key: cord-0894271-6uf9apo7
authors: Thomas, Archana; Messer, William B; Hansel, Donna E; Streblow, Daniel N; Kazmierczak, Steven C; Lyski, Zoe L; Lu, Zhengchun; Slifka, Mark K
title: Establishment of Monoclonal Antibody Standards for Quantitative Serological Diagnosis of SARS-CoV-2 in Low Incidence Settings
date: 2021-02-02
journal: Open Forum Infect Dis
DOI: 10.1093/ofid/ofab061
sha: 1660dcaf36a9997665046641b2dbacb96886b6ef
doc_id: 894271
cord_uid: 6uf9apo7

BACKGROUND: Serological confirmation of coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is critical for understanding the dynamics of the pandemic and determining seroprevalence rates within afflicted communities. Common challenges with SARS-CoV-2 serological assays include poor analytical specificity and sensitivity, and lack of a serological standard for quantitative assessment of antibody titers. METHODS: To overcome these obstacles, we developed a quantitative ELISA based on an optimized 2-dimensional screening assay that utilizes SARS-CoV-2 RBD (Receptor Binding Domain of spike protein) and SARS-CoV-2 spike S1 subunit. RESULTS: A total of 4 SARS-CoV-2-reactive monoclonal antibodies were evaluated for use as serum standards for calibrating assays performed on different days or by different laboratories. This approach provided quantitative analysis of hospitalized RT-PCR-confirmed COVID-19 cases that in some cases reached >100 μg/mL. The assay demonstrated 72% sensitivity based on time points ranging from 2-52 days post-symptom onset, with 100% sensitivity at time points measured ≥13 days post-symptom onset and 100% specificity. DISCUSSION: Using these optimized reagents and serological standards, we believe this approach will be useful for sensitive and specific determination of seroconversion rates and quantitatively measuring the durability of antiviral antibody responses following SARS-CoV-2 infection or vaccination.

A c c e p t e d M a n u s c r i p t Background COVID-19 has spread rapidly across the globe and as of January 20 th , 2020, there have been over 24 million cases reported in the U.S. and over 95 million confirmed cases reported worldwide (https://coronavirus.jhu.edu/map.html). During this same period, more than 403,000 COVID-19related deaths have been identified in the U.S. with nearly 2.1 million COVID-19-related deaths reported worldwide.

It has been just over a year since SARS-CoV-2 began spreading in the human population and COVID-19 seroepidemiology studies are mainly still in their infancy [1] [2] [3] [4] [5] [6] [7] . At the individual level, serology tests may allow retrospective diagnosis of prior infection and this may be particularly useful in cases where virological testing was not performed during the window of early infection or was not available at the time of exposure. If an immune correlate is eventually identified, then quantitative serology tests could also prove valuable for determining immune status. However, the use of serology tests to determine immune status has been viewed with caution in low incidence settings.

For example, if the true seroprevalence of a population is 5%, then an assay with 95% specificity would have a high false-positive rate resulting in only 50% positive predictive value (PPV) and would be unsuitable for clinical decision making or determining potential risk for infection [8] . Serologic assays may also be used to characterize COVID-19 convalescent donor plasma by providing a quantitative estimate of antibody titers prior to blood donation [8] . At the population level, seroepidemiology studies can be used to characterize transmission within COVID-19 clusters as well as during larger outbreaks, in addition to determining the extent of disease burden and/or immunity within a particular community at a given point in time. However, interpretation of many serology studies has been complicated by poor assay sensitivity and specificity, leading some authors to note that seroconversion rates may be driven by the serological test performance characteristics themselves [5] rather than correctly identifying true seropositive and seronegative individuals.

Moreover, because of ongoing SARS-CoV-2 transmission, serosurveys performed in the first few months of the pandemic are no longer representative of the current population. For instance, although an early study [5] identified a 6.9% seroconversion rate in New York City using samples collected March 23-April 1, 2020, another study conducted a few weeks later found seroconversion rates as high as 22.7% [4] . For these reasons, development of quantitative serological assays with high sensitivity and specificity will be important for future COVID-19 research.

