Multiplex Detection of Antibody Landscapes to SARS-CoV-2/Influenza/Common Human Coronaviruses Following Vaccination or Infection with SARS-CoV-2 and Influenza

Abstract Background SARS-CoV-2 and influenza viruses continue to co-circulate, representing two major public health threats from respiratory infections with similar clinical presentations. SARS-CoV-2 and influenza vaccines can also now be co-administered. However, data on antibody responses to SARS-CoV-2 and influenza co-infection, and vaccine co-administration remains limited. Methods We developed a 41-plex antibody immunity assay that can simultaneously characterize antibody landscapes to SARS-CoV-2/influenza/common human coronaviruses. We analyzed sera from 840 individuals (11-93 years), including sera from reverse transcription polymerase chain reaction (RT-PCR) confirmed SARS-CoV-2 positive (n = 218) and negative (n = 120) cases, paired sera from SARS-CoV-2 vaccination (n = 29) and infection (n = 11), and paired sera from influenza vaccination (n = 56) and RT-PCR confirmed influenza infection (n = 158) cases. Lastly, we analyzed sera collected from 377 individual that exhibited acute respiratory illness (ARI) in 2020. Results This 41-plex assay has high sensitivity and specificity in detecting SARS-CoV-2 infections. It differentiated SARS-CoV-2 vaccination (antibody responses only to spike protein) from infection (antibody responses to both spike and nucleoprotein). No cross-reactive antibodies were detected to SARS-CoV-2 from influenza vaccination and infection, and vice versa, suggesting no interaction between SARS-CoV-2 and influenza antibody responses. However, cross-reactive antibodies were detected between spike proteins of SARS-CoV-2 and common human coronaviruses that were removed by serum adsorption. Among 377 individual who exhibited ARI in 2020, 129 were influenza positive, none had serological evidence of SARS-CoV-2/influenza co-infections. Conclusions Multiplex detection of antibody landscapes can provide in-depth analysis of the antibody protective immunity to SARS-CoV-2 in the context of other respiratory viruses including influenza.

The current unprecedented coronavirus disease 2019  global pandemic is caused by 2 severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) [1] . As of March 2022, more than 450 3 million cases and 6 million deaths have been reported worldwide [2]. Assay sensitivity and specificity for each antigen target were measured based on Pan Ig antibody 2 responses. Various cut-off values were analyzed, the cutoff values with the highest j-index ( [22] were used 3 as the positivity threshold (Supplementary Table 1). 4 Comparison of antibody responses were analyzed using two-tailed t tests. Statistical analyses 5 including Pearson correlation were performed using GraphPad Prism 8. 6

RESULTS 7
Sensitivity and specificity of MISHADA assay in detecting SARS-CoV-2 antibodies from RT-PCR-8 confirmed SARS-CoV-2 infections. 9 We developed a 41-plex MISHADA assay that can simultaneously measure antibody landscapes to 10 SARS-CoV-2, influenza, and common human coronavirus antigens using less than 10µl of sera. The serum 11 dilution at 1:500 is well within the wide dynamic linear range [18,23] for most antigens for pan Ig, IgG, IgA detected between 41-plex and 1-plex was less than 20 %. 16 To determine the sensitivity and specificity of MISHADA, ELISA and sVNT assays in detecting 17 antibodies from RT-PCR confirmed SARS-CoV-2 infection, we analyzed convalescent sera collected from 18 218 RT-PCR confirmed SARS-CoV-2 positive patients and baseline sera from 120 RT-PCR confirmed 19 SARS-CoV-2 negative persons (Table 1 and Supplementary Figure 3). In the MISHADA assay, when using 20 both SARS-CoV-2-S-RBD and SARS-CoV-2-S-Ec targets, it has 93.6% sensitivity and 98.3% specificity; 21 when using SARS-CoV-2-N protein alone, it has 93.1% sensitivity and 95.0% specificity, whereas when 22 A C C E P T E D M A N U S C R I P T using all 3 antigen targets (SARS-CoV-2-S-RBD, SARS-CoV-2-S-Ec and SARS-CoV-2-N), it achieved 1 sensitivity of 93.0% and specificity of 99.0% (Table 3). Antibody landscape shifts to SARS-CoV-2/influenza/common human coronavirus antigens following 7

