key: cord-0924174-rhctbi9a
authors: Kim, Chulwoo; Kim, Jae-Deog; Seo, Sang-Uk
title: Nanoparticle and virus-like particle vaccine approaches against SARS-CoV-2
date: 2022-01-28
journal: J Microbiol
DOI: 10.1007/s12275-022-1608-z
sha: 805eaf9f1721208c7f9bdfe753f8b56a10276aad
doc_id: 924174
cord_uid: rhctbi9a

The global spread of coronavirus disease 2019 caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection has provoked an urgent need for prophylactic measures. Several innovative vaccine platforms have been introduced and billions of vaccine doses have been administered worldwide. To enable the creation of safer and more effective vaccines, additional platforms are under development. These include the use of nanoparticle (NP) and virus-like particle (VLP) technology. NP vaccines utilize self-assembling scaffold structures designed to load the entire spike protein or receptor-binding domain of SARS-CoV-2 in a trimeric configuration. In contrast, VLP vaccines are genetically modified recombinant viruses that are considered safe, as they are generally replication-defective. Furthermore, VLPs have indigenous immunogenic potential due to their microbial origin. Importantly, NP and VLP vaccines have shown stronger immunogenicity with greater protection by mimicking the physicochemical characteristics of SARS-CoV-2. The study of NP- and VLP-based coronavirus vaccines will help ensure the development of rapid-response technology against SARS-CoV-2 variants and future coronavirus pandemics.

Coronavirus, a zoonotic RNA virus, can infect mammalian and avian hosts (Woo et al., 2012) . Only four "non-severe" strains (i.e., 229E, OC43, NL63, and HKU1) are known to circulate in humans and usually cause mild airway manifes-tations (Garbino et al., 2006) . However, the impact of coronavirus on public health increased with the emergence of severe acute respiratory syndrome (SARS) in 2002 and Middle East respiratory syndrome (MERS) in 2012, which caused 774 and 858 deaths, respectively (Paules et al., 2020) . Unlike "non-severe" human coronaviruses, SARS coronavirus (SARS-CoV) and MERS coronavirus (MERS-CoV) utilize distinct host receptors and employ evasion mechanisms to escape host immune responses, which eventually contribute to serious infections in humans (Lee et al., 2019a (Lee et al., , 2019b Molaei et al., 2021) . The unprecedented impact of coronavirus on public health has created the need for efficient prophylactic measures.

Vaccines are the most effective and economical way to control infectious diseases. However, they may not be the best option for rapid control of newly emerging pathogens because of the significant time needed to evaluate their safety and effectiveness for human application. Vaccine candidates against SARS-CoV were developed based on inactivated viruses, virus-like particles (VLPs), and DNA vaccine platforms but were not used in human subjects before the outbreak ended in 2003 (Taylor, 2006) . While the MERS outbreak also led to vaccine development when the virus spread globally with a rapidly increasing incidence, the candidate vaccine remains under clinical study (Alharbi et al., 2019; Bosaeed et al., 2022) . However, when SARS-CoV-2 was identified as a causative agent of coronavirus disease 2019 and its gene sequences were reported in January 2020, research groups immediately began vaccine development. It seems obvious that the experience gained in earlier coronavirus outbreaks accelerated vaccine development. The COVID-19 pandemic has resulted in far more infections compared to previous outbreaks, highlighting the need for safe and effective vaccines. The speed of transmission also enabled the acceptance of innovative new vaccine platforms; namely, mRNA and adenoviral DNA vector vaccines, in addition to traditional inactivated virus vaccines. The timeline for clinical studies of new COVID-19 vaccines was exceptionally short to meet the urgent demand. While the first COVID-19 vaccine was licensed in August 2021, vaccines were used in advance under the World Health Organization pre-qualification and emergency use authorization. While it is early yet to conclusively discuss adverse effects related to the new vaccine platforms, safety issues remain of interest as new SARS-CoV-2 variants sustain prolonged spread of infection (Kim et al., 2021a) . Risk-benefit assessment has also grown in importance as vaccination programs are extended to younger age groups. Therefore, safer vaccines using diverse platforms must also be investigated.

