Glanville says his treatment will work well and be more affordable—think $900 per dose instead of the $2,000 per that Regeneron charges.
Peter Prato(Photo)
Glanville says his treatment will work well and be more affordable—think $900 per dose instead of the $2,000 per that Regeneron charges.
Glanville says his treatment will work well and be more affordable—think $900 per dose instead of the $2,000 per that Regeneron charges. (Photo: Peter Prato)

The Antibody Avenger and the Quest for a COVID-19 Cure

Vaccines are rolling out with increasing speed, but we’ll also need effective treatments, because new coronavirus cases will be a worldwide reality for years to come. Enter Jacob Glanville, a maverick San Francisco immunologist who believes he’s found an unparalleled path to healing.

Heading out the door? Read this article on the new Outside+ app available now on iOS devices for members! Download the app.

Since the early days of the pandemic, Jacob Glanville, a 40-year-old skateboarding immunologist with a silver tongue and a knack for self-promotion, has been proclaiming that he would develop the world’s best COVID-19 treatment. Few competitors believed him. Most just scoffed. But Glanville, then the chief executive officer of a tiny San Francisco company called Distributed Bio, has made significant progress, successfully creating an antibody that neutralizes the coronavirus in hamsters, an animal model for humans. He’s far behind the pharmaceutical giants Regeneron and Eli Lilly—both of which won FDA approval of antibody-based treatments in November 2020—but his work is still important. With a total of $9 million invested so far, his discoveries have cost only a fraction of what the big firms have spent on their research. He says his treatment will work well and be more affordable—think $900 per dose instead of the $2,000 per that Regeneron charges. “I’ve got the best-in-class therapeutic,” Glanville told me recently, repeating a line he’s used often since April 2019.

Given the FDA’s approval of the first vaccine on December 11, 2020, you might be wondering why you should care about Glanville or his drug. It’s because vaccines alone won’t end the pandemic. That’s especially true in the U.S., where so many people see their freedom to get sick and pass it along as a basic right. The chilling truth is that, at this point, it’s effectively impossible to vaccinate enough people worldwide to vanquish COVID-19. And so, as with the flu, the deadly virus is likely to remain resident among us, destined to cause breakouts every year. You should care about Glanville—along with the many other drug manufacturers still pushing their creations forward—because, odds are, the best COVID-19 treatments aren’t available yet. If Glanville is right, he could save many lives all over the planet.

To understand how a recovering hacker from Central America bootstrapped a COVID-19 treatment, it’s best to start back in January 2020, when Glanville elbowed his way into an Arlington, Virginia, conference crowded with immunologists and virus hunters, who were starting to look worried. On stage, Dr. Anthony Fauci, the now famous director of the National Institute of Allergy and Infectious Diseases, was telling the room about a spooky new coronavirus preparing to migrate from Wuhan, China, and hopscotch around the globe. This was the moment Glanville had been preparing for his entire life. He met it with eagerness and energy. Before Fauci had finished speaking, Glanville knew he would enter the race to discover a treatment for whatever this disease turned out to be. 

On his phone, he pulled up BioRxiv (pronounced “bio archive”), an online server that has published roughly 100,000 COVID-19 research papers since the virus surfaced, and he found two early studies that confirmed a hunch he had. One noted that the virus that causes COVID-19 is a close relative to the one that causes SARS, the highly infectious coronavirus that killed more than 700 people between 2002 and 2003. The other was about five different antibodies that neutralized the SARS-causing virus. The next day, Glanville was leveraging his knowledge during a meeting with the DARPA, the cutting-edge Department of Defense group that funds infectious-disease research. “We have the technology to adapt that old antibody to hit a new virus,” Glanville told DARPA officials inside a dimly lit conference room. “Nobody else can do that, and it’s what we do every day!”

DARPA’s people nodded politely. Later, so did representatives from the federal government’s other big infectious-disease funding arm, the Biomedical Advancement Research and Development Authority, which is part of the Department of Health and Human Services. Glanville got nothing. Antibodies are just one drug class among many, ranging from immunosuppressants to convalescent plasma to antivirals, that researchers and the government were exploring as possible treatments for the novel coronavirus. Understandably, the billions invested in rapid research ended up going to the relatively few giant drug companies with a proven record of producing effective drugs on short notice: Johnson and Johnson, Pfizer, and Regeneron among them. “My feeling was, fuck it,” Glanville says of that early setback. “We’re nimble. We can move faster than anybody else. The money would come afterward.”

