COVID-19: the biology and effective therapy 
By John Timmer - became Ars Technica's science editor in 2007 after spending 15 years doing biology research at places like Berkeley and Cornell.

Posted April 9, 2020.

This article was selected for the general public because it was written by John Timmer in a format that is easy for most to understand.

A coronavirus vaccine may not arrive for at least a year—so what are the chances of finding a useful therapy that could stave off the worst effects of the virus in the meantime?

Earlier coronavirus outbreaks like SARS and MERS raised warning flags for public health officials. Fortunately, they also alerted the biological research community that this large family of viruses was worth studying in more detail. Recent research has built on a large body of knowledge about coronaviruses that have long caused significant diseases in livestock, and so SARS-CoV-2 does not arrive as a total unknown. Indeed, we are actually in a decent position to understand what might make a good potential therapy.

While some of the therapies being tested may seem random—we're trying chloroquine, an antimalarial drug?—there's serious biology behind what's being done.

Genes without DNA

A basic challenge confronts all viral therapies: most viruses have just a handful of genes, and they rely on proteins in the cells they infect (host cells) to perform many of the functions needed to reproduce. But therapies that target host cell proteins run the risk of killing uninfected cells, making matters worse. So antiviral therapies usually target something unique about the virus—something important enough that a few mutations in the virus won't make the therapy ineffective.

Those of you who didn't sleep through high school biology may remember that genetic information is carried by DNA. When a protein needs to be built, the relevant bit of DNA is read and the cell makes a temporary copy of the information using a very similar chemical called RNA. This piece of RNA is then translated into a sequence of amino acids, which form the protein. While there are some exceptions to this—many RNAs perform important functions without ever being translated into proteins—all RNA in our cells is made by transcribing a DNA sequence.

But we've known for a long time that this process doesn't hold for viruses. Many viruses, including HIV and the influenza virus, use RNA for their basic genetic material. The coronavirus is also an RNA virus; it consists of a single, 30,000-base-long RNA molecule.

This is a problem for the virus. The host cells it infects only have proteins that copy DNA, not RNA, so how can more copies of the virus get made?

Target: reproduction

It turns out that the virus carries its own solution with it. When virus' RNA genome first enters a cell, it interacts with the host's protein-making machinery, using it to make proteins that can copy RNA molecules.

These RNA-copying proteins, called "polymerases," make an enticing target for therapies. Because host cells don't naturally have them, therapies that target these RNA-making proteins should have a lower chance of off-target effects. Block these RNA polymerases, and the virus can no longer reproduce, stopping an infection. That's the good news.

The bad news is that DNA and RNA are so closely related that it can be difficult to make a drug that affects only one type of polymerase. We saw this with some of the first therapies against HIV, which targeted the enzymes that copied the virus' RNA genome: they did slow the virus down, but they also harmed any rapidly dividing cells in the host.

coronavirus genome
coronavirus genome

So the work is tricky. But many such drugs have been developed that don't interact as well with our own DNA polymerases. Some have even been tested for safety in humans, since they were developed for earlier threats like HIV or Ebola. Now, several are being quickly tested against coronavirus.

One such drug, remdesivir, was originally developed in the hope that it would limit Ebola virus and its relatives. While that hasn't worked out, the drug was safe for human use and showed promise in its ability to limit the spread of another coronavirus (MERS-CoV) in cultured cells. As a result, it was quickly tested against SARS-CoV-2, and the results were also positive. The National Institutes of Health started a clinical trial against COVID-19 in February.

Vincent Racaniello is a faculty member at Columbia University and the host of the This Week in Virology podcast. He believes that RNA polymerases are so similar across a range of coronaviruses that we might find a single molecule that inhibits them all. To Racaniello, our response to SARS and MERS wasted a great opportunity.

"We could have had a broadly acting antiviral that targeted RNA polymerase by now," he told Ars. "We could have had people isolating the gene from various bat coronaviruses and doing screens to see if we could find compounds that could have inhibited them all. That's the kind of thing that's doable and should have been done. And if we had such antivirals ready, they could have been used right at the onset in China."

Target: processing

RNA copying polymerases aren't the only potential therapeutic targets for a coronavirus. Their RNA polymerases are initially made in forms that aren't fully functional; instead, they must have small pieces snipped out in order to adopt their mature configuration. Coronavirus RNA therefore encodes two or three proteins that do this cutting. They belong to a class of proteins collectively termed "proteases" for their protein-cutting ability. Proteases typically have a very specific site where the cutting takes place, and any chemicals that can fit into this site might shut the protease down. Not surprisingly, such chemicals are called protease inhibitors.

This approach has been used successfully against other viruses, notably including HIV. Scientists have now found that protease inhibitors targeted to HIV might have activity against coronavirus, despite the fact that these viruses are unrelated.

