HEK293T and HEK293T-ACE2 cells (kindly provided by June Ereño-Orbea from CIC bioGUNE and Dr. Jean-Philippe Julien from The Hospital for Sick Children Research Institute) were cultured in DMEM (41966-029, Gibco). Media was supplemented with 10% FBS (10270106, Thermo Fisher) and 1% Penicillin–Streptomycin (15140122, Thermo Fisher).

Serum samples were divided in four cohorts, based on sample collection date: i) preCOVID cohort (30 serum samples, collected in 2018-2019), ii) COVID cohort (30 serum samples from PCR-positive individuals collected in 2020-2021), iii) pre-Omicron cohort (a representative subset of the population collected in late 2021, before the third dose was administered, n=30 serum samples) and iv) post-Omicron cohort (a representative subset of the population collected after the outbreak of Omicron and the administration of the third dose of the vaccine, n=30 serum samples). Serum samples corresponding to preCOVID and seropositive individuals were provided by the Basque Biobank (https://www.biobancovasco.org) after approval from the corresponding ethics committee (CEIC-E 20-26, 1-2016, CEIm-E PI+CES-BIOEF 2020-04 and PI219130). Some samples were obtained during the yearly medical check-up of the working population of the Basque Country in 2018-2019 in collaboration with Osarten Kooperatiba Elkartea from Mondragon Corporation (20). The COVID cohort corresponding to patients presenting COVID-19 symptomatology and diagnosed by PCR were obtained after written informed consent and approval by the Cantabria Ethics Committee (CEIm Code: 2020.167).

Custom-made Becton Dickinson (BD) functional cytometric bead array (CBA) beads were used in preparation of streptavidin beads as described in the Functional Bead Conjugation Buffer Set Instruction Manual (BD 558556). Briefly, 1 mg of SMCC-modified streptavidin per mL (in 50 mM sodium phosphate, 0.2 M sodium chloride, pH 7.4) was used for conjugation to CBA beads as described in the above protocol. The streptavidin beads were suspended in 1x PBS, 0.5% BSA, pH 7.2 before coating with biotinylated RBD proteins. Streptavidin microbeads were then coated with antigen as described before (18). Briefly, streptavidin coated microbeads were incubated with 11 μg/mL of biotinylated RBD proteins from the following SARS-CoV-2 variants: RBD Wuhan-1 (Acrobiosystems SPD-C82E9), RBD-B.1.351 (β, Acrobiosystems SPD-C82E5), RBD- B.1.617.1 (κ, Acrobiosystems SPD-C82Ec) and RBD-B.1.1.529 (ο, Acrobiosystems SPD-C82E4). Uncoated microbeads were used as negative controls. All microbeads were incubated with an excess of D-biotin to saturate the remaining free streptavidin sites.

Antigen-coupled microbeads were added to LoBind 1.5 mL Eppendorf tubes (Eppendorf 525-0133) in a volume of 50 μL of PBS (Gibco 10270-106) containing a total of 5000-6000 microbeads. After centrifugation (1000 xg for 5 min), microbeads were resuspended with 100 µL of diluted serum samples (1:5 in PBS, 20 µl of serum sample in 80 µl of PBS) or serially diluted commercially available antibodies against S1: clone AM122 (Acrobiosystems) and Imdevimab (Ibian technologies, SRBDC4-MAB1), starting from a concentration of 20 mg/mL. Negative control samples were prepared in PBS with 5% FBS (Gibco 10270-106). After a 30-minute incubation (4°C protected from light), samples were washed three times with PBS. Human recombinant ACE2 with murine IgG1 (rhACE2-mIgG1, recombinant fusion protein of human ACE2 ectodomain fused with murine IgG1 Fc produced in HEK293T eukaryotic cells, SinoBiological 10108-H05H) was added at 0.18 mg/mL (diluted in 100 µL of PBS containing 5% FBS) and incubated for 30 minutes at 4°C protected from light. Three washes were performed, and microbeads were incubated with an anti-mouse IgG1-BV421 antibody (1:200 in 100 µL of PBS supplemented with 5% FBS) for 30 mins at 4°C. A final wash was performed, and microbeads were resuspended in 200 µL of PBS for acquisition. At least 600 events for each type of microbead were acquired in a FACSymphony flow cytometer (BD Biosciences) and geometric mean fluorescence intensities (gMFI) were obtained. Results were analysed using FlowJo version 10 (BD Biosciences). The neutralizing percentage corresponding to each sample against each variant was measured with the following formula: % neutralizing = ([gMFI sample − gMFI max neutralization]/[gMFI max binding − gMFI max neutralization]) × 100). The value of gMFI maximum neutralization corresponds to the signal obtained with microbeads in the absence rhACE2. The value of gMFI max binding corresponds to the signal obtained in the absence of serum samples or blocking antibodies.

