Evaluation of a biotin-based surrogate virus neutralization test for detecting postvaccination antibodies against SARS-CoV-2 variants in sera

Abstract

A severe acute respiratory syndrome coronavirus 2(SARS-CoV-2) surrogate virus neutralization test (sVNT) was used to determine the degree of inhibition of binding between human angiotensin converting enzyme 2 (hACE2) and the receptor binding domain (RBD) of spike protein by neutralizing antibodies in a biosafety level 2 facility. Here, to improve the sensitivity and specificity of the commercial sVNT, we developed a new biotin based sVNT using biotinylated RBD and HRP conjugated streptavidin instead of HRP conjugated RBD for direct detection in an ELISA assay that strongly correlated to the FDA approved cPass sVNT commercial kit (R² = 0.8521) and pseudo virus neutralization test (R² = 0.9006) (pVNT). The biotin based sVNT was evaluated in 535 postvaccination serum samples corresponding to second and third boosts of AZD1222 and BNT162b2 vaccines of the wild type strain. We confirmed that the neutralizing antibodies against SARS-CoV-2 variants in second vaccination sera decreased after a median of 141.5 days. Furthermore, vaccination sera from BNT162b2-BNT162b2 vaccines maintained neutralizing antibodies for longer than those of AZD1222 only vaccination. In addition, both vaccines maintained high neutralizing antibodies in third vaccination sera against Omicron BA.2 after a median of 27 days, but neutralizing antibodies significantly decreased after a median of 141.5 days. Along with the cPass sVNT commercial kit, biotin based sVNTs may also be suitable for specifically detecting neutralizing antibodies against multiple SARS-CoV-2 variants; however, to initially monitor the neutralizing antibodies in vaccinated sera using high throughput screening, conventional PRNT could be replaced by sVNT to circumvent the inconvenience of a long test time.

Nomenclature

sVNT

surrogate virus neutralization test

hACE2

human angiotensin converting enzyme 2

RBD

receptor binding domain

PRNT

plaque reduction neutralization test

pVNT

pseudovirus neutralization test

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

PBS

phosphate buffered saline.

1. Introduction

Since the emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), whose first case was designated as disease X, several COVID-19 vaccines have been developed through emergency use approval to prevent the pandemic [1]. Moreover, SARS-CoV-2 variants of concern (VOCs) continue to emerge, and new COVID-19 vaccines against these mutants are continuously in demand. Since the beginning of vaccinations against the initial Wuhan strain (wild type) with the AZD1222 (Oxford, AstraZeneca) vaccine in February 2021, South Korea has had a high vaccination rate in all ages, particularly in the over 60 year old population [2]. However, in January 2022, as the Omicron variants emerged, the number of infected patients gradually increased and reached 621,266 as of March 16, 2022, the highest number worldwide [3]. At the time, the second and third vaccination rates of the domestic SARS-CoV-2 vaccine were 86 and 63%, respectively [2], but it was difficult to determine the degree of infection because the efficacy of the vaccine had not been accurately confirmed for the Omicron and other SARS-CoV-2 mutations [4].

Robust serological tests to detect neutralizing antibodies against SARS-CoV-2 are urgently needed to determine vaccine efficacy and develop vaccination plans for persistent mutations following COVID-19 vaccination. The current standard test method is a conventional plaque reduction neutralization test (PRNT) used for confirming positive results, which requires over 3 days of testing and live viruses in biosafety level 3 facilities [5]. However, PRNT is complex, time consuming, and more expensive than other antibody detection systems, such as lateral flow rapid diagnostic kits or enzyme linked immunosorbent assays (ELISA). To make up for these shortcomings, alternative methods, such as ELISA based surrogate virus neutralization tests (sVNTs), which can be conducted within 2–3 h, have been widely developed and used at biosafety level 2 without using live viruses directly [6].

cPass (Genscript Biotech Corp, Piscataway, NJ, USA) is the first FDA approved sVNT kit for detecting neutralizing antibodies that inhibit human angiotensin converting enzyme 2 (hACE2) receptor binding domain (RBD) interaction [7]. However, previous studies revealed that the cPass kit showed false positive results for 32.4 and 58.3% of samples identified as negative by PRNT-50 and PRNT-90, respectively [8,9].

