Cancer/testis (CT) antigens are the protein products of germ line-associated genes that are activated in a wide variety of tumors and can elicit autologous cellular and humoral immune responses. CT antigens can be divided between those that are encoded on the X chromosome (CT-X antigens) and those that are not (non-X CT antigens). Among the CT-X antigens, the melanoma antigen gene (MAGE) family, defined by a shared MAGE homology domain (MHD), is the largest. CT-X genes are frequently expressed in a coordinate manner in cancer cells, and their expression appears to be modulated by epigenetic mechanisms. The expression of CT-X genes is associated with advanced disease and poor outcome in different tumor types. We used the yeast two-hybrid system to identify putative MHD-interacting proteins. The MHD of MAGE-C1 (CT7) was used as bait to screen a human testis cDNA library. This study identified NY-ESO-1 (CT6) as a MAGE-C1 binding partner. Immunoprecipitation and immunofluorescence staining confirmed MAGE-C1 interaction with NY-ESO-1, and cytoplasmic co-localization of both proteins in melanoma cells. Co-expression of these two genes was found to occur in cancer cell lines from different origins, as well as in primary tumors (multiple myeloma and non-small cell lung cancer samples). This is the first report of direct interaction between two CT antigens and may be pertinent in the light of the frequently coordinated expression of these proteins.
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.
Cancer/testis antigens (CT antigens) are present in normal germ line tissues, such as testis, placenta, and ovary and in a range of human cancers (1). They constitute a promising class of tumor antigens due to their limited expression in somatic tissues and strong immunogenicity (2). Cancer/testis antigens have been isolated by various methods, including cDNA expression cloning with tumor-reactive CTLs and patients' sera (SEREX), cDNA subtractions, and also by EST or MPSS database mining (2, 3, 4). More than 44 CT genes and/or gene families have been identified to date that can be divided between those that are encoded on the X chromosome (CT-X antigens) and those that are not (non-X CT antigens) (1).
It is estimated that 10% of the genes on the X-chromosome belong to CT-X families (5). CT-X antigens are encoded by genes that have recently undergone rapid evolution and amplification (6) and are usually highly expressed in spermatogonia (7, 8, 9). The CT-X genes are frequently expressed in a coordinate manner in cancer cells (10) and their expression appears to be modulated by epigenetic mechanisms, such as promoter hypermethylation and histone deacetylation (11).
In several different tumor types, the expression of CT-X genes is associated with advanced disease and poor outcome. In non-small cell lung cancer (NSCLC), cancer/testis gene expression, either cumulatively or individually, showed significant associations with advanced tumor type, nodal and pathologic stages, as well as pleural invasion (10). In the same study, the expression of NY-ESO-1 (CTAG1B) and MAGEA3 was each found to be a marker for poor prognosis, independent of confounding variables. Likewise, expression of MAGEA3 in pancreatic ductal adenocarcinoma was found to be a prognostic factor for poor survival (12). In colorectal cancers, NY-ESO-1 gene expression may serve as a marker for local metastasis and advanced disease, while expression of MAGEA4 is significantly associated with vessel emboli (13). Similarly, expression of MAGEA1, MAGEA3, MAGEA4, MAGEC1 (CT-7), and NY-ESO-1 in malignant gammopathies correlates with stage and risk status of disease (14). Co-expression of SSX1, 2, 4, and 5 also correlates with adverse prognosis in multiple myeloma patients (15). Immunohistochemistry (IHC) has demonstrated that 82% of stage III myeloma specimens expressed CT7 and 70% expressed MAGE-A3/6, with CT7 protein expression increasing with advanced stage of disease and higher levels of CT7 and MAGE-A3/6 proteins being correlated with elevated plasma-cell proliferation (16).
