These authors contributed equally to this study.
Multiepitope peptide-loaded virus-like particles as a vaccine against hepatitis B virus–related hepatocellular carcinoma†
Article first published online: 6 JAN 2009
Copyright © 2009 American Association for the Study of Liver Diseases
Volume 49, Issue 5, pages 1492–1502, May 2009
How to Cite
Ding, F.-X., Wang, F., Lu, Y.-M., Li, K., Wang, K.-H., He, X.-W. and Sun, S.-H. (2009), Multiepitope peptide-loaded virus-like particles as a vaccine against hepatitis B virus–related hepatocellular carcinoma. Hepatology, 49: 1492–1502. doi: 10.1002/hep.22816
Potential conflict of interest: Nothing to report.
- Issue published online: 27 APR 2009
- Article first published online: 6 JAN 2009
- Accepted manuscript online: 6 JAN 2009 12:00AM EST
- Manuscript Accepted: 22 DEC 2008
- Manuscript Received: 10 AUG 2008
- National Natural Science Foundation. Grant Numbers: 30530660, 30671920
- National High-Tech Research and Development Program of Chinaork. Grant Number: 2006AA02A238
- Shanghai Rising-Star Program. Grant Number: 06QA14064
- Eleventh Five-Year Plan. Grant Numbers: 06MA179, 06G67
To develop a hepatitis B virus (HBV) therapeutic vaccine that can induce a broad but specific immune response and significant antitumor effects both in vivo and in vitro, we inserted HBV X protein (HBx)-derived epitopes HBx(52-60), HBx(92-100), and HBx(115-123); a novel subdominant cytolytic T lymphocyte (CTL) epitope HBx(140-148); and the universal T helper epitope pan human leukocyte antigen DR-binding epitope into HBV core protein to form multiepitope peptide-loaded virus-like particles (VLPs). CTL responses against epitope-loaded VLPs were elicited by priming with VLP-pulsed dendritic cells in both HLA-A*0201 transgenic (Tg) mice and peripheral blood lymphocytes from HLA-A2+/HBx+ HBV-infected hepatocellular carcinoma (HCC) patients. The multiepitope peptide-loaded VLPs demonstrated significantly higher immunogenicity in Tg mice than any single responsive epitope. Significant antitumor effects were demonstrated both with primary cultured autologous HCC cells in vitro and tumor-bearing Tg mice in vivo in an HLA-A2–restricted and epitope-specific fashion. Conclusion: The significant antitumor effects both in vivo and in vitro demonstrate the potential of multiepitope peptide-loaded VLPs as a vaccine against HCC. (HEPATOLOGY 2009.)
Hepatitis B virus (HBV) infection may cause acute and chronic hepatitis, which in turn may progress to cirrhosis and hepatocellular carcinoma (HCC).1 The HBV vaccines currently used in immunization have lowered the incidence of HBV infection, but they are ineffective as a treatment for patients with hepatitis and HCC.2 In acute HBV infection, both type 1 CD4+ T helper (Th) lymphocytes and CD8+ cytolytic T lymphocytes (CTLs) that are HBV-specific contribute to effective control of HBV.3, 4 However, chronic HBV infection is rarely resolved by the immune system.5 Thus, broadly specific Th lymphocyte and CTL responses against various HBV antigens would be a new approach to improve the efficacy of a candidate therapeutic vaccine.
HBV X protein (HBx) is a multifunctional regulatory protein that may participate in viral pathogenesis and carcinogenesis.6 In HBx transgenic (Tg) mice, the expression of HBx could increase the tumorigenic effects of hepatocarcinogens two- to three-fold.7, 8 In HCC, HBx protein expression has been demonstrated at a higher ratio than preS2/S, HBV core protein (HBc), and HBV surface protein.9 Because HBx was dominantly expressed in hepatitis and hepatoma tissues, it may be applied as a target for immunotherapy when specific CTL responses could be induced via in vitro or in vivo stimulation.