In this study, given that cPass sVNTs could have higher false negative and positive rates for samples than PRNT, we developed a modified sVNT using biotinylated RBDs interacting with a hACE2 6 histidine tag coated on a nickel coated plate and streptavidin HRP for visualizing signals to improve sensitivity and specificity. Furthermore, a biotin based sVNT was employed to detect neutralizing antibodies against multiple SARS-CoV-2 mutants and measure the neutralizing efficacy in postvaccination clinical sera using a high throughput assay. Therefore, we aimed to evaluate the ability of COVID-19 vaccines to form neutralizing antibodies against SARS-CoV-2 variants in 535 vaccinated sera samples using biotin based sVNTs. Finally, we compared the ability to maintain neutralizing antibodies according to the type of vaccine, number of vaccinations, and the period from vaccination to sampling.

2. Materials and methods

2.1. Recombinant protein expression and purification

RBD genes from the spike protein of the Wuhan strain (wild type), B.1.1.7 (Alpha), B.1.351 (Beta), P.1 (Gamma), B.1.617.2 (Delta), and Omicron BA.2 were codon optimized and synthesized by Genscript Biotech Corp. All 323–532 RBD residues of the spike protein were aligned using Clustal X (Fig. 1 A). The hACE2 gene with the 6 histidine tag in place of the transmembrane region at the C-terminus was also synthesized and spanned amino acid positions 16–740. RBD and 6 histidine tagged hACE2 genes were subcloned to a pcDNA3.4 TOPO vector (Thermo Fisher Scientific, Waltham, MA, USA) inserted with the KOZAC sequence, a start codon, and murine Ig κ-chain leader sequence at the N-terminus sequentially for protein expression and secretion. ExpiCHO cells were cultured for 6–10 days at 37 °C, 8% CO₂, and 120 rpm in 50 mL of ExpiCHO expression medium. All transfections were performed using the ExpiCHO Expression System (Thermo Fisher Scientific), applying the Max titer protocol starting at 6 × 10⁶ cells/mL, adding ExpiCHO Enhancer and Feed media after 18–22 h, and transferring to a 32 °C shaking incubator with 5% CO₂. Five days post transfection, ExpiCHO Feed media was added and cultured continuously. After 12–14 days post transfection, when cell viability reached over 75%, cell culturing was stopped, the media was centrifuged at 3500×g at 4 °C for 30 min, and then the supernatant was filtered through 0.8 μm pore size filter paper (Whatman plc, Maidstone, UK).

Fig. 1.

Sequence and structure alignment of SARS-CoV-2 RBDs, and SDS-PAGE of purified RBD proteins and hACE2. (A) Primary structure of SARS-CoV-2 spike protein. SP, signal peptide; NTD, N-terminal domain; RBD, receptor binding domain; TMD, transmembrane domain. Sequence alignments of the SARS-CoV-2 RBD proteins for the wild-type and mutants (B) Structural alignment of SARS-CoV-2 RBDs. Superimposed structures of wild-type RBD(7VYR, green), B.1.1.7 RBD(7EK0, cyan), B.1.351 RBD(7EKG, purple), P.1RBD(7EKC, yellow), B.1.617.2 RBD(7WBQ, pink), and Omicron BA.2 RBD(7ZE7, blue) as a ribbon diagram with electrostatic surface. (C), SDS-PAGE of 5 μg SARS-CoV-2 RBD proteins and hACE2 protein. Lane 1, SARS-CoV-2 wild-type RBD; lane 2, SARS-CoV-2 B.1.1.7 RBD; lane 3, SARS-CoV-2 B.1.351 RBD; lane 4, SARS-CoV-2 P.1RBD; lane 5, SARS-CoV-2 B.1.617.2 RBD; lane 6, hACE2 protein; lane 7, SARS-CoV-2 Omicron BA.2 RBD; M, marker.

