XAGE-1, A New Gene That Is Frequently Expressed in Ewing’s Sarcoma¹

Large numbers of ESTs³have been cloned from various tissues and cancers (1, 2). Each cDNA clone or EST sequence is generated from a single transcript. The frequency and distribution of the many different transcripts in a given tissue depends on the level of gene expression. Therefore, a particular gene expression pattern can be frequently predicted by analysis of the frequency and specificity of various EST sequences. We previously reported a computer screening strategy that identified genes that are preferentially expressed in prostate or prostate tumors (3, 4, 5). From this screen, numerous genes were identified including a novel gene, PAGE4 (previously named PAGE1) (6), and a set of XAGEgenes (7), which are related to the GAGE/MAGE family of melanoma associated CT antigens.

CT antigens are a distinct class of differentiation antigens that have a restricted pattern of expression in normal tissues(8, 9, 10). CT antigens are currently viewed as attractive candidates for cancer vaccines. Some thoroughly studied CT antigens are MAGE, BAGE, GAGE, and LAGE/NY-ESO-1 (9, 10, 11, 12, 13, 14, 15, 16, 17). These genes are primarily expressed in the primitive germ cells, spermatogonia, in the normal testis. Malignant transformation is often associated with activation or derepression of silent CT genes, and this results in the expression of CT antigens in a variable proportion of a wide range of human tumors. Recently, several additional members were added to the CT antigen family. These include various PAGEs, PRAME, SSX, SCP-1, CT7,MAGEC1, and MAGED1 (6, 11, 18, 19, 20, 21, 22). Identification of new CT antigens or new family members is still being actively pursued in the cancer research field.

Previously, we identified three related genes that we termed XAGEs by homology walking using the dbEST database(7). ESTs of the XAGE group were found in cDNA libraries from testis, Ewing’s sarcoma, alveolar rhabdomyosarcoma,fetal tissues, and germ cell tumors by database analysis. In this study, we describe the isolation of a full-length cDNA of XAGE-1 and analysis of the expression of the XAGE-1 gene. We found that the gene is highly expressed in normal testis, with aberrant expression in sarcomas of various types. Although XAGE-1 did not show overall homology with any polypeptide recorded in the data bank, it has strong homology with members of GAGE/PAGE family in the COOH terminus of the predicted protein,indicating that XAGE-1 belongs to a new family of CT antigens.

Ewing’s tumor tissue was obtained from frozen specimens obtained from patients treated at the National Cancer Institute. Rhabdomyosarcoma tumor tissue was obtained from the Cooperative Human Tissue Network,Children’s Cancer Group. All alveolar rhabdomyosarcoma tumor specimens were found to express the PAX-3-FKHR fusion transcript by RT-PCR.⁴Osteosarcoma cell lines were obtained from the American Type Culture Collection. Ewing’s sarcoma cell lines RD-ES, TC-32, TC-71, and 5838 have been described previously, and all contain EWS-FLI-1 fusion transcripts (23).⁴ LD, LG, JM, and SB are cell lines established in our laboratories (by L. J. H.) directly from tumor specimens. The cell line JM does not express an EWS-ETS fusion transcript.

RNA was extracted either from tumor tissue using Trizol (Life Technologies, Inc.) or from cell lines using RNAeasy (Qiagen). Total RNA (20 μg) was used for Northern blot analysis of sarcoma tumors. The multiple tissue mRNA dot blot and the normal tissue Northern blot were purchased from Clontech (Palo Alto, CA). The 450-bp probe used for hybridization was generated from EST clone af89d01.s1 by digestion with EcoRI and NotI. The hybridization was conducted as follows: the RNA-containing membranes were blocked for 3 h at 45°C in hybridization solution. Probes labeled with ³²P either by random primer extension or by end labeling (Lofstrand Labs Limited) were added to the membrane and hybridized for 15 h at 45°C. Membranes were washed twice with 2× SSC/0.1% SDS at room temperature and twice with 0.1× SSC/0.1%SDS at 65°C. The membranes were exposed to X-ray film for 1–2 days before development

The Southern blot of human chromosomes (Oncor, Gaithersburg, MD) was conducted using the same probe and the same hybridization conditions used for Northern blot.

