Breast cancer suppressor candidate-1 (BCSC-1) is a melanoma tumor suppressor that down regulates MITF.

Understanding the molecular aberrations involved in the development and progression of metastatic melanoma (MM) is essential for a better diagnosis and targeted therapy. We identified breast cancer suppressor candidate-1 (BCSC-1) as a novel tumor suppressor in melanoma. BCSC-1 expression is decreased in human MM, and its ectopic expression in MM-derived cell lines blocks tumor formation in vivo and melanoma cell proliferation in vitro while increasing cell migration. We demonstrate that BCSC-1 binds to Sox10, which down regulates MITF, and results in a switch of melanoma cells from a proliferative to a migratory phenotype. In conclusion, we have identified BCSC-1 as a tumor suppressor in melanoma and as a novel regulator of the MITF pathway.

Melan-A/MART-1-specific CD4 T cells in melanoma patients: identification of new epitopes and ex vivo visualization of specific T cells by MHC class II tetramers.

Over the past decade, many efforts have been made to identify MHC class II-restricted epitopes from different tumor-associated Ags. Melan-A/MART-1(26-35) parental or Melan-A/MART-1(26-35(A27L)) analog epitopes have been widely used in melanoma immunotherapy to induce and boost CTL responses, but only one Th epitope is currently known (Melan-A51-73, DRB1*0401 restricted). In this study, we describe two novel Melan-A/MART-1-derived sequences recognized by CD4 T cells from melanoma patients. These epitopes can be mimicked by peptides Melan-A27-40 presented by HLA-DRB1*0101 and HLA-DRB1*0102 and Melan-A25-36 presented by HLA-DQB1*0602 and HLA-DRB1*0301. CD4 T cell clones specific for these epitopes recognize Melan-A/MART-1+ tumor cells and Melan-A/MART-1-transduced EBV-B cells and recognition is reduced by inhibitors of the MHC class II presentation pathway. This suggests that the epitopes are naturally processed and presented by EBV-B cells and melanoma cells. Moreover, Melan-A-specific Abs could be detected in the serum of patients with measurable CD4 T cell responses specific for Melan-A/MART-1. Interestingly, even the short Melan-A/MART-1(26-35(A27L)) peptide was recognized by CD4 T cells from HLA-DQ6+ and HLA-DR3+ melanoma patients. Using Melan-A/MART-1(25-36)/DQ6 tetramers, we could detect Ag-specific CD4 T cells directly ex vivo in circulating lymphocytes of a melanoma patient. Together, these results provide the basis for monitoring of naturally occurring and vaccine-induced Melan-A/MART-1-specific CD4 T cell responses, allowing precise and ex vivo characterization of responding T cells.

