules, and a November shot hit 700 kilojoules. Attempts will continue into 2022.
As NIF edges toward breakeven, private fusion projects are upping the pace. Several predict they will generate energy long before the ITER reactor, a $25 billion publicly funded magnetic fusion effort. This year, Commonwealth Fusion Systems and Tokamak Energy claimed progress with high-temperature superconducting magnets. And General Fusion and TAE Technologies, which use pistons and particle beams, respectively, are planning energy-producing demonstration power plants they say will switch on in 2025.
Whichever approach reaches energy gain first, formidable challenges in materials science and engineering remain before fusion can become a practical power source.
Vaccines have had a starring role in the fight against COVID-19, but a new player is joining them on stage: antiviral pills that prevent symptoms and death if taken early in infection.
In the fall, drugmakers Pfizer and Merck & Co. touted positive clinical trial results via press release. More antivirals are in trials, and existing generic drugs—including the obsessive-compulsive disorder treatment fluvoxamine—may also prove useful. Merck’s antiviral, molnupiravir, reduces the risk of hospitalization or death by 30% in high-risk, unvaccinated individuals, according to final data submitted to regulators. (That figure is lower than the 50% from an interim analysis.) Pfizer’s antiviral, PF-07321332, reduces hospitalization by 89% if started within 3 days of symptoms. The United Kingdom approved molnupiravir in November, and an advisory body to the U.S. Food and Drug Administration has narrowly endorsed it; regulators are also considering Pfizer’s treatment.
Scientists are quick to stress that antivirals can’t replace vaccination. But they are still vital, and may become even more so if the new Omicron variant causes a surge in breakthrough infections. Pfizer is running another PF-07321332 trial in a group that includes vaccinated people, and both Merck and Pfizer are testing whether the drugs can head off illness in people who have recently been exposed.
Still, questions abound. Will the antivirals reduce transmission from infected people? Can lower income countries access the supply they need? (Both companies have pledged to sell the drugs in these places at a steep discount.) Will there be side effects not seen in clinical trials? Despite uncertainties, scientists and doctors are buoyed by the results—and pleased to have multiple therapies, which may help keep the virus from becoming resistant to any one.
The mind-altering power of psychedelic drugs has raised hopes that they can ease psychiatric disease, but few large, rigorous trials have shown they’re effective. This year brought a big win for the field: A multicenter, randomized, controlled trial found that 3,4-methylenedioxymethamphetamine (MDMA), popularly called ecstasy, significantly reduced symptoms in patients with post-traumatic stress disorder (PTSD).
The study, published in Nature Medicine in May, tested an intensive combination of talk therapy and MDMA, which can create a sense of well-being and empathy that may help people process traumatic experiences. The trial’s 76 participants had three 8-hour guided therapy sessions with either the drug or a placebo, plus a set of shorter “preparatory” and “integration” sessions with therapists before and after treatment. After 2 months, 67% of those who got MDMA no longer met the diagnostic criteria for PTSD, versus 32% in the placebo group.
The results prompted enthusiasm, but also caution. Such trials face a “Gordian knot of blinding and placebo effects,” neurologist Matthew Burke and psychiatrist Daniel Blumberger of the University of Toronto warned in Nature Medicine in October. That’s because MDMA’s psychoactive effects are obvious to participants and could influence their expectations—and even the odds they will improve. (Follow-ups with participants after the study suggested as many as 90% correctly guessed which group they were in.) And simply accepting that such expectations are part of the treatment “would require a complete overhaul of how we measure efficacy in psychiatry,” Burke and Blumberger write.
Still, psychedelic research is booming as academic labs and companies explore the potential of MDMA and other psychedelics to treat conditions like depression, anxiety, and addiction. In November, the London-based mental health care company COMPASS Pathways announced positive results from a 233-participant randomized trial of psilocybin, the substance in so-called magic mushrooms, in people with treatment-resistant depression. The company is now planning a larger trial. And if an ongoing follow-up study can confirm the initial MDMA results, its sponsor—the nonprofit Multidisciplinary Association for Psychedelic Studies—plans to seek approval from the U.S. Food and Drug Administration as early as 2023.
Artificial antibodies tame infectious diseases
by Jon Cohen
Labmade antibodies called monoclonals have revolutionized the treatment of some cancers and autoimmune diseases, but they’ve had limited success against infectious diseases. That changed this year, as monoclonal antibodies (mAbs) made inroads against SARS-CoV-2 and other life-threatening pathogens, including respiratory syncytial virus (RSV), HIV, and malaria parasites.
