Prospects for developing an Hepatitis C virus E1E2‐based nanoparticle vaccine

Globally, more than 58 million people are chronically infected with Hepatitis C virus (HCV) with 1.5 million new infections occurring each year. An effective vaccine for HCV is therefore a major unmet medical and public health need. Since HCV rapidly accumulates mutations, vaccines must elicit the production of broadly neutralising antibodies (bnAbs) in a reproducible fashion. Decades of research have generated a number of HCV vaccine candidates. Based on the available data and research through clinical development, a vaccine antigen based on the E1E2 glycoprotein complex appears to be the best choice, but robust induction of humoral and cellular responses leading to virus neutralisation has not yet been achieved. One issue that has arisen in developing an HCV vaccine (and many other vaccines as well) is the platform used for antigen delivery. The majority of viral vaccine trials have employed subunit vaccines. However, subunit vaccines often have limited immunogenicity, as seen for HCV, and thus multiple formats must be examined in order to elicit a robust anti‐HCV immune response. Nanoparticle vaccines are gaining prominence in the field due to their ability to facilitate a controlled multivalent presentation and trafficking to lymph nodes, where they can interact with both arms of the immune system. This review discusses the potential for development of a nanoparticle‐based HCV E1E2 vaccine, with an emphasis on the potential benefits of such an approach along with the major challenges facing the incorporation of E1E2 into nanoparticulate delivery systems and how those challenges can be addressed.


| INTRODUCTION
Hepatitis C virus (HCV) is an RNA virus with highly error-prone replication resulting in high mutation rates, making vaccine development for this pathogen challenging. HCV is a major cause of severe liver diseases and cancer, and the global burden is over 58 million chronically infected individuals with an annual increase of 1.5 million new infections. 1 Moreover, HCV is responsible for more deaths in the USA than all other infectious diseases combined. 2 Nearly 75% of infected individuals progress to chronic HCV infection, which often leads to cirrhosis or hepatocellular carcinoma, a frequently fatal form of liver cancer. The World Health Organization established a goal to reduce new HCV infections by 90% and associated deaths by 65% by 2030. 1 The availability of direct-acting antivirals (DAAs) will play a large part in achieving this goal as they cure existing HCV infections. 3,4 However, most HCV-infected individuals are unaware of their infection status because they are asymptomatic until liver damage is extensive enough to present as liver disease. Moreover, DAAs remain too costly for many in the U.S., particularly for high risk groups such as intravenous drug users 5 and even more so in developing countries with high disease burdens. In addition, DAAtreated individuals can become re-infected which decreases effectiveness in high-risk groups. Therefore, preventive measures such as vaccines are a key priority, 6 and there is a clear and urgent need for a prophylactic HCV vaccine. 7 The high rates of spontaneous clearance and presence of immune memory among individuals who clear their first HCV infection 8-11 point to the feasibility of such a vaccine. 7,[12][13][14] There is broad agreement that both B and T cell immunity contribute to the control of acute HCV infection. [15][16][17] Thus, for a successful vaccine, the immunogen will need to robustly elicit both responses. Consistent with this idea, broadly neutralizing antibodies (bnAbs) infused into humanized mice or chimpanzees protect against HCV infection. [18][19][20] Moreover, viral clearance is associated with a robust induction of bnAbs early in the infection. [21][22][23][24] A robust and timely induction of bnAbs is therefore a requirement for a protective HCV vaccine. A major challenge in developing a successful HCV vaccine is its remarkable genetic diversity, which has six major genotypes (gts 1-6), two less common genotypes 25 (gts 7-8), and intragenotypic diversity resulting in 90 subtypes. 26 The virus actively escapes the B cell and T cell responses in infected individuals. 27 In particular, escape of the B cell response was observed in a clinical trial of a monoclonal antibody (mAb) therapeutic where escape variants accompanied viral rebound. 28 HCV can also escape the immune response via glycan shielding, direct cell-cell transmission, and downregulation of major histocompatibility complex proteins. 29 An ideal vaccine should elicit high antibody titers directed against multiple conserved E1E2 epitopes to ensure broad neutralization, 19,[30][31][32][33] in conjunction with cytotoxic and tissue-resident memory T cells to achieve immunity and protection against a high diversity of HCV isolates.