The prognostic value of minimally involved melanoma sentinel lymph nodes

Introduction: la biopsie du ganglion sentinelle (GS) est une procédure reconnue et fiable pour établir le stade ganglionnaire du mélanome cutané. Le GS est le facteur pronostique le plus puissant pour la survie des patients atteints d'un mélanome à risque intermédiaire, cliniquement localisé. Celui-ci est métastatique dans environ 15-30% des cas. Lorsque le GS est positif, un curage de l'aire ganglionnaire concernée est généralement entrepris. Néanmoins, seuls 20-25% de ces patients présentent des ganglions non-sentinelles (GNS) métastatiques. Ces données suggèrent que le curage, et les risques opératoires qui y sont associés, n'est peut-être pas nécessaire chez le trois-quarts de ces patients. Un autre aspect est que l'impact sur la survie des curages basé sur le résultat du GS n'est pas clairement démontré. La nécessité de ce curage d'emblé est actuellement en cours d'évaluation par un protocole international (Multicenter Selective Lymphadenectomy Trial II : MSLT II). Plusieurs auteurs ont essayé de classifier la charge tumorale du GS afin d'évaluer s'il était possible d'épargner le curage à certains patients et de mieux affiner ce facteur pronostic sans succès. En 2009, le Groupe Mélanome de l'EORTC (European Organisation for Research and Treatment of Cancer) a recommandé un protocole d'évaluation anatomopathologique du GS-positif en trois items: (1) la localisation micro-anatomique des métastases à l'intérieur du ganglion selon Dewar (A = sous-capsulaire, B = combinée sous-capsulaire and parenchymateuse, C = parenchymateuse, D = multifocale, and'E = extensive) ; (2) la mesure de la taille tumorale dans le ganglion selon les critères de Rotterdam pour le diamètre maximal. Le diamètre de la plus grande métastase est exprimé en nombre absolu et (3) la taille tumorale stratifiée par catégories : <0.1mm, 0.1-1.0mm et >1.0 mm. Le but de cette étude rétrospective d'une cohorte de patients, était d'investiguer les résultats des GS-positifs et d'analyser les facteurs pronostiques de la survie à la lumière des recommandations de l'EORTC. Ainsi que de comparer les sous-groupes du GS-positif avec une invasion minimale (taille tumorale <0.1mm et/ou atteinte sous-capsulaire) avec le GS-négatif. Les facteurs pouvant prédire la présence de GNS- positif ont également été analysés. Matériel et méthode : une étude des dossiers a été réalisée pour les 499 patients consécutifs entre 1997 et 2008 qui ont eu une biopsie du GS dans notre institution. Le dégrée d'envahissement du GS-positif a été entièrement revue par l'équipe référente de l'Institut de Pathologie (Dresse E. Saiji et Dresse H. Bouzourène) selon les recommandations de l'EORTC. Des analyses univariées et multivariées des potentiels facteuis pronostics ont été réalisées. Des analyses de survie ont également été effectuées avec des courbes d'estimation de Kaplan-Meier combinées à une régression de Cox. Le protocole a été accepté par la Commission d'Ethique. Résultats: un GS-positif a été trouvé chez 123 (25%) patients panni les 499 qui ont bénéficié d'une biopsie. Avec un suivi médian de 52 mois, la survie à 5 ans sans récidive (SSR), spécifique à la maladie (SS) et globale (SG) étaient de 88%, 94%, et 90% respectivement pour les patients avec GS-négatif. Concernant les GS avec invasion minimale, 21 patients étaient dans le sous-groupe <0.1 mm selon les critères de Rotterdam et 52 patients dans le sous-groupe sous-capsulaire selon Dewar. La survie dans ces deux sous-groupes était de 80% et 57% pour la SSR, 87% et 70% pour la SS, 87% et 68% pour la SG, respectivement. L'analyse multivariée des GS-positifs a montré que les facteuis suivants influençaient significativement la survie (SSR, SS et SG): l'épaisseur selon Breslow de la tumeur primaire (p=0.002, 0.006, 0.004), la taille tumorale du GS-positif >0.1 mm (p= 0.01, 0.04, 0.03), le genre masculin (p=0.06, 0.005, 0.002) et l'ulcération de la tumeur primaire (p=0.05, 0.03, 0.007). L'analyse des sous-groupes avec invasion minimale n'a pas permis d'établir de facteur pour prédire la négativité des GNSs. Conclusion: La classification du GS-positif par la taille tumorale selon les critères de Rotterdam est un facteur pronostique simple et utile pour évaluer la survie des patients atteints de mélanome. Nous avons observé une tendance (non statistiquement significative) d'une survie diminuée pour le sous-groupe des patients avec GS-positif et une taille de la métastase <0.1 mm comparée à celle des patients avec GS-négatif. Ceci nous incite à conclure que ce sous-groupe de patients ne devrait pas être assimilé et traité comme ceux qui ont un GS-négatif. D'autre part nos résultats montrent que la localisation micro-anatomique selon Dewar n'est pas un outil pronostique utile pour évaluer la survie, ni pour prédire le status des GNSs.

Frequent cytolytic T-cell responses to peptide MAGE-A10(254-262) in melanoma.

MAGE genes encode tumor-specific shared antigens that are among the most interesting candidates for cancer vaccines. Despite extensive studies, however, CD8+ T-cell responses to MAGE-derived epitopes have been detected only occasionally in cancer patients, even after vaccination. In contrast with these findings, we report here that HLA-A2 melanoma patients respond frequently to the recently identified peptide MAGE-A10(254-262). Indeed, as assessed by staining with fluorescent HLA-A2/peptide MAGE-A10(254-262) tetramers, CD8+ T cells directed against this peptide were readily detectable in a large proportion of HLA-A2+ melanoma patients. These results provide new insight into the immunogenicity of MAGE antigens and underline the potential usefulness of MAGE-A10 peptide-based cancer vaccines.