The universal HLA-restricted Th epitope pan HLA DR-binding epitope (PADRE) has been found to be as much as 1,000 times more powerful than natural T cell epitopes.10 In addition, PADRE has been shown to be safe and well tolerated in human clinical trials.11 Virus-like particles (VLPs) are capable of inducing strong cellular and humoral responses as direct immunogens.12 VLP size appears to be favorable for uptake by dendritic cells (DCs) via macropinocytosis and endocytosis, which play a central role in activating innate and adaptive immune responses.13 Multiepitope-based vaccines represent a powerful approach to overcome immunodominance and simultaneously generate broad immune responses.14, 15 Fusion VLPs carrying multiepitopes may generate more broad immune responses and lead to more powerful antitumor attack by loading effective Th epitopes and CTL epitopes than single VLPs, multiepitopes, or epitopes.
The aim of this study was to develop fusion HBc–VLPs inserted with four dominant HBx-derived CTL epitopes and the universal Th lymphocyte epitope PADRE and to determine if these particles could induce a broad but specific immune response and significant antitumor effects both in vivo and in vitro.
Materials and Methods
Cells and Mice
HLA-A2.1/Kb Tg mice (6 to 8 weeks old) were purchased from Jackson Laboratory (Bar Harbor, ME) and were bred in specific pathogen-free facilities. Transporter associated with antigen processing–deficient T2, human hepatoma cell line SNU-398 (HLA-A2.1+ HBx+), colorectal carcinoma SW480 (HLA-A2.1+ HBx−), mouse melanoma cell line EL-4, and proerythroblastic cell line K562 (HLA-A2− A24−) were obtained from the American Type Culture Collection (Manassas, VA). Peripheral blood lymphocyte dendritic cells (PBL DCs) were purified from HBV-infected HCC patients (HLA-A2+/HBx+) (from the Eastern Hepatobiliary Surgery Hospital, Second Military Medical University, Shanghai, China) by using human CD14 magnetic microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany).
Peptide Selections by Measurement of HLA-A*0201/Peptide Binding, Stability Assay, and Peptide-Specific CTLs
HBc(18-27) and the irrelevant peptide CAP-1 were evaluated as a positive and negative control, respectively, in the peptide selections. Because HLA-A*0201 is one of the most prominent HLA class I alleles, 10 HLA-A*0201–restricted HBx-derived candidate CD8+ T cell epitopes (Table 1) with the highest estimated half-time of dissociation were selected, based on two epitope prediction algorithms16, 17 from HBV FMU013 strain (adw2 type, GenBank accession number: AY206384) and an epidemic strain in China. The peptides were synthesized by GL Biochem and purified to 98% via reverse-phase high-performance liquid chromatography and further confirmed via mass spectrometry. Lyophilized peptides were dissolved in dimethyl sulfoxide and stored in aliquots at −80°C.
|1||HBx(115-123)||CLFKDWEEL||65.841||1.72 ± 0.43||6-8|
|2||HBx(133-141)||VLGGCRHKL||36.316||0.58 ± 0.21||—|
|3||HBx(7-15)||CQLDPARDV||21.395||0.43 ± 0.29||—|
|4||HBx(92-100)||VLHKRTLGL||36.316||1.51 ± 0.35||4-6|
|5||HBx(140-148)||KLVCSPAPC||17.388||1.20 ± 0.26||4-6|
|6||HBx(15-23)||VLCLRPVGA||8.446||1.08 ± 0.46||0-2|
|7||HBx(97-105)||TLGLAAMST||7.452||0.27 ± 0.25||—|
|8||HBx(102-110)||AMSTTDLEA||3.588||1.42 ± 0.31||4-6|
|9||HBx(52-60)||HLSLRGLPV||2.365||1.58 ± 0.35||6-8|
|10||HBx(84-92)||NAHQVLPKV||2.222||1.09 ± 0.29||—|
|Positive||HBV core-derived HBc(18-27)||YLSGANLNL||98.27||1.9 ± 0.43||6-8|
|Negative||Irrelevant peptide CAP-1||SIINFKEL||0||0.1 ± 0.08||0-2|
Transporter associated with antigen processing–deficient, HLA-A2–expressing T2 cells (1 × 106/200 μL) were incubated with 100 μM peptide in serum-free RPMI 1640 supplemented with 5 μg/mL human β2-microglobulin (β2m) (Fluka, Buchs, Germany) at 26°C for 18 hours and then at 37°C for 2 hours. HLA-A*0201 expression was then measured via flow cytometry using the fluorescein isothiocyanate (FITC)-conjugated anti-HLA-A2.1 monoclonal antibody (mAb) BB7.2 (Biosource, Camarillo, CA).
