Antibodies Targeting Novel Neutralizing Epitopes of Hepatitis C Virus Glycoprotein Preclude Genotype 2 Virus Infection

Currently, there is no effective vaccine to prevent hepatitis C virus (HCV) infection, partly due to our insufficient understanding of the virus glycoprotein immunology. Most neutralizing antibodies (nAbs) were identified using glycoprotein immunogens, such as recombinant E1E2, HCV pseudoparticles or cell culture derived HCV. However, the fact that in the HCV acute infection phase, only a small proportion of patients are self-resolved accompanied with the emergence of nAbs, indicates the limited immunogenicity of glycoprotein itself to induce effective antibodies against a highly evolved virus. Secondly, in previous reports, the immunogen sequence was mostly the genotype of the 1a H77 strain. Rarely, other genotypes/subtypes have been studied, although theoretically one genotype/subtype immunogen is able to induce cross-genotype neutralizing antibodies. To overcome these drawbacks and find potential novel neutralizing epitopes, 57 overlapping peptides encompassing the full-length glycoprotein E1E2 of subtype 1b were synthesized to immunize BALB/c mice, and the neutralizing reactive of the induced antisera against HCVpp genotypes 1–6 was determined. We defined a domain comprising amino acids (aa) 192–221, 232–251, 262–281 and 292–331 of E1, and 421–543, 564–583, 594–618 and 634–673 of E2, as the neutralizing regions of HCV glycoprotein. Peptides PUHI26 (aa 444–463) and PUHI45 (aa 604–618)-induced antisera displayed the most potent broad neutralizing reactive. Two monoclonal antibodies recognizing the PUHI26 and PUHI45 epitopes efficiently precluded genotype 2 viral (HCVcc JFH and J6 strains) infection, but they did not neutralize other genotypes. Our study mapped a neutralizing epitope region of HCV glycoprotein using a novel immunization strategy, and identified two monoclonal antibodies effective in preventing genotype 2 virus infection.


Introduction
Hepatitis C virus (HCV) is one of the major causes of liver disease. An estimated 185 million people worldwide are infected with hepatitis C [1] and have a high risk of liver cirrhosis, hepatocellular cancer and death [2]. There is no prophylactic or therapeutic vaccine available for HCV, although rapid progress in hepatitis C treatment has been made due to the emergence of direct-acting antiviral (DAA) drugs. Once infected with HCV, most patients develop chronic hepatitis and only a small number of individuals clear the virus. Cellular immunity is thought to play a vital role in viral clearance [3][4][5]. Recently, accumulating evidence has highlighted the importance of humoral immunity in controlling infection [6,7]. Neutralizing antibodies (nAbs) were associated with the eradication of the virus both in the acute and chronic infection phases [7,8].
HCV glycoprotein, which mediates virus entry by interplay with host co-receptors, is the natural target of nAbs. Many nAbs with potent cross-genotype neutralizing reactive have been identified based on artificial glycoprotein immunogens, including recombinant E1E2, soluble E2, HCV pseudoparticles (HCVpp) and cell culture-derived HCV (HCVcc), mimicking the spare structure of the wild type virus glycoprotein [9][10][11]. Recently, the crystal structure of E2 was determined. The epitopes of these nAbs were mostly mapped to the "broadly neutralizing face", mainly within the N terminal of E2 and approximately comprising amino acids (aa) 412-453 and 502-535 [12][13][14]. The E2-CD81 interaction region was also thought to be within this domain. The fact that only a few infected patients are resolved during the acute phase in the presence of nAbs implies that the epitopes recognized by the most potent and effective nAbs may be relatively weakly immunogenic and not reactive in most patients with hepatitis C. In the HCV E1E2 steric structure, the epitopes may be buried by adjacent conformation and not accessible for nAbs. On the contrary, variable regions of E2 are immunodominant [15], but they only raise strain-specific protective immunity, which is unable to neutralize highly evolved HCV [16]. Thus, the strategy of solely adopting a glycoprotein immunogen may miss some neutralizing epitopes outside the "broadly neutralizing face". It is of interest to determine whether there are other novel neutralizing epitopes using a different immunization approach.
Another factor deserving attention is that, in previous studies, the glycoprotein sequence was based on the H77 strain, which represented the most prevalent genotype 1a worldwide. Other genotypes/subtypes were rarely studied, although theoretically one genotype/subtype immunogen was capable of inducing a cross-genotype nAbs [9], and the sera of chronic hepatitis C patients of one subtype were reported to have broadly neutralizing potential [17].
To address the issues mentioned above, we employed a different immunization strategy. First, we synthesized overlapping peptides encompassing the full-length glycoprotein E1E2 (not including the transmembrane domain of E2) instead of glycoprotein as the immunogen. Secondly, the immunogen sequence was mostly according to subtype 1b strain H77, which was prevalent globally and was the dominant subtype in China. Our study revealed that peptides of subtype 1b did induce nAbs, and the neutralizing epitopes of HCV glycoprotein were more broadly distributed than expected. Furthermore, we identified two monoclonal antibodies (mAbs), 2O18 and 2C21, recognizing epitopes aa 454-463 and aa 611-618 of E2, respectively, which efficiently blocked genotype 2 virus (HCVcc, JFH and J6 strains) infection in vitro.
Taken together, our study reveals the neutralizing domain of HCV glycoprotein from a new angle and also identifies two monoclonal antibodies that recognize novel glycoprotein epitopes blocking genotype 2 virus infection. These results facilitate future vaccine design and development.

