key: cord-0257976-3mh6ckkj
authors: van der Straten, K.; Guerra, D.; van Gils, M.; Bontjer, I.; Caniels, T. G.; van Willigen, H. D.; Wynberg, E.; Poniman, M.; Burger, J. A.; Bouhuijs, J. H.; Lavell, A.; Appelman, B.; Sikkens, J. J.; Bomers, M.; Han, A. X.; Nichols, B. E.; Prins, M.; Vennema, H.; Reusken, C.; Eggink, D.; de Jong, M.; de Bree, G. J.; Russell, C. A.; Sanders, R. W.
title: Mapping the antigenic diversification of SARS-CoV-2
date: 2022-01-03
journal: nan
DOI: 10.1101/2022.01.03.21268582
sha: 9249ca4d25b80d95fd579043c6b6825de978c390
doc_id: 257976
cord_uid: 3mh6ckkj

Large-scale vaccination campaigns have prevented countless SARS-CoV-2 infections, hospitalizations and deaths. However, the emergence of variants that escape from immunity challenges the effectiveness of current vaccines. Given this continuing evolution, an important question is when and how to update SARS-CoV-2 vaccines to antigenically match circulating variants, similar to seasonal influenza viruses where antigenic drift necessitates periodic vaccine updates. Here, we studied SARS-CoV-2 antigenic drift by assessing neutralizing activity against variants-of-concern (VOCs) of a unique set of sera from patients infected with a range of VOCs. Infections with ancestral or Alpha strains induced the broadest immunity, while individuals infected with other VOCs had more strain-specific responses. Omicron was substantially resistant to neutralization by sera elicited by all other variants. Antigenic cartography revealed that all VOCs preceding Omicron belong to one antigenic cluster, while Omicron forms a new antigenic cluster associated with immune escape and likely requiring vaccine updates to ensure vaccine effectiveness.

Reductions in neutralizing activity against Omicron were substantial in all groups. Omicron neutralization dropped below the limit of detection (2 IU/mL or an ID50 of 20) in 21/51 of the studied individuals. The fold-reduction of Omicron neutralization versus homologous neutralization was 9-fold (range 0.5 to 93-fold) when considering all patients, 10-fold (range 3 to 93-fold) for patients infected with ancestral strains, 56-fold (range 0.5 to 89-fold) for Alpha, 6-fold (range 1 to 22-fold) for Gamma, and 5-fold (range 1 to 51-fold) for Delta infected patients.

To explore the antigenic relationships between the VOCs, we used the neutralization data to construct a SARS-CoV-2 antigenic map (Fig. 2a) . In this map, homologous sera tend to cluster around the infecting strain, reflecting that homologous neutralization is dominant. The ancestral and Alpha viruses cluster tightly together in the centre of the map, while the Beta, Gamma, and Delta variants all lie within 2 antigenic units (1 unit = 2-fold change in neutralization titre) of the ancestral virus suggesting a high degree of antigenic similarity. For influenza viruses, variants are considered to be antigenically similar in case of antigenic distances below 3 antigenic units, i.e. an 8-fold change in neutralization titre, and different when above this threshold 10, 11 . By analogy the ancestral, Alpha, Beta, Gamma and Delta variants belong to one antigenic cluster. However, the distance between this antigenic cluster and Omicron is more than 5 antigenic units (>32-fold change in neutralization) implying that Omicron represents a first widely circulating new major SARS-CoV-2 variant (Fig. 2a) . One caveat is that it is unclear whether 2-fold changes in pseudovirus neutralization titres are directly comparable to 2-fold changes in hemagglutination inhibition assay titres used to characterize influenza viruses. However, the change in neutralization between Omicron and other variants of SARS-CoV-2, including the ancestral strain, is striking.

We next used neutralization data from sera of 122 COVID-19 naïve vaccinees receiving either two Moderna (mRNA-1273, n=30), Pfizer/BioNTech (BNT162b2, n=49), or AstraZeneca (AZD1222, n=30) vaccines, which are all based on the ancestral S sequence to generate a second antigenic map 4 . This map agreed well with the infectee map (Fig. 2b) , and corroborated that Omicron represents a distinct antigenic variant from viruses currently included in vaccines. Interestingly, while the distributions of sera from recipients of different vaccines overlap, there is a skew of sera of mRNA-1273 vaccinees towards Omicron suggesting small differences in antigen stimulation among vaccine formulations considered here.

