Introduction
In early December 2019, a novel zoonotic coronavirus (CoV) caused a cluster of pneumonia cases in Wuhan, China [
1]. Since then, the virus has spread globally and caused a pandemic with over 34.8 million confirmed infections and over 1 million fatalities (as of Oct 4, 2020) [
2]. Due to its phylogenetic similarity to the severe acute respiratory syndrome related coronavirus (SARS-CoV-1), the novel CoV was named SARS-CoV-2 [
3]. The acute respiratory disease induced by SARS-CoV-2 is called coronavirus disease 19 (COVID-19).
The identification of acutely infected individuals by the detection of viral RNA by real-time PCR [
4] was implemented rapidly in the health care of most countries. While this method is highly valuable for the diagnosis of acute COVID-19 cases, specific serological methods are urgently needed to determine seroconversion in general and more specifically to characterize the humoral response against SARS-CoV-2. Robust, validated serological approaches are essential to track transmission events in individuals that have already cleared the infection especially after mild or symptom-free disease. With increasing numbers of immune individuals, serological tests will also help to understand epidemiological aspects of the pandemic and to employ SARS-CoV-2 immune staff in critical frontline positions at hospitals or nursing homes. In addition, validated serological methods are essential to evaluate novel vaccine candidates in clinical studies.
Together with the 2003 SARS-CoV-1 and the 2012 Middle East respiratory syndrome coronavirus (MERS-CoV) epidemic, the SARS-CoV-2 pandemic represents the third betacoronavirus in 20 years that crossed the species barrier and resulted in a significant number of human infections. At the same time, four other CoVs are endemic in the human population (two alphacoronaviruses: CoV-NL63 and -229E, two betacoronaviruses: CoV-OC43 and -HKU1) that cause episodes of common cold in humans in all parts of the world [
5]. CoVs are enveloped single-stranded RNA viruses that contain four structural proteins: membrane (M), envelope (E), spike (S), and nucleocapsid (N). From SARS-CoV-1, it is known that N and S proteins are the most immunogenic viral antigens, while only S-specific antibodies can mediate virus neutralization [
6,
7]. Therefore, N- and S-specific antibody responses should be first-choice parameters for a sensitive serology [
8]. However, depending on the study cohort, up to 90% of the population is seropositive for common cold CoVs [
9‐
11]. Thus, a careful validation of the assay specificity is required in CoV serology.
Here, we describe a novel flow cytometric assay to determine SARS-CoV-2 spike protein-specific antibodies in serum samples. The virus-free assay relies on reagents and devices that are available in many medical and biological research labs and therefore can be easily adopted in a decentral manner without the need for commercial kits or products that are prone to shortage.
Materials and methods
Serum samples
Anonymized, random sera (
n = 180) were selected from the sample repository of the diagnostics department of the Institute for Clinical and Molecular Virology at the University Hospital Erlangen to evaluate the specificity of the novel diagnostic test. Samples were collected until August 2019 (further denominated as pre-COVID-19 era) and no longer needed for diagnostic purposes and assigned for disposal. Those specimens were not characterized in regard to anti-HCoV antibody status. Twenty-one sera from eight patients with PCR-confirmed endemic HCoV infections were additionally included. These samples were collected at least 1 week before and 2 to 4 weeks after HCoV infection. These include 4x HKU-1, 2x 229E, 1x NL63, and 1x OC43 infections. Post-infection sera were sampled twice from some patients (Table
2). Additionally, 116 specimens from 53 individuals with a PCR-confirmed SARS-CoV-2 infection (some sampled longitudinally) were obtained. The majority is derived from a newly established biobank for COVID-19 patients at the University Hospital Erlangen. The data are collected in accordance with ethical requirements (ethics committee UK Erlangen, license number AZ. 174_20 B). Five out of 116 were derived from plasma donors after (patients’ informed consents; approved by local ethics committee of the FAU; AZ. 2020, 49_20B). Another set of sera was collected from thirteen COVID-19 patients at the Hospital Nürnberg Nord at different time points after the PCR confirmation (Table
3). All sera were sampled for recent diagnostic purpose and have been tested for seroconversion in the EuroImmun ELISA at the Institute of Clinical Hygiene, Medical Microbiology and Infectiology, Paracelsus Medical University, Hospital Nürnberg, Germany. All clinical specimens were used in anonymous form for retrospective analyses.