Nanoparticle (NP) and VLP vaccines are terminologically interchangeable because they express the antigen of interest on virus-sized particles. Many NPs are derived from naturally oligomer-forming molecules, whereas VLPs utilize viral capsid proteins or replication-defective virus particles. More details about both nanoscale vaccine platforms against SARS-CoV-2 are covered later in this review (Table 1) . NP and VLP platforms can be classified by type (Fig. 1A) . NPs are subdivided into two groups according to the simultaneous or separate expression of antigen and scaffold proteins. NPs can express scaffolds with antigen domains within a single peptide, enabling antigen display on the outer surface through self-assembly (Lim et al., 2021) . However, many NP approaches, especially for recent COVID-19 vaccine development, have separately expressed scaffold and antigen do-mains, which are later combined using linking systems. One advantage to the separate expression is the Escherichia colibased production of the NP scaffold, while the antigenic portions can be produced in eukaryotic cell lines. E. coli-based expression enables the cheaper and faster production of scaffold proteins compared to eukaryotic expression systems. Another advantage is the flexible use of the NP platform for different antigen combinations based on common linking systems (Brune et al., 2016; Zhang et al., 2020; Boyoglu-Barnum et al., 2021; Cohen et al., 2021a) . The use of a single antigen with linking system allows it to be easily shared by different types of NP scaffolds. Also, multiple antigens in a common linking system can be conjugated within a single particle to generate mosaic NP vaccines. VLPs are intrinsically immunogenic substances, as they usually carry genetic materials and pathogen-associated molecular patterns (Yan et al., 2005; Zepeda-Cervantes et al., 2020; Mohsen et al., 2022) . The essential gene is removed to ensure safety, rendering VLPs replication-defective. For example, non-structural gene knockout influenza particles were produced by reverse genetics to yield replication-defective VLPs that protected mice from lethal challenge (Watanabe et al., 2002) . Based on their immunogenicity and safety advantages, VLPs have been used for vaccine development longer than NPs (Shirbaghaee and Bolhassani, 2016) . VLP technology has also fused foreign antigens to the VLP coat or membrane proteins (Frietze et al., 2016) . VLPs also introduced a linking system and created efficient scaffold systems for foreign antigens. These continuous improvements broaden the scope of the application of VLP platforms for vaccine use. Viral capsid protein cloning and expression is another option for generating VLPs without incorporating genetic material (van Oosten et al., 2021) . Genetic material-free VLP vaccines are closer to the concept of NP vaccines and can also be used for drug delivery.

Unlike whole-virus vaccines, subunit vaccines need to select antigens, including epitopes with higher immunogenic po-tential. At present, the most popular SARS-CoV-2 antigen is the receptor-binding domain (RBD) located in the spike (S) protein, which comprises the S1 and S2 subunits (Ke et al., 2020) . Immunization with RBD can induce protective humoral immunity against viral pathogens, including SARS-CoV, MERS-CoV, and influenza virus (Jiang et al., 2012; Zhou et al., 2018) . SARS-CoV-2 also has an RBD that plays a primary role in virus entry by interacting with angiotensin-converting enzyme 2 (ACE2). RBD-specific antibodies can efficiently block virus entry (Ju et al., 2020; Seydoux et al., 2020; Shang et al., 2020) . One study suggested that the RBD is more immunogenic than the entire S protein when equal amounts of each antigen are loaded and displayed on the same NP scaffold (Guo et al., 2021a ). An additional advantage of using RBD rather than the whole S protein is reduced antigen size, as the expression of smaller antigens can increase production yield (Salzer et al., 2021) . So far, most NP approaches against SARS-CoV-2 use RBD as an antigen, whereas VLP approaches more frequently use the whole S protein. The S2 subunit is more conserved than the S1 subunit and has a greater potential to induce cross-reactive antibody ( (1) Some NP vaccine approaches express antigen fused with scaffold protein which then form NPs that displays antigen on the outer wall of the particle.

(2) Linker protein can be alternatively expressed on the NP surface to conjugate separately expressed antigen.

(3) Likewise, VLP produced from packaging cell lines can express linker on surface for antigen display. (4) Direct display of antigen on VLP is traditional for generating VLP with foreign antigen. (5) A linking system can be used to generate mosaic vaccine by conjugating mixed antigens. Mosaic vaccine can induce immune response against multiple epitopes and can provide broader-spectrum protection. (B) NP and VLP are purified based on size and concentrated. Isolated vaccine candidates are chemicophysically evaluated before testing for biological safety and efficacy in animal models. Shah et al., 2021) . In one study, the heptad repeat antigen within the S2 subunit efficiently induced neutralizing antibodies against seasonal human coronaviruses, including 229E and OC43 . However, another study reported that the S2 subunit expressing baculovirus-based VLP failed to induce neutralizing antibody responses, while whole S and S1 subunit expressing VLPs were successful (Chu et al., 2021) . SpyTag/SpyCatcher is a paired linking system frequently used in SARS-CoV-2 NP vaccine development (Brune et al., 2016) . This "plug-and-display" system can link the SpyTagantigen to the SpyCatcher-scaffold protein (or vice versa) by spontaneous isopeptide bond formation. Recently, another modification was introduced to SpyTag/SpyCatcher to achieve a more efficient antigen display and enhanced stability . Fc-tagged SARS-CoV-2 RBD has also been coupled with a protein A-expressing VLP scaffold (Geng et al., 2021) . The SnoopLigase-mediated linking system is another platform that can be used for antigen-vaccine conjugation (Andersson et al., 2019) . In an earlier study, both SpyTag/SpyCatcher and SnoopLigase systems were introduced into a single scaffold for twin antigen immunization (Brune et al., 2017) . That study expanded the application range of NPs to achieve stronger immunization by expressing additional antigens in a single vaccination.