Pseudotyped lentiviral particles were generated by transfecting HEK 293T cells as described before (21). Briefly, 5 million HEK293T cells were seeded 24h before transfection in a P100 plate. Cells were transfected with plasmids encoding for the lentiviral backbone containing ZsGreen under the control of the CMV promoter (NR-52520, 5.79 µg), the SARS-CoV-2 spike protein (NR-52514, 1.97 µg), HDM-Hgpm2 (NR-52517, 1.27 µg), pRC-CMV-Rev1b (NR-52519, 1.27 µg) and HDM-tat1b (NR-52518, 1.27 µg), using the jetPEI kit (101-10N, Polyplus-transfection). 48 h after transfection, viruses were harvested from the supernatant, filtered through a 0.45 µm filter (514-0063, VWR) and mixed with lentiX (631232, Takara) in a 3:1 proportion at 4°C overnight. Then, viruses were centrifuged at 1500xG for 45 minutes and the pellet was resuspended in 200 µL of DMEM, aliquoted and stored at -80°C.

Viral titration was performed in HEK293T-ACE2 cells as described before (21). Then, 50,000 cells were seeded in 96-well plates, cultured overnight, and incubated with pseudotyped lentiviral particles (MOI: 0.34) for 24 h in the presence of a serial dilution of anti-RBD monoclonal antibodies or serum samples (1:5 dilution in culture media). HEK293T-ACE2 cells were harvested and Zsgreen positive cells were quantified by flow cytometry using a FACSymphony (BD Biosciences). Results were analyzed in Flowjo (BD Biosciences).

96-well ELISA plates (Nunc Maxisorp 44-2404-21) were coated overnight at 4°C with 50 μL of Wuhan-1 RBD (Acrobiosystems SPD-C52H3) (2 µg/mL in PBS). Then, the coating solution was removed, and plates were blocked with 3% non-fat milk in PBS supplemented with 0.1% Tween-20 (P1379-500ML, Sigma) (PBST) for 1 hour at RT. Serum samples were pre-diluted in at 1:50 in 1% non-fat milk in PBST and incubated for 2 hours at RT. After three washes with 250 μL of PBST in a plate washer (Biotek), an anti-human IgG-horseradish peroxidase (HRP)-conjugated secondary antibody (1:5000) (GenScript A01854) was added for 1 hour at RT. Three washes were performed, and after an incubation of 2 min with 100 μL of TMB substrate (Thermo Scientific 34021), the reaction was stopped with 50 μL of stop solution (Thermo Scientific N600). The optical density (OD) was measured at 450 nm in a VictorNivo multimode plate reader (PerkinElmer) and shown OD values were calculated by subtracting the negative control values from all samples.

Statistical analyses were performed with an unpaired two-tailed Student’s t-test. For correlation analyses, Pearson correlation coefficients (r) and their corresponding p-values were calculated. Data were analyzed using Prism 8 (GraphPad). Bland–Altman plots were generated with Graphpad to compare the agreement between nC19BA and the cell-based pseudotyped lentiviral particle assay for the determination of the blocking activity of antibodies.