DNA plasmids
The pCG1_CoV_2019-S plasmid encoding the codon-optimized sequence of the SARS-CoV-2S protein was generated as described elsewhere [
12]. The plasmid pcDNA3.1 (Invitrogen) was used in the mock transfection control. Blue fluorescent protein (BFP)– and red fluorescent protein–encoding (dsRed; from
Discosoma sp.) plasmids were used as marker proteins for transfected 293T cells.
Flow cytometric antibody assay
Human embryonic kidney cells (HEK 293T cells; ECACC 12022001) were maintained in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Cat #11960-044) containing 10% fetal calf serum (Capricorn Scientific, Cat #FBS-12A), 1% GlutaMAX (Gibco, Cat #35050-038), and 1% Penicillin/Streptomycin (Gibco, Cat #15140-122) at 37 °C and 5% CO2. For the assay, 1.12 × 107 cells were plated out (25-ml medium; 175 cm2 cell culture flask) and, 12–24 h later, were transfected with 30 μg pCG1_CoV_2019-S plus 15 μg fluorescent protein (BFP) by standard polyethylenimine transfection (3.5 ml DMEM, 67.5 μg polyethylenimine). As an internal control, a mock transfection was used with 30 μg pcDNA3.1 and 15 μg fluorescent protein (dsRed). Forty-eight hours after the transfection, cells were harvested, resuspended in freeze medium (75% FCS, 10% DMSO, 3% Glucose in DMEM), and stored in 1-ml aliquots of 1 × 107 cells at − 80 °C.
For the assay, aliquots of cells were thawed, washed once with PBS, and then resuspended in FACS buffer (PBS with 0.5% bovine serum albumin and 1 nmol sodium azide). 0.5 × 105 cells of each of the two cell preparations (S- and mock-transfected) were seeded out per sample in a 96-well U-bottom plate. Serial dilutions of the standards or serum samples (1:100) were diluted in 100 μl FACS buffer and given on the cells (30 min, 4 °C). 100 μl FACS buffer was added, cells were centrifuged (500×g, 4 °C, 3 min; used for all following centrifugation steps), washed two times with 180 μl FACS buffer, and bound antibodies were stained with secondary detection antibodies diluted 1:300 in 100 μl FACS buffer (30 min, 4 °C, anti-IgG-AF647, clone HP6017, Biolegend, Cat #409320; anti-IgM-BV711, clone MHM-88, Biolegend, Cat #314540). One hundred microliters of PBS was added, cells were centrifuged, washed two times with 180 μl PBS, and fixed in 200 μl 2% paraformaldehyde in PBS (15 min, 4 °C). Cells were centrifuged and washed once in 180 μl FACS buffer, before resuspended in 200 μl FACS buffer for flow cytometric analysis. Data were acquired on a BD LSRII or Thermo Fisher Attune Nxt cytometer and analysis was performed with FlowJo (Tree Star Inc.) or Flowlogic (Inivai Technologies).
