A story about spit, the NBA, the FDA and Proteinase K! And bull manure… Did you just do a spit take? Well, COVID-19 is putting a whole new meaning on spit take-ing! On August 15, the FDA authorized a saliva-based coronavirus testing method called SalivaDirect that bypasses the labor intensive “RNA extraction” step by using a protein-chewer called Proteinase K which was originally found in a fungus grown in manure made from bull horns. As I’ll explain, this EUA (Emergency Use Authorization) is great, but it’s *not* a total game-changer, whereas cheap, rapid tests *could* be.
Diagnostic tests to detect SARS-CoV-2 (the novel coronavirus that causes the disease COVID-19) typically look for the virus’ genetic information. SARS-CoV-2 holds that info in a single strand of RNA. Tests can “reverse transcribe” that RNA into DNA and then use techniques like PCR or isothermal amplification to make lots of copies of snippets of the viral genetic sequence. If you see copies get made (such as through a fluorescence readout), you know the virus was there to copy. Problem is, that viral RNA is inside of the viral membrane and wrapped up around a nucleocapsid protein (N). So the first step of most diagnostic tests to date has been RNA extraction – isolating that viral RNA so you can make copies.
It doesn’t get that much attention, but this RNA extraction step actually can be one of the most time consuming parts – and “reagent” consuming parts. “Reagent” is just a fancy term we use for the chemicals you put into a reaction. And the reagents for RNA extraction typically include a few proprietary mixes of chemicals that you add sequentially to lyse (break open) the cells and precipitate out the proteins and cell gunk. Several other tests using saliva have been approved for some time, but typically the RNA was still extracted from the saliva in that tedious process before doing the PCR. Now, however, the FDA has approved the use of saliva directly – no extraction required! At least no manual extraction like we typically think of it. And the story of how this now-approved method called SalivaDirect came about is pretty interesting – it even made it to ESPN https://es.pn/2YtbGjR
Especially in the US, coronavirus testing has been a major issue – from supply shortages to long wait times that basically make the tests useless. Early on in the pandemic, celebrities were getting a lot of shade when they’d talk about getting tested (even when they didn’t have symptoms) while the rest of America (including symptomatic people) couldn’t get a test. When the NBA decided to restart their season in a “bubble” where the players from all the teams are “confined” to the Walt Disney World Resort in Orlando, they also committed to testing all the players really frequently, like every day, with a quick turnaround so that anyone who was infected would quickly be isolated before hitting the court and potentially infecting others.
I might be being a bit too cynical here, but the NBA probably knew that this would get a lot more side-eye-ing. They’re using a private company and paying out of pocket so it hopefully isn’t taking too much resources from the community. But they also wanted to help give back and contribute to research on making faster, cheaper, testing a reality. Also, the players likely didn’t want to have one of those giant q-tips called an NP swab that’s not really a q-tip (and which was another supply chain issue early on) stuck up their nose to the point where people were saying it felt like it was touching their brain… So they were likely eager to jump on the saliva train!
The NBA league and players’ union therefore teamed up with a research group at Yale led by Nathan Grubaugh and Anne Wyllie to test (and fund) the group’s SalivaDirect method. Actually, at the time they initiated the collaboration, in April, “SalivaDirect” didn’t even exist – an NBA official had just seen an article from the Yale team showing that RNA could be extracted from saliva & used instead of that nasopharyngeal (NP) swabbing, and he’d gotten in touch.
The Yale group decided to go for it – in a way it was a match made in heaven (or at least a match made in the Happiest Place on Earth…) In addition to funding, the Yale team would need a LOT of samples to be able to verify their method and get FDA approval. And the NBA had a lot of $ and was running a LOT of tests. All the players still got nasal swabs, but, if they chose to participate in the study, they also spit into a tube (in addition to not requiring those nasopharyngeal (NP) swabs, saliva has the added benefit that people can easily collect their own samples). The spit samples are then shipped to the Yale lab, tested with the SalivaDirect method & conventional methods and the results are compared.
The NBA study is important because these tests are mainly on asymptomatic or presymptomatic individuals, who can contribute to the spread, but are less likely to seek out testing. If you want to do this sort of “screening testing” it’s important to make sure that your technique can catch these individuals. I think that the Yale-NBA study which they call the Surveillance with Improved Screening and Health (SWISH) Study is still ongoing. But, Yale recently released findings from a study they did on patients from their hospitals. These findings showed that SalivaDirect worked comparably to conventional methods, so the FDA gave them the EUA. Here’s that paper: https://doi.org/10.1101/2020.08.03.20167791
There are a couple of major benefits to the SalivaDirect method as opposed to other methods (even saliva-based ones). A really crucial part is that you don’t have to go through that long RNA extraction step to isolate the RNA. Instead, the SalivaDirect test uses an enzyme (protein mediator/speed-upper) called Proteinase K to make the proteins go away!
