Potential conflict of interest: Nothing to report.
Immune correlates of protection against hepatitis C virus (HCV) infection are not well understood. Here we investigated 2 naive and 6 immunized chimpanzees before and after intravenous challenge, 12 weeks after the last immunization, with 100 50% chimpanzee infectious doses (CID50) of heterologous genotype 1b HCV. Vaccination with recombinant DNA and adenovirus vaccines expressing HCV core, E1E2, and NS3-5 genes induced long-term HCV-specific antibody and T-cell responses and reduced peak viral load about 100 times compared with controls (5.91 ± 0.38 vs. 3.81 ± 0.71 logs, respectively). There was a statistically significant inverse correlation between peak viral loads and envelope glycoprotein 2 (E2)-specific antibody responses at the time of challenge. Interestingly, one vaccinee that had sterilizing immunity against slightly heterologous virus generated the highest level of E2-specific total and neutralizing antibody responses as well as strong NS3/NS5-specific T-cell proliferative responses. The other four vaccinees with low levels of E2-specific antibody had about 44-fold reduced peak viral loads but eventually developed persistent infections. In conclusion, vaccine-induced E2-specific antibody plays an important role in prevention from nonhomologous virus infection and may provide new insight into the development of an effective HCV vaccine. (HEPATOLOGY 2005;42:1429–1436.)
At least 170 million people worldwide are persistently infected with hepatitis C virus (HCV), which is the most common reason for liver transplantation. An estimated 2.3 to 4.7 million people become newly infected every year, but an effective vaccine is not yet available.1, 2 Six different genotypes and a variety of quasispecies of HCV pose a major challenge for the development of an effective HCV vaccine. At present the chimpanzee is the only reliable experimental animal model in which to investigate the early events after HCV infection and to evaluate the efficacy of vaccine candidates. Since HCV-specific T-cell immunity has been known to be important in the control of HCV infection,3–5 a substantial effort has been focused on the induction of vigorous HCV-specific T-cell immunity. Although a neutralizing antibody was considered to be crucial for vaccine-mediated protection against initial virus infection by blunting the infection at an early stage,6 there have been few reports demonstrating the role of the antibody in the prevention of HCV infection. Vaccination with recombinant E1/E2 proteins in chimpanzees, albeit transient, induced strong antibody responses that were responsible for the prevention of a low dose of homologous virus (HCV-1) infection.7 However, when the protected chimpanzees were boosted with E1/E2 proteins and then re-challenged with the closely related but heterologous HCV strain (H77), only poor cross-protection was observed.8 It was also reported that envelope glycoprotein 2 (E2) protein-encoding DNA vaccination of two chimpanzees induced antibody and T-cell responses to E2 protein, leading to resolution of acute homologous monoclonal HCV infection. However, the correlates of protective immunity were difficult to interpret.9 Therefore, it remains to be determined how broadly cross-reactive long-lasting humoral immune responses can be induced, and their correlation with protection against HCV infection.
Here we demonstrate that vaccination by DNA prime-recombinant adenovirus boost induces HCV-specific T-cell responses and a long-lasting E2-specific antibody response that correlates inversely with peak viral loads. In addition, the strong E2-specific total and neutralizing antibody responses as well as nonstructural protein 3 (NS3)/NS5-specific T-cell proliferative responses appeared to be required for inducing complete protection against closely related but heterologous HCV infection.
DNA102 vaccine contains a mixture of pGX10-sΔST (41–729 amino acids [aa]; 462–2528 nucleotides [nt]), pGX10-NS34 (1027–1971 aa; 3420–6254 nt), and pGX10-NS5 (1972–3009 aa; 6255–9371 nt) plasmids. These HCV genes were derived from the Korean genotype HCV 1b strain (gHCV).10 pGX10-sΔST plasmid encodes structural genes, which are devoid of N-terminal 40 aa to abolish the immunosuppresive effect of core protein.10, 11 The murine interleukin 12 (IL-12) gene of pGX10-mIL-12N220L was replaced with the human IL-12N222L gene to generate pGX10-hIL-12N222L.12–14 pGX10-hIL-12N222L was added to the DNA102 vaccine, and the product was named DNA103 vaccine. We used the AdEasy Vector system (Qbiogene, Carlsbad, CA) to generate recombinant adenoviruses (rAd's) expressing HCV antigens, as previously described.15 rAd102 vaccine consists of equal amounts of rAd sΔST (462–2528 nt), rAd NS34 (3420–6254 nt), and rAd NS5 (6255–9371 nt), which were designed to express the same regions of HCV as the individual components of DNA102 vaccine.
