Purpose: Hepatocellular carcinoma (HCC) can express various cancer-testis antigens including NY-ESO-1, members of the SSX family, members of the MAGE family, SCP-1, and CTP11. Immunotherapy directed against these antigens is a potential alternative treatment for HCC. To date, it remains unclear whether HCC patients have spontaneous immune responses to these tumor antigens. The objectives of this study were to measure immune responses to NY-ESO-1, a promising cancer vaccine candidate, in HCC patients using the HLA-A2–restricted NY-ESO-1b peptide (p157-165) to measure cellular responses and whole protein to measure antibody responses.
Experimental Design: In HLA-A2+ patients with NY-ESO-1+ HCC, we analyzed T-cell antigen-dependent interferon (IFN)-γ and/or Granzyme B release by enzyme-linked immunospot (ELISPOT) assay and IFN-γ–producing intracellular cytokine flow cytometry (CytoSpot). As an assay independent of T-cell function, we performed tetramer staining. Antibodies to whole NY-ESO-1 were assayed by enzyme-linked immunosorbent assay.
Results: The frequency of specific CD8+ T-cell responses to NY-ESO-1b in 28 NY-ESO-1 mRNA+HLA-A2+ HCC patients was 35.7% (10 of 28). The average magnitude of effector CD8+ T cells was 0.3% (89 ± 59 per 2.5 × 104 CD8+ cells) and 1.2% as measured by IFN-γ release ELISPOT and CytoSpot assays, respectively. These in vitro induced NY-ESO-1b–specific CD8+ T cells can also recognize HepG2 cells transfected with pcDNA3.1-NY-ESO-1 in both IFN-γ and Granzyme B ELISPOT assays. Frequencies of NY-ESO-1b–specific T cells in several patients were confirmed by tetramer staining. Nonfunctional tetramer+CD8+ T cells were also present. The CD8+ T-cell response was apparently increased in patients with late-stage HCC. A discordance between antibody and CD8+ T-cell responses in HCC patients was observed.
Conclusions: The elevated frequency of specific CD8+ T-cell responses to NY-ESO-1b in NY-ESO-1 mRNA+HLA-A2+ HCC patients suggests that NY-ESO-1 is appropriate for use in the immunotherapy of HCC patients.
Evidence that tumor antigens can elicit measurable immune responses has opened a new perspective for cancer immunotherapy (1 , 2) . Human tumor cells express diverse types of antigens, depending on the etiology and pathogenesis of the disease (3) . Among these, tumor germ-line gene-encoded cancer-testis (CT) antigens and melanocyte differentiation antigens have proved to be widely immunogenic and are being exploited in the experimental therapeutic vaccination of cancer patients (4) . One such CT antigen in clinical trial, NY-ESO-1, has been shown to induce specific cellular and humoral immune responses in melanoma, ovarian, and breast cancer patients without severe adverse effects (5 , 6) . Furthermore, in some patients with NY-ESO-1–expressing tumors, immunity to the antigen after vaccination has been associated with a complete or partial tumor remission (5) .
Hepatocellular carcinoma (HCC) is one of the most common and pernicious malignancies in China. Despite improvement in sonographic examinations and therapy for HCC, the outcome generally remains unsatisfactory. Recently, the expression of CT antigen-encoding genes in HCC tissues has been reported by several laboratories (7, 8, 9, 10) . Our own data have revealed that mRNA for at least 20 CT antigens can be detected in HCC tissues. Expressed genes include MAGE-A1 (11) , MAGE-A3 (12) , CTP11, SSX1, HCA587 (MAGE-C2), and FATE-BJ-HCC-2 (13) , each of which is expressed in 60% to 80% of HCC specimens, together with NY-ESO-1, MAGE-A4, MAGE-10, SSX-4, SSX2, and SCP-1 mRNAs, each of which is expressed in 30% to 49% of HCC specimens. Immunotherapy directed against these antigens thus represents a potential alternative for the treatment of HCC. Recently, MAGE-A1 p161-169- and MAGE-A3 p271-279–specific CD8+ T cells were identified by tetramers in 2 of 10 samples of tumor-infiltrating cells isolated from MAGE-A1 and/or MAGE-A3 mRNA-positive HCC patients (14) . In this study, the MAGE-A1 p161-169–specific CD8+ T cells were functional and capable of killing target cells. However, in general, it remains unclear whether HCC patients mount spontaneous immune responses against CT antigens and what effect these immune responses have in vivo against HCC. On the other hand, the potential efficacy of induced immunity against HCC has been indicated by a report that immunization of two HCC patients with autologous HCC lysate-loaded dendritic cells (DCs) resulted in prolonged survival extending to >3 years in one patient (15) . Because NY-ESO-1 is one of the most promising cancer vaccine candidates for a variety of tumor types, here we have evaluated the status of spontaneous CD8+ T-cell responses to the NY-ESO-1b peptide by IFN-γ and/or Granzyme B (GrB) release enzyme-linked immunospot (ELISPOT), IFN-γ-producing CytoSpot, and tetramer assays in HCC patients. A CD8+ T-cell response was found in a significant proportion of patients with advanced HCC, encouraging further exploration of the utility of this antigen in HCC immunotherapy.
