We mapped all #SARSCoV2 mutations escaping key antibodies used to treat #COVID19 (biorxiv.org/content/10.110…). Surprising observations: one amino-acid mutation escapes *both* antibodies in @Regeneron cocktail & escape mutations selected in infected patient treated w cocktail. (1/8)
Specifically, @tylernstarr, Allie Greaney, Amin Addetia, and @AdamDingens used a deep mutational scanning system to determine all mutations that escape antibodies in REGN-COV2 and LY-CoV016. You can view these complete escape maps here: jbloomlab.github.io/SARS-CoV-2-RBD… (2/8)
Surprisingly, they found a single amino-acid mutation (E406W) can escapes both antibodies in REGN-COV2 cocktail. E406W isn't in structural footprint of either antibody (see image), so mechanism is unclear. But it reduces cocktail neutralization by 100-fold. (3/8) 
Next, Will Hannon worked with @DrJLi @Dr_MChoudhary to examine viral evolution in a persistently infected patient treated with the Regeneron cocktail at day 145. Patient previously described in their paper here: nejm.org/doi/full/10.10… (4/8)
Analyzing deep sequencing data, found that antibody treatment was followed by changes in frequency of 5 RBD mutations, 4 of which we mapped to escape antibodies. See evolutionary dynamics below: (5/8)
Mutational dynamics show hitchhiking & competition among antibody-escape lineages, similar to persistent flu (elifesciences.org/articles/26875). We don't think comparable escape likely in typical shorter infections, but shows virus can escape via mutations we map given time. (6/8)
Finally, we used maps to see which escape mutations present in circulating #SARSCoV2. Prominent ones include Y453F (associated with mink outbreaks in Denmark / Netherlands), N439K, & others (see below). But fortunately, all low frequency right now (7/8).
Going forward, these prospective escape maps should help with assessing which viral mutations will escape important therapeutic antibodies. All the data available at links above in this thread, so please use them to help with viral surveillance! (8/8)
• • •
Missing some Tweet in this thread? You can try to
force a refresh
