Article

ISSN 0006-2979, Biochemistry (Moscow), 2024, Vol. 89, No. 5, pp. 862-871 © Pleiades Publishing, Ltd., 2024.

862

REVIEW

Antigenic Cartography of SARS-CoV-2

Ekaterina A. Astakhova

1,2,a

*, Alexey A. Morozov

1,2

, Julia D. Vavilova

and Alexander V. Filatov

1,2

National Research Center Institute of Immunology, Federal Medical Biological Agency of Russia,

115522 Moscow, Russia

Department of Immunology, Faculty of Biology, Lomonosov Moscow State University,

119234 Moscow, Russia

Shemyakin–Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences,

117997 Moscow, Russia

e-mail: ast_kat@mail.ru

Received November 7, 2023

Revised January 14, 2024

Accepted January 17, 2024

Abstract— Antigenic cartography is a tool for interpreting and visualizing antigenic differences between virus

variants based on virus neutralization data. This approach has been successfully used in the selection of influen-

za vaccine seed strains. With the emergence of SARS-CoV-2 variants escaping vaccine-induced antibody response,

adjusting COVID-19 vaccines has become essential. This review provides information on the antigenic differences

between SARS-CoV-2 variants revealed by antigenic cartography and explores a potential of antigenic cartogra-

phy-based methods (e.g., building antibody landscapes and neutralization breadth gain plots) for the quantita-

tive assessment of the breadth of the antibody response. Understanding the antigenic differences of SARS-CoV-2

and the possibilities of the formed humoral immunity aids in the prompt modification of preventative vaccines

against COVID-19.

DOI: 10.1134/S0006297924050079

Keywords: antigenic cartography, SARS-CoV-2, COVID-19, breadth of virus-neutralization

Abbreviations: AU,antigenic unit; WT,wild type.

* To whom correspondence should be addressed.

INTRODUCTION

Since the beginning of the COVID-19 pandemic,

more than 40 variants of SARS-CoV-2 have emerged,

of which more than 10 belong to the Omicron family

(https://www.who.int/activities/tracking-SARS-CoV-2-

variants, https://covariants.org/, 01.09.2024). SARS-CoV-2

variants differ in contagiousness, nature of infection

caused, and some other epidemiological characteristics

[1-3]. There are several ways to classify viral variants.

The most common is construction of phylogenetic trees

of viral variants and isolates based on sequencing data

[4]. The phylogenetic trees reveal the evolution of viral

variants and routes of their spread. However, such clas-

sification provides only an indirect information with re-

gard to antigenicity [5]. Obviously, not all mutations are

equally important. The most intriguing of them are mu-

tations in the surface spike protein (S protein) located

within the domains directly involved in the interaction

with host cell receptors or in the virus fusion with the

host cell. The most important mutations are those allow-

ing the virus to escape neutralizing antibodies formed

after vaccination or previous infections. These muta-

tions constitute the antigenic profile of a viral variant.

Quantification of antigenic differences presents

certain difficulties. The arithmetic sum of mutations

cannot be a measure of antigenic differences. Some of

mutations are more immunogenic, while others are

less; they may occur in functionally important regions

of the antigens or happen far from them. This problem

can be solved using the method of antigenic mapping

that was first applied in the study of influenza virus

[6] and is now increasingly used to monitor the emer-

gence of new SARS-CoV-2 variants and to assess their

antigenicity. Here, we reviewed the published data on

the antigenic properties of SARS-CoV-2 obtained using

antigenic maps.

ANTIGENIC CARTOGRAPHY OF SARS-CoV-2 863

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

STEPS IN CONSTRUCTING ANTIGENIC MAPS

Determining the antigenic characteristics of the

virus using antigenic maps begins with serological

tests [7]. While hemagglutination inhibition (HI) test is

commonly used for the influenza virus, the gold stan-

dard for SARS-CoV-2 is the live virus neutralization test

(cVNT, conventional virus-neutralization test). Due to

the increased safety requirements for working with

a live virus, cVNT is often replaced with a test using

virus-like particles pseudotyped with the coronavirus

S protein (pVNT, pseudovirus-based virus-neutralization

test). A high level of correlation between the cVNT and

pVNT results has been repeatedly shown [7-10]. These

tests provide a rich set of multidimensional data that,

due to their large volume, are difficult to interpret.

Biological science has previously encountered sim-

ilar multidimensional problems, for example, when

interpreting transcriptomics data or data obtained by

multicolor flow cytometry [11-13]. The t-SNE (t-sto-

chastic neighborhood embedding) and UMAP (uniform

manifold approximation and projection) algorithms

were developed to reduce the dimensionality of data

[14, 15], which made it possible to represent multidi-

mensional data on a two-dimensional plane. The ad-

vantage of these algorithms is preservation of prox-

imity between the points by taking into account the

distances between them in a multidimensional space.

Antigenic maps are constructed using a method based

on the principle of multidimensional scaling (MDS)

[16], which in its algorithms is close to the traditional

principal component analysis (PCA). The advantage of

MDS is that the distance between the points visualized

on a plane closely matches the distance between them

in the multidimensional space.

The construction of antigenic maps begins with

a creation of tables with sera shown in the columns,

virus variants – in the rows, and neutralization ti-

ters – at the intersection. Next stage, neutralization

titers are converted into antigenic distances (https://

acorg.github.io/Racmacs/articles/intro-to-antigenic-

cartography.html, 09.04.2023). The higher the serum neu-

tralization titer relative to the antigen, the smaller the

antigenic distance between them. Next, the MDS meth-

od is applied and the data are visualized on aplane.

Antigenic maps display both antigens (virus vari-

ants) and sera. The distance between a serum and an

antigen directly depends on the antigenic distance.

Variants with similar antigenic properties are located

next to each other, forming a cluster. For example, the

wild-type (WT) coronavirus co-clusters with the Alpha,

Beta, and Gamma variants [17], but, as a rule, is far

removed from the Omicron variants (Fig.1).

The antigenic evolution of a virus differs from its

genetic evolution, since different mutations make an

unequal contribution to changes in its antigenic char-

acteristics. For example, the E484K/Q mutation is the

most significant for the antigenic differences between

the Beta, Gamma, Kappa, and Mu variants relative to

WT, while the L452R/Q mutation is decisive for the

Delta, Epsilon, and Lambda variants, although other

mutations are also present in these variants [18, 19].