Although these data indicate that CT antigen expression might contribute to tumorigenesis, their biological role in both germ line and tumors remains poorly understood. Recent studies have provided some evidence that CT antigens play a role in human tumorigenesis. Through a yeast two-hybrid assay, the transcriptional regulator SKIP was identified as a MAGE-A1 binding partner (17). SKIP connects DNA-binding proteins to other proteins that either activate or repress transcription, and participates in a range of signaling pathways, including those involving vitamin D, retinoic acid, estrogens, glucocorticoids, Notch1 and transforming growth factor-β. In the Notch1 pathway, MAGE-A1 was found to disrupt SKIP-mediated Notch1 signal transduction by binding to SKIP and recruiting histone deacetylase, therefore acting as a transcriptional repressor. Yeast two-hybrid studies using other cancer-related genes as bait have twice pulled out MAGE proteins: MAGE-A11 and MAGE-A4 (18, 19). MAGE-A11 was found to have a role in the regulation of androgen receptor function by modulating its internal domain interactions and was found to have a dual amplifying effect on androgen signaling (18). MAGE-A4 was identified in a search for binding partners of the oncoprotein gankyrin (19). Gankyrin destabilizes the retinoblastoma tumor suppressor, contributing to unscheduled entry into the cell cycle and escape from cell-cycle arrest and/or apoptosis. MAGE-A4 suppresses the oncogenic activity of gankyrin through the action of a peptide that is naturally cleaved from the carboxyl terminus of MAGE-A4 and which induces p53-dependent and p53-independent apoptosis. Overexpression of MAGEA4 in a human embryonic kidney cell line (293 cells) was found to increase apoptosis as measured by apoptotic index and caspase-3 activity, while MAGEA4 silencing using a small interfering RNA approach resulted in decreased caspase-3 activity in a squamous cell lung cancer and in 293/MAGE-A4 cells (20).
The MAGE family consists of a large group of proteins that harbor the MAGE homology domain (MHD), a well-conserved domain of about 200 amino acids (21). The functional role of most family members remains uncharacterized. The cancer/testis antigens of the MAGE family are localized in clusters on the X chromosome (22): MAGEA genes at Xq28, MAGEB genes at Xp21, and MAGEC genes at Xp26-27; these are classified as Type I MAGE genes (21). Type II MAGE genes (MAGED, MAGEE, MAGEF, MAGEG, MAGEH and necdin) are expressed in many normal tissues at various levels (22). The sequence of the MAGE homology domain is a common feature of the Type I and II MAGE gene families; it can be found in mammalian species, as well as in Xenopus, Drosophila, and zebrafish (23). The MHD does not contain any regions of significant homology with other known proteins, but detailed analysis of a number of type II MAGE proteins shows that this domain is an important site for protein–protein interactions (24). The promoters and first exons of the MAGEA genes show considerable variability, suggesting that they are subject to different transcriptional controls.
To investigate further the function of members of the CT antigen family, we used the yeast two-hybrid system to identify putative MHD-interacting proteins. The MHD of MAGE-C1 (CT7) was used as bait to screen a human testis cDNA library. MAGE-C1 is about 800 amino acids longer than the other MAGE proteins and contains a large number of unique short repetitive sequences in front of the MAGE homologous sequence. As a result of this investigation, we identified another CT antigen, NY-ESO-1 (CT6) as a MAGE-C1 binding partner. This is the first report of direct interaction between two CT antigens and may be pertinent in the light of the frequently coordinated expression of these proteins.
Identification of proteins interacting with the MHD of MAGE-C1 (CT7)
The MHD of MAGE-C1 (CT7) was cloned into a bait construct and used to screen a high complexity human testis cDNA library in yeast two-hybrid assays. Human testis was selected because it is the normal tissue in which type I MAGE genes are expressed. Eighteen clones selected from a pool of positive candidates interacted specifically with the MHD upon reconstitution analysis. Subsequent sequence analysis showed that most of these clones contained genomic DNA or non-coding regions of cDNAs (Table 1). We selected the clone containing a 662 bp insert with 100% identity to the human NY-ESO-1 gene (RefSeq Accession No. NM_001327) for further analysis.
Co-precipitation of MAGE-C1 and NY-ESO-1 from cell lysates
To further analyze protein-protein interactions between MAGE-C1 (CT7) and NY-ESO-1 (CT6) in human cells, we prepared total lysates from the melanoma derived SK-MEL-37 cell line that expresses both proteins. Complexes between endogenous MAGE-C1 and NY-ESO-1 proteins were precipitated when both anti-NY-ESO-1 antibodies (NY-41 and E978) were used, as well as with anti-MAGE-C1 monoclonal antibody (CT7.33). The presence of both proteins in the complex was detected by subsequent Western blotting using antibodies against MAGE-C1 and NY-ESO-1. As shown in Figure 1A, MAGE-C1 and NY-ESO-1 proteins co-precipitated when both antibodies (E978 and NY-41) against NY-ESO-1 were used. Similarly, NY-ESO-1 was detected in Western blots by anti-CT7 mAb CT7-33, although with a much lower efficiency. MAGE-C1 and NY-ESO-1 were also co-precipitated when the anti-NY-ESO-1 antibodies were used in two additional cells lines: SK-LC-17, a NSCLC cell line (data not shown) and U266, a multiple myeloma cell line (Figure 1B). In contrast, co-precipitation reactions with a rabbit polyclonal antibody against actin or a monoclonal against influenza hemagglutinin (HA) failed to precipitate MAGE-C1 or NY-ESO-1, indicating that the MAGE-C1/ NY-ESO-1 interaction was specific (Figure 1, panels A and B). Additionally, we used SK-MEL-10, another melanoma cell line that expresses MAGEC1 but does not express NY-ESO-1. As shown in Figure 1C, CT7.33 could efficiently immunoprecipitate MAGE-C1, but neither of the NY-ESO-1 specific antibodies (E978 and NY-41) could precipitate MAGE-C1 in the absence of NY-ESO-1, further indicating that the interaction between MAGE-C1 and NY-ESO-1 is specific.