Ammonium sulfate and sodium phosphate buffer (pH 6.5) were gently mixed with the clarified RBD culture supernatant to final concentrations of 1.2 M and 25 mM, respectively. Thereafter, the sample was loaded into a HiTrap Phenyl FF column (Cytiva, Marlborough, MA, USA) previously equilibrated with 1.2 M ammonium sulfate and 25 mM sodium phosphate buffer (pH 6.5). Thereafter, the RBD protein was eluted by reverse linear gradient, and then the fractions containing RBD protein were dialyzed in 10 kDa MWCO tubing (Spectrum; Fisher Scientific, Hampton, NH, USA) with 20 mM potassium phosphate buffer (pH 6.8). Next, the dialyzed sample was loaded into a hydroxyapatite column (Bio-Rad Laboratories, Hercules, CA, USA), and the protein was eluted using a 100 mM potassium phosphate buffer (pH 6.8). Finally, gel filtration was performed using a Superdex 200 10/300 column (Cytiva) equilibrated with 2 fold phosphate buffered saline (PBS).

The 6 histidine tagged hACE2 culture supernatant was dialyzed in 30 kDa MWCO tubing (Spectrum; Fisher Scientific) with 25 mM Tris-HCl (pH 7.5) and 0.2 M NaCl and loaded into a Ni-NTA superflow column (Qiagen, Hilden, Germany), after which the imidazole elution fractions of 6 histidine tagged hACE2 proteins were collected and concentrated. For further purification, gel filtration was performed using a Superdex 200 10/300 column (Cytiva) equilibrated with 2 fold PBS. Subsequently, the purified RBD and 6 histidine tagged hACE2 protein fractions were collected and concentrated to 1 mg/mL in 10 MWCO polyethersulfone (PES) using a Vivaspin Ultrafiltration Unit (Sartorius, Göttingen, Germany). All purification steps were performed using the AKTA prime and go system (Cytiva).

2.2. Biotinylation of RBD proteins

All RBD proteins were biotinylated using EZ-Link™ Sulfo–NHS–Biotin (Thermo Fisher Scientific). First, 2.2 mg of Sulfo–NHS–Biotin was dissolved in 500 μL pure water and added to RBD protein solutions at a molar ratio of 20:1. After incubating on ice for 2 h, the unreacted biotin was removed by dialysis overnight against 2 fold PBS buffer. Next, biotin conjugated RBD proteins were concentrated at 1 mg/mL and stored at 4 °C before use.

2.3. cPass and biotin based surrogate virus neutralization test (sVNT)

The FDA approved sVNT kit cPass (Genscript Biotech Corp) was used according to the manufacturer's instructions [7]. Briefly, serum samples and positive and negative controls were diluted at 1:10 in dilution buffer, mixed with 100 μL HRP-RBD solution, and incubated at 37 °C for 30 min. Subsequently, 100 μL of the samples and controls were added to a 96 well plate coated with the hACE2 protein. Thereafter, the plate was incubated at 37 °C for 15 min and washed four times with 260 μL of washing buffer, after which 100 μL of TMB solution was added in the dark for 15 min at 25 °C. Finally, 50 μL of stop solution per well was added, and the absorbance at 450 nm was measured using a microplate reader. (Victor III; PerkinElmer, Waltham, MA, USA).

Our biotin based sVNT used 100 ng immobilized 6 histidine tagged hACE2 proteins on nickel coated plates (Nunc; Thermo Fisher Scientific) incubated overnight at 4 °C (1 μg/mL). In addition, 2 ng of biotin conjugated RBD protein was used to measure the binding affinity of hACE2. A 50 fold diluted serum sample and biotin conjugated RBD mixture were transferred to a 6 histidine tagged hACE2 coated plate and incubated for 10 min at room temperature. Biotin conjugated RBD was unbound from hACE2 by washing twice with phosphate buffered 0.05% Tween-20 (PBST). Afterward, HRP conjugated streptavidin was added to the wells and incubated for 1 h at room temperature to visualize the biotinylated protein. After washing four times with PBST, the unreacted HRP conjugated streptavidin was removed. Subsequently, TMB substrate (Invitrogen, Waltham, MA, USA) was added in the dark, and the absorbance was measured using a microplate reader (Victor III; PerkinElmer, Waltham, MA, USA). The percent binding inhibition of neutralizing antibodies was subsequently calculated as [1 – (optical density of sample/optical density of negative control)] × 100%.

2.4. Pseudo virus neutralization test (pVNT)

For the neutralization test using pseudovirus, 1 × 10⁴ RLU SARS-CoV-2 spike protein based pseudoviruses (Genecopoiea) containing the luciferase gene were incubated with diluted serum samples (1:50 dilution) for 1 h at 37 °C. Next, 2 × 10⁴ hACE2-HEK293T cells were added to each well. hACE2-HEK293T cells without the pseudovirus and the sera of participants were used as blank controls, while cells with the pseudovirus without serum were used as virus controls [10]. Luciferase signals were detected for 24 h after infection (Promega, Madison, WI, USA), with the percentage neutralization calculated as 100% − [(sample signals − blank control signals)/(virus control signals − blank control signals) × 100%].

