| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Molecular Biology and Genetics |
Laboratory of Molecular Medicine, Human Genome Center, Institute of Medical Science, The University of Tokyo, Tokyo 108-8639, Japan [S. N., T. K., Yu. N.]; Laboratory for Medical Informatics, SNP Research Center, RIKEN (Institute of Physical and Chemical Research), Tokyo 108-8639, Japan [Tat. T.]; Department of Tissue Regeneration, Institute for Frontier Medical Sciences [S. N., T. H., J. T.], and Departments of Orthopaedic Surgery [T. H., To. N., Ta. N.], Pathology [Ya. N], and Surgery and Surgical Basic Science [S. N., M. I.], Graduate School of Medicine, College of Medical Technology [Tad. T.], Kyoto University, Kyoto 606-8507, Japan; Department of Orthopedic Surgery, Osaka Medical Center for Cancer and Cardiovascular Diseases [N. A.], Osaka 537-8511, Japan; Department of Orthopedic Surgery, Kyoto Prefectural University of Medicine [K. K.], Kyoto 602-0841, Japan
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Among several histological findings, the most distinctive feature of SS is epithelial differentiation. On the basis of the presence or absence of an epithelial component, SS is classified into two major subtypes: biphasic, which is composed of distinct epithelial and spindle tumor cells; and monophasic, which is composed of fibrosarcoma-like spindle tumor cells and no detectable epithelial components (1) . However, because the proportion and features of the epithelial component vary significantly among biphasic tumors, transition from one to the other subtype may be gradual rather than abrupt.
Although the histogenesis of SS remains unclear, molecular analysis of the mechanisms underlying tumorigenesis of SS progressed markedly with the discovery that a SYT-SSX fusion gene is a SS-specific genetic alteration (11 , 12) . The SYT-SSX fusion product contains both activator and repressor elements for transcription, and the net result seems to be transcriptional repression of certain genes in precursors of SS (13 , 14) . Target genes of the SYT-SSX protein seem to be related not only to oncogenesis, but also to epithelial differentiation, because the SYT-SSX2 fusion protein has been found only in tumors with monophasic morphology, whereas biphasic tumors contain only the SYT-SSX1 subtype (15 , 16) . Therefore, identifying the target genes for the SYT-SSX protein will require determination of what the precursor cells are.
In this study we analyzed the gene expression profiles of a panel of SS cases, using a genome-wide cDNA microarray containing 23,040 genes. This approach has been useful for clarifying molecular mechanisms that underlie disease progression, for identifying novel cancer-related genes, and for classifying human cancers at the molecular level (17, 18, 19, 20, 21, 22) . In addition to SS, we analyzed gene expression profiles among four other types of STS: MFH, LMS, PLS and DLS, and MPNST. These tumors sometimes exhibit histological features that closely resemble those of SS. Although differential diagnosis among these other tumors and SS is now feasible through analysis for the SYT-SSX fusion gene (23) , we judged that comparison of expression profiles among the various STSs should provide information about the histogenesis of SS.
We report here that SS most closely resembles MPNST in terms of gene expression profiles, sharing differential expression of several genes characteristic of neural crest-derived cells. This suggests a neuroectodermal origin of SS. Our results also indicate that SS can be classified into two subgroups irrespective of the histological classification. We also identified several genes expressed commonly in SS, whose products should be suitable targets for development of novel therapeutic drugs.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
RNA Preparation and T7-based RNA Amplification.
Total RNAs were extracted from each frozen specimen and from human MSCs purchased from BioWhittaker, Inc. (Walkersville, MD) as a universal control, by use of TRIzol reagent (Life Technologies, Rockville, MD) according to the manufacturer’s instructions. After treatment with DNase I (Nippon Gene, Osaka, Japan), 10 µg of total RNA from the tumors and MSC were amplified using an Ampliscribe T7 Transcription Kit (Epicentre Technologies, Madison, WI), and 5 µg of the amplified RNAs were labeled with Cy5-dCTP and Cy3-dCTP, respectively, as described previously (24)
. Total RNA was also prepared from three SS cell lines (OUSS, HS-SY-II, and YaFuSS) to reinforce the microarray data. All samples used in this study were analyzed for SYT-SSX fusion transcripts by RT-PCR using the following primers: for the SYT-SSX1 gene, 5'-CAACAGCAAGATGCATACCA-3' and 5'-GGTGCAGTTGTTTCCCATCG-3'; for the SYT-SSX2 gene, 5'-CAACAGCAAGATGCATACCA-3' and 5'-GGCACAGCTCTTTC CCATCA-3'.
cDNA Microarray.
