A Wisconsin City Experiments With a Faster, DIY Covid-19 Test

EACH MORNING AT 7, Brian Wolf, the fire chief of Racine, Wisconsin, arrives for work in a neglected corner of a building called the City Hall Annex. A stately, Depression-era brick structure on a not-so-stately stretch of parking lots and auto glass shops, the space once held the workshops of a vocational college. Later, it became home to the city’s boxing club—a sign still hangs in the room that held the old ring. More recently, the space was used for storage. “It’s basically a warehouse,” Wolf says. But a few weeks ago, the Annex’s ancillary purpose became an asset. Old items were cleared out, the walls and floors scrubbed down, and rooms stocked with an array of lab tools and boxes of test tubes and pipettes. The result is perhaps the world’s most unassuming Covid-19 testing lab, where Wolf now presides over his fire crew as pipetter-in-chief.

At the start of a shift, the firefighters don gowns, goggles, and medical masks. Then they await a stream of city workers—permit issuers, police officers, lifeguards—who drop by on their way to work. Under careful supervision, each visitor spits into a vial. And for the rest of the day, the firefighters perform the tests, heating up the samples and pipetting liquids from one test tube to the next. “We’re slow at it,” says Wolf. “But like any emergency we respond to, we have to work together. No bickering. No errors. We get it done.” If all goes well, they’ll have results by lunch. They await a simple change of color in the test tubes: Pink is the all-clear; no virus detected. Yellow is evidence of SARS-CoV-2.

The technology used in the tests is called LAMP, or loop-mediated isothermal amplification. It’s similar to PCR—or polymerase chain reaction—the common way of diagnosing Covid-19, in that it looks for genetic traces of the virus. But it’s a stripped-down process, requiring fewer chemical ingredients, test tubes, and equipment. Typically, the method is used to test for diseases like Zika and dengue fever in places that don’t have lab access—or else by farmers who want to find out what pathogen ails their soybeans. But in Racine, it’s part of an experiment: to see how a cheaper, simpler test, if done frequently enough, might hold outbreaks at bay. Instead of sending samples off to a hospital or lab, and potentially waiting days for results, could you test for Covid-19 from the back of a van, or outside a school, or in a city-owned former boxing gym?

The tests had found their way to the shores of Lake Michigan by way of Chris Mason, a professor at Cornell-Weill Medical School, Racine native, and younger brother of the city’s mayor. Earlier this spring, Mason had published a preprint of his work on LAMP, and then posted it on Facebook and Twitter. It caught his brother’s attention. Their hometown, much like everywhere else, lacked adequate testing for the new virus, and in April, Mason’s brother called, inquiring about his research. The mayor told him that cases were surging, and asked if the test could be put to use.

The idea behind LAMP testing, according to Mason, is to give every town or city its own capacity to test closer to home. While few places have their own clinically certified labs, “everybody has a fire department,” he says. Because people can have a Covid-19 infection without symptoms, the major barrier to any return to a semblance of normal life—back to school, to work—is the need to test lots of people, regularly, whether they show signs of Covid-19 or not. An estimate from Paul Romer, a Nobel Prize–winning economist at New York University, puts the requirement for doing so safely at more than 35 million tests per day.

Frequent testing would allow for the use of slightly less accurate—and potentially cheaper—tests, since missed infections would likely be caught by subsequent rounds of testing. The key is to return results within hours or even minutes, says Jeremy Kamil, a microbiologist at Louisiana State University. That’s useful for quashing outbreaks before they spread, and also critical for avoiding too much disruption.“The savior test will be something that’s economical and reliable in one hour directly from saliva,” he says. You could intercept infected people before they spend many hours together at work or to pack the gym for the high school basketball game. If they tested positive, they would be asked to self-isolate. The rest could go about their day.

For Mason, LAMP appeared to check all those boxes. He had toyed with the technology as part of his research in viral surveillance, which involves wiping microbes off subway turnstiles and hospital bathroom walls to study which ones are lurking in public spaces. So the brothers devised a plan: City workers would get tested with LAMP, whether they were showing symptoms or not, and the data would offer clues as to whether the test was accurate enough to be used elsewhere. Alison Kriegel, a physiologist at the nearby Medical College of Wisconsin and an RNA specialist who has known Mason since elementary school, would oversee the operation in Racine, training the firefighters in how to safely take samples and run the tests. Mostly, it would involve a lot of pipetting.

