Abstract
Genes MAGE, BAGE, GAGE, and LAGE-1/NY-ESO-1 code for antigens that are recognized on melanoma cells by autologous CTLs. Because the pattern of expression of these genes results in the presence of antigens on many tumors of various histological types and not on normal tissues, these antigens qualify for cancer immunotherapy. To identify new genes with tumor-specific expression, we applied a cDNA subtraction approach, i.e., representational difference analysis, to a human sarcoma cell line. We obtained two cDNA clones that appeared to be tumor specific. The corresponding genes were named SAGE and HAGE because they have the same pattern of expression as genes of the MAGE family. SAGE encodes a putative protein of 904 amino acids and shows no homology to any recorded gene. Like the MAGE-A genes, it is located in the q28 region of chromosome X. Expression of gene SAGE was observed mainly in bladder carcinoma, lung carcinoma, and head and neck carcinoma but not in normal tissues, with the exception of testis. Gene HAGE, which is located on chromosome 6, encodes a putative protein of 648 amino acids. This protein is a new member of the DEAD-box family of ATP-dependent RNA helicases. Gene HAGE is expressed in many tumors of various histological types at a level that is 100-fold higher than the level observed in normal tissues except testis. Because of this tumor-specific expression, genes SAGE and HAGE ought to encode antigens that could be useful for antitumoral therapeutic vaccination.
INTRODUCTION
Several studies have shown that by cultivating irradiated tumor cells with autologous lymphocytes, it is possible to obtain responder cell populations that display a cytotoxic response against tumor cells (1) . The CTL clones derived from such responder populations have been found to recognize several antigens (2) . The genes coding for these antigens have been identified by transfection and detection of the transfectants by the CTLs. A first important class of tumor antigens recognized by CTLs is encoded by genes that are activated in tumors. These genes belong to the MAGE, BAGE, GAGE, and LAGE-1/NY-ESO-1 families. They are all expressed in tumors of different histological types but not in normal tissues, except for spermatogenic cells, and for some of them, placenta (3–8) . A second category contains differentiation antigens encoded by genes expressed in normal melanocytes and in melanoma cells, such as tyrosinase, Melan-A/Mart-1, gp100, and gp75 (9–12) . A third class constitutes antigens produced by point mutations in genes that are expressed ubiquitously, e.g., MUM-1, cyclin-dependent kinase 4,β -catenin, and HLA-A2 (13–16) . Finally, there are antigens derived from genes overexpressed in tumors relative to normal cells, such as HER-2/neu and PRAME (17, 18) .
The antigens encoded by genes that are expressed only in tumors and in germ-line cells appear to be strictly tumor specific, because the spermatogenic cells that express these genes do not express HLA molecules and are therefore incapable of presenting antigens to T cells (19, 20) . Because these genes are expressed in a large proportion of tumors of various histological types, they appear to be excellent potential sources of antigens for cancer immunotherapy. Clinical trials involving MAGE antigens are ongoing, and tumor regressions have been observed (21–24) .
For the purpose of finding new genes that present the same pattern of expression as the MAGE, BAGE, GAGE, and LAGE-1/NY-ESO-1 genes, we have recently applied subtraction of cDNA from a tumor with cDNA from a panel of normal tissues. This approach has led to the identification of a new member of the MAGE family, gene MAGE-C1 (25) , as well as LAGE-1, a cancer germ-line gene mentioned above (7) . We report here that a similar approach, this time applied to a sarcoma cell line, led to the identification of two new genes that have a pattern of expression similar to the MAGE-type genes. These genes are not homologous to the MAGE, BAGE, GAGE, and LAGE-1/NY-ESO-1 families.
MATERIALS AND METHODS
Cell Lines.
Rhabdomyosarcoma cell line LB23-SAR was derived from patient LB23 and cultured in Iscove’s medium (Life Technologies, Inc., Grand Island, NY) containing 10% FCS (Life Technologies) and supplemented with l-asparagine (36 mg/l), l-arginine (116 mg/l), and l-glutamine (216 mg/l).
Representational Difference Analysis.
