Humoral response to SARS‐CoV‐2 and seasonal coronaviruses in COVID‐19 patients

Abstract We used enzyme‐linked immunoassay methods to measure the prevalence and the levels of antibody responses to the nucleocapsid (N) protein of severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) and four seasonal human coronaviruses (HCoV‐OC43, HCoV‐HKU1, HCoV 229E, and HCoV‐NL63) in a cohort of 115 convalescent plasma donors infected with SARS‐CoV‐2 (1–61 days after symptom onset) compared to antibody levels in 114 individuals with no evidence of a recent infection with SARS‐CoV‐2. In the humoral response to the four seasonal coronaviruses, only HCoV‐HKU1‐ and HCoV‐229E‐assays showed slightly elevated antibody levels in the COVID group compared to the control group. While in the COVID‐group the levels of SARS‐CoV‐2 antibodies correlated significantly with disease severity, no association was found in the levels of antibodies against the seasonal coronaviruses. The most striking result in both groups was that the levels of antibodies against all tested coronaviruses, including the new SARS‐CoV‐2 showed a highly significant correlation with each other. There seems to be an individual predisposition to a weaker or stronger humoral immune response against all known seasonal human coronaviruses including the new SARS‐CoV‐2, which could lead to a definition of low and high responders against human coronaviruses with potential impact on the assessment of postinfection antibody levels and protection.


| BACKGROUND
At the end of December 2019, the first cases of severe pneumonia with an unknown cause were reported in Wuhan, China. 1,2 The new virus is very similar to the beta-coronavirus SARS-coronavirus and was, therefore, named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). SARS-CoV-2 is the cause of the current COVID-19 pandemic. 3 In addition to SARS-CoV-2, we know of six other human coronaviruses today. Among them are four seasonal human coronaviruses, that widely circulate in the human population and are responsible for 10%-30% of mild self-limiting upper respiratory tract infections. 4 A protective effect of previous infections with seasonal human coronaviruses against infection with SARS-CoV-2 has been described as well as an association between previous infections with seasonal human coronaviruses and less severe COVID-19 disease. 5

| Study population
Characteristics of the study population were summarized in Table 1. The study was approved by the local ethics committee al Heinrich-Heine University Duesseldorf (study no. 2020-1148).
Informed consent was obtained from all of the SARS-CoV-2 convalescent plasma donors (N = 115) before blood sampling. The control sera (N = 114) derived from routine presurgery serology testing excluding patients with known underlying chronic diseases or immunosuppressive therapy. The basic demographics were similar for both groups ( All blood samples of the study group and control group were collected between March 2020 and July 2020 and stored at −20°C. Additionally, for 10 post-COVID plasma donors it was possible to obtain sequential sera for up to 6 months after onset of symptoms.

| SARS-CoV-2 neutralization antibodies
To determine the SARS-CoV-2 neutralization activity of the plasma samples, a serial endpoint neutralization test for SARS-CoV-2 was developed based on our prior work. 9 Serial dilutions of heatinactivated (56°C, 30 min) serum samples in duplicates were preincubated in cell-free plates with 100 TCID 50 units of SARS-CoV-2 for 1 h at 37°C. After preincubation, 100 µl of cell suspension containing 7 × 10 4 /ml Vero cells (ATTC-CCL-81) were added. Plates were incubated at 37°C, 5% CO 2 for 4 days before microscopic inspection for virus induced cytopathic effect.