MATERIALS AND METHODS
Patients, Tumors, and HLA-A2 Subtyping.
Tumor samples and peripheral blood mononuclear cells (PBMCs) were obtained from 113 HCC patients with stage II–IV disease. Among these patients, 48% (55 of 113) were HLA-A2+ as defined using polymerase chain reaction-sequence specific primer and sequence-based typing as described previously (16) . Of the 55 HLA-A2+ HCC patients, 51 were males, and 4 were females; mean patient age was 52 years (range, 28–70 years). The mean tumor size was 5.8 cm (range, 1.5–12.0 cm). Fifty of these patients (90.9%) were positive for hepatitis B surface antigen. According to the tumor-node-metastasis (TNM) classification, 29% (16 of 55), 33% (18 of 55), and 38% (21 of 55) of the tumors were of stage II, III, and IV, respectively. Fifty-one percent of patients had serum α-fetoprotein above the diagnostic cutoff (400 ng/mL).
Peptides and Tetramers.
The HLA-A2–restricted NY-ESO-1b peptide p157-165 (SLLMWITQC) was used for analysis of the CD8+ T-cell response to NY-ESO-1. The HLA-A2–restricted flu matrix peptide p58-66 (GILGFVFTL) and MAGE-A3 peptide p271-279 (FLWGPRALV) were used as controls. Tetramers of each peptide were prepared as described previously (17) .
Tumor Typing for NY-ESO-1 Messenger RNA.
Expression of NY-ESO-1 mRNA in tumor specimens was assessed by reverse transcription-polymerase chain reaction, using primers as described previously (18) .
Assays for NY-ESO-1 Antibody.
NY-ESO-1 serum antibodies were assayed by enzyme-linked immunosorbent assay using NY-ESO-1 recombinant protein purified from Escherichia coli. All of the sera were tested over a range of serial 4-fold dilution from 1:100 to 1:102,400. A positive reaction was defined as an absorbance value of a 1:400 diluted serum that exceeds the mean absorbance value of the negative control (coated with irrelevant protein) by 3 SDs.
Transfection of NY-ESO-1 Gene into HepG2 Cells.
Because the HLA-A*0201+ HCC cell line HepG2 does not express the NY-ESO-1 antigen, the full-length NY-ESO-1 gene constructed in pcDNA3.1 plasmid (kindly provided by Prof. Yao-Tseng Chen, Cornell University, Ithaca, NY) was transiently transfected into HepG2 cells using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. After a 48-hour incubation with DNA-Lipofectamine 2000 reagent complexes, the cells were harvested and used as target cells for cytotoxic T lymphocyte (CTL) assays.
Peptide Presensitization of CD8+ T Cells.
CD8+ T cells isolated from PBMCs with a purity of >95% were seeded into 48-well plates at a concentration of 5 × 105 cells per well in 10% human AB serum RPMI 1640. For antigen presentation, the autologous CD8+ T-cell–depleted PBMCs (CD8− PBMCs) or DCs were incubated with 2.5 μg/mL β2-microglobulin and 10 μg/mL peptide for 2 hours and then irradiated at 30 Gy. After washing, the CD8− PBMCs or the autologous DCs were added to the plates at ratios of 2:1 or 1:5, respectively. Interleukin (IL)-7 (10 ng/mL; Sigma, St. Louis, MO) was added at the initiation of the cultures for 48 hours, after which IL-2 (10 units/mL; R&D Systems, Minneapolis, MN) and IL-6 (500 units/mL; expressed and purified in our laboratory) were also added. In some cases, the CD8+ T cells were restimulated at day 8 and day 15 with peptide-loaded, irradiated autologous PBMCs. Flu matrix-responsive CD8+ T cells were cultured in parallel as a positive control, whereas CD8+ T cells cultured without peptide stimulation were used as a negative control.
