Nature Nanotechnology
nature nanotechnology
https://doi.org/10.1038/s41565-024-01655-9
Article
Proactive vaccination using multiviral
Quartet Nanocages to elicit broad
anti-coronavirus responses
Rory A. Hills
1,2
, Tiong Kit Tan
3
, Alexander A. Cohen
4
, Jennifer R. Keeffe
4
,
Anthony H. Keeble
1,2
, Priyanthi N. P. Gnanapragasam
4
, Kaya N. Storm
4
,
Annie V. Rorick
4
, Anthony P. West Jr.
4
, Michelle L. Hill
5
, Sai Liu
5
,
Javier Gilbert-Jaramillo
5
, Madeeha Afzal
5
, Amy Napier
5
,
Gabrielle Admans
2
, William S. James
5
, Pamela J. Bjorkman
4
,
Alain R. Townsend
3,6
& Mark R. Howarth
1,2
Defending against future pandemics requires vaccine platforms that
protect across a range of related pathogens. Nanoscale patterning
can be used to address this issue. Here, we produce quartets of linked
receptor-binding domains (RBDs) from a panel of S
AR
S-
like b
et-
acoronaviruses, coupled to a computationally designed nanocage
through SpyTag/SpyCatcher links. These Quartet Nanocages, possessing
a branched morphology, induce a high level of neutralizing antibodies
against several different coronaviruses, including against viruses not
represented in the vaccine. Equivalent antibody responses are raised to
RBDs close to the nanocage or at the tips of the nanoparticle’s branches.
In animals primed with SARS-CoV-2 Spike, boost immunizations with
Quartet Nanocages increase the strength and breadth of an otherwise
narrow immune response. A Quartet Nanocage including the Omicron
XBB.1.5 ‘Kraken’ RBD induced antibodies with binding to a broad range
of sarbecoviruses, as well as neutralizing activity against this variant of
concern. Quartet nanocages are a nanomedicine approach with potential
to confer heterotypic protection against emergent zoonotic pathogens and
facilitate proactive pandemic protection.
Nanoscale organization is a key signal for the programming of immune
responses
1
–
3
. Highly multivalent display of antigens on virus-like par-
ticles (VLPs) or other nanoparticles enhances the strength and per
-
sistence of immune responses, facilitating lymph node uptake and
increasing B cell receptor (BCR) clustering
1
,
2
. VLP manufacturing uses
existing facilities for microbial fermentation to facilitate large-scale
production
4
and can avoid the need for a cold-chain
5
, and VLPs have
shown a good balance of safety and efficacy
6
.
Existing vaccination strategies have shown success in reducing
death and serious illness from SARS-CoV-2 (SARS2)
7
. Nevertheless,
Received: 20 March 2023
Accepted: 15 March 2024
Published online: xx xx xxxx
Check for updates
1
Department of Biochemistry, University of Oxford, Oxford, UK.
2
Department of Pharmacology, University of Cambridge, Cambridge, UK.
3
MRC Human
Immunology Unit, MRC Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford, Oxford, UK.
4
Division of Biology and
Biological Engineering, California Institute of Technology, Pasadena, CA, USA.
5
James & Lillian Martin Centre, Sir William Dunn School of Pathology,
University of Oxford, Oxford, UK.
6
Centre for Translational Immunology, Chinese Academy of Medical Sciences Oxford Institute, University of Oxford,
Oxford, UK.
e-mail:
bjorkman@caltech.edu
;
alain.townsend@imm.ox.ac.uk
;
mh2186@cam.ac.uk
Nature Nanotechnology
Article
https://doi.org/10.1038/s41565-024-01655-9
Quartet band was broad on SDS–PAGE because of natural variation
in glycosylation (Fig.
1d
). Removing N-linked glycans with peptide
N-Glycosidase F (PNGase F) induced a downward shift on the gel
(Fig.
1d
). Quartet-SpyTag gave a uniform peak by size-exclusion chroma
-
tography (Supplementary Fig. 4a). We demonstrated that the Quartet
coupled efficiently to SpyCatcher003-mi3 (Fig.
1e
).