Subcellular co-localization of NY-ESO-1 and MAGE-C1
We next examined whether the association between MAGE-C1 and NY-ESO-1 resulted in co-localization of the two proteins in vivo using double immunofluorescence staining of SK-MEL-37 cells. As shown in Figure 2, MAGE-C1 and NY-ESO-1 presented a diffuse distribution throughout the cytosol in exponentially growing SK-MEL-37 cells. MAGE-C1 was also distributed in the cell nuclei. MAGE-C1 and NY-ESO-1 appear to co-localize to the cytoplasm of these cells.
Analysis of MAGEC1 and NY-ESO-1 expression in cultured cells and tumors
The frequency of expression of MAGEC1 and NY-ESO-1 in 88 human cancer cell lines representing a broad range of solid and hematological malignancies was examined by RT-PCR (Table 2). MAGEC1 transcripts were detected in 42 of 88 (48%) cancer cell lines. NY-ESO-1 transcripts were detected in 38 (43%) cancer cell lines. In 74% of the cases, there were concordant results (either negative or positive) between expression of MAGEC1 and NY-ESO-1. In 28 cell lines (32%), both genes were found to be co-expressed.
In general, these preliminary results suggested that NY-ESO-1 expression may coincide with MAGEC1 expression in many cancer cell types. We next decided to examine primary tumors from cohorts of multiple myeloma and NSCLC samples for the co-expression of MAGEC1 and NY-ESO-1. As shown in Table 3, co-expression of MAGEC1 and NY-ESO-1 also occurs in primary tumors.
Thirty-nine multiple myeloma samples were tested by RT-PCR for both MAGEC1 and NY-ESO-1 expression. Of these, 13 samples were positive for NY-ESO-1 (33%) and 30 positive for MAGEC1 (77%). A total of 456 NSCLC samples were tested by RT-PCR for expression of both MAGEC1 and NY-ESO-1. There were 120 samples positive for NY-ESO-1 (26%) and 101 samples positive for MAGEC1 (22%). From these values, assuming that expression of these antigens is independent, the expected frequency of NSCLC samples positive for both MAGEC1 and NY-ESO-1 is 6%, i.e., about 26 samples. The observed frequency of NY-ESO-1 positive and MAGEC1 positive samples is 13% (59 samples), i.e., more than twice what was expected (P < 0.001).
In this study we report the first evidence of specific protein-protein interaction between two distinct CT-X antigens via the MHD of MAGE-C1. This finding suggests the possibility that cancer/testis antigens exert and/or regulate their activities through specific interactions with other CT-X antigens. However, the positive interaction between MAGE-C1 and NY-ESO-1 could be due to bridging by other proteins and in this case MAGE-C1 and NY-ESO-1 may not directly bind to each other, but may be part of a multi-protein complex.
In normal testis, the CT-X genes are generally expressed in germ cells, particularly spermatogonia, but their function during male germ cell development is not known. Insights into the function of these genes may provide links between spermatogenesis and tumor growth and may identify novel therapeutic targets amenable to immunologic or pharmacologic strategies.
The expression of different CT-X antigens in tumors is positively correlated, suggesting that these genes may be activated by a common mechanism. Both promoter demethylation and non-demethylation dependent induction are responsible for the ectopic expression of several CT antigens (11). CT antigen expression in tumors may be the result of the activation of a coordinated gene-expression program, rather than of independent events (1).
The first evidence of the function of MAGE family members came from recent studies suggesting important roles for type II MAGE proteins via transcriptional regulation in cell cycle control and apoptosis. Clues are also emerging showing that type I MAGE genes have roles as transcriptional regulators. MAGE-A1 was found to bind SKIP and to recruit histone deacetylases, acting as a potent transcriptional repressor (17). MAGE-A11 was identified as an androgen receptor (AR) coregulator that increases AR activity by modulating the AR interdomain interaction (18). MAGE-A2 was shown to interact and repress p53 activity by recruiting transcription repressors to p53 transcription sites, conferring resistance to apoptosis (25).