A c c e p t e d M a n u s c r i p t

In the studies described here, we have optimized a SARS-CoV-2 serological assay that is based on antiviral antibodies binding to two antigen substrates [SARS-CoV-2 Receptor Binding Domain (RBD) and spike protein subunit S1] and provides both high sensitivity and specificity. We have also characterized 4 SARS-CoV-2-reactive monoclonal antibodies for use as ELISA assay standards. The use of a monoclonal IgG standard allows for the quantitative assessment of antiviral antibody levels over time and provides an approach for different laboratories to compare results across assay platforms by using the same readily available antibody reagents. Together, this work provides an important new tool for the assessment of humoral immunity following SARS-CoV-2 infection including longitudinal studies on immunological memory as well as serology studies to determine transmission dynamics, seroprevalence and estimated burden of disease.

Samples were obtained from RT-PCR-confirmed hospitalized COVID-19 patients who provided informed written consent. If subjects were unable to provide written consent, then written consent was obtained from their legally authorized representative (LAR). In some cases, de-identified samples were obtained from a biorepository that included hospitalized COVID-19 patients at OHSU who tested positive for SARS-CoV-2 by nasopharyngeal swab and subsequent RT-PCR. The 23 COVID-19 patients were an average of 64 years old (range, 20-88 years), 52.2% female, 30.4% required mechanical ventilation, and 8.7% required extracorporeal membrane oxygenation (ECMO). A total of 50 plasma samples from the COVID-19 patients were used in these experiments and pre-pandemic plasma or serum samples obtained between November, 1989 to August, 2019 from another 300 adults were included in the study as negative controls. The study was approved by the institutional review board of Oregon Health & Science University.

SARS-CoV-2 recombinant antigens used in the ELISA included the receptor binding domain (RBD, Cat #230-30162, Ray Biotech), S1 subunit of the spike protein (spike S1, Cat#40591-V08H, Sino The 96-well ELISA plates (Cat#9018, Corning) were coated with 100 L of each antigen at a concentration of 1 g/mL prepared in PBS and the plates were incubated overnight at 4 o C and then stored frozen at -20 o C until use. Plates were thawed at room temperature (RT) the coating antigen was removed, and plates were blocked for 1 hour at RT with 5% Omniblok (Cat#AB10109-01000, American Bio) prepared in PBST containing 0.05% Tween (i.e., dilution buffer). Plates were washed once with PBST containing 0.05% Tween (i.e., wash buffer) and for diagnostic screening, 50 L of dilution buffer was added to each well along with 50 L of a 1:50 dilution of heat-inactivated serum or plasma (1:100 dilution final). For quantitative analysis, samples were serially 3-fold diluted in dilution buffer. Plates were incubated at RT for 1 hour, followed by the addition of 50 L of 10% 

Populations that have low rates of SARS-CoV-2 infection can be problematic for seroepidemiology studies unless the diagnostic assay has sufficiently high specificity.