There is a demand for the development of a novel linking system as people may possess linker-specific immune responses or develop such responses by repeated immunizations. Due to its bacterial origin, 98% of the tested human samples were positive for antibodies specific for SpyTag/ SpyCatcher protein . It is not clear if linker-specific antibodies affect the immunogenicity of NP and VLP vaccines. These concerns may be overcome by chemical conjugation methods or by introducing genetic fusion so that antigen and scaffold proteins are expressed in a single system (Patel and Swartz, 2011; Charlton Hume and Lua, 2017; Lim et al., 2021) .

Since NPs and VLPs use diverse cell lines for expression, the process itself can differ by platform. Protein subunits for NPs are purified using affinity tags and assembled into NPs, whereas VLPs are generally produced from packaging cell lines that are transiently transfected with viral gene-coding plasmids. An additional assembly step is required in the case of a linking system used for antigen display. NPs and VLPs separated by size-exclusion chromatography collected particles in the desired size ranges. Once NP and VLP particles are ready, many of their evaluation steps are common (Fig.  1B) . NP and VLP size and morphology are evaluated by electron microscopy. Further biophysical techniques, such as dynamic light scattering and high-performance size-exclusion chromatography, are also used for size evaluation. It is also important to analyze the trimeric antigen formation and glycosylation patterns for COVID-19 vaccine development.

Although the correlation between glycosylation and immunogenicity is not obvious, homogenous glycosylation is a parameters that requires assessment (van Oosten et al., 2021) . These biochemical analyses typically employ mass spectro-metry techniques. In addition, protein stability evaluation is important in NP vaccine development as it affects the storage period and temperature requirements. Biolayer interferometry is useful for assessing antigenic characteristics by measuring the interaction between RBD or whole S antigen displayed on vaccine candidates and immobilized ACE2 (Forssén et al., 2020) .

Finally, vaccine candidates are tested for safety and immunogenicity. The mouse model is most frequently used to evaluate COVID-19 vaccine candidates. As laboratory mice (e.g., BALB/c, C57BL/6) lack ACE2 expression, K18-hACE2 transgenic mice are used instead (Winkler et al., 2020) . Syrian golden hamsters, ferrets, rabbits, and cynomolgus macaques have also been used for vaccine evaluation (Sia et al., 2020; Brouwer et al., 2021; Kim et al., 2021b) . Vaccine candidates are prepared with or without adjuvants and introduced to animals to monitor antibody responses. Enzyme-linked immunosorbent assay (ELISA) is a basic method used to evaluate antigen-specific antibody production. Neutralizing antibody production can be assessed by plaque reduction neutralization test; however, this test requires a biosafety level 3 facility due to the use of live virus. Thus, alternative methods for measuring the levels of neutralizing antibodies, such as competition ELISA have been developed (Wouters et al., 2021) . Levels of protection against SARS-CoV-2 infection are determined by the virus burden from nasal swabs and lungs, histopathology, weight loss, and survival rate.

I53-50, an icosahedral protein scaffold generated by computational engineering, can display 120 antigens in a trimeric configuration (Bale et al., 2016) . Separately expressed trimeric antigens (containing I53-50A or I53dn5B) and scaffolds (I53-50B or I53dn5A) are assembled in vitro. This system was first used for respiratory syncytial virus (RSV), influenza, and human immunodeficiency virus (HIV) vaccine development (Brouwer et al., 2019; Marcandalli et al., 2019; Boyoglu-Barnum et al., 2021) . These I53-50-based vaccine candidates induced robust antibody responses in animal models. Notably, the NP scaffold itself acted as an immunogen and elicited specific antibody responses (Marcandalli et al., 2019) . However, NP scaffold-specific antibody responses were reduced when the scaffolds were masked by antigens. Pre-immunization with the NP scaffold did not affect the antibody response against subsequent immunization with antigen-containing NPs, suggesting that NP scaffold-specific antibodies do not adversely affect vaccine efficacy. For COVID-19 vaccine development, self-assembling scaffold components (I53-50B) were expressed by E. coli and trimeric RBDs were purified from HEK293F supernatant (Walls et al., 2020) . Assembled RBD-I53-50 NPs elicited robust antibodies in immunized mice and nonhuman primates. Although this review does not discuss the role of adjuvants in NP vaccine development, adjuvant selection was not critical for the immunogenicity of the RBD-I53-50 NP vaccine (Arunachalam et al., 2021) . I53-30 is the closest COVID-19 NP vaccine for human use and is currently in clinical trials (NCT04742738, NCT05007951). Another study tested modified S proteins instead of RBD in the I53-50 setting (Brouwer et al., 2021) . Immunization with S-I53-50 NP induced antibodies against multiple neutralizing targets and protected macaques from SARS-CoV-2 infection (Brouwer et al., 2020 (Brouwer et al., , 2021 .

