Testing for COVID-19 antibodies
07/16/2020 - 15:21
In this Q&A, Rahul Roy, Associate Professor at the Department of Chemical Engineering, IISc, discusses how antibody tests work and what they are useful for in the context of the COVID-19 pandemic.
What does an antibody test look for?
If a person gets infected with a pathogen or a foreign body, their adaptive immune system recognizes foreign antigens ‒ which can be a part of the pathogen’s protein, RNA or DNA ‒ based on immune cells that go and bind to them, and release soluble proteins called antibodies. In COVID-19, for example, we know that this usually happens after about 10-25 days (some antibodies can be detected as early as five days from the onset of symptoms). These antibodies are what the test will detect.
Our immune response, and therefore the level of antibodies, goes down over time. However, even if the pathogen is no longer there, its “memory” still remains in our body. The kind of immunity one gains depends on how robust our immune system’s memory cells are and how many we have. This will also depend on the type of antigen and how strongly it invokes the adaptive immune response.
Of course, there will be differences in the level of antibodies based on age and other factors. Older people will have weaker adaptive immune responses. If you test too early or too late, the levels will be low. But when you collect and test samples from a large number of people, antibody levels are a good indicator of the extent of infection in a population.
How are antibody tests useful in the context of COVID-19?
In the case of COVID-19, a large number of people are asymptomatic, so the only way to get an idea of whether they were ever infected or not is via an antibody test. If we find antibodies against the novel coronavirus in a person’s blood sample, it means that they have, at some point of time, been infected with it. The virus may not be detectable at this stage, especially if they have recovered completely. But we will still be able to detect the antibodies that the body produced against the virus when they were infected. Our cells can produce these antibodies for as long as a year after infection, in high enough quantities that we can detect them.
Detecting antibodies can be useful to carry out what is called a serology survey, to find out how many people in a population could have gotten infected at any stage. The Indian Council of Medical Research (ICMR) is trying to do some serological surveys. Insights from such surveys are important for epidemiological reasons.
First, we would like to know if most of the people have already been infected so that we need not worry as much, because then enough people in the population may have developed immunity. Of course, the test is not necessarily proof that one has immunity; just that they were at some point infected with the novel coronavirus. Second, it could help authorities decide where to prioritize vaccination, and focus on high-risk areas with high rates of infection.
How do antibody tests compare to other types?
Unlike the RT-PCR tests [which look for the virus’s RNA], antibody tests can easily be adapted to a rapid diagnostic kit which takes about 10-15 minutes to process. They are easy to interpret: a single line can tell whether the test is positive or negative, similar to a pregnancy test. Hypothetically, such tests can be shipped to an individual’s house and they can test themselves using a prick of blood. They are also cheaper; it would be easy to make them in India for less than Rs. 100.
However, there are some caveats. Interpretation of antibody tests becomes harder and harder with a larger population. They are also not foolproof. There are two parameters we use to assess their accuracy. One is sensitivity. If a person is infected with COVID-19 and produces antibodies, how often will the test detect them correctly? If it recognizes them 100 percent of the time, then the sensitivity is 100 percent.
The other parameter is specificity. We have a lot of similar pathogens and other coronaviruses in the environment. How well will the test recognize antigens that belong to only the COVID-19 virus and not others? The goal, therefore, is to look for antigens that are dramatically different between two viruses. For example, the Receptor Binding Domain (RBD) of the spike protein is about 70 percent similar between the 2003 SARS virus and SARS-CoV-2, whereas the nuclear protein has 90 percent similarity. There is, therefore, a lower chance that antibodies against the spike protein will be mistakenly recognized.
Unfortunately, we do not know exactly which proteins of the virus the body produces antibodies against. This is a unique case because the infection is appearing first in the respiratory tract, not in the blood. The immune cells in the blood will come across the antigens only after they have been degraded by other cells. So we cannot know for sure which viral proteins are the ones that are recognized, but we can guess to an extent.
What kind of antibody tests are being developed or used in India?
ICMR has come up with an antibody test which is based on the whole virus that has been neutralized and can be used to detect antibodies in the blood sample. However, developing tests based on the whole virus is expensive and scaling up is a challenge. Other people are using recombinant proteins.
Normally, we rely on kits developed outside the country. But this is a unique situation where the burden of the disease is equally high in the outside world. It is, therefore, hard to get high quality tests into the country quickly.
Can you tell us about the kit that you are currently developing at IISc?
If we use full proteins, there are very few companies which can scale up mammalian cell cultures to make such proteins in large amounts, and they would not want to do this for a small kit which costs less than Rs. 100. We wanted to avoid this trap of scalability. Therefore, we decided to work on peptides, which are small segments of proteins that will also be seen by the immune system.
We are working on linear peptides that can be chemically synthesized, which makes quality control very easy. We design these peptides based on computational models. Peptides are much more stable than proteins; the latter can sometimes unfold and cannot recognize the antibody. We can easily order them off peptide synthesizers which are commercially available in the country as well as outside. We do not have to worry about batch-to-batch variation in quality.
Proteins can bind to many different antibodies, and so their response will always be much larger. How do we resolve this issue for peptides? We did this by finding several linear peptides across the virus proteome that will evoke immune response and combining them together. We are currently working with physicians at the Bangalore Medical College and Research Institute and Victoria Hospital to access blood samples and validate our tests.
Ranjini Raghunath is Communications Officer at the Office of Communications, Indian Institute of Science (IISc)