Article Text
PDF
Abstract
Objective Despite the development of highly effective direct-acting antivirals, a prophylactic vaccine is needed for eradicating HCV. A major hurdle of HCV vaccine development is to induce immunity against HCV with high genome diversity. We previously demonstrated that a soluble E2 (sE2) expressed from insect cells induces broadly neutralising antibodies (NAbs) and prevents HCV infection. The objective of this study is to develop a multivalent HCV vaccine to increase the antigenic coverage.
Design We designed a trivalent vaccine containing sE2 from genotype 1a, 1b and 3a. Mice and rhesus macaques were immunised with monovalent or trivalent sE2 vaccine, and sera or purified immunoglobulin were assessed for neutralisation against a panel of cell culture-derived virion (HCVcc) of genotype 1–7 in cell culture. Splenocytes from the vaccinated macaques were assessed for HCV-specific T cell response.
Results We showed that the trivalent vaccine elicited pangenotypic NAbs in mice, which neutralised HCVcc of all the seven genotypes more potently than the monovalent vaccine. Further analyses demonstrated that each sE2 component of this trivalent vaccine elicited unique spectrum of NAbs which acted synergistically to inhibit HCV infection. Finally, the trivalent vaccine triggered stronger and more uniform multigenotypic neutralising antibody response than the monovalent vaccine in rhesus macaques.
Conclusions In summary, we developed a trivalent HCV vaccine that induces broad and synergistic-acting neutralising antibodies in mice and non-human primates.
- hepatitis c
- infectious disease
Statistics from Altmetric.com
Significance of this study
What is already known on this subject?
No prophylactic HCV vaccine is available yet. A major hurdle of HCV vaccine development is to induce immunity against HCV with high genome diversity. Although a number of HCV vaccine candidates have been developed, none of them have moved to the late clinical phases.
We previously reported that recombinant soluble E2 (sE2) of a GT1b strain produced from insect cells could induce neutralising antibodies in mice and macaques and also protect humanised mice from HCV infection.
The E2 antigen production is simple and has a high yield (up to 100 mg/L culture supernatants), making it technically possible to explore a multivalent vaccine that consists of E2 of multiple genotypes to increase the antigenic coverage.
What are the new findings?
We developed a trivalent vaccine containing sE2 from genotype 1a, 1b and 3a.
The trivalent vaccine elicited stronger pangenotypic neutralising antibodies than the monovalent vaccine in mice.
We found that each sE2 component of this trivalent vaccine elicited unique spectrum of neutralising antibodies, which acted synergistically to inhibit HCV infection.
We found that the trivalent vaccine triggered stronger and more uniform multigenotypic neutralising antibody responses than the monovalent vaccine in rhesus macaques.
How might it impact on clinical practice in the foreseeable future?
The trivalent sE2 vaccine is a promising prophylactic HCV vaccine candidate for several following reasons. First, it induces broad and synergistic-acting neutralising antibodies in mice and non-human primates. Second, it induces more uniform neutralising activity in rhesus macaques which have varying genetic background. Third, no immunogenic interference between individual sE2 components within the trivalent cocktail was observed. Finally, antigen production can be easily scaled up to the manufacture level.
Introduction
HCV, a plus-strand RNA virus, infects approximately 185 million people worldwide who could develop into liver cirrhosis and hepatocellular carcinoma.1 No vaccine is licensed in the world yet. Although recently approved direct-acting antiviral agents (DAAs) can cure HCV infection, problems remain including therapy accessibility and potential drug resistance.2 In addition, although DAAs could cure HCV infection, some patients may still develop into progressive liver failure.3 Historically, it has been difficult to eliminate viral infectious diseases by therapeutics alone. Therefore, it remains imperative to develop an effective prophylactic HCV vaccine.
The T cell immune response plays an important role in the clearance of HCV infection.4 5 Moreover, accumulating evidence demonstrated that broadly neutralising antibodies (bNAbs) elicited at early stages of infection are associated with spontaneous clearance of HCV infection in people who inject drugs.6 Importantly, HCV reinfection of individuals who spontaneously clear the initial infection is often associated with reductions in the magnitude and duration of viraemia, as well as robust cellular immune responses and cross-reactive antibody responses.7 Therefore, development of a prophylactic vaccine that promotes cellular and humoral immunity to prevent HCV infection would be feasible.