I first met Glanville in 2018, when I was reporting a story for Outside on his efforts to use antibodies to develop an antivenom that treated bites from many different species of venomous snakes. Low-key for somebody as ambitious and smart as he is, Glanville took me to a Kava restaurant in San Francisco’s Mission District. While we sipped earthy tea from wooden cups, he explained how coding, geometry, and probability were central to designing what he hopes will become a universal influenza vaccine, his passion project.

Even then, Glanville had a reputation for taking peculiar approaches to complex immunological problems. (He prefers “innovative” to peculiar, but agrees that it’s “scrappy.”) Research for his flu vaccine included a live pig experiment based inside a jungle shack in Guatemala, and his quest to make the antivenom began as a YouTube search that turned up Tim Friede, an underemployed truck driver from Milwaukee who’d intentionally been letting the world’s deadliest snakes bite him for 20 years. Glanville was fascinated by antibodies. Not by where they came from, but by how good they were at turning off diseases.

Glanville grew up outside a Maya village on Guatemala’s Lago de Atitlán. As a kid, he displayed an unnerving intellect, cooking nitroglycerin in the family bathtub and completing all available high school math classes by the time he was 11. His dad ran a hotel and a restaurant, and his mom, an artist, was the daughter of the scientist who led the development of boosters that sent the first NASA rockets into space.

You should care about Glanville—along with the many other drug manufacturers still pushing their creations forward—because, odds are, the best COVID-19 treatments aren’t available yet. If Glanville is right, he could save many lives all over the planet.

In his late teens, while studying molecular and cell biology at the University of California at Berkeley, Glanville was hacking into friends’ computers, pulling the types of pranks you’d expect from a teenager in a frat house. He was a good student but not a particularly focused one. During the summer between his freshman and sophomore years, Glanville was robbed at knifepoint while climbing a volcano near his family home. The assailant knocked him to the ground with the side of a machete blade, giving him a deep wound and some lasting psychological trauma. He skipped the next year of school and, when he returned to Berkeley, began to study with renewed focus. Immunology enthralled him, and he soon started writing programs that essentially hacked the immune system. Pfizer hired him shortly out of college, in 2008. By then recent advancements in genetic science had pushed antibody research to a new and profitable maturity. Glanville’s job was figuring out how to match antibodies to antigens, the fancy word for any bacteria, virus, or toxin that our bodies fight by producing antibodies. 

Antibodies were discovered in 1890, and drug researchers have long viewed them as immunological silver bullets. To describe the cells’ advantage over other types of drugs, Glanville regularly uses an analogy. Picture a drone buzzing in the sky, he says. A hawk suddenly dives from above and takes it out. The drone and hawk are both machines, but as Glanville points out, “one was built radically better than the other.” In this example, antibodies are the superior machine: the hawk.

After somewhere on the order of 200 million years of evolution, all terrestrial vertebrates and every species of mammal still use antibodies as their primary immunological defense. The challenge for mankind has been figuring out how to harness and improve on nature’s gift. Since drug developers initially broke through in 1986, when the first monoclonal antibody drug was released, health agencies have approved 100 or more for use in the U.S. and European Union. As of 2015, antibody engineering generated $125 billion in annual revenue. Six years later, it’s surely much more. 

“The science is built on the fact that life has these very elegant and simple rules in place that govern how it functions,” Glanville says. “We’re taking advantage of how amazingly rational life is.”

Inside the human body, antibodies work like a personal security team that never sleeps. When a novel antigen infects a person, the immune system awakens its B cells—the cells responsible for secreting antibodies—and starts creating new tiny predators by scrambling different combinations of the 20 different amino acids found in most animals on the planet. 

But the body doesn’t simply print a panacea; if it did, we’d never get sick. It has to learn. To do this, the immune system makes millions of different versions of antibodies and sees what sticks. At first, very few do. But when an antibody that hits and sticks to an antigen is found, the B cells start improving its successful designs by altering one amino acid at a time. The parts of the antibody that work well, nature keeps. The parts that don’t, nature changes. “Evolution in the bloodstream,” Glanville calls it. “Our immune systems are rapid evolution machines. It’s the only way we can fight off pathogens that can evolve every half-hour.” 