Quartet Nanocages raise antibodies to diverse
sarbecoviruses
We then explored the Quartet’s immunogenicity as a soluble protein
or multimerized on nanocages (Fig.
2a
). Doses for all immunizations
were normalized by the number of SpyTags, allowing comparison of a
molar equivalent of SpyCatcher003-mi3 nanocages with similar levels
of occupancy. Two doses were administered to mice 14 d apart using
alum-based adjuvant (Fig.
2b
), before quantifying IgG titre against RBD
antigens by enzyme-linked immunosorbent assay (ELISA). Post-prime,
the Quartet Nanocage elicited the highest antibody titre against SARS2
waning vaccine protection, continuing emergence of new variants and
uncertain efficacy of therapeutics mean that new vaccine strategies are
still urgently needed
8
,
9
. It is also important to protect against new pan
-
demic threats from coronaviruses, which previously led to SARS-CoV
(SARS1) and MERS-CoV outbreaks
10
. Other zoonotic coronaviruses
such as WIV1 and SHC014 have been identified as having pandemic
potential
11
. Immunizing with a single antigen typically induces a nar
-
row strain-specific immune response, which may not protect against
diverse pre-existing strains or newly arising variants of that pathogen
12
.
In a recently introduced approach, VLPs display a panel of protein
variants arranged stochastically on their surface, to drive expansion of
B cells recognizing common features of the different antigens. A mosaic
of different hemagglutinin heads on ferritin nanoparticles elicited
cross-reactive antibodies against diverse influenza strains within the
H1 subtype
13
. This approach has been applied to SARS2, with mosaic
nanoparticles displaying multiple RBDs from the Spike of different
sarbecoviruses
12
,
14
,
15
. Sarbecoviruses are the subgenus of betacorona-
viruses that includes SARS1 and SARS2. RBDs can be multimerized on
VLPs through genetic fusion
15
or isopeptide coupling
12
. Fusion of a set
of sarbecovirus RBDs with SpyTag003 facilitates simple nanoassembly
onto the SpyCatcher003-mi3 VLP
12
(Fig.
1a
). SpyCatcher003 is a protein
that we engineered to rapidly form an isopeptide bond with SpyTag
peptide
16
. mi3 is a 60-mer hollow protein nanocage, computationally
designed to self-assemble into a stable dodecahedron
17
,
18
.
In our previous study, the broadest immune response came from
mosaic particles displaying eight different RBDs
12
,
14
. These Mosaic-8
nanoparticles elicited neutralizing antibodies against a variety of sarbe
-
coviruses in mouse and rhesus macaque models. Critically, responses
were not limited to viruses whose RBDs were represented on Mosaic-8
nanoparticles and included mismatched responses against heterolo
-
gous sarbecoviruses
12
,
14
. Mosaic-8 nanoparticles have gained support
from the Coalition for Epidemic Preparedness Innovations to enter
clinical trials. However, the need to produce nine different components
(eight RBDs and SpyCatcher003-mi3) at Good Manufacturing Practice
level creates a challenge for broad scaling.
Here, we establish the production of multiviral Quartet Nanocages
(Fig.
1a
). Initially we express a multiviral Quartet from RBDs of four
different viruses, concatenated as a single polypeptide chain. These
antigenic Quartets are assembled via a terminal SpyTag to extend out
from SpyCatcher003-mi3 nanocages, creating a protein nanoparticle
with a branched morphology. This nanoassembly route reduces the
number of vaccine components, as well as creating an architecture that
allows a greater number of RBDs to be displayed on each nanocage. We
measure antibody responses to the range of sarbecoviruses displayed
on the Quartet Nanocage, to sarbecoviruses not present within the
chain, as well as to SARS2 variants of concern (VOCs). Comparing dif-
ferent nanoassemblies, we dissect the breadth of antibody binding
to different sarbecoviruses, neutralization potency and the ability to
boost a broad response following focused priming. The magnitude and
breadth of antibody induction show that Quartet Nanocages may pro
-
vide a scalable route to induce neutralizing antibodies across a range
of related viruses, to prepare for emerging outbreak disease threats.