Because in most of these cases regions including the MHD were involved in protein-protein interactions, we decided to use the MAGE-C1 MHD as bait in a yeast two hybrid screen. The physical interaction between the MHD of endogenous MAGE-C1 and NY-ESO-1 could be confirmed by co-immunoprecipitation in a melanoma cell line, and was also observed in multiple myeloma and NSCLC cells. Both proteins were shown to co-localize to the cytoplasm. Moreover, analysis of expression in a broad range of human tumor cell lines also suggests that coordinate expression of MAGEC1 and NY-ESO-1 may be a common event in a subset of cancers. Together, these data identify a novel interaction between cancer/testis antigens. Further studies to identify the functional consequence of this interaction may indicate whether the interaction of these two proteins plays a role in tumorigenesis.
Finally, CT-X gene products have been extensively investigated as targets for cancer vaccines. Insight into the interactions between these proteins may provide a rational basis for polyvalent vaccine formulations that may target specific tumor-promoting regulatory pathways at multiple levels or that target multiple pathways. The understanding of the biology of these proteins in both normal and cancer cells is critical to the development of effective vaccine therapy. It may also lead to other classes of targeted therapy that may be combined with vaccines in order to neutralize the tumorigenic effects of CT-X proteins.
Materials and methods
Yeast two-hybrid screening
The NH2-terminal portion of MAGE-C1 (amino acids 775-1143, nucleotide 2322-3429) encompassing the MAGE homology domain (amino acids 902-1029, nucleotides 2704-3087) of MAGE-C1 (1107 bp) was cloned into the pGBKT7 vector (Clontech, Palo Alto, CA) to generate a fusion protein downstream of the Gal4 DNA binding domain (BD). This plasmid was then used as bait to screen a high complexity human testis cDNA library (Invitrogen, Carlsbad, CA), which was cloned downstream of the Gal4 activation domain (GalAD). The yeast strain AH109 was transformed with the bait plasmid, then with the cDNA library. The transformants were plated on medium lacking leucine, tryptophan and histidine (L, T, H) in the presence of 75 mM 3-amino-1,2,4-triazole (3AT) for up to 3 weeks at 30˚C. Colonies were picked and replated on L, T, H agar plates. Those that grew in the second plate were plated in medium lacking leucine, tryptophan and adenine (L, T, A). The positive colonies were then subjected to several rounds of culture in SC without tryptophan (T) to eliminate the bait plasmid. Each Trp+ clone was tested again for activation of the reporter genes to eliminate those that transactivate the reporter gene in the absence of the bait plasmid. Finally, the plasmid DNA was isolated from the positive yeast clones, amplified in E. coli, and analyzed by automated DNA sequencing. All inserts were identified by homology searching with the NCBI BLAST program (26). These plasmids were also used to reconstruct the in vivo interaction by transforming them back into the original yeast strain containing the bait plasmid.
All cell lines used in this study were obtained from the cell culture bank of the New York Branch of the Ludwig Institute for Cancer Research. They were maintained in RPMI medium containing 10% fetal bovine serum (FBS) and non-essential amino acids.
Western blot analysis
Cell extracts were prepared in RIPA buffer (25 mM Tris-HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS and protease inhibitor Complete™ tablet) and subjected to five cycles of sonication. Fifty microliters of the protein was mixed with an equal volume of 2x loading buffer (125 mM Tris-HCl pH 6.8, 4% SDS, 10% glycerol, 0.006% bromophenol blue, 2% β-mercaptoethanol), incubated at 95˚C for 3 min, and loaded onto 10% SDS Bis-Tris gels (Invitrogen, Carlsbad, CA). After electrophoresis, proteins were transferred to nitrocellulose membranes. The membranes were blocked by incubation in PBST (PBS, 0.1% Tween 20) with 3% bovine serum albumin (BSA) for 1 h, and then incubated with the primary antibody overnight at 4˚C in PBST with 1% BSA. After washing four times in PBST, the membranes were incubated either with peroxidase-conjugated anti-rabbit or anti-mouse IgG (Jackson Immunoresearch, Bar Harbor, ME) for 1 h at room temperature. Antibody binding was detected using the Western Lightening Chemiluminescence Reagent Plus system (Perkin Elmer, Emeryville, CA). The antibodies used were: a rabbit polyclonal anti-full length NY-ESO-1 (NY-41), a monoclonal anti-NY-ESO-1 (E978) (27), a monoclonal anti-CT7 (CT7.33) (8), a rabbit polyclonal anti-actin (20-33, Sigma-Aldrich, St. Louis, MO) and an anti-influenza hemagglutinin (HA-7, Sigma-Aldrich, St. Louis, MO) antibody.