Selecting highly affine and well-expressed TCRs for gene therapy of melanoma.

A recent phase 1 trial has demonstrated that the generation of tumor-reactive T lymphocytes by transfer of specific T-cell receptor (TCR) genes into autologous lymphocytes is feasible. However, compared with results obtained by infusion of tumor-infiltrating lymphocytes, the response rate observed in this first TCR gene therapy trial is low. One strategy that is likely to enhance the success rate of TCR gene therapy is the use of tumor-reactive TCRs with a higher capacity for tumor cell recognition. We therefore sought to develop standardized procedures for the selection of well-expressed, high-affinity, and safe human TCRs. Here we show that TCR surface expression can be improved by modification of TCR alpha and beta sequences and that such improvement has a marked effect on the in vivo function of TCR gene-modified T cells. From a panel of human, melanoma-reactive TCRs we subsequently selected the TCR with the highest affinity. Furthermore, a generally applicable assay was used to assess the lack of alloreactivity of this TCR against a large series of common human leukocyte antigen alleles. The procedures described in this study should be of general value for the selection of well- and stably expressed, high-affinity, and safe human TCRs for subsequent clinical testing.

Toll-like receptor 3 expressed by melanoma cells as a target for therapy?

PURPOSE: The immunomodulatory properties of Toll-like receptors (TLR) agonists have inspired their use as experimental adjuvants for vaccination of cancer patients. However, it is now well recognized that TLR expression is not restricted to immune cells but can also be found in many cell types, including those giving rise to tumors. It is therefore mandatory to explore the potential effects of TLR triggering directly on tumor cells. EXPERIMENTAL DESIGN: In the present work, we have investigated TLR3 protein expression in melanoma cell lines derived from patients, and analyzed the effects of TLR3 agonists on tumor cell survival. Moreover, we used RNA interference to stably knock down TLR3 expression and study the involvement of this receptor in dsRNA-induced effects on melanoma cells viability. RESULTS: Human melanoma cells can express functional TLR3 protein. Interestingly, the engagement of the receptor by TLR3 agonists can directly inhibit cell proliferation and induce tumor cell death when combined to treatment with either type I IFN or protein synthesis inhibitors. These effects were shown by RNA interference to be largely dependent on TLR3. Moreover, TLR3-mediated cell death involves the activation of caspases and engages both extrinsic and intrinsic apoptotic pathways. CONCLUSION: TLR3 protein can be expressed in human melanoma cells, where it can deliver proapoptotic and antiproliferative signaling. Altogether, these results suggest that TLR3 agonists represent very promising adjuvants for cancer vaccines not only based on their well-described immunostimulatory properties, but also due to their newly identified cytostatic and cytotoxic effects directly on tumor cells.

Crystal structures of two closely related but antigenically distinct HLA-A2/melanocyte-melanoma tumor-antigen peptide complexes.

We have determined high-resolution crystal structures of the complexes of HLA-A2 molecules with two modified immunodominant peptides from the melanoma tumor-associated protein Melan-A/Melanoma Ag recognized by T cells-1. The two peptides, a decamer and nonamer with overlapping sequences (ELAGIGILTV and ALGIGILTV), are modified in the second residue to increase their affinity for HLA-A2. The modified decamer is more immunogenic than the natural peptide and a candidate for peptide-based melanoma immunotherapy. The crystal structures at 1.8 and 2.15 A resolution define the differences in binding modes of the modified peptides, including different clusters of water molecules that appear to stabilize the peptide-HLA interaction. The structures suggest both how the wild-type peptides would bind and how three categories of cytotoxic T lymphocytes with differing fine specificity might recognize the two peptides.

Tissue homing and persistence of defined antigen-specific CD8+ tumor-reactive T-cell clones in long-term melanoma survivors.

