Evaluation of the immunogenicity of liposome encapsulated HVR1 and NS3 regions of genotype 3 HCV, either singly or in combination

Background Hepatitis C virus displays a high rate of mutation and exists as a quasispecies in infected patients. In the absence of an effective universal vaccine, genotype-specific vaccine development represents an alternative. We have attempted to develop a genotype 3 based, liposome encapsulated HCV vaccine with hypervariable region-1 (HVR1) and non-structural region-3 (NS3) components. Results HCV RNA extracted from serum samples of 49 chronically infected patients was PCR amplified to obtain HVR1 region. These amplified products were cloned to obtain 20 clones per sample in order to identify the quasispecies pattern. The HVR1 consensus sequence, along with three variants was reverse transcribed to obtain peptides. The peptides were checked for immunoreactivity individually, as a pool or as a single peptide tetramer interspersed with four glycine residues. Anti-HCV positivity varied from 42.6% (tetramer) to 92.2% (variant-4) when 115 anti-HCV positive sera representing genotypes 1, 3, 4 and 6 were screened. All the 95 anti-HCV negatives were scored negative by all antigens. Mice were immunized with different liposome encapsulated or Al(OH)3 adjuvanted formulations of HVR1 variants and recombinant NS3 protein, and monitored for anti-HVR1 and anti-NS3 antibody titres, IgG isotypes and antigen specific cytokine levels. A balanced Th1/Th2 isotyping response with high antibody titres was observed in most of the liposome encapsulated antigen groups. The effect of liposomes and aluminium hydroxide on the expression of immune response genes was studied using Taqman Low Density Array. Both Th1 (IFN-gamma, Il18) and Th2 (Il4) genes were up regulated in the liposome encapsulated HVR1 variant pool-NS3 combination group. In-vitro binding of the virus to anti-HVR1 antibodies was demonstrated. Conclusion The optimum immunogen was identified to be combination of peptides of HVR1 consensus sequence and its variants along with pNS3 encapsulated in liposomes, which could generate both cellular and humoral immune responses in mice deserving further evaluation in a suitable cell culture system/non-human primate model.


Introduction
Hepatitis C virus (HCV), a major causative agent of chronic hepatitis is distributed worldwide with an estimated 170 million carriers. HCV displays a high rate of mutation contributed by both host and viral components [1][2][3][4] and exists in infected patients as a quasispecies, which fluctuate during the course of infection. The extreme sequence variability among different genotypes and the global genotype distribution pattern [5] are major concerns in vaccine development. The observed absence of protection of chimpanzees against reinfection with homologous or heterologous strains is noteworthy [6]. Though the development of an effective vaccine for HCV has been a challenging task, several attempts have been made, a few reaching different clinical trial phases [7]. In the absence of an effective universal vaccine, genotype-specific vaccine development represents an alternate strategy. In India, an estimated 12 to 13 million people are HCV infected [8] with a predominance of genotype 3 (> 70%) [[9], our unpublished data].
The 27 amino acid long N-terminal region of E2 glycoprotein (amino acids 384-410) termed as hypervariable region 1 (HVR1), is the most variable region of the whole HCV polyprotein and contains a neutralizing determinant [10][11][12][13]. Studies have shown that HVR1 can serve as B and T cell epitopes and genetic hypervariability is driven by host's immune pressure. Various studies have proven the direct role of anti-HVR1 antibodies in viral clearance [14][15][16][17]. Therefore, we evaluated the HVR1 region for candidate vaccine development. However, immune response to envelope proteins develops slowly and achieves only modest titres during primary infection and use of an additional HCV protein was considered necessary [18,19]. In this context, non-structural protein-3 (NS3) eliciting strong humoral and cellular immune responses was chosen. Importantly, the T-helper immune response against NS3 is associated with viral clearance and absence of such response leads to viral persistence [20,21]. Hence, NS3 may be effectively used for vaccine development. This study reports our attempts to develop a genotype-3-based HVR1 and NS3 combination vaccine.