To make mAbs, scientists isolate the most powerful antibodies from lab animals and humans and reproduce them in massive quantities. As medicines, they are mostly used to tamp down immune responses or mark tumor cells for destruction. The only mAbs approved for infectious diseases in the United States are limited to rare maladies: Ebola, inhalational anthrax, recurrent Clostridium difficile, RSV in high-risk infants, and HIV in people for whom all drugs have failed. India has approved a mAb for rabies.
With advances in cloning, animal models, and x-ray crystallography, researchers can now make and screen more mAbs than ever before, simplifying their search. SARS-CoV-2 mAbs showed promising results in clinical trials in 2020, and by late this year, the U.S. Food and Drug Administration had granted emergency use authorization to three to treat COVID-19 and, in some cases, prevent infection.
Monoclonals are also being developed against influenza, Zika, and cytomegalovirus. High hopes surround two candidates designed to prevent RSV in all infants. And in one HIV prevention study that failed overall, the candidate worked well against some strains.
High costs and the need to infuse mAbs in a clinic have put them out of reach for many. But as prices plummet, injections replace infusions, and more potent mAbs come to market, they may become standard weapons in the infectious disease arsenal.
NASA lander uncovers the Red Planet’s core
by Paul Voosen
The interior of a rocky planet is a kind of time machine: Its dense core, viscous mantle, and hardened crust can reveal how it coalesced, churned, and settled into what it is today. Until this year, scientists have had access to just two such time capsules: Earth and, briefly during the Apollo missions, the Moon. Now, for the first time, instruments aboard NASA’s InSight lander are bringing Mars’s planetary core into focus.
When InSight arrived on the Red Planet in 2018, Mars seemed reluctant to expose its buried secrets. InSight’s heat probe failed to penetrate the planet’s surprisingly sticky sediments, despite repeated attempts. The lander’s hypersensitive seismic station, designed to monitor the underground rumblings that could help chart the planet’s interior, never picked up a marsquake powerful enough to do the job. And dust piled on the lander’s solar panels, slowly eroding their power output.
But within 1 year, InSight picked up a handful of moderate quakes, including several stemming from Cerberus Fossae, a fissured region 1600 kilometers away. When combined with estimates of the interior’s composition, those readings helped chart the planet’s depths. Offsets in the quakes’ seismic waves revealed that the martian crust is layered and less than 40 kilometers thick—thinner than Earth’s continental crust. That thin shell would have let Mars quickly shed its early internal heat.
Looking deeper, InSight found the martian mantle lacked the insulating lower layer seen in Earth’s. The mantle was also shallow, squeezed between the crust and an unusually large, liquid core that occupies more than half of Mars’s width. Given the planet’s mass, scientists concluded that the core’s density is low, and that a mixture of light elements such as sulfur likely keep its iron and nickel liquid, despite the planet’s rapid heat loss—much as salt prevents icing. The researchers published their findings in Science this year.
Armed with these new data, scientists will be puzzling over Mars’s history for years to come. Questions include whether the planet once had something resembling plate tectonics and when its liquid core stopped churning, shutting off the magnetic field it once generated.
InSight may have still more stories to share: In August and September, the lander heard its largest marsquakes yet. But as red dust continues to build on its solar panels, its time is growing short. It is estimated to run out of power by the end of 2022.
Until then, InSight will wait, and listen.
At last, a crack in particle physics’ standard model?
by Adrian Cho
It may be a sign of particle physicists’ desperation for something new that the biggest result in years confirms an oddity first observed 2 decades ago. A particle called the muon—a heavier, unstable cousin of the electron—is slightly more magnetic than physicists’ prevailing theory of fundamental particles and forces predicts. Reported in April, the 2.5-parts-per-billion discrepancy could signal new particles lurking just over the high energy horizon.
Developed in the 1960s and ’70s, the current theory, known as the standard model, accounts for three forces—electromagnetism, the strong nuclear force, and the weak nuclear force—and two dozen fundamental particles. It cannot be the final description of nature, as it leaves out both gravity and dark matter, the mysterious stuff thought to outweigh the universe’s ordinary matter. Yet, so far, the standard model accounts for every particle blasted into fleeting existence with high energy particle accelerators.
The muon’s magnetism gives scientists an indirect way to search for additional, undiscovered particles. Thanks to quantum uncertainty, empty space around the muon roils with particle-antiparticle pairs popping in and out of “virtual” existence too fast to be directly observed. Those in the standard model increase the muon’s magnetism by a precise amount. New particles could change that calculation in unpredicted ways.