Total RNA was isolated from normal tissues (uterus, breast, colon, and heart) and from cell line LB23-SAR by the guanidine isothiocyanate/cesium chloride procedure (26) . Poly(A)+ RNA was prepared by binding to oligo-dT cellulose (mRNA purification kit; Pharmacia Fine Chemicals, Piscataway, NJ), starting from 500 μg of rhabdomyosarcoma LB23-SAR total RNA and from 455 μg of a mixture of human uterus, breast, heart, and colon total RNA. About 4 μg of each poly(A)+ RNA were used to synthesize double-stranded cDNA (cDNA Synthesis Module; Amersham Life Science, Buckinghamshire, United Kingdom). Two micrograms of each cDNA were then digested with DpnII (New England Biolabs, Beverly, MA). The digested cDNAs were ligated to R-Bgl adapters and PCR-amplified to generate the rhabdomyosarcoma (tester) and uterus-breast-heart-colon (driver) representations, as described by Hubank and Schatz (27) . We performed three rounds of subtractive hybridization and selective amplification with the J-Bgl and N-Bgl adapters, as described (27) .
Cloning of Difference Products.
The second and third difference products were digested with DpnII, separated from the adapters on a 1.2% agarose gel, and purified using the QIAEX II Gel Extraction kit (Qiagen, Westburg, Leusden, the Netherlands). They were ligated to dephosphorylated BamHI ends of pTZ18R (Pharmacia). DH5αF′IQ bacteria (Life Technologies, Gaithersburg, MD) were transformed by electroporation with one-fifth of the ligation product.
Sequencing and Sequence Comparison.
Plasmid DNA was prepared by the boiling minipreparation procedure (28) . Sequencing was performed with the T7 Sequencing kit (Pharmacia LKB) and [α-35S]dATP (>1000 Ci/mmol; Amersham), and reaction products were separated on a 2010 Macrophor manual sequencer (Pharmacia LKB). When long read lengths were required, we performed the sequencing reactions with the BigDye Terminator Cycle Sequencing kit (PE Applied Biosystems, Warrington, United Kingdom). The reaction products were separated on the ABI PRISM 310 Genetic Analyser (Perkin-Elmer).
Sequence homology searches were performed in the databases provided by the National Center for Biotechnology Information (Bethesda, MD) using the BLAST program (29) . Alignments were performed with the GeneWorks computer program (Intelligenetics, Mountain View, CA). For DNA, we used parameters 30-2-10-4000.
RT-PCR Analyses.
The expression of cDNA clones isolated from the difference products was analyzed by RT-PCR 3 . RNA were extracted from tumor cells and normal tissues and reverse transcribed, as described previously (30) . RNA samples from 5-aza-2′-deoxycytidine-treated cells were obtained as described (31) . Twenty-four pairs of PCR primers were derived from the sequences of the fragments of cDNA of the enriched library. The sequence of these primers are available from the authors. PCR amplification of cDNA was performed for 30 or 35 cycles, whereas amplification of β-actin was performed for 21 cycles with primer 5′-GGCATCGTGATGGACTCCG-3′ (exon 3, sense) and primer 5′-GCTGGAAGGTGGACAGCGA-3′ (exon 6, antisense). PCR products were visualized on 1.7% agarose gels stained with ethidium bromide.
For the analysis of SAGE expression, the cDNA produced from 50 ng of total RNA was PCR amplified in a TRIO-Thermoblock (Biometra, Göttingen, Germany) for 30 cycles of 1 min at 94°C, 2 min at 57°C, and 2 min at 72°C with primers sdph3.10S and sdph3.10A.
To analyze HAGE expression, PCR amplification was performed for 30 cycles of 1 min at 94°C, 2 min at 57°C, and 2 min at 72°C with primers sdp3.8.8 and sdp3.8.9.
PCR Analysis of Radiation Hybrids.
The GeneBridge 4 Radiation Hybrid Panel (Research Genetics, Inc., Huntsville, AL) was used to map the SAGE and HAGE genes (32) . Twenty-five ng of genomic DNA from each of the 93 radiation hybrid clones were PCR amplified with primers sdph3.10S (5′-TGTACCTCTTCAAGCAAAAT-3′) and sdph3.10A (5′-GTGACCCACCAGTTACAGTA-3′) that are specific for gene SAGE. PCR amplification was performed for 30 cycles of 1 min at 94°C, 2 min at 57°C, and 2 min at 72°C. PCR results were submitted for analysis to the web site of the Whitehead Institute for Biomedical Research. 4 Positioning of markers from the Whitehead RH map on the chromosome cytogenetic map was performed via the Genome Database web site. 5 The same panel was used to map gene HAGE with primers sdp3.8.8 (5′-TATTCTTCAGATTGACGAAG-3′) and sdp3.8.9 (5′-CCTTTCAATGTTATCCTGAG-3′) using identical PCR conditions.