Antibody landscape shifts following influenza infection and investigation of serological evidence of 18 influenza/SARS-CoV-2 co-infection in persons with acute respiratory illness (ARI). 19
During December 2019 and March 2020, the US Flu VE network collected sera and nasal swabs 20 from 377 outpatient participants exhibiting ARI [8]. Sera and nasal swabs were collected within 7 days from 21 symptom onset. Nasal swabs were tested for influenza infection by RT-PCR, patients that were RT-PCR positive for influenza also provided convalescent sera (Table 1). Although all participants exhibited 23 respiratory illness, RT-PCR for SARS-CoV-2 was not performed at the time. To investigate whether there 24 were SARS-CoV-2/influenza co-infection among these participants with ARI during the early stage of the 1 COVID-19 pandemic in US, we analyzed these sera by the MISHADA assay.  HAs from all A(H1N1) viruses tested including A/South Carolina/1/1918, as well as N1 and influenza A NP 8 ( Figure 4B). We also plotted MFIs from those influenza negative cases (n=248) in the antibody landscapes. 9 Interestingly, antibodies to HAs of B and A(H1N1)pdm09 viruses in influenza negative persons (S1) were 10 significantly higher than those in S1 sera from either influenza B or A(H1N1) infected persons (p<0.05, 11 Figure 4A-B), suggesting high HA antibodies may be associated with protection from influenza infections. 12 No cross-reactivity to SARS-CoV-2 or common human coronavirus antigens were detected from influenza 13 A(H1N1) or B infections. 14 Due to the low A(H3N2) activity in 2019-20, none of the participants were positive for influenza 15 A(H3N2), we therefore also analyzed paired sera from A(H3N2) infected adults in 2018-19 influenza season 16 (n=29). A(H3N2) infection induced antibody rises in MFIs to HAs from all A(H3N2) strains between 1968 17 and 2017, N2 and influenza A NP ( Figure 4C). Furthermore, Influenza A(H3N2) infection did not induce any 18 cross-reactive antibody rises to any antigens from SARS-CoV-2 and common human coronaviruses. 19 Lastly, to investigate whether there were SARS-CoV-2 infection among the 377 participants with 20 ARI, we identified those that had MFI antibody values to SARS-CoV-2 antigens above the thresholds 21 defined in Table 3. These sera were collected before any SARS-CoV-2 vaccines were available, therefore 22 elevated antibody levels to either SARS-CoV-2 spike protein or nucleoprotein could be indicative of 23 infection. Many participants had high pre-existing antibodies to common human coronavirus spike proteins 24 Figure 5A). When using SARS-CoV-2 spike antigen targets, 49 participants were positive to either SARS-1 CoV-2-S-RBD or SARS-CoV-2-S-Ec, 3 were positive to both components of the spike protein ( Figure 5B). 2 However, none was positive for both spike and nucleoprotein targets. To further verify the positivity, we 3 then analyzed the 49 sera by both ELISA and sVNT assays ( Figure 5C). By ELISA, 10 persons were 4 positive to either SARS-CoV-2-S-RBD (n=1) or SARS-CoV-2-S-Ec (n=9), but none was positive to both. 5 None of the 49 persons was positive in the sVNT assay ( Figure 5C). 6 Cross-reactive antibodies to SARS-CoV-2 spike protein can be removed by serum adsorption with 7 spike proteins from common human coronaviruses 8 To further elucidate the nature of the positive signals detected to SARS-CoV-2 spike proteins, we 9 first performed serum adsorption using cocktails of antigens (Table 2) with ten sera that had positive ELISA 10 titers (Table 5C). Following adsorption with spike proteins from 4 common human coronaviruses (H-CoV-S-11

Ec-Ads), ELISA titers to SARS-CoV-2-S-Ec and MFIs to both spike protein SARS-CoV-2-S-Ec and SARS-12
CoV-2-S2 were reduced to baseline levels ( Figure 6A-B), suggesting these positives were likely due to cross-13 reactive responses from past exposures to common human coronaviruses, rather than from SARS-CoV-2 14