We fabricated a "genome-wide" cDNA microarray with 23,040 cDNAs selected from the UniGene database (build no. 131) of the National Center for Biotechnology Information. Construction of the microarray, procedures for hybridization and washing, and photometric quantification of signal intensities of each spot were performed as described previously (24)
, except that all hybridization and washing procedures were carried out with an Automated Slide Processor (Amersham Bioscience, Buckinghamshire, United Kingdom). Each slide contained 52 housekeeping genes, and the Cy5/Cy3 ratio for each gene’s expression was adjusted so that the averaged Cy5/Cy3 ratio of the panel of housekeeping genes was 1.0. We assigned a cutoff value to each microarray slide, using variance analysis. If both Cy3 and Cy5 signal intensities were lower than the cutoff values, the expression of the corresponding gene in that sample was assessed as low or absent. For other genes, we calculated Cy5/Cy3 as a relative expression ratio.
Cluster Analysis of 47 STS Cases According to Gene Expression Profiles.
We applied a hierarchical clustering method to both genes and samples. To obtain reproducible clusters for classification of the 47 STSs, we selected 1204 genes for which data were present in 90% of the experiments and that had expression ratios that varied by SDs >1.0. The analysis was performed using web-available software ("Cluster" and "TreeView") written by M. Eisen.4
Before the clustering algorithm was applied, the fluorescence ratio for each spot was first log-transformed (log2), and then the data for each sample were median-centered to remove experimental biases.
Identification of Up-Regulated Genes Common to SS.
The relative expression ratio of each gene (Cy5/Cy3 intensity ratio) was classified into one of four categories: (A) up-regulated (expression ratio >2.0); (B) down-regulated (expression ratio <0.5); (C) unchanged (expression ratio between 0.5 and 2.0); and (D) not expressed (or slight expression but under the cutoff level for detection). We used these categories to detect a set of genes for which changes in the expression ratios were common among samples as well as specific to a certain subgroup in accordance with Schena et al. (25)
. To detect candidate genes that were commonly up-regulated in SS, the overall expression patterns of 23,040 genes were first screened to select genes with expression ratios >3.0 that were present in >75% of the SS cases categorized as A, B, or C. From a list of selected genes, we then chose those showing slight or no expression (category D) in >80% of non-SS cases.
Cluster Analysis of 13 SS Cases According to Gene Expression Profiles.
To clarify the nature of the histological heterogeneity within the SS group, we focused on differences in expression patterns of 23,040 genes among the 13 original SS cases. From the overall expression profiles of the SS group, we chose 1405 genes for which data were present in 75% of the cases and that had expression ratios that varied by SDs >1.0. Clustering analysis was performed in the manner described above.
Identification of Candidate Genes for Discriminating between SS Subclasses.
We selected 7067 genes for which data were present in >10 of 13 SS cases, and the mean (µ) and SD (
) were calculated from the relative expression ratios of each gene in one of the two subclasses. A discrimination score (DS) for each gene was defined as: DS = µ1 - µ2/(1 + 2), where the subscripts refer to the same groups (26)
. A large DS indicates that a gene’s expression varies greatly between the two groups but little within its own group. We invoked a permutation test to calculate the ability of individual genes to distinguish between the two subclasses; samples were randomly permutated into each of the two groups 10,000 times. Because the DS dataset of each gene showed a normal distribution, a P for the user-defined grouping was calculated. If the P was <0.001, the gene was considered to have the power to distinguish the two groups.
Semiquantitative and Real-Time Quantitative RT-PCR.
A 3-µg aliquot of total RNA from each tissue sample was reverse-transcribed for single-stranded cDNAs, using oligo(dT)12–18 primer and Superscript II (Invitrogen, Carlsbad, CA). Semiquantitative RT-PCR was carried out with the same gene-specific primers as those prepared for constructing our cDNA microarray or with a ß2-microglobulin-specific primer as an internal control as described previously (27)
. The primer sequences are listed in Table 1
. PCR reactions were optimized for the number of cycles to ensure product intensity within the linear phase of amplification. Real-time quantitative RT-PCR (TaqMan PCR; Applied Biosystems, Foster City, CA) was performed with the ABI Prism 7700 Sequence Detection system (Applied Biosystems) as described previously (27)
. The primers and TaqMan probes are shown in Table 2
. The expression levels of each candidate gene were corrected by that of ß2-microglobulin, and relative expression ratios (r) of each sample to MSC were calculated.