Interferon-γ Enzyme-Linked Immunospot Assay.
The IFN-γ release ELISPOT assay was performed as described by Jager et al. (5) , with slight modifications, using a commercial kit (MABTECH, Stockholm, Sweden). Plates (Millipore MAHA S45; Millipore, Bedford, MA) were coated overnight with anti–IFN-γ monoclonal antibody (mAb) and washed six times. After blocking with 10% human AB serum RPMI 1640, peptide-pulsed T2 cells (5 × 104 cells per well) were added together with responder T cells (2.5 × 104 cells per well) to each well. After incubation for 20 hours at 37°C, cells were removed, and the plates were developed with a second (biotinylated) antibody (Ab) to human IFN-γ and streptavidin-alkaline phosphatase. Plates were developed for 10 minutes at room temperature in the dark, and the reaction was stopped by rinsing plates with distilled water. The membranes were air dried, and spots were counted using the Champ Spot II ELISOT reader system (Sage Creation, Beijing, China).
Granzyme B Enzyme-Linked Immunospot Assay.
The GrB release ELISPOT assay was performed in a similar manner to the IFN-γ ELISPOT assay as described previously (19) . Briefly, plates (Millipore MAHA S45; Millipore) were coated overnight at 4°C with 100 μL/well antihuman GrB capture Ab (clone GB-10; U-Cytech Biosciences, Utrecht, the Netherlands) and washed six times with PBS. After blocking with assay medium (200 μL per well), HepG2 cells (transfected with pcDNA3.1-mock or pcDNA3.1-NY-ESO-1 vector) or peptide-pulsed T2 cells (5 × 104 cells per well) were added together with responder T cells (2.5 × 104 cells per well) to each well. Then, after incubation for 20 hours at 37°C, cells were removed, and the plates were developed with a second (biotinylated) Ab (clone GB-11; U-Cytech Biosciences) and streptavidin-alkaline phosphatase. After the reaction was stopped, the spots in each well were counted as described above.
Intracellular Interferon-γ Staining for Fluorescence-Activated Cell-Sorting Analysis (CytoSpot Assay).
PBMCs were cultured with or without peptide presensitization as described above. To analyze the cell surface expression of CD8 and intracellular IFN-γ, the Cytofix/Cytoperm with GolgiStop kit (PharMingen, Becton Dickinson, San Jose, CA) was used according to the manufacturer’s instructions. Briefly, the effector CD8+ T cells generated were activated with peptide-loaded DCs for 1 hour at an effector to target cell ratio of 10:1, followed by the addition of GolgiStop (monesin) for 5 hours. Cells were stained with fluorescein isothiocyanate-conjugated anti-CD8 mAb (Becton Dickinson), and then fixed and permeabilized using Cytoperm for 20 minutes on ice. After washing, the cells were stained with phycoerythrin-labeled antihuman IFN-γ mAb. Two-color analysis was performed on FACSCalibur (Becton Dickinson) to determine the proportion of CD8+IFN-γ+ cells. The positive control of the assay consisted of phytohemagglutinin-stimulated T cells, whereas the negative control comprised T cells cultured in cytokine-containing medium without peptide stimulation. The number of CD8+ T cells responsive to the specific peptide was obtained by subtracting background reactivity from the specific response to the stimulated peptide.
CD8+ T cells were presensitized with peptide as described above. Seven days after each round of stimulation, the responder T cells were expanded with 500 units/mL IL-2 (20) and incubated for an additional 5 days. Cells were then stained with phycoerythrin-labeled NY-ESO-1 p157-165/HLA-A*0201 tetramer and control tetramer MAGE-A3 p271-279/HLA-A*0201 for 15 minutes at room temperature before the addition of fluorescein isothiocyanate-conjugated CD8 mAb for 15 minutes on ice. After washing, stained cells were analyzed by flow cytometry (FACSCalibur; Becton Dickinson).
RESULTS AND DISCUSSION
HLA-A2 Subtypes and NY-ESO-1 Messenger RNA Expression in Hepatocellular Carcinoma Tissue Specimens.