T2 cells (106/mL) were incubated overnight with 100 μM of each peptide in serum-free RPMI 1640 supplemented with 100 ng/mL human β2m at 37°C. Cells were then incubated for 1 hour with 10 mg/mL Brefeldin A (Sigma-Aldrich, Lyon, France) to block cell surface expression of newly synthesized HLA-A*0201 molecules, and incubated at 37°C for 0, 2, 4, 6, or 8 hours. Cells were subsequently stained with anti-HLA-A2.1 mAb BB7.2.
Construction and Production of VLPs (HBc(1-78)-Multiepitope-HBc(79-144) and HBc(1-144)-Multiepitope)
A complementary DNA fragment encompassing N-terminal 144 amino acids of the core protein was obtained via polymerase chain reaction from HBV template DNA with primers 5′-CGGGATCCCATGAGCACGAATCCTAAACC-3′ and 5 ′-GGAATTCTTACAGGAGCCATCCTGCCCA-3 ′, then cloned into pBAD-A. The resulting plasmid, pBAD-HBc(1-144), was used to direct the expression of the truncated HBV core protein HBc(1-144).
The multiple epitopes and Th epitope were determined by computer-assisted analysis, in order that the possible epitope combinations formed the fewest junctional epitopes for human HLA-A*0201 and mouse H-2 Kb alleles. The minigene was synthesized and purified via high-performance liquid chromatography by Shanghai Genebase Gene-Tech Co., Ltd., then cloned into the pBAD-HBc(1–144) (C-terminus behind aa 144, or c/e1 site at aa 78), and identified by restriction endonuclease digestion and sequencing (Fig. 1).
Cultures of Escherichia coli strain BL21 (DE3) harboring the recombinant plasmids were grown and induced by arabinose (0.02%) for 6 hours. Both fusion protein HBc(1-78)-multiepitope-HBc(79-144) and HBc(1-144)-multiepitope were purified with Ni Sepharose High Performance (Amersham Biosciences, CA). Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and western blotting were performed as described.18 Purified VLPs were analyzed via electron microscopy.
In Vitro Functional Assay of VLP-Loaded DCs
HLA-A2/Kb/HBx(52-60), HBx(92-100), HBx(115-123), HBx(140-148), or HBc(18-27) complexes were synthesized and tetramerized using fusion heavy chains with α1 and α2 domains from the HLA-A2 molecule and the α3 domain from the H-2Kb molecule as described.19 Recombinant A2/Kb and human β2m proteins were expressed and refolded at 4°C in the presence of each peptide. Refolded concentrates were filtered and loaded into a gel filtration column for the separation of monomers. The isolated monomers were then biotinylated overnight using recombinant BirA enzyme and repurified in a second fast protein liquid chromatography (FPLC) run. CTLs elicited by epitope- or VLP-pulsed DCs in HLA-A2.1 Tg mice were stained with phycoerythrin-conjugated tetramers and FITC-labeled anti-mouse CD8 mAb (PharMingen, San Diego, CA) and analyzed via flow cytometry.18–23
Enzyme-Linked Immunosorbent Spot Assay.
Tg mice were immunized with VLP-pulsed DCs. T2 cells (1 or 10μg/mL peptide-pulsed) were used as stimulator cells. Enzyme-linked immunosorbent spot assays were performed with the Immunoassay Kit (BioSource International Inc.). Effector cells (1 × 105) and stimulator cells (1 × 105) were seeded into 96-well polyvinylidene fluoride-backed microplates coated with mAb specific for mouse interferon-γ (IFN-γ). After incubation at 37°C for 16 hours, cells were removed and plates were processed. Only brown spots with fuzzy borders were scored as spot-forming cells.
