Increased COVID-19 Mortality and Deficient SARS-CoV-2 Immune Response Are Not Associated with Higher Levels of Endemic Coronavirus Antibodies
, Joshua O. Amimo 2, Sergei A. Raev 2, Soledad Fernández 5, Mirela Anghelina 5, Shan-Lu Liu 3,6,7,8, Mark P. Rubinstein 9,10, Daniel M. Jones 11, Linda J. Saif 1,2 and Anastasia N. Vlasova 1,2,*Round 1
Reviewer 1 Report
In this work, Dr Adhikari studied the link between pre-existing immunity to common cold coronaviruses and COVID-19 outcome in a cohort of 94 patients.
To address this question, the authors determined in silico with bioinformatic tools, immunogenic peptides from spike and nucleoprotein proteins of 7 currently known human coronaviruses. They predicted 37 peptides and after a series of evaluation, they retained 27 peptides for further evaluation. The authors next validated their ELISA assay on panels of human and rabbit sera to finally select a limited set of peptides for the core of the work. The authors classified the cohort of 94 patients (20 non-Covid-19 and 74 Covid-19 positive) into three categories according to disease severity. They then tested for IgA, IgA and IgG antibodies directed against the different peptides and concluded that elevated levels of common cold coronaviruses are not associated to increased mortality due to COVOD-19.
The impact of common cold, or endemic, coronaviruses on COVID-19 severity is a matter of intense scientific debate. It has for example been speculated that endemic coronaviruses is responsible, in part, for the milder impact of COVID-19 in Africa by providing cross-protection. This has to be demonstrated. Hence, the work by Dr Adhikari and colleagues is among the very few to address this issue in a well-documented cohort. They are commended for this endeavour.
The work deserves some comments.
1. The control panel for assay validation is very limited: only 7 pre-pandemic human samples and 10 Covid-19 positive (3 times each patient) samples. An assay validation necessitates more samples in a panel.
2. The assay performances (sensitivity, specificity, accuracy, etc) are not provided. One of the arguments of using synthetic peptides, is to avoid cross-reactions among coronaviruses. It is good practice to provide such information for novel assay.
3. In relation with the item above, it is important to provide raw binding data, here the ODs, and the range of variability for negative and positive samples.
4. Do the authors know the time between infection (first positive PCR) and testing? As they know, the kinetics of the different immunoglobulins are different upon infection and the sensitivity of serological assays varies accordingly. If the information is available, it is worth to analyse they data by stratifying with this parameter.
5. There is really an overuse of heatmaps representation of the data. Is there really no other way to present them?
Overall, the manuscript and the results therein deserve the journal’s consideration. However, the manuscript can be improved in several aspects.
Author Response
Reviewer 1.
RE1: 1. The control panel for assay validation is very limited: only 7 pre-pandemic human samples and 10 Covid-19 positive (3 times each patient) samples. An assay validation necessitates more samples in a panel.
AU: Thank you for your suggestion. This was our oversight: we failed to mention the blinded SeroNet panel we used for our ELISA validation (n=109). These data are now included it in the manuscript (Table 2).
RE1: 2. The assay performances (sensitivity, specificity, accuracy, etc) are not provided. One of the arguments of using synthetic peptides, is to avoid cross-reactions among coronaviruses. It is good practice to provide such information for novel assay.
AU: Thank you for your comment. We have calculated sensitivity, specificity, accuracy for SARS-CoV-2 ELISA (Table 3). However, it is impossible to do this for other CCCoV ELISA tests as characterized panels of human positive/negative samples are not available.
RE1: 3. In relation with the item above, it is important to provide raw binding data, here the ODs, and the range of variability for negative and positive samples.
AU: Per reviewer’s request we are sharing the OD650 value ranges for previously characterized SARS-CoV-2 seropositive/seronegative samples: 0.9-1.64 for seropositive samples, and 0.05-0.14 for seronegative samples. However, we’d like to emphasize again that these data are not available for CCCoV Abs due to the unknown exposure status of each patient and the lack of characterized validation panels. The available commercial seropositive and seronegative CCCoV samples had the OD650 values of 0.03-0.12 (SN) and 0.47-1.08 (SP).
