The core domain of hepatitis C virus glycoprotein E2 generates potent cross‐neutralizing antibodies in guinea pigs

A vaccine that prevents hepatitis C virus (HCV) infection is urgently needed to support an emerging global elimination program. However, vaccine development has been confounded because of HCV's high degree of antigenic variability and the preferential induction of type‐specific immune responses with limited potency against heterologous viral strains and genotypes. We showed previously that deletion of the three variable regions from the E2 receptor‐binding domain (Δ123) increases the ability of human broadly neutralizing antibodies (bNAbs) to inhibit E2‐CD81 receptor interactions, suggesting improved bNAb epitope exposure. In this study, the immunogenicity of Δ123 was examined. We show that high‐molecular‐weight forms of Δ123 elicit distinct antibody specificities with potent and broad neutralizing activity against all seven HCV genotypes. Antibody competition studies revealed that immune sera raised to high‐molecular‐weight Δ123 was poly specific, given that it inhibited the binding of human bNAbs directed to three major neutralization epitopes on E2. By contrast, the immune sera raised to monomeric Δ123 predominantly blocked the binding of a non‐neutralizing antibody to Δ123, while having reduced ability to block bNAb binding to E2, and neutralization was largely toward the homologous genotype. This increased ability of oligomeric Δ123 to generate bNAbs correlates with occlusion of the non‐neutralizing face of E2 in this glycoprotein form. Conclusion: The results from this study reveal new information on the antigenic and immunogenic potential of E2‐based immunogens and provide a pathway for the development of a simple, recombinant protein‐based prophylactic vaccine for HCV with potential for universal protection. (Hepatology 2017;65:1117‐1131).