We often think about membranes, such as the membrane surrounding the viral RNA, as being made up of oily lipids. And they are – but there are also a ton of proteins embedded in there and contributing to the membrane’s integrity. So adding a protein chewer can disrupt the membranes and let out the RNA. The RNA is still surrounded by and coated in protein, but Proteinase K can chew that all up as well.
Seriously – Proteinase K is a biochemical beast! A lot of proteases (protein-cutters) have specific sequences of amino acids (protein letters) that they recognize and cut. But Proteinase K will cut pretty much anywhere (though it prefers to cut next to hydrophobic (water-avoided) amino acids. And it can even cut proteins when those proteins are folded up into their beautiful 3D structures (as opposed to other proteases that can only work on unfolded proteins or super flexible regions of folded ones).
The K in “Proteinase K” stands for keratin because it was discovered as a protein made and excreted by the fungus Tritirachium album Limber that was able to digest keratin. Keratin is the protein in our hair and nails – and animal horns. Scientists at Merck isolated this fungus from “horn chip manure,” further characterized it, and commercialized it. It’s now used a lot in labs, including during some conventional RNA or DNA extractions, to get rid of contaminating proteins.
And of course, it’s used in this SalivaDirect Method. After using Proteinase K to “isolate” the RNA, you then heat things up to make Proteinase K go away (or at least become inactivated). This is really important because the enzymes you’re gonna use in the next step, the reverse transcriptase you’re gonna use to make the DNA copy of the RNA, and the DNA polymerase you’re gonna use to make DNA copies of that DNA, are both proteins, so if you don’t inactivate Proteinase K, Proteinase K will inactivate those enzymes!
From this point on, it’s “continue as usual” onto your amplification/detection step. But you just saved yourself a bunch of time, energy, and “reagents.”
It’s important to keep in mind, however, that going from spit does NOT make this a “rapid test” – it just speeds up the front-end process. The amplification/detection step (e.g. RT-PCR or isothermal amplification technique such as LAMP) still takes however long that part takes. And has to be done in a special machine.
However, a great thing about the SalivaDirect technique is that it’s compatible with all sorts of different amplification mixes & special machines. One of the problems with testing has been that different mixes have to be used with different machines sold by different companies, etc. but since SalivaDirect is just a “method” to get the RNA ready it can be broadly applied. With the major caveat that it can be broadly applied only in “high complexity labs” where you have the machines, etc. to carry out the amplification and detection. In FDA terms: “Because this test does not rely on any proprietary equipment from Yale and can use a variety of commercially available testing components, it can be assembled and used in high-complexity labs throughout the country, provided they comply with the conditions of authorization in the EUA.” https://bit.ly/2CUPEz6
One of the things our country really needs is true rapid tests that don’t need to be done in CLIA (Clinical Laboratory Improvement Amendments)–certified lab. We need tests that are Point Of Care (POC) so they can be done at clinics, etc, and we especially tests that can be done *anywhere* (home, school, work, etc.) without requiring fancy equipment. And we need these tests to be cheap so that people can use them frequently (e.g. 3X a week).
A Harvard physician scientist named Michael Mina has really been leading the push for such tests (or at least leading it in the media, no clue what goes on behind the scenes!)
short tangent note: I heard Mina interviewed on TWiV (This Week in Virology) which is a really great podcast/now videos too FYI – as are the other podcasts put out by the same virologist, Vincent Racaniello. My favorite is Immune. Completely outside my field, but the hosts are terrific and I love that they walk through scientific papers in detail, pointing out pros and cons, what they would have done differently, etc. and it’s totally at my grad school geek level.
Anyways, here’s a link to the Mina interview (TWiV 640) if you want to watch/listen: https://www.microbe.tv/twiv/twiv-640/
And here’s a summary of what the TWiV team’s been referring to as the “Mina Method.”
In late June, Mina was part of a team which put out a preprint (one of those not-yet-peer-reviewed articles in medRxiv modeling how testing frequency is way more important than testing accuracy. Here’s a link to the paper https://doi.org/10.1101/2020.06.22.20136309 & I walk you through a couple of its figures in my figures. But here’s the gist: the problem, as they see it, is that RT-PCR is in a way “too good” at finding viral genetic information (i.e. it’s incredibly sensitive) and the FDA is basing test authorization based on the RT-PCR gold standard. So it’s really hard for less sensitive tests to get approval.