Immunization and Challenge of Chimpanzees.
Chimpanzees (Pan troglodytes) were housed in the New York Blood Center's primate laboratory (Vilab II) as previously described.16 They were maintained in social groups in large outdoor enclosures and were initially selected on the basis of nonexposure to HCV without any previous history of HCV viremia or anti-HCV antibody responses. Six chimpanzees were divided into two groups, and each group was immunized intramuscularly with 6 mg DNA102 vaccine or 8 mg DNA103 vaccine in 3 mL or 4 mL into two sites of gluteus maximus and two sites of deltoid muscles (0.75 mL/site for DNA102 vaccine or 1 mL/site for DNA103 vaccine) at weeks 0, 8, and 16. DNA102 and DNA103 vaccines contain 2 mg of each plasmid in 3 mL and 4 mL of 0.15 mol/L sodium phosphate buffer, respectively. The DNA vaccine-primed chimpanzees were boosted intramuscularly into two sites of gluteus maximus (0.4 mL/site) and two sites of deltoid muscles (0.4 mL/site) with 9 × 109 pfu of rAd102 (3 × 109 pfu of each rAd) in 1.6 mL injection buffer (10 mmol/L Tris, 5% sucrose, 2 mmol/L MgCl2, pH 8.0) at week 46. Control chimpanzees were not immunized. Challenge inoculation was done intravenously with 100 50% chimpanzee infectious doses (CID50) of a different HCV 1b strain, HCV-BK, 12 weeks after the last boost. The New York Blood Center's approved IACUC assurance is on file with the Office for Protection from Research Risks (OPRR) of the National Institutes of Health (NIH).
Quantification of HCV RNA.
HCV RNA was quantitatively measured by real time reverse transcriptase–polymerase chain reaction (RT-PCR) with molecular beacon technology as described.17 The method was sensitive to approximately 100 RNA molecules/mL using a synthetic RNA standard, and it gave linear results between 102.5 and 107 RNA molecules/mL. The HCV viral loads were confirmed two or three times, producing consistent results.
Isolation and Propagation of Peripheral Blood Lymphocytes From Chimpanzees.
Peripheral blood lymphocytes (PBLs) removed at indicated intervals were purified by Ficoll-Hypaque and cryopreserved as previously described.16 These were used for the determination of cell-mediated immune responses. Frozen and thawed PBLs were tested in parallel with fresh isolated PBLs to establish the effects of cryopreservation technique. They showed no difference in ELISPOT reactivity (data not shown).
T-Cell Proliferative Responses.
Briefly, 2 × 105 PBLs were plated onto 96-well plates in triplicate, and stimulated with GST-NS3 (1b, 1205–1615 aa) or recombinant NS5 (rNS5, 1b, 2026–2339 aa) protein at 5 μg/mL.18 Phytohemagglutinin (Difco/Becton Dickinson, Sparks, MD) and recombinant superoxide dismutase (kindly provided by M. Houghton, Chiron, Emeryville, CA) were used as positive and negative controls, respectively. After 5 days of stimulation,3H-thymidine containing growth media was added and further incubated for 18 hours. Antigen-specific stimulation indices (SIs) were calculated as antigen-specific thymidine incorporation/thymidine incorporation with superoxide dismutase, and these values were normalized as a percentage of SI stimulated with phytohemagglutinin; (SI with HCV protein / SI with phytohemagglutinin) × 100 (%). The cutoff was defined as greater than or equal to 3.75.
Synthetic HCV Peptides.
A panel of a total of 93 overlapping peptides was synthesized by Peptron Inc. (Daejon, South Korea), to have 20 aa in length, with 10 aa overlap; 12 peptides for core (43–172 aa), 31 for E2 (384–713 aa), 16 for NS3 protease (1029–1217 aa), and 34 for NS3 helicase (1208–1647 aa). The peptide sequence was derived from the gHCV vaccine strain (genotype 1b),10 and used for stimulation in the interferon gamma (IFN-γ) ELISPOT assay.
IFN-γ ELISPOT Assay.