Significant changes in the antigenic properties of

a new virus variant lead to the inability of antibodies

that neutralize the previous variant to neutralize the

new one. In this case, first, the new variant will be lo-

cated at a great distance from the previous variants on

the antigenic map. Second, there will be no sera locat-

ed on the map between the studied variants.

PREVIOUS EXPERIENCE OF THE IMMUNE

SYSTEM INFLUENCES THE POSITION

OF A VIRAL VARIANT ON THE ANTIGENIC MAP

At first glance, it may seem that the location of

viral variants on the antigenic map depends solely on

the number and quality of amino acid substitutions.

However, this is not entirely true. Prior exposure of

the immune system to either SARS-CoV-2 infection

or vaccination can significantly change the degree of

neutralization of viral variants and, therefore, influ-

ence the appearance of antigenic maps. Indeed, sera

from infected laboratory animals or recovered people,

vaccinated volunteers, and individuals with the hybrid

immunity ultimately produce different antigenic maps

[20-22]. Thus, the position of a viral variant on the anti-

genic map, and, consequently, the distance between the

variants, depend not only on the structure or compo-

sition of their immunodominant epitopes, but also on

the source of sera that was used to construct the map.

Fig. 1. Example of a SARS-CoV-2 antigenic map (adapted

from[18]). Circles, locations of individual antigens; squares,

locations of individual sera; dotted lines, variants with com-

mon specified substitutions.

ASTAKHOVA et al.864

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

QUANTITATIVE ESTIMATION OF ANTIGENIC

DISTANCES BETWEEN VIRUS VARIANTS

Antigenic maps are widely used to make decisions

about changing the strain of the influenza virus vaccine.

Smith etal. [6] retrospectively analyzed the evolution of

the influenza virus from 1968 to 2003 and showed that

vaccine strains changed when the new variant was 2 or

more antigenic units (AU) away from the previous vac-

cine strain, which was used to define the “sufficiency”

of antigenic differences between the viruses for updat-

ing a vaccine strain. For SARS-CoV-2, this criterion has

not yet been clearly determined. However, after con-

structing the first maps with Omicron BA.1, it was no-

ticed it was located “like a lonely island in the middle of

the ocean” on the antigenic map of SARS-CoV-2 (https://

spectrum.ieee.org/omicron-covid-variant#toggle-gdpr,

09.04.2023), i.e., was far from the pre-Omicron variants.

This observation provided a good explanation for the

spread of breakthrough infections among vaccinated

donors and prompted consideration of the necessity to

update the vaccine strain.

At present, significant antigenic differences be-

tween the WT coronavirus and Omicron family vari-

ants cause no doubt. These differences range from 3 to9

AU depending on the variant [18, 20-25], where one

AU corresponds to a twofold serum dilution. The most

distant from D614G are the variants BQ.1.1, BM.1.1.1

and XBB.1, XBB.2, BN.1.3.1 [22, 25]. The distance be-

tween these variants and the D614G is more than 6 AU.

The variants of the Omicron family are antigenically

distant not only from pre-Omicron variants, but also

from each other [22, 25, 26]. For example, XBB.1 and

BM.1.1.1 are located at a distance of about 6 AU from

each other on the Omicron antigenic map [25].

Unfortunately, exact determination of antigenic

distances is not a straightforward procedure. Thean-

tigenic distances between the same variants deter-

mined in different studies may differ (table) likely due

to different methods for obtaining the samples. Stan-

dardization of serum collection procedure and testing

conditions can lead to a more precise quantification of

antigenic variances.

APPLICATION OF ANTIGENIC MAPPING

FOR DETERMINING THE OPTIMAL

VACCINE STRAIN

Once it is discovered that a new variant is anti-

genically distant from the previous one, the question

arises of which particular variant should be used for

Antigenic distances between D614G and SARS-CoV-2 variants

No. Serum source

Antigenic distance between D614G and variant

References

Alpha Beta Delta BA.1 BA.2 BA.4/5

Syrian hamsters, 14 days after

double i.n. SARS-CoV-2 administration

0.8 1.1 0.7 4.4 3.7 4.8 [22]

Syrian hamsters, 26 days after

i.m. SARS-CoV-2 injection

0.8 1.3 1.7 4.7 4.0 ND [18]

Syrian hamsters, 26 days after

i.m. SARS-CoV-2 injection

1.4 0.6 1.5 6.2 5.6

5.0

(BA.5)

[25]

4 Recovered donors 0.9 2.0 1.2 4.7 4.0 ND [21]

Vaccinated (mRNA-1273,

BNT162b or ChAdOx-S1) donors

0.4 2.2 1.7 5.9 ND ND [21]

6 Recovered donors 0.6 3.7 1.4 7.0 6.3 5.8 [20]

Recovered

or vaccinated (mRNA-1273) donors

0.5 3.8 2.2 5.6 5.3 6.7 [19]

Recovered or vaccinated

(mRNA-1273 ×2*, BNT162b ×2,

AZD1222 ×2 or AZD1222/BNT162b) donors

0.8 2.1 1.5 5.5 3.3

4.2

(BA.5)

[27]

Note: i.n., intranasal; i.m., intramuscular injection; ND,not determined; *double homologous vaccination.

ANTIGENIC CARTOGRAPHY OF SARS-CoV-2 865

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

the next vaccine [28]. For this purpose, a “basic” an-

tigenic map is built. Laboratory animals (most often,

Syrian hamsters or mice) are infected with a potential

vaccine strain, and the extent to which the resulting

sera cross-neutralize the variants of interest is deter-

mined.

It was shown [17] that sera from mice infected

with the Gamma variant also neutralized the WT and

Beta variants (these sera are located between the WT,

Beta, and Gamma antigens on the antigenic map).

Atthe same time, sera from the mice infected with the

Beta variant did not neutralize WT and Gamma.

The focus of current research is to identify a suit-

able SARS-CoV-2 vaccine strain from the Omicron fami-

ly. In [29], K18-hACE2 sera from the mice infected with

BA.1 neutralized BA.1 to the greatest extent, but also

showed broad cross-reactivity with BA.2 and BA.2.12.1.