| Virus neutralisation and the E1E2 glycoprotein complex
In developing a prophylactic vaccine for HCV, the choice of antigen has not always been clear due to production issues associated with the E1E2 glycoprotein complex, in part due to its membrane association. The HCV envelope glycoprotein complex, as the target of the protective antibody response, has been the primary focus of most vaccine development to date. The envelope complex of HCV contains two glycoproteins, E1 and E2, which are encoded as part of the HCV polyprotein expressed in infected liver cells. This polyprotein is processed in the ER by signal peptidases and cellular glycosylation machinery to produce the mature E1E2 complex. These glycoproteins are membrane-anchored via their C-terminal transmembrane domains (TMDs), resulting in a membrane-bound E1E2 (mbE1E2) comprise the majority of the CD81-binding interface on the E2 ectodomain 64 and thus mAbs that bind to these domains can block receptor binding as part of their role in virus neutralisation. In addition, studies with a panel of E1-specific mAbs showed that that E1-specific responses can both protect against virus challenge and broadly neutralise HCV pseudoparticles (HCVpp) of various genotypes. 65,66 However, a similar correlation between E1-specific mAbs and viral clearance has not yet been identified.

| Structure-based vaccine design -Past limitations and future prospects
Limited high resolution structural data are available for the E1 glycoprotein. The E1 ectodomain contains 158 residues (residues 192-349 of the H77 sequence; NCBI Refseq ID NC_004102), of which eight are cysteine residues that form four putative disulfide bonds, and five potential N-glycosylation sites. The reported X-ray structure of an N-terminal fragment of E1 (residues 192-270) exhibits domain swapped and disulfide cross-linked oligomers, 67 that are no doubt the result of the C-terminal residue truncation and lack of E2 context. As noted in a recent review, 68 small helical fragments of E1, one a crystal structure comprising the IGH526 binding motif in complex with the IGH526 Fab, 69 and the other and NMR structure of the transmembrane helix 70 have also been reported.
In contrast to E1 and E1E2, the E2 ectodomain has been subjected to a number of structural studies. The initially reported structures of the core E2 ectodomain structure 71,72 revealed a folded domain with significant unstructured regions, with the overall fold stabilised by numerous disulfide bonds between conserved cysteines. The two core structures unexpectedly exhibited different disulfide bonding patterns. This is likely due to different truncations of the E2 sequences used in the two structures, 73  highlighted the potential role of hypervariable region 1 (HVR-1), which resides at the N-terminus of E2 as a region on E1E2 that can control such activity. 82 The authors proposed that HVR-1 acts as an entropic "safety catch" that modulates E1E2 receptor binding activity and thereby antibody sensitivity. Removal of HVR-1 induces an E1E2 "hyperactive" state that is both highly susceptible to antibody neutralisation and prone to spontaneous inactivation. 82 Other viral envelope glycoproteins, including HIV Env 83   was determined by cryoelectron microscopy (cryo-EM), 85 revealing details about the interface between E1 and E2 and the AR4A interface with the complex that will greatly aid in rational design of a vaccine against HCV. In addition, the recent development of a soluble, secreted E1E2 complex (sE1E2) 36,45 and determination of its structure 86 further enables the potential assembly of E1E2-based NPs as vaccine candidates. In particular, the sE1E2 complex adopts a native structure, with an RMSD for Cα atoms between the two structures of 1.34 Å. This structure demonstrates that a native E1E2 complex that is untethered from the membrane can be made with the appropriate scaffold, which opens up the possibility of developing native E1E2 NP vaccine candidates. Moreover, these structures afford opportunities to stabilise the E1E2 interface and modify the E1 and E2 surfaces to make the complex more amenable to NP incorporation. This review will focus on the important considerations relevant to the potential and feasibility of making an E1E2-based NP vaccine.