HVR1 peptide and DNA preparations Clinical specimens
Serum samples of 49 HCV RNA positive chronic hepatitis C patients genotyped earlier and stored at -80°C were used.
ELISA for the evaluation of reactivity of HVR1 peptides with human sera samples ELISA was performed according to the protocol described earlier [23,24] with the following modifications. Maxisorp Nunc ELISA plates were coated with 1 μg/well peptides and incubated for 37°C for 1 hour followed by overnight incubation at 4°C. Anti-HCV positive and anti-HCV negative human serum samples were added at a dilution of 1:50. A total of 115 anti-HCV positive and 75 anti-HCV negative human sera were tested. The genotype distribution of anti-HCV positive human sera is described in Table 1.

Mice immunizations
This study was approved by the Institutional Ethical Committee for animals and experiments were conducted as per the guidelines. Six-to eight-week-old female Balb/ c mice (8 mice/group) were immunized intramuscularly (100 μl) at 0, 4, and 8 weeks interval with various antigens/combinations and different adjuvants ( Table 2). The immunogens included the four HVR1 variants given individually, or as a pool of all four variants or as a tetramer. Similarly, recombinant NS3 was used in different combinations. Mice were bled before immunization, and at regular intervals of 2 weeks after first dose and sera were stored at -20°C till tested.

ELISA for antibodies and isotyping
Recombinant protein based ELISA was performed for detection and titration of anti-NS3 antibodies as described earlier with 1 μg/well pNS3 as the coating antigen [23,24]. For anti-HVR1 antibodies, a similar  protocol was used with the following modifications.
Coating was done at 37°C for 1 hour followed by overnight incubation at 4°C. Blocking and incubation with mice sera were done at 37°C for 1 hour. Isotyping was done as described previously by using the respective coating antigens [23].

Cytokine assay
Cytokine assay was performed as described previously [26]. Cytokines in culture supernatants were measured using BD Cytometric Bead Array Mouse Th1/Th2 cytokine kit (BD Biosciences, USA).

Taqman low density array
Total RNA extraction and gene expression analysis was done from frozen spleen samples as described previously [27]. TLDA card of the mouse immune panel (Applied Biosystems, UK) was used and run on 7900 HT Fast Real-time PCR system (Applied Biosystems, UK) Relative gene expression values were calculated employing comparative Ct method using Applied Biosystems relative quantification (RQ) manager software v1.2. cDNAs from three phosphate buffered saline immunized mice were considered as calibrators. Hprt1 gene was used as endogenous control. Relative quantification values for each study group were used to calculate mean RQ values. For cluster analysis, relative quantification values were log2 transformed and hierarchically clustered with analysis software (cluster 3.0) [28].
Immune capture RT PCR assay for anti-HVR1 antibodies The assay was performed as described earlier [29]. RNA from the captured virus was extracted using Qiagen viral RNA kit (Qiagen, Hilden, Germany) and diagnostic PCR was performed using 5'NCR [30].

Statistical analysis
Student-t test and Mann Whitney tests were performed. SPSS 11.0 software was used for statistical analyses.

Anti-HVR1 antibody responses in mice immunized with HVR1-variant-pool
In mice immunized with either pool-peptides + DNA, or pool-peptides + pNS3, 100% seroconversion was noted after 2 nd dose. No seroconversion was seen when variantpool was given alone or with Al(OH) 3 . Interestingly, 100% seroconversion was recorded when variant-pool was given along with CadB. Immunization with DNA alone failed to produce antibodies. The anti-HVR1 antibody titres of pHVR1-pool-CadB, pVAX-pHVR-pool-CadB and pHVR1 pool-pNS3-CadB were comparable (p > 0.05). HVR-pool + pNS3 administered with CadB produced higher antibody titres than without adjuvant (p = 0.002) (Figure 2e). Addition of DNA or pNS3 along with peptides did not alter antibody titres (p > 0.05).

