Identification of a new HLA-A*0201-restricted CD8⁺ T cell epitope from hepatocellular carcinoma-associated antigen HCA587

Abstract

For the development of peptide-based cancer immunotherapies, we aimed to identify specific HLA-A*0201-restricted CTL epitopes in hepatocellular carcinoma (HCC) associated antigen HCA587, which has been identified as a member of the cancer/testis (CT) antigens highly expressed in HCC. We first combined the use of an HLA-A*0201/peptide binding algorithm and T2 binding assays with the induction of specific CD8⁺ T cell lines from normal donors by in vitro priming with high-affinity peptides, then IFN-γ release and cytotoxicity assays were employed to identify the specific HLA-A*0201 CD8⁺ T cell epitope using peptide-loaded T2 cells or the HCA587 protein⁺ HCC cell line HepG2. In the six candidate synthesized peptides, two peptides showed higher binding ability in T2 binding assays. No. 2 peptide, encompassing amino acid residues FLAKLNNTV (HCA587_317−325), was able to activate a HCA587-specific CD8⁺ T-cell response in human lymphocyte cultures from two normal donors and two HCC patients, and these HCA587-specific CD8⁺ T cells recognized peptide-pulsed T2 cells as well as the HCA587 protein⁺ HCC cell line HepG2 in IFN-γ release and cytotoxicity assays. The results indicate that no. 2 peptide is a new HLA-A*0201-restricted CTL epitope capable of inducing HCA587-specific CTLs. Our data suggest that identification of this new HCA587/HLA-A*0201 peptide FLAKLNNTV may facilitate the design of peptide-based immunotherapies for the treatment of HCA587-bearing HCC patients.

Introduction

Hepatocellular carcinoma (HCC) is one of the most prevalent malignancies worldwide with a global incidence of 1·2 million new cases per year, and the incidence of HCC is rapidly rising in both Asian and western countries due to the global pandemic of hepatitis B and C infections. HCC is often diagnosed late in its development because of a lack of clear symptoms, and the 5-year survival rate is as low as 5%[1]. In view of the poor prognosis for HCC patients with surgery and chemotherapies, an alternative of immunotherapeutic strategies have been pursued in patients in order to improve the treatment [2].

The immune system has long been implicated in controlling tumours. The evidence collected in the last decade indicates that CD8⁺ T cells which recognize peptide fragments derived from tumour-associated antigens (TAA) in a human leucocyte antigen (HLA)-restricted fashion play an important role in this process [3,4]. Such cytotoxic T lymphocytes (CTLs) can directly lyse tumour cells and also secrete cytokines such as interleukin (IL)-2, tumour necrosis factor (TNF), granulocyte macrophage-colony stimulating factor (GM-CSF) and interferon (IFN)-γ, which have indirect anti-tumour effects. In vivo CTL induction in transgenic mice gives rise to responses recognizing tumour cells associated with tumour regression or protection from tumour challenge [5,6]. Adoptive transfer experiments in humans have also demonstrated the efficacy of the CTLs with anti-tumour activity [7]. In addition, human clinical trials have demonstrated that epitope-specific CTLs can be induced in cancer patients and their induction correlated, in multiple instances, with partial or complete tumour responses [8–10].

However, it was not until 1991 that the first report describing the cloning of a gene encoding a human TAA, the melanoma antigen-1 (MAGE-1) was published [11]. The identification of its nonamer peptide, which is recognized by HLA-A1-restricted CTLs, was published in the following year [12]. Identification of TAA peptides expressed by different human tumours [13] provided the basis for antigen-specific active immunotherapy or vaccination and facilitated the design of new vaccination clinical trials. As for HCC, the prerequisite for the success of immunotherapy is to identify the hepatocellular carcinoma associated antigens that are exclusively or preferentially expressed in malignant liver tissues and to prove that these TAA-derived epitopes are potently immunogenic.

With the application of serological analysis of recombinant cDNA libraries (SEREX), several tumour-associated antigens have been identified in HCC [14]. In our previous study, we successfully identified a new CT antigen HCA587 which was highly expressed in human HCC tissues [15]. Because TAA-derived epitopes have been identified by various means [16,17], here we explored the candidate epitopes from HCA587 by their major histocompatibility complex (MHC)-binding motif and MHC class I binding affinity. High-affinity peptides are then tested for in vitro immunogenicity with peripheral blood mononuclear cells (PBMCs) from normal donors and HCC patients, and their ability to induce tumour-reactive CTLs.

