In recent weeks there’s been a lot of discussion in the media of antibody tests relating to SARS-CoV-2, the virus that causes COVID-19.
In the early stages of the outbreak the main focus of testing was to determine whether a subject having characteristic symptoms of COVID-19 had an active infection with SARS-CoV-2, using tests that detect SARS-CoV-2 nucleic acid, typically in samples obtained from the upper respiratory tract of the subject.
However, over time – and especially in countries where the actual number of subjects that have been infected with SARS-CoV-2 is thought to be many times greater than the number of confirmed cases – the focus for testing has shifted to antibody tests.
Unlike the nucleic acid tests for active SARS-CoV-2 infection, the aim of antibody tests is to determine whether someone has previously been infected with SARS-CoV-2. When a subject is infected with SARS-CoV-2, fragments of viral proteins encounter B cells expressing receptor molecules which bind specifically to the fragments. These cells are ultimately stimulated to proliferate and differentiate into e.g. plasma B cells, which produce large quantities of the receptor molecules in soluble form; these are known as antibodies. Antibody tests for SARS-CoV-2 generally involve detecting the presence in blood samples of antibodies capable of binding to a given SARS-CoV-2 protein, such as the spike protein (which the virus uses to gain entry into host cells), e.g. using reporter agents which produce a detectable signal.
If a blood sample from a subject contains antibodies specific for a SARS-CoV-2 protein, then you can be confident that the subject has mounted an immune response to SARS-CoV-2, and has therefore previously been infected with the virus.
This is a very important population to know about in the broader context of monitoring and managing the outbreak at a community level. Subjects having previously mounted an immune response to the virus are likely to be resistant to re-infection. In addition to being highly unlikely to develop COVID-19 for a second time (at least in the short to medium term), they are also therefore unlikely to be able to infect individuals that have not yet encountered the virus. It has been suggested that it could be safe for such subjects to re-enter the community. Information from antibody tests could also be useful for informing future strategy when a vaccine for SARS-CoV-2 becomes available, with prioritised vaccination of people known not to have previously been infected. This may reduce the time it takes to reach the threshold level of immunity in the community to effectively prevent further SARS-CoV-2 infections.
However, so far it is proving much more difficult to develop reliable tests for antibodies to SARS-CoV-2 than it has been to develop nucleic acid tests for the virus itself. The biology of antibody immune responses presents several challenges.
Infection elicits the production of different antibodies in different people, and so the antibodies to SARS-CoV-2 produced by one person’s immune system can be quite different to those produced by another’s. Even within a given individual, primary antibody responses to initial challenge with a given infectious agent – which is what the antibody tests for prior SARS-CoV-2 infection seek to identify – are typically characterised by the production of a diversity of antibodies. Unlike nucleic acid tests which seek to detect one or a small number of very similar molecules of known structure, a reliable antibody test must therefore be able to detect a diversity of antibodies of unknown structure, but having the shared property of binding to the SARS-CoV-2 protein(s) the test uses. Moreover, the antibodies produced during primary immune responses generally bind to their targets with low affinity, and so individual antibody:target binding events will typically be weak. Further still, the extent of antibody responses will differ between subjects, and even for a given individual the levels of antibodies to SARS-CoV-2 in their blood will change over time.
So, an ideal antibody test for prior infection with SARS-CoV-2 will be able to detect a broad diversity of antibodies specific for SARS-CoV-2, which will in many cases display only weak binding to their target, and which will be present at different levels in subject samples, and from the signal detected arrive at a binary determination of whether a subject has or has not previously been infected with SARS-CoV-2, with high confidence.
This is a very tall order, and it is easy to understand why many of the existing tests have encountered issues with sensitivity, such that samples from subjects known have had COVID-19 and recovered are not necessarily positively identified, and also with specificity, where samples from subjects known not to have been infected with SARS-CoV-2 have been determined to contain antibodies to the virus. There’s typically a trade-off between sensitivity and specificity. Highly-sensitive assays which are able to detect a broad range of antibodies to SARS-CoV-2 protein(s) even when present at low levels in a sample will likely suffer from reduced specificity, with antibodies which are not specific for SARS-CoV-2 protein(s) also producing a positive signal. Conversely, highly-specific tests are unlikely to be able to detect the full diversity of antibodies to SARS-CoV-2 protein(s), and at the various different concentrations they might be present at in subject samples.
However, we have reason to be optimistic. We are seeing a truly unprecedented level of innovation and collaboration within the scientific community, with the world’s best minds united behind the common goal of developing a suite of tools for combatting the COVID-19 pandemic. Reliable antibody tests could be key for our understanding and management of the outbreak, and a major step towards life returning to some kind of normality.
