Skip to main content

Q&A: B cells and antibodies in COVID-19: what does 'good' look like?

On Thursday 13 August, our 'Connecting on Coronavirus' webinar was presented by BSI Trustee Professor Deborah Dunn-Walters on the topic of 'B cells and antibodies – what does 'good' look like?' If you missed the live webinar, you can catch up and watch again here.

Professor Deborah Dunn-Walters is Professor of Immunology, Head of Immunology Section and Lead for Lifelong Health Research Theme at the University of Surrey. We hear a lot about the production of antibodies against SARS-CoV-2, but not all B cells go on to produce antibodies, and of those that do, some produce a different type of antibodies than others.  Making good antibodies without making potentially harmful ones is the goal of vaccines, but do we know how to tell the difference easily? In this webinar, Professor Dunn-Walters provided us with a summary of research to date and posed some key questions for areas to research in the future. 

Unfortunately, we lost the internet connection during the Q&A at the end of the webinar.  However, the audience had sent a substantial number of questions in and Professor Dunn-Walters very kindly took the time answer some of these for this blog.

Given what we currently know about the production of antibodies and COVID-19, what implications does this have for understanding the results of antibody tests against the disease? If I have antibodies, does that mean I’m immune to future infection? 

Graphic from BSI resources on COVID-19 testing

Deborah Dunn-Walters: There is a lot of variability in antibody response between people and in the results depending on which method is used to measure the antibody. Antibodies may not be seen early in the disease in SARS-CoV-2 positive patients but are likely to be seen by 20 days after initial infection. IgM and IgA antibodies would not be detected a couple of months after recovery, but IgG would be there for longer.  You are more likely to develop antibodies that can be picked up by a standard antibody test if you had a more severe disease than someone who was asymptomatic or only had mild disease.

A positive test using a method that is 100% specific, means that you have had the disease.  If a test were, for example, 99.7% specific then a positive result would mean that you are highly likely to have had the disease, but 3 in every thousand results might be a false positive. 

A negative test could mean that either you haven’t had the virus at all, or that you did have the virus but it was such a mild/asymptomatic infection you didn’t make antibodies or that any antibodies made are in such low quantity that an antibody test is not sensitive enough to detect them. A recent study of 341 volunteers with a previous positive PCR test showed that even after correcting for the lack of sensitivity of the test there were 4% negative results.1

Either way, because we do not yet have an answer as to whether the presence of antibodies indicates immune protection against infection we can only use antibody tests to estimate disease prevalence, not to assess our level of immunity.

Do you think that any of the antibodies produced in response to COVID-19 might cross-react with antibodies produced for the other coronaviruses that infect humans (and which cause common colds)?

DDW: Yes. The systems serology paper by Butler et al. presented in the webinar shows that there are antibodies produced in COVID-19 patients, in both serum and nasal washes, that react with other common coronaviruses.2 

Do you have any further thoughts about the use of convalescent serum containing high levels of antibodies for severe COVID-19 patients. Is there any evidence that it is helping recovery?

DDW: If my hypothesis is correct, that later antibodies are more matured than the antibodies made early in the disease, then it follows that antibodies from convalescent patients will be useful.  For example, they may be of higher affinity and have different levels of sugars on them. There has been no evidence of harm from convalescent serum and randomised controlled trials are ongoing. Indications are that transfusion at an early stage with convalescent plasma containing high levels of IgG would be of use.3

What determines low or high antibody titers in COVID-19? For example, the amount of virus, or how many B-cells recognise the antigen, or both?

DDW: Firstly, I am very keen to stress the distinction between correlation of two factors versus “determination” or “causation” between one and the other. The observations we make about COVID-19 now are correlations.  Plus, I don’t yet know of any quantitative studies of antigen-specific B cells during the disease, so this is difficult to answer. In cases with high virus titres and severe disease you will see high levels of antibody.

Some reports suggest that there is T cell depletion in COVID-19. Will this affect T-dependent B cell response in COVID-19 and subsequent antibody formation?

DDW: I certainly think this is an issue. Helper T cells are needed to help both Killer T cells and B cells.  A B cell in the germinal centre needs T cell help to survive when it is affinity maturing its antibody.  I should also mention that B cells are excellent antigen presenting cells to help activate T cells.  You need both for the best responses.

Do you think that the generation of appropriate memory B cells is more important than having long-lived antibodies in building immunity to COVID-19?

DDW: I would like to have both. Having long-lived neutralising antibody circulating in the blood (or present in the nose) is good because the antibody can block infection in the first place.  Having memory B cells is also good because the cells can reactivate quickly in response to challenge and replenish the long-lived antibody producing cells.  It may be possible to improve the antibody even further if a memory B cell enters a new germinal centre reaction the second time round.  This might be an opportunity to slightly adjust the antibody in case the virus had changed a little in the meantime.

What implications does the half-life of COVID-19 antibodies have for vaccine development and for the type of approach that might be successful?

DDW: My personal hope is that the response to vaccines would be different than the response to disease. If in the disease the B and T cells decrease and this hinders the type of response that makes high affinity antibodies, perhaps in a vaccine the response is more controlled – more T-dependent response and less extrafollicular response. So better longer-lived antibodies and memory B cells will be made from the beginning.  This is my conjecture and hope not fact.

We’d like to thank Professor Deborah Dunn-Walters for her time and expertise in both presenting and answering these questions. You can watch the full webinar here


  1. Ward et al. 2020. Antibody prevalence for SARS-CoV-2 in England following first peak of the pandemic: REACT2 study in 100,000 adults.
  2. Butler et al. 2020. Features and functions of systemic and mucosal humoral immunity among SARS-CoV-2 convalescent individuals.
  3. Joyner et al. 2020. Effect of convalescent plasma on mortality among hospitalized patients with COVID-19: initial three-month experience.