Self-assembling SARS-CoV-2 spike-HBsAg nanoparticles elicit potent and durable neutralizing antibody responses via genetic delivery

While several COVID-19 vaccines have been in use, more effective and durable vaccines are needed to combat the ongoing COVID-19 pandemic. Here, we report highly immunogenic self-assembling SARS-CoV-2 spike-HBsAg nanoparticles displaying a six-proline-stabilized WA1 (wild type, WT) spike S6P on a HBsAg core. These S6P-HBsAgs bound diverse domain-specific SARS-CoV-2 monoclonal antibodies. In mice with and without a HBV pre-vaccination, DNA immunization with S6P-HBsAgs elicited significantly more potent and durable neutralizing antibody (nAb) responses against diverse SARS-CoV-2 strains than that of soluble S2P or S6P, or full-length S2P with its coding sequence matching mRNA-1273. The nAb responses elicited by S6P-HBsAgs persisted substantially longer than by soluble S2P or S6P and appeared to be enhanced by HBsAg pre-exposure. These data show that genetic delivery of SARS-CoV-2 S6P-HBsAg nanoparticles can elicit greater and more durable nAb responses than non-nanoparticle forms of stabilized spike. Our findings highlight the potential of S6P-HBsAgs as next generation genetic vaccine candidates against SARS-CoV-2.


INTRODUCTION
Since the first known case of Coronavirus disease 2019 (COVID- 19)  identified in December of 2019, COVID-19 infection has been detected in over 500 million people and caused 6 million deaths worldwide (https://coronavirus.jhu.edu/map.html).COVID-19 is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), a beta-coronavirus closely related to SARS 1 .The SARS-CoV-2 spike is a trimeric glycoprotein that coats the surface of SARS-CoV-2 viral particles 2 and consists of an N-terminal signal peptide, S1 and S2 subunits 3,4 .The S1 subunit contains an N-terminal domain (NTD) and a receptor-binding domain (RBD).The RBD domain is responsible for the binding to human angiotensin-converting enzyme 2 (ACE2), the major receptor in humans for SARS-CoV-2.The S1 subunit also contains SD1 and SD2 domains, which are located between the RBD and the furin cleavage site.The S2 subunit is composed of the fusion peptide (FP), heptapeptide repeat sequences 1 and 2 (HR1 and HR2), transmembrane domain (TM), and cytoplasmic domain 4 .The S2 subunit is involved in membrane fusion.Owing to its role in receptor binding, viral attachment, and entry into host cells, SARS-CoV-2 spike is the target of many authorized or licensed COVID-19 vaccines 5 .
Among the leading COVID-19 vaccines, the mRNA vaccines from both Moderna and Pfizer-BioNTech exhibit high protective efficacy 11 .However, the protection offered by these vaccines has been waning over time 12,13 and has declined considerably against the Delta and Omicron sublineages, as compared to the ancestral Wuhan or WA1 strain 14,15 .A third dose of the mRNA vaccines is needed to achieve high level of protection against Omicron variants and a fourth dose of the mRNA vaccines is sometimes recommended to restore the antibody levels and improve clinical protection 16 .Recently, the bivalent mRNA vaccines encoding both WA1 and BA.4/5 S2Ps from Moderna and Pfizer-BioNTech have been authorized as a booster shot (https://www.fda.gov/newsevents/press-announcements/coronavirus-covid-19-update-fdaauthorizes-moderna-pfizer-biontech-bivalent-covid-19-vaccinesuse).While mRNA vaccines have been proven to be effective, there remains room to improve the potency and durability of the elicited immune response.
Nanoparticle vaccine platforms allow multivalent antigen presentation and can elicit more potent immune responses than protein immunogens 17 .Several nanoparticle platforms using lumazine synthase, ferritin or I53-50 displaying SARS-CoV-2 spike or its RBD domain elicit potent nAb responses against SARS-CoV-2 [18][19][20][21] .HBsAg can self-assemble into a 22 nm nanoparticle and has been used successfully in multiple vaccines including Hepaccine B and Recombivax HB 22 .Recently, a HBsAg-based malaria vaccine, R21/MM, has shown 77% efficacy in a phase IIb trial and advanced into phase III 23 .HBsAg has also been used in recent vaccine designs displaying SARS-CoV-2 RBD [24][25][26] .In addition, the hepatitis B vaccine can elicit over 30 years of immunological memory against HBsAg in healthy people after the first vaccination at the age of over 6 months, without the need of a booster shot 27,28 .In this study, we designed and characterized self-assembling HBsAgbased nanoparticles displaying the SARS-CoV-2 spike and compared the immunogenicity of these spike-HBsAg nanoparticles and non-nanoparticle form of stabilized spikes in mice via genetic delivery.