2.5. Correlation among the cPass, biotin based sVNT, and pVNT and negative cut off determination

To more effectively compare and evaluate the relationship between biotin based sVNTs and cPass sVNTs, and between biotin based sVNTs and pVNTs, Pearson's correlation coefficient (r) was calculated for the correlation among each VNT using sera from 100 participants. For the absolute value of Pearson's r, 0.00–0.19 was considered very weak, while 0.20–0.39 was weak, 0.40–0.59 was moderate, 0.6–0.79 was strong, and 0.80–1.00 was a very strong correlation [11]. In addition, a negative cut off value was established using 154 negative samples to distinguish the background reaction from the positive surrogate neutralization reaction of SARS-CoV-2 negative samples.

2.6. Sample collection and measurement of neutralizing antibody inhibition

Clinical participants in three vaccination groups inoculated with second and third doses of AZD1222 (AstraZeneca, Cambridge, UK) and BNT162b2 (Pfizer–BioNTech, New York, NY, USA and Mainz, Germany) vaccines using homologous or heterologous prime boost methods, respectively, visited the Korea Institute of Radiological & Medical Sciences (KIRAMS) within 4–20 weeks after vaccination (Table 1 ). The institutional review board of KIRAMS approved this study (IRB number: KIRAMS2022-01-003-001). Additionally, all participants submitted a written consent form and brief questionnaire before participating in the study. Serum from participants was collected for VNT analysis, excluding serum samples from participants with concurrent active infection, assessed through questionnaires on symptoms of SARS-CoV-2 infection. The neutralizing antibodies of serum from 535 participants (median age 42 years, a short period: median 27 [22 to 32] days after pre boost or after a boost; after a long term period: more than median 141.5 [97 to 168] days after pre boost or after a boost) were measured against the wild type (wild type), B.1.1.7 (Alpha), B.1.351 (Beta), P.1 (Gamma), B.1.617.2 (Delta), and Omicron BA.2 using biotin based sVNT.

Table 1.

Initial data analysis of participants in this study (Korea Institute of Radiological & Medical Sciences).

Vaccination	Prime-pre boost			Prime-pre boost-boost			total
	AZD1222-AZD1222	AZD1222 BNT162b2	BNT162b2 BNT162b2	AZD1222 AZD1222 BNT162b2	AZD1222 BNT162b2 BNT162b2	BNT162b2 BNT162b2 BNT162b2
Total	19	18	93	15	17	103	265
M	5(26%)	5(28%)	10(11%)	6(40%)	4(24%)	9(9%)	39(14.7%)
F	14(74%)	13(72%)	83(89%)	9(60%)	13(76%)	94(91%)	226(85.3%)
Median days since final Dose(IQR)	25 [21–32]	26 [23.50–29.75]	25 [14–35]	26 [22–27]	25 [20.5–26]	28 [24–30]	27 [22–32]
Age median	54	49	39	55	43	40	42
Vaccination	Prime-pre boost			Prime-pre boost-boost			total
	AZD1222-AZD1222	AZD1222-BNT162b2	BNT162b2-BNT162b2	AZD1222-AZD1222-BNT162b2	AZD1222-BNT162b2-BNT162b2	BNT162b2-BNT162b2-BNT162b2
Total	31	19	138	10	9	63	270
M	9(29%)	4(21%)	10(7%)	5(50%)	3(33%)	3(5%)	34(12.6%)
F	22(71%)	15(79%)	128(93%)	5(50%)	6(67%)	60(95%)	236(87.4%)
Median days since final Dose(IQR)	98 [91–158]	95 [88–103]	167 [97–181]	117 [117–130.5]	123 [114–133.5]	141 [139–146]	141.5 [97–168]
Age median	50	44	40	53	42	43	42