Construction and Screening of the cDNA Library.
We used a cDNA library derived from human testis sample LB451 to search for the complete cDNA sequences of genes SAGE and HAGE. Poly(A)+ RNA was extracted from testis cells using the FastTrack mRNA extraction kit (InVitrogen, San Diego, CA). mRNA was converted to cDNA with the SuperScript Choice System (Life Technologies, Inc.) using an oligo-dT primer containing a NotI site at its 5′ end. cDNAs were then ligated to BstXI adaptors and digested with NotI. After size fractionation, the cDNAs were cloned into the BstXI and NotI sites of plasmid pcDNAI/Amp. Recombinant plasmids were electroporated into Escherichia coli DH5αF′IQ and selected with ampicillin (50 μg/ml). The library was divided into three fractions: A (54,000 independent clones); B (150,000 clones); and C (300,000 clones). Recombinant bacteria (250,000) from fractions B and C were screened by colony hybridization with a PCR-amplified fragment of genes SAGE or HAGE, using the same primers as those used for the analysis of radiation hybrids, and labeled with[α -32P]dCTP (3000 Ci/mmol). Plasmid DNA from one positive clone was amplified and purified for further analysis.
Northern Blot Analysis.
Four μg of poly(A)+ RNA were separated by formaldehyde agarose gel electrophoresis, transferred to a nylon membrane by capillary transfer, and fixed to the membrane by heating at 80°C for 2 h. Hybridization with the 1.95-kb insert of SAGE cDNA or the 2.3-kb insert of HAGE cDNA probes was performed overnight at 60°C in 10% dextran sulfate, 1 m NaCl, 1% SDS, and 100 μg/ml denatured herring sperm DNA. Final washing of the membrane was done in 0.2× SSC, 0.1% SDS for 10 min at 60°C. Autoradiography was performed overnight using BioMax MS film (Kodak). The same membrane was stripped and hybridized with a 0.6-kb probe for β-actin obtained in similar conditions. Autoradiography was performed for 2 h.
Southern Blot Analysis.
Ten μg of genomic DNA were digested with restriction enzymes BamHI, EcoRI, HindIII, and PstI. They were separated by agarose gel electrophoresis, transferred to nylon membranes by the capillary transfer method, and fixed by heating at 80°C for 2 h. Hybridization with theα -32P-radiolabeled PCR probes was performed in 5× SSC, 5× Denhardt’s solution, 0.1% SDS, and 100 μg/ml denatured herring sperm DNA for 16 h at 65°C. Final washing of the membranes was done in 2× SSC, 0.1% SDS for 20 min at 65°C. Autoradiography was performed overnight using BioMax MS film (Kodak).
Rapid Amplification of the 5′ cDNA End of Gene SAGE.
The amplification of the 5′ cDNA end of SAGE was performed with the 5′ rapid amplification of cDNA ends system, version 2.0 (Life Technologies, Inc.). We used antisense primers Gsp1 (5′-CTGGTACTCATGGGTGGAAT-3′), Gsp2 (5′-GTTAATAAGCTCTGGTGTAA-3′), and GspN (5′-CATATCAGCTGTTGCCAAAT-3′). The amplified products were cloned using the Original TA Cloning kit (InVitrogen). The resulting clones were screened with the labeled sdph3.10.1 oligonucleotide (5′-ACTCACAATGTCTGTGAAGA-3′). We sequenced 24 positive clones with primer GspN.
RESULTS
Genes Overexpressed in a Rhabdomyosarcoma.