Sera from the mice infected with BA.5 neutralized BA.2

and BA.2.12.1 and, to a lesser extent, BA.4.6 and BA.5

and were located on the antigenic map mainly close

to BA.2.12.1, being distant from BA.5 at 1 or more AU.

Another study showed that sera from the BA.5-infected

hamsters neutralized BA.2 and BQ.1.1 and, to a lesser

extent, BM.1.1.1. None of the sera neutralized XBB.1

[25]. Despite the antigenic similarity of BA.5 and

BQ.1.1, the authors questioned whether inclusion of

BA.5 in a bivalent vaccine would result in a sufficient

level of cross-response against BQ.1.1 due to the fact

that the two variants do not cluster with each other

(distance of more than 3AU).

As confirmed in [30-32], immunization of people

with a bivalent vaccine based on BA.5 did not induce

significant neutralization of BA.2.75.2, BQ.1.1, and XBB.1

variants.

Currently, SARS-CoV-2 variants of interest are

XBB.1.5, XBB.1.16, BA.2.86, and some others (https://

www.who.int/activities/tracking-SARS-CoV-2-variants/

tracking-SARS-CoV-2-variants, 01.09.2024). A number

of studies have shown a relatively close location of

BA.2.86 and XBB.1.5 (distance of approximately 0.5-

4AU depending on the source of sera) [33-35]. Serum

antibodies raised against XBB.1.5 cross-reacted with

BA.2.86, supporting inclusion of XBB.1.5 in the updat-

ed vaccine.

COMPARISON OF “BASIC” ANTIGENIC MAPS

WITH MAPS BUILT WITH SERA

FROM VACCINED DONORS

By considering “basic” maps as a gold standard

for determining virus antigenicity, we assume that

the antibody response in humans and laboratory an-

imals is generated in a similar way, although this is

not entirely true. “Basic” antigenic maps can also be

constructed using sera from donors who had been

previously infected with a known SARS-CoV-2 variant.

However, such sera will become increasingly difficult

to obtain due to the increasing spread of hybrid immu-

nity in the population. That is why it is important to

build antigenic maps using the data from vaccinated

donors and donors with the hybrid immunity and then

to compare them with the “basic” maps.

In the study of the “pre-omicron” period [36],

comparison of antigenic maps from the donors who

had recovered from D614G or one of the SARS-CoV-2

variants, including the L452R variant, with the maps

of those vaccinated with mRNA-1273 revealed the fol-

lowing differences. The distance between D614G and

AY.1 (Delta family) on the map for the vaccinated in-

dividuals was greater than on the map for those who

had recovered from the disease. D614G and Beta, on

the contrary, were closer to each other. The Beta and

AY.1 variants were located at a distance of about 1AU

on the map for the vaccinated individuals, while on

the map of those who had recovered from the disease,

they were more than 4AU apart. The authors discov-

ered a general trend that the studied variants D614G,

Alpha, Beta, Kappa, Lambda, Delta, and others were

located further from each other on the map for the

recovered donors than on the map for vaccinated do-

nors. The same observation was made in a more re-

cent study that included donors who had recovered

from COVID-19, including the Omicron BA.1 and BA.2

variants, as well as those who had received the mRNA-

1273, BNT162b2, or AZD1222 vaccines [21]. On the

map for the vaccinated individuals, D614G, Alpha, and

Gamma were located almost in the same place and

formed a cluster, while on the map of those who had

recovered, the distance between these variants was

about 1 AU. Another difference was location of sera

on the maps. In the map of vaccinated individuals, the

sera were located close to the D614G/Alpha/Gamma

cluster only. On the map of those who had recovered

from COVID-19, the sera were located closer to the

variant that the donor had suffered from.

The study [21] also revealed differences in the

positions of sera from the donors vaccinated with dif-

ferent vaccines. We noticed that the sera located near

the D614G/Alpha/Gamma cluster, but closer to Omicron

BA.1, were predominantly from the donors who re-

ceived the mRNA-1273 vaccine. The sera from those

vaccinated with BNT162b2 or AZD1222 were located near

the D614G/Alpha/Gamma cluster and away from BA.1.

ANTIGENIC MAPS USED TO ASSESS

THE DEVELOPMENT OF CROSS-REACTIVE

ANTIBODIES AFTER BOOSTER VACCINATION

Comparison of antigenic maps before and after vac-

cination allows to monitor the formation of antibody

ASTAKHOVA et al.866

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

cross-reactivity. Booster vaccination leads to changes

in the antigenic distances between the variants [20,

24, 37]. As an antigen used for serum testing remains

constant, any variations in the antigenic distances be-

tween the variants are linked to changes in the compo-

sition of neutralizing antibodies in the sera. A decrease

in the distance between two antigens may indicate

that the serum contains more antibodies that bind to

both antigens [38].

A number of studies have shown that the third

booster vaccination with BNT162b2 leads to a decrease

in the antigenic distance between D614G and BA.1,

BA.2, BA2.12.1, and BA.4/5 [20, 24]. The same studies

showed that after the third immunization with an

mRNA vaccine, the antigenic distances between D614G

and Delta, on the contrary, increased. We demonstrat-

ed that homologous booster vaccination with Sputnik-V

leads to a decrease in the antigenic distance between

WT and BA.1. We also found oppositely directed

changes in the distances between WT and Delta after

homo- and heterologous revaccination. Revaccination

with Sputnik-V led to a decrease in the antigenic dis-

tance between WT and Delta, while vaccination with

BNT162b2 increased this distance. These discrepan-

cies may be due to the fact that Sputnik-V induces

immune response to the full-length S protein, while

BNT162b2– to the pre-fusion stabilized conformation

of the S protein [39-41].

Antigenic maps allow to quantify changes in the

antigenic distances between the viral variants before

and after vaccination. We proposed a convenient way

to visualize these changes using the hat graphs [42].

The edge of the hat on such graph corresponds to the

initial indicator value, i.e., to the antigenic distance be-

tween variants on the antigenic maps before the expo-

sure. The height of the hat corresponds to the change

in the antigenic distance, which can be either positive

or negative (in the latter case the hat will be upside

down). Using hat graphs, we visualized changes in the

antigenic distances between WT and Alpha, Beta, Delta,

Omicron BA.1, and BA.4/5 of the sera from volunteers

revaccinated with Sputnik-V or BNT162b2 after pri-

mary vaccination with Sputnik-V [38]. The largest de-

crease in the antigenic distance occurred between WT

and BA.1 for both types of booster vaccination.