Expression vectors for the production of HCVpp incorporating E1E2 heterodimers from G1a were pE1E2H77c (1).
ELISA assays. Synthetic peptides were synthesized by Genscript corporation (USA) or Auspep (Australia) and their sequences can be found in SI. Details for MAbs can be found in SI. E2-CD81 inhibition assays have been previously described in (13) with details available in SI. For antibody competition assays, a constant amount of MAb and a half log dilution series of each serum were simultaneously added to blocked wells and incubated for 2 hours at room temperature before addition to plate bound monomeric Δ123. Residual MAb binding was detected with anti-human Fab 2 . Curves were fitted by non-linear regression and used to determine the ID50 for each serum sample. Where a serum sample failed to achieve an ID50 at the highest concentration tested (1:10 dilution, log10 1 ) a value of log10 0.5 was assigned to that serum.
Affinity purification methods. Culture supernatant from transfected FS293F cells were affinity purified on 5ml HisTrap columns (GE Healthcare). Proteins were bound in the presence of binding buffer (20mM Sodium Phosphate, 0.5M NaCl, 10mM Imidazole pH7.4), washed (20mM Sodium Phosphate, 0.5M NaCl, 25mM Imidazole pH7.4) and then eluted (20mM Sodium Phosphate, 0.5M NaCl, 500mM Imidazole pH7.4). Immunizations. Groups of 8 age and weight matched Albino Dunkin Hartley guinea pigs were immunized with 100 μg E2 antigen in the presence of 75 ISCO TM units of ISCOMATRIX® adjuvant, subcutaneously at three weekly intervals in all cases, except animals vaccinated with fractionated WT E2 antigens that were vaccinated at two weekly intervals (data shown in Supplemental figures 3 and 12). Animals were sacrificed 2 weeks after the final immunization, blood was collected and the prepared serum stored at 4°C or at -80°C for long term storage. In the case of Figure 1, 10 animals were used for WT and Δ123 immunogens and 5 animals in the no antigen control group. In the case of Figures 2-4, 8 animals per group were used for each Δ123 antigen and 5 animals in the no antigen control group. One serum from the monomer group (animal 4.6) was excluded from all analyses due to haemolysis. For use in serum neutralization assays, serum was heat inactivated at 56°C for 60 min.
Multi angle light scattering (MALS). The SEC separation was performed on a Agilent 1200 series HPLC system with a WTC-030-N5 4.6/300 size exclusion column (Wyatt Technology corp, CA) using a flow rate of 0.2ml/min in PBS (phosphate buffered saline). Online detection included a Agilent UV detector, a DAWN HELEOS II MALS detector and a Wyatt Optilab T-rex differential refractive index (RI) detector. Molecular weight analysis of peaks was generated using MALS detection and analyzed using Astra 6 software (Wyatt Technology corp, CA).
IgG depletion of serum. To determine whether neutralization was due to IgG, protein G-Sepharose (PGS) was washed with sterile PBS 3 times. 50 μl of guinea pig serum was adsorbed with 10 ul sterile PGS for 30 minutes. Beads were then pelleted (3,000g, 1 minute) and the depleted immune serum used in H77c HCVpp neutralization assays.
E2-CD81 inhibition assays. Solid phase immunoassay plates (Maxisorb, Nunc, Roskilde) were coated with dimeric CD81-LEL protein (5 μg/ml) in PBS buffer overnight followed by BSA 10 PBS for 1 h at RT. The immune sera were titrated half-log in BSA 5 PBST and 50ng of WT E2 glycoproteins were added followed by incubation at RT for 1 h before addition to washed dimeric MBP-LEL 113-201 -coated ELISA plates. After 1 h incubation at RT and washing, bound WT E2 was detected using rabbit anti-His tag antibody (Rockland) diluted in BSA 5 PBST for 1 h at RT followed by washing and addition of HRP-labelled goat anti-rabbit (Dako, Glostrup, Denmark) for 1 h at RT. After washing, bound E2 was visualized with TMB substrate. The E2-CD81 inhibition titers were calculated as the reciprocal dilution of immune serum from an 11-point dilution series that reduces E2 binding to CD81-LEL by 50% or 80% (ID50 and ID80, respectively).
Neutralization assays. To perform NAb assays, serial dilutions of immune serum were added to HCVpp and incubated for 1 h at 37°C before addition to Huh 7.5 cells seeded 24 h earlier at 30,000 cells/well in 48 well plates. After 4 h incubation at 37°C, the inoculum was removed and replaced with DMF10 containing non-essential amino acids (NEAA) for 72 h. Cells were washed with PBS before lysis in Cell Lysis Buffer (Promega). Luciferase activity in clarified lysates was measured using luciferase substrate (Promega) and a FLUOstar Optima microplate reader fitted with luminescence optics (BMG Life Technologies, Germany). Neutralization assays using HCVcc were performed by mixing HCVcc virus with an equal amount of serially diluted immune serum. Each experiment was performed in triplicate (G2a) or duplicate (G3-7). The virus/serum mixture was incubated for 1 h at 37°C before addition to Huh7.5 cells seeded 24 h earlier at 30,000 cells/well in 48 well plates for 4 h. Cells were washed at least 4 times and replenished with fresh DMF10NEAA and incubated for a further 48-72 h. For G4a, G5a, G6a and G7a, cells were washed with PBS before lysis in Renilla Cell Lysis Buffer (Promega). Luciferase activity was measured in clarified lysates using Renilla luciferase substrate (Promega) and a FLUOstar Optima microplate reader fitted with luminescence optics (BMG Life Technologies, Germany). In the case of G2a, supernatant fluid collected from cells 48 h after infection were lysed with Renilla Cell Lysis Buffer (Promega) and measured as described for G4a-7a. The neutralization titre was calculated from 6-point dilution curves as the reciprocal dilution of serum to reduce luciferase activity by 50% (ID50) or 80% (ID80) and the data shown is the mean from at least two independent experiments.
In the case of G3a, infected cells were fixed with cold methanol 72 h after infection and infection events measured by staining with anti-NS5A MAb (9E10) and goat-anti-mouse Alexa 488 antibody (Life Technologies) and visualized by immunofluorescence using an Olympus inverted microscope IX51.