To obtain a cDNA library enriched for tumor-specific sequences, we used the representational difference analysis method (27) . Rhabdomyosarcoma LB23-SAR was used as the source of tester cDNA to be subtracted with driver cDNA that was a mixture of human uterus, breast, heart, and colon cDNA. Tester and driver were subjected to PCR amplification. The amplified tester was then ligated to a pair of adapters and hybridized to a large excess of driver. The hybridization product was amplified by PCR using the tester-specific adapters as primers. Under these conditions, only tester-tester homoduplexes, corresponding to LB23-SAR specific sequences, were amplified exponentially. This first difference product was submitted to two additional rounds of subtraction and amplification, generating the second and third difference products, which were cloned.
Forty-two different sequences of the second and third difference products were analyzed. Twenty-two sequences showed ubiquitous expression, 12 of these coding for mitochondrial or ribosomal components or for known abundant proteins involved in cell proliferation. Eighteen clones were transcribed in some normal tissues but had a level of expression in LB23-SAR cells that was significantly higher. Among those, six were not recorded in databases. Finally, two sequences that we named SAGE and HAGE appeared to be highly expressed in testis and in LB23-SAR but were completely or nearly completely silent in normal tissues. Because these genes had a pattern of expression similar to that of the MAGE, BAGE, GAGE, and LAGE-1/NY-ESO-1 genes, we felt that they had the potential to code for tumor-specific antigens, and we characterized them fully.
Gene SAGE.
We analyzed the expression pattern of gene SAGE in tumor samples and normal tissues. The PCR product was 171 bp when derived from cDNA templates. It was easily distinguished from the 750-bp product derived from genomic DNA, eliminating the risk of false positives attributable to DNA contamination of the RNA. SAGE was expressed in many samples of bladder, lung, and head and neck carcinomas (Table 1 ⇓ ). In contrast, normal tissues were found to express <0.1% of the amount expressed by LB23-SAR. The only positive tissue was testis, as has been observed for the other MAGE-type genes (Fig. 1 ⇓ ). The expression of several MAGE genes can be induced with 5-aza-2′-deoxycytidine in normal and tumoral cell types that do not normally express these genes (25, 31) . We examined whether gene SAGE could also be activated in nonexpressing cells after treatment with this agent. We found that gene SAGE was induced in fibroblasts and in tumoral cell lines of various types (data not shown).
RT-PCR analysis of gene SAGE expression. Total RNA from the indicated normal tissues or tumor samples was submitted to 30 cycles of RT-PCR amplification with SAGE-specific primers. The 171-bp PCR products were visualized on a 1.7% agarose gel stained with ethidium bromide.
Expression of gene SAGE tested by RT-PCR in samples of normal tissues and tumors
Total RNA was converted to cDNA; the cDNA corresponding to 50 ng of total RNA was then amplified by PCR with primers sdph3.10S and sdph3.10A (Fig. 3) ⇓ and 0.625 unit of TaKaRa Taq for 30 cycles (94°C for 1 min, 57°C for 2 min, 72°C for 2 min).
To obtain a full-length cDNA for gene SAGE, we screened a testis library. A 1950-bp cDNA clone hybridized with the 171 bp PCR product that we used as a probe. This insert was sequenced and comparison with sequences available in databases revealed that it was unrelated to any known sequence. This 1.95-kb insert was used as a probe and was found to hybridize with a single messenger of about 3.5 kb on a Northern blot prepared with poly(A)+ RNA from tumor cell lines and normal tissues (Fig. 2 ⇓ ) indicating that we had not obtained a full-length cDNA clone. To obtain the full-length cDNA, we used the PCR amplification of the 5′ cDNA end. This led to a 3069 bp sequence. The largest open reading frame comprised 2712 bp and encoded a putative protein of 904 amino acids; this protein is composed of 15 repeated motifs of 47 amino acids (Fig. 3 ⇓ ).
Northern blot hybridized with the 1.95-kb insert of SAGE cDNA. Each lane was loaded with 4 μg of poly(A)+ RNA from tumor cell lines.
Nucleotide and amino acid sequences of gene SAGE. Primers are underlined. The repetitive motif of 47 amino acids is in bold.
In a Southern blot analysis, the labeled 1.95-kb insert hybridized with several bands (Fig. 4 ⇓ ) suggesting that SAGE is a member of a family of genes. Using the GeneBridge 4 Radiation Hybrid Panel, we established that gene SAGE was located between markers WI-6213 and WI-5285 on the chromosome X, in the q28 region, like the MAGE-A and LAGE-1/NY-ESO-1 genes (Fig. 5 ⇓ ).