Design of multiviral Quartet Nanocages
The SARS2 RBD is directly involved in binding to the cell receptor
angiotensin-converting enzyme 2 (ACE2) and is the target of most neu-
tralizing antibodies
15
. We genetically fused RBDs from the evolutionarily
related sarbecoviruses SHC014, Rs4081, RaTG13 and SARS2 Wuhan
(Fig.
1b
and Supplementary Figs. 1 and 2) to produce a multiviral Quartet
(Fig.
1c
). These RBDs allow comparison with the previously described
Mosaic-4 vaccine
12
. The multiviral Quartet was engineered with a signal
sequence for secretion from mammalian cells and a terminal SpyTag,
to enable multivalent display on SpyCatcher003-mi3 (Fig.
1a
). The
Quartet was secreted efficiently by Expi293F cells and affinity-purified
via SpyTag using the SpySwitch system
19
(Supplementary Fig. 3). The
250
130
100
70
55
35
25
20
SARS1
WIV1
SHC014
BM-4831
BtKY72
pang17
SARS2
RaTG13
Rs4081
RmYN02
Rf1
b
a
–
+
PNGase F
d
e
Quartet
PNGase F
Deglycosylated
Quartet
Monomeric
RBDs
RBD
quartet
Sarbecovirus phylogeny
Assembly of multiviral Quartet Nanocage
c
Quartet coupling to nanocage
Nanocage
Nanocage:Quartet
Quartet
250
130
100
70
55
35
25
Nanocage (M)
2
RBD quartet (M)
2
0
2
4
6
Quartet glycosylation
500
1,000
1,500
2,000
2,500
Glycosylation
Signal peptide
1
Nucleotide
2,895
SHC014
RBD
Rs4081
RBD
RaTG13
RBD
SARS2
Wuhan RBD
SpyTag
Clade
1a
3
1b
2
SpyCatcher003-
mi3
Nanocage for
Plug-and-Display
Fig. 1 | Preparation of multiviral Quartet Nanocages.
a
, Plug-and-Display
vaccine assembly of mosaic and Quartet Nanocages. Genetic fusion of
SpyCatcher003 (dark blue) with mi3 (purple) allows efficient multimerization of
single or Quartet RBDs linked to SpyTag (cyan) through spontaneous isopeptide
bond formation (marked in red). Only some antigens are shown in the schematic
for clarity.
b
, Phylogenetic tree of sarbecoviruses used in this study, based
on RBD sequence.
c
, Genetic organization of the multiviral Quartet-SpyTag,
indicating the viral origin of RBDs,
N
-linked glycosylation sites and tag location.
d
, Analysis of Quartet-SpyTag with SDS–PAGE/Coomassie staining, with or
without PNGase F deglycosylation. A representative gel from two independent
experiments. Molecular weight markers are in kDa.
e
, Coupling of RBD Quartet
to SpyCatcher003-mi3 Nanocage at different molar Nanocage:antigen ratios,
analysed by SDS–PAGE/Coomassie. A representative gel from two independent
experiments. Molecular weight markers are in kDa.
Nature Nanotechnology
Article
https://doi.org/10.1038/s41565-024-01655-9
Wuhan RBD, surpassing the Homotypic Nanocage and Uncoupled Quar
-
tet (Supplementary Fig. 5a). The Quartet Nanocage also elicited a strong
post-prime response to SARS1 RBD, not represented on the immunogen
(Supplementary Fig. 5a), with a titre greater than the response against
SARS2 Wuhan by Homotypic Nanocage (Supplementary Fig. 5).
After boosting, we similarly found the strongest response against
SARS2 from Quartet Nanocage, followed by Uncoupled Quartet, Homo
-
typic Nanocage and finally Uncoupled RBD (Fig.
2c
and Supplementary
Figs. 6 and 7). This pattern is retained for SARS2 Wuhan, Beta and
Delta Spikes (Supplementary Fig. 5b). Uncoupled RBD and Homotypic
Nanocage immunization raised low responses against the panel of sar
-
becovirus RBDs, with the greatest Homotypic Nanocage cross-reactive
response against the closely related RaTG13 RBD (Fig.