To measure muon magnetism, scientists fire a beam of them into a magnetic field, where they twirl like compass needles at a rate that depends on how magnetic they are. Physicists ran just such an experiment from 1997 to 2001 at Brookhaven National Laboratory in New York, where they first detected the anomaly. In 2003, they hauled their 15-meter-wide magnet to Fermi National Accelerator Laboratory in Illinois, to obtain a purer muon beam. This year, they proved their previous result was not a fluke.
But a proper comparison depends on the precision of the standard model prediction. On the same day experimenters released their result, one team of theorists published a calculation that, they argued, increases the standard model prediction and closes the observed gap. Other physicists say the theoretical consensus still indicates that the muon is extra magnetic.
Now, the question is why. Other searches for tiny discrepancies from the standard model’s predictions could yield more clues to the hoped-for new physics. Or, if physicists are lucky, the world’s biggest atom smasher, Europe’s Large Hadron Collider, will blast some new particle into plain view when it comes back online next spring after 3 years of upgrades.
The gene-editing tool CRISPR had its first clinical victory in 2020, when it appeared to cure people with two inherited blood disorders, sickle cell disease and beta-thalassemia. Those treatments took place in a lab dish: Scientists removed defective blood stem cells from patients, edited them, and reinfused the cells into patients. This year, scientists took things one step further, deploying CRISPR directly in the body. In small studies, the strategy reduced a toxic liver protein and modestly improved vision in people with inherited blindness.
Gene editing could tackle many more diseases if the therapy could be injected into an organ or the bloodstream. But getting CRISPR to work inside a person, or in vivo, poses significant challenges. Before CRISPR’s molecular components can correctly modify a specific gene, they must be ferried safely to the right cells in the right quantities.
To fight hereditary transthyretin (TTR) amyloidosis, a disease in which a misfolded TTR protein builds up and damages nerves and the heart, researchers at Intellia Therapeutics and Regeneron Pharmaceuticals gave six patients an infusion of tiny fat balls encasing a guide RNA and the RNA instructions for CRISPR’s genome-snipping enzyme. The team hoped the patients’ own liver cells would take up the particles and make the CRISPR components, which would snip both strands of DNA at the TTR gene. The cell’s repair system would mend the cuts imperfectly, leaving the gene disabled. It worked: After 4 weeks, average blood levels of TTR dropped 52% or 87% depending on the dose, researchers reported in June in The New England Journal of Medicine.
It will take many months to learn whether the drop in TTR eases symptoms. The hope is that the one-time treatment will work as well as, if not better than, an RNA-based drug that must be injected every 3 weeks.
In another study, researchers at Editas Medicine injected a harmless virus carrying CRISPR DNA into the eyes of six adults with an inherited vision disorder called Leber congenital amaurosis 10. The scientists hoped to snip out extra DNA that disrupted a mutated eye gene so cells would then make its missing protein. After 3 to 6 months, two patients—who had been almost completely blind—could sense more light, and one could navigate an obstacle course in dim light, the researchers reported at a September meeting. They hope for greater vision gains in adults receiving a higher CRISPR dose, and in young patients.
Embryo ‘husbandry’ opens windows into early development
by Mitch Leslie
Insights into early embryonic development can help scientists understand miscarriages and birth defects—and hone in vitro fertilization (IVF) protocols. But legal, practical, and ethical limitations constrain studies with human embryos. This year, scientists unveiled potential stand-ins: mouse embryos reared far longer than before, and embryo replicas made from human stem cells or reprogrammed adult cells.
Scientists have struggled to grow mouse embryos outside a mother mouse’s body for much longer than 3 or 4 days. But in March, one team reported a recipe for stretching that to 11 days. A key step, they found, was rotating the jars containing the embryos on a device that resembles a miniature Ferris wheel. It continually mixes the nutrient broth that bathes the embryos and ensures that oxygen levels and atmospheric pressure are congenial. The embryos underwent a key stage of cellular reorganization, grew organs, and sprouted hind legs.
Other scientists devised substitutes for a crucial embryonic stage known as the blastocyst. A hollow ball harboring only a few hundred cells, the blastocyst implants into the uterus and is the first embryonic stage to feature specialized cells. It’s also inserted by many IVF clinics into prospective mothers.
One team made blastocyst replicas from human embryonic stem cells and induced pluripotent stem (iPS) cells, stem cells reprogrammed from specialized adult cells. Another discovered that skin cells undergoing the transition to iPS cells produce blastocystlike structures. These ersatz blastocysts aren’t real embryos, but some of them could offer an instructive—and less controversial—alternative.
The field received another boost in May. The international organization that sets stem cell research guidelines relaxed its long-standing prohibition on growing human embryos in the lab for more than 14 days, allowing scientists to probe embryonic events that occur after that time.