The ELISPOT assay was performed according to the manufacturer's instructions in the IFN-γ ELISPOT kit with modifications (MABTECH, Stockholm, Sweden). Briefly, 3 × 105 PBLs were plated onto 96-well plates in triplicate and were stimulated for 18 hours with the indicated peptide pool (1 μg/mL each peptide). The number of IFN-γ–secreting cells was enumerated using an ELISPOT image analyzer and KS ELISPOT 4.2 software (Axioplan 2 imaging, Zeiss System, Oberkochen, Germany). Antigen-specific IFN-γ–secreting cells were determined by subtraction of IFN-γ–secreting cells on culture with medium alone, since stimulation with an irrelevant synthetic peptide, EDRNNSHSEEQNEKQ, which is designed as a synthetic HIV peptide on the basis of a conformational epitope of HIV gp41, showed a number of spots comparable to that with medium alone. The average number of IFN-γ–secreting cells observed with medium alone was 41 ± 9 (SEM) per well. The number of IFN-γ–secreting cells was represented as the number/106 PBLs.
Serum samples were diluted 1:50 and used for the determination of HCV-specific immunoglobulin G (IgG) responses by standard ELISA technique as described.19, 20 Plates were coated with 50 μL (8 μg/mL) of each protein (core, 1–191 aa); hgh-E1t, 192–383 aa; hgh-E2t, 384–718 aa; GST-NS3, 1205–1615 aa; MBP-NS4, 1637–1771 aa; or NS5, 2026–2339 aa) derived from gHCV strain. For semi-quantitative determination of E2-specific antibody responses, endpoint titration of antibodies was performed with two-fold serial dilutions. Positive cutoff was set to have a higher absorbance than that of preimmune sera plus three standard deviations.
A peptide-based ELISA was developed by using a synthetic peptide representing the E2 hypervariable region 1 (HVR1; 384–413 aa) of the BK strain (genotype 1b). Wells were coated with 200 ng peptide, and the ELISA was performed as described above.
Neutralization Assay Using Retroviral HCVpp.
The HCV pseudovirus particles (HCVpp) were prepared as previously described.21, 22 All procedures were performed in the presence of 5% to 10% fetal bovine serum. Because of nonspecific coagulation of chimpanzee plasma during the assay, total IgGs from the plasma drawn at the time of challenge were purified using rProtein A Agarose Fast Flow IgG purification kit (PEPTRON Inc., Daejon, South Korea). HCVpp were incubated with the purified IgG for 30 minutes at room temperature and then were used for infection of Huh-7 cells. After 8 hours of incubation at 37°C, supernatants were removed and cells were further incubated in Dulbecco's modified Eagle medium/10% fetal bovine serum for 72 hours at 37°C. Green fluorescence-positive cells were quantified by FACS analysis. As a negative control, pseudovirus particles (pp) bearing glycoproteins derived from the feline endogenous retrovirus RD114 were used as described.21 The assays were performed two times independently and produced similar a pattern of results.
The difference in immune responses between groups was determined with nonparametric Mann-Whitney U test. The Pearson's correlation coefficients were computed to arrive at the relationship between HCV viral load (Log10 HCV RNA) and individual immune responses using the SPSS statistical package (SPSS Inc., Chicago, IL). A P value of less than .05 is considered statistically significant.
Results and Discussion
In this study, 6 chimpanzees were divided into two groups (3 chimpanzees per group). Group 1 chimpanzees (nos. 393, 376, and 400) and group 2 chimpanzees (nos. 381, 397, and 402) were immunized three times with DNA102, covering HCV structural and NS3-5 genes of genotype 1b (gHCV) and DNA103 (DNA102 + DNA hIL-12N222L), respectively. These DNA vaccine-primed chimpanzees were then given a booster injection of recombinant adenovirus (rAd102) expressing the same HCV genes as DNA102 (Fig. 1A–B). We previously demonstrated that human IL-12N222L (hIL-12N222L) is a N-glycosylation mutant that selectively reduced secretion of the p40 subunit, a natural antagonist of IL-12, and that codelivery of mouse IL-12N220L (mIL-12N220L), the murine homolog of hIL-12N222L, induced higher HCV E2-specific Th1 and CTL immunity than that of native mIL-12.12 To investigate the relative efficacy of HC102 (DNA102 + rAd102) and HC103 (DNA103 + rAd102) vaccines, 6 vaccinated chimpanzees, together with 2 naive controls, were challenged intravenously with 100 CID50 of HCV-BK 12 weeks after the last immunization (Fig. 1B). The vaccine and challenge strains are both genotype 1b, but HCV-BK differs from the gHCV vaccine strain in about 7% of total amino acids. As expected, 2 control chimpanzees (chimpanzees 404 and 406) had high peak viral loads (PVL) (5.53 logs and 6.25 logs), which became undetectable at 13 and 24 weeks after challenge (WPC), as determined by quantitative real time RT-PCR, respectively. However, the virus transiently reappeared in one of the animals (no. 406) during the follow-up period (Fig. 1C). These