HCV glycoproteins E1 and E2 form a heterodimer to constitute viral envelope. E2, directly interacting with cellular receptors CD81 and SR-B1 to mediate HCV entry, possesses majority of epitopes recognised by NAbs8–11 and thus is an ideal antigen candidate for HCV vaccine design.12 Various types of E2-based or E1E2-based vaccine candidates have been tested. A DNA vaccine coding E2 elicited specific antibody responses in rodents.13–15 The virus-like particles (VLPs) containing core, E1 and E2 produced from baculoviral expression system elicited humoral and cellular response in mice and non-human primates16 17 and offered protection against HCV infection in chimpanzees.18 Prime-boost immunisation with VLP displaying E2 and/or E1 of genotype (GT) 1a induced effective NAbs in mice and macaques.19 Cell culture-derived virion (HCVcc) has also been explored for its potentials in HCV vaccine development. IgG from HCVcc-immunised mice prevented HCV infection in chimeric mice engrafted with human livers20 and the inactivated HCVcc induced NAbs and T cell response in marmosets.21 A subunit vaccine candidate using GT1a-derived E1-E2 heterodimer expressed from mammalian cells could induce NAbs in rodents.22 The result of phase I clinical trial of this vaccine showed that 3 out of 16 human volunteers generated strong antibody response to neutralise homologous GT1a HCV, and 1 volunteer generated bNAbs towards all genotypes of HCVcc.23 24 Mammalian cell-produced E2 core domain lacking hypervariable regions has also been reported to elicit efficient bNAbs in rodents.25 26 We recently developed an HCV vaccine candidate based on soluble E2 (E2, aa384–661) of Con1 strain (GT1b). The E2 protein is expressed from Drosophila S2 cells at a remarkably high yield (up to 100 mg/L) and has been shown to be highly immunogenic possibly due to its distinct glycosylation patterns in insect cells.27 28 Importantly, active immunisation with the E2 vaccine efficiently protected genetically humanised mice from heterologous HCV challenge.27 The objective of this study was to improve the breadth and potency of immunity induced by E2, particularly to the HCV strains that were inefficiently neutralised by the Con1sE2-induced antibodies, such as H77 (GT1a) and S52 (GT3a) which are prevalent subgenotypes in the world.29
Materials and methods
Viruses and plasmids
A pangenotypic panel of HCVcc strains used in neutralisation assay were described previously.27 30 To construct the expression plasmid pMT-sE2, the DNA sequence coding for sE2 (aa384-661) of H77 (GenBank accession number JF343780.2), Con1 (AJ238799.1) and S52 (JF343784.2) were PCR amplified using primers (see online supplementary table S1) and inserted into the NcoI and XbaI sites of vector pMT/BiP/V5-HisA (Invitrogen, Carlsbad, California, USA).
Supplementary file 1
sE2 protein production
The expression and purification of recombinant sE2 proteins from various HCV strains were as previously described.27
Receptor and antibody binding assay
The antigen-binding assay was performed to determine the binding of sE2 to CD81 receptor and to E2-specific antibody AR3A and AP33. Briefly, a series of concentrations of human CD81LEL proteins (Sino Biological, Beijing, China) precoated in 96-well plates were incubated with 1 µg of sE2 or bovine serum albumin, followed by incubation with horseradish peroxidase (HRP)-conjugated anti-6xHis mAb for 2 hours. For the antibody-binding assay, a series of concentrations of sE2 proteins precoated in 96-well plates were incubated with 100 ng of E2-specific mAb AR3A or AP33,9 31 followed by incubation with HRP-conjugated antihuman IgG or antimouse IgG for 1 hour. After substrate colour reaction, colorimetric analysis was performed at 450 nm.
Animal immunisation
All animal studies were reviewed and approved by the Laboratory Animal Management and Ethics Committee at the Institut Pasteur of Shanghai. Mice were obtained from Shanghai SLAC Laboratory Animal and bred in specific pathogen-free facility at Institut Pasteur of Shanghai. Rhesus macaques were obtained from Suzhou XiShanZhongke and maintained in clean facility in Non-Human Primates Experiment Platform of Institute of Neuroscience, Chinese Academy of Sciences.
To assess the immunogenicity of trivalent sE2 vaccine, female BALB/c mice were divided into five groups (six mice each group) and intraperitoneal injected at week 0, 2, 4 and 8 with 30 µg of monovalent sE2 antigen of Con1, H77 and S52, respectively, or the trivalent sE2 cocktails (10 µg of each) supplemented with 500 µg of Inject Alum (Invitrogen) plus 25 µg of CpG 7909 (also known as CpG2006 or PF-3512676; 5′-TCGTCGTTTTGTCGTTTTGTCGTT-3′; Sangon Biotech, Shanghai, China).
Adjuvant alone group was set up as a control. Sera were collected at week 0, 10, 22 and 44 for ELISA and neutralisation assays.
To assess the immunogenicity of the trivalent vaccine in rhesus monkey, 12 monkeys aged 3–5 years were divided into three groups (monkey information was shown in online supplementary table S2), two of which were monovalent Con1sE2 group and trivalent sE2 group (3 male and 2 female monkeys per group) that were intramusculary injected at month 0, 1, 2, 3, 4 and 7 with 200 µg of antigens formulated with AS04 adjuvants, respectively. The other group (one male and one female) was injected with adjuvants alone as a control. The sera were collected at month 0, 3, 4, 5, 7 and 8 to determine the antibody titres and neutralisation.