Of a total of 113 HCC patients assayed, 55 (48%) were found to be HLA-A2+. NY-ESO-1 mRNA transcripts were detected in tumors resected from 28 (51%) of these patients. In these 28 HLA-A2+ NY-ESO-1 mRNA+ patients, the percentages of A2 subtypes were as follows: A*0201, 29% (8 of 28); A*0206, 18% (5 of 28); A*0207, 25% (7 of 28); and A*0203, 36% (10 of 28) [two patients, CL70 and CL72, are heterozygotes for A2 subtypes and have been counted twice]. Detailed demographic data and the HLA-A2 subtypes of these patients are shown in Tables 1⇓ and 2⇓ , respectively.
The Frequency and Magnitude of Specific CD8+ T-Cell Responses to the NY-ESO-1b Peptide.
Spontaneous CD8+ T cell responses to the NY-ESO-1b peptide were monitored in 42 of the 55 HLA-A2+ HCC patients (28 with NY-ESO-1 mRNA+ and 14 with NY-ESO-1 mRNA− tumors). These patients were distributed among three groups and assayed by IFN-γ release ELISPOT (group 1), IFN-γ–producing CytoSpot (group 2), or GrB release ELISPOT assay (group 3). Some patients were assessed by two methods. Representative data from these three methods are shown in Figs. 1A and B⇓ and 3⇓ .
Estimation by Interferon-γ Release Enzyme-Linked Immunospot Assays.
HLA-A2+ blood samples were obtained from 25 of the HCC patients (16 with NY-ESO-1mRNA+ tumors and 9 with NY-ESO-1mRNA− tumors) and 13 healthy donors. Purified CD8+ T cells from these blood samples were stimulated with NY-ESO-1b peptide presented by autologous CD8− PBMCs. From a total of 16 NY-ESO-1 mRNA+ patients, 15 CD8+ T-cell samples were stimulated once with peptide, and 11 of these 15 samples were boost stimulated for two or three rounds (Table 2)⇓ . The ELISPOT assay was performed on day 7, 14, and 21 after 1, 2, or 3 rounds of stimulation, respectively. Another CD8+ T-cell sample collected from patient BW01 was stimulated twice with the peptide, but ELSIPOT assay was performed on day 14 only (Table 2)⇓ . Among the 15 samples collected after a single round of peptide stimulation, NY-ESO-1b–specific, IFN-γ–producing CD8+ T cells were detected in cultures from the A*0201+ patients BW15 and BW12 (Fig. 1A⇓ ; Table 2⇓ ), but not in the other 13 NY-ESO-1mRNA+ or 9 NY-ESO-1 mRNA− patients (Table 2)⇓ . After two to three rounds of peptide stimulation, a potent CD8+ T-cell response was observed in cultures from patients BW11 (A*0201), BW15 (A*0201), BW02 (A*0206), and BW01 (A*0207), but not in the other 8 patients or 13 healthy donors (Fig. 2A)⇓ . Thus, after two to three rounds of peptide stimulation, a specific CD8+ T-cell response rate to NY-ESO-1b of 33.3% (4 of 12) was detected, significantly higher than that (13.3%, 2 of 15) observed after a single peptide stimulation. The magnitude of the CD8+ T-cell response as shown for the sample from patient BW15 was also markedly increased by 2.75-fold (Table 2)⇓ . The CD8+ T-cell responses were judged to be specific because (a) the IFN-γ–producing effector CD8+ T cells only responded to the peptide they encountered in the stimulation phase, not to the peptide encountered only in the effector phase, and (b) the secretion of IFN-γ was entirely blocked by anti-major histocompatibility complex class I Ab W6/32 (Fig. 2B)⇓ . These results indicated that the boost-stimulated increase in frequency of the CTL response to the NY-ESO-1b peptide was the consequence of the expansion of the peptide-primed T cells, not the activation of naïve T cells. Thus, for optimal estimation of the preexisting CD8+ T-cell response to NY-ESO-1b peptide in HCC patients, stimulation of the cells with the peptide two to three times appears to be required. Our data also demonstrate that the HLA-A*0201–restricted NY-ESO-1b peptide can also be presented by the A*0206 and A*0207 subtypes. Taking the results of a single round and two to three rounds of stimulation together, the frequency of CD8+ T-cell responses to NY-ESO-1b was 31.3% (5 of 16) in HLA-A2+ patients bearing NY-ESO-1 mRNA+ HCC, and the average numbers of IFN-γ–producing cells were 82 ± 61 per 2.5 × 104 CD8+ cells (0.33 ± 0.24%).
