It’s not easy to spot filopodia. Thin wisps of protein less than a micrometer across, they dangle from the surface of a cell, probing its surroundings much as we might grope our way around a darkened room. Usually, they help the cell explore and communicate with its neighbors. But recently, a team of researchers imaging cells infected with SARS-CoV-2 noticed some strange filopodia behavior. Seen through an electron microscope, the filopodia resembled the sprouting tubers of a long-abandoned potato. The tendrils were gnarled and extended, branching outwards at the tips. And all along them were buds of the virus. They seemed to be hitching a ride—part of their onward journey to nearby cells.
The discovery of these unusual filopodia was, according to Nevan Krogan, a systems biologist at UC San Francisco's Quantitative Biosciences Institute, largely by happenstance: His team had been searching for drugs that might hamper the virus’s ability to turn cells into replication factories. To do that, they look for changes that occur in infected cells—which proteins fade or multiply, which cellular processes are turned off or on—deciphering how the virus goes about its hijacking. Then they look for drugs that throw a wrench into the virus’s plans. In particular: old drugs, which have already been approved to treat other diseases or are on their way to it. The results from the team’s recent search are expected to publish Saturday in the journal Cell.
Krogan’s team—a vast collaboration that includes dozens of researchers around the world—is one of a number of groups engaged in a strategy known as drug repurposing. The idea is to short-circuit the long road to a new treatment. It works, in part, because many existing drugs are fairly blunt instruments, exploiting weaknesses that are common across different pathogens. Take remdesivir, which was developed by Gilead over a decade ago to treat Ebola patients. In that quest, it failed. But now there’s evidence it can help treat Covid-19. Other repurposing efforts take advantage of the common ways the body responds to different pathogens. That’s the story for dexamethasone, an old steroid normally used for ailments like croup that the UK’s Recovery trial recently found effective in calming the overactive immune response in critically ill Covid-19 patients.
To find promising candidates, researchers often comb through libraries of thousands of compounds and apply doses of them to infected cells, looking for ones that get rid of the virus at a safe dosage. But Krogan argues that the process can be more targeted than that. While a brute force approach may eventually yield a hit, the less you know about how a particular drug works, the more hurdles lie ahead. He points to the antimalaria drug hydroxychloroquine, which worked well in a dish, but turned out to be no better than a placebo in preventing Covid-19 infections among exposed people. “There are thousands of people testing hydroxychloroquine without understanding what it does,” he says. “It kind of drives me crazy.” So instead, the team tries to develop leads by examining how the virus rewires a cell’s basic biology, turning off some functions while amplifying others.
In an initial study, published in April in the journal Nature, Krogan’s team examined the interactions between proteins in the virus and human cells, then looked for drugs that would interfere. They published a map of those interactions, which researchers and drug companies have scoured for candidates for further animal studies and clinical trials.
“It’s a lot more critical to identify where that drug is working, because ultimately you don't want to go blindly into a human trial or even an animal experiment,” says Jason Kindrachuk, a microbiologist at the University of Manitoba who isn’t involved in the research. The approach is by no means a guarantee, but it’s a good way to generate leads that researchers can study more deeply. “All of this information is amazingly useful,” he says.
The new UCSF-led research takes a slightly different approach. This time, they measured changes in an activity called phosphorylation, which plays a role in flicking cellular processes on and off: everything from how the cell grows and divides and dies to, yes, how it develops filopodia. The activity is managed by a type of protein called kinases. The researchers infected cells from African green monkeys, known for their high susceptibility to SARS-CoV-2, and, using mass spectronomy, looked at how phosphorylation changed over time. They then mapped the activity to analogous kinases in human cells, finding substantial changes in 49 of them.
Some of the affected pathways seemed intuitive, inhibiting cell processes like division and apoptosis, or cell death. “It’s a cool snapshot,” says Julia Schaletzky, a drug discovery researcher at UC Berkeley who wasn’t involved in the study. “The virus just wants to make sure the cell starts working for you, the virus, ASAP. It drops everything except for making more particles.”
The next task was to find drugs that block the misbehaving kinases. “What’s so great is there are so many different drugs that work against them,” Krogan says. Many of the team’s candidates were cancer drugs, designed to quell cellular processes gone haywire. To source them, Krogan turned to a colleague at UCSF, chemist Kevan Shokat. “You go in his freezer and you can find an inhibitor for any kinase. And if he doesn’t have it, he knows how to get it,” he says.
