key: cord-0259785-bl1wbo5l authors: Solastie, A.; Virta, C.; Haveri, A.; Ekström, N.; Kantele, A.; Miettinen, S.; Lempainen, J.; Jalkanen, P.; Kakkola, L.; Dub, T.; Julkunen, I.; Melin, M. title: A highly sensitive and specific SARS-CoV-2 spike- and nucleoprotein-based fluorescent multiplex immunoassay (FMIA) to measure IgG, IgA and IgM class antibodies date: 2021-07-30 journal: nan DOI: 10.1101/2021.07.28.21260990 sha: e3c3b3181417451d4628f8a6503f7cca9538f377 doc_id: 259785 cord_uid: bl1wbo5l Background. Validation and standardization of accurate serological assays are crucial for the surveillance of the coronavirus disease 2019 (COVID-19) pandemic and population immunity. Methods. We describe the analytical and clinical performance of an in-house fluorescent multiplex immunoassay (FMIA) for simultaneous quantification of antibodies against the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) nucleoprotein and spike glycoprotein. Furthermore, we calibrated IgG-FMIA against World Health Organisation (WHO) International Standard and compared FMIA results to an in-house enzyme immunoassay (EIA) and a microneutralisation test (MNT). We also compared the MNT results of two laboratories. Results. IgG-FMIA displayed 100% specificity and sensitivity for samples collected 13-150 days post-onset of symptoms (DPO). For IgA- and IgM-FMIA 100% specificity and sensitivity were obtained for a shorter time window (13-36 and 13-28 DPO for IgA- and IgM-FMIA, respectively). FMIA and EIA results displayed moderate to strong correlation, but FMIA was overall more specific and sensitive. IgG-FMIA identified 100% of samples with neutralising antibodies (NAbs). Anti-spike IgG concentrations correlated strongly ({rho}=0.77-0.84, P<2.2x10-16) with NAb titers. The NAb titers of the two laboratories displayed a very strong correlation ({rho}=0.95, P<2.2x10-16). Discussion. Our results indicate good correlation and concordance of antibody concentrations measured with different types of in-house SARS-CoV-2 antibody assays. Calibration against WHO international standard did not, however, improve the comparability of FMIA and EIA results. keywords: SARS-CoV-2, COVID-19, antibody, immunoassay, nucleoprotein, spike glycoprotein, WHO international standard, neutralising antibodies, microneutralisation, receptor-binding domain Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative agent of coronavirus 46 disease 2019 (COVID-19), had claimed over 4 million lives and infected 190 million by late July 2021 47 [1] . A large proportion of all infections may go undetected in their acute phase [2,3] for various reasons, 48 such as lack of symptoms [4, 5] or hesitancy of getting tested [6] . Therefore, accurate serological assays 49 are needed to provide more reliable estimates of COVID-19 prevalence. 50 SARS-CoV-2 serological assays are useful anywhere between determining seroprevalence 51 in the general population to the investigation of confined outbreaks. In neutralisation tests and other 52 serological tests alike, the clinical specificity and sensitivity of the assays can vary considerably [7] [8] [9] [10] [11] [12] . 53 Comparisons of methods and their standardisation is urgently needed to properly understand and apply 54 the vast acquired information on the immunity induced by SARS-CoV-2 infection and COVID-19 55 vaccines. 56 We describe the validation and performance of an in-house fluorescent multiplex 57 immunoassay (FMIA) developed for quantification of antibodies produced against three SARS-CoV-2 58 antigens: the full-length spike glycoprotein (SFL), spike receptor-binding domain (RBD) and 59 nucleoprotein (N). Our assay detects antibodies against these three antigens simultaneously, which 60 enables differentiation of natural SARS-CoV-2 infection from vaccine-induced immunity. The assay is 61 based on an FMIA previously described by Trivedi et al. [13] . We have recently reported the performance 62 of the FMIA for measuring IgG antibodies against SARS-CoV-2 nucleoprotein [14] . Here, we report our 63 findings on the analytical and clinical performance of FMIA and compare its IgG, IgA and IgM assay 64 results to another laboratory's in-house enzyme immunoassay (EIA) [15, 16] , both calibrated for IgG with 65 WHO International Standard [17] . Furthermore, we compare FMIA antibody levels to neutralising 66 . 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 July 30, 2021 71 We assessed the analytical and clinical performance of the FMIA by using negative and positive sample SARS-CoV-2. Samples were collected as part of a COVID-19 household transmission study [19] . We 76 used the negative and positive serum panels also to compare FMIA and MNT results. We excluded one 77 sample due to limited sample volume leaving the total sample size to n=548. Details of the panels are 78 presented in Table S1 and Figure S1 . 