In this study, we identify one novel HCA587-derived peptide, FLAKLNNTV (HCA587_317−325), that binds to HLA-A*0201 molecules using PBMCs and GM-CSF- and IL-4-induced autologous dendritic cells (DC), and report the generation of CD8⁺ T cells specific for this peptide that can recognize HCA587 protein⁺ HCC cell line HepG2.

Materials and methods

Cell lines and antibodies

For HLA-A*0201 typing, normal donors, HCC patients and cell lines were prescreened with the HLA-A2-specific antibody BB7·2 (American Type Culture Collection, Manassas, VA, USA) by flow cytometry. HLA-A2-positive cells were subtyped by polymerase chain reaction (PCR) as described previously [18]. The cell lines HepG2 (hepatocellular carcinoma) and T2 (B and T lymphoblast hybrid) were obtained from American Type Culture Collection. Other hepatocellular carcinoma cell lines, HLE, SMMC-7721, Bel-7402, Bel-7405, were purchased from Cell Institute (Shanghai, China). All cell lines were maintained in complete RPMI-1640 [10% heat-inactivated fetal calf serum (FCS), 2 m M glutamine and 100 units/ml penicillin/streptomycin].

Peptides and protein

Peptides, with more than 99% purity, were synthesized using automated solid phase techniques, purified by reversed phase-high performance liquid chromatography (HPLC) and their structures were verified by mass spectrometry. Among the peptides listed in Table 1, the HLA-A*0201 binding peptide GILGFVFTL (influenza matrix protein-derived peptide 58–66) were used as positive control. Peptides were obtained in lyophilized form, dissolved to a final concentration of 1 mg/ml in DMSO and stored at − 70°C. HCA587 recombinant protein was generated successfully and purified from Escherichia coli strain M15 with the more than 95% purity, as described previously [19].

Table 1.

HCA587-derived peptides for HLA-A*0201 binding analysis.

Name	Position	Sequence	Parker score	Tuebingen score	Uohsc score	m.w	FIa
No. 1	228	SLLIIILSV	591·9	31	165	970·3	1·21
No. 2	317	FLAKLNNTV	319·9	27	142	1019·3	3·91
No. 3	140	TLDEKVAEL	″80·6	31	139	1017·3	4·12
No. 4	361	SLSVMSSNV	″69·6	22	133	923·1	1·18
No. 5	188	FMELLFGLAL	″17·3	20	107	1153·6	1·11
No. 6	190	ELLFGLALI	″6·7	24	126	988·3	1·10
Flub	58	GILGFVFTL	550·9	30	107	966·3	3·43

^aFI = (MFI sample – MFI background)/MFI background, where MFI background represents the value without peptide.

^bFlu matrix derived peptide served as a positive control.

T2 binding assays

The binding affinity of each HCA587 peptide and control peptide to HLA-A*0201 molecules was determined as described by Vonderheide et al. [20]. Briefly, T2 cells were incubated overnight with peptide (50 µg/ml) in serum-free medium containing human β₂-microglobulin (β₂-M; 5 µg/ml; Sigma, Deisenhofen, Germany). Cells were washed, stained with BB7·2 and GAM-IgG-FITC (Dianova, Hamburg, Germany), and then analysed on a FACScan flow cytometer (Becton Dickinson, Heidelberg, Germany). The fluorescence index (FI) was calculated using the following formula: FI = [mean fluorescence intensity (MFI) sample-MFI background]/MFI background, where MFI background represents the value without peptide. Samples were tested in duplicate and mean FI was calculated.

Analysis of HCA587 expression in cell lines of HCC

Reverse transcription-polymerase chain reaction (RT-PCR)

RT-PCR was carried out as described previously [15] to detect the expression of HCA587 mRNA in HepG2, SMMC-7721, HLE, Bel-7402 and Bel-7405. Briefly, total RNA was extracted from tumour cell lines, and reverse-transcribed with Avian myoblastosis virus (AMV) reverse transcriptase and oligodT₍₁₈₎ (Clontech, Palo  Alto,  CA,  USA).  Quality  of  the  cDNA  was  confirmed  by PCR for G3PDH. Gene-specific PCR primers used to amplify HCA587 were designed according to our published sequences (HCA587 forward primer: 5′ATCG GATCCCCTCCCGTTCCAGGCGT-3′, HCA587 reverse primer: 5′ACTAAGCTTTCACTCGAAAAGGAGAC3′). RT-PCR was performed with the Advantage Taq (Clontech, Palo Alto, CA, USA) for 30 cycles in a GeneAmp PCR system 9700 (Applied Biosystem, Foster City, CA, USA) at an annealing temperature of 60°C. Products were analysed on 1% agarose gels containing 0·01 mg/ml ethidium bromide.

