Thread by @profshanecrotty

1/ Now for the thornier data.

2/ I covered the great news from the Novavax COVID vaccine UK Phase 3 trial.

https://twitter.com/profshanecrotty/status/1355259172640591872?s=20

3/ I covered the excellent news from the J&J COVID vaccine Phase 3 international trials.

https://twitter.com/profshanecrotty/status/1355294402755739650?s=20

4/ The thornier information was:
Infections with the South Africa variant (SA) ("attack rate") were similar in SA between seronegative subjects and already seropositive subjects.

https://twitter.com/kakape/status/1354922533997473798?s=20

5/ That implies no immunity to the SA variant (B.1.351) from previous infection. If accurate, that is a serious problem. However, the data are difficult to square with a lot of other observations, discussed in tweet clusters below.

6/ The Lumley et al study, the SIREN study, the Qatar study, and others of COVID-19 subjects, have all observed quite high levels of protection against re-infection (COVID-19 a second time) for a period of at least 6-9 months.
nejm.org/doi/10.1056/NE…

7/ Our study of immune memory observed robust and complex immune memory, with slow declines (no particular drop-off at 6 months).
science.sciencemag.org/content/early/…

8/ and thus an expectation that the protective immunity against COVID-19 observed at 6-8 months (in the studies above) will likely be maintained over a period of years, with a slow decline in the interim.

9/ If, let's say, the serological work by Prof. Penny Moore and colleagues is most predictive regarding the SA variant (and Moore is certainly an outstanding scientist),

biorxiv.org/content/10.110…

10/ That study found complete loss of neutralization of the SA variant in about half of subjects (501Y.V2, formally B.1.351). (re-graphed by @trvrb )

https://twitter.com/trvrb/status/1351785356782313473?s=20

11/ At face value, with a strong bias towards inferred importance of neutralizing antibodies in protection, one might then predict a 50% loss in protection. But not 100%.

12/ And there is plenty of evidence that T cells contribute to protective immunity against COVID-19.
doi.org/10.1016/j.cell…

13/ And it is very unlikely any variants are escaping the broad CD4 and CD8 T cell memory to SARS2 generated by COVID-19. (two study refs)
science.sciencemag.org/content/early/…

14/ Including this new COVID T cell work from @SetteLab

https://twitter.com/profshanecrotty/status/1355262019562209282?s=20

15/ Additionally, the Pfizer and Moderna vaccines provide significant protection against COVID after a single immunization. Presuming that protection is mediated by adaptive immunity, it either means protection of humans :

16/

(i) requires very low levels of neutralizing antibodies, because most subjects are below the limit of detection;

or

(ii) the protection is mediated by T cells

or

(iii) the protection is mediated by other antibody functions.

17/ In any of those 3 scenarios, there is a disconnect between the excellent immunity reported after immunization with the Pfizer or Moderna vaccines— or after infection with 'regular' SARS2 — versus the new data from SA.

18/ Additionally, the relatively modest drop in efficacy for the J&J vaccine between America and South Africa doesn't  align with a putative massive loss in natural immunity of COVID-19 subjects against the SA variant.

https://twitter.com/profshanecrotty/status/1355294409508548609?s=20

19/ Is the SA variant intrinsically much more difficult to protect against, independent of immune escape? The virus does appear to gain a great deal of affinity for ACE2 with 501 and 484 mutations. (ref below, consistent with @jbloom_lab data in Starr et.)
biorxiv.org/content/10.110…

20/ But, receptor affinity doesn't automatically correlate with infectivity. (And also see the points above.)

21/ Someone on Twitter did point out classic flu work by Hensley and Yewdell that receptor binding avidity can drive flu escape (antigenic drift), (and I do think that affinity and escape are being linked in a similar biological way for SARS2), but:

22/ But, SARS2 D614G swept the globe and actually has lower affinity for ACE2, for example. (riddle me that!)

23/ And, even if the Spike affinity gain matters, it should be most "baked in" to the neutralizing antibody titers (which looks to be the case for flu)

24/ That said, from a correlates of protection perspective, I think it is important for the field to reconsider how different SARS2 neutralization assays may reveal this to differing degrees, depending on the pseudovirus, and on the amounts of ACE2 expressed on the target cell.

25/ Next, It was also stated that there wasn’t an observable difference in protection between the baseline zero positive and zero negative individuals receiving the Novavax vaccine in South Africa. (Stated in the Q&A). That doesn’t make a lot of sense.
edge.media-server.com/mmc/p/ex338sjy

26/ The Novavax vaccine clearly showed outstanding efficacy in the UK. So why would there be no difference between baseline seropositive and seronegative immunized people? Something about this phase 2b seems to be an outlier.

27/ So, here we are. Not all of the data agree. No firm conclusions can be drawn, to be sure, and this particular chunk of data is certainly worrisome. It is certainly important to keep gathering quality data about this variant and others, and conclusions have been pivoting.

28/ Bottom line is still that the current COVID vaccines should have important efficacy against all SARS2 variants known, and getting as many people vaccinated as possible as quickly as possible is very important to blunt the spread.