Ferritin is a common natural protein cage that stores iron. It is found in animals, including humans and some bacteria. Helicobacter pylori ferritin forms a 24-mer structure containing eight threefold axes that can be used to display trimeric viral spikes, such as influenza hemagglutinin (Kanekiyo et al., 2013) . Ferritin has also been used for HIV (He et al., 2016) , Epstein-Barr virus (Kanekiyo et al., 2015) , RSV (Swanson et al., 2020) , and MERS-CoV (Kim et al., 2018) vaccines. Several potential COVID-19 vaccines have also used ferritin as a scaffold system, mostly using RBD as an antigen He et al., 2021; Kang et al., 2021; Kim et al., 2021b; King et al., 2021; Powell et al., 2021) . Only one study used the entire S subunit (Wuertz et al., 2021) . While most of the COVID-19 vaccine studies used H. pylori ferritin, human and horse ferritin have also been evaluated (Qi et al., 2018; Kalathiya et al., 2021) . Deletion of the 70 aa C-terminal in ferritin showed enhanced neutralizing antibody responses (Powell et al., 2021) . mi3 NP, a mutant form of a rationally designed i301 nanocage, showed increased production yield, particle uniformity, and stability compared to the original structure (Bruun et al., 2018) . This self-assembling NP scaffold can display 20 trimeric antigens (60-mer). The platform is thermally stable and, therefore, useful for vaccine supply, where cold-chain storage is problematic ( IMX313 contains the oligomerization domain of an avian C4-binding protein (Ogun et al., 2008) . During early development, antigen-conjugated IMX313 was designed to be expressed by the modified vaccinia virus Ankara (MVA). MVA-expressing Mycobacterium tuberculosis 85A-IMX313 showed an adjuvant effect and successfully induced CD4 + and CD8 + T cell responses in animal models (Spencer et al., 2012) . Similarly, the Plasmodium falciparum Pfs25-IMX313 complex was developed as a malaria vaccine (Li et al., 2016) . Psf25-IMX313 was delivered intramuscularly using a protein-in-adjuvant regimen, which demonstrated its potential for protein-based vaccine use (Li et al., 2016; Brod et al., 2018) . DNA vaccine approaches have also been used to express HIV Tat-IMX313 (Tomusange et al., 2016) . IMX313-based M. tuberculosis and malaria vaccine candidates showed both safety and immunogenicity in human clinical trials (Minhinnick et al., 2016; de Graaf et al., 2021) . Although IMX313 is a wellestablished NP system, it has not been used for COVID-19 vaccine development, possibly because NP systems that express antigens in trimeric configurations (i.e., I53-50, mi3, ferritin) are more suitable for displaying SARS-CoV-2 S protein or RBD.

VLP vaccines are sometimes indistinguishable from NP vaccines. When genetic material-free viral capsid structures are adopted for vaccine use, VLPs fit perfectly into the self-assembled and protein-based NP category. Due to this NP-like character, the protein coat of the AP205 bacteriophage (AP205 VLP) frequently utilizes the SpyTag/SpyCatcher linking system to display antigens (Brune et al., 2016; Leneghan et al., 2017; Cohen et al., 2021b; Guo et al., 2021b; van Oosten et al., 2021) . While one study comparing AP205 VLP and other NP vaccines, such as mi3 and IMX313, suggested that VLPs induced stronger antibody responses than NPs (Leneghan et al., 2017) , other studies showed comparable antibody responses (Bruun et al., 2018; Cohen et al., 2021b) . More recently, in a COVID-19 vaccine study in animals, AP205 VLPs induced specific antibody responses (Guo et al., 2021a; Liu et al., 2021a; van Oosten et al., 2021) . Phage display is another way to use a bacteriophage for COVID-19 vaccine development. The dual display of the SARS-CoV-2 S protein epitope and the α 3 β 1 integrin ligand on the f88-4 bacteriophage efficiently induced antibody responses against the S protein (Staquicini et al., 2021) . Many other plant virus-and eukaryotic virus-based pseudovirus platforms have been investigated for use in COVID-19 vaccines. Human type 5 adenovirus (Ad5) is the best-studied and is widely used as a replication-defective vector system. Ad5-based VLPs induced immunity against SARS-CoV-2 and protected mice and rhesus macaques from infection (Feng et al., 2020; Wu et al., 2020) . Clinical studies of the Ad5 COVID-19 vaccine have demonstrated its safety and immunogenicity in human adults (Zhu et al., 2020b; . Another study showed that chimpanzee adenovirus (ChAd) also induced robust antibody responses against SARS- CoV-2 (Hassan et al., 2020) , in which the intranasal administration of a ChAd-based vaccine induced both S protein-specific IgG and IgA antibodies. In another study, a gene sequence from tetanus toxin was introduced to cucumber mosaic virus (CuMV) to boost T cell responses (Zeltins et al., 2017) . CuMV VLP-expressing SARS-CoV-2 receptor-binding motif (RBM) induced an antibody response against RBD (Mohsen et al., 2022) . More VLP platforms show promise for COVID-19 vaccines, including those for Newcastle disease virus (Sun et al., 2020; Yang et al., 2021) , rabies virus (RABV) (Kurup et al., 2020) , and vesicular stomatitis virus (VSV) (Case et al., 2020; Yahalom-Ronen et al., 2020; Hennrich et al., 2021) . RABV and VSV belong to the Rhabdoviridae family and have long been used for foreign antigen display. Glycoprotein was deleted to make replication-defective VSV, and RABV VLPs were tested in both live and inactivated settings. Inactivated RABV VLPs showed higher antibody responses than live VLPs (Kurup et al., 2020) . While most VLP approaches are not adjuvanted, this study also tested adjuvants and confirmed that adjuvants induced stronger antibody responses and changed the IgG2a/IgG1 ratio (Kurup et al., 2020) . In another study using NDV VLPs, adjuvant was effective when lower dose of antigen was used (Sun et al., 2020) . While most NP studies used RBD as an antigen, those on VLPs more frequently used the whole S protein. Numerous attempts have also been made to enhance immunogenicity by displaying the S protein. For example, when whole SARS-CoV-2 S or 19 amino acid-deleted S (DS) proteins were displayed on Moloney murine leukemia virus (MLV) VLPs, DS was more efficiently loaded on the VLP surface (Roy et al., 2021) . Although that study did not compare S and DS antigens for immunogenicity in animal studies, more compact antigen presentation will increase vaccine efficacy. Other studies substituted two residues (986 and 987), deleted the polybasic cleavage site, and optimized codons to enhance protein stability and increase VLP yield (Hassan et al., 2020; Sun et al., 2020; Wu et al., 2020) . As the whole S protein can induce antibodies that are cross-reactive with SARS-CoV (Wuertz et al., 2021) , current VLP approaches are somewhat more advantageous than NPs that mainly use RBD in terms of broader protection against SARS-CoV-2 variants and newly emerging highly pathogenic coronaviruses.

