“Coronavirus tests:” the tests that people actually want to take! But what test is best? There isn’t a single answer! Traditional “coronavirus tests” use a molecular biology technique called RT-PCR to make lots of copies of genetic material that’s specific to SARS-Cov-2 (the “novel coronavirus” that causes the disease called Covid-19) and use fluorescent probes to detect if copies get made, indicating the presence of the virus. Serological (blood-based) antibody tests use a technique called ELISA to look for little proteins called antibodies that your immune system makes to specifically target the virus.
Since the PCR test is looking for the virus itself, it is limited to detecting *current* infections, but it can detect them really early on in the infection. Antibody tests, on the other hand, look for evidence of the body’s virus-specific response to the infection. They can NOT detect the virus in the early stages, before this response has started to gain some ground on the virus, but they CAN detect the virus in the later stages (probably ~3-4 days after symptom onset) with a really quick, easy test.
And, potentially more importantly, they can tell if someone has had the virus in the past and recovered, because some of antibodies the body uses to fight the infection stick around after the fact to keep watch. Developing antibodies is part of the adaptive immune response, and it’s a kind of trial and error process where your body experiments with mixing and matching constant and variable parts of antibody proteins to find ones that bind parts of the virus but don’t bind parts of us (antibodies often target the spike protein that juts out of the viral membrane and allows it to dock onto our cells). Your body then uses these antibodies to call in helpers to destroy the virus whenever an antibody bumps into and sticks to that virus.
While a patient is infected, lots of these antibodies are made and they can work on actively fighting the virus. But once the infection’s over, those antibodies don’t just kick back with a beer. Instead, some stick around to keep watch after the fact. This way, if the virus tries to infect them again, their body doesn’t have to go through the “development” stage of creating an effective antibody. This likely provides immunity against re-infection, which could mean that recovered patients could safely interact with infected people – so, importantly, immune healthcare workers could treat patients with less risk to themselves. But, an important caveat is that the commercial tests are likely to be yes/no tests about whether or not antibodies are present – nothing about how much of the antibodies are there and therefore how “ready to attack” the immune system is. So as my immunologist e-friend Dr. Samantha Le Sommer (thanks again!) puts it: “it’s a huge assumption with these tests that exposure = immunity.” There are ways to measure antibody *levels* which I’ll get into later.
Additionally, knowing who has been infected in the past will add a very valuable tool to the epidemiologist’s toolbox. For example, early on in the pandemic, when there were few cases, it was usually “easy” to do “case tracing” – working backwards from an individual known to have the disease to figure out where they got it from (the chain of transmission) – and working forwards to identify people they were in contact with in order to notify and quarantine those people and monitor them to see if they developed the disease. But, since some people can transmit the virus without having symptoms themselves (asymptomatic carriers), there were cases where doctors couldn’t find the direct line of transmission, but antibody tests could solve the puzzle, as was first shown by a team in Singapore: https://bit.ly/2URzxqO
This strategy still can be used in regions not yet severely affected, to cut transmission off at the source. Even in regions with widespread “community transmission” where there’s no real hope of figuring out the exact spread pattern, finding out who has been infected (even unknowingly) can help doctors & scientists figure out more about how the disease spreads and why some people’s bodies are unfazed by it (which could offer potential hints towards treatment strategies).
Currently we don’t have enough tests to just go around testing everyone, so we don’t know the true extent of the disease. This is a large part of the reason why you’ve probably seen a lot of differing statistics when it comes to Covid-19: wide variations in reported death rates, infection numbers, etc. The US in particular has been slow to roll out testing, early tests were faulty, etc. – it was a whole big mess. As a result, even now that we have tests being manufactured and shipped out around the country, we don’t have nearly enough to meet the demand. So, often, only the most severe cases and/or people with known exposures or who are at the front lines are able to get tested, with others being told to self-isolate as if they have the virus without really knowing if they do.
This must be really anxiety-producing for the patient – and it also is anxiety-producing for the public health officials trying to figure out where and how widespread the virus is, and how to combat it. Since almost all of these patients recover just fine, test or no test, there are a lot more cases going undetected – but it isn’t “too late” to detect them – we just need a different test.
The current tests being used are the traditional, RT-PCR ones. You can learn a lot more about them here: https://bit.ly/covid19bbresources
They take a while to perform because they involve actually isolating the viral RNA (RNA extraction), which is a “delicate” procedure that has to be done in a lab. Recently, a new, rapid test has been rolled out by Abbott, called ID NOW. It’s an example of what’s called a “point of care” test, meaning that it can be done in your doctor’s office instead of having to be sent to a lab.
