Gastric cancer has the highest mortality rate and the second-highest morbidity rate of all malignant tumors in China. Since cancer/testis (CT) antigens are expressed in various types of human tumors but generally not in normal tissue except for testis, they are promising antigens for cancer immunotherapy. NY-ESO-1, in particular, is the most immunogenic of the CT antigens. To study the feasibility of developing a CT antigen vaccine for gastric cancer, 101 gastric cancer samples were analyzed for the presence of NY-ESO-1 mRNA and that of 10 other CT antigen genes. Twelve out of 101 samples (11.9%) were found to be NY-ESO-1 mRNA-positive, 11 of them from advanced stage patients. In 7 of the 12 NY-ESO-1 mRNA-positive samples, the NY-ESO-1 protein was also detected by immunohistochemistry. An autologous humoral immune response to NY-ESO-1 was detected in 6 of 12 advanced stage NY-ESO-1 mRNA-positive patients, indicating that NY-ESO-1 is immunogenic in advanced stage gastric cancer. The serum from a patient with an NY-ESO-1 negative but LAGE-1 positive tumor was also found to be NY-ESO-1 antibody positive, possibly due to cross-reactivity between NY-ESO-1 and LAGE-1. All NY-ESO-1 mRNA-positive gastric cancer samples also expressed one to seven additional CT genes, revealing a tendency toward a clustered expression pattern, regardless of disease stage. About 74% of the samples expressed at least one CT antigen, most frequently MAGE-3 (41.6%). NY-ESO-1 and MAGE-3 are thus potential targets for a multivalent CT antigen vaccine.
This article was published in Cancer Immunity, a Cancer Research Institute journal that ceased publication in 2013 and is now provided online in association with Cancer Immunology Research.
Gastric cancer has the highest mortality rate and the second-highest morbidity rate of all malignant tumors in China. Since only 10-15% of patients are diagnosed at an early stage, most hospital cases are in the advanced stages (1). The therapeutic approaches are often limited to radiation therapy and chemotherapy, with poor overall clinical outcome. Owing to their high specificity and low toxicity, cancer vaccines represent a potential alternative therapeutic approach.
NY-ESO-1 is a cancer/testis (CT) antigen originally identified by the serological analysis of recombinant cDNA expression cloning (SEREX) method in the context of esophageal cancer (2). It has been found in a wide variety of cancers, including melanoma, breast cancer (3), lung cancer (4), synovial sarcoma (5), and hepatocellular carcinoma (6, 7). Spontaneous humoral and cellular immune responses to NY-ESO-1 are frequently detected in cancer patients whose tumor expresses this molecule (8). Indeed, NY-ESO-1 is the most immunogenic CT antigen discovered so far, making it a highly promising therapeutic vaccine component.
In order to explore the relevance of the development of a vaccine for gastric cancer based on NY-ESO-1, we analyzed its expression and that of 10 other CT genes in gastric tumor specimens. In addition, the expression of the NY-ESO-1 protein was evaluated by immunohistochemistry, and the autologous humoral immune response to NY-ESO-1 was tested to analyze the immunogenicity of NY-ESO-1 in gastric cancer.
NY-ESO-1 expression in gastric cancer
The presence of NY-ESO-1 mRNA was determined by RT-PCR in 101 gastric cancer specimens that included 13 stage I, 17 stage II, 45 stage III, and 24 stage IV tumors. Two specimens were from patients with recurrent disease. Twelve of the 101 samples (11.9%) were NY-ESO-1 mRNA-positive (Figure 1), of which 11 were from tumors at advanced stages (8 stage III and 3 stage IV). Although the overall number of NY-ESO-1 positive samples was not high, the frequency of NY-ESO-1 mRNA expression increased from the early stages (I and II, 1/30, 3.3%) to the advanced stages (III and IV, 11/69, 15.9%).