2c
). In contrast,
we saw substantial responses against all tested RBDs with Uncou
-
pled Quartet and most substantially Quartet Nanocage immunization
(Fig.
2c
). This included a strong heterotypic response against BM-4831
and SARS1 RBDs, which were absent from the chain and elicited titres
only slightly lower than Homotypic Nanocage against SARS2 Wuhan
(Fig.
2c
). These results suggest the potential of this Quartet Nanocage
approach to induce antibodies against a broad range of sarbecoviruses.
We had hypothesized that RBDs at the tip of the Quartet would give
stronger responses than RBDs nearer the nanocage surface. In fact, we
saw no obvious relationship between RBD chain location and antibody
titre (Fig.
2c
).
Comparison of Quartet Nanocages and Mosaic
nanoparticles
We next compared the multiviral Quartet with leading mosaic nanopar
-
ticle vaccines, which have stochastic arrangements of RBDs. Mosaic-4,
containing the same four RBDs as our Quartet, previously induced
broad antibodies, but the best breadth was obtained with a Mosaic-8
immunogen
12
,
14
. Therefore, we also produced the Alternate Multiviral
Quartet, containing SpyTag followed by RBDs from other sarbecovi
-
ruses: pang17, RmYN02, Rf1 and WIV1 (Supplementary Fig. 8b). Cou
-
pling both the Quartet and Alternate Quartet to SpyCatcher003-mi3
generated the Dual Quartet Nanocage, presenting the same eight
RBDs as Mosaic-8 (Fig.
3a
). We characterized the Alternate Quartet by
SDS–PAGE/Coomassie with and without deglycosylation using PNGase
F (Fig.
3b
) and by size-exclusion chromatography (Supplementary
Fig. 4c). To interrogate further the relationship between chain posi
-
tion and immunogenicity, we produced a Quartet with SpyTag moved
from the C terminus to the N terminus (Supplementary Fig. 8a). This
SpyTag-Quartet was used for all subsequent immunizations.
Dynamic light scattering (DLS) validated that each immuno
-
gen homogeneously assembled with SpyCatcher003-mi3 (Fig.
3c
).
Negative-stain transmission electron microscopy (TEM) confirmed
the integrity of the Quartet Nanocages. The visible particle diameter
was equivalent between uncoupled, Mosaic-8 and Quartet Nanocages,
consistent with dynamic arrangement of the RBD quartet on the nanoc
-
age surface (Supplementary Fig. 9).
Immunizations compared these Mosaic and Quartet Nanocage
antigens (Fig.
3a
). For all RBDs, the two highest post-boost antibody
titres were raised by Quartet Nanocage and Dual Quartet Nanocage
(Fig.
3d
and Supplementary Figs. 10–13). Surprisingly, Quartet Nanoc
-
age and Dual Quartet Nanocage induced a similar response to each
other against WIV1 and pang17 (Fig.
3d
and Supplementary Figs. 10–13),
even though these antigens were present in Dual Quartet Nanocage but
not Quartet Nanocage. In agreement with previous results
12
, Mosaic-4
and Mosaic-8 produced higher titres than SARS2 Homotypic Nanocage
against the panel of sarbecovirus RBDs. Uncoupled Quartet produced
similar titres as both Mosaics against the RBD set (Fig.
3d
and Supple-
mentary Figs. 10–13). These trends were also apparent in post-prime
samples, except Mosaic-8 and Quartet Nanocage raised a similar
anti-SARS1 response (Supplementary Fig. 10b). As previously, there was
no clear relationship between chain position and antibody response
against that RBD (Fig.
3d
). All conditions except Uncoupled Quartet
induced a comparable antibody response against SpyCatcher003-mi3
itself (Supplementary Fig. 10c). SpyTag-Maltose Binding Protein (MBP)
was a negative control, revealing minimal antibodies against SpyTag
itself (Supplementary Fig. 10c).
Induction of neutralizing antibodies by Quartet
Nanocages
To relate antibody level to antibody efficacy, we tested neutralization
of SARS2 Wuhan or Delta virus. We saw that the strongest neutraliza
-
tion was induced by Quartet Nanocage in each case, while Homotypic
Nanocage gave higher responses than Uncoupled Quartet (Fig.