T2 cells (loaded with 10 μg/mL each peptide at 37°C for 1 hour) were used as target cells to raise CTLs. Cytolytic activity was tested based on the measurement of 51Cr or lactate dehydrogenase (LDH) release using the Cytotoxicity Detection Kit (Roche Applied Science). The percentage lysis was calculated as described.25 HLA-A*0201 blocking of CTL activity was performed by preincubating target cells with 10 μg/mL anti–HLA-A2 mAb (BD Pharmingen, San Diego, CA).24, 25 The epitope specificity of tumor cell lysis was assessed in a cold target inhibition assay. Peptide-loaded or unloaded 51Cr-unlabeled SW480 cells were used as cold inhibitors to block lysis of 51Cr-labeled target cells at a ratio of 50:1 (inhibitor/target ratio).26
Antitumor Effects in Mice Produced by Adoptive or Active Immunization with Multiepitope-Loaded VLPs
Adoptive Transfer of Splenocytes from Immunized HLA-A2.1/Kb Tg Mice to C57BL/6nu/nu Mice Bearing Human HCC.
Each group of HLA-A2.1/Kb Tg mice were immunized with VLP- or epitope-pulsed DCs. Splenocytes from each group of immunized Tg mice were restimulated in vitro with 10 μg/mL VLPs or epitope for an additional 6 days. C57BL/6nu/nu mice were inoculated subcutaneously with 5 × 106 SNU-398 tumor cells in the left flank area, and 5 days later were injected intravenously with 1 × 108 stimulated splenocytes derived from immunized HLA-A2.1/Kb Tg mice. The adoptive transfer was performed twice in a 1-week interval followed by intraperitoneal injection of 2,000 IU/mouse interleukin (IL)-2 every 2 days. Control mice received splenocytes from HLA-A2.1/Kb Tg mice immunized with unpulsed DCs or IL-2 only.
EL-4–Based Stable Transfectants.
pIRES (BD Biosciences Clontech, Palo Alto, CA) is a mammamlian expression vector that allows the overexpression of two genes of interest.27 Full-length sequence of HLA-A2/Kb or HBx was cloned into pIRES, which resulted in pIRES–HLA-A2 (HLA-A2/Kb in site A), pIRES-HBx (HBx in site B), or plasmid pIRES–HLA-A2–HBx (both HLA-A2/Kb in A and HBx in B). EL-4 was transfected with the recombinant plasmids above according to the manufacturer's instructions (Lipofectamine Plus Reagent, Life Technologies, Grand Island, NY) and selected in the continuous presence of 600 μg/mL G418 (GIBCO BRL). G418-resistant clones were screened for HLA-A2 molecule expression via flow cytometry or X protein expression via western blotting following standard procedures.28
In Vivo Tumor Protection.
Five mice in each group were immunized intraperitoneally with VLPs or HBc in the absence of adjuvant (50 μg/mouse on day 0, 14, and 28). Three weeks after the last immunization, individual groups of mice were challenged subcutaneously in the left lateral flank with 2 × 106 EL-4 cells transfected with empty pIRES vector, pIRES–HLA-A2, pIRES-HBx, or pIRES–HLA-A2/Kb–HBx. Tumor volume (mm3) was measured and calculated on every third day using the formula [length (mm) × shortest width (mm) × longest width (mm)], in which length was the longest axis.
Generation of CTLs in Patients
PBL-derived DCs were purified from PBLs of patients using human CD14 magnetic microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) and cultured in the presence of IL-4 (1,000 U/mL) and GM-CSF (800 U/mL). On day 5, DCs were pulsed with 20 μg/mL VLPs overnight for 18 hours at 37°C as described.29, 30 IL-6 (1,000 U/mL), IL-1α (10 ng/mL), tumor necrosis factor α (10 ng/mL), and prostaglandin E2 (1 μg/mL) were added on days 6 through 8 for DC maturation.