DNA encoding SARS-CoV
Ten, 2, and 0.4 µg of soluble S2P and S6P elicited similar levels of neutralization potency against each individual pseudovirus at each dose (Fig. 3 and Supplementary Table 5).At all three doses, S6P-12-HBsAg or S6P-16-HBsAg elicited substantially higher binding and antigenicity of S6P-HBsAgs.Plates were coated with 1 µg/ml of SARS-CoV-2 S2P, S6P or S6P-HBsAg nanoparticles.Human ACE2 was assessed at a concentration ranging from 6.1 ng/ml to 100 µg/ml.The mAbs specific to the NTD, SD1, S2 or RBD domain of SARS-CoV-2 spike were tested from 1 pg/ml to 4 µg/ml.All plots were color-coded according to the legends on the right.

DNA encoding SARS-CoV-2 S6P-HBsAgs elicits durable nAb responses
As S6P-HBsAgs elicited potent nAb responses at week 6 against SARS-CoV-2 WA1 and diverse variant pseudoviruses, we assessed the durability of the neutralization potency elicited by S6P-HBsAg.At week 14, the neutralization ID50s and ID80s elicited by S6P-12-HBsAg or S6P-16-HBsAg were maintained at significantly higher levels than those elicited by the same or higher dose of soluble S2P and S6P (Supplementary Fig. 8).The ID50s and ID80s elicited by these S6P-HBsAgs at week 14 are comparable to those at week 6.Notably, the ID50s elicited by 0.4 µg of S6P-12-HBsAg or S6P-16-HBsAg increased from week 6 to week 14, though without statistical significance.
We further monitored binding antibodies in the animal sera to WA1 S2P and HBsAg from week 0 to week 45.Potent anti-WA1 S2P antibodies were detected from week 6 to week 45, for mice immunized twice with S2P(1-1206), S6P(1-1206), S6P-12-HBsAg and S6P-16-HBsAg at 10, 2, and 0.4 µg doses (Supplementary Fig. 9).Neither soluble spike groups nor nanoparticle groups showed significant change in the levels of S2P-binding antibodies over this duration.The anti-HBsAg binding antibodies were only detected from week 6 to week 45 in 10 and 2 µg of S6P-HBsAg groups, with significantly lower titers than the anti-S2P antibodies (Supplementary Figs. 9 and 10).
Encouraged by the durable S2P-binding antibody response elicited by S6P-HBsAg, we further tested the neutralization potency of the pooled sera from each group from week 22 to week 45.At 10 µg dose, the neutralization activity was detected up to week 22 for soluble S2P and up to week 45 for soluble S6P and S6P-HBsAg at comparable levels (Figs. 2 and 4; Supplementary Figs. 8 and 11).While the neutralization activity was only detected up to week 14 for 2 and 0.4 µg soluble S2P and S6P groups, the neutralization activity was maintained at substantially higher levels through week 45 for the same dose of S6P-HBsAg groups, which lasted about 7 months longer than the duration of soluble S2P and S6P groups.The only exception was the 0.4 µg S6P-12-HBsAg group which showed no neutralization potency at week 45.  4.