Immunohistochemical (IHC) staining

IHC staining was performed to detect the HCA587 protein expression in HCA587 mRNA⁺ tumour cells HepG2. HCA587 mRNA^– tumour cells HLE and SMMC-7721 were used as negative controls. Slides containing tumour cells were prepared using a cytocentrifuge and were fixed with acetone. IHC staining was performed using Two-Step Histostaining Reagent (Zymed Laboratories Inc., South San Francisco, CA, USA) according to the manufacturer's instructions. The first antibody to HCA587 was polyclonal antibody generated from purified rabbit serum after four immunizations with recombinant HCA587 protein produced from E. coli. Diaminobenzidine tetrahydrochloride was then used as a chromogen, followed by counterstaining with haematoxylin solution.

Isolation of PBMCs and generation of mature DCs

PBMCs were separated from heparinized venous blood by Ficoll-Hypaque density gradient centrifugation from HLA-A*0201 normal donors and HCC patients. DCs were generated from peripheral blood monocytes as described by Romani et al. [21]. Briefly, PBMCs were seeded into six-well culture plates containing 3 ml RPMI-1640 and 10% FCS at 5–10 × 10⁶/well. Plates were incubated in a 37°C incubator for 2 h, then the non-adherent cells were removed and the adherent cells were cultured at 37°C in RPMI-1640 supplemented with 10% FCS, 1000 U/ml human recombinant GM-CSF and 500 U/ml human recombinant IL-4. All T cell stimulation was with day 7 DCs. As described below, DCs were matured with lipopolysaccharide (LPS) on day 6 and HCA587 antigen was added for effective processing at this time. Peptides do not need the latter, so were added on day 7.

DC pulsing with HCA587 peptide

Day 6 DCs were resuspended in RPMI-1640 with 1000 U/ml GM-CSF, 500 U/ml IL-4 and 10 ng/ml LPS and incubated at 37°C. After 24 h cultivation the DCs were collected, washed and pulsed with 10 µg/ml HCA587 peptide for 3 h at 37°C. The DCs were washed twice for the following stimulation.

DC pulsing with HCA587 protein

The cationic lipid N-[1-(2,3-dioleoyloxy)propyl]N,N,N-trimethylammonium methyl sulphate (DOTAP) transfection reagent (25 µg) (Roche) and E. coli-produced and purified HCA587 protein (10 µg) were mixed at room temperature in polystyrene tubes for 20 min. The DOTAP/protein mixture was added to DCs at 37°C in an incubator with occasional agitation for 3 h. The cells were washed twice, resuspended in RPMI-1640 with 1000 U/ml GM-CSF, 500 U/ml IL-4 and 10 ng/ml LPS and incubated at 37°C. After 24 h cultivation, the antigen-pulsed DCs were harvested and washed twice before being used as antigen-presenting cells (APCs).

In vitro generation of HCA587-specific CTLs

Induction of CTLs with DC pulsed with peptide

CD8⁺ T cells were isolated by positive selection with CD8-Dynal immunomagnetic beads (Dynal PYT Ltd, Carlton South, Australia) according to the manufacturer's instructions. After washing, CD8⁺ T cells were co-cultured with HCA587 peptide-loaded DCs in 2 ml RPMI-1640 medium supplemented with 10% AB serum, recombinant human IL-2 (10 ng/ml) and IL-6 (500 U/ml). Seven days later, the cultured T cells were restimulated with freshly prepared peptide-pulsed DCs and cultured for another 7 days. After four consecutive rounds of stimulation, cultures were tested for the presence of HCA587-specific CTLs.