Antibody titres assay and neutralisation assay
HCV E2-specific antibodies in sera were quantified by ELISA as previously described.27 For neutralisation assay, around 100 focus-forming units of HCVcc were incubated for 1 hour with serially diluted sera or purified IgG. The mixture was then added to 104 Huh7.5.1 cells preseeded in 96-well plates. Viral inocula were removed 4 hours later, and the cells were kept in fresh medium for 3 days. Immunofluorescence assay was performed to quantify HCV infection by counting the number of foci. 50% inhibitory dilution (ID50) was calculated from six-point dilution curves.
Determination of cellular immune response by ELISPOT
Splenocytes were isolated from spleens of macaques at month 8. After sequential filtration by 100 µm and 40 µm of cell strainers (Corning), the ACK lysis buffer (Beyotime, Shanghai, China) was added to clear red blood cells. Splenocytes were obtained by adding complete RPMI-1640 medium to stop reaction and then counted for enzyme-linked immunospot (ELISPOT) experiment.
Interferon gamma (IFN-γ) and interleukin 4 (IL-4) ELISPOT were performed following the manufacture’s protocol. Briefly, the polyvinylidene fluoride plates (Millipore, Bedford, Massachusetts, USA) were precoated with 500 ng of IFN-γ-ELISPOT capturing antibody (51-2555KZ; Becton, Dickinson and Company (BD)) or 200 ng of IL-4 ELISPOT capturing antibody (51-1865KZ, BD) at 4°C overnight and then blocked with 200 µL/well of complete RPMI-1640 medium at 37°C for 3 hours. One million per well of isolated splenocytes were transferred to the precoated 96-well plates and stimulated with 10 µg/mL of sE2 at 37°C for 48 hours. After decanting the cells and washing with ice-cold double-distilled water once and phosphate-buffered saline for five times, the plates were incubated with 2 µg/mL of Biotin-conjugated Human IFN-γ-ELISPOT detection antibody (51-1890KZ, BD) or 4 µg/mL of biotin-conjugated human IL-4-ELISPOT detection antibody (51-1850KZ, BD) at 37°C for 2 hours, followed by incubation of alkaline phosphatase-conjugated streptavidin (Mabtech, Nacka, Sweden) for 1 hour. After washing for three times, nitroblue tetrazolium/5-bromo-4-chloro-3-indolylphosphate substrate (Promega, Madison, Wisconsin, USA) was added for colour development. The cytokine-secreting cell spots were analysed on a CTL Immunospot Reader (Cellular Technology, Shaker Heights, Ohio, USA).
Synergy analysis
We used Chou-Talalay method32 to analyse the cooperation in neutralisation. Briefly, monovalent E2-immunised sera alone or in mixture were twofold diluted around ID50 value. The neutralisation percentages at each dose were entered into CalcuSyn V.2.0 software as fraction affected (Fa) ranging from 0.01 to 0.99, then dose–effect curves were generated, and Combination Index (CI) at ED50, ED75 and ED90 were calculated based on the Fa–CI plots. CI <1 and>1 represent synergism and antagonism, respectively.
Statistical analysis
Significant differences were calculated with Student’s t-test or one-way analysis of variance followed by Tukey’s multiple comparison test. Significant values are shown as: no significant difference P>0.05, *P<0.05, **P<0.01, ***P<0.001. All P value analyses were calculated by GraphPad Prism V.5.0 c.
Results
Comparison of the immunogenicity of trivalent and monovalent sE2 vaccines in mice
We previously reported that a subunit vaccine based on sE2 of Con1 strain (GT1b) induces bNAbs in mice and rhesus monkey.27 28 Here, we designed a trivalent formulation that includes sE2 from Con1 (GT1b), H77 (GT1a) and S52 (GT3a) to increase the antigenic coverage. These three sE2 proteins were expressed and purified from Drosophila S2 cell supernatants (figure 1A–C) as previously reported27 and mixed in an equal ratio and compared with each individual sE2 protein. ELISA analysis shows that these E2 proteins reacted in a dose-dependent manner with E2 receptor CD81 (figure 1D) or E2-specific antibodies AR3A9 and AP33,31 which recognise a conformational and linear E2 epitope, respectively (figure 1E), suggesting that the three sE2 proteins were properly folded.