Because proteases are present in small numbers in infected cells and have a catalytic activity that depends on a single, specific site, Racaniello views them as some of the most promising targets for therapies. We've also got large libraries of chemicals that are known to inhibit similar proteins, many of which are already approved for use in humans. So, while the news around protease inhibitors has been somewhat limited, expect it to pick up dramatically as more of these molecules are screened.

coronavirus structure
coronavirus structure

Target: packaging

After replication, viral RNA can't continue an infection until it is packaged up into a mature virus and gets outside of the host cell. This requires special packaging proteins. (In coronavirus, these proteins do double duty by also helping the viral RNA link up with its copying enzymes.) This packaging step would seem to provide a great opportunity for targeted therapy, as disrupting it should limit the amount of functional virus that gets made and exported from any particular cell.

But drugs that try to block viral packaging are rare—Racaniello can only think of one, a treatment for Hepatitis B that causes the mature virus particles to form without any genetic material inside. "That's been a very unusual antiviral," Racaniello said. "There's no other like it." Part of the problem, he said, is that structural proteins like this are present in high numbers, since they're part of every single virus particle that's produced. And you have to interfere with all these copies to be effective.

Another problem is that the interactions among proteins and genetic material during packaging of a virus tend to involve extensive contacts between multiple molecules. These are a bit harder to disrupt specifically, and doing so may require large molecules that don't diffuse in and out of cells well. So, while we know which protein binds to the RNA and helps package it inside the virus particle, this protein is not an obvious target for therapies.

It's also hard to disrupt newly packaged viruses as they are moved out of the cell. Once packaged, coronaviruses leave their host cell via an export system that's normally used to send material to the cell's surface (a process called exocytosis). This process is fairly generic—it works with a huge variety of proteins in addition to those encoded by coronaviruses—making it vital for cell survival. As a result, there are not many places where we can intervene without shutting down exocytosis in healthy cells as well.

Target: the viral shell

Once we have a mature virus particle, its behavior is controlled by the proteins that form the exterior structure of the virus. In the coronavirus, two of these proteins (called "membrane" and "envelope") combine with some of the cell's membrane to form the virus' shell. There's also the spike protein, which creates a halo (or "corona," meaning "crown") around the virus that gives it its name—and which serves to latch on to cells to enable infection.

In some coronavirus strains, the envelope protein can be eliminated without blocking the virus from infecting cells, which means it's a lousy target for therapies.

The membrane protein is the most abundant protein on the surface of the virus, but it's small and buried within the membrane (as its name implies). Not much of it is accessible to the outside world. Combine that with the fact that it doesn't appear to have an enzymatic function, and it's not an ideal target, either.

That leaves the spike protein. Spike is a complicated protein that provides a wealth of targets for potential therapies. As the most prominent feature of the virus' exterior, spike is the main target of antibodies against the virus produced by the immune system.

coronavirus surfacespike-protein
Coronavirus surfacespike-protein

This reality has already led to one option for therapies: purifying plasma from people who have fought off a coronavirus infection, on the assumption that the plasma contains antibodies that can neutralize the virus. This plasma can then be infused into sick people, where the antibodies should help the immune system clear the virus. While it's only a temporary fix—antibodies don't survive indefinitely in the blood stream—it may give a patient's own immune system sufficient time to develop its own antibodies.

There are unknowns about whether infected individuals produce effective antibodies—more on that immediately below. But the big issue here is scaling, as plasma treatment relies on having enough healthy, formerly infected individuals who are willing to donate blood plasma. If used strategically—on the most at-risk patients, or to help infected health care professionals—it could be a helpful tool, but isn't likely an effective general therapy.

A different approach to antibodies

But antibodies therapies aren't limited to infusing blood plasma. Once the immune system generates cells that produce anti-coronavirus antibodies, we can pull out the genes that encode these antibodies, insert the genes into plant cells, and get those cells to pump out large quantities of the antibodies. With a bit of time, we might even produce a cocktail of several antibodies that all bind to coronavirus, and do so in quantities that could make this an effective therapy for those infected. (This approach was tried during an Ebola virus outbreak.)

This approach will take longer to develop and vet for safety, so it won't be the quick fix provided by blood plasma. But it does offer the promise of scale, producing sufficient quantities of the therapy to treat entire populations.

It also provides us with the ability to carefully select the antibodies we produce. While our immune systems produce antibodies to viruses like HIV and influenza, many of these bind to parts of the virus that can easily change through mutations. That makes them ineffective, since the virus has an opportunity to evolve. What we need are "broadly neutralizing" antibodies, which seem to bind to parts of the virus where mutational changes can't occur without compromising its basic function. In many cases, broadly neutralizing antibodies turn out to stick to the parts of the virus that latch on to human cells to start new infections, and thus they block the virus' ability to infect anything.

At the moment, we simply don't know how much the proteins on the surface of SARS-CoV-2 can change while still retaining their function. We can make some inferences based on what we've seen in other coronaviruses, but experts have reached somewhat different conclusions. This is a research area to watch carefully, because the rate of change in the surface proteins will dictate how effective antibody-based therapies are—and how easy it will be to develop a vaccine.

Fortunately, things are moving quickly, with one company announcing on Wednesday that it has identified hundreds of antibodies that target SARS-CoV-2. It estimates that it could have sufficient production for testing by the summer and be making hundreds of thousands of doses a month by the end of the summer.

