Review 3: "The Majority of SARS-CoV-2 Plasma Cells are Excluded from the Bone Marrow Long-Lived Compartment 33 Months after mRNA Vaccination"

RR:C19 Evidence Scale rating by reviewer:

• Potentially informative. The main claims made are not strongly justified by the methods and data, but may yield some insight. The results and conclusions of the study may resemble those from the hypothetical ideal study, but there is substantial room for doubt. Decision-makers should consider this evidence only with a thorough understanding of its weaknesses, alongside other evidence and theory. Decision-makers should not consider this actionable, unless the weaknesses are clearly understood and there is other theory and evidence to further support it.

***************************************

Review: This manuscript addresses the longevity of SARS-CoV-2 antibody responses by describing the phenotype of human bone marrow plasma cell specific to spike protein. The authors conclude that mRNA vaccine-induced immunity fails to produce CD19^- long-lived plasma cells (LLPC) specific to SARS-CoV-2 spike protein.

The authors set out to determine the heterogeneity of the SARS-CoV-2 spike protein specific plasma cells in the bone marrow of human donors. They applied their previously described flow cytometry protocol for identification of immature and mature plasma cells based on CD19 expression (PMID 26187412). Using this system for defining plasma cell subsets, the authors sorted out plasma cells from human bone marrow aspirates. They focused on 3 populations that consist of 2 CD19⁺ subsets that are considered short-lived and 1 CD19^- subset that they define as LLPC. Spike-specific plasma cells were compared to influenza-HA- and Tetanus toxoid-specific plasma cells. Their main conclusion is that IgG-secreting plasma cells specific to spike protein are largely absent from the putative LLPC population, which contrasts with both HA and tetanus toxoid. Notably, they observe a robust population of CD19⁺plasma cells specific to spike. The authors conclude that SARS-CoV-2 fails to induce LLPC by either infection or vaccination with mRNA vaccines.

This study presents interesting findings that have implications for our understanding of how LLPCs are defined and how they are generated. However, the authors conclusions are entirely supported by their data. The authors themselves acknowledge some of the caveats in their discussion. For example, they note that they observe heterogeneity in the CD19⁺ bone marrow plasma cell population that could include LLPCs. Another issue is the small sample size and variability of timing throughout the study. It is difficult to design an ideal human study with a large sample size of volunteers willing to provide a bone marrow aspirate, but the low number of study participants makes some of the concluding statements too strong. The choice of influenza and tetanus toxoid as controls for LLPC is interesting considering that both are known to be relatively short-lived responses compared to other common vaccines. Additionally, these control antigens have to potential for repeated exposure throughout life, with Tetanus vaccination recommended every 10 years throughout adulthood and influenza vaccination recommended yearly with potential environmental exposure as well. A perhaps more compelling control would be vaccination against Yellow Fever due to the fact that it is not commonly provided in the United States, and it offers robust life-long immunity. This would provide a true primary immune response to compare to the novel coronavirus.

Overall, the data are very interesting and provide a valuable view into the phenotype of bone marrow plasma cells in response to SARS-CoV-2 immunization and infection. However, the variability of the sample collection and immunization/infection history make the sample size too small to support the strong conclusion that mRNA vaccines fail to generate LLPC.