ANTIBODY LANDSCAPES

The breadth of antibody response could be eval-

uated using an approach called antibody landscape

modeling [43]. Antibody landscape represents a three-

dimensional surface, where the xy plane is the “basic”

antigenic map, and the height of the landscape (z-axis)

is determined by the neutralization titer of sera of a

studied group against a specific antigen. The smoothed

surface is constructed using multiple linear regression.

This creates an immunological profile of the sera with

elevations corresponding to the areas on the antigenic

map with higher levels of antibodies (Fig. 2).

Antibody landscapes constructed based on the

data for cross-reactive sera are flatter, with less tilt

Fig. 2. Example of antibody landscapes (adapted from [32]).

ANTIGENIC CARTOGRAPHY OF SARS-CoV-2 867

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

towards one or another antigenic region (top land-

scape in Fig. 2). Antibody landscapes of individuals

who had received two doses of mRNA-1273, BNT162b2,

and AZD1222 vaccines, as well as of those who had

been previously infected with the D614G, Alpha, and

Beta variants, exhibit a distinct decline towards the

Omicron cluster. This decline is attributed to the re-

duced neutralization titers of the Omicron variants

compared to the pre-Omicron variants [19, 27]. In-

terestingly, the landscapes exhibited a flatter profile

three months following double and triple booster with

mRNA-1273, as compared to those obtained after one

month [19], indicating that the decrease in the neutral-

ization titer occurs at different rates – titers against

pre-Omicron variants decreased faster, which may be

due to rapid activation of memory B cells specific for

pre-Omicron variants after booster immunization. The

presence of specific antibodies produced by the plas-

ma cells and activated memory B cells inhibits the de-

velopment of naïve B cells with a similar specificity,

on particular, through direct masking of dominant re-

ceptor-binding domain (RBD) epitopes [44]. However,

naïve B cells do appear to be activated [45] and, during

maturation in the germinal center, acquire a different

specificity useful for neutralization of Omicron vari-

ants. By this moment, antibody secretion by memory

B cells has already declined, so we see flatter antibody

landscapes only 3 months after vaccination.

Flatter antibody landscapes were also observed

in individuals with the hybrid immunity who have re-

ceived bivalent (BA.1 or BA.4/5) vaccines, in contrast to

the donors who had not been previously infected [32].

ASSESSING THE BREADTH

OF NEUTRALIZATION USING

ANTIGENIC CARTOGRAPHY

The concept of breadth of virus neutralization

was introduced in the studies on the effectiveness of

humoral immunity against several variants of SARS-

CoV-2. This parameter shows how efficiently related

and more distant variants of the virus are neutralized.

As a rule, the breadth of virus neutralization is as-

sessed qualitatively. However, we believe that antigen-

ic cartography allows to quantify the breadth of virus

neutralization.

Projecting 3D antibody landscapes on a plane

produces a graph of the increase in the neutralization

breadth [20]. The antigens are placed along the x-axis

according to their antigenic distance (WT is typically

taken as the reference point). The y-axis represents

the neutralization titers of these antigens (Fig. 3). This

type of graph allows to evaluate changes in the shape

and area of the immune profile, which is especially

useful when studying the dynamics of antibody re-

sponse over a long period of time or after several vac-

cinations.

The changes in the shape of the immune profile

on the breadth-gain plot can be assessed only qualita-

tively. It was shown that after triple vaccination with

BNT162b2 or vaccination associated with BA.1 infec-

tion, there was a greater increase in the neutralization

titers against variants antigenically distant from the

vaccine strain [20].

Unlike shape, the area of the immune profile is a

parameter that can be calculated quantitatively. How-

ever, a number of features should be taken into ac-

count. First, for the reasons described above, precise

positioning of antigens on the x-axis is difficult. Sec-

ond, the area will depend on the number of antigens

against which the sera were tested. The two problems

can be solved by standardization of methods used for

the determination of the virus neutralization titers.

Finally, it is important to consider that the greater the

distance from the starting point along the x-axis, the

more significant the contribution to the virus neutral-

ization indicator. In this regard, to evaluate the area

of each site, it is necessary to introduce a certain in-

creasing factor, which, for example, for site B in Fig.3

will be greater than for site A. As far as we know, a

model for calculating the breadth of virus neutraliza-

tion based on a graph of this type has not yet been de-

veloped.

CONCLUSION

Construction of antigenic maps has been tradition-

ally used to choose a vaccine strain against the influen-

za virus [28] (https://www.who.int/publications/m/item/

recommended-composition-of-influenza-virus-vaccines-

Fig. 3. Example of a graph of increasing neutralization

breadth upon homo- and heterologous revaccination

(adapted from [37]).

ASTAKHOVA et al.868

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

for-use-in-the-2023-2024-northern-hemisphere-influenza-

season, 09.04.2023). Typically, antigenic maps confirm

the findings of serological tests (e.g., virus neutral-

ization) and are used to facilitate visualization of the

obtained results. Thus, antigenic maps have clearly

demonstrated that Omicron BA.1 and other variants of

this family are antigenically distant from the ancestral

variants of SARS-CoV-2 [19-25]. World Health Organiza-

tion has incorporated these findings to support the ne-

cessity for the modification of coronavirus vaccine in

order to accommodate new circulating strains (https://

www.who.int/news/item/18-05-2023-statement-on-the-

antigen-composition-of-covid-19-vaccines, 04.09.2023).

When choosing a new vaccine strain, it is crucial

to select a strain whose antigenic characteristics would

be significantly different from characteristics of the

previous one. This helps reduce the impact of the “orig-

inal sin,” when previously generated memory Bcells

are preferentially activated in response to a new anti-

gen instead of naïve B cells, which could produce more

specific antibodies to the new antigen [45].

Using antigenic cartography, it was shown that

variants of the Omicron family significantly differ in

their antigenic characteristics not only from the an-

cestral forms of SARS-CoV-2, but also from each other.