Target: new infections

The final step in the virus' life cycle is infecting a new cell. Typically, what is taught here is a "lie of simplification," which goes: the virus latches on to a protein on the cell's surface, then uses that protein to gain entry into the cell. This is true as far as it goes, but for most viruses, things are considerably more complicated. Coronaviruses definitely fall into the "more complicated" category in this regard.

SARS-CoV-2 does latch on to a protein on the surface of cells in the respiratory tract; we've already confirmed that it's the same protein as the one used by the original SARS-CoV. But that doesn't immediately result in viral contents entering the cell. Instead, the complex of virus and receptors stays on the outside of the cell membrane. That membrane, however, gets pulled into the cell and "pinched off" from the cell's surface, creating a sac within the cell that now contains "outside" material.

coronavirus endocytosis
Image Enlarge

Once this occurs, the virus is technically inside the cell, but it's still on the wrong side of a membrane from everything it needs to reproduce.

The cell takes over this compartment, lowering its pH and adding enzymes to break down its contents. Corona and other viruses actually take advantage of these changes to enable their infection. In the case of coronavirus, a protease made by the host cell cuts the viral spike protein. Once cut, the spike protein triggers a merger between the membrane in the virus' coat and the membrane of the compartment it is trapped in. This finally places the virus' genome inside the cell, where it can proceed with the infection.

This series of events provides potential targets for therapies. One of these targets is the drop in pH. This is the step that's targeted by chloroquine, the antimalarial drug. Chloroquine can cross membranes and so can enter the sac containing the virus. Once there, it can neutralize the pH.

That's significant, because many proteases are only active at lower pH. If the pH inside the sac doesn't change, it's possible that the coronavirus spike protein won't be cut and thus won't be activated. This appears to be the case in cultured cells infected by the virus, and there are anecdotal case reports of chloroquine helping COVID-19 patients.

The host cell proteases themselves also make a tempting target. A paper we mentioned above identified a protease that appears to be essential for the coronavirus spike protein's activation. That team showed that an inhibitor of this protease blocked coronavirus infections in cultured cells. The inhibitor has been approved for use in humans by Japan, so this may be another promising avenue for tests. (Racaniello notes that this protease is also used to activate influenza viruses.) The risk here is that the protease in question might also play an essential role in healthy cells.

Finally, it's tempting to directly target the interactions between the spike protein and the protein it binds. But Racaniello says that these interactions are extensive, and they can be difficult to disrupt with a single molecule. It's been tried with HIV but mostly came up short. The only thing that has worked is a 30-amino-acid-long protein that mimics part of the protein HIV binds to, but that can't be stored in water, and it needs to be mixed up and injected for use—not the sort of thing likely to be helpful when a pandemic is limiting healthcare resources.

Beyond the obvious

There are plenty of options for interfering with coronavirus based on what we already know about its biology. But there are still many things we don't know. A recent article in the New York Times described how scientists have identified hundreds of proteins made by host cells that interact with proteins encoded in the coronavirus genome. We don't know the significance of most of these interactions and whether or not they're important or coincidental, but any of them could potentially lead to a therapy. That would, however, probably take longer than a targeted therapy, since there are more steps involved in screening for effective drugs than there are in, say, screening a library of known protease inhibitors against the coronavirus' proteases.

There's also the potential to intervene at the level of the body's response to the virus, rather than targeting the virus itself. The more damaging consequences of some infections come from an exaggerated immune response to the virus. Biotech giant Genentech, for example, announced on Thursday that it was starting clinical trials of an immune-dampening treatment on hospitalized coronavirus cases.

The potential return for having any useful therapy is so large that it's worth following as many of these paths as we can at once. The easiest way to understand why is to return to the epidemiological model we covered earlier this week. The model indicated that any steps short of extreme isolation measures would likely allow the virus to overwhelm the healthcare system—and any easing off of restrictions could lead to a resurgence within weeks. Extreme restrictions, however, will probably cause severe economic problems, especially if the only hope is a vaccine that might be over a year off.

But the model has an obvious gap: it doesn't account for an effective therapy.

If any of the approaches described above—or one we didn't consider—is even moderately effective, it could radically change our circumstances. It could ensure that far fewer coronavirus cases need hospitalization, and that fewer of those that do require critical care. A country's healthcare system could then continue functioning in the presence of a higher rate of infection, which in turn could mean that less dramatic social restrictions are required. If carefully managed, this might even allow countries to allow enough infections so that they achieve herd immunity before the availability of a vaccine.

We are just beginning clinical trials on a small subset of these ideas now, so we're still facing difficult times in the months to come. And it's important to emphasize that there's no guarantee that any of these approaches will work. But finding a therapy does offer hope that the difficult months of isolation in our immediate future might not stretch to the end of the year.

Coronavirus: what you should know and do:  WebMD: Information coronavirus 2020

Source: Timmer John, "COVID-19: the biology of an effective therapy," ARS TECHNICA, March 30, 2020.  Timmer: Effective therapy for coronavirus 2020