DISCUSSION
In this study, we designed and characterized self-assembling SARS-CoV-2 spike-HBsAg nanoparticles and assessed their immunogenicity in mice via genetic delivery.We show that these HBsAg nanoparticles displaying SARS-CoV-2 S6P elicit potent, broad and durable immune responses compared to non-nanoparticle form of stabilized SARS-CoV-2 spikes.Additionally, preimmunization with a HBV vaccine leads to a further enhancement of the HBsAg nanoparticle-elicited immune responses against SARS-CoV-2 and its variants.
SARS-CoV-2 spike exhibits a metastable prefusion conformation that spontaneously transitions to its post-fusion conformation 3 .The two-proline-stabilized SARS-CoV-2 spike S2P in its prefusion conformation has been used in the current COVID-19 vaccines from Moderna, Pfizer-BioNTech and Johnson & Johnson.We have screened WA1 spike, S2P and S6P for nanoparticle formation on a HBsAg core.S6P succeeded to form well-defined HBsAg nanoparticles whereas small fractions of wild type spike and S2P formed nanoparticles.This may be due to the metastability of SARS-CoV-2 spike.S2P is not as stable as S6P, as four proline substitutions in S6P in addition to the K986P and V987P substitutions further stabilize its prefusion conformation 2,8 .Both S6P-12-HBsAg and S6P-16-HBsAg bound well to human ACE2 and mAbs specific to the RBD or NTD domain of SARS-CoV-2 spike, with comparable or stronger affinity as compared with free S6P, suggesting the accessibility of the neutralizing epitopes in the S6P-HBsAg nanoparticles.The reduced binding of S6P-HBsAg shown by ELISA was likely due to less molarity of and less accessible area to the coated spike on the HBsAg core.
DNA encoding S6P-12-HBsAg and S6P-16-HBsAg elicited potent nAb responses to SARS-CoV-2 and its variants.The lower S2Pbinding antibody levels but higher nAb responses elicited by 0.4 µg S6P-HBsAg than those elicited by the same dose of soluble S2P and S6P are indicative of the higher RBD-specific neutralizing antibodies elicited by S6P-HBsAg.In addition, the nAb responses elicited by S6P-HBsAg persisted through week 45 and is more durable than those elicited by soluble S2P and S6P at low doses via genetic delivery.Though both soluble S2P and S6P as well as S6P-HBsAg elicited persistent S2P-binding antibody responses, S6P-HBsAg elicited significantly higher proportions of RBDbinding antibodies while less NTD-binding antibodies than soluble S2P and S6P, which may result in a higher fraction of nAbs and neutralizing epitope specific long-lived plasma cells, leading to more potent and durable nAb responses.The presence of HBsAg in the S6P-HBsAg nanoparticles might also contribute to this durable memory, as HBV vaccines are able to elicit about 30 years of memory against HBsAg 27,28 .The detailed mechanisms underlying the durable nAb responses elicited by S6P-HBsAg remain to be addressed.Similar to our findings, persistent binding antibody responses along with waning nAb responses over time have been observed in people post mRNA-1273 vaccination or post natural infection of SARS-CoV-2 39,40 .The persistent binding antibody responses accompanied by waning nAb responses are likely due to the presence of large fraction of long-lived plasma cells elicited by non-neutralizing epitopes of SARS-CoV-2 spike, which remain to be investigated.
The neutralization potency induced by DNA delivery of S6P-HBsAgs is comparable to that elicited by intramuscular delivery of 1 µg mRNA-1273 or SARS-CoV-2 S2P protein 7,41 , albeit more uniform responses were induced by mRNA-1273.Using the same route of genetic delivery, S6P-HBsAgs elicited substantially higher neutralization potency than S2P (1-1273), whose coding sequence matches that in mRNA-1273.Similar observations were noted in mice preimmunized with a HBV vaccine, with further significantly enhanced potency.The higher neutralization potency elicited by S6P-HBsAgs than by non-nanoparticle forms of stabilized spike alone could be contributed by several factors.First, S6P-HBsAg nanoparticles with diameters of ~70 nm are much larger than the SARS-CoV-2 spike.The large size of these nanoparticles may result in more efficient internalization of the spike by antigen presenting cells (APCs) and retention of the spike on lymph node follicles 42 .The repetitive array of the spikes on S6P-HBsAg nanoparticles may also enable efficient binding and activation of multiple B cell receptors.In addition, the antibody responses elicited by S6P-HBsAgs were mainly directed against the spikes displayed on the HBsAg nanoparticles as opposed to HBsAg, suggesting efficient presentation of the spike on the surface of these nanoparticles by the HBsAg core.Lastly, it has been shown that the anti-HBsAg seroconversion rates in humans increase for a period up to 8 months post Recombivax HB immunizations 43 , and the GMTs persist at high levels for 13 months as tested 44 , indicative of a prolonged period of immune activation after Recombivax HB vaccination.Such prolonged immune responses elicited by Recombivax HB might have contributed to the remarkably increased neutralization potency in mice preimmunized with Recombivax HB, as compared to immunized mice originally naïve to HBsAg.Two immunizations of S6P-HBsAgs elicited robust neutralization potency against B.1.1.529(Omicron BA.1) pseudovirus, in mice preimmunized with Recombivax HB.While a 3rd dose of mRNA vaccines is needed to offer better protection against Omicron variants 14,15 and a fourth dose retains low efficacy in preventing Omicron infections 16 , and the bivalent COVID-19 vaccines do not resolve the durability issue, S6P-HBsAgs may provide a strategy to offer longer and better protection against SARS-CoV-2 in populations that have previously exposed to HBsAg by infection or vaccination.As per WHO estimates, approximately 1/3 of the global population have been infected with HBV in 2010 37 .The immunization rate with 3 doses of HBV vaccine during infancy has also reached 85% worldwide in 2019 38 .This high rate of exposure to HBsAg in global population further highlights the potential of SARS-CoV-2 S6P-HBsAg as a next generation genetic vaccine platform for inducing more potent and durable protection against SARS-CoV-2.
Many nanoparticle designs require an additional protein conjugation step [18][19][20]25,26 , which does not allow them to be amenable to gene-based delivery. Our SP-HBsAg designs exhibit advantages as genetic vaccine candidates over similar nanoparticle vaccine designs utilizing an additional conjugation step 18,19,25 .The utilization of spike in our designs enables more neutralizing epitopes than RBD-based vaccine designs, as NTD and S2 are known to be able to elicit neutralizing antibodies against SARS-CoV-2 (Supplementary Table 1, references attached).The preexisting immunological memory against HBsAg in large population globally may also enable S6P-HBsAgs to leverage their potency for vaccine development against SAR-CoV-2 and other related coronaviruses.Our DNA constructs encode the HBsAgspike as a single gene product and could readily be encoded by mRNA or viral vector gene delivery vaccine platform.
One limitation of this study is that it was done with DNA, which is highly immunogenic in mice, but less immunogenic than mRNA in nonhuman primates (NHP) and humans.Therefore, further studies should evaluate mRNAs encoding these S6P-HBsAg particles in NHP and investigate the mechanism underlying the enhanced vaccine-induced immune response by a HBV prevaccination.
In conclusion, SARS-CoV-2 S6P-HBsAg can elicit more potent and durable nAb responses against diverse SARS-CoV-2 strains than non-nanoparticle form of stabilized spikes, including SARS-CoV-2 full-length S2P with its coding sequence matching that in mRNA-1273.The nAb responses elicited by S6P-HBsAg can persist 7 months longer than soluble stabilized spikes and appeared to be enhanced by pre-exposure to HBsAg.S6P-HBsAgs represent promising next generation genetic vaccine candidates against SARS-CoV-2.Overall, this platform has the potential to serve as a universal vaccine platform against coronaviruses and other infectious pathogens.