Eventually, when the immune system has finished its job, the bloodstream is swarming with an army of antibodies capable of blocking, with great specificity, the parts of the antigen that bind to our cells and make us sick. Only then does the viral infection blink out.

While Glanville was still at Pfizer, he isolated antibodies from 654 donors and collected their cells inside a single test tube filled with a solution of frigid glycerol. Ultimately, he collected 32 billion different antibodies, which he thought at the time was a rough approximation of the entire range of possible antibodies our immune system can produce. (That was off by a lot. Fifteen years later, Glanville’s library holds 76 billion antibodies and will soon hold upward of 100 billion.) Others had created “libraries” like this before. Indeed, a discovery in the mid-1980s that led to the creation of such antibody repositories won George P. Smith and Gregory P. Winter the 2018 Nobel Prize in Medicine. Glanville the Computer Whiz accelerated the process of scanning those antibodies for potential drug leads.

Finding a few winning specimens inside billion-member antibody libraries is done by repeatedly dribbling a concentration of an antigen into a pipette well—a glass tray with dozens of liquid-holding compartments—that is swarming with antibodies and then scooping up the ones that are attracted to it. The technique is called panning, and it’s both highly effective and very time-consuming. Antibodies are shaped like cellular burrs, and they often stick to random cells or parts of the antigen that do nothing to actually neutralize it. During the first few panning rounds, billions of antibodies may attach to high concentrations of an antigen. But as the panning progresses, researchers lower the concentrations of the added antigen while filling the wells with successful antibodies from the preceding round. Eventually, after many rounds of panning, what’s left is the “gold”: a few thousand antibodies, each of which possesses an almost magnetic attraction to the antigen, which suggests that the antibodies know something about the invaders. These antibodies stand the best chance of being useful as drugs. 

Glanville sped up the panning process with math. All of the possible variations of antibodies exist on a spectrum. Each cell is unique, but it’s also just slightly different from its neighbors. By writing an algorithm that identified the shape of the most likely useful antibody, Glanville shrank the search pool. Math says our winning antibody is probably one of the hundred million variations between antibody X and antibody Y, the reasoning went. We look there first.

Thanks to the work by Glanville and others, within 20 years of the emergence of antibody libraries, the search process became much faster—something that might take a year could now be accomplished in a week. These changes coincided with the dawn of the genetic revolution, and antibody drug researchers began borrowing skills and techniques from that field as well. Today the race isn’t only to find nature’s best version of an antibody. It’s also about who can genetically engineer them to work better.

After isolating a promising antibody, drug researchers start fine-tuning their product by growing many slightly tweaked versions. On one, they might lop off a part of an antibody that’s thought to prompt an immunological response. On another, they might add mutations that make the antibody easier to track, or give it a longer half-life in the immune system, or make it deliverable in a shot instead of an IV. Each designed antibody gets tested in petri dishes. The antibodies that perform best go into live animals. If they clear that hurdle, sick people get the treatments. And should any one of these imperfect blends of nature and genetic nurture cause only mild or acceptable side effects in actual patients—while still performing its job inside the body—the FDA grants an approval for use. Failure can happen at any point: Even with all the advantages of full-service genetic-engineering shops like Glanville’s, drug researchers’ chances of success remain needle-in-a-haystack low. Just one of every 5,000 prescription drugs in development actually makes it to market. With antibody drugs, it’s closer to 2 percent of contenders. 

Like anybody in this field, Glanville was confident he could beat the odds, and the influenza work became his life’s purpose. “I was convinced that a post-pathogen existence was possible,” he says. “We do not have to live with infectious diseases anymore.”

Not long after building an antibody library for Pfizer, Glanville took his algorithms and antibody-library concepts and established Distributed Bio. At first the business only did contract jobs, in which big firms would give Glanville a drug target and have him find the antibodies that hit it. But the goal was always to use Distributed Bio as a stepping-stone to build a therapeutic drug company. In classic Silicon Valley style, he envisioned a profitable company with a sort of Robin Hood ethic: his operation would use technology developed by the rich to make drugs affordable for the poor. His company wouldn’t prioritize profit but instead would treat neglected diseases and maladies, like snakebite envenomations, and would make a huge volume of curative drugs for diseases like the flu and COVID-19 available to all. “Fiduciary obligations are at odds with solving global crises,” Glanville says. “Health care is a human right.” 

