key: cord-0685519-zhbim9ek
authors: Li, Fang Fang; Liu, Aaron; Gibbs, Ebrima; Tanunliong, Guadalein; Marquez, Ana Citlali; Gantt, Soren; Frykman, Hans; Krajden, Mel; Morshed, Muhammed; Prystajecky, Natalie A.; Cashman, Neil; Sekirov, Inna; Jassem, Dr. Agatha N.
title: A novel multiplex electrochemiluminescent immunoassay for detection and quantification of anti-SARS-CoV-2 IgG and anti-seasonal endemic human coronavirus IgG
date: 2021-12-01
journal: J Clin Virol
DOI: 10.1016/j.jcv.2021.105050
sha: e7e23e92b83577fdf76ca3b6a20e569a0c32340e
doc_id: 685519
cord_uid: zhbim9ek

BACKGROUND: Multiplex immunoassays capture a comprehensive profile of the humoral response against SARS-CoV-2 and human endemic coronaviruses. We validated a multiplex panel (V-PLEX Panel 2) from Meso Scale Diagnostics targeting antibodies against nine coronavirus antigens. Performance was compared against alternative single- and multi-antigen immunoassays. METHODS: Sera collected for clinical or public health testing from 2018-2020 (n=135) were used to compare all tested platforms, and inter-test agreement was assessed by Cohen's kappa coefficient. Sample category (positive/negative) was assigned based on collection date relative to the index case in Canada, and SARS-CoV-2 PCR and serology results. 117 out of the 135 samples (31 positive, 86 negative) were assigned a category and were used to calculate sensitivity and specificity, with MSD's test results based upon manufacturer-set cut-offs. RESULTS: We observed SARS-CoV-2 target sensitivities of 100% and specificities >94% for all SARS-CoV-2 antigens (RBD, Nucleocapsid, Spike) in V-PLEX Panel 2. When targets are combined, we found a SARS-CoV-2 sensitivity of 100% and specificity of 98.8% with no difference in performance compared to clinical assays, and Cohen's kappa ranging from 0.798-0.945 compared to surface plasmon resonance imaging (SPRi). Quantitative measurements of antibodies against the Spike protein of endemic human coronaviruses were concordant with SPRi. CONCLUSION: Meso Scale Diagnostics’ V-PLEX Coronavirus Panel 2 allows for highly sensitive and specific detection of anti-coronavirus IgG, and is concordant with other serological assays for detection of antibodies against SARS-CoV-2 and the endemic human coronaviruses, making it a good tool for humoral response characterization after both infection and vaccination.

and Spike, nucleocapsid (NC), and S1 receptor binding domain (RBD) of SARS-CoV-2. The 132 assay was performed per manufacturer's instructions as previously described [6] . 133 Samples diluted 1:5000 were added to the plate, and bound antibodies were then labelled 134 with SULFO-TAG TM Anti-human IgG Antibody. Plates were read using MSD QuickPlex SQ120. 135

Raw data was processed using MSD's Discovery Workbench version 4.0 and imported into R to 136 interpret signal cut-off values. Quantification was reported in Arbitrary Units/mL (AU/mL). As 137 per manufacturer guidelines, cut-off values were: Spike, 1960 AU/mL; NC, 5000 AU/mL; and 138 S1 RBD, 538 AU/mL. To be considered serologically positive for SARS-CoV-2 via natural 139 infection using this assay's interpretive algorithm, samples must be positive for targets as 140 described in Figure 1 . Table 1 ) were run in duplicate on each plate. To assess assay precision and inter-assay variability, 144 readings from 15 independent runs were used to calculate the coefficient of variation (CV). To 145 further assess the precision around the manufacturer-recommended cut-offs, we re-analyzed 146 three serum specimens previously found to be close to the cut-off from our validation panel. We 147 first identified potential candidate sera lying within a pre-determined range (Spike: 1300-2300 148 AU/mL, RBD: 500-600 AU/mL, NC: 4500-5500 AU/mL), and then anonymized the container 149 identifiers before randomly selecting three to assay. To note, ranges were expanded in 150 increments of 100 AU/mL to ensure a minimum of six candidates per target. Readings from four 151 independent runs were used to calculate CVs (Supplementary Table 2) . Polyhistidine-tagged SARS-CoV-2 RBD and Spike of Beta-HCoVs OC43 and HKU1 were 157 captured and immobilized onto a sensor chip and sera diluted 1:10 in dilution buffer (3 mM 158 HEPES buffered saline with EDTA, 0,05% P-20, 0.2 mg/mL BSA) was sequentially injected 159 over the sensor surface, with binding events recorded as a function of time to generate a 160 "sensorgram". Mass changes on the surface of the chip due to antibody binding were translated 161 into response units (RUs), with 1 RU equivalent to 1 pg/mL of bound antibody. Antibody 162 responses were compared to a positivity cut-off defined as the mean value plus 3 standard 163 9 deviations of five pre-pandemic healthy control sera run on each assay, and S1 RBD values 164 above 10 RU were considered positive. Diagnostic performance is shown for both MSD's individual antigen targets and the three 182 HC-approved chemiluminescent assays for clinical diagnosis in Table 1 . SARS-CoV-2 antigens 183 RBD, S, and NC on the MSD assay were able to detect the presence of anti-SARS-CoV-2 IgG 184 with 100% clinical sensitivity and clinical specificities of > 94%. To note, no statistically 185 significant differences were seen in IgG detection between the targets on MSD's panel and when 186 compared to HC-approved chemiluminescent assays across all parameters. Based on serology 187 controls run in duplicate alongside the panel, inter-assay variation for NC and S was found to be 188 between 6-15% while RBD ranged from 24-30% (Supplementary Table 1) . 189 Table 1 . Diagnostic performance for each described assay using the 117 serum specimens from 190 the validation panel. vaccination setting, as anti-NC IgG is known to wane faster than anti-Spike/RBD IgG [11] [12] [13] . 202