ACE-2-Fc standard
A PCR fragment containing the sequence coding for the extracellular domain of human ACE-2 lacking the secretory signal peptide (NM_021804.3, nucleotides 358–2520) fused at the 3′ end with a PCR fragment coding for the Fc-part of human IgG1 and a C-terminal myc/his tag was cloned into the expression vector pCEP4 (Thermo Fisher Scientific). The signal peptide of the murine IgG kappa-chain V-J2 was used instead of the ACE-2 signal peptide. The synthetic intron from pIRES (IVS, Takara Bio) was cloned via NheI restriction sites between the transcription start and the translation start site. Expression and purification of the Fc-fusion protein was done as described before [
13]. Briefly, HEK 293T cells were transfected by calcium phosphate method and kept in culture for 6 days. Cell culture supernatant was then harvested and cell debris removed by centrifugation. The pH of the supernatant was adjusted to 8.0 with NaOH and sterile filtered. The supernatant was then applied to a HiTrap Protein A HP column (GE Healthcare Life Sciences). ACE-2-Fc fusion protein was eluted by a pH step gradient using 0.1 M citrate buffer. ACE-2 Fc fusion protein eluted at pH 4.0 and the pH was immediately neutralized by the addition of 1 M Tris buffer (pH 9).
As an external standard for IgG quantitation, a twofold dilution series starting with 10 μg/ml of ACE-2-Fc was measured in the flow cytometric assay as described above. With this standard, we quantified the amount of ACE-2-binding equivalents in a plasma sample available in larger volume. Adjusting for molecular weight differences between ACE2-Fc and IgG, the anti-SARS-CoV-2S IgG concentration in this plasma sample was determined.
Enzyme-linked immunosorbent assay
Commercially available ELISA for the detection of anti-SARS-CoV-2 IgG (anti-S1-specific, EuroImmun, Cat #EI 2606-9601 G) and IgA (EuroImmun, Cat #EI 2606-9601 A) were performed according to the manufacturer’s protocols. Sera were diluted 1:101 (10 μl sample + 1000 μl sample buffer) and the optical density was detected at 450 nm at a multilabel plate reader (Victor X5, Perkin Elmer). A cut-off for a positive result was according to the manufacturer defined as a ratio of > 1.1 between the specific specimen and the calibrator. Values between 0.8 and 1.1 were defined as “borderline.” The specificity and sensitivity of these assays provided by the manufacturer are listed in Table
1.
Table 1
Specificities and sensitivities of commercial antibody tests used in this study (provided by manufacturers)
EuroImmun | IgA ELISA | 98.3% | 96.9% | |
EuroImmun | IgG ELISA | 99.6% | 94.4% | |
Shenzhen Yhlo Biotech | IgM CLIA | 99.2% | 86.1% | |
Shenzhen Yhlo Biotech | IgG CLIA | 96.3% | 97.3% | |
Chemiluminescent immunoassay
Commercially available magnetic bead–based CLIA for the detection of IgG (N- and S-specific, Shenzhen Yhlo Biotech, iFlash-SARS-CoV-2, Cat #C86095G) and IgM (Shenzhen Yhlo Biotech, iFlash-SARS-CoV-2, Cat #C86095M) was performed on a fully automated iFlash immunoassay analyzer (Shenzhen Yhlo Biotech). The assays were performed according to the manufacturer’s protocols. The IgG and IgM titers were automatically calculated as arbitrary units (AU/ml) and the cut-off value for a positive test was 10 AU/ml. The specificity and sensitivity of these assays provided by the manufacturer are listed in Table
1.
Discussion
While rigorous measures led to a partial control of the recent SARS-CoV-2 pandemic in some countries, validated serological assays are needed to consolidate those achievements and to support the transition to a post-peak phase. This includes for example diagnostic measures for late/post-infection stages, COVID-19 contact tracing, the assessment of epidemiological aspects, and the evaluation of immunity after infection or in potential vaccine trials. In the recent study, we validated an in-house flow cytometric assay for the detection of SARS-CoV-2S-specific IgM and IgG using sera from PCR-confirmed COVID-19 cases and a collection of control serum samples. In regard to specificity and sensitivity, our flow cytometric assay showed a comparable or even better performance compared to commercial CE-marked serological assays (EuroImmun ELISA and Shenzhen Yhlo CLIA) (Table
1).