RT-PCR was performed on cDNA from 24 different human tissues using human rapid-scan gene expression panels (Origene, Inc., Rockville, MD). The thermocycling protocol was initial denaturation at 94°C for 3 min and 35 cycles of denaturation at 94°C for 1 min, annealing at 65°C for 1 min, and elongation at 72°C for 3 min. The PCR reactions were analyzed on agarose gels, and specific products were cloned into TA vectors (Invitrogen) and sequenced on an automated capillary sequencer using Perkin-Elmer’s dRhodamine terminator cycle sequencing kit(Perkin-Elmer Applied System).

The primers used were as follows: (a) xa-1,5′-CAGCTTGTCTTCATTTAAACTTGTGGTTGC-3′; (b) xa-2,5′-TCCCAGGAGCCCAGTAATGGAGA-3′; (c) xa-8,5′-ACCTGGGAAGGAGCATAGGA-3′; and (d) xa-10,5′-CTTTATTGAGATAGTTTAAGTCAAATATCTAA-3′. The oligonucleotides were synthesized by Sigma-Genosys.

To determine the relative expression of XAGE-1 mRNA in different tissues and tumors, we first conducted a mRNA dot blot(Clontech) analysis using a full insert of EST af89d01.s1 as a labeling probe. Among the 61 different samples of normal tissues and 7 fetal tissues including lung, brain, liver, heart, and spleen, the expression of XAGE-1 was detected only in testis (Fig. 1 A). This result indicates that XAGE-1, like other CT antigens, is present in testis.

To verify the specificity of XAGE-1 expression, we conducted a RT-PCR analysis by using the human rapid-scan panel with primers xa-1 and xa-2. We detected a 275-bp fragment in testis among the 24 different tissues analyzed (Fig. 1,B). Unexpectedly, the 275-bp fragment was also present at lower amounts in normal lung and PBLs. Extremely weak expression of XAGE-1 was detected in bone marrow, spleen, and skin. To compare the relative level of XAGE-1 in testis, lung, and PBLs, different dilutions of cDNA were analyzed in the same rapid-scan panel as shown in Fig. 1 C. The mRNA present in testis was about 10–100 times higher than that in the lung and more than 100 times higher than that in PBLs.

Because XAGE-1 is highly abundant in testis and is expressed at a low level in lung and PBLs, we attempted to determine the transcript size in these different tissues. Northern blot analysis was conducted using the same probe used for the RNA dot blot. As shown in Fig. 2,A, a single band of 700 bp was revealed in the testis. However, no signal was detected in lung and PBLs. This result is probably due to the low level of XAGE-1 expression in lung and PBLs because the Northern blot analysis is much less sensitive than RT-PCR in detecting the expression of XAGE-1. These results are consistent with the RNA dot blot analysis described above (Fig. 1 A).

Analysis of the EST database predicts that XAGE-1 is present in Ewing’s sarcoma and alveolar rhabdomyosarcoma. To confirm the database prediction experimentally, we first determined whether XAGE-1 was present in the various Ewing’s cell lines by Northern blot analysis. A single band of 700 bp was detected in seven of eight cell lines (Fig. 2,B). XAGE-1 was not expressed in cell line JM, which is a mouse xenograft tumor derived from a Ewing’s sarcoma that has lost the chromosome translocation t(11;22).⁴ Cell line 5838 had an extra band with a size of 1.2 kb. This band might be due to alternate splicing or use of an alternate polyadenalytion signal in the XAGE-1 gene. XAGE-1 was present in two of five osteosarcoma cell lines, with the SAOS cell line showing relatively low XAGE-1 expression (Fig. 2 B).

To address whether XAGE-1 was present in human patient samples, we conducted a Northern blot hybridization analysis. Of nine patients with Ewing’s sarcoma, four (patients 5–8) expressed XAGE-1 with a single 700-bp band (Fig. 2,C). Patients 1 and 5–9 expressed the EWS-FLI-1 transcript, an indication of chromosome translocation (Ref. 24; data not shown). We noticed that XAGE-1 was not expressed in all of the patient samples with the chromosome translocation. However, samples that did not express either an EWS-FLI-1 or an EWS-ERG fusion transcript also did not express XAGE-1 (Fig. 2,C, Lanes 2–4). The correlation between the absence of XAGE-1 expression and the absence of chromosome translocation needs further study. XAGE-1 was also expressed in one of one patient sample with alveolar rhabdomyosarcoma and in one of three patient samples of embryonal rhabdomyosarcoma but not in the normal controls (Fig. 2 C). Together, these data indicate that XAGE-1 is expressed in nearly half of the sarcoma patient samples.

