Monophasic and biphasic synovial sarcomas abundantly express cancer/testis antigen ny-eso-1 but not mage-a1 or ct7



Synovial sarcomas are high-grade malignant mesenchymal tumors with biphasic (BSS) and monophasic (MSS) variants that carry a pathognomonic cytogenetic alteration, t(X;18), involving the SYT gene on chromosome 18 and one of several SSX genes on chromosome X, usually SSX1 or SSX2. Cancer/testis (CT) antigens are expressed in a variety of malignant neoplasms but, in normal tissues, are restricted to male germ cells. Previous analysis revealed a high incidence and homogeneous expression of MAGE CT antigen in synovial sarcomas. The present study was performed to analyze the expression of 3 CT antigens, NY-ESO-1, MAGE-A1 and CT7, by immunohistochemistry with 3 monoclonal antibodies (MAbs), ES121 (anti-NY-ESO-1), MA454 (anti-MAGE-A1) and CT7-33 (anti-CT7), in 25 synovial sarcomas (12 MSS, 13 BSS) typed for the t(X;18)-derived fusion transcript by RT-PCR (19 SYT-SSX1, 6 SYT-SSX2). NY-ESO-1 immunoreactivity was found in 20/25 (80%) cases, and antigen expression was homogeneous in 14/20 NY-ESO-1-positive cases. Both morphologic variants and both translocation types were NY-ESO-1-positive, whereas 5 SYT-SSX1 tumors (1 MSS, 4 BSS) were NY-ESO-1-negative. MAb MA454 was immunoreactive with 4/25 cases (2 MSS, 2 BSS; 3 SYT-SSX1, 1 SYT-SSX2), and MAb CT7-33 was immunoreactive with only 2/25 cases (both BSS, SYT-SSX1). Expression of MAGE-A1 and CT7 was heterogeneous in all positive cases. Our study shows that NY-ESO-1 is highly expressed in a homogeneous pattern in synovial sarcomas of both morphologic variants and both translocation types, making these tumors an attractive target for NY-ESO-1 antigen-based immunotherapy. © 2001 Wiley-Liss, Inc.

Cancer/testis (CT) antigens are expressed in a variety of malignant tumors and in germ cells in the testis.1 They are immunogenic in humans, eliciting cellular and humoral immune responses in a subset of patients with tumors expressing them.2, 3, 4 Over 10 genes or gene families coding for CT antigens have been identified by epitope-cloning techniques using autologous CD8+ T cells5 or antibody6, 7 or by representational difference analysis.8, 9 Two of the best studied CT systems are MAGE, a large gene family,10–13 and NY-ESO-1,14 a gene with only one closely related homologue, LAGE.15 Another CT antigen, CT7, was identified by cDNA expression cloning using a melanoma cell line and allogeneic sera from a melanoma patient.16CT7 is identical to MAGE-C1, a gene identified by representational difference analysis.9 Serologic reagents identifying MAGE gene products,17–19 NY-ESO-13 and CT7 (data not shown) have been generated and are being used to analyze expression of these antigens in normal and malignant tumors.18, 20–24

In a previous survey of human tumors using the polyvalent MAGE reagent monoclonal antibody (MAb) 57B, synovial sarcomas revealed not only a high number of positive cases but also an exceptional homogeneous expression pattern.23 Synovial sarcomas are high-grade malignant tumors with a poor prognosis, accounting for 5% to 10% of all soft tissue sarcomas, mostly in younger patients.25 They carry a pathognomonic chromosomal alteration, t(X;18),26, 27 representing a rearrangement involving the SYT gene on chromosome 18 and 1 of 5 SSX genes on chromosome X.28–30 The presence of the translocation can be determined by RT-PCR of the chimeric SYT-SSX transcript,28, 29, 31–33 and the translocation type corresponds to the SSX gene involved in the gene rearrangement. Two main morphologic subtypes of synovial sarcoma can be distinguished: a biphasic form (BSS) consisting of mesenchymal/spindle as well as epithelial components and a monophasic type (MSS) composed entirely of spindle cells.34–37 These 2 histologic types have been associated with t(X;18) translocation types and with prognosis.38–42

In the present study, we have analyzed the expression pattern of 3 CT antigens, NY-ESO-1, MAGE-A1 and CT7, in a series of synovial sarcomas that were typed for the SYT-SSX fusion transcript. Our study establishes high and homogeneous expression of NY-ESO-1 and a low frequency of MAGE-A1 and CT7 expression in synovial sarcomas. No correlation was found between NY-ESO-1 expression and morphologic variant or translocation type.