Induction of CTLs with DCs pulsed with recombinant HCA587 protein

Autologous PBLs were co-cultured with HCA587 protein-pulsed DCs at a ratio of 20 : 1 with the same medium and procedure as before. After four consecutive cycles of stimulation with protein-loaded DCs, CD8⁺ T cells were separated from the bulk cultures by positive selection with CD8-Dynabeads and further expanded in number for 3–6 days by the addition of autologous irradiated PBMCs (40 Gy) and 10 ng/ml IL-2 in RPMI-1640 with 10% AB serum and then used for the following assays.

Enzyme-linked immunospot (ELISPOT) assay for IFN-γ

ELISPOT assay was performed as described previously [18] using anti-human IFN-γ capture monoclonal antibody (mAb) (1-D1K, Mabtech, Stockholm, Sweden) and biotinylated anti-human IFN-γ detecting mAb (7-B6-1, Mabtech, Stockholm, Sweden). Ten days after the last stimulation, 2·5 × 10⁴ HCA587-stimulated CD8⁺ T cells were used as effector cells, and 5 × 10⁴ T2 cells or HCC cell lines loaded with HCA587 peptides or not were used as stimulator cells. Control wells contained tumour cells alone, CTLs with or without T2 cells. For HLA-blocking experiments, stimulator cells were preincubated with anti-HLA class I mAb (W6/32) or anti-HLA class II mAb (L243) at a concentration of 25 µg/ml for 1 h at 4°C, and then cultured with effector cells in the ELISPOT assay. The spot numbers per well were determined automatically with the use of a computer-assisted video image analyser (Sage Creation, Beijing, China). Indicated spot numbers per well represent mean values of three replicates.

Cytotoxicity assay

The cytotoxic activity of bulk CTLs was tested by the Granzyme B (GrB) ELISPOT assay as described by Weaver et al. [22]. Briefly, MultiScreen-IP plates (PVDF Membranes, Millipore) were coated with anti-human GrB capture antibody (clone GB-10, Pelicluster, Cell Science, Norwood, MA, USA). Effector cells (2·5 × 10⁴/well) were added to triplicate wells followed by 5 × 10⁴ target cells per well. After effector and target cells were incubated at 37°C for 4 h, the plates were washed and biotinylated anti-human GrB detecting antibody (clone GB-11, Pelicluster, Cell Science, Norwood, MA, USA) was added for 2 h at room temperature (RT). The plates were then washed and incubated with streptavidin–alkaline phosphatase (Gibco-BRL Life Technologies, Rockville, MD, USA) for 1 h at RT. Spots were visualized with filtered BCIP-NBT phosphatase substrate (KPL, Gaitherburg, MD, USA) and enumerated as stated above.

Results

Selection of potential CD8⁺ T-cell epitopes of HCA587 for binding to HLA-A*0201 molecules

To identify the potential HLA-A*0201-restricted epitopes from the 373-aa long open reading frame of HCA587 protein, we used different computer programs available via the internet from the National Center for Biotechnology Information (http://bimas.dcrt.nih.gov/molbio/hla_bind/), the University of Tubingen (http://www.uni-tuebingen.de/uni/kxi/) and the University of Oklahoma Health Science Center (http://hlaligand.ouhsc.edu/prediction.htm) [23,24] to predict binding affinity and stability based on matrices derived from known epitopes. The six peptides, derived from HCA587 protein with relatively higher predicted binding scores in different computer programs, are listed in Table 1. They were synthesized and tested for actual HLA-A*0201 binding capacity using the T2 binding assays. The Flu matrix peptide GILGFVFTL was served as positive control in this assay(Fig. 1). As shown in Table 1, all the synthesized peptides from HCA587 were bound to HLA-A*0201 molecules with different affinities, but peptides no. 2 and no. 3 apparently up-regulated the HLA-A*0201 molecules and showed high affinities to HLA-A*0201 molecules, whereas other synthesized peptides had low affinities to the molecules. Therefore, peptides no. 2 and no. 3 from HCA587 protein were selected to determine whether they were capable of stimulating human HCA587-specific CD8⁺ T cells in vitro.

Fig. 1.

Facs analysis of T2 binding assay. The binding affinity of each HCA587 peptide and control (flu) peptide (50 µg/ml) was determined by incubation with T2 cells overnight in serum-free medium containing human β₂-M (5 µg/ml).