results are similar to those observed in 8 chimpanzees independently challenged with the same inoculum (HCV-BK) at the New York Blood Center in terms of PVL (6.06 ± 0.54 logs) and outcome of acute infection; 5 of the 8 challenged chimpanzees naturally resolved the infections up to 70 WPC. The 6 vaccinees had about 2-logs lower average peak acute phase viral loads than those of the two controls (3.81 ± 0.38 vs. 5.91 ± 0.71 logs, P < .001). This is a typical pattern of acute phase viremia found in re-challenged convalescent chimpanzees.23, 24 Interestingly, chimpanzee 393 in group 1 was completely protected from viral infection, and chimpanzee 381 in group 2 resolved the infection after a short period of a low level of viremia (3.28 logs of PVL), which disappeared at 8 WPC. The remaining 4 vaccinated chimpanzees failed to clear the virus in spite of 44-fold lower PVL (4.26 logs ± 0.26) (Fig. 1C) than the controls. All the chimpanzees infected with HCV-BK had asymptomatic infection without ALT elevation (data not shown). It is worth noting that, compared with humans, chimpanzees have a higher rate of viral clearance from HCV infection and do not develop chronic active hepatitis, cirrhosis, or liver cancer.25
To determine immune correlates of viremia control, cross-sectional analysis of T-cell responses was performed by T-cell proliferation and IFN-γ ELISPOT assays (Fig. 2A–B). T-cell responses to at least one HCV protein became detectable after DNA vaccination in all vaccinees, and were significantly enhanced overall by rAd vaccination. Among vaccinees, chimpanzees 393 and 381, who cleared the virus, appeared to maintain vigorous T-cell proliferative responses to NS3 and NS5, determined at the time of challenge, compared with other chimpanzees (376 and 402) that developed persistent infections (P < .05), suggesting an important role of CD4+ memory T-cell responses in viremia control (Figs. 1C,2A). However, there was no direct correlation of the frequency of peripheral IFN-γ ELISPOT responses with control of viremia (Fig. 2B). These results agree in part with previous reports that T-cell proliferative responses to nonstructural proteins contributed to long-lasting, protective immune responses in chimpanzees,23 and that there is no apparent correlation between the frequency of circulating IFN-γ–producing CD4+ or CD8+ T cells and viral clearance.26 It is likely that the frequency of IFN-γ–secreting T cells after stimulation with peptide pools reflects the early (effector) phase of CD4+ and CD8+ T-cell response generation.27
When we examined E2-specific antibody responses using ELISA, the two controls did not show any E2-specific antibody responses until 24 weeks after the virus challenge (data not shown). In this study, E2-specific antibody responses were not induced even after three DNA vaccinations, but they became detectable 2 weeks after the rAd boost. These antibody responses were further enhanced at the time of HCV challenge, 12 weeks after rAd vaccination, indicating induction of long-term E2-specific antibody responses by adenovirus boost in DNA-primed chimpanzees (Fig. 3A). These results contrast with those seen after an envelope glycoprotein vaccination in which E2-specific antibody responses were only transiently induced.7 E2-specific antibody responses were not boosted by virus challenge, but were maintained at least up to 24 WPC, although their levels were slightly decreased over time except in chimpanzee 381, in which the level of E2-specific antibody increased from 4 to 24 WPC (Fig. 3A). It is possible that trace amounts of HCV, below the detection limit, may continuously stimulate E2-specific memory B cells to proliferate and differentiate. It was reported that minute amounts of HCV may remain at least intermittently detectable by sensitive molecular techniques in chimpanzees.28 Interestingly, there was a strong inverse correlation between PVL and E2-specific antibody responses at the time of challenge (r = −0.910, P < .001) (Fig. 3B). When we examined the cross-reactivity of the antibody against the challenging BK strain with peptide-based ELISA, there was also a strong inverse correlation between PVL and BK-specific anti-HVR1 responses at the time of challenge (r = −0.969, P < .0001), indicating that there is cross-reactivity between vaccine strain (gHCV) and challenge strain (HCV-BK) (Fig. 3C). It is worth noting that amino acid sequence identity between the vaccine strain (gHCV) and the HCV challenging inoculum (HCV-BK) is 87% in the E2 region and 48% in HVR1. Chimpanzee 393 with the highest E2-specific antibody titer achieved sterilizing immunity, and chimpanzee 381 with the second level of antibody titer rapidly cleared virus, following transient low PVL. In contrast, 4 other vaccinees (nos. 376, 397, 400, and 402) with reduced PVL failed to clear the virus (Figs. 1C,3A), although they developed detectable E2-specific antibody responses and HCV-specific T-cell responses.