| ANTIGEN SIZE AND THE IMMUNE RESPONSE
Subunit vaccines, which comprise a limited portion of the infectious virus known to be immunologically important, are typically safer to use than the live attenuated viruses used for most of the history of F I G U R E 1 Conserved antigenic domains on the E1E2 complex. The E1E2 complex (RCSB ID 8FSJ) is represented as a solventexcluded surface. The antigenic domains in the E1 ectodomain are the N-terminal residues 192-202 (dark red) and a conserved αhelix (α2, residues 314-327). This helix appears to have a role in viral entry as mutants in this region result in structurally intact but noninfectious virions. 59   vaccination. This enhanced safety stems in part from the ability to produce highly purified preparations of the subunit vaccine. However, the ease of production and high level of purity comes at the expense of immunogenicity. One factor is that intact live or attenuated viruses often contain immunopotentiating components that facilitate a robust immune response. Another factor is size, which affects antigen uptake and clearance/degradation and also de- where the antibody titre achieved after immunisation of mice with inactivated HCV was five orders of magnitude larger than that for the E2 ectodomain alone. However, it is unclear based on that study and subsequent studies 91,92 whether the enhanced antibody titre translated to enhanced neutralisation potency or protective efficacy. It is possible that an effective HCV vaccine will require precise immunopotentiation of the principal neutralising determinants in order to facilitate a protective response. The varied and variable host cell components contained in an inactivated HCV vaccine 93 might amplify the overall immune response at the expense of HCV-specific targeted Toll-like receptor activation and a balanced T helper 1/2 (Th1/Th2) response. Moreover, an issue that plagues inactivated HCV as a vaccine candidate is the inability to produce the virus in the amounts needed for larger scale studies. However, the recent development of high titre cell culture virus for use as a vaccine candidate 91,92 has made proof-of-concept studies using inactivated HCV more feasible.

| Dispersions, emulsions and supramolecular assemblies
In order to recapture some of the immunogenic potency of the virus vaccine system, a variety of particulate and emulsion-based formulations have been explored as detailed in recent reviews. [94][95][96] One of the most common methods of making a particulate subunit vaccine is to formulate the subunit antigen with a heterogeneous Alum adjuvant, however it typically forms aggregated microparticles ranging from 0.5 to 10 μm in size. 97 Additional platforms for particulate and emulsion formulations, which have been explored with HCV have been recently reviewed 98 ( Figure 2). These include polymer-based nanoparticulates, such as those based on bioerodible poly(D,L-lactic-co-glycolic acid) or PLG 99,100 and micelle-forming pluronics, 101 emulsions, such as MF59 102 and Montanide, 103 and nanovesicles, which include liposomes 104 -a basis for the advanced AS01 105 and CAF01 106 systems.
In addition, for some membrane-associated glycoprotein vaccine candidates, the formulation process creates rosettes of antigen protruding from a hydrophobic micelle core, thereby creating nanoparticulate vaccines without the use of additional protein scaffolds.

F I G U R E 2
Overview of polymer, emulsion and vesicle-based delivery systems (Reprinted with permission from Andrianov AK, Fuerst TR. Immunopotentiating and Delivery Systems for Hepatitis C virus (HCV) Vaccines. Viruses. 2021; 13 (6)).

