Equivalent Binding Of Sera From Omicron And Delta Period To Future Omicron Subvariants

Study design

De-identified residual sera (N=126) received at the Department of Clinical Virology for routine diagnostic testing between April 2021 and December 2022 were included in this study. Samples with collection date between April 2021 to December 2021 (N=33) and January 2022 to December 2022 (N=93) were classified as ‘Delta period’ and ‘Omicron period’ sera, respectively, representing the date range of delta and omicron VoC circulation in Vellore, as per our previous study.¹² The study was approved by the Institutional Review Board of Christian Medical College, Vellore (IRB No. 12917).

Detection of SARS CoV-2 antibodies

Sera were tested for antibodies to nucleoprotein (anti-N) using Elecsys anti-SARS-CoV-2 assay and to spike receptor binding domain [anti-S (RBD)] using Elecsys anti-SARS-CoV-2 S assay on a COBAS e401 analyser (Roche Diagnostics, Switzerland) based on the principle of electrochemiluminescence immunoassay (ECLIA). A cut-off index (COI) ≥1.0 was considered reactive for anti-N and ≥0.8 U/ml (units/ml) considered positive for anti-S (RBD). Antibodies to spike protein were tested using monomeric spike assay (Liaison SARS-CoV-2 S1/S2) referred to as “anti-S (S1/S2)” and trimeric spike assay (Liaison SARS-CoV-2 Trimeric S) referred to as “anti-S (TrimericS)” on the Liaison XL system (Diasorin, Italy). All samples were tested as per manufacturers’ instructions. Values ≥15 AU/ml (arbitrary unit/ml) and ≥33.8 BAU/ml (binding antibody units/ml) were considered positive for anti-S (S1/S2) and anti-S (Trimeric S) antibodies. Raw machine readouts were exported into excel spreadsheets for further analysis. Values above the upper range of detection [>400, >250 and >2080, respectively for anti-S (S1/S2), anti-S (RBD) and anti-S (TrimericS), respectively] were rounded off for the purpose of calculation.

Detection of binding of sera to omicron subvariants by ELISA

Sera were also tested for antibodies to trimeric spike protein of omicron subvariants - BA.1, BA.2, BA.2.12.2, BA.2.75, BA.4/5, and BA.7 [purified trimeric spike protein (ACRO Biosystems) expressed in HEK-293 cells] using an in-house enzyme linked immunosorbent assay (ELISA).¹² Briefly, 0.1 μg/well of trimeric spike protein of each subvariants were coated in high binding 96-well strip plate (Biomat, MT01F4-HB8) with incubation overnight at +4°C, followed by blocking with 3% BSA (bovine serum albumin). Sera (diluted 1/500 with 1% BSA in DPBS-T) were added to each well and incubated 1 hour at 37°C. After washing, goat anti-human IgG specific−HRP conjugate was added and incubated, followed by post-wash substrate addition (TMB). Optical density (OD) was read at 450 nm and 630 nm after addition of stop solution. Each plate included appropriate blank, negative control, positive control and internal quality control samples in addition to test sera. For each antigen, cut-off values for a sample to be considered ‘positive’ or ‘reactive’ indicative of presence of antibody (IgG) binding to trimeric spike of omicron subvariants was computed by calculating 3 standard deviations from the mean OD of negative samples (negative for anti-spike antibodies). The reactivity ratio (RR) i.e., ratio of sample OD to cut-off OD was computed for each antigen and compared between groups. Samples with RR ≥1 considered as ‘positive’ or ‘reactive’. ELISA OD and RR values were exported to excel spreadsheets for further analysis.

Antigen cartography

Serum ELISA OD values to each of the 6 omicron subvariant antigens in the groups with collection dates between April 2021-March 2022 (Delta High, Delta Low, Omicron High and Omicron Low; N=53) were plotted in 2-D antigenic space using the Antigenic Cartography web portal (https://acmacs-web.antigenic-cartography.org/) using default settings for the data analysis. From the session output, distance matrices of sera to the 6 antigens were exported and compared across the different groups. The analysis output coordinates (antigenic map) was plotted using GraphPad Prism version 8 (GraphPad Software, USA).

Hierarchical clustering

Further, we determined the level of hierarchical clustering of ELISA and CLIA (chemiluminescence immunoassay) data of all sera (both delta and omicron periods, N=126) against six omicron subvariants, anti-N, anti-S (RBD), anti-S (S1/S2) and anti-S (TrimericS). Data exploration and clusterability of the data was assessed with respect to assay type/antigen used and individual serum reactivity profile across all antigens (wild type and omicron subvariants). Analysis was performed using packages (hclust, pheatmap, ggplot2) on RStudio (v4.3.0).

