Synovial sarcoma is a clinically and morphologically well defined soft-tissue tumor that occurs predominantly in the extremities of adolescents and young adults. It tends to arise in the vicinity of large joints, especially the knee8. This tumor was designated synovial sarcoma because of its anatomical location and its histological resemblance to normal synovial tissue. However, we do not believe that there is any convincing evidence that synovial sarcoma always originates from synovial tissue. It is possible that the lesion sometimes arises from primitive mesenchymal cells rather than from preformed synovial cells; therefore, the development of synovial sarcoma is no longer thought to require the presence of preexisting synovial tissue. To the best of our knowledge, there have been no reports of synovial sarcoma arising in bone.
The reciprocal translocation t(X;18)(p11.2;q11.2) was noted in all twenty-two reported cases of synovial sarcoma that have been studied cytogenetically5,18. This translocation is considered to be specific to this tumor. Cloning of the DNA adjacent to the breakpoints of the translocation t(X;18) has shown2 that this translocation results in fusion of the SYT gene at 18q11.2 to either of two genes, SSX1 or SSX2, at Xp11.2. This rearranged configuration of DNA is transcribed to messenger RNA and is called a chimeric SYT-SSX transcript. A chimeric SYT-SSX transcript does not occur normally and thus provides a specific marker for synovial sarcoma that can be detected with reverse transcriptase-polymerase chain-reaction testing.
We describe the case of a patient who had synovial sarcoma of bone that was verified by the detection of a chimeric SYT-SSX transcript.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Funds were received in total or partial support of the research or clinical study presented in this article. The funding sources were the Vehicle Racing Commemorative Foundation, Kanazawa Medical University Project (H98-2), the Ministry of Health and Welfare, and the Ministry of Education, Science, and Culture of Japan.
†Divisions of Orthopaedic Surgery (H. H., K. I., and S. Y.) and Clinical Pathology (K. Y.), Department of Clinical Research, Sapporo National Hospital, Sapporo 003, Japan.
‡Department of Pathology, Kanazawa Medical University Hospital, Uchinada 920-02, Japan.
§Department of Pathology, Hokkaido University School of Medicine, Sapporo 060, Japan.
A sixty-seven-year-old man had a four-month history of pain in the right wrist when he was first seen by us. Physical examination revealed moderate swelling and a restricted range of motion of the wrist (30 degrees of flexion, 15 degrees of extension, -5 degrees of radial deviation, 15 degrees of ulnar deviation, 50 degrees of pronation, and 30 degrees of supination). The pertinent laboratory data, including a complete blood-cell count and blood-chemistry levels, were normal except for a slight increase in the number of white blood cells (9600 per cubic millimeter [9.6 x 109 per liter]; normal value, 3900 to 9200 per cubic millimeter [3.9 to 9.2 x 109 per liter]) and in the level of C-reactive protein (3.3 milligrams per deciliter [0.33 milligram per liter]; normal value, less than 0.1 milligram per deciliter [0.01 milligram per liter]). A roentgenogram showed a radiolucent lesion located centrally in the medullary bone of the juxta-articular and metaphyseal portions of the radius. The indistinct margin between the tumor and the normal bone revealed the invasive nature of the tumor (Fig. 1). Although magnetic resonance imaging showed expansion of the tumor to the surrounding soft tissue, no adjacent soft-tissue component was apparent (Fig. 2). No abnormal uptake of technetium-99m methylene diphosphonate was found except in the area of the lesion of the right radius. An incisional biopsy was performed through the dorsal aspect of the tumor, and two specimens, each approximately five by five by five millimeters in size, were obtained. The pathological diagnosis was spindle-cell sarcoma. Preoperative neoadjuvant chemotherapy, including intra-arterial infusion of cis-diammine-dichloroplatinum (100 milligrams) and systemic administration of pirarubicin (two doses of forty milligrams each), was given. A wide resection was performed; the resected specimen, which included the tumor and adjacent normal bone, was nine centimeters long. The radius was reconstructed with a fibular graft. Eight months after the resection, pulmonary metastases were detected on plain roentgenograms. A skip metastasis occurred in the proximal part of the radius ten months later, and it was treated with irradiation.
Although no treatment was instituted for the pulmonary metastases, the patient was alive at the latest follow-up evaluation (two years and five months after the initial operation).