| Virus-like particles
Another platform that is particulate in nature is virus-like particles (VLPs). VLPs are often highly immunogenic and exhibit a marked increase in immunogenicity relative to a subunit vaccine containing a single component. 120  However, VLPs can be difficult to produce, and making a VLP for HCV has been particularly challenging. In particular, the baculovirus Sf9 cell-based HCV VLP production system exhibited less efficient polyprotein cleavage than observed in mammalian cells and as a result two different forms of E2 are present in those VLPs. 121 There is also the challenge of producing uniform HCV VLPs in high yields.
The Sf9-based HCV VLPs have been used in non-human primate studies, but immunisation with the VLPs failed to produce a robust antibody response in chimpanzees. 122 However, a T cell response was observed and the chimpanzees were protected from a homologous virus challenge. Mammalian systems have also been developed for HCV VLPs. 123,124 The VLPs produced in mammalian cells are morphologically similar to native virions and also incorporate apolipoproteins C and E, which are present in native virions. [125][126][127][128] The mammalian VLPs have not yet progressed to non-human primate studies, 123,129 but a recent study in pigs 129 suggests that this system can be produced at a scale amenable to assembling pre-clinical data required for eventual human trials. A third type of particulate assembly that has been developed more recently for use in vaccines is protein-based NPs. NP platforms for the multivalent display of vaccine antigens have proven effective in boosting immunogenicity for antigens such as Ebola GP, 130 HIV env, 131-133 SARS CoV-2 spike, 134 RSV F, 135 and influenza haemagglutinin. 136

| Common protein-based NP systems
The most common NP platforms currently in use contain either 24 potential attachment sites as in the case of ferritin or 60 attachment sites in a variety of other potential scaffolds. Ferritin is a naturally occurring iron storage protein and for NP delivery systems, the Helicobacter pylori ferritin is typically used, 137 although other bacterial ferritins have been used as well. 138 139 or Aquifex aeolicus lumazine synthase (LS). 140 There are also single component variants of KDPG aldolase from Thermotoga maritima called I3 141 and mi3 142 which contain 60 potential attachment sites. There is an additional system that allows 60 potential attachment sites in a two-component platform, I53 (Icosahedral assembly from 5mers and 3mers). This platform takes a pentameric Lumazine synthase RibH2 from Mesorhizobium loti and a trimeric wild-type KDPG aldolase from Thermotoga maritima modified such that the two proteins interact to form an icosahedral shell 143 with the antigen of interest displayed on the KDPG aldolase trimer components. The above platforms all have been used as "cis" genetic fusions with their target antigens. Another trend in NP delivery has been to generate NPs than can be produced separately from their associated antigens and then coupled after purification in what is called a "plug and display" modular assembly. 144 This system, which uses components called SpyTag and SpyCatcher, takes advantage of the propensity of some bacterial cell surface proteins to autocatalyse isopeptide bond formation 145,146 to create a covalent tethering system incorporated on the surface of NPs. The original platform used the AP205 phage coat protein, but it has since been expanded to other formats, including ferritin, 138,147 mi3, 147 and I53-50. 147 The plug-and-display format is particularly useful for antigens that are difficult to produce in NP form as genetic fusions and/or for producing mosaic NPs in which a number of variants of a particular antigen are coupled simultaneously to the same NP in order to increase the breadth of neutralisation of the vaccine. 136,148

| NPs, antigen retention, and the immune response
Of the differently-sized particulate platforms used for vaccination, NPs appear to have an advantage over larger particles such as microparticles at priming cytotoxic T cells. 149,150 Over the past ten to