79 We compared FMIA with EIA using a serum panel that consisted of a subset of 80 samples 80 collected as part of a COVID-19 mRNA vaccine response study [16] . We collected the samples from 81 convalescent-phase PCR-confirmed COVID-19 patients (n=20) and volunteer health care workers 82 (n=20) who received two doses of Pfizer-BioNTech BNT162b2 mRNA vaccine. We recruited the 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 July 30, 2021. ; https://doi.org/10.1101/2021.07.28.21260990 doi: medRxiv preprint 5 collection, three weeks (median 27 days, range 18-29) had also passed from the 2 nd vaccine dose. 89 Participant demographics are presented in Table S1 . The SARS-CoV-2 EIA used in this study has previously been described in detail [15] . We measured IgG, 135 IgA, IgM and total Ig antibodies against SARS-CoV-2 S1 and N proteins from sera diluted 1:300. SARS-136 . 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 July 30, 2021. Analytical performance of the FMIA 199 We calculated the LOQ and LOD separately for each antigen and each antibody class and the data is 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 July 30, 2021. ; https://doi.org/10.1101/2021.07.28.21260990 doi: medRxiv preprint The mean intra-assay variation ranged from 8 to 10% and inter-assay variation from 4 to 208 12% in the different antibody classes (Table S3 ). The mean variation between different technicians was 209 20% for IgG, 21% for IgA and 18% for IgM assays. The variation between four different batches of 210 conjugated microspheres was 16% and the variation between different batches of detection antibodies 211 was 15% for IgA and IgG assays. Overall, both intra-and inter-assay variation were found acceptable 212 for all antibody classes and antigens. SFL to obtain calibrated BAU/ml. In EIA, the calibration factor for N IgG was 12.9, for N total Ig 17.4, 220 for S1 IgG 11.8 and for S1 total Ig assay 12.0. We did not calibrate MNTs but obtained a titer of 192 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 July 30, 2021. this study (n=147) the specificity and sensitivity of the N-IgG FMIA we have previously described [14] 235 was 100% for samples collected 21-51 DPO but the sensitivity of the assay decreased to 98% for samples 236 collected 52-150 DPO (Figure 2 ). The optimal threshold for IgA-FMIA was also based on the simultaneous detection of both (Table 1) . DPO range resulting in 100% sensitivity was 13-28 days for IgM [95% CI for sensitivity: 91.0-100, 248 n=39] ( Table 1) . As DPO increased, the sensitivity of the IgM assay decreased steeply (Figure 2) . 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 July 30, 2021. ; https://doi.org/10.1101/2021.07.28.21260990 doi: medRxiv preprint samples were negative in MNT but positive for S-IgG in FMIA (Table S4) (Table 259 S4, Figure 2) . 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 July 30, 2021. ; https://doi.org/10.1101/2021.07.28.21260990 doi: medRxiv preprint 14 nucleoprotein. Our results indicate that both methods accurately recognise vaccine-induced antibody 302 responses at 6 weeks, but that FMIA is slightly more sensitive than EIA in samples taken 3 weeks after 303 1 st vaccine dose. As we compared the MNTs of two laboratories, we found that NAb titers had a very 304 strong correlation despite some differences in the methodology. Overall, the MNT of one of the 305 laboratories was somewhat more sensitive, which is likely explained by less diluted sera and different 306 viral strains (lineage B vs. B.1). Limitations of this study are related to the sample material used in the validation process of 308 FMIA. In essence, the thresholds and DPOs described here were optimised for the detection of previous 309 infections with mild to moderate symptoms. Despite this, the S-IgG levels correlated strongly with NAb 310 titers. Our results also suggest that FMIA is better at recognising samples with NAbs than many 311 commercial assays [12] . Whilst S-IgG FMIA identified all samples with NAbs, it does not necessarily 312 measure antibodies against neutralising epitopes only, as some samples negative in MNT were 313 considered positive for S-IgG. We conclude that FMIA is valuable in pre-screening of serum samples 314 prior to confirmatory MNT. As the anti-SARS-CoV-2 antibody levels have been found to correlate with 315 the disease severity [23-25], we can expect FMIA to perform well in the serological diagnostics of also 316 previous severe infections. Importantly, while the present data show an excellent sensitivity for IgG-317 FMIA until 150 DPO, its performance after that remains to be investigated. Estimates of the persistence of immunity to COVID-19 appear to depend on the serological 319 assay used [11, 26] . We recently reported that six and twelve months after SARS-CoV-2 infection 98% 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 July 30, 2021. ; https://doi.org/10.1101/2021.07.28.21260990 doi: medRxiv preprint