And no matter what, they thought, it would be better than having no tests.“It was scary, honestly,” says Kriegel of those days this spring. “The testing was in short supply, and no one knew what was going on. What if we can’t get tests? How do we do this on our own?”

First developed in the 1990s by researchers in Japan, LAMP is a slightly younger cousin to PCR. At a high level, it works similarly. As with PCR, the test uses a series of chemical reactions to amplify the genetic material in a sample, making it easier to find a match with a particular viral RNA sequence. But with PCR, those reactions happen at a variety of temperatures—hence the need for automated machinery. In LAMP-based tests, the reactions occur at a constant temperature. That means a warm bath suitable for heating the samples to 95 degrees Celsius, plus some pipettes, tips, and a few chemicals, will do. (In principle, Mason says, you could do it on your kitchen counter, using a microwave.)

An added benefit is that those chemicals differ from the scarce supplies currently used for PCR tests. They’re also, at least currently, cheap—a recent preprint from researchers at Harvard Medical School suggested a sample could be processed for as little as 7 cents, not including inevitable overhead and markups if tests go commercial. (For comparison, a PCR test at a high-throughput lab costs about $40, though $100 is a more typical insurance bill.)

In February, Mason’s team in New York had reached out to a contact at New England Biolabs, a company that makes LAMP reagents. The company sent them a copy of the data and methods developed by a team in Wuhan that had begun experimenting with screening for SARS-CoV-2 there. Mason’s team quickly set up the procedure for themselves, testing the method on a set of eight known positive samples. “It was close to a eureka moment,” he says. “It worked in 20 minutes. Bright yellow.”

They and other research groups quickly set about improving the method, with a goal of whittling down the use of complex lab equipment. Taking diagnostic tests out of clinical labs is no easy task. “LAMP is a simple recipe,” says Melis Anahtar, a microbiologist at Harvard University who has helped develop both PCR and LAMP-based tests for Covid-19. “But it really has to be extremely easy and foolproof.” In a typical clinical lab, every detail is precisely regulated: from the ambient temperature of the room to the calibration and cleaning of each instrument. A contaminated pipette, a hot-water bath that’s off a few degrees, or even just bad record-keeping could easily throw off results. Plus, there’s always a concern that a poorly-trained staffer could mix up the test tubes or expose themselves to an infectious sample.

In recent months, researchers from dozens of universities and companies have come together to discuss those challenges in an informal group Mason calls “gLAMP,” or the Global LAMP Working Group. He calls it a “pre-competitive space,” where people can openly share data and methods. Some are trying to develop procedure usable in mobile or pop-up labs, like the one in Racine; others to see how LAMP could work in centralized labs. “We’re trying to launch as many biotechnology ships as possible and see what reaches the far shore,” Mason says.

On a recent weekly Zoom call, the LAMP researchers discussed one of the major curveballs: How to work with spit. Observing people as they spit into a tube is less dangerous for health care workers, who don’t need to venture close with a long nasopharyngeal swab, which reaches deep into the nose. It’s also critical for avoiding what Dave O’Connor, a virologist at the University of Wisconsin-Madison, calls “prevention fatigue.” For example, if teachers at a school, who otherwise feel perfectly healthy, come to dread their twice-weekly swab, surveillance testing will quickly become unreliable. “They’ll say, ‘I feel fine’ and find a way to skip it,” O’Connor says. “We’re a nation of wusses, myself included.”

For LAMP tests, saliva poses certain technical difficulties. The appeal of “colorimetric” tests is their simplicity: Interpreting the change from pink to yellow is so easy “a monkey or a turtle could do the analysis,” Mason says. But there can be complications. “You might have dirty samples,” Anahtar says. “You might say, ‘Don’t eat anything before the test,’ but then one person gets some broccoli in the saliva.” That extra gunk can change the acidity of the sample, throwing off the final color. Sometimes it might end up a little orange. One solution is to add a fluorescent material to the sample—a step added to the process in Racine—which eliminates the need to eyeball a color change. The cost is an extra piece of machinery to analyze the fluorescence.

The other concern is improving accuracy. PCR testing involves a step called RNA extraction, involving chemicals that help ensure the correct viral material is amplified and ultimately detected. LAMP tests don’t have that step, which results in a less sensitive test that can miss some infections. Even in safeguarded lab settings, some studies have put the figure around 10 percent.