Induction of specific CD8⁺ T cell lines by in vitro priming with high-affinity peptides

To assess the immunogenicity of the selected peptides no. 2 and no. 3 from HCA587 protein, CD8⁺ T cells were purified by immunobeads from the peripheral blood of 15 healthy HLA-A*0201 donors to obtain a population more that 96% pure. They were stimulated in vitro with peptide-pulsed autologous DCs for CTL priming and expansion, as described in Materials and methods. On day 10 after the fourth round of stimulation, the specificity of in vitro primed responder T cells was tested in IFN-γ release ELISPOT assays. As shown in Fig. 2, no. 2 peptide with the high binding affinity (FI 3·91) for the HLA-A*0201 molecule (Table 1) activated peptide-responsive CTLs that could be detected by IFN-γ release ELISPOT assays after restimulated with no. 2 peptide-loaded T2 in two donors. However, the no. 3 peptide, also displaying high binding affinity (FI 4·12) for HLA-A*0201 molecules, was unable to elicit peptide-response CTLs in all donors. This may imply that the peptide binding capacity to HLA is not the only factor to determine its immunogenicity. In this assay, the peptide-specific CTL response was verified because no specific spots were observed when T2 cells were loaded with the other HCA587 peptides distinct from the no. 2 peptide.

Fig. 2.

Effector activity of in vitro peptide-primed CD8⁺ T cells in interferon (IFN)-γ enzyme-linked immunospot (ELISPOT) assays. CD8⁺ T cells from two different normal donors (each donor shown in separate column) were, respectively, stimulated four times with HCA587 no. 2 and no. 3 peptide (each shown in separate rows). The CD8⁺ T cells (2·5 × 10⁴/well) were then restimulated with T2 cells (5 × 10⁴/well) loaded with various peptides (10 µg/ml) in IFN-γ release ELISPOT assays. Number of specific responsive spots per well is shown as mean ± standard deviation from three parallel wells. The peptide-stimulated specific CD8⁺ T cells response was dramatically blocked by anti-HLA class I monoclonal antibody (mAb) (W6/32), but not by anti-HLA class II mAb (L243).

Moreover, T2 cells that loaded with no. 2 peptide were also tested as stimulators in the presence of blocking mAb. The anti-class I mAb W6/32 dramatically decreased spot numbers compared to the control group with the addition of the anti-HLA class II mAb L243, indicating that the in vitro-primed effector CD8⁺ T cells were HLA class I-restricted (Fig. 2).

Determination of the endogenously procession of HCA587 no. 2 peptide

To understand further if HCA587 no. 2 peptide was able to be processed endogenously in APCs, recombinant HCA587 protein was produced in E. coli and used to pulse autologous DCs with the help of DOTAP, which was instrumental in facilitating the cytoplasmic incorporation of exogenous antigen for MHC class I-restricted presentation to CD8⁺ T cells [25]. Then, PBMCs from these two no. 2 peptide ELISPOT positive normal donors were co-cultured with these protein-loaded DC in medium supplemented with IL-2 and IL-6. After four rounds of protein stimulation, the CD8⁺ T cells were purified with Dynabeads and used as effector cells in ELISPOT assay after restimulation with T2 cells loaded with different peptides derived from HCA587 protein. As shown in Fig. 3, only specific spots were formed in the wells where CD8⁺ T cells were restimulated with no. 2 peptide-loaded T2 cells in both donors. Although the numbers of spots were less than those generated from no. 2 peptide-stimulated CD8⁺ T cells, the assay clearly proved that HCA587 protein was able to be endogenously cleaved and processed by autologous DCs to present MHC-no. 2 peptide complex on the DC surface.

Fig. 3.

Effector activity of in vitro recombinant HCA587 protein-primed CD8⁺ T cells in interferon (IFN)-γ enzyme-linked immunospot (ELISPOT) assays. Peripheral blood lymphocytes (PBLs) from two different normal donors (donor 1 and donor 2) were stimulated four times with HCA587 recombinant protein, then the effector CD8⁺ T cells were positively purified with Dynabeads. The effector CD8⁺ T cells (2·5 × 10⁴/well) were restimulated with T2 cells (5 × 10⁴/well) loaded with various peptides(10 µg/ml) in IFN-γ release ELISPOT assays. Number of specific responsive spots per well is shown as mean ± standard deviation from three parallel wells.