Bearing this in mind, we performed routine optimization of assay conditions (time, temperature, antigen concentration, blocking buffer and wash buffer formulations) and prepared an initial pre-pandemic serology panel from 100 individuals and tested this panel for antibodies to seasonal human coronaviruses (HCoVs) in addition to SARS-CoV-2 antigens ( Fig. 1) . Coating ELISA plates with purified full-length spike protein from each of the 4 seasonal human coronaviruses (HCoV-OC43, HCoV-HKU, HCoV-NL63, and HCoV-229E) revealed high seroconversion rates that ranged from 99% (HCoV-229E) to 100% (HCoV-A c c e p t e d M a n u s c r i p t OC43, HCoV-HKU, and HCoV-NL63) with high geometric mean antibody titers that ranged between 1,774 to 4,710 ELISA units (EU) (Fig. 1A) . In contrast, cross-reactive antibodies to the SARS-CoV-2 spike protein were low (geometric mean; 45 EU) with only 17% of the prepandemic control samples scoring above the detection threshold (200 EU). Narrowing the viral antigen from the 1,213 amino acid full-length SARS-CoV-2 spike protein to either the SARS-CoV-2 spike S1 subunit (685 amino acids) or to the smaller SARS-CoV-2 receptor binding domain (RBD; 222 amino acids) further reduced serological cross-reactivity from 17% cross-reactivity to 3% and 8%, respectively. Similarly, serological cross-reactivity to SARS-CoV-2 nucleocapsid protein (NP) was observed among 10% of the pre-pandemic samples. Together, these results indicate that pre-existing cross-reactive antibody responses to SARS-CoV-2 antigens are low or undetectable among people who have not been exposed to the virus and are consistent with prior studies showing a lack of cross-reactive neutralizing antibodies to SARS-CoV-2 [9] [10] [11] [12] . However, based on these results none of the individual SARS-CoV-2 antigens would provide sufficient specificity for seroepidemiology studies in low-incidence settings unless the detection threshold was raised from 200 EU to >1,000 EU in order to achieve 100% specificity. Although this type of approach would improve assay specificity it would also decrease assay sensitivity, which is a critical parameter for detecting low antiviral antibody titers in COVID-19 serology studies [13] . Since SARS-CoV-2 RBD is an important target for neutralizing antibodies and spike/RBD ELISA titers correlate well with neutralizing titers [7, [14] [15] [16] [17] , we focused on RBD seroreactivity for further diagnostic assay development. To improve specificity without reducing sensitivity, we tested a 2dimensional screening approach by comparing RBD vs. NP (Fig. 1B) , RBD vs. full-length spike (Fig. 1C) , or RBD vs. spike S1 (Fig. 1 D) . Using a detection threshold of 200 EU for each antigen, we found that this technique greatly improved the assay with results that reached 99% specificity for RBD vs. NP, 99% specificity for RBD vs. full-length spike, and 100% specificity when using RBD vs. spike S1 (Fig. 1D) . Overall, this indicates that a combination screening assay based on RBD and spike S1 provided the best diagnostic approach for these studies.

There is currently no international serum standard approved for SARS-CoV-2. This can be problematic for making direct comparisons of antibody titers published by individual research groups because each laboratory may obtain disparate antibody titers based on the use of different reagents and assay conditions. Use of an antibody assay standard not only makes it easier to normalize results between experiments performed on different days or performed A c c e p t e d M a n u s c r i p t by different operators within the same laboratory, but this also provides an important bridge for comparing results between independent research groups. As a step towards the development of a serological SARS-CoV-2 standard for quantitation of antiviral antibody titers, we compared 4 human SARS-CoV-2-reactive IgG monoclonal antibodies with respect to their binding profiles to RBD and spike S1 (Fig. 2) . The monoclonal antibodies, ABMX-002, Sanyou, and DA0002 are derived from SARS-CoV-2 infection whereas CR3022 is a SARS-CoV-1-specific clone that cross-reacts with SARS-CoV-2.

Three of these monoclonals had similar RBD-binding curves ( Fig. 2A) well whereas the CR3022 clone demonstrated nearly 2-fold lower binding to spike S1 compared to RBD. This indicates that if the CR3022 monoclonal antibody is used as an ELISA standard, then the experimentally derived antibody titers for SARS-CoV-2 samples will differ markedly (approximately 2-fold) depending on whether the RBD or the spike S1

antigen is used in the assay.

Based on the antibody binding characteristics described in Fig. 2A and 2B , the choice of monoclonal antibody used as the ELISA standard was anticipated to impact the calculated magnitude of antiviral antibody titers. To address this question, RBD-specific antibody levels were calculated from 8 COVID-19 cases based on each monoclonal standard (Fig.   2C ). As expected, use of ABMX-002, Sanyou, or CR3022 monoclonal antibodies provided similar results. This indicates that these monoclonal antibodies may be used interchangeably for quantitating SARS-CoV-2-specific antibody titers as long as the RBD antigen is used in the ELISA. In contrast, when the DA0002 monoclonal antibody was used as the standard, the calculated antibody titers were much higher. For example, when SARS-CoV-2-specific antibody titers from Subject #8 were measured, the antiviral IgG levels were determined to be approximately 145 g/mL when using ABMX-002, Sanyou, or CR3022 but were estimated A c c e p t e d M a n u s c r i p t at 331 g/mL when calibrated against the DA0002 monoclonal antibody. Together, this indicates that the DA0002 monoclonal standard resulted in inflated antibody scores that were approximately 2-to 2.5-fold higher than those obtained with the other three monoclonal standards and in terms of providing conservative estimates of antibody titer, we recommend ABMX-002, Sanyou, or CR3022 as a serological standard when using SARS-CoV-2 RBD to quantitate antiviral antibody levels. If SARS-CoV-2 spike S1 is used in the ELISA, then either ABMX-002 or Sanyou are recommended to quantitate antiviral antibody levels.