Twenty-five samples of synovial sarcoma tissue were obtained from the archives of the Department of Pathology, Memorial Sloan-Kettering Cancer Center (Table I). Slides of all cases were reviewed and representative blocks chosen for immunohistochemic analysis. Three pathologists (CRA, AAJ, KJB) evaluated the slides independently without knowledge of clinical or molecular data.

Table I. Immunophenotyping of 25 Synovial Sarcomas for NY-ESO-1 (MAb ES121), MAGE-A1 (MAb MA454) and CT7 (MAb CT7-33): Comparison of Antigenic Phenotype, Morphology and SYT-SSX Translocation Type
NumberSYT-SSX fusion transcriptClinical stage of sampleHistologic typeMAb ES121MAb MA454MAb CT7-33
  1. Grading: Focal approx. <5%; +, 5% to <25%; ++, 25% to <50%; +++, 50% to <75%; ++++, >75%. P, primary; M, metastasis; R, recurrence; BSS, biphasic synovial sarcoma; MSS, monophasic synovial sarcoma.


Frozen tissue specimens were used for RT-PCR analysis. RNA was extracted by the guanidinium-phenol-chloroform method, 1.0 μg of which was reverse-transcribed (Superscript reverse transcriptase; GIBCO, Gaithersburg, MD). Primers for the detection of translocation were as follows:41 forward SYT (5′-CAACAGCAAGATGCATACCA), reverse SSX consensus (5′-GGTGCAGTTGTTTCCCATCG), SSX1 (5′-GGTGCAGTTGTTCCCATCG) and SSX2 (5′-GGCACAGCTCTTTCCCATCA). RT-PCR products were separated by gel electrophoresis and visualized by ethidium bromide.

Figure 1.

Immunohistochemical staining of synovial sarcomas for NY-ESO-1 (MAb ES121), MAGE-A1 (MAb MA454) and CT7 (MAb CT7-33). (a) Homogeneous immunoreactivity of MSS for ES121, case 5. (b) Same case negative for CT7-33. (c) Same case negative for MA454. (d) BSS, case 8, ES121-negative (thick arrows) and positive (*) glands and a gland (+) with single positive epithelial cells (thin upward arrow). (e) Focal cluster of MA454-positive cells, case 2. (f) CT7-33-positive spindle cells interspersed among mostly negative sarcoma cells, case 6.

MAb ES121, which detects NY-ESO-1,3, 14 and MAb MA454, which detects MAGE-A1,17–24 have been described. MAb CT7-33, with specificity for CT7,16 was generated recently (data not shown). Immunohistochemistry was performed on formalin-fixed, paraffin-embedded tissues. All MAbs were titered on normal testis, and testis was used as a control tissue. The following MAb concentrations were used: ES121 2.5 μg/ml, MA454 1.0 μg/ml and CT7-33 1:500 (supernatant). A heat-based antigen-retrieval method was applied to all antibodies [Black and Decker steamer, Black and Decker, Denver, CO; 90°C, 30 min]. DAKO hipH-solution and EDTA (1 mmol, pH 8.0) were used as the antigen-retrieval solution for ES121 and MA454 respectively, and citrate (10 mmol, pH 6.0) was used for CT7-33. A biotinylated horse anti-mouse secondary antibody (1:200; Vector, Burlingame, CA) was used to detect primary antibody, followed by an avidin-biotin system (ABC-elite kit, Vector). 3,3′-Diaminobenzidine tetrahydrochloride (Biogenex, San Ramon, CA) served as a chromogen. Endogenous peroxidase was suppressed by 1% H2O2 for 20 min. The extent of tumor staining was estimated on the basis of numbers of tumor cells stained and graded as follows: Focal approx. <5%; +, 5% to <25%; ++, 25% to <50%; +++, 50% to <75%; ++++, >75%.


