Bloom Lab Profile picture
May 21, 2022 19 tweets 10 min read Read on X
In this thread, I discuss what are candidates for the next mutational steps in evolution of #SARSCoV2 to evade neutralizing antibodies. TLDR: in addition to mutations at sites 452 & 486 in BA.4/5, watch for mutations at sites 346-348, 356, 444-446, & 468. (1/n)
As background, human CoVs evolve to erode antibody neutralization (). As result, typical person infected with common-cold CoV every few years, which “updates” their immunity to newer strains until a few more years of viral evolution erodes it again (2/n)
As most people know, this process is ongoing for #SARSCoV2, with new variants continuing to erode neutralization by antibodies elicited by old strains (like one in vaccine), which contributes to increasing re-infections & vaccine breakthroughs () (3/n)
Question I address here is which specific mutations are candidates for “next steps” in this antigenic evolution. I’ll focus only on spike’s receptor-binding domain (RBD), which is dominant but not exclusive target of neutralizing antibody response: (4/n)
To prospectively identify antibody escape mutations, @tylernstarr @AllieGreaney developed deep mutational scanning to map all mutations that reduce binding (science.org/doi/10.1126/sc…). Below is map of mutations that escape one monoclonal antibody (LY-CoV555 = bamlanivimab) (5/n)
Of course, human antibody response to infection & vaccination is polyclonal. To understand escape from that, we need to identify how mutations affect the spectrum of different neutralizing antibodies generated by human immune system (6/n)
As simple example, consider mix of 3 antibodies shown below. Full escape from mix requires combining mutations that escape each antibody. Note mutations can be redundant if they escape same antibody (484 & 490) or synergistic if escape different antibodies (484 & 417). (7/n)
We previously formalized this idea into an antibody escape calculator, which leverages deep mutational scanning for a large set of antibodies to estimate effects of mutations on polyclonal serum: academic.oup.com/ve/article/8/1… (8/n)
Originally calculator used data for few dozen antibodies generated by @tylernstarr @AllieGreaney in our group, but recently Sunney Xie, Richard Cao, @facyanOvO et al at @PKU1898 generated HUGE set of data for ~1,500 antibodies! (9/n)
Their data, which @facyanOvO generously posted on GitHub (github.com/jianfcpku/SARS…), now compose vast majority of info used by escape calculator. Such a large data set enables some pretty cool analyses. (10/n)
First, as has now been extensively described, antibodies elicited by early #SARSCoV2 (eg, current vaccine) that neutralize early strains (eg, Wuhan-Hu-1) strongly escaped by mutations at site 484 & also sites like 417, 346 & 446—all of which are mutated in some variants. (11/n)
Escape calculator also shows how Omicron BA.1 and BA.2 both have extensive escape from antibodies elicited by early #SARSCoV2, as is now well described. Importantly, it shows that 486 is site of largest escape from residual antibodies that still neutralize BA.1 / BA.2 (12/n)
In fact, using calculator we predicted back in Dec 2021 that site 486 was one to watch for future evolution (). And just last month, @tuliodna reported it was mutated in BA.4/BA.5, which have largest antibody escape of any variants yet described. (13/n)
So what might be virus’s next steps in antigenic evolution? We can subset on just antibodies elicited by early (pre-Omicron) strains that still neutralize BA.2. In addition to mutations already in BA.4/BA.5, sites of possible future escape include 346, 444-446 & 499. (14/n)
But importantly, sites of escape in BA.2 somewhat different for antibodies elicited by early strains FOLLOWED by BA.1 breakthrough. Comparing below image to that in prior Tweet you can see some different peaks, such as at 347-348, 356 & 468. (15/n)
This divergence in effects of mutations between people +/- prior BA.1 breakthrough is because exposure history shapes immunity. We will increasingly see variation in how #SARSCoV2 mutations impact antibodies of different people, as for influenza: elifesciences.org/articles/49324 (16/n)
Finally, antibody-escape calculator is available at jbloomlab.github.io/SARS2_RBD_Ab_e… You can select antibodies by exposure history (including past infection w SARS-CoV-1) & by what strains they neutralize, & click on sites to see impacts of mutations. (17/n)
Also, some slides going into more detail on the material in this thread are at slides.com/jbloom/escape-… (18/n)
And really embarrassingly, I mis-tagged @fucyanOvO who shared so much of the great data analyzed here. Sorry... 😞 (19/n)

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Bloom Lab

Bloom Lab Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @jbloom_lab

Aug 20
In new study led by Bernadeta Dadonaite, we measure how spike mutations affect function & antigenicity of spike of KP.3.1.1 strain of SARS-CoV-2.

