Immunomic analysis of human sarcoma

Abstract

The screening of cDNA expression libraries from human tumors with serum antibody (SEREX) has proven to be a powerful method for identifying the repertoire of tumor antigens recognized by the immune system of cancer patients, referred to as the cancer immunome. In this regard, cancer/testis (CT) antigens are of particular interest because of their immunogenicity and restricted expression patterns. Synoivial sarcomas are striking with regard to CT antigen expression, with >80% of specimens homogeneously expressing NY-ESO-1 and MAGE-A3. In the present study, 54 sarcoma patients were tested for serum antibodies to NY-ESO-1, SSX2, MAGE-A1, MAGE-A3, MAGE-A4, MAGE-A10, CT7, and CT10. Two patients had detectable antibodies to CT antigens, and this seroreactivity was restricted to NY-ESO-1. Thus, although highly expressed in sarcoma, CT antigens do not induce frequent humoral immune responses in sarcoma patients. Sera from these two patients were used to immunoscreen cDNA libraries from two synovial sarcoma cell lines and normal testis, resulting in the identification of 113 distinct antigens. Thirty-nine antigens were previously identified by SEREX analysis of other tumor types, and 23/39 antigens (59%) had a serological profile that was not restricted to cancer patients, indicating that only a proportion of SEREX-defined antigens are cancer-related. A novel CT antigen, NY-SAR-35, mapping to chromosome Xq28 was identified among the cancer-related antigens, and encodes a putative extracellular protein. In addition to testis-restricted expression, NY-SAR-35 mRNA was expressed in sarcoma, melanoma, esophageal cancer, lung cancer and breast cancer. NY-SAR-35 is therefore a potential target for cancer vaccines and monoclonal antibody-based immunotherapies.

The identification of human tumor antigens recognized by the autologous host is yielding an array of target molecules for the diagnosis, monitoring, and immunotherapy of human cancer (1–4). Studies of the cellular and humoral immune response to cancer have revealed an extensive repertoire of tumor antigens recognized by the immune system, collectively termed the cancer immunome. The immunome comprises antigens defined by T cell epitope cloning (5–8), MHC peptide elution (9–11), and serological expression cloning (SEREX, refs. 12–14), and is catalogued in three databases, the peptide database of T cell-defined tumor antigens (www.cancerimmunity.org/peptidedatabase/Tcellepitopes.htm), SYFPEITHI database of MHC ligands and peptide motifs (www.bmi-heidelberg.com/syfpeithi/), and the cancer immunome database (www2.licr.org/CancerImmunomeDB/).

SEREX is a method of immunoscreening tumor-derived cDNA expression libraries with cancer patient sera to identify molecules recognized by high titered IgG antibodies (12). Approximately 1,100 distinct antigens have been defined by SEREX analysis to date, including a number of etiologically and therapeutically significant cancer antigens, such as mutational antigens (e.g., p53, LKB1; refs. 14 and 15), differentiation antigens (e.g., tyrosinase, NY-BR-1; refs. 12 and 16), overexpressed gene products (e.g., Her2neu, TPD52; ref. 14), and cancer/testis (CT) antigens (e.g., MAGE-1, NY-ESO-1, refs. 12 and 13). CT antigens are the products of transcripts present only in developing germ cells and human cancers of diverse origins (17) that elicit spontaneous cellular (18) and humoral immune (19) responses in some cancer patients. Because of their tissue-restricted expression and immunogenicity, CT antigens are potential targets for vaccine-based immunotherapies. In general, CT antigens are expressed in 20–40% of specimens from a given tumor type (18, 20). One exception to this is synovial sarcoma, in which 80% of specimens express the CT antigens, NY-ESO-1 (21), and MAGE (22).

The current study examines the humoral immune response of sarcoma patients to sarcoma and testicular antigens. Sera from patients actively mounting a humoral immune response to their cancers, as determined by an antibody response against NY-ESO-1, were used to screen cDNA libraries derived from CT-rich synovial sarcoma cell lines, as well as normal testis. Although there was little overlap in the identity of clones isolated with different sarcoma sera, more than 30% of the isolated clones were previously identified during SEREX analysis of other tumor types. In conformity with previous findings (14, 15), ≈60% of these antigens also reacted with sera from normal individuals. Thus, only a fraction of the serologically defined immunome is associated with a cancer-related immune response. A previously uncharacterized CT antigen, NY-SAR-35, was identified among the antigens associated with a cancer-related immune response. This antigen is striking, as it appears to be a rare example of a cell surface/secreted molecule identified as a CT antigen. It is therefore of significant interest as a potential immunotherapeutic target.

Materials and Methods

Cell Lines, Tissues, Sera, and RNA.

SW1045 and SW982 synovial sarcoma cell lines were obtained from the cell repository of the Ludwig Institute for Cancer Research, New York Branch at the Memorial Sloan–Kettering Cancer Center. Tumor tissues and sera were obtained from Memorial Sloan–Kettering Cancer Center, Weill Medical College of Cornell University, and Aichi Cancer Center Research Center, Nagoya, Japan. Normal tissue RNA preparations were purchased from CLONTECH and Ambion (Austin, TX). Total RNA from tumor tissues was prepared with guanidinium thiocyanate.

SEREX Analysis of cDNA Expression Libraries.

Poly(A)+ RNA from SW1045 and SW982 cells was prepared by using the Fast Track mRNA purification kit (Invitrogen, Life Technologies, Carlsbad, CA). Poly(A)+ RNA from normal testis was purchased from CLONTECH. Separate cDNA libraries were constructed for each of these in the ZAP Express vector (Stragene) according to the manufacturer's instructions using 5 μg poly(A)+ mRNA. Libraries containing 1–2 × 10⁶ recombinants were obtained, and were not amplified before imunoscreening.

To remove serum antibodies reactive with vector-related antigens, sera was absorbed with Escherichia coli/bacteriophage lysates prepared in the following manner. Wild-type λ ZAP Express bacteriophage were amplified in E. coli XL1 Blue MRF′ by overnight propagation of 5,000 plaque-forming units in 15-cm Petri dishes containing 100 ml of NZY/0.7% agarose growth media. Ten milliliters of binding buffer (0.1M NaHCO3, pH 8.3) was then added to the plates, and the plates were gently agitated at 4°C for 15 h. The resultant supernatants were collected, and residual E. coli were lysed by sonication. The lysates were then coupled to CNBr-Sepharose 4B (Amersham Pharmacia) as per manufacturer's instructions. Patient sera were absorbed with an equal volume of Sepharose 4B coupled E. coli/phage lysates at 4°C for 6 h. This procedure was repeated a total of three times, and was followed by a 15-h incubation with nitrocellulose filters precoated with proteins derived from E. coli and E. coli/phage lysates.

Library screenings were performed as described (14, 15). A total of five independent SEREX immunoscreenings of the cDNA librarieas were undertaken by using 1.75 × 10⁶ primary (nonamplified) bacteriophage clones and sera from two sarcoma patients. Serum reactive phage clones were converted to plasmid forms and subjected to DNA sequencing (Cornell University DNA Services, Ithaca, NY).

Determination of Serum Antibody Reactivity.

Serum antibody reactivity to CT antigens was determined by ELISA as described (19). Briefly, NY-ESO-1, SSX2, MAGE-A1, MAGE-A3, MAGE-A4, MAGE-A10, CT7, and CT10 were produced in E. coli by transfection with pQE30 expression vectors (Qiagen, Valencia, CA) as per the manufacturer's protocol. Ten nanograms of recombinant protein (1 μg/ml) was absorbed to TC microwell plates and incubated with diluted (1:100 to 1:25,000) patient sera. Bound antibody was detected with an alkaline phosphatase-conjugated goat anti-human IgG secondary antibody (Southern Biotechnology, Birmingham, AL). In the case of SEREX-defined sarcoma antigens, SADA (Serum Antibody Detection Array, refs. 14 and 23) was used to determine serological reactivity in preabsorbed serum samples from 39 sarcoma patients and 33 healthy blood donors. In brief, 5 × 10³ plaque-forming units per μl of bacteriophage encoding individual SEREX-defined tumor antigens were mixed with an equal volume of exponentially growing E. coli XL-1 Blue MRF′, and spotted on NZY coated nitrocellulose membranes by using a 96-pin replicator (Nalge Nunc). Membranes were incubated for 15 h at 37°C, and then processed as per the standard SEREX protocol (14, 15).

RT-PCR Analysis.

The cDNA preparations used as templates in the RT-PCR reactions were prepared by using 2.5 μg of total RNA in conjunction with the Superscript first strand synthesis kit (Invitrogen, Life Technologies). PCR primers specific for select SEREX-defined sarcoma antigens are listed below. The DNA sequences of PCR primers specific for NY-ESO-1, LAGE-1, MAGE-1, MAGE-3, MAGE-4, MAGE-10, SCP-1, BAGE, CT7, SSX1, SSX2, and SSX4, correspond to published primer sequences (5, 6, 8, 13, 20, 25–27). Twenty-five-microliter PCR mixtures, consisting of 2 μl of cDNA, 0.2 mM dNTP, 1.5 mM MgCl₂, 0.25 μM gene specific forward and reverse primers, and 2.5 units of Platinum TaqDNA polymerase (Invitrogen Life Technologies), were heated to 94°C for 2 min, followed by 35 thermal cycles of 94°C for 30 s, 55°C for 30 s and 72°C for 1 min (final cycle: 72°C for 5 min) by using a GeneAmp PCR System 9700 (Applied Biosystems). Resultant products were analyzed in 2% agarose/Tris-acetate-EDTA gels.

Primers were as follows. NY-SAR-12 5′/3′: TGGCGCAGAAAGGAAAAGGAAAAT/AGAGGTAGCTGGCAGGATGTTAG. NY-SAR-35 5′/3′: CTTGGTGCGATCAGCCTTAT/TTGATGCATGAAAACAGAACTC. NY-SAR-41 5′/3′: AGAATTGGCAGAGGCTCGTCATCA/TTCCAATTTTGCCTTCTCTAACTG. NY-SAR-73 5′/3′: CCCGGAGCACGTCGAGGTCTAC/GGTGAGGGGCCCAGGAAGC. NY-SAR-78 5′/3′: CACAATGTATCCTGTTGAAAG/GAGATGATACATTCTTCCAG. NY-SAR-92 5′/3′: CTTCCGCCAACTCCTCCTACC/GATGCCCGTGTCTTGTCCTT. NY-SAR-96 5′/3′: CACTAGGCTGCTGAGGAAGAT/GTTTTGGTGGGCAGCATTGAG. NY-SAR-97 5′/3′: GGACCACCCCAAATAGAA/CCACCAGCTCAGGAAGA. NY-SAR-110 5′/3′: TCTGATGGAGCGGTGGGATGC/GTGTGCCTCGGCTTCTTTCTTC.

Real-Time Quantitative RT-PCR.

The relative amount of NY-SAR-35 mRNA was measured by real-time RT-PCR using normalized cDNA preparations derived from nonmalignant tissues and lung cancer specimens. Gene-specific TaqMan probes and PCR primers were designed by using Primer Express software (PE Biosystems, Foster City, CA), and their sequences are provided below. Triplicate PCR reactions were prepared by using 2.5 μl of cDNA diluted in TaqMan PCR Master Mix (PE Biosystems) supplemented with 200 nM 6-carboxy-fluorescein labeled gene-specific TaqMan probe, and a predetermined, optimum concentration of gene specific forward and reverse primers (300–900 nM). PCR consisted of 40 cycles of 95°C denaturation (15 s) and 60°C annealing/extension (60 s). Thermal cycling and fluorescent monitoring were performed with an ABI 7700 sequence analyzer (PE Biosystems). The point at which the PCR product is first detected above a fixed threshold, termed the cycle threshold (Ct), was determined for each sample. The abundance of gene-specific transcripts in normal tissues was determined by comparison with a standard curve generated from the Ct values of known concentrations of plasmid DNA template encoding NY-SAR-35.

TaqMan primers were as follows. NY-SAR-35 5′ TaqMan primer: TGGTGCGATCAGCCTTATCC. NY-SAR-35 3′ TaqMan primer: CGGTTCGCTCCTCCAGAA. NY-SAR-35 TaqMan probe: TGTCTGCCCATTTATTGCCGCTCTCT.

Results

Frequency of Antibody Responses to CT Antigens in Sarcoma Patients.

Sera from 54 sarcoma patients‖ were tested by ELISA for the presence of antibodies to NY-ESO-1, SSX2, MAGE-A1, MAGE-A3, MAGE-A4, MAGE-A10, CT7, and CT10. Two patients, one with a malignant fibrous histocytoma (patient MFH) and the other with a fibrosarcoma (patient FS), had detectable serum antibodies to NY-ESO-1. All other serum/antigen combinations were negative. Fibrosarcoma tissue from the NY-ESO-1 seropositive patient, FS, was available for CT antigen typing by RT-PCR, and found to express 11/12 different CT antigen transcripts (NY-ESO-1, LAGE-1, MAGE-A1, -A3, -A4, -A10, BAGE, CT7, SSX1, -2, -4). Tissue from the NY-ESO-1 seropositive patient, MFH, was not available for CT antigen typing by RT-PCR.

Identification of Human Sarcoma Antigens by SEREX Analysis.

Serum samples from the two NY-ESO-1 seropositive patients (FS and MFH) were used to immunoscreen cDNA libraries prepared from the SW982 and SW1045 synovial sarcoma cell lines, both of which were shown to express eight or more known CT antigen transcripts (NY-ESO-1, MAGE-A1, -A3, -A4, -A10, BAGE, CT7, and SSX4; and NY-ESO-1, LAGE-1, MAGE-A1, -A3, -A4, -A10, BAGE, CT7, SSX-1, and SSX-4, respectively). Furthermore, sera from the fibrosarcoma patient was also used to immunoscreen a cDNA library derived from normal testis. In total, five independent SEREX immunoscreenings were performed, leading to the identification of 113 distinct sarcoma antigens, designated NY-SAR-1 through NY-SAR-113. These antigens are listed in Table 3, which is published as supporting information on the PNAS web site, www.pnas.org. The 113 SEREX-defined sarcoma antigens represent 91 known proteins and 22 uncharacterized gene products (novel, ESTs, KIAA series, FLJ series, MGC series, ORFs, DKFZ series). The nucleotide sequences of all uncharacterized gene products (NY-SAR-3, -10, -16, -22, -23, -24, -27, -29, -35, -41, -48, -71, -77, -79, -80, -84, -88, -91, -95, -97, -104, -105, and -113) have been deposited in the GenBank database (sequential accession nos. AY211909–AY211931). In terms of serum sources, 27 of the 113 antigens were identified by using sera from a MFH patient and 86 were identified with FS sera. Of the 113 antigens identified, 95 were unique to a particular cDNA library screening, and 18 antigens were identified in more than one library.

Seroepidemiology of SEREX-Defined Sarcoma Antigens.

There was little overlap between the antigens recognized by serum antibodies from the MFH and FS patients. Only three of the 113 antigens, NY-SAR-1/TMF1, NY-SAR-4/FH, NY-SAR-17/LAGE-1, were reactive with both these serum samples. Because serological reactivity to NY-ESO-1 was the criteria used in selecting sera for cDNA library screening, mutual immunoreactivity to the highly homologous (84% amino acid identity) NY-SAR-17/LAGE-1 antigen was expected, and in all likelihood is due to shared epitopes. To determine whether immune recognition of NY-SAR-1/TMF1and NY-SAR-4/FH was cancer-related, allogeneic sera samples obtained from 33 normal blood donors and 39 sarcoma patients (various histologies) were tested for reactivity to these two antigens by SADA. Sera from two normal individuals and three sarcoma patients reacted with NY-SAR-1/TMF1, suggesting the reactivity was unrelated to cancer, whereas NY-SAR-4/FH reactivity was restricted to cancer patients (see Discussion). Five of 39 (13%) sarcoma patients had serum antibodies to NY-SAR-4/FH. The seroreactive sarcoma patients included 2 of 10 osteosarcoma patients, 1 of 6 MFH patients, 1 of 2 FS patients, and 1 of 7 Ewing sarcoma patients.

The cDNA sequences encoding the 113 sarcoma antigens were also compared with sequences deposited in the cancer immunome database (http://www2.licr.org/CancerImmunomeDB). These comparisons showed that 39 of the 113 sarcoma antigens defined in this study (34%) were also identified through SEREX analysis of other tumor types (Table 1). To determine whether immune recognition of these 39 antigens was cancer-related, serum samples from normal individuals (n = 33) were tested for reactivity to these antigens. Twenty-three of the 39 antigens (59%) had a serological profile that was not restricted to cancer patients, whereas the remaining 16 antigens had a cancer-related serological profile, reacting only with sera from cancer patients (sarcoma patients and serum source of SEREX database entry), and not with sera from normal individuals. Fourteen of these 16 antigens reacted only with sera from a single sarcoma patient when tested for reactivity with additional allogeneic sarcoma sera (n = 39). The remaining 2 antigens, NY-SAR-17/LAGE-1 and NY-SAR-80/FLJ12577 (see Discussion), reacted with 2 of 39 and 3 of 39 sarcoma sera, respectively, and not with sera from normal individuals (n = 33).

Table 1.

Immunomic analysis of sarcoma/testis antigens: Reactivity with sera from sarcoma patients, patients with other forms of cancer, and normal individuals

NY-SAR-antigen	Gene identity (ugene cluster)	Cancer patient seroreactivity*	Normal seroreactivity
1	TMF1 (Hs.267632)	GC, BC, CC, SRC	2/33
2	STAU (Hs.6113)	PC, BC, SRC	3/30
3	KIAA1536 (Hs.156667)	BC, SRC	2/33
6	RHAMM (Hs.72550)	OC, SRC	1/33
7	PINCH (Hs.112378)	CC, GC, RC, BC, HN, ESO, AML, SRC	16/21
11	U2AF1RS2 (Hs.171909)	RC, HD, BC, GC, SRC	6/33
13	ACTN1 (Hs.119000)	BC, SRC	5/30
15	RBM6 (Hs.173993)	LC, SRC	0/33
16	FLJ12785 (Hs.192742)	TALL, SRC	0/33
17	LAGE-1a (Hs.87225)	BC, SRC	0/33
18	SSSCA1 (Hs.25723)	CC, SRC	0/33
19	HEF1 (Hs.80261)	RC, SRC	3/33
28	PPIL4 (Hs.11065)	BC, SRC	0/33
29	FLJ13441 (Hs.232146)	PN, SRC	6/33
31	AUANTIG (Hs.75528)	BC, GC, OC, SRC	2/33
39	PSMD4 (Hs.148495)	MEL, SRC	0/33
44	LGALS1 (Hs.227751)	RC, SRC	0/33
45	STIP1 (Hs.75612)	RC, SRC	4/33
47	MIF (Hs.73798)	MEL, SRC	0/33
57	GCN5L2 (Hs.101067)	PC, SRC	0/33
61	ZNF282 (Hs.58167)	RC, SRC	1/33
63	USP19 (Hs.301373)	OC, SRC	0/33
64	USP16 (Hs.99819)	PN, SRC	2/33
66	ROCK1 (Hs.17820)	RC, BC, CC, SRC	1/33
74	RANBP2 (Hs.199179)	BC, GL, BC, SRC	2/33
77	KIAA0992 (Hs.194431)	PC, SRC	4/15
80	FLJ12577 (Hs.87159)	GC, SRC	0/33
81	SDS3 (Hs.20104)	GC, SRC	4/16
82	NYCO45 (Hs.160881)	CC, SRC	0/33
89	SSX2 (Hs.289105)	BC, MEL, SRC	0/33
90	UACA (Hs.49753)	BC, ESO, SRC	4/25
93	NYBR15 (Hs.178175)	BC, SRC	1/12
98	OIP2 (Hs.274170)	BC, SRC	0/33
99	SSX3 (Hs.178749)	BC, MEL, SRC	2/30
101	RANBP2L1 (Hs.179825)	GL, BC, SRC	3/33
102	RBPJK (Hs.356806)	GC, RC, BC, MEL, SRC	1/16
103	Hsp40 (Hs.94)	HN, HCC, SRC	0/33
108	EIF4G (Hs.25732)	GC, SRC	5/27
112	PMSCL1 (Hs.91728)	CC, SRC	0/33

AML, acute myelogenous leukemia; BC, breast cancer; CC, colon cancer; GC, gastric cancer; GL, glioma; HCC, hepatocellular carcinoma; HN, head and neck cancer; LC, lung cancer; MEL, melanoma; OC, ovarian cancer; PC, prostate cancer; PN, pancreatic cancer; RC, renal cancer; SRC, sarcoma; TALL, T cell acute lymphocytic leukemia. 

^*Determined by sequence comparisons with the SEREX database (www2.licr.org/CancerImmunomeDB/). 

Expression Patterns of mRNA Encoding Serologically Defined Sarcoma/Testis Antigens in Normal and Malignant Tissues.

A preliminary in silico mRNA expression profile of all gene products identified in this study was carried out based on the tissue distribution of ESTs in the human EST database. Products with no EST matches, or those having EST matches limited to tumor tissue, fetal tissue, and/or less than three normal adult tissues were further examined by RT-PCR. Gene products with restricted EST profiles included five known CT antigens, NY-SAR-17/LAGE-1, NY-SAR-36/SSX1, NY-SAR-43/SSX4, NY-SAR-89/SSX2 and NY-SAR-99/SSX3, which are expressed exclusively in normal testis and a range of different tumor types (25–27). Nine other putative tissue-restricted antigens were identified, including five other known gene products, NY-SAR-12/nasopharyngeal specific protein 1(NESG1, ref. 28), NY-SAR-73/Protamine 2 (PRM2, ref. 29), NY-SAR-78/TSP-NY (unpublished data, UniGene cluster Hs.97643), NY-SAR-96/mitochondrial capsule selenoprotein (MCSP, ref. 30), and NY-SAR-110/NYD-SP14 (unpublished, Hs.98105), and four uncharacterized gene products, NY-SAR-35 (Hs.128580), NY-SAR-41 (Hs.166670), NY-SAR-92 (Hs.368781), and NY-SAR-97 (not clustered).

Two of the nine putative tissue restricted antigens, NY-SAR-73/PRM2 and NY-SAR-110/NYD-SP14, were ubiquitously expressed in a panel of 20 normal tissues as determined by RT-PCR (Table 2). The remaining seven genes were found to be expressed with frequencies ranging from 1 to 9 of 20 normal tissues. NY-SAR-35 and NY-SAR-78 were both testis-specific. The mRNA expression profiles of NY-SAR-35 and NY-SAR-78 were then analyzed in various malignant tissues by RT-PCR. Transcripts encoding NY-SAR-78/TSP-NY were not detected in cancer. The tumor specimens examined included, lung cancer (0 of 9), colon cancer (0 of 9), breast cancer (0 of 18), renal cancer (0 of 11), esophageal cancer (0 of 12), ovarian cancer (0 of 14), melanoma (0 of 18), and sarcoma (0 of 8). Thus, although NY-SAR-78/TSP-NY is a “virtual CT antigen,” with 100% identity with ESTs derived from prostate cancer and leukemia, its expression in cancer could not be verified in our RT-PCR series.

Table 2.

Analysis of mRNA expression by RT-PCR of 9 of the 113 sarcoma/testis antigens

Tissue	NY-SAR antigen*
	12	35	41	73	78	92	96	97	110
Brain	−	−	−	+	−	−	−	−	+
Kidney	−	−	−	+	−	−	−	−	+
Liver	−	−	−	+	−	−	−	−	+
Pancreas	−	−	−	+	−	−	−	−	+
Placenta	+	−	−	+	−	−	−	−	+
Testis	+	+	+	+	+	+	+	+	+
Fetal brain	−	−	+	+	−	−	+	+	+
Small intestine	−	−	−	+	−	−	−	−	+
Heart	−	−	−	+	−	−	−	−	+
Prostate	−	−	−	+	−	−	+	+	+
Adrenal	−	−	−	+	−	−	+	+	+
Spleen	+	−	−	+	−	+	+	+	+
Colon	+	−	+	+	−	−	−	−	+
Stomach	−	−	−	+	−	−	−	−	+
Lung	+	−	+	+	−	−	−	+	+
Bladder	−	−	+	+	−	−	−	+	+
Ovary	+	−	+	+	−	−	−	+	+
Breast	−	−	−	+	−	−	−	+	+
Cervix	−	−	−	+	−	−	−	−	+
Skeletal muscle	−	−	−	+	−	−	−	−	+
Total no. of positive tissues	6/20	1/20	6/20	20/20	1/20	2/20	5/20	9/20	20/20

^*Unigene clusters: NY-SAR-12, Hs.158450; NY-SAR-35, Hs.128580; NY-SAR-41, Hs.166670; NY-SAR-73, Hs.2324; NY-SAR-78, Hs.97643; NY-SAR-92, Hs.368781; NY-SAR-96, Hs.111850; NY-SAR-97, Hs.128836; NY-SAR-110, Hs.98105. 

NY-SAR-35 mRNA was detected in a variety of tumor specimens, such as melanoma (1 of 16 specimens), sarcoma (2 of 26 specimens), lung cancer (5 of 29 specimens), breast cancer (3 of 13 specimens), bladder cancer (5 of 12 specimens), esophageal cancer (1 of 12 specimens) and ovarian cancer (1 of 12 specimens). NY-SAR-35 mRNA was not detected in colon cancer (n = 9) or renal cancer (n = 8). The CT-restricted expression profile of NY-SAR-35 was confirmed by real time quantitative RT-PCR at 40 amplification cycles (Fig. 1). In these studies, NY-SAR-35 was expressed in normal testis at a level corresponding to 83.2 ag (1 ag = 10⁻¹⁸ g), which was more than 1,000 times the level detected in the remaining 15 normal tissues. In two of nine non-small cell lung cancer specimens tested, NY-SAR-35 was expressed at levels that were 0.13 and 0.15 times the level detected in normal testis. Thus, on the basis of its immunogenicity in a fibrosarcoma patient, and its restricted mRNA expression profile, NY-SAR-35 can be considered a novel CT antigen. In conformity with the proposed nomenclature for CT antigens (31), NY-SAR-35 is designated CT-20.

Figure 1.

Quantitative real-time RT-PCR analysis of NY-SAR-35 mRNA in various normal tissues and non-small cell lung cancer specimens. NY-SAR-35 was expressed in normal testis (83.2 ag) at a level that was >1,000 times the level detected in all other normal tissues. In two of nine cases of non-small cell lung cancer examined, the level of NY-SAR-35 expression was equivalent to 0.15 (12.5 ag) and 0.13 (10.8 ag) times the level detected in normal testis, or ≈100 times the level detected in normal tissues.

The NY-SAR-35 Gene, Transcript, Putative Protein, and Orthologous Gene.

Analysis of the human genome databases (www.ncbi.nlm.nih.gov/genome and www.celera.com) mapped NY-SAR-35 to Xq28. The NY-SAR-35 gene is ≈44 kb in length, spanning six exons. Database analysis revealed no genomic sequences of high similarity, suggesting that it is a single copy gene with no additional family members. These results were supported by probing Southern blots of human genomic DNA with NY-SAR-35 cDNA (data not shown).

The present SEREX analysis provided four overlapping NY-SAR-35 cDNA clones, ranging from 677 to 767 bp in length, all with identical 3′ sequences. The NY-SAR-35 cDNA sequence was identical to three ESTs (GenBank accession nos. AA909915, AA906131, and AW593050) derived from the NFL_T_GBC_S1 mixed tissue (fetal lung, testis, germinal center B cell) cDNA library, and found in UniGene cluster Hs.128580, as well as 4 ESTs (GenBank accession nos. BC034320, AK098602, BG771667, and BI465380) derived from a testis cell line, and found in Unigene cluster Hs.375082. To obtain a full-length NY-SAR-35 transcript, 5′ RACE was undertaken, yielding 262 bp of additional 5′ DNA sequence. Thus, the full-length NY-SAR-35 transcript is ≈1,029 bp (GenBank accession no. AY211917), a size that is in agreement with the 1.1-kb hybridization signal seen in Northern blots of testis mRNA probed with NY-SAR-35 cDNA (data not shown).

The NY-SAR-35 transcript encodes an ORF of 255 aa (bp 68–895) with a predicted molecular mass of 29.2 kDa. It is identical to the hypothetical protein, XM098959, predicted from Genefinder analysis of human chromosome X sequences. The putative NY-SAR-35 protein has a signal peptide, a transmembrane domain, and a cysteine-rich trefoil/P-domain, found in several secreted proteins of the gastrointestinal tract (32). These data suggest that NY-SAR-35 is a secreted or membrane bound protein.

To identify a murine orthologue of NY-SAR-35, the putative human NY-SAR-35 protein sequence was used to search a translated nonredundant nucleotide database by using the TBLASTN tool of the NCBI (www.ncbi.nlm.nih.gov/blast/Blast.cgi). A hypothetical mouse protein, termed XP_150408, generated from a conceptual translation of the mouse X chromosome, was found to have 57% identity (49/85 aa) with NY-SAR-35. Using nucleotide primers corresponding to sequences encoding XP_150408, 5′ and 3′ RACE reactions were undertaken by using mouse testis cDNA. By combining 5′ and 3′ RACE products, a 1,202 bp cDNA was identified (GenBank accession no. AY214130). This cDNA encoded a putative full length mouse protein of 238 aa which is 41% identical to human NY-SAR-35, with conservation of the trefoil and transmembrane domains. This murine NY-SAR-35 (mNY-SAR-35) cDNA sequence was used to search mouse genome sequences (www.ncbi.nlm.nih.gov/genome/seq/MmBlast.html) yielding an identical genome sequence, NW 042622, from mouse chromosome X. Analysis of this sequence showed the mNY-SAR-35 gene is composed of ≈42,600 nucleotides and seven exons.

Discussion

Immunohistochemical analyses have shown that CT antigens, such as NY-ESO-1 and MAGE-A, are expressed in 80% of synovial sarcomas, and that the expression is often homogeneous throughout the tumor (21, 22). Because CT antigen expression rarely exceeds 40% of a given cancer type (17, 18, 20, 26), and is generally heterogenous (33, 34), this CT antigen expression profile in synovial sarcoma was extraordinary. To identify additional immunoreactive antigens in sarcoma, the present comprehensive SEREX analysis was undertaken by using cDNA libraries prepared from two synovial cell sarcoma cell lines, SW982 and SW1045, both of which express multiple CT antigens (e.g., NY-ESO-1, BAGE, MAGE-A3, CT7, and SSX4). A testis cDNA library was also immunoscreened to increase the probability of isolating unidentified CT antigens that are immunogenic in cancer patients. The presence of serum antibody to a panel of CT antigens was used to identify sarcoma patients who were actively mounting a humoral immune response to tumor-related antigens, and who could serve as serum sources for immunoscreening. Although CT antigen expression is frequent in sarcoma tissue (21, 22, 35, 36), we detected antibodies in only two of 54 sarcoma patients (4%) to just one of eight CT antigens (NY-ESO-1). In particular, none of the serum samples from the five synovial sarcoma patients had detectable anti-CT antigen reactivity, and were therefore not used for immunoscreening purposes. Interestingly, humoral immune responses to SSX antigens were detected in sarcoma patients by SADA, but were not detected by ELISA, suggesting SADA may be a more sensitive method compared with ELISA. This may be due to qualitative differences in the assays, such as the use of denatured (ELISA) versus native (SADA) recombinant proteins, or quantitative differences, such as the amount of recombinant protein binding to the plastic (ELISA) or nitrocellulose (SADA) support matrices.

In the current study, 113 distinct antigens were isolated from these three cDNA libraries following SEREX immunoscreening with sera from two NY-ESO-1 seroreactive sarcoma patients. Only 18 of 113 antigens (16%) were isolated from more than one cDNA library, thus underlining the importance of incorporating multiple cDNA libraries into large-scale SEREX analyses of the cancer immunome. In addition to the LAGE-1/NY-ESO-2 antigen (NY-SAR-17), which is highly homologous to NY-ESO-1, only two other antigens, NY-SAR-4/FH, and NY-SAR-80/FLJ12577, were recognized by antibodies present in multiple sarcoma patients (5 of 39 and 3 of 39 sarcoma patients, respectively), but not in normal individuals (n = 33). This serological response to NY-SAR-4/FH is of considerable interest given the recent finding that germ-line mutations in the FH gene are associated with a predisposition to uterine and cutaneous leiomyomata, and also renal cell carcinoma (37) and is a target of somatic mutation in sarcoma (38), suggesting that the immune response is directed against mutated epitopes. NY-SAR-80/FLJ12577 is an uncharacterized member of the Mo25 protein family, an evolutionarily conserved family of proteins with no known function. Analysis of the tissue distribution and frequency of EST sequences homologous to NY-SAR-80/FLJ12577 indicate widespread mRNA expression, with a preponderance of malignant tissue-derived homologous ESTs suggesting possible overexpression in cancer.

Overall, the relative infrequency of overlapping humoral immune responses among the population of sarcoma patients analyzed in the current study is contrary to previous findings for colon (23), breast (14), and renal cancers (15), in which a subset of antigens were mutually seroreactive in a cancer related manner. These results suggest that the immune response to sarcoma is either highly variable, or that distinct sarcoma histiotypes have distinct immunomes. Thirty-nine of the 113 sarcoma antigens (34%) defined in the current study were also identified through SEREX analysis of other tumor types. Twenty-three of these 39 antigens (59%) reacted with sera from normal individuals, indicating that the humoral immune response to a large proportion of SEREX-defined antigens is not related to cancer.

The current SEREX analysis of sarcoma led to the isolation of 12 tissue-restricted gene products, including six known testis-restricted antigens (NY-SAR-17/LAGE-1, NY-SAR-36/SSX1, NY-SAR-43/SSX4, NY-SAR-78/TSP-NY, NY-SAR-89/SSX2, and NY-SAR-99/SSX3). Four of these differentially expressed antigens (NY-SAR-35, NY-SAR-41, NY-SAR-92, and NY-SAR-97) are novel gene products, and the remaining two tissue-restricted antigens (NY-SAR-12/NESG1 and NY-SAR-96/MCSP) have not been previously studied in relation to cancer. Their tissue restricted expression profile and immunogenicity indicate that these 12 antigens should be further analyzed with regard to their immunotherapeutic potential. Of particular interest is NY-SAR-35, which represents a recently defined CT antigen apparently expressed exclusively in normal testis, as well as melanoma, sarcoma, lung cancer, esophageal cancer, ovarian cancer, and breast cancer. Like several other CT antigens, NY-SAR-35, maps to chromosome X, but in contrast to other CT antigens (e.g., MAGE, NY-ESO-1, GAGE, and SSX), NY-SAR-35 is not a member of a multigene family. The presence of a signal sequence, a putative transmembrane domain, and a Trefoil domain suggests that NY-SAR-35 may be an extracellular or plasma membrane associated protein. Thus, in addition to being a potential cancer vaccine target, NY-SAR-35 may also be a target for of therapeutic antibodies.

Abbreviations

SEREX

serological analysis of recombinant cDNA expression libraries

CT

cancer/testis