Southern blot hybridized with the 1.95-kb insert of SAGE cDNA. Each lane was loaded with 10 μg of genomic DNA from LB23-SAR cell line digested with the indicated restriction enzymes.
Localization of gene SAGE and other tumor-specific genes on the human X chromosome. The diagram is a schematic representation of the X chromosome, with its G-banding pattern. Left, name of the cytogenetic bands.
Gene HAGE.
The expression of gene HAGE was tested in RT-PCR assays with a large panel of tumor samples and normal tissues. We used PCR primers that enabled us to distinguish the amplification product derived from cDNA templates (431 bp) from that of genomic DNA (∼2 kb). With the exception of testis, which showed a high level of expression, a very low level of expression was observed with normal tissues (Fig. 6 ⇓ ). The level of expression of gene HAGE in normal tissues, including skeletal muscle, was ∼0.2% of the level of expression found in the LB23-SAR sarcoma cell line. When we analyzed tumor samples, we found that gene HAGE was expressed in a large fraction of tumor samples of various histological types, well above the level observed in normal tissues (Table 2 ⇓ ). About 5% of the positive tumor samples tested showed a level of expression >10% of the level of LB23-SAR, and 7% showed a level of expression between 1 and 10%. Like gene SAGE, the expression of gene HAGE could also be induced in cells treated with 5-aza-2′-deoxycytidine (data not shown).
RT-PCR analysis of gene HAGE expression. Total RNA from the indicated normal tissues or tumor samples were submitted to 30 cycles of RT-PCR amplification with HAGE-specific primers. The 431-bp PCR products were visualized on a 1.7% agarose gel stained with ethidium bromide.
Expression of gene HAGE tested by RT-PCR in normal tissues and tumoral samples
Total RNA was converted to cDNA; the cDNA corresponding to 50 ng of total RNA was then amplified by PCR with primers sdp3.8.8 and sdp3.8.9 (Fig. 8) ⇓ and 0.625 unit of TaKaRa Taq for 30 cycles (94°C for 1 min, 57°C for 2 min, 72°C for 2 min).
To obtain a full-length HAGE cDNA, we used colony hybridization with a HAGE probe to screen the same testis cDNA library that we used for gene SAGE. The only cDNA clone hybridizing with the probe contained a 2.3-kb insert. This insert used as a probe hybridized with a single messenger of ∼2.2 kb on a Northern blot (Fig. 7 ⇓ ) suggesting that the isolated cDNA was full length. The largest open reading frame comprised 1944 bp and encoded a putative protein of 648 amino acids (Fig. 8 ⇓ ).
Northern blot hybridized with the 2.3-kb insert of HAGE cDNA. Each lane was loaded with 4 μg of poly(A)+ RNA from tumor cell lines.
Nucleotide and amino acid sequences of gene HAGE. The boxes correspond to motifs typical of the DEAD-box family of helicases: the D-X-X-X-X-A-X-X-X-X-G-K-T sequence at position 281–293 is the typical A-motif of ATP-binding proteins; the D-E-A-D box at position 394–406 represents a special version of the B-motif of ATP-binding proteins; the S-A-T motif at position 427–429 is also conserved in all DEAD-box proteins, but no function can yet be attributed to this motif; and, finally, the H-R-I-G-R motif at position 573–577 seems to be involved in the polynucleotide binding and/or unwinding activity.
Comparison with sequences available in databases revealed a homology with p68, an ATP-dependent RNA helicase that is a member of the DEAD-box proteins (33) . They share several conserved amino acids, including the D-E-A-D motif, which provides their name (Fig. 8) ⇓ . All of the motifs that are present in all of the members of the DEAD box family are conserved in the HAGE protein.
We used the 2.3-kb insert for Southern analysis of human genomic DNA, and it hybridized with several fragments of DNA obtained with four different restriction enzymes (Fig. 9 ⇓ ). It is probable that the probe hybridizes with other members of the DEAD-box family because their sequences are well conserved.
Southern blot hybridized with the 2.3-kb insert of HAGE cDNA. Each lane was loaded with 10 μg of genomic DNA from the LB23-SAR cell line digested with the indicated restriction enzymes.