NY-ESO-1 protein expression was evaluated by immunohistochemical staining with the NY-ESO-1 specific monoclonal antibody E978. Positive staining was seen in the cytoplasm or in both cytoplasm and nucleus (shown in Figure 2). Seven of the 12 NY-ESO-1 mRNA-positive samples were also positive for NY-ESO-1 protein. All 7 samples were gastric adenocarcinomas. Among them, two cases were accompanied by neuroendocrine differentiation. One of the seven cases positive for NY-ESO-1 protein was stage I, with the dysplastic epithelium adjacent to the tumor also showing positive staining (data not shown); the other six cases were all from patients at advanced stages.
Autologous humoral immunity to NY-ESO-1 in gastric cancer
Serum samples from all 101 patients were analyzed for the presence of antibody to NY-ESO-1. Six of 12 serum samples from patients with stage III and IV NY-ESO-1 mRNA-positive tumors and one of four samples from patients with stage III and IV LAGE-1 positive, NY-ESO-1 negative tumors were found to contain antibodies to NY-ESO-1 (Figure 3, Table 1). The positive cut-off for ELISA was set as 3 standard deviations above the mean of 8 negative controls. The criterion for a serum sample to be positive was two out of three serial dilutions giving a positive result. All data were analyzed after subtraction of the BSA background.
CT gene expression in gastric cancer
The expression of 10 other CT antigens in the 101 gastric cancer specimens was also examined. Examples of positive RT-PCR amplification for these CT genes are shown in Figure 4. All PCR products sequenced were confirmed to be the expected target gene sequences.
Table 2 summarizes CT expression in this group of gastric cancer specimens. MAGE-3 was the most frequently expressed CT gene (41.6%), followed by SSX-4 (26.7%), MAGE-4 (24.8%), MAGE-1 (23.8%), SSX-1 (22.8%), LAGE-1 (16.8%), CT10 (13.9%), and NY-ESO-1 (11.9%). The expression of SCP-1 (5.9%), CT7 (5.9%), and SSX-2 (3%) was relatively low.
The expression of some CT genes has been reported to correlate with tumor stage (9). In this study, advanced gastric tumors expressed NY-ESO-1, SSX-4, and CT10 more frequently (Figure 5). The expression frequency of NY-ESO-1 increased from 3.3% of stage I and II gastric tumors to 15.9% of stage III and IV tumors, while SSX-4 increased from 16.7% to 30.4% and CT10 increased from 6.7% to 15.9% respectively. However, this correlation of CT expression with clinical stage was not universal. For instance, MAGE-3 and SSX-1 gene expression was observed to be higher in lower stage tumors.
Clustered CT gene expression
The phenomenon of clustered CT gene expression, that is, tumor specimens that either simultaneously express multiple CT antigens or do not express any CT antigens, has been described previously (10). Of the gastric cancer samples analyzed, 42.6% were positive for two or more CT genes, whereas no CT antigen expression was detected in 25.7% (Figure 6). Among the 12 NY-ESO-1 mRNA-positive tumors, 10 (83%) expressed 5 or more of the 11 CT genes examined. This high clustering frequency was also seen in tumors positive for LAGE-1, CT7, and CT10, with tumors expressing 5 or more CT genes constituting 70% (12/17), 83% (5/6), and 57% (8/14) of each group, respectively. However, not all CT genes are similar in this regard; MAGE-3 and SSX-4 were frequently expressed alone (Figure 7).
Expression frequency and spontaneous immunogenicity are important criteria for potential targets of cancer vaccines. The expression of tumor antigens is a prerequisite for cancer vaccine therapy, whereas pre-existing autologous immunity clearly demonstrates the potential of the immune system to recognize the antigen, thus making it likely that vaccination might result in immune responses. We examined both of these facets for NY-ESO-1 in patients with gastric cancer in China. A widely expressed and the most immunogenic CT antigen known to date, NY-ESO-1 has been the subject of extensive study in many tumor types (11). There has been relatively little analysis of this cancer antigen in gastric cancer, however, although it was expressed in 7.8% of gastric cancer patients in Japan (12). In the present study, RT-PCR revealed NY-ESO-1 to be expressed in 11.9% of gastric tumors in our samples from China, with the expression frequency being higher in advanced gastric cancers (stages III and IV, 15.9%). We found that NY-ESO-1 and/or LAGE-1 induced a detectable level of antibody in a significant proportion of patients in this group (7/12), demonstrating its well-documented immunogenicity. Six of seven patients had high-titer anti-NY-ESO-1 antibodies, whereas the antibody titer observed in one case was lower than 1:1000. It has previously been shown that patients with demonstrable antibodies to NY-ESO-1 often have simultaneous CD8 and CD4 T cell responses to NY-ESO-1 (13), and studies to evaluate such responses in our patient group are ongoing. Overall, our results suggest that NY-ESO-1 might be considered a candidate component of cancer vaccines for the treatment of a subset of advanced stage gastric cancers.
