BSIP / Universal Images / Getty / Paul Spella / The Atlantic

Editor’s Note: The Atlantic is making vital coverage of the coronavirus available to all readers. Find the collection here.

Twenty-nine. That’s the number of proteins the new coronavirus has, at most, in its arsenal to attack human cells. That’s 29 proteins to go up against upwards of tens of thousands of proteins comprising the vastly more complex and sophisticated human body. Twenty-nine proteins that have taken over enough cells in enough bodies to kill more than 80,000 people and grind the world to a halt.                   

If there is a way—a vaccine, therapy, or drug—to stop the coronavirus, it will be by blocking these proteins from hijacking, suppressing, and evading humans’ cellular machinery. The coronavirus may sound small and simple with its mere 29 proteins, but that is also what makes it hard to fight. It has so few weak spots to exploit. Bacteria, in comparison, might have hundreds of their own proteins.

Scientists have been furiously looking for a weakness in SARS-CoV-2, as the coronavirus that causes COVID-19 is formally known, ever since it was identified as the culprit behind mysterious pneumonia cases in Wuhan, China, in January. In just three months, labs around the globe have homed in on individual proteins, mapping some of their structures atom by atom at a record pace. Others are screening molecular libraries and the blood of COVID-19 survivors for compounds that can tightly bind and inhibit these viral proteins. More than 100 existing and experimental drugs are being tested against COVID-19. A vaccine candidate from Moderna was first injected into the arm of the first volunteer in mid-March.

Yet other researchers are focusing on how these 29 proteins interact with parts of the human cell—with the goal of finding drugs that target the host instead of the virus. While this seems indirect, it follows with the replication cycle of viruses. Unlike bacteria, viruses cannot copy themselves. “Viruses use the machinery of the host,” says Adolfo García-Sastre, a microbiologist at the Icahn School of Medicine at Mount Sinai. They trick host cells into copying their viral genomes and making their viral proteins.

One idea is to stop these virus-ordered functions without interfering with a cell’s normal functions. Here, the best analogy for a potential SARS-CoV-2 drug may not be an antibiotic, which kills foreign bacterial cells rather indiscriminately. “I think it’s much more like a cancer therapy,” Kevan Shokat, a pharmacologist at UC San Francisco, told me. In other words, it may be about selectively killing the human cells that have gone haywire. This opens up the possibility of many more drug targets in the host, but it also adds a challenge: It is much easier for a drug to distinguish between human and bacteria than between human and virus-hijacked human.

Antivirals are thus rarely “miracle cures” the way antibiotics can be against bacteria. Tamiflu, for instance, can shorten the duration of the flu by a day or two, but does not outright cure it. Antivirals for HIV and hepatitis C have to be taken in cocktails of two or three drugs at a time because the viruses can quickly mutate to become resistant. The good news about SARS-CoV-2, at least, is that it does not seem to mutate especially quickly for a virus. A number of different steps in the disease cycle could be lasting targets for a treatment.

Stop the virus from getting into a cell

Let’s begin where the virus starts, which is by tricking its way into a host. SARS-CoV-2 is covered in lollipop-shaped “spike” proteins, whose tips can bind to a receptor found in some human cells called ACE2. These spikes are what give coronaviruses—the group of related viruses that includes SARS-CoV-2 as well MERS and SARS—their name, because they create a crown- or corona-like appearance. The three coronaviruses are similar enough in their spike proteins that scientists are repurposing strategies from SARS and MERS to fight SARS-CoV-2. The vaccine from Moderna, for example, was able to start clinical trials so quickly in March because it is based on previous research into MERS’s spike protein.

The spike protein is also the focus of antibody therapy, which is likely faster to create than a new pill because it harnesses the power of the human immune system. The immune system makes proteins called antibodies to neutralize foreign proteins, such as those from a virus. Several hospitals around the country are trying to infuse antibody-rich plasma from COVID-19 survivors into patients. Currently, research groups as well as biotech companies are also screening the survivor plasma to identity antibodies that can be manufactured en masse in a factory. Spike proteins are a logical target for antibodies because the proteins are so plentiful on the outside of the virus. And again, the similarities between SARS-CoV-2 and SARS helps. “It looked enough like SARS that we had a bit of a head start,” says Amy Jenkins, a program manager at the Defense Advanced Research Projects Agency, the Pentagon’s blue-sky research arm, which is funding four different groups working on antibody therapy against the new virus.