Picture a drone buzzing in the sky, he says. A hawk suddenly dives from above and takes it out. The drone and hawk are both machines, but as Glanville points out, “one was built radically better than the other.” In this example, antibodies are the superior machine: the hawk.

For the first six years, Glanville bolstered Distributed Bio’s assets and technological expertise by focusing on lucrative contract work, collaborating on projects with seven of the ten leading pharmaceutical companies. They’d give him an antigen or a cellular target. He’d find them an antibody that hit it. The big companies would then develop that antibody into a drug or use it for a handful of other health-related research purposes that designer cells can be used for. The profits Glanville earned were put back into new technologies that grew Distributed Bio to a company valued in excess of $100 million. Meanwhile, Glanville finished his Ph.D. from Stanford and was remarkably productive. He continued to write academic papers, placing more than 30 in top scientific journals like Nature, while also filing for patents on nine new technologies related to antibody discovery and winning competitive grants from the Gates Foundation and the National Institutes of Health. While doing all this, he consciously worked to reposition himself as a drug developer. Part of this effort has included branding: Glanville has been featured in several national publications and in a January 2020 Netflix series, Pandemic: How to Prevent an Outbreak, which profiled “the heroes at the front lines” of stopping the next global epidemic. The documentary, which explained Glanville’s concept of a post-pathogen existence, debuted during the week of the BioThreats conference in Virginia.

What Glanville wasn’t, though, was a drug developer. “He’s very, very good at what he does. He makes the raw material that is the first step in putting together an antibody drug,” says one scientist I spoke to who wishes to remain anonymous. “What he has not done yet is make antibody drugs. He still doesn’t know what he doesn’t know.” This drug developer, who has a proven record of bringing products to market, says Glanville’s expectation that he can rise to the top of the pharmaceutical industry with a single moon-shot pharmaceutical is naive, like a world-class sprinter expecting to win his first marathon. “Would you bet that they could be competitive?” she says.

As the quest to create a COVID-19 treatment began, Glanville pursued his vision with a characteristically dogged approach. “Jake just kept calling,” remembers Jack Wang, Distributed Bio’s top virus builder. “Finally I answered, and he started bombarding me with all these interesting ideas.”

It was the afternoon of January 30, 2020, and Wang, who dropped out of an immunology and biophysics Ph.D. program at Yale and was hired by Glanville, had only been at Distributed Bio for three months. Glanville contacted him immediately after his unsuccessful meeting with DARPA. “He wanted me to redesign the SARS antibodies to hit COVID,” Wang recalls. “Yeah, it was a little overwhelming.”

By late January, the outbreak was mostly confined to Asia, with just a few confirmed cases in Europe and fewer than ten in the U.S. The nation, transfixed by the impeachment of President Trump, was still weeks away from lockdowns and school closures, and nobody was panicking yet. Thinking about what lay ahead, the 26-year-old Wang was openly skeptical that a tiny firm would be able to engineer an antibody to fight an emerging pandemic virus. “I’m pessimistic in general,” he says, “but this seemed like work for much larger companies.”

He threw himself into it anyway. Soon Wang was reviewing the blueprints of five antibodies that other research groups had identified as having “potential therapeutic value” against coronaviruses. Two were harvested from serum donated by patients who’d caught SARS in 2002 and survived. A third came from a SARS-infected rat. A fourth was an antibody from 2007 that a Chinese team had found in January, which hit weakly and did not neutralize what was being called SARS-CoV-2—better known now as the virus that causes COVID-19. Finally, Glanville prepared his 76 billion antibody collection to be searched for one that would dock with CoV-2. None of these antibodies worked very well. But looking at the shared regions where the antibodies hit the virus helped serve as a map for designing Glanville’s new creation.

To get started, Wang and Glanville turned to a computer program that depicted the amino-acid sequences in the viruses and antibodies as crystalline-structure shapes, modeling how the proteins look in real life. Individually, each amino acid is infinitesimally small—a strip of joined molecules that are bent, kinked, or looped. When squished together on a screen, the assemblages that form the virus and antibodies looked like bad-hair days: pastel-colored amino acids poking out from a clearly defined nucleus. 

The men wanted to see on-screen how the antibodies were docking to the virus. If they could see that, then maybe they could understand nature’s approach to coronavirus-neutralizing antibody design. And maybe an image of antibody and virus binding would offer clues on how they could shuffle the amino acids to improve their own design.