Detection of viral nucleic acids via real-time PCR is the gold standard in the diagnosis of acute SARS-CoV-2 infections. However, despite its reliability early during infection, confirmation of an infection at later time points becomes less reliable. As early as 8 days post-infection, the diagnostic value of serological assays might therefore outperform nucleic acid–based methods [
20,
21]. Indeed, also our study showed seroconversion in a longitudinal set of sera from one patient 8 days after the first positive PCR test, although the exact infection date is not clearly defined. As reported before [
20,
22,
23], IgM did not generally possess a higher clinical sensitivity compared to IgG, since most of the IgM+ specimen tested in the present study were positive for both isotypes. Detection of SARS-CoV-2-specific IgA was previously reported as more sensitive than detection of IgG in the EuroImmun ELISA kits [
14]. However, while this held also true in our study, IgG and IgM measured by the flow cytometric assay were similarly sensitive compared to the IgA ELISA. This higher sensitivity to detect S-specific IgG might be due to the different viral antigens used in the assays. Our flow cytometric assay exploits full-length S protein in its natural conformation and with the respective post-translational modifications due to the expression in mammalian cells. This enables detection of the full spectrum of S-specific antibodies directed against conformational epitopes and glycosylated sites as well, some immunogenic sites possibly missing in truncated, recombinant S1-only proteins as used in the EuroImmun ELISA.
A potential downside of using full-length S for serological testing might be the detection of cross-reactive antibodies induced by other HCoV. Along this line, some assays detect only antibodies directed against the S1 subunit (like the EuroImmun ELISA) or the receptor-binding domain in order to increase specificity [
14,
15]. However, in a collection of sera from individuals that suffered from an infection with an endemic HCoV shortly before blood collection, none was tested positive for SARS-CoV-2 antibodies. In additional 180 specimens sampled before the COVID-19 outbreak, two sera were found to be reactive in the flow cytometric assay. Since endemic HCoV seroprevalence is high in the general population [
11] and those two individuals were non-reactive in the commercial S1-specific ELISA, a plausible explanation is cross-reaction of antibodies induced by the endemic HCoVs with the S2 subunit of SARS-CoV-2. Although the reactivity of the two specimens needs to be classified as false-positive detection of SARS-CoV-2 antibodies, these cross-reactive antibodies might possess antiviral activity against COVID-19 and the analysis of cross-protection due to these responses might be an interesting topic for further investigations.
Regarding the clinical sensitivity, the flow cytometric serology assay detected 100% of IgM- and IgG-positive samples measured by either of the two commercial assays. Only in cases where blood samples were taken at the same day as PCR sampling, all assays (ELISA, CLIA, cytometry) were negative, probably reflecting acute infections prior to development of detectable antibody responses. The lower analytical detection limit of the flow cytometric assay is consistent with its excellent clinical sensitivity.
Since in the early phase of such a pandemic, there are naturally a limited number of reliable serological test kits available, and there is a high need to expand the portfolio of serology techniques which can be rapidly applied and scaled up. The flow cytometry–based technique to detect SARS-CoV-2 seroconversion presented here fulfills fundamental criteria in regard to sensitivity, specificity, and robustness. The basic requirements needed like cell culture, plastic ware, and a flow cytometer are available in many standard diagnostic and biomedical research labs. Although performing the assay and analyzing the primary data need some trained personal, high-throughput solutions of this method can increase serology testing capacities significantly without competing for ELISA/CLIA kits. Given the large number of antibody assays reaching the market without clearly defined analytical sensitivities, using recombinant ACE-2 Fc protein for standardization is a potential strategy for cross-assay comparisons. Moreover, the quantification of S-specific antibody responses might help to define protective antibody levels as correlate of protective immunity.
In conclusion, our in-house flow cytometry–based serological assay has good specificity and sensitivity for the detection of antibodies to SARS-CoV-2. In addition to the nucleotide sequence of the antigen, only readily available reagents were needed to establish the assay. Therefore, the flow cytometric assay may also serve as a blueprint for rapid-response antibody tests against other emerging viral infections.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.