But O’Connor notes that the accuracy should be considered in the context of how the test will be used. Viral loads are low at two points during an infection: immediately after exposure and later on, during the illness, as the virus peters out. If testing is frequent enough, those in the just-exposed group will be caught by future tests. Those in the nearly recovered group tend to be the least infectious and thus at lowest risk of sparking new outbreaks.

The plan is to bring more real-world data to those assumptions. Last week, O’Connor, who is working independently of the Racine team, began driving a white Dodge minivan around Madison with the equipment for his LAMP test rig, performing tests of staff at the Wisconsin National Primate Research Center on the building’s rooftop. “I did not expect to be pipetting in the hot July sun,” says O’Connor, whose primary expertise is in vaccines, not diagnostics. “We started doing this when we realized labs weren’t going to do it. I see it as a sign of abject failure that my lab is doing this instead of doing animal model work we specialize in.”

But so far he’s been happy with the progress. On Wednesday, his team hauled the equipment to his 11-year-old son’s school, where a group of 28 teachers and staff volunteered to come in to try out the tests—a trial run for what a school reopening might look like. Surveillance testing, O’Connor notes, will work only if the virus is largely brought under control first. It’s designed to catch asymptomatic people before they spark new outbreaks, not stop those that are already raging.

Right now, Wisconsin is in a gray area. A month ago, schools seemed poised to reopen in the fall, but with cases now spiking again, the outlook is uncertain. “The ideal thing would be to have kids here safely. But some of that is out of my control, because it’s a community function,” says Andy Wright, principal of the Eagle School, where O’Connor’s son is a student. “Can I push it further than other places if we have testing like this? I think I can, and I think parents and teachers would feel much safer.” Ideally, O’Connor would test the kids too, if they can get approval from both parents and the University of Wisconsin’s research board. (Kids, he notes, bring their own challenges to saliva-based sampling.)

In the end, the primary barrier for surveillance testing is not likely to be technological, or a matter of getting supplies. (The biggest challenge O’Connor has had obtaining materials, he says, is getting ahold of Lysol to wipe the outsides of the saliva vials.) It’s regulatory and logistical. The key benefit of LAMP—that the recipe is modular and easily adapted to different settings—puts it at a potential disadvantage with regulators. While the Food and Drug Administration has approved some so-called point-of-care tests for Covid-19, they’re typically well-packaged: They run on proprietary machines, usually found only in hospitals, and rely on single-use cartridges that cost as much as $40. So far, the FDA has authorized LAMP-based tests as well as Crispr-based tests, which layer the gene editing technology on top of LAMP procedures to improve the specificity of the test, but only within the context of certified labs.

“In practice, it’s going to be very, very difficult to conform to the current regulatory standards,” O’Connor says. One strategy would involve seeking clinical certification for a mobile lab that could provide diagnoses using a standardized set of instruments and procedures. Another would involve testing that does not involve a diagnosis but serves instead as screening for who should get a PCR test for the final word. But neither case has a clear regulatory framework to follow. “We need this to be part of a coordinated organized response,” he says. “I don’t believe we lack the logistical wherewithal to develop a plan for a test where you spit into a tube and you heat it up and wait for an hour. I could train my 11-year-old to do this in a day.”

For now, the experiments in Madison and Racine are being done as research; the teams will generate data to validate the tests, not provide diagnoses. But in practice, they’ll offer a form of screening. In Racine, if a sample turns yellow—positive—Kriegel calls the person at work and tells them they have a result of “potential clinical significance.” She asks them to come back to the Annex as soon as possible for a nasal swab, which the firefighters perform and then ship off to a lab in Chicago for PCR analysis, to confirm the result. Some people take the news in stride. One worker arrived in fretful tears. “That’s the whole challenge,” Kriegel says. “How do you do this kind of testing without making people constantly fearful?”

But Wolf, the fire chief, says he sees signs that the city is getting more comfortable. Word is getting around, and more volunteers among the city staff are showing up to be tested; Mason’s hope is to open the tests to Racine’s 77,000 residents soon, if they can sort the logistics and funding. In the meantime, Wolf and his crew are trying to get better at their pipetting—and at assuaging the public’s fears. “I just try to reassure them: Wear a mask, isolate as best they can,” he says. “If they’re careful around others, everyone will be OK.”