After a few minutes of digitally rotating the various shapes, a pattern emerged: the antibodies were all hitting the virus on the same spike of the coronavirus’s crowns. (Corona means “crown.”) It made evolutionary sense. The SARS- and COVID-19-causing viruses both picked humans’ cellular locks in the same place and in almost the same way. This little cellular node was the viruses’ key. Now they needed to design an antibody to block it.

We owe good luck and a team of quick-thinking public-health experts a debt of gratitude for stopping the 2003 SARS outbreak in Toronto, where the epidemic ended before erupting the way COVID-19 has. A highly infectious respiratory disease, SARS sickened 8,098 people worldwide and killed around 800.

Once the virus fizzled among people, though, it didn’t disappear or become any less dangerous. It just retreated back into the reservoir species from which it had come, or to another species entirely. Bats certainly carry the virus; palm civets or pangolins likely do, too. Inside the warm bodies of infectious hosts, the organism has kept evolving. It was only a matter of time before a new threat, like CoV-2, made itself known.

Genetic evidence suggests that CoV-2 broke out after the virus that caused SARS met with a member of the coronavirus family. Most likely while adrift in the warm body of some animal host, the two viruses swapped genes. In scientific terms, they probably “recombined,” switching out one protein for another and creating a new and alien progeny, a process that can happen when viruses in the same family meet in the same host. Evolution isn’t a one-way street, and recombination most often results in viruses that are less fit. That didn’t happen this time. Instead, generations of recombined viruses later, CoV produced an offspring that was a nearly perfect match with humans.

Genetically, the viruses that cause SARS and COVID-19 are 80 percent identical, and the COVID bug succeeded where the SARS bug failed because of a yet unknown subset of genetic mutations. By now, most of us are well acquainted with the colorful, furry-looking microscopic images of CoV-2. Picture the flowery red things that festoon the virus, and then zoom in. Each spike is actually a cluster of elegantly arranged proteins that look like a crown. And each spike on that crown is the key that the virus uses to open a very specific lock on very specific human and animal cells. The mutations powering COVID-19’s global spread happened there.

A typical human body contains roughly 37.2 trillion cells. Through a process that’s like a biological game of Tetris, food gets digested, fingers type, and eyes cross when two or more of these intricately shaped molecules meet their perfect match and trigger a cascade of cellular reactions. The places where cells connect are called receptors, and the extreme specificity of receptors is the gate that regulates which proteins get in and which don’t.

Just one of every 5,000 prescription drugs in development actually makes it to market. With antibody drugs, it’s closer to 2 percent of contenders.

Receptors bring (and end) life by ordering the cellular universe. Viruses sabotage our finely tuned molecular machinery by twisting the proteins they’re made of into shapes that match our receptors perfectly, and then hijacking the cell and instructing it to make more copies of the virus. After somewhere between three and four billion years on earth, viruses have become so successful at what they do that they outnumber all other forms of life combined. For good reason, they’ve been called the most successful inhabitants of the biosphere.

The virus that causes COVID-19 is better at connecting than most. When viruses mesh with the right cellular receptor, most spread by hijacking the cell’s machinery and printing off anywhere from 1,000 to 10,000 copies of themselves in a few hours—sometimes as many as 100,000. 

Like the virus that causes SARS, CoV-2 spreads most easily when it hits cells outfitted with a receptor called ACE-2. Those cells aren’t found everywhere in the body. After the SARS outbreak in 2002, researchers began seeking out colonies of cells rich with ACE-2 receptors and found them in the blood vessels, the lower intestines, the kidneys, and the heart.

The fact that the virus hits cells all over the body helps explain the diversity of symptoms it causes, including vomiting, diarrhea, and kidney or heart failures. It may also explain some of the other symptoms that make this strange virus so mysterious. The loss of smell or other neurological symptoms that have been widely documented—hallucinations or delirium, for example—suggest that the virus could be hitting and replicating in the brain via an unidentified backdoor receptor. It’s also notable that viruses don’t cause most symptoms directly. Our immune response produces symptoms to get rid of infections, explaining why drugs like steroids, which ramp up or suppress the immune response, are so often used to treat viral infections.