DPO. Together with data of persisting antibodies, our findings imply, that S-IgG FMIA continues to be 326 reliable in the detection of antibodies produced against SARS-CoV-2 for at least several months after 327 infection. Notably, the participants providing convalescent serum samples for the positive serum CoV-2 infection. The need for highly sensitive and specific antibody tests will continue in the future as new 337 variants emerge fuelling further COVID-19 waves. Here we have described a reliable antibody assay for 338 the simultaneous quantification of multiple SARS-CoV-2 antibodies, with excellent clinical performance 339 and good comparability with NAb titers. We have also described the calibration of FMIA against a WHO 340 international standard, which has been reported to reduce inter-laboratory variation notably [31] . 341 However, we observed that calibrating FMIA and EIA only emphasised the differences between their 342 results. In FMIA the results are calculated from a standard curve and in EIA from the ratio between a 343 sample and two controls. Although standardisation of SARS-CoV-2 antibody assays is urgently needed, 344 our results indicate only a limited value for the international standard in calibrating two assays whose 345 test principles are quite different. The applicability and value brought by calibration should be considered 346 when used in comparisons between methodologically diverse assays. 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 July 30, 2021. ; https://doi.org/10.1101/2021.07.28.21260990 doi: medRxiv preprint COVID-19) Dashboard Comparison of seroprevalence of SARS-CoV-2 375 infections with cumulative and imputed COVID-19 cases: Systematic review Undiagnosed SARS-CoV-2 seropositivity 378 during the first 6 months of the COVID-19 pandemic in the United States Estimating the effectiveness of routine asymptomatic 381 PCR testing at different frequencies for the detection of SARS-CoV-2 infections Estimating asymptomatic, undetected and total cases for the COVID-19 384 outbreak in Wuhan: a mathematical modeling study Analytic comparison between three high-throughput 388 commercial SARS-CoV-2 antibody assays reveals minor discrepancies in a high-incidence 389 population Comparison of six commercially available SARS-391 CoV-2 antibody assays-Choice of assay depends on intended use Evaluation of SARS-CoV-2 serology assays reveals a 394 range of test performance Persistence and decay of human antibody responses to 396 the receptor binding domain of SARS-CoV-2 spike protein in COVID-19 patients Persistence of Anti-SARS-CoV-2 Antibodies Depends on 399 the Analytical Kit: A Report for Up to 10 Months after Performance of six SARS-CoV-2 402 immunoassays in comparison with microneutralisation 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 Development and Evaluation of a Multiplexed 404 Immunoassay for Simultaneous Detection of Serum IgG Antibodies to Six Human Coronaviruses Analytical and clinical evaluation of antibody tests for 407 SARS-CoV-2 serosurveillance studies used in Finland in 2020 A combination of N and S antigens with IgA and IgG 410 measurement strengthens the accuracy of SARS-CoV-2 serodiagnostics COVID-19 mRNA vaccine induced antibody 413 responses against three SARS-CoV-2 variants Reference Panel for anti-SARS-CoV-2 antibody Serological and molecular findings during SARS-CoV-2 417 infection: the first case study in Finland High secondary attack rate and persistence of SARS-CoV-2 420 antibodies in household transmission study participants Act on the National Institute for Health and Welfare 668 Antibody pharmacokinetics and pharmacodynamics Disease severity dictates SARS-CoV-2-specific neutralizing antibody 428 responses in COVID-19 Clinical and immunological assessment of asymptomatic 430 SARS-CoV-2 infections Defining the features and duration of antibody responses to 432 SARS-CoV-2 infection associated with disease severity and outcome Antibody Responses 8 Months after Asymptomatic or Mild 434 SARS-CoV-2 Infection Persistence of neutralizing antibodies a year after SARS-436 CoV-2 infection International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity SARS-CoV-2-directed antibodies persist for more than 438 six months in a cohort with mild to moderate COVID-19 Persistence of Antibody and Cellular Immune Responses in 440 Patients Over Nine Months After Infection Changes in SARS-CoV-2 Spike versus Nucleoprotein 443 Antibody Responses Impact the Estimates of Infections in Population-Based Seroprevalence 444 WHO International Standard for anti-SARS-CoV-2 446 immunoglobulin. The Lancet 449 antibody classes due to exclusion of samples that did not meet DPO of optimal sensitivity criteria. . 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. 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 July 30, 2021. ; https://doi.org/10.1101/2021.07.28.21260990 doi: medRxiv preprint