Virus-like Particle Vaccines and Platforms for Vaccine Development

Virus-like particles (VLPs) have gained a lot of interest within the past two decades. The use of VLP-based vaccines to protect against three infectious agents—hepatitis B virus, human papillomavirus, and hepatitis E virus—has been approved; they are very efficacious and offer long-lasting immune responses. Besides these, VLPs from other viral infectious agents (that infect humans, animals, plants, and bacteria) are under development. These VLPs, especially those from human and animal viruses, serve as stand-alone vaccines to protect against viruses from which the VLPs were derived. Additionally, VLPs, including those derived from plant and bacterial viruses, serve as platforms upon which to display foreign peptide antigens from other infectious agents or metabolic diseases such as cancer, i.e., they can be used to develop chimeric VLPs. The goal of chimeric VLPs is to enhance the immunogenicity of foreign peptides displayed on VLPs and not necessarily the platforms. This review provides a summary of VLP vaccines for human and veterinary use that have been approved and those that are under development. Furthermore, this review summarizes chimeric VLP vaccines that have been developed and tested in pre-clinical studies. Finally, the review concludes with a snapshot of the advantages of VLP-based vaccines such as hybrid/mosaic VLPs over conventional vaccine approaches such as live-attenuated and inactivated vaccines.


Introduction
Viruses infect a wide range of organisms, ranging from humans, plants, animals, birds, and insects to microorganisms (e.g., eukaryotes and prokaryotes). One unique feature of viruses is that their structural proteins, envelope proteins, or capsid proteins, along with other structural proteins, can, either independently or collectively, spontaneously self-assemble to form virus-like particles (VLPs) without the viral genome. Thus, any virus can be utilized to develop VLPs; VLPs can be developed by cloning the structural genes that code for the proteins of a virus of interest into an expression vector ( Figure 1). The expression vector depends on the expression system in which the protein(s) will be expressed and, sometimes, the protein is codon-optimized if the expression system (including that of mammalian, insects, and bacteria) is different to the cells which the virus of interest normally infect. The vector harboring the DNA of the structural protein(s) is then transfected/transformed into cells of interest, where the DNA is transcribed and translated. Translated protein folds and assembles to form VLPs [1]. VLPs have many applications in biomedical sciences, such as: (i) therapy-the delivery of drugs/cargo to specific cancer cells; (ii) in vivo imaging-VLPs loaded with fluorophores; (iii) diagnostic tests-utilizing armored RNA as positive controls for infectious diseases; iv) vaccine development [2][3][4]. This review focuses only on the latter: the application of VLPs in vaccine development.
VLPs have many features, unlike conventional vaccines, which make them very attractive platforms for vaccine design. They mimic the viruses from which the VLPs are derived in terms of size (20-200 nm) [5], geometry (i.e., icosahedral structures with multivalent epitopes) [5][6][7][8], and the ability to activate T-helper cells. VLPs naturally encode T-helper cell epitopes, which are presented to T-helper cells by antigen-presenting cells (APCs) in association with major histocompatibility complex (MHC) class II. The Figure 1. A schematic illustrating the generation of VLPs and chimeric VLPs. A coat protein from a virus is cloned to an expression vector. The vector with the coat protein can also be used to insert a foreign peptide into the coat protein (bottom). Each vector is then transformed or transfected to an expression system where the proteins are expressed and assembled into VLPs. Moreover, VLPs are considered safe, since they do not contain the viral genome, and therefore, they cannot replicate. It is worth mentioning that, like any vaccine, VLPs can cause side effects such as pain and swelling at the injection site. Hence, it is no surprise that VLPs have gained considerable attention over the past two decades as an attractive platform for vaccine design. While VLPs derived from human and animal viruses are developed to protect humans and animals against viruses from which the VLPs are derived, VLPs from other organisms (including plants and prokaryotic microorganisms) are developed to display heterologous antigens from human and animal viruses, i.e., to develop chimeric VLPs (Figure 1, bottom). The goal of chimeric VLPs is to elicit immune responses against heterologous antigens displayed on the platform and not necessarily the platform. It is worth mentioning that VLPs derived from human or animal viruses can also be used to develop chimeric VLPs, whereby a heterologous antigen displayed on the VLP is derived from other human viruses (Figure 1, bottom). A schematic illustrating the generation of VLPs and chimeric VLPs. A coat protein from a virus is cloned to an expression vector. The vector with the coat protein can also be used to insert a foreign peptide into the coat protein (bottom). Each vector is then transformed or transfected to an expression system where the proteins are expressed and assembled into VLPs.