| Statistical analysis
The data were analyzed using SPSS Statistics 25 (IBM © ) and Graph-Pad Prism 5.01 (GraphPad Software). Categorical data were studied using Fisher's exact test or Pearson's chi square test, depending on the sample size. The 95% confidence interval (95% CI) for proportions was calculated using the modified Wald method. Quantitative data were analyzed by the nonparametric Mann-Whitney U test for two groups and by the Kruskal-Wallis test for more than two groups.
Correlation coefficients (r2) were determined to calculate the quality of the fit of the quantitative results from the different ELISA assays.
All of the tests were two tailed, and p < 0.05 was considered to be statistically significant. respectively no significant differences in antibody levels were found between the two groups ( Figure 1). Gender showed no significant influence on antibody levels. In both groups, anitbodies against seasonal coronaviruses were slightly higher in males than in females (mean difference: 27.5 AU/ml, range from 3.3 to 57.0 AU/ml), but the difference was not significant. Likewise, age had no effect on the level of antibodies to seasonal coronaviruses with the exception of HCoV-HKU1 antibodies, which showed a significant increase with age (p = 0.041), but only in the COVID group (data not shown).  (Figure 2). In contrast, the antibody levels against the four seasonal coronaviruses showed no significant difference or even a trend.

|
The median antibody level was around 100 AU/ml in all cases ( Figure 2).
The mean age of the Grade 1-3 groups showed an increase from 39.12 to 48.67 years, but the differences were not statistically significant (data not shown).

| Correlation analysis of all coronavirus antibodies in both groups
The most striking result of this study was a high correlation of antibody titers between all coronavirus types (Tables 2 and 3). In the group of COVID patients, highly significant correlations (p < 0.0001) were found between the level of SARS-CoV-2 antibodies and seasonal coronaviruses ( Table 2) Table 3).
The highest correlation coefficient in the control group was found in the relationship between HCoV-OC43 and HCoV-HKU1 (R = 0.808).
Sera from both groups were tested for the presence of antibodies to the measles virus. As expected, there was a high ser-

| DISCUSSION
We developed a specific two-step capture ELISA protocol to detect antibodies against the nucleocapsid protein of the four seasonal human coronavirus and SARS-CoV-2 based on a technique we previously published for human polyomaviruses. 7,8 As previously shown by others, we also found a correlation between the level of antibodies to SARS-CoV-2 and the severity of the disease. [10][11][12][13] In contrast to this relationship, we found no correlation between the severity and the level of antibodies to seasonal coronaviruses. It is possible that in some viral infections, non-neutralizing antibodies may enhance infection and not be protective against reinfection. Antibodydependent enhancement was discussed earlier with SARS-CoV pathogenicity, was shown with other virus infections, e.g., Dengue virus serotypes 1-4 and is also discussed with SARS-CoV-2 infections. [14][15][16][17] As a caveat, it is noteworthy that we have shown this lack of correlation only in antibodies to the N protein and not in antibodies to other SARS-CoV-2 proteins. However, on the basis of the available data, we have no evidence that previous infections with seasonal coronaviruses have an impact on the severity of SARS-CoV-2 disease.
In the 10 donors with consecutive sera in most cases increased antibody levels against seasonal coronaviruses were found in the initial serum which decreased in the subsequent follow up. We interpreted this finding as a cross stimulation of B cells (Figure 3). This was seen in several cases with antibodies against more than one coronavirus and was also found in individuals with no or only a low level of SARS-CoV-2 antibodies ( Figure 3F-I), thus, it is unlikely to be caused by pure cross-reactivity of antibodies. The rapid decrease in SARS-CoV and SARS-CoV-2 antibodies was described in a large number of papers, but in the present work, we were additionally able to show for the first time that this decrease also affects antibodies against seasonal coronaviruses that had initially been increased. 13,[18][19][20][21][22] We cannot say from our data whether these crossstimulated antibodies are also functionally active, i.e., whether they can take over a protective function for the host. Measles antibodies were chosen due to the greater than 90% seroprevalence in adults.
Another interpretation of our data leads to the term of "trained immunity": There is evidence, that influenza vaccine can induce trained immunity responses against SARS-CoV-2, which may result in relative protection against COVID-19. 25 Accordingly, our results could be interpreted as trained immunity induced by previous infections with seasonal coronaviruses. It could therefore make sense to determine the pre-existing immunity to seasonal coronaviruses before vaccinating against SARS-CoV-2, to assess, whether a low response to a vaccine simply reflects a general constitution of the patient as a coronavirus low-responder.

DATA AVAILABILITY STATEMENT
The data are not publicly available due to privacy and ethical restrictions.