Estimation by Intracellular Interferon-γ–Producing CytoSpot Assay.
In a group of 12 HLA-A2+ HCC patients (7 patients with NY-ESO-I mRNA+ tumors and 5 patients with NY-ESO-1mRNA− tumors), recall CD8+ T-cell responses were assessed by intracellular IFN-γ production using the CytoSpot assay. The T cells in nonadherent PBMCs recovered from these patients were stimulated twice with NY-ESO-1b peptide presented by autologous DCs. In the effector phase, the cultured T cells were restimulated with NY-ESO-1b–pulsed autologous DCs. The intracellular IFN-γ–producing effector CD8+ T cells were detected by a two-color fluorescence-activated cell-sorting analysis. The frequency of CD8+ T-cell response to NY-ESO-1b peptide was 43% (3 of 7; Table 2⇓ , group 2) in NY-ESO-1+HCC patients. The HLA-A2 subtypes of positive response patients were A*0203 in two cases and A*0207 in one case. The magnitude of the response ranged from 0.57% to 2.96%, with an average of 1.2%. A CD8+ T-cell response to NY-ESO-1b was not detected in the five PBMCs samples obtained from patients with NY-ESO-1 mRNA− HCC (Table 2⇓ , group 2).
To test whether the results of the CD8+ T-cell responses were reproducibly measured by both assays, three PBMC samples were assessed by both ELISPOT and CytoSpot assays (Table 2⇓ , group1). In samples obtained from two NY-ESO-1 mRNA+ HCC patients (BW12 and BW17), the CD8+ T-cell response was positive by both assays in BW12, and negative by both assays in BW17. In a PBMC sample obtained from patient BW25, who had NY-ESO-1 mRNA− HCC, a CD8+ T-cell response was not detected by either assay. Thus, both assays gave similar results.
Estimation by Granzyme B Release Enzyme-Linked Immunospot Assay.
In five HLA-A2+ NY-ESO-1 mRNA+ HCC patients, cultures of CD8+ T cells were assayed by the GrB ELISPOT as an alternative to the 51Cr release assay for monitoring cell-mediated cytotoxicity (19) . To compare with IFN-γ secretion function in the same patients, the IFN-γ ELISPOT assay was performed simultaneously. After two or three rounds of peptide stimulation, a potent CD8+ T-cell response was observed in cultures from patients BW41 (A*0201) and CL70 (A*0201/0207) in both assays (Table 2⇓ , group 3). The frequencies of the effector T cells were similar in both assays, indicating that the majority of the IFN-γ–secreting CD8 T cells also have the ability to release GrB.
Combining the results obtained from the three assays, the overall recall CD8+ T-cell response specific to NY-ESO-1b peptide was 35.7% (10 of 28) in HCC patients bearing NY-ESO-1+ tumors.
Recognition of an NY-ESO-1–Expressing Hepatocellular Carcinoma Cell Line by NY-ESO-1b–Specific CD8+ T Cells.
Using cultures from patients BW41 and CL70, the inducible NY-ESO-1b–specific CD8+ T cells were tested for their ability to recognize and lyse an NY-ESO-1+ HCC cell line in both the IFN-γ and GrB ELISPOT assay. Because no HCC cell lines expressing NY-ESO-1 are available, we transiently transfected pcDNA3.1-NY-ESO-1 into HLA-A*0201+ HepG2 cells that were then used as target cells in the ELISPOT assay. NY-ESO-1 protein expression in the transfected HepG2 cells was confirmed by immunohistochemical staining using anti–NY-ESO-1 Ab E978 (data not shown). After incubation with NY-ESO-1–expressing HepG2 cells, a potent response was observed from both assays, suggesting that the in vitro inducible NY-ESO-1b–specific CD8 T cells can recognize the NY-ESO-1+ HCC cells and secret IFN-γ and GrB (Fig. 3)⇓ , a key effector molecule for killing and inducing apoptosis (21) . Comparison with the results obtained with HepG2 cells transfected with and without pcDNA3.1-mock revealed that the response was specific, although the background in the GrB ELISPOT assay was higher than that in the IFN-γ ELISPOT assay. The frequency of the effector CD8 T cells releasing GrB was similar to that of cells releasing IFN-γ in patient CL70 but much higher in patient BW41, indicating that some other cytotoxic cells may also be involved in lysing the HCC cells.