DNA construct design
The HBsAg nanoparticles displaying SARS-CoV-2 spike, S2P or S6P were designed as spike-HBsAg fusion proteins.The coding sequence of the ectodomain (aa 1-1206) of SARS-CoV-2 WA1 spike (NC_045512.2),2P-or 6P-stabilized mutant (S2P or S6P) 2,8 was joined with the coding sequence of HBsAg (aa 1-226, AET06188.1).The linker was made up of varying repeats of GS dipeptide.The coding sequences of SARS-CoV-2 spike-HBsAgs were inserted into the CMVR8400 vector via XbaI and BamHI sites.The ectodomain of SARS-CoV-2 S2P or S6P, as well as the fulllength S2P (aa 1-1273) were also constructed in the CMVR8400 vector via the same restriction sites as above, respectively.Each construct was human codon-optimized, synthesized by GenScript and confirmed by DNA sequencing.

Protein expression and purification
The SARS-CoV-2 spike-HBsAg plasmids were transfected into Expi293F cells using ExpiFectamine 293 transfection kit following the manufacturer's instructions.The transfected culture was grown at 37 °C for 5 days before harvest.The culture was then spun down at 10,000 × g at 20 °C for 30 min.The supernatant was collected and passed through a 0. ).The tags in S2P and S6P were cleaved with HRV-3C and further purified by sizing column purification in PBS.Fctagged human ACE2, His-tagged SARS-CoV-2 RBD, NTD recombinant proteins were expressed likewise and purified using rProtA Sepharose and Ni-NTA resin, respectively, before sizing column purification 7,45 .The purified proteins were aliquoted, frozen in liquid nitrogen, and kept at −80 °C before use.