2.7. Statistical analysis

Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA). Receptor binding inhibitions are expressed as percentages, while inhibition percentages, ages, and days after vaccination are all expressed as means with median and interquartile ranges (IQRs), with 95% confidence intervals (CI). Receptor binding inhibition data were analyzed using one way analysis of variance with Tukey's multiple comparison test. Each experiment was performed in triplicate. Significance levels are indicated as ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

3. Results and discussion

3.1. Preparation of SARS-CoV-2 RBDs and hACE2 protein for biotin based sVNT

We aimed to improve a commercially available sVNT assay for monitoring the neutralizing antibodies in serum that inhibit binding between the hACE2 protein and RDB. To upgrade a previous sVNT, recombinant RBDs of spike protein and the receptor hACE2 were constructed, transfected, and purified using the CHO cell expression system. Since the first virus outbreak in Wuhan, China, in 2019, viral genome sequencing has been conducted and reported rapidly, with the sequence of the spike protein on the surface of the virus being analyzed first [12]. Results showed that it consisted of an S1 head region, including a signal peptide and NTD and RBD domains at the N-terminus and an S2 stalk region with a transmembrane domain at the C-terminus. In addition, an S cleavage site, widely used to activate the fusion machinery of viral glycoproteins, existed between S1 and S2 [13]. Following the initial SARS-CoV-2 outbreak, four representative species, including B.1.1.7 (Alpha), B.1.351 (Beta), P.1 (Gamma), and B.1.617.2 (Delta), and the recent Omicron BA.2 of the RBD amino acid sequence were aligned and highlighted the sequences with a difference (Fig. 1A). The RBD domain was generally defined at amino acid residues 318–541, but the 323–532 region was expressed by deleting both terminals to design a more globular domain structure based on the protein structure (Fig. 1B). All RBDs did not have a tag for affinity chromatography purification; therefore, they were purified using the natural properties of the protein itself. We successfully purified the RBD proteins using phenyl sepharose and hydroxylapatite chromatography and proved that these two methods were exceedingly effective for purifying RBD proteins. The hACE2 protein had a 6 histidine tag attached to the C-terminus and was also purified using Ni-affinity chromatography. RBDs and hACE2 were finally obtained as a single peak fraction by gel filtration chromatography using Superdex 200. RBD proteins were purified at yields of 40–130 mg and hACE2 at 21.6 mg in a liter of culture media (Fig. 1C).

3.2. Comparison of cPass (Genscript) with biotin based sVNT

Neutralizing antibodies are an important indicator for evaluating the efficacy of a vaccine [14], and the methods for measuring neutralizing antibodies include PRNT, cytopathic effect(CPE) and ELISA based assays, and the pVNT method. The first sVNT, an ELISA based assay, was developed by Genscript Biotech as a cPass SARS-CoV-2 sVNT. To measure neutralizing antibody inhibition, serum containing neutralizing antibodies, controls, and HRP conjugated with SARS-CoV-2 RBD were incubated and added into a 96 well microplate pre coated with hACE2, and the degree of the inhibition of RBD binding to hACE2 was measured. Absorbance was measured at 450 nm, and % inhibition was calculated. (Fig. 2 A). For the cut off interpretation of cPass cVNT, more than 20% percent inhibition was regarded as positive, and under 20% was negative according to the first version of the manufacturer's instructions. The 20% positive cut off was recently raised to 30% by the manufacturer to improve specificity. Moreover, to improve the sensitivity and specificity of cPass sVNT, we replaced HRP-RBD with biotin conjugated RBD and developed a method for indirect detection by binding streptavidin to which HRP is conjugated (Fig. 2B). Because biotin (0.556 kDa, Sulfo–NHS–LC-Biotin) has a much smaller molecular weight than HRP (44 kDa), it can be specifically bound to RBD of approximately 26 kDa and interferes less with the receptor binding motif for binding to hACE2 than when HRP is conjugated to RBD. The greatest improvement of our sVNT was observed using hACE2 attached to the 6 histidine tag at the C-terminus such that the orientation of the N-terminus of the hACE2 protein structure was directed upward and attached to the nickel coated plate. Kento et al. previously developed an sVNT that used biotin conjugated hACE2; RBD was coated on 96 well microplates with a random protein orientation and showed a high correlation (R ² = 0.76) compared with that of pVNT [15]. The difference between Kento et al.‘s sVNT and ours is the type of microplate for ELISA reaction and biotinylated protein. The advantage of our sVNT hACE2 protein coating the microplate was that it was attached in an orientation that could bind well to the RBD, improving sensitivity. The amount of coating protein used was 100 ng, the same for both sVNTs; Kento's sVNT used 50 ng biotinylated hACE2, whereas we only used 2 ng biotinylated RBD to detect the degree of neutralizing antibody inhibition. In addition, due to its high sensitivity, our biotin-based sVNT can be detected at a higher serum dilution rate than cPass sVNT, allowing the use of a small amount of sample. These results indicate that our biotin based sVNT is effective, and it is possible to distinguish the extent to which it specifically inhibits the binding of neutralizing antibodies to hACE2 using a small amount of RBD. In addition, by adjusting the amount of RBD, various neutralization activities at high and low levels could be measured.