Thus, we sought to design an algorithm able to differentiate between a recent positive response 203 ("Recent") from a vaccine-induced/remote-infection response ("Vaccine/Remote") based on 204 positivity in anti-Spike/RBD and anti-NC antibodies as described in Figure 1 . We assessed the performance of the algorithm in SARS-CoV-2 diagnosis using the same 212 117 serum specimens as before with results summarized in Table 2 . To note, all serum was 213 collected prior to the start of immunization programs in Canada from patients diagnosed with 214 SARS-CoV-2 infection within three months pre-collection, with no anticipated waning in 215 humoral response. Thus, only "Recent" was defined as positive for the purpose of assessing the 216 diagnostic performance of the algorithm. 217 

HCoV infection is relatively short and self-resolving, and there also exists no current gold 241 standard to discern true positives from true negatives. Furthermore, due to the endemic nature of 242

HCoVs, it is estimated that all individuals 6 years or older would have acquired immunity 243 against one of the four endemic HCoVs [14, 15] , making it difficult to validate an assay's ability 244 to discern true positives from true negatives with sufficient sample sizes. Thus, we validated the 245 detection of anti-HCoV IgG on the MSD assay by comparing the results to literature and against 246

SPRi, which can detect antibodies against Beta-HCoVs OC43 and HKU1. 247

The 117 serum specimens with defined SARS-CoV-2 categories were used for this With vaccine rollout ramping up worldwide, seroprevalence of anti-SARS-CoV-2 S is 307 expected to increase. As many of the vaccines generate immunity via Spike, NC may serve as a 308 marker for natural infection. Here, we have proposed an interpretive algorithm using MSD's V- 

Antibody Immunological 333 Imprinting on COVID-19 Patients

Antibody-dependent enhancement (ADE) of SARS-338

CoV-2 infection in recovered COVID-19 patients: studies based on cellular and structural 339 biology analysis

Viral epitope profiling of COVID-19 patients reveals cross-347 reactivity and correlates of severity

Seasonal human 355 coronavirus antibodies are boosted upon SARS-CoV-2 infection but not associated with 356 protection

A rapid and quantitative serum test for

SARS-CoV-2 antibodies with portable surface plasmon resonance sensing

Safety and immunogenicity of 366 the ChAdOx1 nCoV-19 vaccine against SARS-CoV-2: a preliminary report of a phase 1/2, 367 single-blind, randomised controlled trial

Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine

Efficacy 380 and Safety of the mRNA-1273 SARS-CoV-2 Vaccine

Oxford-AstraZeneca COVID-19 vaccine efficacy

Interim Results of a Phase 1-2a Trial of Ad26.COV2.S Covid-19 Vaccine

CoV-2 Anti-Spike Antibodies More Than Double the Half-Life of Anti-nucleocapsid 392

Antibodies in Healthcare Workers

Immunological 397 memory to SARS-CoV-2 assessed for up to 8 months after infection, Science (80-. )

The 403 duration, dynamics and determinants of SARS-CoV-2 antibody responses in individual 404 healthcare workers

Clinical Presentations of the Four Human Coronaviruses 229E, HKU1, NL63, and OC43

Detected over 3 Years Using a Novel Multiplex Real-Time PCR Method

OC43 strain-related coronavirus antibodies in 410 different age groups

S Protein-Reactive IgG and Memory B Cell Production after Human 414

SARS-CoV-2 Infection Includes Broad Reactivity to the S2 Subunit

Cross-reactive antibody against human coronavirus OC43 spike protein 419 correlates with disease severity in COVID-19 patients: a retrospective study

Evaluation of a novel multiplexed assay for determining IgG 423 levels and functional activity to SARS-CoV-2

A majority of uninfected adults show pre-existing 429 antibody reactivity against SARS-CoV-2