Analysis of antigenic maps revealed that infection

with BA.5 produces an insufficient neutralizing re-

sponse against BA.4.6, BA.5, XBB.1, and BQ.1.1 [25, 29].

Thus, bivalent BA.5-based vaccines were insufficiently

effective against new variants [30-32]. The antigenic

similarity of XBB.1.5 with the modern variant BA.2.86

that was shown using the antigenic maps, supports

inclusion of XBB.1.5 in the updated vaccine [33-35].

Using antigenic characterization of the new variants,

Food and Drug Administration (FDA) has recommend-

ed the development of a monovalent vaccine based on

XBB.1.5 (https://www.fda.gov/vaccines-blood-biologics/

updated-covid-19-vaccines-use-united-states-beginning-

fall-2023, 01.09.2024).

Antigenic cartography and construction of anti-

body landscape allow to predict which viral variant

can evade the formed immunity, i.e., to get ahead of

the virus evolution and to prepare the vaccines in ad-

vance. Finally, antigen mapping provides an oppor-

tunity to quantitatively measure the breadth of virus

neutralization by the sera.

Contributions. A.E.A. prepared the manuscript;

A.A.M., Y.D.V., and A.V.F. edited the text.

Funding. This work was financially supported

by the Russian Science Foundation (project no. 23-25-

00472).

Ethics declarations. This work does not contain

any studies involving human and animal subjects.

Theauthors of this work declare that they have nocon-

flicts of interest.

REFERENCES

1. Sun,C., Xie,C., Bu, G.L., Zhong, L.Y., and Zeng, M.S.

(2022) Molecular characteristics, immune evasion,

and impact of SARS-CoV-2 variants, Signal Transduct.

Target Ther., 7, 202, doi:10.1038/s41392-022-01039-2.

2. Tian, D., Sun, Y., Xu, H., and Ye, Q. (2022) The emer-

gence and epidemic characteristics of the highly

mutated SARS-CoV-2 Omicron variant, J. Med. Virol.,

94, 2376-2383, doi:10.1002/jmv.27643.

3. Guo, Y., Han, J., Zhang, Y., He, J., Yu, W., Zhang, X.,

Wu,J., Zhang,S., Kong,Y., Guo,Y., Lin,Y., and Zhang,J.

(2022) SARS-CoV-2 Omicron variant: epidemiological

features, biological characteristics, and clinical sig-

nificance, Front. Immunol., 13, 877101, doi: 10.3389/

fimmu.2022.877101.

4. Li,J., Lai, S., Gao, G.F., and Shi,W. (2021) The emer-

gence, genomic diversity and global spread of SARS-

CoV-2, Nature, 600, 408-418, doi: 10.1038/s41586-

021-04188-6.

5. Komissarov, A. B., Safina, K. R., Garushyants, S. K.,

Fadeev, A. V., Sergeeva, M. V., Ivanova, A. A., Dani-

lenko, D.M., Lioznov,D., Shneider, O.V., Shvyrev,N.,

Spirin, V., Glyzin, D., Shchur, V., and Bazykin, G. A.

(2021) Genomic epidemiology of the early stages of the

SARS-CoV-2 outbreak in Russia, Nat. Commun., 12, 649,

doi:10.1038/s41467-020-20880-z.

6. Smith, D. J., Lapedes, A. S., De Jong, J. C., Bestebroer,

T.M., Rimmelzwaan, G.F., Osterhaus, A.D.M.E., and

Fouchier, R. A. M. (2004) Mapping the antigenic and

genetic evolution of influenza virus, Science, 305,

371-376, doi:10.1126/science.1097211.

7. Tan, C.W., Chia, W. N., Qin,X., Liu,P., Chen, M.I.-C.,

Tiu,C., Hu,Z., Chen, V.C.W., Young, B.E., Sia, W.R.,

Tan, Y.-J., Foo,R., Yi,Y., Lye, D.C., Anderson, D.E., and

Wang, L.F. (2020) A SARS-CoV-2 surrogate virus neu-

tralization test based on antibody-mediated blockage

of ACE2-spike protein-protein interaction, Nat. Bio-

technol., 38, 1073-1078, doi:10.1038/s41587-020-0631-z.

8. Astakhova, E. A., Byazrova, M. G., Yusubalieva,

G. M., Larichev, V. F., Baklaushev, V. P., and Filatov,

A. V. (2022) High heterogeneity of virus-neutraliz-

ing and RBD-binding activities of COVID-19 conva-

lescent sera, Mol. Biol., 56, 1095-1103, doi: 10.1134/

S002689332206005X.

9. Hyseni, I., Molesti, E., Benincasa, L., Piu, P., Casa, E.,

Temperton, N.J., Manenti,A., and Montomoli,E. (2020)

Characterisation of SARS-CoV-2 lentiviral pseudotypes

and correlation between pseudotype-based neutral-

isation assays and live virus-based micro neutralisa-

tion assays, Viruses, 12, 1011, doi:10.3390/v12091011.

10. Schmidt, F., Weisblum, Y., Muecksch, F., Hoffmann,

H. H., Michailidis, E., Lorenzi, J. C. C., Mendoza, P.,

Rutkowska,M., Bednarski,E., Gaebler,C., Agudelo,M.,

Cho, A., Wang, Z., Gazumyan, A., Cipolla, M., Cas-

key,M., Robbiani, D.F., Nussenzweig, M.C., Rice, C.M.,

ANTIGENIC CARTOGRAPHY OF SARS-CoV-2 869

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

Hatziioannou, T., and Bieniasz, P. D. (2020) Measur-

ing SARS-CoV-2 neutralizing antibody activity using

pseudotyped and chimeric viruses, J. Exp. Med., 217,

e20201181, doi:10.1084/jem.20201181.

11. Saeys,Y., Van Gassen,S., and Lambrecht, B.N. (2016)

Computational flow cytometry: Helping to make sense

of high-dimensional immunology data, Nat. Rev. Im-

munol., 16, 449-462, doi:10.1038/nri.2016.56.

12. Dorrity, M.W., Saunders, L.M., Queitsch,C., Fields,S.,

and Trapnell, C. (2020) Dimensionality reduction

by UMAP to visualize physical and genetic interac-

tions, Nat. Commun., 11, 1537, doi: 10.1038/s41467-

020-15351-4.