Figure 2.
The activation of immune responses by VLPs. Antigen-presenting cells phagocytose and process VLPs into fragments, which are presented to T-helper cells with the help of MHC class II (tope image). This leads to the activation of T-helper cells, which secrete cytokines that activate B-cells (below). B-cells are also activated by the cross-linking of B-cell receptors (BCRs) by VLPs. Activated B-cells divide and differentiate into plasma cells and memory cells (not shown here). Plasma cells secrete antibodies into the body, which neutralize the virus of interest from which the VLPs were derived. The figure is adopted from [5].

VLP Vaccines Derived from Human Viruses
VLPs have been developed (while others remain under development) using viruses that infect humans (Table 1). In fact, VLP-based vaccines have been approved to protect against three human viral infections, namely, hepatitis B virus (HBV), human papillomaviruses (HPV), and hepatitis E (HEV). Among these, at least nine HBV VLP-based vaccines have been approved globally, and these include Engerix-B, Recombivax HB, Euvax B, Hepavax-Gene, GenHevac B, GenVac B, and Heberbiovac HB (this is reviewed in [1,18,19]). The vaccines are based on the HBV surface antigen; some of the vaccines; for example, Engerix-B and Recombivax HB, have been in use since the mid-1980s, and they offer cross-protection against other HBV genotypes, such as A and C [20]. Although the vaccines cross-protect against other genotypes, several vaccine-escape viral mutants have been reported for the vaccines (reviewed in [21]). Regardless of this, HBV protective antibodies last up to 30 years in some individuals [22,23]. Age at the time of vaccination seems to be associated with HBV antibody responses. For example, individuals vaccinated at an average age of 36 years developed fewer anti-HBV antibodies compared to those who were vaccinated at an average age of 32 years. However, a single boost enhanced antibody levels in 94% of individuals who had low levels of antibodies [22,23].
In addition to the approved VLP-based vaccines above, several candidate VLP-based vaccines are under development. An excellent new book by Pumpens and Pusko provides a comprehensive summary of candidate VLP-based vaccines that have been developed from various viruses [1]. VLPs that have been developed and tested include influenza viruses, SARS-CoV-2, human immunodeficiency virus, and the Zika virus. Table 1 summarizes some of the VLPs developed from human viruses in pre-clinical and clinal trails. The VLPs are derived from either the envelope protein, the capsid protein, or both and are expressed using various expression systems, such as mammalian cells, insect cells, yeast cells, and bacterial cells, as well as transgenic plants. The efficacy of some of these VLPs is comparable to those of approved vaccines developed using conventional approaches. For instance, immunization with hybrid VLPs developed using structural proteins from two influenza A subtypes-H1N1 and H3N2-is as immunogenic as an approved inactivate influenza virus vaccine (Vaxigrip; H2N3); IgG responses had a normalized median fluorescence intensity of >10 7 [43]. Similarly, the immunization of mice with VLPs derived from the Zika virus or Ebola virus offers a survival rate of up to 100%, unlike control mice, where there was no survival [44,45].
VLPs derived from human viruses have also been used to develop candidate vaccines against other human viruses. For example, VLPs derived from HBV have been used to develop candidate vaccines against Helicobacter pylori (H. pylori) bacteria, hepatitis C virus (HCV), HPV-associated cancers, and four serotypes of dengue virus ( Table 2). In preclinical studies, mice immunized with chimeric HBV VLPs displaying peptides derived from H. pylori had a reduced bacterial burden [46]. In other studies, sera from mice immunized with HBV VLPs displaying peptides from HCV and dengue viruses neutralized and protected mice from HCV and dengue virus infection, respectively [47,48]. Additionally, HBV VLPs displaying peptides from HPV16 E7 suppress tumors in a mouse model for HPV-associated cancer [49].