| PROOF-OF-PRINCIPLE NP STUDIES FOR HCV
The first HCV NP study 153 was conducted using E2 residues 384-661 from genotype 1b (Con1) covalently linked to H. pylori ferritin.
The NP was expressed in Drosophila S2 cells and purified by sequential sucrose gradient centrifugation. The purified ferritin-E2 NP was antigenically intact as evidenced by binding to the antibodies AR3A and AP33 and the cellular receptor CD81. Ferritin-E2 exhibited enhanced binding to the antibodies and CD81 most likely due to an avidity effect. The NP also exhibited enhanced binding to patient sera relative to sE2 alone. However, relative to sE2 alone, E2ferritin did not produce higher serum endpoint titres at any of the doses used in the study. Despite a lack of enhanced antibody response, at the medium dosage used E2-ferritin exhibited enhanced neutralisation of cell culture virus (HCVcc) strains Con1(GT1b) and JFH1(GT2a). Moreover, when tested against a panel of HCVcc, sera from E2-ferritin inoculated mice exhibited greater neutralisation breadth, with 10 out of 13 strains neutralised with a dilution at which 50% inhibition was observed (ID 50 ) > 100, relative to sE2 alone, which only neutralised 6 out of 13 strains with an ID 50 > 100. One observation from this study that is both cautionary and applicable to all nanoparticle-based vaccine studies is that immunisation of mice elicited a noticeable dose-dependent response to the ferritin nanoparticle. These kinds of responses can divert the immune system and negatively impact efficacy. Moreover, nanoparticle scaffolds should have as little sequence homology to human proteins as possible to avoid immune responses that break tolerance to self-antigens.
The second NP study 154 tested multiple NP platforms and an optimised version of sE2 from two different genotypes (GT1a and GT6a). The sE2 versions used incorporated optimizations that truncated HVR1 (residues 384-411) and both the variable region 2 (450-494) and β-sandwich loops (540-550) in an effort to focus the immune response on only the most highly-conserved antigenic epitopes.
The authors called the designs E2mc3. Moreover, these constructs are truncated at residue 645, so the E2mc3-v1 core used in these studies excludes a number of residues at the N-and C-termini and the two loops that are present in the Li et al. study. The authors initiated the study using three NP platforms: ferritin, E2p, and I3-01.
When the E2 is modelled as displayed on the surface of the ferritin 24mer, the overall particle diameter is estimated to be 24.5 nm, and for the 60mer NPs E2p and I3-01 the estimated diameters are 34.5 and 37.5 nm, respectively. As with the previous study, the NPs exhibited enhanced binding to a panel of E2-specific mAbs, with the 60mer E2p NPs exhibiting a larger increase in general than the 24mer ferritin. This aligns with what one would expect for fully occupied 60mers versus 24mers. The authors were unable to produce NPs for the I3-01 platform using the GT6a E2mc3-v1 antigen, so I3-01 was not used for antigenicity or immunisation experiments.
The authors immunised mice with E2mc3-v1 and the corresponding ferritin and E2p NPs using E2mc3-v1 from GT1a and GT6a.
At early time points, there was a clear trend towards higher antibody titres in the NP groups relative to E2 alone. However, the titres converged and in some cases titres were higher for E2 alone groups at the last time point (week 11). Due to the fact that the NP comprises up to 49% (ferritin) to 58% (E2p) of the mass of the particle, the NP groups got less antigen as the dosage was based on the total protein amount, not the E2 amount. When looking at the ability of mouse sera to neutralise HCV, the E2mc3-v1-E2p NP-inoculated mouse sera exhibited better homologous and heterologous neutralisation activity than sera from mice inoculated with E2mc3-v1 itself.
This neutralisation trend did not correlate with antibody titre, which converges at later time points, indicating a shift in the quality of the immune response in favour of the E2p-E2mc3-v1 group. However, the ferritin NP group did not show better homologous or heterologous neutralisation than E2mc3-v1 alone, indicating that the ferritin platform is not a productive system for this antigen. The authors further examined the antibody response to the E2mc3-v1 and corresponding E2p NP groups. Compared to E2mc3-v1 alone, the E2p-E2mc3-v1 group showed a greater frequency of E2-specific B cells and used significantly more Vh groups. In addition, the E2mc3-v1 alone group exhibited highly skewed germline gene usage relative to the E2p-E2mc3-v1 group based on HCDR3 length distribution. In general, the E2p-E2mc3-v1 group engaged more naïve B cells, and activated diverse germline genes, likely accounting for its enhanced neutralisation potency.
The third NP study 155 utilised a permuted version of E1E2 in which the order of the antigens in the open reading frame was reversed (i.e. E2E1) and the I53-50 platform for NP display. This permutation allowed fusing E1 to the NP scaffold I53-50A subunit, which is trimeric. This arrangement mimicked the putative trimeric E1 configuration on the virion, for which there exists biochemical 156 but not structural evidence thus far. The authors used the GT1a resistant strain AMS0232 for the E2 and E1 ectodomains, and in the NP construct fused E2 residues 384-698 to the E1 ectodomain (residues 192-325) separated by a Gly-Ser rich linker and a hexaarginine furin cleavage site. As expected, the construct exhibited binding to E2-and E1-specific mAbs but not E1E2-specific mAbs.
Interestingly, the E2E1 NP exhibited better binding to E2-specific neutralising antibodies, but sE2 alone exhibited better binding to 6 of 14 -TOTH ET AL. AT1211, which targets domain C and non-neutralising antibodies that target domain A. This behaviour is consistent with that observed previously for both mbE1E2 and sE1E2 36 and with the E1E2 structures that show antigenic domain A in the E1E2 complex abutting the base and stem regions of E2, which are present in both E1E2 and E2E1 complexes but absent in sE2. The authors immunised rabbits with sE2, E2E1-I53-50A trimers, and E2E1-I53-50 NPs. To avoid skewing endpoint titres with antibodies to the NP scaffold, the authors produced an E2E1-foldon trimer for analysis. Serum antigen binding titres for the E2E1-I53-50 NP group were higher to E2E1foldon immobilised on the plate than sE2, but not significantly higher than the sera from rabbits immunised with E2E1-I53-50A trimers. When tested against the homologous pseudovirus, sera from all six rabbits immunised with E2E1-I53-50 NPs neutralised the pseudovirus with a median ID 50 of 457, whereas only one sample in the E2-immunised group was able to neutralise the homologous HCVpp, and sera from five of six rabbits immunised with E2E1-I53050A trimers were able to neutralise the homologous HCVpp (median ID 50~1 00). Heterologous neutralisation was observed, but it was weak for the three single-species groups. Therefore, the authors leveraged the flexibility of the I53-50 NP system to produce mosaic NPs and a NP cocktail with E2E1s from multiple strains.
Incorporating E2E1s from multiple genotypes increased the neutralisation breadth relative to the group immunised with a single strain of E2E1.