PBLs from HCC patients were stimulated with VLP-pulsed irradiated autologous DCs twice in the presence of human IL-7, IL-10, and IL-2. Non–peptide-pulsed, irradiated autologous DCs were also included as a negative control. On day 14 after the last stimulation, CD8+ T lymphocytes were purified form PBLs using human CD8 magnetic microbeads (Miltenyi Biotec) and were assayed for their cytotoxicity effects on autologous HCC cells.
Primary HCC cell culture was performed as described.31 Solid masses from HCC patients were manually minced with sterile surgical scissors into tiny pieces, then digested with 0.8% collagenase I and 0.002% DNase I. Cells were collected and their viability was determined via the trypan blue exclusion method. Polypropylene plates were used instead of polystyrene plates to avoid attachment and growth of fibroblasts. The cytotoxicity assay was obtained as described in the Cytotoxicity Assay section above.
VLPs were assayed for general and specific toxicity in outbred white ICR male mice.26 Before and during the experiment, the animals were kept under standard vivarium and natural lighting conditions and on a balanced diet. Animals were given 10-fold injections of VLPs in doses of 100 μg per 20 g of body weight. Control mice received equivalent volumes of saline. Each experimental group consisted of 20 mice.
HLA-A*0201 Peptide-Binding and Complex Stability Assay and Peptide-Specific CTLs.
As shown in Table 1, HBx(15-23), HBx(52-60), HBx(84-92), HBx(92-100), HBx(102-110), HBx(115-123), and HBx(140-148) were shown to form more stable complexes with HLA-A2 (DC50 ranging from 4 to 8 hours) than the others; and HBx(15-23) and HBx(84-92) were excluded from the additional study because of their low stability (DC50 0-2 hours). In enzyme-linked immunosorbent spot and cytotoxicity assays, HBx(115-123), HBx(92-100), HBx(140-148), or HBx(52-60) epitope-pulsed DCs could elicit a CTL response to T2 cells loaded with the respective epitope peptide in a dose-dependent fashion, with significant IFN-γ production, but not to T2 cells alone or those loaded with the irrelevant peptide CAP-1 (Fig. 2A). The CTLs could also efficiently lyse epitope-pulsed T2 target cells, but did not lyse T2 cells alone or CAP-1–pulsed T2 cells at any effector/target ratio (Fig. 2B). No specific reactivity could be detected in CTLs generated from Tg mice immunized with HBx(102-110) epitope-pulsed DCs.
Based on the above results, HBx(52-60), HBx(92-100), HBx(115-123), and HBx(140-148)were finally selected and loaded onto HBc with PADRE to form fusion VLPs.
Expression and Characterization of VLPs.
As shown in Fig. 2C, lanes 2 and 3, a distinguishable extra band (≈30 kDa) was observed that corresponded to the calculated value of fusion proteins. The purity of both proteins was greater than 95% after a single procedure of affinity chromatography (Fig. 2C, lanes 4 and 5). Western blotting indicated that these bands could be detected by HBc mAb (Fig. 2C, lanes 6 and 7). HBc(1-144)-multiepitope did not form core-like particles, as demonstrated by negative staining under electron microscopy, while another fusion protein, HBc(1-78)-multiepitope-HBc(79-144), readily assembled to high and comparable particle concentrations (Fig. 2D). HBc(1-78)-multiepitope-HBc(79-144) was designated as a VLP for further studies.
In Vitro Functional Assay of VLP-Loaded DCs.
In enzyme-linked immunosorbent spot assay, the average number of IFN-γ–secreting splenocytes was significantly higher than that in mice vaccinated with epitope-pulsed DCs, in response to the stimulation of each epitope (P < 0.01 at a dose of 1 or 10 μg/mL). The results of the cytotoxicity assay demonstrated that VLP-pulsed DCs could simultaneously elicit a broad repertoire of epitope-specific CTL responses, which were also markedly stronger (Fig. 2E,F) than that elicited by epitope-pulsed DCs (Fig. 2A,B), especially for responses to HBx(92-100) and HBx(140-148) (P < 0.01 or P < 0.05 at any effector/target ratio).