Production and characterisation of the sE2 in insect cells. (A) Schematic representation of sE2 expression plasmid. The truncated HCV E2 (aa384–661) lacking the C-terminal transmembrane domain was tagged with the BiP signal peptide at the N-terminal and His-tag at the C-terminal, and its expression was driven by the metallothionein promoter. (B) CBB staining and (C) Western blot analysis of purified sE2 from three subgenotypes. Anti-6xHis mAb was used to detect sE2. (D) CD81 binding ELISA. Precoated CD81of serial concentrations were incubated with 1 µg of sE2 proteins. Mean±SEM of duplicate wells are shown. (E) E2-specific antibody-binding ELISA. Precoated sE2 of serial concentrations were incubated with 0.1 µg of E2-specific antibodies AR3A or AP33. Mean±SEM of duplicate wells are shown. BSA, bovine serum albumin; CBB, Coomasie Brilliant Blue; mAb, monoclonal antibody; MT, metallothionein; sE2, soluble E2.
To determine the immunogenicity of this trivalent sE2 cocktail, mice were immunised with 30 µg of trivalent or monovalent sE2, respectively. Adjuvant alone group was set up as a control. ELISA analysis demonstrated that the monovalent and trivalent sE2-induced E2-specific antibody titres over 105, which persisted for as long as 44 weeks, almost a mouse lifetime (figure 2A). Moreover, monovalent sE2-induced antibodies displayed a stronger binding activity towards homologous than heterologous antigens. However, the trivalent sE2-immunised sera always exhibited strong binding to each of three sE2 antigens, inferring an advantage over the monovalent sE2 in the antigenic breadth.
Immunogenicity analysis of the trivalent and monovalent sE2 vaccines in mice. (A) The E2-specific antibody titres in immunised mouse sera. BALB/c mice (n=6 per group) were immunised intraperitoneally at weeks 0, 2, 4 and 8 (indicated by arrow heads below the x-axis). The E2-specific antibody titres at week 0, 10, 22 and 44 were measured by ELISA and expressed as mean±SEM of six mice within each group. (B) Neutralisation of HCVcc by sE2-induced antisera. The 1:40 diluted antisera at week 10 were tested for neutralisation against 10 HCVcc strains. Each symbol represents one animal and data were expressed as mean±SEM for each group. The data were representative of three independent experiments. (C) ID50 of antisera. ID50 was defined as the dilution of sera able to neutralise 50% of HCVcc infectivity. Antisera were pooled for each group and diluted two times serially from 20 to 10 240 for the neutralisation test. Neutralisation curves were fitted by non-linear regression, and the ID50 values were calculated by GraphPad Prism. HCVcc, cell culture-derived virion; ID50, 50% inhibitory dilution; sE2, soluble E2.
Next, we performed a neutralisation assay using the sera collected at week 10. As shown in figure 2B, the trivalent sE2 induced the strongest neutralising activities against HCVcc strains, such as H77, Con1, PR52B6mt, S52, SA13 and QC69 at 1/40 sera dilution. We noticed that as compared with our previous report,27 Con1sE2-immunised sera neutralised slightly less efficiently some HCVcc strains, such as H77, JFH1, S52, SA13 and HK6a, possibly because we used less antigens and adjuvants (30 µg sE2/0.5 mg aluminium in the current study compared with 50 µg sE2/1 mg aluminium in the previous experiments).27 To compare the neutralising activity more quantitatively, we pooled week 10 sera of 6 mice within the same group and used sera dilution-neutralisation curves to determine the serum dilution inhibiting 50% of viral activity (ID50) (figure 2C). For the purpose of analysis, the ID50 values were classified into four grades, G1 (>810), G2 (270–810), G3 (90–270) and G4 (30–90). The ID50 values of the trivalent sE2-immunised sera towards 3 out of 14 HCVcc were in G1, 4/14 in G2, 4/14 in G3 and 3/14 in G4, while the ID50 values of the monovalent sE2-immunised sera were as follows: anti-H77sE2 sera (G1-1/14, G2-1/14, G3-6/14, G4-6/14), anti-Con1sE2sera (G1-0/14, G2-3/14, G3-5/14, G4-6/14), anti-S52sE2 sera (G1-0/14, G2-2/14, G3-5/14, G4-7/14). The ID50 values of the trivalent sE2-immunised sera towards seven HCVcc strains were higher than 270 (G1-G2), significantly more than those of the monovalent groups (H77-2/14, Con1-3/14, S52-2/14). Of note, the trivalent formulation enhanced neutralisation potency towards homologous strains (Con1, H77 and S52) as well as heterologous strains (JFH1, PR63cc, J8 and ED43). Importantly, the threefold reduced dosage of each sE2 component in the trivalent formulation did not reduce its ability to induce neutralisation against any HCVcc strains, particularly for those that were neutralised much more efficiently by homologous sE2-vaccinated sera, for example, H77. On the contrary, despite the decreased dosage of each sE2 antigen used for immunisation compared with monovalent sE2 vaccine, the trivalent sE2 vaccine induced much more potent neutralisation towards some HCVcc strains, such as Con1, JFH1 and PR63cc.