4a,b
).
We compared SARS1 pseudovirus neutralization induced by Quartet
and Mosaic antigens, giving insight into neutralization breadth, as
SARS1 was a mismatch for all tested immunogens. Pseudotyped virus
neutralization correlates well with neutralization of authentic virus
20
.
Quartet
Nanocage
Homotypic
Nanocage
Uncoupled
Quartet
Uncoupled
RBD
a
c
b
Prime
Day 0
Post-prime
bleed
Day 13
Boost
Day 14
Post-boost
bleed
Day 41
Post-boost SARS2 Wuhan
Post-boost SHC014
Post-boost Rs4081
Post-boost RaTG13
Post-boost SARS1
Post-boost BM-4831
0
0.5
1.0
1.5
2.0
2.5
IgG titre AUC
0
0.5
1.0
1.5
2.0
2.5
IgG titre AUC
0
0.5
1.0
1.5
2.0
2.5
IgG titre AUC
0
0.5
1.0
1.5
2.0
2.5
IgG titre AUC
0
0.5
1.0
1.5
2.0
IgG titre AUC
0
0.5
1.0
1.5
2.0
IgG titre AUC
Mismatched
Matched
SpyTag
SpyTag
SARS2 RBD
SpyTag
SpyTag
SARS2 RBD
SARS2 RBD
SARS2 RBD
RaTG13 RBD
Rs4081 RBD
SHC014 RBD
RaTG13 RBD
Rs4081 RBD
SHC014 RBD
mi3
mi3
SpyCatcher003
SpyCatcher003
***
*
*
***
***
***
***
***
***
***
***
***
***
***
***
***
***
**
***
*
***
***
**
Fig. 2 | Broad immune response from immunization with Quartet Nanocages.
a
, Schematic of antigens for this set of immunizations, comparing uncoupled
proteins or proteins coupled to the SpyCatcher003-mi3 nanocage.
b
, Procedure
for immunization and sampling.
c
, ELISA for post-boost serum IgG binding to
different sarbecovirus RBDs is shown as the area under the curve (AUC) of a serial
dilution. Sera are from mice immunized with uncoupled SARS2 Wuhan RBD
(orange), uncoupled Quartet-SpyTag (yellow), SARS2 Wuhan RBD coupled to
SpyCatcher003-mi3 (green) or Quartet-SpyTag coupled to SpyCatcher003-mi3
(blue). Solid rectangles under samples indicate ELISA against a component of
that vaccine (matched). Striped rectangles indicate ELISA against an antigen
absent in that vaccine (mismatched). Each dot represents one animal. The mean
is denoted by a bar, shown ±1 s.d.;
n
= 6. Significance was calculated with an
ANOVA test using Tukey’s post hoc test. *
P
< 0.05, **
P
< 0.01, ***
P
< 0.001; other
comparisons were non-significant.
Nature Nanotechnology
Article
https://doi.org/10.1038/s41565-024-01655-9
Dual Quartet Nanocage gave the strongest neutralizing response to
SARS1. This was followed Quartet Nanocage and Mosaic-8 which
induced similar, relatively strong anti-SARS1 responses (Fig.
4c
).
We performed immunizations with tenfold higher antigen dose
and the squalene-based adjuvant AddaVax to further enhance neutrali
-
zation (Supplementary Figs. 14–17). These results are outlined further
in the Supplementary Discussion.