Expression of HCA587 in HCC cell lines

To identify whether HCA587 was expressed in HCC cell lines, we collected five cell lines derived from HCC, namely HepG2, HLE, SMMC-7721, Bel-7402 and Bel-7405, and detected the mRNA expression of HCA587 by RT-PCR. The HCA587 mRNA expression was detected only in HepG2 among the five cell lines detected, in which mRNA integrity was confirmed by the housekeeping gene G3PDH. The expected PCR products of 1·2kb HCA587 and 452 bp of G3PDH in HCC were equal to the products in testis which served as positive control, and these products were confirmed further by DNA sequencing (Fig. 4). Moreover, IHC staining showed that the HCA587 protein was expressed in the cytoplasm of HepG2 cells (Fig. 5a), but not expressed in HCA587 mRNA^– cell lines HLE (Fig. 5b) and SMMC-7721(Fig. 5c). Because HLA class I molecular typing analysis showed that HCC cell line HepG2 and HLE carried the same HLA allele in A locus(A2 and A24) [26], while SMMC-7721 was HLA-A2 negative, they were used as the HCA587 natural targets and controls in the following restriction studies.

Fig. 4.

Analysis of HCA587 mRNA expression in HCC cell lines. HCA587 mRNA expression was assessed by reverse transcription-polymerase chain reaction (RT-PCR) from the cDNA samples prepared from HCC cell lines and testis: lane 1, HepG2; lane 2, HLE; lane 3, SMMC-7721; lane 4, Bel-7402; lane 5, Bel-7405; lane 6, testis as positive control; lane 7, H₂O as negative control.

Fig. 5.

Analysis of HCA587 protein expression in hepatocellular carcinoma (HCC) cell lines. HCA587 protein expression was assessed by immunochemical staining in HCA587 mRNA⁺ HepG2 cells (a), HCA587 mRNA^– HLE cells (b) and SMMC-7721 (c) with anti-HCA587 polyclonal antibody. The slides were then stained with the peroxidase-labelled goat-anti-rabbit IgG. The colour was developed with diaminobenzidine (DAB) chromogen and the slides were counterstained with haematoxylin.

HCA587-specifc CD8⁺ T cells recognizing HepG2 cell line

More importantly, we examined whether peptide-specific T cell lines were able to recognize the selected HLA-A2 human HCC cell line HepG2 that endogenously expressed and presented HCA587-derived epitopes. No. 2 peptide-stimulated CD8⁺ T cells from the two normal donors were tested against the HCA587 protein⁺ HCC cell line HepG2, the HCA587 protein^– HCC cell line HLE and SMMC-7721. To increase the processing and presentation of endogenous antigens, tumour cells were cultured in medium supplemented with 200 U/ml IFN-γ over 48 h before being used as target cells in ELISPOT assays. When HepG2 cells were used as targets in ELISPOT assay they were recognized by no. 2 peptide-specific CD8⁺ T cells from the both donors, indicating further that the no. 2 peptide epitope could be endogenously processed and presented by HepG2 expressing natural HCA587 protein(Fig. 6). To test whether the recognition by no. 2 peptide-reactive CTLs was specific for this HCA587 epitope, HLA-A2⁺ and HLA-A2^– HCC cell lines HLE and SMMC-7721, which were loaded with no. 2 peptide or not, were used as targets. In contrast to HepG2 cells, HCA587 protein^– HLE and SMMC-7721 cells were not recognized by peptide-reactive CTLs from the both donors. When these tumour cells loaded with exogenous addition of no. 2 peptide, HLA-A2⁺ HLE cells were able to activate peptide-specific cells while HLA-A2^– SMMC-7721 cells were not, indicating that this no. 2 peptide was presented by HLA-A2 molecules, not by other MHC alleles(Fig. 6). These data showed that the generated CTLs are specific for the peptide FLAKLNNTV encompassing residues 317–325 of HCA587.

Fig. 6.

In vitro HCA587 no. 2 peptide-primed CD8⁺ T cells from the two donors recognized the hepatocellular carcinoma (HCC) cell lines loaded with no. 2 peptide (10 µg/ml) or not. The interferon (IFN)-γ enzyme-linked immunospot (ELISPOT) assays were performed using the peptide-reactive CD8⁺ T cells (2·5 × 10⁴/well) from donor 1 and donor 2 to estimate its activity against HCA587 protein⁺ HepG2 cells (5 × 10⁴/well), HCA587 protein^– HLE cells (5 × 10⁴/well) and SMMC-7721 (5 × 10⁴/well). Number of specific responsive spots per well is shown as mean ± standard deviation from three parallel wells.