Histologic subtype, type of fusion transcript and immunohistochemical staining of the 25 synovial sarcomas are summarized in Table I. Twelve cases showed a monophasic spindle cell type, while 13 tumors revealed a biphasic pattern. Monophasic tumors showed the typical morphology of highly cellular spindle cells arranged in sheets or bundles. Biphasic tumors showed variable areas of mostly gland-like epithelial structures in addition to the sarcomatous component. None of the cases represented a poorly differentiated pattern.35, 37 Fifteen samples were derived from primary tumors, 4 from metastatic sites and 6 from recurrences. The chimeric fusion transcript was SYT-SSX1 in 19 cases and SYT-SSX2 in the remaining 6 cases. All 13 biphasic tumors were of the SYT-SSX1 translocation type, whereas all 6 SYT-SSX2 cases and 6 SYT-SSX1 cases were monophasic.

Immunohistochemical analysis showed strong reactivity, with MAb ES121 detecting NY-ESO-1 in a large proportion of cases; 80% (20/25) of the synovial sarcomas in this series showed strong NY-ESO-1 expression. Both MSS (11 cases) and BSS (9 cases) groups were NY-ESO-1-positive. In biphasic tumors, immunoreactivity was present in the spindle cell as well as the glandular epithelial component, though in some cases there was preferential staining of one component or the other. Staining was mostly cytoplasmic, with nuclear staining being less prominent but present in the epithelial as well as the spindle cell components of tumors of both translocation types. Most tumors revealed a homogeneous staining pattern: 14/20 MAb ES121-positive synovial sarcomas showed immunoreactivity in >50% of the tumor area. All 6 tumors carrying the SYT-SSX2 fusion were MAb ES121-reactive. Of the 19 cases carrying the SYT-SSX1 translocation, 14 were NY-ESO-1-positive (9 BSS, 5 MSS). The 5 NY-ESO-1-negative cases included 4 BSS cases and 1 MSS case. In contrast to the strong and homogeneous expression of NY-ESO-1, MAb CT7-33 immunoreactivity was seen in only 2/25 (8%) cases; staining was focal in 1 case and <25% in the other. Both cases were of the biphasic type, carrying the SYT-SSX1 translocation, and showed cytoplasmic and nuclear staining confined to spindle cell areas. Anti-MAGE-A1 MAb MA454 also showed very restricted immunoreactivity. Staining could be observed in 4/25 (16%) tumors, focal in 1 case and <25% in 3 lesions; immunoreactivity was cytoplasmic and present in spindle and epithelial components. Of the 4 MAGE-A1-positive tumors, 3 had the SYT-SSX1 translocation (2 BSS, 1 MSS), while the other case (MSS) carried the SYT-SSX2 rearrangement.

Comparison of the staining results of the 3 reagents indicated that only synovial sarcomas with ++ to ++++ MAb ES121 immunoreactivity showed staining with CT7-33 and MA454.


CT antigens are attractive targets for immunotherapy because of their highly restricted expression in normal tissue (male germ cells) and broad expression in a wide range of different tumor types. Although much is known about their expression pattern at the mRNA level, a systematic survey of CT antigen expression at the protein level using immunohistochemical analysis has only begun.23, 24 MAbs have been generated against a number of CT antigens, including MAGE-A1,17, 43, 44 MAGE-A3,21, 22 MAGE-A4,44 NY-ESO-1,3, 19a SCP-1, CT7 and SSX1.45 The availability of these serologic reagents will make it possible to develop a comprehensive picture of CT antigen expression in human cancer.