Synergistic neutralising effects of polyclonal antibodies induced by the trivalent vaccine
As shown in figure 2C, the ID50 of the trivalent sE2-vaccinated sera at week 10 was greater than that of each monovalent sE2-vaccinated sera against some HCVcc strains, such as Con1, JFH1, PR63cc and S52. To confirm this observation, we determined the ID50 of mixture of three monovalent sE2-immunised week 22 sera and compared it with that of monovalent or trivalent sE2-immunised sera (figure 3A,B). HCVcc strains, including Con1, JFH1, PR63cc and S52 where significantly higher neutralising efficiencies were achieved with the trivalent sE2-vaccinated sera, as well as strain H77 where no such an effect was observed, were used in the neutralisation assay. Consistent with the results using week 10 sera (figure 2C), the week 22 sera of trivalent sE2-immunised mice displayed significantly higher neutralising activity against HCVcc strains Con1, JFH1, PR63cc and S52 than each monovalent sE2-immunised serum. Importantly, this improved neutralisation effect was recapitulated when the three monovalent sE2-immunised sera were mixed in an equal ratio (H77 +Con1+S52), suggesting that antibodies induced by three sE2 antigens may cooperate in neutralising these HCV strains. On the contrary, for H77 strain that was predominantly neutralised by the H77sE2-immunised sera, the three-serum mixture (ID50: 261.8) had much lower neutralising activity as compared with the H77sE2-and trivalent sE2-immunised sera (ID50: 1394 and 1252, respectively), apparently resulting from dilution of more effective anti-H77sE2 sera by less effective anti-Con1sE2 and anti-S52sE2 sera.
Synergistic neutralising effects induced by the trivalent vaccine. (A) Neutralisation curve of the week 22 vaccinated sera (monovalent, trivalent and three-serum mix). Week 22 sera of 6 mice within each group were pooled for the neutralisation assay. The sera of the three monovalent sE2-vaccinated groups were mixed at an equal ratio (H77 +Con1+S52). All sera were diluted two times serially to test neutralisation against HCV strains H77, Con1, JFH1, PR63cc and S52. Neutralisation curves were fitted by non-linear regression. (B) ID50 of the sera. ID50, defined as the dilution of sera able to neutralise 50% of HCVcc infectivity, was calculated by GraphPad Prism. (C) CI values at ED50, ED75 and ED90. The CI was calculated using the CalcuSyn software V.2.0 from the neutralisation of the three-serum mix versus monovalent sE2-immunised sera. The dotted line represents CI=1 below of which was defined as synergism.32 CI, combination index; ED50, 50% effective dose; HCVcc, cell culture-derived virion; ID50, 50% inhibitory dilution; sE2, soluble E2.
To analyse this putative synergistic neutralising effect, we performed statistical analysis to calculate the CI, a parameter that evaluates the degree of cooperation among sera induced by individual antigens.32–34 The CI of the three-serum mixture versus monovalent sE2-vaccinated sera at ED50, ED75 and ED90 were presented in figure 3C and online supplementary table S3. The results showed that the CI values of sera mixture against Con1, JFH1, PR63cc and S52 were less than 1, a hallmark of synergism,32 suggesting that the improved neutralising activity by the trivalent sE2 vaccinated-serum likely resulted from synergistic effect of polyclonal NAbs induced by the three sE2 antigens.
Next, we investigated which constituents of the trivalent sE2-immunised sera may contribute to the synergism against HCVcc strains JFH1 and PR63cc. To address this question, three different monovalent sE2-immunised sera were mixed two-by-two or three-by-three and analysed for their neutralising activity. As shown in figure 4, in the case of neutralisation against JFH1, all three two-serum mixtures (H77 +Con1, H77 +S52 and Con1 +S52) had higher neutralising activity than the monovalent sE2-immunised sera, but still lower than the trivalent sE2-immunised sera and the three-serum mixture (H77 +Con1+S52), suggesting that each monovalent sE2 induced unique sets of neutralising antibodies that act synergistically to inhibit JFH1 infection. On the contrary, in the case of neutralisation against PR63cc, the H77 +S52 mixture had a better neutralising activity than the other two-serum mixtures (H77 +Con1 and Con1 +S52) and the three-serum mixture (H77 +Con1+S52), suggesting that the synergism against PR63cc was mainly conferred by the antibodies induced by sE2 from H77 and S52. Altogether, our results demonstrated that the trivalent sE2 vaccine induced polyclonal neutralising antibodies which may act synergistically to inhibit HCV infection.