Fig. 2.

Graphical illustration of the principle of a surrogate SARS-CoV-2 virus neutralization test. (A) cPass sVNT kit (Genscript Biotech) and (B) a biotin-based sVNT (in this study). Biotin-based sVNTs use biotinylated RBD for the detection of hACE2 that has not been inhibited by neutralization antibodies via HRP-conjugated streptavidin.

3.3. Correlation of biotin based sVNT to cPass (Genscript) and pVNT

Comparing the correlation between biotin based sVNTs, pVNTs, and cPass sVNTs, we performed a comparative analysis using 100 postvaccination serum samples with neutralizing activity. As can be seen in Fig. 3 , all VNT assays (biotin based sVNT, cPass, and pVNT) had an excellent overall correlation. Biotin based sVNTs had a strong correlation with cPass sVNTs (r = 0.9231, R ² = 0.8521) and pVNTs (r = 0.949, R ² = 0.9006), and cPass sVNTs also had a high correlation with pVNTs (r = 0.9433, R ² = 0.8899). All VNT titers were calculated using the degree of hACE2 RBD or hACE2 spike protein binding inhibition. Our results revealed that biotin based sVNTs showed not only a strong correlation with cPass sVNTs but also higher specificity with pVNTs than that of cPass sVNT. We also established the negative cut off value using 154 negative samples, and the median surrogate neutralization antibodies inhibition was 12.04% (standard deviation = 4.51) in the present study. The cut off value was determined by the median and an SD that was 2 SDs above the median; the cut off value was 21.1% [16]. Based on these results, we measured the neutralization antibody inhibition against SARS-CoV-2 variants in vaccinated sera. Although our negative cut off was 21.1%, which was lower than the 30% of the cPass sVNT method, the criterion was used to compare the relative inhibition of neutralizing antibodies in postvaccination serum and not to distinguish positive from negative results.

Fig. 3.

Correlation among cPass sVNTs, biotin-based sVNTs, and pVNTs and negative cut-off determination (A). Biotin-based sVNT and cPass sVNT (B), Biotin-based sVNT and pVNT (C), and cPass sVNT and pVNT. (D) Negative cut-off determination. Correlation analyses of the biotin-based sVNTs, cPass sVNTs, and pVNTs using the sera of 100 selected participants with different SARS-CoV-2 neutralizing antibody levels and a negative cut-off test performed using a biotin-based sVNT kit. Correlation and linear regression analyses were performed using Pearson's correlation coefficient. Statistical significance was calculated using a two-tailed t-test. The data presented are the inhibition of hACE2 and RBD binding to modified sVNT, cPass, or pVNT and are averages of three independent experiments. The dotted line represents the standard deviation of the linear regression plot.

3.4. Evaluation of biotin based sVNT against multiple SARS-CoV-2 mutants in postvaccination sera

The neutralizing antibodies in the serum of 535 participants (median age 42 years, a short period: median 27 [22 to 32] days after pre boost or after a boost; after a long term period: more than median 141.5 [97 to168] days after pre boost or after a boost; Table 1) were measured against the wild type [WT], B.1.1.7 (Alpha), B.1.351 (Beta), P.1 (Gamma), B.1.617.2 (Delta), and Omicron BA.2 variants using biotin based sVNTs.

