This study was conducted in accordance with the ethical principles of the Declaration of Helsinki and was approved by the Medical Ethics Committee of the First Affiliated Hospital of Guangzhou Medical University (ethics approval no. gyfyy-2021-31) and Guangzhou Eighth People’s Hospital (202002135). Written informed consent was obtained from all participants prior to the study, in strict adherence to the ethical standards outlined in the Declaration of Helsinki.

Forty-four healthy adult volunteers (n = 44) were randomly selected for follow-up visits before, one month after each of the three doses of the inactivated COVID-19 vaccination BBIBP-CorV, and were randomly selected for follow-up visits before immunization (before the first dose, i.e., healthy subjects, V1), one month after the first injection (V1+30), one month after the second injection (V2+30), and one month after the booster injection (V3+30). Out of these 44 volunteers, 22 had a hypoimmune response, indicating that no effective neutralizing antibody (NAb) was produced one month after the second dose, defined as virus neutralization test (VNT) <8, while the remaining 22 volunteers exhibited a hyperimmune response (VNT ≥16).

Furthermore, 61 patients with SARS-CoV-2 infection, confirmed using real-time quantitative polymerase chain reaction (RT-qPCR) and hospitalized in Guangzhou Eighth People’s Hospital, were divided into three groups according to the severity of their disease: 18 asymptomatic patients (AP), 33 mildly-symptomatic patients (MP), and 10 severely-symptomatic patients (SP). As asymptomatic infections do not develop symptoms, the day on which the nucleic acid test was first positive was defined as day 0.

Peripheral blood samples were collected from the antecubital vein after an overnight fast in 10 ml EDTA and 5 ml Serum-Gel tubes. The tubes were immediately centrifuged at 3000 rpm at 4°C for 10 minutes to separate the plasma and serum, respectively. Following centrifugation, all samples were aliquoted into 0.5 ml tubes and stored at -80°C until further processing. Plasma was used for live-virus neutralization assay and serum was used for antibodies and peptide microarrays detection.

Nucleic acids were extracted from samples primarily collected from nasopharyngeal tissue. The extraction was conducted according to the instructions of a commercial viral RNA extraction kit (DaAn Gene Co., Ltd., Sun Yat-sen University, China). Subsequently, reverse transcription-polymerase chain reaction (RT-PCR) assay kits targeting the SARS-CoV-2 open reading frame1ab (ORFlab) and nucleocapsid (N) gene regions were acquired from DaAn Gene Co., Ltd. (Guangzhou, China). All extraction and testing processes were conducted in accordance with scientific reporting standards.

The amino acid sequence of the SARS-CoV-2 strain (MN908947) was analyzed and the 20-mer peptides with an overlap of 10 amino acid (aa) residues, partially covering four structural proteins of SARS-CoV-2 (i.e., Spike, Envelope, Membrane, and N proteins), were chemically synthesized by GenScript (Jiangsu, China). The microarray yielded 131 peptides ( Table S1 ). In order to assess the protein-peptide interactions, a peptide and protein hybrid microarray (PPHM) was designed using RBD (GenScript, Jiangsu, China), (S1+S2) ECD (Sino Biological, Beijing, China) and the Nucleotide protein (N protein, VACURE Biotechnology, Sichuan, China) and 131 peptides of SARS-CoV-2, which finally yielded 136 peptides and proteins in total. Each well of the chip had a 4x4 rectangular microarray, with three human immunoglobulin G (IgG)-positive controls and one negative control in the four corners. The PPHM was screened to obtain the indicated peptides for detecting COVID-19 patients with both high sensitivity and high specificity, detailed descriptions was showed in supplementary material (Appendix 1). The probes included in the center of the microarray were M1, N16, N24, S15, S39, S44, S64, S82, S95, S104, and S115 (see Table 1 ).