Having optimized ELISA protocol parameters to detect as little as 0.5 ng/mL of purified SARS-CoV-2-specific monoclonal IgG (Fig. 2) , we next determined the sensitivity and specificity of this diagnostic approach based on 50 samples from hospitalized RT-PCRconfirmed COVID-19 cases as well as 300 pre-pandemic negative control samples obtained prior to SARS-CoV-2 circulation in the U.S. (Fig. 3) . Samples were screened against SARS-CoV-2 RBD and spike S1 in a 2-dimensional plot (Fig. 3A) . A diagnostic detection threshold set at 120 ng/mL allows detection of low-level antiviral IgG while still maintaining 100% specificity (95% CI: 98.7-100%, Fig. 3B) . Although a small number of pre-pandemic samples scored above the detection threshold for either RBD (Y-axis) or spike S1 (X-axis) alone, none of the 300 negative controls scored positive on both antigens. Screening samples obtained between 2-52 days after symptom onset resulted in 36/50 samples scoring positive against both RBD and spike S1 antigens for an overall sensitivity of 72% (95% CI: 58-83%, Fig. 3B ). Antiviral antibodies are most likely to reach detectable levels by ~2 weeks after symptom onset [14] and for samples obtained at 13 days, we observed 100% sensitivity (95% CI: 89-100%, Fig. 3B ).

Quantitative analysis of SARS-CoV-2-specific antibodies were measured as a function of time post-symptom onset (Fig. 4) . These studies indicated that antibody titers increased rapidly during the first 2-3 weeks and reached titers that, in two of the severe hospitalized cases (requiring mechanical ventilation or ECMO for respiratory support), exceeded 100 g/mL within 4 weeks after symptom onset. Antiviral antibody titers remained stable out to 52 days after symptom onset, indicating that antiviral immunity may be long-lived and more long-term longitudinal studies are currently in progress. Together, these results indicate that we have developed a sensitive and specific diagnostic approach for determining SARS-CoV-2 serostatus with the ability to quantitatively measure antiviral IgG titers over time.

A c c e p t e d M a n u s c r i p t

In these studies, we developed a robust 2-dimensional diagnostic SARS-CoV-2 ELISA assay that at 13 days or more post-symptom onset, provided up to 100% sensitivity as well as 100% specificity for serological detection of SARS-CoV-2 infection. High sensitivity was attained by optimizing the reagents and assay conditions to detect virus-specific monoclonal IgG at concentrations as low as 0.5 ng/mL. High specificity was attained by requiring that samples score above the detection threshold against not one, but two closely related SARS-CoV-2 antigens (RBD and spike S1). Inclusion of a human monoclonal antibody standard in the assay is essential for providing a quantitative assessment of antiviral antibody titers and this not only allows future SARS-CoV-2 serological studies by independent groups to be directly compared using the same calibrated unit values, but may eventually provide a framework for quantitation of an immunological correlate of immunity if one is eventually established for COVID-19.

High assay sensitivity is important for detecting early SARS-CoV-2-specific antibody responses shortly after infection as well as at later time points when antibody levels have declined from their initial peaks. Furthermore, accurate antibody measurements will be critical following the introduction of vaccines in order to determine response rates and confirm antibody production. One limitation is that we measured antibody responses only from hospitalized COVID-19 cases and further studies will be needed to determine assay sensitivity at different time points after mild or asymptomatic SARS-CoV-2 infection. Another limitation of our study is that we focused primarily on IgG measurements because we had IgG monoclonal antibodies available for direct reference/standardization/quantitation but we did not measure IgM or IgA levels since we did not have monoclonal antibodies that matched these specific immunoglobulin isotypes. However, recent studies have shown nearly simultaneous seroconversion for IgG, IgM, and IgA isotypes with the median seroconversion date for each isotype occurring around 11-12 days after symptom onset [14] .