13. Becht,E., McInnes,L., Healy,J., Dutertre, C.A., Kwok,

I. W. H., Ng, L. G., Ginhoux, F., and Newell, E. W.

(2019) Dimensionality reduction for visualizing sin-

gle-cell data using UMAP, Nat. Biotechnol., 37, 38-47,

doi:10.1038/nbt.4314.

14. Van der Maaten,L., and Hinton,G. (2008) Visualizing

data using t-SNE, J.Mach. Learn. Res., 9, 2579-2605.

15. Mcinnes, L., Healy, J., and Melville, J. (2021) UMAP:

Uniform Manifold Approximation and Projec-

tion for dimension reduction, arXiv, doi: 10.48550/

arXiv.1802.03426.

16. Cai, Z., Zhang, T., and Wan, X. (2012) Antigenic dis-

tance measurements for seasonal influenza vac-

cine selection, Vaccine, 30, 448-453, doi: 10.1016/

j.vaccine.2011.10.051.

17. Amanat,F., Strohmeier,S., Meade,P., Dambrauskas,N.,

Mühlemann, B., Smith, D. J., Vigdorovich, V., Noah

Sather, D., Coughlan,L., and Krammer, F. (2021) Vac-

cination with SARS-CoV-2 variants of concern protects

mice from challenge with wild-type virus, PLoS Biol.,

19, e3001384, doi:10.1371/journal.pbio.3001384.

18. Mykytyn, A. Z., Rissmann, M., Kok, A., Rosu, M. E.,

Schipper,D., Breugem, T.I., van den Doel, P.B., Chan-

dler, F., Bestebroer, T., de Wit, M., van Royen, M. E.,

Molenkamp,R., Oude Munnink, B. B., de Vries, R.D.,

GeurtsvanKessel,C., Smith, D.J., Koopmans, M.P. G.,

Rockx,B., Lamers, M.M., Fouchier, R.A.M., and Haag-

mans, B. L. (2022) Antigenic cartography of SARS-

CoV-2 reveals that Omicron BA.1 and BA.2 are antigen-

ically distinct, Sci. Immunol., 7, eabq4450, doi:10.1126/

sciimmunol.abq4450.

19. Wilks, S. H., Mühlemann, B., Shen, X., Türeli, S.,

LeGresley, E. B., Netzl, A., Caniza, M. A., Chacaltana-

Huarcaya, J. N., Corman, V. M., Daniell, X., Datto,

M.B., Dawood, F.S., Denny, T.N., Drosten, C., Fouch-

ier, R. A. M., Garcia, P. J., Halfmann, P. J., Jassem, A.,

Jeworowski, L. M., Jones, T. C., Kawaoka, Y., Kram-

mer, F., McDanal, C., Pajon, R., Simon, V., Stockwell,

M.S., Tang, H., van Bakel,H., Veguilla,V., Webby, R.,

Montefiori, D. C., and Smith, D. J. (2023) Mapping

SARS-CoV-2 antigenic relationships and serological re-

sponses, Science, 382, eadj0070, doi: 10.1126/science.

adj0070.

20. Wang, W., Lusvarghi, S., Subramanian,R., Epsi, N. J.,

Wang, R., Goguet, E., Fries, A. C., Echegaray, F., Vas-

sell,R., Coggins, S.A., Richard, S.A., Lindholm, D.A.,

Mende,K., Ewers, E.C., Larson, D.T., Colombo, R.E.,

Colombo, C. J., Joseph, J. O., Rozman, J. S., Smith, A.,

Lalani, T., Berjohn, C. M., Maves, R. C., Jones, M. U.,

Mody, R., Huprikar, N., Livezey, J., Saunders, D., Hol-

lis-Perry, M., Wang, G., Ganesan, A., Simons, M. P.,

Broder, C. C., Tribble, D. R., Laing, E. D., Agan, B. K.,

Burgess, T. H., Mitre, E., Pollett, S. D., Katzelnick,

L. C., and Weiss, C. D. (2022) Antigenic cartography

of well-characterized human sera shows SARS-CoV-2

neutralization differences based on infection and vac-

cination history, Cell Host Microbe, 30, 1745-1758.e7,

doi:10.1016/j.chom.2022.10.012.

21. Van der Straten,K., Guerra,D., van Gils, M.J., Bont-

jer,I., Caniels, T.G., van Willigen, H.D.G., Wynberg,E.,

Poniman,M., Burger, J.A., Bouhuijs, J.H., van Rijswi-

jk,J., Olijhoek,W., Liesdek, M.H., Lavell, A.H.A., Ap-

pelman, B., Sikkens, J. J., Bomers, M. K., Han, A. X.,

Nichols, B.E., Prins,M., Vennema,H., Reusken,C., de

Jong, M.D., de Bree, G.J., Russell, C.A., Eggink,D., and

Sanders, R. W. (2022) Antigenic cartography using

sera from sequence-confirmed SARS-CoV-2 variants

of concern infections reveals antigenic divergence of

Omicron, Immunity, 55, 1725-1731.e4, doi: 10.1016/

j.immuni.2022.07.018.

22. Mühlemann,B., Trimpert,J., Walper,F., and Schmidt,

M. L. (2023) Antigenic cartography using vari-

ant-specific hamster sera reveals substantial an-

tigenicv among Omicron subvariants, bioRxiv,

doi:10.1101/2023.07.02.547076.

23. Bekliz, M., Adea, K., Vetter, P., Eberhardt, C. S.,

Hosszu-Fellous, K., Vu, D. L., Puhach, O., Essaidi-

Laziosi, M., Waldvogel-Abramowski, S., Stephan, C.,

L’Huillier, A.G., Siegrist, C.A., Didierlaurent, A.M., Kai-

ser,L., Meyer,B., and Eckerle,I. (2022) Neutralization

capacity of antibodies elicited through homologous or

heterologous infection or vaccination against SARS-

CoV-2 VOCs, Nat. Commun., 13, 3840, doi: 10.1038/

s41467-022-31556-1.