Cytokine bead array
For the assessment of the type of immune response (Th1/Th2/balanced) generated by the different formulations, secretion of specific antigen-stimulated cytokines, i.e., TNF (pro-inflammatory), IL-2/IL-4 (Th1) and IFNg/IL-5 (Th2) was determined (Figure 3b). Overall, except for TNF and IFN-g, induction of other molecules was not recorded. For HVR1-variant groups, combination with NS3 did show secretion of IFN-g and TNF. Variants 1 and 4 produced higher levels of TNF than IFNg. However, variant-3-pNS3 induced high levels of IFNg (1348.8 ± 233.8 pg/ml) when compared to TNF (460.9 ± 25.6 pg/ml). IL2 was also detected (30.3 ± 1.9 pg/ml). Addition of corresponding DNA did not influence cytokine stimulation (Figure 3b). In accordance of absence of antibody response, no cytokines were detected for the variant-2 group. On the contrary, despite higher antibody titres, no cytokines were detected in the tetramergroup. Immunization with variant pool + NS3 in the presence/absence of the adjuvant was associated with the no secretion/secretion of TNF respectively. As far the response with respect to pNS3 is concerned, variant-1-NS3, Variant-2-NS3, pool-Ns3-CadB, pool-NS3/ tetramer-NS3 without adjuvant and NS3 with/without adjuvant led to TNF response (Figures 3b). Upon stimulation with pNS3, IFN-g response was noted only with variant-1 (581.9 ± 130 pg/ml) and 2 (1515.5 ± 143.7 pg/ ml).

Taqman low density array (TLDA)
Upon comparing the genes of the TLDA mouse immune panel for control groups immunized with PBS, cadB and aluminium hydroxide with unimmunized control group, no significant up regulation or down regulation was noted. This indicated that the adjuvants alone were not responsible for regulation of the genes.
Overall, it was seen that the immune response was skewed towards Th1 type as suggested by the up regulation of IFN-g, Il18, and Il2 genes. This was further supported by the up regulation of pro-inflammatory TNF and chemokine genes (Ccl19, Ccl2, Ccl3, Ccr2, Ccr4, Ccr7, Cxcl10, Cxcr3). The Th2 genes Il13, Il5 were down regulated. The up regulation of Il4 suggested generation of Th2 immune response as well.
On comparing the groups HVR1 variant pool without adjuvant with HVR1 variant pool given along with cadB or aluminium hydroxide, there was no significant difference in the gene profile (data not shown). This further demonstrates the combined action of NS3 and cadB that led to the immune response pattern in the pool-NS3-cadB group as described above. When Tetramer-NS3 without adjuvant, tetramer-NS3-cadB and tetramer-NS3-aluminium hydroxide groups were calibrated against PBS, the fas gene was up regulated in all three groups. In tetramer-NS3 without adjuvant group, cyp1a (333.3 fold) gene and Il9 (10 fold) were down regulated. In tetramer-pNS3-cadB group cyp1a (1000 fold) and Il9 (142.9 fold) genes were down regulated and fas (2.3 fold) and Il6 (2.4 fold) genes were up regulated (Figure 4c). No significant difference was recorded for other genes. Il9 was consistently down regulated in all three groups (tetramer-NS3: 10 fold, tetramer-NS3-cadB: 142.9 fold, tetramer-NS3-Al(OH) 3 : 1000 fold). Il6 was not significantly altered in the tetramer-NS3 without adjuvant group. However, Il6 expression was marginally up regulated upon addition of either cadB (2.4 fold) or Al(OH) 3 (2.8 fold). Based on the expression pattern of these genes, Th1/Th2 immune response pattern could not be conclusively determined for tetramer-NS3-cadB, tetramer-NS3-Al(OH) 3 or tetramer-NS3 without adjuvant as immunogen groups.

Immune capture-RT PCR
Anti-variant-pool antibodies could bind both genotype 3a and 1b viruses as detected by the immune-capture-RT PCR ( Figure 5).