VLP Candidate Vaccines Derived from Veterinary Viruses
VLPs derived from viruses that infect animals and fish have also been used to develop vaccines against viruses that affect various animals, including pigs, rabbits, sheep, horses, chickens, and fish (Table 3). For instance, VLPs derived from porcine circovirus-2 (PCV-2) elicit better antibody responses in mice, superior to those of a commercial subunit vaccine [86]. PCV-2 is associated with postweaning multisystem wasting syndrome in young pigs. In sheep, immunization with VLPs derived from bluetongue virus offers the same level of protection as immunization as a live-attenuated commercial vaccine (monovalent BTV-8) [87]. In foals, immunization with chimeric VLPs derived from three African horse sickness viruses-AHSV-6, AHSV-3, and AHSV-1-elicited neutralizing antibodies against these viruses, albeit at low levels [88]. In chickens, immunization with mosaic VLPs derived from different subtypes/strains of influenza A or immunization with VLPs derived from infectious bursal disease virus offered better protection than commercial inactivated vaccines (e.g., H6N2 for influenza virus) [89][90][91] (Table 3). Meanwhile, in fish, immunization with VLPs derived from red-spotted grouper nervous necrosis virus (RGNNV) lowered the mortality rate to~3.3%, compared to immunization with a commercially inactivated vaccine, OceanTect viral nervous necrosis, and the control group (which had mortality rates of 10% and~79%, respectively) [92]. Additionally, the immunization of fish with VLPs derived from Atlantic cod nervous necrosis virus lowered the mortality rate to 14%, compared to immunization in the control group, where the mortality rate was lowered to only~80% [93].
VLPs derived from animal viruses have also been used as platforms to develop candidate vaccines against other viral infections (Table 4). For example, VLPs from canine parvovirus have been used to display peptide antigens from Middle East respiratory syndrome coronavirus (MERS-CoV) and the VLPs elicited a balanced immune response in mice; sera from immunized mice neutralized pseudo-MERS-CoV [94]. Furthermore, immunization with VLPs derived from Newcastle disease virus displaying a Brucella antigen-BCSP31-offered protection against a virulent strain (16M) of Brucella melitensis; the protection level was similar to that of a commercial live-attenuated vaccine: Brucella melitensis strain M5 [95]; Brucella melitensis is a zoonotic disease associated with brucellosis.

Nicotiana benthamian
Single, double, and triple chimeric VLPs were developed; anti-VP7 specific responses were detected in foals immunized with triple chimeric VLP (AHSV-6/AHSV-3/AHSV-1). However, single AHSV-6 VLPs elicited a weak neutralizing humoral immune response against homologous AHSV virus. Low neutralization levels were also observed with a control live-attenuated AHSV-6 vaccine. [88] Goose hemorrhagic polyomavirus (GHPV) VP1 with or without VP2 Capsid Sf 9 insect cells and S. cerevisiae The VLPs expressed by yeast were of smaller size. VLPs (as a diagnostic antigen) detected GHPV-specific antibodies in up to 85.7% of geese sera with hemorrhagic nephritis and enteritis. [100] Canine influenza virus (CIV) H3N2 M1 and hemagglutinin proteins Envelope Sf 9 insect cells Dogs vaccinated with the VLPs and later challenged with CIV H3N2 did not show clinical signs of respiratory disease, unlike control dogs. [91]

Nicotiana benthamiana
VLPs elicited an innate immune response in fish, which was associated with reduced viral replication in the heart, spleen, and kidney of salmon, and reduced inflammatory lesions in cardiomyocytes.

VLP Candidate Vaccines Derived from Plant and Bacterial Viruses
VLPs have been developed not only from human and animal viruses, but also from plant viruses and bacteria viruses. Plant-derived VLPs have been developed using various viruses such as alfalfa mosaic virus (AMV), physalis mottle virus (PhMV), potato virus Y (PVY), cucumber mosaic virus, and malva mosaic virus (MaMV) ( Table 5). VLPs from AMVs displaying an antigen (Pfs25 protein) from Plasmodium falciparum have been shown to be safe and tolerable in a phase I clinical trial [110,111]. PhMV VLPs displaying HER2 peptide were found to slow tumor growth in mice and delay their death [112,113]. In addition to this, PVY VLPs displaying a cat allergen peptide-Feline domesticus-elicited high-antibody titers against the antigen [114]. Moreover, immunization with MaMV-displaying canine influenza virus H3N8 antigen protected mice from lethal challenge with homologous and heterologous mouse-adapted influenza virus strains, unlike control mice [115].