Analysis of key constitutes in the trivalent vaccine that contribute to the synergism. (A) Neutralisation curve of vaccinated sera (monovalent, trivalent, two-serum and three-serum mixture). Sera of six mice within each group were pooled for the neutralisation assay. The sera of the three monovalent sE2-vaccinated groups were mixed two-by-two or three-by-three at an equal ratio (H77 +Con1, H77 +S52, Con1 +S52, H77 +Con1+S52). All sera were diluted two times serially from 20 to 10 240 to test neutralisation against HCV strains JFH1 and PR63cc. Neutralisation curves were fitted by non-linear regression. (B) ID50 of the sera were calculated by GraphPad Prism. ID50, 50% inhibitory dilution; sE2, soluble E2.
More potent and uniform bNAbs response induced by trivalent vaccine in rhesus macaques
We next evaluated the immunogenicity of the trivalent sE2 vaccine in non-human primates (NHP). Two groups of five rhesus macaques were immunised with 200 µg of the trivalent sE2 (monkeys T1–T5) and monovalent Con1sE2 (monkeys M1–M5), respectively, complemented with AS04, a commercial adjuvant consisting of aluminium and MPL,35 which we previously showed to be the best adjuvant of sE2 vaccine to induce broadly neutralising antibody.28 As a control, two macaques were injected with adjuvant alone (monkeys A1–A2). Blood samples were collected to monitor the antibody titres (figure 5A). In general, sE2-specific antibody titre increased from the third to fifth immunisation, peaked at month 5, then decreased until the final immunisation was conducted at month 7. Overall, the trivalent and monovalent sE2 induced comparable levels of E2-specific antibody titres.
Antibody response induced by the trivalent vaccine in rhesus macaque. (A) The E2-specific antibody titres in immunised monkey sera. Rhesus macaques (n=5 for monovalent and trivalent group, n=2 for adjuvant group) were immunised intramuscularly at month 0, 1, 2, 3, 4 and 7 (indicated by arrow heads below the x-axis). The E2-specific antibody titres at month 0, 3, 4, 5, 7 and 8 were measured by ELISA and expressed as mean±SEM for each animal. (B) Box and whiskers plot analysis of IC50 of the immunised monkey sera. IgG purified from the month 8 sera were diluted two times serially from 2000 to 62.5 µg/mL for neutralisation test. IC50, defined as the concentration of IgG able to neutralise 50% of HCVcc infectivity, was plotted as box (25th–75th) and whiskers (minimum to maximum) by GraphPad Prism. Each dot represents an IC50 value, and the median is indicated as a line. The dotted line corresponds the maximum dilution which is the detection limit. (C) Coefficient of variation of IC50 values. CV% was calculated from IC50 of five monkeys in each group by GraphPad Prism. Each symbol represents CV% of IC50 against one HCVcc. Symbols against the same HCVcc were paired. CV%, coefficient of variation; HCVcc, cell culture-derived virion; IC50, 50% inhibitory concentration; sE2, soluble E2.
Next, we determined the neutralising activity of immunised macaque sera of month 8 against a panel of HCVcc strains. As shown in online supplementary figure S1, the anti-trivalent sE2 sera neutralised JFH1 (GT2a), J8 (GT2b) and S52 (GT3a) more efficiently (P<0.05) than the anti-monovalent sE2 sera. To more quantitatively analyse the neutralising activity of the both groups, we purified IgG from the sera with protein G affinity resin (see online supplementary figure S2) and determined the half maximal inhibitory concentration (IC50). The IgG from monovalent and trivalent sE2-vaccinated groups exhibited dose-dependent-neutralising activities against HCVcc (see online supplementary figure S3A). Intriguingly, the anti-trivalent sE2 IgG neutralised H77, JFH1, PR63cc, S52, SA13 and HK6a more efficiently (P<0.05) than the anti-monovalent IgG (figure 5B). Of note, the IC50 values of the trivalent group against most of HCVcc strains were less than or equal to 500 µg/mL, two to fivefold lower than that of monovalent group (figure 5B and online supplementary figure S3B). Remarkably, the coefficients of variation of IC50 measurements which evaluate the extent of variability relative to the mean from the trivalent sE2-vaccinated IgG were significantly smaller (P<0.05) than those of the Con1sE2-vaccinated IgG (figure 5C), suggesting that the trivalent sE2 vaccine elicit stronger and broader and more equipotent and uniform NAb response in outbred NHP with different genetic backgrounds.