Discussion
Though the exact mechanisms of protection against HCV infection remain unclear, association of both humoral and cellular immune responses have encouraged evaluation of different approaches for vaccine development. Over the last decade, numerous HCV vaccine approaches have been assessed in mice and primates. So far, majority of the recombinant candidate vaccines have been predominantly generated from genotype 1 virus [31,32]. Many of these vaccine attempts including those reaching clinical trials produced cross-genotypic immune responses [29,31,33]. Still, the fact remains that the vaccine-genotype specific antibodies are mainly produced.
The predominance of genotype 3 in India (> 70%) [9] led us to attempt to develop a genotype 3 based vaccine. The two components of our candidate vaccine were the hyper variable region 1 (HVR1) and non-structural protein-3 (NS3).
The next issue to be addressed was the presence and distribution of HCV quasispecies in the samples studied. A shift over time of the minor quasispecies to the most predominant one has been reported earlier [37]. Understanding the degree of variation and proportion of the different sequences was indeed necessary. For this, the quasispecies pattern was determined for all the 49 patients by sequencing 20 HVR1 clones obtained from every patient. The pattern varied from 2 to 14 quasispecies demonstrating degree of sequence variability in different patients and need to consider these variations for designing appropriate 27 amino acid long HVR1 peptides.
In the HVR1 consensus profile generated during this study, amino acid positions 2, 6, 7, 10, 19, 20, 23, and 26 were found to be more conserved than the rest (91.8% at position 2 to 73% at position 10% conservation). Amino acid positions 11, 12, 14, 22 and 25 were found to be most variable (77.6% at position 14 to 63.3% at position 11). Importantly, except for position 10, all the other relatively more conserved amino acids were also present Gupte and Arankalle Virology Journal 2012, 9:74 http://www.virologyj.com/content/9/1/74 in the consensus sequences reported earlier [33,38]. The preference of amino acids in a given position of HVR1 provides information as to the combinations of permissible mutations in any position of HVR1 [33,38]. This significant observation encouraged us to proceed further and design possible variants of the HVR1 sequence representing genotype 3 and also expect cross-reactivity with the other circulating genotypes.
A total of 4 variants were thought to cover the range of sequence variability. Before assessing the immunogenecity of the 4 peptide variants in mice, we examined immunoreactivity with antibodies from 115 anti-HCV positive patients' sera infected with different genotypes (genoytpes 3, 1, 4 and 6). ELISAs were performed using the peptides individually, as a pool or as a single tetramer where the individual peptide sequences were used in tandem interspersed with 4 glycine residues. Overall, all the variants efficiently detected genotype 3 and a substantial proportion of genotype 1 infected sera ( Table 1). Use of variant pool did not improve ELISA performance, and in fact, resulted in lower reactivity than the individual variants 1, 3 and 4 probably because of the competitive inhibition. The tetramer exhibited lowest reactivity (42.6%) with the anti-HCV positive human serum samples when compared with the peptide pool (72.2%). ELISA results showed that the HVR1 peptides used exhibited a high degree of cross-genotypic reactivity and could be taken forward for the immunogenicity studies. A substantial proportion (26/40) of anti-HCV positives (genotypes 1, 2, 3 and 4) were earlier shown to be non-reactive with a peptide polytope comprising of variants of 11 N-terminal amino acids of HVR1 were used as coating antigens [29]. Our results encouraged us to evaluate the HVR1 consensus and its variants as candidate vaccine. Murine antibody response to HVR1 region was assessed for each variant individually, pool of all the 4 variants or as a tetramer. The variant-2, though immunoreactive, was not immunogenic in mice. This variant differed from the consensus sequence by 5 amino acids (positions 3,8,9,10,11), while the variants 3 and 4 that differed from the consensus by 3 (positions 3,8,9) and 1 (position 9) amino acids respectively were immunogenic. Probably, substitutions at four consecutive positions or substitutions at positions 10 and 11 rendered variant-2 nonimmunogenic. Antibody responses to individual variants 1, 3 and 4 were comparable to the HVR1 variant pool and could be detected by all the variants individually, indicating cross-reactivity in ELISA. The results are in concordance with Mihailova et al [39], where a combination of bacterially expressed HVR1 tetramer, core and NS3 proteins were used as immunogens and high anti-HVR1 tetramer antibody titres and vigorous lymphoproliferative responses were obtained. Similar results were obtained when HVR1 N-terminal polytope [29], and HVR1 DNA [40] were used as immunogens.