While the motivation of this approach is protection against future
zoonotic pathogens, the ideal vaccine candidate would also protect
against circulating SARS2 variants. We produced an updated Kraken
Quartet containing SARS2 XBB.1.5 in place of SARS2 Wuhan (Supple-
mentary Fig. 8d). Mouse immunizations were performed with Homo
-
typic Nanoparticles, Mosaic-8 nanoparticles and Quartet Nanocages
that contained either SARS2 Wuhan or XBB.1.5, in addition to a Dual
Quartet Nanocage that contained only the Wuhan RBD (Supplementary
Fig. 18). All the Quartet and Mosaic immunogens produced greater
antibody binding against zoonotic coronaviruses than their homotypic
counterparts (Supplementary Fig. 18). There was substantial antibody
binding against SARS2 VOCs with no statistically significant difference
between antibody binding raised by Quartet and Mosaic immuno
-
gens against any tested SARS2 VOC (Wuhan, Delta, XBB.1.5 and BQ.1.1)
(Supplementary Fig. 18). Immunogens containing XBB.1.5 provided
substantially improved neutralization of SARS2 XBB.1.5 pseudovirus
relative to Wuhan-containing counterparts (Supplementary Fig. 19).
This result highlights the capacity to update the Quartet Nanocage,
to protect against recently evolved antibody-escape variants. Both
Homotypic
Nanocage
Mosaic-4
Mosaic-8
Quartet
Nanocage
Dual Quartet
Nanocage
Uncoupled
Quartet
a
Immunogens
–
+
PNGase F
250
130
70
55
35
25
20
b
Alternate Quartet
PNGase F
Deglycosylated
Alternate Quartet
c
Alternate Quartet
0
0.2
0.4
0.6
0.8
1.0
1
10
100
1,000
Normalized intensity
Hydrodynamic radius (nm)
d
Immunogen
R
H
(nm)
Dual Quartet Nanocage
Quartet Nanocage
Homotypic Nanocage
Mosaic-4
Mosaic-8
Uncoupled Nanocage
25.3 ± 5.8
25.2 ± 4.9
22.0 ± 5.2
22.0 ± 4.0
19.9 ± 5.2
17.2 ± 4.6
0
IgG titre AUC
1.0
2.0
3.0
4.0
Antigen
Quartet
Nanocage
Homotypic
Nanocage
Mosaic-4
Mosaic-8
Dual Quartet
Nanocage
Uncoupled
Quartet
Matched
Mismatched
SARS2
Wuhan
SHC014
RaTG13
Rs4081
WIV1
pang17
SARS1
BtKY72
MBP
Fig. 3 | Comparison of immunization with Mosaic or Quartet Nanocages.
a
, Schematic of antigens for this set of immunizations.
b
, Validation of the
Alternate Quartet by SDS–PAGE with Coomassie staining, shown ±PNGase F
deglycosylation. A representative gel from two independent experiments.
Molecular weight markers are in kDa.
c
, DLS of SpyCatcher003-mi3 alone
(uncoupled nanocage) or each immunogen. The mean hydrodynamic radius
(
R
H
) is shown ±1 s.d., derived from 20 scans of the sample. Uncoupled Nanocage
is shown in black, with the other particles coloured as in the table inset.
d
, ELISA
for post-boost serum IgG as the AUC of serial dilution, from mice immunized
with Homotypic SARS2 Nanocages (dark blue), Mosaic-4 (green), Mosaic-8 (light
blue), SpyTag-Quartet Nanocage (pink), Dual Quartet Nanocage (purple) or
Uncoupled Quartet (grey). Squares indicate ELISA against a component of that
vaccine (matched) while crosses indicate ELISA against an antigen absent in that
vaccine (mismatched). Responses are shown to the set of sarbecovirus RBDs, with
SpyTag-MBP as a negative control. The mean is shown ±1 s.d.;
n
= 6. Individual
data points and statistics are shown in Supplementary Figs. 10 and 11.
ID
50
ID
50
Uncoupled
RBD
Uncoupled
Quartet
Homotypic
Nanocage
Quartet
Nanocage
Homotypic
Nanocage
Mosaic-4
Mosaic-8
Quartet
Nanocage
Dual Quartet
Nanocage
Uncoupled
Quartet
ID
50
10
4
10
3
10
2
10
1
a
SARS2 Wuhan virus neutralization
b
c
Mismatched
Matched
***
**
*
***
***
***
*
*
***
***
***
***
***
**
*
**
*
10
3
10
2
10
1
10
0
10
3
10
2
10
1
10
0
SARS1 pseudovirus neutralization
SARS2 Delta virus neutralization
Fig. 4 | Neutralization induced by Quartet immunogens.
