At the six-month check-in, the patients in this study were all PCR-negative for the virus – they were no longer infected. Still, 38 of them reported some persistent long-term effects from it (you really, really do not want to catch this if you can avoid it – for more on long-term effects, see this new paper). The more severe the acute infection was, the more likely people were to experience long-term problems. Over a six-month period after recovery from SARS-CoV2, the patients in this work did indeed show lowered antibody levels. The drop in antibody titers was not evenly distributed over the different types (antibodies against the Spike receptor-binding domain (RBD) or the nucleoprotein (N), and IgM for these versus IgA), but overall neutralizing antibody activity was down about five-fold at the six month point as compared to the after-one-month check. But that’s what happens with any infection: the immune system does not crank out high levels of specific antibodies forever. It settles back down to its watch-and-wait mode, and that’s where the memory B cells come in.
And looking at those B cells showed some interesting patterns. There were a variety of them at both the 1-month and 6-month points, but they changed over time. Some of the clonal lines that were present earlier had disappeared, while new ones had continued to show up. The distribution was different as well: at the earlier point, the most common B cell clones were a greater percentage of the whole than at the 6-month point, for example. The authors say: “We conclude that while the magnitude of the RBD-specific memory B cell compartment is conserved between 1.3 and 6.2 months after SARS-CoV-2 infection, there is significant clonal turnover and antibody sequence evolution, consistent with prolonged germinal center reactions.“
What’s that antibody evolution look like, then? The good news is that the ones from the six-month check showed both increased potency and an increased range of responses against various protein mutations. That includes many of the ones that are in the news these days, things like R346S, Q493R, and E484K. (As an aside, did anyone ever imagine that amino acid variant notation would creep into major news stories? Strange days). But while the one-month antibody samples were unable to recognize these and bind to them, the six-month ones were.
How does the immune system do this? It comes down to follicular dendritic cells (FDCs), a specialized cohort that displays antigens on their cell surface for an extended time in the “germinal centers” where B cells replicate. You’ll find them in lymph nodes, spleen, bone marrow, and other tissues (here’s a review article on the subject). They’re quite odd, with unusual cellular structures (very few visible organelles in their cytoplasm and an overall large, spread-out net-like structure). They secrete chemokine signals that attract B cells, allowing them to be exposed repeatedly over the long term to the antigens that they present on the FDC surfaces (which are technically in the form of “immune complexes” there).
In severe autoimmune disease, it appears that this process can go wrong and lead to FDCs and B-cells assembling in other parts of the body entirely as part of a sustained (and completely inappropriate) immune response. But when things are working as they’re supposed to, the reservoir of antigen in the FDCs and its continued presentation to the B cells allows for the generation and selection of new and improved antibodies, as seen in this study. We got this system via evolution, too, of course: the viruses that have been attacking us and our phylogenetic ancestors for hundreds of millions of years have always been mutating in real time, so having a system that keeps pace with them after an initial infection is a clear survival advantage. As always, the immune system inspires awe and a bit of terror as well.
What does this mean for people who are vaccinated, instead of getting their immune response by being really infected by the virus? As that last link mentions, “an efficient vaccine should maximize the deposition of immune complexes on FDCs“. You would expect the Spike protein produced in the body via the mRNA or viral vector vaccines to be handled in just this way, as well as the directly-introduced antigen proteins in a vaccine like Novavax is developing. The new paper being discussed also looks into the possibility of persistence of viral antigens in intestinal tissue, which was seen in some (but not all) of the patients studied, and would be expected to drive the evolution of IgA-type antibodies in particular. But overall, the key would be to make sure that the FDCs are displaying the Spike antigen the way that they should be.
How about the variations on the coronavirus that are out there right now? This has a bearing on some very real-time data from Ravi Gupta’s lab at Cambridge. There’s a manuscript that’s on the way to MedrXiv, but this pre-preprint stuff can be found on Twitter here and here. They’re looking at antibodies from the blood of people who were vaccinated three weeks ago, and seeing how these perform against the B.1.1.7 variant. In short, they still have neutralizing activity, but it’s generally lower (at least, in ten of the fifteen patients studied). Now, this doesn’t mean that the vaccine is ineffective in those people – the antibodies still do their job. But it does mean that as mutations continue to pile up, that escape of such a new variant of the coronavirus is not impossible.
But. . .note that Gupta’s lab is (necessarily!) looking at the antibody profile of people who have been recently vaccinated. The paper I’m discussing today raises the strong possibility of continued B-cell and antibody evolution over a period of months, leading to a different set of antibodies that appears to be able to better deal with some of these mutations. It will be very interesting indeed to combine these two studies – clonal B cell changes and antibody evolution with activity against variants like B.1.1.7 – to see if this is indeed the case. It should be, but we’ll want to check!