Ethics Statements
All immunization procedures in BALB/c mice were conducted by Abmart Inc. (Shanghai, China; http://www.ab-mart.com) according to national guidelines (the Regulations for the Administration of Affairs Concerning Experimental Animals, China) and were approved by the Ethics Committee of Peking University People's Hospital.

Peptide Synthesis
A peptide library consisting of 57 peptides (Tables 1 and 2) averaging 20 amino acid residues long and overlapping by 10 residues encompassing the complete sequence of HCV glycoprotein E1E2 (not including the transmembrane domain of E2, aa 718-746) of a subtype 1b "reference strain" was synthesized by Invitrogen Corp. (Shanghai, China). The "reference strain" was a consensus sequence generated by alignment of 43 chronic hepatitis C patient viral sequences belonging to subtype 1b (S1 File).

Animal Immunization and Antibody Generation
Fifty μg of peptide was used to inoculate BALB/c mice (n = 3 for each peptide) with complete adjuvant to elite polyclonal antibodies (antisera), and repeated at day 7, 14 and 21 with incomplete adjuvant. The antisera were collected 4 weeks post-immunization. The antibody concentration in the sera was validated by ELISA (>1:200). The production of the monoclonal antibody was conducted according to standard hybridoma technology. Ascites from inoculated BALB/c mice were collected at 4-6 weeks post-immunization, and monoclonal antibodies were purified from ascites using a protein A column. After the experiment, all mice were euthanized by CO 2 asphyxiation.

Neutralization Assays
For HCVpp neutralization assays, 8×10 3 Huh7.5 cells were seeded into 96-well plates one day before infection. Ten μL HCVpp stock was incubated with antiserum, monoclonal antibodies or normal mouse serum/IgG (control group) at various concentrations, plus 4 μg/ml polybrene at 37°C for 1 hour. The mixtures (100 μL in total) were then added to each well. After incubation at 37°C for 6 hours, the mixtures were replaced with complete culture medium and incubated for 72 hours. HCV infection was evaluated by measuring luciferase activity (Promega, Cat. E1501). The value of %Neutralization was calculated as (1-luciferase value of experimental group/luciferase value of control group) ×100%. The IC 50 of the antibody (required to neutralize 50% of virus) was determined based on a neutralization curve generated from a series of 2-fold dilutions tested in triplicate.
For the HCVcc neutralization assays, 6×10 3 Huh7.5 cells were seeded into 96-well plates one day before infection. A sample of 100 FFU HCVcc was incubated with monoclonal antibodies or normal mouse IgG (control group) at 37°C for 1 hour. The mixture was then incubated with Huh7.5 cells for 4 hours. Seventy-two hours post-infection, HCV infection was evaluated by counting HCV NS3-positive foci in an indirect immunofluorescence assay [21]. Each test was performed in triplicate. %Neutralization was calculated as (1-foci of experimental group/foci of control group) ×100%.

Epitope Mapping
To map the precise epitopes of mAbs 2O18 and 2C21, three overlapping peptides covering aa 444-463 and aa 604-618, respectively, were synthesized. The plates were coated with 5 μg/ ml of peptide and blocked with 4% PBST. The two mAbs were incubated and binding was detected in an ELISA format as described previously [22]. Irrelevant rabies virus peptide (VNLHDFRSDEIE) served as the negative control. Peptide PUHI26 (aa 444-463) and PUHI45 (aa 604-618) served as the positive controls.

Alanine Replacement Mutagenesis
To identify residues crucial for mAbs binding, alanine mutagenesis of the epitope residues was analyzed in a GNL capture ELISA assay [23]. The OD value of antibody binding to epitope with replacing <50% of wild type residues was defined as positive. Each experiment was performed in triplicate.

Data Analysis and Software
Statistical comparison of the neutralization between experimental and control groups was performed using the χ2 test (GraphPad Prism 5). The neutralizing entropy of 47 peptides inducing antisera against HCVpp was analyzed with Genesis software (http://genome.tugraz.at/ genesisclient/genesisclient_description.shtml) [24].