Epitope-specific CD8+ T cells were able to bind HBx(52-60), HBx(92-100), HBx(115-123), HBx(140-148), or HBc(18-27)-tetramer, which comprised 2.0% to 4.6% of the CD8+ population in splenocytes, after immunization with each group of epitope-pulsed DCs (Fig. 3). Immunization with VLP-pulsed DCs simultaneously elicited a broad spectrum of specific CD8+ T cells that bound to each tetramer, with a higher frequency ranging from 6.2% to 9.8%. The staining was epitope-specific because CD8+ T cells induced by epitope- or VLP-pulsed DCs showed significant binding to the irrelevant peptide CAP-1/tetramer (0.03% or 0.02%).
Antitumor Effects in Mice Produced by Adoptive Transfer or Active Immunization with VLPs.
As indicated in Fig. 4A, adoptive transfer of splenocytes from HLA-A2.1/Kb Tg mice immunized with HBx(52-60) or HBx(115-123)-pulsed DCs was able to significantly inhibit SNU-398 growth in nude mice (P < 0.05 versus control groups receiving IL-2 alone or phosphate-buffered saline (PBS)-pulsed DCs on days 17 to 27 postinoculation) (Fig. 6A). More potent inhibitory effects were observed in the group of mice receiving the splenocytes from Tg mice immunized with VLP-pulsed DCs (P < 0.01 versus any other group on days 17 to 27 postinoculation) (Fig. 4A). The group that received splenocytes from Tg mice immunized with HBx(92-100) or HBx(140-148)-pulsed DCs showed no significant difference in tumor growth from the control groups treated with low-dose IL-2 alone or splenocytes from Tg mice immunized with PBS-pulsed DCs (P > 0.05) (Fig. 4A). The results indicated that at least HBx(52-60) and HBx(115-123) among our four HBx-derived candidate CTL epitopes could be naturally processed and presented by human hepatoma cell line SNU-398. It was also suggested that the stronger inhibitory effects were related to augmented activity and increased frequency of epitope-specific CD8+ T cells elicited by VLPs.
The pIRES–HLA-A2/Kb–HBx–transfected EL-4 cell line (pIRES–HLA-A2/Kb–HBx–EL-4) was established and found to stably express both HLA-A2/Kb and X protein (Fig. 4B,C). CTLs were generated in Tg mice via immunization with HBx(52-60), HBx(92-100), HBx(115-123), or HBx(140-148)-pulsed DCs and tested for the ability to recognize and lyse pIRES–HLA-A2/Kb–HBx–EL-4 in vitro. CTLs elicted by HBx(52-60), HBx(115-123), or HBx(140-148), but not HBx(92-100)-pulsed DCs, were capable of killing pIRES–HLA-A2/Kb–HBx–EL-4 (Fig. 5). However, no lysis of pIRES–HLA-A2/Kb–EL-4 (HLA-A2.1+, HBx−) and pIRES–HBx–EL-4 cells (HLA-A2.1−, HBx+) was observed (Fig. 5). The results indicated that HLA-A2.1–restricted HBx(52-60), HBx(115-123), and HBx(140-148) could be naturally processed and presented from pIRES–HLA-A2/Kb–HBx–EL-4. Therfore, we transplanted pIRES–HLA-A2/Kb–HBx–EL-4 in Tg mice as a novel carcinoma model for evaluation of the antitumor effects produced by immunization with VLPs.
After three immunizations with our VLPs, we observed a significant inhibitory effect on tumor formation and growth in Tg mice for up to 30 days, following challenge with the transplanted pIRES–HLA-A2/Kb–HBx–EL-4 (P < 0.01 versus other groups on days 15 to 30 after tumor challenge) (Fig. 6A). No antitumor effects were observed against the transplanted tumor in the mice vaccinated with the VLP backbone, HBc particles (an irrelevant protein to the target HBx) at the same dosage (P > 0.05 versus the group injected with PBS) (Fig. 6A). Immunization with VLPs did not elicit any inhibitory effects against the transplanted pIRES–HLA-A2–EL-4 or pIRES–HBx–EL-4, which stably expressed HLA-A2 or HBx alone, which had analogical tumor growth rate to the transplanted pIRES-transfected EL-4 cells (Fig. 6B). Thus, all the results indicate that antitumor effects were elicited by the multiepitope-loaded VLPs in an HLA-A2–restricted and antigen-specific fashion, and were not caused by the different immunogenicity of transplanted transfectants.