Negative-stain electron microscopy
The sucrose fractions of SARS-CoV-2 spike-HBsAg fusion proteins were buffer exchanged to PBS containing 5-10% sucrose, and then applied to a freshly glow-discharged carbon-coated grid for about 15 s.The grid was washed with buffer containing 10 mM HEPES, pH 7.0, and 150 mM NaCl, followed by negative staining with 0.7% uranyl formate.Images were collected at a nominal magnification of 57,000 using EPU software on a Thermo Scientific Talos F200C electron microscope operated at 200 kV and equipped with a 4k × 4k Ceta CCD camera or at a nominal magnification of 50,000 using SerialEM 46 on an FEI T20 electron microscope operated at 200 kV and equipped with an Eagle CCD camera.The corresponding pixel sizes were 0.253 and 0.22 nm.Particles were picked using e2boxer from the EMAN2 software package 47 .Reference-free 2D classification was performed using Relion 48 .
ELISA to assess ACE2 binding and antigenicity of SARS-CoV-2 spike-HBsAgs ELISA plates (Thermo Fisher, 442404) were coated with 1 µg/ml of SARS-CoV-2 WA1 or variant S2P, WA1 S6P or S6P-HBsAg in PBS buffer, pH 7.4, 100 µl/well at 4 °C for 16 h 2 .The plates were washed with PBST thrice, 300 µl per well per time and then blocked with 5% skim milk in 1× PBST at room temperature for 1 h.The binding of SARS-CoV-2 WA1 S2P, S6P and spike-HBsAg was done with 4 µg/ml SARS-CoV-2 mAbs (Supplementary Table 1), followed by 7 data points of 4-fold serial dilutions or with 100 µg/ml human ACE2, followed by 11 data points of 4-fold serial dilutions 49 .The goat anti-human IgG Fc-HRP antibody (Invitrogen, A18817, 1/ 5000) was used to detect the ACE2 binding.The primary antibody incubation was done at room temperature for 30 min.Following three washes, the plates were incubated with HRP-conjugated anti-human or anti-mouse antibody (Thermo Fisher, A18811 and G21040, 1/2000) at room temperature for 30 min.The plates were then washed thrice and developed with 3,5,3′,5′-tetramethylbenzidine (TMB) (KPL) at room temperature for 10 min.After quenched with 1 N H 2 SO 4 (Fisher), the plates were read at 450 nm with a SpectraMax Plus 384 microplate reader.The data were plotted and analyzed with GraphPad Prism.

Bio-layer interferometry (BLI)
The BLI binding of SARS-CoV-2 stabilized spike and HBsAg nanoparticles to SARS-CoV-2 mAbs or human ACE2 was measured by an Octet HTX instrument.The binding buffer contained 1x HBS-EP+ buffer (GE) and 5% sucrose.The mAbs or human ACE2 with a Fc tag were captured by AHC or AMC sensors to yield a binding signal of 1-1.3 nm.The SARS-CoV-2 stabilized spike free or displayed on a HBsAg core varied from 0 to 800 nM.The molar concentration of the S6P-HBsAg was determined based on spike (but not based on a nanoparticle).The association and dissociation were monitored for 300 s at 30 °C, respectively.The binding curves were globally fitted with a 1:1 Langmuir binding model using Data Analysis Software v9.0.
Serum ELISA ELISA plates (Thermo Fisher, 442404) were coated with 1 µg/ml of SARS-CoV-2 WA1, variant S2P, or HBsAg (ProspecBio, HBS-875) in PBS, pH 7.4 at 4 °C for 16 h 7 .Standard washes and blocking steps were done as described above.The plates were blocked for 2 h.The sera were diluted by 100-fold in 5% skim milk in PBST.A serial 4-fold dilution for weeks 0 and 2 sera or 6-fold dilution for week 6 sera was applied to the 100-fold dilution preparations.The plates were incubated with diluted sera at room temperature for 1 h.The HRP-conjugated anti-mouse secondary antibody (Thermo Fisher, G21040, 1/2000) was used to detect the antibody responses.The endpoint titers were calculated as the dilution that yielded an optical density equivalent to 4×background (secondary antibody alone).
2 µm filter.The supernatant was further spun through 20% sucrose cushion in a buffer containing 20 mM MES, 150 mM NaCl, pH 6.0.The ultracentrifugation was performed with a Surespin rotor at 71552 × g, 4 °C for 2 h.The pellet was resuspended in a buffer containing 20 mM MES, 150 mM NaCl, pH 6.0, filtered through a 0.45 µm filter, and then spun through a sucrose gradient consisting of 1.5 ml of 20% to 65% sucrose in a buffer containing 20 mM MES, 150 mM NaCl, pH 6.0.This step of ultracentrifugation was done with a Th-641 rotor at 217,339 × g, 4 °C for 8 h.The sucrose fractions were collected and stored at 4 °C.A dialysis against PBS buffer, pH 7.4, was performed before use.