Tumor antigen-specific cytotoxic T cells (CTLs) play a major role in the adaptive immune response to cancers. This CTL response is often insufficient because of functional impairment, tumor escape mechanisms, or inhibitory tumor microenvironment. However, little is known about the fate of given tumor-specific CTL clones in cancer patients. Studies in patients with favorable outcomes may be very informative. In this longitudinal study, we tracked, quantified, and characterized functionally defined antigen-specific T-cell clones ex vivo, in peripheral blood and at tumor sites, in two long-term melanoma survivors. MAGE-A10-specific CD8+ T-cell clones with high avidity to antigenic peptide and tumor lytic capabilities persisted in peripheral blood over more than 10 years, with quantitative variations correlating with the clinical course. These clones were also found in emerging metastases, and, in one patient, circulating clonal T cells displayed a fully differentiated effector phenotype at the time of relapse. Longevity, tumor homing, differentiation phenotype, and quantitative adaptation to the disease phases suggest the contribution of the tracked tumor-reactive clones in the tumor control of these long-term metastatic survivor patients. Focusing research on patients with favorable outcomes may help to identify parameters that are crucial for an efficient antitumor response and to optimize cancer immunotherapy.

Temozolomide combined with bevacizumab in metastatic melanoma. A multicenter phase II trial (SAKK 50/07)

Background: Single agent DTIC is the standard therapy for metastatic melanoma (MM) with response rates of 5−20%. Temozolomide (Tem) as an oral drug has shown equal efficacy in phase III trials. Preclinical models have shown an inhibitory effect for bevacizumab (Bev) on the proliferation of melanoma cells as well as on sprouting endothelial cells. Therefore, a therapeutic approach that combines angiogenesis inhibitors with cytotoxic agents may provide clinical benefit in MM. Methods: Design: Multicenter phase II trial. Primary endpoint: Clinical benefit (CR, PR and SD) at 12 weeks; secondary endpoints: best overall response by RECIST, response duration, progression free survival, adverse events, survival after 6 months and overall survival. Sample size was calculated according to Simon's two stage optimal design (5% significance level and 80% power) with an overall sample size of 62 patients (pts) to test H0: 20% versus H1: 35% rate of clinical benefit. Response assessment was done every 6 weeks (3 cycles). Eligibility: Stage IV MM, ECOG PS 0−2, no prior treatment for metastatic disease. Treatment regimen: One cycle consisted of Tem at 150 mg/m2 days 1−7 po and Bev at 10 mg/kg day 1 over 30 min iv and was repeated every 2 weeks until progression or unacceptable toxicity. Results: Between January 2008 and April 2009, 62 pts (40 male/22 female) at a median age of 61 years (range 30−86) with stage IV (M1a:4, M1b:12, M1c:46) melanoma were enrolled in 9 centers. The first 50 pts, who received 415 cycles are included in this interim report. The overall response rate was 26% (CR: 1 pt, PR: 12 pts; PR not confirmed yet in 3 pts), and 44% (22 pts) had stable disease over 1.5−7.5 months (median: 3). Only 30% (15 pts) had disease progression at the first evaluation at week 6. The hematological grade 3/4 toxicities according to NCI CTAE 3.0 were thrombocytopenia 10% (5 pts), neutropenia 8% (4 pts), lymphopenia and leucocytopenia each 2% (1 pt). Cumulative non-hematological toxicities grade 3/4 were nausea and fatigue each 6% (3 pts), hypertension, vomiting and hemorrhage, each 4% (2 pts), thrombosis/embolism, infection, constipation, anorexia, elevation of alkaline phosphatase, bilirubin, GGT, ALT and AST each 2% (1 pt). Conclusion: In metastatic melanoma the combination of Tem/Bev is a safe regimen with a promising efficacy and few grade 3/4 toxicities. Updated results of all 62 pts will be presented.

Enrichment of human CD4+ V(alpha)24/Vbeta11 invariant NKT cells in intrahepatic malignant tumors.

Invariant NKT cells (iNKT cells) recognize glycolipid Ags via an invariant TCR alpha-chain and play a central role in various immune responses. Although human CD4(+) and CD4(-) iNKT cell subsets both produce Th1 cytokines, the CD4(+) subset displays an enhanced ability to secrete Th2 cytokines and shows regulatory activity. We performed an ex vivo analysis of blood, liver, and tumor iNKT cells from patients with hepatocellular carcinoma and metastases from uveal melanoma or colon carcinoma. Frequencies of Valpha24/Vbeta11 iNKT cells were increased in tumors, especially in patients with hepatocellular carcinoma. The proportions of CD4(+), double negative, and CD8alpha(+) iNKT cell subsets in the blood of patients were similar to those of healthy donors. However, we consistently found that the proportion of CD4(+) iNKT cells increased gradually from blood to liver to tumor. Furthermore, CD4(+) iNKT cell clones generated from healthy donors were functionally distinct from their CD4(-) counterparts, exhibiting higher Th2 cytokine production and lower cytolytic activity. Thus, in the tumor microenvironment the iNKT cell repertoire is modified by the enrichment of CD4(+) iNKT cells, a subset able to generate Th2 cytokines that can inhibit the expansion of tumor Ag-specific CD8(+) T cells. Because CD4(+) iNKT cells appear inefficient in tumor defense and may even favor tumor growth and recurrence, novel iNKT-targeted therapies should restore CD4(-) iNKT cells at the tumor site and specifically induce Th1 cytokine production from all iNKT cell subsets.

Impact of 3 different short-term chemotherapy regimens on lymphocyte-depletion and reconstitution in melanoma patients.

Recent immunotherapy trials have shown that lymphodepletion induced by short-term chemotherapy favors subsequent expansion of adoptively transferred T cells, by homeostatic mechanisms. To take advantage of this effect, novel regimens are being developed with the aim to enhance tumor immunity and reduce treatment toxicity. We have designed a clinical phase I trial combining chemotherapy, reinfusion of PBMC containing Melan-A(MART-1)-specific T cells, and vaccination with Melan-A peptide in Incomplete Freund's Adjuvant. Treatment with Busulfan plus Fludarabine depleted lymphocytes only weakly. Cyclophosphamide (CTX) plus Fludarabine depleted lymphocytes more profoundly, with a maximal effect using high doses of CTX. It is interesting to note that, the degree of homeostatic T-cell proliferation correlated tightly with the extent of lymphodepletion. As compared with CD4 T cells, CD8 T cells showed higher susceptibility to chemotherapy, followed by more rapid homeostatic proliferation and recovery, resulting in strong inversions of CD4/CD8 ratios. Despite efficient homeostatic proliferation of total CD4 and CD8 T cells, the frequency of CD8 T cells specific for Melan-A and cancer-testis antigens remained relatively low. In contrast, EBV-specific T cells expanded and reached high numbers. We conclude that short-term chemotherapy promoted homeostatic lymphocyte proliferation depending on the intensity of lymphocyte depletion, however without preferential expansion of tumor antigen-specific T cells.

Predictive value of the MGMT promoter methylation status in metastatic melanoma patients receiving first-line temozolomide plus bevacizumab in the trial SAKK 50/07.

The O6-methylguanine-DNA-methyltransferase (MGMT) promoter methylation status is a predictive parameter for the response of malignant gliomas to alkylating agents such as temozolomide. First clinical trials with temozolomide plus bevacizumab therapy in metastatic melanoma patients are ongoing, although the predictive value of the MGMT promoter methylation status in this setting remains unclear. We assessed MGMT promoter methylation in formalin-fixed, primary tumor tissue of metastatic melanoma patients treated with first-line temozolomide and bevacizumab from the trial SAKK 50/07 by methylation-specific polymerase chain reaction. In addition, the MGMT expression levels were also analyzed by MGMT immunohistochemistry. Eleven of 42 primary melanomas (26%) revealed a methylated MGMT promoter. Promoter methylation was significantly associated with response rates CR + PR versus SD + PD according to RECIST (response evaluation criteria in solid tumors) (p<0.05) with a trend to prolonged median progression-free survival (8.1 versus 3.4 months, p>0.05). Immunohistochemically different protein expression patterns with heterogeneous and homogeneous nuclear MGMT expression were identified. Negative MGMT expression levels were associated with overall disease stabilization CR + PR + SD versus PD (p=0.05). There was only a poor correlation between MGMT methylation and lack of MGMT expression. A significant proportion of melanomas have a methylated MGMT promoter. The MGMT promoter methylation status may be a promising predictive marker for temozolomide therapy in metastatic melanoma patients. Larger sample sizes may help to validate significant differences in survival type endpoints.