We have started to define the antigenic SARS-CoV-2 landscape after two years of antigenic drift, which should inform risk assessment of re-infections as well as strain selection for COVID-19 vaccine updates. We can draw several conclusions. First, homologous neutralization was usually stronger than heterologous neutralization. Second, heterologous responses were broadest and most potent in individuals infected with Alpha and ancestral strains, while infection with Delta resulted in narrow-specificity responses. Third, crossneutralization of Omicron was weak across all samples, with few exceptions. Moreover, the ancestral and Alpha strains are at the centre of our antigenic map, which supports the use of the current COVID-19 vaccines based on the ancestral strain, in case of circulation of the Alpha, Beta, Gamma and Delta variants. Our data suggest that updated vaccines based on the Beta or Delta variants would not have been appreciably more effective than the ancestral virus-based vaccines. However, the substantial reduction of neutralization in all groups against the Omicron variant indicates a high risk of re-infections and vaccine breakthrough cases when exposed to this VOC. The long antigenic distance between Omicron and the preceding variants in the antigenic map indicates that the current high rates of Omicron infections are at least partially associated with immune escape and that a vaccine update is required.

As in the case of seasonal influenza viruses, the prospect of SARS-CoV-2 becoming an endemic virus with recurring outbreaks implies the need for surveillance of antigenic drift and yearly administration of updated vaccines, especially for individuals at risk for severe COVID-19. Antigenic cartography efforts such as those presented here, can inform future vaccine updates. All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted January 3, 2022. Figure 1. Homologous and cross-neutralization. a. Midpoint neutralization titres against the VOCs in International Units per mL (IU/mL). The individuals are grouped per VOC they were infected with and plotted accordingly. Non-hospitalized patients (dots) showed lower neutralization response compared to hospitalized patients (triangles) as in line with previous research 12 . A Mann-Whitney test is used to test for differences between group medians (line). ns= non-significant, * = p< 0.005, **** = p<0.0001. b. Cross-neutralization is expressed as the geometric mean of the neutralization titres against all VOCs except the autologous strain in IU/mL. A cut-off of 10IU/mL is used for all neutralization titres, as indicated by the grey bar. All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted January 3, 2022. ;

Study population 51 adults with a PCR proven primary SARS-CoV-2 infection were included in the COSCAstudy or the RECOVERED study between June 2020 and July 2021 at Amsterdam UMC and via the Dutch national SARS-CoV-2 sequence surveillance program as described previously 13, 14 . 3-9 weeks after symptom onset, blood, patient demographics, time between symptom onset and sampling, and admission status were collected. The diagnostic oropharyngeal swabs were used to determine the SARS-CoV-2 strain causing the infection. All ancestral strain infected participants (n=20) were assumed to be infected with the ancestral strain as they were sampled before the emergence of any VOC in the Netherlands. Two of the individuals infected with Alpha strains harboured the E484K mutation, which are indicated as squares in all graphs. The Omicron variant did not yet circulate at time of inclusion.

Neutralization data on COVID-19 naive vaccinee sera collected in the S3-study, were kindly provided by van Gils et al 4 . Post-vaccination sera was obtained four weeks after the second doses of either Moderna (mRNA-1273), Pfizer/BioNTech (BNT162b2), or AstraZeneca (AZD1222). Post-vaccination serum after Janssen (Ad26.COV2.S) were excluded from analysis because they did not have enough non-threshold titres to be included in the map.

The COSCA study, the RECoVERED study and the S3-study were conducted at the Amsterdam University Medical Centres, the Netherlands, and approved the medical ethical review board of the Amsterdam University Medical Centres (NL 73281.018.20, NL73759.018.20, NL73478.029.20, respectively). All individuals provided written informed consent before participating.

The Alpha, Beta, Gamma, Delta and Omicron pseudovirus constructs contained the following mutations: deletion (Δ) of H69, V70 and Y144, N501Y, A570D, D614G, P681H, T716I, S982A, and D1118H in Alpha; L18F, D80A, D215G, L242H, R246I, K417N, E484K, N501Y, D614G, and A701V in Beta; and L18F, T20N, P26S, D138Y, R190S, K417T, E484K, N501Y, D614G, H655Y, and T1027I in Gamma; T19R, G142D, E156G, Δ157-158, L452R, T478K, D416G, P681R and D950N in Delta; and A67V, Δ69-70, T95I, G142D, Δ143-145, Δ211, L212I,  ins214EPE, G339D, S371L, S373P, S375F, K417N, N440K, G446S, S477N, T478K, E484A,  Q493K, G496S, Q498R, N501Y, Y505H, T547K, D614G, H655Y, N679K, P681H, N764K,  D796Y , N856K, Q954H, N969K, L981F in Omicron. The ancestral strain included the D614G mutation as this variant emerged early in the pandemic and all ancestral infected participants are most likely to be infected with this variant. The spike constructs were ordered as gBlock gene fragments (Integrated DNA Technologies) and cloned SacI and ApaI in the pCR3 SARS-CoV-2-S Δ19 expression plasmid (GenBank: MT449663.1) using Gibson Assembly (Thermo Fisher Scientific). The pseudovirus constructs were made using the QuikChange Site-Directed Mutagenesis Kit (Agilent Technologies) and verified by using Sanger sequencing. Pseudoviruses were procedures by cotransfecting HEK293T cells (American Type Culture Collection, CRL-11268) with the pCR3 SARS-CoV-2-SΔ19 expression plasmid and the pHIV-1NL43 ΔEnv-NanoLuc reporter virus plasmid. 48 hours after the transfection, cell supernatants containing the pseudovirus were harvested and stored at -80 °C until further use.