infection. Pearson correlation analysis showed MFIs between SARS-CoV-2-S-Ec, SARS-CoV-2-S2 and H-15
CoV-S-Ec correlated well (r: 0.45-0.80) ( Figure 7). As positive controls, we then performed serum 16 adsorption of convalescent sera from 10 SARS-CoV-2 infected participants. High ELISA titers to SARS-17 CoV-2-S-RBD in these sera were completely removed by adsorption with SARS-CoV-2-S-RBD, but not by 18 the cocktail of 4 spike proteins from common human coronaviruses (H-CoV-S-Ec-Ads) suggesting authentic 19 antibody responses to SARS-CoV-2 infection. Moreover, adsorption with a cocktail of 8 influenza HA 20 proteins (Flu-Ads) only reduced MFIs to influenza antigens ( Figure 6D), but not to any SARS-CoV-2 21 antigens ( Figure 6C-D), confirming no cross-reactivity between influenza and SARS-CoV-2 virus antigens. 22 This 41-plex MISHADA assay is a powerful tool that can provide in-depth analysis of antibody 1 responses to multiple antigens that contribute to the protective immunity of SARS-CoV-2 and influenza. 2 The assay can identify SARS-CoV-2, influenza, common human coronavirus infections and co-infections, 3 and differentiate infection from vaccination. 4 Humans have complex immunity to influenza. The antibody immune profile of an individual is 5 often shaped by the initial priming to influenza viruses in childhood, and subsequent exposure to influenza 6 through vaccination and infection later in life. The back boost effect to antigenic-related viruses from the 7 current influenza vaccination/infection is evident in the antibody landscape analysis. For influenza, the 8 antibody landscape of an individual can impact one's susceptibility to infection, and immune response to 9 vaccination [24,25]. Compared to influenza, SARS-CoV-2 viruses were able to spread rapidly across the 10 globe and caused an unprecedented global pandemic, in part, due to naïve population immunity. The 11 seropositivity against SARS-CoV-2 was low (1.0%-6.9%) among the US population in early 2020 during the 12 early stage of the pandemic [26]. Since then, more than 450 million COVID-19 cases have been confirmed antibody profiles in the population to SARS-CoV-2 are becoming more complex [28]. Similar to influenza, 17 antibody landscapes tailored to SARS-CoV-2 antigens may be needed to anticipate population susceptibility 18 to emerging SARS-CoV-2 variants and to inform future vaccination strategies. 19 Using the antibody landscape analysis and serum adsorption, we demonstrated the presence of pre-20 existing, cross-reactive antibodies between the spike proteins of SARS-CoV-2 and common human 21 coronaviruses. Vaccination and infection with SARS-CoV-2 also induced antibody rises to spike proteins 22 from common human coronaviruses, mostly from betacoronaviruses (OC43 and HKU1), which are more 23 closely related to SARS-CoV-2 than alphacoronaviruses (NL63 and 229E) [29]. Others also reported cross-24 reactive antibodies against spike proteins of common human coronaviruses following SARS-CoV-2 infection 25 [30-33], which could be due to immunological imprinting, or shared epitopes. It was hypothesized that high 26 levels of antibodies against common human coronaviruses in children may have contributed to the mild 1 symptoms often observed in this age group [34,35]. Nonetheless, the protective potential of cross-reactive 2 antibodies against SARS-CoV-2 infection is not well understood [36,37]. Moreover, cross-reactive 3 antibodies will also complicate the interpretation of serologic results, thus multiple antigen detection may be 4 necessary to fully assess the antibody immunity to SARS-CoV-2 virus. 5 SARS-CoV-2 and influenza viruses continue to co-circulate, posing a threat to public health. Studies 6 have reported that co-infection with influenza may enhance the SAS-CoV-2 infectivity and disease severity 7 [38] [39,40]. Among the participants who exhibited respiratory illness during the early stage of the 8 pandemic, we did not identify SARS-CoV-2/influenza co-infections. In participants vaccinated or infected 9 with SARS-CoV-2 or influenza, our analysis indicated that there were no interactions or cross-reactivity of 10 antibody responses between these two respiratory viruses. 11 Our study has limitations: first, we were not able to obtain SARS-CoV-2/influenza coinfection sera 12 in the current analysis, in part, due to the low influenza circulation since the onset of the COVID-19 13 pandemic; further studies are warranted. Second, given the timeframe of the sera collection, we only 14 included SARS-CoV-2 antigens from Wuhan-Hu-1 virus. Additional antigens from Delta, Omicron and 15 future emerging variants, can be included in the further analysis of SARS-CoV-2 antibody landscapes. 16 In summary, the multiplex detection of antibody landscapes against SARS-CoV-17 2/influenza/common human coronaviruses is a high throughput tool to investigate the antibody responses to 18 these respiratory pathogens. Our results demonstrated no cross-reactivity between influenza and SARS-19 CoV-2 antibodies following infection and vaccination by either virus, providing scientific evidence to 20 support the co-administration of SARS-CoV-2 and influenza vaccination [6]. As the COVID-19 pandemic is 21 progressing through yet another flu season, it is important to gain better understanding of the humoral 22 protective immunity to SARS-CoV-2 in the context of other respiratory illness, especially influenza and 23 common human coronaviruses, to identify the most effective public health strategies for the control and 24 prevention of these respiratory pathogens.

Acknowledgments 2
This work was funded by the Centers for Disease Control and Prevention (CDC). In Pittsburgh, efforts 3 were funded by CDC (U01-IP1035) and by the National Institutes of Health (UL1TR001857). We thank 4 Makeda Kay from influenza division, CDC for managing the serum samples. We thank CDC's 2019 novel 5 coronavirus response lab task force for providing the reference serum panels. Lastly, we thank the 6 investigators from US Flu VE network and HEROS network sites for providing parts of the study serum 7 samples. 8

Disclaimer: 9
The findings and conclusions in this report are those of the authors and do not necessarily represent the 10 official positions of CDC. 11

Sera panels with known SARS-CoV-2 or influenza status
Reference sera panels with RT-PCR confirmed SARS-CoV-2 status for sensitivity and specificity analysis   Table 3 Sensitivity and specificity

27
Sera were collected from 377 persons with ARI, including 248 persons who were RT-PCR negative for influenza (S1) and 44 RT-PCR  Table 3. C. Positivity in ELISA and sVNT assays: sera from 49 31 persons that were positive for SARS-CoV-2-S-RBD and/or SARS-CoV-2-S-Ec in pan Ig MISHADA were tested by ELISA and sVNT.

32
Positivity was determined by the cutoff values in Table 3.