A broad survey of human tumors and normal tissues has been carried out using 2 MAbs with specificity for CT antigens: MA454, against recombinant MAGE-A1,17 and 57B, against recombinant MAGE-A3.19, 22 Although the specificity of MAb MA454 for MAGE-A1 has been established,17, 24, 46 MAb 57B clearly reacts with more than 1 MAGE gene product when tested on cells transfected with different MAGE genes and should therefore be considered a polyvalent MAGE reagent.47 However, others48 have reported that MAb 57B reactivity with human tumors appears to co-type with MAGE-4 mRNA expression, so further analysis of its specificity is needed. The results of our analysis of human tissues with 57B23 and MA45424 can be summarized as follows: (i) Reactivity with human normal tissue is restricted to germ cells in the testis, with intense staining of spermatogonia and decreasing staining of male germ cells in later stages of maturation. (ii) There is a general correspondence between RT-PCR typing of MAGE expression and immunophenotyping with MAb MA454. Some tumor types, e.g., melanoma and head-and-neck cancers, have a high frequency of MAGE mRNA and MAGE antigen expression, whereas others, e.g., renal and colorectal cancers, show little or no MAGE expression. (iii) The general pattern of MAGE antigen expression in human tumors was heterogeneous, with only a small percentage of tumors showing a homogeneous staining pattern. The common picture was of strongly positive tumor cells or areas of tumor surrounded by tumor cells/areas with no immunoreactivity. The basis for this highly heterogeneous pattern of MAGE antigen expression is unclear, and efforts are under way to determine whether it is also true at the mRNA level.

In our analysis of MAb 57B, one tumor type, synovial sarcomas, stood out because of an exceptionally high frequency of immunoreactivity and a strong homogenous staining pattern.23 To extend this observation, we typed 25 synovial sarcomas for the expression of 3 CT antigens, NY-ESO-1, MAGE-A1 and CT7. Similar to the reactivity of MAb 57B, the NY-ESO-1 MAb reacted with a high percentage of synovial sarcomas (80%) and gave a strong, generally homogeneous pattern of staining. Both morphologic types showed NY-ESO-1 expression, with strong staining of spindle cells as well as epithelial components in the tumor. No definite statement can be made with regard to the relation of translocation type and NY-ESO-1 expression because of the small series and the uneven distribution of SSX-1 and SSX-2 cases. However, it may be significant that all 6 cancers involving SSX2 were ES121-positive, whereas all 5 NY-ESO-1-negative sarcomas had the SYT-SSX1 fusion transcript. In contrast to the high frequency of NY-ESO-1 expression in synovial sarcomas, expression of MAGE-A1 and CT7 was very infrequent, occurring only in tumors with strong NY-ESO-1 expression. Frozen tissue was available in 3 of the 25 synovial sarcoma cases. RT-PCR analysis of these 3 cases was in accordance with previous co-typing analyses for CT7/CT7-33 (data not shown), MAGE-1/MA45424 and NY-ESO-1/ES12119a and corresponded to the immunohistochemical reactivity pattern (data not shown).

The molecular basis for the strong expression of NY-ESO-1 in synovial sarcomas is unknown. Because NY-ESO-1, like a number of other CT antigens, is encoded by genes on the X chromosome, it is possible that the X-linked translocation event associated with synovial sarcoma is involved in NY-ESO-1 expression. However, the SSX genes involved in synovial sarcomas map to the short arm of chromosome X (Xp11), whereas NY-ESO-1 as well as MAGE-A1 and CT7 map to the end of the long arm (Xq26-28), making it unlikely that the SYT-SSX translocation is a direct cause of the activation or de-repression of the NY-ESO-1 gene. A genomewide demethylation process associated with malignancy has been correlated with activation of MAGE genes in cancer.49–52 However, this would not explain the selective activation of different CT genes in cancer or why NY-ESO-1 is activated in synovial sarcoma, whereas MAGE-A1 and CT7 are generally not.

With the availability of MAbs to other CT antigens, it will be important to extend this profiling of CT antigen expression in synovial sarcomas and other sarcomas and to determine whether there are any coordinate patterns of CT antigen expression. In addition, a larger series will need to be analyzed, to establish correlations (if any) between NY-ESO-1 expression, morphologic type, specific translocation and stage and site of disease. However, the finding that NY-ESO-1 has strong and homogeneous expression in synovial sarcomas focuses attention on this tumor type as a target for NY-ESO-1–based immunotherapy.