Breakdowns
The world’s nations gathered in early November for U.N. climate talks in Glasgow, U.K., to keep alive the hope of limiting global warming to 1.5°C above preindustrial levels. It’s an ambitious goal, with the world already at 1.2°C and the global economy still overwhelmingly reliant on fossil fuels. After 2 weeks of discussion, it was clear that—although rapid change has begun—the 1.5°C target is on life support.
The recent, rapid adoption of renewable energy has likely averted the worst-case warming scenarios of 4°C or more that seemed probable 10 years ago. And for the first time, there is a path toward limiting warming to just below 2°C, as long as nations keep their Glasgow pledges. But it would take far more ambitious action—halving present emissions within 10 years—to reach the 1.5°C goal.
Other progress in Glasgow included a deal on the rules for carbon markets and reporting emissions. But given that U.N. agreements are not binding, the true fate of carbon emissions will come down to action at the national level.
To keep the pressure on, the U.N. talks are now encouraging countries to offer revised emissions reduction pledges every year, including at the next meeting, in Sharm el-Sheikh, Egypt, in November 2022.
Under normal circumstances, U.S. approval of the first drug designed to interfere with the biology underlying Alzheimer’s disease would be cause for celebration. But aducanumab, cleared by the Food and Drug Administration (FDA) in June, was not the drug many researchers were hoping for. Its bizarre path to market stoked divisions between scientists, sowed confusion among doctors and patients, and prompted many to question the integrity of the U.S. drug regulatory system.
The drug, an intravenously delivered antibody developed by Biogen and marketed as Aduhelm, clears sticky plaques of the protein beta amyloid, thought to cause damage in the brains of people with Alzheimer’s disease. But only one of two large clinical trials showed the drug was better at slowing cognitive decline than a placebo. (One possible explanation is that removing amyloid plaques isn’t sufficient to reverse a cascade of harmful brain changes already underway by the time Alzheimer’s symptoms emerge.)
In November 2020, an independent advisory committee to FDA recommended overwhelmingly against approving the drug. But 7 months later, the agency stunned scientists by greenlighting aducanumab under an accelerated approval pathway. That process relies on a “surrogate endpoint”—here, beta amyloid reduction—rather than a demonstration of clinical benefit.
Reaction was swift. Three of 10 advisory committee members resigned in June. One of them, Aaron Kesselheim of Harvard Medical School, called the move “probably the worst drug approval decision in recent U.S. history.” Journalists dug into the close relationship between Biogen executives and FDA officials, reporting back-channel communications and unofficial meetings. Investigations into the approval process by the inspector general of the Department of Health and Human Services and by two committees in the U.S. House of Representatives are ongoing.
Some physicians have said they will not prescribe the drug, and medical centers, including the Cleveland Clinic and Mount Sinai, have announced they won’t administer it. Biogen told investors in September that sales of the $56,000-a-year drug had been slower than anticipated. The company announced in November that it was investigating the death of a 75-year-old woman who took aducanumab and developed brain swelling, a known side effect.
Even supporters of the drug say the turbulent rollout has harmed the field. “All of this is confusing and scary for people and families who are facing this devastating and fatal disease,” Maria Carrillo, chief scientific officer of the Alzheimer’s Association, which advocated for approval, said at a conference last month.
Scientists have long come under attack for their work. But this year, political rifts over the COVID-19 pandemic sparked unprecedented public hostility toward scientists, including online and offline intimidation, protests, and death threats.
Those involved in public health suffered the highest profile harassment. U.S. National Institute of Allergy and Infectious Diseases Director Anthony Fauci, for example, needed a full-time security detail as early as April 2020. Chris Whitty, England’s chief medical officer, faced harassment in public and regular demonstrations outside his house before the U.K. government took protective measures. Health workers and officials around the world have reported physical and online attacks—and many have quit their jobs as a result.
Researchers who criticized unproven treatments for COVID-19 also faced abuse. Marcus Lacerda, a clinical researcher who led a large trial in Manaus, Brazil, showing no benefits from high-dose chloroquine, received death threats. Scientific integrity consultant Elisabeth Bik faced online intimidation—and legal threats—after raising concerns about ethical and methodological problems in French microbiologist Didier Raoult’s hydroxychloroquine research.
Such threats have had a chilling effect on scientists: A Nature survey of 321 researchers who spoke to the media about the pandemic found that more than half have had their credibility attacked, and 15% had received death threats. Many said the experiences left them unwilling to give future interviews.
Correction, 16 December, 2 p.m.: An earlier version of this story misstated which drug was used in the Brazil study.