Cytotoxic Effects on Autologous Tumor Cells in an HLA-A2–Restricted and Epitope-Specific Fashion.
Only the tumor specimens from six patients (1, 4, 10, 11, 13, and 16) grew sufficient HCC cells in primary culture for cytotoxicity assay, because some patients were undergoing chemotherapy or radiotherapy. HBC and HBx expression were detected via immunohistochemical staining in the tumor specimens from these six patients (HBx+/HBc−, patients 1, 10, and 13; HBx+/HBc+, patients 4, 11, and 16). As illustrated in Fig. 6C, the primary cultured HCC cells from the six patients were highly susceptible to lysis by autologous CTLs elicited by VLP-pulsed DCs (P < 0.01 versus CD8+ primed with nonpulsed DCs at 25:1 or 12.5:1 effector/target ratio). No cytotoxic effect was observed against natural killer–sensitive K562 cells, which indicated that the cytotoxic activity was not mediated by nonspecific natural killer cell activity.
The lytic activities of CTLs from the six patients were significantly inhibited in the presence of anti–HLA-A2 mAb (P < 0.01 versus in absence of mAb at 25:1 or 12.5:1 effector/target ratio) (Fig. 6C). No cytotoxic activity was observed in primary cultured HCC cells (HBx+/HBc+) from an HLA-A3+ patient at any ratio of effector to target cells. The cytotoxic activity of CTLs elicited from patients 11, 13, and 16 was significantly inhibited by SW480 loaded with different repertoires of the CTL epitopes (P < 0.01 or 0.05 versus unloaded SW480) (Fig. 6D). These results indicated that HBx(52-60), HBx(92-100), HBx(115-123), HBx(140-148), and HBc(18-27) on our VLPs were naturally presented and processed in HBx- or HBc-expressing HCC of three patients, and the functional CTLs elicited were epitope-specific. Taken together, these results also indicate that CTL responses were elicited by the VLPs in an HLA-A2–restricted and epitope-specific fashion.
VLPs at the dose used had no effect on the total protein concentration, hemoglobin concentration, platelet count, alanine aminotransferase, alkaline phosphatase, or urea concentration. In the experimental animals, a slight increase in the erythrocyte and leukocyte count was observed 14 days after injection (Table 2). The examination revealed that the experimental mice had no more dystrophic changes than controls. The effects of the vaccine were not dose-dependent; therefore, it is likely they were nonspecific.