Sheds light on how key neutralizing epitopes are changing & importance of RBD up/down motion.

biorxiv.org/content/10.110…
We examined spike of KP.3.1.1, a strain from late 2024 / early 2025 similar to current variants

KP.3.1.1 & other recent variants have >60 spike amino-acid mutations relative to early pandemic strains, as spike has evolved at extraordinary rate of >10 mutations/year on avg Image
We previously developed pseudovirus deep mutational scanning (), which uses non-replicative viral particles to safely study spike mutations.

Here we used approach to measure how mutations to KP.3.1.1 spike affect five phenotypes, as shown below. pubmed.ncbi.nlm.nih.gov/36868218/Image
Read 12 tweets
May 27
In new study led by @timcyuu, we measure how mutations to H3 flu HA affect cell entry, stability & antibody escape

We find pleiotropic effects of mutations on these phenotypes shape evolution: epistasis alleviates cell-entry but not stability constraints

biorxiv.org/content/10.110…
We used pseudovirus deep mutational scanning to characterize all mutations to a recent H3N2 HA. This approach uses virions that can only undergo one round of cell entry & so are not pathogens capable of causing disease.

All measurements available here: dms-vep.org/Flu_H3_Massach…
As can be seen below, constraint due to mutational impacts on cell entry are widely distributed across HA including receptor-binding pocket and fusion peptide. But mutational constraint due to HA stability is concentrated at trimer and HA1-HA2 interface. Image
Read 8 tweets
Mar 12
In study led by Cassie Simonich & T McMahon, we quantify antigenic evolution of RSV F. Important because:

1⃣ RSV top cause of infant hospitalization in USA

2⃣ New antibodies & vax can prevent hospitalizations

3⃣ But will virus evolution erode efficacy?

biorxiv.org/content/10.110…
RSV has high burden in infants: top cause of infant hospitalization in USA, 2nd-leading cause of infant mortality globally

A monoclonal antibody (nirsevimab) recently recommended for infants born in USA in RSV season. It prevents hospitalizations

pubmed.ncbi.nlm.nih.gov/38457312/
RSV vaccines also now approved to protect infants (via maternal vaccination) & elderly

But some viruses evolve to erode antibodies and vaccines

Will RSV do same? Worryingly, a Regeneron antibody failed phase 3 trials due to resistance in some RSV strains
pubmed.ncbi.nlm.nih.gov/32897368/
Read 11 tweets
Jan 21
In new study, we find dramatic differences in specificities of serum neutralizing antibodies in infants w single infection by a recent SARS-CoV-2 strain versus adults/children imprinted by an early viral strain.

biorxiv.org/content/10.110…
As background, immune response to a virus is “imprinted” by first exposure, since later exposures to new viral strains often activate pre-existing B-cells.

For SARS-CoV-2, most people globally imprinted by an early viral strain from either vaccination or infection in 2020-2021.
However, small but growing fraction of population has instead been imprinted by more recent viral strain.

Specifically, we compared adults/children imprinted by original vaccine then infected w XBB* strain in 2023 vs infants only infected w XBB* in 2023. Image
Read 9 tweets
Nov 21, 2024
I’ve updated SARSCoV2 antibody-escape calculator w new deep mutational scanning data of @yunlong_cao @jianfcpku

My interpretation: antigenic evolution currently constrained by pleiotropic effects of mutations on RBD-ACE2 affinity, RBD up-down position & antibody neutralization
First, the updated escape calculator is at

As shown below, it is remarkable how much antigenicity of RBD has changed over last 4 yrs. jbloomlab.github.io/SARS2-RBD-esca…Image
Updated data for calculator from this paper by @yunlong_cao’s group (nature.com/articles/s4158…), described in this thread by first author @jianfcpku:
x.com/jianfcpku/stat…

Calculator show how much mutations at each RBD site escape binding by set of neutralizing antibodies
Read 13 tweets
Nov 16, 2024
@Nucleocapsoid @HNimanFC @mrmickme2 @0bFuSc8 @PeacockFlu @CVRHutchinson Good observations. See also this thread posted by @SCOTTeHENSLEY:

I have added a few notes to the bottom of that thread.

To recap here:bsky.app/profile/scotte…
@Nucleocapsoid @HNimanFC @mrmickme2 @0bFuSc8 @PeacockFlu @CVRHutchinson @SCOTTeHENSLEY To add to thread linked above, human British Columbia H5 case has a HA sequence (GISAID EPI_ISL_19548836) that is ambiguous at *both* site Q226 and site E190 (H3 numbering)

Both these sites play an important role in sialic acid binding specificity
@Nucleocapsoid @HNimanFC @mrmickme2 @0bFuSc8 @PeacockFlu @CVRHutchinson @SCOTTeHENSLEY If you are searching literature, these sites are E190 and Q226 in H3 numbering, E186 and Q222 in mature H5 numbering, and E202 and Q238 in sequential H5 numbering (see: )dms-vep.org/Flu_H5_America…
Read 6 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(