SARS CoV-2 sequences (India) and structural mapping

High quality whole genome sequences from Indian strains encompassing wild, delta and omicron infection timeline between January 2020 and December 2022 were downloaded from public repositories [N=6147 sequences (2020: 1683 sequences; 2021: 3192 sequences; 2022: 1272 sequences)]. Sequences with gaps and ambiguities were removed and unique full length spike genes extracted (N=702). Sequences were classified into wild-type and alpha (N=374), delta (N=175) and omicron (N=153) using a maximum likelihood tree generated using FastTree (v2.1.11) in Geneious Prime 2023.1.2 (Biomatters Inc.). Each of the wild, delta and omicron datasets were aligned using MAFFT (v1.0). From the translated alignments, individual subdomains – S1 N terminal domain (NTD) (AA 16-303), S1 receptor binding domain (RBD) (AA 319-541) and S2 subunit (AA 686-1273) were extracted using the NC_045512 (SARS CoV-2 isolate Wuhan Hu-1; AA 1-1273) and extent of variability was compared within and between the VOCs. The amino acid conservation across the 3 variants were plotted on the Wu Hu-1 spike protein structural backbone (PDB: 6XRA) to illustrate conservancy across all variants in the S1 and S2 domains.

Statistical analysis

Levels of anti-spike IgG across variants was analysed by nonparametric one-way ANOVA (Friedman test). Statistical analysis and visualization of data was performed in GraphPad Prism version 8 (GraphPad Software, USA). Mann Whitney test was used to analyse significance between groups.

Delta and omicron sera show similarity in antigenic space

Given the heterogeneity in reactivity (P₆ and P_any) and similar levels seen between High S (Delta High and Omicron High, Figure 2F) and Low S (Delta Low and Omicron Low, Figure 2I) groups, we quantified antigenic distances between sera using antigenic cartography. We included 53 serum samples between April 2021-March 2022. Of these, 25 were from the delta period (9 ‘Delta High’ and 16 ‘Delta Low’) and 28 from omicron period (21 ‘Omicron High’ and 7 ‘Omicron Low’). The antigenic maps for ‘High S’ and ‘Low S’ groups as well as mean (SD) distance of sera to the 6 omicron variant antigens are shown in Figure 3A-D. It is noteworthy that all of sera, were placed within a grid of 1 antigenic unit with respect to all antigens. This was seen for both High S (Figure 3A) and Low S (Figure 3B) groups. Further, the distance of individual sera in the Delta High and Omicron High (Figure 3C) to each of the antigens was similar between the groups (ANOVA, p>0.05). The distance of the Low S groups (Figure 3D) from the antigens was also similar (ANOVA, p>0.05). These findings indicate that delta and omicron sera were closely related to the omicron subvariants in antigenic space.

• Download figure
• Open in new tab

Figure 3.

Antigenic cartography of Delta and Omicron sera against Omicron subvariants.

Antigenic Cartography for the ‘High S’ (Delta High and Omicron High) (3A) and ‘Low S’ (Delta Low and Omicron Low) (3B) sera groups against six omicron subvariants in 2D antigenic space. Omicron subvariants are shown as circles and sera are shaped as diamond. Each diamond corresponds to sera of one individual. Both axes of the map are antigenic coordinates. Mean (SD) antigenic distance in between the ‘High S’ groups (3C) and ‘Low S’ groups (3D) against six omicron variants. Sera are coloured as olive (Delta High), gold (Omicron High), red (Delta Low) and orange (Omicron Low). The omicron variants are coloured as pink (BA.1), brown (BA.2), green (BA.2.12.1), sky blue (BA.2.75), blue (BA.4/5) and purple (BF.7). The distance between points in the map can be interpreted as a measure of antigenic similarity, where the points more closely together indicate that they are antigenically more similar.

Hierarchical clustering

Given the observed antigenic distance but heterogeneity in levels/reactivity of sera groups with respect to anti-N, anti-S (RBD), anti-S (TrimericS), anti-S (S1/S2) and omicron subvariant antigens, we performed a hierarchical clustering of all sera (N=126) based on results of CLIA assays or OD values to each antigen (Figure 4). We further noted, heterogeneous spike binding manifested as positivity to 6 omicron variants (P₆, N=93) or any one (P_any, N=14) or none at all (negative, N=19) in delta and omicron periods. High S sera showed completely P6 while Low S sera showed a mix of P₆, P_any and negative. Hierarchical clustering of antibody testing (columns) showed distinct clusters with no delineated segregation by serum group or reactivity. Sera from the Low-S groups clustered together but were also seen to be interspersed with Delta High and Omicron High. Particularly notable was the finding of sera demonstrating P₆ reactivity to have High S (Delta High and Omicron High) and Low S (Delta Low and Omicron Low) groups interspersed. The clustering of assay type (rows), revealed clustering of omicron variants ELISA results together along anti-S (S1/S2) and anti-S (TrimericS) branches, while the anti-S (RBD) and anti-N were on separate branches.