|
|
 |
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. MFH, LMS, DLS, and PLS were scattered into several different clusters and failed to compose a disease-specific cluster. On the other hand, SS cases showed a distinct cluster along with MPNST. Four tumors with biphasic features (SS190, SS334, SS487, and SS582) were clustered into one group, but nine tumors with monophasic features failed to make one cluster. Two SS cases (SS213 and SS438) constituted one subcluster together with a case of MPNST (MPNST248), and one case of MPNST (MPNST558) fell into the major cluster of SS. These data suggested that SS and MPNST are closely related diseases in terms of gene expression, although both are regarded as clearly distinct entities from a histological point of view.
|
. Among them, frizzled homologue 10 (C0671) and one EST (A8647) were up-regulated specifically in SS; the remaining 24 genes were also expressed in MPNST, at the same or lower levels. The results of semiquantitative RT-PCR experiments confirmed the specific expression of these genes in SS or SS and MPNST (Fig. 2)
. In addition, expression of all 26 genes was detected in three SS cell lines, indicating that this activity was intrinsic to SS cells and not induced by the in vivo environment. In most of the known genes that were up-regulated commonly in SS and MPNST, the proposed function and distribution of expression were related to neural tissues, e.g., EphA4, ephrin-B3, and endothelin 3. Moreover, SS expressed additional markers of neural differentiation, such as neurofilament, neuron-specific protein, and fibroblast growth factor 18, which were expressed in MPNST at the same or slightly lower levels. These data suggested that the cellular precursor of SS was very similar to that of MPNST, possibly a cell derived from the neural crest.
|
). As shown in the hierarchical clustering analysis for all tumors (Fig. 1)
, four biphasic tumors (SS190, SS334, SS487, and SS582) were again clustered closely, whereas monophasic tumors were divided into two groups. Therefore, the degree of epithelial differentiation may contribute to this subclassification. However, because a tumor (SS646) that showed minimal epithelial differentiation and carried the SYT-SSX2 fusion gene (data not shown) was classified closely to one of the clusters of biphasic type, it is likely that factors other than epithelial differentiation also contribute to this subclassification.
|
. We applied forward stepwise discriminant analysis using 12 candidate genes and established a predictive formula with discriminant coefficients of 9 predictive genes that were further selected from the 12 genes and a constant value of 6.464 for subclassification of these learning cases (see "Materials and Methods" and Table 3
). To further verify the power of a set of the nine genes to assign a tumor to one subgroup or the other, their expression levels were also determined by real-time RT-PCR analysis in 15 additional SS cases in which SYT-SSX fusion transcripts were identified, SYT-SSX1 gene in 12 cases and SYT-SSX2 in 3 cases. Histologically, 10 of the 15 additional cases were classified into the monophasic subtype, and 5 into the biphasic subtype. By the predictive formula, the 15 test cases also clearly fell into either subclass A (4 cases) or subclass B (11 cases; Fig. 4
). All of the four cases in subclass A were monophasic, whereas all five biphasic cases were assigned to subclass B, confirming that the subclassification reflected the degree of epithelial differentiation. However, a monophasic case with the SYT-SSX2 gene in the test cases, which may correspond to SS646 in the learning cases, fell into subclass B, again suggesting that factors other than epithelial differentiation also contribute to this subclassification. These results indicated a potential for this set of genes to classify SS cases in terms of their biological properties.
|
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our cluster analyses demonstrated that SS also had a feature of neuronal differentiation, irrespective of the presence of an epithelial component. Among the genes commonly up-regulated in SS, several are known to be involved in migration or differentiation of neural crest cells. For example, interactions between ephrin-B3 and EphA4 are important for the proper migration of neural crest cells (34) , and endothelin 3 is essential for development of neural crest-derived cell lineages (35) . In addition to being the major extracellular matrix of cartilage, collagen IX also has a crucial function in migration of neural crest cells (36) ; retinoic acid signaling is also indispensable for this migration, and cellular retinoic acid-binding protein 1 is thought to be a general marker of neural crest cells (37) . Transduction of Wnt signals is involved in the genesis of neural crest cells (38) , and fibroblast growth factor 18 is expressed in such cells adjacent to primitive streaks (39) . All of these findings suggest that SS cells are derived from the neural crest. However, because the expression of some of these genes is not limited in neural tissues, our data are not conclusive for the origin of SS cells, and further investigation should be required.
A previous report demonstrated that both spindle and epithelial components of SS contained the same type of fusion transcripts (40 , 41) , leading to the hypothesis that SS is monoclonal in origin. The transition from uncommitted spindle cells to differentiated epithelial cells is likely to be induced by expression of a particular set of genes, although no clear mechanism has been reported. One of the confounding facts is that there are variations in the degree of epithelial differentiation in different tumor samples, a feature that renders clear-cut classification into one or the other of the two subclasses difficult. We therefore attempted to classify SS cases according to their gene expression patterns, irrespective of histological classification. As a result, although all of the biphasic cases were clearly clustered together, monophasic cases were divided into two subclasses. These data suggested the presence of other factors reflecting the differentiation stage of SS precursor cells or the concomitance of additional genetic alterations during tumorigenesis. We believe that putative subclassification, which could be predicted by a set of genes we selected, should be useful for understanding the histological heterogeneity of SS.
In contrast to the cluster formation of SS and MPNST, other histological types were not well clustered. Although MFH is the most common STS of elder adult life, there has been a long-standing debate whether MFH exists as an independent entity or represents an admixture of various mesenchymal and even nonmesenchymal tumors (1) . Our failure to create a distinct cluster of MFH indicated that MFHs are not only morphologically, but also genetically heterogeneous tumors, supporting the idea that MFHs under the current histological definition may include tumors of various types (42) . Liposarcomas in this study consisted of two rare subtypes, PLS and DLS. The definition of DLS is a neoplasm with well-differentiated liposarcoma juxtaposed to high-grade pleomorphic sarcoma, which usually resembles MFH (43) . Distinction of PLS and MFH has been a difficult problem for pathologists, especially for MFHs without a storiform pattern (44) . Our cluster analyses demonstrated that DLS and PLS shared gene expression profiles with some cases of MFH, which may explain the histopathological similarities. It is notable that three LMS cases (LMS181, LMS407, and LMS551), constituting a distinct group within a large cluster of MFHs, were diagnosed to be poorly differentiated types and clinically revealed very aggressive phenotypes with distant metastases (data not shown). Further analyses of the relationship between the expression profiles and clinicopathological features of tumors used in this study, including SS and MPNST, may provide insight for exploration of disease-specific genes and, hopefully, of potential genes for molecular target therapy.
|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported in part by Research for the Future Program Grant No. 00L01402 from the Japan Society for the Promotion of Science. 
2 To whom requests for reprints should be addressed, at Laboratory of Molecular Medicine, Human Genome Center, Institute of Medical Science, The University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108-8639, Japan. Phone: 81-3-5449-5372; Fax: 81-3-5449-5433; E-mail: yusuke{at}ims.u-tokyo.ac.jp
3 The abbreviations used are: STS, soft tissue sarcoma; SS, synovial sarcoma; MFH, malignant fibrous histiocytoma; LMS, leiomyosarcoma; PLS, pleomorphic liposarcoma, DLS; dedifferentiated liposarcoma; MPNST, malignant peripheral nerve sheath tumor; MSC, mesenchymal stem cell; RT-PCR, reverse transcription-PCR; EST, expressed sequence tag.
4 http://genome-www5.stanford.edu/MicroArray/SMD/restech.html.
Received 1/18/02. Accepted 8/19/02.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Proc. Natl. Acad. Sci. USA, 98: 3843-3848, 2001.This article has been cited by other articles:
|
|
 |
|
  I. Ray-Coquard, A. Le Cesne, J. S. Whelan, P. Schoffski, B. N. Bui, J. Verweij, S. Marreaud, M. van Glabbeke, P. Hogendoorn, and J.-Y. Blay A Phase II Study of Gefitinib for Patients with Advanced HER-1 Expressing Synovial Sarcoma Refractory to Doxorubicin-Containing Regimens Oncologist, April 1, 2008; 13(4): 467 - 473. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Barco, L. B. Hunt, A. L. Frump, C. B. Garcia, A. Benesh, R. L. Caldwell, and J. E. Eid The Synovial Sarcoma SYT-SSX2 Oncogene Remodels the Cytoskeleton through Activation of the Ephrin Pathway Mol. Biol. Cell, October 1, 2007; 18(10): 4003 - 4012. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Kohno, T. Okamoto, T. Ishibe, S. Nagayama, Y. Shima, K. Nishijo, K. R. Shibata, K. Fukiage, S. Otsuka, D. Uejima, et al. Expression of Claudin7 Is Tightly Associated with Epithelial Structures in Synovial Sarcomas and Regulated by an Ets Family Transcription Factor, ELF3 J. Biol. Chem., December 15, 2006; 281(50): 38941 - 38950. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Miller, F. Rangwala, J. Williams, P. Ackerman, S. Kong, A. G. Jegga, S. Kaiser, B. J. Aronow, S. Frahm, L. Kluwe, et al. Large-scale molecular comparison of human schwann cells to malignant peripheral nerve sheath tumor cell lines and tissues. Cancer Res., March 1, 2006; 66(5): 2584 - 2591. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Baird, S. Davis, C. R. Antonescu, U. L. Harper, R. L. Walker, Y. Chen, A. A. Glatfelter, P. H. Duray, and P. S. Meltzer Gene Expression Profiling of Human Sarcomas: Insights into Sarcoma Biology Cancer Res., October 15, 2005; 65(20): 9226 - 9235. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Terry, J. M. Lubieniecka, W. Kwan, S. Liu, and T. O. Nielsen Hsp90 Inhibitor 17-Allylamino-17-Demethoxygeldanamycin Prevents Synovial Sarcoma Proliferation via Apoptosis in In vitro Models Clin. Cancer Res., August 1, 2005; 11(15): 5631 - 5638. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Milyavsky, Y. Tabach, I. Shats, N. Erez, Y. Cohen, X. Tang, M. Kalis, I. Kogan, Y. Buganim, N. Goldfinger, et al. Transcriptional Programs following Genetic Alterations in p53, INK4A, and H-Ras Genes along Defined Stages of Malignant Transformation Cancer Res., June 1, 2005; 65(11): 4530 - 4543. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Togashi, T. Katagiri, S. Ashida, T. Fujioka, O. Maruyama, Y. Wakumoto, Y. Sakamoto, M. Fujime, Y. Kawachi, T. Shuin, et al. Hypoxia-Inducible Protein 2 (HIG2), a Novel Diagnostic Marker for Renal Cell Carcinoma and Potential Target for Molecular Therapy Cancer Res., June 1, 2005; 65(11): 4817 - 4826. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Ishibe, T. Nakayama, T. Okamoto, T. Aoyama, K. Nishijo, K. R. Shibata, Y. Shima, S. Nagayama, T. Katagiri, Y. Nakamura, et al. Disruption of Fibroblast Growth Factor Signal Pathway Inhibits the Growth of Synovial Sarcomas: Potential Application of Signal Inhibitors to Molecular Target Therapy Clin. Cancer Res., April 1, 2005; 11(7): 2702 - 2712. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Kan, Y. Shimada, F. Sato, T. Ito, K. Kondo, G. Watanabe, M. Maeda, S. Yamasaki, S. J. Meltzer, and M. Imamura Prediction of Lymph Node Metastasis with Use of Artificial Neural Networks Based on Gene Expression Profiles in Esophageal Squamous Cell Carcinoma Ann. Surg. Oncol., December 1, 2004; 11(12): 1070 - 1078. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Nakanishi, K. Yoshioka, S. Joyama, N. Araki, A. Myoui, S. Ishiguro, T. Ueda, H. Yoshikawa, and K. Itoh Interleukin-6/Soluble Interleukin-6 Receptor Signaling Attenuates Proliferation and Invasion, and Induces Morphological Changes of a Newly Established Pleomorphic Malignant Fibrous Histiocytoma Cell Line Am. J. Pathol., August 1, 2004; 165(2): 471 - 480. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. O. Nielsen, F. D. Hsu, J. X. O'Connell, C. B. Gilks, P. H.B. Sorensen, S. Linn, R. B. West, C. L. Liu, D. Botstein, P. O. Brown, et al. Tissue Microarray Validation of Epidermal Growth Factor Receptor and SALL2 in Synovial Sarcoma with Comparison to Tumors of Similar Histology Am. J. Pathol., October 1, 2003; 163(4): 1449 - 1456. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. H. Segal, P. Pavlidis, C. R. Antonescu, R. G. Maki, W. S. Noble, D. DeSantis, J. M. Woodruff, J. J. Lewis, M. F. Brennan, A. N. Houghton, et al. Classification and Subtype Prediction of Adult Soft Tissue Sarcoma by Functional Genomics Am. J. Pathol., August 1, 2003; 163(2): 691 - 700. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
) and 15 test cases (
). With the predictive formula determined by forward discriminant analysis in learning cases, 15 test cases were grouped into one of the putative subclasses. Horizontal bars indicate the median value for each group.