An NS3 specific CD4+ T cell immune response is much stronger and more frequently found in patients who resolve acute hepatitis than in patients who develop chronic infection and this response may be necessary for virus clearance [41]. Importantly, several studies have demonstrated the efficacy of the NS3 region as a vaccine candidate either singly or in combination with other HCV proteins [42][43][44][45][46]. In our study 100% antibody seroconversion for anti-NS3 antibodies was seen after second dose. Addition of individual variants, variant pool or tetramer in Cad-B (liposome) yielded optimum and comparable antibody titres in mice. The enhanced anti-NS3 response by the addition of HVR1 is of particular advantage as we intended to use combination of HVR1 and NS3 as vaccine candidates.
As against our earlier observations of enhanced immune response following encapsulation of proteins and corresponding DNAs of hepatitis E and B viruses [25], no advantage was noted with both HVR1 and NS3. Thus, the phenomenon of enhancement of immunogenicity by DNA seems to be immunogen-dependent.
Adjuvanted pNS3 did produce significantly higher antibody titres. The effect with reference to adjuvant could be detected in NS3-HVR1 variant tetramer. 100% seroconversion was noted after third dose for tetramer-pNS3 given either with cadB or aluminium hydroxide, whereas 66.6% seroconversion was observed when NS3 was given with HVR1 variant pool or tetramer without any adjuvant.
On account of the inherent problems of the immunogenecity of peptides and recombinant proteins, use of adjuvants becomes necessary. In view of the crucial role played by various adjuvants in determining immunogenicity, [18,19], we decided to compare, liposomes (Cad-B) with Al(OH) 3 , the time-tested adjuvant. In our lab, liposome was successfully used in the development of recombinant hepatitis E vaccine as judged by experiments in mice and rhesus monkeys [25,47] as well as for a recombinant combination vaccine for hepatitis E and B [25]. Liposome encapsulation did enhance immunogenicity of the HCV antigens [42,48,49]. Though we were using 27-amino acids peptides as immunogens, formulations with Cad-B (HVR1 pool alone, HVR1pool-NS3 and HVR1-pool + corresponding DNAs) yielded high antibody titres as compared to low/no antibodies with Al(OH) 3 . Further, even though the tetramer induced high antibody titres with different formulations, the lower efficiency (42.6%, Table 1) in detecting anti-HCV antibodies in human serum samples argues against its suitability as a vaccine candidate. Thus, liposome encapsulated HVR1 peptides emerged as suitable vaccine candidate.
As our aim was to use HVR1 and NS3 for reasons described earlier, it was important to assess the effect of adjuvants, the individual components and combination of these immunogens on the antibody response. As far as the antibody titres against NS3 component is concerned, addition of individual variants, variant pool or tetramer in Cad-B yielded optimum titres (Figure 2). Addition of pNS3 to the peptides did not alter anti-HVR1 antibody titres significantly (p > 0.05) as compared to the addition of corresponding DNA (Figure 2). This data strongly suggests utility of HVR1-NS3 combination.