The protective efficacy of vaccines is determined by their capacity to induce potent antibody responses, as neutralizing antibody titers have been considered the mechanistic correlates of protection (Hogan and Pardi, 2021) . This has been confirmed in COVID-19 vaccine studies Khoury et al., 2021) . However, the vaccine-induced activation of T cells is equally important. T follicular helper (Tfh) Fig. 2 . Immunological advantages of nanoscale vaccines. Overview of immunogenic parameters of nanoparticle (NP) and virus-like particle (VLP) vaccines in inducing potent antibody, T cell, and B cell responses. (A) Accelerated engagement of nanoscale vaccines with the innate immune system at immunization site. Upon immunization, characteristic features of nanoscale vaccines, including high density of multimeric antigen, size, charge, and hydrophobicity, collectively influence their greater interactions with the innate immune system and trafficking to draining lymph nodes, leading to local inflammation, infiltration of innate immune cells, and uptake and processing of antigen by antigen-presenting cells (APCs). (B) Strong induction of humoral and cellular immunity in draining lymph nodes. Higher antigen density improves B cell receptor (BCR) cross-linking with multiple antigens, which lowers B cell activation threshold and thereby induces potent B cell activation. Also, enhanced antigen presentation by APCs induces strong T cell activation. Vaccine-induced T follicular helper (Tfh) cells provide help to cognate B cells for antibody class switching, affinity maturation, and generation of long-lived antibody-secreting plasma cells and memory B cells. DC, dendritic cells. cells assist B cells in producing high-affinity antibodies as well as differentiation into memory B cells and long-lived antibody-secreting plasma cells (Crotty, 2019). As antibody titers decline and escape variants emerge, neutralizing antibodies often fail to maintain sterilizing immunity against SARS-CoV-2 infection (Noh et al., 2021) . In this scenario, memory T cells are the ultimate agent of protection by eliminating infected cells, thereby limiting the spread of infectious agents. Indeed, the depletion of CD8 T cells impaired the protection of convalescent macaques against SARS-CoV-2 when antibody titers were suboptimal (McMahan et al., 2021) . Accordingly, understanding the precise immune mechanisms elicited by vaccines is essential to determine their protective efficacy (Fig. 2) . During a vaccine response, immunization triggers local inflammation and infiltration of innate immune cells, including antigen-presenting cells (APCs) such as monocytes, macrophages, and dendritic cells (DCs). APCs capture antigens at the site of immunization, migrate to draining lymph nodes, and present processed peptide antigens to cognate T cells that subsequently proliferate and differentiate into effector and memory T cells (Schenten and Bhattacharya, 2021) . Antigens trafficked to draining lymph nodes are also directly recognized by B cells through their B cell receptors (BCRs) on the cell surface, triggering B cell activation, proliferation, and differentiation (Irvine and Read, 2020).