NY-ESO-1b Tetramer+ CD8+ T Cells in Hepatocellular Carcinoma Patients.
In samples of peptide-stimulated cultured CD8+ T cells obtained from 12 HLA-A2+ NY-ESO-1 mRNA+ HCC patients, a tetramer assay was performed in parallel with the ELISPOT assay. Six samples were tetramer positive with a range of 0.17% to 3.4%, whereas six samples were tetramer negative (Table 2⇓ , Figs. 1C⇓ and 3⇓ ). Of the six tetramer-positive samples, two (BW15 and BW12) were positive in an IFN-γ release ELISPOT assay after one round of peptide stimulation (Fig. 1A)⇓ ; two samples (BW41 and CL70) were positive in the IFN-γ and GrB ELISPOT assay after two rounds of peptide stimulation (Fig. 3)⇓ ; one sample (BW02) was positive in the IFN-γ release ELISPOT assay after three rounds of peptide stimulation, but not after two rounds of stimulation (Fig. 2B)⇓ ; and one sample (BW38) remained negative in the functional assay (Fig. 2B)⇓ . In six samples of CD8+ T cells (BW05, BW39, BW40, CL69, CL72, and CL82), neither the tetramer nor IFN-γ release ELISPOT assay was positive (representative data from patient BW40 is shown in Fig. 1C⇓ ). The percentage of CD8+ tetramer+ T cells from some patients did not correlate with the numbers of spot-forming cells counted in the ELISPOT assay in all cases. For example, after a single round of stimulation, patient BW15 had 3.4% NY-ESO-1b–specific T cells as assessed by tetramer analysis compared with about 50 spots of a total of 25,000 CD8+ T cells applied to the ELISPOT assay. Such discrepancies have been reported previously (22) . The most reasonable explanation is that although the antigen-specific T-cell clones were expanded after antigen stimulation, some do not secret cytokines or cytotoxic molecules.
Our results reveal that HLA-A2+ NY-ESO-1+ HCC patients have specific, spontaneous CD8+ T-cell responses to NY-ESO-1b peptide, demonstrating that the frequency of NY-ESO-1b–specific CTL precursor is high in these patients. It is possible that a significant proportion of the specific CD8+ T-cell responses we have observed are attributable to the multiple rounds of peptide stimulation that we used. This is consistent with the report that MAGE-A3 p271-279/HLA-A2 tetramer+ CD8+ T cells were only detected in tumor-infiltrating cells of HCC patients after two rounds of peptide stimulation (14) . In addition, the high overall frequency of detectable CD8 responses may also be influenced by the ongoing viral infection that underlies the cancer in the majority of these patients. Furthermore, we found that these in vitro inducible NY-ESO-1b–specific CD8+ T cells can recognize and lyse NY-ESO-1–expressing HCC cell lines, suggesting that the NY-ESO-1 antigen can be processed and presented by HCC cells for CTL recognition. Also of interest, we have found that the HLA-A2 functional supertype for NY-ESO-1b presentation includes A*0201, A*0206, A*0207, and A*0203 alleles, which comprise 96% of the HLA-A2 northern Chinese population. This significantly expands the proportion of Chinese HCC patients capable of mounting a CD8+ T-cell response to NY-ESO-1b. We found that NY-ESO-1b–specific CD8+ T cells presensitized by autologous CD8− PBMCs from patients bearing non-HLA-A*0201 subtypes can recognize A*0201+ NY-ESO-1–transfected HepG2 cells and peptide-pulsed T2 cells. Because we do not have access to NY-ESO-1–expressing HCC cell lines bearing other HLA-A2 subtypes, we could not perform these assays with target cells expressing matched A2 subtypes. However, from previous studies (16 , 23) , we found that the HLA-A2 subtype of target cells used in CTL assays is not a key factor in influencing recognition by CTLs. Another phenomenon linked to the functional supertype is the background response, which was much lower in the peptide-stimulated CD8+ T cells of the HLA-A*0201 subtype than in the peptide-stimulated CD8+ T cells of the HLA-A*0206, HLA-A*0207, and HLA-A*0203 subtypes. For example, in our ELISPOT assays, the number of IFN-γ–producing CD8+ T cells without peptide stimulation was higher in the cultures from patients BW01 (HLA-A*0207) and BW02 (HLA-A*0206) than in cultures from patients BW11 (HLA-A*0201) and BW12 (HLA-A*0201). This phenomenon was also found in our previous studies measuring specific CD8+ T-cell response to flu matrix peptide p58-66 among 53 healthy individuals bearing different HLA-A2 subtypes (16) .