At first you might think that this is a good thing – I mean, don’t you want your tests to be really sensitive? One of the problems is that there can be a big (like orders of magnitude) difference between the virus levels detectable by RT-PCR and the virus levels that correspond to contagiousness. In the really early stages of the infection, even before a person has symptoms, they’re shedding the most “live” virus. After a couple of days, they’re much less contagious and, even after they’re not contagious at all, they can shed “dead” virus (sometimes for weeks or even months).
And RT-PCR can still detect this. Because all it’s detecting is (DNA copies of) the viral RNA which, on its own isn’t dangerous. The RNA can’t do anything unless it’s encapsulated in a membrane with that Spike protein it needs to dock onto new cells and get inside, etc. So just detecting viral RNA doesn’t mean that there’s “replication competent” virus.
note: viruses, even though they’re super duper clever, are never considered “living” so although we use the terms “live” and “dead” a lot to refer to them, it’s not jargon-ly accurate so I should be using terms like “replication competent” or “active” instead of living and “inactive” instead of dead but “live” and “dead” are used pretty commonly even amongst scientists and I think the terms help make it more understandable to the public (or at least that’s how I justify it to myself…)
The CDC has a great compilation of articles and data from numerous studies showing that although viral RNA has been detected in patients even several months post-infection, “For patients with mild to moderate COVID-19, replication-competent virus has not been recovered after 10 days following symptom onset.” The CDC therefore recommends that, after 10 days post symptom onset, people no longer need to isolate. https://bit.ly/3j7AIge
This was a change from CDC’s original policy of requiring 2 negative tests before declaring someone “safe” to reenter society. And the change came because more and more evidence was mounting that there’s actually a fairly short period when patients are really contagious and that patients who had “high Ct values” (e.g. in the 30s) didn’t actually have “live” virus in their samples. Before you get scared off by terms and acronyms, let me explain…
RT-PCR (our gold standard diagnostic test) typically detects the copies of viral genetic info that get made using fluorescent probes that bind to those copies and let off light. The reaction is run in cycles of separating the DNA strands, making copies, separating those, making copies of the copies…. The more copies you start with, the faster the fluorescence signal will build and the tests report the “Ct” (cycle threshold) value which is how many cycles it took for the signal to reach a certain threshold above noise. The more you start with, the fewer cycles it will take, and thus the lower the Ct value. Since PCR is exponential, a difference in 3 in the Ct value corresponds to about a 10-fold difference in viral genetic info. So what may seem like small Ct value differences actually amount to big differences in amount of viral RNA (aka “viral load”).
RT-PCR can detect patients with really high Ct values (corresponding to really low levels of the viral genetic information). These patients almost certainly not contagious, yet they’re given the same result: “positive” as the person before them who had orders of magnitude more times virus. Yet both patients will (hopefully) be told to isolate themselves and will be followed up with by contact tracers who will then look for the people they’ve been with recently (even though “recently” may have been way too long after the person was actually contagious). In that case, you have some wasted resource usage, but it’s a “better safe than sorry” scenario that isn’t the really big issue.
The really big issue is time. People are waiting days and days for test results, by which time it’s too late. By the time they get their result, the person, even if they were infected, is no longer contagious, and anyone they were in contact with during the actual contagious period – and potentially infected – could have gone on to infect more people. This is especially important for screening tests, where people are asymptomatic so they have no reason to think they need to self-isolate while they wait for their results.
And on the other side of things, in part because tests take so long, some places like camps or even potentially airlines are requiring that people show evidence of a negative test result obtained within the past 12 days. But who knows what the heck that person’s been doing in those 12 days post-testing! And at the time of testing someone might be in that window of time post-exposure but pre-infectiousness when the virus is catching hold and replicating enough to shed enough to be detected. They could test positive the next day, or even the same day because the ramp up (and down) is really fast.
But it’s highly unlikely that the person will get tested the next day (and thus caught). It’s highly unlikely that the person will even have their test results back by the next day!
So, clearly, we need fast, frequent testing. To address the testing backlog problem, instead of pouring more resources into RT-PCR tests which, even with SalivaDirect still take time and fancy equipment & labs, Mina and other scientists are advocating for cheap, truly rapid tests. Tests that aren’t nearly as sensitive as RT-PCR but which are still good enough to detect people who are actually contagious. And which are cheap and easy enough (think a paper strip you can spit on) that they can be used frequently enough (at least 2X a week) to catch contagious people and prevent them from infecting others.