Traditional vaccine platforms, including inactivated and live attenuated viral vaccines, are highly effective in inducing antibody and T cell responses, but are not safe and may be potentially infectious, particularly in older individuals and adults with underlying diseases (Kim et al., 2017) . Conversely, recombinant protein-based subunit vaccines are safe but often less immunogenic and suboptimal, requiring adjuvants to induce inflammation (Nguyen and Tolia, 2021) . Adenovirus-based vaccines are another successful platform. Unfortunately, the prevalence of common adenovirus serotypes in humans along with immunization history could limit the selection of adenovirus serotypes to evade pre-existing immunity against the adenoviral vector itself in the vaccine development of newly emerging infectious diseases (Zhu et al., 2020a) . Recent mRNA vaccines have shown high efficacy in protecting against SARS-CoV-2, with durable antibody and T cell responses (Polack et al., 2020; Baden et al., 2021; Hogan and Pardi, 2021) . However, the nature of inflammatory factors and associated effector molecules responding to synthetic mRNA and lipid-NP carriers remain unresolved (Hogan and Pardi, 2021) and may lead to unexpected side effects such as myocarditis (Kim et al., 2021a) . In addition, as type I interferon is a potent negative regulator of Tfh cell differentiation (Ray et al., 2014) , the intrinsic property of synthetic mRNA in inducing a strong type I interferon response could skew CD4 T cell responses toward T helper type 1 (Th1) cell differentiation and away from desirable Tfh cell development (Guerrera et al., 2021) . NP vaccines have demonstrated enhanced vaccine responses and greater protection compared to traditional subunit vaccines. The high density of repetitive protein antigen is a major factor that contributes to the higher immunogenicity (Kelly et al., 2019; Brouwer et al., 2021) . Multimeric protein antigens have long been associated with improved antibody responses for several reasons (Slifka and Amanna, 2019) . Monomeric soluble antigens have limited interactions with the BCRs of the cognate B cells. In contrast, a high density of surface antigens leads to stronger interactions with multiple BCRs, which promote BCR cross-linking, activation of downstream signaling cascades, and potent B cell activation (Bachmann and Zinkernagel, 1997) . Activated B cells can, in turn, interact with cognate Tfh cells and receive essential signals required for class switching, affinity maturation, and generation of long-lived antibody-secreting plasma cells and memory B cells (Crotty, 2019) . Importantly, stronger interactions of antigens with multiple BCRs increase overall avidity to the cognate antigen and, therefore, lower the activation threshold of B cells (Bachmann and Zinkernagel, 1997) . Accordingly, B cells with relatively weak avidity to antigens may also be recruited into the responses, broadening the repertoires of B cells responding to the antigen, as shown by one vaccine study that used multivalent HIV antigens (Abbott et al., 2018) . Along with B cells, antigen multivalency with optimal size also facilitates an enhanced recognition of antigens by the innate immune system. Rapid recognition of NPs by the complement system induces local inflammation and antigen transfer to APCs. Multivalent antigens are also directly recognized by APCs, including DCs (Kelly et al., 2019) . Collectively, accelerated antigen uptake and processing by APCs leads to greater antigen presentation to cognate T cells in the draining lymph nodes to promote T cell activation and differentiation. Beyond antigen multivalency, other factors, including NP size, charge, and trafficking to draining lymph nodes, contribute to vaccine efficacy (Gause et al., 2017) . For example, NP size controls lymph node trafficking, with sizes of 20-30 nm freely trafficking to draining lymph nodes, while APCs are strictly required for the transport of larger NPs from the injection site to lymph nodes (Reddy et al., 2006; Manolova et al., 2008) . Although results have varied, studies have shown that size also influences the distribution of particles within draining lymph nodes, with 20-nm polystyrene beads preferentially localized in the B cell area to facilitate their interactions with B cells (Manolova et al., 2008) . Such size-dependent changes in antigen trafficking to lymph nodes appear to be functionally important, as 40-nm polystyrene beadconjugated OVA antigen induced the highest CD4 and CD8 T cell responses as well as potent antibody production compared to larger particle sizes in immunized mice, presumably due to the enhanced antigen presentation by DCs in the lymph nodes (Fifis et al., 2004) . The biochemical properties of NP surfaces are important determinants. Positively charged surfaces preferentially interact with negatively charged membranes of innate immune cells (Fromen et al., 2015) . Similarly, hydrophobic NPs lead to increased cellular interactions, resulting in enhanced immune responses (Moyano et al., 2012) . The major advantage of NP vaccines is their ability to modulate these properties, including antigen multivalency, size, charge, and lymph node trafficking, to improve antibody and T cell responses and, ultimately, vaccine efficacy.