P. pastoris
Immunization with VLPs displaying S9 peptide elicited a Th1 response against the peptide. However, immunization with the peptide conjugated to keyhole limpet hemocyanin elicited a Th2 response. [121] PhMV Capsid protein Capsid HER2-derived CH401 epitope E. coli VLPs elicited a strong immune response. Chimeric VLPs slowed the growth of DDHER2 tumor cells in mice and also delayed death by more than 15 days compared to control mice that were immunized with just the CH401 epitope. [112,113] PVY Capsid protein Capsid HBV preS1 epitope E. coli Chimeric VLPs elicited high-titer (1:8620) anti-HBV preS1 antibodies in mice without adjuvant. [122] Cat allergen Feline domesticus (Fel d 1)

Expert Review Commentary and Future Perspectives
VLP vaccines are proteins, and thus, they cannot be amplified in the body (i.e., transcribed and translated to make more copies) like other recombinant vaccines such as viral vector vaccines and mRNA vaccines. Nevertheless, VLP vaccines are better alternatives to the two recombinant vaccines. VLP vaccines do not need freezing conditions of −20 to −80 • C for transportation and storage, like mRNA vaccines. In addition to this, VLPs are immediately processed by APCs once they are injected into the body, unlike DNA and mRNA vaccines, whereby cells first need to transcribe (for DNA vaccines) and translate their mRNAs into proteins before they can be processed by APCs to be presented to the immune system.
VLP vaccines are also better alternatives to conventional vaccines (live-attenuated vaccines and inactivated vaccines) for several reasons: First, VLP vaccines do not replicate, and thus, they can be used for everyone, including women who are pregnant or people with a compromised immune system. Additionally, coat proteins-including other structural proteins-from viruses that have segmented genomes (e.g., influenza viruses, AHSV, and bluetongue virus) can also be used to develop hybrid/mosaic VLP-based vaccines without fear of genetic reassortment, like in live-attenuated vaccines. Viruses with segmented genomes have the ability to undergo genetic reassortment; unfortunately, a live-attenuated polyvalent vaccine derived from an AHSV led to the virulent reversion of AHSV type 1 and genomic reassortments with segments from AHSV types 1, 3, and 4, leading sporadic outbreaks of AHSV [128]. Second, VLPs closely mimic the structure of authentic viruses, unlike inactivated vaccines, whose structural proteins may be modified during inactivation, leading to compromised immunogenicity [129]. VLP vaccines, therefore, serve as a better alternative to inactivated vaccines. They are also immunogenic at lower doses [16,17]; in fact, studies show that immunization with 0.3 µg of a candidate rabies virus VLP vaccine elicits antibody titers that are comparable to immunization with 3 µg of the VLPs or with veterinary and human vaccines [17].
Although VLP vaccines are immunogenic at smaller doses, it is unclear whether smaller doses of VLPs can provide long-term protection against viruses with different genotypes such as HBV (which has at least nine genotypes [130]), HPV (more than 20 types associated with cancers [29]), or HEV (which has at least seven genotypes [131]). As mentioned above, immune responses to VLP-based vaccines offer cross-protection against different HBV, HPV, and HEV genotypes [20,29,41,42] but less protection against other genotypes or escape mutants (e.g., HBV) [21,41,42]. These vaccines are given at doses of at least 10 µg/immunization (for example, Engerix-B for HBV [132]); therefore, reducing the dose/immunization may not enhance cross-protection. We believe that increasing the concentration of VLPs (antigens in general) or the number of booster doses may enhance cross-protection. For example, studies have shown that cross-protection against an influenza virus subtype and mice survival following challenge are enhanced at higher doses of antigen, for example, 90 µg, compared to 10 µg [133]. Similarly, immunization with a high dose (10 6 plaque forming units) of an ASFV vaccine offers complete cross-protection against an ASFV strain compared to partial cross-protection following immunization with half the dose [134].
VLP-based vaccines against viruses of interest should therefore elicit cross-protective immunity against all genotypes of the virus of interest to ensure global efficacy, given the fact that the distribution of genotypes vary from one geographical region to another. For example, HBV genotypes vary from one continent to another, with genotype A being more prevalent in North America, South America, Africa, and Europe, while genotypes B and C are prevalent in Southeast and East Asia. Genotypes D is prevalent in west-ern/central/southern Asia and in Europe [135]. A VLP-based HBV vaccine for global use should offer protection against all HBV genotypes irrespective of where an individual lives. The same applies to HPV-and HEV-based vaccines. Although efforts have been made with the development of Gadarsil-9 to broaden the spectrum of protection against the oncogenic HPV types (HPV16, 18, 31, 33, 45, 52, and 58) prevalent in other parts of the world than those mentioned above, there are still some HPV types (HPV35, 39, 51, 59, etc.) associated with 10% of cervical cancer against which the