• Download figure
• Open in new tab

Figure 4.

Hierarchical clustering of Delta and Omicron sera against spike of Wild-type and Omicron subvariants.

Hierarchical clustering of all sera (N=126) based on results of the CLIA assays or OD values against six omicron subvariants, anti-N, anti-S (RBD), anti-S (S1/S2) and anti-S (TrimericS). The dendrogram was annotated as per High S and Low S group and positivity (categorical) to individual antigens. The clusterability (Simpsons index) was 0.73. The optimal number of clusters based on k-means distance was 8. The heatmap is split according to clusters 1-8 from left to right by white space. Clusters grouped together based on ELISA antigens (i.e. P6 or Pany). Analysis was performed using packages (hclust, pheatmap, ggplot2) on RStudio (v4.3.0). Sera from delta and omicron period (both ‘High S’ and ‘Low S’ groups) are coloured as olive (Delta High), gold (Omicron High), red (Delta Low) and orange (Omicron Low). Individual sera positive to any omicron subvariant are indicated with brown colour and negative to all omicron subvariant are indicated with grey colour. Distance matrix ranged from -2 to +2.

Amino acid conservation in SARS CoV-2 sequences from India

S1 and S2 domain of the downloaded SARS CoV-2 sequences were analysed to assess level of variability. Through our analysis approach, we included high quality publicly available sequence data from India to capture the entire spectrum genetic diversity of SARS CoV-2 in India between 2020 to 2022. Overall, percentage identity at amino acid level across all variants of SARS CoV-2 was 63.5% (185/291), 80.3% (179/223) and 83.2% (489/588) in the S1-NTD, S1-RBD and S2 domain. Higher % conservation was seen in the fusion peptide (FP 1 and 2) at >90%. The % conservation ranged from 75-92%, 87-94% and 81-96% within wild, delta and omicron lineages, respectively, with higher conservancy seen among S2 compared to S1 regions. The conserved regions of S1 and S2 (trimeric configuration of PDB: 6XRA) with highlighted broadly neutralizing epitopes from literature are shown in Figure S1.

Omicron and delta sera show similar reactivity to omicron subvariants and wild-type antigen

In the ‘High S’ groups (i.e. Delta High and Omicron High), reactivity ratio (RR) of sera to the omicron subvariant antigens (BA.1, BA.2, BA.2.12.1, BA.2.75, BA.4/5 and BF.7) ranged from 2.1-4.0 and 0.6-3.9 for Delta High and Omicron High, respectively, and these were found similar (Mann-Whitney U-test, p>0.05). These groups showed similar reactivity levels to ‘wild-type’ spike antigens - anti-S (RBD), anti-S (TrimericS) as per manufacturers’ assay configuration for these commercial kits (CLIA/ECLIA). This indicates antibodies detected by the above methods were similar or had common binding regions.

Delta High groups represent a cumulative immunity to wild type and delta while Omicron High groups, represent cumulative exposure to wild type, delta and omicron and both groups include breakthrough infections among the vaccinated, respectively. Delta High sera were collected prior to the emergence of omicron in India (January 2022) and evidence of binding to omicron subvariant spike points to similarity or commonality of binding region between all variants (wild type, delta and omicron). India experienced multiple omicron waves.^3,17 Reinfections were common in the omicron period¹⁸ and individuals would likely have had multiple exposures to omicron subvariants (BA.1, BA.2, BA.2.12.1, BA.4/5, BA.2.75.2 etc.)^19,20 and thus can be expected to bind to conserved spike regions across omicron sub-lineages. Similarly, in the Low S groups, Delta Low and Omicron Low demonstrated similar RR level between them (Mann Whitney U-test, p>0.05). Likewise, levels were also similar between groups to wild type antigens. Sera in Low S groups, may be considered primary infection, and the observed binding to ancestral and “future” variants suggests antibody repertoire binding to S1 and S2 domains conserved across SARS CoV-2 wild type, delta and omicron variants. Re-infection or breakthrough infection leads to increased depth (antibody levels) as well as breadth (reactivity to heterologous viruses) of spike reactivity.^12,21 In the time course of repeat infections/exposures, S1 particularly NTD shows antigenic divergence between wild-type and successive VoCs, but S2 is relatively conserved^22,23 and contributes to serological cross reactivity.

