A thread as promised on this biorxiv.org/content/10.110… from George Kassiotis’ lab including @kevinWng and others @thecrick, the result of incredible teamwork with @uclh.
This is a paper born of adversity, and contains two really important and unexpected findings. When the pandemic hit London hard – very hard – we had minimal diagnostic capacity. Our main focus was the critically important qPCR pipeline for diagnosing active infection,
3/n but George’s lab and with @RealMcCoyLab and @eleni_nastouli and colleagues at UCLH took on the task of developing diagnostic serology. One very annoying aspect of this is that there is some cross-reactivity between previous seasonal coronavirus infection and SARS-CoV-2.
4/n That means a few percent of people will appear to have been infected with SARS-CoV-2 when they hadn’t. A clever way around this is to look not just at IgG responses, but IgA and IgM simultaneously. Being immunologists, George’s lab went for flow cytometry.
5/n You express Spike (S1S2) in human cells, and look for binding. This works very well: you get some cross-reactive IgG sticking to the cells, but the SARS-CoV-2 samples have IgA & IgM too. It’s exquisitely sensitive & specific, used to great effect in the @SAFERuclh study.
6/n But what is the cross-reactive IgG doing, besides being annoying? In the first version of the pre-print we were able to show some neutralisation of viral pseudoparticles, but we didn’t have a microneutralisation assay for the real virus ready.
7/n In all honesty, we weren’t that hopeful there would be authentic neutralisation, these antibodies aren’t directed at the receptor binding domain. Rachel Ulferts in my lab and the World Influenza Centre (who have huge experience with these assays)
8/n got a very nice one up and running and reproducible, of which more anon. It was a surprise to us that at least some seasonal coronavirus sera neutralised the authentic virus, with pretty reasonable titres too. But there was another huge surprise...
9/n You might expect that children would have more cross-reactive antibodies – but up to 60% was a kind of bombshell. Way above what any of us expected. So what does this all mean? I will start by stating what it emphatically *does not mean*.
10/n It *does not mean* that 60% of children are already immune. It might be *one* reason why children generally get very mild disease, but for all we know most of these antibodies might be useless in vivo, or possibly even pathogenic in some rare cases.
11/n It quite possibly means that a modest proportion of children & very small proportion of adults have a useful degree of pre-existing immunity. What it definitely implies is that we should study these pre-existing antibody responses in detail to work out what they really do.
12/n Huge contributions from many co-authors, the ones I know are on twitter include @LabGandhi @SmallRedOne @ViralHoolahoop @CharlesSwanton and apologies to anyone I’ve missed. Thanks to @AdamJKucharski @stuartjdneil and @BallouxFrancois for publicising this already.
Finally you should follow first author @KevinWNg and look at his great thread here:
