NY-ESO-1 is a ubiquitous immunotherapeutic target antigen for patients with myxoid/round cell liposarcoma


  • The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health.



Myxoid/round cell liposarcoma (MRCL) is the second most common liposarcoma subtype, accounting for >33% of liposarcomas and approximately 10% of all soft tissue sarcomas. Although MRCL is a chemosensitive subtype, patients with metastatic disease have a poor outcome. NY-ESO-1 is a cancer-testis antigen (also known as cancer germ cell antigen) that has been successfully targeted in vaccine trials and in adoptive T-cell therapy trials for the treatment of several solid tumors.


The authors investigated the feasibility of targeting NY-ESO-1 in patients with MRCL by evaluating the prevalence of NY-ESO-1 expression in tumors using immunohistochemistry and quantitative reverse transcriptase-polymerase chain reaction analysis. NY-ESO-1–specific tumor recognition by NY-ESO-1–specific T-cells also was analyzed using a chromium release assay.


A search of the University of Washington Sarcoma Tissue Bank identified paraffin-embedded tumor samples from 25 patients with MRCL. NY-ESO-1 expression was observed in every MRCL tumor assessed (100%); in 18 tumors (72%), staining was homogenous. In all but 2 tumors, staining was sufficiently robust (2+) that such patients would be eligible for clinical trials of NY-ESO-1–directed therapy. By using NY-ESO-1 specific, CD8-positive T-cells, the in vitro sensitivity of myxoid liposarcoma cell lines to antigen-specific lysis was demonstrated.


The current results establish NY-ESO-1 as an important target antigen for the treatment of patients with MRCL. Cancer 2012. © 2012 American Cancer Society.


On the basis of its immunogenicity, the cancer-testis antigen NY-ESO-1 is considered to be among the most attractive antigens for immunotherapy. It has been targeted in several clinical studies, including several vaccine trials that have induced serologic, cluster of differentiation 4 (CD4)-positive, and CD8-positive T-cell responses. Delayed type hypersensitivity responses after NY-ESO-1 vaccination have been associated with long-term survival.1, 2 Objective clinical responses have been observed in patients with melanoma after vaccination against NY-ESO-1, including 1 complete response.3

NY-ESO-1 also has been successfully targeted in trials of antigen-specific adoptive T-cell therapy. For example, transfer of NY-ESO-1–specific CD4-positive cells has been used effectively to treat patients with metastatic melanoma.4 NY-ESO-1 also has been targeted using a class I T-cell receptor (TCR) retrovirally transfected into T cells,5 inducing complete responses in patients with melanoma. To date, there have been no known grade III or grade IV autoimmune toxicities associated with anti-NY-ESO-1 therapy.

NY-ESO-1, a member of the family of cancer testis antigens (CT antigens), was first discovered through serologic analysis in patients with esophageal cancer; subsequently, it was observed that NY-ESO-1 induced a strong cytotoxic T-cell response.6-8 CT antigens (also sometimes referred to as cancer germ-cell antigens), as their name implies, are expressed at the protein level in various malignant tumors and germ cells of the testis but not in other adult tissues.

Soft tissue sarcomas are a heterogeneous group of malignancies of mesenchymal origin with a poor prognosis in the metastatic setting and a median overall survival <1 year.9, 10 Liposarcomas account for approximately 10% to 20% of soft tissue sarcomas and can be classified into 3 subtypes, each with their own distinct clinical behavior: pleomorphic, well-/dedifferentiated, and myxoid/round cell liposarcoma (MRCL). MRCL accounts for 40% to 50% of liposarcomas and is almost always associated with a chromosomal translocation, most commonly t(12;16)(q13;p11), although several less common translocations also have been described.11 The resultant fusion proteins have an activity that is not well understood.12 MRCL is relatively sensitive to frontline chemotherapy, and trabectedin as second-line treatment for metastatic MRCL is promising.10, 13 However, mortality remains high for patients with metastatic disease, suggesting the need for novel approaches.

The discovery that >80% of synovial sarcomas express NY-ESO-1,14 often homogenously, established synovial sarcoma as a malignancy with 1 of the highest rates of NY-ESO-1 expression and led to the perception that synovial sarcoma is a model disease for the study of NY-ESO-1–directed therapy. Supporting this concept, a recently published adoptive therapy trial5 using retrovirally transfected, NY-ESO-1–specific TCR, documented 4 partial responses in 6 patients with synovial sarcoma. Here, we report that another soft tissue sarcoma subtype, MRCL, ubiquitously expresses the cancer-testis antigen NY-ESO-1, most often homogenously, raising the possibility of NY-ESO-1–directed therapy for patients with this challenging disease, for whom there are limited treatment options in the metastatic setting.


Tumor Samples

Both paraffin-embedded and flash-frozen samples from patients with MRCL were obtained through the University of Washington sarcoma tumor bank (Institutional Review Board-approved protocol 21369).


Immunohistochemistry was performed on formalin-fixed, paraffin-embedded tissues. The monoclonal antibody E978 was used to detect NY-ESO-1, as previously described.15 For all assays, appropriate positive controls (normal testis with preserved spermiogenesis) and negative controls (omission of primary antibody and replacement with phosphate buffer saline, pH 7.4) were included.

Tissue sections were deparaffinized in xylene and rehydrated in a series of graded alcohols. Endogenous peroxidase was blocked by incubating slides for 30 minutes at room temperature in 99.7% methanol containing 0.3% hydrogen peroxide. Slides were washed with Tris-buffered saline solution and blocked in 2% bovine serum albumin at room temperature for 5 minutes to prevent unspecific protein interactions. We used a heat-based antigen-retrieval technique with a commercial vegetable steamer to heat the slides in a buffer solution for 30 minutes at approximately 96°C. Tissue slides were incubated with primary antibodies overnight at 4°C in a wet chamber. Primary antibody detection was performed by using of a Novolink polymer detection kit (Leica Microsystems Inc., Bannockburn, Ill) according to the manufacturer's instructions. 3,3′-Diaminobenzidine served as a chromogen, and counterstains were done with Harris hematoxylin. Finally, slides were dehydrated in a series of graded ethanols and coverslipped.

The slides were examined under a light microscope by 2 pathologist, including an experienced bone and soft tissue pathologist. Characteristic morphologic features of myxoid liposarcoma were confirmed, including myxoid stroma, primitive mesenchymal cells, lipoblasts, and an arborizing capillary vasculature. Samples that had staining present in <5% of cells were considered focally positive. The presence of staining in 5% to 25% of cells was considered 1+ staining, the presence of staining in 25% to 50% of cells was considered 2+ staining, the presence of staining in 50% to 75% of cells was considered 3+ staining, and the presence of staining in >75% of cells was considered 4+ staining.

Quantitative Reverse Transcriptase-Polymerase Chain Reaction Analysis

RNA was extracted from frozen tumor samples using Trizol (Invitrogen Life Technologies, Carlsbad, Calif) and from cell lines using the RNeasy kit (Qiagen, Valencia, Calif). Because of various tissue-collection conditions, RNA quality was recorded before analysis using either a gel or a bioanalyzer. Samples with poor-quality RNA were not analyzed further. One nonmyxoid liposarcoma sample was invading into the spermatic cord and, thus, was not included in the analysis because of concerns about a false-positive result.

RNA samples were converted to combinational DNA (cDNA) using the Transcriptor First Strand cDNA Synthesis Kit (Roche, Indianapolis, Ind). The results were analyzed using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as a housekeeping gene and were calculated relative to testis using the standard curve method. The NY-ESO-1 primers were TGCTTGAGTTCTACCTGCCA and TATGTTGCCGGACACAGTGAA.16 GAPDH primers were GAAGGTGAAGGTCGGAGTC and GAAGATGGTGATGGGATTTC.17 In all MRCL tumors, NY-ESO-1 expression also was confirmed and quantified using primers from SA Biosciences (Frederick, Md). Amplification was performed using SYBR Green PCR Master Mix (Applied Biosystems, Foster City, Calif) on an ABI 7900HT (Applied Biosystems).

Antigen-Specific T-Cells

NY-ESO-1–specific effectors were generated from a patient with human leukemic antigen (HLA) A*0201 (HLA-A*0201)-positive synovial sarcoma who underwent leukopheresis according to established protocols (Fred Hutchinson Cancer Research Center, Protocol 1246). Peripheral blood mononuclear cell (PBMC)-derived dendritic cells18 were pulsed with the NY-ESO-1 peptide SLLMWITQC. PBMCs were depleted of CD25-positive T-cells using CliniMACS CD25 MicroBeads (Miltenyi Biotech, Auburn, Calif) according to manufacturer's instructions and were stimulated using interleukin 21 as previously described.19 NY-ESO-1–positive cells were sorted using NY-ESO-1 tetramer, then cloned with limited dilution, and expanded using a rapid expansion protocol.20 T-cell clones that were specific for melanoma antigen recognized by T-cells (MART-1) were used as control effector cells.

Cell Lines

The human myxoid liposarcoma cell lines 402 and 1765 (gifts of Pierre Aman) have been described previously.21 They were maintained in RPMI 1640 medium containing 25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 2 mM L-glutamine, 50 U/mL penicillin, and 50 mg/mL streptomycin (Invitrogen) in 10% fetal bovine serum. The T2 cell line—a peptide transporter (TAP)-deficient T–B-cell hybrid that expresses the HLA-A2 allele—was used unpulsed as a negative control and was pulsed with target peptide as a positive control.

Chromium Release Assay Using Recombinant Vaccinia–HLA-A2–Transfected Targets

Because the target cells (cell lines 402 and 1765) did not express HLA-A*0201, we used recombinant vaccinia to endow target cells with the presenting HLA-restricting element according to established methods.22 The NY-ESO-1–positive MRCL tumor lines 402 and 1765 were transfected with the HLA-A*0201 gene using a vaccinia vector at a multiplicity of infection (MOI) of 2.5, which was identified as the optimal MOI in titration experiments. For a control against cell lysis resulting from the vaccinia infection, the cell lines also were treated with wild-type Vaccinia virus. For the chromium release assay, cell lines were labeled with 100 μCi 51Cr and cocultured with effector cells for 4 to 6 hours at 37°C in a 5% CO2 atmosphere.


Myxoid/Round Cell Liposarcoma Samples Expressed NY-ESO-1 in 100% of Samples, and Homogenous Expression Was Observed in >70% of Patients

We stained 25 MRCL tumors for NY-ESO-1, and all were positive for NY-ESO-1 expression. In 2 tumors, only 1+ staining was observed (5%-25% of cells were positive for staining). In the remaining 23 tumors, at least 2+ staining was observed (2+ staining has been the cutoff for clinical trial eligibility in some trials of NY-ESO-1–directed therapy5). In 6 of 25 tumors, flash-frozen tissue was available for RNA extraction and quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) analysis. In all of these tissues, NY-ESO-1 expression was confirmed (Table 1).

Table 1. NY-ESO-1 Expression and Cytogenetics of MRCL Tumors
PatientCytogeneticsSite of Primary TumorPrimary or Metastatic SpecimenOutcomeNY-ESO-1 StainingaqRT-PCR Confirmed
  • Abbreviations: t, translocation; NY-ESO-1, cancer-testis antigen; qRT-PCR, quantitative reverse transcriptase-polymerase chain reaction.

  • a

    Staining scores were based on the percentage of cells that stained positive for the cancer-testis antigen NY-ESO-1: focal staining, <5%; 1+, 5% to 25%; 2+, 25% to 50%; 3+, 50% to 75%; 4+, >75%.

1UnavailableRight thighPrimaryMetastatic Disease, died3+Not available
2UnavailableRight hipPrimaryMetastatic disease3+Not available
3UnavailableLeft buttockPrimaryLocal recurrence, currently disease free3+Not available
4UnavailableRight thighMetastaticMetastatic disease2+Yes
5UnavailableLeft popliteal fossaPrimaryMetastatic disease, died4+Not available
6t(12;16) With other complex cytogeneticsRight groinPrimaryMetastatic disease, died4+Not available
7t(12;16)(q13;p11)Left popliteal fossaPrimaryDisease free1+Not available
8t(12;16) With other complex cytogeneticsLeft thighPrimaryMetastatic disease, died3+Not available
9t(10;16;12)(q26;p11;q13)Left calfPrimaryDisease free2+Not available
10UnavailableRight thighPrimaryDisease free4+Not available
11t(12;16)(q13;p11)Right thighPrimaryLocal recurrence, currently disease free2+Not available
12t(12;16)(q13;p11)Right thighPrimaryIsolated recurrence, currently disease free4+Not available
13t(12;16)(q13;p11)Left popliteal fossaPrimaryDisease free3+Yes
14t(12;16) With other complex cytogeneticsRight tibiaPrimaryDisease free4+Not available
15t(12;16)(q13;p11)Right thighPrimaryDisease free4+Not available
16t(12;22)(q13;q12)Right great toePrimaryDisease free2+Not available
17UnavailableLeft legMetastaticMetastatic disease4+Not available
18t(12;16) With other complex cytogeneticsRight thighPrimaryDisease free1+Not available
19t(12;16)(q13;p11)Right pelvisPrimaryMetastatic disease4+Not available
20NormalLeft thighMetastaticMetastatic disease3+Not available
21t(12;16) With other complex cytogeneticsLeft axillaPrimaryDisease free4+Yes
22t(12;16)(q13;p11)Left glutealPrimaryMetastatic disease3+Not available
23t(15;17)(q22;q23)Left thighPrimaryDisease free3+Yes
24t(12;16)(q13;p11)Right thighPrimaryDisease free3+Yes
25t(12;16) With other complex cytogeneticsLeft ankle/fibulaPrimaryMetastatic disease2+Yes

In 18 of 25 tumors (72%), staining was homogenous (3+ or 4+), ie, staining was observed in >50% of the tumor. Nine of these tumors had 4+ staining. In all but 3 tumors, the primary tumor was stained from a resection specimen; however, in the 3 tumors from patients who had metastatic disease, staining appeared to have at least similar pattern of expression (4+ staining in 2 tumors, 2+ staining in 1 tumor). At least 11 patients subsequently developed metastatic disease (several patients were lost to follow-up), but we were unable to observe a clear correlation between disease outcome and staining intensity.

It is noteworthy that, although most patients had the classic t(12;16)(q13;p11) translocation, there were 4 patients with different cytogenetics, including t(10;16;12), t(12;22), t(15;17), and 1 patient with normal cytogenetics (for that patient, the evaluation was repeated and confirmed, and the growth of normal host cells could not be excluded). All of these patients had NY-ESO-1 staining regardless of karyotype.

To determine whether NY-ESO-1 expression was limited to the MRCL subtype, we tested 7 nonmyxoid liposarcoma specimens using immunohistochemistry and also tested frozen samples from 3 (additional) patients by qRT-PCR. These 10 tumors included 2 pleomorphic liposarcomas, 4 well differentiated tumors, and 4 dedifferentiated tumors. None of these tumors expressed NY-ESO-1. Although these negative results are not definitive, they do suggest that the high prevalence of NY-ESO-1 expression observed in MRCL tumors is not shared among the other liposarcoma subtypes.

It is noteworthy that, although both myxoid and round cell areas of the tumor tended to stain positive for NY-ESO-1, a more uniform and intense appearance of the staining was observed in the round cell component. Also of interest was that the well differentiated elements of some tumors stained less homogenously (1+ to 2+) than the more typical myxoid and round cell components within the same tumor (Fig. 1).

Figure 1.

(A) Round cell components often had intense and uniform staining; (B) however, myxoid components frequently were stained homogenously. (C) In tumors with both myxoid and round cell components, both segments stained homogenously but sometimes more intensely in the round cell compartment. (D) In some of the lesser staining tumors, well differentiated areas of mature appearing fat were sometimes present.

Myxoid/Round Cell Liposarcoma Cell Lines Expressing NY-ESO-1 Can Be Recognized and Specifically Lysed by NY-ESO-1–Specific Effectors

The MRCL cell lines 402 and 1765 were analyzed by quantitative PCR, and NY-ESO-1 messenger RNA transcripts were expressed at levels even higher than those observed in testis normalized to GAPDH (2.5-fold and 3.5-fold higher). Both cell lines underwent class I typing at the Puget Sound Blood Center and were negative for HLA-A*02. Because HLA-A2–restricted, NY-ESO-1–specific cytotoxic T-lymphocytes were used to evaluate antigen recognition, the 402 and 1765 cells were preinfected with a recombinant Vaccinia virus that expressed HLA-A*0201 (Vac-A2). When treated with Vac-A2, the cells were lysed at >30% after 4 hours with an effector-to-target ratio of 20:1. Control MART-1–specific effectors were unable to kill either cell line. Similarly, MRCL cell lines transfected with wild-type Vaccinia virus were not sensitized to lysis by NY-ESO-1–specific effectors (Fig. 2).

Figure 2.

Effectors of the cancer-testis antigen NY-ESO-1 recognize and specifically lyse myxoid/round cell liposarcoma (MRCL) cell lines at an effector-to-target ratio of 20:1 after transfection with Vaccinia virus expressing human leukemic antigen-A*0201 (HLA-A*0201). Melan-A (melanoma antigen recognized by T-cells [MART-1]) effectors and wild-type (WT) Vaccinia virus were used as controls.


NY-ESO-1 is widely considered an attractive target for immunotherapy. Complete responses have been observed in melanoma trials that targeted NY-ESO-1 using both vaccines as well as adoptively transferred T-cells. The discovery that 80% of synovial sarcomas express NY-ESO-1 was rapidly translated into a clinical trial; the National Cancer Institute surgery branch treated 6 patients with synovial sarcoma using T-cells transfected with a retrovirus that expressed the NY-ESO-1 TCR. Partial responses were observed in 4 of 6 synovial sarcoma patients.3-5

Here, we report another soft tissue sarcoma subtype that demonstrates a pattern of NY-ESO-1 expression that is even more prevalent than the pattern observed in synovial sarcoma. Although other reports have included NY-ESO-1 expression in liposarcomas, including MRCL, generally, this is the first study to specifically examine NY-ESO-1 protein expression in MRCL.23-28 On the basis of our analysis of 25 samples, MRCL appears to express NY-ESO-1 with a frequency that is unmatched by any other malignancy studied to date. Furthermore, a high proportion (9 of 25 tumors) had 4+ staining (>75% of cells), and an additional 9 of 25 tumors had 3+ staining (>50% of cells). It is noteworthy that patients who had histopathologic phenotypes of MRCL that included a variety of chromosomal translocations were included in this analysis, and all had tumors that expressed NY-ESO-1. Furthermore, we demonstrated that MRCL cell lines are capable of presenting NY-ESO-1 peptide, such that it can be recognized by NY-ESO-1–specific effectors in vitro initiating cell-mediated lysis of tumor cells.

MRCL generally is associated with a characteristic fusion protein; however, the role played by the mutation in oncogenesis is not clear. Most tumors contain t(12;16) (q13;p11), producing the fused in sarcoma (FUS)-C/EBP homologous protein (FUS-CHOP) fusion protein, although a notable minority contains t(12;22)(q13;12) translocation-associated Ewing sarcoma (EWS) breakpoint region 1 (EWSR1)-CHOP. Both FUS and EWS (along with TF15 RNA polymerase II, TATA box binding protein-associated factor [TAF15]) are in the FET family (also known as the TET family) of RNA binding proteins.29 However, although the RNA binding profiles of the FET family proteins are remarkably similar to one another,30 Ewing sarcomas, which typically have translocations of EWS,31, 32 do not generally express NY-ESO-1.33

There is evidence to suggest that murine adipocyte-derived mesenchymal stem cells transfected with a FUS-CHOP gene develop an MRCL phenotype; however, the translocation alone was insufficient to induce an MRCL tumor-like phenotype using human adipocyte-derived mesenchymal stem cells transfected with FUS-CHOP, suggesting the need for additional genetic “hits.”34-37 Similar to MRCL, in synovial sarcoma models, the presence of synovial sarcoma translocation t(X;18) (SYT-SSX) alone appears to be insufficient on its own to cause oncogenesis.38 Mesenchymal stem cells also have been postulated as potential cells of origin in synovial sarcoma.39

To our knowledge, there has never been a study of NY-ESO-1–specific serologic response in patients with MRCL.28, 40 An analysis of the serologic response to several CT antigens, including NY-ESO-1, was reported in 54 patients with sarcoma, including 5 patients with synovial sarcoma and 1 patient with MRCL. In that report, serology was negative except for 2 patients (1 with pleomorphic sarcoma and another with fibrosarcoma).41 Serology was also analyzed in a study by Ayyoub et al that included 1 patient with liposarcoma, although the histologic subtype was not mentioned.28

The expression of CT antigens has been correlated with outcomes in several malignancies.42-44 Although the numbers in our current study would be under powered to perform an adequate analysis assessing differences in outcome patterns between patients with strong and weak NY-ESO-1 expression, we are assessing the feasibility of this approach in our patient population. We also are assessing ways to apply this knowledge to preclinical models, such as MRCL xenografts, to advance NY-ESO-1–directed immunotherapy for patients with sarcoma.45

To our knowledge, no other malignancy, including synovial sarcoma, has been described that has NY-ESO-1 expression in 100% of tumors or such a high proportion homogenous expression. We believe that, like synovial sarcoma, these results will establish MRCL as a model disease for the study of NY-ESO-1–directed therapy.


This work is supported by the Bob and Eileen Gilman Family Sarcoma Research Program as well as the Walker Immunotherapy Research Fellowship and the Sarcoma Alliance for Research Through Collaboration (SARC) Career Development Award. The University of Washington Tissue Bank is supported by grant RO1 CA65537-16. Seth M. Pollack is a recipient of the SARC Career Development Award as well as the Walker Immunotherapy Research Fellowship. Marie Bleakley is the Damon Runyon-Richard Lumsden Foundation Clinical Investigator and is supported in part by the Damon Runyon Cancer Research Foundation (CI-57-11) and in part by K23CA154532-01 from the National Cancer Institute. Eve Rodler is supported by Abbott Laboratories. Janet F. Eary is supported grant RO1 CA65537-16. Robin L. Jones is supported by the Bob and Eileen Gilman Family Sarcoma Research Program. Cassian Yee is a recipient of a Burroughs Wellcome Fund Clinical Scientist Award in Translational Research.


The authors made no disclosures.