RE1: 4. Do the authors know the time between infection (first positive PCR) and testing? As they know, the kinetics of the different immunoglobulins are different upon infection and the sensitivity of serological assays varies accordingly. If the information is available, it is worth to analyse they data by stratifying with this parameter.
AU: We have now clarified in the manuscript (P2, LL89-90), that all patients got tested at the admission, and this was the time of the first blood draw (week 1). However, the time that passed since contracting the infection and the hospital admission is not known for any patients. Also, we would like to emphasize that we did include the data on the kinetics of SARS-CoV Ab responses week 1- week 3 (Figure 7). This figure shows that IgM levels remained low and unchanged for most patients, which seems reasonable because patients generally develop severe COVID-19 10-14 days after becoming infected, and IgM Ab levels decline fast. Also, expectedly IgG and IgA Ab levels increased during the observation period.
RE1: 5. There is really an overuse of heatmaps representation of the data. Is there no other way to present them?
AU: We believe that heatmaps are the optimal if not the only possible way to present our data. This is because we present the data for 2 proteins (S and N), three isotypes (IgM, IgA and IgG) and 5 different coronaviruses that are further grouped based on various demographic and clinical parameters. Thus, heatmaps represent the optimal way to present the data. It would require too many graphs to present these data, which would be hard to comprehend. For this reason, heatmaps are often used in similar studies (Miyara et al., 2022; Lin et al., 2022; Hicks et al., 2021; Selva et al., 2021; Song et al., 2020)
Reviewer 2 Report
This is a well written and well documented manuscript. The Authors made a lot of work by synthesizing several peptides which are expression of N and S relevant epitopes in seasonal coronaviruses and SARS-CoV-2 in order to minimize/eliminate the possible cross-reactivity. However, the issue of the immunological and clinical relevance of pre-existing anti seasonal coronavirus antibodies is quite controversial in the literature, as underlined by the Authors themselves, probably because it is multifactorial, thus preventing definitive conclusions be drawn.
The Authors could discuss the compatibility of their results with the "antigenic sin theory", which has been recently summarised (Pathogens 2023, 12, 868 https://doi.org/10.3390/ pathogens12070868).
Finally, in Figures S1 and S7 four asterisks in SARS-CoV-2 IgA and IgG, respectively are reported, whereas in the legend the maximum of only three asterisks is reported.
Author Response
Reviewer 2
RE2: This is a well written and well documented manuscript. The Authors made a lot of work by synthesizing several peptides which are expression of N and S relevant epitopes in seasonal coronaviruses and SARS-CoV-2 in order to minimize/eliminate the possible cross-reactivity. However, the issue of the immunological and clinical relevance of pre-existing anti seasonal coronavirus antibodies is quite controversial in the literature, as underlined by the Authors themselves, probably because it is multifactorial, thus preventing definitive conclusions be drawn.
AU: Thank you. Yes, this field is full controversial findings due to multifactorial confounders. In this study, we wanted to generate evidence for the role of CCCoV Abs in severe COVID-19 using samples from our unique patient cohort. Our data demonstrate that there is no evidence of CCCoV Ab contributing to more severe COVID19.
RE2: The Authors could discuss the compatibility of their results with the "antigenic sin theory", which has been recently summarised (Pathogens 2023, 12, 868 https://doi.org/10.3390/ pathogens12070868).
AU: Thank you for the suggestion, we have included the discussion as suggested (P2, LL62-65, and P15, LL337-339).
RE2: In Figures S1 and S7 four asterisks in SARS-CoV-2 IgA and IgG, respectively are reported, whereas in the legend the maximum of only three asterisks is reported.
AU: Thank you for noticing it, we have included the missing information in our revised manuscript in supplemental section.