Cells were lysed in IP buffer (50 mM Tris-Cl pH 7.4, 0.15 M NaCl, 2 mM EDTA, 1% NP-40), containing protease inhibitors (Protease Inhibitors Cocktail, Roche, Indianapolis, IN). Five hundred micrograms to one milligram of total lysate were incubated at 4˚C for 1 h in a rotator with the polyclonal or monoclonal antibodies, or with normal rabbit serum or mouse IgG1 as negative controls. Immunocomplexes were precipitated with 150 µl of 10% protein A/G sepharose beads (Pierce Biotechnology, Rockford, IL) overnight at 4˚C. After four washes in IP buffer, beads were boiled in 60 µl of 2x loading buffer (125 mM Tris-HCl pH 6.8, 4% SDS, 10% glycerol, 0.006% bromophenol blue, 2% β-mercaptoethanol). Immunoprecipitated proteins were resolved in 12% SDS-PAGE gels, followed by Western blotting.
Cells were grown on Lab-Tek II 4-well chamber slides (Nalge Nunc, Naperville, IL) until they reached 70% confluence, rinsed twice with PBS and fixed with 3% paraformaldehyde in PBS (138 mM NaCl, 2.7 mM KCl, pH 7.4) for 10 min at room temperature, permeabilized with 0.2% Triton X-100 in PBS for 2 min and blocked with 5% goat serum in PBST for 30 min at room temperature. The cells were then incubated with the primary antibodies (NY-41 and CT7.33) overnight at 4˚C and, following three washes in PBST, were incubated with a mixture of Alexa Fluor 488 and 633 labeled secondary antibodies (Molecular Probes, Eugene, OR). After washing, the slides were incubated with 0.02 µg/ml DAPI (Sigma-Aldrich, St. Louis, MO) to visualize the nuclei. The slides were mounted in Vectashield (Vector Laboratories, Burlingame, CA) and analyzed under an epifluorescence microscope.
RT-PCR analysis of tumor cells
Total RNA was prepared from tissues or cell line pellets following homogenization by the guanidinium isothiocyanate method, followed by CsCl gradient centrifugation. Alternatively, the Ribopure kit (Ambion, Austin, TX) or Trizol reagent (Invitrogen, Carlsbad, CA) were used according to manufacturer's instructions. Total RNA (2 µg) was reverse-transcribed with 200 units Moloney murine leukemia virus reverse transcriptase (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions, in the presence of 2 µg random hexamers (Applied Biosystems, Foster City, CA), 20 units RNaseOUT (Invitrogen), and 5 mmol/l DTT in a total volume of 20 µl. For individual PCR reactions, 250 ng of cDNA were amplified with gene-specific oligonucleotides (2 ng per 25 µl reaction) in the presence of 1 unit AmpliTaq Gold (Applied Biosystems, Foster City, CA) and 5 µmol/l of each dNTP (Applied Biosystems, Foster City, CA). The MAGEC1, NY-ESO-1, MAGEA1 and MAGEA3 specific primers used for PCR amplification were the same as those used by Gure et al., 2005 (10) and van Baren et al. (28).
All patients provided written informed consent prior to our study. Four hundred and fifty six tumor tissue specimens sampled from patients with NSCLC who underwent surgery between 1991 and 2004 at the Department of Cardio-Thoracic Surgery, Weill Medical College of Cornell University were included in this study. The study was approved by the Institutional Review Board of Weill Medical College of Cornell University. Thirty nine consecutive, previously untreated multiple myeloma (MM) patients seen in the MM outpatient service of the Discipline of Hematology and Hemotherapy, UNIFESP/EPM, São Paulo, Brazil, between June 2002 and May 2006 were studied. Cells were collected from patients with multiple myeloma by bone marrow aspiration.
HJC was funded by a Clinical Investigation Award from the Cancer Research Institute. VCCA is recipient of a fellowship from Fundação de Amparo à Pesquisa do Estado de São Paulo, Brazil.
- Received September 26, 2006.
- Accepted October 10, 2006.
- Copyright © 2006 by Andrew J. G. Simpson