There is ongoing debate regarding the usefulness of the COVID-19 vaccine in the post-pandemic era as it is difficult to predict the future of COVID-19 (Telenti et al., 2021) . Despite systematic studies suggesting the possibility of seasonal circulation of COVID-19 (Audi et al., 2020; , it is not clear whether virulence will remain sufficiently high to require vaccination programs. SARS-CoV-2 may disappear or become endemic within restricted areas, similar to previous SARS-CoV and MERS-CoV outbreaks (Al-Omari et al., 2019) . Nonetheless, a rapid vaccine production platform is critical to minimize the impact of unexpected respiratory virus pandemics. The introduction of new types of vaccines against SARS-CoV-2 undoubtedly reduced morbidity and mortality rates among high-risk groups (McNamara et al., 2021) . However, young people, who have lower SARS-CoV-2 morbidity and mortality, need assurance of greater safety (Brandt et al., 2021) . The adjuvanted subunit vaccines are already in the vaccine market, and accumulating studies have evaluated their safety and efficacy. Protein-based NP vaccines are believed to be similarly safe and effective for human use. Costeffectiveness will be an ongoing concern once multiple fastresponding vaccine platforms are prepared. Therefore, continuous efforts to reduce the costs for NP and VLP vaccine production will be the key to competitive pricing. Production of the NP vaccines in yeast may lower costs (Dalvie et al., 2021) . Also, for NP and VLP production, use of linking systems that adopt E. coli to express scaffold protein may be advantageous. In these systems, the folding process or posttranslational modification is less strict (Kang et al., 2021) ; however, E. coli lacks post-translational modification which may affect glycosylated antigen expression. Further, E. coli expression may pose a problem with protein folding. However, E. coli-based expression of influenza and MERS-CoV vaccine candidates successfully induced neutralizing antibody responses, suggesting that E. coli is an attractive expression system for NP vaccine production (Kim et al., 2018; Lim et al., 2021) . Further developments employing the innovative chaperone technology will enhance E. coli-based soluble expression systems and provide an efficient method for NP vaccine production (Son et al., 2021) . Similarly, the plant-based generation of VLPs may be useful for high-yield vaccine production (Maharjan and Choe, 2021) .

Efforts are ongoing to develop a universal influenza vaccine to control continuous seasonal outbreaks and occasional pandemics (McMillan et al., 2021) . Likewise, prophylactic vaccination with a universal vaccine may be beneficial in controlling seasonal and new coronavirus pandemic outbreaks. Although some RBD-specific antibody clones isolated from patients with COVID-19 show cross-reactivity to SARS-CoV or MERS-CoV, they have greater reactivity to SARS-CoV-2, suggesting that a single RBD is not sufficient to induce immunity against a broad spectrum of coronaviruses. To overcome this problem, one group recently developed mosaic NPs that display multiple RBDs from distinct clades, and this vaccine candidate elicited immunity against heterologous coronavirus strains (Cohen et al., 2021a) . Alternatively, a mixture of NPs that express different RBDs can induce broader immune responses. NP and VLP vaccines also have inherent hurdles to overcome to make them more attractive for the COVID-19 vac-cine platform. Safety issues that might arise from the pathogen-derived features of VLPs and the use of immune-stimulatory adjuvants are critical for vaccine development. Some adverse effects have been reported following AS04-adjuvanted human papillomavirus VLP vaccination (Harris et al., 2014) . As traditional aluminum adjuvants are less effective in weak immune responders (Miller et al., 1999) , newer adjuvants such as pattern-recognition receptor agonists are necessary despite their pro-inflammatory potential (Wu, 2016) . Compared to monomeric subunit vaccines, NP and VLP approaches demand more thorough validation for antigen and scaffold protein expression because incorrect folding significantly disrupts particle formation. Molecular engineering can enhance antigen stability and solubility. For example, modification of the VLP capsid to remove the surface charge increased solubility and improved assembly (Lu et al., 2015) . Despite these concerns, NP and VLP technologies are flexible platforms with remarkable expandability. Further studies will identify promising options for vaccine development for the current COVID-19 pandemic and future coronavirus outbreaks.

Proc. Natl. Acad. Sci. USA

Nanoparticle vaccines based on the receptor binding domain (RBD) and heptad repeat (HR) of SARS-CoV-2 elicit robust protective immune responses

Plant-based COVID-19 vaccines: current status, design, and development strategies of candidate vaccines

Nanoparticles target distinct dendritic cell populations according to their size

Induction of potent neutralizing antibody responses by a designed protein nanoparticle vaccine for respiratory syncytial virus

Correlates of protection against SARS-CoV-2 in rhesus macaques

The next generation of influenza vaccines: towards a universal solution

Estimating the early impact of the US COVID-19 vaccination programme on COVID-19 cases, emergency department visits, hospital admissions, and deaths among adults aged 65 years and older: an ecological analysis of national surveillance data

Protective effect of hepatitis B vaccine in chronic hemodialysis patients

A first-in-human phase 1 trial to evaluate the safety and immunogenicity of the candidate tuberculosis vaccine MVA85A-IMX313, administered to BCG-vaccinated adults

A scalable and highly immunogenic virus-like particle-based vaccine against SARS-CoV-2

The immune response and immune evasion characteristics in SARS-CoV, MERS-CoV, and SARS-CoV-2: vaccine design strategies

Nanoparticle hydrophobicity dictates immune response

Protein-based antigen presentation platforms for nanoparticle vaccines. npj Vaccines 6, 70

The oligomerization domain of C4-binding protein (C4bp) acts as an adjuvant, and the fusion protein comprised of the 19-kilodalton merozoite surface protein 1 fused with the murine C4bp domain protects mice against malaria