| FUTURE PROSPECTS FOR A NP-BASED HCV VACCINE
From the aggregated data for other viruses and the three HCV NP immunisation studies discussed above, it is likely that a NP-based vaccine containing E1E2 will facilitate an enhanced immune response with consistent and controllable production and solution properties. It will also preserve not only the E2 regions that are strongly associated with viral clearance but also the AR4A epitope that is associated with viral clearance, 32 thereby making a more potent immunogen. The recently-determined cryo-EM structures of mbE1E2 85 and a soluble secreted form of E1E2 86 will allow modifications of the E1E2 complex that both enhance its immunogenicity and enable its efficient incorporation into NP platforms. However, the above studies also contain some data that serve as cautionary tales. In the He et al. study, the I3-01 system was eliminated from consideration because they were unable to produce a NP form for GT6a E2mc3-v1. This observation underscores that, while these systems are designed as modular and adaptable, a fair amount of optimization might be required for some antigens, particularly those containing multiple post-translational modifications or multiple disulfide bonds like E2 and E1E2. E2 is generally well-behaved in solution and can be isolated from a variety of eukaryotic expression systems in monomeric form 40 and yet from the data in this study it is clear that not all NP systems are compatible with this antigen. E1E2, in both membrane-bound and secreted native-like forms, is not as well behaved in solution as E2 36,157 and, thus, the bar for getting native E1E2 NPs will undoubtedly be much higher. Perhaps the SpyTag-SpyCatcher plug-and-display systems 144 will mitigate this issue (Figure 3), but such a study has not yet been completed. Those systems are flexible, but for antigens that are not well-behaved, there is no guarantee of high occupancy on the NP after coupling and postpurification assembly, and as such the preparations could include contamination with empty NPs that dilute the immune response to the antigen.
Moreover, re-purification can be disadvantageous as significant recovery losses can occur. Thus, the presence of unreacted SpyTag-SpyCatcher nanoparticles can potentially complicate manufacturing as the unreacted components might be difficult to remove. In addition, the plug-and-display NPs and two component NPs are often expressed either entirely or in part in bacteria. From a vaccine manufacturing standpoint, it would be advantageous to express the components together in eukaryotic cells. However, efficient secretion to the media can be an issue for NPs. 158 A recent study found that this issue can stem from cryptic transmembrane domains 158 which can then be rectified in some cases via computational "degreasing" which alleviates the propensity of some hydrophobic interfaces in designed NPs to impede secretion. Analysing the secretion efficiency for a given NP system is therefore a critical step that needs to be taken early in the process to avoid potential roadblocks after the preclinical stage. An additional cautionary piece of data from this study is the inferiority of neutralisation for E2mc3-v1-ferritin relative to E2mc3-v1 alone. These data are inconsistent with the data from Li et al. and perhaps result from the modifications to E2 in the context of the ferritin NP system. That is, not all modifications that optimise the antigen on its own will necessarily be directly transferrable to a given NP platform. This complicates the issue considerably, but considering the aggregate data in the HCV vaccine development field, it is not surprizing. E1E2, and even E2, are glycoproteins with unusual structures that are likely highly dynamic and thus any changes to the protein can have unexpected effects on antigenic epitope presentation. Tethering the glycoproteins to a NP adds another variable to the interplay between structure and dynamics that must be tested on a case-by-case basis.
The consistent positive theme from the three HCV NP studies is that successful NPs outperform subunit vaccines, making an E1E2-based NP in any one of the major kinds of platforms