The screening process was conducted following the same procedure as described previously (15) with minor modifications and using the indirect enzyme-linked immunosorbent assay (indirect ELISA) principle. To begin, diluted serum (100-fold) was prepared using a serum-dilution buffer containing 1% bovine serum albumin, 1% casein, 0.5% sucrose, 0.2% polyvinylpyrrolidone, 0.5% Tween 20 in 0.01 M phosphate-buffered saline (PBS, pH 7.4). A 100 μL sample of the diluted serum was then added to each microarray well and incubated with a peptide microarray for 30 minutes on a shaker (500 rpm, 37°C). The microarray well incubated with just serum-dilution buffer served as a negative control. Subsequently, the microarray was washed three times with 0.01 MPBS-Tween (PBST, pH 7.4) and then incubated with 100 μL of horseradish peroxidase (HRP)-conjugated anti-human IgG (ZSGB-BIO, Beijing, China) for a further 30 minutes on a shaker (500 rpm, 37°C). Following this, any unbound HRP-conjugated anti-human IgG was washed away with PBST and 100 μL of 1-step Ultra TMB-Blotting Solution (Thermo Scientific) was used to detect any informative signal of IgGs against peptide probes using a microarray imager (Suzhou Epitope, China). The data were then processed using UVC instrument V1.0 software (version Epitope Company, Suzhou, China). The signal for each dot was calculated by subtracting the background signal from the readout signal: Signal _dot = Signal _readout - Signal _background. The cut-off value for each probe was set at 10.

The FRNT was conducted in a certified biosafety laboratory for live-virus neutralization assay. Briefly, plasma samples were continuously diluted and mixed with 50 µL SARS-CoV-2 virus suspension (100 virus focal forming units, FFU) in a 96-well plate, followed by incubation at 37°C for 1 hour. The mixture was then transferred to a 96-well plate inoculated with Vero E6 cells (ATCC, Manassas, VA) and incubated for an additional hour at 37°C to allow for viral entry into the cells. Subsequently, the cell culture medium was removed and replaced with a covering medium (125 ml 1.6% carboxymethyl cellulose, CMC). The plates were then incubated at 37°C for 24 hours. After removal of the covering, the cells were fixed with 4% paraformaldehyde solution for 30 minutes. Subsequently, cells were permeabilized with 0.2% Triton X-100 and incubated with cross-reactive rabbit Anti-SARS-CoV-N IgG 40143R001 at 37°C for 1 hour, followed by incubation with HRP-conjugated goat anti-rabbit IgG (high+low) antibody (diluted at 1:4,000) (Catalog number:111-035-144, Jackson ImmunoResearch, West Grove, PA). After an additional incubation at 37°C for 1 hour, KPL TrueBlue peroxidase substrate (Seracare Life Sciences Inc., Milford, MA, USA) was used as catalyst for the reaction. Finally, an Elispot reader (Cellular Technology Ltd., Shaker Heights, OH) was employed to calculate the number of SARS-CoV-2 lesions.

Statistical analyses were conducted with SPSS software (version 25.0). The Mann-Whitney U test was used to evaluate the measurement data between the two groups, which are described as median and quartile distances. Spearman’s correlation analysis was used to analyze the correlations between the two groups. Furthermore, the ROC curve and AUC analysis were conducted using R pROC package, while accuracy, sensitivity, specificity, and cutoff value were calculated with R caret and epiR packages. In addition, odds ratio and corresponding 95% confidence intervals (CI) from logistic regression (LR) were calculated using the R-package stats. Finally, Excel and Graph Pad Prism 8.0 (La Jolla, USA) were used for charting. Statistical significance was denoted as follows: * P < 0.05, ** P < 0.01, *** P < 0.001, and **** P < 0.0001.

Forty-four healthy volunteers were enrolled in the study and administered three doses of the BBIBP-CorV vaccine. Follow-up was conducted approximately one month after each dose ( Table 2 ). The volunteers were separated into two groups: a hyperimmune response group (n=22), with an average age of 36.14 ± 9.14 years, and a hypoimmune response group (n=22), with an average age of 41.32 ± 5.96 years. Figure 1 illustrates the corrected sample sizes while the patients returned for follow-up visits.

We conducted an analysis of the peptide-specific antibody responses in 22 hyperimmune and 22 hypoimmune individuals, as indicated by clustered heatmaps ( Figure 2 ). We found that the distribution of the 11 peptides was consistent between the two groups, with no statistically significant difference before the third immunization ( Figure S1 ). Our findings suggest that healthy individuals vaccinated with inactivated COVID-19 vaccines produce similar specific antibody response spectra. However, there was a significant variability in the peptide reactions collected by V3+30, and the hypoimmune group showed significantly lower binding to peptides N22, N24, N40, and S58 and S69 compared to the hyperimmune group, while S82 was the opposite. Therefore, the detection method using short peptide antibodies can effectively distinguish between hypoimmune and hyperimmune vaccinated individuals. As such, antibody testing for COVID-19 peptides may be used as a screening tool for hypoimmune populations, helping to alert them to the need to bolster their immunity.

