Thread by @michaelzlin

In a recent thread, I analyzed Novavax vax data and hypothesized it might elicit broader antibodies than the RNA vaccines and thus be more protective vs. variants.

Now we'll explore differences between Novavax and RNA vaccines that could account for this

https://twitter.com/michaelzlin/status/1488957682874679298

Novavax is S protein assembled with lipids in viral-like nanoparticles. These are taken up by APCs (DCs, macrophages), and S fragments presented on MHC2 to activate CD4+ T cells. B cells with surface Ig recognizing S and coactivation by CD4+ T cells via MHC2 then get expanded.

RNA vax deliver RNA to cells to express S. Some cells die with S at the ER-Golgi and surface. S-containing membranes are digested by APCs and S fragments presented on MHC2 to activate CD4+ T cells. B cells bind S on cells to get their 1st signal.
(Pic: doi.org/10.1038/s41541…)

An additional thing happens with RNA vax that doesn't happen with Novavax, which is that the S protein is cleaved into S1 and S2 fragments at the infamous polybasic furin cleavage site. As illustrated above, that creates a free S1 fragment (containing the RBD) that is secreted.

S1-S2 cleavage doesn't occur in Novavax because Novavax mutated the furin cleavage site to prevent this cleavage. This was probably done so the full-length S could be purified from the insect cells used to express it.
(Pic: science.org/doi/10.1126/sc…)

One possibility is RNA vax could be S1 biased in terms of antibody induction. Secreted S1 is trimeric so can activate BCR in naive B cells, and being soluble could be more easily processed for presentation on MHC2 to get the 2nd signal from CD4+ T cells.

In Novavax, physical linkage of S1 and S2 might assure epitopes across the protein are presented together which might allow proportionally more activation of S2-specific B cells than with RNA vax.

There are conserved antibody epitopes in S1 and S2 that elicit broadly neutralizing antibodies. But most antibodies tend to bind RBD of S1, and the variants that have arisen have been immunoevasive due to RBD mutations.

S2 mutations are less common, and antibodies that target certain sites on S2 can also be neutralizing. So one possibility is that S1/S2 balance in Novavax tends to create more antibodies to S2 which are broadly neutralizing.
frontiersin.org/articles/10.33…

In addition to a potential function in balancing S1 and S2 presentation, @florian_krammer found that removing the FCS also enhances immunogenicity of protein vaccines. That is, protein vax without the FCS protected mice better than those with the FCS.
journals.asm.org/doi/10.1128/mB…

Since full-length S protein was used in both FCS-mutated and FCS-intact protein vaccines in the Krammer study, the superiority of the FCS-mutated protein can't be due to S1-S2 balance. Rather they suggested the protein can just be more stable overall.

The above is pure speculation; we don't have proof Novavax protects more broadly than current RNA vaccines. We would need a head-to-head comparison to know.

But details of antigen composition could matter. It would be useful perhaps to try a RNA vax with the FCS mutated too.

Indeed you can see the S1 bias of the RNA vaccines here. Note how RNA-induced Abs against S1 are higher than an infection (left chart), whereas RNA-induced Abs against S2 are lower than infection (right chart).

nejm.org/doi/full/10.10…

Would be very informative if @Novavax can test their patients' S1 and S2 antibody responses, and compare to RNA vax recipients. It can be done in a day and cost a few dollars a sample.

An academic lab with sample access could do it too.

Just found this recent study that looked at S1 vs S2 responses. Turns out some people have Abs from earlier coronavirus infections (OC43 and HKU1) that cross-react to SARSCoV2 S2 — that's how conserved it is.

RNA vax layers on top a mostly S1 response.
nature.com/articles/s4159…