Supplemental figure 3.
Antigenic and immunogenic characterization of WT E2. A. SEC profile of WT E2 expressed in FS293F cells following NiNTA purification. B. Antibody titres of immune serum raised to unfractionated (Un/frac), monomer, dimer, HMW2 and HMW1 WT E2 against monomeric H77c E2 antigen. C. Antibody titres of immune serum raised to unfractionated (Un/frac), monomer, dimer, HMW2 and HMW1 WT E2 against G1a HVR1 peptide. D. Ability of immune serum raised to unfractionated (Un/frac), monomer, dimer, HMW2 and HMW1 WT E2 to prevent binding of H77c G1a E2 antigen to CD81. E. Ability of immune serum raised to unfractionated (Un/frac), monomer, dimer, HMW2 and HMW1 WT E2 to prevent binding of JFH1 G2a E2 antigen to CD81. F. Ability of immune serum raised to unfractionated (Un/frac), monomer, dimer, HMW2 and HMW1 WT E2 to neutralize G1a HCVpp. G. Ability of immune serum raised to unfractionated (Un/frac), monomer, dimer, HMW2 and HMW1 WT E2 to neutralize G2a HCVcc. p values were determined using the Kruskal Wallis test with Dunn's post-test correction for multiple comparison (Prism v 6.0f). For B-G, the horizontal bar is the geometric mean. For B-C, the dotted line is the background binding. For D-G, the dotted line is the mean neutralization value for 4-5 no antigen control animals. ID50 neutralization data was derived from three independent experiments performed in triplicate.

Supplemental figure 4.
Ability of monoclonal antibodies to bind different species of Δ123. An equivalent amount of each Δ123 species was absorbed to solid-phase plates followed by serial dilutions of each monoclonal antibody as indicated by the reactivity towards the C-terminal 6xHis epitope tag shown for two different experiments. The relative binding of each antibody towards dimer, HMW2 and HMW1 was calculated from the mid-point of each binding curve, between 5-10 times background absorbance to BSA, and the fold difference in binding calculated relative to monomer was used for the construction of Table 1.
Supplemental figure 5. HMW1 forms of Δ123 restrict the presentation of epitopes in two regions of E2. The contact residues for MAbs used in this study are indicated by an X derived from mutagenesis studies or from crystal structures of MAbs with their epitopes as listed in Table 1. The MAbs that show ~5 fold reduction in binding ability are shown in orange whilst those having an ~ 10 fold reduction in binding are shown in red. The Grey shaded areas correspond to uniquely occluded sites in HMW1 Δ123. Regions that do not contain epitopes have been deleted. Figure 6. Ability of immune sera to inhibit the binding between (A) homologous G1a H77 E2 and recombinant CD81-LEL protein and (B) heterologous G2a JFH1 E2 and recombinant CD81-LEL. Serial dilutions of individual immune sera were incubated with a constant amount of E2 and added to CD81-LEL coated ELISA plates. The percentage binding is calculated as the (E2 binding in the presence of immune serum)/(E2 binding in the absence of immune serum)x100. Figure 7. Protein G sepharose (PGS) removes neutralizing antibody activity. Immune sera from (A) HMW1, (B) HMW2 and (C) no antigen control animals was added to H77c HCVpp before (white) and after (black) PGS absorption of IgG. The geometric mean of triplicate 1/40 serum dilutions is shown. Figure 8. Pooled immune sera prepared from each vaccine group was serially diluted and mixed with VSV-G pseudotyped retroviral particles. The mean ± standard deviation of luciferase activity of triplicate samples is shown. Figure 9. ClustalW alignment of the corresponding E2 RBD region of HCV isolates used in neutralization assays. H77c (AF009606, G1a), J6 (AF177036, G2a), S52 (GU814263, G3a), ED43 (GU814265, G4a), SA13 (AF064490, G5a), EUHK2 (Y12083, G6a), QC69 (EF108306, G7a). HVR1, HVR2 and igVR/VR3 are shown in red/orange. Blue residues are CD81 binding regions. Figure 10. ID50 neutralization titers for HMW2 immune sera against G4a, G5a, G6a, and G7a HCVcc viruses. Dashed line represents the mean neutralization observed for the no antigen control group.

Supplemental
Supplemental Figure 11. Competitive binding of human MAbs specific to three major neutralization epitopes on E2 (HCV1, HC84.27 and AR3C) and one major non-neutralizing epitope (CBH4B) in the presence of serial dilutions of guinea pig serum raised to HMW1 or monomeric Δ123 antigen. Binding levels in the presence of guinea pig sera raised to adjuvant alone (no antigen, dotted blue lines) are shown as well as no sera controls (solid blue lines).