The chromosomal location of gene HAGE was established by PCR analysis of DNA samples of the radiation hybrid panel. Gene HAGE was found to be located on the long arm of chromosome 6, between markers GATA11B08 and D6S284 in the region q12-q13.
DISCUSSION
We have identified two new genes, SAGE and HAGE, with patterns of expression similar to that of genes MAGE, BAGE, GAGE, and LAGE-1/NY-ESO-1, i.e., a strong expression in many tumor samples and in testis and either no expression or very little expression (at least 100 times lower) in other normal tissues (3–8) . This confirms the usefulness of representational difference analysis for the identification of genes with tumor-specific expression.
As already observed for other MAGE-type genes, SAGE and HAGE expression was induced by 5-aza-2′-deoxycytidine, suggesting that demethylation plays a role in the activation of these genes in tumors. There is evidence that, in cells containing the relevant transcription factors, DNA methylation can on its own be effective for gene silencing (34, 35) . One mechanism by which methylation can affect gene expression is that binding of transcription factors to their target sequences can be inhibited by methylation of CpGs sites located in these sequences (36) . Another mechanism for transcriptional silencing induced by methylation is through the modification of the structure of the chromatin. Methylation of DNA helps to stabilize the chromatin in an inactive configuration and therefore inhibits gene transcription (37) . This inactive state of the chromatin is associated with underacetylated histones, suggesting that DNA methylation and histone deacetylation are linked. In line with this hypothesis, it has been shown recently that histone deacetylases form complexes with methyl-binding proteins such as MeCP2, MBD2, and MBD3 (38–40) .
We suggested previously that the activation of MAGE genes in tumor cells results from the genome-wide demethylation that occurs in these cells (31) . However, tumor cells with demethylated genomes do not activate all of the MAGE-related genes (41) . This is probably attributable to random demethylation and chromatin modification, which vary from cell to cell and lead to different patterns of expression of MAGE-related genes, even in tumors of the same histological type. We report here that MAGE and SAGE genes are both expressed in head and neck carcinomas, bladder carcinomas, and epidermoid carcinomas of the lung, but that in melanomas MAGE genes are expressed whereas SAGE is rarely expressed. MAGE-1 and MAGE-4 were found to be regulated by ubiquitous transcription factors (42, 43) , but for other genes, such as SAGE, we cannot exclude that some of the factors regulating their expression are tissue specific and hence are present only in certain types of tumors.
Gene SAGE is unrelated to any known sequence and appears to be a member of a new family of several genes. It is not expressed at all in normal tissues, with the exception of testis. The expression of MAGE and LAGE genes in testis appears to be restricted to germ-line cells (19) , 6 and we consider it likely that this will apply to SAGE also. Like MAGE-A genes (44) , gene SAGE maps to the q28 region of chromosome X. It is noteworthy that several other genes that are expressed in a wide array of tumors and in male germ-line cells map to chromosome X. In addition to MAGE-A and LAGE genes, this is the case for P1A in the mouse and for MAGE-B, MAGE-C1, GAGE, and SSX2 genes in humans (25, 45–47) . The significance of this localization is not clear.
The second gene that we identified was named HAGE because it encodes a protein that shows 55% similarity with the human p68 protein, a DEAD-box protein whose ATP-dependent RNA helicase activity has been demonstrated (48, 49) . The protein encoded by gene HAGE seems to be a new member of the family of DEAD-box proteins (33) , which contain the highly conserved Asp-Glu-Ala-Asp (D-E-A-D) motif. These proteins are involved in many aspects of RNA metabolism, spermatogenesis, embryogenesis, and cell growth. The amino acids that are highly conserved in all of the DEAD-box proteins are also present in protein HAGE, suggesting that HAGE may also be an ATP-dependent RNA helicase. Gene HAGE is not located on chromosome X but on chromosome 6.