F I G U R E 3
The SpyCatcher-SpyTag system as a potential platform for making an E1E2-based nanoparticle. The sE1E2 construct with the SpyTag at the C-terminus of one of the subunits (SpyTag attachment to the E2 subunit is shown) and the Spycatcher nanoparticle can be purified separately. The two components can then be coupled by leveraging the spontaneous formation of an isopeptide bond upon the interaction of SpyCatcher and SpyTag, creating an E1E2 nanoparticle assembly.
F I G U R E 4 E1E2-based nanoparticles modelled for three common platforms. A model of sE1E2 derived from a recent cryo-EM structure was superposed onto the attachment points of each monomer for the associated model platforms. Both the full particle and cross-section are shown to provide a view of the approximate size and antigen density on each type of nanoparticle. Another potential avenue of exploration with E1E2 NPs is the potential to create mosaic and cocktail mixtures as a means to enhance neutralisation breadth as in the E2E1 NP study. 155 It is not clear, nor has it been systematically investigated, whether the best choice for an HCV vaccine will be a single patient-derived strain derived from an elite neutraliser or a mixture of representative strains from multiple genotypes, perhaps varying based on the geographical distribution of genotypes and population to be vaccinated. The successful development of mosaic NP vaccine candidates against SARS-CoV-2 148 suggests that once a NP-based E1E2 vaccine system is established, the mosaic versus single-strain immunogen design can be more thoroughly investigated.

CONFLICT OF INTEREST STATEMENT
The authors have no conflicts of interest to declare.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.

ETHICS STATEMENT
Ethics approval was not needed. Figure 2 is reprinted with permission from Andrianov AK, Fuerst TR.

PERMISSION TO REPRODUCE MATERIAL FROM OTHER SOURCES
Immunopotentiating and Delivery Systems for HCV Vaccines. Viruses. 2021; 13 (6).