No. 2 peptide specific CTLs in HCC patients

To determine whether the immunogenicity of HCA587 no. 2 peptide is reproducible in other individuals, particularly in HCC patients. PBMCs from four HLA-A*0201 HCC patients were sensitized in vitro with no. 2 peptide under the same conditions as described before. Cytotoxic activity of no. 2 peptide-specific bulk CTLs derived from these HCC patients against HCC cell lines was detected by the GrB cytotoxicity assay. Of the four HCC patients, the no. 2 peptide induced specific CTLs in two HCA587 positive HCC patients because both these bulk CTLs were capable of mediating cytotoxicity by GrB secretion to HLA-A2⁺ HCA587 protein⁺ HepG2, but not to HLA-A2⁺ HCA587 protein^– HLE, or HLA-A2^– SMMC-7721 cells (Fig. 7). In contrast, CTLs isolated from the other two HCA587-negative HCC patients failed to recognize all the above HCC cell lines. Because CTLs reactive to this epitope from both normal donors and HCC patients can recognize HCA587 protein⁺ HCC cell line HepG2, HCA587 and this epitope should be promising target antigens for vaccination against HCA587-bearing HCC patients.

Fig. 7.

Granzyme B cytotoxic activity of no. 2 peptide-specific bulk CTLs from two HLA-A*0201 hepatocellular carcinoma (HCC) patients on HCC cell lines. Effector cells(2·5 × 10⁴/well) were run against HepG2, HLE, SMMC-7721 targets (5 × 10⁴/well) in a 4-h assay at 37°C. Data is presented as spots per well ± standard deviation and is representative of three experiments with similar results.

Discussion

The integration of molecular and immunological techniques has led to the identification of a large number of human tumour associated antigens recognized by CD8⁺ T cells and antibodies. Most of these identified antigens can be divided into the following categories according to expression patterns or structural features: (i) CT antigens, such as the members of the MAGE gene family, SSX2, SCP1 and NY-ESO-1 [27]; (ii) mutated antigens, e.g. CDK4 [28]; (iii) overexpressed antigens, e.g. HER-2/neu [29]; (iv) differentiation antigens, such as Melan-A/MART1, tyrosinase and gp100 in melanocytes [30]; and (v) virus antigens, such as human papillomavirus (HPV) in cervical cancer [31]. Of all these categories, CT antigens are of major interest as targets for immunotherapy because of their characteristic expression patterns in cancer and testis, but not in other normal tissues.

HCA587 was a CT antigen identified by SEREX from HCC cDNA expression library in our laboratory [15]. The cDNA of HCA587 was cloned in 1999 (GenBank Accession no. AF151378) with the identical sequence to the MAGE-C2 [32] and exhibited only two amino acid differences from CT10 [33]. The gene of HCA587, located in Xq27, belongs to MAGE-C subfamily with a large terminal exon encoding a protein of 373 amino acid residues. Since the first member of the human MAGE families was identified as a gene encoding tumour-specific antigen [11], these MAGE family antigens have been of particular interest for anti-tumour immunotherapy because of their specific expression on tumour cells [34]. Our previous work has proved that HCA587 is highly expressed in HCC patients [15]. Therefore, HCA587 may be a good potential target for immunotherapy to HCC patients, and the identification of HCA587-derived CD8⁺ T-cell epitopes has become a critical step in the potential to develop a peptide-based immunotherapy for HCC patients.

At present, numerous epitopes have been identified by sequencing peptides eluted from purified MHC from the tumour cells of cancer patients and recognized by TIL cells [16,17]. Other epitopes, such as those derived from the MAGE-C2 antigen, have been identified by expression cloning [35]. However, both processes are relatively laborious and inefficient. Recently, a reverse immunology approach has been used widely and proved to be an efficient method to identify tumour-associated peptides, particularly when T cells capable of recognizing tumour antigens are not available [36]. This approach includes a four-step procedure of (i) peptide selection based on specific MHC binding motifs from the amino acid sequence of a candidate protein, (ii) peptide-binding assay to determine the affinity of the predicted peptide with MHC, (iii) in vitro sensitization assay by pulsing peptides onto APCs for CTL induction from PBMCs of healthy donors or cancer patients and (iv) testing the generated CTLs toward tumour target cells endogenously expressing the antigen. Using this approach, a large number of CD8⁺ T-cell epitopes have been identified from several human TAAs, such as IL-13Rα2 [37], prostate stem cell antigen [38] and the glycoprotein B homologue of human herpesvirus 8 [39], etc.

