Alterations in B Cell and Follicular T-Helper Cell Subsets in Patients with Acute COVID-19 and COVID-19 Convalescents

Round 1

Reviewer 1 Report

The work by Kudryavtsev et al. focused on comparing the humoral immunity features by looking at the expression levels of different B cell subsets and TFH cells, from different Covid patients , one group of acute COVID patients, one group of convalescent patients and one group of healthy controls.

Though it exists already several data in the literature, divergent results have been published and this new set of data can expand our knowledge on the current pandemic.

Though the data presented here are very interesting, I have a few comments that I hope would help improved the manuscript.

-In the first part of their results, the authors focused on the peripheral blood B-cell subset composition and presented their results in two tables. I don’t see the relevance of presenting their data in two tables.

The first table is based on the “Bm1-Bm5” classification using the surface IgD and CD38 markers. However this approach has a few weaknesses, as it fails to separate transitional cells from pre-GC cells, for example (see review from Sanz et al., 2019 Frontiers in Immunol). The second table discriminates more subsets with the use of the CD27 and CD38 markers. I would suggest the authors to change their approach and combine instead all their markers (IgD, CD27 and C38) and make only one table (A possible gating strategy could be as described in Sanz et al., 2019 Frontiers in Immunol or Sosa-Hernandez et al., 2020, Frontiers in Immunol (a reference cited by the authors)). The addition of extra markers (CD21, CD24) it could also help but since they work on whole blood, I understand the impossibility to repeat it.

-In the figures 1 and 2, the authors showed results regarding the TFH cells and they represent the group medians as horizontal bars and quartile ranges. The authors should show instead the means of each groups since we could have difficulties to understand some of the statistics displayed. For example, in the figure 1, the group COVID-19 has its horizontal bar higher that the second group CONV, with error bars similar and yet there is no significative difference with the Healthy controls which is quite surprising. Similar changes should be applied to tables as they displayed values from medians and quartile ranges.

The authors should also mention that there is a tendency of acute COVID patients to present a higher proportion of circulating TFH, which they could correlate with their data from convalescent patients.

- With the increasing amount of vaccinated people (and with different vaccines), it would be interesting for the authors to compare their data with data from these individuals. If they don’t have access to samples from vaccinated people, they should at least discuss with the data from the literature.

-I find that the authors could have discusses more in the discussion section, as well as in the results section. It is lacking of interpretations and/or hypothesis. For example, what would a decrease of transitional B cells would imply? Are these results consistent with what has been seen for other viral infections…

Author Response

We thank the Reviewer#1 for finding our study interesting and for valuable comments.

-In the first part of their results, the authors focused on the peripheral blood B-cell subset composition and presented their results in two tables. I don’t see the relevance of presenting their data in two tables.

We understand the concern of the Reviewer #1 regarding subsets of B cells. Currently, there is no commonly accepted classification of B cells based on cell-surface markers, therefore we used two largely independent B cell classification, and the received data were summarized in two independent tables. We think that both mentioned above approaches are still of interest to the reader, since they are intensively used by many researchers.

- The first table is based on the “Bm1-Bm5” classification using the surface IgD and CD38 markers. However this approach has a few weaknesses, as it fails to separate transitional cells from pre-GC cells, for example (see review from Sanz et al., 2019 Frontiers in Immunol). The second table discriminates more subsets with the use of the CD27 and CD38 markers. I would suggest the authors to change their approach and combine instead all their markers (IgD, CD27 and C38) and make only one table (A possible gating strategy could be as described in Sanz et al., 2019 Frontiers in Immunol or Sosa-Hernandez et al., 2020, Frontiers in Immunol (a reference cited by the authors)). The addition of extra markers (CD21, CD24) it could also help but since they work on whole blood, I understand the impossibility to repeat it.

We thank the Reviewer for this comment. When we use “Bm1-Bm5” and CD27-vs.-CD38 classifications we understand the limitations of our current approach, and we clearly understand that the majority of B cell subsets are partly overlapping subsets. For instance, IgD-vs.-CD38 classification separates most important stages in B cell development from naive to memory B cells (it was recommended by Bohnhorst et al. in 2001), and for a long time it was quite useful for main B cell subset identification (especially for activated (naïve and memory) as well as for purification of B cell that able to enter lymph nodes. It was applied in different papers but it also suffers from significant limitations (for example, Bm1 and Bm2 cells also contain IgD+CD27+ unswitched memory B cells, while Bm2’ also contains transitional cells); yet this classification is largely used for patients with autoimmune disorders. But if we plan to get more information on activated B cells and plasmablast precursors we had to use CD27-vs.-CD38 classification that was build on the notion of CD27 as a universal marker of human memory B cells (thus, we able to distinguish between CD27+ memory cells and CD27- “naïve” B cells) and CD38 that reflected the activation status of circulating B cells. Furthermore, interest in this CD27-vs.-CD38 classification increased with the advent of COVID-19, but before COVID-19 pandemic high plasmablast precursors levels were observed in patients with autoimmunity. Furthermore, there were several attempts aimed to unify or generalize somehow different flow cytometric B cell classification (for example, Wei C, Jung J, Sanz I. OMIP-003: phenotypic analysis of human memory B cells. Cytometry A. 2011 Nov;79(11):894-6 or Kaminski DA, Wei C, Qian Y, Rosenberg AF, Sanz I. Advances in human B cell phenotypic profiling. Front Immunol. 2012 Oct 10;3:302), but, however, progress in this has not yet been revealed. That’s why we used several flow cytometry immunophenotyping approaches in the current paper in order to be able to compare our and literary data. We should mention that we plan to add of extra markers for B cell phenotyping, including CD21 and CD24 as well as CD11c, since the latter one was shown to be effective and informative for patients with acute COVID-19.