a
, Neutralization of
Wuhan SARS2 virus by boosted mouse sera. Mice were primed and boosted with
Uncoupled RBD (orange), Uncoupled Quartet (yellow), Homotypic Nanocage
(green) or Quartet Nanocage (blue). Each dot represents one animal, showing
the serum dilution giving 50% inhibition of infection (ID
50
).
b
, Neutralization
of Delta SARS2 virus by boosted mouse sera, as in
a
.
c
, Neutralization of SARS1
pseudovirus (mismatched) by post-boost mouse sera, after immunization with
different Quartet and Mosaic immunogens. Dashed horizontal lines represent
the limit of detection. The mean is denoted by a bar + 1 s.d.;
n
= 6. Significance was
calculated with an ANOVA test, followed by Tukey’s multiple comparison post hoc
test of ID
50
values converted to log
10
scale. *
P
< 0.05, **
P
< 0.01, ***
P
< 0.001; other
comparisons were non-significant.
Nature Nanotechnology
Article
https://doi.org/10.1038/s41565-024-01655-9
the Kraken- and Wuhan-containing Quartet Nanocage and Mosaic-8
provided greater neutralization of the mismatched SARS1 pseudovirus
than their homotypic counterparts (Supplementary Fig. 19).
Quartet Nanocage immunization in mice with
existing immunity
Given the large fraction of the world vaccinated or previously infected
with SARS2 (more than 770 million confirmed cases and 13 billion vac
-
cine doses administered by December 2023)
21
–
24
, an outstanding ques
-
tion was whether a broad antibody response could be achieved in the
face of a prebiased immune response. It is not feasible to match the
pattern of vaccine sources and timings for different people around the
world, but we generated a pre-existing response by priming with SARS2
Wuhan Spike (HexaPro) protein. We then boosted with different immu
-
nogens designed to elicit a broad response (Fig.
5a
). One hypothesis
is that animals with a pre-existing response to SARS2, upon boosting
with Quartet Nanocage, would amplify their SARS2 antibodies from
a memory response and be less stimulated by other antigens, so the
immune response would be narrow. To test this question, we generated
Quartet [SARS1], replacing SARS2 with SARS1 RBD (Supplementary
Fig. 8c). This approach led to the ambitious aim of boosting a SARS2
response using an immunogen lacking any SARS2 sequence. We pro
-
duced Dual Quartet Nanocage [SARS1] by mixing Alternate Quartet
and Quartet [SARS1] (Supplementary Fig. 8).
Priming with SARS2 Wuhan Spike raised the expected narrow
strain-specific response against SARS2 Wuhan RBD (Fig.
5b
) and negligi
-
ble response to SARS1 or BtKY72 (Supplementary Fig. 20). Surprisingly,
the different boosts (Fig.
5b
) raised similar responses against SARS2,
despite SARS2 RBD being absent in Quartet Nanocage [SARS1] and
Dual Quartet Nanocage [SARS1] (Fig.
5b
). As expected, Quartet Nanoc
-
age [SARS1] and Dual Quartet Nanocage [SARS1] raised the strongest
response against SARS1 RBD (Fig.
5c
). Quartet Nanocage and Mosaic-8
raised greater antibody response than Homotypic Nanocage or Spike
boost against SARS1 and BtKY72 (Fig.
5c
). Mismatched responses to
SARS1 and BtKY72 raised by Mosaic-8 and Quartet Nanocage were simi
-
lar to the SARS1 response from a single dose of these candidates in naive
mice (Supplementary Fig. 10b). Overall, Quartet Nanocages achieve
broad anti-sarbecovirus response, despite animals being prebiased to
a specific viral antigen. In addition, Quartet Nanocage lacking SARS2
still induces a good level of anti-SARS2 antibodies, while stimulating
broad responses across sarbecoviruses.