Most of the CT antigens are localized on the X chromosome, with the exception of SCP-1, which is located on chromosome 1 (8, 20). To find where XAGE-1 is localized, Southern blot hybridization was performed on a human chromosome blot using the same probe used for dot blot and Northern blot. We detected only one strong band on the X chromosome, and no other cross-hybridizing bands were found on the blot (data not shown). This result indicates that the XAGE-1 gene is located on the X chromosome and that there is not a very strong homology with the other predicted XAGEmembers, XAGE-2 and XAGE-3, because under stringent hybridization conditions, XAGE-2 and XAGE-3 were not detected.

To obtain the full-length XAGE-1 cDNA sequence, rapid amplification of cDNA ends-PCR was performed using primers localized in the EST contig and total RNA from Ewing’s sarcoma cell line TC-71. The longest rapid amplification of cDNA ends product contains an additional 184 nucleotides at the 5′ end compared with the EST contig sequence. The correct cDNA sequence was confirmed by sequencing the PCR product from primers xa-8 and xa-10. The XAGE-1 cDNA is 611 bp in length, excluding the poly(A) tail, and contains 438 nucleotides in the coding region, flanked by 85 bp in the 5′-untranslated region and 88 bp in the 3′-untranslated region (Fig. 3 A).

The longest open reading frame indicates that the putative XAGE-1 protein consists of 146 amino acid residues with a molecular weight of 16,300. Hydrophilicity analysis (Fig. 3,B) indicates a hydrophobic sequence in the NH₂-terminal end,suggesting the possibility of a transmembrane domain. Analysis of the protein sequence reveals no possible posttranslational modifications by searching GCG Lite. This putative protein did not show overall sequence homology with any peptide recorded in the data banks. However,alignment of the amino acid sequence of XAGE-1 with PAGE4(6) and PAGE1 [Ref. 25 (renamed GAGE9; Ref.15)] reveals a striking homology in the COOH-terminal end of these proteins (Fig. 4), suggesting that XAGE-1 encodes a distinct protein that could share structural or functional features with other GAGE/PAGE family members.

XAGE-1 is a human X-linked gene that is strongly expressed in normal testis, Ewing’s sarcoma, and alveolar rhabdomyosarcoma. It is also expressed in normal lung and PBLs, but at much lower levels. XAGE-1 expression in lung and lymphocytes was not predicted by EST database analysis, probably because of its weak expression and the small size of lung and PBL libraries. The expression pattern of XAGE-1 in Ewing’s sarcoma, lung,PBLs, and bone marrow is interesting. At present, it is unclear in which cell type Ewing’s sarcoma originates and why it metastasizes to the lung. The high level of XAGE-1 expression in many Ewing’s sarcomas suggests that the tumor could arise in cells in the bone or bone marrow that normally express XAGE-1 or metastasize to the region where XAGE-1 is expressed. Additional experiments using in situ hybridization or immunohistochemistry should provide information about this possible relationship.

Until now, there has been no report on the expression of CT antigens in Ewing’s sarcoma and alveolar rhabdomyosarcoma. We found XAGE-1 expressed in seven of eight Ewing’s sarcoma cell lines, 44% (four of nine) of Ewing’s sarcoma tumor specimens, and one of one alveolar rhabdomyosarcoma tumor sample. We are currently screening the expression of XAGE-1 in other tumor cell types. Our preliminary data indicate that the XAGE-1 gene is activated in many tumor cell lines and other tumor types. Studies on the XAGE-1 gene may provide more information on this cancer antigen family that may be of use in the diagnosis or immunotherapy of cancer.

We thank Drs. C. Wolfgang and K. Egland for critical reading of the manuscript. We also thank R. Mann for editorial assistance.