|Index||Control Group||After Injection for 14 Days||Changes||P Value|
|Total protein (g/L)||47.00 ± 18.23||57.00 ± 6.71||—||0.18|
|Hemoglobin (g/L)||166.40 ± 5.27||162.20 ± 6.06||—||0.19|
|Erythrocytes (1012/L)||9.18 ± 0.31||10.73 ± 0.39||0.001|
|Leukocytes (109/L)||8.06 ± 0.35||12.58 ± 2.13||0.005|
|Platelet count (109/L)||949.60 ± 28.20||962.40 ± 29.70||—||0.17|
|Alanine aminotransferase (U/L)||4.60 ± 1.52||5.00 ± 1.58||—||0.37|
|Alkaline phosphatase (U/L)||24.00 ± 10.07||18.20 ± 1.10||—||0.15|
|Urea (mmol/L)||1.64 ± 0.63||2.44 ± 0.46||—||0.056|
Chronic hepatitis B is closely associated with HCC.32 Among the four proteins that originate from the HBV genome, HBx is a multifunctional regulatory protein participating in viral pathogenesis and in HBV-mediated hepatocarcinogenesis.33, 34 HBx is produced very early after infection and throughout chronic infection. In most in vitro and in vivo studies, HBx exhibited an oncogenic effect and can induce HCC in certain lines of transgenic mice.7
Noninfectious VLPs can be generated by heterologous expression of viral structural proteins and their spontaneous self-assembly.29 The core protein of HBV expressed in bacteria forms shells resembling those in HBV-infected liver cells.30E. coli–expressed core has been demonstrated to assemble icosahedral shells with 180 (T = 3) or 240 (T = 4) subunits.16 Due to its advantageous features, HBV core has been extensively exploited as a carrier for foreign epitopes.29 Most importantly, core particles have been demonstrated to drastically improve the immunogenicity of foreign protein segments presented on their surface.31 Moreover, on the basis of HBc, highly promising vaccine candidates have been generated for influenza35 and malaria.36
Immunodominant responses are not necessarily protective,37 whereas some subdominant responses have been associated with lower viral loads.38 Furthermore, narrow immunodominant CTL responses may select viral escape mutants, which suggest that an optimal vaccine should elicit broader CTL responses. Therefore, understanding factors contributing to HBV epitope hierarchy is necessary for rational vaccine design, because it may be required to enhance protective CTL responses and avoid stimulation of nonprotective responses. HBx(52-60), HBx(92-100), and HBx(115-123) were potent HBV antigens for developing immunotherapy against chronic hepatitis and HCC, and they strongly induced specific CTLs in blood taken from patients with chronic HBV.24 HBx(140-148) was a novel subdominant CTL epitope that can be used as a potent antigenic form to induce antitumor or antiviral immune responses in therapeutic approaches such as DNA vaccine, therapeutic proteins including synthetic polypeptides, or DC therapy. Our results show that VLPs correlate with a more efficient endogenous processing and presentation of the immunodominant epitope.
It was noted that each candidate epitope loaded on VLPs elicited a higher frequency of epitope-specific CD8+ T cells and a significantly stronger response in mice compared with a responding single peptide. The close association between the stronger CTL activities with the higher frequencies of epitope-specific CD8+ T cells indicated that the multiepitope-loaded VLPs could augment the immunogenicity of each epitope by eliciting an increased number of specific CTLs in vivo. The immunopotentiation effects of VLPs were demonstrated in other HBc-derived chimeric particles carrying exogenous sequence from Puumala hantavirus or malaria with significantly higher titer of specific antibody and greater protective efficiency.39 The immunoenhancement effects could be ascribed to Th epitopes in the carrier core protein31 and the polymeric or particulate nature of VLPs, which could be taken up rapidly and processed efficiently by antigen-presenting cells.40
The presence of junctional epitopes may elicit an undesired immunodominance effect, completely silencing the recognition of the desired epitopes.41 Great efforts have been made to avoid junctional epitopes by adopting suitable flanking spacers. The compilation of a database of 94 epitope/flanking region combinations has revealed that the presence of asparagine and glycine immediately after the C terminus of the epitope is most frequently associated with optimal CTL responses.41 Such amino acids were adopted as spacers in our study and showed significant heterogenicity in the immunogenicity of each epitope, without evident junctional epitopes. It has also been reported that a GPGPG spacer can disrupt the junctional epitope and finally restore the immunogenicity against all epitopes encompassed within a linear polypeptide.42 These findings have important practical applications in the design of multiepitope vaccines.
In the present study, immunization with VLP-pulsed DCs elicited specific CTL responses to each epitope loaded both in vivo and in vitro. In light of the broad repertoire and strong magnitude of elicited CTL responses and significant antitumor effects in vitro and in vivo, our VLPs appear to be promising candidates for HBV-related HCC. Other HCC-related epitopes such as MAGE-1(278-286),43 MAGE-3(271-279),22 AFP(158-166), and AFP(542-550)21 may be adopted to formulate more potential multiepitope vaccines. The results also have relevance to the process of designing and evaluating multiepitope-based vaccines aimed at eliciting CTL responses for prophylactic and therapeutic applications.