In the present study, we used epitope prediction programs to select six HCA587-derived peptides that bind to HLA-A*0201 molecules because HLA-A*0201 is the most frequent HLA-A allele in Caucasian and Asian individuals. Peptides that bind to HLA-A*0201 molecules with high affinity are predominantly nonamers and decamers with allele-specific anchor residues at positions 2 and 9 for nonamers and positions 2 and 10 for decamers [24]. Typical anchor residues are isoleucine (I) or leucine (L) at position 2 and valine (V), leucine (L) or methionine (M) at the C-terminal position. All the selected HCA587 peptides fulfilled the requirements for HLA-A*0201 binding. Among the synthesized HCA587 peptides, no. 2 and no. 3 peptide were selected for the in vitro priming because of their higher affinity to HLA-A*0201 molecules according to T2-binding assays. Using an in vitro CTL sensitization protocol that allows stimulation of human CD8⁺ T cells with peptide-pulsed DCs in the presence of IL-2 and IL-6, we generated HLA-A*0201-restricted CD8⁺ T cells against HCA587 no. 2 peptide, FLAKLNNTV (HCA587_317−325) in two normal donors based on IFN-γ release ELISPOT assay, which is a favoured method widely employed for quantification of T lymphocytes responsive against TAA-derived epitopes [40].

We further used HCA587 recombinant protein pulsed-DC to prime polyclonal CD8⁺ T cells to see whether the CD8⁺ T cells specific to no. 2 peptide were able to be generated. ELISPOT results showed that T2 cells loaded with no. 2 peptide could trigger protein-stimulated CD8⁺ T cells to release IFN-γ in these two normal donors, while no IFN-γ release was observed by T2 cells loaded with no. 3 peptide in both donors. Our data indicated that HCA587 protein could be endogenously cleaved and processed in autologous DCs, and the specific no. 2 peptide–MHC complex could be presented on the cell surface to induce specific CD8⁺ T cell response. Although binding affinity of peptides in general correlates well with their immunogenicity [41], the HCA587 no. 3 peptide, with the higher affinity to HLA-A*0201 molecules in T2-binding assays, failed to induce specific T cell lines in normal donors. This implies that other factors may also be critical to determine the peptide immunogenicity, such as protein processing by proteosomal cleavage, peptide presentation and susceptibility of T cell response [42,43].

Finally, the HCC HepG2 cell line was proved to express HCA587 not only in mRNA level by RT-PCR, but also in protein level by IHC staining. More importantly, the peptide-specific CD8⁺ T cells induced from the two normal donors and two HCC patients were able to recognize the HLA-A2⁺ HCA587 protein⁺ HepG2 cells in both IFN-γ and GrB ELISPOT assays, while the generated CD8⁺ T cells could not recognize the HLA-A2⁺ HCA587 protein^– HLE cells, as well as the HLA-A2^– SMMC-7721. These data confirm the endogenous origin of no. 2 peptide in HepG2 cells and its presentation to the cell surface in the context of HLA-A2 molecules. Because GrB ELISPOT assay is a viable alternative to the standard ⁵¹Cr-release assay to measure granule-mediated cytotoxicity [44], these epitope-specific CTLs showed specific cytotoxic activity to HLA-A2⁺ HCA587 protein⁺ HepG2 cells by Granzyme B cytotoxic assay. In fact, not all HLA-A*0201 normal donors or HCC patients can mount peptide-specific CD8⁺ T cells in our experiments. These discrepancies might be due to the variability in size and availability of a specific T cell repertoire in different individuals. The ability of individual patients to develop CD8⁺ T cell responses to individual peptides can be strikingly heterogeneous [45].

Taken together, our study suggests that one novel HCA587-derived epitope (FLAKLNNTV aa 317–325) is capable of inducing HLA-A*0201-restricted CD8⁺ T cells. Because peptide-based cancer immunotherapy is tumour-specific, less toxic and could have a long-lasting effect, it is a potentially promising new treatment modality. Here we propose that this CD8⁺ T cell epitope could serve as a good candidate to design the peptide-based immunotherapeutic strategies for the treatment of HCA587-bearing HCC patients.