Tumor Stage and CD8+ T-Cell Response to the NY-ESO-1b Peptide in Hepatocellular Carcinoma Patients.
The frequency of CD8+ T-cell responses in relation to tumor stage was analyzed in 28 cases. As shown in Table 3⇓ , the frequency of CD8+ T-cell responses to NY-ESO-1b peptide in the IFN-γ release ELISPOT assay was approximately 18% (2 of 11), 60% (3 of 5), and 42% (5 of 12) in patients with HCC at stage II, III, and IVa, respectively. Comparing the results of early-stage (stage II; 18%) and late-stage disease (stages III and IV; 47%), the frequencies of responses increased with progression of the tumor and might be related to the presence of metastasis (24) . Considering other clinical pathological data listed in Table 1⇓ , no significant correlation with stage was found with the CD8+ T-cell response analyzed.
There are numerous reports that the immune responses are more frequently detectable at relatively late stages of tumor progression than at early stages, in which the tumor remains relatively small (17 , 25 , 26) . This may be due to lack of metastasis to the draining lymph nodes, in the latter situation, whereby the immune cells would be more likely to be effectively stimulated by tumor antigens. Likewise, as tumors grow, apoptosis and necrosis occur, resulting in the release of tumor components that induce inflammatory responses, including infiltration of T and B cells that would also initiate an adaptive immune response. Although the prognosis of the patients diagnosed at later stages is worse than that seen when early diagnosis is possible, it is only in these late stages that spontaneous responses occur. We do not know to what extent they are able to reduce the speed of subsequent progression in these large tumors. Overall, our aim is to induce similar responses in patients with earlier stage disease, when the immune response may be more effective. However, the presence of large tumors that effectively prime the immune response may also be of benefit to the efficacy of the proposed therapy.
Correlation of Antibody and CD8+ T-Cell Responses to NY-ESO-1.
Serum and PBMC specimens were collected from 22 HLA-A2+ HCC patients who had undergone surgical resection of their tumors. In the resected HCC specimens, 13 were NY-ESO-1 mRNA+, and 9 were NY-ESO-1 mRNA−. As shown in Table 4⇓ , anti-NY-ESO-1 Ab was detected in the serum of 4 of 13 NY-ESO-1 mRNA+ HCC patients with an Ab titer of 1:1,600 in two patients and 1:6,400 in the other two patients. However, overall, only two patients exhibited both an Ab response to NY-ESO-1 protein and effector CD8+ T-cell response to NY-ESO-1b peptide; four patients had an effector CD8+ T-cell response but no detectable Ab, whereas five patients had neither Ab nor a peptide-specific CD8+ T-cell response. There was no Ab response or peptide-specific CD8+ T-cell response in nine patients with NY-ESO-1 mRNA− HCC. These results indicate that a spontaneous CD8+ T-cell response to NY-ESO-1b peptide in HCC is not necessarily concomitant with the presence of Ab.
Taken together, our experiments have demonstrated that nearly one third of HLA-A2+ NY-ESO-1 mRNA+ HCC patients have a specific CD8+ T-cell response to NY-ESO-1b peptide. This peptide is thus a good vaccine candidate for immunotherapy for HCC patients, especially for those bearing advanced HCC but in whom overall physiologic conditions have not deteriorated, allowing sufficient survival time for a potentially effective immune response to be elicited.
Grant support: The James R. Kerr Program of the Ludwig Institute for Cancer Research (KSP003), China National 973 Program (G1999053904), Beijing Municipal Government Foundation for Natural Sciences (7001002), and China National 863 High Technology Program (2001AA217151).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: X-Y. Shang, H-S. Chen, and H-G. Zhang contributed equally to this work.
Requests for reprints: Wei-Feng Chen, Department of Immunology, School of Basic Medical Sciences, Peking University Health Science Center, 38 Xue Yuan Road, Beijing 100083, People’s Republic of China. E-mail:
- Received March 11, 2004.
- Revision received July 13, 2004.
- Accepted July 19, 2004.