Analysis of the Specimens
The specimens were fixed in 10 percent (weight per volume) neutral buffered formalin, embedded in paraffin, and stained with hematoxylin and eosin, periodic acid-Schiff, alcian blue, and silver impregnation for reticulin fibers. Immunohistochemical staining with labeled streptavidin-biotin peroxidase and alkaline phosphatase was performed. The following antibodies were used (with the appropriate dilution, as determined in previous experiments): anti-vimentin (010050; Bioscience Products AG, Emmenbruücke, Switzerland), wide spectrum-screening anti-keratin (Z622; Dako, Copenhagen, Denmark), anti-epithelial keratin AE1 and AE3 (69-145; ICN, Costa Mesa, California), anti-epithelial membrane antigen (clone E29, M613; Dako), anti-S-100 (Z311; Dako), and anti-muscle actin (clone HHF35, M635; Dako). For electron microscopic examination, the biopsy specimens were fixed in 2 percent (weight per volume) glutaraldehyde, postfixed with 1 percent (weight per volume) osmium tetroxide, and embedded in Epon 812. Ultrathin sections were prepared, stained with uranyl acetate and lead citrate, and examined with an electron microscope (model H-800; Hitachi, Tokyo, Japan).
For reverse transcriptase-polymerase chain-reaction studies, total RNA was extracted from the frozen biopsy specimen and from peripheral lymphocytes of our patient with use of Isogen (Nippon Gene, Tokyo, Japan). A biphasic synovial sarcoma was included in the assays as a control specimen. One microgram of RNA was reverse-transcribed with random hexamers and the GeneAmp RNA-polymerase chain-reaction kit (Perkin Elmer, Foster City, California) in accordance with the manufacturer's protocol. The resulting cDNA was amplified by polymerase chain reaction with the use of SYT (5'-CAACAGCAAGATGCATACCA-3') primers as well as SSX (5'-CACTTGCTATGCACCTGATG-3') primers2. The RNA from the tumor in our patient was also amplified without reverse transcription to serve as a negative control. The amplification was performed in a final volume of 100 microliters with AmpliTaq polymerase (Perkin Elmer) and a thermal cycler (Perkin Elmer). We performed thirty amplification cycles, each consisting of one minute at 93 degrees Celsius (denaturation), one minute at 55 degrees Celsius (annealing), and one minute at 72 degrees Celsius (extension). The amplified products were separated by electrophoresis in 3 percent (weight per volume) agarose gel and stained with ethidium bromide. Subcloning of the polymerase chain reaction product was performed with an original TA cloning kit (Invitrogen, San Diego, California) and sequenced with the 4000LS long-read IR DNA sequencing system (LI-COR, Lincoln, Nebraska).
Pathological Findings
Gross examination of a cross section of the specimen revealed a white-to-yellow tumor located in the medullary cavity with central hemorrhagic necrosis (Fig. 3-A). Although both infiltration of the tumor into the soft tissue and periosteal reaction on the radial side of the radius were noted, there was no evidence of tumor nodules in the soft tissue.
Histologically, the biopsy specimen was composed of fascicles of spindle cells (Fig. 3-B). The nuclei of the spindle cells were short, spindle or ovoid-shaped, rather plump, and dark-staining, with one or two small nucleoli (Fig. 3-C). No distinct epithelial structures were found after either staining with hematoxylin and eosin or silver impregnation for reticulin fibers. Staining with periodic acid-Schiff and alcian blue did not show any secretory material. Infiltration of mast cells was noted. There were approximately eighteen mitoses per ten high-power fields. The material that was obtained at the time of the wide resection was composed of fascicles of spindle cells arranged in a monotonous fashion. The nuclei of these cells were dark-staining but slender, much like those of a fibrosarcoma. A herringbone pattern, similar to that associated with fibrosarcoma, was noticed as well. No differentiation to epithelial tissue was found.
Immunohistochemically, most of the tumor cells stained positively for vimentin (which is a marker for mesenchymal cells), and a small number stained positively for anti-epithelial membrane antigen (which represents epithelial differentiation) (Fig. 3-D). The cells stained negatively for the other antibodies, including those for cytokeratin. Electron microscopy revealed short fusiform cells with irregularly shaped nuclei. Intercellular junctions occasionally were noted, and intermediate filaments were seen in the cytoplasm (Fig. 4). The rough endoplasmic reticulum was scanty and, in some instances, dilated. Intercellular spaces were not found.
Detection of Chimeric SYT-SSX Transcript
RNA from the tumor in our patient as well as from a biphasic synovial sarcoma yielded a major, 585-base-pair DNA fragment on reverse transcriptase-polymerase chain-reaction testing (Fig. 5). In contrast, reverse transcription of the RNA from peripheral lymphocytes did not promote DNA amplification with the SYT-specific and SSX-specific primers. Sequencing of the amplified DNA revealed the same configuration of base pairs surrounding the junction between the SYT and SSX1 genes as previously described by Crew et al.3.