Surface functionalization of viruslike particles by direct conjugation using azide-alkyne click chemistry

Coronavirus infections-more than just the common cold

Safety and efficacy of the BNT162b2 mRNA Covid-19 vaccine

A single immunization with spike-functionalized ferritin vaccines elicits neutralizing antibody responses against SARS-CoV-2 in mice

Intranasal nanovaccine confers homo-and hetero-subtypic influenza protection

Overcoming symmetry mismatch in vaccine nanoassembly through spontaneous amidation

Transcription factor STAT3 and type I interferons are corepressive insulators for differentiation of follicular helper and T helper 1 cells

In vivo targeting of dendritic cells in lymph nodes with poly(propylene sulfide) nanoparticles

A stable platform for the production of virus-like particles pseudotyped with the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) spike protein

Single-dose immunisation with a multimerised SARS-CoV-2 receptor binding domain (RBD) induces an enhanced and protective response in mice

Immunology of SARS-CoV-2 infections and vaccines

Analysis of a SARS-CoV-2-infected individual reveals development of potent neutralizing antibodies with limited somatic mutation

The case for S2: the potential benefits of the S2 subunit of the SARS-CoV-2 spike protein as an immunogen in fighting the COVID-19 pandemic

Cell entry mechanisms of SARS-CoV-2

Different applications of virus-like particles in biology and medicine: vaccination and delivery systems

Pathogenesis and transmission of SARS-CoV-2 in golden hamsters

Role of multivalency and antigenic threshold in generating protective antibody responses

Chaperna: linking the ancient RNA and protein worlds

Fusion of the Mycobacterium tuberculosis antigen 85A

to an oligomerization domain enhances its immunogenicity in both mice and non-human primates

Design and proof of concept for targeted phage-based COVID-19 vaccination strategies with a streamlined cold-free supply chain

A respiratory syncytial virus (RSV) F protein nanoparticle vaccine focuses antibody responses to a conserved neutralization domain

A COVID-19 vaccine candidate using SpyCatcher multimerization of the SARS-CoV-2 spike protein receptor-binding domain induces potent neutralising antibody responses

Obstacles and advances in SARS vaccine development

After the pandemic: perspectives on the future trajectory of COVID-19

Two-component nanoparticle vaccine displaying glycosylated spike S1 domain induces neutralizing antibody response against SARS-CoV-2 variants

Elicitation of potent neutralizing antibody responses by designed protein nanoparticle vaccines for SARS-CoV-2

Immunogenicity and protective efficacy of replicationincompetent influenza virus-like particles

SARS-CoV-2 infection of human ACE2-transgenic mice causes severe lung inflammation and impaired function

Discovery of seven novel Mammalian and avian coronaviruses in the genus Deltacoronavirus supports bat coronaviruses as the gene source of Alphacoronavirus and betacoronavirus and avian coronaviruses as the gene source of Gammacoronavirus and Deltacoronavirus

A novel competition ELISA for the rapid quantification of SARS-CoV-2 neutralizing antibodies in convalescent plasma

Strategies for designing synthetic immune agonists

Safety, tolerability, and immunogenicity of an aerosolised adenovirus type-5 vector-based COVID-19 vaccine (Ad5-nCoV) in adults: preliminary report of an openlabel and randomised phase 1 clinical trial

A single dose of an adenovirus-vectored vaccine provides protection against SARS-CoV-2 challenge

A SARS-CoV-2 spike ferritin nanoparticle vaccine protects hamsters against Alpha and Beta virus variant challenge

A single dose of recombinant VSV-G-spike vaccine provides protection against SARS-CoV-2 challenge

Activation of dendritic cells by human papillomavirus-like particles through TLR4 and NF-κB-mediated signalling, moderated by TGF-β

Newcastle disease virus-like particles displaying prefusion-stabilized SARS-CoV-2 spikes elicit potent neutralizing responses

DNA vaccine protection against SARS-CoV-2 in rhesus macaques

Incorporation of tetanus-epitope into virus-like particles achieves vaccine responses even in older recipients in models of psoriasis, Alzheimer's and cat allergy

Interaction between virus-like particles (VLPs) and pattern recognition receptors (PRRs) from dendritic cells (DCs): toward better engineering of VLPs

A platform incorporating trimeric antigens into self-assembling nanoparticles reveals SARS-CoV-2-spike nanoparticles to elicit substantially higher neutralizing responses than spike alone

Prospects for a MERS-CoV spike vaccine

Immunogenicity and safety of a recombinant adenovirus type-5-vectored COVID-19 vaccine in healthy adults aged 18 years or older: a randomised, double-blind, placebo-controlled, phase 2 trial

Safety, tolerability, and immunogenicity of a recombinant adenovirus type-5 vectored COVID-19 vaccine: a dose-escalation, open-label, nonrandomised, first-in-human trial

Assessing sequence plasticity of a virus-like nanoparticle by evo-