The nC19BA is a flow cytometry assay that consists of a multiplexed array containing microbeads with different APC fluorescence intensities covalently coated with different RBD variants. The bead array is first incubated with serum samples, allowing the binding of anti-SARS-CoV-2 antibodies to RBD-coated microbeads, followed by the addition of rhACE2-mIgG1. The binding of rhACE2-mIgG1(rhACE2) to RBD-coated microbeads is then measured by flow cytometry after incubation with a fluorescent anti-mouse IgG1 antibody ( Figure 1A ). The neutralizing ability of antibodies is determined by the reduction on the fluorescence signal corresponding to the binding of rhACE2 to RBD-coated microbeads. Selected RBD variants included in this assay are the original RBD (Wuhan-1), the South African B.1.351 (β), the Indian B.1.167.1 (κ) and the South African B.1.1.529 (ο). The mutations present in the RBD of the variants can affect the binding affinity of ACE2. The variants included in nC19BA are shown in Figure 1B . We first assessed the ability of this assay to detect the interaction between rhACE2 and different RBD variants. Figure 1C shows that rhACE2 binds to RBD-coated microbeads but not negative control beads.

We then interrogated anti-RBD monoclonal antibodies for their ability to block the interaction between rhACE2 and different variants of RBD. To this end, we selected Imdevimab, an anti-RBD therapeutic monoclonal antibody (mAb) that is characterized as a SARS-CoV-2 Wuhan-1 neutralizing antibody without activity against the Omicron BA.1 variant (22–25). These features are clearly observed on Figure 2A , which shows that Imdevimab blocks the binding of rhACE2 to all tested variants except Omicron BA.1.

Another anti-RBD mAb, clone AM122, is able to neutralize all tested variants, including Omicron BA.1 although with reduced potency ( Figure 2B ). Together, these initial results indicate that this assay is a promising tool to support therapeutic antibody development with capability to quickly adapt to emerging variants. Importantly, the assay indicates that Omicron BA.1 is a variant that contains mutations that compromise the neutralizing activity of existing antibodies.

In order to determine the ability of nC19BA to measure the neutralizing activity of antibodies at the population level, we selected different cohorts of serum samples based on the time of sample collection, representing infected and/or vaccinated donors at different time-points before and during the pandemic. These cohorts presented different blocking capacity against the tested variants ( Figure 3A ), with significantly reduced ability to neutralize the Omicron BA.1 variant in the case of samples collected before 2022 ( Figure 3B ). In the case of samples collected after the emergence of Omicron BA.1 and the administration of a third dose of the vaccines in early 2022, neutralization of the Omicron BA.1 variant was significantly enhanced and complete, as previously demonstrated by other methods based on cellular infection assays (26–28).

Previous studies showed that there is a strong correlation between serum levels of anti-RBD antibodies and their blocking capacity (29, 30). In this context, we studied the correlation of antibody levels measured by ELISA ( Supplementary table 1 ) with the blocking activity against variants measured by the nC19BA flow cytometry assay, employing the same cohorts of seropositive individuals shown in Figure 3 . Using this method, we observed a significant correlation between anti-RBD (Wuhan-1) antibody levels and their neutralizing activity against the tested variants ( Table 1 ). This correlation was weaker in the case of the Omicron BA.1 variant, as a result of the different levels of neutralizing antibodies present in the different cohorts of samples ( Figure 3B ) and their binding capacity to Omicron BA.1 RBD ( Table 1 ).

In order to compare the performance of the bead-based neutralization assay with established surrogate methods, we measured the neutralizing activity of Imdevimab against the Wuhan-1 strain in both assays. Figure 4A shows that pseudotyped lentiviral particles expressing the spike of SARS-CoV-2 and encoding for a fluorescent reporter readily infect 293 cells overexpressing ACE2, and infection is effectively blocked by Imdevimad, as expected (31). The blocking activity percentages obtained in a dose response curve in this cellular assay were comparable to those obtained by the bead-based assay ( Figure 4B ), showing a strong correlation ( Figure 4C ). A Brand-Altman analysis demonstrated agreement between both methods ( Figure 4D ).

We then compared the performance of bead-based and cell-based assays for the determination of the neutralizing capacity of serum samples corresponding to preCOVID (n=10) and seropositive (n=18) cohorts. Figure 4E shows correlation in the neutralizing percentage assessed by both methods, and a Brand-Altman analysis was performed to demonstrate agreement ( Figure 4F ).