There is a high level of similarity between NY-ESO-1 and LAGE-1 (94% nucleotide identity and 88% amino acid identity), including shared HLA-A2 binding T cell epitopes (14). Thus, patients with either NY-ESO-1 positive or LAGE-1 positive tumors could be considered eligible for the same vaccine, even in peptide form. We found that LAGE-1 was expressed at a similar frequency to NY-ESO-1 in gastric cancer (16.8% versus 11.9%). Overall, the combined positive rate for NY-ESO-1 and/or LAGE-1 mRNA was 20% of the tumors analyzed, which indeed increases the pool of potential vaccine recipients.
We also evaluated NY-ESO-1 protein expression by immunohistochemical staining with an anti-NY-ESO-1 monoclonal antibody. Our finding of NY-ESO-1 protein expression in some, but not all, of the mRNA-positive tumors is consistent with previous studies (15) and supports the possibility that low-level NY-ESO-1 expression is likely to be undetectable by immunostaining (16). We found a positive correlation between the presence of detectable NY-ESO-1 protein in tumors and the presence of anti-NY-ESO-1 antibodies in the serum, suggesting that antibody stimulation is influenced by protein concentration in the tumors (Table 1). Thus, of 7 cases with detectable NY-ESO-1 protein, 5 had serum antibodies. In contrast, of the five cases that were mRNA-positive but negative by immunostaining, only one had detectable circulating NY-ESO-1 antibody.
The comparatively low frequency of NY-ESO-1 and LAGE-1 expression in gastric tumors found both in this study and in the only earlier study in Japan raises the possibility that a polyvalent CT antigen might be appropriate for this type of cancer. To evaluate the feasibility of this approach, and indeed to gain an indication of whether antigens other than NY-ESO-1 might exist to form the basis of a monovalent but more broadly applicable vaccine, we also examined the expression of a number of other CT antigens in our series of gastric tumor specimens. In all, 74.3% of the cancers we examined were found to express at least 1 of the 11 CT genes analyzed, with 42.6% simultaneously expressing 2 or more CT genes. The most frequently expressed antigen was found to be MAGE-3, with 43% of the specimens analyzed expressing it, and which exhibited a much higher frequency of detection in low-stage tumors than most other antigens. This finding certainly suggests that detailed analysis of the immunogenicity of this antigen in gastric tumors is warranted. The next most frequently expressed antigen was found to be SSX-4, which should also be examined in detail. Nevertheless, no single antigen approached the combined frequency of expression of all the CT antigens tested, and it would appear that for gastric cancer a multivalent approach may indeed be necessary.
The clustering phenomenon observed suggests the possibility that different programs that control the expression of CT genes might be operative in gastric tumors (17). Moreover, the observation that the different CT genes varied in their tendency to exhibit clustered expression suggests that CT genes might be expressed in a cascade that roughly correlates with tumor progression. MAGE-3 and SSX-4, for instance, would appear to be expressed earlier in this cascade since they are most frequently expressed alone. Other CT genes, including NY-ESO-1, LAGE-1, CT7, and CT10, appear to be expressed at a later stage and are most frequently expressed together with other antigens (9). Consistent with this hypothesis, MAGE-3 is the CT gene most frequently found in early stage tumors. From the point of view of cancer vaccine strategy, this may indicate that CT antigens expressed at late stage tumors, such as NY-ESO-1, LAGE-1, CT7, and CT10, could also be used to prevent progression.