Neutralizing Epitope Domain of HCV Glycoprotein E1E2
A series of 57 overlapping peptides covering the full-length glycoprotein E1E2 (subtype 1b) was used as an immunogen to produce antibodies (Tables 1 and 2 To determine the neutralizing reactive of the antibodies induced by 47 peptides, the antisera were 50-fold diluted and tested in HCVpp neutralizing assays (genotypes 1-6). The relative neutralization of antisera was determined (Tables 3 and 4) and analyzed with Genesis software, as described previously [24]. The neutralizing values of the antisera were converted to different kinds and degrees of colors ( Fig 1A). The green color indicated neutralizing reactive antisera, and the red color indicated antisera without neutralizing reactive or promoting virus entry.
There is not a precise value of relative neutralization to define antisera as "neutralizing antisera". In this study, we did not adopt rigorous standards (such as neutralization >50%) because during the neutralizing mAbs screening course, we found that the same peptide immunogen simultaneously induced neutralizing and non-neutralizing antibodies, and the latter somewhat promoted virus infection, which may offset the effect of neutralizing antibodies in antisera (S1 Fig). In our study, the difference between antisera with neutralizing reacitve 30% and the normal mice serum group (5.6%) was statistically significant (P< 0.01, χ2 test). In addition, the sera with neutralizing reactive 30% prevented virus entry in a dose-dependent manner (S2 Fig). Thus, we defined the antisera with relative neutralization 30% as "neutralizing antisera".

Epitope Mapping and Determination of Residues Crucial for Antibody Binding
To map the precise epitopes of 2O18 and 2C21, three overlapping peptides, encompassing aa 444-463 and aa 604-618, were synthesized and their relative binding to antibodies was determined in by ELISA (Fig 5). The results demonstrated that 2O18 and 2C21 recognized epitopes aa 454-463 and aa 611-618, respectively. To determine the residues crucial for mAbs binding, we conducted alanine replacement mutagenesis (Fig 6). The results demonstrated that residues R460, P461, I462 and D463 were crucial for 2O18 binding, while L615 and W616 were responsible for 2C1 binding.

Conservation Analysis of 2O18 and 2C21 Epitopes
To analyze the conservation of 2O18 and 2C21 epitopes, a total of 536 HCV sequences comprising genotypes/subtypes 1-6 were manually retrieved from the NCBI database (Table 5, S2 File). All the sequences were submitted to Clusta Omega (http://www.ebi.ac.uk/Tools/msa/ clustalo/) for alignment, and the conserved epitope residues are marked in Fig 7. Residues M456, C459 and I462 of the 2O18 epitope were conserved, while R614 and L615 of the 2C21 epitope were conserved. Thus, the conserved residue crucial for 2O18 binding was I462, and L615 for 2C21.

Discussion
Development of an effective vaccine against HCV, such as a neutralizing antibody, is an urgent need because of the inaccessibility and unaffordability of DAAs in many developing countries  and regions in the near future. Neutralizing antibody plays an important role in controlling HCV infection and is associated with viral clearance [6,7]. Many nAbs were identified using glycoprotein immunogens, with the immunogen sequence according to the subtype 1a H77 strain. These epitopes were mostly mapped to a "neutralizing face", mainly within the aa 412-535 region [12,25]. In this study, we adopted a different immunization strategy, using overlapping peptides instead of glycoproteins, and the 1b strain sequence instead of H77, to map the neutralizing domain of HCV glycoprotein. Our study revealed that the chief neutralizing     [29]. The use of a particular carrier in combination with a peptide immunogen may improve the production of the antibody [30]. Of note, 2O18 and 2C21 antibodies appeared to promote infections of most tested HCVpp at low doses (<16μg/ml for 2O18, <8μg/ml for 2C21), the antibody Fc receptor was probably account for that [31]. Based on the identified neutralizing epitopes, we isolated two nAbs, 2O18 and 2C21. Both displayed cross-genotype neutralizing reactive in HCVpp neutralizing assays, and potently blocked HCVcc genotype 2 JFH and J6 strain infection. Epitope mapping revealed that some of the residues crucial for antibody binding were highly conserved, indicating that the interplay between antibodies and virus epitopes was relatively stable. Because genotype 2 is prevalent in the Asian-Pacific region and accounts for a large proportion of HCV infection, it is advisable to further test the efficacy of 2O18 and 2C21 in more HCVcc genotype 2 strains, and in vivo.
2O18 and 2C21 did not neutralize HCVcc of genotypes other than genotype 2, but displayed broad neutralizing reactive in HCVpp. The reason for this difference is elusive and needs further elucidation. The strains used in HCVpp and HCVcc were not exactly the same. Recently there were similar observations for other neutralizing antibodies [32].
In summary, our study reveals the neutralizing region of HCV glycoprotein by use of the peptide immunization strategy. We developed two monoclonal antibodies with potent neutralizing reactive against genotype 2 viruses. These findings have important implications for HCV vaccine design and development.  , 1b and 2a). Some of the antibodies (clones 2-6 and 14 in 11365 group, and clones 7-14, CBH-5 in 11368 group) were not tested in genotype 2a HCVpp neutralizing asssay. Clone 2C21 in 11368 group was only tested in genotype 1b HCVpp neutralizing assay. Antibodies CBH-5 (13μg/ml) and Conservation analysis of antibodies 2O18 and 2C21 epitopes. Five hundred and thirty-six HCV protein sequences from Table 5 were retrieved for alignment, and the conserved amino acid residues within 2O18 and 2C21 epitopes were marked. ":" indicates highly conserved residues, while "*" indicates absolutely conserved residues across 536 sequences.