For the vaccination groups, neutralizing antibodies were compared at a median value of 27 and 141.5 days after the second vaccination using the prime boost method of AZD1222-AZD1222, AZD1222-BNT162b2, and BNT162b2-BNT162b2. At a median of 27 days after the second vaccination, neutralizing antibodies against all SARS-CoV-2 variants were not observed in the AZD1222-AZD1222 group. This result differed from a previous study that confirmed the production of neutralizing antibodies after two doses of the ADZ1222 vaccine targeting healthy medical workers [17]. Therefore, we suggested that vaccine efficacy depends on factors such as the age, sex, and underlying diseases of the inoculated person. In particular, the presence or absence of underlying diseases is related to a decrease in antibody production [18].

The neutralizing antibodies of the AZD1222-BNT162b2 group were at a higher level (95% confidence interval [CI], the median value of WT, 54.2%; B.1.1.7, 41.2%; B.1.351, 33.1%; P.1, 41.8%; and B.1.617.2, 64.4%) than those of the AZD1222-AZD1222 and BNT162b2-BNT162b2 groups (Table 2 ). The BNT162b2-BNT162b2 group had approximately 50% neutralizing antibodies against the WT and B.1.617.2 variants, and neutralizing antibodies against Omicron BA.2 were not observed in all vaccination groups (Fig. 4 A). After approximately 141.5 days, very few neutralizing antibodies against the virus were observed in all vaccination groups, except for WT or B.1.617.2, and neutralizing antibodies against Omicron BA.2 were also not observed in all vaccination groups (Fig. 4B).

Table 2.

Median neutralizing antibody inhibition percentage of 95% CI after second and third vaccinations.

Median neutralization antibody inhibition percentage (%) of 95% CI
Median of 27 days after vaccination	Prime-pre boost			Prime-pre boost-boost
	AZD1222 AZD1222	AZD1222-BNT162b2	BNT162b2 BNT162b2	AZD1222-AZD1222-BNT162b2	AZD1222-BNT162b2 BNT162b2	BNT162b2 BNT162b2 BNT162b2
WT	15.3% [3.7–24.4]	54.2% [30.9–86.5]	44.7% [33.1–60.5]	88.8% [74.5–91.1]	88.1% [79.9–91.7]	90.2% [83.3–93.4]
B.1.1.7(Alpha)	9.9% [3.5–13.9]	41.2% [23.4–76.3]	26.4% [15.6–36.9]	85.1% [57.8–93.9]	84.8% [69.6–92.0]	92.3% [83.3–95.1]
B.1.351(Beta)	7.4% [5.2–19.2]	33.1% [19.1–59.8]	20.5% [11.2–31.0]	72.5% [45.0–83.5]	71.5% [38.1–79.7]	79.0% [60.3–89.1]
P.1(Gamma)	17.6% [5.8–23.6]	41.8% [29.6–69.9]	23.1% [16.7–33.9]	78.7% [57.5–88.9]	80.2% [42.9–84.4]	85.3% [70.9–90.8]
B.1.617.2(Delta)	16.8% [7.5–30.3]	64.4% [41.9–88.5]	49.4% [34.1–63.0]	90.7% [73.2–9.6]	90.0% [76.0–93.4]	92.3% [88.3–94.5]
Omicron BA.2	3.1% [-0.5–10.7]	15.4% [5.2–31.9]	5.8% [-20.5–12.9]	43.1% [14.3–57.2]	35.5% [14.8–45.2]	52.0% [34.5–64.8]

Median of 141.5 days after vaccination	Prime-pre boost			Prime-pre boost-boost
	AZD1222-AZD1222	AZD1222-BNT162b2	BNT162b2-BNT162b2	AZD1222-AZD1222-BNT162b2	AZD1222-BNT162b2-BNT162b2	BNT162b2-BNT162b2-BNT162b2
WT	12.7% [5.6–26.0]	13.5% [10.5–37.1]	21.4% [10.5–30.4]	69.1% [52.6–89.4]	77.1% [32.2–90.1]	65.0% [42.9–88.6]
B.1.1.7(Alpha)	7.7% [4.8–11.6]	10.6% [4.9–14.1]	10.0% [5.8–16.7]	39.1% [29.5–82.3]	53.3% [17.5–92.5]	41.2% [25.8–83.0]
B.1.351(Beta)	10.0% [4.5–14.0]	13.4% [0.4–20.2]	9.0% [3.1–17.1]	31.9% [11.6–71.3]	40.0% [-0.6–86.6]	29.8% [13.7–61.6]
P.1(Gamma)	11.4% [4.8–20.5]	15.1% [10.9–26.8]	12.6% [5.9–20.8]	43.7% [27.4–78.8]	34.8% [9.3–88.0]	36.7% [24.0–70.9]
B.1.617.2(Delta)	12.6% [6.6–28.2]	24.4% [17.9–34.9]	19.8% [11.7–31.0]	68.9% [49.6–89.8]	80.9% [33.2–93.9]	69.1% [49.5–90.2]
Omicron BA.2	6.4% [-0.4–14.3]	−0.1% [-19.8–10.2]	6.5% [-20.5–12.9]	14.1% [-1.7–46.6]	18.8% [-15.6–69.7]	12.1% [2.3–38.0]