The virus CoV-2 owes its success to a few tricks. One is that the virus doesn’t make everybody sick. It can be transmitted by asymptomatic patients and could even have a “negative serial window,” meaning that newly infected people can show symptoms before the person who gave them the disease. Also important: the critical troves of ACE-2-equipped cells that it likes so much are found in the upper respiratory tract—the larynx, the pharynx, the nasal cavity—and the lungs. There, the receptors are found on cells that create surfactant, a slippery film that is critical for breathing. 

When CoV-2 first infected a person somewhere in rural China, the new bug was far stickier to the ACE-2 receptor. For the virus, it’s hard to imagine a better evolutionary move. For a human, it’s hard to imagine one that could be worse. Cough, and the virus wafts out of the body on a parachute of aerosolized spit, where it can stay suspended and possibly infectious for three hours. Inhale, and the virus lands among fertile cells.

By the middle of March, COVID-19 had put San Francisco in lockdown. Two days later, Glanville got permission from the city for his team to keep working together in their labs. Distributed Bio halted all other projects except its search for a COVID-19 treatment.

One morning not long after that, Wang drove to the office across the Bay Bridge, which was eerily empty. By then, he and Shahrad Daraeikia, an antibody engineer who was also assigned to the COVID project, had been working 12-hour days for a month. Creating the antibody required nailing down three components: the part of the virus that binds to the cell, the part of the cell that the virus binds to, and the antibodies themselves. Wang and his team delivered the first two pieces after several weeks. It took a bit longer for Daraeikia to make the antibodies.

Antibodies are Y-shaped and, like Lego pieces, they attach to an antigen via pegs on their tips. Each arm of the Y has three pairs of pegs, for a total of 12 pegs per antibody—an array collectively referred to as the complementary binding region (CBR). Once the immune system has identified the best antibody class from the five that our body makes, it starts tweaking the pegs by shuffling the amino acids on their tips. In the shop, Glanville and Daraeikia focused their work on CBRs. There are, as Glanville explained, “e to the 84 different possible combinations” of them—that’s a one with 84 zeros behind it, “or roughly equivalent to the number of known atoms in the universe.”

“The trick was hitting the mathematical sweet spot,” Glanville says. As he’d done at Pfizer, he used math to shrink the number of promising antibodies to a manageable 12 billion. This time he focused on antibody combinations found in nature and those that looked most like the original five SARS neutralizing antibodies. “We wanted to look at all the variations that can exist in finite spaces.”

Once they’d defined that space, it was Daraeikia’s job to fill it with possible antibodies. Though drug development employs a mix of biological tools and supercomputers, Daraeikia used Microsoft Excel.

“Yeah, kind of miserable,” he says. But in this one painful instance, where speed was paramount, Excel happened to be faster than having Distributed Bio’s coder write a new program. On each row of a spreadsheet, Daraeikia wrote out 6-to-14-letter amino-acid sequences that define each CBR on each tip of the antibody candidates. Then, one letter at a time, he manually changed an amino acid in each sequence—a process that took 30 hours. “If it was a Y in the first position of the first antibody, I’d change it to an F,” he says. “Move to the next, and do the same. Then I’d go through and change two letters at a time. It burns a hole through your eyes.”

After hand-writing the genetic codes for hundreds of possible CBR combinations, Daraeikia ordered 400 snippets of DNA from Integrated DNA Technologies, a life-sciences company in Coralville, Iowa, and loaded them into a program called Tumblr. They then infused hundreds of millions of additional antibodies from more than 100 healthy human donors to account for as many possible combinations of CBRs as possible. The Tumblr device cracked the antigen’s code by mixing up DNA combinations that rapidly explored all the combinations of CBRs that fit its locks. With instruction provided by Daraeikia’s spreadsheet, the Tumblr machine grew 12 billion distinct antibodies. Somewhere, in one of the hundreds of pipette wells that they’d filled with engineered antibody parts, was a sequence that could cure a disease that was on its way to killing 2.54 million people and counting. 

Daraeikia, Wang, and their colleagues finished the final tests on March 30. That morning, Wang donned a lab coat and an N95 mask and grew each individual antibody by dribbling the DNA onto tiny flecks of bacteriophage, an organism that, like nature’s 3D printer, creates the physical manifestation of the DNA’s instructions. He then magnetized the parts of the virus that bound to the ACE-2 receptor and dredged them through the antibody soup. By dangling a magnet over the top of the petri dish, Wang separated the antibodies that attacked the virus from those that didn’t—the panning process. Doing this again and again, using increasingly smaller amounts of parts from the CoV-2 spike, he isolated the antibodies that bound to the virus with the highest affinity.