Although E2-specific antibody responses play a key role in the prevention of early infection, it is unlikely that these antibodies are also responsible for clearing the established virus, since control chimpanzees did not develop E2-specific antibodies but did resolve infection. These results indicate that another crucial mechanism is involved in viremia control. Since the innate immune system was reported to be part of the resolution of HCV,29–31 the outcome of HCV infection may be influenced by the frequency of natural killer (NK) and NKT cells as well as the strength of a long-lasting T-cell response. It was reported that uninfected liver contains a significant population of NK and NKT cells and that patients who developed persistent infection have impaired NK cell activity and few intrahepatic NK cells.32, 33 In addition, early NK cell expansion was shown to control viral replication in murine models of viral hepatitis such as lymphocytic choriomeningitis virus (LCMV).34 Thus, liver NK and/or NKT cells may play a key role in the host response against established HCV infection.
To further confirm whether E2-specific antibody responses play an important role in preventing HCV infection, in vitro neutralization assay was performed using two different strains of 1b genotype HCVpp (CG-1B and BK) and 1a genotype HCVpp (H77). Overall, it is likely that the higher E2-specific antibody titer induced, the stronger neutralizing activity against 1b genotype HCVpp observed (r = −0.902, P = .005). In addition, there was an inverse correlation between PVL and in vitro neutralizing activities against 1b HCVpp (r = −0.832, P = .02), indicating an important role for E2-specific antibody response in preventing HCV infection. Specifically, no. 393, which had the strongest E2-specific antibody responses, showed the highest neutralizing activity against all of the tested HCVpp (CG-1B, BK, and H77) (Fig. 4A–B). Taken together, these results suggest that our vaccine can induce neutralizing antibody responses against nonhomologous HCV strains, including challenge strain (HCV-BK). One possible explanation for the induction of cross-reactive antibody responses in this study may be that DNA and rAd vaccines produce proteins with native conformation and thus ensure the integrity of epitopes that stimulate broad neutralizing antibody responses.35 Considering that HCV presents multiple genotypes with a high degree of heterogeneity and circulates as quasispecies within an individual,36 these results are quite encouraging for the development of an effective HCV preventive vaccine.
In contrast to E2-specific antibody responses, antibody responses to other HCV gene products such as core, E1, NS3, NS4, or NS5 do not correlate with PVL, although a significant level of antibody response was detected at the time of challenge (Fig. 3D). Additionally, there was no significant difference between group 1 and group 2 chimpanzees regarding viremia control as well as vaccine-induced antibody and T-cell responses, which indicates that hIL-12N222L DNA vaccine had little adjuvant effect in chimpanzees. These results are in contrast to a previous result that codelivery of mouse IL-12N220L DNA significantly increased HCV E2 DNA vaccine-induced T-cell responses and protection from modified tumor challenge.12 This discrepancy might be caused by the differences in experimental conditions such as vaccination regimen (DNA vs. DNA/rAd) and species (mouse vs. chimpanzee). However, the number of chimpanzees in each group is too small to draw a solid conclusion (Fig. 1C).
In conclusion, DNA prime-recombinant adenovirus booster vaccination can induce long-lasting E2-specific antibody responses, which inversely correlate with PVL after nonhomologous HCV infection, suggesting that cross-reactive antibody responses to HCV envelope could be induced by vaccination. In addition, this study may provide valuable information for designing an effective prophylactic HCV vaccine, which should induce higher long-lasting E2-specific antibody and HCV-specific T-cell proliferative responses than those observed in chimpanzees 393 and 381.
We are grateful to Betsy Brotman, Mussa Konneh, and John Zeonuway for their dedicated assistance in the chimpanzee experiment; Liping Li for technical support in the quantitation of viral loads; and Shibo Jiang for the kind donation of the control peptide for ELISPOT assay.