Fig. 4.

Neutralizing antibody inhibition (%) (NAbI) against SARS-CoV-2 variants from study participants who had received second and third doses of AZD1222 (AstraZeneca) and BNT162b2 (Pfizer–BioNTech) at medians of 27 and 141.5 days following vaccination. (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001)

(A) NAbI of second dose vaccination at a median of 27 days, (B) NAbI of second dose vaccination at a median of 141.5 days, (C) NAbI of third dose vaccination at a median of 27 days, (D) NAbI of third dose vaccination at a median of 141.5 days.

Neutralizing antibodies in the AZD1222-AZD1222-BNT162b2, AZD1222-BNT162b2-BNT162b2, and BNT162b2-BNT162b2-BNT162b2 groups were compared at medians of 27 and 141.5 days after the third vaccination. After a median of 27 days, high neutralizing antibodies against all SARS-CoV-2 variants were observed in all vaccination groups, regardless of the vaccine type (95% CI, WT, B.1.1.7, B.1.351, P.1, and B.1.617.2 was above 71.5%, Omicron BA.2 was above 35.5%) (Table 2, Fig. 4C). However, the neutralizing antibodies for all SARS-CoV-2 variants except Omicron BA.2 remained between 37.7 and 89.8% after a median of 27 days. After a median of 141.5 days, neutralizing antibodies against Omicron BA.2 (P < 0.0001) significantly declined (Fig. 4D).

We found differences in the generation of neutralizing antibodies against various SARS-CoV-2 variants, including Omicron, at different times after the second and third vaccinations with AZD1222 and BNT162b2 vaccines. High levels of neutralizing antibodies were observed within 2–4 weeks after the second and third vaccinations in all vaccine groups owing to high humoral immunogenicity. However, neutralizing antibodies were not observed in the group vaccinated with the AZD1222 vaccine only up to the second round; however, after the third boost with the BNT162b2 vaccine, neutralizing antibodies increased. In addition, from approximately 4 months after vaccination, the persistence of neutralizing antibodies against various SARS-CoV-2 variants decreased rapidly, especially in the case of the second vaccination group compared with the third vaccination group. Moreover, in the case of the third vaccination group, it was confirmed that neutralizing antibodies against Omicron BA.2 significantly decreased. These results are consistent with previous data for the detection of neutralizing antibodies against the Omicron mutant using PRNT in sera following BNT162b2 vaccination [19,20].

In conclusion, we developed a biotin based sVNT for SARS-CoV-2 and evaluated the correlation between cPass sVNTs and pVNTs. A biotin based sVNT microplate was coated with hACE2 protein, considering the orientation of hACE2, such that it was most likely to bind RBD conjugated with biotin, and signals could be detected using streptavidin conjugated with HRP. Biotin based sVNTs, together with other sVNTs, were well suited for the initial high throughput screening of neutralizing antibodies in vaccinated individuals or SARS-CoV-2 convalescent sera. However, because the sVNT assay is limited to detecting neutralizing antibodies that inhibit the binding of hACE2 to RBD proteins, PRNT or CPE methods are required when viruses are neutralized in other mechanisms. Finally, as SARS-CoV-2 Omicron variants continuously emerge, neutralizing antibodies in sera depending on vaccine type and vaccination interval can be monitored using sVNTs changing recombinant RBD proteins without live viruses.

Funding

The Korea Research Institute of Bioscience and Biotechnology (KRIBB, South Korea), National NonoFab Center (NNFC, South Korea) Initiative program, and the National Research Council of Science & Technology (NST, South Korea) provided financial support (grant CAP21061-100).

Declaration of competing interest

All authors declare no competing interests.