It’s similar to the traditional PCR based test in that it looks for pieces of the virus’ genome, but it copies those pieces using a different technique called isothermal amplification. It doesn’t require extracting the RNA, instead it just breaks the cells open within the machine and does it all there – really quickly (10-15 min). Similar tests (using the same machine in fact) are used for detecting other viruses, like flu and strep, but we don’t yet know how sensitive and accurate the Covid-19 version is compared to the PCR test.
Both of those are looking for the virus’ genetic information (which is only there if the virus is in you, but is there even in the early stage of infection) and they typically use samples collected by swabbing your nose or throat. Antibody tests, however, look for antibodies you have produced against the virus, which only show up later in the infection (a team of scientists at Mount Sinai hospital in New York developed a test that they say can detect antibodies as early as 3 days post symptom onset). https://wb.md/2JwBDac
And these tests look for these antibodies in blood samples, so a quick note about all those bloody blood terms! If you take whole blood and centrifuge it (spin it really fast) in a tube coated with an anti-coagulant to prevent clotting, the blood cells will sink to the bottom, and the rest of the stuff that’s in the blood (the blood plasma) will remain on top. It’s mostly water, but it also contains things like secreted antibodies and proteins, including proteins called “clotting factors” that allow blood to clot. If you take whole blood and, before you centrifuge it you allow it to clot, the liquid that remains after you spin it is called serum. It’s like the plasma but without those clotting factors.
Sera or plasma can be used for ELISAs – either way, when antibodies are detectable in a patient’s blood, it’s said that a person has “seroconverted” – so, for example, according to the Mt. Sinai paper, patients “seroconvert” around 3 days after they first notice symptoms.
How do you know if a patient has “seroconverted”? Typically with a serological test called an ELISA (enzyme-linked immunosorbent assay), which traps the virus-specific antibodies on a plate and then gives a signal so we know it’s there.
The tests are typically carried out in wells on a shallow plate. These wells are coated with viral protein to act as an antigen (thing that an antibody binds to). Some of the antigens that have been/are being tested are parts of the nucleocapsid (N) protein, which coats & protects the viral RNA (and is the most abundant viral protein), and the spike (S) protein which is the one that juts out of the viral membrane and binds to receptors on host cells. These can be made recombinantly – scientists can stick the individual genes for making them into expression cells to make them in a lab without having to deal with the actual virus.
When developing the test, the scientists have to be sure to find antigens that are unique to the virus – so, for example, you want to use something from SARS-Cov-2 (the novel coronavirus) that is distinct enough from the original SARS-Cov virus that if a patient has antibodies against SARS-Cov, they won’t get a false positive due to “cross-reaction.” But scientists also have to make sure that the antigen they use is able to detect antibodies that the patient has (the test needs to be sensitive and accurate).
So the development phase can take some time to optimize, but the whole testing process can be automated and isn’t too long.
The first step is adding the sample – the plasma or serum which can be heated up first to kill any live virus, making it safer for the testers. If the sample contains antibodies that can specifically bind to that viral protein, they’ll get stuck. If not, they’ll get washed off in the next step, where you wash the wells several times to remove anything that’s not stuck.
But at this point, you can’t tell if anything *is* stuck – you need some way to detect that the “primary antibody” is bound. You can do this by adding a labeled “secondary antibody” that recognizes the generic adaptor part of the primary antibody (the constant region). For example, the Mt. Sinai group used anti-human IgG as the secondary antibody (there are a few different adapter parts that can be used to give you different isotypes of antibodies and IgG is one of these (Ig stands for immunoglobulin, which is another name for antibody).
Sometimes the secondary antibody labels are directly seeable, but other times you have to add something else to make a seeable product. For example, oftentimes, secondary antibodies are tagged with HRP (horse radish peroxidase). On its own, you still can’t see it, but when, after some more washes, you add another chemical called OPD (o-phenylenediamine dihydrochloride) the HRP converts the OPD into a yellow-orange colored product (2,3-diaminophenazine) whose “coloredness” can be measured using UV-Vis spectroscopy.
Back to that “how much antibody is there” question – to find this out, the patient sample is serially diluted (e.g. 1:10, 1:100, 1:1000) so you can see how low much you can dilute it and still get a positive test response (signal above background) and take the inverse of that to determine the antibody titer – how much antibody the sample contains. This can be really useful because if some now-healthy person has a high level of antibodies against SARS-Cov-2, they might be able to donate blood plasma to help patients with the disease. Such “convalescent plasma” is being used as an experimental treatment.
more Covid-19 resources (some in multiple languages) here: https://bit.ly/covid19bbresources
more on other topics mentioned (& others) #365DaysOfScience All (with topics listed) 👉 http://bit.ly/2OllAB0