The pseudovirus neutralization assay was performed as described previously 2 . Shortly, HEK293T/ACE2 cells were kindly provided by P. Bieniasz 15 were seeded at a density of 20,000 cells per well in a 96-well plate coated with poly-lysine (50 g/ml) 1 day before the start of the neutralization assay. The next day, heat-inactivated sera samples were serial diluted in threefold steps, starting at 1:20 dilution to test for Omicron pseudovirus neutralization and 1:100 for the other variants and the ancestral pseudovirus. Sera was diluted in cell culture medium (DMEM), supplemented with 10% fetal bovine serum, penicillin (100 U/ml), streptomycin (100 g/ml), and GlutaMAX (Gibco), mixed in a 1:1 ratio with pseudovirus, and incubated for 1 hour at 37°C. These mixtures were then added to the cells in a 1:1 ratio and incubated for 48 hours at 37°C, followed by a PBS wash, and lysis buffer was added. The luciferase activity in cell lysates was measured using the Nano-Glo Luciferase Assay System (Promega) and GloMax system (Turner BioSystems). Relative luminescence units were normalized to those from cells infected with SARS-CoCV-2 pseudovirus in the absence of sera. The inhibitory neutralization titres (ID50) were determined as the serum dilution at which infectivity was inhibited by 50%, using a nonlinear regression curve fit (GraphPad Prism software version 8.3). The International Standard for anti-SARS-CoV-2 immunoglobulins provided by the WHO 16 were used to convert the ID50 values into International Units per milliliters (IU/mL). Samples with IU/mL titres <10 were defined as having undetectable neutralization against the ancestral, Alpha, Beta, Gamma and Delta variant. For Omicron neutralization, the start-dilution of 1:20 enables a cut-off of <2 IU/mL.

Antigenic maps were constructed as previously described 6,7 using the antigenic cartography software from https://acmacs-web.antigenic-cartography.org. In brief, this approach to antigenic mapping uses multidimensional scaling to position antigens (viruses) and sera in a map to represent their antigenic relationships. The maps there relied on the first SARS-CoV-2 infection serology data and post-vaccination serology data for the maps shown in Figure 2a and Figure 2b . The positions of antigens and sera were optimized in the map to minimise the error between the target distances set by the observed pairwise virus-serum combinations in the pseudovirus assay described above and the resulting computationally derived map. Maps were constructed in 2, 3, 4, and 5 dimensions to investigate the dimensionality of the antigenic relationships. Both the infectee and vaccinee datasets were strongly two-dimensional with only small improvements in residual mean squared error of the maps as map dimensionality increased.

Data visualization and statistical analyses were performed in GraphPad Prism software (version 8.3). Mann-Whitney test was used for non-paired group comparisons. Wilcoxon signed rank test with Benjamini Hochberg correction was used to compare cross-neutralization titres with the homologous neutralization. Spider plots were made in Excel 2016. Fold-changes were calculated using a cut-off of 10 IU/mL for all VOCs. The antigenic maps were produced using the antigenic cartography software mentioned above. 

World Health Organization. Tracking SARS-CoV-2 variants

Emerging SARS-CoV-2 variants of concern evade humoral immune responses from infection and vaccination

mRNA-based COVID-19 vaccine boosters induce neutralizing immunity against SARS-CoV-2 Omicron variant. medRxiv

Four SARS-CoV-2 vaccines induce quantitatively different antibody responses against SARS-CoV-2 variants. medRxiv

Reduced Neutralization of SARS-CoV-2 Omicron Variant by Vaccine Sera and monoclonal antibodies. medRxiv

Mapping the Antigenic and Genetic Evolution of Influenza Virus

Antibody landscapes after influenza virus infection or vaccination

Reduced neutralization of SARS-CoV-2 B.1.617 by vaccine and convalescent serum

Antibody responses to SARS-CoV-2 in patients with COVID-19

WHO recommendations for the viruses used in the

Northern Hemisphere influenza vaccine: Epidemiology, antigenic and genetic characteristics of influenza A(H1N1)pdm09, A(H3N2) and B influenza viruses collected from

Magnitude and Kinetics of Anti-Severe Acute Respiratory Syndrome Coronavirus 2 Antibody Responses and Their Relationship to Disease Severity

Cross-reactive antibodies after SARS-CoV-2 infection and vaccination. medRxiv

Evolution of COVID-19 symptoms during the first 12 months after illness onset

Measuring SARS-CoV-2 neutralizing antibody activity using

We thank all public health services (GGD) in the Netherlands for their help in contacting participants. We are also thankful to the study personnel and the participants of the COSCA, RECoVERED and the S3-study for their contribution to this research. *** * *** ** ** **