Delta and omicron sera overlap in antigenic space

To examine reactivity to “future” antigens, antigenic cartography was performed in 2D space using ELISA OD values on sera collected between April to December 2021 (N=25, Delta) and January to March 2022 (N=28, Omicron). The dates of blood collection for these selected sera pre-dates the emergence of the viruses - BA.2.12.1, BA.2.75, BA.4/5 and BF.7 which emerged sequentially after May 2022. For Delta High and Omicron High groups, mean distance of sera to antigens ranged from 0.18-0.58 units and 0.14-0.57 units respectively, and were similar between groups. Similarly, in the Low S sera, both groups occupied one unit of antigenic space and mean distances to the 6 antigens were similar (Delta Low range: 0.21-0.35 units; Omicron Low range: 0.17-0.41 units). All sera (both High S and Low S groups) were placed within a narrow antigenic window of 1 unit in relation to subvariant antigens.

Wild, delta and omicron VOCs are placed wide apart due to amino acid mutations resulting in escape from neutralization.^24-26 Antigenic distance is a measure of relatedness between sera and antigens and use multidimensional scaling to generate consensus relatedness in 2D space. As a result, sera raised from homologous viruses do not occupy exact overlapping coordinates, but exist as a cluster. Our finding of aggregation of serum samples in a narrow antigenic space, indicated a close relationship and binding to similar regions on the spike. S2 domain of spike is more conserved compared to S1.^22,27 Taken together, these suggest that the binding regions of the sera are antigenic sites conserved between the variants.

Antibody reactivity profiles to wild and omicron antigens across entire time period of the pandemic

We examined anti-S (wild type) and anti-S (omicron subvariant) reactivity by hierarchical clustering in a Euclidean distance framework. Our observed clustering based on antigen type employed, showed grouping of non-RBD spike antigens represented by S1/S2 and TrimericS with omicron subvariant antigens. Further, we noted a dissimilarity of anti-S (RBD) and anti-N, which are based on wild type antigens with other spike antigens. These suggest that the “future variant” binding antibody repertoire target the conserved, consensus regions across wild and omicron spike but excluding RBD.

Delta Low and Omicron Low sera were interspersed with Delta High and Omicron High among clusters 3-8 (Figure 4). It suggests similarity in reactivity profile. Low S groups, due to primary infection result in reactivity to S1 and S2 subunits of homologous viruses. The reactivity to all antigens (P₆) seen with both Delta Low and Omicron Low sera, can be explained by antibody binding to regions on the spike common between delta and omicron. Hybrid immunity seen in the High S group, showed mainly P₆ reactivity, and is likely due to boosted level of antibodies to conserved regions. We have previously shown that omicron breakthrough results in higher binding to predecessor or ancestral viruses, mediated by antibodies binding to conserved regions.¹²

The infection/exposure history and impact on binding to future variants seen with SARS CoV-2 is analogous to sequential vaccination with a chimeric haemagglutinin (cHA). Booster with a divergent HA head (HA1, heterologous) after a primary (homologous) vaccination induces antibodies to stem (HA2) of influenza A viruses.²⁸ The subdominant stalk is retargeted by immune memory from the primary vaccine and boosted levels of stem-specific antibodies mediate protection against heterologous challenge.²⁹ Similar to influenza, variability of SARS CoV-2 spike is mainly in the S1-NTD compared to S2.

Serological profiles especially the High S group, need further experimental validation of binding specificity. Further studies into these antibodies for SARS CoV-2 may help the development of assays that can measure the efficacy of pan-coronavirus vaccine candidates, which can mediate protection from future variants as well as broadly-neutralizing (bn) antibodies directed in the conserved S2 domain could that offer cross-variant and pan-genus protection. Fusion peptide (FP) and heptad repeat 2 (HR2) sites may induce production of more broadly active nAbs. These same classes of antibodies have been shown to be involved in protection from severe disease.

Public Health impact

Our study showed that antibody binding to future omicron variants was seen to occur even among sera without exposure to these variants. These may partly explain the lack of sustained transmission in India which had a severe epidemic of delta but also implemented a successful vaccination program. Similarly, this existing immunity may prevent future epidemics. This would be of great benefit to populations in whom COVID-19 vaccine is not yet approved for use, i.e. children <12 years of age.

Limitations of the study

Our study used de-identified sera instead of an ideal setting of a temporal follow-up cohort with detailed documentation of infection and vaccination. Confirmation of results with a S2-specific assay would further strengthen the conclusion of this study. The inferences thus represent a sum-total of reactivity to conserved regions. Antigenic cartography depicts the antigenic distance based on neutralizing antibodies. In our study, we used ELISA OD values and not neutralizing end-point titres, potentially leading to inaccurate distance estimation between the omicron variants.