During this work, Glanville was at home with his ten-month-old daughter. Sometime around 3 P.M., his phone started vibrating. Wang and Daraeikia were sending him a barrage of texts.

“Holy shit!”

“We have binders!”

“How many?”


“From every fucking arm!”

“It was sort of like a ship sailing off into the darkness,” Glanville remembers. “That was the moment we saw shore. We were like, OK, we’re fucking there. We’ve got a drug.”

Shortly after, the cold reality of drug development slapped him in the face. He needed $30 million or so just to manufacture enough antibodies to get through clinical trials, and then another $86 million to pay for the trials themselves. “I started harassing all levels of government for funding,” he says. He got less than 1 percent of what he needed, so he appealed to the public, launching a Kickstarter page that raised about $200,000. As it grew clearer that his competitors were outpacing him, Glanville was invited by the editors of Nature Biotechnology to join a panel of luminaries discussing the challenges monoclonal antibodies face in curbing the pandemic. He tipped his hand that he thought his competitors’ soon-to-be-approved antibodies might fail in the clinical setting because they triggered an immune reaction, which is exactly what they did. He also made a case for his work, explaining how he’d designed his drug to be harder for the immune system to detect. Meanwhile, he started making the rounds on TV. Fox Business did a story about Glanville’s potential cure. MSNBC followed. Then Yahoo. CNBC. Good Morning America. Dr. Phil. A news program in Turkey. Ones in Australia, India, New Zealand. In every interview, Glanville repeated the same line: by September of 2020, his cure would be going into patients. Obviously, that hasn’t happened.

To remind herself that hurried work can have consequences, the anonymous virologist I interviewed keeps a quote on her office wall from Richard Feynman, the Nobel Prize–winning physicist. As a lesson in drug development, she often tells the story of Feynman’s devastating conclusions about the 1986 explosion of the space shuttle Challenger. It’s set during an inquiry about the disaster. During a famous line of questioning about the dangerous disconnect between the caution of NASA’s engineers and the ambition of the agency’s management, Feynman took out an O-ring that engineers had identified prelaunch as a part that could fail catastrophically, especially in freezing temperatures. He dropped it in ice water and the part failed. “For a successful technology, reality must take place over public relations,” Feynman said. “For Mother Nature can’t be fooled.”

“Data is king,” the virologist says, echoing Feynman. “In my field, a drug is either going to work or it’s not.”

Basically, she thinks that Glanville, who has yet to publish any results from his coronavirus research in a major scientific publication, has oversold the importance of discovering antibodies that can neutralize CoV-2 in a dish or a hamster, even though he’s succeeded in doing both. In experiments with hamsters, Glanville’s antibodies reduced viral load by 97 percent in rodents that received the drug as a treatment, and even more than that when they were given prophylactically. The virologist says this is a good start, but it still doesn’t demonstrate the ability to neutralize the virus in people; it doesn’t show whether the treatment can cause dangerous side effects; and it doesn’t reveal how much to give in a dose, where and how the dose should be administered, whether the antibody actually disperses to the parts of the body that harbor the virus, and whether the drug can even be manufactured.

“That’s the problem with biology,” says the virologist. “It gets more and more complicated the deeper you get into drug development.” Between the discovery of an antibody, even a potent one, and the development of an actual drug, there is a gauntlet of manufacturing and safety hurdles that, because of the expertise and money needed to navigate them, giant pharmaceutical companies are better equipped to clear. Although Glanville’s team includes researchers with experience shepherding antibodies from discovery to the marketplace, he is having to learn the bureaucracy of drug approval on the fly. His public optimism, the virologist argues, may be dangerously and even cruelly misleading to those outside the industry.

Glanville is now one in a crowded field of researchers trying to improve antibodies’ efficacy against COVID-19. By late 2020, there were at least 21 other monoclonal antibodies in some form of clinical trials, including five knocking on the door of FDA approval in phase three. And after watching the mixed success of the leading antibody drug manufacturer, Glanville decided to stop trying to emulate the front-runners. Regeneron, the multibillion-dollar company whose antibody-based drug was approved for emergency use by the FDA in late November, took all the right steps, but its drug is far from the effective cure it hoped it would be. Before the FDA granted its final approval, early results suggested it could be hugely successful. Because of this, doctors gave an experimental version of it to President Trump, who claimed that it cured him, despite there being no scientific way to know this, since he received several treatments at once.