Fine structural variations of alphabeta TCRs selected by vaccination with natural versus altered self-antigen in melanoma patients

Immunotherapy of cancer is often performed with altered "analog" peptide Ags optimized for HLA class I binding, resulting in enhanced immunogenicity, but the induced T cell responses require further evaluation. Recently, we demonstrated fine specificity differences and enhanced recognition of naturally presented Ag by T cells after vaccination with natural Melan-A/MART-1 peptide, as compared with analog peptide. In this study, we compared the TCR primary structures of 1489 HLA-A*0201/Melan-A26-35-specific CD8 T cells derived from both cohorts of patients. Although a strong preference for TRAV12-2 segment usage was present in nearly all patients, usage of particular TRAJ gene segments and CDR3 composition differed slightly after vaccination with natural vs analog peptide. Moreover, TCR β-chain repertoires were broader after natural than analog peptide vaccination. In all patients, we observed a marked conservation of the CDR3β amino acid composition with recurrent sequences centered on a glycyl-leucyl/valyl/alanyl-glycyl motif. In contrast to viral-specific TCR repertoires, such "public" motifs were primarily expressed by nondominant T cell clonotypes, which contrasted with "private" CDR3β signatures frequently found in T cell clonotypes that dominated repertoires of individual patients. Interestingly, no differences in functional avidity were observed between public and private T cell clonotypes. Collectively, our data indicate that T cell repertoires generated against natural or analog Melan-A peptide exhibited slightly distinct but otherwise overlapping and structurally conserved TCR features, suggesting that the differences in binding affinity/avidity of TCRs toward pMHC observed in the two cohorts of patients are caused by subtle structural TCR variations.

The role of interval nodes in sentinel lymph node mapping and dissection for melanoma patients.

In sentinel node (SN) biopsy, an interval SN is defined as a lymph node or group of lymph nodes located between the primary melanoma and an anatomically well-defined lymph node group directly draining the skin. As shown in previous reports, these interval SNs seem to be at the same metastatic risk as are SNs in the usual, classic areas. This study aimed to review the incidence, lymphatic anatomy, and metastatic risk of interval SNs. METHODS: SN biopsy was performed at a tertiary center by a single surgical team on a cohort of 402 consecutive patients with primary melanoma. The triple technique of localization was used-that is, lymphoscintigraphy, blue dye, and gamma-probe. Otolaryngologic melanoma and mucosal melanoma were excluded from this analysis. SNs were examined by serial sectioning and immunohistochemistry. All patients with metastatic SNs were recommended to undergo a radical selective lymph node dissection. RESULTS: The primary locations of the melanomas included the trunk (188), an upper limb (67), or a lower limb (147). Overall, 97 (24.1%) of the 402 SNs were metastatic. Interval SNs were observed in 18 patients, in all but 2 of whom classic SNs were also found. The location of the primary was truncal in 11 (61%) of the 18, upper limb in 5, and lower limb in 2. One patient with a dorsal melanoma had drainage exclusively in a cervicoscapular area that was shown on removal to contain not lymph node tissue but only a blue lymph channel without tumor cells. Apart from the interval SN, 13 patients had 1 classic SN area and 3 patients 2 classic SN areas. Of the 18 patients, 2 had at least 1 metastatic interval SN and 2 had a classic SN that was metastatic; overall, 4 (22.2%) of 18 patients were node-positive. CONCLUSION: We found that 2 of 18 interval SNs were metastatic: This study showed that preoperative lymphoscintigraphy must review all known lymphatic areas in order to exclude an interval SN.

Proteasome-assisted identification of a SSX-2-derived epitope recognized by tumor-reactive CTL infiltrating metastatic melanoma.

The tumor Ag SSX-2 (HOM-MEL-40) was found by serological identification of Ags by recombinant expression cloning and was shown to be a cancer/testis Ag expressed in a wide variety of tumors. It may therefore represent a source of CD8(+) T cell epitopes useful for specific immunotherapy of cancer. To identify potential SSX-2-derived epitopes that can be recognized by CD8(+) T cells, we used an approach that combined: 1) the in vitro proteasomal digestion of precursor peptides overlapping the complete SSX-2 sequence; 2) the prediction of SSX-2-derived peptides with an appropriate HLA-A2 binding score; and 3) the analysis of a tumor-infiltrated lymph node cell population from an HLA-A2(+) melanoma patient with detectable anti-SSX-2 serum Abs. This strategy allowed us to identify peptide SSX-2(41-49) as an HLA-A2-restricted epitope. SSX2(41-49)-specific CD8(+) T cells were readily detectable in the tumor-infiltrated lymph node population by multimer staining, and CTL clones isolated by multimer-guided cell sorting were able to lyse HLA-A2(+) tumor cells expressing SSX-2.