The tumor in our patient had several notable histopathological features, including fascicles of spindle cells, infiltration of mast cells, and positive staining for anti-epithelial membrane antigen. Although the most likely spindle-cell tumor with these features is monophasic fibrous synovial sarcoma, it was necessary to distinguish the lesion from several other types of neoplasms, including osteosarcoma, malignant fibrous histiocytoma, leiomyosarcoma, fibrosarcoma, and malignant peripheral nerve-sheath tumor. Osteosarcoma was ruled out because it occurs at a younger age and produces neoplastic osteoid, which was not found in our patient. The storiform-pleomorphic type of malignant fibrous histiocytoma was also ruled out because the tumor did not demonstrate a storiform pattern or pleomorphism on histological analysis. Although the tumor was histologically similar to leiomyosarcoma of bone, its nuclei were plumper than those of leiomyosarcoma. In addition, the tumor cells did not stain positively for anti-muscle actin and no myofilaments were observed ultrastructurally. The exclusion of fibrosarcoma and malignant peripheral nerve-sheath tumor was more difficult. Spindle-cell tumors that stain positively for anti-epithelial membrane antigen but negatively for S-100 are known to be synovial sarcomas rather than fibrosarcomas or malignant peripheral nerve-sheath tumors. In addition, although the tumor cells did not stain positively for cytokeratin, it has been reported that some synovial sarcomas stain positively for anti-epithelial membrane antigen but not for cytokeratin1. Infiltration of mast cells is also a typical feature of synovial sarcoma8.
Thus, the tumor in our patient is believed to have been a primary synovial sarcoma of bone in that it was a central medullary, slightly expansile tumor that demonstrated no evidence, on either imaging or pathological studies, of being a contiguous soft-tissue tumor that had invaded bone. No other primary soft-tissue tumor focus was identified to suggest that the lesion represented an osseous metastasis. This view is supported by the fact that, at two years and five months after the resection, no soft-tissue tumor had developed.
Although synovial sarcoma occurs predominantly in the extremities, the recent literature contains reports of some tumors that have arisen at other sites, including the lung19, the heart13, the pleural cavity10, and the small intestinal mesentery12. The diagnosis of these sarcomas has been based on histological findings, immunohistochemical studies, and, in a few instances, cytogenetic features16.
Recently, many types of sarcoma have been characterized by specific chromosomal translocations that are likely to be of etiological importance. Cloning of the breakpoints revealed that these translocations result in the production of novel, tumor-specific chimeric transcripts, such as EWS-FLI1, EWS-ERG, and EWS-ETV1 from t(11;22), t(21;22), and t(7;22) of peripheral primitive neuroectodermal tumors6,15,21; EWS-ATF1 from t(12;22) of clear-cell sarcoma22; TLS-CHOP from t(12;16) of myxoid liposarcoma4; PAX3-FKHR and PAX7-FKHR from t(2;13) and t(1;13) of alveolar rhabdomyosarcoma11; and EWS-WT1 from t(11;22) of desmoplastic small-round-cell tumor17. Clinically, these chimeric transcripts can serve as tumor-specific markers that can be detected with use of reverse transcriptase-polymerase chain-reaction methods20.
All twenty-two cases of synovial sarcoma that have been analyzed cytogenetically5,18 have demonstrated the reciprocal translocation t(X;18)(p11.2;q11.2), which is thought to be the primary cytogenetic abnormality and specific to synovial sarcoma5. The t(X;18) translocation results in fusion of the SYT gene at 18q11 to the SSX gene at Xp11.2, providing a marker for synovial sarcoma2. Subcloning and sequencing of the reverse transcriptase-polymerase chain-reaction products revealed that there are at least two copies of SSX—SSX1 and SSX2—that encode closely related (81 percent identical) proteins of 188 amino acids9. SSX1 and SSX2 were contained in the yeast artificial chromosomes, designated as OATL1 and OATL2, respectively, which are known to contain two distinct breakpoints of synovial sarcoma7. The two breakpoints may correspond to the histological phenotypes of the tumors—that is, those with a breakpoint near the OATL1 region are biphasic, whereas those with a breakpoint near the OATL2 region are primarily monophasic14. However, this relationship has not been verified in other studies3. Clinically, the detection of SYT-SSX1 or SYT-SSX2 with reverse transcriptase-polymerase chain-reaction testing has opened a new avenue for the molecular diagnosis of synovial sarcoma. In the present study, the chimeric transcript SYT-SSX1 was detected with reverse transcriptase-polymerase chain-reaction analysis and verified with nucleotide sequencing. This evidence certainly supports the histological evidence that this tumor is synovial sarcoma.
It is interesting that the histological features of the resected material resembled those of fibrosarcoma. On the basis of analysis of only the resected material and without the detection of the chimeric transcripts, this tumor could have been misdiagnosed as fibrosarcoma. This is particularly true because synovial sarcoma of bone has not been described previously, to our knowledge. Our findings suggest that other cases of synovial sarcoma of bone may also have been misdiagnosed in the past. We believe that molecular biological techniques are not only useful but may be essential for the diagnosis of certain tumors, including synovial sarcoma of bone.