 -In the figures 1 and 2, the authors showed results regarding the TFH cells and they represent the group medians as horizontal bars and quartile ranges. The authors should show instead the means of each groups since we could have difficulties to understand some of the statistics displayed. For example, in the figure 1, the group COVID-19 has its horizontal bar higher that the second group CONV, with error bars similar and yet there is no significative difference with the Healthy controls which is quite surprising. Similar changes should be applied to tables as they displayed values from medians and quartile ranges.

As suggested, we changed the data presentation to provide mean +/- SEM to make the paper more reader friendly. All changes were made both to the figures and the table themselves, and to their legends. Furthermore, all relevant changers were made in the text of ‘Results”.

- The authors should also mention that there is a tendency of acute COVID patients to present a higher proportion of circulating TFH, which they could correlate with their data from convalescent patients.

We added to the text of “Results” the following sentence: A trend toward higher proportions of circulating Tfh was noted in patients with acute COVID compared to controls, but did not reach statistical significance.

- With the increasing amount of vaccinated people (and with different vaccines), it would be interesting for the authors to compare their data with data from these individuals. If they don’t have access to samples from vaccinated people, they should at least discuss with the data from the literature.

We understand the concern of the Reviewer #1 regarding post-vaccine immunity. Notably, this is a very novel line of research, and there are several studies addressing the role of B cells and Tfh cells in SARS-CoV-2-specific immunity in convalescent patients, while the data on immunity after SARS-CoV-2 vaccination are still very limited. From our point-of-view many of these questions require further studies. Unfortunately, we do not have the possibility to perform these experiments now. But in future we plan to focus on the blood samples from vaccinated people and we plan to study the dynamics of B and T cell subsets that were analyzed during the current paper. But, presently, the addition of the data about vaccinated people as it was suggested by the Reviewer#1 could distract the reader from the main topic of our paper – the alterations in B and T cell subsets during the acute phase of SARS-CoV-2 infection and unvaccinated patients who had recovered from COVID-19.

 - I find that the authors could have discusses more in the discussion section, as well as in the results section. It is lacking of interpretations and/or hypothesis. For example, what would a decrease of transitional B cells would imply? Are these results consistent with what has been seen for other viral infections…

We thank the Reviewer #1 for bringing up this important issue. We have now included the following text to ‘Discussion’:

It was showed that the level of circulating plasmablast precursors was peaked consistently at day 6 or 7 after acute virus infection and then dropped to baseline levels within 2–3 weeks post symptom onset [33]. For instance, plasmablast (co-expressing CD27 and CD38 CD19+ cells) frequencies were 10–50% of B cells in patients infected with Ebola virus, compared with less than 1% in healthy individuals [34]. Next, dengue-infected patients the frequency of antibody secreting cells was significantly increased and associated with the lower frequency of naïve B cells in peripheral blood [35]. Moreover, frequency of plasmablasts significantly and positively correlated with the frequency of the total Tfh cells during the acute phase of dengue virus infection [36]. Furthermore, the frequency of Tfh cells was significantly higher during the convalescent phase compared to the frequencies in acute phase and in healthy individuals. Interestingly, vaccine-induced plasmablast (circulating CD27+CD38++CD20- B cells) expansion that followed influenza virus vaccination was detected in peripheral blood 7 days after vaccination [37]. Similarly, it was shown that PD-1+ICOS+ circulating Tfh cell subsets at day 7 correlated with plasma specific IgG functional affinity at their peak levels at days 14 and 21 [38].

Taken together these data indicates that the presence of high frequencies of circulating plasma cell precursors and altered Tfh cell subsets could be closely linked with acute viral infections and might constitute residual effects by which COVID-19 can impact the homeostasis of B cell and Tfh cell interactions in the long-term. Furthermore, the better understandings of these long-term alterations in adaptive humoral immunity (which required participation from different types of B cells and CD4+ T cells) in COVID-19 convalescent patients could help us to develop long-term host protection against SARS-CoV-2, effective secondary response against future SARS-CoV-2 infections and predict the generation of a highly effective humoral and cellular immunity in response to vaccination [39]. Many of these questions linked with SARS-CoV-2-specific immunity require further studies.

Reviewer 2 Report

The authors highlight the potential contribution of T follicular helper cells in COVID-19. The manuscript text is clearly written and well organized. The introduction and the discussion are reasonable given the premise of the article. Figures and tables are comprehensive and helpful. I have one minor comment for the author to address.
The authors should indicate the working dilutions of the antibodies used in the Flow cytometry staining protocol.

Author Response

We thank the Reviewer#2 for finding our study interesting and well organized.

- The authors should indicate the working dilutions of the antibodies used in the Flow cytometry staining protocol.

We apologize for this confusing presentation. As suggested by the Reviewer#2, we have now included this information to “Materials and methods”.