Many reports of vaccine development for hepatitis C had used ELISPOT, LPA, intracellular staining and chromium release assays to determine the cell mediated responses against candidate vaccines [29,43,50]. In our experiments, we used cytokine bead array to determine the Th1/Th2 response pattern. We also substantiated this data by serum antibody isotype pattern, an indicator of the type of immune response obtained. Earlier studies reported detection of IFN-gamma (Th1 response) and IL4 (Th2 response) using ELISPOT for NS3 and other non-structural proteins [51,52]. However, we did not observe strong T-cell mediated immune response as judged by the induction of cytokines by antigen-stimulated spleenocytes using CBA. In the cytokine bead array analysis, activation of both Th1 and Th2 cytokines was seen though at very low concentrations. This was probably because mice spleens were harvested at a later stage (15 days after 3 rd dose). The possibility of higher elevation in these cytokines at early time points cannot be ruled out. Incidentally TNF-alpha the pro-inflammatory cytokine was detected at higher concentrations (to a maximum of 583.68 ± 51.34 pg/ml).
Isotyping analysis documented involvement of both Th1 and Th2 type immune responses with NS3 and Th2 with HVR1 component. Variant-1 and 4 with 1 amino acid substitution behaved similarly i.e., Th2 response with 25 μg variant + corresponding DNA switching to balanced Th1/Th2 isotype pattern after the increase of peptide to 50 μg or addition of 1 μg pNS3. The effect of Al(OH) 3 and Cad-B adjuvants on isotype pattern could be studied for variant-pool and tetramer. Al(OH) 3 -variant-pool group remained antibody negative while addition of CadB led to balanced response. With tetramer, Al (OH) 3 gave balanced response while Th2 response was observed with CadB adjuvant. It has been shown earlier that the composition of cationic liposomes resulted in a shift from Th1 to Th2 anti-NS3 antibody response. NS3 DNA given along with DDAB/DOEPC a highly charged cationic liposome combination was seen to induce Th2 response in comparison to DDAB/EPC liposome-NS3 DNA formulation which induced Th1 response. It was further hypothesized by Jiao et al [48] that plasmids adsorbed on muscle cells induced Th2 type of immune response while those captured directly by APCs would induce Th1 type of immune response. Thus the type of response was influenced by the adjuvant used and the site of injection [48].
We further checked the gene expression profile for mice immunized with various antigen formulations either with or without adjuvants using Taqman Low Density Array (TLDA) in order to confirm the results obtained in isotyping analysis.
The adjuvants, when given alone without antigens, did not show significant change in gene profiling when compared with PBS control, thereby indicating that the immune response of the antigen was enhanced/modified because of the adjuvanted antigens. The adjuvants (Al (OH) 3 /CadB) by themselves did not cause alteration in gene regulation.
Both Th1 and Th2 genes were up regulated indicating a balanced immune response and its enhancement upon liposome encapsulation of HVR1 pool-NS3-cadB group. This conclusion was further supported by serum antibody isotypes. The absence of IL5 rise in spleen was accompanied by the down regulation of the gene.
On comparing HVR1 variant pool without adjuvant and HVR1 variant pool given either with CadB or aluminium hydroxide, no significant difference in the gene profile was recorded. Thus the observed gene profile could be attributed to the additive effect of NS3, CadB and HVR1 pool. A similar phenomenon was not observed with the tetramer-NS3 combinations, indicating that individual peptides rather than the tetramer could generate better cytokine response, despite similar antibody titres produced by anti-HVR1 pool and anti-HVR1 tetramer. This further strengthens our conclusion that HVR1 variant pool, and not HVR1 tetramer should be used for further studies. Overall, it was observed that the HVR1 pool-NS3 combination encapsulated in CadB was the most promising vaccine candidate to be studied in depth.
In the absence of an easy, efficient and convenient invitro/in-vivo neutralization test system, we performed the immune capture RT-PCR for the demonstration of binding of the virus to the antibodies generated against the HVR1. It is indeed important to note that these antibodies were able to bind viruses of genotypes 3a and 1b. Whether this binding is indicative of neutralization cannot be ascertained. Similar results were reported with HVR1 based polytope vaccine candidate [29].
In conclusion, the combination of peptides of HVR1 consensus sequence and its variants along with pNS3 encapsulated in liposomes could generate both cellular and humoral immune responses in mice deserving further evaluation in a suitable cell culture system/nonhuman primate model.