Such tests are in development, many of which are “antigen tests” – more on these types of tests here: https://bit.ly/antigentests but the basic idea is that they use antibodies on a strip, etc. to detect viral *proteins* present in a sample from a patient instead of viral genetic information. There’s no amplification step with these tests, so they’re less sensitive than RT-PCR (you need larger amounts of starting viral material) – BUT they can detect people who have the level of virus needed to actually be contagious! And they can be done in a paper strip type tests (lateral flow assay) (think pregnancy test style).
The antigen tests that are currently approved (such as the Quidel one I talk about in that post) detect proteins that are inside the viral particle, so the virus has to be broken open first and more time & equipment is needed – especially since they have a fluorescence readout. What we really need are tests with no fancy ingredients and with a colored readout. Quidel uses fluorescence because you get a strong signal from it, stronger than you can get with just a colored dye. So that bumps up their sensitivity, but makes it more expensive and less user-friendly. But if we accept lower sensitivity, then color can work – and a company called e25 bio has such a test waiting for approval. It detects the Spike protein, which is on the outside of the virus, so you don’t have to break it open. And it has a colored readout. But it doesn’t have FDA approval.
And a similar situation is faced by other companies working on similar tests – they’ve all had a hard time getting FDA approval because they’re being judged in comparison to RT-PCR. Mina thinks there should be a separate category in the approval process for something like “contagiousness tests” where the sensitivity is good enough to identify infectious people but won’t detect people like that super-high-Ct count person who’s only shedding harmless viral RNA.
Early on in the pandemic, the FDA was super super strict about authorizing tests – only the official CDC tests could be used. And those tests were faulty… https://bit.ly/cdctestproblems
The FDA eventually approved more tests made by multiple companies, but they might have “overcorrected” when it came to antigen tests. They allowed companies to sell pretty bad tests (way worse than the pretty good but not perfect tests Mina is advocating for) prior to getting FDA approval (basically you can sell them til we have a chance to check them and if they’re bad we’ll withdraw them). The FDA realized their problem and course-corrected, but back to a potentially “too strict” mode – at least for “contagiousness tests” – it’s still important to have really accurate gold-standard-level diagnostic tests for places like hospitals, clinics, and high-risk environments.
I think when we talk about lowering the sensitivity requirements for “contagiousness tests”, it’s important that we reach the right balance. We don’t want the wild wild west all over again. Tests much be required to prove what they say they’re designed to do – they need to be able to detect contagious people. They need to be regulated and they need to clearly convey to consumers what they are designed to do.
There are of course some other potential problems with the mass testing strategy, as are outlined nicely in this article from the Atlantic by Robinson Meyer and Alexis C. Madrigal https://bit.ly/31kRa6N
Firstly, even if you get approval, you’re going to need to really mass produce these things, which is going to take a lot of $ and resources, potentially even use of the Defense Production Act as the Rockefeller Foundation’s COVID-19 Testing Solutions Group calls for. https://bit.ly/2FIyCoF
Some other important questions: What happens if a person tests positive? Do they then go get a RT-PCR test to confirm? Will they self-report and get contact-traced or will governments lose any idea (other than hospitalizations) about viral prevalence?
If a person who gets a false positive does so because they have something cross-reactive in them, will they *always* get a false positive? How will they be tested – will they have to get special permission from their school, work, etc. to use a different test?
Who’s footing the bill? Will everyone have access, or just the wealthy? Or will the wealthy poo-poo these tests and insist on conventional tests, maybe even purchasing their own little machines to run some of the more sophisticated “rapid tests” like the Abbot ID-NOW?
At this point, it’s all just speculation because this rapid, cheap, testing isn’t available. But it’s not just Mina raising attention. Here are a couple more good articles:
“‘Instant Coffee’ COVID-19 Tests Could Be the Answer to Reopening the U.S.”, Carolyn Barber, Scientific American: https://bit.ly/32f2Yaa
“Even Covid-19 tests that aren’t perfect can help control the pandemic,” JEFFREY L. SCHNIPPER and PAUL E. SAX, STAT: https://bit.ly/2YmmNef
And a website with more resources, etc. https://www.rapidtests.org/
Sorry this post kinda turned into a story in 2 parts…
more coronavirus-related posts: https://bit.ly/covid19bbresources
more on topics mentioned (& others) #365DaysOfScience All (with topics listed) 👉 http://bit.ly/2OllAB0⠀