E2-specific T cell response elicited by sE2 vaccine in rhesus macaques
Finally, we analysed T cell response in the vaccinated macaques. Splenocytes collected at month 8 were stimulated with sE2 of H77, Con1 or S52 and measured for the levels of IFN-γ and IL-4. As shown in figure 6A, splenocytes of the monovalent and trivalent groups produced IFN-γ on the stimulation of each of the three sE2 proteins (P<0.05), The IL-4 T cell response of vaccinated macaques was generally weaker. The splenocytes from both monovalent and trivalent vaccinated groups stimulated with Con1sE2 secreted similar levels of IL-4 (P>0.05) (figure 6B). However, IL-4 was induced by the stimulation with H77sE2 or S52sE2 only in the trivalent groups (P<0.05) but not in monovalent group (P=0.0984 and P=0.0631, respectively). Next, we analysed the CD4 and CD8 T cell responses using intracellular IL-4, IFN-γ, tumour necrosis factor alpha (TNF-α) and IL-2 cytokine staining assays. As shown in online supplementary figure S4A–F, both monovalent and trivalent vaccines elicited TNF-α and IL-2-producing CD4 T cell response on the stimulation of Con1sE2 or cocktail sE2, while IFN-γ-producing CD4 T cell response was only observed in the trivalent vaccine group stimulated with the cocktail sE2 (P<0.05). On the contrary, the CD8 T cell response was much weaker in the both monovalent and trivalent vaccine groups, and only TNF-α-producing-CD8 T cell response was detected. The vaccine-induced CD4 and CD8 T cell responses were also confirmed by the carboxyfluorescein succinimidyl ester (CFSE)-labelled T cell proliferation assay (data not shown). Altogether, these results demonstrated that trivalent sE2 vaccine induces E2-specific T cell response in rhesus macaques.
E2-specific T cell response in splenocytes of the vaccinated rhesus macaque. (A) IFN-γ and (B) IL-4 ELISPOT of monkey splenocytes. Splenocytes collected at month 8 were analysed by ELISPOT for IFN-γ- and IL-4 production on the sE2 stimulation. E2-specific ELISPOT counts were plotted as box (25th-75th) and whiskers (minimum to maximum) by GraphPad Prism. Each dot represents an ELISPOT count, and the median is indicated as a line. ELISPOT, enzyme-linked immunospot; HCVcc, cell culture-derived virion; IC50, 50% inhibitory concentration; IFN-γ, interferon gamma; IL-4, interleukin 4; sE2, soluble E2.
Discussion
We previously developed a subunit HCV vaccine based on insect cells-expressed recombinant sE2 proteins. The sE2 antigen production is simple and has a high yield (up to 100 mg/L culture supernatants), making it technically possible to explore a multivalent vaccine that consists of sE2 of various genotypes. Here, we developed a trivalent HCV vaccine, containing sE2 from three subgenotypes (GT1a, GT1b and GT3a), that is capable of inducing broad and synergistic-acting polyclonal neutralising antibodies in mice and rhesus macaques. Although the sE2 vaccine-induced neutralising titres may be still low compared with those in chronically infected patients, they could be sufficient to prevent HCV infection in an acute infection setting as our previous study demonstrated that the immunisation of sE2 vaccine protects against heterologous HCV challenge in a genetically humanised mouse model.27
Our analyses have revealed several advantages of the trivalent sE2 vaccine. First, each sE2 component of the trivalent cocktail may elicit unique sets of NAbs that work synergistically to inhibit some HCVcc infection, including Con1, JFH1, PR63cc and S52 (figure 3). Detailed analysis demonstrated that this synergism is conferred by different sE2 combinations. For example, the synergistic neutralisation against JFH1 by the trivalent sE2 vaccine requires all three components, while the H77/S52sE2 combination is sufficient to achieve the synergism in neutralising PR63cc (figure 4). It was reported that combination of NAbs that recognise epitopes located in distinct regions of E2 could inhibit HCV entry.36 This synergistic neutralising effect may occur when NAbs elicited by different sE2 bind to distinct E2 epitopes simultaneously to achieve the maximum blockade of HCV entry. Alternatively, initial binding of antibodies induced by one sE2 component may alter the conformation of E2 on the virions so that originally hidden neutralising epitopes become accessible to NAbs induced by another sE2 component. It is important to point out that the improved neutralising activity by the trivalent sE2 vaccine may also possibly result from increased ability to induce NAbs by the three-antigen formulation, given that the ID50 value of trivalent sE2-immunised sera against JFH1 is slightly higher than that of sera mixture (figure 3B). More studies will be needed to investigate mechanisms underlying this synergistic neutralising effect.