Further characterization of Quartet Nanocage
immunogens
To investigate responses to RBDs at different distances from Spy
-
Catcher003-mi3, we performed ELISAs on Quartet antigens using a
panel of anti-SARS2 monoclonal antibodies
25
. We found minimal differ
-
ence between binding to SpyTag-Quartet with or without coupling to
SpyCatcher003-mi3 (Supplementary Fig. 21a,b). There was consistent
reduction in anti-SARS2 antibody binding when SARS2 was the inner-
most RBD (Quartet-SpyTag) compared with being the outermost RBD
(SpyTag-Quartet) (Supplementary Fig. 21c). Despite this difference
in antibody binding, there remains no relationship between location
on the chain and antibody response elicited by immunization in any
condition that we tested.
Our hypothesis is that the flexible linkers facilitate a dynamic
surface that displays multiple different RBDs. To this end, we produced
a Quartet without flexible Gly-Ser linkers separating the different
RBDs and performed immunizations comparing the No Linker Quar-
tet Nanocage with the conventional Quartet Nanocage (Supplemen
-
tary Fig. 22a). There was no statistically significant difference in the
immune responses raised by the No Linker Quartet Nanocage and the
conventional Quartet Nanocage to any antigen tested (Supplementary
Fig. 22b). There remained no apparent relationship between location
on the chain and antibody response for the No Linker Quartet Nanocage
(Supplementary Fig. 22b). It remains possible that the flexible region
at the N and C termini of the RBD protein maintained overall flexibility
for the polyprotein.
We employed yeast-display deep mutational scanning to map
mutations in SARS2 Wuhan RBD that escape antisera binding, giving
insight where elicited antibodies bind (Supplementary Fig. 23)
26
,
27
.
Homotypic Nanocage immunization produces a response dominated
by narrowly focussed class 1 and class 2 antibodies
25
. The Quartet
Nanocage showed variable responses: in one case class 1 dominated
and other cases were dominated by class 3 or class 4 (ref.
25
). Both
the Dual Quartet Nanocage and Mosaic-8 elicited antisera binding
Dual Quartet
Nanocage [SARS1]
Quartet Nanocage
[SARS1]
Quartet
Nanocage
Homotypic
Nanocage
IgG titre AUC
Post-boost
SARS2 Wuhan
IgG titre AUC
1.5
1.0
0.5
0
2.0
2.5
IgG titre AUC
b
c
a
Prime
Day 0
Post-prime
bleed
Day 13
Boost
Day 14
Post-boost
bleed
Day 35
Same
prime
SARS2
Spike
SARS2
Spike
Mosaic-8
1.5
1.0
0.5
0
2.0
2.5
Dierent
boost
Dual Quartet
Nanocage
1.5
1.0
0.5
0
2.0
Mismatched
Matched
Post-boost BtKY72
Post-prime
SARS2 Wuhan
Post-boost SARS1
IgG titre AUC
1.5
1.0
0.5
0
2.0
***
***
**
***
***
***
***
***
***
**
***
***
*
*
*
Fig. 5 | Quartet immunization induces broad antibodies even after a
preprimed SARS2 response.
a
, Summary of timeline and antigens for this
set of immunizations.
b
, ELISA for serum IgG to SARS2 Wuhan RBD presented
as the AUC of a serial dilution. All mice were primed with Wuhan SARS2 Spike,
before boosting with Wuhan SARS2 Spike protein (light green), Homotypic
Nanocage (pink), Mosaic-8 (dark blue), SpyTag-Quartet Nanocage (red), Dual
Quartet Nanocage (orange), Quartet Nanocage with SARS1 RBD replacing SARS2
(purple) or Dual Quartet Nanocage with SARS1 RBD replacing SARS2 (cyan). Solid
rectangles under samples indicate ELISA against a component of that vaccine
(matched). Striped rectangles indicate ELISA against an antigen absent in that
vaccine (mismatched). Each dot represents one animal. The mean is denoted by a
bar ±1 s.d.;
n
= 6.
c
, ELISA for serum IgG to other sarbecovirus RBDs, as for
b
, with
each dot representing one animal and the mean being denoted by a bar ±1 s.d.;
n
= 6. Significance was calculated with an ANOVA test using Tukey’s post hoc test.
*
P
< 0.05, **
P
< 0.01, ***
P
< 0.001; other comparisons were non-significant.