Overall, the frequency of CT gene expression and the immunogenicity of NY-ESO-1 is encouraging, and therapeutic vaccines based on one or more CT antigens can be envisaged as playing a future role in the control of gastric cancer. Given the dismal outcome of current therapeutic modalities, the systematic evaluation of this possibility is a high priority.
Materials and methods
Tumor tissue samples and sera
Collection of tissue samples and sera was contingent upon written consent from all patients and was approved by the Hospital Ethics Review Committee. A total of 101 samples of gastric cancer tissue with paired adjacent stomach tissue were obtained during surgical resection at the Beijing Cancer Hospital, China. The resected tissue samples were immediately cut into small pieces, snap frozen in liquid nitrogen, and transferred the following day to -80°C until use. All gastric cancer and adjacent noncancerous stomach tissue samples were confirmed pathologically. Serum was also obtained from each patient prior to surgery.
Extraction of total RNA and RT-PCR analysis
TRIzol™ reagent (Invitrogen, Carlsbad, CA, USA) was used to extract total RNA. First-strand cDNA for the RT-PCR was synthesized from 2.0 µg of total RNA using the SuperScript First-Strand Synthesis System (Invitrogen, Carlsbad, CA, USA). The presence of NY-ESO-1, LAGE-1, MAGE-1 (MAGE-A1), MAGE-3 (MAGE-A3), MAGE-4 (MAGE-A4), SSX-1, SSX-2, SSX-4, SCP-1, CT7, and CT10 mRNA was determined by PCR using AmpliTaq Gold® (Applied Biosystems, Foster City, CA, USA) under the conditions listed in Table 3. All the primers were designed to be intron-spanning, i.e. to cover at least one intron of each gene in order to exclude PCR products that might result from the contaminated genomic DNA. Controls for the integrity and quantity of the cDNA samples were provided by the amplification of glyceraldehyde-3-phosphate dehydrogenase (G3PDH). Additional positive and negative controls were included in each PCR analysis experiment.
Sample PCR products for each individual gene were cloned into the pGEM®-T Easy vector (Promega, Madison, WI, USA) and sequenced using an ABI Prism automated DNA sequencer at the Shanghai Genecore Bioscience Company, China.
E978 generated against NY-ESO-1 was used as the primary antibody (15). Following heat-based antigen retrieval in EDTA buffer (1 mM, pH 8.0), paraffin-embedded specimens were incubated overnight with E978 at a concentration of 1 µg/ml at 4°C. Detection of the primary antibody was performed with the PowerVision™ two-step histostaining reagent (Zhongshan Biotechnology, Beijing, China).
NY-ESO-1 autologous antibodies were detected by ELISA, as described by Stockert et al. (18) with minor modifications. Briefly, 30 µl/well of either 1 µg/ml NY-ESO-1 recombinant protein or 3 µg/ml BSA (background control) in coating buffer (15 mM Na2CO3, 30 mM NaHCO3, pH 9.6) was absorbed to 96-well half-area plates (Costa, Acton, MA, USA) at 4°C overnight. After blocking with 2% BSA in PBS and washing, the plates were incubated for 1 h with 1:25, 1:125, and 1:625 dilutions of patient sera. Peroxidase conjugated rabbit antihuman immunoglobulin (IgA, IgG, and IgM) (Sigma, St. Louis, MO, USA) was used as secondary antibody and the reaction allowed to proceed for 45 min. The plates were then incubated with substrate (OPD) for 30 min and analyzed using an ELISA reader (Bio-Rad Laboratories, Hercules, CA, USA). One positive, eight negative (normal donors), and background (BSA) controls were included in each test.
This study was supported by the Cancer Research Institute (Designated Grant), USA, the New Star Program of Science and Technology of the Beijing Science and Technology Committee (H020821370130), the "211" Program of Peking University, and the Scientific Foundation of Beijing Cancer Hospital, China.
- Received June 7, 2004.
- Accepted August 4, 2004.
- Copyright © 2004 by Jia-Fu Ji