Highly Specific Memory B Cells Generation after the 2nd Dose of BNT162b2 Vaccine Compensate for the Decline of Serum Antibodies and Absence of Mucosal IgA

Round 1

Reviewer 1 Report

This study demonstrated that BNT162b2 vaccine (from Pfizer) elicited higher antigen-specific memory B cell responses which can provide better protection despite of decline of serum antibody levels and mucosal IgA. It was a clinical human study to provide valuable information in vaccine efficacy and protection against SARS-CoV2 infection.

1. Line 61, not all COVID-19 vaccines require 2 doses (eg. Janssen).
2. It may be needed to provide the general information of BNT162b2 vaccine (how it was produced, what it made of, suggested immunization schedule, the official protection rate, what we know about the vaccine efficacy until now, etc.).
3. Part 3.1 (Vaccination of HCWs) should be in Materials and Methods rather than Results.
4. The authors performed surrogate neutralization assay rather than using actual virus. Generally, for neutralization assay, the live virus is incubated with serum antibodies and then the virus titer is measured. Can you explain the surrogate neutralization assay methods and principles with more details?
5. And memory B cell analysis by using flow cytometry, the authors have used spike-PE/spike-BUV395 to discriminate low and high affinity cells. If these two antibodies bind to the same epitope, the two antibodies have competition rather than showing low/high affinity. More detailed experiment design and principles are needed for better understanding.
6. The last paragraph (line 678-683) described that memory B cell response is the most important defense system for protection against the viral infection. It may be “true” somehow, but still there are many studies showing antigen-specific T cell responses are also required for the protection. This paragraph needs to be rephrased.

Author Response

Please find attached our responses. 

Thank you for your suggestions 

Sincerely,

Rita Carsetti

Author Response File: Author Response.pdf

Reviewer 2 Report

Overall

The authors studied the induction of SARS-CoV-2 spike and receptor binding domain (RBD)-specific binding antibodies and spike-specific memory B cells of IgM, IgG and IgA isotypes longitudinally after 2 immunizations with the BNT162b2 Vaccine in a cohort of health care workers. They found that 2 doses were required to induce significant serum IgG and IgA specific for the spike and RBD.  This also corresponded with the induction of spike and RBD specific IgG and IgA memory B cells (MBC) (by Elispot).  However, no spike-specific IgA was detected in the saliva of vaccinated subjects, in contrast to convalescent subjects who have recovered from COVID-19.  The authors hypothesized that perhaps the vaccine-induced spike+ IgA+ MBC detected in the peripheral blood transit to the mucosa upon viral exposure and secrete IgA. They address this by studying IgA in breastmilk of recently vaccinated lactating mothers, and indeed they do detect a significant increase of S1-specific IgA in the milk of vaccinated mothers, thus supporting the hypothesis. Lastly, they look at 6 subjects who were infected after vaccination and show an increase in saliva IgA after infection.

The demonstration of IgA in the milk of vaccinated mothers is an interesting finding. However, the claims related to the boosting of the vaccine-induced humoral response after infection are tenuous and poorly substantiated and should be revised or removed as they are anecdotal and the data presented do not directly prove the conclusions.

Major comments:

1. The use of single positive versus double positive spike staining for inferring B cell receptor (BCR) affinity is not adequately scientifically validated in this manuscript and the authors do not reference any previous validation of this inference.  The level of expression of Ig/BCR on the surface of the cell will significantly impact the level of staining/fluorescence intensity and the ability to separate the single spike-PE positive from the double spike-positive (BUV395 & PE).  In addition, the fluorophore used for the single positive spike standing is known to have significant background which isn’t always sufficiently removed by the decoy method employed by the authors. Perhaps the authors have tested this using streptavidin reagents without spike protein to prove that this eliminates the background, but the authors do not describe any control experiments.  It would be necessary to cite appropriate references that validate the inference of BCR “affinity” through this staining method preferably with actual sorting of single vs double positive B cells and generation of monoclonal antibodies with measurements of affinity by established methods.
2. The authors describe results from “a surrogate neutralization assay”. However, the assay that they refer to is the SARS-Cov-2 IgG II kit from Abbott and is not a surrogate neutralization assay. It is a binding assay which detects IgG specific for RBD. It should not be described as a neutralization assay or a surrogate neutralization assay. While SARS-CoV-2 IgG titers to RBD usually correlate with SARS-CoV-2 serum neutralization, reported alone they do not sufficiently capture the variability seen in reported correlations, possibly a result of contributions of neutralizing responses targeting other epitopes or other antibody isotypes.

Minor:

1. The demographics table is very helpful. However, since most of the assays were only performed on a subset of participants, it is critical to add the demographics of the assay subsets, as age and sex have been repeatedly shown to modulate one’s response to vaccination.
2. In the elispot data: The finding that 10% of IgM MBC are specific for spike is highly concerning and brings into question the validity of the assay. It would be helpful if the authors included an irrelevant protein or some other controls.  
3. Also, the MBC gating strategy in the methods suggests that MBCs were gated on CD27+. However, in figure 4A and B, it appears that MBCs were only gated on CD24+/CD38dim and then the CD27 expression was shown for this population (about 90% CD27+ in the example). Please clarify which is the case and would be helpful if you could include a citation for the rationale of usin.g this approach as it differs from the cited Dan etal. publication. It would also be important to understand whether these S+ IgM are also IgD+ and CD27+/- as this would distinguish between naïve and transitional/marginal zone B cells. It would also be important to confirm that these B cells are indeed specific for spike by sorting and making monoclonal antibodies. Please add citations if this has been done by your group and reported elsewhere.
4. In the description of the IgA ELISA, please clarify in the methods that the ELISA is only to the S1 subunit.
5. Ideally, it would be nice to confirm that the IgA being detected is secretory IgA, but not necessary. Maybe good to address this assumption in the discussion
6. Figure 5C, for the >6 days sample, please provide more specifics about the time frame which participants were sampled.
7. Regarding Fig 5D-F, and the conclusion at lines 538 and 539: “A further evidence of the function of 538 MBCs is the increase of serum specific IgG after the positive swab in 5 out of 6 HCWs 539 (Figure 5F).” The presented data doesn’t support this conclusion. It is a reasonable hypothesis and should be reworded as such. Alternatively, there could be de novo spike-specific B cells induced by infection.  The authors would need to conduct BCR sequencing to address this.
8. Figure 5 states that Wilcoxon matched-pairs signed-rank test was used to measure significance.  However,  there is a mix of comparisons presented in the figure. For instance, this would be appropriate for comparing the 6 infected subjects between the 2 timepoints but not comparing that 6 to the other columns where a Mann-Whitney would be more appropriate.  Also, the last 2 columns “infection after vaccination” appear to have more than 6 data points?

Author Response

Dear, please find attached our responses. 

Thank you for your suggestions.

Sincerely, 

Rita Carsetti

Author Response File: Author Response.pdf