key: cord-0923418-maxwqgma authors: Hunt, D. T.; Yates, J. L.; Kulas, K. E.; Carson, K.; Lamson, T.; Demarest, V.; Furuya, A.; Howard, K.; Marchewka, M.; Stone, R.; Tucker, H.; Warszycki, C.; Yee, J.; Yang, H. S.; Racine-Brzostek, S.; Zhao, Z.; Ejemel, M.; Li, Q.; Wang, Y.; Fernando, S.; La Carpia, F.; Hod, E. A.; McDonough, K. A.; Lee, W. T. title: COVID-19 Serology in New York State Using a Multiplex Microsphere Immunoassay date: 2021-05-15 journal: nan DOI: 10.1101/2021.05.12.21257125 sha: fa340af6abdcf0e4c8844a2186ce950dcc3d67a1 doc_id: 923418 cord_uid: maxwqgma The emergence of SARS-CoV-2, leading to COVID-19, necessitated the development of new molecular and serological tests. Here, we describe a multiplexed serological assay developed as the global pandemic moved into New York State in the spring of 2020. The original microsphere immunoassay used a target antigen from the SARS-CoV-1 virus responsible for the 2003 SARS outbreak, but evolved to incorporate multiple SARS-CoV-2 protein antigens (nucleocapsid, spike and spike domains, spike and nucleocapsid proteins from seasonal human coronaviruses). Besides being highly versatile due to multiplex capabilities, the assay was highly specific and sensitive and adaptable to measuring both total antibodies and antibody isotypes. While determining the assay performance characteristics, we were able to identify antibody production patterns (e.g., kinetics of isotypes, individual variations) for total antibodies and individual antibody classes. Overall, the results provide insights into the laboratory response to new serology needs, and how the evolution and fine-tuning of a serology assay helped contribute to a better understanding of the antibody response to SARS-CoV-2. The coronavirus disease 2019 (COVID-19) (1) , caused by the Severe Acute Respiratory 58 Syndrome Coronavirus 2 (SARS-CoV-2)(2), is historically one of the most widespread 59 and devastating pandemics with severe health, economic, and social consequences (3, (Table 1) . Including sera containing Abs to known pathogens or 134 autoantigens did not reduce the specificity (Table 1) . For clinical test reporting, the 135 specimens with MFI values that fell between 3 SD and 6 SD of the mean MFI of the 136 normal serum specimens were listed as "Indeterminate" ( Figure 2C ). In the second-137 generation MIA, that included the SARS-CoV-2 RBD, when the two target antigens gave 138 different results, the overall result was based on the most positive of either antigen (e.g., 139 indeterminate plus reactive was reported as reactive). Subsequent analyses of the pre- to us was 25 days, in which only 88% of the specimens had N-specific antibodies detected 154 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 15, 2021. ; https://doi.org/10.1101/2021.05.12.21257125 doi: medRxiv preprint by the MIA above the 6SD cutoff (EUA: https://www.fda.gov/media/137541/download). testing of specimens with >30 days post symptom onset, showed that the assay could 157 detect a much higher percentage of positive specimens. For example, an analysis of 158 8659 COVID-19 convalescent sera (collected 30-50 days post symptom onset) showed 159 that the assay had a 92.2% (N antigen-reactive) sensitivity using the 6 SD cutoff, 160 compared to 95.9% using 3 SD (Table 3 ). This study also showed the impact of adding 161 additional antigens, such as the RBD, into the MIA. Comparison of separate reactivity to 162 the N and the RBD proteins within the same specimens showed that, while most sera 163 contained antibodies reactive with both antigens, several had reactivity only to the N 164 protein (4.2%, Table 3 ). Reactivity to RBD protein without reactivity to N did occur but at 165 a lower frequency (<1.2% of reactive specimens) (Table 3) . When examining individual 166 specimens, although more often a high N reactivity coincided with high RBD reactivity, 167 many specimens had high Ab levels to one antigen but lower levels to the other (data not 168 shown). Our recent studies showed that by adding the complete spike antigen, or by 169 replacing the RBD with the spike antigen, the sensitivity of the MIA could be even further 170 increased (Supplementary Table 1 ). However, either with just the RBD or with the spike 171 antigen, the MIA performs comparably to other COVID-19 serology assays. In a study 172 looking at 500 sera, including 100 pre-COVID sera, the MIA had similar or better 173 sensitivity and specificity as three highly used commercial laboratory tests performed on 174 the Ortho Vitros, Biomerieux Vidas, or Abbot Architect platforms (Supplementary Table 175 1). . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 15, 2021. ; https://doi.org/10.1101/2021.05.12.21257125 doi: medRxiv preprint measured total antibody, the assay proved readily adaptable to both greater multiplexing 178 of antigens and to the measurement of individual Ab isotypes. For example, the study in 179 Supplementary Figure 1 shows the multiplexing capability with simultaneous 180 measurement of Abs to the N and spike antigens of the four seasonal coronaviruses and 181 the same antigens of SARS-CoV-2 (10-plex). The study shown in Supplementary Figure 4A demonstrates that there is a wide range in individual production of total Ab 187 made to both the N and RBD antigens and shows that, even early on, some individuals 188 were highly reactive in the MIA. We more deeply investigated the onset of Ab production 189 by determining the kinetics and distribution of antibody classes. We assessed sera from 190 individuals at different times from onset of COVID-19 symptoms using the isotype-specific 191 version of the MIA while examining the responses to multiple SARS-CoV-2 antigens. For 192 these studies, we not only looked at reactivity to the RBD of the spike antigen, but also 193 the