Collaborators in Paris and New York then tested 68 compounds known to inhibit the kinases, searching for those that could eliminate the virus at a dose that didn’t also kill the monkey cells. In that respect, most didn't pass muster. But six or seven of the compounds, Krogan says, were particularly promising, including a handful that were “more potent than remdesivir,” which they used as a control.
Among them was a compound called silmitasertib, which inhibits a kinase called CK2, and is being tested as a cancer drug. The researchers had already identified it in their prior Nature study based on protein interactions. And in the course of that research, they also noted interesting structures emanating from the infected cells. They turned to yet more collaborators—this time at the University of Freiburg in Germany and at Rocky Mountain Labs in Montana—to image the infected cells in greater detail using electron microscopes. That’s when they found those unusually spindly filopodia, along with clusters of both the virus and CK2.
“It’s exciting but preliminary,” says Jenny Gallop, a biochemist at the University of Cambridge who studies filopodia. With her own lab temporarily shuttered due to the pandemic, she’d been scouring the internet for signs that filopodia might play a role in SARS-CoV-2. She had reason to suspect they might. Viruses, including Marburg and Ebola, have been found to hijack filopodia, forcing them to grow longer, branch, and ferry virus particles into their neighbors. The mechanism is especially well studied in another virus called vaccinia, she notes, where research has suggested it plays a role in “superspreading” the virus from one cell to others.
The difficulty, Gallop says, is that imaging tiny filopodia is an arduous process. The level of detail the researchers managed in a short time is no small feat, but more experimentation and imaging is needed to confirm whether the researchers have identified the right mechanism. CK2 is an important kinase that plays a role in many cell processes, she notes; plus, other kinases have roles in filopodia development. A next logical step for firming up the link would be to see if silmitasertib, as a CK2 inhibitor, actually limits the growth of filopodia in infected cells.
That’s one limitation of basing a search for drugs on cellular changes, Schaletzky notes. While the underlying biology can be useful, it can be hard to draw causative links, because we only know so much about how cellular processes actually work. “It’s an educated guess model,” she says. “Not surprisingly, a lot of things change in the cell when you expose it to something new.” But she adds that the more potent compounds in the Cell paper could be good leads for further study. “It’s exciting to see tangible outcomes from an approach like this,” she says.
Moving forward, one challenge for the kinase inhibitors, in particular, is their toxicity—they may end up interfering with processes that certain cells need to survive. Plus, some drugs might not prove appropriate in the context of Covid-19 treatment. Because much of the viral replication happens early in the disease, kinase inhibitors are more likely to work as an early intervention, Schaletzky points out. Later on in disease, the greater danger becomes inflammation from a haywire immune system reaction. (That’s one reason why steroids like dexamethasone are useful for treating the severely ill.)
“Ultimately what’s really going on during infection—which tissues are infected, the immune system reacting, the timing, the dosing—it’s a different story,” she says. What works in a dish of monkey cells may not work in a living animal, much less human lungs. “Even if you find a perfect inhibitor against some obscure kinase that might help mess up SARS-CoV-2’s plan in monkey cells, will it work in an animal model?” she asks. At a certain point, the only way to know is to test.
Krogan says he plans to do just that. Some of the compounds identified in the study are already in the clinical pipeline for Covid-19—and for those, he’s happy to have companies take on the expensive task of testing them. For the others, he says, “if they don’t, we will at some point.” Of particular interest, he adds, is testing combinations of compounds, noting the success of the three-drug cocktail currently used to treat HIV. Perhaps they’ll try one of these compounds they identified along with remdesivir, to see if it improves the virus-clearing effects at a lower dose. “By doing the biology first, you can be so much more intelligent when you get to specific combinations,” he says. “The more you know about your foe, the easier it is to defeat.”
Update 6-27-2020 7:00 PM EST: This story was updated to correctly identify the UCSF researcher who sourced the kinase inhibitors tested in the study.
- The country is reopening. I’m still on lockdown
- What's confusing about calling cases “asymptomatic“
- Should I send my kid back to day care?
- If the virus slows this summer, it may be time to worry
- Glossary: Too many buzzwords? These are the ones to know
- Read all of our coronavirus coverage here