But simply attaching its spike protein to a receptor is not enough for SARS-CoV-2 to gain entry into a cell. In fact, the spike protein is not active until it is cut in two. The virus takes advantage of another human enzyme—such as furin or the inelegantly named TMPRSS2—which can unwittingly come along and activate the spike protein. Several candidate drugs are meant to prevent these enzymes from unknowingly doing the virus’s work. One possible mechanism for the much-hyped hydroxychloroquine, the malaria drug Trump is fixated on, may be inhibiting this spike-activation process.

Once the spike protein is activated, SARS-CoV-2 fuses itself with the membrane of the host cell. It injects its genome, and it’s in.

Stop the virus from replicating

To a human cell, a naked SARS-CoV-2 genome looks like a specific type of RNA, a molecule that normally functions as instructions for making new proteins. So like a soldier who has gotten new orders, the human cell dutifully begins churning out viral proteins to make more viruses.

Replication is a relatively complicated step, which makes it a ripe target for antivirals. “There’s many, many proteins involved … there’s many potential targets,” says Melanie Ott, a virologist at the Gladstone Institutes and UCSF. For example, remdesivir, an experimental antiviral that is in clinical trials for COVID-19, targets the viral protein that copies the RNA, so the genome-copying step goes awry. Other viral proteins called proteases are necessary to free individual viral proteins that are linked together in one long strand, so they can go off and help the virus replicate as well. And still other proteins might help remodel the internal membranes of the human cell, creating bubbles of membrane that get turned into little virus factories. “The replication machinery sits on these membranes, and then it just starts making tons of viral RNA over and over and over again,” Matthew Frieman, a virologist at the University of Maryland School of Medicine, told me.

In addition to proteins that help it replicate and the spike proteins that make up a portion of the virus’s outer capsule, SARS-CoV-2 has a set of relatively mysterious “accessory proteins” that are unique to this virus. Figuring out what these accessory proteins are doing, Frieman said, could help scientists figure out other ways SARS-CoV-2 interacts with the human cell. These accessory proteins might allow the virus to evade the human cell’s natural antiviral defense in some way—another potential target for a drug. “If you can target that process,” Frieman said, “you can help the cell inhibit the virus.”

Stop the immune system from going haywire

Antivirals are likely to work best early in an infection, when the virus has not infected many cells nor made too many copies of itself yet. “When you give antivirals too late, the risk is the immune component has already taken over,” Ott says. In COVID-19 specifically, patients who become critically and fatally ill seem to experience what’s known as a cytokine storm, in which the disease sets off an indiscriminate and runaway immune response. Perversely, cytokine storms can also further damage the lungs, sometimes permanently, by allowing fluid to build up in the tissue, says Stephen Gottschalk, an immunologist at St. Jude Children’s Research Hospital. Another way to treat COVID-19, then, is by treating the immune response, rather than the virus itself.

Cytokine storms are not unique to SARS-CoV-2 or even infectious diseases. They can happen in patients with a genetic disorder, an autoimmune disease, or a bone marrow transplant. Drugs for quelling the immune system in these patients are now being repurposed in clinical trials for COVID-19. Randy Cron, a rheumatologist at the University of Alabama, is planning a small trial for Anakinra, an immunosuppressant currently approved to treat rheumatoid arthritis. Other trials are repurposing yet other drugs on the market, such as tocilizumab and ruxolitinib, which were originally developed for arthritis and diseases of the bone marrow, respectively. Treating a viral infection by tamping down the immune system is especially tricky to balance, because the patient still needs to clear the virus.

Moreover, Cron says, the reports of COVID-19 patients suggest that the cytokine storm within this disease is unique, even compared to another respiratory disease like influenza. “This one really starts fast in the lungs,” Cron says, but with less damage to other organs. Biomarkers of cytokine storms aren’t as “screaming” high as they usually are, he adds, despite the high level of lung damage. COVID-19 and the virus that causes it are, after all, still incredibly new to science.

Much of the early research into drugs against COVID-19 has focused on repurposing existing drugs because they are the fastest way to get something to a patient in a hospital bed. Doctors already know their side effects, and companies already know how to manufacture them. Unless researchers get very lucky, though, these repurposed drugs are unlikely to be a cure-all for COVID-19. Still, they might just work well enough to keep a mildly ill person from becoming severely ill, which is enough to free up a ventilator. “We can do better probably as time goes by,” says García-Sastre, “but right now we need something to start.”

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.