24. Lusvarghi, S., Pollett, S. D., Neerukonda, S. N.,

Wang,W., Wang,R., Vassell,R., Epsi, N.J., Fries, A.C.,

Agan, B.K., Lindholm, D.A., Colombo, C.J., Mody,R.,

Ewers, E.C., Lalani,T., Ganesan,A., Goguet,E., Hollis-

Perry,M., Coggins, S.A., Simons, M.P., Katzelnick, L.C.,

Wang,G., Tribble, D.R., Bentley,L., Eakin, A.E., Brod-

er, C. C., Erlandson, K. J., Laing, E. D., Burgess, T. H.,

Mitre,E., and Weiss, C.D. (2022) SARS-CoV-2 BA.1 vari-

ant is neutralized by vaccine booster-elicited serum

but evades most convalescent serum and therapeutic

antibodies, Sci. Transl. Med., 14, 8543, doi: 10.1126/

scitranslmed.abn8543.

25. Mykytyn, A.Z., Rosu, M.E., Kok,A., Rissmann,M., van

Amerongen, G., Geurtsvankessel, C., de Vries, R. D.,

Munnink, B. B. O., Smith, D. J., Koopmans, M. P. G.,

ASTAKHOVA et al.870

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

Lamers, M. M., Fouchier, R. A. M., and Haagmans,

B. L. (2023) Antigenic mapping of emerging SARS-

CoV-2 omicron variants BM.1.1.1, BQ.1.1, and XBB.1,

Lancet Microbe, 4, e294-e295, doi:10.1016/S2666-5247

(22)00384-6.

26. Wang,X., Jiang,S., Jiang,S., Li,X., Ai,J., Lin,K., Lv,S.,

Zhang, S., Li, M., Li, J., Dai, L., Hu, Z., Zhang, W.,

Zhang,Y., and Wang,P. (2023) Neutralization of SARS-

CoV-2 BQ.1.1, CH.1.1, and XBB.1.5 by breakthrough in-

fection sera from previous and recent waves in China,

Cell Discov., 9, 5-8, doi:10.1038/s41421-023-00569-5.

27. Rössler,A., Netzl,A., Knabl,L., Schäfer,H., Wilks, S.H.,

Bante,D., Falkensammer,B., Borena,W., von Laer,D.,

Smith, D.J., and Kimpel,J. (2022) BA.2 and BA.5 omi-

cron differ immunologically from both BA.1 omicron

and pre-omicron variants, Nat. Commun., 13, 7701,

doi:10.1038/s41467-022-35312-3.

28. Fouchier, R.A.M., and Smith, D.J. (2010) Use of anti-

genic cartography in vaccine seed strain selection,

Avian Dis., 54, 220-223, doi: 10.1637/8740-032509-

ResNote.1.

29. Xia,H., Yeung,J., Kalveram,B., Bills, C.J., Chen, J.Y.C.,

Kurhade, C., Zou, J., Widen, S. G., Mann, B. R., and

Kondor, R. (2023) Cross-neutralization and viral fit-

ness of SARS-CoV-2 Omicron sublineages, Emerg. Mi-

crobes Infect., 12, e2161422, doi: 10.1080/22221751.

2022.2161422.

30. Kurhade, C., Zou, J., Xia, H., Liu, M., Chang, H. C.,

Ren, P., Xie, X., and Shi, P. Y. (2023) Low neutraliza-

tion of SARS-CoV-2 Omicron BA.2.75.2, BQ.1.1 and

XBB.1 by parental mRNA vaccine or a BA.5 bivalent

booster, Nat. Med., 29, 344-347, doi: 10.1038/s41591-

022-02162-x.

31. Hoffmann,M., Behrens, G.M.N., Arora,P., Kempf,A.,

Nehlmeier, I., Cossmann, A., Manthey, L., Dopfer-

Jablonka, A., and Pöhlmann, S. (2022) Effect of hy-

brid immunity and bivalent booster vaccination on

Omicron sublineage neutralization, Lancet, 23, 25-28,

doi:10.1016/S1473-3099(22)00792-7.

32. Roessler,A., Netzl,A., Knabl,L., Bante,D., Wilks, S.H.,

Borena, W., Laer, D. vo., Smith, D. J., and Kimpel, J.

(2023) Characterizing SARS-CoV-2 neutralization pro-

files after bivalent boosting using antigenic cartog-

raphy, Nat. Commun., 14, 5224, doi: 10.1038/s41467-

023-41049-4.

33. Coombes, N., Bewley, K. R., Duff, Y. L., Alami-

Rahmouni,N., Ryan, K.A., Kempster,S., Ferguson,D.,

Davies, E. R., Weldon, T. M., and Cross Eleanor, S.

(2023) Evaluation of the neutralising antibody re-

sponse in human and hamster sera against SARS-

CoV-2 variants up to and including BA.2.86 using

an authentic virus neutralisation assay, bioRxiv,

doi:10.1101/2023.10.21.563398.

34. Qu,P., Xu,K., Faraone, J.N., Goodarzi,N., Zheng, Y.-M.,

Carlin, C., Bednash, J. S., Horowitz, J. C., Mallampall,

R.K., and Saif, L.J. (2023) Immune evasion, infectivity,

and fusogenicity of SARS-CoV-2 Omicron BA.2.86 and

flip variants, bioRxiv, doi:10.1101/2023.09.11.557206.

35. Ba,S., Ho,J., Zhang, R.M., Iketani,S., Yu,J., Huang,Y.,

and Qu, Y. (2023) Antigenicity and receptor affini-

ty of SARS-CoV-2 BA.2.86 spike, Nature, 624, 639-644,

doi:10.1038/s41586-023-06750-w.

36. Neerukonda, S.N., Vassell,R., Lusvarghi,S., Wang,R.,

Echegaray,F., Bentley,L., Eakin, A.E., Erlandson, K.J.,

Katzelnick, L. C., Weiss, C. D., and Wang, W. (2021)

SARS-CoV-2 Delta variant displays moderate resis-

tance to neutralizing antibodies and Spike protein

properties of higher soluble ACE2 sensitivity, en-

hanced cleavage and fusogenic activity, Viruses, 13,

2485, doi:10.3390/v13122485.