This suggests that the timing of antiviral IgG production after symptom onset does not lag appreciably behind IgM or IgA and so the impact on overall assay sensitivity in minimized. In those studies [14] , predictive accuracy for identifying SARS-CoV-2 cases by serology using a high IgG detection threshold of 570 ng/mL and indicated that SARS-CoV-2 RBD-specific IgG could be detected among 7% of cases at <7 days, 51% of cases between 8-14 days, and 95% of cases at >15 days postsymptom onset [14] . Our IgG detection threshold is 120 ng/mL and using this approach we identified 31% seroconversion at <7 days, 44% seroconversion between 8-14 days and 100% seroconversion at 13 days post-symptom onset. The differences between these two studies may A c c e p t e d M a n u s c r i p t be due in part to sample size or differences in COVID-19 disease severity but we believe that the improved assay sensitivity observed at both the early and late time periods is likely to be attributable to having nearly a 5-fold lower limit of detection obtained through the use of the 2dimensional ELISA screening format (Fig. 1C, Fig. 3A) .

Over specificity [1] . However, a more recent study that analyzed 2,204 serum samples found what appears to be a much higher false-positive rate and substantial discordance with antibodies to RBD and spike S1 from the same samples and this has raised concerns for its continued use in the U.K. M a n u s c r i p t (A) ELISA titers from hospitalized RT-PCR+ COVID-19 cases (red symbols, n = 50) and pre-pandemic naïve controls (black symbols, n = 300) were assayed on SARS-CoV-2 RBD (Y axis) versus SARS-CoV-2 spike S1 (X axis) with samples that score >200 EU/120 ng/mL on each axis considered seropositive and samples that score <200 EU/120 ng/mL on one or both antigens considered seronegative. (B)

Diagnostic sensitivity was 72% for all samples obtained from days 2-52 after symptom onset and reached 100% sensitivity for detecting virus-specific antibodies at time points at 13 days after A c c e p t e d M a n u s c r i p t symptom onset, with 100% specificity based on 300 pre-pandemic negative controls. Error bars represent 95% confidence intervals. A c c e p t e d M a n u s c r i p t

Performance Characteristics of the Abbott Architect SARS-CoV-2 IgG Assay and Seroprevalence in

Seroprevalence of SARS-CoV-2-Specific Antibodies Among Adults

Estimated Community Seroprevalence of SARS-CoV-2 Antibodies -Two Georgia Counties

Cumulative incidence and diagnosis of SARS-CoV-2 infection in New York

Seroprevalence of Antibodies to SARS-CoV-2 in 10 Sites in the United States

Comparative assessment of multiple COVID-19 serological technologies supports continued evaluation of point-of-care lateral flow assays in hospital and community healthcare settings

Orthogonal SARS-CoV-2 Serological Assays Enable Surveillance of Low Prevalence Communities and Reveal Durable Humoral Immunity

Serodiagnostics for Severe Acute Respiratory Syndrome-Related Coronavirus-2: A Narrative Review

The RBD Of The Spike Protein Of SARS-Group Coronaviruses Is A Highly Specific Target Of SARS-CoV-2 Antibodies But Not Other Pathogenic Human and Animal Coronavirus Antibodies

The receptor binding domain of the viral spike protein is an immunodominant and highly specific target of antibodies in SARS-CoV-2 patients

A highly conserved cryptic epitope in the receptor binding domains of SARS-CoV-2 and SARS-CoV

Broad neutralization of SARS-related viruses by human monoclonal antibodies

Whole nucleocapsid protein of SARS-CoV-2 may cause false positive results in serological assays

Persistence and decay of human antibody responses to the receptor binding domain of SARS-CoV-2 spike protein in COVID-19 patients

Severe Acute Respiratory Syndrome Coronavirus 2-Specific Antibody Responses in Coronavirus Disease Patients

A serological assay to detect SARS-CoV-2 seroconversion in humans

Rapid Generation of Neutralizing Antibody Responses in COVID-19 Patients

The following reagent was obtained through BEI Resources, NIAID, NIH:Monoclonal Anti-SARS Coronavirus Recombinant Human IgG1, Clone CR3022 (produced in Nicotiana benthamiana), NR-52392.

M a n u s c r i p t A c c e p t e d M a n u s c r i p t