A cluster heatmap (Fig 3A was used to illustrate the distribution of peptide-specific responses in 44 vaccinated (V1, V1+30, V2+30, and V3+30) and 61 patients (AP, MP, and SP). Analysis revealed that specific IgG responses to the 11 epitopes were significantly weaker in all vaccinated populations than those in all types of patients. Subsequent analysis of S64, S15, and S104 peptides showed statistically significant differences in the specific IgG responses among healthy persons, infected persons, and vaccinees ( Figures 3B–D ).

As shown in Figure 3A , the specific reactivity of peptides differed between vaccinated and patients, and among patients of different affliction grades. To further investigate this, 18 AP, 33 MP, and 10 SP patients were enrolled and subjected to cluster analysis. Results revealed that the specific reactivity of S64, S15, and S104 peptides in AP was significantly higher than in MP and SP (P ≤ 0.001). However, there was no statistical difference between MP and SP ( Figures 4B–D ). On the other hand, the expression of M1, N24, S82, and S115 peptides in AP was significantly lower than in MP and SP (P ≤ 0.05), with no significant difference between MP and SP observed ( Figures 4E–H ).

Screening the antibody-binding activity of the SARS-COV-2 antigen allows for the exploration of existing potential epitopes and characteristics of the infection mechanism, providing a reference for COVID-19 treatment and peptide vaccine development. Therefore, we analyzed the correlation between the SARS-COV-2 peptide-specific IgG response and PRNT50 during infection. According to the correlation heatmap, peptide-specific IgG of M1, N24, S82 and S115 displayed a significant correlation with PRNT50 (0.55, 0.58, 0.58 and 0.73, respectively) (P ≤ 0.05, Figure 5A ). Multiple linear regression analysis confirmed the significance of the correlation with F = 47.758, P ≤ 0.001, and R = 0.649. Moreover, the effects of N24 and S115 peptides included in the model on the PRNT50 results were also found to be statistically significant (P ≤ 0.05).

It has been observed that antibody levels in vaccinated individuals typically peak one month after vaccination. Therefore, we analyzed the antibody levels of these four peptides one month after the first, second , and third doses included in the study. Additionally, we analyzed the variation trend of these four peptides in vaccinated individuals and found that N24 and S115 were most consistent with the variation trend of antibody levels ( Figure 5B ).

With the recent rise in screening efforts in China, an increasing number of patients have been identified as nucleic acid-positive. However, the viral load can reach undetectable levels within 1–2 weeks of symptom onset, making the diagnosis of AP essential for effective epidemic control. Unfortunately, antigen detection is currently plagued by low sensitivity and specificity. Identifying differences in peptide-specific IgG antibody responses between vaccinated, AP, and symptomatic patients (MP and SP) is thus of great importance for disease discovery, diagnosis, and treatment.

To this end, we used three markers, S64, S15, and S104, to conduct ROC analysis on patients after vaccination and on infected patients ( Figure 6A ). We found that these markers had both good sensitivity and specificity for distinguishing between patients and vaccinees, with combined diagnosis achieving a sensitivity of 91.3% and a specificity of 99.2% ( Table 3 ).

In addition, to differentiate AP from symptomatic patients, ROC analysis was conducted for S64, S15, and S104 with high expression in AP ( Figure 6B ) and M1, N24, S82, and S115 with low expression ( Figure 6C ). The sensitivities of S64, S15, and S104 to distinguish AP from symptomatic patients were 94.1%, 86.3%, and 96.1%, respectively, while the specificities of M1, N24, S82, and S115 were 94.1%, 92.2%, 86.3%, and 98%, respectively ( Table 4 ). The combined predictive sensitivity and specificity of the diagnostics reached 94.1% and 85.3%, respectively ( Figure 6D ; Table 4 ), making them suitable biomarkers to distinguish AP from symptomatic patients.