For MAGE-1, it has been shown that tumor cell lines could be lysed by a relevant CTL clone if the level of expression of this gene exceeded the threshold of 10% of that found in the MZ2-MEL reference tumor cell line, i.e., more than three RNA molecules/cell (50) . We have indications (data not shown) that the level of expression of gene HAGE in cell line LB23-SAR is ∼10 times lower than the level of expression of MAGE-1 in MZ2-MEL. This level may be sufficient to produce antigenic peptides, but it will only be possible to test this once a CTL clone that recognizes a peptide encoded by gene HAGE has been obtained. This CTL clone could then also be used to test the presence of antigens on normal tissues, where the level of expression of gene HAGE is at least 100 times lower than in LB23-SAR. The HAGE protein is not the first case of a DEAD-box protein that is overexpressed in tumors; this has already been described in retinoblastoma and neuroblastoma cell lines (51–53) . Moreover, it is worth noting that of 42 tumor antigens discovered until now, two are derived from mutated helicases. A mutated murine helicase, named p68, was found to lead to the expression of an antigen recognized by a CTL clone on a UV-induced sarcoma (54) . The human melanoma antigen LB33-A is produced by a point mutation in a new gene, MUM-3, which is expressed ubiquitously and shows homology with members of the RNA helicase gene family. 7 These observations suggest that mutated or overexpressed helicases may contribute to tumoral transformation or progression.
Genes such as SAGE and HAGE, with expression that appears to be restricted to tumors and germ-line cells, are potentially coding for tumor-specific antigens recognized by T lymphocytes. But these antigens remain to be identified. Approaches to establish which antigenic peptides recognized by T cells are encoded by such genes have recently made considerable progress. In vitro stimulation of CD8+ T lymphocytes with dendritic cells infected with canarypoxes or adenoviruses carrying sequences of MAGE genes has led to the definition of a large number of new epitopes encoded by these genes (55, 56) . Moreover, stimulation of CD4+ T cells with dendritic cells pulsed with MAGE proteins has led to the identification of several new antigenic peptides presented by class II molecules (57) . These results significantly enlarge our possibilities for therapeutic vaccination with tumor-specific antigenic peptides, because it is now certain that almost every cancer patient whose tumor expresses a MAGE gene will have at least one HLA molecule presenting a MAGE antigenic peptide. The same approach could be applied to genes SAGE and HAGE to identify tumor-specific antigens derived from them. Consecutive immunization with several antigens might ensure a more effective rejection of tumor cells because this should reduce the emergence of antigen-loss variants arising by mutations in the genes producing the antigenic peptides or by down-regulation of their expression. SAGE and HAGE antigens could also be used to immunize patients against tumors that do not express MAGE genes.
Acknowledgments
We thank Marie-Claire Letellier and Claudine Blondiaux for technical assistance, Pierre van der Bruggen for critical reading of the manuscript, and Saïda Khaoulali and Simon Mapp for editorial assistance.
Footnotes
-
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.
-
↵This work was supported by grants from Fonds J. Maisin (Belgium), Fédération Belge contre le Cancer (Belgium), Caisse Générale d’Epargne et de Retraite (CGER)-Assurances and VIVA (Belgium), the Belgian program on Interuniversity Poles of Attraction initiated by the Belgian State, Prime Minister’s Office, Office for Science, Technology and Culture. V. M. is supported by Fonds pour la Recherche Scientifique dans l’Industrie et l’Agriculture (Belgium).
-
↵To whom requests for reprints should be addressed, at Ludwig Institute for Cancer Research, Université Catholique de Louvain, 74 avenue Hippocrate, UCL 7459, B-1200 Brussels, Belgium. Phone: 32-2-764-75-80; Fax: 32-2-762-94-05; E-mail: boon{at}licr.ucl.ac.be
-
↵The abbreviations used are: RT-PCR, reverse transcription-PCR; SAGE, sarcoma antigen gene; HAGE, helicose antigen gene.
-
↵Internet address: http://www-genome.wi.mit.edu/cgi-bin/contig/rhmapper.pl.
-
↵Internet address: http://www.ncbi.nlm.nih.gov/SCIENCE96.
-
↵B. Lethé, personal communication.
-
↵J. F. Baurain, D. Colau, N. van Baren, C. Landry, V. Martelange, M. Vikkula, T. Boon, and P. G. Coulie. High Frequency of Autologous Anti-Melanoma CTL Directed Against an Antigen Generated by a Point Mutation in a New Helicase Gene, submitted for publication.<./>
- Received December 17, 1999.
- Accepted May 16, 2000.
- ©2000 American Association for Cancer Research.
References
Volume 60, Issue 14