37. Astakhova, E. A., Morozov, A. A., Byazrova, M. G.,

Sukhova, M.M., Mikhailov, A.A., Minnegalieva, A.R.,

Gorchakov, A. A., and Filatov, A. V. (2023) Antigenic

cartography indicates that the Omicron BA.1 and BA.4/

BA.5 variants remain antigenically distant to ancestral

SARS-CoV-2 after Sputnik V vaccination followed by

homologous (SputnikV) or heterologous (Comirnaty)

revaccination, Int.J. Mol. Sci., 24, 10493, doi:10.3390/

ijms241310493.

38. Anderson, C.S., Sangster, M.Y., Yang,H., Mariani, T.J.,

Chaudhury,S., and Topham, D.J. (2020) Implementing

sequence-based antigenic distance calculation into

immunological shape space model, BMC Bioinformat-

ics, 21, 256, doi:10.1186/s12859-020-03594-3.

39. Anderson, E.J., Rouphael, N.G., Widge, A.T., Jackson,

L.A., Roberts, P.C., Makhene,M., Chappell, J.D., Den-

ison, M.R., Stevens, L.J., Pruijssers, A.J., McDermott,

A.B., Flach,B., Lin, B.C., Doria-Rose, N.A., O’Dell,S.,

Schmidt, S.D., Corbett, K.S., Swanson, P.A., Padilla,M.,

Neuzil, K.M., Bennett,H., Leav,B., Makowski,M., Al-

bert,J., Cross,K., Edara, V.V., Floyd,K., Suthar, M.S.,

Martinez, D. R., Baric, R., Buchanan, W., Luke, C. J.,

Phadke, V. K., Rostad, C. A., Ledgerwood, J. E., Gra-

ham, B.S., and Beigel, J.H. (2020) Safety and immu-

nogenicity of SARS-CoV-2 mRNA-1273 vaccine in older

adults, N. Engl. J. Med., 383, 2427-2438, doi: 10.1056/

NEJMoa2028436.

40. Walsh, E. E., Frenck, R. W., Falsey, A. R., Kitchin, N.,

Absalon,J., Gurtman,A., Lockhart,S., Neuzil,K., Mul-

ligan, M.J., Bailey,R., Swanson, K.A., Li,P., Koury,K.,

Kalina, W., Cooper, D., Fontes-Garfias, C., Shi, P.-Y.,

Türeci,Ö., Tompkins, K.R., Lyke, K.E., Raabe,V., Dor-

mitzer, P.R., Jansen, K.U., Şahin,U., and Gruber, W.C.

(2020) Safety and immunogenicity of two RNA-based

COVID-19 vaccine candidates, N. Engl. J. Med., 383,

2439-2450, doi:10.1056/NEJMoa2027906.

41. Logunov, D.Y., Dolzhikova, I.V., Zubkova, O.V., Tukh-

vatullin, A.I., Shcheblyakov, D.V., Dzharullaeva, A.S.,

Grousova, D.M., Erokhova, A.S., Kovyrshina, A.V., Bo-

tikov, A.G., Izhaeva, F.M., Popova,O., Ozharovskaya,

T. A., Esmagambetov, I. B., Favorskaya, I. A., Zrelkin,

D.I., Voronina, D.V., Shcherbinin, D.N., Semikhin,A.S.,

ANTIGENIC CARTOGRAPHY OF SARS-CoV-2 871

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

Simakova, Y. V., Tokarskaya, E. A., Lubenets, N. L.,

Egorova, D. A., Shmarov, M. M., Nikitenko, N. A.,

Morozova, L. F., Smolyarchuk, E. A., Kryukov, E. V.,

Babira, V. F., Borisevich, S. V., Naroditsky, B. S., and

Gintsburg, A. L. (2020) Safety and immunogenici-

ty of an rAd26 and rAd5 vector-based heterologous

prime-boost COVID-19 vaccine in two formulations:

two open, non-randomised phase 1/2 studies from

Russia, Lancet, 396, 887-897, doi: 10.1016/S0140-6736

(20)31866-3.

42. Witt, J. K. (2019) Introducing hat graphs, Cogn. Res.

Princ. Implic., 4, 31, doi:10.1186/s41235-019-0182-3.

43. Fonville, J.M., Wilks, S.H., James, S.L., Fox,A., Ventres-

ca,M., Aban,M., Xue,L., Jones, T.C., Le, N.M.H., Pham,

Q. T., Tran, N. D., Wong, Y., Mosterin, A., Katzelnick,

L.C., Labonte,D., Le, T.T., Van Der Net,G., Skepner,E.,

Russell, C.A., Kaplan, T. D., Rimmelzwaan, G.F., Ma-

surel,N., De Jong, J.C., Palache,A., Beyer, W.E.P., Le,

Q.M., Nguyen, T.H., Wertheim, H.F.L., Hurt, A.C., Os-

terhaus, A.D.M.E., Barr, I.G., Fouchier, R.A.M., Hor-

by, P.W., and Smith, D.J. (2014) Antibody landscapes

after influenza virus infection or vaccination, Science,

346, 996-1000, doi:10.1126/science.1256427.

44. Schaefer-Babajew,D., Wang,Z., Muecksch,F., Cho,A.,

Loewe, M., Cipolla, M., Raspe, R., Johnson, B., Ca-

nis, M., DaSilva, J., Ramos, V., Turroja, M., Millard,

K.G., Schmidt,F., Witte,L., Dizon,J., Shimeliovich,I.,

Yao, K. H., Oliveira, T. Y., Gazumyan, A., Gaebler, C.,

Bieniasz, P.D., Hatziioannou,T., Caskey,M., and Nus-

senzweig, M. C. (2023) Antibody feedback regulates

immune memory after SARS-CoV-2 mRNA vaccination,

Nature, 613, 735-742, doi:10.1038/s41586-022-05609-w.

45. Schiepers, A., van ’t Wout, M. F. L., Greaney, A. J.,

Zang,T., Muramatsu,H., Lin, P.J.C., Tam, Y.K., Mes-

in,L., Starr, T.N., Bieniasz, P.D., Pardi,N., Bloom, J.D.,

and Victora, G.D. (2023) Molecular fate-mapping of se-

rum antibody responses to repeat immunization, Na-

ture, 615, 482-489, doi:10.1038/s41586-023-05715-3.

Publisher’s Note. Pleiades Publishing remains

neutral with regard to jurisdictional claims in pub-

lished maps and institutional affiliations.