entire spike and its two major domains: the amino-terminal S1 subunit of the spike, 194 which contains the RBD, and the carboxy-terminal S2 subunit. Although most of the anti-195 spike Abs appear to be directed toward S1 in most individuals, the S2 is also 196 immunogenic ( Figure 4B and Supplementary Figure 2) . The higher reactivity towards the 197 S1 domain, as compared to the RBD, suggests that of the S1-reactive Abs, the RBD 198 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 15, 2021. appears to be one, but not the only, focus of reactivity as measured by this assay. There is clearly individual variation, with some specimens having copious amounts of 208 antibodies and early production of antibodies. However, overall there was a progressive 209 increase of total antibodies that did not peak until after 40 days post symptom onset 210 ( Figure 3 ). In contrast with canonical kinetics of secreted Ab isotypes, we found that IgM 211 did not significantly precede the production of IgG; indeed, a higher percentage of IgG 212 positive specimens were found at the earlier times after symptom onset, as compared to 213 IgM ( Figure 5 ). The rapid rise of IgG can be more clearly seen when looking at serial 214 specimens from the same individual ( Figure 6 ). We also examined the production of IgG 215 subclasses. We found a robust IgG1 response to all the target antigens that was induced . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 15, 2021. Multiplexing the N and spike antigens may also be useful for distinguishing natural 287 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 329 Further, for the convalescent plasma screening effort, we currently are aligning the MIA 330 Index values with virus neutralization values to better identify "high titer" neutralizing 331 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 15, 2021. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 15, 2021. ; https://doi.org/10.1101/2021.05.12.21257125 doi: medRxiv preprint 4. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 15, 2021. ; https://doi.org/10.1101/2021.05.12.21257125 doi: medRxiv preprint below. Recombinant SARS-1 nucleocapsid was produced at the Wadsworth Center as 381 described below. For some experiments, SARS CoV-2 RBD was also a kind gift from Dr. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 15, 2021. ; https://doi.org/10.1101/2021.05.12.21257125 doi: medRxiv preprint All graphs and statistical analyses to determine Spearman's rank correlation coefficients 444 were done using Prism 9.0 (Graphpad, San Diego, CA). Spearman's correlations were 445 calculated using all complete pairs of variables in the dataset. Associations with Age and Disease Severity. Cell. 2020. Sette A, Crotty S. Adaptive immunity to SARS-CoV-2 and COVID-19. Cell. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted May 15, 2021. ; https://doi.org/10.1101/2021.05.12.21257125 doi: medRxiv preprint 20. McMahan K, Yu J, Mercado NB, Loos C, Tostanoski LH, Chandrashekar A, et al. Indeterminate or (gray) Nonreactive. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) Equine Encephalitis, and Yellow Fever viruses), as well as sera with Antinuclear antibodies and Rheumatoid Factor. Shown are the numbers (percentages) of MIA negative/total specimens and the 95% Confidence Intervals. "Negative" means nonreactive to both target antigens in the specimen; "indeterminate" means indeterminant to one target antigen and indeterminant or non-reactive to the other target antigen in the specimen. For this analysis, indeterminates and negatives are counted together, using a threshold of 6 SD above the mean MFI of 92 normal sera to determine positivity. IgM IgG IgA Table 2 . Determination of MIA Sensitivity. A total of 93 serum specimens from individuals with SARS-CoV-2 infections, confirmed using RT-PCR, and reported symptom onset dates were assessed for total antibodies to the SARS-CoV-2 Nucleocapsid antigen and the RBD. A positive specimen is ≥6 SD above the cutoff for either one of the target antigens; an indeterminate is ≥3 SD above the cutoff for either or both of the target antigens; a negative is a sample with no reactivity to either target antigen. Sensitivity at different period post symptom onset is shown on the right as the number of MIA-positive specimens/total specimens tested. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) Using N and RBD**, Positive @6 SD (clinical*) 8219 94.9 Using N and RBD**, Positive @3 SD 8411 97.1 Table 3 . Reactivity to the SARS-COV-2 N and RBD antigens. 8659 serum specimens from convalescent COVID-19 donors who were at least 14 days removed from symptom onset were tested using the MIA for total antibody reactivity to the SARS-COV-2 and RBD antigens. A reactive specimen is ≥6 SD above the cutoff; an indeterminate is ≥3 SD above the cutoff. The bottom four rows indicate assay reported results calculated to show positivity at 3SD or 6 SD thresholds and shows the effect of using the single N target Ag versus both target Ags. * Threshold used for reporting "Reactive" specimens in diagnostic assays. **One or both target Ags reactive. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 15, 2021. ; https://doi.org/10.1101/2021.05.12.21257125 doi: medRxiv preprint A novel coronavirus outbreak of 470 global health concern A pneumonia 472 outbreak associated with a new coronavirus of probable bat origin The socio-475 economic implications of the coronavirus pandemic (COVID-19): A review Evaluation of SARS-CoV-2 neutralization assays for antibody monitoring in natural 510 infection and vaccine trials. medRxiv A serological assay to detect SARS-CoV-2 seroconversion in humans Neutralizing Antibody Responses in COVID-19 Convalescent Sera Serologic Cross-Reactivity of SARS-CoV-2 with Endemic and Seasonal 519 Acknowledgments.