What has become clear is that Regeneron’s cocktail, like Eli Lilly’s drug bamlanivimab, only works well against milder cases of COVID-19. These drugs aren’t being widely used by hospitals, because when people fall critically ill, even massive doses of the antibodies delivered intravenously do little to revive them. Antibodies only target the virus, and once an infection is established, there is simply too much virus for the administered antibodies to control, and they can do nothing to tamp down the symptoms that ultimately cause death. This fact, plus issues related to storage and cost, explains why many in the industry no longer pin their hopes of taming COVID-19 on antibodies.

That Glanville’s competitors haven’t been huge successes might seem like a good reason for him to abandon his project. So, too, that by midwinter no agencies or private investors had come forward to fund his efforts, despite almost a full year of persistent, exhausting, and ultimately deflating lobbying efforts. By early March, Glanville estimated he’d met with almost a dozen government agencies funding COVID research, from the Army and Navy to Operation Warp Speed. The Gates Foundation turned him down. So did a handful of other big-dollar foundations. He raised only $9 million, barely enough to get his antibodies through animal trials. The challenge seems to have only hardened his resolve. Reality, he says, is driving him forward. “Very rarely in the history of pathogens have we vaccinated enough people worldwide to eliminate them,” he says (smallpox being the lone example). “COVID is here to stay.”

When CoV-2 first infected a person somewhere in rural China, the new bug was far stickier to the ACE-2 receptor. For the virus, it’s hard to imagine a better evolutionary move. For a human, it’s hard to imagine one that could be worse.

Glanville maintains that his antibody is one answer. His sales pitch is as convincing as ever: an antibody potent enough that doses can be smaller; capable of being delivered in a shot rather than an IV; engineered to cause fewer side effects in the immune-system response than his competitors’; and, because it targets a part of the virus that hasn’t changed even as the human pandemic has spawned new viral mutations in Brazil, South Africa, and England, effective against new variants. True to his Robin Hood style, Glanville also wants his drug to be widely available and relatively cheap. He has mapped out a sort of Walmart distribution method for his drug, a model in which bulk production will keep the price down. Instead of $2,000 a dose, it will be $800, maybe $900, but certainly “less than the cost of an iPhone,” he says. (Glanville isn’t alone in his pharmaceutical goodwill. AstraZeneca is trying to sell its vaccine for $4 a dose.) Driving the cost savings for Glanville is smaller overhead—30 employees versus 30,000 at a company like Eli Lilly—and a novel manufacturing approach. Glanville had a team of interns identify more than 500 companies around the world with bioreactors that are capable of brewing his antibodies. Instead of cooking drugs through in-house bioreactors or subcontractors with restrictive terms, as the big companies have done, his plan is for many hands to make light work. By increasing supply, Glanville will fill the need and lower the costs.

The virologist who asked to remain anonymous is unwaveringly skeptical that this will play out as Glanville is willing it to, especially with so many researchers on pace or way out ahead of him. “Skeptical is the safe bet,” Glanville said of her take. “Odds are we fail.”

And that looked to be his antibody’s fate. But then, in early February, Glanville got a few pieces of good news. He refused to call them unexpected. The first was that Nature Biotechnology, an esteemed journal in his field, agreed to publish his work on the coronavirus. And in late February, Merck bought Pandion for $1.9 billion. The significance to Glanville was that Pandion used his patented technologies for some of its drug-discovery work. The announcement demonstrates that antibodies he has designed have clinical value. Most exciting for him is that he is finalizing an agreement with a federal entity—which he won’t name until the deal is final—that will fund his phase-one research.

Whether his antibody becomes a drug or not, entering the race to find a COVID-19 treatment clarified for Glanville why he got into this business—to help people. To that end, in the first week of January, he and his partners sold Distributed Bio to a much larger pharmaceutical company called Charles River Labs for more than $100 million. He’s since founded a new firm called Centivax that will focus solely on making therapeutic drugs and vaccines and getting the ones he’s already developed to market. “The time is nigh,” he says. “This work needs the best version of me possible.” As such, at 40, he quit drinking and started swimming in the ocean each day. To get just enough of the altered reality he needs to maintain sanity, he smokes three cigars daily on his rooftop office, looking out over the ocean and thinking about where the next bad bug might emerge.