Another advantage of the trivalent sE2 vaccine is that it induces stronger and more uniform neutralising activity in rhesus macaques which have varying genetic background (figure 5C). This reduction of variability in neutralisation is crucial for human vaccine development and may be caused by several factors. One explanation could be that the weaker neutralising activities by the monovalent sE2-vaccinated sera are generally more subject to variation in the neutralisation assay simply due to the poor reproducibility associated with inefficient neutralisation. Another explanation is that the variation of neutralising titres in outbred rhesus macaques can be influenced by the selective antigen presentation by varying major histocompatibility complex (MHC) in different animals. Increasing coverage of T-cell epitopes in a multivalent vaccine will increase the chance of antigen presentation by a given MHC, and thus help reduce variations among different MHC in their abilities to assist B cell activation. Similar observation has been made in a study of simian immunodeficiency lentiviruses (SIVs). Heterologous cocktail peptides derived from SIV envelope protein can overcome restricted single peptide presentation by MHC in four different inbred mouse strains.37
A potential problem for a multivalent vaccine is interference between individual components. H77 infection was efficiently neutralised by anti-H77sE2 sera, but not so by anti-Con1sE2 or anti-S52sE2 (figures 2C and 3B). Interestingly, the ID50 value of the three-serum mix (261.8) is even lower than one-third of ID50 value of anti-H77sE2 sera (1394) (figure 3B), implying that Con1sE2, S52sE2 or both may induce antibodies that interfere with H77sE2-induced NAb. Cooperation analysis also indicates antagonism in neutralisation against H77 strain by the three-serum mix (CI >1) (figure 3C). However, such interference was not observed with the trivalent sE2 vaccine as the ID50 values of the trivalent sE2-vacccinated sera (week 10, 1218; week 22, 1252) were comparable to those of anti-H77sE2 sera (week 10, 1130; week 22, 1394) (figures 2C and 3B). This suggests that the putative interfering antibodies are not or inadequately induced by Con1sE2 or S52sE2 in the three-antigen formulation, possibly due to antigenic competition during the vaccination process.
Lentivirus-based HCV pseudo-type particles that express HCV E1 and E2 glycoproteins on the virus surface are the first cell culture system to be used for assessing the bNAb-inducing activity of HCV vaccine candidates. The development of HCVcc system has provided a more suitable surrogate system in that the structure arrangement and lipidation of envelope glycoproteins in HCVcc represent more closely to those in authentic virions.38–40 We used a pangenotypic HCVcc panel in our study. This panel contains up to 14 HCVcc strains across all seven genotypes and major subgenotypes, including some prototype strains constructed based on viral consensus sequences41–43 and some strains constructed directly from clinical isolates (PR52B6mt, PR79L9 and PR26C3mt and PR63cc).30 44 To our knowledge, this is one of largest HCVcc panel that has been used for assessing bNAb activity of HCV vaccine candidates. These HCVcc strains displayed distinct patterns of sensitivity to neutralisation. Of 14 HCVcc strains, 6 (Con1, PR79L9, JFH1, PR63cc, J8 and S52) were neutralised more efficiently by the trivalent sE2-immunised mouse sera than 3 monovalent sE2-immunised sera (figure 2C). The neutralisation efficiency against another four HCVcc strains (H77, PR52B6mt, ED43 and HK6a) by the trivalent sE2-immunised sera matched the most efficient monovalent sE2-immunised sera (H77sE2 for strains H77 and PR52B6mt, S52sE2 for strains ED43 and HK6a). Neither the trivalent nor monovalent sE2-immunised sera were efficient in neutralising four strains (PR26c3mt, Jc1, SA13 and QC69). Interestingly, although the infections of H77 and Con1 were neutralised more efficiently by homologous sE2-immunised sera, the neutralisation against S52 was more efficient by heterologous sE2-immunised sera (Con1sE2). It is important to point out that no obvious genotype-specific neutralisation was observed in our analysis. For example, GT1b-derived Con1sE2 induced effective neutralisation against JFH1 (GT2a) and S52 (GT3a), but not against GT1b strains PR26c3mt and PR79L9. Therefore, as many HCV strains as possible should be included to test the bNAb-inducing activity of HCV vaccine candidates in the future.
In summary, we developed a trivalent HCV vaccine containing sE2 from three subgenotypes that displays several advantages over monovalent sE2 vaccine. Future efforts should be invested to further improve potency against HCV strains that remain to be poorly neutralised, such as SA13 (GT5a) and HK6a (GT6a).
Acknowledgments
We thank Dr Jens Bukh (University of Copenhagen) for providing chimeric HCVcc plasmids, Drs Mansun Law and Dennis Burton (The Scripps Research Institute) for providing MAb AR3A, Dr Arvind Patel (University of Glasgow) for providing MAb AP33, Yan Wang and Yanhong Nie (Suzhou Non-human Primate Facility) for the technical assistance in immunisation and blood sampling of macaques, Chao Zhang and Ruihong Zhu (Institut Pasteur of Shanghai) for the technical assistance in flow cytometry analysis.
References
Footnotes
XW and YY contributed equally.
Contributors XW and YY designed/conducted the experiments, acquired/analysed the data and wrote the manuscript. TG conducted the experiments. XY and DZ helped conduct the experiments and provided critical reagents. DL generated critical reagents. QS provided experimental support. JZ and ZH designed the study, analysed the data and wrote the manuscript.
Funding This work was supported by grants from National Natural Science Foundation of China (81330039), Chinese National 973 Program (2015CB554300) to JZ, CAS-SAFEA International Partnership Program for Creative Research Teams to JZ and ZH.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement All data described in the manuscript can be shared on request.