vaccine does not protect (this is reviewed in [29]). For example, while Gadarsil-9 offers protection against the majority of HPV types associated with~90% of cervical cancer cases in Africa, Europe, North America, and Latin America and the Caribbean, it only provides 86.5 and 87.5% protection in Australia and Asia, respectively (this is reviewed in [29]). The vaccine does not protect against HPV35, which is associated with 3.4% of cervical cancer cases in Africa, 1.6% of those in Asia, 2.3% of those in Latin America and the Caribbean, 1.8% of those in Australia, and 1.4% of those in Europe. HPV56 is another prevalent HPV type against which the vaccine does not protect, and it is associated with 1.4% of cervical cancer cases in both Europe and Africa and 1.0% of those in North America (this is reviewed in [29]). The development of a 14-valent candidate vaccine (including HPV35, 39, 51, 56, and 59, which are not covered by Gadarsail-9) will protect against the aforementioned HPV types worldwide.
To be effective globally, an HEV-based VLP vaccine such as Hecolin must provide protection against all four HEV genotypes (HEV1-4) prevalent in humans in different geographical regions around the world. As mentioned above, the Hecolin vaccine is derived from HEV genotype 1 and it cross-protects against genotypes 2 and 4. The vaccine is also expected to offer some degree of cross-protection to genotype 3 [41,42,136]. However, since genotype 3 is prevalent worldwide [137], while genotypes 1 and 2 are prevalent mainly in developing countries in Asia and Africa, where the vaccine is currently being tested (i.e., in Bangladesh, but not in Africa), it is crucial to conduct protective studies that include genotype 3 in other regions of the world to ensure the vaccine's effectiveness on a global scale; genotype 4 is also prevalent in China. It is worth mentioning that the Hecolin vaccine is developed using the bacterial expression system. The use of the bacterial expression system is cheaper than the use of eukaryotic expression systems. However, bacterial expression systems, unlike eukaryotic expression systems, lack post-translational modifications such as glycosylation. Although the lack of glycosylation has been associated with the poor efficacy of some vaccines [138], this does not seem to be the case with the Hecolin vaccine; as a matter of fact, monkeys vaccinated with the vaccine were completely protected from infection with HEV genotype 1 [41].
As mentioned earlier, VLPs (especially those from HBV and HPV) have also been used as display platforms to enhance the immunogenicity of peptides derived from other infectious agents (e.g., bacteria and other viruses) and non-infectious agents (e.g., cancer) [46,49,[80][81][82]139], with the ultimate goal of protecting against these agents. However, using VLPs from human viruses as display platforms for foreign peptides presents a challenge due to the presence of pre-existing antibodies in the human population, either from natural infection or vaccination, against the platforms (this is reviewed in [140]). In fact, some studies have shown that high levels of pre-existing antibodies to some platforms can reduce the immunogenicity of the platforms as well as the efficacy of heterologous antigens displayed by the platforms [141][142][143][144][145][146]. For example, pre-existing maternal antibodies to poliovirus in children have been shown to reduce vaccine efficacy by up to 28% following the immunization of the children with the same antigen [144]. To overcome this limitation, VLPs from animal viruses (which do not colonize humans), plant viruses, and bacterial viruses can be used to display foreign peptides.
Overall, VLPs are versatile and efficient platforms for vaccine development. They can be used to develop vaccines which act not only against the viruses from which they are derived, but also against other infectious and non-infectious agents. Furthermore, they are excellent platforms for the development of hybrid/mosaic vaccines against viruses with segmented genomes, especially those that are transmitted by insects (e.g., AHSV). Their inability to replicate in insects reduces the risk of reassortment events that can occur with live-attenuated vaccines developed from viruses with segmented genomes. Additionally, a VLP-based vaccine for viruses such as AHSV may also help in differentiating vaccinated horses from infected horses, which may be challenging with a live-attenuated vaccine.
Author Contributions: Conceptualization, M.K. and E.T.; writing-original draft preparation, M.K., H.L. and E.T.; writing-review and editing, M.K., H.L. and E.T. All authors have read and agreed to the published version of the manuscript.
Funding: This work was supported in part by start-up funding from the School of Veterinary Medicine at Texas Tech University and by grant number 7R15AI146982-02 from the US National Institutes of Health (National Institute of Allergy and Infectious Diseases). The content is solely the responsibility of the authors and does not necessarily represent the views of the National Institutes of Health.

Conflicts of Interest:
Ebenezer Tumban is a co-inventor of an L2-bacteriophage VLP-related patent, managed by the University of New Mexico in accordance with its conflict of interest policies.